1 2013-11-07 Richard Biener <rguenther@suse.de>
3 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
4 (canonicalize_value): ... this. Also handle stripping of
6 (get_value, set_lattice_value, get_value_for_expr): Adjust.
7 * gimple-fold.c (canonicalize_constructor_val): Strip
9 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
11 2013-11-07 Richard Biener <rguenther@suse.de>
13 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
15 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
18 * config/i386/i386.md (push peepholer/splitter): Use Pmode
19 with stack_pointer_rtx.
21 2013-11-07 Bin Cheng <bin.cheng@arm.com>
23 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
24 using operand_equal_p.
26 2013-11-07 Bin Cheng <bin.cheng@arm.com>
28 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
29 * tree-affine.c (get_inner_reference_aff): Return base.
30 * tree-affine.h (get_inner_reference_aff): Change prototype.
32 2013-11-06 Tobias Burnus <burnus@net-b.de>
34 * doc/invoke.texi (Wdate-time): Fix typo.
36 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
38 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
39 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
40 fp_arith_reg_operand instead of arith_reg_operand.
42 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
44 * config/sh/sh.md (adddi3): Remove empty constraints.
45 Remove can_create_pseudo_p and arith_reg_operand check.
46 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
49 2013-11-06 Jeff Law <law@redhat.com>
50 Tom Tromey <tromey@redhat.com>
52 * gdbinit.in: Disable strict type checking.
54 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
56 * tree-pass.h (make_pass_live_range_shrinkage): New external.
57 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
58 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
59 (rest_of_handle_live_range_shrinkage): Ditto
60 (class pass_live_range_shrinkage): Ditto.
61 (pass_data_live_range_shrinkage): Ditto.
62 (make_pass_live_range_shrinkage): Ditto.
63 * sched-int.h (initialize_live_range_shrinkage): New prototype.
64 (finish_live_range_shrinkage): Ditto.
65 * sched-deps.c (create_insn_reg_set): Make void return value.
66 * passes.def: Add pass_live_range_shrinkage.
67 * ira.c (update_equiv_regs): Don't move if
68 flag_live_range_shrinkage.
69 * haifa-sched.c (live_range_shrinkage_p): New.
70 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
72 (rank_for_schedule): Add code for pressure relief through live
74 (schedule_insn): Print more debug info.
75 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
76 through live range shrinkage.
77 * doc/invoke.texi (-flive-range-shrinkage): New.
78 * common.opt (flive-range-shrinkage): New.
80 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
83 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
84 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
85 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
86 RTXes that return in AVX256 register.
88 2013-11-06 Richard Biener <rguenther@suse.de>
90 PR tree-optimization/58653
91 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
92 (prepare_initializers_chain): Adjust.
94 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
96 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
97 (gimple_in_transaction): Use bb_in_transaction.
98 * basic-block.h (bb_in_transaction): Relocate here and rename.
99 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
101 2013-11-06 Richard Biener <rguenther@suse.de>
103 * tree.c (drop_tree_overflow): New function.
104 * tree.h (drop_tree_overflow): Declare.
105 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
106 * tree-vrp.c (range_int_cst_singleton_p): Use
107 is_overflow_infinity instead of testing TREE_OVERFLOW.
108 (extract_range_from_assert): Likewise.
109 (zero_nonzero_bits_from_vr): Likewise.
110 (extract_range_basic): Likewise.
111 (register_new_assert_for): Use drop_tree_overflow.
112 (vrp_visit_phi_node): Likewise.
114 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
116 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
117 checking for leaf functions without dynamic stack allocation.
118 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
119 (ia64_expand_prologue): Likewise.
120 * config/mips/mips.c (mips_expand_prologue): Likewise.
121 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
122 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
123 (sparc_flat_expand_prologue): Likewise.
125 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
127 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
128 (__ST3_LANE_FUNC): Likewise.
129 (__ST4_LANE_FUNC): Likewise.
131 2013-11-06 Nick Clifton <nickc@redhat.com>
133 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
134 name returned by msp430_mcu_name.
135 (LIB_SPEC): If a -T option has not been specified then set a
136 default, mcu-specific, linker script.
137 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
138 * config/msp430/msp430.c (msp430x_names): Likewise.
139 Alpha sort the names for ease of comparison.
140 (msp430_mcu_name): New function: Returns a string suitable for
141 use as a C preprocessor symbol based upon the name of the MCU
143 (msp430_option_override): Accept msp430x and msp430xv2 as generic
145 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
147 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
148 unless it is the last part of a spec substring.
149 * doc/invoke.texi (Spec Files): Document space insertion
152 2013-11-06 Christian Bruel <christian.bruel@st.com>
154 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
155 Factorize probabilities, Use adjust_address instead of
156 adjust_automodify_address when possible. Enable for optimize.
157 (sh_expand_strlen): New function.
158 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
159 * config/sh/sh.md (strlensi): New pattern.
160 (UNSPEC_BUILTIN_STRLEN): Define.
162 2013-11-06 Jakub Jelinek <jakub@redhat.com>
165 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
166 (expand_assignment): If *bitpos is negative, set *offset
167 and adjust *bitpos, so that it is not negative.
169 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
171 * config/i386/bdver3.md : Added two additional decoder units
172 to support issue rate of 4 and remodeled vector unit.
173 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
174 architectures is set to 4.
175 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
176 lookahead is set to 4 for BD architectures.
178 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
180 * config/rs6000/rs6000.c (rs6000_option_override_internal):
181 Remove restriction against use of VSX instructions when generating
182 code for little endian mode.
184 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
186 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
187 for both big and little endian.
188 (mulv8hi3): Swap input operands for merge high and merge low
189 instructions for little endian.
191 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
193 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
194 define_insn to define_expand that uses even patterns for big
195 endian and odd patterns for little endian.
196 (vec_widen_smult_even_v16qi): Likewise.
197 (vec_widen_umult_even_v8hi): Likewise.
198 (vec_widen_smult_even_v8hi): Likewise.
199 (vec_widen_umult_odd_v16qi): Likewise.
200 (vec_widen_smult_odd_v16qi): Likewise.
201 (vec_widen_umult_odd_v8hi): Likewise.
202 (vec_widen_smult_odd_v8hi): Likewise.
203 (altivec_vmuleub): New define_insn.
204 (altivec_vmuloub): Likewise.
205 (altivec_vmulesb): Likewise.
206 (altivec_vmulosb): Likewise.
207 (altivec_vmuleuh): Likewise.
208 (altivec_vmulouh): Likewise.
209 (altivec_vmulesh): Likewise.
210 (altivec_vmulosh): Likewise.
212 2013-11-05 Mike Stump <mikestump@comcast.net>
214 * Makefile.in (mostlyclean): Remove c-family objects.
216 2013-11-05 Ian Lance Taylor <iant@google.com>
218 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
219 If possible, add .cfi directives to record change to bx.
220 * config/i386/i386.c (ix86_emit_cfi): New function.
221 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
223 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
225 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
226 immediately after a label for a tablejump pattern.
228 * config/arm/arm.c (is_jump_table): Remove.
229 (create_fix_barrier): Use tablejump_p instead.
230 (arm_reorg): Likewise.
231 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
232 (thumb2_output_casesi): Likewise.
233 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
234 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
235 casesi_shift_media, casesi_load_media): Likewise.
236 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
237 * config/microblaze/microblaze.md: Likewise.
239 2013-11-05 Tobias Burnus <burnus@net-b.de>
241 * doc/invoke.texi (-Wdate-time): Document.
243 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
245 * double-int.c (lshift_double, rshift_double): Remove
246 SHIFT_COUNT_TRUNCATED handling.
248 2013-11-05 Jeff Law <law@redhat.com>
250 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
251 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
252 * gimple-ssa-isolate-paths.c: New file.
253 * gimple.c (check_loadstore): New function.
254 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
255 Verify OP is in the argument list and the argument corresponding
256 to OP is a pointer type. Use operand_equal_p rather than
257 pointer equality when testing if OP is on the nonnull list.
258 Use check_loadstore rather than count_ptr_derefs. Handle
259 GIMPLE_RETURN statements.
260 * tree-vrp.c (infer_nonnull_range): Remove.
261 * gimple.h (infer_nonnull_range): Declare.
262 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
263 * passes.def: Add pass_isolate_erroneous_paths.
264 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
265 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
266 * tree-ssa.c (struct count_ptr_d): Remove.
267 (count_ptr_derefs, count_uses_and_derefs): Remove.
268 * tree-ssa.h (count_uses_and_derefs): Remove.
270 2013-11-05 Jakub Jelinek <jakub@redhat.com>
272 PR rtl-optimization/58997
273 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
274 get_iv_value to be in iv->mode rather than iv->extend_mode.
275 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
276 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
277 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
280 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
282 * gimple.h: Move some prototypes to gimple-expr.h and add to include
284 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
285 Move to gimple-expr.h.
286 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
287 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
288 is_gimple_id, virtual_operand_p, is_gimple_addressable,
289 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
291 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
292 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
293 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
294 is_gimple_address, is_gimple_invariant_address,
295 is_gimple_ip_invariant_address, is_gimple_min_invariant,
296 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
297 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
298 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
299 gimple_decl_printable_name, useless_type_conversion_p,
300 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
302 * gimple-expr.c: New File.
303 (useless_type_conversion_p, gimple_set_body, gimple_body,
304 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
305 gimple_can_coalesce_p, extract_ops_from_tree_1,
306 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
307 is_gimple_address, is_gimple_invariant_address,
308 is_gimple_ip_invariant_address, is_gimple_min_invariant,
309 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
310 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
311 is_gimple_mem_ref_addr): Relocate here.
312 * Makefile.in (OBJS): Add gimple-expr.o.
314 2013-11-05 David Malcolm <dmalcolm@redhat.com>
316 * gengtype-parse.c (struct_field_seq): Support empty structs.
318 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
320 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
322 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
324 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
326 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
328 (LONG_DOUBLE_TYPE_SIZE): New define.
329 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
331 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
334 * doc/extend.texi [visibility ("visibility_type")]: Add example
335 about visibility attribute on namespace declaration.
337 2013-11-05 Richard Biener <rguenther@suse.de>
340 * passes.def (all_passes): Start with pass_fixup_cfg again.
342 2013-11-05 Richard Biener <rguenther@suse.de>
344 PR tree-optimization/58955
345 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
348 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
350 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
352 (vec_pack_ufix_trunc_v2df): Likewise.
354 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
357 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
358 pattern, instead of word_mode.
360 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
362 (set_storage_via_setmem): Likewise.
364 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
366 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
367 gimple.h and the rest of the condition in eliminate_build.
368 (eliminate_build): Call new routine.
369 * gimple.h (phi_ssa_name_p): Delete.
371 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
373 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
374 case of no prefix and reserving zero slots, because when that's the
375 case we'll never get here.
376 * vec.h (va_heap::reserve): Don't try and handle
377 vec_prefix::calculate_allocation returning zero because that should
380 2013-11-05 Richard Biener <rguenther@suse.de>
383 * tree-dfa.c (get_ref_base_and_extent): Merge common code
384 in MEM_REF and TARGET_MEM_REF handling. Make sure to
385 process trailing array detection before diving into the
386 view-converted object (and possibly apply some extra offset).
388 2013-11-05 Joseph Myers <joseph@codesourcery.com>
390 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
392 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
394 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
396 PR tree-optimization/58958
397 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
398 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
400 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
402 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
403 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
405 2013-11-05 Jakub Jelinek <jakub@redhat.com>
407 PR tree-optimization/58984
408 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
409 set *SIZE_P if non-NULL on success.
410 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
412 (ipcp_transform_function): Likewise. Punt if size of access
413 is different from TYPE_SIZE on v->value's type.
415 2013-11-05 Tobias Burnus <burnus@net-b.de>
417 * doc/invoke.texi (-fopenmp-simd): Document new option.
418 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
419 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
420 * tree.c (attribute_value_equal): Ditto.
422 2013-11-04 Wei Mi <wmi@google.com>
424 * sched-rgn.c (add_branch_dependences): Keep insns in
425 a SCHED_GROUP at the end of BB to remain their location.
427 2013-11-04 Wei Mi <wmi@google.com>
429 * config/i386/i386.c (memory_address_length): Extract a part
430 of code to rip_relative_addr_p.
431 (rip_relative_addr_p): New Function.
432 (ix86_macro_fusion_p): Ditto.
433 (ix86_macro_fusion_pair_p): Ditto.
434 * config/i386/i386.h: Add new tune features about macro-fusion.
435 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
436 * doc/tm.texi: Generated.
437 * doc/tm.texi.in: Ditto.
438 * haifa-sched.c (try_group_insn): New Function.
439 (group_insns_for_macro_fusion): Ditto.
440 (sched_init): Call group_insns_for_macro_fusion.
441 * target.def: Add two hooks: macro_fusion_p and
444 2013-11-04 Kostya Serebryany <kcc@google.com>
446 Update to match the changed asan API.
447 * asan.c (asan_function_start): New function.
448 (asan_emit_stack_protection): Update the string stored in the
449 stack red zone to match new API. Store the PC of the current
450 function in the red zone.
451 (asan_global_struct): Update the __asan_global definition to match
453 (asan_add_global): Ditto.
454 * asan.h (asan_function_start): New prototype.
455 * final.c (final_start_function): Call asan_function_start.
456 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
459 2013-11-04 Wei Mi <wmi@google.com>
461 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
462 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
463 * config/i386/i386.c (ix86_option_override_internal): Ditto.
464 (ix86_issue_rate): Ditto.
465 (ix86_adjust_cost): Ditto.
466 (ia32_multipass_dfa_lookahead): Ditto.
467 (ix86_sched_init_global): Ditto.
468 (get_builtin_code_for_version): Ditto.
469 * config/i386/i386.h (enum target_cpu_default): Ditto.
470 (enum processor_type): Ditto.
471 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
473 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
475 PR rtl-optimization/58967
476 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
477 !lra_in_progress for mode sizes bigger word.
479 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
481 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
482 arguments to merge instruction for little endian.
483 (vec_widen_umult_lo_v16qi): Likewise.
484 (vec_widen_smult_hi_v16qi): Likewise.
485 (vec_widen_smult_lo_v16qi): Likewise.
486 (vec_widen_umult_hi_v8hi): Likewise.
487 (vec_widen_umult_lo_v8hi): Likewise.
488 (vec_widen_smult_hi_v8hi): Likewise.
489 (vec_widen_smult_lo_v8hi): Likewise.
491 2013-11-04 Ian Lance Taylor <iant@google.com>
493 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
494 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
495 builtins that take pointers.
496 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
497 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
498 (run_gcc): Pass -fnon-call-exceptions.
500 2013-11-04 Jakub Jelinek <jakub@redhat.com>
502 * optabs.c (expand_vec_perm): Revert one incorrect line from
505 PR tree-optimization/58978
506 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
507 use_stmt by single_imm_use directly. Only call single_imm_use
510 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
512 PR rtl-optimization/58968
513 * lra-spills.c (return_regno_p): New function.
514 (lra_final_code_change): Use it.
516 2013-11-04 Joseph Myers <joseph@codesourcery.com>
518 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
519 * target.def (float_exceptions_rounding_supported_p): New hook.
520 * targhooks.c (default_float_exceptions_rounding_supported_p): New
522 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
523 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
525 * doc/tm.texi: Regenerate.
526 * config.gcc (powerpc*-*-linux*): Set extra_objs.
527 * config/rs6000/rs6000-linux.c: New file.
528 * config/rs6000/rs6000-protos.h
529 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
530 * config/rs6000/linux.h
531 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
532 * config/rs6000/linux64.h
533 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
534 * config/rs6000/t-linux (rs6000-linux.o): New rule.
535 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
537 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
539 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
540 Replace the define_insn_and_split with a define_insn and two
541 define_splits, with the split after reload re-permuting the source
542 register to its original value.
543 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
544 (*vsx_le_perm_store_v8hi): Likewise.
545 (*vsx_le_perm_store_v16qi): Likewise.
547 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
549 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
552 2013-11-04 Jakub Jelinek <jakub@redhat.com>
554 PR tree-optimization/58946
555 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
556 bbs with bbinfo[idx].op != NULL before all blocks with
557 bbinfo[idx].op == NULL.
559 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
561 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
562 (avr_dump_double_int_hex): Likewise.
563 (avr_log_vadump): Remove %D and %X handling.
564 * config/avr/avr.c (avr_double_int_push_digit): Delete.
565 (avr_map_op_t): Change map from double_int to unsigned int.
566 (avr_map_op): Update accordingly.
567 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
568 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
569 unsigned ints rather than double_ints.
571 2013-11-03 Marek Polacek <polacek@redhat.com>
573 Implement -fsanitize=vla-bound.
574 * opts.c (common_handle_option): Handle vla-bound.
575 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
576 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
577 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
579 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
581 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
584 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
586 * config/i386/constraints.md (Ts, Tv): New address constrains.
587 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
588 constraint for address_no_seg_operand.
589 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
590 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
591 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
592 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
593 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
594 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
595 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
596 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
597 constraint for vsib_address_operand.
599 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
601 * gcse.c (pre_delete): Remove references to regmove from comments.
602 * recog.c: (validate_replace_rtx_1): Likewise.
603 * config/rl78/rl78.c: Likewise.
604 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
605 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
606 * common/config/mmix/mmix-common.c: Likewise.
608 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
610 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
611 * gimplify.c (gimplify_compound_lval): Likewise.
612 * graphite-clast-to-gimple.c (gloog): Likewise.
613 * graphite-dependences.c (loop_is_parallel_p): Likewise.
614 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
615 (limit_scop); Likewise.
616 (build_scops): Likewise.
617 (dot_scop): Likewise.
618 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
619 (build_scop_drs): Likewise.
620 (insert_stmts): Likewise.
621 (insert_out_of_ssa_copy): Likewise.
622 (remove_phi): Likewise.
623 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
624 * hw-doloop.c (discover_loop): Likewise.
625 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
626 * tree-dfa.c (dump_enumerated_decls): Likewise.
627 * tree-if-conv.c (if_convertable_loop_p): Likewise.
628 * tree-inline.c (tree_function_versioning): Likewise.
629 * tree-loop-distribution.c (build_rdg): Likewise.
630 (rdg_flag_vertex_and_dependent): Likewise.
631 (distribute_loop): Likewise.
632 * tree-parloops.c (loop_parallel_p): Likewise.
633 (eliminate_local_variables): Likewise.
634 (separate_decls_in_region): Likewise.
635 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
636 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
637 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
638 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
639 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
640 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
641 (vectorizable_condition): Likewise.
643 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
645 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
646 * configure: Regenerate.
647 * config/i386/i386.md (*movdi_internal): Change
648 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
649 (*movdf_internal): Ditto.
650 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
651 * config/i386/sse.md (vec_concatv2di): Output interunit movq
652 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
654 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
656 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
659 2013-10-31 David Malcolm <dmalcolm@redhat.com>
661 Automated part of renaming of symtab_node_base to symtab_node.
663 Patch autogenerated by rename_symtab.py from
664 https://github.com/davidmalcolm/gcc-refactoring-scripts
665 revision 58bb219cc090b2f4516a9297d868c245495ee622
666 with ChangeLog entry fixed up by hand.
668 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
670 (cgraph_node_for_asm): Likewise.
671 * cgraph.h (symtab_node_base): Likewise.
672 (cgraph_node): Likewise.
673 (varpool_node): Likewise.
674 (is_a_helper <cgraph_node>::test): Likewise.
675 (is_a_helper <varpool_node>::test): Likewise.
676 (symtab_nodes): Likewise.
677 (symtab_register_node): Likewise.
678 (symtab_unregister_node): Likewise.
679 (symtab_remove_node): Likewise.
680 (symtab_get_node): Likewise.
681 (symtab_node_for_asm): Likewise.
682 (symtab_node_asm_name): Likewise.
683 (symtab_node_name): Likewise.
684 (symtab_insert_node_to_hashtable): Likewise.
685 (symtab_add_to_same_comdat_group): Likewise.
686 (symtab_dissolve_same_comdat_group_list): Likewise.
687 (dump_symtab_node): Likewise.
688 (debug_symtab_node): Likewise.
689 (dump_symtab_base): Likewise.
690 (verify_symtab_node): Likewise.
691 (verify_symtab_base): Likewise.
692 (symtab_used_from_object_file_p): Likewise.
693 (symtab_alias_ultimate_target): Likewise.
694 (symtab_resolve_alias): Likewise.
695 (fixup_same_cpp_alias_visibility): Likewise.
696 (symtab_for_node_and_aliases): Likewise.
697 (symtab_nonoverwritable_alias): Likewise.
698 (availability symtab_node_availability): Likewise.
699 (symtab_semantically_equivalent_p): Likewise.
700 (fixup_same_cpp_alias_visibility): Likewise.
701 (symtab_prevail_in_asm_name_hash): Likewise.
704 (varpool_first_variable): Likewise.
705 (varpool_next_variable): Likewise.
706 (varpool_first_static_initializer): Likewise.
707 (varpool_next_static_initializer): Likewise.
708 (varpool_first_defined_variable): Likewise.
709 (varpool_next_defined_variable): Likewise.
710 (cgraph_first_defined_function): Likewise.
711 (cgraph_next_defined_function): Likewise.
712 (cgraph_first_function): Likewise.
713 (cgraph_next_function): Likewise.
714 (cgraph_first_function_with_gimple_body): Likewise.
715 (cgraph_next_function_with_gimple_body): Likewise.
716 (symtab_alias_target): Likewise.
717 (symtab_real_symbol_p): Likewise.
718 (symtab_can_be_discarded): Likewise.
719 * cgraphbuild.c (mark_address): Likewise.
720 (mark_load): Likewise.
721 (mark_store): Likewise.
722 * cgraphunit.c (decide_is_symbol_needed): Likewise.
724 (enqueue_node): Likewise.
725 (referred_to_p): Likewise.
726 (cgraph_process_same_body_aliases): Likewise.
727 (analyze_functions): Likewise.
728 (handle_alias_pairs): Likewise.
729 (output_weakrefs): Likewise.
731 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
732 * ipa-inline-analysis.c (inline_write_summary): Likewise.
733 * ipa-prop.c (remove_described_reference): Likewise.
734 (try_decrement_rdesc_refcount): Likewise.
735 (ipa_edge_duplication_hook): Likewise.
736 * ipa-ref.c (ipa_record_reference): Likewise.
737 (ipa_maybe_record_reference): Likewise.
738 (ipa_clone_ref): Likewise.
739 (ipa_clone_references): Likewise.
740 (ipa_clone_referring): Likewise.
741 (ipa_find_reference): Likewise.
742 (ipa_remove_stmt_references): Likewise.
743 (ipa_clear_stmts_in_references): Likewise.
744 * ipa-ref.h (symtab_node_base): Likewise.
746 (ipa_record_reference): Likewise.
747 (ipa_maybe_record_reference): Likewise.
748 (ipa_clone_references): Likewise.
749 (ipa_clone_referring): Likewise.
750 (ipa_clone_ref): Likewise.
751 (ipa_find_reference): Likewise.
752 (ipa_remove_stmt_references): Likewise.
753 (ipa_clear_stmts_in_references): Likewise.
754 * ipa-reference.c (ipa_reference_write_optimization_summary):
756 * ipa.c (enqueue_node): Likewise.
757 (process_references): Likewise.
758 (walk_polymorphic_call_targets): Likewise.
759 (symtab_remove_unreachable_nodes): Likewise.
760 (address_taken_from_non_vtable_p): Likewise.
761 (comdat_can_be_unshared_p_1): Likewise.
762 (comdat_can_be_unshared_p): Likewise.
763 (can_replace_by_local_alias): Likewise.
764 (function_and_variable_visibility): Likewise.
765 * is-a.h: Likewise (within example in comment).
766 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
767 (lto_symtab_encoder_encode): Likewise.
768 (lto_symtab_encoder_delete_node): Likewise.
769 (lto_symtab_encoder_in_partition_p): Likewise.
770 (lto_set_symtab_encoder_in_partition): Likewise.
771 (output_refs): Likewise.
772 (compute_ltrans_boundary): Likewise.
773 (output_symtab): Likewise.
774 (input_node): Likewise.
775 (input_ref): Likewise.
776 (input_edge): Likewise.
777 (input_cgraph_1): Likewise.
778 (input_refs): Likewise.
779 (output_cgraph_opt_summary): Likewise.
780 (input_node_opt_summary): Likewise.
781 (input_cgraph_opt_section): Likewise.
782 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
784 * lto-streamer-out.c (lto_output): Likewise.
785 (output_symbol_p): Likewise.
786 (produce_symtab): Likewise.
787 * lto-streamer.h (lto_encoder_entry): Likewise.
788 (lto_free_function_in_decl_state_for_node): Likewise.
789 (lto_symtab_encoder_encode): Likewise.
790 (lto_symtab_encoder_delete_node): Likewise.
791 (lto_symtab_encoder_in_partition_p): Likewise.
792 (lto_set_symtab_encoder_in_partition): Likewise.
793 (lto_symtab_encoder_lookup): Likewise.
794 (lsei_node): Likewise.
795 (lto_symtab_encoder_deref): Likewise.
796 * symtab.c (symtab_hash): Likewise.
797 (assembler_name_hash): Likewise.
798 (symtab_nodes): Likewise.
799 (hash_node): Likewise.
801 (hash_node_by_assembler_name): Likewise.
802 (eq_assembler_name): Likewise.
803 (insert_to_assembler_name_hash): Likewise.
804 (unlink_from_assembler_name_hash): Likewise.
805 (symtab_prevail_in_asm_name_hash): Likewise.
806 (symtab_register_node): Likewise.
807 (symtab_insert_node_to_hashtable): Likewise.
808 (symtab_unregister_node): Likewise.
809 (symtab_get_node): Likewise.
810 (symtab_remove_node): Likewise.
811 (symtab_initialize_asm_name_hash): Likewise.
812 (symtab_node_for_asm): Likewise.
813 (symtab_add_to_same_comdat_group): Likewise.
814 (symtab_dissolve_same_comdat_group_list): Likewise.
815 (symtab_node_asm_name): Likewise.
816 (symtab_node_name): Likewise.
817 (dump_symtab_base): Likewise.
818 (dump_symtab_node): Likewise.
819 (dump_symtab): Likewise.
820 (debug_symtab_node): Likewise.
821 (verify_symtab_base): Likewise.
822 (verify_symtab_node): Likewise.
823 (verify_symtab): Likewise.
824 (symtab_used_from_object_file_p): Likewise.
825 (symtab_node_availability): Likewise.
826 (symtab_alias_ultimate_target): Likewise.
827 (fixup_same_cpp_alias_visibility): Likewise.
828 (symtab_resolve_alias): Likewise.
829 (symtab_for_node_and_aliases): Likewise.
830 (symtab_for_node_and_aliases): Likewise.
831 (symtab_nonoverwritable_alias_1): Likewise.
832 (symtab_nonoverwritable_alias): Likewise.
833 (symtab_semantically_equivalent_p): Likewise.
834 * value-prof.c (init_node_map): Likewise.
835 * varasm.c (find_decl): Likewise.
836 * varpool.c (varpool_node_for_asm): Likewise.
837 (varpool_remove_unreferenced_decls): Likewise.
839 2013-10-31 David Malcolm <dmalcolm@redhat.com>
841 Manual part of renaming of symtab_node_base to symtab_node.
843 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
844 clashes with the preferred name for the base class.
845 (const_symtab_node): Remove redundant typedef.
847 2013-10-31 Jakub Jelinek <jakub@redhat.com>
849 * optabs.c (expand_vec_perm): Avoid vector mode punning
851 * expmed.c (store_bit_field_1): Likewise.
852 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
853 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
854 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
855 * config/i386/i386.c (ix86_expand_vector_move_misalign,
856 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
857 ix86_expand_sse_unpack, ix86_expand_args_builtin,
858 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
859 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
860 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
861 expand_vec_perm_vpshufb2_vpermq,
862 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
863 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
864 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
865 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
868 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
870 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
871 Use gcc_unreachable for unhandled modes. Do not check results of
872 expand_simple_binop. If not expanded to target, move the result.
874 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
875 Shiva Chen <shiva0217@gmail.com>
877 * config.gcc (nds32*-*-*): Add nds32 target.
878 * config/nds32/nds32.c: New file.
879 * config/nds32/nds32.h: New file.
880 * config/nds32/nds32.md: New file.
881 * config/nds32/constants.md: New file.
882 * config/nds32/constraints.md: New file.
883 * config/nds32/iterators.md: New file.
884 * config/nds32/nds32-doubleword.md: New file.
885 * config/nds32/nds32-intrinsic.md: New file.
886 * config/nds32/nds32_intrinsic.h: New file.
887 * config/nds32/nds32-modes.def: New file.
888 * config/nds32/nds32-multiple.md: New file.
889 * config/nds32/nds32.opt: New file.
890 * config/nds32/nds32-opts.h: New file.
891 * config/nds32/nds32-protos.h: New file.
892 * config/nds32/nds32-peephole2.md: New file.
893 * config/nds32/pipelines.md: New file.
894 * config/nds32/predicates.md: New file.
895 * config/nds32/t-mlibs: New file.
896 * common/config/nds32: New directory and files.
898 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
899 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
900 * doc/install.texi (Cross-Compiler-Specific Options): Document
901 --with-nds32-lib for nds32 target.
902 * doc/extend.texi (Function Attributes, Target Builtins): Document
903 nds32 specific attributes.
905 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
907 * lra-constraints (process_alt_operands): Use the result
908 elimination register for operand when matching constraints.
910 2013-10-31 Jakub Jelinek <jakub@redhat.com>
912 * tree-vrp.c (maybe_set_nonzero_bits): New function.
913 (remove_range_assertions): Call it.
915 * tree.c (tree_ctz): New function.
916 * tree.h (tree_ctz): New prototype.
917 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
918 first argument from tree to const_tree.
919 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
920 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
921 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
923 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
924 * tree-vect-loop.c (vect_analyze_loop_operations,
925 vect_transform_loop): Don't force scalar loop for bound just because
926 number of iterations is unknown, only do it if it is not known to be
927 a multiple of vectorization_factor.
928 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
930 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
931 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
932 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
933 (set_nonzero_bits, get_nonzero_bits): New prototypes.
934 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
935 a default def isn't partially constant.
936 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
937 is known to be partially zero.
938 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
939 to see if a default def isn't partially constant.
940 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
941 creation of a range, if VR_RANGE, try to improve nonzero_bits from
943 (set_nonzero_bits, get_nonzero_bits): New functions.
945 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
946 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
947 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
948 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
949 uses but in the condition and ASSERT_EXPR and the other successor of
950 the predecessor bb is __builtin_unreachable (), set_range_info of the
951 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
953 2013-10-31 Martin Jambor <mjambor@suse.cz>
955 PR rtl-optimization/58934
957 2013-10-30 Martin Jambor <mjambor@suse.cz>
958 PR rtl-optimization/10474
959 * ira.c (find_moveable_pseudos): Do not calculate dominance info
961 (interesting_dest_for_shprep): New function.
962 (split_live_ranges_for_shrink_wrap): Likewise.
963 (ira): Calculate dominance info and df analysis. Call
964 split_live_ranges_for_shrink_wrap.
966 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
967 Yury Gribov <y.gribov@samsung.com>
970 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
971 shadow pointer to avoid clobbering the main one.
973 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
975 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
977 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
980 * ira-color.c (update_costs_from_copies): Add new parameter. Use
981 it for calling update_costs_from_allocno.
982 (assign_hard_reg): Call restore_costs_from_copies only for
983 !retry_p. Pass new argument to update_costs_from_copies.
984 (color_pass): Pass new argument to update_costs_from_copies.
985 (ira_mark_allocation_change): Ditto.
987 2013-10-30 Sharad Singhai <singhai@google.com>
990 * opts.c (common_handle_option): Remove deprecated option
991 -ftree-vectorizer-verbose.
992 * doc/invoke.texi (Debugging Options): Ditto.
993 * opts-global.c (handle_common_deferred_options): Ditto.
994 (dump_remap_tree_vectorizer_verbose): Delete.
995 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
997 2013-10-30 DJ Delorie <dj@redhat.com>
999 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1000 that take no arguments.
1002 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1005 * reload1.c (update_eliminables_and_spill): New function, broken
1007 (reload): Use it. Check for frame size change after frame size
1008 alignment, and call update_eliminables_and_spill first if continue-ing.
1010 2013-10-30 Cong Hou <congh@google.com>
1013 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1014 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1015 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1017 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1019 PR rtl-optimization/58369
1020 * reload1.c (compute_reload_subreg_offset): New function.
1021 (choose_reload_regs): Use it to pass endian-correct
1022 offset to subreg_regno_offset.
1024 2013-10-30 Tobias Burnus <burnus@net-b.de>
1027 * tree-cfg.c (replace_loop_annotate): Replace warning by
1030 2013-10-30 Jason Merrill <jason@redhat.com>
1032 * configure.ac (loose_warn): Add -Wno-format if
1033 --disable-build-format-warnings.
1035 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1037 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1038 onto multiple lines to make things easier for rename_symtab.py.
1040 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1041 (symtab_semantically_equivalent_p): Likewise.
1043 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1046 * lra.c (check_rtl): Remove address check before LRA work.
1048 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1050 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1051 POINTER_PLUS_EXPR in the defining statement.
1053 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1055 * regmove.c: Remove.
1056 * tree-pass.h (make_pass_regmove): Remove.
1057 * timevar.def (TV_REGMOVE): Remove.
1058 * passes.def (pass_regmove): Remove.
1059 * opts.c (default_options_table): Remove entry for regmove.
1060 * doc/passes.texi: Remove regmove pass description.
1061 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1063 (-fdump-rtl-regmove): Ditto.
1064 * common.opt (foptimize-register-move, fregmove): Ignore.
1065 * Makefile.in (OBJS): Remove regmove.o.
1066 * regmove.c: Remove.
1067 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1069 (struct ira_allocno) New member allocno_prefs.
1070 (ALLOCNO_PREFS): New macro.
1071 (ira_prefs, ira_prefs_num): New external vars.
1072 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1073 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1074 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1076 (ira_add_allocno_copy_to_list): Remove prototype.
1077 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1078 (ira_pref_iterator): New type.
1079 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1080 (FOR_EACH_PREF): New macro.
1081 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1082 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1084 (ira_setup_alts): New function.
1085 (decrease_live_ranges_number): New function.
1086 (ira): Call the above function.
1087 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1088 (ira_create_allocno): Initialize allocno prefs.
1089 (pref_pool, pref_vec): New static vars.
1090 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1092 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1093 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1094 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1095 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1096 (ira_add_allocno_copy_to_list): Make static. Rename to
1097 add_allocno_copy_to_list.
1098 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1099 swap_allocno_copy_ends_if_necessary.
1100 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1101 ira_remove_allocno_prefs.
1102 (ira_flattening): Ditto.
1103 (ira_build): Call initiate_prefs, print_prefs.
1104 (ira_destroy): Call finish_prefs.
1105 * ira-color.c (struct update_cost_record): New.
1106 (struct allocno_color_data): Add new member update_cost_records.
1107 (update_cost_record_pool): New static var.
1108 (init_update_cost_records, get_update_cost_record): New functions.
1109 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1110 (struct update_cost_queue_elem): Add member from.
1111 (initiate_cost_update): Call init_update_cost_records.
1112 (finish_cost_update): Call finish_update_cost_records.
1113 (queue_update_cost, get_next_update_cost): Add new param from.
1114 (Update_allocno_cost, update_costs_from_allocno): New functions.
1115 (update_costs_from_prefs): Ditto.
1116 (update_copy_costs): Rename to update_costs_from_copies.
1117 (restore_costs_from_copies): New function.
1118 (update_conflict_hard_regno_costs): Don't go back.
1119 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1121 (pop_allocnos): Add priniting more debug info.
1122 (color_allocnos): Remove prefs for conflicting hard regs.
1123 Call update_costs_from_prefs.
1124 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1125 (get_dup_num): Rename, modify, and move to ira.c
1126 (process_regs_for_copy): Add prefs.
1127 (add_insn_allocno_copies): Put src as first arg of
1128 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1129 * ira-costs.c (record_reg_classes): Modify and move code into
1130 record_operands_costs.
1131 (find_costs_and_classes): Create prefs for the hard reg of small
1133 (process_bb_node_for_hard_reg_moves): Add prefs.
1135 2013-10-30 Richard Biener <rguenther@suse.de>
1138 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1139 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1141 (pre_and_rev_post_order_compute): ... which now wraps it.
1142 * graph.c (draw_cfg_nodes_no_loops): Use
1143 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1146 2013-10-30 Christian Bruel <christian.bruel@st.com>
1148 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1149 (sh_expand_cmpstr): Handle known align and schedule improvements.
1150 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1151 * config/sh/sh.md (cmpstrnsi): New pattern.
1153 2013-10-30 Martin Jambor <mjambor@suse.cz>
1155 PR rtl-optimization/10474
1156 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1158 (interesting_dest_for_shprep): New function.
1159 (split_live_ranges_for_shrink_wrap): Likewise.
1160 (ira): Calculate dominance info and df analysis. Call
1161 split_live_ranges_for_shrink_wrap.
1163 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1166 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1168 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1170 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1171 * tree.h (POINTER_BOUNDS_P): New.
1172 (BOUNDED_TYPE_P): New.
1174 (pointer_bounds_type_node): New.
1175 * tree.c (build_common_tree_nodes): Initialize
1176 pointer_bounds_type_node.
1177 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1178 (gimple_call_num_nobnd_args): New.
1179 (gimple_call_nobnd_arg): New.
1180 (gimple_return_retbnd): New.
1181 (gimple_return_set_retbnd): New
1182 * gimple.c (gimple_build_return): Increase number of ops
1183 for return statement.
1184 (gimple_call_get_nobnd_arg_index): New.
1185 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1187 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1189 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1190 with "chkp ctor" and "bnd_legacy" attributes.
1191 * gimplify.c (gimplify_init_constructor): Avoid infinite
1192 loop during gimplification of bounds initializer.
1194 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1196 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1197 (handle_bnd_legacy): New.
1198 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1199 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1202 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1204 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1205 (BT_FN_PTR_CONST_PTR): New.
1206 (BT_FN_CONST_PTR_CONST_PTR): New.
1207 (BT_FN_PTR_CONST_PTR_SIZE): New.
1208 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1209 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1210 (BT_FN_VOID_CONST_PTR_SIZE): New.
1211 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1212 * chkp-builtins.def: New.
1213 * builtins.def: include chkp-builtins.def.
1214 (DEF_CHKP_BUILTIN): New.
1215 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1216 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1217 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1218 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1219 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1220 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1221 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1222 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1223 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1224 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1225 * common.opt (fcheck-pointer-bounds): New.
1226 * toplev.c (process_options): Check Pointer Bounds Checker is
1228 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1230 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1232 * target.def (builtin_chkp_function): New.
1233 (chkp_bound_type): New.
1234 (chkp_bound_mode): New.
1235 (fn_abi_va_list_bounds_size): New.
1236 (load_bounds_for_arg): New.
1237 (store_bounds_for_arg): New.
1238 * targhooks.h (default_load_bounds_for_arg): New.
1239 (default_store_bounds_for_arg): New.
1240 (default_fn_abi_va_list_bounds_size): New.
1241 (default_chkp_bound_type): New.
1242 (default_chkp_bound_mode): New.
1243 (default_builtin_chkp_function): New.
1244 * targhooks.c (default_load_bounds_for_arg): New.
1245 (default_store_bounds_for_arg): New.
1246 (default_fn_abi_va_list_bounds_size): New.
1247 (default_chkp_bound_type): New.
1248 (default_chkp_bound_mode); New.
1249 (default_builtin_chkp_function): New.
1250 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1251 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1252 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1253 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1254 (TARGET_CHKP_BOUND_TYPE): New.
1255 (TARGET_CHKP_BOUND_MODE): New.
1256 * doc/tm.texi: Regenerated.
1257 * langhooks.h (lang_hooks): Add chkp_supported field.
1258 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1259 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1261 2013-10-29 Andrew Pinski <apinski@cavium.com>
1263 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1264 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1265 fails, combining the branches anyways.
1266 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1267 increases the number of combinings.
1268 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1270 2013-10-29 Mike Stump <mikestump@comcast.net>
1272 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1273 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1274 (make_vector_mode): Increase namebuf to 16.
1275 (emit_insn_modes_h): When processing BImode, don't
1276 also match partial int modes.
1277 (emit_class_narrowest_mode): Likewise.
1279 * config/bfin/bfin-modes.def: Add precision to PDI.
1280 * config/m32c/m32c-modes.def: Add precision to PSI.
1281 * config/msp430/msp430-modes.def: Add precision to PSI.
1282 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1283 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1285 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1288 * config/sh/sh.md (*addc): Rename existing variations to ...
1289 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
1290 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
1291 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
1292 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
1294 2013-10-29 Teresa Johnson <tejohnson@google.com>
1297 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
1298 insanities when updating probabilities.
1300 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1302 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
1303 prettyprinter for cgraph_node to reflect the conversion of the
1304 symtable types to a C++ class hierarchy: it now *is* a
1305 symtab_node_base, rather than having one (named "symbol").
1307 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
1309 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
1310 __cilkrts_pop_frame. If matched, then return true for built-in
1312 (expand_builtin): Added BUILT_IN_CILK_DETACH and
1313 BUILT_IN_CILK_POP_FRAME case.
1314 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
1315 (lhs_cilk_detect_spawn): Likewise.
1316 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
1317 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
1318 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
1319 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
1320 (LANG_HOOKS_CILKPLUS): Likewise.
1321 * tree.h (CILK_SPAWN_FN): Likewise.
1322 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
1323 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
1324 (OBJS): Added cilk-common.o.
1325 (BUILTINS_DEF): Added cilk-builtins.def.
1326 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
1327 (lhd_cilk_detect_spawn): Likewise.
1328 * langhooks.h (lang_hooks_for_cilkplus): New struct.
1329 (struct lang_hooks): Added new field called "cilkplus."
1330 * cilk-common.c: New file.
1332 * cilk-builtins.def: Likewise.
1333 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
1334 "__cilk" macro and set it to 200.
1335 * function.h (struct function::cilk_frame_decl): New field.
1336 (struct function::is_cilk_function): Likewise.
1337 (struct function::calls_cilk_spawn): Likewise.
1338 * gimplify.c (gimplify_call_expr): Added a check if the function call
1339 being gimplified is a spawn detach point. If so, then add pop_frame
1340 and detach function calls.
1341 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
1342 for gimplifying _Cilk_spawn and _Cilk_sync statements.
1343 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
1344 function. If so, added a _Cilk_sync and gimplified it.
1345 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
1346 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
1347 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
1349 (can_inline_edge_p): Prevent inling of spawnee function.
1350 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
1351 for functions that use Cilk keywords.
1352 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
1353 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
1354 CILK_SYNC_STMT cases.
1355 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
1357 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
1358 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1359 * passes.texi (Cilk Keywords): New section that describes the compiler
1360 code changes for handling Cilk Keywords.
1362 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1364 Patch autogenerated by refactor_symtab.py from
1365 https://github.com/davidmalcolm/gcc-refactoring-scripts
1366 revision 58bb219cc090b2f4516a9297d868c245495ee622
1368 * asan.c (asan_finish_file): Update for conversion of symtab types to
1369 a true class hierarchy.
1370 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1371 * cgraph.c (cgraph_get_body): Likewise.
1372 (cgraph_get_create_real_symbol_node): Likewise.
1373 (verify_cgraph_node): Likewise.
1374 (verify_edge_corresponds_to_fndecl): Likewise.
1375 (verify_edge_count_and_frequency): Likewise.
1376 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1377 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1378 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1379 (cgraph_node_cannot_return): Likewise.
1380 (cgraph_set_pure_flag_1): Likewise.
1381 (cgraph_set_const_flag_1): Likewise.
1382 (cgraph_set_nothrow_flag_1): Likewise.
1383 (cgraph_make_node_local_1): Likewise.
1384 (cgraph_for_node_and_aliases): Likewise.
1385 (cgraph_for_node_thunks_and_aliases): Likewise.
1386 (cgraph_node_can_be_local_p): Likewise.
1387 (cgraph_node_cannot_be_local_p_1): Likewise.
1388 (cgraph_function_body_availability): Likewise.
1389 (dump_cgraph_node): Likewise.
1390 (cgraph_rtl_info): Likewise.
1391 (cgraph_mark_address_taken_node): Likewise.
1392 (cgraph_remove_node): Likewise.
1393 (cgraph_release_function_body): Likewise.
1394 (cgraph_update_edges_for_call_stmt_node): Likewise.
1395 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1396 (cgraph_make_edge_direct): Likewise.
1397 (cgraph_resolve_speculation): Likewise.
1398 (cgraph_speculative_call_info): Likewise.
1399 (cgraph_turn_edge_to_speculative): Likewise.
1400 (cgraph_create_edge_1): Likewise.
1401 (cgraph_set_call_stmt): Likewise.
1402 (cgraph_node_for_asm): Likewise.
1403 (cgraph_add_thunk): Likewise.
1404 (cgraph_same_body_alias): Likewise.
1405 (cgraph_create_function_alias): Likewise.
1406 (cgraph_create_node): Likewise.
1407 (cgraph_create_empty_node): Likewise.
1408 (record_function_versions): Likewise.
1409 (used_from_object_file_p): Likewise.
1410 * cgraph.h (symtab_can_be_discarded): Likewise.
1411 (symtab_real_symbol_p): Likewise.
1412 (cgraph_mark_force_output_node): Likewise.
1413 (cgraph_edge_recursive_p): Likewise.
1414 (symtab_alias_target): Likewise.
1415 (varpool_all_refs_explicit_p): Likewise.
1416 (varpool_can_remove_if_no_refs): Likewise.
1417 (cgraph_only_called_directly_or_aliased_p): Likewise.
1418 (cgraph_next_function_with_gimple_body): Likewise.
1419 (cgraph_first_function_with_gimple_body): Likewise.
1420 (cgraph_function_with_gimple_body_p): Likewise.
1421 (cgraph_next_function): Likewise.
1422 (cgraph_first_function): Likewise.
1423 (cgraph_next_defined_function): Likewise.
1424 (cgraph_first_defined_function): Likewise.
1425 (varpool_next_defined_variable): Likewise.
1426 (varpool_first_defined_variable): Likewise.
1427 (varpool_next_static_initializer): Likewise.
1428 (varpool_first_static_initializer): Likewise.
1429 (varpool_next_variable): Likewise.
1430 (varpool_first_variable): Likewise.
1431 (varpool_node_name): Likewise.
1432 (varpool): Likewise.
1434 (is_a_helper <varpool_node>::test): Likewise.
1435 (is_a_helper <cgraph_node>::test): Likewise.
1436 (varpool_variable_node): Likewise.
1437 (cgraph_function_or_thunk_node): Likewise.
1438 (varpool_alias_target): Likewise.
1439 (cgraph_alias_target): Likewise.
1440 (cgraph_node_name): Likewise.
1441 (varpool_node_asm_name): Likewise.
1442 (cgraph_node_asm_name): Likewise.
1443 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1444 (cgraph_rebuild_references): Likewise.
1445 (rebuild_cgraph_edges): Likewise.
1446 (record_eh_tables): Likewise.
1447 (build_cgraph_edges): Likewise.
1448 (mark_store): Likewise.
1449 (mark_load): Likewise.
1450 (mark_address): Likewise.
1451 (record_type_list): Likewise.
1452 (record_reference): Likewise.
1453 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1454 (cgraph_materialize_clone): Likewise.
1455 (cgraph_function_versioning): Likewise.
1456 (cgraph_copy_node_for_versioning): Likewise.
1457 (update_call_expr): Likewise.
1458 (cgraph_find_replacement_node): Likewise.
1459 (cgraph_create_virtual_clone): Likewise.
1460 (cgraph_clone_node): Likewise.
1461 * cgraphunit.c (compile): Likewise.
1462 (output_weakrefs): Likewise.
1463 (output_in_order): Likewise.
1464 (expand_function): Likewise.
1465 (assemble_thunks_and_aliases): Likewise.
1466 (expand_thunk): Likewise.
1467 (mark_functions_to_output): Likewise.
1468 (handle_alias_pairs): Likewise.
1469 (analyze_functions): Likewise.
1470 (walk_polymorphic_call_targets): Likewise.
1471 (varpool_finalize_decl): Likewise.
1472 (process_function_and_variable_attributes): Likewise.
1473 (cgraph_process_same_body_aliases): Likewise.
1474 (analyze_function): Likewise.
1475 (cgraph_add_new_function): Likewise.
1476 (cgraph_finalize_function): Likewise.
1477 (referred_to_p): Likewise.
1478 (cgraph_reset_node): Likewise.
1479 (cgraph_process_new_functions): Likewise.
1480 (enqueue_node): Likewise.
1481 (decide_is_symbol_needed): Likewise.
1482 * coverage.c (coverage_compute_profile_id): Likewise.
1483 * dbxout.c (dbxout_expand_expr): Likewise.
1484 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1485 (reference_to_unused): Likewise.
1486 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1487 * gimplify.c (unvisit_body): Likewise.
1488 (unshare_body): Likewise.
1489 * ipa-cp.c (ipcp_generate_summary): Likewise.
1490 (ipcp_decision_stage): Likewise.
1491 (identify_dead_nodes): Likewise.
1492 (decide_whether_version_node): Likewise.
1493 (decide_about_value): Likewise.
1494 (perhaps_add_new_callers): Likewise.
1495 (create_specialized_node): Likewise.
1496 (update_profiling_info): Likewise.
1497 (ipcp_propagate_stage): Likewise.
1498 (estimate_local_effects): Likewise.
1499 (good_cloning_opportunity_p): Likewise.
1500 (devirtualization_time_bonus): Likewise.
1501 (propagate_constants_accross_call): Likewise.
1502 (initialize_node_lattices): Likewise.
1503 (ipcp_cloning_candidate_p): Likewise.
1504 (determine_versionability): Likewise.
1505 (print_all_lattices): Likewise.
1506 (print_lattice): Likewise.
1507 (ipcp_discover_new_direct_edges): Likewise.
1508 * ipa-devirt.c (ipa_devirt): Likewise.
1509 (likely_target_p): Likewise.
1510 (update_type_inheritance_graph): Likewise.
1511 (possible_polymorphic_call_target_p): Likewise.
1512 (dump_possible_polymorphic_call_targets): Likewise.
1513 (devirt_variable_node_removal_hook): Likewise.
1514 (record_binfo): Likewise.
1515 (maybe_record_node): Likewise.
1516 (build_type_inheritance_graph): Likewise.
1517 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1518 (inline_generate_summary): Likewise.
1519 (inline_analyze_function): Likewise.
1520 (do_estimate_growth): Likewise.
1521 (simple_edge_hints): Likewise.
1522 (estimate_node_size_and_time): Likewise.
1523 (estimate_edge_devirt_benefit): Likewise.
1524 (compute_inline_parameters): Likewise.
1525 (estimate_function_body_sizes): Likewise.
1526 (compute_bb_predicates): Likewise.
1527 (initialize_inline_failed): Likewise.
1528 (dump_inline_summary): Likewise.
1529 (dump_inline_edge_summary): Likewise.
1530 * ipa-inline-transform.c (inline_transform): Likewise.
1531 (preserve_function_body_p): Likewise.
1532 (save_inline_function_body): Likewise.
1533 (inline_call): Likewise.
1534 (clone_inlined_nodes): Likewise.
1535 (can_remove_node_now_p): Likewise.
1536 (can_remove_node_now_p_1): Likewise.
1537 * ipa-inline.c (early_inliner): Likewise.
1538 (early_inline_small_functions): Likewise.
1539 (inline_always_inline_functions): Likewise.
1540 (ipa_inline): Likewise.
1541 (flatten_function): Likewise.
1542 (inline_small_functions): Likewise.
1543 (speculation_useful_p): Likewise.
1544 (recursive_inlining): Likewise.
1545 (update_caller_keys): Likewise.
1546 (reset_edge_caches): Likewise.
1547 (update_edge_key): Likewise.
1548 (edge_badness): Likewise.
1549 (relative_time_benefit): Likewise.
1550 (want_inline_self_recursive_call_p): Likewise.
1551 (want_inline_small_function_p): Likewise.
1552 (want_early_inline_function_p): Likewise.
1553 (num_calls): Likewise.
1554 (can_early_inline_edge_p): Likewise.
1555 (can_inline_edge_p): Likewise.
1556 (report_inline_failed_reason): Likewise.
1557 * ipa-profile.c (ipa_profile): Likewise.
1558 (ipa_propagate_frequency): Likewise.
1559 (ipa_propagate_frequency_1): Likewise.
1560 (ipa_profile_generate_summary): Likewise.
1561 * ipa-prop.c (ipcp_transform_function): Likewise.
1562 (read_replacements_section): Likewise.
1563 (ipa_prop_read_section): Likewise.
1564 (ipa_modify_call_arguments): Likewise.
1565 (ipa_print_node_params): Likewise.
1566 (propagate_controlled_uses): Likewise.
1567 (update_indirect_edges_after_inlining): Likewise.
1568 (remove_described_reference): Likewise.
1569 (ipa_make_edge_direct_to_target): Likewise.
1570 (ipa_analyze_node): Likewise.
1571 (ipa_analyze_params_uses): Likewise.
1572 (ipa_compute_jump_functions): Likewise.
1573 (ipa_get_callee_param_type): Likewise.
1574 (ipa_print_node_jump_functions): Likewise.
1575 (ipa_initialize_node_params): Likewise.
1576 (ipa_populate_param_decls): Likewise.
1577 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1578 (write_agg_replacement_chain): Likewise.
1579 (ipa_write_node_info): Likewise.
1580 (ipa_edge_duplication_hook): Likewise.
1581 (try_decrement_rdesc_refcount): Likewise.
1582 * ipa-pure-const.c (propagate_nothrow): Likewise.
1583 (propagate_pure_const): Likewise.
1584 (pure_const_read_summary): Likewise.
1585 (pure_const_write_summary): Likewise.
1586 (analyze_function): Likewise.
1587 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1588 (ipa_ref_referring_ref_list): Likewise.
1589 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1590 (ipa_remove_stmt_references): Likewise.
1591 (ipa_find_reference): Likewise.
1592 (ipa_dump_referring): Likewise.
1593 (ipa_dump_references): Likewise.
1594 (ipa_record_reference): Likewise.
1595 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1596 (ipa_reference_write_optimization_summary): Likewise.
1597 (write_node_summary_p): Likewise.
1598 (propagate): Likewise.
1599 (read_write_all_from_decl): Likewise.
1600 (generate_summary): Likewise.
1601 (analyze_function): Likewise.
1602 (propagate_bits): Likewise.
1603 (ipa_reference_get_not_written_global): Likewise.
1604 (ipa_reference_get_not_read_global): Likewise.
1605 * ipa-split.c (execute_split_functions): Likewise.
1606 (split_function): Likewise.
1607 * ipa-utils.c (ipa_merge_profiles): Likewise.
1608 (dump_cgraph_node_set): Likewise.
1609 (ipa_reverse_postorder): Likewise.
1610 (ipa_edge_within_scc): Likewise.
1611 (ipa_get_nodes_in_cycle): Likewise.
1612 (ipa_free_postorder_info): Likewise.
1613 (ipa_reduced_postorder): Likewise.
1614 (searchc): Likewise.
1615 (recursive_call_p): Likewise.
1616 * ipa.c (ipa_cdtor_merge): Likewise.
1617 (record_cdtor_fn): Likewise.
1618 (function_and_variable_visibility): Likewise.
1619 (varpool_externally_visible_p): Likewise.
1620 (cgraph_externally_visible_p): Likewise.
1621 (comdat_can_be_unshared_p): Likewise.
1622 (comdat_can_be_unshared_p_1): Likewise.
1623 (address_taken_from_non_vtable_p): Likewise.
1624 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1625 (symtab_remove_unreachable_nodes): Likewise.
1626 (walk_polymorphic_call_targets): Likewise.
1627 (process_references): Likewise.
1628 (enqueue_node): Likewise.
1629 (has_addr_references_p): Likewise.
1630 (cgraph_non_local_node_p_1): Likewise.
1631 * is-a.h (varpool_analyze_node): Likewise.
1632 * lto-cgraph.c (input_symtab): Likewise.
1633 (merge_profile_summaries): Likewise.
1634 (input_cgraph_1): Likewise.
1635 (input_edge): Likewise.
1636 (input_varpool_node): Likewise.
1637 (input_node): Likewise.
1638 (input_overwrite_node): Likewise.
1639 (compute_ltrans_boundary): Likewise.
1640 (output_refs): Likewise.
1641 (lto_output_varpool_node): Likewise.
1642 (lto_output_node): Likewise.
1643 (reachable_from_other_partition_p): Likewise.
1644 (referenced_from_other_partition_p): Likewise.
1645 (lto_output_edge): Likewise.
1646 (output_node_opt_summary): Likewise.
1647 (add_node_to): Likewise.
1648 (reachable_from_this_partition_p): Likewise.
1649 (lto_set_symtab_encoder_in_partition): Likewise.
1650 (lto_symtab_encoder_in_partition_p): Likewise.
1651 (lto_set_symtab_encoder_encode_initializer): Likewise.
1652 (lto_symtab_encoder_encode_initializer_p): Likewise.
1653 (lto_set_symtab_encoder_encode_body): Likewise.
1654 (lto_symtab_encoder_encode_body_p): Likewise.
1655 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1657 * lto-streamer-in.c (lto_read_body): Likewise.
1658 (fixup_call_stmt_edges): Likewise.
1659 (fixup_call_stmt_edges_1): Likewise.
1660 * lto-streamer-out.c (produce_symtab): Likewise.
1661 (output_symbol_p): Likewise.
1662 (write_symbol): Likewise.
1663 (lto_output): Likewise.
1664 (copy_function): Likewise.
1665 (output_function): Likewise.
1666 * passes.c (function_called_by_processed_nodes_p): Likewise.
1667 (ipa_write_optimization_summaries): Likewise.
1668 (ipa_write_summaries): Likewise.
1669 (do_per_function_toporder): Likewise.
1670 (do_per_function): Likewise.
1671 (dump_passes): Likewise.
1672 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1673 (symtab_nonoverwritable_alias): Likewise.
1674 (symtab_nonoverwritable_alias_1): Likewise.
1675 (symtab_for_node_and_aliases): Likewise.
1676 (symtab_resolve_alias): Likewise.
1677 (fixup_same_cpp_alias_visibility): Likewise.
1678 (symtab_alias_ultimate_target): Likewise.
1679 (symtab_used_from_object_file_p): Likewise.
1680 (verify_symtab_base): Likewise.
1681 (dump_symtab_base): Likewise.
1682 (symtab_node_name): Likewise.
1683 (symtab_node_asm_name): Likewise.
1684 (symtab_dissolve_same_comdat_group_list): Likewise.
1685 (symtab_add_to_same_comdat_group): Likewise.
1686 (symtab_unregister_node): Likewise.
1687 (symtab_insert_node_to_hashtable): Likewise.
1688 (symtab_register_node): Likewise.
1689 (unlink_from_assembler_name_hash): Likewise.
1690 (insert_to_assembler_name_hash): Likewise.
1691 (eq_assembler_name): Likewise.
1692 (hash_node_by_assembler_name): Likewise.
1693 (eq_node): Likewise.
1694 (hash_node): Likewise.
1695 * toplev.c (wrapup_global_declaration_2): Likewise.
1696 * trans-mem.c (ipa_tm_execute): Likewise.
1697 (ipa_tm_transform_clone): Likewise.
1698 (ipa_tm_transform_transaction): Likewise.
1699 (ipa_tm_transform_calls_redirect): Likewise.
1700 (ipa_tm_insert_gettmclone_call): Likewise.
1701 (ipa_tm_insert_irr_call): Likewise.
1702 (ipa_tm_create_version): Likewise.
1703 (ipa_tm_create_version_alias): Likewise.
1704 (ipa_tm_mark_forced_by_abi_node): Likewise.
1705 (ipa_tm_mark_force_output_node): Likewise.
1706 (ipa_tm_diagnose_tm_safe): Likewise.
1707 (ipa_tm_mayenterirr_function): Likewise.
1708 (ipa_tm_scan_irr_function): Likewise.
1709 (ipa_tm_note_irrevocable): Likewise.
1710 (ipa_tm_scan_calls_clone): Likewise.
1711 (get_cg_data): Likewise.
1712 * tree-eh.c (tree_could_trap_p): Likewise.
1713 * tree-emutls.c (ipa_lower_emutls): Likewise.
1714 (create_emultls_var): Likewise.
1715 (lower_emutls_function_body): Likewise.
1716 (gen_emutls_addr): Likewise.
1717 (emutls_decl): Likewise.
1718 (new_emutls_decl): Likewise.
1719 * tree-inline.c (tree_function_versioning): Likewise.
1720 (optimize_inline_calls): Likewise.
1721 (expand_call_inline): Likewise.
1722 (estimate_num_insns): Likewise.
1723 (copy_bb): Likewise.
1724 (delete_unreachable_blocks_update_callgraph): Likewise.
1725 * tree-nested.c (gimplify_all_functions): Likewise.
1726 (create_nesting_tree): Likewise.
1727 (check_for_nested_with_variably_modified): Likewise.
1728 * tree-pretty-print.c (dump_function_header): Likewise.
1729 * tree-profile.c (tree_profiling): Likewise.
1730 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1731 (modify_function): Likewise.
1732 (convert_callers): Likewise.
1733 (convert_callers_for_node): Likewise.
1734 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1735 (associate_varinfo_to_alias): Likewise.
1736 (create_variable_info_for): Likewise.
1737 (get_constraint_for_ssa_var): Likewise.
1738 * tree-vectorizer.c (increase_alignment): Likewise.
1739 * tree.c (find_decls_types_in_var): Likewise.
1740 (find_decls_types_in_node): Likewise.
1741 (free_lang_data_in_decl): Likewise.
1742 * value-prof.c (gimple_ic_transform): Likewise.
1743 (gimple_ic): Likewise.
1744 (check_ic_target): Likewise.
1745 (init_node_map): Likewise.
1746 * varasm.c (decl_binds_to_current_def_p): Likewise.
1747 (default_binds_local_p_1): Likewise.
1748 (dump_tm_clone_pairs): Likewise.
1749 (assemble_alias): Likewise.
1750 (find_decl): Likewise.
1751 (mark_decl_referenced): Likewise.
1752 * varpool.c (varpool_for_node_and_aliases): Likewise.
1753 (varpool_extra_name_alias): Likewise.
1754 (varpool_create_variable_alias): Likewise.
1755 (add_new_static_var): Likewise.
1756 (varpool_finalize_named_section_flags): Likewise.
1757 (varpool_remove_unreferenced_decls): Likewise.
1758 (enqueue_node): Likewise.
1759 (varpool_assemble_decl): Likewise.
1760 (assemble_aliases): Likewise.
1761 (varpool_analyze_node): Likewise.
1762 (cgraph_variable_initializer_availability): Likewise.
1763 (varpool_add_new_variable): Likewise.
1764 (ctor_for_folding): Likewise.
1765 (dump_varpool_node): Likewise.
1766 (varpool_remove_initializer): Likewise.
1767 (varpool_remove_node): Likewise.
1768 (varpool_node_for_decl): Likewise.
1769 (varpool_create_empty_node): Likewise.
1770 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1771 (ix86_get_function_versions_dispatcher): Likewise.
1773 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1775 * cgraph.h (symtab_node_base): Convert to a class;
1776 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1777 chain_next/prev from symtab_node_def.
1778 (cgraph_node): Inherit from symtab_node; add GTY option
1779 tag ("SYMTAB_FUNCTION").
1780 (varpool_node): Inherit from symtab_node; add GTY option
1781 tag ("SYMTAB_VARIABLE").
1782 (symtab_node_def): Remove.
1783 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1784 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1785 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1786 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1788 * ipa-ref.h (symtab_node_def): Drop.
1789 (symtab_node): Change underlying type from symtab_node_def to
1791 (const_symtab_node): Likwise.
1793 * is-a.h: Update examples in comment.
1795 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1796 (assembler_name_hash): Likewise.
1798 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1800 * doc/tree-ssa.texi (gimple_phi_result): Document.
1801 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1802 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1803 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1804 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1806 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1808 * expr.h: Revert change and include tree-core.h.
1809 * rtl.h: Revert change and don't include tree-core.h.
1811 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1813 * config/darwin.c: Include gimple.h.
1814 * config/i386/winnt.c: Likewise.
1816 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1818 PR tree-optimization/19831
1819 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1821 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1823 * tree-outof-ssa.h: Remove include files.
1824 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1826 * tree-ssa-coalesce.c: Likewise.
1827 * cfgexpand.c: Likewise.
1828 * tree-ssa-ter.c: Likewise.
1829 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1830 * lto-streamer.h: Likewise.
1831 * cgraphbuild.c: Add gimple.h to include list.
1832 * data-streamer-in.c: Likewise.
1833 * ipa-cp.c: Likewise.
1834 * tree-streamer.c: Likewise.
1835 * lto-compress.c: Likewise.
1836 * ipa-reference.c: Likewise.
1837 * data-streamer-out.c: Likewise.
1838 * lto-cgraph.c: Likewise.
1839 * cgraphclones.c: Likewise.
1840 * ipa-utils.c: Likewise.
1841 * data-streamer.c: Likewise.
1842 * ipa-split.c: Likewise.
1843 * lto-section-in.c: Likewise.
1844 * tree-streamer-out.c: Likewise.
1845 * ipa-prop.c: Likewise.
1846 * tree-streamer-in.c: Likewise.
1847 * symtab.c: Likewise.
1848 * opts-global.c: Likewise.
1849 * lto-opts.c: Likewise.
1850 * lto-section-out.c: Likewise.
1851 * lto-streamer.c: Likewise.
1852 * rtl.h: Add tree-core.h to include list.
1853 * expr.h: Remove tree-core.h from include list.
1854 * gimple.h: Likewise.
1855 * ipa-utils.h: Likewise.
1856 * streamer-hooks.h: Likewise.
1857 * streamer-hooks.c: Include input.h.
1859 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1861 * config/arm/arm.c (cortexa7_extra_costs): New table.
1862 (arm_cortex_a7_tune): New.
1863 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1865 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1867 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1869 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1871 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1872 to use autogenerated markers for a class-hierarchy, every class
1873 must have a GTY marker.
1874 * gengtype.h (struct type): Add linked list of subclasses to
1875 the "s" member of the union.
1876 (add_subclass): New decl.
1877 * gengtype-state.c (read_state_struct_type): Set up subclass
1879 * gengtype.c (get_ultimate_base_class): New.
1880 (add_subclass): New.
1881 (new_structure): Set up subclass linked list.
1882 (set_gc_used_type): Propagate usage information to subclasses.
1883 (output_mangled_typename): Use get_ultimate_base_class.
1884 (walk_subclasses): Use the subclass linked list, avoiding an
1885 O(N^2) when writing out all types.
1886 (walk_type): Issue an error if the base class is missing a tag,
1887 rather than generating bogus C code. Add a gcc_unreachable
1888 default case, in case people omit tags from concrete subclasses,
1889 or get the values wrong.
1890 (write_func_for_structure): Issue an error for subclasses for
1891 which the base doesn't have a "desc", since otherwise the
1892 autogenerated routines for the base would silently fail to visit
1893 any subclass fields.
1894 (write_root): Use get_ultimate_base_class, tweaking constness of
1895 tp to match that function's signature.
1897 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1899 * doc/gty.texi (GTY Options): Add note about inheritance to
1900 description of desc and tag.
1901 (Inheritance and GTY): New.
1903 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1905 * gengtype-parse.c (opts_have): Drop "static" so that
1906 we can use this from gengtype.c.
1907 * gengtype.c (set_gc_used_type): Mark any base class as used;
1908 update field traversal to visit inherited fields.
1909 (output_mangled_typename): Convert references to classes within
1910 an inheritance hierarchy to reference the ultimate base class,
1911 since only it will have gt_ functions.
1912 (get_string_option): New.
1913 (walk_subclasses): New.
1914 (walk_type): Treat GTY structs that have a "desc" as being the
1915 root of an inheritance hierarchy. Generate a switch on it
1916 within the marking function which walks all subclasses, adding
1917 cases for them via walk_subclasses. For subclasses, visit all
1918 fields of the type (including inherited ones).
1919 (write_func_for_structure): Don't write fns for subclasses, only
1920 for the ultimate base class within an inheritance hierarchy.
1921 Subclasses-marking will be handled by the base class marking functions.
1922 (write_types): Likewise.
1923 (write_local_func_for_structure): Likewise.
1924 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1925 a "tag" option (and are thus concrete subclasses).
1926 (write_root): Use the marker function for the ultimate base class.
1927 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1928 (opts_have): Add declaration.
1930 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1932 * lra-spills.c (lra_final_code_change): Remove useless move insns
1933 originated from moves of pseudos.
1935 2013-10-28 Jeff Law <law@redhat.com>
1937 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1939 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1941 * df-scan.c (df_collection_rec): Adjust.
1942 (copy_defs): New constant.
1943 (copy_uses): Likewise.
1944 (copy_eq_uses): Likewise.
1945 (copy_mw): Likewise.
1946 (copy_all): Likewise.
1947 (df_insn_rescan): Adjust.
1948 (df_notes_rescan): Likewise.
1949 (df_swap_refs): Likewise.
1950 (df_sort_and_compress_refs): Likewise.
1951 (df_sort_and_compress_mws): Likewise.
1952 (df_install_refs): Likewise.
1953 (df_install_mws): Likewise.
1954 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1956 (df_bb_refs_record): Adjust.
1957 (df_record_entry_block_defs): Likewise.
1958 (df_record_exit_block_defs): Likewise.
1959 (df_refs_verify): Likewise.
1960 (df_mws_verify): Likewise.
1961 (df_insn_refs_verify): Likewise.
1962 (df_bb_verify): Likewise.
1963 * ipa-pure-const.c (finish_state): Remove.
1964 (propagate): Adjust.
1965 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1966 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1967 var-tracking.c: Adjust.
1968 * vec.c (stack_vecs): Remove.
1969 (register_stack_vec): Likewise.
1970 (stack_vec_register_index): Likewise.
1971 (unregister_stack_vec): Likewise.
1972 * vec.h (struct va_stack): Remove.
1973 (struct vec<T, A, vl_ptr>): Specialize as
1974 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1975 allocation strategy compatable with the vl_ptr layout.
1976 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1977 specialization anyway.
1978 (class stack_vec): New class.
1979 (vec_stack_alloc): Remove.
1980 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1982 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1983 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1984 Sergey Lega <sergey.s.lega@intel.com>
1985 Anna Tikhonova <anna.tikhonova@intel.com>
1986 Ilya Tocar <ilya.tocar@intel.com>
1987 Andrey Turetskiy <andrey.turetskiy@intel.com>
1988 Ilya Verbin <ilya.verbin@intel.com>
1989 Kirill Yukhin <kirill.yukhin@intel.com>
1990 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1992 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1993 (*prefetch_avx512pf_<mode>): New.
1994 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1995 (avx512f_maskcmp<mode>3): Ditto.
1996 (vashrv16si3): Ditto.
1998 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1999 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2000 Sergey Lega <sergey.s.lega@intel.com>
2001 Anna Tikhonova <anna.tikhonova@intel.com>
2002 Ilya Tocar <ilya.tocar@intel.com>
2003 Andrey Turetskiy <andrey.turetskiy@intel.com>
2004 Ilya Verbin <ilya.verbin@intel.com>
2005 Kirill Yukhin <kirill.yukhin@intel.com>
2006 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2008 * config/i386/i386.md (any_truncate): New.
2009 (trunsuffix): Ditto.
2010 * config/i386/predicates.md (const_8_to_9_operand): New.
2011 (const_10_to_11_operand): Ditto.
2012 (const_12_to_13_operand): Ditto.
2013 (const_14_to_15_operand): Ditto.
2014 (const_16_to_19_operand): Ditto.
2015 (const_20_to_23_operand): Ditto.
2016 (const_24_to_27_operand): Ditto.
2017 (const_28_to_31_operand): Ditto.
2018 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2019 (cvtusi2<ssescalarmodesuffix>32): New.
2020 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2021 (ufloatv16siv16sf2): Ditto.
2022 (avx512f_fix_notruncv16sfv16si): Ditto.
2023 (avx512f_ufix_notruncv16sfv16si): Ditto.
2024 (avx512f_vcvtss2usi): Ditto.
2025 (avx512f_vcvtss2usiq): Ditto.
2026 (avx512f_vcvttss2usi): Ditto.
2027 (avx512f_vcvttss2usiq): Ditto.
2028 (avx512f_vcvtsd2usi): Ditto.
2029 (avx512f_vcvtsd2usiq): Ditto.
2030 (avx512f_vcvttsd2usi): Ditto.
2031 (avx512f_vcvttsd2usiq): Ditto.
2032 (ufloatv8siv8df): Ditto.
2033 (avx512f_cvtdq2pd512_2): Ditto.
2034 (avx512f_cvtpd2dq512): Ditto.
2035 (avx512f_ufix_notruncv8dfv8si): Ditto.
2036 (avx512f_cvtpd2ps512): Ditto.
2037 (vec_unpacks_lo_v16sf): Ditto.
2038 (vec_unpacks_hi_v16sf): Ditto.
2039 (vec_unpacks_float_hi_v16si): Ditto.
2040 (vec_unpacks_float_lo_v16si): Ditto.
2041 (avx512f_unpckhps512): Ditto.
2042 (avx512f_unpcklps512): Ditto.
2043 (avx512f_movshdup512): Ditto.
2044 (avx512f_movsldup512): Ditto.
2045 (vec_extract_lo_v32hi): Ditto.
2046 (vec_extract_hi_v32hi): Ditto.
2047 (vec_extract_lo_v64qi): Ditto.
2048 (vec_extract_hi_v64qi): Ditto.
2049 (avx512f_unpckhpd512): Ditto.
2050 (avx512f_movddup512): Ditto.
2051 (avx512f_unpcklpd512): Ditto.
2052 (*avx512f_unpcklpd512): Ditto.
2053 (avx512f_shufps512_1): Ditto.
2054 (avx512f_shufpd512_1): Ditto.
2055 (avx512f_interleave_highv8di): Ditto.
2056 (avx512f_interleave_lowv8di): Ditto.
2057 (PMOV_DST_MODE): Ditto.
2058 (pmov_src_mode): Ditto.
2059 (pmov_src_lower): Ditto.
2061 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2062 (*avx512f_<code>v8div16qi2): Ditto.
2063 (*avx512f_<code>v8div16qi2_store): Ditto.
2064 (vec_widen_umult_even_v16si): Ditto.
2065 (*vec_widen_umult_even_v16si): Ditto.
2066 (vec_widen_smult_even_v16si): Ditto.
2067 (*vec_widen_smult_even_v16si): Ditto.
2068 (avx512f_interleave_highv16si): Ditto.
2069 (avx512f_interleave_lowv16si): Ditto.
2070 (avx512f_<code>v16qiv16si2): Ditto.
2071 (avx512f_<code>v16hiv16si2): Ditto.
2072 (avx512f_<code>v8qiv8di2): Ditto.
2073 (avx512f_<code>v8hiv8di2): Ditto.
2074 (avx512f_<code>v8siv8di2): Ditto.
2075 (avx512cd_maskb_vec_dupv8di): Ditto.
2076 (avx512cd_maskw_vec_dupv16si): Ditto.
2077 (avx512f_vcvtph2ps512): Ditto.
2078 (avx512f_vcvtps2ph512): Ditto.
2079 (VEC_EXTRACT_MODE): Extened with wider modes.
2080 (VEC_PERM_AVX2): Ditto.
2081 (VEC_PERM_CONST): Ditto.
2083 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2085 * config/arc/arc.c (arc_ccfsm_post_advance):
2086 Add comment about TYPE_RETURN.
2088 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2090 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2091 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2092 (strip_offset): Convert offset to unsigned number.
2094 2013-10-27 Tom de Vries <tom@codesourcery.com>
2096 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2097 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2099 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2101 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2102 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2103 (TARGET_BIG_ENDIAN): New macro.
2104 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2105 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2106 * config/sh/sh.c: Likewise.
2107 * config/sh/sh.h: Likewise.
2109 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2111 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2112 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2114 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2116 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2118 (cmpstr_t, cmpstrsi): Fix formatting.
2120 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2123 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2124 addressing, do not use general_operand for memory operands.
2126 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2129 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2130 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2132 2013-10-26 Jeff Law <law@redhat.com>
2134 * predict.c (PRED_MUDFLAP): Remove.
2135 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2137 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2138 (OBJS): Remove tree-nomudflap.o
2139 (GTFILES): Remove tree-mudflap.c
2140 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2141 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2142 (mfwrap_spec, mflib_spec): Likewise.
2143 (cpp_unique_options, cc1_options, static_specs): Likewise.
2144 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2145 * passes.def: Likewise.
2146 * toplev.c (compile_file, process_options): Likewise.
2147 * tree-inline.c (copy_tree_r): Likewise.
2148 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2149 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2150 (build_constant_desc, output_constant_def_contents): Likewise.
2151 (categorize_decl_for_section): Likewise.
2152 * tree-mudflap.c: Removed.
2153 * tree-mudflap.h: Removed.
2154 * tree-nomudflap.c: Removed.
2155 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2156 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2157 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2158 * config/sol2.h (MFLIB_SPEC): Likewise.
2159 * doc/install.texi: Remove mudflap references.
2160 * doc/passes.texi: Similarly.
2161 * doc/sourcebuild.texi: Similarly.
2162 * doc/invoke.texi: Remove mudlfap related options.
2164 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2166 PR rtl-optimization/58759
2167 * lra-constraints.c (lra_constraints): Remove wrong condition to
2168 remove insn setting up an equivalent pseudo.
2170 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2172 * config/rs6000/rs6000-protos.h
2173 (rs6000_secondary_memory_needed_mode): New prototype.
2174 * config/rs6000/rs6000.c: Include ira.h.
2175 (TARGET_LRA_P): Redefine.
2176 (rs6000_legitimate_offset_address_p): Call
2177 legitimate_constant_pool_address_p in strict mode for LRA.
2178 (rs6000_legitimate_address_p): Ditto.
2179 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2180 (rs6000_emit_move): Add LRA version of code to generate load/store
2182 (rs6000_secondary_memory_needed_mode): New.
2183 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2184 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2185 constants, memory, and FP registers.
2186 (rs6000_lra_p): New.
2187 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2188 * config/rs6000/rs6000.opt (mlra): New option.
2189 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2191 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2193 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2194 has_single_use () and not do the conversion if has_single_use ()
2195 returns false for the multiplication result.
2197 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2199 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2200 to "NODE", since this works on a "tree", not an
2202 (CONSTANT_CLASS_P): Likewise.
2205 (INDIRECT_REF_P): Likewise.
2206 (REFERENCE_CLASS_P): Likewise.
2207 (COMPARISON_CLASS_P): Likewise.
2208 (UNARY_CLASS_P): Likewise.
2209 (BINARY_CLASS_P): Likewise.
2210 (STATEMENT_CLASS_P): Likewise.
2211 (VL_EXP_CLASS_P): Likewise.
2212 (EXPRESSION_CLASS_P): Likewise.
2213 (IS_TYPE_OR_DECL_P): Likewise.
2215 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2217 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2218 ADDR_EXPR in the defining statement.
2220 2013-10-25 Richard Biener <rguenther@suse.de>
2222 PR tree-optimization/58626
2223 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2224 anti_dd, output_dd and input_dd.
2225 (struct rdg_edge): Remove level and relation members.
2226 (RDGE_LEVEL, RDGE_RELATION): Remove.
2227 (dot_rdg_1): Adjust.
2228 (create_rdg_edge_for_ddr): Remove.
2229 (create_rdg_edges_for_scalar): Adjust.
2230 (create_edge_for_control_dependence): Likewise.
2231 (create_rdg_edges): Split into ...
2232 (create_rdg_flow_edges): ... this
2233 (create_rdg_cd_edges): ... and this.
2235 (build_rdg): Likewise, do not compute data dependences or
2237 (pg_add_dependence_edges): New function.
2239 (distribute_loop): First apply all non-dependence based
2240 partition mergings. Then compute dependences between partitions
2241 and merge and order partitions according to them.
2243 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2245 PR rtl-optimization/58831
2246 * alias.c (init_alias_analysis): At the beginning of each iteration,
2247 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2249 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2251 * recog.c (search_ofs): New static variable moved from...
2252 (peep2_find_free_register): ...here.
2253 (peephole2_optimize): Initialize it.
2255 2013-10-25 Tobias Burnus <burnus@net-b.de>
2257 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2259 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2261 * config/i386/i386.h (TARGET_MPX): New define.
2262 (TARGET_MPX_P): Ditto.
2264 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2266 * config/i386/constraints.md (B): New.
2269 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2270 * config/i386/i386-modes.def (BND32): New.
2272 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2273 * config/i386/i386.c (isa_opts): Add mmpx.
2274 (regclass_map): Add bound registers.
2275 (dbx_register_map): Likewise.
2276 (dbx64_register_map): Likewise.
2277 (svr4_dbx_register_map): Likewise.
2279 (ix86_option_override_internal): Support MPX ISA.
2280 (ix86_conditional_register_usage): Support bound registers.
2281 (print_reg): Likewise.
2282 (ix86_code_end): Add MPX bnd prefix.
2283 (output_set_got): Likewise.
2284 (ix86_output_call_insn): Likewise.
2285 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2286 (ix86_print_operand_punct_valid_p): Likewise.
2287 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
2289 (ix86_class_likely_spilled_p): Add bound regs support.
2290 (ix86_hard_regno_mode_ok): Likewise.
2291 (x86_order_regs_for_local_alloc): Likewise.
2292 (ix86_bnd_prefixed_insn_p): New.
2293 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
2294 (FIXED_REGISTERS): Add bound registers.
2295 (CALL_USED_REGISTERS): Likewise.
2296 (REG_ALLOC_ORDER): Likewise.
2297 (HARD_REGNO_NREGS): Likewise.
2299 (VALID_BND_REG_MODE): New.
2300 (FIRST_BND_REG): New.
2301 (LAST_BND_REG): New.
2302 (reg_class): Add BND_REGS.
2303 (REG_CLASS_NAMES): Likewise.
2304 (REG_CLASS_CONTENTS): Likewise.
2306 (ANY_BND_REG_P): New.
2308 (HI_REGISTER_NAMES): Add bound registers.
2309 * config/i386/i386.md (UNSPEC_BNDMK): New.
2310 (UNSPEC_BNDMK_ADDR): New.
2311 (UNSPEC_BNDSTX): New.
2312 (UNSPEC_BNDLDX): New.
2313 (UNSPEC_BNDLDX_ADDR): New.
2314 (UNSPEC_BNDCL): New.
2315 (UNSPEC_BNDCU): New.
2316 (UNSPEC_BNDCN): New.
2317 (UNSPEC_MPX_FENCE): New.
2320 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
2321 (length_immediate): Likewise.
2322 (prefix_0f): Likewise.
2324 (prefix_rep): Check for bnd prefix.
2325 (length_nobnd): New.
2326 (length): Use length_nobnd if specified.
2331 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
2334 (simple_return_internal): Likewise.
2335 (simple_return_pop_internal): Likewise.
2336 (*indirect_jump): Add MPX bnd prefix.
2337 (*tablejump_1): Likewise.
2338 (simple_return_internal_long): Likewise.
2339 (simple_return_indirect_internal): Likewise.
2343 (*mov<mode>_internal_mpx): New.
2344 (<mode>_<bndcheck>): New.
2345 (*<mode>_<bndcheck>): New.
2350 * config/i386/predicates.md (lea_address_operand): Rename to...
2351 (address_no_seg_operand): ... this.
2352 (address_mpx_no_base_operand): New.
2353 (address_mpx_no_index_operand): New.
2354 (bnd_mem_operator): New.
2355 * config/i386/i386.opt (mmpx): New.
2356 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
2357 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
2359 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2361 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2362 * tree.def (POINTER_BOUNDS_TYPE): New.
2363 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2364 (POINTER_BOUNDS_MODE): New.
2365 (make_pointer_bounds_mode): New.
2366 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2367 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2368 (layout_type): Support POINTER_BOUNDS_TYPE.
2369 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2370 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2371 (type_contains_placeholder_1): Likewise.
2372 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2373 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2374 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2376 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2378 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2379 mode of address computation when calling memory_address_addr_space.
2381 2013-08-24 Richard Henderson <rth@twiddle.net>
2384 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2385 instead of create_convert_operand_to.
2386 (maybe_emit_sync_lock_test_and_set): Likewise.
2387 (expand_atomic_compare_and_swap): Likewise.
2388 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2390 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2392 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2393 Change param_is to use the struct and not the pointer to the struct.
2395 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2397 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2398 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2399 Move to targhooks.c.
2400 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2401 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2402 Relocate from gimplify.c.
2403 * targhooks.h: Add 2 prototypes.
2404 * tree.h: Delete 2 prototypes.
2406 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2408 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2409 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2411 2013-10-24 Cong Hou <congh@google.com>
2413 * convert.c (convert_to_real): Guard those unsafe math function
2414 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2417 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2420 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2422 (cgraph_create_edge): Use the new argument.
2423 (cgraph_create_indirect_edge): Likewise.
2425 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2427 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2429 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2430 changes statep->state.
2432 2013-10-24 Tobias Burnus <burnus@net-b.de>
2435 * tree-cfg.c (replace_loop_annotate): New function.
2436 (execute_build_cfg): Call it.
2437 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2438 * internal-fn.c (expand_ANNOTATE): New function.
2439 * internal-fn.def (ANNOTATE): Define as new internal function.
2440 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2441 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2442 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2443 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2444 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2446 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2447 Marcus Shawcroft <marcus.shawcroft@arm.com>
2449 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2450 Special case reload SP+C into none GENERAL_REGS.
2452 2013-10-24 Michael Matz <matz@suse.de>
2454 * gengtype.c (is_file_equal): Check that files will be same length.
2456 2013-10-25 Christian Bruel <christian.bruel@st.com>
2458 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2459 * config/sh/t-sh (sh-mem.o): New rule.
2460 * config/sh/sh-mem.cc (expand_block_move): Moved here.
2461 (sh_expand_cmpstr): New function.
2462 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2463 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2464 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2465 (rotlhi3_8): Rename.
2467 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2469 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2470 * Makefile.in (CXXDEPMODE): Assign and change users.
2471 (CCDEPMODE): Delete.
2472 * configure: Regenerate.
2474 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2476 * gengtype-parse.c (require_without_advance): New.
2477 (type): For GTY-marked types that are not GTY((user)), parse any
2478 base classes, requiring them to be single-inheritance, and not
2479 be templates. For non-GTY-marked types and GTY((user)),
2480 continue to skip over any C++ inheritance specification.
2481 * gengtype-state.c (state_writer::write_state_struct_type):
2482 Write base class of type (if any).
2483 (read_state_struct_type): Read base class of type (if any).
2484 * gengtype.c (new_structure): Add a "base_class" parameter.
2485 (create_optional_field_): Update for new parameter to new_structure.
2486 (adjust_field_rtx_def): Likewise.
2487 (adjust_field_tree_exp): Likewise.
2488 * gengtype.h (struct type): Add "base_class" field to the s
2490 (new_structure): Add "base" parameter.
2492 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2495 * config/i386/i386.c (ix86_option_override_internal):
2496 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2497 (ix86_valid_target_attribute_tree):
2498 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2499 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2501 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2503 * tree-ssa-loop.h: Remove include files.
2504 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2505 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2506 * graphite-clast-to-gimple.c: Likewise.
2507 * graphite-scop-detection.c: Likewise.
2508 * graphite-sese-to-poly.c: Likewise.
2509 * ipa-inline-analysis.c: Likewise.
2510 * ipa-pure-const.c: Likewise.
2511 * loop-init.c: Likewise.
2512 * passes.c: Likewise.
2513 * predict.c: Likewise.
2514 * tree-cfg.c: Likewise.
2515 * tree-cfgcleanup.c: Likewise.
2516 * tree-chrec.c: Likewise.
2517 * tree-data-ref.c: Likewise.
2518 * tree-loop-distribution.c: Likewise.
2519 * tree-parloops.c: Likewise.
2520 * tree-predcom.c: Likewise.
2521 * tree-scalar-evolution.c: Likewise.
2522 * tree-ssa-address.c: Likewise.
2523 * tree-ssa.c: Likewise.
2524 * tree-ssa-dce.c: Likewise.
2525 * tree-ssa-loop.c: Likewise.
2526 * tree-ssa-loop-im.c: Likewise.
2527 * tree-ssa-loop-ivcanon.c: Likewise.
2528 * tree-ssa-loop-ivopts.c: Likewise.
2529 * tree-ssa-loop-manip.c: Likewise.
2530 * tree-ssa-loop-niter.c: Likewise.
2531 * tree-ssa-loop-prefetch.c: Likewise.
2532 * tree-ssa-loop-unswitch.c: Likewise.
2533 * tree-ssa-reassoc.c: Likewise.
2534 * tree-vect-data-refs.c: Likewise.
2535 * tree-vect-loop.c: Likewise.
2536 * tree-vect-loop-manip.c: Likewise.
2537 * tree-vectorizer.c: Likewise.
2538 * tree-vect-stmts.c: Likewise.
2539 * tree-vrp.c: Likewise.
2541 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2543 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2545 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2547 PR tree-optimization/58775
2548 PR tree-optimization/58791
2549 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2550 (insert_stmt_after): Rewritten, don't move the stmt, but really
2552 (get_stmt_uid_with_default): Remove.
2553 (build_and_add_sum): Use insert_stmt_after and
2554 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2555 (update_range_test): Set uid on stmts added by
2556 force_gimple_operand_gsi. Don't immediately modify statements
2557 in inter-bb optimization, just update oe->op values.
2558 (optimize_range_tests): Return bool whether any changed have been made.
2559 (update_ops): New function.
2560 (struct inter_bb_range_test_entry): New type.
2561 (maybe_optimize_range_tests): Perform statement changes here.
2562 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2563 ensure_ops_are_available): Remove.
2564 (find_insert_point): Rewritten.
2565 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2566 return LHS of the (new resp. old) stmt. Don't call
2567 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2568 instead of last, move new stmt at the right place.
2569 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2570 (negate_value): Likewise. Set uids.
2571 (break_up_subtract_bb): Initialize uids.
2572 (reassociate_bb): Adjust rewrite_expr_tree caller.
2573 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2575 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2578 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2579 TARGET_32BIT final condition.
2580 (mulsi3_internal2 and splitter): Same.
2582 2013-10-23 Jeff Law <law@redhat.com>
2584 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2585 through joiner blocks with abnormal outgoing edges.
2587 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2588 Add parameter JOINERS, to allow/disallow threading through joiner
2590 (thread_block): New. Call thread_block_1.
2591 (mark_threaded_blocks): Remove code to filter out certain cases
2592 of threading through joiner blocks.
2593 (thread_through_all_blocks): Document how we can have a dangling
2594 edge AUX field and clear it.
2596 2013-10-23 Ian Lance Taylor <iant@google.com>
2598 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2599 (C++ Dialect Options): Likewise.
2600 (Optimize Options): Likewise.
2602 2013-10-23 Tom de Vries <tom@codesourcery.com>
2604 PR tree-optimization/58805
2605 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2607 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2609 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2610 sequence based on get_range_info returned range.
2612 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2614 * tree-ssa.h: Remove all #include's
2615 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2616 * alias.c: Move required includes from tree-ssa.h.
2618 * builtins.c: Likewise.
2619 * calls.c: Likewise.
2620 * cfgexpand.c: Likewise.
2621 * cfghooks.c: Likewise.
2622 * cfgloop.c: Likewise.
2623 * cfgloopmanip.c: Likewise.
2624 * cgraph.c: Likewise.
2625 * cgraphbuild.c: Likewise.
2626 * cgraphclones.c: Likewise.
2627 * cgraphunit.c: Likewise.
2629 * except.c: Likewise.
2631 * final.c: Likewise.
2632 * fold-const.c: Likewise.
2633 * ggc-page.c: Likewise.
2634 * gimple-builder.c: Likewise.
2635 * gimple-fold.c: Likewise.
2636 * gimple-iterator.c: Likewise.
2637 * gimple-low.c: Likewise.
2638 * gimple-pretty-print.c: Likewise.
2639 * gimple-ssa-strength-reduction.c: Likewise.
2640 * gimple-streamer-in.c: Likewise.
2641 * gimple-streamer-out.c: Likewise.
2642 * gimplify.c: Likewise.
2643 * graphite-blocking.c: Likewise.
2644 * graphite-clast-to-gimple.c: Likewise.
2645 * graphite-dependences.c: Likewise.
2646 * graphite-interchange.c: Likewise.
2647 * graphite-optimize-isl.c: Likewise.
2648 * graphite-poly.c: Likewise.
2649 * graphite-scop-detection.c: Likewise.
2650 * graphite-sese-to-poly.c: Likewise.
2651 * graphite.c: Likewise.
2652 * ipa-cp.c: Likewise.
2653 * ipa-inline-analysis.c: Likewise.
2654 * ipa-inline-transform.c: Likewise.
2655 * ipa-inline.c: Likewise.
2656 * ipa-prop.c: Likewise.
2657 * ipa-pure-const.c: Likewise.
2658 * ipa-reference.c: Likewise.
2659 * ipa-split.c: Likewise.
2660 * ipa-utils.c: Likewise.
2661 * loop-init.c: Likewise.
2662 * lto-cgraph.c: Likewise.
2663 * lto-section-in.c: Likewise.
2664 * lto-section-out.c: Likewise.
2665 * lto-streamer-in.c: Likewise.
2666 * lto-streamer-out.c: Likewise.
2667 * lto-streamer.c: Likewise.
2668 * omp-low.c: Likewise.
2669 * passes.c: Likewise.
2670 * predict.c: Likewise.
2671 * print-tree.c: Likewise.
2672 * profile.c: Likewise.
2674 * targhooks.c: Likewise.
2675 * tracer.c: Likewise.
2676 * trans-mem.c: Likewise.
2677 * tree-call-cdce.c: Likewise.
2678 * tree-cfg.c: Likewise.
2679 * tree-cfgcleanup.c: Likewise.
2680 * tree-chrec.c: Likewise.
2681 * tree-complex.c: Likewise.
2682 * tree-data-ref.c: Likewise.
2683 * tree-dfa.c: Likewise.
2684 * tree-eh.c: Likewise.
2685 * tree-emutls.c: Likewise.
2686 * tree-if-conv.c: Likewise.
2687 * tree-inline.c: Likewise.
2688 * tree-into-ssa.c: Likewise.
2689 * tree-loop-distribution.c: Likewise.
2690 * tree-mudflap.c: Likewise.
2691 * tree-nested.c: Likewise.
2692 * tree-nrv.c: Likewise.
2693 * tree-object-size.c: Likewise.
2694 * tree-outof-ssa.c: Likewise.
2695 * tree-parloops.c: Likewise.
2696 * tree-phinodes.c: Likewise.
2697 * tree-predcom.c: Likewise.
2698 * tree-pretty-print.c: Likewise.
2699 * tree-profile.c: Likewise.
2700 * tree-scalar-evolution.c: Likewise.
2701 * tree-sra.c: Likewise.
2702 * tree-ssa-address.c: Likewise.
2703 * tree-ssa-alias.c: Likewise.
2704 * tree-ssa-ccp.c: Likewise.
2705 * tree-ssa-coalesce.c: Likewise.
2706 * tree-ssa-copy.c: Likewise.
2707 * tree-ssa-copyrename.c: Likewise.
2708 * tree-ssa-dce.c: Likewise.
2709 * tree-ssa-dom.c: Likewise.
2710 * tree-ssa-dse.c: Likewise.
2711 * tree-ssa-forwprop.c: Likewise.
2712 * tree-ssa-ifcombine.c: Likewise.
2713 * tree-ssa-live.c: Likewise.
2714 * tree-ssa-loop-ch.c: Likewise.
2715 * tree-ssa-loop-im.c: Likewise.
2716 * tree-ssa-loop-ivcanon.c: Likewise.
2717 * tree-ssa-loop-ivopts.c: Likewise.
2718 * tree-ssa-loop-manip.c: Likewise.
2719 * tree-ssa-loop-niter.c: Likewise.
2720 * tree-ssa-loop-prefetch.c: Likewise.
2721 * tree-ssa-loop-unswitch.c: Likewise.
2722 * tree-ssa-loop.c: Likewise.
2723 * tree-ssa-math-opts.c: Likewise.
2724 * tree-ssa-operands.c: Likewise.
2725 * tree-ssa-phiopt.c: Likewise.
2726 * tree-ssa-phiprop.c: Likewise.
2727 * tree-ssa-pre.c: Likewise.
2728 * tree-ssa-propagate.c: Likewise.
2729 * tree-ssa-reassoc.c: Likewise.
2730 * tree-ssa-sccvn.c: Likewise.
2731 * tree-ssa-sink.c: Likewise.
2732 * tree-ssa-strlen.c: Likewise.
2733 * tree-ssa-structalias.c: Likewise.
2734 * tree-ssa-tail-merge.c: Likewise.
2735 * tree-ssa-ter.c: Likewise.
2736 * tree-ssa-threadedge.c: Likewise.
2737 * tree-ssa-threadupdate.c: Likewise.
2738 * tree-ssa-uncprop.c: Likewise.
2739 * tree-ssa-uninit.c: Likewise.
2740 * tree-ssa.c: Likewise.
2741 * tree-ssanames.c: Likewise.
2742 * tree-stdarg.c: Likewise.
2743 * tree-streamer-in.c: Likewise.
2744 * tree-switch-conversion.c: Likewise.
2745 * tree-tailcall.c: Likewise.
2746 * tree-vect-data-refs.c: Likewise.
2747 * tree-vect-generic.c: Likewise.
2748 * tree-vect-loop-manip.c: Likewise.
2749 * tree-vect-loop.c: Likewise.
2750 * tree-vect-patterns.c: Likewise.
2751 * tree-vect-slp.c: Likewise.
2752 * tree-vect-stmts.c: Likewise.
2753 * tree-vectorizer.c: Likewise.
2754 * tree-vrp.c: Likewise.
2757 * value-prof.c: Likewise.
2758 * var-tracking.c: Likewise.
2759 * varpool.c: Likewise.
2760 * vtable-verify.c: Likewise.
2762 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2764 * config/tilegx/tilegx.c: Include "tree.h".
2766 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2768 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2769 the info, not after it.
2770 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2771 dump_ssaname_info call.
2772 (pp_gimple_stmt_1): Adjust caller.
2773 (dump_phi_nodes): Likewise. Don't print "# " here.
2775 2013-10-22 Jan Hubicka <jh@suse.cz>
2777 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2779 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2780 * i386.c (expand_small_movmem_or_setmem): New function.
2781 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2783 (alg_usable_p): Add support for value ranges; cleanup.
2784 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2786 2013-10-22 Sterling Augustine <saugustine@google.com>
2788 * doc/invoke.texi: Document -ggnu-pubnames.
2789 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2791 * dwarf2out.c: Include gdb/gdb-index.h.
2792 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2793 debug_generate_pub_sections.
2794 (is_java, output_pubtables, output_pubname): New functions.
2795 (include_pubname_in_output): Handle debug_generate_pub_sections at
2797 (size_of_pubnames): Use new local space_for_flags based on
2798 debug_generate_pub_sections.
2799 (output_pubnames): Unify pubnames and pubtypes output logic.
2800 Genericize comments. Call output_pubname.
2801 (dwarf2out_finish): Move logic to output_pubnames and call it.
2803 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2806 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2807 Remove CCCmode handling.
2808 <case LTU>: Return 'c' suffix for CCCmode.
2809 <case GEU>: Return 'nc' suffix for CCCmode.
2810 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2811 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2812 (*sub<mode>3_cc_overflow): Ditto.
2813 (*subsi3_zext_cc_overflow): Ditto.
2815 2013-10-22 Steve Ellcey <sellcey@mips.com>
2817 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2818 (AND (NOT OP1) (NOT OP2)).
2820 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2822 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2823 meaning of merge-high and merge-low masks for little endian; avoid
2824 use of vector-pack masks for little endian for mismatched modes.
2826 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2828 * config/tilepro/tilepro.c: Include "tree.h".
2830 2013-10-22 Andreas Schwab <schwab@suse.de>
2832 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2835 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2837 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2838 [Ed Smith-Rowland]): New entries.
2839 ([Stephen M. Webb]): Update.
2841 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2843 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2845 2013-10-21 Marek Polacek <polacek@redhat.com>
2848 * fold-const.c (fold_range_test): Return 0 if the type is not
2851 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2853 * system.h: Move hwint.h include further down.
2854 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2855 gcc_checking_asserts.
2856 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2858 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2860 Fix volatile issues in optimize_bit_field_compare.
2861 * fold-const.c (optimize_bit_field_compare): Bail out if
2862 lvolatilep or rvolatilep.
2864 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2866 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2867 and get_inner_reference returning different pmode for non-volatile
2868 bit-field members dependent on flag_strict_volatile_bitfields.
2869 * stor-layout.c (layout_decl): Remove special handling of
2870 flag_strict_volatile_bitfields.
2871 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2872 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2874 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2876 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2879 2013-10-21 Jeff Law <law@redhat.com>
2881 * tree-ssa-threadedge.c (thread_through_normal_block): New
2882 argument VISITED. Remove VISISTED as a local variable. When we
2883 have a threadable jump, verify the destination of the jump has not
2885 (thread_across_edge): Allocate VISITED bitmap once at function
2886 scope and use it throughout. Make sure to set appropriate bits in
2887 VISITED for E (start of jump thread path).
2888 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2889 through a joiner if any edge on the path has a recorded jump thread.
2891 2013-10-21 Ian Lance Taylor <iant@google.com>
2893 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2894 don't imply that attributes can solve all problems.
2895 (Directory Options): Fix typo.
2897 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2899 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2900 extend and extend_add.
2902 2013-10-21 Richard Biener <rguenther@suse.de>
2904 PR tree-optimization/58794
2905 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2906 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2908 2013-10-21 Richard Biener <rguenther@suse.de>
2911 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2912 to (T) X for sign-changing conversions (or no conversion).
2914 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2916 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2918 2013-10-20 Jan Hubicka <jh@suse.cz>
2920 * config/i386/i386-tune.def: Add comment; organize into categories
2922 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2924 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2925 argument. Update function comment.
2926 (expand_set_or_movmem_via_rep): New function combining
2927 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2928 (expand_movmem_via_rep_mov): Remove.
2929 expand_setmem_via_rep_stos): Remove.
2930 (expand_movmem_epilogue): Update calls correspondingly.
2931 (expand_setmem_epilogue_via_loop): Likewise.
2933 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2934 (expand_set_or_movmem_prologue): New function combining
2935 expand_movmem_prologue and expand_setmem_prologue.
2936 (expand_movmem_prologue): Remove.
2937 (expand_setmem_prologue): Remove.
2938 (expand_set_or_movmem_constant_prologue): New function combining
2939 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2940 (expand_constant_movmem_prologue): Remove.
2941 (expand_constant_setmem_prologue): Remove.
2942 (promote_duplicated_reg): Allow vector-const0 value.
2943 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2944 and ix86_expand_setmem.
2945 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2946 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2948 2013-10-21 Diego Novillo <dnovillo@google.com>
2950 * asan.c: Include tree.h
2951 * bb-reorder.c: Likewise.
2952 * cfgcleanup.c: Likewise.
2953 * cfgloopmanip.c: Likewise.
2954 * data-streamer-in.c: Likewise.
2955 * data-streamer-out.c: Likewise.
2956 * data-streamer.c: Likewise.
2957 * dwarf2cfi.c: Likewise.
2958 * graphite-blocking.c: Likewise.
2959 * graphite-clast-to-gimple.c: Likewise.
2960 * graphite-dependences.c: Likewise.
2961 * graphite-interchange.c: Likewise.
2962 * graphite-optimize-isl.c: Likewise.
2963 * graphite-poly.c: Likewise.
2964 * graphite-scop-detection.c: Likewise.
2965 * graphite-sese-to-poly.c: Likewise.
2966 * graphite.c: Likewise.
2967 * ipa-devirt.c: Likewise.
2968 * ipa-profile.c: Likewise.
2971 * loop-init.c: Likewise.
2972 * loop-unroll.c: Likewise.
2973 * lower-subreg.c: Likewise.
2974 * lto/lto-object.c: Likewise.
2975 * recog.c: Likewise.
2976 * reginfo.c: Likewise.
2977 * tree-loop-distribution.c: Likewise.
2978 * tree-parloops.c: Likewise.
2979 * tree-ssa-strlen.c: Likewise.
2980 * tree-streamer.c: Likewise.
2981 * value-prof.c: Likewise.
2982 * target-globals.c: Likewise.
2983 * expr.h: Include tree-core.h instead of tree.h.
2984 * gimple.h: Likewise.
2985 * ipa-prop.h: Likewise.
2986 * ipa-utils.h: Likewise.
2987 * lto-streamer.h: Likewise.
2988 * streamer-hooks.h: Likewise.
2989 * ipa-reference.h: Include cgraph.h instead of tree.h.
2990 * cgraph.h: Include basic-block.h instead of tree.h.
2991 * tree-streamer.h: Do not include tree.h.
2992 * genattrtab.c (write_header): Generate inclusion of
2994 * genautomata.c (main): Likewise.
2995 * genemit.c: Likewise.
2996 * genopinit.c: Likewise.
2997 * genoutput.c (output_prologue): Likewise.
2998 * genpeep.c: Likewise.
3000 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3002 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3004 (vec_unpacku_hi_v8hi): Likewise.
3005 (vec_unpacku_lo_v16qi): Likewise.
3006 (vec_unpacku_lo_v8hi): Likewise.
3008 2013-10-20 Jan Hubicka <jh@suse.cz>
3010 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3011 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3012 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3014 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3016 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3017 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3019 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3021 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3024 * system.h: Undef m_slot.
3026 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3028 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3029 all elements for both endian flavors.
3031 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3034 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3035 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3036 and SI_REG for 64bit SYSV ABI targets.
3038 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3040 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3041 to multi_reg_return. Clarify that we are skipping USEs of multiple
3042 return registers. Use bool type where appropriate.
3044 2013-10-18 Jan Hubicka <jh@suse.cz>
3046 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3048 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3049 (X86_TUNE_PUSH_MEMORY): Likewise.
3050 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3051 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3052 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3054 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3055 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3057 (ix86_option_override_internal): Update to use tune features instead
3060 2013-10-18 Cong Hou <congh@google.com>
3062 PR tree-optimization/58508
3063 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3064 statement that contains data refs with zero-step.
3066 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3068 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3069 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3070 * gimple-low.c (gimple_check_call_arg,
3071 gimple_check_call_matching_types): Move to cgraph.c.
3072 * gimple-low.h: Remove prototype.
3073 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3074 Relocate from gimple-low.c.
3075 * cgraph.h: Add prototype. Don't include basic-block.h.
3076 * gimplify.c: Add gimple-low to include list.
3077 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3078 * tree-eh.c: Add gimple-low to include list.
3079 * tree-nested.c: Likewise.
3080 * cfgexpand.c: Add tree-ssa-address.h to include list.
3082 * gimple-fold.c: Likewise.
3083 * gimple-ssa-strength-reduction.c: Likewise.
3084 * trans-mem.c: Likewise.
3085 * tree-mudflap.c: Likewise.
3086 * tree-ssa-loop-ivopts.c: Likewise.
3087 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3088 (degenerate_phi_result): Move to tree-phinodes.c.
3089 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3090 * tree-ssa-threadedge.c: Likewise.
3091 * tree-vrp.c: Likewise.
3092 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3093 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3094 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3095 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3096 * tree-ssa-forwprop.c: Likewise.
3097 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3098 pass_data_cleanup_cfg_post_optimizing,
3099 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3100 * tree-optimize.c: Delete File.
3101 * graphite.c: Include tree-cfgcleanup.h.
3102 * passes.c: Likewise.
3103 * tree-cfg.c: Likewise.
3104 * tree-profile.c: Likewise.
3105 * tree-ssa-dse.c: Likewise.
3106 * tree-ssa-loop-ivcanon.c: Likewise.
3107 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3108 * tree-outof-ssa.c: Include sbitmap.h.
3109 * tree-ssa-live.c: Likewise.
3110 * tree-ssa-propagate.c: Likewise.
3111 * tree-ssa-structalias.c: Likewise.
3112 * tree-stdarg.c: Likewise.
3113 * Makefile.in (OBJS): Delete tree-optimize.o.
3114 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3115 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3116 * varasm.c: Include basic-block.h.
3117 * cfgloop.h: Include function.h instead of basic-block.h
3118 (bb_loop_depth): Move to cfgloop.c.
3119 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3121 2013-10-18 Teresa Johnson <tejohnson@google.com>
3123 * predict.c (probably_never_executed): Compare frequency-based
3124 count to number of training runs.
3125 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3127 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3129 * config/arm/arm.c (cortexa9_extra_costs): New table.
3130 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3132 2013-10-18 Jeff Law <law@redhat.com>
3134 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3136 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3137 (register_jump_thread): Add "registered_jump_thread" debug
3139 * dbgcnt.def (registered_jump_thread): New debug counter.
3141 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3143 * config/rs6000/rs6000.c: Include cgraph.h.
3145 2013-10-18 Teresa Johnson <tejohnson@google.com>
3147 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3148 weights during merging.
3150 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3152 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3153 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3154 * tree-ssa-operands.h: Remove prototype.
3155 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3156 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3157 * gimple.h: Add prototype.
3158 * gimple-ssa.h: Include tree-ssa-operands.h.
3159 * tree-dump.c: Add tree-cfg.h to include list.
3160 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3161 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3162 * config/i386/i386.c: Don't include tree-flow.h.
3163 * config/rs6000/rs6000.c: Likewise.
3165 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3167 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3169 2013-10-18 Richard Biener <rguenther@suse.de>
3171 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3172 or TYPE_UNSIGNED of integral types.
3173 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3174 NULL_TREE for zero-precision integral types.
3176 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3178 * config/aarch64/arm_neon.h
3179 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3181 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3184 * opth-gen.awk: Define target_flags_explicit.
3186 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3188 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3189 fields to the reg_addr array that describes the valid addressing
3190 mode for any register, general purpose registers, floating point
3191 registers, and Altivec registers.
3192 (FIRST_RELOAD_REG_CLASS): Likewise.
3193 (LAST_RELOAD_REG_CLASS): Likewise.
3194 (struct reload_reg_map_type): Likewise.
3195 (reload_reg_map_type): Likewise.
3196 (RELOAD_REG_VALID): Likewise.
3197 (RELOAD_REG_MULTIPLE): Likewise.
3198 (RELOAD_REG_INDEXED): Likewise.
3199 (RELOAD_REG_OFFSET): Likewise.
3200 (RELOAD_REG_PRE_INCDEC): Likewise.
3201 (RELOAD_REG_PRE_MODIFY): Likewise.
3202 (reg_addr): Likewise.
3203 (mode_supports_pre_incdec_p): New helper functions to say whether
3204 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3205 (mode_supports_pre_modify_p): Likewise.
3206 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3207 print the valid address mode bits for each mode.
3208 (rs6000_debug_print_mode): Likewise.
3209 (rs6000_debug_reg_global): Likewise.
3210 (rs6000_setup_reg_addr_masks): New function to set up the address
3211 mask bits for each type.
3212 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3213 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3214 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3215 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3216 PRE_MODIFY are supported.
3217 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3218 registers, instead of {src,dest}_av_p.
3219 (rs6000_print_options_internal): Tweak the debug output slightly.
3221 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3223 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3226 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3228 * tree-flow.h (struct omp_region): Move to omp-low.c.
3229 Remove omp_ prototypes and variables.
3230 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3231 (copy_var_decl): Relocate prototype from tree-flow.h.
3232 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3233 * tree.h: Move prototype to omp-low.h.
3234 * omp-low.h: New File. Relocate prototypes here.
3235 * omp-low.c (struct omp_region): Make local here.
3236 (root_omp_region): Make static.
3237 (copy_var_decl) Move to gimple.c.
3238 (new_omp_region): Make static.
3239 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3240 * tree-cfg.c: Include omp-low.h.
3241 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3242 * gimplify.c: Include omp-low.h.
3243 * tree-parloops.c: Likewise.
3245 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3247 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3248 operands are in memory, prefer to force non-matched operand 1 to
3251 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3254 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3255 restrict TImode addresses to single indirect registers if both
3256 -mquad-memory and -mvsx-timode are used.
3257 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3258 we should emit load/store quad. Remove using %y for quad memory
3261 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3262 constraints to allow load/store quad on machines where TImode is
3263 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3264 for TImode to load integer constants.
3266 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3268 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3270 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3272 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3273 handling of STACK_REG.
3275 2013-10-17 Richard Biener <rguenther@suse.de>
3277 PR tree-optimization/58143
3278 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3280 (rewrite_to_defined_overflow): Likewise.
3281 (move_computations_dom_walker::before_dom): Rewrite stmts
3282 with undefined signed overflow that are not always executed
3283 into unsigned arithmetic.
3285 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3288 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
3289 explicit isa flag to be an options variable, instead of using
3290 global_options_set. Remove define from rs6000.h.
3291 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
3293 * config/rs6000/rs6000.c (rs6000_option_override_internal):
3294 Initialize rs6000_isa_flags_explicit.
3295 (rs6000_function_specific_save): Add gcc_options* parameter, so
3296 that the powerpc builds after the 2013-10-15 changes.
3297 (rs6000_function_specific_restore): Likewise.
3299 2013-10-16 DJ Delorie <dj@redhat.com>
3301 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
3302 op is a REG before checking REGNO.
3303 (rl78_alloc_physical_registers): Verify pattern is a SET before
3306 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3308 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
3310 (vec_unpacks_lo_v4sf): Likewise.
3311 (vec_unpacks_float_hi_v4si): Likewise.
3312 (vec_unpacks_float_lo_v4si): Likewise.
3313 (vec_unpacku_float_hi_v4si): Likewise.
3314 (vec_unpacku_float_lo_v4si): Likewise.
3316 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3318 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
3319 (vsx_concat_v2sf): Likewise.
3321 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
3323 * config/aarch64/aarch64.md
3324 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
3326 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
3328 PR tree-optimization/58697
3329 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
3330 estimated_loop_iterations_int.
3331 (max_stmt_executions_int): Call get_max_loop_iterations_int.
3332 (get_max_loop_iterations_int): New. HWINT version of
3333 get_max_loop_iterations.
3334 * cfgloop.h: Add prototypes.
3335 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
3336 * loop-unroll.c (decide_peel_once_rolling): Call
3337 get_estimated_loop_iterations_int.
3338 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
3339 * tree-ssa-loop-niter.h: Tweak prototypes.
3341 2013-10-16 David Malcolm <dmalcolm@redhat.com>
3343 * gengtype-parse.c (struct_field_seq): Ignore access-control
3344 keywords ("public:" etc).
3346 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
3348 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
3349 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
3351 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
3353 * config/arm/arm.opt (mlra): New option.
3354 * config/arm/arm.c (arm_lra_p): New function.
3355 (TARGET_LRA_P): Define.
3357 2013-10-16 Paulo Matos <pmatos@broadcom.com>
3359 * tree-core.h (tree_code_name): Remove.
3360 * tree.h (get_tree_code_name): New prototype.
3361 * tree.c (tree_code_name): Make static.
3362 (get_tree_code_name): New function.
3363 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3364 tree_class_check_failed, tree_range_check_failed,
3365 tree_not_class_check_failed, omp_clause_check_failed,
3366 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3367 wrapper get_tree_code_name instead of calling tree_code_name directly.
3368 * tree-vrp.c (dump_asserts_for): Likewise.
3369 * tree-dump.c (dequeue_and_dump): Likewise.
3370 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3371 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3372 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3373 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3374 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3375 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3376 dump_gimple_omp_for): Likewise.
3377 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3378 * tree-ssa-pre.c (print_pre_expr): Likewise.
3379 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3380 * print-tree.c (print_node_brief, print_node): Likewise.
3381 * gimple.c (gimple_check_failed): Likewise.
3382 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3383 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3384 * config/mep/mep.c (mep_validate_vliw): Likewise.
3385 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3386 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3388 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3390 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3393 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3395 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3396 (MULTILIB_MATCHES): Add multilib for -march=v8.
3398 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3401 * optc-save-gen.awk: Add extra parameter to the save and restore
3403 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3404 * tree.c (build_optimization_node): New parameter. Add extra parameter
3405 to call to cl_optimization_save.
3406 (build_target_option_node): New parameter. Add extra parameter
3407 to call to cl_target_option_save.
3408 * tree.h (build_optimization_node): New parameter.
3409 (build_target_option_node): New parameter.
3410 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3411 build_optimization_node and build_target_option_node.
3412 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3413 (handle_pragma_push_options): Ditto.
3414 * toplev.c (process_options): Ditto.
3415 * opts.c (init_options_struct): Check for opts_set non-null.
3416 * target.def (target_option.save): New parameter.
3417 (target_option.restore): New parameter.
3418 * tm.texi: Generate.
3419 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3420 (ix86_pragma_target_parse): Ditto.
3421 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3423 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3424 to build_optimization_node and build_target_option_node.
3425 (rs6000_valid_attribute_p): Ditto.
3426 (rs6000_pragma_target_parse): Ditto.
3427 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3429 * config/i386/i386.h:
3430 TARGET_64BIT_P: New Macro
3431 TARGET_MMX_P: New Macro.
3432 TARGET_3DNOW_P: New Macro.
3433 TARGET_3DNOW_A_P: New Macro.
3434 TARGET_SSE_P: New Macro.
3435 TARGET_SSE2_P: New Macro.
3436 TARGET_SSE3_P: New Macro.
3437 TARGET_SSSE3_P: New Macro.
3438 TARGET_SSE4_1_P: New Macro.
3439 TARGET_SSE4_2_P: New Macro.
3440 TARGET_AVX_P: New Macro.
3441 TARGET_AVX2_P: New Macro.
3442 TARGET_AVX512F_P: New Macro.
3443 TARGET_AVX512PF_P: New Macro.
3444 TARGET_AVX512ER_P: New Macro.
3445 TARGET_AVX512CD_P: New Macro.
3446 TARGET_FMA_P: New Macro.
3447 TARGET_SSE4A_P: New Macro.
3448 TARGET_FMA4_P: New Macro.
3449 TARGET_XOP_P: New Macro.
3450 TARGET_LWP_P: New Macro.
3451 TARGET_ABM_P: New Macro.
3452 TARGET_BMI_P: New Macro.
3453 TARGET_BMI2_P: New Macro.
3454 TARGET_LZCNT_P: New Macro.
3455 TARGET_TBM_P: New Macro.
3456 TARGET_POPCNT_P: New Macro.
3457 TARGET_SAHF_P: New Macro.
3458 TARGET_MOVBE_P: New Macro.
3459 TARGET_CRC32_P: New Macro.
3460 TARGET_AES_P: New Macro.
3461 TARGET_PCLMUL_P: New Macro.
3462 TARGET_CMPXCHG16B_P: New Macro.
3463 TARGET_FSGSBASE_P: New Macro.
3464 TARGET_RDRND_P: New Macro.
3465 TARGET_F16C_P: New Macro.
3466 TARGET_RTM_P: New Macro.
3467 TARGET_HLE_P: New Macro.
3468 TARGET_RDSEED_P: New Macro.
3469 TARGET_PRFCHW_P: New Macro.
3470 TARGET_ADX_P: New Macro.
3471 TARGET_FXSR_P: New Macro.
3472 TARGET_XSAVE_P: New Macro.
3473 TARGET_XSAVEOPT_P: New Macro.
3474 TARGET_LP64_P: New Macro.
3475 TARGET_X32_P: New Macro.
3476 TARGET_FPMATH_DEFAULT_P: New Macro.
3477 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3478 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3480 Change ix86_tune_string to access opts->x_ix86_tune_string.
3481 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3482 Change ix86_arch_string to access opts->x_ix86_arch_string.
3483 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3484 Change ix86_pmode to access opts->x_ix86_pmode.
3485 Change ix86_abi to access opts->x_ix86_abi.
3486 Change ix86_cmodel to access opts->x_ix86_cmodel.
3487 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3488 Change ix86_isa_flags_explicit to access
3489 opts->x_ix86_isa_flags_explicit.
3490 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3491 Change ix86_regparm to access opts->x_ix86_regparm.
3492 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3493 Change ix86_preferred_stack_boundary_arg to access
3494 opts->x_ix86_preferred_stack_boundary_arg.
3495 Change ix86_force_align_arg_pointer to access
3496 opts->x_ix86_force_align_arg_pointer.
3497 Change ix86_incoming_stack_boundar_arg to access
3498 opts->x_ix86_incoming_stack_boundar_arg.
3499 Change ix86_fpmath to access opts->x_ix86_fpmath.
3500 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3501 Change ix86_recip_name to access opts->x_ix86_recip_name.
3502 Change ix86_stack_protector_guard to access
3503 opts->x_ix86_stack_protector_guard.
3504 Change ix86_tune_memcpy_strategy to access
3505 opts->x_ix86_tune_memcpy_strategy.
3506 Change ix86_tune_memset_strategy to access
3507 opts->x_ix86_tune_memset_strategy.
3508 Change global_options to access opts.
3509 Change global_options_set to access opts_set.
3510 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3511 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3512 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3513 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3514 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3515 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3516 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3517 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3518 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3519 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3520 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3521 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3522 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3523 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3524 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3525 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3526 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3527 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3528 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3529 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3530 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3531 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3532 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3533 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3534 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3535 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3536 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3537 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3538 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3539 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3540 Change TARGET_AES to TARGET_AES_P (opts->...).
3541 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3542 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3543 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3544 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3545 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3546 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3547 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3548 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3549 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3550 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3551 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3552 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3553 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3554 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3555 Change TARGET_X32 to TARGET_X32_P (opts->...).
3556 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3557 Change TARGET_FLOAT_RETURNS_IN_80387 to
3558 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3559 (ix86_function_specific_save): New parameter. Use opts-> fields
3560 to replace global fields.
3561 (ix86_function_specific_restore): Ditto.
3562 (ix86_valid_target_attribute_inner_p): New parameters.
3564 Fix call to ix86_handle_option and set_option.
3565 (ix86_valid_target_attribute_tree): New parameters.
3566 Change global_options to access opts.
3567 Change global_options_set to access opts_set.
3568 Fix call to ix86_valid_target_attribute_inner_p.
3569 Change ix86_tune_string to access opts->x_ix86_tune_string.
3570 Change ix86_arch_string to access opts->x_ix86_arch_string.
3571 Change ix86_fpmath to access opts->x_ix86_fpmath
3572 Fix call to ix86_option_override_internal.
3573 Fix call to ix86_add_new_builtins.
3574 Fix calls to build_optimization_node and build_target_option_node.
3575 (ix86_valid_target_attribute_p): Remove access to global_options.
3576 Use new gcc_options structure func_options.
3577 Fix call to ix86_valid_target_attribute_tree.
3578 Fix call to build_optimization_node.
3579 (get_builtin_code_for_version): Fix call to
3580 ix86_valid_target_attribute_tree.
3582 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3584 * Makefile.in (PICFLAG): New.
3585 (enable_host_shared): New.
3586 (INTERNAL_CFLAGS): Use PICFLAG.
3587 (LIBIBERTY): Use pic build of libiberty.a if configured with
3588 --enable-host-shared.
3589 * configure.ac: Add --enable-host-shared, setting up new
3591 * configure: Regenerate.
3592 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3593 (--enable-host-shared): New option.
3595 2013-10-15 Richard Biener <rguenther@suse.de>
3597 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3598 for built-in functions.
3600 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3602 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3603 (optimize_range_tests_1): New function,
3604 extracted from optimize_range_tests.
3605 (optimize_range_tests_xor): Similarly.
3606 (optimize_range_tests_diff): New function.
3607 (optimize_range_tests): Use optimize_range_tests_1.
3609 2013-10-15 Cong Hou <congh@google.com>
3611 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3612 requirement of the reduction pattern so that one operand of the
3613 reduction operation can come from outside of the loop.
3615 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3617 * config/arm/neon-schedgen.ml: Remove.
3618 * config/arm/cortex-a9-neon.md: Remove comment regarding
3621 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3623 * config/arm/types: Remove old neon types.
3625 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3627 * config/arm/cortex-a7.md
3628 (cortex_a7_neon_type): New.
3629 (cortex_a7_neon_mul): Update for new types.
3630 (cortex_a7_neon_mla): Likewise.
3631 (cortex_a7_neon): Likewise.
3633 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3635 * config/arm/cortex-a15-neon.md
3636 (cortex_a15_neon_type): New,
3638 (cortex_a15_neon_int_1): Remove.
3639 (cortex_a15_neon_int_2): Likewise.
3640 (cortex_a15_neon_int_3): Likewise.
3641 (cortex_a15_neon_int_4): Likewise.
3642 (cortex_a15_neon_int_5): Likewise.
3643 (cortex_a15_neon_vqneg_vqabs): Likewise.
3644 (cortex_a15_neon_vmov): Likewise.
3645 (cortex_a15_neon_vaba): Likewise.
3646 (cortex_a15_neon_vaba_qqq): Likewise.
3647 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3648 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3649 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3651 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3652 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3653 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3654 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3655 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3656 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3657 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3658 (cortex_a15_neon_shift_1): Likewise.
3659 (cortex_a15_neon_shift_2): Likewise.
3660 (cortex_a15_neon_shift_3): Likewise.
3661 (cortex_a15_neon_vshl_ddd): Likewise.
3662 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3663 (cortex_a15_neon_vsra_vrsra): Likewise.
3664 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3665 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3666 (cortex_a15_neon_bp_3cycle): Likewise.
3667 (cortex_a15_neon_ldm_2): Likewise.
3668 (cortex_a15_neon_stm_2): Likewise.
3669 (cortex_a15_neon_mcr): Likewise.
3670 (cortex_a15_neon_mrc): Likewise.
3671 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3672 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3673 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3674 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3675 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3676 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3677 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3678 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3679 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3680 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3681 (cortex_a15_neon_bp_simple): Likewise.
3682 (cortex_a15_neon_bp_2cycle): Likewise.
3683 (cortex_a15_neon_bp_3cycle): Likewise.
3684 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3685 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3686 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3687 (cortex_a15_neon_vld2_4_regs): Likewise.
3688 (cortex_a15_neon_vld3_vld4): Likewise.
3689 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3690 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3691 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3692 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3693 (cortex_a15_neon_vst3_vst4): Rename to...
3694 (cortex_a15_neon_vst4): This, update for new attributes.
3695 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3696 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3697 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3698 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3699 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3700 (cortex_a15_neon_ldm_2): Likewise.
3701 (cortex_a15_neon_stm_2): Likewise.
3702 (cortex_a15_neon_mcr): Likewise.
3703 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3704 (cortex_a15_neon_mrc): Likewise.
3705 (cortex_a15_neon_mrrc): Likewise.
3707 (cortex_a15_neon_abd): New.
3708 (cortex_a15_neon_abd_q): Likewise.
3709 (cortex_a15_neon_aba): Likewise.
3710 (cortex_a15_neon_aba_q): Likewise.
3711 (cortex_a15_neon_acc): Likewise.
3712 (cortex_a15_neon_acc_q): Likewise.
3713 (cortex_a15_neon_arith_basic): Likewise.
3714 (cortex_a15_neon_arith_complex): Likewise.
3715 (cortex_a15_neon_multiply): Likewise.
3716 (cortex_a15_neon_multiply_q): Likewise.
3717 (cortex_a15_neon_mla): Likewise.
3718 (cortex_a15_neon_mla_q): Likewise.
3719 (cortex_a15_neon_sat_mla_long): Likewise.
3720 (cortex_a15_neon_shift_acc): Likewise.
3721 (cortex_a15_neon_shift_imm_basic): Likewise.
3722 (cortex_a15_neon_shift_imm_complex): Likewise.
3723 (cortex_a15_neon_shift_reg_basic): Likewise.
3724 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3725 (cortex_a15_neon_shift_reg_complex): Likewise.
3726 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3727 (cortex_a15_neon_fp_negabs): Likewise
3728 (cortex_a15_neon_fp_arith): Likewise
3729 (cortex_a15_neon_fp_arith_q): Likewise
3730 (cortex_a15_neon_fp_cvt_int): Likewise
3731 (cortex_a15_neon_fp_cvt_int_q): Likewise
3732 (cortex_a15_neon_fp_cvt_16): Likewise
3733 (cortex_a15_neon_fp_mul): Likewise
3734 (cortex_a15_neon_fp_mul_q): Likewise
3735 (cortex_a15_neon_fp_mla): Likewise
3736 (cortex_a15_neon_fp_mla_q): Likewise
3737 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3738 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3739 (cortex_a15_neon_bitops): Likewise.
3740 (cortex_a15_neon_bitops_q): Likewise.
3741 (cortex_a15_neon_from_gp): Likewise.
3742 (cortex_a15_neon_from_gp_q): Likewise.
3743 (cortex_a15_neon_tbl3_tbl4): Likewise.
3744 (cortex_a15_neon_zip_q): Likewise.
3745 (cortex_a15_neon_to_gp): Likewise.
3746 (cortex_a15_neon_load_a): Likewise.
3747 (cortex_a15_neon_load_b): Likewise.
3748 (cortex_a15_neon_load_c): Likewise.
3749 (cortex_a15_neon_load_d): Likewise.
3750 (cortex_a15_neon_load_e): Likewise.
3751 (cortex_a15_neon_load_f): Likewise.
3752 (cortex_a15_neon_store_a): Likewise.
3753 (cortex_a15_neon_store_b): Likewise.
3754 (cortex_a15_neon_store_c): Likewise.
3755 (cortex_a15_neon_store_d): Likewise.
3756 (cortex_a15_neon_store_e): Likewise.
3757 (cortex_a15_neon_store_f): Likewise.
3758 (cortex_a15_neon_store_g): Likewise.
3759 (cortex_a15_neon_store_h): Likewise.
3760 (cortex_a15_vfp_to_from_gp): Likewise.
3762 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3764 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3766 (cortex_a9_neon_vshl_ddd): Remove.
3767 (cortex_a9_neon_vst3_vst4): Likewise.
3768 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3770 (cortex_a9_neon_bit_ops_q): New.
3772 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3773 (cortex_a9_neon_int_2): Likewise.
3774 (cortex_a9_neon_int_3): Likewise.
3775 (cortex_a9_neon_int_4): Likewise.
3776 (cortex_a9_neon_int_5): Likewise.
3777 (cortex_a9_neon_vqneg_vqabs): Likewise.
3778 (cortex_a9_neon_vmov): Likewise.
3779 (cortex_a9_neon_vaba): Likewise.
3780 (cortex_a9_neon_vaba_qqq): Likewise.
3781 (cortex_a9_neon_shift_1): Likewise.
3782 (cortex_a9_neon_shift_2): Likewise.
3783 (cortex_a9_neon_shift_3): Likewise.
3784 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3785 (cortex_a9_neon_vsra_vrsra): Likewise.
3786 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3787 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3788 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3790 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3791 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3792 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3794 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3795 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3796 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3797 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3798 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3799 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3800 (cortex_a9_neon_fp_vsum): Likewise.
3801 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3802 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3803 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3804 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3805 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3806 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3807 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3808 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3809 (cortex_a9_neon_bp_simple): Likewise.
3810 (cortex_a9_neon_bp_2cycle): Likewise.
3811 (cortex_a9_neon_bp_3cycle): Likewise.
3812 (cortex_a9_neon_ldr): Likewise.
3813 (cortex_a9_neon_str): Likewise.
3814 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3815 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3816 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3817 (cortex_a9_neon_vld2_4_regs): Likewise.
3818 (cortex_a9_neon_vld3_vld4): Likewise.
3819 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3820 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3821 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3822 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3823 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3824 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3825 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3826 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3827 (cortex_a9_neon_mcr): Likewise.
3828 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3829 (cortex_a9_neon_mrc): Likewise.
3830 (cortex_a9_neon_mrrc): Likewise.
3832 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3834 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3836 (cortex_a8_neon_vshl_ddd): Remove.
3837 (cortex_a8_neon_vst3_vst4): Likewise.
3838 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3840 (cortex_a8_neon_bit_ops_q): New.
3842 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3843 (cortex_a8_neon_int_2): Likewise..
3844 (cortex_a8_neon_int_3): Likewise.
3845 (cortex_a8_neon_int_5): Likewise.
3846 (cortex_a8_neon_vqneg_vqabs): Likewise.
3847 (cortex_a8_neon_int_4): Likewise.
3848 (cortex_a8_neon_vaba): Likewise.
3849 (cortex_a8_neon_vaba_qqq): Likewise.
3850 (cortex_a8_neon_shift_1): Likewise.
3851 (cortex_a8_neon_shift_2): Likewise.
3852 (cortex_a8_neon_shift_3): Likewise.
3853 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3854 (cortex_a8_neon_vsra_vrsra): Likewise.
3855 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3856 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3857 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3859 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3860 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3861 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3863 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3864 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3865 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3866 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3867 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3868 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3869 (cortex_a8_neon_fp_vsum): Likewise.
3870 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3871 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3872 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3873 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3874 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3875 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3876 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3877 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3878 (cortex_a8_neon_bp_simple): Likewise.
3879 (cortex_a8_neon_bp_2cycle): Likewise.
3880 (cortex_a8_neon_bp_3cycle): Likewise.
3881 (cortex_a8_neon_ldr): Likewise.
3882 (cortex_a8_neon_str): Likewise.
3883 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3884 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3885 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3886 (cortex_a8_neon_vld2_4_regs): Likewise.
3887 (cortex_a8_neon_vld3_vld4): Likewise.
3888 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3889 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3890 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3891 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3892 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3893 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3894 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3895 (cortex_a8_neon_mcr): Likewise.
3896 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3897 (cortex_a8_neon_mrc): Likewise.
3898 (cortex_a8_neon_mrrc): Likewise.
3900 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3902 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3904 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3905 (aarch64_simd_dup<mode>): Add "type" attribute.
3906 (aarch64_dup_lane<mode>): Likewise.
3907 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3908 (*aarch64_simd_mov<mode>): Likewise.
3909 (aarch64_simd_mov_from_<mode>low): Likewise.
3910 (aarch64_simd_mov_from_<mode>high): Likewise.
3911 (orn<mode>3): Likewise.
3912 (bic<mode>3): Likewise.
3913 (add<mode>3): Likewise.
3914 (sub<mode>3): Likewise.
3915 (mul<mode>3): Likewise.
3916 (*aarch64_mul3_elt<mode>): Likewise.
3917 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3918 (*aarch64_mul3_elt_to_128df): Likewise.
3919 (*aarch64_mul3_elt_to_64v2df): Likewise.
3920 (neg<mode>2): Likewise.
3921 (abs<mode>2): Likewise.
3922 (abd<mode>_3): Likewise.
3923 (aba<mode>_3): Likewise.
3924 (fabd<mode>_3): Likewise.
3925 (*fabd_scalar<mode>3): Likewise.
3926 (and<mode>3): Likewise.
3927 (ior<mode>3): Likewise.
3928 (xor<mode>3): Likewise.
3929 (one_cmpl<mode>2): Likewise.
3930 (aarch64_simd_vec_set<mode>): Likewise.
3931 (aarch64_simd_lshr<mode>): Likewise.
3932 (aarch64_simd_ashr<mode>): Likewise.
3933 (aarch64_simd_imm_shl<mode>): Likewise.
3934 (aarch64_simd_reg_sshl<mode): Likewise.
3935 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3936 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3937 (aarch64_simd_vec_setv2di): Likewise.
3938 (aarch64_simd_vec_set<mode>): Likewise.
3939 (aarch64_mla<mode>): Likewise.
3940 (*aarch64_mla_elt<mode>): Likewise.
3941 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3942 (aarch64_mls<mode>): Likewise.
3943 (*aarch64_mls_elt<mode>): Likewise.
3944 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3945 (<su><maxmin><mode>3): Likewise.
3946 (move_lo_quad_<mode>): Likewise.
3947 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3948 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3949 (vec_pack_trunc_<mode>): Likewise.
3950 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3951 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3952 (*aarch64_<su>mlal_lo<mode>): Likewise.
3953 (*aarch64_<su>mlal_hi<mode>): Likewise.
3954 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3955 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3956 (*aarch64_<su>mlal<mode>): Likewise.
3957 (*aarch64_<su>mlsl<mode>): Likewise.
3958 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3959 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3960 (add<mode>3): Likewise.
3961 (sub<mode>3): Likewise.
3962 (mul<mode>3): Likewise.
3963 (div<mode>3): Likewise.
3964 (neg<mode>2): Likewise.
3965 (abs<mode>2): Likewise.
3966 (fma<mode>4): Likewise.
3967 (*aarch64_fma4_elt<mode>): Likewise.
3968 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3969 (*aarch64_fma4_elt_to_128df): Likewise.
3970 (*aarch64_fma4_elt_to_64v2df): Likewise.
3971 (fnma<mode>4): Likewise.
3972 (*aarch64_fnma4_elt<mode>): Likewise.
3973 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3974 (*aarch64_fnma4_elt_to_128df): Likewise.
3975 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3976 (<frint_pattern><mode>2): Likewise.
3977 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3978 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3979 (vec_unpacks_lo_v4sf): Likewise.
3980 (aarch64_float_extend_lo_v2df): Likewise.
3981 (vec_unpacks_hi_v4sf): Likewise.
3982 (aarch64_float_truncate_lo_v2sf): Likewise.
3983 (aarch64_float_truncate_hi_v4sf): Likewise.
3984 (aarch64_vmls<mode>): Likewise.
3985 (<su><maxmin><mode>3): Likewise.
3986 (<maxmin_uns><mode>3): Likewise.
3987 (reduc_<sur>plus_<mode>): Likewise.
3988 (reduc_<sur>plus_v2di): Likewise.
3989 (reduc_<sur>plus_v2si): Likewise.
3990 (reduc_<sur>plus_<mode>): Likewise.
3991 (aarch64_addpv4sf): Likewise.
3992 (clz<mode>2): Likewise.
3993 (reduc_<maxmin_uns>_<mode>): Likewise.
3994 (reduc_<maxmin_uns>_v2di): Likewise.
3995 (reduc_<maxmin_uns>_v2si): Likewise.
3996 (reduc_<maxmin_uns>_<mode>): Likewise.
3997 (reduc_<maxmin_uns>_v4sf): Likewise.
3998 (aarch64_simd_bsl<mode>_internal): Likewise.
3999 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4000 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4001 (aarch64_get_lane<mode>): Likewise.
4002 (*aarch64_combinez<mode>): Likewise.
4003 (aarch64_combine<mode>): Likewise.
4004 (aarch64_simd_combine<mode>): Likewise.
4005 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4006 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4007 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4008 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4009 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4010 (aarch64_<sur>h<addsub><mode>): Likewise.
4011 (aarch64_<sur><addsub>hn<mode>): Likewise.
4012 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4013 (aarch64_pmul<mode>): Likewise.
4014 (aarch64_<su_optab><optab><mode>): Likewise.
4015 (aarch64_<sur>qadd<mode>): Likewise.
4016 (aarch64_sqmovun<mode>): Likewise.
4017 (aarch64_<sur>qmovn<mode>): Likewise.
4018 (aarch64_s<optab><mode>): Likewise.
4019 (aarch64_sq<r>dmulh<mode>): Likewise.
4020 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4021 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4022 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4023 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4024 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4025 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4026 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4027 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4028 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4029 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4030 (aarch64_sqdmull<mode>): Likewise.
4031 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4032 (aarch64_sqdmull_n<mode>): Likewise.
4033 (aarch64_sqdmull2<mode>_internal): Likewise.
4034 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4035 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4036 (aarch64_<sur>shl<mode>): Likewise.
4037 (aarch64_<sur>q<r>shl<mode>
4038 (aarch64_<sur>shll_n<mode>): Likewise.
4039 (aarch64_<sur>shll2_n<mode>): Likewise.
4040 (aarch64_<sur>shr_n<mode>): Likewise.
4041 (aarch64_<sur>sra_n<mode>): Likewise.
4042 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4043 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4044 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4045 (aarch64_cm<optab><mode>): Likewise.
4046 (aarch64_cm<optab>di): Likewise.
4047 (aarch64_cm<optab><mode>): Likewise.
4048 (aarch64_cm<optab>di): Likewise.
4049 (aarch64_cmtst<mode>): Likewise.
4050 (aarch64_cmtstdi): Likewise.
4051 (aarch64_cm<optab><mode>): Likewise.
4052 (*aarch64_fac<optab><mode>): Likewise.
4053 (aarch64_addp<mode>): Likewise.
4054 (aarch64_addpdi): Likewise.
4055 (sqrt<mode>2): Likewise.
4056 (vec_load_lanesoi<mode>): Likewise.
4057 (vec_store_lanesoi<mode>): Likewise.
4058 (vec_load_lanesci<mode>): Likewise.
4059 (vec_store_lanesci<mode>): Likewise.
4060 (vec_load_lanesxi<mode>): Likewise.
4061 (vec_store_lanesxi<mode>): Likewise.
4062 (*aarch64_mov<mode>): Likewise.
4063 (aarch64_ld2<mode>_dreg): Likewise.
4064 (aarch64_ld2<mode>_dreg): Likewise.
4065 (aarch64_ld3<mode>_dreg): Likewise.
4066 (aarch64_ld3<mode>_dreg): Likewise.
4067 (aarch64_ld4<mode>_dreg): Likewise.
4068 (aarch64_ld4<mode>_dreg): Likewise.
4069 (aarch64_tbl1<mode>): Likewise.
4070 (aarch64_tbl2v16qi): Likewise.
4071 (aarch64_combinev16qi): Likewise.
4072 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4073 (aarch64_st2<mode>_dreg): Likewise.
4074 (aarch64_st2<mode>_dreg): Likewise.
4075 (aarch64_st3<mode>_dreg): Likewise.
4076 (aarch64_st3<mode>_dreg): Likewise.
4077 (aarch64_st4<mode>_dreg): Likewise.
4078 (aarch64_st4<mode>_dreg): Likewise.
4079 (*aarch64_simd_ld1r<mode>): Likewise.
4080 (aarch64_frecpe<mode>): Likewise.
4081 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4082 (aarch64_frecps<mode>): Likewise.
4084 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4086 * config/arm/iterators.md (V_elem_ch): New.
4088 (VQH_type): Likewise.
4089 * config/arm/arm.md (is_neon_type): New.
4090 (conds): Use is_neon_type.
4091 (anddi3_insn): Update type attribute.
4092 (xordi3_insn): Likewise.
4093 (one_cmpldi2): Likewise.
4094 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4095 * config/arm/neon.md (neon_mov): Update type attribute.
4096 (*movmisalign<mode>_neon_store): Likewise.
4097 (*movmisalign<mode>_neon_load): Likewise.
4098 (vec_set<mode>_internal): Likewise.
4099 (vec_set<mode>_internal): Likewise.
4100 (vec_setv2di_internal): Likewise.
4101 (vec_extract<mode>): Likewise.
4102 (vec_extract<mode>): Likewise.
4103 (vec_extractv2di): Likewise.
4104 (*add<mode>3_neon): Likewise.
4105 (adddi3_neon): Likewise.
4106 (*sub<mode>3_neon): Likewise.
4107 (subdi3_neon): Likewise.
4108 (fma<VCVTF:mode>4): Likewise.
4109 (fma<VCVTF:mode>4_intrinsic): Likewise.
4110 (*fmsub<VCVTF:mode>4): Likewise.
4111 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4112 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4113 (ior<mode>3): Likewise.
4114 (and<mode>3): Likewise.
4115 (orn<mode>3_neon): Likewise.
4116 (orndi3_neon): Likewise.
4117 (bic<mode>3_neon): Likewise.
4118 (bicdi3_neon): Likewise.
4119 (xor<mode>3): Likewise.
4120 (one_cmpl<mode>2): Likewise.
4121 (abs<mode>2): Likewise.
4122 (neg<mode>2): Likewise.
4123 (negdi2_neon): Likewise.
4124 (*umin<mode>3_neon): Likewise.
4125 (*umax<mode>3_neon): Likewise.
4126 (*smin<mode>3_neon): Likewise.
4127 (*smax<mode>3_neon): Likewise.
4128 (vashl<mode>3): Likewise.
4129 (vashr<mode>3_imm): Likewise.
4130 (vlshr<mode>3_imm): Likewise.
4131 (ashl<mode>3_signed): Likewise.
4132 (ashl<mode>3_unsigned): Likewise.
4133 (neon_load_count): Likewise.
4134 (ashldi3_neon_noclobber): Likewise.
4135 (ashldi3_neon): Likewise.
4136 (signed_shift_di3_neon): Likewise.
4137 (unsigned_shift_di3_neon): Likewise.
4138 (ashrdi3_neon_imm_noclobber): Likewise.
4139 (lshrdi3_neon_imm_noclobber): Likewise.
4140 (<shift>di3_neon): Likewise.
4141 (widen_ssum<mode>3): Likewise.
4142 (widen_usum<mode>3): Likewise.
4143 (quad_halves_<code>v4si): Likewise.
4144 (quad_halves_<code>v4sf): Likewise.
4145 (quad_halves_<code>v8hi): Likewise.
4146 (quad_halves_<code>v16qi): Likewise.
4147 (reduc_splus_v2di): Likewise.
4148 (neon_vpadd_internal<mode>): Likewise.
4149 (neon_vpsmin<mode>): Likewise.
4150 (neon_vpsmax<mode>): Likewise.
4151 (neon_vpumin<mode>): Likewise.
4152 (neon_vpumax<mode>): Likewise.
4153 (*ss_add<mode>_neon): Likewise.
4154 (*us_add<mode>_neon): Likewise.
4155 (*ss_sub<mode>_neon): Likewise.
4156 (*us_sub<mode>_neon): Likewise.
4157 (neon_vadd<mode>_unspec): Likewise.
4158 (neon_vaddl<mode>): Likewise.
4159 (neon_vaddw<mode>): Likewise.
4160 (neon_vhadd<mode>): Likewise.
4161 (neon_vqadd<mode>): Likewise.
4162 (neon_vaddhn<mode>): Likewise.
4163 (neon_vmul<mode>): Likewise.
4164 (neon_vfms<VCVTF:mode>): Likewise.
4165 (neon_vmlal<mode>): Likewise.
4166 (neon_vmls<mode>): Likewise.
4167 (neon_vmlsl<mode>): Likewise.
4168 (neon_vqdmulh<mode>): Likewise.
4169 (neon_vqdmlal<mode>): Likewise.
4170 (neon_vqdmlsl<mode>): Likewise.
4171 (neon_vmull<mode>): Likewise.
4172 (neon_vqdmull<mode>): Likewise.
4173 (neon_vsub<mode>_unspec): Likewise.
4174 (neon_vsubl<mode>): Likewise.
4175 (neon_vsubw<mode>): Likewise.
4176 (neon_vqsub<mode>): Likewise.
4177 (neon_vhsub<mode>): Likewise.
4178 (neon_vsubhn<mode>): Likewise.
4179 (neon_vceq<mode>): Likewise.
4180 (neon_vcge<mode>): Likewise.
4181 (neon_vcgeu<mode>): Likewise.
4182 (neon_vcgt<mode>): Likewise.
4183 (neon_vcgtu<mode>): Likewise.
4184 (neon_vcle<mode>): Likewise.
4185 (neon_vclt<mode>): Likewise.
4186 (neon_vcage<mode>): Likewise.
4187 (neon_vcagt<mode>): Likewise.
4188 (neon_vtst<mode>): Likewise.
4189 (neon_vabd<mode>): Likewise.
4190 (neon_vabdl<mode>): Likewise.
4191 (neon_vaba<mode>): Likewise.
4192 (neon_vabal<mode>): Likewise.
4193 (neon_vmax<mode>): Likewise.
4194 (neon_vmin<mode>): Likewise.
4195 (neon_vpaddl<mode>): Likewise.
4196 (neon_vpadal<mode>): Likewise.
4197 (neon_vpmax<mode>): Likewise.
4198 (neon_vpmin<mode>): Likewise.
4199 (neon_vrecps<mode>): Likewise.
4200 (neon_vrsqrts<mode>): Likewise.
4201 (neon_vqabs<mode>): Likewise.
4202 (neon_vqneg<mode>): Likewise.
4203 (neon_vcls<mode>): Likewise.
4204 (clz<mode>2): Likewise.
4205 (popcount<mode>2): Likewise.
4206 (neon_vrecpe<mode>): Likewise.
4207 (neon_vrsqrte<mode>): Likewise.
4208 (neon_vget_lane<mode>_sext_internal): Likewise.
4209 (neon_vget_lane<mode>_zext_internal): Likewise.
4210 (neon_vdup_n<mode>): Likewise.
4211 (neon_vdup_n<mode>): Likewise.
4212 (neon_vdup_nv2di): Likewise.
4213 (neon_vdup_lane<mode>_interal): Likewise.
4214 (*neon_vswp<mode>): Likewise.
4215 (neon_vcombine<mode>): Likewise.
4216 (float<mode><V_cvtto>2): Likewise.
4217 (floatuns<mode><V_cvtto>2): Likewise.
4218 (fix_trunc<mode><V_cvtto>2): Likewise.
4219 (fixuns_trunc<mode><V_cvtto>2
4220 (neon_vcvt<mode>): Likewise.
4221 (neon_vcvt<mode>): Likewise.
4222 (neon_vcvtv4sfv4hf): Likewise.
4223 (neon_vcvtv4hfv4sf): Likewise.
4224 (neon_vcvt_n<mode>): Likewise.
4225 (neon_vcvt_n<mode>): Likewise.
4226 (neon_vmovn<mode>): Likewise.
4227 (neon_vqmovn<mode>): Likewise.
4228 (neon_vqmovun<mode>): Likewise.
4229 (neon_vmovl<mode>): Likewise.
4230 (neon_vmul_lane<mode>): Likewise.
4231 (neon_vmul_lane<mode>): Likewise.
4232 (neon_vmull_lane<mode>): Likewise.
4233 (neon_vqdmull_lane<mode>): Likewise.
4234 (neon_vqdmulh_lane<mode>): Likewise.
4235 (neon_vqdmulh_lane<mode>): Likewise.
4236 (neon_vmla_lane<mode>): Likewise.
4237 (neon_vmla_lane<mode>): Likewise.
4238 (neon_vmlal_lane<mode>): Likewise.
4239 (neon_vqdmlal_lane<mode>): Likewise.
4240 (neon_vmls_lane<mode>): Likewise.
4241 (neon_vmls_lane<mode>): Likewise.
4242 (neon_vmlsl_lane<mode>): Likewise.
4243 (neon_vqdmlsl_lane<mode>): Likewise.
4244 (neon_vext<mode>): Likewise.
4245 (neon_vrev64<mode>): Likewise.
4246 (neon_vrev32<mode>): Likewise.
4247 (neon_vrev16<mode>): Likewise.
4248 (neon_vbsl<mode>_internal): Likewise.
4249 (neon_vshl<mode>): Likewise.
4250 (neon_vqshl<mode>): Likewise.
4251 (neon_vshr_n<mode>): Likewise.
4252 (neon_vshrn_n<mode>): Likewise.
4253 (neon_vqshrn_n<mode>): Likewise.
4254 (neon_vqshrun_n<mode>): Likewise.
4255 (neon_vshl_n<mode>): Likewise.
4256 (neon_vqshl_n<mode>): Likewise.
4257 (neon_vqshlu_n<mode>): Likewise.
4258 (neon_vshll_n<mode>): Likewise.
4259 (neon_vsra_n<mode>): Likewise.
4260 (neon_vsri_n<mode>): Likewise.
4261 (neon_vsli_n<mode>): Likewise.
4262 (neon_vtbl1v8qi): Likewise.
4263 (neon_vtbl2v8qi): Likewise.
4264 (neon_vtbl3v8qi): Likewise.
4265 (neon_vtbl4v8qi): Likewise.
4266 (neon_vtbl1v16qi): Likewise.
4267 (neon_vtbl2v16qi): Likewise.
4268 (neon_vcombinev16qi): Likewise.
4269 (neon_vtbx1v8qi): Likewise.
4270 (neon_vtbx2v8qi): Likewise.
4271 (neon_vtbx3v8qi): Likewise.
4272 (neon_vtbx4v8qi): Likewise.
4273 (*neon_vtrn<mode>_insn): Likewise.
4274 (*neon_vzip<mode>_insn): Likewise.
4275 (*neon_vuzp<mode>_insn): Likewise.
4276 (neon_vld1<mode>): Likewise.
4277 (neon_vld1_lane<mode>): Likewise.
4278 (neon_vld1_lane<mode>): Likewise.
4279 (neon_vld1_dup<mode>): Likewise.
4280 (neon_vld1_dup<mode>): Likewise.
4281 (neon_vld1_dupv2di): Likewise.
4282 (neon_vst1<mode>): Likewise.
4283 (neon_vst1_lane<mode>): Likewise.
4284 (neon_vst1_lane<mode>): Likewise.
4285 (neon_vld2<mode>): Likewise.
4286 (neon_vld2<mode>): Likewise.
4287 (neon_vld2_lane<mode>): Likewise.
4288 (neon_vld2_lane<mode>): Likewise.
4289 (neon_vld2_dup<mode>): Likewise.
4290 (neon_vst2<mode>): Likewise.
4291 (neon_vst2<mode>): Likewise.
4292 (neon_vst2_lane<mode>): Likewise.
4293 (neon_vst2_lane<mode>): Likewise.
4294 (neon_vld3<mode>): Likewise.
4295 (neon_vld3qa<mode>): Likewise.
4296 (neon_vld3qb<mode>): Likewise.
4297 (neon_vld3_lane<mode>): Likewise.
4298 (neon_vld3_lane<mode>): Likewise.
4299 (neon_vld3_dup<mode>): Likewise.
4300 (neon_vst3<mode>): Likewise.
4301 (neon_vst3qa<mode>): Likewise.
4302 (neon_vst3qb<mode>): Likewise.
4303 (neon_vst3_lane<mode>): Likewise.
4304 (neon_vst3_lane<mode>): Likewise.
4305 (neon_vld4<mode>): Likewise.
4306 (neon_vld4qa<mode>): Likewise.
4307 (neon_vld4qb<mode>): Likewise.
4308 (neon_vld4_lane<mode>): Likewise.
4309 (neon_vld4_lane<mode>): Likewise.
4310 (neon_vld4_dup<mode>): Likewise.
4311 (neon_vst4<mode>): Likewise.
4312 (neon_vst4qa<mode>): Likewise.
4313 (neon_vst4qb<mode>): Likewise.
4314 (neon_vst4_lane<mode>): Likewise.
4315 (neon_vst4_lane<mode>): Likewise.
4316 (neon_vec_unpack<US>_lo_<mode>): Likewise.
4317 (neon_vec_unpack<US>_hi_<mode>): Likewise.
4318 (neon_vec_<US>mult_lo_<mode>): Likewise.
4319 (neon_vec_<US>mult_hi_<mode>): Likewise.
4320 (neon_vec_<US>shiftl_<mode>): Likewise.
4321 (neon_unpack<US>_<mode>): Likewise.
4322 (neon_vec_<US>mult_<mode>): Likewise.
4323 (vec_pack_trunc_<mode>): Likewise.
4324 (neon_vec_pack_trunc_<mode>): Likewise.
4325 (neon_vabd<mode>_2): Likewise.
4326 (neon_vabd<mode>_3): Likewise.
4328 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4330 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
4331 (load_pair): Update type attribute.
4332 (store_pair): Update type attribute.
4333 * config/aarch64/iterators.md (q): New.
4335 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4337 * config/arm/types.md: Add new types for Neon insns.
4339 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4340 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4341 Sergey Lega <sergey.s.lega@intel.com>
4342 Anna Tikhonova <anna.tikhonova@intel.com>
4343 Ilya Tocar <ilya.tocar@intel.com>
4344 Andrey Turetskiy <andrey.turetskiy@intel.com>
4345 Ilya Verbin <ilya.verbin@intel.com>
4346 Kirill Yukhin <kirill.yukhin@intel.com>
4347 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4349 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
4350 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
4351 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
4353 (srcp14<mode>): Ditto.
4354 (rsqrt14<mode>): Ditto.
4355 (rsqrt14<mode>): Ditto.
4356 (avx512f_vmscalef<mode>): Ditto.
4357 (avx512f_scalef<mode>): Ditto.
4358 (avx512f_getexp<mode>): Ditto.
4359 (avx512f_sgetexp<mode>): Ditto.
4360 (avx512f_fixupimm<mode>): Ditto.
4361 (avx512f_sfixupimm<mode>): Ditto.
4362 (avx512f_rndscale<mode>): Ditto.
4363 (*avx512er_exp2<mode>): Ditto.
4364 (*avx512er_rcp28<mode>): Ditto.
4365 (avx512er_rsqrt28<mode>): Ditto.
4366 (avx512f_getmant<mode>): Ditto.
4367 (avx512f_getmant<mode>): Ditto.
4368 (avx512f_rndscale<mode>): Fix formatting.
4370 2013-10-15 Martin Jambor <mjambor@suse.cz>
4372 * ipa-utils.h (ipa_edge_within_scc): Declare.
4373 * ipa-cp.c (edge_within_scc): Moved...
4374 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4376 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4377 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4378 Sergey Lega <sergey.s.lega@intel.com>
4379 Anna Tikhonova <anna.tikhonova@intel.com>
4380 Ilya Tocar <ilya.tocar@intel.com>
4381 Andrey Turetskiy <andrey.turetskiy@intel.com>
4382 Ilya Verbin <ilya.verbin@intel.com>
4383 Kirill Yukhin <kirill.yukhin@intel.com>
4384 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4386 * config/i386/predicates.md (const_8_to_15_operand): New.
4387 (const_16_to_31_operand): Ditto.
4388 * config/i386/sse.md (V8FI): New.
4390 (reduc_splus_v8df): Ditto.
4391 (reduc_splus_v16sf): Ditto.
4392 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4393 (vec_extract_hi_<mode>): Ditto.
4394 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4395 (vec_set_lo_<mode>): Ditto.
4396 (vec_set_hi_<mode>): Ditto.
4397 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4398 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4399 (avx512f_pshufd_1): Ditto.
4400 (avx512f_broadcast<mode>): Ditto.
4401 (avx512f_broadcast<mode>): Ditto.
4402 (define_split): Split vec_extract_lo into move.
4403 (ssequartermode): Ditto.
4404 (ssedoublemode): Extened with wider modes.
4405 (vec_extract_lo_<mode>): Ditto.
4407 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4408 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4409 Sergey Lega <sergey.s.lega@intel.com>
4410 Anna Tikhonova <anna.tikhonova@intel.com>
4411 Ilya Tocar <ilya.tocar@intel.com>
4412 Andrey Turetskiy <andrey.turetskiy@intel.com>
4413 Ilya Verbin <ilya.verbin@intel.com>
4414 Kirill Yukhin <kirill.yukhin@intel.com>
4415 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4417 * config/i386/predicates.md (register_or_constm1_operand): New.
4418 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4419 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4420 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4421 UNSPEC_SCATTER_PREFETCH
4423 (avx512f_ucmp<mode>3): Ditto.
4424 (avx512f_vternlog<mode>): Ditto.
4425 (avx512f_align<mode>): Ditto.
4426 (<shift_insn><mode>3): Ditto.
4427 (avx512f_<rotate>v<mode>): Ditto.
4428 (avx512f_<rotate><mode>): Ditto.
4429 (avx512f_eq<mode>3): Ditto.
4430 (avx512f_eq<mode>3_1): Ditto.
4431 (avx512f_gt<mode>3): Ditto.
4432 (avx512f_testm<mode>3): Ditto.
4433 (avx512f_testnm<mode>3): Ditto.
4434 (avx512pf_gatherpf<mode>): Ditto.
4435 (*avx512pf_gatherpf<mode>_mask): Ditto.
4436 (*avx512pf_gatherpf<mode>): Ditto.
4437 (avx512pf_scatterpf<mode>): Ditto.
4438 (*avx512pf_scatterpf<mode>_mask): Ditto.
4439 (*avx512pf_scatterpf<mode>): Ditto.
4440 (avx512f_vec_dup_gpr<mode>): Ditto.
4441 (clz<mode>2): Ditto.
4442 (conflict<mode>): Ditto.
4443 (REDUC_SMINMAX_MODE): Extened with wider modes.
4444 (reduc_<code>_<mode>): Ditto.
4445 (vlshr<mode>3): Ditto.
4446 (vashl<mode>3): Ditto.
4448 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4449 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4450 Sergey Lega <sergey.s.lega@intel.com>
4451 Anna Tikhonova <anna.tikhonova@intel.com>
4452 Ilya Tocar <ilya.tocar@intel.com>
4453 Andrey Turetskiy <andrey.turetskiy@intel.com>
4454 Ilya Verbin <ilya.verbin@intel.com>
4455 Kirill Yukhin <kirill.yukhin@intel.com>
4456 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4458 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4461 (avx512fmaskmode): Ditto.
4462 (bcstscalarsuff): Ditto.
4463 (avx512f_blendm<mode>): Ditto.
4464 (cmp_imm_predicate): Ditto.
4465 (avx512f_cmp<mode>3): Ditto.
4466 (avx512f_vec_dup<mode>): Ditto.
4467 (avx512f_vec_dup_mem<mode>): Ditto.
4468 (avx512f_vpermi2var<mode>3): Ditto.
4469 (avx512f_vpermt2var<mode>3): Ditto.
4470 (vec_init<mode>): Ditto.
4471 (avx512f_gathersi<mode>): Ditto.
4472 (*avx512f_gathersi<mode>): Ditto.
4473 (*avx512f_gathersi<mode>_2): Ditto.
4474 (avx512f_gatherdi<mode>): Ditto.
4475 (*avx512f_gatherdi<mode>): Ditto.
4476 (*avx512f_gatherdi<mode>_2): Ditto.
4477 (avx512f_scattersi<mode>): Ditto.
4478 (*avx512f_scattersi<mode>): Ditto.
4479 (avx512f_scatterdi<mode>): Ditto.
4480 (*avx512f_scatterdi<mode>): Ditto.
4481 (sseintprefix): Extened with wider modes.
4482 (VEC_GATHER_IDXSI): Ditto.
4483 (VEC_GATHER_IDXDI): Ditto.
4484 (VEC_GATHER_SRCDI): Ditto.
4486 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4487 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4489 * config/arm/t-aprofile: New file.
4490 * config.gcc: Handle --with-multilib-list option.
4492 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4494 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4497 2013-10-15 Richard Biener <rguenther@suse.de>
4499 * tree-loop-distribution.c (build_empty_rdg): Inline into
4501 (rdg_flag_vertex): Inline into single user.
4502 (rdg_flag_vertex_and_dependent): Likewise.
4503 (build_rdg_partition_for_vertex): Remove processed bitmap.
4504 (rdg_build_partitions): Simplify.
4506 2013-10-15 Richard Biener <rguenther@suse.de>
4508 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4509 Restructure forwarding through conversions and copies to
4510 avoid performing copy-propagation the wrong way. Adjust
4511 recursion invocations.
4512 (forward_propagate_addr_expr): Add argument stating if we
4513 are recursing from a single-use.
4514 (ssa_forward_propagate_and_combine): Adjust.
4516 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4518 * dumpfile.h (gcc::dump_manager): New class, to hold state
4519 relating to dumpfile management.
4520 (get_dump_file_name): Remove in favor of method of dump_manager.
4521 (dump_initialized_p): Likewise.
4522 (dump_start): Likewise.
4523 (dump_finish): Likewise.
4524 (dump_switch_p): Likewise.
4525 (dump_register): Likewise.
4526 (get_dump_file_info): Likewise.
4527 * context.c (gcc::context::context): Construct the dump_manager
4529 * context.h (gcc::context::get_dumps): New.
4530 (gcc::context::m_dumps): New.
4531 * coverage.c (coverage_init): Port to dump_manager API.
4532 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4533 (extra_dump_files_in_use): Likewise.
4534 (extra_dump_files_alloced): Likewise.
4535 (gcc::dump_manager::dump_manager): New.
4536 (dump_register): Convert to...
4537 (gcc::dump_manager::dump_register): ...method, replacing
4538 function-static next_dump with m_next_dump field.
4539 (get_dump_file_info): Convert to...
4540 (gcc::dump_manager::get_dump_file_info): ...method.
4541 (get_dump_file_name): Convert to...
4542 (gcc::dump_manager::get_dump_file_name): ...method.
4543 (dump_start): Convert to...
4544 (gcc::dump_manager::dump_start): ...method.
4545 (dump_finish): Convert to...
4546 (gcc::dump_manager::dump_finish): ...method.
4547 (dump_begin): Replace body with...
4548 (gcc::dump_manager::dump_begin): ...new method.
4549 (dump_phase_enabled_p): Convert to...
4550 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4551 (dump_phase_enabled_p): Convert to...
4552 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4553 (dump_initialized_p): Convert to...
4554 (gcc::dump_manager::dump_initialized_p): ...method.
4555 (dump_flag_name): Replace body with...
4556 (gcc::dump_manager::dump_flag_name): ...new method.
4557 (dump_enable_all): Convert to...
4558 (gcc::dump_manager::dump_enable_all): ...new method.
4559 (opt_info_enable_passes): Convert to...
4560 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4561 (dump_switch_p_1): Convert to...
4562 (gcc::dump_manager::dump_switch_p_1): ...new method.
4563 (dump_switch_p): Convert to...
4564 (gcc::dump_manager::dump_switch_p): ...new method.
4565 (opt_info_switch_p): Port to dump_manager API.
4566 (enable_rtl_dump_file): Likewise.
4567 * opts-global.c (handle_common_deferred_options): Port to new
4569 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4570 (pass_manager::register_one_dump_file): Likewise.
4571 (pass_manager::register_pass): Likewise.
4572 (pass_init_dump_file): Likewise.
4573 (pass_fini_dump_file): Likewise.
4574 * statistics.c (statistics_early_init): Likewise.
4576 2013-10-14 Richard Biener <rguenther@suse.de>
4578 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4579 iterative_hash_canonical_type, gimple_canonical_type_hash,
4580 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4581 gimple_register_canonical_type, print_gimple_types_stats,
4582 free_gimple_type_tables): Move to lto/lto.c
4583 (gt-gimple.h): Do not include.
4584 * gimple.h (gimple_register_canonical_type,
4585 print_gimple_types_stats, free_gimple_type_tables): Remove.
4586 * Makefile.in (GTFILES): Remove gimple.c.
4588 2013-10-14 Travis Snoozy <quandary@remstate.com>
4591 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4592 scanning for msp430x targets.
4594 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4597 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4598 (registers_ok_for_ldd_peep): Move around.
4599 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4600 * config/sparc/sparc.md (widening peepholes): Use it.
4602 2013-10-14 Richard Biener <rguenther@suse.de>
4606 * gimple.c (iterative_hash_canonical_type): Make sure to
4607 record the hash into the correct hashtable slot.
4609 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4611 PR rtl-optimization/58662
4612 * combine.c (try_combine): Take into account death nodes on I2 when
4613 splitting a PARALLEL of two independent SETs. Fix dump message.
4615 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4618 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4619 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4620 * config/sh/t-sh (sh_treg_combine.o): New entry.
4621 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4622 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4623 (register_sh_passes): New function. Register sh_treg_combine pass.
4624 (sh_option_override): Invoke it.
4625 (sh_canonicalize_comparison): Handle op0_preserve_value.
4626 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4627 opportunities. Canonicalize branch condition.
4628 (nott): Allow only if pseudos can be created for non-SH2A.
4630 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4633 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4634 (ix86_expand_movmem): Replace copy_addr_to_reg with
4635 ix86_copy_addr_to_reg.
4636 (ix86_expand_setmem): Likewise.
4638 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4640 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4641 provided by ix86_fp_compare_mode instead of CCFPUmode.
4643 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4645 * config/aarch64/arm_neon.h
4646 (vtbx<1,3>_<psu>8): Fix register constriants.
4648 2013-10-11 Jeff Law <law@redhat.com>
4650 PR tree-optimization/58640
4651 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4652 threading paths that cross over two loop entry points.
4654 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4656 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4657 handle vector float as well.
4658 (*vsx_le_perm_load_v4si): Likewise.
4659 (*vsx_le_perm_store_v2di): Likewise.
4660 (*vsx_le_perm_store_v4si): Likewise.
4662 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4664 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4665 directly to circumvent subtract from splat{31} workaround.
4666 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4668 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4669 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4670 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4671 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4672 separate big and little endian logic.
4673 (*altivec_vperm_<mode>_internal): New define_insn.
4674 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4675 separate big and little endian logic.
4676 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4677 (vec_permv16qi): Add little endian logic.
4679 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4681 * doc/extend.texi (returns_nonnull): Remove arguments.
4683 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4684 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4685 Sergey Lega <sergey.s.lega@intel.com>
4686 Anna Tikhonova <anna.tikhonova@intel.com>
4687 Ilya Tocar <ilya.tocar@intel.com>
4688 Andrey Turetskiy <andrey.turetskiy@intel.com>
4689 Ilya Verbin <ilya.verbin@intel.com>
4690 Kirill Yukhin <kirill.yukhin@intel.com>
4691 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4693 * config/i386/sse.md (VI48F_256_512): New.
4694 (avx2_permvar<mode>): Change to ...
4695 (<avx2_avx512f>_permvar<mode>): This.
4697 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4698 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4699 Sergey Lega <sergey.s.lega@intel.com>
4700 Anna Tikhonova <anna.tikhonova@intel.com>
4701 Ilya Tocar <ilya.tocar@intel.com>
4702 Andrey Turetskiy <andrey.turetskiy@intel.com>
4703 Ilya Verbin <ilya.verbin@intel.com>
4704 Kirill Yukhin <kirill.yukhin@intel.com>
4705 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4707 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4708 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4709 * config/i386/sse.md (VI4_AVX): New.
4710 (sf2simodelower): Ditto.
4711 (sse2_cvtps2dq): Change to ...
4712 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4714 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4715 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4716 Sergey Lega <sergey.s.lega@intel.com>
4717 Anna Tikhonova <anna.tikhonova@intel.com>
4718 Ilya Tocar <ilya.tocar@intel.com>
4719 Andrey Turetskiy <andrey.turetskiy@intel.com>
4720 Ilya Verbin <ilya.verbin@intel.com>
4721 Kirill Yukhin <kirill.yukhin@intel.com>
4722 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4724 * config/i386/sse.md (V_512): New.
4726 (vcond<V_512:mode><VF_512:mode>): Ditto.
4727 (vcond<V_512:mode><VI_512:mode>): Ditto.
4728 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4730 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4731 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4732 Sergey Lega <sergey.s.lega@intel.com>
4733 Anna Tikhonova <anna.tikhonova@intel.com>
4734 Ilya Tocar <ilya.tocar@intel.com>
4735 Andrey Turetskiy <andrey.turetskiy@intel.com>
4736 Ilya Verbin <ilya.verbin@intel.com>
4737 Kirill Yukhin <kirill.yukhin@intel.com>
4738 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4740 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4741 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4743 (fma<mode>4): Removed condition.
4744 (fms<mode>4): Ditto.
4745 (fnma<mode>4): Ditto.
4746 (fnms<mode>4): Ditto.
4747 (fma4i_fmadd_<mode>): Ditto.
4748 (*fma_fmadd_<mode>): Ditto.
4749 (*fma_fmsub_<mode>): Ditto.
4750 (*fma_fnmadd_<mode>): Ditto.
4751 (*fma_fnmsub_<mode>): Ditto.
4752 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4753 (*fma_fmaddsub_<mode>): Ditto.
4754 (*fma_fmsubadd_<mode>): Ditto.
4755 (*fmai_fmadd_<mode>): Ditto.
4756 (*fmai_fmsub_<mode>): Ditto.
4757 (*fmai_fnmadd_<mode>): Ditto.
4758 (*fmai_fnmsub_<mode>): Ditto.
4760 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4761 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4762 Sergey Lega <sergey.s.lega@intel.com>
4763 Anna Tikhonova <anna.tikhonova@intel.com>
4764 Ilya Tocar <ilya.tocar@intel.com>
4765 Andrey Turetskiy <andrey.turetskiy@intel.com>
4766 Ilya Verbin <ilya.verbin@intel.com>
4767 Kirill Yukhin <kirill.yukhin@intel.com>
4768 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4770 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4771 (VI124_256): Changed to ...
4772 (VI124_256_48_512): This.
4773 (ssepackmode): Extended with wider modes.
4774 (<code><mode>3): Changed iterator.
4775 (*avx2_<code><mode>3): Ditto.
4776 (vec_pack_trunc_<mode>): Ditto.
4778 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4779 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4780 Sergey Lega <sergey.s.lega@intel.com>
4781 Anna Tikhonova <anna.tikhonova@intel.com>
4782 Ilya Tocar <ilya.tocar@intel.com>
4783 Andrey Turetskiy <andrey.turetskiy@intel.com>
4784 Ilya Verbin <ilya.verbin@intel.com>
4785 Kirill Yukhin <kirill.yukhin@intel.com>
4786 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4788 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4789 (VI8F_256_512): Ditto.
4790 (abs<mode>2): Changed iterator.
4791 (avx2_perm<mode>): Changed to ...
4792 (<avx2_avx512f>_perm<mode>): This.
4793 (avx2_perm<mode>_1): Changed to ...
4794 (<avx2_avx512f>_perm<mode>_1): This.
4796 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4797 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4798 Sergey Lega <sergey.s.lega@intel.com>
4799 Anna Tikhonova <anna.tikhonova@intel.com>
4800 Ilya Tocar <ilya.tocar@intel.com>
4801 Andrey Turetskiy <andrey.turetskiy@intel.com>
4802 Ilya Verbin <ilya.verbin@intel.com>
4803 Kirill Yukhin <kirill.yukhin@intel.com>
4804 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4806 * config/i386/sse.md (VI48_AVX512F): New.
4807 (VI48_AVX2): Changed to ...
4808 (VI48_AVX2_48_AVX512F): This.
4809 (avx2_ashrv<mode>): Changed to ...
4810 (<avx2_avx512f>_ashrv<mode>): This.
4811 (avx2_<shift_insn>v<mode>): Changed to ...
4812 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4814 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4815 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4816 Sergey Lega <sergey.s.lega@intel.com>
4817 Anna Tikhonova <anna.tikhonova@intel.com>
4818 Ilya Tocar <ilya.tocar@intel.com>
4819 Andrey Turetskiy <andrey.turetskiy@intel.com>
4820 Ilya Verbin <ilya.verbin@intel.com>
4821 Kirill Yukhin <kirill.yukhin@intel.com>
4822 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4824 * config/i386/sse.md (VI4_AVX512F): New.
4825 (VI8_AVX2_AVX512F): Ditto.
4826 (mul<mode>3): Extended with wider modes.
4827 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4828 (mul<mode>3): Ditto.
4829 (vec_widen_<s>mult_odd_<mode>): Ditto.
4831 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4832 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4833 Sergey Lega <sergey.s.lega@intel.com>
4834 Anna Tikhonova <anna.tikhonova@intel.com>
4835 Ilya Tocar <ilya.tocar@intel.com>
4836 Andrey Turetskiy <andrey.turetskiy@intel.com>
4837 Ilya Verbin <ilya.verbin@intel.com>
4838 Kirill Yukhin <kirill.yukhin@intel.com>
4839 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4841 * config/i386/sse.md (VI2_AVX512F): New.
4842 (VI124_AVX512F): Ditto.
4843 (sseunpackmode): Extended with wider modes.
4844 (sseunpackfltmode): Ditto.
4845 (vec_unpacks_float_hi_<mode>): Ditto.
4846 (vec_unpacks_float_lo_<mode>): Ditto.
4847 (vec_unpacku_float_hi_<mode>): Ditto.
4848 (vec_unpacku_float_lo_<mode>): Ditto.
4849 (vec_unpacks_lo_<mode>): Ditto.
4850 (vec_unpacks_hi_<mode>): Ditto.
4851 (vec_unpacku_lo_<mode>): Ditto.
4852 (vec_unpacku_hi_<mode>): Ditto.
4854 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4855 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4856 Sergey Lega <sergey.s.lega@intel.com>
4857 Anna Tikhonova <anna.tikhonova@intel.com>
4858 Ilya Tocar <ilya.tocar@intel.com>
4859 Andrey Turetskiy <andrey.turetskiy@intel.com>
4860 Ilya Verbin <ilya.verbin@intel.com>
4861 Kirill Yukhin <kirill.yukhin@intel.com>
4862 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4864 * config/i386/i386.md (multdiv): New.
4865 (multdiv_mnemonic): Ditto.
4866 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4867 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4868 (<sse>_vmdiv<mode>3): Removed.
4870 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4871 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4872 Sergey Lega <sergey.s.lega@intel.com>
4873 Anna Tikhonova <anna.tikhonova@intel.com>
4874 Ilya Tocar <ilya.tocar@intel.com>
4875 Andrey Turetskiy <andrey.turetskiy@intel.com>
4876 Ilya Verbin <ilya.verbin@intel.com>
4877 Kirill Yukhin <kirill.yukhin@intel.com>
4878 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4880 * config/i386/sse.md (V): Extended with wider modes.
4882 (ssehalfvecmode): Ditto.
4884 (ssepackfltmode): Ditto.
4885 (avx_vec_concat<mode>): Ditto.
4886 (V_256_512): New iterator.
4887 (VF2_512_256): Ditto.
4888 (si2dfmode): New attribute.
4889 (si2dfmodelower): Ditto.
4891 (concat_tg_mode): Ditto.
4892 (floatv4siv4df2): Changed to ...
4893 (float<si2dfmodelower><mode>2): This.
4894 (avx_cvtps2pd256): Changed to ...
4895 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4896 (vec_pack_trunc_v4df): Changed to ...
4897 (vec_pack_trunc_<mode>): This.
4898 (avx_vpermil<mode>): Changed to ...
4899 (<sse2_avx_avx512f>_vpermil<mode>): This.
4900 (<fixsuffix>fix_truncv8dfv8si2): New.
4901 (vec_pack_sfix_trunc_v8df): Ditto.
4902 (avx512f_rndscale<mode>): Ditto.
4903 (avx512f_roundpd512): Ditto.
4904 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4906 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4907 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4908 Sergey Lega <sergey.s.lega@intel.com>
4909 Anna Tikhonova <anna.tikhonova@intel.com>
4910 Ilya Tocar <ilya.tocar@intel.com>
4911 Andrey Turetskiy <andrey.turetskiy@intel.com>
4912 Ilya Verbin <ilya.verbin@intel.com>
4913 Kirill Yukhin <kirill.yukhin@intel.com>
4914 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4916 * config/i386/i386.md (any_fix): New iterator.
4917 (fixsuffix): New attribute.
4918 * config/i386/sse.md (VF1): Extened with wider modes.
4922 (sseintvecmodelower): Ditto.
4923 (ssescalarmode): Ditto.
4924 (ssescalarnum): Ditto.
4926 (ssexmmmode): Ditto.
4927 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4928 (<sse>_rcp<mode>2): Change iterator.
4929 (rsqrt<mode>2): Ditto.
4930 (<sse>_rsqrt<mode>2): Ditto.
4931 (avx2_vec_dup<mode>): Ditto.
4932 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4933 (round<mode>2_sfix): Ditto.
4934 (avx2_pbroadcast<mode>): Ditto.
4935 (*andnot<mode>3): Handle XI mode.
4936 (*<code><mode>3): Ditto.
4937 (AVXTOSSEMODE): Removed.
4938 (avx_vpermil<mode>): Changed to ...
4939 (<sse2_avx_avx512f>_vpermil<mode>): This.
4941 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4942 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4943 Sergey Lega <sergey.s.lega@intel.com>
4944 Anna Tikhonova <anna.tikhonova@intel.com>
4945 Ilya Tocar <ilya.tocar@intel.com>
4946 Andrey Turetskiy <andrey.turetskiy@intel.com>
4947 Ilya Verbin <ilya.verbin@intel.com>
4948 Kirill Yukhin <kirill.yukhin@intel.com>
4949 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4951 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4952 (<sse>_comi): Ditto.
4953 (<sse>_ucomi): Ditto.
4954 (sse_cvtss2siq_2): Ditto.
4955 (sse2_cvtsd2si): Ditto.
4956 (sse2_cvtsd2siq): Ditto.
4957 (sse2_cvttsd2si): Ditto.
4958 (sse2_cvttsd2siq): Ditto.
4959 (<shift_insn><mode>3): Ditto.
4960 (sse2_cvtsi2sdq): Update constraint and prefix.
4961 (sse_cvtsi2ss): Update prefix.
4962 (sse_cvtsi2ssq): Ditto.
4964 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4966 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4967 ignore internal calls.
4969 2013-10-11 Richard Biener <rguenther@suse.de>
4971 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4972 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4973 TREE_OVERFLOW is set.
4975 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4977 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4979 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4980 * gimple.h: Likewise.
4982 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4984 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4987 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4989 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4990 * configure: Regenerate.
4992 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4994 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4995 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4996 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4997 similarly pp_right_paren.
4998 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4999 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5000 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5001 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5003 (omp_declare_simd_clauses_equal,
5004 omp_remove_redundant_declare_simd_attrs): New functions.
5005 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5006 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5007 * tree.h (OMP_LOOP_CHECK): Define.
5008 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5009 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5010 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5011 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5012 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5013 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5014 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5015 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5016 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5017 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5018 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5019 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5020 (omp_remove_redundant_declare_simd_attrs): New prototype.
5021 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5022 GIMPLE_OMP_TEAMS): New codes.
5023 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5024 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5026 (target_nesting_level): New variable.
5027 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5028 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5029 for collapse > 1 static schedule unless ordered.
5030 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5031 (determine_parallel_type): Adjust get_ws_args_for caller.
5032 (install_var_field): Handle mask & 4 for double indirection.
5033 (scan_sharing_clauses): Ignore shared clause on teams construct.
5034 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5035 (create_omp_child_function): If inside target or declare target
5036 constructs, set "omp declare target" attribute on the child function.
5037 (find_combined_for): New function.
5038 (scan_omp_parallel): Handle combined loops.
5039 (scan_omp_target, scan_omp_teams): New functions.
5040 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5041 restrictions and set ctx->cancellable for cancellable constructs.
5042 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5043 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5044 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5045 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5046 (omp_clause_aligned_alignment): New function.
5047 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5048 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5049 on teams constructs. Handle user defined reductions and new
5051 (lower_reduction_clauses): Don't set placeholder to address of ref
5052 if it has already the right type.
5053 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5054 (expand_parallel_call): Use the new non-_start suffixed builtins,
5055 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5056 and GOMP_parallel_end after the call.
5057 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5058 (expand_omp_for_init_counts): Handle combined loops.
5059 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5061 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5062 end of cancellable loops.
5063 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5064 Likewise. Handle collapse > 1 loops.
5065 (expand_omp_simd): Handle combined loops.
5066 (expand_omp_for): Add inner_stmt argument, adjust callers of
5067 expand_omp_for* functions, use expand_omp_for_static*chunk even
5068 for collapse > 1 unless ordered.
5069 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5070 of cancellable sections.
5071 (expand_omp_single): Remove need_barrier variable, just rely on
5072 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5073 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5074 (expand_omp_atomic_load, expand_omp_atomic_store,
5075 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5076 (expand_omp_target): New function.
5077 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5078 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5079 (build_omp_regions_1): Immediately close region for
5080 GF_OMP_TARGET_KIND_UPDATE.
5081 (maybe_add_implicit_barrier_cancel): New function.
5082 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5084 (lower_omp_single): Likewise. Add clobber after the barrier.
5085 (lower_omp_taskgroup): New function.
5086 (lower_omp_for): Handle combined loops. Adjust
5087 lower_rec_input_clauses caller. Handle cancellation.
5088 (lower_depend_clauses): New function.
5089 (lower_omp_taskreg): Lower depend clauses. Adjust
5090 lower_rec_input_clauses caller. Add clobber after the call. Handle
5092 (lower_omp_target, lower_omp_teams): New functions.
5093 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5094 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5095 and GOMP_cancellation_point calls.
5096 (lower_omp): Fold stmts inside of target region.
5097 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5098 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5099 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5100 (BT_FN_VOID_OMPFN_PTR_UINT,
5101 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5102 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5103 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5104 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5105 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5106 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5107 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5108 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5109 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5110 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5111 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5112 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5113 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5114 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5115 BUILT_IN_GOMP_PARALLEL_END.
5116 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5117 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5118 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5119 GF_OMP_FOR_KIND_DISTRIBUTE.
5120 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5121 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5122 (dump_gimple_omp_return): Print lhs if it has any.
5123 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5124 gimple_omp_atomic_seq_cst_p.
5125 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5126 and GIMPLE_OMP_TEAMS.
5127 * langhooks.c (lhd_omp_mappable_type): New function.
5128 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5129 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5131 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5132 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5133 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5134 (struct gimplify_omp_ctx): Add combined_loop field.
5135 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5136 on stmts inside of target region.
5137 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5138 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5139 ORT_TARGET and ORT_TARGET_DATA.
5140 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5142 (omp_notice_threadprivate_variable): Complain about threadprivate
5143 variables in target region.
5144 (omp_notice_variable): Complain about vars with non-mappable type
5145 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5146 (omp_check_private): Ignore ORT_TARGET* regions.
5147 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5148 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5149 (find_combined_omp_for): New function.
5150 (gimplify_omp_for): Handle gimplification of combined loops.
5151 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5153 (gimplify_omp_target_update): New function.
5154 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5155 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5156 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5157 (gimplify_body): If fndecl has "omp declare target" attribute, add
5158 implicit ORT_TARGET context around it.
5159 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5160 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5161 * tree-nested.c (convert_nonlocal_reference_stmt,
5162 convert_local_reference_stmt, convert_gimple_call): Handle
5163 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5164 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5165 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5166 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5167 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5168 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5169 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5170 BUILT_IN_GOMP_LOOP_END_CANCEL,
5171 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5172 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5173 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5174 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5175 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5176 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5177 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5178 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5179 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5180 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5181 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5182 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5183 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5184 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5185 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5186 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5187 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5188 gimple_build_omp_teams): New functions.
5189 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5190 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5191 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5192 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5193 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5194 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5195 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5196 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5197 GF_OMP_ATOMIC_SEQ_CST): New.
5198 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5199 gimple_build_omp_teams): New prototypes.
5200 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5201 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5202 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5203 GIMPLE_OMP_SINGLE as end of range.
5204 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5205 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5206 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5207 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5208 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5209 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5210 gimple_omp_target_kind, gimple_omp_target_set_kind,
5211 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5212 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5213 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5214 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5215 gimple_omp_teams_set_clauses): New inlines.
5216 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5217 and GIMPLE_OMP_TASKGROUP.
5218 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5219 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5220 enum omp_clause_proc_bind_kind): New.
5221 (union omp_clause_subcode): Add depend_kind, map_kind and
5222 proc_bind_kind fields.
5223 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5224 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5225 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5226 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5227 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5229 2013-10-10 Teresa Johnson <tejohnson@google.com>
5231 * predict.c (tree_estimate_probability): Add new parameter
5232 for estimate_bb_frequencies.
5233 (estimate_bb_frequencies): Add new parameter to force estimation.
5234 (rebuild_frequencies): When max frequency in function is small,
5235 recompute counts from frequencies.
5236 * predict.h (estimate_bb_frequencies): New parameter.
5238 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5240 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5241 optimizations are disabled.
5243 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5245 * coverage.c (coverage_finish): Fix leak of da_file_name.
5247 2013-10-10 Jan Hubicka <jh@suse.cz>
5249 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5250 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5253 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5256 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5257 if any labels are in FALLTHRU_BB, use a special label emitted
5258 immediately after the asm goto insn rather than label_rtx
5260 (expand_asm_stmt): Adjust caller.
5261 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5262 edge if the last insn in predecessor is a jump with single successor,
5263 but it isn't simplejump_p.
5265 2013-10-10 Richard Biener <rguenther@suse.de>
5267 PR tree-optimization/58656
5268 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5270 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5272 * gimplify.c: Include expr.h and tm_p.h for targets with special
5273 va-arg padding requirements.
5275 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5277 * tree-flow.h: Move some prototypes to gimple.h.
5278 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5279 * gimple.h: Relocate some prototypes from tree-flow.h
5280 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5282 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5283 (build_va_arg_indirect_ref): Relocate and make static.
5284 (std_gimplify_va_arg_expr): Relocate here.
5285 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5286 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5288 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5290 * doc/md.texi: Document the mnemonic attribute.
5292 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5295 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
5296 (set_attr_alternative x ...) when searching for user defined
5299 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5301 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
5303 2013-10-09 Easwaran Raman <eraman@google.com>
5305 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
5306 * cfgexpand.c (defer_stack_allocation): ...use here
5307 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
5309 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
5311 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
5312 (operand_equal_for_value_replacement): New function, extracted from
5313 value_replacement and enhanced to catch more cases.
5314 (value_replacement): Use operand_equal_for_value_replacement.
5316 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5318 * loop-doloop.c (doloop_modify, doloop_optimize): Use
5319 get_max_loop_iterations.
5321 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5323 * config/arm/aarch-common.c (arm_early_load_addr_dep):
5324 Place comment above function.
5326 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5328 * tree-flow.h: Remove all remaining prototypes, enums and structs that
5329 are not related to tree-cfg.c.
5330 * tree-ssa-address.h: New file. Relocate prototypes.
5331 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
5332 (addr_for_mem_ref): New. Combine call to get_address_description and
5333 return addr_for_mem_ref.
5334 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
5335 * tree-ssa-live.h: Adjust prototypes.
5336 * passes.c: Include tree-ssa-live.h.
5337 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
5338 * graphite.c (graphite_transform_loops): Make static.
5339 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5340 make_pass_graphite, pass_data_graphite_transforms,
5341 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
5342 * ipa-pure-const.c (warn_function_noreturn): Make static.
5343 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5344 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5345 Relocate from tree-cfg.c
5346 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
5348 (execute_warn_function_noreturn, gate_warn_function_noreturn,
5349 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
5350 Move to ipa-pure-const.c.
5351 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
5352 Relocate from tree-optimize.c.
5353 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
5354 make_pass_fixup_cfg): Move to tree-cfg.c.
5355 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
5356 Relocate some prototypes.
5357 * tree-data-ref.h (tree_check_data_deps) Add prototype.
5358 * tree-dump.c (dump_function_to_file): Remove prototype.
5359 Add tree-flow.h to the include file.
5360 * tree-dump.h: Remove prototype.
5361 * tree-parloops.h: New File. Add prototypes.
5362 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5363 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5364 from tree-ssa-loop.c.
5365 * tree-predcom.c (run_tree_predictive_commoning,
5366 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5367 Relocate here from tree-ssa-loop.c.
5368 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5369 ssa_name_values.release ().
5370 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5371 (ssa_name_values): Relocate from tree-flow.h.
5372 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5373 * tree-ssa-loop.c (run_tree_predictive_commoning,
5374 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5375 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5376 make_pass_graphite, pass_data_graphite_transforms,
5377 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5378 tree_parallelize_loops, pass_data_parallelize_loops,
5379 make_pass_parallelize_loops): Move to other files.
5380 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5382 * tree.h: Remove prototypes from tree-address.c.
5384 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5386 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5387 (struct int_tree_map): Move to tree-hasher.h
5388 (SCALE, LABEL, PERCENT): Move to gimple.h
5389 * tree-flow-inline.h: Delete. Move functions to other files.
5390 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5391 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5392 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5393 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5394 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5395 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5396 * tree-hasher.h: Don't include tree-flow.h.
5397 (struct int_tree_map): Relocate from tree-flow.h.
5398 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5399 tree-flow-inline.h and make static.
5400 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5402 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5403 tree-flow-inline.h and make static.
5404 * tree.h (is_global_var, may_be_aliased): Relocate from
5406 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5407 * value-prof.c: No longer include tree-flow-inline.h.
5408 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5410 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5412 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5413 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5414 (enum move_pos): Move to tree-ssa-loop-im.h
5415 * cfgloop.h: Move some prototypes.
5416 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5417 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5418 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5419 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5420 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5421 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5422 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5423 (enum move_pos): Relocate here.
5424 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5426 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5427 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5428 make_pass_lim): Relocate here from tree-ssa-loop.c.
5429 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5431 (loop_edge_to_cancel, unloop_loops): Make static.
5432 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5433 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5434 (tree_complete_unroll, gate_tree_complete_unroll,
5435 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5436 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5437 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5438 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5439 (stmt_invariant_in_loop_p): Remove unused function.
5440 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5441 * tree-ssa-loop-manip.h: New file. Add prototypes.
5442 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5443 (gcov_type_to_double_int): Move to cfgloop.h.
5444 (double_int_cmp, bound_index,
5445 estimate_numbers_of_iterations_loop): Make static.
5446 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5447 (max_loop_iterations): Factor out get_max_loop_iterations.
5448 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5450 * tree-ssa-loop-niter.h: New file. Add prototypes.
5451 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5452 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5453 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5454 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5455 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5456 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5457 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5458 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5459 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5460 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5461 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5462 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5463 pass_data_complete_unroll, make_pass_complete_unroll,
5464 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5465 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5466 tree-ssa-loop-ivcanon.c.
5467 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5468 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5469 tree-ssa-loop-prefetch.c.
5470 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5472 (get_lsm_tmp_name): Relocate and add suffix parameter.
5473 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5474 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5475 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5476 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5477 (get_estimated_loop_iterations): Factor out accessor from
5478 estimated_loop_iterations in tree-ssa-loop-niter.c.
5479 (get_max_loop_iterations): Factor out accessor from
5480 _max_loop_iterations in tree-ssa-niter.c.
5481 * loop-unroll.c (decide_unroll_constant_iterations,
5482 decide_unroll_runtime_iterations, decide_peel_simple,
5483 decide_unroll_stupid): Use new get_* accessors.
5485 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5487 PR tree-optimization/20318
5488 * doc/extend.texi (returns_nonnull): New function attribute.
5489 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5491 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5492 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5494 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5497 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5498 false if both components are bitfields.
5500 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5502 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5503 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5504 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5505 * config/aarch64/aarch64.h
5506 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5507 * config/aarch64/aarch64-simd-builtins.def
5508 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5510 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5512 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5514 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5516 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5518 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5520 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5521 (vneg_f64): New intrinsic.
5522 (vneg_s8): Asm replaced with C.
5523 (vneg_s16): Likewise.
5524 (vneg_s32): Likewise.
5525 (vneg_s64): New intrinsic.
5526 (vnegq_f32): Asm replaced with C.
5527 (vnegq_f64): Likewise.
5528 (vnegq_s8): Likewise.
5529 (vnegq_s16): Likewise.
5530 (vnegq_s32): Likewise.
5531 (vnegq_s64): Likewise.
5533 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5535 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5537 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5539 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5540 packed stack special handling.
5541 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5542 back to fixed stack slots for FPRs saved due to stdarg.
5544 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5546 * config/s390/s390.c (s390_frame_info): Restructure function.
5548 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5550 * config/s390/s390.c (struct s390_frame_layout): New field
5552 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5553 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5554 regs_ever_clobbered to char*.
5555 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5556 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5557 (s390_register_info): Enable FPR save slots. Move/Copy some
5558 functionality into ...
5559 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5560 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5562 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5563 not imply a stack frame.
5564 (s390_init_frame_layout): Remove variable clobbered_regs.
5565 (s390_update_register_info): Remove function.
5566 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5567 cannot be used for register renaming.
5568 (s390_hard_regno_scratch_ok): New function.
5569 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5570 (s390_initial_elimination_offset): Change offset calculation of
5571 the return address pointer.
5572 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5573 (restore_gprs): Set frame related flag.
5574 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5575 (s390_emit_prologue): Call s390_register_info instead of
5576 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5577 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5578 (s390_optimize_prologue): Call s390_optimize_register_info.
5579 Try to remove also FPR slot save/restore INSNs. Remove frame
5580 related flags from restore INSNs.
5582 2013-10-08 DJ Delorie <dj@redhat.com>
5584 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5587 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5588 avoid conflict with the MI use of %c.
5589 * config/rl78/rl78-real.md: change %c to %C throughout.
5590 * config/rl78/rl78-virt.md: Likewise.
5592 2013-10-08 Jan Hubicka <jh@suse.cz>
5594 * config/i386/i386.c (ix86_option_override_internal): Switch
5595 to SSE math for -ffast-math when target ISA supports SSE2.
5597 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5599 * tree-flow.h: Remove some prototypes.
5600 * tree.h: Remove some protypes, add a couple.
5601 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5602 using_eh_for_cleanups_p): Add interface routines for front ends.
5603 * tree-eh.h: New file. Add protoptyes.
5604 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5605 (add_stmt_to_eh_lp_fn): Make static.
5606 (lower_try_finally): Use new using_eh_for_cleanups_p.
5607 * emit-rtl.c: Include tree-eh.h.
5608 * gimple.h: Include tree-eh.h.
5610 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5612 PR tree-optimization/58480
5613 * tree-vrp.c (infer_nonnull_range): New function.
5614 (infer_value_range): Call infer_nonnull_range.
5616 2013-10-08 Dehao Chen <dehao@google.com>
5618 PR tree-optimization/58619
5619 * tree-inline.c (copy_phis_for_bb): Combine location data
5622 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5625 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5626 RTX_FRAME_RELATED_P on INSN.
5628 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5630 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5631 (altivec_expand_vec_perm_const): Call it.
5633 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5635 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5636 sequences for LE VSX loads and stores at expand time.
5637 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5639 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5640 (rs6000_gen_le_vsx_permute): New.
5641 (rs6000_gen_le_vsx_load): New.
5642 (rs6000_gen_le_vsx_store): New.
5643 (rs6000_gen_le_vsx_move): New.
5644 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5645 (*vsx_le_perm_load_v4si): New.
5646 (*vsx_le_perm_load_v8hi): New.
5647 (*vsx_le_perm_load_v16qi): New.
5648 (*vsx_le_perm_store_v2di): New.
5649 (*vsx_le_perm_store_v4si): New.
5650 (*vsx_le_perm_store_v8hi): New.
5651 (*vsx_le_perm_store_v16qi): New.
5652 (*vsx_xxpermdi2_le_<mode>): New.
5653 (*vsx_xxpermdi4_le_<mode>): New.
5654 (*vsx_xxpermdi8_le_V8HI): New.
5655 (*vsx_xxpermdi16_le_V16QI): New.
5656 (*vsx_lxvd2x2_le_<mode>): New.
5657 (*vsx_lxvd2x4_le_<mode>): New.
5658 (*vsx_lxvd2x8_le_V8HI): New.
5659 (*vsx_lxvd2x16_le_V16QI): New.
5660 (*vsx_stxvd2x2_le_<mode>): New.
5661 (*vsx_stxvd2x4_le_<mode>): New.
5662 (*vsx_stxvd2x8_le_V8HI): New.
5663 (*vsx_stxvd2x16_le_V16QI): New.
5665 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5667 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5669 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5671 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5672 loop to work also for 31bit ABI.
5673 Save the stack pointer for frame_size > 0.
5675 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5677 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5678 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5679 constraint letters from expanders.
5680 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5681 retry count to general_operand.
5682 ("tabort"): Give operand 0 a mode.
5683 ("tabort_1"): Add mode and constraint letter for operand 0.
5684 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5686 2013-10-04 Jeff Law <law@redhat.com>
5688 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5690 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5692 (thread_across_edge): Here. Call it.
5694 2013-10-04 Cary Coutant <ccoutant@google.com>
5696 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5697 discarding a location list expression (or a piece of one).
5699 2013-10-03 Jan Hubicka <jh@suse.cz>
5701 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5702 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5703 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5705 2013-10-03 Jan Hubicka <jh@suse.cz>
5707 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5708 accumulate-outgoing-args when producing unwind info.
5710 2013-10-03 Wei Mi <wmi@google.com>
5712 * lra-constraints.c (insert_move_for_subreg): New function
5713 extracted from simplify_operand_subreg.
5714 (simplify_operand_subreg): Add reload for paradoxical subreg.
5716 2013-10-03 Rong Xu <xur@google.com>
5718 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5719 the candidate of builtin_expect such that we should fix the
5720 size/time estimation.
5721 (estimate_function_body_sizes): Do the acutally size/time fix-up
5724 2013-10-03 Rong Xu <xur@google.com>
5726 * predict.c (tree_predict_by_opcode): Get the probability
5727 for builtin_expect from param builtin_expect_probability.
5728 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5729 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5730 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5732 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5735 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5736 * calls.c (alloca_call_p): Use get_callee_fndecl.
5737 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5738 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5740 (vrp_visit_stmt): Remove duplicated code.
5742 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5744 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5745 ceildf2, btruncdf2, instead of vsx_* name.
5747 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5748 iterators to only do V2DF and V4SF here. Move the DF code to
5749 rs6000.md where it is combined with SF mode. Replace <VSv> with
5750 just 'v' since only vector operations are handled with these insns
5751 after moving the DF support to rs6000.md.
5752 (vsx_sub<mode>3): Likewise.
5753 (vsx_mul<mode>3): Likewise.
5754 (vsx_div<mode>3): Likewise.
5755 (vsx_fre<mode>2): Likewise.
5756 (vsx_neg<mode>2): Likewise.
5757 (vsx_abs<mode>2): Likewise.
5758 (vsx_nabs<mode>2): Likewise.
5759 (vsx_smax<mode>3): Likewise.
5760 (vsx_smin<mode>3): Likewise.
5761 (vsx_sqrt<mode>2): Likewise.
5762 (vsx_rsqrte<mode>2): Likewise.
5763 (vsx_fms<mode>4): Likewise.
5764 (vsx_nfma<mode>4): Likewise.
5765 (vsx_copysign<mode>3): Likewise.
5766 (vsx_btrunc<mode>2): Likewise.
5767 (vsx_floor<mode>2): Likewise.
5768 (vsx_ceil<mode>2): Likewise.
5769 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5770 (vsx_sminsf3): Likewise.
5771 (vsx_fmadf4): Likewise.
5772 (vsx_fmsdf4): Likewise.
5773 (vsx_nfmadf4): Likewise.
5774 (vsx_nfmsdf4): Likewise.
5775 (vsx_cmpdf_internal1): Likewise.
5777 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5778 simpler to select whether a target has SPE or traditional floating
5779 point support in iterators.
5780 (TARGET_DF_SPE): Likewise.
5781 (TARGET_SF_FPR): Likewise.
5782 (TARGET_DF_FPR): Likewise.
5783 (TARGET_SF_INSN): Macros to say whether floating point support
5784 exists for a given operation for expanders.
5785 (TARGET_DF_INSN): Likewise.
5787 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5788 combining of SF/DF mode operations, using both traditional and VSX
5796 (abs<mode>2): Combine SF/DF modes using traditional floating point
5797 instructions. Add support for using the upper DF registers with
5798 VSX support, and SF registers with power8-vector support. Update
5799 expanders for operations supported by both the SPE and traditional
5800 floating point units.
5801 (abs<mode>2_fpr): Likewise.
5802 (nabs<mode>2): Likewise.
5803 (nabs<mode>2_fpr): Likewise.
5804 (neg<mode>2): Likewise.
5805 (neg<mode>2_fpr): Likewise.
5806 (add<mode>3): Likewise.
5807 (add<mode>3_fpr): Likewise.
5808 (sub<mode>3): Likewise.
5809 (sub<mode>3_fpr): Likewise.
5810 (mul<mode>3): Likewise.
5811 (mul<mode>3_fpr): Likewise.
5812 (div<mode>3): Likewise.
5813 (div<mode>3_fpr): Likewise.
5814 (sqrt<mode>3): Likewise.
5815 (sqrt<mode>3_fpr): Likewise.
5816 (fre<Fs>): Likewise.
5817 (rsqrt<mode>2): Likewise.
5818 (cmp<mode>_fpr): Likewise.
5819 (smax<mode>3): Likewise.
5820 (smin<mode>3): Likewise.
5821 (smax<mode>3_vsx): Likewise.
5822 (smin<mode>3_vsx): Likewise.
5823 (negsf2): Delete SF operations that are merged with DF.
5830 (fmasf4_fpr): Likewise.
5831 (fmssf4_fpr): Likewise.
5832 (nfmasf4_fpr): Likewise.
5833 (nfmssf4_fpr): Likewise.
5834 (sqrtsf2): Likewise.
5835 (rsqrtsf_internal1): Likewise.
5836 (smaxsf3): Likewise.
5837 (sminsf3): Likewise.
5838 (cmpsf_internal1): Likewise.
5839 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5840 (negdf2): Delete DF operations that are merged with SF.
5842 (nabsdf2): Likewise.
5848 (rsqrtdf_internal1): Likewise.
5849 (fmadf4_fpr): Likewise.
5850 (fmsdf4_fpr): Likewise.
5851 (nfmadf4_fpr): Likewise.
5852 (nfmsdf4_fpr): Likewise.
5853 (sqrtdf2): Likewise.
5854 (smaxdf3): Likewise.
5855 (smindf3): Likewise.
5856 (cmpdf_internal1): Likewise.
5857 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5858 (btrunc<mode>2): Delete separate expander, and combine with the
5859 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5860 (btrunc<mode>2_fpr): Likewise.
5861 (ceil<mode>2): Likewise.
5862 (ceil<mode>2_fpr): Likewise.
5863 (floor<mode>2): Likewise.
5864 (floor<mode>2_fpr): Likewise.
5865 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5866 Add support for using the upper registers with VSX and
5867 power8-vector. Move insns to be closer to the define_expands. On
5868 VSX systems, prefer the traditional form of FMA over the VSX
5869 version, since the traditional form allows the target not to
5870 overlap with the inputs.
5871 (fms<mode>4_fpr): Likewise.
5872 (nfma<mode>4_fpr): Likewise.
5873 (nfms<mode>4_fpr): Likewise.
5875 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5876 Richard Earnshaw <richard.earnshaw@arm.com>
5878 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5879 (struct mult_cost_table): Likewise.
5880 (struct mem_cost_table): Likewise.
5881 (struct fp_cost_table): Likewise.
5882 (struct vector_cost_table): Likewise.
5883 (cpu_cost_table): Likewise.
5884 * config/arm/arm.opt (mold-rts-costs): New option.
5885 (mnew-generic-costs): Likewise.
5886 * config/arm/arm.c (generic_extra_costs): New table.
5887 (cortexa15_extra_costs): Likewise.
5888 (arm_slowmul_tune): Use NULL as new costs.
5889 (arm_fastmul_tune): Likewise.
5890 (arm_strongarm_tune): Likewise.
5891 (arm_xscale_tune): Likewise.
5892 (arm_9e_tune): Likewise.
5893 (arm_v6t2_tune): Likewise.
5894 (arm_cortex_a5_tune): Likewise.
5895 (arm_cortex_a9_tune): Likewise.
5896 (arm_v6m_tune): Likewise.
5897 (arm_fa726te_tune): Likewise.
5898 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5899 (arm_cortex_tune): Use generict_extra_costs.
5900 (shifter_op_p): New function.
5901 (arm_unspec_cost): Likewise.
5902 (LIBCALL_COST): Define.
5903 (arm_new_rtx_costs): New function.
5904 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5905 table is available. Use old costs otherwise unless mnew-generic-costs
5907 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5908 (cpu_cost_table): Declare.
5910 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5913 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5914 (*subs_mul_imm_<mode>)
5915 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5916 (*sub_<shift>_<mode>)
5917 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5918 Remove k constraint.
5920 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5922 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5924 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5926 2013-10-02 Teresa Johnson <tejohnson@google.com>
5928 * predict.c (probably_never_executed): New function.
5929 (probably_never_executed_bb_p): Invoke probably_never_executed.
5930 (probably_never_executed_edge_p): Ditto.
5931 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5932 Treat profile insanities conservatively.
5934 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5936 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5938 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5940 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5941 use it. Use smaller increase for scratch. Don't increase reject
5942 for early clobber scratch.
5943 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5944 setting eliminated regs except setting fp from hfp.
5945 (lra_eliminate): Check lra_insn_recog_data on NULL.
5947 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5950 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5951 setting -mvsx-timode by default until the underlying problem is fixed.
5952 (RS6000_CPU, power7 defaults): Likewise.
5954 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5956 * config/x-linux (host-linux.o): Remove header dependencies.
5957 Use $(COMPILE) and $(POSTCOMPILE).
5958 * config/t-linux-android (linux-android.o): Ditto.
5960 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5962 * Makefile.in (expmed.o-warn): Remove.
5964 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5966 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5967 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5969 2013-10-02 Teresa Johnson <tejohnson@google.com>
5971 * dojump.c (do_jump_1): Divide probability between
5972 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5974 2013-10-02 Tom Tromey <tromey@redhat.com>
5976 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5978 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5980 * tree-flow.h: Remove some prototypes.
5981 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5982 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5983 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5984 mark_virtual_phi_result_for_renaming): Relocate here.
5985 * tree-into-ssa.h: Add prototypes.
5986 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5987 single_pred_before_succ_order.
5988 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5989 (nonfreeing_call_p) Move to gimple.c.
5990 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5992 * basic-block.h (single_pred_before_succ_order): Add prototype.
5993 * gimple.c (nonfreeing_call_p): Relocate here.
5994 * gimple.h: Add prototype.
5995 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5996 * tree-ssa-dom.h: New file. Relocate prototypes here.
5997 * tree-ssa.h: Include tree-ssa-dom.h.
5999 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6001 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6002 Use $(COMPILE) and $(POSTCOMPILE).
6004 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6006 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6008 * tree-flow.h: Remove some prototypes.
6009 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6010 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6011 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6012 propagate_tree_value*): Move from here to...
6013 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6014 propagate_tree_value*): Relocate here.
6015 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6016 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6017 * gimple-fold.c: Remove gimple-fold.h from include list.
6018 * tree-vrp.c: Remove gimple-fold.h from include list.
6019 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6020 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6021 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6022 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6023 * sese.c: Add tree-ssa-propagate.h to include list.
6025 2013-10-02 Richard Biener <rguenther@suse.de>
6027 * tree-loop-distribution.c: Include tree-vectorizer.h for
6029 (enum partition_kind): Remove PKIND_REDUCTION.
6030 (struct partition_s): Remove has_writes member, add reduction_p member.
6031 (partition_alloc): Adjust.
6032 (partition_builtin_p): Likewise.
6033 (partition_has_writes): Remove.
6034 (partition_reduction_p): New function.
6035 (partition_merge_into): Likewise.
6036 (generate_code_for_partition): Commonize builtin partition
6038 (rdg_cannot_recompute_vertex_p): Remove.
6039 (already_processed_vertex_p): Likewise.
6040 (rdg_flag_vertex): Do not set has_writes.
6041 (classify_partition): Adjust.
6042 (rdg_build_partitions): Do not set has_writes, treat all
6043 partitions as useful.
6044 (distribute_loop): Record number of library calls generated. Adjust.
6045 (tree_loop_distribution): Report number of loops and library
6046 calls generated as opt-info.
6048 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6050 * tree-flow.h: Include new .h files. Move prototypes.
6051 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6052 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6053 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6054 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6055 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6056 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6057 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6058 * tree.h (get_ref_base_and_extent): Move prototype out.
6059 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6061 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6062 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6063 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6064 * tree-scalar-evolution.c: Include tree.h.
6065 * sese.c: Include tree.h.
6066 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6067 * dwarf2out.c: Include tree-dfa.h.
6068 * tree-chrec.c: Include tree.h.
6069 * tree-data-ref.c: Include tree.h.
6071 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6073 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6076 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6078 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6079 Use $(COMPILE) and $(POSTCOMPILE).
6080 (sol2-cxx.o): Likewise.
6081 (sol2-stubs.o): Likewise.
6083 * config/x-solaris (host-solaris.o): Likewise.
6085 * config/sparc/t-sparc (sparc.o): Remove.
6086 (sparc-c.o): Remove header dependencies.
6087 Use $(COMPILE) and $(POSTCOMPILE).
6088 * config/sparc/x-sparc: Likewise.
6090 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6092 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6093 * config/arc/arc700.md: Likewise.
6094 * config/arc/arc-modes.def: Likewise.
6095 * config/arc/arc-simd.h: Likewise.
6096 * config/arc/t-arc-uClibc: Likewise.
6097 * config/arc/t-arc-newlib: Likewise.
6099 2013-10-02 Renlin Li <renlin.li@arm.com>
6101 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6103 (aarch64_expand_epilogue): Likewise.
6105 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6106 Yufeng Zhang <yufeng.zhang@arm.com>
6108 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6110 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6111 'base_in' represent a conversion and legal_cast_p_1 holds; set
6112 'base_in' with the returned value from get_unwidened.
6114 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6116 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6117 plus_constant is not used.
6119 2013-10-01 Wei Mi <wmi@google.com>
6121 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6122 * config/i386/i386.md: Add define_peephole2 to
6123 break partial reg stall for cvtss2sd/cvtsd2ss.
6125 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6127 * config/arc/arc.c (pass_arc_ifcvt::clone):
6128 Update for ctxt_ -> m_ctxt change.
6130 2013-10-01 Jeff Law <law@redhat.com>
6132 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6133 outgoing_edge and intermediate_edge fields. Instead store the path.
6134 (redirection_data::hash): Hash on the last edge's destination index.
6135 (redirection_data::equal): Check the entire thread path.
6136 (lookup_redirectio_data): Corresponding changes.
6137 (create_edge_and_update_destination_phis): Likewise.
6138 (thread_single_edge): Likewise.
6140 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6141 Diego Novillo <dnovillo@google.com>
6143 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6144 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6145 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6146 versions of these patterns.
6148 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6149 (__builtin_arc_aligned): Likewise.
6151 * config/arc/arc.md: Expand adc_0 comment stating the intended
6152 purpose and why it isn't ready.
6153 Replace commented out call_value_via_label_mixed with a
6154 plain comment about bl_s.
6156 * config/arc/arc.c (stdio.h): Don't include directly.
6157 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6158 Remove if (1) condition.
6159 (arc_encode_section_info): Fix comment.
6161 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6163 * config/arc/arc.c (arc_conditional_register_usage):
6164 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6165 Also set reg_alloc_order for DMA config regs.
6167 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6168 Jeremy Bennett <jeremy.bennett@embecosm.com>
6170 * doc/install.texi (--with-cpu): Mention ARC.
6171 (arc-*-elf32): New paragraph.
6172 (arc-linux-uclibc): Likewise.
6173 * doc/md.texi (Machine Constraints): Add ARC part.
6174 * doc/invoke.texi: (menu): Add ARC Options.
6175 (Machine Dependent Options) <ARC Options>: Add synopsis.
6176 (node ARC Options): Add.
6177 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6178 (ARC Built-in Functions): New section defining
6179 generic ARC built-in functions.
6180 (ARC SIMD Built-in Functions): New section defining SIMD specific
6182 (Declaring Attributes of Functions): Extended
6183 description of short_call and long_call attributes for ARC and
6184 added index entries.
6186 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6187 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6188 Joern Rennecke <joern.rennecke@embecosm.com>
6189 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6190 Brendan Kehoe <brendan@zen.org>
6191 Michael Eager <eager@eagercon.com>
6192 Simon Cook <simon.cook@embecosm.com>
6193 Jeremy Bennett <jeremy.bennett@embecosm.com>
6195 * config/arc, common/config/arc: New directories.
6197 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6198 Brendan Kehoe <brendan@zen.org>
6199 Simon Cook <simon.cook@embecosm.com>
6201 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6203 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6205 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6206 * tree-ssa-coalesce.h: New. Move prototype to here.
6207 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6208 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6209 (gimple_can_coalesce_p): Move to...
6210 * gimple.c (gimple_can_coalesce_p): Here.
6212 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6214 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6215 (dump_decl_set): Move to gimple.c.
6216 * gimple.h: Don't include tree-ssa-operands.h.
6217 (dump_decl_set): Add prototype.
6218 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6219 Move to gimple-ssa.h.
6220 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6221 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6222 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6223 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6224 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6226 (dump_decl_set): Relocate here.
6227 * gimple-ssa.h: New file.
6228 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6229 Relocate from gimple.h.
6230 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6231 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6232 swap_tree_operands and remove non-ssa path.
6233 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6234 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6236 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6237 vect_is_simple_reduction_1): Use swap_ssa_operands.
6238 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6239 (enum need_phi_state): Move to tree-into-ssa.c.
6240 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6241 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6242 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6243 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6244 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6245 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6246 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6247 num_imm_uses): Move to ssa-iterators.h.
6248 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6249 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6251 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6252 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6253 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6254 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6255 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6256 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6257 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6258 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6259 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6260 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6261 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6262 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6263 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6264 (set_phi_nodes): Move to tree-phinodes.h.
6265 * tree-ssa-operands.h (enum ssa_op_iter_type,
6266 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6267 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6268 (dump_decl_set): Remove prototype.
6269 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6270 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6271 (set_phi_nodes): Relocate from tree-flow-inline.h.
6272 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6274 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6275 include list. Temporarily add gimple.h to include list.
6276 * ssa-iterators.h: New file.
6277 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6278 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6279 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6280 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6281 Relocate from tree-ssa-operands.h.
6282 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6283 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6284 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6285 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6286 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6287 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6288 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
6289 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
6290 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
6291 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
6292 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
6293 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
6294 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
6295 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
6296 Relocate from tree-flow-inline.h.
6297 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
6299 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
6302 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
6303 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
6304 Insert '\t' to output template.
6305 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
6306 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
6307 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
6308 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
6310 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
6312 * doc/install.texi (Host/target specific installation notes for GCC):
6313 Put @anchor before @heading.
6314 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
6315 Use @email for email addresses.
6317 2013-10-01 Jeff Law <law@redhat.com>
6319 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
6320 a vec. Only delete the path if we create one without successfully
6321 registering a jump thread.
6322 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
6324 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
6325 (paths): New vector of jump threading paths.
6326 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
6327 (THREAD_PATH): New accessor macro for the entire thread path.
6328 (lookup_redirection_data): Get intermediate and final outgoing edge
6329 from the thread path.
6330 (create_edge_and_update_destination_phis): Copy the threading path.
6331 (ssa_fix_duplicate_block_edges): Get edges and block types from the
6332 jump threading path.
6333 (ssa_redirect_edges): Get edges and block types from the jump threading
6334 path. Free the path vector.
6335 (thread_block): Get edges from the jump threading path. Look at the
6336 entire path to see if we thread to a loop exit. If we cancel a jump
6337 thread request, then free the path vector.
6338 (thread_single_edge): Get edges and block types from the jump threading
6339 path. Free the path vector.
6340 (thread_through_loop_header): Get edges and block types from the jump
6341 threading path. Free the path vector.
6342 (mark_threaded_blocks): Iterate over the vector of paths and store
6343 the path on the appropriate edge. Get edges and block types from the
6344 jump threading path.
6345 (mark_threaded_blocks): Get edges and block types from the jump
6346 threading path. Free the path vector.
6347 (thread_through_all_blocks): Use the vector of paths rather than
6348 a vector of 3-edge sets.
6349 (register_jump_thread): Accept pointer to a path vector rather
6350 than the path vector itself. Store the path vector for later use.
6353 2013-10-01 Jakub Jelinek <jakub@redhat.com>
6354 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6357 * config/s390/s390.c (s390_split_branches): Modify check for table
6359 (s390_chunkify_start): Rearrange table jump insn check in order to
6360 deal with compare and branch insns correctly.
6362 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6366 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6367 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6368 define_insn_and_split.
6369 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6370 (shiftsi3_compare): New pattern.
6372 * config/arm/unspecs.md (UNSPEC_RRX): New.
6374 2013-10-01 Alan Modra <amodra@gmail.com>
6376 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6377 casing inout operands.
6379 2013-10-01 Richard Biener <rguenther@suse.de>
6381 PR tree-optimization/58553
6382 * tree-loop-distribution.c (struct partition_s): Add niter member.
6383 (classify_partition): Populate niter member for the partition
6384 and properly identify whether the relevant store happens before
6385 or after the loop exit.
6386 (generate_memset_builtin): Use niter member from the partition.
6387 (generate_memcpy_builtin): Likewise.
6389 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6391 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6392 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6394 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6396 * basic-block.h (edge_list): Prefix member names with "m_".
6397 * context.h (context): Likewise.
6398 * domwalk.h (dom_walker): Likewise.
6399 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6400 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6401 * hash-table.h (hash_table): Likewise.
6402 * machmode.h (bit_field_mode_iterator): Likewise.
6403 * pass_manager.h (pass_list): Likewise.
6404 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6405 * tree-pass.h (pass_data): Likewise.
6406 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6407 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6408 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6409 * asan.c (pass_data_asan): Update accordingly.
6410 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6411 (control_dependences::control_dependences): Likewise.
6412 (control_dependences::~control_dependences): Likewise.
6413 (control_dependences::~control_dependences): Likewise.
6414 (control_dependences::get_edges_dependent_on): Likewise.
6415 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6416 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6417 * context.c (gcc::context::context): Likewise.
6418 * cprop.c (pass_rtl_cprop::clone): Likewise.
6419 * domwalk.c (dom_walker::walk): Likewise.
6420 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6421 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6422 * mode-switching.c (pass_mode_switching::clone): Likewise.
6423 * passes.c (opt_pass::opt_pass): Likewise.
6424 (pass_manager::pass_manager): Likewise.
6425 * predict.c (pass_strip_predict_hints::clone): Likewise.
6426 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6427 (pass_split_all_insns::clone): Likewise.
6428 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6430 (bit_field_mode_iterator::next_mode): Likewise.
6431 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6432 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6433 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6434 * tree-complex.c (pass_lower_complex::clone): Likewise.
6435 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6436 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6437 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6438 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6439 (pass_fold_builtins::clone): Likewise.
6440 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6441 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6442 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6443 (pass_cd_dce::clone): Likewise.
6444 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6445 (pass_phi_only_cprop::clone): Likewise.
6446 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6447 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6448 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6449 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6450 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6451 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6452 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6453 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6454 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6455 * tree-vrp.c (pass_vrp::clone): Likewise.
6456 * tsan.c (pass_tsan::clone): Likewise.
6458 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6461 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6462 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6465 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6466 optimization, punt if sign_bit_p looked through any zero extension.
6468 2013-09-30 Teresa Johnson <tejohnson@google.com>
6470 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6471 Update redirected out edge count in joiner case.
6472 (ssa_redirect_edges): Common the joiner and non-joiner cases
6473 so that joiner case gets profile updates.
6475 2013-09-30 Richard Biener <rguenther@suse.de>
6477 PR tree-optimization/58554
6478 * tree-loop-distribution.c (classify_partition): Require
6479 unconditionally executed stores for memcpy and memset recognition.
6480 (tree_loop_distribution): Calculate dominance info.
6482 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6484 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6485 (NO_PROFILE_COUNTERS): Likewise.
6486 (PROFILE_HOOK): Likewise.
6487 (FUNCTION_PROFILER): Likewise.
6488 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6490 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6492 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6493 calls and defines in TARGET_MACHO conditional.
6494 (load_macho_picbase_di): Likewise.
6495 (reload_macho_picbase): Likewise.
6496 (reload_macho_picbase_si): Likewise.
6497 (reload_macho_picbase_di): Likewise.
6498 (nonlocal_goto_receiver): Likewise.
6500 2013-09-30 Nick Clifton <nickc@redhat.com>
6502 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6503 that use the MSP430X ISA.
6504 (msp430_option_override): Scan -mmcu command line option for any
6505 MCU name that supports the MSP430X ISA.
6506 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6507 -mmcu options which enable the MSP430X ISA.
6509 2013-09-30 Richard Biener <rguenther@suse.de>
6512 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6513 before looking for setjmp-like calls.
6515 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6518 * config/darwin-protos.h (machopic_get_function_picbase): New.
6519 * config/darwin.c (machopic_get_function_picbase): New.
6520 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6521 label for a new func. (load_macho_picbase_di): Likewise.
6522 (reload_macho_picbase): New expand.
6523 (reload_macho_picbase_si): New insn.
6524 (reload_macho_picbase_di): New insn.
6525 (nonlocal_goto_receiver): New define and split.
6526 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6527 (unspecv enum): Add UNSPECV_NLGR.
6529 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6531 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6532 that altivec registers are correctly sized on Darwin.
6534 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6536 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6537 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6538 * config/x-darwin (host-darwin.o): Likewise.
6539 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6540 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6541 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6543 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6545 * doc/invoke.texi: Fix usage of @tie{} command.
6547 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6549 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6551 2013-09-28 Jan Hubicka <jh@suse.cz>
6553 * config/i386/x86-tune.def: Add documentation for each of the options;
6556 2013-09-28 Jan Hubicka <jh@suse.cz>
6558 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6560 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6561 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6562 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6563 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6564 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6566 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6568 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6569 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6570 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6571 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6572 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6573 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6574 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6575 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6576 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6577 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6578 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6579 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6580 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6581 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6582 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6583 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6584 gimple.h, godump.c, graphite-clast-to-gimple.c,
6585 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6586 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6587 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6588 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6589 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6590 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6591 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6592 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6593 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6594 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6595 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6596 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6597 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6598 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6599 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6600 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6601 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6602 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6603 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6604 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6605 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6606 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6607 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6608 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6609 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6610 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6611 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6612 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6613 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6614 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6615 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6616 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6617 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6618 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6619 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6620 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6621 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6622 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6623 whitespace before "(".
6625 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6627 * expr.h (extract_bit_field): Remove packedp parameter.
6628 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6629 from forward declaration.
6630 (store_split_bit_field): Remove packedp arg from calls to
6631 extract_fixed_bit_field.
6632 (extract_bit_field_1): Remove packedp parameter and packedp
6633 argument from recursive calls and calls to extract_fixed_bit_field.
6634 (extract_bit_field): Remove packedp parameter and corresponding
6635 arg to extract_bit_field_1.
6636 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6638 (extract_split_bit_field): Remove packedp arg from call to
6639 extract_fixed_bit_field.
6640 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6641 (copy_blkmode_from_reg): Likewise.
6642 (copy_blkmode_to_reg): Likewise.
6643 (read_complex_part): Likewise.
6644 (store_field): Likewise.
6645 (expand_expr_real_1): Likewise.
6646 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6647 to extract_bit_field.
6648 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6649 call to extract_bit_field.
6650 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6651 call to extract_bit_field.
6652 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6653 and special packedp behavior from -fstrict-volatile-bitfields
6656 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6658 * lra-eliminations.c (init_elim_table): Guard value_p.
6660 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6662 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6663 DFmode, DImode, and SFmode in the upper VSX registers based on the
6664 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6665 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6666 of the VSX register handling.
6668 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6670 (zero_extendsidi2_lfiwzx): Add support for loading into the
6671 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6672 only do VSX memory options on Altivec registers.
6673 (extendsidi2_lfiwax): Likewise.
6674 (extendsfdf2_fpr): Likewise.
6675 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6676 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6677 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6678 (movdi_internal64): Likewise.
6680 2013-09-27 Xinliang David Li <davidxl@google.com>
6682 * opts.c (finish_options): Adjust parameters
6683 according to vect cost model.
6684 (common_handle_option): Set dynamic vect cost
6686 targhooks.c (default_add_stmt_cost): Compute stmt cost
6688 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6689 Use helper function.
6690 * tree-vectorizer.h (unlimited_cost_model): New function.
6691 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6692 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6694 (vect_enhance_data_refs_alignment): Ditto.
6695 * flag-types.h: New enum.
6696 * common/config/i386/i386-common.c (ix86_option_init_struct):
6697 No need to initialize vect_cost_model flag.
6698 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6701 2013-09-27 Diego Novillo <dnovillo@google.com>
6703 * gimple.h (enum ssa_mode): Remove.
6705 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6707 * cfgloop.h (number_of_loops): Fix typo in check for null.
6709 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6712 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6713 are children of outermost saved_cfun's loop, and set it up to
6714 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6715 if loop != loop0 and SESE region contains bbs that belong to loop0.
6717 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6719 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6720 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6721 (set_address_segment, set_address_base, set_address_index)
6722 (set_address_disp): Accept the argument unconditionally.
6723 (baseness): Remove must_be_base_p and must_be_index_p checks.
6724 (decompose_normal_address): Classify as much as possible in the
6727 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6729 * cse.c (count_reg_usage): Handle INT_LIST.
6730 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6731 * reginfo.c (reg_scan_mark_refs): Likewise.
6732 * reload1.c (eliminate_regs_1): Likewise.
6734 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6737 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6740 2013-09-27 Richard Biener <rguenther@suse.de>
6742 PR tree-optimization/58459
6743 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6744 restriction not propagating into loops.
6746 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6748 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6749 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6750 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6752 2013-09-26 Richard Biener <rguenther@suse.de>
6754 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6756 2013-09-26 Richard Biener <rguenther@suse.de>
6758 * alias.h (component_uses_parent_alias_set): Rename to ...
6759 (component_uses_parent_alias_set_from): ... this.
6760 * alias.c (component_uses_parent_alias_set): Rename to ...
6761 (component_uses_parent_alias_set_from): ... this and return
6763 (reference_alias_ptr_type_1): Use the result from
6764 component_uses_parent_alias_set_from instead of stripping
6765 components one at a time.
6766 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6768 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6770 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6771 Move prototypes to...
6772 * tree-ssa-ter.h: New File. Move prototypes here.
6773 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6774 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6775 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6776 from is_replaceable_p.
6777 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6778 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6779 (process_replaceable): Use ter_is_replaceable_p.
6780 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6781 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6782 newly refactored ssa_is_replaceable_p.
6783 * cfgexpand.c: Include tree-outof-ssa.h.
6784 * ssaexpand.h: Delete.
6786 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6788 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6789 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6791 (create_gimple_tmp): Delete.
6792 (get_expr_type, build_assign, build_type_cast): Move to...
6793 * gimple-builder.c: New File.
6794 (get_expr_type): Relocate from gimple.c.
6795 (build_assign, build_type_cast): Change to only create ssanames.
6796 * gimple.h: Move prototypes to...
6797 * gimple-builder.h: New File. Here.
6798 * tree-ssa.h: And here.
6799 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6800 count_uses_and_derefs): Relocate from gimple.c.
6801 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6802 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6803 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6804 gimple_replace_ssa_lhs.
6805 * asan.c: Include gimple-builder.h.
6806 * Makefile.in: Add gimple-builder.o.
6808 2013-09-26 Richard Biener <rguenther@suse.de>
6810 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6811 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6812 (loe_visit_block): Use gcc_checking_assert.
6813 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6814 gimple_assign_ssa_name_copy_p.
6815 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6818 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6820 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6821 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6823 2013-09-26 Richard Biener <rguenther@suse.de>
6825 PR tree-optimization/58539
6826 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6827 the fact that debug statements are not taking part in loop-closed
6830 2013-09-26 Nick Clifton <nickc@redhat.com>
6832 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6833 time warning message.
6834 (msp430_print_operand_raw): Delete unused letter parameter.
6835 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6836 (msp430_print_operand_address): New function.
6837 (msp430_print_operand): Move address printing code from here to
6839 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6841 (zero_extendpsisi2): Likewise.
6842 (extendpsisi2): New pattern.
6843 (andneghi3): New pattern.
6845 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6847 * config/aarch64/aarch64.opt (mlra): New option.
6848 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6849 (TARGET_LRA_P): Define.
6851 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6853 * expr.c (expand_assignment): Remove obsolete comment.
6855 2013-09-25 Jeff Law <law@redhat.com>
6857 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6858 tree-ssa-threadupdate.h.
6859 (register_jump_thread): Similarly.
6860 * tree-ssa-threadupdate.h: New header file.
6861 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6862 * tree-vrp.c: Likewise.
6863 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6864 (thread_around_empty_blocks): Change type of path vector argument to
6865 an edge,type pair from just an edge. Initialize both elements when
6866 appending to a jump threading path. Tweak references to elements
6868 (thread_across_edge): Similarly. Release memory for the elements
6870 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6871 (dump_jump_thread_path): New function broken out from
6872 register_jump_thread.
6873 (register_jump_thread): Use dump_jump_thread_path. Change type of
6874 path vector entries. Search the path for NULL edges and dump
6875 the path if one is found. Tweak the conversion of path to 3-edge
6876 form to use the block copy type information embedded in the path.
6878 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6880 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6882 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6883 Vladimir Makarov <vmakarov@redhat.com>
6885 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6886 from the least significant bit.
6887 (strip_address_mutations): Add bitfield operations handling.
6888 (must_be_index_p): Add shifting and rotate operations handling.
6889 (set_address_base): Use must_be_base_p predicate.
6890 (set_address_index): Use must_be_index_p predicate.
6892 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6893 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6894 Sergey Lega <sergey.s.lega@intel.com>
6895 Anna Tikhonova <anna.tikhonova@intel.com>
6896 Ilya Tocar <ilya.tocar@intel.com>
6897 Andrey Turetskiy <andrey.turetskiy@intel.com>
6898 Ilya Verbin <ilya.verbin@intel.com>
6899 Kirill Yukhin <kirill.yukhin@intel.com>
6900 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6902 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6904 (ix86_expand_vector_move_misalign): Support new unaligned load and
6905 stores and use new names.
6906 (CODE_FOR_sse2_storedqu): Rename to ...
6907 (CODE_FOR_sse2_storedquv16qi): ... this.
6908 (CODE_FOR_sse2_loaddqu): Rename to ...
6909 (CODE_FOR_sse2_loaddquv16qi): ... this.
6910 (CODE_FOR_avx_loaddqu256): Rename to ...
6911 (CODE_FOR_avx_loaddquv32qi): ... this.
6912 (CODE_FOR_avx_storedqu256): Rename to ...
6913 (CODE_FOR_avx_storedquv32qi): ... this.
6914 * config/i386/i386.md (fpint_logic): New.
6915 * config/i386/sse.md (VMOVE): Extend for AVX512.
6919 (VI_UNALIGNED_LOADSTORE): Ditto.
6920 (sse2_avx_avx512f): Ditto.
6921 (sse2_avx2): Extend for AVX512.
6922 (sse4_1_avx2): Ditto.
6923 (avx2_avx512f): New.
6924 (sse): Extend for AVX512.
6927 (avxsizesuffix): Ditto.
6928 (sseintvecmode): Ditto.
6930 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6931 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6932 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6933 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6934 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6935 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6936 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6937 (STORENT_MODE): Extend for AVX512.
6938 (*absneg<mode>2): Replace constraint "x" with "v".
6939 (*mul<mode>3): Ditto.
6940 (*ieee_smin<mode>3): Ditto.
6941 (*ieee_smax<mode>3): Ditto.
6942 (avx_cmp<mode>3): Replace VF with VF_128_256.
6943 (*<sse>_maskcmp<mode>3_comm): Ditto.
6944 (<sse>_maskcmp<mode>3): Ditto.
6945 (<sse>_andnot<mode>3): Extend for AVX512.
6946 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6947 (<code><mode>3, fpint_logic): New.
6948 (*<code><mode>3): Extend for AVX512.
6949 (avx512flogicsuff): New.
6950 (avx512f_<logic><mode>): Ditto.
6951 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6953 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6954 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6955 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6956 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6957 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6958 (xop_vpermil2<mode>3): Ditto.
6959 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6960 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6961 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6962 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6964 2013-09-25 Tom Tromey <tromey@redhat.com>
6966 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6967 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6968 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6969 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6970 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6971 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6972 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6973 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6974 (GRAPHITE_HTAB_H): Remove.
6976 2013-09-25 Tom Tromey <tromey@redhat.com>
6978 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6980 2013-09-25 Tom Tromey <tromey@redhat.com>
6982 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6984 2013-09-25 Tom Tromey <tromey@redhat.com>
6986 * config/i386/t-i386 (i386.o): Remove.
6987 (i386-c.o): Use COMPILE and POSTCOMPILE.
6989 2013-09-25 Tom Tromey <tromey@redhat.com>
6991 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6993 2013-09-25 Tom Tromey <tromey@redhat.com>
6995 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6996 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6997 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6998 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6999 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7000 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7001 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7002 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7003 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7004 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7005 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7006 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7007 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7008 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7009 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7010 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7011 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7012 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7013 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7014 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7015 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7016 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7017 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7018 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7019 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7020 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7021 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7022 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7023 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7024 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7025 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7026 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7027 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7028 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7029 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7030 (graphite-clast-to-gimple.o, graphite-dependences.o)
7031 (graphite-interchange.o, graphite-poly.o)
7032 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7033 (graphite-optimize-isl.o, tree-vect-loop.o)
7034 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7035 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7036 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7037 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7038 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7039 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7040 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7041 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7042 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7043 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7044 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7045 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7046 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7047 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7048 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7049 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7050 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7051 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7052 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7053 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7054 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7055 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7056 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7057 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7058 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7059 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7060 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7061 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7062 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7063 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7064 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7065 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7066 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7067 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7068 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7069 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7070 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7071 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7072 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7073 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7074 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7075 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7076 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7077 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7078 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7079 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7080 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7081 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7082 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7083 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7084 (gcov-dump.o): Remove.
7085 (default-c.o): Use COMPILE and POSTCOMPILE.
7086 (CFLAGS-gcc.o): New variable.
7087 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7089 2013-09-25 Tom Tromey <tromey@redhat.com>
7091 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7092 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7093 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7094 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7095 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7096 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7097 (c-family/array-notation-common.o, c-family/stub-objc.o)
7098 (c-family/c-ubsan.o): Remove.
7100 2013-09-25 Tom Tromey <tromey@redhat.com>
7102 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7104 2013-09-25 Tom Tromey <tromey@redhat.com>
7106 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7107 to add -DENABLE_SHARED_LIBGCC.
7108 (gcc.o): Don't use subshell.
7110 2013-09-25 Tom Tromey <tromey@redhat.com>
7112 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7113 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7114 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7115 * configure, config.in: Rebuild.
7117 2013-09-25 Tom Tromey <tromey@redhat.com>
7119 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7120 (COMPILE, POSTCOMPILE): New variables.
7121 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7122 (DEPFILES): New variable.
7123 Include ".Po" files.
7124 * configure.ac: Add checks for dependency checking.
7125 * configure, aclocal.m4: Regenerate.
7127 2013-09-25 Tom Tromey <tromey@redhat.com>
7129 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7130 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7132 2013-09-25 Tom Tromey <tromey@redhat.com>
7134 * Makefile.in (generated_files): Add options.h,
7135 target-hooks-def.h, insn-opinit.h,
7136 common/common-target-hooks-def.h, pass-instances.def,
7137 c-family/c-target-hooks-def.h.
7139 2013-09-25 Jeff Law <law@redhat.com>
7141 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7142 than foo[foo.length () - 1] to access last member in a vec.
7143 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7145 2013-09-25 Richard Biener <rguenther@suse.de>
7148 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7150 2013-09-25 Jan Hubicka <jh@suse.cz>
7152 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7155 2013-09-25 Marek Polacek <polacek@redhat.com>
7158 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7159 when determining the type name.
7161 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7163 * config/sh/sh.md: Fix formatting.
7165 2013-09-24 Xinliang David Li <davidxl@google.com>
7167 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7168 max peel iterations parameter.
7169 * param.def: New parameter.
7170 * doc/invoke.texi: Document New parameter.
7172 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7174 * gimple-pretty-print.c: Various whitespace tweaks.
7175 * tree-core.h: Likewise.
7176 * tree-pretty-print.c: Likewise.
7177 * tree-ssa-alias.c: Likewise.
7178 * tree-ssa-copy.c: Likewise.
7179 * tree-ssanames.c: Likewise.
7180 * tree-ssanames.h: Likewise.
7181 * tree-vrp.c: Likewise.
7183 2013-09-24 Alan Modra <amodra@gmail.com>
7187 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7188 for output operands that disallow regs. Don't use EXPAND_WRITE on
7191 2013-09-24 Richard Biener <rguenther@suse.de>
7194 * tree.c (reference_alias_ptr_type): Move ...
7195 * alias.c (reference_alias_ptr_type): ... here and implement
7196 in terms of the new reference_alias_ptr_type_1.
7197 (ref_all_alias_ptr_type_p): New helper.
7198 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7199 use ref_all_alias_ptr_type_p.
7200 (get_deref_alias_set): Add flag_strict_aliasing check here.
7201 (reference_alias_ptr_type_1): New function, split out from ...
7202 (get_alias_set): ... here.
7203 (alias_ptr_types_compatible_p): New function.
7204 * alias.h (reference_alias_ptr_type): Declare.
7205 (alias_ptr_types_compatible_p): Likewise.
7206 * tree.h (reference_alias_ptr_type): Remove.
7207 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7208 to compare MEM_REF alias types.
7210 2013-09-24 Richard Biener <rguenther@suse.de>
7212 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7214 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7216 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7217 reload helper function arrays into a single array reg_addr.
7218 (reload_fpr_gpr): Likewise.
7219 (reload_gpr_vsx): Likewise.
7220 (reload_vsx_gpr): Likewise.
7221 (struct rs6000_reg_addr): Likewise.
7222 (reg_addr): Likewise.
7223 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7224 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7225 (rs6000_init_hard_regno_mode_ok): Likewise.
7226 (rs6000_secondary_reload_direct_move): Likewise.
7227 (rs6000_secondary_reload): Likewise.
7229 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7230 constraints: wu, ww, and wy. Repurpose wv constraint added during
7231 power8 changes. Put wg constraint in alphabetical order.
7233 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7234 for future work to add ISA 2.07 VSX single precision support.
7235 (-mvsx-scalar-double): Change default from -1 to 1, update
7236 documentation comment.
7237 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7238 (-mupper-regs-df): New debug switch to control whether DF values
7239 can go in the traditional Altivec registers.
7240 (-mupper-regs-sf): New debug switch to control whether SF values
7241 can go in the traditional Altivec registers.
7243 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7245 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7246 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7247 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7248 (rs6000_debug_legitimate_address_p): Print if we are running
7249 before, during, or after reload.
7250 (rs6000_secondary_reload): Add a comment.
7251 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7253 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7254 constraints. Update documentation string.
7255 (wd constraint): Likewise.
7256 (wf constraint): Likewise.
7257 (wg constraint): Likewise.
7258 (wn constraint): Likewise.
7259 (ws constraint): Likewise.
7260 (wt constraint): Likewise.
7261 (wx constraint): Likewise.
7262 (wz constraint): Likewise.
7263 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7265 (ww constraint): Likewise.
7266 (wy constraint): Likewise.
7267 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7268 in the previous submissions.
7269 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7271 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7273 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7274 (REG_BR_PROB): Say that the probability is stored in an int_list.
7275 * reg-notes.def: Update commentary to mention INT_LIST.
7276 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7277 (INT_LIST): New rtx.
7278 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7279 * rtlanal.c (int_reg_note_p): New function.
7280 (alloc_reg_note): Assert that the note does not have an int argument.
7281 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7282 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7283 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7284 rather than an INSN_LIST. Handle INT_LIST.
7285 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7286 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7287 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7288 Manipulate them as ints rather than rtxes.
7289 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
7290 * regmove.c (copy_src_to_dest): Likewise.
7291 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
7293 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
7294 into the cases that need it.
7295 * config/arm/arm.c (arm_unwind_emit): Likewise.
7297 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
7298 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
7299 * loop-doloop.c (add_test, doloop_modify): Likewise.
7300 * loop-unswitch.c (compare_and_jump_seq): Likewise.
7301 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
7302 * predict.c (combine_predictions_for_insn): Likewise.
7303 * print-rtl.c (print_rtx): Handle INT_LIST.
7304 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
7305 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
7306 * config/arm/arm.c (emit_unlikely_jump): Likewise.
7307 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
7308 (ix86_split_fp_branch, predict_jump): Likewise.
7309 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
7310 * config/sh/sh.c (expand_cbranchsi4): Likewise.
7311 * config/spu/spu.c (ea_load_store_inline): Likewise.
7313 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
7314 value of a REG_BR_PROB note.
7315 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
7316 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
7317 * emit-rtl.c (try_split): Likewise.
7318 * predict.c (br_prob_note_reliable_p): Likewise.
7319 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
7320 * reorg.c (mostly_true_jump): Likewise.
7321 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
7322 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
7323 * config/i386/i386.c (ix86_print_operand): Likewise.
7324 * config/ia64/ia64.c (ia64_print_operand): Likewise.
7325 * config/mmix/mmix.c (mmix_print_operand): Likewise.
7326 * config/rs6000/rs6000.c (output_cbranch): Likewise.
7327 * config/s390/s390.c (s390_expand_tbegin): Likewise.
7328 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
7329 * config/sparc/sparc.c (output_cbranch): Likewise.
7330 * config/spu/spu.c (get_branch_target): Likewise.
7331 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
7332 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
7334 2013-09-23 Jan Hubicka <jh@suse.cz>
7336 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
7338 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
7339 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
7341 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
7342 (gimple_fold_call): Dump inconsistent devirtualizations; add
7343 sanity check for type based devirtualizations.
7344 * ipa-prop.c: Include ipa-utils.h
7345 (ipa_intraprocedural_devirtualization): Add sanity check.
7346 (try_make_edge_direct_virtual_call): Likewise.
7348 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
7350 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
7351 assignment statements.
7353 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
7355 * gimple-pretty-print.c (dump_ssaname_info): New function.
7356 (dump_gimple_phi): Call it.
7357 (pp_gimple_stmt_1): Likewise.
7358 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7359 (range_info_def): Declare.
7360 * tree-pretty-print.c (pp_double_int): New function.
7361 (dump_generic_node): Call it.
7362 * tree-pretty-print.h (pp_double_int): Declare.
7363 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7364 * tree-ssanames.h (range_info_def): New structure.
7365 (value_range_type): Move definition here.
7366 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7368 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7370 (set_range_info): New function.
7371 (get_range_info): Likewise.
7372 (duplicate_ssa_name_range_info): Likewise.
7373 (duplicate_ssa_name_fn): Check pointer type and call
7374 duplicate_ssa_name_range_info.
7375 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7376 * tree-vrp.c (value_range_type): Remove definition, now in
7378 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7379 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7380 (SSA_NAME_RANGE_INFO): New macro.
7382 2013-09-23 Richard Biener <rguenther@suse.de>
7384 PR tree-optimization/58464
7385 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7386 (phi_trans_add): Change to add conditionally on being not present.
7387 (phi_translate_1): Remove recursion detection here.
7388 (phi_translate): Pre-seed the cache with NULL to catch
7389 recursion here in a more generic way.
7390 (bitmap_find_leader): Adjust comment.
7391 (get_representative_for): Dump value-numbers.
7392 (create_expression_by_pieces): Likewise.
7393 (insert_into_preds_of_block): Likewise.
7395 2013-09-23 Christian Bruel <christian.bruel@st.com>
7398 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7399 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7401 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7404 2013-09-20 Renlin Li <renlin.li@arm.com>
7406 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7408 (aarch64_expand_epilogue): Likewise.
7409 (aarch64_legitimize_reload_address): Likewise.
7411 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7413 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7414 NULL_TREE before pushing them onto the vector. Likewise for labels.
7416 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7418 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7419 comparison operators when -fno-trapping-math is in effect.
7420 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7421 comparison operators in TFmode and assert that unsupported operators
7423 (ia64_print_operand): Likewise.
7425 2013-09-21 Jan Hubicka <jh@suse.cz>
7427 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7428 (sse_typeless_stores): Enable for core
7429 (sse_load0_by_pxor): Likewise.
7430 (four_jump_limit): Disable for core.
7431 (pad_returns): Likewise.
7432 (avoid_vector_decode): Likewise.
7433 (fuse_cmp_and_branch): Enable for cores.
7434 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7436 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7439 * config/pa/pa.c (pa_option_override): Disable auto increment and
7440 decrement instructions until reload is completed.
7442 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7443 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7444 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7446 2013-09-20 DJ Delorie <dj@redhat.com>
7447 Nick Clifton <nickc@redhat.com>
7449 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7450 (need_to_save): Save bank 0 on interrupts.
7451 (characterize_address): Strip far address wrappers.
7452 (rl78_as_legitimate_address): Likewise.
7453 (transcode_memory_rtx): Likewise.
7454 (rl78_peep_movhi_p): Disable this peephole after devirt.
7455 (rl78_propogate_register_origins): Forget all origins when a
7457 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7458 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7460 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7461 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7462 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7464 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7466 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7468 * config/pa/pa.md: In "scc" insn patterns, change output template to
7469 handle const0_rtx in reg_or_0_operand operands.
7471 2013-09-20 Martin Husemann <martin@NetBSD.org>
7474 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7475 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7477 2013-09-20 Richard Biener <rguenther@suse.de>
7480 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7481 remembering SSA name version and block index.
7482 (new_scev_info_str): Adjust.
7483 (hash_scev_info): Likewise. Also hash the block index.
7484 (eq_scev_info): Adjust.
7485 (find_var_scev_info): Likewise.
7486 (struct instantiate_cache_entry): Remove.
7487 (struct instantiate_cache_type): Use a htab to map name, block
7489 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7490 (get_instantiated_value_entry): Likewise.
7491 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7492 (instantiate_scev_name): Adjust.
7494 2013-09-20 Jeff Law <law@redhat.com>
7496 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7498 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7500 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7501 Call aarch64_simd_expand_args to update op[argc].
7503 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7505 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7508 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7510 * gengtype.c (file_rules): Added rule for *.cc files.
7511 (get_output_file_with_visibility): Give fatal message when no
7514 2013-09-20 Renlin Li <renlin.li@arm.com>
7516 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7517 (aarch64_expand_epilogue): Likewise.
7518 (aarch64_legitimize_reload_address): Likewise.
7520 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7523 * expr.c (expand_assignment): Remove misalignp code path.
7525 2013-09-20 Marek Polacek <polacek@redhat.com>
7528 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7529 TYPE_PRECISION. Add asserts.
7531 2013-09-20 Richard Biener <rguenther@suse.de>
7533 PR tree-optimization/58453
7534 * tree-loop-distribution.c (distribute_loop): Apply the cost
7535 model for -ftree-loop-distribute-patterns, too.
7537 2013-09-20 Richard Biener <rguenther@suse.de>
7540 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7541 make type comparison less strict.
7543 2013-09-20 Alan Modra <amodra@gmail.com>
7545 * configure: Regenerate.
7546 * aclocal.m4: Regenerate.
7548 2013-09-20 Marek Polacek <polacek@redhat.com>
7551 * doc/extend.texi: Document that attribute used is meant to be used
7552 on variables with static storage duration.
7554 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7556 PR tree-optimization/58472
7557 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7558 simd_lane_access set inv_p = false.
7559 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7560 the simduid magic VAR_DECL.
7562 2013-09-19 Jan Hubicka <jh@suse.cz>
7564 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7566 2013-09-17 Jeff Law <law@redhat.com>
7568 * tree-ssa-dom.c (record_temporary_equivalences): New function
7569 split out of dom_opt_dom_walker::after_dom_children.
7570 (dom_opt_dom_walker::thread_across_edge): Move common code
7571 in here from dom_opt_dom_walker::after_dom_children.
7572 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7574 2013-09-19 Jan Hubicka <jh@suse.cz>
7576 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7577 (TARGET_GENERIC): Use PROCESOR_GENERIC
7578 (enum processor_type): Unify generic32 and 64.
7579 * i386.md (cpu): Likewise.
7580 * x86-tune.def (use_leave): Enable for generic32.
7581 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7582 * athlon.md: Change generic64 to generic in all occurences.
7583 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7584 (ix86_target_macros_internal): Likewise.
7585 * driver-i386.c (host_detect_local_cpu): Likewise.
7586 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7588 (generic_memcpy, generic_memset, generic_cost): This one.
7589 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7590 (m_GENERIC32, m_GENERIC64): Remove.
7591 (m_GENERIC): Turn into one flag.
7592 (processor_target): Unify generic tunnings.
7593 (ix86_option_override_internal): Replace generic32/64 by generic.
7594 (ix86_issue_rate): Likewise.
7595 (ix86_adjust_cost): Likewise.
7597 2013-09-19 Jan Hubicka <jh@suse.cz>
7599 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7600 of speculative flag.
7602 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7604 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7605 GOMP_sections_start, even if !exit_reachable.
7607 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7609 * lra-constraints.c (need_for_all_save_p): Use macro
7610 HARD_REGNO_CALL_PART_CLOBBERED.
7611 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7612 set up pseudo conflict hard regs.
7614 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7617 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7620 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7622 PR rtl-optimization/58438
7623 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7624 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7625 from previous subpasses.
7627 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7629 * arm.c (arm_get_frame_offsets): Validate architecture supports
7630 LDRD/STRD before accepting the tuning preference.
7631 (arm_expand_prologue): Likewise.
7632 (arm_expand_epilogue): Likewise.
7634 2013-09-18 Richard Biener <rguenther@suse.de>
7636 PR tree-optimization/58417
7637 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7638 have chrecs with symbols defined in the loop as operands.
7639 (chrec_fold_multiply): Likewise.
7640 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7641 parameters before folding binary operations.
7642 (struct instantiate_cache_entry_hasher): Remove.
7643 (struct instantiate_cache_type): Use a pointer-map.
7644 (instantiate_cache_type::instantiate_cache_type): New function.
7645 (instantiate_cache_type::get): Likewise.
7646 (instantiate_cache_type::set): Likewise.
7647 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7648 (get_instantiated_value_entry): Likewise.
7649 (global_cache): New global.
7650 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7651 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7652 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7653 (instantiate_scev_name): Adjust.
7654 (instantiate_scev): Construct global instead of local cache.
7655 (resolve_mixers): Likewise.
7657 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7658 Paolo Carlini <paolo.carlini@oracle.com>
7661 * doc/implement-cxx.texi: Fix references to the C++ standards.
7663 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7665 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7666 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7667 simd_lane_access drs, update also DR_ALIGNED_TO.
7669 2013-09-18 Marek Polacek <polacek@redhat.com>
7672 * doc/extend.texi: Document no_sanitize_undefined attribute.
7673 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7674 no_sanitize_undefined attribute.
7676 2013-09-18 Nick Clifton <nickc@redhat.com>
7678 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7679 (ASM_DECLARE_FUNCTION_NAME): Define.
7681 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7683 * compare-elim.c (find_comparison_dom_walker): New class
7684 (find_comparisons_in_bb): Rename to
7685 find_comparison_dom_walker::before_dom_children
7686 (find_comparisons): Adjust
7687 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7689 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7690 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7691 (init_walk_dominator_tree): Remove declaration.
7692 (fini_walk_dominator_tree): Remove declaration.
7693 * fwprop.c (single_def_use_dom_walker): New class
7694 (single_def_use_enter_block): Convert to
7695 single_def_use_dom_walker::before_dom_children.
7696 (single_def_use_leave_block): Convert to
7697 single_def_use_dom_walker::after_dom_children.
7698 (build_single_def_use_links): Adjust.
7699 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7701 (find_candidates_in_block): Convert to
7702 find_candidates_dom_walker::before_dom_children.
7703 (execute_strength_reduction): Adjust.
7704 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7705 (sese_dom_walker): New class.
7706 (sese_dom_walker::sese_dom_walker): New constructor.
7707 (sese_dom_walker::~sese_dom_walker): New destructor.
7708 (build_sese_conditions_before): Convert to
7709 sese_dom_walker::before_dom_children.
7710 (build_sese_conditions_after): Convert to
7711 sese_dom_walker::after_dom_children.
7712 (build_poly_scop): Adjust
7713 * tree-into-ssa.c (rewrite_dom_walker): New class
7714 (rewrite_enter_block): Convert to
7715 rewrite_dom_walker::before_dom_children.
7716 (rewrite_leave_block): Convert to
7717 rewrite_dom_walker::after_dom_children.
7718 (rewrite_update_dom_walker): New class.
7719 (rewrite_update_enter_block): Convert to
7720 rewrite_update_dom_walker::before_dom_children.
7721 (rewrite_update_leave_block): Convert to
7722 rewrite_update_dom_walker::after_dom_children.
7723 (rewrite_blocks, rewrite_into_ssa): Adjust.
7724 (mark_def_dom_walker): New class.
7725 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7726 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7727 (mark_def_sites_blocks): Convert to
7728 mark_def_dom_walker::before_dom_children.
7729 (mark_def_site_blocks): Remove.
7730 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7731 (tree_ssa_dominator_optimize): Adjust.
7732 (dom_thread_across_edge): Convert to method
7733 dom_opt_dom_walker::thread_across_edge.
7734 (dom_opt_enter_block): Convert to member function
7735 dom_opt_dom_walker::before_dom_children.
7736 (dom_opt_leave_block): Convert to member function
7737 dom_opt_dom_walker::after_dom_children.
7738 * tree-ssa-dse.c (dse_dom_walker): New class.
7739 (dse_enter_block): Convert to member function
7740 dse_dom_walker::before_dom_children.
7741 (tree_ssa_dse): Adjust.
7742 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7743 (determine_invariantness_stmt): Convert to method
7744 invariantness_dom_walker::before_dom_children.
7745 (determine_invariantness): Remove
7746 (move_computations_dom_walker): New class.
7747 (move_computations_stmt): Convert to method
7748 move_computations_dom_walker::before_dom_children.
7749 (move_computations, tree_ssa_lim): Adjust.
7750 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7751 (nt_init_block): Convert to method
7752 notrappping_dom_walker::before_dom_children.
7753 (nt_fini_block): Convert to method
7754 method nontrapping_dom_walker::after_dom_children.
7755 (get_non_trapping): Adjust.
7756 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7757 (eliminate_bb): Convert to method
7758 eliminate_dom_walker::before_dom_children.
7759 (eliminate_leave_block): Convert to method
7760 eliminate_dom_walker::after_dom_children.
7761 (eliminate): Adjust.
7762 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7763 (strlen_enter_block): Convert to method
7764 strlen_dom_walker::before_dom_children.
7765 (strlen_leave_block): Convert to method
7766 method strlen_dom_walker::after_dom_children.
7767 (tree_ssa_strlen): Adjust.
7768 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7769 (tree_ssa_uncprop): Adjust.
7770 (uncprop_leave_block): Convert to method
7771 uncprop_dom_walker::after_dom_children.
7772 (uncprop_leave_block): Convert to method
7773 uncprop_dom_walker::before_dom_children.
7775 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7777 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7778 branch in same basic block. Check both src and dest of the move insn.
7780 2013-09-17 Nick Clifton <nickc@redhat.com>
7782 * config/rl78/rl78-real.md (bf): New pattern.
7784 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7785 (rl78_print_operand): Do not put a # before a %B.
7786 * config/rl78/rl78.opt: Tweak doc strings.
7788 2013-09-17 DJ Delorie <dj@redhat.com>
7790 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7791 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7792 (rl78_option_override): Likewise, if -mallregs.
7793 (is_virtual_register): Likewise.
7794 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7795 (REGNO_OK_FOR_BASE_P): Likewise.
7796 * config/rl78/rl78.opt (-mallregs): New.
7798 2013-09-17 Nick Clifton <nickc@redhat.com>
7800 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7801 For interrupt functions: save all call clobbered registers if the
7802 interrupt handler is not a leaf function.
7803 (rl78_expand_prologue): Always recompute the frame information.
7804 For interrupt functions: only select bank 0 if one of the bank 0
7805 registers is going to be psuhed.
7807 2013-09-17 DJ Delorie <dj@redhat.com>
7809 * config/rl78/constraints.md: For each W* constraint, rename to C*
7810 and create a W* constraint that checks for an optional ES: prefix
7812 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7813 (es_addr): New. Used to wrap far addresses.
7814 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7815 (rl78_es_base): New.
7816 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7817 wrapped far addresses.
7818 (rl78_print_operand_1): Unwrap far addresses before processing.
7819 (rl78_lo16): Wrap far addresses in unspecs.
7820 (rl78_es_addr): New.
7821 (rl78_es_base): New.
7822 (insn_ok_now): Check for not-yet-wrapped far addresses.
7823 (transcode_memory_rtx): Properly re-wrap far addresses.
7825 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7827 * config/sparc/t-rtems: Add leon3 multilibs.
7829 2013-09-17 Cong Hou <congh@google.com>
7831 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7832 when checking the dot production pattern. The type of rhs operand
7833 of multiply is now checked correctly.
7835 2013-09-17 Jeff Law <law@redhat.com>
7837 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7838 edge implied equivalences into successor phis.
7839 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7840 here from tree-ssa-threadedge.c.
7841 (mark_threaded_blocks): When threading through a joiner, if both
7842 successors of the joiner's clone reach the same block, verify the
7843 PHI arguments are equal. If not, cancel the jump threading request.
7844 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7845 tree-ssa-threadupdate.c
7846 (thread_across_edge): Don't check PHI argument equality when
7847 threading through joiner block here.
7849 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7851 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7852 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7853 (warn_uninit, warn_uninitialized_vars,
7854 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7855 Move to tree-ssa-uninit.c.
7856 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7857 (has_undefined_value_p): New. Pass dependant parts of
7858 ssa_undefined_value_p.
7859 (uninit_undefined_value_p): Use has_undefined_value_p.
7860 (warn_uninit, warn_uninitialized_vars,
7861 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7862 Move from tree-ssa.c.
7863 * tree-ssa.h: Adjust prototypes.
7865 2013-09-17 Jan Hubicka <jh@suse.cz>
7868 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7869 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7870 FUNCTION_NOT_OPTIMIZED.
7871 * ipa-inline-analysis.c (compute_inline_parameters): Function
7872 not optimized is not inlinable unless it is alwaysinline.
7873 (inline_analyze_function): Force calls in not optimized
7874 function not inlinable.
7876 2013-09-17 Jan Hubicka <jh@suse.cz>
7879 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7881 * ipa.c (function_and_variable_visibility): Likewise.
7882 * ipa-profile.c (ipa_profile): Likewise.
7884 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7887 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7888 attribute, and don't inline the resolver in this case.
7890 2013-09-17 Teresa Johnson <tejohnson@google.com>
7892 * coverage.c (get_coverage_counts): Add missing newline.
7894 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7896 PR tree-optimization/58088
7897 * fold-const.c (mask_with_trailing_zeros): New function.
7898 (fold_binary_loc): Make sure we don't recurse infinitely
7899 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7900 Use mask_with_trailing_zeros where appropriate.
7902 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7904 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7905 of 'prev' var to get better distance estimation.
7907 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7909 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7910 * tree-inline.c (is_parameter_of): New predicate.
7911 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7912 a parameter has been remapped.
7913 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7914 (optimize_inline_calls): Initialize transform_parameter.
7915 (copy_gimple_seq_and_replace_locals): Likewise.
7916 (tree_function_versioning): Likewise.
7917 (maybe_inline_call_in_expr): Likewise.
7919 2013-09-17 Nick Clifton <nickc@redhat.com>
7921 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7922 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7924 (is_attr_func): New function.
7925 (msp430_is_interrupt_func): New function.
7926 (is_naked_func): New function.
7927 (is_reentrant_func): New function.
7928 (is_critical_func): New function.
7929 (msp430_start_function): Add annotations for function attributes.
7930 (msp430_attr): New function.
7931 (msp430_attribute_table): New.
7932 (msp430_function_section): New function.
7933 (TARGET_ASM_FUNCTION_SECTION): Define.
7934 (msp430_builtin): New enum.
7935 (msp430_init_builtins): New function.
7936 (msp430_builtin_devl): New function.
7937 (msp430_expand_builtin): New function.
7938 (TARGET_INIT_BUILTINS): Define.
7939 (TARGET_EXPAND_BUILTINS): Define.
7940 (TARGET_BUILTIN_DECL): Define.
7941 (msp430_expand_prologue): Add support for naked, interrupt,
7942 critical and reentrant functions.
7943 (msp430_expand_epilogue): Likewise.
7944 (msp430_print_operand): Handle 'O' character.
7945 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7947 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7948 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7949 (pushm): Use a 'n' rather than an 'i' constraint.
7950 (msp_return): Add generation of the interrupt return instruction.
7951 (disable_interrupts): New pattern.
7952 (enable_interrupts): New pattern.
7953 (push_intr_state): New pattern.
7954 (pop_intr_state): New pattern.
7955 (bic_SR): New pattern.
7956 (bis_SR): New pattern.
7957 * doc/extend.texi: Document MSP430 function attributes and builtin
7960 2013-09-17 Richard Biener <rguenther@suse.de>
7962 PR tree-optimization/58432
7963 * tree-loop-distribution.c (tree_loop_distribution): Also
7964 scan PHIs for outside loop uses and seed a partition from them.
7966 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7968 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7969 (restructure_reference): Call backtrace_base_for_ref.
7971 2013-09-17 Alan Modra <amodra@gmail.com>
7974 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7977 2013-09-16 DJ Delorie <dj@redhat.com>
7979 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7980 vregs location for RL78/G10.
7981 (rl78_expand_prologue): Avoid SEL on G10.
7982 (rl78_expand_epilogue): Likewise.
7983 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7984 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7985 __RL78_G10__ when appropriate.
7986 (ASM_SPEC): Pass -mg10 along to the assembler.
7987 * config/rl78/rl78.md (sel_rb): Disable for G10.
7988 * config/rl78/rl78.opt: Add -mg10 option.
7989 * config/rl78/t-rl78: Add -mg10 multilib.
7991 2013-09-16 Xinliang David Li <davidxl@google.com>
7993 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7994 * omp-low.c (omp_max_vf): Ditto.
7995 (expand_omp_simd): Ditto.
7996 * tree-vectorizer.c (vectorize_loops): Ditto.
7997 (gate_vect_slp): Ditto.
7998 (gate_increase_alignment): Ditto.
7999 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8000 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8001 (gate_tree_vectorize): Name change.
8002 (tree_vectorize): Ditto.
8003 (pass_vectorize::gate): Call new function.
8004 (pass_vectorize::execute): Ditto.
8005 * opts.c: O3 default setting change.
8006 (finish_options): Check new flag.
8007 * doc/invoke.texi: Document new flags.
8008 * common.opt: New flags.
8010 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8012 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8013 (Condition Code): Don't reference it.
8014 * doc/tm.texi: Regenerate.
8016 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8019 * lra-constraints.c (undo_optional_reloads): Consider all optional
8020 reload even if it did not get a hard reg.
8022 2013-09-16 Teresa Johnson <tejohnson@google.com>
8024 * dumpfile.c (dump_loc): Remove newline emission.
8025 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8026 emission to dump_printf_loc calls where missing.
8027 (vect_mark_for_runtime_alias_test): Ditto.
8028 (vect_analyze_data_ref_dependence): Ditto.
8029 (vect_analyze_data_ref_dependences): Ditto.
8030 (vect_slp_analyze_data_ref_dependence): Ditto.
8031 (vect_slp_analyze_data_ref_dependences): Ditto.
8032 (vect_compute_data_ref_alignment): Ditto.
8033 (vect_update_misalignment_for_peel): Ditto.
8034 (vect_verify_datarefs_alignment): Ditto.
8035 (vector_alignment_reachable_p): Ditto.
8036 (vect_get_data_access_cost): Ditto.
8037 (vect_enhance_data_refs_alignment): Ditto.
8038 (vect_find_same_alignment_drs): Ditto.
8039 (vect_analyze_data_refs_alignment): Ditto.
8040 (vect_analyze_group_access): Ditto.
8041 (vect_analyze_data_ref_access): Ditto.
8042 (vect_analyze_data_ref_accesses): Ditto.
8043 (vect_prune_runtime_alias_test_list): Ditto.
8044 (vect_analyze_data_refs): Ditto.
8045 (vect_create_addr_base_for_vector_ref): Ditto.
8046 (vect_create_data_ref_ptr): Ditto.
8047 (vect_grouped_store_supported): Ditto.
8048 (vect_grouped_load_supported): Ditto.
8049 * value-prof.c (check_counter): Ditto.
8050 (check_ic_target): Ditto.
8051 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8052 (vect_recog_widen_mult_pattern): Ditto.
8053 (vect_recog_widen_sum_pattern): Ditto.
8054 (vect_recog_over_widening_pattern): Ditto.
8055 (vect_recog_widen_shift_pattern): Ditto.
8056 (vect_recog_rotate_pattern): Ditto.
8057 (vect_recog_vector_vector_shift_pattern): Ditto.
8058 (vect_recog_divmod_pattern): Ditto.
8059 (vect_recog_mixed_size_cond_pattern): Ditto.
8060 (vect_recog_bool_pattern): Ditto.
8061 (vect_pattern_recog_1): Ditto.
8062 (vect_pattern_recog): Ditto.
8063 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8064 (vect_is_simple_iv_evolution): Ditto.
8065 (vect_analyze_scalar_cycles_1): Ditto.
8066 (vect_get_loop_niters): Ditto.
8067 (vect_analyze_loop_1): Ditto.
8068 (vect_analyze_loop_form): Ditto.
8069 (vect_analyze_loop_operations): Ditto.
8070 (vect_analyze_loop_2): Ditto.
8071 (vect_analyze_loop): Ditto.
8072 (report_vect_op): Ditto.
8073 (vect_is_slp_reduction): Ditto.
8074 (vect_is_simple_reduction_1): Ditto.
8075 (vect_get_known_peeling_cost): Ditto.
8076 (vect_estimate_min_profitable_iters): Ditto.
8077 (vect_model_reduction_cost): Ditto.
8078 (vect_model_induction_cost): Ditto.
8079 (get_initial_def_for_induction): Ditto.
8080 (vect_create_epilog_for_reduction): Ditto.
8081 (vectorizable_reduction): Ditto.
8082 (vectorizable_induction): Ditto.
8083 (vectorizable_live_operation): Ditto.
8084 (vect_loop_kill_debug_uses): Ditto.
8085 (vect_transform_loop): Ditto.
8086 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8087 (vect_stmt_relevant_p): Ditto.
8088 (process_use): Ditto.
8089 (vect_mark_stmts_to_be_vectorized): Ditto.
8090 (vect_model_simple_cost): Ditto.
8091 (vect_model_promotion_demotion_cost): Ditto.
8092 (vect_model_store_cost): Ditto.
8093 (vect_get_store_cost): Ditto.
8094 (vect_model_load_cost): Ditto.
8095 (vect_get_load_cost): Ditto.
8096 (vect_init_vector_1): Ditto.
8097 (vect_get_vec_def_for_operand): Ditto.
8098 (vect_finish_stmt_generation): Ditto.
8099 (vectorizable_call): Ditto.
8100 (vectorizable_conversion): Ditto.
8101 (vectorizable_assignment): Ditto.
8102 (vectorizable_shift): Ditto.
8103 (vectorizable_operation): Ditto.
8104 (vectorizable_store): Ditto.
8105 (vectorizable_load): Ditto.
8106 (vectorizable_condition): Ditto.
8107 (vect_analyze_stmt): Ditto.
8108 (vect_transform_stmt): Ditto.
8109 (vect_is_simple_use): Ditto.
8110 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8111 (vect_can_advance_ivs_p): Ditto.
8112 (vect_update_ivs_after_vectorizer): Ditto.
8113 (vect_do_peeling_for_loop_bound): Ditto.
8114 (vect_gen_niters_for_prolog_loop): Ditto.
8115 (vect_update_inits_of_drs): Ditto.
8116 (vect_create_cond_for_alias_checks): Ditto.
8117 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8118 (vect_build_slp_tree_1): Ditto.
8119 (vect_supported_load_permutation_p): Ditto.
8120 (vect_analyze_slp_instance): Ditto.
8121 (vect_analyze_slp): Ditto.
8122 (vect_make_slp_decision): Ditto.
8123 (vect_detect_hybrid_slp): Ditto.
8124 (vect_bb_vectorization_profitable_p): Ditto.
8125 (vect_slp_analyze_bb_1): Ditto.
8126 (vect_update_slp_costs_according_to_vf): Ditto.
8127 (vect_get_mask_element): Ditto.
8128 (vect_transform_slp_perm_load): Ditto.
8129 (vect_schedule_slp_instance): Ditto.
8130 (vect_schedule_slp): Ditto.
8131 (vect_slp_transform_bb): Ditto.
8132 * profile.c (read_profile_edge_counts): Ditto.
8133 (compute_branch_probabilities): Ditto.
8134 * coverage.c (get_coverage_counts): Ditto.
8136 2013-09-16 Diego Novillo <dnovillo@google.com>
8138 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8140 2013-09-16 Jan Hubicka <jh@suse.cz>
8142 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8143 abstract functions; for static functions check the presence of body.
8145 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8147 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8148 * config/aarch64/aarch64-simd.md
8149 (aarch64_mla_elt<mode>): New.
8150 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8151 (aarch64_mls_elt<mode>): Likewise.
8152 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8153 (aarch64_fma4_elt<mode>): Likewise.
8154 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8155 (aarch64_fma4_elt_to_128v2df): Likewise.
8156 (aarch64_fma4_elt_to_64df): Likewise.
8157 (fnma<mode>4): Likewise.
8158 (aarch64_fnma4_elt<mode>): Likewise.
8159 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8160 (aarch64_fnma4_elt_to_128v2df): Likewise.
8161 (aarch64_fnma4_elt_to_64df): Likewise.
8162 * config/aarch64/iterators.md (VDQSF): New.
8163 * config/aarch64/arm_neon.h
8164 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8165 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8167 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8169 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8170 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8171 (aarch64_mul3_elt_to_128df): Likewise.
8172 (aarch64_mul3_elt_to_64v2df): Likewise.
8173 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8175 (VMUL_CHANGE_NLANES) Likewise.
8178 * config/aarch64/arm_neon.h
8179 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8181 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8183 * config/aarch64/arm_neon.h
8184 (vcvtx_high_f32_f64): Fix parameters.
8186 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8187 Uros Bizjak <ubizjak@gmail.com>
8189 * config/alpha.c: Include tree-ssanames.h.
8191 2013-09-16 Richard Biener <rguenther@suse.de>
8193 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8194 (dot_rdg_1): Handle control_dd.
8195 (create_edge_for_control_dependence): New function.
8196 (create_rdg_edges): Add control dependences if asked for.
8197 (build_rdg): Likewise.
8198 (generate_loops_for_partition): If there are not necessary
8199 control stmts remove all their dependencies.
8200 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8201 (distribute_loop): Pass on control dependences.
8202 (tree_loop_distribution): Compute control dependences and remove
8203 restriction on number of loop nodes.
8205 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8207 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8210 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8212 * cse.c (try_const_anchors): Punt on CC modes.
8214 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8216 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8218 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8221 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8224 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8227 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8229 (darwin_objc1_section): Likewise.
8230 (darwin_file_end): Emit Image Info section when required.
8232 2013-09-14 Jan Hubicka <jh@suse.cz>
8234 * tree-into-ssa.c (gate_into_ssa): New.
8235 (pass_data_build_ssa): Use it.
8236 * cgraph.h (expand_thunk): Update prototype.
8237 * cgraphunit.c (analyze_function): Expand thunks early.
8238 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8239 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8240 set lowered flag; do not add new function.
8241 (assemble_thunks_and_aliases): Update.
8242 * tree-ssa.c (gate_init_datastructures): New gate.
8243 (pass_data_init_datastructures): Use it.
8245 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8248 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8249 xmm register set described in the psABI.
8251 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8253 * dwarf2out.c (should_emit_struct_debug): Add check
8254 for type_decl variable is not NULL.
8256 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8258 * config.gcc: Use new winnt-c.c target hooks
8259 * config/t-winnt: New file
8260 * config/winnt-c.c: New file
8261 * doc/tm.texi.in: Document new hook
8262 * doc/tm.texi: Regenerated
8264 2013-09-13 Jan Hubicka <jh@suse.cz>
8267 * ipa-inline.c (check_callers): New function.
8268 (check_caller_edge): Remove.
8269 (want_inline_function_to_all_callers_p): Also permit alises that are
8271 (inline_to_all_callers): Terminate the walk when devirtualization
8272 introduce new calls.
8274 2013-09-13 Jan Hubicka <jh@suse.cz>
8276 * ipa-inline-analysis.c (struct growth_data): Add node.
8277 (do_estimate_growth_1): Fix detection of recursion.
8279 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8281 PR tree-optimization/58392
8282 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8283 to avoid shadowing of outer loop variable. If
8284 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8285 replace_by_duplicate_decl simduid of loops that have it set and
8286 set dest_cfun->has_simduid_loops and/or
8287 dest_cfun->has_force_vect_loops.
8288 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
8289 instead of maybe_lookup_decl.
8290 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
8291 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
8292 call. Copy over force_vect and copy and remap simduid. Set
8293 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
8294 (copy_cfg_body): Remove blocks_to_copy argument. Use
8295 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
8296 caller. Don't set cfun->has_simduid_loops and/or
8297 cfun->has_force_vect_loops here.
8298 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
8300 (expand_call_inline, tree_function_versioning): Adjust copy_body
8303 2013-09-13 Martin Jambor <mjambor@suse.cz>
8306 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
8307 the assert if the edge was a speculative one.
8309 2013-09-13 Richard Biener <rguenther@suse.de>
8311 * tree-data-ref.h (known_dependences_p): Move here ...
8312 * tree-loop-distribution.c (known_dependences_p): ... from here.
8313 (dump_rdg_component, debug_rdg_component): Remove.
8315 (generate_loops_for_partition): Use gimple_uid instead of
8316 relying on matching stmt visit order.
8317 (rdg_build_partitions): Take starting stmt vector.
8318 (ldist_gen): Merge into ...
8319 (distribute_loop): ... this function. Do not compute starting
8321 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
8323 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8325 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
8326 Set type attribute correctly. Set predicable_short_it attribute.
8327 (cmpsi_shiftsi): Remove %? from output template.
8329 2013-09-13 Richard Biener <rguenther@suse.de>
8331 * tree-loop-distribution.c (struct rdg_component,
8332 rdg_defs_used_in_other_loops_p, free_rdg_components,
8333 rdg_build_components): Remove.
8334 (stmts_from_loop): Do not record virtual PHIs.
8335 (generate_loops_for_partition): Skip virtual PHIs.
8336 (build_rdg_partition_for_component): Rename to ...
8337 (build_rdg_partition_for_vertex): ... this and adjust.
8338 (rdg_build_partitions): Take a vector of starting vertices
8339 instead of components. Remove unnecessary leftover handling.
8340 (ldist_gen): Do not build components or record other stores.
8341 (distribute_loop): Do not distribute loops containing stmts
8344 2013-09-13 Christian Bruel <christian.bruel@st.com>
8347 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
8349 2013-09-13 Kai Tietz <ktietz@redhat.com>
8351 * config.gcc: Separate cases for mingw and cygwin targets,
8352 and add 64-bit cygwin target case.
8354 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
8355 dll-export inline-functions.
8356 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
8358 2013-09-13 Jeff Law <law@redhat.com>
8362 2013-09-06 Jeff Law <law@redhat.com>
8364 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8365 edge implied equivalences into successor phis.
8367 2013-09-12 DJ Delorie <dj@redhat.com>
8369 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8372 2013-09-12 Brooks Moses <bmoses@google.com>
8375 * Makefile.in: Do not install driver binaries in $(target)/bin.
8377 2013-09-12 DJ Delorie <dj@redhat.com>
8379 * config/rl78/rl78.opt (mrelax): New.
8380 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8381 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8383 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8385 (rl78_expand_epilogue): Likewise.
8387 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8390 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8391 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8392 pointer elimination with using remove_reg_equal_offset_note.
8394 2013-09-12 DJ Delorie <dj@redhat.com>
8396 * config/msp430/: New port.
8397 * config.gcc (msp430): Added.
8398 * doc/invoke.texi: Document MSP430 options.
8399 * doc/install.texi: Document msp430-elf
8400 * doc/md.texi: Document msp430-elf
8401 * doc/contrib.texi: Document msp430-elf
8403 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8406 2013-09-12 Martin Jambor <mjambor@suse.cz>
8409 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8410 reference descriptor is NULL.
8411 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8414 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8416 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8417 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8418 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8419 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8420 make_temp_ssa_name): move to tree-ssanames.h
8421 * tree-ssa-alias.h: Move prototype.
8422 * tree-ssa.h: Include tree-ssanames.h.
8423 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8424 * tree-ssanames.h: New. Move items from tree-flow*.h
8425 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8427 2013-09-12 Richard Biener <rguenther@suse.de>
8429 PR tree-optimization/58404
8430 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8431 propagate non-invariant addresses into dereferences wrapped
8432 in component references.
8434 2013-09-12 Richard Biener <rguenther@suse.de>
8436 PR tree-optimization/58402
8437 * passes.def: Move pass_late_warn_uninitialized later.
8439 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8441 * tree-ssa.h: New. Move content from tree-flow.h and
8443 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8444 Move prototypes belonging to tree-ssa.c.
8445 * tree-flow-inline.h (redirect_edge_var_map_def,
8446 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8448 * gimple.h: Adjust prototypes.
8449 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8451 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8452 * tree.h: Move prototype to tree-ssa.h.
8453 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8454 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8455 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8456 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8457 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8458 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8459 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8460 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8461 graphite-blocking.c, graphite-clast-to-gimple.c,
8462 graphite-dependences.c, graphite-interchange.c,
8463 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8464 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8465 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8466 ipa-reference.c, ipa-split.c, ipa-utils.c,
8467 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8468 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8469 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8470 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8471 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8472 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8473 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8474 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8475 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8476 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8477 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8478 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8479 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8480 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8481 value-prof.c, var-tracking.c,
8482 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8484 2013-09-12 Richard Biener <rguenther@suse.de>
8486 PR tree-optimization/58396
8487 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8488 (build_rdg): Take a loop-nest parameter, fix memleaks.
8489 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8491 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8493 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8496 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8498 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8499 The FMA instruction names should have a 'v' prefix.
8501 2013-09-12 Richard Biener <rguenther@suse.de>
8503 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8504 (dot_rdg): Use popen instead of system in optional code.
8505 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8506 (already_processed_vertex_p): Adjust.
8507 (has_anti_or_output_dependence, predecessor_has_mem_write,
8508 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8509 rdg_flag_uses): Remove.
8510 (rdg_flag_vertex): Simplify.
8511 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8513 (build_rdg_partition_for_component): Process the first vertex
8514 of a component only.
8515 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8517 2013-09-12 Alan Modra <amodra@gmail.com>
8519 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8521 2013-09-11 DJ Delorie <dj@redhat.com>
8522 Nick Clifton <nickc@redhat.com>
8524 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8525 (rl78_stack_based_mem): New.
8526 * config/rl78/constraints.md (Iv08): New.
8535 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8537 (movsi): Change from expand to insn-and-split.
8538 (ashrsi3): Clobber AX.
8542 * config/rl78/rl78.md (CC_REG): Fix.
8543 (addsi3): Allow memory and immediate operands.
8544 (addsi3_internal): Split into...
8545 (addsi3_internal_virt): ...new, and ...
8546 (addsi3_internal_real): ...new.
8548 (subsi3_internal_virt): New.
8549 (subsi3_internal_real): New.
8550 (mulsi3): Add memory operand.
8551 (mulsi3_rl78): Likewise.
8552 (mulsi3_g13): Likewise.
8553 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8554 (cbranchqi4_real): Add more constraint options.
8555 (cbranchhi4_real): Expand pattern.
8556 (cbranchhi4_real_signed): New.
8557 (cbranchhi4_real_inverted): New.
8558 (cbranchsi4_real_lt): New.
8559 (cbranchsi4_real_ge): New.
8560 (cbranchsi4_real_signed): New.
8561 (cbranchsi4_real): New.
8563 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8565 (lshrsi3_virt): Likewise.
8566 (ashlsi3_virt): Likewise.
8567 (cbranchqi4_virt_signed): New.
8568 (cbranchhi4_virt_signed): New.
8569 (cbranchsi4_virt): New.
8570 * config/rl78/rl78.c: Whitespace fixes throughout.
8571 (move_elim_pass): New.
8572 (pass_data_rl78_move_elim): New.
8573 (pass_rl78_move_elim): New.
8574 (make_pass_rl78_move_elim): New.
8575 (rl78_devirt_info): Run devirt earlier.
8576 (rl78_move_elim_info): New.
8577 (rl78_asm_file_start): Register it.
8578 (rl78_split_movsi): New.
8579 (rl78_as_legitimate_address): Allow virtual base registers when
8581 (rl78_addr_space_convert): Remove spurious debug stuff.
8582 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8583 (rl78_print_operand): More cases for not printing '#'.
8584 (rl78_expand_compare): Remove most of the logic.
8585 (content_memory): New.
8586 (clear_content_memory): New.
8587 (get_content_index): New.
8588 (get_content_name): New.
8589 (display_content_memory): New.
8590 (update_content): New.
8591 (record_content): New.
8592 (already_contains): New.
8593 (insn_ok_now): Re-recog insns with virtual registers.
8594 (add_postponed_content_update): New.
8595 (process_postponed_content_update): New.
8596 (gen_and_emit_move): New.
8597 (transcode_memory_rtx): Record new location content.
8598 Use gen_and_emit_move.
8599 (force_into_acc): New.
8600 (move_to_acc): Use gen_and_emit_move.
8601 (move_from_acc): Likewise.
8602 (move_acc_to_reg): Likewise.
8603 (move_to_x): Likewise.
8604 (move_to_hl): Likewise.
8605 (move_to_de): Likewise.
8606 (rl78_alloc_physical_registers_op1): Record location content.
8607 (has_constraint): New.
8608 (rl78_alloc_physical_registers_op2): Record location content.
8610 (rl78_alloc_physical_registers_ro1): Likewise.
8611 (rl78_alloc_physical_registers_cmp): Likewise.
8612 (rl78_alloc_physical_registers_umul): Likewise.
8613 (rl78_alloc_address_registers_macax): New.
8614 (rl78_alloc_physical_registers): Initialize and set location
8615 content memory as needed.
8616 (rl78_reorg): Make sure split2 is called.
8617 (rl78_rtx_costs): New.
8619 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8621 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8622 for (not (neg ...)) and (neg (not ...)) cases.
8624 2013-09-11 Richard Biener <rguenther@suse.de>
8627 * passes.def: Split critical edges before late uninit warning passes.
8628 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8630 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8632 PR tree-optimization/58385
8633 * fold-const.c (build_range_check): If both low and high are NULL,
8634 use omit_one_operand_loc to preserve exp side-effects.
8636 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8638 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8640 2013-09-11 Richard Biener <rguenther@suse.de>
8642 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8643 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8644 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8645 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8646 stmts_from_loop, known_dependences_p, build_empty_rdg,
8647 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8648 * tree-loop-distribution.c: ... here.
8649 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8650 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8651 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8652 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8653 * tree-loop-distribution.c: ... here.
8654 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8655 (struct partition_s): Add loops member.
8656 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8657 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8658 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8660 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8661 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8662 Sergey Lega <sergey.s.lega@intel.com>
8663 Anna Tikhonova <anna.tikhonova@intel.com>
8664 Ilya Tocar <ilya.tocar@intel.com>
8665 Andrey Turetskiy <andrey.turetskiy@intel.com>
8666 Ilya Verbin <ilya.verbin@intel.com>
8667 Kirill Yukhin <kirill.yukhin@intel.com>
8668 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8670 * config/i386/constraints.md (k): New.
8672 * config/i386/i386.c (const regclass_map): Add new mask registers.
8673 (dbx_register_map): Ditto.
8674 (dbx64_register_map): Ditto.
8675 (svr4_dbx_register_map): Ditto.
8676 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8678 (ix86_preferred_reload_class): Disable constants for mask registers.
8679 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8680 (ix86_hard_regno_mode_ok): Support new mask registers.
8681 (x86_order_regs_for_local_alloc): Ditto.
8682 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8683 (FIXED_REGISTERS): Add new mask registers.
8684 (CALL_USED_REGISTERS): Ditto.
8685 (REG_ALLOC_ORDER): Ditto.
8686 (VALID_MASK_REG_MODE): New.
8687 (FIRST_MASK_REG): Ditto.
8688 (LAST_MASK_REG): Ditto.
8689 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8690 (MAYBE_MASK_CLASS_P): New.
8691 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8692 (REG_CLASS_CONTENTS): Ditto.
8693 (MASK_REGNO_P): New.
8694 (ANY_MASK_REG_P): Ditto.
8695 (HI_REGISTER_NAMES): Add new mask registers.
8696 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8697 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8699 (attribute "type"): Add mskmov, msklog.
8700 (attribute "length_immediate"): Support them.
8701 (attribute "memory"): Ditto.
8702 (attribute "prefix_0f"): Ditto.
8703 (*movhi_internal): Support new mask registers.
8704 (*movqi_internal): Ditto.
8705 (define_split): Split out clobber pattern is a logic
8706 insn on mask registers.
8707 (*k<logic><mode>): New.
8708 (*andhi_1): Extend to support mask regs.
8709 (*andqi_1): Extend to support mask regs.
8711 (define_split): Split and-not to and and not if operands
8713 (*<code><mode>_1): Separate HI mode to new pattern...
8714 (*<code>hi_1): This.
8715 (*<code>qi_1): Extend to support mask regs.
8717 (kortestzhi): Ditto.
8718 (kortestchi): Ditto.
8720 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8721 (*one_cmplhi2_1): ...Here, now with mask registers support.
8722 (*one_cmplqi2_1): Support new mask registers.
8723 (HI/QImode arithmetics splitter): Don't split if mask registers
8725 (HI/QImode not splitter): Ditto.
8726 * config/i386/predicated.md (mask_reg_operand): New.
8727 (general_reg_operand): Ditto.
8729 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8731 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8732 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8734 2013-09-10 Jeff Law <law@redhat.com>
8736 PR tree-optimization/58380
8737 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8738 of threading through a buried loop header.
8740 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8741 return value for single successor case.
8743 2013-09-10 Jan Hubicka <jh@suse.cz>
8745 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8747 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8750 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8751 support conditional execution.
8752 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8754 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8756 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8758 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8759 Check destination too to check move insn.
8760 (undo_optional_reloads): Add check that the original peudo did not
8761 changed its allocation and the optional reload was inherited on last
8762 inheritance pass. Break loop after deciding to keep optional reload.
8763 (lra_undo_inheritance): Add check that inherited pseudo still in
8766 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8768 * config/aarch64/aarch64.md (generic_sched): New.
8769 * config/aarch64/aarch64-generic.md (load): Make conditional
8770 on generic_sched attribute.
8771 (nonload): Likewise.
8773 2013-09-10 Jan Hubicka <jh@suse.cz>
8775 * lto-cgraph.c: Include ipa-utils.h.
8776 (compute_ltrans_boundary): Also add possible targets into the boundary.
8778 2013-09-10 Jan Hubicka <jh@suse.cz>
8780 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8781 VAR_DECL of vtable rather than full expression.
8783 2013-09-10 Jan Hubicka <jh@suse.cz>
8784 Paolo Carlini <paolo.carlini@oracle.com>
8786 * cgraphunit.c (analyze_functions): Save input_location, set it
8787 to UNKNOWN_LOCATION and restore it at the end.
8789 2013-09-10 Martin Jambor <mjambor@suse.cz>
8791 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8792 represented by a thunk.
8794 2013-09-10 Jeff Law <law@redhat.com>
8796 PR tree-optimization/58343
8797 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8798 jump threading requests through loop headers buried in the middle
8799 of a jump threading path.
8801 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8802 in return value/type.
8804 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8806 PR rtl-optimization/58365
8807 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8808 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8811 2013-09-10 Richard Biener <rguenther@suse.de>
8813 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8814 * tree-data-ref.c (create_rdg_vertices): Collect all data
8815 references, signal failure to the caller, use data-ref API.
8816 (build_rdg): Compute data references only once. Maintain lifetime
8817 of data references and data dependences from within RDG.
8818 (free_rdg): Free dependence relations.
8819 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8820 inventing extra dependences.
8821 (distribute_loop): Update for RDG API changes.
8823 2013-09-10 Kai Tietz <ktietz@redhat.com>
8825 * doc/invoke.texi (fms-extensions): Document changed
8826 behavior for ms-abi targets.
8827 * config/i386/i386.c (ix86_option_override_internal):
8828 Set default value of option -fms-extension for ms-abi targets.
8830 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8832 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8834 2013-09-10 Alan Modra <amodra@gmail.com>
8837 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8839 2013-09-10 Alan Modra <amodra@gmail.com>
8841 * config/rs6000/predicates.md (add_cint_operand): New.
8842 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8843 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8844 using add_cint_operand.
8845 (largetoc_high_plus_aix): Likewise.
8847 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8849 PR tree-optimization/58364
8850 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8851 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8852 the current range can't be an unconditional true or false.
8854 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8856 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8858 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8860 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8862 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8865 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8867 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8869 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8870 NO_REGS for immediate that can't be moved directly into FP_REGS.
8872 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8874 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8875 comparison with negated operand.
8876 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8879 2013-09-09 Richard Biener <rguenther@suse.de>
8882 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8883 of a subloop record all its block as affecting loop-closed SSA form.
8885 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8887 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8888 of an rtx/bitpos pair.
8889 (store_fixed_bit_field): Update accordingly.
8891 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8893 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8895 * builtins.c (expand_errno_check): Likewise.
8896 * dwarf2cfi.c (init_return_column_size): Likewise.
8897 * except.c (sjlj_mark_call_sites): Likewise.
8898 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8899 * lra-constraints.c (emit_inc): Likewise.
8900 * ree.c (combine_set_extension): Likewise.
8901 * regmove.c (fixup_match_2): Likewise.
8902 * reload1.c (inc_for_reload): Likewise.
8904 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8906 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8907 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8908 Use gen_int_mode with the mode of the associated simplify_* call.
8909 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8910 * expmed.c (expand_shift_1): Likewise.
8911 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8912 * loop-iv.c (iv_number_of_iterations): Likewise.
8913 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8914 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8916 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8918 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8919 of the associated expand_* call.
8920 (asan_emit_stack_protection): Likewise.
8921 * builtins.c (round_trampoline_addr): Likewise.
8922 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8923 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8924 (emit_store_flag): Likewise.
8925 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8926 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8928 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8929 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8930 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8931 * modulo-sched.c (generate_prolog_epilog): Likewise.
8932 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8933 (expand_ctz, expand_ffs, expand_unop): Likewise.
8935 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8937 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8938 of the associated gen_rtx_* call.
8939 * caller-save.c (init_caller_save): Likewise.
8940 * combine.c (find_split_point, make_extraction): Likewise.
8941 (make_compound_operation): Likewise.
8942 * dwarf2out.c (mem_loc_descriptor): Likewise.
8943 * explow.c (plus_constant, probe_stack_range): Likewise.
8944 * expmed.c (expand_mult_const): Likewise.
8945 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8946 * reload1.c (init_reload): Likewise.
8947 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8948 * var-tracking.c (adjust_mems): Likewise.
8949 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8950 rather than gen_rtx_fmt_ee.
8952 2013-09-09 Jan Hubicka <jh@suse.cz>
8955 * value-prof.c (gimple_ic): Copy also abnormal edges.
8957 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8959 * asan.c (asan_shadow_cst): Use gen_int_mode.
8961 2013-09-08 Jan Hubicka <jh@suse.cz>
8963 * ipa-profile.c: Add toplevel comment.
8964 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8965 (contains_hot_call_p): New function.
8966 (ipa_propagate_frequency): Set frequencies based on counts when
8968 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8969 profile; do not tamper with profile after inlining if it is read.
8971 2013-09-08 Jan Hubicka <jh@suse.cz>
8973 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8976 2013-09-08 Jan Hubicka <jh@suse.cz>
8978 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8981 2013-09-08 Jeff Law <law@redhat.com>
8984 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8987 2013-09-08 Andi Kleen <ak@linux.intel.com>
8989 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8991 2013-09-08 Jan Hubicka <jh@suse.cz>
8993 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8995 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8997 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8998 for non-debug insns.
8999 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9000 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9001 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9004 2013-09-08 Jan Hubicka <jh@suse.cz>
9006 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9007 targets and devirtualize to BUILT_IN_UNREACHABLE.
9008 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9009 * ipa.c (walk_polymorphic_call_targets): New function.
9010 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9011 functions; use the new timevar.
9012 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9013 was removed from the program.
9014 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9015 not consider it in the walk when its vtable is dead.
9016 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9018 (devirt_variable_node_removal_hook): New function.
9019 (possible_polymorphic_call_targets): Also register
9020 devirt_variable_node_removal_hook.
9021 (ipa_devirt): Do not do non-speculative devirtualization.
9022 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9024 2013-09-08 Jan Hubicka <jh@suse.cz>
9026 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9027 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9028 varpool_remove_variable_insertion_hook): Declare.
9029 * varpool.c (varpool_node_hook_list): New structure.
9030 (first_varpool_node_removal_hook,
9031 first_varpool_variable_insertion_hook): New variables.
9032 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9033 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9034 varpool_remove_variable_insertion_hook,
9035 varpool_call_variable_insertion_hooks): New functions.
9036 (varpool_remove_node): Use it.
9038 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9041 * diagnostic.c (diagnostic_build_prefix): When s.file is
9042 "<built-in>" don't output line and column numbers.
9044 2013-09-06 Jan Hubicka <jh@suse.cz>
9046 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9047 * lto-streamer-out.c: Stream thunks, too.
9048 * lto-streamer-in.c (input_function): Pop cfun here
9049 (lto_read_body): Instead of here.
9051 2013-09-06 Caroline Tice <cmtice@google.com>
9053 * doc/install.texi: Add documentation for the --enable-vtable-verify
9054 and the --disable-libvtv configure options.
9056 2013-09-06 Jeff Law <law@redhat.com>
9058 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9059 edge implied equivalences into successor phis.
9061 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9063 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9065 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9067 * resource.c (mark_target_live_regs): Compute resources taking
9068 into account if a call is predicated or not.
9070 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9072 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9073 the compiler in the function names.
9075 2013-09-06 Jan Hubicka <jh@suse.cz>
9078 * ipa-inline.c (has_caller_p): New function.
9079 (want_inline_function_to_all_callers_p): Use it.
9080 (sum_callers, inline_to_all_callers): Break out from ...
9081 (ipa_inline): ... here.
9083 2013-09-06 Jan Hubicka <jh@suse.cz>
9085 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9086 only when AVX is enabled.
9088 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9090 * config/aarch64/aarch64.md
9091 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9092 is fpsimd_<load/store>2.
9093 (load_pair<mode>): Likewise.
9094 (store_pair<mode>): Likewise.
9096 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9098 * config/arm/types.md (type): Add "mrs" type.
9099 * config/aarch64/aarch64.md
9100 (aarch64_load_tp_hard): Make type "mrs".
9102 (load_tp_hard): Make type "mrs".
9103 * config/arm/cortex-a15.md: Update with new attributes.
9104 * config/arm/cortex-a5.md: Update with new attributes.
9105 * config/arm/cortex-a53.md: Update with new attributes.
9106 * config/arm/cortex-a7.md: Update with new attributes.
9107 * config/arm/cortex-a8.md: Update with new attributes.
9108 * config/arm/cortex-a9.md: Update with new attributes.
9109 * config/arm/cortex-m4.md: Update with new attributes.
9110 * config/arm/cortex-r4.md: Update with new attributes.
9111 * config/arm/fa526.md: Update with new attributes.
9112 * config/arm/fa606te.md: Update with new attributes.
9113 * config/arm/fa626te.md: Update with new attributes.
9114 * config/arm/fa726te.md: Update with new attributes.
9116 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9118 * config/aarch64/aarch64.md
9119 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9120 (*movtf_aarch64): Likewise.
9122 (thumb1_movdi_insn): Use "multiple" for type where more than one
9123 instruction is used for a move.
9124 (*arm32_movhf): Likewise.
9125 (*thumb_movdf_insn): Likewise.
9127 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9129 * config/arm/types.md (type): Rename fcpys to fmov.
9131 (*arm_movsi_vfp): Rename type fcpys as fmov.
9132 (*thumb2_movsi_vfp): Likewise
9133 (*movhf_vfp_neon): Likewise
9134 (*movhf_vfp): Likewise
9135 (*movsf_vfp): Likewise
9136 (*thumb2_movsf_vfp): Likewise
9137 (*movsfcc_vfp): Likewise
9138 (*thumb2_movsfcc_vfp): Likewise
9139 * config/aarch64/aarch64-simd.md
9140 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9141 * config/aarch64/aarch64.md
9142 (*movsi_aarch64): Replace type mov_reg with fmovs.
9143 (*movdi_aarch64): Likewise
9144 (*movsf_aarch64): Likewise
9145 (*movdf_aarch64): Likewise
9147 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9148 * config/arm/iwmmxt.md
9149 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9150 * config/arm/arm1020e.md: Update with new attributes.
9151 * config/arm/cortex-a15-neon.md: Update with new attributes.
9152 * config/arm/cortex-a5.md: Update with new attributes.
9153 * config/arm/cortex-a53.md: Update with new attributes.
9154 * config/arm/cortex-a7.md: Update with new attributes.
9155 * config/arm/cortex-a8-neon.md: Update with new attributes.
9156 * config/arm/cortex-a9.md: Update with new attributes.
9157 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9158 * config/arm/cortex-r4f.md: Update with new attributes.
9159 * config/arm/marvell-pj4.md: Update with new attributes.
9160 * config/arm/vfp11.md: Update with new attributes.
9162 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9164 * config/aarch64/aarch64.md
9165 (*madd<mode>): Fix type attribute.
9166 (*maddsi_uxtw): Likewise.
9167 (*msub<mode>): Likewise.
9168 (*msubsi_uxtw): Likewise.
9169 (<su_optab>maddsidi4): Likewise.
9170 (<su_optab>msubsidi4): Likewise.
9172 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9174 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9175 * config/arm/arm.md (core_cycles): Remove fdiv.
9176 * config/arm/vfp.md:
9177 (*sqrtsf2_vfp): Update for attribute changes.
9178 (*sqrtdf2_vfp): Likewise.
9179 * config/aarch64/aarch64.md:
9180 (sqrt<mode>2): Update for attribute changes.
9181 * config/arm/arm1020e.md: Update with new attributes.
9182 * config/arm/cortex-a15-neon.md: Update with new attributes.
9183 * config/arm/cortex-a5.md: Update with new attributes.
9184 * config/arm/cortex-a53.md: Update with new attributes.
9185 * config/arm/cortex-a7.md: Update with new attributes.
9186 * config/arm/cortex-a8-neon.md: Update with new attributes.
9187 * config/arm/cortex-a9.md: Update with new attributes.
9188 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9189 * config/arm/cortex-r4f.md: Update with new attributes.
9190 * config/arm/marvell-pj4.md: Update with new attributes.
9191 * config/arm/vfp11.md: Update with new attributes.
9193 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9195 * config/arm/types.md
9196 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9197 * config/aarch64/aarch64.md
9198 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9200 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9201 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9202 (float<GPI:mode><GPF:mode>2): Likewise.
9204 (*truncsisf2_vfp): Update with new attributes.
9205 (*truncsidf2_vfp): Likewise.
9206 (fixuns_truncsfsi2): Likewise.
9207 (fixuns_truncdfsi2): Likewise.
9208 (*floatsisf2_vfp): Likewise.
9209 (*floatsidf2_vfp): Likewise.
9210 (floatunssisf2): Likewise.
9211 (floatunssidf2): Likewise.
9212 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9213 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9214 * config/arm/arm1020e.md: Update with new attributes.
9215 * config/arm/cortex-a15-neon.md: Update with new attributes.
9216 * config/arm/cortex-a5.md: Update with new attributes.
9217 * config/arm/cortex-a53.md: Update with new attributes.
9218 * config/arm/cortex-a7.md: Update with new attributes.
9219 * config/arm/cortex-a8-neon.md: Update with new attributes.
9220 * config/arm/cortex-a9.md: Update with new attributes.
9221 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9222 * config/arm/cortex-r4f.md: Update with new attributes.
9223 * config/arm/marvell-pj4.md: Update with new attributes.
9224 * config/arm/vfp11.md: Update with new attributes.
9226 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9228 * config/aarch64/arm_neon.h
9229 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9230 (vqtbx<1,2,3,4><q>_s8): Likewise.
9232 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9234 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9235 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9236 (add<mode>3): Add type attribute.
9237 (add<mode>3): Likewise.
9238 (usadd<mode>3): Likewise.
9239 (ssadd<mode>3): Likewise.
9240 (sub<mode>3): Likewise.
9241 (sub<mode>3): Likewise.
9242 (ussub<mode>3): Likewise.
9243 (sssub<mode>3): Likewise.
9244 (ssmulsa3): Likewise.
9245 (usmulusa3): Likewise.
9246 (arm_usatsihi): Likewise.
9248 (*movdi_vfp): Add types for all instructions.
9249 (*movdi_vfp_cortexa8): Likewise.
9250 (*movhf_vfp_neon): Likewise.
9251 (*movhf_vfp): Likewise.
9252 (*movdf_vfp): Likewise.
9253 (*thumb2_movdf_vfp): Likewise.
9254 (*thumb2_movdfcc_vfp): Likewise.
9255 * config/arm/arm.md: Add type attribute to all insn patterns.
9256 (*thumb1_adddi3): Add type attribute.
9257 (*arm_adddi3): Likewise.
9258 (*adddi_sesidi_di): Likewise.
9259 (*adddi_zesidi_di): Likewise.
9260 (*thumb1_addsi3): Likewise.
9261 (addsi3_compare0): Likewise.
9262 (*addsi3_compare0_scratch): Likewise.
9263 (*compare_negsi_si): Likewise.
9264 (cmpsi2_addneg): Likewise.
9265 (*addsi3_carryin_<optab>): Likewise.
9266 (*addsi3_carryin_alt2_<optab>): Likewise.
9267 (*addsi3_carryin_clobercc_<optab>): Likewise.
9268 (*subsi3_carryin): Likewise.
9269 (*subsi3_carryin_const): Likewise.
9270 (*subsi3_carryin_compare): Likewise.
9271 (*subsi3_carryin_compare_const): Likewise.
9272 (*arm_subdi3): Likewise.
9273 (*thumb_subdi3): Likewise.
9274 (*subdi_di_zesidi): Likewise.
9275 (*subdi_di_sesidi): Likewise.
9276 (*subdi_zesidi_di): Likewise.
9277 (*subdi_sesidi_di): Likewise.
9278 (*subdi_zesidi_ze): Likewise.
9279 (thumb1_subsi3_insn): Likewise.
9280 (*arm_subsi3_insn): Likewise.
9281 (*anddi3_insn): Likewise.
9282 (*anddi_zesidi_di): Likewise.
9283 (*anddi_sesdi_di): Likewise.
9284 (*ne_zeroextracts): Likewise.
9285 (*ne_zeroextracts): Likewise.
9286 (*ite_ne_zeroextr): Likewise.
9287 (*ite_ne_zeroextr): Likewise.
9288 (*anddi_notdi_di): Likewise.
9289 (*anddi_notzesidi): Likewise.
9290 (*anddi_notsesidi): Likewise.
9291 (andsi_notsi_si): Likewise.
9292 (thumb1_bicsi3): Likewise.
9293 (*iordi3_insn): Likewise.
9294 (*iordi_zesidi_di): Likewise.
9295 (*iordi_sesidi_di): Likewise.
9296 (*thumb1_iorsi3_insn): Likewise.
9297 (*xordi3_insn): Likewise.
9298 (*xordi_zesidi_di): Likewise.
9299 (*xordi_sesidi_di): Likewise.
9300 (*arm_xorsi3): Likewise.
9301 (*andsi_iorsi3_no): Likewise.
9302 (*smax_0): Likewise.
9303 (*smax_m1): Likewise.
9304 (*arm_smax_insn): Likewise.
9305 (*smin_0): Likewise.
9306 (*arm_smin_insn): Likewise.
9307 (*arm_umaxsi3): Likewise.
9308 (*arm_uminsi3): Likewise.
9309 (*minmax_arithsi): Likewise.
9310 (*minmax_arithsi_): Likewise.
9311 (*satsi_<SAT:code>): Likewise.
9312 (arm_ashldi3_1bit): Likewise.
9313 (arm_ashrdi3_1bit): Likewise.
9314 (arm_lshrdi3_1bit): Likewise.
9315 (*arm_negdi2): Likewise.
9316 (*thumb1_negdi2): Likewise.
9317 (*arm_negsi2): Likewise.
9318 (*thumb1_negsi2): Likewise.
9319 (*negdi_extendsid): Likewise.
9320 (*negdi_zero_extend): Likewise.
9321 (*arm_abssi2): Likewise.
9322 (*thumb1_abssi2): Likewise.
9323 (*arm_neg_abssi2): Likewise.
9324 (*thumb1_neg_abss): Likewise.
9325 (one_cmpldi2): Likewise.
9326 (extend<mode>di2): Likewise.
9327 (*compareqi_eq0): Likewise.
9328 (*arm_extendhisi2addsi): Likewise.
9329 (*arm_movdi): Likewise.
9330 (*thumb1_movdi_insn): Likewise.
9331 (*arm_movt): Likewise.
9332 (*thumb1_movsi_insn): Likewise.
9333 (pic_add_dot_plus_four): Likewise.
9334 (pic_add_dot_plus_eight): Likewise.
9335 (tls_load_dot_plus_eight): Likewise.
9336 (*thumb1_movhi_insn): Likewise.
9337 (*thumb1_movsf_insn): Likewise.
9338 (*movdf_soft_insn): Likewise.
9339 (*thumb_movdf_insn): Likewise.
9340 (cbranchsi4_insn): Likewise.
9341 (cbranchsi4_scratch): Likewise.
9342 (*negated_cbranchsi4): Likewise.
9343 (*tbit_cbranch): Likewise.
9344 (*tlobits_cbranch): Likewise.
9345 (*tstsi3_cbranch): Likewise.
9346 (*cbranchne_decr1): Likewise.
9347 (*addsi3_cbranch): Likewise.
9348 (*addsi3_cbranch_scratch): Likewise.
9349 (*arm_cmpdi_insn): Likewise.
9350 (*arm_cmpdi_unsig): Likewise.
9351 (*arm_cmpdi_zero): Likewise.
9352 (*thumb_cmpdi_zero): Likewise.
9353 (*deleted_compare): Likewise.
9354 (*mov_scc): Likewise.
9355 (*mov_negscc): Likewise.
9356 (*mov_notscc): Likewise.
9357 (*cstoresi_eq0_thumb1_insn): Likewise.
9358 (cstoresi_nltu_thumb1): Likewise.
9359 (cstoresi_ltu_thu): Likewise.
9360 (thumb1_addsi3_addgeu): Likewise.
9361 (*arm_jump): Likewise.
9362 (*thumb_jump): Likewise.
9363 (*check_arch2): Likewise.
9364 (arm_casesi_internal): Likewise.
9365 (thumb1_casesi_dispatch): Likewise.
9366 (*arm_indirect_jump): Likewise.
9367 (*thumb1_indirect_jump): Likewise.
9369 (*and_scc): Likewise.
9370 (*ior_scc): Likewise.
9371 (*compare_scc): Likewise.
9372 (*cond_move): Likewise.
9373 (*cond_arith): Likewise.
9374 (*cond_sub): Likewise.
9375 (*cmp_ite0): Likewise.
9376 (*cmp_ite1): Likewise.
9377 (*cmp_and): Likewise.
9378 (*cmp_ior): Likewise.
9379 (*ior_scc_scc): Likewise.
9380 (*ior_scc_scc_cmp): Likewise.
9381 (*and_scc_scc): Likewise.
9382 (*and_scc_scc_cmp): Likewise.
9383 (*and_scc_scc_nod): Likewise.
9384 (*negscc): Likewise.
9385 (movcond_addsi): Likewise.
9386 (movcond): Likewise.
9387 (*ifcompare_plus_move): Likewise.
9388 (*if_plus_move): Likewise.
9389 (*ifcompare_move_plus): Likewise.
9390 (*if_move_plus): Likewise.
9391 (*ifcompare_arith_arith): Likewise.
9392 (*if_arith_arith): Likewise.
9393 (*ifcompare_arith_move): Likewise.
9394 (*if_arith_move): Likewise.
9395 (*ifcompare_move_arith): Likewise.
9396 (*if_move_arith): Likewise.
9397 (*ifcompare_move_not): Likewise.
9398 (*if_move_not): Likewise.
9399 (*ifcompare_not_move): Likewise.
9400 (*if_not_move): Likewise.
9401 (*ifcompare_shift_move): Likewise.
9402 (*if_shift_move): Likewise.
9403 (*ifcompare_move_shift): Likewise.
9404 (*if_move_shift): Likewise.
9405 (*ifcompare_shift_shift): Likewise.
9406 (*ifcompare_not_arith): Likewise.
9407 (*ifcompare_arith_not): Likewise.
9408 (*if_arith_not): Likewise.
9409 (*ifcompare_neg_move): Likewise.
9410 (*if_neg_move): Likewise.
9411 (*ifcompare_move_neg): Likewise.
9412 (*if_move_neg): Likewise.
9413 (prologue_thumb1_interwork): Likewise.
9414 (*cond_move_not): Likewise.
9415 (*sign_extract_onebit): Likewise.
9416 (*not_signextract_onebit): Likewise.
9417 (stack_tie): Likewise.
9418 (align_4): Likewise.
9419 (align_8): Likewise.
9420 (consttable_end): Likewise.
9421 (consttable_1): Likewise.
9422 (consttable_2): Likewise.
9423 (consttable_4): Likewise.
9424 (consttable_8): Likewise.
9425 (consttable_16): Likewise.
9426 (*thumb1_tablejump): Likewise.
9427 (prefetch): Likewise.
9428 (force_register_use): Likewise.
9429 (thumb_eh_return): Likewise.
9430 (load_tp_hard): Likewise.
9431 (load_tp_soft): Likewise.
9432 (tlscall): Likewise.
9433 (*arm_movtas_ze): Likewise.
9434 (*arm_rev): Likewise.
9435 (*arm_revsh): Likewise.
9436 (*arm_rev16): Likewise.
9437 * config/arm/thumb2.md
9438 (*thumb2_smaxsi3): Likewise.
9439 (*thumb2_sminsi3): Likewise.
9440 (*thumb32_umaxsi3): Likewise.
9441 (*thumb2_uminsi3): Likewise.
9442 (*thumb2_negdi2): Likewise.
9443 (*thumb2_abssi2): Likewise.
9444 (*thumb2_neg_abss): Likewise.
9445 (*thumb2_movsi_insn): Likewise.
9446 (tls_load_dot_plus_four): Likewise.
9447 (*thumb2_movhi_insn): Likewise.
9448 (*thumb2_mov_scc): Likewise.
9449 (*thumb2_mov_negs): Likewise.
9450 (*thumb2_mov_negs): Likewise.
9451 (*thumb2_mov_nots): Likewise.
9452 (*thumb2_mov_nots): Likewise.
9453 (*thumb2_movsicc_): Likewise.
9454 (*thumb2_movsfcc_soft_insn): Likewise.
9455 (*thumb2_indirect_jump): Likewise.
9456 (*thumb2_and_scc): Likewise.
9457 (*thumb2_ior_scc): Likewise.
9458 (*thumb2_ior_scc_strict_it): Likewise.
9459 (*thumb2_cond_move): Likewise.
9460 (*thumb2_cond_arith): Likewise.
9461 (*thumb2_cond_ari): Likewise.
9462 (*thumb2_cond_sub): Likewise.
9463 (*thumb2_negscc): Likewise.
9464 (*thumb2_movcond): Likewise.
9465 (thumb2_casesi_internal): Likewise.
9466 (thumb2_casesi_internal_pic): Likewise.
9467 (*thumb2_alusi3_short): Likewise.
9468 (*thumb2_mov<mode>_shortim): Likewise.
9469 (*thumb2_addsi_short): Likewise.
9470 (*thumb2_subsi_short): Likewise.
9471 (thumb2_addsi3_compare0): Likewise.
9472 (*thumb2_cbz): Likewise.
9473 (*thumb2_cbnz): Likewise.
9474 (*thumb2_one_cmplsi2_short): Likewise.
9475 (*thumb2_negsi2_short): Likewise.
9476 (*orsi_notsi_si): Likewise.
9477 * config/arm/arm1020e.md: Update with new attributes.
9478 * config/arm/arm1026ejs.md: Update with new attributes.
9479 * config/arm/arm1136jfs.md: Update with new attributes.
9480 * config/arm/arm926ejs.md: Update with new attributes.
9481 * config/arm/cortex-a15.md: Update with new attributes.
9482 * config/arm/cortex-a5.md: Update with new attributes.
9483 * config/arm/cortex-a53.md: Update with new attributes.
9484 * config/arm/cortex-a7.md: Update with new attributes.
9485 * config/arm/cortex-a8.md: Update with new attributes.
9486 * config/arm/cortex-a9.md: Update with new attributes.
9487 * config/arm/cortex-m4.md: Update with new attributes.
9488 * config/arm/cortex-r4.md: Update with new attributes.
9489 * config/arm/fa526.md: Update with new attributes.
9490 * config/arm/fa606te.md: Update with new attributes.
9491 * config/arm/fa626te.md: Update with new attributes.
9492 * config/arm/fa726te.md: Update with new attributes.
9494 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9496 * config/aarch64/aarch64-simd.md
9497 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9498 <vwx> iterator to ensure correct register choice.
9499 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9500 (aarch64_sqdmull_n<mode>): Likewise.
9501 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9502 * config/aarch64/arm_neon.h
9503 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9504 (vml<as><q>_n_<su>16): Likewise.
9505 (vml<as>l_high_lane<q>_<su>16): Likewise.
9506 (vml<as>l_high_n_<su>16): Likewise.
9507 (vml<as>l_lane<q>_<su>16): Likewise.
9508 (vml<as>l_n_<su>16): Likewise.
9509 (vmul<q>_lane<q>_<su>16): Likewise.
9510 (vmul<q>_n_<su>16): Likewise.
9511 (vmull_lane<q>_<su>16): Likewise.
9512 (vmull_n_<su>16): Likewise.
9513 (vmull_high_lane<q>_<su>16): Likewise.
9514 (vmull_high_n_<su>16): Likewise.
9515 (vqrdmulh<q>_n_s16): Likewise.
9517 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9519 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9520 have the correct lane parameter.
9522 2013-09-06 Richard Biener <rguenther@suse.de>
9524 * cfganal.c (control_dependences::~control_dependences):
9525 Properly free all of the vector.
9527 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9530 * config/i386/i386.c (ix86_conditional_register_usage):
9531 Proper initialize extended SSE registers.
9533 2013-09-06 Jan Hubicka <jh@suse.cz>
9535 PR tree-optimization/58311
9536 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9538 2013-09-06 Jan Hubicka <jh@suse.cz>
9540 * Makefile.in (tree-sra.o): Update dependencies.
9541 * tree-sra.c: Include ipa-utils.h
9542 (scan_function): Use recursive_call_p.
9543 (has_caller_p): New function.
9544 (cgraph_for_node_and_aliases): Count also callers of aliases.
9546 2013-09-06 Jan Hubicka <jh@suse.cz>
9549 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9550 * tree-tailcall.c: Include ipa-utils.h.
9551 (find_tail_calls): Use it.
9552 * ipa-pure-const.c (check_call): Likewise.
9553 * ipa-utils.c (recursive_call_p): New function.
9554 * ipa-utils.h (recursive_call_p): Dclare.
9555 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9556 (symtab_semantically_equivalent_p): New function.
9557 * Makefile.in (tree-tailcall.o): Update dependencies.
9559 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9561 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9564 2013-09-06 Richard Biener <rguenther@suse.de>
9566 * lto-streamer.h (lto_global_var_decls): Remove.
9567 * Makefile.in (OBJS): Remove lto-symtab.o.
9568 (lto-symtab.o): Remove.
9569 (GTFILES): Remove lto-symtab.c
9570 * lto-symtab.c: Move to lto/
9572 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9574 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9575 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9576 (UNSPEC_FPINT_RINT): New constant definitions.
9577 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9578 definition with 2 attributes.
9579 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9580 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9583 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9585 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9586 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9587 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9588 attribute to "z196_alone".
9589 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9592 2013-09-06 Richard Biener <rguenther@suse.de>
9594 * basic-block.h (class control_dependences): New.
9595 * tree-ssa-dce.c (control_dependence_map): Remove.
9597 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9598 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9599 find_pdom, find_control_dependence, find_all_control_dependences):
9601 (mark_control_dependent_edges_necessary,
9602 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9603 tree_dce_done, perform_tree_ssa_dce): Adjust.
9604 * cfganal.c (set_control_dependence_map_bit,
9605 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9606 find_all_control_dependences): Move from tree-ssa-dce.c and
9607 implement as methods of control_dependences class.
9608 (control_dependences::control_dependences): New.
9609 (control_dependences::~control_dependences): Likewise.
9610 (control_dependences::get_edges_dependent_on): Likewise.
9611 (control_dependences::get_edge): Likewise.
9613 2013-09-04 Jan Hubicka <jh@suse.cz>
9615 * tree.c (types_same_for_odr): Drop overactive check.
9616 * ipa-devirt.c (hash_type_name): Likewise.
9618 2013-09-04 Jan Hubicka <jh@suse.cz>
9620 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9621 (analyze_functions): ... here.
9623 2013-09-04 Jan Hubicka <jh@suse.cz>
9626 * cgraphunit.c (analyze_functions): Clear AUX fields
9627 after processing; initialize assembler name has.
9629 2013-09-05 Jeff Law <law@redhat.com>
9631 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9632 from thread_around_empty_block. Record threading path into PATH.
9633 Recurse if threading through the initial block is successful.
9634 (thread_across_edge): Corresponding changes to slightly simplify.
9636 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9638 * config/aarch64/aarch64.md
9639 (type): Remove frecpe, frecps, frecpx.
9640 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9641 fix to be a TARGET_SIMD instruction.
9642 (aarch64_frecps): Remove.
9643 * config/aarch64/aarch64-simd.md
9644 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9645 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9647 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9648 Sofiane Naci <sofiane.naci@arm.com>
9650 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9651 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9652 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9653 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9654 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9655 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9656 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9657 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9658 into "clz, "rbit". Rename "shift" to "shift_imm".
9659 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9660 changes. Update for attribute changes all occurrences of arlo_* and
9662 * config/arm/arm-fixed.md: Update for attribute changes
9663 all occurrences of arlo_* types.
9664 * config/arm/thumb2.md: Update for attribute changes all occurrences
9666 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9667 (cortexa7_older_only): Likewise.
9668 (cortexa7_younger): Likewise.
9669 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9670 (1020alu_shift_op): Likewise.
9671 (1020alu_shift_reg_op): Likewise.
9672 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9673 (alu_shift_op): Likewise.
9674 (alu_shift_reg_op): Likewise.
9675 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9676 (11_alu_shift_op): Likewise.
9677 (11_alu_shift_reg_op): Likewise.
9678 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9679 (9_alu_shift_reg_op): Likewise.
9680 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9682 (cortex_a15_alu_shift): Likewise.
9683 (cortex_a15_alu_shift_reg): Likewise.
9684 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9686 (cortex_a5_alu_shift): Likewise.
9687 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9689 (cortex_a53_alu_shift): Likewise.
9690 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9692 (cortex_a7_alu_reg): Likewise.
9693 (cortex_a7_alu_shift): Likewise.
9694 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9696 (cortex_a8_alu_shift): Likewise.
9697 (cortex_a8_alu_shift_reg): Likewise.
9698 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9699 (cortex_a9_dp_shift): Likewise.
9700 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9702 * config/arm/cortex-r4.md
9703 (cortex_r4_alu): Update for attribute changes.
9704 (cortex_r4_mov): Likewise.
9705 (cortex_r4_alu_shift_reg): Likewise.
9706 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9707 (526_alu_shift_op): Likewise.
9708 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9709 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9710 (626te_alu_shift_op): Likewise.
9711 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9712 (726te_alu_shift_op): Likewise.
9713 (726te_alu_shift_reg_op): Likewise.
9714 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9715 (mp626_alu_shift_op): Likewise.
9716 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9717 (pj4_alu_conds): Likewise.
9718 (pj4_shift): Likewise.
9719 (pj4_shift_conds): Likewise.
9720 (pj4_alu_shift): Likewise.
9721 (pj4_alu_shift_conds): Likewise.
9722 * config/aarch64/aarch64.md: Update for attribute change
9723 all occurrences of arlo_* and shift* types.
9725 2013-09-05 Mike Stump <mikestump@comcast.net>
9727 * tree.h: Move documentation for tree_function_decl to tree-core.h
9728 with the declaration.
9730 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9733 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9734 looking for widest mode.
9736 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9738 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9740 2013-09-05 Richard Biener <rguenther@suse.de>
9742 PR tree-optimization/58137
9743 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9744 Do not create vectors of pointers.
9745 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9746 types for the components of the vector initializer.
9747 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9748 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9750 2013-09-05 Martin Jambor <mjambor@suse.cz>
9752 * ipa-prop.c (remove_described_reference): Accept missing references,
9753 return false if that hppens, otherwise return true.
9754 (cgraph_node_for_jfunc): New function.
9755 (try_decrement_rdesc_refcount): Likewise.
9756 (try_make_edge_direct_simple_call): Use them.
9757 (ipa_edge_removal_hook): Remove references from rdescs.
9758 (ipa_edge_duplication_hook): Clone rdescs and their references
9759 when the new edge has the same caller as the old one.
9760 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9761 reference before removing any edges.
9763 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9765 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9767 * thumb2.md (thumb2_storewb_parisi): New pattern.
9769 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9771 * config/aarch64/aarch64-option-extensions.def: Add
9772 AARCH64_OPT_EXTENSION of 'crc'.
9773 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9774 (AARCH64_ISA_CRC): Ditto.
9775 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9776 description of the CRC extension.
9778 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9780 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9781 * config/rs6000/linux.h: Ditto.
9782 * alpha/linux.h: Ditto.
9783 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9784 no_c99_libc_has_function.
9785 * config/c6x/uclinux-elf.h: Ditto.
9786 * config/lm32/uclinux-elf.h: Ditto.
9787 * config/m68k/uclinux.h: Ditto.
9788 * config/moxie/uclinux.h: Ditto.
9789 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9790 (crisv32-*-linux*, cris-*-linux*): Ditto.
9791 * config/bfin/bfin.c: Include "tm_p.h".
9793 2013-09-05 Richard Biener <rguenther@suse.de>
9795 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9796 check for a definition without a basic-block.
9798 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9799 Sofiane Naci <sofiane.naci@arm.com>
9801 * config/aarch64/aarch64.md
9802 (*movti_aarch64): Rename r_2_f and f_2_r.
9803 (*movsf_aarch64): Likewise.
9804 (*movdf_aarch64): Likewise.
9805 (*movtf_aarch64): Likewise.
9806 (aarch64_movdi_<mode>low): Likewise.
9807 (aarch64_movdi_<mode>high): Likewise.
9808 (aarch64_mov<mode>high_di): Likewise.
9809 (aarch64_mov<mode>low_di): Likewise.
9810 (aarch64_movtilow_tilow): Likewise.
9811 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9812 values to config/arm/types.md
9813 (attribute "conds"): Update for attribute change.
9814 (anddi3_insn): Likewise.
9815 (iordi3_insn): Likewise.
9816 (xordi3_insn): Likewise.
9817 (one_cmpldi2): Likewise.
9818 * config/arm/types.md (type): Add Neon types.
9819 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9820 use "type" attribute.
9821 (movmisalign<mode>_neon_store): Likewise.
9822 (movmisalign<mode>_neon_load): Likewise.
9823 (vec_set<mode>_internal): Likewise.
9824 (vec_setv2di_internal): Likewise.
9825 (vec_extract<mode>): Likewise.
9826 (vec_extractv2di): Likewise.
9827 (add<mode>3_neon): Likewise.
9828 (adddi3_neon): Likewise.
9829 (sub<mode>3_neon): Likewise.
9830 (subdi3_neon): Likewise.
9831 (mul<mode>3_neon): Likewise.
9832 (mul<mode>3add<mode>_neon): Likewise.
9833 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9834 (fma<VCVTF:mode>4)): Likewise.
9835 (fma<VCVTF:mode>4_intrinsic): Likewise.
9836 (fmsub<VCVTF:mode>4)): Likewise.
9837 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9838 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9839 (ior<mode>3): Likewise.
9840 (and<mode>3): Likewise.
9841 (anddi3_neon): Likewise.
9842 (orn<mode>3_neon): Likewise.
9843 (orndi3_neon): Likewise.
9844 (bic<mode>3_neon): Likewise.
9845 (bicdi3_neon): Likewise.
9846 (xor<mode>3): Likewise.
9847 (one_cmpl<mode>2): Likewise.
9848 (abs<mode>2): Likewise.
9849 (neg<mode>2): Likewise.
9850 (umin<mode>3_neon): Likewise.
9851 (umax<mode>3_neon): Likewise.
9852 (smin<mode>3_neon): Likewise.
9853 (smax<mode>3_neon): Likewise.
9854 (vashl<mode>3): Likewise.
9855 (vashr<mode>3_imm): Likewise.
9856 (vlshr<mode>3_imm): Likewise.
9857 (ashl<mode>3_signed): Likewise.
9858 (ashl<mode>3_unsigned): Likewise.
9859 (neon_load_count): Likewise.
9860 (ashldi3_neon_noclobber): Likewise.
9861 (signed_shift_di3_neon): Likewise.
9862 (unsigned_shift_di3_neon): Likewise.
9863 (ashrdi3_neon_imm_noclobber): Likewise.
9864 (lshrdi3_neon_imm_noclobber): Likewise.
9865 (widen_ssum<mode>3): Likewise.
9866 (widen_usum<mode>3): Likewise.
9867 (quad_halves_<code>v4si): Likewise.
9868 (quad_halves_<code>v4sf): Likewise.
9869 (quad_halves_<code>v8hi): Likewise.
9870 (quad_halves_<code>v16qi): Likewise.
9871 (reduc_splus_v2di): Likewise.
9872 (neon_vpadd_internal<mode>): Likewise.
9873 (neon_vpsmin<mode>): Likewise.
9874 (neon_vpsmax<mode>): Likewise.
9875 (neon_vpumin<mode>): Likewise.
9876 (neon_vpumax<mode>): Likewise.
9877 (ss_add<mode>_neon): Likewise.
9878 (us_add<mode>_neon): Likewise.
9879 (ss_sub<mode>_neon): Likewise.
9880 (us_sub<mode>_neon): Likewise.
9881 (neon_vadd<mode>_unspec): Likewise.
9882 (neon_vaddl<mode>): Likewise.
9883 (neon_vaddw<mode>): Likewise.
9884 (neon_vhadd<mode>): Likewise.
9885 (neon_vqadd<mode>): Likewise.
9886 (neon_vaddhn<mode>): Likewise.
9887 (neon_vmul<mode>): Likewise.
9888 (neon_vmla<mode>): Likewise.
9889 (neon_vmlal<mode>): Likewise.
9890 (neon_vmls<mode>): Likewise.
9891 (neon_vmlsl<mode>): Likewise.
9892 (neon_vqdmulh<mode>): Likewise.
9893 (neon_vqdmlal<mode>): Likewise.
9894 (neon_vqdmlsl<mode>): Likewise.
9895 (neon_vmull<mode>): Likewise.
9896 (neon_vqdmull<mode>): Likewise.
9897 (neon_vsub<mode>_unspec): Likewise.
9898 (neon_vsubl<mode>): Likewise.
9899 (neon_vsubw<mode>): Likewise.
9900 (neon_vqsub<mode>): Likewise.
9901 (neon_vhsub<mode>): Likewise.
9902 (neon_vsubhn<mode>): Likewise.
9903 (neon_vceq<mode>): Likewise.
9904 (neon_vcge<mode>): Likewise.
9905 (neon_vcgeu<mode>): Likewise.
9906 (neon_vcgt<mode>): Likewise.
9907 (neon_vcgtu<mode>): Likewise.
9908 (neon_vcle<mode>): Likewise.
9909 (neon_vclt<mode>): Likewise.
9910 (neon_vcage<mode>): Likewise.
9911 (neon_vcagt<mode>): Likewise.
9912 (neon_vtst<mode>): Likewise.
9913 (neon_vabd<mode>): Likewise.
9914 (neon_vabdl<mode>): Likewise.
9915 (neon_vaba<mode>): Likewise.
9916 (neon_vabal<mode>): Likewise.
9917 (neon_vmax<mode>): Likewise.
9918 (neon_vmin<mode>): Likewise.
9919 (neon_vpaddl<mode>): Likewise.
9920 (neon_vpadal<mode>): Likewise.
9921 (neon_vpmax<mode>): Likewise.
9922 (neon_vpmin<mode>): Likewise.
9923 (neon_vrecps<mode>): Likewise.
9924 (neon_vrsqrts<mode>): Likewise.
9925 (neon_vqabs<mode>): Likewise.
9926 (neon_vqneg<mode>): Likewise.
9927 (neon_vcls<mode>): Likewise.
9928 (clz<mode>2): Likewise.
9929 (popcount<mode>2): Likewise.
9930 (neon_vrecpe): Likewise.
9931 (neon_vrsqrte): Likewise.
9932 (neon_vget_lane<mode>_sext_internal): Likewise.
9933 (neon_vget_lane<mode>_zext_internal): Likewise.
9934 (neon_vdup_n<mode>): Likewise.
9935 (neon_vdup_nv2di): Likewise.
9936 (neon_vdpu_lane<mode>_internal): Likewise.
9937 (neon_vswp<mode>): Likewise.
9938 (float<mode><V_cvtto>2): Likewise.
9939 (floatuns<mode><V_cvtto>2): Likewise.
9940 (fix_trunc<mode><V_cvtto>)2): Likewise
9941 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9942 (neon_vcvt<mode>): Likewise.
9943 (neon_vcvtv4sfv4hf): Likewise.
9944 (neon_vcvtv4hfv4sf): Likewise.
9945 (neon_vcvt_n<mode>): Likewise.
9946 (neon_vmovn<mode>): Likewise.
9947 (neon_vqmovn<mode>): Likewise.
9948 (neon_vqmovun<mode>): Likewise.
9949 (neon_vmovl<mode>): Likewise.
9950 (neon_vmul_lane<mode>): Likewise.
9951 (neon_vmull_lane<mode>): Likewise.
9952 (neon_vqdmull_lane<mode>): Likewise.
9953 (neon_vqdmulh_lane<mode>): Likewise.
9954 (neon_vmla_lane<mode>): Likewise.
9955 (neon_vmlal_lane<mode>): Likewise.
9956 (neon_vqdmlal_lane<mode>): Likewise.
9957 (neon_vmls_lane<mode>): Likewise.
9958 (neon_vmlsl_lane<mode>): Likewise.
9959 (neon_vqdmlsl_lane<mode>): Likewise.
9960 (neon_vext<mode>): Likewise.
9961 (neon_vrev64<mode>): Likewise.
9962 (neon_vrev32<mode>): Likewise.
9963 (neon_vrev16<mode>): Likewise.
9964 (neon_vbsl<mode>_internal): Likewise.
9965 (neon_vshl<mode>): Likewise.
9966 (neon_vqshl<mode>): Likewise.
9967 (neon_vshr_n<mode>): Likewise.
9968 (neon_vshrn_n<mode>): Likewise.
9969 (neon_vqshrn_n<mode>): Likewise.
9970 (neon_vqshrun_n<mode>): Likewise.
9971 (neon_vshl_n<mode>): Likewise.
9972 (neon_vqshl_n<mode>): Likewise.
9973 (neon_vqshlu_n<mode>): Likewise.
9974 (neon_vshll_n<mode>): Likewise.
9975 (neon_vsra_n<mode>): Likewise.
9976 (neon_vsri_n<mode>): Likewise.
9977 (neon_vsli_n<mode>): Likewise.
9978 (neon_vtbl1v8qi): Likewise.
9979 (neon_vtbl2v8qi): Likewise.
9980 (neon_vtbl3v8qi): Likewise.
9981 (neon_vtbl4v8qi): Likewise.
9982 (neon_vtbx1v8qi): Likewise.
9983 (neon_vtbx2v8qi): Likewise.
9984 (neon_vtbx3v8qi): Likewise.
9985 (neon_vtbx4v8qi): Likewise.
9986 (neon_vtrn<mode>_internal): Likewise.
9987 (neon_vzip<mode>_internal): Likewise.
9988 (neon_vuzp<mode>_internal): Likewise.
9989 (neon_vld1<mode>): Likewise.
9990 (neon_vld1_lane<mode>): Likewise.
9991 (neon_vld1_dup<mode>): Likewise.
9992 (neon_vld1_dupv2di): Likewise.
9993 (neon_vst1<mode>): Likewise.
9994 (neon_vst1_lane<mode>): Likewise.
9995 (neon_vld2<mode>): Likewise.
9996 (neon_vld2_lane<mode>): Likewise.
9997 (neon_vld2_dup<mode>): Likewise.
9998 (neon_vst2<mode>): Likewise.
9999 (neon_vst2_lane<mode>): Likewise.
10000 (neon_vld3<mode>): Likewise.
10001 (neon_vld3qa<mode>): Likewise.
10002 (neon_vld3qb<mode>): Likewise.
10003 (neon_vld3_lane<mode>): Likewise.
10004 (neon_vld3_dup<mode>): Likewise.
10005 (neon_vst3<mode>): Likewise.
10006 (neon_vst3qa<mode>): Likewise.
10007 (neon_vst3qb<mode>): Likewise.
10008 (neon_vst3_lane<mode>): Likewise.
10009 (neon_vld4<mode>): Likewise.
10010 (neon_vld4qa<mode>): Likewise.
10011 (neon_vld4qb<mode>): Likewise.
10012 (neon_vld4_lane<mode>): Likewise.
10013 (neon_vld4_dup<mode>): Likewise.
10014 (neon_vst4<mode>): Likewise.
10015 (neon_vst4qa<mode>): Likewise.
10016 (neon_vst4qb<mode>): Likewise.
10017 (neon_vst4_lane<mode>): Likewise.
10018 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10019 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10020 (neon_vec_<US>mult_lo_<mode>): Likewise.
10021 (neon_vec_<US>mult_hi_<mode>): Likewise.
10022 (neon_vec_<US>shiftl_<mode>): Likewise.
10023 (neon_unpack<US>_<mode>): Likewise.
10024 (neon_vec_<US>mult_<mode>): Likewise.
10025 (vec_pack_trunc_<mode>): Likewise.
10026 (neon_vec_pack_trunk_<mode>): Likewise.
10027 (neon_vabd<mode>_2): Likewise.
10028 (neon_vabd<mode>_3): Likewise.
10029 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10030 (thumb2_movsi_vfp): Likewise.
10031 (movdi_vfp): Likewise.
10032 (movdi_vfp_cortexa8): Likewise.
10033 (movhf_vfp_neon): Likewise.
10034 (movhf_vfp): Likewiwse.
10035 (movsf_vfp): Likewiwse.
10036 (thumb2_movsf_vfp): Likewiwse.
10037 (movdf_vfp): Likewise.
10038 (thumb2_movdf_vfp): Likewise.
10039 (movsfcc_vfp): Likewise.
10040 (thumb2_movsfcc_vfp): Likewise.
10041 (movdfcc_vfp): Likewise.
10042 (thumb2_movdfcc_vfp): Likewise.
10043 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10044 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10045 (v10_v2c): Likewise.
10046 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10048 (cortex_a15_neon_int_2): Likewise.
10049 (cortex_a15_neon_int_3): Likewise.
10050 (cortex_a15_neon_int_4): Likewise.
10051 (cortex_a15_neon_int_5): Likewise.
10052 (cortex_a15_neon_vqneg_vqabs): Likewise.
10053 (cortex_a15_neon_vmov): Likewise.
10054 (cortex_a15_neon_vaba): Likewise.
10055 (cortex_a15_neon_vaba_qqq): Likewise.
10056 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10057 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10058 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10059 scalar_64_32_long_scalar): Likewise.
10060 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10061 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10062 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10063 lotype_qdd_64_32_long): Likewise.
10064 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10065 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10066 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10067 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10068 (cortex_a15_neon_shift_1): Likewise.
10069 (cortex_a15_neon_shift_2): Likewise.
10070 (cortex_a15_neon_shift_3): Likewise.
10071 (cortex_a15_neon_vshl_ddd): Likewise.
10072 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10073 (cortex_a15_neon_vsra_vrsra): Likewise.
10074 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10075 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10076 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10077 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10078 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10079 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10080 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10081 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10082 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10083 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10084 (cortex_a15_neon_bp_simple): Likewise.
10085 (cortex_a15_neon_bp_2cycle): Likewise.
10086 (cortex_a15_neon_bp_3cycle): Likewise.
10087 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10088 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10089 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10090 (cortex_a15_neon_vld2_4_regs): Likewise.
10091 (cortex_a15_neon_vld3_vld4): Likewise.
10092 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10093 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10094 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10095 (cortex_a15_neon_vst3_vst4): Likewise.
10096 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10097 (cortex_a15_neon_vld3_vld4_lane" 10
10098 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10099 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10100 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10101 (cortex_a15_neon_ldm_2): Likewise.0
10102 (cortex_a15_neon_stm_2): Likewise.
10103 (cortex_a15_neon_mcr): Likewise.
10104 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10105 (cortex_a15_neon_mrc): Likewise.
10106 (cortex_a15_neon_mrrc): Likewise.
10107 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10109 (cortex_a15_alu_shift): Likewise.
10110 (cortex_a15_alu_shift_reg): Likewise.
10111 (cortex_a15_mult32): Likewise.
10112 (cortex_a15_mult64): Likewise.
10113 (cortex_a15_block): Likewise.
10114 (cortex_a15_branch): Likewise.
10115 (cortex_a15_load1): Likewise.
10116 (cortex_a15_load3): Likewise.
10117 (cortex_a15_store1): Likewise.
10118 (cortex_a15_store3): Likewise.
10119 (cortex_a15_call): Likewise.
10120 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10121 (cortex_a5_f2r): Likewise.
10122 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10124 (cortex_a53_f2r): Likewise.
10125 * config/arm/cortex-a7.md
10126 (cortex_a7_branch): Update for attribute change.
10127 (cortex_a7_call): Likewise.
10128 (cortex_a7_alu_imm): Likewise.
10129 (cortex_a7_alu_reg): Likewise.
10130 (cortex_a7_alu_shift): Likewise.
10131 (cortex_a7_mul): Likewise.
10132 (cortex_a7_load1): Likewise.
10133 (cortex_a7_store1): Likewise.
10134 (cortex_a7_load2): Likewise.
10135 (cortex_a7_store2): Likewise.
10136 (cortex_a7_load3): Likewise.
10137 (cortex_a7_store3): Likewise.
10138 (cortex_a7_load4): Likewise.
10139 (cortex_a7_store4): Likewise.
10140 (cortex_a7_fpalu): Likewise.
10141 (cortex_a7_fconst): Likewise.
10142 (cortex_a7_fpmuls): Likewise.
10143 (cortex_a7_neon_mul): Likewise.
10144 (cortex_a7_fpmacs): Likewise.
10145 (cortex_a7_neon_mla: Likewise.
10146 (cortex_a7_fpmuld: Likewise.
10147 (cortex_a7_fpmacd: Likewise.
10148 (cortex_a7_fpfmad: Likewise.
10149 (cortex_a7_fdivs: Likewise.
10150 (cortex_a7_fdivd: Likewise.
10151 (cortex_a7_r2f: Likewise.
10152 (cortex_a7_f2r: Likewise.
10153 (cortex_a7_f_flags: Likewise.
10154 (cortex_a7_f_loads: Likewise.
10155 (cortex_a7_f_loadd: Likewise.
10156 (cortex_a7_f_stores: Likewise.
10157 (cortex_a7_f_stored: Likewise.
10158 (cortex_a7_neon): Likewise.
10159 * config/arm/cortex-a8-neon.md
10160 (cortex_a8_neon_mrc): Update for attribute change.
10161 (cortex_a8_neon_mrrc): Likewise.
10162 (cortex_a8_neon_int_1): Likewise.
10163 (cortex_a8_neon_int_2): Likewise.
10164 (cortex_a8_neon_int_3): Likewise.
10165 (cortex_a8_neon_int_4): Likewise.
10166 (cortex_a8_neon_int_5): Likewise.
10167 (cortex_a8_neon_vqneg_vqabs): Likewise.
10168 (cortex_a8_neon_vmov): Likewise.
10169 (cortex_a8_neon_vaba): Likewise.
10170 (cortex_a8_neon_vaba_qqq): Likewise.
10171 (cortex_a8_neon_vsma): Likewise.
10172 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10173 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10174 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10175 long_scalar): Likewise.
10176 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10177 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10178 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10179 long_scalar_qdd_64_32_long): Likewise.
10180 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10181 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10182 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10183 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10184 (cortex_a8_neon_shift_1): Likewise.
10185 (cortex_a8_neon_shift_2): Likewise.
10186 (cortex_a8_neon_shift_3): Likewise.
10187 (cortex_a8_neon_vshl_ddd): Likewise.
10188 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10189 (cortex_a8_neon_vsra_vrsra): Likewise.
10190 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10191 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10192 (cortex_a8_neon_fp_vsum): Likewise.
10193 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10194 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10195 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10196 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10197 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10198 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10199 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10200 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10201 (cortex_a8_neon_bp_simple): Likewise.
10202 (cortex_a8_neon_bp_2cycle): Likewise.
10203 (cortex_a8_neon_bp_3cycle): Likewise.
10204 (cortex_a8_neon_ldr): Likewise.
10205 (cortex_a8_neon_str): Likewise.
10206 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10207 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10208 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10209 (cortex_a8_neon_vld2_4_regs): Likewise.
10210 (cortex_a8_neon_vld3_vld4): Likewise.
10211 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10212 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10213 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10214 (cortex_a8_neon_vst3_vst4): Likewise.
10215 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10216 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10217 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10218 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10219 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10220 (cortex_a8_neon_mcr): Likewise.
10221 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10222 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10223 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10225 (ca9_neon_mrrc): Likewise.
10226 (cortex_a9_neon_int_1): Likewise.
10227 (cortex_a9_neon_int_2): Likewise.
10228 (cortex_a9_neon_int_3): Likewise.
10229 (cortex_a9_neon_int_4): Likewise.
10230 (cortex_a9_neon_int_5): Likewise.
10231 (cortex_a9_neon_vqneg_vqabs): Likewise.
10232 (cortex_a9_neon_vmov): Likewise.
10233 (cortex_a9_neon_vaba): Likewise.
10234 (cortex_a9_neon_vaba_qqq): Likewise.
10235 (cortex_a9_neon_vsma): Likewise.
10236 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10237 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10238 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10239 long_scalar): Likewise.
10240 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10241 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10242 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10243 long_scalar_qdd_64_32_long): Likewise.
10244 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10245 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10246 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10247 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10248 (cortex_a9_neon_shift_1): Likewise.
10249 (cortex_a9_neon_shift_2): Likewise.
10250 (cortex_a9_neon_shift_3): Likewise.
10251 (cortex_a9_neon_vshl_ddd): Likewise.
10252 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10253 (cortex_a9_neon_vsra_vrsra): Likewise.
10254 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10255 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10256 (cortex_a9_neon_fp_vsum): Likewise.
10257 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10258 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10259 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10260 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10261 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10262 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10263 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10264 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10265 (cortex_a9_neon_bp_simple): Likewise.
10266 (cortex_a9_neon_bp_2cycle): Likewise.
10267 (cortex_a9_neon_bp_3cycle): Likewise.
10268 (cortex_a9_neon_ldr): Likewise.
10269 (cortex_a9_neon_str): Likewise.
10270 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10271 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10272 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10273 (cortex_a9_neon_vld2_4_regs): Likewise.
10274 (cortex_a9_neon_vld3_vld4): Likewise.
10275 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10276 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10277 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10278 (cortex_a9_neon_vst3_vst4): Likewise.
10279 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10280 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10281 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10282 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10283 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10284 (cortex_a9_neon_mcr): Likewise.
10285 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10286 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10287 (cortex_a9_fps): Likewise.
10288 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
10290 (cortex_m4_fmuls): Likewise.
10291 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
10293 (cortex_r4_mrc): Likewise.
10294 * config/arm/iterators.md: Update comment referring to neon_type.
10295 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
10296 (iwmmxt_movsi_insn): Likewise.
10297 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
10299 (pj4_core_to_vfp): Likewise.
10300 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
10302 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
10303 (vfp_fstore): Likewise.
10304 * doc/md.texi: Change references to neon_type to refer to type.
10306 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10308 * tree.h (DECL_BUILT_IN): Fix typo in comment.
10310 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
10312 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
10313 lglobl if not weak.
10315 2013-09-04 Easwaran Raman <eraman@google.com>
10317 PR middle-end/57370
10318 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
10319 (build_and_add_sum): Use it.
10320 (appears_later_in_bb): Simplify code.
10322 2013-09-04 Teresa Johnson <tejohnson@google.com>
10324 * dumpfile.c (dump_finish): Don't close stderr/stdout.
10326 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
10328 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
10330 2013-09-04 Jan Hubicka <jh@suse.cz>
10332 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
10333 * ipa-devirt.c: Include diganostic.h
10334 (odr_type_d): Add types and types_set.
10335 (hash_type_name): Work for types with vtables during LTO.
10336 (odr_hasher::remove): Fix comment; destroy types_set.
10337 (add_type_duplicate): New function,
10338 (get_odr_type): Use it.
10339 (dump_type_inheritance_graph): Dump type duplicates.
10340 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
10342 * tree.c (types_same_for_odr): Give exact answers on types with
10345 2013-09-04 Dodji Seketeli <dodji@redhat.com>
10347 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
10348 explaining their differences.
10350 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
10352 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
10354 2013-09-03 Jeff Law <law@redhat.com>
10356 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
10357 when not threading through a joiner block. Pass joiner/no joiner
10358 state to register_jump_thread.
10359 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10360 state from argument rather than implying on path length.
10361 Dump the entire jump thread path into debugging dump.
10362 * tree-flow.h (register_jump_thread): Update prototype.
10364 2013-08-29 Xinliang David Li <davidxl@google.com>
10366 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10367 Remove a trivial gcc_assert.
10369 2013-08-29 Xinliang David Li <davidxl@google.com>
10371 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10372 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10373 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10374 Delay base decl alignment adjustment.
10375 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10376 * tree-vectorizer.h: New data structure.
10377 (set_dr_misalignment): New function.
10378 (dr_misalignment): Ditto.
10379 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10380 (vectorizable_load): Ditto.
10381 (ensure_base_align): New function.
10382 (vectorize_loops): Add dbg_cnt support.
10383 (execute_vect_slp): Ditto.
10384 * dbgcnt.def: New debug counter.
10385 * Makefile: New dependency.
10387 2013-09-03 Meador Inge <meadori@codesourcery.com>
10391 2013-08-30 Meador Inge <meadori@codesourcery.com>
10393 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10395 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10397 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10398 function descriptor.
10400 2013-09-03 Richard Biener <rguenther@suse.de>
10402 * tree-affine.c (add_elt_to_tree): Fix association issue,
10403 avoid useless converts and make sure to always return a
10404 properly typed result.
10406 2013-09-03 Richard Biener <rguenther@suse.de>
10408 PR middle-end/57656
10409 * fold-const.c (negate_expr_p): Fix division case.
10410 (negate_expr): Likewise.
10412 2013-09-03 Richard Biener <rguenther@suse.de>
10415 * tree-streamer-out.c: Include tm.h.
10416 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10418 2013-09-03 Jan Hubicka <jh@suse.cz>
10420 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10422 2013-09-03 Alan Modra <amodra@gmail.com>
10424 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10425 multilibs to big-endian target and vice versa.
10426 * config/rs6000/t-linux64: Use := assignment on all vars.
10427 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10428 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10429 * config/rs6000/t-linux64le: New file.
10430 * config/rs6000/t-linux64bele: New file.
10431 * config/rs6000/t-linux64lebe: New file.
10433 2013-09-02 Jan Hubicka <jh@suse.cz>
10435 * ipa-inline-transform.c (inline_transform): Do not
10436 optimize_inline_calls when not optimizing.
10438 2013-09-02 Jan Hubicka <jh@suse.cz>
10440 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10441 duplicated nodes for assembler names.
10442 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10443 hard registers from assembler name hash.
10445 2013-09-02 Jan Hubicka <jh@suse.cz>
10447 * ipa-split.c (execute_split_functions): Split externally visible
10448 functions called once.
10450 2013-09-02 Martin Jambor <mjambor@suse.cz>
10453 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10454 linked list. When finding the correct duplicate, also consider also
10455 the caller in additon to its inlined_to node.
10457 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10459 * config/aarch64/aarch64-simd-builtins.def
10460 (dup_lane_scalar): Remove.
10461 * config/aarch64/aarch64-simd.md
10462 (aarch64_simd_dup): Add 'w->w' alternative.
10463 (aarch64_dup_lane<mode>): Allow for VALL.
10464 (aarch64_dup_lane_scalar<mode>): Remove.
10465 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10466 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10467 (aarch64_get_lane_unsigned<mode>): Likewise.
10468 (aarch64_get_lane<mode>): Likewise.
10469 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10470 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10471 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10472 (VCON): Change container of V2SF.
10473 (vswap_width_name): Likewise.
10474 * config/aarch64/arm_neon.h
10475 (__aarch64_vdup_lane_any): New.
10476 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10477 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10478 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10480 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10482 PR middle-end/56382
10483 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10484 the source or the destination is a single hard register.
10486 2013-09-02 Richard Biener <rguenther@suse.de>
10488 PR middle-end/57511
10489 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10492 2013-09-02 Richard Biener <rguenther@suse.de>
10494 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10495 arithmetic to sizetype.
10497 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10499 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10500 Find auto-increment use both before and after candidate.
10502 2013-09-02 Marek Polacek <polacek@redhat.com>
10504 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10506 2013-09-01 Jan Hubicka <jh@suse.cz>
10508 * Makefile.in: Add ipa-profile.o
10509 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10510 * cgraph.c (struct cgraph_propagate_frequency_data,
10511 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10512 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10513 * cgraph.h (cgraph_propagate_frequency): Remove.
10514 * ipa-inline-analysis.c: Include ipa-utils.h;
10515 drop duplicated cfgloop.h.
10516 (inline_update_callee_summaries): Update.
10517 * ipa-profile.c: New file.
10518 * ipa-utils.h (ipa_propagate_frequency): Declare.
10519 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10520 data-streamer.h, value-prof.h.
10521 (symtab_remove_unreachable_nodes): Update profile.
10522 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10523 account_time_size, cmp_counts, dump_histogram,
10524 ipa_profile_generate_summary, ipa_profile_write_summary,
10525 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10526 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10527 Move to ipa-profile.c.
10529 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10531 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10533 2013-09-01 Jan Hubicka <jh@suse.cz>
10535 * common.opt (fdevirtualize-speculatively): New function.
10536 * invoke.texi (fdevirtualize-speculatively): Document.
10537 * ipa-devirt.c: Include ipa-inline.h
10538 (likely_target_p): New function.
10539 (ipa_devirt): New function.
10540 (gate_ipa_devirt): New function.
10541 (pass_data_ipa_devirt): New static var.
10542 (pass_ipa_devirt): Likewise.
10543 (make_pass_ipa_devirt): New function.
10544 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10545 (common_handle_option): Disable devirtualization when
10546 value range profiling is available.
10547 * passes.def (pass_ipa_devirt): Add.
10548 * timever.def (TV_IPA_DEVIRT): New timevar.
10549 * tree-pass.h (make_pass_ipa_devirt):
10551 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10553 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10554 include sanitize(undefined).
10556 2013-08-31 Diego Novillo <dnovillo@google.com>
10558 * Makefile.in (TREE_CORE_H): Define.
10560 (GTFILES): Add tree-core.h.
10561 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10563 * tree-core.h: New file.
10564 Move all data structures, enum, typedefs, global
10565 declarations and constants from ...
10566 * tree.h: ... here.
10568 2013-08-31 Jan Hubicka <jh@suse.cz>
10570 * bulitins.c (expand_builtin): Do not early exit for gcov
10571 instrumented functions.
10573 2013-08-31 Marek Polacek <polacek@redhat.com>
10575 * ubsan.c: Include tm_p.h.
10577 2013-08-31 Jan Hubicka <jh@suse.cz>
10579 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10582 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10583 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10584 * tree-cfg.c (verify_gimple_label): ... here.
10585 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10586 (ipa_merge_profiles): New function.
10587 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10588 (lto_input_function_body): Likewise.
10589 * ipa-utils.h (ipa_merge_profiles): Declare.
10590 * lto-streamer.h (lto_input_function_body): Update prototype.
10591 (emit_label_in_global_context_p): Remove.
10592 * lto-symtab.c: Include ipa-utils.h
10593 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10595 2013-08-31 Jan Hubicka <jh@suse.cz>
10597 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10599 2013-08-31 Jan Hubicka <jh@suse.cz>
10601 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10603 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10605 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10606 "cmp" RTX before signed_comparison_operator check to account
10607 for "code" changes.
10609 2013-08-30 Jan Hubicka <jh@suse.cz>
10611 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10612 (detect_type_change_1): Rename to ...
10613 (detect_type_change): ... this one; early return on non-polymorphic
10615 (detect_type_change_ssa): Add comp_type parameter; update
10616 use of detect_type_change.
10617 (compute_complex_assign_jump_func): Add param_type parameter;
10618 update use of detect_type_change_ssa.
10619 (compute_complex_ancestor_jump_func): Likewise.
10620 (ipa_get_callee_param_type): New function.
10621 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10622 update calls to the jump function computation functions.
10624 2013-08-30 Teresa Johnson <tejohnson@google.com>
10625 Steven Bosscher <steven@gcc.gnu.org>
10627 * cfgrtl.c (fixup_new_cold_bb): New routine.
10628 (commit_edge_insertions): Invoke fixup_partitions.
10629 (find_partition_fixes): New routine.
10630 (fixup_partitions): Ditto.
10631 (verify_hot_cold_block_grouping): Update comments.
10632 (rtl_verify_edges): Invoke find_partition_fixes.
10633 (rtl_verify_bb_pointers): Update comments.
10634 (rtl_verify_bb_layout): Ditto.
10635 * basic-block.h (probably_never_executed_edge_p): Declare.
10636 (fixup_partitions): Ditto.
10637 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10638 * bb-reorder.c (sanitize_hot_paths): New function.
10639 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10640 sanitize_hot_paths.
10641 * predict.c (probably_never_executed_edge_p): New routine.
10642 * cfg.c (check_bb_profile): Add partition insanity warnings.
10644 2013-08-30 Meador Inge <meadori@codesourcery.com>
10646 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10648 2013-08-30 Marek Polacek <polacek@redhat.com>
10650 * Makefile.in (ubsan.o): Add.
10651 (c-family/c-ubsan.o): Add.
10652 (builtins.o): Add ubsan.h dependency.
10653 * ubsan.h: New file.
10654 * ubsan.c: New file.
10655 * common.opt: Add -fsanitize=undefined option.
10656 (flag_sanitize): Add variable.
10657 (fsanitize=): Add option. Add Driver.
10658 (fsanitize=thread): Remove option.
10659 (fsanitize=address): Likewise.
10660 (static-libubsan): New option.
10661 * doc/invoke.texi: Document the new flag and -static-libubsan.
10662 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10663 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10664 * builtin-attrs.def (ATTR_COLD): Define.
10665 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10666 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10667 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10668 * flag-types.h (sanitize_code): New enum.
10669 * opts.c (common_handle_option): Parse command line arguments
10670 of -fsanitize=. Add -fsanitize=unreachable option.
10671 * varasm.c (get_variable_section): Adjust.
10672 (assemble_noswitch_variable): Likewise.
10673 (assemble_variable): Likewise.
10674 (output_constant_def_contents): Likewise.
10675 (categorize_decl_for_section): Likewise.
10676 (place_block_symbol): Likewise.
10677 (output_object_block): Likewise.
10678 * builtins.def: Likewise.
10679 * toplev.c (compile_file): Likewise.
10680 (process_options): Likewise.
10681 * cppbuiltin.c: Likewise.
10682 * tsan.c (tsan_pass): Likewise.
10683 (tsan_gate): Likewise.
10684 (tsan_gate_O0): Likewise.
10685 * cfgexpand.c (partition_stack_vars): Likewise.
10686 (expand_stack_vars): Likewise.
10687 (defer_stack_allocation): Likewise.
10688 (expand_used_vars): Likewise.
10689 * cfgcleanup.c (old_insns_match_p): Likewise.
10690 * asan.c (asan_finish_file): Likewise.
10691 (asan_instrument): Likewise.
10692 (gate_asan): Likewise.
10693 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10694 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10695 (asan_global_struct): Use pointer_sized_int_node instead
10696 calling build_nonstandard_integer_type.
10697 (initialize_sanitizer_builtins): Likewise.
10698 (asan_finish_file): Likewise.
10699 * gcc.c: Document %{%:function(args):X}.
10700 (static_spec_functions): Add sanitize.
10701 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10702 store funcval != NULL there.
10703 (do_spec_1): Adjust handle_spec_function caller.
10704 (handle_braces): Allow %:function(args) as condition.
10705 (sanitize_spec_function): New function.
10706 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10707 (LIBUBSAN_SPEC): Likewise.
10708 (LIBUBSAN_EARLY_SPEC): Likewise.
10709 (SANITIZER_SPEC): Handle libubsan.
10710 (SANITIZER_EARLY_SPEC): Likewise.
10711 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10712 instead of fsanitize=address.
10713 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10714 instead of fsanitize=address*.
10715 * builtins.c: Include ubsan.h.
10716 (fold_builtin_0): Instrument __builtin_unreachable.
10717 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10718 instead of flag_asan.
10719 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10720 (pointer_sized_int_node): Define.
10721 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10723 2013-08-30 Mike Stump <mikestump@comcast.net>
10725 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10728 2013-08-29 Jan Hubicka <jh@suse.cz>
10730 * cgraph.c (cgraph_function_body_availability): Handle weakref
10732 * passes.def: Remove pass_fixup_cfg.
10733 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10734 track when we need to remove functions.
10735 (gate_ipa_inline): Execute inlining always; add comment why.
10736 (pass_data_ipa_inline): Remove TODO_remove_functions.
10737 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10738 do not produce summaries.
10739 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10740 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10741 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10742 constant pool and vtable.
10744 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10746 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10747 New arm_neon.h's internal macros to specify 64-bit constants.
10748 Avoid using stdint.h's macros.
10750 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10752 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10753 before applying REG_P.
10755 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10757 PR tree-optimization/58277
10758 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10759 after seeing too many stmts with vdef in between dombb and current
10760 bb, invalidate everything.
10762 2013-08-30 Richard Biener <rguenther@suse.de>
10764 * fold-const.c (fold_single_bit_test): Fix overflow test.
10766 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10768 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10769 and which can live in a register, always retrieve the value on entry.
10770 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10771 passed by invisible reference specially.
10772 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10773 (vt_add_function_parameter): Correctly deal with a parameter passed by
10774 invisible reference.
10776 2013-08-30 Jan Hubicka <jh@suse.cz>
10778 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10780 2013-08-30 Richard Biener <rguenther@suse.de>
10782 PR tree-optimization/58228
10783 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10784 allow invariant loads in nested loop vectorization.
10786 2013-08-30 Richard Biener <rguenther@suse.de>
10788 PR tree-optimization/58223
10789 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10790 (has_anti_or_output_dependence): ... this and adjust to also
10791 look for output dependences.
10792 (mark_nodes_having_upstream_mem_writes): Adjust.
10793 (rdg_flag_uses): Likewise.
10795 2013-08-30 Richard Biener <rguenther@suse.de>
10797 PR tree-optimization/58010
10798 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10799 assert that we have a loop-closed PHI.
10801 2013-08-29 Jan Hubicka <jh@suse.cz>
10803 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10804 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10805 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10806 (lto_free_function_in_decl_state_for_node): New function.
10808 2013-08-29 Xinliang David Li <davidxl@google.com>
10810 * loop-unroll.c (report_unroll_peel): Minor message change.
10811 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10812 Emit alignment peeling message with default -fopt-info.
10813 (vect_loop_versioning): Emit loop version info message.
10814 * tree-vectorizer.c (vectorize_loops): Minor message change.
10815 (execute_vect_slp): Ditto.
10817 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10819 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10820 of the clone from the DECL_NAME of the original function.
10822 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10824 * passes.c (register_pass): Add overload.
10825 * tree-pass.h (register_pass): Forward declare it. Add comment.
10827 2013-08-29 Jan Hubicka <jh@suse.cz>
10829 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10830 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10831 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10832 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10833 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10834 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10835 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10836 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10838 2013-08-29 Teresa Johnson <tejohnson@google.com>
10840 * dumpfile.c (dump_loc): Output column number.
10841 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10842 * doc/invoke.texi: Document optall -fopt-info flag.
10843 * profile.c (read_profile_edge_counts): Use new dump framework.
10844 (compute_branch_probabilities): Ditto.
10845 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10846 when pass not in any opt group.
10847 * pass_manager.h (pass_manager::get_pass_profile): New method.
10848 * value-prof.c (check_counter): Use new dump framework.
10849 (check_ic_target): Ditto.
10850 * coverage.c (get_coverage_counts): Ditto.
10851 (coverage_init): Setup new dump framework.
10853 2013-08-29 Richard Biener <rguenther@suse.de>
10855 PR tree-optimization/58246
10856 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10857 handle the dominance check inside a basic-block.
10859 2013-08-29 Richard Biener <rguenther@suse.de>
10861 PR middle-end/57287
10862 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10863 of default defs that appear in abnormal PHI nodes.
10865 2013-08-29 Richard Biener <rguenther@suse.de>
10867 PR tree-optimization/57685
10868 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10869 single-use operands to avoid exponential complexity.
10871 2013-08-28 Dehao Chen <dehao@google.com>
10873 * ipa-inline.c (edge_badness): Fix integer underflow.
10875 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10877 * gtm-builtins.def (_ITM_free): Declare leaf.
10879 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10882 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10883 (*tls_local_dynamic_base_64_largepic): Likewise.
10884 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10885 Remove predicate from call operand.
10886 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10887 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10889 2013-08-28 Jeff Law <law@redhat.com>
10891 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10892 checks for the number of predecessors and successors allowed.
10893 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10894 which require copying a joiner block if there is a request which
10895 is a subpath that requires no joiner block copying.
10897 2013-08-28 Jan Hubicka <jh@suse.cz>
10899 * lto-streamer-out.c (DFS_write_tree_body): Drop
10900 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10901 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10902 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10903 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10904 Do not read DECL_ERROR_ISSUED.
10905 (unpack_ts_decl_with_vis_value_fields): Do not read
10907 (lto_input_ts_binfo_tree_pointers): Do not read
10908 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10909 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10910 write DECL_ERROR_ISSUED..
10911 (pack_ts_decl_with_vis_value_fields): Do not write
10913 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10914 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10915 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10916 * tree.h (tree_decl_common): Update comment.
10917 (DECL_ERROR_ISSUED): Remove.
10919 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10921 PR middle-end/58257
10922 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10924 2013-08-28 Jan Hubicka <jh@suse.cz>
10926 * builtins.def (free): Declare leaf.
10928 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10930 * gdbhooks.py: New.
10931 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10932 * configure: Regenerate.
10934 2013-08-27 Martin Jambor <mjambor@suse.cz>
10936 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10937 (ipa_ancestor_jf_data): Likewise.
10938 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10939 (ipa_get_jf_pass_through_type_preserved): New function.
10940 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10941 (ipa_get_jf_ancestor_type_preserved): New function.
10942 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10943 (ipa_get_jf_ancestor_result): Likewise.
10944 (propagate_vals_accross_pass_through): Use
10945 ipa_get_jf_pass_through_result to do all the value mappings.
10946 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10947 type_preserved flag.
10948 (ipa_set_jf_cst_copy): New function.
10949 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10950 (ipa_set_jf_arith_pass_through): Likewise.
10951 (ipa_set_ancestor_jf): Likewise.
10952 (compute_complex_assign_jump_func): Set type_preserved instead of
10954 (ipa_compute_jump_functions_for_edge): Likewise.
10955 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10956 (update_jump_functions_after_inlining): Update type_preserved.
10957 Explicitely create jump functions when combining one with pass_through.
10958 (ipa_write_jump_function): Stream the type_preserved flags.
10959 (ipa_read_jump_function): Likewise.
10961 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10962 Aldy Hernandez <aldyh@redhat.com>
10964 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10965 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10966 * function.h (struct function): Add has_force_vect_loops and
10968 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10969 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10971 * gimple.def: Update CLAUSES comments.
10972 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10973 (gimple_build_omp_for): Add argument to prototype.
10974 (gimple_omp_for_kind): New.
10975 (gimple_omp_for_set_kind): New.
10976 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10977 GOVD_DATA_SHARE_CLASS.
10978 (enum omp_region_type): Add ORT_SIMD.
10979 (gimple_add_tmp_var): Handle ORT_SIMD.
10980 (gimplify_var_or_parm_decl): Same.
10981 (is_gimple_stmt): Same.
10982 (omp_firstprivatize_variable): Same.
10983 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10984 (omp_notice_variable): Handle ORT_SIMD.
10985 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10986 (omp_check_private): Handle ORT_SIMD.
10987 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10988 OMP_CLAUSE_SAFELEN.
10989 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10990 Handle OMP_CLAUSE_LASTPRIVATE.
10991 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10992 OMP_CLAUSE_SAFELEN.
10993 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10994 (gimplify_expr): Handle OMP_SIMD.
10995 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10996 (expand_GOMP_SIMD_VF): New.
10997 (expand_GOMP_SIMD_LAST_LANE): New.
10998 * internal-fn.def (GOMP_SIMD_LANE): New.
10999 (GOMP_SIMD_VF): New.
11000 (GOMP_SIMD_LAST_LANE): New.
11001 * omp-low.c: Include target.h.
11002 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11003 OMP_CLAUSE_SAFELEN.
11004 (check_omp_nesting_restrictions): Same.
11006 (lower_rec_simd_input_clauses): New.
11007 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11009 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11010 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11011 (expand_omp_build_assign): New.
11012 (expand_omp_for_init_counts): New.
11013 (expand_omp_for_init_vars): New.
11014 (extract_omp_for_update_vars): New.
11015 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11016 and rewrite accordingly.
11017 (expand_omp_simd): New.
11018 (expand_omp_for): Use expand_omp_simd.
11019 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11020 (lower_omp_for): Do not lower the body.
11021 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11022 in their own loops.
11023 * tree-flow.h (find_omp_clause): Remove prototype.
11024 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11025 forcing vectorization of the loop, or if flag_tree_vectorize.
11026 (gate_tree_if_conversion): Similarly.
11027 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11028 gimple_build_omp_for.
11029 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11030 * tree-parloops (create_parallel_loop): Pass kind argument to
11031 gimple_build_omp_for.
11032 * tree-pretty-print.c (dump_omp_clause): Add cases for
11033 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11034 OMP_CLAUSE__SIMDUID_.
11035 (dump_generic_node): Handle OMP_SIMD.
11036 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11037 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11038 unroll OMP_SIMD loops here.
11039 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11040 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11042 (vect_analyze_data_refs): Handle simd loops.
11043 * tree-vect-loop.c (vectorizable_live_operation): Handle
11045 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11046 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11047 (vectorizable_load): Same.
11048 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11049 (struct simduid_to_vf): New.
11050 (simduid_to_vf::hash): New.
11051 (simduid_to-vf::equal): New.
11052 (struct simd_array_to_simduid): New.
11053 (simd_array_to_simduid::hash): New.
11054 (simd_array_to_simduid::equal): New.
11055 (adjust_simduid_builtins): New.
11056 (struct note_simd_array_uses_struct): New.
11057 (note_simd_array_uses_cb): New.
11058 (note_simd_array_uses): New.
11059 (vectorize_loops): Handle simd hints and adjust simd builtins
11061 * tree-vectorizer.h (struct _stmt_vec_info): Add
11062 simd_lane_access_p field.
11063 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11064 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11065 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11066 (omp_clause_code_name): Same.
11067 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11068 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11069 * tree.def (OMP_SIMD): New entry.
11070 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11071 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11072 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11073 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11074 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11075 (OMP_CLAUSE_LINEAR_STEP): New.
11076 (OMP_CLAUSE_SAFELEN_EXPR): New.
11077 (OMP_CLAUSE__SIMDUID__DECL): New.
11078 (find_omp_clause): New prototype.
11080 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11082 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11083 Haswell processor detection.
11085 2013-08-27 Christian Widmer <shadow@umbrox.de>
11088 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11089 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11090 AVX2 capable processors.
11092 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11094 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11095 of vget_low_* with implementations in terms of other intrinsics.
11097 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11099 PR middle-end/57219
11100 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11101 values to -1, 0 and 1.
11103 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11105 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11106 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11107 (<optab><mode>3_insn): Remove.
11108 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11109 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11110 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11111 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11112 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11113 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11114 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11115 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11116 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11117 (ror<mode>3_insn): Likewise.
11118 * config/aarch64/predicates.md (aarch64_simd_register): New.
11120 2013-08-27 Richard Biener <rguenther@suse.de>
11122 PR tree-optimization/57521
11123 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11124 one edge is non-critical.
11125 (find_phi_replacement_condition): Make sure to use a non-critical
11126 edge. Cleanup and remove old bug workarounds.
11127 (bb_postdominates_preds): Remove.
11128 (if_convertible_loop_p_1): Do not compute post-dominators.
11129 (combine_blocks): Do not free post-dominators.
11130 (main_tree_if_conversion): Likewise.
11131 (pass_data_if_conversion): Add TODO_verify_ssa.
11133 2013-08-27 DJ Delorie <dj@redhat.com>
11135 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11137 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11139 * function.c (assign_parm_find_data_types): Set passed_mode and
11140 nominal_mode to the TYPE_MODE of nominal_type for the built
11141 pointer type in case of the struct-pass-by-reference.
11143 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11145 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11146 (UINT16_TYPE): Change default to "unsigned int".
11148 * config/avr/avr.opt (mfract-convert-truncate): New option.
11149 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11150 is set, round negative fractional integers according to n1169
11151 when converting to integer types.
11153 2013-08-26 Jan Hubicka <jh@suse.cz>
11155 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11156 methods can not be called indirectly when their address is not taken.
11158 2013-08-26 Jan Hubicka <jh@suse.cz>
11160 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11163 2013-08-26 Jan Hubicka <jh@suse.cz>
11165 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11168 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11170 * reload.c (find_valid_class): Allow classes that do not include
11171 FIRST_PSEUDO_REGISTER - 1.
11173 2013-08-26 Jan Hubicka <jh@suse.cz>
11175 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11176 fix edge count/frequency when speculation failed; fix type check
11177 for the direct call.
11179 2013-08-26 Jan Hubicka <jh@suse.cz>
11181 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11183 2013-08-26 Jan Hubicka <jh@suse.cz>
11185 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11186 to be changed by edge redirection.
11188 2013-08-26 Jan Hubicka <jh@suse.cz>
11190 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11191 formating; add sanity check.
11192 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11193 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11194 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11195 (resolve_noninline_speculation): Update callee keys, too.
11197 2013-08-26 Jan Hubicka <jh@suse.cz>
11199 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11200 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11202 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11204 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11207 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11209 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11210 (x86_64_elf_select_section): Ditto.
11211 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11213 (ix86_pass_by_reference): Ditto.
11214 (output_set_got): Ditto.
11215 (ix86_unary_operator_ok): Ditto.
11216 (ix86_expand_builtin): Ditto.
11218 2013-08-23 Jan Hubicka <jh@suse.cz>
11220 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11222 2013-08-23 Jan Hubicka <jh@suse.cz>
11224 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11225 (tree_decl_with_vis): Add FINAL field.
11227 2013-08-23 Jeff Law <law@redhat.com>
11229 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11230 the debugging dump when the expression is fully redundant.
11232 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11234 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11235 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11236 * pretty-print.c (pp_formatted_text_data): Likewise.
11237 (pp_write_text_to_stream): Likewise.
11238 (pp_write_text_as_dot_label_to_stream): Likewise.
11239 (pp_append_r): Likewise.
11240 (pp_format): Likewise.
11241 (pp_flush): Likewise.
11242 (pp_clear_output_area): Likewise.
11243 (pp_append_text): Likewise.
11244 (pp_formatted_text): Likewise.
11245 (pp_remaining_character_count_for_line): Likewise.
11246 (pp_newline): Likewise.
11247 (pp_character): Likewise.
11248 (output_buffer::~output_buffer): Define.
11249 (pretty_printer::~pretty_printer): Destruct output buffer.
11250 * pretty-print.h (output_buffer::~output_buffer): Declare.
11251 (pretty_printer::~pretty_printer): Declare virtual.
11253 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11256 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11257 HOST_WIDE_INT_M1U): New macros.
11258 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11259 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11260 unsigned -1 for lshift.
11261 * cse.c (cse_insn): Likewise.
11262 * double-int.c (rshift_double, lshift_double): Likewise.
11263 * builtins.c (fold_builtin_bitop): Likewise.
11264 * combine.c (force_to_mode): Likewise.
11265 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11266 * simplify-rtx.c (simplify_const_unary_operation,
11267 simplify_const_binary_operation): Likewise.
11268 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11269 check_va_list_escapes): Likewise.
11270 * rtlanal.c (nonzero_bits1): Likewise.
11271 * expmed.c (expand_smod_pow2): Likewise.
11272 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11274 2013-08-23 Jan Hubicka <jh@suse.cz>
11276 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11277 as having address taken.
11279 2013-08-23 Jan Hubicka <jh@suse.cz>
11281 * ipa-utils.h (method_class_type): Declare.
11282 * ipa-devirt.c (method_class_type): Export.
11284 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11285 do not walk base classes of anonymous types.
11287 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
11289 PR rtl-optimization/58220
11290 PR regression/58221
11291 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
11292 handle SEQUENCE insns properly.
11294 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11296 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
11298 (pp_newline_and_indent): Likewise.
11299 (pp_separate_with): Likewise.
11300 * pretty-print.c (pp_newline_and_flush): Define.
11301 (pp_newline_and_indent): Likewise.
11302 (pp_separate_with): Likewise.
11304 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11307 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
11308 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
11310 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
11312 * config/i386/predicates.md (ext_sse_reg_operand): New.
11313 * config/i386/i386.md (*movti_internal): Use
11314 predicate to determine if EVEX is needed.
11315 (*movsi_internal): Ditto.
11316 (*movdf_internal): Ditto.
11317 (*movsf_internal): Ditto.
11318 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
11320 2013-08-23 Jakub Jelinek <jakub@redhat.com>
11322 PR tree-optimization/58209
11323 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
11324 (find_tail_calls): Give up for pointer result types if m is non-NULL.
11325 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
11326 emit POINTER_PLUS_EXPR.
11327 (create_tailcall_accumulator): For pointer result type accumulate in
11330 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
11332 * configure.ac: Add backslashes missing from the last change.
11333 * configure: Regenerate.
11335 2013-08-22 Jan Hubicka <jh@susue.cz>
11337 * ipa.c (function_and_variable_visibility): First remember function
11338 was global and then make it local.
11340 2013-08-22 Julian Brown <julian@codesourcery.com>
11342 * configure.ac: Add aarch64 to list of arches which use "nop" in
11344 * configure: Regenerate.
11346 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11348 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
11349 gnu_libc_has_function.
11350 * config/s390/tpf.h: Likewise.
11352 2013-08-22 Jan Hubicka <jh@susue.cz>
11354 * timevar.c (validate_phases): Add cast.
11356 2013-08-22 Jan Hubicka <jh@susue.cz>
11358 * timevar.c (validate_phases): Use size_t for memory.
11359 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11361 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11363 * pretty-print.h (output_buffer::output_buffer): Declare.
11364 (pretty_printer::pretty_printer): Likewise.
11365 (pp_construct): Remove.
11366 * pretty-print.c (output_buffer::output_buffer): Define.
11367 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11368 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11369 (print_gimple_expr): Likewise.
11370 (print_gimple_seq): Likewise.
11371 (gimple_dump_bb): Likewise.
11372 * sched-vis.c (dump_value_slim): Likewise.
11373 (dump_insn_slim): Likewise.
11374 (dump_rtl_slim): Likewise.
11375 (str_pattern_slim): Likewise.
11376 * tree-mudflap.c (mf_varname_tree): Likewise.
11377 * graph.c (print_graph_cfg): Likewise.
11378 (start_graph_dump): Likewise.
11379 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11381 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11382 pretty printer initialization.
11383 * coretypes.h (diagnostic_context): Remove superflous type alias
11385 (pretty_printer): Likewise. Declare directly as a class.
11386 (pretty_print_info): Remove declaration as class.
11387 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11388 and pp_clear_output_area.
11389 (asan_add_global): Likewise.
11391 2013-08-22 Jan Hubicka <jh@suse.cz>
11393 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11394 * ipa-utils.h (update_type_inheritance_graph): Declare.
11395 (possible_polymorphic_call_target_p): Declare.
11396 (possible_polymorphic_call_target_p): New.
11397 * ipa-devirt.c: Update toplevel comments.
11398 (cached_polymorphic_call_targets): Move up.
11399 (odr_type_d): Move ID down.
11400 (polymorphic_type_binfo_p): Update comment.
11401 (odr_hasher::remove): Likewise;
11402 (get_odr_type): Set anonymous_namespace.
11403 (dump_odr_type): Dump it.
11404 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11405 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11406 (record_binfo): Add comment.
11407 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11409 (devirt_node_removal_hook): Do not iCE when cache is freed.
11410 (possible_polymorphic_call_target_p): New predicate.
11411 (update_type_inheritance_graph): New function.
11413 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11414 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11415 Sergey Lega <sergey.s.lega@intel.com>
11416 Anna Tikhonova <anna.tikhonova@intel.com>
11417 Ilya Tocar <ilya.tocar@intel.com>
11418 Andrey Turetskiy <andrey.turetskiy@intel.com>
11419 Ilya Verbin <ilya.verbin@intel.com>
11420 Kirill Yukhin <kirill.yukhin@intel.com>
11421 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11423 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11424 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11425 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11426 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11427 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11428 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11429 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11430 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11431 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11432 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11433 OPT_mavx512pf, OPT_mavx512er cases.
11434 * config/i386/constraints.md (v): New constraint.
11435 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11436 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11437 (bit_AVX512CD): New.
11438 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11439 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11440 * config/i386/i386-c.c (ix86_target_macros_internal):
11441 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11443 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11444 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11445 * config/i386/i386.c (regclass_map, dbx_register_map)
11446 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11447 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11448 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11449 -mavx512pf options.
11450 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11451 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11452 -mavx512pf options. Fix formatting.
11453 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11454 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11455 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11456 -mavx512cd, -mavx512pf options.
11457 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11458 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11459 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11461 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11462 (ix86_set_reg_reg_cost): Ditto.
11463 (x86_order_regs_for_local_alloc): Ditto.
11464 (MAX_VECT_LEN): Extend to 64-byte.
11465 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11466 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11467 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11468 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11469 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11470 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11471 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11472 (SSE_REG_MODE_P): Support new modes.
11473 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11474 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11475 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11476 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11477 (REG_CLASS_CONTENTS): Add new registers.
11478 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11479 (EXT_REX_SSE_REGNO_P): New.
11480 (HI_REGISTER_NAMES): Add new registers.
11481 * config/i386/i386.md: Define constants for new registers.
11482 (mode): Add new 512-bit modes.
11483 (prefix): Support evex prefix.
11484 (isa): Support avx512f, noavx512f, fma_avx512f.
11485 (ssemodesuffix): Add new 512-bit modes.
11487 (*movxi_internal_avx512f): Ditto.
11488 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11490 (*movsi_internal): Likewise.
11491 (*movdf_internal): Likewise.
11492 (*movsf_internal): Likewise.
11493 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11494 (<code><mode>3): Likewise.
11495 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11497 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11498 with the new constraint "v".
11499 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11501 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11502 with the new constraint "v".
11503 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11504 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11505 (*<plusminus_insn><mode>3): Likewise.
11506 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11507 (*mul<mode>3): Likewise.
11508 (<sse>_vmmul<mode>3): Likewise.
11509 (<sse>_div<mode>3): Likewise.
11510 (<sse>_vmdiv<mode>3): Likewise.
11511 (<sse>_sqrt<mode>2): Likewise.
11512 (<sse>_vmsqrt<mode>2): Likewise.
11513 (*<code><mode>3_finite): Likewise.
11514 (*<code><mode>3) <smaxmin>: Likewise.
11515 (<sse>_vm<code><mode>3): Likewise.
11516 (*<code><mode>3) <any_logic>: Likewise.
11517 (*fma_fmadd_<mode>): Likewise.
11518 (*fma_fmsub_<mode>): Likewise.
11519 (*fma_fnmadd_<mode>): Likewise.
11520 (*fma_fnmsub_<mode>): Likewise.
11521 (*fma_fmaddsub_<mode>): Likewise.
11522 (*fma_fmsubadd_<mode>): Likewise.
11523 (*fmai_fmadd_<mode>): Likewise.
11524 (*fmai_fmsub_<mode>): Likewise.
11525 (*fmai_fnmadd_<mode>): Likewise.
11526 (*fmai_fnmsub_<mode>): Likewise.
11527 (sse_cvtsi2ss): Likewise.
11528 (sse_cvtsi2ssq): Likewise.
11529 (sse_cvtss2si): Likewise.
11530 (sse_cvtss2si_2): Likewise.
11531 (sse_cvtss2siq): Likewise.
11532 (sse_cvtss2siq_2): Likewise.
11533 (sse_cvttss2si): Likewise.
11534 (sse_cvtss2siq_2): Likewise.
11535 (float<sseintvecmodelower><mode>2): Likewise.
11536 (sse2_cvtsd2si_2): Likewise.
11537 (sse2_cvtsd2siq_2): Likewise.
11538 (*<plusminus_insn><mode>3): Likewise.
11539 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11540 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11541 (ashr<mode>3): Likewise.
11542 (<shift_insn><mode>3): Likewise.
11543 (avx2_<code><mode>3): Likewise.
11544 (*avx2_<code><mode>3): Likewise.
11545 (*andnot<mode>3): Likewise.
11546 (*<code><mode>3) <any_logic>: Likewise.
11547 (abs<mode>2): Likewise.
11548 (avx2_permvar<mode>): Likewise.
11549 (avx2_perm<mode>_1): Likewise.
11550 (*avx_vpermilp<mode>): Likewise.
11551 (avx_vpermilvar<mode>3): Likewise.
11552 (avx2_ashrv<mode>): Likewise.
11553 (avx2_<shift_insn>v<mode>): Likewise.
11554 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11556 * doc/rtl.texi: Document XImode.
11558 2013-08-21 Jeff Law <law@redhat.com>
11560 * tree-flow.h (register_jump_thread): Pass vector of edges
11561 instead of each important edge.
11562 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11563 thread path into a vector and pass that to register_jump_thread.
11564 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11565 passed in edge vector to the current 3-edge form.
11568 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11570 * dce.c (fini_dce): Call df_analyze again just in case
11571 delete_unmarked_insns removed anything.
11573 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11575 * reload.h (struct reg_equivs): Rename to ..
11576 (struct reg_equivs_s): .. this.
11578 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11580 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11582 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11584 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11586 2013-08-21 Jeff Law <law@redhat.com>
11588 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11589 simplify assignments too. If the RHS collapses to a singleton
11590 range, then return the value for the range.
11592 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11594 * config/i386/sse.md (V16): Rename to...
11596 (mov<mode>): Update iterator name.
11597 (*mov<mode>_internal): Ditto.
11598 (push<mode>1): Ditto.
11599 (movmisalign<mode>): Ditto.
11601 2013-08-20 Jan Hubicka <jh@suse.cz>
11604 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11605 entry for direct edges.
11606 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11608 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11610 Revert my last two changes, r201865 and r201864:
11613 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11615 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11616 instances can own GC refs.
11618 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11619 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11620 (gcc::context::gt_pch_nx): Likewise.
11621 (gcc::context::gt_pch_nx): Likewise.
11622 * ggc.h (gt_ggc_mx <T>): New.
11623 (gt_pch_nx_with_op <T>): New.
11624 (gt_pch_nx <T>): New.
11625 * passes.c (opt_pass::gt_ggc_mx): New.
11626 (opt_pass::gt_pch_nx): New.
11627 (opt_pass::gt_pch_nx_with_op): New.
11628 (pass_manager::gt_ggc_mx): New.
11629 (pass_manager::gt_pch_nx): New.
11630 (pass_manager::gt_pch_nx_with_op): New.
11631 (pass_manager::operator new): Use
11632 ggc_internal_cleared_alloc_stat rather than xcalloc.
11633 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11634 (pass_manager::gt_ggc_mx): New.
11635 (pass_manager::gt_pch_nx): New.
11636 (pass_manager::gt_pch_nx_with_op): New.
11637 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11638 (opt_pass::operator new): New.
11639 (opt_pass::gt_ggc_mx): New.
11640 (opt_pass::gt_pch_nx): New.
11641 (opt_pass::gt_pch_nx_with_op): New.
11644 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11646 * Makefile.in (GTFILES): Add context.h.
11647 * context.c (gcc::context::operator new): New.
11648 (gcc::context::gt_ggc_mx): New.
11649 (gcc::context::gt_pch_nx): New.
11650 (gcc::context::gt_pch_nx): New.
11651 * context.h (gcc::context): Add GTY((user)) marking.
11652 (gcc::context::operator new): New.
11653 (gcc::context::gt_ggc_mx): New.
11654 (gcc::context::gt_pch_nx): New.
11655 (gcc::context::gt_pch_nx): New.
11656 (g): Add GTY marking.
11657 (gt_ggc_mx (gcc::context *)): New.
11658 (gt_pch_nx (gcc::context *)): New.
11659 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11660 void *cookie)): New.
11661 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11663 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11665 * dce.c (fini_dce): Call df_analyze again just in case
11666 delete_unmarked_insns removed anything.
11668 2013-08-20 Teresa Johnson <tejohnson@google.com>
11670 PR rtl-optimizations/57451
11671 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11672 from crossing split section boundaries.
11674 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11676 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11677 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11678 with MULTLIB_DEFAULTS.
11680 2013-08-20 Nick Clifton <nickc@redhat.com>
11682 * target.def (narrow_volatile_bitfield): Note that the default
11683 value is false, not !TARGET_STRICT_ALIGN.
11684 * doc/tm.texi: Regenerate.
11686 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11688 Fix LIB_SPEC for systems without libpthread.
11690 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11691 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11693 * config/i386/linux-common.h: Likewise.
11694 * config/mips/linux-common.h: Likewise.
11696 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11698 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11701 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11703 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11704 instances can own GC refs.
11706 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11707 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11708 (gcc::context::gt_pch_nx): Likewise.
11709 (gcc::context::gt_pch_nx): Likewise.
11710 * ggc.h (gt_ggc_mx <T>): New.
11711 (gt_pch_nx_with_op <T>): New.
11712 (gt_pch_nx <T>): New.
11713 * passes.c (opt_pass::gt_ggc_mx): New.
11714 (opt_pass::gt_pch_nx): New.
11715 (opt_pass::gt_pch_nx_with_op): New.
11716 (pass_manager::gt_ggc_mx): New.
11717 (pass_manager::gt_pch_nx): New.
11718 (pass_manager::gt_pch_nx_with_op): New.
11719 (pass_manager::operator new): Use
11720 ggc_internal_cleared_alloc_stat rather than xcalloc.
11721 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11722 (pass_manager::gt_ggc_mx): New.
11723 (pass_manager::gt_pch_nx): New.
11724 (pass_manager::gt_pch_nx_with_op): New.
11725 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11726 (opt_pass::operator new): New.
11727 (opt_pass::gt_ggc_mx): New.
11728 (opt_pass::gt_pch_nx): New.
11729 (opt_pass::gt_pch_nx_with_op): New.
11731 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11733 * Makefile.in (GTFILES): Add context.h.
11734 * context.c (gcc::context::operator new): New.
11735 (gcc::context::gt_ggc_mx): New.
11736 (gcc::context::gt_pch_nx): New.
11737 (gcc::context::gt_pch_nx): New.
11738 * context.h (gcc::context): Add GTY((user)) marking.
11739 (gcc::context::operator new): New.
11740 (gcc::context::gt_ggc_mx): New.
11741 (gcc::context::gt_pch_nx): New.
11742 (gcc::context::gt_pch_nx): New.
11743 (g): Add GTY marking.
11744 (gt_ggc_mx (gcc::context *)): New.
11745 (gt_pch_nx (gcc::context *)): New.
11746 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11747 void *cookie)): New.
11748 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11750 2013-08-20 Alan Modra <amodra@gmail.com>
11753 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11754 (rs6000_emit_epilogue): Likewise.
11756 2013-08-19 Dehao Chen <dehao@google.com>
11758 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11760 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11761 Jakub Jelinek <jakub@redhat.com>
11763 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11764 (BUILT_IN_FABSD64): Likewise.
11765 (BUILT_IN_FABSD128): Likewise.
11766 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11767 (fold_builtin_1): Likewise.
11768 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11769 destination and source operands.
11770 (*abstd2_fpr): Likewise.
11771 (*nabstd2_fpr): Likewise.
11773 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11775 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11778 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11780 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11783 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11785 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11786 * builtins.c (default_libc_has_function): New.
11787 (gnu_libc_has_function): Ditto.
11788 (no_c99_libc_has_function): Ditto.
11789 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11790 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11791 (fold_builtin_sincos): Likewise.
11792 (fold_builtin_cexp): Likewise.
11793 * builtins.def (DEF_C94_BUILTIN): Likewise.
11794 (DEF_C99_BUILTIN): Likewise.
11795 (DEF_C99_C90RES_BUILTIN): Likewise.
11796 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11797 definitions to using this define.
11798 * config/darwin-protos.h (darwin_libc_has_function): New.
11799 * config/darwin.c (darwin_libc_has_function): Ditto.
11800 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11801 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11802 * config/darwin.h: Ditto.
11803 * config/elfos.h: Ditto.
11804 * config/freebsd.h: Ditto.
11805 * config/i386/cygming.h: Ditto.
11806 * config/i386/djgpp.h: Ditto.
11807 * config/i386/i386-interix.h: Ditto.
11808 * config/microblaze/microblaze.h: Ditto.
11809 * config/mmix/mmix.h: Ditto.
11810 * config/gnu-user.h: Ditto.
11811 * config/ia64/hpux.h: Ditto.
11812 * config/pa/pa-hpux.h: Ditto.
11813 * config/pdp11/pdp11.h: Ditto.
11814 * config/picochip/picochip.h: Ditto.
11815 * config/linux.h: Ditto.
11816 * config/netbsd.h: Ditto.
11817 * config/openbsd.h: Ditto.
11818 * config/rs6000/aix43.h: Ditto.
11819 * config/rs6000/aix51.h: Ditto.
11820 * config/rs6000/aix52.h: Ditto.
11821 * config/rs6000/aix53.h: Ditto.
11822 * config/rs6000/aix61.h: Ditto.
11823 * config/rs6000/darwin.h: Ditto.
11824 * config/rs6000/linux.h: Ditto.
11825 * config/rs6000/linux64.h: Ditto.
11826 * config/s390/tpf.h: Ditto.
11827 * config/sol2-10.h: Ditto.
11828 * config/sol2.h: Ditto.
11829 * config/vms/vms.h: Ditto.
11830 * config/vxworks.h: Ditto.
11831 * config/linux-android.c (linux_android_libc_has_function):
11832 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11833 * config/linux-protos.h (linux_android_libc_has_function):
11835 * config/i386/i386.c (ix86_libc_has_function): New.
11836 * config/i386/i386-protos.h
11837 (ix86_libc_has_function): New declaration.
11838 * config/i386/i386.md
11839 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11840 ("isinf<mode>2): Likewise.
11841 * convert.c (convert_to_integer): Using new target hook
11842 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11843 TARGET_C99_FUNCTIONS.
11844 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11845 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11846 * coretypes.h (function_class): New enum for different
11847 classes of functions.
11848 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11849 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11850 (TARGET_HAS_SINCOS): Likewise.
11851 (TARGET_LIBC_HAS_FUNCTION): New.
11852 * doc/tm.texi: Regenerated.
11853 * targhooks.h (default_libc_has_function): New declaration.
11854 (no_c99_libc_has_function): Ditto.
11855 (gnu_libc_has_function): Ditto.
11856 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11857 and TARGET_HAS_SINCOS.
11859 2013-08-18 Jan Hubicka <jh@suse.cz>
11861 * Makeifle-in (ipa-devirt.o): New.
11862 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11863 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11864 (analyze_functions): Look into possible targets of polymorphic call.
11865 * dumpfile.c (dump_files): Add type-inheritance dump.
11866 * dumpfile.h (TDI_inheritance): New.
11867 * ipa-devirt.c: New file.
11868 * ipa-utils.h (odr_type_d): Forward declare.
11869 (odr_type): New type.
11870 (build_type_inheritance_graph): Declare.
11871 (possible_polymorphic_call_targets): Declare and introduce inline
11872 variant when only edge is pased.
11873 (dump_possible_polymorphic_call_targets): Likewise.
11874 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11875 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11876 (types_same_for_odr): ... here.
11877 * tree.h (type_in_anonymous_namespace_p): Declare.
11879 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11881 PR tree-optimization/58006
11882 * tree-parloops.c (take_address_of): Don't ICE if get_name
11884 (eliminate_local_variables_stmt): Remove clobber stmts.
11886 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11888 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11889 error message is issued for an alias to undefined symbol.
11891 2013-08-18 Jan Hubicka <jh@suse.cz>
11893 * cgraph.c (cgraph_create_indirect_edge): Discover
11894 polymorphic calls and record basic info into indirect_info.
11895 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11896 devirtualization, ignore objc function calls.
11897 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11898 call with no parm index info.
11899 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11900 * tree.c (virtual_method_call_p): New function.
11901 * tree.h (virtual_method_call_p): Declare.
11903 2013-08-16 Jan Hubicka <jh@suse.cz>
11905 PR middle-end/58179
11906 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11908 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11910 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11913 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11915 * gengtype.c (type_for_name): Add special-case support for
11916 locating types within the "gcc::" namespace.
11917 (open_base_files): Emit a "using namespace gcc" directive.
11919 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11922 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11923 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11925 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11926 array instead of each individual operand as a separate argument.
11927 (emit_fusion_gpr_load): Likewise.
11928 (expand_fusion_gpr_load): Add new function declaration.
11930 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11931 signature to have the operands passed as an array, instead of as
11932 separate arguments. Allow ZERO_EXTEND to be in the memory
11933 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11934 depend on the register live/dead flags when peepholes are run.
11935 (expand_fusion_gpr_load): New function to be called from the
11936 peephole2 pass, to change the register that addis sets to be the
11938 (emit_fusion_gpr_load): Change the calling signature to have the
11939 operands passed as an array, instead of as separate arguments.
11940 Allow ZERO_EXTEND to be in the memory address, and also
11941 SIGN_EXTEND if -mpower8-fusion-sign.
11943 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11944 unspec enumeration.
11945 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11946 adjust the register addis loads up in the peephole2 pass. Do not
11947 depend on the register live/dead state when the peephole pass is done.
11949 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11951 * gengtype.c (create_user_defined_type): Ensure that the kind
11952 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11953 declaration is seen before the GTY((user)) marking.
11955 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11958 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11960 2013-08-16 Jan Hubicka <jh@suse.cz>
11962 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11963 arugment expected_type.
11964 (gimple_fold_call): Use it.
11965 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11966 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11967 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11968 (try_make_edge_direct_virtual_call): Likewise.
11969 * tree.c (obj_type_ref_class): New.
11970 * tree.h (obj_type_ref_class): Use it.
11972 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11974 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11975 (rtl_slim_pp): Likewise.
11976 (init_rtl_slim_pretty_print): Likewise.
11977 (dump_value_slim): Don't call it. Use local pretty printer.
11978 (dump_insn_slim): Likewise.
11979 (dump_rtl_slim): Likewise.
11980 (str_pattern_slim): Likewise.
11981 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11984 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11986 PR tree-optimization/58164
11987 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11988 walk gimple_goto_dest of GIMPLE_GOTO.
11990 PR tree-optimization/58165
11991 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11992 bi_call must be the last stmt in a bb, don't split_block, instead
11993 use fallthru edge from it and give up if there is none.
11994 Release conds vector when returning early.
11996 2013-08-14 Xinliang David Li <davidxl@google.com>
11998 * config/i386/i386.c (ix86_option_override_internal):
11999 Remove unused variable and field.
12001 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12004 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12005 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12006 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12007 and Linux, correct BLKmode alignment when 128-bit alignment is
12008 required and compatibility flag is not set.
12009 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12010 for zero-size arguments when compatibility flag is not set.
12012 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12014 PR tree-optimization/58145
12015 * tree-sra.c (build_ref_for_offset): If prev_base has
12016 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12018 2013-08-14 Xinliang David Li <davidxl@google.com>
12020 * config/i386/i386.c (ix86_option_override_internal):
12021 Fix uninitialized variable error.
12023 2013-08-14 Xinliang David Li <davidxl@google.com>
12025 * config/i386/i386.opt: Define two new options.
12026 * config/i386/x86-tune.def: Add arch selector field in macros.
12027 * config/i386/i386.h: Adjust macro definition.
12028 * config/i386/i386.c (ix86_option_override_internal):
12030 (parse_mtune_ctrl_str): New function.
12031 (set_ix86_tune_features): New function.
12032 (ix86_function_specific_restore): Call the new helper function.
12034 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12036 PR rtl-optimization/57662
12037 * sel-sched.c (code_motion_process_successors): When the current insn
12038 is removed after the recursive traversal, break from the loop.
12039 Add comments and debug printouts.
12041 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12042 Alexandre Oliva <aoliva@redhat.com>
12045 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12046 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12047 there also UNSPEC_PLTOFF.
12049 2013-08-14 Marek Polacek <polacek@redhat.com>
12051 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12052 NUM_CONDITIONS bit positions.
12054 2013-08-13 Cary Coutant <ccoutant@google.com>
12056 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12057 (attr_checksum): Hash vector contents instead of pointer.
12058 (attr_checksum_ordered): Likewise.
12060 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12062 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12063 when Pmode != word_mode. Add length_address attribute.
12064 (sse3_monitor_<mode>): Merge from sse3_monitor and
12065 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12066 Pmode != word_mode. Update insn length attribute.
12067 * config/i386/i386.c (ix86_option_override_internal): Update
12068 ix86_gen_monitor selection for merged sse3_monitor insn.
12070 2013-08-13 Julian Brown <julian@codesourcery.com>
12072 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12073 perform invalid legitimization on greater-than-word-size modes for
12074 TARGET_E500_DOUBLE.
12076 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12078 * ira.c (setup_class_translate_array): Use aclass instead of cl
12079 for classes not fully covered by allocno classes.
12081 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12083 PR tree-optimization/57661
12084 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12085 * tree-inline.c (tree_function_versioning): Initialize it.
12086 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12087 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12088 that is not being copied.
12091 * asan.c (instrument_strlen_call): Fix typo in comment.
12092 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12094 2013-08-13 Steve Ellcey <sellcey@mips.com>
12096 * config/mips/mips.md (prefetch): Use lw instead of ld on
12097 loongson in 32bit mode.
12099 2013-08-13 Nick Clifton <nickc@redhat.com>
12101 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12103 2013-08-13 Jan Hubicka <jh@suse.cz>
12105 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12106 introduced edge; fix typo in sanity check.
12107 (cgraph_resolve_speculation): Export; improve diagnostic.
12108 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12109 speculation at type mismatch.
12110 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12111 (cgraph_resolve_speculation): Declare.
12112 (symtab_can_be_discarded): New function.
12113 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12114 * ipa-inline-transform.c (speculation_removed): New global var.
12115 (clone_inlined_nodes): See if speculation can be removed.
12116 (inline_call): If speculations was removed, we growths may not match.
12117 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12118 (speculation_useful_p): New function.
12119 (resolve_noninline_speculation): New function.
12120 (inline_small_functions): Resolve useless speculations.
12121 * ipa-inline.h (speculation_useful_p): Declare
12122 * ipa.c (can_replace_by_local_alias): Simplify.
12123 (ipa_profile): Produce speculative calls in non-lto, too;
12124 add simple cost model; produce local aliases.
12126 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12128 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12131 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12133 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12135 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12137 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12138 iterator instead of X87MODEF.
12140 2013-08-12 Perez Read <netfirewall@gmail.com>
12143 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12144 operand 0 for intel asm alternative.
12145 (*movabs<mode>_2): Ditto for operand 1.
12147 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12149 * config/aarch64/arm_none.h
12150 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12152 2013-08-12 Nick Clifton <nickc@redhat.com>
12154 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12156 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12158 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12159 expand for QI/HImode operand to produce more effictive code for
12160 unsigned char(short) --> float(double) conversion.
12162 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12164 * doc/invoke.texi: Mention that -ftls-model does not force the final
12167 2013-08-12 Marek Polacek <polacek@redhat.com>
12168 Marc Glisse <marc.glisse@inria.fr>
12170 PR tree-optimization/57980
12171 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12172 when creating -1 constant.
12174 2013-08-10 Jan Hubicka <jh@suse.cz>
12176 Workaround binutils PR14342.
12177 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12178 (gimple_init_edge_profiler): Likewise.
12179 (gimple_gen_ic_func_profiler): Likewise.
12181 2013-08-09 Jan Hubicka <jh@suse.cz>
12183 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12185 2013-08-09 Xinliang David Li <davidxl@google.com>
12187 * config/i386/stringop.def: New file.
12188 * config/i386/stringop.opt: New file.
12189 * config/i386/i386-opts.h: Include stringopt.def.
12190 * config/i386/i386.opt: Include stringopt.opt.
12191 * config/i386/i386.c (ix86_option_override_internal):
12192 Override default size based stringop inline strategies with options.
12193 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12196 2013-08-09 Jan Hubicka <jh@suse.cz>
12198 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12200 2013-08-09 Jan Hubicka <jh@suse.cz>
12202 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12204 (dump_cgraph_node): Dump profile-id.
12205 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12206 and common_target_probability.
12207 * lto-cgraph.c (lto_output_edge): Stream common targets.
12208 (lto_output_node): Stream profile ids.
12209 (input_node): Stream profile ids.
12210 (input_edge): Stream common targets.
12211 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12212 * ipa.c: Include value-prof.h
12213 (ipa_profile_generate_summary): Turn indirect call statement histograms
12214 into common targets.
12215 (ipa_profile): Turn common targets into speculative edges.
12217 2013-08-09 Jan Hubicka <jh@suse.cz>
12219 * cgraph.h (cgraph_node): Add profile_id.
12220 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12221 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12222 (del_node_map): Update.
12223 (find_func_by_funcdef_no): Replace by ...
12224 (find_func_by_profile_id): ... this one.
12225 (gimple_ic_transform): Do not remove useful histograms when
12226 speculation is not done; dump info when indirect call removal
12228 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12229 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12230 (__gcov_indirect_call_profiler_v2): .. this one.
12231 * profile.h (init_node_map): Update.
12232 * coverage.c (coverage_compute_profile_id): New function.
12233 * coverage.h (coverage_compute_profile_id): Declare.
12234 * tree-profile.c (init_ic_make_global_vars): Make
12235 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12236 (gimple_init_edge_profiler): Update prototype of
12237 __gcov_indirect_call_profiler.
12238 (gimple_gen_ic_func_profiler): Simplify.
12239 (tree_profiling): Use init_node_map
12241 2013-08-09 Jan Hubicka <jh@suse.cz>
12243 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12244 non-speculative refs.
12245 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12246 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12247 (cgraph_set_call_stmt): Likewise.
12248 (cgraph_create_edge_1): Fix release checking compilatoin;
12249 clear lto_stmt_uid.
12250 (cgraph_free_edge): Free indirect info.
12251 (cgraph_turn_edge_to_speculative): New function.
12252 (cgraph_speculative_call_info): New function.
12253 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12254 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12255 (dump_cgraph_node): Dump speculation.
12256 (verify_edge_count_and_frequency): Accept speculative edges.
12257 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12258 (verify_cgraph_node): Handle speculation.
12259 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12260 (cgraph_set_call_stmt): Update prototype.
12261 (cgraph_make_edge_direct): Update prototype.
12262 (cgraph_speculative_call_info): Declare.
12263 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12264 to change; update call of ipa_find_references.
12265 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12266 lto_stmt_uid and speculative flags.
12267 (ipa_dump_references): Dump speculation.
12268 (ipa_clone_references): Clone speculative flag.
12269 (ipa_clone_referring): Likewise.
12270 (ipa_clone_ref): New function.
12271 (ipa_find_reference): Look into lto_stmt_uids
12272 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12273 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12274 (ipa_find_reference): Update declaration.
12275 (ipa_clone_ref): Declare.
12276 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12277 stream speculative flag.
12278 (lto_output_ref): Stream statements uids and speculation.
12279 (input_ref): Likewise.
12280 (input_edge): Stream speuclation.
12281 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12282 (cgraph_set_call_stmt_including_clones): Handle speculation.
12283 * ipa-inline.c (heap_edge_removal_hook): New function.
12284 (inline_small_functions): Register it.
12285 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12286 also initialize refs.
12287 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
12289 (try_make_edge_direct_simple_call): Likewise.
12290 (try_make_edge_direct_simple_call): Likewise.
12291 (update_indirect_edges_after_inlining): Likewise.
12292 (remove_described_reference): Look proper lto_stmt_uid.
12293 (propagate_controlled_uses): Likewise.
12294 (propagate_controlled_uses): Liekwise.
12295 * tree-inline.c (copy_bb): Copy speculative edges.
12296 (redirect_all_calls): New function.
12297 (copy_cfg_body): Do redirection after loop info is updated.
12298 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
12300 2013-08-09 Jan Hubicka <jh@suse.cz>
12302 * lto-streamer-out.c (output_function): Renumber PHIs.
12303 * lto-streamer-in.c (input_function): Likewise.
12305 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
12307 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
12308 (get_lane_unsigned): Likewise.
12309 (dup_lane_scalar): Likewise.
12310 (get_lane): enable for VALL.
12311 * config/aarch64/aarch64-simd.md
12312 (aarch64_dup_lane_scalar<mode>): Remove.
12313 (aarch64_get_lane_signed<mode>): Likewise.
12314 (aarch64_get_lane_unsigned<mode>): Likewise.
12315 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
12316 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
12317 (aarch64_get_lane<mode>): Enable for all vector modes.
12318 (aarch64_get_lanedi): Remove misleading constraints.
12319 * config/aarch64/arm_neon.h
12320 (__aarch64_vget_lane_any): Define.
12321 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
12322 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
12323 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
12324 * config/aarch64/iterators.md (VDQQH): New.
12325 (VDQQHS): Likewise.
12326 (vwcore): Likewise.
12328 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
12330 * configure.ac: Add GAS check for LEON instructions on SPARC.
12331 * configure: Regenerate.
12332 * config.in: Likewise.
12333 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
12335 * config/sparc/sparc.opt (LEON, LEON3): New masks.
12336 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
12338 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
12339 (AS_LEON_FLAG): New macro.
12340 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
12341 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
12342 Deal with LEON and LEON3 for the memory model.
12343 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
12344 (atomic_compare_and_swap<mode>_1): Likewise.
12345 (*atomic_compare_and_swap<mode>_1): Likewise.
12347 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12349 * config/arm/neon.md (vcond): Fix floating-point vector
12350 comparisons against 0.
12352 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
12354 * lra-constraints.c (emit_spill_move): Remove assert.
12355 (process_alt_operands): Add more debugging
12356 output. Increase reject for spilling into memory. Decrease
12357 reject for reloading scratch.
12358 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12360 2013-08-08 Steve Ellcey <sellcey@mips.com>
12362 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12363 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12364 micromips incompatible. Add nan2008.
12365 (MULTILIB_DIRNAMES): Add nan2008.
12366 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12367 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12368 and micromips incompatible. Add nan2008.
12369 (MULTILIB_DIRNAMES): Add nan2008.
12370 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12372 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12374 PR rtl-optimization/58079
12375 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12376 simplify_comparison has widened a comparison with an integer.
12378 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12380 * config/arm/neon.md (movmisalign<mode>): Disable when we
12381 don't allow unaligned accesses.
12382 (*movmisalign<mode>_neon_store): Likewise.
12383 (*movmisalign<mode>_neon_load): Likewise.
12384 (*movmisalign<mode>_neon_store): Likewise.
12385 (*movmisalign<mode>_neon_load): Likewise.
12387 2013-08-08 Jan Hubicka <jh@suse.cz>
12389 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12390 (make_pass_rebuild_cgraph_edges): Also clear references.
12391 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12392 * ipa-inline-transform.c (inline_transform): Remove all references
12394 * cgraphunit.c (expand_function): Remove all references after
12396 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12397 (ipa_find_reference): Rewrite to iterator.
12398 (remove_stmt_references): Likewise.
12399 (ipa_clear_stmts_in_references): New function.
12400 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12401 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12403 * ipa-split.c (split_function): Remove references in split function.
12405 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12408 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12409 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12411 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12414 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12415 suggested mode for the assignment isn't compatible with the
12416 registers required.
12418 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12421 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12423 2013-08-07 Xinliang David Li <davidxl@google.com>
12425 * config/i386/i386.opt: New option -mtune-ctrl=.
12426 * config/i386/x86-tune.def: New file.
12427 * config/i386/i386.h: include x86-tune.def.
12428 * config/i386/i386.c (ix86_option_override_internal):
12429 Parsing -mtune-ctrl= option and set tune features.
12431 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12434 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12435 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12438 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12440 * rtl.h (update_alignments): Declare.
12441 * final.c (grow_label_align): New function extracted from...
12442 (shorten_branches): ...here. Call it.
12443 (update_alignments): New function.
12444 * reorg.c (sibling_labels): New variable.
12445 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12446 the new label along with it onto the sibling_labels vector.
12447 (fill_simple_delay_slots): Adjust call to get_label_before.
12448 (fill_slots_from_thread): Likewise.
12449 (relax_delay_slots): Likewise.
12450 (make_return_insns): Likewise.
12451 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12453 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12455 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12456 document its semantics.
12457 (diagnostic_report_diagnostic): Adjust accordingly.
12459 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12461 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12462 (sparc_option_override): ...and port to new C++ pass API.
12463 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12465 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12467 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12469 2013-08-06 Caroline Tice <cmtice@google.com>
12471 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12472 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12473 * tree-pass.h: Add pass_vtable_verify.
12474 * varasm.c (assemble_variable): Add code to properly set the comdat
12475 section and name for the .vtable_map_vars section.
12476 (assemble_vtyv_preinit_initializer): New function.
12477 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12479 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12480 * vtable-verify.c: New file.
12481 * vtable-verify.h: New file.
12482 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12483 initialiation levels.
12484 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12485 * passes.def: Insert pass_vtable_verify.
12486 * aclocal.m4: Reorder includes.
12487 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12488 -fvtv-counts options.
12489 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12490 as appropriate, if -fvtable-verify=... is used.
12491 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12492 -fvtable-verify=... is used.
12493 * Makefile.in (OBJS): Add vtable-verify.o to list.
12494 (vtable-verify.o): Add new build rule.
12495 (GTFILES): Add vtable-verify.c to list.
12496 * common.opt (fvtable-verify=): New flag.
12497 (vtv_priority): Values for fvtable-verify= flag.
12498 (fvtv-counts): New flag.
12499 (fvtv-debug): New flag.
12500 * tree.h (save_vtable_map_decl): New extern function decl.
12502 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12504 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12505 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12506 (pass_data_rl78_devirt): ...new pass_data instance and...
12507 (make_pass_rl78_devirt): ...new function.
12508 (rl78_asm_file_start): Port pass registration to new C++ API.
12510 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12512 * coretypes.h (rtl_opt_pass): Add.
12513 (gcc::context): Add.
12514 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12515 (epiphany_init): Port to new C++ pass API.
12516 (epiphany_optimize_mode_switching): Likewise.
12517 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12518 (pass_manager::get_pass_mode_switching): New.
12519 (pass_manager::get_pass_peephole2): New.
12520 * mode-switching.c (pass_mode_switching): Add clone method.
12521 * recog.c (pass_peephole2): Add clone method.
12522 (pass_split_all_insns): Add clone method.
12524 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12526 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12527 (mips_option_override): ...here, porting to new C++ API for passes.
12529 2013-08-06 Jan Hubicka <jh@suse.cz>
12531 * cgraph.c (cgraph_get_body): New function based on lto.c
12533 * cgraph.h (cgraph_get_body): Declare.
12534 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12536 * cgraphunit.c (expand_function): Get body prior expanding.
12537 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12538 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12540 * passes.c (do_per_function_toporder): Get body.
12541 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12542 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12544 2013-08-06 Martin Jambor <mjambor@suse.cz>
12547 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12548 is not re-finalized. Rename second parameter to no_collect.
12550 2013-08-06 Martin Jambor <mjambor@suse.cz>
12552 PR middle-end/58041
12553 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12554 MEM_REF has proper alignment information.
12556 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12559 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12561 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12562 rtx (*) (rtx, ...) with insn_gen_fn.
12563 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12564 insn_gen_fn::stored_funcptr. Add initializer braces.
12566 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12568 Rewrite how instances of passes are cloned to remove assumptions
12569 about their sizes (thus allowing pass subclasses to have
12570 additional data fields, albeit non-GC-managed ones at this point).
12572 * passes.c (make_pass_instance): Now that passes have clone
12573 methods, rewrite this function to eliminate XNEW and memcpy
12574 calls that used hardcoded sizes. Since this function no longer
12575 creates pass instances, rename it to...
12576 (add_pass_instance): ...this. Document the old way that passes were
12577 numbered and flagged, and rework this function to continue using it.
12578 (next_pass_1): Add an initial_pass argument for use by
12580 (position_pass): When adding multiple instances of a pass, use
12581 the pass's clone method, rather than relying on the XNEW/memcpy
12582 within the former make_pass_instance (now add_pass_instance).
12583 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12584 the initial instance of the current pass within the pass manager.
12586 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12588 This is the automated part of the conversion of passes from C
12589 structs to C++ classes.
12591 Patch autogenerated by refactor_passes.py from
12592 https://github.com/davidmalcolm/gcc-refactoring-scripts
12593 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12595 * asan.c (pass_asan): Convert from a global struct to a subclass of
12596 gimple_opt_pass along with...
12597 (pass_data_asan): ...new pass_data instance and...
12598 (make_pass_asan): ...new function.
12599 (pass_asan_O0): Convert from a global struct to a subclass of
12600 gimple_opt_pass along with...
12601 (pass_data_asan_O0): ...new pass_data instance and...
12602 (make_pass_asan_O0): ...new function.
12603 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12604 subclass of rtl_opt_pass along with...
12605 (pass_data_inc_dec): ...new pass_data instance and...
12606 (make_pass_inc_dec): ...new function.
12607 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12608 a subclass of rtl_opt_pass along with...
12609 (pass_data_reorder_blocks): ...new pass_data instance and...
12610 (make_pass_reorder_blocks): ...new function.
12611 (pass_duplicate_computed_gotos): Convert from a global struct to a
12612 subclass of rtl_opt_pass along with...
12613 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12614 (make_pass_duplicate_computed_gotos): ...new function.
12615 (pass_partition_blocks): Convert from a global struct to a subclass of
12616 rtl_opt_pass along with...
12617 (pass_data_partition_blocks): ...new pass_data instance and...
12618 (make_pass_partition_blocks): ...new function.
12619 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12620 struct to a subclass of rtl_opt_pass along with...
12621 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12623 (make_pass_branch_target_load_optimize1): ...new function.
12624 (pass_branch_target_load_optimize2): Convert from a global struct to a
12625 subclass of rtl_opt_pass along with...
12626 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12628 (make_pass_branch_target_load_optimize2): ...new function.
12629 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12630 of rtl_opt_pass along with...
12631 (pass_data_jump): ...new pass_data instance and...
12632 (make_pass_jump): ...new function.
12633 (pass_jump2): Convert from a global struct to a subclass of
12634 rtl_opt_pass along with...
12635 (pass_data_jump2): ...new pass_data instance and...
12636 (make_pass_jump2): ...new function.
12637 * cfgexpand.c (pass_expand): Convert from a global struct to a
12638 subclass of rtl_opt_pass along with...
12639 (pass_data_expand): ...new pass_data instance and...
12640 (make_pass_expand): ...new function.
12641 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12642 of rtl_opt_pass along with...
12643 (pass_data_free_cfg): ...new pass_data instance and...
12644 (make_pass_free_cfg): ...new function.
12645 (pass_into_cfg_layout_mode): Convert from a global struct to a
12646 subclass of rtl_opt_pass along with...
12647 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12648 (make_pass_into_cfg_layout_mode): ...new function.
12649 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12650 subclass of rtl_opt_pass along with...
12651 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12652 (make_pass_outof_cfg_layout_mode): ...new function.
12653 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12654 struct to a subclass of gimple_opt_pass along with...
12655 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12656 (make_pass_build_cgraph_edges): ...new function.
12657 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12658 subclass of gimple_opt_pass along with...
12659 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12660 (make_pass_rebuild_cgraph_edges): ...new function.
12661 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12662 subclass of gimple_opt_pass along with...
12663 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12665 (make_pass_remove_cgraph_callee_edges): ...new function.
12666 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12667 struct to a subclass of rtl_opt_pass along with...
12668 (pass_data_stack_adjustments): ...new pass_data instance and...
12669 (make_pass_stack_adjustments): ...new function.
12670 * combine.c (pass_combine): Convert from a global struct to a subclass
12671 of rtl_opt_pass along with...
12672 (pass_data_combine): ...new pass_data instance and...
12673 (make_pass_combine): ...new function.
12674 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12675 global struct to a subclass of rtl_opt_pass along with...
12676 (pass_data_compare_elim_after_reload): ...new pass_data instance
12678 (make_pass_compare_elim_after_reload): ...new function.
12679 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12680 of rtl_opt_pass along with...
12681 (pass_data_rtl_cprop): ...new pass_data instance and...
12682 (make_pass_rtl_cprop): ...new function.
12683 * cse.c (pass_cse): Convert from a global struct to a subclass of
12684 rtl_opt_pass along with...
12685 (pass_data_cse): ...new pass_data instance and...
12686 (make_pass_cse): ...new function.
12687 (pass_cse2): Convert from a global struct to a subclass of
12688 rtl_opt_pass along with...
12689 (pass_data_cse2): ...new pass_data instance and...
12690 (make_pass_cse2): ...new function.
12691 (pass_cse_after_global_opts): Convert from a global struct to a
12692 subclass of rtl_opt_pass along with...
12693 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12694 (make_pass_cse_after_global_opts): ...new function.
12695 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12696 of rtl_opt_pass along with...
12697 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12698 (make_pass_ud_rtl_dce): ...new function.
12699 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12700 rtl_opt_pass along with...
12701 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12702 (make_pass_fast_rtl_dce): ...new function.
12703 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12704 a subclass of rtl_opt_pass along with...
12705 (pass_data_df_initialize_opt): ...new pass_data instance and...
12706 (make_pass_df_initialize_opt): ...new function.
12707 (pass_df_initialize_no_opt): Convert from a global struct to a
12708 subclass of rtl_opt_pass along with...
12709 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12710 (make_pass_df_initialize_no_opt): ...new function.
12711 (pass_df_finish): Convert from a global struct to a subclass of
12712 rtl_opt_pass along with...
12713 (pass_data_df_finish): ...new pass_data instance and...
12714 (make_pass_df_finish): ...new function.
12715 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12716 rtl_opt_pass along with...
12717 (pass_data_rtl_dse1): ...new pass_data instance and...
12718 (make_pass_rtl_dse1): ...new function.
12719 (pass_rtl_dse2): Convert from a global struct to a subclass of
12720 rtl_opt_pass along with...
12721 (pass_data_rtl_dse2): ...new pass_data instance and...
12722 (make_pass_rtl_dse2): ...new function.
12723 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12724 subclass of rtl_opt_pass along with...
12725 (pass_data_dwarf2_frame): ...new pass_data instance and...
12726 (make_pass_dwarf2_frame): ...new function.
12727 * except.c (pass_set_nothrow_function_flags): Convert from a global
12728 struct to a subclass of rtl_opt_pass along with...
12729 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12731 (make_pass_set_nothrow_function_flags): ...new function.
12732 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12733 subclass of rtl_opt_pass along with...
12734 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12736 (make_pass_convert_to_eh_region_ranges): ...new function.
12737 * final.c (pass_compute_alignments): Convert from a global struct to a
12738 subclass of rtl_opt_pass along with...
12739 (pass_data_compute_alignments): ...new pass_data instance and...
12740 (make_pass_compute_alignments): ...new function.
12741 (pass_final): Convert from a global struct to a subclass of
12742 rtl_opt_pass along with...
12743 (pass_data_final): ...new pass_data instance and...
12744 (make_pass_final): ...new function.
12745 (pass_shorten_branches): Convert from a global struct to a subclass of
12746 rtl_opt_pass along with...
12747 (pass_data_shorten_branches): ...new pass_data instance and...
12748 (make_pass_shorten_branches): ...new function.
12749 (pass_clean_state): Convert from a global struct to a subclass of
12750 rtl_opt_pass along with...
12751 (pass_data_clean_state): ...new pass_data instance and...
12752 (make_pass_clean_state): ...new function.
12753 * function.c (pass_instantiate_virtual_regs): Convert from a global
12754 struct to a subclass of rtl_opt_pass along with...
12755 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12756 (make_pass_instantiate_virtual_regs): ...new function.
12757 (pass_leaf_regs): Convert from a global struct to a subclass of
12758 rtl_opt_pass along with...
12759 (pass_data_leaf_regs): ...new pass_data instance and...
12760 (make_pass_leaf_regs): ...new function.
12761 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12762 subclass of rtl_opt_pass along with...
12763 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12765 (make_pass_thread_prologue_and_epilogue): ...new function.
12766 (pass_match_asm_constraints): Convert from a global struct to a
12767 subclass of rtl_opt_pass along with...
12768 (pass_data_match_asm_constraints): ...new pass_data instance and...
12769 (make_pass_match_asm_constraints): ...new function.
12770 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12771 subclass of rtl_opt_pass along with...
12772 (pass_data_rtl_fwprop): ...new pass_data instance and...
12773 (make_pass_rtl_fwprop): ...new function.
12774 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12775 rtl_opt_pass along with...
12776 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12777 (make_pass_rtl_fwprop_addr): ...new function.
12778 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12779 rtl_opt_pass along with...
12780 (pass_data_rtl_pre): ...new pass_data instance and...
12781 (make_pass_rtl_pre): ...new function.
12782 (pass_rtl_hoist): Convert from a global struct to a subclass of
12783 rtl_opt_pass along with...
12784 (pass_data_rtl_hoist): ...new pass_data instance and...
12785 (make_pass_rtl_hoist): ...new function.
12786 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12787 subclass of gimple_opt_pass along with...
12788 (pass_data_lower_cf): ...new pass_data instance and...
12789 (make_pass_lower_cf): ...new function.
12790 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12791 from a global struct to a subclass of gimple_opt_pass along with...
12792 (pass_data_strength_reduction): ...new pass_data instance and...
12793 (make_pass_strength_reduction): ...new function.
12794 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12795 of rtl_opt_pass along with...
12796 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12797 (make_pass_rtl_ifcvt): ...new function.
12798 (pass_if_after_combine): Convert from a global struct to a subclass of
12799 rtl_opt_pass along with...
12800 (pass_data_if_after_combine): ...new pass_data instance and...
12801 (make_pass_if_after_combine): ...new function.
12802 (pass_if_after_reload): Convert from a global struct to a subclass of
12803 rtl_opt_pass along with...
12804 (pass_data_if_after_reload): ...new pass_data instance and...
12805 (make_pass_if_after_reload): ...new function.
12806 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12807 a subclass of rtl_opt_pass along with...
12808 (pass_data_initialize_regs): ...new pass_data instance and...
12809 (make_pass_initialize_regs): ...new function.
12810 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12811 of ipa_opt_pass_d along with...
12812 (pass_data_ipa_cp): ...new pass_data instance and...
12813 (make_pass_ipa_cp): ...new function.
12814 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12815 global struct to a subclass of gimple_opt_pass along with...
12816 (pass_data_inline_parameters): ...new pass_data instance and...
12817 (make_pass_inline_parameters): ...new function.
12818 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12819 subclass of gimple_opt_pass along with...
12820 (pass_data_early_inline): ...new pass_data instance and...
12821 (make_pass_early_inline): ...new function.
12822 (pass_ipa_inline): Convert from a global struct to a subclass of
12823 ipa_opt_pass_d along with...
12824 (pass_data_ipa_inline): ...new pass_data instance and...
12825 (make_pass_ipa_inline): ...new function.
12826 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12827 struct to a subclass of gimple_opt_pass along with...
12828 (pass_data_local_pure_const): ...new pass_data instance and...
12829 (make_pass_local_pure_const): ...new function.
12830 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12831 ipa_opt_pass_d along with...
12832 (pass_data_ipa_pure_const): ...new pass_data instance and...
12833 (make_pass_ipa_pure_const): ...new function.
12834 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12835 to a subclass of ipa_opt_pass_d along with...
12836 (pass_data_ipa_reference): ...new pass_data instance and...
12837 (make_pass_ipa_reference): ...new function.
12838 * ipa-split.c (pass_split_functions): Convert from a global struct to
12839 a subclass of gimple_opt_pass along with...
12840 (pass_data_split_functions): ...new pass_data instance and...
12841 (make_pass_split_functions): ...new function.
12842 (pass_feedback_split_functions): Convert from a global struct to a
12843 subclass of gimple_opt_pass along with...
12844 (pass_data_feedback_split_functions): ...new pass_data instance and...
12845 (make_pass_feedback_split_functions): ...new function.
12846 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12847 global struct to a subclass of simple_ipa_opt_pass along with...
12848 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12850 (make_pass_ipa_function_and_variable_visibility): ...new function.
12851 (pass_ipa_free_inline_summary): Convert from a global struct to a
12852 subclass of simple_ipa_opt_pass along with...
12853 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12854 (make_pass_ipa_free_inline_summary): ...new function.
12855 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12856 subclass of ipa_opt_pass_d along with...
12857 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12859 (make_pass_ipa_whole_program_visibility): ...new function.
12860 (pass_ipa_profile): Convert from a global struct to a subclass of
12861 ipa_opt_pass_d along with...
12862 (pass_data_ipa_profile): ...new pass_data instance and...
12863 (make_pass_ipa_profile): ...new function.
12864 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12865 ipa_opt_pass_d along with...
12866 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12867 (make_pass_ipa_cdtor_merge): ...new function.
12868 * ira.c (pass_ira): Convert from a global struct to a subclass of
12869 rtl_opt_pass along with...
12870 (pass_data_ira): ...new pass_data instance and...
12871 (make_pass_ira): ...new function.
12872 (pass_reload): Convert from a global struct to a subclass of
12873 rtl_opt_pass along with...
12874 (pass_data_reload): ...new pass_data instance and...
12875 (make_pass_reload): ...new function.
12876 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12877 subclass of rtl_opt_pass along with...
12878 (pass_data_cleanup_barriers): ...new pass_data instance and...
12879 (make_pass_cleanup_barriers): ...new function.
12880 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12881 of rtl_opt_pass along with...
12882 (pass_data_loop2): ...new pass_data instance and...
12883 (make_pass_loop2): ...new function.
12884 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12885 rtl_opt_pass along with...
12886 (pass_data_rtl_loop_init): ...new pass_data instance and...
12887 (make_pass_rtl_loop_init): ...new function.
12888 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12889 rtl_opt_pass along with...
12890 (pass_data_rtl_loop_done): ...new pass_data instance and...
12891 (make_pass_rtl_loop_done): ...new function.
12892 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12893 subclass of rtl_opt_pass along with...
12894 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12895 (make_pass_rtl_move_loop_invariants): ...new function.
12896 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12897 rtl_opt_pass along with...
12898 (pass_data_rtl_unswitch): ...new pass_data instance and...
12899 (make_pass_rtl_unswitch): ...new function.
12900 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12901 subclass of rtl_opt_pass along with...
12902 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12904 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12905 (pass_rtl_doloop): Convert from a global struct to a subclass of
12906 rtl_opt_pass along with...
12907 (pass_data_rtl_doloop): ...new pass_data instance and...
12908 (make_pass_rtl_doloop): ...new function.
12909 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12910 a subclass of rtl_opt_pass along with...
12911 (pass_data_lower_subreg): ...new pass_data instance and...
12912 (make_pass_lower_subreg): ...new function.
12913 (pass_lower_subreg2): Convert from a global struct to a subclass of
12914 rtl_opt_pass along with...
12915 (pass_data_lower_subreg2): ...new pass_data instance and...
12916 (make_pass_lower_subreg2): ...new function.
12917 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12918 struct to a subclass of ipa_opt_pass_d along with...
12919 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12920 (make_pass_ipa_lto_gimple_out): ...new function.
12921 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12922 of ipa_opt_pass_d along with...
12923 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12924 (make_pass_ipa_lto_finish_out): ...new function.
12925 * mode-switching.c (pass_mode_switching): Convert from a global struct
12926 to a subclass of rtl_opt_pass along with...
12927 (pass_data_mode_switching): ...new pass_data instance and...
12928 (make_pass_mode_switching): ...new function.
12929 * modulo-sched.c (pass_sms): Convert from a global struct to a
12930 subclass of rtl_opt_pass along with...
12931 (pass_data_sms): ...new pass_data instance and...
12932 (make_pass_sms): ...new function.
12933 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12934 subclass of gimple_opt_pass along with...
12935 (pass_data_expand_omp): ...new pass_data instance and...
12936 (make_pass_expand_omp): ...new function.
12937 (pass_lower_omp): Convert from a global struct to a subclass of
12938 gimple_opt_pass along with...
12939 (pass_data_lower_omp): ...new pass_data instance and...
12940 (make_pass_lower_omp): ...new function.
12941 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12942 of gimple_opt_pass along with...
12943 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12944 (make_pass_diagnose_omp_blocks): ...new function.
12945 * passes.c (pass_early_local_passes): Convert from a global struct to
12946 a subclass of simple_ipa_opt_pass along with...
12947 (pass_data_early_local_passes): ...new pass_data instance and...
12948 (make_pass_early_local_passes): ...new function.
12949 (pass_all_early_optimizations): Convert from a global struct to a
12950 subclass of gimple_opt_pass along with...
12951 (pass_data_all_early_optimizations): ...new pass_data instance and...
12952 (make_pass_all_early_optimizations): ...new function.
12953 (pass_all_optimizations): Convert from a global struct to a subclass
12954 of gimple_opt_pass along with...
12955 (pass_data_all_optimizations): ...new pass_data instance and...
12956 (make_pass_all_optimizations): ...new function.
12957 (pass_all_optimizations_g): Convert from a global struct to a subclass
12958 of gimple_opt_pass along with...
12959 (pass_data_all_optimizations_g): ...new pass_data instance and...
12960 (make_pass_all_optimizations_g): ...new function.
12961 (pass_rest_of_compilation): Convert from a global struct to a subclass
12962 of rtl_opt_pass along with...
12963 (pass_data_rest_of_compilation): ...new pass_data instance and...
12964 (make_pass_rest_of_compilation): ...new function.
12965 (pass_postreload): Convert from a global struct to a subclass of
12966 rtl_opt_pass along with...
12967 (pass_data_postreload): ...new pass_data instance and...
12968 (make_pass_postreload): ...new function.
12969 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12970 subclass of rtl_opt_pass along with...
12971 (pass_data_gcse2): ...new pass_data instance and...
12972 (make_pass_gcse2): ...new function.
12973 * postreload.c (pass_postreload_cse): Convert from a global struct to
12974 a subclass of rtl_opt_pass along with...
12975 (pass_data_postreload_cse): ...new pass_data instance and...
12976 (make_pass_postreload_cse): ...new function.
12977 * predict.c (pass_profile): Convert from a global struct to a subclass
12978 of gimple_opt_pass along with...
12979 (pass_data_profile): ...new pass_data instance and...
12980 (make_pass_profile): ...new function.
12981 (pass_strip_predict_hints): Convert from a global struct to a subclass
12982 of gimple_opt_pass along with...
12983 (pass_data_strip_predict_hints): ...new pass_data instance and...
12984 (make_pass_strip_predict_hints): ...new function.
12985 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12986 of rtl_opt_pass along with...
12987 (pass_data_peephole2): ...new pass_data instance and...
12988 (make_pass_peephole2): ...new function.
12989 (pass_split_all_insns): Convert from a global struct to a subclass of
12990 rtl_opt_pass along with...
12991 (pass_data_split_all_insns): ...new pass_data instance and...
12992 (make_pass_split_all_insns): ...new function.
12993 (pass_split_after_reload): Convert from a global struct to a subclass
12994 of rtl_opt_pass along with...
12995 (pass_data_split_after_reload): ...new pass_data instance and...
12996 (make_pass_split_after_reload): ...new function.
12997 (pass_split_before_regstack): Convert from a global struct to a
12998 subclass of rtl_opt_pass along with...
12999 (pass_data_split_before_regstack): ...new pass_data instance and...
13000 (make_pass_split_before_regstack): ...new function.
13001 (pass_split_before_sched2): Convert from a global struct to a subclass
13002 of rtl_opt_pass along with...
13003 (pass_data_split_before_sched2): ...new pass_data instance and...
13004 (make_pass_split_before_sched2): ...new function.
13005 (pass_split_for_shorten_branches): Convert from a global struct to a
13006 subclass of rtl_opt_pass along with...
13007 (pass_data_split_for_shorten_branches): ...new pass_data instance
13009 (make_pass_split_for_shorten_branches): ...new function.
13010 * ree.c (pass_ree): Convert from a global struct to a subclass of
13011 rtl_opt_pass along with...
13012 (pass_data_ree): ...new pass_data instance and...
13013 (make_pass_ree): ...new function.
13014 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13015 subclass of rtl_opt_pass along with...
13016 (pass_data_stack_regs): ...new pass_data instance and...
13017 (make_pass_stack_regs): ...new function.
13018 (pass_stack_regs_run): Convert from a global struct to a subclass of
13019 rtl_opt_pass along with...
13020 (pass_data_stack_regs_run): ...new pass_data instance and...
13021 (make_pass_stack_regs_run): ...new function.
13022 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13023 subclass of rtl_opt_pass along with...
13024 (pass_data_cprop_hardreg): ...new pass_data instance and...
13025 (make_pass_cprop_hardreg): ...new function.
13026 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13027 subclass of rtl_opt_pass along with...
13028 (pass_data_reginfo_init): ...new pass_data instance and...
13029 (make_pass_reginfo_init): ...new function.
13030 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13031 of rtl_opt_pass along with...
13032 (pass_data_regmove): ...new pass_data instance and...
13033 (make_pass_regmove): ...new function.
13034 * regrename.c (pass_regrename): Convert from a global struct to a
13035 subclass of rtl_opt_pass along with...
13036 (pass_data_regrename): ...new pass_data instance and...
13037 (make_pass_regrename): ...new function.
13038 * reorg.c (pass_delay_slots): Convert from a global struct to a
13039 subclass of rtl_opt_pass along with...
13040 (pass_data_delay_slots): ...new pass_data instance and...
13041 (make_pass_delay_slots): ...new function.
13042 (pass_machine_reorg): Convert from a global struct to a subclass of
13043 rtl_opt_pass along with...
13044 (pass_data_machine_reorg): ...new pass_data instance and...
13045 (make_pass_machine_reorg): ...new function.
13046 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13047 of rtl_opt_pass along with...
13048 (pass_data_sched): ...new pass_data instance and...
13049 (make_pass_sched): ...new function.
13050 (pass_sched2): Convert from a global struct to a subclass of
13051 rtl_opt_pass along with...
13052 (pass_data_sched2): ...new pass_data instance and...
13053 (make_pass_sched2): ...new function.
13054 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13055 to a subclass of rtl_opt_pass along with...
13056 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13057 (make_pass_stack_ptr_mod): ...new function.
13058 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13059 to a subclass of rtl_opt_pass along with...
13060 (pass_data_rtl_store_motion): ...new pass_data instance and...
13061 (make_pass_rtl_store_motion): ...new function.
13062 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13063 of gimple_opt_pass along with...
13064 (pass_data_tracer): ...new pass_data instance and...
13065 (make_pass_tracer): ...new function.
13066 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13067 to a subclass of gimple_opt_pass along with...
13068 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13069 (make_pass_diagnose_tm_blocks): ...new function.
13070 (pass_lower_tm): Convert from a global struct to a subclass of
13071 gimple_opt_pass along with...
13072 (pass_data_lower_tm): ...new pass_data instance and...
13073 (make_pass_lower_tm): ...new function.
13074 (pass_tm_init): Convert from a global struct to a subclass of
13075 gimple_opt_pass along with...
13076 (pass_data_tm_init): ...new pass_data instance and...
13077 (make_pass_tm_init): ...new function.
13078 (pass_tm_mark): Convert from a global struct to a subclass of
13079 gimple_opt_pass along with...
13080 (pass_data_tm_mark): ...new pass_data instance and...
13081 (make_pass_tm_mark): ...new function.
13082 (pass_tm_edges): Convert from a global struct to a subclass of
13083 gimple_opt_pass along with...
13084 (pass_data_tm_edges): ...new pass_data instance and...
13085 (make_pass_tm_edges): ...new function.
13086 (pass_tm_memopt): Convert from a global struct to a subclass of
13087 gimple_opt_pass along with...
13088 (pass_data_tm_memopt): ...new pass_data instance and...
13089 (make_pass_tm_memopt): ...new function.
13090 (pass_ipa_tm): Convert from a global struct to a subclass of
13091 simple_ipa_opt_pass along with...
13092 (pass_data_ipa_tm): ...new pass_data instance and...
13093 (make_pass_ipa_tm): ...new function.
13094 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13095 subclass of gimple_opt_pass along with...
13096 (pass_data_call_cdce): ...new pass_data instance and...
13097 (make_pass_call_cdce): ...new function.
13098 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13099 subclass of gimple_opt_pass along with...
13100 (pass_data_build_cfg): ...new pass_data instance and...
13101 (make_pass_build_cfg): ...new function.
13102 (pass_split_crit_edges): Convert from a global struct to a subclass of
13103 gimple_opt_pass along with...
13104 (pass_data_split_crit_edges): ...new pass_data instance and...
13105 (make_pass_split_crit_edges): ...new function.
13106 (pass_warn_function_return): Convert from a global struct to a
13107 subclass of gimple_opt_pass along with...
13108 (pass_data_warn_function_return): ...new pass_data instance and...
13109 (make_pass_warn_function_return): ...new function.
13110 (pass_warn_function_noreturn): Convert from a global struct to a
13111 subclass of gimple_opt_pass along with...
13112 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13113 (make_pass_warn_function_noreturn): ...new function.
13114 (pass_warn_unused_result): Convert from a global struct to a subclass
13115 of gimple_opt_pass along with...
13116 (pass_data_warn_unused_result): ...new pass_data instance and...
13117 (make_pass_warn_unused_result): ...new function.
13118 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13119 a subclass of gimple_opt_pass along with...
13120 (pass_data_merge_phi): ...new pass_data instance and...
13121 (make_pass_merge_phi): ...new function.
13122 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13123 a subclass of gimple_opt_pass along with...
13124 (pass_data_lower_complex): ...new pass_data instance and...
13125 (make_pass_lower_complex): ...new function.
13126 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13127 gimple_opt_pass along with...
13128 (pass_data_lower_complex_O0): ...new pass_data instance and...
13129 (make_pass_lower_complex_O0): ...new function.
13130 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13131 subclass of gimple_opt_pass along with...
13132 (pass_data_lower_eh): ...new pass_data instance and...
13133 (make_pass_lower_eh): ...new function.
13134 (pass_refactor_eh): Convert from a global struct to a subclass of
13135 gimple_opt_pass along with...
13136 (pass_data_refactor_eh): ...new pass_data instance and...
13137 (make_pass_refactor_eh): ...new function.
13138 (pass_lower_resx): Convert from a global struct to a subclass of
13139 gimple_opt_pass along with...
13140 (pass_data_lower_resx): ...new pass_data instance and...
13141 (make_pass_lower_resx): ...new function.
13142 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13143 of gimple_opt_pass along with...
13144 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13145 (make_pass_lower_eh_dispatch): ...new function.
13146 (pass_cleanup_eh): Convert from a global struct to a subclass of
13147 gimple_opt_pass along with...
13148 (pass_data_cleanup_eh): ...new pass_data instance and...
13149 (make_pass_cleanup_eh): ...new function.
13150 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13151 to a subclass of simple_ipa_opt_pass along with...
13152 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13153 (make_pass_ipa_lower_emutls): ...new function.
13154 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13155 a subclass of gimple_opt_pass along with...
13156 (pass_data_if_conversion): ...new pass_data instance and...
13157 (make_pass_if_conversion): ...new function.
13158 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13159 subclass of gimple_opt_pass along with...
13160 (pass_data_build_ssa): ...new pass_data instance and...
13161 (make_pass_build_ssa): ...new function.
13162 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13163 global struct to a subclass of gimple_opt_pass along with...
13164 (pass_data_loop_distribution): ...new pass_data instance and...
13165 (make_pass_loop_distribution): ...new function.
13166 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13167 subclass of gimple_opt_pass along with...
13168 (pass_data_mudflap_1): ...new pass_data instance and...
13169 (make_pass_mudflap_1): ...new function.
13170 (pass_mudflap_2): Convert from a global struct to a subclass of
13171 gimple_opt_pass along with...
13172 (pass_data_mudflap_2): ...new pass_data instance and...
13173 (make_pass_mudflap_2): ...new function.
13174 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13175 subclass of gimple_opt_pass along with...
13176 (pass_data_mudflap_1): ...new pass_data instance and...
13177 (make_pass_mudflap_1): ...new function.
13178 (pass_mudflap_2): Convert from a global struct to a subclass of
13179 gimple_opt_pass along with...
13180 (pass_data_mudflap_2): ...new pass_data instance and...
13181 (make_pass_mudflap_2): ...new function.
13182 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13183 gimple_opt_pass along with...
13184 (pass_data_nrv): ...new pass_data instance and...
13185 (make_pass_nrv): ...new function.
13186 (pass_return_slot): Convert from a global struct to a subclass of
13187 gimple_opt_pass along with...
13188 (pass_data_return_slot): ...new pass_data instance and...
13189 (make_pass_return_slot): ...new function.
13190 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13191 to a subclass of gimple_opt_pass along with...
13192 (pass_data_object_sizes): ...new pass_data instance and...
13193 (make_pass_object_sizes): ...new function.
13194 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13195 global struct to a subclass of gimple_opt_pass along with...
13196 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13198 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13199 (pass_fixup_cfg): Convert from a global struct to a subclass of
13200 gimple_opt_pass along with...
13201 (pass_data_fixup_cfg): ...new pass_data instance and...
13202 (make_pass_fixup_cfg): ...new function.
13203 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13204 (make_pass_mudflap_1): ...new function.
13205 (pass_mudflap_2): Replace declaration with that of...
13206 (make_pass_mudflap_2): ...new function.
13207 (pass_asan): Replace declaration with that of...
13208 (make_pass_asan): ...new function.
13209 (pass_asan_O0): Replace declaration with that of...
13210 (make_pass_asan_O0): ...new function.
13211 (pass_tsan): Replace declaration with that of...
13212 (make_pass_tsan): ...new function.
13213 (pass_tsan_O0): Replace declaration with that of...
13214 (make_pass_tsan_O0): ...new function.
13215 (pass_lower_cf): Replace declaration with that of...
13216 (make_pass_lower_cf): ...new function.
13217 (pass_refactor_eh): Replace declaration with that of...
13218 (make_pass_refactor_eh): ...new function.
13219 (pass_lower_eh): Replace declaration with that of...
13220 (make_pass_lower_eh): ...new function.
13221 (pass_lower_eh_dispatch): Replace declaration with that of...
13222 (make_pass_lower_eh_dispatch): ...new function.
13223 (pass_lower_resx): Replace declaration with that of...
13224 (make_pass_lower_resx): ...new function.
13225 (pass_build_cfg): Replace declaration with that of...
13226 (make_pass_build_cfg): ...new function.
13227 (pass_early_tree_profile): Replace declaration with that of...
13228 (make_pass_early_tree_profile): ...new function.
13229 (pass_cleanup_eh): Replace declaration with that of...
13230 (make_pass_cleanup_eh): ...new function.
13231 (pass_sra): Replace declaration with that of...
13232 (make_pass_sra): ...new function.
13233 (pass_sra_early): Replace declaration with that of...
13234 (make_pass_sra_early): ...new function.
13235 (pass_early_ipa_sra): Replace declaration with that of...
13236 (make_pass_early_ipa_sra): ...new function.
13237 (pass_tail_recursion): Replace declaration with that of...
13238 (make_pass_tail_recursion): ...new function.
13239 (pass_tail_calls): Replace declaration with that of...
13240 (make_pass_tail_calls): ...new function.
13241 (pass_tree_loop): Replace declaration with that of...
13242 (make_pass_tree_loop): ...new function.
13243 (pass_tree_loop_init): Replace declaration with that of...
13244 (make_pass_tree_loop_init): ...new function.
13245 (pass_lim): Replace declaration with that of...
13246 (make_pass_lim): ...new function.
13247 (pass_tree_unswitch): Replace declaration with that of...
13248 (make_pass_tree_unswitch): ...new function.
13249 (pass_predcom): Replace declaration with that of...
13250 (make_pass_predcom): ...new function.
13251 (pass_iv_canon): Replace declaration with that of...
13252 (make_pass_iv_canon): ...new function.
13253 (pass_scev_cprop): Replace declaration with that of...
13254 (make_pass_scev_cprop): ...new function.
13255 (pass_empty_loop): Replace declaration with that of...
13256 (make_pass_empty_loop): ...new function.
13257 (pass_record_bounds): Replace declaration with that of...
13258 (make_pass_record_bounds): ...new function.
13259 (pass_graphite): Replace declaration with that of...
13260 (make_pass_graphite): ...new function.
13261 (pass_graphite_transforms): Replace declaration with that of...
13262 (make_pass_graphite_transforms): ...new function.
13263 (pass_if_conversion): Replace declaration with that of...
13264 (make_pass_if_conversion): ...new function.
13265 (pass_loop_distribution): Replace declaration with that of...
13266 (make_pass_loop_distribution): ...new function.
13267 (pass_vectorize): Replace declaration with that of...
13268 (make_pass_vectorize): ...new function.
13269 (pass_slp_vectorize): Replace declaration with that of...
13270 (make_pass_slp_vectorize): ...new function.
13271 (pass_complete_unroll): Replace declaration with that of...
13272 (make_pass_complete_unroll): ...new function.
13273 (pass_complete_unrolli): Replace declaration with that of...
13274 (make_pass_complete_unrolli): ...new function.
13275 (pass_parallelize_loops): Replace declaration with that of...
13276 (make_pass_parallelize_loops): ...new function.
13277 (pass_loop_prefetch): Replace declaration with that of...
13278 (make_pass_loop_prefetch): ...new function.
13279 (pass_iv_optimize): Replace declaration with that of...
13280 (make_pass_iv_optimize): ...new function.
13281 (pass_tree_loop_done): Replace declaration with that of...
13282 (make_pass_tree_loop_done): ...new function.
13283 (pass_ch): Replace declaration with that of...
13284 (make_pass_ch): ...new function.
13285 (pass_ccp): Replace declaration with that of...
13286 (make_pass_ccp): ...new function.
13287 (pass_phi_only_cprop): Replace declaration with that of...
13288 (make_pass_phi_only_cprop): ...new function.
13289 (pass_build_ssa): Replace declaration with that of...
13290 (make_pass_build_ssa): ...new function.
13291 (pass_build_alias): Replace declaration with that of...
13292 (make_pass_build_alias): ...new function.
13293 (pass_build_ealias): Replace declaration with that of...
13294 (make_pass_build_ealias): ...new function.
13295 (pass_dominator): Replace declaration with that of...
13296 (make_pass_dominator): ...new function.
13297 (pass_dce): Replace declaration with that of...
13298 (make_pass_dce): ...new function.
13299 (pass_dce_loop): Replace declaration with that of...
13300 (make_pass_dce_loop): ...new function.
13301 (pass_cd_dce): Replace declaration with that of...
13302 (make_pass_cd_dce): ...new function.
13303 (pass_call_cdce): Replace declaration with that of...
13304 (make_pass_call_cdce): ...new function.
13305 (pass_merge_phi): Replace declaration with that of...
13306 (make_pass_merge_phi): ...new function.
13307 (pass_split_crit_edges): Replace declaration with that of...
13308 (make_pass_split_crit_edges): ...new function.
13309 (pass_pre): Replace declaration with that of...
13310 (make_pass_pre): ...new function.
13311 (pass_profile): Replace declaration with that of...
13312 (make_pass_profile): ...new function.
13313 (pass_strip_predict_hints): Replace declaration with that of...
13314 (make_pass_strip_predict_hints): ...new function.
13315 (pass_lower_complex_O0): Replace declaration with that of...
13316 (make_pass_lower_complex_O0): ...new function.
13317 (pass_lower_complex): Replace declaration with that of...
13318 (make_pass_lower_complex): ...new function.
13319 (pass_lower_vector): Replace declaration with that of...
13320 (make_pass_lower_vector): ...new function.
13321 (pass_lower_vector_ssa): Replace declaration with that of...
13322 (make_pass_lower_vector_ssa): ...new function.
13323 (pass_lower_omp): Replace declaration with that of...
13324 (make_pass_lower_omp): ...new function.
13325 (pass_diagnose_omp_blocks): Replace declaration with that of...
13326 (make_pass_diagnose_omp_blocks): ...new function.
13327 (pass_expand_omp): Replace declaration with that of...
13328 (make_pass_expand_omp): ...new function.
13329 (pass_expand_omp_ssa): Replace declaration with that of...
13330 (make_pass_expand_omp_ssa): ...new function.
13331 (pass_object_sizes): Replace declaration with that of...
13332 (make_pass_object_sizes): ...new function.
13333 (pass_strlen): Replace declaration with that of...
13334 (make_pass_strlen): ...new function.
13335 (pass_fold_builtins): Replace declaration with that of...
13336 (make_pass_fold_builtins): ...new function.
13337 (pass_stdarg): Replace declaration with that of...
13338 (make_pass_stdarg): ...new function.
13339 (pass_early_warn_uninitialized): Replace declaration with that of...
13340 (make_pass_early_warn_uninitialized): ...new function.
13341 (pass_late_warn_uninitialized): Replace declaration with that of...
13342 (make_pass_late_warn_uninitialized): ...new function.
13343 (pass_cse_reciprocals): Replace declaration with that of...
13344 (make_pass_cse_reciprocals): ...new function.
13345 (pass_cse_sincos): Replace declaration with that of...
13346 (make_pass_cse_sincos): ...new function.
13347 (pass_optimize_bswap): Replace declaration with that of...
13348 (make_pass_optimize_bswap): ...new function.
13349 (pass_optimize_widening_mul): Replace declaration with that of...
13350 (make_pass_optimize_widening_mul): ...new function.
13351 (pass_warn_function_return): Replace declaration with that of...
13352 (make_pass_warn_function_return): ...new function.
13353 (pass_warn_function_noreturn): Replace declaration with that of...
13354 (make_pass_warn_function_noreturn): ...new function.
13355 (pass_cselim): Replace declaration with that of...
13356 (make_pass_cselim): ...new function.
13357 (pass_phiopt): Replace declaration with that of...
13358 (make_pass_phiopt): ...new function.
13359 (pass_forwprop): Replace declaration with that of...
13360 (make_pass_forwprop): ...new function.
13361 (pass_phiprop): Replace declaration with that of...
13362 (make_pass_phiprop): ...new function.
13363 (pass_tree_ifcombine): Replace declaration with that of...
13364 (make_pass_tree_ifcombine): ...new function.
13365 (pass_dse): Replace declaration with that of...
13366 (make_pass_dse): ...new function.
13367 (pass_nrv): Replace declaration with that of...
13368 (make_pass_nrv): ...new function.
13369 (pass_rename_ssa_copies): Replace declaration with that of...
13370 (make_pass_rename_ssa_copies): ...new function.
13371 (pass_sink_code): Replace declaration with that of...
13372 (make_pass_sink_code): ...new function.
13373 (pass_fre): Replace declaration with that of...
13374 (make_pass_fre): ...new function.
13375 (pass_check_data_deps): Replace declaration with that of...
13376 (make_pass_check_data_deps): ...new function.
13377 (pass_copy_prop): Replace declaration with that of...
13378 (make_pass_copy_prop): ...new function.
13379 (pass_vrp): Replace declaration with that of...
13380 (make_pass_vrp): ...new function.
13381 (pass_uncprop): Replace declaration with that of...
13382 (make_pass_uncprop): ...new function.
13383 (pass_return_slot): Replace declaration with that of...
13384 (make_pass_return_slot): ...new function.
13385 (pass_reassoc): Replace declaration with that of...
13386 (make_pass_reassoc): ...new function.
13387 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13388 (make_pass_rebuild_cgraph_edges): ...new function.
13389 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13390 (make_pass_remove_cgraph_callee_edges): ...new function.
13391 (pass_build_cgraph_edges): Replace declaration with that of...
13392 (make_pass_build_cgraph_edges): ...new function.
13393 (pass_local_pure_const): Replace declaration with that of...
13394 (make_pass_local_pure_const): ...new function.
13395 (pass_tracer): Replace declaration with that of...
13396 (make_pass_tracer): ...new function.
13397 (pass_warn_unused_result): Replace declaration with that of...
13398 (make_pass_warn_unused_result): ...new function.
13399 (pass_diagnose_tm_blocks): Replace declaration with that of...
13400 (make_pass_diagnose_tm_blocks): ...new function.
13401 (pass_lower_tm): Replace declaration with that of...
13402 (make_pass_lower_tm): ...new function.
13403 (pass_tm_init): Replace declaration with that of...
13404 (make_pass_tm_init): ...new function.
13405 (pass_tm_mark): Replace declaration with that of...
13406 (make_pass_tm_mark): ...new function.
13407 (pass_tm_memopt): Replace declaration with that of...
13408 (make_pass_tm_memopt): ...new function.
13409 (pass_tm_edges): Replace declaration with that of...
13410 (make_pass_tm_edges): ...new function.
13411 (pass_split_functions): Replace declaration with that of...
13412 (make_pass_split_functions): ...new function.
13413 (pass_feedback_split_functions): Replace declaration with that of...
13414 (make_pass_feedback_split_functions): ...new function.
13415 (pass_strength_reduction): Replace declaration with that of...
13416 (make_pass_strength_reduction): ...new function.
13417 (pass_ipa_lower_emutls): Replace declaration with that of...
13418 (make_pass_ipa_lower_emutls): ...new function.
13419 (pass_ipa_function_and_variable_visibility): Replace declaration with
13421 (make_pass_ipa_function_and_variable_visibility): ...new function.
13422 (pass_ipa_tree_profile): Replace declaration with that of...
13423 (make_pass_ipa_tree_profile): ...new function.
13424 (pass_early_local_passes): Replace declaration with that of...
13425 (make_pass_early_local_passes): ...new function.
13426 (pass_ipa_whole_program_visibility): Replace declaration with that
13428 (make_pass_ipa_whole_program_visibility): ...new function.
13429 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13430 (make_pass_ipa_lto_gimple_out): ...new function.
13431 (pass_ipa_increase_alignment): Replace declaration with that of...
13432 (make_pass_ipa_increase_alignment): ...new function.
13433 (pass_ipa_inline): Replace declaration with that of...
13434 (make_pass_ipa_inline): ...new function.
13435 (pass_ipa_free_lang_data): Replace declaration with that of...
13436 (make_pass_ipa_free_lang_data): ...new function.
13437 (pass_ipa_free_inline_summary): Replace declaration with that of...
13438 (make_pass_ipa_free_inline_summary): ...new function.
13439 (pass_ipa_cp): Replace declaration with that of...
13440 (make_pass_ipa_cp): ...new function.
13441 (pass_ipa_reference): Replace declaration with that of...
13442 (make_pass_ipa_reference): ...new function.
13443 (pass_ipa_pure_const): Replace declaration with that of...
13444 (make_pass_ipa_pure_const): ...new function.
13445 (pass_ipa_pta): Replace declaration with that of...
13446 (make_pass_ipa_pta): ...new function.
13447 (pass_ipa_lto_finish_out): Replace declaration with that of...
13448 (make_pass_ipa_lto_finish_out): ...new function.
13449 (pass_ipa_tm): Replace declaration with that of...
13450 (make_pass_ipa_tm): ...new function.
13451 (pass_ipa_profile): Replace declaration with that of...
13452 (make_pass_ipa_profile): ...new function.
13453 (pass_ipa_cdtor_merge): Replace declaration with that of...
13454 (make_pass_ipa_cdtor_merge): ...new function.
13455 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13457 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13458 (pass_init_datastructures): Replace declaration with that of...
13459 (make_pass_init_datastructures): ...new function.
13460 (pass_fixup_cfg): Replace declaration with that of...
13461 (make_pass_fixup_cfg): ...new function.
13462 (pass_expand): Replace declaration with that of...
13463 (make_pass_expand): ...new function.
13464 (pass_instantiate_virtual_regs): Replace declaration with that of...
13465 (make_pass_instantiate_virtual_regs): ...new function.
13466 (pass_rtl_fwprop): Replace declaration with that of...
13467 (make_pass_rtl_fwprop): ...new function.
13468 (pass_rtl_fwprop_addr): Replace declaration with that of...
13469 (make_pass_rtl_fwprop_addr): ...new function.
13470 (pass_jump): Replace declaration with that of...
13471 (make_pass_jump): ...new function.
13472 (pass_jump2): Replace declaration with that of...
13473 (make_pass_jump2): ...new function.
13474 (pass_lower_subreg): Replace declaration with that of...
13475 (make_pass_lower_subreg): ...new function.
13476 (pass_cse): Replace declaration with that of...
13477 (make_pass_cse): ...new function.
13478 (pass_fast_rtl_dce): Replace declaration with that of...
13479 (make_pass_fast_rtl_dce): ...new function.
13480 (pass_ud_rtl_dce): Replace declaration with that of...
13481 (make_pass_ud_rtl_dce): ...new function.
13482 (pass_rtl_dce): Replace declaration with that of...
13483 (make_pass_rtl_dce): ...new function.
13484 (pass_rtl_dse1): Replace declaration with that of...
13485 (make_pass_rtl_dse1): ...new function.
13486 (pass_rtl_dse2): Replace declaration with that of...
13487 (make_pass_rtl_dse2): ...new function.
13488 (pass_rtl_dse3): Replace declaration with that of...
13489 (make_pass_rtl_dse3): ...new function.
13490 (pass_rtl_cprop): Replace declaration with that of...
13491 (make_pass_rtl_cprop): ...new function.
13492 (pass_rtl_pre): Replace declaration with that of...
13493 (make_pass_rtl_pre): ...new function.
13494 (pass_rtl_hoist): Replace declaration with that of...
13495 (make_pass_rtl_hoist): ...new function.
13496 (pass_rtl_store_motion): Replace declaration with that of...
13497 (make_pass_rtl_store_motion): ...new function.
13498 (pass_cse_after_global_opts): Replace declaration with that of...
13499 (make_pass_cse_after_global_opts): ...new function.
13500 (pass_rtl_ifcvt): Replace declaration with that of...
13501 (make_pass_rtl_ifcvt): ...new function.
13502 (pass_into_cfg_layout_mode): Replace declaration with that of...
13503 (make_pass_into_cfg_layout_mode): ...new function.
13504 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13505 (make_pass_outof_cfg_layout_mode): ...new function.
13506 (pass_loop2): Replace declaration with that of...
13507 (make_pass_loop2): ...new function.
13508 (pass_rtl_loop_init): Replace declaration with that of...
13509 (make_pass_rtl_loop_init): ...new function.
13510 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13511 (make_pass_rtl_move_loop_invariants): ...new function.
13512 (pass_rtl_unswitch): Replace declaration with that of...
13513 (make_pass_rtl_unswitch): ...new function.
13514 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13515 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13516 (pass_rtl_doloop): Replace declaration with that of...
13517 (make_pass_rtl_doloop): ...new function.
13518 (pass_rtl_loop_done): Replace declaration with that of...
13519 (make_pass_rtl_loop_done): ...new function.
13520 (pass_web): Replace declaration with that of...
13521 (make_pass_web): ...new function.
13522 (pass_cse2): Replace declaration with that of...
13523 (make_pass_cse2): ...new function.
13524 (pass_df_initialize_opt): Replace declaration with that of...
13525 (make_pass_df_initialize_opt): ...new function.
13526 (pass_df_initialize_no_opt): Replace declaration with that of...
13527 (make_pass_df_initialize_no_opt): ...new function.
13528 (pass_reginfo_init): Replace declaration with that of...
13529 (make_pass_reginfo_init): ...new function.
13530 (pass_inc_dec): Replace declaration with that of...
13531 (make_pass_inc_dec): ...new function.
13532 (pass_stack_ptr_mod): Replace declaration with that of...
13533 (make_pass_stack_ptr_mod): ...new function.
13534 (pass_initialize_regs): Replace declaration with that of...
13535 (make_pass_initialize_regs): ...new function.
13536 (pass_combine): Replace declaration with that of...
13537 (make_pass_combine): ...new function.
13538 (pass_if_after_combine): Replace declaration with that of...
13539 (make_pass_if_after_combine): ...new function.
13540 (pass_ree): Replace declaration with that of...
13541 (make_pass_ree): ...new function.
13542 (pass_partition_blocks): Replace declaration with that of...
13543 (make_pass_partition_blocks): ...new function.
13544 (pass_match_asm_constraints): Replace declaration with that of...
13545 (make_pass_match_asm_constraints): ...new function.
13546 (pass_regmove): Replace declaration with that of...
13547 (make_pass_regmove): ...new function.
13548 (pass_split_all_insns): Replace declaration with that of...
13549 (make_pass_split_all_insns): ...new function.
13550 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13551 (make_pass_fast_rtl_byte_dce): ...new function.
13552 (pass_lower_subreg2): Replace declaration with that of...
13553 (make_pass_lower_subreg2): ...new function.
13554 (pass_mode_switching): Replace declaration with that of...
13555 (make_pass_mode_switching): ...new function.
13556 (pass_sms): Replace declaration with that of...
13557 (make_pass_sms): ...new function.
13558 (pass_sched): Replace declaration with that of...
13559 (make_pass_sched): ...new function.
13560 (pass_ira): Replace declaration with that of...
13561 (make_pass_ira): ...new function.
13562 (pass_reload): Replace declaration with that of...
13563 (make_pass_reload): ...new function.
13564 (pass_clean_state): Replace declaration with that of...
13565 (make_pass_clean_state): ...new function.
13566 (pass_branch_prob): Replace declaration with that of...
13567 (make_pass_branch_prob): ...new function.
13568 (pass_value_profile_transformations): Replace declaration with that
13570 (make_pass_value_profile_transformations): ...new function.
13571 (pass_postreload_cse): Replace declaration with that of...
13572 (make_pass_postreload_cse): ...new function.
13573 (pass_gcse2): Replace declaration with that of...
13574 (make_pass_gcse2): ...new function.
13575 (pass_split_after_reload): Replace declaration with that of...
13576 (make_pass_split_after_reload): ...new function.
13577 (pass_branch_target_load_optimize1): Replace declaration with that
13579 (make_pass_branch_target_load_optimize1): ...new function.
13580 (pass_thread_prologue_and_epilogue): Replace declaration with that
13582 (make_pass_thread_prologue_and_epilogue): ...new function.
13583 (pass_stack_adjustments): Replace declaration with that of...
13584 (make_pass_stack_adjustments): ...new function.
13585 (pass_peephole2): Replace declaration with that of...
13586 (make_pass_peephole2): ...new function.
13587 (pass_if_after_reload): Replace declaration with that of...
13588 (make_pass_if_after_reload): ...new function.
13589 (pass_regrename): Replace declaration with that of...
13590 (make_pass_regrename): ...new function.
13591 (pass_cprop_hardreg): Replace declaration with that of...
13592 (make_pass_cprop_hardreg): ...new function.
13593 (pass_reorder_blocks): Replace declaration with that of...
13594 (make_pass_reorder_blocks): ...new function.
13595 (pass_branch_target_load_optimize2): Replace declaration with that
13597 (make_pass_branch_target_load_optimize2): ...new function.
13598 (pass_leaf_regs): Replace declaration with that of...
13599 (make_pass_leaf_regs): ...new function.
13600 (pass_split_before_sched2): Replace declaration with that of...
13601 (make_pass_split_before_sched2): ...new function.
13602 (pass_compare_elim_after_reload): Replace declaration with that of...
13603 (make_pass_compare_elim_after_reload): ...new function.
13604 (pass_sched2): Replace declaration with that of...
13605 (make_pass_sched2): ...new function.
13606 (pass_stack_regs): Replace declaration with that of...
13607 (make_pass_stack_regs): ...new function.
13608 (pass_stack_regs_run): Replace declaration with that of...
13609 (make_pass_stack_regs_run): ...new function.
13610 (pass_df_finish): Replace declaration with that of...
13611 (make_pass_df_finish): ...new function.
13612 (pass_compute_alignments): Replace declaration with that of...
13613 (make_pass_compute_alignments): ...new function.
13614 (pass_duplicate_computed_gotos): Replace declaration with that of...
13615 (make_pass_duplicate_computed_gotos): ...new function.
13616 (pass_variable_tracking): Replace declaration with that of...
13617 (make_pass_variable_tracking): ...new function.
13618 (pass_free_cfg): Replace declaration with that of...
13619 (make_pass_free_cfg): ...new function.
13620 (pass_machine_reorg): Replace declaration with that of...
13621 (make_pass_machine_reorg): ...new function.
13622 (pass_cleanup_barriers): Replace declaration with that of...
13623 (make_pass_cleanup_barriers): ...new function.
13624 (pass_delay_slots): Replace declaration with that of...
13625 (make_pass_delay_slots): ...new function.
13626 (pass_split_for_shorten_branches): Replace declaration with that of...
13627 (make_pass_split_for_shorten_branches): ...new function.
13628 (pass_split_before_regstack): Replace declaration with that of...
13629 (make_pass_split_before_regstack): ...new function.
13630 (pass_convert_to_eh_region_ranges): Replace declaration with that
13632 (make_pass_convert_to_eh_region_ranges): ...new function.
13633 (pass_shorten_branches): Replace declaration with that of...
13634 (make_pass_shorten_branches): ...new function.
13635 (pass_set_nothrow_function_flags): Replace declaration with that of...
13636 (make_pass_set_nothrow_function_flags): ...new function.
13637 (pass_dwarf2_frame): Replace declaration with that of...
13638 (make_pass_dwarf2_frame): ...new function.
13639 (pass_final): Replace declaration with that of...
13640 (make_pass_final): ...new function.
13641 (pass_rtl_seqabstr): Replace declaration with that of...
13642 (make_pass_rtl_seqabstr): ...new function.
13643 (pass_release_ssa_names): Replace declaration with that of...
13644 (make_pass_release_ssa_names): ...new function.
13645 (pass_early_inline): Replace declaration with that of...
13646 (make_pass_early_inline): ...new function.
13647 (pass_inline_parameters): Replace declaration with that of...
13648 (make_pass_inline_parameters): ...new function.
13649 (pass_update_address_taken): Replace declaration with that of...
13650 (make_pass_update_address_taken): ...new function.
13651 (pass_convert_switch): Replace declaration with that of...
13652 (make_pass_convert_switch): ...new function.
13653 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13654 to a subclass of simple_ipa_opt_pass along with...
13655 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13656 (make_pass_ipa_tree_profile): ...new function.
13657 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13658 subclass of gimple_opt_pass along with...
13659 (pass_data_sra_early): ...new pass_data instance and...
13660 (make_pass_sra_early): ...new function.
13661 (pass_sra): Convert from a global struct to a subclass of
13662 gimple_opt_pass along with...
13663 (pass_data_sra): ...new pass_data instance and...
13664 (make_pass_sra): ...new function.
13665 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13666 gimple_opt_pass along with...
13667 (pass_data_early_ipa_sra): ...new pass_data instance and...
13668 (make_pass_early_ipa_sra): ...new function.
13669 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13670 subclass of gimple_opt_pass along with...
13671 (pass_data_ccp): ...new pass_data instance and...
13672 (make_pass_ccp): ...new function.
13673 (pass_fold_builtins): Convert from a global struct to a subclass of
13674 gimple_opt_pass along with...
13675 (pass_data_fold_builtins): ...new pass_data instance and...
13676 (make_pass_fold_builtins): ...new function.
13677 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13678 subclass of gimple_opt_pass along with...
13679 (pass_data_copy_prop): ...new pass_data instance and...
13680 (make_pass_copy_prop): ...new function.
13681 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13682 global struct to a subclass of gimple_opt_pass along with...
13683 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13684 (make_pass_rename_ssa_copies): ...new function.
13685 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13686 subclass of gimple_opt_pass along with...
13687 (pass_data_dce): ...new pass_data instance and...
13688 (make_pass_dce): ...new function.
13689 (pass_dce_loop): Convert from a global struct to a subclass of
13690 gimple_opt_pass along with...
13691 (pass_data_dce_loop): ...new pass_data instance and...
13692 (make_pass_dce_loop): ...new function.
13693 (pass_cd_dce): Convert from a global struct to a subclass of
13694 gimple_opt_pass along with...
13695 (pass_data_cd_dce): ...new pass_data instance and...
13696 (make_pass_cd_dce): ...new function.
13697 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13698 subclass of gimple_opt_pass along with...
13699 (pass_data_dominator): ...new pass_data instance and...
13700 (make_pass_dominator): ...new function.
13701 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13702 gimple_opt_pass along with...
13703 (pass_data_phi_only_cprop): ...new pass_data instance and...
13704 (make_pass_phi_only_cprop): ...new function.
13705 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13706 subclass of gimple_opt_pass along with...
13707 (pass_data_dse): ...new pass_data instance and...
13708 (make_pass_dse): ...new function.
13709 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13710 a subclass of gimple_opt_pass along with...
13711 (pass_data_forwprop): ...new pass_data instance and...
13712 (make_pass_forwprop): ...new function.
13713 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13714 struct to a subclass of gimple_opt_pass along with...
13715 (pass_data_tree_ifcombine): ...new pass_data instance and...
13716 (make_pass_tree_ifcombine): ...new function.
13717 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13718 subclass of gimple_opt_pass along with...
13719 (pass_data_ch): ...new pass_data instance and...
13720 (make_pass_ch): ...new function.
13721 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13722 subclass of gimple_opt_pass along with...
13723 (pass_data_tree_loop): ...new pass_data instance and...
13724 (make_pass_tree_loop): ...new function.
13725 (pass_tree_loop_init): Convert from a global struct to a subclass of
13726 gimple_opt_pass along with...
13727 (pass_data_tree_loop_init): ...new pass_data instance and...
13728 (make_pass_tree_loop_init): ...new function.
13729 (pass_lim): Convert from a global struct to a subclass of
13730 gimple_opt_pass along with...
13731 (pass_data_lim): ...new pass_data instance and...
13732 (make_pass_lim): ...new function.
13733 (pass_tree_unswitch): Convert from a global struct to a subclass of
13734 gimple_opt_pass along with...
13735 (pass_data_tree_unswitch): ...new pass_data instance and...
13736 (make_pass_tree_unswitch): ...new function.
13737 (pass_predcom): Convert from a global struct to a subclass of
13738 gimple_opt_pass along with...
13739 (pass_data_predcom): ...new pass_data instance and...
13740 (make_pass_predcom): ...new function.
13741 (pass_vectorize): Convert from a global struct to a subclass of
13742 gimple_opt_pass along with...
13743 (pass_data_vectorize): ...new pass_data instance and...
13744 (make_pass_vectorize): ...new function.
13745 (pass_graphite): Convert from a global struct to a subclass of
13746 gimple_opt_pass along with...
13747 (pass_data_graphite): ...new pass_data instance and...
13748 (make_pass_graphite): ...new function.
13749 (pass_graphite_transforms): Convert from a global struct to a subclass
13750 of gimple_opt_pass along with...
13751 (pass_data_graphite_transforms): ...new pass_data instance and...
13752 (make_pass_graphite_transforms): ...new function.
13753 (pass_check_data_deps): Convert from a global struct to a subclass of
13754 gimple_opt_pass along with...
13755 (pass_data_check_data_deps): ...new pass_data instance and...
13756 (make_pass_check_data_deps): ...new function.
13757 (pass_iv_canon): Convert from a global struct to a subclass of
13758 gimple_opt_pass along with...
13759 (pass_data_iv_canon): ...new pass_data instance and...
13760 (make_pass_iv_canon): ...new function.
13761 (pass_scev_cprop): Convert from a global struct to a subclass of
13762 gimple_opt_pass along with...
13763 (pass_data_scev_cprop): ...new pass_data instance and...
13764 (make_pass_scev_cprop): ...new function.
13765 (pass_record_bounds): Convert from a global struct to a subclass of
13766 gimple_opt_pass along with...
13767 (pass_data_record_bounds): ...new pass_data instance and...
13768 (make_pass_record_bounds): ...new function.
13769 (pass_complete_unroll): Convert from a global struct to a subclass of
13770 gimple_opt_pass along with...
13771 (pass_data_complete_unroll): ...new pass_data instance and...
13772 (make_pass_complete_unroll): ...new function.
13773 (pass_complete_unrolli): Convert from a global struct to a subclass of
13774 gimple_opt_pass along with...
13775 (pass_data_complete_unrolli): ...new pass_data instance and...
13776 (make_pass_complete_unrolli): ...new function.
13777 (pass_parallelize_loops): Convert from a global struct to a subclass
13778 of gimple_opt_pass along with...
13779 (pass_data_parallelize_loops): ...new pass_data instance and...
13780 (make_pass_parallelize_loops): ...new function.
13781 (pass_loop_prefetch): Convert from a global struct to a subclass of
13782 gimple_opt_pass along with...
13783 (pass_data_loop_prefetch): ...new pass_data instance and...
13784 (make_pass_loop_prefetch): ...new function.
13785 (pass_iv_optimize): Convert from a global struct to a subclass of
13786 gimple_opt_pass along with...
13787 (pass_data_iv_optimize): ...new pass_data instance and...
13788 (make_pass_iv_optimize): ...new function.
13789 (pass_tree_loop_done): Convert from a global struct to a subclass of
13790 gimple_opt_pass along with...
13791 (pass_data_tree_loop_done): ...new pass_data instance and...
13792 (make_pass_tree_loop_done): ...new function.
13793 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13794 struct to a subclass of gimple_opt_pass along with...
13795 (pass_data_cse_reciprocals): ...new pass_data instance and...
13796 (make_pass_cse_reciprocals): ...new function.
13797 (pass_cse_sincos): Convert from a global struct to a subclass of
13798 gimple_opt_pass along with...
13799 (pass_data_cse_sincos): ...new pass_data instance and...
13800 (make_pass_cse_sincos): ...new function.
13801 (pass_optimize_bswap): Convert from a global struct to a subclass of
13802 gimple_opt_pass along with...
13803 (pass_data_optimize_bswap): ...new pass_data instance and...
13804 (make_pass_optimize_bswap): ...new function.
13805 (pass_optimize_widening_mul): Convert from a global struct to a
13806 subclass of gimple_opt_pass along with...
13807 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13808 (make_pass_optimize_widening_mul): ...new function.
13809 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13810 subclass of gimple_opt_pass along with...
13811 (pass_data_phiopt): ...new pass_data instance and...
13812 (make_pass_phiopt): ...new function.
13813 (pass_cselim): Convert from a global struct to a subclass of
13814 gimple_opt_pass along with...
13815 (pass_data_cselim): ...new pass_data instance and...
13816 (make_pass_cselim): ...new function.
13817 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13818 subclass of gimple_opt_pass along with...
13819 (pass_data_phiprop): ...new pass_data instance and...
13820 (make_pass_phiprop): ...new function.
13821 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13822 subclass of gimple_opt_pass along with...
13823 (pass_data_pre): ...new pass_data instance and...
13824 (make_pass_pre): ...new function.
13825 (pass_fre): Convert from a global struct to a subclass of
13826 gimple_opt_pass along with...
13827 (pass_data_fre): ...new pass_data instance and...
13828 (make_pass_fre): ...new function.
13829 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13830 subclass of gimple_opt_pass along with...
13831 (pass_data_reassoc): ...new pass_data instance and...
13832 (make_pass_reassoc): ...new function.
13833 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13834 subclass of gimple_opt_pass along with...
13835 (pass_data_sink_code): ...new pass_data instance and...
13836 (make_pass_sink_code): ...new function.
13837 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13838 subclass of gimple_opt_pass along with...
13839 (pass_data_strlen): ...new pass_data instance and...
13840 (make_pass_strlen): ...new function.
13841 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13842 struct to a subclass of gimple_opt_pass along with...
13843 (pass_data_build_alias): ...new pass_data instance and...
13844 (make_pass_build_alias): ...new function.
13845 (pass_build_ealias): Convert from a global struct to a subclass of
13846 gimple_opt_pass along with...
13847 (pass_data_build_ealias): ...new pass_data instance and...
13848 (make_pass_build_ealias): ...new function.
13849 (pass_ipa_pta): Convert from a global struct to a subclass of
13850 simple_ipa_opt_pass along with...
13851 (pass_data_ipa_pta): ...new pass_data instance and...
13852 (make_pass_ipa_pta): ...new function.
13853 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13854 subclass of gimple_opt_pass along with...
13855 (pass_data_uncprop): ...new pass_data instance and...
13856 (make_pass_uncprop): ...new function.
13857 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13858 global struct to a subclass of gimple_opt_pass along with...
13859 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13860 (make_pass_late_warn_uninitialized): ...new function.
13861 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13862 to a subclass of gimple_opt_pass along with...
13863 (pass_data_init_datastructures): ...new pass_data instance and...
13864 (make_pass_init_datastructures): ...new function.
13865 (pass_early_warn_uninitialized): Convert from a global struct to a
13866 subclass of gimple_opt_pass along with...
13867 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13868 (make_pass_early_warn_uninitialized): ...new function.
13869 (pass_update_address_taken): Convert from a global struct to a
13870 subclass of gimple_opt_pass along with...
13871 (pass_data_update_address_taken): ...new pass_data instance and...
13872 (make_pass_update_address_taken): ...new function.
13873 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13874 struct to a subclass of gimple_opt_pass along with...
13875 (pass_data_release_ssa_names): ...new pass_data instance and...
13876 (make_pass_release_ssa_names): ...new function.
13877 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13878 subclass of gimple_opt_pass along with...
13879 (pass_data_stdarg): ...new pass_data instance and...
13880 (make_pass_stdarg): ...new function.
13881 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13882 global struct to a subclass of gimple_opt_pass along with...
13883 (pass_data_convert_switch): ...new pass_data instance and...
13884 (make_pass_convert_switch): ...new function.
13885 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13886 to a subclass of gimple_opt_pass along with...
13887 (pass_data_tail_recursion): ...new pass_data instance and...
13888 (make_pass_tail_recursion): ...new function.
13889 (pass_tail_calls): Convert from a global struct to a subclass of
13890 gimple_opt_pass along with...
13891 (pass_data_tail_calls): ...new pass_data instance and...
13892 (make_pass_tail_calls): ...new function.
13893 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13894 struct to a subclass of gimple_opt_pass along with...
13895 (pass_data_lower_vector): ...new pass_data instance and...
13896 (make_pass_lower_vector): ...new function.
13897 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13898 gimple_opt_pass along with...
13899 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13900 (make_pass_lower_vector_ssa): ...new function.
13901 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13902 to a subclass of gimple_opt_pass along with...
13903 (pass_data_slp_vectorize): ...new pass_data instance and...
13904 (make_pass_slp_vectorize): ...new function.
13905 (pass_ipa_increase_alignment): Convert from a global struct to a
13906 subclass of simple_ipa_opt_pass along with...
13907 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13908 (make_pass_ipa_increase_alignment): ...new function.
13909 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13910 gimple_opt_pass along with...
13911 (pass_data_vrp): ...new pass_data instance and...
13912 (make_pass_vrp): ...new function.
13913 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13914 subclass of simple_ipa_opt_pass along with...
13915 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13916 (make_pass_ipa_free_lang_data): ...new function.
13917 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13918 gimple_opt_pass along with...
13919 (pass_data_tsan): ...new pass_data instance and...
13920 (make_pass_tsan): ...new function.
13921 (pass_tsan_O0): Convert from a global struct to a subclass of
13922 gimple_opt_pass along with...
13923 (pass_data_tsan_O0): ...new pass_data instance and...
13924 (make_pass_tsan_O0): ...new function.
13925 * var-tracking.c (pass_variable_tracking): Convert from a global
13926 struct to a subclass of rtl_opt_pass along with...
13927 (pass_data_variable_tracking): ...new pass_data instance and...
13928 (make_pass_variable_tracking): ...new function.
13929 * web.c (pass_web): Convert from a global struct to a subclass of
13930 rtl_opt_pass along with...
13931 (pass_data_web): ...new pass_data instance and...
13932 (make_pass_web): ...new function.
13933 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13934 declaration with that of...
13935 (make_pass_mode_switch_use): ...new function.
13936 (pass_resolve_sw_modes): Replace declaration with that of...
13937 (make_pass_resolve_sw_modes): ...new function.
13938 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13939 from a global struct to a subclass of rtl_opt_pass along with...
13940 (pass_data_mode_switch_use): ...new pass_data instance and...
13941 (make_pass_mode_switch_use): ...new function.
13942 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13943 from a global struct to a subclass of rtl_opt_pass along with...
13944 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13945 (make_pass_resolve_sw_modes): ...new function.
13946 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13947 struct to a subclass of rtl_opt_pass along with...
13948 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13949 (make_pass_insert_vzeroupper): ...new function.
13950 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13951 global struct to a subclass of rtl_opt_pass along with...
13952 (pass_data_work_around_errata): ...new pass_data instance and...
13953 (make_pass_work_around_errata): ...new function.
13954 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13955 struct to a subclass of rtl_opt_pass along with...
13956 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13957 (make_pass_mips_machine_reorg2): ...new function.
13959 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13961 * passes.c (pass_manager::operator new): New.
13963 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13965 Handwritten part of conversion of passes to C++ classes.
13967 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13968 (toplev.o): Add dep on PASS_MANAGER_H.
13969 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13970 of early local pases to reflect this moving from a global to a
13971 member of gcc::pass_manager.
13972 (cgraph_add_new_function): Likewise.
13973 * lto-cgraph.c (lto_output_node): Update for conversion of
13974 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13975 * passes.c (opt_pass::clone): New.
13976 (opt_pass::gate): New.
13977 (opt_pass::execute): New.
13978 (opt_pass::opt_pass): New.
13979 (pass_manager::execute_early_local_passes): New.
13980 (pass_manager::execute_pass_mode_switching): new.
13981 (finish_optimization_passes): Convert to...
13982 (pass_manager::finish_optimization_passes): ...this.
13983 (finish_optimization_passes): Update for conversion of passes to
13985 (register_dump_files_1): Use has_gate since we cannot portably
13986 check a vtable entry against NULL.
13987 (dump_one_pass): Likewise.
13988 (ipa_write_summaries_2): Likewise.
13989 (ipa_write_optimization_summaries_1): Likewise.
13990 (ipa_read_summaries_1): Likewise.
13991 (ipa_read_optimization_summaries_1): Likewise.
13992 (execute_ipa_stmt_fixups): Likewise.
13993 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13994 pass-creation functions rather than wiring up globals, and
13995 storing the results in fields of pass_manager generated using
13996 pass-instances.def.
13997 (pass_manager::dump_profile_report): Update for conversion of
13998 passes to C++ classes.
13999 (pass_manager::execute_ipa_summary_passes): Likewise.
14000 (execute_one_ipa_transform_pass): Likewise.
14001 (execute_one_pass): Use has_gate and has_execute since we cannot
14002 portably check a vtable entry against NULL.
14003 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14004 (pass_manager): Use pass-instances.def to add fields for the
14005 various pass instances.
14006 * toplev.c (finalize): Update for move of
14007 finish_optimization_passes to a method of gcc::pass_manager.
14008 * toplev.h (finish_optimization_passes): Move to method of class
14010 * tree-pass.h (struct pass_data): New.
14011 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14012 (opt_pass::gate): Convert to virtual function.
14013 (opt_pass::~opt_pass): New.
14014 (opt_pass::clone): New.
14015 (opt_pass::execute): Convert to virtual function.
14016 (opt_pass::opt_pass): New.
14017 (opt_pass::ctxt_): new.
14018 (gimple_opt_pass): Convert to subclass of opt_pass.
14019 (gimple_opt_pass::gimple_opt_pass): New.
14020 (rtl_opt_pass): Convert to subclass of opt_pass.
14021 (rtl_opt_pass::rtl_opt_pass): New.
14022 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14023 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14024 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14025 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14026 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14027 invocation of pass_mode_switching to reflect this moving from a
14028 global to a member of gcc::pass_manager.
14029 (ix86_option_override): Rework how pass_insert_vzeroupper is
14030 added to the pass_manager to reflect autogenerated changes.
14031 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14033 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14035 PR rtl-optimization/57708
14036 * recog.c (peep2_find_free_register): Validate all regs in a
14039 2013-08-05 Jan Hubicka <jh@suse.cz>
14042 * cgraph.c (verify_cgraph_node): Accept local flags from other
14044 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14045 (function_and_variable_visibility): Likewise.
14046 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14048 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14050 * graph.c (init_graph_slim_pretty_print): Remove.
14051 (print_graph_cfg): Do not call it. Use local pretty printer.
14052 (start_graph_dump): Likewise.
14054 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14056 * gimple-pretty-print.c (buffer): Remove.
14057 (initialized): Likewise.
14058 (maybe_init_pretty_print): Likewise.
14059 (print_gimple_stmt): Do not call it. Use non-static local
14060 pretty_printer variable.
14061 (print_gimple_expr): Likewise.
14062 (print_gimple_seq): Likewise.
14063 (gimple_dump_bb): Likewise.
14065 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14067 * asan.c (asan_pp): Remove.
14068 (asan_pp_initialized): Likewise.
14069 (asan_pp_initialize): Likewise.
14070 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14071 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14072 (asan_add_global): Likewise.
14074 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14076 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14077 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14078 * pretty-print.c (pp_formatted_text_data): Likewise.
14079 (pp_indent): Rename from pp_base_indent.
14080 (pp_format): Rename from pp_base_format.
14081 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14082 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14083 (pp_flush): Rename from pp_base_flush.
14084 (pp_set_line_maximum_length): Rename from
14085 pp_base_set_line_maximum_length.
14086 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14087 (pp_set_prefix): Rename from pp_base_set_prefix.
14088 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14089 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14090 (pp_append_text): Rename from pp_base_append_text.
14091 (pp_formatted_text): Rename from pp_base_formatted_text.
14092 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14093 (pp_remaining_character_count_for_line): Rename from
14094 pp_base_remaining_character_count_for_line.
14095 (pp_newline): Rename from pp_base_newline.
14096 (pp_character): Rename from pp_base_character.
14097 (pp_string): Rename from pp_base_string.
14098 (pp_maybe_space): Rename from pp_base_maybe_space.
14099 * asan.c (asan_pp_string): Adjust.
14100 (asan_emit_stack_protection): Likewise.
14101 (asan_add_global): Likewise.
14102 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14103 * tree-mudflap.c (mf_varname_tree): Likewise.
14104 * tree-pretty-print.c (pp_tree_identifier): Rename from
14105 pp_base_tree_identifier.
14106 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14107 Declare as function.
14109 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14111 * pretty-print.h (pp_bar_bar): New.
14112 (pp_ampersand_ampersand): Likewise.
14113 (pp_less_equal): Likewise.
14114 (pp_greater_equal): Likewise.
14115 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14116 printer functions instead of pp_string or operators and punctuators.
14117 (dump_gimple_call): Likewise.
14118 (dump_gimple_omp_for): Likewise.
14119 (dump_gimple_transaction): Likewise.
14120 (dump_gimple_phi): Likewise.
14121 (pp_gimple_stmt_1): Likewise.
14122 * sched-vis.c (print_insn): Likewise.
14123 * tree-mudflap.c (mf_varname_tree): Likewise.
14124 * tree-pretty-print.c (dump_block_node): Likewise.
14125 (dump_generic_node): Likewise.
14127 2013-08-02 Jan Hubicka <jh@suse.cz>
14129 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14131 * lto-streamer-out.c (tree_is_indexable): Results decls and
14132 parm decls are not indexable.
14133 (DFS_write_tree_body): Do not follow args and results.
14134 (hash_tree): Likewise.
14135 (output_functions): Rearrange so struct function is needed
14136 only when real body is output; be able to also ouptut abstract
14137 functions; output DECL_ARGUMENTS and DECL_RESULT.
14138 (lto_output): When not in WPA, ale store abstract functions.
14139 (write_symbol): Do not care about RESULT_DECL.
14140 (output_symbol_p): Handle correctly sbtract decls.
14141 * lto-streamer-in.c (input_function): Rearrange so struct
14142 function can be NULL at entry; allow streaming of
14143 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14144 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14145 sanity check during LTO.
14146 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14147 RESULT_DECl and DECL_ARGUMENTS.
14148 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14151 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14153 * pretty-print.h (pp_underscore): New.
14155 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14156 printer functions instead of pp_character.
14157 (dump_binary_rhs): Likewise.
14158 (dump_ternary_rhs): Likewise.
14159 (dump_gimple_call_args): Likewise.
14160 (pp_points_to_solution): Likewise.
14161 (dump_gimple_call): Likewise.
14162 (dump_gimple_switch): Likewise.
14163 (dump_gimple_cond): Likewise.
14164 (dump_gimple_bind): Likewise.
14165 (dump_gimple_try): Likewise.
14166 (dump_gimple_omp_for): Likewise.
14167 (dump_gimple_omp_continue): Likewise.
14168 (dump_gimple_omp_single): Likewise.
14169 (dump_gimple_omp_sections): Likewise.
14170 (dump_gimple_omp_block): Likewise.
14171 (dump_gimple_omp_critical): Likewise.
14172 (dump_gimple_transaction): Likewise.
14173 (dump_gimple_asm): Likewise.
14174 (dump_gimple_phi): Likewise.
14175 (dump_gimple_omp_parallel): Likewise.
14176 (dump_gimple_omp_task): Likewise.
14177 (dump_gimple_omp_atomic_load): Likewise.
14178 (dump_gimple_omp_atomic_store): Likewise.
14179 (dump_gimple_mem_ops): Likewise.
14180 (pp_gimple_stmt_1): Likewise.
14181 (pp_cfg_jump): Likewise.
14182 (dump_implicit_edges): Likewise.
14183 (gimple_dump_bb_for_graph): Likewise.
14184 * graph.c (draw_cfg_node): Likewise.
14185 * langhooks.c (lhd_print_error_function): Likewise.
14186 * sched-vis.c (print_exp): Likewise.
14187 (print_value): Likewise.
14188 (print_pattern): Likewise.
14189 (print_insn): Likewise.
14190 (rtl_dump_bb_for_graph): Likewise.
14191 * tree-pretty-print.c (dump_function_declaration): Likewise.
14192 (dump_array_domain): Likewise.
14193 (dump_omp_clause): Likewise.
14194 (dump_location): Likewise.
14195 (dump_generic_node): Likewise.
14196 (print_struct_decl): Likewise.
14197 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14199 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14201 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14202 candidate table when replacing a candidate statement.
14203 (replace_rhs_if_not_dup): Likewise.
14204 (replace_one_candidate): Likewise.
14206 2013-08-02 Jan Hubicka <jh@suse.cz>
14207 Martin Liska <marxin.liska@gmail.com>
14209 * cgraphunit.c (add_new_function): Fix logic when adding from
14211 (assemble_thunk): Rename to ...
14212 (expand_thunk); .. this one; export; get it working with
14213 general functions; make produced gimple valid.
14214 * cgraph.h (expand_thunk): Declare.
14216 2013-08-02 Jan Hubicka <jh@suse.cz>
14218 * ipa-cp.c (gather_context_independent_values): Use
14219 ipa_get_param_move_cost.
14220 (get_replacement_map): Remove PARAM; move parameter folding
14222 (create_specialized_node): Update.
14223 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14224 assert that we have gimple body; update move_cost.
14225 (count_formal_params): Assert that we have gimple body.
14226 (ipa_dump_param): New function.
14227 (ipa_alloc_node_params): Break out from ...
14228 (ipa_initialize_node_params): ... here.
14229 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14230 (ipa_write_node_info): Stream move costs.
14231 (ipa_read_node_info): Read move costs.
14232 (ipa_update_after_lto_read): Do not recompute node params.
14233 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14234 (ipa_get_param): Check we are not in WPA.
14235 (ipa_get_param_move_cost): New.
14236 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14237 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14238 parm numbers to be present.
14240 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14242 PR rtl-optimization/58048
14243 * lra-constraints.c (process_alt_operands): Don't check asm
14244 operand on register.
14246 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14248 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14249 the implied StoreLoad barrier for atomic operations if before.
14251 2013-08-02 Jan Hubicka <jh@suse.cz>
14252 Martin Liska <marxin.liska@gmail.com>
14254 * cgraph.c (cgraph_function_body_availability): Do not check
14256 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14257 symtab_node_availability): Declare.
14258 * ipa.c (can_replace_by_local_alias): New.
14259 (function_and_variable_visibility): Use it.
14260 * symtab.c (symtab_for_node_and_aliases,
14261 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14263 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14265 PR rtl-optimization/57963
14266 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14267 (lra_constraints): Use them.
14269 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14271 * config/arm/types.md (define_attr "type"): Add "load_acq"
14273 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14275 (cortex_a53_store1): Likewise.
14277 2013-08-01 Jan Hubicka <jh@suse.cz>
14279 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14280 partitions are not needed.
14282 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14284 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14285 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14286 MAYBE_NON_Q_CLASS_P where appropriate.
14288 2013-08-01 Jan Hubicka <jh@suse.cz>
14290 * cgraph.h (release_function_body): Declare.
14291 * tree.c (free_lang_data_in_decl): Free, parameters and return values
14292 of unused delcarations.
14294 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14296 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
14297 RTL form when subtracting a constant.
14299 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14301 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
14302 Generate canonical plus rtx with negated immediate instead of minus
14304 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
14306 2013-08-01 Jan Hubicka <jh@suse.cz>
14308 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
14309 (cgraph_release_function_body): Likewise.
14310 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
14311 * cgraph.h (cgrpah_node): Rename abstract_and_needed
14312 to used_as_abstract_origin.
14313 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
14314 symbols used as abstract origins.
14315 * cgraphunit.c (analyze_functions): Update.
14316 * ipa.c (symtab_remove_unreachable_nodes): Recompute
14317 used_as_abstract_origin.
14318 * tree-inline.c (tree_function_versioning): Update
14319 used_as_abstract_origin; be ready for DECL_RESULT and
14320 DECL_ARGUMENTS to be NULL.
14322 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
14323 for abstract functions.
14324 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
14327 2013-08-01 Jan Hubicka <jh@suse.cz>
14329 * profile.c (compute_value_histograms): Fix thinko.
14331 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14333 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
14334 aarch-common-protos.h to extra_headers.
14335 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14336 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
14337 * config/aarch64/t-aarch64 (aarch-common.o): Define.
14339 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
14341 * config/aarch64/aarch64.md (define_attr "type"): Delete.
14342 Include "../arm/types.md". Define "type" attribute for all patterns.
14343 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
14346 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
14348 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
14349 to support power8 load fusion.
14350 (fusion_gpr_mem_load): Likewise.
14352 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
14354 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
14355 declarations for power8 load fusion.
14356 (emit_fusion_gpr_load): Likewise.
14358 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14359 tuning for power8, turn on fusion mode by default. Turn on sign
14360 extending fusion mode if normal fusion mode is on, and we are at
14362 (fusion_gpr_load_p): New function, return true if we can fuse an
14363 addis instruction with a dependent load to a GPR.
14364 (emit_fusion_gpr_load): Emit the instructions for power8 load
14367 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14368 (VSX load fusion peepholes): New peepholes to fuse together an
14369 addi instruction with a VSX load instruction.
14371 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14372 peepholes to fuse an addis instruction with a load to a GPR base
14373 register. If we are supporting sign extending fusions, convert
14374 sign extending loads to zero extending loads and add an explicit
14377 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14379 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14380 aarch-common-protos.h to extra_headers.
14381 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14382 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14383 (arm_early_store_addr_dep): Likewise.
14384 (arm_no_early_alu_shift_dep): Likewise.
14385 (arm_no_early_alu_shift_value_dep): Likewise.
14386 (arm_no_early_mul_dep): Likewise.
14387 (arm_no_early_store_addr_dep): Likewise.
14388 (arm_mac_accumulator_is_mul_result): Likewise.
14389 (arm_mac_accumulator_is_result): Likewise.
14390 * config/arm/aarch-common.c: ... here. New file.
14391 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14393 (arm_early_store_addr_dep): Likewise.
14394 (arm_no_early_alu_shift_dep): Likewise.
14395 (arm_no_early_alu_shift_value_dep): Likewise.
14396 (arm_no_early_mul_dep): Likewise.
14397 (arm_no_early_store_addr_dep): Likewise.
14398 (arm_mac_accumulator_is_mul_result): Likewise.
14399 (arm_mac_accumulator_is_result): Likewise.
14400 * config/arm/aarch-common-protos.h: ... here. New file.
14401 * config/arm/t-arm (aarch-common.o): Define.
14403 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14405 * config/arm/arm.md: Include new file "types.md".
14406 (define_attr "type"): Move from here to ...
14407 (define_attr "mul32"): Likewise.
14408 (define_attr "mul64"): Likewise.
14409 * config/arm/types.md: ... here. New file.
14411 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14413 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14414 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14416 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14418 * gen-pass-instances.awk: Fix offset of substr().
14420 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14422 * Makefile.in (pass-instances.def): New.
14423 (passes.o): Replace dependency on passes.def with one on
14426 * gen-pass-instances.awk: New.
14428 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14429 rather than passes.def, updating local definition of NEXT_PASS
14430 macro to add an extra NUM parameter (currently unused).
14432 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14434 * Makefile.in (PASS_MANAGER_H): New.
14435 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14436 (passes.o): Likewise.
14437 (statistics.o): Likewise.
14438 (cgraphunit.o): Likewise.
14439 (context.o): Depend on PASS_MANAGER_H.
14441 * pass_manager.h: New.
14443 * cgraphunit.c (cgraph_add_new_function): Update for moves
14444 of globals to fields of pass_manager.
14445 (analyze_function): Likewise.
14446 (expand_function): Likewise.
14447 (ipa_passes): Likewise.
14448 (compile): Likewise.
14450 * context.c (context::context): New.
14451 * context.h (context::context): New.
14452 (context::get_passes): New.
14453 (context::passes_): New.
14455 * lto-cgraph.c (input_node): Update for moves of globals to
14456 fields of pass_manager.
14458 * passes.c (all_passes): Remove, in favor of a field of the
14459 same name within the new class pass_manager.
14460 (all_small_ipa_passes): Likewise.
14461 (all_lowering_passes): Likewise.
14462 (all_regular_ipa_passes): Likewise.
14463 (all_late_ipa_passes): Likewise.
14464 (all_lto_gen_passes): Likewise.
14465 (passes_by_id): Likewise.
14466 (passes_by_id_size): Likewise.
14467 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14468 the new class pass_manager.
14469 (set_pass_for_id): Convert to...
14470 (pass_manager::set_pass_for_id): ...method.
14471 (get_pass_for_id): Convert to...
14472 (pass_manager::get_pass_for_id): ...method.
14473 (register_one_dump_file): Move body of implementation into...
14474 (pass_manager::register_one_dump_file): ...here.
14475 (register_dump_files_1): Convert to...
14476 (pass_manager::register_dump_files_1): ...method.
14477 (register_dump_files): Convert to...
14478 (pass_manager::register_dump_files): ...method.
14479 (create_pass_tab): Update for moves of globals to fields of
14481 (dump_passes): Move body of implementation into...
14482 (pass_manager::dump_passes): ...here.
14483 (register_pass): Move body of implementation into...
14484 (pass_manager::register_pass): ...here.
14485 (init_optimization_passes): Convert into...
14486 (pass_manager::pass_manager): ...constructor for new
14487 pass_manager class, and initialize the pass_lists array.
14488 (check_profile_consistency): Update for moves of globals to
14489 fields of pass_manager.
14490 (dump_profile_report): Move body of implementation into...
14491 (pass_manager::dump_profile_report): ...here.
14492 (ipa_write_summaries_1): Update for moves of pass lists from
14493 being globals to fields of pass_manager.
14494 (ipa_write_optimization_summaries): Likewise.
14495 (ipa_read_summaries): Likewise.
14496 (ipa_read_optimization_summaries): Likewise.
14497 (execute_all_ipa_stmt_fixups): Likewise.
14499 * statistics.c (statistics_fini): Update for moves of globals to
14500 fields of pass_manager.
14502 * toplev.c (general_init): Replace call to
14503 init_optimization_passes with construction of the pass_manager
14506 * tree-pass.h (all_passes): Remove, in favor of a field of the
14507 same name within the new class pass_manager.
14508 (all_small_ipa_passes): Likewise.
14509 (all_lowering_passes): Likewise.
14510 (all_regular_ipa_passes): Likewise.
14511 (all_lto_gen_passes): Likewise.
14512 (all_late_ipa_passes): Likewise.
14513 (passes_by_id): Likewise.
14514 (passes_by_id_size): Likewise.
14515 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14516 the new class pass_manager.
14517 (get_pass_for_id): Remove.
14519 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14521 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14524 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14526 * arm.md (mulhi3): New expand pattern.
14528 2013-07-30 Jan Hubicka <jh@suse.cz>
14529 Martin Liska <marxin.liska@gmail.com>
14531 * profile.c (compute_value_histograms): Do not ICE when
14532 there is mismatch only on some counters.
14534 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14536 PR rtl-optimization/57637
14537 * function.c (move_insn_for_shrink_wrap): Also check the
14538 GEN set of the LIVE problem for the liveness analysis
14539 if it exists, otherwise give up.
14541 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14543 PR tree-optimization/57993
14544 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14545 replaced statement in the candidate table.
14546 (phi_add_costs): Return infinite cost when the hidden basis does
14547 not dominate all phis on which the candidate is dependent.
14548 (replace_one_candidate): Record replaced statement in the
14551 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14553 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14554 (ashlv2si3): New expander.
14555 (*ashlv2si3_i): New define_insn_and_split.
14556 * predicates.md (float_operation): Allow patterns with three
14557 basic sub-patterns.
14559 PR rtl-optimization/58021
14560 * mode-switching.c (create_pre_exit): Always split off preceding
14561 insns if we are not at the basic block head.
14563 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14565 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14566 (UCLIBC_DYNAMIC_LINKER): New macro.
14567 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14569 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14570 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14572 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14573 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14574 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14575 for SF and DF modes. Use ieee_quad_format for TF mode.
14576 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14577 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14578 (mips_option_override): Handle `-mnan=legacy'.
14579 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14580 `-mabs=2008' and `-mnan=2008'.
14581 (OPTION_DEFAULT_SPECS): Add "nan" default.
14582 (ASM_SPEC): Handle `-mnan='.
14583 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14584 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14585 comment accordingly.
14586 (neg<mode>2): Likewise.
14587 * config/mips/mips.opt (mabs, mnan): New options.
14588 * doc/install.texi (Configuration): Document `--with-nan=' option.
14589 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14591 (MIPS Options): Document them.
14592 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14593 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14594 * configure: Regenerate.
14595 * config.in: Regenerate.
14597 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14599 * config/i386/i386.md (float post-reload splitters): Do not check
14600 for subregs of SSE registers.
14602 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14603 H.J. Lu <hongjiu.lu@intel.com>
14607 * config/i386/i386.md (post-reload splitter
14608 to avoid partial SSE reg dependency stalls): New pattern.
14610 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14612 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14614 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14616 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14617 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14619 2013-07-26 Cary Coutant <ccoutant@google.com>
14621 * dwarf2out.c (die_checksum_ordered): Don't include template
14622 instantiations in signature.
14623 (is_template_parameter): New function.
14624 (is_template_instantiation): New function.
14625 (generate_skeleton_bottom_up): Don't include template instantiations
14627 (generate_skeleton): Likewise.
14628 (break_out_comdat_types): Move recursive call to break out nested
14630 (prune_unused_types_mark_generic_parms_dies): Call
14631 is_template_parameter.
14633 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14635 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14636 uses vector registers.
14637 * config/aarch64/iterators.md: Add attributes rtn and vas.
14639 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14640 Richard Earnshaw <richard.earnshaw@arm.com>
14642 * combine.c (simplify_comparison): Re-canonicalize operands
14644 * config/arm/arm.md (movcond_addsi): New splitter.
14646 2013-07-25 Sterling Augustine <saugustine@google.com>
14648 * dwarf2out.c (size_of_pubnames): Move code to...
14649 (include_pubname_in_output): ...here. New.
14650 (want_pubnames): Rearrange.
14651 (output_pubnames): Call include_pubname_in_output. Move assertion.
14653 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14655 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14657 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14660 * doc/extend.texi: Remove obsolete builtins. Fix
14661 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14663 2013-07-25 Jan Hubicka <jh@suse.cz>
14665 * cgraph.c (release_function_body): Break out from ...
14666 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14667 and DECL_ARGUMENTS.
14668 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14669 old_tree in the map.
14670 (create_specialized_node): Update.
14671 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14673 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14674 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14675 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14676 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14679 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14681 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14682 addsi3_carryin_alt2_<optab>): Correct output template.
14684 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14686 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14687 Adjust for arm_restrict_it.
14688 Remove trailing whitespace.
14690 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14692 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14693 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14695 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14697 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14699 PR rtl-optimization/57960
14700 * lra-constraints.c (process_alt_operands): Use the right mode
14701 when checking strict_low.
14703 2013-07-25 Jan Hubicka <jh@suse.cz>
14705 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14706 * cgraph.c (cgraph_remove_node): Do not release function body
14707 when in cgraph streaming.
14708 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14709 in other partitions are not considered reachable; fix handling of
14712 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14714 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14716 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14721 * config/arm/arm.md ("*sibcall_insn): Replace use of
14722 Ss with US. Adjust output for v5 and v4t.
14723 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14725 * config/arm/constraints.md ("Ss"): Rename to US.
14727 2013-07-25 Terry Guo <terry.guo@arm.com>
14729 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14730 shift_add/shift_sub0/shift_sub1 RTXs.
14732 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14733 Anton Blanchard <anton@au1.ibm.com>
14735 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14736 (altivec_vpks<VI_char>ss): Likewise.
14737 (altivec_vpks<VI_char>us): Likewise.
14738 (altivec_vpku<VI_char>us): Likewise.
14739 (altivec_vpku<VI_char>um): Likewise.
14741 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14743 Introduce context class.
14745 * Makefile.in (CONTEXT_H): New.
14746 (OBJS): Add context.o.
14747 (toplev.o): Add CONTEXT_H to dependencies.
14750 * toplev.c (general_init): Create the singleton gcc::context instance.
14756 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14758 PR rtl-optimization/57968
14759 * mode-switching.c (create_pre_exit): Allow instructions that
14760 don't set a return register to need a non-exit mode.
14762 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14763 Anton Blanchard <anton@au1.ibm.com>
14765 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14766 operands to vperm for little endian.
14767 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14768 of lvsl to create the control mask for a vperm for little endian.
14770 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14771 Anton Blanchard <anton@au1.ibm.com>
14773 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14774 two operands for little-endian.
14776 2013-07-23 Steve Ellcey <sellcey@mips.com>
14778 * config/mips/mips.c (mips_case_values_threshold): New.
14779 (TARGET_CASE_VALUES_THRESHOLD): Define.
14781 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14782 Anton Blanchard <anton@au1.ibm.com>
14784 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14785 selection of field for vector splat in little endian mode.
14787 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14789 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14790 expanders to rs6000.md.
14791 (ior<mode>3): Likewise.
14792 (and<mode>3): Likewise.
14793 (one_cmpl<mode>2): Likewise.
14794 (nor<mode>3): Likewise.
14795 (andc<mode>3): Likewise.
14796 (eqv<mode>3): Likewise.
14797 (nand<mode>3): Likewise.
14798 (orc<mode>3): Likewise.
14800 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14803 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14804 to split multi-word logical operations.
14805 (rs6000_split_logical_di): Likewise.
14806 (rs6000_split_logical): Likewise.
14808 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14809 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14810 and allow TImode operations in 32-bit.
14811 (vsx_and<mode>3_64bit): Likewise.
14812 (vsx_ior<mode>3_32bit): Likewise.
14813 (vsx_ior<mode>3_64bit): Likewise.
14814 (vsx_xor<mode>3_32bit): Likewise.
14815 (vsx_xor<mode>3_64bit): Likewise.
14816 (vsx_one_cmpl<mode>2_32bit): Likewise.
14817 (vsx_one_cmpl<mode>2_64bit): Likewise.
14818 (vsx_nor<mode>3_32bit): Likewise.
14819 (vsx_nor<mode>3_64bit): Likewise.
14820 (vsx_andc<mode>3_32bit): Likewise.
14821 (vsx_andc<mode>3_64bit): Likewise.
14822 (vsx_eqv<mode>3_32bit): Likewise.
14823 (vsx_eqv<mode>3_64bit): Likewise.
14824 (vsx_nand<mode>3_32bit): Likewise.
14825 (vsx_nand<mode>3_64bit): Likewise.
14826 (vsx_orc<mode>3_32bit): Likewise.
14827 (vsx_orc<mode>3_64bit): Likewise.
14829 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14830 logical types in GPRs.
14832 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14833 logical insns to rs6000.md, and allow TImode operations in
14835 (altivec_ior<mode>3): Likewise.
14836 (altivec_xor<mode>3): Likewise.
14837 (altivec_one_cmpl<mode>2): Likewise.
14838 (altivec_nor<mode>3): Likewise.
14839 (altivec_andc<mode>3): Likewise.
14841 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14842 attributes for moving the 128-bit logical operations into
14844 (BOOL_REGS_OUTPUT): Likewise.
14845 (BOOL_REGS_OP1): Likewise.
14846 (BOOL_REGS_OP2): Likewise.
14847 (BOOL_REGS_UNARY): Likewise.
14848 (BOOL_REGS_AND_CR0): Likewise.
14849 (one_cmpl<mode>2): Add support for DI logical operations on
14850 32-bit, splitting the operations to 32-bit.
14851 (anddi3): Likewise.
14852 (iordi3): Likewise.
14853 (xordi3): Likewise.
14854 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14855 changes to combine the 32/64-bit code, allow logical operations on
14856 TI mode in 32-bit, and to use similar match_operator patterns like
14857 scalar mode uses. Combine the Altivec and VSX code for logical
14858 operations, and move it here.
14859 (ior<mode>3, 128-bit types): Likewise.
14860 (xor<mode>3, 128-bit types): Likewise.
14861 (one_cmpl<mode>3, 128-bit types): Likewise.
14862 (nor<mode>3, 128-bit types): Likewise.
14863 (andc<mode>3, 128-bit types): Likewise.
14864 (eqv<mode>3, 128-bit types): Likewise.
14865 (nand<mode>3, 128-bit types): Likewise.
14866 (orc<mode>3, 128-bit types): Likewise.
14867 (and<mode>3_internal): Likewise.
14868 (bool<mode>3_internal): Likewise.
14869 (boolc<mode>3_internal1): Likewise.
14870 (boolc<mode>3_internal2): Likewise.
14871 (boolcc<mode>3_internal1): Likewise.
14872 (boolcc<mode>3_internal2): Likewise.
14873 (eqv<mode>3_internal1): Likewise.
14874 (eqv<mode>3_internal2): Likewise.
14875 (one_cmpl1<mode>3_internal): Likewise.
14877 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14879 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14880 Rename flag_stack_usage to flag_stack_usage_info.
14882 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14884 * config/microblaze/sync.md: New file.
14885 * config/microblaze/microblaze.md: Include sync.md
14886 * config/microblaze/microblaze.c: Add print_operand 'y'.
14887 * config/microblaze/constraints.md: Add memory_contraint
14888 'Q' which is a single register.
14890 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14892 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14894 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14896 * reload.c (find_reloads): Exit loop once we find this operand
14897 cannot be reloaded somehow for this alternative.
14899 * reload.c (find_reloads): Exit loop once we find a hard register.
14901 * rtlanal.c (computed_jump_p): Exit loop once we find label
14904 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14906 * cfgloopmanip.c (remove_path): Exit loop after setting
14909 * gensupport.c (subst_dup): Avoid loop if code is not
14910 MATCH_DUP nor MATCH_OP_DUP.
14912 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14914 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14916 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14918 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14919 true for SP_REGNUM if mode == ptr_mode.
14920 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14921 with value R0_REGNUM + 31.
14923 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14925 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14926 pad pointer-typed argument downward.
14928 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14930 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14931 and __ILP32__ when the ILP32 model is in use.
14933 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14935 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14936 (aarch64_load_symref_appropriately): In the case of
14937 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14938 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14939 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14940 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14941 if 'mode' doesn't equal to 'ptr_mode'.
14942 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14943 'vcall_offset'; change to call aarch64_emit_move differently depending
14944 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14945 to calculate the upper bound of 'vcall_offset'.
14946 (aarch64_cannot_force_const_mem): Change to also return true if
14948 (aarch64_legitimize_reload_address): In the case of large
14949 displacements, add new local variable 'xmode' and an assertion
14950 based on it; change to use 'xmode' to generate the new rtx and
14952 (aarch64_asm_trampoline_template): Change to generate the template
14953 differently depending on TARGET_ILP32 or not; change to use
14954 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14955 (aarch64_trampoline_size): Removed.
14956 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14957 and replace immediate literals with it. Change to use 'ptr_mode'
14958 instead of 'DImode' and call convert_memory_address if the mode
14959 of 'fnaddr' doesn't equal to 'ptr_mode'.
14960 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14962 (aarch64_elf_asm_destructor): Likewise.
14963 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14964 on TARGET_ILP32 instead of aarch64_trampoline_size.
14965 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14966 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14967 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14968 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14969 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14970 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14971 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14972 depending on the value of 'mode'.
14973 (add_losym_<mode>): New.
14974 (ldr_got_small_<mode>): New, based on ldr_got_small.
14975 (ldr_got_small): Remove.
14976 (ldr_got_small_sidi): New.
14977 * config/aarch64/iterators.md (P): New.
14978 (PTR): Change to 'ptr_mode' in the condition.
14980 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14982 * config.gcc (aarch64*-*-*): Support --with-abi.
14983 (aarch64*-*-elf): Support --with-multilib-list.
14984 (aarch64*-*-linux*): Likewise.
14985 (supported_defaults): Add abi to aarch64*-*-*.
14986 * configure.ac: Mention AArch64 for --with-multilib-list.
14987 * configure: Re-generated.
14988 * config/aarch64/biarchilp32.h: New file.
14989 * config/aarch64/biarchlp64.h: New file.
14990 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14992 (MULTILIB_DEFAULTS): Ditto.
14993 (DRIVER_SELF_SPECS): Ditto.
14994 (ASM_SPEC): Update to also substitute -mabi.
14995 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14996 file whose name depends on -mabi= and -mbig-endian.
14997 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14999 (POINTER_SIZE): New define.
15000 (POINTERS_EXTEND_UNSIGNED): Ditto.
15001 (enum aarch64_abi_type): New enumeration tag.
15002 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15003 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15004 (TARGET_ILP32): New define.
15005 * config/aarch64/aarch64.opt (mabi): New.
15006 (aarch64_abi): New.
15007 (ilp32, lp64): New values for -mabi.
15008 * config/aarch64/t-aarch64 (comma): New define.
15009 (MULTILIB_OPTIONS): Ditto.
15010 (MULTILIB_DIRNAMES): Ditto.
15011 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15012 * doc/invoke.texi: Document -mabi for AArch64.
15014 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15016 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15018 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15019 Anton Blanchard <anton@au1.ibm.com>
15021 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15022 endianness when selecting field to splat.
15024 2013-07-22 Eric Christopher <echristo@gmail.com>
15026 * dwarf2out.c (die_odr_checksum): New function to use
15027 CHECKSUM_ macros and ULEB128 for DIE tag.
15028 (generate_type_signature): Use.
15030 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15032 * config.gcc (sparc*-*-*): Accept leon3 processor.
15033 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15034 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15035 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15036 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15037 (mfix-ut699): Adjust comment.
15038 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15039 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15040 (CPP_CPU_SPEC): Likewise.
15041 (ASM_CPU_SPEC): Likewise.
15042 * config/sparc/sparc.c (leon3_cost): New constant.
15043 (sparc_option_override): Add leon3 support.
15044 (mem_ref): New function.
15045 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15046 (sparc_do_work_around_errata): Look into the instruction in the delay
15047 slot and adjust accordingly. Add fix for the data cache nullify issues
15048 of the UT699. Change insertion position for the NOP.
15049 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15050 (leon3_load): New reservation.
15051 (leon_store): Bump latency to 2.
15052 (grfpu): New automaton.
15053 (grfpu_alu): New unit.
15054 (grfpu_ds): Likewise.
15055 (leon_fp_alu): Adjust.
15056 (leon_fp_mult): Delete.
15057 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15058 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15059 * config/sparc/sparc.md (cpu): Add leon3.
15060 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15061 (swapsi): Likewise.
15062 (atomic_test_and_set): Likewise.
15063 (ldstub): Likewise.
15065 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15067 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15068 default for R5900 targets.
15069 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15070 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15071 * config/mips/mips.c (mips_option_override): Report an error for
15072 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15073 for -march=r5900 -mhard-float.
15075 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15077 * df-problems.c (can_move_insns_across): Exit loop once we
15078 find a non-fixed, non-global register.
15080 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15083 * omega.c (omega_eliminate_red): Break after setting red_found.
15084 (omega_problem_has_red_equations): Similarly after setting found.
15085 (omega_query_variable): Similarly after setting coupled.
15087 2013-07-22 Marek Polacek <polacek@redhat.com>
15089 * gimplify.c: Don't include gimple.h twice.
15091 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15093 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15094 instead of TARGET_THUMB1.
15095 (Pz): New constraint.
15096 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15098 (compare_negsi_si): Likewise.
15099 (compare_addsi2_op0): Likewise.
15100 (compare_addsi2_op1): Likewise.
15101 (addsi3_carryin_<optab>): Likewise.
15102 (addsi3_carryin_alt2_<optab>): Likewise.
15103 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15104 for arm_restrict_it.
15105 (subsi3_carryin): Likewise.
15106 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15107 (minmax_arithsi): Disable for arm_restrict_it.
15108 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15109 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15110 (satsi_<SAT:code>_shift): Likewise.
15111 (arm_shiftsi3): Add alternative for 16-bit encoding.
15112 (arm32_movhf): Disable for arm_restrict_it.
15113 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15114 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15116 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15118 * config/arm/arm.md (attribute "insn"): Delete.
15119 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15120 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15121 (not_shiftsi): Update for attribute change.
15122 (not_shiftsi_compare0): Likewise.
15123 (not_shiftsi_compare0_scratch): Likewise.
15124 (arm_one_cmplsi2): Likewise.
15125 (thumb1_one_cmplsi2): Likewise.
15126 (notsi_compare0): Likewise.
15127 (notsi_compare0_scratch): Likewise.
15128 (thumb1_movdi_insn): Likewise.
15129 (arm_movsi_insn): Likewise.
15130 (movhi_insn_arch4): Likewise.
15131 (movhi_bytes): Likewise.
15132 (arm_movqi_insn): Likewise.
15133 (thumb1_movqi_insn): Likewise.
15134 (arm32_movhf): Likewise.
15135 (thumb1_movhf): Likewise.
15136 (arm_movsf_soft_insn): Likewise.
15137 (thumb1_movsf_insn): Likewise.
15138 (thumb_movdf_insn): Likewise.
15139 (movsicc_insn): Likewise.
15140 (movsfcc_soft_insn): Likewise.
15141 (and_scc): Likewise.
15142 (cond_move): Likewise.
15143 (if_move_not): Likewise.
15144 (if_not_move): Likewise.
15145 (if_shift_move): Likewise.
15146 (if_move_shift): Likewise.
15147 (if_shift_shift): Likewise.
15148 (if_not_arith): Likewise.
15149 (if_arith_not): Likewise.
15150 (cond_move_not): Likewise.
15151 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15152 (neon_mov<mode>): Likewise.
15153 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15154 (thumb2_movsi_vfp): Likewise.
15155 (movsf_vfp): Likewise.
15156 (thumb2_movsf_vfp): Likewise.
15157 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15159 (cortexa7_older_only): Likewise.
15160 (cortexa7_younger): Likewise.
15161 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15162 (1020alu_shift_op): Likewise.
15163 (1020alu_shift_reg_op): Likewise.
15164 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15165 (alu_shift_op): Likewise.
15166 (alu_shift_reg_op): Likewise.
15167 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15168 (11_alu_shift_op): Likewise.
15169 (11_alu_shift_reg_op): Likewise.
15170 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15171 (9_alu_shift_reg_op): Likewise.
15172 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15174 (cortex_a15_alu_shift): Likewise.
15175 (cortex_a15_alu_shift_reg): Likewise.
15176 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15177 (cortex_a5_alu_shift): Likewise.
15178 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15180 (cortex_a53_alu_shift): Likewise.
15181 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15183 (cortex_a7_alu_reg): Likewise.
15184 (cortex_a7_alu_shift): Likewise.
15185 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15186 (cortex_a8_alu_shift): Likewise.
15187 (cortex_a8_alu_shift_reg): Likewise.
15188 (cortex_a8_mov): Likewise.
15189 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15190 (cortex_a9_dp_shift): Likewise.
15191 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15192 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15193 (cortex_r4_mov): Likewise.
15194 (cortex_r4_alu_shift): Likewise.
15195 (cortex_r4_alu_shift_reg): Likewise.
15196 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15197 (526_alu_shift_op): Likewise.
15198 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15199 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15200 (626te_alu_shift_op): Likewise.
15201 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15202 (726te_alu_op): Likewise.
15203 (726te_alu_shift_op): Likewise.
15204 (726te_alu_shift_reg_op): Likewise.
15205 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15206 (mp626_alu_shift_op): Likewise.
15207 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15208 (pj4_alu_e1_conds): Likewise.
15209 (pj4_alu): Likewise.
15210 (pj4_alu_conds): Likewise.
15211 (pj4_shift): Likewise.
15212 (pj4_shift_conds): Likewise.
15213 (pj4_alu_shift): Likewise.
15214 (pj4_alu_shift_conds): Likewise.
15216 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15218 * config/arm/predicates.md (shiftable_operator_strict_it):
15220 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15221 Disable cond_exec version for arm_restrict_it.
15222 (thumb2_smaxsi3): Convert to generate cond_exec.
15223 (thumb2_sminsi3): Likewise.
15224 (thumb32_umaxsi3): Likewise.
15225 (thumb2_uminsi3): Likewise.
15226 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15227 (thumb2_neg_abssi2): Likewise.
15228 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15229 (thumb2_movsicc_insn): Adjust alternatives.
15230 (thumb2_mov_negscc): Disable for arm_restrict_it.
15231 (thumb2_mov_negscc_strict_it): New pattern.
15232 (thumb2_mov_notscc_strict_it): New pattern.
15233 (thumb2_mov_notscc): Disable for arm_restrict_it.
15234 (thumb2_ior_scc): Likewise.
15235 (thumb2_ior_scc_strict_it): New pattern.
15236 (thumb2_cond_move): Adjust for arm_restrict_it.
15237 (thumb2_cond_arith): Disable for arm_restrict_it.
15238 (thumb2_cond_arith_strict_it): New pattern.
15239 (thumb2_cond_sub): Adjust for arm_restrict_it.
15240 (thumb2_movcond): Likewise.
15241 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15242 (thumb2_zero_extendhisi2_v6): Likewise.
15243 (thumb2_zero_extendqisi2_v6): Likewise.
15244 (orsi_notsi_si): Likewise.
15245 (orsi_not_shiftsi_si): Likewise.
15247 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15249 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15250 instruction sequence is 1 byte shorter.
15252 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15254 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15255 it is not needed after split.
15257 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15260 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15261 second label for nonlocal goto receivers. Don't output pic base labels
15262 unless we're producing PIC; mark that action unreachable().
15263 (ix86_save_reg): If the function contains a nonlocal label, save the
15265 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15266 * config/darwin.c (emitted_pic_label_num): New GTY.
15267 (update_pic_label_number_if_needed): New.
15268 (machopic_output_function_base_name): Adjust for nonlocal receiver
15270 (machopic_should_output_picbase_label): New.
15271 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15272 (nonlocal_goto_receiver): New insn and split.
15274 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15276 * config/aarch64/aarch64-builtins.c
15277 (aarch64_fold_builtin): Fold abs in all modes.
15278 * config/aarch64/aarch64-simd-builtins.def
15279 (abs): Enable for all modes.
15280 * config/aarch64/arm_neon.h
15281 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15282 (vabs_f64): Add missing intrinsic.
15284 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15286 * config/aarch64/arm_neon.h (vabs_s64): New function
15288 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
15291 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
15292 * config/avr/avr.md (adjust_len): Add `round'.
15293 * config/avr/avr-protos.h (avr_out_round): New prototype.
15294 (avr_out_plus): Add `out_label' argument.
15295 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
15296 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
15297 Handle the case where `insn' is just a pattern.
15298 (avr_out_bitop): Handle the case where `insn' is just a pattern.
15299 (avr_out_round): New function.
15300 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
15302 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
15304 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15305 Add check for flag_stack_usage to handle -fstack-usage support
15307 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
15309 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
15310 interaction for new Power8 flags and VSX.
15312 2013-07-18 Sriraman Tallam <tmsriram@google.com>
15314 PR middle-end/57698
15315 * tree-inline.c (expand_call_inline): Emit errors during
15316 early_inlining only if optimization is not turned on.
15318 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15322 * passes.c (init_optimization_passes): Move the construction of
15323 the pass hierarchy into a new passes.def file.
15325 * Makefile.in (passes.o): Add dependency on passes.def.
15327 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15329 * passes.c (init_optimization_passes): Introduce macros for
15330 constructing the tree of passes (INSERT_PASSES_AFTER,
15331 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
15332 TERMINATE_PASS_LIST).
15334 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
15335 Wei Mi <wmi@google.com>
15337 PR rtl-optimization/57878
15338 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
15340 (reload_pseudo_compare_func): Check nregs first for reload
15343 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15345 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
15347 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
15349 * read-rtl.c (validate_const_int): Once an invalid character is
15350 seen, quit the loop.
15352 * gengtype.c (write_roots): Similarly once we find the "deletable"
15353 or "if_marked" option.
15355 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15357 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15358 "xtab" and "sat". Move value "clz" from here to ...
15359 (attriubte "type"): ... here.
15360 (satsi_<SAT:code>): Delete "insn" attribute.
15361 (satsi_<SAT:code>_shift): Likewise.
15362 (arm_zero_extendqisi2addsi): Likewise.
15363 (arm_extendqisi2addsi): Likewise.
15364 (clzsi2): Update for attribute changes.
15365 (rbitsi2): Likewise.
15366 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15368 (arm_usatsihi): Likewise.
15369 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15371 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15373 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15374 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15375 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15376 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15377 in alphabetical order.
15378 (attribute "core_cycles"): Update for attribute changes.
15379 (arm_addsi3): Likewise.
15380 (addsi3_compare0): Likewise.
15381 (addsi3_compare0_scratch): Likewise.
15382 (addsi3_compare_op1): Likewise.
15383 (addsi3_compare_op2): Likewise.
15384 (compare_addsi2_op0): Likewise.
15385 (compare_addsi2_op1): Likewise.
15386 (addsi3_carryin_shift_<optab>): Likewise.
15387 (subsi3_carryin_shift): Likewise.
15388 (rsbsi3_carryin_shift): Likewise.
15389 (arm_subsi3_insn): Likewise.
15390 (subsi3_compare0): Likewise.
15391 (subsi3_compare): Likewise.
15392 (arm_andsi3_insn): Likewise.
15393 (thumb1_andsi3_insn): Likewise.
15394 (andsi3_compare0): Likewise.
15395 (andsi3_compare0_scratch): Likewise.
15396 (zeroextractsi_compare0_scratch
15397 (andsi_not_shiftsi_si): Likewise.
15398 (iorsi3_insn): Likewise.
15399 (iorsi3_compare0): Likewise.
15400 (iorsi3_compare0_scratch): Likewise.
15401 (arm_xorsi3): Likewise.
15402 (thumb1_xorsi3_insn): Likewise.
15403 (xorsi3_compare0): Likewise.
15404 (xorsi3_compare0_scratch): Likewise.
15405 (satsi_<SAT:code>_shift): Likewise.
15407 (arm_shiftsi3): Likewise.
15408 (shiftsi3_compare0): Likewise.
15409 (not_shiftsi): Likewise.
15410 (not_shiftsi_compare0): Likewise.
15411 (not_shiftsi_compare0_scratch): Likewise.
15412 (arm_one_cmplsi2): Likewise.
15413 (thumb_one_complsi2): Likewise.
15414 (notsi_compare0): Likewise.
15415 (notsi_compare0_scratch): Likewise.
15416 (thumb1_zero_extendhisi2): Likewise.
15417 (arm_zero_extendhisi2): Likewise.
15418 (arm_zero_extendhisi2_v6): Likewise.
15419 (arm_zero_extendhisi2addsi): Likewise.
15420 (thumb1_zero_extendqisi2): Likewise.
15421 (thumb1_zero_extendqisi2_v6): Likewise.
15422 (arm_zero_extendqisi2): Likewise.
15423 (arm_zero_extendqisi2_v6): Likewise.
15424 (arm_zero_extendqisi2addsi): Likewise.
15425 (thumb1_extendhisi2): Likewise.
15426 (arm_extendhisi2): Likewise.
15427 (arm_extendhisi2_v6): Likewise.
15428 (arm_extendqisi): Likewise.
15429 (arm_extendqisi_v6): Likewise.
15430 (arm_extendqisi2addsi): Likewise.
15431 (thumb1_extendqisi2): Likewise.
15432 (thumb1_movdi_insn): Likewise.
15433 (arm_movsi_insn): Likewise.
15434 (movsi_compare0): Likewise.
15435 (movhi_insn_arch4): Likewise.
15436 (movhi_bytes): Likewise.
15437 (arm_movqi_insn): Likewise.
15438 (thumb1_movqi_insn): Likewise.
15439 (arm32_movhf): Likewise.
15440 (thumb1_movhf): Likewise.
15441 (arm_movsf_soft_insn): Likewise.
15442 (thumb1_movsf_insn): Likewise.
15443 (movdf_soft_insn): Likewise.
15444 (thumb_movdf_insn): Likewise.
15445 (arm_cmpsi_insn): Likewise.
15446 (cmpsi_shiftsi): Likewise.
15447 (cmpsi_shiftsi_swp): Likewise.
15448 (arm_cmpsi_negshiftsi_si): Likewise.
15449 (movsicc_insn): Likewise.
15450 (movsfcc_soft_insn): Likewise.
15451 (arith_shiftsi): Likewise.
15452 (arith_shiftsi_compare0
15453 (arith_shiftsi_compare0_scratch
15454 (sub_shiftsi): Likewise.
15455 (sub_shiftsi_compare0
15456 (sub_shiftsi_compare0_scratch
15457 (and_scc): Likewise.
15458 (cond_move): Likewise.
15459 (if_plus_move): Likewise.
15460 (if_move_plus): Likewise.
15461 (if_move_not): Likewise.
15462 (if_not_move): Likewise.
15463 (if_shift_move): Likewise.
15464 (if_move_shift): Likewise.
15465 (if_shift_shift): Likewise.
15466 (if_not_arith): Likewise.
15467 (if_arith_not): Likewise.
15468 (cond_move_not): Likewise.
15469 (thumb1_ashlsi3): Set type attribute.
15470 (thumb1_ashrsi3): Likewise.
15471 (thumb1_lshrsi3): Likewise.
15472 (thumb1_rotrsi3): Likewise.
15473 (shiftsi3_compare0_scratch): Likewise.
15474 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15475 (neon_mov<mode>): Likewise.
15476 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15478 (thumb2_movsi_insn): Likewise.
15479 (thumb2_cmpsi_neg_shiftsi): Likewise.
15480 (thumb2_extendqisi_v6): Likewise.
15481 (thumb2_zero_extendhisi2_v6): Likewise.
15482 (thumb2_zero_extendqisi2_v6): Likewise.
15483 (thumb2_shiftsi3_short): Likewise.
15484 (thumb2_addsi3_compare0_scratch): Likewise.
15485 (orsi_not_shiftsi_si): Likewise.
15486 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15487 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15489 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15490 (1020alu_shift_op): Likewise.
15491 (1020alu_shift_reg_op): Likewise.
15492 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15493 (alu_shift_op): Likewise.
15494 (alu_shift_reg_op): Likewise.
15495 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15496 (11_alu_shift_op): Likewise.
15497 (11_alu_shift_reg_op): Likewise.
15498 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15499 (9_alu_shift_reg_op): Likewise.
15500 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15502 (cortex_a15_alu_shift): Likewise.
15503 (cortex_a15_alu_shift_reg): Likewise.
15504 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15506 (cortex_a5_alu_shift): Likewise.
15507 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15509 (cortex_a53_alu_shift): Likewise.
15510 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15512 (cortex_a7_alu_reg): Likewise.
15513 (cortex_a7_alu_shift): Likewise.
15514 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15516 (cortex_a8_alu_shift): Likewise.
15517 (cortex_a8_alu_shift_reg): Likewise.
15518 (cortex_a8_mov): Likewise.
15519 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15520 (cortex_a9_dp_shift): Likewise.
15521 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15523 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15525 (cortex_r4_mov): Likewise.
15526 (cortex_r4_alu_shift): Likewise.
15527 (cortex_r4_alu_shift_reg): Likewise.
15528 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15529 (526_alu_shift_op): Likewise.
15530 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15531 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15532 (626te_alu_shift_op): Likewise.
15533 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15534 (726te_alu_op): Likewise.
15535 (726te_alu_shift_op): Likewise.
15536 (726te_alu_shift_reg_op): Likewise.
15537 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15538 (mp626_alu_shift_op): Likewise.
15539 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15540 (pj4_alu_e1_conds): Likewise.
15541 (pj4_alu): Likewise.
15542 (pj4_alu_conds): Likewise.
15543 (pj4_shift): Likewise.
15544 (pj4_shift_conds): Likewise.
15545 (pj4_alu_shift): Likewise.
15546 (pj4_alu_shift_conds): Likewise.
15547 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15549 (cortexa7_older_only): Likewise.
15550 (cortexa7_younger): Likewise.
15552 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15554 * ipa-pure-const.c (generate_summary): Rename to...
15555 (pure_const_generate_summary): ... this.
15557 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15559 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15561 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15564 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15567 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15569 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15570 enabled without -march=zEC12.
15571 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15574 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15576 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15577 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15578 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15579 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15580 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15581 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15582 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15584 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15585 and ISA_HAS_NMADD3_NMSUB3.
15586 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15587 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15588 (nmsub4<mode>, nmsub3<mode>): Likewise.
15589 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15591 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15593 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15594 TARGET_MIPS5400 checking.
15596 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15597 Peter Bergner <bergner@vnet.ibm.com>
15599 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15600 registers in the comment.
15601 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15602 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15603 rather than FIRST_PSEUDO_REGISTERS.
15605 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15607 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15608 enable extra ISA flags with TARGET_HTM.
15610 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15612 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15615 2013-07-15 Cong Hou <congh@google.com>
15617 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15618 in compare function for sorting.
15620 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15622 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15623 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15624 * config/rs6000/rs6000.opt: Add -mhtm option.
15625 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15626 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15627 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15628 __HTM__ if the HTM instructions are available.
15629 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15630 htm_spr_reg_operand): New define_predicates.
15631 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15632 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15634 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15635 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15636 HTM builtin functions.
15637 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15638 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15639 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15640 (rs6000_builtin_mask_calculate): Likewise.
15641 (rs6000_option_override_internal): Likewise.
15642 (bdesc_htm): Add new HTM builtin support.
15643 (htm_spr_num): New function.
15644 (htm_spr_regno): Likewise.
15645 (rs6000_htm_spr_icode): Likewise.
15646 (htm_expand_builtin): Likewise.
15647 (htm_init_builtins): Likewise.
15648 (rs6000_expand_builtin): Add support for HTM builtin functions.
15649 (rs6000_init_builtins): Likewise.
15650 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15652 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15653 (TARGET_HTM, MASK_HTM): Define macros.
15654 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15655 (FIXED_REGISTERS): Likewise.
15656 (CALL_USED_REGISTERS): Likewise.
15657 (CALL_REALLY_USED_REGISTERS): Likewise.
15658 (REG_ALLOC_ORDER): Likewise.
15659 (enum reg_class): Likewise.
15660 (REG_CLASS_NAMES): Likewise.
15661 (REG_CLASS_CONTENTS): Likewise.
15662 (REGISTER_NAMES): Likewise.
15663 (ADDITIONAL_REGISTER_NAMES): Likewise.
15664 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15665 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15666 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15667 * config/rs6000/htm.md: New file.
15668 * config/rs6000/htmintrin.h: New file.
15669 * config/rs6000/htmxlintrin.h: New file.
15671 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15673 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15674 Define SYMBOL_TINY_GOT, update comment.
15675 * config/aarch64/aarch64.c
15676 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15677 (aarch64_expand_mov_immediate): Likewise.
15678 (aarch64_print_operand): Likewise.
15679 (aarch64_classify_symbol): Likewise.
15680 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15681 (ldr_got_tiny): Define.
15683 2013-07-13 Tobias Grosser <tobias@grosser.es>
15685 PR tree-optimization/54094
15686 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15687 scheduling dimension for the parallelism check from the polyhedral
15688 information in the AST.
15689 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15692 2013-07-13 Jason Merrill <jason@redhat.com>
15694 * print-tree.c (debug_vec_tree): Use debug_raw.
15695 (debug_raw (vec<tree, va_gc> &)): New.
15696 (debug_raw (vec<tree, va_gc> *)): New.
15697 * tree.h: Declare them.
15699 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15701 * ifcvt.c (ifcvt_after_combine): New static variable.
15702 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15704 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15705 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15706 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15708 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15710 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15712 2013-07-12 Michael Matz <matz@suse.de>
15714 PR middle-end/55771
15715 * convert.c (convert_to_real): Reject non-float inner types.
15717 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15719 * config/aarch64/aarch64-protos.h
15720 (aarch64_simd_immediate_valid_for_move): Remove.
15721 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15722 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15724 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15726 2013-07-11 Steve Ellcey <sellcey@mips.com>
15728 * config/mips/mips.c (mips_conditional_register_usage): Do not
15729 use t[0-7] registers in MIPS16 mode when optimizing for size.
15731 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15733 * config/i386/i386.c (dispatch_function_versions): Fix array
15734 indexing of function_version_info to match actual_versions.
15736 2013-07-11 Teresa Johnson <tejohnson@google.com>
15738 * vec.h (struct va_gc): Move release out-of-line.
15739 (va_gc::release): Call ggc_free on released vec.
15741 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15743 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15744 Require GOT register as additional operand in UNSPEC.
15745 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15746 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15747 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15748 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15749 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15750 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15751 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15753 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15756 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15757 name seen by assembler/linker rather if available.
15759 2013-07-11 Andreas Schwab <schwab@suse.de>
15761 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15763 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15765 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15767 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15769 * doc/tm.texi.in: Move hook documentation to ....
15770 * target.def: ... here.
15772 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15773 text on @hook line.
15774 * doc/tm.texi: Regenerate.
15776 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15779 * doc/invoke.texi: Document -Wconditionally-supported.
15781 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15784 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15787 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15790 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15791 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15792 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15793 Remove duplicate devices.
15794 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15795 * config/avr/t-multilib: Regenerate.
15796 * config/avr/avr-tables.opt: Regenerate.
15797 * doc/avr-mmcu.texi: Regenerate.
15799 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15802 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15804 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15806 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15807 the cost of MULT when optimizing for size.
15809 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15811 * config/cr16/cr16-protos.h: Don't include target.h.
15813 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15815 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15816 adjust register size for TDmode and TFmode for VSX registers.
15818 2013-07-08 Kai Tietz <ktietz@redhat.com>
15821 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15822 hook_bool_const_tree_true.
15824 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15826 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15827 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15828 * config/s390/s390.md: Define FPR*_REGNUM constants.
15829 Fix FPR2_REGNUM constant (18 -> 17).
15830 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15831 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15832 ("trunc<BFP:mode><DFP_ALL:mode>2")
15833 ("trunc<DFP_ALL:mode><BFP:mode>2")
15834 ("*extend<BFP:mode><DFP_ALL:mode>2")
15835 ("*extend<DFP_ALL:mode><BFP:mode>2")
15836 ("extend<BFP:mode><DFP_ALL:mode>2")
15837 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15840 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15842 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15844 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15846 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15847 and cfun_fpr_bit_p to cfun_fpr_save_p.
15848 (s390_frame_area, s390_register_info, s390_frame_info)
15849 (s390_emit_prologue, s390_emit_epilogue)
15850 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15852 * config/s390/s390.h: Define *_REGNUM macros for floating point
15855 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15857 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15859 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15861 PR rtl-optimization/57786
15862 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15863 and break out of the loop when it is set to false.
15865 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15868 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15869 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15870 (const_int 63)) 0)).
15871 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15872 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15873 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15875 PR rtl-optimization/57829
15876 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15877 mask bits outside of mode are just sign-extension from mode to HWI.
15879 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15881 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15882 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15883 adjust_address instead of change_address to keep info about alignment.
15884 (emit_strmov): Remove.
15885 (emit_memmov): New function.
15886 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15887 (expand_movmem_epilogue): Likewise and return updated rtx for
15889 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15890 destination and source.
15891 (decide_alignment): Refactor, handle vector_loop.
15892 (ix86_expand_movmem): Likewise.
15893 (ix86_expand_setmem): Likewise.
15894 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15896 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15898 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15899 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15901 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15903 * config/i386/sse.md (sse_movlhps): Change alternative 3
15904 of operand 2 to "m".
15906 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15909 * config/i386/sse.md (iptr): New mode attribute.
15910 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15911 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15912 (<sse>_vmmul<mode>3): Ditto.
15913 (<sse>_vmdiv<mode>3): Ditto.
15914 (sse_vmrcpv4sf2): Ditto.
15915 (<sse>_vmsqrt<mode>2): Ditto.
15916 (sse_vmrsqrtv4sf2): Ditto.
15917 (<sse>_vm<code><mode>3): Ditto.
15918 (avx_vmcmp<mode>3): Ditto.
15919 (<sse>_vmmaskcmp<mode>3): Ditto.
15920 (<sse>_comi): Ditto.
15921 (<sse>_ucomi): Ditto.
15922 (*xop_vmfrcz_<mode>): Ditto.
15923 (*fmai_fmadd_<mode>): Ditto.
15924 (*fmai_fmsub_<mode>): Ditto.
15925 (*fmai_fnmadd_<mode>): Ditto.
15926 (*fmai_fnmsub_<mode>): Ditto.
15927 (*fma4i_vmfmadd_<mode>): Ditto.
15928 (*fma4i_vmfmsub_<mode>): Ditto.
15929 (*fma4i_vmfnmadd_<mode>): Ditto.
15930 (*fma4i_vmfnmsub_<mode>): Ditto.
15931 (*xop_vmfrcz_<mode>): Ditto.
15932 (sse_cvtps2pi): Ditto.
15933 (sse_cvttps2pi): Ditto.
15934 (sse_cvtss2si): Ditto.
15935 (sse_cvtss2si_2): Ditto.
15936 (sse_cvtss2siq_2): Ditto.
15937 (sse_cvttss2si): Ditto.
15938 (sse_cvttss2siq): Ditto.
15939 (sse_cvtsd2si): Ditto.
15940 (sse_cvtsd2si_2): Ditto.
15941 (sse_cvtsd2siq_2): Ditto.
15942 (sse_cvttsd2si): Ditto.
15943 (sse_cvttsd2siq): Ditto.
15944 (sse_cvtsd2ss): Ditto.
15945 (sse_cvtss2sd): Ditto.
15946 (avx2_pbroadcast<mode>): Ditto.
15947 (avx2_pbroadcast<mode>_1): Ditto.
15948 (*avx_vperm_broadcast_v4sf): Ditto.
15950 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15951 (sse_movlhps): Ditto.
15952 (sse_storehps): Ditto.
15953 (sse_loadhps): Ditto.
15954 (sse_storelps): Ditto.
15955 (sse_loadlps): Ditto.
15956 (*vec_concatv4sf): Ditto.
15957 (*vec_interleave_highv2df): Ditto.
15958 (*vec_interleave_lowv2df): Ditto.
15959 (*vec_extractv2df_1_sse): Ditto.
15960 (*vec_extractv2df_0_sse): Ditto.
15961 (sse2_storelpd): Ditto.
15962 (sse2_loadlpd): Ditto.
15963 (sse2_movsd): Ditto.
15964 (*vec_concatv4si): Ditto.
15965 (vec_concatv2di): Ditto.
15967 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15968 for Intel asm dialect.
15969 (mmx_punpcklwd): Ditto.
15970 (mmx_punpckldq): Ditto.
15972 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15973 for intel assembler dialect.
15975 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15978 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15979 for BUILT_IN_C{LZ,LRSB}*.
15980 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15981 * tree-vrp.c (extract_range_basic): Handle
15982 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15983 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15984 fall thru to code calling set_value*.
15985 * builtins.c (expand_builtin): Remove *IMAX cases.
15986 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15987 if width is bigger than 2*HWI.
15989 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15991 PR rtl-optimization/55342
15992 * lra-int.h (lra_subreg_reload_pseudos): New.
15993 * lra.c: Add undoing optional reloads to the block diagram.
15994 (lra_subreg_reload_pseudos): New.
15995 (lra_optional_reload_pseudos): Change comments.
15996 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15997 lra_optional_reload_pseudos after undo transformations.
15998 * lra-assigns.c (pseudo_prefix_title): New.
15999 (lra_setup_reg_renumber): Use it.
16000 (spill_for): Ditto. Check subreg reload pseudos too.
16001 (assign_by_spills): Consider subreg reload pseudos too.
16002 * lra-constraints.c (simplify_operand_subreg): Use
16003 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16004 (curr_insn_transform): Recognize and do optional reloads.
16005 (undo_optional_reloads): New.
16006 (lra_undo_inheritance): Call undo_optional_reloads.
16008 2013-07-05 Thomas Quinot <quinot@adacore.com>
16010 * tree-complex.c (expand_complex_operations_1): Fix typo.
16012 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16014 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16015 (tune_params): New member 'const vec_costs'.
16016 * config/aarch64/aarch64.c (generic_vector_cost): New.
16017 (generic_tunings): New member 'generic_vector_cost'.
16018 (aarch64_builtin_vectorization_cost): New.
16019 (aarch64_add_stmt_cost): New.
16020 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16021 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16023 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16026 * config/i386/predicates.md (vsib_address_operand): Disallow
16027 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16029 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16031 PR middle-end/55030
16032 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16033 expand_builtin_setjmp_receiver.
16034 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16035 with NULL for the label parameter.
16036 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16037 the frame-pointer. Adjust comments.
16038 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16039 only if LABEL is non-NULL.
16041 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16043 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16044 (ARM_ABI_AAPCS64): Ditto.
16046 (ARM_DEFAULT_ABI): Ditto.
16048 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16050 * config/aarch64/aarch64-builtins.c
16051 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16052 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16054 * config/aarch64/aarch64-simd.md
16055 (aarch64_ld1<VALL:mode>): New.
16056 (aarch64_st1<VALL:mode>): Likewise.
16057 * config/aarch64/arm_neon.h
16058 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16060 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16062 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16064 (ix86_option_override_internal):Turn on all -mavx target flags by
16065 default as they are dependent on AVX anyway.
16067 2013-07-02 Cary Coutant <ccoutant@google.com>
16069 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16070 deterministic hash.
16071 (loc_checksum_ordered): Likewise.
16072 (hash_loc_operands): Remove inline keyword.
16074 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16076 PR tree-optimization/57741
16077 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16078 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16079 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16080 Allow REAL_CST step_exprs if flag_associative_math.
16081 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16083 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16085 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16087 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16089 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16091 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16093 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16095 (iorsi3_insn): Likewise.
16096 (arm_xorsi3): Likewise.
16098 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16100 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16102 (attribute "type"): ... here, and prefix with "wmmx_".
16103 (attribute "core_cycles"): Update for attribute changes.
16104 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16105 (tbcstv4hi): Likewise.
16106 (tbcstv2si): Likewise.
16107 (iwmmxt_iordi3): Likewise.
16108 (iwmmxt_xordi3): Likewise.
16109 (iwmmxt_anddi3): Likewise.
16110 (iwmmxt_nanddi3): Likewise.
16111 (iwmmxt_arm_movdi): Likewise.
16112 (iwmmxt_movsi_insn): Likewise.
16113 (mov<mode>_internal): Likewise.
16114 (and<mode>3_iwmmxt): Likewise.
16115 (ior<mode>3_iwmmxt): Likewise.
16116 (xor<mode>3_iwmmxt): Likewise.
16117 (add<mode>3_iwmmxt): Likewise.
16118 (ssaddv8qi3): Likewise.
16119 (ssaddv4hi3): Likewise.
16120 (ssaddv2si3): Likewise.
16121 (usaddv8qi3): Likewise.
16122 (usaddv4hi3): Likewise.
16123 (usaddv2si3): Likewise.
16124 (sub<mode>3_iwmmxt): Likewise.
16125 (sssubv8qi3): Likewise.
16126 (sssubv4hi3): Likewise.
16127 (sssubv2si3): Likewise.
16128 (ussubv8qi3): Likewise.
16129 (ussubv4hi3): Likewise.
16130 (ussubv2si3): Likewise.
16131 (mulv4hi3_iwmmxt): Likewise.
16132 (smulv4hi3_highpart): Likewise.
16133 (umulv4hi3_highpart): Likewise.
16134 (iwmmxt_wmacs): Likewise.
16135 (iwmmxt_wmacsz): Likewise.
16136 (iwmmxt_wmacu): Likewise.
16137 (iwmmxt_wmacuz): Likewise.
16138 (iwmmxt_clrdi): Likewise.
16139 (iwmmxt_clrv8qi): Likewise.
16140 (iwmmxt_clr4hi): Likewise.
16141 (iwmmxt_clr2si): Likewise.
16142 (iwmmxt_uavgrndv8qi3): Likewise.
16143 (iwmmxt_uavgrndv4hi3): Likewise.
16144 (iwmmxt_uavgv8qi3): Likewise.
16145 (iwmmxt_uavgv4hi3): Likewise.
16146 (iwmmxt_tinsrb): Likewise.
16147 (iwmmxt_tinsrh): Likewise.
16148 (iwmmxt_tinsrw): Likewise.
16149 (iwmmxt_textrmub): Likewise.
16150 (iwmmxt_textrmsb): Likewise.
16151 (iwmmxt_textrmuh): Likewise.
16152 (iwmmxt_textrmsh): Likewise.
16153 (iwmmxt_textrmw): Likewise.
16154 (iwmxxt_wshufh): Likewise.
16155 (eqv8qi3): Likewise.
16156 (eqv4hi3): Likewise.
16157 (eqv2si3): Likewise.
16158 (gtuv8qi3): Likewise.
16159 (gtuv4hi3): Likewise.
16160 (gtuv2si3): Likewise.
16161 (gtv8qi3): Likewise.
16162 (gtv4hi3): Likewise.
16163 (gtv2si3): Likewise.
16164 (smax<mode>3_iwmmxt): Likewise.
16165 (umax<mode>3_iwmmxt): Likewise.
16166 (smin<mode>3_iwmmxt): Likewise.
16167 (umin<mode>3_iwmmxt): Likewise.
16168 (iwmmxt_wpackhss): Likewise.
16169 (iwmmxt_wpackwss): Likewise.
16170 (iwmmxt_wpackdss): Likewise.
16171 (iwmmxt_wpackhus): Likewise.
16172 (iwmmxt_wpackwus): Likewise.
16173 (iwmmxt_wpackdus): Likewise.
16174 (iwmmxt_wunpckihb): Likewise.
16175 (iwmmxt_wunpckihh): Likewise.
16176 (iwmmxt_wunpckihw): Likewise.
16177 (iwmmxt_wunpckilb): Likewise.
16178 (iwmmxt_wunpckilh): Likewise.
16179 (iwmmxt_wunpckilw): Likewise.
16180 (iwmmxt_wunpckehub): Likewise.
16181 (iwmmxt_wunpckehuh): Likewise.
16182 (iwmmxt_wunpckehuw): Likewise.
16183 (iwmmxt_wunpckehsb): Likewise.
16184 (iwmmxt_wunpckehsh): Likewise.
16185 (iwmmxt_wunpckehsw): Likewise.
16186 (iwmmxt_wunpckelub): Likewise.
16187 (iwmmxt_wunpckeluh): Likewise.
16188 (iwmmxt_wunpckeluw): Likewise.
16189 (iwmmxt_wunpckelsb): Likewise.
16190 (iwmmxt_wunpckelsh): Likewise.
16191 (iwmmxt_wunpckelsw): Likewise.
16192 (ror<mode>3): Likewise.
16193 (ashr<mode>3_iwmmxt): Likewise.
16194 (lshr<mode>3_iwmmxt): Likewise.
16195 (ashl<mode>3_iwmmxt): Likewise.
16196 (ror<mode>3_di): Likewise.
16197 (ashr<mode>3_di): Likewise.
16198 (lshr<mode>3_di): Likewise.
16199 (ashl<mode>3_di): Likewise.
16200 (iwmmxt_wmadds): Likewise.
16201 (iwmmxt_wmaddu): Likewise.
16202 (iwmmxt_tmia): Likewise.
16203 (iwmmxt_tmiaph): Likewise.
16204 (iwmmxt_tmiabb): Likewise.
16205 (iwmmxt_tmiatb): Likewise.
16206 (iwmmxt_tmiabt): Likewise.
16207 (iwmmxt_tmiatt): Likewise.
16208 (iwmmxt_tmovmskb): Likewise.
16209 (iwmmxt_tmovmskh): Likewise.
16210 (iwmmxt_tmovmskw): Likewise.
16211 (iwmmxt_waccb): Likewise.
16212 (iwmmxt_wacch): Likewise.
16213 (iwmmxt_waccw): Likewise.
16214 (iwmmxt_waligni): Likewise.
16215 (iwmmxt_walignr): Likewise.
16216 (iwmmxt_walignr0): Likewise.
16217 (iwmmxt_walignr1): Likewise.
16218 (iwmmxt_walignr2): Likewise.
16219 (iwmmxt_walignr3): Likewise.
16220 (iwmmxt_wsadb): Likewise.
16221 (iwmmxt_wsadh): Likewise.
16222 (iwmmxt_wsadbz): Likewise.
16223 (iwmmxt_wsadhz): Likewise.
16224 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16225 (iwmmxt_wabsdiffb): Likewise.
16226 (iwmmxt_wabsdiffh): Likewise.
16227 (iwmmxt_wabsdiffw): Likewise.
16228 (iwmmxt_waddsubhx): Likewise
16229 (iwmmxt_wsubaddhx): Likewise.
16230 (addc<mode>3): Likewise.
16231 (iwmmxt_avg4): Likewise.
16232 (iwmmxt_avg4r): Likewise.
16233 (iwmmxt_wmaddsx): Likewise.
16234 (iwmmxt_wmaddux): Likewise.
16235 (iwmmxt_wmaddsn): Likewise.
16236 (iwmmxt_wmaddun): Likewise.
16237 (iwmmxt_wmulwsm): Likewise.
16238 (iwmmxt_wmulwum): Likewise.
16239 (iwmmxt_wmulsmr): Likewise.
16240 (iwmmxt_wmulumr): Likewise.
16241 (iwmmxt_wmulwsmr): Likewise.
16242 (iwmmxt_wmulwumr): Likewise.
16243 (iwmmxt_wmulwl): Likewise.
16244 (iwmmxt_wqmulm): Likewise.
16245 (iwmmxt_wqmulwm): Likewise.
16246 (iwmmxt_wqmulmr): Likewise.
16247 (iwmmxt_wqmulwmr): Likewise.
16248 (iwmmxt_waddbhusm): Likewise.
16249 (iwmmxt_waddbhusl): Likewise.
16250 (iwmmxt_wqmiabb): Likewise.
16251 (iwmmxt_wqmiabt): Likewise.
16252 (iwmmxt_wqmiatb): Likewise.
16253 (iwmmxt_wqmiatt): Likewise.
16254 (iwmmxt_wqmiabbn): Likewise.
16255 (iwmmxt_wqmiabtn): Likewise.
16256 (iwmmxt_wqmiatbn): Likewise.
16257 (iwmmxt_wqmiattn): Likewise.
16258 (iwmmxt_wmiabb): Likewise.
16259 (iwmmxt_wmiabt): Likewise.
16260 (iwmmxt_wmiatb): Likewise.
16261 (iwmmxt_wmiatt): Likewise.
16262 (iwmmxt_wmiabbn): Likewise.
16263 (iwmmxt_wmiabtn): Likewise.
16264 (iwmmxt_wmiatbn): Likewise.
16265 (iwmmxt_wmiattn): Likewise.
16266 (iwmmxt_wmiawbb): Likewise.
16267 (iwmmxt_wmiawbt): Likewise.
16268 (iwmmxt_wmiawtb): Likewise.
16269 (iwmmxt_wmiawtt): Likewise.
16270 (iwmmxt_wmiawbbn): Likewise.
16271 (iwmmxt_wmiawbtn): Likewise.
16272 (iwmmxt_wmiawtbn): Likewise.
16273 (iwmmxt_wmiawttn): Likewise.
16274 (iwmmxt_wmerge): Likewise.
16275 (iwmmxt_tandc<mode>3): Likewise.
16276 (iwmmxt_torc<mode>3): Likewise.
16277 (iwmmxt_torvsc<mode>3): Likewise.
16278 (iwmmxt_textrc<mode>3): Likewise.
16279 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16280 (wmmxt_pack): Likewise.
16281 (wmmxt_mult_c1): Likewise.
16282 (wmmxt_mult_c2): Likewise.
16283 (wmmxt_alu_c1): Likewise.
16284 (wmmxt_alu_c2): Likewise.
16285 (wmmxt_alu_c3): Likewise.
16286 (wmmxt_transfer_c1): Likewise.
16287 (wmmxt_transfer_c2): Likewise.
16288 (wmmxt_transfer_c3): Likewise.
16289 (marvell_f_iwmmxt_wstr): Likewise.
16290 (marvell_f_iwmmxt_wldr): Likewise.
16292 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
16294 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
16296 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16299 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16300 * lra-constraints.c (need_for_split_p): Check call used hard regs
16301 living through calls.
16303 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
16304 call used regs for call insn.
16306 2013-06-28 Jakub Jelinek <jakub@redhat.com>
16309 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
16310 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
16312 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16314 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
16316 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
16318 * lra-constraints.c (need_for_split_p): Check call used hard regs
16319 living through calls.
16321 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
16324 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
16325 to tie with any other modes. Eliminate Altivec vector mode tests,
16326 since these are a subset of ALTIVEC or VSX vector modes. Simplify
16327 code, to return 0 if testing MODE2 for a condition, if we've
16328 already tested MODE1 for the same condition.
16330 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16332 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
16335 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16337 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16338 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
16340 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
16342 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
16344 (aarch64_symbolic_constant_p): Remove.
16345 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
16346 static. Fix line length and white space.
16347 (aarch64_symbolic_constant_p): Remove.
16348 * config/aarch64/predicates.md (aarch64_valid_symref):
16349 Use aarch64_classify_symbol_expression.
16351 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16353 * config/arm/constraints.md (Ts): New constraint.
16354 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
16356 (compare_scc): Use "Ts" constraint for operand 0.
16357 (ior_scc_scc): Likewise.
16358 (and_scc_scc): Likewise.
16359 (and_scc_scc_nodom): Likewise.
16360 (ior_scc_scc_cmp): Likewise for operand 7.
16361 (and_scc_scc_cmp): Likewise.
16362 * config/arm/thumb2.md (thumb2_movsi_insn):
16363 Add alternatives for 16-bit encodings.
16364 (thumb2_movhi_insn): Likewise.
16365 (thumb2_movsicc_insn): Likewise.
16366 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16367 (thumb2_negscc): Use "Ts" constraint.
16368 Move mvn instruction outside cond_exec block.
16369 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16370 for 16-bit encodings.
16372 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16374 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16376 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16377 (mulsi3subsi): Likewise.
16378 (mulsidi3adddi): Likewise.
16379 (mulsidi3_v6): Likewise.
16380 (umulsidi3_v6): Likewise.
16381 (umulsidi3adddi_v6): Likewise.
16382 (smulsi3_highpart_v6): Likewise.
16383 (umulsi3_highpart_v6): Likewise.
16384 (mulhisi3tb): Likewise.
16385 (mulhisi3bt): Likewise.
16386 (mulhisi3tt): Likewise.
16387 (maddhisi4): Likewise.
16388 (maddhisi4tb): Likewise.
16389 (maddhisi4tt): Likewise.
16390 (maddhidi4): Likewise.
16391 (maddhidi4tb): Likewise.
16392 (maddhidi4tt): Likewise.
16393 (zeroextractsi_compare0_scratch): Likewise.
16394 (insv_zero): Likewise.
16395 (insv_t2): Likewise.
16396 (anddi_notzesidi_di): Likewise.
16397 (anddi_notsesidi_di): Likewise.
16398 (andsi_notsi_si): Likewise.
16399 (iordi_zesidi_di): Likewise.
16400 (xordi_zesidi_di): Likewise.
16401 (andsi_iorsi3_notsi): Likewise.
16402 (smax_0): Likewise.
16403 (smax_m1): Likewise.
16404 (smin_0): Likewise.
16405 (not_shiftsi): Likewise.
16406 (unaligned_loadsi): Likewise.
16407 (unaligned_loadhis): Likewise.
16408 (unaligned_loadhiu): Likewise.
16409 (unaligned_storesi): Likewise.
16410 (unaligned_storehi): Likewise.
16411 (extv_reg): Likewise.
16412 (extzv_t2): Likewise.
16413 (divsi3): Likewise.
16414 (udivsi3): Likewise.
16415 (arm_zero_extendhisi2addsi): Likewise.
16416 (arm_zero_extendqisi2addsi): Likewise.
16417 (compareqi_eq0): Likewise.
16418 (arm_extendhisi2_v6): Likewise.
16419 (arm_extendqisi2addsi): Likewise.
16420 (arm_movt): Likewise.
16421 (thumb2_ldrd): Likewise.
16422 (thumb2_ldrd_base): Likewise.
16423 (thumb2_ldrd_base_neg): Likewise.
16424 (thumb2_strd): Likewise.
16425 (thumb2_strd_base): Likewise.
16426 (thumb2_strd_base_neg): Likewise.
16427 (arm_negsi2): Add alternative for 16-bit encoding.
16428 (arm_one_cmplsi2): Likewise.
16430 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16432 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16433 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16434 (movdfcc): Likewise.
16435 * config/arm/vfp.md (*thumb2_movsf_vfp):
16436 Disable predication for arm_restrict_it.
16437 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16438 (*thumb2_movdfcc_vfp): Likewise.
16439 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16440 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16441 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16442 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16443 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16444 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16445 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16446 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16447 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16448 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16449 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16450 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16451 Disable predication for arm_restrict_it.
16453 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16455 * config/i386/bmiintrin.h (_bextr_u32): New.
16456 (_bextr_u64): Ditto.
16458 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16460 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16461 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16462 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16463 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16464 * config/mips/n32-elf.h: ...this new file.
16466 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16469 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16470 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16472 2013-06-27 Catherine Moore <clm@codesourcery.com>
16474 * config/mips/mips-tables.opt: Regenerate.
16475 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16476 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16477 * doc/invoke.texi: Add -m14kc.
16479 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16482 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16483 constraints of operand 1 and 2.
16486 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16487 to match RTL canonicalization. Swap predicates and
16488 constraints of operand 1 and 2.
16490 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16492 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16493 Process OP_INOUT regs for splitting too.
16495 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16497 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16498 decl before the loop, initialize to NULL.
16499 (vectorizable_load): Initialize ptr_incr to NULL.
16501 2013-06-27 Martin Jambor <mjambor@suse.cz>
16504 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16505 * ipa-ref.c (ipa_maybe_record_reference): New function.
16506 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16507 * ipa-cp.c (create_specialized_node): Record potential references from
16509 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16511 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16513 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16514 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16515 'mode' to force_reg.
16516 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16517 (aarch64_expand_mov_immediate): Likewise.
16519 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16521 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16522 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16524 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16526 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16527 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16528 (struct machine_function): Add tbegin_p.
16529 (s390_canonicalize_comparison): Fold CC mode compares to
16530 conditional jump if possible.
16531 (s390_emit_jump): Return the emitted jump.
16532 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16533 Handle CCRAWmode compares.
16534 (s390_option_override): Default to -mhtm if available.
16535 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16536 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16537 FPRs instead of df_regs_ever_live_p.
16538 (s390_optimize_nonescaping_tx): New function.
16539 (s390_init_frame_layout): Extend clobbered_regs array to cover
16541 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16542 (s390_expand_tbegin): New function.
16543 (enum s390_builtin): New enum definition.
16544 (code_for_builtin): New array definition.
16545 (s390_init_builtins): New function.
16546 (s390_expand_builtin): New function.
16547 (TARGET_INIT_BUILTINS): Define.
16548 (TARGET_EXPAND_BUILTIN): Define.
16549 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16550 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16551 (s390_alc_comparison): Likewise.
16552 * config/s390/s390-modes.def: Add CCRAWmode.
16553 * config/s390/s390.h (processor_flags): Add PF_TX.
16554 (TARGET_CPU_HTM): Define macro.
16555 (TARGET_HTM): Define macro.
16556 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16557 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16558 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16559 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16560 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16561 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16562 ("*cc_to_int"): Move up.
16563 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16564 constants other than 0.
16565 ("*ccraw_to_int"): New insn and splitter definition.
16566 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16567 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16568 ("tx_assist"): New expander.
16569 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16570 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16571 * config/s390/s390.opt: Add -mhtm option.
16572 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16573 * config/s390/htmxlintrin.h: New file.
16574 * config/s390/htmintrin.h: New file.
16575 * config/s390/s390intrin.h: New file.
16576 * doc/extend.texi: Document htm builtins.
16577 * config.gcc: Add the new header files to extra_headers.
16579 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16581 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16582 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16584 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16585 Pat Haugen <pthaugen@us.ibm.com>
16586 Peter Bergner <bergner@vnet.ibm.com>
16588 * config/rs6000/power8.md: New.
16589 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16590 setting for power8 entry.
16591 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16592 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16593 test for Power4/Power5 only.
16594 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16596 (force_new_group): Adjust comment.
16597 * config/rs6000/rs6000.md: Include power8.md.
16599 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16601 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16602 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16604 (tune_params): Update comment.
16605 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16606 (arm_max_conditional_execute): New function.
16607 (thumb2_final_prescan_insn): Use max_insn_skipped and
16608 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16610 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16612 PR tree-optimization/57705
16613 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16614 SSA_NAME step, provided that it is not defined inside the loop.
16615 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16616 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16618 2013-06-25 Martin Jambor <mjambor@suse.cz>
16620 PR middle-end/57670
16621 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16622 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16624 (ipa_note_param_call): Initialize member_ptr flag.
16625 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16626 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16627 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16628 (ipa_read_indirect_edge_info): Likewise.
16630 2013-06-25 Richard Biener <rguenther@suse.de>
16632 PR middle-end/56977
16633 * passes.c (init_optimization_passes): Move pass_fold_builtins
16634 and pass_dce earlier with -Og.
16636 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16638 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16639 <BIT_FIELD_REF>: Remove trailing TAB.
16640 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16643 2013-06-24 Martin Jambor <mjambor@suse.cz>
16645 PR tree-optimization/57358
16646 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16647 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16648 (ipa_analyze_params_uses): Generate pessimistic info when true.
16650 2013-06-24 Martin Jambor <mjambor@suse.cz>
16652 PR tree-optimization/57539
16653 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16654 global.inlined_to of the new node to it. All callers changed.
16655 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16656 inlining_into, pass it to cgraph_clone_node.
16657 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16658 ipa_free_edge_args_substructures.
16659 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16660 rdesc linked list. Do not assert rdesc edges have inlined caller.
16661 Assert we have found an rdesc in the rdesc list.
16663 2013-06-24 Richard Biener <rguenther@suse.de>
16665 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16666 (pointer_set_lookup): Declare.
16667 (class pointer_map): New template class implementing a
16668 generic pointer to T map.
16669 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16670 pointer_map<T>::contains, pointer_map<T>::insert,
16671 pointer_map<T>::traverse): New functions.
16672 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16673 (pointer_set_lookup): New function.
16674 (pointer_set_contains): Use pointer_set_lookup.
16675 (pointer_set_insert): Likewise.
16676 (insert_aux): Remove.
16677 (struct pointer_map_t): Embed a pointer_set_t.
16678 (pointer_map_create): Adjust.
16679 (pointer_map_destroy): Likewise.
16680 (pointer_map_contains): Likewise.
16681 (pointer_map_insert): Likewise.
16682 (pointer_map_traverse): Likewise.
16683 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16684 pointer_map<unsigned> instead of a pointer_map_t.
16685 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16686 (streamer_tree_cache_lookup): Likewise.
16687 (streamer_tree_cache_create): Likewise.
16688 (streamer_tree_cache_delete): Likewise.
16689 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16690 pointer_map<unsigned> instead of a pointer_map_t.
16691 (lto_init_tree_ref_encoder): Adjust.
16692 (lto_destroy_tree_ref_encoder): Likewise.
16693 * lto-section-out.c (lto_output_decl_index): Likewise.
16694 (lto_record_function_out_decl_state): Likewise.
16695 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16697 2013-06-24 Richard Biener <rguenther@suse.de>
16699 PR tree-optimization/57488
16700 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16702 2013-06-24 Alan Modra <amodra@gmail.com>
16704 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16705 (gen_easy_altivec_constant): Likewise.
16706 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16707 easy_vector_constant_msb): Likewise.
16709 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16712 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16713 add missing return true.
16715 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16718 * config/sh/predicates.md (general_extend_operand): Invoke
16719 general_movsrc_operand for memory operands.
16720 (general_movsrc_operand): Allow reg+reg addressing, do not use
16721 general_operand for memory operands.
16723 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16725 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16726 when current target options does not apply.
16727 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16728 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16729 * config/i386/bmiintrin.h: Pass appropriate target
16730 attributes to header.
16731 * config/i386/mmintrin.h: Ditto.
16732 * config/i386/nmmintrin.h: Ditto.
16733 * config/i386/avx2intrin.h: Ditto.
16734 * config/i386/fxsrintrin.h: Ditto.
16735 * config/i386/tbmintrin.h: Ditto.
16736 * config/i386/xsaveintrin.h: Ditto.
16737 * config/i386/f16cintrin.h: Ditto.
16738 * config/i386/xtestintrin.h: Ditto.
16739 * config/i386/xsaveoptintrin.h: Ditto.
16740 * config/i386/bmi2intrin.h: Ditto.
16741 * config/i386/lzcntintrin.h: Ditto.
16742 * config/i386/smmintrin.h: Ditto.
16743 * config/i386/wmmintrin.h: Ditto.
16744 * config/i386/x86intrin.h: Remove all header include guards.
16745 * config/i386/prfchwintrin.h: Ditto.
16746 * config/i386/pmmintrin.h: Ditto.
16747 * config/i386/tmmintrin.h: Ditto.
16748 * config/i386/xmmintrin.h: Ditto.
16749 * config/i386/popcntintrin.h: Ditto.
16750 * config/i386/rdseedintrin.h: Ditto.
16751 * config/i386/ammintrin.h: Ditto.
16752 * config/i386/emmintrin.h: Ditto.
16753 * config/i386/immintrin.h: Remove all header include guards.
16754 * config/i386/fma4intrin.h: Ditto.
16755 * config/i386/lwpintrin.h: Ditto.
16756 * config/i386/xopintrin.h: Ditto.
16757 * config/i386/ia32intrin.h: Ditto.
16758 * config/i386/avxintrin.h: Ditto.
16759 * config/i386/rtmintrin.h: Ditto.
16760 * config/i386/fmaintrin.h: Ditto.
16761 * config/i386/mm3dnow.h: Ditto.
16763 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16765 * common/config/i386/i386-common.c: Handle LZCNT.
16767 2013-06-22 Andi Kleen <ak@linux.intel.com>
16769 * doc/extend.texi: Use __atomic_store_n instead of
16770 __atomic_store in HLE example.
16772 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16774 * config/sh/sh.c: Remove <cstdlib> workaround.
16776 2013-06-21 Andi Kleen <ak@linux.intel.com>
16778 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16780 2013-06-21 Andi Kleen <ak@linux.intel.com>
16782 * doc/extend.texi: Document that __atomic_clear and
16783 __atomic_test_and_set should only be used with bool.
16785 2013-06-20 Jan Hubicka <jh@suse.cz>
16787 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16788 types_same_for_odr.
16789 * tree.c (decls_same_for_odr): New function.
16790 (same_for_edr): New function.
16791 (types_same_for_odr): New function.
16792 (get_binfo_at_offset): Use it.
16793 * tree.h (types_same_for_odr): Declare.
16795 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16796 Jason Merrill <jason@redhat.com>
16798 * system.h: Include <cstdlib> as well as <stdlib.h>.
16800 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16803 * config/i386/i386.c (construct_container): Report error if
16804 long double is used with disabled x87 float returns.
16806 2013-06-20 Jan Hubicka <jh@suse.cz>
16808 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16810 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16812 PR rtl-optimization/57425
16813 PR rtl-optimization/57569
16814 * alias.c (write_dependence_p): Remove parameters mem_mode and
16815 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16816 Changed all callers.
16817 (canon_anti_dependence): Get comments and semantics in sync.
16818 Add parameter mem_canonicalized. Changed all callers.
16819 * rtl.h (canon_anti_dependence): Update prototype.
16821 2013-06-20 Richard Biener <rguenther@suse.de>
16823 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16824 byte case, inline streamer_read_uchar and defer section
16827 2013-06-20 Richard Biener <rguenther@suse.de>
16829 PR tree-optimization/57584
16830 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16831 SSA names into the expanded expression that take part in
16832 abnormal coalescing.
16834 2013-06-19 Sharad Singhai <singhai@google.com>
16836 * gcov.c (print_usage): Handle new option.
16837 (process_args): Ditto.
16838 (get_gcov_intermediate_filename): New function.
16839 (output_intermediate_file): New function.
16840 (output_gcov_file): New function
16841 (generate_results): Handle new option.
16842 (release_function): Relase demangled name.
16843 (read_graph_file): Handle demangled name.
16844 (output_lines): Ditto.
16845 * doc/gcov.texi: Document gcov intermediate format.
16847 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16850 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16851 (lra_emit_add): Use the functions. Add comment about Y as an
16854 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16857 * collect2.c (collect_atexit): New.
16858 (collect_exit): Delete.
16859 (main): Register collect_atexit with atexit.
16860 (collect_wait): Change collect_exit to exit.
16862 * collect2.h (collect_exit): Delete.
16863 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16865 2013-06-19 Wei Mi <wmi@google.com>
16867 PR rtl-optimization/57518
16868 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16869 if regno is used in paradoxical subreg.
16870 (update_equiv_regs): Check pdx_subregs[regno] before
16871 set a reg to be equivalent with a mem.
16873 2013-06-19 Matthias Klose <doko@ubuntu.com>
16876 * file-find.h (find_a_file): Add a mode parameter.
16877 * file-find.c (find_a_file): Likewise.
16878 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16879 with X_OK for the executables.
16880 * collect2.c (main): Call find_a_file with X_OK.
16882 2013-06-19 Steve Ellcey <sellcey@mips.com>
16885 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16886 Use NEXT_INSN instead of next_real_insn.
16888 2013-06-19 Jan Hubicka <jh@suse.cz>
16890 * cgraph.h (const_value_known_p): Replace by ...
16891 (ctor_for_folding): .. this one.
16892 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16893 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16894 * expr.c (expand_expr_real_1): Likewise.
16895 (string_constant): Likewise.
16896 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16897 * ipa.c (process_references): Likewise.
16898 (symtab_remove_unreachable_nodes): Likewise.
16899 * ipa-inline-analysis.c (param_change_prob): Likewise.
16900 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16901 (get_base_constructor): Likwise.
16902 * varpool.c (varpool_remove_node): Likewise.
16903 (varpool_remove_initializer): LIkewise.
16904 (dump_varpool_node): LIkwise.
16905 (const_value_known_p): Rewrite to ...
16906 (ctor_for_folding): ... this one.
16908 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16911 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16912 PERSONALITY in $PATH derived prefixes.
16914 2013-06-19 Jeff Law <law@redhat.com>
16916 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16919 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16920 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16923 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16925 * config/arm/vfp.md: Move VFP instruction classification documentation
16927 * config/arm/arm.md: ... here. Update instruction classification
16930 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16932 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16933 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16934 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16936 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16939 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16940 with NEXT_INSN. Use tablejump_p to check for jump table data
16943 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16946 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16947 that now in C++ the value is correct per the C++ standards.
16949 2013-06-19 Richard Biener <rguenther@suse.de>
16951 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16952 for global context.
16954 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16957 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16960 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16961 with next_active_insn.
16963 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16965 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16966 functions are inlined during failures to flag an error.
16967 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16968 in early inline pass.
16970 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16972 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16975 2013-06-18 Julian Brown <julian@codesourcery.com>
16977 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16978 Permit virtual register pre-reload if !strict.
16979 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16981 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16983 * config/arm/neon.md (movmisalign<mode>): Use
16984 neon_perm_struct_or_reg_operand instead of
16985 neon_struct_or_register_operand.
16986 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16987 neon_permissive_struct_operand instead of neon_struct_operand.
16988 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16989 neon_vector_mem_operand.
16990 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16991 neon_vector_mem_operand.
16992 (neon_permissive_struct_operand): New.
16993 (neon_struct_or_register_operand): Rename to...
16994 (neon_perm_struct_or_reg_operand): This. Adjust call to
16995 neon_vector_mem_operand.
16997 2013-06-18 Richard Biener <rguenther@suse.de>
16999 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17000 * lto-streamer.h: Include pointer-set.h.
17001 (struct lto_decl_slot): Remove.
17002 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17003 Remove next_index entry.
17004 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17005 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17006 (lto_init_tree_ref_encoder): Adjust.
17007 (lto_destroy_tree_ref_encoder): Likewise.
17008 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17009 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17010 (lto_output_decl_index): Adjust.
17011 (lto_new_out_decl_state): Likewise.
17012 (lto_record_function_out_decl_state): Likewise.
17013 * lto-streamer-out.c (copy_function): Likewise.
17015 2013-06-18 Richard Biener <rguenther@suse.de>
17017 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17018 * cgraphunit.c: Include cfgloop.h.
17019 (init_lowered_empty_function): Initialize the loop tree.
17020 (assemble_thunk): Insert new BBs into loops.
17022 2013-06-18 Richard Biener <rguenther@suse.de>
17024 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17025 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17026 the map from cache entry to cache index optional.
17027 (streamer_tree_cache_replace_tree): Adjust accordingly.
17028 (streamer_tree_cache_append): Likewise.
17029 (streamer_tree_cache_delete): Likewise.
17030 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17031 streamer cache map from cache entry to cache index.
17032 * lto-streamer-out.c (create_output_block): Adjust.
17034 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17036 * config/arm/arm.md (attribute "insn"): Move multiplication and
17037 division attributes to...
17038 (attribute "type"): ... here. Remove mult.
17039 (attribute "mul32"): New attribute.
17040 (attribute "mul64"): Add umaal.
17041 (*arm_mulsi3): Update attributes.
17042 (*arm_mulsi3_v6): Likewise.
17043 (*thumb_mulsi3): Likewise.
17044 (*thumb_mulsi3_v6): Likewise.
17045 (*mulsi3_compare0): Likewise.
17046 (*mulsi3_compare0_v6): Likewise.
17047 (*mulsi_compare0_scratch): Likewise.
17048 (*mulsi_compare0_scratch_v6): Likewise.
17049 (*mulsi3addsi): Likewise.
17050 (*mulsi3addsi_v6): Likewise.
17051 (*mulsi3addsi_compare0): Likewise.
17052 (*mulsi3addsi_compare0_v6): Likewise.
17053 (*mulsi3addsi_compare0_scratch): Likewise.
17054 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17055 (*mulsi3subsi): Likewise.
17056 (*mulsidi3adddi): Likewise.
17057 (*mulsi3addsi_v6): Likewise.
17058 (*mulsidi3adddi_v6): Likewise.
17059 (*mulsidi3_nov6): Likewise.
17060 (*mulsidi3_v6): Likewise.
17061 (*umulsidi3_nov6): Likewise.
17062 (*umulsidi3_v6): Likewise.
17063 (*umulsidi3adddi): Likewise.
17064 (*umulsidi3adddi_v6): Likewise.
17065 (*smulsi3_highpart_nov6): Likewise.
17066 (*smulsi3_highpart_v6): Likewise.
17067 (*umulsi3_highpart_nov6): Likewise.
17068 (*umulsi3_highpart_v6): Likewise.
17069 (mulhisi3): Likewise.
17070 (*mulhisi3tb): Likewise.
17071 (*mulhisi3bt): Likewise.
17072 (*mulhisi3tt): Likewise.
17073 (maddhisi4): Likewise.
17074 (*maddhisi4tb): Likewise.
17075 (*maddhisi4tt): Likewise.
17076 (maddhidi4): Likewise.
17077 (*maddhidi4tb): Likewise.
17078 (*maddhidi4tt): Likewise.
17079 (divsi3): Likewise.
17080 (udivsi3): Likewise.
17081 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17082 (thumb2_mulsi_short_compare0): Likewise.
17083 (thumb2_mulsi_short_compare0_scratch): Likewise.
17084 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17085 (1020mult2): Likewise.
17086 (1020mult3): Likewise.
17087 (1020mult4): Likewise.
17088 (1020mult5): Likewise.
17089 (1020mult6): Likewise.
17090 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17092 (cortex_a15_mult64): Likewise.
17093 (cortex_a15_sdiv): Likewise.
17094 (cortex_a15_udiv): Likewise.
17095 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17101 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17102 (pj4_ir_div): Likewise.
17103 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17104 (11_mult2): Likewise.
17105 (11_mult3): Likewise.
17106 (11_mult4): Likewise.
17107 (11_mult5): Likewise.
17108 (11_mult6): Likewise.
17109 (11_mult7): Likewise.
17110 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17111 (cortex_a8_mla): Likewise.
17112 (cortex_a8_mull): Likewise.
17113 (cortex_a8_smulwy): Likewise.
17114 (cortex_a8_smlald): Likewise.
17115 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17116 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17117 (cortex_r4_mul_3): Likewise.
17118 (cortex_r4_mla_4): Likewise.
17119 (cortex_r4_mla_3): Likewise.
17120 (cortex_r4_smlald): Likewise.
17121 (cortex_r4_mull): Likewise.
17122 (cortex_r4_sdiv): Likewise.
17123 (cortex_r4_udiv): Likewise.
17124 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17125 (cortex_a7_idiv): Likewise.
17126 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17127 (9_mult2): Likewise.
17128 (9_mult3): Likewise.
17129 (9_mult4): Likewise.
17130 (9_mult5): Likewise.
17131 (9_mult6): Likewise.
17132 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17133 (cortex_a53_sdiv): Likewise.
17134 (cortex_a53_udiv): Likewise.
17135 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17136 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17137 (mp626_mult2): Likewise.
17138 (mp626_mult3): Likewise.
17139 (mp626_mult4): Likewise.
17140 * config/arm/fa526.md (526_mult1): Update attribute change.
17141 (526_mult2): Likewise.
17142 * config/arm/arm-generic.md (mult): Update attribute change.
17143 (mult_ldsched_strongarm): Likewise.
17144 (mult_ldsched): Likewise.
17145 (multi_cycle): Likewise.
17146 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17147 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17148 (606te_mult2): Likewise.
17149 (606te_mult3): Likewise.
17150 (606te_mult4): Likewise.
17151 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17152 (cortex_a9_mac16): Likewise.
17153 (cortex_a9_multiply): Likewise.
17154 (cortex_a9_mac): Likewise.
17155 (cortex_a9_multiply_long): Likewise.
17156 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17157 (626te_mult2): Likewise.
17158 (626te_mult3): Likewise.
17159 (626te_mult4): Likewise.
17161 2013-06-18 Richard Biener <rguenther@suse.de>
17164 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17166 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17169 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17170 with next_active_insn.
17172 2013-06-18 Alan Modra <amodra@gmail.com>
17174 * config/rs6000/rs6000.h (enum data_align): New.
17175 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17176 (DATA_ABI_ALIGNMENT): Define.
17177 (CONSTANT_ALIGNMENT): Correct comment.
17178 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17179 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17181 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17183 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17184 ATTRIBUTE_UNUSED marking.
17186 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17188 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17189 alternative and update.
17190 (aarch64_dup_lanedi): Delete.
17191 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17192 * config/aarch64/aarch64-simd-builtins.def: Update.
17194 2013-06-17 Richard Biener <rguenther@suse.de>
17196 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17197 (lto_input_scc): Declare.
17198 (lto_input_tree_1): Likewise.
17199 (struct lto_stats_d): Add num_tree_bodies_output and
17200 num_pickle_refs_output.
17201 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17202 (lto_read_tree_1): Split out from ...
17203 (lto_read_tree): ... this.
17204 (lto_input_scc): New function.
17205 (lto_input_tree_1): Split out from ...
17206 (lto_input_tree): ... this. Handle LTO_tree_scc.
17207 (lto_data_in_create): Create the streamer cache without hashes.
17208 * lto-streamer-out.c (create_output_block): Create the streamer
17209 cache with hashes when not doing WPA.
17210 (lto_write_tree_1): Split out from ...
17211 (lto_write_tree): ... this.
17212 (get_symbol_initial_value): New function.
17213 (lto_output_tree_1): Split out from ...
17214 (lto_output_tree): ... this. Write trees as series of SCCs
17215 using a DFS walk via DFS_write_tree.
17216 (struct sccs, struct scc_entry): New types.
17217 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17218 (DFS_write_tree_body): New function.
17219 (DFS_write_tree): Likewise.
17220 (hash_tree): Likewise.
17221 (scc_entry_compare): Likewise.
17222 (hash_scc): Likewise.
17223 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17224 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17225 TREE_CHAIN as regular reference.
17226 (streamer_read_integer_cst): Remove.
17227 (streamer_get_pickled_tree): Adjust.
17228 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17229 of DECL_EXTERNALs in BLOCK_VARS for now.
17230 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17232 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17233 Add hash value argument and record that if hashes are recorded
17235 (streamer_tree_cache_insert_1): Adjust.
17236 (streamer_tree_cache_insert): Likewise.
17237 (streamer_tree_cache_insert_at): Rename to ...
17238 (streamer_tree_cache_replace_tree): ... this and adjust.
17239 (streamer_tree_cache_append): Adjust.
17240 (record_common_node): Likewise.
17241 (streamer_tree_cache_create): Add argument whether to
17242 record hash values together with trees.
17243 (streamer_tree_cache_delete): Adjust.
17244 * tree-streamer.h (struct streamer_tree_cache_d): Add
17246 (streamer_read_integer_cst): Remove.
17247 (streamer_tree_cache_insert): Adjust.
17248 (streamer_tree_cache_append): Likewise.
17249 (streamer_tree_cache_insert_at): Rename to ...
17250 (streamer_tree_cache_replace_tree): ... this and adjust.
17251 (streamer_tree_cache_create): Add argument whether to record hashes.
17252 (streamer_tree_cache_get): Rename to ...
17253 (streamer_tree_cache_get_tree): ... this.
17254 (streamer_tree_cache_get_hash): New function.
17255 * tree.c (cache_integer_cst): New function.
17256 * tree.h (cache_integer_cst): Declare.
17257 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17258 * lto-symtab.c (lto_varpool_replace_node): Only release
17259 DECL_INITIAL of non-prevailing decls.
17260 * varpool.c (varpool_remove_initializer): Do not release
17261 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17263 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17265 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17266 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17267 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17268 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17269 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17270 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17271 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17272 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17273 instead of TARGET_64BIT.
17274 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17275 Require ISA_HAS_<D>DIV.
17277 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17279 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17280 (mips*-*-linux*): Move default with_llsc setting to where other
17282 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17284 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17285 Likewise. Remove default with_tune setting. Move default float
17286 setting to its own block. Handle with_llsc in the same block as above.
17288 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
17290 PR rtl-optimization/57425
17291 PR rtl-optimization/57569
17292 * alias.c (write_dependence_p): Add new parameters mem_mode,
17293 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
17294 Changed all callers.
17295 (canon_anti_dependence): New function.
17296 * cse.c (check_dependence): Use canon_anti_dependence.
17297 * cselib.c (cselib_invalidate_mem): Likewise.
17298 * rtl.h (canon_anti_dependence): Declare.
17300 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17302 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
17303 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
17304 ".set mips3" for 64-bit targets.
17306 2013-06-15 Dehao Chen <dehao@google.com>
17308 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
17309 * gimple-low.c (gimple_check_call_matching_types): Likewise.
17310 (gimple_check_call_args): Likewise.
17311 * value-prof.c (check_ic_target): Likewise.
17312 * ipa-inline.c (early_inliner): Likewise.
17313 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
17314 * cgraph.c (cgraph_create_edge_1): Likewise.
17315 (cgraph_make_edge_direct): Likewise.
17317 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
17320 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
17321 rs6000_output_move_128bit to handle emitting quad memory
17322 operations. Set attribute length to 8 bytes.
17324 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
17326 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
17328 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
17329 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
17330 (aarch64_<su>mlsl<mode>): Likewise.
17332 2013-06-14 Mike Stump <mikestump@comcast.net>
17334 * Makefile.in (TARGET_H): Add insn-codes.h.
17336 2013-06-14 Alan Modra <amodra@gmail.com>
17338 PR middle-end/57134
17339 PR middle-end/57586
17340 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
17341 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
17342 bitfield expansion when EXPAND_MEMORY.
17343 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
17345 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
17347 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
17348 test for clearing quad memory on 32-bit later.
17350 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17352 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
17353 (fold_negate_expr): Likewise.
17354 (fold_real_zero_addition_p): Handle vectors.
17355 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
17357 2013-06-14 Alan Modra <amodra@gmail.com>
17359 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17361 2013-06-13 Jan Hubicka <jh@suse.cz>
17363 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17364 Local comdats are not externally visible.
17365 * symtab.c (dump_symtab_base): Dump externally visible.
17366 (verify_symtab_base): Verify back links in the symtab hash.
17368 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17370 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17371 CONVERT_EXPR as equal nodes.
17373 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17375 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17377 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17379 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17380 Generalize to complex and vector.
17381 * tree.c (build_all_ones_cst): New function.
17382 * tree.h (build_all_ones_cst): Declare it.
17384 2013-06-13 Alan Modra <amodra@gmail.com>
17386 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17387 * config/rs6000/rs6000.md (signbittf2): New insn.
17388 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17389 (abstf2_internal, cmptf_internal2): Likewise.
17390 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17392 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17393 Pat Haugen <pthaugen@us.ibm.com>
17394 Peter Bergner <bergner@vnet.ibm.com>
17396 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17397 power8 byte, half-word, and quad-word atomic instructions.
17398 (emit_store_conditional): Likewise.
17399 (rs6000_expand_atomic_compare_and_swap): Likewise.
17400 (rs6000_expand_atomic_op): Likewise.
17402 * config/rs6000/sync.md (larx): Add new modes for power8.
17404 (AINT): New mode iterator to include TImode as well as normal
17405 integer modes on power8.
17406 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17407 that VSX registers are not considered. Use AINT mode iterator
17408 instead of INT1 to allow inclusion of quad word atomic operations
17410 (load_locked<mode>): Likewise.
17411 (store_conditional<mode>): Likewise.
17412 (atomic_compare_and_swap<mode>): Likewise.
17413 (atomic_exchange<mode>): Likewise.
17414 (atomic_nand<mode>): Likewise.
17415 (atomic_fetch_<fetchop_name><mode>): Likewise.
17416 (atomic_nand_fetch<mode>): Likewise.
17417 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17419 (ATOMIC): On power8, add QImode, HImode modes.
17420 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17421 modes that promote to SImode.
17422 (load_lockedti): Convert TImode arguments to PTImode, so that we
17423 get a guaranteed even/odd register pair.
17424 (load_lockedpti): Likewise.
17425 (store_conditionalti): Likewise.
17426 (store_conditionalpti): Likewise.
17428 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17429 atomic load/store instructions.
17432 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17434 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17436 (insn_count): New attribute, with most cases extracted from...
17437 (length): ...here. Redefine most cases in terms of insn_count.
17438 (single_insn): Delete.
17439 (can_delay): Use insn_count to check for single instructions.
17440 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17441 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17442 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17443 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17444 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17445 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17446 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17447 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17448 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17449 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17450 rather than "length".
17451 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17452 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17453 Use "insn_count" rather than "length".
17454 * config/mips/mips-dsp.md
17455 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17456 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17459 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17461 PR tree-optimization/57361
17462 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17464 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17466 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17468 (aarch64_simd_combine<mode>): New instruction expansion.
17469 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17470 function prototype.
17471 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17472 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17474 2013-06-12 Jan Hubicka <jh@suse.cz>
17476 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17477 decl has when in streaming stage.
17478 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17479 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17481 2013-06-12 Roland Stigge <stigge@antcom.de>
17484 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17486 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17488 PR tree-optimization/57537
17489 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17490 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17492 2013-06-12 Richard Biener <rguenther@suse.de>
17494 * data-streamer.h (streamer_write_char_stream): CSE
17495 obs->current_pointer.
17496 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17497 streamer_write_char_stream manually and optimize the resulting loop.
17498 (streamer_write_hwi_stream): Likewise.
17500 2013-06-12 Jan Hubicka <jh@suse.cz>
17502 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17503 * cgraph.h (varpool_create_empty_node): Declare.
17504 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17506 * symtab.c (symtab_unregister_node): Be lax about missin entries
17508 (symtab_get_node): Update comment.
17509 * varpool.c (varpool_create_empty_node): Break out from ...
17510 (varpool_node_for_decl): ... here.
17511 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17513 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17515 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17516 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17517 part. Use straight-line flow at the end.
17518 <COMPONENT_REF>: Remove superfluous else.
17519 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17521 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17524 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17525 target hook even for !TREE_PUBLIC decls. If no resolution info
17526 is available, return false for common and external decls.
17528 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17530 * config/rl78/constraints.md (U): New constraint.
17531 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17534 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17537 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17538 to allow returning address to AT_PLATFORM name.
17540 2013-06-11 Jan Hubicka <jh@suse.cz>
17542 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17543 * cgraph.h (symtab_node_base): Add weakref flag.
17544 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17545 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17546 (output_weakrefs): Use weakref flag.
17547 * fold-const.c (simple_operand_p): Handle WEAK.
17548 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17549 * ipa.c (varpool_externally_visible_p): Drop weakref.
17550 (function_and_variable_visibility): Update comment; fix weakref
17551 sanity checks; do not clear DECL_WEAK on them.
17552 * lto-cgraph.c (lto_output_node): update.
17553 (lto_output_varpool_node): Update.
17554 (input_overwrite_node): Update.
17555 (input_node): Update.
17556 (input_varpool_node): Update.
17557 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17558 (lto_symtab_merge_symbols): Add sanity check.
17559 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17560 * passes.c (rest_of_decl_compilation): Set static flag, too.
17561 * symtab.c (dump_symtab_base): Dump weakref.
17562 (verify_symtab_base): Sanity check weakrefs.
17563 (symtab_make_decl_local): Remove duplicated code.
17564 (symtab_alias_ultimate_target): Simplify.
17565 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17567 2013-06-11 Tom de Vries <tom@codesourcery.com>
17569 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17570 sequence_vect == NULL.
17572 2013-06-11 DJ Delorie <dj@redhat.com>
17574 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17575 (rl78_unwind_word_mode): New.
17577 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17579 * final.c (debug_prefix_maps): Make static.
17581 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17583 * function.c (initial_trampoline): Remove stray copy.
17585 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17587 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17589 2013-06-11 Martin Jambor <mjambor@suse.cz>
17591 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17592 within bounds at the beginning of the function.
17594 2013-06-11 Alan Modra <amodra@gmail.com>
17596 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17598 (get_named_section): Don't NULL !DECL_P decl.
17600 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17602 * doc/invoke.texi (core-avx2): Document.
17604 (atom): Updated with MOVBE.
17606 2013-06-11 Richard Biener <rguenther@suse.de>
17608 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17610 2013-06-11 Anton Blanchard <anton@samba.org>
17612 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17613 correct shift value in little-endian mode.
17615 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17618 * varasm.c (get_variable_align): Move #endif to the right place.
17620 2013-06-10 Cary Coutant <ccoutant@google.com>
17622 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17623 for hash so that hash table traversal order is deterministic.
17625 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17626 Pat Haugen <pthaugen@us.ibm.com>
17627 Peter Bergner <bergner@vnet.ibm.com>
17629 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17630 of vectors in GPRS if they are direct moves or quad word load or
17633 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17635 (direct_move_p): Likewise.
17636 (quad_load_store_p): Likewise.
17638 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17639 classes into bins based on the physical register type.
17640 (reg_class_to_reg_type): Likewise.
17641 (IS_STD_REG_TYPE): Likewise.
17642 (IS_FP_VECT_REG_TYPE): Likewise.
17643 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17644 use direct move instructions.
17645 (reload_gpr_vsx): Likewise.
17646 (reload_vsx_gpr): Likewise.
17647 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17648 information that is a simplification of register classes. Also
17649 precalculate direct move reload helpers.
17650 (direct_move_p): New function to return true if the operation can
17651 be done as a direct move instruciton.
17652 (quad_load_store_p): New function to return true if the operation
17653 is a quad memory operation.
17654 (rs6000_legitimize_address): If quad memory, only allow register
17655 indirect for TImode addresses.
17656 (rs6000_legitimate_address_p): Likewise.
17657 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17658 (rs6000_reload_register_type): Likewise.
17659 (register_to_reg_type): Return register type.
17660 (rs6000_secondary_reload_simple_move): New helper function for
17661 secondary reload and secondary memory needed to identify anything
17662 that is a simple move, and does not need reloading.
17663 (rs6000_secondary_reload_direct_move): New helper function for
17664 secondary reload to identify cases that can be done with several
17665 instructions via the direct move instructions.
17666 (rs6000_secondary_reload_move): New helper function for secondary
17667 reload to identify moves between register types that can be done.
17668 (rs6000_secondary_reload): Add support for quad memory operations
17669 and for direct move.
17670 (rs6000_secondary_memory_needed): Likewise.
17671 (rs6000_debug_secondary_memory_needed): Change argument names.
17672 (rs6000_output_move_128bit): New function to return the move to
17673 use for 128-bit moves, including knowing about the various
17674 limitations of quad memory operations.
17676 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17677 memory operations. call rs6000_output_move_128bit for the actual
17678 instruciton(s) to generate.
17679 (vsx_movti_64bit): Likewise.
17681 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17682 (UNSPEC_P8V_MTVSRWZ): Likewise.
17683 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17684 (UNSPEC_P8V_MTVSRD): Likewise.
17685 (UNSPEC_P8V_XXPERMDI): Likewise.
17686 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17687 (UNSPEC_FUSION_GPR): Likewise.
17688 (FMOVE128_GPR): New iterator for direct move.
17689 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17690 (f32_sv): Likewise.
17691 (f32_dm): Likewise.
17692 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17693 loads and direct move instructions.
17694 (zero_extendsidi2_lfiwzx): Likewise.
17695 (extendsidi2_lfiwax): Likewise.
17696 (extendsidi2_nocell): Likewise.
17697 (floatsi<mode>2_lfiwax): Likewise.
17698 (lfiwax): Likewise.
17699 (floatunssi<mode>2_lfiwzx): Likewise.
17700 (lfiwzx): Likewise.
17701 (fix_trunc<mode>_stfiwx): Likewise.
17702 (fixuns_trunc<mode>_stfiwx): Likewise.
17703 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17704 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17705 (parity<mode>2_cmpb): Set length/type attr.
17706 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17707 for 'mr.' to fast_compare.
17708 (bpermd_<mode>): Change type attr to popcnt.
17709 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17710 (p8_mtvsrwz_1): Likewise.
17711 (p8_mtvsrwz_2): Likewise.
17712 (reload_fpr_from_gpr<mode>): Likewise.
17713 (p8_mtvsrd_1): Likewise.
17714 (p8_mtvsrd_2): Likewise.
17715 (p8_xxpermdi_<mode>): Likewise.
17716 (reload_vsx_from_gpr<mode>): Likewise.
17717 (reload_vsx_from_gprsf): Likewise.
17718 (p8_mfvsrd_3_<mode>): LIkewise.
17719 (reload_gpr_from_vsx<mode>): Likewise.
17720 (reload_gpr_from_vsxsf): Likewise.
17721 (p8_mfvsrd_4_disf): Likewise.
17722 (multi-word GPR splits): Do not split direct moves or quad memory
17725 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17727 * tree-into-ssa.c (interesting_blocks): Make static.
17729 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17732 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17733 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17734 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17735 (get_variable_align): New function.
17736 (get_variable_section, emit_bss, emit_common,
17737 assemble_variable_contents, place_block_symbol): Use
17738 get_variable_align instead of DECL_ALIGN.
17739 (assemble_noswitch_variable): Add align argument, use it
17740 instead of DECL_ALIGN.
17741 (assemble_variable): Adjust caller. Use get_variable_align
17742 instead of DECL_ALIGN.
17743 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17745 (DATA_ABI_ALIGNMENT): Define.
17746 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17747 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17748 opt is false, only return the psABI mandated alignment increase.
17749 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17750 (DATA_ABI_ALIGNMENT): ... this.
17751 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17752 (DATA_ABI_ALIGNMENT): ... this.
17753 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17754 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17755 (DATA_ABI_ALIGNMENT): ... this.
17756 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17757 * doc/tm.texi: Regenerated.
17759 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17761 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17762 cmp_code to construct REG_EQUAL note.
17764 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17767 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17768 that operands[2] doesn't overlap with operands[0].
17770 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17771 Jan Hubicka <jh@suse.cz>
17773 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17774 hack to mark symbols as used.
17776 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17778 PR rtl-optimization/57559
17779 * lra-constraints.c (process_alt_operands): Don't discourage
17780 memory with known offset for offsetable memory constraint.
17781 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17783 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17785 * varasm.c (struct oc_local_state): Reorder fields.
17786 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17787 and adjust accordingly.
17788 (output_constructor): Reorder initialization code and adjust call to
17789 output_constructor_bitfield.
17791 2013-06-07 Jan Hubicka <jh@suse.cz>
17793 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17795 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17797 * tree-object-size.c (unknown): Make const.
17799 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17801 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17802 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17803 for last alternative in the cpu_facility attribute.
17805 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17808 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17809 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17810 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17811 * config/arm/constraints.md (Dg): New constraint.
17812 * config/arm/neon.md (xordi3_neon): Remove.
17813 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17814 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17816 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17818 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17819 Clean up alternatives.
17821 2013-06-07 Alan Modra <amodra@gmail.com>
17823 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17826 2013-06-07 Alan Modra <amodra@gmail.com>
17828 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17830 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17832 * config/arm/constraints.md (Df): New constraint.
17833 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17834 Correct length attribute for last two alternatives.
17836 2013-06-07 Alan Modra <amodra@gmail.com>
17838 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17839 override user -mfp-in-toc.
17840 (offsettable_ok_by_alignment): Consider just the current access
17841 rather than the whole object, unless BLKmode. Handle
17842 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17843 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17844 for -mcmodel=medium.
17845 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17846 override user -mfp-in-toc or -msum-in-toc. Default to
17847 -mno-fp-in-toc for -mcmodel=medium.
17849 2013-06-06 DJ Delorie <dj@redhat.com>
17851 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17852 TARGET_VALID_POINTER_MODE.
17854 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17855 Pat Haugen <pthaugen@us.ibm.com>
17856 Peter Bergner <bergner@vnet.ibm.com>
17858 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17859 Document new power8 builtins.
17861 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17862 condition code register, to allow 128-bit logical operations to be
17863 done in the VSX or GPR registers.
17864 (nor<mode>3): Use the canonical form for nor.
17865 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17866 vclz*, and vpopcnt* vector instructions.
17867 (nand<mode>3): Likewise.
17868 (orc<mode>3): Likewise.
17869 (clz<mode>2): LIkewise.
17870 (popcount<mode>2): Likewise.
17872 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17873 that only the GPRs are recognized.
17875 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17876 support for new power8 builtins.
17878 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17880 (xscvdpspn): Likewise.
17886 (vpopcnt): Likewise.
17887 (vpopcntb): Likewise.
17888 (vpopcnth): Likewise.
17889 (vpopcntw): Likewise.
17890 (vpopcntd): Likewise.
17892 (vmrgew): Likewise.
17893 (vmrgow): Likewise.
17895 (eqv_v16qi3): Likewise.
17896 (eqv_v8hi3): Likewise.
17897 (eqv_v4si3): Likewise.
17898 (eqv_v2di3): Likewise.
17899 (eqv_v4sf3): Likewise.
17900 (eqv_v2df3): Likewise.
17902 (nand_v16qi3): Likewise.
17903 (nand_v8hi3): Likewise.
17904 (nand_v4si3): Likewise.
17905 (nand_v2di3): Likewise.
17906 (nand_v4sf3): Likewise.
17907 (nand_v2df3): Likewise.
17909 (orc_v16qi3): Likewise.
17910 (orc_v8hi3): Likewise.
17911 (orc_v4si3): Likewise.
17912 (orc_v2di3): Likewise.
17913 (orc_v4sf3): Likewise.
17914 (orc_v2df3): Likewise.
17916 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17917 allow power8 quad mode in 64-bit.
17918 (rs6000_builtin_vectorized_function): Add support to vectorize
17919 ISA 2.07 count leading zeros, population count builtins.
17920 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17921 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17922 (builtin_function_type): Add vgbbd builtin function which takes an
17924 (altivec_expand_vec_perm_const): Add support for new power8 merge
17927 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17928 that does not include TImdoe for use with 32-bit.
17929 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17931 (UNSPEC_VSX_CVDPSPN): Likewise.
17932 (vsx_xscvdpspn): Likewise.
17933 (vsx_xscvspdpn): Likewise.
17934 (vsx_xscvdpspn_scalar): Likewise.
17935 (vsx_xscvspdpn_directmove): Likewise.
17936 (vsx_and<mode>3): Split logical operations into 32-bit and
17937 64-bit. Add support to do logical operations on TImode as well as
17938 VSX vector types. Allow logical operations to be done in either
17939 VSX registers or in general purpose registers in 64-bit mode. Add
17940 splitters if GPRs were used. For AND, add clobber of CCmode to
17941 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17943 (vsx_and<mode>3_32bit): Likewise.
17944 (vsx_and<mode>3_64bit): Likewise.
17945 (vsx_ior<mode>3): Likewise.
17946 (vsx_ior<mode>3_32bit): Likewise.
17947 (vsx_ior<mode>3_64bit): Likewise.
17948 (vsx_xor<mode>3): Likewise.
17949 (vsx_xor<mode>3_32bit): Likewise.
17950 (vsx_xor<mode>3_64bit): Likewise.
17951 (vsx_one_cmpl<mode>2): Likewise.
17952 (vsx_one_cmpl<mode>2_32bit): Likewise.
17953 (vsx_one_cmpl<mode>2_64bit): Likewise.
17954 (vsx_nor<mode>3): Likewise.
17955 (vsx_nor<mode>3_32bit): Likewise.
17956 (vsx_nor<mode>3_64bit): Likewise.
17957 (vsx_andc<mode>3): Likewise.
17958 (vsx_andc<mode>3_32bit): Likewise.
17959 (vsx_andc<mode>3_64bit): Likewise.
17960 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17961 and xxlorc instructions.
17962 (vsx_eqv<mode>3_64bit): Likewise.
17963 (vsx_nand<mode>3_32bit): Likewise.
17964 (vsx_nand<mode>3_64bit): Likewise.
17965 (vsx_orc<mode>3_32bit): Likewise.
17966 (vsx_orc<mode>3_64bit): Likewise.
17968 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17970 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17972 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17973 (p8_vmrgow): Likewise.
17974 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17975 GPRs to be split under VSX.
17976 (p8v_clz<mode>2): Add power8 count leading zero support.
17977 (p8v_popcount<mode>2): Add power8 population count support.
17978 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17981 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17984 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17986 (vec_nand): Likewise.
17987 (vec_vclz): Likewise.
17988 (vec_vclzb): Likewise.
17989 (vec_vclzd): Likewise.
17990 (vec_vclzh): Likewise.
17991 (vec_vclzw): Likewise.
17992 (vec_vgbbd): Likewise.
17993 (vec_vmrgew): Likewise.
17994 (vec_vmrgow): Likewise.
17995 (vec_vpopcnt): Likewise.
17996 (vec_vpopcntb): Likewise.
17997 (vec_vpopcntd): Likewise.
17998 (vec_vpopcnth): Likewise.
17999 (vec_vpopcntw): Likewise.
18001 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18003 PR rtl-optimization/57468
18004 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18007 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18009 PR rtl-optimization/57459
18010 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18011 type when setting live regs.
18013 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18015 * config/s390/s390.opt (mlra): New option.
18016 * config/s390/s390.c (s390_decompose_address): Check displacement
18017 for all registers for LRA.
18018 (s390_secondary_reload): Don't used secondary reloads for LRA.
18019 (s390_lra_p): New function.
18020 (TARGET_LRA_P): Define.
18021 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18022 of attribute cpu_facility to zarch for the last alternative.
18023 (*cmpmem_short): Ditto.
18025 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18027 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18028 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18029 (arm_expand_prologue): Likewise.
18031 2013-06-06 Teresa Johnson <tejohnson@google.com>
18034 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18035 as this is now done by redirect_edge_and_branch_force.
18036 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18037 barriers, and fix interaction with splitting.
18038 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18039 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18040 reflect changes made in the routine.
18041 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18042 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18043 since this is called in cfglayout mode, and replace partition fixup
18044 with assert as that is now done by force_nonfallthru_and_redirect.
18045 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18046 already be marked with region crossing note.
18047 (insert_section_boundary_note): Make non-static, gate on flag
18048 has_bb_partition, rewrite to also check for multiple partitions.
18049 (rest_of_handle_reorder_blocks): Remove call to
18050 insert_section_boundary_note, now done later during free_cfg.
18051 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18052 * bb-reorder.h (insert_section_boundary_note): Declare.
18053 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18054 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18055 invoke insert_section_boundary_note.
18056 (try_redirect_by_replacing_jump): Remove unnecessary
18057 check for region crossing note.
18058 (fixup_partition_crossing): New function.
18059 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18060 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18061 in non-cfglayout mode.
18062 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18063 remove old code that tried to do this. Emit barrier correctly
18064 when we are in cfglayout mode.
18065 (last_bb_in_partition): New function.
18066 (rtl_split_edge): Correctly fixup partition boundaries.
18067 (commit_one_edge_insertion): Remove old code that tried to
18068 fixup region crossing edge since this is now handled in
18069 split_block, and set up insertion point correctly since
18070 block may now end in a jump.
18071 (verify_hot_cold_block_grouping): Guard against checking when not in
18072 linearized RTL mode.
18073 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18075 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18076 rtl_verify_flow_info, so not called in cfglayout mode.
18077 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18078 (fixup_reorder_chain): Remove old code that attempted to fixup region
18079 crossing note as this is now handled in force_nonfallthru_and_redirect.
18080 (duplicate_insn_chain): Don't duplicate switch section notes.
18081 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18083 * basic-block.h (emit_barrier_after_bb): Declare.
18085 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18087 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18088 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18089 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18091 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18093 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18095 * config/arm/ldmstm.md: Regenerate.
18097 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18099 * config/arm/sync.md (atomic_loaddi_1):
18100 Disable predication for arm_restrict_it.
18101 (arm_load_exclusive<mode>): Likewise.
18102 (arm_load_exclusivesi): Likewise.
18103 (arm_load_exclusivedi): Likewise.
18104 (arm_load_acquire_exclusive<mode>): Likewise.
18105 (arm_load_acquire_exclusivesi): Likewise.
18106 (arm_load_acquire_exclusivedi): Likewise.
18107 (arm_store_exclusive<mode>): Likewise.
18108 (arm_store_exclusive<mode>): Likewise.
18109 (arm_store_release_exclusivedi): Likewise.
18110 (arm_store_release_exclusive<mode>): Likewise.
18112 2013-06-06 Richard Biener <rguenther@suse.de>
18114 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18116 (lto_tag_is_tree_code_p): Adjust.
18117 (lto_tag_is_gimple_code_p): Likewise.
18118 (lto_gimple_code_to_tag): Likewise.
18119 (lto_tag_to_gimple_code): Likewise.
18120 (lto_tree_code_to_tag): Likewise.
18121 (lto_tag_to_tree_code): Likewise.
18122 * data-streamer.h (streamer_write_hwi_in_range): Use
18123 uhwi streaming to stream the normalized range.
18124 (streamer_read_hwi_in_range): Likewise.
18126 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18128 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18129 (predicable_short_it): Likewise.
18130 (predicated): Likewise.
18131 (enabled): Handle above.
18132 (define_cond_exec): Set predicated attribute to yes.
18134 2013-06-05 Mike Stump <mikestump@comcast.net>
18136 * gdbinit.in (__FUNCTION__): Add.
18138 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18140 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18141 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18143 2013-06-05 Jan Hubicka <jh@suse.cz>
18145 * varasm.c (mark_decl_referenced): Revert the removal until targets
18148 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18150 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18151 instead of mark_decl_referenced.
18153 2013-06-05 Jan Hubicka <jh@suse.cz>
18155 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18156 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18157 and symtab_used_from_object_file_p.
18158 (cgraph_make_node_local_1): Clear forced_by_abi.
18159 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18160 * cgraph.h (symtab_node_base): Add forced_by_abi.
18161 (decide_is_variable_needed): Remove.
18162 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18163 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18164 (decide_is_symbol_needed): ... this one; handle symbols in general;
18165 always analyze virtuals; honnor forced_by_abi.
18166 (cgraph_finalize_function): Update.
18167 (varpool_finalize_decl): Update.
18168 (symbol_defined_and_needed): Remove.
18169 (analyze_functions): Update.
18170 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18171 output_refs, input_overwrite_node): Handle forced_by_abi.
18172 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18173 (address_taken_from_non_vtable_p): ... this one.
18174 (comdat_can_be_unshared_p_1): New function.
18175 (cgraph_comdat_can_be_unshared_p): Rename to ...
18176 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18177 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18178 (function_and_variable_visibility): Clear forced_by_abi as needed.
18179 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18180 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18181 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18182 * varpool.c (decide_is_variable_needed): Remove.
18184 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18186 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18187 (arm_option_override): Override arm_restrict_it where appropriate.
18188 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18189 * config/arm/arm.opt (mrestrict-it): New command-line option.
18190 * doc/invoke.texi: Document -mrestrict-it.
18192 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18194 * tsan.c (tsan_atomic_table): Make const.
18196 2013-06-05 Richard Biener <rguenther@suse.de>
18198 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18199 index associated with the tree we are supposed to replace.
18200 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18201 TREE_ASM_WRITTEN as zero for everything but SSA names.
18203 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18205 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18207 2013-06-04 Jan Hubicka <jh@suse.cz>
18209 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18210 (input_node, input_varpool_node): Handle correctly external same
18212 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18213 nodes at ltrans stage.
18215 2013-06-04 Jan Hubicka <jh@suse.cz>
18217 * ipa-inline.c (update_caller_keys): Fix availability test.
18218 (update_callee_keys): Likewise.
18219 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18220 to follow ELF standard.
18222 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18224 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18225 (mips64r5900el-*-elf*): New configurations.
18226 * config/mips/mips-cpus.def (r5900): New processor.
18227 * config/mips/mips-tables.opt: Regenerate.
18228 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18229 (mips_issue_rate): Handle PROCESSOR_R5900.
18230 (mips_reorg_process_insns): Force reorder mode for the R5900.
18231 * config/mips/mips.h (TARGET_MIPS5900): Define.
18232 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18234 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18236 * config/mips/mips.md (processor): Add r5900.
18237 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18239 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18241 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18242 into function to generate MOVI instruction.
18243 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18244 (aarch64_preferred_simd_mode): Turn into wrapper.
18245 (aarch64_output_scalar_simd_mov_immediate): New function.
18246 * config/aarch64/aarch64-protos.h: Add prototype for above.
18248 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18250 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18251 element_char member.
18252 (sizetochar): Return signed char.
18253 (aarch64_simd_valid_immediate): Remove elchar and other
18254 unnecessary variables.
18255 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18256 Calculate element_char as required.
18257 * config/aarch64/aarch64-protos.h: Update and move prototype
18258 for aarch64_output_simd_mov_immediate.
18259 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18262 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18264 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18265 information completed by aarch64_simd_valid_immediate.
18266 (aarch64_legitimate_constant_p): Update arguments.
18267 (aarch64_simd_valid_immediate): Work with struct rather than many
18269 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18270 (aarch64_simd_make_constant): Update arguments.
18271 (aarch64_output_simd_mov_immediate): Work with struct rather than
18272 many pointers. Output immediate directly rather than as operand.
18273 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18275 * config/aarch64/constraints.md (Dn): Update arguments.
18277 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18279 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18281 (aarch64_simd_immediate_valid_for_move): Remove.
18282 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18283 (aarch64_simd_make_constant): Update call.
18284 (aarch64_output_simd_mov_immediate): Update call.
18285 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18287 * config/aarch64/constraints.md (Dn): Update call.
18289 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18291 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
18292 return type to bool for prototype.
18293 (aarch64_legitimate_constant_p): Check for true instead of not -1.
18294 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
18295 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
18297 2013-06-04 Catherine Moore <clm@codesourcery.com>
18299 * config/mips/mips.opt (meva): New.
18300 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
18301 (ASM_SPEC): Handle -meva.
18302 * doc/invoke.texi (meva): Document.
18304 2013-06-04 Alan Modra <amodra@gmail.com>
18306 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
18309 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18311 * rtl.def: Add extra fourth optional field to define_cond_exec.
18312 * gensupport.c (process_one_cond_exec): Process attributes from
18314 * doc/md.texi: Document fourth field in define_cond_exec.
18316 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
18318 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
18319 out the processing order as in store_bit_field_1.
18321 2013-06-04 Jan Hubicka <jh@suse.cz>
18323 PR middle-end/57500
18324 * cgraphunit.c (cgraph_process_same_body_aliases): Create
18325 non-VAR_DECL node if it does not exist yet.
18327 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
18329 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
18330 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
18331 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
18332 target_cpu_default setting.
18334 2013-06-03 Teresa Johnson <tejohnson@google.com>
18336 * dumpfile.c (opt_info_switch_p): Change -fopt-info
18337 default to -fopt-info=optimized instead of all.
18338 * doc/invoke.texi: Ditto.
18339 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
18340 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
18341 (execute_vect_slp): Emit BB vectorization success under
18342 MSG_OPTIMIZED_LOCATIONS.
18343 * tree-vect-slp.c (vect_slp_transform_bb): Change
18344 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
18345 * tree-vect-loop.c (vect_transform_loop): Ditto.
18347 2013-06-03 Jason Merrill <jason@redhat.com>
18350 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
18351 Use TARGET_EXPR for C++.
18353 2013-06-03 Jakub Jelinek <jakub@redhat.com>
18355 PR rtl-optimization/57268
18356 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
18357 if DEBUG_INSN_P (insn).
18360 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18362 PR rtl-optimization/57268
18363 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18364 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18366 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18368 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18369 (ix86_avoid_lea_for_addr): Likewise.
18370 (exact_dependency_1): Likewise.
18371 (ix86_adjust_cost): Likewise.
18372 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18374 (do_reorder_for_imul): Fix typo, formatting and
18375 !reload_completed check removed.
18376 (ix86_sched_reorder): Fix typo and formatting.
18377 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18380 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18382 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18384 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18386 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18387 <STRING_CST>: Likewise.
18388 <VECTOR_CST>: Likewise.
18390 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18391 Mikael Morin <mikael@gcc.gnu.org>
18393 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18394 * config.in: Regenerated.
18395 * configure: Regenerated.
18397 2013-06-01 Jan Hubicka <jh@suse.cz>
18399 PR middle-end/57366
18400 * cgraphunit.c (compile): When weakref is not supported,
18401 set up transparent aliases before final output pass.
18402 * varasm.c (assemble_alias): Do not try to do it here.
18404 2013-06-01 Jan Hubicka <jh@suse.cz>
18406 PR middle-end/57467
18407 * passes.c (for_per_function): Skip unanalyzed functions.
18409 2013-06-01 Jan Hubicka <jh@suse.cz>
18411 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18412 (lto_symtab_merge_symbols_1): ... this one.
18413 (lto_symtab_merge_cgraph_nodes): Rename to ...
18414 (lto_symtab_merge_symbols): ... this one; simplify.
18415 * cgraph.c (same_body_aliases_done): Rename to ...
18416 (cpp_implicit_aliases_done): ... this one.
18417 (cgraph_create_function_alias): Update.
18418 (cgraph_same_body_alias): Update.
18419 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18420 (verify_edge_corresponds_to_fndecl): Simplify.
18421 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18422 (cgraph_node): Remove same_body_alias.
18423 (varpool_node): Remove alias_of and extra_name_alias.
18424 (same_body_aliases_done): Rename to ..
18425 (cpp_implicit_aliases_done): ... this one.
18426 (symtab_alias_ultimate_target): Add default parameter.
18427 (symtab_resolve_alias): New function.
18428 (fixup_same_cpp_alias_visibility): Declare.
18429 (cgraph_function_node): Add default parameter.
18430 (cgraph_node_asm_name): Likewise.
18431 (cgraph_function_or_thunk_node): Add default parameter; do
18432 not ICE when it is NULL.
18433 (varpool_variable_node): Likewise.
18434 * tree-emutls.c (create_emultls_var): Update.
18435 (ipa_lower_emutls): Update.
18436 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18437 (cgraph_reset_node): Reset alias info.
18438 (cgraph_finalize_function): Update.
18439 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18440 (analyze_function): Simplify.
18441 (cgraph_process_same_body_aliases): Simplify.
18442 (analyze_functions): Fixup same body aliases.
18443 (handle_alias_pairs): Simplify.
18444 (assemble_thunk): Update.
18445 (assemble_thunks_and_aliases): Update.
18446 (output_weakrefs): Rewrite.
18447 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18448 (lto_output_varpool_node): Likewise.
18449 (compute_ltrans_boundary): Remve assert.
18450 (get_alias_symbol): New functoin.
18451 (input_node): Rewrite alias handling.
18452 (input_varpool_node): Likewise.
18453 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18454 * ipa.c (process_references): Handle weakrefs correctly.
18455 (symtab_remove_unreachable_nodes): Likewise.
18456 * trans-mem.c (get_cg_data): Update.
18457 (ipa_tm_create_version_alias): Update.
18458 (ipa_tm_execute): Update.
18459 * symtab.c (dump_symtab_base): Dump aliases.
18460 (verify_symtab_base): Verify aliases.
18461 (symtab_node_availability): New function.
18462 (symtab_alias_ultimate_target): Simplify.
18463 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18464 handle all the fixup cases.
18465 (symtab_resolve_alias): New function.
18466 * passes.c (ipa_write_summaries): Handle weakrefs.
18467 * varpool.c (varpool_analyze_node): Simplify.
18468 (assemble_aliases): Update.
18469 (varpool_create_variable_alias): Simplify.
18470 (varpool_extra_name_alias): Simplify.
18471 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18472 (lto_symtab_merge_symbols): ... this one.
18474 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18477 PR rtl-optimization/57268
18478 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18479 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18481 2013-06-01 Tobias Burnus <burnus@net-b.de>
18483 Partially reverted:
18484 2013-05-31 Tobias Burnus <burnus@net-b.de>
18486 PR middle-end/57073
18487 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18490 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18492 PR rtl-optimization/57268
18493 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18494 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18496 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18498 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18499 unordered comparison operators when -fno-trapping-math is in effect
18501 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18502 and implement unordered comparison operators properly on the e500.
18504 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18506 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18507 for constant scalar integers.
18508 (simplify_relational_operation_1): Likewise.
18510 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18512 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18513 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18516 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18517 Igor Zamyatin <igor.zamyatin@intel.com>
18519 Silvermont (SLM) architecture performance tuning.
18520 * config/i386/i386.h (enum ix86_tune_indices): Add
18521 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18522 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18524 * config/i386/i386.c (initial_ix86_tune_features)
18525 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18526 (ix86_lea_outperforms): Handle Silvermont tuning.
18527 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18529 (ix86_use_lea_for_mov): Likewise.
18530 (ix86_avoid_lea_for_addr): Likewise.
18531 (ix86_lea_for_add_ok): Likewise.
18532 (exact_dependency_1): New function.
18533 (exact_store_load_dependency): Likewise.
18534 (ix86_adjust_cost): Handle Silvermont tuning.
18535 (do_reoder_for_imul): Likewise.
18536 (swap_top_of_ready_list): New function.
18537 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18539 * config/i386/i386.md (peepholes that split memory operand in fp
18542 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18544 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18545 Remove un-necessary braces.
18547 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18549 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18550 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18552 2013-05-31 Tobias Burnus <burnus@net-b.de>
18554 PR middle-end/57073
18555 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18557 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18560 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18561 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18562 * config/arm/neon.md (iordi3_neon): Remove.
18563 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18564 * config/arm/predicates.md (imm_for_neon_logic_operand):
18565 Move to earlier in the file.
18566 (neon_logic_op2): Likewise.
18567 (arm_iordi_operand_neon): New predicate.
18569 2013-05-31 Richard Biener <rguenther@suse.de>
18571 PR tree-optimization/57478
18572 PR tree-optimization/57453
18573 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18576 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18578 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18579 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18581 2013-05-30 Tobias Burnus <burnus@net-b.de>
18582 Thomas Koenig <tkoenig@gcc.gnu.org>
18584 PR middle-end/57073
18585 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18586 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18588 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18590 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18592 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18594 * target.def (register_usage_leveling_p): New hook.
18595 * targhooks.c (default_register_usage_leveling_p): New.
18596 * targhooks.h (default_register_usage_leveling_p): New prototype.
18597 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18598 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18599 * doc/tm.texi: Update.
18600 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18602 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18604 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18605 (*insv_reg<mode>): New define_insn.
18607 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18609 PR rtl-optimization/57439
18610 * postreload.c (move2add_valid_value_p): Check that we have
18611 a zero subreg_regno_offset when accessing the register in
18612 the requested mode.
18614 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18615 Igor Zamyatin <igor.zamyatin@intel.com>
18617 Silvermont (SLM) architecture pipeline model, tuning and
18619 * config.gcc: Add slm config options and target.
18621 * config/i386/slm.md: New.
18623 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18625 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18627 (ix86_target_macros_internal): Likewise.
18629 * config/i386/i386.c (slm_cost): New cost.
18630 (m_SLM): New macro flag.
18631 (initial_ix86_tune_features): Set m_SLM.
18632 (x86_accumulate_outgoing_args): Likewise.
18633 (x86_arch_always_fancy_math_387): Likewise.
18634 (processor_target_table): Add slm cost.
18635 (cpu_names): Add slm cpu name.
18636 (x86_option_override_internal): Set SLM ISA.
18637 (ix86_issue_rate): New case PROCESSOR_SLM.
18638 (ia32_multipass_dfa_lookahead): Likewise.
18639 (fold_builtin_cpu): Add slm.
18641 * config/i386/i386.h (TARGET_SLM): New target macro.
18642 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18643 (processor_type): Add PROCESSOR_SLM.
18645 * config/i386/i386.md (cpu): Add new value "slm".
18646 (slm.md): Include slm.md.
18648 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18649 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18651 * config/arm/arm-protos.h: Add and update function protos.
18652 * config/arm/arm.c (use_simple_return_p): New added.
18653 (thumb2_expand_return): Check simple_return flag.
18654 * config/arm/arm.md: Add simple_return and conditional simple_return.
18655 * config/arm/iterators.md: Add iterator for return and simple_return.
18657 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18659 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18660 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18661 (arm_emit_vfp_multi_reg_pop): Likewise.
18662 (thumb2_emit_ldrd_pop): Likewise.
18663 (arm_expand_epilogue): Add misc REG_CFA notes.
18664 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18666 2013-05-29 Lawrence Crowl <crowl@google.com>
18668 * config/arm/t-arm: Update for below.
18670 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18671 Change type to hash_table. Update dependent calls and types.
18673 * config/i386/t-cygming: Update for below.
18675 * config/i386/t-interix: Update for below.
18677 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18678 Change type to hash_table. Update dependent calls and types.
18679 (i386_find_on_wrapper_list::wrappers): Likewise.
18681 * config/ia64/t-ia64: Update for below.
18683 * config/ia64/ia64.c (bundle_state_table):
18684 Change type to hash_table. Update dependent calls and types.
18686 * config/mips/mips.c (mips_reorg_process_insns::htab):
18687 Change type to hash_table. Update dependent calls and types.
18689 * config/sol2.c (solaris_comdat_htab):
18690 Change type to hash_table. Update dependent calls and types.
18692 * config/t-sol2: Update for above.
18694 2013-05-29 Teresa Johnson <tejohnson@google.com>
18696 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18697 functions are not yet marked as defined.
18699 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18700 Pat Haugen <pthaugen@us.ibm.com>
18701 Peter Bergner <bergner@vnet.ibm.com>
18703 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18707 (vrotl<mode>3): Likewise.
18708 (vashl<mode>3): Likewise.
18709 (vlshr<mode>3): Likewise.
18710 (vashr<mode>3): Likewise.
18712 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18713 support for power8 V2DI builtins.
18715 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18716 power8 V2DI builtins.
18717 (vupkhsw): Likewise.
18718 (vupklsw): Likewise.
18719 (vaddudm): Likewise.
18720 (vminsd): Likewise.
18721 (vmaxsd): Likewise.
18722 (vminud): Likewise.
18723 (vmaxud): Likewise.
18724 (vpkudum): Likewise.
18725 (vpksdss): Likewise.
18726 (vpkudus): Likewise.
18727 (vpksdus): Likewise.
18732 (vsubudm): Likewise.
18733 (vcmpequd): Likewise.
18734 (vcmpgtsd): Likewise.
18735 (vcmpgtud): Likewise.
18736 (vcmpequd_p): Likewise.
18737 (vcmpgtsd_p): Likewise.
18738 (vcmpgtud_p): Likewise.
18739 (vupkhsw): Likewise.
18740 (vupklsw): Likewise.
18741 (vaddudm): Likewise.
18742 (vmaxsd): Likewise.
18743 (vmaxud): Likewise.
18744 (vminsd): Likewise.
18745 (vminud): Likewise.
18746 (vpksdss): Likewise.
18747 (vpksdus): Likewise.
18748 (vpkudum): Likewise.
18749 (vpkudus): Likewise.
18754 (vsubudm): Likewise.
18756 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18757 support for power8 V2DI instructions.
18759 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18760 power8 V2DI instructions. Combine pack and unpack insns to use an
18761 iterator for each mode. Check whether a particular mode supports
18762 Altivec instructions instead of just checking TARGET_ALTIVEC.
18763 (UNSPEC_VPKUWUM): Likewise.
18764 (UNSPEC_VPKSHSS): Likewise.
18765 (UNSPEC_VPKSWSS): Likewise.
18766 (UNSPEC_VPKUHUS): Likewise.
18767 (UNSPEC_VPKSHUS): Likewise.
18768 (UNSPEC_VPKUWUS): Likewise.
18769 (UNSPEC_VPKSWUS): Likewise.
18770 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18771 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18772 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18773 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18774 (UNSPEC_VUPKHSB): Likewise.
18775 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18776 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18777 (UNSPEC_VUPKHSH): Likewise.
18778 (UNSPEC_VUPKLSB): Likewise.
18779 (UNSPEC_VUPKLSH): Likewise.
18781 (VI_char): Likewise.
18782 (VI_scalar): Likewise.
18783 (VI_unit): Likewise.
18785 (VP_small): Likewise.
18786 (VP_small_lc): Likewise.
18787 (VU_char): Likewise.
18788 (add<mode>3): Likewise.
18789 (altivec_vaddcuw): Likewise.
18790 (altivec_vaddu<VI_char>s): Likewise.
18791 (altivec_vadds<VI_char>s): Likewise.
18792 (sub<mode>3): Likewise.
18793 (altivec_vsubcuw): Likewise.
18794 (altivec_vsubu<VI_char>s): Likewise.
18795 (altivec_vsubs<VI_char>s): Likewise.
18796 (altivec_vavgs<VI_char>): Likewise.
18797 (altivec_vcmpbfp): Likewise.
18798 (altivec_eq<mode>): Likewise.
18799 (altivec_gt<mode>): Likewise.
18800 (altivec_gtu<mode>): Likewise.
18801 (umax<mode>3): Likewise.
18802 (smax<mode>3): Likewise.
18803 (umin<mode>3): Likewise.
18804 (smin<mode>3): Likewise.
18805 (altivec_vpkuhum): Likewise.
18806 (altivec_vpkuwum): Likewise.
18807 (altivec_vpkshss): Likewise.
18808 (altivec_vpkswss): Likewise.
18809 (altivec_vpkuhus): Likewise.
18810 (altivec_vpkshus): Likewise.
18811 (altivec_vpkuwus): Likewise.
18812 (altivec_vpkswus): Likewise.
18813 (altivec_vpks<VI_char>ss): Likewise.
18814 (altivec_vpks<VI_char>us): Likewise.
18815 (altivec_vpku<VI_char>us): Likewise.
18816 (altivec_vpku<VI_char>um): Likewise.
18817 (altivec_vrl<VI_char>): Likewise.
18818 (altivec_vsl<VI_char>): Likewise.
18819 (altivec_vsr<VI_char>): Likewise.
18820 (altivec_vsra<VI_char>): Likewise.
18821 (altivec_vsldoi_<mode>): Likewise.
18822 (altivec_vupkhsb): Likewise.
18823 (altivec_vupkhs<VU_char>): Likewise.
18824 (altivec_vupkls<VU_char>): Likewise.
18825 (altivec_vupkhsh): Likewise.
18826 (altivec_vupklsb): Likewise.
18827 (altivec_vupklsh): Likewise.
18828 (altivec_vcmpequ<VI_char>_p): Likewise.
18829 (altivec_vcmpgts<VI_char>_p): Likewise.
18830 (altivec_vcmpgtu<VI_char>_p): Likewise.
18831 (abs<mode>2): Likewise.
18832 (vec_unpacks_hi_v16qi): Likewise.
18833 (vec_unpacks_hi_v8hi): Likewise.
18834 (vec_unpacks_lo_v16qi): Likewise.
18835 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18836 (vec_unpacks_lo_v8hi): Likewise.
18837 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18838 (vec_pack_trunc_v8h): Likewise.
18839 (vec_pack_trunc_v4si): Likewise.
18840 (vec_pack_trunc_<mode>): Likewise.
18842 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18844 (vec_vmaxsd): Likewise.
18845 (vec_vmaxud): Likewise.
18846 (vec_vminsd): Likewise.
18847 (vec_vminud): Likewise.
18848 (vec_vpksdss): Likewise.
18849 (vec_vpksdus): Likewise.
18850 (vec_vpkudum): Likewise.
18851 (vec_vpkudus): Likewise.
18852 (vec_vrld): Likewise.
18853 (vec_vsld): Likewise.
18854 (vec_vsrad): Likewise.
18855 (vec_vsrd): Likewise.
18856 (vec_vsubudm): Likewise.
18857 (vec_vupkhsw): Likewise.
18858 (vec_vupklsw): Likewise.
18860 2013-05-29 Jan Hubicka <jh@suse.cz>
18862 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18863 flags; reorder rest of fields in more consistent way.
18864 (varpool_node): Remove analyzed, finalized and alias.
18865 (cgraph_ndoe): Likewise.
18866 (symtab_alias_ultimate_target): New function.
18867 (cgraph_function_node): Move offline.
18868 (cgraph_reset_node): Declare.
18869 (cgraph_comdat_can_be_unshared_p): Remove.
18870 (varpool_remove_initializer): Declare.
18871 (varpool_first_defined_variable, varpool_next_defined_variable
18872 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18873 (cgraph_function_with_gimple_body_p): Update.
18874 (varpool_all_refs_explicit_p): Update.
18875 (symtab_alias_target): New function.
18876 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18877 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18878 (cgraph_function_or_thunk_node): Simplify using
18879 symtab_alias_ultimate_target.
18880 (varpool_variable_node): Likewise.
18881 * cgraph.c (cgraph_create_function_alias): Update.
18882 (cgraph_add_thunk): Update.
18883 (cgraph_remove_node): Update.
18884 (dump_cgraph_node): Do not dump removed flags.
18885 (cgraph_function_body_availability): Update.
18886 (cgraph_propagate_frequency): Update.
18887 (verify_cgraph_node): Check sanity of local flag.
18888 (cgraph_function_node): Move here from cgraph.h; revamp for
18889 cgraph_function_or_thunk_node.
18890 * lto-symtab.c (lto_varpool_replace_node): Update.
18891 (lto_symtab_resolve_can_prevail_p): Update.
18892 (lto_symtab_merge_cgraph_nodes): Update.
18893 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18894 propagate_constants_accross_call, devirtualization_time_bonus,
18895 ipcp_propagate_stage): Update.
18896 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18897 * ipa-inline-transform.c (clone_inlined_nodes,
18898 preserve_function_body_p): Update.
18899 * ipa-reference.c (propagate): Update.
18900 (write_node_summary_p): Update.
18901 * toplev.c (wrapup_global_declaration_2): Update.
18902 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18903 (analyze_function) ... this one.
18904 (cgraph_process_new_functions): Update.
18905 (cgraph_reset_node): Export.
18906 (cgraph_finalize_function): Update.
18907 (cgraph_add_new_function): Update.
18908 (process_function_and_variable_attributes): Update.
18909 (varpool_finalize_decl): Update.
18910 (symbol_finalized): Remove.
18911 (symbol_finalized_and_needed): Rename to ...
18912 (symbol_defined_and_needed): ... update.
18913 (cgraph_analyze_functions): Update.
18914 (handle_alias_pairs): Update.
18915 (mark_functions_to_output): Update.
18916 (assemble_thunk): Update.
18917 (output_in_order): Update.
18918 (output_weakrefs): Update.
18919 (finalize_compilation_unit): Update.
18920 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18921 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18922 input_node, input_varpool_node): Update.
18923 * dbxout.c (dbxout_expand_expr): Update.
18924 * cgraphclones.c (cgraph_clone_node): Update.
18925 (cgraph_copy_node_for_versioning): Update.
18926 (cgraph_materialize_clone): Update.
18927 (cgraph_materialize_all_clones): Update.
18928 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18929 propagate_pure_const, propagate_nothrow): Update.
18930 * lto-streamer-out.c (lto_output, write_symbol): Update.
18931 * ipa-utils.c (ipa_reverse_postorder): Update.
18932 * ipa-inline.c (can_inline_edge_p): Update.
18933 (update_caller_keys, ipa_inline): Update.
18934 * dwarf2out.c (reference_to_unused,
18935 premark_types_used_by_global_vars_helper): Update.
18936 * tree-eh.c (tree_could_trap_p): Update.
18937 * ipa-split.c (consider_split, execute_split_functions): Update.
18938 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18939 has_addr_references_p): Update; move ahead in file for better
18941 (process_references): Simplify.
18942 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18943 bodies are removed.
18944 (cgraph_comdat_can_be_unshared_p): Make static.
18945 (cgraph_externally_visible_p): Update.
18946 (varpool_externally_visible_p): Update.
18947 (function_and_variable_visibility): Update.
18948 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18949 ipa_tm_mark_force_output_node): Update.
18950 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18951 estimate_edge_devirt_benefit, inline_generate_summary,
18952 inline_write_summary): Update.
18953 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18954 * ipa-prop.c (ipa_compute_jump_functions): Update.
18955 (ipa_print_node_params, ipa_prop_read_section,
18956 ipa_update_after_lto_read, read_replacements_section): Update.
18957 * varasm.c (mark_decl_referenced): Update.
18958 (assemble_alias, dump_tm_clone_pairs): Update.
18959 * tree-inline.c (copy_bb): Update.
18960 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18962 * symtab.c (dump_symtab_base): Print new flags.
18963 (verify_symtab_base): Verify new flags.
18964 (symtab_alias_ultimate_target): New function.
18965 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18966 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18968 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18970 * i386.c (ix86_get_function_versions_dispatcher,
18971 ix86_generate_version_dispatcher_body): Update.
18972 (fold_builtin_cpu): Use varpool_add_new_variable.
18973 * varpool.c (varpool_remove_initializer): Break out from ...
18974 (varpool_remove_node): ... this one.
18975 (dump_varpool_node, varpool_node_for_asm,
18976 cgraph_variable_initializer_availability, varpool_analyze_node,
18977 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18978 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18981 2013-05-29 Jan Hubicka <jh@suse.cz>
18983 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18985 2013-05-29 Easwaran Raman <eraman@google.com>
18987 PR tree-optimization/57442
18988 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18989 when control exits the main loop.
18991 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18993 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18995 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18996 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18997 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19000 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19002 PR tree-optimization/57441
19003 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19004 Don't limit size of incr_vec to number of candidates.
19006 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19008 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19009 and mips16 directories.
19010 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19011 (MULTILIB_DIRNAMES): Ditto.
19012 (MULTILIB_EXCEPTIONS): Add new exceptions.
19013 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19014 (MULTILIB_DIRNAMES): Ditto.
19015 (MULTILIB_EXCEPTIONS): Add new exceptions.
19017 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19018 Marcus Shawcroft <marcus.shawcroft@arm.com>
19020 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19021 SYMBOL_TINY_ABSOLUTE.
19022 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19023 SYMBOL_TINY_ABSOLUTE.
19024 (aarch64_expand_mov_immediate): Likewise.
19025 (aarch64_classify_symbol): Likewise.
19026 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19027 Permit SYMBOL_TINY_ABSOLUTE.
19028 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19030 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19031 Marcus Shawcroft <marcus.shawcroft@arm.com>
19033 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19034 Refactor if/switch. Replace gcc_assert with if.
19036 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19038 * config/i386/i386.c (initial_ix86_tune_features): Enable
19039 FP Reassociation for AMD bdver1 and bdver2.
19041 2013-05-29 Martin Jambor <mjambor@suse.cz>
19043 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19044 and IMAGPART_EXPR do not occur within other handled_components.
19046 2013-05-29 Richard Biener <rguenther@suse.de>
19048 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19049 access on whether the use is in the BB we currently try to
19051 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19052 vectorize to vect_bb_slp_scalar_cost.
19054 2013-05-29 Richard Biener <rguenther@suse.de>
19056 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19057 computing scalar cost offsetted by stmts that are kept live
19059 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19060 for computation of scalar cost.
19062 2013-05-28 Steve Ellcey <sellcey@mips.com>
19064 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19066 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19068 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19069 array notation built-in reduction functions.
19070 * doc/passes.texi (Passes): Added documentation about changes done
19072 * doc/invoke.texi (C Dialect Options): Added documentation about
19073 the -fcilkplus flag.
19074 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19075 (BUILTINS_DEF): Depend on cilkplus.def.
19076 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19077 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19078 * cilkplus.def: New file.
19080 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19082 PR rtl-optimization/57439
19083 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19085 2013-05-28 Easwaran Raman <eraman@google.com>
19087 PR tree-optimization/57337
19088 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19089 (find_insert_point): Correctly identify the insertion point
19090 when two statements with the same UID is compared.
19092 2013-05-28 Richard Biener <rguenther@suse.de>
19094 PR tree-optimization/56787
19095 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19096 from the list of data references.
19097 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19099 (vect_analyze_loop_operations): Likewise.
19100 (vect_transform_loop): Remove clobbers.
19102 2013-05-28 Martin Jambor <mjambor@suse.cz>
19104 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19105 and REALPART_EXPRs have scalar type.
19107 2013-05-28 Richard Biener <rguenther@suse.de>
19109 PR tree-optimization/57411
19110 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19112 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19113 virtual operand propagation.
19115 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19117 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19118 destination register for bmasksi_vis.
19119 (vector_init_bshuffle): Likewise.
19120 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19122 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19124 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19125 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19126 mode if the instruction isn't available in the original mode.
19127 * config/sparc/sparc.opt (mfix-ut699): New option.
19128 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19129 (divdf3): Turn into expander.
19130 (divdf3_nofix): New insn.
19131 (divdf3_fix): Likewise.
19132 (divsf3): Disable if -mfix-ut699.
19133 (sqrtdf2): Turn into expander.
19134 (sqrtdf2_nofix): New insn.
19135 (sqrtdf2_fix): Likewise.
19136 (sqrtsf2): Disable if -mfix-ut699.
19138 2013-05-27 Richard Biener <rguenther@suse.de>
19140 PR middle-end/57412
19141 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19142 block for the new loop.
19144 2013-05-27 Richard Biener <rguenther@suse.de>
19146 PR tree-optimization/57343
19147 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19148 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19149 (number_of_iterations_cond): Do not build the folded tree.
19151 2013-05-27 Richard Biener <rguenther@suse.de>
19154 PR middle-end/57381
19155 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19156 OEP_CONSTANT_ADDRESS_OF retained.
19158 PR tree-optimization/57417
19159 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19160 for unchanged base.
19161 (set_ssa_val_to): Compare addresses using
19162 get_addr_base_and_unit_offset.
19164 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19166 PR rtl-optimization/56833
19167 * postreload.c (move2add_record_mode): New function.
19168 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19169 (move2add_use_add2_insn): Use move2add_record_sym_value.
19170 (move2add_use_add3_insn): Likewise.
19171 (reload_cse_move2add): Use move2add_valid_value_p and
19172 move2add_record_mode. Invalidate call-clobbered and REG_INC
19173 affected regs by setting reg_mode to VOIDmode.
19174 (move2add_note_store): Don't pretend the inside of a SUBREG is
19175 the actual destination. Invalidate single/leading registers by
19176 setting reg_mode to VOIDmode.
19177 Use move2add_record_sym_value, move2add_valid_value_p and
19178 move2add_record_mode.
19180 2013-05-27 Richard Biener <rguenther@suse.de>
19182 PR tree-optimization/57396
19183 * tree-affine.c (double_int_constant_multiple_p): Properly
19184 return false for val == 0 and div != 0.
19186 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19188 * config/mips/mips.h: Use #elif in preprocessor conditions.
19190 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19193 * config/mips/constraints.md (kl): New constraint.
19194 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19195 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19196 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19197 from LO for MIPS16.
19198 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19200 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19203 * config/mips/mips.c (mips_can_inline_p): New function.
19204 (TARGET_CAN_INLINE_P): Define.
19206 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19208 * sched-int.h (ds_t, dw_t): Make unsigned int.
19209 Fix documentation that describes how all the ds_t bits are used.
19210 Reserve the last bit for delayed-branch scheduling.
19211 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19212 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19213 (gen_dep_weak_1): Remove prototype.
19214 * sched-deps.c (get_dep_weak_1): Make static.
19215 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19216 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19217 * doc/tm.texi: Regenerate.
19218 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19220 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19223 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19225 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19226 Sandra Loosemore <sandra@codesourcery.com>
19228 * config.gcc (powerpc-*): Allow native for with-cpu.
19230 2013-05-24 Jeff Law <law@redhat.com>
19232 PR tree-optimization/57124
19233 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19234 conversion feeding a condition if the range has an overflow
19235 if -fstrict-overflow. Add warnings for when we do make the
19238 2013-05-24 Dehao Chen <dehao@google.com>
19240 * tree-cfg.c (locus_discrim_map): Fix the typo.
19241 (locus_discrim_hasher): Likewise.
19242 (locus_discrim_hasher::hash): Likewise.
19243 (locus_discrim_hasher::equal): Likewise.
19245 2013-05-24 Martin Jambor <mjambor@suse.cz>
19247 PR tree-optimization/57294
19248 * cgraph.h (ipa_record_stmt_references): Declare.
19249 * cgraphbuild.c (ipa_record_stmt_references): New function.
19250 (build_cgraph_edges): Use ipa_record_stmt_references.
19251 (rebuild_cgraph_edges): Likewise.
19252 (cgraph_rebuild_references): Likewise.
19253 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19254 associated with the old statement and build references from the
19255 newly built statements.
19256 * ipa-ref.c (ipa_remove_stmt_references): New function.
19257 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19259 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19261 * lra-constraints.c (emit_spill_move): Use smaller mode for
19263 (check_and_process_move): Consider mem-reg moves for secondary
19265 (curr_insn_transform): Don't lose insns emitted before for
19266 secondary memory moves.
19267 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19268 reg set up in the current insn.
19270 2013-05-24 Dehao Chen <dehao@google.com>
19272 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19274 (locus_descrim_hasher::equal): Likewise.
19275 (build_gimple_cfg): New discriminator assignment algorithm.
19276 (make_edges): Likewise.
19277 (next_discriminator_for_locus): Likewise.
19278 (same_line_p): Likewise.
19279 (assign_discriminators): Likewise.
19280 (make_cond_expr_edges): Likewise.
19281 (make_gimple_switch_edges): Likewise.
19282 (make_goto_expr_edges): Likewise.
19283 (make_gimple_asm_edges): Likewise.
19285 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19287 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19288 X format specifier to only display bottom 16 bits.
19289 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
19290 immediate to match for operand 2, since it will be masked.
19292 2013-05-24 Richard Biener <rguenther@suse.de>
19294 PR tree-optimization/57287
19295 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
19296 all SSA names that occur in abnormal PHIs.
19298 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
19301 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
19302 that index is not negative.
19304 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
19306 PR rtl-optimization/55177
19307 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
19308 (simplify_byte_swapping_operation): New.
19309 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
19310 (simplify_relational_operation_1): Deal with BSWAP.
19312 2013-05-23 Richard Henderson <rth@redhat.com>
19315 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
19316 (ix86_reorg): Call it.
19318 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
19321 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
19322 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
19323 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
19325 2013-05-23 Christian Bruel <christian.bruel@st.com>
19328 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
19330 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19331 Marcus Shawcroft <marcus.shawcroft@arm.com>
19333 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
19334 * config/aarch64/constraints.md (Usa): Remove.
19335 * doc/md.texi (AArch64 Usa): Remove.
19337 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19338 Marcus Shawcroft <marcus.shawcroft@arm.com>
19340 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
19341 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
19342 * config/aarch64/predicates.md (aarch64_const_address): Remove.
19343 (aarch64_mov_operand): Use aarch64_mov_operand_p.
19345 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
19347 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
19348 instruction (AdvSIMD).
19349 * config/aarch64/aarch64-builtins.c
19350 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
19351 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
19353 2013-05-23 Martin Jambor <mjambor@suse.cz>
19355 PR middle-end/57347
19356 * tree.h (contains_bitfld_component_ref_p): Declare.
19357 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19358 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19360 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19361 not access a bit-field. Assert all final offsets are byte-aligned.
19363 2013-05-23 Richard Biener <rguenther@suse.de>
19365 PR tree-optimization/57380
19366 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19367 least one invariant or re-used load.
19368 * passes.c (init_optimization_passes): Move pass_phiprop before
19371 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19373 * config/aarch64/aarch64-simd.md
19374 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19376 2013-05-23 Richard Biener <rguenther@suse.de>
19378 PR middle-end/57381
19379 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19380 OEP_CONSTANT_ADDRESS_OF retained.
19382 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19384 PR middle-end/57344
19385 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19386 don't lower unit. Handle unit not being always BITS_PER_WORD.
19388 2013-05-23 Richard Biener <rguenther@suse.de>
19390 PR rtl-optimization/57341
19391 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19392 instead of true_dependence.
19394 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19396 * bb-reorder.c (branch_threshold): Make const.
19397 (exec_threshold): Ditto.
19399 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19400 Pat Haugen <pthaugen@us.ibm.com>
19401 Peter Bergner <bergner@vnet.ibm.com>
19403 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19404 documentation for the power8 crypto builtins.
19406 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19408 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19409 macros for defining power8 builtin functions.
19410 (BU_P8V_AV_2): Likewise.
19411 (BU_P8V_AV_P): Likewise.
19412 (BU_P8V_VSX_1): Likewise.
19413 (BU_P8V_OVERLOAD_1): Likewise.
19414 (BU_P8V_OVERLOAD_2): Likewise.
19415 (BU_CRYPTO_1): Likewise.
19416 (BU_CRYPTO_2): Likewise.
19417 (BU_CRYPTO_3): Likewise.
19418 (BU_CRYPTO_OVERLOAD_1): Likewise.
19419 (BU_CRYPTO_OVERLOAD_2): Likewise.
19420 (XSCVSPDP): Fix typo, point to the correct instruction.
19421 (VCIPHER): Add power8 crypto builtins.
19422 (VCIPHERLAST): Likewise.
19423 (VNCIPHER): Likewise.
19424 (VNCIPHERLAST): Likewise.
19425 (VPMSUMB): Likewise.
19426 (VPMSUMH): Likewise.
19427 (VPMSUMW): Likewise.
19428 (VPERMXOR_V2DI): Likewise.
19429 (VPERMXOR_V4SI: Likewise.
19430 (VPERMXOR_V8HI: Likewise.
19431 (VPERMXOR_V16QI: Likewise.
19432 (VSHASIGMAW): Likewise.
19433 (VSHASIGMAD): Likewise.
19434 (VPMSUM): Likewise.
19435 (VPERMXOR): Likewise.
19436 (VSHASIGMA): Likewise.
19438 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19439 __CRYPTO__ if the crypto instructions are available.
19440 (altivec_overloaded_builtins): Add support for overloaded power8
19443 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19444 support for power8 crypto builtins.
19445 (builtin_function_type): Likewise.
19446 (altivec_init_builtins): Add support for builtins that take vector
19447 long long (V2DI) arguments.
19449 * config/rs6000/crypto.md: New file, define power8 crypto
19452 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19453 Pat Haugen <pthaugen@us.ibm.com>
19454 Peter Bergner <bergner@vnet.ibm.com>
19456 * doc/invoke.texi (Option Summary): Add power8 options.
19457 (RS/6000 and PowerPC Options): Likewise.
19459 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19460 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19461 wm, wn, wr documentation.
19463 * config/rs6000/constraints.md (wm): New constraint for VSX
19464 registers if direct move instructions are enabled.
19465 (wn): New constraint for no registers.
19466 (wq): New constraint for quad word even GPR registers.
19467 (wr): New constraint if 64-bit instructions are enabled.
19468 (wv): New constraint if power8 vector instructions are enabled.
19469 (wQ): New constraint for quad word memory locations.
19471 * config/rs6000/predicates.md (const_0_to_15_operand): New
19472 constraint for 0..15 for crypto instructions.
19473 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19474 as GPR and floating point registers.
19475 (int_reg_operand): New predicate to match only GPR registers.
19476 (base_reg_operand): New predicate to match base registers.
19477 (quad_int_reg_operand): New predicate to match even GPR registers
19478 for quad memory operations.
19479 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19480 operations in both GPR and VSX registers.
19481 (quad_memory_operand): New predicate for quad memory operations.
19482 (reg_or_indexed_operand): New predicate for direct move support.
19484 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19485 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19486 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19487 (POWERPC_MASKS): Add power8 options.
19488 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19491 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19492 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19494 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19495 (-mpower8-fusion): New power8 options.
19496 (-mpower8-fusion-sign): Likewise.
19497 (-mpower8-vector): Likewise.
19498 (-mcrypto): Likewise.
19499 (-mdirect-move): Likewise.
19500 (-mquad-memory): Likewise.
19502 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19503 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19505 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19506 (rs6000_debug_vector_unit): Add p8_vector.
19507 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19508 definitions. Also print fusion state.
19509 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19510 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19511 (rs6000_option_override_internal): Add support for power8.
19512 (rs6000_common_init_builtins): Add debugging for skipped builtins
19513 if -mdebug=builtin.
19514 (rs6000_adjust_cost): Add power8 support.
19515 (rs6000_issue_rate): Likewise.
19516 (insn_must_be_first_in_group): Likewise.
19517 (insn_must_be_last_in_group): Likewise.
19518 (force_new_group): Likewise.
19519 (rs6000_register_move_cost): Likewise.
19520 (rs6000_opt_masks): Likewise.
19522 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19523 power8 capable assembler, default to power7 options.
19524 (TARGET_DIRECT_MOVE): Likewise.
19525 (TARGET_CRYPTO): Likewise.
19526 (TARGET_P8_VECTOR): Likewise.
19527 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19528 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19529 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19530 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19531 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19532 (TARGET_XSCVDPSPN): Likewise.
19533 (TARGET_XSCVSPDPN): Likewsie.
19534 (TARGET_SYNC_HI_QI): Likewise.
19535 (TARGET_SYNC_TI): Likewise.
19536 (MASK_CRYPTO): Likewise.
19537 (MASK_DIRECT_MOVE): Likewise.
19538 (MASK_P8_FUSION): Likewise.
19539 (MASK_P8_VECTOR): Likewise.
19540 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19541 temporary used by some of the direct move instructions to get two FP
19542 temporary registers does not force creation of a stack frame.
19543 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19544 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19545 that any VSX registers are tieable, even if they are also an
19546 Altivec vector mode.
19547 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19548 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19549 (RS6000_BTM_CRYPTO): Likewise.
19550 (RS6000_BTM_COMMON): Likewise.
19552 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19553 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19554 (enum rs6000_vector): Add power8 vector support.
19556 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19560 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19561 (any_sibcall_could_use_r3): this and handle indirect calls.
19562 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19564 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19566 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19568 2013-05-22 Richard Biener <rguenther@suse.de>
19570 PR middle-end/57349
19571 * profile.c (branch_prob): Do not split blocks that are
19572 abnormally receiving from ECF_RETURNS_TWICE functions.
19574 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19576 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19577 address mode. Move pointer mode initialization to the same place.
19579 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19581 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19582 while it has any effect.
19584 2013-05-21 Easwaran Raman <eraman@google.com>
19586 PR tree-optimization/57322
19587 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19588 UID of the statement added to the BB to be 1.
19590 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19592 PR tree-optimization/57331
19593 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19594 of conversion from pointer type to integral type with integer.
19596 2013-05-21 Martin Jambor <mjambor@suse.cz>
19599 * ipa-prop.c (ipa_read_node_info): Process param_used and
19600 controlled_uses in the same order as when writing.
19602 2013-05-21 Magnus Granberg <baldrick@free.fr>
19605 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19607 2013-05-21 Richard Biener <rguenther@suse.de>
19609 PR tree-optimization/57318
19610 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19611 estimate stmts with side-effects as likely eliminated.
19613 2013-05-21 Richard Biener <rguenther@suse.de>
19615 PR tree-optimization/57330
19616 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19617 preserve the call stmts fntype.
19619 2013-05-21 Richard Biener <rguenther@suse.de>
19621 PR tree-optimization/57303
19622 * tree-ssa-sink.c (statement_sink_location): Improve killing
19623 stmt detection and properly handle self-assignments.
19625 2013-05-21 Christian Bruel <christian.bruel@st.com>
19627 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19628 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19629 registers. Set register size out of the PARALLEL loop.
19631 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19634 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19635 (*fmasf4, *fmasf4_media): New insns.
19637 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19639 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19640 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19641 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19642 (mips_idiv_insns): Update the comments to say that the returned
19643 instruction counts are in units of BASE_INSN_LENGTH.
19644 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19645 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19646 using 2 rather than 4 as the length of indirect MIPS16 and
19647 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19648 length of a NOP. Don't divide MIPS16 lengths by 2.
19649 (mips16_split_long_branches): Assume a branch is long if the
19650 length is greater than 4 rather than 8.
19651 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19652 rather than multiplying them by 2. Multiply instruction counts
19653 by BASE_INSN_LENGTH rather than 4.
19654 (*jump_mips16, tls_get_tp_mips16_<mode>)
19655 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19657 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19659 * config/mips/mips.md (extended_mips16): Remove branch case.
19660 (length): Remove duplicated extended_mips16 test.
19662 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19664 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19666 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19668 * recog.h (Recog_data): Rename to...
19669 (recog_data_d): ...this.
19670 (recog_data): Update accordingly.
19671 * recog.c (recog_data): Likewise.
19672 * reload.c (save_recog_data): Likewise.
19673 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19674 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19676 2013-05-17 Julian Brown <julian@codesourcery.com>
19678 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19679 found in a REG_EQUAL note, invalidate it.
19681 2013-05-17 Easwaran Raman <eraman@google.com>
19683 * tree-ssa-reassoc.c (find_insert_point): New function.
19684 (insert_stmt_after): Likewise.
19685 (get_def_stmt): Likewise.
19686 (ensure_ops_are_available): Likewise.
19687 (not_dominated_by): Likewise.
19688 (rewrite_expr_tree): Do not move statements beyond what is
19689 necessary. Remove call to swap_ops_for_binary_stmt...
19690 (reassociate_bb): ... and move it here.
19691 (build_and_add_sum): Assign UIDs for new statements.
19692 (linearize_expr): Likewise.
19693 (do_reassoc): Renumber gimple statement UIDs.
19695 2013-05-17 Jan Hubicka <jh@suse.cz>
19697 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19699 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19700 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19702 (output_weakrefs): Update.
19704 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19705 Martin Jambor <mjambor@suse.cz>
19707 PR middle-end/57276
19708 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19709 value that corresponds to the given aggval is found in values vector.
19711 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19713 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19714 sse, sse2, sse3, ssse3 and sse4a flags to options.
19716 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19718 * gengtype-state.c: (s_expr_writer): New class, to handle
19719 prettifying of output layout of s-expressions.
19720 (state_writer): New class, to write out gtype.state.
19721 (state_written_type_count): Move this variable into member data of
19723 (s_expr_writer::s_expr_writer): New code: constructor for new class
19724 (state_writer::state_writer(): ditto
19725 (s_expr_writer::write_new_line): New function
19726 (s_expr_writer::write_any_indent): ditto
19727 (s_expr_writer::begin_s_expr): ditto
19728 (s_expr_writer::end_s_expr): ditto
19729 (write_state_fileloc): convert to method of state_writer...
19730 (state_writer:: write_state_fileloc): ...and use methods of
19731 s_expr_writer to write indentation into the gtype.state output file
19732 to visually represent the hierarchical structure of the list
19734 (write_state_fields): ditto, renaming to...
19735 (state_writer::write_state_fields)
19736 (write_state_a_string): ditto, renaming to...
19737 (state_writer::write_state_a_string)
19738 (write_state_string_option): ditto, renaming to...
19739 (state_writer::write_state_string_option)
19740 (write_state_type_option): ditto, renaming to...
19741 (state_writer::write_state_type_option)
19742 (write_state_nested_option): ditto, renaming to...
19743 (state_writer::write_state_nested_option)
19744 (write_state_option): ditto, renaming to...
19745 (state_writer::write_state_option)
19746 (write_state_options): ditto, renaming to...
19747 (state_writer::write_state_options)
19748 (write_state_lang_bitmap): ditto, renaming to...
19749 (state_writer::write_state_lang_bitmap)
19750 (write_state_version): ditto, renaming to...
19751 (state_writer::write_state_version)
19752 (write_state_scalar_type): ditto, renaming to...
19753 (state_writer::write_state_scalar_type)
19754 (write_state_string_type): ditto, renaming to...
19755 (state_writer::write_state_string_type)
19756 (write_state_undefined_type): ditto, renaming to...
19757 (state_writer::write_state_undefined_type)
19758 (write_state_struct_union_type): ditto, renaming to...
19759 (state_writer::write_state_struct_union_type)
19760 (write_state_struct_type): ditto, renaming to...
19761 (state_writer::write_state_struct_type)
19762 (write_state_user_struct_type): ditto, renaming to...
19763 (state_writer::write_state_user_struct_type)
19764 (write_state_lang_struct_type): ditto, renaming to...
19765 (state_writer::write_state_lang_struct_type)
19766 (write_state_param_struct_type): ditto, renaming to...
19767 (state_writer::write_state_param_struct_type)
19768 (write_state_pointer_type): ditto, renaming to...
19769 (state_writer::write_state_pointer_type)
19770 (write_state_array_type): ditto, renaming to...
19771 (state_writer::write_state_array_type)
19772 (write_state_gc_used): ditto, renaming to...
19773 (state_writer::write_state_gc_used)
19774 (write_state_common_type_content): ditto, renaming to...
19775 (state_writer::write_state_common_type_content)
19776 (write_state_type): ditto, renaming to...
19777 (state_writer::write_state_type)
19778 (write_state_pair_list): ditto, renaming to...
19779 (state_writer::write_state_pair_list)
19780 (write_state_pair): ditto, renaming to...
19781 (state_writer::write_state_pair)
19782 (write_state_typedefs): ditto, renaming to...
19783 (state_writer::write_state_typedefs)
19784 (write_state_structures): ditto, renaming to...
19785 (state_writer::write_state_structures)
19786 (write_state_param_structs): ditto, renaming to...
19787 (state_writer::write_state_param_structs)
19788 (write_state_variables): ditto, renaming to...
19789 (state_writer::write_state_variables)
19790 (write_state_srcdir): ditto, renaming to...
19791 (state_writer::write_state_srcdir)
19792 (write_state_files_list): ditto, renaming to...
19793 (state_writer::write_state_files_list)
19794 (write_state_languages): ditto, renaming to...
19795 (state_writer::write_state_languages)
19796 (write_state): create a state_writer instance and use it when
19797 writing out the state file
19799 2013-05-17 Mike Stump <mikestump@comcast.net>
19801 PR rtl-optimization/57304
19802 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19803 accessing DF_REF_REAL_LOC.
19805 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19807 PR rtl-optimization/57281
19808 PR rtl-optimization/57300
19809 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19810 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19811 what the other splitter did if the registers are dead.
19813 2013-05-17 Richard Biener <rguenther@suse.de>
19815 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19818 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19820 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19823 2013-05-17 Marek Polacek <polacek@redhat.com>
19825 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19826 length when doing non-zero store of storing '\0' to '\0'.
19828 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19830 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19831 vect_external_def oprnd1 with loop_vinfo, try to emit
19832 optional cast, negation and and stmts on the loop preheader
19833 edge instead of into the pattern def seq.
19835 PR tree-optimization/57051
19836 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19837 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19839 2013-05-16 Nick Clifton <nickc@redhat.com>
19841 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19842 (rl78_is_naked_func): New function.
19843 (rl78_expand_prologue): Skip prologue generation for naked functions.
19844 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19845 * doc/extend.texi (naked): Add RL78 to the list of processors
19846 that supports this attribute.
19848 2013-05-16 Jeff Law <law@redhat.com>
19850 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19852 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19854 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19855 cache parameters using detect_caches_amd also for CYRIX,
19856 NSC and TM2 signatures.
19858 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19859 Dzianis Kahanovich <mahatma@eu.by>
19863 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19864 VIA/Centaur processors and determine their cache parameters
19865 using detect_caches_amd.
19867 2013-05-16 Teresa Johnson <tejohnson@google.com>
19869 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19870 (rtl_verify_edges): New function.
19871 (rtl_verify_bb_insns): Ditto.
19872 (rtl_verify_bb_pointers): Ditto.
19873 (rtl_verify_bb_insn_chain): Ditto.
19874 (rtl_verify_fallthru): Ditto.
19875 (rtl_verify_bb_layout): Ditto.
19876 (rtl_verify_flow_info_1): Outline checks into new functions.
19877 (rtl_verify_flow_info): Ditto.
19879 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19881 * cfghooks.c (copy_bbs): Add update_dominance argument.
19882 * cfghooks.h (copy_bbs): Update prototype.
19883 * tree-cfg.c (gimple_duplicate_sese_region):
19884 Add update_dominance argument.
19885 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19886 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19887 gimple_duplicate_sese_region call.
19888 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19889 Update copy_bbs call.
19890 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19891 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19893 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19895 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19896 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19897 vect_recog_rotate_pattern.
19898 (vect_recog_rotate_pattern): New function.
19900 2013-05-16 Jason Merrill <jason@redhat.com>
19902 * Makefile.in (LLINKER): New variable.
19903 (mostlyclean): Remove link mutex.
19904 * configure.ac: Handle --enable-link-mutex.
19905 * lock-and-run.sh: New script.
19907 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19910 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19913 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19915 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19917 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19919 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19920 * config/arm/arm.c (next_consecutive_mem): New function.
19921 (gen_movmem_ldrd_strd): Likewise.
19922 * config/arm/arm.md (movmemqi): Update condition and code.
19923 (unaligned_loaddi, unaligned_storedi): New patterns.
19925 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19927 * config.gcc: Obsolete *-*-solaris2.9*.
19928 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19930 2013-05-16 Richard Biener <rguenther@suse.de>
19932 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19933 earlier, after GRAPHITE transforms and IV canonicalization.
19935 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19937 * omp-low.c (extract_omp_for_data): For collapsed loops,
19938 if at least one of the loops is known at compile time to
19939 iterate zero times, set count to 0.
19940 (expand_omp_regimplify_p): New function.
19941 (expand_omp_for_generic): For collapsed loops, if at least
19942 one of the loops isn't known to iterate at least once,
19943 add runtime check with setting count to 0.
19944 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19945 For unsigned types if it isn't known at compile time that
19946 the loop will iterate at least once, add runtime check to bypass
19947 the whole loop if initial condition isn't true.
19949 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19951 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19953 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19955 PR middle-end/57286
19956 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19957 transformations to avoid an infinite loop.
19959 2013-05-16 Marek Polacek <polacek@redhat.com>
19961 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19963 2013-05-15 Leif Ekblad <leif@rdos.net>
19965 * config/i386/i386.c (ix86_decompose_address): Use
19966 DEFAULT_TLS_SEG_REG to access TLS segment register.
19967 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19968 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19969 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19971 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19974 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19975 sibling calls to functions that would normally be lazily bound,
19976 unless $gp is call-clobbered.
19978 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19980 * config/i386/i386.c (ix86_option_override_internal): Update
19981 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19982 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19983 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19984 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19986 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19988 2013-05-15 Andreas Schwab <schwab@suse.de>
19990 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19991 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19994 2013-05-15 Teresa Johnson <tejohnson@google.com>
19996 * loop-unroll.c (report_unroll_peel): Check decision before
19997 emitting unroll/peel message.
19999 2013-05-15 Teresa Johnson <tejohnson@google.com>
20001 * function.h (has_bb_partition): New rtl_data flag.
20002 (bb_reorder_complete): Ditto.
20003 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20004 instead of flag_reorder_blocks_and_partition.
20005 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20006 with some enhancements.
20007 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20008 * bb-reorder.c (connect_traces): Check for has_bb_partition
20009 instead of flag_reorder_blocks_and_partition.
20010 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20011 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20012 verify_hot_cold_block_grouping.
20013 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20015 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20018 * config/arm/predicates.md (call_insn_operand): New predicate.
20019 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20020 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20021 if insn is not a tail call.
20022 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20024 * config/arm/arm.h (enum reg_class): New caller save register class.
20025 (REG_CLASS_NAMES): Likewise.
20026 (REG_CLASS_CONTENTS): Likewise.
20027 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20030 2013-05-15 Richard Biener <rguenther@suse.de>
20032 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20033 of MSG_OPTIMIZED_LOCATIONS.
20034 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20035 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20037 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20038 of MSG_OPTIMIZED_LOCATIONS.
20039 (execute_vect_slp): Likewise.
20040 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20041 (vect_create_cond_for_alias_checks): Likewise.
20042 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20043 (vect_recog_widen_mult_pattern): Likewise.
20044 (vect_recog_widen_sum_pattern): Likewise.
20045 (vect_recog_over_widening_pattern): Likewise.
20046 (vect_recog_widen_shift_pattern): Likewise.
20047 (vect_recog_vector_vector_shift_pattern): Likewise.
20048 (vect_recog_divmod_pattern): Likewise.
20049 (vect_recog_mixed_size_cond_pattern): Likewise.
20050 (vect_recog_bool_pattern): Likewise.
20051 (vect_pattern_recog_1): Likewise.
20053 2013-05-15 Martin Jambor <mjambor@suse.cz>
20055 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20056 non-functions to builtin_unreachable.
20057 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20058 correct when new direct edges were discovered.
20060 2013-05-15 Martin Jambor <mjambor@suse.cz>
20062 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20063 header, print symbol order instead of node uid, print more information
20064 about indirect edge targets.
20065 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20067 (ipa_make_edge_direct_to_target): Likewise.
20068 (remove_described_reference): Likewise.
20069 (propagate_controlled_uses): Likewise.
20070 (ipa_print_node_params): Also print symbol order.
20071 (ipcp_transform_function): Print symbol order instead of node uids.
20072 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20073 (cgraph_get_create_real_symbol_node): Likewise.
20074 * ipa-cp.c (print_lattice): Likewise.
20075 (print_all_lattices): Likewise.
20076 (determine_versionability): Likewise.
20077 (initialize_node_lattices): Likewise.
20078 (estimate_local_effects): Likewise.
20079 (update_profiling_info): Likewise.
20080 (create_specialized_node): Likewise.
20081 (perhaps_add_new_callers): Likewise.
20082 (decide_about_value): Likewise.
20083 (decide_whether_version_node): Likewise.
20084 (identify_dead_nodes): Likewise.
20085 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20086 (dump_inline_summary): Likewise.
20087 (estimate_node_size_and_time): Likewise.
20088 (inline_analyze_function): Likewise.
20089 * ipa-inline.c (report_inline_failed_reason): Likewise.
20090 (want_early_inline_function_p): Likewise.
20091 (edge_badness): Likewise.
20092 (update_edge_key): Likewise.
20093 (inline_small_functions): Likewise. Add dumping of order to two other
20095 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20096 instead of node uids.
20097 (propagate_pure_const): Likewise.
20098 (propagate_pure_const): Likewise.
20099 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20100 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20101 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20103 * tree-pretty-print.c (dump_function_header): Likewise.
20104 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20105 Print symbol order instead of node uids.
20107 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20109 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20110 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20112 2013-05-15 Richard Biener <rguenther@suse.de>
20114 PR tree-optimization/57275
20115 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20116 return value for fail to do runtime alias checks for gather loads.
20118 2013-05-15 Jan Hubicka <jh@suse.cz>
20122 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20123 weakrefs are not external.
20124 (lto_symtab_merge_decls): Fix thinko when dealing with
20125 non-lto_symtab decls.
20126 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20127 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20128 * varpool.c (dump_varpool_node): Dump more flags.
20130 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20132 * config/i386/i386.c (processor_alias_table): Add instruction
20133 FSGSBASE for AMD bdver3 architecture.
20135 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20137 * tree.c (warn_deprecated_use): Print file:line using locus color.
20138 * diagnostic.c (diagnostic_report_current_module): Print file:line
20139 and file:line:column using locus color.
20141 2013-05-14 Mike Stump <mikestump@comcast.net>
20143 * gdbinit.in: Add __null.
20145 2013-05-14 Mike Stump <mikestump@comcast.net>
20147 * recog.h: Rename struct recog_data to Recog_data.
20148 * recog.c: Likewise.
20149 * reload.c (can_reload_into): Likewise.
20150 * config/picochip/picochip.c: Likewise.
20152 2013-05-14 Mike Stump <mikestump@comcast.net>
20154 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20156 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20158 * resource.h (struct resources): Remove unch_memory member.
20159 (CLEAR_RESOURCE): Don't clear unch_memory.
20160 * resource.c (mark_referenced_resources): Don't set it.
20161 (mark_set_resources): Likewise.
20162 (mark_target_live_regs): Don't clear it.
20163 (init_resource_info): Likewise.
20164 * reorg.c (resource_conflicts_p): Don't compare it.
20165 (redundant_insn): Don't set it.
20167 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20169 * emit-rtl.c (next_label): Remove unused function.
20170 (skip_consecutive_labels, link_cc0_insns): Move to ...
20171 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20172 only place where these functions are used, and make them static.
20174 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20176 * fold-const.c (fold_negate_expr): Handle vectors.
20177 (fold_truth_not_expr): Make it static.
20178 (fold_invert_truthvalue): New static function.
20179 (invert_truthvalue_loc): Handle vectors. Do not call
20180 fold_truth_not_expr directly.
20181 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20182 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20183 (fold_comparison): Handle vector constants.
20184 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20185 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20186 * tree.h (fold_truth_not_expr): Remove declaration.
20188 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20190 * config/aarch64/aarch64-simd.md
20191 (aarch64_vcond_internal<mode>): Rename to...
20192 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20193 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20194 float modes. Clarify all iterator modes.
20195 (vcond<mode><mode>): Use new name for vcond expanders.
20196 (vcond<v_cmp_result><mode>): Likewise.
20197 (vcondu<mode><mode>: Likewise.
20198 * config/aarch64/iterators.md (VDQF_COND): New.
20200 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20203 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20204 variable for the shift amount. Check that we shift by non-negative
20207 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20210 * config/arm/arm.h (EPILOGUE_USES): Only return true
20211 for LR_REGNUM after epilogue_completed.
20213 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20215 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20216 is error_mark_node.
20218 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20221 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20222 and Solaris 11+/x86 with gld.
20223 * configure: Regenerate.
20225 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20227 * expmed.c (expand_shift_1): Canonicalize rotates by
20228 constant bitsize / 2 to bitsize - 1.
20229 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20230 case ROTATERT>: Likewise.
20233 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20235 * config/i386/i386.md (rotateinv): New code attr.
20236 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20237 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20238 roll $31, %eax, etc.
20240 2013-05-14 Richard Biener <rguenther@suse.de>
20242 PR middle-end/57235
20243 * tree-eh.c (sink_clobbers): Give up for successors with
20244 multiple predecessors and no virtual uses.
20246 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20248 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20249 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20251 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20253 PR middle-end/57251
20254 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20255 the case when both op0 and op1 have VOIDmode.
20257 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20259 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20260 in multiply-accumulate mode.
20262 2013-05-13 Guozhi Wei <carrot@google.com>
20264 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20266 2013-05-13 Kai Tietz <ktietz@redhat.com>
20269 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20270 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20271 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20272 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20273 (ix86_expand_prologue): Likewise.
20274 (ix86_expand_split_stack_prologue): Likewise.
20275 (legitimate_pic_address_disp_p): Likewise.
20276 (legitimize_pic_address): Likewise.
20277 (legitimize_tls_address): Likewise.
20278 (legitimize_pe_coff_symbol): Likewise.
20279 (output_pic_addr_const): Likewise.
20280 (construct_plt_address): Likewise.
20281 (ix86_expand_call): Likewise.
20282 (x86_output_mi_thunk): Likewise.
20283 (x86_function_profiler): Likewise.
20285 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20287 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20288 similar switch cases.
20289 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
20290 (aarch64_simd_mov_to_<mode>low): Delete.
20291 (aarch64_simd_mov_to_<mode>high): Delete.
20292 (move_lo_quad_<mode>): Add w<-r alternative.
20293 (aarch64_simd_move_hi_quad_<mode>): Likewise.
20294 (aarch64_simd_mov_from_*): Update type attribute.
20295 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
20298 2013-05-13 Jan Hubicka <jh@suse.cz>
20300 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
20301 * config/i386/i386.c (ix86_compute_frame_layout,
20302 ix86_expand_epilogue, emit_i387_cw_initialization,
20303 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
20304 ix86_local_alignment): Fix use of size/speed predicates.
20306 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20308 PR tree-optimization/45216
20309 PR tree-optimization/57157
20310 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
20311 the (-Y) & (B - 1) variant if OP is |.
20312 * expmed.c (expand_shift_1): For rotations by const0_rtx just
20313 return shifted. Use (-op1) & (prec - 1) as other_amount
20314 instead of prec - op1.
20316 2013-05-13 Martin Jambor <mjambor@suse.cz>
20318 PR middle-end/42371
20319 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
20320 (ipa_constant_data): New type.
20321 (ipa_jump_func): Use ipa_constant_data to hold information about
20322 constant jump functions.
20323 (ipa_get_jf_constant): Adjust to jump function type changes.
20324 (ipa_get_jf_constant_rdesc): New function.
20325 (ipa_param_descriptor): New field controlled_uses.
20326 (ipa_get_controlled_uses): New function.
20327 (ipa_set_controlled_uses): Likewise.
20328 * ipa-ref.h (ipa_find_reference): Declare.
20329 * ipa-prop.c (ipa_cst_ref_desc): New type.
20330 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
20332 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
20333 New parameter cs. Adjust all callers.
20334 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
20335 (remove_described_reference): New function.
20336 (jfunc_rdesc_usable): Likewise.
20337 (try_make_edge_direct_simple_call): Decrement controlled use count,
20338 attempt to remove reference if it hits zero.
20339 (combine_controlled_uses_counters): New function.
20340 (propagate_controlled_uses): Likewise.
20341 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
20342 (ipa_edge_duplication_hook): Duplicate reference descriptions.
20343 (ipa_print_node_params): Print described use counter.
20344 (ipa_write_jump_function): Adjust to jump function type changes.
20345 (ipa_read_jump_function): New parameter CS, pass it to
20346 ipa_set_jf_constant. Adjust caller.
20347 (ipa_write_node_info): Stream controlled use count
20348 (ipa_read_node_info): Likewise.
20349 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
20351 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
20352 count. Remove cloning-added reference if it reaches zero.
20353 * ipa-ref.c (ipa_find_reference): New function.
20355 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20357 * config/i386/i386.c (processor_target_table): Modified default
20358 alignment values for AMD BD and BT architectures.
20360 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20362 * tree-vect-generic.c (uniform_vector_p): Move ...
20363 * tree.c (uniform_vector_p): ... here.
20364 * tree.h (uniform_vector_p): Declare it.
20365 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20368 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20370 PR tree-optimization/57230
20371 * tree-ssa-strlen.c (handle_char_store): Record length for
20372 array store from STRING_CST.
20374 PR tree-optimization/57230
20375 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20378 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20380 * config/epiphany/epiphany.c (epiphany_init): Check size of
20381 NUM_MODES_FOR_MODE_SWITCHING.
20382 (epiphany_expand_prologue):
20383 Remove CONFIG_REGNUM initial value handling code.
20384 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20385 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20386 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20387 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20388 Don't return 1 for FP_MODE_NONE.
20389 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20390 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20391 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20392 * config/epiphany/epiphany.md (save_config): New pattern.
20394 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20396 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20398 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20400 * config/i386/i386.md (memory): Handle sseishft1.
20401 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20402 (*vec_extractv2di_1): Ditto.
20404 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20406 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20409 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20411 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20412 Add mthumb/march=armv7-a multilib.
20413 Add mthumb/march=armv7-r multilib.
20414 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20416 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20418 * config/v850/t-rtems: Add more multilibs.
20420 2013-05-10 Richard Biener <rguenther@suse.de>
20422 PR tree-optimization/57214
20423 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20424 not propagate from SSA names that occur in abnormal PHI nodes.
20426 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20428 * stor-layout.c (element_precision): New function.
20429 * machmode.h (element_precision): Declare it.
20430 * tree.c (build_minus_one_cst): New function.
20431 (element_precision): Likewise.
20432 * tree.h (build_minus_one_cst): Declare new function.
20433 (element_precision): Likewise.
20434 * fold-const.c (operand_equal_p): Use element_precision.
20435 (fold_binary_loc): Handle vector types.
20436 * convert.c (convert_to_integer): Use element_precision.
20437 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20440 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20442 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20443 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20444 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20445 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20446 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20447 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20448 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20449 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20450 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20451 (Uuw8): New constraints.
20452 (Usb4): Move into alphabetical order.
20453 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20454 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20455 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20456 previously unnamed patterns.
20457 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20458 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20459 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20460 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20461 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20462 of specific lengths.
20464 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20466 * config/i386/i386.md (rotateinv): New code attr.
20467 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20468 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20469 roll $31, %eax, etc.
20471 PR tree-optimization/45216
20472 PR tree-optimization/57157
20473 * tree-ssa-forwprop.c (simplify_rotate): New function.
20474 (ssa_forward_propagate_and_combine): Call it.
20476 2013-05-10 Richard Biener <rguenther@suse.de>
20478 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20479 disable peeling when we version for aliasing.
20480 (vector_alignment_reachable_p): Honor explicit user alignment.
20481 (vect_supportable_dr_alignment): Likewise.
20482 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20483 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20484 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20485 then peeling to arrange for the cost-model check to come first.
20487 2013-05-10 Alan Modra <amodra@gmail.com>
20489 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20490 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20491 * configure: Regenerate.
20493 2013-05-10 Alan Modra <amodra@gmail.com>
20496 * varasm.c (default_elf_select_section): Move !DECL_P check..
20497 (get_named_section): ..to here before calling get_section_name.
20499 (default_section_type_flags): Add DECL_P check.
20500 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20501 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20503 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20505 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20506 When using gen_stack_adjust_str with a register offset, add a
20507 REG_FRAME_RELATED_EXPR note.
20509 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20511 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20512 (*vec_extractv4si_zext_mem): Ditto.
20513 (*vec_extractv2di): Add 0->x and x->x alternatives.
20514 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20515 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20517 2013-05-09 Jason Merrill <jason@redhat.com>
20519 N3639 C++1y VLA support
20520 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20523 * tree.c (build_constructor_va): New.
20524 * tree.h: Declare it.
20526 2013-05-09 Martin Jambor <mjambor@suse.cz>
20529 * gimple-fold.c (canonicalize_constructor_val): Call
20530 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20532 2013-05-09 Jan Hubicka <jh@suse.cz>
20533 Richard Biener <rguenther@suse.de>
20536 * symtab.c (symtab_make_decl_local): Do not add private names.
20538 2013-05-09 Jan Hubicka <jh@suse.cz>
20541 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20542 (unlink_from_assembler_name_hash): Likewise.
20543 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20544 symtab_unregister_node, symtab_initialize_asm_name_hash,
20545 change_decl_assembler_name): Update.
20547 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20549 * config/aarch64/aarch64.md: New movtf split.
20550 (*movtf_aarch64): Update.
20551 (aarch64_movdi_tilow): Handle TF modes and rename to
20552 aarch64_movdi_<mode>low.
20553 (aarch64_movdi_tihigh): Handle TF modes and rename to
20554 aarch64_movdi_<mode>high
20555 (aarch64_movtihigh_di): Handle TF modes and rename to
20556 aarch64_mov<mode>high_di
20557 (aarch64_movtilow_di): Handle TF modes and rename to
20558 aarch64_mov<mode>low_di
20559 (aarch64_movtilow_tilow): Remove spurious whitespace.
20560 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20562 (aarch64_print_operand): Update.
20564 2013-05-09 Alan Modra <amodra@gmail.com>
20566 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20568 * configure: Regenerate.
20570 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20572 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20573 splitter preparation statements.
20574 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20575 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20578 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20580 * gimple-ssa-strength-reduction.c (count_candidates): Change
20581 return value to int.
20582 (analyze_candidates_and_replace): Change type of length to int.
20584 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20586 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20587 (*vec_extract<mode>): Use VI12_128 mode iterator.
20588 (*vec_extract<mode>_mem): Ditto.
20589 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20592 2013-05-08 Diego Novillo <dnovillo@google.com>
20597 2012-08-17 Diego Novillo <dnovillo@google.com>
20600 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20601 * config.in: Regenerate.
20602 * configure: Regenerate.
20603 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20605 2013-05-08 Jan Hubicka <jh@suse.cz>
20608 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20609 * cgraph.h (symtab_node_base): Add unique_name.
20610 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20611 input_overwrite_node, input_varpool_node): Stream unique_name.
20612 * cgraphclones.c (cgraph_create_virtual_clone,
20613 cgraph_function_versioning): Set unique_name.
20614 * ipa.c (function_and_variable_visibility): Set unique_name.
20616 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20618 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20619 (alloc_cand_and_find_basis): Restrict conditional candidate
20620 processing to CAND_MULTs.
20622 2013-05-08 Jan Hubicka <jh@suse.cz>
20625 lto-symtab.c (lto_symtab_symbol_p): New function.
20626 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20627 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20628 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20629 Skip static symbols.
20631 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20633 PR tree-optimization/57200
20634 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20635 Only call inform if the preceding warning_at returns true.
20637 2013-05-07 Han Shen <shenhan@google.com>
20639 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20640 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20641 * common.opt (fstack-protector-strong): New option.
20642 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20643 * doc/invoke.texi (Optimization Options): Document
20644 "-fstack-protector-strong".
20645 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20647 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20649 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20651 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20653 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20654 it is not lhs of the insn.
20655 (setup_reg_equiv): Remove insn having equiv memory which it is not
20657 * lra-constraints.c (process_address): Try to improve generation
20658 code for address base + disp.
20659 (lra_constraints): Make correct the code for checking insn setting
20660 up backward equivalence. Remove insn only if it is in the init
20662 * lra-eliminations.c (update_reg_eliminate): Change return value.
20663 (lra_eliminate): Use the result.
20665 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20667 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20668 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20669 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20670 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20671 register target operands.
20672 (*vec_extractv8hi_sse2): New pattern.
20673 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20674 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20675 (*vec_extract<mode>_mem): New insn and split pattern.
20677 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20679 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20680 (TARGET_ASAN_SHADOW_OFFSET): Define.
20681 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20682 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20684 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20686 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20687 (incr_vec_index): Return -1 if increment not found.
20688 (create_add_on_incoming_edge): Assert if increment not found.
20689 (record_increment): Limit number of increments recorded.
20690 (all_phi_incrs_profitable): Return false if an increment not found.
20691 (replace_profitable_candidates): Don't process increments that were
20693 (analyze_candidates_and_replace): Limit size of incr_vec.
20695 2013-05-07 Richard Biener <rguenther@suse.de>
20697 * calls.c (special_function_p): setjmp-like functions are leaf.
20698 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20699 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20701 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20703 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20704 (aarch64_simd_mov<mode>): New expander.
20705 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20706 (aarch64_simd_mov_to_<mode>high): Likewise.
20707 (aarch64_simd_mov_from_<mode>low): Likewise.
20708 (aarch64_simd_mov_from_<mode>high): Likewise.
20709 (aarch64_dup_lane<mode>): Update.
20710 (aarch64_dup_lanedi): New instruction pattern.
20711 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20712 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20714 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20716 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20717 (replace_mult_candidate): Remove unnecessary argument; remove
20718 unnecessary parameter from call to introduce_cast_before_cand.
20719 (replace_unconditional_candidate): Remove unnecessary parameter
20720 from call to replace_mult_candidate.
20721 (replace_conditional_candidate): Likewise.
20722 (insert_initializers): Use make_temp_ssa_name.
20723 (introduce_cast_before_cand): Remove unnecessary argument; use
20724 make_temp_ssa_name.
20725 (replace_one_candidate): Remove unnecessary argument; remove
20726 unnecessary parameter from calls to introduce_cast_before_cand.
20727 (replace_profitable_candidates): Remove unnecessary parameters
20728 from calls to replace_one_candidate.
20730 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20732 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20733 phi def as possibly hiding a basis for a CAND_ADD whose operands
20734 have been commuted in the analysis.
20735 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20737 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20739 * config/aarch64/aarch64.md
20740 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20741 shift value between 0-4.
20743 2013-05-07 Richard Biener <rguenther@suse.de>
20745 * double-int.h (rshift): New overload.
20746 * double-int.c (rshift): New function.
20747 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20748 (create_reference_ops_from_ref): Remove.
20749 (vn_reference_insert): Use shared ops for constructing the
20750 reference and copy it.
20752 2013-05-07 Richard Biener <rguenther@suse.de>
20754 PR middle-end/57190
20755 * tree-eh.c (sink_clobbers): Properly propagate
20756 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20758 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20760 PR tree-optimization/57149
20761 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20762 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20763 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20764 uninit_undefined_value_p instead of ssa_undefined_value_p.
20767 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20768 for modifier == EXPAND_INITIALIZER.
20770 2013-05-07 Anton Blanchard <anton@samba.org>
20772 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20773 for powerpc64 little endian.
20774 * configure: Regenerate.
20776 2013-05-06 Graham Stott <grahams@btinternet.com>
20778 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20779 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20780 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20781 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20783 2013-05-06 Graham Stott <grahams@btinternet.com>
20785 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20786 codes which allow non-lvalues.
20788 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20790 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20791 components are all 1s.
20792 (integer_minus_onep): New function.
20793 * tree.h (integer_minus_onep): Declare it.
20794 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20795 integer_minus_onep instead of integer_all_onesp.
20797 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20800 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20801 variations of these patterns.
20803 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20805 * config/i386/i386.md (isa): Add x64_sse4 member.
20806 (enabled): Handle x64_sse4.
20807 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20808 instruction for 64bit SSE4_1 targets. Update insn attributes.
20809 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20810 instruction for SSE4_1 targets. Update insn attributes.
20811 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20812 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20813 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20815 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20816 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20817 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20819 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20822 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20824 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20826 * final.c (do_assembler_dialects): Don't handle curly braces and
20827 vertical bar escaped by % as dialect delimiters.
20828 (output_asm_insn): Print curly braces and vertical bar if escaped
20829 by % and ASSEMBLER_DIALECT defined.
20830 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20831 * doc/tm.texi: Regenerated.
20833 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20835 * config/mips/mips.c: Include tree-pass.h.
20836 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20837 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20838 (pass_mips_machine_reorg2): New machine specific pass.
20839 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20840 (mips_option_override): Register the new pass.
20841 * rtl.h (cleanup_barriers): Remove prototype.
20842 (dbr_schedule): Likewise.
20843 * jump.c (cleanup_barriers): Make static.
20844 * reorg.c (dbr_schedule): Likewise.
20846 2013-05-06 Richard Biener <rguenther@suse.de>
20848 PR tree-optimization/57185
20849 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20850 SSA names properly.
20852 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20855 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20856 between "first_arg" and "insn", not "last" and "insn".
20858 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20860 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20861 (find_candidates_in_block): Re-enable slsr_process_phi.
20862 (create_phi_basis): Fix double counting of candidate adjustment.
20864 2013-05-06 Richard Biener <rguenther@suse.de>
20866 PR middle-end/57147
20867 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20868 the edge is also fallthru, preserve it and just clear the
20870 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20871 also complex, preserve that and just clear the fallthru flag.
20872 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20873 update virtual operands.
20875 2013-05-06 Alan Modra <amodra@gmail.com>
20877 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20878 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20879 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20880 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20881 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20882 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20884 2013-05-06 Alan Modra <amodra@gmail.com>
20886 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20887 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20888 (DEFAULT_ASM_ENDIAN): Define.
20889 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20890 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20891 Update -K PIC clause from sysv4.h.
20892 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20893 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20895 2013-05-06 Alan Modra <amodra@gmail.com>
20897 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20898 twice for little-endian.
20899 (ashrdi3_no_power, ashrdi3): Support little-endian.
20901 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20904 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20905 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20907 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20908 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20909 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20912 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20913 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20915 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20916 * config/pa/pa.opt: Make mbig-switch a no-op.
20917 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20918 (CASE_VECTOR_MODE): Always return SImode.
20919 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20920 for the !TARGET_BIG_SWITCH case.
20921 * config/pa/pa-linux.h: Likewise.
20922 * config/pa/pa-openbsd.h: Likewise.
20923 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20924 * config/pa/pa.md (short_jump): Remove define_insn.
20925 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20926 (casesi0): Remove define_insn.
20927 (type): Remove btable_branch.
20928 (pa_combine_type): Likewise.
20929 (in_nullified_branch_delay): Likewise.
20930 (in_call_delay): Likewise.
20931 (define_delay): Likewise.
20932 (define_insn_reservation "Z3"): Likewise.
20933 (define_insn_reservation "Z4"): Likewise.
20934 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20935 (pa_adjust_insn_length): Remove adjustment for btable branches.
20936 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20939 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20941 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20942 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20943 Add m->r,x alternatives.
20944 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20945 splitters using SWI48x mode iterator.
20946 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20947 TARGET_64BIT. Add m->x alternative.
20948 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20949 Add o->x alternative. Enable for TARGET_SSE.
20950 (sse_storeq): Remove expander.
20951 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20952 with memory input operand.
20953 (*vec_extractv2di_1 splitter): New.
20954 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20955 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20957 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20959 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20960 (INT_LOWPART): Delete.
20961 (extract_MB): Adjust.
20962 (extract_ME): Adjust.
20963 (print_operand): Adjust.
20965 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20967 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20968 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20969 (reg_or_logical_cint_operand, easy_fp_constant,
20970 logical_const_operand): Delete "CONST_DOUBLE" case.
20971 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20972 "HOST_BITS_PER_WIDE_INT == 64" test.
20973 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20974 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20975 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20976 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20978 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20979 CONST_DOUBLE DImode/VOIDmode case.
20980 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20981 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20982 CONST_DOUBLE VOIDmode case.
20983 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20984 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20985 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20986 Delete CONST_DOUBLE case.
20987 (splitters for mov FMOVE64 const_double): Delete
20988 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20989 "HOST_BITS_PER_WIDE_INT >= 64" test.
20990 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20992 (mov DI const_double): Delete.
20994 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20996 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20997 on op shows all bits zero in mode of a lowpart subreg, return zero.
20999 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21002 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21003 to save TFmode registers and DImode to save TImode registers for
21004 caller save operations.
21005 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21006 mark being partially clobbered since they only use the first
21009 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21010 and TDmode only use the upper 64-bits of each VSX register.
21012 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21014 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21015 (find_candidates_in_block): Disable slsr_process_phi.
21017 2013-05-03 Guozhi Wei <carrot@google.com>
21019 * coverage.c (coverage_obj_init): Move the construction of gcov
21021 (build_init_ctor): ... here.
21023 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21025 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21026 (slsr_cand_d): Redefine def_phi.
21027 (stride_status, phi_adjust_status, count_phis_status): New enums.
21028 (find_phi_def): New.
21029 (find_basis_for_base_expr): New.
21030 (find_basis_for_candidate): Handle hidden bases.
21031 (alloc_cand_and_find_basis): Handle phi candidates.
21032 (slsr_process_phi): New.
21033 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21034 (create_mul_imm_cand): Likewise.
21035 (create_add_ssa_cand): Exclude phi base candidates.
21036 (create_add_imm_cand): Likewise.
21037 (slsr_process_cast): Likewise.
21038 (slsr_process_copy): Likewise.
21039 (find_candidates_in_block): Handle phi candidates.
21040 (dump_candidate): Likewise.
21041 (unconditional_cands): Delete.
21042 (unconditional_cands_with_known_stride_p): Delete.
21043 (phi_dependent_cand_p): New.
21044 (cand_increment): Handle phi-dependent candidates.
21045 (replace_dependent): Delete.
21046 (replace_mult_candidate): New.
21047 (replace_unconditional_candidate): New.
21048 (incr_vec_index): Move to avoid forward reference.
21049 (create_add_on_incoming_edge): New.
21050 (create_phi_basis): New.
21051 (replace_dependents): Delete.
21052 (replace_conditional_candidate): New.
21053 (phi_add_costs): New.
21054 (replace_uncond_cands_and_profitable_phis): New.
21055 (record_increment): Handle phi adjustments.
21056 (record_phi_increments): New.
21057 (record_increments): Handle phi adjustments.
21058 (phi_incr_cost): New.
21059 (lowest_cost_path): Handle phis.
21060 (total_savings): Likewise.
21061 (analyze_increments): Likewise.
21062 (ncd_with_phi): New.
21063 (ncd_of_cand_and_phis): New.
21064 (nearest_common_dominator_for_cands): Handle phi increments.
21065 (all_phi_incrs_profitable): New.
21066 (replace_profitable_candidates): Handle phi-dependent candidates.
21067 (analyze_candidates_and_replace): Likewise.
21069 2013-05-03 Teresa Johnson <tejohnson@google.com>
21072 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21073 do not exceed REG_BR_PROB_BASE.
21075 2013-05-03 Jeff Law <law@redhat.com>
21077 PR tree-optimization/57144
21078 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21079 operand of the condition will bit into the new type when eliminating
21080 a cast feeding a condition.
21082 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21084 PR rtl-optimization/57130
21085 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21086 of COMPARE as in_code to the recursive call if needed.
21088 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21090 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21091 (enabled): Handle new members.
21092 * config/i386/sse.md (*vec_concatv2si): Merge from
21093 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21094 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21096 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21098 PR tree-optimization/57027
21099 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21100 for fnms opportunity, check we got the prerequisite kind
21101 of tree / gimple before using accessor functions.
21103 2013-05-03 Richard Biener <rguenther@suse.de>
21105 * double-int.h (lshift): New overload without precision
21106 and arith argument.
21107 (operator *=, operator +=, operator -=): Move ...
21108 * double-int.c (operator *=, operator +=, operator -=): ... here
21109 and implement more efficiently.
21110 (mul_double_with_sign): Remove.
21111 (lshift_double): Adjust to take unsinged shift argument, push
21112 dispatching code to callers.
21113 (mul_double_wide_with_sign): Add early out for callers that
21114 are not interested in high parts or overflow.
21115 (lshift): New function.
21116 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21117 dispatch code here.
21118 (lrotate, rrotate): Use logical shifts.
21119 * expr.c (get_inner_reference): Use lshift.
21120 * fixed-value.c (do_fixed_divide): Likewise.
21121 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21122 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21123 (indirect_refs_may_alias_p): Likewise.
21124 (stmt_kills_ref_p_1): Likewise.
21126 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21128 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21130 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21132 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21133 scalar form of FABD instruction.
21135 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21137 * lra-constraints.c (process_alt_operands): Add checking alt
21138 number to choose the best alternative.
21140 2013-05-02 Richard Biener <rguenther@suse.de>
21142 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21143 bitmap and its handling.
21144 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21146 2013-05-02 Richard Biener <rguenther@suse.de>
21148 PR middle-end/57140
21149 * tree-inline.c (copy_loops): Properly handle removed loops.
21150 (copy_cfg_body): Mark destination loops for fixup if source
21151 loops needed fixup.
21153 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21156 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21157 generating simple_return for naked functions.
21159 2013-05-02 Martin Jambor <mjambor@suse.cz>
21161 PR middle-end/56988
21162 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21163 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21165 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21166 ipa_agg_replacement_value structures.
21167 (known_aggs_to_agg_replacement_list): Likewise.
21168 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21169 (read_agg_replacement_chain): Likewise.
21170 (ipcp_transform_function): Also check that by_ref flags match.
21172 2013-05-02 Richard Biener <rguenther@suse.de>
21174 * graphds.h (struct graph): Add obstack member.
21175 * graphds.c (new_graph): Initialize obstack and allocate
21177 (add_edge): Allocate edge from the obstack.
21178 (free_graph): Free the obstack instead of all edges and vertices.
21180 2013-05-02 Teresa Johnson <tejohnson@google.com>
21182 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21184 * cfg.c (update_bb_profile_for_threading): Ditto.
21185 * tree-inline.c (copy_bb): Ditto.
21186 (copy_edges_for_bb): Ditto.
21187 (initialize_cfun): Ditto.
21188 (copy_cfg_body): Ditto.
21189 (expand_call_inline): Ditto.
21190 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21191 (estimate_node_size_and_time): Ditto.
21192 (inline_merge_summary): Ditto.
21193 * cgraphclones.c (cgraph_clone_edge): Ditto.
21194 (cgraph_clone_node): Ditto.
21195 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21196 (compute_trg_info): Ditto.
21198 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21200 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21201 S reg when fp attribute set.
21202 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21204 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21206 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21208 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21209 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21210 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21212 2013-05-02 Richard Biener <rguenther@suse.de>
21214 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21215 (struct instantiate_cache_entry): New type.
21216 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21217 (struct instantiate_cache_type): New type.
21218 (set_instantiated_value, get_instantiated_value): Remove.
21219 (get_instantiated_value_entry): New function.
21220 (instantiate_scev_name): Use the new cache and adjust.
21221 (instantiate_scev_poly): Adjust.
21222 (instantiate_scev_binary): Likewise.
21223 (instantiate_array_ref): Likewise.
21224 (instantiate_scev_convert): Likewise.
21225 (instantiate_scev_not): Likewise.
21226 (instantiate_scev_3): Likewise.
21227 (instantiate_scev_2): Likewise.
21228 (instantiate_scev_r): Likewise.
21229 (instantiate_scev): Likewise.
21230 (resolve_mixers): Likewise.
21232 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21235 * lra-constraints.c (best_small_class_operands_num): Remove.
21236 (process_alt_operands): Remove small_class_operands_num. Take
21237 small classes operands into losers and only if the operand is not
21238 matched. Modify debugging output.
21239 (curr_insn_transform): Remove best_small_class_operands_num.
21242 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21244 * config/aarch64/aarch64-builtins.c
21245 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21246 * config/aarch64/aarch64-simd-builtins.def
21247 (reduc_splus_): Add new modes.
21248 (reduc_uplus_): New.
21249 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21250 (reduc_uplus_v4sf): Likewise.
21251 (reduc_splus_v4sf): Likewise.
21252 (aarch64_addv<mode>): Likewise.
21253 (reduc_uplus_<mode>): Likewise.
21254 (reduc_splus_<mode>): Likewise.
21255 (aarch64_addvv2di): Likewise.
21256 (reduc_uplus_v2di): Likewise.
21257 (reduc_splus_v2di): Likewise.
21258 (aarch64_addvv2si): Likewise.
21259 (reduc_uplus_v2si): Likewise.
21260 (reduc_splus_v2si): Likewise.
21261 (reduc_<sur>plus_<mode>): New.
21262 (reduc_<sur>plus_v2di): Likewise.
21263 (reduc_<sur>plus_v2si): Likewise.
21264 (reduc_<sur>plus_v4sf): Likewise.
21265 (aarch64_addpv4sf): Likewise.
21266 * config/aarch64/arm_neon.h
21267 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21268 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21269 add UNSPEC_SADDV, UNSPEC_UADDV.
21271 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21273 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21275 * config/aarch64/arm_neon.h
21276 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21278 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21280 * config/aarch64/aarch64-builtins
21281 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21283 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21285 * config/aarch64/aarch64-simd-builtins.def
21286 (reduc_smax_): New.
21287 (reduc_smin_): Likewise.
21288 (reduc_umax_): Likewise.
21289 (reduc_umin_): Likewise.
21290 (reduc_smax_nan_): Likewise.
21291 (reduc_smin_nan_): Likewise.
21294 (smax): Update for V2SF, V4SF and V2DF modes.
21297 (smin_nan): Likewise.
21298 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
21299 (<su><maxmin><mode>3): ...This, refactor.
21300 (s<maxmin><mode>3): New.
21301 (<maxmin_uns><mode>3): Likewise.
21302 (reduc_<maxmin_uns>_<mode>): Refactor.
21303 (reduc_<maxmin_uns>_v4sf): Likewise.
21304 (reduc_<maxmin_uns>_v2si): Likewise.
21305 (aarch64_<fmaxmin><mode>: Remove.
21306 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
21308 (vmin<q>_f<32,64>): Likewise.
21309 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21311 (su): Add mappings for smax, smin, umax, umin.
21313 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
21314 (FMAXMIN): Rename as...
21315 (FMAXMIN_UNS): ...This.
21317 (fmaxminv): Likewise.
21318 (fmaxmin): Likewise.
21320 (maxmin_uns_op): Likewise.
21322 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21324 * config/aarch64/arm_neon.h
21325 (vac<ge, gt><sd>_f<32, 64>): Rename to...
21326 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
21327 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
21329 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21331 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
21332 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
21334 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21336 * config/aarch64/aarch64-simd.md
21337 (vcond<mode>_internal): Handle special cases for constant masks.
21338 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
21339 (vcondu<mode><mode>): Likewise.
21340 (vcond<v_cmp_result><mode>): New.
21342 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21344 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
21345 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
21346 * config/aarch64/aarch64-simd-builtins.def
21347 (cmeq): Update to BUILTIN_VALLDI.
21352 * config/aarch64/arm_neon.h
21353 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
21354 to builtins or C as appropriate.
21356 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21358 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21360 (cmhi): Rename to...
21362 * config/aarch64/aarch64-simd.md
21363 (simd_mode): Add SF.
21364 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21365 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21366 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21367 (cstore<mode>_neg): ...This.
21368 * config/aarch64/iterators.md
21370 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21371 (COMPARISONS): New.
21372 (UCOMPARISONS): Likewise.
21373 (optab): Add missing comparisons.
21379 (VCMP_S): Likewise.
21380 (VCMP_U): Likewise.
21381 (V_cmp_result): Add DF, SF modes.
21382 (v_cmp_result): Likewise.
21384 (vmtype): Likewise.
21385 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21387 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21389 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21390 define_insn to define_insn_and_split.
21391 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21392 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21393 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21394 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21395 (thumb2_negscc): Likewise.
21397 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21399 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21401 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21403 * config/arm/thumb2.md: Remove trailing whitespaces.
21405 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21407 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21408 Use gen_int_mode rather than GEN_INT.
21410 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21412 * value-prof.c (stream_in_histogram_value): Remove the strayed
21415 2013-04-30 Richard Biener <rguenther@suse.de>
21417 PR middle-end/57122
21418 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21420 2013-04-30 Richard Biener <rguenther@suse.de>
21422 PR middle-end/57107
21423 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21425 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21427 PR rtl-optimization/56957
21428 PR rtl-optimization/57105
21429 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21430 variable. Use just INSN_UID for determining whether an insn
21431 should be only disconnected from the insn stream.
21432 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21434 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21436 PR tree-optimization/57104
21437 * tsan.c (instrument_expr): Don't instrument accesses to
21438 DECL_HARD_REGISTER VAR_DECLs.
21440 2013-04-30 Richard Biener <rguenther@suse.de>
21442 * function.h (loops_for_fn): New inline function.
21443 (set_loops_for_fn): Likewise.
21444 * cfgloop.h (place_new_loop): Add struct function parameter.
21445 (get_loop): Likewise.
21446 (get_loops): Likewise.
21447 (number_of_loops): Likewise.
21448 (fel_next): Adjust.
21449 (fel_init): Likewise.
21450 * cfg.c (get_loop_copy): Adjust.
21451 * cfgloop.c (flow_loops_dump): Likewise.
21452 (record_loop_exits): Likewise.
21453 (verify_loop_structure): Likewise.
21454 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21455 (estimate_reg_pressure_cost): Likewise.
21456 (mark_loop_exit_edges): Likewise.
21457 * cfgloopmanip.c (place_new_loop): Likewise.
21458 (add_loop): Likewise.
21459 (duplicate_loop): Likewise.
21460 * graph.c (draw_cfg_nodes): Likewise.
21461 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21462 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21463 (extract_affine_chrec): Likewise.
21464 (build_scop_iteration_domain): Likewise.
21465 * graphite.c (graphite_initialize): Likewise.
21466 * ira-build.c (create_loop_tree_nodes): Likewise.
21467 (more_one_region_p): Likewise.
21468 (rebuild_regno_allocno_maps): Likewise.
21469 (mark_loops_for_removal): Likewise.
21470 (mark_all_loops_for_removal): Likewise.
21471 (remove_unnecessary_regions): Likewise.
21472 (ira_build): Likewise.
21473 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21474 * loop-init.c (fix_loop_structure): Likewise.
21475 (gate_rtl_move_loop_invariants): Likewise.
21476 (gate_rtl_unswitch): Likewise.
21477 (gate_rtl_unroll_and_peel_loops): Likewise.
21478 (rtl_doloop): Likewise.
21479 * lto-streamer-in.c (input_cfg): Likewise.
21480 * lto-streamer-out.c (output_cfg): Likewise.
21481 * modulo-sched.c (sms_schedule): Likewise.
21482 * predict.c (tree_estimate_probability): Likewise.
21483 (tree_estimate_probability_driver): Likewise.
21484 (estimate_loops): Likewise.
21485 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21486 (move_sese_region_to_fn): Likewise.
21487 (debug_loop_num): Likewise.
21488 * tree-chrec.c (chrec_evaluate): Likewise.
21489 (hide_evolution_in_other_loops_than_loop): Likewise.
21490 (chrec_component_in_loop_num): Likewise.
21491 (reset_evolution_in_loop): Likewise.
21492 (evolution_function_is_invariant_rec_p): Likewise.
21493 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21494 * tree-inline.c (copy_loops): Likewise.
21495 (copy_cfg_body): Likewise.
21496 (tree_function_versioning): Likewise.
21497 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21498 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21500 (add_to_evolution_1): Likewise.
21501 (scev_const_prop): Likewise.
21502 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21503 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21504 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21505 (tree_ssa_lim_initialize): Likewise.
21506 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21507 (verify_loop_closed_ssa): Likewise.
21508 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21509 (tree_ssa_loop_im): Likewise.
21510 (tree_ssa_loop_unswitch): Likewise.
21511 (tree_vectorize): Likewise.
21512 (check_data_deps): Likewise.
21513 (tree_ssa_loop_ivcanon): Likewise.
21514 (tree_ssa_loop_bounds): Likewise.
21515 (tree_complete_unroll): Likewise.
21516 (tree_complete_unroll_inner): Likewise.
21517 (tree_parallelize_loops): Likewise.
21518 (tree_ssa_loop_prefetch): Likewise.
21519 (tree_ssa_loop_ivopts): Likewise.
21520 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21521 * tree-vectorizer.c (vectorize_loops): Likewise.
21523 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21525 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21526 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21527 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21528 with EABI_LINK_SPEC.
21530 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21533 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21536 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21539 * lra-constraints.c (process_alt_operands): Discourage a bit more
21540 using memory for pseudos. Print cost dump for alternatives.
21541 Modify cost values for conflicts with early clobbers.
21542 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21544 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21547 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21549 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21551 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21552 from/to S register.
21553 (movdi_aarch64): Support LDR/STR from/to D register.
21555 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21557 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21558 or higher by default.
21560 2013-04-29 Richard Biener <rguenther@suse.de>
21562 PR middle-end/57075
21563 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21564 even if not adding abnormal edges for calls that can make
21567 2013-04-29 Richard Biener <rguenther@suse.de>
21569 PR middle-end/57103
21570 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21572 (move_stmt_r): Remove redundant checking.
21574 2013-04-29 Teresa Johnson <tejohnson@google.com>
21577 * basic-block.h (apply_scale): New function.
21578 (apply_probability): Use apply_scale.
21579 * gimple-streamer-in.c (input_bb): Ditto.
21580 * lto-streamer-in.c (input_cfg): Ditto.
21581 * lto-cgraph.c (merge_profile_summaries): Ditto.
21582 * tree-optimize.c (execute_fixup_cfg): Ditto.
21583 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21584 (copy_edges_for_bb): Ditto.
21585 (copy_cfg_body): Ditto.
21587 2013-04-29 Tom de Vries <tom@codesourcery.com>
21589 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21590 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21591 (tail_merge_optimize): Handle current_loops == NULL.
21593 2013-04-26 Jeff Law <law@redhat.com>
21595 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21596 (simplify_cond_using_ranges): Generalize code to simplify
21597 COND_EXPRs where one argument is a constant and the other
21598 is an SSA_NAME created by an integral type conversion.
21600 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21602 * config/arm/arm.md (store_minmaxsi): Use only when
21603 optimize_insn_for_size_p.
21605 2013-04-29 Christian Bruel <christian.bruel@st.com>
21608 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21610 2013-04-29 Richard Biener <rguenther@suse.de>
21612 PR middle-end/57089
21613 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21614 loop tree make sure to schedule a fixup for the child as well.
21615 (expand_omp_for_generic): Properly add loops.
21616 (expand_omp_for_static_nochunk): Likewise.
21617 (expand_omp_for_static_chunk): Likewise.
21618 (expand_omp_for): For the degenerate case fixup loops.
21619 (expand_omp_sections): Fix default bb placement in loops.
21620 (expand_omp_atomic_pipeline): Properly add loops.
21622 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21624 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21626 2013-04-29 Tom de Vries <tom@codesourcery.com>
21628 * tree-ssa-tail-merge.c: Update header comment.
21630 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21632 * config/aarch64/arm_neon.h
21633 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21634 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21635 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21636 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21637 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21638 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21639 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21640 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21642 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21644 * config/aarch64/aarch64-simd.md
21645 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21646 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21647 fix_trunc, fixuns_trunc.
21648 (ftrunc<VDQF:mode>2): New.
21649 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21650 (fix_trunc_optab): New.
21652 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21654 * config/aarch64/aarch64-builtins.c
21655 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21656 iceilf, lround, iroundf.
21658 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21661 * config/i386/i386.h (enum ix86_tune_indices)
21662 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21663 New, split from X86_TUNE_INTER_UNIT_MOVES.
21664 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21665 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21666 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21667 (TARGET_INTER_UNIT_MOVES): Remove.
21668 * config/i386/i386.c (initial_ix86_tune_features): Update.
21669 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21670 (ix86_expand_convert_uns_didf_sse): Use
21671 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21672 (ix86_expand_vector_init_one_nonzero): Ditto.
21673 (ix86_expand_vector_init_interleave): Ditto.
21674 (inline_secondary_memory_needed): Return true for moves from SSE class
21675 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21676 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21677 * config/i386/constraints.md (Yi, Ym): Depend on
21678 TARGET_INTER_UNIT_MOVES_TO_VEC.
21679 (Yj, Yn): New constraints.
21680 * config/i386/i386.md (*movdi_internal): Change constraints of
21681 operand 1 from Yi to Yj and from Ym to Yn.
21682 (*movsi_internal): Ditto.
21683 (*movdf_internal): Ditto.
21684 (*movsf_internal): Ditto.
21685 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21686 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21687 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21688 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21689 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21690 * config/i386/sse.md (movdi_to_sse): Ditto.
21691 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21692 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21693 TARGET_INTER_UNIT_MOVES.
21694 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21695 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21696 instead of TARGET_INTER_UNIT_MOVES.
21697 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21698 operand 1 from Yi to Yj and from Ym to Yn.
21700 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21702 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21703 (float_truncate_hi_): Likewise.
21704 (float_extend_lo_): Likewise.
21705 (float_truncate_lo_): Likewise.
21706 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21707 (aarch64_float_extend_lo_v2df): Likewise.
21708 (vec_unpacks_hi_v4sf): Likewise.
21709 (aarch64_float_truncate_lo_v2sf): Likewise.
21710 (aarch64_float_truncate_hi_v4sf): Likewise.
21711 (vec_pack_trunc_v2df): Likewise.
21712 (vec_pack_trunc_df): Likewise.
21714 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21716 * config/aarch64/aarch64-builtins.c
21717 (aarch64_fold_builtin): Fold float conversions.
21718 * config/aarch64/aarch64-simd-builtins.def
21719 (floatv2si, floatv4si, floatv2di): New.
21720 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21721 * config/aarch64/aarch64-simd.md
21722 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21723 * config/aarch64/iterators.md (FLOATUORS): New.
21724 (optab): Add float, floatuns.
21725 (su_optab): Likewise.
21727 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21729 * config/aarch64/aarch64-builtins.c
21730 (aarch64_builtin_vectorized_function): Use new names for
21732 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21733 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21734 (fcvtzu): Split as...
21735 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21736 (fcvtas): Split as...
21737 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21738 (fcvtau): Split as...
21739 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21740 (fcvtps): Split as...
21741 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21742 (fcvtpu): Split as...
21743 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21744 (fcvtms): Split as...
21745 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21746 (fcvtmu): Split as...
21747 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21748 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21749 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21750 (lfrintnusf, lfrintnudf): Likewise.
21751 * config/aarch64/aarch64-simd.md
21752 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21754 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21755 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21756 (fcvt_pattern): Likewise.
21758 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21760 * config/aarch64/aarch64-simd.md
21761 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21762 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21764 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21766 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21767 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21768 (vrnd<a,m,n,p>_f32): Implement using builtins.
21769 (vrnd<i,x><q>_f<32, 64>): New.
21771 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21773 * config/aarch64/aarch64-builtins.c
21774 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21775 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21776 (frintz): Rename to...
21778 (frintp): Rename to...
21780 (frintm): Rename to...
21782 (frinti): Rename to...
21783 (nearbyint): ...this.
21784 (frintx): Rename to...
21786 (frinta): Rename to...
21788 * config/aarch64/aarch64-simd.md
21789 (aarch64_frint<frint_suffix><mode>): Delete.
21790 (<frint_pattern><mode>2): Convert to insn.
21791 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21792 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21793 (frint_pattern): Likewise.
21794 (frint_suffix): Likewise.
21796 2013-04-29 Richard Biener <rguenther@suse.de>
21798 PR tree-optimization/57081
21799 * loop-init.c: Include tree-flow.h.
21800 (loop_optimizer_finalize): Free number of iteration estimates.
21801 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21803 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21805 PR tree-optimization/57083
21806 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21807 non-singleton shift count range, zero extend low_bound for uns case.
21809 * config/i386/predicates.md (general_vector_operand): New predicate.
21810 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21811 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21812 if they aren't nonimmediate operands. If their original values
21813 satisfy const_vector_equal_evenodd_p, don't shift them.
21814 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21815 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21816 if not nonimmediate_operand.
21817 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21818 instead of register_operand.
21819 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21821 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21823 * stor-layout.c (finalize_size_functions): Allocate a structure and
21824 reset cfun before dumping the functions.
21826 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21828 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21831 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21832 use xop_pmacsdqh if uns_p.
21833 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21834 the immediate rotate count.
21836 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21838 * rtl.h (struct rtx_def): Add comment for field jump.
21839 (LRA_SUBREG_P): New macro.
21840 * recog.c (register_operand): Check LRA_SUBREG_P.
21841 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21843 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21844 (lra_final_code_change): Skip subreg change for operators.
21845 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21846 if there are no operand changes.
21847 * lra-constraints.c (curr_insn_set): New.
21848 (match_reload): Set LRA_SUBREG_P.
21849 (emit_spill_move): Ditto.
21850 (check_and_process_move): Use curr_insn_set. Process only single
21851 set insns. Don't initialize sec_mem_p and change_p.
21852 (simplify_operand_subreg): Use LRA_SUBREG_P.
21853 (reg_in_class_p): New function.
21854 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21855 of #ifdef. Add code to remove cycling.
21856 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21857 non-null disp. Reload inner instead of disp when base and index
21858 are null. Try to put lo_sum into register.
21859 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21860 (check_and_process_move): Move code for move cost check to
21861 simple_move_p. Remove equiv_substitution.
21862 (simple_move_p): New function.
21863 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21864 curr_insn_set. Call check_and_process_move only for single set
21865 insns. Use the new function. Move call of check_and_process_move
21866 after operand equiv substitution and address process.
21868 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21871 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21872 with nonlocal goto receivers or returns twice calls, ignore
21873 unininitialized values from abnormal edges to nl goto receiver
21874 or returns twice call.
21876 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21878 PR tree-optimization/57051
21879 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21880 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21883 2013-04-26 Richard Biener <rguenther@suse.de>
21885 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21886 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21887 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21888 (fixup_loop_arrays_after_move): New function.
21889 (move_sese_region_to_fn): Properly outline the loop tree parts
21890 of the SESE region.
21892 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21894 * config/i386/i386.md (type, unit): Fix long lines.
21896 2013-04-26 Richard Biener <rguenther@suse.de>
21898 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21899 (lto-streamer-out.o): Likewise.
21900 * cfgloop.c (init_loops_structure): Export, add struct function
21901 argument and adjust.
21902 (flow_loops_find): Adjust.
21903 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21904 (init_loops_structure): Declare.
21905 * lto-streamer-in.c: Include cfgloop.h.
21906 (input_cfg): Input the loop tree.
21907 * lto-streamer-out.c: Include cfgloop.h.
21908 (output_cfg): Output the loop tree.
21909 (output_struct_function_base): Do not drop PROP_loops.
21911 2013-03-26 Richard Biener <rguenther@suse.de>
21913 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21914 (pass_build_cfg): Provide PROP_loops.
21915 (move_sese_region_to_fn): Remove loops that are outlined into fn
21917 * tree-inline.c: Include cfgloop.h.
21918 (initialize_cfun): Do not drop PROP_loops.
21919 (copy_loops): New function.
21920 (copy_cfg_body): Copy loop structure.
21921 (tree_function_versioning): Initialize destination loop tree.
21922 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21923 (pass_parallelize_loops): Do IL verification.
21924 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21925 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21926 the CFG make sure we fixup loops as well.
21927 * tree-ssa-tail-merge.c: Include cfgloop.h.
21928 (replace_block_by): When merging loop latches mark loops for fixup.
21929 * lto-streamer-out.c (output_struct_function_base): Drop
21930 PROP_loops for now.
21931 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21932 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21933 * ipa-split.c: Include cfgloop.h.
21934 (split_function): Add the new return block to the loop tree root.
21935 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21936 whether we have removed the forwarder block.
21937 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21938 * cfgloop.h (place_new_loop): Declare.
21939 * cfgloopmanip.c (place_new_loop): Export.
21940 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21941 (tree-switch-conversion.o): Likewise.
21942 (tree-complex.o): Likewise.
21943 (tree-inline.o): Likewise.
21944 (tree-ssa-tailmerge.o): Likewise.
21945 (ipa-split.o): Likewise.
21946 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21947 (tree-ssa-copy.o): Likewise.
21948 * tree-switch-conversion.c: Include cfgloop.h
21949 (process_switch): If we emit a bit-test cascade, schedule loops
21951 * tree-complex.c: Include cfgloop.h.
21952 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21953 * asan.c: Include cfgloop.h.
21954 (create_cond_insert_point): Properly add new basic-blocks to
21955 loops, schedule loop fixup.
21956 * cfgloop.c (verify_loop_structure): Check that looks are not
21958 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21960 (expand_omp_for_generic): Likewise.
21961 (expand_omp_sections): Likewise.
21962 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21963 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21964 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21965 is initialized, not when loops are present.
21966 * tree-parloops.c (parallelize_loops): Remove checking here.
21967 * passes.c (init_optimization_passes): Schedule a copy-propagation
21968 pass before complete unrolling of inner loops.
21970 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21972 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21973 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21974 (colorize_init): Add argument to _WIN32 version.
21975 * toplev.c: Include diagnostic-color.h.
21976 (process_options): Default to -fdiagnostics-color=auto if
21977 GCC_COLORS env var is in the environment.
21978 * common.opt (fdiagnostics-color=): Add Var and Init.
21979 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21980 env var is in the environment, the default is auto rather than never.
21982 * diagnostic.h (file_name_as_prefix): Add context argument.
21983 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21984 the string as locus.
21985 * langhooks.c (lhd_print_error_function): Adjust caller.
21987 2013-04-25 Lawrence Crowl <crowl@google.com>
21989 * var-tracking.c (shared_hash_def::htab):
21990 Change type to hash_table. Update dependent calls and types.
21992 2013-04-25 Lawrence Crowl <crowl@google.com>
21994 * Makefile.in: Update as needed below.
21996 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21997 Move declaration to after the type's method definitons.
21999 * attribs.c (htab_t scoped_attributes::attribute_hash):
22000 Change type to hash_table. Update dependent calls and types.
22002 * bitmap.c (htab_t bitmap_desc_hash):
22003 Change type to hash_table. Update dependent calls and types.
22005 * cselib.c (htab_t cselib_hash_table):
22006 Change type to hash_table. Update dependent calls and types.
22008 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22009 (hash_string_slot_node): Move implementation into lto-streamer.h
22010 struct string_slot_hasher.
22011 (eq_string_slot_node): Likewise.
22013 * data-streamer-out.c: Update output_block::string_hash_table
22014 dependent calls and types.
22016 * dwarf2cfi.c (htab_t trace_index):
22017 Change type to hash_table. Update dependent calls and types.
22019 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22020 Change type to hash_table. Update dependent calls and types.
22021 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22022 (htab_t optimize_external_refs::map): Likewise.
22023 (htab_t output_comp_unit::extern_map): Likewise.
22024 (htab_t output_comdat_type_unit::extern_map): Likewise.
22025 (htab_t output_macinfo::macinfo_htab): Likewise.
22026 (htab_t optimize_location_lists::htab): Likewise.
22027 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22029 * except.c (htab_t ehspec_hash_type):
22030 Change type to hash_table. Update dependent calls and types.
22031 (assign_filter_values::ttypes): Likewise.
22032 (assign_filter_values::ehspec): Likewise.
22033 (sjlj_assign_call_site_values::ar_hash): Likewise.
22034 (convert_to_eh_region_ranges::ar_hash): Likewise.
22036 * gcse.c (htab_t pre_ldst_table):
22037 Change type to hash_table. Update dependent calls and types.
22039 * ggc-common.c (htab_t saving_htab):
22040 Change type to hash_table. Update dependent calls and types.
22041 (htab_t loc_hash): Likewise.
22042 (htab_t ptr_hash): Likewise.
22043 (call_count): Rename ggc_call_count.
22044 (call_alloc): Rename ggc_call_alloc.
22045 (loc_descriptor): Rename make_loc_descriptor.
22046 (add_statistics): Rename ggc_add_statistics.
22048 * ggc-common.c (saving_htab):
22049 Change type to hash_table. Update dependent calls and types.
22051 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22052 (push_gimplify_context): Likewise.
22053 (pop_gimplify_context): Likewise.
22054 (struct gimple_temp_hash_elt): Added.
22055 (struct gimplify_hasher): Likewise.
22056 (struct gimplify_ctx.temp_htab):
22057 Change type to hash_table. Update dependent calls and types.
22059 * gimple-fold.c: Include gimplify-ctx.h.
22061 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22062 Change type to hash_table. Update dependent calls and types.
22063 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22064 avoid potential global name collision.
22066 * gimplify.c: Include gimplify-ctx.h.
22067 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22068 (htab_t gimplify_ctx::temp_htab):
22069 Update dependent calls and types for new type hash_table.
22070 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22071 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22073 * gimplify-ctx.h: New.
22074 (struct gimple_temp_hash_elt): Move from gimplify.c.
22075 (class gimplify_hasher): New.
22076 (struct gimplify_ctx): Move from gimple.h.
22077 (htab_t gimplify_ctx::temp_htab):
22078 Change type to hash_table. Update dependent calls and types.
22080 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22081 (htab_t ivs_params::newivs_index):
22082 Change type to hash_table. Update dependent calls and types.
22083 (htab_t ivs_params::params_index): Likewise.
22084 (htab_t print_generated_program::params_index): Likewise.
22085 (htab_t gloog::newivs_index): Likewise.
22086 (htab_t gloog::params_index): Likewise.
22088 * graphite.c: Include graphite-htab.h.
22089 4htab_t graphite_transform_loops::bb_pbb_mapping):
22090 Change type to hash_table. Update dependent calls and types.
22092 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22093 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22094 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22096 * graphite-dependences.c: Include graphite-htab.h.
22097 (loop_is_parallel_p): Change hash table type of parameter.
22099 * graphite-htab.h: New.
22100 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22101 (extern find_pbb_via_hash): Move from graphite-poly.h.
22102 (extern loop_is_parallel_p): Move from graphite-poly.h.
22103 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22105 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22106 (extern loop_is_parallel_p): Move to graphite-htab.h.
22107 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22109 * haifa-sched.c (htab_t delay_htab):
22110 Change type to hash_table. Update dependent calls and types.
22111 (htab_t delay_htab_i2): Likewise.
22113 * ira-color.c (htab_t allocno_hard_regs_htab):
22114 Change type to hash_table. Update dependent calls and types.
22116 * ira-costs.c (htab_t cost_classes_htab):
22117 Change type to hash_table. Update dependent calls and types.
22119 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22120 Change type to hash_table. Update dependent calls and types.
22122 * loop-iv.c (htab_t bivs):
22123 Change type to hash_table. Update dependent calls and types.
22125 * loop-unroll.c (htab_t opt_info::insns_to_split):
22126 Change type to hash_table. Update dependent calls and types.
22127 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22129 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22130 (struct string_slot_hasher): New.
22131 (htab_t output_block::string_hash_table):
22132 Change type to hash_table. Update dependent calls and types.
22134 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22135 (htab_t file_name_hash_table):
22136 Change type to hash_table. Update dependent calls and types.
22138 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22141 * lto-streamer.c (htab_t tree_htab):
22142 Change type to hash_table. Update dependent calls and types.
22144 * omp-low.c: Include gimplify-ctx.h.
22146 * passes.c (htab_t name_to_pass_map):
22147 Change type to hash_table. Update dependent calls and types.
22148 (pass_traverse): Rename to passes_pass_traverse.
22150 * plugin.c (htab_t event_tab):
22151 Change type to hash_table. Update dependent calls and types.
22153 * postreload-gcse.c (htab_t expr_table):
22154 Change type to hash_table. Update dependent calls and types.
22155 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22157 * sese.c (debug_rename_map_1): Make extern.
22158 (htab_t copy_bb_and_scalar_dependences::rename_map):
22159 Change type to hash_table. Update dependent calls and types.
22161 * sese.h (extern debug_rename_map): Move to .c file.
22163 * store-motion.c (htab_t store_motion_mems_table):
22164 Change type to hash_table. Update dependent calls and types.
22166 * trans-mem.c (htab_t tm_new_mem_hash):
22167 Change type to hash_table. Update dependent calls and types.
22169 * tree-browser.c (htab_t TB_up_ht):
22170 Change type to hash_table. Update dependent calls and types.
22172 * tree-cfg.c (htab_t discriminator_per_locus):
22173 Change type to hash_table. Update dependent calls and types.
22175 * tree-complex.c: Include tree-hasher.h
22176 (htab_t complex_variable_components):
22177 Change type to hash_table. Update dependent calls and types.
22179 * tree-eh.c (htab_t finally_tree):
22180 Change type to hash_table. Update dependent calls and types.
22182 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22183 struct int_tree_hasher.
22184 (extern int_tree_map_eq): Likewise.
22185 (uid_decl_map_hash): Removed.
22186 (extern decl_tree_map_eq): Likewise.
22188 * tree-hasher.h: New.
22189 (struct int_tree_hasher): New.
22190 (typedef int_tree_htab_type): New.
22192 * tree-inline.c: Include gimplify-ctx.h.
22194 * tree-mudflap.c: Include gimplify-ctx.h.
22196 * tree-parloops.c: Include tree-hasher.h.
22197 (htab_t eliminate_local_variables_stmt::decl_address):
22198 Change type to hash_table. Update dependent calls and types.
22199 (htab_t separate_decls_in_region::decl_copies): Likewise.
22201 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22202 Change type to hash_table. Update dependent calls and types.
22204 * tree-sra.c (candidates):
22205 Change type to hash_table. Update dependent calls and types.
22207 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22209 (int_tree_map_hash): Likewise.
22211 * tree-ssa-dom.c (htab_t avail_exprs):
22212 Change type to hash_table. Update dependent calls and types.
22214 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22215 Change type to hash_table. Update dependent calls and types.
22217 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22218 Change type to hash_table. Update dependent calls and types.
22220 * tree-ssa-phiopt.c (seen_ssa_names):
22221 Change type to hash_table. Update dependent calls and types.
22223 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22224 Change type to hash_table. Update dependent calls and types.
22226 * tree-ssa-uncprop.c (equiv):
22227 Change type to hash_table. Update dependent calls and types.
22229 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22231 PR rtl-optimization/57003
22232 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22233 call note_stores with kill_clobbered_value callback again after
22234 killing regs_invalidated_by_call.
22236 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22238 * config/aarch64/aarch64-simd.md
22239 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22240 (aarch64_simd_bsl<mode>): Likewise.
22241 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22243 2013-04-25 Marek Polacek <polacek@redhat.com>
22245 PR tree-optimization/57066
22246 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22248 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22250 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22252 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22254 * config/aarch64/aarch64-builtins.c
22255 (aarch64_fold_builtin): New.
22256 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22257 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22258 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22259 * config/aarch64/arm_neon.h
22260 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22262 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22263 Tejas Belagod <tejas.belagod@arm.com>
22265 * config/aarch64/aarch64-builtins.c
22266 (aarch64_gimple_fold_builtin): New.
22267 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22268 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22269 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22270 (addvv4sf): Update.
22271 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22273 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22275 * config/aarch64/aarch64.md
22276 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22278 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22280 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22281 (*ngcsi_uxtw): New pattern.
22283 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22284 Julian Brown <julian@codesourcery.com>
22286 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22287 (TB_DREG): Add T_V4HF.
22288 (v4hf_UP): New macro.
22289 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22290 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22291 Handle initialisation of V4HF. Adjust initialisation of reinterpret
22293 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
22294 (arm_vector_mode_supported_p): Handle V4HF.
22295 (arm_mangle_map): Handle V4HFmode.
22296 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
22297 * config/arm/arm_neon_builtins.def: Add entries for
22298 vcvtv4hfv4sf, vcvtv4sfv4hf.
22299 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
22300 (neon_vcvtv4hfv4sf): Likewise.
22301 * config/arm/neon-gen.ml: Handle half-precision floating point
22303 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
22304 * config/arm/arm_neon.h: Regenerate.
22305 * config/arm/neon.ml (type elts): Add F16.
22306 (type vectype): Add T_float16x4, T_floatHF.
22307 (type vecmode): Add V4HF.
22308 (type features): Add Requires_FP_bit feature.
22309 (elt_width): Handle F16.
22310 (elt_class): Likewise.
22311 (elt_of_class_width): Likewise.
22312 (mode_of_elt): Refactor.
22313 (type_for_elt): Handle F16, fix error messages.
22314 (vectype_size): Handle T_float16x4.
22315 (vcvt_sh): New function.
22316 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
22317 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
22318 (string_of_mode): Handle V4HF.
22319 * doc/arm-neon-intrinsics.texi: Regenerate.
22321 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22323 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
22324 format specifier in 'X' case.
22326 2013-04-25 Alan Modra <amodra@gmail.com>
22329 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
22330 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
22331 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
22332 Repeat for many other rotate/shift and mask patterns using subregs.
22333 Name lshiftrt insns.
22334 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
22335 on WORDS_BIG_ENDIAN.
22337 2013-04-25 Alan Modra <amodra@gmail.com>
22339 * config.gcc: Support little-endian powerpc-linux targets.
22340 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
22341 (LINK_OS_LINUX_SPEC): Define.
22342 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
22343 Preserve MASK_LITTLE_ENDIAN.
22344 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
22345 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
22346 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
22347 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
22348 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
22349 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
22350 Correct fp word order for little-endian. Don't shift toc entries
22351 smaller than a word for little-endian.
22352 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
22353 (bswapdi2 splits): Correct low-part subreg for little-endian.
22354 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
22355 low/high where such is correct only for be.
22356 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
22357 little-endian for -mcall-aixdesc.
22359 2013-04-25 Alan Modra <amodra@gmail.com>
22361 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22362 replace_equiv_address_nv.
22364 2013-04-25 Alan Modra <amodra@gmail.com>
22366 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22368 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22371 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22372 * rtl.h (struct rtx_def): ...
22374 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22376 PR rtl-optimizations/57046
22377 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22378 for multi-reg splits.
22380 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22382 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22384 2013-04-24 Sterling Augustine <saugustine@google.com>
22386 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22387 (comp_dir_string, debug_str_dwo_section): New.
22388 (DEBUG_STR_DWO_SECTION): Rename to ...
22389 (DEBUG_DWO_STR_SECTION): ... this.
22390 (DEBUG_NORM_STR_SECTION): Delete.
22391 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22392 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22393 (find_AT_string): Move most logic to ...
22394 (find_AT_string_in_table): ... here. New.
22395 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22396 add_skeleton_AT_string. Delete logic.
22397 (output_skeleton_debug_sections): Remove call to
22398 add_top_level_skeleton_die_attrs.
22399 (add_comp_dir_attribute): Move logic to comp_dir_string.
22400 (dwarf2out_init): Initialize debug_str_dwo_section.
22401 (output_indirect_string): Call find_string_form.
22402 (output_indirect_strings): Rewrite.
22403 (prune_unused_types): Empty skeleton_debug_str_hash.
22404 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22405 (dwarf2out_finish): Call output_indirect_strings.
22407 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22409 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22411 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22413 * rtl.h (struct rtx_def): Add comment for field jump.
22414 (LRA_SUBREG_P): New macro.
22415 * recog.c (register_operand): Check LRA_SUBREG_P.
22416 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22418 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22419 (lra_final_code_change): Skip subreg change for operators.
22420 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22421 if there are no operand changes.
22422 * lra-constraints.c (curr_insn_set): New.
22423 (match_reload): Set LRA_SUBREG_P.
22424 (emit_spill_move): Ditto.
22425 (check_and_process_move): Use curr_insn_set. Process only single
22426 set insns. Don't initialize sec_mem_p and change_p.
22427 (simplify_operand_subreg): Use LRA_SUBREG_P.
22428 (reg_in_class_p): New function.
22429 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22430 of #ifdef. Add code to remove cycling.
22431 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22432 non-null disp. Reload inner instead of disp when base and index
22433 are null. Try to put lo_sum into register.
22434 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22435 (check_and_process_move): Move code for move cost check to
22436 simple_move_p. Remove equiv_substitution.
22437 (simple_move_p): New function.
22438 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22439 curr_insn_set. Call check_and_process_move only for single set
22440 insns. Use the new function. Move call of check_and_process_move
22441 after operand equiv substitution and address process.
22443 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22445 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22446 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22447 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22449 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22451 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22453 2013-04-24 Marek Polacek <polacek@redhat.com>
22455 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22456 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22457 (select_loops_exit_conditions): Likewise.
22458 (number_of_iterations_for_all_loops): Likewise.
22459 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22460 (scev_analysis): Likewise.
22462 2013-04-02 Catherine Moore <clm@codesourcery.com>
22463 Chao-ying Fu <fu@mips.com>
22465 * config/mips/micromips.md (jraddiusp): New pattern.
22466 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22467 instruction if possible.
22469 2013-04-24 Alan Modra <amodra@gmail.com>
22471 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22473 2013-04-24 Julian Brown <julian@codesourcery.com>
22474 Chung-Lin Tang <cltang@codesourcery.com>
22476 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22477 dependency behavior in enumeration type DIE generation. Add TODO note
22478 to comments about future DW_FORM_sdata/udata re-work of related code.
22480 2013-04-23 Lawrence Crowl <crowl@google.com>
22482 * Makefile.in: Update as needed below.
22484 * hash-table.h (class hash_table):
22485 Correct many methods with parameter types compare_type to the correct
22486 value_type. (Correct code was unlikely to notice the change.)
22487 (hash_table::elements_with_deleted) New.
22488 (class hashtable::iterator): New.
22489 (hashtable::begin()): New.
22490 (hashtable::end()): New.
22491 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22493 * statistics.c (statistics_hashes):
22494 Change type to hash_table. Update dependent calls and types.
22496 * tree-into-ssa.c (var_infos):
22497 Change type to hash_table. Update dependent calls and types.
22499 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22500 Change type to hash_table. Update dependent calls and types.
22502 * tree-ssa-loop-im.c (struct mem_ref.refs):
22503 Change type to hash_table. Update dependent calls and types.
22505 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22506 Change type to hash_table. Update dependent calls and types.
22508 * tree-ssa-sccvn.c (vn_tables_s::nary):
22509 Change type to hash_table. Update dependent calls and types.
22510 (vn_tables_s::phis): Likewise.
22511 (vn_tables_s::references): Likewise.
22513 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22514 (vn_reference_eq): Update parameter and return types.
22516 * tree-ssa-structalias.c (pointer_equiv_class_table):
22517 Change type to hash_table. Update dependent calls and types.
22518 (location_equiv_class_table): Likewise.
22520 * tree-vect-data-refs.c: Consequential changes for making
22521 peeling a hash_table.
22523 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22524 (destroy_loop_vec_info): Dependent hash_table update.
22526 * tree-vectorizer.h (peeling_htab):
22527 Change type to hash_table. Update dependent calls and types.
22529 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22531 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22532 to check the register content is equal or not.
22533 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22534 to assign register content record.
22535 * lra-eliminations.c (update_reg_eliminate): Use
22536 lra_update_reg_val_offset to update register content offset.
22537 * lra-int.h (struct lra_reg): Add offset member.
22538 (lra_reg_val_equal_p): New static inline function.
22539 (lra_update_reg_val_offset): New static inline function.
22540 (lra_assign_reg_val): New static inline function.
22541 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22542 to assign register content record.
22543 (initialize_lra_reg_info_element): Initial offset to zero.
22545 2013-04-23 Catherine Moore <clm@codesourcery.com>
22547 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22548 operands. Record compression.
22550 2013-04-23 Xinliang David Li <davidxl@google.com>
22552 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22554 2013-04-23 Richard Biener <rguenther@suse.de>
22556 PR middle-end/57036
22557 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22558 parameter, only add abnormal goto edges from the copied body
22559 if the call could perform abnormal gotos.
22560 (copy_cfg_body): Adjust.
22562 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22564 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22566 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22568 * coretypes.h (gimple_stmt_iterator): Add struct to make
22571 2013-04-23 Richard Biener <rguenther@suse.de>
22573 PR tree-optimization/57026
22574 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22575 from SSA names occuring in abnormal PHI nodes.
22577 2013-04-22 Andi Kleen <ak@linux.intel.com>
22579 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22581 2013-04-22 Andi Kleen <ak@linux.intel.com>
22583 * lto/lto.c (print_lto_report_1): Declare early.
22584 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22586 2013-04-22 Andi Kleen <ak@linux.intel.com>
22588 * common.opt (-flto-report-wpa): Add.
22589 * doc/invoke.texi (-flto-report-wpa): Add.
22590 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22593 2013-04-22 Xinliang David Li <davidxl@google.com>
22595 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22596 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22597 * Makefile.in: New dependency
22599 David Daney <ddaney.cavm@gmail.com>
22601 * configure.ac (gcc_cv_as_micromips_support): Use the
22602 --fatal-warnings option.
22603 * configure: Regenerate.
22605 2013-04-22 Marek Polacek <polacek@redhat.com>
22608 * tsan.c (instrument_expr): Don't instrument expression
22609 in case its size is zero.
22611 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22615 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22617 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22619 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22621 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22622 (gimple_stmt_iterator): New typedef.
22623 * gimple.h (gimple_stmt_iterator): Rename to...
22624 (gimple_stmt_iterator_d): ... This.
22625 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22626 trees be valid for GIMPLE and GENERIC.
22627 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22628 * gimple-fold.c (gimple_fold_call): Call target hook
22629 gimple_fold_builtin.
22630 * hooks.c (hook_bool_gsiptr_false): New.
22631 * hooks.h (hook_bool_gsiptr_false): New.
22632 * target.def (fold_stmt): New.
22633 * doc/tm.texi: Regenerate.
22635 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22638 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22639 a set sp if no stack realignment.
22641 2013-04-22 Nick Clifton <nickc@redhat.com>
22643 * config.gcc (tilegx-linux): Extend extra_objs rather than
22645 (tilepro-linux): Likewise.
22647 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22649 * config/aarch64/aarch64-builtins.c
22651 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22652 (VAR<1-12>): Add MAP parameter.
22653 (BUILTIN_*): Likewise.
22654 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22655 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22656 (aarch64_ushl_n<mode>): Likewise.
22657 (aarch64_sshr_n<mode>): Likewise.
22658 (aarch64_ushr_n<mode>): Likewise.
22659 (aarch64_<maxmin><mode>): Likewise.
22660 (aarch64_sqrt<mode>): Likewise.
22661 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22662 (vshr<q>_n_*): Likewise.
22664 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22666 * config/aarch64/aarch64-builtins.c
22667 (aarch64_simd_builtin_type_mode): Handle SF types.
22669 (BUILTIN_GPF): Define.
22670 (aarch64_init_simd_builtins): Handle SF types.
22671 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22672 (frecps): Likewise.
22673 (frecpx): Likewise.
22674 * config/aarch64/aarch64-simd.md
22675 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22676 (aarch64_frecpe<mode>): New.
22677 (aarch64_frecps<mode>): Likewise.
22678 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22679 (v8type): Add frecp<esx>.
22680 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22681 (aarch64_frecps<mode>): Likewise.
22682 * config/aarch64/iterators.md (FRECP): New.
22683 (frecp_suffix): Likewise.
22684 * config/aarch64/arm_neon.h
22685 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22687 2013-04-22 Christian Bruel <christian.bruel@st.com>
22690 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22691 (REG_CLASS_NAMES): Idem.
22692 (REG_CLASS_CONTENTS): Idem.
22693 (REGCLASS_HAS_FP_REG): Idem.
22694 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22695 (sh_conditional_register_usage): Idem.
22697 2013-04-21 Jeff Law <law@redhat.com>
22699 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22700 (ssa_forward_propagate_and_combine): Use it.
22702 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22704 * lra.c: Update the flow chart diagram.
22706 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22708 PR rtl-optimization/56847
22709 * lra-constraints.c (process_alt_operands): Discourage alternative
22710 with non-matche doffsettable memory constraint fro memory with
22713 2013-04-19 Richard Biener <rguenther@suse.de>
22715 PR tree-optimization/56982
22716 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22718 * gimplify.c (gimplify_call_expr): Notice special calls.
22719 (gimplify_modify_expr): Likewise.
22720 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22721 abnormal control flow receivers.
22722 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22723 in the same way as cfun->has_nonlocal_labels.
22724 (gimple_purge_dead_abnormal_call_edges): Likewise.
22725 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22726 receivers start a basic-block.
22728 2013-04-19 Richard Biener <rguenther@suse.de>
22730 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22732 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22733 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22734 (SLP_TREE_LOAD_PERMUTATION): Add.
22735 (vect_transform_slp_perm_load): Adjust prototype.
22736 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22737 (vect_free_slp_instance): Likewise.
22738 (vect_create_new_slp_node): Likewise.
22739 (vect_supported_slp_permutation_p): Remove.
22740 (vect_slp_rearrange_stmts): Adjust.
22741 (vect_supported_load_permutation_p): Likewise. Inline
22742 vect_supported_slp_permutation_p here.
22743 (vect_analyze_slp_instance): Compute load permutations per
22744 slp node instead of per instance.
22745 (vect_get_slp_defs): Adjust.
22746 (vect_transform_slp_perm_load): Likewise.
22747 (vect_schedule_slp_instance): Remove redundant code.
22748 (vect_schedule_slp): Remove hack for PR56270, add it ...
22749 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22750 CSE loads for SLP. Adjust.
22752 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22754 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22755 spelling in two comments.
22757 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22760 * config/arm/arm.c (load_multiple_sequence): Require SP
22761 as base register for loads if SP is in the register list.
22763 2013-04-19 Martin Jambor <mjambor@suse.cz>
22765 PR tree-optimization/56718
22766 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22767 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22768 and made public. Adjusted all callers.
22769 (ipa_intraprocedural_devirtualization): New function.
22770 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22771 (ipa_intraprocedural_devirtualization): Likewise.
22772 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22774 2013-04-19 Richard Biener <rguenther@suse.de>
22776 PR tree-optimization/57000
22777 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22779 2013-04-19 Terry Guo <terry.guo@arm.com>
22781 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22783 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22784 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22785 (cortex_m4_fmacs): Use new reservations.
22786 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22788 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22790 PR rtl-optimization/56999
22791 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22793 (lra_coalesce): Remove split_origin_bitmap and related code.
22794 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22795 ranges if necessary.
22797 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22799 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22801 (ix86_expand_call): Remove clobbered_registers array and use
22802 x86_64_ms_sysv_extra_clobbered_registers instead.
22803 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22805 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22807 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22808 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22809 (*call_value_rex64_ms_sysv): Ditto.
22811 2013-04-18 Cary Coutant <ccoutant@google.com>
22813 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22814 parent instead of die_mark.
22816 2013-04-18 Diego Novillo <dnovillo@google.com>
22818 * gimple.c (create_gimple_tmp): New.
22819 (get_expr_type): New.
22820 (build_assign): New.
22821 (build_type_cast): New.
22822 * gimple.h (enum ssa_mode): Define.
22823 (gimple_seq_set_location): New.
22824 * asan.c (build_check_stmt): Change some gimple_build_* calls
22825 to use build_assign and build_type_cast.
22827 2013-04-18 Richard Biener <rguenther@suse.de>
22829 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22830 handle negative step. Remove redundant checks.
22831 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22832 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22833 for negative step and grouped loads fail to vectorize.
22835 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22837 * emit-rtl.c (reset_insn_used_flags): New function.
22838 (reset_all_used_flags): Use it.
22839 (verify_insn_sharing): New function.
22840 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22842 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22844 PR tree-optimization/56984
22845 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22846 and (x >> M) >= N don't register any assertion if N << M is the
22849 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22851 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22852 do not delete_insn insns that have not yet been emitted, only
22853 unlink them with remove_insn.
22854 * df-scan.c (df_insn_delete): Revert r197492.
22856 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22858 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22859 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22861 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22863 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22864 define_insn_and_split.
22865 (and_scc,ior_scc,negscc): Likewise.
22866 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22868 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22870 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22871 can benefit from using a sequence of LDRD instructions in epilogue
22872 instead of a single LDM instruction.
22874 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22877 * doc/extend.texi: Fix typo.
22879 2013-04-17 Richard Biener <rguenther@suse.de>
22881 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22882 (vect_build_slp_tree): ... here.
22883 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22884 match. Remove special-casing of mismatched complex loads.
22885 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22886 re-try the match with swapped commutative operands.
22887 (vect_supported_load_permutation_p): Remove special-casing of
22888 mismatched complex loads.
22889 (vect_analyze_slp_instance): Adjust.
22891 2013-04-17 Richard Biener <rguenther@suse.de>
22893 PR rtl-optimization/56921
22894 * cfgloop.h (struct loop): Add simple_loop_desc member.
22895 (struct niter_desc): Mark with GTY(()).
22896 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22897 * loop-iv.c (get_simple_loop_desc): Likewise.
22898 (free_simple_loop_desc): Likewise.
22901 2013-04-16 Richard Biener <rguenther@suse.de>
22903 PR rtl-optimization/56921
22904 * loop-init.c (pass_rtl_move_loop_invariants): Add
22905 TODO_do_not_ggc_collect to todo_flags_finish.
22906 (pass_rtl_unswitch): Same.
22907 (pass_rtl_unroll_and_peel_loops): Same.
22908 (pass_rtl_doloop): Same.
22910 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22912 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22913 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22914 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22916 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22917 * tree-streamer.c (record_common_node): Adjust reference in comment.
22919 2013-04-17 Terry Guo <terry.guo@arm.com>
22921 * config/arm/cortex-m4.md: Add a new bypass.
22923 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22925 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22927 (*subs_<optab><mode>_multp2): New pattern.
22928 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22929 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22931 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22933 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22934 (*subs_mul_imm_<mode>): New pattern.
22936 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22939 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22940 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22941 (vsx_movti_32bit): Change j->wa to O->wa.
22943 2013-04-16 Richard Biener <rguenther@suse.de>
22945 PR rtl-optimization/56921
22946 * loop-init.c (pass_rtl_move_loop_invariants): Add
22947 TODO_do_not_ggc_collect to todo_flags_finish.
22948 (pass_rtl_unswitch): Same.
22949 (pass_rtl_unroll_and_peel_loops): Same.
22950 (pass_rtl_doloop): Same.
22952 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22954 * config/arm/arm.c (emit_multi_reg_push): New declaration
22955 for an existing function.
22956 (arm_emit_strd_push): New function.
22957 (arm_expand_prologue): Used here.
22958 (arm_emit_ldrd_pop): New function.
22959 (arm_expand_epilogue): Used here.
22960 (arm_get_frame_offsets): Update condition.
22961 (arm_emit_multi_reg_pop): Add a special case for load of a single
22962 register with writeback.
22964 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22966 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22969 2013-04-16 Richard Biener <rguenther@suse.de>
22971 PR tree-optimization/56756
22972 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22973 (first_mem_ref_loc): New.
22974 (execute_sm): Place the load temporarily before a previous
22975 access instead of in the latch edge to ensure its SSA dependencies
22976 are defined at points dominating the load.
22978 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22980 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22981 correct fix by moving header and footer insn to the footer of
22982 the merged basic block. Clear BB_END of the merged-away block.
22984 PR middle-end/43631
22985 * emit-rtl.c (make_note_raw): New function.
22986 (link_insn_into_chain): New static inline function.
22987 (add_insn): Use it.
22988 (add_insn_before, add_insn_after): Factor insn chain linking code...
22989 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22990 using link_insn_into_chain.
22991 (note_outside_basic_block_p): New helper function for emit_note_after
22992 and emit_note_before.
22993 (emit_note_after): Use nobb variant of add_insn_after if the note
22994 should not be contained in a basic block.
22995 (emit_note_before): Use nobb variant of add_insn_before if the note
22996 should not be contained in a basic block.
22997 (emit_note_copy): Use make_note_raw.
22998 (emit_note): Likewise.
22999 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23000 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23001 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23002 the moved barrier the tail of the basic block it follows.
23003 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23005 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23007 PR tree-optimization/56962
23008 * gimple-ssa-strength-reduction.c (record_increment): Only set
23009 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23010 either rhs1 or rhs2 is equal to c->base_expr.
23012 2013-04-15 Richard Biener <rguenther@suse.de>
23014 PR tree-optimization/56933
23015 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23017 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23018 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23019 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23020 dependence check ...
23021 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23023 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23024 GROUP_READ_WRITE_DEPENDENCE.
23026 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23028 * emit-rtl.c (reset_all_used_flags): New function.
23029 (verify_rtl_sharing): Call reset_all_used_flags before and after
23030 performing the checks.
23032 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23034 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23035 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23036 * config/arm/constraints.md (De): New constraint.
23037 * config/arm/neon.md (anddi3_neon): Delete.
23038 (neon_vand<mode>): Expand to standard anddi3 pattern.
23039 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23040 Move earlier in the file.
23041 (neon_inv_logic_op2): Likewise.
23042 (arm_anddi_operand_neon): New predicate.
23044 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23046 * configure.ac (gcc_cv_ld_as_needed): Set
23047 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23048 Use -z ignore, -z record on *-*-solaris2*.
23049 (HAVE_LD_AS_NEEDED): Update comment.
23050 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23051 * configure: Regenerate.
23052 * config.in: Regenerate.
23053 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23054 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23055 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23056 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23057 equivalents. Fix markup.
23058 * doc/tm.texi: Regenerate.
23060 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23062 * config/i386/i386.opt: New option mstack-protector-guard=.
23063 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23064 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23065 TARGET_SSP_TLS_GUARD.
23066 * config/i386/i386.c (ix86_option_override_internal): Set
23067 ix86_stack_protector_guard.
23068 * config/i386/i386.md (stack_protect_set): Enable for
23069 TARGET_SSP_TLS_GUARD only.
23070 (stack_protect_set_<mode>): Ditto.
23071 (stack_protect_test): Ditto.
23072 (stack_protect_test_<mode>): Ditto.
23073 * doc/invoke.texi (i386 Option): Document.
23075 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23078 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23079 (S_MODES): Set H_MODE bit.
23080 (SF_MODES): Set only S_MODE and SF_MODE bits.
23081 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23082 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23083 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23084 <MODE_FLOAT>: Likewise.
23086 2013-04-15 Joey Ye <joey.ye@arm.com>
23088 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23090 2013-04-15 Joey Ye <joey.ye@arm.com>
23092 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23094 (thumb_far_jump_used_p): Count instruction size and set
23097 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23099 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23100 * resource.c (mark_target_live_regs): Reformat conditional block.
23102 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23104 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23105 notes, they are emitted only just before final.
23106 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23108 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23110 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23111 * cfgrtl.c (delete_insn): Call it here instead.
23112 * lra-spills.c (lra_final_code_change): Use delete_insn.
23113 * haifa-sched.c (sched_remove_insn): Likewise.
23114 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23115 returning to the nop pool.
23116 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23117 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23119 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23121 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23122 * doc/tm.texi: Regenerated.
23124 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23126 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23129 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23131 * df-core.c (df_find_def): Compare register numbers.
23132 (df_find_use): Likewise.
23134 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23137 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23138 lra_in_progress for return.
23140 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23142 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23143 define_insn into define_insn_and_split and emit movsicc patterns.
23145 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23147 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23149 2013-04-12 Richard Biener <rguenther@suse.de>
23151 * tree-pass.h (TODO_do_not_ggc_collect): New.
23152 * passes.c (execute_one_ipa_transform_pass): Honor
23153 TODO_do_not_ggc_collect.
23154 (execute_one_pass): Likewise.
23157 2013-04-10 Richard Biener <rguenther@suse.de>
23159 * passes.c (init_optimization_passes): Remove reload pass.
23160 * ira.c (do_reload): Merge into ...
23162 (rest_of_handle_reload): Remove.
23163 (pass_reload): Likewise.
23164 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23165 of reload for vzeroupper pass placement.
23167 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23169 PR tree-optimization/56918
23170 PR tree-optimization/56920
23171 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23172 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23173 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23174 use wide_mul_with_sign method.
23176 2013-04-12 Richard Biener <rguenther@suse.de>
23178 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23179 not be considered a gimple constant.
23181 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23183 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23184 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23187 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23188 Jakub Jelinek <jakub@redhat.com>
23190 * opts.c: Include diagnostic-color.h.
23191 (common_handle_option): Handle OPT_fdiagnostics_color_.
23192 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23193 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23194 (diagnostic-color.o): New.
23195 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23196 (diagnostic_color_rule): New enum.
23197 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23198 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23199 the location string.
23200 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23201 either NULL, or color kind.
23202 * diagnostic-color.c: New file.
23203 * diagnostic-color.h: New file.
23204 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23206 * doc/invoke.texi (-fdiagnostics-color): Document.
23207 * pretty-print.h (pp_show_color): Define.
23208 (struct pretty_print_info): Add show_color field.
23209 * diagnostic.c: Include diagnostic-color.h.
23210 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23211 macros. Colorize error:, warning: etc. strings and also the location
23213 (diagnostic_show_locus): Colorize the caret line.
23214 * pretty-print.c: Include diagnostic-color.h.
23215 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23216 inside of %< %> quotes or quoted through q format modifier.
23218 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23220 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23222 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23224 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23225 code in CC_NZ mode.
23226 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23229 2013-04-11 Marek Polacek <polacek@redhat.com>
23231 PR tree-optimization/48184
23232 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23234 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23236 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23237 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23238 (skip_simple_arithmetic): Tidy up.
23239 * tree.h (skip_simple_constant_arithmetic): Declare.
23241 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23243 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23245 2013-04-11 Richard Biener <rguenther@suse.de>
23247 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23248 generate vector constants.
23250 2013-04-11 Richard Biener <rguenther@suse.de>
23252 PR tree-optimization/56878
23253 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23254 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23256 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23257 Prefer to align the DR with the most invariant base address.
23259 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23261 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23264 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23266 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23267 floating-point vector comparisons against 0.
23269 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23271 PR tree-optimization/56899
23272 * fold-const.c (extract_muldiv_1): Apply distributive law
23273 only if TYPE_OVERFLOW_WRAPS (ctype).
23275 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23278 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23279 loading has side effect.
23281 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23283 * config/sparc/sparc.c: Include tree-pass.h.
23284 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23285 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23286 head of file. Change return type. Split off gate function.
23287 (sparc_gate_work_around_errata): New function.
23288 (pass_work_around_errata): New pass definition.
23289 (insert_pass_work_around_errata) New pass insert definition to
23290 insert pass_work_around_errata just after delayed-branch scheduling.
23291 (sparc_option_override): Insert the pass.
23292 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
23294 2013-04-10 David S. Miller <davem@davemloft.net>
23296 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
23297 or -mcpu=hypersparc.
23299 * target.def (cstore_mode): New hook.
23300 * target.h: Include insn-codes.h
23301 * targhooks.c: Likewise.
23302 (default_cstore_mode): New function.
23303 * targhooks.h: Declare it.
23304 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
23305 * doc/tm.texi: Rebuild.
23306 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
23307 target hook, rather than inspecting the insn_data.
23308 * config/sparc/sparc.c (sparc_cstore_mode): New function.
23309 (TARGET_CSTORE_MODE): Redefine.
23310 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
23312 * config/sparc/predicates.md (cstore_result_operand): New special
23314 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
23315 Use it for operand 0.
23316 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
23317 (*snesi_special): Likewise.
23318 (*snesi_zero): Likewise.
23319 (*seqsi_zero): Likewise.
23320 (*sltu_insn): Likewise.
23321 (*sgeu_insn): Likewise.
23322 (*seqdi_special): Make operand 0 and comparison operation be of
23324 (*snedi_special): Likewise.
23325 (*snedi_special_vis3): Likewise.
23326 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
23327 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
23328 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
23329 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
23330 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
23331 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
23332 (*sltu_extend_sp64): Likewise.
23333 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
23334 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
23335 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
23336 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
23337 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
23339 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
23341 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
23342 (aarch64_start_file): Use the new function.
23344 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23345 Jason Merrill <jason@redhat.com>
23347 * common.opt: Add -gdwarf.
23348 * opts.c (common_handle_option): Handle it.
23349 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
23351 2013-04-10 Richard Biener <rguenther@suse.de>
23353 * passes.c (execute_todo): Do not call ggc_collect conditional here.
23354 (execute_one_ipa_transform_pass): But unconditionally here.
23355 (execute_one_pass): And here.
23356 (init_optimization_passes): Remove reload pass.
23357 * tree-pass.h (TODO_ggc_collect): Remove.
23358 (pass_reload): Likewise.
23359 * ira.c (do_reload): Merge into ...
23361 (rest_of_handle_reload): Remove.
23362 (pass_reload): Likewise.
23363 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23364 of reload for vzeroupper pass placement.
23365 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23366 and todo_flags_finish of all passes.
23368 2013-04-10 Richard Biener <rguenther@suse.de>
23370 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23371 first_const_oprnd field, rename first_def_type to first_op_type.
23372 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23373 (vect_get_and_check_slp_defs): Always use the type of the
23374 operand. Allow mixed vect_external_def, vect_constant_def types.
23375 (vect_get_constant_vectors): Handle mixed vect_external_def,
23376 vect_constant_def types.
23378 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23380 PR tree-optimization/55524
23381 * tree-ssa-math-opts.c
23382 (convert_mult_to_fma): Don't use an fms construct
23383 when we don't have an fms operation, but fnma, and it looks
23384 likely that we'll be able to use the latter.
23386 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23388 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23390 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23391 inline fail caused by overwritable functions.
23393 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23395 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23396 unnecessary bits in the constant power of two case.
23398 2013-04-10 Richard Biener <rguenther@suse.de>
23400 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23401 broken code swapping operands.
23402 (vect_build_slp_tree): Do not compute load permutations here.
23403 (vect_analyze_slp_instance): Compute load permutations here,
23404 after building the SLP tree.
23406 2013-04-09 Christian Bruel <christian.bruel@st.com>
23408 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23409 of next/prev_real_insn.
23411 2013-04-09 Jan Hubicka <jh@suse.cz>
23413 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23414 Drop aliased parameter.
23415 (function_and_variable_visibility): Do not handle alias pairs.
23416 * cgraph.c (varpool_externally_visible_p): Update prototype.
23417 * varpool.c (varpool_add_new_variable): Update.
23419 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23421 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23423 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23425 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23427 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23429 2013-04-09 Marek Polacek <polacek@redhat.com>
23431 PR tree-optimization/48762
23432 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23434 2013-04-09 Richard Biener <rguenther@suse.de>
23436 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23438 (vect_build_slp_tree): Likewise.
23439 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23440 calculating the cost of a SLP instance.
23441 (vect_analyze_slp_instance): Use it from here, after building
23444 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23446 PR middle-end/56883
23447 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23448 expand_omp_for_static_chunk): Use simple_p = true in
23449 force_gimple_operand_gsi calls when assigning to addressable decls.
23451 2013-04-09 Jeff Law <law@redhat.com>
23453 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23454 when the boolean was created by converting a wider object which
23455 had a boolean range.
23457 2013-04-09 Richard Biener <rguenther@suse.de>
23459 * tree-vectorizer.h (slp_void_p): Remove.
23460 (slp_tree): Typedef before _slp_tree declaration.
23461 (struct _slp_tree): Use a vector of slp_tree as children.
23462 (vect_get_place_in_interleaving_chain): Remove.
23463 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23465 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23467 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23468 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23469 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23470 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23471 Use slp_node instead of slp_void_p and adjust.
23473 2013-04-09 Richard Biener <rguenther@suse.de>
23475 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23476 work that is not necessary.
23478 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23480 PR tree-optimization/56854
23481 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23482 forward into clobber stmts if it would change MEM_REF lhs into
23485 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23487 * tree.c (type_hash_lookup, type_hash_add): Make static.
23488 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23490 2013-04-09 Richard Biener <rguenther@suse.de>
23492 * tree.h (unsave_expr_now): Remove.
23493 * tree-inline.c (mark_local_for_remap_r): Remove.
23494 (unsave_expr_1): Likewise.
23495 (unsave_r): Likewise.
23496 (unsave_expr_now): Likewise.
23497 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23498 (propagate_tree_value): Likewise.
23500 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23502 * doc/rtl.texi (sequence): Rewrite documentation to match the
23503 current use of SEQUENCE rtl objects.
23504 * rtl.def (SEQUENCE): Likewise.
23506 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23507 Update documentation.
23508 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23509 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23511 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23513 2013-04-08 Teresa Johnson <tejohnson@google.com>
23515 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23516 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23518 (estimate_edge_size_and_time): Add comment to suggest using rounding
23520 (estimate_node_size_and_time): Ditto.
23521 (remap_edge_change_prob): Use helper rounding divide methods.
23522 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23523 (gimple_mod_pow2_value_transform): Ditto.
23524 (gimple_mod_subtract_transform): Ditto.
23525 (gimple_ic_transform): Ditto.
23526 (gimple_stringops_transform): Ditto.
23527 * stmt.c (conditional_probability): Ditto.
23528 (emit_case_dispatch_table): Ditto.
23529 * lto-cgraph.c (merge_profile_summaries): Ditto.
23530 * tree-optimize.c (execute_fixup_cfg): Ditto.
23531 * cfgcleanup.c (try_forward_edges): Ditto.
23532 * cfgloopmanip.c (scale_loop_profile): Ditto.
23534 (duplicate_loop_to_header_edge): Ditto.
23535 (lv_adjust_loop_entry_edge): Ditto.
23536 * tree-vect-loop.c (vect_transform_loop): Ditto.
23537 * profile.c (compute_branch_probabilities): Ditto.
23538 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23539 * lto-streamer-in.c (input_cfg): Ditto.
23540 * gimple-streamer-in.c (input_bb): Ditto.
23541 * ipa-cp.c (update_profiling_info): Ditto.
23542 (update_specialized_profile): Ditto.
23543 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23544 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23546 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23547 (compute_trg_info): Ditto.
23548 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23549 (purge_dead_edges): Ditto.
23550 * loop-unswitch.c (unswitch_loop): Ditto.
23551 * cgraphclones.c (cgraph_clone_edge): Ditto.
23552 (cgraph_clone_node): Ditto.
23553 * tree-inline.c (copy_bb): Ditto.
23554 (copy_edges_for_bb): Ditto.
23555 (initialize_cfun): Ditto.
23556 (copy_cfg_body): Ditto.
23557 (expand_call_inline): Ditto.
23559 2013-04-08 Kai Tietz <ktietz@redhat.com>
23561 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23562 TARGET_CYGWIN64 by TARGET_64BIT.
23564 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23566 * config/epiphany/epiphany.md (GPR_1): New constant.
23567 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23568 * config/epiphany/epiphany.c (gen_compare_reg):
23569 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23570 is already in place.
23571 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23572 Don't require being called during rtl expansion; If y operlaps r0,
23574 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23575 (epiphany_expand_epilogue): Likewise.
23577 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23578 Don't use CC_FPmode for ORDERED / UNORDERED.
23579 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23581 * config/epiphany/constraints.md (CnL): New constraint.
23582 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23583 * config/epiphany/predicates.md (add_operand): Allow 1024.
23585 * config/epiphany/epiphany.md (logical_op): New code iterator.
23586 (op_mnc): New code attribute.
23587 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23588 (mov_f+1, mov_f+2): New peephole2 patterns.
23590 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23591 (cstoresi4): Also allow re-use of zero result when doing a NE
23592 comparison to a non-zero operand.
23593 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23595 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23596 Use gen_addsi3_i / gen_subsi3_i.
23598 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23602 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23603 contain anything but clobbers, at most one __builtin_stack_restore,
23604 optionally debug stmts and final resx, and if it has at least one
23605 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23606 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23607 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23608 which isn't defaut definition, remove them.
23609 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23610 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23611 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23612 with MEM_REF LHS with SSA_NAME address.
23614 2013-04-08 Jeff Law <law@redhat.com>
23616 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23618 2013-04-08 Richard Biener <rguenther@suse.de>
23620 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23622 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23623 determined vector type.
23624 (vect_analyze_data_refs): Likewise.
23625 (vect_get_new_vect_var): Adjust.
23626 (vect_create_destination_var): Preserve SSA name versions.
23627 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23628 not dump anything here.
23630 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23632 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23633 Add member lr_slot_known.
23634 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23636 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23637 Remove code that sets lr_slot_offset according to what a previous
23638 version of epiphany_emit_save_restore used to do.
23639 (epiphany_emit_save_restore): When doing an lr save or restore,
23640 set/verify lr_slot_known and lr_slot_offset.
23642 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23645 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23648 2013-04-08 Richard Biener <rguenther@suse.de>
23650 * alias.c (find_base_term): Fix thinko in previous change.
23652 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23654 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23655 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23656 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23657 if possible to compute val.
23658 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23659 For QImode integers don't require anything about precision. Use
23660 const_with_all_bytes_same to find out if the constant doesn't have
23661 repeated bytes in it.
23663 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23665 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23668 2013-04-08 Marek Polacek <polacek@redhat.com>
23670 PR rtl-optimization/48182
23671 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23674 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23677 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23678 nuses, make sure we have a label.
23680 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23683 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23684 (rs6000_emit_swdiv_low_precision): Remove.
23685 (rs6000_emit_swdiv): Rewrite to handle between one and four
23686 iterations of Newton-Raphson generally; modify required number of
23687 iterations for some cases.
23688 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23690 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23692 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23693 set-but-unused variable.
23695 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23696 basic blocks of released function bodies garbage-collectable.
23698 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23699 (struct rtl_opt_pass): Add TODO_df_finish.
23701 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23703 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23705 * config/arm/constraints.md (q): New constraint.
23706 * config/arm/ldrdstrd.md: New file.
23707 * config/arm/arm.md (ldrdstrd.md) New include.
23708 (arm_movdi): Use "q" instead of "r" constraint
23709 for double-word memory access.
23710 (movdf_soft_insn): Likewise.
23711 * config/arm/vfp.md (movdi_vfp): Likewise.
23712 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23713 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23714 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23715 (mem_ok_for_ldrd_strd): Likewise.
23716 (output_move_double): Update assertion.
23718 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23720 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23722 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23724 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23725 define_insn_and_split.
23726 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23728 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23730 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23731 define_insn_and_split.
23732 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23733 (shiftsi3_compare): New pattern.
23734 (rrx): New pattern.
23735 * config/arm/unspecs.md (UNSPEC_RRX): New.
23737 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23739 * config/arm/arm.md (negdi_extendsidi): New pattern.
23740 (negdi_zero_extendsidi): Likewise.
23742 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23744 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23745 define_insn_and_split.
23746 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23747 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23749 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23751 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23752 define_insn_and_split.
23753 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23754 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23756 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23758 * config/arm/arm.md (subsi3_carryin): New pattern.
23759 (subsi3_carryin_const): Likewise.
23760 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23761 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23763 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23765 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23767 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23769 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23770 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23772 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23774 * config/arm/arm.c (arm_expand_builtin): Change fcode
23775 type to unsigned int.
23777 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23779 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23781 2013-04-04 Ian Lance Taylor <iant@google.com>
23783 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23786 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23788 PR middle-end/56729
23789 * df-scan.c (df_insn_delete): Disable failing assert.
23791 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23793 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23794 New function prototype.
23795 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23796 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23797 (arm_builtin_vectorized_function): New function.
23799 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23801 * config/arm/arm_neon_builtins.def: New file.
23802 * config/arm/arm.c (neon_builtin_data): Move contents to
23803 arm_neon_builtins.def.
23804 (enum arm_builtins): Include neon builtin definitions.
23805 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23806 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23808 2013-04-04 Marek Polacek <polacek@redhat.com>
23810 PR tree-optimization/48186
23811 * predict.c (maybe_hot_frequency_p): Return false if
23812 HOT_BB_FREQUENCY_FRACTION is 0.
23813 (cgraph_maybe_hot_edge_p): Likewise.
23815 2013-04-04 Richard Biener <rguenther@suse.de>
23817 PR tree-optimization/56826
23818 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23821 2013-04-04 Richard Biener <rguenther@suse.de>
23823 PR tree-optimization/56213
23824 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23825 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23827 2013-04-04 Richard Biener <rguenther@suse.de>
23829 PR tree-optimization/56837
23830 * tree-loop-distribution.c (classify_partition): For non-zero
23831 values require that the value has the same precision as its
23832 mode to be useful as memset value.
23834 2013-04-03 Nick Clifton <nickc@redhat.com>
23836 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23837 (fmssf4): Use fmsf.s on E3V5 architectures.
23838 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23839 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23841 2013-04-03 Jeff Law <law@redhat.com>
23843 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23844 (lra-eliminations.o): Likewise.
23846 2013-04-03 Teresa Johnson <tejohnson@google.com>
23848 * gcov-io.c (compute_working_sets): Moved most of body of old
23849 compute_working_sets here from profile.c.
23850 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23851 (gcov_working_set_t): Moved typedef here from basic-block.h
23852 (compute_working_set): Declare.
23853 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23854 (get_working_sets): Renamed from compute_working_set,
23855 replace most of body with call to new compute_working_sets.
23856 (get_exec_counts): Replace call to compute_working_sets
23857 to get_working_sets.
23858 * profile.h (get_working_sets): Renamed from compute_working_set.
23859 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23860 to get_working_sets.
23861 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23862 * gcov-dump.c (dump_working_sets): New function.
23864 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23866 * hwint.c (sext_hwi, zext_hwi): New functions.
23867 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23868 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23869 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23870 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23871 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23872 (sext_hwi, zext_hwi): New functions.
23874 2013-04-03 Jeff Law <law@redhat.com>
23876 PR tree-optimization/56799
23877 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23878 back test for widening conversion erroneously dropped in prior change.
23880 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23883 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23884 instead of next_real_insn.
23886 2013-04-03 Marek Polacek <polacek@redhat.com>
23889 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23891 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23894 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23896 (thumb1_output_casesi): Likewise.
23897 (thumb2_output_casesi): Likewise.
23899 2013-04-03 Richard Biener <rguenther@suse.de>
23901 PR tree-optimization/56817
23902 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23904 (tree_unroll_loops_completely_1): ... new function to manually
23905 walk the loop tree, properly defering outer loops of unrolled
23906 loops to later iterations.
23908 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23910 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23911 (vectorizable_load): Likewise.
23912 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23913 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23915 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23917 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23920 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23922 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23924 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23926 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23928 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23930 PR tree-optimization/56790
23931 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23934 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23936 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23938 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23939 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23942 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23945 * tree.h (force_folding_builtin_constant_p): New decl.
23946 * builtins.c (force_folding_builtin_constant_p): New variable.
23947 (fold_builtin_constant_p): Fold immediately also if
23948 force_folding_builtin_constant_p.
23950 2013-04-03 Richard Biener <rguenther@suse.de>
23952 PR tree-optimization/56812
23953 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23954 DRs of the same interleaving chain are independent.
23956 2013-04-02 Jason Merrill <jason@redhat.com>
23958 * gdbinit.in (pbb): Use debug fn.
23960 2013-04-02 Lawrence Crowl <crowl@google.com>
23962 * sese.h (struct ivtype_map_elt_s): Remove unused.
23963 (extern debug_ivtype_map): Remove unused.
23964 (extern eq_ivtype_map_elts): Remove unused.
23965 * sese.c (debug_ivtype_map): Removed unused.
23966 (debug_ivtype_map_1): Removed unused.
23967 (debug_ivtype_elt): Remove unused.
23968 (eq_ivtype_map_elts): Remove unused.
23970 2013-04-02 Kai Tietz <ktietz@redhat.com>
23973 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23974 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23975 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23977 (legitimize_pe_coff_symbol): Likewise.
23978 (is_imported_p): New helper-function.
23979 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23980 for Windows x64 targets.
23981 (ix86_expand_prologue): Optimize for pe-coff targets.
23982 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23983 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23984 medium/large code-model.
23985 (legitimize_pic_address): Likewise.
23986 (legitimize_tls_address): Likewise.
23987 (ix86_expand_call): Likewise.
23988 (x86_output_mi_thunk): Likewise.
23989 (get_dllimport_decl): Add new beimport argument.
23990 (construct_plt_address): Don't assert for x64 pe-coff targets.
23991 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23993 (SYMBOL_FLAG_STUBVAR): New macro.
23994 (SYMBOL_REF_STUBVAR_P): Likewise.
23995 * config/i386/winnt.c (stub_list): New structure.
23996 (stub_head): New local variable.
23997 (i386_pe_record_stub): New function.
23998 (i386_pe_file_end): Emit refptr-stubs.
24000 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24002 PR rtl-optimization/56745
24003 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24004 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24007 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24008 and both of them are MEM_REFs, just compare first argument for
24009 equality and attempt to deal even with differing offsets.
24012 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24013 of gimple_clobber_p to be MEM_REF.
24014 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24015 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24016 after gimplification.
24017 * asan.c (get_mem_ref_of_assignment): Don't instrument
24018 gimple_clobber_p stmts.
24019 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24020 gimple_clobber_p stmt if they have MEM_REF lhs and
24021 are dead because of another gimple_clobber_p stmt.
24022 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24023 gimple_clobber_p stmts like debug stmts.
24024 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24025 that refer to unused VAR_DECLs or uninitialized values.
24026 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24027 gimple_clobber_p stmts if they refer to removed parameters.
24028 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24031 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24033 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24034 using SWI48 mode attribute.
24036 2013-04-02 Wei Mi <wmi@google.com>
24038 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24039 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24040 *<rotate_insn><mode>3_mask in i386.md.
24042 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24044 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24046 2013-04-02 Richard Biener <rguenther@suse.de>
24048 PR tree-optimization/56778
24049 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24050 Runtime alias tests are not supported for gather loads.
24051 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24052 stmts referenced from SSA operands before updating SSA form.
24054 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24055 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24057 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24058 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24059 * config/arm/cortex-a53.md: New file.
24060 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24061 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24062 * config/arm/arm.c (arm_issue_rate): Likewise.
24063 * config/arm/arm-tune.md: Regenerate
24064 * config/arm/arm-tables.opt: Regenerate.
24065 * config/arm/arm-cores.def: Add cortex-a53.
24067 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24069 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24072 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24074 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24075 scalar load/store operations using B/H registers.
24076 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24078 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24080 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24082 * config/aarch64/aarch64.c
24083 (aarch64_simd_scalar_immediate_valid_for_move): New.
24084 * config/aarch64/aarch64-protos.h
24085 (aarch64_simd_scalar_immediate_valid_for_move): New.
24086 * config/aarch64/constraints.md (Dh, Dq): New.
24087 * config/aarch64/iterators.md (hq): New.
24089 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24091 * reorg.c (get_branch_condition): Deal with conditional returns.
24092 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24094 2013-04-01 Wei Mi <wmi@google.com>
24096 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24097 Truncate operand 2 using %b asm operand modifier.
24098 (*<shift_insn><mode>3_mask): Ditto.
24099 (*<rotate_insn><mode>3_mask): Ditto.
24101 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24103 PR middle-end/56798
24104 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24106 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24108 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24110 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24112 2013-03-30 Lawrence Crowl <crowl@google.com>
24114 * dse.c (clear_alias_sets): Remove never set.
24115 (disqualified_clear_alias_sets): Remove never set.
24116 (clear_alias_mode_pool): Remove never set.
24117 (dse_step0): Remove condition that is never true.
24118 (canon_address): Remove condition that is never true.
24119 (dse_step7): Remove condition that is never true.
24120 (rest_of_handle_dse): Remove condition that is never true.
24121 (rest_of_handle_dse::did_global): Remove never read from above.
24122 (dse_step2_spill): Remove never called from above.
24123 (dse_step5_spill): Remove never called from above.
24125 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24127 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24128 JUMP_TABLE_DATA changes.
24129 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24130 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24131 (Insns) <jump_table_data>: New entry.
24132 * doc/tm.texi: Regenerate.
24134 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24136 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24137 for table jump at the end of a basic block using tablejump_p.
24138 * targhooks.c (default_invalid_within_doloop): Likewise.
24139 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24140 target hook implementation that is identical to the default hook.
24141 (rs6000_invalid_within_doloop): Remove.
24143 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24144 unused variable from tablejump_p call.
24146 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24147 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24148 (INSN_DELETED_P): Likewise.
24149 (emit_jump_table_data): New prototype.
24150 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24151 after 4th as unused.
24152 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24153 * sched-vis.c (print_insn): Likewise.
24154 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24155 insn for compatibility with back ends that use next_active_insn to
24156 identify jump table data.
24157 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24158 (remove_insn): Likewise.
24159 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24161 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24162 (emit_jump_table_data): New function.
24164 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24165 basic block, a JUMP_TABLE_DATA never is.
24166 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24167 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24168 off from code handling real insns.
24169 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24170 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24171 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24172 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24173 is not a NONDEBUG_INSN_P.
24174 * ira-costs.c (scan_one_insn): Likewise.
24175 * jump.c (mark_all_labels): Likewise.
24176 (mark_jump_label_1): Likewise.
24177 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24178 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24180 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24181 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24182 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24183 JUMP_TABLE_DATA_P insns.
24184 (calculate_elim_costs_all_insns): Likewise.
24185 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24186 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24187 (delete_output_reload): Code style fixups.
24188 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24189 insn flags on this non-insn.
24190 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24191 as scheduling barriers, for pre-change compatibility.
24192 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24193 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24195 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24196 redundant JUMP_TABLE_DATA_P test.
24197 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24198 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24199 (frv_for_each_packet): Likewise.
24200 * config/i386/i386.c (min_insn_size): Likewise.
24201 (ix86_avoid_jump_mispredicts): Likewise.
24202 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24203 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24204 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24205 (mips16_insn_length): Robustify.
24206 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24207 (mips16_split_long_branches): Likewise.
24208 * config/pa/pa.c (pa_combine_instructions): Likewise.
24209 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24210 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24211 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24212 as contributing to pool range lengths.
24213 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24214 Remove redundant JUMP_TABLE_DATA_P test.
24215 (sh_loop_align): Likewise.
24216 (split_branches): Likewise.
24217 (sh_insn_length_adjustment): Likewise.
24218 * config/spu/spu.c (get_branch_target): Likewise.
24220 2013-03-29 Jan Hubicka <jh@suse.cz>
24222 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24223 gcov streaming; stream hot bb threshold to ltrans.
24224 * predict.c (get_hot_bb_threshold): Break out from ....
24225 (maybe_hot_count_p): ... here.
24226 (set_hot_bb_threshold): New function.
24227 * lto-section-in.c (lto_section_name): Add profile.
24228 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24229 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24230 and data-streamer.h
24231 (histogram_entry): New structure.
24232 (histogram, histogram_pool): New global vars.
24233 (histogram_hash): New structure.
24234 (histogram_hash::hash): New method.
24235 (histogram_hash::equal): Likewise.
24236 (account_time_size): New function.
24237 (cmp_counts): New function.
24238 (dump_histogram): New function.
24239 (ipa_profile_generate_summary): New function.
24240 (ipa_profile_write_summary): New function.
24241 (ipa_profile_read_summary): New function.
24242 (ipa_profile): Decide on threshold.
24243 (pass_ipa_profile): Add ipa_profile_write_summary and
24244 ipa_profile_read_summary.
24245 * Makefile.in (ipa.o): Update dependencies.
24246 * lto-streamer.h (LTO_section_ipa_profile): New section.
24248 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24250 * tree.h (VAR_P): New.
24252 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24255 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24257 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24259 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24260 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24261 (control_flow_insn_p): Likewise.
24262 * cfgrtl.c (duplicate_insn_chain): Likewise.
24263 * final.c (get_attr_length_1): Likewise.
24264 (shorten_branches): Likewise.
24265 (final_scan_insn): Likewise.
24266 * function.c (instantiate_virtual_regs): Likewise.
24267 * gcse.c (insert_insn_end_basic_block): Likewise.
24268 * ira-costs.c (scan_one_insn): Likewise.
24269 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24270 * lra.c (check_rtl): Likewise.
24271 * reload1.c (elimination_costs_in_insn): Likewise.
24272 * reorg.c (follow_jumps): Likewise.
24274 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24275 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24276 (thumb_far_jump_used_p): Likewise.
24277 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24278 (workaround_speculation): Likewise.
24279 (add_sched_insns_for_speculation): Likewise.
24280 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24281 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24282 (frv_for_each_packet): Likewise.
24283 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24284 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24285 (final_emit_insn_group_barriers): Likewise.
24286 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24287 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24288 (mips16_insn_length): Likewise.
24289 * config/pa/pa.c (pa_reorg): Likewise.
24290 (pa_combine_instructions): Likewise.
24291 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
24292 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
24293 (sh_reorg): Likewise.
24294 (split_branches): Likewise.
24295 * config/spu/spu.c (get_branch_target): Likewise.
24297 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
24300 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
24302 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
24303 Fix declaration name.
24305 2013-03-28 Lawrence Crowl <crowl@google.com>
24307 * graphds.h (struct graph.indicies): Remove unused.
24308 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
24309 (SCOP_ORIGINAL_PDDRS): Remove unused.
24310 * sese.h (extern insert_loop_close_phis): Removed unused.
24311 (extern insert_guard_phis): Removed unused.
24312 (extern ivtype_map_elt_info): Removed unused.
24313 (new_ivtype_map_elt): Removed unused.
24314 * sese.c (ivtype_map_elt_info): Removed unused.
24316 2013-03-28 Lawrence Crowl <crowl@google.com>
24318 * Makefile.in: Add several missing include dependences.
24320 (test-dump.o): New. This object is not added to any executable,
24321 but is present for ad-hoc testing.
24323 (debug (const bitmap_head_def &)): New.
24324 (debug (const bitmap_head_def *)): New.
24326 (extern debug (const bitmap_head_def &)): New.
24327 (extern debug (const bitmap_head_def *)): New.
24329 (debug (edge_def &)): New.
24330 (debug (edge_def *)): New.
24332 (debug (basic_block_def &)): New.
24333 (debug (basic_block_def *)): New.
24335 (dump_node (const_tree, int, FILE *)): Correct source file.
24337 (debug (die_struct &)): New.
24338 (debug (die_struct *)): New.
24340 (extern debug (die_struct &)): New.
24341 (extern debug (die_struct *)): New.
24342 * gimple-pretty-print.c
24343 (debug (gimple_statement_d &)): New.
24344 (debug (gimple_statement_d *)): New.
24345 * gimple-pretty-print.h
24346 (extern debug (gimple_statement_d &)): New.
24347 (extern debug (gimple_statement_d *)): New.
24349 (debug (ira_allocno_copy &)): New.
24350 (debug (ira_allocno_copy *)): New.
24351 (debug (ira_allocno &)): New.
24352 (debug (ira_allocno *)): New.
24354 (extern debug (ira_allocno_copy &)): New.
24355 (extern debug (ira_allocno_copy *)): New.
24356 (extern debug (ira_allocno &)): New.
24357 (extern debug (ira_allocno *)): New.
24359 (debug (live_range &)): New.
24360 (debug (live_range *)): New.
24362 (debug (lra_live_range &)): New.
24363 (debug (lra_live_range *)): New.
24365 (debug (lra_live_range &)): New.
24366 (debug (lra_live_range *)): New.
24368 (debug (omega_pb_d &)): New.
24369 (debug (omega_pb_d *)): New.
24371 (extern debug (omega_pb_d &)): New.
24372 (extern debug (omega_pb_d *)): New.
24374 (debug (const rtx_def &)): New.
24375 (debug (const rtx_def *)): New.
24377 (debug_tree (tree): Move within file.
24378 (debug_raw (const tree_node &)): New.
24379 (debug_raw (const tree_node *)): New.
24380 (dump_tree_via_hooks (const tree_node *, int)): New.
24381 (debug (const tree_node &)): New.
24382 (debug (const tree_node *)): New.
24383 (debug_verbose (const tree_node &)): New.
24384 (debug_verbose (const tree_node *)): New.
24385 (debug_head (const tree_node &)): New.
24386 (debug_head (const tree_node *)): New.
24387 (debug_body (const tree_node &)): New.
24388 (debug_body (const tree_node *)): New.
24389 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24390 (debug (vec<tree, va_gc> &)): New.
24391 (debug (vec<tree, va_gc> *)): New.
24393 (extern debug (const rtx_def &)): New.
24394 (extern debug (const rtx_def *)): New.
24396 (debug_raw (simple_bitmap_def &)): New.
24397 (debug_raw (simple_bitmap_def *)): New.
24398 (debug (simple_bitmap_def &)): New.
24399 (debug (simple_bitmap_def *)): New.
24401 (extern debug (simple_bitmap_def &)): New.
24402 (extern debug (simple_bitmap_def *)): New.
24403 (extern debug_raw (simple_bitmap_def &)): New.
24404 (extern debug_raw (simple_bitmap_def *)): New.
24406 (debug (vinsn_def &)): New.
24407 (debug (vinsn_def *)): New.
24408 (debug_verbose (vinsn_def &)): New.
24409 (debug_verbose (vinsn_def *)): New.
24410 (debug (expr_def &)): New.
24411 (debug (expr_def *)): New.
24412 (debug_verbose (expr_def &)): New.
24413 (debug_verbose (expr_def *)): New.
24414 (debug (vec<rtx> &)): New.
24415 (debug (vec<rtx> *)): New.
24417 (extern debug (vinsn_def &)): New.
24418 (extern debug (vinsn_def *)): New.
24419 (extern debug_verbose (vinsn_def &)): New.
24420 (extern debug_verbose (vinsn_def *)): New.
24421 (extern debug (expr_def &)): New.
24422 (extern debug (expr_def *)): New.
24423 (extern debug_verbose (expr_def &)): New.
24424 (extern debug_verbose (expr_def *)): New.
24425 (extern debug (vec<rtx> &)): New.
24426 (extern debug (vec<rtx> *)): New.
24428 (_list_iter_cond_expr): Make inline instead of static.
24430 (debug (sreal &)): New.
24431 (debug (sreal *)): New.
24433 (extern debug (sreal &)): New.
24434 (extern debug (sreal *)): New.
24436 (extern debug_raw (const tree_node &)): New.
24437 (extern debug_raw (const tree_node *)): New.
24438 (extern debug (const tree_node &)): New.
24439 (extern debug (const tree_node *)): New.
24440 (extern debug_verbose (const tree_node &)): New.
24441 (extern debug_verbose (const tree_node *)): New.
24442 (extern debug_head (const tree_node &)): New.
24443 (extern debug_head (const tree_node *)): New.
24444 (extern debug_body (const tree_node &)): New.
24445 (extern debug_body (const tree_node *)): New.
24446 (extern debug (vec<tree, va_gc> &)): New.
24447 (extern debug (vec<tree, va_gc> *)): New.
24449 (debug (struct loop &)): New.
24450 (debug (struct loop *)): New.
24451 (debug_verbose (struct loop &)): New.
24452 (debug_verbose (struct loop *)): New.
24453 * tree-dump.c: Add header dependence.
24455 (extern debug (struct loop &)): New.
24456 (extern debug (struct loop *)): New.
24457 (extern debug_verbose (struct loop &)): New.
24458 (extern debug_verbose (struct loop *)): New.
24460 (debug (data_reference &)): New.
24461 (debug (data_reference *)): New.
24462 (debug (vec<data_reference_p> &)): New.
24463 (debug (vec<data_reference_p> *)): New.
24464 (debug (vec<ddr_p> &)): New.
24465 (debug (vec<ddr_p> *)): New.
24467 (extern debug (data_reference &)): New.
24468 (extern debug (data_reference *)): New.
24469 (extern debug (vec<data_reference_p> &)): New.
24470 (extern debug (vec<data_reference_p> *)): New.
24471 (extern debug (vec<ddr_p> &)): New.
24472 (extern debug (vec<ddr_p> *)): New.
24474 (debug (pt_solution &)): New.
24475 (debug (pt_solution *)): New.
24477 (extern debug (pt_solution &)): New.
24478 (extern debug (pt_solution *)): New.
24480 (debug (_var_map &)): New.
24481 (debug (_var_map *)): New.
24482 (debug (tree_live_info_d &)): New.
24483 (debug (tree_live_info_d *)): New.
24485 (extern debug (_var_map &)): New.
24486 (extern debug (_var_map *)): New.
24487 (extern debug (tree_live_info_d &)): New.
24488 (extern debug (tree_live_info_d *)): New.
24490 2013-03-28 Jan Hubicka <jh@suse.cz>
24492 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24494 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24496 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24497 record only when desired or required.
24499 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24501 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24502 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24504 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24506 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24507 (*andsi3_compare0_uxtw): New pattern.
24508 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24509 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24511 2013-03-28 Jan Hubicka <jh@suse.cz>
24513 * data-streamer-in.c (streamer_read_gcov_count): New function.
24514 * gimple-streamer-out.c: Include value-prof.h.
24515 (output_gimple_stmt): Output histogram.
24516 (output_bb): Use streamer_write_gcov_count.
24517 * value-prof.c: Include data-streamer.h
24518 (dump_histogram_value): Add HIST_TYPE_MAX.
24519 (stream_out_histogram_value): New function.
24520 (stream_in_histogram_value): New function.
24521 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24522 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24523 * data-streamer-out.c (streamer_write_gcov_count): New function.
24524 (streamer_write_gcov_count_stream): New function.
24525 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24526 (lto_output_node): Likewise.
24527 (input_node, input_edge): Likewise.
24528 * lto-streamer-out.c (output_cfg): Update streaming.
24529 * lto-streamer-in.c (input_cfg): Likewise.
24530 * data-streamer.h (streamer_write_gcov_count,
24531 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24532 * gimple-streamer-in.c: Include value-prof.h
24533 (input_gimple_stmt): Input histograms.
24534 (input_bb): Update profile streaming.
24536 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24538 * genmodes.c (emit_max_int): New function.
24539 (emit_insn_modes_h): Added call to emit_max_function.
24540 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24542 * machmode.def: Fixed comment.
24544 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24546 * combine.c (try_combine): Removed useless assert.
24547 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24549 2013-03-28 Marek Polacek <polacek@redhat.com>
24550 Richard Biener <rguenther@suse.de>
24552 PR tree-optimization/56695
24553 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24554 build signed result of a vector comparison.
24555 * tree-cfg.c (verify_gimple_comparison): Check that a result
24556 of a vector comparison has signed type.
24558 2013-03-28 Richard Biener <rguenther@suse.de>
24560 PR tree-optimization/37021
24561 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24562 do not restrict gaps between groups.
24563 * tree-vect-stmts.c (vectorizable_load): Properly account for
24564 a gap between groups.
24566 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24568 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24569 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24572 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24574 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24575 * doc/extend.texi (Named Address Spaces): Ditto.
24576 (Variable Attributes): Ditto.
24578 2013-03-27 Kai Tietz <ktietz@redhat.com>
24580 * config.build: Add support for cygwin x64 target.
24581 * config.gcc: Likewise.
24582 * config.host: Likewise.
24583 * configure.ac: Likewise
24584 * configure: Regenerated.
24586 2013-03-27 Kai Tietz <ktietz@redhat.com>
24588 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24589 * config/i386/t-cygwin-w64: New file.
24590 * config/i386/cygwin-w64.h: New file.
24591 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24592 and add support for x64-cygwin target.
24593 (CPP_SPEC): Likewise.
24594 (CXX_WRAP_SPEC_LIST): Undefine before define.
24595 (LIBGCJ_SONAME): Use 15 as version.
24597 2013-03-27 Richard Biener <rguenther@suse.de>
24599 PR tree-optimization/56716
24600 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24601 dumping for ref nodes.
24603 2013-03-27 Martin Jambor <mjambor@suse.cz>
24605 PR tree-optimization/55334
24606 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24607 restricted pointers to arrays.
24609 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24611 * Makefile.in (.SUFFIXES): Add .cc.
24612 (.c.o): Apply same recipe for implicit rule .cc.o.
24614 2013-03-27 Richard Biener <rguenther@suse.de>
24616 PR tree-optimization/37021
24617 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24618 REALPART/IMAGPART_EXPRs around the supported refs.
24619 * tree-ssa-structalias.c (find_func_aliases): Assume that
24620 floating-point values are not used to transfer pointers.
24622 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24624 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24625 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24626 * doc/tm.texi: Regenerate.
24627 * targhooks.h (default_has_ifunc_p): New.
24628 * targhooks.c (default_has_ifunc_p): Ditto.
24629 * config/linux-protos.h: New file.
24630 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24631 hook for linux which disables support of indirect functions in android.
24632 * config/linux-android.c: New file.
24633 * config/t-linux-android.c: Ditto.
24634 * config.gcc: Added new object file linux-android.o.
24635 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24636 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24637 * varasm.c (do_assemble_alias): Likewise.
24638 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24639 doesn't support indirect functions.
24640 * configure: Regenerate.
24642 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24645 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24646 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24649 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24651 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24653 2013-03-27 Terry Guo <terry.guo@arm.com>
24655 * config/arm/arm-cores.def: Added core cortex-r7.
24656 * config/arm/arm-tune.md: Regenerated.
24657 * config/arm/arm-tables.opt: Regenerated.
24658 * doc/invoke.texi: Added entry for core cortex-r7.
24660 2013-03-27 Walter Lee <walt@tilera.com>
24662 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24663 double-decrement of next_scratch_regno.
24665 2013-03-27 Walter Lee <walt@tilera.com>
24667 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24669 (insn_v1mulus): Ditto.
24670 (insn_v2muls): Ditto.
24672 2013-03-27 Walter Lee <walt@tilera.com>
24674 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24675 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24677 2013-03-27 Walter Lee <walt@tilera.com>
24679 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24680 (*sibcall_value): Ditto.
24682 2013-03-27 Walter Lee <walt@tilera.com>
24684 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24685 (insn_mnz_v8qi): ... this ...
24686 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24688 (insn_v<n>mnz): Replaced by ...
24689 (insn_v1mnz): ... this ...
24690 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24692 (insn_mz_<mode>): Replaced by ...
24693 (insn_mz_v8qi): ... this ...
24694 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24696 (insn_v<n>mz): Replaced by ...
24697 (insn_v1mz): ... this ...
24698 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24701 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24703 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24705 2013-03-26 Roland McGrath <mcgrathr@google.com>
24707 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24708 than fprintf with a non-constant, non-format string.
24710 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24712 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24713 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24714 operand 0 predicate.
24715 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24716 attribute. Use general_x64nomem_operand as operand 1 predicate.
24717 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24718 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24719 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24720 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24721 (mov<mode>_insv_1): Remove expander. Merge insn with
24722 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24723 Use general_x64nomem_operand as operand 1 predicate.
24724 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24725 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24726 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24727 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24728 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24729 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24730 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24731 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24732 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24733 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24734 (general_x64nomem_operand): Ditto.
24736 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24738 * config/rtems.opt: Add -pthread option.
24740 2013-03-26 Richard Biener <rguenther@suse.de>
24742 * alias.c (find_base_term): Avoid redundant and not used recursion.
24743 (base_alias_check): Get the initial base term from the caller.
24744 (true_dependence_1): Compute and pass base terms to base_alias_check.
24745 (write_dependence_p): Likewise.
24746 (may_alias_p): Likewise.
24748 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24750 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24751 PC-relative load in SI modes and above only.
24753 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24755 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24756 * config/arm/iwmmxt.md (WCGR0): Update.
24757 (WCGR1, WCGR2, WCGR3): Likewise.
24759 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24761 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24762 Use x64 and nox64 isa attributes.
24764 2013-03-26 Richard Biener <rguenther@suse.de>
24766 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24767 alignment computations and rely on get_object_alignment_1
24768 for the !TYPE_P case.
24769 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24771 2013-03-26 Walter Lee <walt@tilera.com>
24773 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24774 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24776 2013-03-25 Jeff Law <law@redhat.com>
24778 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24779 check for INTEGRAL_TYPE_P that was missing due to checking in
24780 wrong version of prior patch.
24782 2013-03-25 Walter Lee <walt@tilera.com>
24784 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24785 TILEGX_INSN_SHUFFLEBYTES1.
24786 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24788 (tilegx_builtins): Ditto.
24789 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24791 2013-03-25 Walter Lee <walt@tilera.com>
24793 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24794 (floatunssisf2): New pattern.
24795 (floatsidf2): New pattern.
24796 (floatunssidf2): New pattern.
24798 2013-03-25 Walter Lee <walt@tilera.com>
24800 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24801 tests for constraint J, K, N, P.
24803 2013-03-25 Walter Lee <walt@tilera.com>
24805 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24806 Use indirect/pcrel encoding.
24807 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24810 2013-03-25 Steve Ellcey <sellcey@mips.com>
24812 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24813 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24814 * config/mips/mips.c (mips_option_override): Set IMADD default.
24815 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24816 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24817 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24818 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24819 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24821 2013-03-25 Jeff Law <law@redhat.com>
24823 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24824 slightly to avoid creating and folding useless trees. Simplify
24825 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24827 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24829 * config/i386/i386.md (*zero_extendsidi2): Merge with
24830 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24831 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24832 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24833 targets. Clarify comment.
24835 2013-03-25 Martin Jambor <mjambor@suse.cz>
24837 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24838 pass-through jump functions differently.
24839 (ipa_read_jump_function): Likewise. Also use setter functions to set
24842 2013-03-25 Martin Jambor <mjambor@suse.cz>
24844 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24845 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24847 (ipa_get_indirect_edge_target): New function.
24848 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24849 ipa_get_indirect_edge_target. Update all callers.
24850 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24851 ipa_get_indirect_edge_target_1 instead of calling
24852 ipa_get_indirect_edge_target.
24853 (create_specialized_node): Pass aggvlas to
24854 ipcp_discover_new_direct_edges.
24856 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24858 * config/arm/arm.md (f_sels, f_seld): New types.
24859 (*cmov<mode>): New pattern.
24860 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24863 2013-03-25 Kai Tietz <ktietz@redhat.com>
24865 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24866 POSIX-printf for mingw-hosted builds.
24868 2013-03-25 Richard Biener <rguenther@suse.de>
24870 PR middle-end/56694
24871 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24872 must-not-throw stmt location.
24874 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24876 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24877 Emit load-acquire versions when acq is true.
24878 (arm_emit_store_exclusive): Add rel parameter.
24879 Emit store-release versions when rel is true.
24880 (arm_split_compare_and_swap): Use acquire-release instructions
24882 of barriers when appropriate.
24883 (arm_split_atomic_op): Likewise.
24884 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24885 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24886 (VUNSPEC_SLX): Likewise.
24887 (VUNSPEC_LDA): Likewise.
24888 (VUNSPEC_STL): Likewise.
24889 * config/arm/sync.md (atomic_load<mode>): New pattern.
24890 (atomic_store<mode>): Likewise.
24891 (arm_load_acquire_exclusive<mode>): Likewise.
24892 (arm_load_acquire_exclusivesi): Likewise.
24893 (arm_load_acquire_exclusivedi): Likewise.
24894 (arm_store_release_exclusive<mode>): Likewise.
24896 2013-03-25 Catherine Moore <clm@codesourcery.com>
24898 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24899 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24900 * config/mip/predicates.md (lwsp_swsp_operand,
24901 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24902 sb16_operand, db4_operand, db7_operand, ib3_operand,
24903 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24904 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24905 andi16_operand): New predicates.
24906 * config/mips/mips.md (compression): New attribute.
24907 (enabled): New attribute.
24908 (length): Consider compression in computing length.
24909 (shift_compression): New code attribute.
24910 (*add<mode>3): New operands. Record compression.
24911 (sub<mode>3): Likewise.
24912 (one_cmpl<mode>2): Likewise.
24913 (*and<mode>3): Likewise.
24914 (*ior<mode>3): Likewise.
24915 (unnamed pattern for xor): Likewise.
24916 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24917 (*<optab><mode>3): Likewise.
24918 (*mov<mode>_internal: Likewise.
24919 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24920 (mips_unsigned_immediate_p): New.
24921 (umips_lwsp_swsp_address_p): New.
24922 (m16_based_address_p): New.
24923 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24924 (mips_unsigned_immediate_p): New prototype.
24925 (lwsp_swsp_address_p): New prototype.
24926 (m16_based_address_p): New prototype.
24927 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24928 (mips_signed_immediate_p): New function.
24929 (m16_based_address_p): New function.
24930 (lwsp_swsp_address_p): New function.
24931 (mips_print_operand_punctuation): Recognize short delay slot insns
24932 for microMIPS.add<mode>3"
24934 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24937 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24938 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24940 2013-03-25 Richard Biener <rguenther@suse.de>
24942 PR tree-optimization/56689
24943 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24946 2013-03-25 Richard Biener <rguenther@suse.de>
24948 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24950 (memory_references): Likewise.
24951 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24952 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24953 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24954 (gather_mem_refs_in_loops): Fold into ...
24955 (analyze_memory_references): ... this. Move initialization
24956 to tree_ssa_lim_initialize.
24957 (fill_always_executed_in): Rename to ...
24958 (fill_always_executed_in_1): ... this.
24959 (fill_always_executed_in): Move contains_call computation to
24960 this new function from ...
24961 (tree_ssa_lim_initialize): ... here.
24962 (tree_ssa_lim): Call fill_always_executed_in.
24964 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24966 * postreload.c (reload_combine): Fix code detecting returns.
24968 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24970 * function.c (emit_use_return_register_into_block): On cc0 targets,
24971 do not emit the sequence between cc0 setter and user.
24973 2013-03-25 Kai Tietz <ktietz@redhat.com>
24975 * config/i386/predicates.md (local_symbolic_operand): Interpret
24976 dll-imported symbols as none-local.
24978 2013-03-25 Richard Biener <rguenther@suse.de>
24980 * tree-ssa-loop-im.c (struct depend): Remove.
24981 (struct lim_aux_data): Make depends a vec of gimples.
24982 (free_lim_aux_data): Adjust.
24983 (add_dependency): Likewise.
24984 (set_level): Likewise.
24986 2013-03-25 Richard Biener <rguenther@suse.de>
24988 PR middle-end/56434
24989 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24990 the pointer returned by calls with ECF_MALLOC set.
24992 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24994 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24996 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24998 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24999 using MMXMODE mode iterator.
25000 (*move<mode>_internal): Merge with *movv2sf_internal and
25001 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25003 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25005 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25006 (record_last_mem_set_info): Likewise.
25008 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25009 of XNEWVEC followed by memset.
25010 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25012 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25014 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25015 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25016 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25017 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25018 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25019 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25020 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25021 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25022 BARRIER_P instead of GET_CODE.
25024 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25026 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25027 inaccuracy in the probing code.
25029 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25030 (ctrapdi4): Likewise.
25032 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25034 * calls.c (expand_call): Add missing guard to code handling return
25035 of non-BLKmode structures in MSB.
25036 * function.c (expand_function_end): Likewise.
25038 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25040 * combine.c (try_combine): Adjust comment. Do not add the set of
25041 insn #0 if the destination indirectly is set or dies in insn #2.
25042 Tidy up code to distribute a new note.
25044 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25046 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25047 also for alternatives 16 and 17.
25049 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25051 * config/i386/sse.md (*mov<mode>_internal): Merge with
25052 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25053 Emit insn template depending on type attribute. Use
25054 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25055 movd instead of movq mnemonic for interunit moves. Rewrite mode
25056 attribute calculation. Remove unit attribute calculation.
25057 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25058 Set prefix_data16 attribute for DImode ssemov types.
25059 Use Ym instead of y for SSE-MMX conversion alternatives.
25060 Reorder operand constraints.
25062 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25064 * df.h (df_insn_delete): Adjust prototype.
25065 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25066 and let it decide whether mark the basic block dirty.
25067 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25068 * df-scan.c (df_insn_info_delete): New helper function, split
25069 off from df_insn_delete.
25070 (df_scan_free_bb_info): Use it.
25071 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25073 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25074 that the insn is actually an insn and it has a non-NULL basic block.
25075 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25077 2013-03-22 Richard Biener <rguenther@suse.de>
25079 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25081 (mem_ref_alloc): Do not allocate them.
25082 (refs_independent_p): Do not query or maintain a cache.
25084 2013-03-22 Richard Biener <rguenther@suse.de>
25086 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25087 (gather_mem_refs_in_loops): Do not compute it.
25088 (analyze_memory_references): Do not allocate it.
25089 (tree_ssa_lim_finalize): Do not free it.
25090 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25092 2013-03-22 Richard Biener <rguenther@suse.de>
25094 * is-a.h (as_a): Use gcc_checking_assert.
25096 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25098 * config/aarch64/aarch64.c (aarch64_print_operand): New
25099 format specifier for printing a constant in hex.
25100 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25101 format specifier for printing second operand.
25103 2013-03-22 Richard Biener <rguenther@suse.de>
25105 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25107 (gather_mem_refs_in_loops): Perform store accumulation here.
25108 (create_vop_ref_mapping_loop): Remove.
25109 (create_vop_ref_mapping): Likewise.
25110 (analyze_memory_references): Initialize refs_stored_in_loop.
25111 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25112 (record_indep_loop): Remove.
25113 (record_dep_loop): New function.
25114 (ref_indep_loop_p_1): Adjust to only walk over references
25115 in the loop, not its subloops.
25116 (ref_indep_loop_p): Rename to ...
25117 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25118 maintaining a more fine-grained cache.
25119 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25120 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25122 2013-03-22 Richard Biener <rguenther@suse.de>
25124 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25125 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25126 aggregate mem_ref_loc.
25127 (free_mem_ref_locs): Inline into ...
25128 (memref_free): ... this and adjust.
25129 (mem_ref_alloc): Adjust.
25130 (mem_ref_locs_alloc): Remove.
25131 (record_mem_ref_loc): Adjust.
25132 (get_all_locs_in_loop): Rewrite into ...
25133 (for_all_locs_in_loop): ... this iterator.
25134 (rewrite_mem_ref_loc): New functor.
25135 (rewrite_mem_refs): Use for_all_locs_in_loop.
25136 (sm_set_flag_if_changed): New functor.
25137 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25138 (ref_always_accessed): New functor.
25139 (ref_always_accessed_p): Use for_all_locs_in_loop.
25141 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25143 * tree-pass.h (PROP_gimple_lvec): New.
25144 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25145 (init_optimization_passes): Move pass_lower_vector.
25146 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25148 (pass_lower_vector): Provide PROP_gimple_lvec.
25149 (pass_lower_vector_ssa): Likewise.
25150 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25152 2013-03-21 Mark Wielaard <mjw@redhat.com>
25154 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25156 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25158 * config/i386/i386.md (*movdi_internal): Disparage slightly
25159 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25160 conversion alternatives.
25162 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25164 PR middle-end/48087
25165 * diagnostic.def (DK_WERROR): New kind.
25166 * diagnostic.h (werrorcount): Define.
25167 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25168 promoted to DK_ERROR, increment DK_WERROR counter instead of
25170 * toplev.c (toplev_main): Call print_ignored_options even if
25171 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25172 even if just werrorcount is non-zero.
25175 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25177 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25178 (string_cst_pool_decl): New function.
25179 (optimize_one_addr_into_implicit_ptr): New function.
25180 (resolve_addr_in_expr): Optimize DWARF location expression
25181 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25182 which doesn't live in memory, but has DW_AT_location or
25183 DW_AT_const_value, or refers to a string literal, into
25184 DW_OP_GNU_implicit_pointer.
25185 (optimize_location_into_implicit_ptr): New function.
25186 (resolve_addr): If removing DW_AT_location of a variable because
25187 it was DW_OP_addr of address of the variable, but the variable doesn't
25188 live in memory, try to emit const value attribute for the initializer.
25190 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25192 * tree.h (VECTOR_TYPE_P): New macro.
25193 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25194 TYPE_MODE): Use it.
25195 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25196 VEC_COND_EXPR cannot be lvalues.
25197 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25199 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25201 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25202 Restrict the transformation to equal modes.
25204 2013-03-21 Richard Biener <rguenther@suse.de>
25206 PR tree-optimization/39326
25207 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25208 (MEM_ANALYZABLE): Adjust.
25209 (record_mem_ref_loc): Move bitmap ops ...
25210 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25211 unanalyzable refs, do not record locations for it.
25212 (analyze_memory_references): Allocate ref zero as shared
25214 (refs_independent_p): Do not test for unanalyzed mems here.
25215 (ref_indep_loop_p_1): Special-case disambiguation against
25216 the unanalyzed ref.
25217 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25219 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25221 * config/arm/arm-protos.h (tune_params): Add
25222 prefer_neon_for_64bits field.
25223 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25224 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25225 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25226 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25227 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25228 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25229 (arm_option_override): Handle -mneon-for-64bits new option.
25230 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25231 (prefer_neon_for_64bits): Declare new variable.
25232 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25233 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25234 (arch_enabled): Handle new arch types. Remove support for onlya8
25236 (one_cmpldi2): Use new arch names.
25237 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25238 * config/arm/arm.opt (mneon-for-64bits): Add option.
25239 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25240 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25241 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25243 * doc/invoke.texi (-mneon-for-64bits): Document.
25245 2013-03-21 Richard Biener <rguenther@suse.de>
25247 PR tree-optimization/39326
25248 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25249 (sort_bbs_in_loop_postorder_cmp): New function.
25250 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25252 2013-03-21 Richard Biener <rguenther@suse.de>
25254 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25255 (vect_insert_into_interleaving_chain): Likewise.
25256 (vect_drs_dependent_in_basic_block): Inline ...
25257 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25259 (vect_analyze_data_ref_dependence): ... here. Simplify.
25260 (vect_check_interleaving): Simplify.
25261 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25262 (vect_slp_analyze_data_ref_dependences): ... this new function.
25263 (dr_group_sort_cmp): New function.
25264 (vect_analyze_data_ref_accesses): Compute data-reference groups
25265 here instead of in vect_analyze_data_ref_dependence. Use
25266 a more efficient algorithm.
25267 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25268 vect_slp_analyze_data_ref_dependences. Call
25269 vect_analyze_data_ref_accesses earlier.
25270 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25271 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25272 (vect_slp_analyze_data_ref_dependences): New prototype.
25274 2013-03-21 Richard Biener <rguenther@suse.de>
25276 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25277 ref is stored in the loop.
25278 (find_refs_for_sm): Walk only over all stores.
25279 (store_motion_loop): Allocate from lim_bitmap_obstack.
25280 (store_motion): Likewise.
25282 2013-03-21 Richard Biener <rguenther@suse.de>
25284 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25285 Update virtual SSA form.
25287 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25289 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
25290 * configure: Regenerate.
25291 * config.in: Regenerate.
25292 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
25293 if !HAVE_LD_EH_FRAME_CIEV3.
25295 2013-03-21 Richard Biener <rguenther@suse.de>
25297 * tree-cfg.c (verify_expr_no_block): New function.
25298 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
25299 nor DECL_VALUE_EXPR have locations with associated blocks.
25300 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
25301 (clear_unused_block_pointer): Remove code dealing with
25302 blocks in DECL_DEBUG_EXPR locations.
25304 2013-03-21 Richard Biener <rguenther@suse.de>
25306 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
25307 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
25308 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
25309 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
25310 instead of DECL_DEBUG_EXPR_IS_FROM.
25311 * gimplify.c (gimplify_modify_expr): Likewise.
25312 * tree-cfg.c (verify_expr_location_1): Likewise.
25313 * tree-complex.c (create_one_component_var): Likewise.
25314 * tree-sra.c (create_access_replacement): Likewise.
25315 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
25316 (clear_unused_block_pointer): Likewise.
25317 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
25318 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
25319 * var-tracking.c (var_debug_decl): Likewise.
25320 (track_expr_p): Likewise.
25321 * tree-inline.c (add_local_variables): Likewise. Set
25322 DECL_HAS_DEBUG_EXPR_P after copying it.
25323 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
25324 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
25326 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25329 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
25330 * configure: Regenerate.
25331 * config.in: Regenerate.
25332 * config/i386/i386.md (*movdf_internal): Use
25333 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25334 movd instead of movq mnemonic for interunit moves.
25335 (*movdi_internal): Ditto.
25337 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25339 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
25340 (abd<mode>_3): New pattern.
25341 (aba<mode>_3): New pattern.
25342 (fabd<mode>_3): New pattern.
25344 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25346 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
25347 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
25348 occurrence of REGISTER_PREFIX as its empty string.
25350 2013-03-20 Jeff Law <law@redhat.com>
25352 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
25353 addititional equivalences for equality comparisons between an SSA_NAME
25354 and a constant where the SSA_NAME was set from a widening conversion.
25356 2013-03-20 Walter Lee <walt@tilera.com>
25358 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25360 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25362 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25363 depending on type attribute.
25364 (*movti_internal): Ditto.
25365 (*movtf_internal): Ditto.
25366 (*movxf_internal): Ditto.
25367 (*movdf_internal): Ditto.
25368 (*movsf_internal): Ditto.
25370 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25372 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25373 maybe_vex for sselog1 and ssemov types.
25374 (*movdi_internal): Reorder operand constraints.
25375 (*movsi_internal): Ditto. Set prefix attribute to
25376 maybe_vex for sselog1 and ssemov types.
25377 (*movtf_internal): Set prefix attribute to maybe_vex
25378 for sselog1 and ssemov types.
25379 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25380 DImode ssemov types. Reorder operand constraints.
25381 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25382 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25383 attribute for SImode ssemov types. Reorder operand constraints.
25385 2013-03-20 Martin Jambor <mjambor@suse.cz>
25387 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25388 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25390 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25392 * config/rs6000/predicates.md (indexed_address, update_address_mem
25393 update_indexed_address_mem): New predicates.
25394 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25395 attribute for load/store instructions.
25396 * config/rs6000/dfp.md (movsd_store): Likewise.
25397 (movsd_load): Likewise.
25398 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25399 (unnamed HI->DI extend define_insn): Likewise.
25400 (unnamed SI->DI extend define_insn): Likewise.
25401 (unnamed QI->SI extend define_insn): Likewise.
25402 (unnamed QI->HI extend define_insn): Likewise.
25403 (unnamed HI->SI extend define_insn): Likewise.
25404 (unnamed HI->SI extend define_insn): Likewise.
25405 (extendsfdf2_fpr): Likewise.
25406 (movsi_internal1): Likewise.
25407 (movsi_internal1_single): Likewise.
25408 (movhi_internal): Likewise.
25409 (movqi_internal): Likewise.
25410 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25411 attribute for load/store instructions.
25412 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25414 (mov<mode>_softfloat): Likewise.
25415 (mov<mode>_hardfloat32): Likewise.
25416 (mov<mode>_hardfloat64): Likewise.
25417 (mov<mode>_softfloat64): Likewise.
25418 (movdi_internal32): Likewise.
25419 (movdi_internal64): Likewise.
25420 (probe_stack_<mode>): Likewise.
25422 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25424 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25425 floating point, and decimal floating point to reload iterator.
25427 * config/rs6000/constraints.md (wl constraint): New constraints to
25428 return FLOAT_REGS if certain options are used to reduce the number
25429 of separate patterns that exist in the file.
25430 (wx constraint): Likewise.
25431 (wz constraint): Likewise.
25433 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25434 -mdebug=reg, print wg, wl, wx, and wz constraints.
25435 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25436 Initialize the reload functions for 64-bit binary/decimal floating
25438 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25439 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25440 create the buffer on the stack to overcome not having a 32-bit
25442 (rs6000_emit_move): Likewise.
25443 (rs6000_secondary_memory_needed_rtx): Likewise.
25444 (rs6000_alloc_sdmode_stack_slot): Likewise.
25445 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25446 via xxlxor, just like DFmode 0.0.
25448 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25449 define as 1 if we are running on a power7 or newer.
25450 (enum r6000_reg_class_enum): Add new constraints.
25452 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25453 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25454 with other moves by using conditional constraits (wg). Use LFIWZX
25455 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25456 (movsd splitter): Likewise.
25457 (movsd_hardfloat): Likewise.
25458 (movsd_softfloat): Likewise.
25460 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25461 binary and decimal floating point moves.
25462 (fmove_ok): New attributes to combine binary and decimal floating
25463 point moves, and to combine power6x (mfpgpr) moves along normal
25465 (real_value_to_target): Likewise.
25466 (f32_lr): Likewise.
25467 (f32_lm): Likewise.
25468 (f32_li): Likewise.
25469 (f32_sr): Likewise.
25470 (f32_sm): Likewise.
25471 (f32_si): Likewise.
25472 (movsf): Combine binary and decimal floating point moves. Combine
25473 power6x (mfpgpr) moves with other moves by using conditional
25474 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25475 (mov<mode> for SFmode/SDmode); Likewise.
25476 (SFmode/SDmode splitters): Likewise.
25477 (movsf_hardfloat): Likewise.
25478 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25479 (movsf_softfloat): Likewise.
25480 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25482 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25483 wx and wz constraints.
25485 * config/rs6000/constraints.md (wg constraint): New constraint to
25486 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25488 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25491 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25492 -mdebug=reg, print wg, wl, wx, and wz constraints.
25493 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25494 Initialize the reload functions for 64-bit binary/decimal floating
25496 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25497 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25498 create the buffer on the stack to overcome not having a 32-bit
25500 (rs6000_emit_move): Likewise.
25501 (rs6000_secondary_memory_needed_rtx): Likewise.
25502 (rs6000_alloc_sdmode_stack_slot): Likewise.
25503 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25504 via xxlxor, just like DFmode 0.0.
25506 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25507 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25508 with other moves by using conditional constraits (wg). Use LFIWZX
25509 and STFIWX for loading SDmode on power7.
25510 (movdd splitters): Likewise.
25511 (movdd_hardfloat32): Likewise.
25512 (movdd_softfloat32): Likewise.
25513 (movdd_hardfloat64_mfpgpr): Likewise.
25514 (movdd_hardfloat64): Likewise.
25515 (movdd_softfloat64): Likewise.
25517 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25518 64-bit binary and decimal floating point moves.
25519 (FMOVE64X): Likewise.
25520 (movdf): Combine 64-bit binary and decimal floating point moves.
25521 Combine power6x (mfpgpr) moves with other moves by using
25522 conditional constraits (wg).
25523 (mov<mode> for DFmode/DDmode): Likewise.
25524 (DFmode/DDmode splitters): Likewise.
25525 (movdf_hardfloat32): Likewise.
25526 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25527 (movdf_softfloat32): Likewise.
25528 (movdf_hardfloat64_mfpgpr): Likewise.
25529 (movdf_hardfloat64): Likewise.
25530 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25531 (movdf_softfloat64): Likewise.
25532 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25533 (reload_<mode>_load): Move to later in the file so they aren't in
25534 the middle of the floating point move insns.
25535 (reload_<mode>_store): Likewise.
25537 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25540 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25541 constraint if -mdebug=reg.
25542 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25543 Enable using dd reload support if needed.
25545 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25546 binary and decimal floating point moves in rs6000.md.
25547 (movtd_internal): Likewise.
25549 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25550 decimal floating point moves.
25552 (movtf_internal): Likewise.
25553 (mov<mode>_internal, TDmode/TFmode): Likewise.
25554 (movtf_softfloat): Likewise.
25555 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25557 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25558 movdi_internal64, using wg constraint for move direct operations.
25559 (movdi_internal64): Likewise.
25561 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25562 MODES_TIEABLE_P for selected modes. Print the numerical value of
25563 the various virtual registers. Use GPR/FPR first/last values,
25564 instead of hard coding the register numbers. Print which modes
25565 have reload functions registered.
25566 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25567 settings before/after setting cpu, target and subtarget settings.
25568 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25569 and for secondary reload failures in rs6000_secondary_reload_inner.
25570 (rs6000_secondary_reload_fail): Likewise.
25571 (rs6000_secondary_reload_inner): Likewise.
25573 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25574 macros for first/last GPR and FPR registers.
25575 (LAST_GPR_REGNO): Likewise.
25576 (FIRST_FPR_REGNO): Likewise.
25577 (LAST_FPR_REGNO): Likewise.
25579 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25580 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25581 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25582 (vcond<mode><mode>): Likewise.
25583 (vcondu<mode><mode>): Likewise.
25584 (vector_gtu<mode>): Likewise.
25585 (vector_gte<mode>): Likewise.
25586 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25587 to prevent the compiler from converting DImode operations to TImode.
25588 (ior<mode>3): Likewise.
25589 (and<mode>3): Likewise.
25590 (one_cmpl<mode>2): Likewise.
25591 (nor<mode>3): Likewise.
25592 (andc<mode>3): Likewise.
25594 * config/rs6000/constraints.md (wt constraint): New constraint
25595 that returns VSX_REGS if TImode is allowed in VSX registers.
25597 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25598 constant under VSX.
25600 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25601 similar to TImode, but it is restricted to being in the GPRs.
25603 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25604 TImode to occupy a single VSX register.
25606 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25607 -mvsx-timode for power7/power8.
25608 (power7 cpu): Likewise.
25609 (power8 cpu): Likewise.
25611 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25612 sure that TFmode/TDmode take up two registers if they are ever
25613 allowed in the upper VSX registers.
25614 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25616 (rs6000_init_hard_regno_mode_ok): Likewise.
25617 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25618 constraint. Print if LRA is turned on.
25619 (rs6000_option_override_internal): Give an error if -mvsx-timode
25620 and VSX is not enabled.
25621 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25622 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25623 to reg+offset addressing. Use PTImode when checking offset
25624 addresses for validity.
25625 (reg_offset_addressing_ok_p): Likewise.
25626 (rs6000_legitimate_offset_address_p): Likewise.
25627 (rs6000_legitimize_address): Likewise.
25628 (rs6000_legitimize_reload_address): Likewise.
25629 (rs6000_legitimate_address_p): Likewise.
25630 (rs6000_eliminate_indexed_memrefs): Likewise.
25631 (rs6000_emit_move): Likewise.
25632 (rs6000_secondary_reload): Likewise.
25633 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25634 reloads to fpr registers to continue to use reg+offset addressing,
25635 but 64-bit reloads to altivec registers need reg+reg addressing.
25636 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25637 it. Treat LO_SUM like a PLUS operation.
25638 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25639 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25640 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25641 registers to share a register with a smaller sized type, since VSX
25642 puts scalars in the upper 64-bits.
25643 (print_operand): Add support for PTImode.
25644 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25645 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25646 registers, but don't have arithmetic support.
25647 (rs6000_memory_move_cost): Add test for VSX.
25648 (rs6000_opt_masks): Add -mvsx-timode.
25650 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25653 (VSr): Use wt constraint for TImode.
25654 (VSv): Drop TImode support.
25655 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25656 (vsx_movti_64bit): Likewise.
25657 (vsx_movti_32bit): Likewise.
25658 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25659 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25660 one '?' on the appropriate output constraint. Do not allow TImode
25661 logical operations on 32-bit systems.
25662 (vsx_ior<mode>3): Likewise.
25663 (vsx_xor<mode>3): Likewise.
25664 (vsx_one_cmpl<mode>2): Likewise.
25665 (vsx_nor<mode>3): Likewise.
25666 (vsx_andc<mode>3): Likewise.
25667 (vsx_concat_<mode>): Likewise.
25668 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25670 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25671 OPTION_MASK_VSX_TIMODE.
25672 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25673 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25675 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25676 (TI2 iterator): New iterator for TImode, PTImode.
25677 (wd mode attribute): Add values for vector types.
25678 (movti_string): Replace TI move operations with operations for TImode
25679 and PTImode. Add support for TImode being allowed in VSX registers.
25680 (mov<mode>_string, TImode/PTImode): Likewise.
25681 (movti_ppc64): Likewise.
25682 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25683 (TI mode splitters): Likewise.
25685 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25688 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25690 PR tree-optimization/56355
25691 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25692 Also handle integers with undefined overflow.
25694 2013-03-20 Catherine Moore <clm@codesourcery.com>
25695 Maciej W. Rozycki <macro@codesourcery.com>
25696 Tom de Vries <tom@codesourcery.com>
25697 Nathan Sidwell <nathan@codesourcery.com>
25698 Iain Sandoe <iain@codesourcery.com>
25699 Nathan Froyd <froydnj@codesourcery.com>
25700 Chao-ying Fu <fu@mips.com>
25702 * doc/extend.texi: (micromips, nomicromips, nocompression):
25703 Document new function attributes.
25704 * doc/invoke.texi (minterlink-compressed, mmicromips,
25705 m14k, m14ke, m14kec): Document new options.
25706 (minterlink-mips16): Update documentation.
25707 * doc/md.texi (ZC, ZD): Document new constraints.
25708 * configure.ac (gcc_cv_as_micromips): Check if linker
25709 supports the .set micromips directive.
25710 * configure: Regenerate.
25711 * config.in: Regenerate.
25712 * config/mips/mips-tables.opt: Regenerate.
25713 * config/mips/micromips.md: New file.
25714 * constraints.md (ZC, ZD): New constraints.
25715 * config/mips/predicates.md (movep_src_register): New predicate.
25716 (movep_src_operand): New predicate.
25717 (non_volatile_mem_operand): New predicate.
25718 * config/mips/mips.md (multimem): New type.
25719 (length): Differentiate between 17-bit and 18-bit branch offsets.
25720 (MOVEP1, MOVEP2): New mode iterator.
25721 (mov_<load>l): Use ZC constraint.
25722 (mov_<load>r): Likewise.
25723 (mov_<store>l): Likewise.
25724 (mov_<store>r): Likewise.
25725 (*branch_equality<mode>_inverted): Add microMIPS support.
25726 (*branch_equality<mode>): Likewise.
25727 (*jump_absolute): Likewise.
25728 (indirect_jump_<mode>): Likewise.
25729 (tablejump_<mode>): Likewise.
25730 (<optab>_internal): Likewise.
25731 (sibcall_internal): Likewise.
25732 (sibcall_value_internal): Likewise.
25733 (prefetch): Use constraint ZD.
25734 * config/mips/mips.opt (minterlink-compressed): New option.
25735 (minterlink-mips16): Now an alias for minterlink-compressed.
25736 (mmicromips): New option.
25737 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25738 (compare_and_swap_12): Likewise.
25739 (sync_add<mode>): Likewise.
25740 (sync_<optab>_12): Likewise.
25741 (sync_old_<optab>_12): Likewise.
25742 (sync_new_<optab>_12): Likewise.
25743 (sync_nand_12): Likewise.
25744 (sync_old_nand_12): Likewise.
25745 (sync_new_nand_12): Likewise.
25746 (sync_sub<mode>): Likewise.
25747 (sync_old_add<mode>): Likewise.
25748 (sync_old_sub<mode>): Likewise.
25749 (sync_new_add<mode>): Likewise.
25750 (sync_new_sub<mode>): Likewise.
25751 (sync_<optab><mode>): Likewise.
25752 (sync_old_<optab><mode>): Likewise.
25753 (sync_new_<optab><mode>): Likewise.
25754 (sync_nand<mode>): Likewise.
25755 (sync_old_nand<mode>): Likewise.
25756 (sync_new_nand<mode>): Likewise.
25757 (sync_lock_test_and_set<mode>): Likewise.
25758 (test_and_set_12): Likewise.
25759 (atomic_compare_and_swap<mode>): Likewise.
25760 (atomic_exchange<mode>_llsc): Likewise.
25761 (atomic_fetch_add<mode>_llsc): Likewise.
25762 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25763 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25764 (umips_save_restore_pattern_p): Likewise.
25765 (umips_load_store_pair_p): Likewise.
25766 (umips_output_load_store_pair): Likewise.
25767 (umips_movep_target_p): Likewise.
25768 (umips_12bit_offset_address_p): Likewise.
25769 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25770 (mips_base_mips16): Rename this...
25771 (mips_base_compression_flags): ...to this. Update all uses.
25772 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25773 (mips_mips16_decl_p): Delete.
25774 (mips_nomips16_decl_p): Delete.
25775 (mips_get_compress_on_flags): New function.
25776 (mips_get_compress_off_flags): New function.
25777 (mips_get_compress_mode): New function.
25778 (mips_get_compress_on_name): New function.
25779 (mips_get_compress_off_name): New function.
25780 (mips_insert_attributes): Support multiple compression types.
25781 (mips_merge_decl_attributes): Likewise.
25782 (umips_12bit_offset_address_p): New function.
25783 (mips_start_function_definition): Emit .set micromips directive.
25784 (mips_call_may_need_jalx_p): New function.
25785 (mips_function_ok_for_sibcall): Add microMIPS support.
25786 (mips_print_operand_punctuation): Support short delay slots and
25788 (umips_swm_mask, umips_swm_encoding): New.
25789 (umips_build_save_restore): New function.
25790 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25791 (was_mips16_p): Remove.
25792 (old_compression_mode): New.
25793 (mips_set_compression_mode): New function.
25794 (mips_set_current_function): Add microMIPS support.
25795 (mips_option_override): Likewise.
25796 (umips_save_restore_pattern_p): New function.
25797 (umips_output_save_restore): New function.
25798 (umips_load_store_pair_p_1): New function.
25799 (umips_load_store_pair_p): New function.
25800 (umips_output_load_store_pair_1): New function.
25801 (umips_output_load_store_pair): New function.
25802 (umips_movep_target_p) New function.
25803 (mips_prepare_pch_save): Add microMIPS support.
25804 * config/mips/mips.h (TARGET_COMPRESSION): New.
25805 (TARGET_CPU_CPP_BUILTINS): Update macro
25806 to use new compression flags and to support microMIPS.
25807 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25808 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25809 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25810 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25811 (ASM_SPEC): Support mmicromips and mno-micromips.
25812 (M16STORE_REG_P): New macro.
25813 (MIPS_CALL): Support TARGET_MICROMIPS.
25814 (MICROMIPS_J): New macro.
25815 (mips_base_mips16): Rename this...
25816 (mips_base_compression_flags): ...to this.
25817 (UMIPS_12BIT_OFFSET_P): New macro.
25818 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25819 (MULTILIB_DIRNAMES): Likewise.
25820 2013-03-20 Richard Biener <rguenther@suse.de>
25822 PR tree-optimization/56661
25823 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25824 the result does not have to be distinct.
25826 2013-03-20 Richard Biener <rguenther@suse.de>
25828 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25831 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25832 Steven Bosscher <steven@gcc.gnu.org>
25834 PR rtl-optimization/56605
25835 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25837 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25840 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25841 that require movd instead of movq.
25843 2013-03-20 Richard Biener <rguenther@suse.de>
25845 * tree-ssa-structalias.c (struct variable_info): Add pointer
25846 to the first field of an aggregate with sub-vars. Make
25847 this and the pointer to the next subfield its ID.
25848 (vi_next): New function.
25849 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25850 storedanything_id, integer_id): Increment by one.
25851 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25852 get_call_clobber_vi): Adjust.
25853 (solution_set_expand): Simplify and speedup.
25854 (solution_set_add): Inline into ...
25855 (set_union_with_increment): ... this. Adjust accordingly.
25856 (do_sd_constraint): Likewise.
25857 (do_ds_constraint): Likewise.
25858 (do_complex_constraint): Simplify.
25859 (build_pred_graph): Adjust.
25860 (solve_graph): Likewise. Simplify and speedup.
25861 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25862 get_constraint_for_component_ref, get_constraint_for_1,
25863 first_vi_for_offset, first_or_preceding_vi_for_offset,
25864 create_function_info_for, create_variable_info_for_1,
25865 create_variable_info_for, intra_create_variable_infos): Adjust.
25866 (init_base_vars): Push NULL for ID zero.
25867 (compute_points_to_sets): Adjust.
25869 2013-03-20 Richard Biener <rguenther@suse.de>
25871 * cfgloop.c (verify_loop_structure): Streamline and avoid
25872 ICEing on corrupt loop tree.
25873 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25876 2013-03-20 Richard Biener <rguenther@suse.de>
25878 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25879 check whether an SSA update is needed.
25881 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25883 * config/mips/constraints.md (T): Rename to...
25887 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25888 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25890 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25892 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25893 (*subsi3_carryin_uxtw): Likewise.
25895 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25897 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25898 (*rorsi3_insn_uxtw): Likewise.
25900 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25902 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25903 (*extrsi5_insn_uxtw): Likewise.
25905 2013-03-19 Richard Biener <rguenther@suse.de>
25907 PR tree-optimization/56273
25908 * passes.c (init_optimization_passes): Move second VRP after DOM.
25910 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25912 * config/i386/i386.md (*movti_internal): Merge from
25913 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25914 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25915 nox64 isa attributes.
25917 2013-03-18 Richard Biener <rguenther@suse.de>
25919 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25921 (merge_node_constraints): Likewise.
25922 (build_succ_graph): Likewise.
25923 (valid_graph_edge): Inline into single caller.
25924 (unify_nodes): Likewise. Use bitmap_set_bit return value
25926 (scc_visit): Fix formatting and variable use.
25927 (do_sd_constraint): Use gcc_checking_assert.
25928 (do_ds_constraint): Likewise.
25929 (do_complex_constraint): Likewise.
25930 (condense_visit): Likewise. Cleanup.
25931 (dump_pred_graph): New function.
25932 (perform_var_substitution): Dump the pred-graph before
25933 variable substitution.
25934 (find_equivalent_node): Use gcc_checking_assert.
25935 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25937 2013-03-18 Richard Biener <rguenther@suse.de>
25939 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25940 Remove cond_expr_stmt_list argument and do not gimplify the
25942 (vect_loop_versioning): Adjust.
25943 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25944 Cleanup to use less temporaries.
25945 (vect_create_data_ref_ptr): Cleanup.
25947 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25949 PR tree-optimization/56635
25950 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25951 require types_compatible_p types.
25953 2013-03-18 Nick Clifton <nickc@redhat.com>
25955 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25956 spurious backslash.
25958 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25959 Add missing line to comment describing function.
25961 2013-03-18 Richard Biener <rguenther@suse.de>
25963 PR tree-optimization/56210
25964 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25965 Handle string / character search functions.
25966 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25968 2013-03-18 Richard Biener <rguenther@suse.de>
25970 PR middle-end/56483
25971 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25972 and implement properly.
25973 * gimple.h (gimple_cond_single_var_p): Remove.
25975 2013-03-18 Richard Biener <rguenther@suse.de>
25977 * tree-data-ref.h (find_data_references_in_loop): Declare.
25978 * tree-data-ref.c (get_references_in_stmt): Use a stack
25979 vector pre-allocated in the callers.
25980 (find_data_references_in_stmt): Adjust.
25981 (graphite_find_data_references_in_stmt): Likewise.
25982 (create_rdg_vertices): Likewise.
25983 (find_data_references_in_loop): Export.
25984 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25985 Compute dependences here...
25986 (vect_analyze_data_refs): ...not here. When we encounter
25987 a non-vectorizable data reference in basic-block vectorization
25988 truncate the data reference vector. Do not bother to
25989 fixup data-dependence information for gather loads.
25990 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25991 of data references, as reported.
25993 2013-03-18 Richard Biener <rguenther@suse.de>
25995 PR tree-optimization/3713
25996 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25997 has_constants and expr.
25998 (stmt_has_constants): Properly valueize SSA names when deciding
25999 whether the stmt has constants.
26001 2013-03-18 Richard Biener <rguenther@suse.de>
26003 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26004 whole function when there is nothing to do.
26005 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26006 * tree-vectorizer.c (vectorize_loops): Update virtual and
26007 loop-closed SSA once.
26008 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26010 2013-03-18 Richard Biener <rguenther@suse.de>
26012 PR middle-end/56113
26013 * domwalk.c (bb_postorder): New global static.
26014 (cmp_bb_postorder): New function.
26015 (walk_dominator_tree): Replace scheme imposing an order for
26016 visiting dominator sons by one sorting them at the time they
26017 are pushed on the stack.
26019 2013-03-18 Richard Biener <rguenther@suse.de>
26021 PR tree-optimization/39326
26022 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26023 (struct mem_ref): Replace mem member with ao_ref typed member.
26024 (MEM_ANALYZABLE): Adjust.
26025 (memref_eq): Likewise.
26026 (mem_ref_alloc): Likewise.
26027 (gather_mem_refs_stmt): Likewise.
26028 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26029 (execute_sm_if_changed_flag_set): Adjust.
26030 (execute_sm): Likewise.
26031 (ref_always_accessed_p): Likewise.
26032 (refs_independent_p): Likewise.
26033 (can_sm_ref_p): Likewise.
26035 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26038 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26039 return 1 even for !unsignedp.
26041 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26043 * config/i386/i386.md (isa): Add x64 and nox64.
26044 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26045 (*pushtf): Enable *roF alternative for x64 isa only.
26046 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26047 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26048 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26049 (*movtf_internal): Merge from *movtf_internal_rex64 and
26050 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26051 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26052 nox64 isa attributes.
26053 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26054 nox64 isa attributes.
26055 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26057 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26059 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26061 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26064 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26065 class names. Remove trailing comma after "ALL_REGS".
26067 2013-03-16 Jan Hubicka <jh@suse.cz>
26069 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26070 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26071 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26072 of cgraph_get_create_node.
26073 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26075 2013-03-16 Jason Merrill <jason@redhat.com>
26078 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26079 with DW_AT_default_value.
26081 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26083 * BASE-VER: Set to 4.9.0.
26085 2013-03-14 Andi Kleen <ak@linux.intel.com>
26088 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26089 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26090 Document _x* TSX intrinsics.
26092 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26093 David Holsgrove <david.holsgrove@xilinx.com>
26095 * configure.ac: Add MicroBlaze TLS support detection.
26096 * configure: Regenerate.
26097 * config/microblaze/microblaze-protos.h
26098 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26099 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26100 * config/microblaze/microblaze.c (microblaze_address_type): Add
26101 ADDRESS_TLS and tls_reloc address types.
26102 (microblaze_address_info): Add tls_reloc.
26103 (TARGET_HAVE_TLS): Define.
26104 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26105 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26106 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26107 load_tls_operand, microblaze_call_tls_get_addr,
26108 microblaze_legitimize_tls_address): New functions.
26109 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26110 (get_base_reg): Use microblaze_tls_symbol_p.
26111 (microblaze_classify_address): Handle TLS.
26112 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26113 label_mentioned_p and microblaze_tls_referenced_p.
26114 (microblaze_legitimize_address): Handle TLS.
26115 (microblaze_address_insns): Handle ADDRESS_TLS.
26116 (pic_address_needs_scratch): Handle TLS.
26117 (print_operand_address): Handle TLS.
26118 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26119 (microblaze_expand_move): Handle TLS.
26120 (microblaze_legitimate_constant_p): Check
26121 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26122 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26123 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26124 (PIC_OFFSET_TABLE_REGNUM): Set.
26125 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26126 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26127 (addsi3, movsi_internal2, movdf_internal): Update constraints
26128 * config/microblaze/predicates.md (arith_plus_operand): Define
26129 (move_operand): Redefine as move_src_operand,
26130 check microblaze_tls_referenced_p.
26132 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26134 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26135 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26137 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26139 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26142 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26144 PR tree-optimization/53265
26145 * common.opt (Waggressive-loop-optimizations): New option.
26146 * tree-ssa-loop-niter.c: Include tree-pass.h.
26147 (do_warn_aggressive_loop_optimizations): New function.
26148 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26149 if number_of_latch_executions returned constant.
26150 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26151 early. If number_of_latch_executions returned constant, set
26152 nb_iterations_upper_bound back to it.
26153 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26155 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26156 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26158 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26159 (MULTILIB_OSDIRNAMES): Set.
26160 * genmultilib: If defaultosdirname doesn't start with :: , set
26161 defaultosdirname2 instead, clear it and emit two . multilib_raw
26162 entries instead of just one.
26164 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26166 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26167 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26168 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26169 (SUBTARGET_OVERRIDE_OPTIONS): New.
26171 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26174 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26175 (musermode): Convert to Var(TARGET_USERMODE).
26176 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26177 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26178 * config/sh/sh.c (sh_option_override): Use
26179 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26180 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26182 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26184 (udivsi3_i4_single, divsi3_i4_single): Use
26185 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26187 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26189 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26192 2013-03-13 Richard Biener <rguenther@suse.de>
26194 PR tree-optimization/56608
26195 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26196 calls when vectorizing basic-blocks.
26198 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26201 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26204 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26206 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26208 2013-03-11 Jan Hubicka <jh@suse.cz>
26211 * lto-streamer-out.c (output_symbol_p): Skip references from
26212 constructors of external variables.
26214 2013-03-11 Jan Hubicka <jh@suse.cz>
26216 PR middle-end/56571
26217 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26219 * emit-rtl.c (verify_rtx_sharing): Likewise.
26220 (copy_insn_1): Likewise.
26221 * rtl.c (copy_rtx): Likewise.
26223 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26226 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26227 output_operand_lossage message.
26229 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26232 * arm.c (shift_op): Validate RTL pattern on the fly.
26233 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26236 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26239 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26240 new call patterns that clobber %fr12.
26241 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26242 split and postreload patterns.
26243 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26244 registers %fr12 and %fr12R as call used.
26246 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26248 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26249 (canon_address, record_store, replace_read, check_mem_read_rtx,
26250 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26251 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26252 rest_of_handle_dse): Likewise.
26254 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26256 PR middle-end/56524
26257 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26259 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26260 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26261 (save_optabs_if_changed): Replace with...
26262 (init_tree_optimization_optabs): ...this.
26263 * optabs.c (save_optabs_if_changed): Rename to...
26264 (init_tree_optimization_optabs): ...this. Take the optimization node
26265 as argument. Do nothing if the base optabs are already correct.
26266 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26267 to recompute optabs.
26268 * function.h (function): Remove optabs field.
26269 * function.c (invoke_set_current_function_hook): Call
26270 init_tree_optimization_optabs. Use the result to initialize
26273 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26275 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26276 if GTMA_HAS_NO_INSTRUMENTATION.
26277 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26278 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26279 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26280 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26281 GTMA_HAS_NO_INSTRUMENTATION.
26283 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26285 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26288 2013-03-08 Marek Polacek <polacek@redhat.com>
26289 Jakub Jelinek <jakub@redhat.com>
26291 PR tree-optimization/56478
26292 * predict.c (is_comparison_with_loop_invariant_p): Change the
26293 type of loop_step to tree.
26294 (predict_loops): Adjust.
26295 (predict_iv_comparison): Perform the computations on double_ints.
26297 2013-03-08 Richard Biener <rguenther@suse.de>
26299 PR tree-optimization/56570
26300 * tree-cfg.c (verify_expr_location_1): Verify locations for
26302 * tree-sra.c (create_access_replacement): Strip locations
26303 from DECL_DEBUG_EXPRs.
26305 2013-03-08 Richard Biener <rguenther@suse.de>
26307 * tree-inline.c (expand_call_inline): Do not associate
26308 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
26309 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
26311 2013-03-08 Richard Biener <rguenther@suse.de>
26313 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
26314 or block changes with -Og. Fix for location / block encoding
26315 changes and PHI arguments with locations.
26317 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
26319 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
26321 (struct output_info): Likewise.
26322 (register_overhead): Remove bad gcc_assert.
26323 (bitmap_find_bit): If there is only a single bitmap element, do not
26324 count a miss as a search.
26325 (print_statistics): Update for counter type changes.
26326 (dump_bitmap_statistics): Likewise. Print headers such that they
26327 are properly lined up with the printed counters.
26329 2013-03-07 Jakub Jelinek <jakub@redhat.com>
26331 PR tree-optimization/56559
26332 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
26333 check that it has only a single use.
26335 2013-03-07 Richard Biener <rguenther@suse.de>
26337 * doc/invoke.texi (fwhole-program): Discourage use in combination
26340 2013-03-06 Jakub Jelinek <jakub@redhat.com>
26342 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
26344 PR tree-optimization/56539
26345 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
26346 instead of GSI_CONTINUE_LINKING as last argument to
26347 force_gimple_operand_gsi. Adjust function comment.
26349 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
26352 PR middle-end/56548
26353 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
26354 promoted mode, convert the result back to the original mode.
26356 2013-03-06 Richard Biener <rguenther@suse.de>
26358 PR middle-end/56294
26359 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26360 (insert_updated_phi_nodes_compare_uids): New function.
26361 (update_ssa): Sort symbols_to_rename after UID before
26362 traversing it to insert PHI nodes.
26364 2013-03-06 Richard Biener <rguenther@suse.de>
26366 PR middle-end/50494
26367 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26368 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26371 2013-02-13 Richard Biener <rguenther@suse.de>
26374 * varasm.c (output_constant_def_1): Get the decl representing
26375 the constant as argument.
26376 (output_constant_def): Wrap output_constant_def_1.
26377 (make_decl_rtl): Use output_constant_def_1 with the decl
26378 representing the constant.
26379 (build_constant_desc): Optionally re-use a decl already
26380 representing the constant.
26381 (tree_output_constant_def): Adjust.
26383 2013-03-06 Joey Ye <joey.ye@arm.com>
26386 * gcc.c (convert_white_space): New function.
26387 (main): Handles white space in function name.
26389 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26392 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26393 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26394 to SH_DIV_CALL_TABLE for TARGET_SH2.
26395 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26397 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26398 call-table options.
26400 2013-03-05 Sterling Augustine <saugustine@google.com>
26401 Cary Coutant <ccoutant@google.com>
26404 * dwarf2out.c (resolve_addr): Don't call
26405 remove_loc_list_addr_table_entries a second time for the same
26408 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26411 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26412 (avoid_complex_debug_insns): New function.
26413 (expand_debug_locations): Call it.
26415 PR rtl-optimization/56484
26416 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26417 lifetimes of hard registers on small register class machines.
26419 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26421 * config/microblaze/microblaze-protos.h: Rename
26422 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26423 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26425 (microblaze_fast_interrupt_function_p): New function.
26426 (microblaze_is_interrupt_handler): Rename to
26427 microblaze_is_interrupt_variant and add fast_interrupt check.
26428 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26429 (save_restore_insns): Likewise.
26430 (compute_frame_size): Likewise.
26431 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26432 (microblaze_globalize_label): Likewise.
26433 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26434 * config/microblaze/microblaze.md: Use wrapper
26435 microblaze_is_interrupt_variant.
26437 2013-03-05 Kai Tietz <ktietz@redhat.com>
26439 * sdbout.c (sdbout_one_type): Switch to current function's section
26440 supporting cold/hot.
26442 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26444 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26447 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26449 PR middle-end/56461
26450 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26451 if VALGRIND_GET_VBITS is defined, temporarily make object
26452 memory all defined, and restore previous valgrind addressability
26453 and definability afterwards. Free this_object at the end.
26455 PR middle-end/56461
26456 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26457 right before calling lra_create_live_ranges, also call it
26458 when clearing live_p. Only call lra_clear_live_ranges
26459 at the end if live_p.
26461 PR middle-end/56461
26462 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26464 2013-03-05 Richard Biener <rguenther@suse.de>
26466 PR tree-optimization/56521
26467 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26470 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26473 * except.h (remove_unreachable_eh_regions): New prototype.
26474 * except.c (remove_eh_handler_splicer): New function, split out
26475 of remove_eh_handler.
26476 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26477 warning about running it on many EH regions one at a time.
26478 (remove_unreachable_eh_regions_worker): New function, walk the
26479 EH tree in depth-first order and remove non-marked regions.
26480 (remove_unreachable_eh_regions): New function.
26481 * tree-eh.c (mark_reachable_handlers): New function, split out
26482 from remove_unreachable_handlers.
26483 (remove_unreachable_handlers): Use mark_reachable_handlers and
26484 remove_unreachable_eh_regions.
26485 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26486 and remove_unreachable_eh_regions.
26488 2013-03-05 Richard Biener <rguenther@suse.de>
26490 PR middle-end/56525
26491 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26492 not freeing them until the end.
26494 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26496 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26498 2013-03-05 Richard Biener <rguenther@suse.de>
26500 PR tree-optimization/56270
26501 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26502 of loads after scheduling an SLP instance.
26504 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26506 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26508 (check_gcc_parallelize): Run guality.exp as a separate job from
26509 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26510 struct-layout-1.exp with stackalign.exp.
26512 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26514 PR middle-end/56461
26515 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26516 load_index sbitmap even if some bit in it isn't set.
26518 PR middle-end/56461
26519 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26520 (discover_iteration_bound_by_body_walk): Change queues to
26521 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26522 spelling in comment. Call safe_push on queues[bound_index] directly.
26523 Release queues[queue_index] in every iteration unconditionally.
26524 Release bounds vector.
26526 PR middle-end/56461
26527 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26528 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26529 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26532 2013-03-05 Richard Biener <rguenther@suse.de>
26535 * tree-inline.c (remap_blocks_to_null): New function.
26536 (expand_call_inline): When expanding a call stmt without
26537 an associated block inline remap all callee blocks to NULL.
26539 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26541 PR rtl-optimization/56494
26542 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26543 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26544 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26546 PR middle-end/56461
26547 * sel-sched-ir.c (free_sched_pools): Release
26548 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26549 if succs_info_pool.max_top isn't -1.
26552 * opts.c (opts_obstack, opts_concat): Moved to...
26553 * opts-common.c (opts_obstack, opts_concat): ... here.
26555 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26557 PR middle-end/56461
26558 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26560 2013-03-04 Martin Jambor <mjambor@suse.cz>
26562 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26563 all appropriate places.
26565 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26567 PR tree-optimization/56424
26568 * ipa-split.c (split_function): Do not set the RSO flag if result is
26569 not by reference and its type is a register type.
26571 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26573 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26574 (microblaze_legitimate_pic_operand): Likewise
26575 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26576 new function microblaze_legitimate_pic_operand
26577 * config/microblaze/microblaze-protos.h
26578 (microblaze_legitimate_pic_operand): Declare.
26580 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26582 * config/microblaze/predicates.md (call_insn_simple_operand):
26583 New predicate for supported rtx code types.
26584 * config/microblaze/microblaze.md (call_internal1): Use
26585 call_insn_simple_operand predicate.
26587 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26589 PR middle-end/56461
26590 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26591 partitions.ordered_remove.
26593 PR middle-end/56461
26594 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26595 vec_oprnds0.create (1) for modifier == NONE.
26597 PR middle-end/56461
26598 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26599 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26600 vec_oprnds1 right before pushing anything to it for
26603 PR middle-end/56461
26604 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26605 set nbbs to 0 instead of having separate code path.
26606 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26607 instead of false as last argument if returning NULL.
26609 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26611 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26612 the attribute is now called "target" instead of "option".
26613 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26614 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26615 attribute/pragma name for TARGET_OPTION_VALID_P and
26616 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26617 * doc/tm.texi: Regenerated.
26619 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26621 * config/microblaze/microblaze.c:
26622 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26623 * config/microblaze/microblaze.h: Add -mxl-reorder to
26625 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26626 instructions emitted if TARGET_REORDER.
26627 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26628 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26631 2013-03-01 Xinliang David Li <davidxl@google.com>
26633 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26636 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26638 PR middle-end/56461
26639 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26640 vector even when returning true. Fix up function comment formatting.
26642 PR middle-end/56461
26643 * ira-build.c (ira_loop_nodes_count): New variable.
26644 (create_loop_tree_nodes): Initialize it.
26645 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26647 PR middle-end/56461
26648 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26649 method on dr_chain and result_chain.
26650 * tree-vect-stmts.c (vectorizable_store): Only call
26651 result_chain.create if j == 0.
26653 PR middle-end/56461
26654 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26655 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26656 before overwriting it.
26658 2013-03-01 Tobias Burnus <burnus@net-b.de>
26660 * doc/extended.texi (C Extensions): Change order in @menu
26662 (Other MIPS Built-in Functions): Move last MIPS entry before
26663 "picoChip Built-in Functions".
26664 (SH Built-in Functions): Move after RX Built-in Functions.
26665 * doc/gcc.texi (Introduction): Change order in @menu
26667 * doc/md.texi (Constraints): Ditto.
26668 * gty.texi (Type Information): Ditto.
26669 (User-provided marking routines for template types): Make
26671 * doc/invoke.texi (AArch64 Options): Move before
26672 "Adapteva Epiphany Options".
26674 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26675 Jakub Jelinek <jakub@redhat.com>
26678 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26679 no_address_safety_analysis attribute.
26680 * doc/extend.texi (no_address_safety_attribute): Rename to
26681 no_sanitize_address attribute, mention no_address_safety_analysis
26682 attribute as deprecated alias.
26684 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26686 PR middle-end/56461
26687 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26688 type to vec<vec<tree> > *.
26689 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26690 to be vec<tree> instead of vec<tree> *, set vec_defs
26691 to vNULL and call vec_defs.create (number_of_vects), adjust other
26693 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26694 vectorizable_condition): Adjust vect_get_slp_defs callers.
26696 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26698 * config/aarch64/aarch64.c
26699 (aarch64_float_const_representable): Remove unused variable.
26701 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26703 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26705 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26707 * config/aarch64/aarch64-builtins.c
26708 (aarch64_init_simd_builtins): Make static.
26710 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26712 * config/aarch64/aarch64.c
26713 (aarch64_simd_make_constant): Make static.
26715 2013-02-28 Martin Jambor <mjambor@suse.cz>
26717 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26718 with no initialization to the RHS of debug statements.
26720 2013-02-28 Martin Jambor <mjambor@suse.cz>
26722 PR tree-optimization/56294
26723 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26725 (get_access_replacement): Do not call create_access_replacement.
26726 Assert a replacement exists.
26727 (get_repl_default_def_ssa_name): Create the replacement declaration
26730 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26732 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26733 final_end_function.
26735 2013-02-28 Marek Polacek <polacek@redhat.com>
26737 PR rtl-optimization/56466
26738 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26739 if we're changing a loop.
26740 (peel_loops_completely): Likewise.
26742 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26745 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26747 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26750 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26751 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26752 INTX_FTYPE_FX, FX_FTYPE_INTX.
26753 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26755 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26757 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26758 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26759 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26760 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26761 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26762 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26763 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26764 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26765 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26766 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26767 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26768 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26769 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26770 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26771 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26772 (avrxmega6): Increase max flash segments from 5 to 6.
26773 * config/avr/t-multilib: Regenerate.
26774 * config/avr/avr-tables.opt: Regenerate.
26775 * doc/avr-mmcu.texi: Regenerate.
26777 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26779 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26780 (avr_device_to_arch): Rename to avr_device_to_ld.
26781 (avr_device_to_as): New prototype.
26782 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26783 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26784 * config/avr/driver-avr.c (avr_device_to_as): New.
26785 (avr_device_to_arch): Rename to avr_device_to_ld.
26787 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26789 PR middle-end/56461
26790 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26791 method on dr_chain and result_chain.
26793 PR middle-end/56461
26794 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26795 pointer_set_destroy on not_executed_last_iteration.
26797 PR middle-end/56461
26798 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26800 PR middle-end/56461
26801 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26802 FOR_EACH_DEFINED_FUNCTION when freeing state.
26804 PR middle-end/56461
26805 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26807 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26810 PR middle-end/56461
26811 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26812 known_aggs[i].items and release known_aggs vector.
26814 PR middle-end/56461
26815 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26817 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26819 * config/microblaze/microblaze.c (microblaze_emit_compare):
26820 Use xor for EQ/NE comparisions.
26821 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26822 (cbranchsf4): Adjust operator to comparison_operator.
26824 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26826 PR middle-end/56461
26827 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26829 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26830 vec_safe_push, always update *slot.
26831 (redirect_edge_var_map_clear): Use vec_free.
26832 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26833 (free_var_map_entry): Use vec_free.
26834 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26835 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26837 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26839 PR middle-end/45472
26840 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26841 when the may_trap_p bit of the exprs being merged differs.
26842 Reorder tests for speculativeness in the logical and operator.
26844 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26846 * incpath.c (add_standard_paths): Use reconcat instead of concat
26847 where appropriate and avoid leaking memory.
26849 * opts.h: Include obstack.h.
26850 (opts_concat): New prototype.
26851 (opts_obstack): New declaration.
26852 * opts.c (opts_concat): New function.
26853 (opts_obstack): New variable.
26854 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26855 (finish_options): Use opts_concat instead of concat
26856 and XOBNEWVEC instead of XNEWVEC.
26857 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26858 generate_option): Likewise.
26859 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26860 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26863 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26864 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26866 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26868 PR middle-end/56461
26869 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26871 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26873 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26874 (arm_block_move_unaligned_straight): Likewise.
26875 (arm_adjust_block_mem): Likewise.
26877 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26880 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26881 temp, cond and label.
26882 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26885 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26886 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26889 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26890 of prologue/epilogue functions.
26893 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26894 Remove unused variable cfa_offset.
26895 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26898 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26899 type promotion to unsigned.
26902 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26903 for HOST_WIDE_INT of 32 bit / same size as int.
26904 (arm_block_move_unaligned_straight): Likewise.
26905 (arm_adjust_block_mem): Likewise.
26908 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26911 2013-02-26 Marek Polacek <polacek@redhat.com>
26913 PR tree-optimization/56426
26914 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26916 2013-02-26 Richard Biener <rguenther@suse.de>
26919 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26920 unused variable loops.
26922 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26924 PR tree-optimization/56448
26925 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26926 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26927 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26928 later operands of the references, or even first operand for
26929 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26931 PR tree-optimization/56443
26932 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26933 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26934 to type_for_mode langhook.
26936 2013-02-25 Matt Turner <mattst88@gmail.com>
26938 * doc/invoke.texi: Document r4700.
26940 2013-02-25 Richard Biener <rguenther@suse.de>
26942 PR tree-optimization/56175
26943 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26945 (simplify_bitwise_binary): ... here. Also guard the conversion
26946 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26948 2013-02-25 Catherine Moore <clm@codesourcery.com>
26951 2013-02-24 Catherine Moore <clm@codesourcery.com>
26952 Maciej W. Rozycki <macro@codesourcery.com>
26953 Tom de Vries <tom@codesourcery.com>
26954 Nathan Sidwell <nathan@codesourcery.com>
26955 Iain Sandoe <iain@codesourcery.com>
26956 Nathan Froyd <froydnj@codesourcery.com>
26957 Chao-ying Fu <fu@mips.com>
26959 * doc/extend.texi: (micromips, nomicromips, nocompression):
26960 Document new function attributes.
26961 * doc/invoke.texi (minterlink-compressed, mmicromips,
26962 m14k, m14ke, m14kec): Document new options.
26963 (minterlink-mips16): Update documentation.
26964 * doc/md.texi (ZC, ZD): Document new constraints.
26965 * configure.ac (gcc_cv_as_micromips): Check if linker
26966 supports the .set micromips directive.
26967 * configure: Regenerate.
26968 * config.in: Regenerate.
26969 * config/mips/mips-tables.opt: Regenerate.
26970 * config/mips/micromips.md: New file.
26971 * constraints.md (ZC, AD): New constraints.
26972 * config/mips/predicates.md (movep_src_register): New predicate.
26973 (movep_src_operand): New predicate.
26974 (non_volatile_mem_operand): New predicate.
26975 * config/mips/mips.md (multimem): New type.
26976 (length): Differentiate between 17-bit and 18-bit branch offsets.
26977 (MOVEP1, MOVEP2): New mode iterator.
26978 (mov_<load>l): Use ZC constraint.
26979 (mov_<load>r): Likewise.
26980 (mov_<store>l): Likewise.
26981 (mov_<store>r): Likewise.
26982 (*branch_equality<mode>_inverted): Add microMIPS support.
26983 (*branch_equality<mode>): Likewise.
26984 (*jump_absolute): Likewise.
26985 (indirect_jump_<mode>): Likewise.
26986 (tablejump_<mode>): Likewise.
26987 (<optab>_internal): Likewise.
26988 (sibcall_internal): Likewise.
26989 (sibcall_value_internal): Likewise.
26990 (prefetch): Use constraint ZD.
26991 * config/mips/mips.opt (minterlink-compressed): New option.
26992 (minterlink-mips16): Now an alias for minterlink-compressed.
26993 (mmicromips): New option.
26994 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26995 (compare_and_swap_12): Likewise.
26996 (sync_add<mode>): Likewise.
26997 (sync_<optab>_12): Likewise.
26998 (sync_old_<optab>_12): Likewise.
26999 (sync_new_<optab>_12): Likewise.
27000 (sync_nand_12): Likewise.
27001 (sync_old_nand_12): Likewise.
27002 (sync_new_nand_12): Likewise.
27003 (sync_sub<mode>): Likewise.
27004 (sync_old_add<mode>): Likewise.
27005 (sync_old_sub<mode>): Likewise.
27006 (sync_new_add<mode>): Likewise.
27007 (sync_new_sub<mode>): Likewise.
27008 (sync_<optab><mode>): Likewise.
27009 (sync_old_<optab><mode>): Likewise.
27010 (sync_new_<optab><mode>): Likewise.
27011 (sync_nand<mode>): Likewise.
27012 (sync_old_nand<mode>): Likewise.
27013 (sync_new_nand<mode>): Likewise.
27014 (sync_lock_test_and_set<mode>): Likewise.
27015 (test_and_set_12): Likewise.
27016 (atomic_compare_and_swap<mode>): Likewise.
27017 (atomic_exchange<mode>_llsc): Likewise.
27018 (atomic_fetch_add<mode>_llsc): Likewise.
27019 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27020 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27021 (umips_save_restore_pattern_p): Likewise.
27022 (umips_load_store_pair_p): Likewise.
27023 (umips_output_load_store_pair): Likewise.
27024 (umips_movep_target_p): Likewise.
27025 (umips_12bit_offset_address_p): Likewise.
27026 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27027 (mips_base_mips16): Rename this...
27028 (mips_base_compression_flags): ...to this. Update all uses.
27029 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27030 (mips_mips16_decl_p): Delete.
27031 (mips_nomips16_decl_p): Delete.
27032 (mips_get_compress_on_flags): New function.
27033 (mips_get_compress_off_flags): New function.
27034 (mips_get_compress_mode): New function.
27035 (mips_get_compress_on_name): New function.
27036 (mips_get_compress_off_name): New function.
27037 (mips_insert_attributes): Support multiple compression types.
27038 (mips_merge_decl_attributes): Likewise.
27039 (umips_12bit_offset_address_p): New function.
27040 (mips_start_function_definition): Emit .set micromips directive.
27041 (mips_call_may_need_jalx_p): New function.
27042 (mips_function_ok_for_sibcall): Add microMIPS support.
27043 (mips_print_operand_punctuation): Support short delay slots and
27045 (umips_swm_mask, umips_swm_encoding): New.
27046 (umips_build_save_restore): New function.
27047 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27048 (was_mips16_p): Remove.
27049 (old_compression_mode): New.
27050 (mips_set_compression_mode): New function.
27051 (mips_set_current_function): Add microMIPS support.
27052 (mips_option_override): Likewise.
27053 (umips_save_restore_pattern_p): New function.
27054 (umips_output_save_restore): New function.
27055 (umips_load_store_pair_p_1): New function.
27056 (umips_load_store_pair_p): New function.
27057 (umips_output_load_store_pair_1): New function.
27058 (umips_output_load_store_pair): New function.
27059 (umips_movep_target_p) New function.
27060 (mips_prepare_pch_save): Add microMIPS support.
27061 * config/mips/mips.h (TARGET_COMPRESSION): New.
27062 (TARGET_CPU_CPP_BUILTINS): Update macro
27063 to use new compression flags and to support microMIPS.
27064 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27065 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27066 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27067 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27068 (ASM_SPEC): Support mmicromips and mno-micromips.
27069 (M16STORE_REG_P): New macro.
27070 (MIPS_CALL): Support TARGET_MICROMIPS.
27071 (MICROMIPS_J): New macro.
27072 (mips_base_mips16): Rename this...
27073 (mips_base_compression_flags): ...to this.
27074 (UMIPS_12BIT_OFFSET_P): New macro.
27075 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27076 (MULTILIB_DIRNAMES): Likewise.
27078 2013-02-25 Tom de Vries <tom@codesourcery.com>
27080 PR rtl-optimization/56131
27081 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27082 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27083 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27085 2013-02-25 Tobias Burnus <burnus@net-b.de>
27087 * doc/invoke.texi (-fsanitize=): Move from optimization
27088 to debugging options.
27090 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27092 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27094 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27095 Alexander Monakov <amonakov@ispras.ru>
27097 PR middle-end/56077
27098 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27099 flush pending lists also on non-jumps. Adjust comment.
27101 2013-02-24 Catherine Moore <clm@codesourcery.com>
27102 Maciej W. Rozycki <macro@codesourcery.com>
27103 Tom de Vries <tom@codesourcery.com>
27104 Nathan Sidwell <nathan@codesourcery.com>
27105 Iain Sandoe <iain@codesourcery.com>
27106 Nathan Froyd <froydnj@codesourcery.com>
27107 Chao-ying Fu <fu@mips.com>
27109 * doc/extend.texi: (micromips, nomicromips, nocompression):
27110 Document new function attributes.
27111 * doc/invoke.texi (minterlink-compressed, mmicromips,
27112 m14k, m14ke, m14kec): Document new options.
27113 (minterlink-mips16): Update documentation.
27114 * doc/md.texi (ZC, ZD): Document new constraints.
27115 * configure.ac (gcc_cv_as_micromips): Check if linker
27116 supports the .set micromips directive.
27117 * configure: Regenerate.
27118 * config.in: Regenerate.
27119 * config/mips/mips-tables.opt: Regenerate.
27120 * config/mips/micromips.md: New file.
27121 * constraints.md (ZC, AD): New constraints.
27122 * config/mips/predicates.md (movep_src_register): New predicate.
27123 (movep_src_operand): New predicate.
27124 (non_volatile_mem_operand): New predicate.
27125 * config/mips/mips.md (multimem): New type.
27126 (length): Differentiate between 17-bit and 18-bit branch offsets.
27127 (MOVEP1, MOVEP2): New mode iterator.
27128 (mov_<load>l): Use ZC constraint.
27129 (mov_<load>r): Likewise.
27130 (mov_<store>l): Likewise.
27131 (mov_<store>r): Likewise.
27132 (*branch_equality<mode>_inverted): Add microMIPS support.
27133 (*branch_equality<mode>): Likewise.
27134 (*jump_absolute): Likewise.
27135 (indirect_jump_<mode>): Likewise.
27136 (tablejump_<mode>): Likewise.
27137 (<optab>_internal): Likewise.
27138 (sibcall_internal): Likewise.
27139 (sibcall_value_internal): Likewise.
27140 (prefetch): Use constraint ZD.
27141 * config/mips/mips.opt (minterlink-compressed): New option.
27142 (minterlink-mips16): Now an alias for minterlink-compressed.
27143 (mmicromips): New option.
27144 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27145 (compare_and_swap_12): Likewise.
27146 (sync_add<mode>): Likewise.
27147 (sync_<optab>_12): Likewise.
27148 (sync_old_<optab>_12): Likewise.
27149 (sync_new_<optab>_12): Likewise.
27150 (sync_nand_12): Likewise.
27151 (sync_old_nand_12): Likewise.
27152 (sync_new_nand_12): Likewise.
27153 (sync_sub<mode>): Likewise.
27154 (sync_old_add<mode>): Likewise.
27155 (sync_old_sub<mode>): Likewise.
27156 (sync_new_add<mode>): Likewise.
27157 (sync_new_sub<mode>): Likewise.
27158 (sync_<optab><mode>): Likewise.
27159 (sync_old_<optab><mode>): Likewise.
27160 (sync_new_<optab><mode>): Likewise.
27161 (sync_nand<mode>): Likewise.
27162 (sync_old_nand<mode>): Likewise.
27163 (sync_new_nand<mode>): Likewise.
27164 (sync_lock_test_and_set<mode>): Likewise.
27165 (test_and_set_12): Likewise.
27166 (atomic_compare_and_swap<mode>): Likewise.
27167 (atomic_exchange<mode>_llsc): Likewise.
27168 (atomic_fetch_add<mode>_llsc): Likewise.
27169 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27170 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27171 (umips_save_restore_pattern_p): Likewise.
27172 (umips_load_store_pair_p): Likewise.
27173 (umips_output_load_store_pair): Likewise.
27174 (umips_movep_target_p): Likewise.
27175 (umips_12bit_offset_address_p): Likewise.
27176 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27177 (mips_base_mips16): Rename this...
27178 (mips_base_compression_flags): ...to this. Update all uses.
27179 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27180 (mips_mips16_decl_p): Delete.
27181 (mips_nomips16_decl_p): Delete.
27182 (mips_get_compress_on_flags): New function.
27183 (mips_get_compress_off_flags): New function.
27184 (mips_get_compress_mode): New function.
27185 (mips_get_compress_on_name): New function.
27186 (mips_get_compress_off_name): New function.
27187 (mips_insert_attributes): Support multiple compression types.
27188 (mips_merge_decl_attributes): Likewise.
27189 (umips_12bit_offset_address_p): New function.
27190 (mips_start_function_definition): Emit .set micromips directive.
27191 (mips_call_may_need_jalx_p): New function.
27192 (mips_function_ok_for_sibcall): Add microMIPS support.
27193 (mips_print_operand_punctuation): Support short delay slots and
27195 (umips_swm_mask, umips_swm_encoding): New.
27196 (umips_build_save_restore): New function.
27197 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27198 (was_mips16_p): Remove.
27199 (old_compression_mode): New.
27200 (mips_set_compression_mode): New function.
27201 (mips_set_current_function): Add microMIPS support.
27202 (mips_option_override): Likewise.
27203 (umips_save_restore_pattern_p): New function.
27204 (umips_output_save_restore): New function.
27205 (umips_load_store_pair_p_1): New function.
27206 (umips_load_store_pair_p): New function.
27207 (umips_output_load_store_pair_1): New function.
27208 (umips_output_load_store_pair): New function.
27209 (umips_movep_target_p) New function.
27210 (mips_prepare_pch_save): Add microMIPS support.
27211 * config/mips/mips.h (TARGET_COMPRESSION): New.
27212 (TARGET_CPU_CPP_BUILTINS): Update macro
27213 to use new compression flags and to support microMIPS.
27214 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27215 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27216 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27217 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27218 (ASM_SPEC): Support mmicromips and mno-micromips.
27219 (M16STORE_REG_P): New macro.
27220 (MIPS_CALL): Support TARGET_MICROMIPS.
27221 (MICROMIPS_J): New macro.
27222 (mips_base_mips16): Rename this...
27223 (mips_base_compression_flags): ...to this.
27224 (UMIPS_12BIT_OFFSET_P): New macro.
27225 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27226 (MULTILIB_DIRNAMES): Likewise.
27228 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27231 * target-globals.c (save_target_globals): For init_reg_sets and
27232 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27234 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27236 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27237 * config/aarch64/t-aarch64
27238 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27240 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27242 PR inline-asm/56148
27243 * lra-constraints.c (process_alt_operands): Reload operand
27244 conflicting with earlier clobber only if no more other conflicting
27247 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27250 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27251 if not linking a shared library.
27253 2013-02-22 Seth LaForge <sethml@google.com>
27255 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27257 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27259 * config/arm/arm.md (split for extendsidi): Update condition.
27260 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27261 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27262 (qhs_zextenddi_cstr): Likewise.
27264 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27266 PR middle-end/56420
27267 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27268 avoid signed wrapping.
27269 (expand_mult): Handle properly multiplication by
27270 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27271 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27272 in the compiler if coeff is HOST_WIDE_INT_MIN.
27273 (expand_divmod): Don't make ext_op1 static, change it's type to
27274 uhwi. Avoid undefined behavior in -INTVAL (op1).
27276 PR rtl-optimization/50339
27277 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27279 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27280 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27281 into splitting_ashiftrt field.
27282 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27284 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27287 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
27289 PR middle-end/56108
27290 * trans-mem.c (execute_tm_mark): Do not expand transactions that
27291 are sure to go irrevocable.
27293 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
27295 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
27296 scalars are valid operands.
27298 2013-02-21 Martin Jambor <mjambor@suse.cz>
27300 PR tree-optimization/56310
27301 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
27302 only matching indices and non-negative final offsets.
27303 (intersect_aggregates_with_edge): Pass src_idx to
27304 agg_replacements_to_vector. Pass src_idx insstead of index to
27305 intersect_with_agg_replacements.
27307 2013-02-21 Martin Jambor <mjambor@suse.cz>
27309 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
27310 instead of hard-wired defaults.
27312 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
27314 * doc/invoke.texi (MIPS Options): Update documentation of the
27315 floating-point multiply-accumulate instruction restrictions.
27317 2013-02-21 Kostya Serebryany <kcc@google.com>
27319 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
27320 asan_shadow_offset on x86_64 linux.
27322 2013-02-21 Richard Biener <rguenther@suse.de>
27324 PR tree-optimization/56415
27326 2013-02-11 Richard Biener <rguenther@suse.de>
27328 PR tree-optimization/56273
27329 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27332 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27335 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
27338 PR inline-asm/56405
27339 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
27340 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
27342 2013-02-20 Jan Hubicka <jh@suse.cz>
27344 PR tree-optimization/56265
27345 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
27346 when target is referenced for first time.
27348 2013-02-20 Richard Biener <rguenther@suse.de>
27350 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
27351 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
27352 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
27353 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
27354 not return anything.
27355 (rename_ssa_copies): Do not remove unused locals.
27356 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27357 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27358 * passes.c (execute_function_todo): Do not schedule unused locals
27359 removal if cleanup_tree_cfg did something.
27360 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27361 about the number of removed locals.
27363 2013-02-20 Richard Biener <rguenther@suse.de>
27365 PR tree-optimization/56398
27366 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27368 2013-02-20 Martin Jambor <mjambor@suse.cz>
27370 PR tree-optimization/55334
27371 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27372 restricted pointers to arrays.
27374 2013-02-20 Richard Biener <rguenther@suse.de>
27375 Jakub Jelinek <jakub@redhat.com>
27377 PR tree-optimization/56396
27378 * tree-ssa-ccp.c (n_const_val): New static variable.
27379 (get_value): Return NULL for SSA names we don't have a lattice
27381 (ccp_initialize): Initialize n_const_val.
27382 * tree-ssa-copy.c (n_copy_of): New static variable.
27383 (init_copy_prop): Initialize n_copy_of.
27384 (get_value): Return NULL_TREE for SSA names we don't have a
27387 2013-02-20 Martin Jambor <mjambor@suse.cz>
27389 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27391 2013-02-20 Richard Biener <rguenther@suse.de>
27393 * genpreds.c (write_lookup_constraint): Do not compare first
27394 letter of the constraint again.
27396 2013-02-20 Richard Biener <rguenther@suse.de>
27398 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27400 (get_use_iv_cost): Terminate hashtable walk when coming across
27403 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27405 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27406 reassociation for avx2 targets.
27408 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27410 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27411 Add version check for v8.10.a to enable microblaze_has_clz
27412 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27413 version and TARGET_PATTERN_COMPARE check
27414 * config/microblaze/microblaze.md: New clzsi2 instruction
27416 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27418 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27419 function before branching.
27421 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27423 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27425 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27427 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27429 PR middle-end/55889
27430 * sel-sched.c: Include ira.h.
27431 (implicit_clobber_conflict_p): New function.
27432 (moveup_expr): Use it.
27433 * Makefile.in (sel-sched.o): Depend on ira.h.
27435 2013-02-19 Richard Biener <rguenther@suse.de>
27437 PR tree-optimization/56384
27438 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27439 (vn_hash_type): Split out from ...
27440 (vn_hash_constant_with_type): ... here.
27441 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27442 (vn_phi_eq): Compare types from vn_phi_s structure.
27443 (vn_phi_lookup): Populate vn_phi_s type.
27444 (vn_phi_insert): Likewise.
27446 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27448 PR tree-optimization/56350
27449 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27450 if haven't found reduction or nested cycle operand, rather than
27451 asserting we must find it.
27453 PR tree-optimization/56381
27454 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27457 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27458 Jakub Jelinek <jakub@redhat.com>
27461 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27462 (swap_optab_enable): Same.
27463 (init_all_optabs): Use argument instead of global.
27464 * tree.h (struct tree_optimization_option): New field target_optabs.
27465 * expr.h (init_all_optabs): Add argument to prototype.
27466 (TREE_OPTIMIZATION_OPTABS): New.
27467 (save_optabs_if_changed): Protoize.
27468 * optabs.h: Declare this_fn_optabs.
27469 * optabs.c (save_optabs_if_changed): New.
27470 Declare this_fn_optabs.
27471 (init_optabs): Add argument to init_all_optabs() call.
27472 * function.c (invoke_set_current_function_hook): Handle per
27474 * function.h (struct function): New field optabs.
27475 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27476 optimization_current_node has changed.
27477 * target-globals.h (save_target_globals_default_opts): Protoize.
27478 * target-globals.c (save_target_globals_default_opts): New.
27480 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27483 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27484 registers %fr12 and %fr12R as call used.
27487 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27488 and HImode, require all displacements to be an integer multiple of
27490 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27491 only allow QImode and HImode when reload is in progress and strict is
27492 true. Likewise for symbolic addresses. Use base14_operand to check
27493 displacements in REG+BASE addresses.
27495 2013-02-18 Richard Biener <rguenther@suse.de>
27497 PR tree-optimization/56366
27498 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27499 handle sign-conversion of outer-loop initial induction value.
27501 2013-02-18 Richard Biener <rguenther@suse.de>
27503 PR middle-end/56349
27504 * cfghooks.c (merge_blocks): If we merge a latch into another
27505 block adjust references to it.
27506 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27507 (verify_loop_structure): Verify that a recorded latch is in fact
27510 2013-02-18 Richard Biener <rguenther@suse.de>
27512 PR tree-optimization/56321
27513 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27514 order SSA name release and virtual operand unlinking.
27516 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27518 * config/microblaze/microblaze.md (save_stack_block): Define.
27519 (restore_stack_block): Likewise.
27521 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27523 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27524 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27525 * config/microblaze/microblaze.c (microblaze_option_override):
27526 Bail out early for PIC modes when target does not support PIC.
27528 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27530 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27531 Replace with a microblaze version.
27532 (microblaze_trampoline_init): Adapt for microblaze.
27533 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27536 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27537 Dodji Seketeli <dodji@redhat.com>
27540 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27541 (instrument_mem_region_access): Do not forget to always put
27542 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27543 0) statement, even for cases where either 'base' or 'base + len'
27544 are not instrumented -- because they have been previously
27545 instrumented. Simplify the logic by putting all the statements
27546 instrument 'base + len' inside a sequence, and then insert that
27547 sequence right before the current insertion point. Then, to
27548 instrument 'base + len', just get an iterator on that statement.
27549 And do not forget to update the pointer to iterator the function
27550 received as argument.
27552 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27554 PR rtl-optimization/56348
27555 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27557 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27559 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27560 (clean_graph_dump_file): Pass base to start_graph_dump.
27562 2013-02-14 Richard Henderson <rth@redhat.com>
27565 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27567 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27569 * collect2-aix.h: Define F_LOADONLY.
27571 2013-02-14 Richard Biener <rguenther@suse.de>
27574 * varasm.c (output_constant_def_1): Get the decl representing
27575 the constant as argument.
27576 (output_constant_def): Wrap output_constant_def_1.
27577 (make_decl_rtl): Use output_constant_def_1 with the decl
27578 representing the constant.
27579 (build_constant_desc): Optionally re-use a decl already
27580 representing the constant.
27581 (tree_output_constant_def): Adjust.
27583 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27586 * asan.c (instrument_builtin_call): Really put the length of the
27587 second source argument into src1_len.
27589 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27591 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27592 argument. If it is false, don't create edge from then_bb to
27594 (insert_if_then_before_iter): Pass true to it.
27595 (build_check_stmt): Pass false to it.
27596 (transform_statements): Flush hash table only on extended basic
27597 block boundaries, rather than at the beginning of every bb.
27598 Don't flush hash table on nonfreeing_call_p calls.
27599 * tree-flow.h (nonfreeing_call_p): New prototype.
27600 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27602 2013-02-13 David S. Miller <davem@davemloft.net>
27604 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27606 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27609 * ira.c (max_regno_before_ira): Move from ...
27611 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27614 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27616 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27618 2013-02-13 Richard Biener <rguenther@suse.de>
27621 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27622 globals in MEM_REFs.
27624 2013-02-13 Richard Biener <rguenther@suse.de>
27626 * loop-init.c (loop_optimizer_init): Clear loop state when
27627 re-initializing preserved loops.
27628 * loop-unswitch.c (unswitch_single_loop): Return whether
27629 we unswitched the loop. Do not verify loop state here.
27630 (unswitch_loops): When we unswitched a loop discover new loops.
27632 2013-02-13 Kostya Serebryany <kcc@google.com>
27634 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27636 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27638 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27640 Avoid instrumenting duplicated memory access in the same basic block
27641 * Makefile.in (asan.o): Add new dependency on hash-table.h
27642 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27643 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27644 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27645 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27646 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27647 (get_mem_ref_of_assignment): New functions.
27648 (get_mem_refs_of_builtin_call): Extract from
27649 instrument_builtin_call and tweak a little bit to make it fit with
27651 (instrument_builtin_call): Use the new
27652 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27653 of is_gimple_builtin_call.
27654 (instrument_derefs, instrument_mem_region_access): Insert the
27655 instrumented memory reference into the hash table.
27656 (maybe_instrument_assignment): Renamed instrument_assignment into
27657 this, and change it to advance the iterator when instrumentation
27658 actually happened and return true in that case. This makes it
27659 homogeneous with maybe_instrument_assignment, and thus give a
27660 chance to callers to be more 'regular'.
27661 (transform_statements): Clear the memory reference hash table
27662 whenever we enter a new BB, when we cross a function call, or when
27663 we are done transforming statements. Use
27664 maybe_instrument_assignment instead of instrumentation. No more
27665 need to special case maybe_instrument_assignment and advance the
27666 iterator after calling it; it's now handled just like
27667 maybe_instrument_call. Update comment.
27669 2013-02-13 Richard Biener <rguenther@suse.de>
27671 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27672 Fix loop discovery code.
27674 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27676 PR inline-asm/56148
27677 * lra-constraints.c (process_alt_operands): Match early clobber
27678 operand with itself. Check conflicts with earlyclobber only if
27679 the operand is not reloaded. Prefer to reload conflicting operand
27680 if earlyclobber and matching operands are the same.
27682 2013-02-12 Richard Biener <rguenther@suse.de>
27685 * lto-streamer-out.c (write_symbol): Do not output symbols
27686 for hard register variables.
27688 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27691 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27692 (umulsidi3_insn, mulsidi3_insn): New insns.
27694 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27696 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27697 (struct tune_params): Add vec_costs field.
27698 * config/arm/arm.c (arm_builtin_vectorization_cost)
27699 (arm_add_stmt_cost): New functions.
27700 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27701 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27702 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27703 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27704 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27705 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27706 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27708 2013-02-12 Richard Biener <rguenther@suse.de>
27711 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27712 decls again if possible.
27714 2013-02-12 Richard Biener <rguenther@suse.de>
27716 PR middle-end/56288
27717 * tree-ssa.c (verify_ssa_name): Fix check, move
27718 SSA_NAME_IN_FREE_LIST check up.
27720 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27721 Steven Bosscher <steven@gcc.gnu.org>
27723 PR rtl-optimization/56151
27724 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27725 equal to op0 or op1, and last_insn pattern is CODE operation
27726 with MEM dest and one of the operands matches that MEM.
27728 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27730 * doc/extend.texi: Document Function Multiversioning and "default"
27731 parameter string to target attribute.
27732 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27733 target attribute parameter is "default".
27734 (ix86_compare_version_priority): Remove checks for target attribute.
27735 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27736 Remove check for target attribute equal to NULL. Add assert.
27737 (ix86_generate_version_dispatcher_body): Change error to sorry.
27739 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27740 Jack Howarth <howarth@bromo.med.uc.edu>
27741 Patrick Marlier <patrick.marlier@gmail.com>
27744 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27745 define ENDFILE_SPEC as TM_DESTRUCTOR.
27746 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27748 2013-02-11 Alexander Potapenko <glider@google.com>
27749 Jack Howarth <howarth@bromo.med.uc.edu>
27750 Jakub Jelinek <jakub@redhat.com>
27753 * config/darwin.c (cdtor_record): Rename ctor_record.
27754 (sort_cdtor_records): Rename sort_ctor_records.
27755 (finalize_dtors): New routine to sort destructors by
27756 priority before use in assemble_integer.
27757 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27759 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27761 PR rtl-optimization/56275
27762 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27763 offset is non-negative and less than cmode size before
27764 calling simplify_subreg.
27766 2013-02-11 Richard Biener <rguenther@suse.de>
27768 PR tree-optimization/56264
27769 * cfgloop.h (fix_loop_structure): Adjust prototype.
27770 * loop-init.c (fix_loop_structure): Return the number of
27771 newly discovered loops.
27772 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27773 are discovered, do a full loop-closed SSA rewrite.
27775 2013-02-11 Richard Biener <rguenther@suse.de>
27777 PR tree-optimization/56273
27778 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27780 (check_array_ref): Fix missing newline in dumps.
27781 (search_for_addr_array): Likewise.
27783 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27785 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27787 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27790 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27792 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27794 PR rtl-optimization/56246
27795 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27797 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27798 constraints are satisfied.
27800 2013-02-08 Jeff Law <law@redhat.com>
27803 * emit-rtl.c (reg_is_parm_p): New function.
27804 * regs.h (reg_is_parm_p): New prototype.
27805 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27806 callee-clobbered registers.
27808 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27811 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27812 If there is no implicit builtin declaration, just return NULL.
27814 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27816 * config/i386/sse.md (FMAMODEM): New mode iterator.
27817 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27818 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27820 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27822 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27823 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27824 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27826 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27828 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27829 (microblaze*-*-elf): Likewise.
27830 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27832 * config/microblaze/microblaze-c.c: Add builtin defines for
27833 _LITTLE_ENDIAN and _BIG_ENDIAN.
27834 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27835 add to TARGET_DEFAULT flags.
27836 Expand ASM_SPEC and LINK_SPEC.
27837 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27838 * config/microblaze/microblaze.md: Update extendsidi2 and
27839 movdi_internal instructions to use low-order / high-order reg
27841 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27842 options and inversemask / mask of LITTLE_ENDIAN.
27843 * config/microblaze/t-microblaze: Expand multilib options to
27844 include mlittle-endian (le) and update exceptions patterns.
27846 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27848 PR rtl-optimization/56195
27849 * lra-constraints.c (get_reload_reg): Don't reuse regs
27850 if they have smaller mode than requested, if they have
27851 wider mode than requested, try to return a SUBREG.
27853 PR tree-optimization/56250
27854 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27855 if type is unsigned and code isn't MULT_EXPR.
27857 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27859 PR tree-optimization/56064
27860 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27861 bits according to mode.
27862 * fixed-value.h (fixed_from_double_int)
27863 (const_fixed_from_double_int): Adjust comments.
27865 2013-02-08 Richard Biener <rguenther@suse.de>
27868 * lto-streamer.h (struct data_in): Remove current_file, current_line
27869 and current_col members.
27870 * lto-streamer-out.c (lto_output_location): Stream changed bits
27871 en-block for efficiency.
27872 * lto-streamer-in.c (clear_line_info): Remove.
27873 (lto_input_location): Cache current file, line and column
27874 globally via local statics. Read changed bits en-block.
27875 (input_function): Do not call clear_line_info.
27876 (lto_read_body): Likewise.
27877 (lto_input_toplevel_asms): Likewise.
27879 2013-02-08 Michael Matz <matz@suse.de>
27881 PR tree-optimization/52448
27882 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27883 (nt_call_phase): New static.
27884 (add_or_mark_expr): Only mark accesses with newer phase than any
27886 (nonfreeing_call_p): New.
27887 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27888 (nt_fini_block): Keep blocks marked as visited.
27889 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27891 2013-02-08 Richard Biener <rguenther@suse.de>
27893 * ira.c (ira): Free broken dominator information.
27895 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27897 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27899 2013-02-08 Marek Polacek <polacek@redhat.com>
27901 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27903 2013-02-08 Richard Biener <rguenther@suse.de>
27905 PR middle-end/56181
27906 * cfgloop.h (flow_loops_find): Adjust.
27907 (bb_loop_header_p): Declare.
27908 * cfgloop.c (bb_loop_header_p): New function split out from ...
27909 (flow_loops_find): ... here. Adjust function signature,
27910 support incremental loop structure update.
27911 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27912 * cfgloopmanip.c (fix_loop_structure): Move ...
27913 * loop-init.c (fix_loop_structure): ... here.
27914 (apply_loop_flags): Split out from ...
27915 (loop_optimizer_init): ... here.
27916 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27917 in incremental mode, only remove dead loops here.
27919 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27922 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27923 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27924 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27925 (*round<mode>3.libgcc): New insns for fixed-modes.
27926 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27927 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27928 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27929 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27930 implementations. Define to __builtin_avr_absFX,
27931 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27932 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27933 __builtin_avr_countlsFX, respectively.
27934 * config/avr/avr-c.c (target.h): Include it.
27935 (enum avr_builtin_id): New enum.
27936 (avr_resolve_overloaded_builtin): New static function.
27937 (avr_register_target_pragmas): Use it to set
27938 targetm.resolve_overloaded_builtin.
27939 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27940 tree nodes used by DEF_BUILTIN.
27941 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27942 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27943 <AVR_BUILTIN_xxBITS>: Same.
27945 2013-02-08 Richard Biener <rguenther@suse.de>
27947 * cfgloop.c (verify_loop_structure): Properly handle
27948 a loop exiting to another loop header.
27949 * ira-int.h (ira_loops): Remove.
27950 * ira.c (ira_loops): Remove.
27951 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27952 (do_reload): Use loop_optimizer_finalize.
27953 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27954 number_of_loops to access the loop tree.
27955 (more_one_region_p): Likewise.
27956 (finish_loop_tree_nodes): Likewise.
27957 (rebuild_regno_allocno_maps): Likewise.
27958 (mark_loops_for_removal): Likewise.
27959 (mark_all_loops_for_removal): Likewise.
27960 (remove_unnecessary_regions): Likewise.
27961 (ira_build): Likewise.
27962 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27964 2013-02-08 Richard Biener <rguenther@suse.de>
27966 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27967 * ipa-pure-const.c (analyze_function): Avoid calling
27968 mark_irreducible_loops twice.
27969 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27971 2013-02-07 David S. Miller <davem@davemloft.net>
27973 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27975 * var-tracking.c (vt_add_function_parameter): Test the presence of
27976 HAVE_window_save properly and do not remap argument registers when
27977 we have a leaf function.
27979 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27982 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27984 * config/i386/i386.c (ix86_print_operand): Ditto.
27986 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27988 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27990 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27992 PR rtl-optimization/56225
27993 * lra-constraints.c (process_alt_operands): Check that reload hard
27994 reg can hold value for strict_low_part.
27996 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27999 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28000 dwarf2out_end_function.
28001 (in_first_function_p, maybe_at_text_label_p,
28002 first_loclabel_num_not_at_text_label): New variables.
28003 (dwarf2out_var_location): In the first function find out
28004 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28005 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28008 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28010 PR rtl-optimization/56178
28011 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28012 SUBREG of a register. Tidy up related block of code.
28013 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28014 note if the source is a register or a SUBREG of a register.
28016 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28019 * config/rs6000/rs6000.md (ptrm): New mode attr.
28020 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28021 call_value_indirect_aix<pttrsize>,
28022 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28025 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28027 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28028 if -bnortl. Convert to strcmp and strncmp.
28030 2013-02-07 Alan Modra <amodra@gmail.com>
28033 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28034 addresses won't wrap when offsetting.
28035 (rs6000_secondary_reload): Provide secondary reloads needed for
28036 wrapping LO_SUM addresses.
28038 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28040 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28041 MACH, just __MACH__.
28043 2013-02-06 Richard Biener <rguenther@suse.de>
28045 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28046 instead of calling fix_loop_structure.
28048 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28050 PR middle-end/56217
28051 * omp-low.c (use_pointer_for_field): Return false if
28052 lower_send_shared_vars doesn't generate any copy-out code.
28054 2013-02-06 Tom de Vries <tom@codesourcery.com>
28056 PR rtl-optimization/56131
28057 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28058 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28059 of the label is NULL. Add comment.
28061 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28063 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28066 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28067 (STATIC_LIBTSAN_LIBS): Likewise.
28068 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28069 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28070 is defined, don't add anything else beyond that.
28071 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28072 (LINK_COMMAND_SPEC): Use them.
28074 PR tree-optimization/56205
28075 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28076 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28077 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28079 2013-02-05 Richard Biener <rguenther@suse.de>
28081 PR tree-optimization/53342
28082 PR tree-optimization/53185
28083 * tree-vectorizer.h (vect_check_strided_load): Remove.
28084 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28085 not disallow peeling for vectorized strided loads.
28086 (vect_check_strided_load): Make static and simplify.
28087 (vect_analyze_data_refs): Adjust.
28088 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28089 correctly when vectorizing strided loads.
28091 2013-02-05 Richard Biener <rguenther@suse.de>
28093 * doc/install.texi: Refer to ISL, not PPL.
28095 2013-02-05 Jan Hubicka <jh@suse.cz>
28097 PR tree-optimization/55789
28098 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28100 2013-02-05 Jan Hubicka <jh@suse.cz>
28102 PR tree-optimization/55789
28103 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28104 the dead call anyway.
28106 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28109 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28111 2013-02-04 Alexander Potapenko <glider@google.com>
28112 Jack Howarth <howarth@bromo.med.uc.edu>
28113 Jakub Jelinek <jakub@redhat.com>
28116 * config/darwin.c (sort_ctor_records): Stabilized qsort
28117 on constructor priority by using original position.
28118 (finalize_ctors): New routine to sort constructors by
28119 priority before use in assemble_integer.
28120 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28122 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28125 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28126 about visibility on artificial decls.
28127 * config/sol2.c (solaris_assemble_visibility): Likewise.
28129 2013-02-04 Kai Tietz <ktietz@redhat.com>
28132 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28133 argument and improve checking of return-argument types for 16-byte
28135 (ix86_function_value_1): Add additional valtype argument on call
28136 of function_value_64.
28137 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28138 handling infunction_value_64 function.
28140 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28142 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28144 2013-02-04 Richard Biener <rguenther@suse.de>
28146 PR tree-optimization/56188
28147 * tree-ssa-structalias.c (label_visit): Consider case with
28148 initially non-empty points-to set.
28149 (perform_var_substitution): Dump node mapping and clean up.
28151 2013-02-04 Richard Guenther <rguenther@suse.de>
28154 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28155 node prevail as last resort.
28156 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28157 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28159 2013-02-04 Richard Biener <rguenther@suse.de>
28161 PR tree-optimization/56113
28162 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28164 (equiv_class_lookup_or_add): ... this.
28165 (label_visit): Adjust and fix error in previous patch.
28166 (perform_var_substitution): Adjust.
28168 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28170 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28171 * config/sh/sh4-300.md: Likewise.
28172 * config/sh/sh4a.md: Likewise.
28173 * config/sh/constraints.md: Likewise.
28174 * config/sh/sh.md: Likewise.
28175 * config/sh/netbsd-elf.h: Likewise.
28176 * config/sh/predicates.md: Likewise.
28177 * config/sh/sh-protos.h: Likewise.
28178 * config/sh/ushmedia.h: Likewise.
28179 * config/sh/linux.h: Likewise.
28180 * config/sh/sh.c: Likewise.
28181 * config/sh/superh.h: Likewise.
28182 * config/sh/elf.h: Likewise.
28183 * config/sh/sh4.md: Likewise.
28184 * config/sh/sh.h: Likewise.
28186 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28188 * config/pa/constraints.md: Adjust unused letters. Change "T"
28189 constraint to match_test floating_point_store_memory_operand().
28190 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28191 (base14_operand): New.
28192 (floating_point_store_memory_operand): New.
28193 (integer_store_memory_operand): Revise to use base14_operand and
28194 reg_plus_base_memory_operand.
28195 (move_dest_operand): Allow symbolic_memory_operands.
28196 (symbolic_memory_operand): Check for LO_SOM.
28197 (symbolic_operand): Change default case to break.
28198 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28199 CONST_DOUBLE values to be reloaded by putting them into memory when
28200 the destination is a floating point register.
28201 (movdf): Remove code to handle CONST_DOUBLE.
28203 (reload_indf_r1): New.
28204 (reload_insf_r1): New.
28205 Consistently use "Q" and "T" constraints with integer and floating
28206 point move instructions, respectively.
28207 (movdi): Remove FAIL.
28208 Change predicate for source operand unamed DImode move from
28209 general_operand to move_src_operand.
28210 (umulsidi3): Change predicate for destination operand to
28212 Likewise for similar unamed patterns.
28213 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28214 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28215 (hppa_legitimize_address): Simplify mask calculation.
28216 (pa_emit_move_sequence): Revised handling of secondary reloads from
28217 REG+D addresses for floating point loads and stores. Directly handle
28218 loading CONST0_RTX (mode) to a floating point register.
28219 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28220 to floating point registers. Don't restrict secondary reloads to
28221 floating point registers to integer modes. Revise some comments and
28223 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28224 (pa_legitimate_address_p): New.
28225 (pa_legitimize_reload_address): New.
28226 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28227 (STRICT_REG_OK_FOR_BASE_P): New.
28228 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28229 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28231 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28232 Andrew Dixie <andrewd@gentrack.com>
28234 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28237 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28239 * expmed.c (extract_bit_field_1): Pass the full width of the
28240 structure to get_best_reg_extraction_insn.
28242 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28245 * configure.ac (use_cxa_atexit): Add AIX.
28246 * configure: Regenerate.
28248 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28250 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28253 * final.c (need_profile_function): New variable.
28254 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28255 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28256 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28257 notes, targetm.asm_out.function_prologue doesn't emit anything,
28258 HAVE_prologue and profiler should be emitted before prologue,
28259 set need_profile_function instead of emitting it.
28260 (final_scan_insn): If need_profile_function, emit
28261 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28262 NOTE_INSN_FUNCTION_BEG note.
28264 2013-02-01 Richard Henderson <rth@redhat.com>
28266 * config/rs6000/rs6000.md (smulditi3): New.
28269 * config/alpha/alpha.md (umulditi3): New.
28271 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28273 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28274 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28276 2013-02-01 Richard Biener <rguenther@suse.de>
28278 PR tree-optimization/56113
28279 * tree-ssa-structalias.c (label_visit): Reduce work for
28280 single-predecessor nodes.
28282 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28284 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28285 range isn't testing for zero.
28287 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
28289 PR middle-end/56113
28290 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
28292 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
28293 Nick Clifton <nickc@redhat.com>
28295 * config/v850/constraints.md (Q): Define as a memory constraint.
28296 * config/v850/predicates.md (label_ref_operand): New predicate.
28297 (e3v5_shift_operand): New predicate.
28298 (ior_operator): New predicate.
28299 * config/v850/t-v850: Add e3v5 multilib.
28300 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
28301 (v850_gen_movdi): Prototype.
28302 * config/v850/v850.c: Add support for e3v5 architecture.
28303 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
28305 (construct_save_jarl): Add e3v5 long JARL support.
28306 (v850_adjust_insn_length): New function. Adjust length of call
28307 insns when using e3v5 instructions.
28308 (v850_gen_movdi): New function: Generate instructions to move a
28310 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
28311 (CPP_SPEC): Define __v850e3v5__ as appropriate.
28312 (TARGET_USE_FPU): Enable for e3v5.
28313 (CONST_OK_FOR_W): New macro.
28314 (ADJUST_INSN_LENGTH): Define.
28315 * config/v850/v850.md (UNSPEC_LOOP): Define.
28316 (attr cpu): Add v850e3v5.
28317 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
28318 (movdi): New pattern.
28319 (movdi_internal): New pattern.
28320 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
28321 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
28322 (cstoresf4): Likewise.
28323 (cstoredf4): Likewise.
28324 (insv): New pattern.
28325 (rotlso3_a): New pattern.
28326 (rotlsi3_b): New pattern
28327 (rotlsi3_v850e3v5): New pattern.
28328 (doloop_begin): New pattern.
28329 (fix_loop_counter): New pattern.
28330 (doloop_end): New pattern.
28331 (branch_normal): Add e3v5 long branch support.
28332 (branch_invert): Likewise.
28333 (branch_z_normal): Likewise.
28334 (branch_z_invert): Likewise.
28335 (branch_nz_normal): Likewise.
28336 (branch_nz_invert): Likewise.
28337 (call_internal_short): Add e3v5 register-indirect JARL support.
28338 (call_internal_long): Likewise.
28339 (call_value_internal_short): Likewise.
28340 (call_value_internal_long): Likewise.
28341 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
28342 (mloop): New option.
28343 * config.gcc: Add support for configuring v840e3v5 target.
28344 * doc/invoke.texi: Document new v850 specific command line options.
28346 2013-01-31 Paul Koning <ni1d@arrl.net>
28350 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
28351 children if parent is a class.
28352 (prune_unused_types_prune): Don't add DW_AT_declaration.
28354 2013-01-31 Richard Biener <rguenther@suse.de>
28356 PR tree-optimization/56157
28357 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28358 match up operand with SLP child.
28360 2013-01-31 Jason Merrill <jason@redhat.com>
28363 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28364 parameters the first time.
28365 (gen_scheduled_generic_parms_dies): Check completeness here.
28367 2013-01-31 Richard Biener <rguenther@suse.de>
28369 PR middle-end/53073
28370 * common.opt (faggressive-loop-optimizations): New flag,
28371 enabled by default.
28372 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28373 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28374 infer_loop_bounds_from_undefined by it.
28376 2013-01-31 Richard Biener <rguenther@suse.de>
28378 PR tree-optimization/56150
28379 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28380 visit virtual operands.
28381 (find_uses_to_rename_bb): Likewise.
28383 2013-01-31 Richard Biener <rguenther@suse.de>
28385 PR tree-optimization/56150
28386 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28387 mixed store non-store stmts.
28389 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28392 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28393 LIBASAN_EARLY_SPEC is defined.
28394 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28395 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28397 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28400 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28401 invalid args instead of ICEing on it.
28402 (ix86_valid_target_attribute_tree): Return error_mark_node if
28403 ix86_valid_target_attribute_inner_p failed.
28404 (ix86_valid_target_attribute_p): Return false only if
28405 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28406 target("default") attribute.
28407 (sorted_attr_string): Change argument from const char * to tree,
28408 merge in all target attribute arguments rather than just one.
28409 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28410 instead of free. Avoid using strcat.
28411 (ix86_mangle_function_version_assembler_name): Mangle
28412 target("default") as if no target attribute is present. Adjust
28413 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28414 instead of xmalloc and XDELETEVEC instead of free.
28415 (ix86_function_versions): Don't return true if one of the decls
28416 doesn't have target attribute. If they don't and one of the decls
28417 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28418 sorted_attr_string caller. Use XDELETEVEC instead of free.
28419 (ix86_supports_function_versions): Remove.
28420 (make_name): Fix up formatting.
28421 (make_dispatcher_decl): Remove resolver_name and its initialization.
28422 Avoid leaking memory.
28423 (is_function_default_version): Return true if there is
28424 target("default") attribute rather than no target attribute at all.
28425 (make_resolver_func): Avoid leaking memory.
28426 (ix86_generate_version_dispatcher_body): Likewise.
28427 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28428 * target.def (supports_function_versions): Remove.
28429 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28430 * doc/tm.texi: Regenerated.
28432 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28434 PR rtl-optimization/56144
28435 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28436 for values with side effects.
28438 2013-01-30 Richard Biener <rguenther@suse.de>
28440 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28441 (sparseset_pop): Likewise.
28442 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28443 to be able to use quick_push in the worker loop.
28445 2013-01-30 Marek Polacek <polacek@redhat.com>
28447 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28449 2013-01-30 Richard Biener <rguenther@suse.de>
28452 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28454 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28456 PR tree-optimization/56064
28457 * fixed-value.c (fixed_from_double_int): New function.
28458 * fixed-value.h (fixed_from_double_int): New prototype.
28459 (const_fixed_from_double_int): New static inline function.
28460 * fold-const.c (native_interpret_fixed): New static function.
28461 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28462 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28463 (native_encode_fixed): New static function.
28464 (native_encode_expr) <FIXED_CST>: Use it.
28465 (native_interpret_int): Move double_int worker code to...
28466 * double-int.c (double_int::from_buffer): ...this new static method.
28467 * double-int.h (double_int::from_buffer): Prototype it.
28469 2013-01-30 Richard Biener <rguenther@suse.de>
28471 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28472 New pointer-map and obstack.
28473 (init_alias_vars): Allocate pointer-map and obstack.
28474 (delete_points_to_sets): Free them.
28475 (find_what_var_points_to): Cache result.
28476 (find_what_p_points_to): Adjust for changed interface of
28477 find_what_var_points_to.
28478 (compute_points_to_sets): Likewise.
28479 (ipa_pta_execute): Likewise.
28481 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28483 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28484 * configure: Regenerate.
28485 * config.in: Regenerate.
28486 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28487 #nobits/#progbits if supported.
28489 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28492 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28493 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28494 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28496 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28498 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28499 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28501 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28503 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28504 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28506 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28508 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28510 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28511 * config/arm/cortex-a7.md: New bypasses using
28512 arm_mac_accumulator_is_result.
28514 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28516 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28517 (cortex_a7_neon_mla): Likewise.
28518 (cortex_a7_fpfmad): New reservation.
28519 (cortex_a7_fpmacs): Use ffmas and update required units.
28520 (cortex_a7_fpmuld): Update required units and latency.
28521 (cortex_a7_fpmacd): Likewise.
28522 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28523 (cortex_a7_neon). Likewise.
28524 (bypass) Update participating units.
28526 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28528 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28529 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28531 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28532 (vfp_fmul): Use ffmad.
28533 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28534 (cortex_r4_fmacd): Use ffmad.
28535 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28536 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28537 (cortex_a9_fmacd): Use ffmad.
28538 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28539 (cortex_a8_vfp_macd): Use ffmad.
28540 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28541 (cortex_a5_fpmacd): Use ffmad.
28542 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28543 (cortex_a15_vfp_macd): Use ffmad.
28544 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28546 2013-01-29 Jason Merrill <jason@redhat.com>
28549 * varasm.c (default_assemble_visibility): Don't warn about
28550 visibility on artificial decls.
28552 2013-01-29 Richard Biener <rguenther@suse.de>
28554 PR tree-optimization/56113
28555 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28557 (label_visit): Free duplicate bitmaps and record the leader instead.
28558 (perform_var_substitution): Adjust.
28560 2013-01-29 Richard Biener <rguenther@suse.de>
28562 PR tree-optimization/55270
28563 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28564 the CFG, schedule loops for fixup.
28566 2013-01-29 Nick Clifton <nickc@redhat.com>
28568 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28571 2013-01-28 Leif Ekblad <leif@rdos.net>
28573 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28574 * config/i386/i386.h (TARGET_RDOS): New macro.
28575 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28576 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28577 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28578 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28579 DEFAULT_LARGE_SECTION_THRESHOLD.
28580 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28581 * config/i386/rdos.h: New file.
28582 * config/i386/rdos64.h: New file.
28584 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28587 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28590 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28592 PR rtl-optimization/56117
28593 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28594 call cselib_lookup_from_insn on the MEM before calling
28595 add_insn_mem_dependence.
28597 2013-01-28 Richard Biener <rguenther@suse.de>
28599 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28600 to a stmt that didn't have one.
28601 (copy_phis_for_bb): Likewise for PHI arguments.
28602 (copy_debug_stmt): Likewise for debug stmts.
28604 2013-01-28 Richard Biener <rguenther@suse.de>
28606 PR tree-optimization/56034
28607 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28608 (partition_builtin_p): Adjust.
28609 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28610 it is the last partition.
28611 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28612 up the vertex for the definition.
28613 (classify_partition): Classify whether a partition is a
28614 PKIND_REDUCTION, thus has uses outside of the loop.
28615 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28616 Merge all PKIND_REDUCTION partitions into the last partition.
28617 (tree_loop_distribution): Seed partitions from reductions as well.
28619 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28621 PR tree-optimization/56125
28622 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28623 pow(x,c) into sqrt(x) * powi(x, n/2) or
28624 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28625 optimizing for size.
28626 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28627 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28630 PR tree-optimization/56094
28631 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28632 to UNKNOWN_LOCATION while gimplifying expr.
28634 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28637 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28638 operand 0 in movabs insn template for -masm=intel asm alternative.
28639 (*movabs<mode>_2): Ditto for operand 1.
28641 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28644 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28647 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28649 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28652 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28654 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28655 Undef to avoid warning.
28657 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28659 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28660 * configure: Regenerate.
28662 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28664 PR tree-optimization/56098
28665 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28666 for stmts with volatile ops.
28667 (cond_store_replacement): Don't optimize if assign has volatile ops.
28668 (cond_if_else_store_replacement_1): Don't optimize if either
28669 then_assign or else_assign have volatile ops.
28670 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28673 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28675 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28677 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28679 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28680 missing ':' in asm example.
28682 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28684 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28685 entries into lane and laneq entries.
28686 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28687 Remove AdvSIMD scalar modes.
28688 (aarch64_sq<r>dmulh_laneq<mode>): New.
28689 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28691 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28692 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28693 * config/aarch64/iterators.md (VCOND): New.
28696 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28699 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28700 Add NULL LIBNAME argument to existing definitions.
28701 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28702 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28703 * config/avr/avr.c (DEF_BUILTIN): Same.
28704 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28705 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28706 implementation is available (DECL_ASSEMBLER_NAME is set).
28707 (avr_fold_absfx): New static function.
28708 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28709 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28710 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28711 AVR_BUILTIN_ABSLLK.
28712 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28713 (abshk, absk, abslk, absllk): Provide as static inline functions.
28715 2013-01-25 Marek Polacek <polacek@redhat.com>
28717 PR tree-optimization/56035
28718 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28720 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28722 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28723 (*movtf_internal_rex64): Add (!o,C) alternative
28724 (*movxf_internal_rex64): Ditto.
28725 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28727 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28729 * doc/invoke.texi: fix typo.
28730 * doc/objc.texi: fix typo.
28732 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28734 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28735 for the first two alternatives.
28737 2013-01-24 Diego Novillo <dnovillo@google.com>
28739 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28740 (ggc-zone.o): Remove.
28741 * configure.ac: Remove option --with-gc.
28742 * configure: Re-generate.
28743 * doc/install.texi: Remove documentation for --with-gc.
28744 * gengtype.c (write_enum_defn): Remove. Update all users.
28745 (write_Types_process_field): Remove generation of gt_e_* argument.
28746 (output_type_enum): Remove. Update all users.
28747 (write_enum_defn): Remove. Update all users.
28748 (enum alloc_zone): Remove. Update all users.
28749 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28750 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28751 Update all callers.
28752 (struct ptr_data): Remove field TYPE. Update all users.
28753 (gt_pch_note_object): Remove argument TYPE. Update all users.
28754 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28756 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28757 (struct alloc_zone): Remove.
28758 (ggc_internal_alloc_zone_stat): Remove.
28759 (ggc_internal_cleared_alloc_zone_stat): Remove.
28760 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28761 (ggc_pch_count_object): Remove last argument. Update all users.
28762 (ggc_pch_alloc_object): Remove last argument. Update all users.
28763 (struct alloc_zone): Remove.
28764 * ggc-zone.c: Remove.
28765 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28766 (struct alloc_zone): Remove.
28767 (ggc_alloc_typed_stat): Remove.
28768 (ggc_alloc_typed): Remove.
28769 (ggc_splay_alloc): Remove first argument.
28770 (rtl_zone): Remove. Update all users.
28771 (tree_zone): Remove. Update all users.
28772 (tree_id_zone): Remove. Update all users.
28773 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28774 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28775 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28776 * tree-ssanames.c: Remove references to zone allocator in comments.
28778 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28780 * config/avr/avr.c (avr_out_fract): Make register numbers that
28781 might be outside of source operand signed.
28783 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28785 * config/i386/constraints.md (Yf): New constraint.
28786 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28787 of f constraint to conditionaly disable x87 register preferences.
28788 (*movdf_internal): Ditto.
28789 (*movsf_internal): Ditto.
28791 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28793 PR inline-asm/55934
28794 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28795 that have operands with impossible constraints.
28796 Add a FIXME for a speed-up opportunity.
28797 * lra-constraints.c (process_alt_operands): Verify that a class
28798 selected from constraints on asms is valid for the operand mode.
28799 (curr_insn_transform): Remove incorrect comment.
28801 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28803 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28804 TOC operand is a valid symbol ref in the constant pool.
28806 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28808 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28810 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28813 * config/avr/stdfix.h: New file.
28814 * t-avr (stdfix-gcc.h): New rule to build it.
28815 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28817 2013-01-23 Kostya Serebryany <kcc@google.com>
28819 * config/darwin.h: remove dependency on
28820 CoreFoundation (asan on Mac OS).
28822 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28825 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28826 instead of cmpdi_operand for first comparison operand.
28827 Don't assert that comparison operands aren't both constants.
28829 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28831 * doc/install.texi (Downloading the Source): Update references to
28832 downloading separate components.
28834 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28836 * doc/extend.texi (__int128): Improve grammar.
28838 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28841 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28842 alternative to (o,r).
28843 (*movdi_internal_rex64): Remove (!o,n) alternative.
28844 (DImode immediate->memory splitter): Remove.
28845 (DImode immediate->memory peephole2): Remove.
28846 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28847 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28848 alternative to (!o,*r).
28849 (*movtf_internal_sse): New pattern.
28850 (*movxf_internal_rex64): New pattern.
28851 (*movxf_internal): Disable for TARGET_64BIT.
28852 (*movdf_internal_rex64): Remove (!o,F) alternative.
28854 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28856 PR middle-end/56074
28857 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28858 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28859 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28860 stmt locations where LOCATION_LOCUS of the stmt location is
28861 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28864 * config/i386/i386.md (UNSPEC_STOS): New.
28865 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28866 *strsetqi_1): Add UNSPEC_STOS.
28868 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28871 * doc/invoke.texi: Remove left over -Wsynth example.
28873 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28875 PR tree-optimization/56051
28876 * fold-const.c (fold_binary_loc): Don't fold
28877 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28878 a narrowing conversion, or widening conversion from signed
28881 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28883 PR rtl-optimization/56023
28884 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28885 dependent on debug instruction.
28887 2013-01-21 Martin Jambor <mjambor@suse.cz>
28889 PR middle-end/56022
28890 * function.c (allocate_struct_function): Call
28891 invoke_set_current_function_hook earlier.
28893 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28895 * reload1.c (init_reload): Only initialize reload_obstack
28896 during the first call.
28898 2013-01-21 Marek Polacek <polacek@redhat.com>
28900 * cfgloop.c (verify_loop_structure): Fix up grammar.
28902 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28904 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28905 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28907 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28910 * config/arm/marvell-pj4.md: Update copyright year.
28911 Fix up use of alu to alu_reg and simple_alu_imm.
28913 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28915 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28917 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28920 * lra-constraints.c (curr_insn_transform): Don't reuse original
28921 insn for secondary memory move when memory mode should be different.
28923 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28925 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28926 atomic_storedi_1): New patterns.
28928 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28930 btver2 pipeline descriptions.
28931 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28933 * config/i386/i386.md (btver2_decode): New type attributes.
28934 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28936 * config/i386/btver2.md: New file describing btver2 pipelines.
28938 2013-01-19 Andrew Pinski <apinski@cavium.com>
28940 PR tree-optimization/52631
28941 * tree-ssa-sccvn (visit_use): Before looking up the original
28942 statement, try looking up the simplified expression.
28944 2013-01-19 Anthony Green <green@moxielogic.com>
28946 * config/moxie/moxie.c (moxie_expand_prologue): Set
28947 current_function_static_stack_size.
28949 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28951 PR tree-optimization/56029
28952 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28953 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28955 2013-01-18 Sharad Singhai <singhai@google.com>
28957 PR tree-optimization/55995
28958 * dumpfile.c (dump_loc): Print location only if available.
28959 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28961 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28964 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28965 secondary memory move.
28966 (inherit_reload_reg): Use rclass instead of cl for
28967 check_secondary_memory_needed_p.
28969 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28971 PR middle-end/56015
28972 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28973 the case where writing real complex part of target modifies op1.
28975 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28977 * config/aarch64/aarch64-simd.md
28978 (aarch64_vcond_internal<mode>): Handle unordered cases.
28979 * config/aarch64/iterators.md (v_cmp_result): New.
28981 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28982 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28984 * config/arm/marvell-pj4.md: New file.
28985 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28986 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28987 (generic_vfp): Likewise.
28988 * config/arm/arm-cores.def: Add marvell-pj4.
28989 * config/arm/arm-tune.md: Regenerate.
28990 * config/arm/arm-tables.opt: Regenerate.
28991 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28992 * doc/invoke.texi: Document marvell-pj4.
28994 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28996 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28998 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29003 * var-tracking.c (negative_power_of_two_p): New.
29004 (global_get_addr_cache, local_get_addr_cache): New.
29005 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29006 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29008 (vt_stack_offset_p): Remove.
29009 (vt_canon_true_dep): Always canonicalize loc's address.
29010 (clobber_overlapping_mems): Make sure we have a MEM.
29011 (local_get_addr_clear_given_value): New.
29012 (val_reset): Clear local cached entries.
29013 (compute_bb_dataflow): Create and release the local cache.
29014 Disable duplicate MEMs clobbering.
29015 (emit_notes_in_bb): Clobber MEMs likewise.
29016 (vt_emit_notes): Create and release the local cache.
29017 (vt_initialize, vt_finalize): Create and release the global
29018 cache, respectively.
29019 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29021 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29023 PR libmudflap/53359
29024 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29025 not found in the symtab.
29027 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29030 PR rtl-optimization/55547
29031 PR rtl-optimization/53827
29034 * alias.c (offset_overlap_p): New, factored out of...
29035 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29036 the conservative special case for symbolic constants. Don't
29037 adjust zero sizes on alignment.
29039 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29041 PR rtl-optimization/52573
29042 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29043 REG_UNUSED for the same register.
29045 2013-01-17 Richard Biener <rguenther@suse.de>
29046 Marek Polacek <polacek@redhat.com>
29048 PR rtl-optimization/55833
29049 * loop-unswitch.c (unswitch_loops): Move loop verification...
29050 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29051 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29052 Set it to true when we're removing a loop from hierarchy tree in
29053 an irreducible region.
29054 (fix_bb_placements): Adjust caller.
29055 (fix_loop_placements): Likewise.
29057 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29059 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29060 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29061 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29062 Remove ID. Adjust comments.
29063 * config/avr/avr-c.c (avr_builtin_name): Remove.
29064 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29065 * config/avr/avr.c (avr_tolower): New static function.
29066 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29067 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29068 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29071 2013-01-17 Jan Hubicka <jh@suse.cz>
29073 PR tree-optimization/55273
29074 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29076 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29079 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29080 store through atomic_store<mode>_1.
29081 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29083 2013-01-17 Martin Jambor <mjambor@suse.cz>
29085 PR tree-optimizations/55264
29086 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29087 for virtual methods.
29088 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29089 virtual methods before inlining is over.
29090 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29092 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29095 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29097 PR rtl-optimization/56005
29098 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29099 pending reads for prefetch.
29101 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29103 * config/aarch64/aarch64.md
29104 (*cstoresi_neg_uxtw): New pattern.
29105 (*cmovsi_insn_uxtw): New pattern.
29106 (*<optab>si3_uxtw): New pattern.
29107 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29108 (*<optab>si3_insn_uxtw): New pattern.
29109 (*bswapsi2_uxtw): New pattern.
29111 2013-01-16 Richard Biener <rguenther@suse.de>
29113 * tree-inline.c (tree_function_versioning): Remove set but
29114 never used variable.
29116 2013-01-16 Richard Biener <rguenther@suse.de>
29118 PR tree-optimization/55964
29119 * tree-flow.h (rename_variables_in_loop): Remove.
29120 (rename_variables_in_bb): Likewise.
29121 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29122 (copy_loop_before): Adjust and delete update-ssa status.
29123 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29124 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29125 (rename_variables_in_loop): Remove.
29126 (slpeel_update_phis_for_duplicate_loop): Likewise.
29127 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29128 use available cfg machinery instead of duplicating it.
29129 Update PHI nodes and perform poor-mans SSA update here.
29130 (slpeel_tree_peel_loop_to_edge): Adjust.
29132 2013-01-16 Richard Biener <rguenther@suse.de>
29134 PR tree-optimization/54767
29135 PR tree-optimization/53465
29136 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29137 (vrp_visit_phi_node): For PHI arguments coming via backedges
29138 drop all symbolical range information.
29139 (execute_vrp): Compute backedges.
29141 2013-01-16 Richard Biener <rguenther@suse.de>
29143 * doc/install.texi: Update CLooG and ISL requirements to
29146 2013-01-16 Christian Bruel <christian.bruel@st.com>
29149 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29150 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29151 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29153 2013-01-16 DJ Delorie <dj@redhat.com>
29155 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29156 (UNSPECV_SP_SWITCH_E): New.
29157 (sp_switch_1): Change to an unspec.
29158 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29161 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29163 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29164 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29165 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29166 (expand_mem_thread_fence): Ditto.
29167 (expand_mem_signal_fence): Ditto.
29168 (expand_atomic_load): Ditto.
29169 (expand_atomic_store): Ditto.
29171 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29173 PR rtl-optimization/55547
29174 PR rtl-optimization/53827
29177 * alias.c (memrefs_conflict_p): Set sizes to negative after
29180 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29183 * function.c (thread_prologue_and_epilogue_insns): Always
29184 add crtl->drap_reg to set_up_by_prologue.set, even if
29185 stack_realign_drap is false.
29187 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29189 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29190 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29191 *call): Fix indention.
29193 2013-01-15 Tom de Vries <tom@codesourcery.com>
29196 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29199 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29201 PR rtl-optimization/55153
29202 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29204 2013-01-15 Martin Jambor <mjambor@suse.cz>
29206 PR tree-optimization/55920
29207 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29208 accesses as grp_to_be_debug_replaced.
29210 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29212 PR tree-optimization/55920
29213 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29214 there is non-useless type conversion needed from debug rhs to lhs,
29215 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29217 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29218 Mikael Pettersson <mikpe@it.uu.se>
29221 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29223 (ASM_OUTPUT_CASE_LABEL): Remove.
29224 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29225 * final.c (shorten_branches): Update alignment of labels before
29226 jump tables if CASE_VECTOR_SHORTEN_MODE.
29228 2013-01-15 Richard Biener <rguenther@suse.de>
29231 * system.h: Do not include gmp.h for building host tools.
29233 2013-01-15 Richard Biener <rguenther@suse.de>
29235 PR middle-end/55882
29236 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29237 account for bitpos when computing alignment.
29239 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29241 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29242 (ix86_target_macros_internal): Likewise.
29244 * config/i386/i386.c (m_CORE2I7): Removed.
29245 (m_CORE_HASWELL): New macro.
29246 (m_CORE_ALL): Likewise.
29247 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29248 (initial_ix86_arch_features): Likewise.
29249 (processor_target_table): Initializations for Core avx2.
29250 (cpu_names): New names "core-avx2".
29251 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29252 PROCESSOR_CORE_HASWELL.
29253 (ix86_issue_rate): New case.
29254 (ia32_multipass_dfa_lookahead): Likewise.
29255 (ix86_sched_init_global): Likewise.
29257 * config/i386/i386.h (TARGET_HASWELL): New macro.
29258 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29259 (processor_type): New PROCESSOR_HASWELL.
29261 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29263 PR tree-optimization/55955
29264 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29265 *SHIFT_EXPR and *ROTATE_EXPR codes.
29267 PR tree-optimization/48766
29268 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29269 -ftrapv disable -fwrapv.
29271 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29274 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29275 etc. to 1 and not to __flash.
29276 Use LL suffix for __INT24_MAX__ with -mint8.
29277 Use ULL suffix for __UINT24_MAX__ with -mint8.
29279 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29281 * config/avr/avr-arch.h
29282 (struct base_arch_s): Use typedef avr_arch_t instead.
29283 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29284 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29285 * config/avr/avr.c: Same.
29286 * config/avr/avr-devices.c: Same.
29287 * config/avr/driver-avr.c: Same.
29288 * config/avr/gen-avr-mmcu-texi.c: Same.
29289 * config/avr/avr-mcus.def: Adjust comment.
29291 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
29293 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
29294 * config/aarch64/iterators.md (VALLDI): New.
29296 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
29297 Andi Kleen <ak@linux.intel.com>
29300 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
29301 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
29304 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29306 * config/avr/avr-stdint.h: Remove trailing blanks.
29307 * config/avr/avr-log.h: Same.
29308 * config/avr/avr-arch.h: Same.
29309 * config/avr/avr-devices.c: Same.
29310 * config/avr/avr-dimode.md: Same.
29311 * config/avr/predicates.md: Same.
29312 * config/avr/avr-c.c: Same. And fix typo.
29314 * config/avr/avr-protos.h: Same. And:
29315 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
29316 (init_cumulative_args): Rename to avr_init_cumulative_args.
29317 (expand_prologue): Rename to avr_expand_prologue.
29318 (expand_epilogue): Rename to avr_expand_epilogue.
29319 (adjust_insn_length): Rename to avr_adjust_insn_length.
29320 (notice_update_cc): Rename to avr_notice_update_cc.
29321 (final_prescan_insn): Rename to avr_final_prescan_insn.
29322 * config/avr/avr.c: Same.
29323 * config/avr/avr.h: Same.
29324 * config/avr/avr.md: Remove trailing blanks.
29325 (prologue): Use avr_expand_prologue.
29326 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
29328 2013-01-14 Richard Biener <rguenther@suse.de>
29330 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
29331 verify_location, collect_subblocks): New functions.
29332 (verify_gimple_in_cfg): Verify that locations only reference
29333 BLOCKs in the functions BLOCK tree.
29335 2013-01-14 Richard Biener <rguenther@suse.de>
29337 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
29339 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
29341 (insert_out_of_ssa_copy_on_edge): Likewise.
29342 (rewrite_close_phi_out_of_ssa): Likewise.
29343 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
29345 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
29346 propagated constants.
29347 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
29350 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29352 * config/avr/avr-modes.def: Add GPL copyright notice.
29354 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
29356 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29357 MEMMODEL_MASK to determine memory model.
29358 (atomic_store<mode>): Ditto from operands[2].
29359 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29361 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29364 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29365 (fold_gimple_assign): Don't call unshare_expr here.
29366 (fold_ctor_reference): Call unshare_expr.
29368 2013-01-13 Terry Guo <terry.guo@arm.com>
29370 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29371 * doc/fragments.texi: Document MULTILIB_REUSE.
29372 * gcc.c (multilib_reuse): New internal spec.
29373 (set_multilib_dir): Also search multilib from multilib_reuse.
29374 * genmultilib (tmpmultilib3): Refactor code.
29375 (tmpmultilib4): Ditto.
29376 (multilib_reuse): New multilib argument.
29378 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29380 * Makefile.in: Update copyright.
29382 2013-01-12 Tom de Vries <tom@codesourcery.com>
29384 PR middle-end/55890
29385 * calls.c (expand_call): Check if arg_nr is valid.
29387 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29389 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29390 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29391 documentation. Add missing '__' in front of
29392 __builtin_ia32_packssdw256.
29394 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29397 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29398 NO_REGS for larl operands.
29399 (s390_reload_larl_operand): Use s390_load_address instead of
29402 2013-01-11 Richard Biener <rguenther@suse.de>
29404 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29405 (verify_node_sharing): ... here.
29406 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29408 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29410 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29411 Substitute TREECHECKING.
29412 * configure: Regenerate.
29413 * Makefile.in (TREECHECKING): New.
29415 2013-01-11 Richard Guenther <rguenther@suse.de>
29417 PR tree-optimization/44061
29418 * tree-vrp.c (extract_range_basic): Compute zero as
29419 value-range for __builtin_constant_p of function parameters.
29421 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29423 Update copyright years.
29425 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29427 PR rtl-optimization/55672
29428 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29429 const to be eliminable.
29431 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29433 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29434 * configure: Regenerate.
29436 2013-01-10 Richard Biener <rguenther@suse.de>
29438 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29440 2013-01-10 Richard Biener <rguenther@suse.de>
29443 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29444 locations for virtual PHI arguments.
29445 (rewrite_update_phi_arguments): Likewise.
29447 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29449 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29452 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29454 PR tree-optimization/55921
29455 * tree-complex.c (expand_complex_asm): New function.
29456 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29458 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29461 * config/s390/s390.c (s390_symref_operand_p)
29462 (s390_loadrelative_operand_p): Merge the two functions.
29463 (s390_check_qrst_address, print_operand_address): Add parameters
29464 to s390_loadrelative_operand_p invokation.
29465 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29466 (s390_reload_larl_operand, s390_secondary_reload): Use
29467 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29468 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29470 2013-01-09 Mike Stump <mikestump@comcast.net>
29472 * dse.c (record_store): Remove unnecessary assert.
29474 2013-01-09 Jan Hubicka <jh@suse.cz>
29476 PR tree-optimization/55569
29477 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29478 * cfgloop.h (scale_loop_profile): Likewise.
29480 2013-01-09 Jan Hubicka <jh@suse.cz>
29483 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29485 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29487 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29489 PR middle-end/55114
29490 * expr.h (maybe_emit_group_store): Declare.
29491 * expr.c (maybe_emit_group_store): New function.
29492 * builtins.c (expand_builtin_int_roundingfn): Call it.
29493 (expand_builtin_int_roundingfn_2): Likewise.
29495 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29497 PR rtl-optimization/55829
29498 * lra-constraints.c (match_reload): Add code for absent output.
29499 (curr_insn_transform): Add code for reloads of matched inputs
29502 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29504 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29505 attribute of movddup insn to DF.
29506 (*vec_interleave_lowv2df): Ditto.
29507 (vec_dupv2df): Ditto.
29509 2013-01-09 Jan Hubicka <jh@suse.cz>
29511 PR tree-optimiation/55875
29512 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29513 EVERY_ITERATION parameter.
29514 (number_of_iterations_exit): Check if exit is executed every iteration.
29515 (idx_infer_loop_bounds): Similarly here.
29516 (n_of_executions_at_most): Simplify
29517 to only test for cases where statement is dominated by the
29518 particular bound; handle correctly the "postdominance" test.
29519 (scev_probably_wraps_p): Use max loop iterations info
29520 as a global bound first.
29522 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29523 Nick Clifton <nickc@redhat.com>
29525 * config/v850/v850.md (cbranchsf4): New pattern.
29526 (cstoresf4): New pattern.
29527 (cbranchdf4): New pattern.
29528 (cstoredf4): New pattern.
29529 (movsicc): Disallow floating point comparisons.
29530 (cmpsf_le_insn): Fix order of operators.
29531 (cmpsf_lt_insn): Likewise.
29532 (cmpsf_eq_insn): Likewise.
29533 (cmpdf_le_insn): Likewise.
29534 (cmpdf_lt_insn): Likewise.
29535 (cmpdf_eq_insn): Likewise.
29536 (cmpsf_ge_insn): Use LE comparison.
29537 (cmpdf_ge_insn): Likewise.
29538 (cmpsf_gt_insn): Use LT comparison.
29539 (cmpdf_gt_insn): Likewise.
29540 (cmpsf_ne_insn): Delete pattern.
29541 (cmpdf_ne_insn): Delete pattern.
29542 * config/v850/v850.c (v850_gen_float_compare): Use
29543 gen_cmpdf_eq_insn for NE comparison.
29544 (v850_float_z_comparison_operator)
29545 (v850_float_nz_comparison_operator): Move from here ...
29546 * config/v850/predicates.md: ... to here. Move GT and GE
29547 comparisons into v850_float_z_comparison_operator.
29548 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29550 (v850_float_nz_comparison_operator): Likewise.
29552 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29554 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29555 with calls to gen_insvsi/gen_insvdi.
29557 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29559 * config/i386/i386.c (initial_ix86_tune_features): Set up
29560 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29562 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29563 Jakub Jelinek <jakub@redhat.com>
29565 PR tree-optimization/48189
29566 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29567 If nitercst is 0, don't predict the exit edge.
29569 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29571 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29572 in asm_fprintf with reg_names.
29573 (aarch64_print_operand_address): Likewise.
29574 (aarch64_return_addr): Likewise.
29575 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29577 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29579 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29580 (INT_U6_BITS): Likewise.
29581 * config/pa/predicates.md (uint6_operand): New predicate.
29582 (shift5_operand, shift6_operand): Likewise.
29583 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29585 (lshrdi3): Use shift6_operand.
29586 (shrpsi4, shrpdi4): New insn patterns.
29587 (extzv): Delete expander.
29588 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29589 predicates in unamed zero extract patterns. Tighten common constraint.
29590 (extv): Delete expander.
29591 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29592 predicates in unamed sign extract patterns. Tighten common constraint.
29593 (insv): Delete expander.
29594 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29595 predicates in unamed insert patterns. Tighten common constraint.
29596 Change uint32_operand predicate to uint6_operand predicate in unamed
29597 DImode pattern to insert constant values of type 1...1xxxx.
29599 2013-01-04 Jan Hubicka <jh@suse.cz>
29601 PR tree-optimization/55823
29602 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29605 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29606 Uros Bizjak <ubizjak@gmail.com>
29608 PR rtl-optimization/55845
29609 * df-problems.c (can_move_insns_across): Stop scanning at
29610 volatile_insn_p source instruction or give up if
29611 across_from .. across_to range contains any volatile_insn_p
29614 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29616 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29617 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29619 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29620 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29622 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29625 * asan.c (asan_clear_shadow): New function.
29626 (asan_emit_stack_protection): Use it.
29628 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29630 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29631 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29632 with tab instead of space.
29634 2013-01-08 Nick Clifton <nickc@redhat.com>
29636 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29637 register bank 0 at the start of an interrupt handler.
29638 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29641 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29643 * config/aarch64/aarch64-simd.md
29644 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29645 (aarch64_simd_bsl): Likewise.
29646 (aarch64_vcond_internal<mode>): Likewise.
29647 (vcond<mode><mode>): Likewise.
29648 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29649 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29651 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29653 * config/aarch64/aarch64-builtins.c
29654 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29656 2013-01-08 Martin Jambor <mjambor@suse.cz>
29659 * tree-sra.c (analyze_access_subtree): Return true also after
29660 potentially creating a debug-only replacement.
29662 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29664 PR middle-end/55890
29665 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29667 PR tree-optimization/54120
29668 * tree-vrp.c (range_fits_type_p): Don't allow
29669 src_precision < precision from signed vr to unsigned_p
29670 if vr->min or vr->max is negative.
29671 (simplify_float_conversion_using_ranges): Test can_float_p
29672 against CODE_FOR_nothing.
29674 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29675 Richard Biener <rguenther@suse.de>
29677 PR middle-end/55851
29678 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29679 types instead of just INTEGER_TYPE types.
29681 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29683 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29686 2013-01-07 Steve Ellcey <sellcey@mips.com>
29689 * config/mips/mips.opt: Change mad to mmad to match documentation.
29691 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29694 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29695 .progmemx.data now.
29697 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29700 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29701 (avr_addrspace_t): Add .section_name field.
29702 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29704 (avr_addrspace): Same. Initialize .section_name. Remove last
29705 NULL entry. Put __memx into .progmemx.data.
29706 (progmem_section_prefix): Remove.
29707 (avr_asm_init_sections): No need to initialize progmem_section.
29708 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29709 section name prefix.
29710 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29711 retrieve the progmem section.
29712 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29713 boundary to run over avr_addrspace[].
29714 (avr_register_target_pragmas): Ditto.
29716 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29718 * varasm.c (output_constant_def_contents): For asan_protect_global
29719 protected strings, adjust DECL_ALIGN if needed, before testing for
29721 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29722 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29724 (output_object_block): For asan protected decls, emit asan padding
29725 after their contents.
29726 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29727 (asan_finish_file): Test it here instead.
29729 2013-01-07 Nick Clifton <nickc@redhat.com>
29730 Matthias Klose <doko@debian.org>
29731 Doug Kwan <dougkwan@google.com>
29732 H.J. Lu <hongjiu.lu@intel.com>
29735 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29737 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29739 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29741 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29743 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29745 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29748 * doc/install.texi (Cross-Compiler-Specific Options): Document
29751 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29753 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29754 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29755 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29756 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29757 vqmovun_high_s64): Fix source operand number and update copyright.
29759 2013-01-07 Richard Biener <rguenther@suse.de>
29761 PR middle-end/55890
29762 * gimple.h (gimple_call_builtin_p): New overload.
29763 * gimple.c (validate_call): New function.
29764 (gimple_call_builtin_p): Likewise.
29765 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29766 Use gimple_call_builtin_p.
29767 (find_func_clobbers): Likewise.
29768 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29769 (strlen_optimize_stmt): Likewise.
29771 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29773 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29774 (vld1q_dup_*): Likewise.
29775 (vld1_*): Likewise.
29776 (vld1q_*): Likewise.
29777 (vld1_lane_*): Likewise.
29778 (vld1q_lane_*): Likewise.
29780 2013-01-07 Richard Biener <rguenther@suse.de>
29782 * lto-streamer.h (LTO_minor_version): Bump to 2.
29784 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29786 * config/aarch64/aarch64-protos.h
29787 (aarch64_const_double_zero_rtx_p): Rename to...
29788 (aarch64_float_const_zero_rtx_p): ...this.
29789 (aarch64_float_const_representable_p): New.
29790 (aarch64_output_simd_mov_immediate): Likewise.
29791 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29792 move immediate case.
29793 * config/aarch64/aarch64.c
29794 (aarch64_const_double_zero_rtx_p): Rename to...
29795 (aarch64_float_const_zero_rtx_p): ...this.
29796 (aarch64_print_operand): Allow printing of new constants.
29797 (aarch64_valid_floating_const): New.
29798 (aarch64_legitimate_constant_p): Check for valid floating-point
29800 (aarch64_simd_valid_immediate): Likewise.
29801 (aarch64_vect_float_const_representable_p): New.
29802 (aarch64_float_const_representable_p): Likewise.
29803 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29804 (aarch64_output_simd_mov_immediate): New.
29805 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29806 (*movdf_aarch64): Likewise.
29807 * config/aarch64/constraints.md (Ufc): New.
29808 (Y): call aarch64_float_const_zero_rtx.
29809 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29811 2013-01-07 Richard Biener <rguenther@suse.de>
29813 PR tree-optimization/55888
29814 PR tree-optimization/55862
29815 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29816 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29817 not if it is contained therein.
29819 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29821 * config/avr/t-avr: Typo.
29823 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29826 * config/avr/t-avr: Don't automatically rebuild
29827 $(srcdir)/config/avr/t-multilib
29828 $(srcdir)/config/avr/avr-tables.opt
29829 $(srcdir)/doc/avr-mmcu.texi
29830 (avr-mcus): New phony target to build them on request.
29831 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29832 * avr/avr-mcus.def: Adjust comments.
29834 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29836 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29838 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29840 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29842 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29844 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29846 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29848 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29849 to generate profiling.
29850 * config/rs6000/aix64.h (LIB_SPEC): Same.
29852 2013-01-04 Andrew Pinski <apinski@cavium.com>
29854 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29856 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29858 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29860 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29862 (ix86_expand_move): Ditto.
29863 (ix86_zero_extend_to_Pmode): Ditto.
29864 (ix86_expand_call): Ditto.
29865 (ix86_expand_special_args_builtin): Ditto.
29866 (ix86_expand_builtin): Ditto.
29868 2013-01-04 Richard Biener <rguenther@suse.de>
29870 PR tree-optimization/55862
29871 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29872 translating them through PHI nodes.
29874 2013-01-04 Martin Jambor <mjambor@suse.cz>
29876 PR tree-optimization/55755
29877 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29878 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29880 2013-01-04 Marek Polacek <polacek@redhat.com>
29882 PR middle-end/55859
29883 * opts.c (default_options_optimization): Clarify error message.
29885 2013-01-04 Richard Biener <rguenther@suse.de>
29887 PR middle-end/55863
29888 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29891 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29894 * config/pa/pa.md (movsi): Revert previous change.
29895 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29898 2013-01-03 Richard Henderson <rth@redhat.com>
29900 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29901 after eliminating TLS symbols.
29903 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29906 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29907 * graphite-poly.c (debug_gmp_value): Likewise.
29909 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29912 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29913 selected code model, define __code_mode_small__, __code_model_medium__,
29914 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29915 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29916 xchg temporary register with %k. Declare temporary register as
29918 [__x86_64__]: For medium and large code models, preserve %rbx register.
29920 2013-01-03 Richard Biener <rguenther@suse.de>
29922 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29923 (dump_subscript): Adjust.
29924 (finalize_ddr_dependent): Do not dump redundant info.
29925 (analyze_siv_subscript): Adjust.
29926 (subscript_dependence_tester): Likewise.
29927 (compute_affine_dependence): Likewise.
29929 2013-01-03 Richard Biener <rguenther@suse.de>
29932 2013-01-03 Richard Biener <rguenther@suse.de>
29934 PR tree-optimization/55857
29935 * tree-vect-stmts.c (vectorizable_load): Do not setup
29936 re-alignment for invariant loads.
29938 2013-01-02 Richard Biener <rguenther@suse.de>
29940 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29941 invariant load do not generate a vector load from the scalar location.
29943 2013-01-03 Richard Biener <rguenther@suse.de>
29945 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29946 for not vectorizing.
29947 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29948 not build INDIRECT_REFs, call get_name once only.
29949 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29950 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29952 2013-01-03 Richard Biener <rguenther@suse.de>
29954 PR tree-optimization/55857
29955 * tree-vect-stmts.c (vectorizable_load): Do not setup
29956 re-alignment for invariant loads.
29958 2013-01-03 Richard Biener <rguenther@suse.de>
29961 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29962 prefer a built-in decl.
29964 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29966 * gcc.c (process_command): Update copyright notice dates.
29967 * gcov.c (print_version): Likewise.
29968 * gcov-dump.c (print_version): Likewise.
29970 PR rtl-optimization/55838
29971 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29972 iv0.step, iv1.step and step.
29974 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29975 Marc Glisse <marc.glisse@inria.fr>
29977 PR tree-optimization/55832
29978 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29979 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29980 integer_{one,zero}_node.
29982 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29985 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29986 * var-tracking.c (reverse_op): Don't add reverse ops to
29987 VALUEs that have already
29988 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29990 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29992 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29994 2013-01-02 Teresa Johnson <tejohnson@google.com>
29996 * dumpfile.c (dump_loc): Print filename with location.
29997 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29998 new location_t parameter to emit complete unroll message with
29999 new dump framework.
30000 (canonicalize_loop_induction_variables): Compute loops location
30001 and pass to try_unroll_loop_completely.
30002 * loop-unroll.c (report_unroll_peel): New function.
30003 (peel_loops_completely): Use new dump format with location
30004 for main dumpfile message, and invoke report_unroll_peel on success.
30005 (decide_unrolling_and_peeling): Ditto.
30006 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30007 by report_unroll_peel.
30008 (decide_peel_completely): Ditto.
30009 (decide_unroll_constant_iterations): Ditto.
30010 (decide_unroll_runtime_iterations): Ditto.
30011 (decide_peel_simple): Ditto.
30012 (decide_unroll_stupid): Ditto.
30013 * cfgloop.c (get_loop_location): New function.
30014 * cfgloop.h (get_loop_location): Declare.
30016 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30018 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30021 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30023 PR middle-end/55198
30024 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30025 BLKmode objects when EXPAND_MEMORY is specified.
30027 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30029 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30031 (fold_builtin_cpu): Do not share cpu model decls across statements.
30033 2013-01-02 Jason Merrill <jason@redhat.com>
30036 * tree.c (build_array_type_1): Revert earlier change.
30038 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30040 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30042 * config/aarch64/aarch64-tune.md: Re-generate.
30044 2013-01-02 Richard Biener <rguenther@suse.de>
30046 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30047 invariant load do not generate a vector load from the scalar location.
30049 2013-01-02 Richard Biener <rguenther@suse.de>
30052 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30053 * configure: Regenerate.
30055 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30057 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30058 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30059 (expand_builtin_int_roundingfn_2): Keep the original target around
30060 for the fallback case.
30062 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30064 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30065 to be clear for sign changes.
30067 2013-01-01 Jan Hubicka <jh@suse.cz>
30069 * ipa-inline-analysis.c: Fix formatting.
30071 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30073 PR tree-optimization/55831
30074 * tree-vect-loop.c (get_initial_def_for_induction): Use
30075 gsi_after_labels instead of gsi_start_bb.
30077 Copyright (C) 2013 Free Software Foundation, Inc.
30079 Copying and distribution of this file, with or without modification,
30080 are permitted in any medium without royalty provided the copyright
30081 notice and this notice are preserved.