1 2014-05-06 Richard Biener <rguenther@suse.de>
4 * bitmap.c (debug_bitmap): Dump to stderr, not stdout.
5 * tree-ssa-structalias.c (dump_solution_for_var): Likewise.
7 2014-05-05 Jan Hubicka <hubicka@ucw.cz>
10 * ipa-devirt.c (get_class_context): Allow POD to change to non-POD.
12 2014-05-05 Radovan Obradovic <robradovic@mips.com>
13 Tom de Vries <tom@codesourcery.com>
15 * target.def (call_fusage_contains_non_callee_clobbers): New DEFHOOKPOD.
16 * doc/tm.texi.in (@node Stack and Calling): Add Miscellaneous Register
18 (@node Miscellaneous Register Hooks): New node.
19 (@hook TARGET_CALL_FUSAGE_CONTAINS_NON_CALLEE_CLOBBERS): New hook.
20 * doc/tm.texi: Regenerate.
22 2014-05-05 Marek Polacek <polacek@redhat.com>
25 * opts.c (common_handle_option): Call error_at instead of warning_at.
27 2014-05-05 Richard Biener <rguenther@suse.de>
29 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
30 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
31 under the TODO_verify_il umbrella.
33 2014-05-05 Richard Biener <rguenther@suse.de>
35 * passes.c (execute_function_todo): Move TODO_verify_flow under
36 the TODO_verify_ul umbrella.
38 2014-05-05 Richard Biener <rguenther@suse.de>
41 * fold-const.c (fold_binary_loc): Consistently avoid
42 canonicalizing X & CST away from a CST that is the mask
45 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
47 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
48 int argument to enum machine_mode.
49 (picochip_class_max_nregs): Ditto.
50 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
51 (picochip_class_max_nregs): Ditto.
53 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
55 * target.def: Add new target hook.
56 * doc/tm.texi: Regenerate.
57 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
58 * targhooks.c (default_keep_leaf_when_profiled): New function.
60 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
61 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
63 2014-05-05 Bin Cheng <bin.cheng@arm.com>
65 PR tree-optimization/60363
66 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
67 (copy_phi_args): New parameters. Call get_value_locus_in_path.
68 (update_destination_phis): New parameter.
69 (create_edge_and_update_destination_phis): Ditto.
70 (ssa_fix_duplicate_block_edges): Pass new arguments.
71 (thread_single_edge): Ditto.
73 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
75 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
76 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
77 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
78 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
79 Use RS6000_BTM_HARD_FLOAT.
80 (BU_MISC_2): Likewise.
81 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
82 RS6000_BTM_HARD_FLOAT.
83 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
85 (rs6000_invalid_builtin): Add hard floating builtin support.
86 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
88 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
90 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
92 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
93 Add missing function* argument.
95 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
97 * lra-constraints.c (valid_address_p): Move earlier in file.
98 Add a constraint argument to the address_info version.
99 (satisfies_memory_constraint_p): New function.
100 (satisfies_address_constraint_p): Likewise.
101 (process_alt_operands, curr_insn_transform): Use them.
102 (process_address): Pass the constraint to valid_address_p when
103 checking address operands.
105 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
107 * config/mips/mips.c (mips_isa_rev): New variable.
108 (mips_set_architecture): Set it.
109 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
111 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
112 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
113 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
114 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
115 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
116 conditions in terms of mips_isa_rev.
117 (mips_isa_rev): Declare.
119 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
121 * config/sh/sh-mem.cc: Use tabs instead of spaces.
122 (prob_unlikely, prob_likely): Make variables const.
124 2014-05-03 Denis Chertykov <chertykov@gmail.com>
126 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
128 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
130 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
132 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
134 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
135 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
136 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
138 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
139 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
141 Replace usage of ROUND_REG with sh_round_reg.
142 Use CEIL instead of ROUND_ADVANCE.
144 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
147 * config/sh/sh.c: Include stdlib headers before everything else.
149 2014-05-02 Jakub Jelinek <jakub@redhat.com>
151 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
152 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
153 (gimplify_adjust_omp_clauses): Simd region is never
154 directly nested in combined parallel. Instead, for linear
155 with copyin/copyout, if in combined for simd loop, make decl
156 firstprivate/lastprivate on OMP_FOR.
157 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
158 expand_omp_for_static_chunk): When setting endvar, also set
159 fd->loop.v to the same value.
161 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
163 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
165 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
167 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
170 2014-05-02 Marek Polacek <polacek@redhat.com>
172 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
174 2014-02-26 Kito Cheng <kito@0xlab.org>
176 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
177 to a C expression marco.
178 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
179 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
180 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
181 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
182 HONOR_REG_ALLOC_ORDER.
183 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
185 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
187 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
189 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
191 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
193 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
195 * tree-if-conv.c (is_cond_scalar_reduction): New function.
196 (convert_scalar_cond_reduction): Likewise.
197 (predicate_scalar_phi): Add recognition and transformation
198 of simple conditioanl reduction to be vectorizable.
200 2014-05-01 Marek Polacek <polacek@redhat.com>
203 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
205 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
207 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
208 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
209 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
210 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
211 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
212 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
213 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
214 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
216 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
218 * config/arc/arc.opt (mlra): Move comment above option name
219 to avoid mis-parsing as language options.
221 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
223 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
224 * config/sol2.h: ... here.
225 * config/sol2-10.h: Remove.
227 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
228 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
229 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
230 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
231 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
232 * config/sol2.h: ... here.
233 (SECTION_NAME_FORMAT): Don't redefine.
234 (STARTFILE_ARCH32_SPEC): Rename to ...
235 (STARTFILE_ARCH_SPEC): ... this.
236 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
237 * config/sparc/sol2.h: ... here.
238 (SECTION_NAME_FORMAT): Don't undef.
239 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
240 (SUBTARGET_EXTRA_SPECS): Remove.
241 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
243 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
244 (MD_STARTFILE_PREFIX): Remove.
245 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
246 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
247 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
248 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
249 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
250 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
251 * config/i386/sol2.h: ... here.
252 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
253 * config/i386/sol2-bi.h: Remove.
254 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
255 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
257 * config/i386/t-sol2-64: Rename to ...
258 * config/i386/t-sol2: ... this.
259 * config/sparc/t-sol2-64: Rename to ...
260 * config/sparc/t-sol2: ... this.
262 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
263 sol2_tm_file_head, sol2_tm_file_tail.
264 Include ${cpu_type}/sol2.h before sol2.h.
266 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
267 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
268 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
269 Reflect i386/t-sol2-64 renaming.
270 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
271 Reflect sparc/t-sol2-64 renaming.
273 2014-04-30 Richard Biener <rguenther@suse.de>
275 * passes.c (execute_function_todo): Move TODO_verify_stmts
276 and TODO_verify_ssa under the TODO_verify_il umbrella.
277 * tree-ssa.h (verify_ssa): Adjust prototype.
278 * tree-ssa.c (verify_ssa): Add parameter to tell whether
279 we should verify SSA operands.
280 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
281 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
282 whether we should verify whether not throwing stmts have EH info.
283 * graphite-scop-detection.c (create_sese_edges): Adjust.
284 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
285 * tree-eh.c (lower_try_finally_switch): Do not add the
286 default case label twice.
288 2014-04-30 Marek Polacek <polacek@redhat.com>
290 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
291 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
292 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
293 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
295 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
297 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
298 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
299 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
300 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
301 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
302 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
303 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
304 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
306 2014-04-29 David Malcolm <dmalcolm@redhat.com>
308 * tree-cfg.c (dump_function_to_file): Dump the return type of
309 functions, in a line to itself before the function body, mimicking
310 the layout of a C function.
312 2014-04-29 Jakub Jelinek <jakub@redhat.com>
314 PR tree-optimization/60971
315 * tree-tailcall.c (process_assignment): Reject conversions which
318 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
320 * calls.c (initialize_argument_information): Always treat
321 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
322 (expand_call): Likewise.
323 (emit_library_call_calue_1): Likewise.
324 * expr.c (PUSH_ARGS_REVERSED): Do not define.
325 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
328 2014-04-29 Nick Clifton <nickc@redhat.com>
330 * config/msp430/msp430.md (umulsidi): Fix typo.
331 (mulhisi3): Enable even inside interrupt handlers.
332 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
333 bigger return address pushed in large mode.
335 2014-04-29 Nick Clifton <nickc@redhat.com>
337 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
338 (arc_init_reg_tables): Use a machine_mode enum to iterate over
340 * config/m32r/m32r.c (init_reg_tables): Likewise.
341 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
342 enum to hold the modes.
344 2014-04-29 Richard Biener <rguenther@suse.de>
346 * dominance.c (free_dominance_info): Add overload with
348 (dom_info_state): Likewise.
349 (dom_info_available_p): Likewise.
350 * basic-block.h (free_dominance_info, dom_info_state,
351 dom_info_available_p): Declare overloads.
352 * passes.c (execute_function_todo): Verify that verifiers
353 don't change dominator info state. Drop dominator info
354 for IPA pass invocations.
355 * cgraph.c (release_function_body): Restore asserts that
356 dominator information is released.
358 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
360 * doc/invoke.texi: Fix typo.
361 * tree-vrp.c: Fix typos.
362 * gimple.c (infer_nonnull_range): Reorder operands of an &&
365 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
367 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
369 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
371 * config/aarch64/aarch64-builtins.c
372 (aarch64_types_storestruct_lane_qualifiers): New.
373 (TYPES_STORESTRUCT_LANE): Likewise.
374 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
375 (st3_lane): Likewise.
376 (st4_lane): Likewise.
377 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
378 (vec_store_lanesci_lane<mode>): Likewise.
379 (vec_store_lanesxi_lane<mode>): Likewise.
380 (aarch64_st2_lane<VQ:mode>): Likewise.
381 (aarch64_st3_lane<VQ:mode>): Likewise.
382 (aarch64_st4_lane<VQ:mode>): Likewise.
383 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
384 * config/aarch64/arm_neon.h
385 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
386 use new macro arguments.
387 (__ST3_LANE_FUNC): Likewise.
388 (__ST4_LANE_FUNC): Likewise.
389 * config/aarch64/iterators.md (V_TWO_ELEM): New.
390 (V_THREE_ELEM): Likewise.
391 (V_FOUR_ELEM): Likewise.
393 2014-04-28 David Malcolm <dmalcolm@redhat.com>
395 * doc/gimple.texi: Replace the description of the now-defunct
396 union gimple_statement_d with a diagram showing the
397 gimple_statement_base class hierarchy and its relationships to
398 the GSS_ and GIMPLE_ enums.
400 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
402 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
403 * config/aarch64/aarch64.c
404 (aarch64_cannot_change_mode_class): Weaken conditions.
405 (aarch64_modes_tieable_p): New.
406 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
408 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
410 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
411 (loadsync_<mode>): Change mode.
412 (load_quadpti, store_quadpti): New.
413 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
414 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
416 2014-04-28 Martin Jambor <mjambor@suse.cz>
418 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
419 same alias type as the original statement.
420 (subreplacement_assignment_data): New type.
421 (handle_unscalarized_data_in_subtree): New type of parameter,
422 generate new memory accesses with same alias type as the original
424 (load_assign_lhs_subreplacements): Likewise.
425 (sra_modify_constructor_assign): Generate new memory accesses with
426 same alias type as the original statement.
428 2014-04-28 Richard Biener <rguenther@suse.de>
430 * tree-pass.h (TODO_verify_il): Define.
431 (TODO_verify_all): Complete properly.
432 * passes.c (execute_function_todo): Move existing loop-closed
433 SSA verification under TODO_verify_il.
434 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
435 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
436 Fix tree sharing issue.
438 2014-04-28 Richard Biener <rguenther@suse.de>
441 * builtins.def (DEF_C11_BUILTIN): Add.
442 (BUILT_IN_ALIGNED_ALLOC): Likewise.
443 * coretypes.h (enum function_class): Add function_c11_misc.
444 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
445 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
446 (call_may_clobber_ref_p_1): Likewise.
447 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
448 (mark_all_reaching_defs_necessary_1): Likewise.
449 (propagate_necessity): Likewise.
450 (eliminate_unnecessary_stmts): Likewise.
451 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
453 2014-04-28 Richard Biener <rguenther@suse.de>
455 * tree-vrp.c (vrp_var_may_overflow): Remove.
456 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
457 with overflow immediately bump to one before that value and
458 let iteration figure out overflow status.
460 2014-04-28 Richard Biener <rguenther@suse.de>
462 * configure.ac: Do valgrind header checks unconditionally.
463 Add --enable-valgrind-annotations.
464 * system.h: Guard valgrind header inclusion with
465 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
466 * alloc-pool.c (pool_alloc, pool_free): Use
467 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
468 to guard possibly dead code.
469 * config.in: Regenerated.
470 * configure: Likewise.
472 2014-04-28 Jeff Law <law@redhat.com>
474 PR tree-optimization/60902
475 * tree-ssa-threadedge.c
476 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
477 over real defs when invalidating outputs from statements that do not
478 produce useful outputs for threading.
480 2014-04-28 Richard Biener <rguenther@suse.de>
482 PR tree-optimization/60979
483 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
484 SCOPs that end in a block with a successor with abnormal
487 2014-04-28 Richard Biener <rguenther@suse.de>
489 * tree-pass.h (execute_pass_list): Adjust prototype.
490 * passes.c (pass_manager::execute_early_local_passes):
492 (do_per_function): Change callback signature, push all actual
493 work to the callbals.
494 (do_per_function_toporder): Likewise.
495 (execute_function_dump): Adjust.
496 (execute_function_todo): Likewise.
497 (clear_last_verified): Likewise.
498 (verify_curr_properties): Likewise.
499 (update_properties_after_pass): Likewise.
500 (execute_pass_list_1): Split out from ...
501 (execute_pass_list): ... here. Adjust.
502 (execute_ipa_pass_list): Likewise.
503 * cgraphunit.c (cgraph_add_new_function): Adjust.
504 (analyze_function): Likewise.
505 (expand_function): Likewise.
506 * cgraph.c (release_function_body): Free dominance info
507 here instead of asserting it was magically freed elsewhere.
509 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
511 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
512 * configure: Regenerate.
513 * config/sparc/sparc.opt (muser-mode): New option.
514 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
516 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
517 * doc/invoke.texi (SPARC options): Document -muser-mode.
519 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
521 * cselib.c (find_slot_memmode): Delete.
522 (cselib_hasher): Change compare_type to a struct.
523 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
525 (preserve_constants_and_equivs): Adjust for new compare_type.
526 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
527 (wrap_constant): Delete.
528 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
530 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
532 * doc/install.texi (Building with profile feedback): Remove
535 2014-04-26 Tom de Vries <tom@codesourcery.com>
537 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
540 2014-04-25 Cary Coutant <ccoutant@google.com>
543 * dwarf2out.c (should_move_die_to_comdat): A type definition
544 can contain a subprogram definition, but don't move it to a
546 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
547 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
549 (clone_tree_hash): Rename to...
550 (clone_tree_partial): ...this; change callers. Copy
551 DW_TAG_subprogram DIEs as declarations.
552 (copy_decls_walk): Don't copy children of a declaration into a
555 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
558 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
561 2014-04-25 Jiong Wang <jiong.wang@arm.com>
563 * config/arm/predicates.md (call_insn_operand): Add long_call check.
564 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
566 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
569 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
571 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
574 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
576 PR tree-optimization/60930
577 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
578 creating a multiply candidate by folding two constant
579 multiplicands when the result overflows.
581 2014-04-25 Jakub Jelinek <jakub@redhat.com>
583 PR tree-optimization/60960
584 * tree-vect-generic.c (expand_vector_operation): Only call
585 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
587 2014-04-25 Tom de Vries <tom@codesourcery.com>
589 * expr.c (clobber_reg_mode): New function.
590 * expr.h (clobber_reg): New function.
592 2014-04-25 Tom de Vries <tom@codesourcery.com>
594 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
597 2014-04-25 Radovan Obradovic <robradovic@mips.com>
598 Tom de Vries <tom@codesourcery.com>
600 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
602 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
603 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
604 new argument to find_all_hard_reg_sets call.
606 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
608 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
609 Use HOST_WIDE_INT_C for mask literal.
610 (aarch_rev16_shleft_mask_imm_p): Likewise.
612 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
615 * config/sparc/sparc.md (ashlsi3_extend): Delete.
617 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
619 PR preprocessor/56540
620 * config/i386/i386-c.c (ix86_target_macros): Define
621 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
623 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
625 * configure.ac (tga_func): Remove.
626 (LIB_TLS_SPEC): Remove.
627 * configure: Regenerate.
628 * config.in: Regenerate.
629 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
631 2014-04-25 Richard Biener <rguenther@suse.de>
634 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
635 call stmt use/clobber sets during stmt walk instead of
636 walking the possibly incomplete set of caller edges.
638 2014-04-25 Richard Biener <rguenther@suse.de>
641 * passes.c (apply_ipa_transforms): Inline into only caller ...
642 (execute_one_pass): ... here. Properly bring in function
643 bodies for nodes we want to apply IPA transforms to.
645 2014-04-24 Cong Hou <congh@google.com>
647 PR tree-optimization/60896
648 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
649 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
650 (vect_mark_pattern_stmts): Set the def type of all statements in
651 PATTERN_DEF_SEQ as vect_internal_def.
653 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
655 * doc/extend.texi (PowerPC Built-in Functions): Document new
656 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
658 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
660 * config/rs6000/predicates.md (const_0_to_3_operand): New
661 predicate to match 0..3 integer constants.
663 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
664 to support adding miscellaneous builtin functions.
665 (BU_DFP_MISC_2): Likewise.
666 (BU_P7_MISC_1): Likewise.
667 (BU_P7_MISC_2): Likewise.
668 (BU_P8V_MISC_3): Likewise.
669 (BU_MISC_1): Likewise.
670 (BU_MISC_2): Likewise.
671 (DIVWE): Add extended divide builtin functions.
679 (DXEX): Add decimal floating-point builtin functions.
691 (CDTBCD): Add new BCD builtin functions.
695 (BCDADD_LT): Likewise.
696 (BCDADD_EQ): Likewise.
697 (BCDADD_GT): Likewise.
698 (BCDADD_OV): Likewise.
700 (BCDSUB_LT): Likewise.
701 (BCDSUB_EQ): Likewise.
702 (BCDSUB_GT): Likewise.
703 (BCDSUB_OV): Likewise.
704 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
705 (UNPACK_TD): Likewise.
707 (UNPACK_TF): Likewise.
708 (UNPACK_TF_0): Likewise.
709 (UNPACK_TF_1): Likewise.
710 (PACK_V1TI): Likewise.
711 (UNPACK_V1TI): Likewise.
713 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
714 support for decimal floating point builtin functions.
715 (rs6000_expand_ternop_builtin): Add checks for the new builtin
716 functions that take constant arguments.
717 (rs6000_invalid_builtin): Add decimal floating point builtin
719 (rs6000_init_builtins): Setup long double, _Decimal64, and
720 _Decimal128 types for new builtin functions.
721 (builtin_function_type): Set the unsigned flags appropriately for
722 the new builtin functions.
723 (rs6000_opt_masks): Add support for decimal floating point builtin
726 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
727 floating point builtin functions.
728 (RS6000_BTM_COMMON): Likewise.
729 (RS6000_BTI_long_double): Likewise.
730 (RS6000_BTI_dfloat64): Likewise.
731 (RS6000_BTI_dfloat128): Likewise.
732 (long_double_type_internal_node): Likewise.
733 (dfloat64_type_internal_node): Likewise.
734 (dfloat128_type_internal_node): Likewise.
736 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
737 2.07 bcd arithmetic instructions.
738 (UNSPEC_BCDSUB): Likewise.
739 (UNSPEC_BCD_OVERFLOW): Likewise.
740 (UNSPEC_BCD_ADD_SUB): Likewise.
741 (bcd_add_sub): Likewise.
742 (BCD_TEST): Likewise.
743 (bcd<bcd_add_sub>): Likewise.
744 (bcd<bcd_add_sub>_test): Likewise.
745 (bcd<bcd_add_sub>_test2): Likewise.
746 (bcd<bcd_add_sub>_<code>): Likewise.
747 (peephole2 for combined bcd ops): Likewise.
749 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
750 decimal floating point builtin functions.
751 (UNSPEC_DENBCD): Likewise.
752 (UNSPEC_DXEX): Likewise.
753 (UNSPEC_DIEX): Likewise.
754 (UNSPEC_DSCLI): Likewise.
755 (UNSPEC_DSCRI): Likewise.
756 (D64_D128): Likewise.
757 (dfp_suffix): Likewise.
758 (dfp_ddedpd_<mode>): Likewise.
759 (dfp_denbcd_<mode>): Likewise.
760 (dfp_dxex_<mode>): Likewise.
761 (dfp_diex_<mode>): Likewise.
762 (dfp_dscli_<mode>): Likewise.
763 (dfp_dscri_<mode>): Likewise.
765 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
767 (UNSPEC_CDTBCD): Likewise.
768 (UNSPEC_CBCDTD): Likewise.
769 (UNSPEC_DIVE): Add support for new extended divide builtin
771 (UNSPEC_DIVEO): Likewise.
772 (UNSPEC_DIVEU): Likewise.
773 (UNSPEC_DIVEUO): Likewise.
774 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
775 pack/unpack 128-bit types.
776 (UNSPEC_PACK_128BIT): Likewise.
777 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
778 (udiv<mode>3): Use idiv_ldiv mode attribute.
779 (div<mode>3): Likewise.
780 (addg6s): Add new BCD builtin functions.
783 (UNSPEC_DIV_EXTEND): Add support for new extended divide
785 (div_extend): Likewise.
786 (div<div_extend>_<mode>"): Likewise.
787 (FP128_64): Add support for new builtin functions to pack/unpack
789 (unpack<mode>): Likewise.
790 (unpacktf_0): Likewise.
791 (unpacktf_1): Likewise.
792 (unpack<mode>_dm): Likewise.
793 (unpack<mode>_nodm): Likewise.
794 (pack<mode>): Likewise.
795 (unpackv1ti): Likewise.
796 (packv1ti): Likewise.
798 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
800 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
803 2014-04-24 Jakub Jelinek <jakub@redhat.com>
805 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
806 * gimplify.c (omp_is_private): Change last argument's type to int.
807 Only diagnose lastprivate if the simd argument is 1, only diagnose
808 linear if the simd argument is 2.
809 (gimplify_omp_for): Adjust omp_is_private callers. When adding
810 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
811 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
812 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
813 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
814 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
815 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
816 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
817 * tree-nested.c (convert_nonlocal_omp_clauses,
818 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
820 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
823 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
826 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
828 * flag-types.h (enum ivar_visibility): Add.
830 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
832 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
835 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
837 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
839 2014-04-24 Radovan Obradovic <robradovic@mips.com>
840 Tom de Vries <tom@codesourcery.com>
842 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
843 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
845 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
846 * emit-rtl.c (try_split): Same.
848 2014-04-24 Radovan Obradovic <robradovic@mips.com>
849 Tom de Vries <tom@codesourcery.com>
851 * common.opt (fuse-caller-save): New option.
853 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
855 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
858 2014-04-24 Richard Biener <rguenther@suse.de>
860 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
861 during TER and instead use the sepops interface for expanding
862 non-GIMPLE_SINGLE_RHS.
864 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
866 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
867 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
869 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
871 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
872 assembler 64-bit option.
873 * configure: Regenerate.
875 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
877 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
878 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
879 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
880 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
881 (TARGET_CRYPTO): Take TARGET_SIMD into account.
883 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
885 * config/aarch64/aarch64-builtins.c
886 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
887 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
888 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
889 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
891 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
892 (Vrevsuff): New mode attribute.
894 2014-04-24 Terry Guo <terry.guo@arm.com>
896 * config/arm/arm.h (machine_function): Define variable
897 after_arm_reorg here.
898 * config/arm/arm.c (after_arm_reorg): Remove the definition.
899 (arm_split_constant): Update the way to access variable
902 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
904 2014-04-23 Tom de Vries <tom@codesourcery.com>
906 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
908 2014-04-23 David Malcolm <dmalcolm@redhat.com>
910 * is-a.h: Update comments to reflect the following changes to the
911 "pointerness" of the API, making the template parameter match the
912 return type, allowing use of is-a.h with typedefs of pointers.
913 (is_a_helper::cast): Return a T rather then a pointer to a T, so
914 that the return type matches the parameter to the is_a_helper.
916 (dyn_cast): Likewise.
918 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
919 pointer from the is-a.h API.
921 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
922 (is_a_helper <cgraph_node *>::test): ...this, matching change to
924 (is_a_helper <varpool_node>::test): Likewise, convert to...
925 (is_a_helper <varpool_node *>::test): ...this.
927 (varpool_first_variable): Update for removal of implicit pointer
929 (varpool_next_variable): Likewise.
930 (varpool_first_static_initializer): Likewise.
931 (varpool_next_static_initializer): Likewise.
932 (varpool_first_defined_variable): Likewise.
933 (varpool_next_defined_variable): Likewise.
934 (cgraph_first_defined_function): Likewise.
935 (cgraph_next_defined_function): Likewise.
936 (cgraph_first_function): Likewise.
937 (cgraph_next_function): Likewise.
938 (cgraph_first_function_with_gimple_body): Likewise.
939 (cgraph_next_function_with_gimple_body): Likewise.
940 (cgraph_alias_target): Likewise.
941 (varpool_alias_target): Likewise.
942 (cgraph_function_or_thunk_node): Likewise.
943 (varpool_variable_node): Likewise.
944 (symtab_real_symbol_p): Likewise.
945 * cgraphunit.c (referred_to_p): Likewise.
946 (analyze_functions): Likewise.
947 (handle_alias_pairs): Likewise.
948 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
949 * gimple-ssa.h (gimple_vuse_op): Likewise.
950 (gimple_vdef_op): Likewise.
951 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
952 * gimple.c (gimple_build_asm_1): Likewise.
953 (gimple_build_try): Likewise.
954 (gimple_build_resx): Likewise.
955 (gimple_build_eh_dispatch): Likewise.
956 (gimple_build_omp_for): Likewise.
957 (gimple_omp_for_set_clauses): Likewise.
959 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
960 (is_a_helper <gimple_statement_asm *>::test): ...this.
961 (is_a_helper <gimple_statement_bind>::test): Convert to...
962 (is_a_helper <gimple_statement_bind *>::test): ...this.
963 (is_a_helper <gimple_statement_call>::test): Convert to...
964 (is_a_helper <gimple_statement_call *>::test): ...this.
965 (is_a_helper <gimple_statement_catch>::test): Convert to...
966 (is_a_helper <gimple_statement_catch *>::test): ...this.
967 (is_a_helper <gimple_statement_resx>::test): Convert to...
968 (is_a_helper <gimple_statement_resx *>::test): ...this.
969 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
970 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
971 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
972 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
973 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
974 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
975 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
976 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
977 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
978 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
979 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
980 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
981 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
982 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
983 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
984 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
985 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
986 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
987 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
988 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
989 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
990 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
991 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
992 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
993 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
994 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
995 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
996 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
997 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
998 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
999 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
1000 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
1001 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
1002 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
1003 (is_a_helper <gimple_statement_phi>::test): Convert to...
1004 (is_a_helper <gimple_statement_phi *>::test): ...this.
1005 (is_a_helper <gimple_statement_transaction>::test): Convert to...
1006 (is_a_helper <gimple_statement_transaction *>::test): ...this.
1007 (is_a_helper <gimple_statement_try>::test): Convert to...
1008 (is_a_helper <gimple_statement_try *>::test): ...this.
1009 (is_a_helper <gimple_statement_wce>::test): Convert to...
1010 (is_a_helper <gimple_statement_wce *>::test): ...this.
1011 (is_a_helper <const gimple_statement_asm>::test): Convert to...
1012 (is_a_helper <const gimple_statement_asm *>::test): ...this.
1013 (is_a_helper <const gimple_statement_bind>::test): Convert to...
1014 (is_a_helper <const gimple_statement_bind *>::test): ...this.
1015 (is_a_helper <const gimple_statement_call>::test): Convert to...
1016 (is_a_helper <const gimple_statement_call *>::test): ...this.
1017 (is_a_helper <const gimple_statement_catch>::test): Convert to...
1018 (is_a_helper <const gimple_statement_catch *>::test): ...this.
1019 (is_a_helper <const gimple_statement_resx>::test): Convert to...
1020 (is_a_helper <const gimple_statement_resx *>::test): ...this.
1021 (is_a_helper <const gimple_statement_eh_dispatch>::test):
1023 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
1025 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1027 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1028 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1030 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1032 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1034 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1036 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1038 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1039 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1041 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1042 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1044 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1045 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1046 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1047 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1049 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1050 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1052 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1053 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1055 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1056 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1058 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1059 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1061 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1062 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1064 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1065 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1067 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1068 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1069 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1070 (is_a_helper <const gimple_statement_transaction>::test): Convert
1072 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1073 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1075 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1076 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1077 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1078 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1080 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1082 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1084 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1086 (gimple_use_ops): Update for removal of implicit pointer from the
1088 (gimple_set_use_ops): Likewise.
1089 (gimple_vuse): Likewise.
1090 (gimple_vdef): Likewise.
1091 (gimple_vuse_ptr): Likewise.
1092 (gimple_vdef_ptr): Likewise.
1093 (gimple_set_vuse): Likewise.
1094 (gimple_set_vdef): Likewise.
1095 (gimple_omp_return_set_lhs): Likewise.
1096 (gimple_omp_return_lhs): Likewise.
1097 (gimple_omp_return_lhs_ptr): Likewise.
1098 (gimple_call_fntype): Likewise.
1099 (gimple_call_set_fntype): Likewise.
1100 (gimple_call_set_internal_fn): Likewise.
1101 (gimple_call_use_set): Likewise.
1102 (gimple_call_clobber_set): Likewise.
1103 (gimple_bind_vars): Likewise.
1104 (gimple_bind_set_vars): Likewise.
1105 (gimple_bind_body_ptr): Likewise.
1106 (gimple_bind_set_body): Likewise.
1107 (gimple_bind_add_stmt): Likewise.
1108 (gimple_bind_block): Likewise.
1109 (gimple_bind_set_block): Likewise.
1110 (gimple_asm_ninputs): Likewise.
1111 (gimple_asm_noutputs): Likewise.
1112 (gimple_asm_nclobbers): Likewise.
1113 (gimple_asm_nlabels): Likewise.
1114 (gimple_asm_input_op): Likewise.
1115 (gimple_asm_input_op_ptr): Likewise.
1116 (gimple_asm_output_op): Likewise.
1117 (gimple_asm_output_op_ptr): Likewise.
1118 (gimple_asm_set_output_op): Likewise.
1119 (gimple_asm_clobber_op): Likewise.
1120 (gimple_asm_set_clobber_op): Likewise.
1121 (gimple_asm_label_op): Likewise.
1122 (gimple_asm_set_label_op): Likewise.
1123 (gimple_asm_string): Likewise.
1124 (gimple_catch_types): Likewise.
1125 (gimple_catch_types_ptr): Likewise.
1126 (gimple_catch_handler_ptr): Likewise.
1127 (gimple_catch_set_types): Likewise.
1128 (gimple_catch_set_handler): Likewise.
1129 (gimple_eh_filter_types): Likewise.
1130 (gimple_eh_filter_types_ptr): Likewise.
1131 (gimple_eh_filter_failure_ptr): Likewise.
1132 (gimple_eh_filter_set_types): Likewise.
1133 (gimple_eh_filter_set_failure): Likewise.
1134 (gimple_eh_must_not_throw_fndecl): Likewise.
1135 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1136 (gimple_eh_else_n_body_ptr): Likewise.
1137 (gimple_eh_else_e_body_ptr): Likewise.
1138 (gimple_eh_else_set_n_body): Likewise.
1139 (gimple_eh_else_set_e_body): Likewise.
1140 (gimple_try_eval_ptr): Likewise.
1141 (gimple_try_cleanup_ptr): Likewise.
1142 (gimple_try_set_eval): Likewise.
1143 (gimple_try_set_cleanup): Likewise.
1144 (gimple_wce_cleanup_ptr): Likewise.
1145 (gimple_wce_set_cleanup): Likewise.
1146 (gimple_phi_capacity): Likewise.
1147 (gimple_phi_num_args): Likewise.
1148 (gimple_phi_result): Likewise.
1149 (gimple_phi_result_ptr): Likewise.
1150 (gimple_phi_set_result): Likewise.
1151 (gimple_phi_arg): Likewise.
1152 (gimple_phi_set_arg): Likewise.
1153 (gimple_resx_region): Likewise.
1154 (gimple_resx_set_region): Likewise.
1155 (gimple_eh_dispatch_region): Likewise.
1156 (gimple_eh_dispatch_set_region): Likewise.
1157 (gimple_omp_critical_name): Likewise.
1158 (gimple_omp_critical_name_ptr): Likewise.
1159 (gimple_omp_critical_set_name): Likewise.
1160 (gimple_omp_for_clauses): Likewise.
1161 (gimple_omp_for_clauses_ptr): Likewise.
1162 (gimple_omp_for_set_clauses): Likewise.
1163 (gimple_omp_for_collapse): Likewise.
1164 (gimple_omp_for_index): Likewise.
1165 (gimple_omp_for_index_ptr): Likewise.
1166 (gimple_omp_for_set_index): Likewise.
1167 (gimple_omp_for_initial): Likewise.
1168 (gimple_omp_for_initial_ptr): Likewise.
1169 (gimple_omp_for_set_initial): Likewise.
1170 (gimple_omp_for_final): Likewise.
1171 (gimple_omp_for_final_ptr): Likewise.
1172 (gimple_omp_for_set_final): Likewise.
1173 (gimple_omp_for_incr): Likewise.
1174 (gimple_omp_for_incr_ptr): Likewise.
1175 (gimple_omp_for_set_incr): Likewise.
1176 (gimple_omp_for_pre_body_ptr): Likewise.
1177 (gimple_omp_for_set_pre_body): Likewise.
1178 (gimple_omp_parallel_clauses): Likewise.
1179 (gimple_omp_parallel_clauses_ptr): Likewise.
1180 (gimple_omp_parallel_set_clauses): Likewise.
1181 (gimple_omp_parallel_child_fn): Likewise.
1182 (gimple_omp_parallel_child_fn_ptr): Likewise.
1183 (gimple_omp_parallel_set_child_fn): Likewise.
1184 (gimple_omp_parallel_data_arg): Likewise.
1185 (gimple_omp_parallel_data_arg_ptr): Likewise.
1186 (gimple_omp_parallel_set_data_arg): Likewise.
1187 (gimple_omp_task_clauses): Likewise.
1188 (gimple_omp_task_clauses_ptr): Likewise.
1189 (gimple_omp_task_set_clauses): Likewise.
1190 (gimple_omp_task_child_fn): Likewise.
1191 (gimple_omp_task_child_fn_ptr): Likewise.
1192 (gimple_omp_task_set_child_fn): Likewise.
1193 (gimple_omp_task_data_arg): Likewise.
1194 (gimple_omp_task_data_arg_ptr): Likewise.
1195 (gimple_omp_task_set_data_arg): Likewise.
1196 (gimple_omp_taskreg_clauses): Likewise.
1197 (gimple_omp_taskreg_clauses_ptr): Likewise.
1198 (gimple_omp_taskreg_set_clauses): Likewise.
1199 (gimple_omp_taskreg_child_fn): Likewise.
1200 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1201 (gimple_omp_taskreg_set_child_fn): Likewise.
1202 (gimple_omp_taskreg_data_arg): Likewise.
1203 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1204 (gimple_omp_taskreg_set_data_arg): Likewise.
1205 (gimple_omp_task_copy_fn): Likewise.
1206 (gimple_omp_task_copy_fn_ptr): Likewise.
1207 (gimple_omp_task_set_copy_fn): Likewise.
1208 (gimple_omp_task_arg_size): Likewise.
1209 (gimple_omp_task_arg_size_ptr): Likewise.
1210 (gimple_omp_task_set_arg_size): Likewise.
1211 (gimple_omp_task_arg_align): Likewise.
1212 (gimple_omp_task_arg_align_ptr): Likewise.
1213 (gimple_omp_task_set_arg_align): Likewise.
1214 (gimple_omp_single_clauses): Likewise.
1215 (gimple_omp_single_clauses_ptr): Likewise.
1216 (gimple_omp_single_set_clauses): Likewise.
1217 (gimple_omp_target_clauses): Likewise.
1218 (gimple_omp_target_clauses_ptr): Likewise.
1219 (gimple_omp_target_set_clauses): Likewise.
1220 (gimple_omp_target_child_fn): Likewise.
1221 (gimple_omp_target_child_fn_ptr): Likewise.
1222 (gimple_omp_target_set_child_fn): Likewise.
1223 (gimple_omp_target_data_arg): Likewise.
1224 (gimple_omp_target_data_arg_ptr): Likewise.
1225 (gimple_omp_target_set_data_arg): Likewise.
1226 (gimple_omp_teams_clauses): Likewise.
1227 (gimple_omp_teams_clauses_ptr): Likewise.
1228 (gimple_omp_teams_set_clauses): Likewise.
1229 (gimple_omp_sections_clauses): Likewise.
1230 (gimple_omp_sections_clauses_ptr): Likewise.
1231 (gimple_omp_sections_set_clauses): Likewise.
1232 (gimple_omp_sections_control): Likewise.
1233 (gimple_omp_sections_control_ptr): Likewise.
1234 (gimple_omp_sections_set_control): Likewise.
1235 (gimple_omp_for_set_cond): Likewise.
1236 (gimple_omp_for_cond): Likewise.
1237 (gimple_omp_atomic_store_set_val): Likewise.
1238 (gimple_omp_atomic_store_val): Likewise.
1239 (gimple_omp_atomic_store_val_ptr): Likewise.
1240 (gimple_omp_atomic_load_set_lhs): Likewise.
1241 (gimple_omp_atomic_load_lhs): Likewise.
1242 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1243 (gimple_omp_atomic_load_set_rhs): Likewise.
1244 (gimple_omp_atomic_load_rhs): Likewise.
1245 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1246 (gimple_omp_continue_control_def): Likewise.
1247 (gimple_omp_continue_control_def_ptr): Likewise.
1248 (gimple_omp_continue_set_control_def): Likewise.
1249 (gimple_omp_continue_control_use): Likewise.
1250 (gimple_omp_continue_control_use_ptr): Likewise.
1251 (gimple_omp_continue_set_control_use): Likewise.
1252 (gimple_transaction_body_ptr): Likewise.
1253 (gimple_transaction_label): Likewise.
1254 (gimple_transaction_label_ptr): Likewise.
1255 (gimple_transaction_set_body): Likewise.
1256 (gimple_transaction_set_label): Likewise.
1258 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1259 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1260 * ipa-ref.c (ipa_record_reference): Likewise.
1261 * ipa-reference.c (analyze_function): Likewise.
1262 (ipa_reference_write_optimization_summary): Likewise.
1263 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1264 (address_taken_from_non_vtable_p): Likewise.
1265 (comdat_can_be_unshared_p_1): Likewise.
1266 * lto-cgraph.c (lto_output_ref): Likewise.
1267 (add_references): Likewise.
1268 (compute_ltrans_boundary): Likewise.
1269 (output_symtab): Likewise.
1270 (input_ref): Likewise.
1271 (input_cgraph_1): Likewise.
1272 (output_cgraph_opt_summary): Likewise.
1273 * lto-streamer-out.c (lto_output): Likewise.
1274 (output_symbol_p): Likewise.
1275 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1276 (lsei_start_function_in_partition): Likewise.
1277 (lsei_next_variable_in_partition): Likewise.
1278 (lsei_start_variable_in_partition): Likewise.
1279 * symtab.c (insert_to_assembler_name_hash): Likewise.
1280 (unlink_from_assembler_name_hash): Likewise.
1281 (symtab_unregister_node): Likewise.
1282 (symtab_remove_node): Likewise.
1283 (dump_symtab_node): Likewise.
1284 (verify_symtab_base): Likewise.
1285 (verify_symtab_node): Likewise.
1286 (symtab_make_decl_local): Likewise.
1287 (symtab_alias_ultimate_target): Likewise.
1288 (symtab_resolve_alias): Likewise.
1289 (symtab_get_symbol_partitioning_class): Likewise.
1290 * tree-phinodes.c (allocate_phi_node): Likewise.
1291 (reserve_phi_args_for_new_edge): Likewise.
1292 (remove_phi_args): Likewise.
1293 * varpool.c (varpool_node_for_asm): Likewise.
1294 (varpool_remove_unreferenced_decls): Likewise.
1296 2014-04-23 Jeff Law <law@redhat.com>
1298 PR tree-optimization/60902
1299 * tree-ssa-threadedge.c
1300 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1301 invalidate outputs from statements that do not produce useful
1302 outputs for threading.
1304 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1306 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1307 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1308 machine descriptions for Stack Smashing Protector.
1310 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1312 * aarch64.md (<optab>_rol<mode>3): New pattern.
1313 (<optab>_rolsi3_uxtw): Likewise.
1314 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1316 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1318 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1319 (arm_cortex_a12_tune): Likewise.
1321 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1323 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1325 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1327 * config/arm/arm.md (arm_rev16si2): New pattern.
1328 (arm_rev16si2_alt): Likewise.
1329 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1331 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1333 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1334 (rev16<mode>2_alt): Likewise.
1335 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1336 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1337 (aarch_rev16_shleft_mask_imm_p): Likewise.
1338 (aarch_rev16_p_1): Likewise.
1339 (aarch_rev16_p): Likewise.
1340 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1341 (aarch_rev16_shright_mask_imm_p): Likewise.
1342 (aarch_rev16_shleft_mask_imm_p): Likewise.
1344 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1346 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1347 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1349 (cortex_a53_extra_costs): Likewise.
1350 (cortex_a57_extra_costs): Likewise.
1351 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1352 (cortexa7_extra_costs): Likewise.
1353 (cortexa8_extra_costs): Likewise.
1354 (cortexa12_extra_costs): Likewise.
1355 (cortexa15_extra_costs): Likewise.
1356 (v7m_extra_costs): Likewise.
1357 (arm_new_rtx_costs): Handle BSWAP.
1359 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1361 * config/arm/arm.c (cortexa8_extra_costs): New table.
1362 (arm_cortex_a8_tune): New tuning struct.
1363 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1365 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1367 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1369 2014-04-23 Richard Biener <rguenther@suse.de>
1371 * Makefile.in (OBJS): Remove loop-unswitch.o.
1372 * loop-unswitch.c: Delete.
1373 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1374 * passes.def (pass_rtl_unswitch): Likewise.
1375 * loop-init.c (gate_rtl_unswitch): Likewise.
1376 (rtl_unswitch): Likewise.
1377 (pass_data_rtl_unswitch): Likewise.
1378 (pass_rtl_unswitch): Likewise.
1379 (make_pass_rtl_unswitch): Likewise.
1380 * rtl.h (reversed_condition): Likewise.
1381 (compare_and_jump_seq): Likewise.
1382 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1384 * loop-unroll.c (compare_and_jump_seq): Likewise.
1386 2014-04-23 Richard Biener <rguenther@suse.de>
1388 PR tree-optimization/60903
1389 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1390 commented code block.
1391 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1392 loop flags to newly created BBs and edges.
1394 2014-04-23 Nick Clifton <nickc@redhat.com>
1396 * config/msp430/msp430.c (msp430_handle_option): Move function
1398 (msp430_option_override): Simplify mcu and mcpu option handling.
1399 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1400 support for -mhwmult command line option.
1401 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1402 -mhwmult command line option.
1403 (msp430_hwmult_enabled): Delete.
1404 (msp43o_output_labelref): Add support for -mhwmult command line
1406 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1407 (umulsidi3): Likewise.
1408 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1409 (mcpu, mlarge, msmall): Likewise.
1410 (mhwmult): New option.
1411 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1413 (msp430_is_f5_mcu): Remove prototype.
1414 (msp430_use_f5_series_hwmult): Add prototype.
1415 * config/msp430/msp430-opts.h: New file.
1416 * common/config/msp430: New directory.
1417 * common/config/msp430/msp430-common.c: New file.
1418 * config.gcc (msp430): Remove target_has_targetm_common.
1419 * doc/invoke.texi: Document -mhwmult command line option.
1421 2014-04-23 Nick Clifton <nickc@redhat.com>
1423 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1424 default-manifest.o if it can be found in the search path.
1425 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1427 2014-04-23 Terry Guo <terry.guo@arm.com>
1429 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1431 2014-04-23 Richard Biener <rguenther@suse.de>
1434 * tree-inline.c (declare_return_variable): Use mark_addressable.
1436 2014-04-23 Richard Biener <rguenther@suse.de>
1439 * loop-init.c (loop_optimizer_init): Make sure to apply
1440 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1442 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1445 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1447 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1448 if flag_sanitize_undefined_trap_on_error.
1449 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1450 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1451 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1452 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1453 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1454 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1455 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1456 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1457 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1458 * ubsan.c (ubsan_instrument_unreachable): Return
1459 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1460 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1461 if flag_sanitize_undefined_trap_on_error and
1462 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1463 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1464 instrument_bool_enum_load): Emit __builtin_trap () if
1465 flag_sanitize_undefined_trap_on_error and
1466 __builtin_handle_*_abort () if !flag_sanitize_recover.
1467 * doc/invoke.texi (-fsanitize-recover,
1468 -fsanitize-undefined-trap-on-error): Document.
1470 2014-04-22 Christian Bruel <christian.bruel@st.com>
1472 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1473 Force immediates to SImode.
1475 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1477 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1479 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1480 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1481 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1482 (nios2_fpu_insn): Add entry for round.
1483 (N2FPU_NO_ERRNO_P): Define.
1484 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1486 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1488 2014-04-22 Richard Henderson <rth@redhat.com>
1490 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1491 (add<GPI>3_compare0): Remove leading * from name.
1492 (add<GPI>3_carryin): Likewise.
1493 (sub<GPI>3_compare0): Likewise.
1494 (sub<GPI>3_carryin): Likewise.
1495 (<su_optab>mulditi3): New expander.
1496 (multi3): New expander.
1497 (madd<GPI>): Remove leading * from name.
1499 2014-04-22 Martin Jambor <mjambor@suse.cz>
1501 * cgraphclones.c (cgraph_function_versioning): Copy
1502 ipa_transforms_to_apply instead of asserting it is empty.
1504 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1507 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1508 on count_exp to get mode.
1510 2014-04-22 Andrew Pinski <apinski@cavium.com>
1512 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1513 Handle TLS for ILP32.
1514 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1515 (tlsie_small_<mode>): this and handle PTR.
1516 (tlsie_small_sidi): New pattern.
1517 (tlsle_small): Change to an expand to handle ILP32.
1518 (tlsle_small_<mode>): New pattern.
1519 (tlsdesc_small): Rename to ...
1520 (tlsdesc_small_<mode>): this and handle PTR.
1522 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1524 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1526 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1528 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1529 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1530 (aarch64_types_signed_poly_qualifiers): Likewise.
1531 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1532 (aarch64_types_poly_signed_qualifiers): Likewise.
1533 (TYPES_REINTERP_SS): Type macro added.
1534 (TYPES_REINTERP_SU): Likewise.
1535 (TYPES_REINTERP_SP): Likewise.
1536 (TYPES_REINTERP_US): Likewise.
1537 (TYPES_REINTERP_PS): Likewise.
1538 (aarch64_fold_builtin): New expression folding added.
1539 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1540 Declarations removed.
1541 (REINTERP_SS): Declarations added.
1542 (REINTERP_US): Likewise.
1543 (REINTERP_PS): Likewise.
1544 (REINTERP_SU): Likewise.
1545 (REINTERP_SP): Likewise.
1546 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1547 (vreinterpretq_p8_f64): Likewise.
1548 (vreinterpret_p16_f64): Likewise.
1549 (vreinterpretq_p16_f64): Likewise.
1550 (vreinterpret_f32_f64): Likewise.
1551 (vreinterpretq_f32_f64): Likewise.
1552 (vreinterpret_f64_f32): Likewise.
1553 (vreinterpret_f64_p8): Likewise.
1554 (vreinterpret_f64_p16): Likewise.
1555 (vreinterpret_f64_s8): Likewise.
1556 (vreinterpret_f64_s16): Likewise.
1557 (vreinterpret_f64_s32): Likewise.
1558 (vreinterpret_f64_s64): Likewise.
1559 (vreinterpret_f64_u8): Likewise.
1560 (vreinterpret_f64_u16): Likewise.
1561 (vreinterpret_f64_u32): Likewise.
1562 (vreinterpret_f64_u64): Likewise.
1563 (vreinterpretq_f64_f32): Likewise.
1564 (vreinterpretq_f64_p8): Likewise.
1565 (vreinterpretq_f64_p16): Likewise.
1566 (vreinterpretq_f64_s8): Likewise.
1567 (vreinterpretq_f64_s16): Likewise.
1568 (vreinterpretq_f64_s32): Likewise.
1569 (vreinterpretq_f64_s64): Likewise.
1570 (vreinterpretq_f64_u8): Likewise.
1571 (vreinterpretq_f64_u16): Likewise.
1572 (vreinterpretq_f64_u32): Likewise.
1573 (vreinterpretq_f64_u64): Likewise.
1574 (vreinterpret_s64_f64): Likewise.
1575 (vreinterpretq_s64_f64): Likewise.
1576 (vreinterpret_u64_f64): Likewise.
1577 (vreinterpretq_u64_f64): Likewise.
1578 (vreinterpret_s8_f64): Likewise.
1579 (vreinterpretq_s8_f64): Likewise.
1580 (vreinterpret_s16_f64): Likewise.
1581 (vreinterpretq_s16_f64): Likewise.
1582 (vreinterpret_s32_f64): Likewise.
1583 (vreinterpretq_s32_f64): Likewise.
1584 (vreinterpret_u8_f64): Likewise.
1585 (vreinterpretq_u8_f64): Likewise.
1586 (vreinterpret_u16_f64): Likewise.
1587 (vreinterpretq_u16_f64): Likewise.
1588 (vreinterpret_u32_f64): Likewise.
1589 (vreinterpretq_u32_f64): Likewise.
1591 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1593 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1594 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1595 (vreinterpret_p8_s8): Likewise.
1596 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1597 (vreinterpret_p8_s16): Likewise.
1598 (vreinterpret_p8_s32): Likewise.
1599 (vreinterpret_p8_s64): Likewise.
1600 (vreinterpret_p8_f32): Likewise.
1601 (vreinterpret_p8_u8): Likewise.
1602 (vreinterpret_p8_u16): Likewise.
1603 (vreinterpret_p8_u32): Likewise.
1604 (vreinterpret_p8_u64): Likewise.
1605 (vreinterpret_p8_p16): Likewise.
1606 (vreinterpretq_p8_s8): Likewise.
1607 (vreinterpretq_p8_s16): Likewise.
1608 (vreinterpretq_p8_s32): Likewise.
1609 (vreinterpretq_p8_s64): Likewise.
1610 (vreinterpretq_p8_f32): Likewise.
1611 (vreinterpretq_p8_u8): Likewise.
1612 (vreinterpretq_p8_u16): Likewise.
1613 (vreinterpretq_p8_u32): Likewise.
1614 (vreinterpretq_p8_u64): Likewise.
1615 (vreinterpretq_p8_p16): Likewise.
1616 (vreinterpret_p16_s8): Likewise.
1617 (vreinterpret_p16_s16): Likewise.
1618 (vreinterpret_p16_s32): Likewise.
1619 (vreinterpret_p16_s64): Likewise.
1620 (vreinterpret_p16_f32): Likewise.
1621 (vreinterpret_p16_u8): Likewise.
1622 (vreinterpret_p16_u16): Likewise.
1623 (vreinterpret_p16_u32): Likewise.
1624 (vreinterpret_p16_u64): Likewise.
1625 (vreinterpret_p16_p8): Likewise.
1626 (vreinterpretq_p16_s8): Likewise.
1627 (vreinterpretq_p16_s16): Likewise.
1628 (vreinterpretq_p16_s32): Likewise.
1629 (vreinterpretq_p16_s64): Likewise.
1630 (vreinterpretq_p16_f32): Likewise.
1631 (vreinterpretq_p16_u8): Likewise.
1632 (vreinterpretq_p16_u16): Likewise.
1633 (vreinterpretq_p16_u32): Likewise.
1634 (vreinterpretq_p16_u64): Likewise.
1635 (vreinterpretq_p16_p8): Likewise.
1636 (vreinterpret_f32_s8): Likewise.
1637 (vreinterpret_f32_s16): Likewise.
1638 (vreinterpret_f32_s32): Likewise.
1639 (vreinterpret_f32_s64): Likewise.
1640 (vreinterpret_f32_u8): Likewise.
1641 (vreinterpret_f32_u16): Likewise.
1642 (vreinterpret_f32_u32): Likewise.
1643 (vreinterpret_f32_u64): Likewise.
1644 (vreinterpret_f32_p8): Likewise.
1645 (vreinterpret_f32_p16): Likewise.
1646 (vreinterpretq_f32_s8): Likewise.
1647 (vreinterpretq_f32_s16): Likewise.
1648 (vreinterpretq_f32_s32): Likewise.
1649 (vreinterpretq_f32_s64): Likewise.
1650 (vreinterpretq_f32_u8): Likewise.
1651 (vreinterpretq_f32_u16): Likewise.
1652 (vreinterpretq_f32_u32): Likewise.
1653 (vreinterpretq_f32_u64): Likewise.
1654 (vreinterpretq_f32_p8): Likewise.
1655 (vreinterpretq_f32_p16): Likewise.
1656 (vreinterpret_s64_s8): Likewise.
1657 (vreinterpret_s64_s16): Likewise.
1658 (vreinterpret_s64_s32): Likewise.
1659 (vreinterpret_s64_f32): Likewise.
1660 (vreinterpret_s64_u8): Likewise.
1661 (vreinterpret_s64_u16): Likewise.
1662 (vreinterpret_s64_u32): Likewise.
1663 (vreinterpret_s64_u64): Likewise.
1664 (vreinterpret_s64_p8): Likewise.
1665 (vreinterpret_s64_p16): Likewise.
1666 (vreinterpretq_s64_s8): Likewise.
1667 (vreinterpretq_s64_s16): Likewise.
1668 (vreinterpretq_s64_s32): Likewise.
1669 (vreinterpretq_s64_f32): Likewise.
1670 (vreinterpretq_s64_u8): Likewise.
1671 (vreinterpretq_s64_u16): Likewise.
1672 (vreinterpretq_s64_u32): Likewise.
1673 (vreinterpretq_s64_u64): Likewise.
1674 (vreinterpretq_s64_p8): Likewise.
1675 (vreinterpretq_s64_p16): Likewise.
1676 (vreinterpret_u64_s8): Likewise.
1677 (vreinterpret_u64_s16): Likewise.
1678 (vreinterpret_u64_s32): Likewise.
1679 (vreinterpret_u64_s64): Likewise.
1680 (vreinterpret_u64_f32): Likewise.
1681 (vreinterpret_u64_u8): Likewise.
1682 (vreinterpret_u64_u16): Likewise.
1683 (vreinterpret_u64_u32): Likewise.
1684 (vreinterpret_u64_p8): Likewise.
1685 (vreinterpret_u64_p16): Likewise.
1686 (vreinterpretq_u64_s8): Likewise.
1687 (vreinterpretq_u64_s16): Likewise.
1688 (vreinterpretq_u64_s32): Likewise.
1689 (vreinterpretq_u64_s64): Likewise.
1690 (vreinterpretq_u64_f32): Likewise.
1691 (vreinterpretq_u64_u8): Likewise.
1692 (vreinterpretq_u64_u16): Likewise.
1693 (vreinterpretq_u64_u32): Likewise.
1694 (vreinterpretq_u64_p8): Likewise.
1695 (vreinterpretq_u64_p16): Likewise.
1696 (vreinterpret_s8_s16): Likewise.
1697 (vreinterpret_s8_s32): Likewise.
1698 (vreinterpret_s8_s64): Likewise.
1699 (vreinterpret_s8_f32): Likewise.
1700 (vreinterpret_s8_u8): Likewise.
1701 (vreinterpret_s8_u16): Likewise.
1702 (vreinterpret_s8_u32): Likewise.
1703 (vreinterpret_s8_u64): Likewise.
1704 (vreinterpret_s8_p8): Likewise.
1705 (vreinterpret_s8_p16): Likewise.
1706 (vreinterpretq_s8_s16): Likewise.
1707 (vreinterpretq_s8_s32): Likewise.
1708 (vreinterpretq_s8_s64): Likewise.
1709 (vreinterpretq_s8_f32): Likewise.
1710 (vreinterpretq_s8_u8): Likewise.
1711 (vreinterpretq_s8_u16): Likewise.
1712 (vreinterpretq_s8_u32): Likewise.
1713 (vreinterpretq_s8_u64): Likewise.
1714 (vreinterpretq_s8_p8): Likewise.
1715 (vreinterpretq_s8_p16): Likewise.
1716 (vreinterpret_s16_s8): Likewise.
1717 (vreinterpret_s16_s32): Likewise.
1718 (vreinterpret_s16_s64): Likewise.
1719 (vreinterpret_s16_f32): Likewise.
1720 (vreinterpret_s16_u8): Likewise.
1721 (vreinterpret_s16_u16): Likewise.
1722 (vreinterpret_s16_u32): Likewise.
1723 (vreinterpret_s16_u64): Likewise.
1724 (vreinterpret_s16_p8): Likewise.
1725 (vreinterpret_s16_p16): Likewise.
1726 (vreinterpretq_s16_s8): Likewise.
1727 (vreinterpretq_s16_s32): Likewise.
1728 (vreinterpretq_s16_s64): Likewise.
1729 (vreinterpretq_s16_f32): Likewise.
1730 (vreinterpretq_s16_u8): Likewise.
1731 (vreinterpretq_s16_u16): Likewise.
1732 (vreinterpretq_s16_u32): Likewise.
1733 (vreinterpretq_s16_u64): Likewise.
1734 (vreinterpretq_s16_p8): Likewise.
1735 (vreinterpretq_s16_p16): Likewise.
1736 (vreinterpret_s32_s8): Likewise.
1737 (vreinterpret_s32_s16): Likewise.
1738 (vreinterpret_s32_s64): Likewise.
1739 (vreinterpret_s32_f32): Likewise.
1740 (vreinterpret_s32_u8): Likewise.
1741 (vreinterpret_s32_u16): Likewise.
1742 (vreinterpret_s32_u32): Likewise.
1743 (vreinterpret_s32_u64): Likewise.
1744 (vreinterpret_s32_p8): Likewise.
1745 (vreinterpret_s32_p16): Likewise.
1746 (vreinterpretq_s32_s8): Likewise.
1747 (vreinterpretq_s32_s16): Likewise.
1748 (vreinterpretq_s32_s64): Likewise.
1749 (vreinterpretq_s32_f32): Likewise.
1750 (vreinterpretq_s32_u8): Likewise.
1751 (vreinterpretq_s32_u16): Likewise.
1752 (vreinterpretq_s32_u32): Likewise.
1753 (vreinterpretq_s32_u64): Likewise.
1754 (vreinterpretq_s32_p8): Likewise.
1755 (vreinterpretq_s32_p16): Likewise.
1756 (vreinterpret_u8_s8): Likewise.
1757 (vreinterpret_u8_s16): Likewise.
1758 (vreinterpret_u8_s32): Likewise.
1759 (vreinterpret_u8_s64): Likewise.
1760 (vreinterpret_u8_f32): Likewise.
1761 (vreinterpret_u8_u16): Likewise.
1762 (vreinterpret_u8_u32): Likewise.
1763 (vreinterpret_u8_u64): Likewise.
1764 (vreinterpret_u8_p8): Likewise.
1765 (vreinterpret_u8_p16): Likewise.
1766 (vreinterpretq_u8_s8): Likewise.
1767 (vreinterpretq_u8_s16): Likewise.
1768 (vreinterpretq_u8_s32): Likewise.
1769 (vreinterpretq_u8_s64): Likewise.
1770 (vreinterpretq_u8_f32): Likewise.
1771 (vreinterpretq_u8_u16): Likewise.
1772 (vreinterpretq_u8_u32): Likewise.
1773 (vreinterpretq_u8_u64): Likewise.
1774 (vreinterpretq_u8_p8): Likewise.
1775 (vreinterpretq_u8_p16): Likewise.
1776 (vreinterpret_u16_s8): Likewise.
1777 (vreinterpret_u16_s16): Likewise.
1778 (vreinterpret_u16_s32): Likewise.
1779 (vreinterpret_u16_s64): Likewise.
1780 (vreinterpret_u16_f32): Likewise.
1781 (vreinterpret_u16_u8): Likewise.
1782 (vreinterpret_u16_u32): Likewise.
1783 (vreinterpret_u16_u64): Likewise.
1784 (vreinterpret_u16_p8): Likewise.
1785 (vreinterpret_u16_p16): Likewise.
1786 (vreinterpretq_u16_s8): Likewise.
1787 (vreinterpretq_u16_s16): Likewise.
1788 (vreinterpretq_u16_s32): Likewise.
1789 (vreinterpretq_u16_s64): Likewise.
1790 (vreinterpretq_u16_f32): Likewise.
1791 (vreinterpretq_u16_u8): Likewise.
1792 (vreinterpretq_u16_u32): Likewise.
1793 (vreinterpretq_u16_u64): Likewise.
1794 (vreinterpretq_u16_p8): Likewise.
1795 (vreinterpretq_u16_p16): Likewise.
1796 (vreinterpret_u32_s8): Likewise.
1797 (vreinterpret_u32_s16): Likewise.
1798 (vreinterpret_u32_s32): Likewise.
1799 (vreinterpret_u32_s64): Likewise.
1800 (vreinterpret_u32_f32): Likewise.
1801 (vreinterpret_u32_u8): Likewise.
1802 (vreinterpret_u32_u16): Likewise.
1803 (vreinterpret_u32_u64): Likewise.
1804 (vreinterpret_u32_p8): Likewise.
1805 (vreinterpret_u32_p16): Likewise.
1806 (vreinterpretq_u32_s8): Likewise.
1807 (vreinterpretq_u32_s16): Likewise.
1808 (vreinterpretq_u32_s32): Likewise.
1809 (vreinterpretq_u32_s64): Likewise.
1810 (vreinterpretq_u32_f32): Likewise.
1811 (vreinterpretq_u32_u8): Likewise.
1812 (vreinterpretq_u32_u16): Likewise.
1813 (vreinterpretq_u32_u64): Likewise.
1814 (vreinterpretq_u32_p8): Likewise.
1815 (vreinterpretq_u32_p16): Likewise.
1817 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1819 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1821 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1824 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1825 (vqnegd_s64): Likewise.
1826 (vqabs_s64): Likewise.
1827 (vqabsd_s64): Likewise.
1829 2014-04-22 Richard Henderson <rth@redhat.com>
1831 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1832 computation to the top of the loop.
1834 2014-04-22 Renlin <renlin.li@arm.com>
1835 Jiong Wang <jiong.wang@arm.com>
1837 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1838 * config/aarch64/aarch64.c (aarch64_layout_frame)
1839 (aarch64_initial_elimination_offset): Likewise.
1841 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1843 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1846 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1848 * machmode.h (bitwise_mode_for_mode): Declare.
1849 * stor-layout.h (bitwise_type_for_mode): Likewise.
1850 * stor-layout.c (bitwise_mode_for_mode): New function.
1851 (bitwise_type_for_mode): Likewise.
1852 * builtins.c (fold_builtin_memory_op): Use it instead of
1853 int_mode_for_mode and build_nonstandard_integer_type.
1855 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1857 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1858 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1859 (*-*-solaris2*): Simplify.
1860 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1861 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1862 *-*-solaris2.9* handling.
1864 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1866 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1867 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1869 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1870 * configure: Regenerate.
1872 * config/i386/sol2-9.h: Remove.
1874 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1875 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1876 Remove Solaris 9 references.
1878 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1880 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1881 (floatuns<GPI:mode><GPF:mode>2): Remove.
1882 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1883 and floatuns conversions.
1884 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1885 and floatuns conversions.
1886 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1887 (w1,w2): New mode attributes for inequal width conversions.
1889 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1891 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1892 the output asm format.
1894 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1896 * config/aarch64/aarch64-simd.md
1897 (aarch64_cm<optab>di): Always split.
1898 (*aarch64_cm<optab>di): New.
1899 (aarch64_cmtstdi): Always split.
1900 (*aarch64_cmtstdi): New.
1902 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1904 PR tree-optimization/60823
1905 * omp-low.c (ipa_simd_modify_function_body): Go through
1906 all SSA_NAMEs and for those refering to vector arguments
1907 which are going to be replaced adjust SSA_NAME_VAR and,
1908 if it is a default definition, change it into a non-default
1909 definition assigned at the beginning of function from new_decl.
1910 (ipa_simd_modify_stmt_ops): Rewritten.
1911 * tree-dfa.c (set_ssa_default_def): When removing default def,
1912 check for NULL loc instead of NULL *loc.
1914 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1916 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1917 restrictions on core registers for DImode values in Thumb2.
1919 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1921 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1922 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1924 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1926 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1927 (*iordi_notzesidi_di): Likewise.
1928 (*iordi_notsesidi_di): Likewise.
1930 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1932 * config/arm/arm-protos.h (tune_params): New struct members.
1933 * config/arm/arm.c: Initialise tune_params per processor.
1934 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1935 for speed, based on new tune_params.
1937 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1939 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1941 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1943 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1945 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1946 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1947 (vrnda_f64): Likewise.
1948 (vrndi_f64): Likewise.
1949 (vrndm_f64): Likewise.
1950 (vrndn_f64): Likewise.
1951 (vrndp_f64): Likewise.
1952 (vrndx_f64): Likewise.
1954 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1956 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1957 GET_MODE_SIZE argument is enum machine_mode.
1959 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1962 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1963 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1965 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1968 * asan.c (asan_emit_stack_protection): Force the base to align to
1969 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1970 appropriate bits if STRICT_ALIGNMENT.
1971 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1973 (expand_used_vars): Leave a space in the stack frame for alignment
1974 if STRICT_ALIGNMENT.
1976 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1978 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1980 (gimple_store_p): Likewise.
1981 (gimple_assign_load_p): Likewise.
1982 (gimple_assign_cast_p): Likewise.
1983 (gimple_clobber_p): Likewise.
1985 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1986 rather than a gimple.
1987 (gimple_assign_cast_p): Likewise.
1989 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1992 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1993 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1995 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1996 more debug information for E500 if -mdebug=reg.
1998 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
2001 * config/i386/i386.c (ix86_expand_builtin)
2002 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
2003 register for target RTX.
2004 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
2006 2014-04-18 Cong Hou <congh@google.com>
2008 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
2009 the widen-mult pattern by handling two operands with different sizes,
2010 and operands whose size is smaller than half of the result type.
2012 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2014 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
2015 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
2016 (do_estimate_edge_time): Compute it.
2017 * ipa-inline.c (want_inline_small_function_p): Bypass
2018 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
2020 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
2022 * ipa-inline.c (spec_rem): New static variable.
2023 (dump_overall_stats): New function.
2024 (dump_inline_stats): New function.
2026 2014-04-18 Richard Henderson <rth@redhat.com>
2028 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2029 to GET_MODE_SIZE, not a reg_class_t.
2031 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2033 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2034 (vsx_xxmrglw_<mode>): Likewise.
2036 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2039 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2040 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2041 (rs6000_init_hard_regno_mode_ok): Likewise.
2043 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2045 * ipa-inline.c (inline_small_functions): Account only non-cold
2047 * doc/invoke.texi (inline-unit-growth): Update documentation.
2049 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2051 * config/rs6000/rs6000.md (addti3, subti3): New.
2053 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2056 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2057 comment. Check optimize_insn_for_size_p instead of
2058 optimize_insn_for_speed_p.
2060 2014-04-17 Martin Jambor <mjambor@suse.cz>
2062 * gimple-iterator.c (gsi_start_edge): New function.
2063 * gimple-iterator.h (gsi_start_edge): Declare.
2064 * tree-sra.c (single_non_eh_succ): New function.
2065 (disqualify_ops_if_throwing_stmt): Renamed to
2066 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2067 having one non-EH successor BB.
2068 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2069 generate loads into replacements.
2070 (sra_modify_assign): Likewise and and also use the simple path for
2072 (sra_modify_function_body): Commit statements on edges.
2074 2014-04-17 Richard Biener <rguenther@suse.de>
2077 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2078 comparison results and add clarifying comment.
2080 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2082 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2083 (blank_mode): Initialize it.
2084 (emit_mode_size_inline, emit_mode_nunits_inline,
2085 emit_mode_inner_inline): New functions.
2086 (emit_insn_modes_h): Call them and surround their output with
2087 #if GCC_VERSION >= 4001 ... #endif.
2088 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2089 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2090 mode_* arrays if the argument is __builtin_constant_p.
2091 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2092 is enum machine_mode.
2094 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2096 * passes.c (opt_pass::execute): Adjust.
2097 (pass_manager::execute_pass_mode_switching): Likewise.
2098 (early_local_passes::execute): Likewise.
2099 (execute_one_pass): Pass cfun to the pass's execute method.
2100 * tree-pass.h (opt_pass::execute): Add function * argument.
2101 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2102 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2103 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2104 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2105 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2106 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2107 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2108 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2109 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2110 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
2111 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
2112 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
2113 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
2114 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
2115 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2116 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2117 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2118 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
2119 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
2120 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2121 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2122 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2123 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2124 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2125 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2126 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2127 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2128 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2129 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2132 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2134 * passes.c (opt_pass::gate): Take function * argument.
2135 (gate_all_early_local_passes): Merge into
2136 (early_local_passes::gate): this.
2137 (gate_all_early_optimizations): Merge into
2138 (all_early_optimizations::gate): this.
2139 (gate_all_optimizations): Mege into
2140 (all_optimizations::gate): this.
2141 (gate_all_optimizations_g): Merge into
2142 (all_optimizations_g::gate): this.
2143 (gate_rest_of_compilation): Mege into
2144 (rest_of_compilation::gate): this.
2145 (gate_postreload): Merge into
2146 (postreload::gate): this.
2147 (dump_one_pass): Pass cfun to the pass's gate method.
2148 (execute_ipa_summary_passes): Likewise.
2149 (execute_one_pass): Likewise.
2150 (ipa_write_summaries_2): Likewise.
2151 (ipa_write_optimization_summaries_1): Likewise.
2152 (ipa_read_summaries_1): Likewise.
2153 (ipa_read_optimization_summaries_1): Likewise.
2154 (execute_ipa_stmt_fixups): Likewise.
2155 * tree-pass.h (opt_pass::gate): Add function * argument.
2156 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2157 combine-stack-adj.c, combine.c, compare-elim.c,
2158 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2159 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2160 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2161 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2162 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2163 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2164 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2165 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2166 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2167 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2168 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2169 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2170 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2171 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2172 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2173 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2174 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2175 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2176 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2177 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2178 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2179 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2180 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2181 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2182 var-tracking.c, vtable-verify.c, web.c: Adjust.
2184 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2186 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2187 * configure: Regenerate.
2189 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2191 * passes.c (dump_one_pass): don't check pass->has_gate.
2192 (execute_ipa_summary_passes): Likewise.
2193 (execute_one_pass): Likewise.
2194 (ipa_write_summaries_2): Likewise.
2195 (ipa_write_optimization_summaries_1): Likewise.
2196 (ipa_read_optimization_summaries_1): Likewise.
2197 (execute_ipa_stmt_fixups): Likewise.
2198 * tree-pass.h (pass_data::has_gate): Remove.
2199 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2200 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2201 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2202 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2203 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2204 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2205 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2206 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2207 gimple-low.c, gimple-ssa-isolate-paths.c,
2208 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2209 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2210 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2211 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2212 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2213 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2214 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2215 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2216 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2217 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2218 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2219 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2220 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2221 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2222 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2223 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2224 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2225 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2226 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2227 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2228 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2229 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2232 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2234 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2236 * passes.c (pass_manager::register_dump_files_1): Merge into
2237 (pass_manager::register_dump_files): this, and remove its handling of
2238 properties since the pass always has the properties anyway.
2239 (pass_manager::pass_manager): Adjust.
2241 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2243 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2244 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2245 dealing with properties.
2246 (pass_manager::register_dump_files): Adjust.
2248 2014-03-20 Mark Wielaard <mjw@redhat.com>
2250 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2251 then represent the bound as normal constant value.
2253 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2256 Forward port from 4.8 branch
2257 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2259 * config/i386/bmiintrin.h (_blsi_u32): New.
2263 (_blsmsk_u32): Ditto.
2264 (_blsmsk_u64): Ditto.
2265 (_tzcnt_u32): Ditto.
2266 (_tzcnt_u64): Ditto.
2268 2014-04-17 Kito Cheng <kito@0xlab.org>
2270 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2272 2014-04-17 Richard Biener <rguenther@suse.de>
2275 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2276 boolean results for comparisons.
2278 2014-04-17 Richard Biener <rguenther@suse.de>
2280 PR tree-optimization/60836
2281 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2282 initial PHI args to be gimple values.
2284 2014-04-17 Richard Biener <rguenther@suse.de>
2286 PR tree-optimization/60841
2287 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2288 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2289 of stmts to SLP build.
2290 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2291 (vect_analyze_slp): Likewise.
2292 (vect_analyze_slp_instance): Likewise.
2293 (vect_build_slp_tree): Limit overall SLP tree growth.
2294 * tree-vectorizer.h (vect_analyze_data_refs,
2295 vect_analyze_slp): Adjust prototypes.
2297 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2299 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2302 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2304 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2305 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2306 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2307 for TARGET_SLOW_PSHUFB
2309 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2311 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2312 * config/i386/i386.c (intel_cost): Ditto.
2314 2014-04-17 Joey Ye <joey.ye@arm.com>
2316 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2318 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2320 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2323 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2325 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2326 (type_all_derivations_known_p): New predicate.
2327 (type_all_ctors_visible_p): New predicate.
2328 (type_possibly_instantiated_p): New predicate.
2329 (get_odr_type): Compute all_derivations_known.
2330 (dump_odr_type): Dump the flag.
2331 (maybe_record_type): Cleanup.
2332 (record_target_from_binfo): Add bases_to_consider array;
2333 record bases for types w/o instances and skip CXX destructor.
2334 (possible_polymorphic_call_targets_1): Add bases_to_consider
2335 and consider_construction parameters; check if type may have instance.
2336 (get_polymorphic_call_info): Set maybe_in_construction to true
2337 when we know nothing.
2338 (record_targets_from_bases): Skip CXX destructors; they are
2339 never called for types in construction.
2340 (possible_polymorphic_call_targets): Do not record target when
2341 type may not have instance.
2343 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2346 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2347 external aliases alive, too.
2349 2014-04-16 Andrew Pinski <apinski@cavium.com>
2351 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2354 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2356 * final.c (compute_alignments): Do not apply loop alignment to a block
2357 falling through to the exit.
2359 2014-04-16 Catherine Moore <clm@codesourcery.com>
2361 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2362 Adjust constraints for microMIPS store patterns.
2364 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2366 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2368 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2370 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2371 (append_use): Run at -O0.
2372 (append_vdef): Likewise.
2373 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2374 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2376 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2378 PR tree-optimization/60844
2379 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2380 (propagate_op_to_single_use, remove_visited_stmt_chain,
2381 linearize_expr, repropagate_negates, reassociate_bb): Use it
2382 instead of gsi_remove.
2384 2014-04-16 Martin Jambor <mjambor@suse.cz>
2386 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2387 ipa_transforms_to_apply.
2388 (cgraph_function_versioning): Assert that old_node has empty
2389 ipa_transforms_to_apply.
2390 * trans-mem.c (ipa_tm_create_version): Likewise.
2391 * tree-inline.c (tree_function_versioning): Do not duplicate
2392 ipa_transforms_to_apply.
2394 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2397 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2399 Pass necessary as flags on 64-bit Solaris/x86.
2400 Use lowercase relocs for x86_64-*-*.
2401 * configure: Regenerate.
2403 2014-04-15 Jan Hubicka <jh@suse.cz>
2405 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2406 (maybe_record_node, likely_target_p): Use it.
2408 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2411 Revert following patch
2413 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2416 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2417 software floating point or no floating point registers, do not
2418 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2419 in GPRs that occurs after we tested for GPRs that would never be
2422 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2423 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2424 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2425 specifically allow DDmode, since that does not use the SPE SIMD
2428 2014-03-21 Mark Wielaard <mjw@redhat.com>
2430 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2431 as unsigned or int depending on type and value used.
2433 2014-04-15 Richard Biener <rguenther@suse.de>
2435 PR rtl-optimization/56965
2436 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2437 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2439 * alias.c (true_dependence_1): Do not call
2440 nonoverlapping_component_refs_p.
2441 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2442 nonoverlapping_component_refs_p.
2443 (indirect_refs_may_alias_p): Likewise.
2445 2014-04-15 Teresa Johnson <tejohnson@google.com>
2447 * cfg.c (dump_bb_info): Fix flags check.
2448 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2450 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2452 PR rtl-optimization/60663
2453 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2456 2014-04-15 Richard Biener <rguenther@suse.de>
2458 * lto-streamer.h (LTO_major_version): Bump to 4.
2460 2014-04-15 Richard Biener <rguenther@suse.de>
2462 * common.opt (lto_partition_model): New enum.
2463 (flto-partition=): Merge separate options with a single with argument,
2464 add -flto-partition=one support.
2465 * flag-types.h (enum lto_partition_model): Declare.
2466 * opts.c (finish_options): Remove duplicate -flto-partition=
2468 * lto-wrapper.c (run_gcc): Adjust.
2470 2014-04-15 Richard Biener <rguenther@suse.de>
2472 * alias.c (ncr_compar): New function.
2473 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2475 2014-04-15 Richard Biener <rguenther@suse.de>
2477 * alias.c (record_component_aliases): Do not walk BINFOs.
2479 2014-04-15 Richard Biener <rguenther@suse.de>
2481 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2482 Add struct function argument and adjust.
2483 (find_func_aliases_for_call): Likewise.
2484 (find_func_aliases): Likewise.
2485 (find_func_clobbers): Likewise.
2486 (intra_create_variable_infos): Likewise.
2487 (compute_points_to_sets): Likewise.
2488 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2490 2014-04-15 Richard Biener <rguenther@suse.de>
2492 * tree.c (iterative_hash_expr): Use enum tree_code_class
2493 to store TREE_CODE_CLASS.
2494 (tree_block): Likewise.
2495 (tree_set_block): Likewise.
2496 * tree.h (fold_build_pointer_plus_loc): Use
2497 convert_to_ptrofftype_loc.
2499 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2502 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2505 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2507 * cfgloop.h (struct loop): Move force_vectorize down.
2508 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2509 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2510 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2511 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2512 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2513 * tree-core.h (enum annot_expr_kind): Add new kind values.
2514 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2515 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2517 * tree.def (ANNOTATE_EXPR): Tweak comment.
2519 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2521 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2524 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2526 * tree.h (TYPE_IDENTIFIER): Declare.
2527 * tree.c (subrange_type_for_debug_p): Use it.
2528 * godump.c (go_format_type): Likewise.
2529 * dwarf2out.c (is_cxx_auto, modified_type_die,
2530 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2531 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2533 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2536 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2538 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2540 * config/i386/i386.c (examine_argument): Return bool. Return true if
2541 parameter should be passed in memory.
2542 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2543 (construct_container): Update calls to examine_argument.
2544 (function_arg_advance_64): Ditto.
2545 (return_in_memory_32): Merge with ix86_return_in_memory.
2546 (return_in_memory_64): Ditto.
2547 (return_in_memory_ms_64): Ditto.
2549 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2551 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2552 * coverage.c (coverage_compute_profile_id): Handle externally visible
2555 2014-04-14 Martin Jambor <mjambor@suse.cz>
2557 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2558 DECL_DISREGARD_INLINE_LIMITS functions.
2560 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2563 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2565 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2568 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2569 optimize_insn_for_speed_p instead of
2570 optimize_function_for_speed_p.
2572 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2574 * doc/invoke.texi (free): Document AArch64.
2576 2014-04-14 Richard Biener <rguenther@suse.de>
2578 PR tree-optimization/60042
2579 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2580 (insert_into_preds_of_block): Do not prevent PHI insertion
2581 for REFERENCE exprs here ...
2582 (eliminate_dom_walker::before_dom_children): ... but prevent
2583 their use here under similar conditions when applied to the
2584 IL after PRE optimizations.
2586 2014-04-14 Richard Biener <rguenther@suse.de>
2588 * passes.def: Move early points-to after early SRA.
2590 2014-04-14 Richard Biener <rguenther@suse.de>
2592 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2593 check for which sign-changes we allow when forwarding
2594 a converted value into a switch.
2596 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2598 * stor-layout.c (place_field): Finalize non-constant offset for the
2601 2014-04-14 Richard Biener <rguenther@suse.de>
2603 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2605 (expand_switch_using_bit_tests_p): Likewise.
2606 (process_switch): Compute and pass on speed_p based on the
2608 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2609 optimize_bb_for_speed_p.
2611 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2613 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2614 * function.h (struct function): Rename has_force_vect_loops into
2615 has_force_vectorize_loops.
2616 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2617 (input_struct_function_base): Likewise.
2618 * lto-streamer-out.c (output_cfg): Likewise.
2619 (output_struct_function_base): Likewise.
2620 * omp-low.c (expand_omp_simd): Likewise.
2621 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2622 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2623 (version_loop_for_if_conversion): Likewise.
2624 (tree_if_conversion): Likewise.
2625 (main_tree_if_conversion): Likewise.
2626 (gate_tree_if_conversion): Likewise.
2627 * tree-inline.c (copy_loops): Likewise.
2628 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2629 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2630 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2631 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2632 * tree-vectorizer.c (vectorize_loops): Likewise.
2633 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2635 2014-04-14 Richard Biener <rguenther@suse.de>
2638 * lto-streamer-out.c (wrap_refs): New function.
2639 (lto_output): Wrap symbol references in global initializes in
2640 type-preserving MEM_REFs.
2642 2014-04-14 Christian Bruel <christian.bruel@st.com>
2644 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2646 2014-04-14 Christian Bruel <christian.bruel@st.com>
2648 * config/sh/sh.md (setmemqi): New expand pattern.
2649 * config/sh/sh.h (CLEAR_RATIO): Define.
2650 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2651 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2653 2014-04-14 Richard Biener <rguenther@suse.de>
2656 * fold-const.c (negate_expr_p): Don't negate directional rounding
2658 (fold_negate_expr): Likewise.
2660 2014-04-14 Richard Biener <rguenther@suse.de>
2662 PR tree-optimization/59817
2663 PR tree-optimization/60453
2664 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2665 recursion to catch all CHRECs in the scalar evolution and restrict
2666 the predicate for the remains appropriately.
2668 2014-04-12 Catherine Moore <clm@codesourcery.com>
2670 * config/mips/constraints.md: Add new register constraint "kb".
2671 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2672 (*movhi_internal): Likewise.
2673 (*movqi_internal): Likewise.
2674 * config/mips/mips.h (M16_STORE_REGS): New register class.
2675 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2676 (REG_CLASS_CONTENTS): Likewise.
2677 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2679 2014-04-11 Tobias Burnus <burnus@net-b.de>
2682 * doc/invoke.texi (-Wformat-signedness): Document it.
2683 (Wformat=2): Mention that this enables -Wformat-signedness.
2685 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2687 * common/config/epiphany/epiphany-common.c
2688 (epiphany_option_optimization_table): Enable section anchors by
2689 default at -O1 or higher.
2690 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2691 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2692 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2693 carries no extra cost.
2694 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2695 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2696 * config/epiphany/predicates.md (memclob_operand): New predicate.
2697 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2698 Use memclob_operand predicate and X constraint for operand 3.
2700 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2702 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2703 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2706 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2708 PR rtl-optimization/60651
2709 * mode-switching.c (optimize_mode_switching): Make sure to emit
2710 sets of a lower numbered entity before sets of a higher numbered
2711 entity to a mode of the same or lower priority.
2712 When creating a seginfo for a basic block that starts with a code
2713 label, move the insertion point past the code label.
2714 (new_seginfo): Document and enforce requirement that
2715 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2716 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2717 * doc/tm.texi: Regenerate.
2719 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2722 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2724 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2726 * BASE-VER: Set to 4.10.0.
2728 2014-04-11 Tobias Burnus <burnus@net-b.de>
2731 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2732 * doc/gcc.texi (Service): Update description in the @menu
2733 * doc/invoke.texi (Option Summary): Remove misplaced and
2736 2014-04-11 Steve Ellcey <sellcey@mips.com>
2737 Jakub Jelinek <jakub@redhat.com>
2740 * expr.c (convert_move): Use emit_store_flag_force instead of
2741 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2744 2014-04-11 Richard Biener <rguenther@suse.de>
2747 * varasm.c (assemble_alias): Avoid endless error reporting
2748 recursion by setting TREE_ASM_WRITTEN.
2750 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2752 * config/s390/s390.md: Add a splitter for NOT rtx.
2754 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2756 PR rtl-optimization/60663
2757 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2759 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2760 Jakub Jelinek <jakub@redhat.com>
2763 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2764 flag from decl_node to node.
2766 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2769 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2770 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2771 ameliorating the cases where it can be.
2773 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2776 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2778 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2779 (loadsync_<mode>): Change mode.
2780 (load_quadpti, store_quadpti): New.
2781 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2782 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2783 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2785 2014-04-09 Cong Hou <congh@google.com>
2788 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2791 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2793 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2794 instead of vnor to exploit possible fusion opportunity in the
2796 (altivec_expand_vec_perm_const_le): Likewise.
2798 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2800 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2801 (loadsync_<mode>): Change mode.
2802 (load_quadpti, store_quadpti): New.
2803 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2804 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2806 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2809 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2810 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2811 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2813 2014-04-08 Richard Biener <rguenther@suse.de>
2816 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2817 a 64bit widest int print double-int similar to on HWI64 hosts.
2819 2014-04-08 Richard Biener <rguenther@suse.de>
2821 PR tree-optimization/60785
2822 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2823 default defs properly.
2825 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2827 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2828 (Weffc++): Likewise.
2830 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2832 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2833 set completep to false rather than true.
2835 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2838 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2839 ARM_TARGET2_DWARF_FORMAT.
2841 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2844 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2845 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2848 2014-04-07 Richard Biener <rguenther@suse.de>
2850 PR tree-optimization/60766
2851 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2852 (may_eliminate_iv): Convert cand_value_at result to desired type.
2854 2014-04-07 Jason Merrill <jason@redhat.com>
2857 * common.opt (-fno-gnu-unique): Add.
2858 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2860 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2862 * haifa-sched.c: Fix outdated function reference and minor
2863 grammar errors in introductory comment.
2865 2014-04-07 Richard Biener <rguenther@suse.de>
2868 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2870 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2872 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2875 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2876 size accounting for thunks.
2877 (pa_asm_output_mi_thunk): Use final_start_function() and
2878 final_end_function() to output function start and end directives.
2880 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2882 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2883 device specific ISA/ feature information. Remove short_sp and
2884 errata_skip ds. Add avr_device_specific_features enum to have device
2886 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2887 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2888 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2889 updated device specific info.
2890 * config/avr/avr-mcus.def: Merge device specific details to
2891 dev_attribute field.
2892 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2894 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2895 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2896 assembler if RMW isa supported by current device.
2897 * config/avr/genmultilib.awk: Update as device info structure changed.
2898 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2900 2014-04-04 Cong Hou <congh@google.com>
2902 PR tree-optimization/60656
2903 * tree-vect-stmts.c (supportable_widening_operation):
2904 Fix a bug that elements in a vector with vect_used_by_reduction
2905 property are incorrectly reordered when the operation on it is not
2906 consistant with the one in reduction operation.
2908 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2910 PR rtl-optimization/60155
2911 * gcse.c (record_set_data): New function.
2912 (single_set_gcse): New function.
2913 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2914 (hoist_code): Likewise.
2915 (get_pressure_class_and_nregs): Likewise.
2917 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2919 * explow.c (probe_stack_range): Emit a final optimization blockage.
2921 2014-04-04 Anthony Green <green@moxielogic.com>
2923 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2926 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2929 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2930 flags are set only during streaming.
2931 * ipa.c (process_references, walk_polymorphic_call_targets,
2932 symtab_remove_unreachable_nodes): Drop bodies of always inline
2933 after early inlining.
2934 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2936 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2937 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2940 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2943 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2946 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2948 (cortex_a53_fdivd): Likewise.
2950 2014-04-04 Martin Jambor <mjambor@suse.cz>
2953 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2955 * cgraph.c (clone_of_p): Also return true if thunks match.
2956 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2957 cgraph_function_or_thunk_node and an obsolete comment.
2958 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2960 (build_function_decl_skip_args): Likewise.
2961 (set_new_clone_decl_and_node_flags): New function.
2962 (duplicate_thunk_for_node): Likewise.
2963 (redirect_edge_duplicating_thunks): Likewise.
2964 (cgraph_clone_node): New parameter args_to_skip, pass it to
2965 redirect_edge_duplicating_thunks which is called instead of
2966 cgraph_redirect_edge_callee.
2967 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2968 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2970 2014-04-04 Jeff Law <law@redhat.com>
2973 * config/arm/predicates.md (const_int_I_operand): New predicate.
2974 (const_int_M_operand): Similarly.
2975 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2977 (insv_t2, extv_reg, extzv_t2): Likewise.
2978 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2979 (pop_multiple_with_writeback_and_return): Likewise.
2980 (vfp_pop_multiple_with_writeback): Likewise
2982 2014-04-04 Richard Biener <rguenther@suse.de>
2985 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2986 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2988 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2989 * gimplify.c (gimple_add_tmp_var_fn): New function.
2990 * gimple-expr.h (create_tmp_reg_fn): Declare.
2991 * gimple-expr.c (create_tmp_reg_fn): New function.
2992 * gimple-low.c (record_vars_into): Don't change cfun.
2993 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2994 code generation without cfun.
2996 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2999 * Makefile.in (install-driver): Fix shell scripting.
3001 2014-04-03 Cong Hou <congh@google.com>
3003 PR tree-optimization/60505
3004 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
3005 threshold of number of iterations below which no vectorization
3007 * tree-vect-loop.c (new_loop_vec_info):
3008 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
3009 * tree-vect-loop.c (vect_analyze_loop_operations):
3010 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
3011 * tree-vect-loop.c (vect_transform_loop):
3012 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
3013 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
3014 of iterations of the loop and see if we should build the epilogue.
3016 2014-04-03 Richard Biener <rguenther@suse.de>
3018 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
3019 (streamer_tree_cache_create): Adjust.
3020 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
3021 to allow optional nodes array.
3022 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
3023 (streamer_tree_cache_append): Likewise.
3024 (streamer_tree_cache_create): Create nodes array optionally
3025 as specified by parameter.
3026 * lto-streamer-out.c (create_output_block): Avoid maintaining
3027 the node array in the writer cache.
3028 (DFS_write_tree): Remove assertion.
3029 (produce_asm_for_decls): Free the out decl state hash table early.
3030 * lto-streamer-in.c (lto_data_in_create): Adjust for
3031 streamer_tree_cache_create prototype change.
3033 2014-04-03 Richard Biener <rguenther@suse.de>
3035 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3036 set TREE_CHAIN to NULL_TREE.
3038 2014-04-03 Richard Biener <rguenther@suse.de>
3040 PR tree-optimization/60740
3041 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3042 over all GIMPLE_COND operands.
3044 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3046 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3047 (Weffc++): Remove Scott's numbering, merge lists and reference
3050 2014-04-03 Nick Clifton <nickc@redhat.com>
3052 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3055 2014-04-03 Martin Jambor <mjambor@suse.cz>
3057 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3058 mention gcc_unreachable before failing.
3059 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3062 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3065 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3066 inconsistent code and instead mark the context inconsistent.
3067 (possible_polymorphic_call_targets): For inconsistent contexts
3068 return empty complete list.
3070 2014-04-02 Anthony Green <green@moxielogic.com>
3072 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3073 (extendqisi2, extendhisi2): Define.
3074 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3075 (WCHAR_TYPE): Change to unsigned int.
3077 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3079 PR tree-optimization/60733
3080 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3081 insertion point for PHI candidates to be the end of the feeding
3082 block for the PHI argument.
3084 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3086 PR rtl-optimization/60650
3087 * lra-constraints.c (process_alt_operands): Decrease reject for
3088 earlyclobber matching.
3090 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3092 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3094 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3096 * config/spu/spu.c (pad_bb): Do not crash when the last
3097 insn is CODE_FOR_blockage.
3099 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3101 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3102 lies outside the target mode.
3104 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3107 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3108 software floating point or no floating point registers, do not
3109 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3110 in GPRs that occurs after we tested for GPRs that would never be
3113 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3114 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3115 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3116 specifically allow DDmode, since that does not use the SPE SIMD
3119 2014-04-02 Richard Biener <rguenther@suse.de>
3122 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
3123 MODE_INTs. Properly use negv_optab.
3124 (expand_abs): Likewise.
3126 2014-04-02 Richard Biener <rguenther@suse.de>
3129 * Makefile.in (install-driver): Guard extra installs with special
3132 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3134 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3137 2014-04-01 Richard Henderson <rth@redhat.com>
3140 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3141 alternative enabled before register allocation.
3143 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3145 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3146 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3148 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3149 (nios2_got_address): Update nios2_large_got_address call site.
3150 (nios2_delegitimize_address): New function.
3151 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3152 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3153 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3155 2014-04-01 Martin Husemann <martin@duskware.de>
3157 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3160 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3162 PR rtl-optimization/60604
3163 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3164 check from register_operand.
3165 (register_operand): Redefine in terms of general_operand.
3166 (nonmemory_operand): Use register_operand for the non-constant cases.
3168 2014-04-01 Richard Biener <rguenther@suse.de>
3170 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3172 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3174 * doc/invoke.texi (mapp-regs): Clarify.
3176 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3178 * config/i386/avx512fintrin.h (__v32hi): Define type.
3179 (__v64qi): Likewise.
3180 (_mm512_set1_epi8): Define.
3181 (_mm512_set1_epi16): Define.
3182 (_mm512_set4_epi32): Define.
3183 (_mm512_set4_epi64): Define.
3184 (_mm512_set4_pd): Define.
3185 (_mm512_set4_ps): Define.
3186 (_mm512_setr4_epi64): Define.
3187 (_mm512_setr4_epi32): Define.
3188 (_mm512_setr4_pd): Define.
3189 (_mm512_setr4_ps): Define.
3190 (_mm512_setzero_epi32): Define.
3192 2014-03-31 Martin Jambor <mjambor@suse.cz>
3195 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3196 callsite_arguments_match_p. Updated all callers. Also check types of
3197 corresponding formal parameters and actual arguments.
3198 (not_all_callers_have_enough_arguments_p) Renamed to
3199 some_callers_have_mismatched_arguments_p.
3201 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3203 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3205 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3208 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3211 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3213 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3214 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3216 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3217 Use FMAMODE_NOVF512 mode iterator.
3218 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3219 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3220 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3221 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3223 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3224 Use VF_128_256 mode iterator.
3225 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3228 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3230 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3231 static chain if needed.
3233 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3236 * lra-constraints.c (index_part_to_reg): New.
3237 (process_address): Use it.
3239 2014-03-27 Jeff Law <law@redhat.com>
3240 Jakub Jelinek <jakub@redhat.com>
3243 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3244 gen_rtx_{PLUS,MULT} to build up the address expression.
3246 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3247 creating non-canonical RTL.
3249 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3252 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3253 functions; reorganize to make cheap checks first.
3254 (inline_small_functions): Do not estimate growth when dumping;
3256 * ipa-inline.h (inline_summary): Add min_size.
3257 (growth_likely_positive): New function.
3258 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3259 (set_cond_stmt_execution_predicate): Cleanup.
3260 (estimate_edge_size_and_time): Compute min_size.
3261 (estimate_calls_size_and_time): Likewise.
3262 (estimate_node_size_and_time): Likewise.
3263 (inline_update_overall_summary): Update min_size.
3264 (do_estimate_edge_time): Likewise.
3265 (do_estimate_edge_size): Update.
3266 (do_estimate_edge_hints): Update.
3267 (growth_likely_positive): New function.
3269 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3272 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3273 also if addr has VOIDmode.
3275 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3277 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3278 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3280 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3281 instructions as well as AdvancedSIMD loads.
3283 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3285 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3286 Use crypto_aese type.
3287 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3288 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3289 crypto_aese, crypto_aesmc. Move to types.md.
3290 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3292 * config/arm/iterators.md (crypto_type): Likewise.
3294 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3296 * cgraph.c: Include expr.h and tree-dfa.h.
3297 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3300 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3303 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3304 regs from checking multi-reg pseudos.
3306 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3308 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3310 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3312 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3313 if it would clobber the stack pointer, even temporarily.
3315 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3317 * mode-switching.c: Make small adjustments to the top comment.
3319 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3321 * config/rs6000/constraints.md (wD constraint): New constraint to
3322 match the constant integer to get the top DImode/DFmode out of a
3323 vector in a VSX register.
3325 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3326 match the constant integer to get the top DImode/DFmode out of a
3327 vector in a VSX register.
3329 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3332 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3335 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3336 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3338 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3339 Optimize vec_extract of 64-bit values, where the value being
3340 extracted is in the top word, where we can use scalar
3341 instructions. Add direct move and store support. Combine the big
3342 endian/little endian vector select load support into a single insn.
3343 (vsx_extract_<mode>_internal1): Likewise.
3344 (vsx_extract_<mode>_internal2): Likewise.
3345 (vsx_extract_<mode>_load): Likewise.
3346 (vsx_extract_<mode>_store): Likewise.
3347 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3348 combined into vsx_extract_<mode>_load.
3349 (vsx_extract_<mode>_one_le): Likewise.
3351 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3352 define the top 64-bit vector element.
3354 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3357 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3358 Document vec_vbpermq builtin.
3361 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3362 enable use of xxsldwi and xxpermdi builtin functions.
3363 (vec_xxpermdi): Likewise.
3365 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3366 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3368 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3370 PR rtl-optimization/60650
3371 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3373 (find_spills_for): New.
3374 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3375 Spill all pseudos on the second iteration.
3377 2014-03-27 Marek Polacek <polacek@redhat.com>
3380 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3383 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3385 * config/s390/s390.c (s390_can_use_return_insn): Check for
3386 call-saved FPRs on 31 bit.
3388 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3391 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3392 if they need regimplification, just drop them instead of
3393 calling gimple_regimplify_operands on them.
3395 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3398 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3399 (aarch64_frame_pointer_required): Adjust logic.
3400 (aarch64_can_eliminate): Adjust logic.
3401 (aarch64_override_options_after_change): Adjust logic.
3403 2014-03-27 Dehao Chen <dehao@google.com>
3405 * ipa-inline.c (early_inliner): Update node's inline info.
3407 2014-03-26 Dehao Chen <dehao@google.com>
3409 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3410 compiler inserted conditional jumps for NAN float check.
3412 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3414 * ubsan.h (ubsan_create_data): Change second argument's type
3415 to const location_t *.
3416 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3418 (ubsan_create_data): Change second argument to const location_t *PLOC.
3419 Create Loc field whenever PLOC is non-NULL.
3420 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3421 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3425 * real.c (real_to_integer2): Change type of low to UHWI.
3427 2014-03-26 Tobias Burnus <burnus@net-b.de>
3429 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3430 (CILK_SELF_SPECS): New define.
3431 (driver_self_specs): Use it.
3433 2014-03-26 Richard Biener <rguenther@suse.de>
3435 * tree-pretty-print.c (percent_K_format): Implement special
3436 case for LTO and its stripped down BLOCK tree.
3438 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3441 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3443 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3444 one range is range_int_cst_p, but not both, at least optimize
3445 addition/subtraction of 0 and multiplication by 0 or 1.
3446 * gimple-fold.c (gimple_fold_call): Fold
3447 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3448 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3449 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3451 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3453 PR rtl-optimization/60452
3454 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3455 <case REG>: Return 1 for invalid offsets from the frame pointer.
3457 2014-03-26 Marek Polacek <polacek@redhat.com>
3460 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3463 2014-03-26 Marek Polacek <polacek@redhat.com>
3466 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3469 2014-03-26 Marek Polacek <polacek@redhat.com>
3472 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3473 multiplication in unsigned type.
3475 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3477 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3479 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3481 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3483 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3486 * cif-code.def (UNREACHABLE) New code.
3487 * ipa-inline.c (inline_small_functions): Skip edges to
3488 __builtlin_unreachable.
3489 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3490 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3491 predicate to __bulitin_unreachable.
3492 (set_cond_stmt_execution_predicate): Fix issue when
3493 invert_tree_comparison returns ERROR_MARK.
3494 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3495 propagate to inline clones.
3496 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3498 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3499 * cgraphclones.c (cgraph_clone_node): If call destination is already
3500 ureachable, do not redirect it back.
3501 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3504 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3506 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3507 Do not modify inline clones.
3509 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3511 * config/i386/i386.md (general_sext_operand): New mode attr.
3512 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3513 don't generate (sign_extend (const_int)).
3514 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3515 operands[2]. Use We constraint instead of <i> and
3516 <general_sext_operand> predicate instead of <general_operand>.
3517 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3518 * config/i386/constraints.md (We): New constraint.
3519 * config/i386/predicates.md (x86_64_sext_operand,
3520 sext_operand): New predicates.
3522 2014-03-25 Martin Jambor <mjambor@suse.cz>
3525 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3526 inconsistent devirtualizations to __builtin_unreachable.
3528 2014-03-25 Marek Polacek <polacek@redhat.com>
3531 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3533 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3535 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3536 order of elements for big-endian.
3538 2014-03-25 Richard Biener <rguenther@suse.de>
3541 * gimplify-me.c (gimple_regimplify_operands): Update the
3544 2014-03-25 Martin Jambor <mjambor@suse.cz>
3547 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3548 (lto_output_varpool_node): Likewise.
3549 (input_overwrite_node): Likewise.
3550 (input_varpool_node): Likewise.
3552 2014-03-25 Richard Biener <rguenther@suse.de>
3554 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3555 (run_gcc): Likewise.
3557 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3559 * combine.c (simplify_compare_const): Add MODE argument.
3560 Handle mode_width 0 as very large mode_width.
3561 (try_combine, simplify_comparison): Adjust callers.
3563 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3564 type to avoid signed integer overflow.
3565 * explow.c (plus_constant): Likewise.
3567 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3569 * doc/generic.texi: Correct typos.
3571 2014-03-24 Tobias Burnus <burnus@net-b.de>
3573 * doc/invoke.texi (-flto): Expand section about
3574 using static libraries with LTO.
3576 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3578 PR rtl-optimization/60501
3579 * optabs.def (addptr3_optab): New optab.
3580 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3581 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3582 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3584 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3586 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3588 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3590 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3593 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3594 (_mm256_undefined_ps): Define.
3595 (_mm256_undefined_pd): Define.
3596 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3597 (_mm_undefined_pd): Define.
3598 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3599 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3600 (_mm512_undefined_ps): Define.
3601 (_mm512_undefined_pd): Define.
3602 Use _mm*_undefined_*.
3603 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3605 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3607 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3608 (lshr_simd): DI mode added.
3609 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3610 (aarch64_ushr_simddi): Likewise.
3611 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3612 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3613 (vshrd_n_u64): Likewise.
3615 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3617 * Makefile.in (s-macro_list): Depend on cc1.
3619 2014-03-23 Teresa Johnson <tejohnson@google.com>
3621 * ipa-utils.c (ipa_print_order): Use specified dump file.
3623 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3625 PR rtl-optimization/60601
3626 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3628 * gcc.c (eval_spec_function): Initialize save_growing_value.
3630 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3633 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3634 code == MINUS_EXPR, never swap op0 with op1.
3636 * toplev.c (init_local_tick): Avoid signed integer multiplication
3638 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3639 shift by first operand's bitsize.
3641 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3644 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3646 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3647 TARGET_ISA_64BIT_P(x).
3649 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3651 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3652 pattern for vector nor instead of subtract from splat(-1).
3653 (altivec_expand_vec_perm_const_le): Likewise.
3655 2014-03-21 Richard Henderson <rth@twiddle.net>
3658 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3659 related insns after epilogue_completed.
3661 2014-03-21 Martin Jambor <mjambor@suse.cz>
3664 * cgraph.h (symtab_node): New flag body_removed.
3665 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3666 when removing bodies.
3667 * symtab.c (dump_symtab_base): Dump body_removed flag.
3668 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3669 had their bodies removed.
3671 2014-03-21 Martin Jambor <mjambor@suse.cz>
3674 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3677 2014-03-21 Richard Biener <rguenther@suse.de>
3679 PR tree-optimization/60577
3680 * tree-core.h (struct tree_base): Document nothrow_flag use
3681 in VAR_DECL_NONALIASED.
3682 * tree.h (VAR_DECL_NONALIASED): New.
3683 (may_be_aliased): Adjust.
3684 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3686 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3688 * expr.c (expand_expr_real_1): Remove outdated comment.
3690 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3693 * ira.c (adjust_cleared_regs): Call copy_rtx on
3694 *reg_equiv[REGNO (loc)].src_p before passing it to
3695 simplify_replace_fn_rtx.
3698 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3699 into CONST, put pic register as first operand of PLUS. Use
3700 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3702 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3704 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3706 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3708 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3709 around for store forwarding issue in the FPU on the UT699.
3710 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3711 loads and operations if -mfix-ut699 is specified.
3712 (divtf3_hq): Tweak attribute.
3713 (sqrttf2_hq): Likewise.
3715 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3717 * calls.c (store_one_arg): Remove incorrect const qualification on the
3718 type of the temporary.
3719 * cfgexpand.c (expand_return): Likewise.
3720 * expr.c (expand_constructor): Likewise.
3721 (expand_expr_real_1): Likewise.
3723 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3725 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3728 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3731 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3733 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3735 * config/arm/aarch-common-protos.h
3736 (alu_cost_table): Fix spelling of "extend".
3737 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3739 2014-03-19 Richard Biener <rguenther@suse.de>
3742 * tree-core.h (tree_type_common): Re-order pointer members
3743 to reduce recursion depth during GC walks.
3745 2014-03-19 Marek Polacek <polacek@redhat.com>
3748 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3749 before accessing it.
3751 2014-03-19 Richard Biener <rguenther@suse.de>
3754 * lto-streamer-in.c (input_function): In WPA stage do not drop
3757 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3759 PR tree-optimization/60559
3760 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3761 with build_zero_cst assignment.
3763 2014-03-18 Kai Tietz <ktietz@redhat.com>
3765 PR rtl-optimization/56356
3766 * sdbout.c (sdbout_parms): Verify that parms'
3767 incoming argument is valid.
3768 (sdbout_reg_parms): Likewise.
3770 2014-03-18 Richard Henderson <rth@redhat.com>
3773 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3774 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3775 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3777 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3779 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3780 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3781 Italicize plugin event names in description. Explain that
3782 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3783 Remind that no GCC functions should be called after PLUGIN_FINISH.
3784 Explain what pragmas with expansion are.
3786 2014-03-18 Martin Liska <mliska@suse.cz>
3788 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3789 gimple call statement is update.
3790 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3791 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3793 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3796 * ubsan.c (ubsan_instrument_unreachable): Call
3797 initialize_sanitizer_builtins.
3798 (ubsan_pass): Likewise.
3801 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3802 varpool_finalize_decl instead of rest_of_decl_compilation.
3804 2014-03-18 Richard Biener <rguenther@suse.de>
3806 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3807 by using bitmap_and_compl instead of bitmap_and_compl_into.
3808 (df_rd_transfer_function): Likewise.
3810 2014-03-18 Richard Biener <rguenther@suse.de>
3812 * doc/lto.texi (fresolution): Fix typo.
3814 2014-03-18 Richard Biener <rguenther@suse.de>
3816 * doc/invoke.texi (flto): Update for changes in 4.9.
3818 2014-03-18 Richard Biener <rguenther@suse.de>
3820 * doc/loop.texi: Remove section on the removed lambda framework.
3821 Update loop docs with recent changes in preserving loop structure.
3823 2014-03-18 Richard Biener <rguenther@suse.de>
3825 * doc/lto.texi (-fresolution): Document.
3827 2014-03-18 Richard Biener <rguenther@suse.de>
3829 * doc/contrib.texi: Adjust my name.
3831 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3834 * internal-fn.c: Include diagnostic-core.h.
3835 (expand_BUILTIN_EXPECT): New function.
3836 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3837 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3838 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3839 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3841 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3842 Revert 3 argument __builtin_expect code.
3843 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3844 * gimple-fold.c (gimple_fold_call): Likewise.
3845 * tree.h (fold_builtin_expect): New prototype.
3846 * builtins.c (build_builtin_expect_predicate): Add predictor
3847 argument, if non-NULL, create 3 argument __builtin_expect.
3848 (fold_builtin_expect): No longer static. Add ARG2 argument,
3849 pass it through to build_builtin_expect_predicate.
3850 (fold_builtin_2): Adjust caller.
3851 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3852 * internal-fn.def (BUILTIN_EXPECT): New.
3854 2014-03-18 Tobias Burnus <burnus@net-b.de>
3857 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3858 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3859 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3861 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3864 * predict.c (combine_predictions_for_bb): Fix up formatting.
3865 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3866 fill what it points to if non-NULL.
3867 (tree_predict_by_opcode): Adjust caller, use the predictor.
3868 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3870 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3872 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3873 proper constant for the store mode.
3875 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3877 * symtab.c (change_decl_assembler_name): Fix transparent alias
3880 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3882 * config/aarch64/aarch64.c: Correct the comments about the
3883 aarch64 stack layout.
3885 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3887 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3888 check for GF_OMP_FOR_KIND_FOR.
3890 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3892 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3893 ymm and zmm register names.
3895 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3898 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3899 note creation for the 2010-08-31 changes.
3901 2014-03-17 Marek Polacek <polacek@redhat.com>
3904 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3905 as -fno-tree-loop-vectorize.
3906 (expand_omp_simd): Likewise.
3908 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3910 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3911 (eligible_for_call_delay): New prototype.
3912 * config/sparc/sparc.c (tls_call_delay): Rename into...
3913 (eligible_for_call_delay): ...this. Return false if the instruction
3914 cannot be put in the delay slot of a branch.
3915 (eligible_for_restore_insn): Simplify.
3916 (eligible_for_return_delay): Return false if the instruction cannot be
3917 put in the delay slot of a branch and simplify.
3918 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3919 put in the delay slot of a branch.
3920 * config/sparc/sparc.md (fix_ut699): New attribute.
3921 (tls_call_delay): Delete.
3922 (in_call_delay): Reimplement.
3923 (eligible_for_sibcall_delay): Rename into...
3924 (in_sibcall_delay): ...this.
3925 (eligible_for_return_delay): Rename into...
3926 (in_return_delay): ...this.
3927 (in_branch_delay): Reimplement.
3928 (in_uncond_branch_delay): Delete.
3929 (in_annul_branch_delay): Delete.
3931 2014-03-14 Richard Henderson <rth@redhat.com>
3934 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3935 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3936 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3937 (floathi splitters): Remove.
3938 (float<SWI48x>xf2): New pattern.
3939 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3940 code that tried to handle DImode for 32-bit, but which was excluded
3941 by the pattern's condition. Drop allocation of stack temporary.
3942 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3943 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3944 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3945 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3946 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3947 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3948 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3949 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3950 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3951 (*float<SWI48x><X87MODEF>2_i387): Remove.
3952 (all float _with_temp splitters): Remove.
3953 (*float<SWI48x><MODEF>2_i387): New pattern.
3954 (*float<SWI48><MODEF>2_sse): New pattern.
3955 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3956 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3958 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3959 Marek Polacek <polacek@redhat.com>
3962 * common.opt (dump_base_name_prefixed): New Variable.
3963 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3964 if x_dump_base_name_prefixed is already set, set it at the end.
3966 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3968 PR rtl-optimization/60508
3969 * lra-constraints.c (get_reload_reg): Add new parameter
3971 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3972 Pass the new parameter values.
3974 2014-03-14 Richard Biener <rguenther@suse.de>
3976 * common.opt: Revert unintented changes from r205065.
3979 2014-03-14 Richard Biener <rguenther@suse.de>
3982 * cfghooks.c (split_block): Properly adjust all loops the
3983 block was a latch of.
3985 2014-03-14 Martin Jambor <mjambor@suse.cz>
3988 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3991 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3994 * config/avr/avr.c (avr_set_current_function): Pass function name
3995 through default_strip_name_encoding before sanity checking instead
3996 of skipping the first char of the assembler name.
3998 2014-03-13 Richard Henderson <rth@redhat.com>
4001 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
4002 (ix86_force_to_memory, ix86_free_from_memory): Remove.
4003 * config/i386/i386-protos.h: Likewise.
4004 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
4005 in the expander instead of a splitter.
4006 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
4007 any possibility of requiring a memory.
4008 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
4009 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
4010 (fp branch splitters): Update for ix86_split_fp_branch.
4011 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
4012 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
4013 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
4014 (*fop_<MODEF>_2_i387): Remove f/r alternative.
4015 (*fop_<MODEF>_3_i387): Likewise.
4016 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
4017 (splitters for the fop_* register patterns): Remove.
4018 (fscalexf4_i387): Rename from *fscalexf4_i387.
4019 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4021 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4023 PR tree-optimization/59779
4024 * tree-dfa.c (get_ref_base_and_extent): Use double_int
4025 type for bitsize and maxsize instead of HOST_WIDE_INT.
4027 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4029 PR rtl-optimization/57320
4030 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4031 the CFG after thread_prologue_and_epilogue_insns.
4033 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4035 PR rtl-optimization/57189
4036 * lra-constraints.c (process_alt_operands): Disfavor spilling
4039 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4041 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4043 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4045 PR tree-optimization/59025
4047 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4048 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4050 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4053 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4054 calls of avr_out_plus_1.
4056 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4058 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4059 BB's single pred and update the father loop's latch info later.
4061 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4063 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4067 (VEC_base): Likewise.
4068 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4069 registers, we need to swap double words in little endian mode.
4071 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4072 to be a container mode for 128-bit integer operations added in ISA
4073 2.07. Unlike TImode and PTImode, the preferred register set is
4074 the Altivec/VMX registers for the 128-bit operations.
4076 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4078 (rs6000_split_128bit_ok_p): Likewise.
4080 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4081 macros for creating ISA 2.07 normal and overloaded builtin
4082 functions with 3 arguments.
4083 (BU_P8V_OVERLOAD_3): Likewise.
4084 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4085 for use as overloaded functions.
4086 (VPERM_1TI_UNS): Likewise.
4087 (VSEL_1TI): Likewise.
4088 (VSEL_1TI_UNS): Likewise.
4089 (ST_INTERNAL_1ti): Likewise.
4090 (LD_INTERNAL_1ti): Likewise.
4091 (XXSEL_1TI): Likewise.
4092 (XXSEL_1TI_UNS): Likewise.
4093 (VPERM_1TI): Likewise.
4094 (VPERM_1TI_UNS): Likewise.
4095 (XXPERMDI_1TI): Likewise.
4096 (SET_1TI): Likewise.
4097 (LXVD2X_V1TI): Likewise.
4098 (STXVD2X_V1TI): Likewise.
4099 (VEC_INIT_V1TI): Likewise.
4100 (VEC_SET_V1TI): Likewise.
4101 (VEC_EXT_V1TI): Likewise.
4102 (EQV_V1TI): Likewise.
4103 (NAND_V1TI): Likewise.
4104 (ORC_V1TI): Likewise.
4105 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4106 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4108 (VADDUQM): Likewise.
4109 (VSUBCUQ): Likewise.
4110 (VADDEUQM): Likewise.
4111 (VADDECUQ): Likewise.
4112 (VSUBEUQM): Likewise.
4113 (VSUBECUQ): Likewise.
4115 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
4116 __int128_t and __uint128_t types.
4117 (__uint128_type): Likewise.
4118 (altivec_categorize_keyword): Add support for vector __int128_t,
4119 vector __uint128_t, vector __int128, and vector unsigned __int128
4120 as a container type for TImode operations that need to be done in
4121 VSX/Altivec registers.
4122 (rs6000_macro_to_expand): Likewise.
4123 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
4124 to support 128-bit integer instructions vaddcuq, vadduqm,
4125 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
4126 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4128 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4129 for V1TImode, and set up preferences to use VSX/Altivec registers.
4130 Setup VSX reload handlers.
4131 (rs6000_debug_reg_global): Likewise.
4132 (rs6000_init_hard_regno_mode_ok): Likewise.
4133 (rs6000_preferred_simd_mode): Likewise.
4134 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4135 (easy_altivec_constant): Likewise.
4136 (output_vec_const_move): Likewise.
4137 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4139 (rs6000_expand_vector_extract): Likewise.
4140 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4142 (rs6000_const_vec): Add support for V1TImode.
4143 (rs6000_emit_le_vsx_load): Swap double words when loading or
4144 storing TImode/V1TImode.
4145 (rs6000_emit_le_vsx_store): Likewise.
4146 (rs6000_emit_le_vsx_move): Likewise.
4147 (rs6000_emit_move): Add support for V1TImode.
4148 (altivec_expand_ld_builtin): Likewise.
4149 (altivec_expand_st_builtin): Likewise.
4150 (altivec_expand_vec_init_builtin): Likewise.
4151 (altivec_expand_builtin): Likewise.
4152 (rs6000_init_builtins): Add support for V1TImode type. Add
4153 support for ISA 2.07 128-bit integer builtins. Define type names
4154 for the VSX/Altivec vector types.
4155 (altivec_init_builtins): Add support for overloaded vector
4156 functions with V1TImode type.
4157 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4158 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4160 (rs6000_split_128bit_ok_p): Likewise.
4161 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4162 __int128_t and vector __uint128_t.
4164 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4165 and mode attributes.
4172 (VS_scalar): Likewise.
4173 (VS_double): Likewise.
4174 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4176 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4177 we support the ISA 2.07 128-bit integer arithmetic instructions.
4178 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4179 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4180 and TImode types for use with the builtin functions.
4181 (V1TI_type_node): Likewise.
4182 (unsigned_V1TI_type_node): Likewise.
4183 (intTI_type_internal_node): Likewise.
4184 (uintTI_type_internal_node): Likewise.
4186 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4187 128-bit builtin functions.
4188 (UNSPEC_VADDEUQM): Likewise.
4189 (UNSPEC_VADDECUQ): Likewise.
4190 (UNSPEC_VSUBCUQ): Likewise.
4191 (UNSPEC_VSUBEUQM): Likewise.
4192 (UNSPEC_VSUBECUQ): Likewise.
4193 (VM): Add V1TImode to vector mode iterators.
4195 (VI_unit): Likewise.
4196 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4197 (altivec_vaddcuq): Likewise.
4198 (altivec_vsubuqm): Likewise.
4199 (altivec_vsubcuq): Likewise.
4200 (altivec_vaddeuqm): Likewise.
4201 (altivec_vaddecuq): Likewise.
4202 (altivec_vsubeuqm): Likewise.
4203 (altivec_vsubecuq): Likewise.
4205 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4207 (BOOL_128): Likewise.
4208 (BOOL_REGS_OUTPUT): Likewise.
4209 (BOOL_REGS_OP1): Likewise.
4210 (BOOL_REGS_OP2): Likewise.
4211 (BOOL_REGS_UNARY): Likewise.
4212 (BOOL_REGS_AND_CR0): Likewise.
4214 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4215 128-bit integer builtin support.
4216 (vec_vadduqm): Likewise.
4217 (vec_vaddecuq): Likewise.
4218 (vec_vaddeuqm): Likewise.
4219 (vec_vsubecuq): Likewise.
4220 (vec_vsubeuqm): Likewise.
4221 (vec_vsubcuq): Likewise.
4222 (vec_vsubuqm): Likewise.
4224 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4225 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4226 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4227 128-bit integer add/subtract to ISA 2.07.
4229 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4231 * config/arc/arc.c (arc_predicate_delay_insns):
4232 Fix third argument passed to conditionalize_nonjump.
4234 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4236 * config/aarch64/aarch64-builtins.c
4237 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4238 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4239 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4240 instead of __builtin_lfloor.
4241 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4243 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4245 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4246 (tree_ssa_ifcombine_bb_1): New function.
4247 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4248 is an empty forwarder block to then_bb or vice versa and then_bb
4249 and else_bb are effectively swapped.
4251 2014-03-12 Christian Bruel <christian.bruel@st.com>
4254 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4255 REG_CFA_DEF_CFA note.
4256 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4257 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4259 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4261 PR tree-optimization/60454
4262 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4264 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4266 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4267 Do not define target_cpu_default2 to generic.
4268 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4269 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4270 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4272 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4273 Marc Glisse <marc.glisse@inria.fr>
4275 PR tree-optimization/60502
4276 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4277 instead of build_low_bits_mask.
4279 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4282 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4283 if there are multiple uses, but op doesn't live on E edge.
4284 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4285 clobber stmts before __builtin_unreachable.
4287 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4289 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4290 hard_frame_pointer_rtx.
4291 * cse.c (cse_insn): Remove volatile check.
4292 * cselib.c (cselib_process_insn): Likewise.
4293 * dse.c (scan_insn): Likewise.
4295 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4297 * config/arc/arc.c (conditionalize_nonjump): New function,
4299 (arc_ifcvt): ... this.
4300 (arc_predicate_delay_insns): Use it.
4302 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4304 * config/arc/predicates.md (extend_operand): During/after reload,
4305 allow const_int_operand.
4306 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4307 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4308 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4310 (umulsi3_highpart_i): Likewise.
4312 2014-03-11 Richard Biener <rguenther@suse.de>
4314 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4315 Add asserts to guard possible wrong-code bugs.
4317 2014-03-11 Richard Biener <rguenther@suse.de>
4319 PR tree-optimization/60429
4320 PR tree-optimization/60485
4321 * tree-ssa-structalias.c (set_union_with_increment): Properly
4322 take into account all fields that overlap the shifted vars.
4323 (do_sd_constraint): Likewise.
4324 (do_ds_constraint): Likewise.
4325 (get_constraint_for_ptr_offset): Likewise.
4327 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4329 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4330 (nios2_compute_frame_layout):
4331 Add calculation of cfun->machine->fp_save_offset.
4332 (nios2_expand_prologue): Correct setting of frame pointer register
4334 (nios2_expand_epilogue): Update recovery of stack pointer from
4335 frame pointer accordingly.
4336 (nios2_initial_elimination_offset): Update calculation of offset
4337 for eliminating to HARD_FRAME_POINTER_REGNUM.
4339 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4342 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4343 cgraph_get_create_node on VAR_DECLs.
4345 2014-03-10 Richard Biener <rguenther@suse.de>
4348 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4350 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4352 * config/vms/vms.opt (vms_float_format): New variable.
4354 2014-03-08 Tobias Burnus <burnus@net-b.de>
4356 * doc/invoke.texi (-fcilkplus): Update implementation status.
4358 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4359 Richard Biener <rguenther@suse.de>
4361 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4362 consistently accross all TUs.
4363 (run_gcc): Enable -fshort-double automatically at link at link-time
4364 and disallow override.
4366 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4369 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4370 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4371 if they can't be used.
4373 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4375 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4376 for Solaris 11/x86 ld.
4377 * configure: Regenerate.
4379 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4381 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4382 (LIB_TLS_SPEC): Save as ld_tls_libs.
4383 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4384 (HAVE_AS_IX86_TLSLDM): New test.
4385 * configure, config.in: Regenerate.
4386 * config/i386/i386.c (legitimize_tls_address): Fall back to
4387 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4388 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4389 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4390 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4392 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4394 * common.opt (fira-loop-pressure): Mark as optimization.
4396 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4398 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4399 an OpenMP mappable type.
4401 2014-03-06 Matthias Klose <doko@ubuntu.com>
4403 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4404 MULTILIB_OSDIRNAMES is not defined.
4406 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4407 Meador Inge <meadori@codesourcery.com>
4410 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4411 (arm_legitimize_address): Call legitimize_tls_address for any
4412 arm_tls_referenced_p expression, handle constant addend. Call it
4413 before testing for !TARGET_ARM.
4414 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4416 2014-03-06 Richard Biener <rguenther@suse.de>
4422 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4424 * tree-streamer.c (record_common_node): Assert we don't record
4425 nodes with type double.
4426 (preload_common_node): Skip type double, complex double and double
4427 pointer since it is now frontend dependent due to fshort-double option.
4429 2014-03-06 Richard Biener <rguenther@suse.de>
4431 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4432 or -fno-lto is specified and the linker has full plugin support.
4433 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4434 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4435 * lto-wrapper.c (merge_and_complain): Merge compile-time
4436 optimization levels.
4437 (run_gcc): And pass it through to the link options.
4439 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4443 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4445 * cselib.c (remove_useless_values): Skip to avoid quadratic
4446 behavior if the condition moved from...
4447 (cselib_process_insn): ... here holds.
4449 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4452 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4453 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4456 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4457 (TM_H): Add x86-tune.def.
4459 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4461 * config/aarch64/aarch64.c (generic_tunings):
4462 Use cortexa57_extra_costs.
4464 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4467 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4468 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4469 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4472 2014-03-04 Heiher <r@hev.cc>
4474 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4475 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4477 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4479 * config/i386/predicates.md (const2356_operand): Change to ...
4480 (const2367_operand): ... this.
4481 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4483 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4484 (*avx512pf_scatterpf<mode>sf): Ditto.
4485 (avx512pf_scatterpf<mode>df): Ditto.
4486 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4487 (*avx512pf_scatterpf<mode>df): Ditto.
4488 * config/i386/i386.c (ix86_expand_builtin): Update
4489 incorrect hint operand error message.
4491 2014-03-04 Richard Biener <rguenther@suse.de>
4493 * lto-section-in.c (lto_get_section_data): Fix const cast.
4495 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4497 * tree-streamer.c (record_common_node): Assert we don't record
4498 nodes with type double.
4499 (preload_common_node): Skip type double, complex double and double
4500 pointer since it is now frontend dependent due to fshort-double option.
4502 2014-03-04 Richard Biener <rguenther@suse.de>
4505 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4506 (lto_input_toplevel_asms): Likewise.
4507 * lto-section-in.c (lto_get_section_data): Instead do it here
4510 2014-03-04 Richard Biener <rguenther@suse.de>
4512 PR tree-optimization/60382
4513 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4514 dead PHIs a reduction.
4516 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4518 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4520 (_mm_prefetch): Move out of GCC target("sse") pragma.
4521 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4522 GCC target("prfchw") pragma.
4523 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4525 * config/i386/i386.c (ix86_option_override_internal): Enable
4526 -mprfchw with -mprefetchwt1.
4528 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4530 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4533 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4535 * opts.h (CL_PCH_IGNORE): Define.
4536 * targhooks.c (option_affects_pch_p):
4537 Return false for options that have CL_PCH_IGNORE set.
4538 * opt-functions.awk: Process PchIgnore.
4539 * doc/options.texi: Document PchIgnore.
4541 * config/arc/arc.opt (misize): Add PchIgnore property.
4543 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4545 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4546 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4547 constraint on constants to permit them being loaded into
4548 GENERAL_REGS or BASE_REGS.
4550 2014-03-03 Nick Clifton <nickc@redhat.com>
4552 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4553 anti-cacnonical alternatives.
4554 (negandhi3_real): New pattern.
4555 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4557 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4559 * config/avr/avr-mcus.def: Remove atxmega16x1.
4560 * config/avr/avr-tables.opt: Regenerate.
4561 * config/avr/t-multilib: Regenerate.
4562 * doc/avr-mmcu.texi: Regenerate.
4564 2014-03-03 Tobias Grosser <tobias@grosser.es>
4565 Mircea Namolaru <mircea.namolaru@inria.fr>
4567 PR tree-optimization/58028
4568 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4571 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4573 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4574 not handled by recognizers.
4576 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4579 * function.c (expand_function_end): Don't emit
4580 clobber_return_register sequence if clobber_after is a BARRIER.
4581 * cfgexpand.c (construct_exit_block): Append instructions before
4582 return_label to prev_bb.
4584 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4586 * config/rs6000/constraints.md: Document reserved use of "wc".
4588 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4591 * ipa.c (function_and_variable_visibility): When dissolving comdat
4592 group, also set all symbols to local.
4594 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4599 2013-12-14 Jan Hubicka <jh@suse.cz>
4601 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4603 2014-03-02 Jon Beniston <jon@beniston.com>
4609 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4610 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4611 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4612 (simple_return, *simple_return): New patterns
4613 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4614 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4616 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4618 * dwarf2out.c (gen_subprogram_die): Tidy.
4620 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4623 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4624 (*mov_t_msb_neg_negc): ... this new insn.
4626 2014-02-28 Jason Merrill <jason@redhat.com>
4629 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4632 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4635 * dwarf2out.c (decltype_auto_die): New static.
4636 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4637 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4638 (is_cxx_auto): Likewise.
4640 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4642 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4643 we are not using general regs only.
4645 2014-02-28 Richard Biener <rguenther@suse.de>
4648 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4649 previous fix and only allow to remove trivial pre-headers
4650 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4651 (remove_forwarder_block): Properly update the latch of a loop.
4653 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4656 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4657 (cselib_preserved_hash_table): New.
4658 (preserve_constants_and_equivs): Move preserved vals to it.
4659 (cselib_find_slot): Look it up first.
4660 (cselib_init): Initialize it.
4661 (cselib_finish): Release it.
4662 (dump_cselib_table): Dump it.
4664 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4667 * cselib.c (remove_useless_values): Skip to avoid quadratic
4668 behavior if the condition moved from...
4669 (cselib_process_insn): ... here holds.
4671 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4674 * var-tracking.c (vt_initialize): Apply the same condition to
4675 preserve the CFA base value.
4677 2014-02-28 Joey Ye <joey.ye@arm.com>
4680 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4681 if reload in progress or completed.
4683 2014-02-28 Tobias Burnus <burnus@net-b.de>
4686 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4689 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4691 * doc/tm.texi.in (Condition Code Status): Update documention for
4692 relative locations of cc0-setter and cc0-user.
4694 2014-02-27 Jeff Law <law@redhat.com>
4696 PR rtl-optimization/52714
4697 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4698 into two independent simple sets, if I3 is a jump, ensure the
4699 pattern we place into I3 is a (set (pc) ...).
4701 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4702 Jeff Law <law@redhat.com>
4704 PR rtl-optimization/49847
4705 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4706 are in different blocks.
4707 * doc/tm.texi (Condition Code Status): Update documention for
4708 relative locations of cc0-setter and cc0-user.
4710 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4713 * lra.c (lra_emit_add): Check SUBREG too.
4715 2014-02-27 Andreas Schwab <schwab@suse.de>
4717 * config/m68k/m68k.c (m68k_option_override): Disable
4718 -flive-range-shrinkage for classic m68k.
4719 (m68k_override_options_after_change): Likewise.
4721 2014-02-27 Marek Polacek <polacek@redhat.com>
4724 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4725 -Wmaybe-uninitialized.
4727 2014-02-27 Alan Modra <amodra@gmail.com>
4730 * reload1.c (emit_input_reload_insns): When reload_override_in,
4731 set old to rl->in_reg when rl->in_reg is a subreg.
4733 2014-02-26 Richard Biener <rguenther@suse.de>
4736 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4738 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4740 * common/config/i386/predicates.md (const1256_operand): Remove.
4741 (const2356_operand): New.
4742 (const_1_to_2_operand): Remove.
4743 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4744 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4745 (*avx512pf_gatherpf<mode>sf): Ditto.
4746 (avx512pf_gatherpf<mode>df): Ditto.
4747 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4748 (*avx512pf_gatherpf<mode>df): Ditto.
4749 (avx512pf_scatterpf<mode>sf): Ditto.
4750 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4751 (*avx512pf_scatterpf<mode>sf): Ditto.
4752 (avx512pf_scatterpf<mode>df): Ditto.
4753 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4754 (*avx512pf_scatterpf<mode>df): Ditto.
4755 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4757 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4759 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4760 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4761 (_mm512_mask_testn_epi64_mask): Move to ...
4762 * config/i386/avx512cdintrin.h: Here.
4763 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4764 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4765 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4766 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4767 TARGET_AVX512F from TARGET_AVX512CD.
4769 2014-02-26 Richard Biener <rguenther@suse.de>
4772 * ipa.c (walk_polymorphic_call_targets): Properly guard
4773 call to inline_update_overall_summary.
4775 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4778 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4779 and latches only if requested. Fix latch if it is removed.
4780 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4781 LOOPS_HAVE_PREHEADERS.
4783 2014-02-25 Andrew Pinski <apinski@cavium.com>
4785 * builtins.c (expand_builtin_thread_pointer): Create a new target
4786 when the target is NULL.
4788 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4790 PR rtl-optimization/60317
4791 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4792 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4793 * lra-assigns.c: Include params.h.
4794 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4795 other reload pseudos considerations.
4797 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4799 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4800 to use canonical form for nor<mode>3.
4802 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4805 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4808 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4810 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4811 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4812 (ix86_handle_option): Handle OPT_mprefetchwt1.
4813 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4814 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4816 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4817 OPTION_MASK_ISA_PREFETCHWT1.
4818 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4819 (PTA_PREFETCHWT1): New.
4820 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4821 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4822 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4823 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4824 (*prefetch_avx512pf_<mode>_: Change into ...
4825 (*prefetch_prefetchwt1_<mode>: This.
4826 * config/i386/i386.opt (mprefetchwt1): New.
4827 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4828 (_mm_prefetch): Handle intent to write.
4829 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4831 2014-02-25 Richard Biener <rguenther@suse.de>
4834 * emit-rtl.c (mem_attrs_htab): Remove.
4835 (mem_attrs_htab_hash): Likewise.
4836 (mem_attrs_htab_eq): Likewise.
4837 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4838 (init_emit_once): Do not allocate mem_attrs_htab.
4840 2014-02-25 Richard Biener <rguenther@suse.de>
4843 * lto-opts.c (lto_write_options): Output non-explicit conservative
4844 -fwrapv, -fno-trapv and -fno-strict-overflow.
4845 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4846 (run_gcc): And pass them through.
4848 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4850 * sel-sched.c (calculate_new_fences): New parameter ptime.
4851 Calculate it as a maximum over all fence cycles.
4852 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4853 Print the final schedule timing when sched_verbose.
4855 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4857 PR rtl-optimization/60292
4858 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4859 bit fot the fence instruction.
4861 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4863 * calls.h: Fix typo in comment.
4865 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4867 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4868 adjusting offsetable addresses.
4870 2014-02-24 Guozhi Wei <carrot@google.com>
4872 * sparseset.h (sparseset_pop): Fix the wrong index.
4874 2014-02-24 Walter Lee <walt@tilera.com>
4876 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4877 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4879 * common/config/tilegx/tilegx-common.c
4880 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4881 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4883 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4884 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4885 (tilegx_gimplify_va_arg_expr): Handle big endian.
4886 (tilegx_expand_unaligned_load): Ditto.
4887 (tilegx_expand_unaligned_store): Ditto.
4888 (TARGET_RETURN_IN_MSB): New.
4889 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4890 (TARGET_ENDIAN_DEFAULT): New.
4891 (TARGET_BIG_ENDIAN): Handle big endian.
4892 (BYTES_BIG_ENDIAN): Ditto.
4893 (WORDS_BIG_ENDIAN): Ditto.
4894 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4897 * config/tilegx/tilegx.md (extv): Handle big endian.
4899 (insn_st<n>): Ditto.
4900 (insn_st<n>_add<bitsuffix>): Ditto.
4901 (insn_stnt<n>): Ditto.
4902 (insn_stnt<n>_add<bitsuffix>):Ditto.
4903 (vec_interleave_highv8qi): Handle big endian.
4904 (vec_interleave_highv8qi_be): New.
4905 (vec_interleave_highv8qi_le): New.
4906 (insn_v1int_h): Handle big endian.
4907 (vec_interleave_lowv8qi): Handle big endian.
4908 (vec_interleave_lowv8qi_be): New.
4909 (vec_interleave_lowv8qi_le): New.
4910 (insn_v1int_l): Handle big endian.
4911 (vec_interleave_highv4hi): Handle big endian.
4912 (vec_interleave_highv4hi_be): New.
4913 (vec_interleave_highv4hi_le): New.
4914 (insn_v2int_h): Handle big endian.
4915 (vec_interleave_lowv4hi): Handle big endian.
4916 (vec_interleave_lowv4hi_be): New.
4917 (vec_interleave_lowv4hi_le): New.
4918 (insn_v2int_l): Handle big endian.
4919 (vec_interleave_highv2si): Handle big endian.
4920 (vec_interleave_highv2si_be): New.
4921 (vec_interleave_highv2si_le): New.
4922 (insn_v4int_h): Handle big endian.
4923 (vec_interleave_lowv2si): Handle big endian.
4924 (vec_interleave_lowv2si_be): New.
4925 (vec_interleave_lowv2si_le): New.
4926 (insn_v4int_l): Handle big endian.
4927 * config/tilegx/tilegx.opt (mbig-endian): New option.
4928 (mlittle-endian): New option.
4929 * doc/install.texi: Document tilegxbe-linux.
4930 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4932 2014-02-24 Martin Jambor <mjambor@suse.cz>
4935 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4936 there are no parameter descriptors.
4938 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4940 PR rtl-optimization/60268
4941 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4942 initialization to ...
4943 (sched_rgn_init): ... here.
4944 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4946 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4948 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4951 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4953 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4956 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4958 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4959 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4961 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4963 * config/microblaze/predicates.md: Add cmp_op predicate.
4964 * config/microblaze/microblaze.md: Add branch_compare instruction
4965 which uses cmp_op predicate and emits cmp insn before branch.
4966 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4967 to microblaze_expand_conditional_branch and consolidate logic.
4968 (microblaze_expand_conditional_branch): emit branch_compare
4969 insn instead of handling cmp op separate from branch insn.
4971 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4973 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4976 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4978 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4979 define_insn with define_expand and new define_insn
4980 *altivec_lve<VI_char>x_internal.
4981 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4982 and new define_insn *altivec_stve<VI_char>x_internal.
4983 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4985 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4987 (altivec_expand_stvex_be): New function.
4989 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4991 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4992 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4993 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4994 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4996 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4999 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
5000 instead of emit_move_insn.
5002 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5004 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
5006 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
5007 gen_altivec_vsumsws.
5009 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5011 * config/rs6000/altivec.md (altivec_lvxl): Rename as
5012 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
5013 (altivec_lvxl_<mode>): New define_expand incorporating
5014 -maltivec=be semantics where needed.
5015 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
5016 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
5017 semantics where needed.
5018 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
5019 (altivec_stvx_<mode>): New define_expand incorporating
5020 -maltivec=be semantics where needed.
5021 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
5022 VM2 iterator instead of V4SI.
5023 (altivec_stvxl_<mode>): New define_expand incorporating
5024 -maltivec=be semantics where needed.
5025 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5026 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5027 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5028 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5029 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5030 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5031 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5032 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5033 ALTIVEC_BUILTIN_STVXL.
5034 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5035 (altivec_expand_stvx_be): Likewise.
5036 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5037 (altivec_expand_lvx_be): Likewise.
5038 (altivec_expand_stvx_be): Likewise.
5039 (altivec_expand_builtin): Add cases for
5040 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5041 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5042 (altivec_init_builtins): Add definitions for
5043 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5044 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5046 2014-02-21 Catherine Moore <clm@codesourcery.com>
5048 * doc/invoke.texi (mvirt, mno-virt): Document.
5049 * config/mips/mips.opt (mvirt): New option.
5050 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5052 2014-02-21 Richard Biener <rguenther@suse.de>
5054 PR tree-optimization/60276
5055 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5056 (STMT_VINFO_MIN_NEG_DIST): New macro.
5057 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5058 STMT_VINFO_MIN_NEG_DIST.
5059 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5060 made for negative dependence distances still hold.
5062 2014-02-21 Richard Biener <rguenther@suse.de>
5065 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5066 DECL_INITIAL for globals not in the current function context.
5068 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5070 PR tree-optimization/56490
5071 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5072 * tree-ssa-uninit.c: Include params.h.
5073 (compute_control_dep_chain): Add num_calls argument, return false
5074 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5075 num_calls to recursive call.
5076 (find_predicates): Change dep_chain into normal array,
5077 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5078 variable and adjust compute_control_dep_chain caller.
5079 (find_def_preds): Likewise.
5081 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5083 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5084 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5086 2014-02-21 Nick Clifton <nickc@redhat.com>
5088 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5089 (pushhi1): Likewise.
5090 (popqi1): Add mode to pre_dec.
5093 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5095 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5096 mode for mask of V8SFmode permutation.
5098 2014-02-20 Richard Henderson <rth@redhat.com>
5101 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5102 a new pseudo for OLDVAL.
5104 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5107 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5108 gen_reg_rtx if d->testing_p.
5109 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5110 if d->testing_p and we will certainly return true.
5111 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
5114 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
5116 * emit-rtl.c (gen_reg_rtx): Assert that
5117 crtl->emit.regno_pointer_align_length is non-zero.
5119 2014-02-20 Richard Henderson <rth@redhat.com>
5122 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
5123 on failure the store back into EXPECT.
5125 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
5126 Sandra Loosemore <sandra@codesourcery.com>
5128 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5129 * config/nios2/nios2.c (nios2_function_profiler): Add
5130 -fPIC (flag_pic == 2) support.
5131 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5132 (nios2_large_offset_p): New function.
5133 (nios2_unspec_reloc_p): Move up position, update to use
5134 nios2_large_offset_p.
5135 (nios2_unspec_address): Remove function.
5136 (nios2_unspec_offset): New function.
5137 (nios2_large_got_address): New function.
5138 (nios2_got_address): Add large offset support.
5139 (nios2_legitimize_tls_address): Update usage of removed and new
5141 (nios2_symbol_binds_local_p): New function.
5142 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5143 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5144 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5145 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5146 processing for (const (unspec ...)).
5147 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5149 2014-02-20 Richard Biener <rguenther@suse.de>
5151 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5152 doing the substitution.
5153 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5155 2014-02-20 Martin Jambor <mjambor@suse.cz>
5158 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5159 info when checking whether lattices are bottom.
5161 2014-02-20 Richard Biener <rguenther@suse.de>
5164 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5167 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5170 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5171 parameter specifying the scaling.
5172 (inline_call): Update.
5173 (want_inline_recursively): Guard division by zero.
5174 (recursive_inlining): Update.
5175 * ipa-inline.h (clone_inlined_nodes): Update.
5177 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5180 * config/i386/i386.c (classify_argument): Pass structures of size
5181 64 bytes or less in register.
5183 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5184 Kirill Yukhin <kirill.yukhin@intel.com>
5186 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5187 (_mm_rcp28_round_ss): Ditto.
5188 (_mm_rsqrt28_round_sd): Ditto.
5189 (_mm_rsqrt28_round_ss): Ditto.
5190 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5191 (_mm_rcp14_round_ss): Ditto.
5192 (_mm_rsqrt14_round_sd): Ditto.
5193 (_mm_rsqrt14_round_ss): Ditto.
5194 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5195 the first input operand, get rid of match_dup.
5196 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5198 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5200 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5201 operand as the first input operand, set type attribute.
5202 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5204 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5205 operand as the first input operand, set type attribute.
5207 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5209 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5212 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5215 * config/i386/i386.c (construct_container): Remove TFmode check
5216 for X86_64_INTEGER_CLASS.
5218 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5221 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5222 only when -Wpsabi is enabled.
5224 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5227 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5228 passing arrays in registers are the same as for structs, so remove the
5229 special case for them.
5231 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5233 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5234 destination type, extract only the valid bits if the source type is not
5235 integral and has a different mode.
5237 2014-02-19 Richard Biener <rguenther@suse.de>
5240 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5243 2014-02-19 Richard Biener <rguenther@suse.de>
5246 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5247 (ipa_modify_call_arguments): Emit an argument load explicitely and
5248 preserve virtual SSA form there and for the replacement call.
5249 Do not update SSA form nor free dominance info.
5251 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5253 * ipa.c (function_and_variable_visibility): Also clear WEAK
5254 flag when disolving COMDAT_GROUP.
5256 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5258 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5259 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5261 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5262 do more sanity checks.
5263 (detect_type_change): Return true when giving up early.
5264 (compute_complex_assign_jump_func): Fix type parameter of
5265 ipa_set_ancestor_jf.
5266 (compute_complex_ancestor_jump_func): Likewise.
5267 (update_jump_functions_after_inlining): Fix updating of
5269 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5271 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5273 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5274 inline clones when edge disappears.
5276 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5279 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5280 Split 64-bit moves into 2 patterns. Do not allow the use of
5281 direct move for TDmode in little endian, since the decimal value
5282 has little endian bytes within a word, but the 64-bit pieces are
5283 ordered in a big endian fashion, and normal subreg's of TDmode are
5285 (mov<mode>_64bit_dm): Likewise.
5286 (movtd_64bit_nodm): Likewise.
5288 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5290 PR tree-optimization/60174
5291 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5292 statement of an SSA_NAME that occurs in an abnormal PHI node.
5294 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5297 * final.c (SEEN_BB): Remove.
5298 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5299 (final_scan_insn): Don't force_source_line on second
5300 NOTE_INSN_BASIC_BLOCK.
5302 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5305 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5306 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5307 (type_natural_mode): Warn ABI change when %zmm register is not
5308 available for AVX512F vector value passing.
5310 2014-02-18 Kai Tietz <ktietz@redhat.com>
5313 * config/i386/i386.c (ix86_expand_prologue): Use value in
5314 rax register as displacement when restoring %r10 or %rax.
5315 Fix wrong offset when restoring both registers.
5317 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5319 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5320 assertion with conditional return.
5322 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5323 Uros Bizjak <ubizjak@gmail.com>
5326 * config/i386/driver-i386.c (host_detect_local_cpu): If
5327 YMM state is not saved by the OS, also clear has_f16c. Move
5328 CPUID 0x80000001 handling before YMM state saving checking.
5330 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5332 PR rtl-optimization/58960
5333 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5334 factored out from ...
5335 (sched_init): ... here.
5336 (free_global_sched_pressure_data): New, factored out from ...
5337 (sched_finish): ... here.
5338 * sched-int.h (free_global_sched_pressure_data): Declare.
5339 * sched-rgn.c (nr_regions_initial): New static global.
5340 (haifa_find_rgns): Initialize it.
5341 (schedule_region): Disable sched-pressure for the newly
5344 2014-02-17 Richard Biener <rguenther@suse.de>
5346 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5347 release SSA defs of pattern stmts.
5349 2014-02-17 Richard Biener <rguenther@suse.de>
5351 * tree-inline.c (expand_call_inline): Release the virtual
5352 operand defined by the call we are about to inline.
5354 2014-02-17 Richard Biener <rguenther@suse.de>
5356 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5358 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5359 Ilya Tocar <ilya.tocar@intel.com>
5361 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5362 arguments order in builtin.
5363 (_mm512_permutexvar_epi64): Ditto.
5364 (_mm512_mask_permutexvar_epi64): Ditto
5365 (_mm512_maskz_permutexvar_epi32): Ditto
5366 (_mm512_permutexvar_epi32): Ditto
5367 (_mm512_mask_permutexvar_epi32): Ditto
5369 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5371 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5372 (p8_vmrgow): Likewise.
5374 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5376 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5379 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5382 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5383 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5384 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5385 using direct move instructions on ISA 2.07. Also adjust
5386 instruction length for 64-bit.
5387 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5388 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5390 2014-02-15 Alan Modra <amodra@gmail.com>
5394 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5395 find_replacement on parts of insn rtl that might be reloaded.
5397 2014-02-15 Richard Biener <rguenther@suse.de>
5399 PR tree-optimization/60183
5400 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5401 (tree_ssa_phiprop): Calculate and free post-dominators.
5403 2014-02-14 Jeff Law <law@redhat.com>
5405 PR rtl-optimization/60131
5406 * ree.c (get_extended_src_reg): New function.
5407 (combine_reaching_defs): Use it rather than assuming location of REG.
5408 (find_and_remove_re): Verify first operand of extension is
5409 a REG before adding the insns to the copy list.
5411 2014-02-14 Roland McGrath <mcgrathr@google.com>
5413 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5414 * configure: Regenerated.
5415 * config.in: Regenerated.
5416 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5417 instead of ASM_SHORT.
5419 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5420 Richard Earnshaw <rearnsha@arm.com>
5422 PR rtl-optimization/59535
5423 * lra-constraints.c (process_alt_operands): Encourage alternative
5424 when unassigned pseudo class is superset of the alternative class.
5425 (inherit_reload_reg): Don't inherit when optimizing for code size.
5426 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5427 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5428 modes not less than 4 for Thumb1.
5430 2014-02-14 Kyle McMartin <kyle@redhat.com>
5433 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5435 2014-02-14 Richard Biener <rguenther@suse.de>
5437 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5438 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5439 Do not strip INDIRECT_REFs.
5441 2014-02-14 Richard Biener <rguenther@suse.de>
5444 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5445 DECL_FUNCTION_SPECIFIC_TARGET.
5446 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5447 * tree-streamer-out.c (pack_ts_target_option): Remove.
5448 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5449 (write_ts_function_decl_tree_pointers): Do not stream
5450 DECL_FUNCTION_SPECIFIC_TARGET.
5451 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5452 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5453 (lto_input_ts_function_decl_tree_pointers): Do not stream
5454 DECL_FUNCTION_SPECIFIC_TARGET.
5456 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5458 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5459 (get_initial_def_for_induction, vectorizable_induction): Ignore
5460 debug stmts when looking for exit_phi.
5461 (vectorizable_live_operation): Fix up condition.
5463 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5465 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5466 nreverse() because it changes the content of original tree list.
5468 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5470 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5471 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5473 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5475 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5476 GNU coding standards.
5478 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5481 * dwarf2out.c (gen_subprogram_die): Don't call
5482 add_calling_convention_attribute if subr_die is old_die.
5484 2014-02-13 Sharad Singhai <singhai@google.com>
5486 * doc/optinfo.texi: Fix order of nodes.
5488 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5490 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5491 operands[2], not operands[3].
5493 2014-02-13 Richard Biener <rguenther@suse.de>
5496 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5497 mention the possibility of an in-tree build.
5498 (CLooG): Update recommended version to 0.18.1, mention the
5499 possibility of an in-tree build and clarify that the ISL
5500 bundled with CLooG does not work.
5502 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5505 * expr.c (compress_float_constant): If x is a hard register,
5506 extend into a pseudo and then move to x.
5508 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5510 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5511 caused by bad second argument to warning_at() with -mhotpatch and
5512 nested functions (e.g. with gfortran).
5514 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5516 * opts.c (option_name): Remove "enabled by default" rider.
5518 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5520 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5522 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5523 Uros Bizjak <ubizjak@gmail.com>
5526 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5527 * configure: Regenerated.
5529 2014-02-12 Richard Biener <rguenther@suse.de>
5531 * vec.c (vec_prefix::calculate_allocation): Move as
5532 inline variant to vec.h.
5533 (vec_prefix::calculate_allocation_1): New out-of-line version.
5534 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5535 (vec_prefix::m_has_auto_buf): Rename to ...
5536 (vec_prefix::m_using_auto_storage): ... this.
5537 (vec_prefix::calculate_allocation): Inline the easy cases
5538 and dispatch to calculate_allocation_1 which doesn't need the
5540 (va_heap::reserve): Use gcc_checking_assert.
5541 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5542 m_using_auto_storage.
5543 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5545 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5546 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5547 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5549 2014-02-12 Richard Biener <rguenther@suse.de>
5551 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5552 when we found a dependence.
5554 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5556 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5558 (maybe_fold_stmt): ... into this new function.
5559 * omp-low.c (lower_omp): Update comment.
5561 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5564 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5567 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5569 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5570 identifiers in comments.
5571 (cortexa53_extra_costs): Likewise.
5572 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5573 (cortexa7_extra_costs): Likewise.
5574 (cortexa12_extra_costs): Likewise.
5575 (cortexa15_extra_costs): Likewise.
5576 (v7m_extra_costs): Likewise.
5578 2014-02-12 Richard Biener <rguenther@suse.de>
5581 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5582 of posix_memalign being successful.
5583 (lower_stmt): Restrict lowering of posix_memalign to when
5584 -ftree-bit-ccp is enabled.
5586 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5588 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5590 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5592 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5594 PR rtl-optimization/60116
5595 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5596 other_insn once the combination has been validated.
5598 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5601 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5603 * ipa-devirt.c: Include demangle.h
5604 (odr_violation_reported): New static variable.
5605 (add_type_duplicate): Update odr_violations.
5606 (maybe_record_node): Add completep parameter; update it.
5607 (record_target_from_binfo): Add COMPLETEP parameter;
5608 update it as needed.
5609 (possible_polymorphic_call_targets_1): Likewise.
5610 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5611 rename FINAL to COMPLETE.
5612 (record_targets_from_bases): Sanity check we found the binfo;
5613 fix COMPLETEP updating.
5614 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5615 parameter, fix computing of COMPLETEP.
5616 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5617 at LTO time do demangling.
5618 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5619 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5621 (gimple_get_virt_method_for_binfo): Likewise.
5622 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5623 gimple_get_virt_method_for_vtable): Update prototypes.
5625 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5628 * genautomata.c (add_presence_absence): Fix typo with
5629 {final_}presence_list.
5631 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5634 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5635 for VSX/Altivec vectors that land in GPR registers.
5637 2014-02-11 Richard Henderson <rth@redhat.com>
5638 Jakub Jelinek <jakub@redhat.com>
5641 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5642 around drhs if type conversion to lacc->type is not useless.
5644 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5646 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5648 (cortex-a57.cortex-a53): Likewise.
5649 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5651 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5653 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5656 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5658 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5659 add_options_for_arm_vfp3.
5661 2014-02-11 Jeff Law <law@redhat.com>
5664 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5665 object with an undesirable mode.
5667 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5670 * config/i386/sol2-9.h: New file.
5671 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5672 *-*-solaris2.9*): Use it.
5674 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5676 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5677 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5679 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5681 * config/microblaze/microblaze.c: Extend mcpu version format
5683 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5685 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5687 2014-02-10 Richard Henderson <rth@redhat.com>
5690 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5691 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5692 ms-abi vs -mno-accumulate-outgoing-args.
5693 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5694 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5697 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5700 * cfgexpand.c (expand_asm_operands): Attach source location to
5701 ASM_INPUT rtx objects.
5702 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5704 2014-02-10 Nick Clifton <nickc@redhat.com>
5706 * config/mn10300/mn10300.c (popcount): New function.
5707 (mn10300_expand_prologue): Include saved registers in stack usage
5710 2014-02-10 Jeff Law <law@redhat.com>
5713 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5714 when changing the SET_DEST of a prior insn to avoid an input reload.
5716 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5718 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5719 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5720 -mcall-openbsd, or -mcall-linux.
5721 (CC1_ENDIAN_BIG_SPEC): Remove.
5722 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5723 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5724 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5725 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5726 and %cc1_endian_default.
5727 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5729 2014-02-10 Richard Biener <rguenther@suse.de>
5731 PR tree-optimization/60115
5732 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5733 MEM_REF handling. Properly verify that the accesses are not
5734 out of the objects bound.
5736 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5738 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5741 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5743 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5744 proper constants and fix formatting.
5745 (possible_polymorphic_call_targets): Fix formatting.
5747 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5748 Ilya Tocar <ilya.tocar@intel.com>
5750 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5751 (_mm512_loadu_epi32): Renamed into...
5752 (_mm512_loadu_si512): This.
5753 (_mm512_storeu_epi32): Renamed into...
5754 (_mm512_storeu_si512): This.
5755 (_mm512_maskz_ceil_ps): Removed.
5756 (_mm512_maskz_ceil_pd): Ditto.
5757 (_mm512_maskz_floor_ps): Ditto.
5758 (_mm512_maskz_floor_pd): Ditto.
5759 (_mm512_floor_round_ps): Ditto.
5760 (_mm512_floor_round_pd): Ditto.
5761 (_mm512_ceil_round_ps): Ditto.
5762 (_mm512_ceil_round_pd): Ditto.
5763 (_mm512_mask_floor_round_ps): Ditto.
5764 (_mm512_mask_floor_round_pd): Ditto.
5765 (_mm512_mask_ceil_round_ps): Ditto.
5766 (_mm512_mask_ceil_round_pd): Ditto.
5767 (_mm512_maskz_floor_round_ps): Ditto.
5768 (_mm512_maskz_floor_round_pd): Ditto.
5769 (_mm512_maskz_ceil_round_ps): Ditto.
5770 (_mm512_maskz_ceil_round_pd): Ditto.
5771 (_mm512_expand_pd): Ditto.
5772 (_mm512_expand_ps): Ditto.
5773 * config/i386/i386.c (ix86_builtins): Remove
5774 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5775 (bdesc_args): Ditto.
5776 * config/i386/predicates.md (const1256_operand): New.
5777 (const_1_to_2_operand): Ditto.
5778 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5779 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5780 (*avx512pf_gatherpf<mode>sf): Ditto.
5781 (avx512pf_gatherpf<mode>df): Ditto.
5782 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5783 (*avx512pf_gatherpf<mode>df): Ditto.
5784 (avx512pf_scatterpf<mode>sf): Ditto.
5785 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5786 (*avx512pf_scatterpf<mode>sf): Ditto.
5787 (avx512pf_scatterpf<mode>df): Ditto.
5788 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5789 (*avx512pf_scatterpf<mode>df): Ditto.
5790 (avx512f_expand<mode>): Removed.
5791 (<shift_insn><mode>3<mask_name>): Change predicate type.
5793 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5795 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5796 not at the end of datarefs vector use ordered_remove to avoid
5797 reordering datarefs vector.
5800 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5801 mark local addressable non-static vars as GOVD_PRIVATE
5802 instead of GOVD_LOCAL.
5803 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5804 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5808 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5809 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5810 assume_aligned or alloc_align attributes.
5811 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5812 arguments. Handle also assume_aligned and alloc_align attributes.
5813 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5814 calls to functions with assume_aligned or alloc_align attributes.
5815 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5817 2014-02-08 Terry Guo <terry.guo@arm.com>
5819 * doc/invoke.texi: Document ARM -march=armv7e-m.
5821 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5823 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5824 flag on __cilkrts_rethrow builtin.
5827 * ipa-cp.c (determine_versionability): Fail at -O0
5828 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5829 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5832 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5835 * tree-inline.c (copy_forbidden): Fail for
5836 __attribute__((optimize (0))) functions.
5838 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5840 * varpool.c: Include pointer-set.h.
5841 (varpool_remove_unreferenced_decls): Variables in other partitions
5842 will not be output; be however careful to not lose information
5845 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5847 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5848 lookup in the vtable constructor.
5850 2014-02-07 Jeff Law <law@redhat.com>
5853 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5854 define_insn_and_split.
5856 * ipa-inline.c (inline_small_functions): Fix typos.
5858 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5860 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5861 (s390_can_use_return_insn): Declare.
5862 * config/s390/s390.h (EPILOGUE_USES): Define.
5863 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5865 (s390_chunkify_start): Handle return JUMP_LABELs.
5866 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5867 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5868 (s390_can_use_return_insn): New functions.
5869 (s390_fix_long_loop_prediction): Handle conditional returns.
5870 (TARGET_SET_UP_BY_PROLOGUE): Define.
5871 * config/s390/s390.md (ANY_RETURN): New code iterator.
5872 (*creturn, *csimple_return, return, simple_return): New patterns.
5874 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5876 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5877 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5878 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5879 REG_CFA_RESTORE list when deciding not to restore a register.
5881 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5883 * config/s390/s390.c: Include tree-pass.h and context.h.
5884 (s390_early_mach): New function, split out from...
5885 (s390_emit_prologue): ...here.
5886 (pass_data_s390_early_mach): New pass structure.
5887 (pass_s390_early_mach): New class.
5888 (s390_option_override): Create and register early_mach pass.
5889 Move to end of file.
5891 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5893 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5894 to match for the exit block.
5896 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5898 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5899 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5900 Reject misaligned operands.
5902 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5904 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5906 2014-02-07 Richard Biener <rguenther@suse.de>
5909 * gimple-low.c (lower_builtin_posix_memalign): New function.
5910 (lower_stmt): Call it to lower posix_memalign in a way
5911 to make alignment info accessible.
5913 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5916 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5917 __builtin_setjmp_receiver.
5919 2014-02-07 Richard Biener <rguenther@suse.de>
5922 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5923 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5924 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5925 Handle BUILT_IN_POSIX_MEMALIGN.
5926 (find_func_clobbers): Likewise.
5927 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5928 (call_may_clobber_ref_p_1): Likewise.
5930 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5933 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5936 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5939 * lto-cgraph.c (lto_output_node): Use
5940 symtab_get_symbol_partitioning_class.
5941 (lto_output_varpool_node): likewise.
5942 (symtab_get_symbol_partitioning_class): Move here from
5944 * cgraph.h (symbol_partitioning_class): Likewise.
5945 (symtab_get_symbol_partitioning_class): Declare.
5947 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5949 * ggc.h (ggc_internal_cleared_alloc): New macro.
5950 * vec.h (vec_safe_copy): Handle memory stats.
5951 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5952 * target-globals.c (save_target_globals): Likewise.
5954 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5957 * expr.c (emit_move_resolve_push): Export; be bit more selective
5958 on when to clear alias set.
5959 * expr.h (emit_move_resolve_push): Declare.
5960 * function.h (struct function): Add tail_call_marked.
5961 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5962 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5963 * config/i386/i386.md (TImode move expander): De not call
5965 (FP push expanders): Preserve memory attributes.
5966 * config/i386/sse.md (push<mode>1): Remove.
5967 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5968 (ix86_expand_push): Remove.
5969 * config/i386/mmx.md (push<mode>1): Remove.
5971 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5973 PR rtl-optimization/60030
5974 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5975 lopart with paradoxical subreg before shifting it up by hprec.
5977 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5979 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5980 Remove extra newline at end of file.
5981 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5982 (arm_issue_rate): Handle cortexa57.
5983 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5984 (cortex-a57.cortex-a53): Likewise.
5986 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5989 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5990 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5992 (arm_expand_prologue): Adjust all callers.
5993 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5994 info, registers also at the lowest numbered registers side. Use
5995 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5999 * var-tracking.c (adjust_mems): Before adding a SET to
6000 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
6002 2014-02-06 Alan Modra <amodra@gmail.com>
6005 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
6006 change SDmode to DDmode when lra_in_progress.
6008 2014-02-06 Jakub Jelinek <jakub@redhat.com>
6011 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
6012 free_data_ref on the dr first, and before goto again also set dr
6013 to the next dr. For simd_lane_access, free old datarefs[i] before
6014 overwriting it. For get_vectype_for_scalar_type failure, don't
6015 free_data_ref if simd_lane_access.
6017 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
6020 * tree.h (opts_for_fn): New inline function.
6021 (opt_for_fn): Define.
6022 * config/i386/i386.c (ix86_function_regparm): Use
6023 opt_for_fn (decl, optimize) instead of optimize.
6025 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6027 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6028 for SYMBOL_REF in large memory model.
6030 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6032 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6034 (cortex-a57): Likewise.
6035 (cortex-a57.cortex-a53): Likewise.
6037 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6038 Kugan Vivekanandarajah <kuganv@linaro.org>
6040 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6042 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6044 2014-02-06 Richard Biener <rguenther@suse.de>
6046 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6047 set_loop_copy and initialize_original_copy_tables.
6049 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6051 * config/aarch64/aarch64-simd.md
6052 (aarch64_ashr_simddi): Change QI to SI.
6054 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6055 Jakub Jelinek <jakub@redhat.com>
6058 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6061 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6063 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6064 CODE_FOR_altivec_vpku[hw]um to
6065 CODE_FOR_altivec_vpku[hw]um_direct.
6066 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6067 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6068 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6069 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6071 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6073 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6074 generation for -maltivec=be.
6075 (altivec_vsumsws): Simplify redundant test.
6077 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6079 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6080 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6081 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6082 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6083 gen_altivec_vpkuwum.
6084 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6086 (altivec_vpks<VI_char>ss): Likewise.
6087 (altivec_vpks<VI_char>us): Likewise.
6088 (altivec_vpku<VI_char>us): Likewise.
6089 (altivec_vpku<VI_char>um): Likewise.
6090 (altivec_vpku<VI_char>um_direct): New (copy of
6091 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6093 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6094 target is little endian and -maltivec=be is not specified.
6095 (*altivec_vupkhs<VU_char>_direct): New (copy of
6096 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6097 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6098 target is little endian and -maltivec=be is not specified.
6099 (*altivec_vupkls<VU_char>_direct): New (copy of
6100 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6101 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6102 little endian and -maltivec=be is not specified.
6103 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6104 little endian and -maltivec=be is not specified.
6106 2014-02-05 Richard Henderson <rth@redhat.com>
6109 * combine-stack-adj.c: Revert r206943.
6110 * sched-int.h (struct deps_desc): Add last_args_size.
6111 * sched-deps.c (init_deps): Initialize it.
6112 (sched_analyze_insn): Add OUTPUT dependencies between insns that
6113 contain REG_ARGS_SIZE notes.
6115 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6117 * lto-cgraph.c (asm_nodes_output): Make global.
6118 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
6119 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
6120 (driver_handle_option): Handle OPT_fwpa.
6122 2014-02-05 Jakub Jelinek <jakub@redhat.com>
6125 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
6126 a comment typo and formatting issue. If odr_hash hasn't been
6127 created, return vNULL and set *completep to false.
6130 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6131 bb with no successors.
6133 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6136 * doc/invoke.texi (-march): Clarify documentation for ARM.
6140 2014-02-05 Richard Biener <rguenther@suse.de>
6142 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6143 when not vectorizing because of too many alias checks.
6144 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6145 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6147 2014-02-05 Nick Clifton <nickc@redhat.com>
6149 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6150 accept extended registers in any mode when compiling for the MN10300.
6152 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6154 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6155 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6156 sanitization attributes.
6157 (can_inline_edge_p): Likewise.
6158 (sanitize_attrs_match_for_inline_p): New function.
6160 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6162 * ipa-prop.c (detect_type_change): Shor circuit testing of
6163 type changes on THIS pointer.
6165 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6168 * config/pa/pa.c (legitimize_tls_address): Return original address
6169 if not passed a SYMBOL_REF rtx.
6170 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6172 (pa_emit_move_sequence): Simplify TLS source operands.
6173 (pa_legitimate_constant_p): Reject all TLS constants.
6174 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6175 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6177 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6179 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6180 groups when we know they are controlled by LTO.
6181 * varasm.c (default_binds_local_p_1): If object is in other partition,
6182 it will be resolved locally.
6184 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6186 * config/host-linux.c (linux_gt_pch_use_address): Don't
6187 use SSIZE_MAX because it is not always defined.
6189 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6192 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6193 threshold for pseudo splitting.
6194 (update_ebb_live_info): Process call argument hard registers and
6195 hard registers from insn definition too.
6196 (max_small_class_regs_num): New constant.
6197 (inherit_in_ebb): Update live hard regs through EBBs. Update
6198 reloads_num only for small register classes. Don't split for
6201 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6204 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6207 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6209 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6212 2014-02-04 Marek Polacek <polacek@redhat.com>
6214 * gdbinit.in (pel): Define.
6216 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6218 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6221 2014-02-04 Richard Biener <rguenther@suse.de>
6224 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6225 in function context local.
6226 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6227 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6228 similar to LTO_imported_decl_ref.
6230 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6232 PR tree-optimization/60002
6233 * cgraphclones.c (build_function_decl_skip_args): Clear
6236 PR tree-optimization/60023
6237 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6238 false to gsi_replace.
6239 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6240 has been in some EH region and vec_stmt could throw, add
6241 vec_stmt into the same EH region.
6242 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6243 has no lhs, ignore it.
6244 * internal-fn.c (expand_MASK_LOAD): Likewise.
6247 * tree-inline.c (copy_forbidden): Fail for
6248 __attribute__((optimize (0))) functions.
6251 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6252 is set, copy one less argument.
6253 (expand_simd_clones): Don't subtract clone_info->inbranch
6254 from simd_clone_struct_alloc argument.
6256 PR rtl-optimization/57915
6257 * recog.c (simplify_while_replacing): If all unary/binary/relational
6258 operation arguments are constant, attempt to simplify those.
6261 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6262 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6264 2014-02-04 Richard Biener <rguenther@suse.de>
6266 PR tree-optimization/60012
6267 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6268 TBAA disambiguation to all DDRs.
6270 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6273 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6274 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6276 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6279 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6281 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6283 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6284 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6286 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6289 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6290 to figure out targets of polymorphic calls with known decl.
6291 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6292 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6293 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6294 (get_polymorphic_call_info): ... here.
6295 (get_polymorphic_call_info_from_invariant): New function.
6297 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6299 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6300 lookup via vtable pointer; check for type consistency
6301 and turn inconsitent facts into UNREACHABLE.
6302 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6303 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6304 type inconsistent querries; return UNREACHABLE instead.
6306 2014-02-03 Richard Henderson <rth@twiddle.net>
6309 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6310 already processed this node.
6311 (normalize_one_pred_1): Pass along mark_set.
6312 (normalize_one_pred): Create and destroy a pointer_set_t.
6313 (normalize_one_pred_chain): Likewise.
6315 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6317 PR gcov-profile/58602
6318 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6320 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6323 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6324 -fno-devirtualize; try to devirtualize by the knowledge of
6325 virtual table pointer given by aggregate propagation.
6326 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6327 (ipa_print_node_jump_functions): Dump also offset that
6328 is relevant for polymorphic calls.
6329 (determine_known_aggregate_parts): Add arg_type parameter; use it
6330 instead of determining the type from pointer type.
6331 (ipa_compute_jump_functions_for_edge): Update call of
6332 determine_known_aggregate_parts.
6333 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6334 (gimple_get_virt_method_for_binfo): ... here; simplify using
6335 vtable_pointer_value_to_vtable.
6336 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6337 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6338 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6339 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6341 (vtable_pointer_value_to_binfo): ... here.
6342 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6344 2014-02-03 Teresa Johnson <tejohnson@google.com>
6346 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6347 redef of outer loop index variable.
6349 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6353 * doc/extend.texi (Function Attributes): Typo.
6355 2014-02-03 Cong Hou <congh@google.com>
6357 PR tree-optimization/60000
6358 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6359 if the vectorized statement is a store. A store statement can only
6360 appear at the end of pattern statements.
6362 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6364 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6365 (ix86_option_override_internal): Default long double to 64-bit for
6366 32-bit Bionic and to 128-bit for 64-bit Bionic.
6368 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6369 TARGET_LONG_DOUBLE_128 is true.
6370 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6372 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6373 (mlong-double-64): Negate -mlong-double-128.
6374 (mlong-double-128): New option.
6376 * config/i386/i386-c.c (ix86_target_macros): Define
6377 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6379 * doc/invoke.texi: Document -mlong-double-128.
6381 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6383 PR rtl-optimization/60024
6384 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6386 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6388 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6390 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6392 PR rtl-optimization/57662
6393 * sel-sched.c (code_motion_path_driver): Do not mark already not
6394 existing blocks in the visiting bitmap.
6396 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6398 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6399 on the insn being emitted.
6401 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6402 Will Deacon <will.deacon@arm.com>
6404 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6406 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6408 * config/arm/arm-tables.opt: Regenerate.
6410 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6412 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6413 for vector types other than V16QImode.
6414 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6415 define_expand, and call altivec_expand_vec_perm_le when producing
6416 code with little endian element order.
6417 (*altivec_vperm_<mode>_internal): New insn having previous
6418 behavior of altivec_vperm_<mode>.
6419 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6420 altivec_expand_vec_perm_le when producing code with little endian
6422 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6423 behavior of altivec_vperm_<mode>_uns.
6425 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6427 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6428 (altivec_vsumsws): Add handling for -maltivec=be with a little
6430 (altivec_vsumsws_direct): New.
6431 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6432 gen_altivec_vsumsws.
6434 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6436 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6437 vtable_pointer_value_to_binfo): New functions.
6438 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6439 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6441 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6443 * config/nios2/nios2.md (load_got_register): Initialize GOT
6444 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6445 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6447 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6449 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6450 preserverd by passthrough, do not propagate the type.
6452 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6454 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6455 (mips_atomic_assign_expand_fenv): New function.
6456 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6458 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6460 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6461 (__builtin_mips_set_fcsr): Likewise.
6462 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6463 MIPS_USI_FTYPE_VOID.
6464 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6465 (mips16_expand_set_fcsr): Likewise.
6466 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6467 (mips16_set_fcsr_stub): Likewise.
6468 (mips16_get_fcsr_one_only_stub): New class.
6469 (mips16_set_fcsr_one_only_stub): Likewise.
6470 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6471 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6472 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6473 (hard_float): New availability predicate.
6474 (mips_builtins): Add get_fcsr and set_fcsr.
6475 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6476 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6477 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6478 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6479 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6482 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6484 * config/mips/mips.c (mips_one_only_stub): New class.
6485 (mips_need_mips16_rdhwr_p): Replace with...
6486 (mips16_rdhwr_stub): ...this new variable.
6487 (mips16_stub_call_address): New function.
6488 (mips16_rdhwr_one_only_stub): New class.
6489 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6490 (mips_output_mips16_rdhwr): Delete.
6491 (mips_finish_stub): New function.
6492 (mips_code_end): Use it to handle rdhwr stubs.
6494 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6497 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6498 when calculating size of integer atomic types.
6500 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6502 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6504 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6506 PR tree-optimization/60003
6507 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6508 * profile.c (branch_prob): Use gimple_call_builtin_p
6509 to check for BUILT_IN_SETJMP_RECEIVER.
6510 * tree-inline.c (copy_bb): Call notice_special_calls.
6512 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6515 * lra-constraints.c (process_alt_operands): Update reload_sum only
6518 2014-01-31 Richard Henderson <rth@redhat.com>
6521 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6522 until after else_eh is processed.
6524 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6526 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6527 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6528 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6529 in smmintrin.h, remove them.
6530 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6531 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6532 * config/i386/i386.md (ROUND_SAE): Fix value.
6533 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6534 (const48_operand): New.
6535 * config/i386/subst.md (round), (round_expand): Use
6536 const_4_or_8_to_11_operand.
6537 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6539 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6541 * config/i386/constraints.md (Yk): Swap meaning with k.
6542 * config/i386/i386.md (movhi_internal): Change Yk to k.
6543 (movqi_internal): Ditto.
6544 (*k<logic><mode>): Ditto.
6547 (kandn<mode>): Ditto.
6548 (*<code>hi_1): Ditto.
6549 (*<code>qi_1): Ditto.
6550 (kxnor<mode>): Ditto.
6551 (kortestzhi): Ditto.
6552 (kortestchi): Ditto.
6554 (*one_cmplhi2_1): Ditto.
6555 (*one_cmplqi2_1): Ditto.
6556 * config/i386/sse.md (): Change k to Yk.
6557 (avx512f_load<mode>_mask): Ditto.
6558 (avx512f_blendm<mode>): Ditto.
6559 (avx512f_store<mode>_mask): Ditto.
6560 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6561 (avx512f_storedqu<mode>_mask): Ditto.
6562 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6564 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6565 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6566 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6567 (avx512f_maskcmp<mode>3): Ditto.
6568 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6569 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6570 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6571 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6572 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6573 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6574 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6575 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6576 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6577 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6578 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6579 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6580 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6581 (vec_extract_lo_<mode>_maskm): Ditto.
6582 (vec_extract_hi_<mode>_maskm): Ditto.
6583 (avx512f_vternlog<mode>_mask): Ditto.
6584 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6585 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6586 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6587 (avx512f_<code>v8div16qi2_mask): Ditto.
6588 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6589 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6590 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6591 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6592 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6593 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6594 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6595 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6596 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6597 (avx512cd_maskb_vec_dupv8di): Ditto.
6598 (avx512cd_maskw_vec_dupv16si): Ditto.
6599 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6600 (avx512f_vpermi2var<mode>3_mask): Ditto.
6601 (avx512f_vpermi2var<mode>3_mask): Ditto.
6602 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6603 (*avx512f_gathersi<mode>): Ditto.
6604 (*avx512f_gathersi<mode>_2): Ditto.
6605 (*avx512f_gatherdi<mode>): Ditto.
6606 (*avx512f_gatherdi<mode>_2): Ditto.
6607 (*avx512f_scattersi<mode>): Ditto.
6608 (*avx512f_scatterdi<mode>): Ditto.
6609 (avx512f_compress<mode>_mask): Ditto.
6610 (avx512f_compressstore<mode>_mask): Ditto.
6611 (avx512f_expand<mode>_mask): Ditto.
6612 * config/i386/subst.md (mask): Change k to Yk.
6613 (mask_scalar_merge): Ditto.
6616 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6618 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6620 2014-01-31 Richard Biener <rguenther@suse.de>
6623 * builtins.c (fold_builtin_memory_op): Make sure to not
6624 use a floating-point mode or a boolean or enumeral type for
6627 2014-01-30 DJ Delorie <dj@redhat.com>
6629 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6630 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6631 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6632 whenever main() has an epilogue.
6634 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6636 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6637 unused variable "field".
6638 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6639 (vsx_mergeh_<mode>): Likewise.
6640 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6641 (altivec_vmrghh): Likewise.
6642 (altivec_vmrghw): Likewise.
6643 (altivec_vmrglb): Likewise.
6644 (altivec_vmrglh): Likewise.
6645 (altivec_vmrglw): Likewise.
6646 (altivec_vspltb): Add missing uses.
6647 (altivec_vsplth): Likewise.
6648 (altivec_vspltw): Likewise.
6649 (altivec_vspltsf): Likewise.
6651 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6654 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6655 frame related instructions.
6657 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6659 PR rtl-optimization/59959
6660 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6661 any reload of register whose subreg is invalid.
6663 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6665 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6666 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6667 Add missing return type - void.
6669 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6671 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6672 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6673 remove element index adjustment for endian (now handled in vsx.md
6675 (altivec_expand_vec_perm_const): Use
6676 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6677 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6678 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6679 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6680 define_expand and a new define_insn *altivec_vspltb_internal;
6681 adjust for -maltivec=be on a little endian target.
6682 (altivec_vspltb_direct): New.
6683 (altivec_vsplth): Divide into a define_expand and a new
6684 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6685 little endian target.
6686 (altivec_vsplth_direct): New.
6687 (altivec_vspltw): Divide into a define_expand and a new
6688 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6689 little endian target.
6690 (altivec_vspltw_direct): New.
6691 (altivec_vspltsf): Divide into a define_expand and a new
6692 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6693 a little endian target.
6695 2014-01-30 Richard Biener <rguenther@suse.de>
6697 PR tree-optimization/59993
6698 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6699 can propagate form the earlier stmt and avoid the transform
6700 when the intermediate result is needed.
6702 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6704 * README.Portability: Fix typo.
6706 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6708 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6709 comparison_operator with ordered_comparison_operator.
6711 2014-01-30 Nick Clifton <nickc@redhat.com>
6713 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6714 Rename to mn10300_store_multiple_regs.
6715 * config/mn10300/mn10300.c: Likewise.
6716 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6717 store_multiple_regs.
6718 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6719 Call mn10300_store_multiple_regs.
6721 2014-01-30 Nick Clifton <nickc@redhat.com>
6722 DJ Delorie <dj@redhat.com>
6724 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6725 %fp 2 to keep registers after it properly word-aligned.
6726 (rl78_alloc_physical_registers_umul): Handle the case where both
6727 input operands are the same.
6729 2014-01-30 Richard Biener <rguenther@suse.de>
6731 PR tree-optimization/59903
6732 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6735 2014-01-30 Jason Merrill <jason@redhat.com>
6738 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6741 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6743 2014-01-30 Richard Biener <rguenther@suse.de>
6745 PR tree-optimization/59951
6746 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6748 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6751 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6752 SFmode to DFmode case.
6754 2014-01-29 DJ Delorie <dj@redhat.com>
6756 * config/msp430/msp430.opt (-minrt): New.
6757 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6759 (ENDFILE_SPEC): Likewise.
6761 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6763 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6764 (estimate_function_body_sizes): Use it.
6766 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6769 * dwarf2out.c (is_cxx_auto): New.
6770 (is_base_type): Use it.
6771 (gen_type_die_with_usage): Likewise.
6773 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6775 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6776 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6777 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6778 -maltivec=be with LE targets.
6779 (vsx_mergeh_<mode>): Likewise.
6780 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6781 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6782 (altivec_vmrghb): Replace with define_expand and new
6783 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6784 (altivec_vmrghb_direct): New define_insn.
6785 (altivec_vmrghh): Replace with define_expand and new
6786 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6787 (altivec_vmrghh_direct): New define_insn.
6788 (altivec_vmrghw): Replace with define_expand and new
6789 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6790 (altivec_vmrghw_direct): New define_insn.
6791 (*altivec_vmrghsf): Adjust for endianness.
6792 (altivec_vmrglb): Replace with define_expand and new
6793 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6794 (altivec_vmrglb_direct): New define_insn.
6795 (altivec_vmrglh): Replace with define_expand and new
6796 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6797 (altivec_vmrglh_direct): New define_insn.
6798 (altivec_vmrglw): Replace with define_expand and new
6799 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6800 (altivec_vmrglw_direct): New define_insn.
6801 (*altivec_vmrglsf): Adjust for endianness.
6802 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6803 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6804 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6805 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6806 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6807 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6808 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6809 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6811 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6813 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6814 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6817 2014-01-29 Richard Biener <rguenther@suse.de>
6819 PR tree-optimization/58742
6820 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6821 associate_pointerplus_align.
6822 (associate_pointerplus_diff): New function.
6823 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6824 and associate_pointerplus_diff.
6826 2014-01-29 Richard Biener <rguenther@suse.de>
6828 * lto-streamer.h (LTO_major_version): Bump to 3.
6829 (LTO_minor_version): Reset to 0.
6831 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6833 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6834 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6835 (arm_file_start): Generate correct asm header for armv7ve.
6836 * config/arm/bpabi.h: Add multilib support for armv7ve.
6837 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6838 and cortex-a15 to armv7ve.
6839 * config/arm/t-aprofile: Add multilib support for armv7ve.
6840 * doc/invoke.texi: Document -march=armv7ve.
6842 2014-01-29 Richard Biener <rguenther@suse.de>
6844 PR tree-optimization/58742
6845 * tree-ssa-forwprop.c (associate_plusminus): Return true
6846 if we changed sth, defer EH cleanup to ...
6847 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6848 (simplify_mult): New function.
6850 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6853 PR tree-optimization/59920
6854 * tree.c (build_common_builtin_nodes): Remove
6855 __builtin_setjmp_dispatcher initialization.
6856 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6857 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6858 instead of gsi_after_labels + manually skipping debug stmts.
6859 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6860 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6861 * tree-inline.c (copy_edges_for_bb): Remove
6862 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6863 argument. Ignore computed_goto_p stmts. Don't call
6864 make_abnormal_goto_edges. If a call might need abnormal edges
6865 for non-local gotos, see if it already has an edge to
6866 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6867 with true argument, don't do anything then, otherwise add
6868 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6869 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6871 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6872 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6873 (lower_stmt): Don't set data->calls_builtin_setjmp.
6874 (lower_builtin_setjmp): Adjust comment.
6875 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6876 * tree-cfg.c (found_computed_goto): Remove.
6877 (factor_computed_gotos): Remove.
6878 (make_goto_expr_edges): Return bool, true for computed gotos.
6879 Don't call make_abnormal_goto_edges.
6880 (build_gimple_cfg): Don't set found_computed_goto, don't call
6881 factor_computed_gotos.
6882 (computed_goto_p): No longer static.
6883 (make_blocks): Don't set found_computed_goto.
6884 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6885 (make_edges): If make_goto_expr_edges returns true, push bb
6886 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6887 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6888 vector. Record mapping between bbs and OpenMP regions if there
6889 are any, adjust make_gimple_omp_edges caller. Call
6890 handle_abnormal_edges.
6891 (make_abnormal_goto_edges): Remove.
6892 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6893 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6894 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6895 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6896 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6897 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6898 filling *region also set *region_idx to (*region)->entry->index.
6901 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6902 For REGs set ORIGINAL_REGNO.
6904 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6906 * doc/md.texi: Mention that a target shouldn't implement
6907 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6910 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6912 PR tree-optimization/59594
6913 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6914 a copy of the datarefs vector rather than the vector itself.
6916 2014-01-28 Jason Merrill <jason@redhat.com>
6919 * dwarf2out.c (auto_die): New static.
6920 (gen_type_die_with_usage): Handle C++1y 'auto'.
6921 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6924 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6927 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6928 (SPEC_X32): Likewise.
6929 (SPEC_64): Likewise.
6930 * config/i386/i386.c (ix86_option_override_internal): Turn off
6931 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6933 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6934 * config/i386/i386.h (TARGET_16BIT): New macro.
6935 (TARGET_16BIT_P): Likewise.
6936 * config/i386/i386.opt: Add m16.
6937 * doc/invoke.texi: Document -m16.
6939 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6941 PR preprocessor/59935
6942 * input.c (location_get_source_line): Bail out on when line number
6943 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6945 2014-01-28 Richard Biener <rguenther@suse.de>
6947 PR tree-optimization/58742
6948 * tree-ssa-forwprop.c (associate_plusminus): Handle
6949 pointer subtraction of the form (T)(P + A) - (T)P.
6951 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6953 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6956 2014-01-28 Richard Biener <rguenther@suse.de>
6959 2014-01-28 Richard Biener <rguenther@suse.de>
6961 PR rtl-optimization/45364
6962 PR rtl-optimization/59890
6963 * var-tracking.c (local_get_addr_clear_given_value): Handle
6964 already cleared slot.
6965 (val_reset): Handle not allocated local_get_addr_cache.
6966 (vt_find_locations): Use post-order on the inverted CFG.
6968 2014-01-28 Richard Biener <rguenther@suse.de>
6970 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6972 2014-01-28 Richard Biener <rguenther@suse.de>
6974 PR rtl-optimization/45364
6975 PR rtl-optimization/59890
6976 * var-tracking.c (local_get_addr_clear_given_value): Handle
6977 already cleared slot.
6978 (val_reset): Handle not allocated local_get_addr_cache.
6979 (vt_find_locations): Use post-order on the inverted CFG.
6981 2014-01-28 Alan Modra <amodra@gmail.com>
6983 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6984 * configure.ac <recursive call for build != host>: Define
6985 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6986 and LD_FOR_BUILD too.
6987 * configure: Regenerate.
6989 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6991 * config/i386/i386.c (get_builtin_code_for_version): Separate
6992 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6993 Broadwell from Haswell.
6995 2014-01-27 Steve Ellcey <sellcey@mips.com>
6997 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6998 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6999 * config/mips/mips.c (mips_option_override): Change setting
7001 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
7002 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
7003 Change from Mask to Var.
7005 2014-01-27 Jeff Law <law@redhat.com>
7007 * ipa-inline.c (inline_small_functions): Fix typo.
7009 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
7011 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
7012 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
7013 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
7014 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
7015 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
7016 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
7017 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
7018 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
7019 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
7020 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
7021 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
7022 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
7023 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
7024 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
7025 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7026 (_mm512_storeu_epi64): Ditto.
7027 (_mm512_cmpge_epi32_mask): Ditto.
7028 (_mm512_cmpge_epu32_mask): Ditto.
7029 (_mm512_cmpge_epi64_mask): Ditto.
7030 (_mm512_cmpge_epu64_mask): Ditto.
7031 (_mm512_cmple_epi32_mask): Ditto.
7032 (_mm512_cmple_epu32_mask): Ditto.
7033 (_mm512_cmple_epi64_mask): Ditto.
7034 (_mm512_cmple_epu64_mask): Ditto.
7035 (_mm512_cmplt_epi32_mask): Ditto.
7036 (_mm512_cmplt_epu32_mask): Ditto.
7037 (_mm512_cmplt_epi64_mask): Ditto.
7038 (_mm512_cmplt_epu64_mask): Ditto.
7039 (_mm512_cmpneq_epi32_mask): Ditto.
7040 (_mm512_cmpneq_epu32_mask): Ditto.
7041 (_mm512_cmpneq_epi64_mask): Ditto.
7042 (_mm512_cmpneq_epu64_mask): Ditto.
7043 (_mm512_expand_pd): Ditto.
7044 (_mm512_expand_ps): Ditto.
7045 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7046 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7047 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7048 * config/i386/i386.c (ix86_builtins): Add
7049 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7050 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7051 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7052 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7053 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7054 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7055 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7056 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7057 IX86_BUILTIN_PMOVUSQW512_MEM.
7058 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7059 __builtin_ia32_pmovsqd512mem_mask,
7060 __builtin_ia32_pmovqd512mem_mask,
7061 __builtin_ia32_pmovusqw512mem_mask,
7062 __builtin_ia32_pmovsqw512mem_mask,
7063 __builtin_ia32_pmovqw512mem_mask,
7064 __builtin_ia32_pmovusdw512mem_mask,
7065 __builtin_ia32_pmovsdw512mem_mask,
7066 __builtin_ia32_pmovdw512mem_mask,
7067 __builtin_ia32_pmovqb512mem_mask,
7068 __builtin_ia32_pmovusqb512mem_mask,
7069 __builtin_ia32_pmovsqb512mem_mask,
7070 __builtin_ia32_pmovusdb512mem_mask,
7071 __builtin_ia32_pmovsdb512mem_mask,
7072 __builtin_ia32_pmovdb512mem_mask.
7073 (bdesc_args): Add __builtin_ia32_expanddf512,
7074 __builtin_ia32_expandsf512.
7075 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7076 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7077 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7078 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7079 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7080 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7081 (avx512f_<code>v8div16qi2_mask_store): This.
7082 (avx512f_expand<mode>): New.
7084 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7086 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7088 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7089 (_mm512_prefetch_i32scatter_pd): Ditto.
7090 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7091 (_mm512_prefetch_i64scatter_pd): Ditto.
7092 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7093 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7094 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7095 (_mm512_prefetch_i32scatter_ps): Ditto.
7096 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7097 (_mm512_prefetch_i64scatter_ps): Ditto.
7098 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7099 * config/i386/i386-builtin-types.def: Define
7100 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7101 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7102 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7103 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7104 IX86_BUILTIN_SCATTERPFQPD.
7105 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7106 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7107 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7108 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7109 __builtin_ia32_scatterpfqps.
7110 (ix86_expand_builtin): Expand new built-ins.
7111 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
7112 fix memory access data type.
7113 (*avx512pf_gatherpf<mode>_mask): Ditto.
7114 (*avx512pf_gatherpf<mode>): Ditto.
7115 (avx512pf_scatterpf<mode>): Ditto.
7116 (*avx512pf_scatterpf<mode>_mask): Ditto.
7117 (*avx512pf_scatterpf<mode>): Ditto.
7118 (GATHER_SCATTER_SF_MEM_MODE): New.
7119 (avx512pf_gatherpf<mode>df): Ditto.
7120 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7121 (*avx512pf_scatterpf<mode>df): Ditto.
7123 2014-01-27 Jakub Jelinek <jakub@redhat.com>
7126 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7127 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7130 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7132 * common/config/arm/arm-common.c
7133 (arm_rewrite_mcpu): Handle multiple names.
7135 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7137 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7139 * gimple-builder.h (create_gimple_tmp): Delete.
7141 2014-01-27 Christian Bruel <christian.bruel@st.com>
7143 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7146 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7148 * config/pa/pa.md (call): Generate indirect long calls to non-local
7149 functions when outputing 32-bit code.
7150 (call_value): Likewise except for special call to buggy powf function.
7152 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7153 portable runtime and PIC indirect calls.
7154 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7155 and PIC call sequences. Use ldo instead of blr to set return register
7156 in PIC call sequence.
7158 2014-01-25 Walter Lee <walt@tilera.com>
7160 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7161 avoid clobbering a live register.
7163 2014-01-25 Walter Lee <walt@tilera.com>
7165 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7166 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7167 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7168 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7170 2014-01-25 Walter Lee <walt@tilera.com>
7172 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7173 arguments on even registers.
7174 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7176 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7177 (BIGGEST_ALIGNMENT): Ditto.
7178 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7180 2014-01-25 Walter Lee <walt@tilera.com>
7182 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7183 insns before bundling.
7184 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7186 2014-01-25 Walter Lee <walt@tilera.com>
7188 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7189 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7190 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7192 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7194 * config/mips/constraints.md (kl): Delete.
7195 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7196 define expands, using...
7197 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7198 instructions for MIPS16.
7199 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7200 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7202 2014-01-25 Walter Lee <walt@tilera.com>
7204 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7208 2014-01-25 Walter Lee <walt@tilera.com>
7210 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7211 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7213 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7215 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7218 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7220 * print-rtl.c (in_call_function_usage): New var.
7221 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7222 EXPR_LIST mode as mode and not as reg note name.
7225 * cfgloopmanip.c (copy_loop_info): If
7226 loop->warned_aggressive_loop_optimizations, make sure
7227 the flag is set in target loop too.
7229 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7231 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7233 * builtins.def: Likewise.
7234 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7235 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7236 * ira.c (ira_setup_eliminable_regset): Likewise.
7237 * omp-low.c (gate_expand_omp): Likewise.
7238 (execute_lower_omp): Likewise.
7239 (diagnose_sb_0): Likewise.
7240 (gate_diagnose_omp_blocks): Likewise.
7241 (simd_clone_clauses_extract): Likewise.
7244 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7246 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7247 correction for little endian...
7248 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7251 2014-01-24 Jeff Law <law@redhat.com>
7253 PR tree-optimization/59919
7254 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7255 for non-returning calls.
7257 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7259 * common/config/aarch64/aarch64-common.c
7260 (aarch64_rewrite_mcpu): Handle multiple names.
7261 * config/aarch64/aarch64.h
7262 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7264 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7266 * input.c (add_file_to_cache_tab): Handle the case where fopen
7269 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7272 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7273 from push operand if code of push isn't PRE_DEC.
7275 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7278 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7279 -mquad-memory-atomic. Update -mquad-memory documentation to say
7280 it is only used for non-atomic loads/stores.
7282 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7283 -mquad-memory or -mquad-memory-atomic switches.
7285 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7286 -mquad-memory-atomic to ISA 2.07 support.
7288 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7289 to separate support of normal quad word memory operations (ldq, stq)
7290 from the atomic quad word memory operations.
7292 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7293 support to separate non-atomic quad word operations from atomic
7294 quad word operations. Disable non-atomic quad word operations in
7295 little endian mode so that we don't have to swap words after the
7296 load and before the store.
7297 (quad_load_store_p): Add comment about atomic quad word support.
7298 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7299 options printed with -mdebug=reg.
7301 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7302 -mquad-memory-atomic as the test for whether we have quad word
7303 atomic instructions.
7304 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7305 or -mp8-vector are used, allow byte/half-word atomic operations.
7307 * config/rs6000/sync.md (load_lockedti): Insure that the address
7308 is a proper indexed or indirect address for the lqarx instruction.
7309 On little endian systems, swap the hi/lo registers after the lqarx
7311 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7312 insure the address is valid for the lqarx instruction.
7313 (store_conditionalti): Insure that the address is a proper indexed
7314 or indirect address for the stqcrx. instruction. On little endian
7315 systems, swap the hi/lo registers before doing the stqcrx.
7317 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7318 insure the address is valid for the stqcrx. instruction.
7320 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7321 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7322 type of quad memory support is available.
7324 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7327 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7328 there is a danger of looping.
7330 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7332 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7333 force flag_ira_loop_pressure if set via command line.
7335 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7337 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7338 (ashr_simd): New builtin handling DI mode.
7339 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7340 (aarch64_sshr_simddi): New match pattern.
7341 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7342 (vshrd_n_s64): Likewise.
7343 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7345 2014-01-23 Nick Clifton <nickc@redhat.com>
7347 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7348 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7349 favour of mcu specific scripts.
7350 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7353 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7354 Alex Velenko <Alex.Velenko@arm.com>
7356 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7357 (vaddv_s16): Likewise.
7358 (vaddv_s32): Likewise.
7359 (vaddv_u8): Likewise.
7360 (vaddv_u16): Likewise.
7361 (vaddv_u32): Likewise.
7362 (vaddvq_s8): Likewise.
7363 (vaddvq_s16): Likewise.
7364 (vaddvq_s32): Likewise.
7365 (vaddvq_s64): Likewise.
7366 (vaddvq_u8): Likewise.
7367 (vaddvq_u16): Likewise.
7368 (vaddvq_u32): Likewise.
7369 (vaddvq_u64): Likewise.
7370 (vaddv_f32): Likewise.
7371 (vaddvq_f32): Likewise.
7372 (vaddvq_f64): Likewise.
7373 (vmaxv_f32): Likewise.
7374 (vmaxv_s8): Likewise.
7375 (vmaxv_s16): Likewise.
7376 (vmaxv_s32): Likewise.
7377 (vmaxv_u8): Likewise.
7378 (vmaxv_u16): Likewise.
7379 (vmaxv_u32): Likewise.
7380 (vmaxvq_f32): Likewise.
7381 (vmaxvq_f64): Likewise.
7382 (vmaxvq_s8): Likewise.
7383 (vmaxvq_s16): Likewise.
7384 (vmaxvq_s32): Likewise.
7385 (vmaxvq_u8): Likewise.
7386 (vmaxvq_u16): Likewise.
7387 (vmaxvq_u32): Likewise.
7388 (vmaxnmv_f32): Likewise.
7389 (vmaxnmvq_f32): Likewise.
7390 (vmaxnmvq_f64): Likewise.
7391 (vminv_f32): Likewise.
7392 (vminv_s8): Likewise.
7393 (vminv_s16): Likewise.
7394 (vminv_s32): Likewise.
7395 (vminv_u8): Likewise.
7396 (vminv_u16): Likewise.
7397 (vminv_u32): Likewise.
7398 (vminvq_f32): Likewise.
7399 (vminvq_f64): Likewise.
7400 (vminvq_s8): Likewise.
7401 (vminvq_s16): Likewise.
7402 (vminvq_s32): Likewise.
7403 (vminvq_u8): Likewise.
7404 (vminvq_u16): Likewise.
7405 (vminvq_u32): Likewise.
7406 (vminnmv_f32): Likewise.
7407 (vminnmvq_f32): Likewise.
7408 (vminnmvq_f64): Likewise.
7410 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7412 * config/aarch64/aarch64-simd.md
7413 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7414 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7415 (*aarch64_mul3_elt<mode>): Likewise.
7416 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7417 (*aarch64_mul3_elt_to_64v2df): Likewise.
7418 (*aarch64_mla_elt<mode>): Likewise.
7419 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7420 (*aarch64_mls_elt<mode>): Likewise.
7421 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7422 (*aarch64_fma4_elt<mode>): Likewise.
7423 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7424 (*aarch64_fma4_elt_to_64v2df): Likewise.
7425 (*aarch64_fnma4_elt<mode>): Likewise.
7426 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7427 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7428 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7429 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7430 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7431 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7432 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7433 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7434 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7436 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7438 * config/aarch64/aarch64-simd.md
7439 (aarch64_be_checked_get_lane<mode>): New define_expand.
7440 * config/aarch64/aarch64-simd-builtins.def
7441 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7442 New builtin definition.
7443 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7444 Use new safe be builtin.
7446 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7448 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7450 (aarch64_be_st1<mode>): Likewise.
7451 (aarch_ld1<VALL:mode>): Define_expand modified.
7452 (aarch_st1<VALL:mode>): Likewise.
7453 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7454 (UNSPEC_ST1): Likewise.
7456 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7458 * config/microblaze/microblaze.md: Add trap insn and attribute
7460 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7462 PR preprocessor/58580
7463 * input.h (location_get_source_line): Take an additional line_size
7465 (void diagnostics_file_cache_fini): Declare new function.
7466 * input.c (struct fcache): New type.
7467 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7468 New static constants.
7469 (diagnostic_file_cache_init, total_lines_num)
7470 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7471 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7472 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7473 (get_next_line, read_next_line, goto_next_line, read_line_num):
7474 New static function definitions.
7475 (diagnostic_file_cache_fini): New function.
7476 (location_get_source_line): Take an additional output line_len
7477 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7479 * diagnostic.c (diagnostic_finish): Call
7480 diagnostic_file_cache_fini.
7481 (adjust_line): Take an additional input parameter for the length
7482 of the line, rather than calculating it with strlen.
7483 (diagnostic_show_locus): Adjust the use of
7484 location_get_source_line and adjust_line with respect to their new
7485 signature. While displaying a line now, do not stop at the first
7486 null byte. Rather, display the zero byte as a space and keep
7487 going until we reach the size of the line.
7488 * Makefile.in: Add vec.o to OBJS-libcommon
7490 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7491 Ilya Tocar <ilya.tocar@intel.com>
7493 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7494 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7495 (__builtin_ia32_kmov16): Ditto.
7496 * config/i386/i386.md (UNSPEC_KMOV): New.
7499 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7501 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7502 (_mm512_storeu_si512): Ditto.
7504 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7507 * rtl.h (get_referenced_operands): Declare.
7508 * recog.c (get_referenced_operands): New function.
7509 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7510 operands have been referenced when recording LO_SUM references.
7512 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7514 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7516 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7518 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7519 Enable for generic and recent AMD targets.
7521 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7523 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7524 ARG_SIZE note when adjustment was eliminated.
7526 2014-01-22 Jeff Law <law@redhat.com>
7528 PR tree-optimization/59597
7529 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7530 in file. Accept new argument REGISTERING and use it to modify
7531 dump output appropriately.
7532 (register_jump_thread): Corresponding changes.
7533 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7534 thread paths involving joiner blocks. Add code to dump cancelled
7535 jump threading paths.
7537 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7539 PR rtl-optimization/59477
7540 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7541 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7543 2014-01-22 Tom Tromey <tromey@redhat.com>
7545 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7547 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7549 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7551 PR rtl-optimization/59896
7552 * lra-constraints.c (process_alt_operands): Check unused note for
7553 matched operands of insn with no output reloads.
7555 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7557 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7558 (mips_move_from_gpr_cost): Likewise.
7560 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7562 PR rtl-optimization/59858
7563 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7564 ira_class_hard_regs_num.
7565 (process_alt_operands): Increase reject for dying matched operand.
7567 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7570 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7571 smaller than size, perform several stores or loads and stores
7572 at dst + count - size to store or copy all of size bytes, rather
7573 than just last modesize bytes.
7575 2014-01-20 DJ Delorie <dj@redhat.com>
7577 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7578 that CLOBBERs are REGs before propogating their values.
7580 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7583 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7584 (cgraph_inline_failed_type): New function.
7585 * cgraph.h (DEFCIFCODE): Add type.
7586 (cgraph_inline_failed_type_t): New enum.
7587 (cgraph_inline_failed_type): New prototype.
7588 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7589 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7590 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7591 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7592 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7593 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7594 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7595 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7596 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7597 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7598 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7599 OPTIMIZATION_MISMATCH.
7600 * tree-inline.c (expand_call_inline): Emit errors during
7601 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7603 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7606 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7607 mode attribute in insn output.
7609 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7611 * output.h (output_constant): Delete.
7612 * varasm.c (output_constant): Make private.
7614 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7616 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7618 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7621 * tree.h (fold_builtin_strcat): New prototype.
7622 * builtins.c (fold_builtin_strcat): No longer static. Add len
7623 argument, if non-NULL, don't call c_strlen. Optimize
7624 directly into __builtin_memcpy instead of __builtin_strcpy.
7625 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7626 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7628 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7630 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7631 for SImode_address_operand operands, having only a REG argument.
7633 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7635 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7636 loader name using mbig-endian.
7637 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7639 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7641 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7645 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7647 * config/aarch64/aarch64-protos.h
7648 (aarch64_cannot_change_mode_class_ptr): Declare.
7649 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7650 aarch64_cannot_change_mode_class_ptr): New.
7651 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7652 backend hook aarch64_cannot_change_mode_class.
7654 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7656 * common/config/aarch64/aarch64-common.c
7657 (aarch64_handle_option): Don't handle any option order logic here.
7658 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7659 selected_cpu, warn on architecture version mismatch.
7660 (aarch64_override_options): Fix parsing order for option strings.
7662 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7663 Iain Sandoe <iain@codesourcery.com>
7666 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7667 warning. Amend comment to reflect current functionality.
7669 2014-01-20 Richard Biener <rguenther@suse.de>
7672 * builtins.c (fold_builtin_strcat): Remove case better handled
7673 by tree-ssa-strlen.c.
7675 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7677 * config/aarch64/aarch64.opt
7678 (mcpu, march, mtune): Make case-insensitive.
7680 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7683 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7684 if operands[1] is a REG or ZERO_EXTEND of a REG.
7686 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7688 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7690 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7692 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7693 long non-pic millicode calls.
7695 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7697 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7699 2014-01-19 Kito Cheng <kito@0xlab.org>
7701 * builtins.c (expand_movstr): Check movstr expand done or fail.
7703 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7704 H.J. Lu <hongjiu.lu@intel.com>
7707 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7708 to DImode for zero-extended addresses.
7710 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7712 PR rtl-optimization/57763
7713 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7714 on the new indirect jump_insn and increment LABEL_NUSES (label).
7716 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7720 * config.gcc (x86_archs): New variable.
7721 (x86_64_archs): Likewise.
7722 (x86_cpus): Likewise.
7723 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7724 --with-arch/--with-cpu= options.
7725 Support --with-arch=/--with-cpu={nehalem,westmere,
7726 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7728 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7730 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7731 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7733 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7735 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7737 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7740 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7741 clear cpp_get_options (parse_in)->warn_unused_macros for
7742 ix86_target_macros_internal with cpp_define.
7744 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7746 * jump.c (delete_related_insns): Keep (use (insn))s.
7747 * reorg.c (redundant_insn): Check for barriers too.
7749 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7751 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7753 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7755 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7756 call to $$dyncall when TARGET_LONG_CALLS is true.
7758 2014-01-17 Jeff Law <law@redhat.com>
7760 * ree.c (combine_set_extension): Temporarily disable test for
7761 changing number of hard registers.
7763 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7766 * ipa-inline-analysis.c (inline_free_summary):
7767 Do not free summary of aliases.
7769 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7772 * gimplify.c (gimplify_expr): Use create_tmp_var
7773 instead of create_tmp_var_raw. If cond doesn't have
7774 integral type, don't add the IFN_ANNOTATE builtin at all.
7776 2014-01-17 Martin Jambor <mjambor@suse.cz>
7779 * ipa-cp.c (prev_edge_clone): New variable.
7780 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7781 Also resize prev_edge_clone vector.
7782 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7783 (ipcp_edge_removal_hook): New function.
7784 (ipcp_driver): Register ipcp_edge_removal_hook.
7786 2014-01-17 Andrew Pinski <apinski@cavium.com>
7787 Steve Ellcey <sellcey@mips.com>
7790 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7793 2014-01-17 Jeff Law <law@redhat.com>
7796 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7797 so that pass_ccp runs first.
7799 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7801 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7802 (ix86_adjust_cost): Use !TARGET_XXX.
7803 (do_reorder_for_imul): Likewise.
7804 (swap_top_of_ready_list): Likewise.
7805 (ix86_sched_reorder): Likewise.
7807 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7809 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7810 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7811 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7812 (intel_memset): New. Duplicate slm_memset.
7813 (intel_cost): New. Duplicate slm_cost.
7814 (m_INTEL): New macro.
7815 (processor_target_table): Add "intel".
7816 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7817 with PROCESSOR_INTEL for "intel".
7818 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7819 PROCESSOR_SILVERMONT.
7820 (ix86_issue_rate): Likewise.
7821 (ix86_adjust_cost): Likewise.
7822 (ia32_multipass_dfa_lookahead): Likewise.
7823 (swap_top_of_ready_list): Likewise.
7824 (ix86_sched_reorder): Likewise.
7825 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7826 instead of TARGET_OPT_AGU.
7827 * config/i386/i386.h (TARGET_INTEL): New.
7828 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7829 (processor_type): Add PROCESSOR_INTEL.
7830 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7831 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7833 2014-01-17 Marek Polacek <polacek@redhat.com>
7836 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7839 2014-01-17 Richard Biener <rguenther@suse.de>
7841 PR tree-optimization/46590
7842 * opts.c (default_options_table): Add entries for
7843 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7844 all enabled at -O1 but not for -Og.
7845 * common.opt (fbranch-count-reg): Remove Init(1).
7846 (fmove-loop-invariants): Likewise.
7847 (ftree-pta): Likewise.
7849 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7851 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7852 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7853 decls to at least the GCC 4.8 used alignments.
7856 * tree-nested.c (convert_nonlocal_reference_stmt,
7857 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7858 of GIMPLE_BIND stmts, adjust associated decls.
7860 2014-01-17 Richard Biener <rguenther@suse.de>
7862 PR tree-optimization/46590
7863 * vec.h (vec<>::bseach): New member function implementing
7864 binary search according to C89 bsearch.
7865 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7866 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7867 bitmap pointer again. Make accesses_in_loop a flat array.
7868 (mem_ref_obstack): New global.
7869 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7870 (mark_ref_stored): Likewise.
7871 (ref_indep_loop_p_2): Likewise.
7872 (set_ref_stored_in_loop): New helper function.
7873 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7874 (memref_free): Adjust.
7875 (record_mem_ref_loc): Simplify.
7876 (gather_mem_refs_stmt): Adjust.
7877 (sort_locs_in_loop_postorder_cmp): New function.
7878 (analyze_memory_references): Sort accesses_in_loop after
7879 loop postorder number.
7880 (find_ref_loc_in_loop_cmp): New function.
7881 (for_all_locs_in_loop): Find relevant cluster of locs in
7882 accesses_in_loop and iterate without recursion.
7883 (execute_sm): Avoid uninit warning.
7884 (struct ref_always_accessed): Simplify.
7885 (ref_always_accessed::operator ()): Likewise.
7886 (ref_always_accessed_p): Likewise.
7887 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7888 loop postorder numbers here.
7889 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7892 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7895 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7896 on decls for which assemble_alias has been called.
7898 2014-01-17 Nick Clifton <nickc@redhat.com>
7900 * config/msp430/msp430.opt: (mcpu): New option.
7901 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7902 (msp430_option_override): Parse target_cpu. If the MCU name
7903 matches a generic string, clear target_mcu.
7904 (msp430_attr): Allow numeric interrupt values up to 63.
7905 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7906 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7908 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7910 * config/msp430/msp430.md (popm): Use %J rather than %I.
7911 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7912 (addhi_cy_i): Use immediate_operand for operand 2.
7913 * doc/invoke.texi: Document -mcpu option.
7915 2014-01-17 Richard Biener <rguenther@suse.de>
7917 PR rtl-optimization/38518
7918 * df.h (df_analyze_loop): Declare.
7919 * df-core.c: Include cfgloop.h.
7920 (df_analyze_1): Split out main part of df_analyze.
7921 (df_analyze): Adjust.
7922 (loop_inverted_post_order_compute): New function.
7923 (loop_post_order_compute): Likewise.
7924 (df_analyze_loop): New function avoiding whole-function
7926 * loop-invariant.c (find_defs): Use df_analyze_loop.
7927 (find_invariants): Adjust.
7928 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7930 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7932 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7933 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7935 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7937 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7938 traversal when removing references.
7940 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7943 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7945 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7948 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7949 pushing a reload for an autoinc when we had previously reloaded an
7950 inner part of the address.
7952 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7954 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7956 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7957 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7958 when not giving up or versioning for alias only because of
7960 (vect_analyze_data_ref_dependences): Set to true.
7961 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7963 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7964 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7968 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7971 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7972 operand 0 predicate for gathers, use a new pseudo as subtarget.
7974 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7977 * lra-constraints.c (process_alt_operands): Add printing debug info.
7978 Check absence of input/output reloads for matched operands too.
7980 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7982 PR rtl-optimization/59835
7983 * ira.c (ira_init_register_move_cost): Increase cost for
7986 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7988 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7990 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7993 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7994 non-register objects. Use gen_(high/low)part more consistently.
7997 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8000 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
8001 endian support, remove tests for WORDS_BIG_ENDIAN.
8002 (p8_mfvsrd_3_<mode>): Likewise.
8003 (reload_gpr_from_vsx<mode>): Likewise.
8004 (reload_gpr_from_vsxsf): Likewise.
8005 (p8_mfvsrd_4_disf): Likewise.
8007 2014-01-16 Richard Biener <rguenther@suse.de>
8009 PR rtl-optimization/46590
8010 * lcm.c (compute_antinout_edge): Use postorder iteration.
8011 (compute_laterin): Use inverted postorder iteration.
8013 2014-01-16 Nick Clifton <nickc@redhat.com>
8016 * varasm.c (output_constant): Return the number of bytes actually
8018 (output_constructor_array_range): Update the field size with the
8019 number of bytes emitted by output_constant.
8020 (output_constructor_regular_field): Likewise. Also do not
8021 complain if the total number of bytes emitted is now greater
8022 than the expected fieldpos.
8023 * output.h (output_constant): Update prototype and descriptive comment.
8025 2014-01-16 Marek Polacek <polacek@redhat.com>
8028 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8029 it is error_mark_node.
8031 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8033 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8034 VALID_AVX256_REG_OR_OI_MODE.
8036 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8038 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8039 current procedure should be profiled.
8041 2014-01-15 Andrew Pinski <apinski@cavium.com>
8043 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8044 of moving from/to the STACK_REG register class.
8046 2014-01-15 Richard Henderson <rth@redhat.com>
8049 * reginfo.c (global_regs_decl): Globalize.
8050 * rtl.h (global_regs_decl): Declare.
8051 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8052 reserved via global_regs.
8054 2014-01-15 Teresa Johnson <tejohnson@google.com>
8056 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8058 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8060 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8061 and vmulosh rather than call gen_vec_widen_smult_*.
8062 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8063 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8064 (vec_widen_smult_even_v16qi): Likewise.
8065 (vec_widen_umult_even_v8hi): Likewise.
8066 (vec_widen_smult_even_v8hi): Likewise.
8067 (vec_widen_umult_odd_v16qi): Likewise.
8068 (vec_widen_smult_odd_v16qi): Likewise.
8069 (vec_widen_umult_odd_v8hi): Likewise.
8070 (vec_widen_smult_odd_v8hi): Likewise.
8071 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8072 vmuloub rather than call gen_vec_widen_umult_*.
8073 (vec_widen_umult_lo_v16qi): Likewise.
8074 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8075 vmulosb rather than call gen_vec_widen_smult_*.
8076 (vec_widen_smult_lo_v16qi): Likewise.
8077 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8078 rather than call gen_vec_widen_umult_*.
8079 (vec_widen_umult_lo_v8hi): Likewise.
8080 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8081 rather than call gen_vec_widen_smult_*.
8082 (vec_widen_smult_lo_v8hi): Likewise.
8084 2014-01-15 Jeff Law <law@redhat.com>
8086 PR tree-optimization/59747
8087 * ree.c (find_and_remove_re): Properly handle case where a second
8088 eliminated extension requires widening a copy created for elimination
8089 of a prior extension.
8090 (combine_set_extension): Ensure that the number of hard regs needed
8091 for a destination register does not change when we widen it.
8093 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8095 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8096 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8097 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8098 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8099 (avr-*-rtems*): Likewise.
8100 (bfin*-rtems*): Likewise.
8101 (moxie-*-rtems*): Likewise.
8102 (h8300-*-rtems*): Likewise.
8103 (i[34567]86-*-rtems*): Likewise.
8104 (lm32-*-rtems*): Likewise.
8105 (m32r-*-rtems*): Likewise.
8106 (m68k-*-rtems*): Likewise.
8107 (microblaze*-*-rtems*): Likewise.
8108 (mips*-*-rtems*): Likewise.
8109 (powerpc-*-rtems*): Likewise.
8110 (sh-*-rtems*): Likewise.
8111 (sparc-*-rtems*): Likewise.
8112 (sparc64-*-rtems*): Likewise.
8113 (v850-*-rtems*): Likewise.
8114 (m32c-*-rtems*): Likewise.
8116 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
8118 PR rtl-optimization/59511
8119 * ira.c (ira_init_register_move_cost): Use memory costs for some
8120 cases of register move cost calculations.
8121 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
8122 instead of BB frequency.
8123 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
8124 * lra-assigns.c (find_hard_regno_for): Ditto.
8126 2014-01-15 Richard Biener <rguenther@suse.de>
8128 PR tree-optimization/59822
8129 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8130 (vectorizable_load): Use it to hoist defs of uses of invariant
8131 loads out of the loop.
8133 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8134 Kugan Vivekanandarajah <kuganv@linaro.org>
8137 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8140 2014-01-15 Richard Biener <rguenther@suse.de>
8142 PR rtl-optimization/59802
8143 * lcm.c (compute_available): Use inverted postorder to seed
8144 the initial worklist.
8146 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8149 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8150 ADDR_REGS for invalid symrefs in non-PIC code.
8152 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8155 * builtins.c (determine_block_size): Initialize *probable_max_size
8156 even if len_rtx is CONST_INT.
8158 2014-01-14 Andrew Pinski <apinski@cavium.com>
8160 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8161 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8162 (cortexa53_tunings): Likewise.
8163 (aarch64_sched_issue_rate): New function.
8164 (TARGET_SCHED_ISSUE_RATE): Define.
8166 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8168 * ira-costs.c (find_costs_and_classes): Add missed
8169 ira_init_register_move_cost_if_necessary.
8171 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8174 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8176 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8179 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8180 to indicate if type is used for function return value. Warn ABI
8181 change if the vector mode isn't available for function return value.
8182 (ix86_function_arg_advance): Pass false to type_natural_mode.
8183 (ix86_function_arg): Likewise.
8184 (ix86_gimplify_va_arg): Likewise.
8185 (function_arg_32): Don't warn ABI change.
8186 (ix86_function_value): Pass true to type_natural_mode.
8187 (ix86_return_in_memory): Likewise.
8188 (ix86_struct_value_rtx): Removed.
8189 (TARGET_STRUCT_VALUE_RTX): Likewise.
8191 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8193 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8194 converting a conditional jump into a conditional return.
8196 2014-01-14 Richard Biener <rguenther@suse.de>
8198 PR tree-optimization/58921
8199 PR tree-optimization/59006
8200 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8201 hoisting invariant stmts.
8202 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8203 invariant loads on the preheader edge if possible.
8205 2014-01-14 Joey Ye <joey.ye@arm.com>
8207 * doc/plugin.texi (Building GCC plugins): Update to C++.
8209 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8211 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8212 (_mm_rcp28_round_ss): Ditto.
8213 (_mm_rsqrt28_round_sd): Ditto.
8214 (_mm_rsqrt28_round_ss): Ditto.
8215 (_mm_rcp28_sd): Ditto.
8216 (_mm_rcp28_ss): Ditto.
8217 (_mm_rsqrt28_sd): Ditto.
8218 (_mm_rsqrt28_ss): Ditto.
8219 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8220 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8221 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8222 (IX86_BUILTIN_RCP28SD): Ditto.
8223 (IX86_BUILTIN_RCP28SS): Ditto.
8224 (IX86_BUILTIN_RSQRT28SD): Ditto.
8225 (IX86_BUILTIN_RSQRT28SS): Ditto.
8226 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8227 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8228 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8229 (ix86_expand_special_args_builtin): Expand new FTYPE.
8230 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8231 (srcp14<mode>): Make insn unary.
8232 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8233 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8234 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8235 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8236 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8237 Fix rounding: make it SAE only.
8238 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8240 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8242 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8243 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8244 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8245 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8246 (round_saeonly_mask_scalar_operand4): Ditto.
8247 (round_saeonly_mask_scalar_op3): Ditto.
8248 (round_saeonly_mask_scalar_op4): Ditto.
8250 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8252 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8253 Implement -maltivec=be for vec_insert and vec_extract.
8255 2014-01-10 DJ Delorie <dj@redhat.com>
8257 * config/msp430/msp430.md (call_internal): Don't allow memory
8258 references with SP as the base register.
8259 (call_value_internal): Likewise.
8260 * config/msp430/constraints.md (Yc): New. For memory references
8261 that don't use SP as a base register.
8263 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8264 "an integer without a # prefix"
8265 * config/msp430/msp430.md (epilogue_helper): Use it.
8267 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8270 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8271 AVX512F gather builtins.
8272 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8273 on gather decls with INTEGER_TYPE masktype.
8274 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8275 directly into the builtin rather than hoisting it before loop.
8277 PR tree-optimization/59387
8278 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8279 (scev_const_prop): If folded_casts and type has undefined overflow,
8280 use force_gimple_operand instead of force_gimple_operand_gsi and
8281 for each added stmt if it is assign with
8282 arith_code_with_undefined_signed_overflow, call
8283 rewrite_to_defined_overflow.
8284 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8285 gimple-fold.h instead.
8286 (arith_code_with_undefined_signed_overflow,
8287 rewrite_to_defined_overflow): Moved to ...
8288 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8289 rewrite_to_defined_overflow): ... here. No longer static.
8290 Include gimplify-me.h.
8291 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8292 rewrite_to_defined_overflow): New prototypes.
8294 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8296 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8298 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8300 * builtins.c (get_object_alignment_2): Minor tweak.
8301 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8303 2014-01-13 Christian Bruel <christian.bruel@st.com>
8305 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8306 optimized non constant lengths.
8308 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8311 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8312 load as __atomic_load_N if possible.
8314 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8316 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8318 (rs6000_expand_builtin): Adjust call.
8320 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8323 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8324 * config/rs6000/rs6000.c: Include target-globals.h.
8325 (rs6000_set_current_function): Instead of doing target_reinit
8326 unconditionally, use save_target_globals_default_opts and
8327 restore_target_globals.
8329 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8331 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8332 (rs6000_expand_builtin): Handle mffs and mtfsf.
8333 (rs6000_init_builtins): Define mffs and mtfsf.
8334 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8335 (rs6000_mffs): New pattern.
8336 (rs6000_mtfsf): New pattern.
8338 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8340 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8341 Start narrowing with START. Apply candidate-use pair
8342 and check overall cost in narrowing.
8343 (iv_ca_prune): Pass new argument.
8345 2014-01-10 Jeff Law <law@redhat.com>
8348 * ree.c (combine_reaching_defs): Ensure the defining statement
8349 occurs before the extension when optimizing extensions with
8350 different source and destination hard registers.
8352 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8355 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8356 vtables into the type inheritance graph.
8358 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8360 PR rtl-optimization/59754
8361 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8362 modes in the REGNO != REGNO case.
8364 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8366 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8368 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8370 PR tree-optimization/59745
8371 * tree-predcom.c (tree_predictive_commoning_loop): Call
8372 free_affine_expand_cache if giving up because components is NULL.
8374 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8375 GC in payload of target_globals struct instead of allocating them on
8376 the heap and the larger structs separately using GC.
8377 * target-globals.h (struct target_globals): Make regs, hard_regs,
8378 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8379 of GTY((skip)) and change type to void *.
8380 (reset_target_globals): Cast loads from those fields to corresponding
8383 2014-01-10 Steve Ellcey <sellcey@mips.com>
8386 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8387 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8388 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8390 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8393 * aarch64-modes.def (CC_Zmode): New flags mode.
8394 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8395 represents an equality.
8396 (aarch64_get_condition_code): Handle CC_Zmode.
8397 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8399 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8401 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8402 extraction in good case.
8404 2014-01-10 Richard Biener <rguenther@suse.de>
8406 PR tree-optimization/59374
8407 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8408 checking after SLP discovery. Mark stmts not participating
8409 in any SLP instance properly.
8411 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8413 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8414 when handling a SET rtx.
8416 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8418 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8419 (cortex-a57): Likewise.
8420 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8422 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8424 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8425 non-iwmmxt builtins.
8427 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8431 * ipa-devirt.c record_target_from_binfo): Take as argument
8432 stack of binfos and lookup matching one for virtual inheritance.
8433 (possible_polymorphic_call_targets_1): Update.
8435 2014-01-10 Huacai Chen <chenhc@lemote.com>
8437 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8438 kernel strings for Loongson-2E/2F/3A.
8440 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8443 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8444 is_gimple_call before calling gimple_call_internal_p.
8446 2014-01-09 Steve Ellcey <sellcey@mips.com>
8448 * Makefile.in (TREE_FLOW_H): Remove.
8449 (TREE_SSA_H): Add file names from tree-flow.h.
8450 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8451 * tree.h: Remove tree-flow.h reference.
8452 * hash-table.h: Remove tree-flow.h reference.
8453 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8454 reference with tree-ssa-loop.h.
8456 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8458 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8459 default element-order behavior for -maltivec.
8460 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8461 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8462 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8463 when targeting big endian, at least for now.
8464 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8466 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8469 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8470 var satisfies use_register_for_decl, just take into account type
8471 alignment, rather than decl alignment.
8473 PR tree-optimization/59622
8474 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8475 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8476 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8477 Don't devirtualize for inplace at all. For targets.length () == 1,
8478 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8480 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8482 * config/i386/i386.md (cpu): Remove the unused btver1.
8484 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8486 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8488 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8491 * tree-core.h (struct target_globals): New forward declaration.
8492 (struct tree_target_option): Add globals field.
8493 * tree.h (TREE_TARGET_GLOBALS): Define.
8494 (prepare_target_option_nodes_for_pch): New prototype.
8495 * target-globals.h (struct target_globals): Define even if
8497 * tree.c (prepare_target_option_node_for_pch,
8498 prepare_target_option_nodes_for_pch): New functions.
8499 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8500 * config/i386/i386.c: Include target-globals.h.
8501 (ix86_set_current_function): Instead of doing target_reinit
8502 unconditionally, use save_target_globals_default_opts and
8503 restore_target_globals.
8505 2014-01-09 Richard Biener <rguenther@suse.de>
8507 PR tree-optimization/59715
8508 * tree-cfg.h (split_critical_edges): Declare.
8509 * tree-cfg.c (split_critical_edges): Export.
8510 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8512 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8514 * cfgexpand.c (expand_stack_vars): Optionally disable
8515 asan stack protection.
8516 (expand_used_vars): Likewise.
8517 (partition_stack_vars): Likewise.
8518 * asan.c (asan_emit_stack_protection): Optionally disable
8519 after return stack usage.
8520 (instrument_derefs): Optionally disable memory access instrumentation.
8521 (instrument_builtin_call): Likewise.
8522 (instrument_strlen_call): Likewise.
8523 (asan_protect_global): Optionally disable global variables protection.
8524 * doc/invoke.texi: Added doc for new options.
8525 * params.def: Added new options.
8526 * params.h: Likewise.
8528 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8530 PR rtl-optimization/59724
8531 * ifcvt.c (cond_exec_process_if_block): Don't call
8532 flow_find_head_matching_sequence with 0 longest_match.
8533 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8534 non-active insns if !stop_after.
8535 (try_head_merge_bb): Revert 2014-01-07 changes.
8537 2014-01-08 Jeff Law <law@redhat.com>
8539 * ree.c (get_sub_rtx): New function, extracted from...
8540 (merge_def_and_ext): Here.
8541 (combine_reaching_defs): Use get_sub_rtx.
8543 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8545 * cgraph.h (varpool_variable_node): Do not choke on null node.
8547 2014-01-08 Catherine Moore <clm@codesourcery.com>
8549 * config/mips/mips.md (simple_return): Attempt to use JRC
8551 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8553 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8555 PR rtl-optimization/59137
8556 * reorg.c (steal_delay_list_from_target): Call update_block for
8558 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8560 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8562 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8563 two duplicate entries.
8565 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8568 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8570 * config/mips/mips.c (mips_truncated_op_cost): New function.
8571 (mips_rtx_costs): Adjust test for BADDU.
8572 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8574 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8576 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8577 (*baddu_si): ...this new pattern.
8579 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8582 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8584 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8587 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8589 (expand_expr, expand_normal): Adjust.
8590 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8591 inner_reference_p. Use inner_reference_p to expand inner references.
8592 (store_expr): Adjust.
8593 * cfgexpand.c (expand_call_stmt): Adjust.
8595 2014-01-08 Rong Xu <xur@google.com>
8597 * gcov-io.c (gcov_var): Move from gcov-io.h.
8598 (gcov_position): Ditto.
8599 (gcov_is_error): Ditto.
8600 (gcov_rewrite): Ditto.
8601 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8602 only part to libgcc/libgcov.h.
8604 2014-01-08 Marek Polacek <polacek@redhat.com>
8607 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8609 2014-01-08 Marek Polacek <polacek@redhat.com>
8612 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8614 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8616 PR rtl-optimization/59649
8617 * stor-layout.c (get_mode_bounds): For BImode return
8618 0 and STORE_FLAG_VALUE.
8620 2014-01-08 Richard Biener <rguenther@suse.de>
8623 * gimple.h (is_gimple_builtin_call): Remove.
8624 (gimple_builtin_call_types_compatible_p): New.
8625 (gimple_call_builtin_p): New overload.
8626 * gimple.c (is_gimple_builtin_call): Remove.
8627 (validate_call): Rename to ...
8628 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8630 (validate_type): New static function.
8631 (gimple_call_builtin_p): New overload and adjust.
8632 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8633 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8634 (gimple_fold_stmt_to_constant_1): Likewise.
8635 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8637 2014-01-08 Richard Biener <rguenther@suse.de>
8640 * gimplify.c (gimplify_expr): Gimplify register-register type
8641 VIEW_CONVERT_EXPRs to separate stmts.
8643 2014-01-07 Jeff Law <law@redhat.com>
8646 * ree.c (combine_set_extension): Handle case where source
8647 and destination registers in an extension insn are different.
8648 (combine_reaching_defs): Allow source and destination registers
8649 in extension to be different under limited circumstances.
8650 (add_removable_extension): Remove restriction that the
8651 source and destination registers in the extension are the same.
8652 (find_and_remove_re): Emit a copy from the extension's
8653 destination to its source after the defining insn if
8654 the source and destination registers are different.
8657 * ifcvt.c (merge_if_block): If we are merging a block with more than
8658 one successor with a block with no successors, remove any BARRIER
8659 after the second block.
8661 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8663 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8665 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8668 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8669 for 14-bit register offsets when INT14_OK_STRICT is false.
8671 2014-01-07 Roland Stigge <stigge@antcom.de>
8672 Michael Meissner <meissner@linux.vnet.ibm.com>
8675 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8676 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8678 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8680 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8683 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8685 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8686 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8687 rtx is const0_rtx or not.
8689 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8692 * target-globals.c (save_target_globals): Remove this_fn_optab
8694 * toplev.c: Include optabs.h.
8695 (target_reinit): Temporarily restore the global options if another
8696 set of options are in force.
8698 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8700 PR rtl-optimization/58668
8701 * cfgcleanup.c (flow_find_cross_jump): Don't count
8702 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8703 to determine what is counted.
8704 (flow_find_head_matching_sequence): Use active_insn_p to determine
8706 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8708 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8709 determine what is counted.
8711 PR tree-optimization/59643
8712 * tree-predcom.c (split_data_refs_to_components): If one dr is
8713 read and one write, determine_offset fails and the write isn't
8714 in the bad component, just put the read into the bad component.
8716 2014-01-07 Mike Stump <mikestump@comcast.net>
8717 Jakub Jelinek <jakub@redhat.com>
8720 * tree-core.h (struct tree_optimization_option): Change optabs
8721 type from unsigned char * to void *.
8722 * optabs.c (init_tree_optimization_optabs): Adjust
8723 TREE_OPTIMIZATION_OPTABS initialization.
8725 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8728 * config/i386/i386.h (struct machine_function): Add
8729 no_drap_save_restore field.
8730 * config/i386/i386.c (ix86_save_reg): Use
8731 !cfun->machine->no_drap_save_restore instead of
8732 crtl->stack_realign_needed.
8733 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8734 this function clears frame_pointer_needed. Set
8735 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8736 and DRAP reg is needed.
8738 2014-01-06 Marek Polacek <polacek@redhat.com>
8741 * doc/implement-c.texi: Mention that other integer types are
8742 permitted as bit-field types in strictly conforming mode.
8744 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8746 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8749 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8751 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8753 2014-01-06 Martin Jambor <mjambor@suse.cz>
8756 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8758 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8760 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8764 * var-tracking.c (add_stores): Preserve the value of the source even if
8765 we don't record the store.
8767 2014-01-06 Terry Guo <terry.guo@arm.com>
8769 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8771 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8774 * config/darwin.c (darwin_function_section): Adjust return values to
8775 correspond to optimisation changes made in r206070.
8777 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8779 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8780 from prefetch_block tune setting.
8781 (nocona_cost): Correct size of prefetch block to 64.
8783 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8785 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8786 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8787 used to save the static chain register in the computation of the offset
8788 from which the FP registers need to be restored.
8790 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8792 PR tree-optimization/59519
8793 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8794 ICE if get_current_def (current_new_name) is already non-NULL, as long
8795 as it is a phi result of some other phi in *new_exit_bb that has
8798 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8799 or vmovdqu* for misaligned_operand.
8800 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8801 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8802 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8803 aligned_mem for AVX512F masked aligned load and store builtins and for
8806 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8808 PR tree-optimization/59651
8809 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8810 Address range for negative step should be added by TYPE_SIZE_UNIT.
8812 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8814 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8815 overlapping registers also for registers other than the stack pointer.
8817 2014-01-03 Marek Polacek <polacek@redhat.com>
8820 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8823 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8826 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8829 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8830 (push splitter): Use <P:MODE_SIZE> instead of
8831 GET_MODE_SIZE (<P:MODE>mode).
8832 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8833 (mov -1, reg peephole2): Likewise.
8834 * config/i386/sse.md (*mov<mode>_internal,
8835 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8836 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8837 *<code><mode>3, *andnot<mode>3<mask_name>,
8838 <mask_codefor><code><mode>3<mask_name>): Likewise.
8839 * config/i386/subst.md (mask_mode512bit_condition,
8840 sd_mask_mode512bit_condition): Likewise.
8842 2014-01-02 Xinliang David Li <davidxl@google.com>
8844 PR tree-optimization/59303
8845 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8846 (dump_predicates): Better output format.
8847 (pred_equal_p): New function.
8848 (is_neq_relop_p): Ditto.
8849 (is_neq_zero_form_p): Ditto.
8850 (pred_expr_equal_p): Ditto.
8851 (pred_neg_p): Ditto.
8852 (simplify_pred): Ditto.
8853 (simplify_preds_2): Ditto.
8854 (simplify_preds_3): Ditto.
8855 (simplify_preds_4): Ditto.
8856 (simplify_preds): Ditto.
8858 (push_to_worklist): Ditto.
8859 (get_pred_info_from_cmp): Ditto.
8860 (is_degenerated_phi): Ditto.
8861 (normalize_one_pred_1): Ditto.
8862 (normalize_one_pred): Ditto.
8863 (normalize_one_pred_chain): Ditto.
8864 (normalize_preds): Ditto.
8865 (normalize_cond_1): Remove function.
8866 (normalize_cond): Ditto.
8867 (is_gcond_subset_of): Ditto.
8868 (is_subset_of_any): Ditto.
8869 (is_or_set_subset_of): Ditto.
8870 (is_and_set_subset_of): Ditto.
8871 (is_norm_cond_subset_of): Ditto.
8872 (pred_chain_length_cmp): Ditto.
8873 (convert_control_dep_chain_into_preds): Type change.
8874 (find_predicates): Ditto.
8875 (find_def_preds): Ditto.
8876 (destroy_predicates_vecs): Ditto.
8877 (find_matching_predicates_in_rest_chains): Ditto.
8878 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8879 (is_pred_expr_subset): Ditto.
8880 (is_pred_chain_subset_of): Ditto.
8881 (is_included_in): Ditto.
8882 (is_superset_of): Ditto.
8884 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8886 Update copyright years.
8888 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8890 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8891 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8892 config/arc/arc.md, config/arc/arc.opt,
8893 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8894 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8895 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8896 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8897 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8898 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8899 vtable-verify.c, vtable-verify.h: Use the standard form for the
8902 2014-01-02 Tobias Burnus <burnus@net-b.de>
8904 * gcc.c (process_command): Update copyright notice dates.
8905 * gcov-dump.c: Ditto.
8907 * doc/cpp.texi: Bump @copying's copyright year.
8908 * doc/cppinternals.texi: Ditto.
8909 * doc/gcc.texi: Ditto.
8910 * doc/gccint.texi: Ditto.
8911 * doc/gcov.texi: Ditto.
8912 * doc/install.texi: Ditto.
8913 * doc/invoke.texi: Ditto.
8915 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8917 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8919 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8921 * config/i386/sse.md (*mov<mode>_internal): Guard
8922 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8924 PR rtl-optimization/59647
8925 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8926 new_rtx into UNSIGNED_FLOAT rtxes.
8928 Copyright (C) 2014 Free Software Foundation, Inc.
8930 Copying and distribution of this file, with or without modification,
8931 are permitted in any medium without royalty provided the copyright
8932 notice and this notice are preserved.