1 2014-05-05 Richard Biener <rguenther@suse.de>
4 * fold-const.c (fold_binary_loc): Consistently avoid
5 canonicalizing X & CST away from a CST that is the mask
8 2014-05-05 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10 * config/picochip/picochip-protos.h (picochip_regno_nregs): Change
11 int argument to enum machine_mode.
12 (picochip_class_max_nregs): Ditto.
13 * config/picochip/picochip.c (picochip_regno_nregs): Ditto.
14 (picochip_class_max_nregs): Ditto.
16 2014-05-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18 * target.def: Add new target hook.
19 * doc/tm.texi: Regenerate.
20 * targhooks.h (default_keep_leaf_when_profiled): Add prototype.
21 * targhooks.c (default_keep_leaf_when_profiled): New function.
23 * config/s390/s390.c (s390_keep_leaf_when_profiled): New function.
24 (TARGET_KEEP_LEAF_WHEN_PROFILED): Define.
26 2014-05-05 Bin Cheng <bin.cheng@arm.com>
28 PR tree-optimization/60363
29 * gcc/tree-ssa-threadupdate.c (get_value_locus_in_path): New.
30 (copy_phi_args): New parameters. Call get_value_locus_in_path.
31 (update_destination_phis): New parameter.
32 (create_edge_and_update_destination_phis): Ditto.
33 (ssa_fix_duplicate_block_edges): Pass new arguments.
34 (thread_single_edge): Ditto.
36 2014-05-04 Peter Bergner <bergner@vnet.ibm.com>
38 * config/rs6000/rs6000.h (RS6000_BTM_HARD_FLOAT): New define.
39 (RS6000_BTM_COMMON): Add RS6000_BTM_HARD_FLOAT.
40 (TARGET_EXTRA_BUILTINS): Add TARGET_HARD_FLOAT.
41 * config/rs6000/rs6000-builtin.def (BU_MISC_1):
42 Use RS6000_BTM_HARD_FLOAT.
43 (BU_MISC_2): Likewise.
44 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Handle
45 RS6000_BTM_HARD_FLOAT.
46 (rs6000_option_override_internal): Enforce -mhard-float if -mhard-dfp
48 (rs6000_invalid_builtin): Add hard floating builtin support.
49 (rs6000_expand_builtin): Relax the gcc_assert to allow the new
51 (rs6000_builtin_mask_names): Add RS6000_BTM_HARD_FLOAT.
53 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
55 * config/sh/sh_optimize_sett_clrt.cc (sh_optimize_sett_clrt::execute):
56 Add missing function* argument.
58 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
60 * lra-constraints.c (valid_address_p): Move earlier in file.
61 Add a constraint argument to the address_info version.
62 (satisfies_memory_constraint_p): New function.
63 (satisfies_address_constraint_p): Likewise.
64 (process_alt_operands, curr_insn_transform): Use them.
65 (process_address): Pass the constraint to valid_address_p when
66 checking address operands.
68 2014-05-03 Richard Sandiford <rdsandiford@googlemail.com>
70 * config/mips/mips.c (mips_isa_rev): New variable.
71 (mips_set_architecture): Set it.
72 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Set __mips_isa_rev
74 (ISA_HAS_MUL3, ISA_HAS_FP_CONDMOVE, ISA_HAS_8CC, ISA_HAS_FP4)
75 (ISA_HAS_PAIRED_SINGLE, ISA_HAS_MADD_MSUB, ISA_HAS_FP_RECIP_RSQRT)
76 (ISA_HAS_CLZ_CLO, ISA_HAS_ROR, ISA_HAS_WSBH, ISA_HAS_PREFETCH)
77 (ISA_HAS_SEB_SEH, ISA_HAS_EXT_INS, ISA_HAS_MXHC1)
78 (ISA_HAS_HILO_INTERLOCKS, ISA_HAS_SYNCI, MIN_FPRS_PER_FMT): Reexpress
79 conditions in terms of mips_isa_rev.
80 (mips_isa_rev): Declare.
82 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
84 * config/sh/sh-mem.cc: Use tabs instead of spaces.
85 (prob_unlikely, prob_likely): Make variables const.
87 2014-05-03 Denis Chertykov <chertykov@gmail.com>
89 * config/avr/avr.c (avr_adjust_insn_length): Handle JUMP_TABLE_DATA.
91 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
93 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
95 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
97 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
98 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
99 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
101 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
102 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
104 Replace usage of ROUND_REG with sh_round_reg.
105 Use CEIL instead of ROUND_ADVANCE.
107 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
110 * config/sh/sh.c: Include stdlib headers before everything else.
112 2014-05-02 Jakub Jelinek <jakub@redhat.com>
114 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
115 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
116 (gimplify_adjust_omp_clauses): Simd region is never
117 directly nested in combined parallel. Instead, for linear
118 with copyin/copyout, if in combined for simd loop, make decl
119 firstprivate/lastprivate on OMP_FOR.
120 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
121 expand_omp_for_static_chunk): When setting endvar, also set
122 fd->loop.v to the same value.
124 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
126 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
128 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
130 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
133 2014-05-02 Marek Polacek <polacek@redhat.com>
135 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
137 2014-02-26 Kito Cheng <kito@0xlab.org>
139 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
140 to a C expression marco.
141 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
142 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
143 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
144 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
145 HONOR_REG_ALLOC_ORDER.
146 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
148 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
150 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
152 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
154 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
156 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
158 * tree-if-conv.c (is_cond_scalar_reduction): New function.
159 (convert_scalar_cond_reduction): Likewise.
160 (predicate_scalar_phi): Add recognition and transformation
161 of simple conditioanl reduction to be vectorizable.
163 2014-05-01 Marek Polacek <polacek@redhat.com>
166 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
168 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
170 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
171 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
172 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
173 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
174 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
175 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
176 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
177 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
179 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
181 * config/arc/arc.opt (mlra): Move comment above option name
182 to avoid mis-parsing as language options.
184 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
186 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
187 * config/sol2.h: ... here.
188 * config/sol2-10.h: Remove.
190 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
191 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
192 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
193 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
194 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
195 * config/sol2.h: ... here.
196 (SECTION_NAME_FORMAT): Don't redefine.
197 (STARTFILE_ARCH32_SPEC): Rename to ...
198 (STARTFILE_ARCH_SPEC): ... this.
199 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
200 * config/sparc/sol2.h: ... here.
201 (SECTION_NAME_FORMAT): Don't undef.
202 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
203 (SUBTARGET_EXTRA_SPECS): Remove.
204 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
206 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
207 (MD_STARTFILE_PREFIX): Remove.
208 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
209 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
210 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
211 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
212 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
213 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
214 * config/i386/sol2.h: ... here.
215 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
216 * config/i386/sol2-bi.h: Remove.
217 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
218 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
220 * config/i386/t-sol2-64: Rename to ...
221 * config/i386/t-sol2: ... this.
222 * config/sparc/t-sol2-64: Rename to ...
223 * config/sparc/t-sol2: ... this.
225 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
226 sol2_tm_file_head, sol2_tm_file_tail.
227 Include ${cpu_type}/sol2.h before sol2.h.
229 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
230 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
231 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
232 Reflect i386/t-sol2-64 renaming.
233 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
234 Reflect sparc/t-sol2-64 renaming.
236 2014-04-30 Richard Biener <rguenther@suse.de>
238 * passes.c (execute_function_todo): Move TODO_verify_stmts
239 and TODO_verify_ssa under the TODO_verify_il umbrella.
240 * tree-ssa.h (verify_ssa): Adjust prototype.
241 * tree-ssa.c (verify_ssa): Add parameter to tell whether
242 we should verify SSA operands.
243 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
244 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
245 whether we should verify whether not throwing stmts have EH info.
246 * graphite-scop-detection.c (create_sese_edges): Adjust.
247 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
248 * tree-eh.c (lower_try_finally_switch): Do not add the
249 default case label twice.
251 2014-04-30 Marek Polacek <polacek@redhat.com>
253 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
254 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
255 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
256 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
258 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
260 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
261 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
262 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
263 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
264 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
265 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
266 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
267 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
269 2014-04-29 David Malcolm <dmalcolm@redhat.com>
271 * tree-cfg.c (dump_function_to_file): Dump the return type of
272 functions, in a line to itself before the function body, mimicking
273 the layout of a C function.
275 2014-04-29 Jakub Jelinek <jakub@redhat.com>
277 PR tree-optimization/60971
278 * tree-tailcall.c (process_assignment): Reject conversions which
281 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
283 * calls.c (initialize_argument_information): Always treat
284 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
285 (expand_call): Likewise.
286 (emit_library_call_calue_1): Likewise.
287 * expr.c (PUSH_ARGS_REVERSED): Do not define.
288 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
291 2014-04-29 Nick Clifton <nickc@redhat.com>
293 * config/msp430/msp430.md (umulsidi): Fix typo.
294 (mulhisi3): Enable even inside interrupt handlers.
295 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
296 bigger return address pushed in large mode.
298 2014-04-29 Nick Clifton <nickc@redhat.com>
300 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
301 (arc_init_reg_tables): Use a machine_mode enum to iterate over
303 * config/m32r/m32r.c (init_reg_tables): Likewise.
304 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
305 enum to hold the modes.
307 2014-04-29 Richard Biener <rguenther@suse.de>
309 * dominance.c (free_dominance_info): Add overload with
311 (dom_info_state): Likewise.
312 (dom_info_available_p): Likewise.
313 * basic-block.h (free_dominance_info, dom_info_state,
314 dom_info_available_p): Declare overloads.
315 * passes.c (execute_function_todo): Verify that verifiers
316 don't change dominator info state. Drop dominator info
317 for IPA pass invocations.
318 * cgraph.c (release_function_body): Restore asserts that
319 dominator information is released.
321 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
323 * doc/invoke.texi: Fix typo.
324 * tree-vrp.c: Fix typos.
325 * gimple.c (infer_nonnull_range): Reorder operands of an &&
328 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
330 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
332 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
334 * config/aarch64/aarch64-builtins.c
335 (aarch64_types_storestruct_lane_qualifiers): New.
336 (TYPES_STORESTRUCT_LANE): Likewise.
337 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
338 (st3_lane): Likewise.
339 (st4_lane): Likewise.
340 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
341 (vec_store_lanesci_lane<mode>): Likewise.
342 (vec_store_lanesxi_lane<mode>): Likewise.
343 (aarch64_st2_lane<VQ:mode>): Likewise.
344 (aarch64_st3_lane<VQ:mode>): Likewise.
345 (aarch64_st4_lane<VQ:mode>): Likewise.
346 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
347 * config/aarch64/arm_neon.h
348 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
349 use new macro arguments.
350 (__ST3_LANE_FUNC): Likewise.
351 (__ST4_LANE_FUNC): Likewise.
352 * config/aarch64/iterators.md (V_TWO_ELEM): New.
353 (V_THREE_ELEM): Likewise.
354 (V_FOUR_ELEM): Likewise.
356 2014-04-28 David Malcolm <dmalcolm@redhat.com>
358 * doc/gimple.texi: Replace the description of the now-defunct
359 union gimple_statement_d with a diagram showing the
360 gimple_statement_base class hierarchy and its relationships to
361 the GSS_ and GIMPLE_ enums.
363 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
365 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
366 * config/aarch64/aarch64.c
367 (aarch64_cannot_change_mode_class): Weaken conditions.
368 (aarch64_modes_tieable_p): New.
369 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
371 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
373 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
374 (loadsync_<mode>): Change mode.
375 (load_quadpti, store_quadpti): New.
376 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
377 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
379 2014-04-28 Martin Jambor <mjambor@suse.cz>
381 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
382 same alias type as the original statement.
383 (subreplacement_assignment_data): New type.
384 (handle_unscalarized_data_in_subtree): New type of parameter,
385 generate new memory accesses with same alias type as the original
387 (load_assign_lhs_subreplacements): Likewise.
388 (sra_modify_constructor_assign): Generate new memory accesses with
389 same alias type as the original statement.
391 2014-04-28 Richard Biener <rguenther@suse.de>
393 * tree-pass.h (TODO_verify_il): Define.
394 (TODO_verify_all): Complete properly.
395 * passes.c (execute_function_todo): Move existing loop-closed
396 SSA verification under TODO_verify_il.
397 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
398 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
399 Fix tree sharing issue.
401 2014-04-28 Richard Biener <rguenther@suse.de>
404 * builtins.def (DEF_C11_BUILTIN): Add.
405 (BUILT_IN_ALIGNED_ALLOC): Likewise.
406 * coretypes.h (enum function_class): Add function_c11_misc.
407 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
408 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
409 (call_may_clobber_ref_p_1): Likewise.
410 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
411 (mark_all_reaching_defs_necessary_1): Likewise.
412 (propagate_necessity): Likewise.
413 (eliminate_unnecessary_stmts): Likewise.
414 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
416 2014-04-28 Richard Biener <rguenther@suse.de>
418 * tree-vrp.c (vrp_var_may_overflow): Remove.
419 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
420 with overflow immediately bump to one before that value and
421 let iteration figure out overflow status.
423 2014-04-28 Richard Biener <rguenther@suse.de>
425 * configure.ac: Do valgrind header checks unconditionally.
426 Add --enable-valgrind-annotations.
427 * system.h: Guard valgrind header inclusion with
428 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
429 * alloc-pool.c (pool_alloc, pool_free): Use
430 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
431 to guard possibly dead code.
432 * config.in: Regenerated.
433 * configure: Likewise.
435 2014-04-28 Jeff Law <law@redhat.com>
437 PR tree-optimization/60902
438 * tree-ssa-threadedge.c
439 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
440 over real defs when invalidating outputs from statements that do not
441 produce useful outputs for threading.
443 2014-04-28 Richard Biener <rguenther@suse.de>
445 PR tree-optimization/60979
446 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
447 SCOPs that end in a block with a successor with abnormal
450 2014-04-28 Richard Biener <rguenther@suse.de>
452 * tree-pass.h (execute_pass_list): Adjust prototype.
453 * passes.c (pass_manager::execute_early_local_passes):
455 (do_per_function): Change callback signature, push all actual
456 work to the callbals.
457 (do_per_function_toporder): Likewise.
458 (execute_function_dump): Adjust.
459 (execute_function_todo): Likewise.
460 (clear_last_verified): Likewise.
461 (verify_curr_properties): Likewise.
462 (update_properties_after_pass): Likewise.
463 (execute_pass_list_1): Split out from ...
464 (execute_pass_list): ... here. Adjust.
465 (execute_ipa_pass_list): Likewise.
466 * cgraphunit.c (cgraph_add_new_function): Adjust.
467 (analyze_function): Likewise.
468 (expand_function): Likewise.
469 * cgraph.c (release_function_body): Free dominance info
470 here instead of asserting it was magically freed elsewhere.
472 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
474 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
475 * configure: Regenerate.
476 * config/sparc/sparc.opt (muser-mode): New option.
477 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
479 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
480 * doc/invoke.texi (SPARC options): Document -muser-mode.
482 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
484 * cselib.c (find_slot_memmode): Delete.
485 (cselib_hasher): Change compare_type to a struct.
486 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
488 (preserve_constants_and_equivs): Adjust for new compare_type.
489 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
490 (wrap_constant): Delete.
491 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
493 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
495 * doc/install.texi (Building with profile feedback): Remove
498 2014-04-26 Tom de Vries <tom@codesourcery.com>
500 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
503 2014-04-25 Cary Coutant <ccoutant@google.com>
506 * dwarf2out.c (should_move_die_to_comdat): A type definition
507 can contain a subprogram definition, but don't move it to a
509 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
510 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
512 (clone_tree_hash): Rename to...
513 (clone_tree_partial): ...this; change callers. Copy
514 DW_TAG_subprogram DIEs as declarations.
515 (copy_decls_walk): Don't copy children of a declaration into a
518 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
521 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
524 2014-04-25 Jiong Wang <jiong.wang@arm.com>
526 * config/arm/predicates.md (call_insn_operand): Add long_call check.
527 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
529 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
532 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
534 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
537 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
539 PR tree-optimization/60930
540 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
541 creating a multiply candidate by folding two constant
542 multiplicands when the result overflows.
544 2014-04-25 Jakub Jelinek <jakub@redhat.com>
546 PR tree-optimization/60960
547 * tree-vect-generic.c (expand_vector_operation): Only call
548 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
550 2014-04-25 Tom de Vries <tom@codesourcery.com>
552 * expr.c (clobber_reg_mode): New function.
553 * expr.h (clobber_reg): New function.
555 2014-04-25 Tom de Vries <tom@codesourcery.com>
557 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
560 2014-04-25 Radovan Obradovic <robradovic@mips.com>
561 Tom de Vries <tom@codesourcery.com>
563 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
565 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
566 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
567 new argument to find_all_hard_reg_sets call.
569 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
571 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
572 Use HOST_WIDE_INT_C for mask literal.
573 (aarch_rev16_shleft_mask_imm_p): Likewise.
575 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
578 * config/sparc/sparc.md (ashlsi3_extend): Delete.
580 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
582 PR preprocessor/56540
583 * config/i386/i386-c.c (ix86_target_macros): Define
584 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
586 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
588 * configure.ac (tga_func): Remove.
589 (LIB_TLS_SPEC): Remove.
590 * configure: Regenerate.
591 * config.in: Regenerate.
592 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
594 2014-04-25 Richard Biener <rguenther@suse.de>
597 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
598 call stmt use/clobber sets during stmt walk instead of
599 walking the possibly incomplete set of caller edges.
601 2014-04-25 Richard Biener <rguenther@suse.de>
604 * passes.c (apply_ipa_transforms): Inline into only caller ...
605 (execute_one_pass): ... here. Properly bring in function
606 bodies for nodes we want to apply IPA transforms to.
608 2014-04-24 Cong Hou <congh@google.com>
610 PR tree-optimization/60896
611 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
612 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
613 (vect_mark_pattern_stmts): Set the def type of all statements in
614 PATTERN_DEF_SEQ as vect_internal_def.
616 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
618 * doc/extend.texi (PowerPC Built-in Functions): Document new
619 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
621 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
623 * config/rs6000/predicates.md (const_0_to_3_operand): New
624 predicate to match 0..3 integer constants.
626 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
627 to support adding miscellaneous builtin functions.
628 (BU_DFP_MISC_2): Likewise.
629 (BU_P7_MISC_1): Likewise.
630 (BU_P7_MISC_2): Likewise.
631 (BU_P8V_MISC_3): Likewise.
632 (BU_MISC_1): Likewise.
633 (BU_MISC_2): Likewise.
634 (DIVWE): Add extended divide builtin functions.
642 (DXEX): Add decimal floating-point builtin functions.
654 (CDTBCD): Add new BCD builtin functions.
658 (BCDADD_LT): Likewise.
659 (BCDADD_EQ): Likewise.
660 (BCDADD_GT): Likewise.
661 (BCDADD_OV): Likewise.
663 (BCDSUB_LT): Likewise.
664 (BCDSUB_EQ): Likewise.
665 (BCDSUB_GT): Likewise.
666 (BCDSUB_OV): Likewise.
667 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
668 (UNPACK_TD): Likewise.
670 (UNPACK_TF): Likewise.
671 (UNPACK_TF_0): Likewise.
672 (UNPACK_TF_1): Likewise.
673 (PACK_V1TI): Likewise.
674 (UNPACK_V1TI): Likewise.
676 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
677 support for decimal floating point builtin functions.
678 (rs6000_expand_ternop_builtin): Add checks for the new builtin
679 functions that take constant arguments.
680 (rs6000_invalid_builtin): Add decimal floating point builtin
682 (rs6000_init_builtins): Setup long double, _Decimal64, and
683 _Decimal128 types for new builtin functions.
684 (builtin_function_type): Set the unsigned flags appropriately for
685 the new builtin functions.
686 (rs6000_opt_masks): Add support for decimal floating point builtin
689 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
690 floating point builtin functions.
691 (RS6000_BTM_COMMON): Likewise.
692 (RS6000_BTI_long_double): Likewise.
693 (RS6000_BTI_dfloat64): Likewise.
694 (RS6000_BTI_dfloat128): Likewise.
695 (long_double_type_internal_node): Likewise.
696 (dfloat64_type_internal_node): Likewise.
697 (dfloat128_type_internal_node): Likewise.
699 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
700 2.07 bcd arithmetic instructions.
701 (UNSPEC_BCDSUB): Likewise.
702 (UNSPEC_BCD_OVERFLOW): Likewise.
703 (UNSPEC_BCD_ADD_SUB): Likewise.
704 (bcd_add_sub): Likewise.
705 (BCD_TEST): Likewise.
706 (bcd<bcd_add_sub>): Likewise.
707 (bcd<bcd_add_sub>_test): Likewise.
708 (bcd<bcd_add_sub>_test2): Likewise.
709 (bcd<bcd_add_sub>_<code>): Likewise.
710 (peephole2 for combined bcd ops): Likewise.
712 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
713 decimal floating point builtin functions.
714 (UNSPEC_DENBCD): Likewise.
715 (UNSPEC_DXEX): Likewise.
716 (UNSPEC_DIEX): Likewise.
717 (UNSPEC_DSCLI): Likewise.
718 (UNSPEC_DSCRI): Likewise.
719 (D64_D128): Likewise.
720 (dfp_suffix): Likewise.
721 (dfp_ddedpd_<mode>): Likewise.
722 (dfp_denbcd_<mode>): Likewise.
723 (dfp_dxex_<mode>): Likewise.
724 (dfp_diex_<mode>): Likewise.
725 (dfp_dscli_<mode>): Likewise.
726 (dfp_dscri_<mode>): Likewise.
728 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
730 (UNSPEC_CDTBCD): Likewise.
731 (UNSPEC_CBCDTD): Likewise.
732 (UNSPEC_DIVE): Add support for new extended divide builtin
734 (UNSPEC_DIVEO): Likewise.
735 (UNSPEC_DIVEU): Likewise.
736 (UNSPEC_DIVEUO): Likewise.
737 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
738 pack/unpack 128-bit types.
739 (UNSPEC_PACK_128BIT): Likewise.
740 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
741 (udiv<mode>3): Use idiv_ldiv mode attribute.
742 (div<mode>3): Likewise.
743 (addg6s): Add new BCD builtin functions.
746 (UNSPEC_DIV_EXTEND): Add support for new extended divide
748 (div_extend): Likewise.
749 (div<div_extend>_<mode>"): Likewise.
750 (FP128_64): Add support for new builtin functions to pack/unpack
752 (unpack<mode>): Likewise.
753 (unpacktf_0): Likewise.
754 (unpacktf_1): Likewise.
755 (unpack<mode>_dm): Likewise.
756 (unpack<mode>_nodm): Likewise.
757 (pack<mode>): Likewise.
758 (unpackv1ti): Likewise.
759 (packv1ti): Likewise.
761 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
763 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
766 2014-04-24 Jakub Jelinek <jakub@redhat.com>
768 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
769 * gimplify.c (omp_is_private): Change last argument's type to int.
770 Only diagnose lastprivate if the simd argument is 1, only diagnose
771 linear if the simd argument is 2.
772 (gimplify_omp_for): Adjust omp_is_private callers. When adding
773 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
774 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
775 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
776 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
777 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
778 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
779 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
780 * tree-nested.c (convert_nonlocal_omp_clauses,
781 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
783 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
786 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
789 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
791 * flag-types.h (enum ivar_visibility): Add.
793 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
795 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
798 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
800 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
802 2014-04-24 Radovan Obradovic <robradovic@mips.com>
803 Tom de Vries <tom@codesourcery.com>
805 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
806 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
808 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
809 * emit-rtl.c (try_split): Same.
811 2014-04-24 Radovan Obradovic <robradovic@mips.com>
812 Tom de Vries <tom@codesourcery.com>
814 * common.opt (fuse-caller-save): New option.
816 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
818 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
821 2014-04-24 Richard Biener <rguenther@suse.de>
823 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
824 during TER and instead use the sepops interface for expanding
825 non-GIMPLE_SINGLE_RHS.
827 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
829 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
830 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
832 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
834 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
835 assembler 64-bit option.
836 * configure: Regenerate.
838 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
840 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
841 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
842 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
843 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
844 (TARGET_CRYPTO): Take TARGET_SIMD into account.
846 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
848 * config/aarch64/aarch64-builtins.c
849 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
850 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
851 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
852 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
854 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
855 (Vrevsuff): New mode attribute.
857 2014-04-24 Terry Guo <terry.guo@arm.com>
859 * config/arm/arm.h (machine_function): Define variable
860 after_arm_reorg here.
861 * config/arm/arm.c (after_arm_reorg): Remove the definition.
862 (arm_split_constant): Update the way to access variable
865 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
867 2014-04-23 Tom de Vries <tom@codesourcery.com>
869 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
871 2014-04-23 David Malcolm <dmalcolm@redhat.com>
873 * is-a.h: Update comments to reflect the following changes to the
874 "pointerness" of the API, making the template parameter match the
875 return type, allowing use of is-a.h with typedefs of pointers.
876 (is_a_helper::cast): Return a T rather then a pointer to a T, so
877 that the return type matches the parameter to the is_a_helper.
879 (dyn_cast): Likewise.
881 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
882 pointer from the is-a.h API.
884 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
885 (is_a_helper <cgraph_node *>::test): ...this, matching change to
887 (is_a_helper <varpool_node>::test): Likewise, convert to...
888 (is_a_helper <varpool_node *>::test): ...this.
890 (varpool_first_variable): Update for removal of implicit pointer
892 (varpool_next_variable): Likewise.
893 (varpool_first_static_initializer): Likewise.
894 (varpool_next_static_initializer): Likewise.
895 (varpool_first_defined_variable): Likewise.
896 (varpool_next_defined_variable): Likewise.
897 (cgraph_first_defined_function): Likewise.
898 (cgraph_next_defined_function): Likewise.
899 (cgraph_first_function): Likewise.
900 (cgraph_next_function): Likewise.
901 (cgraph_first_function_with_gimple_body): Likewise.
902 (cgraph_next_function_with_gimple_body): Likewise.
903 (cgraph_alias_target): Likewise.
904 (varpool_alias_target): Likewise.
905 (cgraph_function_or_thunk_node): Likewise.
906 (varpool_variable_node): Likewise.
907 (symtab_real_symbol_p): Likewise.
908 * cgraphunit.c (referred_to_p): Likewise.
909 (analyze_functions): Likewise.
910 (handle_alias_pairs): Likewise.
911 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
912 * gimple-ssa.h (gimple_vuse_op): Likewise.
913 (gimple_vdef_op): Likewise.
914 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
915 * gimple.c (gimple_build_asm_1): Likewise.
916 (gimple_build_try): Likewise.
917 (gimple_build_resx): Likewise.
918 (gimple_build_eh_dispatch): Likewise.
919 (gimple_build_omp_for): Likewise.
920 (gimple_omp_for_set_clauses): Likewise.
922 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
923 (is_a_helper <gimple_statement_asm *>::test): ...this.
924 (is_a_helper <gimple_statement_bind>::test): Convert to...
925 (is_a_helper <gimple_statement_bind *>::test): ...this.
926 (is_a_helper <gimple_statement_call>::test): Convert to...
927 (is_a_helper <gimple_statement_call *>::test): ...this.
928 (is_a_helper <gimple_statement_catch>::test): Convert to...
929 (is_a_helper <gimple_statement_catch *>::test): ...this.
930 (is_a_helper <gimple_statement_resx>::test): Convert to...
931 (is_a_helper <gimple_statement_resx *>::test): ...this.
932 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
933 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
934 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
935 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
936 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
937 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
938 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
939 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
940 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
941 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
942 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
943 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
944 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
945 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
946 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
947 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
948 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
949 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
950 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
951 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
952 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
953 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
954 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
955 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
956 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
957 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
958 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
959 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
960 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
961 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
962 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
963 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
964 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
965 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
966 (is_a_helper <gimple_statement_phi>::test): Convert to...
967 (is_a_helper <gimple_statement_phi *>::test): ...this.
968 (is_a_helper <gimple_statement_transaction>::test): Convert to...
969 (is_a_helper <gimple_statement_transaction *>::test): ...this.
970 (is_a_helper <gimple_statement_try>::test): Convert to...
971 (is_a_helper <gimple_statement_try *>::test): ...this.
972 (is_a_helper <gimple_statement_wce>::test): Convert to...
973 (is_a_helper <gimple_statement_wce *>::test): ...this.
974 (is_a_helper <const gimple_statement_asm>::test): Convert to...
975 (is_a_helper <const gimple_statement_asm *>::test): ...this.
976 (is_a_helper <const gimple_statement_bind>::test): Convert to...
977 (is_a_helper <const gimple_statement_bind *>::test): ...this.
978 (is_a_helper <const gimple_statement_call>::test): Convert to...
979 (is_a_helper <const gimple_statement_call *>::test): ...this.
980 (is_a_helper <const gimple_statement_catch>::test): Convert to...
981 (is_a_helper <const gimple_statement_catch *>::test): ...this.
982 (is_a_helper <const gimple_statement_resx>::test): Convert to...
983 (is_a_helper <const gimple_statement_resx *>::test): ...this.
984 (is_a_helper <const gimple_statement_eh_dispatch>::test):
986 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
988 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
990 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
991 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
993 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
995 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
997 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
999 (is_a_helper <const gimple_statement_omp_return>::test): Convert
1001 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
1002 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
1004 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
1005 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
1007 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
1008 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
1009 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
1010 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
1012 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
1013 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
1015 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
1016 (is_a_helper <const gimple_statement_omp_target>::test): Convert
1018 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
1019 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
1021 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
1022 (is_a_helper <const gimple_statement_omp_single>::test): Convert
1024 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
1025 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
1027 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
1028 (is_a_helper <const gimple_statement_omp_task>::test): Convert
1030 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
1031 (is_a_helper <const gimple_statement_phi>::test): Convert to...
1032 (is_a_helper <const gimple_statement_phi *>::test): ...this.
1033 (is_a_helper <const gimple_statement_transaction>::test): Convert
1035 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
1036 (is_a_helper <const gimple_statement_with_ops>::test): Convert
1038 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
1039 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
1040 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
1041 (is_a_helper <const gimple_statement_with_memory_ops>::test):
1043 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
1045 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
1047 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
1049 (gimple_use_ops): Update for removal of implicit pointer from the
1051 (gimple_set_use_ops): Likewise.
1052 (gimple_vuse): Likewise.
1053 (gimple_vdef): Likewise.
1054 (gimple_vuse_ptr): Likewise.
1055 (gimple_vdef_ptr): Likewise.
1056 (gimple_set_vuse): Likewise.
1057 (gimple_set_vdef): Likewise.
1058 (gimple_omp_return_set_lhs): Likewise.
1059 (gimple_omp_return_lhs): Likewise.
1060 (gimple_omp_return_lhs_ptr): Likewise.
1061 (gimple_call_fntype): Likewise.
1062 (gimple_call_set_fntype): Likewise.
1063 (gimple_call_set_internal_fn): Likewise.
1064 (gimple_call_use_set): Likewise.
1065 (gimple_call_clobber_set): Likewise.
1066 (gimple_bind_vars): Likewise.
1067 (gimple_bind_set_vars): Likewise.
1068 (gimple_bind_body_ptr): Likewise.
1069 (gimple_bind_set_body): Likewise.
1070 (gimple_bind_add_stmt): Likewise.
1071 (gimple_bind_block): Likewise.
1072 (gimple_bind_set_block): Likewise.
1073 (gimple_asm_ninputs): Likewise.
1074 (gimple_asm_noutputs): Likewise.
1075 (gimple_asm_nclobbers): Likewise.
1076 (gimple_asm_nlabels): Likewise.
1077 (gimple_asm_input_op): Likewise.
1078 (gimple_asm_input_op_ptr): Likewise.
1079 (gimple_asm_output_op): Likewise.
1080 (gimple_asm_output_op_ptr): Likewise.
1081 (gimple_asm_set_output_op): Likewise.
1082 (gimple_asm_clobber_op): Likewise.
1083 (gimple_asm_set_clobber_op): Likewise.
1084 (gimple_asm_label_op): Likewise.
1085 (gimple_asm_set_label_op): Likewise.
1086 (gimple_asm_string): Likewise.
1087 (gimple_catch_types): Likewise.
1088 (gimple_catch_types_ptr): Likewise.
1089 (gimple_catch_handler_ptr): Likewise.
1090 (gimple_catch_set_types): Likewise.
1091 (gimple_catch_set_handler): Likewise.
1092 (gimple_eh_filter_types): Likewise.
1093 (gimple_eh_filter_types_ptr): Likewise.
1094 (gimple_eh_filter_failure_ptr): Likewise.
1095 (gimple_eh_filter_set_types): Likewise.
1096 (gimple_eh_filter_set_failure): Likewise.
1097 (gimple_eh_must_not_throw_fndecl): Likewise.
1098 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1099 (gimple_eh_else_n_body_ptr): Likewise.
1100 (gimple_eh_else_e_body_ptr): Likewise.
1101 (gimple_eh_else_set_n_body): Likewise.
1102 (gimple_eh_else_set_e_body): Likewise.
1103 (gimple_try_eval_ptr): Likewise.
1104 (gimple_try_cleanup_ptr): Likewise.
1105 (gimple_try_set_eval): Likewise.
1106 (gimple_try_set_cleanup): Likewise.
1107 (gimple_wce_cleanup_ptr): Likewise.
1108 (gimple_wce_set_cleanup): Likewise.
1109 (gimple_phi_capacity): Likewise.
1110 (gimple_phi_num_args): Likewise.
1111 (gimple_phi_result): Likewise.
1112 (gimple_phi_result_ptr): Likewise.
1113 (gimple_phi_set_result): Likewise.
1114 (gimple_phi_arg): Likewise.
1115 (gimple_phi_set_arg): Likewise.
1116 (gimple_resx_region): Likewise.
1117 (gimple_resx_set_region): Likewise.
1118 (gimple_eh_dispatch_region): Likewise.
1119 (gimple_eh_dispatch_set_region): Likewise.
1120 (gimple_omp_critical_name): Likewise.
1121 (gimple_omp_critical_name_ptr): Likewise.
1122 (gimple_omp_critical_set_name): Likewise.
1123 (gimple_omp_for_clauses): Likewise.
1124 (gimple_omp_for_clauses_ptr): Likewise.
1125 (gimple_omp_for_set_clauses): Likewise.
1126 (gimple_omp_for_collapse): Likewise.
1127 (gimple_omp_for_index): Likewise.
1128 (gimple_omp_for_index_ptr): Likewise.
1129 (gimple_omp_for_set_index): Likewise.
1130 (gimple_omp_for_initial): Likewise.
1131 (gimple_omp_for_initial_ptr): Likewise.
1132 (gimple_omp_for_set_initial): Likewise.
1133 (gimple_omp_for_final): Likewise.
1134 (gimple_omp_for_final_ptr): Likewise.
1135 (gimple_omp_for_set_final): Likewise.
1136 (gimple_omp_for_incr): Likewise.
1137 (gimple_omp_for_incr_ptr): Likewise.
1138 (gimple_omp_for_set_incr): Likewise.
1139 (gimple_omp_for_pre_body_ptr): Likewise.
1140 (gimple_omp_for_set_pre_body): Likewise.
1141 (gimple_omp_parallel_clauses): Likewise.
1142 (gimple_omp_parallel_clauses_ptr): Likewise.
1143 (gimple_omp_parallel_set_clauses): Likewise.
1144 (gimple_omp_parallel_child_fn): Likewise.
1145 (gimple_omp_parallel_child_fn_ptr): Likewise.
1146 (gimple_omp_parallel_set_child_fn): Likewise.
1147 (gimple_omp_parallel_data_arg): Likewise.
1148 (gimple_omp_parallel_data_arg_ptr): Likewise.
1149 (gimple_omp_parallel_set_data_arg): Likewise.
1150 (gimple_omp_task_clauses): Likewise.
1151 (gimple_omp_task_clauses_ptr): Likewise.
1152 (gimple_omp_task_set_clauses): Likewise.
1153 (gimple_omp_task_child_fn): Likewise.
1154 (gimple_omp_task_child_fn_ptr): Likewise.
1155 (gimple_omp_task_set_child_fn): Likewise.
1156 (gimple_omp_task_data_arg): Likewise.
1157 (gimple_omp_task_data_arg_ptr): Likewise.
1158 (gimple_omp_task_set_data_arg): Likewise.
1159 (gimple_omp_taskreg_clauses): Likewise.
1160 (gimple_omp_taskreg_clauses_ptr): Likewise.
1161 (gimple_omp_taskreg_set_clauses): Likewise.
1162 (gimple_omp_taskreg_child_fn): Likewise.
1163 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1164 (gimple_omp_taskreg_set_child_fn): Likewise.
1165 (gimple_omp_taskreg_data_arg): Likewise.
1166 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1167 (gimple_omp_taskreg_set_data_arg): Likewise.
1168 (gimple_omp_task_copy_fn): Likewise.
1169 (gimple_omp_task_copy_fn_ptr): Likewise.
1170 (gimple_omp_task_set_copy_fn): Likewise.
1171 (gimple_omp_task_arg_size): Likewise.
1172 (gimple_omp_task_arg_size_ptr): Likewise.
1173 (gimple_omp_task_set_arg_size): Likewise.
1174 (gimple_omp_task_arg_align): Likewise.
1175 (gimple_omp_task_arg_align_ptr): Likewise.
1176 (gimple_omp_task_set_arg_align): Likewise.
1177 (gimple_omp_single_clauses): Likewise.
1178 (gimple_omp_single_clauses_ptr): Likewise.
1179 (gimple_omp_single_set_clauses): Likewise.
1180 (gimple_omp_target_clauses): Likewise.
1181 (gimple_omp_target_clauses_ptr): Likewise.
1182 (gimple_omp_target_set_clauses): Likewise.
1183 (gimple_omp_target_child_fn): Likewise.
1184 (gimple_omp_target_child_fn_ptr): Likewise.
1185 (gimple_omp_target_set_child_fn): Likewise.
1186 (gimple_omp_target_data_arg): Likewise.
1187 (gimple_omp_target_data_arg_ptr): Likewise.
1188 (gimple_omp_target_set_data_arg): Likewise.
1189 (gimple_omp_teams_clauses): Likewise.
1190 (gimple_omp_teams_clauses_ptr): Likewise.
1191 (gimple_omp_teams_set_clauses): Likewise.
1192 (gimple_omp_sections_clauses): Likewise.
1193 (gimple_omp_sections_clauses_ptr): Likewise.
1194 (gimple_omp_sections_set_clauses): Likewise.
1195 (gimple_omp_sections_control): Likewise.
1196 (gimple_omp_sections_control_ptr): Likewise.
1197 (gimple_omp_sections_set_control): Likewise.
1198 (gimple_omp_for_set_cond): Likewise.
1199 (gimple_omp_for_cond): Likewise.
1200 (gimple_omp_atomic_store_set_val): Likewise.
1201 (gimple_omp_atomic_store_val): Likewise.
1202 (gimple_omp_atomic_store_val_ptr): Likewise.
1203 (gimple_omp_atomic_load_set_lhs): Likewise.
1204 (gimple_omp_atomic_load_lhs): Likewise.
1205 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1206 (gimple_omp_atomic_load_set_rhs): Likewise.
1207 (gimple_omp_atomic_load_rhs): Likewise.
1208 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1209 (gimple_omp_continue_control_def): Likewise.
1210 (gimple_omp_continue_control_def_ptr): Likewise.
1211 (gimple_omp_continue_set_control_def): Likewise.
1212 (gimple_omp_continue_control_use): Likewise.
1213 (gimple_omp_continue_control_use_ptr): Likewise.
1214 (gimple_omp_continue_set_control_use): Likewise.
1215 (gimple_transaction_body_ptr): Likewise.
1216 (gimple_transaction_label): Likewise.
1217 (gimple_transaction_label_ptr): Likewise.
1218 (gimple_transaction_set_body): Likewise.
1219 (gimple_transaction_set_label): Likewise.
1221 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1222 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1223 * ipa-ref.c (ipa_record_reference): Likewise.
1224 * ipa-reference.c (analyze_function): Likewise.
1225 (ipa_reference_write_optimization_summary): Likewise.
1226 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1227 (address_taken_from_non_vtable_p): Likewise.
1228 (comdat_can_be_unshared_p_1): Likewise.
1229 * lto-cgraph.c (lto_output_ref): Likewise.
1230 (add_references): Likewise.
1231 (compute_ltrans_boundary): Likewise.
1232 (output_symtab): Likewise.
1233 (input_ref): Likewise.
1234 (input_cgraph_1): Likewise.
1235 (output_cgraph_opt_summary): Likewise.
1236 * lto-streamer-out.c (lto_output): Likewise.
1237 (output_symbol_p): Likewise.
1238 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1239 (lsei_start_function_in_partition): Likewise.
1240 (lsei_next_variable_in_partition): Likewise.
1241 (lsei_start_variable_in_partition): Likewise.
1242 * symtab.c (insert_to_assembler_name_hash): Likewise.
1243 (unlink_from_assembler_name_hash): Likewise.
1244 (symtab_unregister_node): Likewise.
1245 (symtab_remove_node): Likewise.
1246 (dump_symtab_node): Likewise.
1247 (verify_symtab_base): Likewise.
1248 (verify_symtab_node): Likewise.
1249 (symtab_make_decl_local): Likewise.
1250 (symtab_alias_ultimate_target): Likewise.
1251 (symtab_resolve_alias): Likewise.
1252 (symtab_get_symbol_partitioning_class): Likewise.
1253 * tree-phinodes.c (allocate_phi_node): Likewise.
1254 (reserve_phi_args_for_new_edge): Likewise.
1255 (remove_phi_args): Likewise.
1256 * varpool.c (varpool_node_for_asm): Likewise.
1257 (varpool_remove_unreferenced_decls): Likewise.
1259 2014-04-23 Jeff Law <law@redhat.com>
1261 PR tree-optimization/60902
1262 * tree-ssa-threadedge.c
1263 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1264 invalidate outputs from statements that do not produce useful
1265 outputs for threading.
1267 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1269 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1270 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1271 machine descriptions for Stack Smashing Protector.
1273 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1275 * aarch64.md (<optab>_rol<mode>3): New pattern.
1276 (<optab>_rolsi3_uxtw): Likewise.
1277 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1279 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1281 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1282 (arm_cortex_a12_tune): Likewise.
1284 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1286 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1288 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1290 * config/arm/arm.md (arm_rev16si2): New pattern.
1291 (arm_rev16si2_alt): Likewise.
1292 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1294 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1296 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1297 (rev16<mode>2_alt): Likewise.
1298 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1299 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1300 (aarch_rev16_shleft_mask_imm_p): Likewise.
1301 (aarch_rev16_p_1): Likewise.
1302 (aarch_rev16_p): Likewise.
1303 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1304 (aarch_rev16_shright_mask_imm_p): Likewise.
1305 (aarch_rev16_shleft_mask_imm_p): Likewise.
1307 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1309 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1310 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1312 (cortex_a53_extra_costs): Likewise.
1313 (cortex_a57_extra_costs): Likewise.
1314 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1315 (cortexa7_extra_costs): Likewise.
1316 (cortexa8_extra_costs): Likewise.
1317 (cortexa12_extra_costs): Likewise.
1318 (cortexa15_extra_costs): Likewise.
1319 (v7m_extra_costs): Likewise.
1320 (arm_new_rtx_costs): Handle BSWAP.
1322 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1324 * config/arm/arm.c (cortexa8_extra_costs): New table.
1325 (arm_cortex_a8_tune): New tuning struct.
1326 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1328 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1330 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1332 2014-04-23 Richard Biener <rguenther@suse.de>
1334 * Makefile.in (OBJS): Remove loop-unswitch.o.
1335 * loop-unswitch.c: Delete.
1336 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1337 * passes.def (pass_rtl_unswitch): Likewise.
1338 * loop-init.c (gate_rtl_unswitch): Likewise.
1339 (rtl_unswitch): Likewise.
1340 (pass_data_rtl_unswitch): Likewise.
1341 (pass_rtl_unswitch): Likewise.
1342 (make_pass_rtl_unswitch): Likewise.
1343 * rtl.h (reversed_condition): Likewise.
1344 (compare_and_jump_seq): Likewise.
1345 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1347 * loop-unroll.c (compare_and_jump_seq): Likewise.
1349 2014-04-23 Richard Biener <rguenther@suse.de>
1351 PR tree-optimization/60903
1352 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1353 commented code block.
1354 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1355 loop flags to newly created BBs and edges.
1357 2014-04-23 Nick Clifton <nickc@redhat.com>
1359 * config/msp430/msp430.c (msp430_handle_option): Move function
1361 (msp430_option_override): Simplify mcu and mcpu option handling.
1362 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1363 support for -mhwmult command line option.
1364 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1365 -mhwmult command line option.
1366 (msp430_hwmult_enabled): Delete.
1367 (msp43o_output_labelref): Add support for -mhwmult command line
1369 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1370 (umulsidi3): Likewise.
1371 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1372 (mcpu, mlarge, msmall): Likewise.
1373 (mhwmult): New option.
1374 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1376 (msp430_is_f5_mcu): Remove prototype.
1377 (msp430_use_f5_series_hwmult): Add prototype.
1378 * config/msp430/msp430-opts.h: New file.
1379 * common/config/msp430: New directory.
1380 * common/config/msp430/msp430-common.c: New file.
1381 * config.gcc (msp430): Remove target_has_targetm_common.
1382 * doc/invoke.texi: Document -mhwmult command line option.
1384 2014-04-23 Nick Clifton <nickc@redhat.com>
1386 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1387 default-manifest.o if it can be found in the search path.
1388 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1390 2014-04-23 Terry Guo <terry.guo@arm.com>
1392 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1394 2014-04-23 Richard Biener <rguenther@suse.de>
1397 * tree-inline.c (declare_return_variable): Use mark_addressable.
1399 2014-04-23 Richard Biener <rguenther@suse.de>
1402 * loop-init.c (loop_optimizer_init): Make sure to apply
1403 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1405 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1408 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1410 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1411 if flag_sanitize_undefined_trap_on_error.
1412 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1413 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1414 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1415 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1416 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1417 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1418 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1419 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1420 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1421 * ubsan.c (ubsan_instrument_unreachable): Return
1422 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1423 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1424 if flag_sanitize_undefined_trap_on_error and
1425 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1426 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1427 instrument_bool_enum_load): Emit __builtin_trap () if
1428 flag_sanitize_undefined_trap_on_error and
1429 __builtin_handle_*_abort () if !flag_sanitize_recover.
1430 * doc/invoke.texi (-fsanitize-recover,
1431 -fsanitize-undefined-trap-on-error): Document.
1433 2014-04-22 Christian Bruel <christian.bruel@st.com>
1435 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1436 Force immediates to SImode.
1438 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1440 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1442 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1443 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1444 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1445 (nios2_fpu_insn): Add entry for round.
1446 (N2FPU_NO_ERRNO_P): Define.
1447 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1449 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1451 2014-04-22 Richard Henderson <rth@redhat.com>
1453 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1454 (add<GPI>3_compare0): Remove leading * from name.
1455 (add<GPI>3_carryin): Likewise.
1456 (sub<GPI>3_compare0): Likewise.
1457 (sub<GPI>3_carryin): Likewise.
1458 (<su_optab>mulditi3): New expander.
1459 (multi3): New expander.
1460 (madd<GPI>): Remove leading * from name.
1462 2014-04-22 Martin Jambor <mjambor@suse.cz>
1464 * cgraphclones.c (cgraph_function_versioning): Copy
1465 ipa_transforms_to_apply instead of asserting it is empty.
1467 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1470 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1471 on count_exp to get mode.
1473 2014-04-22 Andrew Pinski <apinski@cavium.com>
1475 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1476 Handle TLS for ILP32.
1477 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1478 (tlsie_small_<mode>): this and handle PTR.
1479 (tlsie_small_sidi): New pattern.
1480 (tlsle_small): Change to an expand to handle ILP32.
1481 (tlsle_small_<mode>): New pattern.
1482 (tlsdesc_small): Rename to ...
1483 (tlsdesc_small_<mode>): this and handle PTR.
1485 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1487 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1489 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1491 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1492 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1493 (aarch64_types_signed_poly_qualifiers): Likewise.
1494 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1495 (aarch64_types_poly_signed_qualifiers): Likewise.
1496 (TYPES_REINTERP_SS): Type macro added.
1497 (TYPES_REINTERP_SU): Likewise.
1498 (TYPES_REINTERP_SP): Likewise.
1499 (TYPES_REINTERP_US): Likewise.
1500 (TYPES_REINTERP_PS): Likewise.
1501 (aarch64_fold_builtin): New expression folding added.
1502 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1503 Declarations removed.
1504 (REINTERP_SS): Declarations added.
1505 (REINTERP_US): Likewise.
1506 (REINTERP_PS): Likewise.
1507 (REINTERP_SU): Likewise.
1508 (REINTERP_SP): Likewise.
1509 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1510 (vreinterpretq_p8_f64): Likewise.
1511 (vreinterpret_p16_f64): Likewise.
1512 (vreinterpretq_p16_f64): Likewise.
1513 (vreinterpret_f32_f64): Likewise.
1514 (vreinterpretq_f32_f64): Likewise.
1515 (vreinterpret_f64_f32): Likewise.
1516 (vreinterpret_f64_p8): Likewise.
1517 (vreinterpret_f64_p16): Likewise.
1518 (vreinterpret_f64_s8): Likewise.
1519 (vreinterpret_f64_s16): Likewise.
1520 (vreinterpret_f64_s32): Likewise.
1521 (vreinterpret_f64_s64): Likewise.
1522 (vreinterpret_f64_u8): Likewise.
1523 (vreinterpret_f64_u16): Likewise.
1524 (vreinterpret_f64_u32): Likewise.
1525 (vreinterpret_f64_u64): Likewise.
1526 (vreinterpretq_f64_f32): Likewise.
1527 (vreinterpretq_f64_p8): Likewise.
1528 (vreinterpretq_f64_p16): Likewise.
1529 (vreinterpretq_f64_s8): Likewise.
1530 (vreinterpretq_f64_s16): Likewise.
1531 (vreinterpretq_f64_s32): Likewise.
1532 (vreinterpretq_f64_s64): Likewise.
1533 (vreinterpretq_f64_u8): Likewise.
1534 (vreinterpretq_f64_u16): Likewise.
1535 (vreinterpretq_f64_u32): Likewise.
1536 (vreinterpretq_f64_u64): Likewise.
1537 (vreinterpret_s64_f64): Likewise.
1538 (vreinterpretq_s64_f64): Likewise.
1539 (vreinterpret_u64_f64): Likewise.
1540 (vreinterpretq_u64_f64): Likewise.
1541 (vreinterpret_s8_f64): Likewise.
1542 (vreinterpretq_s8_f64): Likewise.
1543 (vreinterpret_s16_f64): Likewise.
1544 (vreinterpretq_s16_f64): Likewise.
1545 (vreinterpret_s32_f64): Likewise.
1546 (vreinterpretq_s32_f64): Likewise.
1547 (vreinterpret_u8_f64): Likewise.
1548 (vreinterpretq_u8_f64): Likewise.
1549 (vreinterpret_u16_f64): Likewise.
1550 (vreinterpretq_u16_f64): Likewise.
1551 (vreinterpret_u32_f64): Likewise.
1552 (vreinterpretq_u32_f64): Likewise.
1554 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1556 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1557 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1558 (vreinterpret_p8_s8): Likewise.
1559 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1560 (vreinterpret_p8_s16): Likewise.
1561 (vreinterpret_p8_s32): Likewise.
1562 (vreinterpret_p8_s64): Likewise.
1563 (vreinterpret_p8_f32): Likewise.
1564 (vreinterpret_p8_u8): Likewise.
1565 (vreinterpret_p8_u16): Likewise.
1566 (vreinterpret_p8_u32): Likewise.
1567 (vreinterpret_p8_u64): Likewise.
1568 (vreinterpret_p8_p16): Likewise.
1569 (vreinterpretq_p8_s8): Likewise.
1570 (vreinterpretq_p8_s16): Likewise.
1571 (vreinterpretq_p8_s32): Likewise.
1572 (vreinterpretq_p8_s64): Likewise.
1573 (vreinterpretq_p8_f32): Likewise.
1574 (vreinterpretq_p8_u8): Likewise.
1575 (vreinterpretq_p8_u16): Likewise.
1576 (vreinterpretq_p8_u32): Likewise.
1577 (vreinterpretq_p8_u64): Likewise.
1578 (vreinterpretq_p8_p16): Likewise.
1579 (vreinterpret_p16_s8): Likewise.
1580 (vreinterpret_p16_s16): Likewise.
1581 (vreinterpret_p16_s32): Likewise.
1582 (vreinterpret_p16_s64): Likewise.
1583 (vreinterpret_p16_f32): Likewise.
1584 (vreinterpret_p16_u8): Likewise.
1585 (vreinterpret_p16_u16): Likewise.
1586 (vreinterpret_p16_u32): Likewise.
1587 (vreinterpret_p16_u64): Likewise.
1588 (vreinterpret_p16_p8): Likewise.
1589 (vreinterpretq_p16_s8): Likewise.
1590 (vreinterpretq_p16_s16): Likewise.
1591 (vreinterpretq_p16_s32): Likewise.
1592 (vreinterpretq_p16_s64): Likewise.
1593 (vreinterpretq_p16_f32): Likewise.
1594 (vreinterpretq_p16_u8): Likewise.
1595 (vreinterpretq_p16_u16): Likewise.
1596 (vreinterpretq_p16_u32): Likewise.
1597 (vreinterpretq_p16_u64): Likewise.
1598 (vreinterpretq_p16_p8): Likewise.
1599 (vreinterpret_f32_s8): Likewise.
1600 (vreinterpret_f32_s16): Likewise.
1601 (vreinterpret_f32_s32): Likewise.
1602 (vreinterpret_f32_s64): Likewise.
1603 (vreinterpret_f32_u8): Likewise.
1604 (vreinterpret_f32_u16): Likewise.
1605 (vreinterpret_f32_u32): Likewise.
1606 (vreinterpret_f32_u64): Likewise.
1607 (vreinterpret_f32_p8): Likewise.
1608 (vreinterpret_f32_p16): Likewise.
1609 (vreinterpretq_f32_s8): Likewise.
1610 (vreinterpretq_f32_s16): Likewise.
1611 (vreinterpretq_f32_s32): Likewise.
1612 (vreinterpretq_f32_s64): Likewise.
1613 (vreinterpretq_f32_u8): Likewise.
1614 (vreinterpretq_f32_u16): Likewise.
1615 (vreinterpretq_f32_u32): Likewise.
1616 (vreinterpretq_f32_u64): Likewise.
1617 (vreinterpretq_f32_p8): Likewise.
1618 (vreinterpretq_f32_p16): Likewise.
1619 (vreinterpret_s64_s8): Likewise.
1620 (vreinterpret_s64_s16): Likewise.
1621 (vreinterpret_s64_s32): Likewise.
1622 (vreinterpret_s64_f32): Likewise.
1623 (vreinterpret_s64_u8): Likewise.
1624 (vreinterpret_s64_u16): Likewise.
1625 (vreinterpret_s64_u32): Likewise.
1626 (vreinterpret_s64_u64): Likewise.
1627 (vreinterpret_s64_p8): Likewise.
1628 (vreinterpret_s64_p16): Likewise.
1629 (vreinterpretq_s64_s8): Likewise.
1630 (vreinterpretq_s64_s16): Likewise.
1631 (vreinterpretq_s64_s32): Likewise.
1632 (vreinterpretq_s64_f32): Likewise.
1633 (vreinterpretq_s64_u8): Likewise.
1634 (vreinterpretq_s64_u16): Likewise.
1635 (vreinterpretq_s64_u32): Likewise.
1636 (vreinterpretq_s64_u64): Likewise.
1637 (vreinterpretq_s64_p8): Likewise.
1638 (vreinterpretq_s64_p16): Likewise.
1639 (vreinterpret_u64_s8): Likewise.
1640 (vreinterpret_u64_s16): Likewise.
1641 (vreinterpret_u64_s32): Likewise.
1642 (vreinterpret_u64_s64): Likewise.
1643 (vreinterpret_u64_f32): Likewise.
1644 (vreinterpret_u64_u8): Likewise.
1645 (vreinterpret_u64_u16): Likewise.
1646 (vreinterpret_u64_u32): Likewise.
1647 (vreinterpret_u64_p8): Likewise.
1648 (vreinterpret_u64_p16): Likewise.
1649 (vreinterpretq_u64_s8): Likewise.
1650 (vreinterpretq_u64_s16): Likewise.
1651 (vreinterpretq_u64_s32): Likewise.
1652 (vreinterpretq_u64_s64): Likewise.
1653 (vreinterpretq_u64_f32): Likewise.
1654 (vreinterpretq_u64_u8): Likewise.
1655 (vreinterpretq_u64_u16): Likewise.
1656 (vreinterpretq_u64_u32): Likewise.
1657 (vreinterpretq_u64_p8): Likewise.
1658 (vreinterpretq_u64_p16): Likewise.
1659 (vreinterpret_s8_s16): Likewise.
1660 (vreinterpret_s8_s32): Likewise.
1661 (vreinterpret_s8_s64): Likewise.
1662 (vreinterpret_s8_f32): Likewise.
1663 (vreinterpret_s8_u8): Likewise.
1664 (vreinterpret_s8_u16): Likewise.
1665 (vreinterpret_s8_u32): Likewise.
1666 (vreinterpret_s8_u64): Likewise.
1667 (vreinterpret_s8_p8): Likewise.
1668 (vreinterpret_s8_p16): Likewise.
1669 (vreinterpretq_s8_s16): Likewise.
1670 (vreinterpretq_s8_s32): Likewise.
1671 (vreinterpretq_s8_s64): Likewise.
1672 (vreinterpretq_s8_f32): Likewise.
1673 (vreinterpretq_s8_u8): Likewise.
1674 (vreinterpretq_s8_u16): Likewise.
1675 (vreinterpretq_s8_u32): Likewise.
1676 (vreinterpretq_s8_u64): Likewise.
1677 (vreinterpretq_s8_p8): Likewise.
1678 (vreinterpretq_s8_p16): Likewise.
1679 (vreinterpret_s16_s8): Likewise.
1680 (vreinterpret_s16_s32): Likewise.
1681 (vreinterpret_s16_s64): Likewise.
1682 (vreinterpret_s16_f32): Likewise.
1683 (vreinterpret_s16_u8): Likewise.
1684 (vreinterpret_s16_u16): Likewise.
1685 (vreinterpret_s16_u32): Likewise.
1686 (vreinterpret_s16_u64): Likewise.
1687 (vreinterpret_s16_p8): Likewise.
1688 (vreinterpret_s16_p16): Likewise.
1689 (vreinterpretq_s16_s8): Likewise.
1690 (vreinterpretq_s16_s32): Likewise.
1691 (vreinterpretq_s16_s64): Likewise.
1692 (vreinterpretq_s16_f32): Likewise.
1693 (vreinterpretq_s16_u8): Likewise.
1694 (vreinterpretq_s16_u16): Likewise.
1695 (vreinterpretq_s16_u32): Likewise.
1696 (vreinterpretq_s16_u64): Likewise.
1697 (vreinterpretq_s16_p8): Likewise.
1698 (vreinterpretq_s16_p16): Likewise.
1699 (vreinterpret_s32_s8): Likewise.
1700 (vreinterpret_s32_s16): Likewise.
1701 (vreinterpret_s32_s64): Likewise.
1702 (vreinterpret_s32_f32): Likewise.
1703 (vreinterpret_s32_u8): Likewise.
1704 (vreinterpret_s32_u16): Likewise.
1705 (vreinterpret_s32_u32): Likewise.
1706 (vreinterpret_s32_u64): Likewise.
1707 (vreinterpret_s32_p8): Likewise.
1708 (vreinterpret_s32_p16): Likewise.
1709 (vreinterpretq_s32_s8): Likewise.
1710 (vreinterpretq_s32_s16): Likewise.
1711 (vreinterpretq_s32_s64): Likewise.
1712 (vreinterpretq_s32_f32): Likewise.
1713 (vreinterpretq_s32_u8): Likewise.
1714 (vreinterpretq_s32_u16): Likewise.
1715 (vreinterpretq_s32_u32): Likewise.
1716 (vreinterpretq_s32_u64): Likewise.
1717 (vreinterpretq_s32_p8): Likewise.
1718 (vreinterpretq_s32_p16): Likewise.
1719 (vreinterpret_u8_s8): Likewise.
1720 (vreinterpret_u8_s16): Likewise.
1721 (vreinterpret_u8_s32): Likewise.
1722 (vreinterpret_u8_s64): Likewise.
1723 (vreinterpret_u8_f32): Likewise.
1724 (vreinterpret_u8_u16): Likewise.
1725 (vreinterpret_u8_u32): Likewise.
1726 (vreinterpret_u8_u64): Likewise.
1727 (vreinterpret_u8_p8): Likewise.
1728 (vreinterpret_u8_p16): Likewise.
1729 (vreinterpretq_u8_s8): Likewise.
1730 (vreinterpretq_u8_s16): Likewise.
1731 (vreinterpretq_u8_s32): Likewise.
1732 (vreinterpretq_u8_s64): Likewise.
1733 (vreinterpretq_u8_f32): Likewise.
1734 (vreinterpretq_u8_u16): Likewise.
1735 (vreinterpretq_u8_u32): Likewise.
1736 (vreinterpretq_u8_u64): Likewise.
1737 (vreinterpretq_u8_p8): Likewise.
1738 (vreinterpretq_u8_p16): Likewise.
1739 (vreinterpret_u16_s8): Likewise.
1740 (vreinterpret_u16_s16): Likewise.
1741 (vreinterpret_u16_s32): Likewise.
1742 (vreinterpret_u16_s64): Likewise.
1743 (vreinterpret_u16_f32): Likewise.
1744 (vreinterpret_u16_u8): Likewise.
1745 (vreinterpret_u16_u32): Likewise.
1746 (vreinterpret_u16_u64): Likewise.
1747 (vreinterpret_u16_p8): Likewise.
1748 (vreinterpret_u16_p16): Likewise.
1749 (vreinterpretq_u16_s8): Likewise.
1750 (vreinterpretq_u16_s16): Likewise.
1751 (vreinterpretq_u16_s32): Likewise.
1752 (vreinterpretq_u16_s64): Likewise.
1753 (vreinterpretq_u16_f32): Likewise.
1754 (vreinterpretq_u16_u8): Likewise.
1755 (vreinterpretq_u16_u32): Likewise.
1756 (vreinterpretq_u16_u64): Likewise.
1757 (vreinterpretq_u16_p8): Likewise.
1758 (vreinterpretq_u16_p16): Likewise.
1759 (vreinterpret_u32_s8): Likewise.
1760 (vreinterpret_u32_s16): Likewise.
1761 (vreinterpret_u32_s32): Likewise.
1762 (vreinterpret_u32_s64): Likewise.
1763 (vreinterpret_u32_f32): Likewise.
1764 (vreinterpret_u32_u8): Likewise.
1765 (vreinterpret_u32_u16): Likewise.
1766 (vreinterpret_u32_u64): Likewise.
1767 (vreinterpret_u32_p8): Likewise.
1768 (vreinterpret_u32_p16): Likewise.
1769 (vreinterpretq_u32_s8): Likewise.
1770 (vreinterpretq_u32_s16): Likewise.
1771 (vreinterpretq_u32_s32): Likewise.
1772 (vreinterpretq_u32_s64): Likewise.
1773 (vreinterpretq_u32_f32): Likewise.
1774 (vreinterpretq_u32_u8): Likewise.
1775 (vreinterpretq_u32_u16): Likewise.
1776 (vreinterpretq_u32_u64): Likewise.
1777 (vreinterpretq_u32_p8): Likewise.
1778 (vreinterpretq_u32_p16): Likewise.
1780 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1782 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1784 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1787 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1788 (vqnegd_s64): Likewise.
1789 (vqabs_s64): Likewise.
1790 (vqabsd_s64): Likewise.
1792 2014-04-22 Richard Henderson <rth@redhat.com>
1794 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1795 computation to the top of the loop.
1797 2014-04-22 Renlin <renlin.li@arm.com>
1798 Jiong Wang <jiong.wang@arm.com>
1800 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1801 * config/aarch64/aarch64.c (aarch64_layout_frame)
1802 (aarch64_initial_elimination_offset): Likewise.
1804 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1806 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1809 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1811 * machmode.h (bitwise_mode_for_mode): Declare.
1812 * stor-layout.h (bitwise_type_for_mode): Likewise.
1813 * stor-layout.c (bitwise_mode_for_mode): New function.
1814 (bitwise_type_for_mode): Likewise.
1815 * builtins.c (fold_builtin_memory_op): Use it instead of
1816 int_mode_for_mode and build_nonstandard_integer_type.
1818 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1820 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1821 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1822 (*-*-solaris2*): Simplify.
1823 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1824 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1825 *-*-solaris2.9* handling.
1827 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1829 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1830 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1832 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1833 * configure: Regenerate.
1835 * config/i386/sol2-9.h: Remove.
1837 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1838 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1839 Remove Solaris 9 references.
1841 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1843 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1844 (floatuns<GPI:mode><GPF:mode>2): Remove.
1845 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1846 and floatuns conversions.
1847 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1848 and floatuns conversions.
1849 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1850 (w1,w2): New mode attributes for inequal width conversions.
1852 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1854 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1855 the output asm format.
1857 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1859 * config/aarch64/aarch64-simd.md
1860 (aarch64_cm<optab>di): Always split.
1861 (*aarch64_cm<optab>di): New.
1862 (aarch64_cmtstdi): Always split.
1863 (*aarch64_cmtstdi): New.
1865 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1867 PR tree-optimization/60823
1868 * omp-low.c (ipa_simd_modify_function_body): Go through
1869 all SSA_NAMEs and for those refering to vector arguments
1870 which are going to be replaced adjust SSA_NAME_VAR and,
1871 if it is a default definition, change it into a non-default
1872 definition assigned at the beginning of function from new_decl.
1873 (ipa_simd_modify_stmt_ops): Rewritten.
1874 * tree-dfa.c (set_ssa_default_def): When removing default def,
1875 check for NULL loc instead of NULL *loc.
1877 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1879 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1880 restrictions on core registers for DImode values in Thumb2.
1882 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1884 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1885 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1887 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1889 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1890 (*iordi_notzesidi_di): Likewise.
1891 (*iordi_notsesidi_di): Likewise.
1893 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1895 * config/arm/arm-protos.h (tune_params): New struct members.
1896 * config/arm/arm.c: Initialise tune_params per processor.
1897 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1898 for speed, based on new tune_params.
1900 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1902 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1904 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1906 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1908 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1909 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1910 (vrnda_f64): Likewise.
1911 (vrndi_f64): Likewise.
1912 (vrndm_f64): Likewise.
1913 (vrndn_f64): Likewise.
1914 (vrndp_f64): Likewise.
1915 (vrndx_f64): Likewise.
1917 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1919 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1920 GET_MODE_SIZE argument is enum machine_mode.
1922 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1925 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1926 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1928 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1931 * asan.c (asan_emit_stack_protection): Force the base to align to
1932 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1933 appropriate bits if STRICT_ALIGNMENT.
1934 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1936 (expand_used_vars): Leave a space in the stack frame for alignment
1937 if STRICT_ALIGNMENT.
1939 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1941 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1943 (gimple_store_p): Likewise.
1944 (gimple_assign_load_p): Likewise.
1945 (gimple_assign_cast_p): Likewise.
1946 (gimple_clobber_p): Likewise.
1948 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1949 rather than a gimple.
1950 (gimple_assign_cast_p): Likewise.
1952 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1955 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1956 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1958 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1959 more debug information for E500 if -mdebug=reg.
1961 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1964 * config/i386/i386.c (ix86_expand_builtin)
1965 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1966 register for target RTX.
1967 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1969 2014-04-18 Cong Hou <congh@google.com>
1971 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1972 the widen-mult pattern by handling two operands with different sizes,
1973 and operands whose size is smaller than half of the result type.
1975 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1977 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1978 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1979 (do_estimate_edge_time): Compute it.
1980 * ipa-inline.c (want_inline_small_function_p): Bypass
1981 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1983 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1985 * ipa-inline.c (spec_rem): New static variable.
1986 (dump_overall_stats): New function.
1987 (dump_inline_stats): New function.
1989 2014-04-18 Richard Henderson <rth@redhat.com>
1991 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1992 to GET_MODE_SIZE, not a reg_class_t.
1994 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1996 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1997 (vsx_xxmrglw_<mode>): Likewise.
1999 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2002 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
2003 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
2004 (rs6000_init_hard_regno_mode_ok): Likewise.
2006 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
2008 * ipa-inline.c (inline_small_functions): Account only non-cold
2010 * doc/invoke.texi (inline-unit-growth): Update documentation.
2012 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
2014 * config/rs6000/rs6000.md (addti3, subti3): New.
2016 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
2019 * config/i386/i386.c (ix86_expand_clear): Remove outdated
2020 comment. Check optimize_insn_for_size_p instead of
2021 optimize_insn_for_speed_p.
2023 2014-04-17 Martin Jambor <mjambor@suse.cz>
2025 * gimple-iterator.c (gsi_start_edge): New function.
2026 * gimple-iterator.h (gsi_start_edge): Declare.
2027 * tree-sra.c (single_non_eh_succ): New function.
2028 (disqualify_ops_if_throwing_stmt): Renamed to
2029 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
2030 having one non-EH successor BB.
2031 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
2032 generate loads into replacements.
2033 (sra_modify_assign): Likewise and and also use the simple path for
2035 (sra_modify_function_body): Commit statements on edges.
2037 2014-04-17 Richard Biener <rguenther@suse.de>
2040 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
2041 comparison results and add clarifying comment.
2043 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2045 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
2046 (blank_mode): Initialize it.
2047 (emit_mode_size_inline, emit_mode_nunits_inline,
2048 emit_mode_inner_inline): New functions.
2049 (emit_insn_modes_h): Call them and surround their output with
2050 #if GCC_VERSION >= 4001 ... #endif.
2051 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
2052 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
2053 mode_* arrays if the argument is __builtin_constant_p.
2054 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
2055 is enum machine_mode.
2057 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2059 * passes.c (opt_pass::execute): Adjust.
2060 (pass_manager::execute_pass_mode_switching): Likewise.
2061 (early_local_passes::execute): Likewise.
2062 (execute_one_pass): Pass cfun to the pass's execute method.
2063 * tree-pass.h (opt_pass::execute): Add function * argument.
2064 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2065 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2066 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2067 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2068 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2069 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
2070 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
2071 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2072 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2073 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
2074 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
2075 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
2076 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
2077 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
2078 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2079 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2080 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2081 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
2082 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
2083 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2084 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2085 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2086 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2087 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2088 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2089 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2090 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2091 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2092 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2095 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2097 * passes.c (opt_pass::gate): Take function * argument.
2098 (gate_all_early_local_passes): Merge into
2099 (early_local_passes::gate): this.
2100 (gate_all_early_optimizations): Merge into
2101 (all_early_optimizations::gate): this.
2102 (gate_all_optimizations): Mege into
2103 (all_optimizations::gate): this.
2104 (gate_all_optimizations_g): Merge into
2105 (all_optimizations_g::gate): this.
2106 (gate_rest_of_compilation): Mege into
2107 (rest_of_compilation::gate): this.
2108 (gate_postreload): Merge into
2109 (postreload::gate): this.
2110 (dump_one_pass): Pass cfun to the pass's gate method.
2111 (execute_ipa_summary_passes): Likewise.
2112 (execute_one_pass): Likewise.
2113 (ipa_write_summaries_2): Likewise.
2114 (ipa_write_optimization_summaries_1): Likewise.
2115 (ipa_read_summaries_1): Likewise.
2116 (ipa_read_optimization_summaries_1): Likewise.
2117 (execute_ipa_stmt_fixups): Likewise.
2118 * tree-pass.h (opt_pass::gate): Add function * argument.
2119 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2120 combine-stack-adj.c, combine.c, compare-elim.c,
2121 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2122 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2123 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2124 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2125 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2126 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2127 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2128 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2129 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2130 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2131 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2132 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2133 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2134 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2135 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2136 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2137 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2138 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2139 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2140 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2141 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2142 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2143 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2144 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2145 var-tracking.c, vtable-verify.c, web.c: Adjust.
2147 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2149 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2150 * configure: Regenerate.
2152 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2154 * passes.c (dump_one_pass): don't check pass->has_gate.
2155 (execute_ipa_summary_passes): Likewise.
2156 (execute_one_pass): Likewise.
2157 (ipa_write_summaries_2): Likewise.
2158 (ipa_write_optimization_summaries_1): Likewise.
2159 (ipa_read_optimization_summaries_1): Likewise.
2160 (execute_ipa_stmt_fixups): Likewise.
2161 * tree-pass.h (pass_data::has_gate): Remove.
2162 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2163 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2164 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2165 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2166 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2167 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2168 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2169 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2170 gimple-low.c, gimple-ssa-isolate-paths.c,
2171 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2172 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2173 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2174 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2175 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2176 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2177 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2178 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2179 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2180 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2181 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2182 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2183 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2184 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2185 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2186 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2187 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2188 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2189 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2190 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2191 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2192 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2195 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2197 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2199 * passes.c (pass_manager::register_dump_files_1): Merge into
2200 (pass_manager::register_dump_files): this, and remove its handling of
2201 properties since the pass always has the properties anyway.
2202 (pass_manager::pass_manager): Adjust.
2204 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2206 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2207 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2208 dealing with properties.
2209 (pass_manager::register_dump_files): Adjust.
2211 2014-03-20 Mark Wielaard <mjw@redhat.com>
2213 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2214 then represent the bound as normal constant value.
2216 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2219 Forward port from 4.8 branch
2220 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2222 * config/i386/bmiintrin.h (_blsi_u32): New.
2226 (_blsmsk_u32): Ditto.
2227 (_blsmsk_u64): Ditto.
2228 (_tzcnt_u32): Ditto.
2229 (_tzcnt_u64): Ditto.
2231 2014-04-17 Kito Cheng <kito@0xlab.org>
2233 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2235 2014-04-17 Richard Biener <rguenther@suse.de>
2238 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2239 boolean results for comparisons.
2241 2014-04-17 Richard Biener <rguenther@suse.de>
2243 PR tree-optimization/60836
2244 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2245 initial PHI args to be gimple values.
2247 2014-04-17 Richard Biener <rguenther@suse.de>
2249 PR tree-optimization/60841
2250 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2251 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2252 of stmts to SLP build.
2253 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2254 (vect_analyze_slp): Likewise.
2255 (vect_analyze_slp_instance): Likewise.
2256 (vect_build_slp_tree): Limit overall SLP tree growth.
2257 * tree-vectorizer.h (vect_analyze_data_refs,
2258 vect_analyze_slp): Adjust prototypes.
2260 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2262 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2265 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2267 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2268 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2269 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2270 for TARGET_SLOW_PSHUFB
2272 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2274 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2275 * config/i386/i386.c (intel_cost): Ditto.
2277 2014-04-17 Joey Ye <joey.ye@arm.com>
2279 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2281 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2283 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2286 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2288 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2289 (type_all_derivations_known_p): New predicate.
2290 (type_all_ctors_visible_p): New predicate.
2291 (type_possibly_instantiated_p): New predicate.
2292 (get_odr_type): Compute all_derivations_known.
2293 (dump_odr_type): Dump the flag.
2294 (maybe_record_type): Cleanup.
2295 (record_target_from_binfo): Add bases_to_consider array;
2296 record bases for types w/o instances and skip CXX destructor.
2297 (possible_polymorphic_call_targets_1): Add bases_to_consider
2298 and consider_construction parameters; check if type may have instance.
2299 (get_polymorphic_call_info): Set maybe_in_construction to true
2300 when we know nothing.
2301 (record_targets_from_bases): Skip CXX destructors; they are
2302 never called for types in construction.
2303 (possible_polymorphic_call_targets): Do not record target when
2304 type may not have instance.
2306 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2309 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2310 external aliases alive, too.
2312 2014-04-16 Andrew Pinski <apinski@cavium.com>
2314 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2317 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2319 * final.c (compute_alignments): Do not apply loop alignment to a block
2320 falling through to the exit.
2322 2014-04-16 Catherine Moore <clm@codesourcery.com>
2324 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2325 Adjust constraints for microMIPS store patterns.
2327 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2329 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2331 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2333 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2334 (append_use): Run at -O0.
2335 (append_vdef): Likewise.
2336 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2337 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2339 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2341 PR tree-optimization/60844
2342 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2343 (propagate_op_to_single_use, remove_visited_stmt_chain,
2344 linearize_expr, repropagate_negates, reassociate_bb): Use it
2345 instead of gsi_remove.
2347 2014-04-16 Martin Jambor <mjambor@suse.cz>
2349 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2350 ipa_transforms_to_apply.
2351 (cgraph_function_versioning): Assert that old_node has empty
2352 ipa_transforms_to_apply.
2353 * trans-mem.c (ipa_tm_create_version): Likewise.
2354 * tree-inline.c (tree_function_versioning): Do not duplicate
2355 ipa_transforms_to_apply.
2357 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2360 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2362 Pass necessary as flags on 64-bit Solaris/x86.
2363 Use lowercase relocs for x86_64-*-*.
2364 * configure: Regenerate.
2366 2014-04-15 Jan Hubicka <jh@suse.cz>
2368 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2369 (maybe_record_node, likely_target_p): Use it.
2371 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2374 Revert following patch
2376 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2379 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2380 software floating point or no floating point registers, do not
2381 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2382 in GPRs that occurs after we tested for GPRs that would never be
2385 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2386 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2387 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2388 specifically allow DDmode, since that does not use the SPE SIMD
2391 2014-03-21 Mark Wielaard <mjw@redhat.com>
2393 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2394 as unsigned or int depending on type and value used.
2396 2014-04-15 Richard Biener <rguenther@suse.de>
2398 PR rtl-optimization/56965
2399 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2400 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2402 * alias.c (true_dependence_1): Do not call
2403 nonoverlapping_component_refs_p.
2404 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2405 nonoverlapping_component_refs_p.
2406 (indirect_refs_may_alias_p): Likewise.
2408 2014-04-15 Teresa Johnson <tejohnson@google.com>
2410 * cfg.c (dump_bb_info): Fix flags check.
2411 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2413 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2415 PR rtl-optimization/60663
2416 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2419 2014-04-15 Richard Biener <rguenther@suse.de>
2421 * lto-streamer.h (LTO_major_version): Bump to 4.
2423 2014-04-15 Richard Biener <rguenther@suse.de>
2425 * common.opt (lto_partition_model): New enum.
2426 (flto-partition=): Merge separate options with a single with argument,
2427 add -flto-partition=one support.
2428 * flag-types.h (enum lto_partition_model): Declare.
2429 * opts.c (finish_options): Remove duplicate -flto-partition=
2431 * lto-wrapper.c (run_gcc): Adjust.
2433 2014-04-15 Richard Biener <rguenther@suse.de>
2435 * alias.c (ncr_compar): New function.
2436 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2438 2014-04-15 Richard Biener <rguenther@suse.de>
2440 * alias.c (record_component_aliases): Do not walk BINFOs.
2442 2014-04-15 Richard Biener <rguenther@suse.de>
2444 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2445 Add struct function argument and adjust.
2446 (find_func_aliases_for_call): Likewise.
2447 (find_func_aliases): Likewise.
2448 (find_func_clobbers): Likewise.
2449 (intra_create_variable_infos): Likewise.
2450 (compute_points_to_sets): Likewise.
2451 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2453 2014-04-15 Richard Biener <rguenther@suse.de>
2455 * tree.c (iterative_hash_expr): Use enum tree_code_class
2456 to store TREE_CODE_CLASS.
2457 (tree_block): Likewise.
2458 (tree_set_block): Likewise.
2459 * tree.h (fold_build_pointer_plus_loc): Use
2460 convert_to_ptrofftype_loc.
2462 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2465 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2468 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2470 * cfgloop.h (struct loop): Move force_vectorize down.
2471 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2472 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2473 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2474 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2475 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2476 * tree-core.h (enum annot_expr_kind): Add new kind values.
2477 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2478 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2480 * tree.def (ANNOTATE_EXPR): Tweak comment.
2482 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2484 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2487 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2489 * tree.h (TYPE_IDENTIFIER): Declare.
2490 * tree.c (subrange_type_for_debug_p): Use it.
2491 * godump.c (go_format_type): Likewise.
2492 * dwarf2out.c (is_cxx_auto, modified_type_die,
2493 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2494 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2496 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2499 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2501 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2503 * config/i386/i386.c (examine_argument): Return bool. Return true if
2504 parameter should be passed in memory.
2505 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2506 (construct_container): Update calls to examine_argument.
2507 (function_arg_advance_64): Ditto.
2508 (return_in_memory_32): Merge with ix86_return_in_memory.
2509 (return_in_memory_64): Ditto.
2510 (return_in_memory_ms_64): Ditto.
2512 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2514 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2515 * coverage.c (coverage_compute_profile_id): Handle externally visible
2518 2014-04-14 Martin Jambor <mjambor@suse.cz>
2520 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2521 DECL_DISREGARD_INLINE_LIMITS functions.
2523 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2526 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2528 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2531 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2532 optimize_insn_for_speed_p instead of
2533 optimize_function_for_speed_p.
2535 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2537 * doc/invoke.texi (free): Document AArch64.
2539 2014-04-14 Richard Biener <rguenther@suse.de>
2541 PR tree-optimization/60042
2542 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2543 (insert_into_preds_of_block): Do not prevent PHI insertion
2544 for REFERENCE exprs here ...
2545 (eliminate_dom_walker::before_dom_children): ... but prevent
2546 their use here under similar conditions when applied to the
2547 IL after PRE optimizations.
2549 2014-04-14 Richard Biener <rguenther@suse.de>
2551 * passes.def: Move early points-to after early SRA.
2553 2014-04-14 Richard Biener <rguenther@suse.de>
2555 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2556 check for which sign-changes we allow when forwarding
2557 a converted value into a switch.
2559 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2561 * stor-layout.c (place_field): Finalize non-constant offset for the
2564 2014-04-14 Richard Biener <rguenther@suse.de>
2566 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2568 (expand_switch_using_bit_tests_p): Likewise.
2569 (process_switch): Compute and pass on speed_p based on the
2571 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2572 optimize_bb_for_speed_p.
2574 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2576 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2577 * function.h (struct function): Rename has_force_vect_loops into
2578 has_force_vectorize_loops.
2579 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2580 (input_struct_function_base): Likewise.
2581 * lto-streamer-out.c (output_cfg): Likewise.
2582 (output_struct_function_base): Likewise.
2583 * omp-low.c (expand_omp_simd): Likewise.
2584 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2585 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2586 (version_loop_for_if_conversion): Likewise.
2587 (tree_if_conversion): Likewise.
2588 (main_tree_if_conversion): Likewise.
2589 (gate_tree_if_conversion): Likewise.
2590 * tree-inline.c (copy_loops): Likewise.
2591 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2592 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2593 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2594 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2595 * tree-vectorizer.c (vectorize_loops): Likewise.
2596 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2598 2014-04-14 Richard Biener <rguenther@suse.de>
2601 * lto-streamer-out.c (wrap_refs): New function.
2602 (lto_output): Wrap symbol references in global initializes in
2603 type-preserving MEM_REFs.
2605 2014-04-14 Christian Bruel <christian.bruel@st.com>
2607 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2609 2014-04-14 Christian Bruel <christian.bruel@st.com>
2611 * config/sh/sh.md (setmemqi): New expand pattern.
2612 * config/sh/sh.h (CLEAR_RATIO): Define.
2613 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2614 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2616 2014-04-14 Richard Biener <rguenther@suse.de>
2619 * fold-const.c (negate_expr_p): Don't negate directional rounding
2621 (fold_negate_expr): Likewise.
2623 2014-04-14 Richard Biener <rguenther@suse.de>
2625 PR tree-optimization/59817
2626 PR tree-optimization/60453
2627 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2628 recursion to catch all CHRECs in the scalar evolution and restrict
2629 the predicate for the remains appropriately.
2631 2014-04-12 Catherine Moore <clm@codesourcery.com>
2633 * config/mips/constraints.md: Add new register constraint "kb".
2634 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2635 (*movhi_internal): Likewise.
2636 (*movqi_internal): Likewise.
2637 * config/mips/mips.h (M16_STORE_REGS): New register class.
2638 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2639 (REG_CLASS_CONTENTS): Likewise.
2640 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2642 2014-04-11 Tobias Burnus <burnus@net-b.de>
2645 * doc/invoke.texi (-Wformat-signedness): Document it.
2646 (Wformat=2): Mention that this enables -Wformat-signedness.
2648 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2650 * common/config/epiphany/epiphany-common.c
2651 (epiphany_option_optimization_table): Enable section anchors by
2652 default at -O1 or higher.
2653 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2654 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2655 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2656 carries no extra cost.
2657 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2658 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2659 * config/epiphany/predicates.md (memclob_operand): New predicate.
2660 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2661 Use memclob_operand predicate and X constraint for operand 3.
2663 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2665 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2666 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2669 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2671 PR rtl-optimization/60651
2672 * mode-switching.c (optimize_mode_switching): Make sure to emit
2673 sets of a lower numbered entity before sets of a higher numbered
2674 entity to a mode of the same or lower priority.
2675 When creating a seginfo for a basic block that starts with a code
2676 label, move the insertion point past the code label.
2677 (new_seginfo): Document and enforce requirement that
2678 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2679 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2680 * doc/tm.texi: Regenerate.
2682 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2685 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2687 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2689 * BASE-VER: Set to 4.10.0.
2691 2014-04-11 Tobias Burnus <burnus@net-b.de>
2694 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2695 * doc/gcc.texi (Service): Update description in the @menu
2696 * doc/invoke.texi (Option Summary): Remove misplaced and
2699 2014-04-11 Steve Ellcey <sellcey@mips.com>
2700 Jakub Jelinek <jakub@redhat.com>
2703 * expr.c (convert_move): Use emit_store_flag_force instead of
2704 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2707 2014-04-11 Richard Biener <rguenther@suse.de>
2710 * varasm.c (assemble_alias): Avoid endless error reporting
2711 recursion by setting TREE_ASM_WRITTEN.
2713 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2715 * config/s390/s390.md: Add a splitter for NOT rtx.
2717 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2719 PR rtl-optimization/60663
2720 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2722 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2723 Jakub Jelinek <jakub@redhat.com>
2726 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2727 flag from decl_node to node.
2729 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2732 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2733 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2734 ameliorating the cases where it can be.
2736 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2739 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2741 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2742 (loadsync_<mode>): Change mode.
2743 (load_quadpti, store_quadpti): New.
2744 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2745 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2746 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2748 2014-04-09 Cong Hou <congh@google.com>
2751 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2754 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2756 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2757 instead of vnor to exploit possible fusion opportunity in the
2759 (altivec_expand_vec_perm_const_le): Likewise.
2761 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2763 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2764 (loadsync_<mode>): Change mode.
2765 (load_quadpti, store_quadpti): New.
2766 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2767 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2769 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2772 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2773 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2774 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2776 2014-04-08 Richard Biener <rguenther@suse.de>
2779 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2780 a 64bit widest int print double-int similar to on HWI64 hosts.
2782 2014-04-08 Richard Biener <rguenther@suse.de>
2784 PR tree-optimization/60785
2785 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2786 default defs properly.
2788 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2790 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2791 (Weffc++): Likewise.
2793 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2795 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2796 set completep to false rather than true.
2798 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2801 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2802 ARM_TARGET2_DWARF_FORMAT.
2804 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2807 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2808 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2811 2014-04-07 Richard Biener <rguenther@suse.de>
2813 PR tree-optimization/60766
2814 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2815 (may_eliminate_iv): Convert cand_value_at result to desired type.
2817 2014-04-07 Jason Merrill <jason@redhat.com>
2820 * common.opt (-fno-gnu-unique): Add.
2821 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2823 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2825 * haifa-sched.c: Fix outdated function reference and minor
2826 grammar errors in introductory comment.
2828 2014-04-07 Richard Biener <rguenther@suse.de>
2831 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2833 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2835 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2838 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2839 size accounting for thunks.
2840 (pa_asm_output_mi_thunk): Use final_start_function() and
2841 final_end_function() to output function start and end directives.
2843 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2845 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2846 device specific ISA/ feature information. Remove short_sp and
2847 errata_skip ds. Add avr_device_specific_features enum to have device
2849 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2850 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2851 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2852 updated device specific info.
2853 * config/avr/avr-mcus.def: Merge device specific details to
2854 dev_attribute field.
2855 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2857 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2858 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2859 assembler if RMW isa supported by current device.
2860 * config/avr/genmultilib.awk: Update as device info structure changed.
2861 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2863 2014-04-04 Cong Hou <congh@google.com>
2865 PR tree-optimization/60656
2866 * tree-vect-stmts.c (supportable_widening_operation):
2867 Fix a bug that elements in a vector with vect_used_by_reduction
2868 property are incorrectly reordered when the operation on it is not
2869 consistant with the one in reduction operation.
2871 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2873 PR rtl-optimization/60155
2874 * gcse.c (record_set_data): New function.
2875 (single_set_gcse): New function.
2876 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2877 (hoist_code): Likewise.
2878 (get_pressure_class_and_nregs): Likewise.
2880 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2882 * explow.c (probe_stack_range): Emit a final optimization blockage.
2884 2014-04-04 Anthony Green <green@moxielogic.com>
2886 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2889 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2892 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2893 flags are set only during streaming.
2894 * ipa.c (process_references, walk_polymorphic_call_targets,
2895 symtab_remove_unreachable_nodes): Drop bodies of always inline
2896 after early inlining.
2897 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2899 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2900 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2903 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2906 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2909 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2911 (cortex_a53_fdivd): Likewise.
2913 2014-04-04 Martin Jambor <mjambor@suse.cz>
2916 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2918 * cgraph.c (clone_of_p): Also return true if thunks match.
2919 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2920 cgraph_function_or_thunk_node and an obsolete comment.
2921 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2923 (build_function_decl_skip_args): Likewise.
2924 (set_new_clone_decl_and_node_flags): New function.
2925 (duplicate_thunk_for_node): Likewise.
2926 (redirect_edge_duplicating_thunks): Likewise.
2927 (cgraph_clone_node): New parameter args_to_skip, pass it to
2928 redirect_edge_duplicating_thunks which is called instead of
2929 cgraph_redirect_edge_callee.
2930 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2931 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2933 2014-04-04 Jeff Law <law@redhat.com>
2936 * config/arm/predicates.md (const_int_I_operand): New predicate.
2937 (const_int_M_operand): Similarly.
2938 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2940 (insv_t2, extv_reg, extzv_t2): Likewise.
2941 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2942 (pop_multiple_with_writeback_and_return): Likewise.
2943 (vfp_pop_multiple_with_writeback): Likewise
2945 2014-04-04 Richard Biener <rguenther@suse.de>
2948 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2949 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2951 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2952 * gimplify.c (gimple_add_tmp_var_fn): New function.
2953 * gimple-expr.h (create_tmp_reg_fn): Declare.
2954 * gimple-expr.c (create_tmp_reg_fn): New function.
2955 * gimple-low.c (record_vars_into): Don't change cfun.
2956 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2957 code generation without cfun.
2959 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2962 * Makefile.in (install-driver): Fix shell scripting.
2964 2014-04-03 Cong Hou <congh@google.com>
2966 PR tree-optimization/60505
2967 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2968 threshold of number of iterations below which no vectorization
2970 * tree-vect-loop.c (new_loop_vec_info):
2971 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2972 * tree-vect-loop.c (vect_analyze_loop_operations):
2973 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2974 * tree-vect-loop.c (vect_transform_loop):
2975 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2976 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2977 of iterations of the loop and see if we should build the epilogue.
2979 2014-04-03 Richard Biener <rguenther@suse.de>
2981 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2982 (streamer_tree_cache_create): Adjust.
2983 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2984 to allow optional nodes array.
2985 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2986 (streamer_tree_cache_append): Likewise.
2987 (streamer_tree_cache_create): Create nodes array optionally
2988 as specified by parameter.
2989 * lto-streamer-out.c (create_output_block): Avoid maintaining
2990 the node array in the writer cache.
2991 (DFS_write_tree): Remove assertion.
2992 (produce_asm_for_decls): Free the out decl state hash table early.
2993 * lto-streamer-in.c (lto_data_in_create): Adjust for
2994 streamer_tree_cache_create prototype change.
2996 2014-04-03 Richard Biener <rguenther@suse.de>
2998 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2999 set TREE_CHAIN to NULL_TREE.
3001 2014-04-03 Richard Biener <rguenther@suse.de>
3003 PR tree-optimization/60740
3004 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
3005 over all GIMPLE_COND operands.
3007 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
3009 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
3010 (Weffc++): Remove Scott's numbering, merge lists and reference
3013 2014-04-03 Nick Clifton <nickc@redhat.com>
3015 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
3018 2014-04-03 Martin Jambor <mjambor@suse.cz>
3020 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
3021 mention gcc_unreachable before failing.
3022 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
3025 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
3028 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
3029 inconsistent code and instead mark the context inconsistent.
3030 (possible_polymorphic_call_targets): For inconsistent contexts
3031 return empty complete list.
3033 2014-04-02 Anthony Green <green@moxielogic.com>
3035 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
3036 (extendqisi2, extendhisi2): Define.
3037 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
3038 (WCHAR_TYPE): Change to unsigned int.
3040 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3042 PR tree-optimization/60733
3043 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
3044 insertion point for PHI candidates to be the end of the feeding
3045 block for the PHI argument.
3047 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
3049 PR rtl-optimization/60650
3050 * lra-constraints.c (process_alt_operands): Decrease reject for
3051 earlyclobber matching.
3053 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3055 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
3057 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3059 * config/spu/spu.c (pad_bb): Do not crash when the last
3060 insn is CODE_FOR_blockage.
3062 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3064 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
3065 lies outside the target mode.
3067 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3070 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
3071 software floating point or no floating point registers, do not
3072 allow any type in the FPRs. Eliminate a test for SPE SIMD types
3073 in GPRs that occurs after we tested for GPRs that would never be
3076 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
3077 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
3078 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
3079 specifically allow DDmode, since that does not use the SPE SIMD
3082 2014-04-02 Richard Biener <rguenther@suse.de>
3085 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
3086 MODE_INTs. Properly use negv_optab.
3087 (expand_abs): Likewise.
3089 2014-04-02 Richard Biener <rguenther@suse.de>
3092 * Makefile.in (install-driver): Guard extra installs with special
3095 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3097 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3100 2014-04-01 Richard Henderson <rth@redhat.com>
3103 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3104 alternative enabled before register allocation.
3106 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3108 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3109 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3111 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3112 (nios2_got_address): Update nios2_large_got_address call site.
3113 (nios2_delegitimize_address): New function.
3114 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3115 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3116 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3118 2014-04-01 Martin Husemann <martin@duskware.de>
3120 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3123 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3125 PR rtl-optimization/60604
3126 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3127 check from register_operand.
3128 (register_operand): Redefine in terms of general_operand.
3129 (nonmemory_operand): Use register_operand for the non-constant cases.
3131 2014-04-01 Richard Biener <rguenther@suse.de>
3133 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3135 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3137 * doc/invoke.texi (mapp-regs): Clarify.
3139 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3141 * config/i386/avx512fintrin.h (__v32hi): Define type.
3142 (__v64qi): Likewise.
3143 (_mm512_set1_epi8): Define.
3144 (_mm512_set1_epi16): Define.
3145 (_mm512_set4_epi32): Define.
3146 (_mm512_set4_epi64): Define.
3147 (_mm512_set4_pd): Define.
3148 (_mm512_set4_ps): Define.
3149 (_mm512_setr4_epi64): Define.
3150 (_mm512_setr4_epi32): Define.
3151 (_mm512_setr4_pd): Define.
3152 (_mm512_setr4_ps): Define.
3153 (_mm512_setzero_epi32): Define.
3155 2014-03-31 Martin Jambor <mjambor@suse.cz>
3158 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3159 callsite_arguments_match_p. Updated all callers. Also check types of
3160 corresponding formal parameters and actual arguments.
3161 (not_all_callers_have_enough_arguments_p) Renamed to
3162 some_callers_have_mismatched_arguments_p.
3164 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3166 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3168 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3171 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3174 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3176 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3177 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3179 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3180 Use FMAMODE_NOVF512 mode iterator.
3181 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3182 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3183 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3184 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3186 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3187 Use VF_128_256 mode iterator.
3188 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3191 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3193 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3194 static chain if needed.
3196 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3199 * lra-constraints.c (index_part_to_reg): New.
3200 (process_address): Use it.
3202 2014-03-27 Jeff Law <law@redhat.com>
3203 Jakub Jelinek <jakub@redhat.com>
3206 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3207 gen_rtx_{PLUS,MULT} to build up the address expression.
3209 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3210 creating non-canonical RTL.
3212 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3215 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3216 functions; reorganize to make cheap checks first.
3217 (inline_small_functions): Do not estimate growth when dumping;
3219 * ipa-inline.h (inline_summary): Add min_size.
3220 (growth_likely_positive): New function.
3221 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3222 (set_cond_stmt_execution_predicate): Cleanup.
3223 (estimate_edge_size_and_time): Compute min_size.
3224 (estimate_calls_size_and_time): Likewise.
3225 (estimate_node_size_and_time): Likewise.
3226 (inline_update_overall_summary): Update min_size.
3227 (do_estimate_edge_time): Likewise.
3228 (do_estimate_edge_size): Update.
3229 (do_estimate_edge_hints): Update.
3230 (growth_likely_positive): New function.
3232 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3235 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3236 also if addr has VOIDmode.
3238 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3240 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3241 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3243 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3244 instructions as well as AdvancedSIMD loads.
3246 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3248 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3249 Use crypto_aese type.
3250 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3251 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3252 crypto_aese, crypto_aesmc. Move to types.md.
3253 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3255 * config/arm/iterators.md (crypto_type): Likewise.
3257 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3259 * cgraph.c: Include expr.h and tree-dfa.h.
3260 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3263 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3266 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3267 regs from checking multi-reg pseudos.
3269 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3271 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3273 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3275 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3276 if it would clobber the stack pointer, even temporarily.
3278 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3280 * mode-switching.c: Make small adjustments to the top comment.
3282 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3284 * config/rs6000/constraints.md (wD constraint): New constraint to
3285 match the constant integer to get the top DImode/DFmode out of a
3286 vector in a VSX register.
3288 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3289 match the constant integer to get the top DImode/DFmode out of a
3290 vector in a VSX register.
3292 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3295 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3298 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3299 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3301 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3302 Optimize vec_extract of 64-bit values, where the value being
3303 extracted is in the top word, where we can use scalar
3304 instructions. Add direct move and store support. Combine the big
3305 endian/little endian vector select load support into a single insn.
3306 (vsx_extract_<mode>_internal1): Likewise.
3307 (vsx_extract_<mode>_internal2): Likewise.
3308 (vsx_extract_<mode>_load): Likewise.
3309 (vsx_extract_<mode>_store): Likewise.
3310 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3311 combined into vsx_extract_<mode>_load.
3312 (vsx_extract_<mode>_one_le): Likewise.
3314 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3315 define the top 64-bit vector element.
3317 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3320 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3321 Document vec_vbpermq builtin.
3324 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3325 enable use of xxsldwi and xxpermdi builtin functions.
3326 (vec_xxpermdi): Likewise.
3328 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3329 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3331 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3333 PR rtl-optimization/60650
3334 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3336 (find_spills_for): New.
3337 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3338 Spill all pseudos on the second iteration.
3340 2014-03-27 Marek Polacek <polacek@redhat.com>
3343 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3346 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3348 * config/s390/s390.c (s390_can_use_return_insn): Check for
3349 call-saved FPRs on 31 bit.
3351 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3354 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3355 if they need regimplification, just drop them instead of
3356 calling gimple_regimplify_operands on them.
3358 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3361 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3362 (aarch64_frame_pointer_required): Adjust logic.
3363 (aarch64_can_eliminate): Adjust logic.
3364 (aarch64_override_options_after_change): Adjust logic.
3366 2014-03-27 Dehao Chen <dehao@google.com>
3368 * ipa-inline.c (early_inliner): Update node's inline info.
3370 2014-03-26 Dehao Chen <dehao@google.com>
3372 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3373 compiler inserted conditional jumps for NAN float check.
3375 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3377 * ubsan.h (ubsan_create_data): Change second argument's type
3378 to const location_t *.
3379 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3381 (ubsan_create_data): Change second argument to const location_t *PLOC.
3382 Create Loc field whenever PLOC is non-NULL.
3383 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3384 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3388 * real.c (real_to_integer2): Change type of low to UHWI.
3390 2014-03-26 Tobias Burnus <burnus@net-b.de>
3392 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3393 (CILK_SELF_SPECS): New define.
3394 (driver_self_specs): Use it.
3396 2014-03-26 Richard Biener <rguenther@suse.de>
3398 * tree-pretty-print.c (percent_K_format): Implement special
3399 case for LTO and its stripped down BLOCK tree.
3401 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3404 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3406 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3407 one range is range_int_cst_p, but not both, at least optimize
3408 addition/subtraction of 0 and multiplication by 0 or 1.
3409 * gimple-fold.c (gimple_fold_call): Fold
3410 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3411 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3412 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3414 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3416 PR rtl-optimization/60452
3417 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3418 <case REG>: Return 1 for invalid offsets from the frame pointer.
3420 2014-03-26 Marek Polacek <polacek@redhat.com>
3423 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3426 2014-03-26 Marek Polacek <polacek@redhat.com>
3429 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3432 2014-03-26 Marek Polacek <polacek@redhat.com>
3435 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3436 multiplication in unsigned type.
3438 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3440 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3442 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3444 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3446 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3449 * cif-code.def (UNREACHABLE) New code.
3450 * ipa-inline.c (inline_small_functions): Skip edges to
3451 __builtlin_unreachable.
3452 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3453 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3454 predicate to __bulitin_unreachable.
3455 (set_cond_stmt_execution_predicate): Fix issue when
3456 invert_tree_comparison returns ERROR_MARK.
3457 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3458 propagate to inline clones.
3459 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3461 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3462 * cgraphclones.c (cgraph_clone_node): If call destination is already
3463 ureachable, do not redirect it back.
3464 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3467 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3469 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3470 Do not modify inline clones.
3472 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3474 * config/i386/i386.md (general_sext_operand): New mode attr.
3475 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3476 don't generate (sign_extend (const_int)).
3477 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3478 operands[2]. Use We constraint instead of <i> and
3479 <general_sext_operand> predicate instead of <general_operand>.
3480 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3481 * config/i386/constraints.md (We): New constraint.
3482 * config/i386/predicates.md (x86_64_sext_operand,
3483 sext_operand): New predicates.
3485 2014-03-25 Martin Jambor <mjambor@suse.cz>
3488 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3489 inconsistent devirtualizations to __builtin_unreachable.
3491 2014-03-25 Marek Polacek <polacek@redhat.com>
3494 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3496 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3498 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3499 order of elements for big-endian.
3501 2014-03-25 Richard Biener <rguenther@suse.de>
3504 * gimplify-me.c (gimple_regimplify_operands): Update the
3507 2014-03-25 Martin Jambor <mjambor@suse.cz>
3510 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3511 (lto_output_varpool_node): Likewise.
3512 (input_overwrite_node): Likewise.
3513 (input_varpool_node): Likewise.
3515 2014-03-25 Richard Biener <rguenther@suse.de>
3517 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3518 (run_gcc): Likewise.
3520 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3522 * combine.c (simplify_compare_const): Add MODE argument.
3523 Handle mode_width 0 as very large mode_width.
3524 (try_combine, simplify_comparison): Adjust callers.
3526 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3527 type to avoid signed integer overflow.
3528 * explow.c (plus_constant): Likewise.
3530 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3532 * doc/generic.texi: Correct typos.
3534 2014-03-24 Tobias Burnus <burnus@net-b.de>
3536 * doc/invoke.texi (-flto): Expand section about
3537 using static libraries with LTO.
3539 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3541 PR rtl-optimization/60501
3542 * optabs.def (addptr3_optab): New optab.
3543 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3544 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3545 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3547 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3549 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3551 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3553 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3556 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3557 (_mm256_undefined_ps): Define.
3558 (_mm256_undefined_pd): Define.
3559 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3560 (_mm_undefined_pd): Define.
3561 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3562 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3563 (_mm512_undefined_ps): Define.
3564 (_mm512_undefined_pd): Define.
3565 Use _mm*_undefined_*.
3566 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3568 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3570 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3571 (lshr_simd): DI mode added.
3572 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3573 (aarch64_ushr_simddi): Likewise.
3574 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3575 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3576 (vshrd_n_u64): Likewise.
3578 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3580 * Makefile.in (s-macro_list): Depend on cc1.
3582 2014-03-23 Teresa Johnson <tejohnson@google.com>
3584 * ipa-utils.c (ipa_print_order): Use specified dump file.
3586 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3588 PR rtl-optimization/60601
3589 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3591 * gcc.c (eval_spec_function): Initialize save_growing_value.
3593 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3596 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3597 code == MINUS_EXPR, never swap op0 with op1.
3599 * toplev.c (init_local_tick): Avoid signed integer multiplication
3601 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3602 shift by first operand's bitsize.
3604 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3607 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3609 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3610 TARGET_ISA_64BIT_P(x).
3612 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3614 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3615 pattern for vector nor instead of subtract from splat(-1).
3616 (altivec_expand_vec_perm_const_le): Likewise.
3618 2014-03-21 Richard Henderson <rth@twiddle.net>
3621 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3622 related insns after epilogue_completed.
3624 2014-03-21 Martin Jambor <mjambor@suse.cz>
3627 * cgraph.h (symtab_node): New flag body_removed.
3628 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3629 when removing bodies.
3630 * symtab.c (dump_symtab_base): Dump body_removed flag.
3631 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3632 had their bodies removed.
3634 2014-03-21 Martin Jambor <mjambor@suse.cz>
3637 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3640 2014-03-21 Richard Biener <rguenther@suse.de>
3642 PR tree-optimization/60577
3643 * tree-core.h (struct tree_base): Document nothrow_flag use
3644 in VAR_DECL_NONALIASED.
3645 * tree.h (VAR_DECL_NONALIASED): New.
3646 (may_be_aliased): Adjust.
3647 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3649 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3651 * expr.c (expand_expr_real_1): Remove outdated comment.
3653 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3656 * ira.c (adjust_cleared_regs): Call copy_rtx on
3657 *reg_equiv[REGNO (loc)].src_p before passing it to
3658 simplify_replace_fn_rtx.
3661 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3662 into CONST, put pic register as first operand of PLUS. Use
3663 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3665 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3667 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3669 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3671 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3672 around for store forwarding issue in the FPU on the UT699.
3673 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3674 loads and operations if -mfix-ut699 is specified.
3675 (divtf3_hq): Tweak attribute.
3676 (sqrttf2_hq): Likewise.
3678 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3680 * calls.c (store_one_arg): Remove incorrect const qualification on the
3681 type of the temporary.
3682 * cfgexpand.c (expand_return): Likewise.
3683 * expr.c (expand_constructor): Likewise.
3684 (expand_expr_real_1): Likewise.
3686 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3688 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3691 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3694 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3696 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3698 * config/arm/aarch-common-protos.h
3699 (alu_cost_table): Fix spelling of "extend".
3700 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3702 2014-03-19 Richard Biener <rguenther@suse.de>
3705 * tree-core.h (tree_type_common): Re-order pointer members
3706 to reduce recursion depth during GC walks.
3708 2014-03-19 Marek Polacek <polacek@redhat.com>
3711 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3712 before accessing it.
3714 2014-03-19 Richard Biener <rguenther@suse.de>
3717 * lto-streamer-in.c (input_function): In WPA stage do not drop
3720 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3722 PR tree-optimization/60559
3723 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3724 with build_zero_cst assignment.
3726 2014-03-18 Kai Tietz <ktietz@redhat.com>
3728 PR rtl-optimization/56356
3729 * sdbout.c (sdbout_parms): Verify that parms'
3730 incoming argument is valid.
3731 (sdbout_reg_parms): Likewise.
3733 2014-03-18 Richard Henderson <rth@redhat.com>
3736 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3737 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3738 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3740 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3742 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3743 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3744 Italicize plugin event names in description. Explain that
3745 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3746 Remind that no GCC functions should be called after PLUGIN_FINISH.
3747 Explain what pragmas with expansion are.
3749 2014-03-18 Martin Liska <mliska@suse.cz>
3751 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3752 gimple call statement is update.
3753 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3754 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3756 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3759 * ubsan.c (ubsan_instrument_unreachable): Call
3760 initialize_sanitizer_builtins.
3761 (ubsan_pass): Likewise.
3764 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3765 varpool_finalize_decl instead of rest_of_decl_compilation.
3767 2014-03-18 Richard Biener <rguenther@suse.de>
3769 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3770 by using bitmap_and_compl instead of bitmap_and_compl_into.
3771 (df_rd_transfer_function): Likewise.
3773 2014-03-18 Richard Biener <rguenther@suse.de>
3775 * doc/lto.texi (fresolution): Fix typo.
3777 2014-03-18 Richard Biener <rguenther@suse.de>
3779 * doc/invoke.texi (flto): Update for changes in 4.9.
3781 2014-03-18 Richard Biener <rguenther@suse.de>
3783 * doc/loop.texi: Remove section on the removed lambda framework.
3784 Update loop docs with recent changes in preserving loop structure.
3786 2014-03-18 Richard Biener <rguenther@suse.de>
3788 * doc/lto.texi (-fresolution): Document.
3790 2014-03-18 Richard Biener <rguenther@suse.de>
3792 * doc/contrib.texi: Adjust my name.
3794 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3797 * internal-fn.c: Include diagnostic-core.h.
3798 (expand_BUILTIN_EXPECT): New function.
3799 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3800 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3801 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3802 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3804 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3805 Revert 3 argument __builtin_expect code.
3806 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3807 * gimple-fold.c (gimple_fold_call): Likewise.
3808 * tree.h (fold_builtin_expect): New prototype.
3809 * builtins.c (build_builtin_expect_predicate): Add predictor
3810 argument, if non-NULL, create 3 argument __builtin_expect.
3811 (fold_builtin_expect): No longer static. Add ARG2 argument,
3812 pass it through to build_builtin_expect_predicate.
3813 (fold_builtin_2): Adjust caller.
3814 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3815 * internal-fn.def (BUILTIN_EXPECT): New.
3817 2014-03-18 Tobias Burnus <burnus@net-b.de>
3820 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3821 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3822 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3824 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3827 * predict.c (combine_predictions_for_bb): Fix up formatting.
3828 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3829 fill what it points to if non-NULL.
3830 (tree_predict_by_opcode): Adjust caller, use the predictor.
3831 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3833 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3835 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3836 proper constant for the store mode.
3838 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3840 * symtab.c (change_decl_assembler_name): Fix transparent alias
3843 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3845 * config/aarch64/aarch64.c: Correct the comments about the
3846 aarch64 stack layout.
3848 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3850 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3851 check for GF_OMP_FOR_KIND_FOR.
3853 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3855 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3856 ymm and zmm register names.
3858 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3861 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3862 note creation for the 2010-08-31 changes.
3864 2014-03-17 Marek Polacek <polacek@redhat.com>
3867 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3868 as -fno-tree-loop-vectorize.
3869 (expand_omp_simd): Likewise.
3871 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3873 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3874 (eligible_for_call_delay): New prototype.
3875 * config/sparc/sparc.c (tls_call_delay): Rename into...
3876 (eligible_for_call_delay): ...this. Return false if the instruction
3877 cannot be put in the delay slot of a branch.
3878 (eligible_for_restore_insn): Simplify.
3879 (eligible_for_return_delay): Return false if the instruction cannot be
3880 put in the delay slot of a branch and simplify.
3881 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3882 put in the delay slot of a branch.
3883 * config/sparc/sparc.md (fix_ut699): New attribute.
3884 (tls_call_delay): Delete.
3885 (in_call_delay): Reimplement.
3886 (eligible_for_sibcall_delay): Rename into...
3887 (in_sibcall_delay): ...this.
3888 (eligible_for_return_delay): Rename into...
3889 (in_return_delay): ...this.
3890 (in_branch_delay): Reimplement.
3891 (in_uncond_branch_delay): Delete.
3892 (in_annul_branch_delay): Delete.
3894 2014-03-14 Richard Henderson <rth@redhat.com>
3897 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3898 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3899 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3900 (floathi splitters): Remove.
3901 (float<SWI48x>xf2): New pattern.
3902 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3903 code that tried to handle DImode for 32-bit, but which was excluded
3904 by the pattern's condition. Drop allocation of stack temporary.
3905 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3906 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3907 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3908 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3909 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3910 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3911 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3912 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3913 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3914 (*float<SWI48x><X87MODEF>2_i387): Remove.
3915 (all float _with_temp splitters): Remove.
3916 (*float<SWI48x><MODEF>2_i387): New pattern.
3917 (*float<SWI48><MODEF>2_sse): New pattern.
3918 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3919 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3921 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3922 Marek Polacek <polacek@redhat.com>
3925 * common.opt (dump_base_name_prefixed): New Variable.
3926 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3927 if x_dump_base_name_prefixed is already set, set it at the end.
3929 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3931 PR rtl-optimization/60508
3932 * lra-constraints.c (get_reload_reg): Add new parameter
3934 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3935 Pass the new parameter values.
3937 2014-03-14 Richard Biener <rguenther@suse.de>
3939 * common.opt: Revert unintented changes from r205065.
3942 2014-03-14 Richard Biener <rguenther@suse.de>
3945 * cfghooks.c (split_block): Properly adjust all loops the
3946 block was a latch of.
3948 2014-03-14 Martin Jambor <mjambor@suse.cz>
3951 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3954 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3957 * config/avr/avr.c (avr_set_current_function): Pass function name
3958 through default_strip_name_encoding before sanity checking instead
3959 of skipping the first char of the assembler name.
3961 2014-03-13 Richard Henderson <rth@redhat.com>
3964 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3965 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3966 * config/i386/i386-protos.h: Likewise.
3967 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3968 in the expander instead of a splitter.
3969 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3970 any possibility of requiring a memory.
3971 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3972 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3973 (fp branch splitters): Update for ix86_split_fp_branch.
3974 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3975 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3976 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3977 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3978 (*fop_<MODEF>_3_i387): Likewise.
3979 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3980 (splitters for the fop_* register patterns): Remove.
3981 (fscalexf4_i387): Rename from *fscalexf4_i387.
3982 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3984 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3986 PR tree-optimization/59779
3987 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3988 type for bitsize and maxsize instead of HOST_WIDE_INT.
3990 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3992 PR rtl-optimization/57320
3993 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3994 the CFG after thread_prologue_and_epilogue_insns.
3996 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3998 PR rtl-optimization/57189
3999 * lra-constraints.c (process_alt_operands): Disfavor spilling
4002 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
4004 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
4006 2014-03-13 Jakub Jelinek <jakub@redhat.com>
4008 PR tree-optimization/59025
4010 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
4011 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
4013 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
4016 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
4017 calls of avr_out_plus_1.
4019 2014-03-13 Bin Cheng <bin.cheng@arm.com>
4021 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
4022 BB's single pred and update the father loop's latch info later.
4024 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
4026 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
4030 (VEC_base): Likewise.
4031 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
4032 registers, we need to swap double words in little endian mode.
4034 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
4035 to be a container mode for 128-bit integer operations added in ISA
4036 2.07. Unlike TImode and PTImode, the preferred register set is
4037 the Altivec/VMX registers for the 128-bit operations.
4039 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
4041 (rs6000_split_128bit_ok_p): Likewise.
4043 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
4044 macros for creating ISA 2.07 normal and overloaded builtin
4045 functions with 3 arguments.
4046 (BU_P8V_OVERLOAD_3): Likewise.
4047 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
4048 for use as overloaded functions.
4049 (VPERM_1TI_UNS): Likewise.
4050 (VSEL_1TI): Likewise.
4051 (VSEL_1TI_UNS): Likewise.
4052 (ST_INTERNAL_1ti): Likewise.
4053 (LD_INTERNAL_1ti): Likewise.
4054 (XXSEL_1TI): Likewise.
4055 (XXSEL_1TI_UNS): Likewise.
4056 (VPERM_1TI): Likewise.
4057 (VPERM_1TI_UNS): Likewise.
4058 (XXPERMDI_1TI): Likewise.
4059 (SET_1TI): Likewise.
4060 (LXVD2X_V1TI): Likewise.
4061 (STXVD2X_V1TI): Likewise.
4062 (VEC_INIT_V1TI): Likewise.
4063 (VEC_SET_V1TI): Likewise.
4064 (VEC_EXT_V1TI): Likewise.
4065 (EQV_V1TI): Likewise.
4066 (NAND_V1TI): Likewise.
4067 (ORC_V1TI): Likewise.
4068 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
4069 added in ISA 2.07. Add both normal 'altivec' builtins, and the
4071 (VADDUQM): Likewise.
4072 (VSUBCUQ): Likewise.
4073 (VADDEUQM): Likewise.
4074 (VADDECUQ): Likewise.
4075 (VSUBEUQM): Likewise.
4076 (VSUBECUQ): Likewise.
4078 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
4079 __int128_t and __uint128_t types.
4080 (__uint128_type): Likewise.
4081 (altivec_categorize_keyword): Add support for vector __int128_t,
4082 vector __uint128_t, vector __int128, and vector unsigned __int128
4083 as a container type for TImode operations that need to be done in
4084 VSX/Altivec registers.
4085 (rs6000_macro_to_expand): Likewise.
4086 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
4087 to support 128-bit integer instructions vaddcuq, vadduqm,
4088 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
4089 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4091 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4092 for V1TImode, and set up preferences to use VSX/Altivec registers.
4093 Setup VSX reload handlers.
4094 (rs6000_debug_reg_global): Likewise.
4095 (rs6000_init_hard_regno_mode_ok): Likewise.
4096 (rs6000_preferred_simd_mode): Likewise.
4097 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4098 (easy_altivec_constant): Likewise.
4099 (output_vec_const_move): Likewise.
4100 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4102 (rs6000_expand_vector_extract): Likewise.
4103 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4105 (rs6000_const_vec): Add support for V1TImode.
4106 (rs6000_emit_le_vsx_load): Swap double words when loading or
4107 storing TImode/V1TImode.
4108 (rs6000_emit_le_vsx_store): Likewise.
4109 (rs6000_emit_le_vsx_move): Likewise.
4110 (rs6000_emit_move): Add support for V1TImode.
4111 (altivec_expand_ld_builtin): Likewise.
4112 (altivec_expand_st_builtin): Likewise.
4113 (altivec_expand_vec_init_builtin): Likewise.
4114 (altivec_expand_builtin): Likewise.
4115 (rs6000_init_builtins): Add support for V1TImode type. Add
4116 support for ISA 2.07 128-bit integer builtins. Define type names
4117 for the VSX/Altivec vector types.
4118 (altivec_init_builtins): Add support for overloaded vector
4119 functions with V1TImode type.
4120 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4121 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4123 (rs6000_split_128bit_ok_p): Likewise.
4124 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4125 __int128_t and vector __uint128_t.
4127 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4128 and mode attributes.
4135 (VS_scalar): Likewise.
4136 (VS_double): Likewise.
4137 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4139 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4140 we support the ISA 2.07 128-bit integer arithmetic instructions.
4141 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4142 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4143 and TImode types for use with the builtin functions.
4144 (V1TI_type_node): Likewise.
4145 (unsigned_V1TI_type_node): Likewise.
4146 (intTI_type_internal_node): Likewise.
4147 (uintTI_type_internal_node): Likewise.
4149 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4150 128-bit builtin functions.
4151 (UNSPEC_VADDEUQM): Likewise.
4152 (UNSPEC_VADDECUQ): Likewise.
4153 (UNSPEC_VSUBCUQ): Likewise.
4154 (UNSPEC_VSUBEUQM): Likewise.
4155 (UNSPEC_VSUBECUQ): Likewise.
4156 (VM): Add V1TImode to vector mode iterators.
4158 (VI_unit): Likewise.
4159 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4160 (altivec_vaddcuq): Likewise.
4161 (altivec_vsubuqm): Likewise.
4162 (altivec_vsubcuq): Likewise.
4163 (altivec_vaddeuqm): Likewise.
4164 (altivec_vaddecuq): Likewise.
4165 (altivec_vsubeuqm): Likewise.
4166 (altivec_vsubecuq): Likewise.
4168 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4170 (BOOL_128): Likewise.
4171 (BOOL_REGS_OUTPUT): Likewise.
4172 (BOOL_REGS_OP1): Likewise.
4173 (BOOL_REGS_OP2): Likewise.
4174 (BOOL_REGS_UNARY): Likewise.
4175 (BOOL_REGS_AND_CR0): Likewise.
4177 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4178 128-bit integer builtin support.
4179 (vec_vadduqm): Likewise.
4180 (vec_vaddecuq): Likewise.
4181 (vec_vaddeuqm): Likewise.
4182 (vec_vsubecuq): Likewise.
4183 (vec_vsubeuqm): Likewise.
4184 (vec_vsubcuq): Likewise.
4185 (vec_vsubuqm): Likewise.
4187 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4188 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4189 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4190 128-bit integer add/subtract to ISA 2.07.
4192 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4194 * config/arc/arc.c (arc_predicate_delay_insns):
4195 Fix third argument passed to conditionalize_nonjump.
4197 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4199 * config/aarch64/aarch64-builtins.c
4200 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4201 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4202 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4203 instead of __builtin_lfloor.
4204 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4206 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4208 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4209 (tree_ssa_ifcombine_bb_1): New function.
4210 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4211 is an empty forwarder block to then_bb or vice versa and then_bb
4212 and else_bb are effectively swapped.
4214 2014-03-12 Christian Bruel <christian.bruel@st.com>
4217 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4218 REG_CFA_DEF_CFA note.
4219 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4220 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4222 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4224 PR tree-optimization/60454
4225 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4227 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4229 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4230 Do not define target_cpu_default2 to generic.
4231 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4232 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4233 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4235 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4236 Marc Glisse <marc.glisse@inria.fr>
4238 PR tree-optimization/60502
4239 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4240 instead of build_low_bits_mask.
4242 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4245 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4246 if there are multiple uses, but op doesn't live on E edge.
4247 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4248 clobber stmts before __builtin_unreachable.
4250 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4252 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4253 hard_frame_pointer_rtx.
4254 * cse.c (cse_insn): Remove volatile check.
4255 * cselib.c (cselib_process_insn): Likewise.
4256 * dse.c (scan_insn): Likewise.
4258 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4260 * config/arc/arc.c (conditionalize_nonjump): New function,
4262 (arc_ifcvt): ... this.
4263 (arc_predicate_delay_insns): Use it.
4265 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4267 * config/arc/predicates.md (extend_operand): During/after reload,
4268 allow const_int_operand.
4269 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4270 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4271 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4273 (umulsi3_highpart_i): Likewise.
4275 2014-03-11 Richard Biener <rguenther@suse.de>
4277 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4278 Add asserts to guard possible wrong-code bugs.
4280 2014-03-11 Richard Biener <rguenther@suse.de>
4282 PR tree-optimization/60429
4283 PR tree-optimization/60485
4284 * tree-ssa-structalias.c (set_union_with_increment): Properly
4285 take into account all fields that overlap the shifted vars.
4286 (do_sd_constraint): Likewise.
4287 (do_ds_constraint): Likewise.
4288 (get_constraint_for_ptr_offset): Likewise.
4290 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4292 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4293 (nios2_compute_frame_layout):
4294 Add calculation of cfun->machine->fp_save_offset.
4295 (nios2_expand_prologue): Correct setting of frame pointer register
4297 (nios2_expand_epilogue): Update recovery of stack pointer from
4298 frame pointer accordingly.
4299 (nios2_initial_elimination_offset): Update calculation of offset
4300 for eliminating to HARD_FRAME_POINTER_REGNUM.
4302 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4305 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4306 cgraph_get_create_node on VAR_DECLs.
4308 2014-03-10 Richard Biener <rguenther@suse.de>
4311 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4313 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4315 * config/vms/vms.opt (vms_float_format): New variable.
4317 2014-03-08 Tobias Burnus <burnus@net-b.de>
4319 * doc/invoke.texi (-fcilkplus): Update implementation status.
4321 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4322 Richard Biener <rguenther@suse.de>
4324 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4325 consistently accross all TUs.
4326 (run_gcc): Enable -fshort-double automatically at link at link-time
4327 and disallow override.
4329 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4332 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4333 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4334 if they can't be used.
4336 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4338 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4339 for Solaris 11/x86 ld.
4340 * configure: Regenerate.
4342 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4344 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4345 (LIB_TLS_SPEC): Save as ld_tls_libs.
4346 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4347 (HAVE_AS_IX86_TLSLDM): New test.
4348 * configure, config.in: Regenerate.
4349 * config/i386/i386.c (legitimize_tls_address): Fall back to
4350 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4351 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4352 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4353 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4355 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4357 * common.opt (fira-loop-pressure): Mark as optimization.
4359 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4361 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4362 an OpenMP mappable type.
4364 2014-03-06 Matthias Klose <doko@ubuntu.com>
4366 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4367 MULTILIB_OSDIRNAMES is not defined.
4369 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4370 Meador Inge <meadori@codesourcery.com>
4373 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4374 (arm_legitimize_address): Call legitimize_tls_address for any
4375 arm_tls_referenced_p expression, handle constant addend. Call it
4376 before testing for !TARGET_ARM.
4377 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4379 2014-03-06 Richard Biener <rguenther@suse.de>
4385 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4387 * tree-streamer.c (record_common_node): Assert we don't record
4388 nodes with type double.
4389 (preload_common_node): Skip type double, complex double and double
4390 pointer since it is now frontend dependent due to fshort-double option.
4392 2014-03-06 Richard Biener <rguenther@suse.de>
4394 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4395 or -fno-lto is specified and the linker has full plugin support.
4396 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4397 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4398 * lto-wrapper.c (merge_and_complain): Merge compile-time
4399 optimization levels.
4400 (run_gcc): And pass it through to the link options.
4402 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4406 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4408 * cselib.c (remove_useless_values): Skip to avoid quadratic
4409 behavior if the condition moved from...
4410 (cselib_process_insn): ... here holds.
4412 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4415 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4416 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4419 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4420 (TM_H): Add x86-tune.def.
4422 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4424 * config/aarch64/aarch64.c (generic_tunings):
4425 Use cortexa57_extra_costs.
4427 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4430 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4431 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4432 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4435 2014-03-04 Heiher <r@hev.cc>
4437 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4438 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4440 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4442 * config/i386/predicates.md (const2356_operand): Change to ...
4443 (const2367_operand): ... this.
4444 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4446 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4447 (*avx512pf_scatterpf<mode>sf): Ditto.
4448 (avx512pf_scatterpf<mode>df): Ditto.
4449 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4450 (*avx512pf_scatterpf<mode>df): Ditto.
4451 * config/i386/i386.c (ix86_expand_builtin): Update
4452 incorrect hint operand error message.
4454 2014-03-04 Richard Biener <rguenther@suse.de>
4456 * lto-section-in.c (lto_get_section_data): Fix const cast.
4458 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4460 * tree-streamer.c (record_common_node): Assert we don't record
4461 nodes with type double.
4462 (preload_common_node): Skip type double, complex double and double
4463 pointer since it is now frontend dependent due to fshort-double option.
4465 2014-03-04 Richard Biener <rguenther@suse.de>
4468 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4469 (lto_input_toplevel_asms): Likewise.
4470 * lto-section-in.c (lto_get_section_data): Instead do it here
4473 2014-03-04 Richard Biener <rguenther@suse.de>
4475 PR tree-optimization/60382
4476 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4477 dead PHIs a reduction.
4479 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4481 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4483 (_mm_prefetch): Move out of GCC target("sse") pragma.
4484 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4485 GCC target("prfchw") pragma.
4486 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4488 * config/i386/i386.c (ix86_option_override_internal): Enable
4489 -mprfchw with -mprefetchwt1.
4491 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4493 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4496 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4498 * opts.h (CL_PCH_IGNORE): Define.
4499 * targhooks.c (option_affects_pch_p):
4500 Return false for options that have CL_PCH_IGNORE set.
4501 * opt-functions.awk: Process PchIgnore.
4502 * doc/options.texi: Document PchIgnore.
4504 * config/arc/arc.opt (misize): Add PchIgnore property.
4506 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4508 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4509 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4510 constraint on constants to permit them being loaded into
4511 GENERAL_REGS or BASE_REGS.
4513 2014-03-03 Nick Clifton <nickc@redhat.com>
4515 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4516 anti-cacnonical alternatives.
4517 (negandhi3_real): New pattern.
4518 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4520 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4522 * config/avr/avr-mcus.def: Remove atxmega16x1.
4523 * config/avr/avr-tables.opt: Regenerate.
4524 * config/avr/t-multilib: Regenerate.
4525 * doc/avr-mmcu.texi: Regenerate.
4527 2014-03-03 Tobias Grosser <tobias@grosser.es>
4528 Mircea Namolaru <mircea.namolaru@inria.fr>
4530 PR tree-optimization/58028
4531 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4534 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4536 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4537 not handled by recognizers.
4539 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4542 * function.c (expand_function_end): Don't emit
4543 clobber_return_register sequence if clobber_after is a BARRIER.
4544 * cfgexpand.c (construct_exit_block): Append instructions before
4545 return_label to prev_bb.
4547 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4549 * config/rs6000/constraints.md: Document reserved use of "wc".
4551 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4554 * ipa.c (function_and_variable_visibility): When dissolving comdat
4555 group, also set all symbols to local.
4557 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4562 2013-12-14 Jan Hubicka <jh@suse.cz>
4564 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4566 2014-03-02 Jon Beniston <jon@beniston.com>
4572 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4573 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4574 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4575 (simple_return, *simple_return): New patterns
4576 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4577 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4579 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4581 * dwarf2out.c (gen_subprogram_die): Tidy.
4583 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4586 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4587 (*mov_t_msb_neg_negc): ... this new insn.
4589 2014-02-28 Jason Merrill <jason@redhat.com>
4592 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4595 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4598 * dwarf2out.c (decltype_auto_die): New static.
4599 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4600 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4601 (is_cxx_auto): Likewise.
4603 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4605 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4606 we are not using general regs only.
4608 2014-02-28 Richard Biener <rguenther@suse.de>
4611 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4612 previous fix and only allow to remove trivial pre-headers
4613 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4614 (remove_forwarder_block): Properly update the latch of a loop.
4616 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4619 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4620 (cselib_preserved_hash_table): New.
4621 (preserve_constants_and_equivs): Move preserved vals to it.
4622 (cselib_find_slot): Look it up first.
4623 (cselib_init): Initialize it.
4624 (cselib_finish): Release it.
4625 (dump_cselib_table): Dump it.
4627 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4630 * cselib.c (remove_useless_values): Skip to avoid quadratic
4631 behavior if the condition moved from...
4632 (cselib_process_insn): ... here holds.
4634 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4637 * var-tracking.c (vt_initialize): Apply the same condition to
4638 preserve the CFA base value.
4640 2014-02-28 Joey Ye <joey.ye@arm.com>
4643 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4644 if reload in progress or completed.
4646 2014-02-28 Tobias Burnus <burnus@net-b.de>
4649 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4652 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4654 * doc/tm.texi.in (Condition Code Status): Update documention for
4655 relative locations of cc0-setter and cc0-user.
4657 2014-02-27 Jeff Law <law@redhat.com>
4659 PR rtl-optimization/52714
4660 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4661 into two independent simple sets, if I3 is a jump, ensure the
4662 pattern we place into I3 is a (set (pc) ...).
4664 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4665 Jeff Law <law@redhat.com>
4667 PR rtl-optimization/49847
4668 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4669 are in different blocks.
4670 * doc/tm.texi (Condition Code Status): Update documention for
4671 relative locations of cc0-setter and cc0-user.
4673 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4676 * lra.c (lra_emit_add): Check SUBREG too.
4678 2014-02-27 Andreas Schwab <schwab@suse.de>
4680 * config/m68k/m68k.c (m68k_option_override): Disable
4681 -flive-range-shrinkage for classic m68k.
4682 (m68k_override_options_after_change): Likewise.
4684 2014-02-27 Marek Polacek <polacek@redhat.com>
4687 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4688 -Wmaybe-uninitialized.
4690 2014-02-27 Alan Modra <amodra@gmail.com>
4693 * reload1.c (emit_input_reload_insns): When reload_override_in,
4694 set old to rl->in_reg when rl->in_reg is a subreg.
4696 2014-02-26 Richard Biener <rguenther@suse.de>
4699 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4701 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4703 * common/config/i386/predicates.md (const1256_operand): Remove.
4704 (const2356_operand): New.
4705 (const_1_to_2_operand): Remove.
4706 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4707 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4708 (*avx512pf_gatherpf<mode>sf): Ditto.
4709 (avx512pf_gatherpf<mode>df): Ditto.
4710 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4711 (*avx512pf_gatherpf<mode>df): Ditto.
4712 (avx512pf_scatterpf<mode>sf): Ditto.
4713 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4714 (*avx512pf_scatterpf<mode>sf): Ditto.
4715 (avx512pf_scatterpf<mode>df): Ditto.
4716 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4717 (*avx512pf_scatterpf<mode>df): Ditto.
4718 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4720 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4722 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4723 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4724 (_mm512_mask_testn_epi64_mask): Move to ...
4725 * config/i386/avx512cdintrin.h: Here.
4726 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4727 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4728 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4729 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4730 TARGET_AVX512F from TARGET_AVX512CD.
4732 2014-02-26 Richard Biener <rguenther@suse.de>
4735 * ipa.c (walk_polymorphic_call_targets): Properly guard
4736 call to inline_update_overall_summary.
4738 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4741 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4742 and latches only if requested. Fix latch if it is removed.
4743 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4744 LOOPS_HAVE_PREHEADERS.
4746 2014-02-25 Andrew Pinski <apinski@cavium.com>
4748 * builtins.c (expand_builtin_thread_pointer): Create a new target
4749 when the target is NULL.
4751 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4753 PR rtl-optimization/60317
4754 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4755 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4756 * lra-assigns.c: Include params.h.
4757 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4758 other reload pseudos considerations.
4760 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4762 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4763 to use canonical form for nor<mode>3.
4765 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4768 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4771 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4773 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4774 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4775 (ix86_handle_option): Handle OPT_mprefetchwt1.
4776 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4777 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4779 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4780 OPTION_MASK_ISA_PREFETCHWT1.
4781 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4782 (PTA_PREFETCHWT1): New.
4783 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4784 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4785 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4786 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4787 (*prefetch_avx512pf_<mode>_: Change into ...
4788 (*prefetch_prefetchwt1_<mode>: This.
4789 * config/i386/i386.opt (mprefetchwt1): New.
4790 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4791 (_mm_prefetch): Handle intent to write.
4792 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4794 2014-02-25 Richard Biener <rguenther@suse.de>
4797 * emit-rtl.c (mem_attrs_htab): Remove.
4798 (mem_attrs_htab_hash): Likewise.
4799 (mem_attrs_htab_eq): Likewise.
4800 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4801 (init_emit_once): Do not allocate mem_attrs_htab.
4803 2014-02-25 Richard Biener <rguenther@suse.de>
4806 * lto-opts.c (lto_write_options): Output non-explicit conservative
4807 -fwrapv, -fno-trapv and -fno-strict-overflow.
4808 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4809 (run_gcc): And pass them through.
4811 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4813 * sel-sched.c (calculate_new_fences): New parameter ptime.
4814 Calculate it as a maximum over all fence cycles.
4815 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4816 Print the final schedule timing when sched_verbose.
4818 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4820 PR rtl-optimization/60292
4821 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4822 bit fot the fence instruction.
4824 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4826 * calls.h: Fix typo in comment.
4828 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4830 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4831 adjusting offsetable addresses.
4833 2014-02-24 Guozhi Wei <carrot@google.com>
4835 * sparseset.h (sparseset_pop): Fix the wrong index.
4837 2014-02-24 Walter Lee <walt@tilera.com>
4839 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4840 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4842 * common/config/tilegx/tilegx-common.c
4843 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4844 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4846 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4847 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4848 (tilegx_gimplify_va_arg_expr): Handle big endian.
4849 (tilegx_expand_unaligned_load): Ditto.
4850 (tilegx_expand_unaligned_store): Ditto.
4851 (TARGET_RETURN_IN_MSB): New.
4852 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4853 (TARGET_ENDIAN_DEFAULT): New.
4854 (TARGET_BIG_ENDIAN): Handle big endian.
4855 (BYTES_BIG_ENDIAN): Ditto.
4856 (WORDS_BIG_ENDIAN): Ditto.
4857 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4860 * config/tilegx/tilegx.md (extv): Handle big endian.
4862 (insn_st<n>): Ditto.
4863 (insn_st<n>_add<bitsuffix>): Ditto.
4864 (insn_stnt<n>): Ditto.
4865 (insn_stnt<n>_add<bitsuffix>):Ditto.
4866 (vec_interleave_highv8qi): Handle big endian.
4867 (vec_interleave_highv8qi_be): New.
4868 (vec_interleave_highv8qi_le): New.
4869 (insn_v1int_h): Handle big endian.
4870 (vec_interleave_lowv8qi): Handle big endian.
4871 (vec_interleave_lowv8qi_be): New.
4872 (vec_interleave_lowv8qi_le): New.
4873 (insn_v1int_l): Handle big endian.
4874 (vec_interleave_highv4hi): Handle big endian.
4875 (vec_interleave_highv4hi_be): New.
4876 (vec_interleave_highv4hi_le): New.
4877 (insn_v2int_h): Handle big endian.
4878 (vec_interleave_lowv4hi): Handle big endian.
4879 (vec_interleave_lowv4hi_be): New.
4880 (vec_interleave_lowv4hi_le): New.
4881 (insn_v2int_l): Handle big endian.
4882 (vec_interleave_highv2si): Handle big endian.
4883 (vec_interleave_highv2si_be): New.
4884 (vec_interleave_highv2si_le): New.
4885 (insn_v4int_h): Handle big endian.
4886 (vec_interleave_lowv2si): Handle big endian.
4887 (vec_interleave_lowv2si_be): New.
4888 (vec_interleave_lowv2si_le): New.
4889 (insn_v4int_l): Handle big endian.
4890 * config/tilegx/tilegx.opt (mbig-endian): New option.
4891 (mlittle-endian): New option.
4892 * doc/install.texi: Document tilegxbe-linux.
4893 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4895 2014-02-24 Martin Jambor <mjambor@suse.cz>
4898 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4899 there are no parameter descriptors.
4901 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4903 PR rtl-optimization/60268
4904 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4905 initialization to ...
4906 (sched_rgn_init): ... here.
4907 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4909 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4911 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4914 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4916 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4919 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4921 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4922 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4924 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4926 * config/microblaze/predicates.md: Add cmp_op predicate.
4927 * config/microblaze/microblaze.md: Add branch_compare instruction
4928 which uses cmp_op predicate and emits cmp insn before branch.
4929 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4930 to microblaze_expand_conditional_branch and consolidate logic.
4931 (microblaze_expand_conditional_branch): emit branch_compare
4932 insn instead of handling cmp op separate from branch insn.
4934 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4936 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4939 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4941 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4942 define_insn with define_expand and new define_insn
4943 *altivec_lve<VI_char>x_internal.
4944 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4945 and new define_insn *altivec_stve<VI_char>x_internal.
4946 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4948 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4950 (altivec_expand_stvex_be): New function.
4952 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4954 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4955 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4956 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4957 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4959 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4962 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4963 instead of emit_move_insn.
4965 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4967 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4969 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4970 gen_altivec_vsumsws.
4972 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4974 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4975 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4976 (altivec_lvxl_<mode>): New define_expand incorporating
4977 -maltivec=be semantics where needed.
4978 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4979 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4980 semantics where needed.
4981 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4982 (altivec_stvx_<mode>): New define_expand incorporating
4983 -maltivec=be semantics where needed.
4984 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4985 VM2 iterator instead of V4SI.
4986 (altivec_stvxl_<mode>): New define_expand incorporating
4987 -maltivec=be semantics where needed.
4988 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4989 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4990 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4991 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4992 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4993 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4994 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4995 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4996 ALTIVEC_BUILTIN_STVXL.
4997 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4998 (altivec_expand_stvx_be): Likewise.
4999 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
5000 (altivec_expand_lvx_be): Likewise.
5001 (altivec_expand_stvx_be): Likewise.
5002 (altivec_expand_builtin): Add cases for
5003 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
5004 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
5005 (altivec_init_builtins): Add definitions for
5006 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
5007 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
5009 2014-02-21 Catherine Moore <clm@codesourcery.com>
5011 * doc/invoke.texi (mvirt, mno-virt): Document.
5012 * config/mips/mips.opt (mvirt): New option.
5013 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
5015 2014-02-21 Richard Biener <rguenther@suse.de>
5017 PR tree-optimization/60276
5018 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
5019 (STMT_VINFO_MIN_NEG_DIST): New macro.
5020 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
5021 STMT_VINFO_MIN_NEG_DIST.
5022 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
5023 made for negative dependence distances still hold.
5025 2014-02-21 Richard Biener <rguenther@suse.de>
5028 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
5029 DECL_INITIAL for globals not in the current function context.
5031 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5033 PR tree-optimization/56490
5034 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
5035 * tree-ssa-uninit.c: Include params.h.
5036 (compute_control_dep_chain): Add num_calls argument, return false
5037 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
5038 num_calls to recursive call.
5039 (find_predicates): Change dep_chain into normal array,
5040 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
5041 variable and adjust compute_control_dep_chain caller.
5042 (find_def_preds): Likewise.
5044 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
5046 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
5047 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
5049 2014-02-21 Nick Clifton <nickc@redhat.com>
5051 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
5052 (pushhi1): Likewise.
5053 (popqi1): Add mode to pre_dec.
5056 2014-02-21 Jakub Jelinek <jakub@redhat.com>
5058 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
5059 mode for mask of V8SFmode permutation.
5061 2014-02-20 Richard Henderson <rth@redhat.com>
5064 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
5065 a new pseudo for OLDVAL.
5067 2014-02-20 Jakub Jelinek <jakub@redhat.com>
5070 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
5071 gen_reg_rtx if d->testing_p.
5072 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
5073 if d->testing_p and we will certainly return true.
5074 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
5077 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
5079 * emit-rtl.c (gen_reg_rtx): Assert that
5080 crtl->emit.regno_pointer_align_length is non-zero.
5082 2014-02-20 Richard Henderson <rth@redhat.com>
5085 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
5086 on failure the store back into EXPECT.
5088 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
5089 Sandra Loosemore <sandra@codesourcery.com>
5091 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5092 * config/nios2/nios2.c (nios2_function_profiler): Add
5093 -fPIC (flag_pic == 2) support.
5094 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5095 (nios2_large_offset_p): New function.
5096 (nios2_unspec_reloc_p): Move up position, update to use
5097 nios2_large_offset_p.
5098 (nios2_unspec_address): Remove function.
5099 (nios2_unspec_offset): New function.
5100 (nios2_large_got_address): New function.
5101 (nios2_got_address): Add large offset support.
5102 (nios2_legitimize_tls_address): Update usage of removed and new
5104 (nios2_symbol_binds_local_p): New function.
5105 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5106 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5107 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5108 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5109 processing for (const (unspec ...)).
5110 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5112 2014-02-20 Richard Biener <rguenther@suse.de>
5114 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5115 doing the substitution.
5116 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5118 2014-02-20 Martin Jambor <mjambor@suse.cz>
5121 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5122 info when checking whether lattices are bottom.
5124 2014-02-20 Richard Biener <rguenther@suse.de>
5127 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5130 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5133 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5134 parameter specifying the scaling.
5135 (inline_call): Update.
5136 (want_inline_recursively): Guard division by zero.
5137 (recursive_inlining): Update.
5138 * ipa-inline.h (clone_inlined_nodes): Update.
5140 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5143 * config/i386/i386.c (classify_argument): Pass structures of size
5144 64 bytes or less in register.
5146 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5147 Kirill Yukhin <kirill.yukhin@intel.com>
5149 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5150 (_mm_rcp28_round_ss): Ditto.
5151 (_mm_rsqrt28_round_sd): Ditto.
5152 (_mm_rsqrt28_round_ss): Ditto.
5153 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5154 (_mm_rcp14_round_ss): Ditto.
5155 (_mm_rsqrt14_round_sd): Ditto.
5156 (_mm_rsqrt14_round_ss): Ditto.
5157 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5158 the first input operand, get rid of match_dup.
5159 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5161 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5163 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5164 operand as the first input operand, set type attribute.
5165 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5167 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5168 operand as the first input operand, set type attribute.
5170 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5172 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5175 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5178 * config/i386/i386.c (construct_container): Remove TFmode check
5179 for X86_64_INTEGER_CLASS.
5181 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5184 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5185 only when -Wpsabi is enabled.
5187 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5190 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5191 passing arrays in registers are the same as for structs, so remove the
5192 special case for them.
5194 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5196 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5197 destination type, extract only the valid bits if the source type is not
5198 integral and has a different mode.
5200 2014-02-19 Richard Biener <rguenther@suse.de>
5203 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5206 2014-02-19 Richard Biener <rguenther@suse.de>
5209 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5210 (ipa_modify_call_arguments): Emit an argument load explicitely and
5211 preserve virtual SSA form there and for the replacement call.
5212 Do not update SSA form nor free dominance info.
5214 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5216 * ipa.c (function_and_variable_visibility): Also clear WEAK
5217 flag when disolving COMDAT_GROUP.
5219 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5221 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5222 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5224 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5225 do more sanity checks.
5226 (detect_type_change): Return true when giving up early.
5227 (compute_complex_assign_jump_func): Fix type parameter of
5228 ipa_set_ancestor_jf.
5229 (compute_complex_ancestor_jump_func): Likewise.
5230 (update_jump_functions_after_inlining): Fix updating of
5232 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5234 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5236 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5237 inline clones when edge disappears.
5239 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5242 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5243 Split 64-bit moves into 2 patterns. Do not allow the use of
5244 direct move for TDmode in little endian, since the decimal value
5245 has little endian bytes within a word, but the 64-bit pieces are
5246 ordered in a big endian fashion, and normal subreg's of TDmode are
5248 (mov<mode>_64bit_dm): Likewise.
5249 (movtd_64bit_nodm): Likewise.
5251 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5253 PR tree-optimization/60174
5254 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5255 statement of an SSA_NAME that occurs in an abnormal PHI node.
5257 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5260 * final.c (SEEN_BB): Remove.
5261 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5262 (final_scan_insn): Don't force_source_line on second
5263 NOTE_INSN_BASIC_BLOCK.
5265 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5268 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5269 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5270 (type_natural_mode): Warn ABI change when %zmm register is not
5271 available for AVX512F vector value passing.
5273 2014-02-18 Kai Tietz <ktietz@redhat.com>
5276 * config/i386/i386.c (ix86_expand_prologue): Use value in
5277 rax register as displacement when restoring %r10 or %rax.
5278 Fix wrong offset when restoring both registers.
5280 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5282 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5283 assertion with conditional return.
5285 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5286 Uros Bizjak <ubizjak@gmail.com>
5289 * config/i386/driver-i386.c (host_detect_local_cpu): If
5290 YMM state is not saved by the OS, also clear has_f16c. Move
5291 CPUID 0x80000001 handling before YMM state saving checking.
5293 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5295 PR rtl-optimization/58960
5296 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5297 factored out from ...
5298 (sched_init): ... here.
5299 (free_global_sched_pressure_data): New, factored out from ...
5300 (sched_finish): ... here.
5301 * sched-int.h (free_global_sched_pressure_data): Declare.
5302 * sched-rgn.c (nr_regions_initial): New static global.
5303 (haifa_find_rgns): Initialize it.
5304 (schedule_region): Disable sched-pressure for the newly
5307 2014-02-17 Richard Biener <rguenther@suse.de>
5309 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5310 release SSA defs of pattern stmts.
5312 2014-02-17 Richard Biener <rguenther@suse.de>
5314 * tree-inline.c (expand_call_inline): Release the virtual
5315 operand defined by the call we are about to inline.
5317 2014-02-17 Richard Biener <rguenther@suse.de>
5319 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5321 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5322 Ilya Tocar <ilya.tocar@intel.com>
5324 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5325 arguments order in builtin.
5326 (_mm512_permutexvar_epi64): Ditto.
5327 (_mm512_mask_permutexvar_epi64): Ditto
5328 (_mm512_maskz_permutexvar_epi32): Ditto
5329 (_mm512_permutexvar_epi32): Ditto
5330 (_mm512_mask_permutexvar_epi32): Ditto
5332 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5334 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5335 (p8_vmrgow): Likewise.
5337 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5339 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5342 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5345 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5346 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5347 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5348 using direct move instructions on ISA 2.07. Also adjust
5349 instruction length for 64-bit.
5350 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5351 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5353 2014-02-15 Alan Modra <amodra@gmail.com>
5357 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5358 find_replacement on parts of insn rtl that might be reloaded.
5360 2014-02-15 Richard Biener <rguenther@suse.de>
5362 PR tree-optimization/60183
5363 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5364 (tree_ssa_phiprop): Calculate and free post-dominators.
5366 2014-02-14 Jeff Law <law@redhat.com>
5368 PR rtl-optimization/60131
5369 * ree.c (get_extended_src_reg): New function.
5370 (combine_reaching_defs): Use it rather than assuming location of REG.
5371 (find_and_remove_re): Verify first operand of extension is
5372 a REG before adding the insns to the copy list.
5374 2014-02-14 Roland McGrath <mcgrathr@google.com>
5376 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5377 * configure: Regenerated.
5378 * config.in: Regenerated.
5379 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5380 instead of ASM_SHORT.
5382 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5383 Richard Earnshaw <rearnsha@arm.com>
5385 PR rtl-optimization/59535
5386 * lra-constraints.c (process_alt_operands): Encourage alternative
5387 when unassigned pseudo class is superset of the alternative class.
5388 (inherit_reload_reg): Don't inherit when optimizing for code size.
5389 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5390 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5391 modes not less than 4 for Thumb1.
5393 2014-02-14 Kyle McMartin <kyle@redhat.com>
5396 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5398 2014-02-14 Richard Biener <rguenther@suse.de>
5400 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5401 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5402 Do not strip INDIRECT_REFs.
5404 2014-02-14 Richard Biener <rguenther@suse.de>
5407 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5408 DECL_FUNCTION_SPECIFIC_TARGET.
5409 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5410 * tree-streamer-out.c (pack_ts_target_option): Remove.
5411 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5412 (write_ts_function_decl_tree_pointers): Do not stream
5413 DECL_FUNCTION_SPECIFIC_TARGET.
5414 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5415 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5416 (lto_input_ts_function_decl_tree_pointers): Do not stream
5417 DECL_FUNCTION_SPECIFIC_TARGET.
5419 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5421 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5422 (get_initial_def_for_induction, vectorizable_induction): Ignore
5423 debug stmts when looking for exit_phi.
5424 (vectorizable_live_operation): Fix up condition.
5426 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5428 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5429 nreverse() because it changes the content of original tree list.
5431 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5433 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5434 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5436 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5438 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5439 GNU coding standards.
5441 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5444 * dwarf2out.c (gen_subprogram_die): Don't call
5445 add_calling_convention_attribute if subr_die is old_die.
5447 2014-02-13 Sharad Singhai <singhai@google.com>
5449 * doc/optinfo.texi: Fix order of nodes.
5451 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5453 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5454 operands[2], not operands[3].
5456 2014-02-13 Richard Biener <rguenther@suse.de>
5459 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5460 mention the possibility of an in-tree build.
5461 (CLooG): Update recommended version to 0.18.1, mention the
5462 possibility of an in-tree build and clarify that the ISL
5463 bundled with CLooG does not work.
5465 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5468 * expr.c (compress_float_constant): If x is a hard register,
5469 extend into a pseudo and then move to x.
5471 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5473 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5474 caused by bad second argument to warning_at() with -mhotpatch and
5475 nested functions (e.g. with gfortran).
5477 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5479 * opts.c (option_name): Remove "enabled by default" rider.
5481 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5483 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5485 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5486 Uros Bizjak <ubizjak@gmail.com>
5489 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5490 * configure: Regenerated.
5492 2014-02-12 Richard Biener <rguenther@suse.de>
5494 * vec.c (vec_prefix::calculate_allocation): Move as
5495 inline variant to vec.h.
5496 (vec_prefix::calculate_allocation_1): New out-of-line version.
5497 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5498 (vec_prefix::m_has_auto_buf): Rename to ...
5499 (vec_prefix::m_using_auto_storage): ... this.
5500 (vec_prefix::calculate_allocation): Inline the easy cases
5501 and dispatch to calculate_allocation_1 which doesn't need the
5503 (va_heap::reserve): Use gcc_checking_assert.
5504 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5505 m_using_auto_storage.
5506 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5508 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5509 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5510 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5512 2014-02-12 Richard Biener <rguenther@suse.de>
5514 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5515 when we found a dependence.
5517 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5519 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5521 (maybe_fold_stmt): ... into this new function.
5522 * omp-low.c (lower_omp): Update comment.
5524 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5527 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5530 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5532 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5533 identifiers in comments.
5534 (cortexa53_extra_costs): Likewise.
5535 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5536 (cortexa7_extra_costs): Likewise.
5537 (cortexa12_extra_costs): Likewise.
5538 (cortexa15_extra_costs): Likewise.
5539 (v7m_extra_costs): Likewise.
5541 2014-02-12 Richard Biener <rguenther@suse.de>
5544 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5545 of posix_memalign being successful.
5546 (lower_stmt): Restrict lowering of posix_memalign to when
5547 -ftree-bit-ccp is enabled.
5549 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5551 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5553 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5555 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5557 PR rtl-optimization/60116
5558 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5559 other_insn once the combination has been validated.
5561 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5564 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5566 * ipa-devirt.c: Include demangle.h
5567 (odr_violation_reported): New static variable.
5568 (add_type_duplicate): Update odr_violations.
5569 (maybe_record_node): Add completep parameter; update it.
5570 (record_target_from_binfo): Add COMPLETEP parameter;
5571 update it as needed.
5572 (possible_polymorphic_call_targets_1): Likewise.
5573 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5574 rename FINAL to COMPLETE.
5575 (record_targets_from_bases): Sanity check we found the binfo;
5576 fix COMPLETEP updating.
5577 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5578 parameter, fix computing of COMPLETEP.
5579 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5580 at LTO time do demangling.
5581 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5582 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5584 (gimple_get_virt_method_for_binfo): Likewise.
5585 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5586 gimple_get_virt_method_for_vtable): Update prototypes.
5588 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5591 * genautomata.c (add_presence_absence): Fix typo with
5592 {final_}presence_list.
5594 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5597 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5598 for VSX/Altivec vectors that land in GPR registers.
5600 2014-02-11 Richard Henderson <rth@redhat.com>
5601 Jakub Jelinek <jakub@redhat.com>
5604 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5605 around drhs if type conversion to lacc->type is not useless.
5607 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5609 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5611 (cortex-a57.cortex-a53): Likewise.
5612 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5614 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5616 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5619 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5621 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5622 add_options_for_arm_vfp3.
5624 2014-02-11 Jeff Law <law@redhat.com>
5627 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5628 object with an undesirable mode.
5630 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5633 * config/i386/sol2-9.h: New file.
5634 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5635 *-*-solaris2.9*): Use it.
5637 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5639 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5640 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5642 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5644 * config/microblaze/microblaze.c: Extend mcpu version format
5646 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5648 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5650 2014-02-10 Richard Henderson <rth@redhat.com>
5653 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5654 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5655 ms-abi vs -mno-accumulate-outgoing-args.
5656 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5657 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5660 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5663 * cfgexpand.c (expand_asm_operands): Attach source location to
5664 ASM_INPUT rtx objects.
5665 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5667 2014-02-10 Nick Clifton <nickc@redhat.com>
5669 * config/mn10300/mn10300.c (popcount): New function.
5670 (mn10300_expand_prologue): Include saved registers in stack usage
5673 2014-02-10 Jeff Law <law@redhat.com>
5676 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5677 when changing the SET_DEST of a prior insn to avoid an input reload.
5679 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5681 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5682 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5683 -mcall-openbsd, or -mcall-linux.
5684 (CC1_ENDIAN_BIG_SPEC): Remove.
5685 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5686 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5687 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5688 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5689 and %cc1_endian_default.
5690 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5692 2014-02-10 Richard Biener <rguenther@suse.de>
5694 PR tree-optimization/60115
5695 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5696 MEM_REF handling. Properly verify that the accesses are not
5697 out of the objects bound.
5699 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5701 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5704 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5706 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5707 proper constants and fix formatting.
5708 (possible_polymorphic_call_targets): Fix formatting.
5710 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5711 Ilya Tocar <ilya.tocar@intel.com>
5713 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5714 (_mm512_loadu_epi32): Renamed into...
5715 (_mm512_loadu_si512): This.
5716 (_mm512_storeu_epi32): Renamed into...
5717 (_mm512_storeu_si512): This.
5718 (_mm512_maskz_ceil_ps): Removed.
5719 (_mm512_maskz_ceil_pd): Ditto.
5720 (_mm512_maskz_floor_ps): Ditto.
5721 (_mm512_maskz_floor_pd): Ditto.
5722 (_mm512_floor_round_ps): Ditto.
5723 (_mm512_floor_round_pd): Ditto.
5724 (_mm512_ceil_round_ps): Ditto.
5725 (_mm512_ceil_round_pd): Ditto.
5726 (_mm512_mask_floor_round_ps): Ditto.
5727 (_mm512_mask_floor_round_pd): Ditto.
5728 (_mm512_mask_ceil_round_ps): Ditto.
5729 (_mm512_mask_ceil_round_pd): Ditto.
5730 (_mm512_maskz_floor_round_ps): Ditto.
5731 (_mm512_maskz_floor_round_pd): Ditto.
5732 (_mm512_maskz_ceil_round_ps): Ditto.
5733 (_mm512_maskz_ceil_round_pd): Ditto.
5734 (_mm512_expand_pd): Ditto.
5735 (_mm512_expand_ps): Ditto.
5736 * config/i386/i386.c (ix86_builtins): Remove
5737 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5738 (bdesc_args): Ditto.
5739 * config/i386/predicates.md (const1256_operand): New.
5740 (const_1_to_2_operand): Ditto.
5741 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5742 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5743 (*avx512pf_gatherpf<mode>sf): Ditto.
5744 (avx512pf_gatherpf<mode>df): Ditto.
5745 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5746 (*avx512pf_gatherpf<mode>df): Ditto.
5747 (avx512pf_scatterpf<mode>sf): Ditto.
5748 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5749 (*avx512pf_scatterpf<mode>sf): Ditto.
5750 (avx512pf_scatterpf<mode>df): Ditto.
5751 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5752 (*avx512pf_scatterpf<mode>df): Ditto.
5753 (avx512f_expand<mode>): Removed.
5754 (<shift_insn><mode>3<mask_name>): Change predicate type.
5756 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5758 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5759 not at the end of datarefs vector use ordered_remove to avoid
5760 reordering datarefs vector.
5763 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5764 mark local addressable non-static vars as GOVD_PRIVATE
5765 instead of GOVD_LOCAL.
5766 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5767 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5771 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5772 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5773 assume_aligned or alloc_align attributes.
5774 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5775 arguments. Handle also assume_aligned and alloc_align attributes.
5776 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5777 calls to functions with assume_aligned or alloc_align attributes.
5778 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5780 2014-02-08 Terry Guo <terry.guo@arm.com>
5782 * doc/invoke.texi: Document ARM -march=armv7e-m.
5784 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5786 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5787 flag on __cilkrts_rethrow builtin.
5790 * ipa-cp.c (determine_versionability): Fail at -O0
5791 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5792 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5795 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5798 * tree-inline.c (copy_forbidden): Fail for
5799 __attribute__((optimize (0))) functions.
5801 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5803 * varpool.c: Include pointer-set.h.
5804 (varpool_remove_unreferenced_decls): Variables in other partitions
5805 will not be output; be however careful to not lose information
5808 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5810 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5811 lookup in the vtable constructor.
5813 2014-02-07 Jeff Law <law@redhat.com>
5816 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5817 define_insn_and_split.
5819 * ipa-inline.c (inline_small_functions): Fix typos.
5821 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5823 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5824 (s390_can_use_return_insn): Declare.
5825 * config/s390/s390.h (EPILOGUE_USES): Define.
5826 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5828 (s390_chunkify_start): Handle return JUMP_LABELs.
5829 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5830 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5831 (s390_can_use_return_insn): New functions.
5832 (s390_fix_long_loop_prediction): Handle conditional returns.
5833 (TARGET_SET_UP_BY_PROLOGUE): Define.
5834 * config/s390/s390.md (ANY_RETURN): New code iterator.
5835 (*creturn, *csimple_return, return, simple_return): New patterns.
5837 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5839 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5840 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5841 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5842 REG_CFA_RESTORE list when deciding not to restore a register.
5844 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5846 * config/s390/s390.c: Include tree-pass.h and context.h.
5847 (s390_early_mach): New function, split out from...
5848 (s390_emit_prologue): ...here.
5849 (pass_data_s390_early_mach): New pass structure.
5850 (pass_s390_early_mach): New class.
5851 (s390_option_override): Create and register early_mach pass.
5852 Move to end of file.
5854 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5856 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5857 to match for the exit block.
5859 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5861 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5862 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5863 Reject misaligned operands.
5865 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5867 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5869 2014-02-07 Richard Biener <rguenther@suse.de>
5872 * gimple-low.c (lower_builtin_posix_memalign): New function.
5873 (lower_stmt): Call it to lower posix_memalign in a way
5874 to make alignment info accessible.
5876 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5879 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5880 __builtin_setjmp_receiver.
5882 2014-02-07 Richard Biener <rguenther@suse.de>
5885 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5886 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5887 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5888 Handle BUILT_IN_POSIX_MEMALIGN.
5889 (find_func_clobbers): Likewise.
5890 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5891 (call_may_clobber_ref_p_1): Likewise.
5893 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5896 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5899 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5902 * lto-cgraph.c (lto_output_node): Use
5903 symtab_get_symbol_partitioning_class.
5904 (lto_output_varpool_node): likewise.
5905 (symtab_get_symbol_partitioning_class): Move here from
5907 * cgraph.h (symbol_partitioning_class): Likewise.
5908 (symtab_get_symbol_partitioning_class): Declare.
5910 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5912 * ggc.h (ggc_internal_cleared_alloc): New macro.
5913 * vec.h (vec_safe_copy): Handle memory stats.
5914 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5915 * target-globals.c (save_target_globals): Likewise.
5917 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5920 * expr.c (emit_move_resolve_push): Export; be bit more selective
5921 on when to clear alias set.
5922 * expr.h (emit_move_resolve_push): Declare.
5923 * function.h (struct function): Add tail_call_marked.
5924 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5925 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5926 * config/i386/i386.md (TImode move expander): De not call
5928 (FP push expanders): Preserve memory attributes.
5929 * config/i386/sse.md (push<mode>1): Remove.
5930 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5931 (ix86_expand_push): Remove.
5932 * config/i386/mmx.md (push<mode>1): Remove.
5934 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5936 PR rtl-optimization/60030
5937 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5938 lopart with paradoxical subreg before shifting it up by hprec.
5940 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5942 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5943 Remove extra newline at end of file.
5944 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5945 (arm_issue_rate): Handle cortexa57.
5946 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5947 (cortex-a57.cortex-a53): Likewise.
5949 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5952 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5953 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5955 (arm_expand_prologue): Adjust all callers.
5956 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5957 info, registers also at the lowest numbered registers side. Use
5958 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5962 * var-tracking.c (adjust_mems): Before adding a SET to
5963 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5965 2014-02-06 Alan Modra <amodra@gmail.com>
5968 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5969 change SDmode to DDmode when lra_in_progress.
5971 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5974 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5975 free_data_ref on the dr first, and before goto again also set dr
5976 to the next dr. For simd_lane_access, free old datarefs[i] before
5977 overwriting it. For get_vectype_for_scalar_type failure, don't
5978 free_data_ref if simd_lane_access.
5980 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5983 * tree.h (opts_for_fn): New inline function.
5984 (opt_for_fn): Define.
5985 * config/i386/i386.c (ix86_function_regparm): Use
5986 opt_for_fn (decl, optimize) instead of optimize.
5988 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5990 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5991 for SYMBOL_REF in large memory model.
5993 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5995 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5997 (cortex-a57): Likewise.
5998 (cortex-a57.cortex-a53): Likewise.
6000 2014-02-06 Yury Gribov <y.gribov@samsung.com>
6001 Kugan Vivekanandarajah <kuganv@linaro.org>
6003 * config/arm/arm.c (arm_vector_alignment_reachable): Check
6005 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
6007 2014-02-06 Richard Biener <rguenther@suse.de>
6009 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
6010 set_loop_copy and initialize_original_copy_tables.
6012 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
6014 * config/aarch64/aarch64-simd.md
6015 (aarch64_ashr_simddi): Change QI to SI.
6017 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6018 Jakub Jelinek <jakub@redhat.com>
6021 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
6024 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6026 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
6027 CODE_FOR_altivec_vpku[hw]um to
6028 CODE_FOR_altivec_vpku[hw]um_direct.
6029 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
6030 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
6031 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
6032 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
6034 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6036 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
6037 generation for -maltivec=be.
6038 (altivec_vsumsws): Simplify redundant test.
6040 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6042 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
6043 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
6044 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
6045 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
6046 gen_altivec_vpkuwum.
6047 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
6049 (altivec_vpks<VI_char>ss): Likewise.
6050 (altivec_vpks<VI_char>us): Likewise.
6051 (altivec_vpku<VI_char>us): Likewise.
6052 (altivec_vpku<VI_char>um): Likewise.
6053 (altivec_vpku<VI_char>um_direct): New (copy of
6054 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
6056 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
6057 target is little endian and -maltivec=be is not specified.
6058 (*altivec_vupkhs<VU_char>_direct): New (copy of
6059 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
6060 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
6061 target is little endian and -maltivec=be is not specified.
6062 (*altivec_vupkls<VU_char>_direct): New (copy of
6063 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
6064 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
6065 little endian and -maltivec=be is not specified.
6066 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
6067 little endian and -maltivec=be is not specified.
6069 2014-02-05 Richard Henderson <rth@redhat.com>
6072 * combine-stack-adj.c: Revert r206943.
6073 * sched-int.h (struct deps_desc): Add last_args_size.
6074 * sched-deps.c (init_deps): Initialize it.
6075 (sched_analyze_insn): Add OUTPUT dependencies between insns that
6076 contain REG_ARGS_SIZE notes.
6078 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
6080 * lto-cgraph.c (asm_nodes_output): Make global.
6081 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
6082 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
6083 (driver_handle_option): Handle OPT_fwpa.
6085 2014-02-05 Jakub Jelinek <jakub@redhat.com>
6088 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
6089 a comment typo and formatting issue. If odr_hash hasn't been
6090 created, return vNULL and set *completep to false.
6093 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6094 bb with no successors.
6096 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6099 * doc/invoke.texi (-march): Clarify documentation for ARM.
6103 2014-02-05 Richard Biener <rguenther@suse.de>
6105 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6106 when not vectorizing because of too many alias checks.
6107 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6108 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6110 2014-02-05 Nick Clifton <nickc@redhat.com>
6112 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6113 accept extended registers in any mode when compiling for the MN10300.
6115 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6117 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6118 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6119 sanitization attributes.
6120 (can_inline_edge_p): Likewise.
6121 (sanitize_attrs_match_for_inline_p): New function.
6123 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6125 * ipa-prop.c (detect_type_change): Shor circuit testing of
6126 type changes on THIS pointer.
6128 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6131 * config/pa/pa.c (legitimize_tls_address): Return original address
6132 if not passed a SYMBOL_REF rtx.
6133 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6135 (pa_emit_move_sequence): Simplify TLS source operands.
6136 (pa_legitimate_constant_p): Reject all TLS constants.
6137 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6138 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6140 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6142 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6143 groups when we know they are controlled by LTO.
6144 * varasm.c (default_binds_local_p_1): If object is in other partition,
6145 it will be resolved locally.
6147 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6149 * config/host-linux.c (linux_gt_pch_use_address): Don't
6150 use SSIZE_MAX because it is not always defined.
6152 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6155 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6156 threshold for pseudo splitting.
6157 (update_ebb_live_info): Process call argument hard registers and
6158 hard registers from insn definition too.
6159 (max_small_class_regs_num): New constant.
6160 (inherit_in_ebb): Update live hard regs through EBBs. Update
6161 reloads_num only for small register classes. Don't split for
6164 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6167 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6170 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6172 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6175 2014-02-04 Marek Polacek <polacek@redhat.com>
6177 * gdbinit.in (pel): Define.
6179 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6181 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6184 2014-02-04 Richard Biener <rguenther@suse.de>
6187 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6188 in function context local.
6189 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6190 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6191 similar to LTO_imported_decl_ref.
6193 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6195 PR tree-optimization/60002
6196 * cgraphclones.c (build_function_decl_skip_args): Clear
6199 PR tree-optimization/60023
6200 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6201 false to gsi_replace.
6202 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6203 has been in some EH region and vec_stmt could throw, add
6204 vec_stmt into the same EH region.
6205 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6206 has no lhs, ignore it.
6207 * internal-fn.c (expand_MASK_LOAD): Likewise.
6210 * tree-inline.c (copy_forbidden): Fail for
6211 __attribute__((optimize (0))) functions.
6214 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6215 is set, copy one less argument.
6216 (expand_simd_clones): Don't subtract clone_info->inbranch
6217 from simd_clone_struct_alloc argument.
6219 PR rtl-optimization/57915
6220 * recog.c (simplify_while_replacing): If all unary/binary/relational
6221 operation arguments are constant, attempt to simplify those.
6224 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6225 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6227 2014-02-04 Richard Biener <rguenther@suse.de>
6229 PR tree-optimization/60012
6230 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6231 TBAA disambiguation to all DDRs.
6233 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6236 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6237 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6239 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6242 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6244 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6246 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6247 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6249 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6252 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6253 to figure out targets of polymorphic calls with known decl.
6254 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6255 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6256 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6257 (get_polymorphic_call_info): ... here.
6258 (get_polymorphic_call_info_from_invariant): New function.
6260 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6262 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6263 lookup via vtable pointer; check for type consistency
6264 and turn inconsitent facts into UNREACHABLE.
6265 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6266 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6267 type inconsistent querries; return UNREACHABLE instead.
6269 2014-02-03 Richard Henderson <rth@twiddle.net>
6272 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6273 already processed this node.
6274 (normalize_one_pred_1): Pass along mark_set.
6275 (normalize_one_pred): Create and destroy a pointer_set_t.
6276 (normalize_one_pred_chain): Likewise.
6278 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6280 PR gcov-profile/58602
6281 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6283 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6286 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6287 -fno-devirtualize; try to devirtualize by the knowledge of
6288 virtual table pointer given by aggregate propagation.
6289 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6290 (ipa_print_node_jump_functions): Dump also offset that
6291 is relevant for polymorphic calls.
6292 (determine_known_aggregate_parts): Add arg_type parameter; use it
6293 instead of determining the type from pointer type.
6294 (ipa_compute_jump_functions_for_edge): Update call of
6295 determine_known_aggregate_parts.
6296 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6297 (gimple_get_virt_method_for_binfo): ... here; simplify using
6298 vtable_pointer_value_to_vtable.
6299 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6300 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6301 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6302 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6304 (vtable_pointer_value_to_binfo): ... here.
6305 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6307 2014-02-03 Teresa Johnson <tejohnson@google.com>
6309 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6310 redef of outer loop index variable.
6312 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6316 * doc/extend.texi (Function Attributes): Typo.
6318 2014-02-03 Cong Hou <congh@google.com>
6320 PR tree-optimization/60000
6321 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6322 if the vectorized statement is a store. A store statement can only
6323 appear at the end of pattern statements.
6325 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6327 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6328 (ix86_option_override_internal): Default long double to 64-bit for
6329 32-bit Bionic and to 128-bit for 64-bit Bionic.
6331 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6332 TARGET_LONG_DOUBLE_128 is true.
6333 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6335 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6336 (mlong-double-64): Negate -mlong-double-128.
6337 (mlong-double-128): New option.
6339 * config/i386/i386-c.c (ix86_target_macros): Define
6340 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6342 * doc/invoke.texi: Document -mlong-double-128.
6344 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6346 PR rtl-optimization/60024
6347 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6349 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6351 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6353 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6355 PR rtl-optimization/57662
6356 * sel-sched.c (code_motion_path_driver): Do not mark already not
6357 existing blocks in the visiting bitmap.
6359 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6361 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6362 on the insn being emitted.
6364 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6365 Will Deacon <will.deacon@arm.com>
6367 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6369 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6371 * config/arm/arm-tables.opt: Regenerate.
6373 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6375 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6376 for vector types other than V16QImode.
6377 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6378 define_expand, and call altivec_expand_vec_perm_le when producing
6379 code with little endian element order.
6380 (*altivec_vperm_<mode>_internal): New insn having previous
6381 behavior of altivec_vperm_<mode>.
6382 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6383 altivec_expand_vec_perm_le when producing code with little endian
6385 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6386 behavior of altivec_vperm_<mode>_uns.
6388 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6390 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6391 (altivec_vsumsws): Add handling for -maltivec=be with a little
6393 (altivec_vsumsws_direct): New.
6394 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6395 gen_altivec_vsumsws.
6397 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6399 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6400 vtable_pointer_value_to_binfo): New functions.
6401 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6402 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6404 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6406 * config/nios2/nios2.md (load_got_register): Initialize GOT
6407 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6408 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6410 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6412 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6413 preserverd by passthrough, do not propagate the type.
6415 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6417 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6418 (mips_atomic_assign_expand_fenv): New function.
6419 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6421 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6423 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6424 (__builtin_mips_set_fcsr): Likewise.
6425 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6426 MIPS_USI_FTYPE_VOID.
6427 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6428 (mips16_expand_set_fcsr): Likewise.
6429 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6430 (mips16_set_fcsr_stub): Likewise.
6431 (mips16_get_fcsr_one_only_stub): New class.
6432 (mips16_set_fcsr_one_only_stub): Likewise.
6433 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6434 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6435 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6436 (hard_float): New availability predicate.
6437 (mips_builtins): Add get_fcsr and set_fcsr.
6438 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6439 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6440 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6441 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6442 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6445 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6447 * config/mips/mips.c (mips_one_only_stub): New class.
6448 (mips_need_mips16_rdhwr_p): Replace with...
6449 (mips16_rdhwr_stub): ...this new variable.
6450 (mips16_stub_call_address): New function.
6451 (mips16_rdhwr_one_only_stub): New class.
6452 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6453 (mips_output_mips16_rdhwr): Delete.
6454 (mips_finish_stub): New function.
6455 (mips_code_end): Use it to handle rdhwr stubs.
6457 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6460 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6461 when calculating size of integer atomic types.
6463 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6465 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6467 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6469 PR tree-optimization/60003
6470 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6471 * profile.c (branch_prob): Use gimple_call_builtin_p
6472 to check for BUILT_IN_SETJMP_RECEIVER.
6473 * tree-inline.c (copy_bb): Call notice_special_calls.
6475 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6478 * lra-constraints.c (process_alt_operands): Update reload_sum only
6481 2014-01-31 Richard Henderson <rth@redhat.com>
6484 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6485 until after else_eh is processed.
6487 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6489 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6490 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6491 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6492 in smmintrin.h, remove them.
6493 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6494 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6495 * config/i386/i386.md (ROUND_SAE): Fix value.
6496 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6497 (const48_operand): New.
6498 * config/i386/subst.md (round), (round_expand): Use
6499 const_4_or_8_to_11_operand.
6500 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6502 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6504 * config/i386/constraints.md (Yk): Swap meaning with k.
6505 * config/i386/i386.md (movhi_internal): Change Yk to k.
6506 (movqi_internal): Ditto.
6507 (*k<logic><mode>): Ditto.
6510 (kandn<mode>): Ditto.
6511 (*<code>hi_1): Ditto.
6512 (*<code>qi_1): Ditto.
6513 (kxnor<mode>): Ditto.
6514 (kortestzhi): Ditto.
6515 (kortestchi): Ditto.
6517 (*one_cmplhi2_1): Ditto.
6518 (*one_cmplqi2_1): Ditto.
6519 * config/i386/sse.md (): Change k to Yk.
6520 (avx512f_load<mode>_mask): Ditto.
6521 (avx512f_blendm<mode>): Ditto.
6522 (avx512f_store<mode>_mask): Ditto.
6523 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6524 (avx512f_storedqu<mode>_mask): Ditto.
6525 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6527 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6528 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6529 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6530 (avx512f_maskcmp<mode>3): Ditto.
6531 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6532 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6533 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6534 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6535 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6536 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6537 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6538 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6539 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6540 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6541 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6542 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6543 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6544 (vec_extract_lo_<mode>_maskm): Ditto.
6545 (vec_extract_hi_<mode>_maskm): Ditto.
6546 (avx512f_vternlog<mode>_mask): Ditto.
6547 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6548 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6549 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6550 (avx512f_<code>v8div16qi2_mask): Ditto.
6551 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6552 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6553 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6554 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6555 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6556 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6557 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6558 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6559 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6560 (avx512cd_maskb_vec_dupv8di): Ditto.
6561 (avx512cd_maskw_vec_dupv16si): Ditto.
6562 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6563 (avx512f_vpermi2var<mode>3_mask): Ditto.
6564 (avx512f_vpermi2var<mode>3_mask): Ditto.
6565 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6566 (*avx512f_gathersi<mode>): Ditto.
6567 (*avx512f_gathersi<mode>_2): Ditto.
6568 (*avx512f_gatherdi<mode>): Ditto.
6569 (*avx512f_gatherdi<mode>_2): Ditto.
6570 (*avx512f_scattersi<mode>): Ditto.
6571 (*avx512f_scatterdi<mode>): Ditto.
6572 (avx512f_compress<mode>_mask): Ditto.
6573 (avx512f_compressstore<mode>_mask): Ditto.
6574 (avx512f_expand<mode>_mask): Ditto.
6575 * config/i386/subst.md (mask): Change k to Yk.
6576 (mask_scalar_merge): Ditto.
6579 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6581 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6583 2014-01-31 Richard Biener <rguenther@suse.de>
6586 * builtins.c (fold_builtin_memory_op): Make sure to not
6587 use a floating-point mode or a boolean or enumeral type for
6590 2014-01-30 DJ Delorie <dj@redhat.com>
6592 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6593 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6594 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6595 whenever main() has an epilogue.
6597 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6599 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6600 unused variable "field".
6601 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6602 (vsx_mergeh_<mode>): Likewise.
6603 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6604 (altivec_vmrghh): Likewise.
6605 (altivec_vmrghw): Likewise.
6606 (altivec_vmrglb): Likewise.
6607 (altivec_vmrglh): Likewise.
6608 (altivec_vmrglw): Likewise.
6609 (altivec_vspltb): Add missing uses.
6610 (altivec_vsplth): Likewise.
6611 (altivec_vspltw): Likewise.
6612 (altivec_vspltsf): Likewise.
6614 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6617 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6618 frame related instructions.
6620 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6622 PR rtl-optimization/59959
6623 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6624 any reload of register whose subreg is invalid.
6626 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6628 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6629 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6630 Add missing return type - void.
6632 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6634 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6635 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6636 remove element index adjustment for endian (now handled in vsx.md
6638 (altivec_expand_vec_perm_const): Use
6639 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6640 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6641 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6642 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6643 define_expand and a new define_insn *altivec_vspltb_internal;
6644 adjust for -maltivec=be on a little endian target.
6645 (altivec_vspltb_direct): New.
6646 (altivec_vsplth): Divide into a define_expand and a new
6647 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6648 little endian target.
6649 (altivec_vsplth_direct): New.
6650 (altivec_vspltw): Divide into a define_expand and a new
6651 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6652 little endian target.
6653 (altivec_vspltw_direct): New.
6654 (altivec_vspltsf): Divide into a define_expand and a new
6655 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6656 a little endian target.
6658 2014-01-30 Richard Biener <rguenther@suse.de>
6660 PR tree-optimization/59993
6661 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6662 can propagate form the earlier stmt and avoid the transform
6663 when the intermediate result is needed.
6665 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6667 * README.Portability: Fix typo.
6669 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6671 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6672 comparison_operator with ordered_comparison_operator.
6674 2014-01-30 Nick Clifton <nickc@redhat.com>
6676 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6677 Rename to mn10300_store_multiple_regs.
6678 * config/mn10300/mn10300.c: Likewise.
6679 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6680 store_multiple_regs.
6681 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6682 Call mn10300_store_multiple_regs.
6684 2014-01-30 Nick Clifton <nickc@redhat.com>
6685 DJ Delorie <dj@redhat.com>
6687 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6688 %fp 2 to keep registers after it properly word-aligned.
6689 (rl78_alloc_physical_registers_umul): Handle the case where both
6690 input operands are the same.
6692 2014-01-30 Richard Biener <rguenther@suse.de>
6694 PR tree-optimization/59903
6695 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6698 2014-01-30 Jason Merrill <jason@redhat.com>
6701 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6704 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6706 2014-01-30 Richard Biener <rguenther@suse.de>
6708 PR tree-optimization/59951
6709 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6711 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6714 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6715 SFmode to DFmode case.
6717 2014-01-29 DJ Delorie <dj@redhat.com>
6719 * config/msp430/msp430.opt (-minrt): New.
6720 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6722 (ENDFILE_SPEC): Likewise.
6724 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6726 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6727 (estimate_function_body_sizes): Use it.
6729 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6732 * dwarf2out.c (is_cxx_auto): New.
6733 (is_base_type): Use it.
6734 (gen_type_die_with_usage): Likewise.
6736 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6738 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6739 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6740 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6741 -maltivec=be with LE targets.
6742 (vsx_mergeh_<mode>): Likewise.
6743 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6744 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6745 (altivec_vmrghb): Replace with define_expand and new
6746 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6747 (altivec_vmrghb_direct): New define_insn.
6748 (altivec_vmrghh): Replace with define_expand and new
6749 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6750 (altivec_vmrghh_direct): New define_insn.
6751 (altivec_vmrghw): Replace with define_expand and new
6752 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6753 (altivec_vmrghw_direct): New define_insn.
6754 (*altivec_vmrghsf): Adjust for endianness.
6755 (altivec_vmrglb): Replace with define_expand and new
6756 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6757 (altivec_vmrglb_direct): New define_insn.
6758 (altivec_vmrglh): Replace with define_expand and new
6759 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6760 (altivec_vmrglh_direct): New define_insn.
6761 (altivec_vmrglw): Replace with define_expand and new
6762 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6763 (altivec_vmrglw_direct): New define_insn.
6764 (*altivec_vmrglsf): Adjust for endianness.
6765 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6766 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6767 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6768 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6769 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6770 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6771 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6772 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6774 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6776 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6777 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6780 2014-01-29 Richard Biener <rguenther@suse.de>
6782 PR tree-optimization/58742
6783 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6784 associate_pointerplus_align.
6785 (associate_pointerplus_diff): New function.
6786 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6787 and associate_pointerplus_diff.
6789 2014-01-29 Richard Biener <rguenther@suse.de>
6791 * lto-streamer.h (LTO_major_version): Bump to 3.
6792 (LTO_minor_version): Reset to 0.
6794 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6796 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6797 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6798 (arm_file_start): Generate correct asm header for armv7ve.
6799 * config/arm/bpabi.h: Add multilib support for armv7ve.
6800 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6801 and cortex-a15 to armv7ve.
6802 * config/arm/t-aprofile: Add multilib support for armv7ve.
6803 * doc/invoke.texi: Document -march=armv7ve.
6805 2014-01-29 Richard Biener <rguenther@suse.de>
6807 PR tree-optimization/58742
6808 * tree-ssa-forwprop.c (associate_plusminus): Return true
6809 if we changed sth, defer EH cleanup to ...
6810 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6811 (simplify_mult): New function.
6813 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6816 PR tree-optimization/59920
6817 * tree.c (build_common_builtin_nodes): Remove
6818 __builtin_setjmp_dispatcher initialization.
6819 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6820 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6821 instead of gsi_after_labels + manually skipping debug stmts.
6822 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6823 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6824 * tree-inline.c (copy_edges_for_bb): Remove
6825 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6826 argument. Ignore computed_goto_p stmts. Don't call
6827 make_abnormal_goto_edges. If a call might need abnormal edges
6828 for non-local gotos, see if it already has an edge to
6829 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6830 with true argument, don't do anything then, otherwise add
6831 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6832 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6834 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6835 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6836 (lower_stmt): Don't set data->calls_builtin_setjmp.
6837 (lower_builtin_setjmp): Adjust comment.
6838 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6839 * tree-cfg.c (found_computed_goto): Remove.
6840 (factor_computed_gotos): Remove.
6841 (make_goto_expr_edges): Return bool, true for computed gotos.
6842 Don't call make_abnormal_goto_edges.
6843 (build_gimple_cfg): Don't set found_computed_goto, don't call
6844 factor_computed_gotos.
6845 (computed_goto_p): No longer static.
6846 (make_blocks): Don't set found_computed_goto.
6847 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6848 (make_edges): If make_goto_expr_edges returns true, push bb
6849 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6850 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6851 vector. Record mapping between bbs and OpenMP regions if there
6852 are any, adjust make_gimple_omp_edges caller. Call
6853 handle_abnormal_edges.
6854 (make_abnormal_goto_edges): Remove.
6855 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6856 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6857 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6858 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6859 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6860 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6861 filling *region also set *region_idx to (*region)->entry->index.
6864 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6865 For REGs set ORIGINAL_REGNO.
6867 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6869 * doc/md.texi: Mention that a target shouldn't implement
6870 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6873 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6875 PR tree-optimization/59594
6876 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6877 a copy of the datarefs vector rather than the vector itself.
6879 2014-01-28 Jason Merrill <jason@redhat.com>
6882 * dwarf2out.c (auto_die): New static.
6883 (gen_type_die_with_usage): Handle C++1y 'auto'.
6884 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6887 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6890 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6891 (SPEC_X32): Likewise.
6892 (SPEC_64): Likewise.
6893 * config/i386/i386.c (ix86_option_override_internal): Turn off
6894 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6896 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6897 * config/i386/i386.h (TARGET_16BIT): New macro.
6898 (TARGET_16BIT_P): Likewise.
6899 * config/i386/i386.opt: Add m16.
6900 * doc/invoke.texi: Document -m16.
6902 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6904 PR preprocessor/59935
6905 * input.c (location_get_source_line): Bail out on when line number
6906 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6908 2014-01-28 Richard Biener <rguenther@suse.de>
6910 PR tree-optimization/58742
6911 * tree-ssa-forwprop.c (associate_plusminus): Handle
6912 pointer subtraction of the form (T)(P + A) - (T)P.
6914 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6916 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6919 2014-01-28 Richard Biener <rguenther@suse.de>
6922 2014-01-28 Richard Biener <rguenther@suse.de>
6924 PR rtl-optimization/45364
6925 PR rtl-optimization/59890
6926 * var-tracking.c (local_get_addr_clear_given_value): Handle
6927 already cleared slot.
6928 (val_reset): Handle not allocated local_get_addr_cache.
6929 (vt_find_locations): Use post-order on the inverted CFG.
6931 2014-01-28 Richard Biener <rguenther@suse.de>
6933 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6935 2014-01-28 Richard Biener <rguenther@suse.de>
6937 PR rtl-optimization/45364
6938 PR rtl-optimization/59890
6939 * var-tracking.c (local_get_addr_clear_given_value): Handle
6940 already cleared slot.
6941 (val_reset): Handle not allocated local_get_addr_cache.
6942 (vt_find_locations): Use post-order on the inverted CFG.
6944 2014-01-28 Alan Modra <amodra@gmail.com>
6946 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6947 * configure.ac <recursive call for build != host>: Define
6948 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6949 and LD_FOR_BUILD too.
6950 * configure: Regenerate.
6952 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6954 * config/i386/i386.c (get_builtin_code_for_version): Separate
6955 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6956 Broadwell from Haswell.
6958 2014-01-27 Steve Ellcey <sellcey@mips.com>
6960 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6961 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6962 * config/mips/mips.c (mips_option_override): Change setting
6964 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6965 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6966 Change from Mask to Var.
6968 2014-01-27 Jeff Law <law@redhat.com>
6970 * ipa-inline.c (inline_small_functions): Fix typo.
6972 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6974 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6975 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6976 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6977 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6978 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6979 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6980 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6981 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6982 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6983 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6984 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6985 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6986 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6987 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6988 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6989 (_mm512_storeu_epi64): Ditto.
6990 (_mm512_cmpge_epi32_mask): Ditto.
6991 (_mm512_cmpge_epu32_mask): Ditto.
6992 (_mm512_cmpge_epi64_mask): Ditto.
6993 (_mm512_cmpge_epu64_mask): Ditto.
6994 (_mm512_cmple_epi32_mask): Ditto.
6995 (_mm512_cmple_epu32_mask): Ditto.
6996 (_mm512_cmple_epi64_mask): Ditto.
6997 (_mm512_cmple_epu64_mask): Ditto.
6998 (_mm512_cmplt_epi32_mask): Ditto.
6999 (_mm512_cmplt_epu32_mask): Ditto.
7000 (_mm512_cmplt_epi64_mask): Ditto.
7001 (_mm512_cmplt_epu64_mask): Ditto.
7002 (_mm512_cmpneq_epi32_mask): Ditto.
7003 (_mm512_cmpneq_epu32_mask): Ditto.
7004 (_mm512_cmpneq_epi64_mask): Ditto.
7005 (_mm512_cmpneq_epu64_mask): Ditto.
7006 (_mm512_expand_pd): Ditto.
7007 (_mm512_expand_ps): Ditto.
7008 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
7009 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
7010 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
7011 * config/i386/i386.c (ix86_builtins): Add
7012 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
7013 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
7014 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
7015 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
7016 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
7017 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
7018 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
7019 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
7020 IX86_BUILTIN_PMOVUSQW512_MEM.
7021 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
7022 __builtin_ia32_pmovsqd512mem_mask,
7023 __builtin_ia32_pmovqd512mem_mask,
7024 __builtin_ia32_pmovusqw512mem_mask,
7025 __builtin_ia32_pmovsqw512mem_mask,
7026 __builtin_ia32_pmovqw512mem_mask,
7027 __builtin_ia32_pmovusdw512mem_mask,
7028 __builtin_ia32_pmovsdw512mem_mask,
7029 __builtin_ia32_pmovdw512mem_mask,
7030 __builtin_ia32_pmovqb512mem_mask,
7031 __builtin_ia32_pmovusqb512mem_mask,
7032 __builtin_ia32_pmovsqb512mem_mask,
7033 __builtin_ia32_pmovusdb512mem_mask,
7034 __builtin_ia32_pmovsdb512mem_mask,
7035 __builtin_ia32_pmovdb512mem_mask.
7036 (bdesc_args): Add __builtin_ia32_expanddf512,
7037 __builtin_ia32_expandsf512.
7038 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
7039 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
7040 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
7041 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
7042 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
7043 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
7044 (avx512f_<code>v8div16qi2_mask_store): This.
7045 (avx512f_expand<mode>): New.
7047 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
7049 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
7051 (_mm512_mask_prefetch_i64gather_pd): Ditto.
7052 (_mm512_prefetch_i32scatter_pd): Ditto.
7053 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
7054 (_mm512_prefetch_i64scatter_pd): Ditto.
7055 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
7056 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
7057 (_mm512_mask_prefetch_i64gather_ps): Ditto.
7058 (_mm512_prefetch_i32scatter_ps): Ditto.
7059 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
7060 (_mm512_prefetch_i64scatter_ps): Ditto.
7061 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
7062 * config/i386/i386-builtin-types.def: Define
7063 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
7064 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
7065 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
7066 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
7067 IX86_BUILTIN_SCATTERPFQPD.
7068 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
7069 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
7070 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
7071 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
7072 __builtin_ia32_scatterpfqps.
7073 (ix86_expand_builtin): Expand new built-ins.
7074 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
7075 fix memory access data type.
7076 (*avx512pf_gatherpf<mode>_mask): Ditto.
7077 (*avx512pf_gatherpf<mode>): Ditto.
7078 (avx512pf_scatterpf<mode>): Ditto.
7079 (*avx512pf_scatterpf<mode>_mask): Ditto.
7080 (*avx512pf_scatterpf<mode>): Ditto.
7081 (GATHER_SCATTER_SF_MEM_MODE): New.
7082 (avx512pf_gatherpf<mode>df): Ditto.
7083 (*avx512pf_gatherpf<mode>df_mask): Ditto.
7084 (*avx512pf_scatterpf<mode>df): Ditto.
7086 2014-01-27 Jakub Jelinek <jakub@redhat.com>
7089 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7090 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7093 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7095 * common/config/arm/arm-common.c
7096 (arm_rewrite_mcpu): Handle multiple names.
7098 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7100 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7102 * gimple-builder.h (create_gimple_tmp): Delete.
7104 2014-01-27 Christian Bruel <christian.bruel@st.com>
7106 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7109 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7111 * config/pa/pa.md (call): Generate indirect long calls to non-local
7112 functions when outputing 32-bit code.
7113 (call_value): Likewise except for special call to buggy powf function.
7115 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7116 portable runtime and PIC indirect calls.
7117 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7118 and PIC call sequences. Use ldo instead of blr to set return register
7119 in PIC call sequence.
7121 2014-01-25 Walter Lee <walt@tilera.com>
7123 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7124 avoid clobbering a live register.
7126 2014-01-25 Walter Lee <walt@tilera.com>
7128 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7129 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7130 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7131 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7133 2014-01-25 Walter Lee <walt@tilera.com>
7135 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7136 arguments on even registers.
7137 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7139 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7140 (BIGGEST_ALIGNMENT): Ditto.
7141 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7143 2014-01-25 Walter Lee <walt@tilera.com>
7145 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7146 insns before bundling.
7147 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7149 2014-01-25 Walter Lee <walt@tilera.com>
7151 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7152 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7153 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7155 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7157 * config/mips/constraints.md (kl): Delete.
7158 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7159 define expands, using...
7160 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7161 instructions for MIPS16.
7162 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7163 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7165 2014-01-25 Walter Lee <walt@tilera.com>
7167 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7171 2014-01-25 Walter Lee <walt@tilera.com>
7173 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7174 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7176 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7178 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7181 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7183 * print-rtl.c (in_call_function_usage): New var.
7184 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7185 EXPR_LIST mode as mode and not as reg note name.
7188 * cfgloopmanip.c (copy_loop_info): If
7189 loop->warned_aggressive_loop_optimizations, make sure
7190 the flag is set in target loop too.
7192 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7194 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7196 * builtins.def: Likewise.
7197 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7198 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7199 * ira.c (ira_setup_eliminable_regset): Likewise.
7200 * omp-low.c (gate_expand_omp): Likewise.
7201 (execute_lower_omp): Likewise.
7202 (diagnose_sb_0): Likewise.
7203 (gate_diagnose_omp_blocks): Likewise.
7204 (simd_clone_clauses_extract): Likewise.
7207 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7209 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7210 correction for little endian...
7211 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7214 2014-01-24 Jeff Law <law@redhat.com>
7216 PR tree-optimization/59919
7217 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7218 for non-returning calls.
7220 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7222 * common/config/aarch64/aarch64-common.c
7223 (aarch64_rewrite_mcpu): Handle multiple names.
7224 * config/aarch64/aarch64.h
7225 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7227 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7229 * input.c (add_file_to_cache_tab): Handle the case where fopen
7232 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7235 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7236 from push operand if code of push isn't PRE_DEC.
7238 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7241 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7242 -mquad-memory-atomic. Update -mquad-memory documentation to say
7243 it is only used for non-atomic loads/stores.
7245 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7246 -mquad-memory or -mquad-memory-atomic switches.
7248 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7249 -mquad-memory-atomic to ISA 2.07 support.
7251 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7252 to separate support of normal quad word memory operations (ldq, stq)
7253 from the atomic quad word memory operations.
7255 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7256 support to separate non-atomic quad word operations from atomic
7257 quad word operations. Disable non-atomic quad word operations in
7258 little endian mode so that we don't have to swap words after the
7259 load and before the store.
7260 (quad_load_store_p): Add comment about atomic quad word support.
7261 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7262 options printed with -mdebug=reg.
7264 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7265 -mquad-memory-atomic as the test for whether we have quad word
7266 atomic instructions.
7267 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7268 or -mp8-vector are used, allow byte/half-word atomic operations.
7270 * config/rs6000/sync.md (load_lockedti): Insure that the address
7271 is a proper indexed or indirect address for the lqarx instruction.
7272 On little endian systems, swap the hi/lo registers after the lqarx
7274 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7275 insure the address is valid for the lqarx instruction.
7276 (store_conditionalti): Insure that the address is a proper indexed
7277 or indirect address for the stqcrx. instruction. On little endian
7278 systems, swap the hi/lo registers before doing the stqcrx.
7280 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7281 insure the address is valid for the stqcrx. instruction.
7283 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7284 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7285 type of quad memory support is available.
7287 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7290 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7291 there is a danger of looping.
7293 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7295 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7296 force flag_ira_loop_pressure if set via command line.
7298 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7300 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7301 (ashr_simd): New builtin handling DI mode.
7302 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7303 (aarch64_sshr_simddi): New match pattern.
7304 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7305 (vshrd_n_s64): Likewise.
7306 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7308 2014-01-23 Nick Clifton <nickc@redhat.com>
7310 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7311 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7312 favour of mcu specific scripts.
7313 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7316 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7317 Alex Velenko <Alex.Velenko@arm.com>
7319 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7320 (vaddv_s16): Likewise.
7321 (vaddv_s32): Likewise.
7322 (vaddv_u8): Likewise.
7323 (vaddv_u16): Likewise.
7324 (vaddv_u32): Likewise.
7325 (vaddvq_s8): Likewise.
7326 (vaddvq_s16): Likewise.
7327 (vaddvq_s32): Likewise.
7328 (vaddvq_s64): Likewise.
7329 (vaddvq_u8): Likewise.
7330 (vaddvq_u16): Likewise.
7331 (vaddvq_u32): Likewise.
7332 (vaddvq_u64): Likewise.
7333 (vaddv_f32): Likewise.
7334 (vaddvq_f32): Likewise.
7335 (vaddvq_f64): Likewise.
7336 (vmaxv_f32): Likewise.
7337 (vmaxv_s8): Likewise.
7338 (vmaxv_s16): Likewise.
7339 (vmaxv_s32): Likewise.
7340 (vmaxv_u8): Likewise.
7341 (vmaxv_u16): Likewise.
7342 (vmaxv_u32): Likewise.
7343 (vmaxvq_f32): Likewise.
7344 (vmaxvq_f64): Likewise.
7345 (vmaxvq_s8): Likewise.
7346 (vmaxvq_s16): Likewise.
7347 (vmaxvq_s32): Likewise.
7348 (vmaxvq_u8): Likewise.
7349 (vmaxvq_u16): Likewise.
7350 (vmaxvq_u32): Likewise.
7351 (vmaxnmv_f32): Likewise.
7352 (vmaxnmvq_f32): Likewise.
7353 (vmaxnmvq_f64): Likewise.
7354 (vminv_f32): Likewise.
7355 (vminv_s8): Likewise.
7356 (vminv_s16): Likewise.
7357 (vminv_s32): Likewise.
7358 (vminv_u8): Likewise.
7359 (vminv_u16): Likewise.
7360 (vminv_u32): Likewise.
7361 (vminvq_f32): Likewise.
7362 (vminvq_f64): Likewise.
7363 (vminvq_s8): Likewise.
7364 (vminvq_s16): Likewise.
7365 (vminvq_s32): Likewise.
7366 (vminvq_u8): Likewise.
7367 (vminvq_u16): Likewise.
7368 (vminvq_u32): Likewise.
7369 (vminnmv_f32): Likewise.
7370 (vminnmvq_f32): Likewise.
7371 (vminnmvq_f64): Likewise.
7373 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7375 * config/aarch64/aarch64-simd.md
7376 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7377 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7378 (*aarch64_mul3_elt<mode>): Likewise.
7379 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7380 (*aarch64_mul3_elt_to_64v2df): Likewise.
7381 (*aarch64_mla_elt<mode>): Likewise.
7382 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7383 (*aarch64_mls_elt<mode>): Likewise.
7384 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7385 (*aarch64_fma4_elt<mode>): Likewise.
7386 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7387 (*aarch64_fma4_elt_to_64v2df): Likewise.
7388 (*aarch64_fnma4_elt<mode>): Likewise.
7389 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7390 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7391 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7392 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7393 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7394 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7395 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7396 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7397 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7399 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7401 * config/aarch64/aarch64-simd.md
7402 (aarch64_be_checked_get_lane<mode>): New define_expand.
7403 * config/aarch64/aarch64-simd-builtins.def
7404 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7405 New builtin definition.
7406 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7407 Use new safe be builtin.
7409 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7411 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7413 (aarch64_be_st1<mode>): Likewise.
7414 (aarch_ld1<VALL:mode>): Define_expand modified.
7415 (aarch_st1<VALL:mode>): Likewise.
7416 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7417 (UNSPEC_ST1): Likewise.
7419 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7421 * config/microblaze/microblaze.md: Add trap insn and attribute
7423 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7425 PR preprocessor/58580
7426 * input.h (location_get_source_line): Take an additional line_size
7428 (void diagnostics_file_cache_fini): Declare new function.
7429 * input.c (struct fcache): New type.
7430 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7431 New static constants.
7432 (diagnostic_file_cache_init, total_lines_num)
7433 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7434 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7435 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7436 (get_next_line, read_next_line, goto_next_line, read_line_num):
7437 New static function definitions.
7438 (diagnostic_file_cache_fini): New function.
7439 (location_get_source_line): Take an additional output line_len
7440 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7442 * diagnostic.c (diagnostic_finish): Call
7443 diagnostic_file_cache_fini.
7444 (adjust_line): Take an additional input parameter for the length
7445 of the line, rather than calculating it with strlen.
7446 (diagnostic_show_locus): Adjust the use of
7447 location_get_source_line and adjust_line with respect to their new
7448 signature. While displaying a line now, do not stop at the first
7449 null byte. Rather, display the zero byte as a space and keep
7450 going until we reach the size of the line.
7451 * Makefile.in: Add vec.o to OBJS-libcommon
7453 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7454 Ilya Tocar <ilya.tocar@intel.com>
7456 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7457 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7458 (__builtin_ia32_kmov16): Ditto.
7459 * config/i386/i386.md (UNSPEC_KMOV): New.
7462 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7464 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7465 (_mm512_storeu_si512): Ditto.
7467 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7470 * rtl.h (get_referenced_operands): Declare.
7471 * recog.c (get_referenced_operands): New function.
7472 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7473 operands have been referenced when recording LO_SUM references.
7475 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7477 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7479 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7481 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7482 Enable for generic and recent AMD targets.
7484 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7486 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7487 ARG_SIZE note when adjustment was eliminated.
7489 2014-01-22 Jeff Law <law@redhat.com>
7491 PR tree-optimization/59597
7492 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7493 in file. Accept new argument REGISTERING and use it to modify
7494 dump output appropriately.
7495 (register_jump_thread): Corresponding changes.
7496 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7497 thread paths involving joiner blocks. Add code to dump cancelled
7498 jump threading paths.
7500 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7502 PR rtl-optimization/59477
7503 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7504 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7506 2014-01-22 Tom Tromey <tromey@redhat.com>
7508 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7510 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7512 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7514 PR rtl-optimization/59896
7515 * lra-constraints.c (process_alt_operands): Check unused note for
7516 matched operands of insn with no output reloads.
7518 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7520 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7521 (mips_move_from_gpr_cost): Likewise.
7523 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7525 PR rtl-optimization/59858
7526 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7527 ira_class_hard_regs_num.
7528 (process_alt_operands): Increase reject for dying matched operand.
7530 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7533 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7534 smaller than size, perform several stores or loads and stores
7535 at dst + count - size to store or copy all of size bytes, rather
7536 than just last modesize bytes.
7538 2014-01-20 DJ Delorie <dj@redhat.com>
7540 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7541 that CLOBBERs are REGs before propogating their values.
7543 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7546 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7547 (cgraph_inline_failed_type): New function.
7548 * cgraph.h (DEFCIFCODE): Add type.
7549 (cgraph_inline_failed_type_t): New enum.
7550 (cgraph_inline_failed_type): New prototype.
7551 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7552 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7553 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7554 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7555 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7556 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7557 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7558 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7559 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7560 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7561 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7562 OPTIMIZATION_MISMATCH.
7563 * tree-inline.c (expand_call_inline): Emit errors during
7564 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7566 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7569 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7570 mode attribute in insn output.
7572 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7574 * output.h (output_constant): Delete.
7575 * varasm.c (output_constant): Make private.
7577 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7579 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7581 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7584 * tree.h (fold_builtin_strcat): New prototype.
7585 * builtins.c (fold_builtin_strcat): No longer static. Add len
7586 argument, if non-NULL, don't call c_strlen. Optimize
7587 directly into __builtin_memcpy instead of __builtin_strcpy.
7588 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7589 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7591 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7593 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7594 for SImode_address_operand operands, having only a REG argument.
7596 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7598 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7599 loader name using mbig-endian.
7600 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7602 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7604 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7608 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7610 * config/aarch64/aarch64-protos.h
7611 (aarch64_cannot_change_mode_class_ptr): Declare.
7612 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7613 aarch64_cannot_change_mode_class_ptr): New.
7614 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7615 backend hook aarch64_cannot_change_mode_class.
7617 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7619 * common/config/aarch64/aarch64-common.c
7620 (aarch64_handle_option): Don't handle any option order logic here.
7621 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7622 selected_cpu, warn on architecture version mismatch.
7623 (aarch64_override_options): Fix parsing order for option strings.
7625 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7626 Iain Sandoe <iain@codesourcery.com>
7629 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7630 warning. Amend comment to reflect current functionality.
7632 2014-01-20 Richard Biener <rguenther@suse.de>
7635 * builtins.c (fold_builtin_strcat): Remove case better handled
7636 by tree-ssa-strlen.c.
7638 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7640 * config/aarch64/aarch64.opt
7641 (mcpu, march, mtune): Make case-insensitive.
7643 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7646 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7647 if operands[1] is a REG or ZERO_EXTEND of a REG.
7649 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7651 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7653 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7655 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7656 long non-pic millicode calls.
7658 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7660 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7662 2014-01-19 Kito Cheng <kito@0xlab.org>
7664 * builtins.c (expand_movstr): Check movstr expand done or fail.
7666 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7667 H.J. Lu <hongjiu.lu@intel.com>
7670 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7671 to DImode for zero-extended addresses.
7673 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7675 PR rtl-optimization/57763
7676 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7677 on the new indirect jump_insn and increment LABEL_NUSES (label).
7679 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7683 * config.gcc (x86_archs): New variable.
7684 (x86_64_archs): Likewise.
7685 (x86_cpus): Likewise.
7686 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7687 --with-arch/--with-cpu= options.
7688 Support --with-arch=/--with-cpu={nehalem,westmere,
7689 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7691 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7693 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7694 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7696 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7698 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7700 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7703 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7704 clear cpp_get_options (parse_in)->warn_unused_macros for
7705 ix86_target_macros_internal with cpp_define.
7707 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7709 * jump.c (delete_related_insns): Keep (use (insn))s.
7710 * reorg.c (redundant_insn): Check for barriers too.
7712 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7714 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7716 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7718 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7719 call to $$dyncall when TARGET_LONG_CALLS is true.
7721 2014-01-17 Jeff Law <law@redhat.com>
7723 * ree.c (combine_set_extension): Temporarily disable test for
7724 changing number of hard registers.
7726 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7729 * ipa-inline-analysis.c (inline_free_summary):
7730 Do not free summary of aliases.
7732 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7735 * gimplify.c (gimplify_expr): Use create_tmp_var
7736 instead of create_tmp_var_raw. If cond doesn't have
7737 integral type, don't add the IFN_ANNOTATE builtin at all.
7739 2014-01-17 Martin Jambor <mjambor@suse.cz>
7742 * ipa-cp.c (prev_edge_clone): New variable.
7743 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7744 Also resize prev_edge_clone vector.
7745 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7746 (ipcp_edge_removal_hook): New function.
7747 (ipcp_driver): Register ipcp_edge_removal_hook.
7749 2014-01-17 Andrew Pinski <apinski@cavium.com>
7750 Steve Ellcey <sellcey@mips.com>
7753 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7756 2014-01-17 Jeff Law <law@redhat.com>
7759 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7760 so that pass_ccp runs first.
7762 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7764 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7765 (ix86_adjust_cost): Use !TARGET_XXX.
7766 (do_reorder_for_imul): Likewise.
7767 (swap_top_of_ready_list): Likewise.
7768 (ix86_sched_reorder): Likewise.
7770 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7772 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7773 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7774 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7775 (intel_memset): New. Duplicate slm_memset.
7776 (intel_cost): New. Duplicate slm_cost.
7777 (m_INTEL): New macro.
7778 (processor_target_table): Add "intel".
7779 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7780 with PROCESSOR_INTEL for "intel".
7781 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7782 PROCESSOR_SILVERMONT.
7783 (ix86_issue_rate): Likewise.
7784 (ix86_adjust_cost): Likewise.
7785 (ia32_multipass_dfa_lookahead): Likewise.
7786 (swap_top_of_ready_list): Likewise.
7787 (ix86_sched_reorder): Likewise.
7788 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7789 instead of TARGET_OPT_AGU.
7790 * config/i386/i386.h (TARGET_INTEL): New.
7791 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7792 (processor_type): Add PROCESSOR_INTEL.
7793 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7794 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7796 2014-01-17 Marek Polacek <polacek@redhat.com>
7799 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7802 2014-01-17 Richard Biener <rguenther@suse.de>
7804 PR tree-optimization/46590
7805 * opts.c (default_options_table): Add entries for
7806 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7807 all enabled at -O1 but not for -Og.
7808 * common.opt (fbranch-count-reg): Remove Init(1).
7809 (fmove-loop-invariants): Likewise.
7810 (ftree-pta): Likewise.
7812 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7814 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7815 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7816 decls to at least the GCC 4.8 used alignments.
7819 * tree-nested.c (convert_nonlocal_reference_stmt,
7820 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7821 of GIMPLE_BIND stmts, adjust associated decls.
7823 2014-01-17 Richard Biener <rguenther@suse.de>
7825 PR tree-optimization/46590
7826 * vec.h (vec<>::bseach): New member function implementing
7827 binary search according to C89 bsearch.
7828 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7829 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7830 bitmap pointer again. Make accesses_in_loop a flat array.
7831 (mem_ref_obstack): New global.
7832 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7833 (mark_ref_stored): Likewise.
7834 (ref_indep_loop_p_2): Likewise.
7835 (set_ref_stored_in_loop): New helper function.
7836 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7837 (memref_free): Adjust.
7838 (record_mem_ref_loc): Simplify.
7839 (gather_mem_refs_stmt): Adjust.
7840 (sort_locs_in_loop_postorder_cmp): New function.
7841 (analyze_memory_references): Sort accesses_in_loop after
7842 loop postorder number.
7843 (find_ref_loc_in_loop_cmp): New function.
7844 (for_all_locs_in_loop): Find relevant cluster of locs in
7845 accesses_in_loop and iterate without recursion.
7846 (execute_sm): Avoid uninit warning.
7847 (struct ref_always_accessed): Simplify.
7848 (ref_always_accessed::operator ()): Likewise.
7849 (ref_always_accessed_p): Likewise.
7850 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7851 loop postorder numbers here.
7852 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7855 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7858 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7859 on decls for which assemble_alias has been called.
7861 2014-01-17 Nick Clifton <nickc@redhat.com>
7863 * config/msp430/msp430.opt: (mcpu): New option.
7864 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7865 (msp430_option_override): Parse target_cpu. If the MCU name
7866 matches a generic string, clear target_mcu.
7867 (msp430_attr): Allow numeric interrupt values up to 63.
7868 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7869 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7871 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7873 * config/msp430/msp430.md (popm): Use %J rather than %I.
7874 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7875 (addhi_cy_i): Use immediate_operand for operand 2.
7876 * doc/invoke.texi: Document -mcpu option.
7878 2014-01-17 Richard Biener <rguenther@suse.de>
7880 PR rtl-optimization/38518
7881 * df.h (df_analyze_loop): Declare.
7882 * df-core.c: Include cfgloop.h.
7883 (df_analyze_1): Split out main part of df_analyze.
7884 (df_analyze): Adjust.
7885 (loop_inverted_post_order_compute): New function.
7886 (loop_post_order_compute): Likewise.
7887 (df_analyze_loop): New function avoiding whole-function
7889 * loop-invariant.c (find_defs): Use df_analyze_loop.
7890 (find_invariants): Adjust.
7891 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7893 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7895 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7896 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7898 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7900 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7901 traversal when removing references.
7903 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7906 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7908 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7911 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7912 pushing a reload for an autoinc when we had previously reloaded an
7913 inner part of the address.
7915 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7917 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7919 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7920 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7921 when not giving up or versioning for alias only because of
7923 (vect_analyze_data_ref_dependences): Set to true.
7924 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7926 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7927 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7931 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7934 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7935 operand 0 predicate for gathers, use a new pseudo as subtarget.
7937 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7940 * lra-constraints.c (process_alt_operands): Add printing debug info.
7941 Check absence of input/output reloads for matched operands too.
7943 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7945 PR rtl-optimization/59835
7946 * ira.c (ira_init_register_move_cost): Increase cost for
7949 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7951 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7953 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7956 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7957 non-register objects. Use gen_(high/low)part more consistently.
7960 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7963 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7964 endian support, remove tests for WORDS_BIG_ENDIAN.
7965 (p8_mfvsrd_3_<mode>): Likewise.
7966 (reload_gpr_from_vsx<mode>): Likewise.
7967 (reload_gpr_from_vsxsf): Likewise.
7968 (p8_mfvsrd_4_disf): Likewise.
7970 2014-01-16 Richard Biener <rguenther@suse.de>
7972 PR rtl-optimization/46590
7973 * lcm.c (compute_antinout_edge): Use postorder iteration.
7974 (compute_laterin): Use inverted postorder iteration.
7976 2014-01-16 Nick Clifton <nickc@redhat.com>
7979 * varasm.c (output_constant): Return the number of bytes actually
7981 (output_constructor_array_range): Update the field size with the
7982 number of bytes emitted by output_constant.
7983 (output_constructor_regular_field): Likewise. Also do not
7984 complain if the total number of bytes emitted is now greater
7985 than the expected fieldpos.
7986 * output.h (output_constant): Update prototype and descriptive comment.
7988 2014-01-16 Marek Polacek <polacek@redhat.com>
7991 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7992 it is error_mark_node.
7994 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7996 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7997 VALID_AVX256_REG_OR_OI_MODE.
7999 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
8001 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
8002 current procedure should be profiled.
8004 2014-01-15 Andrew Pinski <apinski@cavium.com>
8006 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
8007 of moving from/to the STACK_REG register class.
8009 2014-01-15 Richard Henderson <rth@redhat.com>
8012 * reginfo.c (global_regs_decl): Globalize.
8013 * rtl.h (global_regs_decl): Declare.
8014 * ira.c (do_reload): Diagnose frame_pointer_needed and it
8015 reserved via global_regs.
8017 2014-01-15 Teresa Johnson <tejohnson@google.com>
8019 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
8021 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
8023 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
8024 and vmulosh rather than call gen_vec_widen_smult_*.
8025 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
8026 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
8027 (vec_widen_smult_even_v16qi): Likewise.
8028 (vec_widen_umult_even_v8hi): Likewise.
8029 (vec_widen_smult_even_v8hi): Likewise.
8030 (vec_widen_umult_odd_v16qi): Likewise.
8031 (vec_widen_smult_odd_v16qi): Likewise.
8032 (vec_widen_umult_odd_v8hi): Likewise.
8033 (vec_widen_smult_odd_v8hi): Likewise.
8034 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
8035 vmuloub rather than call gen_vec_widen_umult_*.
8036 (vec_widen_umult_lo_v16qi): Likewise.
8037 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
8038 vmulosb rather than call gen_vec_widen_smult_*.
8039 (vec_widen_smult_lo_v16qi): Likewise.
8040 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
8041 rather than call gen_vec_widen_umult_*.
8042 (vec_widen_umult_lo_v8hi): Likewise.
8043 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
8044 rather than call gen_vec_widen_smult_*.
8045 (vec_widen_smult_lo_v8hi): Likewise.
8047 2014-01-15 Jeff Law <law@redhat.com>
8049 PR tree-optimization/59747
8050 * ree.c (find_and_remove_re): Properly handle case where a second
8051 eliminated extension requires widening a copy created for elimination
8052 of a prior extension.
8053 (combine_set_extension): Ensure that the number of hard regs needed
8054 for a destination register does not change when we widen it.
8056 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
8058 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
8059 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
8060 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
8061 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
8062 (avr-*-rtems*): Likewise.
8063 (bfin*-rtems*): Likewise.
8064 (moxie-*-rtems*): Likewise.
8065 (h8300-*-rtems*): Likewise.
8066 (i[34567]86-*-rtems*): Likewise.
8067 (lm32-*-rtems*): Likewise.
8068 (m32r-*-rtems*): Likewise.
8069 (m68k-*-rtems*): Likewise.
8070 (microblaze*-*-rtems*): Likewise.
8071 (mips*-*-rtems*): Likewise.
8072 (powerpc-*-rtems*): Likewise.
8073 (sh-*-rtems*): Likewise.
8074 (sparc-*-rtems*): Likewise.
8075 (sparc64-*-rtems*): Likewise.
8076 (v850-*-rtems*): Likewise.
8077 (m32c-*-rtems*): Likewise.
8079 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
8081 PR rtl-optimization/59511
8082 * ira.c (ira_init_register_move_cost): Use memory costs for some
8083 cases of register move cost calculations.
8084 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
8085 instead of BB frequency.
8086 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
8087 * lra-assigns.c (find_hard_regno_for): Ditto.
8089 2014-01-15 Richard Biener <rguenther@suse.de>
8091 PR tree-optimization/59822
8092 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8093 (vectorizable_load): Use it to hoist defs of uses of invariant
8094 loads out of the loop.
8096 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8097 Kugan Vivekanandarajah <kuganv@linaro.org>
8100 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8103 2014-01-15 Richard Biener <rguenther@suse.de>
8105 PR rtl-optimization/59802
8106 * lcm.c (compute_available): Use inverted postorder to seed
8107 the initial worklist.
8109 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8112 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8113 ADDR_REGS for invalid symrefs in non-PIC code.
8115 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8118 * builtins.c (determine_block_size): Initialize *probable_max_size
8119 even if len_rtx is CONST_INT.
8121 2014-01-14 Andrew Pinski <apinski@cavium.com>
8123 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8124 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8125 (cortexa53_tunings): Likewise.
8126 (aarch64_sched_issue_rate): New function.
8127 (TARGET_SCHED_ISSUE_RATE): Define.
8129 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8131 * ira-costs.c (find_costs_and_classes): Add missed
8132 ira_init_register_move_cost_if_necessary.
8134 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8137 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8139 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8142 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8143 to indicate if type is used for function return value. Warn ABI
8144 change if the vector mode isn't available for function return value.
8145 (ix86_function_arg_advance): Pass false to type_natural_mode.
8146 (ix86_function_arg): Likewise.
8147 (ix86_gimplify_va_arg): Likewise.
8148 (function_arg_32): Don't warn ABI change.
8149 (ix86_function_value): Pass true to type_natural_mode.
8150 (ix86_return_in_memory): Likewise.
8151 (ix86_struct_value_rtx): Removed.
8152 (TARGET_STRUCT_VALUE_RTX): Likewise.
8154 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8156 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8157 converting a conditional jump into a conditional return.
8159 2014-01-14 Richard Biener <rguenther@suse.de>
8161 PR tree-optimization/58921
8162 PR tree-optimization/59006
8163 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8164 hoisting invariant stmts.
8165 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8166 invariant loads on the preheader edge if possible.
8168 2014-01-14 Joey Ye <joey.ye@arm.com>
8170 * doc/plugin.texi (Building GCC plugins): Update to C++.
8172 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8174 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8175 (_mm_rcp28_round_ss): Ditto.
8176 (_mm_rsqrt28_round_sd): Ditto.
8177 (_mm_rsqrt28_round_ss): Ditto.
8178 (_mm_rcp28_sd): Ditto.
8179 (_mm_rcp28_ss): Ditto.
8180 (_mm_rsqrt28_sd): Ditto.
8181 (_mm_rsqrt28_ss): Ditto.
8182 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8183 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8184 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8185 (IX86_BUILTIN_RCP28SD): Ditto.
8186 (IX86_BUILTIN_RCP28SS): Ditto.
8187 (IX86_BUILTIN_RSQRT28SD): Ditto.
8188 (IX86_BUILTIN_RSQRT28SS): Ditto.
8189 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8190 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8191 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8192 (ix86_expand_special_args_builtin): Expand new FTYPE.
8193 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8194 (srcp14<mode>): Make insn unary.
8195 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8196 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8197 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8198 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8199 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8200 Fix rounding: make it SAE only.
8201 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8203 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8205 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8206 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8207 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8208 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8209 (round_saeonly_mask_scalar_operand4): Ditto.
8210 (round_saeonly_mask_scalar_op3): Ditto.
8211 (round_saeonly_mask_scalar_op4): Ditto.
8213 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8215 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8216 Implement -maltivec=be for vec_insert and vec_extract.
8218 2014-01-10 DJ Delorie <dj@redhat.com>
8220 * config/msp430/msp430.md (call_internal): Don't allow memory
8221 references with SP as the base register.
8222 (call_value_internal): Likewise.
8223 * config/msp430/constraints.md (Yc): New. For memory references
8224 that don't use SP as a base register.
8226 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8227 "an integer without a # prefix"
8228 * config/msp430/msp430.md (epilogue_helper): Use it.
8230 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8233 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8234 AVX512F gather builtins.
8235 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8236 on gather decls with INTEGER_TYPE masktype.
8237 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8238 directly into the builtin rather than hoisting it before loop.
8240 PR tree-optimization/59387
8241 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8242 (scev_const_prop): If folded_casts and type has undefined overflow,
8243 use force_gimple_operand instead of force_gimple_operand_gsi and
8244 for each added stmt if it is assign with
8245 arith_code_with_undefined_signed_overflow, call
8246 rewrite_to_defined_overflow.
8247 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8248 gimple-fold.h instead.
8249 (arith_code_with_undefined_signed_overflow,
8250 rewrite_to_defined_overflow): Moved to ...
8251 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8252 rewrite_to_defined_overflow): ... here. No longer static.
8253 Include gimplify-me.h.
8254 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8255 rewrite_to_defined_overflow): New prototypes.
8257 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8259 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8261 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8263 * builtins.c (get_object_alignment_2): Minor tweak.
8264 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8266 2014-01-13 Christian Bruel <christian.bruel@st.com>
8268 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8269 optimized non constant lengths.
8271 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8274 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8275 load as __atomic_load_N if possible.
8277 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8279 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8281 (rs6000_expand_builtin): Adjust call.
8283 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8286 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8287 * config/rs6000/rs6000.c: Include target-globals.h.
8288 (rs6000_set_current_function): Instead of doing target_reinit
8289 unconditionally, use save_target_globals_default_opts and
8290 restore_target_globals.
8292 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8294 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8295 (rs6000_expand_builtin): Handle mffs and mtfsf.
8296 (rs6000_init_builtins): Define mffs and mtfsf.
8297 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8298 (rs6000_mffs): New pattern.
8299 (rs6000_mtfsf): New pattern.
8301 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8303 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8304 Start narrowing with START. Apply candidate-use pair
8305 and check overall cost in narrowing.
8306 (iv_ca_prune): Pass new argument.
8308 2014-01-10 Jeff Law <law@redhat.com>
8311 * ree.c (combine_reaching_defs): Ensure the defining statement
8312 occurs before the extension when optimizing extensions with
8313 different source and destination hard registers.
8315 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8318 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8319 vtables into the type inheritance graph.
8321 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8323 PR rtl-optimization/59754
8324 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8325 modes in the REGNO != REGNO case.
8327 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8329 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8331 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8333 PR tree-optimization/59745
8334 * tree-predcom.c (tree_predictive_commoning_loop): Call
8335 free_affine_expand_cache if giving up because components is NULL.
8337 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8338 GC in payload of target_globals struct instead of allocating them on
8339 the heap and the larger structs separately using GC.
8340 * target-globals.h (struct target_globals): Make regs, hard_regs,
8341 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8342 of GTY((skip)) and change type to void *.
8343 (reset_target_globals): Cast loads from those fields to corresponding
8346 2014-01-10 Steve Ellcey <sellcey@mips.com>
8349 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8350 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8351 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8353 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8356 * aarch64-modes.def (CC_Zmode): New flags mode.
8357 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8358 represents an equality.
8359 (aarch64_get_condition_code): Handle CC_Zmode.
8360 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8362 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8364 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8365 extraction in good case.
8367 2014-01-10 Richard Biener <rguenther@suse.de>
8369 PR tree-optimization/59374
8370 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8371 checking after SLP discovery. Mark stmts not participating
8372 in any SLP instance properly.
8374 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8376 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8377 when handling a SET rtx.
8379 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8381 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8382 (cortex-a57): Likewise.
8383 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8385 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8387 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8388 non-iwmmxt builtins.
8390 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8394 * ipa-devirt.c record_target_from_binfo): Take as argument
8395 stack of binfos and lookup matching one for virtual inheritance.
8396 (possible_polymorphic_call_targets_1): Update.
8398 2014-01-10 Huacai Chen <chenhc@lemote.com>
8400 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8401 kernel strings for Loongson-2E/2F/3A.
8403 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8406 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8407 is_gimple_call before calling gimple_call_internal_p.
8409 2014-01-09 Steve Ellcey <sellcey@mips.com>
8411 * Makefile.in (TREE_FLOW_H): Remove.
8412 (TREE_SSA_H): Add file names from tree-flow.h.
8413 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8414 * tree.h: Remove tree-flow.h reference.
8415 * hash-table.h: Remove tree-flow.h reference.
8416 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8417 reference with tree-ssa-loop.h.
8419 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8421 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8422 default element-order behavior for -maltivec.
8423 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8424 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8425 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8426 when targeting big endian, at least for now.
8427 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8429 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8432 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8433 var satisfies use_register_for_decl, just take into account type
8434 alignment, rather than decl alignment.
8436 PR tree-optimization/59622
8437 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8438 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8439 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8440 Don't devirtualize for inplace at all. For targets.length () == 1,
8441 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8443 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8445 * config/i386/i386.md (cpu): Remove the unused btver1.
8447 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8449 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8451 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8454 * tree-core.h (struct target_globals): New forward declaration.
8455 (struct tree_target_option): Add globals field.
8456 * tree.h (TREE_TARGET_GLOBALS): Define.
8457 (prepare_target_option_nodes_for_pch): New prototype.
8458 * target-globals.h (struct target_globals): Define even if
8460 * tree.c (prepare_target_option_node_for_pch,
8461 prepare_target_option_nodes_for_pch): New functions.
8462 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8463 * config/i386/i386.c: Include target-globals.h.
8464 (ix86_set_current_function): Instead of doing target_reinit
8465 unconditionally, use save_target_globals_default_opts and
8466 restore_target_globals.
8468 2014-01-09 Richard Biener <rguenther@suse.de>
8470 PR tree-optimization/59715
8471 * tree-cfg.h (split_critical_edges): Declare.
8472 * tree-cfg.c (split_critical_edges): Export.
8473 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8475 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8477 * cfgexpand.c (expand_stack_vars): Optionally disable
8478 asan stack protection.
8479 (expand_used_vars): Likewise.
8480 (partition_stack_vars): Likewise.
8481 * asan.c (asan_emit_stack_protection): Optionally disable
8482 after return stack usage.
8483 (instrument_derefs): Optionally disable memory access instrumentation.
8484 (instrument_builtin_call): Likewise.
8485 (instrument_strlen_call): Likewise.
8486 (asan_protect_global): Optionally disable global variables protection.
8487 * doc/invoke.texi: Added doc for new options.
8488 * params.def: Added new options.
8489 * params.h: Likewise.
8491 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8493 PR rtl-optimization/59724
8494 * ifcvt.c (cond_exec_process_if_block): Don't call
8495 flow_find_head_matching_sequence with 0 longest_match.
8496 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8497 non-active insns if !stop_after.
8498 (try_head_merge_bb): Revert 2014-01-07 changes.
8500 2014-01-08 Jeff Law <law@redhat.com>
8502 * ree.c (get_sub_rtx): New function, extracted from...
8503 (merge_def_and_ext): Here.
8504 (combine_reaching_defs): Use get_sub_rtx.
8506 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8508 * cgraph.h (varpool_variable_node): Do not choke on null node.
8510 2014-01-08 Catherine Moore <clm@codesourcery.com>
8512 * config/mips/mips.md (simple_return): Attempt to use JRC
8514 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8516 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8518 PR rtl-optimization/59137
8519 * reorg.c (steal_delay_list_from_target): Call update_block for
8521 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8523 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8525 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8526 two duplicate entries.
8528 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8531 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8533 * config/mips/mips.c (mips_truncated_op_cost): New function.
8534 (mips_rtx_costs): Adjust test for BADDU.
8535 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8537 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8539 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8540 (*baddu_si): ...this new pattern.
8542 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8545 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8547 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8550 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8552 (expand_expr, expand_normal): Adjust.
8553 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8554 inner_reference_p. Use inner_reference_p to expand inner references.
8555 (store_expr): Adjust.
8556 * cfgexpand.c (expand_call_stmt): Adjust.
8558 2014-01-08 Rong Xu <xur@google.com>
8560 * gcov-io.c (gcov_var): Move from gcov-io.h.
8561 (gcov_position): Ditto.
8562 (gcov_is_error): Ditto.
8563 (gcov_rewrite): Ditto.
8564 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8565 only part to libgcc/libgcov.h.
8567 2014-01-08 Marek Polacek <polacek@redhat.com>
8570 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8572 2014-01-08 Marek Polacek <polacek@redhat.com>
8575 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8577 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8579 PR rtl-optimization/59649
8580 * stor-layout.c (get_mode_bounds): For BImode return
8581 0 and STORE_FLAG_VALUE.
8583 2014-01-08 Richard Biener <rguenther@suse.de>
8586 * gimple.h (is_gimple_builtin_call): Remove.
8587 (gimple_builtin_call_types_compatible_p): New.
8588 (gimple_call_builtin_p): New overload.
8589 * gimple.c (is_gimple_builtin_call): Remove.
8590 (validate_call): Rename to ...
8591 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8593 (validate_type): New static function.
8594 (gimple_call_builtin_p): New overload and adjust.
8595 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8596 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8597 (gimple_fold_stmt_to_constant_1): Likewise.
8598 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8600 2014-01-08 Richard Biener <rguenther@suse.de>
8603 * gimplify.c (gimplify_expr): Gimplify register-register type
8604 VIEW_CONVERT_EXPRs to separate stmts.
8606 2014-01-07 Jeff Law <law@redhat.com>
8609 * ree.c (combine_set_extension): Handle case where source
8610 and destination registers in an extension insn are different.
8611 (combine_reaching_defs): Allow source and destination registers
8612 in extension to be different under limited circumstances.
8613 (add_removable_extension): Remove restriction that the
8614 source and destination registers in the extension are the same.
8615 (find_and_remove_re): Emit a copy from the extension's
8616 destination to its source after the defining insn if
8617 the source and destination registers are different.
8620 * ifcvt.c (merge_if_block): If we are merging a block with more than
8621 one successor with a block with no successors, remove any BARRIER
8622 after the second block.
8624 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8626 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8628 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8631 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8632 for 14-bit register offsets when INT14_OK_STRICT is false.
8634 2014-01-07 Roland Stigge <stigge@antcom.de>
8635 Michael Meissner <meissner@linux.vnet.ibm.com>
8638 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8639 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8641 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8643 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8646 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8648 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8649 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8650 rtx is const0_rtx or not.
8652 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8655 * target-globals.c (save_target_globals): Remove this_fn_optab
8657 * toplev.c: Include optabs.h.
8658 (target_reinit): Temporarily restore the global options if another
8659 set of options are in force.
8661 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8663 PR rtl-optimization/58668
8664 * cfgcleanup.c (flow_find_cross_jump): Don't count
8665 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8666 to determine what is counted.
8667 (flow_find_head_matching_sequence): Use active_insn_p to determine
8669 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8671 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8672 determine what is counted.
8674 PR tree-optimization/59643
8675 * tree-predcom.c (split_data_refs_to_components): If one dr is
8676 read and one write, determine_offset fails and the write isn't
8677 in the bad component, just put the read into the bad component.
8679 2014-01-07 Mike Stump <mikestump@comcast.net>
8680 Jakub Jelinek <jakub@redhat.com>
8683 * tree-core.h (struct tree_optimization_option): Change optabs
8684 type from unsigned char * to void *.
8685 * optabs.c (init_tree_optimization_optabs): Adjust
8686 TREE_OPTIMIZATION_OPTABS initialization.
8688 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8691 * config/i386/i386.h (struct machine_function): Add
8692 no_drap_save_restore field.
8693 * config/i386/i386.c (ix86_save_reg): Use
8694 !cfun->machine->no_drap_save_restore instead of
8695 crtl->stack_realign_needed.
8696 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8697 this function clears frame_pointer_needed. Set
8698 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8699 and DRAP reg is needed.
8701 2014-01-06 Marek Polacek <polacek@redhat.com>
8704 * doc/implement-c.texi: Mention that other integer types are
8705 permitted as bit-field types in strictly conforming mode.
8707 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8709 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8712 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8714 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8716 2014-01-06 Martin Jambor <mjambor@suse.cz>
8719 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8721 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8723 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8727 * var-tracking.c (add_stores): Preserve the value of the source even if
8728 we don't record the store.
8730 2014-01-06 Terry Guo <terry.guo@arm.com>
8732 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8734 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8737 * config/darwin.c (darwin_function_section): Adjust return values to
8738 correspond to optimisation changes made in r206070.
8740 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8742 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8743 from prefetch_block tune setting.
8744 (nocona_cost): Correct size of prefetch block to 64.
8746 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8748 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8749 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8750 used to save the static chain register in the computation of the offset
8751 from which the FP registers need to be restored.
8753 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8755 PR tree-optimization/59519
8756 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8757 ICE if get_current_def (current_new_name) is already non-NULL, as long
8758 as it is a phi result of some other phi in *new_exit_bb that has
8761 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8762 or vmovdqu* for misaligned_operand.
8763 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8764 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8765 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8766 aligned_mem for AVX512F masked aligned load and store builtins and for
8769 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8771 PR tree-optimization/59651
8772 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8773 Address range for negative step should be added by TYPE_SIZE_UNIT.
8775 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8777 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8778 overlapping registers also for registers other than the stack pointer.
8780 2014-01-03 Marek Polacek <polacek@redhat.com>
8783 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8786 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8789 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8792 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8793 (push splitter): Use <P:MODE_SIZE> instead of
8794 GET_MODE_SIZE (<P:MODE>mode).
8795 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8796 (mov -1, reg peephole2): Likewise.
8797 * config/i386/sse.md (*mov<mode>_internal,
8798 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8799 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8800 *<code><mode>3, *andnot<mode>3<mask_name>,
8801 <mask_codefor><code><mode>3<mask_name>): Likewise.
8802 * config/i386/subst.md (mask_mode512bit_condition,
8803 sd_mask_mode512bit_condition): Likewise.
8805 2014-01-02 Xinliang David Li <davidxl@google.com>
8807 PR tree-optimization/59303
8808 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8809 (dump_predicates): Better output format.
8810 (pred_equal_p): New function.
8811 (is_neq_relop_p): Ditto.
8812 (is_neq_zero_form_p): Ditto.
8813 (pred_expr_equal_p): Ditto.
8814 (pred_neg_p): Ditto.
8815 (simplify_pred): Ditto.
8816 (simplify_preds_2): Ditto.
8817 (simplify_preds_3): Ditto.
8818 (simplify_preds_4): Ditto.
8819 (simplify_preds): Ditto.
8821 (push_to_worklist): Ditto.
8822 (get_pred_info_from_cmp): Ditto.
8823 (is_degenerated_phi): Ditto.
8824 (normalize_one_pred_1): Ditto.
8825 (normalize_one_pred): Ditto.
8826 (normalize_one_pred_chain): Ditto.
8827 (normalize_preds): Ditto.
8828 (normalize_cond_1): Remove function.
8829 (normalize_cond): Ditto.
8830 (is_gcond_subset_of): Ditto.
8831 (is_subset_of_any): Ditto.
8832 (is_or_set_subset_of): Ditto.
8833 (is_and_set_subset_of): Ditto.
8834 (is_norm_cond_subset_of): Ditto.
8835 (pred_chain_length_cmp): Ditto.
8836 (convert_control_dep_chain_into_preds): Type change.
8837 (find_predicates): Ditto.
8838 (find_def_preds): Ditto.
8839 (destroy_predicates_vecs): Ditto.
8840 (find_matching_predicates_in_rest_chains): Ditto.
8841 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8842 (is_pred_expr_subset): Ditto.
8843 (is_pred_chain_subset_of): Ditto.
8844 (is_included_in): Ditto.
8845 (is_superset_of): Ditto.
8847 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8849 Update copyright years.
8851 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8853 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8854 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8855 config/arc/arc.md, config/arc/arc.opt,
8856 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8857 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8858 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8859 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8860 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8861 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8862 vtable-verify.c, vtable-verify.h: Use the standard form for the
8865 2014-01-02 Tobias Burnus <burnus@net-b.de>
8867 * gcc.c (process_command): Update copyright notice dates.
8868 * gcov-dump.c: Ditto.
8870 * doc/cpp.texi: Bump @copying's copyright year.
8871 * doc/cppinternals.texi: Ditto.
8872 * doc/gcc.texi: Ditto.
8873 * doc/gccint.texi: Ditto.
8874 * doc/gcov.texi: Ditto.
8875 * doc/install.texi: Ditto.
8876 * doc/invoke.texi: Ditto.
8878 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8880 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8882 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8884 * config/i386/sse.md (*mov<mode>_internal): Guard
8885 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8887 PR rtl-optimization/59647
8888 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8889 new_rtx into UNSIGNED_FLOAT rtxes.
8891 Copyright (C) 2014 Free Software Foundation, Inc.
8893 Copying and distribution of this file, with or without modification,
8894 are permitted in any medium without royalty provided the copyright
8895 notice and this notice are preserved.