1 2014-05-05 Richard Biener <rguenther@suse.de>
3 * passes.c (execute_function_todo): Don't reset TODO_verify_ssa
4 from last_verified if update_ssa ran. Move TODO_verify_rtl_sharing
5 under the TODO_verify_il umbrella.
7 2014-05-05 Richard Biener <rguenther@suse.de>
9 * passes.c (execute_function_todo): Move TODO_verify_flow under
10 the TODO_verify_ul umbrella.
12 2014-05-05 Richard Biener <rguenther@suse.de>
15 * fold-const.c (fold_binary_loc): Consistently avoid
16 canonicalizing X & CST away from a CST that is the mask
19 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
21 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
22 int argument to enum machine_mode.
23 (picochip_class_max_nregs): Ditto.
24 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
25 (picochip_class_max_nregs): Ditto.
27 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29 * target.def: Add new target hook.
30 * doc/tm.texi: Regenerate.
31 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
32 * targhooks.c (default_keep_leaf_when_profiled): New function.
34 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
35 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
37 2014-05-05 Bin Cheng <bin.cheng@arm.com>
39 PR tree-optimization/60363
40 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
41 (copy_phi_args): New parameters. Call get_value_locus_in_path.
42 (update_destination_phis): New parameter.
43 (create_edge_and_update_destination_phis): Ditto.
44 (ssa_fix_duplicate_block_edges): Pass new arguments.
45 (thread_single_edge): Ditto.
47 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
49 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
50 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
51 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
52 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
53 Use RS6000_BTM_HARD_FLOAT.
54 (BU_MISC_2): Likewise.
55 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
56 RS6000_BTM_HARD_FLOAT.
57 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
59 (rs6000_invalid_builtin): Add hard floating builtin support.
60 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
62 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
64 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
66 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
67 Add missing function* argument.
69 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
71 * lra-constraints.c (valid_address_p): Move earlier in file.
72 Add a constraint argument to the address_info version.
73 (satisfies_memory_constraint_p): New function.
74 (satisfies_address_constraint_p): Likewise.
75 (process_alt_operands, curr_insn_transform): Use them.
76 (process_address): Pass the constraint to valid_address_p when
77 checking address operands.
79 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
81 * config/mips/mips.c (mips_isa_rev): New variable.
82 (mips_set_architecture): Set it.
83 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
85 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
86 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
87 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
88 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
89 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
90 conditions in terms of mips_isa_rev.
91 (mips_isa_rev): Declare.
93 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
95 * config/sh/sh-mem.cc: Use tabs instead of spaces.
96 (prob_unlikely, prob_likely): Make variables const.
98 2014-05-03 Denis Chertykov <chertykov@gmail.com>
100 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
102 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
104 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
106 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
108 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
109 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
110 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
112 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
113 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
115 Replace usage of ROUND_REG with sh_round_reg.
116 Use CEIL instead of ROUND_ADVANCE.
118 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
121 * config/sh/sh.c: Include stdlib headers before everything else.
123 2014-05-02 Jakub Jelinek <jakub@redhat.com>
125 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
126 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
127 (gimplify_adjust_omp_clauses): Simd region is never
128 directly nested in combined parallel. Instead, for linear
129 with copyin/copyout, if in combined for simd loop, make decl
130 firstprivate/lastprivate on OMP_FOR.
131 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
132 expand_omp_for_static_chunk): When setting endvar, also set
133 fd->loop.v to the same value.
135 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
137 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
139 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
141 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
144 2014-05-02 Marek Polacek <polacek@redhat.com>
146 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
148 2014-02-26 Kito Cheng <kito@0xlab.org>
150 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
151 to a C expression marco.
152 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
153 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
154 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
155 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
156 HONOR_REG_ALLOC_ORDER.
157 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
159 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
161 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
163 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
165 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
167 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
169 * tree-if-conv.c (is_cond_scalar_reduction): New function.
170 (convert_scalar_cond_reduction): Likewise.
171 (predicate_scalar_phi): Add recognition and transformation
172 of simple conditioanl reduction to be vectorizable.
174 2014-05-01 Marek Polacek <polacek@redhat.com>
177 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
179 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
181 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
182 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
183 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
184 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
185 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
186 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
187 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
188 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
190 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
192 * config/arc/arc.opt (mlra): Move comment above option name
193 to avoid mis-parsing as language options.
195 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
197 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
198 * config/sol2.h: ... here.
199 * config/sol2-10.h: Remove.
201 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
202 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
203 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
204 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
205 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
206 * config/sol2.h: ... here.
207 (SECTION_NAME_FORMAT): Don't redefine.
208 (STARTFILE_ARCH32_SPEC): Rename to ...
209 (STARTFILE_ARCH_SPEC): ... this.
210 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
211 * config/sparc/sol2.h: ... here.
212 (SECTION_NAME_FORMAT): Don't undef.
213 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
214 (SUBTARGET_EXTRA_SPECS): Remove.
215 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
217 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
218 (MD_STARTFILE_PREFIX): Remove.
219 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
220 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
221 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
222 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
223 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
224 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
225 * config/i386/sol2.h: ... here.
226 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
227 * config/i386/sol2-bi.h: Remove.
228 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
229 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
231 * config/i386/t-sol2-64: Rename to ...
232 * config/i386/t-sol2: ... this.
233 * config/sparc/t-sol2-64: Rename to ...
234 * config/sparc/t-sol2: ... this.
236 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
237 sol2_tm_file_head, sol2_tm_file_tail.
238 Include ${cpu_type}/sol2.h before sol2.h.
240 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
241 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
242 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
243 Reflect i386/t-sol2-64 renaming.
244 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
245 Reflect sparc/t-sol2-64 renaming.
247 2014-04-30 Richard Biener <rguenther@suse.de>
249 * passes.c (execute_function_todo): Move TODO_verify_stmts
250 and TODO_verify_ssa under the TODO_verify_il umbrella.
251 * tree-ssa.h (verify_ssa): Adjust prototype.
252 * tree-ssa.c (verify_ssa): Add parameter to tell whether
253 we should verify SSA operands.
254 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
255 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
256 whether we should verify whether not throwing stmts have EH info.
257 * graphite-scop-detection.c (create_sese_edges): Adjust.
258 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
259 * tree-eh.c (lower_try_finally_switch): Do not add the
260 default case label twice.
262 2014-04-30 Marek Polacek <polacek@redhat.com>
264 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
265 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
266 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
267 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
269 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
271 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
272 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
273 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
274 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
275 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
276 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
277 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
278 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
280 2014-04-29 David Malcolm <dmalcolm@redhat.com>
282 * tree-cfg.c (dump_function_to_file): Dump the return type of
283 functions, in a line to itself before the function body, mimicking
284 the layout of a C function.
286 2014-04-29 Jakub Jelinek <jakub@redhat.com>
288 PR tree-optimization/60971
289 * tree-tailcall.c (process_assignment): Reject conversions which
292 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
294 * calls.c (initialize_argument_information): Always treat
295 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
296 (expand_call): Likewise.
297 (emit_library_call_calue_1): Likewise.
298 * expr.c (PUSH_ARGS_REVERSED): Do not define.
299 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
302 2014-04-29 Nick Clifton <nickc@redhat.com>
304 * config/msp430/msp430.md (umulsidi): Fix typo.
305 (mulhisi3): Enable even inside interrupt handlers.
306 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
307 bigger return address pushed in large mode.
309 2014-04-29 Nick Clifton <nickc@redhat.com>
311 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
312 (arc_init_reg_tables): Use a machine_mode enum to iterate over
314 * config/m32r/m32r.c (init_reg_tables): Likewise.
315 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
316 enum to hold the modes.
318 2014-04-29 Richard Biener <rguenther@suse.de>
320 * dominance.c (free_dominance_info): Add overload with
322 (dom_info_state): Likewise.
323 (dom_info_available_p): Likewise.
324 * basic-block.h (free_dominance_info, dom_info_state,
325 dom_info_available_p): Declare overloads.
326 * passes.c (execute_function_todo): Verify that verifiers
327 don't change dominator info state. Drop dominator info
328 for IPA pass invocations.
329 * cgraph.c (release_function_body): Restore asserts that
330 dominator information is released.
332 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
334 * doc/invoke.texi: Fix typo.
335 * tree-vrp.c: Fix typos.
336 * gimple.c (infer_nonnull_range): Reorder operands of an &&
339 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
341 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
343 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
345 * config/aarch64/aarch64-builtins.c
346 (aarch64_types_storestruct_lane_qualifiers): New.
347 (TYPES_STORESTRUCT_LANE): Likewise.
348 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
349 (st3_lane): Likewise.
350 (st4_lane): Likewise.
351 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
352 (vec_store_lanesci_lane<mode>): Likewise.
353 (vec_store_lanesxi_lane<mode>): Likewise.
354 (aarch64_st2_lane<VQ:mode>): Likewise.
355 (aarch64_st3_lane<VQ:mode>): Likewise.
356 (aarch64_st4_lane<VQ:mode>): Likewise.
357 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
358 * config/aarch64/arm_neon.h
359 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
360 use new macro arguments.
361 (__ST3_LANE_FUNC): Likewise.
362 (__ST4_LANE_FUNC): Likewise.
363 * config/aarch64/iterators.md (V_TWO_ELEM): New.
364 (V_THREE_ELEM): Likewise.
365 (V_FOUR_ELEM): Likewise.
367 2014-04-28 David Malcolm <dmalcolm@redhat.com>
369 * doc/gimple.texi: Replace the description of the now-defunct
370 union gimple_statement_d with a diagram showing the
371 gimple_statement_base class hierarchy and its relationships to
372 the GSS_ and GIMPLE_ enums.
374 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
376 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
377 * config/aarch64/aarch64.c
378 (aarch64_cannot_change_mode_class): Weaken conditions.
379 (aarch64_modes_tieable_p): New.
380 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
382 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
384 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
385 (loadsync_<mode>): Change mode.
386 (load_quadpti, store_quadpti): New.
387 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
388 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
390 2014-04-28 Martin Jambor <mjambor@suse.cz>
392 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
393 same alias type as the original statement.
394 (subreplacement_assignment_data): New type.
395 (handle_unscalarized_data_in_subtree): New type of parameter,
396 generate new memory accesses with same alias type as the original
398 (load_assign_lhs_subreplacements): Likewise.
399 (sra_modify_constructor_assign): Generate new memory accesses with
400 same alias type as the original statement.
402 2014-04-28 Richard Biener <rguenther@suse.de>
404 * tree-pass.h (TODO_verify_il): Define.
405 (TODO_verify_all): Complete properly.
406 * passes.c (execute_function_todo): Move existing loop-closed
407 SSA verification under TODO_verify_il.
408 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
409 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
410 Fix tree sharing issue.
412 2014-04-28 Richard Biener <rguenther@suse.de>
415 * builtins.def (DEF_C11_BUILTIN): Add.
416 (BUILT_IN_ALIGNED_ALLOC): Likewise.
417 * coretypes.h (enum function_class): Add function_c11_misc.
418 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
419 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
420 (call_may_clobber_ref_p_1): Likewise.
421 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
422 (mark_all_reaching_defs_necessary_1): Likewise.
423 (propagate_necessity): Likewise.
424 (eliminate_unnecessary_stmts): Likewise.
425 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
427 2014-04-28 Richard Biener <rguenther@suse.de>
429 * tree-vrp.c (vrp_var_may_overflow): Remove.
430 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
431 with overflow immediately bump to one before that value and
432 let iteration figure out overflow status.
434 2014-04-28 Richard Biener <rguenther@suse.de>
436 * configure.ac: Do valgrind header checks unconditionally.
437 Add --enable-valgrind-annotations.
438 * system.h: Guard valgrind header inclusion with
439 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
440 * alloc-pool.c (pool_alloc, pool_free): Use
441 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
442 to guard possibly dead code.
443 * config.in: Regenerated.
444 * configure: Likewise.
446 2014-04-28 Jeff Law <law@redhat.com>
448 PR tree-optimization/60902
449 * tree-ssa-threadedge.c
450 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
451 over real defs when invalidating outputs from statements that do not
452 produce useful outputs for threading.
454 2014-04-28 Richard Biener <rguenther@suse.de>
456 PR tree-optimization/60979
457 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
458 SCOPs that end in a block with a successor with abnormal
461 2014-04-28 Richard Biener <rguenther@suse.de>
463 * tree-pass.h (execute_pass_list): Adjust prototype.
464 * passes.c (pass_manager::execute_early_local_passes):
466 (do_per_function): Change callback signature, push all actual
467 work to the callbals.
468 (do_per_function_toporder): Likewise.
469 (execute_function_dump): Adjust.
470 (execute_function_todo): Likewise.
471 (clear_last_verified): Likewise.
472 (verify_curr_properties): Likewise.
473 (update_properties_after_pass): Likewise.
474 (execute_pass_list_1): Split out from ...
475 (execute_pass_list): ... here. Adjust.
476 (execute_ipa_pass_list): Likewise.
477 * cgraphunit.c (cgraph_add_new_function): Adjust.
478 (analyze_function): Likewise.
479 (expand_function): Likewise.
480 * cgraph.c (release_function_body): Free dominance info
481 here instead of asserting it was magically freed elsewhere.
483 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
485 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
486 * configure: Regenerate.
487 * config/sparc/sparc.opt (muser-mode): New option.
488 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
490 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
491 * doc/invoke.texi (SPARC options): Document -muser-mode.
493 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
495 * cselib.c (find_slot_memmode): Delete.
496 (cselib_hasher): Change compare_type to a struct.
497 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
499 (preserve_constants_and_equivs): Adjust for new compare_type.
500 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
501 (wrap_constant): Delete.
502 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
504 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
506 * doc/install.texi (Building with profile feedback): Remove
509 2014-04-26 Tom de Vries <tom@codesourcery.com>
511 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
514 2014-04-25 Cary Coutant <ccoutant@google.com>
517 * dwarf2out.c (should_move_die_to_comdat): A type definition
518 can contain a subprogram definition, but don't move it to a
520 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
521 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
523 (clone_tree_hash): Rename to...
524 (clone_tree_partial): ...this; change callers. Copy
525 DW_TAG_subprogram DIEs as declarations.
526 (copy_decls_walk): Don't copy children of a declaration into a
529 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
532 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
535 2014-04-25 Jiong Wang <jiong.wang@arm.com>
537 * config/arm/predicates.md (call_insn_operand): Add long_call check.
538 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
540 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
543 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
545 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
548 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
550 PR tree-optimization/60930
551 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
552 creating a multiply candidate by folding two constant
553 multiplicands when the result overflows.
555 2014-04-25 Jakub Jelinek <jakub@redhat.com>
557 PR tree-optimization/60960
558 * tree-vect-generic.c (expand_vector_operation): Only call
559 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
561 2014-04-25 Tom de Vries <tom@codesourcery.com>
563 * expr.c (clobber_reg_mode): New function.
564 * expr.h (clobber_reg): New function.
566 2014-04-25 Tom de Vries <tom@codesourcery.com>
568 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
571 2014-04-25 Radovan Obradovic <robradovic@mips.com>
572 Tom de Vries <tom@codesourcery.com>
574 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
576 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
577 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
578 new argument to find_all_hard_reg_sets call.
580 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
582 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
583 Use HOST_WIDE_INT_C for mask literal.
584 (aarch_rev16_shleft_mask_imm_p): Likewise.
586 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
589 * config/sparc/sparc.md (ashlsi3_extend): Delete.
591 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
593 PR preprocessor/56540
594 * config/i386/i386-c.c (ix86_target_macros): Define
595 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
597 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
599 * configure.ac (tga_func): Remove.
600 (LIB_TLS_SPEC): Remove.
601 * configure: Regenerate.
602 * config.in: Regenerate.
603 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
605 2014-04-25 Richard Biener <rguenther@suse.de>
608 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
609 call stmt use/clobber sets during stmt walk instead of
610 walking the possibly incomplete set of caller edges.
612 2014-04-25 Richard Biener <rguenther@suse.de>
615 * passes.c (apply_ipa_transforms): Inline into only caller ...
616 (execute_one_pass): ... here. Properly bring in function
617 bodies for nodes we want to apply IPA transforms to.
619 2014-04-24 Cong Hou <congh@google.com>
621 PR tree-optimization/60896
622 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
623 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
624 (vect_mark_pattern_stmts): Set the def type of all statements in
625 PATTERN_DEF_SEQ as vect_internal_def.
627 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
629 * doc/extend.texi (PowerPC Built-in Functions): Document new
630 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
632 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
634 * config/rs6000/predicates.md (const_0_to_3_operand): New
635 predicate to match 0..3 integer constants.
637 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
638 to support adding miscellaneous builtin functions.
639 (BU_DFP_MISC_2): Likewise.
640 (BU_P7_MISC_1): Likewise.
641 (BU_P7_MISC_2): Likewise.
642 (BU_P8V_MISC_3): Likewise.
643 (BU_MISC_1): Likewise.
644 (BU_MISC_2): Likewise.
645 (DIVWE): Add extended divide builtin functions.
653 (DXEX): Add decimal floating-point builtin functions.
665 (CDTBCD): Add new BCD builtin functions.
669 (BCDADD_LT): Likewise.
670 (BCDADD_EQ): Likewise.
671 (BCDADD_GT): Likewise.
672 (BCDADD_OV): Likewise.
674 (BCDSUB_LT): Likewise.
675 (BCDSUB_EQ): Likewise.
676 (BCDSUB_GT): Likewise.
677 (BCDSUB_OV): Likewise.
678 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
679 (UNPACK_TD): Likewise.
681 (UNPACK_TF): Likewise.
682 (UNPACK_TF_0): Likewise.
683 (UNPACK_TF_1): Likewise.
684 (PACK_V1TI): Likewise.
685 (UNPACK_V1TI): Likewise.
687 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
688 support for decimal floating point builtin functions.
689 (rs6000_expand_ternop_builtin): Add checks for the new builtin
690 functions that take constant arguments.
691 (rs6000_invalid_builtin): Add decimal floating point builtin
693 (rs6000_init_builtins): Setup long double, _Decimal64, and
694 _Decimal128 types for new builtin functions.
695 (builtin_function_type): Set the unsigned flags appropriately for
696 the new builtin functions.
697 (rs6000_opt_masks): Add support for decimal floating point builtin
700 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
701 floating point builtin functions.
702 (RS6000_BTM_COMMON): Likewise.
703 (RS6000_BTI_long_double): Likewise.
704 (RS6000_BTI_dfloat64): Likewise.
705 (RS6000_BTI_dfloat128): Likewise.
706 (long_double_type_internal_node): Likewise.
707 (dfloat64_type_internal_node): Likewise.
708 (dfloat128_type_internal_node): Likewise.
710 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
711 2.07 bcd arithmetic instructions.
712 (UNSPEC_BCDSUB): Likewise.
713 (UNSPEC_BCD_OVERFLOW): Likewise.
714 (UNSPEC_BCD_ADD_SUB): Likewise.
715 (bcd_add_sub): Likewise.
716 (BCD_TEST): Likewise.
717 (bcd<bcd_add_sub>): Likewise.
718 (bcd<bcd_add_sub>_test): Likewise.
719 (bcd<bcd_add_sub>_test2): Likewise.
720 (bcd<bcd_add_sub>_<code>): Likewise.
721 (peephole2 for combined bcd ops): Likewise.
723 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
724 decimal floating point builtin functions.
725 (UNSPEC_DENBCD): Likewise.
726 (UNSPEC_DXEX): Likewise.
727 (UNSPEC_DIEX): Likewise.
728 (UNSPEC_DSCLI): Likewise.
729 (UNSPEC_DSCRI): Likewise.
730 (D64_D128): Likewise.
731 (dfp_suffix): Likewise.
732 (dfp_ddedpd_<mode>): Likewise.
733 (dfp_denbcd_<mode>): Likewise.
734 (dfp_dxex_<mode>): Likewise.
735 (dfp_diex_<mode>): Likewise.
736 (dfp_dscli_<mode>): Likewise.
737 (dfp_dscri_<mode>): Likewise.
739 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
741 (UNSPEC_CDTBCD): Likewise.
742 (UNSPEC_CBCDTD): Likewise.
743 (UNSPEC_DIVE): Add support for new extended divide builtin
745 (UNSPEC_DIVEO): Likewise.
746 (UNSPEC_DIVEU): Likewise.
747 (UNSPEC_DIVEUO): Likewise.
748 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
749 pack/unpack 128-bit types.
750 (UNSPEC_PACK_128BIT): Likewise.
751 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
752 (udiv<mode>3): Use idiv_ldiv mode attribute.
753 (div<mode>3): Likewise.
754 (addg6s): Add new BCD builtin functions.
757 (UNSPEC_DIV_EXTEND): Add support for new extended divide
759 (div_extend): Likewise.
760 (div<div_extend>_<mode>"): Likewise.
761 (FP128_64): Add support for new builtin functions to pack/unpack
763 (unpack<mode>): Likewise.
764 (unpacktf_0): Likewise.
765 (unpacktf_1): Likewise.
766 (unpack<mode>_dm): Likewise.
767 (unpack<mode>_nodm): Likewise.
768 (pack<mode>): Likewise.
769 (unpackv1ti): Likewise.
770 (packv1ti): Likewise.
772 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
774 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
777 2014-04-24 Jakub Jelinek <jakub@redhat.com>
779 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
780 * gimplify.c (omp_is_private): Change last argument's type to int.
781 Only diagnose lastprivate if the simd argument is 1, only diagnose
782 linear if the simd argument is 2.
783 (gimplify_omp_for): Adjust omp_is_private callers. When adding
784 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
785 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
786 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
787 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
788 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
789 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
790 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
791 * tree-nested.c (convert_nonlocal_omp_clauses,
792 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
794 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
797 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
800 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
802 * flag-types.h (enum ivar_visibility): Add.
804 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
806 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
809 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
811 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
813 2014-04-24 Radovan Obradovic <robradovic@mips.com>
814 Tom de Vries <tom@codesourcery.com>
816 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
817 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
819 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
820 * emit-rtl.c (try_split): Same.
822 2014-04-24 Radovan Obradovic <robradovic@mips.com>
823 Tom de Vries <tom@codesourcery.com>
825 * common.opt (fuse-caller-save): New option.
827 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
829 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
832 2014-04-24 Richard Biener <rguenther@suse.de>
834 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
835 during TER and instead use the sepops interface for expanding
836 non-GIMPLE_SINGLE_RHS.
838 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
840 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
841 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
843 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
845 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
846 assembler 64-bit option.
847 * configure: Regenerate.
849 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
851 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
852 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
853 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
854 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
855 (TARGET_CRYPTO): Take TARGET_SIMD into account.
857 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
859 * config/aarch64/aarch64-builtins.c
860 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
861 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
862 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
863 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
865 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
866 (Vrevsuff): New mode attribute.
868 2014-04-24 Terry Guo <terry.guo@arm.com>
870 * config/arm/arm.h (machine_function): Define variable
871 after_arm_reorg here.
872 * config/arm/arm.c (after_arm_reorg): Remove the definition.
873 (arm_split_constant): Update the way to access variable
876 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
878 2014-04-23 Tom de Vries <tom@codesourcery.com>
880 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
882 2014-04-23 David Malcolm <dmalcolm@redhat.com>
884 * is-a.h: Update comments to reflect the following changes to the
885 "pointerness" of the API, making the template parameter match the
886 return type, allowing use of is-a.h with typedefs of pointers.
887 (is_a_helper::cast): Return a T rather then a pointer to a T, so
888 that the return type matches the parameter to the is_a_helper.
890 (dyn_cast): Likewise.
892 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
893 pointer from the is-a.h API.
895 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
896 (is_a_helper <cgraph_node *>::test): ...this, matching change to
898 (is_a_helper <varpool_node>::test): Likewise, convert to...
899 (is_a_helper <varpool_node *>::test): ...this.
901 (varpool_first_variable): Update for removal of implicit pointer
903 (varpool_next_variable): Likewise.
904 (varpool_first_static_initializer): Likewise.
905 (varpool_next_static_initializer): Likewise.
906 (varpool_first_defined_variable): Likewise.
907 (varpool_next_defined_variable): Likewise.
908 (cgraph_first_defined_function): Likewise.
909 (cgraph_next_defined_function): Likewise.
910 (cgraph_first_function): Likewise.
911 (cgraph_next_function): Likewise.
912 (cgraph_first_function_with_gimple_body): Likewise.
913 (cgraph_next_function_with_gimple_body): Likewise.
914 (cgraph_alias_target): Likewise.
915 (varpool_alias_target): Likewise.
916 (cgraph_function_or_thunk_node): Likewise.
917 (varpool_variable_node): Likewise.
918 (symtab_real_symbol_p): Likewise.
919 * cgraphunit.c (referred_to_p): Likewise.
920 (analyze_functions): Likewise.
921 (handle_alias_pairs): Likewise.
922 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
923 * gimple-ssa.h (gimple_vuse_op): Likewise.
924 (gimple_vdef_op): Likewise.
925 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
926 * gimple.c (gimple_build_asm_1): Likewise.
927 (gimple_build_try): Likewise.
928 (gimple_build_resx): Likewise.
929 (gimple_build_eh_dispatch): Likewise.
930 (gimple_build_omp_for): Likewise.
931 (gimple_omp_for_set_clauses): Likewise.
933 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
934 (is_a_helper <gimple_statement_asm *>::test): ...this.
935 (is_a_helper <gimple_statement_bind>::test): Convert to...
936 (is_a_helper <gimple_statement_bind *>::test): ...this.
937 (is_a_helper <gimple_statement_call>::test): Convert to...
938 (is_a_helper <gimple_statement_call *>::test): ...this.
939 (is_a_helper <gimple_statement_catch>::test): Convert to...
940 (is_a_helper <gimple_statement_catch *>::test): ...this.
941 (is_a_helper <gimple_statement_resx>::test): Convert to...
942 (is_a_helper <gimple_statement_resx *>::test): ...this.
943 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
944 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
945 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
946 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
947 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
948 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
949 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
950 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
951 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
952 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
953 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
954 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
955 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
956 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
957 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
958 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
959 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
960 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
961 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
962 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
963 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
964 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
965 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
966 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
967 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
968 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
969 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
970 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
971 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
972 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
973 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
974 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
975 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
976 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
977 (is_a_helper <gimple_statement_phi>::test): Convert to...
978 (is_a_helper <gimple_statement_phi *>::test): ...this.
979 (is_a_helper <gimple_statement_transaction>::test): Convert to...
980 (is_a_helper <gimple_statement_transaction *>::test): ...this.
981 (is_a_helper <gimple_statement_try>::test): Convert to...
982 (is_a_helper <gimple_statement_try *>::test): ...this.
983 (is_a_helper <gimple_statement_wce>::test): Convert to...
984 (is_a_helper <gimple_statement_wce *>::test): ...this.
985 (is_a_helper <const gimple_statement_asm>::test): Convert to...
986 (is_a_helper <const gimple_statement_asm *>::test): ...this.
987 (is_a_helper <const gimple_statement_bind>::test): Convert to...
988 (is_a_helper <const gimple_statement_bind *>::test): ...this.
989 (is_a_helper <const gimple_statement_call>::test): Convert to...
990 (is_a_helper <const gimple_statement_call *>::test): ...this.
991 (is_a_helper <const gimple_statement_catch>::test): Convert to...
992 (is_a_helper <const gimple_statement_catch *>::test): ...this.
993 (is_a_helper <const gimple_statement_resx>::test): Convert to...
994 (is_a_helper <const gimple_statement_resx *>::test): ...this.
995 (is_a_helper <const gimple_statement_eh_dispatch>::test):
997 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
999 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
1001 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
1002 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
1004 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
1006 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
1008 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
1010 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1012 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1013 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1015 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1016 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1018 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1019 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1020 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1021 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1023 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1024 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1026 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1027 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1029 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1030 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1032 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1033 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1035 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1036 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1038 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1039 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1041 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1042 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1043 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1044 (is_a_helper <const gimple_statement_transaction>::test): Convert
1046 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1047 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1049 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1050 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1051 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1052 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1054 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1056 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1058 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1060 (gimple_use_ops): Update for removal of implicit pointer from the
1062 (gimple_set_use_ops): Likewise.
1063 (gimple_vuse): Likewise.
1064 (gimple_vdef): Likewise.
1065 (gimple_vuse_ptr): Likewise.
1066 (gimple_vdef_ptr): Likewise.
1067 (gimple_set_vuse): Likewise.
1068 (gimple_set_vdef): Likewise.
1069 (gimple_omp_return_set_lhs): Likewise.
1070 (gimple_omp_return_lhs): Likewise.
1071 (gimple_omp_return_lhs_ptr): Likewise.
1072 (gimple_call_fntype): Likewise.
1073 (gimple_call_set_fntype): Likewise.
1074 (gimple_call_set_internal_fn): Likewise.
1075 (gimple_call_use_set): Likewise.
1076 (gimple_call_clobber_set): Likewise.
1077 (gimple_bind_vars): Likewise.
1078 (gimple_bind_set_vars): Likewise.
1079 (gimple_bind_body_ptr): Likewise.
1080 (gimple_bind_set_body): Likewise.
1081 (gimple_bind_add_stmt): Likewise.
1082 (gimple_bind_block): Likewise.
1083 (gimple_bind_set_block): Likewise.
1084 (gimple_asm_ninputs): Likewise.
1085 (gimple_asm_noutputs): Likewise.
1086 (gimple_asm_nclobbers): Likewise.
1087 (gimple_asm_nlabels): Likewise.
1088 (gimple_asm_input_op): Likewise.
1089 (gimple_asm_input_op_ptr): Likewise.
1090 (gimple_asm_output_op): Likewise.
1091 (gimple_asm_output_op_ptr): Likewise.
1092 (gimple_asm_set_output_op): Likewise.
1093 (gimple_asm_clobber_op): Likewise.
1094 (gimple_asm_set_clobber_op): Likewise.
1095 (gimple_asm_label_op): Likewise.
1096 (gimple_asm_set_label_op): Likewise.
1097 (gimple_asm_string): Likewise.
1098 (gimple_catch_types): Likewise.
1099 (gimple_catch_types_ptr): Likewise.
1100 (gimple_catch_handler_ptr): Likewise.
1101 (gimple_catch_set_types): Likewise.
1102 (gimple_catch_set_handler): Likewise.
1103 (gimple_eh_filter_types): Likewise.
1104 (gimple_eh_filter_types_ptr): Likewise.
1105 (gimple_eh_filter_failure_ptr): Likewise.
1106 (gimple_eh_filter_set_types): Likewise.
1107 (gimple_eh_filter_set_failure): Likewise.
1108 (gimple_eh_must_not_throw_fndecl): Likewise.
1109 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1110 (gimple_eh_else_n_body_ptr): Likewise.
1111 (gimple_eh_else_e_body_ptr): Likewise.
1112 (gimple_eh_else_set_n_body): Likewise.
1113 (gimple_eh_else_set_e_body): Likewise.
1114 (gimple_try_eval_ptr): Likewise.
1115 (gimple_try_cleanup_ptr): Likewise.
1116 (gimple_try_set_eval): Likewise.
1117 (gimple_try_set_cleanup): Likewise.
1118 (gimple_wce_cleanup_ptr): Likewise.
1119 (gimple_wce_set_cleanup): Likewise.
1120 (gimple_phi_capacity): Likewise.
1121 (gimple_phi_num_args): Likewise.
1122 (gimple_phi_result): Likewise.
1123 (gimple_phi_result_ptr): Likewise.
1124 (gimple_phi_set_result): Likewise.
1125 (gimple_phi_arg): Likewise.
1126 (gimple_phi_set_arg): Likewise.
1127 (gimple_resx_region): Likewise.
1128 (gimple_resx_set_region): Likewise.
1129 (gimple_eh_dispatch_region): Likewise.
1130 (gimple_eh_dispatch_set_region): Likewise.
1131 (gimple_omp_critical_name): Likewise.
1132 (gimple_omp_critical_name_ptr): Likewise.
1133 (gimple_omp_critical_set_name): Likewise.
1134 (gimple_omp_for_clauses): Likewise.
1135 (gimple_omp_for_clauses_ptr): Likewise.
1136 (gimple_omp_for_set_clauses): Likewise.
1137 (gimple_omp_for_collapse): Likewise.
1138 (gimple_omp_for_index): Likewise.
1139 (gimple_omp_for_index_ptr): Likewise.
1140 (gimple_omp_for_set_index): Likewise.
1141 (gimple_omp_for_initial): Likewise.
1142 (gimple_omp_for_initial_ptr): Likewise.
1143 (gimple_omp_for_set_initial): Likewise.
1144 (gimple_omp_for_final): Likewise.
1145 (gimple_omp_for_final_ptr): Likewise.
1146 (gimple_omp_for_set_final): Likewise.
1147 (gimple_omp_for_incr): Likewise.
1148 (gimple_omp_for_incr_ptr): Likewise.
1149 (gimple_omp_for_set_incr): Likewise.
1150 (gimple_omp_for_pre_body_ptr): Likewise.
1151 (gimple_omp_for_set_pre_body): Likewise.
1152 (gimple_omp_parallel_clauses): Likewise.
1153 (gimple_omp_parallel_clauses_ptr): Likewise.
1154 (gimple_omp_parallel_set_clauses): Likewise.
1155 (gimple_omp_parallel_child_fn): Likewise.
1156 (gimple_omp_parallel_child_fn_ptr): Likewise.
1157 (gimple_omp_parallel_set_child_fn): Likewise.
1158 (gimple_omp_parallel_data_arg): Likewise.
1159 (gimple_omp_parallel_data_arg_ptr): Likewise.
1160 (gimple_omp_parallel_set_data_arg): Likewise.
1161 (gimple_omp_task_clauses): Likewise.
1162 (gimple_omp_task_clauses_ptr): Likewise.
1163 (gimple_omp_task_set_clauses): Likewise.
1164 (gimple_omp_task_child_fn): Likewise.
1165 (gimple_omp_task_child_fn_ptr): Likewise.
1166 (gimple_omp_task_set_child_fn): Likewise.
1167 (gimple_omp_task_data_arg): Likewise.
1168 (gimple_omp_task_data_arg_ptr): Likewise.
1169 (gimple_omp_task_set_data_arg): Likewise.
1170 (gimple_omp_taskreg_clauses): Likewise.
1171 (gimple_omp_taskreg_clauses_ptr): Likewise.
1172 (gimple_omp_taskreg_set_clauses): Likewise.
1173 (gimple_omp_taskreg_child_fn): Likewise.
1174 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1175 (gimple_omp_taskreg_set_child_fn): Likewise.
1176 (gimple_omp_taskreg_data_arg): Likewise.
1177 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1178 (gimple_omp_taskreg_set_data_arg): Likewise.
1179 (gimple_omp_task_copy_fn): Likewise.
1180 (gimple_omp_task_copy_fn_ptr): Likewise.
1181 (gimple_omp_task_set_copy_fn): Likewise.
1182 (gimple_omp_task_arg_size): Likewise.
1183 (gimple_omp_task_arg_size_ptr): Likewise.
1184 (gimple_omp_task_set_arg_size): Likewise.
1185 (gimple_omp_task_arg_align): Likewise.
1186 (gimple_omp_task_arg_align_ptr): Likewise.
1187 (gimple_omp_task_set_arg_align): Likewise.
1188 (gimple_omp_single_clauses): Likewise.
1189 (gimple_omp_single_clauses_ptr): Likewise.
1190 (gimple_omp_single_set_clauses): Likewise.
1191 (gimple_omp_target_clauses): Likewise.
1192 (gimple_omp_target_clauses_ptr): Likewise.
1193 (gimple_omp_target_set_clauses): Likewise.
1194 (gimple_omp_target_child_fn): Likewise.
1195 (gimple_omp_target_child_fn_ptr): Likewise.
1196 (gimple_omp_target_set_child_fn): Likewise.
1197 (gimple_omp_target_data_arg): Likewise.
1198 (gimple_omp_target_data_arg_ptr): Likewise.
1199 (gimple_omp_target_set_data_arg): Likewise.
1200 (gimple_omp_teams_clauses): Likewise.
1201 (gimple_omp_teams_clauses_ptr): Likewise.
1202 (gimple_omp_teams_set_clauses): Likewise.
1203 (gimple_omp_sections_clauses): Likewise.
1204 (gimple_omp_sections_clauses_ptr): Likewise.
1205 (gimple_omp_sections_set_clauses): Likewise.
1206 (gimple_omp_sections_control): Likewise.
1207 (gimple_omp_sections_control_ptr): Likewise.
1208 (gimple_omp_sections_set_control): Likewise.
1209 (gimple_omp_for_set_cond): Likewise.
1210 (gimple_omp_for_cond): Likewise.
1211 (gimple_omp_atomic_store_set_val): Likewise.
1212 (gimple_omp_atomic_store_val): Likewise.
1213 (gimple_omp_atomic_store_val_ptr): Likewise.
1214 (gimple_omp_atomic_load_set_lhs): Likewise.
1215 (gimple_omp_atomic_load_lhs): Likewise.
1216 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1217 (gimple_omp_atomic_load_set_rhs): Likewise.
1218 (gimple_omp_atomic_load_rhs): Likewise.
1219 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1220 (gimple_omp_continue_control_def): Likewise.
1221 (gimple_omp_continue_control_def_ptr): Likewise.
1222 (gimple_omp_continue_set_control_def): Likewise.
1223 (gimple_omp_continue_control_use): Likewise.
1224 (gimple_omp_continue_control_use_ptr): Likewise.
1225 (gimple_omp_continue_set_control_use): Likewise.
1226 (gimple_transaction_body_ptr): Likewise.
1227 (gimple_transaction_label): Likewise.
1228 (gimple_transaction_label_ptr): Likewise.
1229 (gimple_transaction_set_body): Likewise.
1230 (gimple_transaction_set_label): Likewise.
1232 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1233 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1234 * ipa-ref.c (ipa_record_reference): Likewise.
1235 * ipa-reference.c (analyze_function): Likewise.
1236 (ipa_reference_write_optimization_summary): Likewise.
1237 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1238 (address_taken_from_non_vtable_p): Likewise.
1239 (comdat_can_be_unshared_p_1): Likewise.
1240 * lto-cgraph.c (lto_output_ref): Likewise.
1241 (add_references): Likewise.
1242 (compute_ltrans_boundary): Likewise.
1243 (output_symtab): Likewise.
1244 (input_ref): Likewise.
1245 (input_cgraph_1): Likewise.
1246 (output_cgraph_opt_summary): Likewise.
1247 * lto-streamer-out.c (lto_output): Likewise.
1248 (output_symbol_p): Likewise.
1249 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1250 (lsei_start_function_in_partition): Likewise.
1251 (lsei_next_variable_in_partition): Likewise.
1252 (lsei_start_variable_in_partition): Likewise.
1253 * symtab.c (insert_to_assembler_name_hash): Likewise.
1254 (unlink_from_assembler_name_hash): Likewise.
1255 (symtab_unregister_node): Likewise.
1256 (symtab_remove_node): Likewise.
1257 (dump_symtab_node): Likewise.
1258 (verify_symtab_base): Likewise.
1259 (verify_symtab_node): Likewise.
1260 (symtab_make_decl_local): Likewise.
1261 (symtab_alias_ultimate_target): Likewise.
1262 (symtab_resolve_alias): Likewise.
1263 (symtab_get_symbol_partitioning_class): Likewise.
1264 * tree-phinodes.c (allocate_phi_node): Likewise.
1265 (reserve_phi_args_for_new_edge): Likewise.
1266 (remove_phi_args): Likewise.
1267 * varpool.c (varpool_node_for_asm): Likewise.
1268 (varpool_remove_unreferenced_decls): Likewise.
1270 2014-04-23 Jeff Law <law@redhat.com>
1272 PR tree-optimization/60902
1273 * tree-ssa-threadedge.c
1274 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1275 invalidate outputs from statements that do not produce useful
1276 outputs for threading.
1278 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1280 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1281 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1282 machine descriptions for Stack Smashing Protector.
1284 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1286 * aarch64.md (<optab>_rol<mode>3): New pattern.
1287 (<optab>_rolsi3_uxtw): Likewise.
1288 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1290 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1292 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1293 (arm_cortex_a12_tune): Likewise.
1295 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1297 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1299 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1301 * config/arm/arm.md (arm_rev16si2): New pattern.
1302 (arm_rev16si2_alt): Likewise.
1303 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1305 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1307 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1308 (rev16<mode>2_alt): Likewise.
1309 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1310 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1311 (aarch_rev16_shleft_mask_imm_p): Likewise.
1312 (aarch_rev16_p_1): Likewise.
1313 (aarch_rev16_p): Likewise.
1314 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1315 (aarch_rev16_shright_mask_imm_p): Likewise.
1316 (aarch_rev16_shleft_mask_imm_p): Likewise.
1318 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1320 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1321 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1323 (cortex_a53_extra_costs): Likewise.
1324 (cortex_a57_extra_costs): Likewise.
1325 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1326 (cortexa7_extra_costs): Likewise.
1327 (cortexa8_extra_costs): Likewise.
1328 (cortexa12_extra_costs): Likewise.
1329 (cortexa15_extra_costs): Likewise.
1330 (v7m_extra_costs): Likewise.
1331 (arm_new_rtx_costs): Handle BSWAP.
1333 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1335 * config/arm/arm.c (cortexa8_extra_costs): New table.
1336 (arm_cortex_a8_tune): New tuning struct.
1337 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1339 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1341 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1343 2014-04-23 Richard Biener <rguenther@suse.de>
1345 * Makefile.in (OBJS): Remove loop-unswitch.o.
1346 * loop-unswitch.c: Delete.
1347 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1348 * passes.def (pass_rtl_unswitch): Likewise.
1349 * loop-init.c (gate_rtl_unswitch): Likewise.
1350 (rtl_unswitch): Likewise.
1351 (pass_data_rtl_unswitch): Likewise.
1352 (pass_rtl_unswitch): Likewise.
1353 (make_pass_rtl_unswitch): Likewise.
1354 * rtl.h (reversed_condition): Likewise.
1355 (compare_and_jump_seq): Likewise.
1356 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1358 * loop-unroll.c (compare_and_jump_seq): Likewise.
1360 2014-04-23 Richard Biener <rguenther@suse.de>
1362 PR tree-optimization/60903
1363 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1364 commented code block.
1365 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1366 loop flags to newly created BBs and edges.
1368 2014-04-23 Nick Clifton <nickc@redhat.com>
1370 * config/msp430/msp430.c (msp430_handle_option): Move function
1372 (msp430_option_override): Simplify mcu and mcpu option handling.
1373 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1374 support for -mhwmult command line option.
1375 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1376 -mhwmult command line option.
1377 (msp430_hwmult_enabled): Delete.
1378 (msp43o_output_labelref): Add support for -mhwmult command line
1380 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1381 (umulsidi3): Likewise.
1382 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1383 (mcpu, mlarge, msmall): Likewise.
1384 (mhwmult): New option.
1385 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1387 (msp430_is_f5_mcu): Remove prototype.
1388 (msp430_use_f5_series_hwmult): Add prototype.
1389 * config/msp430/msp430-opts.h: New file.
1390 * common/config/msp430: New directory.
1391 * common/config/msp430/msp430-common.c: New file.
1392 * config.gcc (msp430): Remove target_has_targetm_common.
1393 * doc/invoke.texi: Document -mhwmult command line option.
1395 2014-04-23 Nick Clifton <nickc@redhat.com>
1397 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1398 default-manifest.o if it can be found in the search path.
1399 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1401 2014-04-23 Terry Guo <terry.guo@arm.com>
1403 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1405 2014-04-23 Richard Biener <rguenther@suse.de>
1408 * tree-inline.c (declare_return_variable): Use mark_addressable.
1410 2014-04-23 Richard Biener <rguenther@suse.de>
1413 * loop-init.c (loop_optimizer_init): Make sure to apply
1414 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1416 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1419 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1421 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1422 if flag_sanitize_undefined_trap_on_error.
1423 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1424 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1425 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1426 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1427 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1428 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1429 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1430 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1431 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1432 * ubsan.c (ubsan_instrument_unreachable): Return
1433 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1434 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1435 if flag_sanitize_undefined_trap_on_error and
1436 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1437 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1438 instrument_bool_enum_load): Emit __builtin_trap () if
1439 flag_sanitize_undefined_trap_on_error and
1440 __builtin_handle_*_abort () if !flag_sanitize_recover.
1441 * doc/invoke.texi (-fsanitize-recover,
1442 -fsanitize-undefined-trap-on-error): Document.
1444 2014-04-22 Christian Bruel <christian.bruel@st.com>
1446 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1447 Force immediates to SImode.
1449 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1451 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1453 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1454 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1455 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1456 (nios2_fpu_insn): Add entry for round.
1457 (N2FPU_NO_ERRNO_P): Define.
1458 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1460 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1462 2014-04-22 Richard Henderson <rth@redhat.com>
1464 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1465 (add<GPI>3_compare0): Remove leading * from name.
1466 (add<GPI>3_carryin): Likewise.
1467 (sub<GPI>3_compare0): Likewise.
1468 (sub<GPI>3_carryin): Likewise.
1469 (<su_optab>mulditi3): New expander.
1470 (multi3): New expander.
1471 (madd<GPI>): Remove leading * from name.
1473 2014-04-22 Martin Jambor <mjambor@suse.cz>
1475 * cgraphclones.c (cgraph_function_versioning): Copy
1476 ipa_transforms_to_apply instead of asserting it is empty.
1478 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1481 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1482 on count_exp to get mode.
1484 2014-04-22 Andrew Pinski <apinski@cavium.com>
1486 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1487 Handle TLS for ILP32.
1488 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1489 (tlsie_small_<mode>): this and handle PTR.
1490 (tlsie_small_sidi): New pattern.
1491 (tlsle_small): Change to an expand to handle ILP32.
1492 (tlsle_small_<mode>): New pattern.
1493 (tlsdesc_small): Rename to ...
1494 (tlsdesc_small_<mode>): this and handle PTR.
1496 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1498 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1500 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1502 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1503 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1504 (aarch64_types_signed_poly_qualifiers): Likewise.
1505 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1506 (aarch64_types_poly_signed_qualifiers): Likewise.
1507 (TYPES_REINTERP_SS): Type macro added.
1508 (TYPES_REINTERP_SU): Likewise.
1509 (TYPES_REINTERP_SP): Likewise.
1510 (TYPES_REINTERP_US): Likewise.
1511 (TYPES_REINTERP_PS): Likewise.
1512 (aarch64_fold_builtin): New expression folding added.
1513 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1514 Declarations removed.
1515 (REINTERP_SS): Declarations added.
1516 (REINTERP_US): Likewise.
1517 (REINTERP_PS): Likewise.
1518 (REINTERP_SU): Likewise.
1519 (REINTERP_SP): Likewise.
1520 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1521 (vreinterpretq_p8_f64): Likewise.
1522 (vreinterpret_p16_f64): Likewise.
1523 (vreinterpretq_p16_f64): Likewise.
1524 (vreinterpret_f32_f64): Likewise.
1525 (vreinterpretq_f32_f64): Likewise.
1526 (vreinterpret_f64_f32): Likewise.
1527 (vreinterpret_f64_p8): Likewise.
1528 (vreinterpret_f64_p16): Likewise.
1529 (vreinterpret_f64_s8): Likewise.
1530 (vreinterpret_f64_s16): Likewise.
1531 (vreinterpret_f64_s32): Likewise.
1532 (vreinterpret_f64_s64): Likewise.
1533 (vreinterpret_f64_u8): Likewise.
1534 (vreinterpret_f64_u16): Likewise.
1535 (vreinterpret_f64_u32): Likewise.
1536 (vreinterpret_f64_u64): Likewise.
1537 (vreinterpretq_f64_f32): Likewise.
1538 (vreinterpretq_f64_p8): Likewise.
1539 (vreinterpretq_f64_p16): Likewise.
1540 (vreinterpretq_f64_s8): Likewise.
1541 (vreinterpretq_f64_s16): Likewise.
1542 (vreinterpretq_f64_s32): Likewise.
1543 (vreinterpretq_f64_s64): Likewise.
1544 (vreinterpretq_f64_u8): Likewise.
1545 (vreinterpretq_f64_u16): Likewise.
1546 (vreinterpretq_f64_u32): Likewise.
1547 (vreinterpretq_f64_u64): Likewise.
1548 (vreinterpret_s64_f64): Likewise.
1549 (vreinterpretq_s64_f64): Likewise.
1550 (vreinterpret_u64_f64): Likewise.
1551 (vreinterpretq_u64_f64): Likewise.
1552 (vreinterpret_s8_f64): Likewise.
1553 (vreinterpretq_s8_f64): Likewise.
1554 (vreinterpret_s16_f64): Likewise.
1555 (vreinterpretq_s16_f64): Likewise.
1556 (vreinterpret_s32_f64): Likewise.
1557 (vreinterpretq_s32_f64): Likewise.
1558 (vreinterpret_u8_f64): Likewise.
1559 (vreinterpretq_u8_f64): Likewise.
1560 (vreinterpret_u16_f64): Likewise.
1561 (vreinterpretq_u16_f64): Likewise.
1562 (vreinterpret_u32_f64): Likewise.
1563 (vreinterpretq_u32_f64): Likewise.
1565 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1567 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1568 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1569 (vreinterpret_p8_s8): Likewise.
1570 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1571 (vreinterpret_p8_s16): Likewise.
1572 (vreinterpret_p8_s32): Likewise.
1573 (vreinterpret_p8_s64): Likewise.
1574 (vreinterpret_p8_f32): Likewise.
1575 (vreinterpret_p8_u8): Likewise.
1576 (vreinterpret_p8_u16): Likewise.
1577 (vreinterpret_p8_u32): Likewise.
1578 (vreinterpret_p8_u64): Likewise.
1579 (vreinterpret_p8_p16): Likewise.
1580 (vreinterpretq_p8_s8): Likewise.
1581 (vreinterpretq_p8_s16): Likewise.
1582 (vreinterpretq_p8_s32): Likewise.
1583 (vreinterpretq_p8_s64): Likewise.
1584 (vreinterpretq_p8_f32): Likewise.
1585 (vreinterpretq_p8_u8): Likewise.
1586 (vreinterpretq_p8_u16): Likewise.
1587 (vreinterpretq_p8_u32): Likewise.
1588 (vreinterpretq_p8_u64): Likewise.
1589 (vreinterpretq_p8_p16): Likewise.
1590 (vreinterpret_p16_s8): Likewise.
1591 (vreinterpret_p16_s16): Likewise.
1592 (vreinterpret_p16_s32): Likewise.
1593 (vreinterpret_p16_s64): Likewise.
1594 (vreinterpret_p16_f32): Likewise.
1595 (vreinterpret_p16_u8): Likewise.
1596 (vreinterpret_p16_u16): Likewise.
1597 (vreinterpret_p16_u32): Likewise.
1598 (vreinterpret_p16_u64): Likewise.
1599 (vreinterpret_p16_p8): Likewise.
1600 (vreinterpretq_p16_s8): Likewise.
1601 (vreinterpretq_p16_s16): Likewise.
1602 (vreinterpretq_p16_s32): Likewise.
1603 (vreinterpretq_p16_s64): Likewise.
1604 (vreinterpretq_p16_f32): Likewise.
1605 (vreinterpretq_p16_u8): Likewise.
1606 (vreinterpretq_p16_u16): Likewise.
1607 (vreinterpretq_p16_u32): Likewise.
1608 (vreinterpretq_p16_u64): Likewise.
1609 (vreinterpretq_p16_p8): Likewise.
1610 (vreinterpret_f32_s8): Likewise.
1611 (vreinterpret_f32_s16): Likewise.
1612 (vreinterpret_f32_s32): Likewise.
1613 (vreinterpret_f32_s64): Likewise.
1614 (vreinterpret_f32_u8): Likewise.
1615 (vreinterpret_f32_u16): Likewise.
1616 (vreinterpret_f32_u32): Likewise.
1617 (vreinterpret_f32_u64): Likewise.
1618 (vreinterpret_f32_p8): Likewise.
1619 (vreinterpret_f32_p16): Likewise.
1620 (vreinterpretq_f32_s8): Likewise.
1621 (vreinterpretq_f32_s16): Likewise.
1622 (vreinterpretq_f32_s32): Likewise.
1623 (vreinterpretq_f32_s64): Likewise.
1624 (vreinterpretq_f32_u8): Likewise.
1625 (vreinterpretq_f32_u16): Likewise.
1626 (vreinterpretq_f32_u32): Likewise.
1627 (vreinterpretq_f32_u64): Likewise.
1628 (vreinterpretq_f32_p8): Likewise.
1629 (vreinterpretq_f32_p16): Likewise.
1630 (vreinterpret_s64_s8): Likewise.
1631 (vreinterpret_s64_s16): Likewise.
1632 (vreinterpret_s64_s32): Likewise.
1633 (vreinterpret_s64_f32): Likewise.
1634 (vreinterpret_s64_u8): Likewise.
1635 (vreinterpret_s64_u16): Likewise.
1636 (vreinterpret_s64_u32): Likewise.
1637 (vreinterpret_s64_u64): Likewise.
1638 (vreinterpret_s64_p8): Likewise.
1639 (vreinterpret_s64_p16): Likewise.
1640 (vreinterpretq_s64_s8): Likewise.
1641 (vreinterpretq_s64_s16): Likewise.
1642 (vreinterpretq_s64_s32): Likewise.
1643 (vreinterpretq_s64_f32): Likewise.
1644 (vreinterpretq_s64_u8): Likewise.
1645 (vreinterpretq_s64_u16): Likewise.
1646 (vreinterpretq_s64_u32): Likewise.
1647 (vreinterpretq_s64_u64): Likewise.
1648 (vreinterpretq_s64_p8): Likewise.
1649 (vreinterpretq_s64_p16): Likewise.
1650 (vreinterpret_u64_s8): Likewise.
1651 (vreinterpret_u64_s16): Likewise.
1652 (vreinterpret_u64_s32): Likewise.
1653 (vreinterpret_u64_s64): Likewise.
1654 (vreinterpret_u64_f32): Likewise.
1655 (vreinterpret_u64_u8): Likewise.
1656 (vreinterpret_u64_u16): Likewise.
1657 (vreinterpret_u64_u32): Likewise.
1658 (vreinterpret_u64_p8): Likewise.
1659 (vreinterpret_u64_p16): Likewise.
1660 (vreinterpretq_u64_s8): Likewise.
1661 (vreinterpretq_u64_s16): Likewise.
1662 (vreinterpretq_u64_s32): Likewise.
1663 (vreinterpretq_u64_s64): Likewise.
1664 (vreinterpretq_u64_f32): Likewise.
1665 (vreinterpretq_u64_u8): Likewise.
1666 (vreinterpretq_u64_u16): Likewise.
1667 (vreinterpretq_u64_u32): Likewise.
1668 (vreinterpretq_u64_p8): Likewise.
1669 (vreinterpretq_u64_p16): Likewise.
1670 (vreinterpret_s8_s16): Likewise.
1671 (vreinterpret_s8_s32): Likewise.
1672 (vreinterpret_s8_s64): Likewise.
1673 (vreinterpret_s8_f32): Likewise.
1674 (vreinterpret_s8_u8): Likewise.
1675 (vreinterpret_s8_u16): Likewise.
1676 (vreinterpret_s8_u32): Likewise.
1677 (vreinterpret_s8_u64): Likewise.
1678 (vreinterpret_s8_p8): Likewise.
1679 (vreinterpret_s8_p16): Likewise.
1680 (vreinterpretq_s8_s16): Likewise.
1681 (vreinterpretq_s8_s32): Likewise.
1682 (vreinterpretq_s8_s64): Likewise.
1683 (vreinterpretq_s8_f32): Likewise.
1684 (vreinterpretq_s8_u8): Likewise.
1685 (vreinterpretq_s8_u16): Likewise.
1686 (vreinterpretq_s8_u32): Likewise.
1687 (vreinterpretq_s8_u64): Likewise.
1688 (vreinterpretq_s8_p8): Likewise.
1689 (vreinterpretq_s8_p16): Likewise.
1690 (vreinterpret_s16_s8): Likewise.
1691 (vreinterpret_s16_s32): Likewise.
1692 (vreinterpret_s16_s64): Likewise.
1693 (vreinterpret_s16_f32): Likewise.
1694 (vreinterpret_s16_u8): Likewise.
1695 (vreinterpret_s16_u16): Likewise.
1696 (vreinterpret_s16_u32): Likewise.
1697 (vreinterpret_s16_u64): Likewise.
1698 (vreinterpret_s16_p8): Likewise.
1699 (vreinterpret_s16_p16): Likewise.
1700 (vreinterpretq_s16_s8): Likewise.
1701 (vreinterpretq_s16_s32): Likewise.
1702 (vreinterpretq_s16_s64): Likewise.
1703 (vreinterpretq_s16_f32): Likewise.
1704 (vreinterpretq_s16_u8): Likewise.
1705 (vreinterpretq_s16_u16): Likewise.
1706 (vreinterpretq_s16_u32): Likewise.
1707 (vreinterpretq_s16_u64): Likewise.
1708 (vreinterpretq_s16_p8): Likewise.
1709 (vreinterpretq_s16_p16): Likewise.
1710 (vreinterpret_s32_s8): Likewise.
1711 (vreinterpret_s32_s16): Likewise.
1712 (vreinterpret_s32_s64): Likewise.
1713 (vreinterpret_s32_f32): Likewise.
1714 (vreinterpret_s32_u8): Likewise.
1715 (vreinterpret_s32_u16): Likewise.
1716 (vreinterpret_s32_u32): Likewise.
1717 (vreinterpret_s32_u64): Likewise.
1718 (vreinterpret_s32_p8): Likewise.
1719 (vreinterpret_s32_p16): Likewise.
1720 (vreinterpretq_s32_s8): Likewise.
1721 (vreinterpretq_s32_s16): Likewise.
1722 (vreinterpretq_s32_s64): Likewise.
1723 (vreinterpretq_s32_f32): Likewise.
1724 (vreinterpretq_s32_u8): Likewise.
1725 (vreinterpretq_s32_u16): Likewise.
1726 (vreinterpretq_s32_u32): Likewise.
1727 (vreinterpretq_s32_u64): Likewise.
1728 (vreinterpretq_s32_p8): Likewise.
1729 (vreinterpretq_s32_p16): Likewise.
1730 (vreinterpret_u8_s8): Likewise.
1731 (vreinterpret_u8_s16): Likewise.
1732 (vreinterpret_u8_s32): Likewise.
1733 (vreinterpret_u8_s64): Likewise.
1734 (vreinterpret_u8_f32): Likewise.
1735 (vreinterpret_u8_u16): Likewise.
1736 (vreinterpret_u8_u32): Likewise.
1737 (vreinterpret_u8_u64): Likewise.
1738 (vreinterpret_u8_p8): Likewise.
1739 (vreinterpret_u8_p16): Likewise.
1740 (vreinterpretq_u8_s8): Likewise.
1741 (vreinterpretq_u8_s16): Likewise.
1742 (vreinterpretq_u8_s32): Likewise.
1743 (vreinterpretq_u8_s64): Likewise.
1744 (vreinterpretq_u8_f32): Likewise.
1745 (vreinterpretq_u8_u16): Likewise.
1746 (vreinterpretq_u8_u32): Likewise.
1747 (vreinterpretq_u8_u64): Likewise.
1748 (vreinterpretq_u8_p8): Likewise.
1749 (vreinterpretq_u8_p16): Likewise.
1750 (vreinterpret_u16_s8): Likewise.
1751 (vreinterpret_u16_s16): Likewise.
1752 (vreinterpret_u16_s32): Likewise.
1753 (vreinterpret_u16_s64): Likewise.
1754 (vreinterpret_u16_f32): Likewise.
1755 (vreinterpret_u16_u8): Likewise.
1756 (vreinterpret_u16_u32): Likewise.
1757 (vreinterpret_u16_u64): Likewise.
1758 (vreinterpret_u16_p8): Likewise.
1759 (vreinterpret_u16_p16): Likewise.
1760 (vreinterpretq_u16_s8): Likewise.
1761 (vreinterpretq_u16_s16): Likewise.
1762 (vreinterpretq_u16_s32): Likewise.
1763 (vreinterpretq_u16_s64): Likewise.
1764 (vreinterpretq_u16_f32): Likewise.
1765 (vreinterpretq_u16_u8): Likewise.
1766 (vreinterpretq_u16_u32): Likewise.
1767 (vreinterpretq_u16_u64): Likewise.
1768 (vreinterpretq_u16_p8): Likewise.
1769 (vreinterpretq_u16_p16): Likewise.
1770 (vreinterpret_u32_s8): Likewise.
1771 (vreinterpret_u32_s16): Likewise.
1772 (vreinterpret_u32_s32): Likewise.
1773 (vreinterpret_u32_s64): Likewise.
1774 (vreinterpret_u32_f32): Likewise.
1775 (vreinterpret_u32_u8): Likewise.
1776 (vreinterpret_u32_u16): Likewise.
1777 (vreinterpret_u32_u64): Likewise.
1778 (vreinterpret_u32_p8): Likewise.
1779 (vreinterpret_u32_p16): Likewise.
1780 (vreinterpretq_u32_s8): Likewise.
1781 (vreinterpretq_u32_s16): Likewise.
1782 (vreinterpretq_u32_s32): Likewise.
1783 (vreinterpretq_u32_s64): Likewise.
1784 (vreinterpretq_u32_f32): Likewise.
1785 (vreinterpretq_u32_u8): Likewise.
1786 (vreinterpretq_u32_u16): Likewise.
1787 (vreinterpretq_u32_u64): Likewise.
1788 (vreinterpretq_u32_p8): Likewise.
1789 (vreinterpretq_u32_p16): Likewise.
1791 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1793 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1795 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1798 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1799 (vqnegd_s64): Likewise.
1800 (vqabs_s64): Likewise.
1801 (vqabsd_s64): Likewise.
1803 2014-04-22 Richard Henderson <rth@redhat.com>
1805 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1806 computation to the top of the loop.
1808 2014-04-22 Renlin <renlin.li@arm.com>
1809 Jiong Wang <jiong.wang@arm.com>
1811 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1812 * config/aarch64/aarch64.c (aarch64_layout_frame)
1813 (aarch64_initial_elimination_offset): Likewise.
1815 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1817 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1820 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1822 * machmode.h (bitwise_mode_for_mode): Declare.
1823 * stor-layout.h (bitwise_type_for_mode): Likewise.
1824 * stor-layout.c (bitwise_mode_for_mode): New function.
1825 (bitwise_type_for_mode): Likewise.
1826 * builtins.c (fold_builtin_memory_op): Use it instead of
1827 int_mode_for_mode and build_nonstandard_integer_type.
1829 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1831 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1832 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1833 (*-*-solaris2*): Simplify.
1834 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1835 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1836 *-*-solaris2.9* handling.
1838 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1840 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1841 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1843 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1844 * configure: Regenerate.
1846 * config/i386/sol2-9.h: Remove.
1848 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1849 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1850 Remove Solaris 9 references.
1852 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1854 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1855 (floatuns<GPI:mode><GPF:mode>2): Remove.
1856 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1857 and floatuns conversions.
1858 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1859 and floatuns conversions.
1860 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1861 (w1,w2): New mode attributes for inequal width conversions.
1863 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1865 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1866 the output asm format.
1868 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1870 * config/aarch64/aarch64-simd.md
1871 (aarch64_cm<optab>di): Always split.
1872 (*aarch64_cm<optab>di): New.
1873 (aarch64_cmtstdi): Always split.
1874 (*aarch64_cmtstdi): New.
1876 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1878 PR tree-optimization/60823
1879 * omp-low.c (ipa_simd_modify_function_body): Go through
1880 all SSA_NAMEs and for those refering to vector arguments
1881 which are going to be replaced adjust SSA_NAME_VAR and,
1882 if it is a default definition, change it into a non-default
1883 definition assigned at the beginning of function from new_decl.
1884 (ipa_simd_modify_stmt_ops): Rewritten.
1885 * tree-dfa.c (set_ssa_default_def): When removing default def,
1886 check for NULL loc instead of NULL *loc.
1888 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1890 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1891 restrictions on core registers for DImode values in Thumb2.
1893 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1895 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1896 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1898 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1900 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1901 (*iordi_notzesidi_di): Likewise.
1902 (*iordi_notsesidi_di): Likewise.
1904 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1906 * config/arm/arm-protos.h (tune_params): New struct members.
1907 * config/arm/arm.c: Initialise tune_params per processor.
1908 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1909 for speed, based on new tune_params.
1911 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1913 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1915 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1917 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1919 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1920 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1921 (vrnda_f64): Likewise.
1922 (vrndi_f64): Likewise.
1923 (vrndm_f64): Likewise.
1924 (vrndn_f64): Likewise.
1925 (vrndp_f64): Likewise.
1926 (vrndx_f64): Likewise.
1928 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1930 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1931 GET_MODE_SIZE argument is enum machine_mode.
1933 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1936 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1937 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1939 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1942 * asan.c (asan_emit_stack_protection): Force the base to align to
1943 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1944 appropriate bits if STRICT_ALIGNMENT.
1945 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1947 (expand_used_vars): Leave a space in the stack frame for alignment
1948 if STRICT_ALIGNMENT.
1950 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1952 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1954 (gimple_store_p): Likewise.
1955 (gimple_assign_load_p): Likewise.
1956 (gimple_assign_cast_p): Likewise.
1957 (gimple_clobber_p): Likewise.
1959 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1960 rather than a gimple.
1961 (gimple_assign_cast_p): Likewise.
1963 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1966 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1967 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1969 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1970 more debug information for E500 if -mdebug=reg.
1972 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1975 * config/i386/i386.c (ix86_expand_builtin)
1976 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1977 register for target RTX.
1978 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1980 2014-04-18 Cong Hou <congh@google.com>
1982 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1983 the widen-mult pattern by handling two operands with different sizes,
1984 and operands whose size is smaller than half of the result type.
1986 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1988 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1989 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1990 (do_estimate_edge_time): Compute it.
1991 * ipa-inline.c (want_inline_small_function_p): Bypass
1992 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1994 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1996 * ipa-inline.c (spec_rem): New static variable.
1997 (dump_overall_stats): New function.
1998 (dump_inline_stats): New function.
2000 2014-04-18 Richard Henderson <rth@redhat.com>
2002 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
2003 to GET_MODE_SIZE, not a reg_class_t.
2005 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2007 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
2008 (vsx_xxmrglw_<mode>): Likewise.
2010 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2013 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2014 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2015 (rs6000_init_hard_regno_mode_ok): Likewise.
2017 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2019 * ipa-inline.c (inline_small_functions): Account only non-cold
2021 * doc/invoke.texi (inline-unit-growth): Update documentation.
2023 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2025 * config/rs6000/rs6000.md (addti3, subti3): New.
2027 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2030 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2031 comment. Check optimize_insn_for_size_p instead of
2032 optimize_insn_for_speed_p.
2034 2014-04-17 Martin Jambor <mjambor@suse.cz>
2036 * gimple-iterator.c (gsi_start_edge): New function.
2037 * gimple-iterator.h (gsi_start_edge): Declare.
2038 * tree-sra.c (single_non_eh_succ): New function.
2039 (disqualify_ops_if_throwing_stmt): Renamed to
2040 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2041 having one non-EH successor BB.
2042 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2043 generate loads into replacements.
2044 (sra_modify_assign): Likewise and and also use the simple path for
2046 (sra_modify_function_body): Commit statements on edges.
2048 2014-04-17 Richard Biener <rguenther@suse.de>
2051 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2052 comparison results and add clarifying comment.
2054 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2056 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2057 (blank_mode): Initialize it.
2058 (emit_mode_size_inline, emit_mode_nunits_inline,
2059 emit_mode_inner_inline): New functions.
2060 (emit_insn_modes_h): Call them and surround their output with
2061 #if GCC_VERSION >= 4001 ... #endif.
2062 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2063 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2064 mode_* arrays if the argument is __builtin_constant_p.
2065 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2066 is enum machine_mode.
2068 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2070 * passes.c (opt_pass::execute): Adjust.
2071 (pass_manager::execute_pass_mode_switching): Likewise.
2072 (early_local_passes::execute): Likewise.
2073 (execute_one_pass): Pass cfun to the pass's execute method.
2074 * tree-pass.h (opt_pass::execute): Add function * argument.
2075 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2076 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2077 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2078 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2079 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2080 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2081 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2082 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2083 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2084 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
2085 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
2086 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
2087 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
2088 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
2089 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2090 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2091 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2092 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
2093 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
2094 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2095 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2096 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2097 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2098 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2099 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2100 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2101 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2102 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2103 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2106 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2108 * passes.c (opt_pass::gate): Take function * argument.
2109 (gate_all_early_local_passes): Merge into
2110 (early_local_passes::gate): this.
2111 (gate_all_early_optimizations): Merge into
2112 (all_early_optimizations::gate): this.
2113 (gate_all_optimizations): Mege into
2114 (all_optimizations::gate): this.
2115 (gate_all_optimizations_g): Merge into
2116 (all_optimizations_g::gate): this.
2117 (gate_rest_of_compilation): Mege into
2118 (rest_of_compilation::gate): this.
2119 (gate_postreload): Merge into
2120 (postreload::gate): this.
2121 (dump_one_pass): Pass cfun to the pass's gate method.
2122 (execute_ipa_summary_passes): Likewise.
2123 (execute_one_pass): Likewise.
2124 (ipa_write_summaries_2): Likewise.
2125 (ipa_write_optimization_summaries_1): Likewise.
2126 (ipa_read_summaries_1): Likewise.
2127 (ipa_read_optimization_summaries_1): Likewise.
2128 (execute_ipa_stmt_fixups): Likewise.
2129 * tree-pass.h (opt_pass::gate): Add function * argument.
2130 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2131 combine-stack-adj.c, combine.c, compare-elim.c,
2132 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2133 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2134 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2135 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2136 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2137 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2138 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2139 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2140 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2141 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2142 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2143 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2144 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2145 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2146 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2147 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2148 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2149 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2150 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2151 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2152 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2153 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2154 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2155 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2156 var-tracking.c, vtable-verify.c, web.c: Adjust.
2158 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2160 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2161 * configure: Regenerate.
2163 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2165 * passes.c (dump_one_pass): don't check pass->has_gate.
2166 (execute_ipa_summary_passes): Likewise.
2167 (execute_one_pass): Likewise.
2168 (ipa_write_summaries_2): Likewise.
2169 (ipa_write_optimization_summaries_1): Likewise.
2170 (ipa_read_optimization_summaries_1): Likewise.
2171 (execute_ipa_stmt_fixups): Likewise.
2172 * tree-pass.h (pass_data::has_gate): Remove.
2173 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2174 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2175 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2176 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2177 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2178 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2179 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2180 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2181 gimple-low.c, gimple-ssa-isolate-paths.c,
2182 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2183 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2184 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2185 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2186 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2187 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2188 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2189 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2190 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2191 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2192 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2193 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2194 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2195 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2196 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2197 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2198 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2199 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2200 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2201 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2202 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2203 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2206 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2208 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2210 * passes.c (pass_manager::register_dump_files_1): Merge into
2211 (pass_manager::register_dump_files): this, and remove its handling of
2212 properties since the pass always has the properties anyway.
2213 (pass_manager::pass_manager): Adjust.
2215 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2217 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2218 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2219 dealing with properties.
2220 (pass_manager::register_dump_files): Adjust.
2222 2014-03-20 Mark Wielaard <mjw@redhat.com>
2224 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2225 then represent the bound as normal constant value.
2227 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2230 Forward port from 4.8 branch
2231 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2233 * config/i386/bmiintrin.h (_blsi_u32): New.
2237 (_blsmsk_u32): Ditto.
2238 (_blsmsk_u64): Ditto.
2239 (_tzcnt_u32): Ditto.
2240 (_tzcnt_u64): Ditto.
2242 2014-04-17 Kito Cheng <kito@0xlab.org>
2244 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2246 2014-04-17 Richard Biener <rguenther@suse.de>
2249 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2250 boolean results for comparisons.
2252 2014-04-17 Richard Biener <rguenther@suse.de>
2254 PR tree-optimization/60836
2255 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2256 initial PHI args to be gimple values.
2258 2014-04-17 Richard Biener <rguenther@suse.de>
2260 PR tree-optimization/60841
2261 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2262 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2263 of stmts to SLP build.
2264 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2265 (vect_analyze_slp): Likewise.
2266 (vect_analyze_slp_instance): Likewise.
2267 (vect_build_slp_tree): Limit overall SLP tree growth.
2268 * tree-vectorizer.h (vect_analyze_data_refs,
2269 vect_analyze_slp): Adjust prototypes.
2271 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2273 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2276 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2278 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2279 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2280 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2281 for TARGET_SLOW_PSHUFB
2283 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2285 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2286 * config/i386/i386.c (intel_cost): Ditto.
2288 2014-04-17 Joey Ye <joey.ye@arm.com>
2290 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2292 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2294 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2297 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2299 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2300 (type_all_derivations_known_p): New predicate.
2301 (type_all_ctors_visible_p): New predicate.
2302 (type_possibly_instantiated_p): New predicate.
2303 (get_odr_type): Compute all_derivations_known.
2304 (dump_odr_type): Dump the flag.
2305 (maybe_record_type): Cleanup.
2306 (record_target_from_binfo): Add bases_to_consider array;
2307 record bases for types w/o instances and skip CXX destructor.
2308 (possible_polymorphic_call_targets_1): Add bases_to_consider
2309 and consider_construction parameters; check if type may have instance.
2310 (get_polymorphic_call_info): Set maybe_in_construction to true
2311 when we know nothing.
2312 (record_targets_from_bases): Skip CXX destructors; they are
2313 never called for types in construction.
2314 (possible_polymorphic_call_targets): Do not record target when
2315 type may not have instance.
2317 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2320 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2321 external aliases alive, too.
2323 2014-04-16 Andrew Pinski <apinski@cavium.com>
2325 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2328 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2330 * final.c (compute_alignments): Do not apply loop alignment to a block
2331 falling through to the exit.
2333 2014-04-16 Catherine Moore <clm@codesourcery.com>
2335 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2336 Adjust constraints for microMIPS store patterns.
2338 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2340 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2342 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2344 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2345 (append_use): Run at -O0.
2346 (append_vdef): Likewise.
2347 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2348 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2350 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2352 PR tree-optimization/60844
2353 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2354 (propagate_op_to_single_use, remove_visited_stmt_chain,
2355 linearize_expr, repropagate_negates, reassociate_bb): Use it
2356 instead of gsi_remove.
2358 2014-04-16 Martin Jambor <mjambor@suse.cz>
2360 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2361 ipa_transforms_to_apply.
2362 (cgraph_function_versioning): Assert that old_node has empty
2363 ipa_transforms_to_apply.
2364 * trans-mem.c (ipa_tm_create_version): Likewise.
2365 * tree-inline.c (tree_function_versioning): Do not duplicate
2366 ipa_transforms_to_apply.
2368 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2371 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2373 Pass necessary as flags on 64-bit Solaris/x86.
2374 Use lowercase relocs for x86_64-*-*.
2375 * configure: Regenerate.
2377 2014-04-15 Jan Hubicka <jh@suse.cz>
2379 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2380 (maybe_record_node, likely_target_p): Use it.
2382 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2385 Revert following patch
2387 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2390 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2391 software floating point or no floating point registers, do not
2392 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2393 in GPRs that occurs after we tested for GPRs that would never be
2396 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2397 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2398 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2399 specifically allow DDmode, since that does not use the SPE SIMD
2402 2014-03-21 Mark Wielaard <mjw@redhat.com>
2404 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2405 as unsigned or int depending on type and value used.
2407 2014-04-15 Richard Biener <rguenther@suse.de>
2409 PR rtl-optimization/56965
2410 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2411 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2413 * alias.c (true_dependence_1): Do not call
2414 nonoverlapping_component_refs_p.
2415 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2416 nonoverlapping_component_refs_p.
2417 (indirect_refs_may_alias_p): Likewise.
2419 2014-04-15 Teresa Johnson <tejohnson@google.com>
2421 * cfg.c (dump_bb_info): Fix flags check.
2422 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2424 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2426 PR rtl-optimization/60663
2427 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2430 2014-04-15 Richard Biener <rguenther@suse.de>
2432 * lto-streamer.h (LTO_major_version): Bump to 4.
2434 2014-04-15 Richard Biener <rguenther@suse.de>
2436 * common.opt (lto_partition_model): New enum.
2437 (flto-partition=): Merge separate options with a single with argument,
2438 add -flto-partition=one support.
2439 * flag-types.h (enum lto_partition_model): Declare.
2440 * opts.c (finish_options): Remove duplicate -flto-partition=
2442 * lto-wrapper.c (run_gcc): Adjust.
2444 2014-04-15 Richard Biener <rguenther@suse.de>
2446 * alias.c (ncr_compar): New function.
2447 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2449 2014-04-15 Richard Biener <rguenther@suse.de>
2451 * alias.c (record_component_aliases): Do not walk BINFOs.
2453 2014-04-15 Richard Biener <rguenther@suse.de>
2455 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2456 Add struct function argument and adjust.
2457 (find_func_aliases_for_call): Likewise.
2458 (find_func_aliases): Likewise.
2459 (find_func_clobbers): Likewise.
2460 (intra_create_variable_infos): Likewise.
2461 (compute_points_to_sets): Likewise.
2462 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2464 2014-04-15 Richard Biener <rguenther@suse.de>
2466 * tree.c (iterative_hash_expr): Use enum tree_code_class
2467 to store TREE_CODE_CLASS.
2468 (tree_block): Likewise.
2469 (tree_set_block): Likewise.
2470 * tree.h (fold_build_pointer_plus_loc): Use
2471 convert_to_ptrofftype_loc.
2473 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2476 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2479 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2481 * cfgloop.h (struct loop): Move force_vectorize down.
2482 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2483 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2484 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2485 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2486 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2487 * tree-core.h (enum annot_expr_kind): Add new kind values.
2488 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2489 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2491 * tree.def (ANNOTATE_EXPR): Tweak comment.
2493 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2495 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2498 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2500 * tree.h (TYPE_IDENTIFIER): Declare.
2501 * tree.c (subrange_type_for_debug_p): Use it.
2502 * godump.c (go_format_type): Likewise.
2503 * dwarf2out.c (is_cxx_auto, modified_type_die,
2504 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2505 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2507 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2510 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2512 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2514 * config/i386/i386.c (examine_argument): Return bool. Return true if
2515 parameter should be passed in memory.
2516 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2517 (construct_container): Update calls to examine_argument.
2518 (function_arg_advance_64): Ditto.
2519 (return_in_memory_32): Merge with ix86_return_in_memory.
2520 (return_in_memory_64): Ditto.
2521 (return_in_memory_ms_64): Ditto.
2523 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2525 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2526 * coverage.c (coverage_compute_profile_id): Handle externally visible
2529 2014-04-14 Martin Jambor <mjambor@suse.cz>
2531 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2532 DECL_DISREGARD_INLINE_LIMITS functions.
2534 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2537 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2539 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2542 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2543 optimize_insn_for_speed_p instead of
2544 optimize_function_for_speed_p.
2546 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2548 * doc/invoke.texi (free): Document AArch64.
2550 2014-04-14 Richard Biener <rguenther@suse.de>
2552 PR tree-optimization/60042
2553 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2554 (insert_into_preds_of_block): Do not prevent PHI insertion
2555 for REFERENCE exprs here ...
2556 (eliminate_dom_walker::before_dom_children): ... but prevent
2557 their use here under similar conditions when applied to the
2558 IL after PRE optimizations.
2560 2014-04-14 Richard Biener <rguenther@suse.de>
2562 * passes.def: Move early points-to after early SRA.
2564 2014-04-14 Richard Biener <rguenther@suse.de>
2566 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2567 check for which sign-changes we allow when forwarding
2568 a converted value into a switch.
2570 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2572 * stor-layout.c (place_field): Finalize non-constant offset for the
2575 2014-04-14 Richard Biener <rguenther@suse.de>
2577 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2579 (expand_switch_using_bit_tests_p): Likewise.
2580 (process_switch): Compute and pass on speed_p based on the
2582 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2583 optimize_bb_for_speed_p.
2585 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2587 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2588 * function.h (struct function): Rename has_force_vect_loops into
2589 has_force_vectorize_loops.
2590 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2591 (input_struct_function_base): Likewise.
2592 * lto-streamer-out.c (output_cfg): Likewise.
2593 (output_struct_function_base): Likewise.
2594 * omp-low.c (expand_omp_simd): Likewise.
2595 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2596 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2597 (version_loop_for_if_conversion): Likewise.
2598 (tree_if_conversion): Likewise.
2599 (main_tree_if_conversion): Likewise.
2600 (gate_tree_if_conversion): Likewise.
2601 * tree-inline.c (copy_loops): Likewise.
2602 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2603 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2604 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2605 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2606 * tree-vectorizer.c (vectorize_loops): Likewise.
2607 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2609 2014-04-14 Richard Biener <rguenther@suse.de>
2612 * lto-streamer-out.c (wrap_refs): New function.
2613 (lto_output): Wrap symbol references in global initializes in
2614 type-preserving MEM_REFs.
2616 2014-04-14 Christian Bruel <christian.bruel@st.com>
2618 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2620 2014-04-14 Christian Bruel <christian.bruel@st.com>
2622 * config/sh/sh.md (setmemqi): New expand pattern.
2623 * config/sh/sh.h (CLEAR_RATIO): Define.
2624 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2625 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2627 2014-04-14 Richard Biener <rguenther@suse.de>
2630 * fold-const.c (negate_expr_p): Don't negate directional rounding
2632 (fold_negate_expr): Likewise.
2634 2014-04-14 Richard Biener <rguenther@suse.de>
2636 PR tree-optimization/59817
2637 PR tree-optimization/60453
2638 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2639 recursion to catch all CHRECs in the scalar evolution and restrict
2640 the predicate for the remains appropriately.
2642 2014-04-12 Catherine Moore <clm@codesourcery.com>
2644 * config/mips/constraints.md: Add new register constraint "kb".
2645 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2646 (*movhi_internal): Likewise.
2647 (*movqi_internal): Likewise.
2648 * config/mips/mips.h (M16_STORE_REGS): New register class.
2649 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2650 (REG_CLASS_CONTENTS): Likewise.
2651 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2653 2014-04-11 Tobias Burnus <burnus@net-b.de>
2656 * doc/invoke.texi (-Wformat-signedness): Document it.
2657 (Wformat=2): Mention that this enables -Wformat-signedness.
2659 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2661 * common/config/epiphany/epiphany-common.c
2662 (epiphany_option_optimization_table): Enable section anchors by
2663 default at -O1 or higher.
2664 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2665 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2666 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2667 carries no extra cost.
2668 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2669 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2670 * config/epiphany/predicates.md (memclob_operand): New predicate.
2671 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2672 Use memclob_operand predicate and X constraint for operand 3.
2674 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2676 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2677 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2680 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2682 PR rtl-optimization/60651
2683 * mode-switching.c (optimize_mode_switching): Make sure to emit
2684 sets of a lower numbered entity before sets of a higher numbered
2685 entity to a mode of the same or lower priority.
2686 When creating a seginfo for a basic block that starts with a code
2687 label, move the insertion point past the code label.
2688 (new_seginfo): Document and enforce requirement that
2689 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2690 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2691 * doc/tm.texi: Regenerate.
2693 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2696 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2698 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2700 * BASE-VER: Set to 4.10.0.
2702 2014-04-11 Tobias Burnus <burnus@net-b.de>
2705 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2706 * doc/gcc.texi (Service): Update description in the @menu
2707 * doc/invoke.texi (Option Summary): Remove misplaced and
2710 2014-04-11 Steve Ellcey <sellcey@mips.com>
2711 Jakub Jelinek <jakub@redhat.com>
2714 * expr.c (convert_move): Use emit_store_flag_force instead of
2715 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2718 2014-04-11 Richard Biener <rguenther@suse.de>
2721 * varasm.c (assemble_alias): Avoid endless error reporting
2722 recursion by setting TREE_ASM_WRITTEN.
2724 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2726 * config/s390/s390.md: Add a splitter for NOT rtx.
2728 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2730 PR rtl-optimization/60663
2731 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2733 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2734 Jakub Jelinek <jakub@redhat.com>
2737 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2738 flag from decl_node to node.
2740 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2743 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2744 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2745 ameliorating the cases where it can be.
2747 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2750 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2752 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2753 (loadsync_<mode>): Change mode.
2754 (load_quadpti, store_quadpti): New.
2755 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2756 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2757 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2759 2014-04-09 Cong Hou <congh@google.com>
2762 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2765 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2767 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2768 instead of vnor to exploit possible fusion opportunity in the
2770 (altivec_expand_vec_perm_const_le): Likewise.
2772 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2774 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2775 (loadsync_<mode>): Change mode.
2776 (load_quadpti, store_quadpti): New.
2777 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2778 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2780 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2783 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2784 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2785 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2787 2014-04-08 Richard Biener <rguenther@suse.de>
2790 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2791 a 64bit widest int print double-int similar to on HWI64 hosts.
2793 2014-04-08 Richard Biener <rguenther@suse.de>
2795 PR tree-optimization/60785
2796 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2797 default defs properly.
2799 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2801 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2802 (Weffc++): Likewise.
2804 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2806 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2807 set completep to false rather than true.
2809 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2812 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2813 ARM_TARGET2_DWARF_FORMAT.
2815 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2818 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2819 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2822 2014-04-07 Richard Biener <rguenther@suse.de>
2824 PR tree-optimization/60766
2825 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2826 (may_eliminate_iv): Convert cand_value_at result to desired type.
2828 2014-04-07 Jason Merrill <jason@redhat.com>
2831 * common.opt (-fno-gnu-unique): Add.
2832 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2834 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2836 * haifa-sched.c: Fix outdated function reference and minor
2837 grammar errors in introductory comment.
2839 2014-04-07 Richard Biener <rguenther@suse.de>
2842 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2844 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2846 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2849 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2850 size accounting for thunks.
2851 (pa_asm_output_mi_thunk): Use final_start_function() and
2852 final_end_function() to output function start and end directives.
2854 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2856 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2857 device specific ISA/ feature information. Remove short_sp and
2858 errata_skip ds. Add avr_device_specific_features enum to have device
2860 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2861 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2862 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2863 updated device specific info.
2864 * config/avr/avr-mcus.def: Merge device specific details to
2865 dev_attribute field.
2866 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2868 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2869 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2870 assembler if RMW isa supported by current device.
2871 * config/avr/genmultilib.awk: Update as device info structure changed.
2872 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2874 2014-04-04 Cong Hou <congh@google.com>
2876 PR tree-optimization/60656
2877 * tree-vect-stmts.c (supportable_widening_operation):
2878 Fix a bug that elements in a vector with vect_used_by_reduction
2879 property are incorrectly reordered when the operation on it is not
2880 consistant with the one in reduction operation.
2882 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2884 PR rtl-optimization/60155
2885 * gcse.c (record_set_data): New function.
2886 (single_set_gcse): New function.
2887 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2888 (hoist_code): Likewise.
2889 (get_pressure_class_and_nregs): Likewise.
2891 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2893 * explow.c (probe_stack_range): Emit a final optimization blockage.
2895 2014-04-04 Anthony Green <green@moxielogic.com>
2897 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2900 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2903 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2904 flags are set only during streaming.
2905 * ipa.c (process_references, walk_polymorphic_call_targets,
2906 symtab_remove_unreachable_nodes): Drop bodies of always inline
2907 after early inlining.
2908 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2910 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2911 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2914 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2917 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2920 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2922 (cortex_a53_fdivd): Likewise.
2924 2014-04-04 Martin Jambor <mjambor@suse.cz>
2927 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2929 * cgraph.c (clone_of_p): Also return true if thunks match.
2930 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2931 cgraph_function_or_thunk_node and an obsolete comment.
2932 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2934 (build_function_decl_skip_args): Likewise.
2935 (set_new_clone_decl_and_node_flags): New function.
2936 (duplicate_thunk_for_node): Likewise.
2937 (redirect_edge_duplicating_thunks): Likewise.
2938 (cgraph_clone_node): New parameter args_to_skip, pass it to
2939 redirect_edge_duplicating_thunks which is called instead of
2940 cgraph_redirect_edge_callee.
2941 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2942 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2944 2014-04-04 Jeff Law <law@redhat.com>
2947 * config/arm/predicates.md (const_int_I_operand): New predicate.
2948 (const_int_M_operand): Similarly.
2949 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2951 (insv_t2, extv_reg, extzv_t2): Likewise.
2952 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2953 (pop_multiple_with_writeback_and_return): Likewise.
2954 (vfp_pop_multiple_with_writeback): Likewise
2956 2014-04-04 Richard Biener <rguenther@suse.de>
2959 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2960 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2962 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2963 * gimplify.c (gimple_add_tmp_var_fn): New function.
2964 * gimple-expr.h (create_tmp_reg_fn): Declare.
2965 * gimple-expr.c (create_tmp_reg_fn): New function.
2966 * gimple-low.c (record_vars_into): Don't change cfun.
2967 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2968 code generation without cfun.
2970 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2973 * Makefile.in (install-driver): Fix shell scripting.
2975 2014-04-03 Cong Hou <congh@google.com>
2977 PR tree-optimization/60505
2978 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2979 threshold of number of iterations below which no vectorization
2981 * tree-vect-loop.c (new_loop_vec_info):
2982 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2983 * tree-vect-loop.c (vect_analyze_loop_operations):
2984 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2985 * tree-vect-loop.c (vect_transform_loop):
2986 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2987 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2988 of iterations of the loop and see if we should build the epilogue.
2990 2014-04-03 Richard Biener <rguenther@suse.de>
2992 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2993 (streamer_tree_cache_create): Adjust.
2994 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2995 to allow optional nodes array.
2996 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2997 (streamer_tree_cache_append): Likewise.
2998 (streamer_tree_cache_create): Create nodes array optionally
2999 as specified by parameter.
3000 * lto-streamer-out.c (create_output_block): Avoid maintaining
3001 the node array in the writer cache.
3002 (DFS_write_tree): Remove assertion.
3003 (produce_asm_for_decls): Free the out decl state hash table early.
3004 * lto-streamer-in.c (lto_data_in_create): Adjust for
3005 streamer_tree_cache_create prototype change.
3007 2014-04-03 Richard Biener <rguenther@suse.de>
3009 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
3010 set TREE_CHAIN to NULL_TREE.
3012 2014-04-03 Richard Biener <rguenther@suse.de>
3014 PR tree-optimization/60740
3015 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3016 over all GIMPLE_COND operands.
3018 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3020 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3021 (Weffc++): Remove Scott's numbering, merge lists and reference
3024 2014-04-03 Nick Clifton <nickc@redhat.com>
3026 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3029 2014-04-03 Martin Jambor <mjambor@suse.cz>
3031 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3032 mention gcc_unreachable before failing.
3033 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3036 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3039 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3040 inconsistent code and instead mark the context inconsistent.
3041 (possible_polymorphic_call_targets): For inconsistent contexts
3042 return empty complete list.
3044 2014-04-02 Anthony Green <green@moxielogic.com>
3046 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3047 (extendqisi2, extendhisi2): Define.
3048 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3049 (WCHAR_TYPE): Change to unsigned int.
3051 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3053 PR tree-optimization/60733
3054 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3055 insertion point for PHI candidates to be the end of the feeding
3056 block for the PHI argument.
3058 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3060 PR rtl-optimization/60650
3061 * lra-constraints.c (process_alt_operands): Decrease reject for
3062 earlyclobber matching.
3064 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3066 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3068 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3070 * config/spu/spu.c (pad_bb): Do not crash when the last
3071 insn is CODE_FOR_blockage.
3073 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3075 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3076 lies outside the target mode.
3078 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3081 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3082 software floating point or no floating point registers, do not
3083 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3084 in GPRs that occurs after we tested for GPRs that would never be
3087 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3088 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3089 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3090 specifically allow DDmode, since that does not use the SPE SIMD
3093 2014-04-02 Richard Biener <rguenther@suse.de>
3096 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
3097 MODE_INTs. Properly use negv_optab.
3098 (expand_abs): Likewise.
3100 2014-04-02 Richard Biener <rguenther@suse.de>
3103 * Makefile.in (install-driver): Guard extra installs with special
3106 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3108 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3111 2014-04-01 Richard Henderson <rth@redhat.com>
3114 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3115 alternative enabled before register allocation.
3117 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3119 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3120 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3122 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3123 (nios2_got_address): Update nios2_large_got_address call site.
3124 (nios2_delegitimize_address): New function.
3125 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3126 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3127 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3129 2014-04-01 Martin Husemann <martin@duskware.de>
3131 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3134 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3136 PR rtl-optimization/60604
3137 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3138 check from register_operand.
3139 (register_operand): Redefine in terms of general_operand.
3140 (nonmemory_operand): Use register_operand for the non-constant cases.
3142 2014-04-01 Richard Biener <rguenther@suse.de>
3144 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3146 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3148 * doc/invoke.texi (mapp-regs): Clarify.
3150 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3152 * config/i386/avx512fintrin.h (__v32hi): Define type.
3153 (__v64qi): Likewise.
3154 (_mm512_set1_epi8): Define.
3155 (_mm512_set1_epi16): Define.
3156 (_mm512_set4_epi32): Define.
3157 (_mm512_set4_epi64): Define.
3158 (_mm512_set4_pd): Define.
3159 (_mm512_set4_ps): Define.
3160 (_mm512_setr4_epi64): Define.
3161 (_mm512_setr4_epi32): Define.
3162 (_mm512_setr4_pd): Define.
3163 (_mm512_setr4_ps): Define.
3164 (_mm512_setzero_epi32): Define.
3166 2014-03-31 Martin Jambor <mjambor@suse.cz>
3169 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3170 callsite_arguments_match_p. Updated all callers. Also check types of
3171 corresponding formal parameters and actual arguments.
3172 (not_all_callers_have_enough_arguments_p) Renamed to
3173 some_callers_have_mismatched_arguments_p.
3175 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3177 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3179 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3182 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3185 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3187 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3188 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3190 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3191 Use FMAMODE_NOVF512 mode iterator.
3192 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3193 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3194 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3195 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3197 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3198 Use VF_128_256 mode iterator.
3199 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3202 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3204 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3205 static chain if needed.
3207 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3210 * lra-constraints.c (index_part_to_reg): New.
3211 (process_address): Use it.
3213 2014-03-27 Jeff Law <law@redhat.com>
3214 Jakub Jelinek <jakub@redhat.com>
3217 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3218 gen_rtx_{PLUS,MULT} to build up the address expression.
3220 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3221 creating non-canonical RTL.
3223 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3226 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3227 functions; reorganize to make cheap checks first.
3228 (inline_small_functions): Do not estimate growth when dumping;
3230 * ipa-inline.h (inline_summary): Add min_size.
3231 (growth_likely_positive): New function.
3232 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3233 (set_cond_stmt_execution_predicate): Cleanup.
3234 (estimate_edge_size_and_time): Compute min_size.
3235 (estimate_calls_size_and_time): Likewise.
3236 (estimate_node_size_and_time): Likewise.
3237 (inline_update_overall_summary): Update min_size.
3238 (do_estimate_edge_time): Likewise.
3239 (do_estimate_edge_size): Update.
3240 (do_estimate_edge_hints): Update.
3241 (growth_likely_positive): New function.
3243 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3246 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3247 also if addr has VOIDmode.
3249 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3251 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3252 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3254 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3255 instructions as well as AdvancedSIMD loads.
3257 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3259 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3260 Use crypto_aese type.
3261 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3262 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3263 crypto_aese, crypto_aesmc. Move to types.md.
3264 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3266 * config/arm/iterators.md (crypto_type): Likewise.
3268 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3270 * cgraph.c: Include expr.h and tree-dfa.h.
3271 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3274 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3277 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3278 regs from checking multi-reg pseudos.
3280 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3282 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3284 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3286 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3287 if it would clobber the stack pointer, even temporarily.
3289 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3291 * mode-switching.c: Make small adjustments to the top comment.
3293 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3295 * config/rs6000/constraints.md (wD constraint): New constraint to
3296 match the constant integer to get the top DImode/DFmode out of a
3297 vector in a VSX register.
3299 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3300 match the constant integer to get the top DImode/DFmode out of a
3301 vector in a VSX register.
3303 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3306 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3309 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3310 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3312 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3313 Optimize vec_extract of 64-bit values, where the value being
3314 extracted is in the top word, where we can use scalar
3315 instructions. Add direct move and store support. Combine the big
3316 endian/little endian vector select load support into a single insn.
3317 (vsx_extract_<mode>_internal1): Likewise.
3318 (vsx_extract_<mode>_internal2): Likewise.
3319 (vsx_extract_<mode>_load): Likewise.
3320 (vsx_extract_<mode>_store): Likewise.
3321 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3322 combined into vsx_extract_<mode>_load.
3323 (vsx_extract_<mode>_one_le): Likewise.
3325 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3326 define the top 64-bit vector element.
3328 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3331 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3332 Document vec_vbpermq builtin.
3335 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3336 enable use of xxsldwi and xxpermdi builtin functions.
3337 (vec_xxpermdi): Likewise.
3339 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3340 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3342 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3344 PR rtl-optimization/60650
3345 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3347 (find_spills_for): New.
3348 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3349 Spill all pseudos on the second iteration.
3351 2014-03-27 Marek Polacek <polacek@redhat.com>
3354 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3357 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3359 * config/s390/s390.c (s390_can_use_return_insn): Check for
3360 call-saved FPRs on 31 bit.
3362 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3365 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3366 if they need regimplification, just drop them instead of
3367 calling gimple_regimplify_operands on them.
3369 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3372 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3373 (aarch64_frame_pointer_required): Adjust logic.
3374 (aarch64_can_eliminate): Adjust logic.
3375 (aarch64_override_options_after_change): Adjust logic.
3377 2014-03-27 Dehao Chen <dehao@google.com>
3379 * ipa-inline.c (early_inliner): Update node's inline info.
3381 2014-03-26 Dehao Chen <dehao@google.com>
3383 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3384 compiler inserted conditional jumps for NAN float check.
3386 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3388 * ubsan.h (ubsan_create_data): Change second argument's type
3389 to const location_t *.
3390 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3392 (ubsan_create_data): Change second argument to const location_t *PLOC.
3393 Create Loc field whenever PLOC is non-NULL.
3394 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3395 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3399 * real.c (real_to_integer2): Change type of low to UHWI.
3401 2014-03-26 Tobias Burnus <burnus@net-b.de>
3403 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3404 (CILK_SELF_SPECS): New define.
3405 (driver_self_specs): Use it.
3407 2014-03-26 Richard Biener <rguenther@suse.de>
3409 * tree-pretty-print.c (percent_K_format): Implement special
3410 case for LTO and its stripped down BLOCK tree.
3412 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3415 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3417 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3418 one range is range_int_cst_p, but not both, at least optimize
3419 addition/subtraction of 0 and multiplication by 0 or 1.
3420 * gimple-fold.c (gimple_fold_call): Fold
3421 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3422 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3423 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3425 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3427 PR rtl-optimization/60452
3428 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3429 <case REG>: Return 1 for invalid offsets from the frame pointer.
3431 2014-03-26 Marek Polacek <polacek@redhat.com>
3434 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3437 2014-03-26 Marek Polacek <polacek@redhat.com>
3440 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3443 2014-03-26 Marek Polacek <polacek@redhat.com>
3446 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3447 multiplication in unsigned type.
3449 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3451 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3453 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3455 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3457 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3460 * cif-code.def (UNREACHABLE) New code.
3461 * ipa-inline.c (inline_small_functions): Skip edges to
3462 __builtlin_unreachable.
3463 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3464 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3465 predicate to __bulitin_unreachable.
3466 (set_cond_stmt_execution_predicate): Fix issue when
3467 invert_tree_comparison returns ERROR_MARK.
3468 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3469 propagate to inline clones.
3470 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3472 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3473 * cgraphclones.c (cgraph_clone_node): If call destination is already
3474 ureachable, do not redirect it back.
3475 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3478 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3480 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3481 Do not modify inline clones.
3483 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3485 * config/i386/i386.md (general_sext_operand): New mode attr.
3486 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3487 don't generate (sign_extend (const_int)).
3488 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3489 operands[2]. Use We constraint instead of <i> and
3490 <general_sext_operand> predicate instead of <general_operand>.
3491 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3492 * config/i386/constraints.md (We): New constraint.
3493 * config/i386/predicates.md (x86_64_sext_operand,
3494 sext_operand): New predicates.
3496 2014-03-25 Martin Jambor <mjambor@suse.cz>
3499 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3500 inconsistent devirtualizations to __builtin_unreachable.
3502 2014-03-25 Marek Polacek <polacek@redhat.com>
3505 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3507 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3509 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3510 order of elements for big-endian.
3512 2014-03-25 Richard Biener <rguenther@suse.de>
3515 * gimplify-me.c (gimple_regimplify_operands): Update the
3518 2014-03-25 Martin Jambor <mjambor@suse.cz>
3521 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3522 (lto_output_varpool_node): Likewise.
3523 (input_overwrite_node): Likewise.
3524 (input_varpool_node): Likewise.
3526 2014-03-25 Richard Biener <rguenther@suse.de>
3528 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3529 (run_gcc): Likewise.
3531 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3533 * combine.c (simplify_compare_const): Add MODE argument.
3534 Handle mode_width 0 as very large mode_width.
3535 (try_combine, simplify_comparison): Adjust callers.
3537 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3538 type to avoid signed integer overflow.
3539 * explow.c (plus_constant): Likewise.
3541 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3543 * doc/generic.texi: Correct typos.
3545 2014-03-24 Tobias Burnus <burnus@net-b.de>
3547 * doc/invoke.texi (-flto): Expand section about
3548 using static libraries with LTO.
3550 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3552 PR rtl-optimization/60501
3553 * optabs.def (addptr3_optab): New optab.
3554 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3555 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3556 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3558 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3560 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3562 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3564 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3567 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3568 (_mm256_undefined_ps): Define.
3569 (_mm256_undefined_pd): Define.
3570 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3571 (_mm_undefined_pd): Define.
3572 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3573 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3574 (_mm512_undefined_ps): Define.
3575 (_mm512_undefined_pd): Define.
3576 Use _mm*_undefined_*.
3577 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3579 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3581 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3582 (lshr_simd): DI mode added.
3583 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3584 (aarch64_ushr_simddi): Likewise.
3585 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3586 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3587 (vshrd_n_u64): Likewise.
3589 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3591 * Makefile.in (s-macro_list): Depend on cc1.
3593 2014-03-23 Teresa Johnson <tejohnson@google.com>
3595 * ipa-utils.c (ipa_print_order): Use specified dump file.
3597 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3599 PR rtl-optimization/60601
3600 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3602 * gcc.c (eval_spec_function): Initialize save_growing_value.
3604 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3607 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3608 code == MINUS_EXPR, never swap op0 with op1.
3610 * toplev.c (init_local_tick): Avoid signed integer multiplication
3612 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3613 shift by first operand's bitsize.
3615 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3618 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3620 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3621 TARGET_ISA_64BIT_P(x).
3623 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3625 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3626 pattern for vector nor instead of subtract from splat(-1).
3627 (altivec_expand_vec_perm_const_le): Likewise.
3629 2014-03-21 Richard Henderson <rth@twiddle.net>
3632 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3633 related insns after epilogue_completed.
3635 2014-03-21 Martin Jambor <mjambor@suse.cz>
3638 * cgraph.h (symtab_node): New flag body_removed.
3639 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3640 when removing bodies.
3641 * symtab.c (dump_symtab_base): Dump body_removed flag.
3642 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3643 had their bodies removed.
3645 2014-03-21 Martin Jambor <mjambor@suse.cz>
3648 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3651 2014-03-21 Richard Biener <rguenther@suse.de>
3653 PR tree-optimization/60577
3654 * tree-core.h (struct tree_base): Document nothrow_flag use
3655 in VAR_DECL_NONALIASED.
3656 * tree.h (VAR_DECL_NONALIASED): New.
3657 (may_be_aliased): Adjust.
3658 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3660 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3662 * expr.c (expand_expr_real_1): Remove outdated comment.
3664 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3667 * ira.c (adjust_cleared_regs): Call copy_rtx on
3668 *reg_equiv[REGNO (loc)].src_p before passing it to
3669 simplify_replace_fn_rtx.
3672 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3673 into CONST, put pic register as first operand of PLUS. Use
3674 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3676 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3678 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3680 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3682 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3683 around for store forwarding issue in the FPU on the UT699.
3684 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3685 loads and operations if -mfix-ut699 is specified.
3686 (divtf3_hq): Tweak attribute.
3687 (sqrttf2_hq): Likewise.
3689 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3691 * calls.c (store_one_arg): Remove incorrect const qualification on the
3692 type of the temporary.
3693 * cfgexpand.c (expand_return): Likewise.
3694 * expr.c (expand_constructor): Likewise.
3695 (expand_expr_real_1): Likewise.
3697 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3699 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3702 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3705 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3707 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3709 * config/arm/aarch-common-protos.h
3710 (alu_cost_table): Fix spelling of "extend".
3711 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3713 2014-03-19 Richard Biener <rguenther@suse.de>
3716 * tree-core.h (tree_type_common): Re-order pointer members
3717 to reduce recursion depth during GC walks.
3719 2014-03-19 Marek Polacek <polacek@redhat.com>
3722 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3723 before accessing it.
3725 2014-03-19 Richard Biener <rguenther@suse.de>
3728 * lto-streamer-in.c (input_function): In WPA stage do not drop
3731 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3733 PR tree-optimization/60559
3734 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3735 with build_zero_cst assignment.
3737 2014-03-18 Kai Tietz <ktietz@redhat.com>
3739 PR rtl-optimization/56356
3740 * sdbout.c (sdbout_parms): Verify that parms'
3741 incoming argument is valid.
3742 (sdbout_reg_parms): Likewise.
3744 2014-03-18 Richard Henderson <rth@redhat.com>
3747 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3748 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3749 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3751 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3753 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3754 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3755 Italicize plugin event names in description. Explain that
3756 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3757 Remind that no GCC functions should be called after PLUGIN_FINISH.
3758 Explain what pragmas with expansion are.
3760 2014-03-18 Martin Liska <mliska@suse.cz>
3762 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3763 gimple call statement is update.
3764 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3765 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3767 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3770 * ubsan.c (ubsan_instrument_unreachable): Call
3771 initialize_sanitizer_builtins.
3772 (ubsan_pass): Likewise.
3775 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3776 varpool_finalize_decl instead of rest_of_decl_compilation.
3778 2014-03-18 Richard Biener <rguenther@suse.de>
3780 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3781 by using bitmap_and_compl instead of bitmap_and_compl_into.
3782 (df_rd_transfer_function): Likewise.
3784 2014-03-18 Richard Biener <rguenther@suse.de>
3786 * doc/lto.texi (fresolution): Fix typo.
3788 2014-03-18 Richard Biener <rguenther@suse.de>
3790 * doc/invoke.texi (flto): Update for changes in 4.9.
3792 2014-03-18 Richard Biener <rguenther@suse.de>
3794 * doc/loop.texi: Remove section on the removed lambda framework.
3795 Update loop docs with recent changes in preserving loop structure.
3797 2014-03-18 Richard Biener <rguenther@suse.de>
3799 * doc/lto.texi (-fresolution): Document.
3801 2014-03-18 Richard Biener <rguenther@suse.de>
3803 * doc/contrib.texi: Adjust my name.
3805 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3808 * internal-fn.c: Include diagnostic-core.h.
3809 (expand_BUILTIN_EXPECT): New function.
3810 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3811 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3812 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3813 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3815 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3816 Revert 3 argument __builtin_expect code.
3817 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3818 * gimple-fold.c (gimple_fold_call): Likewise.
3819 * tree.h (fold_builtin_expect): New prototype.
3820 * builtins.c (build_builtin_expect_predicate): Add predictor
3821 argument, if non-NULL, create 3 argument __builtin_expect.
3822 (fold_builtin_expect): No longer static. Add ARG2 argument,
3823 pass it through to build_builtin_expect_predicate.
3824 (fold_builtin_2): Adjust caller.
3825 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3826 * internal-fn.def (BUILTIN_EXPECT): New.
3828 2014-03-18 Tobias Burnus <burnus@net-b.de>
3831 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3832 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3833 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3835 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3838 * predict.c (combine_predictions_for_bb): Fix up formatting.
3839 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3840 fill what it points to if non-NULL.
3841 (tree_predict_by_opcode): Adjust caller, use the predictor.
3842 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3844 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3846 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3847 proper constant for the store mode.
3849 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3851 * symtab.c (change_decl_assembler_name): Fix transparent alias
3854 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3856 * config/aarch64/aarch64.c: Correct the comments about the
3857 aarch64 stack layout.
3859 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3861 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3862 check for GF_OMP_FOR_KIND_FOR.
3864 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3866 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3867 ymm and zmm register names.
3869 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3872 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3873 note creation for the 2010-08-31 changes.
3875 2014-03-17 Marek Polacek <polacek@redhat.com>
3878 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3879 as -fno-tree-loop-vectorize.
3880 (expand_omp_simd): Likewise.
3882 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3884 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3885 (eligible_for_call_delay): New prototype.
3886 * config/sparc/sparc.c (tls_call_delay): Rename into...
3887 (eligible_for_call_delay): ...this. Return false if the instruction
3888 cannot be put in the delay slot of a branch.
3889 (eligible_for_restore_insn): Simplify.
3890 (eligible_for_return_delay): Return false if the instruction cannot be
3891 put in the delay slot of a branch and simplify.
3892 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3893 put in the delay slot of a branch.
3894 * config/sparc/sparc.md (fix_ut699): New attribute.
3895 (tls_call_delay): Delete.
3896 (in_call_delay): Reimplement.
3897 (eligible_for_sibcall_delay): Rename into...
3898 (in_sibcall_delay): ...this.
3899 (eligible_for_return_delay): Rename into...
3900 (in_return_delay): ...this.
3901 (in_branch_delay): Reimplement.
3902 (in_uncond_branch_delay): Delete.
3903 (in_annul_branch_delay): Delete.
3905 2014-03-14 Richard Henderson <rth@redhat.com>
3908 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3909 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3910 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3911 (floathi splitters): Remove.
3912 (float<SWI48x>xf2): New pattern.
3913 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3914 code that tried to handle DImode for 32-bit, but which was excluded
3915 by the pattern's condition. Drop allocation of stack temporary.
3916 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3917 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3918 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3919 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3920 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3921 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3922 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3923 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3924 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3925 (*float<SWI48x><X87MODEF>2_i387): Remove.
3926 (all float _with_temp splitters): Remove.
3927 (*float<SWI48x><MODEF>2_i387): New pattern.
3928 (*float<SWI48><MODEF>2_sse): New pattern.
3929 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3930 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3932 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3933 Marek Polacek <polacek@redhat.com>
3936 * common.opt (dump_base_name_prefixed): New Variable.
3937 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3938 if x_dump_base_name_prefixed is already set, set it at the end.
3940 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3942 PR rtl-optimization/60508
3943 * lra-constraints.c (get_reload_reg): Add new parameter
3945 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3946 Pass the new parameter values.
3948 2014-03-14 Richard Biener <rguenther@suse.de>
3950 * common.opt: Revert unintented changes from r205065.
3953 2014-03-14 Richard Biener <rguenther@suse.de>
3956 * cfghooks.c (split_block): Properly adjust all loops the
3957 block was a latch of.
3959 2014-03-14 Martin Jambor <mjambor@suse.cz>
3962 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3965 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3968 * config/avr/avr.c (avr_set_current_function): Pass function name
3969 through default_strip_name_encoding before sanity checking instead
3970 of skipping the first char of the assembler name.
3972 2014-03-13 Richard Henderson <rth@redhat.com>
3975 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3976 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3977 * config/i386/i386-protos.h: Likewise.
3978 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3979 in the expander instead of a splitter.
3980 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3981 any possibility of requiring a memory.
3982 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3983 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3984 (fp branch splitters): Update for ix86_split_fp_branch.
3985 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3986 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3987 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3988 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3989 (*fop_<MODEF>_3_i387): Likewise.
3990 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3991 (splitters for the fop_* register patterns): Remove.
3992 (fscalexf4_i387): Rename from *fscalexf4_i387.
3993 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3995 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3997 PR tree-optimization/59779
3998 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3999 type for bitsize and maxsize instead of HOST_WIDE_INT.
4001 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
4003 PR rtl-optimization/57320
4004 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
4005 the CFG after thread_prologue_and_epilogue_insns.
4007 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
4009 PR rtl-optimization/57189
4010 * lra-constraints.c (process_alt_operands): Disfavor spilling
4013 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4015 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4017 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4019 PR tree-optimization/59025
4021 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4022 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4024 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4027 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4028 calls of avr_out_plus_1.
4030 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4032 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4033 BB's single pred and update the father loop's latch info later.
4035 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4037 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4041 (VEC_base): Likewise.
4042 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4043 registers, we need to swap double words in little endian mode.
4045 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4046 to be a container mode for 128-bit integer operations added in ISA
4047 2.07. Unlike TImode and PTImode, the preferred register set is
4048 the Altivec/VMX registers for the 128-bit operations.
4050 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4052 (rs6000_split_128bit_ok_p): Likewise.
4054 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4055 macros for creating ISA 2.07 normal and overloaded builtin
4056 functions with 3 arguments.
4057 (BU_P8V_OVERLOAD_3): Likewise.
4058 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4059 for use as overloaded functions.
4060 (VPERM_1TI_UNS): Likewise.
4061 (VSEL_1TI): Likewise.
4062 (VSEL_1TI_UNS): Likewise.
4063 (ST_INTERNAL_1ti): Likewise.
4064 (LD_INTERNAL_1ti): Likewise.
4065 (XXSEL_1TI): Likewise.
4066 (XXSEL_1TI_UNS): Likewise.
4067 (VPERM_1TI): Likewise.
4068 (VPERM_1TI_UNS): Likewise.
4069 (XXPERMDI_1TI): Likewise.
4070 (SET_1TI): Likewise.
4071 (LXVD2X_V1TI): Likewise.
4072 (STXVD2X_V1TI): Likewise.
4073 (VEC_INIT_V1TI): Likewise.
4074 (VEC_SET_V1TI): Likewise.
4075 (VEC_EXT_V1TI): Likewise.
4076 (EQV_V1TI): Likewise.
4077 (NAND_V1TI): Likewise.
4078 (ORC_V1TI): Likewise.
4079 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4080 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4082 (VADDUQM): Likewise.
4083 (VSUBCUQ): Likewise.
4084 (VADDEUQM): Likewise.
4085 (VADDECUQ): Likewise.
4086 (VSUBEUQM): Likewise.
4087 (VSUBECUQ): Likewise.
4089 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
4090 __int128_t and __uint128_t types.
4091 (__uint128_type): Likewise.
4092 (altivec_categorize_keyword): Add support for vector __int128_t,
4093 vector __uint128_t, vector __int128, and vector unsigned __int128
4094 as a container type for TImode operations that need to be done in
4095 VSX/Altivec registers.
4096 (rs6000_macro_to_expand): Likewise.
4097 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
4098 to support 128-bit integer instructions vaddcuq, vadduqm,
4099 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
4100 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4102 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4103 for V1TImode, and set up preferences to use VSX/Altivec registers.
4104 Setup VSX reload handlers.
4105 (rs6000_debug_reg_global): Likewise.
4106 (rs6000_init_hard_regno_mode_ok): Likewise.
4107 (rs6000_preferred_simd_mode): Likewise.
4108 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4109 (easy_altivec_constant): Likewise.
4110 (output_vec_const_move): Likewise.
4111 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4113 (rs6000_expand_vector_extract): Likewise.
4114 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4116 (rs6000_const_vec): Add support for V1TImode.
4117 (rs6000_emit_le_vsx_load): Swap double words when loading or
4118 storing TImode/V1TImode.
4119 (rs6000_emit_le_vsx_store): Likewise.
4120 (rs6000_emit_le_vsx_move): Likewise.
4121 (rs6000_emit_move): Add support for V1TImode.
4122 (altivec_expand_ld_builtin): Likewise.
4123 (altivec_expand_st_builtin): Likewise.
4124 (altivec_expand_vec_init_builtin): Likewise.
4125 (altivec_expand_builtin): Likewise.
4126 (rs6000_init_builtins): Add support for V1TImode type. Add
4127 support for ISA 2.07 128-bit integer builtins. Define type names
4128 for the VSX/Altivec vector types.
4129 (altivec_init_builtins): Add support for overloaded vector
4130 functions with V1TImode type.
4131 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4132 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4134 (rs6000_split_128bit_ok_p): Likewise.
4135 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4136 __int128_t and vector __uint128_t.
4138 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4139 and mode attributes.
4146 (VS_scalar): Likewise.
4147 (VS_double): Likewise.
4148 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4150 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4151 we support the ISA 2.07 128-bit integer arithmetic instructions.
4152 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4153 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4154 and TImode types for use with the builtin functions.
4155 (V1TI_type_node): Likewise.
4156 (unsigned_V1TI_type_node): Likewise.
4157 (intTI_type_internal_node): Likewise.
4158 (uintTI_type_internal_node): Likewise.
4160 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4161 128-bit builtin functions.
4162 (UNSPEC_VADDEUQM): Likewise.
4163 (UNSPEC_VADDECUQ): Likewise.
4164 (UNSPEC_VSUBCUQ): Likewise.
4165 (UNSPEC_VSUBEUQM): Likewise.
4166 (UNSPEC_VSUBECUQ): Likewise.
4167 (VM): Add V1TImode to vector mode iterators.
4169 (VI_unit): Likewise.
4170 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4171 (altivec_vaddcuq): Likewise.
4172 (altivec_vsubuqm): Likewise.
4173 (altivec_vsubcuq): Likewise.
4174 (altivec_vaddeuqm): Likewise.
4175 (altivec_vaddecuq): Likewise.
4176 (altivec_vsubeuqm): Likewise.
4177 (altivec_vsubecuq): Likewise.
4179 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4181 (BOOL_128): Likewise.
4182 (BOOL_REGS_OUTPUT): Likewise.
4183 (BOOL_REGS_OP1): Likewise.
4184 (BOOL_REGS_OP2): Likewise.
4185 (BOOL_REGS_UNARY): Likewise.
4186 (BOOL_REGS_AND_CR0): Likewise.
4188 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4189 128-bit integer builtin support.
4190 (vec_vadduqm): Likewise.
4191 (vec_vaddecuq): Likewise.
4192 (vec_vaddeuqm): Likewise.
4193 (vec_vsubecuq): Likewise.
4194 (vec_vsubeuqm): Likewise.
4195 (vec_vsubcuq): Likewise.
4196 (vec_vsubuqm): Likewise.
4198 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4199 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4200 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4201 128-bit integer add/subtract to ISA 2.07.
4203 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4205 * config/arc/arc.c (arc_predicate_delay_insns):
4206 Fix third argument passed to conditionalize_nonjump.
4208 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4210 * config/aarch64/aarch64-builtins.c
4211 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4212 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4213 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4214 instead of __builtin_lfloor.
4215 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4217 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4219 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4220 (tree_ssa_ifcombine_bb_1): New function.
4221 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4222 is an empty forwarder block to then_bb or vice versa and then_bb
4223 and else_bb are effectively swapped.
4225 2014-03-12 Christian Bruel <christian.bruel@st.com>
4228 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4229 REG_CFA_DEF_CFA note.
4230 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4231 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4233 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4235 PR tree-optimization/60454
4236 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4238 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4240 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4241 Do not define target_cpu_default2 to generic.
4242 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4243 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4244 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4246 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4247 Marc Glisse <marc.glisse@inria.fr>
4249 PR tree-optimization/60502
4250 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4251 instead of build_low_bits_mask.
4253 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4256 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4257 if there are multiple uses, but op doesn't live on E edge.
4258 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4259 clobber stmts before __builtin_unreachable.
4261 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4263 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4264 hard_frame_pointer_rtx.
4265 * cse.c (cse_insn): Remove volatile check.
4266 * cselib.c (cselib_process_insn): Likewise.
4267 * dse.c (scan_insn): Likewise.
4269 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4271 * config/arc/arc.c (conditionalize_nonjump): New function,
4273 (arc_ifcvt): ... this.
4274 (arc_predicate_delay_insns): Use it.
4276 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4278 * config/arc/predicates.md (extend_operand): During/after reload,
4279 allow const_int_operand.
4280 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4281 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4282 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4284 (umulsi3_highpart_i): Likewise.
4286 2014-03-11 Richard Biener <rguenther@suse.de>
4288 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4289 Add asserts to guard possible wrong-code bugs.
4291 2014-03-11 Richard Biener <rguenther@suse.de>
4293 PR tree-optimization/60429
4294 PR tree-optimization/60485
4295 * tree-ssa-structalias.c (set_union_with_increment): Properly
4296 take into account all fields that overlap the shifted vars.
4297 (do_sd_constraint): Likewise.
4298 (do_ds_constraint): Likewise.
4299 (get_constraint_for_ptr_offset): Likewise.
4301 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4303 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4304 (nios2_compute_frame_layout):
4305 Add calculation of cfun->machine->fp_save_offset.
4306 (nios2_expand_prologue): Correct setting of frame pointer register
4308 (nios2_expand_epilogue): Update recovery of stack pointer from
4309 frame pointer accordingly.
4310 (nios2_initial_elimination_offset): Update calculation of offset
4311 for eliminating to HARD_FRAME_POINTER_REGNUM.
4313 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4316 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4317 cgraph_get_create_node on VAR_DECLs.
4319 2014-03-10 Richard Biener <rguenther@suse.de>
4322 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4324 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4326 * config/vms/vms.opt (vms_float_format): New variable.
4328 2014-03-08 Tobias Burnus <burnus@net-b.de>
4330 * doc/invoke.texi (-fcilkplus): Update implementation status.
4332 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4333 Richard Biener <rguenther@suse.de>
4335 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4336 consistently accross all TUs.
4337 (run_gcc): Enable -fshort-double automatically at link at link-time
4338 and disallow override.
4340 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4343 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4344 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4345 if they can't be used.
4347 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4349 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4350 for Solaris 11/x86 ld.
4351 * configure: Regenerate.
4353 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4355 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4356 (LIB_TLS_SPEC): Save as ld_tls_libs.
4357 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4358 (HAVE_AS_IX86_TLSLDM): New test.
4359 * configure, config.in: Regenerate.
4360 * config/i386/i386.c (legitimize_tls_address): Fall back to
4361 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4362 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4363 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4364 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4366 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4368 * common.opt (fira-loop-pressure): Mark as optimization.
4370 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4372 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4373 an OpenMP mappable type.
4375 2014-03-06 Matthias Klose <doko@ubuntu.com>
4377 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4378 MULTILIB_OSDIRNAMES is not defined.
4380 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4381 Meador Inge <meadori@codesourcery.com>
4384 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4385 (arm_legitimize_address): Call legitimize_tls_address for any
4386 arm_tls_referenced_p expression, handle constant addend. Call it
4387 before testing for !TARGET_ARM.
4388 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4390 2014-03-06 Richard Biener <rguenther@suse.de>
4396 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4398 * tree-streamer.c (record_common_node): Assert we don't record
4399 nodes with type double.
4400 (preload_common_node): Skip type double, complex double and double
4401 pointer since it is now frontend dependent due to fshort-double option.
4403 2014-03-06 Richard Biener <rguenther@suse.de>
4405 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4406 or -fno-lto is specified and the linker has full plugin support.
4407 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4408 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4409 * lto-wrapper.c (merge_and_complain): Merge compile-time
4410 optimization levels.
4411 (run_gcc): And pass it through to the link options.
4413 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4417 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4419 * cselib.c (remove_useless_values): Skip to avoid quadratic
4420 behavior if the condition moved from...
4421 (cselib_process_insn): ... here holds.
4423 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4426 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4427 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4430 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4431 (TM_H): Add x86-tune.def.
4433 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4435 * config/aarch64/aarch64.c (generic_tunings):
4436 Use cortexa57_extra_costs.
4438 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4441 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4442 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4443 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4446 2014-03-04 Heiher <r@hev.cc>
4448 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4449 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4451 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4453 * config/i386/predicates.md (const2356_operand): Change to ...
4454 (const2367_operand): ... this.
4455 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4457 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4458 (*avx512pf_scatterpf<mode>sf): Ditto.
4459 (avx512pf_scatterpf<mode>df): Ditto.
4460 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4461 (*avx512pf_scatterpf<mode>df): Ditto.
4462 * config/i386/i386.c (ix86_expand_builtin): Update
4463 incorrect hint operand error message.
4465 2014-03-04 Richard Biener <rguenther@suse.de>
4467 * lto-section-in.c (lto_get_section_data): Fix const cast.
4469 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4471 * tree-streamer.c (record_common_node): Assert we don't record
4472 nodes with type double.
4473 (preload_common_node): Skip type double, complex double and double
4474 pointer since it is now frontend dependent due to fshort-double option.
4476 2014-03-04 Richard Biener <rguenther@suse.de>
4479 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4480 (lto_input_toplevel_asms): Likewise.
4481 * lto-section-in.c (lto_get_section_data): Instead do it here
4484 2014-03-04 Richard Biener <rguenther@suse.de>
4486 PR tree-optimization/60382
4487 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4488 dead PHIs a reduction.
4490 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4492 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4494 (_mm_prefetch): Move out of GCC target("sse") pragma.
4495 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4496 GCC target("prfchw") pragma.
4497 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4499 * config/i386/i386.c (ix86_option_override_internal): Enable
4500 -mprfchw with -mprefetchwt1.
4502 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4504 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4507 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4509 * opts.h (CL_PCH_IGNORE): Define.
4510 * targhooks.c (option_affects_pch_p):
4511 Return false for options that have CL_PCH_IGNORE set.
4512 * opt-functions.awk: Process PchIgnore.
4513 * doc/options.texi: Document PchIgnore.
4515 * config/arc/arc.opt (misize): Add PchIgnore property.
4517 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4519 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4520 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4521 constraint on constants to permit them being loaded into
4522 GENERAL_REGS or BASE_REGS.
4524 2014-03-03 Nick Clifton <nickc@redhat.com>
4526 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4527 anti-cacnonical alternatives.
4528 (negandhi3_real): New pattern.
4529 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4531 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4533 * config/avr/avr-mcus.def: Remove atxmega16x1.
4534 * config/avr/avr-tables.opt: Regenerate.
4535 * config/avr/t-multilib: Regenerate.
4536 * doc/avr-mmcu.texi: Regenerate.
4538 2014-03-03 Tobias Grosser <tobias@grosser.es>
4539 Mircea Namolaru <mircea.namolaru@inria.fr>
4541 PR tree-optimization/58028
4542 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4545 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4547 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4548 not handled by recognizers.
4550 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4553 * function.c (expand_function_end): Don't emit
4554 clobber_return_register sequence if clobber_after is a BARRIER.
4555 * cfgexpand.c (construct_exit_block): Append instructions before
4556 return_label to prev_bb.
4558 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4560 * config/rs6000/constraints.md: Document reserved use of "wc".
4562 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4565 * ipa.c (function_and_variable_visibility): When dissolving comdat
4566 group, also set all symbols to local.
4568 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4573 2013-12-14 Jan Hubicka <jh@suse.cz>
4575 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4577 2014-03-02 Jon Beniston <jon@beniston.com>
4583 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4584 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4585 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4586 (simple_return, *simple_return): New patterns
4587 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4588 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4590 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4592 * dwarf2out.c (gen_subprogram_die): Tidy.
4594 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4597 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4598 (*mov_t_msb_neg_negc): ... this new insn.
4600 2014-02-28 Jason Merrill <jason@redhat.com>
4603 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4606 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4609 * dwarf2out.c (decltype_auto_die): New static.
4610 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4611 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4612 (is_cxx_auto): Likewise.
4614 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4616 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4617 we are not using general regs only.
4619 2014-02-28 Richard Biener <rguenther@suse.de>
4622 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4623 previous fix and only allow to remove trivial pre-headers
4624 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4625 (remove_forwarder_block): Properly update the latch of a loop.
4627 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4630 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4631 (cselib_preserved_hash_table): New.
4632 (preserve_constants_and_equivs): Move preserved vals to it.
4633 (cselib_find_slot): Look it up first.
4634 (cselib_init): Initialize it.
4635 (cselib_finish): Release it.
4636 (dump_cselib_table): Dump it.
4638 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4641 * cselib.c (remove_useless_values): Skip to avoid quadratic
4642 behavior if the condition moved from...
4643 (cselib_process_insn): ... here holds.
4645 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4648 * var-tracking.c (vt_initialize): Apply the same condition to
4649 preserve the CFA base value.
4651 2014-02-28 Joey Ye <joey.ye@arm.com>
4654 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4655 if reload in progress or completed.
4657 2014-02-28 Tobias Burnus <burnus@net-b.de>
4660 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4663 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4665 * doc/tm.texi.in (Condition Code Status): Update documention for
4666 relative locations of cc0-setter and cc0-user.
4668 2014-02-27 Jeff Law <law@redhat.com>
4670 PR rtl-optimization/52714
4671 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4672 into two independent simple sets, if I3 is a jump, ensure the
4673 pattern we place into I3 is a (set (pc) ...).
4675 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4676 Jeff Law <law@redhat.com>
4678 PR rtl-optimization/49847
4679 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4680 are in different blocks.
4681 * doc/tm.texi (Condition Code Status): Update documention for
4682 relative locations of cc0-setter and cc0-user.
4684 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4687 * lra.c (lra_emit_add): Check SUBREG too.
4689 2014-02-27 Andreas Schwab <schwab@suse.de>
4691 * config/m68k/m68k.c (m68k_option_override): Disable
4692 -flive-range-shrinkage for classic m68k.
4693 (m68k_override_options_after_change): Likewise.
4695 2014-02-27 Marek Polacek <polacek@redhat.com>
4698 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4699 -Wmaybe-uninitialized.
4701 2014-02-27 Alan Modra <amodra@gmail.com>
4704 * reload1.c (emit_input_reload_insns): When reload_override_in,
4705 set old to rl->in_reg when rl->in_reg is a subreg.
4707 2014-02-26 Richard Biener <rguenther@suse.de>
4710 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4712 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4714 * common/config/i386/predicates.md (const1256_operand): Remove.
4715 (const2356_operand): New.
4716 (const_1_to_2_operand): Remove.
4717 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4718 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4719 (*avx512pf_gatherpf<mode>sf): Ditto.
4720 (avx512pf_gatherpf<mode>df): Ditto.
4721 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4722 (*avx512pf_gatherpf<mode>df): Ditto.
4723 (avx512pf_scatterpf<mode>sf): Ditto.
4724 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4725 (*avx512pf_scatterpf<mode>sf): Ditto.
4726 (avx512pf_scatterpf<mode>df): Ditto.
4727 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4728 (*avx512pf_scatterpf<mode>df): Ditto.
4729 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4731 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4733 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4734 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4735 (_mm512_mask_testn_epi64_mask): Move to ...
4736 * config/i386/avx512cdintrin.h: Here.
4737 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4738 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4739 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4740 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4741 TARGET_AVX512F from TARGET_AVX512CD.
4743 2014-02-26 Richard Biener <rguenther@suse.de>
4746 * ipa.c (walk_polymorphic_call_targets): Properly guard
4747 call to inline_update_overall_summary.
4749 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4752 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4753 and latches only if requested. Fix latch if it is removed.
4754 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4755 LOOPS_HAVE_PREHEADERS.
4757 2014-02-25 Andrew Pinski <apinski@cavium.com>
4759 * builtins.c (expand_builtin_thread_pointer): Create a new target
4760 when the target is NULL.
4762 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4764 PR rtl-optimization/60317
4765 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4766 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4767 * lra-assigns.c: Include params.h.
4768 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4769 other reload pseudos considerations.
4771 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4773 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4774 to use canonical form for nor<mode>3.
4776 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4779 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4782 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4784 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4785 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4786 (ix86_handle_option): Handle OPT_mprefetchwt1.
4787 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4788 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4790 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4791 OPTION_MASK_ISA_PREFETCHWT1.
4792 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4793 (PTA_PREFETCHWT1): New.
4794 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4795 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4796 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4797 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4798 (*prefetch_avx512pf_<mode>_: Change into ...
4799 (*prefetch_prefetchwt1_<mode>: This.
4800 * config/i386/i386.opt (mprefetchwt1): New.
4801 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4802 (_mm_prefetch): Handle intent to write.
4803 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4805 2014-02-25 Richard Biener <rguenther@suse.de>
4808 * emit-rtl.c (mem_attrs_htab): Remove.
4809 (mem_attrs_htab_hash): Likewise.
4810 (mem_attrs_htab_eq): Likewise.
4811 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4812 (init_emit_once): Do not allocate mem_attrs_htab.
4814 2014-02-25 Richard Biener <rguenther@suse.de>
4817 * lto-opts.c (lto_write_options): Output non-explicit conservative
4818 -fwrapv, -fno-trapv and -fno-strict-overflow.
4819 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4820 (run_gcc): And pass them through.
4822 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4824 * sel-sched.c (calculate_new_fences): New parameter ptime.
4825 Calculate it as a maximum over all fence cycles.
4826 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4827 Print the final schedule timing when sched_verbose.
4829 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4831 PR rtl-optimization/60292
4832 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4833 bit fot the fence instruction.
4835 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4837 * calls.h: Fix typo in comment.
4839 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4841 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4842 adjusting offsetable addresses.
4844 2014-02-24 Guozhi Wei <carrot@google.com>
4846 * sparseset.h (sparseset_pop): Fix the wrong index.
4848 2014-02-24 Walter Lee <walt@tilera.com>
4850 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4851 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4853 * common/config/tilegx/tilegx-common.c
4854 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4855 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4857 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4858 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4859 (tilegx_gimplify_va_arg_expr): Handle big endian.
4860 (tilegx_expand_unaligned_load): Ditto.
4861 (tilegx_expand_unaligned_store): Ditto.
4862 (TARGET_RETURN_IN_MSB): New.
4863 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4864 (TARGET_ENDIAN_DEFAULT): New.
4865 (TARGET_BIG_ENDIAN): Handle big endian.
4866 (BYTES_BIG_ENDIAN): Ditto.
4867 (WORDS_BIG_ENDIAN): Ditto.
4868 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4871 * config/tilegx/tilegx.md (extv): Handle big endian.
4873 (insn_st<n>): Ditto.
4874 (insn_st<n>_add<bitsuffix>): Ditto.
4875 (insn_stnt<n>): Ditto.
4876 (insn_stnt<n>_add<bitsuffix>):Ditto.
4877 (vec_interleave_highv8qi): Handle big endian.
4878 (vec_interleave_highv8qi_be): New.
4879 (vec_interleave_highv8qi_le): New.
4880 (insn_v1int_h): Handle big endian.
4881 (vec_interleave_lowv8qi): Handle big endian.
4882 (vec_interleave_lowv8qi_be): New.
4883 (vec_interleave_lowv8qi_le): New.
4884 (insn_v1int_l): Handle big endian.
4885 (vec_interleave_highv4hi): Handle big endian.
4886 (vec_interleave_highv4hi_be): New.
4887 (vec_interleave_highv4hi_le): New.
4888 (insn_v2int_h): Handle big endian.
4889 (vec_interleave_lowv4hi): Handle big endian.
4890 (vec_interleave_lowv4hi_be): New.
4891 (vec_interleave_lowv4hi_le): New.
4892 (insn_v2int_l): Handle big endian.
4893 (vec_interleave_highv2si): Handle big endian.
4894 (vec_interleave_highv2si_be): New.
4895 (vec_interleave_highv2si_le): New.
4896 (insn_v4int_h): Handle big endian.
4897 (vec_interleave_lowv2si): Handle big endian.
4898 (vec_interleave_lowv2si_be): New.
4899 (vec_interleave_lowv2si_le): New.
4900 (insn_v4int_l): Handle big endian.
4901 * config/tilegx/tilegx.opt (mbig-endian): New option.
4902 (mlittle-endian): New option.
4903 * doc/install.texi: Document tilegxbe-linux.
4904 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4906 2014-02-24 Martin Jambor <mjambor@suse.cz>
4909 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4910 there are no parameter descriptors.
4912 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4914 PR rtl-optimization/60268
4915 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4916 initialization to ...
4917 (sched_rgn_init): ... here.
4918 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4920 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4922 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4925 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4927 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4930 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4932 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4933 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4935 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4937 * config/microblaze/predicates.md: Add cmp_op predicate.
4938 * config/microblaze/microblaze.md: Add branch_compare instruction
4939 which uses cmp_op predicate and emits cmp insn before branch.
4940 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4941 to microblaze_expand_conditional_branch and consolidate logic.
4942 (microblaze_expand_conditional_branch): emit branch_compare
4943 insn instead of handling cmp op separate from branch insn.
4945 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4947 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4950 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4952 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4953 define_insn with define_expand and new define_insn
4954 *altivec_lve<VI_char>x_internal.
4955 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4956 and new define_insn *altivec_stve<VI_char>x_internal.
4957 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4959 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4961 (altivec_expand_stvex_be): New function.
4963 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4965 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4966 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4967 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4968 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4970 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4973 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4974 instead of emit_move_insn.
4976 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4978 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4980 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4981 gen_altivec_vsumsws.
4983 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4985 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4986 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4987 (altivec_lvxl_<mode>): New define_expand incorporating
4988 -maltivec=be semantics where needed.
4989 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4990 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4991 semantics where needed.
4992 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4993 (altivec_stvx_<mode>): New define_expand incorporating
4994 -maltivec=be semantics where needed.
4995 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4996 VM2 iterator instead of V4SI.
4997 (altivec_stvxl_<mode>): New define_expand incorporating
4998 -maltivec=be semantics where needed.
4999 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
5000 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
5001 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
5002 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
5003 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
5004 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
5005 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
5006 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
5007 ALTIVEC_BUILTIN_STVXL.
5008 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
5009 (altivec_expand_stvx_be): Likewise.
5010 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5011 (altivec_expand_lvx_be): Likewise.
5012 (altivec_expand_stvx_be): Likewise.
5013 (altivec_expand_builtin): Add cases for
5014 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5015 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5016 (altivec_init_builtins): Add definitions for
5017 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5018 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5020 2014-02-21 Catherine Moore <clm@codesourcery.com>
5022 * doc/invoke.texi (mvirt, mno-virt): Document.
5023 * config/mips/mips.opt (mvirt): New option.
5024 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5026 2014-02-21 Richard Biener <rguenther@suse.de>
5028 PR tree-optimization/60276
5029 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5030 (STMT_VINFO_MIN_NEG_DIST): New macro.
5031 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5032 STMT_VINFO_MIN_NEG_DIST.
5033 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5034 made for negative dependence distances still hold.
5036 2014-02-21 Richard Biener <rguenther@suse.de>
5039 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5040 DECL_INITIAL for globals not in the current function context.
5042 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5044 PR tree-optimization/56490
5045 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5046 * tree-ssa-uninit.c: Include params.h.
5047 (compute_control_dep_chain): Add num_calls argument, return false
5048 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5049 num_calls to recursive call.
5050 (find_predicates): Change dep_chain into normal array,
5051 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5052 variable and adjust compute_control_dep_chain caller.
5053 (find_def_preds): Likewise.
5055 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5057 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5058 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5060 2014-02-21 Nick Clifton <nickc@redhat.com>
5062 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5063 (pushhi1): Likewise.
5064 (popqi1): Add mode to pre_dec.
5067 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5069 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5070 mode for mask of V8SFmode permutation.
5072 2014-02-20 Richard Henderson <rth@redhat.com>
5075 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5076 a new pseudo for OLDVAL.
5078 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5081 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5082 gen_reg_rtx if d->testing_p.
5083 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5084 if d->testing_p and we will certainly return true.
5085 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
5088 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
5090 * emit-rtl.c (gen_reg_rtx): Assert that
5091 crtl->emit.regno_pointer_align_length is non-zero.
5093 2014-02-20 Richard Henderson <rth@redhat.com>
5096 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
5097 on failure the store back into EXPECT.
5099 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
5100 Sandra Loosemore <sandra@codesourcery.com>
5102 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5103 * config/nios2/nios2.c (nios2_function_profiler): Add
5104 -fPIC (flag_pic == 2) support.
5105 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5106 (nios2_large_offset_p): New function.
5107 (nios2_unspec_reloc_p): Move up position, update to use
5108 nios2_large_offset_p.
5109 (nios2_unspec_address): Remove function.
5110 (nios2_unspec_offset): New function.
5111 (nios2_large_got_address): New function.
5112 (nios2_got_address): Add large offset support.
5113 (nios2_legitimize_tls_address): Update usage of removed and new
5115 (nios2_symbol_binds_local_p): New function.
5116 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5117 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5118 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5119 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5120 processing for (const (unspec ...)).
5121 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5123 2014-02-20 Richard Biener <rguenther@suse.de>
5125 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5126 doing the substitution.
5127 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5129 2014-02-20 Martin Jambor <mjambor@suse.cz>
5132 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5133 info when checking whether lattices are bottom.
5135 2014-02-20 Richard Biener <rguenther@suse.de>
5138 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5141 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5144 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5145 parameter specifying the scaling.
5146 (inline_call): Update.
5147 (want_inline_recursively): Guard division by zero.
5148 (recursive_inlining): Update.
5149 * ipa-inline.h (clone_inlined_nodes): Update.
5151 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5154 * config/i386/i386.c (classify_argument): Pass structures of size
5155 64 bytes or less in register.
5157 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5158 Kirill Yukhin <kirill.yukhin@intel.com>
5160 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5161 (_mm_rcp28_round_ss): Ditto.
5162 (_mm_rsqrt28_round_sd): Ditto.
5163 (_mm_rsqrt28_round_ss): Ditto.
5164 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5165 (_mm_rcp14_round_ss): Ditto.
5166 (_mm_rsqrt14_round_sd): Ditto.
5167 (_mm_rsqrt14_round_ss): Ditto.
5168 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5169 the first input operand, get rid of match_dup.
5170 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5172 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5174 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5175 operand as the first input operand, set type attribute.
5176 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5178 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5179 operand as the first input operand, set type attribute.
5181 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5183 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5186 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5189 * config/i386/i386.c (construct_container): Remove TFmode check
5190 for X86_64_INTEGER_CLASS.
5192 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5195 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5196 only when -Wpsabi is enabled.
5198 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5201 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5202 passing arrays in registers are the same as for structs, so remove the
5203 special case for them.
5205 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5207 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5208 destination type, extract only the valid bits if the source type is not
5209 integral and has a different mode.
5211 2014-02-19 Richard Biener <rguenther@suse.de>
5214 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5217 2014-02-19 Richard Biener <rguenther@suse.de>
5220 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5221 (ipa_modify_call_arguments): Emit an argument load explicitely and
5222 preserve virtual SSA form there and for the replacement call.
5223 Do not update SSA form nor free dominance info.
5225 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5227 * ipa.c (function_and_variable_visibility): Also clear WEAK
5228 flag when disolving COMDAT_GROUP.
5230 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5232 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5233 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5235 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5236 do more sanity checks.
5237 (detect_type_change): Return true when giving up early.
5238 (compute_complex_assign_jump_func): Fix type parameter of
5239 ipa_set_ancestor_jf.
5240 (compute_complex_ancestor_jump_func): Likewise.
5241 (update_jump_functions_after_inlining): Fix updating of
5243 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5245 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5247 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5248 inline clones when edge disappears.
5250 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5253 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5254 Split 64-bit moves into 2 patterns. Do not allow the use of
5255 direct move for TDmode in little endian, since the decimal value
5256 has little endian bytes within a word, but the 64-bit pieces are
5257 ordered in a big endian fashion, and normal subreg's of TDmode are
5259 (mov<mode>_64bit_dm): Likewise.
5260 (movtd_64bit_nodm): Likewise.
5262 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5264 PR tree-optimization/60174
5265 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5266 statement of an SSA_NAME that occurs in an abnormal PHI node.
5268 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5271 * final.c (SEEN_BB): Remove.
5272 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5273 (final_scan_insn): Don't force_source_line on second
5274 NOTE_INSN_BASIC_BLOCK.
5276 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5279 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5280 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5281 (type_natural_mode): Warn ABI change when %zmm register is not
5282 available for AVX512F vector value passing.
5284 2014-02-18 Kai Tietz <ktietz@redhat.com>
5287 * config/i386/i386.c (ix86_expand_prologue): Use value in
5288 rax register as displacement when restoring %r10 or %rax.
5289 Fix wrong offset when restoring both registers.
5291 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5293 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5294 assertion with conditional return.
5296 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5297 Uros Bizjak <ubizjak@gmail.com>
5300 * config/i386/driver-i386.c (host_detect_local_cpu): If
5301 YMM state is not saved by the OS, also clear has_f16c. Move
5302 CPUID 0x80000001 handling before YMM state saving checking.
5304 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5306 PR rtl-optimization/58960
5307 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5308 factored out from ...
5309 (sched_init): ... here.
5310 (free_global_sched_pressure_data): New, factored out from ...
5311 (sched_finish): ... here.
5312 * sched-int.h (free_global_sched_pressure_data): Declare.
5313 * sched-rgn.c (nr_regions_initial): New static global.
5314 (haifa_find_rgns): Initialize it.
5315 (schedule_region): Disable sched-pressure for the newly
5318 2014-02-17 Richard Biener <rguenther@suse.de>
5320 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5321 release SSA defs of pattern stmts.
5323 2014-02-17 Richard Biener <rguenther@suse.de>
5325 * tree-inline.c (expand_call_inline): Release the virtual
5326 operand defined by the call we are about to inline.
5328 2014-02-17 Richard Biener <rguenther@suse.de>
5330 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5332 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5333 Ilya Tocar <ilya.tocar@intel.com>
5335 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5336 arguments order in builtin.
5337 (_mm512_permutexvar_epi64): Ditto.
5338 (_mm512_mask_permutexvar_epi64): Ditto
5339 (_mm512_maskz_permutexvar_epi32): Ditto
5340 (_mm512_permutexvar_epi32): Ditto
5341 (_mm512_mask_permutexvar_epi32): Ditto
5343 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5345 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5346 (p8_vmrgow): Likewise.
5348 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5350 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5353 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5356 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5357 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5358 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5359 using direct move instructions on ISA 2.07. Also adjust
5360 instruction length for 64-bit.
5361 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5362 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5364 2014-02-15 Alan Modra <amodra@gmail.com>
5368 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5369 find_replacement on parts of insn rtl that might be reloaded.
5371 2014-02-15 Richard Biener <rguenther@suse.de>
5373 PR tree-optimization/60183
5374 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5375 (tree_ssa_phiprop): Calculate and free post-dominators.
5377 2014-02-14 Jeff Law <law@redhat.com>
5379 PR rtl-optimization/60131
5380 * ree.c (get_extended_src_reg): New function.
5381 (combine_reaching_defs): Use it rather than assuming location of REG.
5382 (find_and_remove_re): Verify first operand of extension is
5383 a REG before adding the insns to the copy list.
5385 2014-02-14 Roland McGrath <mcgrathr@google.com>
5387 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5388 * configure: Regenerated.
5389 * config.in: Regenerated.
5390 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5391 instead of ASM_SHORT.
5393 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5394 Richard Earnshaw <rearnsha@arm.com>
5396 PR rtl-optimization/59535
5397 * lra-constraints.c (process_alt_operands): Encourage alternative
5398 when unassigned pseudo class is superset of the alternative class.
5399 (inherit_reload_reg): Don't inherit when optimizing for code size.
5400 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5401 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5402 modes not less than 4 for Thumb1.
5404 2014-02-14 Kyle McMartin <kyle@redhat.com>
5407 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5409 2014-02-14 Richard Biener <rguenther@suse.de>
5411 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5412 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5413 Do not strip INDIRECT_REFs.
5415 2014-02-14 Richard Biener <rguenther@suse.de>
5418 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5419 DECL_FUNCTION_SPECIFIC_TARGET.
5420 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5421 * tree-streamer-out.c (pack_ts_target_option): Remove.
5422 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5423 (write_ts_function_decl_tree_pointers): Do not stream
5424 DECL_FUNCTION_SPECIFIC_TARGET.
5425 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5426 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5427 (lto_input_ts_function_decl_tree_pointers): Do not stream
5428 DECL_FUNCTION_SPECIFIC_TARGET.
5430 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5432 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5433 (get_initial_def_for_induction, vectorizable_induction): Ignore
5434 debug stmts when looking for exit_phi.
5435 (vectorizable_live_operation): Fix up condition.
5437 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5439 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5440 nreverse() because it changes the content of original tree list.
5442 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5444 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5445 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5447 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5449 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5450 GNU coding standards.
5452 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5455 * dwarf2out.c (gen_subprogram_die): Don't call
5456 add_calling_convention_attribute if subr_die is old_die.
5458 2014-02-13 Sharad Singhai <singhai@google.com>
5460 * doc/optinfo.texi: Fix order of nodes.
5462 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5464 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5465 operands[2], not operands[3].
5467 2014-02-13 Richard Biener <rguenther@suse.de>
5470 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5471 mention the possibility of an in-tree build.
5472 (CLooG): Update recommended version to 0.18.1, mention the
5473 possibility of an in-tree build and clarify that the ISL
5474 bundled with CLooG does not work.
5476 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5479 * expr.c (compress_float_constant): If x is a hard register,
5480 extend into a pseudo and then move to x.
5482 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5484 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5485 caused by bad second argument to warning_at() with -mhotpatch and
5486 nested functions (e.g. with gfortran).
5488 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5490 * opts.c (option_name): Remove "enabled by default" rider.
5492 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5494 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5496 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5497 Uros Bizjak <ubizjak@gmail.com>
5500 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5501 * configure: Regenerated.
5503 2014-02-12 Richard Biener <rguenther@suse.de>
5505 * vec.c (vec_prefix::calculate_allocation): Move as
5506 inline variant to vec.h.
5507 (vec_prefix::calculate_allocation_1): New out-of-line version.
5508 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5509 (vec_prefix::m_has_auto_buf): Rename to ...
5510 (vec_prefix::m_using_auto_storage): ... this.
5511 (vec_prefix::calculate_allocation): Inline the easy cases
5512 and dispatch to calculate_allocation_1 which doesn't need the
5514 (va_heap::reserve): Use gcc_checking_assert.
5515 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5516 m_using_auto_storage.
5517 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5519 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5520 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5521 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5523 2014-02-12 Richard Biener <rguenther@suse.de>
5525 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5526 when we found a dependence.
5528 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5530 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5532 (maybe_fold_stmt): ... into this new function.
5533 * omp-low.c (lower_omp): Update comment.
5535 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5538 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5541 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5543 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5544 identifiers in comments.
5545 (cortexa53_extra_costs): Likewise.
5546 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5547 (cortexa7_extra_costs): Likewise.
5548 (cortexa12_extra_costs): Likewise.
5549 (cortexa15_extra_costs): Likewise.
5550 (v7m_extra_costs): Likewise.
5552 2014-02-12 Richard Biener <rguenther@suse.de>
5555 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5556 of posix_memalign being successful.
5557 (lower_stmt): Restrict lowering of posix_memalign to when
5558 -ftree-bit-ccp is enabled.
5560 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5562 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5564 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5566 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5568 PR rtl-optimization/60116
5569 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5570 other_insn once the combination has been validated.
5572 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5575 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5577 * ipa-devirt.c: Include demangle.h
5578 (odr_violation_reported): New static variable.
5579 (add_type_duplicate): Update odr_violations.
5580 (maybe_record_node): Add completep parameter; update it.
5581 (record_target_from_binfo): Add COMPLETEP parameter;
5582 update it as needed.
5583 (possible_polymorphic_call_targets_1): Likewise.
5584 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5585 rename FINAL to COMPLETE.
5586 (record_targets_from_bases): Sanity check we found the binfo;
5587 fix COMPLETEP updating.
5588 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5589 parameter, fix computing of COMPLETEP.
5590 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5591 at LTO time do demangling.
5592 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5593 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5595 (gimple_get_virt_method_for_binfo): Likewise.
5596 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5597 gimple_get_virt_method_for_vtable): Update prototypes.
5599 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5602 * genautomata.c (add_presence_absence): Fix typo with
5603 {final_}presence_list.
5605 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5608 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5609 for VSX/Altivec vectors that land in GPR registers.
5611 2014-02-11 Richard Henderson <rth@redhat.com>
5612 Jakub Jelinek <jakub@redhat.com>
5615 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5616 around drhs if type conversion to lacc->type is not useless.
5618 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5620 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5622 (cortex-a57.cortex-a53): Likewise.
5623 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5625 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5627 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5630 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5632 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5633 add_options_for_arm_vfp3.
5635 2014-02-11 Jeff Law <law@redhat.com>
5638 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5639 object with an undesirable mode.
5641 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5644 * config/i386/sol2-9.h: New file.
5645 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5646 *-*-solaris2.9*): Use it.
5648 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5650 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5651 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5653 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5655 * config/microblaze/microblaze.c: Extend mcpu version format
5657 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5659 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5661 2014-02-10 Richard Henderson <rth@redhat.com>
5664 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5665 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5666 ms-abi vs -mno-accumulate-outgoing-args.
5667 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5668 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5671 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5674 * cfgexpand.c (expand_asm_operands): Attach source location to
5675 ASM_INPUT rtx objects.
5676 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5678 2014-02-10 Nick Clifton <nickc@redhat.com>
5680 * config/mn10300/mn10300.c (popcount): New function.
5681 (mn10300_expand_prologue): Include saved registers in stack usage
5684 2014-02-10 Jeff Law <law@redhat.com>
5687 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5688 when changing the SET_DEST of a prior insn to avoid an input reload.
5690 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5692 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5693 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5694 -mcall-openbsd, or -mcall-linux.
5695 (CC1_ENDIAN_BIG_SPEC): Remove.
5696 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5697 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5698 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5699 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5700 and %cc1_endian_default.
5701 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5703 2014-02-10 Richard Biener <rguenther@suse.de>
5705 PR tree-optimization/60115
5706 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5707 MEM_REF handling. Properly verify that the accesses are not
5708 out of the objects bound.
5710 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5712 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5715 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5717 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5718 proper constants and fix formatting.
5719 (possible_polymorphic_call_targets): Fix formatting.
5721 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5722 Ilya Tocar <ilya.tocar@intel.com>
5724 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5725 (_mm512_loadu_epi32): Renamed into...
5726 (_mm512_loadu_si512): This.
5727 (_mm512_storeu_epi32): Renamed into...
5728 (_mm512_storeu_si512): This.
5729 (_mm512_maskz_ceil_ps): Removed.
5730 (_mm512_maskz_ceil_pd): Ditto.
5731 (_mm512_maskz_floor_ps): Ditto.
5732 (_mm512_maskz_floor_pd): Ditto.
5733 (_mm512_floor_round_ps): Ditto.
5734 (_mm512_floor_round_pd): Ditto.
5735 (_mm512_ceil_round_ps): Ditto.
5736 (_mm512_ceil_round_pd): Ditto.
5737 (_mm512_mask_floor_round_ps): Ditto.
5738 (_mm512_mask_floor_round_pd): Ditto.
5739 (_mm512_mask_ceil_round_ps): Ditto.
5740 (_mm512_mask_ceil_round_pd): Ditto.
5741 (_mm512_maskz_floor_round_ps): Ditto.
5742 (_mm512_maskz_floor_round_pd): Ditto.
5743 (_mm512_maskz_ceil_round_ps): Ditto.
5744 (_mm512_maskz_ceil_round_pd): Ditto.
5745 (_mm512_expand_pd): Ditto.
5746 (_mm512_expand_ps): Ditto.
5747 * config/i386/i386.c (ix86_builtins): Remove
5748 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5749 (bdesc_args): Ditto.
5750 * config/i386/predicates.md (const1256_operand): New.
5751 (const_1_to_2_operand): Ditto.
5752 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5753 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5754 (*avx512pf_gatherpf<mode>sf): Ditto.
5755 (avx512pf_gatherpf<mode>df): Ditto.
5756 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5757 (*avx512pf_gatherpf<mode>df): Ditto.
5758 (avx512pf_scatterpf<mode>sf): Ditto.
5759 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5760 (*avx512pf_scatterpf<mode>sf): Ditto.
5761 (avx512pf_scatterpf<mode>df): Ditto.
5762 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5763 (*avx512pf_scatterpf<mode>df): Ditto.
5764 (avx512f_expand<mode>): Removed.
5765 (<shift_insn><mode>3<mask_name>): Change predicate type.
5767 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5769 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5770 not at the end of datarefs vector use ordered_remove to avoid
5771 reordering datarefs vector.
5774 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5775 mark local addressable non-static vars as GOVD_PRIVATE
5776 instead of GOVD_LOCAL.
5777 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5778 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5782 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5783 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5784 assume_aligned or alloc_align attributes.
5785 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5786 arguments. Handle also assume_aligned and alloc_align attributes.
5787 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5788 calls to functions with assume_aligned or alloc_align attributes.
5789 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5791 2014-02-08 Terry Guo <terry.guo@arm.com>
5793 * doc/invoke.texi: Document ARM -march=armv7e-m.
5795 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5797 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5798 flag on __cilkrts_rethrow builtin.
5801 * ipa-cp.c (determine_versionability): Fail at -O0
5802 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5803 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5806 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5809 * tree-inline.c (copy_forbidden): Fail for
5810 __attribute__((optimize (0))) functions.
5812 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5814 * varpool.c: Include pointer-set.h.
5815 (varpool_remove_unreferenced_decls): Variables in other partitions
5816 will not be output; be however careful to not lose information
5819 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5821 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5822 lookup in the vtable constructor.
5824 2014-02-07 Jeff Law <law@redhat.com>
5827 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5828 define_insn_and_split.
5830 * ipa-inline.c (inline_small_functions): Fix typos.
5832 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5834 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5835 (s390_can_use_return_insn): Declare.
5836 * config/s390/s390.h (EPILOGUE_USES): Define.
5837 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5839 (s390_chunkify_start): Handle return JUMP_LABELs.
5840 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5841 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5842 (s390_can_use_return_insn): New functions.
5843 (s390_fix_long_loop_prediction): Handle conditional returns.
5844 (TARGET_SET_UP_BY_PROLOGUE): Define.
5845 * config/s390/s390.md (ANY_RETURN): New code iterator.
5846 (*creturn, *csimple_return, return, simple_return): New patterns.
5848 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5850 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5851 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5852 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5853 REG_CFA_RESTORE list when deciding not to restore a register.
5855 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5857 * config/s390/s390.c: Include tree-pass.h and context.h.
5858 (s390_early_mach): New function, split out from...
5859 (s390_emit_prologue): ...here.
5860 (pass_data_s390_early_mach): New pass structure.
5861 (pass_s390_early_mach): New class.
5862 (s390_option_override): Create and register early_mach pass.
5863 Move to end of file.
5865 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5867 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5868 to match for the exit block.
5870 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5872 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5873 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5874 Reject misaligned operands.
5876 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5878 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5880 2014-02-07 Richard Biener <rguenther@suse.de>
5883 * gimple-low.c (lower_builtin_posix_memalign): New function.
5884 (lower_stmt): Call it to lower posix_memalign in a way
5885 to make alignment info accessible.
5887 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5890 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5891 __builtin_setjmp_receiver.
5893 2014-02-07 Richard Biener <rguenther@suse.de>
5896 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5897 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5898 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5899 Handle BUILT_IN_POSIX_MEMALIGN.
5900 (find_func_clobbers): Likewise.
5901 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5902 (call_may_clobber_ref_p_1): Likewise.
5904 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5907 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5910 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5913 * lto-cgraph.c (lto_output_node): Use
5914 symtab_get_symbol_partitioning_class.
5915 (lto_output_varpool_node): likewise.
5916 (symtab_get_symbol_partitioning_class): Move here from
5918 * cgraph.h (symbol_partitioning_class): Likewise.
5919 (symtab_get_symbol_partitioning_class): Declare.
5921 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5923 * ggc.h (ggc_internal_cleared_alloc): New macro.
5924 * vec.h (vec_safe_copy): Handle memory stats.
5925 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5926 * target-globals.c (save_target_globals): Likewise.
5928 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5931 * expr.c (emit_move_resolve_push): Export; be bit more selective
5932 on when to clear alias set.
5933 * expr.h (emit_move_resolve_push): Declare.
5934 * function.h (struct function): Add tail_call_marked.
5935 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5936 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5937 * config/i386/i386.md (TImode move expander): De not call
5939 (FP push expanders): Preserve memory attributes.
5940 * config/i386/sse.md (push<mode>1): Remove.
5941 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5942 (ix86_expand_push): Remove.
5943 * config/i386/mmx.md (push<mode>1): Remove.
5945 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5947 PR rtl-optimization/60030
5948 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5949 lopart with paradoxical subreg before shifting it up by hprec.
5951 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5953 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5954 Remove extra newline at end of file.
5955 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5956 (arm_issue_rate): Handle cortexa57.
5957 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5958 (cortex-a57.cortex-a53): Likewise.
5960 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5963 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5964 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5966 (arm_expand_prologue): Adjust all callers.
5967 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5968 info, registers also at the lowest numbered registers side. Use
5969 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5973 * var-tracking.c (adjust_mems): Before adding a SET to
5974 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5976 2014-02-06 Alan Modra <amodra@gmail.com>
5979 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5980 change SDmode to DDmode when lra_in_progress.
5982 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5985 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5986 free_data_ref on the dr first, and before goto again also set dr
5987 to the next dr. For simd_lane_access, free old datarefs[i] before
5988 overwriting it. For get_vectype_for_scalar_type failure, don't
5989 free_data_ref if simd_lane_access.
5991 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5994 * tree.h (opts_for_fn): New inline function.
5995 (opt_for_fn): Define.
5996 * config/i386/i386.c (ix86_function_regparm): Use
5997 opt_for_fn (decl, optimize) instead of optimize.
5999 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
6001 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
6002 for SYMBOL_REF in large memory model.
6004 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6006 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
6008 (cortex-a57): Likewise.
6009 (cortex-a57.cortex-a53): Likewise.
6011 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6012 Kugan Vivekanandarajah <kuganv@linaro.org>
6014 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6016 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6018 2014-02-06 Richard Biener <rguenther@suse.de>
6020 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6021 set_loop_copy and initialize_original_copy_tables.
6023 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6025 * config/aarch64/aarch64-simd.md
6026 (aarch64_ashr_simddi): Change QI to SI.
6028 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6029 Jakub Jelinek <jakub@redhat.com>
6032 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6035 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6037 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6038 CODE_FOR_altivec_vpku[hw]um to
6039 CODE_FOR_altivec_vpku[hw]um_direct.
6040 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6041 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6042 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6043 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6045 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6047 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6048 generation for -maltivec=be.
6049 (altivec_vsumsws): Simplify redundant test.
6051 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6053 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6054 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6055 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6056 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6057 gen_altivec_vpkuwum.
6058 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6060 (altivec_vpks<VI_char>ss): Likewise.
6061 (altivec_vpks<VI_char>us): Likewise.
6062 (altivec_vpku<VI_char>us): Likewise.
6063 (altivec_vpku<VI_char>um): Likewise.
6064 (altivec_vpku<VI_char>um_direct): New (copy of
6065 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6067 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6068 target is little endian and -maltivec=be is not specified.
6069 (*altivec_vupkhs<VU_char>_direct): New (copy of
6070 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6071 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6072 target is little endian and -maltivec=be is not specified.
6073 (*altivec_vupkls<VU_char>_direct): New (copy of
6074 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6075 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6076 little endian and -maltivec=be is not specified.
6077 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6078 little endian and -maltivec=be is not specified.
6080 2014-02-05 Richard Henderson <rth@redhat.com>
6083 * combine-stack-adj.c: Revert r206943.
6084 * sched-int.h (struct deps_desc): Add last_args_size.
6085 * sched-deps.c (init_deps): Initialize it.
6086 (sched_analyze_insn): Add OUTPUT dependencies between insns that
6087 contain REG_ARGS_SIZE notes.
6089 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6091 * lto-cgraph.c (asm_nodes_output): Make global.
6092 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
6093 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
6094 (driver_handle_option): Handle OPT_fwpa.
6096 2014-02-05 Jakub Jelinek <jakub@redhat.com>
6099 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
6100 a comment typo and formatting issue. If odr_hash hasn't been
6101 created, return vNULL and set *completep to false.
6104 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6105 bb with no successors.
6107 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6110 * doc/invoke.texi (-march): Clarify documentation for ARM.
6114 2014-02-05 Richard Biener <rguenther@suse.de>
6116 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6117 when not vectorizing because of too many alias checks.
6118 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6119 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6121 2014-02-05 Nick Clifton <nickc@redhat.com>
6123 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6124 accept extended registers in any mode when compiling for the MN10300.
6126 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6128 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6129 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6130 sanitization attributes.
6131 (can_inline_edge_p): Likewise.
6132 (sanitize_attrs_match_for_inline_p): New function.
6134 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6136 * ipa-prop.c (detect_type_change): Shor circuit testing of
6137 type changes on THIS pointer.
6139 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6142 * config/pa/pa.c (legitimize_tls_address): Return original address
6143 if not passed a SYMBOL_REF rtx.
6144 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6146 (pa_emit_move_sequence): Simplify TLS source operands.
6147 (pa_legitimate_constant_p): Reject all TLS constants.
6148 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6149 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6151 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6153 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6154 groups when we know they are controlled by LTO.
6155 * varasm.c (default_binds_local_p_1): If object is in other partition,
6156 it will be resolved locally.
6158 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6160 * config/host-linux.c (linux_gt_pch_use_address): Don't
6161 use SSIZE_MAX because it is not always defined.
6163 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6166 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6167 threshold for pseudo splitting.
6168 (update_ebb_live_info): Process call argument hard registers and
6169 hard registers from insn definition too.
6170 (max_small_class_regs_num): New constant.
6171 (inherit_in_ebb): Update live hard regs through EBBs. Update
6172 reloads_num only for small register classes. Don't split for
6175 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6178 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6181 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6183 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6186 2014-02-04 Marek Polacek <polacek@redhat.com>
6188 * gdbinit.in (pel): Define.
6190 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6192 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6195 2014-02-04 Richard Biener <rguenther@suse.de>
6198 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6199 in function context local.
6200 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6201 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6202 similar to LTO_imported_decl_ref.
6204 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6206 PR tree-optimization/60002
6207 * cgraphclones.c (build_function_decl_skip_args): Clear
6210 PR tree-optimization/60023
6211 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6212 false to gsi_replace.
6213 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6214 has been in some EH region and vec_stmt could throw, add
6215 vec_stmt into the same EH region.
6216 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6217 has no lhs, ignore it.
6218 * internal-fn.c (expand_MASK_LOAD): Likewise.
6221 * tree-inline.c (copy_forbidden): Fail for
6222 __attribute__((optimize (0))) functions.
6225 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6226 is set, copy one less argument.
6227 (expand_simd_clones): Don't subtract clone_info->inbranch
6228 from simd_clone_struct_alloc argument.
6230 PR rtl-optimization/57915
6231 * recog.c (simplify_while_replacing): If all unary/binary/relational
6232 operation arguments are constant, attempt to simplify those.
6235 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6236 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6238 2014-02-04 Richard Biener <rguenther@suse.de>
6240 PR tree-optimization/60012
6241 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6242 TBAA disambiguation to all DDRs.
6244 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6247 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6248 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6250 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6253 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6255 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6257 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6258 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6260 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6263 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6264 to figure out targets of polymorphic calls with known decl.
6265 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6266 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6267 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6268 (get_polymorphic_call_info): ... here.
6269 (get_polymorphic_call_info_from_invariant): New function.
6271 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6273 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6274 lookup via vtable pointer; check for type consistency
6275 and turn inconsitent facts into UNREACHABLE.
6276 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6277 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6278 type inconsistent querries; return UNREACHABLE instead.
6280 2014-02-03 Richard Henderson <rth@twiddle.net>
6283 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6284 already processed this node.
6285 (normalize_one_pred_1): Pass along mark_set.
6286 (normalize_one_pred): Create and destroy a pointer_set_t.
6287 (normalize_one_pred_chain): Likewise.
6289 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6291 PR gcov-profile/58602
6292 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6294 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6297 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6298 -fno-devirtualize; try to devirtualize by the knowledge of
6299 virtual table pointer given by aggregate propagation.
6300 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6301 (ipa_print_node_jump_functions): Dump also offset that
6302 is relevant for polymorphic calls.
6303 (determine_known_aggregate_parts): Add arg_type parameter; use it
6304 instead of determining the type from pointer type.
6305 (ipa_compute_jump_functions_for_edge): Update call of
6306 determine_known_aggregate_parts.
6307 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6308 (gimple_get_virt_method_for_binfo): ... here; simplify using
6309 vtable_pointer_value_to_vtable.
6310 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6311 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6312 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6313 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6315 (vtable_pointer_value_to_binfo): ... here.
6316 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6318 2014-02-03 Teresa Johnson <tejohnson@google.com>
6320 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6321 redef of outer loop index variable.
6323 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6327 * doc/extend.texi (Function Attributes): Typo.
6329 2014-02-03 Cong Hou <congh@google.com>
6331 PR tree-optimization/60000
6332 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6333 if the vectorized statement is a store. A store statement can only
6334 appear at the end of pattern statements.
6336 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6338 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6339 (ix86_option_override_internal): Default long double to 64-bit for
6340 32-bit Bionic and to 128-bit for 64-bit Bionic.
6342 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6343 TARGET_LONG_DOUBLE_128 is true.
6344 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6346 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6347 (mlong-double-64): Negate -mlong-double-128.
6348 (mlong-double-128): New option.
6350 * config/i386/i386-c.c (ix86_target_macros): Define
6351 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6353 * doc/invoke.texi: Document -mlong-double-128.
6355 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6357 PR rtl-optimization/60024
6358 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6360 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6362 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6364 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6366 PR rtl-optimization/57662
6367 * sel-sched.c (code_motion_path_driver): Do not mark already not
6368 existing blocks in the visiting bitmap.
6370 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6372 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6373 on the insn being emitted.
6375 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6376 Will Deacon <will.deacon@arm.com>
6378 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6380 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6382 * config/arm/arm-tables.opt: Regenerate.
6384 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6386 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6387 for vector types other than V16QImode.
6388 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6389 define_expand, and call altivec_expand_vec_perm_le when producing
6390 code with little endian element order.
6391 (*altivec_vperm_<mode>_internal): New insn having previous
6392 behavior of altivec_vperm_<mode>.
6393 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6394 altivec_expand_vec_perm_le when producing code with little endian
6396 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6397 behavior of altivec_vperm_<mode>_uns.
6399 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6401 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6402 (altivec_vsumsws): Add handling for -maltivec=be with a little
6404 (altivec_vsumsws_direct): New.
6405 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6406 gen_altivec_vsumsws.
6408 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6410 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6411 vtable_pointer_value_to_binfo): New functions.
6412 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6413 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6415 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6417 * config/nios2/nios2.md (load_got_register): Initialize GOT
6418 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6419 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6421 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6423 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6424 preserverd by passthrough, do not propagate the type.
6426 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6428 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6429 (mips_atomic_assign_expand_fenv): New function.
6430 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6432 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6434 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6435 (__builtin_mips_set_fcsr): Likewise.
6436 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6437 MIPS_USI_FTYPE_VOID.
6438 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6439 (mips16_expand_set_fcsr): Likewise.
6440 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6441 (mips16_set_fcsr_stub): Likewise.
6442 (mips16_get_fcsr_one_only_stub): New class.
6443 (mips16_set_fcsr_one_only_stub): Likewise.
6444 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6445 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6446 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6447 (hard_float): New availability predicate.
6448 (mips_builtins): Add get_fcsr and set_fcsr.
6449 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6450 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6451 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6452 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6453 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6456 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6458 * config/mips/mips.c (mips_one_only_stub): New class.
6459 (mips_need_mips16_rdhwr_p): Replace with...
6460 (mips16_rdhwr_stub): ...this new variable.
6461 (mips16_stub_call_address): New function.
6462 (mips16_rdhwr_one_only_stub): New class.
6463 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6464 (mips_output_mips16_rdhwr): Delete.
6465 (mips_finish_stub): New function.
6466 (mips_code_end): Use it to handle rdhwr stubs.
6468 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6471 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6472 when calculating size of integer atomic types.
6474 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6476 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6478 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6480 PR tree-optimization/60003
6481 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6482 * profile.c (branch_prob): Use gimple_call_builtin_p
6483 to check for BUILT_IN_SETJMP_RECEIVER.
6484 * tree-inline.c (copy_bb): Call notice_special_calls.
6486 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6489 * lra-constraints.c (process_alt_operands): Update reload_sum only
6492 2014-01-31 Richard Henderson <rth@redhat.com>
6495 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6496 until after else_eh is processed.
6498 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6500 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6501 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6502 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6503 in smmintrin.h, remove them.
6504 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6505 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6506 * config/i386/i386.md (ROUND_SAE): Fix value.
6507 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6508 (const48_operand): New.
6509 * config/i386/subst.md (round), (round_expand): Use
6510 const_4_or_8_to_11_operand.
6511 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6513 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6515 * config/i386/constraints.md (Yk): Swap meaning with k.
6516 * config/i386/i386.md (movhi_internal): Change Yk to k.
6517 (movqi_internal): Ditto.
6518 (*k<logic><mode>): Ditto.
6521 (kandn<mode>): Ditto.
6522 (*<code>hi_1): Ditto.
6523 (*<code>qi_1): Ditto.
6524 (kxnor<mode>): Ditto.
6525 (kortestzhi): Ditto.
6526 (kortestchi): Ditto.
6528 (*one_cmplhi2_1): Ditto.
6529 (*one_cmplqi2_1): Ditto.
6530 * config/i386/sse.md (): Change k to Yk.
6531 (avx512f_load<mode>_mask): Ditto.
6532 (avx512f_blendm<mode>): Ditto.
6533 (avx512f_store<mode>_mask): Ditto.
6534 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6535 (avx512f_storedqu<mode>_mask): Ditto.
6536 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6538 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6539 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6540 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6541 (avx512f_maskcmp<mode>3): Ditto.
6542 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6543 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6544 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6545 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6546 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6547 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6548 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6549 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6550 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6551 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6552 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6553 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6554 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6555 (vec_extract_lo_<mode>_maskm): Ditto.
6556 (vec_extract_hi_<mode>_maskm): Ditto.
6557 (avx512f_vternlog<mode>_mask): Ditto.
6558 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6559 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6560 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6561 (avx512f_<code>v8div16qi2_mask): Ditto.
6562 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6563 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6564 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6565 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6566 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6567 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6568 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6569 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6570 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6571 (avx512cd_maskb_vec_dupv8di): Ditto.
6572 (avx512cd_maskw_vec_dupv16si): Ditto.
6573 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6574 (avx512f_vpermi2var<mode>3_mask): Ditto.
6575 (avx512f_vpermi2var<mode>3_mask): Ditto.
6576 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6577 (*avx512f_gathersi<mode>): Ditto.
6578 (*avx512f_gathersi<mode>_2): Ditto.
6579 (*avx512f_gatherdi<mode>): Ditto.
6580 (*avx512f_gatherdi<mode>_2): Ditto.
6581 (*avx512f_scattersi<mode>): Ditto.
6582 (*avx512f_scatterdi<mode>): Ditto.
6583 (avx512f_compress<mode>_mask): Ditto.
6584 (avx512f_compressstore<mode>_mask): Ditto.
6585 (avx512f_expand<mode>_mask): Ditto.
6586 * config/i386/subst.md (mask): Change k to Yk.
6587 (mask_scalar_merge): Ditto.
6590 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6592 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6594 2014-01-31 Richard Biener <rguenther@suse.de>
6597 * builtins.c (fold_builtin_memory_op): Make sure to not
6598 use a floating-point mode or a boolean or enumeral type for
6601 2014-01-30 DJ Delorie <dj@redhat.com>
6603 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6604 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6605 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6606 whenever main() has an epilogue.
6608 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6610 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6611 unused variable "field".
6612 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6613 (vsx_mergeh_<mode>): Likewise.
6614 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6615 (altivec_vmrghh): Likewise.
6616 (altivec_vmrghw): Likewise.
6617 (altivec_vmrglb): Likewise.
6618 (altivec_vmrglh): Likewise.
6619 (altivec_vmrglw): Likewise.
6620 (altivec_vspltb): Add missing uses.
6621 (altivec_vsplth): Likewise.
6622 (altivec_vspltw): Likewise.
6623 (altivec_vspltsf): Likewise.
6625 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6628 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6629 frame related instructions.
6631 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6633 PR rtl-optimization/59959
6634 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6635 any reload of register whose subreg is invalid.
6637 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6639 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6640 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6641 Add missing return type - void.
6643 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6645 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6646 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6647 remove element index adjustment for endian (now handled in vsx.md
6649 (altivec_expand_vec_perm_const): Use
6650 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6651 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6652 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6653 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6654 define_expand and a new define_insn *altivec_vspltb_internal;
6655 adjust for -maltivec=be on a little endian target.
6656 (altivec_vspltb_direct): New.
6657 (altivec_vsplth): Divide into a define_expand and a new
6658 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6659 little endian target.
6660 (altivec_vsplth_direct): New.
6661 (altivec_vspltw): Divide into a define_expand and a new
6662 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6663 little endian target.
6664 (altivec_vspltw_direct): New.
6665 (altivec_vspltsf): Divide into a define_expand and a new
6666 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6667 a little endian target.
6669 2014-01-30 Richard Biener <rguenther@suse.de>
6671 PR tree-optimization/59993
6672 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6673 can propagate form the earlier stmt and avoid the transform
6674 when the intermediate result is needed.
6676 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6678 * README.Portability: Fix typo.
6680 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6682 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6683 comparison_operator with ordered_comparison_operator.
6685 2014-01-30 Nick Clifton <nickc@redhat.com>
6687 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6688 Rename to mn10300_store_multiple_regs.
6689 * config/mn10300/mn10300.c: Likewise.
6690 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6691 store_multiple_regs.
6692 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6693 Call mn10300_store_multiple_regs.
6695 2014-01-30 Nick Clifton <nickc@redhat.com>
6696 DJ Delorie <dj@redhat.com>
6698 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6699 %fp 2 to keep registers after it properly word-aligned.
6700 (rl78_alloc_physical_registers_umul): Handle the case where both
6701 input operands are the same.
6703 2014-01-30 Richard Biener <rguenther@suse.de>
6705 PR tree-optimization/59903
6706 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6709 2014-01-30 Jason Merrill <jason@redhat.com>
6712 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6715 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6717 2014-01-30 Richard Biener <rguenther@suse.de>
6719 PR tree-optimization/59951
6720 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6722 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6725 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6726 SFmode to DFmode case.
6728 2014-01-29 DJ Delorie <dj@redhat.com>
6730 * config/msp430/msp430.opt (-minrt): New.
6731 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6733 (ENDFILE_SPEC): Likewise.
6735 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6737 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6738 (estimate_function_body_sizes): Use it.
6740 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6743 * dwarf2out.c (is_cxx_auto): New.
6744 (is_base_type): Use it.
6745 (gen_type_die_with_usage): Likewise.
6747 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6749 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6750 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6751 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6752 -maltivec=be with LE targets.
6753 (vsx_mergeh_<mode>): Likewise.
6754 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6755 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6756 (altivec_vmrghb): Replace with define_expand and new
6757 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6758 (altivec_vmrghb_direct): New define_insn.
6759 (altivec_vmrghh): Replace with define_expand and new
6760 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6761 (altivec_vmrghh_direct): New define_insn.
6762 (altivec_vmrghw): Replace with define_expand and new
6763 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6764 (altivec_vmrghw_direct): New define_insn.
6765 (*altivec_vmrghsf): Adjust for endianness.
6766 (altivec_vmrglb): Replace with define_expand and new
6767 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6768 (altivec_vmrglb_direct): New define_insn.
6769 (altivec_vmrglh): Replace with define_expand and new
6770 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6771 (altivec_vmrglh_direct): New define_insn.
6772 (altivec_vmrglw): Replace with define_expand and new
6773 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6774 (altivec_vmrglw_direct): New define_insn.
6775 (*altivec_vmrglsf): Adjust for endianness.
6776 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6777 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6778 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6779 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6780 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6781 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6782 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6783 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6785 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6787 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6788 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6791 2014-01-29 Richard Biener <rguenther@suse.de>
6793 PR tree-optimization/58742
6794 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6795 associate_pointerplus_align.
6796 (associate_pointerplus_diff): New function.
6797 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6798 and associate_pointerplus_diff.
6800 2014-01-29 Richard Biener <rguenther@suse.de>
6802 * lto-streamer.h (LTO_major_version): Bump to 3.
6803 (LTO_minor_version): Reset to 0.
6805 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6807 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6808 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6809 (arm_file_start): Generate correct asm header for armv7ve.
6810 * config/arm/bpabi.h: Add multilib support for armv7ve.
6811 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6812 and cortex-a15 to armv7ve.
6813 * config/arm/t-aprofile: Add multilib support for armv7ve.
6814 * doc/invoke.texi: Document -march=armv7ve.
6816 2014-01-29 Richard Biener <rguenther@suse.de>
6818 PR tree-optimization/58742
6819 * tree-ssa-forwprop.c (associate_plusminus): Return true
6820 if we changed sth, defer EH cleanup to ...
6821 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6822 (simplify_mult): New function.
6824 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6827 PR tree-optimization/59920
6828 * tree.c (build_common_builtin_nodes): Remove
6829 __builtin_setjmp_dispatcher initialization.
6830 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6831 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6832 instead of gsi_after_labels + manually skipping debug stmts.
6833 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6834 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6835 * tree-inline.c (copy_edges_for_bb): Remove
6836 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6837 argument. Ignore computed_goto_p stmts. Don't call
6838 make_abnormal_goto_edges. If a call might need abnormal edges
6839 for non-local gotos, see if it already has an edge to
6840 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6841 with true argument, don't do anything then, otherwise add
6842 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6843 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6845 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6846 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6847 (lower_stmt): Don't set data->calls_builtin_setjmp.
6848 (lower_builtin_setjmp): Adjust comment.
6849 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6850 * tree-cfg.c (found_computed_goto): Remove.
6851 (factor_computed_gotos): Remove.
6852 (make_goto_expr_edges): Return bool, true for computed gotos.
6853 Don't call make_abnormal_goto_edges.
6854 (build_gimple_cfg): Don't set found_computed_goto, don't call
6855 factor_computed_gotos.
6856 (computed_goto_p): No longer static.
6857 (make_blocks): Don't set found_computed_goto.
6858 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6859 (make_edges): If make_goto_expr_edges returns true, push bb
6860 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6861 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6862 vector. Record mapping between bbs and OpenMP regions if there
6863 are any, adjust make_gimple_omp_edges caller. Call
6864 handle_abnormal_edges.
6865 (make_abnormal_goto_edges): Remove.
6866 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6867 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6868 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6869 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6870 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6871 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6872 filling *region also set *region_idx to (*region)->entry->index.
6875 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6876 For REGs set ORIGINAL_REGNO.
6878 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6880 * doc/md.texi: Mention that a target shouldn't implement
6881 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6884 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6886 PR tree-optimization/59594
6887 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6888 a copy of the datarefs vector rather than the vector itself.
6890 2014-01-28 Jason Merrill <jason@redhat.com>
6893 * dwarf2out.c (auto_die): New static.
6894 (gen_type_die_with_usage): Handle C++1y 'auto'.
6895 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6898 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6901 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6902 (SPEC_X32): Likewise.
6903 (SPEC_64): Likewise.
6904 * config/i386/i386.c (ix86_option_override_internal): Turn off
6905 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6907 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6908 * config/i386/i386.h (TARGET_16BIT): New macro.
6909 (TARGET_16BIT_P): Likewise.
6910 * config/i386/i386.opt: Add m16.
6911 * doc/invoke.texi: Document -m16.
6913 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6915 PR preprocessor/59935
6916 * input.c (location_get_source_line): Bail out on when line number
6917 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6919 2014-01-28 Richard Biener <rguenther@suse.de>
6921 PR tree-optimization/58742
6922 * tree-ssa-forwprop.c (associate_plusminus): Handle
6923 pointer subtraction of the form (T)(P + A) - (T)P.
6925 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6927 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6930 2014-01-28 Richard Biener <rguenther@suse.de>
6933 2014-01-28 Richard Biener <rguenther@suse.de>
6935 PR rtl-optimization/45364
6936 PR rtl-optimization/59890
6937 * var-tracking.c (local_get_addr_clear_given_value): Handle
6938 already cleared slot.
6939 (val_reset): Handle not allocated local_get_addr_cache.
6940 (vt_find_locations): Use post-order on the inverted CFG.
6942 2014-01-28 Richard Biener <rguenther@suse.de>
6944 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6946 2014-01-28 Richard Biener <rguenther@suse.de>
6948 PR rtl-optimization/45364
6949 PR rtl-optimization/59890
6950 * var-tracking.c (local_get_addr_clear_given_value): Handle
6951 already cleared slot.
6952 (val_reset): Handle not allocated local_get_addr_cache.
6953 (vt_find_locations): Use post-order on the inverted CFG.
6955 2014-01-28 Alan Modra <amodra@gmail.com>
6957 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6958 * configure.ac <recursive call for build != host>: Define
6959 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6960 and LD_FOR_BUILD too.
6961 * configure: Regenerate.
6963 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6965 * config/i386/i386.c (get_builtin_code_for_version): Separate
6966 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6967 Broadwell from Haswell.
6969 2014-01-27 Steve Ellcey <sellcey@mips.com>
6971 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6972 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6973 * config/mips/mips.c (mips_option_override): Change setting
6975 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6976 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6977 Change from Mask to Var.
6979 2014-01-27 Jeff Law <law@redhat.com>
6981 * ipa-inline.c (inline_small_functions): Fix typo.
6983 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6985 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6986 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6987 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6988 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6989 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6990 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6991 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6992 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6993 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6994 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6995 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6996 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6997 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6998 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6999 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
7000 (_mm512_storeu_epi64): Ditto.
7001 (_mm512_cmpge_epi32_mask): Ditto.
7002 (_mm512_cmpge_epu32_mask): Ditto.
7003 (_mm512_cmpge_epi64_mask): Ditto.
7004 (_mm512_cmpge_epu64_mask): Ditto.
7005 (_mm512_cmple_epi32_mask): Ditto.
7006 (_mm512_cmple_epu32_mask): Ditto.
7007 (_mm512_cmple_epi64_mask): Ditto.
7008 (_mm512_cmple_epu64_mask): Ditto.
7009 (_mm512_cmplt_epi32_mask): Ditto.
7010 (_mm512_cmplt_epu32_mask): Ditto.
7011 (_mm512_cmplt_epi64_mask): Ditto.
7012 (_mm512_cmplt_epu64_mask): Ditto.
7013 (_mm512_cmpneq_epi32_mask): Ditto.
7014 (_mm512_cmpneq_epu32_mask): Ditto.
7015 (_mm512_cmpneq_epi64_mask): Ditto.
7016 (_mm512_cmpneq_epu64_mask): Ditto.
7017 (_mm512_expand_pd): Ditto.
7018 (_mm512_expand_ps): Ditto.
7019 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7020 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7021 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7022 * config/i386/i386.c (ix86_builtins): Add
7023 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7024 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7025 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7026 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7027 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7028 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7029 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7030 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7031 IX86_BUILTIN_PMOVUSQW512_MEM.
7032 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7033 __builtin_ia32_pmovsqd512mem_mask,
7034 __builtin_ia32_pmovqd512mem_mask,
7035 __builtin_ia32_pmovusqw512mem_mask,
7036 __builtin_ia32_pmovsqw512mem_mask,
7037 __builtin_ia32_pmovqw512mem_mask,
7038 __builtin_ia32_pmovusdw512mem_mask,
7039 __builtin_ia32_pmovsdw512mem_mask,
7040 __builtin_ia32_pmovdw512mem_mask,
7041 __builtin_ia32_pmovqb512mem_mask,
7042 __builtin_ia32_pmovusqb512mem_mask,
7043 __builtin_ia32_pmovsqb512mem_mask,
7044 __builtin_ia32_pmovusdb512mem_mask,
7045 __builtin_ia32_pmovsdb512mem_mask,
7046 __builtin_ia32_pmovdb512mem_mask.
7047 (bdesc_args): Add __builtin_ia32_expanddf512,
7048 __builtin_ia32_expandsf512.
7049 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7050 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7051 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7052 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7053 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7054 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7055 (avx512f_<code>v8div16qi2_mask_store): This.
7056 (avx512f_expand<mode>): New.
7058 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7060 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7062 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7063 (_mm512_prefetch_i32scatter_pd): Ditto.
7064 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7065 (_mm512_prefetch_i64scatter_pd): Ditto.
7066 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7067 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7068 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7069 (_mm512_prefetch_i32scatter_ps): Ditto.
7070 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7071 (_mm512_prefetch_i64scatter_ps): Ditto.
7072 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7073 * config/i386/i386-builtin-types.def: Define
7074 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7075 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7076 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7077 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7078 IX86_BUILTIN_SCATTERPFQPD.
7079 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7080 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7081 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7082 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7083 __builtin_ia32_scatterpfqps.
7084 (ix86_expand_builtin): Expand new built-ins.
7085 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
7086 fix memory access data type.
7087 (*avx512pf_gatherpf<mode>_mask): Ditto.
7088 (*avx512pf_gatherpf<mode>): Ditto.
7089 (avx512pf_scatterpf<mode>): Ditto.
7090 (*avx512pf_scatterpf<mode>_mask): Ditto.
7091 (*avx512pf_scatterpf<mode>): Ditto.
7092 (GATHER_SCATTER_SF_MEM_MODE): New.
7093 (avx512pf_gatherpf<mode>df): Ditto.
7094 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7095 (*avx512pf_scatterpf<mode>df): Ditto.
7097 2014-01-27 Jakub Jelinek <jakub@redhat.com>
7100 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7101 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7104 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7106 * common/config/arm/arm-common.c
7107 (arm_rewrite_mcpu): Handle multiple names.
7109 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7111 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7113 * gimple-builder.h (create_gimple_tmp): Delete.
7115 2014-01-27 Christian Bruel <christian.bruel@st.com>
7117 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7120 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7122 * config/pa/pa.md (call): Generate indirect long calls to non-local
7123 functions when outputing 32-bit code.
7124 (call_value): Likewise except for special call to buggy powf function.
7126 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7127 portable runtime and PIC indirect calls.
7128 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7129 and PIC call sequences. Use ldo instead of blr to set return register
7130 in PIC call sequence.
7132 2014-01-25 Walter Lee <walt@tilera.com>
7134 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7135 avoid clobbering a live register.
7137 2014-01-25 Walter Lee <walt@tilera.com>
7139 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7140 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7141 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7142 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7144 2014-01-25 Walter Lee <walt@tilera.com>
7146 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7147 arguments on even registers.
7148 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7150 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7151 (BIGGEST_ALIGNMENT): Ditto.
7152 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7154 2014-01-25 Walter Lee <walt@tilera.com>
7156 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7157 insns before bundling.
7158 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7160 2014-01-25 Walter Lee <walt@tilera.com>
7162 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7163 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7164 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7166 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7168 * config/mips/constraints.md (kl): Delete.
7169 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7170 define expands, using...
7171 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7172 instructions for MIPS16.
7173 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7174 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7176 2014-01-25 Walter Lee <walt@tilera.com>
7178 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7182 2014-01-25 Walter Lee <walt@tilera.com>
7184 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7185 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7187 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7189 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7192 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7194 * print-rtl.c (in_call_function_usage): New var.
7195 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7196 EXPR_LIST mode as mode and not as reg note name.
7199 * cfgloopmanip.c (copy_loop_info): If
7200 loop->warned_aggressive_loop_optimizations, make sure
7201 the flag is set in target loop too.
7203 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7205 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7207 * builtins.def: Likewise.
7208 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7209 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7210 * ira.c (ira_setup_eliminable_regset): Likewise.
7211 * omp-low.c (gate_expand_omp): Likewise.
7212 (execute_lower_omp): Likewise.
7213 (diagnose_sb_0): Likewise.
7214 (gate_diagnose_omp_blocks): Likewise.
7215 (simd_clone_clauses_extract): Likewise.
7218 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7220 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7221 correction for little endian...
7222 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7225 2014-01-24 Jeff Law <law@redhat.com>
7227 PR tree-optimization/59919
7228 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7229 for non-returning calls.
7231 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7233 * common/config/aarch64/aarch64-common.c
7234 (aarch64_rewrite_mcpu): Handle multiple names.
7235 * config/aarch64/aarch64.h
7236 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7238 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7240 * input.c (add_file_to_cache_tab): Handle the case where fopen
7243 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7246 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7247 from push operand if code of push isn't PRE_DEC.
7249 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7252 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7253 -mquad-memory-atomic. Update -mquad-memory documentation to say
7254 it is only used for non-atomic loads/stores.
7256 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7257 -mquad-memory or -mquad-memory-atomic switches.
7259 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7260 -mquad-memory-atomic to ISA 2.07 support.
7262 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7263 to separate support of normal quad word memory operations (ldq, stq)
7264 from the atomic quad word memory operations.
7266 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7267 support to separate non-atomic quad word operations from atomic
7268 quad word operations. Disable non-atomic quad word operations in
7269 little endian mode so that we don't have to swap words after the
7270 load and before the store.
7271 (quad_load_store_p): Add comment about atomic quad word support.
7272 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7273 options printed with -mdebug=reg.
7275 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7276 -mquad-memory-atomic as the test for whether we have quad word
7277 atomic instructions.
7278 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7279 or -mp8-vector are used, allow byte/half-word atomic operations.
7281 * config/rs6000/sync.md (load_lockedti): Insure that the address
7282 is a proper indexed or indirect address for the lqarx instruction.
7283 On little endian systems, swap the hi/lo registers after the lqarx
7285 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7286 insure the address is valid for the lqarx instruction.
7287 (store_conditionalti): Insure that the address is a proper indexed
7288 or indirect address for the stqcrx. instruction. On little endian
7289 systems, swap the hi/lo registers before doing the stqcrx.
7291 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7292 insure the address is valid for the stqcrx. instruction.
7294 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7295 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7296 type of quad memory support is available.
7298 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7301 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7302 there is a danger of looping.
7304 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7306 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7307 force flag_ira_loop_pressure if set via command line.
7309 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7311 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7312 (ashr_simd): New builtin handling DI mode.
7313 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7314 (aarch64_sshr_simddi): New match pattern.
7315 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7316 (vshrd_n_s64): Likewise.
7317 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7319 2014-01-23 Nick Clifton <nickc@redhat.com>
7321 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7322 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7323 favour of mcu specific scripts.
7324 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7327 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7328 Alex Velenko <Alex.Velenko@arm.com>
7330 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7331 (vaddv_s16): Likewise.
7332 (vaddv_s32): Likewise.
7333 (vaddv_u8): Likewise.
7334 (vaddv_u16): Likewise.
7335 (vaddv_u32): Likewise.
7336 (vaddvq_s8): Likewise.
7337 (vaddvq_s16): Likewise.
7338 (vaddvq_s32): Likewise.
7339 (vaddvq_s64): Likewise.
7340 (vaddvq_u8): Likewise.
7341 (vaddvq_u16): Likewise.
7342 (vaddvq_u32): Likewise.
7343 (vaddvq_u64): Likewise.
7344 (vaddv_f32): Likewise.
7345 (vaddvq_f32): Likewise.
7346 (vaddvq_f64): Likewise.
7347 (vmaxv_f32): Likewise.
7348 (vmaxv_s8): Likewise.
7349 (vmaxv_s16): Likewise.
7350 (vmaxv_s32): Likewise.
7351 (vmaxv_u8): Likewise.
7352 (vmaxv_u16): Likewise.
7353 (vmaxv_u32): Likewise.
7354 (vmaxvq_f32): Likewise.
7355 (vmaxvq_f64): Likewise.
7356 (vmaxvq_s8): Likewise.
7357 (vmaxvq_s16): Likewise.
7358 (vmaxvq_s32): Likewise.
7359 (vmaxvq_u8): Likewise.
7360 (vmaxvq_u16): Likewise.
7361 (vmaxvq_u32): Likewise.
7362 (vmaxnmv_f32): Likewise.
7363 (vmaxnmvq_f32): Likewise.
7364 (vmaxnmvq_f64): Likewise.
7365 (vminv_f32): Likewise.
7366 (vminv_s8): Likewise.
7367 (vminv_s16): Likewise.
7368 (vminv_s32): Likewise.
7369 (vminv_u8): Likewise.
7370 (vminv_u16): Likewise.
7371 (vminv_u32): Likewise.
7372 (vminvq_f32): Likewise.
7373 (vminvq_f64): Likewise.
7374 (vminvq_s8): Likewise.
7375 (vminvq_s16): Likewise.
7376 (vminvq_s32): Likewise.
7377 (vminvq_u8): Likewise.
7378 (vminvq_u16): Likewise.
7379 (vminvq_u32): Likewise.
7380 (vminnmv_f32): Likewise.
7381 (vminnmvq_f32): Likewise.
7382 (vminnmvq_f64): Likewise.
7384 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7386 * config/aarch64/aarch64-simd.md
7387 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7388 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7389 (*aarch64_mul3_elt<mode>): Likewise.
7390 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7391 (*aarch64_mul3_elt_to_64v2df): Likewise.
7392 (*aarch64_mla_elt<mode>): Likewise.
7393 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7394 (*aarch64_mls_elt<mode>): Likewise.
7395 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7396 (*aarch64_fma4_elt<mode>): Likewise.
7397 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7398 (*aarch64_fma4_elt_to_64v2df): Likewise.
7399 (*aarch64_fnma4_elt<mode>): Likewise.
7400 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7401 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7402 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7403 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7404 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7405 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7406 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7407 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7408 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7410 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7412 * config/aarch64/aarch64-simd.md
7413 (aarch64_be_checked_get_lane<mode>): New define_expand.
7414 * config/aarch64/aarch64-simd-builtins.def
7415 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7416 New builtin definition.
7417 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7418 Use new safe be builtin.
7420 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7422 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7424 (aarch64_be_st1<mode>): Likewise.
7425 (aarch_ld1<VALL:mode>): Define_expand modified.
7426 (aarch_st1<VALL:mode>): Likewise.
7427 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7428 (UNSPEC_ST1): Likewise.
7430 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7432 * config/microblaze/microblaze.md: Add trap insn and attribute
7434 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7436 PR preprocessor/58580
7437 * input.h (location_get_source_line): Take an additional line_size
7439 (void diagnostics_file_cache_fini): Declare new function.
7440 * input.c (struct fcache): New type.
7441 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7442 New static constants.
7443 (diagnostic_file_cache_init, total_lines_num)
7444 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7445 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7446 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7447 (get_next_line, read_next_line, goto_next_line, read_line_num):
7448 New static function definitions.
7449 (diagnostic_file_cache_fini): New function.
7450 (location_get_source_line): Take an additional output line_len
7451 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7453 * diagnostic.c (diagnostic_finish): Call
7454 diagnostic_file_cache_fini.
7455 (adjust_line): Take an additional input parameter for the length
7456 of the line, rather than calculating it with strlen.
7457 (diagnostic_show_locus): Adjust the use of
7458 location_get_source_line and adjust_line with respect to their new
7459 signature. While displaying a line now, do not stop at the first
7460 null byte. Rather, display the zero byte as a space and keep
7461 going until we reach the size of the line.
7462 * Makefile.in: Add vec.o to OBJS-libcommon
7464 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7465 Ilya Tocar <ilya.tocar@intel.com>
7467 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7468 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7469 (__builtin_ia32_kmov16): Ditto.
7470 * config/i386/i386.md (UNSPEC_KMOV): New.
7473 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7475 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7476 (_mm512_storeu_si512): Ditto.
7478 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7481 * rtl.h (get_referenced_operands): Declare.
7482 * recog.c (get_referenced_operands): New function.
7483 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7484 operands have been referenced when recording LO_SUM references.
7486 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7488 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7490 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7492 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7493 Enable for generic and recent AMD targets.
7495 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7497 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7498 ARG_SIZE note when adjustment was eliminated.
7500 2014-01-22 Jeff Law <law@redhat.com>
7502 PR tree-optimization/59597
7503 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7504 in file. Accept new argument REGISTERING and use it to modify
7505 dump output appropriately.
7506 (register_jump_thread): Corresponding changes.
7507 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7508 thread paths involving joiner blocks. Add code to dump cancelled
7509 jump threading paths.
7511 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7513 PR rtl-optimization/59477
7514 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7515 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7517 2014-01-22 Tom Tromey <tromey@redhat.com>
7519 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7521 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7523 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7525 PR rtl-optimization/59896
7526 * lra-constraints.c (process_alt_operands): Check unused note for
7527 matched operands of insn with no output reloads.
7529 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7531 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7532 (mips_move_from_gpr_cost): Likewise.
7534 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7536 PR rtl-optimization/59858
7537 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7538 ira_class_hard_regs_num.
7539 (process_alt_operands): Increase reject for dying matched operand.
7541 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7544 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7545 smaller than size, perform several stores or loads and stores
7546 at dst + count - size to store or copy all of size bytes, rather
7547 than just last modesize bytes.
7549 2014-01-20 DJ Delorie <dj@redhat.com>
7551 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7552 that CLOBBERs are REGs before propogating their values.
7554 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7557 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7558 (cgraph_inline_failed_type): New function.
7559 * cgraph.h (DEFCIFCODE): Add type.
7560 (cgraph_inline_failed_type_t): New enum.
7561 (cgraph_inline_failed_type): New prototype.
7562 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7563 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7564 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7565 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7566 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7567 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7568 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7569 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7570 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7571 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7572 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7573 OPTIMIZATION_MISMATCH.
7574 * tree-inline.c (expand_call_inline): Emit errors during
7575 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7577 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7580 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7581 mode attribute in insn output.
7583 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7585 * output.h (output_constant): Delete.
7586 * varasm.c (output_constant): Make private.
7588 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7590 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7592 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7595 * tree.h (fold_builtin_strcat): New prototype.
7596 * builtins.c (fold_builtin_strcat): No longer static. Add len
7597 argument, if non-NULL, don't call c_strlen. Optimize
7598 directly into __builtin_memcpy instead of __builtin_strcpy.
7599 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7600 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7602 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7604 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7605 for SImode_address_operand operands, having only a REG argument.
7607 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7609 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7610 loader name using mbig-endian.
7611 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7613 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7615 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7619 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7621 * config/aarch64/aarch64-protos.h
7622 (aarch64_cannot_change_mode_class_ptr): Declare.
7623 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7624 aarch64_cannot_change_mode_class_ptr): New.
7625 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7626 backend hook aarch64_cannot_change_mode_class.
7628 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7630 * common/config/aarch64/aarch64-common.c
7631 (aarch64_handle_option): Don't handle any option order logic here.
7632 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7633 selected_cpu, warn on architecture version mismatch.
7634 (aarch64_override_options): Fix parsing order for option strings.
7636 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7637 Iain Sandoe <iain@codesourcery.com>
7640 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7641 warning. Amend comment to reflect current functionality.
7643 2014-01-20 Richard Biener <rguenther@suse.de>
7646 * builtins.c (fold_builtin_strcat): Remove case better handled
7647 by tree-ssa-strlen.c.
7649 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7651 * config/aarch64/aarch64.opt
7652 (mcpu, march, mtune): Make case-insensitive.
7654 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7657 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7658 if operands[1] is a REG or ZERO_EXTEND of a REG.
7660 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7662 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7664 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7666 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7667 long non-pic millicode calls.
7669 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7671 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7673 2014-01-19 Kito Cheng <kito@0xlab.org>
7675 * builtins.c (expand_movstr): Check movstr expand done or fail.
7677 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7678 H.J. Lu <hongjiu.lu@intel.com>
7681 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7682 to DImode for zero-extended addresses.
7684 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7686 PR rtl-optimization/57763
7687 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7688 on the new indirect jump_insn and increment LABEL_NUSES (label).
7690 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7694 * config.gcc (x86_archs): New variable.
7695 (x86_64_archs): Likewise.
7696 (x86_cpus): Likewise.
7697 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7698 --with-arch/--with-cpu= options.
7699 Support --with-arch=/--with-cpu={nehalem,westmere,
7700 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7702 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7704 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7705 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7707 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7709 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7711 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7714 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7715 clear cpp_get_options (parse_in)->warn_unused_macros for
7716 ix86_target_macros_internal with cpp_define.
7718 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7720 * jump.c (delete_related_insns): Keep (use (insn))s.
7721 * reorg.c (redundant_insn): Check for barriers too.
7723 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7725 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7727 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7729 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7730 call to $$dyncall when TARGET_LONG_CALLS is true.
7732 2014-01-17 Jeff Law <law@redhat.com>
7734 * ree.c (combine_set_extension): Temporarily disable test for
7735 changing number of hard registers.
7737 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7740 * ipa-inline-analysis.c (inline_free_summary):
7741 Do not free summary of aliases.
7743 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7746 * gimplify.c (gimplify_expr): Use create_tmp_var
7747 instead of create_tmp_var_raw. If cond doesn't have
7748 integral type, don't add the IFN_ANNOTATE builtin at all.
7750 2014-01-17 Martin Jambor <mjambor@suse.cz>
7753 * ipa-cp.c (prev_edge_clone): New variable.
7754 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7755 Also resize prev_edge_clone vector.
7756 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7757 (ipcp_edge_removal_hook): New function.
7758 (ipcp_driver): Register ipcp_edge_removal_hook.
7760 2014-01-17 Andrew Pinski <apinski@cavium.com>
7761 Steve Ellcey <sellcey@mips.com>
7764 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7767 2014-01-17 Jeff Law <law@redhat.com>
7770 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7771 so that pass_ccp runs first.
7773 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7775 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7776 (ix86_adjust_cost): Use !TARGET_XXX.
7777 (do_reorder_for_imul): Likewise.
7778 (swap_top_of_ready_list): Likewise.
7779 (ix86_sched_reorder): Likewise.
7781 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7783 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7784 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7785 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7786 (intel_memset): New. Duplicate slm_memset.
7787 (intel_cost): New. Duplicate slm_cost.
7788 (m_INTEL): New macro.
7789 (processor_target_table): Add "intel".
7790 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7791 with PROCESSOR_INTEL for "intel".
7792 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7793 PROCESSOR_SILVERMONT.
7794 (ix86_issue_rate): Likewise.
7795 (ix86_adjust_cost): Likewise.
7796 (ia32_multipass_dfa_lookahead): Likewise.
7797 (swap_top_of_ready_list): Likewise.
7798 (ix86_sched_reorder): Likewise.
7799 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7800 instead of TARGET_OPT_AGU.
7801 * config/i386/i386.h (TARGET_INTEL): New.
7802 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7803 (processor_type): Add PROCESSOR_INTEL.
7804 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7805 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7807 2014-01-17 Marek Polacek <polacek@redhat.com>
7810 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7813 2014-01-17 Richard Biener <rguenther@suse.de>
7815 PR tree-optimization/46590
7816 * opts.c (default_options_table): Add entries for
7817 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7818 all enabled at -O1 but not for -Og.
7819 * common.opt (fbranch-count-reg): Remove Init(1).
7820 (fmove-loop-invariants): Likewise.
7821 (ftree-pta): Likewise.
7823 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7825 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7826 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7827 decls to at least the GCC 4.8 used alignments.
7830 * tree-nested.c (convert_nonlocal_reference_stmt,
7831 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7832 of GIMPLE_BIND stmts, adjust associated decls.
7834 2014-01-17 Richard Biener <rguenther@suse.de>
7836 PR tree-optimization/46590
7837 * vec.h (vec<>::bseach): New member function implementing
7838 binary search according to C89 bsearch.
7839 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7840 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7841 bitmap pointer again. Make accesses_in_loop a flat array.
7842 (mem_ref_obstack): New global.
7843 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7844 (mark_ref_stored): Likewise.
7845 (ref_indep_loop_p_2): Likewise.
7846 (set_ref_stored_in_loop): New helper function.
7847 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7848 (memref_free): Adjust.
7849 (record_mem_ref_loc): Simplify.
7850 (gather_mem_refs_stmt): Adjust.
7851 (sort_locs_in_loop_postorder_cmp): New function.
7852 (analyze_memory_references): Sort accesses_in_loop after
7853 loop postorder number.
7854 (find_ref_loc_in_loop_cmp): New function.
7855 (for_all_locs_in_loop): Find relevant cluster of locs in
7856 accesses_in_loop and iterate without recursion.
7857 (execute_sm): Avoid uninit warning.
7858 (struct ref_always_accessed): Simplify.
7859 (ref_always_accessed::operator ()): Likewise.
7860 (ref_always_accessed_p): Likewise.
7861 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7862 loop postorder numbers here.
7863 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7866 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7869 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7870 on decls for which assemble_alias has been called.
7872 2014-01-17 Nick Clifton <nickc@redhat.com>
7874 * config/msp430/msp430.opt: (mcpu): New option.
7875 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7876 (msp430_option_override): Parse target_cpu. If the MCU name
7877 matches a generic string, clear target_mcu.
7878 (msp430_attr): Allow numeric interrupt values up to 63.
7879 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7880 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7882 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7884 * config/msp430/msp430.md (popm): Use %J rather than %I.
7885 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7886 (addhi_cy_i): Use immediate_operand for operand 2.
7887 * doc/invoke.texi: Document -mcpu option.
7889 2014-01-17 Richard Biener <rguenther@suse.de>
7891 PR rtl-optimization/38518
7892 * df.h (df_analyze_loop): Declare.
7893 * df-core.c: Include cfgloop.h.
7894 (df_analyze_1): Split out main part of df_analyze.
7895 (df_analyze): Adjust.
7896 (loop_inverted_post_order_compute): New function.
7897 (loop_post_order_compute): Likewise.
7898 (df_analyze_loop): New function avoiding whole-function
7900 * loop-invariant.c (find_defs): Use df_analyze_loop.
7901 (find_invariants): Adjust.
7902 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7904 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7906 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7907 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7909 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7911 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7912 traversal when removing references.
7914 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7917 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7919 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7922 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7923 pushing a reload for an autoinc when we had previously reloaded an
7924 inner part of the address.
7926 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7928 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7930 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7931 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7932 when not giving up or versioning for alias only because of
7934 (vect_analyze_data_ref_dependences): Set to true.
7935 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7937 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7938 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7942 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7945 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7946 operand 0 predicate for gathers, use a new pseudo as subtarget.
7948 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7951 * lra-constraints.c (process_alt_operands): Add printing debug info.
7952 Check absence of input/output reloads for matched operands too.
7954 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7956 PR rtl-optimization/59835
7957 * ira.c (ira_init_register_move_cost): Increase cost for
7960 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7962 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7964 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7967 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7968 non-register objects. Use gen_(high/low)part more consistently.
7971 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7974 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7975 endian support, remove tests for WORDS_BIG_ENDIAN.
7976 (p8_mfvsrd_3_<mode>): Likewise.
7977 (reload_gpr_from_vsx<mode>): Likewise.
7978 (reload_gpr_from_vsxsf): Likewise.
7979 (p8_mfvsrd_4_disf): Likewise.
7981 2014-01-16 Richard Biener <rguenther@suse.de>
7983 PR rtl-optimization/46590
7984 * lcm.c (compute_antinout_edge): Use postorder iteration.
7985 (compute_laterin): Use inverted postorder iteration.
7987 2014-01-16 Nick Clifton <nickc@redhat.com>
7990 * varasm.c (output_constant): Return the number of bytes actually
7992 (output_constructor_array_range): Update the field size with the
7993 number of bytes emitted by output_constant.
7994 (output_constructor_regular_field): Likewise. Also do not
7995 complain if the total number of bytes emitted is now greater
7996 than the expected fieldpos.
7997 * output.h (output_constant): Update prototype and descriptive comment.
7999 2014-01-16 Marek Polacek <polacek@redhat.com>
8002 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
8003 it is error_mark_node.
8005 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
8007 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
8008 VALID_AVX256_REG_OR_OI_MODE.
8010 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8012 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8013 current procedure should be profiled.
8015 2014-01-15 Andrew Pinski <apinski@cavium.com>
8017 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8018 of moving from/to the STACK_REG register class.
8020 2014-01-15 Richard Henderson <rth@redhat.com>
8023 * reginfo.c (global_regs_decl): Globalize.
8024 * rtl.h (global_regs_decl): Declare.
8025 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8026 reserved via global_regs.
8028 2014-01-15 Teresa Johnson <tejohnson@google.com>
8030 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8032 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8034 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8035 and vmulosh rather than call gen_vec_widen_smult_*.
8036 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8037 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8038 (vec_widen_smult_even_v16qi): Likewise.
8039 (vec_widen_umult_even_v8hi): Likewise.
8040 (vec_widen_smult_even_v8hi): Likewise.
8041 (vec_widen_umult_odd_v16qi): Likewise.
8042 (vec_widen_smult_odd_v16qi): Likewise.
8043 (vec_widen_umult_odd_v8hi): Likewise.
8044 (vec_widen_smult_odd_v8hi): Likewise.
8045 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8046 vmuloub rather than call gen_vec_widen_umult_*.
8047 (vec_widen_umult_lo_v16qi): Likewise.
8048 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8049 vmulosb rather than call gen_vec_widen_smult_*.
8050 (vec_widen_smult_lo_v16qi): Likewise.
8051 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8052 rather than call gen_vec_widen_umult_*.
8053 (vec_widen_umult_lo_v8hi): Likewise.
8054 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8055 rather than call gen_vec_widen_smult_*.
8056 (vec_widen_smult_lo_v8hi): Likewise.
8058 2014-01-15 Jeff Law <law@redhat.com>
8060 PR tree-optimization/59747
8061 * ree.c (find_and_remove_re): Properly handle case where a second
8062 eliminated extension requires widening a copy created for elimination
8063 of a prior extension.
8064 (combine_set_extension): Ensure that the number of hard regs needed
8065 for a destination register does not change when we widen it.
8067 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8069 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8070 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8071 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8072 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8073 (avr-*-rtems*): Likewise.
8074 (bfin*-rtems*): Likewise.
8075 (moxie-*-rtems*): Likewise.
8076 (h8300-*-rtems*): Likewise.
8077 (i[34567]86-*-rtems*): Likewise.
8078 (lm32-*-rtems*): Likewise.
8079 (m32r-*-rtems*): Likewise.
8080 (m68k-*-rtems*): Likewise.
8081 (microblaze*-*-rtems*): Likewise.
8082 (mips*-*-rtems*): Likewise.
8083 (powerpc-*-rtems*): Likewise.
8084 (sh-*-rtems*): Likewise.
8085 (sparc-*-rtems*): Likewise.
8086 (sparc64-*-rtems*): Likewise.
8087 (v850-*-rtems*): Likewise.
8088 (m32c-*-rtems*): Likewise.
8090 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
8092 PR rtl-optimization/59511
8093 * ira.c (ira_init_register_move_cost): Use memory costs for some
8094 cases of register move cost calculations.
8095 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
8096 instead of BB frequency.
8097 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
8098 * lra-assigns.c (find_hard_regno_for): Ditto.
8100 2014-01-15 Richard Biener <rguenther@suse.de>
8102 PR tree-optimization/59822
8103 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8104 (vectorizable_load): Use it to hoist defs of uses of invariant
8105 loads out of the loop.
8107 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8108 Kugan Vivekanandarajah <kuganv@linaro.org>
8111 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8114 2014-01-15 Richard Biener <rguenther@suse.de>
8116 PR rtl-optimization/59802
8117 * lcm.c (compute_available): Use inverted postorder to seed
8118 the initial worklist.
8120 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8123 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8124 ADDR_REGS for invalid symrefs in non-PIC code.
8126 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8129 * builtins.c (determine_block_size): Initialize *probable_max_size
8130 even if len_rtx is CONST_INT.
8132 2014-01-14 Andrew Pinski <apinski@cavium.com>
8134 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8135 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8136 (cortexa53_tunings): Likewise.
8137 (aarch64_sched_issue_rate): New function.
8138 (TARGET_SCHED_ISSUE_RATE): Define.
8140 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8142 * ira-costs.c (find_costs_and_classes): Add missed
8143 ira_init_register_move_cost_if_necessary.
8145 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8148 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8150 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8153 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8154 to indicate if type is used for function return value. Warn ABI
8155 change if the vector mode isn't available for function return value.
8156 (ix86_function_arg_advance): Pass false to type_natural_mode.
8157 (ix86_function_arg): Likewise.
8158 (ix86_gimplify_va_arg): Likewise.
8159 (function_arg_32): Don't warn ABI change.
8160 (ix86_function_value): Pass true to type_natural_mode.
8161 (ix86_return_in_memory): Likewise.
8162 (ix86_struct_value_rtx): Removed.
8163 (TARGET_STRUCT_VALUE_RTX): Likewise.
8165 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8167 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8168 converting a conditional jump into a conditional return.
8170 2014-01-14 Richard Biener <rguenther@suse.de>
8172 PR tree-optimization/58921
8173 PR tree-optimization/59006
8174 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8175 hoisting invariant stmts.
8176 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8177 invariant loads on the preheader edge if possible.
8179 2014-01-14 Joey Ye <joey.ye@arm.com>
8181 * doc/plugin.texi (Building GCC plugins): Update to C++.
8183 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8185 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8186 (_mm_rcp28_round_ss): Ditto.
8187 (_mm_rsqrt28_round_sd): Ditto.
8188 (_mm_rsqrt28_round_ss): Ditto.
8189 (_mm_rcp28_sd): Ditto.
8190 (_mm_rcp28_ss): Ditto.
8191 (_mm_rsqrt28_sd): Ditto.
8192 (_mm_rsqrt28_ss): Ditto.
8193 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8194 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8195 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8196 (IX86_BUILTIN_RCP28SD): Ditto.
8197 (IX86_BUILTIN_RCP28SS): Ditto.
8198 (IX86_BUILTIN_RSQRT28SD): Ditto.
8199 (IX86_BUILTIN_RSQRT28SS): Ditto.
8200 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8201 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8202 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8203 (ix86_expand_special_args_builtin): Expand new FTYPE.
8204 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8205 (srcp14<mode>): Make insn unary.
8206 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8207 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8208 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8209 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8210 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8211 Fix rounding: make it SAE only.
8212 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8214 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8216 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8217 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8218 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8219 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8220 (round_saeonly_mask_scalar_operand4): Ditto.
8221 (round_saeonly_mask_scalar_op3): Ditto.
8222 (round_saeonly_mask_scalar_op4): Ditto.
8224 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8226 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8227 Implement -maltivec=be for vec_insert and vec_extract.
8229 2014-01-10 DJ Delorie <dj@redhat.com>
8231 * config/msp430/msp430.md (call_internal): Don't allow memory
8232 references with SP as the base register.
8233 (call_value_internal): Likewise.
8234 * config/msp430/constraints.md (Yc): New. For memory references
8235 that don't use SP as a base register.
8237 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8238 "an integer without a # prefix"
8239 * config/msp430/msp430.md (epilogue_helper): Use it.
8241 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8244 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8245 AVX512F gather builtins.
8246 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8247 on gather decls with INTEGER_TYPE masktype.
8248 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8249 directly into the builtin rather than hoisting it before loop.
8251 PR tree-optimization/59387
8252 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8253 (scev_const_prop): If folded_casts and type has undefined overflow,
8254 use force_gimple_operand instead of force_gimple_operand_gsi and
8255 for each added stmt if it is assign with
8256 arith_code_with_undefined_signed_overflow, call
8257 rewrite_to_defined_overflow.
8258 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8259 gimple-fold.h instead.
8260 (arith_code_with_undefined_signed_overflow,
8261 rewrite_to_defined_overflow): Moved to ...
8262 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8263 rewrite_to_defined_overflow): ... here. No longer static.
8264 Include gimplify-me.h.
8265 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8266 rewrite_to_defined_overflow): New prototypes.
8268 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8270 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8272 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8274 * builtins.c (get_object_alignment_2): Minor tweak.
8275 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8277 2014-01-13 Christian Bruel <christian.bruel@st.com>
8279 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8280 optimized non constant lengths.
8282 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8285 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8286 load as __atomic_load_N if possible.
8288 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8290 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8292 (rs6000_expand_builtin): Adjust call.
8294 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8297 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8298 * config/rs6000/rs6000.c: Include target-globals.h.
8299 (rs6000_set_current_function): Instead of doing target_reinit
8300 unconditionally, use save_target_globals_default_opts and
8301 restore_target_globals.
8303 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8305 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8306 (rs6000_expand_builtin): Handle mffs and mtfsf.
8307 (rs6000_init_builtins): Define mffs and mtfsf.
8308 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8309 (rs6000_mffs): New pattern.
8310 (rs6000_mtfsf): New pattern.
8312 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8314 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8315 Start narrowing with START. Apply candidate-use pair
8316 and check overall cost in narrowing.
8317 (iv_ca_prune): Pass new argument.
8319 2014-01-10 Jeff Law <law@redhat.com>
8322 * ree.c (combine_reaching_defs): Ensure the defining statement
8323 occurs before the extension when optimizing extensions with
8324 different source and destination hard registers.
8326 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8329 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8330 vtables into the type inheritance graph.
8332 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8334 PR rtl-optimization/59754
8335 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8336 modes in the REGNO != REGNO case.
8338 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8340 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8342 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8344 PR tree-optimization/59745
8345 * tree-predcom.c (tree_predictive_commoning_loop): Call
8346 free_affine_expand_cache if giving up because components is NULL.
8348 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8349 GC in payload of target_globals struct instead of allocating them on
8350 the heap and the larger structs separately using GC.
8351 * target-globals.h (struct target_globals): Make regs, hard_regs,
8352 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8353 of GTY((skip)) and change type to void *.
8354 (reset_target_globals): Cast loads from those fields to corresponding
8357 2014-01-10 Steve Ellcey <sellcey@mips.com>
8360 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8361 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8362 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8364 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8367 * aarch64-modes.def (CC_Zmode): New flags mode.
8368 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8369 represents an equality.
8370 (aarch64_get_condition_code): Handle CC_Zmode.
8371 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8373 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8375 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8376 extraction in good case.
8378 2014-01-10 Richard Biener <rguenther@suse.de>
8380 PR tree-optimization/59374
8381 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8382 checking after SLP discovery. Mark stmts not participating
8383 in any SLP instance properly.
8385 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8387 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8388 when handling a SET rtx.
8390 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8392 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8393 (cortex-a57): Likewise.
8394 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8396 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8398 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8399 non-iwmmxt builtins.
8401 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8405 * ipa-devirt.c record_target_from_binfo): Take as argument
8406 stack of binfos and lookup matching one for virtual inheritance.
8407 (possible_polymorphic_call_targets_1): Update.
8409 2014-01-10 Huacai Chen <chenhc@lemote.com>
8411 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8412 kernel strings for Loongson-2E/2F/3A.
8414 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8417 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8418 is_gimple_call before calling gimple_call_internal_p.
8420 2014-01-09 Steve Ellcey <sellcey@mips.com>
8422 * Makefile.in (TREE_FLOW_H): Remove.
8423 (TREE_SSA_H): Add file names from tree-flow.h.
8424 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8425 * tree.h: Remove tree-flow.h reference.
8426 * hash-table.h: Remove tree-flow.h reference.
8427 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8428 reference with tree-ssa-loop.h.
8430 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8432 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8433 default element-order behavior for -maltivec.
8434 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8435 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8436 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8437 when targeting big endian, at least for now.
8438 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8440 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8443 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8444 var satisfies use_register_for_decl, just take into account type
8445 alignment, rather than decl alignment.
8447 PR tree-optimization/59622
8448 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8449 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8450 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8451 Don't devirtualize for inplace at all. For targets.length () == 1,
8452 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8454 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8456 * config/i386/i386.md (cpu): Remove the unused btver1.
8458 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8460 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8462 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8465 * tree-core.h (struct target_globals): New forward declaration.
8466 (struct tree_target_option): Add globals field.
8467 * tree.h (TREE_TARGET_GLOBALS): Define.
8468 (prepare_target_option_nodes_for_pch): New prototype.
8469 * target-globals.h (struct target_globals): Define even if
8471 * tree.c (prepare_target_option_node_for_pch,
8472 prepare_target_option_nodes_for_pch): New functions.
8473 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8474 * config/i386/i386.c: Include target-globals.h.
8475 (ix86_set_current_function): Instead of doing target_reinit
8476 unconditionally, use save_target_globals_default_opts and
8477 restore_target_globals.
8479 2014-01-09 Richard Biener <rguenther@suse.de>
8481 PR tree-optimization/59715
8482 * tree-cfg.h (split_critical_edges): Declare.
8483 * tree-cfg.c (split_critical_edges): Export.
8484 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8486 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8488 * cfgexpand.c (expand_stack_vars): Optionally disable
8489 asan stack protection.
8490 (expand_used_vars): Likewise.
8491 (partition_stack_vars): Likewise.
8492 * asan.c (asan_emit_stack_protection): Optionally disable
8493 after return stack usage.
8494 (instrument_derefs): Optionally disable memory access instrumentation.
8495 (instrument_builtin_call): Likewise.
8496 (instrument_strlen_call): Likewise.
8497 (asan_protect_global): Optionally disable global variables protection.
8498 * doc/invoke.texi: Added doc for new options.
8499 * params.def: Added new options.
8500 * params.h: Likewise.
8502 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8504 PR rtl-optimization/59724
8505 * ifcvt.c (cond_exec_process_if_block): Don't call
8506 flow_find_head_matching_sequence with 0 longest_match.
8507 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8508 non-active insns if !stop_after.
8509 (try_head_merge_bb): Revert 2014-01-07 changes.
8511 2014-01-08 Jeff Law <law@redhat.com>
8513 * ree.c (get_sub_rtx): New function, extracted from...
8514 (merge_def_and_ext): Here.
8515 (combine_reaching_defs): Use get_sub_rtx.
8517 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8519 * cgraph.h (varpool_variable_node): Do not choke on null node.
8521 2014-01-08 Catherine Moore <clm@codesourcery.com>
8523 * config/mips/mips.md (simple_return): Attempt to use JRC
8525 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8527 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8529 PR rtl-optimization/59137
8530 * reorg.c (steal_delay_list_from_target): Call update_block for
8532 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8534 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8536 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8537 two duplicate entries.
8539 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8542 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8544 * config/mips/mips.c (mips_truncated_op_cost): New function.
8545 (mips_rtx_costs): Adjust test for BADDU.
8546 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8548 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8550 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8551 (*baddu_si): ...this new pattern.
8553 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8556 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8558 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8561 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8563 (expand_expr, expand_normal): Adjust.
8564 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8565 inner_reference_p. Use inner_reference_p to expand inner references.
8566 (store_expr): Adjust.
8567 * cfgexpand.c (expand_call_stmt): Adjust.
8569 2014-01-08 Rong Xu <xur@google.com>
8571 * gcov-io.c (gcov_var): Move from gcov-io.h.
8572 (gcov_position): Ditto.
8573 (gcov_is_error): Ditto.
8574 (gcov_rewrite): Ditto.
8575 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8576 only part to libgcc/libgcov.h.
8578 2014-01-08 Marek Polacek <polacek@redhat.com>
8581 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8583 2014-01-08 Marek Polacek <polacek@redhat.com>
8586 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8588 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8590 PR rtl-optimization/59649
8591 * stor-layout.c (get_mode_bounds): For BImode return
8592 0 and STORE_FLAG_VALUE.
8594 2014-01-08 Richard Biener <rguenther@suse.de>
8597 * gimple.h (is_gimple_builtin_call): Remove.
8598 (gimple_builtin_call_types_compatible_p): New.
8599 (gimple_call_builtin_p): New overload.
8600 * gimple.c (is_gimple_builtin_call): Remove.
8601 (validate_call): Rename to ...
8602 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8604 (validate_type): New static function.
8605 (gimple_call_builtin_p): New overload and adjust.
8606 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8607 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8608 (gimple_fold_stmt_to_constant_1): Likewise.
8609 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8611 2014-01-08 Richard Biener <rguenther@suse.de>
8614 * gimplify.c (gimplify_expr): Gimplify register-register type
8615 VIEW_CONVERT_EXPRs to separate stmts.
8617 2014-01-07 Jeff Law <law@redhat.com>
8620 * ree.c (combine_set_extension): Handle case where source
8621 and destination registers in an extension insn are different.
8622 (combine_reaching_defs): Allow source and destination registers
8623 in extension to be different under limited circumstances.
8624 (add_removable_extension): Remove restriction that the
8625 source and destination registers in the extension are the same.
8626 (find_and_remove_re): Emit a copy from the extension's
8627 destination to its source after the defining insn if
8628 the source and destination registers are different.
8631 * ifcvt.c (merge_if_block): If we are merging a block with more than
8632 one successor with a block with no successors, remove any BARRIER
8633 after the second block.
8635 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8637 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8639 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8642 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8643 for 14-bit register offsets when INT14_OK_STRICT is false.
8645 2014-01-07 Roland Stigge <stigge@antcom.de>
8646 Michael Meissner <meissner@linux.vnet.ibm.com>
8649 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8650 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8652 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8654 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8657 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8659 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8660 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8661 rtx is const0_rtx or not.
8663 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8666 * target-globals.c (save_target_globals): Remove this_fn_optab
8668 * toplev.c: Include optabs.h.
8669 (target_reinit): Temporarily restore the global options if another
8670 set of options are in force.
8672 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8674 PR rtl-optimization/58668
8675 * cfgcleanup.c (flow_find_cross_jump): Don't count
8676 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8677 to determine what is counted.
8678 (flow_find_head_matching_sequence): Use active_insn_p to determine
8680 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8682 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8683 determine what is counted.
8685 PR tree-optimization/59643
8686 * tree-predcom.c (split_data_refs_to_components): If one dr is
8687 read and one write, determine_offset fails and the write isn't
8688 in the bad component, just put the read into the bad component.
8690 2014-01-07 Mike Stump <mikestump@comcast.net>
8691 Jakub Jelinek <jakub@redhat.com>
8694 * tree-core.h (struct tree_optimization_option): Change optabs
8695 type from unsigned char * to void *.
8696 * optabs.c (init_tree_optimization_optabs): Adjust
8697 TREE_OPTIMIZATION_OPTABS initialization.
8699 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8702 * config/i386/i386.h (struct machine_function): Add
8703 no_drap_save_restore field.
8704 * config/i386/i386.c (ix86_save_reg): Use
8705 !cfun->machine->no_drap_save_restore instead of
8706 crtl->stack_realign_needed.
8707 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8708 this function clears frame_pointer_needed. Set
8709 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8710 and DRAP reg is needed.
8712 2014-01-06 Marek Polacek <polacek@redhat.com>
8715 * doc/implement-c.texi: Mention that other integer types are
8716 permitted as bit-field types in strictly conforming mode.
8718 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8720 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8723 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8725 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8727 2014-01-06 Martin Jambor <mjambor@suse.cz>
8730 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8732 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8734 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8738 * var-tracking.c (add_stores): Preserve the value of the source even if
8739 we don't record the store.
8741 2014-01-06 Terry Guo <terry.guo@arm.com>
8743 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8745 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8748 * config/darwin.c (darwin_function_section): Adjust return values to
8749 correspond to optimisation changes made in r206070.
8751 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8753 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8754 from prefetch_block tune setting.
8755 (nocona_cost): Correct size of prefetch block to 64.
8757 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8759 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8760 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8761 used to save the static chain register in the computation of the offset
8762 from which the FP registers need to be restored.
8764 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8766 PR tree-optimization/59519
8767 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8768 ICE if get_current_def (current_new_name) is already non-NULL, as long
8769 as it is a phi result of some other phi in *new_exit_bb that has
8772 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8773 or vmovdqu* for misaligned_operand.
8774 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8775 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8776 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8777 aligned_mem for AVX512F masked aligned load and store builtins and for
8780 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8782 PR tree-optimization/59651
8783 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8784 Address range for negative step should be added by TYPE_SIZE_UNIT.
8786 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8788 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8789 overlapping registers also for registers other than the stack pointer.
8791 2014-01-03 Marek Polacek <polacek@redhat.com>
8794 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8797 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8800 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8803 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8804 (push splitter): Use <P:MODE_SIZE> instead of
8805 GET_MODE_SIZE (<P:MODE>mode).
8806 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8807 (mov -1, reg peephole2): Likewise.
8808 * config/i386/sse.md (*mov<mode>_internal,
8809 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8810 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8811 *<code><mode>3, *andnot<mode>3<mask_name>,
8812 <mask_codefor><code><mode>3<mask_name>): Likewise.
8813 * config/i386/subst.md (mask_mode512bit_condition,
8814 sd_mask_mode512bit_condition): Likewise.
8816 2014-01-02 Xinliang David Li <davidxl@google.com>
8818 PR tree-optimization/59303
8819 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8820 (dump_predicates): Better output format.
8821 (pred_equal_p): New function.
8822 (is_neq_relop_p): Ditto.
8823 (is_neq_zero_form_p): Ditto.
8824 (pred_expr_equal_p): Ditto.
8825 (pred_neg_p): Ditto.
8826 (simplify_pred): Ditto.
8827 (simplify_preds_2): Ditto.
8828 (simplify_preds_3): Ditto.
8829 (simplify_preds_4): Ditto.
8830 (simplify_preds): Ditto.
8832 (push_to_worklist): Ditto.
8833 (get_pred_info_from_cmp): Ditto.
8834 (is_degenerated_phi): Ditto.
8835 (normalize_one_pred_1): Ditto.
8836 (normalize_one_pred): Ditto.
8837 (normalize_one_pred_chain): Ditto.
8838 (normalize_preds): Ditto.
8839 (normalize_cond_1): Remove function.
8840 (normalize_cond): Ditto.
8841 (is_gcond_subset_of): Ditto.
8842 (is_subset_of_any): Ditto.
8843 (is_or_set_subset_of): Ditto.
8844 (is_and_set_subset_of): Ditto.
8845 (is_norm_cond_subset_of): Ditto.
8846 (pred_chain_length_cmp): Ditto.
8847 (convert_control_dep_chain_into_preds): Type change.
8848 (find_predicates): Ditto.
8849 (find_def_preds): Ditto.
8850 (destroy_predicates_vecs): Ditto.
8851 (find_matching_predicates_in_rest_chains): Ditto.
8852 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8853 (is_pred_expr_subset): Ditto.
8854 (is_pred_chain_subset_of): Ditto.
8855 (is_included_in): Ditto.
8856 (is_superset_of): Ditto.
8858 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8860 Update copyright years.
8862 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8864 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8865 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8866 config/arc/arc.md, config/arc/arc.opt,
8867 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8868 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8869 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8870 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8871 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8872 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8873 vtable-verify.c, vtable-verify.h: Use the standard form for the
8876 2014-01-02 Tobias Burnus <burnus@net-b.de>
8878 * gcc.c (process_command): Update copyright notice dates.
8879 * gcov-dump.c: Ditto.
8881 * doc/cpp.texi: Bump @copying's copyright year.
8882 * doc/cppinternals.texi: Ditto.
8883 * doc/gcc.texi: Ditto.
8884 * doc/gccint.texi: Ditto.
8885 * doc/gcov.texi: Ditto.
8886 * doc/install.texi: Ditto.
8887 * doc/invoke.texi: Ditto.
8889 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8891 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8893 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8895 * config/i386/sse.md (*mov<mode>_internal): Guard
8896 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8898 PR rtl-optimization/59647
8899 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8900 new_rtx into UNSIGNED_FLOAT rtxes.
8902 Copyright (C) 2014 Free Software Foundation, Inc.
8904 Copying and distribution of this file, with or without modification,
8905 are permitted in any medium without royalty provided the copyright
8906 notice and this notice are preserved.