1 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
3 * config/arc/arc.opt (mlra): Move comment above option name
4 to avoid mis-parsing as language options.
6 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
9 * config/sol2.h: ... here.
10 * config/sol2-10.h: Remove.
12 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
13 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
14 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
15 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
16 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
17 * config/sol2.h: ... here.
18 (SECTION_NAME_FORMAT): Don't redefine.
19 (STARTFILE_ARCH32_SPEC): Rename to ...
20 (STARTFILE_ARCH_SPEC): ... this.
21 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
22 * config/sparc/sol2.h: ... here.
23 (SECTION_NAME_FORMAT): Don't undef.
24 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
25 (SUBTARGET_EXTRA_SPECS): Remove.
26 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
28 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
29 (MD_STARTFILE_PREFIX): Remove.
30 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
31 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
32 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
33 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
34 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
35 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
36 * config/i386/sol2.h: ... here.
37 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
38 * config/i386/sol2-bi.h: Remove.
39 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
40 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
42 * config/i386/t-sol2-64: Rename to ...
43 * config/i386/t-sol2: ... this.
44 * config/sparc/t-sol2-64: Rename to ...
45 * config/sparc/t-sol2: ... this.
47 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
48 sol2_tm_file_head, sol2_tm_file_tail.
49 Include ${cpu_type}/sol2.h before sol2.h.
51 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
52 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
53 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
54 Reflect i386/t-sol2-64 renaming.
55 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
56 Reflect sparc/t-sol2-64 renaming.
58 2014-04-30 Richard Biener <rguenther@suse.de>
60 * passes.c (execute_function_todo): Move TODO_verify_stmts
61 and TODO_verify_ssa under the TODO_verify_il umbrella.
62 * tree-ssa.h (verify_ssa): Adjust prototype.
63 * tree-ssa.c (verify_ssa): Add parameter to tell whether
64 we should verify SSA operands.
65 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
66 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
67 whether we should verify whether not throwing stmts have EH info.
68 * graphite-scop-detection.c (create_sese_edges): Adjust.
69 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
70 * tree-eh.c (lower_try_finally_switch): Do not add the
71 default case label twice.
73 2014-04-30 Marek Polacek <polacek@redhat.com>
75 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
76 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
77 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
78 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
80 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
82 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
83 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
84 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
85 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
86 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
87 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
88 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
89 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
91 2014-04-29 David Malcolm <dmalcolm@redhat.com>
93 * tree-cfg.c (dump_function_to_file): Dump the return type of
94 functions, in a line to itself before the function body, mimicking
95 the layout of a C function.
97 2014-04-29 Jakub Jelinek <jakub@redhat.com>
99 PR tree-optimization/60971
100 * tree-tailcall.c (process_assignment): Reject conversions which
103 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
105 * calls.c (initialize_argument_information): Always treat
106 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
107 (expand_call): Likewise.
108 (emit_library_call_calue_1): Likewise.
109 * expr.c (PUSH_ARGS_REVERSED): Do not define.
110 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
113 2014-04-29 Nick Clifton <nickc@redhat.com>
115 * config/msp430/msp430.md (umulsidi): Fix typo.
116 (mulhisi3): Enable even inside interrupt handlers.
117 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
118 bigger return address pushed in large mode.
120 2014-04-29 Nick Clifton <nickc@redhat.com>
122 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
123 (arc_init_reg_tables): Use a machine_mode enum to iterate over
125 * config/m32r/m32r.c (init_reg_tables): Likewise.
126 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
127 enum to hold the modes.
129 2014-04-29 Richard Biener <rguenther@suse.de>
131 * dominance.c (free_dominance_info): Add overload with
133 (dom_info_state): Likewise.
134 (dom_info_available_p): Likewise.
135 * basic-block.h (free_dominance_info, dom_info_state,
136 dom_info_available_p): Declare overloads.
137 * passes.c (execute_function_todo): Verify that verifiers
138 don't change dominator info state. Drop dominator info
139 for IPA pass invocations.
140 * cgraph.c (release_function_body): Restore asserts that
141 dominator information is released.
143 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
145 * doc/invoke.texi: Fix typo.
146 * tree-vrp.c: Fix typos.
147 * gimple.c (infer_nonnull_range): Reorder operands of an &&
150 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
152 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
154 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
156 * config/aarch64/aarch64-builtins.c
157 (aarch64_types_storestruct_lane_qualifiers): New.
158 (TYPES_STORESTRUCT_LANE): Likewise.
159 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
160 (st3_lane): Likewise.
161 (st4_lane): Likewise.
162 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
163 (vec_store_lanesci_lane<mode>): Likewise.
164 (vec_store_lanesxi_lane<mode>): Likewise.
165 (aarch64_st2_lane<VQ:mode>): Likewise.
166 (aarch64_st3_lane<VQ:mode>): Likewise.
167 (aarch64_st4_lane<VQ:mode>): Likewise.
168 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
169 * config/aarch64/arm_neon.h
170 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
171 use new macro arguments.
172 (__ST3_LANE_FUNC): Likewise.
173 (__ST4_LANE_FUNC): Likewise.
174 * config/aarch64/iterators.md (V_TWO_ELEM): New.
175 (V_THREE_ELEM): Likewise.
176 (V_FOUR_ELEM): Likewise.
178 2014-04-28 David Malcolm <dmalcolm@redhat.com>
180 * doc/gimple.texi: Replace the description of the now-defunct
181 union gimple_statement_d with a diagram showing the
182 gimple_statement_base class hierarchy and its relationships to
183 the GSS_ and GIMPLE_ enums.
185 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
187 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
188 * config/aarch64/aarch64.c
189 (aarch64_cannot_change_mode_class): Weaken conditions.
190 (aarch64_modes_tieable_p): New.
191 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
193 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
195 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
196 (loadsync_<mode>): Change mode.
197 (load_quadpti, store_quadpti): New.
198 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
199 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
201 2014-04-28 Martin Jambor <mjambor@suse.cz>
203 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
204 same alias type as the original statement.
205 (subreplacement_assignment_data): New type.
206 (handle_unscalarized_data_in_subtree): New type of parameter,
207 generate new memory accesses with same alias type as the original
209 (load_assign_lhs_subreplacements): Likewise.
210 (sra_modify_constructor_assign): Generate new memory accesses with
211 same alias type as the original statement.
213 2014-04-28 Richard Biener <rguenther@suse.de>
215 * tree-pass.h (TODO_verify_il): Define.
216 (TODO_verify_all): Complete properly.
217 * passes.c (execute_function_todo): Move existing loop-closed
218 SSA verification under TODO_verify_il.
219 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
220 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
221 Fix tree sharing issue.
223 2014-04-28 Richard Biener <rguenther@suse.de>
226 * builtins.def (DEF_C11_BUILTIN): Add.
227 (BUILT_IN_ALIGNED_ALLOC): Likewise.
228 * coretypes.h (enum function_class): Add function_c11_misc.
229 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
230 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
231 (call_may_clobber_ref_p_1): Likewise.
232 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
233 (mark_all_reaching_defs_necessary_1): Likewise.
234 (propagate_necessity): Likewise.
235 (eliminate_unnecessary_stmts): Likewise.
236 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
238 2014-04-28 Richard Biener <rguenther@suse.de>
240 * tree-vrp.c (vrp_var_may_overflow): Remove.
241 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
242 with overflow immediately bump to one before that value and
243 let iteration figure out overflow status.
245 2014-04-28 Richard Biener <rguenther@suse.de>
247 * configure.ac: Do valgrind header checks unconditionally.
248 Add --enable-valgrind-annotations.
249 * system.h: Guard valgrind header inclusion with
250 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
251 * alloc-pool.c (pool_alloc, pool_free): Use
252 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
253 to guard possibly dead code.
254 * config.in: Regenerated.
255 * configure: Likewise.
257 2014-04-28 Jeff Law <law@redhat.com>
259 PR tree-optimization/60902
260 * tree-ssa-threadedge.c
261 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
262 over real defs when invalidating outputs from statements that do not
263 produce useful outputs for threading.
265 2014-04-28 Richard Biener <rguenther@suse.de>
267 PR tree-optimization/60979
268 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
269 SCOPs that end in a block with a successor with abnormal
272 2014-04-28 Richard Biener <rguenther@suse.de>
274 * tree-pass.h (execute_pass_list): Adjust prototype.
275 * passes.c (pass_manager::execute_early_local_passes):
277 (do_per_function): Change callback signature, push all actual
278 work to the callbals.
279 (do_per_function_toporder): Likewise.
280 (execute_function_dump): Adjust.
281 (execute_function_todo): Likewise.
282 (clear_last_verified): Likewise.
283 (verify_curr_properties): Likewise.
284 (update_properties_after_pass): Likewise.
285 (execute_pass_list_1): Split out from ...
286 (execute_pass_list): ... here. Adjust.
287 (execute_ipa_pass_list): Likewise.
288 * cgraphunit.c (cgraph_add_new_function): Adjust.
289 (analyze_function): Likewise.
290 (expand_function): Likewise.
291 * cgraph.c (release_function_body): Free dominance info
292 here instead of asserting it was magically freed elsewhere.
294 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
296 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
297 * configure: Regenerate.
298 * config/sparc/sparc.opt (muser-mode): New option.
299 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
301 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
302 * doc/invoke.texi (SPARC options): Document -muser-mode.
304 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
306 * cselib.c (find_slot_memmode): Delete.
307 (cselib_hasher): Change compare_type to a struct.
308 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
310 (preserve_constants_and_equivs): Adjust for new compare_type.
311 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
312 (wrap_constant): Delete.
313 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
315 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
317 * doc/install.texi (Building with profile feedback): Remove
320 2014-04-26 Tom de Vries <tom@codesourcery.com>
322 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
325 2014-04-25 Cary Coutant <ccoutant@google.com>
328 * dwarf2out.c (should_move_die_to_comdat): A type definition
329 can contain a subprogram definition, but don't move it to a
331 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
332 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
334 (clone_tree_hash): Rename to...
335 (clone_tree_partial): ...this; change callers. Copy
336 DW_TAG_subprogram DIEs as declarations.
337 (copy_decls_walk): Don't copy children of a declaration into a
340 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
343 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
346 2014-04-25 Jiong Wang <jiong.wang@arm.com>
348 * config/arm/predicates.md (call_insn_operand): Add long_call check.
349 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
351 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
354 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
356 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
359 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
361 PR tree-optimization/60930
362 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
363 creating a multiply candidate by folding two constant
364 multiplicands when the result overflows.
366 2014-04-25 Jakub Jelinek <jakub@redhat.com>
368 PR tree-optimization/60960
369 * tree-vect-generic.c (expand_vector_operation): Only call
370 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
372 2014-04-25 Tom de Vries <tom@codesourcery.com>
374 * expr.c (clobber_reg_mode): New function.
375 * expr.h (clobber_reg): New function.
377 2014-04-25 Tom de Vries <tom@codesourcery.com>
379 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
382 2014-04-25 Radovan Obradovic <robradovic@mips.com>
383 Tom de Vries <tom@codesourcery.com>
385 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
387 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
388 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
389 new argument to find_all_hard_reg_sets call.
391 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
393 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
394 Use HOST_WIDE_INT_C for mask literal.
395 (aarch_rev16_shleft_mask_imm_p): Likewise.
397 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
400 * config/sparc/sparc.md (ashlsi3_extend): Delete.
402 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
404 PR preprocessor/56540
405 * config/i386/i386-c.c (ix86_target_macros): Define
406 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
408 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
410 * configure.ac (tga_func): Remove.
411 (LIB_TLS_SPEC): Remove.
412 * configure: Regenerate.
413 * config.in: Regenerate.
414 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
416 2014-04-25 Richard Biener <rguenther@suse.de>
419 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
420 call stmt use/clobber sets during stmt walk instead of
421 walking the possibly incomplete set of caller edges.
423 2014-04-25 Richard Biener <rguenther@suse.de>
426 * passes.c (apply_ipa_transforms): Inline into only caller ...
427 (execute_one_pass): ... here. Properly bring in function
428 bodies for nodes we want to apply IPA transforms to.
430 2014-04-24 Cong Hou <congh@google.com>
432 PR tree-optimization/60896
433 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
434 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
435 (vect_mark_pattern_stmts): Set the def type of all statements in
436 PATTERN_DEF_SEQ as vect_internal_def.
438 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
440 * doc/extend.texi (PowerPC Built-in Functions): Document new
441 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
443 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
445 * config/rs6000/predicates.md (const_0_to_3_operand): New
446 predicate to match 0..3 integer constants.
448 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
449 to support adding miscellaneous builtin functions.
450 (BU_DFP_MISC_2): Likewise.
451 (BU_P7_MISC_1): Likewise.
452 (BU_P7_MISC_2): Likewise.
453 (BU_P8V_MISC_3): Likewise.
454 (BU_MISC_1): Likewise.
455 (BU_MISC_2): Likewise.
456 (DIVWE): Add extended divide builtin functions.
464 (DXEX): Add decimal floating-point builtin functions.
476 (CDTBCD): Add new BCD builtin functions.
480 (BCDADD_LT): Likewise.
481 (BCDADD_EQ): Likewise.
482 (BCDADD_GT): Likewise.
483 (BCDADD_OV): Likewise.
485 (BCDSUB_LT): Likewise.
486 (BCDSUB_EQ): Likewise.
487 (BCDSUB_GT): Likewise.
488 (BCDSUB_OV): Likewise.
489 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
490 (UNPACK_TD): Likewise.
492 (UNPACK_TF): Likewise.
493 (UNPACK_TF_0): Likewise.
494 (UNPACK_TF_1): Likewise.
495 (PACK_V1TI): Likewise.
496 (UNPACK_V1TI): Likewise.
498 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
499 support for decimal floating point builtin functions.
500 (rs6000_expand_ternop_builtin): Add checks for the new builtin
501 functions that take constant arguments.
502 (rs6000_invalid_builtin): Add decimal floating point builtin
504 (rs6000_init_builtins): Setup long double, _Decimal64, and
505 _Decimal128 types for new builtin functions.
506 (builtin_function_type): Set the unsigned flags appropriately for
507 the new builtin functions.
508 (rs6000_opt_masks): Add support for decimal floating point builtin
511 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
512 floating point builtin functions.
513 (RS6000_BTM_COMMON): Likewise.
514 (RS6000_BTI_long_double): Likewise.
515 (RS6000_BTI_dfloat64): Likewise.
516 (RS6000_BTI_dfloat128): Likewise.
517 (long_double_type_internal_node): Likewise.
518 (dfloat64_type_internal_node): Likewise.
519 (dfloat128_type_internal_node): Likewise.
521 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
522 2.07 bcd arithmetic instructions.
523 (UNSPEC_BCDSUB): Likewise.
524 (UNSPEC_BCD_OVERFLOW): Likewise.
525 (UNSPEC_BCD_ADD_SUB): Likewise.
526 (bcd_add_sub): Likewise.
527 (BCD_TEST): Likewise.
528 (bcd<bcd_add_sub>): Likewise.
529 (bcd<bcd_add_sub>_test): Likewise.
530 (bcd<bcd_add_sub>_test2): Likewise.
531 (bcd<bcd_add_sub>_<code>): Likewise.
532 (peephole2 for combined bcd ops): Likewise.
534 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
535 decimal floating point builtin functions.
536 (UNSPEC_DENBCD): Likewise.
537 (UNSPEC_DXEX): Likewise.
538 (UNSPEC_DIEX): Likewise.
539 (UNSPEC_DSCLI): Likewise.
540 (UNSPEC_DSCRI): Likewise.
541 (D64_D128): Likewise.
542 (dfp_suffix): Likewise.
543 (dfp_ddedpd_<mode>): Likewise.
544 (dfp_denbcd_<mode>): Likewise.
545 (dfp_dxex_<mode>): Likewise.
546 (dfp_diex_<mode>): Likewise.
547 (dfp_dscli_<mode>): Likewise.
548 (dfp_dscri_<mode>): Likewise.
550 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
552 (UNSPEC_CDTBCD): Likewise.
553 (UNSPEC_CBCDTD): Likewise.
554 (UNSPEC_DIVE): Add support for new extended divide builtin
556 (UNSPEC_DIVEO): Likewise.
557 (UNSPEC_DIVEU): Likewise.
558 (UNSPEC_DIVEUO): Likewise.
559 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
560 pack/unpack 128-bit types.
561 (UNSPEC_PACK_128BIT): Likewise.
562 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
563 (udiv<mode>3): Use idiv_ldiv mode attribute.
564 (div<mode>3): Likewise.
565 (addg6s): Add new BCD builtin functions.
568 (UNSPEC_DIV_EXTEND): Add support for new extended divide
570 (div_extend): Likewise.
571 (div<div_extend>_<mode>"): Likewise.
572 (FP128_64): Add support for new builtin functions to pack/unpack
574 (unpack<mode>): Likewise.
575 (unpacktf_0): Likewise.
576 (unpacktf_1): Likewise.
577 (unpack<mode>_dm): Likewise.
578 (unpack<mode>_nodm): Likewise.
579 (pack<mode>): Likewise.
580 (unpackv1ti): Likewise.
581 (packv1ti): Likewise.
583 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
585 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
588 2014-04-24 Jakub Jelinek <jakub@redhat.com>
590 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
591 * gimplify.c (omp_is_private): Change last argument's type to int.
592 Only diagnose lastprivate if the simd argument is 1, only diagnose
593 linear if the simd argument is 2.
594 (gimplify_omp_for): Adjust omp_is_private callers. When adding
595 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
596 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
597 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
598 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
599 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
600 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
601 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
602 * tree-nested.c (convert_nonlocal_omp_clauses,
603 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
605 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
608 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
611 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
613 * flag-types.h (enum ivar_visibility): Add.
615 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
617 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
620 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
622 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
624 2014-04-24 Radovan Obradovic <robradovic@mips.com>
625 Tom de Vries <tom@codesourcery.com>
627 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
628 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
630 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
631 * emit-rtl.c (try_split): Same.
633 2014-04-24 Radovan Obradovic <robradovic@mips.com>
634 Tom de Vries <tom@codesourcery.com>
636 * common.opt (fuse-caller-save): New option.
638 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
640 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
643 2014-04-24 Richard Biener <rguenther@suse.de>
645 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
646 during TER and instead use the sepops interface for expanding
647 non-GIMPLE_SINGLE_RHS.
649 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
651 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
652 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
654 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
656 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
657 assembler 64-bit option.
658 * configure: Regenerate.
660 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
662 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
663 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
664 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
665 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
666 (TARGET_CRYPTO): Take TARGET_SIMD into account.
668 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
670 * config/aarch64/aarch64-builtins.c
671 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
672 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
673 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
674 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
676 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
677 (Vrevsuff): New mode attribute.
679 2014-04-24 Terry Guo <terry.guo@arm.com>
681 * config/arm/arm.h (machine_function): Define variable
682 after_arm_reorg here.
683 * config/arm/arm.c (after_arm_reorg): Remove the definition.
684 (arm_split_constant): Update the way to access variable
687 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
689 2014-04-23 Tom de Vries <tom@codesourcery.com>
691 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
693 2014-04-23 David Malcolm <dmalcolm@redhat.com>
695 * is-a.h: Update comments to reflect the following changes to the
696 "pointerness" of the API, making the template parameter match the
697 return type, allowing use of is-a.h with typedefs of pointers.
698 (is_a_helper::cast): Return a T rather then a pointer to a T, so
699 that the return type matches the parameter to the is_a_helper.
701 (dyn_cast): Likewise.
703 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
704 pointer from the is-a.h API.
706 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
707 (is_a_helper <cgraph_node *>::test): ...this, matching change to
709 (is_a_helper <varpool_node>::test): Likewise, convert to...
710 (is_a_helper <varpool_node *>::test): ...this.
712 (varpool_first_variable): Update for removal of implicit pointer
714 (varpool_next_variable): Likewise.
715 (varpool_first_static_initializer): Likewise.
716 (varpool_next_static_initializer): Likewise.
717 (varpool_first_defined_variable): Likewise.
718 (varpool_next_defined_variable): Likewise.
719 (cgraph_first_defined_function): Likewise.
720 (cgraph_next_defined_function): Likewise.
721 (cgraph_first_function): Likewise.
722 (cgraph_next_function): Likewise.
723 (cgraph_first_function_with_gimple_body): Likewise.
724 (cgraph_next_function_with_gimple_body): Likewise.
725 (cgraph_alias_target): Likewise.
726 (varpool_alias_target): Likewise.
727 (cgraph_function_or_thunk_node): Likewise.
728 (varpool_variable_node): Likewise.
729 (symtab_real_symbol_p): Likewise.
730 * cgraphunit.c (referred_to_p): Likewise.
731 (analyze_functions): Likewise.
732 (handle_alias_pairs): Likewise.
733 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
734 * gimple-ssa.h (gimple_vuse_op): Likewise.
735 (gimple_vdef_op): Likewise.
736 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
737 * gimple.c (gimple_build_asm_1): Likewise.
738 (gimple_build_try): Likewise.
739 (gimple_build_resx): Likewise.
740 (gimple_build_eh_dispatch): Likewise.
741 (gimple_build_omp_for): Likewise.
742 (gimple_omp_for_set_clauses): Likewise.
744 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
745 (is_a_helper <gimple_statement_asm *>::test): ...this.
746 (is_a_helper <gimple_statement_bind>::test): Convert to...
747 (is_a_helper <gimple_statement_bind *>::test): ...this.
748 (is_a_helper <gimple_statement_call>::test): Convert to...
749 (is_a_helper <gimple_statement_call *>::test): ...this.
750 (is_a_helper <gimple_statement_catch>::test): Convert to...
751 (is_a_helper <gimple_statement_catch *>::test): ...this.
752 (is_a_helper <gimple_statement_resx>::test): Convert to...
753 (is_a_helper <gimple_statement_resx *>::test): ...this.
754 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
755 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
756 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
757 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
758 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
759 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
760 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
761 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
762 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
763 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
764 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
765 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
766 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
767 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
768 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
769 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
770 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
771 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
772 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
773 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
774 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
775 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
776 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
777 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
778 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
779 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
780 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
781 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
782 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
783 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
784 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
785 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
786 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
787 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
788 (is_a_helper <gimple_statement_phi>::test): Convert to...
789 (is_a_helper <gimple_statement_phi *>::test): ...this.
790 (is_a_helper <gimple_statement_transaction>::test): Convert to...
791 (is_a_helper <gimple_statement_transaction *>::test): ...this.
792 (is_a_helper <gimple_statement_try>::test): Convert to...
793 (is_a_helper <gimple_statement_try *>::test): ...this.
794 (is_a_helper <gimple_statement_wce>::test): Convert to...
795 (is_a_helper <gimple_statement_wce *>::test): ...this.
796 (is_a_helper <const gimple_statement_asm>::test): Convert to...
797 (is_a_helper <const gimple_statement_asm *>::test): ...this.
798 (is_a_helper <const gimple_statement_bind>::test): Convert to...
799 (is_a_helper <const gimple_statement_bind *>::test): ...this.
800 (is_a_helper <const gimple_statement_call>::test): Convert to...
801 (is_a_helper <const gimple_statement_call *>::test): ...this.
802 (is_a_helper <const gimple_statement_catch>::test): Convert to...
803 (is_a_helper <const gimple_statement_catch *>::test): ...this.
804 (is_a_helper <const gimple_statement_resx>::test): Convert to...
805 (is_a_helper <const gimple_statement_resx *>::test): ...this.
806 (is_a_helper <const gimple_statement_eh_dispatch>::test):
808 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
810 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
812 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
813 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
815 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
817 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
819 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
821 (is_a_helper <const gimple_statement_omp_return>::test): Convert
823 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
824 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
826 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
827 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
829 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
830 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
831 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
832 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
834 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
835 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
837 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
838 (is_a_helper <const gimple_statement_omp_target>::test): Convert
840 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
841 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
843 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
844 (is_a_helper <const gimple_statement_omp_single>::test): Convert
846 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
847 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
849 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
850 (is_a_helper <const gimple_statement_omp_task>::test): Convert
852 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
853 (is_a_helper <const gimple_statement_phi>::test): Convert to...
854 (is_a_helper <const gimple_statement_phi *>::test): ...this.
855 (is_a_helper <const gimple_statement_transaction>::test): Convert
857 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
858 (is_a_helper <const gimple_statement_with_ops>::test): Convert
860 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
861 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
862 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
863 (is_a_helper <const gimple_statement_with_memory_ops>::test):
865 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
867 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
869 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
871 (gimple_use_ops): Update for removal of implicit pointer from the
873 (gimple_set_use_ops): Likewise.
874 (gimple_vuse): Likewise.
875 (gimple_vdef): Likewise.
876 (gimple_vuse_ptr): Likewise.
877 (gimple_vdef_ptr): Likewise.
878 (gimple_set_vuse): Likewise.
879 (gimple_set_vdef): Likewise.
880 (gimple_omp_return_set_lhs): Likewise.
881 (gimple_omp_return_lhs): Likewise.
882 (gimple_omp_return_lhs_ptr): Likewise.
883 (gimple_call_fntype): Likewise.
884 (gimple_call_set_fntype): Likewise.
885 (gimple_call_set_internal_fn): Likewise.
886 (gimple_call_use_set): Likewise.
887 (gimple_call_clobber_set): Likewise.
888 (gimple_bind_vars): Likewise.
889 (gimple_bind_set_vars): Likewise.
890 (gimple_bind_body_ptr): Likewise.
891 (gimple_bind_set_body): Likewise.
892 (gimple_bind_add_stmt): Likewise.
893 (gimple_bind_block): Likewise.
894 (gimple_bind_set_block): Likewise.
895 (gimple_asm_ninputs): Likewise.
896 (gimple_asm_noutputs): Likewise.
897 (gimple_asm_nclobbers): Likewise.
898 (gimple_asm_nlabels): Likewise.
899 (gimple_asm_input_op): Likewise.
900 (gimple_asm_input_op_ptr): Likewise.
901 (gimple_asm_output_op): Likewise.
902 (gimple_asm_output_op_ptr): Likewise.
903 (gimple_asm_set_output_op): Likewise.
904 (gimple_asm_clobber_op): Likewise.
905 (gimple_asm_set_clobber_op): Likewise.
906 (gimple_asm_label_op): Likewise.
907 (gimple_asm_set_label_op): Likewise.
908 (gimple_asm_string): Likewise.
909 (gimple_catch_types): Likewise.
910 (gimple_catch_types_ptr): Likewise.
911 (gimple_catch_handler_ptr): Likewise.
912 (gimple_catch_set_types): Likewise.
913 (gimple_catch_set_handler): Likewise.
914 (gimple_eh_filter_types): Likewise.
915 (gimple_eh_filter_types_ptr): Likewise.
916 (gimple_eh_filter_failure_ptr): Likewise.
917 (gimple_eh_filter_set_types): Likewise.
918 (gimple_eh_filter_set_failure): Likewise.
919 (gimple_eh_must_not_throw_fndecl): Likewise.
920 (gimple_eh_must_not_throw_set_fndecl): Likewise.
921 (gimple_eh_else_n_body_ptr): Likewise.
922 (gimple_eh_else_e_body_ptr): Likewise.
923 (gimple_eh_else_set_n_body): Likewise.
924 (gimple_eh_else_set_e_body): Likewise.
925 (gimple_try_eval_ptr): Likewise.
926 (gimple_try_cleanup_ptr): Likewise.
927 (gimple_try_set_eval): Likewise.
928 (gimple_try_set_cleanup): Likewise.
929 (gimple_wce_cleanup_ptr): Likewise.
930 (gimple_wce_set_cleanup): Likewise.
931 (gimple_phi_capacity): Likewise.
932 (gimple_phi_num_args): Likewise.
933 (gimple_phi_result): Likewise.
934 (gimple_phi_result_ptr): Likewise.
935 (gimple_phi_set_result): Likewise.
936 (gimple_phi_arg): Likewise.
937 (gimple_phi_set_arg): Likewise.
938 (gimple_resx_region): Likewise.
939 (gimple_resx_set_region): Likewise.
940 (gimple_eh_dispatch_region): Likewise.
941 (gimple_eh_dispatch_set_region): Likewise.
942 (gimple_omp_critical_name): Likewise.
943 (gimple_omp_critical_name_ptr): Likewise.
944 (gimple_omp_critical_set_name): Likewise.
945 (gimple_omp_for_clauses): Likewise.
946 (gimple_omp_for_clauses_ptr): Likewise.
947 (gimple_omp_for_set_clauses): Likewise.
948 (gimple_omp_for_collapse): Likewise.
949 (gimple_omp_for_index): Likewise.
950 (gimple_omp_for_index_ptr): Likewise.
951 (gimple_omp_for_set_index): Likewise.
952 (gimple_omp_for_initial): Likewise.
953 (gimple_omp_for_initial_ptr): Likewise.
954 (gimple_omp_for_set_initial): Likewise.
955 (gimple_omp_for_final): Likewise.
956 (gimple_omp_for_final_ptr): Likewise.
957 (gimple_omp_for_set_final): Likewise.
958 (gimple_omp_for_incr): Likewise.
959 (gimple_omp_for_incr_ptr): Likewise.
960 (gimple_omp_for_set_incr): Likewise.
961 (gimple_omp_for_pre_body_ptr): Likewise.
962 (gimple_omp_for_set_pre_body): Likewise.
963 (gimple_omp_parallel_clauses): Likewise.
964 (gimple_omp_parallel_clauses_ptr): Likewise.
965 (gimple_omp_parallel_set_clauses): Likewise.
966 (gimple_omp_parallel_child_fn): Likewise.
967 (gimple_omp_parallel_child_fn_ptr): Likewise.
968 (gimple_omp_parallel_set_child_fn): Likewise.
969 (gimple_omp_parallel_data_arg): Likewise.
970 (gimple_omp_parallel_data_arg_ptr): Likewise.
971 (gimple_omp_parallel_set_data_arg): Likewise.
972 (gimple_omp_task_clauses): Likewise.
973 (gimple_omp_task_clauses_ptr): Likewise.
974 (gimple_omp_task_set_clauses): Likewise.
975 (gimple_omp_task_child_fn): Likewise.
976 (gimple_omp_task_child_fn_ptr): Likewise.
977 (gimple_omp_task_set_child_fn): Likewise.
978 (gimple_omp_task_data_arg): Likewise.
979 (gimple_omp_task_data_arg_ptr): Likewise.
980 (gimple_omp_task_set_data_arg): Likewise.
981 (gimple_omp_taskreg_clauses): Likewise.
982 (gimple_omp_taskreg_clauses_ptr): Likewise.
983 (gimple_omp_taskreg_set_clauses): Likewise.
984 (gimple_omp_taskreg_child_fn): Likewise.
985 (gimple_omp_taskreg_child_fn_ptr): Likewise.
986 (gimple_omp_taskreg_set_child_fn): Likewise.
987 (gimple_omp_taskreg_data_arg): Likewise.
988 (gimple_omp_taskreg_data_arg_ptr): Likewise.
989 (gimple_omp_taskreg_set_data_arg): Likewise.
990 (gimple_omp_task_copy_fn): Likewise.
991 (gimple_omp_task_copy_fn_ptr): Likewise.
992 (gimple_omp_task_set_copy_fn): Likewise.
993 (gimple_omp_task_arg_size): Likewise.
994 (gimple_omp_task_arg_size_ptr): Likewise.
995 (gimple_omp_task_set_arg_size): Likewise.
996 (gimple_omp_task_arg_align): Likewise.
997 (gimple_omp_task_arg_align_ptr): Likewise.
998 (gimple_omp_task_set_arg_align): Likewise.
999 (gimple_omp_single_clauses): Likewise.
1000 (gimple_omp_single_clauses_ptr): Likewise.
1001 (gimple_omp_single_set_clauses): Likewise.
1002 (gimple_omp_target_clauses): Likewise.
1003 (gimple_omp_target_clauses_ptr): Likewise.
1004 (gimple_omp_target_set_clauses): Likewise.
1005 (gimple_omp_target_child_fn): Likewise.
1006 (gimple_omp_target_child_fn_ptr): Likewise.
1007 (gimple_omp_target_set_child_fn): Likewise.
1008 (gimple_omp_target_data_arg): Likewise.
1009 (gimple_omp_target_data_arg_ptr): Likewise.
1010 (gimple_omp_target_set_data_arg): Likewise.
1011 (gimple_omp_teams_clauses): Likewise.
1012 (gimple_omp_teams_clauses_ptr): Likewise.
1013 (gimple_omp_teams_set_clauses): Likewise.
1014 (gimple_omp_sections_clauses): Likewise.
1015 (gimple_omp_sections_clauses_ptr): Likewise.
1016 (gimple_omp_sections_set_clauses): Likewise.
1017 (gimple_omp_sections_control): Likewise.
1018 (gimple_omp_sections_control_ptr): Likewise.
1019 (gimple_omp_sections_set_control): Likewise.
1020 (gimple_omp_for_set_cond): Likewise.
1021 (gimple_omp_for_cond): Likewise.
1022 (gimple_omp_atomic_store_set_val): Likewise.
1023 (gimple_omp_atomic_store_val): Likewise.
1024 (gimple_omp_atomic_store_val_ptr): Likewise.
1025 (gimple_omp_atomic_load_set_lhs): Likewise.
1026 (gimple_omp_atomic_load_lhs): Likewise.
1027 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1028 (gimple_omp_atomic_load_set_rhs): Likewise.
1029 (gimple_omp_atomic_load_rhs): Likewise.
1030 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1031 (gimple_omp_continue_control_def): Likewise.
1032 (gimple_omp_continue_control_def_ptr): Likewise.
1033 (gimple_omp_continue_set_control_def): Likewise.
1034 (gimple_omp_continue_control_use): Likewise.
1035 (gimple_omp_continue_control_use_ptr): Likewise.
1036 (gimple_omp_continue_set_control_use): Likewise.
1037 (gimple_transaction_body_ptr): Likewise.
1038 (gimple_transaction_label): Likewise.
1039 (gimple_transaction_label_ptr): Likewise.
1040 (gimple_transaction_set_body): Likewise.
1041 (gimple_transaction_set_label): Likewise.
1043 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1044 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1045 * ipa-ref.c (ipa_record_reference): Likewise.
1046 * ipa-reference.c (analyze_function): Likewise.
1047 (ipa_reference_write_optimization_summary): Likewise.
1048 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1049 (address_taken_from_non_vtable_p): Likewise.
1050 (comdat_can_be_unshared_p_1): Likewise.
1051 * lto-cgraph.c (lto_output_ref): Likewise.
1052 (add_references): Likewise.
1053 (compute_ltrans_boundary): Likewise.
1054 (output_symtab): Likewise.
1055 (input_ref): Likewise.
1056 (input_cgraph_1): Likewise.
1057 (output_cgraph_opt_summary): Likewise.
1058 * lto-streamer-out.c (lto_output): Likewise.
1059 (output_symbol_p): Likewise.
1060 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1061 (lsei_start_function_in_partition): Likewise.
1062 (lsei_next_variable_in_partition): Likewise.
1063 (lsei_start_variable_in_partition): Likewise.
1064 * symtab.c (insert_to_assembler_name_hash): Likewise.
1065 (unlink_from_assembler_name_hash): Likewise.
1066 (symtab_unregister_node): Likewise.
1067 (symtab_remove_node): Likewise.
1068 (dump_symtab_node): Likewise.
1069 (verify_symtab_base): Likewise.
1070 (verify_symtab_node): Likewise.
1071 (symtab_make_decl_local): Likewise.
1072 (symtab_alias_ultimate_target): Likewise.
1073 (symtab_resolve_alias): Likewise.
1074 (symtab_get_symbol_partitioning_class): Likewise.
1075 * tree-phinodes.c (allocate_phi_node): Likewise.
1076 (reserve_phi_args_for_new_edge): Likewise.
1077 (remove_phi_args): Likewise.
1078 * varpool.c (varpool_node_for_asm): Likewise.
1079 (varpool_remove_unreferenced_decls): Likewise.
1081 2014-04-23 Jeff Law <law@redhat.com>
1083 PR tree-optimization/60902
1084 * tree-ssa-threadedge.c
1085 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1086 invalidate outputs from statements that do not produce useful
1087 outputs for threading.
1089 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1091 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1092 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1093 machine descriptions for Stack Smashing Protector.
1095 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1097 * aarch64.md (<optab>_rol<mode>3): New pattern.
1098 (<optab>_rolsi3_uxtw): Likewise.
1099 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1101 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1103 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1104 (arm_cortex_a12_tune): Likewise.
1106 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1108 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1110 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1112 * config/arm/arm.md (arm_rev16si2): New pattern.
1113 (arm_rev16si2_alt): Likewise.
1114 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1116 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1118 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1119 (rev16<mode>2_alt): Likewise.
1120 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1121 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1122 (aarch_rev16_shleft_mask_imm_p): Likewise.
1123 (aarch_rev16_p_1): Likewise.
1124 (aarch_rev16_p): Likewise.
1125 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1126 (aarch_rev16_shright_mask_imm_p): Likewise.
1127 (aarch_rev16_shleft_mask_imm_p): Likewise.
1129 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1131 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1132 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1134 (cortex_a53_extra_costs): Likewise.
1135 (cortex_a57_extra_costs): Likewise.
1136 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1137 (cortexa7_extra_costs): Likewise.
1138 (cortexa8_extra_costs): Likewise.
1139 (cortexa12_extra_costs): Likewise.
1140 (cortexa15_extra_costs): Likewise.
1141 (v7m_extra_costs): Likewise.
1142 (arm_new_rtx_costs): Handle BSWAP.
1144 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1146 * config/arm/arm.c (cortexa8_extra_costs): New table.
1147 (arm_cortex_a8_tune): New tuning struct.
1148 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1150 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1152 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1154 2014-04-23 Richard Biener <rguenther@suse.de>
1156 * Makefile.in (OBJS): Remove loop-unswitch.o.
1157 * loop-unswitch.c: Delete.
1158 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1159 * passes.def (pass_rtl_unswitch): Likewise.
1160 * loop-init.c (gate_rtl_unswitch): Likewise.
1161 (rtl_unswitch): Likewise.
1162 (pass_data_rtl_unswitch): Likewise.
1163 (pass_rtl_unswitch): Likewise.
1164 (make_pass_rtl_unswitch): Likewise.
1165 * rtl.h (reversed_condition): Likewise.
1166 (compare_and_jump_seq): Likewise.
1167 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1169 * loop-unroll.c (compare_and_jump_seq): Likewise.
1171 2014-04-23 Richard Biener <rguenther@suse.de>
1173 PR tree-optimization/60903
1174 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1175 commented code block.
1176 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1177 loop flags to newly created BBs and edges.
1179 2014-04-23 Nick Clifton <nickc@redhat.com>
1181 * config/msp430/msp430.c (msp430_handle_option): Move function
1183 (msp430_option_override): Simplify mcu and mcpu option handling.
1184 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1185 support for -mhwmult command line option.
1186 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1187 -mhwmult command line option.
1188 (msp430_hwmult_enabled): Delete.
1189 (msp43o_output_labelref): Add support for -mhwmult command line
1191 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1192 (umulsidi3): Likewise.
1193 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1194 (mcpu, mlarge, msmall): Likewise.
1195 (mhwmult): New option.
1196 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1198 (msp430_is_f5_mcu): Remove prototype.
1199 (msp430_use_f5_series_hwmult): Add prototype.
1200 * config/msp430/msp430-opts.h: New file.
1201 * common/config/msp430: New directory.
1202 * common/config/msp430/msp430-common.c: New file.
1203 * config.gcc (msp430): Remove target_has_targetm_common.
1204 * doc/invoke.texi: Document -mhwmult command line option.
1206 2014-04-23 Nick Clifton <nickc@redhat.com>
1208 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1209 default-manifest.o if it can be found in the search path.
1210 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1212 2014-04-23 Terry Guo <terry.guo@arm.com>
1214 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1216 2014-04-23 Richard Biener <rguenther@suse.de>
1219 * tree-inline.c (declare_return_variable): Use mark_addressable.
1221 2014-04-23 Richard Biener <rguenther@suse.de>
1224 * loop-init.c (loop_optimizer_init): Make sure to apply
1225 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1227 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1230 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1232 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1233 if flag_sanitize_undefined_trap_on_error.
1234 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1235 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1236 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1237 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1238 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1239 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1240 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1241 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1242 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1243 * ubsan.c (ubsan_instrument_unreachable): Return
1244 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1245 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1246 if flag_sanitize_undefined_trap_on_error and
1247 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1248 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1249 instrument_bool_enum_load): Emit __builtin_trap () if
1250 flag_sanitize_undefined_trap_on_error and
1251 __builtin_handle_*_abort () if !flag_sanitize_recover.
1252 * doc/invoke.texi (-fsanitize-recover,
1253 -fsanitize-undefined-trap-on-error): Document.
1255 2014-04-22 Christian Bruel <christian.bruel@st.com>
1257 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1258 Force immediates to SImode.
1260 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1262 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1264 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1265 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1266 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1267 (nios2_fpu_insn): Add entry for round.
1268 (N2FPU_NO_ERRNO_P): Define.
1269 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1271 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1273 2014-04-22 Richard Henderson <rth@redhat.com>
1275 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1276 (add<GPI>3_compare0): Remove leading * from name.
1277 (add<GPI>3_carryin): Likewise.
1278 (sub<GPI>3_compare0): Likewise.
1279 (sub<GPI>3_carryin): Likewise.
1280 (<su_optab>mulditi3): New expander.
1281 (multi3): New expander.
1282 (madd<GPI>): Remove leading * from name.
1284 2014-04-22 Martin Jambor <mjambor@suse.cz>
1286 * cgraphclones.c (cgraph_function_versioning): Copy
1287 ipa_transforms_to_apply instead of asserting it is empty.
1289 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1292 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1293 on count_exp to get mode.
1295 2014-04-22 Andrew Pinski <apinski@cavium.com>
1297 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1298 Handle TLS for ILP32.
1299 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1300 (tlsie_small_<mode>): this and handle PTR.
1301 (tlsie_small_sidi): New pattern.
1302 (tlsle_small): Change to an expand to handle ILP32.
1303 (tlsle_small_<mode>): New pattern.
1304 (tlsdesc_small): Rename to ...
1305 (tlsdesc_small_<mode>): this and handle PTR.
1307 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1309 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1311 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1313 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1314 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1315 (aarch64_types_signed_poly_qualifiers): Likewise.
1316 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1317 (aarch64_types_poly_signed_qualifiers): Likewise.
1318 (TYPES_REINTERP_SS): Type macro added.
1319 (TYPES_REINTERP_SU): Likewise.
1320 (TYPES_REINTERP_SP): Likewise.
1321 (TYPES_REINTERP_US): Likewise.
1322 (TYPES_REINTERP_PS): Likewise.
1323 (aarch64_fold_builtin): New expression folding added.
1324 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1325 Declarations removed.
1326 (REINTERP_SS): Declarations added.
1327 (REINTERP_US): Likewise.
1328 (REINTERP_PS): Likewise.
1329 (REINTERP_SU): Likewise.
1330 (REINTERP_SP): Likewise.
1331 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1332 (vreinterpretq_p8_f64): Likewise.
1333 (vreinterpret_p16_f64): Likewise.
1334 (vreinterpretq_p16_f64): Likewise.
1335 (vreinterpret_f32_f64): Likewise.
1336 (vreinterpretq_f32_f64): Likewise.
1337 (vreinterpret_f64_f32): Likewise.
1338 (vreinterpret_f64_p8): Likewise.
1339 (vreinterpret_f64_p16): Likewise.
1340 (vreinterpret_f64_s8): Likewise.
1341 (vreinterpret_f64_s16): Likewise.
1342 (vreinterpret_f64_s32): Likewise.
1343 (vreinterpret_f64_s64): Likewise.
1344 (vreinterpret_f64_u8): Likewise.
1345 (vreinterpret_f64_u16): Likewise.
1346 (vreinterpret_f64_u32): Likewise.
1347 (vreinterpret_f64_u64): Likewise.
1348 (vreinterpretq_f64_f32): Likewise.
1349 (vreinterpretq_f64_p8): Likewise.
1350 (vreinterpretq_f64_p16): Likewise.
1351 (vreinterpretq_f64_s8): Likewise.
1352 (vreinterpretq_f64_s16): Likewise.
1353 (vreinterpretq_f64_s32): Likewise.
1354 (vreinterpretq_f64_s64): Likewise.
1355 (vreinterpretq_f64_u8): Likewise.
1356 (vreinterpretq_f64_u16): Likewise.
1357 (vreinterpretq_f64_u32): Likewise.
1358 (vreinterpretq_f64_u64): Likewise.
1359 (vreinterpret_s64_f64): Likewise.
1360 (vreinterpretq_s64_f64): Likewise.
1361 (vreinterpret_u64_f64): Likewise.
1362 (vreinterpretq_u64_f64): Likewise.
1363 (vreinterpret_s8_f64): Likewise.
1364 (vreinterpretq_s8_f64): Likewise.
1365 (vreinterpret_s16_f64): Likewise.
1366 (vreinterpretq_s16_f64): Likewise.
1367 (vreinterpret_s32_f64): Likewise.
1368 (vreinterpretq_s32_f64): Likewise.
1369 (vreinterpret_u8_f64): Likewise.
1370 (vreinterpretq_u8_f64): Likewise.
1371 (vreinterpret_u16_f64): Likewise.
1372 (vreinterpretq_u16_f64): Likewise.
1373 (vreinterpret_u32_f64): Likewise.
1374 (vreinterpretq_u32_f64): Likewise.
1376 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1378 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1379 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1380 (vreinterpret_p8_s8): Likewise.
1381 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1382 (vreinterpret_p8_s16): Likewise.
1383 (vreinterpret_p8_s32): Likewise.
1384 (vreinterpret_p8_s64): Likewise.
1385 (vreinterpret_p8_f32): Likewise.
1386 (vreinterpret_p8_u8): Likewise.
1387 (vreinterpret_p8_u16): Likewise.
1388 (vreinterpret_p8_u32): Likewise.
1389 (vreinterpret_p8_u64): Likewise.
1390 (vreinterpret_p8_p16): Likewise.
1391 (vreinterpretq_p8_s8): Likewise.
1392 (vreinterpretq_p8_s16): Likewise.
1393 (vreinterpretq_p8_s32): Likewise.
1394 (vreinterpretq_p8_s64): Likewise.
1395 (vreinterpretq_p8_f32): Likewise.
1396 (vreinterpretq_p8_u8): Likewise.
1397 (vreinterpretq_p8_u16): Likewise.
1398 (vreinterpretq_p8_u32): Likewise.
1399 (vreinterpretq_p8_u64): Likewise.
1400 (vreinterpretq_p8_p16): Likewise.
1401 (vreinterpret_p16_s8): Likewise.
1402 (vreinterpret_p16_s16): Likewise.
1403 (vreinterpret_p16_s32): Likewise.
1404 (vreinterpret_p16_s64): Likewise.
1405 (vreinterpret_p16_f32): Likewise.
1406 (vreinterpret_p16_u8): Likewise.
1407 (vreinterpret_p16_u16): Likewise.
1408 (vreinterpret_p16_u32): Likewise.
1409 (vreinterpret_p16_u64): Likewise.
1410 (vreinterpret_p16_p8): Likewise.
1411 (vreinterpretq_p16_s8): Likewise.
1412 (vreinterpretq_p16_s16): Likewise.
1413 (vreinterpretq_p16_s32): Likewise.
1414 (vreinterpretq_p16_s64): Likewise.
1415 (vreinterpretq_p16_f32): Likewise.
1416 (vreinterpretq_p16_u8): Likewise.
1417 (vreinterpretq_p16_u16): Likewise.
1418 (vreinterpretq_p16_u32): Likewise.
1419 (vreinterpretq_p16_u64): Likewise.
1420 (vreinterpretq_p16_p8): Likewise.
1421 (vreinterpret_f32_s8): Likewise.
1422 (vreinterpret_f32_s16): Likewise.
1423 (vreinterpret_f32_s32): Likewise.
1424 (vreinterpret_f32_s64): Likewise.
1425 (vreinterpret_f32_u8): Likewise.
1426 (vreinterpret_f32_u16): Likewise.
1427 (vreinterpret_f32_u32): Likewise.
1428 (vreinterpret_f32_u64): Likewise.
1429 (vreinterpret_f32_p8): Likewise.
1430 (vreinterpret_f32_p16): Likewise.
1431 (vreinterpretq_f32_s8): Likewise.
1432 (vreinterpretq_f32_s16): Likewise.
1433 (vreinterpretq_f32_s32): Likewise.
1434 (vreinterpretq_f32_s64): Likewise.
1435 (vreinterpretq_f32_u8): Likewise.
1436 (vreinterpretq_f32_u16): Likewise.
1437 (vreinterpretq_f32_u32): Likewise.
1438 (vreinterpretq_f32_u64): Likewise.
1439 (vreinterpretq_f32_p8): Likewise.
1440 (vreinterpretq_f32_p16): Likewise.
1441 (vreinterpret_s64_s8): Likewise.
1442 (vreinterpret_s64_s16): Likewise.
1443 (vreinterpret_s64_s32): Likewise.
1444 (vreinterpret_s64_f32): Likewise.
1445 (vreinterpret_s64_u8): Likewise.
1446 (vreinterpret_s64_u16): Likewise.
1447 (vreinterpret_s64_u32): Likewise.
1448 (vreinterpret_s64_u64): Likewise.
1449 (vreinterpret_s64_p8): Likewise.
1450 (vreinterpret_s64_p16): Likewise.
1451 (vreinterpretq_s64_s8): Likewise.
1452 (vreinterpretq_s64_s16): Likewise.
1453 (vreinterpretq_s64_s32): Likewise.
1454 (vreinterpretq_s64_f32): Likewise.
1455 (vreinterpretq_s64_u8): Likewise.
1456 (vreinterpretq_s64_u16): Likewise.
1457 (vreinterpretq_s64_u32): Likewise.
1458 (vreinterpretq_s64_u64): Likewise.
1459 (vreinterpretq_s64_p8): Likewise.
1460 (vreinterpretq_s64_p16): Likewise.
1461 (vreinterpret_u64_s8): Likewise.
1462 (vreinterpret_u64_s16): Likewise.
1463 (vreinterpret_u64_s32): Likewise.
1464 (vreinterpret_u64_s64): Likewise.
1465 (vreinterpret_u64_f32): Likewise.
1466 (vreinterpret_u64_u8): Likewise.
1467 (vreinterpret_u64_u16): Likewise.
1468 (vreinterpret_u64_u32): Likewise.
1469 (vreinterpret_u64_p8): Likewise.
1470 (vreinterpret_u64_p16): Likewise.
1471 (vreinterpretq_u64_s8): Likewise.
1472 (vreinterpretq_u64_s16): Likewise.
1473 (vreinterpretq_u64_s32): Likewise.
1474 (vreinterpretq_u64_s64): Likewise.
1475 (vreinterpretq_u64_f32): Likewise.
1476 (vreinterpretq_u64_u8): Likewise.
1477 (vreinterpretq_u64_u16): Likewise.
1478 (vreinterpretq_u64_u32): Likewise.
1479 (vreinterpretq_u64_p8): Likewise.
1480 (vreinterpretq_u64_p16): Likewise.
1481 (vreinterpret_s8_s16): Likewise.
1482 (vreinterpret_s8_s32): Likewise.
1483 (vreinterpret_s8_s64): Likewise.
1484 (vreinterpret_s8_f32): Likewise.
1485 (vreinterpret_s8_u8): Likewise.
1486 (vreinterpret_s8_u16): Likewise.
1487 (vreinterpret_s8_u32): Likewise.
1488 (vreinterpret_s8_u64): Likewise.
1489 (vreinterpret_s8_p8): Likewise.
1490 (vreinterpret_s8_p16): Likewise.
1491 (vreinterpretq_s8_s16): Likewise.
1492 (vreinterpretq_s8_s32): Likewise.
1493 (vreinterpretq_s8_s64): Likewise.
1494 (vreinterpretq_s8_f32): Likewise.
1495 (vreinterpretq_s8_u8): Likewise.
1496 (vreinterpretq_s8_u16): Likewise.
1497 (vreinterpretq_s8_u32): Likewise.
1498 (vreinterpretq_s8_u64): Likewise.
1499 (vreinterpretq_s8_p8): Likewise.
1500 (vreinterpretq_s8_p16): Likewise.
1501 (vreinterpret_s16_s8): Likewise.
1502 (vreinterpret_s16_s32): Likewise.
1503 (vreinterpret_s16_s64): Likewise.
1504 (vreinterpret_s16_f32): Likewise.
1505 (vreinterpret_s16_u8): Likewise.
1506 (vreinterpret_s16_u16): Likewise.
1507 (vreinterpret_s16_u32): Likewise.
1508 (vreinterpret_s16_u64): Likewise.
1509 (vreinterpret_s16_p8): Likewise.
1510 (vreinterpret_s16_p16): Likewise.
1511 (vreinterpretq_s16_s8): Likewise.
1512 (vreinterpretq_s16_s32): Likewise.
1513 (vreinterpretq_s16_s64): Likewise.
1514 (vreinterpretq_s16_f32): Likewise.
1515 (vreinterpretq_s16_u8): Likewise.
1516 (vreinterpretq_s16_u16): Likewise.
1517 (vreinterpretq_s16_u32): Likewise.
1518 (vreinterpretq_s16_u64): Likewise.
1519 (vreinterpretq_s16_p8): Likewise.
1520 (vreinterpretq_s16_p16): Likewise.
1521 (vreinterpret_s32_s8): Likewise.
1522 (vreinterpret_s32_s16): Likewise.
1523 (vreinterpret_s32_s64): Likewise.
1524 (vreinterpret_s32_f32): Likewise.
1525 (vreinterpret_s32_u8): Likewise.
1526 (vreinterpret_s32_u16): Likewise.
1527 (vreinterpret_s32_u32): Likewise.
1528 (vreinterpret_s32_u64): Likewise.
1529 (vreinterpret_s32_p8): Likewise.
1530 (vreinterpret_s32_p16): Likewise.
1531 (vreinterpretq_s32_s8): Likewise.
1532 (vreinterpretq_s32_s16): Likewise.
1533 (vreinterpretq_s32_s64): Likewise.
1534 (vreinterpretq_s32_f32): Likewise.
1535 (vreinterpretq_s32_u8): Likewise.
1536 (vreinterpretq_s32_u16): Likewise.
1537 (vreinterpretq_s32_u32): Likewise.
1538 (vreinterpretq_s32_u64): Likewise.
1539 (vreinterpretq_s32_p8): Likewise.
1540 (vreinterpretq_s32_p16): Likewise.
1541 (vreinterpret_u8_s8): Likewise.
1542 (vreinterpret_u8_s16): Likewise.
1543 (vreinterpret_u8_s32): Likewise.
1544 (vreinterpret_u8_s64): Likewise.
1545 (vreinterpret_u8_f32): Likewise.
1546 (vreinterpret_u8_u16): Likewise.
1547 (vreinterpret_u8_u32): Likewise.
1548 (vreinterpret_u8_u64): Likewise.
1549 (vreinterpret_u8_p8): Likewise.
1550 (vreinterpret_u8_p16): Likewise.
1551 (vreinterpretq_u8_s8): Likewise.
1552 (vreinterpretq_u8_s16): Likewise.
1553 (vreinterpretq_u8_s32): Likewise.
1554 (vreinterpretq_u8_s64): Likewise.
1555 (vreinterpretq_u8_f32): Likewise.
1556 (vreinterpretq_u8_u16): Likewise.
1557 (vreinterpretq_u8_u32): Likewise.
1558 (vreinterpretq_u8_u64): Likewise.
1559 (vreinterpretq_u8_p8): Likewise.
1560 (vreinterpretq_u8_p16): Likewise.
1561 (vreinterpret_u16_s8): Likewise.
1562 (vreinterpret_u16_s16): Likewise.
1563 (vreinterpret_u16_s32): Likewise.
1564 (vreinterpret_u16_s64): Likewise.
1565 (vreinterpret_u16_f32): Likewise.
1566 (vreinterpret_u16_u8): Likewise.
1567 (vreinterpret_u16_u32): Likewise.
1568 (vreinterpret_u16_u64): Likewise.
1569 (vreinterpret_u16_p8): Likewise.
1570 (vreinterpret_u16_p16): Likewise.
1571 (vreinterpretq_u16_s8): Likewise.
1572 (vreinterpretq_u16_s16): Likewise.
1573 (vreinterpretq_u16_s32): Likewise.
1574 (vreinterpretq_u16_s64): Likewise.
1575 (vreinterpretq_u16_f32): Likewise.
1576 (vreinterpretq_u16_u8): Likewise.
1577 (vreinterpretq_u16_u32): Likewise.
1578 (vreinterpretq_u16_u64): Likewise.
1579 (vreinterpretq_u16_p8): Likewise.
1580 (vreinterpretq_u16_p16): Likewise.
1581 (vreinterpret_u32_s8): Likewise.
1582 (vreinterpret_u32_s16): Likewise.
1583 (vreinterpret_u32_s32): Likewise.
1584 (vreinterpret_u32_s64): Likewise.
1585 (vreinterpret_u32_f32): Likewise.
1586 (vreinterpret_u32_u8): Likewise.
1587 (vreinterpret_u32_u16): Likewise.
1588 (vreinterpret_u32_u64): Likewise.
1589 (vreinterpret_u32_p8): Likewise.
1590 (vreinterpret_u32_p16): Likewise.
1591 (vreinterpretq_u32_s8): Likewise.
1592 (vreinterpretq_u32_s16): Likewise.
1593 (vreinterpretq_u32_s32): Likewise.
1594 (vreinterpretq_u32_s64): Likewise.
1595 (vreinterpretq_u32_f32): Likewise.
1596 (vreinterpretq_u32_u8): Likewise.
1597 (vreinterpretq_u32_u16): Likewise.
1598 (vreinterpretq_u32_u64): Likewise.
1599 (vreinterpretq_u32_p8): Likewise.
1600 (vreinterpretq_u32_p16): Likewise.
1602 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1604 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1606 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1609 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1610 (vqnegd_s64): Likewise.
1611 (vqabs_s64): Likewise.
1612 (vqabsd_s64): Likewise.
1614 2014-04-22 Richard Henderson <rth@redhat.com>
1616 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1617 computation to the top of the loop.
1619 2014-04-22 Renlin <renlin.li@arm.com>
1620 Jiong Wang <jiong.wang@arm.com>
1622 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1623 * config/aarch64/aarch64.c (aarch64_layout_frame)
1624 (aarch64_initial_elimination_offset): Likewise.
1626 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1628 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1631 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1633 * machmode.h (bitwise_mode_for_mode): Declare.
1634 * stor-layout.h (bitwise_type_for_mode): Likewise.
1635 * stor-layout.c (bitwise_mode_for_mode): New function.
1636 (bitwise_type_for_mode): Likewise.
1637 * builtins.c (fold_builtin_memory_op): Use it instead of
1638 int_mode_for_mode and build_nonstandard_integer_type.
1640 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1642 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1643 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1644 (*-*-solaris2*): Simplify.
1645 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1646 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1647 *-*-solaris2.9* handling.
1649 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1651 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1652 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1654 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1655 * configure: Regenerate.
1657 * config/i386/sol2-9.h: Remove.
1659 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1660 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1661 Remove Solaris 9 references.
1663 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1665 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1666 (floatuns<GPI:mode><GPF:mode>2): Remove.
1667 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1668 and floatuns conversions.
1669 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1670 and floatuns conversions.
1671 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1672 (w1,w2): New mode attributes for inequal width conversions.
1674 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1676 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1677 the output asm format.
1679 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1681 * config/aarch64/aarch64-simd.md
1682 (aarch64_cm<optab>di): Always split.
1683 (*aarch64_cm<optab>di): New.
1684 (aarch64_cmtstdi): Always split.
1685 (*aarch64_cmtstdi): New.
1687 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1689 PR tree-optimization/60823
1690 * omp-low.c (ipa_simd_modify_function_body): Go through
1691 all SSA_NAMEs and for those refering to vector arguments
1692 which are going to be replaced adjust SSA_NAME_VAR and,
1693 if it is a default definition, change it into a non-default
1694 definition assigned at the beginning of function from new_decl.
1695 (ipa_simd_modify_stmt_ops): Rewritten.
1696 * tree-dfa.c (set_ssa_default_def): When removing default def,
1697 check for NULL loc instead of NULL *loc.
1699 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1701 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1702 restrictions on core registers for DImode values in Thumb2.
1704 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1706 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1707 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1709 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1711 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1712 (*iordi_notzesidi_di): Likewise.
1713 (*iordi_notsesidi_di): Likewise.
1715 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1717 * config/arm/arm-protos.h (tune_params): New struct members.
1718 * config/arm/arm.c: Initialise tune_params per processor.
1719 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1720 for speed, based on new tune_params.
1722 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1724 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1726 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1728 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1730 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1731 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1732 (vrnda_f64): Likewise.
1733 (vrndi_f64): Likewise.
1734 (vrndm_f64): Likewise.
1735 (vrndn_f64): Likewise.
1736 (vrndp_f64): Likewise.
1737 (vrndx_f64): Likewise.
1739 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1741 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1742 GET_MODE_SIZE argument is enum machine_mode.
1744 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1747 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1748 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1750 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1753 * asan.c (asan_emit_stack_protection): Force the base to align to
1754 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1755 appropriate bits if STRICT_ALIGNMENT.
1756 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1758 (expand_used_vars): Leave a space in the stack frame for alignment
1759 if STRICT_ALIGNMENT.
1761 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1763 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1765 (gimple_store_p): Likewise.
1766 (gimple_assign_load_p): Likewise.
1767 (gimple_assign_cast_p): Likewise.
1768 (gimple_clobber_p): Likewise.
1770 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1771 rather than a gimple.
1772 (gimple_assign_cast_p): Likewise.
1774 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1777 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1778 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1780 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1781 more debug information for E500 if -mdebug=reg.
1783 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1786 * config/i386/i386.c (ix86_expand_builtin)
1787 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1788 register for target RTX.
1789 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1791 2014-04-18 Cong Hou <congh@google.com>
1793 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1794 the widen-mult pattern by handling two operands with different sizes,
1795 and operands whose size is smaller than half of the result type.
1797 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1799 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1800 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1801 (do_estimate_edge_time): Compute it.
1802 * ipa-inline.c (want_inline_small_function_p): Bypass
1803 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1805 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1807 * ipa-inline.c (spec_rem): New static variable.
1808 (dump_overall_stats): New function.
1809 (dump_inline_stats): New function.
1811 2014-04-18 Richard Henderson <rth@redhat.com>
1813 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1814 to GET_MODE_SIZE, not a reg_class_t.
1816 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1818 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1819 (vsx_xxmrglw_<mode>): Likewise.
1821 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1824 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1825 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1826 (rs6000_init_hard_regno_mode_ok): Likewise.
1828 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1830 * ipa-inline.c (inline_small_functions): Account only non-cold
1832 * doc/invoke.texi (inline-unit-growth): Update documentation.
1834 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1836 * config/rs6000/rs6000.md (addti3, subti3): New.
1838 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1841 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1842 comment. Check optimize_insn_for_size_p instead of
1843 optimize_insn_for_speed_p.
1845 2014-04-17 Martin Jambor <mjambor@suse.cz>
1847 * gimple-iterator.c (gsi_start_edge): New function.
1848 * gimple-iterator.h (gsi_start_edge): Declare.
1849 * tree-sra.c (single_non_eh_succ): New function.
1850 (disqualify_ops_if_throwing_stmt): Renamed to
1851 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1852 having one non-EH successor BB.
1853 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1854 generate loads into replacements.
1855 (sra_modify_assign): Likewise and and also use the simple path for
1857 (sra_modify_function_body): Commit statements on edges.
1859 2014-04-17 Richard Biener <rguenther@suse.de>
1862 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1863 comparison results and add clarifying comment.
1865 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1867 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1868 (blank_mode): Initialize it.
1869 (emit_mode_size_inline, emit_mode_nunits_inline,
1870 emit_mode_inner_inline): New functions.
1871 (emit_insn_modes_h): Call them and surround their output with
1872 #if GCC_VERSION >= 4001 ... #endif.
1873 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1874 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1875 mode_* arrays if the argument is __builtin_constant_p.
1876 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1877 is enum machine_mode.
1879 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1881 * passes.c (opt_pass::execute): Adjust.
1882 (pass_manager::execute_pass_mode_switching): Likewise.
1883 (early_local_passes::execute): Likewise.
1884 (execute_one_pass): Pass cfun to the pass's execute method.
1885 * tree-pass.h (opt_pass::execute): Add function * argument.
1886 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1887 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1888 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1889 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1890 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1891 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1892 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1893 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1894 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1895 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1896 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1897 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1898 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1899 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1900 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1901 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1902 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1903 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1904 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1905 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1906 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1907 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1908 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1909 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1910 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1911 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1912 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1913 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1914 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1917 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1919 * passes.c (opt_pass::gate): Take function * argument.
1920 (gate_all_early_local_passes): Merge into
1921 (early_local_passes::gate): this.
1922 (gate_all_early_optimizations): Merge into
1923 (all_early_optimizations::gate): this.
1924 (gate_all_optimizations): Mege into
1925 (all_optimizations::gate): this.
1926 (gate_all_optimizations_g): Merge into
1927 (all_optimizations_g::gate): this.
1928 (gate_rest_of_compilation): Mege into
1929 (rest_of_compilation::gate): this.
1930 (gate_postreload): Merge into
1931 (postreload::gate): this.
1932 (dump_one_pass): Pass cfun to the pass's gate method.
1933 (execute_ipa_summary_passes): Likewise.
1934 (execute_one_pass): Likewise.
1935 (ipa_write_summaries_2): Likewise.
1936 (ipa_write_optimization_summaries_1): Likewise.
1937 (ipa_read_summaries_1): Likewise.
1938 (ipa_read_optimization_summaries_1): Likewise.
1939 (execute_ipa_stmt_fixups): Likewise.
1940 * tree-pass.h (opt_pass::gate): Add function * argument.
1941 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1942 combine-stack-adj.c, combine.c, compare-elim.c,
1943 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1944 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1945 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1946 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1947 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1948 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1949 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1950 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1951 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1952 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1953 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1954 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1955 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1956 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1957 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1958 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1959 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1960 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1961 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1962 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1963 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1964 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1965 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1966 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1967 var-tracking.c, vtable-verify.c, web.c: Adjust.
1969 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1971 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
1972 * configure: Regenerate.
1974 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1976 * passes.c (dump_one_pass): don't check pass->has_gate.
1977 (execute_ipa_summary_passes): Likewise.
1978 (execute_one_pass): Likewise.
1979 (ipa_write_summaries_2): Likewise.
1980 (ipa_write_optimization_summaries_1): Likewise.
1981 (ipa_read_optimization_summaries_1): Likewise.
1982 (execute_ipa_stmt_fixups): Likewise.
1983 * tree-pass.h (pass_data::has_gate): Remove.
1984 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1985 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1986 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1987 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1988 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1989 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
1990 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
1991 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
1992 gimple-low.c, gimple-ssa-isolate-paths.c,
1993 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
1994 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
1995 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1996 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
1997 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
1998 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
1999 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2000 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2001 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2002 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2003 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2004 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2005 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2006 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2007 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2008 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2009 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2010 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2011 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2012 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2013 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2014 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2017 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2019 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2021 * passes.c (pass_manager::register_dump_files_1): Merge into
2022 (pass_manager::register_dump_files): this, and remove its handling of
2023 properties since the pass always has the properties anyway.
2024 (pass_manager::pass_manager): Adjust.
2026 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2028 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2029 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2030 dealing with properties.
2031 (pass_manager::register_dump_files): Adjust.
2033 2014-03-20 Mark Wielaard <mjw@redhat.com>
2035 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2036 then represent the bound as normal constant value.
2038 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2041 Forward port from 4.8 branch
2042 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2044 * config/i386/bmiintrin.h (_blsi_u32): New.
2048 (_blsmsk_u32): Ditto.
2049 (_blsmsk_u64): Ditto.
2050 (_tzcnt_u32): Ditto.
2051 (_tzcnt_u64): Ditto.
2053 2014-04-17 Kito Cheng <kito@0xlab.org>
2055 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2057 2014-04-17 Richard Biener <rguenther@suse.de>
2060 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2061 boolean results for comparisons.
2063 2014-04-17 Richard Biener <rguenther@suse.de>
2065 PR tree-optimization/60836
2066 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2067 initial PHI args to be gimple values.
2069 2014-04-17 Richard Biener <rguenther@suse.de>
2071 PR tree-optimization/60841
2072 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2073 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2074 of stmts to SLP build.
2075 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2076 (vect_analyze_slp): Likewise.
2077 (vect_analyze_slp_instance): Likewise.
2078 (vect_build_slp_tree): Limit overall SLP tree growth.
2079 * tree-vectorizer.h (vect_analyze_data_refs,
2080 vect_analyze_slp): Adjust prototypes.
2082 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2084 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2087 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2089 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2090 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2091 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2092 for TARGET_SLOW_PSHUFB
2094 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2096 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2097 * config/i386/i386.c (intel_cost): Ditto.
2099 2014-04-17 Joey Ye <joey.ye@arm.com>
2101 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2103 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2105 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2108 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2110 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2111 (type_all_derivations_known_p): New predicate.
2112 (type_all_ctors_visible_p): New predicate.
2113 (type_possibly_instantiated_p): New predicate.
2114 (get_odr_type): Compute all_derivations_known.
2115 (dump_odr_type): Dump the flag.
2116 (maybe_record_type): Cleanup.
2117 (record_target_from_binfo): Add bases_to_consider array;
2118 record bases for types w/o instances and skip CXX destructor.
2119 (possible_polymorphic_call_targets_1): Add bases_to_consider
2120 and consider_construction parameters; check if type may have instance.
2121 (get_polymorphic_call_info): Set maybe_in_construction to true
2122 when we know nothing.
2123 (record_targets_from_bases): Skip CXX destructors; they are
2124 never called for types in construction.
2125 (possible_polymorphic_call_targets): Do not record target when
2126 type may not have instance.
2128 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2131 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2132 external aliases alive, too.
2134 2014-04-16 Andrew Pinski <apinski@cavium.com>
2136 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2139 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2141 * final.c (compute_alignments): Do not apply loop alignment to a block
2142 falling through to the exit.
2144 2014-04-16 Catherine Moore <clm@codesourcery.com>
2146 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2147 Adjust constraints for microMIPS store patterns.
2149 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2151 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2153 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2155 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2156 (append_use): Run at -O0.
2157 (append_vdef): Likewise.
2158 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2159 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2161 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2163 PR tree-optimization/60844
2164 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2165 (propagate_op_to_single_use, remove_visited_stmt_chain,
2166 linearize_expr, repropagate_negates, reassociate_bb): Use it
2167 instead of gsi_remove.
2169 2014-04-16 Martin Jambor <mjambor@suse.cz>
2171 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2172 ipa_transforms_to_apply.
2173 (cgraph_function_versioning): Assert that old_node has empty
2174 ipa_transforms_to_apply.
2175 * trans-mem.c (ipa_tm_create_version): Likewise.
2176 * tree-inline.c (tree_function_versioning): Do not duplicate
2177 ipa_transforms_to_apply.
2179 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2182 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2184 Pass necessary as flags on 64-bit Solaris/x86.
2185 Use lowercase relocs for x86_64-*-*.
2186 * configure: Regenerate.
2188 2014-04-15 Jan Hubicka <jh@suse.cz>
2190 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2191 (maybe_record_node, likely_target_p): Use it.
2193 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2196 Revert following patch
2198 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2201 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2202 software floating point or no floating point registers, do not
2203 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2204 in GPRs that occurs after we tested for GPRs that would never be
2207 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2208 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2209 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2210 specifically allow DDmode, since that does not use the SPE SIMD
2213 2014-03-21 Mark Wielaard <mjw@redhat.com>
2215 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2216 as unsigned or int depending on type and value used.
2218 2014-04-15 Richard Biener <rguenther@suse.de>
2220 PR rtl-optimization/56965
2221 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2222 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2224 * alias.c (true_dependence_1): Do not call
2225 nonoverlapping_component_refs_p.
2226 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2227 nonoverlapping_component_refs_p.
2228 (indirect_refs_may_alias_p): Likewise.
2230 2014-04-15 Teresa Johnson <tejohnson@google.com>
2232 * cfg.c (dump_bb_info): Fix flags check.
2233 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2235 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2237 PR rtl-optimization/60663
2238 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2241 2014-04-15 Richard Biener <rguenther@suse.de>
2243 * lto-streamer.h (LTO_major_version): Bump to 4.
2245 2014-04-15 Richard Biener <rguenther@suse.de>
2247 * common.opt (lto_partition_model): New enum.
2248 (flto-partition=): Merge separate options with a single with argument,
2249 add -flto-partition=one support.
2250 * flag-types.h (enum lto_partition_model): Declare.
2251 * opts.c (finish_options): Remove duplicate -flto-partition=
2253 * lto-wrapper.c (run_gcc): Adjust.
2255 2014-04-15 Richard Biener <rguenther@suse.de>
2257 * alias.c (ncr_compar): New function.
2258 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2260 2014-04-15 Richard Biener <rguenther@suse.de>
2262 * alias.c (record_component_aliases): Do not walk BINFOs.
2264 2014-04-15 Richard Biener <rguenther@suse.de>
2266 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2267 Add struct function argument and adjust.
2268 (find_func_aliases_for_call): Likewise.
2269 (find_func_aliases): Likewise.
2270 (find_func_clobbers): Likewise.
2271 (intra_create_variable_infos): Likewise.
2272 (compute_points_to_sets): Likewise.
2273 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2275 2014-04-15 Richard Biener <rguenther@suse.de>
2277 * tree.c (iterative_hash_expr): Use enum tree_code_class
2278 to store TREE_CODE_CLASS.
2279 (tree_block): Likewise.
2280 (tree_set_block): Likewise.
2281 * tree.h (fold_build_pointer_plus_loc): Use
2282 convert_to_ptrofftype_loc.
2284 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2287 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2290 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2292 * cfgloop.h (struct loop): Move force_vectorize down.
2293 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2294 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2295 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2296 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2297 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2298 * tree-core.h (enum annot_expr_kind): Add new kind values.
2299 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2300 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2302 * tree.def (ANNOTATE_EXPR): Tweak comment.
2304 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2306 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2309 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2311 * tree.h (TYPE_IDENTIFIER): Declare.
2312 * tree.c (subrange_type_for_debug_p): Use it.
2313 * godump.c (go_format_type): Likewise.
2314 * dwarf2out.c (is_cxx_auto, modified_type_die,
2315 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2316 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2318 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2321 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2323 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2325 * config/i386/i386.c (examine_argument): Return bool. Return true if
2326 parameter should be passed in memory.
2327 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2328 (construct_container): Update calls to examine_argument.
2329 (function_arg_advance_64): Ditto.
2330 (return_in_memory_32): Merge with ix86_return_in_memory.
2331 (return_in_memory_64): Ditto.
2332 (return_in_memory_ms_64): Ditto.
2334 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2336 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2337 * coverage.c (coverage_compute_profile_id): Handle externally visible
2340 2014-04-14 Martin Jambor <mjambor@suse.cz>
2342 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2343 DECL_DISREGARD_INLINE_LIMITS functions.
2345 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2348 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2350 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2353 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2354 optimize_insn_for_speed_p instead of
2355 optimize_function_for_speed_p.
2357 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2359 * doc/invoke.texi (free): Document AArch64.
2361 2014-04-14 Richard Biener <rguenther@suse.de>
2363 PR tree-optimization/60042
2364 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2365 (insert_into_preds_of_block): Do not prevent PHI insertion
2366 for REFERENCE exprs here ...
2367 (eliminate_dom_walker::before_dom_children): ... but prevent
2368 their use here under similar conditions when applied to the
2369 IL after PRE optimizations.
2371 2014-04-14 Richard Biener <rguenther@suse.de>
2373 * passes.def: Move early points-to after early SRA.
2375 2014-04-14 Richard Biener <rguenther@suse.de>
2377 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2378 check for which sign-changes we allow when forwarding
2379 a converted value into a switch.
2381 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2383 * stor-layout.c (place_field): Finalize non-constant offset for the
2386 2014-04-14 Richard Biener <rguenther@suse.de>
2388 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2390 (expand_switch_using_bit_tests_p): Likewise.
2391 (process_switch): Compute and pass on speed_p based on the
2393 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2394 optimize_bb_for_speed_p.
2396 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2398 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2399 * function.h (struct function): Rename has_force_vect_loops into
2400 has_force_vectorize_loops.
2401 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2402 (input_struct_function_base): Likewise.
2403 * lto-streamer-out.c (output_cfg): Likewise.
2404 (output_struct_function_base): Likewise.
2405 * omp-low.c (expand_omp_simd): Likewise.
2406 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2407 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2408 (version_loop_for_if_conversion): Likewise.
2409 (tree_if_conversion): Likewise.
2410 (main_tree_if_conversion): Likewise.
2411 (gate_tree_if_conversion): Likewise.
2412 * tree-inline.c (copy_loops): Likewise.
2413 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2414 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2415 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2416 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2417 * tree-vectorizer.c (vectorize_loops): Likewise.
2418 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2420 2014-04-14 Richard Biener <rguenther@suse.de>
2423 * lto-streamer-out.c (wrap_refs): New function.
2424 (lto_output): Wrap symbol references in global initializes in
2425 type-preserving MEM_REFs.
2427 2014-04-14 Christian Bruel <christian.bruel@st.com>
2429 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2431 2014-04-14 Christian Bruel <christian.bruel@st.com>
2433 * config/sh/sh.md (setmemqi): New expand pattern.
2434 * config/sh/sh.h (CLEAR_RATIO): Define.
2435 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2436 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2438 2014-04-14 Richard Biener <rguenther@suse.de>
2441 * fold-const.c (negate_expr_p): Don't negate directional rounding
2443 (fold_negate_expr): Likewise.
2445 2014-04-14 Richard Biener <rguenther@suse.de>
2447 PR tree-optimization/59817
2448 PR tree-optimization/60453
2449 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2450 recursion to catch all CHRECs in the scalar evolution and restrict
2451 the predicate for the remains appropriately.
2453 2014-04-12 Catherine Moore <clm@codesourcery.com>
2455 * config/mips/constraints.md: Add new register constraint "kb".
2456 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2457 (*movhi_internal): Likewise.
2458 (*movqi_internal): Likewise.
2459 * config/mips/mips.h (M16_STORE_REGS): New register class.
2460 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2461 (REG_CLASS_CONTENTS): Likewise.
2462 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2464 2014-04-11 Tobias Burnus <burnus@net-b.de>
2467 * doc/invoke.texi (-Wformat-signedness): Document it.
2468 (Wformat=2): Mention that this enables -Wformat-signedness.
2470 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2472 * common/config/epiphany/epiphany-common.c
2473 (epiphany_option_optimization_table): Enable section anchors by
2474 default at -O1 or higher.
2475 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2476 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2477 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2478 carries no extra cost.
2479 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2480 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2481 * config/epiphany/predicates.md (memclob_operand): New predicate.
2482 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2483 Use memclob_operand predicate and X constraint for operand 3.
2485 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2487 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2488 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2491 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2493 PR rtl-optimization/60651
2494 * mode-switching.c (optimize_mode_switching): Make sure to emit
2495 sets of a lower numbered entity before sets of a higher numbered
2496 entity to a mode of the same or lower priority.
2497 When creating a seginfo for a basic block that starts with a code
2498 label, move the insertion point past the code label.
2499 (new_seginfo): Document and enforce requirement that
2500 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2501 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2502 * doc/tm.texi: Regenerate.
2504 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2507 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2509 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2511 * BASE-VER: Set to 4.10.0.
2513 2014-04-11 Tobias Burnus <burnus@net-b.de>
2516 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2517 * doc/gcc.texi (Service): Update description in the @menu
2518 * doc/invoke.texi (Option Summary): Remove misplaced and
2521 2014-04-11 Steve Ellcey <sellcey@mips.com>
2522 Jakub Jelinek <jakub@redhat.com>
2525 * expr.c (convert_move): Use emit_store_flag_force instead of
2526 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2529 2014-04-11 Richard Biener <rguenther@suse.de>
2532 * varasm.c (assemble_alias): Avoid endless error reporting
2533 recursion by setting TREE_ASM_WRITTEN.
2535 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2537 * config/s390/s390.md: Add a splitter for NOT rtx.
2539 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2541 PR rtl-optimization/60663
2542 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2544 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2545 Jakub Jelinek <jakub@redhat.com>
2548 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2549 flag from decl_node to node.
2551 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2554 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2555 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2556 ameliorating the cases where it can be.
2558 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2561 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2563 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2564 (loadsync_<mode>): Change mode.
2565 (load_quadpti, store_quadpti): New.
2566 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2567 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2568 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2570 2014-04-09 Cong Hou <congh@google.com>
2573 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2576 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2578 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2579 instead of vnor to exploit possible fusion opportunity in the
2581 (altivec_expand_vec_perm_const_le): Likewise.
2583 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2585 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2586 (loadsync_<mode>): Change mode.
2587 (load_quadpti, store_quadpti): New.
2588 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2589 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2591 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2594 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2595 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2596 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2598 2014-04-08 Richard Biener <rguenther@suse.de>
2601 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2602 a 64bit widest int print double-int similar to on HWI64 hosts.
2604 2014-04-08 Richard Biener <rguenther@suse.de>
2606 PR tree-optimization/60785
2607 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2608 default defs properly.
2610 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2612 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2613 (Weffc++): Likewise.
2615 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2617 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2618 set completep to false rather than true.
2620 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2623 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2624 ARM_TARGET2_DWARF_FORMAT.
2626 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2629 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2630 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2633 2014-04-07 Richard Biener <rguenther@suse.de>
2635 PR tree-optimization/60766
2636 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2637 (may_eliminate_iv): Convert cand_value_at result to desired type.
2639 2014-04-07 Jason Merrill <jason@redhat.com>
2642 * common.opt (-fno-gnu-unique): Add.
2643 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2645 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2647 * haifa-sched.c: Fix outdated function reference and minor
2648 grammar errors in introductory comment.
2650 2014-04-07 Richard Biener <rguenther@suse.de>
2653 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2655 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2657 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2660 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2661 size accounting for thunks.
2662 (pa_asm_output_mi_thunk): Use final_start_function() and
2663 final_end_function() to output function start and end directives.
2665 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2667 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2668 device specific ISA/ feature information. Remove short_sp and
2669 errata_skip ds. Add avr_device_specific_features enum to have device
2671 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2672 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2673 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2674 updated device specific info.
2675 * config/avr/avr-mcus.def: Merge device specific details to
2676 dev_attribute field.
2677 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2679 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2680 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2681 assembler if RMW isa supported by current device.
2682 * config/avr/genmultilib.awk: Update as device info structure changed.
2683 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2685 2014-04-04 Cong Hou <congh@google.com>
2687 PR tree-optimization/60656
2688 * tree-vect-stmts.c (supportable_widening_operation):
2689 Fix a bug that elements in a vector with vect_used_by_reduction
2690 property are incorrectly reordered when the operation on it is not
2691 consistant with the one in reduction operation.
2693 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2695 PR rtl-optimization/60155
2696 * gcse.c (record_set_data): New function.
2697 (single_set_gcse): New function.
2698 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2699 (hoist_code): Likewise.
2700 (get_pressure_class_and_nregs): Likewise.
2702 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2704 * explow.c (probe_stack_range): Emit a final optimization blockage.
2706 2014-04-04 Anthony Green <green@moxielogic.com>
2708 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2711 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2714 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2715 flags are set only during streaming.
2716 * ipa.c (process_references, walk_polymorphic_call_targets,
2717 symtab_remove_unreachable_nodes): Drop bodies of always inline
2718 after early inlining.
2719 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2721 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2722 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2725 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2728 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2731 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2733 (cortex_a53_fdivd): Likewise.
2735 2014-04-04 Martin Jambor <mjambor@suse.cz>
2738 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2740 * cgraph.c (clone_of_p): Also return true if thunks match.
2741 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2742 cgraph_function_or_thunk_node and an obsolete comment.
2743 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2745 (build_function_decl_skip_args): Likewise.
2746 (set_new_clone_decl_and_node_flags): New function.
2747 (duplicate_thunk_for_node): Likewise.
2748 (redirect_edge_duplicating_thunks): Likewise.
2749 (cgraph_clone_node): New parameter args_to_skip, pass it to
2750 redirect_edge_duplicating_thunks which is called instead of
2751 cgraph_redirect_edge_callee.
2752 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2753 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2755 2014-04-04 Jeff Law <law@redhat.com>
2758 * config/arm/predicates.md (const_int_I_operand): New predicate.
2759 (const_int_M_operand): Similarly.
2760 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2762 (insv_t2, extv_reg, extzv_t2): Likewise.
2763 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2764 (pop_multiple_with_writeback_and_return): Likewise.
2765 (vfp_pop_multiple_with_writeback): Likewise
2767 2014-04-04 Richard Biener <rguenther@suse.de>
2770 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2771 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2773 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2774 * gimplify.c (gimple_add_tmp_var_fn): New function.
2775 * gimple-expr.h (create_tmp_reg_fn): Declare.
2776 * gimple-expr.c (create_tmp_reg_fn): New function.
2777 * gimple-low.c (record_vars_into): Don't change cfun.
2778 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2779 code generation without cfun.
2781 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2784 * Makefile.in (install-driver): Fix shell scripting.
2786 2014-04-03 Cong Hou <congh@google.com>
2788 PR tree-optimization/60505
2789 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2790 threshold of number of iterations below which no vectorization
2792 * tree-vect-loop.c (new_loop_vec_info):
2793 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2794 * tree-vect-loop.c (vect_analyze_loop_operations):
2795 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2796 * tree-vect-loop.c (vect_transform_loop):
2797 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2798 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2799 of iterations of the loop and see if we should build the epilogue.
2801 2014-04-03 Richard Biener <rguenther@suse.de>
2803 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2804 (streamer_tree_cache_create): Adjust.
2805 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2806 to allow optional nodes array.
2807 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2808 (streamer_tree_cache_append): Likewise.
2809 (streamer_tree_cache_create): Create nodes array optionally
2810 as specified by parameter.
2811 * lto-streamer-out.c (create_output_block): Avoid maintaining
2812 the node array in the writer cache.
2813 (DFS_write_tree): Remove assertion.
2814 (produce_asm_for_decls): Free the out decl state hash table early.
2815 * lto-streamer-in.c (lto_data_in_create): Adjust for
2816 streamer_tree_cache_create prototype change.
2818 2014-04-03 Richard Biener <rguenther@suse.de>
2820 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2821 set TREE_CHAIN to NULL_TREE.
2823 2014-04-03 Richard Biener <rguenther@suse.de>
2825 PR tree-optimization/60740
2826 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2827 over all GIMPLE_COND operands.
2829 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2831 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2832 (Weffc++): Remove Scott's numbering, merge lists and reference
2835 2014-04-03 Nick Clifton <nickc@redhat.com>
2837 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2840 2014-04-03 Martin Jambor <mjambor@suse.cz>
2842 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2843 mention gcc_unreachable before failing.
2844 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2847 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2850 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2851 inconsistent code and instead mark the context inconsistent.
2852 (possible_polymorphic_call_targets): For inconsistent contexts
2853 return empty complete list.
2855 2014-04-02 Anthony Green <green@moxielogic.com>
2857 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2858 (extendqisi2, extendhisi2): Define.
2859 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2860 (WCHAR_TYPE): Change to unsigned int.
2862 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2864 PR tree-optimization/60733
2865 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2866 insertion point for PHI candidates to be the end of the feeding
2867 block for the PHI argument.
2869 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2871 PR rtl-optimization/60650
2872 * lra-constraints.c (process_alt_operands): Decrease reject for
2873 earlyclobber matching.
2875 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2877 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2879 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2881 * config/spu/spu.c (pad_bb): Do not crash when the last
2882 insn is CODE_FOR_blockage.
2884 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2886 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2887 lies outside the target mode.
2889 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2892 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2893 software floating point or no floating point registers, do not
2894 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2895 in GPRs that occurs after we tested for GPRs that would never be
2898 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2899 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2900 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2901 specifically allow DDmode, since that does not use the SPE SIMD
2904 2014-04-02 Richard Biener <rguenther@suse.de>
2907 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2908 MODE_INTs. Properly use negv_optab.
2909 (expand_abs): Likewise.
2911 2014-04-02 Richard Biener <rguenther@suse.de>
2914 * Makefile.in (install-driver): Guard extra installs with special
2917 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2919 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2922 2014-04-01 Richard Henderson <rth@redhat.com>
2925 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2926 alternative enabled before register allocation.
2928 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2930 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2931 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2933 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2934 (nios2_got_address): Update nios2_large_got_address call site.
2935 (nios2_delegitimize_address): New function.
2936 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2937 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2938 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2940 2014-04-01 Martin Husemann <martin@duskware.de>
2942 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2945 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2947 PR rtl-optimization/60604
2948 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2949 check from register_operand.
2950 (register_operand): Redefine in terms of general_operand.
2951 (nonmemory_operand): Use register_operand for the non-constant cases.
2953 2014-04-01 Richard Biener <rguenther@suse.de>
2955 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2957 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2959 * doc/invoke.texi (mapp-regs): Clarify.
2961 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2963 * config/i386/avx512fintrin.h (__v32hi): Define type.
2964 (__v64qi): Likewise.
2965 (_mm512_set1_epi8): Define.
2966 (_mm512_set1_epi16): Define.
2967 (_mm512_set4_epi32): Define.
2968 (_mm512_set4_epi64): Define.
2969 (_mm512_set4_pd): Define.
2970 (_mm512_set4_ps): Define.
2971 (_mm512_setr4_epi64): Define.
2972 (_mm512_setr4_epi32): Define.
2973 (_mm512_setr4_pd): Define.
2974 (_mm512_setr4_ps): Define.
2975 (_mm512_setzero_epi32): Define.
2977 2014-03-31 Martin Jambor <mjambor@suse.cz>
2980 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
2981 callsite_arguments_match_p. Updated all callers. Also check types of
2982 corresponding formal parameters and actual arguments.
2983 (not_all_callers_have_enough_arguments_p) Renamed to
2984 some_callers_have_mismatched_arguments_p.
2986 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
2988 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
2990 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
2993 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
2996 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
2998 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
2999 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3001 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3002 Use FMAMODE_NOVF512 mode iterator.
3003 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3004 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3005 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3006 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3008 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3009 Use VF_128_256 mode iterator.
3010 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3013 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3015 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3016 static chain if needed.
3018 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3021 * lra-constraints.c (index_part_to_reg): New.
3022 (process_address): Use it.
3024 2014-03-27 Jeff Law <law@redhat.com>
3025 Jakub Jelinek <jakub@redhat.com>
3028 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3029 gen_rtx_{PLUS,MULT} to build up the address expression.
3031 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3032 creating non-canonical RTL.
3034 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3037 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3038 functions; reorganize to make cheap checks first.
3039 (inline_small_functions): Do not estimate growth when dumping;
3041 * ipa-inline.h (inline_summary): Add min_size.
3042 (growth_likely_positive): New function.
3043 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3044 (set_cond_stmt_execution_predicate): Cleanup.
3045 (estimate_edge_size_and_time): Compute min_size.
3046 (estimate_calls_size_and_time): Likewise.
3047 (estimate_node_size_and_time): Likewise.
3048 (inline_update_overall_summary): Update min_size.
3049 (do_estimate_edge_time): Likewise.
3050 (do_estimate_edge_size): Update.
3051 (do_estimate_edge_hints): Update.
3052 (growth_likely_positive): New function.
3054 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3057 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3058 also if addr has VOIDmode.
3060 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3062 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3063 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3065 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3066 instructions as well as AdvancedSIMD loads.
3068 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3070 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3071 Use crypto_aese type.
3072 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3073 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3074 crypto_aese, crypto_aesmc. Move to types.md.
3075 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3077 * config/arm/iterators.md (crypto_type): Likewise.
3079 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3081 * cgraph.c: Include expr.h and tree-dfa.h.
3082 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3085 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3088 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3089 regs from checking multi-reg pseudos.
3091 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3093 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3095 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3097 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3098 if it would clobber the stack pointer, even temporarily.
3100 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3102 * mode-switching.c: Make small adjustments to the top comment.
3104 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3106 * config/rs6000/constraints.md (wD constraint): New constraint to
3107 match the constant integer to get the top DImode/DFmode out of a
3108 vector in a VSX register.
3110 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3111 match the constant integer to get the top DImode/DFmode out of a
3112 vector in a VSX register.
3114 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3117 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3120 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3121 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3123 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3124 Optimize vec_extract of 64-bit values, where the value being
3125 extracted is in the top word, where we can use scalar
3126 instructions. Add direct move and store support. Combine the big
3127 endian/little endian vector select load support into a single insn.
3128 (vsx_extract_<mode>_internal1): Likewise.
3129 (vsx_extract_<mode>_internal2): Likewise.
3130 (vsx_extract_<mode>_load): Likewise.
3131 (vsx_extract_<mode>_store): Likewise.
3132 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3133 combined into vsx_extract_<mode>_load.
3134 (vsx_extract_<mode>_one_le): Likewise.
3136 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3137 define the top 64-bit vector element.
3139 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3142 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3143 Document vec_vbpermq builtin.
3146 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3147 enable use of xxsldwi and xxpermdi builtin functions.
3148 (vec_xxpermdi): Likewise.
3150 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3151 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3153 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3155 PR rtl-optimization/60650
3156 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3158 (find_spills_for): New.
3159 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3160 Spill all pseudos on the second iteration.
3162 2014-03-27 Marek Polacek <polacek@redhat.com>
3165 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3168 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3170 * config/s390/s390.c (s390_can_use_return_insn): Check for
3171 call-saved FPRs on 31 bit.
3173 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3176 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3177 if they need regimplification, just drop them instead of
3178 calling gimple_regimplify_operands on them.
3180 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3183 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3184 (aarch64_frame_pointer_required): Adjust logic.
3185 (aarch64_can_eliminate): Adjust logic.
3186 (aarch64_override_options_after_change): Adjust logic.
3188 2014-03-27 Dehao Chen <dehao@google.com>
3190 * ipa-inline.c (early_inliner): Update node's inline info.
3192 2014-03-26 Dehao Chen <dehao@google.com>
3194 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3195 compiler inserted conditional jumps for NAN float check.
3197 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3199 * ubsan.h (ubsan_create_data): Change second argument's type
3200 to const location_t *.
3201 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3203 (ubsan_create_data): Change second argument to const location_t *PLOC.
3204 Create Loc field whenever PLOC is non-NULL.
3205 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3206 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3210 * real.c (real_to_integer2): Change type of low to UHWI.
3212 2014-03-26 Tobias Burnus <burnus@net-b.de>
3214 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3215 (CILK_SELF_SPECS): New define.
3216 (driver_self_specs): Use it.
3218 2014-03-26 Richard Biener <rguenther@suse.de>
3220 * tree-pretty-print.c (percent_K_format): Implement special
3221 case for LTO and its stripped down BLOCK tree.
3223 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3226 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3228 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3229 one range is range_int_cst_p, but not both, at least optimize
3230 addition/subtraction of 0 and multiplication by 0 or 1.
3231 * gimple-fold.c (gimple_fold_call): Fold
3232 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3233 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3234 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3236 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3238 PR rtl-optimization/60452
3239 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3240 <case REG>: Return 1 for invalid offsets from the frame pointer.
3242 2014-03-26 Marek Polacek <polacek@redhat.com>
3245 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3248 2014-03-26 Marek Polacek <polacek@redhat.com>
3251 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3254 2014-03-26 Marek Polacek <polacek@redhat.com>
3257 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3258 multiplication in unsigned type.
3260 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3262 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3264 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3266 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3268 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3271 * cif-code.def (UNREACHABLE) New code.
3272 * ipa-inline.c (inline_small_functions): Skip edges to
3273 __builtlin_unreachable.
3274 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3275 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3276 predicate to __bulitin_unreachable.
3277 (set_cond_stmt_execution_predicate): Fix issue when
3278 invert_tree_comparison returns ERROR_MARK.
3279 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3280 propagate to inline clones.
3281 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3283 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3284 * cgraphclones.c (cgraph_clone_node): If call destination is already
3285 ureachable, do not redirect it back.
3286 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3289 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3291 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3292 Do not modify inline clones.
3294 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3296 * config/i386/i386.md (general_sext_operand): New mode attr.
3297 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3298 don't generate (sign_extend (const_int)).
3299 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3300 operands[2]. Use We constraint instead of <i> and
3301 <general_sext_operand> predicate instead of <general_operand>.
3302 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3303 * config/i386/constraints.md (We): New constraint.
3304 * config/i386/predicates.md (x86_64_sext_operand,
3305 sext_operand): New predicates.
3307 2014-03-25 Martin Jambor <mjambor@suse.cz>
3310 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3311 inconsistent devirtualizations to __builtin_unreachable.
3313 2014-03-25 Marek Polacek <polacek@redhat.com>
3316 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3318 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3320 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3321 order of elements for big-endian.
3323 2014-03-25 Richard Biener <rguenther@suse.de>
3326 * gimplify-me.c (gimple_regimplify_operands): Update the
3329 2014-03-25 Martin Jambor <mjambor@suse.cz>
3332 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3333 (lto_output_varpool_node): Likewise.
3334 (input_overwrite_node): Likewise.
3335 (input_varpool_node): Likewise.
3337 2014-03-25 Richard Biener <rguenther@suse.de>
3339 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3340 (run_gcc): Likewise.
3342 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3344 * combine.c (simplify_compare_const): Add MODE argument.
3345 Handle mode_width 0 as very large mode_width.
3346 (try_combine, simplify_comparison): Adjust callers.
3348 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3349 type to avoid signed integer overflow.
3350 * explow.c (plus_constant): Likewise.
3352 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3354 * doc/generic.texi: Correct typos.
3356 2014-03-24 Tobias Burnus <burnus@net-b.de>
3358 * doc/invoke.texi (-flto): Expand section about
3359 using static libraries with LTO.
3361 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3363 PR rtl-optimization/60501
3364 * optabs.def (addptr3_optab): New optab.
3365 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3366 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3367 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3369 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3371 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3373 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3375 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3378 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3379 (_mm256_undefined_ps): Define.
3380 (_mm256_undefined_pd): Define.
3381 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3382 (_mm_undefined_pd): Define.
3383 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3384 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3385 (_mm512_undefined_ps): Define.
3386 (_mm512_undefined_pd): Define.
3387 Use _mm*_undefined_*.
3388 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3390 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3392 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3393 (lshr_simd): DI mode added.
3394 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3395 (aarch64_ushr_simddi): Likewise.
3396 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3397 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3398 (vshrd_n_u64): Likewise.
3400 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3402 * Makefile.in (s-macro_list): Depend on cc1.
3404 2014-03-23 Teresa Johnson <tejohnson@google.com>
3406 * ipa-utils.c (ipa_print_order): Use specified dump file.
3408 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3410 PR rtl-optimization/60601
3411 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3413 * gcc.c (eval_spec_function): Initialize save_growing_value.
3415 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3418 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3419 code == MINUS_EXPR, never swap op0 with op1.
3421 * toplev.c (init_local_tick): Avoid signed integer multiplication
3423 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3424 shift by first operand's bitsize.
3426 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3429 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3431 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3432 TARGET_ISA_64BIT_P(x).
3434 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3436 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3437 pattern for vector nor instead of subtract from splat(-1).
3438 (altivec_expand_vec_perm_const_le): Likewise.
3440 2014-03-21 Richard Henderson <rth@twiddle.net>
3443 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3444 related insns after epilogue_completed.
3446 2014-03-21 Martin Jambor <mjambor@suse.cz>
3449 * cgraph.h (symtab_node): New flag body_removed.
3450 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3451 when removing bodies.
3452 * symtab.c (dump_symtab_base): Dump body_removed flag.
3453 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3454 had their bodies removed.
3456 2014-03-21 Martin Jambor <mjambor@suse.cz>
3459 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3462 2014-03-21 Richard Biener <rguenther@suse.de>
3464 PR tree-optimization/60577
3465 * tree-core.h (struct tree_base): Document nothrow_flag use
3466 in VAR_DECL_NONALIASED.
3467 * tree.h (VAR_DECL_NONALIASED): New.
3468 (may_be_aliased): Adjust.
3469 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3471 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3473 * expr.c (expand_expr_real_1): Remove outdated comment.
3475 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3478 * ira.c (adjust_cleared_regs): Call copy_rtx on
3479 *reg_equiv[REGNO (loc)].src_p before passing it to
3480 simplify_replace_fn_rtx.
3483 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3484 into CONST, put pic register as first operand of PLUS. Use
3485 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3487 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3489 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3491 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3493 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3494 around for store forwarding issue in the FPU on the UT699.
3495 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3496 loads and operations if -mfix-ut699 is specified.
3497 (divtf3_hq): Tweak attribute.
3498 (sqrttf2_hq): Likewise.
3500 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3502 * calls.c (store_one_arg): Remove incorrect const qualification on the
3503 type of the temporary.
3504 * cfgexpand.c (expand_return): Likewise.
3505 * expr.c (expand_constructor): Likewise.
3506 (expand_expr_real_1): Likewise.
3508 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3510 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3513 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3516 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3518 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3520 * config/arm/aarch-common-protos.h
3521 (alu_cost_table): Fix spelling of "extend".
3522 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3524 2014-03-19 Richard Biener <rguenther@suse.de>
3527 * tree-core.h (tree_type_common): Re-order pointer members
3528 to reduce recursion depth during GC walks.
3530 2014-03-19 Marek Polacek <polacek@redhat.com>
3533 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3534 before accessing it.
3536 2014-03-19 Richard Biener <rguenther@suse.de>
3539 * lto-streamer-in.c (input_function): In WPA stage do not drop
3542 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3544 PR tree-optimization/60559
3545 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3546 with build_zero_cst assignment.
3548 2014-03-18 Kai Tietz <ktietz@redhat.com>
3550 PR rtl-optimization/56356
3551 * sdbout.c (sdbout_parms): Verify that parms'
3552 incoming argument is valid.
3553 (sdbout_reg_parms): Likewise.
3555 2014-03-18 Richard Henderson <rth@redhat.com>
3558 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3559 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3560 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3562 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3564 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3565 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3566 Italicize plugin event names in description. Explain that
3567 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3568 Remind that no GCC functions should be called after PLUGIN_FINISH.
3569 Explain what pragmas with expansion are.
3571 2014-03-18 Martin Liska <mliska@suse.cz>
3573 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3574 gimple call statement is update.
3575 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3576 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3578 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3581 * ubsan.c (ubsan_instrument_unreachable): Call
3582 initialize_sanitizer_builtins.
3583 (ubsan_pass): Likewise.
3586 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3587 varpool_finalize_decl instead of rest_of_decl_compilation.
3589 2014-03-18 Richard Biener <rguenther@suse.de>
3591 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3592 by using bitmap_and_compl instead of bitmap_and_compl_into.
3593 (df_rd_transfer_function): Likewise.
3595 2014-03-18 Richard Biener <rguenther@suse.de>
3597 * doc/lto.texi (fresolution): Fix typo.
3599 2014-03-18 Richard Biener <rguenther@suse.de>
3601 * doc/invoke.texi (flto): Update for changes in 4.9.
3603 2014-03-18 Richard Biener <rguenther@suse.de>
3605 * doc/loop.texi: Remove section on the removed lambda framework.
3606 Update loop docs with recent changes in preserving loop structure.
3608 2014-03-18 Richard Biener <rguenther@suse.de>
3610 * doc/lto.texi (-fresolution): Document.
3612 2014-03-18 Richard Biener <rguenther@suse.de>
3614 * doc/contrib.texi: Adjust my name.
3616 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3619 * internal-fn.c: Include diagnostic-core.h.
3620 (expand_BUILTIN_EXPECT): New function.
3621 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3622 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3623 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3624 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3626 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3627 Revert 3 argument __builtin_expect code.
3628 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3629 * gimple-fold.c (gimple_fold_call): Likewise.
3630 * tree.h (fold_builtin_expect): New prototype.
3631 * builtins.c (build_builtin_expect_predicate): Add predictor
3632 argument, if non-NULL, create 3 argument __builtin_expect.
3633 (fold_builtin_expect): No longer static. Add ARG2 argument,
3634 pass it through to build_builtin_expect_predicate.
3635 (fold_builtin_2): Adjust caller.
3636 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3637 * internal-fn.def (BUILTIN_EXPECT): New.
3639 2014-03-18 Tobias Burnus <burnus@net-b.de>
3642 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3643 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3644 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3646 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3649 * predict.c (combine_predictions_for_bb): Fix up formatting.
3650 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3651 fill what it points to if non-NULL.
3652 (tree_predict_by_opcode): Adjust caller, use the predictor.
3653 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3655 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3657 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3658 proper constant for the store mode.
3660 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3662 * symtab.c (change_decl_assembler_name): Fix transparent alias
3665 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3667 * config/aarch64/aarch64.c: Correct the comments about the
3668 aarch64 stack layout.
3670 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3672 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3673 check for GF_OMP_FOR_KIND_FOR.
3675 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3677 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3678 ymm and zmm register names.
3680 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3683 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3684 note creation for the 2010-08-31 changes.
3686 2014-03-17 Marek Polacek <polacek@redhat.com>
3689 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3690 as -fno-tree-loop-vectorize.
3691 (expand_omp_simd): Likewise.
3693 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3695 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3696 (eligible_for_call_delay): New prototype.
3697 * config/sparc/sparc.c (tls_call_delay): Rename into...
3698 (eligible_for_call_delay): ...this. Return false if the instruction
3699 cannot be put in the delay slot of a branch.
3700 (eligible_for_restore_insn): Simplify.
3701 (eligible_for_return_delay): Return false if the instruction cannot be
3702 put in the delay slot of a branch and simplify.
3703 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3704 put in the delay slot of a branch.
3705 * config/sparc/sparc.md (fix_ut699): New attribute.
3706 (tls_call_delay): Delete.
3707 (in_call_delay): Reimplement.
3708 (eligible_for_sibcall_delay): Rename into...
3709 (in_sibcall_delay): ...this.
3710 (eligible_for_return_delay): Rename into...
3711 (in_return_delay): ...this.
3712 (in_branch_delay): Reimplement.
3713 (in_uncond_branch_delay): Delete.
3714 (in_annul_branch_delay): Delete.
3716 2014-03-14 Richard Henderson <rth@redhat.com>
3719 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3720 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3721 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3722 (floathi splitters): Remove.
3723 (float<SWI48x>xf2): New pattern.
3724 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3725 code that tried to handle DImode for 32-bit, but which was excluded
3726 by the pattern's condition. Drop allocation of stack temporary.
3727 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3728 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3729 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3730 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3731 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3732 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3733 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3734 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3735 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3736 (*float<SWI48x><X87MODEF>2_i387): Remove.
3737 (all float _with_temp splitters): Remove.
3738 (*float<SWI48x><MODEF>2_i387): New pattern.
3739 (*float<SWI48><MODEF>2_sse): New pattern.
3740 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3741 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3743 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3744 Marek Polacek <polacek@redhat.com>
3747 * common.opt (dump_base_name_prefixed): New Variable.
3748 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3749 if x_dump_base_name_prefixed is already set, set it at the end.
3751 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3753 PR rtl-optimization/60508
3754 * lra-constraints.c (get_reload_reg): Add new parameter
3756 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3757 Pass the new parameter values.
3759 2014-03-14 Richard Biener <rguenther@suse.de>
3761 * common.opt: Revert unintented changes from r205065.
3764 2014-03-14 Richard Biener <rguenther@suse.de>
3767 * cfghooks.c (split_block): Properly adjust all loops the
3768 block was a latch of.
3770 2014-03-14 Martin Jambor <mjambor@suse.cz>
3773 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3776 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3779 * config/avr/avr.c (avr_set_current_function): Pass function name
3780 through default_strip_name_encoding before sanity checking instead
3781 of skipping the first char of the assembler name.
3783 2014-03-13 Richard Henderson <rth@redhat.com>
3786 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3787 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3788 * config/i386/i386-protos.h: Likewise.
3789 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3790 in the expander instead of a splitter.
3791 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3792 any possibility of requiring a memory.
3793 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3794 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3795 (fp branch splitters): Update for ix86_split_fp_branch.
3796 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3797 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3798 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3799 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3800 (*fop_<MODEF>_3_i387): Likewise.
3801 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3802 (splitters for the fop_* register patterns): Remove.
3803 (fscalexf4_i387): Rename from *fscalexf4_i387.
3804 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3806 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3808 PR tree-optimization/59779
3809 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3810 type for bitsize and maxsize instead of HOST_WIDE_INT.
3812 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3814 PR rtl-optimization/57320
3815 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3816 the CFG after thread_prologue_and_epilogue_insns.
3818 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3820 PR rtl-optimization/57189
3821 * lra-constraints.c (process_alt_operands): Disfavor spilling
3824 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3826 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3828 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3830 PR tree-optimization/59025
3832 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3833 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3835 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3838 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3839 calls of avr_out_plus_1.
3841 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3843 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3844 BB's single pred and update the father loop's latch info later.
3846 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3848 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3852 (VEC_base): Likewise.
3853 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3854 registers, we need to swap double words in little endian mode.
3856 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3857 to be a container mode for 128-bit integer operations added in ISA
3858 2.07. Unlike TImode and PTImode, the preferred register set is
3859 the Altivec/VMX registers for the 128-bit operations.
3861 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3863 (rs6000_split_128bit_ok_p): Likewise.
3865 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3866 macros for creating ISA 2.07 normal and overloaded builtin
3867 functions with 3 arguments.
3868 (BU_P8V_OVERLOAD_3): Likewise.
3869 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3870 for use as overloaded functions.
3871 (VPERM_1TI_UNS): Likewise.
3872 (VSEL_1TI): Likewise.
3873 (VSEL_1TI_UNS): Likewise.
3874 (ST_INTERNAL_1ti): Likewise.
3875 (LD_INTERNAL_1ti): Likewise.
3876 (XXSEL_1TI): Likewise.
3877 (XXSEL_1TI_UNS): Likewise.
3878 (VPERM_1TI): Likewise.
3879 (VPERM_1TI_UNS): Likewise.
3880 (XXPERMDI_1TI): Likewise.
3881 (SET_1TI): Likewise.
3882 (LXVD2X_V1TI): Likewise.
3883 (STXVD2X_V1TI): Likewise.
3884 (VEC_INIT_V1TI): Likewise.
3885 (VEC_SET_V1TI): Likewise.
3886 (VEC_EXT_V1TI): Likewise.
3887 (EQV_V1TI): Likewise.
3888 (NAND_V1TI): Likewise.
3889 (ORC_V1TI): Likewise.
3890 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3891 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3893 (VADDUQM): Likewise.
3894 (VSUBCUQ): Likewise.
3895 (VADDEUQM): Likewise.
3896 (VADDECUQ): Likewise.
3897 (VSUBEUQM): Likewise.
3898 (VSUBECUQ): Likewise.
3900 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3901 __int128_t and __uint128_t types.
3902 (__uint128_type): Likewise.
3903 (altivec_categorize_keyword): Add support for vector __int128_t,
3904 vector __uint128_t, vector __int128, and vector unsigned __int128
3905 as a container type for TImode operations that need to be done in
3906 VSX/Altivec registers.
3907 (rs6000_macro_to_expand): Likewise.
3908 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3909 to support 128-bit integer instructions vaddcuq, vadduqm,
3910 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3911 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3913 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3914 for V1TImode, and set up preferences to use VSX/Altivec registers.
3915 Setup VSX reload handlers.
3916 (rs6000_debug_reg_global): Likewise.
3917 (rs6000_init_hard_regno_mode_ok): Likewise.
3918 (rs6000_preferred_simd_mode): Likewise.
3919 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3920 (easy_altivec_constant): Likewise.
3921 (output_vec_const_move): Likewise.
3922 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3924 (rs6000_expand_vector_extract): Likewise.
3925 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3927 (rs6000_const_vec): Add support for V1TImode.
3928 (rs6000_emit_le_vsx_load): Swap double words when loading or
3929 storing TImode/V1TImode.
3930 (rs6000_emit_le_vsx_store): Likewise.
3931 (rs6000_emit_le_vsx_move): Likewise.
3932 (rs6000_emit_move): Add support for V1TImode.
3933 (altivec_expand_ld_builtin): Likewise.
3934 (altivec_expand_st_builtin): Likewise.
3935 (altivec_expand_vec_init_builtin): Likewise.
3936 (altivec_expand_builtin): Likewise.
3937 (rs6000_init_builtins): Add support for V1TImode type. Add
3938 support for ISA 2.07 128-bit integer builtins. Define type names
3939 for the VSX/Altivec vector types.
3940 (altivec_init_builtins): Add support for overloaded vector
3941 functions with V1TImode type.
3942 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3943 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3945 (rs6000_split_128bit_ok_p): Likewise.
3946 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3947 __int128_t and vector __uint128_t.
3949 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3950 and mode attributes.
3957 (VS_scalar): Likewise.
3958 (VS_double): Likewise.
3959 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3961 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3962 we support the ISA 2.07 128-bit integer arithmetic instructions.
3963 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3964 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3965 and TImode types for use with the builtin functions.
3966 (V1TI_type_node): Likewise.
3967 (unsigned_V1TI_type_node): Likewise.
3968 (intTI_type_internal_node): Likewise.
3969 (uintTI_type_internal_node): Likewise.
3971 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
3972 128-bit builtin functions.
3973 (UNSPEC_VADDEUQM): Likewise.
3974 (UNSPEC_VADDECUQ): Likewise.
3975 (UNSPEC_VSUBCUQ): Likewise.
3976 (UNSPEC_VSUBEUQM): Likewise.
3977 (UNSPEC_VSUBECUQ): Likewise.
3978 (VM): Add V1TImode to vector mode iterators.
3980 (VI_unit): Likewise.
3981 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
3982 (altivec_vaddcuq): Likewise.
3983 (altivec_vsubuqm): Likewise.
3984 (altivec_vsubcuq): Likewise.
3985 (altivec_vaddeuqm): Likewise.
3986 (altivec_vaddecuq): Likewise.
3987 (altivec_vsubeuqm): Likewise.
3988 (altivec_vsubecuq): Likewise.
3990 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
3992 (BOOL_128): Likewise.
3993 (BOOL_REGS_OUTPUT): Likewise.
3994 (BOOL_REGS_OP1): Likewise.
3995 (BOOL_REGS_OP2): Likewise.
3996 (BOOL_REGS_UNARY): Likewise.
3997 (BOOL_REGS_AND_CR0): Likewise.
3999 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4000 128-bit integer builtin support.
4001 (vec_vadduqm): Likewise.
4002 (vec_vaddecuq): Likewise.
4003 (vec_vaddeuqm): Likewise.
4004 (vec_vsubecuq): Likewise.
4005 (vec_vsubeuqm): Likewise.
4006 (vec_vsubcuq): Likewise.
4007 (vec_vsubuqm): Likewise.
4009 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4010 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4011 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4012 128-bit integer add/subtract to ISA 2.07.
4014 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4016 * config/arc/arc.c (arc_predicate_delay_insns):
4017 Fix third argument passed to conditionalize_nonjump.
4019 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4021 * config/aarch64/aarch64-builtins.c
4022 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4023 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4024 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4025 instead of __builtin_lfloor.
4026 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4028 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4030 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4031 (tree_ssa_ifcombine_bb_1): New function.
4032 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4033 is an empty forwarder block to then_bb or vice versa and then_bb
4034 and else_bb are effectively swapped.
4036 2014-03-12 Christian Bruel <christian.bruel@st.com>
4039 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4040 REG_CFA_DEF_CFA note.
4041 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4042 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4044 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4046 PR tree-optimization/60454
4047 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4049 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4051 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4052 Do not define target_cpu_default2 to generic.
4053 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4054 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4055 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4057 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4058 Marc Glisse <marc.glisse@inria.fr>
4060 PR tree-optimization/60502
4061 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4062 instead of build_low_bits_mask.
4064 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4067 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4068 if there are multiple uses, but op doesn't live on E edge.
4069 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4070 clobber stmts before __builtin_unreachable.
4072 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4074 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4075 hard_frame_pointer_rtx.
4076 * cse.c (cse_insn): Remove volatile check.
4077 * cselib.c (cselib_process_insn): Likewise.
4078 * dse.c (scan_insn): Likewise.
4080 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4082 * config/arc/arc.c (conditionalize_nonjump): New function,
4084 (arc_ifcvt): ... this.
4085 (arc_predicate_delay_insns): Use it.
4087 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4089 * config/arc/predicates.md (extend_operand): During/after reload,
4090 allow const_int_operand.
4091 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4092 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4093 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4095 (umulsi3_highpart_i): Likewise.
4097 2014-03-11 Richard Biener <rguenther@suse.de>
4099 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4100 Add asserts to guard possible wrong-code bugs.
4102 2014-03-11 Richard Biener <rguenther@suse.de>
4104 PR tree-optimization/60429
4105 PR tree-optimization/60485
4106 * tree-ssa-structalias.c (set_union_with_increment): Properly
4107 take into account all fields that overlap the shifted vars.
4108 (do_sd_constraint): Likewise.
4109 (do_ds_constraint): Likewise.
4110 (get_constraint_for_ptr_offset): Likewise.
4112 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4114 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4115 (nios2_compute_frame_layout):
4116 Add calculation of cfun->machine->fp_save_offset.
4117 (nios2_expand_prologue): Correct setting of frame pointer register
4119 (nios2_expand_epilogue): Update recovery of stack pointer from
4120 frame pointer accordingly.
4121 (nios2_initial_elimination_offset): Update calculation of offset
4122 for eliminating to HARD_FRAME_POINTER_REGNUM.
4124 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4127 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4128 cgraph_get_create_node on VAR_DECLs.
4130 2014-03-10 Richard Biener <rguenther@suse.de>
4133 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4135 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4137 * config/vms/vms.opt (vms_float_format): New variable.
4139 2014-03-08 Tobias Burnus <burnus@net-b.de>
4141 * doc/invoke.texi (-fcilkplus): Update implementation status.
4143 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4144 Richard Biener <rguenther@suse.de>
4146 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4147 consistently accross all TUs.
4148 (run_gcc): Enable -fshort-double automatically at link at link-time
4149 and disallow override.
4151 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4154 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4155 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4156 if they can't be used.
4158 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4160 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4161 for Solaris 11/x86 ld.
4162 * configure: Regenerate.
4164 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4166 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4167 (LIB_TLS_SPEC): Save as ld_tls_libs.
4168 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4169 (HAVE_AS_IX86_TLSLDM): New test.
4170 * configure, config.in: Regenerate.
4171 * config/i386/i386.c (legitimize_tls_address): Fall back to
4172 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4173 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4174 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4175 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4177 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4179 * common.opt (fira-loop-pressure): Mark as optimization.
4181 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4183 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4184 an OpenMP mappable type.
4186 2014-03-06 Matthias Klose <doko@ubuntu.com>
4188 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4189 MULTILIB_OSDIRNAMES is not defined.
4191 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4192 Meador Inge <meadori@codesourcery.com>
4195 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4196 (arm_legitimize_address): Call legitimize_tls_address for any
4197 arm_tls_referenced_p expression, handle constant addend. Call it
4198 before testing for !TARGET_ARM.
4199 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4201 2014-03-06 Richard Biener <rguenther@suse.de>
4207 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4209 * tree-streamer.c (record_common_node): Assert we don't record
4210 nodes with type double.
4211 (preload_common_node): Skip type double, complex double and double
4212 pointer since it is now frontend dependent due to fshort-double option.
4214 2014-03-06 Richard Biener <rguenther@suse.de>
4216 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4217 or -fno-lto is specified and the linker has full plugin support.
4218 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4219 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4220 * lto-wrapper.c (merge_and_complain): Merge compile-time
4221 optimization levels.
4222 (run_gcc): And pass it through to the link options.
4224 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4228 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4230 * cselib.c (remove_useless_values): Skip to avoid quadratic
4231 behavior if the condition moved from...
4232 (cselib_process_insn): ... here holds.
4234 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4237 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4238 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4241 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4242 (TM_H): Add x86-tune.def.
4244 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4246 * config/aarch64/aarch64.c (generic_tunings):
4247 Use cortexa57_extra_costs.
4249 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4252 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4253 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4254 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4257 2014-03-04 Heiher <r@hev.cc>
4259 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4260 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4262 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4264 * config/i386/predicates.md (const2356_operand): Change to ...
4265 (const2367_operand): ... this.
4266 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4268 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4269 (*avx512pf_scatterpf<mode>sf): Ditto.
4270 (avx512pf_scatterpf<mode>df): Ditto.
4271 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4272 (*avx512pf_scatterpf<mode>df): Ditto.
4273 * config/i386/i386.c (ix86_expand_builtin): Update
4274 incorrect hint operand error message.
4276 2014-03-04 Richard Biener <rguenther@suse.de>
4278 * lto-section-in.c (lto_get_section_data): Fix const cast.
4280 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4282 * tree-streamer.c (record_common_node): Assert we don't record
4283 nodes with type double.
4284 (preload_common_node): Skip type double, complex double and double
4285 pointer since it is now frontend dependent due to fshort-double option.
4287 2014-03-04 Richard Biener <rguenther@suse.de>
4290 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4291 (lto_input_toplevel_asms): Likewise.
4292 * lto-section-in.c (lto_get_section_data): Instead do it here
4295 2014-03-04 Richard Biener <rguenther@suse.de>
4297 PR tree-optimization/60382
4298 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4299 dead PHIs a reduction.
4301 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4303 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4305 (_mm_prefetch): Move out of GCC target("sse") pragma.
4306 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4307 GCC target("prfchw") pragma.
4308 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4310 * config/i386/i386.c (ix86_option_override_internal): Enable
4311 -mprfchw with -mprefetchwt1.
4313 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4315 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4318 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4320 * opts.h (CL_PCH_IGNORE): Define.
4321 * targhooks.c (option_affects_pch_p):
4322 Return false for options that have CL_PCH_IGNORE set.
4323 * opt-functions.awk: Process PchIgnore.
4324 * doc/options.texi: Document PchIgnore.
4326 * config/arc/arc.opt (misize): Add PchIgnore property.
4328 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4330 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4331 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4332 constraint on constants to permit them being loaded into
4333 GENERAL_REGS or BASE_REGS.
4335 2014-03-03 Nick Clifton <nickc@redhat.com>
4337 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4338 anti-cacnonical alternatives.
4339 (negandhi3_real): New pattern.
4340 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4342 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4344 * config/avr/avr-mcus.def: Remove atxmega16x1.
4345 * config/avr/avr-tables.opt: Regenerate.
4346 * config/avr/t-multilib: Regenerate.
4347 * doc/avr-mmcu.texi: Regenerate.
4349 2014-03-03 Tobias Grosser <tobias@grosser.es>
4350 Mircea Namolaru <mircea.namolaru@inria.fr>
4352 PR tree-optimization/58028
4353 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4356 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4358 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4359 not handled by recognizers.
4361 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4364 * function.c (expand_function_end): Don't emit
4365 clobber_return_register sequence if clobber_after is a BARRIER.
4366 * cfgexpand.c (construct_exit_block): Append instructions before
4367 return_label to prev_bb.
4369 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4371 * config/rs6000/constraints.md: Document reserved use of "wc".
4373 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4376 * ipa.c (function_and_variable_visibility): When dissolving comdat
4377 group, also set all symbols to local.
4379 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4384 2013-12-14 Jan Hubicka <jh@suse.cz>
4386 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4388 2014-03-02 Jon Beniston <jon@beniston.com>
4394 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4395 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4396 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4397 (simple_return, *simple_return): New patterns
4398 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4399 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4401 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4403 * dwarf2out.c (gen_subprogram_die): Tidy.
4405 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4408 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4409 (*mov_t_msb_neg_negc): ... this new insn.
4411 2014-02-28 Jason Merrill <jason@redhat.com>
4414 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4417 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4420 * dwarf2out.c (decltype_auto_die): New static.
4421 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4422 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4423 (is_cxx_auto): Likewise.
4425 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4427 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4428 we are not using general regs only.
4430 2014-02-28 Richard Biener <rguenther@suse.de>
4433 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4434 previous fix and only allow to remove trivial pre-headers
4435 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4436 (remove_forwarder_block): Properly update the latch of a loop.
4438 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4441 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4442 (cselib_preserved_hash_table): New.
4443 (preserve_constants_and_equivs): Move preserved vals to it.
4444 (cselib_find_slot): Look it up first.
4445 (cselib_init): Initialize it.
4446 (cselib_finish): Release it.
4447 (dump_cselib_table): Dump it.
4449 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4452 * cselib.c (remove_useless_values): Skip to avoid quadratic
4453 behavior if the condition moved from...
4454 (cselib_process_insn): ... here holds.
4456 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4459 * var-tracking.c (vt_initialize): Apply the same condition to
4460 preserve the CFA base value.
4462 2014-02-28 Joey Ye <joey.ye@arm.com>
4465 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4466 if reload in progress or completed.
4468 2014-02-28 Tobias Burnus <burnus@net-b.de>
4471 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4474 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4476 * doc/tm.texi.in (Condition Code Status): Update documention for
4477 relative locations of cc0-setter and cc0-user.
4479 2014-02-27 Jeff Law <law@redhat.com>
4481 PR rtl-optimization/52714
4482 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4483 into two independent simple sets, if I3 is a jump, ensure the
4484 pattern we place into I3 is a (set (pc) ...).
4486 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4487 Jeff Law <law@redhat.com>
4489 PR rtl-optimization/49847
4490 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4491 are in different blocks.
4492 * doc/tm.texi (Condition Code Status): Update documention for
4493 relative locations of cc0-setter and cc0-user.
4495 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4498 * lra.c (lra_emit_add): Check SUBREG too.
4500 2014-02-27 Andreas Schwab <schwab@suse.de>
4502 * config/m68k/m68k.c (m68k_option_override): Disable
4503 -flive-range-shrinkage for classic m68k.
4504 (m68k_override_options_after_change): Likewise.
4506 2014-02-27 Marek Polacek <polacek@redhat.com>
4509 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4510 -Wmaybe-uninitialized.
4512 2014-02-27 Alan Modra <amodra@gmail.com>
4515 * reload1.c (emit_input_reload_insns): When reload_override_in,
4516 set old to rl->in_reg when rl->in_reg is a subreg.
4518 2014-02-26 Richard Biener <rguenther@suse.de>
4521 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4523 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4525 * common/config/i386/predicates.md (const1256_operand): Remove.
4526 (const2356_operand): New.
4527 (const_1_to_2_operand): Remove.
4528 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4529 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4530 (*avx512pf_gatherpf<mode>sf): Ditto.
4531 (avx512pf_gatherpf<mode>df): Ditto.
4532 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4533 (*avx512pf_gatherpf<mode>df): Ditto.
4534 (avx512pf_scatterpf<mode>sf): Ditto.
4535 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4536 (*avx512pf_scatterpf<mode>sf): Ditto.
4537 (avx512pf_scatterpf<mode>df): Ditto.
4538 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4539 (*avx512pf_scatterpf<mode>df): Ditto.
4540 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4542 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4544 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4545 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4546 (_mm512_mask_testn_epi64_mask): Move to ...
4547 * config/i386/avx512cdintrin.h: Here.
4548 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4549 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4550 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4551 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4552 TARGET_AVX512F from TARGET_AVX512CD.
4554 2014-02-26 Richard Biener <rguenther@suse.de>
4557 * ipa.c (walk_polymorphic_call_targets): Properly guard
4558 call to inline_update_overall_summary.
4560 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4563 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4564 and latches only if requested. Fix latch if it is removed.
4565 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4566 LOOPS_HAVE_PREHEADERS.
4568 2014-02-25 Andrew Pinski <apinski@cavium.com>
4570 * builtins.c (expand_builtin_thread_pointer): Create a new target
4571 when the target is NULL.
4573 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4575 PR rtl-optimization/60317
4576 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4577 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4578 * lra-assigns.c: Include params.h.
4579 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4580 other reload pseudos considerations.
4582 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4584 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4585 to use canonical form for nor<mode>3.
4587 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4590 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4593 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4595 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4596 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4597 (ix86_handle_option): Handle OPT_mprefetchwt1.
4598 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4599 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4601 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4602 OPTION_MASK_ISA_PREFETCHWT1.
4603 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4604 (PTA_PREFETCHWT1): New.
4605 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4606 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4607 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4608 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4609 (*prefetch_avx512pf_<mode>_: Change into ...
4610 (*prefetch_prefetchwt1_<mode>: This.
4611 * config/i386/i386.opt (mprefetchwt1): New.
4612 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4613 (_mm_prefetch): Handle intent to write.
4614 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4616 2014-02-25 Richard Biener <rguenther@suse.de>
4619 * emit-rtl.c (mem_attrs_htab): Remove.
4620 (mem_attrs_htab_hash): Likewise.
4621 (mem_attrs_htab_eq): Likewise.
4622 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4623 (init_emit_once): Do not allocate mem_attrs_htab.
4625 2014-02-25 Richard Biener <rguenther@suse.de>
4628 * lto-opts.c (lto_write_options): Output non-explicit conservative
4629 -fwrapv, -fno-trapv and -fno-strict-overflow.
4630 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4631 (run_gcc): And pass them through.
4633 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4635 * sel-sched.c (calculate_new_fences): New parameter ptime.
4636 Calculate it as a maximum over all fence cycles.
4637 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4638 Print the final schedule timing when sched_verbose.
4640 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4642 PR rtl-optimization/60292
4643 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4644 bit fot the fence instruction.
4646 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4648 * calls.h: Fix typo in comment.
4650 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4652 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4653 adjusting offsetable addresses.
4655 2014-02-24 Guozhi Wei <carrot@google.com>
4657 * sparseset.h (sparseset_pop): Fix the wrong index.
4659 2014-02-24 Walter Lee <walt@tilera.com>
4661 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4662 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4664 * common/config/tilegx/tilegx-common.c
4665 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4666 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4668 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4669 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4670 (tilegx_gimplify_va_arg_expr): Handle big endian.
4671 (tilegx_expand_unaligned_load): Ditto.
4672 (tilegx_expand_unaligned_store): Ditto.
4673 (TARGET_RETURN_IN_MSB): New.
4674 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4675 (TARGET_ENDIAN_DEFAULT): New.
4676 (TARGET_BIG_ENDIAN): Handle big endian.
4677 (BYTES_BIG_ENDIAN): Ditto.
4678 (WORDS_BIG_ENDIAN): Ditto.
4679 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4682 * config/tilegx/tilegx.md (extv): Handle big endian.
4684 (insn_st<n>): Ditto.
4685 (insn_st<n>_add<bitsuffix>): Ditto.
4686 (insn_stnt<n>): Ditto.
4687 (insn_stnt<n>_add<bitsuffix>):Ditto.
4688 (vec_interleave_highv8qi): Handle big endian.
4689 (vec_interleave_highv8qi_be): New.
4690 (vec_interleave_highv8qi_le): New.
4691 (insn_v1int_h): Handle big endian.
4692 (vec_interleave_lowv8qi): Handle big endian.
4693 (vec_interleave_lowv8qi_be): New.
4694 (vec_interleave_lowv8qi_le): New.
4695 (insn_v1int_l): Handle big endian.
4696 (vec_interleave_highv4hi): Handle big endian.
4697 (vec_interleave_highv4hi_be): New.
4698 (vec_interleave_highv4hi_le): New.
4699 (insn_v2int_h): Handle big endian.
4700 (vec_interleave_lowv4hi): Handle big endian.
4701 (vec_interleave_lowv4hi_be): New.
4702 (vec_interleave_lowv4hi_le): New.
4703 (insn_v2int_l): Handle big endian.
4704 (vec_interleave_highv2si): Handle big endian.
4705 (vec_interleave_highv2si_be): New.
4706 (vec_interleave_highv2si_le): New.
4707 (insn_v4int_h): Handle big endian.
4708 (vec_interleave_lowv2si): Handle big endian.
4709 (vec_interleave_lowv2si_be): New.
4710 (vec_interleave_lowv2si_le): New.
4711 (insn_v4int_l): Handle big endian.
4712 * config/tilegx/tilegx.opt (mbig-endian): New option.
4713 (mlittle-endian): New option.
4714 * doc/install.texi: Document tilegxbe-linux.
4715 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4717 2014-02-24 Martin Jambor <mjambor@suse.cz>
4720 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4721 there are no parameter descriptors.
4723 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4725 PR rtl-optimization/60268
4726 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4727 initialization to ...
4728 (sched_rgn_init): ... here.
4729 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4731 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4733 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4736 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4738 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4741 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4743 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4744 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4746 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4748 * config/microblaze/predicates.md: Add cmp_op predicate.
4749 * config/microblaze/microblaze.md: Add branch_compare instruction
4750 which uses cmp_op predicate and emits cmp insn before branch.
4751 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4752 to microblaze_expand_conditional_branch and consolidate logic.
4753 (microblaze_expand_conditional_branch): emit branch_compare
4754 insn instead of handling cmp op separate from branch insn.
4756 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4758 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4761 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4763 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4764 define_insn with define_expand and new define_insn
4765 *altivec_lve<VI_char>x_internal.
4766 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4767 and new define_insn *altivec_stve<VI_char>x_internal.
4768 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4770 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4772 (altivec_expand_stvex_be): New function.
4774 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4776 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4777 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4778 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4779 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4781 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4784 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4785 instead of emit_move_insn.
4787 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4789 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4791 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4792 gen_altivec_vsumsws.
4794 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4796 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4797 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4798 (altivec_lvxl_<mode>): New define_expand incorporating
4799 -maltivec=be semantics where needed.
4800 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4801 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4802 semantics where needed.
4803 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4804 (altivec_stvx_<mode>): New define_expand incorporating
4805 -maltivec=be semantics where needed.
4806 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4807 VM2 iterator instead of V4SI.
4808 (altivec_stvxl_<mode>): New define_expand incorporating
4809 -maltivec=be semantics where needed.
4810 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4811 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4812 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4813 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4814 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4815 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4816 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4817 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4818 ALTIVEC_BUILTIN_STVXL.
4819 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4820 (altivec_expand_stvx_be): Likewise.
4821 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4822 (altivec_expand_lvx_be): Likewise.
4823 (altivec_expand_stvx_be): Likewise.
4824 (altivec_expand_builtin): Add cases for
4825 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4826 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4827 (altivec_init_builtins): Add definitions for
4828 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4829 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4831 2014-02-21 Catherine Moore <clm@codesourcery.com>
4833 * doc/invoke.texi (mvirt, mno-virt): Document.
4834 * config/mips/mips.opt (mvirt): New option.
4835 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4837 2014-02-21 Richard Biener <rguenther@suse.de>
4839 PR tree-optimization/60276
4840 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4841 (STMT_VINFO_MIN_NEG_DIST): New macro.
4842 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4843 STMT_VINFO_MIN_NEG_DIST.
4844 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4845 made for negative dependence distances still hold.
4847 2014-02-21 Richard Biener <rguenther@suse.de>
4850 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4851 DECL_INITIAL for globals not in the current function context.
4853 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4855 PR tree-optimization/56490
4856 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4857 * tree-ssa-uninit.c: Include params.h.
4858 (compute_control_dep_chain): Add num_calls argument, return false
4859 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4860 num_calls to recursive call.
4861 (find_predicates): Change dep_chain into normal array,
4862 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4863 variable and adjust compute_control_dep_chain caller.
4864 (find_def_preds): Likewise.
4866 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4868 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4869 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4871 2014-02-21 Nick Clifton <nickc@redhat.com>
4873 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4874 (pushhi1): Likewise.
4875 (popqi1): Add mode to pre_dec.
4878 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4880 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4881 mode for mask of V8SFmode permutation.
4883 2014-02-20 Richard Henderson <rth@redhat.com>
4886 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4887 a new pseudo for OLDVAL.
4889 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4892 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4893 gen_reg_rtx if d->testing_p.
4894 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4895 if d->testing_p and we will certainly return true.
4896 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4899 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4901 * emit-rtl.c (gen_reg_rtx): Assert that
4902 crtl->emit.regno_pointer_align_length is non-zero.
4904 2014-02-20 Richard Henderson <rth@redhat.com>
4907 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4908 on failure the store back into EXPECT.
4910 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4911 Sandra Loosemore <sandra@codesourcery.com>
4913 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4914 * config/nios2/nios2.c (nios2_function_profiler): Add
4915 -fPIC (flag_pic == 2) support.
4916 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4917 (nios2_large_offset_p): New function.
4918 (nios2_unspec_reloc_p): Move up position, update to use
4919 nios2_large_offset_p.
4920 (nios2_unspec_address): Remove function.
4921 (nios2_unspec_offset): New function.
4922 (nios2_large_got_address): New function.
4923 (nios2_got_address): Add large offset support.
4924 (nios2_legitimize_tls_address): Update usage of removed and new
4926 (nios2_symbol_binds_local_p): New function.
4927 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4928 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4929 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4930 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4931 processing for (const (unspec ...)).
4932 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4934 2014-02-20 Richard Biener <rguenther@suse.de>
4936 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4937 doing the substitution.
4938 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4940 2014-02-20 Martin Jambor <mjambor@suse.cz>
4943 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4944 info when checking whether lattices are bottom.
4946 2014-02-20 Richard Biener <rguenther@suse.de>
4949 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4952 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4955 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4956 parameter specifying the scaling.
4957 (inline_call): Update.
4958 (want_inline_recursively): Guard division by zero.
4959 (recursive_inlining): Update.
4960 * ipa-inline.h (clone_inlined_nodes): Update.
4962 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4965 * config/i386/i386.c (classify_argument): Pass structures of size
4966 64 bytes or less in register.
4968 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4969 Kirill Yukhin <kirill.yukhin@intel.com>
4971 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
4972 (_mm_rcp28_round_ss): Ditto.
4973 (_mm_rsqrt28_round_sd): Ditto.
4974 (_mm_rsqrt28_round_ss): Ditto.
4975 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
4976 (_mm_rcp14_round_ss): Ditto.
4977 (_mm_rsqrt14_round_sd): Ditto.
4978 (_mm_rsqrt14_round_ss): Ditto.
4979 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
4980 the first input operand, get rid of match_dup.
4981 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
4983 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
4985 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
4986 operand as the first input operand, set type attribute.
4987 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
4989 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
4990 operand as the first input operand, set type attribute.
4992 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4994 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
4997 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5000 * config/i386/i386.c (construct_container): Remove TFmode check
5001 for X86_64_INTEGER_CLASS.
5003 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5006 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5007 only when -Wpsabi is enabled.
5009 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5012 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5013 passing arrays in registers are the same as for structs, so remove the
5014 special case for them.
5016 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5018 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5019 destination type, extract only the valid bits if the source type is not
5020 integral and has a different mode.
5022 2014-02-19 Richard Biener <rguenther@suse.de>
5025 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5028 2014-02-19 Richard Biener <rguenther@suse.de>
5031 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5032 (ipa_modify_call_arguments): Emit an argument load explicitely and
5033 preserve virtual SSA form there and for the replacement call.
5034 Do not update SSA form nor free dominance info.
5036 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5038 * ipa.c (function_and_variable_visibility): Also clear WEAK
5039 flag when disolving COMDAT_GROUP.
5041 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5043 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5044 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5046 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5047 do more sanity checks.
5048 (detect_type_change): Return true when giving up early.
5049 (compute_complex_assign_jump_func): Fix type parameter of
5050 ipa_set_ancestor_jf.
5051 (compute_complex_ancestor_jump_func): Likewise.
5052 (update_jump_functions_after_inlining): Fix updating of
5054 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5056 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5058 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5059 inline clones when edge disappears.
5061 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5064 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5065 Split 64-bit moves into 2 patterns. Do not allow the use of
5066 direct move for TDmode in little endian, since the decimal value
5067 has little endian bytes within a word, but the 64-bit pieces are
5068 ordered in a big endian fashion, and normal subreg's of TDmode are
5070 (mov<mode>_64bit_dm): Likewise.
5071 (movtd_64bit_nodm): Likewise.
5073 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5075 PR tree-optimization/60174
5076 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5077 statement of an SSA_NAME that occurs in an abnormal PHI node.
5079 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5082 * final.c (SEEN_BB): Remove.
5083 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5084 (final_scan_insn): Don't force_source_line on second
5085 NOTE_INSN_BASIC_BLOCK.
5087 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5090 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5091 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5092 (type_natural_mode): Warn ABI change when %zmm register is not
5093 available for AVX512F vector value passing.
5095 2014-02-18 Kai Tietz <ktietz@redhat.com>
5098 * config/i386/i386.c (ix86_expand_prologue): Use value in
5099 rax register as displacement when restoring %r10 or %rax.
5100 Fix wrong offset when restoring both registers.
5102 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5104 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5105 assertion with conditional return.
5107 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5108 Uros Bizjak <ubizjak@gmail.com>
5111 * config/i386/driver-i386.c (host_detect_local_cpu): If
5112 YMM state is not saved by the OS, also clear has_f16c. Move
5113 CPUID 0x80000001 handling before YMM state saving checking.
5115 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5117 PR rtl-optimization/58960
5118 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5119 factored out from ...
5120 (sched_init): ... here.
5121 (free_global_sched_pressure_data): New, factored out from ...
5122 (sched_finish): ... here.
5123 * sched-int.h (free_global_sched_pressure_data): Declare.
5124 * sched-rgn.c (nr_regions_initial): New static global.
5125 (haifa_find_rgns): Initialize it.
5126 (schedule_region): Disable sched-pressure for the newly
5129 2014-02-17 Richard Biener <rguenther@suse.de>
5131 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5132 release SSA defs of pattern stmts.
5134 2014-02-17 Richard Biener <rguenther@suse.de>
5136 * tree-inline.c (expand_call_inline): Release the virtual
5137 operand defined by the call we are about to inline.
5139 2014-02-17 Richard Biener <rguenther@suse.de>
5141 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5143 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5144 Ilya Tocar <ilya.tocar@intel.com>
5146 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5147 arguments order in builtin.
5148 (_mm512_permutexvar_epi64): Ditto.
5149 (_mm512_mask_permutexvar_epi64): Ditto
5150 (_mm512_maskz_permutexvar_epi32): Ditto
5151 (_mm512_permutexvar_epi32): Ditto
5152 (_mm512_mask_permutexvar_epi32): Ditto
5154 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5156 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5157 (p8_vmrgow): Likewise.
5159 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5161 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5164 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5167 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5168 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5169 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5170 using direct move instructions on ISA 2.07. Also adjust
5171 instruction length for 64-bit.
5172 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5173 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5175 2014-02-15 Alan Modra <amodra@gmail.com>
5179 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5180 find_replacement on parts of insn rtl that might be reloaded.
5182 2014-02-15 Richard Biener <rguenther@suse.de>
5184 PR tree-optimization/60183
5185 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5186 (tree_ssa_phiprop): Calculate and free post-dominators.
5188 2014-02-14 Jeff Law <law@redhat.com>
5190 PR rtl-optimization/60131
5191 * ree.c (get_extended_src_reg): New function.
5192 (combine_reaching_defs): Use it rather than assuming location of REG.
5193 (find_and_remove_re): Verify first operand of extension is
5194 a REG before adding the insns to the copy list.
5196 2014-02-14 Roland McGrath <mcgrathr@google.com>
5198 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5199 * configure: Regenerated.
5200 * config.in: Regenerated.
5201 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5202 instead of ASM_SHORT.
5204 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5205 Richard Earnshaw <rearnsha@arm.com>
5207 PR rtl-optimization/59535
5208 * lra-constraints.c (process_alt_operands): Encourage alternative
5209 when unassigned pseudo class is superset of the alternative class.
5210 (inherit_reload_reg): Don't inherit when optimizing for code size.
5211 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5212 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5213 modes not less than 4 for Thumb1.
5215 2014-02-14 Kyle McMartin <kyle@redhat.com>
5218 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5220 2014-02-14 Richard Biener <rguenther@suse.de>
5222 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5223 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5224 Do not strip INDIRECT_REFs.
5226 2014-02-14 Richard Biener <rguenther@suse.de>
5229 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5230 DECL_FUNCTION_SPECIFIC_TARGET.
5231 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5232 * tree-streamer-out.c (pack_ts_target_option): Remove.
5233 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5234 (write_ts_function_decl_tree_pointers): Do not stream
5235 DECL_FUNCTION_SPECIFIC_TARGET.
5236 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5237 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5238 (lto_input_ts_function_decl_tree_pointers): Do not stream
5239 DECL_FUNCTION_SPECIFIC_TARGET.
5241 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5243 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5244 (get_initial_def_for_induction, vectorizable_induction): Ignore
5245 debug stmts when looking for exit_phi.
5246 (vectorizable_live_operation): Fix up condition.
5248 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5250 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5251 nreverse() because it changes the content of original tree list.
5253 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5255 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5256 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5258 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5260 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5261 GNU coding standards.
5263 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5266 * dwarf2out.c (gen_subprogram_die): Don't call
5267 add_calling_convention_attribute if subr_die is old_die.
5269 2014-02-13 Sharad Singhai <singhai@google.com>
5271 * doc/optinfo.texi: Fix order of nodes.
5273 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5275 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5276 operands[2], not operands[3].
5278 2014-02-13 Richard Biener <rguenther@suse.de>
5281 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5282 mention the possibility of an in-tree build.
5283 (CLooG): Update recommended version to 0.18.1, mention the
5284 possibility of an in-tree build and clarify that the ISL
5285 bundled with CLooG does not work.
5287 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5290 * expr.c (compress_float_constant): If x is a hard register,
5291 extend into a pseudo and then move to x.
5293 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5295 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5296 caused by bad second argument to warning_at() with -mhotpatch and
5297 nested functions (e.g. with gfortran).
5299 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5301 * opts.c (option_name): Remove "enabled by default" rider.
5303 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5305 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5307 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5308 Uros Bizjak <ubizjak@gmail.com>
5311 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5312 * configure: Regenerated.
5314 2014-02-12 Richard Biener <rguenther@suse.de>
5316 * vec.c (vec_prefix::calculate_allocation): Move as
5317 inline variant to vec.h.
5318 (vec_prefix::calculate_allocation_1): New out-of-line version.
5319 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5320 (vec_prefix::m_has_auto_buf): Rename to ...
5321 (vec_prefix::m_using_auto_storage): ... this.
5322 (vec_prefix::calculate_allocation): Inline the easy cases
5323 and dispatch to calculate_allocation_1 which doesn't need the
5325 (va_heap::reserve): Use gcc_checking_assert.
5326 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5327 m_using_auto_storage.
5328 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5330 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5331 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5332 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5334 2014-02-12 Richard Biener <rguenther@suse.de>
5336 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5337 when we found a dependence.
5339 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5341 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5343 (maybe_fold_stmt): ... into this new function.
5344 * omp-low.c (lower_omp): Update comment.
5346 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5349 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5352 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5354 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5355 identifiers in comments.
5356 (cortexa53_extra_costs): Likewise.
5357 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5358 (cortexa7_extra_costs): Likewise.
5359 (cortexa12_extra_costs): Likewise.
5360 (cortexa15_extra_costs): Likewise.
5361 (v7m_extra_costs): Likewise.
5363 2014-02-12 Richard Biener <rguenther@suse.de>
5366 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5367 of posix_memalign being successful.
5368 (lower_stmt): Restrict lowering of posix_memalign to when
5369 -ftree-bit-ccp is enabled.
5371 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5373 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5375 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5377 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5379 PR rtl-optimization/60116
5380 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5381 other_insn once the combination has been validated.
5383 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5386 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5388 * ipa-devirt.c: Include demangle.h
5389 (odr_violation_reported): New static variable.
5390 (add_type_duplicate): Update odr_violations.
5391 (maybe_record_node): Add completep parameter; update it.
5392 (record_target_from_binfo): Add COMPLETEP parameter;
5393 update it as needed.
5394 (possible_polymorphic_call_targets_1): Likewise.
5395 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5396 rename FINAL to COMPLETE.
5397 (record_targets_from_bases): Sanity check we found the binfo;
5398 fix COMPLETEP updating.
5399 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5400 parameter, fix computing of COMPLETEP.
5401 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5402 at LTO time do demangling.
5403 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5404 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5406 (gimple_get_virt_method_for_binfo): Likewise.
5407 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5408 gimple_get_virt_method_for_vtable): Update prototypes.
5410 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5413 * genautomata.c (add_presence_absence): Fix typo with
5414 {final_}presence_list.
5416 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5419 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5420 for VSX/Altivec vectors that land in GPR registers.
5422 2014-02-11 Richard Henderson <rth@redhat.com>
5423 Jakub Jelinek <jakub@redhat.com>
5426 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5427 around drhs if type conversion to lacc->type is not useless.
5429 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5431 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5433 (cortex-a57.cortex-a53): Likewise.
5434 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5436 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5438 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5441 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5443 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5444 add_options_for_arm_vfp3.
5446 2014-02-11 Jeff Law <law@redhat.com>
5449 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5450 object with an undesirable mode.
5452 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5455 * config/i386/sol2-9.h: New file.
5456 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5457 *-*-solaris2.9*): Use it.
5459 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5461 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5462 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5464 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5466 * config/microblaze/microblaze.c: Extend mcpu version format
5468 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5470 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5472 2014-02-10 Richard Henderson <rth@redhat.com>
5475 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5476 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5477 ms-abi vs -mno-accumulate-outgoing-args.
5478 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5479 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5482 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5485 * cfgexpand.c (expand_asm_operands): Attach source location to
5486 ASM_INPUT rtx objects.
5487 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5489 2014-02-10 Nick Clifton <nickc@redhat.com>
5491 * config/mn10300/mn10300.c (popcount): New function.
5492 (mn10300_expand_prologue): Include saved registers in stack usage
5495 2014-02-10 Jeff Law <law@redhat.com>
5498 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5499 when changing the SET_DEST of a prior insn to avoid an input reload.
5501 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5503 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5504 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5505 -mcall-openbsd, or -mcall-linux.
5506 (CC1_ENDIAN_BIG_SPEC): Remove.
5507 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5508 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5509 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5510 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5511 and %cc1_endian_default.
5512 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5514 2014-02-10 Richard Biener <rguenther@suse.de>
5516 PR tree-optimization/60115
5517 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5518 MEM_REF handling. Properly verify that the accesses are not
5519 out of the objects bound.
5521 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5523 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5526 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5528 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5529 proper constants and fix formatting.
5530 (possible_polymorphic_call_targets): Fix formatting.
5532 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5533 Ilya Tocar <ilya.tocar@intel.com>
5535 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5536 (_mm512_loadu_epi32): Renamed into...
5537 (_mm512_loadu_si512): This.
5538 (_mm512_storeu_epi32): Renamed into...
5539 (_mm512_storeu_si512): This.
5540 (_mm512_maskz_ceil_ps): Removed.
5541 (_mm512_maskz_ceil_pd): Ditto.
5542 (_mm512_maskz_floor_ps): Ditto.
5543 (_mm512_maskz_floor_pd): Ditto.
5544 (_mm512_floor_round_ps): Ditto.
5545 (_mm512_floor_round_pd): Ditto.
5546 (_mm512_ceil_round_ps): Ditto.
5547 (_mm512_ceil_round_pd): Ditto.
5548 (_mm512_mask_floor_round_ps): Ditto.
5549 (_mm512_mask_floor_round_pd): Ditto.
5550 (_mm512_mask_ceil_round_ps): Ditto.
5551 (_mm512_mask_ceil_round_pd): Ditto.
5552 (_mm512_maskz_floor_round_ps): Ditto.
5553 (_mm512_maskz_floor_round_pd): Ditto.
5554 (_mm512_maskz_ceil_round_ps): Ditto.
5555 (_mm512_maskz_ceil_round_pd): Ditto.
5556 (_mm512_expand_pd): Ditto.
5557 (_mm512_expand_ps): Ditto.
5558 * config/i386/i386.c (ix86_builtins): Remove
5559 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5560 (bdesc_args): Ditto.
5561 * config/i386/predicates.md (const1256_operand): New.
5562 (const_1_to_2_operand): Ditto.
5563 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5564 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5565 (*avx512pf_gatherpf<mode>sf): Ditto.
5566 (avx512pf_gatherpf<mode>df): Ditto.
5567 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5568 (*avx512pf_gatherpf<mode>df): Ditto.
5569 (avx512pf_scatterpf<mode>sf): Ditto.
5570 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5571 (*avx512pf_scatterpf<mode>sf): Ditto.
5572 (avx512pf_scatterpf<mode>df): Ditto.
5573 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5574 (*avx512pf_scatterpf<mode>df): Ditto.
5575 (avx512f_expand<mode>): Removed.
5576 (<shift_insn><mode>3<mask_name>): Change predicate type.
5578 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5580 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5581 not at the end of datarefs vector use ordered_remove to avoid
5582 reordering datarefs vector.
5585 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5586 mark local addressable non-static vars as GOVD_PRIVATE
5587 instead of GOVD_LOCAL.
5588 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5589 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5593 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5594 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5595 assume_aligned or alloc_align attributes.
5596 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5597 arguments. Handle also assume_aligned and alloc_align attributes.
5598 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5599 calls to functions with assume_aligned or alloc_align attributes.
5600 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5602 2014-02-08 Terry Guo <terry.guo@arm.com>
5604 * doc/invoke.texi: Document ARM -march=armv7e-m.
5606 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5608 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5609 flag on __cilkrts_rethrow builtin.
5612 * ipa-cp.c (determine_versionability): Fail at -O0
5613 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5614 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5617 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5620 * tree-inline.c (copy_forbidden): Fail for
5621 __attribute__((optimize (0))) functions.
5623 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5625 * varpool.c: Include pointer-set.h.
5626 (varpool_remove_unreferenced_decls): Variables in other partitions
5627 will not be output; be however careful to not lose information
5630 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5632 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5633 lookup in the vtable constructor.
5635 2014-02-07 Jeff Law <law@redhat.com>
5638 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5639 define_insn_and_split.
5641 * ipa-inline.c (inline_small_functions): Fix typos.
5643 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5645 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5646 (s390_can_use_return_insn): Declare.
5647 * config/s390/s390.h (EPILOGUE_USES): Define.
5648 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5650 (s390_chunkify_start): Handle return JUMP_LABELs.
5651 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5652 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5653 (s390_can_use_return_insn): New functions.
5654 (s390_fix_long_loop_prediction): Handle conditional returns.
5655 (TARGET_SET_UP_BY_PROLOGUE): Define.
5656 * config/s390/s390.md (ANY_RETURN): New code iterator.
5657 (*creturn, *csimple_return, return, simple_return): New patterns.
5659 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5661 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5662 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5663 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5664 REG_CFA_RESTORE list when deciding not to restore a register.
5666 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5668 * config/s390/s390.c: Include tree-pass.h and context.h.
5669 (s390_early_mach): New function, split out from...
5670 (s390_emit_prologue): ...here.
5671 (pass_data_s390_early_mach): New pass structure.
5672 (pass_s390_early_mach): New class.
5673 (s390_option_override): Create and register early_mach pass.
5674 Move to end of file.
5676 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5678 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5679 to match for the exit block.
5681 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5683 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5684 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5685 Reject misaligned operands.
5687 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5689 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5691 2014-02-07 Richard Biener <rguenther@suse.de>
5694 * gimple-low.c (lower_builtin_posix_memalign): New function.
5695 (lower_stmt): Call it to lower posix_memalign in a way
5696 to make alignment info accessible.
5698 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5701 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5702 __builtin_setjmp_receiver.
5704 2014-02-07 Richard Biener <rguenther@suse.de>
5707 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5708 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5709 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5710 Handle BUILT_IN_POSIX_MEMALIGN.
5711 (find_func_clobbers): Likewise.
5712 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5713 (call_may_clobber_ref_p_1): Likewise.
5715 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5718 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5721 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5724 * lto-cgraph.c (lto_output_node): Use
5725 symtab_get_symbol_partitioning_class.
5726 (lto_output_varpool_node): likewise.
5727 (symtab_get_symbol_partitioning_class): Move here from
5729 * cgraph.h (symbol_partitioning_class): Likewise.
5730 (symtab_get_symbol_partitioning_class): Declare.
5732 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5734 * ggc.h (ggc_internal_cleared_alloc): New macro.
5735 * vec.h (vec_safe_copy): Handle memory stats.
5736 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5737 * target-globals.c (save_target_globals): Likewise.
5739 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5742 * expr.c (emit_move_resolve_push): Export; be bit more selective
5743 on when to clear alias set.
5744 * expr.h (emit_move_resolve_push): Declare.
5745 * function.h (struct function): Add tail_call_marked.
5746 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5747 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5748 * config/i386/i386.md (TImode move expander): De not call
5750 (FP push expanders): Preserve memory attributes.
5751 * config/i386/sse.md (push<mode>1): Remove.
5752 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5753 (ix86_expand_push): Remove.
5754 * config/i386/mmx.md (push<mode>1): Remove.
5756 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5758 PR rtl-optimization/60030
5759 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5760 lopart with paradoxical subreg before shifting it up by hprec.
5762 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5764 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5765 Remove extra newline at end of file.
5766 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5767 (arm_issue_rate): Handle cortexa57.
5768 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5769 (cortex-a57.cortex-a53): Likewise.
5771 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5774 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5775 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5777 (arm_expand_prologue): Adjust all callers.
5778 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5779 info, registers also at the lowest numbered registers side. Use
5780 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5784 * var-tracking.c (adjust_mems): Before adding a SET to
5785 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5787 2014-02-06 Alan Modra <amodra@gmail.com>
5790 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5791 change SDmode to DDmode when lra_in_progress.
5793 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5796 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5797 free_data_ref on the dr first, and before goto again also set dr
5798 to the next dr. For simd_lane_access, free old datarefs[i] before
5799 overwriting it. For get_vectype_for_scalar_type failure, don't
5800 free_data_ref if simd_lane_access.
5802 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5805 * tree.h (opts_for_fn): New inline function.
5806 (opt_for_fn): Define.
5807 * config/i386/i386.c (ix86_function_regparm): Use
5808 opt_for_fn (decl, optimize) instead of optimize.
5810 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5812 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5813 for SYMBOL_REF in large memory model.
5815 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5817 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5819 (cortex-a57): Likewise.
5820 (cortex-a57.cortex-a53): Likewise.
5822 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5823 Kugan Vivekanandarajah <kuganv@linaro.org>
5825 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5827 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5829 2014-02-06 Richard Biener <rguenther@suse.de>
5831 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5832 set_loop_copy and initialize_original_copy_tables.
5834 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5836 * config/aarch64/aarch64-simd.md
5837 (aarch64_ashr_simddi): Change QI to SI.
5839 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5840 Jakub Jelinek <jakub@redhat.com>
5843 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5846 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5848 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5849 CODE_FOR_altivec_vpku[hw]um to
5850 CODE_FOR_altivec_vpku[hw]um_direct.
5851 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5852 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5853 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5854 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5856 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5858 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5859 generation for -maltivec=be.
5860 (altivec_vsumsws): Simplify redundant test.
5862 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5864 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5865 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5866 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5867 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5868 gen_altivec_vpkuwum.
5869 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5871 (altivec_vpks<VI_char>ss): Likewise.
5872 (altivec_vpks<VI_char>us): Likewise.
5873 (altivec_vpku<VI_char>us): Likewise.
5874 (altivec_vpku<VI_char>um): Likewise.
5875 (altivec_vpku<VI_char>um_direct): New (copy of
5876 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5878 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5879 target is little endian and -maltivec=be is not specified.
5880 (*altivec_vupkhs<VU_char>_direct): New (copy of
5881 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5882 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5883 target is little endian and -maltivec=be is not specified.
5884 (*altivec_vupkls<VU_char>_direct): New (copy of
5885 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5886 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5887 little endian and -maltivec=be is not specified.
5888 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5889 little endian and -maltivec=be is not specified.
5891 2014-02-05 Richard Henderson <rth@redhat.com>
5894 * combine-stack-adj.c: Revert r206943.
5895 * sched-int.h (struct deps_desc): Add last_args_size.
5896 * sched-deps.c (init_deps): Initialize it.
5897 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5898 contain REG_ARGS_SIZE notes.
5900 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5902 * lto-cgraph.c (asm_nodes_output): Make global.
5903 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5904 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5905 (driver_handle_option): Handle OPT_fwpa.
5907 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5910 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5911 a comment typo and formatting issue. If odr_hash hasn't been
5912 created, return vNULL and set *completep to false.
5915 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5916 bb with no successors.
5918 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5921 * doc/invoke.texi (-march): Clarify documentation for ARM.
5925 2014-02-05 Richard Biener <rguenther@suse.de>
5927 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5928 when not vectorizing because of too many alias checks.
5929 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5930 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5932 2014-02-05 Nick Clifton <nickc@redhat.com>
5934 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5935 accept extended registers in any mode when compiling for the MN10300.
5937 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5939 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5940 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5941 sanitization attributes.
5942 (can_inline_edge_p): Likewise.
5943 (sanitize_attrs_match_for_inline_p): New function.
5945 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5947 * ipa-prop.c (detect_type_change): Shor circuit testing of
5948 type changes on THIS pointer.
5950 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5953 * config/pa/pa.c (legitimize_tls_address): Return original address
5954 if not passed a SYMBOL_REF rtx.
5955 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5957 (pa_emit_move_sequence): Simplify TLS source operands.
5958 (pa_legitimate_constant_p): Reject all TLS constants.
5959 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5960 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5962 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5964 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5965 groups when we know they are controlled by LTO.
5966 * varasm.c (default_binds_local_p_1): If object is in other partition,
5967 it will be resolved locally.
5969 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
5971 * config/host-linux.c (linux_gt_pch_use_address): Don't
5972 use SSIZE_MAX because it is not always defined.
5974 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
5977 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
5978 threshold for pseudo splitting.
5979 (update_ebb_live_info): Process call argument hard registers and
5980 hard registers from insn definition too.
5981 (max_small_class_regs_num): New constant.
5982 (inherit_in_ebb): Update live hard regs through EBBs. Update
5983 reloads_num only for small register classes. Don't split for
5986 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
5989 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
5992 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5994 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
5997 2014-02-04 Marek Polacek <polacek@redhat.com>
5999 * gdbinit.in (pel): Define.
6001 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6003 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6006 2014-02-04 Richard Biener <rguenther@suse.de>
6009 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6010 in function context local.
6011 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6012 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6013 similar to LTO_imported_decl_ref.
6015 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6017 PR tree-optimization/60002
6018 * cgraphclones.c (build_function_decl_skip_args): Clear
6021 PR tree-optimization/60023
6022 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6023 false to gsi_replace.
6024 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6025 has been in some EH region and vec_stmt could throw, add
6026 vec_stmt into the same EH region.
6027 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6028 has no lhs, ignore it.
6029 * internal-fn.c (expand_MASK_LOAD): Likewise.
6032 * tree-inline.c (copy_forbidden): Fail for
6033 __attribute__((optimize (0))) functions.
6036 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6037 is set, copy one less argument.
6038 (expand_simd_clones): Don't subtract clone_info->inbranch
6039 from simd_clone_struct_alloc argument.
6041 PR rtl-optimization/57915
6042 * recog.c (simplify_while_replacing): If all unary/binary/relational
6043 operation arguments are constant, attempt to simplify those.
6046 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6047 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6049 2014-02-04 Richard Biener <rguenther@suse.de>
6051 PR tree-optimization/60012
6052 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6053 TBAA disambiguation to all DDRs.
6055 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6058 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6059 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6061 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6064 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6066 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6068 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6069 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6071 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6074 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6075 to figure out targets of polymorphic calls with known decl.
6076 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6077 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6078 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6079 (get_polymorphic_call_info): ... here.
6080 (get_polymorphic_call_info_from_invariant): New function.
6082 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6084 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6085 lookup via vtable pointer; check for type consistency
6086 and turn inconsitent facts into UNREACHABLE.
6087 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6088 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6089 type inconsistent querries; return UNREACHABLE instead.
6091 2014-02-03 Richard Henderson <rth@twiddle.net>
6094 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6095 already processed this node.
6096 (normalize_one_pred_1): Pass along mark_set.
6097 (normalize_one_pred): Create and destroy a pointer_set_t.
6098 (normalize_one_pred_chain): Likewise.
6100 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6102 PR gcov-profile/58602
6103 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6105 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6108 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6109 -fno-devirtualize; try to devirtualize by the knowledge of
6110 virtual table pointer given by aggregate propagation.
6111 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6112 (ipa_print_node_jump_functions): Dump also offset that
6113 is relevant for polymorphic calls.
6114 (determine_known_aggregate_parts): Add arg_type parameter; use it
6115 instead of determining the type from pointer type.
6116 (ipa_compute_jump_functions_for_edge): Update call of
6117 determine_known_aggregate_parts.
6118 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6119 (gimple_get_virt_method_for_binfo): ... here; simplify using
6120 vtable_pointer_value_to_vtable.
6121 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6122 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6123 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6124 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6126 (vtable_pointer_value_to_binfo): ... here.
6127 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6129 2014-02-03 Teresa Johnson <tejohnson@google.com>
6131 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6132 redef of outer loop index variable.
6134 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6138 * doc/extend.texi (Function Attributes): Typo.
6140 2014-02-03 Cong Hou <congh@google.com>
6142 PR tree-optimization/60000
6143 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6144 if the vectorized statement is a store. A store statement can only
6145 appear at the end of pattern statements.
6147 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6149 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6150 (ix86_option_override_internal): Default long double to 64-bit for
6151 32-bit Bionic and to 128-bit for 64-bit Bionic.
6153 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6154 TARGET_LONG_DOUBLE_128 is true.
6155 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6157 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6158 (mlong-double-64): Negate -mlong-double-128.
6159 (mlong-double-128): New option.
6161 * config/i386/i386-c.c (ix86_target_macros): Define
6162 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6164 * doc/invoke.texi: Document -mlong-double-128.
6166 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6168 PR rtl-optimization/60024
6169 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6171 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6173 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6175 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6177 PR rtl-optimization/57662
6178 * sel-sched.c (code_motion_path_driver): Do not mark already not
6179 existing blocks in the visiting bitmap.
6181 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6183 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6184 on the insn being emitted.
6186 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6187 Will Deacon <will.deacon@arm.com>
6189 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6191 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6193 * config/arm/arm-tables.opt: Regenerate.
6195 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6197 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6198 for vector types other than V16QImode.
6199 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6200 define_expand, and call altivec_expand_vec_perm_le when producing
6201 code with little endian element order.
6202 (*altivec_vperm_<mode>_internal): New insn having previous
6203 behavior of altivec_vperm_<mode>.
6204 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6205 altivec_expand_vec_perm_le when producing code with little endian
6207 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6208 behavior of altivec_vperm_<mode>_uns.
6210 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6212 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6213 (altivec_vsumsws): Add handling for -maltivec=be with a little
6215 (altivec_vsumsws_direct): New.
6216 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6217 gen_altivec_vsumsws.
6219 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6221 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6222 vtable_pointer_value_to_binfo): New functions.
6223 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6224 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6226 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6228 * config/nios2/nios2.md (load_got_register): Initialize GOT
6229 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6230 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6232 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6234 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6235 preserverd by passthrough, do not propagate the type.
6237 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6239 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6240 (mips_atomic_assign_expand_fenv): New function.
6241 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6243 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6245 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6246 (__builtin_mips_set_fcsr): Likewise.
6247 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6248 MIPS_USI_FTYPE_VOID.
6249 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6250 (mips16_expand_set_fcsr): Likewise.
6251 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6252 (mips16_set_fcsr_stub): Likewise.
6253 (mips16_get_fcsr_one_only_stub): New class.
6254 (mips16_set_fcsr_one_only_stub): Likewise.
6255 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6256 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6257 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6258 (hard_float): New availability predicate.
6259 (mips_builtins): Add get_fcsr and set_fcsr.
6260 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6261 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6262 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6263 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6264 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6267 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6269 * config/mips/mips.c (mips_one_only_stub): New class.
6270 (mips_need_mips16_rdhwr_p): Replace with...
6271 (mips16_rdhwr_stub): ...this new variable.
6272 (mips16_stub_call_address): New function.
6273 (mips16_rdhwr_one_only_stub): New class.
6274 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6275 (mips_output_mips16_rdhwr): Delete.
6276 (mips_finish_stub): New function.
6277 (mips_code_end): Use it to handle rdhwr stubs.
6279 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6282 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6283 when calculating size of integer atomic types.
6285 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6287 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6289 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6291 PR tree-optimization/60003
6292 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6293 * profile.c (branch_prob): Use gimple_call_builtin_p
6294 to check for BUILT_IN_SETJMP_RECEIVER.
6295 * tree-inline.c (copy_bb): Call notice_special_calls.
6297 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6300 * lra-constraints.c (process_alt_operands): Update reload_sum only
6303 2014-01-31 Richard Henderson <rth@redhat.com>
6306 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6307 until after else_eh is processed.
6309 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6311 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6312 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6313 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6314 in smmintrin.h, remove them.
6315 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6316 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6317 * config/i386/i386.md (ROUND_SAE): Fix value.
6318 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6319 (const48_operand): New.
6320 * config/i386/subst.md (round), (round_expand): Use
6321 const_4_or_8_to_11_operand.
6322 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6324 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6326 * config/i386/constraints.md (Yk): Swap meaning with k.
6327 * config/i386/i386.md (movhi_internal): Change Yk to k.
6328 (movqi_internal): Ditto.
6329 (*k<logic><mode>): Ditto.
6332 (kandn<mode>): Ditto.
6333 (*<code>hi_1): Ditto.
6334 (*<code>qi_1): Ditto.
6335 (kxnor<mode>): Ditto.
6336 (kortestzhi): Ditto.
6337 (kortestchi): Ditto.
6339 (*one_cmplhi2_1): Ditto.
6340 (*one_cmplqi2_1): Ditto.
6341 * config/i386/sse.md (): Change k to Yk.
6342 (avx512f_load<mode>_mask): Ditto.
6343 (avx512f_blendm<mode>): Ditto.
6344 (avx512f_store<mode>_mask): Ditto.
6345 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6346 (avx512f_storedqu<mode>_mask): Ditto.
6347 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6349 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6350 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6351 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6352 (avx512f_maskcmp<mode>3): Ditto.
6353 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6354 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6355 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6356 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6357 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6358 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6359 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6360 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6361 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6362 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6363 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6364 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6365 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6366 (vec_extract_lo_<mode>_maskm): Ditto.
6367 (vec_extract_hi_<mode>_maskm): Ditto.
6368 (avx512f_vternlog<mode>_mask): Ditto.
6369 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6370 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6371 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6372 (avx512f_<code>v8div16qi2_mask): Ditto.
6373 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6374 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6375 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6376 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6377 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6378 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6379 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6380 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6381 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6382 (avx512cd_maskb_vec_dupv8di): Ditto.
6383 (avx512cd_maskw_vec_dupv16si): Ditto.
6384 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6385 (avx512f_vpermi2var<mode>3_mask): Ditto.
6386 (avx512f_vpermi2var<mode>3_mask): Ditto.
6387 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6388 (*avx512f_gathersi<mode>): Ditto.
6389 (*avx512f_gathersi<mode>_2): Ditto.
6390 (*avx512f_gatherdi<mode>): Ditto.
6391 (*avx512f_gatherdi<mode>_2): Ditto.
6392 (*avx512f_scattersi<mode>): Ditto.
6393 (*avx512f_scatterdi<mode>): Ditto.
6394 (avx512f_compress<mode>_mask): Ditto.
6395 (avx512f_compressstore<mode>_mask): Ditto.
6396 (avx512f_expand<mode>_mask): Ditto.
6397 * config/i386/subst.md (mask): Change k to Yk.
6398 (mask_scalar_merge): Ditto.
6401 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6403 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6405 2014-01-31 Richard Biener <rguenther@suse.de>
6408 * builtins.c (fold_builtin_memory_op): Make sure to not
6409 use a floating-point mode or a boolean or enumeral type for
6412 2014-01-30 DJ Delorie <dj@redhat.com>
6414 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6415 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6416 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6417 whenever main() has an epilogue.
6419 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6421 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6422 unused variable "field".
6423 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6424 (vsx_mergeh_<mode>): Likewise.
6425 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6426 (altivec_vmrghh): Likewise.
6427 (altivec_vmrghw): Likewise.
6428 (altivec_vmrglb): Likewise.
6429 (altivec_vmrglh): Likewise.
6430 (altivec_vmrglw): Likewise.
6431 (altivec_vspltb): Add missing uses.
6432 (altivec_vsplth): Likewise.
6433 (altivec_vspltw): Likewise.
6434 (altivec_vspltsf): Likewise.
6436 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6439 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6440 frame related instructions.
6442 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6444 PR rtl-optimization/59959
6445 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6446 any reload of register whose subreg is invalid.
6448 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6450 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6451 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6452 Add missing return type - void.
6454 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6456 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6457 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6458 remove element index adjustment for endian (now handled in vsx.md
6460 (altivec_expand_vec_perm_const): Use
6461 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6462 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6463 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6464 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6465 define_expand and a new define_insn *altivec_vspltb_internal;
6466 adjust for -maltivec=be on a little endian target.
6467 (altivec_vspltb_direct): New.
6468 (altivec_vsplth): Divide into a define_expand and a new
6469 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6470 little endian target.
6471 (altivec_vsplth_direct): New.
6472 (altivec_vspltw): Divide into a define_expand and a new
6473 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6474 little endian target.
6475 (altivec_vspltw_direct): New.
6476 (altivec_vspltsf): Divide into a define_expand and a new
6477 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6478 a little endian target.
6480 2014-01-30 Richard Biener <rguenther@suse.de>
6482 PR tree-optimization/59993
6483 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6484 can propagate form the earlier stmt and avoid the transform
6485 when the intermediate result is needed.
6487 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6489 * README.Portability: Fix typo.
6491 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6493 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6494 comparison_operator with ordered_comparison_operator.
6496 2014-01-30 Nick Clifton <nickc@redhat.com>
6498 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6499 Rename to mn10300_store_multiple_regs.
6500 * config/mn10300/mn10300.c: Likewise.
6501 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6502 store_multiple_regs.
6503 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6504 Call mn10300_store_multiple_regs.
6506 2014-01-30 Nick Clifton <nickc@redhat.com>
6507 DJ Delorie <dj@redhat.com>
6509 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6510 %fp 2 to keep registers after it properly word-aligned.
6511 (rl78_alloc_physical_registers_umul): Handle the case where both
6512 input operands are the same.
6514 2014-01-30 Richard Biener <rguenther@suse.de>
6516 PR tree-optimization/59903
6517 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6520 2014-01-30 Jason Merrill <jason@redhat.com>
6523 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6526 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6528 2014-01-30 Richard Biener <rguenther@suse.de>
6530 PR tree-optimization/59951
6531 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6533 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6536 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6537 SFmode to DFmode case.
6539 2014-01-29 DJ Delorie <dj@redhat.com>
6541 * config/msp430/msp430.opt (-minrt): New.
6542 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6544 (ENDFILE_SPEC): Likewise.
6546 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6548 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6549 (estimate_function_body_sizes): Use it.
6551 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6554 * dwarf2out.c (is_cxx_auto): New.
6555 (is_base_type): Use it.
6556 (gen_type_die_with_usage): Likewise.
6558 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6560 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6561 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6562 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6563 -maltivec=be with LE targets.
6564 (vsx_mergeh_<mode>): Likewise.
6565 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6566 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6567 (altivec_vmrghb): Replace with define_expand and new
6568 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6569 (altivec_vmrghb_direct): New define_insn.
6570 (altivec_vmrghh): Replace with define_expand and new
6571 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6572 (altivec_vmrghh_direct): New define_insn.
6573 (altivec_vmrghw): Replace with define_expand and new
6574 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6575 (altivec_vmrghw_direct): New define_insn.
6576 (*altivec_vmrghsf): Adjust for endianness.
6577 (altivec_vmrglb): Replace with define_expand and new
6578 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6579 (altivec_vmrglb_direct): New define_insn.
6580 (altivec_vmrglh): Replace with define_expand and new
6581 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6582 (altivec_vmrglh_direct): New define_insn.
6583 (altivec_vmrglw): Replace with define_expand and new
6584 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6585 (altivec_vmrglw_direct): New define_insn.
6586 (*altivec_vmrglsf): Adjust for endianness.
6587 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6588 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6589 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6590 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6591 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6592 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6593 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6594 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6596 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6598 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6599 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6602 2014-01-29 Richard Biener <rguenther@suse.de>
6604 PR tree-optimization/58742
6605 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6606 associate_pointerplus_align.
6607 (associate_pointerplus_diff): New function.
6608 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6609 and associate_pointerplus_diff.
6611 2014-01-29 Richard Biener <rguenther@suse.de>
6613 * lto-streamer.h (LTO_major_version): Bump to 3.
6614 (LTO_minor_version): Reset to 0.
6616 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6618 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6619 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6620 (arm_file_start): Generate correct asm header for armv7ve.
6621 * config/arm/bpabi.h: Add multilib support for armv7ve.
6622 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6623 and cortex-a15 to armv7ve.
6624 * config/arm/t-aprofile: Add multilib support for armv7ve.
6625 * doc/invoke.texi: Document -march=armv7ve.
6627 2014-01-29 Richard Biener <rguenther@suse.de>
6629 PR tree-optimization/58742
6630 * tree-ssa-forwprop.c (associate_plusminus): Return true
6631 if we changed sth, defer EH cleanup to ...
6632 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6633 (simplify_mult): New function.
6635 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6638 PR tree-optimization/59920
6639 * tree.c (build_common_builtin_nodes): Remove
6640 __builtin_setjmp_dispatcher initialization.
6641 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6642 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6643 instead of gsi_after_labels + manually skipping debug stmts.
6644 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6645 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6646 * tree-inline.c (copy_edges_for_bb): Remove
6647 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6648 argument. Ignore computed_goto_p stmts. Don't call
6649 make_abnormal_goto_edges. If a call might need abnormal edges
6650 for non-local gotos, see if it already has an edge to
6651 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6652 with true argument, don't do anything then, otherwise add
6653 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6654 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6656 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6657 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6658 (lower_stmt): Don't set data->calls_builtin_setjmp.
6659 (lower_builtin_setjmp): Adjust comment.
6660 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6661 * tree-cfg.c (found_computed_goto): Remove.
6662 (factor_computed_gotos): Remove.
6663 (make_goto_expr_edges): Return bool, true for computed gotos.
6664 Don't call make_abnormal_goto_edges.
6665 (build_gimple_cfg): Don't set found_computed_goto, don't call
6666 factor_computed_gotos.
6667 (computed_goto_p): No longer static.
6668 (make_blocks): Don't set found_computed_goto.
6669 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6670 (make_edges): If make_goto_expr_edges returns true, push bb
6671 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6672 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6673 vector. Record mapping between bbs and OpenMP regions if there
6674 are any, adjust make_gimple_omp_edges caller. Call
6675 handle_abnormal_edges.
6676 (make_abnormal_goto_edges): Remove.
6677 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6678 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6679 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6680 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6681 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6682 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6683 filling *region also set *region_idx to (*region)->entry->index.
6686 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6687 For REGs set ORIGINAL_REGNO.
6689 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6691 * doc/md.texi: Mention that a target shouldn't implement
6692 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6695 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6697 PR tree-optimization/59594
6698 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6699 a copy of the datarefs vector rather than the vector itself.
6701 2014-01-28 Jason Merrill <jason@redhat.com>
6704 * dwarf2out.c (auto_die): New static.
6705 (gen_type_die_with_usage): Handle C++1y 'auto'.
6706 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6709 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6712 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6713 (SPEC_X32): Likewise.
6714 (SPEC_64): Likewise.
6715 * config/i386/i386.c (ix86_option_override_internal): Turn off
6716 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6718 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6719 * config/i386/i386.h (TARGET_16BIT): New macro.
6720 (TARGET_16BIT_P): Likewise.
6721 * config/i386/i386.opt: Add m16.
6722 * doc/invoke.texi: Document -m16.
6724 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6726 PR preprocessor/59935
6727 * input.c (location_get_source_line): Bail out on when line number
6728 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6730 2014-01-28 Richard Biener <rguenther@suse.de>
6732 PR tree-optimization/58742
6733 * tree-ssa-forwprop.c (associate_plusminus): Handle
6734 pointer subtraction of the form (T)(P + A) - (T)P.
6736 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6738 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6741 2014-01-28 Richard Biener <rguenther@suse.de>
6744 2014-01-28 Richard Biener <rguenther@suse.de>
6746 PR rtl-optimization/45364
6747 PR rtl-optimization/59890
6748 * var-tracking.c (local_get_addr_clear_given_value): Handle
6749 already cleared slot.
6750 (val_reset): Handle not allocated local_get_addr_cache.
6751 (vt_find_locations): Use post-order on the inverted CFG.
6753 2014-01-28 Richard Biener <rguenther@suse.de>
6755 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6757 2014-01-28 Richard Biener <rguenther@suse.de>
6759 PR rtl-optimization/45364
6760 PR rtl-optimization/59890
6761 * var-tracking.c (local_get_addr_clear_given_value): Handle
6762 already cleared slot.
6763 (val_reset): Handle not allocated local_get_addr_cache.
6764 (vt_find_locations): Use post-order on the inverted CFG.
6766 2014-01-28 Alan Modra <amodra@gmail.com>
6768 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6769 * configure.ac <recursive call for build != host>: Define
6770 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6771 and LD_FOR_BUILD too.
6772 * configure: Regenerate.
6774 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6776 * config/i386/i386.c (get_builtin_code_for_version): Separate
6777 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6778 Broadwell from Haswell.
6780 2014-01-27 Steve Ellcey <sellcey@mips.com>
6782 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6783 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6784 * config/mips/mips.c (mips_option_override): Change setting
6786 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6787 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6788 Change from Mask to Var.
6790 2014-01-27 Jeff Law <law@redhat.com>
6792 * ipa-inline.c (inline_small_functions): Fix typo.
6794 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6796 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6797 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6798 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6799 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6800 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6801 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6802 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6803 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6804 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6805 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6806 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6807 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6808 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6809 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6810 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6811 (_mm512_storeu_epi64): Ditto.
6812 (_mm512_cmpge_epi32_mask): Ditto.
6813 (_mm512_cmpge_epu32_mask): Ditto.
6814 (_mm512_cmpge_epi64_mask): Ditto.
6815 (_mm512_cmpge_epu64_mask): Ditto.
6816 (_mm512_cmple_epi32_mask): Ditto.
6817 (_mm512_cmple_epu32_mask): Ditto.
6818 (_mm512_cmple_epi64_mask): Ditto.
6819 (_mm512_cmple_epu64_mask): Ditto.
6820 (_mm512_cmplt_epi32_mask): Ditto.
6821 (_mm512_cmplt_epu32_mask): Ditto.
6822 (_mm512_cmplt_epi64_mask): Ditto.
6823 (_mm512_cmplt_epu64_mask): Ditto.
6824 (_mm512_cmpneq_epi32_mask): Ditto.
6825 (_mm512_cmpneq_epu32_mask): Ditto.
6826 (_mm512_cmpneq_epi64_mask): Ditto.
6827 (_mm512_cmpneq_epu64_mask): Ditto.
6828 (_mm512_expand_pd): Ditto.
6829 (_mm512_expand_ps): Ditto.
6830 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6831 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6832 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6833 * config/i386/i386.c (ix86_builtins): Add
6834 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6835 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6836 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6837 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6838 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6839 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6840 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6841 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6842 IX86_BUILTIN_PMOVUSQW512_MEM.
6843 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6844 __builtin_ia32_pmovsqd512mem_mask,
6845 __builtin_ia32_pmovqd512mem_mask,
6846 __builtin_ia32_pmovusqw512mem_mask,
6847 __builtin_ia32_pmovsqw512mem_mask,
6848 __builtin_ia32_pmovqw512mem_mask,
6849 __builtin_ia32_pmovusdw512mem_mask,
6850 __builtin_ia32_pmovsdw512mem_mask,
6851 __builtin_ia32_pmovdw512mem_mask,
6852 __builtin_ia32_pmovqb512mem_mask,
6853 __builtin_ia32_pmovusqb512mem_mask,
6854 __builtin_ia32_pmovsqb512mem_mask,
6855 __builtin_ia32_pmovusdb512mem_mask,
6856 __builtin_ia32_pmovsdb512mem_mask,
6857 __builtin_ia32_pmovdb512mem_mask.
6858 (bdesc_args): Add __builtin_ia32_expanddf512,
6859 __builtin_ia32_expandsf512.
6860 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6861 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6862 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6863 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6864 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6865 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6866 (avx512f_<code>v8div16qi2_mask_store): This.
6867 (avx512f_expand<mode>): New.
6869 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6871 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6873 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6874 (_mm512_prefetch_i32scatter_pd): Ditto.
6875 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6876 (_mm512_prefetch_i64scatter_pd): Ditto.
6877 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6878 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6879 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6880 (_mm512_prefetch_i32scatter_ps): Ditto.
6881 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6882 (_mm512_prefetch_i64scatter_ps): Ditto.
6883 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6884 * config/i386/i386-builtin-types.def: Define
6885 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6886 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6887 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6888 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6889 IX86_BUILTIN_SCATTERPFQPD.
6890 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6891 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6892 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6893 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6894 __builtin_ia32_scatterpfqps.
6895 (ix86_expand_builtin): Expand new built-ins.
6896 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6897 fix memory access data type.
6898 (*avx512pf_gatherpf<mode>_mask): Ditto.
6899 (*avx512pf_gatherpf<mode>): Ditto.
6900 (avx512pf_scatterpf<mode>): Ditto.
6901 (*avx512pf_scatterpf<mode>_mask): Ditto.
6902 (*avx512pf_scatterpf<mode>): Ditto.
6903 (GATHER_SCATTER_SF_MEM_MODE): New.
6904 (avx512pf_gatherpf<mode>df): Ditto.
6905 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6906 (*avx512pf_scatterpf<mode>df): Ditto.
6908 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6911 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6912 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6915 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6917 * common/config/arm/arm-common.c
6918 (arm_rewrite_mcpu): Handle multiple names.
6920 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6922 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6924 * gimple-builder.h (create_gimple_tmp): Delete.
6926 2014-01-27 Christian Bruel <christian.bruel@st.com>
6928 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6931 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6933 * config/pa/pa.md (call): Generate indirect long calls to non-local
6934 functions when outputing 32-bit code.
6935 (call_value): Likewise except for special call to buggy powf function.
6937 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6938 portable runtime and PIC indirect calls.
6939 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6940 and PIC call sequences. Use ldo instead of blr to set return register
6941 in PIC call sequence.
6943 2014-01-25 Walter Lee <walt@tilera.com>
6945 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6946 avoid clobbering a live register.
6948 2014-01-25 Walter Lee <walt@tilera.com>
6950 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6951 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6952 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6953 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6955 2014-01-25 Walter Lee <walt@tilera.com>
6957 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6958 arguments on even registers.
6959 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6961 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6962 (BIGGEST_ALIGNMENT): Ditto.
6963 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6965 2014-01-25 Walter Lee <walt@tilera.com>
6967 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6968 insns before bundling.
6969 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
6971 2014-01-25 Walter Lee <walt@tilera.com>
6973 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
6974 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
6975 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
6977 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
6979 * config/mips/constraints.md (kl): Delete.
6980 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
6981 define expands, using...
6982 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
6983 instructions for MIPS16.
6984 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
6985 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
6987 2014-01-25 Walter Lee <walt@tilera.com>
6989 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
6993 2014-01-25 Walter Lee <walt@tilera.com>
6995 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
6996 (TARGET_EXPAND_TO_RTL_HOOK): Define.
6998 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7000 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7003 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7005 * print-rtl.c (in_call_function_usage): New var.
7006 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7007 EXPR_LIST mode as mode and not as reg note name.
7010 * cfgloopmanip.c (copy_loop_info): If
7011 loop->warned_aggressive_loop_optimizations, make sure
7012 the flag is set in target loop too.
7014 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7016 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7018 * builtins.def: Likewise.
7019 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7020 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7021 * ira.c (ira_setup_eliminable_regset): Likewise.
7022 * omp-low.c (gate_expand_omp): Likewise.
7023 (execute_lower_omp): Likewise.
7024 (diagnose_sb_0): Likewise.
7025 (gate_diagnose_omp_blocks): Likewise.
7026 (simd_clone_clauses_extract): Likewise.
7029 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7031 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7032 correction for little endian...
7033 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7036 2014-01-24 Jeff Law <law@redhat.com>
7038 PR tree-optimization/59919
7039 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7040 for non-returning calls.
7042 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7044 * common/config/aarch64/aarch64-common.c
7045 (aarch64_rewrite_mcpu): Handle multiple names.
7046 * config/aarch64/aarch64.h
7047 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7049 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7051 * input.c (add_file_to_cache_tab): Handle the case where fopen
7054 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7057 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7058 from push operand if code of push isn't PRE_DEC.
7060 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7063 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7064 -mquad-memory-atomic. Update -mquad-memory documentation to say
7065 it is only used for non-atomic loads/stores.
7067 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7068 -mquad-memory or -mquad-memory-atomic switches.
7070 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7071 -mquad-memory-atomic to ISA 2.07 support.
7073 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7074 to separate support of normal quad word memory operations (ldq, stq)
7075 from the atomic quad word memory operations.
7077 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7078 support to separate non-atomic quad word operations from atomic
7079 quad word operations. Disable non-atomic quad word operations in
7080 little endian mode so that we don't have to swap words after the
7081 load and before the store.
7082 (quad_load_store_p): Add comment about atomic quad word support.
7083 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7084 options printed with -mdebug=reg.
7086 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7087 -mquad-memory-atomic as the test for whether we have quad word
7088 atomic instructions.
7089 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7090 or -mp8-vector are used, allow byte/half-word atomic operations.
7092 * config/rs6000/sync.md (load_lockedti): Insure that the address
7093 is a proper indexed or indirect address for the lqarx instruction.
7094 On little endian systems, swap the hi/lo registers after the lqarx
7096 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7097 insure the address is valid for the lqarx instruction.
7098 (store_conditionalti): Insure that the address is a proper indexed
7099 or indirect address for the stqcrx. instruction. On little endian
7100 systems, swap the hi/lo registers before doing the stqcrx.
7102 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7103 insure the address is valid for the stqcrx. instruction.
7105 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7106 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7107 type of quad memory support is available.
7109 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7112 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7113 there is a danger of looping.
7115 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7117 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7118 force flag_ira_loop_pressure if set via command line.
7120 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7122 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7123 (ashr_simd): New builtin handling DI mode.
7124 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7125 (aarch64_sshr_simddi): New match pattern.
7126 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7127 (vshrd_n_s64): Likewise.
7128 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7130 2014-01-23 Nick Clifton <nickc@redhat.com>
7132 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7133 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7134 favour of mcu specific scripts.
7135 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7138 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7139 Alex Velenko <Alex.Velenko@arm.com>
7141 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7142 (vaddv_s16): Likewise.
7143 (vaddv_s32): Likewise.
7144 (vaddv_u8): Likewise.
7145 (vaddv_u16): Likewise.
7146 (vaddv_u32): Likewise.
7147 (vaddvq_s8): Likewise.
7148 (vaddvq_s16): Likewise.
7149 (vaddvq_s32): Likewise.
7150 (vaddvq_s64): Likewise.
7151 (vaddvq_u8): Likewise.
7152 (vaddvq_u16): Likewise.
7153 (vaddvq_u32): Likewise.
7154 (vaddvq_u64): Likewise.
7155 (vaddv_f32): Likewise.
7156 (vaddvq_f32): Likewise.
7157 (vaddvq_f64): Likewise.
7158 (vmaxv_f32): Likewise.
7159 (vmaxv_s8): Likewise.
7160 (vmaxv_s16): Likewise.
7161 (vmaxv_s32): Likewise.
7162 (vmaxv_u8): Likewise.
7163 (vmaxv_u16): Likewise.
7164 (vmaxv_u32): Likewise.
7165 (vmaxvq_f32): Likewise.
7166 (vmaxvq_f64): Likewise.
7167 (vmaxvq_s8): Likewise.
7168 (vmaxvq_s16): Likewise.
7169 (vmaxvq_s32): Likewise.
7170 (vmaxvq_u8): Likewise.
7171 (vmaxvq_u16): Likewise.
7172 (vmaxvq_u32): Likewise.
7173 (vmaxnmv_f32): Likewise.
7174 (vmaxnmvq_f32): Likewise.
7175 (vmaxnmvq_f64): Likewise.
7176 (vminv_f32): Likewise.
7177 (vminv_s8): Likewise.
7178 (vminv_s16): Likewise.
7179 (vminv_s32): Likewise.
7180 (vminv_u8): Likewise.
7181 (vminv_u16): Likewise.
7182 (vminv_u32): Likewise.
7183 (vminvq_f32): Likewise.
7184 (vminvq_f64): Likewise.
7185 (vminvq_s8): Likewise.
7186 (vminvq_s16): Likewise.
7187 (vminvq_s32): Likewise.
7188 (vminvq_u8): Likewise.
7189 (vminvq_u16): Likewise.
7190 (vminvq_u32): Likewise.
7191 (vminnmv_f32): Likewise.
7192 (vminnmvq_f32): Likewise.
7193 (vminnmvq_f64): Likewise.
7195 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7197 * config/aarch64/aarch64-simd.md
7198 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7199 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7200 (*aarch64_mul3_elt<mode>): Likewise.
7201 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7202 (*aarch64_mul3_elt_to_64v2df): Likewise.
7203 (*aarch64_mla_elt<mode>): Likewise.
7204 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7205 (*aarch64_mls_elt<mode>): Likewise.
7206 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7207 (*aarch64_fma4_elt<mode>): Likewise.
7208 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7209 (*aarch64_fma4_elt_to_64v2df): Likewise.
7210 (*aarch64_fnma4_elt<mode>): Likewise.
7211 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7212 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7213 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7214 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7215 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7216 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7217 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7218 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7219 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7221 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7223 * config/aarch64/aarch64-simd.md
7224 (aarch64_be_checked_get_lane<mode>): New define_expand.
7225 * config/aarch64/aarch64-simd-builtins.def
7226 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7227 New builtin definition.
7228 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7229 Use new safe be builtin.
7231 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7233 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7235 (aarch64_be_st1<mode>): Likewise.
7236 (aarch_ld1<VALL:mode>): Define_expand modified.
7237 (aarch_st1<VALL:mode>): Likewise.
7238 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7239 (UNSPEC_ST1): Likewise.
7241 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7243 * config/microblaze/microblaze.md: Add trap insn and attribute
7245 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7247 PR preprocessor/58580
7248 * input.h (location_get_source_line): Take an additional line_size
7250 (void diagnostics_file_cache_fini): Declare new function.
7251 * input.c (struct fcache): New type.
7252 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7253 New static constants.
7254 (diagnostic_file_cache_init, total_lines_num)
7255 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7256 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7257 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7258 (get_next_line, read_next_line, goto_next_line, read_line_num):
7259 New static function definitions.
7260 (diagnostic_file_cache_fini): New function.
7261 (location_get_source_line): Take an additional output line_len
7262 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7264 * diagnostic.c (diagnostic_finish): Call
7265 diagnostic_file_cache_fini.
7266 (adjust_line): Take an additional input parameter for the length
7267 of the line, rather than calculating it with strlen.
7268 (diagnostic_show_locus): Adjust the use of
7269 location_get_source_line and adjust_line with respect to their new
7270 signature. While displaying a line now, do not stop at the first
7271 null byte. Rather, display the zero byte as a space and keep
7272 going until we reach the size of the line.
7273 * Makefile.in: Add vec.o to OBJS-libcommon
7275 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7276 Ilya Tocar <ilya.tocar@intel.com>
7278 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7279 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7280 (__builtin_ia32_kmov16): Ditto.
7281 * config/i386/i386.md (UNSPEC_KMOV): New.
7284 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7286 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7287 (_mm512_storeu_si512): Ditto.
7289 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7292 * rtl.h (get_referenced_operands): Declare.
7293 * recog.c (get_referenced_operands): New function.
7294 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7295 operands have been referenced when recording LO_SUM references.
7297 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7299 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7301 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7303 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7304 Enable for generic and recent AMD targets.
7306 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7308 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7309 ARG_SIZE note when adjustment was eliminated.
7311 2014-01-22 Jeff Law <law@redhat.com>
7313 PR tree-optimization/59597
7314 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7315 in file. Accept new argument REGISTERING and use it to modify
7316 dump output appropriately.
7317 (register_jump_thread): Corresponding changes.
7318 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7319 thread paths involving joiner blocks. Add code to dump cancelled
7320 jump threading paths.
7322 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7324 PR rtl-optimization/59477
7325 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7326 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7328 2014-01-22 Tom Tromey <tromey@redhat.com>
7330 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7332 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7334 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7336 PR rtl-optimization/59896
7337 * lra-constraints.c (process_alt_operands): Check unused note for
7338 matched operands of insn with no output reloads.
7340 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7342 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7343 (mips_move_from_gpr_cost): Likewise.
7345 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7347 PR rtl-optimization/59858
7348 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7349 ira_class_hard_regs_num.
7350 (process_alt_operands): Increase reject for dying matched operand.
7352 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7355 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7356 smaller than size, perform several stores or loads and stores
7357 at dst + count - size to store or copy all of size bytes, rather
7358 than just last modesize bytes.
7360 2014-01-20 DJ Delorie <dj@redhat.com>
7362 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7363 that CLOBBERs are REGs before propogating their values.
7365 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7368 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7369 (cgraph_inline_failed_type): New function.
7370 * cgraph.h (DEFCIFCODE): Add type.
7371 (cgraph_inline_failed_type_t): New enum.
7372 (cgraph_inline_failed_type): New prototype.
7373 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7374 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7375 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7376 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7377 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7378 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7379 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7380 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7381 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7382 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7383 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7384 OPTIMIZATION_MISMATCH.
7385 * tree-inline.c (expand_call_inline): Emit errors during
7386 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7388 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7391 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7392 mode attribute in insn output.
7394 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7396 * output.h (output_constant): Delete.
7397 * varasm.c (output_constant): Make private.
7399 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7401 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7403 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7406 * tree.h (fold_builtin_strcat): New prototype.
7407 * builtins.c (fold_builtin_strcat): No longer static. Add len
7408 argument, if non-NULL, don't call c_strlen. Optimize
7409 directly into __builtin_memcpy instead of __builtin_strcpy.
7410 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7411 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7413 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7415 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7416 for SImode_address_operand operands, having only a REG argument.
7418 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7420 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7421 loader name using mbig-endian.
7422 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7424 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7426 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7430 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7432 * config/aarch64/aarch64-protos.h
7433 (aarch64_cannot_change_mode_class_ptr): Declare.
7434 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7435 aarch64_cannot_change_mode_class_ptr): New.
7436 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7437 backend hook aarch64_cannot_change_mode_class.
7439 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7441 * common/config/aarch64/aarch64-common.c
7442 (aarch64_handle_option): Don't handle any option order logic here.
7443 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7444 selected_cpu, warn on architecture version mismatch.
7445 (aarch64_override_options): Fix parsing order for option strings.
7447 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7448 Iain Sandoe <iain@codesourcery.com>
7451 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7452 warning. Amend comment to reflect current functionality.
7454 2014-01-20 Richard Biener <rguenther@suse.de>
7457 * builtins.c (fold_builtin_strcat): Remove case better handled
7458 by tree-ssa-strlen.c.
7460 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7462 * config/aarch64/aarch64.opt
7463 (mcpu, march, mtune): Make case-insensitive.
7465 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7468 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7469 if operands[1] is a REG or ZERO_EXTEND of a REG.
7471 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7473 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7475 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7477 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7478 long non-pic millicode calls.
7480 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7482 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7484 2014-01-19 Kito Cheng <kito@0xlab.org>
7486 * builtins.c (expand_movstr): Check movstr expand done or fail.
7488 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7489 H.J. Lu <hongjiu.lu@intel.com>
7492 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7493 to DImode for zero-extended addresses.
7495 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7497 PR rtl-optimization/57763
7498 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7499 on the new indirect jump_insn and increment LABEL_NUSES (label).
7501 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7505 * config.gcc (x86_archs): New variable.
7506 (x86_64_archs): Likewise.
7507 (x86_cpus): Likewise.
7508 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7509 --with-arch/--with-cpu= options.
7510 Support --with-arch=/--with-cpu={nehalem,westmere,
7511 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7513 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7515 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7516 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7518 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7520 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7522 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7525 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7526 clear cpp_get_options (parse_in)->warn_unused_macros for
7527 ix86_target_macros_internal with cpp_define.
7529 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7531 * jump.c (delete_related_insns): Keep (use (insn))s.
7532 * reorg.c (redundant_insn): Check for barriers too.
7534 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7536 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7538 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7540 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7541 call to $$dyncall when TARGET_LONG_CALLS is true.
7543 2014-01-17 Jeff Law <law@redhat.com>
7545 * ree.c (combine_set_extension): Temporarily disable test for
7546 changing number of hard registers.
7548 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7551 * ipa-inline-analysis.c (inline_free_summary):
7552 Do not free summary of aliases.
7554 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7557 * gimplify.c (gimplify_expr): Use create_tmp_var
7558 instead of create_tmp_var_raw. If cond doesn't have
7559 integral type, don't add the IFN_ANNOTATE builtin at all.
7561 2014-01-17 Martin Jambor <mjambor@suse.cz>
7564 * ipa-cp.c (prev_edge_clone): New variable.
7565 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7566 Also resize prev_edge_clone vector.
7567 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7568 (ipcp_edge_removal_hook): New function.
7569 (ipcp_driver): Register ipcp_edge_removal_hook.
7571 2014-01-17 Andrew Pinski <apinski@cavium.com>
7572 Steve Ellcey <sellcey@mips.com>
7575 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7578 2014-01-17 Jeff Law <law@redhat.com>
7581 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7582 so that pass_ccp runs first.
7584 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7586 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7587 (ix86_adjust_cost): Use !TARGET_XXX.
7588 (do_reorder_for_imul): Likewise.
7589 (swap_top_of_ready_list): Likewise.
7590 (ix86_sched_reorder): Likewise.
7592 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7594 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7595 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7596 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7597 (intel_memset): New. Duplicate slm_memset.
7598 (intel_cost): New. Duplicate slm_cost.
7599 (m_INTEL): New macro.
7600 (processor_target_table): Add "intel".
7601 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7602 with PROCESSOR_INTEL for "intel".
7603 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7604 PROCESSOR_SILVERMONT.
7605 (ix86_issue_rate): Likewise.
7606 (ix86_adjust_cost): Likewise.
7607 (ia32_multipass_dfa_lookahead): Likewise.
7608 (swap_top_of_ready_list): Likewise.
7609 (ix86_sched_reorder): Likewise.
7610 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7611 instead of TARGET_OPT_AGU.
7612 * config/i386/i386.h (TARGET_INTEL): New.
7613 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7614 (processor_type): Add PROCESSOR_INTEL.
7615 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7616 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7618 2014-01-17 Marek Polacek <polacek@redhat.com>
7621 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7624 2014-01-17 Richard Biener <rguenther@suse.de>
7626 PR tree-optimization/46590
7627 * opts.c (default_options_table): Add entries for
7628 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7629 all enabled at -O1 but not for -Og.
7630 * common.opt (fbranch-count-reg): Remove Init(1).
7631 (fmove-loop-invariants): Likewise.
7632 (ftree-pta): Likewise.
7634 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7636 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7637 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7638 decls to at least the GCC 4.8 used alignments.
7641 * tree-nested.c (convert_nonlocal_reference_stmt,
7642 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7643 of GIMPLE_BIND stmts, adjust associated decls.
7645 2014-01-17 Richard Biener <rguenther@suse.de>
7647 PR tree-optimization/46590
7648 * vec.h (vec<>::bseach): New member function implementing
7649 binary search according to C89 bsearch.
7650 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7651 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7652 bitmap pointer again. Make accesses_in_loop a flat array.
7653 (mem_ref_obstack): New global.
7654 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7655 (mark_ref_stored): Likewise.
7656 (ref_indep_loop_p_2): Likewise.
7657 (set_ref_stored_in_loop): New helper function.
7658 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7659 (memref_free): Adjust.
7660 (record_mem_ref_loc): Simplify.
7661 (gather_mem_refs_stmt): Adjust.
7662 (sort_locs_in_loop_postorder_cmp): New function.
7663 (analyze_memory_references): Sort accesses_in_loop after
7664 loop postorder number.
7665 (find_ref_loc_in_loop_cmp): New function.
7666 (for_all_locs_in_loop): Find relevant cluster of locs in
7667 accesses_in_loop and iterate without recursion.
7668 (execute_sm): Avoid uninit warning.
7669 (struct ref_always_accessed): Simplify.
7670 (ref_always_accessed::operator ()): Likewise.
7671 (ref_always_accessed_p): Likewise.
7672 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7673 loop postorder numbers here.
7674 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7677 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7680 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7681 on decls for which assemble_alias has been called.
7683 2014-01-17 Nick Clifton <nickc@redhat.com>
7685 * config/msp430/msp430.opt: (mcpu): New option.
7686 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7687 (msp430_option_override): Parse target_cpu. If the MCU name
7688 matches a generic string, clear target_mcu.
7689 (msp430_attr): Allow numeric interrupt values up to 63.
7690 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7691 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7693 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7695 * config/msp430/msp430.md (popm): Use %J rather than %I.
7696 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7697 (addhi_cy_i): Use immediate_operand for operand 2.
7698 * doc/invoke.texi: Document -mcpu option.
7700 2014-01-17 Richard Biener <rguenther@suse.de>
7702 PR rtl-optimization/38518
7703 * df.h (df_analyze_loop): Declare.
7704 * df-core.c: Include cfgloop.h.
7705 (df_analyze_1): Split out main part of df_analyze.
7706 (df_analyze): Adjust.
7707 (loop_inverted_post_order_compute): New function.
7708 (loop_post_order_compute): Likewise.
7709 (df_analyze_loop): New function avoiding whole-function
7711 * loop-invariant.c (find_defs): Use df_analyze_loop.
7712 (find_invariants): Adjust.
7713 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7715 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7717 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7718 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7720 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7722 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7723 traversal when removing references.
7725 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7728 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7730 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7733 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7734 pushing a reload for an autoinc when we had previously reloaded an
7735 inner part of the address.
7737 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7739 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7741 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7742 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7743 when not giving up or versioning for alias only because of
7745 (vect_analyze_data_ref_dependences): Set to true.
7746 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7748 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7749 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7753 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7756 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7757 operand 0 predicate for gathers, use a new pseudo as subtarget.
7759 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7762 * lra-constraints.c (process_alt_operands): Add printing debug info.
7763 Check absence of input/output reloads for matched operands too.
7765 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7767 PR rtl-optimization/59835
7768 * ira.c (ira_init_register_move_cost): Increase cost for
7771 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7773 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7775 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7778 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7779 non-register objects. Use gen_(high/low)part more consistently.
7782 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7785 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7786 endian support, remove tests for WORDS_BIG_ENDIAN.
7787 (p8_mfvsrd_3_<mode>): Likewise.
7788 (reload_gpr_from_vsx<mode>): Likewise.
7789 (reload_gpr_from_vsxsf): Likewise.
7790 (p8_mfvsrd_4_disf): Likewise.
7792 2014-01-16 Richard Biener <rguenther@suse.de>
7794 PR rtl-optimization/46590
7795 * lcm.c (compute_antinout_edge): Use postorder iteration.
7796 (compute_laterin): Use inverted postorder iteration.
7798 2014-01-16 Nick Clifton <nickc@redhat.com>
7801 * varasm.c (output_constant): Return the number of bytes actually
7803 (output_constructor_array_range): Update the field size with the
7804 number of bytes emitted by output_constant.
7805 (output_constructor_regular_field): Likewise. Also do not
7806 complain if the total number of bytes emitted is now greater
7807 than the expected fieldpos.
7808 * output.h (output_constant): Update prototype and descriptive comment.
7810 2014-01-16 Marek Polacek <polacek@redhat.com>
7813 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7814 it is error_mark_node.
7816 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7818 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7819 VALID_AVX256_REG_OR_OI_MODE.
7821 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7823 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7824 current procedure should be profiled.
7826 2014-01-15 Andrew Pinski <apinski@cavium.com>
7828 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7829 of moving from/to the STACK_REG register class.
7831 2014-01-15 Richard Henderson <rth@redhat.com>
7834 * reginfo.c (global_regs_decl): Globalize.
7835 * rtl.h (global_regs_decl): Declare.
7836 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7837 reserved via global_regs.
7839 2014-01-15 Teresa Johnson <tejohnson@google.com>
7841 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7843 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7845 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7846 and vmulosh rather than call gen_vec_widen_smult_*.
7847 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7848 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7849 (vec_widen_smult_even_v16qi): Likewise.
7850 (vec_widen_umult_even_v8hi): Likewise.
7851 (vec_widen_smult_even_v8hi): Likewise.
7852 (vec_widen_umult_odd_v16qi): Likewise.
7853 (vec_widen_smult_odd_v16qi): Likewise.
7854 (vec_widen_umult_odd_v8hi): Likewise.
7855 (vec_widen_smult_odd_v8hi): Likewise.
7856 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7857 vmuloub rather than call gen_vec_widen_umult_*.
7858 (vec_widen_umult_lo_v16qi): Likewise.
7859 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7860 vmulosb rather than call gen_vec_widen_smult_*.
7861 (vec_widen_smult_lo_v16qi): Likewise.
7862 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7863 rather than call gen_vec_widen_umult_*.
7864 (vec_widen_umult_lo_v8hi): Likewise.
7865 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7866 rather than call gen_vec_widen_smult_*.
7867 (vec_widen_smult_lo_v8hi): Likewise.
7869 2014-01-15 Jeff Law <law@redhat.com>
7871 PR tree-optimization/59747
7872 * ree.c (find_and_remove_re): Properly handle case where a second
7873 eliminated extension requires widening a copy created for elimination
7874 of a prior extension.
7875 (combine_set_extension): Ensure that the number of hard regs needed
7876 for a destination register does not change when we widen it.
7878 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7880 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7881 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7882 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7883 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7884 (avr-*-rtems*): Likewise.
7885 (bfin*-rtems*): Likewise.
7886 (moxie-*-rtems*): Likewise.
7887 (h8300-*-rtems*): Likewise.
7888 (i[34567]86-*-rtems*): Likewise.
7889 (lm32-*-rtems*): Likewise.
7890 (m32r-*-rtems*): Likewise.
7891 (m68k-*-rtems*): Likewise.
7892 (microblaze*-*-rtems*): Likewise.
7893 (mips*-*-rtems*): Likewise.
7894 (powerpc-*-rtems*): Likewise.
7895 (sh-*-rtems*): Likewise.
7896 (sparc-*-rtems*): Likewise.
7897 (sparc64-*-rtems*): Likewise.
7898 (v850-*-rtems*): Likewise.
7899 (m32c-*-rtems*): Likewise.
7901 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7903 PR rtl-optimization/59511
7904 * ira.c (ira_init_register_move_cost): Use memory costs for some
7905 cases of register move cost calculations.
7906 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7907 instead of BB frequency.
7908 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7909 * lra-assigns.c (find_hard_regno_for): Ditto.
7911 2014-01-15 Richard Biener <rguenther@suse.de>
7913 PR tree-optimization/59822
7914 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7915 (vectorizable_load): Use it to hoist defs of uses of invariant
7916 loads out of the loop.
7918 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7919 Kugan Vivekanandarajah <kuganv@linaro.org>
7922 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7925 2014-01-15 Richard Biener <rguenther@suse.de>
7927 PR rtl-optimization/59802
7928 * lcm.c (compute_available): Use inverted postorder to seed
7929 the initial worklist.
7931 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7934 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7935 ADDR_REGS for invalid symrefs in non-PIC code.
7937 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7940 * builtins.c (determine_block_size): Initialize *probable_max_size
7941 even if len_rtx is CONST_INT.
7943 2014-01-14 Andrew Pinski <apinski@cavium.com>
7945 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7946 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7947 (cortexa53_tunings): Likewise.
7948 (aarch64_sched_issue_rate): New function.
7949 (TARGET_SCHED_ISSUE_RATE): Define.
7951 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7953 * ira-costs.c (find_costs_and_classes): Add missed
7954 ira_init_register_move_cost_if_necessary.
7956 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7959 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7961 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7964 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7965 to indicate if type is used for function return value. Warn ABI
7966 change if the vector mode isn't available for function return value.
7967 (ix86_function_arg_advance): Pass false to type_natural_mode.
7968 (ix86_function_arg): Likewise.
7969 (ix86_gimplify_va_arg): Likewise.
7970 (function_arg_32): Don't warn ABI change.
7971 (ix86_function_value): Pass true to type_natural_mode.
7972 (ix86_return_in_memory): Likewise.
7973 (ix86_struct_value_rtx): Removed.
7974 (TARGET_STRUCT_VALUE_RTX): Likewise.
7976 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
7978 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
7979 converting a conditional jump into a conditional return.
7981 2014-01-14 Richard Biener <rguenther@suse.de>
7983 PR tree-optimization/58921
7984 PR tree-optimization/59006
7985 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
7986 hoisting invariant stmts.
7987 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
7988 invariant loads on the preheader edge if possible.
7990 2014-01-14 Joey Ye <joey.ye@arm.com>
7992 * doc/plugin.texi (Building GCC plugins): Update to C++.
7994 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
7996 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
7997 (_mm_rcp28_round_ss): Ditto.
7998 (_mm_rsqrt28_round_sd): Ditto.
7999 (_mm_rsqrt28_round_ss): Ditto.
8000 (_mm_rcp28_sd): Ditto.
8001 (_mm_rcp28_ss): Ditto.
8002 (_mm_rsqrt28_sd): Ditto.
8003 (_mm_rsqrt28_ss): Ditto.
8004 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8005 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8006 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8007 (IX86_BUILTIN_RCP28SD): Ditto.
8008 (IX86_BUILTIN_RCP28SS): Ditto.
8009 (IX86_BUILTIN_RSQRT28SD): Ditto.
8010 (IX86_BUILTIN_RSQRT28SS): Ditto.
8011 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8012 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8013 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8014 (ix86_expand_special_args_builtin): Expand new FTYPE.
8015 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8016 (srcp14<mode>): Make insn unary.
8017 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8018 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8019 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8020 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8021 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8022 Fix rounding: make it SAE only.
8023 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8025 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8027 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8028 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8029 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8030 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8031 (round_saeonly_mask_scalar_operand4): Ditto.
8032 (round_saeonly_mask_scalar_op3): Ditto.
8033 (round_saeonly_mask_scalar_op4): Ditto.
8035 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8037 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8038 Implement -maltivec=be for vec_insert and vec_extract.
8040 2014-01-10 DJ Delorie <dj@redhat.com>
8042 * config/msp430/msp430.md (call_internal): Don't allow memory
8043 references with SP as the base register.
8044 (call_value_internal): Likewise.
8045 * config/msp430/constraints.md (Yc): New. For memory references
8046 that don't use SP as a base register.
8048 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8049 "an integer without a # prefix"
8050 * config/msp430/msp430.md (epilogue_helper): Use it.
8052 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8055 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8056 AVX512F gather builtins.
8057 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8058 on gather decls with INTEGER_TYPE masktype.
8059 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8060 directly into the builtin rather than hoisting it before loop.
8062 PR tree-optimization/59387
8063 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8064 (scev_const_prop): If folded_casts and type has undefined overflow,
8065 use force_gimple_operand instead of force_gimple_operand_gsi and
8066 for each added stmt if it is assign with
8067 arith_code_with_undefined_signed_overflow, call
8068 rewrite_to_defined_overflow.
8069 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8070 gimple-fold.h instead.
8071 (arith_code_with_undefined_signed_overflow,
8072 rewrite_to_defined_overflow): Moved to ...
8073 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8074 rewrite_to_defined_overflow): ... here. No longer static.
8075 Include gimplify-me.h.
8076 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8077 rewrite_to_defined_overflow): New prototypes.
8079 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8081 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8083 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8085 * builtins.c (get_object_alignment_2): Minor tweak.
8086 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8088 2014-01-13 Christian Bruel <christian.bruel@st.com>
8090 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8091 optimized non constant lengths.
8093 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8096 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8097 load as __atomic_load_N if possible.
8099 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8101 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8103 (rs6000_expand_builtin): Adjust call.
8105 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8108 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8109 * config/rs6000/rs6000.c: Include target-globals.h.
8110 (rs6000_set_current_function): Instead of doing target_reinit
8111 unconditionally, use save_target_globals_default_opts and
8112 restore_target_globals.
8114 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8116 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8117 (rs6000_expand_builtin): Handle mffs and mtfsf.
8118 (rs6000_init_builtins): Define mffs and mtfsf.
8119 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8120 (rs6000_mffs): New pattern.
8121 (rs6000_mtfsf): New pattern.
8123 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8125 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8126 Start narrowing with START. Apply candidate-use pair
8127 and check overall cost in narrowing.
8128 (iv_ca_prune): Pass new argument.
8130 2014-01-10 Jeff Law <law@redhat.com>
8133 * ree.c (combine_reaching_defs): Ensure the defining statement
8134 occurs before the extension when optimizing extensions with
8135 different source and destination hard registers.
8137 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8140 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8141 vtables into the type inheritance graph.
8143 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8145 PR rtl-optimization/59754
8146 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8147 modes in the REGNO != REGNO case.
8149 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8151 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8153 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8155 PR tree-optimization/59745
8156 * tree-predcom.c (tree_predictive_commoning_loop): Call
8157 free_affine_expand_cache if giving up because components is NULL.
8159 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8160 GC in payload of target_globals struct instead of allocating them on
8161 the heap and the larger structs separately using GC.
8162 * target-globals.h (struct target_globals): Make regs, hard_regs,
8163 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8164 of GTY((skip)) and change type to void *.
8165 (reset_target_globals): Cast loads from those fields to corresponding
8168 2014-01-10 Steve Ellcey <sellcey@mips.com>
8171 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8172 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8173 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8175 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8178 * aarch64-modes.def (CC_Zmode): New flags mode.
8179 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8180 represents an equality.
8181 (aarch64_get_condition_code): Handle CC_Zmode.
8182 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8184 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8186 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8187 extraction in good case.
8189 2014-01-10 Richard Biener <rguenther@suse.de>
8191 PR tree-optimization/59374
8192 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8193 checking after SLP discovery. Mark stmts not participating
8194 in any SLP instance properly.
8196 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8198 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8199 when handling a SET rtx.
8201 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8203 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8204 (cortex-a57): Likewise.
8205 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8207 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8209 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8210 non-iwmmxt builtins.
8212 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8216 * ipa-devirt.c record_target_from_binfo): Take as argument
8217 stack of binfos and lookup matching one for virtual inheritance.
8218 (possible_polymorphic_call_targets_1): Update.
8220 2014-01-10 Huacai Chen <chenhc@lemote.com>
8222 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8223 kernel strings for Loongson-2E/2F/3A.
8225 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8228 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8229 is_gimple_call before calling gimple_call_internal_p.
8231 2014-01-09 Steve Ellcey <sellcey@mips.com>
8233 * Makefile.in (TREE_FLOW_H): Remove.
8234 (TREE_SSA_H): Add file names from tree-flow.h.
8235 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8236 * tree.h: Remove tree-flow.h reference.
8237 * hash-table.h: Remove tree-flow.h reference.
8238 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8239 reference with tree-ssa-loop.h.
8241 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8243 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8244 default element-order behavior for -maltivec.
8245 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8246 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8247 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8248 when targeting big endian, at least for now.
8249 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8251 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8254 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8255 var satisfies use_register_for_decl, just take into account type
8256 alignment, rather than decl alignment.
8258 PR tree-optimization/59622
8259 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8260 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8261 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8262 Don't devirtualize for inplace at all. For targets.length () == 1,
8263 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8265 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8267 * config/i386/i386.md (cpu): Remove the unused btver1.
8269 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8271 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8273 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8276 * tree-core.h (struct target_globals): New forward declaration.
8277 (struct tree_target_option): Add globals field.
8278 * tree.h (TREE_TARGET_GLOBALS): Define.
8279 (prepare_target_option_nodes_for_pch): New prototype.
8280 * target-globals.h (struct target_globals): Define even if
8282 * tree.c (prepare_target_option_node_for_pch,
8283 prepare_target_option_nodes_for_pch): New functions.
8284 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8285 * config/i386/i386.c: Include target-globals.h.
8286 (ix86_set_current_function): Instead of doing target_reinit
8287 unconditionally, use save_target_globals_default_opts and
8288 restore_target_globals.
8290 2014-01-09 Richard Biener <rguenther@suse.de>
8292 PR tree-optimization/59715
8293 * tree-cfg.h (split_critical_edges): Declare.
8294 * tree-cfg.c (split_critical_edges): Export.
8295 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8297 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8299 * cfgexpand.c (expand_stack_vars): Optionally disable
8300 asan stack protection.
8301 (expand_used_vars): Likewise.
8302 (partition_stack_vars): Likewise.
8303 * asan.c (asan_emit_stack_protection): Optionally disable
8304 after return stack usage.
8305 (instrument_derefs): Optionally disable memory access instrumentation.
8306 (instrument_builtin_call): Likewise.
8307 (instrument_strlen_call): Likewise.
8308 (asan_protect_global): Optionally disable global variables protection.
8309 * doc/invoke.texi: Added doc for new options.
8310 * params.def: Added new options.
8311 * params.h: Likewise.
8313 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8315 PR rtl-optimization/59724
8316 * ifcvt.c (cond_exec_process_if_block): Don't call
8317 flow_find_head_matching_sequence with 0 longest_match.
8318 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8319 non-active insns if !stop_after.
8320 (try_head_merge_bb): Revert 2014-01-07 changes.
8322 2014-01-08 Jeff Law <law@redhat.com>
8324 * ree.c (get_sub_rtx): New function, extracted from...
8325 (merge_def_and_ext): Here.
8326 (combine_reaching_defs): Use get_sub_rtx.
8328 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8330 * cgraph.h (varpool_variable_node): Do not choke on null node.
8332 2014-01-08 Catherine Moore <clm@codesourcery.com>
8334 * config/mips/mips.md (simple_return): Attempt to use JRC
8336 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8338 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8340 PR rtl-optimization/59137
8341 * reorg.c (steal_delay_list_from_target): Call update_block for
8343 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8345 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8347 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8348 two duplicate entries.
8350 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8353 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8355 * config/mips/mips.c (mips_truncated_op_cost): New function.
8356 (mips_rtx_costs): Adjust test for BADDU.
8357 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8359 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8361 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8362 (*baddu_si): ...this new pattern.
8364 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8367 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8369 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8372 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8374 (expand_expr, expand_normal): Adjust.
8375 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8376 inner_reference_p. Use inner_reference_p to expand inner references.
8377 (store_expr): Adjust.
8378 * cfgexpand.c (expand_call_stmt): Adjust.
8380 2014-01-08 Rong Xu <xur@google.com>
8382 * gcov-io.c (gcov_var): Move from gcov-io.h.
8383 (gcov_position): Ditto.
8384 (gcov_is_error): Ditto.
8385 (gcov_rewrite): Ditto.
8386 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8387 only part to libgcc/libgcov.h.
8389 2014-01-08 Marek Polacek <polacek@redhat.com>
8392 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8394 2014-01-08 Marek Polacek <polacek@redhat.com>
8397 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8399 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8401 PR rtl-optimization/59649
8402 * stor-layout.c (get_mode_bounds): For BImode return
8403 0 and STORE_FLAG_VALUE.
8405 2014-01-08 Richard Biener <rguenther@suse.de>
8408 * gimple.h (is_gimple_builtin_call): Remove.
8409 (gimple_builtin_call_types_compatible_p): New.
8410 (gimple_call_builtin_p): New overload.
8411 * gimple.c (is_gimple_builtin_call): Remove.
8412 (validate_call): Rename to ...
8413 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8415 (validate_type): New static function.
8416 (gimple_call_builtin_p): New overload and adjust.
8417 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8418 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8419 (gimple_fold_stmt_to_constant_1): Likewise.
8420 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8422 2014-01-08 Richard Biener <rguenther@suse.de>
8425 * gimplify.c (gimplify_expr): Gimplify register-register type
8426 VIEW_CONVERT_EXPRs to separate stmts.
8428 2014-01-07 Jeff Law <law@redhat.com>
8431 * ree.c (combine_set_extension): Handle case where source
8432 and destination registers in an extension insn are different.
8433 (combine_reaching_defs): Allow source and destination registers
8434 in extension to be different under limited circumstances.
8435 (add_removable_extension): Remove restriction that the
8436 source and destination registers in the extension are the same.
8437 (find_and_remove_re): Emit a copy from the extension's
8438 destination to its source after the defining insn if
8439 the source and destination registers are different.
8442 * ifcvt.c (merge_if_block): If we are merging a block with more than
8443 one successor with a block with no successors, remove any BARRIER
8444 after the second block.
8446 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8448 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8450 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8453 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8454 for 14-bit register offsets when INT14_OK_STRICT is false.
8456 2014-01-07 Roland Stigge <stigge@antcom.de>
8457 Michael Meissner <meissner@linux.vnet.ibm.com>
8460 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8461 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8463 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8465 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8468 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8470 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8471 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8472 rtx is const0_rtx or not.
8474 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8477 * target-globals.c (save_target_globals): Remove this_fn_optab
8479 * toplev.c: Include optabs.h.
8480 (target_reinit): Temporarily restore the global options if another
8481 set of options are in force.
8483 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8485 PR rtl-optimization/58668
8486 * cfgcleanup.c (flow_find_cross_jump): Don't count
8487 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8488 to determine what is counted.
8489 (flow_find_head_matching_sequence): Use active_insn_p to determine
8491 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8493 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8494 determine what is counted.
8496 PR tree-optimization/59643
8497 * tree-predcom.c (split_data_refs_to_components): If one dr is
8498 read and one write, determine_offset fails and the write isn't
8499 in the bad component, just put the read into the bad component.
8501 2014-01-07 Mike Stump <mikestump@comcast.net>
8502 Jakub Jelinek <jakub@redhat.com>
8505 * tree-core.h (struct tree_optimization_option): Change optabs
8506 type from unsigned char * to void *.
8507 * optabs.c (init_tree_optimization_optabs): Adjust
8508 TREE_OPTIMIZATION_OPTABS initialization.
8510 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8513 * config/i386/i386.h (struct machine_function): Add
8514 no_drap_save_restore field.
8515 * config/i386/i386.c (ix86_save_reg): Use
8516 !cfun->machine->no_drap_save_restore instead of
8517 crtl->stack_realign_needed.
8518 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8519 this function clears frame_pointer_needed. Set
8520 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8521 and DRAP reg is needed.
8523 2014-01-06 Marek Polacek <polacek@redhat.com>
8526 * doc/implement-c.texi: Mention that other integer types are
8527 permitted as bit-field types in strictly conforming mode.
8529 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8531 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8534 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8536 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8538 2014-01-06 Martin Jambor <mjambor@suse.cz>
8541 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8543 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8545 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8549 * var-tracking.c (add_stores): Preserve the value of the source even if
8550 we don't record the store.
8552 2014-01-06 Terry Guo <terry.guo@arm.com>
8554 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8556 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8559 * config/darwin.c (darwin_function_section): Adjust return values to
8560 correspond to optimisation changes made in r206070.
8562 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8564 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8565 from prefetch_block tune setting.
8566 (nocona_cost): Correct size of prefetch block to 64.
8568 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8570 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8571 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8572 used to save the static chain register in the computation of the offset
8573 from which the FP registers need to be restored.
8575 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8577 PR tree-optimization/59519
8578 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8579 ICE if get_current_def (current_new_name) is already non-NULL, as long
8580 as it is a phi result of some other phi in *new_exit_bb that has
8583 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8584 or vmovdqu* for misaligned_operand.
8585 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8586 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8587 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8588 aligned_mem for AVX512F masked aligned load and store builtins and for
8591 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8593 PR tree-optimization/59651
8594 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8595 Address range for negative step should be added by TYPE_SIZE_UNIT.
8597 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8599 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8600 overlapping registers also for registers other than the stack pointer.
8602 2014-01-03 Marek Polacek <polacek@redhat.com>
8605 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8608 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8611 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8614 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8615 (push splitter): Use <P:MODE_SIZE> instead of
8616 GET_MODE_SIZE (<P:MODE>mode).
8617 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8618 (mov -1, reg peephole2): Likewise.
8619 * config/i386/sse.md (*mov<mode>_internal,
8620 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8621 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8622 *<code><mode>3, *andnot<mode>3<mask_name>,
8623 <mask_codefor><code><mode>3<mask_name>): Likewise.
8624 * config/i386/subst.md (mask_mode512bit_condition,
8625 sd_mask_mode512bit_condition): Likewise.
8627 2014-01-02 Xinliang David Li <davidxl@google.com>
8629 PR tree-optimization/59303
8630 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8631 (dump_predicates): Better output format.
8632 (pred_equal_p): New function.
8633 (is_neq_relop_p): Ditto.
8634 (is_neq_zero_form_p): Ditto.
8635 (pred_expr_equal_p): Ditto.
8636 (pred_neg_p): Ditto.
8637 (simplify_pred): Ditto.
8638 (simplify_preds_2): Ditto.
8639 (simplify_preds_3): Ditto.
8640 (simplify_preds_4): Ditto.
8641 (simplify_preds): Ditto.
8643 (push_to_worklist): Ditto.
8644 (get_pred_info_from_cmp): Ditto.
8645 (is_degenerated_phi): Ditto.
8646 (normalize_one_pred_1): Ditto.
8647 (normalize_one_pred): Ditto.
8648 (normalize_one_pred_chain): Ditto.
8649 (normalize_preds): Ditto.
8650 (normalize_cond_1): Remove function.
8651 (normalize_cond): Ditto.
8652 (is_gcond_subset_of): Ditto.
8653 (is_subset_of_any): Ditto.
8654 (is_or_set_subset_of): Ditto.
8655 (is_and_set_subset_of): Ditto.
8656 (is_norm_cond_subset_of): Ditto.
8657 (pred_chain_length_cmp): Ditto.
8658 (convert_control_dep_chain_into_preds): Type change.
8659 (find_predicates): Ditto.
8660 (find_def_preds): Ditto.
8661 (destroy_predicates_vecs): Ditto.
8662 (find_matching_predicates_in_rest_chains): Ditto.
8663 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8664 (is_pred_expr_subset): Ditto.
8665 (is_pred_chain_subset_of): Ditto.
8666 (is_included_in): Ditto.
8667 (is_superset_of): Ditto.
8669 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8671 Update copyright years.
8673 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8675 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8676 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8677 config/arc/arc.md, config/arc/arc.opt,
8678 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8679 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8680 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8681 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8682 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8683 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8684 vtable-verify.c, vtable-verify.h: Use the standard form for the
8687 2014-01-02 Tobias Burnus <burnus@net-b.de>
8689 * gcc.c (process_command): Update copyright notice dates.
8690 * gcov-dump.c: Ditto.
8692 * doc/cpp.texi: Bump @copying's copyright year.
8693 * doc/cppinternals.texi: Ditto.
8694 * doc/gcc.texi: Ditto.
8695 * doc/gccint.texi: Ditto.
8696 * doc/gcov.texi: Ditto.
8697 * doc/install.texi: Ditto.
8698 * doc/invoke.texi: Ditto.
8700 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8702 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8704 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8706 * config/i386/sse.md (*mov<mode>_internal): Guard
8707 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8709 PR rtl-optimization/59647
8710 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8711 new_rtx into UNSIGNED_FLOAT rtxes.
8713 Copyright (C) 2014 Free Software Foundation, Inc.
8715 Copying and distribution of this file, with or without modification,
8716 are permitted in any medium without royalty provided the copyright
8717 notice and this notice are preserved.