1 2014-05-05 Bin Cheng <bin.cheng@arm.com>
3 PR tree-optimization/60363
4 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
5 (copy_phi_args): New parameters. Call get_value_locus_in_path.
6 (update_destination_phis): New parameter.
7 (create_edge_and_update_destination_phis): Ditto.
8 (ssa_fix_duplicate_block_edges): Pass new arguments.
9 (thread_single_edge): Ditto.
11 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
13 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
14 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
15 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
16 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
17 Use RS6000_BTM_HARD_FLOAT.
18 (BU_MISC_2): Likewise.
19 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
20 RS6000_BTM_HARD_FLOAT.
21 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
23 (rs6000_invalid_builtin): Add hard floating builtin support.
24 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
26 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
28 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
30 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
31 Add missing function* argument.
33 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
35 * lra-constraints.c (valid_address_p): Move earlier in file.
36 Add a constraint argument to the address_info version.
37 (satisfies_memory_constraint_p): New function.
38 (satisfies_address_constraint_p): Likewise.
39 (process_alt_operands, curr_insn_transform): Use them.
40 (process_address): Pass the constraint to valid_address_p when
41 checking address operands.
43 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
45 * config/mips/mips.c (mips_isa_rev): New variable.
46 (mips_set_architecture): Set it.
47 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
49 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
50 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
51 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
52 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
53 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
54 conditions in terms of mips_isa_rev.
55 (mips_isa_rev): Declare.
57 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
59 * config/sh/sh-mem.cc: Use tabs instead of spaces.
60 (prob_unlikely, prob_likely): Make variables const.
62 2014-05-03 Denis Chertykov <chertykov@gmail.com>
64 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
66 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
68 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
70 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
72 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
73 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
74 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
76 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
77 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
79 Replace usage of ROUND_REG with sh_round_reg.
80 Use CEIL instead of ROUND_ADVANCE.
82 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
85 * config/sh/sh.c: Include stdlib headers before everything else.
87 2014-05-02 Jakub Jelinek <jakub@redhat.com>
89 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
90 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
91 (gimplify_adjust_omp_clauses): Simd region is never
92 directly nested in combined parallel. Instead, for linear
93 with copyin/copyout, if in combined for simd loop, make decl
94 firstprivate/lastprivate on OMP_FOR.
95 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
96 expand_omp_for_static_chunk): When setting endvar, also set
97 fd->loop.v to the same value.
99 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
101 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
103 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
105 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
108 2014-05-02 Marek Polacek <polacek@redhat.com>
110 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
112 2014-02-26 Kito Cheng <kito@0xlab.org>
114 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
115 to a C expression marco.
116 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
117 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
118 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
119 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
120 HONOR_REG_ALLOC_ORDER.
121 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
123 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
125 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
127 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
129 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
131 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
133 * tree-if-conv.c (is_cond_scalar_reduction): New function.
134 (convert_scalar_cond_reduction): Likewise.
135 (predicate_scalar_phi): Add recognition and transformation
136 of simple conditioanl reduction to be vectorizable.
138 2014-05-01 Marek Polacek <polacek@redhat.com>
141 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
143 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
145 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
146 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
147 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
148 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
149 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
150 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
151 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
152 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
154 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
156 * config/arc/arc.opt (mlra): Move comment above option name
157 to avoid mis-parsing as language options.
159 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
161 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
162 * config/sol2.h: ... here.
163 * config/sol2-10.h: Remove.
165 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
166 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
167 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
168 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
169 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
170 * config/sol2.h: ... here.
171 (SECTION_NAME_FORMAT): Don't redefine.
172 (STARTFILE_ARCH32_SPEC): Rename to ...
173 (STARTFILE_ARCH_SPEC): ... this.
174 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
175 * config/sparc/sol2.h: ... here.
176 (SECTION_NAME_FORMAT): Don't undef.
177 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
178 (SUBTARGET_EXTRA_SPECS): Remove.
179 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
181 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
182 (MD_STARTFILE_PREFIX): Remove.
183 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
184 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
185 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
186 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
187 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
188 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
189 * config/i386/sol2.h: ... here.
190 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
191 * config/i386/sol2-bi.h: Remove.
192 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
193 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
195 * config/i386/t-sol2-64: Rename to ...
196 * config/i386/t-sol2: ... this.
197 * config/sparc/t-sol2-64: Rename to ...
198 * config/sparc/t-sol2: ... this.
200 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
201 sol2_tm_file_head, sol2_tm_file_tail.
202 Include ${cpu_type}/sol2.h before sol2.h.
204 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
205 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
206 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
207 Reflect i386/t-sol2-64 renaming.
208 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
209 Reflect sparc/t-sol2-64 renaming.
211 2014-04-30 Richard Biener <rguenther@suse.de>
213 * passes.c (execute_function_todo): Move TODO_verify_stmts
214 and TODO_verify_ssa under the TODO_verify_il umbrella.
215 * tree-ssa.h (verify_ssa): Adjust prototype.
216 * tree-ssa.c (verify_ssa): Add parameter to tell whether
217 we should verify SSA operands.
218 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
219 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
220 whether we should verify whether not throwing stmts have EH info.
221 * graphite-scop-detection.c (create_sese_edges): Adjust.
222 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
223 * tree-eh.c (lower_try_finally_switch): Do not add the
224 default case label twice.
226 2014-04-30 Marek Polacek <polacek@redhat.com>
228 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
229 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
230 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
231 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
233 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
235 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
236 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
237 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
238 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
239 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
240 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
241 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
242 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
244 2014-04-29 David Malcolm <dmalcolm@redhat.com>
246 * tree-cfg.c (dump_function_to_file): Dump the return type of
247 functions, in a line to itself before the function body, mimicking
248 the layout of a C function.
250 2014-04-29 Jakub Jelinek <jakub@redhat.com>
252 PR tree-optimization/60971
253 * tree-tailcall.c (process_assignment): Reject conversions which
256 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
258 * calls.c (initialize_argument_information): Always treat
259 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
260 (expand_call): Likewise.
261 (emit_library_call_calue_1): Likewise.
262 * expr.c (PUSH_ARGS_REVERSED): Do not define.
263 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
266 2014-04-29 Nick Clifton <nickc@redhat.com>
268 * config/msp430/msp430.md (umulsidi): Fix typo.
269 (mulhisi3): Enable even inside interrupt handlers.
270 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
271 bigger return address pushed in large mode.
273 2014-04-29 Nick Clifton <nickc@redhat.com>
275 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
276 (arc_init_reg_tables): Use a machine_mode enum to iterate over
278 * config/m32r/m32r.c (init_reg_tables): Likewise.
279 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
280 enum to hold the modes.
282 2014-04-29 Richard Biener <rguenther@suse.de>
284 * dominance.c (free_dominance_info): Add overload with
286 (dom_info_state): Likewise.
287 (dom_info_available_p): Likewise.
288 * basic-block.h (free_dominance_info, dom_info_state,
289 dom_info_available_p): Declare overloads.
290 * passes.c (execute_function_todo): Verify that verifiers
291 don't change dominator info state. Drop dominator info
292 for IPA pass invocations.
293 * cgraph.c (release_function_body): Restore asserts that
294 dominator information is released.
296 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
298 * doc/invoke.texi: Fix typo.
299 * tree-vrp.c: Fix typos.
300 * gimple.c (infer_nonnull_range): Reorder operands of an &&
303 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
305 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
307 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
309 * config/aarch64/aarch64-builtins.c
310 (aarch64_types_storestruct_lane_qualifiers): New.
311 (TYPES_STORESTRUCT_LANE): Likewise.
312 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
313 (st3_lane): Likewise.
314 (st4_lane): Likewise.
315 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
316 (vec_store_lanesci_lane<mode>): Likewise.
317 (vec_store_lanesxi_lane<mode>): Likewise.
318 (aarch64_st2_lane<VQ:mode>): Likewise.
319 (aarch64_st3_lane<VQ:mode>): Likewise.
320 (aarch64_st4_lane<VQ:mode>): Likewise.
321 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
322 * config/aarch64/arm_neon.h
323 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
324 use new macro arguments.
325 (__ST3_LANE_FUNC): Likewise.
326 (__ST4_LANE_FUNC): Likewise.
327 * config/aarch64/iterators.md (V_TWO_ELEM): New.
328 (V_THREE_ELEM): Likewise.
329 (V_FOUR_ELEM): Likewise.
331 2014-04-28 David Malcolm <dmalcolm@redhat.com>
333 * doc/gimple.texi: Replace the description of the now-defunct
334 union gimple_statement_d with a diagram showing the
335 gimple_statement_base class hierarchy and its relationships to
336 the GSS_ and GIMPLE_ enums.
338 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
340 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
341 * config/aarch64/aarch64.c
342 (aarch64_cannot_change_mode_class): Weaken conditions.
343 (aarch64_modes_tieable_p): New.
344 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
346 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
348 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
349 (loadsync_<mode>): Change mode.
350 (load_quadpti, store_quadpti): New.
351 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
352 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
354 2014-04-28 Martin Jambor <mjambor@suse.cz>
356 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
357 same alias type as the original statement.
358 (subreplacement_assignment_data): New type.
359 (handle_unscalarized_data_in_subtree): New type of parameter,
360 generate new memory accesses with same alias type as the original
362 (load_assign_lhs_subreplacements): Likewise.
363 (sra_modify_constructor_assign): Generate new memory accesses with
364 same alias type as the original statement.
366 2014-04-28 Richard Biener <rguenther@suse.de>
368 * tree-pass.h (TODO_verify_il): Define.
369 (TODO_verify_all): Complete properly.
370 * passes.c (execute_function_todo): Move existing loop-closed
371 SSA verification under TODO_verify_il.
372 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
373 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
374 Fix tree sharing issue.
376 2014-04-28 Richard Biener <rguenther@suse.de>
379 * builtins.def (DEF_C11_BUILTIN): Add.
380 (BUILT_IN_ALIGNED_ALLOC): Likewise.
381 * coretypes.h (enum function_class): Add function_c11_misc.
382 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
383 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
384 (call_may_clobber_ref_p_1): Likewise.
385 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
386 (mark_all_reaching_defs_necessary_1): Likewise.
387 (propagate_necessity): Likewise.
388 (eliminate_unnecessary_stmts): Likewise.
389 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
391 2014-04-28 Richard Biener <rguenther@suse.de>
393 * tree-vrp.c (vrp_var_may_overflow): Remove.
394 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
395 with overflow immediately bump to one before that value and
396 let iteration figure out overflow status.
398 2014-04-28 Richard Biener <rguenther@suse.de>
400 * configure.ac: Do valgrind header checks unconditionally.
401 Add --enable-valgrind-annotations.
402 * system.h: Guard valgrind header inclusion with
403 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
404 * alloc-pool.c (pool_alloc, pool_free): Use
405 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
406 to guard possibly dead code.
407 * config.in: Regenerated.
408 * configure: Likewise.
410 2014-04-28 Jeff Law <law@redhat.com>
412 PR tree-optimization/60902
413 * tree-ssa-threadedge.c
414 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
415 over real defs when invalidating outputs from statements that do not
416 produce useful outputs for threading.
418 2014-04-28 Richard Biener <rguenther@suse.de>
420 PR tree-optimization/60979
421 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
422 SCOPs that end in a block with a successor with abnormal
425 2014-04-28 Richard Biener <rguenther@suse.de>
427 * tree-pass.h (execute_pass_list): Adjust prototype.
428 * passes.c (pass_manager::execute_early_local_passes):
430 (do_per_function): Change callback signature, push all actual
431 work to the callbals.
432 (do_per_function_toporder): Likewise.
433 (execute_function_dump): Adjust.
434 (execute_function_todo): Likewise.
435 (clear_last_verified): Likewise.
436 (verify_curr_properties): Likewise.
437 (update_properties_after_pass): Likewise.
438 (execute_pass_list_1): Split out from ...
439 (execute_pass_list): ... here. Adjust.
440 (execute_ipa_pass_list): Likewise.
441 * cgraphunit.c (cgraph_add_new_function): Adjust.
442 (analyze_function): Likewise.
443 (expand_function): Likewise.
444 * cgraph.c (release_function_body): Free dominance info
445 here instead of asserting it was magically freed elsewhere.
447 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
449 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
450 * configure: Regenerate.
451 * config/sparc/sparc.opt (muser-mode): New option.
452 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
454 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
455 * doc/invoke.texi (SPARC options): Document -muser-mode.
457 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
459 * cselib.c (find_slot_memmode): Delete.
460 (cselib_hasher): Change compare_type to a struct.
461 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
463 (preserve_constants_and_equivs): Adjust for new compare_type.
464 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
465 (wrap_constant): Delete.
466 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
468 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
470 * doc/install.texi (Building with profile feedback): Remove
473 2014-04-26 Tom de Vries <tom@codesourcery.com>
475 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
478 2014-04-25 Cary Coutant <ccoutant@google.com>
481 * dwarf2out.c (should_move_die_to_comdat): A type definition
482 can contain a subprogram definition, but don't move it to a
484 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
485 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
487 (clone_tree_hash): Rename to...
488 (clone_tree_partial): ...this; change callers. Copy
489 DW_TAG_subprogram DIEs as declarations.
490 (copy_decls_walk): Don't copy children of a declaration into a
493 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
496 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
499 2014-04-25 Jiong Wang <jiong.wang@arm.com>
501 * config/arm/predicates.md (call_insn_operand): Add long_call check.
502 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
504 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
507 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
509 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
512 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
514 PR tree-optimization/60930
515 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
516 creating a multiply candidate by folding two constant
517 multiplicands when the result overflows.
519 2014-04-25 Jakub Jelinek <jakub@redhat.com>
521 PR tree-optimization/60960
522 * tree-vect-generic.c (expand_vector_operation): Only call
523 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
525 2014-04-25 Tom de Vries <tom@codesourcery.com>
527 * expr.c (clobber_reg_mode): New function.
528 * expr.h (clobber_reg): New function.
530 2014-04-25 Tom de Vries <tom@codesourcery.com>
532 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
535 2014-04-25 Radovan Obradovic <robradovic@mips.com>
536 Tom de Vries <tom@codesourcery.com>
538 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
540 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
541 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
542 new argument to find_all_hard_reg_sets call.
544 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
546 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
547 Use HOST_WIDE_INT_C for mask literal.
548 (aarch_rev16_shleft_mask_imm_p): Likewise.
550 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
553 * config/sparc/sparc.md (ashlsi3_extend): Delete.
555 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
557 PR preprocessor/56540
558 * config/i386/i386-c.c (ix86_target_macros): Define
559 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
561 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
563 * configure.ac (tga_func): Remove.
564 (LIB_TLS_SPEC): Remove.
565 * configure: Regenerate.
566 * config.in: Regenerate.
567 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
569 2014-04-25 Richard Biener <rguenther@suse.de>
572 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
573 call stmt use/clobber sets during stmt walk instead of
574 walking the possibly incomplete set of caller edges.
576 2014-04-25 Richard Biener <rguenther@suse.de>
579 * passes.c (apply_ipa_transforms): Inline into only caller ...
580 (execute_one_pass): ... here. Properly bring in function
581 bodies for nodes we want to apply IPA transforms to.
583 2014-04-24 Cong Hou <congh@google.com>
585 PR tree-optimization/60896
586 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
587 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
588 (vect_mark_pattern_stmts): Set the def type of all statements in
589 PATTERN_DEF_SEQ as vect_internal_def.
591 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
593 * doc/extend.texi (PowerPC Built-in Functions): Document new
594 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
596 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
598 * config/rs6000/predicates.md (const_0_to_3_operand): New
599 predicate to match 0..3 integer constants.
601 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
602 to support adding miscellaneous builtin functions.
603 (BU_DFP_MISC_2): Likewise.
604 (BU_P7_MISC_1): Likewise.
605 (BU_P7_MISC_2): Likewise.
606 (BU_P8V_MISC_3): Likewise.
607 (BU_MISC_1): Likewise.
608 (BU_MISC_2): Likewise.
609 (DIVWE): Add extended divide builtin functions.
617 (DXEX): Add decimal floating-point builtin functions.
629 (CDTBCD): Add new BCD builtin functions.
633 (BCDADD_LT): Likewise.
634 (BCDADD_EQ): Likewise.
635 (BCDADD_GT): Likewise.
636 (BCDADD_OV): Likewise.
638 (BCDSUB_LT): Likewise.
639 (BCDSUB_EQ): Likewise.
640 (BCDSUB_GT): Likewise.
641 (BCDSUB_OV): Likewise.
642 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
643 (UNPACK_TD): Likewise.
645 (UNPACK_TF): Likewise.
646 (UNPACK_TF_0): Likewise.
647 (UNPACK_TF_1): Likewise.
648 (PACK_V1TI): Likewise.
649 (UNPACK_V1TI): Likewise.
651 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
652 support for decimal floating point builtin functions.
653 (rs6000_expand_ternop_builtin): Add checks for the new builtin
654 functions that take constant arguments.
655 (rs6000_invalid_builtin): Add decimal floating point builtin
657 (rs6000_init_builtins): Setup long double, _Decimal64, and
658 _Decimal128 types for new builtin functions.
659 (builtin_function_type): Set the unsigned flags appropriately for
660 the new builtin functions.
661 (rs6000_opt_masks): Add support for decimal floating point builtin
664 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
665 floating point builtin functions.
666 (RS6000_BTM_COMMON): Likewise.
667 (RS6000_BTI_long_double): Likewise.
668 (RS6000_BTI_dfloat64): Likewise.
669 (RS6000_BTI_dfloat128): Likewise.
670 (long_double_type_internal_node): Likewise.
671 (dfloat64_type_internal_node): Likewise.
672 (dfloat128_type_internal_node): Likewise.
674 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
675 2.07 bcd arithmetic instructions.
676 (UNSPEC_BCDSUB): Likewise.
677 (UNSPEC_BCD_OVERFLOW): Likewise.
678 (UNSPEC_BCD_ADD_SUB): Likewise.
679 (bcd_add_sub): Likewise.
680 (BCD_TEST): Likewise.
681 (bcd<bcd_add_sub>): Likewise.
682 (bcd<bcd_add_sub>_test): Likewise.
683 (bcd<bcd_add_sub>_test2): Likewise.
684 (bcd<bcd_add_sub>_<code>): Likewise.
685 (peephole2 for combined bcd ops): Likewise.
687 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
688 decimal floating point builtin functions.
689 (UNSPEC_DENBCD): Likewise.
690 (UNSPEC_DXEX): Likewise.
691 (UNSPEC_DIEX): Likewise.
692 (UNSPEC_DSCLI): Likewise.
693 (UNSPEC_DSCRI): Likewise.
694 (D64_D128): Likewise.
695 (dfp_suffix): Likewise.
696 (dfp_ddedpd_<mode>): Likewise.
697 (dfp_denbcd_<mode>): Likewise.
698 (dfp_dxex_<mode>): Likewise.
699 (dfp_diex_<mode>): Likewise.
700 (dfp_dscli_<mode>): Likewise.
701 (dfp_dscri_<mode>): Likewise.
703 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
705 (UNSPEC_CDTBCD): Likewise.
706 (UNSPEC_CBCDTD): Likewise.
707 (UNSPEC_DIVE): Add support for new extended divide builtin
709 (UNSPEC_DIVEO): Likewise.
710 (UNSPEC_DIVEU): Likewise.
711 (UNSPEC_DIVEUO): Likewise.
712 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
713 pack/unpack 128-bit types.
714 (UNSPEC_PACK_128BIT): Likewise.
715 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
716 (udiv<mode>3): Use idiv_ldiv mode attribute.
717 (div<mode>3): Likewise.
718 (addg6s): Add new BCD builtin functions.
721 (UNSPEC_DIV_EXTEND): Add support for new extended divide
723 (div_extend): Likewise.
724 (div<div_extend>_<mode>"): Likewise.
725 (FP128_64): Add support for new builtin functions to pack/unpack
727 (unpack<mode>): Likewise.
728 (unpacktf_0): Likewise.
729 (unpacktf_1): Likewise.
730 (unpack<mode>_dm): Likewise.
731 (unpack<mode>_nodm): Likewise.
732 (pack<mode>): Likewise.
733 (unpackv1ti): Likewise.
734 (packv1ti): Likewise.
736 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
738 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
741 2014-04-24 Jakub Jelinek <jakub@redhat.com>
743 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
744 * gimplify.c (omp_is_private): Change last argument's type to int.
745 Only diagnose lastprivate if the simd argument is 1, only diagnose
746 linear if the simd argument is 2.
747 (gimplify_omp_for): Adjust omp_is_private callers. When adding
748 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
749 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
750 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
751 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
752 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
753 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
754 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
755 * tree-nested.c (convert_nonlocal_omp_clauses,
756 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
758 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
761 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
764 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
766 * flag-types.h (enum ivar_visibility): Add.
768 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
770 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
773 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
775 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
777 2014-04-24 Radovan Obradovic <robradovic@mips.com>
778 Tom de Vries <tom@codesourcery.com>
780 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
781 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
783 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
784 * emit-rtl.c (try_split): Same.
786 2014-04-24 Radovan Obradovic <robradovic@mips.com>
787 Tom de Vries <tom@codesourcery.com>
789 * common.opt (fuse-caller-save): New option.
791 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
793 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
796 2014-04-24 Richard Biener <rguenther@suse.de>
798 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
799 during TER and instead use the sepops interface for expanding
800 non-GIMPLE_SINGLE_RHS.
802 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
804 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
805 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
807 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
809 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
810 assembler 64-bit option.
811 * configure: Regenerate.
813 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
815 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
816 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
817 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
818 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
819 (TARGET_CRYPTO): Take TARGET_SIMD into account.
821 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
823 * config/aarch64/aarch64-builtins.c
824 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
825 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
826 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
827 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
829 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
830 (Vrevsuff): New mode attribute.
832 2014-04-24 Terry Guo <terry.guo@arm.com>
834 * config/arm/arm.h (machine_function): Define variable
835 after_arm_reorg here.
836 * config/arm/arm.c (after_arm_reorg): Remove the definition.
837 (arm_split_constant): Update the way to access variable
840 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
842 2014-04-23 Tom de Vries <tom@codesourcery.com>
844 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
846 2014-04-23 David Malcolm <dmalcolm@redhat.com>
848 * is-a.h: Update comments to reflect the following changes to the
849 "pointerness" of the API, making the template parameter match the
850 return type, allowing use of is-a.h with typedefs of pointers.
851 (is_a_helper::cast): Return a T rather then a pointer to a T, so
852 that the return type matches the parameter to the is_a_helper.
854 (dyn_cast): Likewise.
856 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
857 pointer from the is-a.h API.
859 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
860 (is_a_helper <cgraph_node *>::test): ...this, matching change to
862 (is_a_helper <varpool_node>::test): Likewise, convert to...
863 (is_a_helper <varpool_node *>::test): ...this.
865 (varpool_first_variable): Update for removal of implicit pointer
867 (varpool_next_variable): Likewise.
868 (varpool_first_static_initializer): Likewise.
869 (varpool_next_static_initializer): Likewise.
870 (varpool_first_defined_variable): Likewise.
871 (varpool_next_defined_variable): Likewise.
872 (cgraph_first_defined_function): Likewise.
873 (cgraph_next_defined_function): Likewise.
874 (cgraph_first_function): Likewise.
875 (cgraph_next_function): Likewise.
876 (cgraph_first_function_with_gimple_body): Likewise.
877 (cgraph_next_function_with_gimple_body): Likewise.
878 (cgraph_alias_target): Likewise.
879 (varpool_alias_target): Likewise.
880 (cgraph_function_or_thunk_node): Likewise.
881 (varpool_variable_node): Likewise.
882 (symtab_real_symbol_p): Likewise.
883 * cgraphunit.c (referred_to_p): Likewise.
884 (analyze_functions): Likewise.
885 (handle_alias_pairs): Likewise.
886 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
887 * gimple-ssa.h (gimple_vuse_op): Likewise.
888 (gimple_vdef_op): Likewise.
889 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
890 * gimple.c (gimple_build_asm_1): Likewise.
891 (gimple_build_try): Likewise.
892 (gimple_build_resx): Likewise.
893 (gimple_build_eh_dispatch): Likewise.
894 (gimple_build_omp_for): Likewise.
895 (gimple_omp_for_set_clauses): Likewise.
897 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
898 (is_a_helper <gimple_statement_asm *>::test): ...this.
899 (is_a_helper <gimple_statement_bind>::test): Convert to...
900 (is_a_helper <gimple_statement_bind *>::test): ...this.
901 (is_a_helper <gimple_statement_call>::test): Convert to...
902 (is_a_helper <gimple_statement_call *>::test): ...this.
903 (is_a_helper <gimple_statement_catch>::test): Convert to...
904 (is_a_helper <gimple_statement_catch *>::test): ...this.
905 (is_a_helper <gimple_statement_resx>::test): Convert to...
906 (is_a_helper <gimple_statement_resx *>::test): ...this.
907 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
908 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
909 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
910 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
911 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
912 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
913 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
914 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
915 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
916 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
917 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
918 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
919 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
920 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
921 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
922 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
923 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
924 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
925 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
926 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
927 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
928 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
929 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
930 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
931 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
932 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
933 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
934 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
935 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
936 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
937 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
938 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
939 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
940 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
941 (is_a_helper <gimple_statement_phi>::test): Convert to...
942 (is_a_helper <gimple_statement_phi *>::test): ...this.
943 (is_a_helper <gimple_statement_transaction>::test): Convert to...
944 (is_a_helper <gimple_statement_transaction *>::test): ...this.
945 (is_a_helper <gimple_statement_try>::test): Convert to...
946 (is_a_helper <gimple_statement_try *>::test): ...this.
947 (is_a_helper <gimple_statement_wce>::test): Convert to...
948 (is_a_helper <gimple_statement_wce *>::test): ...this.
949 (is_a_helper <const gimple_statement_asm>::test): Convert to...
950 (is_a_helper <const gimple_statement_asm *>::test): ...this.
951 (is_a_helper <const gimple_statement_bind>::test): Convert to...
952 (is_a_helper <const gimple_statement_bind *>::test): ...this.
953 (is_a_helper <const gimple_statement_call>::test): Convert to...
954 (is_a_helper <const gimple_statement_call *>::test): ...this.
955 (is_a_helper <const gimple_statement_catch>::test): Convert to...
956 (is_a_helper <const gimple_statement_catch *>::test): ...this.
957 (is_a_helper <const gimple_statement_resx>::test): Convert to...
958 (is_a_helper <const gimple_statement_resx *>::test): ...this.
959 (is_a_helper <const gimple_statement_eh_dispatch>::test):
961 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
963 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
965 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
966 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
968 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
970 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
972 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
974 (is_a_helper <const gimple_statement_omp_return>::test): Convert
976 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
977 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
979 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
980 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
982 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
983 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
984 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
985 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
987 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
988 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
990 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
991 (is_a_helper <const gimple_statement_omp_target>::test): Convert
993 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
994 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
996 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
997 (is_a_helper <const gimple_statement_omp_single>::test): Convert
999 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1000 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1002 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1003 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1005 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1006 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1007 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1008 (is_a_helper <const gimple_statement_transaction>::test): Convert
1010 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1011 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1013 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1014 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1015 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1016 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1018 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1020 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1022 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1024 (gimple_use_ops): Update for removal of implicit pointer from the
1026 (gimple_set_use_ops): Likewise.
1027 (gimple_vuse): Likewise.
1028 (gimple_vdef): Likewise.
1029 (gimple_vuse_ptr): Likewise.
1030 (gimple_vdef_ptr): Likewise.
1031 (gimple_set_vuse): Likewise.
1032 (gimple_set_vdef): Likewise.
1033 (gimple_omp_return_set_lhs): Likewise.
1034 (gimple_omp_return_lhs): Likewise.
1035 (gimple_omp_return_lhs_ptr): Likewise.
1036 (gimple_call_fntype): Likewise.
1037 (gimple_call_set_fntype): Likewise.
1038 (gimple_call_set_internal_fn): Likewise.
1039 (gimple_call_use_set): Likewise.
1040 (gimple_call_clobber_set): Likewise.
1041 (gimple_bind_vars): Likewise.
1042 (gimple_bind_set_vars): Likewise.
1043 (gimple_bind_body_ptr): Likewise.
1044 (gimple_bind_set_body): Likewise.
1045 (gimple_bind_add_stmt): Likewise.
1046 (gimple_bind_block): Likewise.
1047 (gimple_bind_set_block): Likewise.
1048 (gimple_asm_ninputs): Likewise.
1049 (gimple_asm_noutputs): Likewise.
1050 (gimple_asm_nclobbers): Likewise.
1051 (gimple_asm_nlabels): Likewise.
1052 (gimple_asm_input_op): Likewise.
1053 (gimple_asm_input_op_ptr): Likewise.
1054 (gimple_asm_output_op): Likewise.
1055 (gimple_asm_output_op_ptr): Likewise.
1056 (gimple_asm_set_output_op): Likewise.
1057 (gimple_asm_clobber_op): Likewise.
1058 (gimple_asm_set_clobber_op): Likewise.
1059 (gimple_asm_label_op): Likewise.
1060 (gimple_asm_set_label_op): Likewise.
1061 (gimple_asm_string): Likewise.
1062 (gimple_catch_types): Likewise.
1063 (gimple_catch_types_ptr): Likewise.
1064 (gimple_catch_handler_ptr): Likewise.
1065 (gimple_catch_set_types): Likewise.
1066 (gimple_catch_set_handler): Likewise.
1067 (gimple_eh_filter_types): Likewise.
1068 (gimple_eh_filter_types_ptr): Likewise.
1069 (gimple_eh_filter_failure_ptr): Likewise.
1070 (gimple_eh_filter_set_types): Likewise.
1071 (gimple_eh_filter_set_failure): Likewise.
1072 (gimple_eh_must_not_throw_fndecl): Likewise.
1073 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1074 (gimple_eh_else_n_body_ptr): Likewise.
1075 (gimple_eh_else_e_body_ptr): Likewise.
1076 (gimple_eh_else_set_n_body): Likewise.
1077 (gimple_eh_else_set_e_body): Likewise.
1078 (gimple_try_eval_ptr): Likewise.
1079 (gimple_try_cleanup_ptr): Likewise.
1080 (gimple_try_set_eval): Likewise.
1081 (gimple_try_set_cleanup): Likewise.
1082 (gimple_wce_cleanup_ptr): Likewise.
1083 (gimple_wce_set_cleanup): Likewise.
1084 (gimple_phi_capacity): Likewise.
1085 (gimple_phi_num_args): Likewise.
1086 (gimple_phi_result): Likewise.
1087 (gimple_phi_result_ptr): Likewise.
1088 (gimple_phi_set_result): Likewise.
1089 (gimple_phi_arg): Likewise.
1090 (gimple_phi_set_arg): Likewise.
1091 (gimple_resx_region): Likewise.
1092 (gimple_resx_set_region): Likewise.
1093 (gimple_eh_dispatch_region): Likewise.
1094 (gimple_eh_dispatch_set_region): Likewise.
1095 (gimple_omp_critical_name): Likewise.
1096 (gimple_omp_critical_name_ptr): Likewise.
1097 (gimple_omp_critical_set_name): Likewise.
1098 (gimple_omp_for_clauses): Likewise.
1099 (gimple_omp_for_clauses_ptr): Likewise.
1100 (gimple_omp_for_set_clauses): Likewise.
1101 (gimple_omp_for_collapse): Likewise.
1102 (gimple_omp_for_index): Likewise.
1103 (gimple_omp_for_index_ptr): Likewise.
1104 (gimple_omp_for_set_index): Likewise.
1105 (gimple_omp_for_initial): Likewise.
1106 (gimple_omp_for_initial_ptr): Likewise.
1107 (gimple_omp_for_set_initial): Likewise.
1108 (gimple_omp_for_final): Likewise.
1109 (gimple_omp_for_final_ptr): Likewise.
1110 (gimple_omp_for_set_final): Likewise.
1111 (gimple_omp_for_incr): Likewise.
1112 (gimple_omp_for_incr_ptr): Likewise.
1113 (gimple_omp_for_set_incr): Likewise.
1114 (gimple_omp_for_pre_body_ptr): Likewise.
1115 (gimple_omp_for_set_pre_body): Likewise.
1116 (gimple_omp_parallel_clauses): Likewise.
1117 (gimple_omp_parallel_clauses_ptr): Likewise.
1118 (gimple_omp_parallel_set_clauses): Likewise.
1119 (gimple_omp_parallel_child_fn): Likewise.
1120 (gimple_omp_parallel_child_fn_ptr): Likewise.
1121 (gimple_omp_parallel_set_child_fn): Likewise.
1122 (gimple_omp_parallel_data_arg): Likewise.
1123 (gimple_omp_parallel_data_arg_ptr): Likewise.
1124 (gimple_omp_parallel_set_data_arg): Likewise.
1125 (gimple_omp_task_clauses): Likewise.
1126 (gimple_omp_task_clauses_ptr): Likewise.
1127 (gimple_omp_task_set_clauses): Likewise.
1128 (gimple_omp_task_child_fn): Likewise.
1129 (gimple_omp_task_child_fn_ptr): Likewise.
1130 (gimple_omp_task_set_child_fn): Likewise.
1131 (gimple_omp_task_data_arg): Likewise.
1132 (gimple_omp_task_data_arg_ptr): Likewise.
1133 (gimple_omp_task_set_data_arg): Likewise.
1134 (gimple_omp_taskreg_clauses): Likewise.
1135 (gimple_omp_taskreg_clauses_ptr): Likewise.
1136 (gimple_omp_taskreg_set_clauses): Likewise.
1137 (gimple_omp_taskreg_child_fn): Likewise.
1138 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1139 (gimple_omp_taskreg_set_child_fn): Likewise.
1140 (gimple_omp_taskreg_data_arg): Likewise.
1141 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1142 (gimple_omp_taskreg_set_data_arg): Likewise.
1143 (gimple_omp_task_copy_fn): Likewise.
1144 (gimple_omp_task_copy_fn_ptr): Likewise.
1145 (gimple_omp_task_set_copy_fn): Likewise.
1146 (gimple_omp_task_arg_size): Likewise.
1147 (gimple_omp_task_arg_size_ptr): Likewise.
1148 (gimple_omp_task_set_arg_size): Likewise.
1149 (gimple_omp_task_arg_align): Likewise.
1150 (gimple_omp_task_arg_align_ptr): Likewise.
1151 (gimple_omp_task_set_arg_align): Likewise.
1152 (gimple_omp_single_clauses): Likewise.
1153 (gimple_omp_single_clauses_ptr): Likewise.
1154 (gimple_omp_single_set_clauses): Likewise.
1155 (gimple_omp_target_clauses): Likewise.
1156 (gimple_omp_target_clauses_ptr): Likewise.
1157 (gimple_omp_target_set_clauses): Likewise.
1158 (gimple_omp_target_child_fn): Likewise.
1159 (gimple_omp_target_child_fn_ptr): Likewise.
1160 (gimple_omp_target_set_child_fn): Likewise.
1161 (gimple_omp_target_data_arg): Likewise.
1162 (gimple_omp_target_data_arg_ptr): Likewise.
1163 (gimple_omp_target_set_data_arg): Likewise.
1164 (gimple_omp_teams_clauses): Likewise.
1165 (gimple_omp_teams_clauses_ptr): Likewise.
1166 (gimple_omp_teams_set_clauses): Likewise.
1167 (gimple_omp_sections_clauses): Likewise.
1168 (gimple_omp_sections_clauses_ptr): Likewise.
1169 (gimple_omp_sections_set_clauses): Likewise.
1170 (gimple_omp_sections_control): Likewise.
1171 (gimple_omp_sections_control_ptr): Likewise.
1172 (gimple_omp_sections_set_control): Likewise.
1173 (gimple_omp_for_set_cond): Likewise.
1174 (gimple_omp_for_cond): Likewise.
1175 (gimple_omp_atomic_store_set_val): Likewise.
1176 (gimple_omp_atomic_store_val): Likewise.
1177 (gimple_omp_atomic_store_val_ptr): Likewise.
1178 (gimple_omp_atomic_load_set_lhs): Likewise.
1179 (gimple_omp_atomic_load_lhs): Likewise.
1180 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1181 (gimple_omp_atomic_load_set_rhs): Likewise.
1182 (gimple_omp_atomic_load_rhs): Likewise.
1183 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1184 (gimple_omp_continue_control_def): Likewise.
1185 (gimple_omp_continue_control_def_ptr): Likewise.
1186 (gimple_omp_continue_set_control_def): Likewise.
1187 (gimple_omp_continue_control_use): Likewise.
1188 (gimple_omp_continue_control_use_ptr): Likewise.
1189 (gimple_omp_continue_set_control_use): Likewise.
1190 (gimple_transaction_body_ptr): Likewise.
1191 (gimple_transaction_label): Likewise.
1192 (gimple_transaction_label_ptr): Likewise.
1193 (gimple_transaction_set_body): Likewise.
1194 (gimple_transaction_set_label): Likewise.
1196 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1197 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1198 * ipa-ref.c (ipa_record_reference): Likewise.
1199 * ipa-reference.c (analyze_function): Likewise.
1200 (ipa_reference_write_optimization_summary): Likewise.
1201 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1202 (address_taken_from_non_vtable_p): Likewise.
1203 (comdat_can_be_unshared_p_1): Likewise.
1204 * lto-cgraph.c (lto_output_ref): Likewise.
1205 (add_references): Likewise.
1206 (compute_ltrans_boundary): Likewise.
1207 (output_symtab): Likewise.
1208 (input_ref): Likewise.
1209 (input_cgraph_1): Likewise.
1210 (output_cgraph_opt_summary): Likewise.
1211 * lto-streamer-out.c (lto_output): Likewise.
1212 (output_symbol_p): Likewise.
1213 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1214 (lsei_start_function_in_partition): Likewise.
1215 (lsei_next_variable_in_partition): Likewise.
1216 (lsei_start_variable_in_partition): Likewise.
1217 * symtab.c (insert_to_assembler_name_hash): Likewise.
1218 (unlink_from_assembler_name_hash): Likewise.
1219 (symtab_unregister_node): Likewise.
1220 (symtab_remove_node): Likewise.
1221 (dump_symtab_node): Likewise.
1222 (verify_symtab_base): Likewise.
1223 (verify_symtab_node): Likewise.
1224 (symtab_make_decl_local): Likewise.
1225 (symtab_alias_ultimate_target): Likewise.
1226 (symtab_resolve_alias): Likewise.
1227 (symtab_get_symbol_partitioning_class): Likewise.
1228 * tree-phinodes.c (allocate_phi_node): Likewise.
1229 (reserve_phi_args_for_new_edge): Likewise.
1230 (remove_phi_args): Likewise.
1231 * varpool.c (varpool_node_for_asm): Likewise.
1232 (varpool_remove_unreferenced_decls): Likewise.
1234 2014-04-23 Jeff Law <law@redhat.com>
1236 PR tree-optimization/60902
1237 * tree-ssa-threadedge.c
1238 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1239 invalidate outputs from statements that do not produce useful
1240 outputs for threading.
1242 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1244 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1245 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1246 machine descriptions for Stack Smashing Protector.
1248 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1250 * aarch64.md (<optab>_rol<mode>3): New pattern.
1251 (<optab>_rolsi3_uxtw): Likewise.
1252 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1254 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1256 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1257 (arm_cortex_a12_tune): Likewise.
1259 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1261 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1263 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1265 * config/arm/arm.md (arm_rev16si2): New pattern.
1266 (arm_rev16si2_alt): Likewise.
1267 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1269 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1271 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1272 (rev16<mode>2_alt): Likewise.
1273 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1274 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1275 (aarch_rev16_shleft_mask_imm_p): Likewise.
1276 (aarch_rev16_p_1): Likewise.
1277 (aarch_rev16_p): Likewise.
1278 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1279 (aarch_rev16_shright_mask_imm_p): Likewise.
1280 (aarch_rev16_shleft_mask_imm_p): Likewise.
1282 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1284 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1285 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1287 (cortex_a53_extra_costs): Likewise.
1288 (cortex_a57_extra_costs): Likewise.
1289 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1290 (cortexa7_extra_costs): Likewise.
1291 (cortexa8_extra_costs): Likewise.
1292 (cortexa12_extra_costs): Likewise.
1293 (cortexa15_extra_costs): Likewise.
1294 (v7m_extra_costs): Likewise.
1295 (arm_new_rtx_costs): Handle BSWAP.
1297 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1299 * config/arm/arm.c (cortexa8_extra_costs): New table.
1300 (arm_cortex_a8_tune): New tuning struct.
1301 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1303 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1305 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1307 2014-04-23 Richard Biener <rguenther@suse.de>
1309 * Makefile.in (OBJS): Remove loop-unswitch.o.
1310 * loop-unswitch.c: Delete.
1311 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1312 * passes.def (pass_rtl_unswitch): Likewise.
1313 * loop-init.c (gate_rtl_unswitch): Likewise.
1314 (rtl_unswitch): Likewise.
1315 (pass_data_rtl_unswitch): Likewise.
1316 (pass_rtl_unswitch): Likewise.
1317 (make_pass_rtl_unswitch): Likewise.
1318 * rtl.h (reversed_condition): Likewise.
1319 (compare_and_jump_seq): Likewise.
1320 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1322 * loop-unroll.c (compare_and_jump_seq): Likewise.
1324 2014-04-23 Richard Biener <rguenther@suse.de>
1326 PR tree-optimization/60903
1327 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1328 commented code block.
1329 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1330 loop flags to newly created BBs and edges.
1332 2014-04-23 Nick Clifton <nickc@redhat.com>
1334 * config/msp430/msp430.c (msp430_handle_option): Move function
1336 (msp430_option_override): Simplify mcu and mcpu option handling.
1337 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1338 support for -mhwmult command line option.
1339 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1340 -mhwmult command line option.
1341 (msp430_hwmult_enabled): Delete.
1342 (msp43o_output_labelref): Add support for -mhwmult command line
1344 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1345 (umulsidi3): Likewise.
1346 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1347 (mcpu, mlarge, msmall): Likewise.
1348 (mhwmult): New option.
1349 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1351 (msp430_is_f5_mcu): Remove prototype.
1352 (msp430_use_f5_series_hwmult): Add prototype.
1353 * config/msp430/msp430-opts.h: New file.
1354 * common/config/msp430: New directory.
1355 * common/config/msp430/msp430-common.c: New file.
1356 * config.gcc (msp430): Remove target_has_targetm_common.
1357 * doc/invoke.texi: Document -mhwmult command line option.
1359 2014-04-23 Nick Clifton <nickc@redhat.com>
1361 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1362 default-manifest.o if it can be found in the search path.
1363 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1365 2014-04-23 Terry Guo <terry.guo@arm.com>
1367 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1369 2014-04-23 Richard Biener <rguenther@suse.de>
1372 * tree-inline.c (declare_return_variable): Use mark_addressable.
1374 2014-04-23 Richard Biener <rguenther@suse.de>
1377 * loop-init.c (loop_optimizer_init): Make sure to apply
1378 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1380 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1383 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1385 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1386 if flag_sanitize_undefined_trap_on_error.
1387 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1388 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1389 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1390 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1391 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1392 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1393 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1394 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1395 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1396 * ubsan.c (ubsan_instrument_unreachable): Return
1397 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1398 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1399 if flag_sanitize_undefined_trap_on_error and
1400 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1401 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1402 instrument_bool_enum_load): Emit __builtin_trap () if
1403 flag_sanitize_undefined_trap_on_error and
1404 __builtin_handle_*_abort () if !flag_sanitize_recover.
1405 * doc/invoke.texi (-fsanitize-recover,
1406 -fsanitize-undefined-trap-on-error): Document.
1408 2014-04-22 Christian Bruel <christian.bruel@st.com>
1410 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1411 Force immediates to SImode.
1413 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1415 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1417 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1418 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1419 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1420 (nios2_fpu_insn): Add entry for round.
1421 (N2FPU_NO_ERRNO_P): Define.
1422 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1424 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1426 2014-04-22 Richard Henderson <rth@redhat.com>
1428 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1429 (add<GPI>3_compare0): Remove leading * from name.
1430 (add<GPI>3_carryin): Likewise.
1431 (sub<GPI>3_compare0): Likewise.
1432 (sub<GPI>3_carryin): Likewise.
1433 (<su_optab>mulditi3): New expander.
1434 (multi3): New expander.
1435 (madd<GPI>): Remove leading * from name.
1437 2014-04-22 Martin Jambor <mjambor@suse.cz>
1439 * cgraphclones.c (cgraph_function_versioning): Copy
1440 ipa_transforms_to_apply instead of asserting it is empty.
1442 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1445 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1446 on count_exp to get mode.
1448 2014-04-22 Andrew Pinski <apinski@cavium.com>
1450 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1451 Handle TLS for ILP32.
1452 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1453 (tlsie_small_<mode>): this and handle PTR.
1454 (tlsie_small_sidi): New pattern.
1455 (tlsle_small): Change to an expand to handle ILP32.
1456 (tlsle_small_<mode>): New pattern.
1457 (tlsdesc_small): Rename to ...
1458 (tlsdesc_small_<mode>): this and handle PTR.
1460 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1462 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1464 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1466 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1467 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1468 (aarch64_types_signed_poly_qualifiers): Likewise.
1469 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1470 (aarch64_types_poly_signed_qualifiers): Likewise.
1471 (TYPES_REINTERP_SS): Type macro added.
1472 (TYPES_REINTERP_SU): Likewise.
1473 (TYPES_REINTERP_SP): Likewise.
1474 (TYPES_REINTERP_US): Likewise.
1475 (TYPES_REINTERP_PS): Likewise.
1476 (aarch64_fold_builtin): New expression folding added.
1477 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1478 Declarations removed.
1479 (REINTERP_SS): Declarations added.
1480 (REINTERP_US): Likewise.
1481 (REINTERP_PS): Likewise.
1482 (REINTERP_SU): Likewise.
1483 (REINTERP_SP): Likewise.
1484 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1485 (vreinterpretq_p8_f64): Likewise.
1486 (vreinterpret_p16_f64): Likewise.
1487 (vreinterpretq_p16_f64): Likewise.
1488 (vreinterpret_f32_f64): Likewise.
1489 (vreinterpretq_f32_f64): Likewise.
1490 (vreinterpret_f64_f32): Likewise.
1491 (vreinterpret_f64_p8): Likewise.
1492 (vreinterpret_f64_p16): Likewise.
1493 (vreinterpret_f64_s8): Likewise.
1494 (vreinterpret_f64_s16): Likewise.
1495 (vreinterpret_f64_s32): Likewise.
1496 (vreinterpret_f64_s64): Likewise.
1497 (vreinterpret_f64_u8): Likewise.
1498 (vreinterpret_f64_u16): Likewise.
1499 (vreinterpret_f64_u32): Likewise.
1500 (vreinterpret_f64_u64): Likewise.
1501 (vreinterpretq_f64_f32): Likewise.
1502 (vreinterpretq_f64_p8): Likewise.
1503 (vreinterpretq_f64_p16): Likewise.
1504 (vreinterpretq_f64_s8): Likewise.
1505 (vreinterpretq_f64_s16): Likewise.
1506 (vreinterpretq_f64_s32): Likewise.
1507 (vreinterpretq_f64_s64): Likewise.
1508 (vreinterpretq_f64_u8): Likewise.
1509 (vreinterpretq_f64_u16): Likewise.
1510 (vreinterpretq_f64_u32): Likewise.
1511 (vreinterpretq_f64_u64): Likewise.
1512 (vreinterpret_s64_f64): Likewise.
1513 (vreinterpretq_s64_f64): Likewise.
1514 (vreinterpret_u64_f64): Likewise.
1515 (vreinterpretq_u64_f64): Likewise.
1516 (vreinterpret_s8_f64): Likewise.
1517 (vreinterpretq_s8_f64): Likewise.
1518 (vreinterpret_s16_f64): Likewise.
1519 (vreinterpretq_s16_f64): Likewise.
1520 (vreinterpret_s32_f64): Likewise.
1521 (vreinterpretq_s32_f64): Likewise.
1522 (vreinterpret_u8_f64): Likewise.
1523 (vreinterpretq_u8_f64): Likewise.
1524 (vreinterpret_u16_f64): Likewise.
1525 (vreinterpretq_u16_f64): Likewise.
1526 (vreinterpret_u32_f64): Likewise.
1527 (vreinterpretq_u32_f64): Likewise.
1529 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1531 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1532 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1533 (vreinterpret_p8_s8): Likewise.
1534 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1535 (vreinterpret_p8_s16): Likewise.
1536 (vreinterpret_p8_s32): Likewise.
1537 (vreinterpret_p8_s64): Likewise.
1538 (vreinterpret_p8_f32): Likewise.
1539 (vreinterpret_p8_u8): Likewise.
1540 (vreinterpret_p8_u16): Likewise.
1541 (vreinterpret_p8_u32): Likewise.
1542 (vreinterpret_p8_u64): Likewise.
1543 (vreinterpret_p8_p16): Likewise.
1544 (vreinterpretq_p8_s8): Likewise.
1545 (vreinterpretq_p8_s16): Likewise.
1546 (vreinterpretq_p8_s32): Likewise.
1547 (vreinterpretq_p8_s64): Likewise.
1548 (vreinterpretq_p8_f32): Likewise.
1549 (vreinterpretq_p8_u8): Likewise.
1550 (vreinterpretq_p8_u16): Likewise.
1551 (vreinterpretq_p8_u32): Likewise.
1552 (vreinterpretq_p8_u64): Likewise.
1553 (vreinterpretq_p8_p16): Likewise.
1554 (vreinterpret_p16_s8): Likewise.
1555 (vreinterpret_p16_s16): Likewise.
1556 (vreinterpret_p16_s32): Likewise.
1557 (vreinterpret_p16_s64): Likewise.
1558 (vreinterpret_p16_f32): Likewise.
1559 (vreinterpret_p16_u8): Likewise.
1560 (vreinterpret_p16_u16): Likewise.
1561 (vreinterpret_p16_u32): Likewise.
1562 (vreinterpret_p16_u64): Likewise.
1563 (vreinterpret_p16_p8): Likewise.
1564 (vreinterpretq_p16_s8): Likewise.
1565 (vreinterpretq_p16_s16): Likewise.
1566 (vreinterpretq_p16_s32): Likewise.
1567 (vreinterpretq_p16_s64): Likewise.
1568 (vreinterpretq_p16_f32): Likewise.
1569 (vreinterpretq_p16_u8): Likewise.
1570 (vreinterpretq_p16_u16): Likewise.
1571 (vreinterpretq_p16_u32): Likewise.
1572 (vreinterpretq_p16_u64): Likewise.
1573 (vreinterpretq_p16_p8): Likewise.
1574 (vreinterpret_f32_s8): Likewise.
1575 (vreinterpret_f32_s16): Likewise.
1576 (vreinterpret_f32_s32): Likewise.
1577 (vreinterpret_f32_s64): Likewise.
1578 (vreinterpret_f32_u8): Likewise.
1579 (vreinterpret_f32_u16): Likewise.
1580 (vreinterpret_f32_u32): Likewise.
1581 (vreinterpret_f32_u64): Likewise.
1582 (vreinterpret_f32_p8): Likewise.
1583 (vreinterpret_f32_p16): Likewise.
1584 (vreinterpretq_f32_s8): Likewise.
1585 (vreinterpretq_f32_s16): Likewise.
1586 (vreinterpretq_f32_s32): Likewise.
1587 (vreinterpretq_f32_s64): Likewise.
1588 (vreinterpretq_f32_u8): Likewise.
1589 (vreinterpretq_f32_u16): Likewise.
1590 (vreinterpretq_f32_u32): Likewise.
1591 (vreinterpretq_f32_u64): Likewise.
1592 (vreinterpretq_f32_p8): Likewise.
1593 (vreinterpretq_f32_p16): Likewise.
1594 (vreinterpret_s64_s8): Likewise.
1595 (vreinterpret_s64_s16): Likewise.
1596 (vreinterpret_s64_s32): Likewise.
1597 (vreinterpret_s64_f32): Likewise.
1598 (vreinterpret_s64_u8): Likewise.
1599 (vreinterpret_s64_u16): Likewise.
1600 (vreinterpret_s64_u32): Likewise.
1601 (vreinterpret_s64_u64): Likewise.
1602 (vreinterpret_s64_p8): Likewise.
1603 (vreinterpret_s64_p16): Likewise.
1604 (vreinterpretq_s64_s8): Likewise.
1605 (vreinterpretq_s64_s16): Likewise.
1606 (vreinterpretq_s64_s32): Likewise.
1607 (vreinterpretq_s64_f32): Likewise.
1608 (vreinterpretq_s64_u8): Likewise.
1609 (vreinterpretq_s64_u16): Likewise.
1610 (vreinterpretq_s64_u32): Likewise.
1611 (vreinterpretq_s64_u64): Likewise.
1612 (vreinterpretq_s64_p8): Likewise.
1613 (vreinterpretq_s64_p16): Likewise.
1614 (vreinterpret_u64_s8): Likewise.
1615 (vreinterpret_u64_s16): Likewise.
1616 (vreinterpret_u64_s32): Likewise.
1617 (vreinterpret_u64_s64): Likewise.
1618 (vreinterpret_u64_f32): Likewise.
1619 (vreinterpret_u64_u8): Likewise.
1620 (vreinterpret_u64_u16): Likewise.
1621 (vreinterpret_u64_u32): Likewise.
1622 (vreinterpret_u64_p8): Likewise.
1623 (vreinterpret_u64_p16): Likewise.
1624 (vreinterpretq_u64_s8): Likewise.
1625 (vreinterpretq_u64_s16): Likewise.
1626 (vreinterpretq_u64_s32): Likewise.
1627 (vreinterpretq_u64_s64): Likewise.
1628 (vreinterpretq_u64_f32): Likewise.
1629 (vreinterpretq_u64_u8): Likewise.
1630 (vreinterpretq_u64_u16): Likewise.
1631 (vreinterpretq_u64_u32): Likewise.
1632 (vreinterpretq_u64_p8): Likewise.
1633 (vreinterpretq_u64_p16): Likewise.
1634 (vreinterpret_s8_s16): Likewise.
1635 (vreinterpret_s8_s32): Likewise.
1636 (vreinterpret_s8_s64): Likewise.
1637 (vreinterpret_s8_f32): Likewise.
1638 (vreinterpret_s8_u8): Likewise.
1639 (vreinterpret_s8_u16): Likewise.
1640 (vreinterpret_s8_u32): Likewise.
1641 (vreinterpret_s8_u64): Likewise.
1642 (vreinterpret_s8_p8): Likewise.
1643 (vreinterpret_s8_p16): Likewise.
1644 (vreinterpretq_s8_s16): Likewise.
1645 (vreinterpretq_s8_s32): Likewise.
1646 (vreinterpretq_s8_s64): Likewise.
1647 (vreinterpretq_s8_f32): Likewise.
1648 (vreinterpretq_s8_u8): Likewise.
1649 (vreinterpretq_s8_u16): Likewise.
1650 (vreinterpretq_s8_u32): Likewise.
1651 (vreinterpretq_s8_u64): Likewise.
1652 (vreinterpretq_s8_p8): Likewise.
1653 (vreinterpretq_s8_p16): Likewise.
1654 (vreinterpret_s16_s8): Likewise.
1655 (vreinterpret_s16_s32): Likewise.
1656 (vreinterpret_s16_s64): Likewise.
1657 (vreinterpret_s16_f32): Likewise.
1658 (vreinterpret_s16_u8): Likewise.
1659 (vreinterpret_s16_u16): Likewise.
1660 (vreinterpret_s16_u32): Likewise.
1661 (vreinterpret_s16_u64): Likewise.
1662 (vreinterpret_s16_p8): Likewise.
1663 (vreinterpret_s16_p16): Likewise.
1664 (vreinterpretq_s16_s8): Likewise.
1665 (vreinterpretq_s16_s32): Likewise.
1666 (vreinterpretq_s16_s64): Likewise.
1667 (vreinterpretq_s16_f32): Likewise.
1668 (vreinterpretq_s16_u8): Likewise.
1669 (vreinterpretq_s16_u16): Likewise.
1670 (vreinterpretq_s16_u32): Likewise.
1671 (vreinterpretq_s16_u64): Likewise.
1672 (vreinterpretq_s16_p8): Likewise.
1673 (vreinterpretq_s16_p16): Likewise.
1674 (vreinterpret_s32_s8): Likewise.
1675 (vreinterpret_s32_s16): Likewise.
1676 (vreinterpret_s32_s64): Likewise.
1677 (vreinterpret_s32_f32): Likewise.
1678 (vreinterpret_s32_u8): Likewise.
1679 (vreinterpret_s32_u16): Likewise.
1680 (vreinterpret_s32_u32): Likewise.
1681 (vreinterpret_s32_u64): Likewise.
1682 (vreinterpret_s32_p8): Likewise.
1683 (vreinterpret_s32_p16): Likewise.
1684 (vreinterpretq_s32_s8): Likewise.
1685 (vreinterpretq_s32_s16): Likewise.
1686 (vreinterpretq_s32_s64): Likewise.
1687 (vreinterpretq_s32_f32): Likewise.
1688 (vreinterpretq_s32_u8): Likewise.
1689 (vreinterpretq_s32_u16): Likewise.
1690 (vreinterpretq_s32_u32): Likewise.
1691 (vreinterpretq_s32_u64): Likewise.
1692 (vreinterpretq_s32_p8): Likewise.
1693 (vreinterpretq_s32_p16): Likewise.
1694 (vreinterpret_u8_s8): Likewise.
1695 (vreinterpret_u8_s16): Likewise.
1696 (vreinterpret_u8_s32): Likewise.
1697 (vreinterpret_u8_s64): Likewise.
1698 (vreinterpret_u8_f32): Likewise.
1699 (vreinterpret_u8_u16): Likewise.
1700 (vreinterpret_u8_u32): Likewise.
1701 (vreinterpret_u8_u64): Likewise.
1702 (vreinterpret_u8_p8): Likewise.
1703 (vreinterpret_u8_p16): Likewise.
1704 (vreinterpretq_u8_s8): Likewise.
1705 (vreinterpretq_u8_s16): Likewise.
1706 (vreinterpretq_u8_s32): Likewise.
1707 (vreinterpretq_u8_s64): Likewise.
1708 (vreinterpretq_u8_f32): Likewise.
1709 (vreinterpretq_u8_u16): Likewise.
1710 (vreinterpretq_u8_u32): Likewise.
1711 (vreinterpretq_u8_u64): Likewise.
1712 (vreinterpretq_u8_p8): Likewise.
1713 (vreinterpretq_u8_p16): Likewise.
1714 (vreinterpret_u16_s8): Likewise.
1715 (vreinterpret_u16_s16): Likewise.
1716 (vreinterpret_u16_s32): Likewise.
1717 (vreinterpret_u16_s64): Likewise.
1718 (vreinterpret_u16_f32): Likewise.
1719 (vreinterpret_u16_u8): Likewise.
1720 (vreinterpret_u16_u32): Likewise.
1721 (vreinterpret_u16_u64): Likewise.
1722 (vreinterpret_u16_p8): Likewise.
1723 (vreinterpret_u16_p16): Likewise.
1724 (vreinterpretq_u16_s8): Likewise.
1725 (vreinterpretq_u16_s16): Likewise.
1726 (vreinterpretq_u16_s32): Likewise.
1727 (vreinterpretq_u16_s64): Likewise.
1728 (vreinterpretq_u16_f32): Likewise.
1729 (vreinterpretq_u16_u8): Likewise.
1730 (vreinterpretq_u16_u32): Likewise.
1731 (vreinterpretq_u16_u64): Likewise.
1732 (vreinterpretq_u16_p8): Likewise.
1733 (vreinterpretq_u16_p16): Likewise.
1734 (vreinterpret_u32_s8): Likewise.
1735 (vreinterpret_u32_s16): Likewise.
1736 (vreinterpret_u32_s32): Likewise.
1737 (vreinterpret_u32_s64): Likewise.
1738 (vreinterpret_u32_f32): Likewise.
1739 (vreinterpret_u32_u8): Likewise.
1740 (vreinterpret_u32_u16): Likewise.
1741 (vreinterpret_u32_u64): Likewise.
1742 (vreinterpret_u32_p8): Likewise.
1743 (vreinterpret_u32_p16): Likewise.
1744 (vreinterpretq_u32_s8): Likewise.
1745 (vreinterpretq_u32_s16): Likewise.
1746 (vreinterpretq_u32_s32): Likewise.
1747 (vreinterpretq_u32_s64): Likewise.
1748 (vreinterpretq_u32_f32): Likewise.
1749 (vreinterpretq_u32_u8): Likewise.
1750 (vreinterpretq_u32_u16): Likewise.
1751 (vreinterpretq_u32_u64): Likewise.
1752 (vreinterpretq_u32_p8): Likewise.
1753 (vreinterpretq_u32_p16): Likewise.
1755 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1757 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1759 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1762 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1763 (vqnegd_s64): Likewise.
1764 (vqabs_s64): Likewise.
1765 (vqabsd_s64): Likewise.
1767 2014-04-22 Richard Henderson <rth@redhat.com>
1769 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1770 computation to the top of the loop.
1772 2014-04-22 Renlin <renlin.li@arm.com>
1773 Jiong Wang <jiong.wang@arm.com>
1775 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1776 * config/aarch64/aarch64.c (aarch64_layout_frame)
1777 (aarch64_initial_elimination_offset): Likewise.
1779 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1781 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1784 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1786 * machmode.h (bitwise_mode_for_mode): Declare.
1787 * stor-layout.h (bitwise_type_for_mode): Likewise.
1788 * stor-layout.c (bitwise_mode_for_mode): New function.
1789 (bitwise_type_for_mode): Likewise.
1790 * builtins.c (fold_builtin_memory_op): Use it instead of
1791 int_mode_for_mode and build_nonstandard_integer_type.
1793 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1795 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1796 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1797 (*-*-solaris2*): Simplify.
1798 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1799 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1800 *-*-solaris2.9* handling.
1802 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1804 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1805 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1807 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1808 * configure: Regenerate.
1810 * config/i386/sol2-9.h: Remove.
1812 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1813 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1814 Remove Solaris 9 references.
1816 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1818 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1819 (floatuns<GPI:mode><GPF:mode>2): Remove.
1820 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1821 and floatuns conversions.
1822 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1823 and floatuns conversions.
1824 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1825 (w1,w2): New mode attributes for inequal width conversions.
1827 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1829 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1830 the output asm format.
1832 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1834 * config/aarch64/aarch64-simd.md
1835 (aarch64_cm<optab>di): Always split.
1836 (*aarch64_cm<optab>di): New.
1837 (aarch64_cmtstdi): Always split.
1838 (*aarch64_cmtstdi): New.
1840 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1842 PR tree-optimization/60823
1843 * omp-low.c (ipa_simd_modify_function_body): Go through
1844 all SSA_NAMEs and for those refering to vector arguments
1845 which are going to be replaced adjust SSA_NAME_VAR and,
1846 if it is a default definition, change it into a non-default
1847 definition assigned at the beginning of function from new_decl.
1848 (ipa_simd_modify_stmt_ops): Rewritten.
1849 * tree-dfa.c (set_ssa_default_def): When removing default def,
1850 check for NULL loc instead of NULL *loc.
1852 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1854 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1855 restrictions on core registers for DImode values in Thumb2.
1857 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1859 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1860 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1862 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1864 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1865 (*iordi_notzesidi_di): Likewise.
1866 (*iordi_notsesidi_di): Likewise.
1868 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1870 * config/arm/arm-protos.h (tune_params): New struct members.
1871 * config/arm/arm.c: Initialise tune_params per processor.
1872 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1873 for speed, based on new tune_params.
1875 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1877 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1879 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1881 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1883 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1884 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1885 (vrnda_f64): Likewise.
1886 (vrndi_f64): Likewise.
1887 (vrndm_f64): Likewise.
1888 (vrndn_f64): Likewise.
1889 (vrndp_f64): Likewise.
1890 (vrndx_f64): Likewise.
1892 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1894 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1895 GET_MODE_SIZE argument is enum machine_mode.
1897 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1900 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1901 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1903 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1906 * asan.c (asan_emit_stack_protection): Force the base to align to
1907 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1908 appropriate bits if STRICT_ALIGNMENT.
1909 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1911 (expand_used_vars): Leave a space in the stack frame for alignment
1912 if STRICT_ALIGNMENT.
1914 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1916 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1918 (gimple_store_p): Likewise.
1919 (gimple_assign_load_p): Likewise.
1920 (gimple_assign_cast_p): Likewise.
1921 (gimple_clobber_p): Likewise.
1923 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1924 rather than a gimple.
1925 (gimple_assign_cast_p): Likewise.
1927 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1930 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1931 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1933 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1934 more debug information for E500 if -mdebug=reg.
1936 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1939 * config/i386/i386.c (ix86_expand_builtin)
1940 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1941 register for target RTX.
1942 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1944 2014-04-18 Cong Hou <congh@google.com>
1946 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1947 the widen-mult pattern by handling two operands with different sizes,
1948 and operands whose size is smaller than half of the result type.
1950 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1952 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1953 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1954 (do_estimate_edge_time): Compute it.
1955 * ipa-inline.c (want_inline_small_function_p): Bypass
1956 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1958 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1960 * ipa-inline.c (spec_rem): New static variable.
1961 (dump_overall_stats): New function.
1962 (dump_inline_stats): New function.
1964 2014-04-18 Richard Henderson <rth@redhat.com>
1966 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1967 to GET_MODE_SIZE, not a reg_class_t.
1969 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1971 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1972 (vsx_xxmrglw_<mode>): Likewise.
1974 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1977 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1978 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1979 (rs6000_init_hard_regno_mode_ok): Likewise.
1981 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1983 * ipa-inline.c (inline_small_functions): Account only non-cold
1985 * doc/invoke.texi (inline-unit-growth): Update documentation.
1987 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1989 * config/rs6000/rs6000.md (addti3, subti3): New.
1991 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1994 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1995 comment. Check optimize_insn_for_size_p instead of
1996 optimize_insn_for_speed_p.
1998 2014-04-17 Martin Jambor <mjambor@suse.cz>
2000 * gimple-iterator.c (gsi_start_edge): New function.
2001 * gimple-iterator.h (gsi_start_edge): Declare.
2002 * tree-sra.c (single_non_eh_succ): New function.
2003 (disqualify_ops_if_throwing_stmt): Renamed to
2004 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2005 having one non-EH successor BB.
2006 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2007 generate loads into replacements.
2008 (sra_modify_assign): Likewise and and also use the simple path for
2010 (sra_modify_function_body): Commit statements on edges.
2012 2014-04-17 Richard Biener <rguenther@suse.de>
2015 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2016 comparison results and add clarifying comment.
2018 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2020 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2021 (blank_mode): Initialize it.
2022 (emit_mode_size_inline, emit_mode_nunits_inline,
2023 emit_mode_inner_inline): New functions.
2024 (emit_insn_modes_h): Call them and surround their output with
2025 #if GCC_VERSION >= 4001 ... #endif.
2026 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2027 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2028 mode_* arrays if the argument is __builtin_constant_p.
2029 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2030 is enum machine_mode.
2032 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2034 * passes.c (opt_pass::execute): Adjust.
2035 (pass_manager::execute_pass_mode_switching): Likewise.
2036 (early_local_passes::execute): Likewise.
2037 (execute_one_pass): Pass cfun to the pass's execute method.
2038 * tree-pass.h (opt_pass::execute): Add function * argument.
2039 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2040 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2041 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2042 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2043 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2044 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2045 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2046 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2047 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2048 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
2049 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
2050 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
2051 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
2052 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
2053 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2054 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2055 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2056 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
2057 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
2058 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2059 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2060 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2061 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2062 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2063 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2064 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2065 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2066 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2067 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2070 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2072 * passes.c (opt_pass::gate): Take function * argument.
2073 (gate_all_early_local_passes): Merge into
2074 (early_local_passes::gate): this.
2075 (gate_all_early_optimizations): Merge into
2076 (all_early_optimizations::gate): this.
2077 (gate_all_optimizations): Mege into
2078 (all_optimizations::gate): this.
2079 (gate_all_optimizations_g): Merge into
2080 (all_optimizations_g::gate): this.
2081 (gate_rest_of_compilation): Mege into
2082 (rest_of_compilation::gate): this.
2083 (gate_postreload): Merge into
2084 (postreload::gate): this.
2085 (dump_one_pass): Pass cfun to the pass's gate method.
2086 (execute_ipa_summary_passes): Likewise.
2087 (execute_one_pass): Likewise.
2088 (ipa_write_summaries_2): Likewise.
2089 (ipa_write_optimization_summaries_1): Likewise.
2090 (ipa_read_summaries_1): Likewise.
2091 (ipa_read_optimization_summaries_1): Likewise.
2092 (execute_ipa_stmt_fixups): Likewise.
2093 * tree-pass.h (opt_pass::gate): Add function * argument.
2094 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2095 combine-stack-adj.c, combine.c, compare-elim.c,
2096 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2097 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2098 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2099 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2100 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2101 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2102 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2103 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2104 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2105 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2106 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2107 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2108 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2109 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2110 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2111 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2112 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2113 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2114 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2115 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2116 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2117 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2118 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2119 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2120 var-tracking.c, vtable-verify.c, web.c: Adjust.
2122 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2124 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2125 * configure: Regenerate.
2127 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2129 * passes.c (dump_one_pass): don't check pass->has_gate.
2130 (execute_ipa_summary_passes): Likewise.
2131 (execute_one_pass): Likewise.
2132 (ipa_write_summaries_2): Likewise.
2133 (ipa_write_optimization_summaries_1): Likewise.
2134 (ipa_read_optimization_summaries_1): Likewise.
2135 (execute_ipa_stmt_fixups): Likewise.
2136 * tree-pass.h (pass_data::has_gate): Remove.
2137 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2138 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2139 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2140 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2141 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2142 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2143 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2144 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2145 gimple-low.c, gimple-ssa-isolate-paths.c,
2146 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2147 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2148 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2149 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2150 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2151 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2152 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2153 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2154 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2155 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2156 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2157 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2158 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2159 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2160 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2161 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2162 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2163 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2164 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2165 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2166 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2167 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2170 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2172 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2174 * passes.c (pass_manager::register_dump_files_1): Merge into
2175 (pass_manager::register_dump_files): this, and remove its handling of
2176 properties since the pass always has the properties anyway.
2177 (pass_manager::pass_manager): Adjust.
2179 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2181 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2182 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2183 dealing with properties.
2184 (pass_manager::register_dump_files): Adjust.
2186 2014-03-20 Mark Wielaard <mjw@redhat.com>
2188 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2189 then represent the bound as normal constant value.
2191 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2194 Forward port from 4.8 branch
2195 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2197 * config/i386/bmiintrin.h (_blsi_u32): New.
2201 (_blsmsk_u32): Ditto.
2202 (_blsmsk_u64): Ditto.
2203 (_tzcnt_u32): Ditto.
2204 (_tzcnt_u64): Ditto.
2206 2014-04-17 Kito Cheng <kito@0xlab.org>
2208 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2210 2014-04-17 Richard Biener <rguenther@suse.de>
2213 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2214 boolean results for comparisons.
2216 2014-04-17 Richard Biener <rguenther@suse.de>
2218 PR tree-optimization/60836
2219 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2220 initial PHI args to be gimple values.
2222 2014-04-17 Richard Biener <rguenther@suse.de>
2224 PR tree-optimization/60841
2225 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2226 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2227 of stmts to SLP build.
2228 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2229 (vect_analyze_slp): Likewise.
2230 (vect_analyze_slp_instance): Likewise.
2231 (vect_build_slp_tree): Limit overall SLP tree growth.
2232 * tree-vectorizer.h (vect_analyze_data_refs,
2233 vect_analyze_slp): Adjust prototypes.
2235 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2237 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2240 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2242 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2243 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2244 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2245 for TARGET_SLOW_PSHUFB
2247 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2249 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2250 * config/i386/i386.c (intel_cost): Ditto.
2252 2014-04-17 Joey Ye <joey.ye@arm.com>
2254 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2256 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2258 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2261 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2263 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2264 (type_all_derivations_known_p): New predicate.
2265 (type_all_ctors_visible_p): New predicate.
2266 (type_possibly_instantiated_p): New predicate.
2267 (get_odr_type): Compute all_derivations_known.
2268 (dump_odr_type): Dump the flag.
2269 (maybe_record_type): Cleanup.
2270 (record_target_from_binfo): Add bases_to_consider array;
2271 record bases for types w/o instances and skip CXX destructor.
2272 (possible_polymorphic_call_targets_1): Add bases_to_consider
2273 and consider_construction parameters; check if type may have instance.
2274 (get_polymorphic_call_info): Set maybe_in_construction to true
2275 when we know nothing.
2276 (record_targets_from_bases): Skip CXX destructors; they are
2277 never called for types in construction.
2278 (possible_polymorphic_call_targets): Do not record target when
2279 type may not have instance.
2281 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2284 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2285 external aliases alive, too.
2287 2014-04-16 Andrew Pinski <apinski@cavium.com>
2289 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2292 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2294 * final.c (compute_alignments): Do not apply loop alignment to a block
2295 falling through to the exit.
2297 2014-04-16 Catherine Moore <clm@codesourcery.com>
2299 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2300 Adjust constraints for microMIPS store patterns.
2302 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2304 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2306 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2308 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2309 (append_use): Run at -O0.
2310 (append_vdef): Likewise.
2311 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2312 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2314 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2316 PR tree-optimization/60844
2317 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2318 (propagate_op_to_single_use, remove_visited_stmt_chain,
2319 linearize_expr, repropagate_negates, reassociate_bb): Use it
2320 instead of gsi_remove.
2322 2014-04-16 Martin Jambor <mjambor@suse.cz>
2324 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2325 ipa_transforms_to_apply.
2326 (cgraph_function_versioning): Assert that old_node has empty
2327 ipa_transforms_to_apply.
2328 * trans-mem.c (ipa_tm_create_version): Likewise.
2329 * tree-inline.c (tree_function_versioning): Do not duplicate
2330 ipa_transforms_to_apply.
2332 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2335 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2337 Pass necessary as flags on 64-bit Solaris/x86.
2338 Use lowercase relocs for x86_64-*-*.
2339 * configure: Regenerate.
2341 2014-04-15 Jan Hubicka <jh@suse.cz>
2343 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2344 (maybe_record_node, likely_target_p): Use it.
2346 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2349 Revert following patch
2351 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2354 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2355 software floating point or no floating point registers, do not
2356 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2357 in GPRs that occurs after we tested for GPRs that would never be
2360 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2361 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2362 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2363 specifically allow DDmode, since that does not use the SPE SIMD
2366 2014-03-21 Mark Wielaard <mjw@redhat.com>
2368 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2369 as unsigned or int depending on type and value used.
2371 2014-04-15 Richard Biener <rguenther@suse.de>
2373 PR rtl-optimization/56965
2374 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2375 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2377 * alias.c (true_dependence_1): Do not call
2378 nonoverlapping_component_refs_p.
2379 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2380 nonoverlapping_component_refs_p.
2381 (indirect_refs_may_alias_p): Likewise.
2383 2014-04-15 Teresa Johnson <tejohnson@google.com>
2385 * cfg.c (dump_bb_info): Fix flags check.
2386 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2388 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2390 PR rtl-optimization/60663
2391 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2394 2014-04-15 Richard Biener <rguenther@suse.de>
2396 * lto-streamer.h (LTO_major_version): Bump to 4.
2398 2014-04-15 Richard Biener <rguenther@suse.de>
2400 * common.opt (lto_partition_model): New enum.
2401 (flto-partition=): Merge separate options with a single with argument,
2402 add -flto-partition=one support.
2403 * flag-types.h (enum lto_partition_model): Declare.
2404 * opts.c (finish_options): Remove duplicate -flto-partition=
2406 * lto-wrapper.c (run_gcc): Adjust.
2408 2014-04-15 Richard Biener <rguenther@suse.de>
2410 * alias.c (ncr_compar): New function.
2411 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2413 2014-04-15 Richard Biener <rguenther@suse.de>
2415 * alias.c (record_component_aliases): Do not walk BINFOs.
2417 2014-04-15 Richard Biener <rguenther@suse.de>
2419 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2420 Add struct function argument and adjust.
2421 (find_func_aliases_for_call): Likewise.
2422 (find_func_aliases): Likewise.
2423 (find_func_clobbers): Likewise.
2424 (intra_create_variable_infos): Likewise.
2425 (compute_points_to_sets): Likewise.
2426 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2428 2014-04-15 Richard Biener <rguenther@suse.de>
2430 * tree.c (iterative_hash_expr): Use enum tree_code_class
2431 to store TREE_CODE_CLASS.
2432 (tree_block): Likewise.
2433 (tree_set_block): Likewise.
2434 * tree.h (fold_build_pointer_plus_loc): Use
2435 convert_to_ptrofftype_loc.
2437 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2440 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2443 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2445 * cfgloop.h (struct loop): Move force_vectorize down.
2446 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2447 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2448 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2449 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2450 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2451 * tree-core.h (enum annot_expr_kind): Add new kind values.
2452 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2453 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2455 * tree.def (ANNOTATE_EXPR): Tweak comment.
2457 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2459 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2462 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2464 * tree.h (TYPE_IDENTIFIER): Declare.
2465 * tree.c (subrange_type_for_debug_p): Use it.
2466 * godump.c (go_format_type): Likewise.
2467 * dwarf2out.c (is_cxx_auto, modified_type_die,
2468 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2469 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2471 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2474 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2476 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2478 * config/i386/i386.c (examine_argument): Return bool. Return true if
2479 parameter should be passed in memory.
2480 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2481 (construct_container): Update calls to examine_argument.
2482 (function_arg_advance_64): Ditto.
2483 (return_in_memory_32): Merge with ix86_return_in_memory.
2484 (return_in_memory_64): Ditto.
2485 (return_in_memory_ms_64): Ditto.
2487 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2489 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2490 * coverage.c (coverage_compute_profile_id): Handle externally visible
2493 2014-04-14 Martin Jambor <mjambor@suse.cz>
2495 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2496 DECL_DISREGARD_INLINE_LIMITS functions.
2498 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2501 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2503 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2506 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2507 optimize_insn_for_speed_p instead of
2508 optimize_function_for_speed_p.
2510 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2512 * doc/invoke.texi (free): Document AArch64.
2514 2014-04-14 Richard Biener <rguenther@suse.de>
2516 PR tree-optimization/60042
2517 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2518 (insert_into_preds_of_block): Do not prevent PHI insertion
2519 for REFERENCE exprs here ...
2520 (eliminate_dom_walker::before_dom_children): ... but prevent
2521 their use here under similar conditions when applied to the
2522 IL after PRE optimizations.
2524 2014-04-14 Richard Biener <rguenther@suse.de>
2526 * passes.def: Move early points-to after early SRA.
2528 2014-04-14 Richard Biener <rguenther@suse.de>
2530 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2531 check for which sign-changes we allow when forwarding
2532 a converted value into a switch.
2534 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2536 * stor-layout.c (place_field): Finalize non-constant offset for the
2539 2014-04-14 Richard Biener <rguenther@suse.de>
2541 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2543 (expand_switch_using_bit_tests_p): Likewise.
2544 (process_switch): Compute and pass on speed_p based on the
2546 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2547 optimize_bb_for_speed_p.
2549 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2551 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2552 * function.h (struct function): Rename has_force_vect_loops into
2553 has_force_vectorize_loops.
2554 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2555 (input_struct_function_base): Likewise.
2556 * lto-streamer-out.c (output_cfg): Likewise.
2557 (output_struct_function_base): Likewise.
2558 * omp-low.c (expand_omp_simd): Likewise.
2559 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2560 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2561 (version_loop_for_if_conversion): Likewise.
2562 (tree_if_conversion): Likewise.
2563 (main_tree_if_conversion): Likewise.
2564 (gate_tree_if_conversion): Likewise.
2565 * tree-inline.c (copy_loops): Likewise.
2566 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2567 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2568 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2569 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2570 * tree-vectorizer.c (vectorize_loops): Likewise.
2571 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2573 2014-04-14 Richard Biener <rguenther@suse.de>
2576 * lto-streamer-out.c (wrap_refs): New function.
2577 (lto_output): Wrap symbol references in global initializes in
2578 type-preserving MEM_REFs.
2580 2014-04-14 Christian Bruel <christian.bruel@st.com>
2582 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2584 2014-04-14 Christian Bruel <christian.bruel@st.com>
2586 * config/sh/sh.md (setmemqi): New expand pattern.
2587 * config/sh/sh.h (CLEAR_RATIO): Define.
2588 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2589 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2591 2014-04-14 Richard Biener <rguenther@suse.de>
2594 * fold-const.c (negate_expr_p): Don't negate directional rounding
2596 (fold_negate_expr): Likewise.
2598 2014-04-14 Richard Biener <rguenther@suse.de>
2600 PR tree-optimization/59817
2601 PR tree-optimization/60453
2602 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2603 recursion to catch all CHRECs in the scalar evolution and restrict
2604 the predicate for the remains appropriately.
2606 2014-04-12 Catherine Moore <clm@codesourcery.com>
2608 * config/mips/constraints.md: Add new register constraint "kb".
2609 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2610 (*movhi_internal): Likewise.
2611 (*movqi_internal): Likewise.
2612 * config/mips/mips.h (M16_STORE_REGS): New register class.
2613 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2614 (REG_CLASS_CONTENTS): Likewise.
2615 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2617 2014-04-11 Tobias Burnus <burnus@net-b.de>
2620 * doc/invoke.texi (-Wformat-signedness): Document it.
2621 (Wformat=2): Mention that this enables -Wformat-signedness.
2623 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2625 * common/config/epiphany/epiphany-common.c
2626 (epiphany_option_optimization_table): Enable section anchors by
2627 default at -O1 or higher.
2628 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2629 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2630 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2631 carries no extra cost.
2632 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2633 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2634 * config/epiphany/predicates.md (memclob_operand): New predicate.
2635 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2636 Use memclob_operand predicate and X constraint for operand 3.
2638 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2640 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2641 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2644 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2646 PR rtl-optimization/60651
2647 * mode-switching.c (optimize_mode_switching): Make sure to emit
2648 sets of a lower numbered entity before sets of a higher numbered
2649 entity to a mode of the same or lower priority.
2650 When creating a seginfo for a basic block that starts with a code
2651 label, move the insertion point past the code label.
2652 (new_seginfo): Document and enforce requirement that
2653 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2654 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2655 * doc/tm.texi: Regenerate.
2657 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2660 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2662 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2664 * BASE-VER: Set to 4.10.0.
2666 2014-04-11 Tobias Burnus <burnus@net-b.de>
2669 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2670 * doc/gcc.texi (Service): Update description in the @menu
2671 * doc/invoke.texi (Option Summary): Remove misplaced and
2674 2014-04-11 Steve Ellcey <sellcey@mips.com>
2675 Jakub Jelinek <jakub@redhat.com>
2678 * expr.c (convert_move): Use emit_store_flag_force instead of
2679 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2682 2014-04-11 Richard Biener <rguenther@suse.de>
2685 * varasm.c (assemble_alias): Avoid endless error reporting
2686 recursion by setting TREE_ASM_WRITTEN.
2688 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2690 * config/s390/s390.md: Add a splitter for NOT rtx.
2692 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2694 PR rtl-optimization/60663
2695 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2697 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2698 Jakub Jelinek <jakub@redhat.com>
2701 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2702 flag from decl_node to node.
2704 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2707 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2708 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2709 ameliorating the cases where it can be.
2711 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2714 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2716 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2717 (loadsync_<mode>): Change mode.
2718 (load_quadpti, store_quadpti): New.
2719 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2720 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2721 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2723 2014-04-09 Cong Hou <congh@google.com>
2726 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2729 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2731 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2732 instead of vnor to exploit possible fusion opportunity in the
2734 (altivec_expand_vec_perm_const_le): Likewise.
2736 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2738 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2739 (loadsync_<mode>): Change mode.
2740 (load_quadpti, store_quadpti): New.
2741 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2742 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2744 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2747 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2748 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2749 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2751 2014-04-08 Richard Biener <rguenther@suse.de>
2754 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2755 a 64bit widest int print double-int similar to on HWI64 hosts.
2757 2014-04-08 Richard Biener <rguenther@suse.de>
2759 PR tree-optimization/60785
2760 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2761 default defs properly.
2763 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2765 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2766 (Weffc++): Likewise.
2768 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2770 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2771 set completep to false rather than true.
2773 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2776 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2777 ARM_TARGET2_DWARF_FORMAT.
2779 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2782 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2783 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2786 2014-04-07 Richard Biener <rguenther@suse.de>
2788 PR tree-optimization/60766
2789 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2790 (may_eliminate_iv): Convert cand_value_at result to desired type.
2792 2014-04-07 Jason Merrill <jason@redhat.com>
2795 * common.opt (-fno-gnu-unique): Add.
2796 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2798 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2800 * haifa-sched.c: Fix outdated function reference and minor
2801 grammar errors in introductory comment.
2803 2014-04-07 Richard Biener <rguenther@suse.de>
2806 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2808 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2810 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2813 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2814 size accounting for thunks.
2815 (pa_asm_output_mi_thunk): Use final_start_function() and
2816 final_end_function() to output function start and end directives.
2818 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2820 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2821 device specific ISA/ feature information. Remove short_sp and
2822 errata_skip ds. Add avr_device_specific_features enum to have device
2824 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2825 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2826 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2827 updated device specific info.
2828 * config/avr/avr-mcus.def: Merge device specific details to
2829 dev_attribute field.
2830 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2832 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2833 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2834 assembler if RMW isa supported by current device.
2835 * config/avr/genmultilib.awk: Update as device info structure changed.
2836 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2838 2014-04-04 Cong Hou <congh@google.com>
2840 PR tree-optimization/60656
2841 * tree-vect-stmts.c (supportable_widening_operation):
2842 Fix a bug that elements in a vector with vect_used_by_reduction
2843 property are incorrectly reordered when the operation on it is not
2844 consistant with the one in reduction operation.
2846 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2848 PR rtl-optimization/60155
2849 * gcse.c (record_set_data): New function.
2850 (single_set_gcse): New function.
2851 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2852 (hoist_code): Likewise.
2853 (get_pressure_class_and_nregs): Likewise.
2855 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2857 * explow.c (probe_stack_range): Emit a final optimization blockage.
2859 2014-04-04 Anthony Green <green@moxielogic.com>
2861 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2864 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2867 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2868 flags are set only during streaming.
2869 * ipa.c (process_references, walk_polymorphic_call_targets,
2870 symtab_remove_unreachable_nodes): Drop bodies of always inline
2871 after early inlining.
2872 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2874 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2875 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2878 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2881 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2884 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2886 (cortex_a53_fdivd): Likewise.
2888 2014-04-04 Martin Jambor <mjambor@suse.cz>
2891 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2893 * cgraph.c (clone_of_p): Also return true if thunks match.
2894 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2895 cgraph_function_or_thunk_node and an obsolete comment.
2896 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2898 (build_function_decl_skip_args): Likewise.
2899 (set_new_clone_decl_and_node_flags): New function.
2900 (duplicate_thunk_for_node): Likewise.
2901 (redirect_edge_duplicating_thunks): Likewise.
2902 (cgraph_clone_node): New parameter args_to_skip, pass it to
2903 redirect_edge_duplicating_thunks which is called instead of
2904 cgraph_redirect_edge_callee.
2905 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2906 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2908 2014-04-04 Jeff Law <law@redhat.com>
2911 * config/arm/predicates.md (const_int_I_operand): New predicate.
2912 (const_int_M_operand): Similarly.
2913 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2915 (insv_t2, extv_reg, extzv_t2): Likewise.
2916 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2917 (pop_multiple_with_writeback_and_return): Likewise.
2918 (vfp_pop_multiple_with_writeback): Likewise
2920 2014-04-04 Richard Biener <rguenther@suse.de>
2923 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2924 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2926 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2927 * gimplify.c (gimple_add_tmp_var_fn): New function.
2928 * gimple-expr.h (create_tmp_reg_fn): Declare.
2929 * gimple-expr.c (create_tmp_reg_fn): New function.
2930 * gimple-low.c (record_vars_into): Don't change cfun.
2931 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2932 code generation without cfun.
2934 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2937 * Makefile.in (install-driver): Fix shell scripting.
2939 2014-04-03 Cong Hou <congh@google.com>
2941 PR tree-optimization/60505
2942 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2943 threshold of number of iterations below which no vectorization
2945 * tree-vect-loop.c (new_loop_vec_info):
2946 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2947 * tree-vect-loop.c (vect_analyze_loop_operations):
2948 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2949 * tree-vect-loop.c (vect_transform_loop):
2950 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2951 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2952 of iterations of the loop and see if we should build the epilogue.
2954 2014-04-03 Richard Biener <rguenther@suse.de>
2956 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2957 (streamer_tree_cache_create): Adjust.
2958 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2959 to allow optional nodes array.
2960 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2961 (streamer_tree_cache_append): Likewise.
2962 (streamer_tree_cache_create): Create nodes array optionally
2963 as specified by parameter.
2964 * lto-streamer-out.c (create_output_block): Avoid maintaining
2965 the node array in the writer cache.
2966 (DFS_write_tree): Remove assertion.
2967 (produce_asm_for_decls): Free the out decl state hash table early.
2968 * lto-streamer-in.c (lto_data_in_create): Adjust for
2969 streamer_tree_cache_create prototype change.
2971 2014-04-03 Richard Biener <rguenther@suse.de>
2973 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2974 set TREE_CHAIN to NULL_TREE.
2976 2014-04-03 Richard Biener <rguenther@suse.de>
2978 PR tree-optimization/60740
2979 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2980 over all GIMPLE_COND operands.
2982 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2984 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2985 (Weffc++): Remove Scott's numbering, merge lists and reference
2988 2014-04-03 Nick Clifton <nickc@redhat.com>
2990 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2993 2014-04-03 Martin Jambor <mjambor@suse.cz>
2995 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2996 mention gcc_unreachable before failing.
2997 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3000 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3003 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3004 inconsistent code and instead mark the context inconsistent.
3005 (possible_polymorphic_call_targets): For inconsistent contexts
3006 return empty complete list.
3008 2014-04-02 Anthony Green <green@moxielogic.com>
3010 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3011 (extendqisi2, extendhisi2): Define.
3012 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3013 (WCHAR_TYPE): Change to unsigned int.
3015 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3017 PR tree-optimization/60733
3018 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3019 insertion point for PHI candidates to be the end of the feeding
3020 block for the PHI argument.
3022 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3024 PR rtl-optimization/60650
3025 * lra-constraints.c (process_alt_operands): Decrease reject for
3026 earlyclobber matching.
3028 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3030 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3032 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3034 * config/spu/spu.c (pad_bb): Do not crash when the last
3035 insn is CODE_FOR_blockage.
3037 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3039 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3040 lies outside the target mode.
3042 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3045 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3046 software floating point or no floating point registers, do not
3047 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3048 in GPRs that occurs after we tested for GPRs that would never be
3051 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3052 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3053 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3054 specifically allow DDmode, since that does not use the SPE SIMD
3057 2014-04-02 Richard Biener <rguenther@suse.de>
3060 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
3061 MODE_INTs. Properly use negv_optab.
3062 (expand_abs): Likewise.
3064 2014-04-02 Richard Biener <rguenther@suse.de>
3067 * Makefile.in (install-driver): Guard extra installs with special
3070 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3072 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3075 2014-04-01 Richard Henderson <rth@redhat.com>
3078 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3079 alternative enabled before register allocation.
3081 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3083 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3084 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3086 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3087 (nios2_got_address): Update nios2_large_got_address call site.
3088 (nios2_delegitimize_address): New function.
3089 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3090 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3091 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3093 2014-04-01 Martin Husemann <martin@duskware.de>
3095 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3098 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3100 PR rtl-optimization/60604
3101 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3102 check from register_operand.
3103 (register_operand): Redefine in terms of general_operand.
3104 (nonmemory_operand): Use register_operand for the non-constant cases.
3106 2014-04-01 Richard Biener <rguenther@suse.de>
3108 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3110 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3112 * doc/invoke.texi (mapp-regs): Clarify.
3114 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3116 * config/i386/avx512fintrin.h (__v32hi): Define type.
3117 (__v64qi): Likewise.
3118 (_mm512_set1_epi8): Define.
3119 (_mm512_set1_epi16): Define.
3120 (_mm512_set4_epi32): Define.
3121 (_mm512_set4_epi64): Define.
3122 (_mm512_set4_pd): Define.
3123 (_mm512_set4_ps): Define.
3124 (_mm512_setr4_epi64): Define.
3125 (_mm512_setr4_epi32): Define.
3126 (_mm512_setr4_pd): Define.
3127 (_mm512_setr4_ps): Define.
3128 (_mm512_setzero_epi32): Define.
3130 2014-03-31 Martin Jambor <mjambor@suse.cz>
3133 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3134 callsite_arguments_match_p. Updated all callers. Also check types of
3135 corresponding formal parameters and actual arguments.
3136 (not_all_callers_have_enough_arguments_p) Renamed to
3137 some_callers_have_mismatched_arguments_p.
3139 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3141 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3143 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3146 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3149 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3151 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3152 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3154 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3155 Use FMAMODE_NOVF512 mode iterator.
3156 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3157 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3158 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3159 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3161 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3162 Use VF_128_256 mode iterator.
3163 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3166 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3168 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3169 static chain if needed.
3171 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3174 * lra-constraints.c (index_part_to_reg): New.
3175 (process_address): Use it.
3177 2014-03-27 Jeff Law <law@redhat.com>
3178 Jakub Jelinek <jakub@redhat.com>
3181 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3182 gen_rtx_{PLUS,MULT} to build up the address expression.
3184 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3185 creating non-canonical RTL.
3187 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3190 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3191 functions; reorganize to make cheap checks first.
3192 (inline_small_functions): Do not estimate growth when dumping;
3194 * ipa-inline.h (inline_summary): Add min_size.
3195 (growth_likely_positive): New function.
3196 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3197 (set_cond_stmt_execution_predicate): Cleanup.
3198 (estimate_edge_size_and_time): Compute min_size.
3199 (estimate_calls_size_and_time): Likewise.
3200 (estimate_node_size_and_time): Likewise.
3201 (inline_update_overall_summary): Update min_size.
3202 (do_estimate_edge_time): Likewise.
3203 (do_estimate_edge_size): Update.
3204 (do_estimate_edge_hints): Update.
3205 (growth_likely_positive): New function.
3207 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3210 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3211 also if addr has VOIDmode.
3213 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3215 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3216 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3218 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3219 instructions as well as AdvancedSIMD loads.
3221 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3223 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3224 Use crypto_aese type.
3225 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3226 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3227 crypto_aese, crypto_aesmc. Move to types.md.
3228 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3230 * config/arm/iterators.md (crypto_type): Likewise.
3232 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3234 * cgraph.c: Include expr.h and tree-dfa.h.
3235 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3238 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3241 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3242 regs from checking multi-reg pseudos.
3244 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3246 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3248 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3250 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3251 if it would clobber the stack pointer, even temporarily.
3253 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3255 * mode-switching.c: Make small adjustments to the top comment.
3257 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3259 * config/rs6000/constraints.md (wD constraint): New constraint to
3260 match the constant integer to get the top DImode/DFmode out of a
3261 vector in a VSX register.
3263 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3264 match the constant integer to get the top DImode/DFmode out of a
3265 vector in a VSX register.
3267 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3270 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3273 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3274 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3276 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3277 Optimize vec_extract of 64-bit values, where the value being
3278 extracted is in the top word, where we can use scalar
3279 instructions. Add direct move and store support. Combine the big
3280 endian/little endian vector select load support into a single insn.
3281 (vsx_extract_<mode>_internal1): Likewise.
3282 (vsx_extract_<mode>_internal2): Likewise.
3283 (vsx_extract_<mode>_load): Likewise.
3284 (vsx_extract_<mode>_store): Likewise.
3285 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3286 combined into vsx_extract_<mode>_load.
3287 (vsx_extract_<mode>_one_le): Likewise.
3289 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3290 define the top 64-bit vector element.
3292 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3295 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3296 Document vec_vbpermq builtin.
3299 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3300 enable use of xxsldwi and xxpermdi builtin functions.
3301 (vec_xxpermdi): Likewise.
3303 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3304 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3306 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3308 PR rtl-optimization/60650
3309 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3311 (find_spills_for): New.
3312 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3313 Spill all pseudos on the second iteration.
3315 2014-03-27 Marek Polacek <polacek@redhat.com>
3318 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3321 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3323 * config/s390/s390.c (s390_can_use_return_insn): Check for
3324 call-saved FPRs on 31 bit.
3326 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3329 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3330 if they need regimplification, just drop them instead of
3331 calling gimple_regimplify_operands on them.
3333 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3336 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3337 (aarch64_frame_pointer_required): Adjust logic.
3338 (aarch64_can_eliminate): Adjust logic.
3339 (aarch64_override_options_after_change): Adjust logic.
3341 2014-03-27 Dehao Chen <dehao@google.com>
3343 * ipa-inline.c (early_inliner): Update node's inline info.
3345 2014-03-26 Dehao Chen <dehao@google.com>
3347 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3348 compiler inserted conditional jumps for NAN float check.
3350 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3352 * ubsan.h (ubsan_create_data): Change second argument's type
3353 to const location_t *.
3354 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3356 (ubsan_create_data): Change second argument to const location_t *PLOC.
3357 Create Loc field whenever PLOC is non-NULL.
3358 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3359 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3363 * real.c (real_to_integer2): Change type of low to UHWI.
3365 2014-03-26 Tobias Burnus <burnus@net-b.de>
3367 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3368 (CILK_SELF_SPECS): New define.
3369 (driver_self_specs): Use it.
3371 2014-03-26 Richard Biener <rguenther@suse.de>
3373 * tree-pretty-print.c (percent_K_format): Implement special
3374 case for LTO and its stripped down BLOCK tree.
3376 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3379 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3381 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3382 one range is range_int_cst_p, but not both, at least optimize
3383 addition/subtraction of 0 and multiplication by 0 or 1.
3384 * gimple-fold.c (gimple_fold_call): Fold
3385 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3386 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3387 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3389 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3391 PR rtl-optimization/60452
3392 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3393 <case REG>: Return 1 for invalid offsets from the frame pointer.
3395 2014-03-26 Marek Polacek <polacek@redhat.com>
3398 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3401 2014-03-26 Marek Polacek <polacek@redhat.com>
3404 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3407 2014-03-26 Marek Polacek <polacek@redhat.com>
3410 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3411 multiplication in unsigned type.
3413 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3415 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3417 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3419 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3421 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3424 * cif-code.def (UNREACHABLE) New code.
3425 * ipa-inline.c (inline_small_functions): Skip edges to
3426 __builtlin_unreachable.
3427 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3428 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3429 predicate to __bulitin_unreachable.
3430 (set_cond_stmt_execution_predicate): Fix issue when
3431 invert_tree_comparison returns ERROR_MARK.
3432 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3433 propagate to inline clones.
3434 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3436 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3437 * cgraphclones.c (cgraph_clone_node): If call destination is already
3438 ureachable, do not redirect it back.
3439 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3442 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3444 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3445 Do not modify inline clones.
3447 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3449 * config/i386/i386.md (general_sext_operand): New mode attr.
3450 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3451 don't generate (sign_extend (const_int)).
3452 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3453 operands[2]. Use We constraint instead of <i> and
3454 <general_sext_operand> predicate instead of <general_operand>.
3455 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3456 * config/i386/constraints.md (We): New constraint.
3457 * config/i386/predicates.md (x86_64_sext_operand,
3458 sext_operand): New predicates.
3460 2014-03-25 Martin Jambor <mjambor@suse.cz>
3463 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3464 inconsistent devirtualizations to __builtin_unreachable.
3466 2014-03-25 Marek Polacek <polacek@redhat.com>
3469 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3471 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3473 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3474 order of elements for big-endian.
3476 2014-03-25 Richard Biener <rguenther@suse.de>
3479 * gimplify-me.c (gimple_regimplify_operands): Update the
3482 2014-03-25 Martin Jambor <mjambor@suse.cz>
3485 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3486 (lto_output_varpool_node): Likewise.
3487 (input_overwrite_node): Likewise.
3488 (input_varpool_node): Likewise.
3490 2014-03-25 Richard Biener <rguenther@suse.de>
3492 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3493 (run_gcc): Likewise.
3495 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3497 * combine.c (simplify_compare_const): Add MODE argument.
3498 Handle mode_width 0 as very large mode_width.
3499 (try_combine, simplify_comparison): Adjust callers.
3501 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3502 type to avoid signed integer overflow.
3503 * explow.c (plus_constant): Likewise.
3505 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3507 * doc/generic.texi: Correct typos.
3509 2014-03-24 Tobias Burnus <burnus@net-b.de>
3511 * doc/invoke.texi (-flto): Expand section about
3512 using static libraries with LTO.
3514 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3516 PR rtl-optimization/60501
3517 * optabs.def (addptr3_optab): New optab.
3518 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3519 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3520 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3522 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3524 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3526 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3528 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3531 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3532 (_mm256_undefined_ps): Define.
3533 (_mm256_undefined_pd): Define.
3534 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3535 (_mm_undefined_pd): Define.
3536 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3537 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3538 (_mm512_undefined_ps): Define.
3539 (_mm512_undefined_pd): Define.
3540 Use _mm*_undefined_*.
3541 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3543 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3545 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3546 (lshr_simd): DI mode added.
3547 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3548 (aarch64_ushr_simddi): Likewise.
3549 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3550 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3551 (vshrd_n_u64): Likewise.
3553 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3555 * Makefile.in (s-macro_list): Depend on cc1.
3557 2014-03-23 Teresa Johnson <tejohnson@google.com>
3559 * ipa-utils.c (ipa_print_order): Use specified dump file.
3561 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3563 PR rtl-optimization/60601
3564 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3566 * gcc.c (eval_spec_function): Initialize save_growing_value.
3568 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3571 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3572 code == MINUS_EXPR, never swap op0 with op1.
3574 * toplev.c (init_local_tick): Avoid signed integer multiplication
3576 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3577 shift by first operand's bitsize.
3579 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3582 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3584 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3585 TARGET_ISA_64BIT_P(x).
3587 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3589 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3590 pattern for vector nor instead of subtract from splat(-1).
3591 (altivec_expand_vec_perm_const_le): Likewise.
3593 2014-03-21 Richard Henderson <rth@twiddle.net>
3596 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3597 related insns after epilogue_completed.
3599 2014-03-21 Martin Jambor <mjambor@suse.cz>
3602 * cgraph.h (symtab_node): New flag body_removed.
3603 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3604 when removing bodies.
3605 * symtab.c (dump_symtab_base): Dump body_removed flag.
3606 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3607 had their bodies removed.
3609 2014-03-21 Martin Jambor <mjambor@suse.cz>
3612 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3615 2014-03-21 Richard Biener <rguenther@suse.de>
3617 PR tree-optimization/60577
3618 * tree-core.h (struct tree_base): Document nothrow_flag use
3619 in VAR_DECL_NONALIASED.
3620 * tree.h (VAR_DECL_NONALIASED): New.
3621 (may_be_aliased): Adjust.
3622 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3624 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3626 * expr.c (expand_expr_real_1): Remove outdated comment.
3628 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3631 * ira.c (adjust_cleared_regs): Call copy_rtx on
3632 *reg_equiv[REGNO (loc)].src_p before passing it to
3633 simplify_replace_fn_rtx.
3636 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3637 into CONST, put pic register as first operand of PLUS. Use
3638 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3640 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3642 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3644 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3646 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3647 around for store forwarding issue in the FPU on the UT699.
3648 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3649 loads and operations if -mfix-ut699 is specified.
3650 (divtf3_hq): Tweak attribute.
3651 (sqrttf2_hq): Likewise.
3653 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3655 * calls.c (store_one_arg): Remove incorrect const qualification on the
3656 type of the temporary.
3657 * cfgexpand.c (expand_return): Likewise.
3658 * expr.c (expand_constructor): Likewise.
3659 (expand_expr_real_1): Likewise.
3661 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3663 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3666 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3669 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3671 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3673 * config/arm/aarch-common-protos.h
3674 (alu_cost_table): Fix spelling of "extend".
3675 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3677 2014-03-19 Richard Biener <rguenther@suse.de>
3680 * tree-core.h (tree_type_common): Re-order pointer members
3681 to reduce recursion depth during GC walks.
3683 2014-03-19 Marek Polacek <polacek@redhat.com>
3686 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3687 before accessing it.
3689 2014-03-19 Richard Biener <rguenther@suse.de>
3692 * lto-streamer-in.c (input_function): In WPA stage do not drop
3695 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3697 PR tree-optimization/60559
3698 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3699 with build_zero_cst assignment.
3701 2014-03-18 Kai Tietz <ktietz@redhat.com>
3703 PR rtl-optimization/56356
3704 * sdbout.c (sdbout_parms): Verify that parms'
3705 incoming argument is valid.
3706 (sdbout_reg_parms): Likewise.
3708 2014-03-18 Richard Henderson <rth@redhat.com>
3711 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3712 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3713 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3715 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3717 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3718 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3719 Italicize plugin event names in description. Explain that
3720 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3721 Remind that no GCC functions should be called after PLUGIN_FINISH.
3722 Explain what pragmas with expansion are.
3724 2014-03-18 Martin Liska <mliska@suse.cz>
3726 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3727 gimple call statement is update.
3728 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3729 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3731 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3734 * ubsan.c (ubsan_instrument_unreachable): Call
3735 initialize_sanitizer_builtins.
3736 (ubsan_pass): Likewise.
3739 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3740 varpool_finalize_decl instead of rest_of_decl_compilation.
3742 2014-03-18 Richard Biener <rguenther@suse.de>
3744 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3745 by using bitmap_and_compl instead of bitmap_and_compl_into.
3746 (df_rd_transfer_function): Likewise.
3748 2014-03-18 Richard Biener <rguenther@suse.de>
3750 * doc/lto.texi (fresolution): Fix typo.
3752 2014-03-18 Richard Biener <rguenther@suse.de>
3754 * doc/invoke.texi (flto): Update for changes in 4.9.
3756 2014-03-18 Richard Biener <rguenther@suse.de>
3758 * doc/loop.texi: Remove section on the removed lambda framework.
3759 Update loop docs with recent changes in preserving loop structure.
3761 2014-03-18 Richard Biener <rguenther@suse.de>
3763 * doc/lto.texi (-fresolution): Document.
3765 2014-03-18 Richard Biener <rguenther@suse.de>
3767 * doc/contrib.texi: Adjust my name.
3769 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3772 * internal-fn.c: Include diagnostic-core.h.
3773 (expand_BUILTIN_EXPECT): New function.
3774 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3775 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3776 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3777 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3779 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3780 Revert 3 argument __builtin_expect code.
3781 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3782 * gimple-fold.c (gimple_fold_call): Likewise.
3783 * tree.h (fold_builtin_expect): New prototype.
3784 * builtins.c (build_builtin_expect_predicate): Add predictor
3785 argument, if non-NULL, create 3 argument __builtin_expect.
3786 (fold_builtin_expect): No longer static. Add ARG2 argument,
3787 pass it through to build_builtin_expect_predicate.
3788 (fold_builtin_2): Adjust caller.
3789 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3790 * internal-fn.def (BUILTIN_EXPECT): New.
3792 2014-03-18 Tobias Burnus <burnus@net-b.de>
3795 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3796 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3797 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3799 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3802 * predict.c (combine_predictions_for_bb): Fix up formatting.
3803 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3804 fill what it points to if non-NULL.
3805 (tree_predict_by_opcode): Adjust caller, use the predictor.
3806 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3808 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3810 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3811 proper constant for the store mode.
3813 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3815 * symtab.c (change_decl_assembler_name): Fix transparent alias
3818 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3820 * config/aarch64/aarch64.c: Correct the comments about the
3821 aarch64 stack layout.
3823 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3825 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3826 check for GF_OMP_FOR_KIND_FOR.
3828 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3830 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3831 ymm and zmm register names.
3833 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3836 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3837 note creation for the 2010-08-31 changes.
3839 2014-03-17 Marek Polacek <polacek@redhat.com>
3842 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3843 as -fno-tree-loop-vectorize.
3844 (expand_omp_simd): Likewise.
3846 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3848 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3849 (eligible_for_call_delay): New prototype.
3850 * config/sparc/sparc.c (tls_call_delay): Rename into...
3851 (eligible_for_call_delay): ...this. Return false if the instruction
3852 cannot be put in the delay slot of a branch.
3853 (eligible_for_restore_insn): Simplify.
3854 (eligible_for_return_delay): Return false if the instruction cannot be
3855 put in the delay slot of a branch and simplify.
3856 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3857 put in the delay slot of a branch.
3858 * config/sparc/sparc.md (fix_ut699): New attribute.
3859 (tls_call_delay): Delete.
3860 (in_call_delay): Reimplement.
3861 (eligible_for_sibcall_delay): Rename into...
3862 (in_sibcall_delay): ...this.
3863 (eligible_for_return_delay): Rename into...
3864 (in_return_delay): ...this.
3865 (in_branch_delay): Reimplement.
3866 (in_uncond_branch_delay): Delete.
3867 (in_annul_branch_delay): Delete.
3869 2014-03-14 Richard Henderson <rth@redhat.com>
3872 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3873 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3874 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3875 (floathi splitters): Remove.
3876 (float<SWI48x>xf2): New pattern.
3877 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3878 code that tried to handle DImode for 32-bit, but which was excluded
3879 by the pattern's condition. Drop allocation of stack temporary.
3880 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3881 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3882 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3883 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3884 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3885 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3886 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3887 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3888 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3889 (*float<SWI48x><X87MODEF>2_i387): Remove.
3890 (all float _with_temp splitters): Remove.
3891 (*float<SWI48x><MODEF>2_i387): New pattern.
3892 (*float<SWI48><MODEF>2_sse): New pattern.
3893 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3894 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3896 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3897 Marek Polacek <polacek@redhat.com>
3900 * common.opt (dump_base_name_prefixed): New Variable.
3901 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3902 if x_dump_base_name_prefixed is already set, set it at the end.
3904 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3906 PR rtl-optimization/60508
3907 * lra-constraints.c (get_reload_reg): Add new parameter
3909 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3910 Pass the new parameter values.
3912 2014-03-14 Richard Biener <rguenther@suse.de>
3914 * common.opt: Revert unintented changes from r205065.
3917 2014-03-14 Richard Biener <rguenther@suse.de>
3920 * cfghooks.c (split_block): Properly adjust all loops the
3921 block was a latch of.
3923 2014-03-14 Martin Jambor <mjambor@suse.cz>
3926 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3929 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3932 * config/avr/avr.c (avr_set_current_function): Pass function name
3933 through default_strip_name_encoding before sanity checking instead
3934 of skipping the first char of the assembler name.
3936 2014-03-13 Richard Henderson <rth@redhat.com>
3939 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3940 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3941 * config/i386/i386-protos.h: Likewise.
3942 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3943 in the expander instead of a splitter.
3944 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3945 any possibility of requiring a memory.
3946 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3947 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3948 (fp branch splitters): Update for ix86_split_fp_branch.
3949 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3950 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3951 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3952 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3953 (*fop_<MODEF>_3_i387): Likewise.
3954 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3955 (splitters for the fop_* register patterns): Remove.
3956 (fscalexf4_i387): Rename from *fscalexf4_i387.
3957 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3959 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3961 PR tree-optimization/59779
3962 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3963 type for bitsize and maxsize instead of HOST_WIDE_INT.
3965 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3967 PR rtl-optimization/57320
3968 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3969 the CFG after thread_prologue_and_epilogue_insns.
3971 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3973 PR rtl-optimization/57189
3974 * lra-constraints.c (process_alt_operands): Disfavor spilling
3977 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3979 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3981 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3983 PR tree-optimization/59025
3985 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3986 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3988 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3991 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3992 calls of avr_out_plus_1.
3994 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3996 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3997 BB's single pred and update the father loop's latch info later.
3999 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4001 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4005 (VEC_base): Likewise.
4006 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4007 registers, we need to swap double words in little endian mode.
4009 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4010 to be a container mode for 128-bit integer operations added in ISA
4011 2.07. Unlike TImode and PTImode, the preferred register set is
4012 the Altivec/VMX registers for the 128-bit operations.
4014 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4016 (rs6000_split_128bit_ok_p): Likewise.
4018 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4019 macros for creating ISA 2.07 normal and overloaded builtin
4020 functions with 3 arguments.
4021 (BU_P8V_OVERLOAD_3): Likewise.
4022 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4023 for use as overloaded functions.
4024 (VPERM_1TI_UNS): Likewise.
4025 (VSEL_1TI): Likewise.
4026 (VSEL_1TI_UNS): Likewise.
4027 (ST_INTERNAL_1ti): Likewise.
4028 (LD_INTERNAL_1ti): Likewise.
4029 (XXSEL_1TI): Likewise.
4030 (XXSEL_1TI_UNS): Likewise.
4031 (VPERM_1TI): Likewise.
4032 (VPERM_1TI_UNS): Likewise.
4033 (XXPERMDI_1TI): Likewise.
4034 (SET_1TI): Likewise.
4035 (LXVD2X_V1TI): Likewise.
4036 (STXVD2X_V1TI): Likewise.
4037 (VEC_INIT_V1TI): Likewise.
4038 (VEC_SET_V1TI): Likewise.
4039 (VEC_EXT_V1TI): Likewise.
4040 (EQV_V1TI): Likewise.
4041 (NAND_V1TI): Likewise.
4042 (ORC_V1TI): Likewise.
4043 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4044 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4046 (VADDUQM): Likewise.
4047 (VSUBCUQ): Likewise.
4048 (VADDEUQM): Likewise.
4049 (VADDECUQ): Likewise.
4050 (VSUBEUQM): Likewise.
4051 (VSUBECUQ): Likewise.
4053 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
4054 __int128_t and __uint128_t types.
4055 (__uint128_type): Likewise.
4056 (altivec_categorize_keyword): Add support for vector __int128_t,
4057 vector __uint128_t, vector __int128, and vector unsigned __int128
4058 as a container type for TImode operations that need to be done in
4059 VSX/Altivec registers.
4060 (rs6000_macro_to_expand): Likewise.
4061 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
4062 to support 128-bit integer instructions vaddcuq, vadduqm,
4063 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
4064 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4066 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4067 for V1TImode, and set up preferences to use VSX/Altivec registers.
4068 Setup VSX reload handlers.
4069 (rs6000_debug_reg_global): Likewise.
4070 (rs6000_init_hard_regno_mode_ok): Likewise.
4071 (rs6000_preferred_simd_mode): Likewise.
4072 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4073 (easy_altivec_constant): Likewise.
4074 (output_vec_const_move): Likewise.
4075 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4077 (rs6000_expand_vector_extract): Likewise.
4078 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4080 (rs6000_const_vec): Add support for V1TImode.
4081 (rs6000_emit_le_vsx_load): Swap double words when loading or
4082 storing TImode/V1TImode.
4083 (rs6000_emit_le_vsx_store): Likewise.
4084 (rs6000_emit_le_vsx_move): Likewise.
4085 (rs6000_emit_move): Add support for V1TImode.
4086 (altivec_expand_ld_builtin): Likewise.
4087 (altivec_expand_st_builtin): Likewise.
4088 (altivec_expand_vec_init_builtin): Likewise.
4089 (altivec_expand_builtin): Likewise.
4090 (rs6000_init_builtins): Add support for V1TImode type. Add
4091 support for ISA 2.07 128-bit integer builtins. Define type names
4092 for the VSX/Altivec vector types.
4093 (altivec_init_builtins): Add support for overloaded vector
4094 functions with V1TImode type.
4095 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4096 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4098 (rs6000_split_128bit_ok_p): Likewise.
4099 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4100 __int128_t and vector __uint128_t.
4102 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4103 and mode attributes.
4110 (VS_scalar): Likewise.
4111 (VS_double): Likewise.
4112 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4114 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4115 we support the ISA 2.07 128-bit integer arithmetic instructions.
4116 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4117 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4118 and TImode types for use with the builtin functions.
4119 (V1TI_type_node): Likewise.
4120 (unsigned_V1TI_type_node): Likewise.
4121 (intTI_type_internal_node): Likewise.
4122 (uintTI_type_internal_node): Likewise.
4124 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4125 128-bit builtin functions.
4126 (UNSPEC_VADDEUQM): Likewise.
4127 (UNSPEC_VADDECUQ): Likewise.
4128 (UNSPEC_VSUBCUQ): Likewise.
4129 (UNSPEC_VSUBEUQM): Likewise.
4130 (UNSPEC_VSUBECUQ): Likewise.
4131 (VM): Add V1TImode to vector mode iterators.
4133 (VI_unit): Likewise.
4134 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4135 (altivec_vaddcuq): Likewise.
4136 (altivec_vsubuqm): Likewise.
4137 (altivec_vsubcuq): Likewise.
4138 (altivec_vaddeuqm): Likewise.
4139 (altivec_vaddecuq): Likewise.
4140 (altivec_vsubeuqm): Likewise.
4141 (altivec_vsubecuq): Likewise.
4143 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4145 (BOOL_128): Likewise.
4146 (BOOL_REGS_OUTPUT): Likewise.
4147 (BOOL_REGS_OP1): Likewise.
4148 (BOOL_REGS_OP2): Likewise.
4149 (BOOL_REGS_UNARY): Likewise.
4150 (BOOL_REGS_AND_CR0): Likewise.
4152 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4153 128-bit integer builtin support.
4154 (vec_vadduqm): Likewise.
4155 (vec_vaddecuq): Likewise.
4156 (vec_vaddeuqm): Likewise.
4157 (vec_vsubecuq): Likewise.
4158 (vec_vsubeuqm): Likewise.
4159 (vec_vsubcuq): Likewise.
4160 (vec_vsubuqm): Likewise.
4162 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4163 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4164 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4165 128-bit integer add/subtract to ISA 2.07.
4167 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4169 * config/arc/arc.c (arc_predicate_delay_insns):
4170 Fix third argument passed to conditionalize_nonjump.
4172 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4174 * config/aarch64/aarch64-builtins.c
4175 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4176 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4177 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4178 instead of __builtin_lfloor.
4179 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4181 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4183 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4184 (tree_ssa_ifcombine_bb_1): New function.
4185 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4186 is an empty forwarder block to then_bb or vice versa and then_bb
4187 and else_bb are effectively swapped.
4189 2014-03-12 Christian Bruel <christian.bruel@st.com>
4192 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4193 REG_CFA_DEF_CFA note.
4194 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4195 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4197 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4199 PR tree-optimization/60454
4200 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4202 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4204 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4205 Do not define target_cpu_default2 to generic.
4206 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4207 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4208 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4210 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4211 Marc Glisse <marc.glisse@inria.fr>
4213 PR tree-optimization/60502
4214 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4215 instead of build_low_bits_mask.
4217 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4220 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4221 if there are multiple uses, but op doesn't live on E edge.
4222 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4223 clobber stmts before __builtin_unreachable.
4225 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4227 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4228 hard_frame_pointer_rtx.
4229 * cse.c (cse_insn): Remove volatile check.
4230 * cselib.c (cselib_process_insn): Likewise.
4231 * dse.c (scan_insn): Likewise.
4233 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4235 * config/arc/arc.c (conditionalize_nonjump): New function,
4237 (arc_ifcvt): ... this.
4238 (arc_predicate_delay_insns): Use it.
4240 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4242 * config/arc/predicates.md (extend_operand): During/after reload,
4243 allow const_int_operand.
4244 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4245 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4246 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4248 (umulsi3_highpart_i): Likewise.
4250 2014-03-11 Richard Biener <rguenther@suse.de>
4252 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4253 Add asserts to guard possible wrong-code bugs.
4255 2014-03-11 Richard Biener <rguenther@suse.de>
4257 PR tree-optimization/60429
4258 PR tree-optimization/60485
4259 * tree-ssa-structalias.c (set_union_with_increment): Properly
4260 take into account all fields that overlap the shifted vars.
4261 (do_sd_constraint): Likewise.
4262 (do_ds_constraint): Likewise.
4263 (get_constraint_for_ptr_offset): Likewise.
4265 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4267 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4268 (nios2_compute_frame_layout):
4269 Add calculation of cfun->machine->fp_save_offset.
4270 (nios2_expand_prologue): Correct setting of frame pointer register
4272 (nios2_expand_epilogue): Update recovery of stack pointer from
4273 frame pointer accordingly.
4274 (nios2_initial_elimination_offset): Update calculation of offset
4275 for eliminating to HARD_FRAME_POINTER_REGNUM.
4277 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4280 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4281 cgraph_get_create_node on VAR_DECLs.
4283 2014-03-10 Richard Biener <rguenther@suse.de>
4286 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4288 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4290 * config/vms/vms.opt (vms_float_format): New variable.
4292 2014-03-08 Tobias Burnus <burnus@net-b.de>
4294 * doc/invoke.texi (-fcilkplus): Update implementation status.
4296 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4297 Richard Biener <rguenther@suse.de>
4299 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4300 consistently accross all TUs.
4301 (run_gcc): Enable -fshort-double automatically at link at link-time
4302 and disallow override.
4304 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4307 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4308 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4309 if they can't be used.
4311 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4313 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4314 for Solaris 11/x86 ld.
4315 * configure: Regenerate.
4317 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4319 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4320 (LIB_TLS_SPEC): Save as ld_tls_libs.
4321 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4322 (HAVE_AS_IX86_TLSLDM): New test.
4323 * configure, config.in: Regenerate.
4324 * config/i386/i386.c (legitimize_tls_address): Fall back to
4325 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4326 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4327 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4328 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4330 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4332 * common.opt (fira-loop-pressure): Mark as optimization.
4334 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4336 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4337 an OpenMP mappable type.
4339 2014-03-06 Matthias Klose <doko@ubuntu.com>
4341 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4342 MULTILIB_OSDIRNAMES is not defined.
4344 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4345 Meador Inge <meadori@codesourcery.com>
4348 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4349 (arm_legitimize_address): Call legitimize_tls_address for any
4350 arm_tls_referenced_p expression, handle constant addend. Call it
4351 before testing for !TARGET_ARM.
4352 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4354 2014-03-06 Richard Biener <rguenther@suse.de>
4360 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4362 * tree-streamer.c (record_common_node): Assert we don't record
4363 nodes with type double.
4364 (preload_common_node): Skip type double, complex double and double
4365 pointer since it is now frontend dependent due to fshort-double option.
4367 2014-03-06 Richard Biener <rguenther@suse.de>
4369 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4370 or -fno-lto is specified and the linker has full plugin support.
4371 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4372 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4373 * lto-wrapper.c (merge_and_complain): Merge compile-time
4374 optimization levels.
4375 (run_gcc): And pass it through to the link options.
4377 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4381 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4383 * cselib.c (remove_useless_values): Skip to avoid quadratic
4384 behavior if the condition moved from...
4385 (cselib_process_insn): ... here holds.
4387 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4390 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4391 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4394 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4395 (TM_H): Add x86-tune.def.
4397 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4399 * config/aarch64/aarch64.c (generic_tunings):
4400 Use cortexa57_extra_costs.
4402 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4405 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4406 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4407 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4410 2014-03-04 Heiher <r@hev.cc>
4412 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4413 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4415 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4417 * config/i386/predicates.md (const2356_operand): Change to ...
4418 (const2367_operand): ... this.
4419 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4421 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4422 (*avx512pf_scatterpf<mode>sf): Ditto.
4423 (avx512pf_scatterpf<mode>df): Ditto.
4424 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4425 (*avx512pf_scatterpf<mode>df): Ditto.
4426 * config/i386/i386.c (ix86_expand_builtin): Update
4427 incorrect hint operand error message.
4429 2014-03-04 Richard Biener <rguenther@suse.de>
4431 * lto-section-in.c (lto_get_section_data): Fix const cast.
4433 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4435 * tree-streamer.c (record_common_node): Assert we don't record
4436 nodes with type double.
4437 (preload_common_node): Skip type double, complex double and double
4438 pointer since it is now frontend dependent due to fshort-double option.
4440 2014-03-04 Richard Biener <rguenther@suse.de>
4443 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4444 (lto_input_toplevel_asms): Likewise.
4445 * lto-section-in.c (lto_get_section_data): Instead do it here
4448 2014-03-04 Richard Biener <rguenther@suse.de>
4450 PR tree-optimization/60382
4451 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4452 dead PHIs a reduction.
4454 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4456 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4458 (_mm_prefetch): Move out of GCC target("sse") pragma.
4459 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4460 GCC target("prfchw") pragma.
4461 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4463 * config/i386/i386.c (ix86_option_override_internal): Enable
4464 -mprfchw with -mprefetchwt1.
4466 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4468 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4471 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4473 * opts.h (CL_PCH_IGNORE): Define.
4474 * targhooks.c (option_affects_pch_p):
4475 Return false for options that have CL_PCH_IGNORE set.
4476 * opt-functions.awk: Process PchIgnore.
4477 * doc/options.texi: Document PchIgnore.
4479 * config/arc/arc.opt (misize): Add PchIgnore property.
4481 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4483 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4484 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4485 constraint on constants to permit them being loaded into
4486 GENERAL_REGS or BASE_REGS.
4488 2014-03-03 Nick Clifton <nickc@redhat.com>
4490 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4491 anti-cacnonical alternatives.
4492 (negandhi3_real): New pattern.
4493 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4495 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4497 * config/avr/avr-mcus.def: Remove atxmega16x1.
4498 * config/avr/avr-tables.opt: Regenerate.
4499 * config/avr/t-multilib: Regenerate.
4500 * doc/avr-mmcu.texi: Regenerate.
4502 2014-03-03 Tobias Grosser <tobias@grosser.es>
4503 Mircea Namolaru <mircea.namolaru@inria.fr>
4505 PR tree-optimization/58028
4506 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4509 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4511 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4512 not handled by recognizers.
4514 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4517 * function.c (expand_function_end): Don't emit
4518 clobber_return_register sequence if clobber_after is a BARRIER.
4519 * cfgexpand.c (construct_exit_block): Append instructions before
4520 return_label to prev_bb.
4522 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4524 * config/rs6000/constraints.md: Document reserved use of "wc".
4526 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4529 * ipa.c (function_and_variable_visibility): When dissolving comdat
4530 group, also set all symbols to local.
4532 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4537 2013-12-14 Jan Hubicka <jh@suse.cz>
4539 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4541 2014-03-02 Jon Beniston <jon@beniston.com>
4547 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4548 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4549 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4550 (simple_return, *simple_return): New patterns
4551 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4552 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4554 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4556 * dwarf2out.c (gen_subprogram_die): Tidy.
4558 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4561 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4562 (*mov_t_msb_neg_negc): ... this new insn.
4564 2014-02-28 Jason Merrill <jason@redhat.com>
4567 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4570 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4573 * dwarf2out.c (decltype_auto_die): New static.
4574 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4575 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4576 (is_cxx_auto): Likewise.
4578 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4580 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4581 we are not using general regs only.
4583 2014-02-28 Richard Biener <rguenther@suse.de>
4586 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4587 previous fix and only allow to remove trivial pre-headers
4588 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4589 (remove_forwarder_block): Properly update the latch of a loop.
4591 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4594 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4595 (cselib_preserved_hash_table): New.
4596 (preserve_constants_and_equivs): Move preserved vals to it.
4597 (cselib_find_slot): Look it up first.
4598 (cselib_init): Initialize it.
4599 (cselib_finish): Release it.
4600 (dump_cselib_table): Dump it.
4602 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4605 * cselib.c (remove_useless_values): Skip to avoid quadratic
4606 behavior if the condition moved from...
4607 (cselib_process_insn): ... here holds.
4609 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4612 * var-tracking.c (vt_initialize): Apply the same condition to
4613 preserve the CFA base value.
4615 2014-02-28 Joey Ye <joey.ye@arm.com>
4618 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4619 if reload in progress or completed.
4621 2014-02-28 Tobias Burnus <burnus@net-b.de>
4624 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4627 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4629 * doc/tm.texi.in (Condition Code Status): Update documention for
4630 relative locations of cc0-setter and cc0-user.
4632 2014-02-27 Jeff Law <law@redhat.com>
4634 PR rtl-optimization/52714
4635 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4636 into two independent simple sets, if I3 is a jump, ensure the
4637 pattern we place into I3 is a (set (pc) ...).
4639 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4640 Jeff Law <law@redhat.com>
4642 PR rtl-optimization/49847
4643 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4644 are in different blocks.
4645 * doc/tm.texi (Condition Code Status): Update documention for
4646 relative locations of cc0-setter and cc0-user.
4648 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4651 * lra.c (lra_emit_add): Check SUBREG too.
4653 2014-02-27 Andreas Schwab <schwab@suse.de>
4655 * config/m68k/m68k.c (m68k_option_override): Disable
4656 -flive-range-shrinkage for classic m68k.
4657 (m68k_override_options_after_change): Likewise.
4659 2014-02-27 Marek Polacek <polacek@redhat.com>
4662 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4663 -Wmaybe-uninitialized.
4665 2014-02-27 Alan Modra <amodra@gmail.com>
4668 * reload1.c (emit_input_reload_insns): When reload_override_in,
4669 set old to rl->in_reg when rl->in_reg is a subreg.
4671 2014-02-26 Richard Biener <rguenther@suse.de>
4674 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4676 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4678 * common/config/i386/predicates.md (const1256_operand): Remove.
4679 (const2356_operand): New.
4680 (const_1_to_2_operand): Remove.
4681 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4682 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4683 (*avx512pf_gatherpf<mode>sf): Ditto.
4684 (avx512pf_gatherpf<mode>df): Ditto.
4685 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4686 (*avx512pf_gatherpf<mode>df): Ditto.
4687 (avx512pf_scatterpf<mode>sf): Ditto.
4688 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4689 (*avx512pf_scatterpf<mode>sf): Ditto.
4690 (avx512pf_scatterpf<mode>df): Ditto.
4691 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4692 (*avx512pf_scatterpf<mode>df): Ditto.
4693 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4695 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4697 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4698 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4699 (_mm512_mask_testn_epi64_mask): Move to ...
4700 * config/i386/avx512cdintrin.h: Here.
4701 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4702 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4703 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4704 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4705 TARGET_AVX512F from TARGET_AVX512CD.
4707 2014-02-26 Richard Biener <rguenther@suse.de>
4710 * ipa.c (walk_polymorphic_call_targets): Properly guard
4711 call to inline_update_overall_summary.
4713 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4716 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4717 and latches only if requested. Fix latch if it is removed.
4718 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4719 LOOPS_HAVE_PREHEADERS.
4721 2014-02-25 Andrew Pinski <apinski@cavium.com>
4723 * builtins.c (expand_builtin_thread_pointer): Create a new target
4724 when the target is NULL.
4726 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4728 PR rtl-optimization/60317
4729 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4730 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4731 * lra-assigns.c: Include params.h.
4732 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4733 other reload pseudos considerations.
4735 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4737 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4738 to use canonical form for nor<mode>3.
4740 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4743 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4746 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4748 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4749 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4750 (ix86_handle_option): Handle OPT_mprefetchwt1.
4751 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4752 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4754 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4755 OPTION_MASK_ISA_PREFETCHWT1.
4756 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4757 (PTA_PREFETCHWT1): New.
4758 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4759 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4760 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4761 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4762 (*prefetch_avx512pf_<mode>_: Change into ...
4763 (*prefetch_prefetchwt1_<mode>: This.
4764 * config/i386/i386.opt (mprefetchwt1): New.
4765 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4766 (_mm_prefetch): Handle intent to write.
4767 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4769 2014-02-25 Richard Biener <rguenther@suse.de>
4772 * emit-rtl.c (mem_attrs_htab): Remove.
4773 (mem_attrs_htab_hash): Likewise.
4774 (mem_attrs_htab_eq): Likewise.
4775 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4776 (init_emit_once): Do not allocate mem_attrs_htab.
4778 2014-02-25 Richard Biener <rguenther@suse.de>
4781 * lto-opts.c (lto_write_options): Output non-explicit conservative
4782 -fwrapv, -fno-trapv and -fno-strict-overflow.
4783 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4784 (run_gcc): And pass them through.
4786 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4788 * sel-sched.c (calculate_new_fences): New parameter ptime.
4789 Calculate it as a maximum over all fence cycles.
4790 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4791 Print the final schedule timing when sched_verbose.
4793 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4795 PR rtl-optimization/60292
4796 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4797 bit fot the fence instruction.
4799 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4801 * calls.h: Fix typo in comment.
4803 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4805 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4806 adjusting offsetable addresses.
4808 2014-02-24 Guozhi Wei <carrot@google.com>
4810 * sparseset.h (sparseset_pop): Fix the wrong index.
4812 2014-02-24 Walter Lee <walt@tilera.com>
4814 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4815 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4817 * common/config/tilegx/tilegx-common.c
4818 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4819 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4821 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4822 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4823 (tilegx_gimplify_va_arg_expr): Handle big endian.
4824 (tilegx_expand_unaligned_load): Ditto.
4825 (tilegx_expand_unaligned_store): Ditto.
4826 (TARGET_RETURN_IN_MSB): New.
4827 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4828 (TARGET_ENDIAN_DEFAULT): New.
4829 (TARGET_BIG_ENDIAN): Handle big endian.
4830 (BYTES_BIG_ENDIAN): Ditto.
4831 (WORDS_BIG_ENDIAN): Ditto.
4832 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4835 * config/tilegx/tilegx.md (extv): Handle big endian.
4837 (insn_st<n>): Ditto.
4838 (insn_st<n>_add<bitsuffix>): Ditto.
4839 (insn_stnt<n>): Ditto.
4840 (insn_stnt<n>_add<bitsuffix>):Ditto.
4841 (vec_interleave_highv8qi): Handle big endian.
4842 (vec_interleave_highv8qi_be): New.
4843 (vec_interleave_highv8qi_le): New.
4844 (insn_v1int_h): Handle big endian.
4845 (vec_interleave_lowv8qi): Handle big endian.
4846 (vec_interleave_lowv8qi_be): New.
4847 (vec_interleave_lowv8qi_le): New.
4848 (insn_v1int_l): Handle big endian.
4849 (vec_interleave_highv4hi): Handle big endian.
4850 (vec_interleave_highv4hi_be): New.
4851 (vec_interleave_highv4hi_le): New.
4852 (insn_v2int_h): Handle big endian.
4853 (vec_interleave_lowv4hi): Handle big endian.
4854 (vec_interleave_lowv4hi_be): New.
4855 (vec_interleave_lowv4hi_le): New.
4856 (insn_v2int_l): Handle big endian.
4857 (vec_interleave_highv2si): Handle big endian.
4858 (vec_interleave_highv2si_be): New.
4859 (vec_interleave_highv2si_le): New.
4860 (insn_v4int_h): Handle big endian.
4861 (vec_interleave_lowv2si): Handle big endian.
4862 (vec_interleave_lowv2si_be): New.
4863 (vec_interleave_lowv2si_le): New.
4864 (insn_v4int_l): Handle big endian.
4865 * config/tilegx/tilegx.opt (mbig-endian): New option.
4866 (mlittle-endian): New option.
4867 * doc/install.texi: Document tilegxbe-linux.
4868 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4870 2014-02-24 Martin Jambor <mjambor@suse.cz>
4873 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4874 there are no parameter descriptors.
4876 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4878 PR rtl-optimization/60268
4879 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4880 initialization to ...
4881 (sched_rgn_init): ... here.
4882 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4884 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4886 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4889 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4891 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4894 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4896 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4897 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4899 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4901 * config/microblaze/predicates.md: Add cmp_op predicate.
4902 * config/microblaze/microblaze.md: Add branch_compare instruction
4903 which uses cmp_op predicate and emits cmp insn before branch.
4904 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4905 to microblaze_expand_conditional_branch and consolidate logic.
4906 (microblaze_expand_conditional_branch): emit branch_compare
4907 insn instead of handling cmp op separate from branch insn.
4909 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4911 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4914 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4916 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4917 define_insn with define_expand and new define_insn
4918 *altivec_lve<VI_char>x_internal.
4919 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4920 and new define_insn *altivec_stve<VI_char>x_internal.
4921 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4923 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4925 (altivec_expand_stvex_be): New function.
4927 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4929 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4930 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4931 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4932 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4934 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4937 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4938 instead of emit_move_insn.
4940 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4942 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4944 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4945 gen_altivec_vsumsws.
4947 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4949 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4950 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4951 (altivec_lvxl_<mode>): New define_expand incorporating
4952 -maltivec=be semantics where needed.
4953 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4954 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4955 semantics where needed.
4956 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4957 (altivec_stvx_<mode>): New define_expand incorporating
4958 -maltivec=be semantics where needed.
4959 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4960 VM2 iterator instead of V4SI.
4961 (altivec_stvxl_<mode>): New define_expand incorporating
4962 -maltivec=be semantics where needed.
4963 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4964 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4965 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4966 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4967 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4968 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4969 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4970 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4971 ALTIVEC_BUILTIN_STVXL.
4972 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4973 (altivec_expand_stvx_be): Likewise.
4974 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4975 (altivec_expand_lvx_be): Likewise.
4976 (altivec_expand_stvx_be): Likewise.
4977 (altivec_expand_builtin): Add cases for
4978 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4979 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4980 (altivec_init_builtins): Add definitions for
4981 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4982 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4984 2014-02-21 Catherine Moore <clm@codesourcery.com>
4986 * doc/invoke.texi (mvirt, mno-virt): Document.
4987 * config/mips/mips.opt (mvirt): New option.
4988 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4990 2014-02-21 Richard Biener <rguenther@suse.de>
4992 PR tree-optimization/60276
4993 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4994 (STMT_VINFO_MIN_NEG_DIST): New macro.
4995 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4996 STMT_VINFO_MIN_NEG_DIST.
4997 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4998 made for negative dependence distances still hold.
5000 2014-02-21 Richard Biener <rguenther@suse.de>
5003 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5004 DECL_INITIAL for globals not in the current function context.
5006 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5008 PR tree-optimization/56490
5009 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5010 * tree-ssa-uninit.c: Include params.h.
5011 (compute_control_dep_chain): Add num_calls argument, return false
5012 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5013 num_calls to recursive call.
5014 (find_predicates): Change dep_chain into normal array,
5015 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5016 variable and adjust compute_control_dep_chain caller.
5017 (find_def_preds): Likewise.
5019 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5021 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5022 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5024 2014-02-21 Nick Clifton <nickc@redhat.com>
5026 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5027 (pushhi1): Likewise.
5028 (popqi1): Add mode to pre_dec.
5031 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5033 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5034 mode for mask of V8SFmode permutation.
5036 2014-02-20 Richard Henderson <rth@redhat.com>
5039 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5040 a new pseudo for OLDVAL.
5042 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5045 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5046 gen_reg_rtx if d->testing_p.
5047 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5048 if d->testing_p and we will certainly return true.
5049 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
5052 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
5054 * emit-rtl.c (gen_reg_rtx): Assert that
5055 crtl->emit.regno_pointer_align_length is non-zero.
5057 2014-02-20 Richard Henderson <rth@redhat.com>
5060 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
5061 on failure the store back into EXPECT.
5063 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
5064 Sandra Loosemore <sandra@codesourcery.com>
5066 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5067 * config/nios2/nios2.c (nios2_function_profiler): Add
5068 -fPIC (flag_pic == 2) support.
5069 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5070 (nios2_large_offset_p): New function.
5071 (nios2_unspec_reloc_p): Move up position, update to use
5072 nios2_large_offset_p.
5073 (nios2_unspec_address): Remove function.
5074 (nios2_unspec_offset): New function.
5075 (nios2_large_got_address): New function.
5076 (nios2_got_address): Add large offset support.
5077 (nios2_legitimize_tls_address): Update usage of removed and new
5079 (nios2_symbol_binds_local_p): New function.
5080 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5081 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5082 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5083 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5084 processing for (const (unspec ...)).
5085 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5087 2014-02-20 Richard Biener <rguenther@suse.de>
5089 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5090 doing the substitution.
5091 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5093 2014-02-20 Martin Jambor <mjambor@suse.cz>
5096 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5097 info when checking whether lattices are bottom.
5099 2014-02-20 Richard Biener <rguenther@suse.de>
5102 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5105 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5108 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5109 parameter specifying the scaling.
5110 (inline_call): Update.
5111 (want_inline_recursively): Guard division by zero.
5112 (recursive_inlining): Update.
5113 * ipa-inline.h (clone_inlined_nodes): Update.
5115 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5118 * config/i386/i386.c (classify_argument): Pass structures of size
5119 64 bytes or less in register.
5121 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5122 Kirill Yukhin <kirill.yukhin@intel.com>
5124 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5125 (_mm_rcp28_round_ss): Ditto.
5126 (_mm_rsqrt28_round_sd): Ditto.
5127 (_mm_rsqrt28_round_ss): Ditto.
5128 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5129 (_mm_rcp14_round_ss): Ditto.
5130 (_mm_rsqrt14_round_sd): Ditto.
5131 (_mm_rsqrt14_round_ss): Ditto.
5132 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5133 the first input operand, get rid of match_dup.
5134 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5136 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5138 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5139 operand as the first input operand, set type attribute.
5140 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5142 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5143 operand as the first input operand, set type attribute.
5145 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5147 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5150 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5153 * config/i386/i386.c (construct_container): Remove TFmode check
5154 for X86_64_INTEGER_CLASS.
5156 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5159 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5160 only when -Wpsabi is enabled.
5162 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5165 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5166 passing arrays in registers are the same as for structs, so remove the
5167 special case for them.
5169 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5171 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5172 destination type, extract only the valid bits if the source type is not
5173 integral and has a different mode.
5175 2014-02-19 Richard Biener <rguenther@suse.de>
5178 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5181 2014-02-19 Richard Biener <rguenther@suse.de>
5184 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5185 (ipa_modify_call_arguments): Emit an argument load explicitely and
5186 preserve virtual SSA form there and for the replacement call.
5187 Do not update SSA form nor free dominance info.
5189 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5191 * ipa.c (function_and_variable_visibility): Also clear WEAK
5192 flag when disolving COMDAT_GROUP.
5194 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5196 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5197 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5199 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5200 do more sanity checks.
5201 (detect_type_change): Return true when giving up early.
5202 (compute_complex_assign_jump_func): Fix type parameter of
5203 ipa_set_ancestor_jf.
5204 (compute_complex_ancestor_jump_func): Likewise.
5205 (update_jump_functions_after_inlining): Fix updating of
5207 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5209 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5211 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5212 inline clones when edge disappears.
5214 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5217 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5218 Split 64-bit moves into 2 patterns. Do not allow the use of
5219 direct move for TDmode in little endian, since the decimal value
5220 has little endian bytes within a word, but the 64-bit pieces are
5221 ordered in a big endian fashion, and normal subreg's of TDmode are
5223 (mov<mode>_64bit_dm): Likewise.
5224 (movtd_64bit_nodm): Likewise.
5226 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5228 PR tree-optimization/60174
5229 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5230 statement of an SSA_NAME that occurs in an abnormal PHI node.
5232 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5235 * final.c (SEEN_BB): Remove.
5236 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5237 (final_scan_insn): Don't force_source_line on second
5238 NOTE_INSN_BASIC_BLOCK.
5240 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5243 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5244 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5245 (type_natural_mode): Warn ABI change when %zmm register is not
5246 available for AVX512F vector value passing.
5248 2014-02-18 Kai Tietz <ktietz@redhat.com>
5251 * config/i386/i386.c (ix86_expand_prologue): Use value in
5252 rax register as displacement when restoring %r10 or %rax.
5253 Fix wrong offset when restoring both registers.
5255 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5257 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5258 assertion with conditional return.
5260 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5261 Uros Bizjak <ubizjak@gmail.com>
5264 * config/i386/driver-i386.c (host_detect_local_cpu): If
5265 YMM state is not saved by the OS, also clear has_f16c. Move
5266 CPUID 0x80000001 handling before YMM state saving checking.
5268 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5270 PR rtl-optimization/58960
5271 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5272 factored out from ...
5273 (sched_init): ... here.
5274 (free_global_sched_pressure_data): New, factored out from ...
5275 (sched_finish): ... here.
5276 * sched-int.h (free_global_sched_pressure_data): Declare.
5277 * sched-rgn.c (nr_regions_initial): New static global.
5278 (haifa_find_rgns): Initialize it.
5279 (schedule_region): Disable sched-pressure for the newly
5282 2014-02-17 Richard Biener <rguenther@suse.de>
5284 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5285 release SSA defs of pattern stmts.
5287 2014-02-17 Richard Biener <rguenther@suse.de>
5289 * tree-inline.c (expand_call_inline): Release the virtual
5290 operand defined by the call we are about to inline.
5292 2014-02-17 Richard Biener <rguenther@suse.de>
5294 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5296 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5297 Ilya Tocar <ilya.tocar@intel.com>
5299 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5300 arguments order in builtin.
5301 (_mm512_permutexvar_epi64): Ditto.
5302 (_mm512_mask_permutexvar_epi64): Ditto
5303 (_mm512_maskz_permutexvar_epi32): Ditto
5304 (_mm512_permutexvar_epi32): Ditto
5305 (_mm512_mask_permutexvar_epi32): Ditto
5307 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5309 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5310 (p8_vmrgow): Likewise.
5312 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5314 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5317 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5320 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5321 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5322 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5323 using direct move instructions on ISA 2.07. Also adjust
5324 instruction length for 64-bit.
5325 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5326 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5328 2014-02-15 Alan Modra <amodra@gmail.com>
5332 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5333 find_replacement on parts of insn rtl that might be reloaded.
5335 2014-02-15 Richard Biener <rguenther@suse.de>
5337 PR tree-optimization/60183
5338 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5339 (tree_ssa_phiprop): Calculate and free post-dominators.
5341 2014-02-14 Jeff Law <law@redhat.com>
5343 PR rtl-optimization/60131
5344 * ree.c (get_extended_src_reg): New function.
5345 (combine_reaching_defs): Use it rather than assuming location of REG.
5346 (find_and_remove_re): Verify first operand of extension is
5347 a REG before adding the insns to the copy list.
5349 2014-02-14 Roland McGrath <mcgrathr@google.com>
5351 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5352 * configure: Regenerated.
5353 * config.in: Regenerated.
5354 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5355 instead of ASM_SHORT.
5357 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5358 Richard Earnshaw <rearnsha@arm.com>
5360 PR rtl-optimization/59535
5361 * lra-constraints.c (process_alt_operands): Encourage alternative
5362 when unassigned pseudo class is superset of the alternative class.
5363 (inherit_reload_reg): Don't inherit when optimizing for code size.
5364 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5365 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5366 modes not less than 4 for Thumb1.
5368 2014-02-14 Kyle McMartin <kyle@redhat.com>
5371 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5373 2014-02-14 Richard Biener <rguenther@suse.de>
5375 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5376 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5377 Do not strip INDIRECT_REFs.
5379 2014-02-14 Richard Biener <rguenther@suse.de>
5382 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5383 DECL_FUNCTION_SPECIFIC_TARGET.
5384 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5385 * tree-streamer-out.c (pack_ts_target_option): Remove.
5386 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5387 (write_ts_function_decl_tree_pointers): Do not stream
5388 DECL_FUNCTION_SPECIFIC_TARGET.
5389 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5390 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5391 (lto_input_ts_function_decl_tree_pointers): Do not stream
5392 DECL_FUNCTION_SPECIFIC_TARGET.
5394 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5396 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5397 (get_initial_def_for_induction, vectorizable_induction): Ignore
5398 debug stmts when looking for exit_phi.
5399 (vectorizable_live_operation): Fix up condition.
5401 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5403 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5404 nreverse() because it changes the content of original tree list.
5406 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5408 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5409 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5411 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5413 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5414 GNU coding standards.
5416 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5419 * dwarf2out.c (gen_subprogram_die): Don't call
5420 add_calling_convention_attribute if subr_die is old_die.
5422 2014-02-13 Sharad Singhai <singhai@google.com>
5424 * doc/optinfo.texi: Fix order of nodes.
5426 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5428 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5429 operands[2], not operands[3].
5431 2014-02-13 Richard Biener <rguenther@suse.de>
5434 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5435 mention the possibility of an in-tree build.
5436 (CLooG): Update recommended version to 0.18.1, mention the
5437 possibility of an in-tree build and clarify that the ISL
5438 bundled with CLooG does not work.
5440 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5443 * expr.c (compress_float_constant): If x is a hard register,
5444 extend into a pseudo and then move to x.
5446 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5448 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5449 caused by bad second argument to warning_at() with -mhotpatch and
5450 nested functions (e.g. with gfortran).
5452 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5454 * opts.c (option_name): Remove "enabled by default" rider.
5456 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5458 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5460 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5461 Uros Bizjak <ubizjak@gmail.com>
5464 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5465 * configure: Regenerated.
5467 2014-02-12 Richard Biener <rguenther@suse.de>
5469 * vec.c (vec_prefix::calculate_allocation): Move as
5470 inline variant to vec.h.
5471 (vec_prefix::calculate_allocation_1): New out-of-line version.
5472 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5473 (vec_prefix::m_has_auto_buf): Rename to ...
5474 (vec_prefix::m_using_auto_storage): ... this.
5475 (vec_prefix::calculate_allocation): Inline the easy cases
5476 and dispatch to calculate_allocation_1 which doesn't need the
5478 (va_heap::reserve): Use gcc_checking_assert.
5479 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5480 m_using_auto_storage.
5481 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5483 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5484 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5485 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5487 2014-02-12 Richard Biener <rguenther@suse.de>
5489 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5490 when we found a dependence.
5492 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5494 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5496 (maybe_fold_stmt): ... into this new function.
5497 * omp-low.c (lower_omp): Update comment.
5499 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5502 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5505 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5507 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5508 identifiers in comments.
5509 (cortexa53_extra_costs): Likewise.
5510 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5511 (cortexa7_extra_costs): Likewise.
5512 (cortexa12_extra_costs): Likewise.
5513 (cortexa15_extra_costs): Likewise.
5514 (v7m_extra_costs): Likewise.
5516 2014-02-12 Richard Biener <rguenther@suse.de>
5519 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5520 of posix_memalign being successful.
5521 (lower_stmt): Restrict lowering of posix_memalign to when
5522 -ftree-bit-ccp is enabled.
5524 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5526 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5528 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5530 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5532 PR rtl-optimization/60116
5533 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5534 other_insn once the combination has been validated.
5536 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5539 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5541 * ipa-devirt.c: Include demangle.h
5542 (odr_violation_reported): New static variable.
5543 (add_type_duplicate): Update odr_violations.
5544 (maybe_record_node): Add completep parameter; update it.
5545 (record_target_from_binfo): Add COMPLETEP parameter;
5546 update it as needed.
5547 (possible_polymorphic_call_targets_1): Likewise.
5548 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5549 rename FINAL to COMPLETE.
5550 (record_targets_from_bases): Sanity check we found the binfo;
5551 fix COMPLETEP updating.
5552 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5553 parameter, fix computing of COMPLETEP.
5554 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5555 at LTO time do demangling.
5556 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5557 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5559 (gimple_get_virt_method_for_binfo): Likewise.
5560 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5561 gimple_get_virt_method_for_vtable): Update prototypes.
5563 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5566 * genautomata.c (add_presence_absence): Fix typo with
5567 {final_}presence_list.
5569 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5572 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5573 for VSX/Altivec vectors that land in GPR registers.
5575 2014-02-11 Richard Henderson <rth@redhat.com>
5576 Jakub Jelinek <jakub@redhat.com>
5579 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5580 around drhs if type conversion to lacc->type is not useless.
5582 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5584 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5586 (cortex-a57.cortex-a53): Likewise.
5587 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5589 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5591 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5594 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5596 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5597 add_options_for_arm_vfp3.
5599 2014-02-11 Jeff Law <law@redhat.com>
5602 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5603 object with an undesirable mode.
5605 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5608 * config/i386/sol2-9.h: New file.
5609 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5610 *-*-solaris2.9*): Use it.
5612 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5614 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5615 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5617 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5619 * config/microblaze/microblaze.c: Extend mcpu version format
5621 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5623 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5625 2014-02-10 Richard Henderson <rth@redhat.com>
5628 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5629 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5630 ms-abi vs -mno-accumulate-outgoing-args.
5631 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5632 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5635 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5638 * cfgexpand.c (expand_asm_operands): Attach source location to
5639 ASM_INPUT rtx objects.
5640 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5642 2014-02-10 Nick Clifton <nickc@redhat.com>
5644 * config/mn10300/mn10300.c (popcount): New function.
5645 (mn10300_expand_prologue): Include saved registers in stack usage
5648 2014-02-10 Jeff Law <law@redhat.com>
5651 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5652 when changing the SET_DEST of a prior insn to avoid an input reload.
5654 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5656 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5657 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5658 -mcall-openbsd, or -mcall-linux.
5659 (CC1_ENDIAN_BIG_SPEC): Remove.
5660 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5661 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5662 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5663 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5664 and %cc1_endian_default.
5665 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5667 2014-02-10 Richard Biener <rguenther@suse.de>
5669 PR tree-optimization/60115
5670 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5671 MEM_REF handling. Properly verify that the accesses are not
5672 out of the objects bound.
5674 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5676 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5679 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5681 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5682 proper constants and fix formatting.
5683 (possible_polymorphic_call_targets): Fix formatting.
5685 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5686 Ilya Tocar <ilya.tocar@intel.com>
5688 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5689 (_mm512_loadu_epi32): Renamed into...
5690 (_mm512_loadu_si512): This.
5691 (_mm512_storeu_epi32): Renamed into...
5692 (_mm512_storeu_si512): This.
5693 (_mm512_maskz_ceil_ps): Removed.
5694 (_mm512_maskz_ceil_pd): Ditto.
5695 (_mm512_maskz_floor_ps): Ditto.
5696 (_mm512_maskz_floor_pd): Ditto.
5697 (_mm512_floor_round_ps): Ditto.
5698 (_mm512_floor_round_pd): Ditto.
5699 (_mm512_ceil_round_ps): Ditto.
5700 (_mm512_ceil_round_pd): Ditto.
5701 (_mm512_mask_floor_round_ps): Ditto.
5702 (_mm512_mask_floor_round_pd): Ditto.
5703 (_mm512_mask_ceil_round_ps): Ditto.
5704 (_mm512_mask_ceil_round_pd): Ditto.
5705 (_mm512_maskz_floor_round_ps): Ditto.
5706 (_mm512_maskz_floor_round_pd): Ditto.
5707 (_mm512_maskz_ceil_round_ps): Ditto.
5708 (_mm512_maskz_ceil_round_pd): Ditto.
5709 (_mm512_expand_pd): Ditto.
5710 (_mm512_expand_ps): Ditto.
5711 * config/i386/i386.c (ix86_builtins): Remove
5712 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5713 (bdesc_args): Ditto.
5714 * config/i386/predicates.md (const1256_operand): New.
5715 (const_1_to_2_operand): Ditto.
5716 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5717 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5718 (*avx512pf_gatherpf<mode>sf): Ditto.
5719 (avx512pf_gatherpf<mode>df): Ditto.
5720 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5721 (*avx512pf_gatherpf<mode>df): Ditto.
5722 (avx512pf_scatterpf<mode>sf): Ditto.
5723 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5724 (*avx512pf_scatterpf<mode>sf): Ditto.
5725 (avx512pf_scatterpf<mode>df): Ditto.
5726 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5727 (*avx512pf_scatterpf<mode>df): Ditto.
5728 (avx512f_expand<mode>): Removed.
5729 (<shift_insn><mode>3<mask_name>): Change predicate type.
5731 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5733 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5734 not at the end of datarefs vector use ordered_remove to avoid
5735 reordering datarefs vector.
5738 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5739 mark local addressable non-static vars as GOVD_PRIVATE
5740 instead of GOVD_LOCAL.
5741 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5742 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5746 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5747 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5748 assume_aligned or alloc_align attributes.
5749 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5750 arguments. Handle also assume_aligned and alloc_align attributes.
5751 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5752 calls to functions with assume_aligned or alloc_align attributes.
5753 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5755 2014-02-08 Terry Guo <terry.guo@arm.com>
5757 * doc/invoke.texi: Document ARM -march=armv7e-m.
5759 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5761 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5762 flag on __cilkrts_rethrow builtin.
5765 * ipa-cp.c (determine_versionability): Fail at -O0
5766 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5767 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5770 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5773 * tree-inline.c (copy_forbidden): Fail for
5774 __attribute__((optimize (0))) functions.
5776 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5778 * varpool.c: Include pointer-set.h.
5779 (varpool_remove_unreferenced_decls): Variables in other partitions
5780 will not be output; be however careful to not lose information
5783 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5785 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5786 lookup in the vtable constructor.
5788 2014-02-07 Jeff Law <law@redhat.com>
5791 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5792 define_insn_and_split.
5794 * ipa-inline.c (inline_small_functions): Fix typos.
5796 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5798 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5799 (s390_can_use_return_insn): Declare.
5800 * config/s390/s390.h (EPILOGUE_USES): Define.
5801 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5803 (s390_chunkify_start): Handle return JUMP_LABELs.
5804 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5805 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5806 (s390_can_use_return_insn): New functions.
5807 (s390_fix_long_loop_prediction): Handle conditional returns.
5808 (TARGET_SET_UP_BY_PROLOGUE): Define.
5809 * config/s390/s390.md (ANY_RETURN): New code iterator.
5810 (*creturn, *csimple_return, return, simple_return): New patterns.
5812 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5814 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5815 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5816 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5817 REG_CFA_RESTORE list when deciding not to restore a register.
5819 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5821 * config/s390/s390.c: Include tree-pass.h and context.h.
5822 (s390_early_mach): New function, split out from...
5823 (s390_emit_prologue): ...here.
5824 (pass_data_s390_early_mach): New pass structure.
5825 (pass_s390_early_mach): New class.
5826 (s390_option_override): Create and register early_mach pass.
5827 Move to end of file.
5829 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5831 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5832 to match for the exit block.
5834 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5836 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5837 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5838 Reject misaligned operands.
5840 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5842 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5844 2014-02-07 Richard Biener <rguenther@suse.de>
5847 * gimple-low.c (lower_builtin_posix_memalign): New function.
5848 (lower_stmt): Call it to lower posix_memalign in a way
5849 to make alignment info accessible.
5851 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5854 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5855 __builtin_setjmp_receiver.
5857 2014-02-07 Richard Biener <rguenther@suse.de>
5860 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5861 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5862 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5863 Handle BUILT_IN_POSIX_MEMALIGN.
5864 (find_func_clobbers): Likewise.
5865 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5866 (call_may_clobber_ref_p_1): Likewise.
5868 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5871 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5874 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5877 * lto-cgraph.c (lto_output_node): Use
5878 symtab_get_symbol_partitioning_class.
5879 (lto_output_varpool_node): likewise.
5880 (symtab_get_symbol_partitioning_class): Move here from
5882 * cgraph.h (symbol_partitioning_class): Likewise.
5883 (symtab_get_symbol_partitioning_class): Declare.
5885 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5887 * ggc.h (ggc_internal_cleared_alloc): New macro.
5888 * vec.h (vec_safe_copy): Handle memory stats.
5889 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5890 * target-globals.c (save_target_globals): Likewise.
5892 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5895 * expr.c (emit_move_resolve_push): Export; be bit more selective
5896 on when to clear alias set.
5897 * expr.h (emit_move_resolve_push): Declare.
5898 * function.h (struct function): Add tail_call_marked.
5899 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5900 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5901 * config/i386/i386.md (TImode move expander): De not call
5903 (FP push expanders): Preserve memory attributes.
5904 * config/i386/sse.md (push<mode>1): Remove.
5905 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5906 (ix86_expand_push): Remove.
5907 * config/i386/mmx.md (push<mode>1): Remove.
5909 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5911 PR rtl-optimization/60030
5912 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5913 lopart with paradoxical subreg before shifting it up by hprec.
5915 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5917 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5918 Remove extra newline at end of file.
5919 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5920 (arm_issue_rate): Handle cortexa57.
5921 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5922 (cortex-a57.cortex-a53): Likewise.
5924 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5927 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5928 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5930 (arm_expand_prologue): Adjust all callers.
5931 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5932 info, registers also at the lowest numbered registers side. Use
5933 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5937 * var-tracking.c (adjust_mems): Before adding a SET to
5938 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5940 2014-02-06 Alan Modra <amodra@gmail.com>
5943 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5944 change SDmode to DDmode when lra_in_progress.
5946 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5949 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5950 free_data_ref on the dr first, and before goto again also set dr
5951 to the next dr. For simd_lane_access, free old datarefs[i] before
5952 overwriting it. For get_vectype_for_scalar_type failure, don't
5953 free_data_ref if simd_lane_access.
5955 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5958 * tree.h (opts_for_fn): New inline function.
5959 (opt_for_fn): Define.
5960 * config/i386/i386.c (ix86_function_regparm): Use
5961 opt_for_fn (decl, optimize) instead of optimize.
5963 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5965 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5966 for SYMBOL_REF in large memory model.
5968 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5970 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5972 (cortex-a57): Likewise.
5973 (cortex-a57.cortex-a53): Likewise.
5975 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5976 Kugan Vivekanandarajah <kuganv@linaro.org>
5978 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5980 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5982 2014-02-06 Richard Biener <rguenther@suse.de>
5984 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5985 set_loop_copy and initialize_original_copy_tables.
5987 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5989 * config/aarch64/aarch64-simd.md
5990 (aarch64_ashr_simddi): Change QI to SI.
5992 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5993 Jakub Jelinek <jakub@redhat.com>
5996 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5999 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6001 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6002 CODE_FOR_altivec_vpku[hw]um to
6003 CODE_FOR_altivec_vpku[hw]um_direct.
6004 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6005 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6006 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6007 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6009 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6011 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6012 generation for -maltivec=be.
6013 (altivec_vsumsws): Simplify redundant test.
6015 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6017 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6018 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6019 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6020 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6021 gen_altivec_vpkuwum.
6022 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6024 (altivec_vpks<VI_char>ss): Likewise.
6025 (altivec_vpks<VI_char>us): Likewise.
6026 (altivec_vpku<VI_char>us): Likewise.
6027 (altivec_vpku<VI_char>um): Likewise.
6028 (altivec_vpku<VI_char>um_direct): New (copy of
6029 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6031 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6032 target is little endian and -maltivec=be is not specified.
6033 (*altivec_vupkhs<VU_char>_direct): New (copy of
6034 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6035 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6036 target is little endian and -maltivec=be is not specified.
6037 (*altivec_vupkls<VU_char>_direct): New (copy of
6038 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6039 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6040 little endian and -maltivec=be is not specified.
6041 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6042 little endian and -maltivec=be is not specified.
6044 2014-02-05 Richard Henderson <rth@redhat.com>
6047 * combine-stack-adj.c: Revert r206943.
6048 * sched-int.h (struct deps_desc): Add last_args_size.
6049 * sched-deps.c (init_deps): Initialize it.
6050 (sched_analyze_insn): Add OUTPUT dependencies between insns that
6051 contain REG_ARGS_SIZE notes.
6053 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6055 * lto-cgraph.c (asm_nodes_output): Make global.
6056 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
6057 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
6058 (driver_handle_option): Handle OPT_fwpa.
6060 2014-02-05 Jakub Jelinek <jakub@redhat.com>
6063 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
6064 a comment typo and formatting issue. If odr_hash hasn't been
6065 created, return vNULL and set *completep to false.
6068 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6069 bb with no successors.
6071 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6074 * doc/invoke.texi (-march): Clarify documentation for ARM.
6078 2014-02-05 Richard Biener <rguenther@suse.de>
6080 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6081 when not vectorizing because of too many alias checks.
6082 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6083 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6085 2014-02-05 Nick Clifton <nickc@redhat.com>
6087 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6088 accept extended registers in any mode when compiling for the MN10300.
6090 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6092 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6093 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6094 sanitization attributes.
6095 (can_inline_edge_p): Likewise.
6096 (sanitize_attrs_match_for_inline_p): New function.
6098 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6100 * ipa-prop.c (detect_type_change): Shor circuit testing of
6101 type changes on THIS pointer.
6103 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6106 * config/pa/pa.c (legitimize_tls_address): Return original address
6107 if not passed a SYMBOL_REF rtx.
6108 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6110 (pa_emit_move_sequence): Simplify TLS source operands.
6111 (pa_legitimate_constant_p): Reject all TLS constants.
6112 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6113 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6115 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6117 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6118 groups when we know they are controlled by LTO.
6119 * varasm.c (default_binds_local_p_1): If object is in other partition,
6120 it will be resolved locally.
6122 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6124 * config/host-linux.c (linux_gt_pch_use_address): Don't
6125 use SSIZE_MAX because it is not always defined.
6127 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6130 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6131 threshold for pseudo splitting.
6132 (update_ebb_live_info): Process call argument hard registers and
6133 hard registers from insn definition too.
6134 (max_small_class_regs_num): New constant.
6135 (inherit_in_ebb): Update live hard regs through EBBs. Update
6136 reloads_num only for small register classes. Don't split for
6139 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6142 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6145 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6147 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6150 2014-02-04 Marek Polacek <polacek@redhat.com>
6152 * gdbinit.in (pel): Define.
6154 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6156 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6159 2014-02-04 Richard Biener <rguenther@suse.de>
6162 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6163 in function context local.
6164 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6165 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6166 similar to LTO_imported_decl_ref.
6168 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6170 PR tree-optimization/60002
6171 * cgraphclones.c (build_function_decl_skip_args): Clear
6174 PR tree-optimization/60023
6175 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6176 false to gsi_replace.
6177 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6178 has been in some EH region and vec_stmt could throw, add
6179 vec_stmt into the same EH region.
6180 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6181 has no lhs, ignore it.
6182 * internal-fn.c (expand_MASK_LOAD): Likewise.
6185 * tree-inline.c (copy_forbidden): Fail for
6186 __attribute__((optimize (0))) functions.
6189 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6190 is set, copy one less argument.
6191 (expand_simd_clones): Don't subtract clone_info->inbranch
6192 from simd_clone_struct_alloc argument.
6194 PR rtl-optimization/57915
6195 * recog.c (simplify_while_replacing): If all unary/binary/relational
6196 operation arguments are constant, attempt to simplify those.
6199 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6200 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6202 2014-02-04 Richard Biener <rguenther@suse.de>
6204 PR tree-optimization/60012
6205 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6206 TBAA disambiguation to all DDRs.
6208 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6211 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6212 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6214 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6217 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6219 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6221 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6222 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6224 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6227 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6228 to figure out targets of polymorphic calls with known decl.
6229 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6230 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6231 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6232 (get_polymorphic_call_info): ... here.
6233 (get_polymorphic_call_info_from_invariant): New function.
6235 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6237 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6238 lookup via vtable pointer; check for type consistency
6239 and turn inconsitent facts into UNREACHABLE.
6240 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6241 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6242 type inconsistent querries; return UNREACHABLE instead.
6244 2014-02-03 Richard Henderson <rth@twiddle.net>
6247 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6248 already processed this node.
6249 (normalize_one_pred_1): Pass along mark_set.
6250 (normalize_one_pred): Create and destroy a pointer_set_t.
6251 (normalize_one_pred_chain): Likewise.
6253 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6255 PR gcov-profile/58602
6256 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6258 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6261 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6262 -fno-devirtualize; try to devirtualize by the knowledge of
6263 virtual table pointer given by aggregate propagation.
6264 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6265 (ipa_print_node_jump_functions): Dump also offset that
6266 is relevant for polymorphic calls.
6267 (determine_known_aggregate_parts): Add arg_type parameter; use it
6268 instead of determining the type from pointer type.
6269 (ipa_compute_jump_functions_for_edge): Update call of
6270 determine_known_aggregate_parts.
6271 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6272 (gimple_get_virt_method_for_binfo): ... here; simplify using
6273 vtable_pointer_value_to_vtable.
6274 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6275 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6276 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6277 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6279 (vtable_pointer_value_to_binfo): ... here.
6280 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6282 2014-02-03 Teresa Johnson <tejohnson@google.com>
6284 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6285 redef of outer loop index variable.
6287 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6291 * doc/extend.texi (Function Attributes): Typo.
6293 2014-02-03 Cong Hou <congh@google.com>
6295 PR tree-optimization/60000
6296 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6297 if the vectorized statement is a store. A store statement can only
6298 appear at the end of pattern statements.
6300 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6302 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6303 (ix86_option_override_internal): Default long double to 64-bit for
6304 32-bit Bionic and to 128-bit for 64-bit Bionic.
6306 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6307 TARGET_LONG_DOUBLE_128 is true.
6308 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6310 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6311 (mlong-double-64): Negate -mlong-double-128.
6312 (mlong-double-128): New option.
6314 * config/i386/i386-c.c (ix86_target_macros): Define
6315 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6317 * doc/invoke.texi: Document -mlong-double-128.
6319 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6321 PR rtl-optimization/60024
6322 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6324 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6326 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6328 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6330 PR rtl-optimization/57662
6331 * sel-sched.c (code_motion_path_driver): Do not mark already not
6332 existing blocks in the visiting bitmap.
6334 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6336 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6337 on the insn being emitted.
6339 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6340 Will Deacon <will.deacon@arm.com>
6342 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6344 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6346 * config/arm/arm-tables.opt: Regenerate.
6348 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6350 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6351 for vector types other than V16QImode.
6352 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6353 define_expand, and call altivec_expand_vec_perm_le when producing
6354 code with little endian element order.
6355 (*altivec_vperm_<mode>_internal): New insn having previous
6356 behavior of altivec_vperm_<mode>.
6357 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6358 altivec_expand_vec_perm_le when producing code with little endian
6360 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6361 behavior of altivec_vperm_<mode>_uns.
6363 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6365 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6366 (altivec_vsumsws): Add handling for -maltivec=be with a little
6368 (altivec_vsumsws_direct): New.
6369 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6370 gen_altivec_vsumsws.
6372 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6374 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6375 vtable_pointer_value_to_binfo): New functions.
6376 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6377 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6379 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6381 * config/nios2/nios2.md (load_got_register): Initialize GOT
6382 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6383 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6385 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6387 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6388 preserverd by passthrough, do not propagate the type.
6390 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6392 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6393 (mips_atomic_assign_expand_fenv): New function.
6394 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6396 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6398 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6399 (__builtin_mips_set_fcsr): Likewise.
6400 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6401 MIPS_USI_FTYPE_VOID.
6402 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6403 (mips16_expand_set_fcsr): Likewise.
6404 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6405 (mips16_set_fcsr_stub): Likewise.
6406 (mips16_get_fcsr_one_only_stub): New class.
6407 (mips16_set_fcsr_one_only_stub): Likewise.
6408 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6409 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6410 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6411 (hard_float): New availability predicate.
6412 (mips_builtins): Add get_fcsr and set_fcsr.
6413 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6414 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6415 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6416 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6417 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6420 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6422 * config/mips/mips.c (mips_one_only_stub): New class.
6423 (mips_need_mips16_rdhwr_p): Replace with...
6424 (mips16_rdhwr_stub): ...this new variable.
6425 (mips16_stub_call_address): New function.
6426 (mips16_rdhwr_one_only_stub): New class.
6427 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6428 (mips_output_mips16_rdhwr): Delete.
6429 (mips_finish_stub): New function.
6430 (mips_code_end): Use it to handle rdhwr stubs.
6432 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6435 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6436 when calculating size of integer atomic types.
6438 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6440 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6442 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6444 PR tree-optimization/60003
6445 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6446 * profile.c (branch_prob): Use gimple_call_builtin_p
6447 to check for BUILT_IN_SETJMP_RECEIVER.
6448 * tree-inline.c (copy_bb): Call notice_special_calls.
6450 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6453 * lra-constraints.c (process_alt_operands): Update reload_sum only
6456 2014-01-31 Richard Henderson <rth@redhat.com>
6459 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6460 until after else_eh is processed.
6462 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6464 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6465 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6466 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6467 in smmintrin.h, remove them.
6468 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6469 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6470 * config/i386/i386.md (ROUND_SAE): Fix value.
6471 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6472 (const48_operand): New.
6473 * config/i386/subst.md (round), (round_expand): Use
6474 const_4_or_8_to_11_operand.
6475 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6477 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6479 * config/i386/constraints.md (Yk): Swap meaning with k.
6480 * config/i386/i386.md (movhi_internal): Change Yk to k.
6481 (movqi_internal): Ditto.
6482 (*k<logic><mode>): Ditto.
6485 (kandn<mode>): Ditto.
6486 (*<code>hi_1): Ditto.
6487 (*<code>qi_1): Ditto.
6488 (kxnor<mode>): Ditto.
6489 (kortestzhi): Ditto.
6490 (kortestchi): Ditto.
6492 (*one_cmplhi2_1): Ditto.
6493 (*one_cmplqi2_1): Ditto.
6494 * config/i386/sse.md (): Change k to Yk.
6495 (avx512f_load<mode>_mask): Ditto.
6496 (avx512f_blendm<mode>): Ditto.
6497 (avx512f_store<mode>_mask): Ditto.
6498 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6499 (avx512f_storedqu<mode>_mask): Ditto.
6500 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6502 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6503 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6504 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6505 (avx512f_maskcmp<mode>3): Ditto.
6506 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6507 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6508 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6509 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6510 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6511 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6512 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6513 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6514 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6515 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6516 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6517 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6518 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6519 (vec_extract_lo_<mode>_maskm): Ditto.
6520 (vec_extract_hi_<mode>_maskm): Ditto.
6521 (avx512f_vternlog<mode>_mask): Ditto.
6522 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6523 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6524 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6525 (avx512f_<code>v8div16qi2_mask): Ditto.
6526 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6527 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6528 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6529 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6530 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6531 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6532 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6533 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6534 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6535 (avx512cd_maskb_vec_dupv8di): Ditto.
6536 (avx512cd_maskw_vec_dupv16si): Ditto.
6537 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6538 (avx512f_vpermi2var<mode>3_mask): Ditto.
6539 (avx512f_vpermi2var<mode>3_mask): Ditto.
6540 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6541 (*avx512f_gathersi<mode>): Ditto.
6542 (*avx512f_gathersi<mode>_2): Ditto.
6543 (*avx512f_gatherdi<mode>): Ditto.
6544 (*avx512f_gatherdi<mode>_2): Ditto.
6545 (*avx512f_scattersi<mode>): Ditto.
6546 (*avx512f_scatterdi<mode>): Ditto.
6547 (avx512f_compress<mode>_mask): Ditto.
6548 (avx512f_compressstore<mode>_mask): Ditto.
6549 (avx512f_expand<mode>_mask): Ditto.
6550 * config/i386/subst.md (mask): Change k to Yk.
6551 (mask_scalar_merge): Ditto.
6554 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6556 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6558 2014-01-31 Richard Biener <rguenther@suse.de>
6561 * builtins.c (fold_builtin_memory_op): Make sure to not
6562 use a floating-point mode or a boolean or enumeral type for
6565 2014-01-30 DJ Delorie <dj@redhat.com>
6567 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6568 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6569 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6570 whenever main() has an epilogue.
6572 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6574 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6575 unused variable "field".
6576 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6577 (vsx_mergeh_<mode>): Likewise.
6578 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6579 (altivec_vmrghh): Likewise.
6580 (altivec_vmrghw): Likewise.
6581 (altivec_vmrglb): Likewise.
6582 (altivec_vmrglh): Likewise.
6583 (altivec_vmrglw): Likewise.
6584 (altivec_vspltb): Add missing uses.
6585 (altivec_vsplth): Likewise.
6586 (altivec_vspltw): Likewise.
6587 (altivec_vspltsf): Likewise.
6589 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6592 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6593 frame related instructions.
6595 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6597 PR rtl-optimization/59959
6598 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6599 any reload of register whose subreg is invalid.
6601 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6603 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6604 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6605 Add missing return type - void.
6607 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6609 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6610 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6611 remove element index adjustment for endian (now handled in vsx.md
6613 (altivec_expand_vec_perm_const): Use
6614 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6615 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6616 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6617 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6618 define_expand and a new define_insn *altivec_vspltb_internal;
6619 adjust for -maltivec=be on a little endian target.
6620 (altivec_vspltb_direct): New.
6621 (altivec_vsplth): Divide into a define_expand and a new
6622 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6623 little endian target.
6624 (altivec_vsplth_direct): New.
6625 (altivec_vspltw): Divide into a define_expand and a new
6626 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6627 little endian target.
6628 (altivec_vspltw_direct): New.
6629 (altivec_vspltsf): Divide into a define_expand and a new
6630 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6631 a little endian target.
6633 2014-01-30 Richard Biener <rguenther@suse.de>
6635 PR tree-optimization/59993
6636 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6637 can propagate form the earlier stmt and avoid the transform
6638 when the intermediate result is needed.
6640 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6642 * README.Portability: Fix typo.
6644 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6646 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6647 comparison_operator with ordered_comparison_operator.
6649 2014-01-30 Nick Clifton <nickc@redhat.com>
6651 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6652 Rename to mn10300_store_multiple_regs.
6653 * config/mn10300/mn10300.c: Likewise.
6654 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6655 store_multiple_regs.
6656 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6657 Call mn10300_store_multiple_regs.
6659 2014-01-30 Nick Clifton <nickc@redhat.com>
6660 DJ Delorie <dj@redhat.com>
6662 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6663 %fp 2 to keep registers after it properly word-aligned.
6664 (rl78_alloc_physical_registers_umul): Handle the case where both
6665 input operands are the same.
6667 2014-01-30 Richard Biener <rguenther@suse.de>
6669 PR tree-optimization/59903
6670 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6673 2014-01-30 Jason Merrill <jason@redhat.com>
6676 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6679 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6681 2014-01-30 Richard Biener <rguenther@suse.de>
6683 PR tree-optimization/59951
6684 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6686 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6689 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6690 SFmode to DFmode case.
6692 2014-01-29 DJ Delorie <dj@redhat.com>
6694 * config/msp430/msp430.opt (-minrt): New.
6695 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6697 (ENDFILE_SPEC): Likewise.
6699 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6701 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6702 (estimate_function_body_sizes): Use it.
6704 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6707 * dwarf2out.c (is_cxx_auto): New.
6708 (is_base_type): Use it.
6709 (gen_type_die_with_usage): Likewise.
6711 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6713 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6714 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6715 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6716 -maltivec=be with LE targets.
6717 (vsx_mergeh_<mode>): Likewise.
6718 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6719 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6720 (altivec_vmrghb): Replace with define_expand and new
6721 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6722 (altivec_vmrghb_direct): New define_insn.
6723 (altivec_vmrghh): Replace with define_expand and new
6724 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6725 (altivec_vmrghh_direct): New define_insn.
6726 (altivec_vmrghw): Replace with define_expand and new
6727 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6728 (altivec_vmrghw_direct): New define_insn.
6729 (*altivec_vmrghsf): Adjust for endianness.
6730 (altivec_vmrglb): Replace with define_expand and new
6731 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6732 (altivec_vmrglb_direct): New define_insn.
6733 (altivec_vmrglh): Replace with define_expand and new
6734 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6735 (altivec_vmrglh_direct): New define_insn.
6736 (altivec_vmrglw): Replace with define_expand and new
6737 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6738 (altivec_vmrglw_direct): New define_insn.
6739 (*altivec_vmrglsf): Adjust for endianness.
6740 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6741 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6742 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6743 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6744 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6745 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6746 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6747 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6749 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6751 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6752 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6755 2014-01-29 Richard Biener <rguenther@suse.de>
6757 PR tree-optimization/58742
6758 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6759 associate_pointerplus_align.
6760 (associate_pointerplus_diff): New function.
6761 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6762 and associate_pointerplus_diff.
6764 2014-01-29 Richard Biener <rguenther@suse.de>
6766 * lto-streamer.h (LTO_major_version): Bump to 3.
6767 (LTO_minor_version): Reset to 0.
6769 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6771 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6772 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6773 (arm_file_start): Generate correct asm header for armv7ve.
6774 * config/arm/bpabi.h: Add multilib support for armv7ve.
6775 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6776 and cortex-a15 to armv7ve.
6777 * config/arm/t-aprofile: Add multilib support for armv7ve.
6778 * doc/invoke.texi: Document -march=armv7ve.
6780 2014-01-29 Richard Biener <rguenther@suse.de>
6782 PR tree-optimization/58742
6783 * tree-ssa-forwprop.c (associate_plusminus): Return true
6784 if we changed sth, defer EH cleanup to ...
6785 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6786 (simplify_mult): New function.
6788 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6791 PR tree-optimization/59920
6792 * tree.c (build_common_builtin_nodes): Remove
6793 __builtin_setjmp_dispatcher initialization.
6794 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6795 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6796 instead of gsi_after_labels + manually skipping debug stmts.
6797 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6798 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6799 * tree-inline.c (copy_edges_for_bb): Remove
6800 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6801 argument. Ignore computed_goto_p stmts. Don't call
6802 make_abnormal_goto_edges. If a call might need abnormal edges
6803 for non-local gotos, see if it already has an edge to
6804 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6805 with true argument, don't do anything then, otherwise add
6806 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6807 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6809 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6810 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6811 (lower_stmt): Don't set data->calls_builtin_setjmp.
6812 (lower_builtin_setjmp): Adjust comment.
6813 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6814 * tree-cfg.c (found_computed_goto): Remove.
6815 (factor_computed_gotos): Remove.
6816 (make_goto_expr_edges): Return bool, true for computed gotos.
6817 Don't call make_abnormal_goto_edges.
6818 (build_gimple_cfg): Don't set found_computed_goto, don't call
6819 factor_computed_gotos.
6820 (computed_goto_p): No longer static.
6821 (make_blocks): Don't set found_computed_goto.
6822 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6823 (make_edges): If make_goto_expr_edges returns true, push bb
6824 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6825 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6826 vector. Record mapping between bbs and OpenMP regions if there
6827 are any, adjust make_gimple_omp_edges caller. Call
6828 handle_abnormal_edges.
6829 (make_abnormal_goto_edges): Remove.
6830 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6831 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6832 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6833 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6834 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6835 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6836 filling *region also set *region_idx to (*region)->entry->index.
6839 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6840 For REGs set ORIGINAL_REGNO.
6842 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6844 * doc/md.texi: Mention that a target shouldn't implement
6845 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6848 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6850 PR tree-optimization/59594
6851 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6852 a copy of the datarefs vector rather than the vector itself.
6854 2014-01-28 Jason Merrill <jason@redhat.com>
6857 * dwarf2out.c (auto_die): New static.
6858 (gen_type_die_with_usage): Handle C++1y 'auto'.
6859 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6862 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6865 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6866 (SPEC_X32): Likewise.
6867 (SPEC_64): Likewise.
6868 * config/i386/i386.c (ix86_option_override_internal): Turn off
6869 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6871 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6872 * config/i386/i386.h (TARGET_16BIT): New macro.
6873 (TARGET_16BIT_P): Likewise.
6874 * config/i386/i386.opt: Add m16.
6875 * doc/invoke.texi: Document -m16.
6877 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6879 PR preprocessor/59935
6880 * input.c (location_get_source_line): Bail out on when line number
6881 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6883 2014-01-28 Richard Biener <rguenther@suse.de>
6885 PR tree-optimization/58742
6886 * tree-ssa-forwprop.c (associate_plusminus): Handle
6887 pointer subtraction of the form (T)(P + A) - (T)P.
6889 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6891 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6894 2014-01-28 Richard Biener <rguenther@suse.de>
6897 2014-01-28 Richard Biener <rguenther@suse.de>
6899 PR rtl-optimization/45364
6900 PR rtl-optimization/59890
6901 * var-tracking.c (local_get_addr_clear_given_value): Handle
6902 already cleared slot.
6903 (val_reset): Handle not allocated local_get_addr_cache.
6904 (vt_find_locations): Use post-order on the inverted CFG.
6906 2014-01-28 Richard Biener <rguenther@suse.de>
6908 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6910 2014-01-28 Richard Biener <rguenther@suse.de>
6912 PR rtl-optimization/45364
6913 PR rtl-optimization/59890
6914 * var-tracking.c (local_get_addr_clear_given_value): Handle
6915 already cleared slot.
6916 (val_reset): Handle not allocated local_get_addr_cache.
6917 (vt_find_locations): Use post-order on the inverted CFG.
6919 2014-01-28 Alan Modra <amodra@gmail.com>
6921 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6922 * configure.ac <recursive call for build != host>: Define
6923 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6924 and LD_FOR_BUILD too.
6925 * configure: Regenerate.
6927 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6929 * config/i386/i386.c (get_builtin_code_for_version): Separate
6930 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6931 Broadwell from Haswell.
6933 2014-01-27 Steve Ellcey <sellcey@mips.com>
6935 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6936 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6937 * config/mips/mips.c (mips_option_override): Change setting
6939 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6940 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6941 Change from Mask to Var.
6943 2014-01-27 Jeff Law <law@redhat.com>
6945 * ipa-inline.c (inline_small_functions): Fix typo.
6947 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6949 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6950 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6951 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6952 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6953 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6954 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6955 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6956 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6957 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6958 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6959 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6960 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6961 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6962 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6963 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6964 (_mm512_storeu_epi64): Ditto.
6965 (_mm512_cmpge_epi32_mask): Ditto.
6966 (_mm512_cmpge_epu32_mask): Ditto.
6967 (_mm512_cmpge_epi64_mask): Ditto.
6968 (_mm512_cmpge_epu64_mask): Ditto.
6969 (_mm512_cmple_epi32_mask): Ditto.
6970 (_mm512_cmple_epu32_mask): Ditto.
6971 (_mm512_cmple_epi64_mask): Ditto.
6972 (_mm512_cmple_epu64_mask): Ditto.
6973 (_mm512_cmplt_epi32_mask): Ditto.
6974 (_mm512_cmplt_epu32_mask): Ditto.
6975 (_mm512_cmplt_epi64_mask): Ditto.
6976 (_mm512_cmplt_epu64_mask): Ditto.
6977 (_mm512_cmpneq_epi32_mask): Ditto.
6978 (_mm512_cmpneq_epu32_mask): Ditto.
6979 (_mm512_cmpneq_epi64_mask): Ditto.
6980 (_mm512_cmpneq_epu64_mask): Ditto.
6981 (_mm512_expand_pd): Ditto.
6982 (_mm512_expand_ps): Ditto.
6983 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6984 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6985 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6986 * config/i386/i386.c (ix86_builtins): Add
6987 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6988 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6989 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6990 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6991 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6992 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6993 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6994 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6995 IX86_BUILTIN_PMOVUSQW512_MEM.
6996 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6997 __builtin_ia32_pmovsqd512mem_mask,
6998 __builtin_ia32_pmovqd512mem_mask,
6999 __builtin_ia32_pmovusqw512mem_mask,
7000 __builtin_ia32_pmovsqw512mem_mask,
7001 __builtin_ia32_pmovqw512mem_mask,
7002 __builtin_ia32_pmovusdw512mem_mask,
7003 __builtin_ia32_pmovsdw512mem_mask,
7004 __builtin_ia32_pmovdw512mem_mask,
7005 __builtin_ia32_pmovqb512mem_mask,
7006 __builtin_ia32_pmovusqb512mem_mask,
7007 __builtin_ia32_pmovsqb512mem_mask,
7008 __builtin_ia32_pmovusdb512mem_mask,
7009 __builtin_ia32_pmovsdb512mem_mask,
7010 __builtin_ia32_pmovdb512mem_mask.
7011 (bdesc_args): Add __builtin_ia32_expanddf512,
7012 __builtin_ia32_expandsf512.
7013 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7014 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7015 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7016 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7017 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7018 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7019 (avx512f_<code>v8div16qi2_mask_store): This.
7020 (avx512f_expand<mode>): New.
7022 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7024 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7026 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7027 (_mm512_prefetch_i32scatter_pd): Ditto.
7028 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7029 (_mm512_prefetch_i64scatter_pd): Ditto.
7030 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7031 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7032 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7033 (_mm512_prefetch_i32scatter_ps): Ditto.
7034 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7035 (_mm512_prefetch_i64scatter_ps): Ditto.
7036 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7037 * config/i386/i386-builtin-types.def: Define
7038 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7039 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7040 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7041 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7042 IX86_BUILTIN_SCATTERPFQPD.
7043 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7044 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7045 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7046 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7047 __builtin_ia32_scatterpfqps.
7048 (ix86_expand_builtin): Expand new built-ins.
7049 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
7050 fix memory access data type.
7051 (*avx512pf_gatherpf<mode>_mask): Ditto.
7052 (*avx512pf_gatherpf<mode>): Ditto.
7053 (avx512pf_scatterpf<mode>): Ditto.
7054 (*avx512pf_scatterpf<mode>_mask): Ditto.
7055 (*avx512pf_scatterpf<mode>): Ditto.
7056 (GATHER_SCATTER_SF_MEM_MODE): New.
7057 (avx512pf_gatherpf<mode>df): Ditto.
7058 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7059 (*avx512pf_scatterpf<mode>df): Ditto.
7061 2014-01-27 Jakub Jelinek <jakub@redhat.com>
7064 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7065 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7068 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7070 * common/config/arm/arm-common.c
7071 (arm_rewrite_mcpu): Handle multiple names.
7073 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7075 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7077 * gimple-builder.h (create_gimple_tmp): Delete.
7079 2014-01-27 Christian Bruel <christian.bruel@st.com>
7081 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7084 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7086 * config/pa/pa.md (call): Generate indirect long calls to non-local
7087 functions when outputing 32-bit code.
7088 (call_value): Likewise except for special call to buggy powf function.
7090 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7091 portable runtime and PIC indirect calls.
7092 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7093 and PIC call sequences. Use ldo instead of blr to set return register
7094 in PIC call sequence.
7096 2014-01-25 Walter Lee <walt@tilera.com>
7098 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7099 avoid clobbering a live register.
7101 2014-01-25 Walter Lee <walt@tilera.com>
7103 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7104 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7105 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7106 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7108 2014-01-25 Walter Lee <walt@tilera.com>
7110 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7111 arguments on even registers.
7112 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7114 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7115 (BIGGEST_ALIGNMENT): Ditto.
7116 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7118 2014-01-25 Walter Lee <walt@tilera.com>
7120 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7121 insns before bundling.
7122 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7124 2014-01-25 Walter Lee <walt@tilera.com>
7126 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7127 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7128 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7130 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7132 * config/mips/constraints.md (kl): Delete.
7133 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7134 define expands, using...
7135 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7136 instructions for MIPS16.
7137 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7138 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7140 2014-01-25 Walter Lee <walt@tilera.com>
7142 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7146 2014-01-25 Walter Lee <walt@tilera.com>
7148 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7149 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7151 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7153 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7156 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7158 * print-rtl.c (in_call_function_usage): New var.
7159 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7160 EXPR_LIST mode as mode and not as reg note name.
7163 * cfgloopmanip.c (copy_loop_info): If
7164 loop->warned_aggressive_loop_optimizations, make sure
7165 the flag is set in target loop too.
7167 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7169 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7171 * builtins.def: Likewise.
7172 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7173 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7174 * ira.c (ira_setup_eliminable_regset): Likewise.
7175 * omp-low.c (gate_expand_omp): Likewise.
7176 (execute_lower_omp): Likewise.
7177 (diagnose_sb_0): Likewise.
7178 (gate_diagnose_omp_blocks): Likewise.
7179 (simd_clone_clauses_extract): Likewise.
7182 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7184 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7185 correction for little endian...
7186 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7189 2014-01-24 Jeff Law <law@redhat.com>
7191 PR tree-optimization/59919
7192 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7193 for non-returning calls.
7195 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7197 * common/config/aarch64/aarch64-common.c
7198 (aarch64_rewrite_mcpu): Handle multiple names.
7199 * config/aarch64/aarch64.h
7200 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7202 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7204 * input.c (add_file_to_cache_tab): Handle the case where fopen
7207 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7210 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7211 from push operand if code of push isn't PRE_DEC.
7213 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7216 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7217 -mquad-memory-atomic. Update -mquad-memory documentation to say
7218 it is only used for non-atomic loads/stores.
7220 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7221 -mquad-memory or -mquad-memory-atomic switches.
7223 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7224 -mquad-memory-atomic to ISA 2.07 support.
7226 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7227 to separate support of normal quad word memory operations (ldq, stq)
7228 from the atomic quad word memory operations.
7230 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7231 support to separate non-atomic quad word operations from atomic
7232 quad word operations. Disable non-atomic quad word operations in
7233 little endian mode so that we don't have to swap words after the
7234 load and before the store.
7235 (quad_load_store_p): Add comment about atomic quad word support.
7236 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7237 options printed with -mdebug=reg.
7239 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7240 -mquad-memory-atomic as the test for whether we have quad word
7241 atomic instructions.
7242 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7243 or -mp8-vector are used, allow byte/half-word atomic operations.
7245 * config/rs6000/sync.md (load_lockedti): Insure that the address
7246 is a proper indexed or indirect address for the lqarx instruction.
7247 On little endian systems, swap the hi/lo registers after the lqarx
7249 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7250 insure the address is valid for the lqarx instruction.
7251 (store_conditionalti): Insure that the address is a proper indexed
7252 or indirect address for the stqcrx. instruction. On little endian
7253 systems, swap the hi/lo registers before doing the stqcrx.
7255 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7256 insure the address is valid for the stqcrx. instruction.
7258 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7259 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7260 type of quad memory support is available.
7262 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7265 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7266 there is a danger of looping.
7268 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7270 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7271 force flag_ira_loop_pressure if set via command line.
7273 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7275 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7276 (ashr_simd): New builtin handling DI mode.
7277 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7278 (aarch64_sshr_simddi): New match pattern.
7279 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7280 (vshrd_n_s64): Likewise.
7281 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7283 2014-01-23 Nick Clifton <nickc@redhat.com>
7285 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7286 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7287 favour of mcu specific scripts.
7288 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7291 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7292 Alex Velenko <Alex.Velenko@arm.com>
7294 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7295 (vaddv_s16): Likewise.
7296 (vaddv_s32): Likewise.
7297 (vaddv_u8): Likewise.
7298 (vaddv_u16): Likewise.
7299 (vaddv_u32): Likewise.
7300 (vaddvq_s8): Likewise.
7301 (vaddvq_s16): Likewise.
7302 (vaddvq_s32): Likewise.
7303 (vaddvq_s64): Likewise.
7304 (vaddvq_u8): Likewise.
7305 (vaddvq_u16): Likewise.
7306 (vaddvq_u32): Likewise.
7307 (vaddvq_u64): Likewise.
7308 (vaddv_f32): Likewise.
7309 (vaddvq_f32): Likewise.
7310 (vaddvq_f64): Likewise.
7311 (vmaxv_f32): Likewise.
7312 (vmaxv_s8): Likewise.
7313 (vmaxv_s16): Likewise.
7314 (vmaxv_s32): Likewise.
7315 (vmaxv_u8): Likewise.
7316 (vmaxv_u16): Likewise.
7317 (vmaxv_u32): Likewise.
7318 (vmaxvq_f32): Likewise.
7319 (vmaxvq_f64): Likewise.
7320 (vmaxvq_s8): Likewise.
7321 (vmaxvq_s16): Likewise.
7322 (vmaxvq_s32): Likewise.
7323 (vmaxvq_u8): Likewise.
7324 (vmaxvq_u16): Likewise.
7325 (vmaxvq_u32): Likewise.
7326 (vmaxnmv_f32): Likewise.
7327 (vmaxnmvq_f32): Likewise.
7328 (vmaxnmvq_f64): Likewise.
7329 (vminv_f32): Likewise.
7330 (vminv_s8): Likewise.
7331 (vminv_s16): Likewise.
7332 (vminv_s32): Likewise.
7333 (vminv_u8): Likewise.
7334 (vminv_u16): Likewise.
7335 (vminv_u32): Likewise.
7336 (vminvq_f32): Likewise.
7337 (vminvq_f64): Likewise.
7338 (vminvq_s8): Likewise.
7339 (vminvq_s16): Likewise.
7340 (vminvq_s32): Likewise.
7341 (vminvq_u8): Likewise.
7342 (vminvq_u16): Likewise.
7343 (vminvq_u32): Likewise.
7344 (vminnmv_f32): Likewise.
7345 (vminnmvq_f32): Likewise.
7346 (vminnmvq_f64): Likewise.
7348 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7350 * config/aarch64/aarch64-simd.md
7351 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7352 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7353 (*aarch64_mul3_elt<mode>): Likewise.
7354 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7355 (*aarch64_mul3_elt_to_64v2df): Likewise.
7356 (*aarch64_mla_elt<mode>): Likewise.
7357 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7358 (*aarch64_mls_elt<mode>): Likewise.
7359 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7360 (*aarch64_fma4_elt<mode>): Likewise.
7361 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7362 (*aarch64_fma4_elt_to_64v2df): Likewise.
7363 (*aarch64_fnma4_elt<mode>): Likewise.
7364 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7365 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7366 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7367 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7368 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7369 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7370 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7371 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7372 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7374 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7376 * config/aarch64/aarch64-simd.md
7377 (aarch64_be_checked_get_lane<mode>): New define_expand.
7378 * config/aarch64/aarch64-simd-builtins.def
7379 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7380 New builtin definition.
7381 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7382 Use new safe be builtin.
7384 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7386 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7388 (aarch64_be_st1<mode>): Likewise.
7389 (aarch_ld1<VALL:mode>): Define_expand modified.
7390 (aarch_st1<VALL:mode>): Likewise.
7391 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7392 (UNSPEC_ST1): Likewise.
7394 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7396 * config/microblaze/microblaze.md: Add trap insn and attribute
7398 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7400 PR preprocessor/58580
7401 * input.h (location_get_source_line): Take an additional line_size
7403 (void diagnostics_file_cache_fini): Declare new function.
7404 * input.c (struct fcache): New type.
7405 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7406 New static constants.
7407 (diagnostic_file_cache_init, total_lines_num)
7408 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7409 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7410 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7411 (get_next_line, read_next_line, goto_next_line, read_line_num):
7412 New static function definitions.
7413 (diagnostic_file_cache_fini): New function.
7414 (location_get_source_line): Take an additional output line_len
7415 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7417 * diagnostic.c (diagnostic_finish): Call
7418 diagnostic_file_cache_fini.
7419 (adjust_line): Take an additional input parameter for the length
7420 of the line, rather than calculating it with strlen.
7421 (diagnostic_show_locus): Adjust the use of
7422 location_get_source_line and adjust_line with respect to their new
7423 signature. While displaying a line now, do not stop at the first
7424 null byte. Rather, display the zero byte as a space and keep
7425 going until we reach the size of the line.
7426 * Makefile.in: Add vec.o to OBJS-libcommon
7428 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7429 Ilya Tocar <ilya.tocar@intel.com>
7431 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7432 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7433 (__builtin_ia32_kmov16): Ditto.
7434 * config/i386/i386.md (UNSPEC_KMOV): New.
7437 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7439 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7440 (_mm512_storeu_si512): Ditto.
7442 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7445 * rtl.h (get_referenced_operands): Declare.
7446 * recog.c (get_referenced_operands): New function.
7447 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7448 operands have been referenced when recording LO_SUM references.
7450 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7452 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7454 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7456 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7457 Enable for generic and recent AMD targets.
7459 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7461 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7462 ARG_SIZE note when adjustment was eliminated.
7464 2014-01-22 Jeff Law <law@redhat.com>
7466 PR tree-optimization/59597
7467 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7468 in file. Accept new argument REGISTERING and use it to modify
7469 dump output appropriately.
7470 (register_jump_thread): Corresponding changes.
7471 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7472 thread paths involving joiner blocks. Add code to dump cancelled
7473 jump threading paths.
7475 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7477 PR rtl-optimization/59477
7478 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7479 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7481 2014-01-22 Tom Tromey <tromey@redhat.com>
7483 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7485 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7487 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7489 PR rtl-optimization/59896
7490 * lra-constraints.c (process_alt_operands): Check unused note for
7491 matched operands of insn with no output reloads.
7493 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7495 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7496 (mips_move_from_gpr_cost): Likewise.
7498 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7500 PR rtl-optimization/59858
7501 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7502 ira_class_hard_regs_num.
7503 (process_alt_operands): Increase reject for dying matched operand.
7505 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7508 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7509 smaller than size, perform several stores or loads and stores
7510 at dst + count - size to store or copy all of size bytes, rather
7511 than just last modesize bytes.
7513 2014-01-20 DJ Delorie <dj@redhat.com>
7515 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7516 that CLOBBERs are REGs before propogating their values.
7518 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7521 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7522 (cgraph_inline_failed_type): New function.
7523 * cgraph.h (DEFCIFCODE): Add type.
7524 (cgraph_inline_failed_type_t): New enum.
7525 (cgraph_inline_failed_type): New prototype.
7526 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7527 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7528 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7529 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7530 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7531 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7532 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7533 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7534 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7535 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7536 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7537 OPTIMIZATION_MISMATCH.
7538 * tree-inline.c (expand_call_inline): Emit errors during
7539 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7541 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7544 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7545 mode attribute in insn output.
7547 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7549 * output.h (output_constant): Delete.
7550 * varasm.c (output_constant): Make private.
7552 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7554 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7556 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7559 * tree.h (fold_builtin_strcat): New prototype.
7560 * builtins.c (fold_builtin_strcat): No longer static. Add len
7561 argument, if non-NULL, don't call c_strlen. Optimize
7562 directly into __builtin_memcpy instead of __builtin_strcpy.
7563 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7564 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7566 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7568 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7569 for SImode_address_operand operands, having only a REG argument.
7571 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7573 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7574 loader name using mbig-endian.
7575 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7577 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7579 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7583 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7585 * config/aarch64/aarch64-protos.h
7586 (aarch64_cannot_change_mode_class_ptr): Declare.
7587 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7588 aarch64_cannot_change_mode_class_ptr): New.
7589 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7590 backend hook aarch64_cannot_change_mode_class.
7592 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7594 * common/config/aarch64/aarch64-common.c
7595 (aarch64_handle_option): Don't handle any option order logic here.
7596 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7597 selected_cpu, warn on architecture version mismatch.
7598 (aarch64_override_options): Fix parsing order for option strings.
7600 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7601 Iain Sandoe <iain@codesourcery.com>
7604 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7605 warning. Amend comment to reflect current functionality.
7607 2014-01-20 Richard Biener <rguenther@suse.de>
7610 * builtins.c (fold_builtin_strcat): Remove case better handled
7611 by tree-ssa-strlen.c.
7613 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7615 * config/aarch64/aarch64.opt
7616 (mcpu, march, mtune): Make case-insensitive.
7618 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7621 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7622 if operands[1] is a REG or ZERO_EXTEND of a REG.
7624 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7626 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7628 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7630 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7631 long non-pic millicode calls.
7633 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7635 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7637 2014-01-19 Kito Cheng <kito@0xlab.org>
7639 * builtins.c (expand_movstr): Check movstr expand done or fail.
7641 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7642 H.J. Lu <hongjiu.lu@intel.com>
7645 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7646 to DImode for zero-extended addresses.
7648 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7650 PR rtl-optimization/57763
7651 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7652 on the new indirect jump_insn and increment LABEL_NUSES (label).
7654 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7658 * config.gcc (x86_archs): New variable.
7659 (x86_64_archs): Likewise.
7660 (x86_cpus): Likewise.
7661 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7662 --with-arch/--with-cpu= options.
7663 Support --with-arch=/--with-cpu={nehalem,westmere,
7664 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7666 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7668 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7669 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7671 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7673 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7675 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7678 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7679 clear cpp_get_options (parse_in)->warn_unused_macros for
7680 ix86_target_macros_internal with cpp_define.
7682 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7684 * jump.c (delete_related_insns): Keep (use (insn))s.
7685 * reorg.c (redundant_insn): Check for barriers too.
7687 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7689 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7691 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7693 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7694 call to $$dyncall when TARGET_LONG_CALLS is true.
7696 2014-01-17 Jeff Law <law@redhat.com>
7698 * ree.c (combine_set_extension): Temporarily disable test for
7699 changing number of hard registers.
7701 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7704 * ipa-inline-analysis.c (inline_free_summary):
7705 Do not free summary of aliases.
7707 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7710 * gimplify.c (gimplify_expr): Use create_tmp_var
7711 instead of create_tmp_var_raw. If cond doesn't have
7712 integral type, don't add the IFN_ANNOTATE builtin at all.
7714 2014-01-17 Martin Jambor <mjambor@suse.cz>
7717 * ipa-cp.c (prev_edge_clone): New variable.
7718 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7719 Also resize prev_edge_clone vector.
7720 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7721 (ipcp_edge_removal_hook): New function.
7722 (ipcp_driver): Register ipcp_edge_removal_hook.
7724 2014-01-17 Andrew Pinski <apinski@cavium.com>
7725 Steve Ellcey <sellcey@mips.com>
7728 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7731 2014-01-17 Jeff Law <law@redhat.com>
7734 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7735 so that pass_ccp runs first.
7737 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7739 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7740 (ix86_adjust_cost): Use !TARGET_XXX.
7741 (do_reorder_for_imul): Likewise.
7742 (swap_top_of_ready_list): Likewise.
7743 (ix86_sched_reorder): Likewise.
7745 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7747 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7748 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7749 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7750 (intel_memset): New. Duplicate slm_memset.
7751 (intel_cost): New. Duplicate slm_cost.
7752 (m_INTEL): New macro.
7753 (processor_target_table): Add "intel".
7754 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7755 with PROCESSOR_INTEL for "intel".
7756 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7757 PROCESSOR_SILVERMONT.
7758 (ix86_issue_rate): Likewise.
7759 (ix86_adjust_cost): Likewise.
7760 (ia32_multipass_dfa_lookahead): Likewise.
7761 (swap_top_of_ready_list): Likewise.
7762 (ix86_sched_reorder): Likewise.
7763 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7764 instead of TARGET_OPT_AGU.
7765 * config/i386/i386.h (TARGET_INTEL): New.
7766 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7767 (processor_type): Add PROCESSOR_INTEL.
7768 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7769 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7771 2014-01-17 Marek Polacek <polacek@redhat.com>
7774 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7777 2014-01-17 Richard Biener <rguenther@suse.de>
7779 PR tree-optimization/46590
7780 * opts.c (default_options_table): Add entries for
7781 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7782 all enabled at -O1 but not for -Og.
7783 * common.opt (fbranch-count-reg): Remove Init(1).
7784 (fmove-loop-invariants): Likewise.
7785 (ftree-pta): Likewise.
7787 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7789 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7790 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7791 decls to at least the GCC 4.8 used alignments.
7794 * tree-nested.c (convert_nonlocal_reference_stmt,
7795 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7796 of GIMPLE_BIND stmts, adjust associated decls.
7798 2014-01-17 Richard Biener <rguenther@suse.de>
7800 PR tree-optimization/46590
7801 * vec.h (vec<>::bseach): New member function implementing
7802 binary search according to C89 bsearch.
7803 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7804 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7805 bitmap pointer again. Make accesses_in_loop a flat array.
7806 (mem_ref_obstack): New global.
7807 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7808 (mark_ref_stored): Likewise.
7809 (ref_indep_loop_p_2): Likewise.
7810 (set_ref_stored_in_loop): New helper function.
7811 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7812 (memref_free): Adjust.
7813 (record_mem_ref_loc): Simplify.
7814 (gather_mem_refs_stmt): Adjust.
7815 (sort_locs_in_loop_postorder_cmp): New function.
7816 (analyze_memory_references): Sort accesses_in_loop after
7817 loop postorder number.
7818 (find_ref_loc_in_loop_cmp): New function.
7819 (for_all_locs_in_loop): Find relevant cluster of locs in
7820 accesses_in_loop and iterate without recursion.
7821 (execute_sm): Avoid uninit warning.
7822 (struct ref_always_accessed): Simplify.
7823 (ref_always_accessed::operator ()): Likewise.
7824 (ref_always_accessed_p): Likewise.
7825 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7826 loop postorder numbers here.
7827 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7830 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7833 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7834 on decls for which assemble_alias has been called.
7836 2014-01-17 Nick Clifton <nickc@redhat.com>
7838 * config/msp430/msp430.opt: (mcpu): New option.
7839 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7840 (msp430_option_override): Parse target_cpu. If the MCU name
7841 matches a generic string, clear target_mcu.
7842 (msp430_attr): Allow numeric interrupt values up to 63.
7843 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7844 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7846 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7848 * config/msp430/msp430.md (popm): Use %J rather than %I.
7849 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7850 (addhi_cy_i): Use immediate_operand for operand 2.
7851 * doc/invoke.texi: Document -mcpu option.
7853 2014-01-17 Richard Biener <rguenther@suse.de>
7855 PR rtl-optimization/38518
7856 * df.h (df_analyze_loop): Declare.
7857 * df-core.c: Include cfgloop.h.
7858 (df_analyze_1): Split out main part of df_analyze.
7859 (df_analyze): Adjust.
7860 (loop_inverted_post_order_compute): New function.
7861 (loop_post_order_compute): Likewise.
7862 (df_analyze_loop): New function avoiding whole-function
7864 * loop-invariant.c (find_defs): Use df_analyze_loop.
7865 (find_invariants): Adjust.
7866 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7868 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7870 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7871 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7873 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7875 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7876 traversal when removing references.
7878 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7881 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7883 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7886 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7887 pushing a reload for an autoinc when we had previously reloaded an
7888 inner part of the address.
7890 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7892 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7894 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7895 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7896 when not giving up or versioning for alias only because of
7898 (vect_analyze_data_ref_dependences): Set to true.
7899 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7901 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7902 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7906 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7909 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7910 operand 0 predicate for gathers, use a new pseudo as subtarget.
7912 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7915 * lra-constraints.c (process_alt_operands): Add printing debug info.
7916 Check absence of input/output reloads for matched operands too.
7918 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7920 PR rtl-optimization/59835
7921 * ira.c (ira_init_register_move_cost): Increase cost for
7924 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7926 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7928 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7931 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7932 non-register objects. Use gen_(high/low)part more consistently.
7935 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7938 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7939 endian support, remove tests for WORDS_BIG_ENDIAN.
7940 (p8_mfvsrd_3_<mode>): Likewise.
7941 (reload_gpr_from_vsx<mode>): Likewise.
7942 (reload_gpr_from_vsxsf): Likewise.
7943 (p8_mfvsrd_4_disf): Likewise.
7945 2014-01-16 Richard Biener <rguenther@suse.de>
7947 PR rtl-optimization/46590
7948 * lcm.c (compute_antinout_edge): Use postorder iteration.
7949 (compute_laterin): Use inverted postorder iteration.
7951 2014-01-16 Nick Clifton <nickc@redhat.com>
7954 * varasm.c (output_constant): Return the number of bytes actually
7956 (output_constructor_array_range): Update the field size with the
7957 number of bytes emitted by output_constant.
7958 (output_constructor_regular_field): Likewise. Also do not
7959 complain if the total number of bytes emitted is now greater
7960 than the expected fieldpos.
7961 * output.h (output_constant): Update prototype and descriptive comment.
7963 2014-01-16 Marek Polacek <polacek@redhat.com>
7966 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7967 it is error_mark_node.
7969 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7971 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7972 VALID_AVX256_REG_OR_OI_MODE.
7974 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7976 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7977 current procedure should be profiled.
7979 2014-01-15 Andrew Pinski <apinski@cavium.com>
7981 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7982 of moving from/to the STACK_REG register class.
7984 2014-01-15 Richard Henderson <rth@redhat.com>
7987 * reginfo.c (global_regs_decl): Globalize.
7988 * rtl.h (global_regs_decl): Declare.
7989 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7990 reserved via global_regs.
7992 2014-01-15 Teresa Johnson <tejohnson@google.com>
7994 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7996 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7998 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7999 and vmulosh rather than call gen_vec_widen_smult_*.
8000 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8001 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8002 (vec_widen_smult_even_v16qi): Likewise.
8003 (vec_widen_umult_even_v8hi): Likewise.
8004 (vec_widen_smult_even_v8hi): Likewise.
8005 (vec_widen_umult_odd_v16qi): Likewise.
8006 (vec_widen_smult_odd_v16qi): Likewise.
8007 (vec_widen_umult_odd_v8hi): Likewise.
8008 (vec_widen_smult_odd_v8hi): Likewise.
8009 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8010 vmuloub rather than call gen_vec_widen_umult_*.
8011 (vec_widen_umult_lo_v16qi): Likewise.
8012 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8013 vmulosb rather than call gen_vec_widen_smult_*.
8014 (vec_widen_smult_lo_v16qi): Likewise.
8015 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8016 rather than call gen_vec_widen_umult_*.
8017 (vec_widen_umult_lo_v8hi): Likewise.
8018 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8019 rather than call gen_vec_widen_smult_*.
8020 (vec_widen_smult_lo_v8hi): Likewise.
8022 2014-01-15 Jeff Law <law@redhat.com>
8024 PR tree-optimization/59747
8025 * ree.c (find_and_remove_re): Properly handle case where a second
8026 eliminated extension requires widening a copy created for elimination
8027 of a prior extension.
8028 (combine_set_extension): Ensure that the number of hard regs needed
8029 for a destination register does not change when we widen it.
8031 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8033 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8034 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8035 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8036 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8037 (avr-*-rtems*): Likewise.
8038 (bfin*-rtems*): Likewise.
8039 (moxie-*-rtems*): Likewise.
8040 (h8300-*-rtems*): Likewise.
8041 (i[34567]86-*-rtems*): Likewise.
8042 (lm32-*-rtems*): Likewise.
8043 (m32r-*-rtems*): Likewise.
8044 (m68k-*-rtems*): Likewise.
8045 (microblaze*-*-rtems*): Likewise.
8046 (mips*-*-rtems*): Likewise.
8047 (powerpc-*-rtems*): Likewise.
8048 (sh-*-rtems*): Likewise.
8049 (sparc-*-rtems*): Likewise.
8050 (sparc64-*-rtems*): Likewise.
8051 (v850-*-rtems*): Likewise.
8052 (m32c-*-rtems*): Likewise.
8054 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
8056 PR rtl-optimization/59511
8057 * ira.c (ira_init_register_move_cost): Use memory costs for some
8058 cases of register move cost calculations.
8059 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
8060 instead of BB frequency.
8061 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
8062 * lra-assigns.c (find_hard_regno_for): Ditto.
8064 2014-01-15 Richard Biener <rguenther@suse.de>
8066 PR tree-optimization/59822
8067 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8068 (vectorizable_load): Use it to hoist defs of uses of invariant
8069 loads out of the loop.
8071 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8072 Kugan Vivekanandarajah <kuganv@linaro.org>
8075 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8078 2014-01-15 Richard Biener <rguenther@suse.de>
8080 PR rtl-optimization/59802
8081 * lcm.c (compute_available): Use inverted postorder to seed
8082 the initial worklist.
8084 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8087 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8088 ADDR_REGS for invalid symrefs in non-PIC code.
8090 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8093 * builtins.c (determine_block_size): Initialize *probable_max_size
8094 even if len_rtx is CONST_INT.
8096 2014-01-14 Andrew Pinski <apinski@cavium.com>
8098 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8099 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8100 (cortexa53_tunings): Likewise.
8101 (aarch64_sched_issue_rate): New function.
8102 (TARGET_SCHED_ISSUE_RATE): Define.
8104 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8106 * ira-costs.c (find_costs_and_classes): Add missed
8107 ira_init_register_move_cost_if_necessary.
8109 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8112 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8114 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8117 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8118 to indicate if type is used for function return value. Warn ABI
8119 change if the vector mode isn't available for function return value.
8120 (ix86_function_arg_advance): Pass false to type_natural_mode.
8121 (ix86_function_arg): Likewise.
8122 (ix86_gimplify_va_arg): Likewise.
8123 (function_arg_32): Don't warn ABI change.
8124 (ix86_function_value): Pass true to type_natural_mode.
8125 (ix86_return_in_memory): Likewise.
8126 (ix86_struct_value_rtx): Removed.
8127 (TARGET_STRUCT_VALUE_RTX): Likewise.
8129 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8131 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8132 converting a conditional jump into a conditional return.
8134 2014-01-14 Richard Biener <rguenther@suse.de>
8136 PR tree-optimization/58921
8137 PR tree-optimization/59006
8138 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8139 hoisting invariant stmts.
8140 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8141 invariant loads on the preheader edge if possible.
8143 2014-01-14 Joey Ye <joey.ye@arm.com>
8145 * doc/plugin.texi (Building GCC plugins): Update to C++.
8147 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8149 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8150 (_mm_rcp28_round_ss): Ditto.
8151 (_mm_rsqrt28_round_sd): Ditto.
8152 (_mm_rsqrt28_round_ss): Ditto.
8153 (_mm_rcp28_sd): Ditto.
8154 (_mm_rcp28_ss): Ditto.
8155 (_mm_rsqrt28_sd): Ditto.
8156 (_mm_rsqrt28_ss): Ditto.
8157 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8158 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8159 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8160 (IX86_BUILTIN_RCP28SD): Ditto.
8161 (IX86_BUILTIN_RCP28SS): Ditto.
8162 (IX86_BUILTIN_RSQRT28SD): Ditto.
8163 (IX86_BUILTIN_RSQRT28SS): Ditto.
8164 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8165 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8166 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8167 (ix86_expand_special_args_builtin): Expand new FTYPE.
8168 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8169 (srcp14<mode>): Make insn unary.
8170 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8171 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8172 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8173 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8174 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8175 Fix rounding: make it SAE only.
8176 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8178 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8180 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8181 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8182 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8183 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8184 (round_saeonly_mask_scalar_operand4): Ditto.
8185 (round_saeonly_mask_scalar_op3): Ditto.
8186 (round_saeonly_mask_scalar_op4): Ditto.
8188 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8190 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8191 Implement -maltivec=be for vec_insert and vec_extract.
8193 2014-01-10 DJ Delorie <dj@redhat.com>
8195 * config/msp430/msp430.md (call_internal): Don't allow memory
8196 references with SP as the base register.
8197 (call_value_internal): Likewise.
8198 * config/msp430/constraints.md (Yc): New. For memory references
8199 that don't use SP as a base register.
8201 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8202 "an integer without a # prefix"
8203 * config/msp430/msp430.md (epilogue_helper): Use it.
8205 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8208 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8209 AVX512F gather builtins.
8210 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8211 on gather decls with INTEGER_TYPE masktype.
8212 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8213 directly into the builtin rather than hoisting it before loop.
8215 PR tree-optimization/59387
8216 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8217 (scev_const_prop): If folded_casts and type has undefined overflow,
8218 use force_gimple_operand instead of force_gimple_operand_gsi and
8219 for each added stmt if it is assign with
8220 arith_code_with_undefined_signed_overflow, call
8221 rewrite_to_defined_overflow.
8222 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8223 gimple-fold.h instead.
8224 (arith_code_with_undefined_signed_overflow,
8225 rewrite_to_defined_overflow): Moved to ...
8226 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8227 rewrite_to_defined_overflow): ... here. No longer static.
8228 Include gimplify-me.h.
8229 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8230 rewrite_to_defined_overflow): New prototypes.
8232 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8234 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8236 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8238 * builtins.c (get_object_alignment_2): Minor tweak.
8239 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8241 2014-01-13 Christian Bruel <christian.bruel@st.com>
8243 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8244 optimized non constant lengths.
8246 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8249 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8250 load as __atomic_load_N if possible.
8252 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8254 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8256 (rs6000_expand_builtin): Adjust call.
8258 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8261 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8262 * config/rs6000/rs6000.c: Include target-globals.h.
8263 (rs6000_set_current_function): Instead of doing target_reinit
8264 unconditionally, use save_target_globals_default_opts and
8265 restore_target_globals.
8267 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8269 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8270 (rs6000_expand_builtin): Handle mffs and mtfsf.
8271 (rs6000_init_builtins): Define mffs and mtfsf.
8272 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8273 (rs6000_mffs): New pattern.
8274 (rs6000_mtfsf): New pattern.
8276 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8278 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8279 Start narrowing with START. Apply candidate-use pair
8280 and check overall cost in narrowing.
8281 (iv_ca_prune): Pass new argument.
8283 2014-01-10 Jeff Law <law@redhat.com>
8286 * ree.c (combine_reaching_defs): Ensure the defining statement
8287 occurs before the extension when optimizing extensions with
8288 different source and destination hard registers.
8290 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8293 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8294 vtables into the type inheritance graph.
8296 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8298 PR rtl-optimization/59754
8299 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8300 modes in the REGNO != REGNO case.
8302 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8304 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8306 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8308 PR tree-optimization/59745
8309 * tree-predcom.c (tree_predictive_commoning_loop): Call
8310 free_affine_expand_cache if giving up because components is NULL.
8312 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8313 GC in payload of target_globals struct instead of allocating them on
8314 the heap and the larger structs separately using GC.
8315 * target-globals.h (struct target_globals): Make regs, hard_regs,
8316 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8317 of GTY((skip)) and change type to void *.
8318 (reset_target_globals): Cast loads from those fields to corresponding
8321 2014-01-10 Steve Ellcey <sellcey@mips.com>
8324 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8325 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8326 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8328 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8331 * aarch64-modes.def (CC_Zmode): New flags mode.
8332 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8333 represents an equality.
8334 (aarch64_get_condition_code): Handle CC_Zmode.
8335 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8337 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8339 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8340 extraction in good case.
8342 2014-01-10 Richard Biener <rguenther@suse.de>
8344 PR tree-optimization/59374
8345 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8346 checking after SLP discovery. Mark stmts not participating
8347 in any SLP instance properly.
8349 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8351 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8352 when handling a SET rtx.
8354 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8356 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8357 (cortex-a57): Likewise.
8358 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8360 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8362 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8363 non-iwmmxt builtins.
8365 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8369 * ipa-devirt.c record_target_from_binfo): Take as argument
8370 stack of binfos and lookup matching one for virtual inheritance.
8371 (possible_polymorphic_call_targets_1): Update.
8373 2014-01-10 Huacai Chen <chenhc@lemote.com>
8375 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8376 kernel strings for Loongson-2E/2F/3A.
8378 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8381 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8382 is_gimple_call before calling gimple_call_internal_p.
8384 2014-01-09 Steve Ellcey <sellcey@mips.com>
8386 * Makefile.in (TREE_FLOW_H): Remove.
8387 (TREE_SSA_H): Add file names from tree-flow.h.
8388 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8389 * tree.h: Remove tree-flow.h reference.
8390 * hash-table.h: Remove tree-flow.h reference.
8391 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8392 reference with tree-ssa-loop.h.
8394 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8396 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8397 default element-order behavior for -maltivec.
8398 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8399 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8400 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8401 when targeting big endian, at least for now.
8402 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8404 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8407 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8408 var satisfies use_register_for_decl, just take into account type
8409 alignment, rather than decl alignment.
8411 PR tree-optimization/59622
8412 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8413 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8414 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8415 Don't devirtualize for inplace at all. For targets.length () == 1,
8416 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8418 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8420 * config/i386/i386.md (cpu): Remove the unused btver1.
8422 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8424 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8426 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8429 * tree-core.h (struct target_globals): New forward declaration.
8430 (struct tree_target_option): Add globals field.
8431 * tree.h (TREE_TARGET_GLOBALS): Define.
8432 (prepare_target_option_nodes_for_pch): New prototype.
8433 * target-globals.h (struct target_globals): Define even if
8435 * tree.c (prepare_target_option_node_for_pch,
8436 prepare_target_option_nodes_for_pch): New functions.
8437 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8438 * config/i386/i386.c: Include target-globals.h.
8439 (ix86_set_current_function): Instead of doing target_reinit
8440 unconditionally, use save_target_globals_default_opts and
8441 restore_target_globals.
8443 2014-01-09 Richard Biener <rguenther@suse.de>
8445 PR tree-optimization/59715
8446 * tree-cfg.h (split_critical_edges): Declare.
8447 * tree-cfg.c (split_critical_edges): Export.
8448 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8450 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8452 * cfgexpand.c (expand_stack_vars): Optionally disable
8453 asan stack protection.
8454 (expand_used_vars): Likewise.
8455 (partition_stack_vars): Likewise.
8456 * asan.c (asan_emit_stack_protection): Optionally disable
8457 after return stack usage.
8458 (instrument_derefs): Optionally disable memory access instrumentation.
8459 (instrument_builtin_call): Likewise.
8460 (instrument_strlen_call): Likewise.
8461 (asan_protect_global): Optionally disable global variables protection.
8462 * doc/invoke.texi: Added doc for new options.
8463 * params.def: Added new options.
8464 * params.h: Likewise.
8466 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8468 PR rtl-optimization/59724
8469 * ifcvt.c (cond_exec_process_if_block): Don't call
8470 flow_find_head_matching_sequence with 0 longest_match.
8471 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8472 non-active insns if !stop_after.
8473 (try_head_merge_bb): Revert 2014-01-07 changes.
8475 2014-01-08 Jeff Law <law@redhat.com>
8477 * ree.c (get_sub_rtx): New function, extracted from...
8478 (merge_def_and_ext): Here.
8479 (combine_reaching_defs): Use get_sub_rtx.
8481 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8483 * cgraph.h (varpool_variable_node): Do not choke on null node.
8485 2014-01-08 Catherine Moore <clm@codesourcery.com>
8487 * config/mips/mips.md (simple_return): Attempt to use JRC
8489 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8491 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8493 PR rtl-optimization/59137
8494 * reorg.c (steal_delay_list_from_target): Call update_block for
8496 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8498 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8500 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8501 two duplicate entries.
8503 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8506 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8508 * config/mips/mips.c (mips_truncated_op_cost): New function.
8509 (mips_rtx_costs): Adjust test for BADDU.
8510 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8512 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8514 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8515 (*baddu_si): ...this new pattern.
8517 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8520 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8522 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8525 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8527 (expand_expr, expand_normal): Adjust.
8528 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8529 inner_reference_p. Use inner_reference_p to expand inner references.
8530 (store_expr): Adjust.
8531 * cfgexpand.c (expand_call_stmt): Adjust.
8533 2014-01-08 Rong Xu <xur@google.com>
8535 * gcov-io.c (gcov_var): Move from gcov-io.h.
8536 (gcov_position): Ditto.
8537 (gcov_is_error): Ditto.
8538 (gcov_rewrite): Ditto.
8539 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8540 only part to libgcc/libgcov.h.
8542 2014-01-08 Marek Polacek <polacek@redhat.com>
8545 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8547 2014-01-08 Marek Polacek <polacek@redhat.com>
8550 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8552 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8554 PR rtl-optimization/59649
8555 * stor-layout.c (get_mode_bounds): For BImode return
8556 0 and STORE_FLAG_VALUE.
8558 2014-01-08 Richard Biener <rguenther@suse.de>
8561 * gimple.h (is_gimple_builtin_call): Remove.
8562 (gimple_builtin_call_types_compatible_p): New.
8563 (gimple_call_builtin_p): New overload.
8564 * gimple.c (is_gimple_builtin_call): Remove.
8565 (validate_call): Rename to ...
8566 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8568 (validate_type): New static function.
8569 (gimple_call_builtin_p): New overload and adjust.
8570 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8571 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8572 (gimple_fold_stmt_to_constant_1): Likewise.
8573 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8575 2014-01-08 Richard Biener <rguenther@suse.de>
8578 * gimplify.c (gimplify_expr): Gimplify register-register type
8579 VIEW_CONVERT_EXPRs to separate stmts.
8581 2014-01-07 Jeff Law <law@redhat.com>
8584 * ree.c (combine_set_extension): Handle case where source
8585 and destination registers in an extension insn are different.
8586 (combine_reaching_defs): Allow source and destination registers
8587 in extension to be different under limited circumstances.
8588 (add_removable_extension): Remove restriction that the
8589 source and destination registers in the extension are the same.
8590 (find_and_remove_re): Emit a copy from the extension's
8591 destination to its source after the defining insn if
8592 the source and destination registers are different.
8595 * ifcvt.c (merge_if_block): If we are merging a block with more than
8596 one successor with a block with no successors, remove any BARRIER
8597 after the second block.
8599 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8601 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8603 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8606 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8607 for 14-bit register offsets when INT14_OK_STRICT is false.
8609 2014-01-07 Roland Stigge <stigge@antcom.de>
8610 Michael Meissner <meissner@linux.vnet.ibm.com>
8613 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8614 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8616 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8618 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8621 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8623 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8624 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8625 rtx is const0_rtx or not.
8627 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8630 * target-globals.c (save_target_globals): Remove this_fn_optab
8632 * toplev.c: Include optabs.h.
8633 (target_reinit): Temporarily restore the global options if another
8634 set of options are in force.
8636 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8638 PR rtl-optimization/58668
8639 * cfgcleanup.c (flow_find_cross_jump): Don't count
8640 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8641 to determine what is counted.
8642 (flow_find_head_matching_sequence): Use active_insn_p to determine
8644 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8646 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8647 determine what is counted.
8649 PR tree-optimization/59643
8650 * tree-predcom.c (split_data_refs_to_components): If one dr is
8651 read and one write, determine_offset fails and the write isn't
8652 in the bad component, just put the read into the bad component.
8654 2014-01-07 Mike Stump <mikestump@comcast.net>
8655 Jakub Jelinek <jakub@redhat.com>
8658 * tree-core.h (struct tree_optimization_option): Change optabs
8659 type from unsigned char * to void *.
8660 * optabs.c (init_tree_optimization_optabs): Adjust
8661 TREE_OPTIMIZATION_OPTABS initialization.
8663 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8666 * config/i386/i386.h (struct machine_function): Add
8667 no_drap_save_restore field.
8668 * config/i386/i386.c (ix86_save_reg): Use
8669 !cfun->machine->no_drap_save_restore instead of
8670 crtl->stack_realign_needed.
8671 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8672 this function clears frame_pointer_needed. Set
8673 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8674 and DRAP reg is needed.
8676 2014-01-06 Marek Polacek <polacek@redhat.com>
8679 * doc/implement-c.texi: Mention that other integer types are
8680 permitted as bit-field types in strictly conforming mode.
8682 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8684 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8687 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8689 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8691 2014-01-06 Martin Jambor <mjambor@suse.cz>
8694 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8696 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8698 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8702 * var-tracking.c (add_stores): Preserve the value of the source even if
8703 we don't record the store.
8705 2014-01-06 Terry Guo <terry.guo@arm.com>
8707 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8709 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8712 * config/darwin.c (darwin_function_section): Adjust return values to
8713 correspond to optimisation changes made in r206070.
8715 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8717 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8718 from prefetch_block tune setting.
8719 (nocona_cost): Correct size of prefetch block to 64.
8721 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8723 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8724 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8725 used to save the static chain register in the computation of the offset
8726 from which the FP registers need to be restored.
8728 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8730 PR tree-optimization/59519
8731 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8732 ICE if get_current_def (current_new_name) is already non-NULL, as long
8733 as it is a phi result of some other phi in *new_exit_bb that has
8736 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8737 or vmovdqu* for misaligned_operand.
8738 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8739 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8740 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8741 aligned_mem for AVX512F masked aligned load and store builtins and for
8744 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8746 PR tree-optimization/59651
8747 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8748 Address range for negative step should be added by TYPE_SIZE_UNIT.
8750 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8752 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8753 overlapping registers also for registers other than the stack pointer.
8755 2014-01-03 Marek Polacek <polacek@redhat.com>
8758 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8761 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8764 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8767 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8768 (push splitter): Use <P:MODE_SIZE> instead of
8769 GET_MODE_SIZE (<P:MODE>mode).
8770 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8771 (mov -1, reg peephole2): Likewise.
8772 * config/i386/sse.md (*mov<mode>_internal,
8773 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8774 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8775 *<code><mode>3, *andnot<mode>3<mask_name>,
8776 <mask_codefor><code><mode>3<mask_name>): Likewise.
8777 * config/i386/subst.md (mask_mode512bit_condition,
8778 sd_mask_mode512bit_condition): Likewise.
8780 2014-01-02 Xinliang David Li <davidxl@google.com>
8782 PR tree-optimization/59303
8783 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8784 (dump_predicates): Better output format.
8785 (pred_equal_p): New function.
8786 (is_neq_relop_p): Ditto.
8787 (is_neq_zero_form_p): Ditto.
8788 (pred_expr_equal_p): Ditto.
8789 (pred_neg_p): Ditto.
8790 (simplify_pred): Ditto.
8791 (simplify_preds_2): Ditto.
8792 (simplify_preds_3): Ditto.
8793 (simplify_preds_4): Ditto.
8794 (simplify_preds): Ditto.
8796 (push_to_worklist): Ditto.
8797 (get_pred_info_from_cmp): Ditto.
8798 (is_degenerated_phi): Ditto.
8799 (normalize_one_pred_1): Ditto.
8800 (normalize_one_pred): Ditto.
8801 (normalize_one_pred_chain): Ditto.
8802 (normalize_preds): Ditto.
8803 (normalize_cond_1): Remove function.
8804 (normalize_cond): Ditto.
8805 (is_gcond_subset_of): Ditto.
8806 (is_subset_of_any): Ditto.
8807 (is_or_set_subset_of): Ditto.
8808 (is_and_set_subset_of): Ditto.
8809 (is_norm_cond_subset_of): Ditto.
8810 (pred_chain_length_cmp): Ditto.
8811 (convert_control_dep_chain_into_preds): Type change.
8812 (find_predicates): Ditto.
8813 (find_def_preds): Ditto.
8814 (destroy_predicates_vecs): Ditto.
8815 (find_matching_predicates_in_rest_chains): Ditto.
8816 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8817 (is_pred_expr_subset): Ditto.
8818 (is_pred_chain_subset_of): Ditto.
8819 (is_included_in): Ditto.
8820 (is_superset_of): Ditto.
8822 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8824 Update copyright years.
8826 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8828 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8829 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8830 config/arc/arc.md, config/arc/arc.opt,
8831 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8832 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8833 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8834 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8835 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8836 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8837 vtable-verify.c, vtable-verify.h: Use the standard form for the
8840 2014-01-02 Tobias Burnus <burnus@net-b.de>
8842 * gcc.c (process_command): Update copyright notice dates.
8843 * gcov-dump.c: Ditto.
8845 * doc/cpp.texi: Bump @copying's copyright year.
8846 * doc/cppinternals.texi: Ditto.
8847 * doc/gcc.texi: Ditto.
8848 * doc/gccint.texi: Ditto.
8849 * doc/gcov.texi: Ditto.
8850 * doc/install.texi: Ditto.
8851 * doc/invoke.texi: Ditto.
8853 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8855 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8857 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8859 * config/i386/sse.md (*mov<mode>_internal): Guard
8860 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8862 PR rtl-optimization/59647
8863 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8864 new_rtx into UNSIGNED_FLOAT rtxes.
8866 Copyright (C) 2014 Free Software Foundation, Inc.
8868 Copying and distribution of this file, with or without modification,
8869 are permitted in any medium without royalty provided the copyright
8870 notice and this notice are preserved.