arc.c (TARGET_LRA_P): Undef before redefine.
[gcc.git] / gcc / ChangeLog
1 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2
3 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
4
5 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6
7 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
8
9 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
10
11 * tree-if-conv.c (is_cond_scalar_reduction): New function.
12 (convert_scalar_cond_reduction): Likewise.
13 (predicate_scalar_phi): Add recognition and transformation
14 of simple conditioanl reduction to be vectorizable.
15
16 2014-05-01 Marek Polacek <polacek@redhat.com>
17
18 PR c/43245
19 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
20
21 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
22
23 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
24 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
25 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
26 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
27 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
28 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
29 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
30 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
31
32 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
33
34 * config/arc/arc.opt (mlra): Move comment above option name
35 to avoid mis-parsing as language options.
36
37 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
38
39 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
40 * config/sol2.h: ... here.
41 * config/sol2-10.h: Remove.
42
43 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
44 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
45 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
46 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
47 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
48 * config/sol2.h: ... here.
49 (SECTION_NAME_FORMAT): Don't redefine.
50 (STARTFILE_ARCH32_SPEC): Rename to ...
51 (STARTFILE_ARCH_SPEC): ... this.
52 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
53 * config/sparc/sol2.h: ... here.
54 (SECTION_NAME_FORMAT): Don't undef.
55 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
56 (SUBTARGET_EXTRA_SPECS): Remove.
57 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
58
59 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
60 (MD_STARTFILE_PREFIX): Remove.
61 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
62 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
63 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
64 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
65 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
66 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
67 * config/i386/sol2.h: ... here.
68 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
69 * config/i386/sol2-bi.h: Remove.
70 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
71 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
72
73 * config/i386/t-sol2-64: Rename to ...
74 * config/i386/t-sol2: ... this.
75 * config/sparc/t-sol2-64: Rename to ...
76 * config/sparc/t-sol2: ... this.
77
78 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
79 sol2_tm_file_head, sol2_tm_file_tail.
80 Include ${cpu_type}/sol2.h before sol2.h.
81 Remove sol2-10.h.
82 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
83 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
84 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
85 Reflect i386/t-sol2-64 renaming.
86 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
87 Reflect sparc/t-sol2-64 renaming.
88
89 2014-04-30 Richard Biener <rguenther@suse.de>
90
91 * passes.c (execute_function_todo): Move TODO_verify_stmts
92 and TODO_verify_ssa under the TODO_verify_il umbrella.
93 * tree-ssa.h (verify_ssa): Adjust prototype.
94 * tree-ssa.c (verify_ssa): Add parameter to tell whether
95 we should verify SSA operands.
96 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
97 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
98 whether we should verify whether not throwing stmts have EH info.
99 * graphite-scop-detection.c (create_sese_edges): Adjust.
100 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
101 * tree-eh.c (lower_try_finally_switch): Do not add the
102 default case label twice.
103
104 2014-04-30 Marek Polacek <polacek@redhat.com>
105
106 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
107 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
108 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
109 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
110
111 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
112
113 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
114 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
115 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
116 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
117 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
118 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
119 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
120 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
121
122 2014-04-29 David Malcolm <dmalcolm@redhat.com>
123
124 * tree-cfg.c (dump_function_to_file): Dump the return type of
125 functions, in a line to itself before the function body, mimicking
126 the layout of a C function.
127
128 2014-04-29 Jakub Jelinek <jakub@redhat.com>
129
130 PR tree-optimization/60971
131 * tree-tailcall.c (process_assignment): Reject conversions which
132 reduce precision.
133
134 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
135
136 * calls.c (initialize_argument_information): Always treat
137 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
138 (expand_call): Likewise.
139 (emit_library_call_calue_1): Likewise.
140 * expr.c (PUSH_ARGS_REVERSED): Do not define.
141 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
142 code accordingly.
143
144 2014-04-29 Nick Clifton <nickc@redhat.com>
145
146 * config/msp430/msp430.md (umulsidi): Fix typo.
147 (mulhisi3): Enable even inside interrupt handlers.
148 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
149 bigger return address pushed in large mode.
150
151 2014-04-29 Nick Clifton <nickc@redhat.com>
152
153 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
154 (arc_init_reg_tables): Use a machine_mode enum to iterate over
155 available modes.
156 * config/m32r/m32r.c (init_reg_tables): Likewise.
157 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
158 enum to hold the modes.
159
160 2014-04-29 Richard Biener <rguenther@suse.de>
161
162 * dominance.c (free_dominance_info): Add overload with
163 function parameter.
164 (dom_info_state): Likewise.
165 (dom_info_available_p): Likewise.
166 * basic-block.h (free_dominance_info, dom_info_state,
167 dom_info_available_p): Declare overloads.
168 * passes.c (execute_function_todo): Verify that verifiers
169 don't change dominator info state. Drop dominator info
170 for IPA pass invocations.
171 * cgraph.c (release_function_body): Restore asserts that
172 dominator information is released.
173
174 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
175
176 * doc/invoke.texi: Fix typo.
177 * tree-vrp.c: Fix typos.
178 * gimple.c (infer_nonnull_range): Reorder operands of an &&
179 condition.
180
181 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
182
183 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
184
185 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
186
187 * config/aarch64/aarch64-builtins.c
188 (aarch64_types_storestruct_lane_qualifiers): New.
189 (TYPES_STORESTRUCT_LANE): Likewise.
190 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
191 (st3_lane): Likewise.
192 (st4_lane): Likewise.
193 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
194 (vec_store_lanesci_lane<mode>): Likewise.
195 (vec_store_lanesxi_lane<mode>): Likewise.
196 (aarch64_st2_lane<VQ:mode>): Likewise.
197 (aarch64_st3_lane<VQ:mode>): Likewise.
198 (aarch64_st4_lane<VQ:mode>): Likewise.
199 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
200 * config/aarch64/arm_neon.h
201 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
202 use new macro arguments.
203 (__ST3_LANE_FUNC): Likewise.
204 (__ST4_LANE_FUNC): Likewise.
205 * config/aarch64/iterators.md (V_TWO_ELEM): New.
206 (V_THREE_ELEM): Likewise.
207 (V_FOUR_ELEM): Likewise.
208
209 2014-04-28 David Malcolm <dmalcolm@redhat.com>
210
211 * doc/gimple.texi: Replace the description of the now-defunct
212 union gimple_statement_d with a diagram showing the
213 gimple_statement_base class hierarchy and its relationships to
214 the GSS_ and GIMPLE_ enums.
215
216 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
217
218 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
219 * config/aarch64/aarch64.c
220 (aarch64_cannot_change_mode_class): Weaken conditions.
221 (aarch64_modes_tieable_p): New.
222 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
223
224 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
225
226 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
227 (loadsync_<mode>): Change mode.
228 (load_quadpti, store_quadpti): New.
229 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
230 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
231
232 2014-04-28 Martin Jambor <mjambor@suse.cz>
233
234 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
235 same alias type as the original statement.
236 (subreplacement_assignment_data): New type.
237 (handle_unscalarized_data_in_subtree): New type of parameter,
238 generate new memory accesses with same alias type as the original
239 statement.
240 (load_assign_lhs_subreplacements): Likewise.
241 (sra_modify_constructor_assign): Generate new memory accesses with
242 same alias type as the original statement.
243
244 2014-04-28 Richard Biener <rguenther@suse.de>
245
246 * tree-pass.h (TODO_verify_il): Define.
247 (TODO_verify_all): Complete properly.
248 * passes.c (execute_function_todo): Move existing loop-closed
249 SSA verification under TODO_verify_il.
250 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
251 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
252 Fix tree sharing issue.
253
254 2014-04-28 Richard Biener <rguenther@suse.de>
255
256 PR middle-end/60092
257 * builtins.def (DEF_C11_BUILTIN): Add.
258 (BUILT_IN_ALIGNED_ALLOC): Likewise.
259 * coretypes.h (enum function_class): Add function_c11_misc.
260 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
261 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
262 (call_may_clobber_ref_p_1): Likewise.
263 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
264 (mark_all_reaching_defs_necessary_1): Likewise.
265 (propagate_necessity): Likewise.
266 (eliminate_unnecessary_stmts): Likewise.
267 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
268
269 2014-04-28 Richard Biener <rguenther@suse.de>
270
271 * tree-vrp.c (vrp_var_may_overflow): Remove.
272 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
273 with overflow immediately bump to one before that value and
274 let iteration figure out overflow status.
275
276 2014-04-28 Richard Biener <rguenther@suse.de>
277
278 * configure.ac: Do valgrind header checks unconditionally.
279 Add --enable-valgrind-annotations.
280 * system.h: Guard valgrind header inclusion with
281 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
282 * alloc-pool.c (pool_alloc, pool_free): Use
283 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
284 to guard possibly dead code.
285 * config.in: Regenerated.
286 * configure: Likewise.
287
288 2014-04-28 Jeff Law <law@redhat.com>
289
290 PR tree-optimization/60902
291 * tree-ssa-threadedge.c
292 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
293 over real defs when invalidating outputs from statements that do not
294 produce useful outputs for threading.
295
296 2014-04-28 Richard Biener <rguenther@suse.de>
297
298 PR tree-optimization/60979
299 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
300 SCOPs that end in a block with a successor with abnormal
301 predecessors.
302
303 2014-04-28 Richard Biener <rguenther@suse.de>
304
305 * tree-pass.h (execute_pass_list): Adjust prototype.
306 * passes.c (pass_manager::execute_early_local_passes):
307 Adjust.
308 (do_per_function): Change callback signature, push all actual
309 work to the callbals.
310 (do_per_function_toporder): Likewise.
311 (execute_function_dump): Adjust.
312 (execute_function_todo): Likewise.
313 (clear_last_verified): Likewise.
314 (verify_curr_properties): Likewise.
315 (update_properties_after_pass): Likewise.
316 (execute_pass_list_1): Split out from ...
317 (execute_pass_list): ... here. Adjust.
318 (execute_ipa_pass_list): Likewise.
319 * cgraphunit.c (cgraph_add_new_function): Adjust.
320 (analyze_function): Likewise.
321 (expand_function): Likewise.
322 * cgraph.c (release_function_body): Free dominance info
323 here instead of asserting it was magically freed elsewhere.
324
325 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
326
327 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
328 * configure: Regenerate.
329 * config/sparc/sparc.opt (muser-mode): New option.
330 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
331 for LEON3.
332 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
333 * doc/invoke.texi (SPARC options): Document -muser-mode.
334
335 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
336
337 * cselib.c (find_slot_memmode): Delete.
338 (cselib_hasher): Change compare_type to a struct.
339 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
340 constants.
341 (preserve_constants_and_equivs): Adjust for new compare_type.
342 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
343 (wrap_constant): Delete.
344 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
345
346 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
347
348 * doc/install.texi (Building with profile feedback): Remove
349 outdated sentence.
350
351 2014-04-26 Tom de Vries <tom@codesourcery.com>
352
353 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
354 array accesses.
355
356 2014-04-25 Cary Coutant <ccoutant@google.com>
357
358 PR debug/60929
359 * dwarf2out.c (should_move_die_to_comdat): A type definition
360 can contain a subprogram definition, but don't move it to a
361 comdat unit.
362 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
363 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
364 from original DIE.
365 (clone_tree_hash): Rename to...
366 (clone_tree_partial): ...this; change callers. Copy
367 DW_TAG_subprogram DIEs as declarations.
368 (copy_decls_walk): Don't copy children of a declaration into a
369 type unit.
370
371 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
372
373 PR target/60969
374 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
375 alternative 12.
376
377 2014-04-25 Jiong Wang <jiong.wang@arm.com>
378
379 * config/arm/predicates.md (call_insn_operand): Add long_call check.
380 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
381 reg for long_call.
382 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
383 restriction.
384
385 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
386
387 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
388 T16-related fields.
389
390 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
391
392 PR tree-optimization/60930
393 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
394 creating a multiply candidate by folding two constant
395 multiplicands when the result overflows.
396
397 2014-04-25 Jakub Jelinek <jakub@redhat.com>
398
399 PR tree-optimization/60960
400 * tree-vect-generic.c (expand_vector_operation): Only call
401 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
402
403 2014-04-25 Tom de Vries <tom@codesourcery.com>
404
405 * expr.c (clobber_reg_mode): New function.
406 * expr.h (clobber_reg): New function.
407
408 2014-04-25 Tom de Vries <tom@codesourcery.com>
409
410 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
411 clobbers.
412
413 2014-04-25 Radovan Obradovic <robradovic@mips.com>
414 Tom de Vries <tom@codesourcery.com>
415
416 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
417 handle.
418 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
419 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
420 new argument to find_all_hard_reg_sets call.
421
422 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
423
424 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
425 Use HOST_WIDE_INT_C for mask literal.
426 (aarch_rev16_shleft_mask_imm_p): Likewise.
427
428 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
429
430 PR target/60941
431 * config/sparc/sparc.md (ashlsi3_extend): Delete.
432
433 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
434
435 PR preprocessor/56540
436 * config/i386/i386-c.c (ix86_target_macros): Define
437 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
438
439 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
440
441 * configure.ac (tga_func): Remove.
442 (LIB_TLS_SPEC): Remove.
443 * configure: Regenerate.
444 * config.in: Regenerate.
445 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
446
447 2014-04-25 Richard Biener <rguenther@suse.de>
448
449 PR ipa/60912
450 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
451 call stmt use/clobber sets during stmt walk instead of
452 walking the possibly incomplete set of caller edges.
453
454 2014-04-25 Richard Biener <rguenther@suse.de>
455
456 PR ipa/60911
457 * passes.c (apply_ipa_transforms): Inline into only caller ...
458 (execute_one_pass): ... here. Properly bring in function
459 bodies for nodes we want to apply IPA transforms to.
460
461 2014-04-24 Cong Hou <congh@google.com>
462
463 PR tree-optimization/60896
464 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
465 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
466 (vect_mark_pattern_stmts): Set the def type of all statements in
467 PATTERN_DEF_SEQ as vect_internal_def.
468
469 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
470
471 * doc/extend.texi (PowerPC Built-in Functions): Document new
472 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
473 functions.
474 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
475
476 * config/rs6000/predicates.md (const_0_to_3_operand): New
477 predicate to match 0..3 integer constants.
478
479 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
480 to support adding miscellaneous builtin functions.
481 (BU_DFP_MISC_2): Likewise.
482 (BU_P7_MISC_1): Likewise.
483 (BU_P7_MISC_2): Likewise.
484 (BU_P8V_MISC_3): Likewise.
485 (BU_MISC_1): Likewise.
486 (BU_MISC_2): Likewise.
487 (DIVWE): Add extended divide builtin functions.
488 (DIVWEO): Likewise.
489 (DIVWEU): Likewise.
490 (DIVWEUO): Likewise.
491 (DIVDE): Likewise.
492 (DIVDEO): Likewise.
493 (DIVDEU): Likewise.
494 (DIVDEUO): Likewise.
495 (DXEX): Add decimal floating-point builtin functions.
496 (DXEXQ): Likewise.
497 (DDEDPD): Likewise.
498 (DDEDPDQ): Likewise.
499 (DENBCD): Likewise.
500 (DENBCDQ): Likewise.
501 (DIEX): Likewise.
502 (DIEXQ): Likewise.
503 (DSCLI): Likewise.
504 (DSCLIQ): Likewise.
505 (DSCRI): Likewise.
506 (DSCRIQ): Likewise.
507 (CDTBCD): Add new BCD builtin functions.
508 (CBCDTD): Likewise.
509 (ADDG6S): Likewise.
510 (BCDADD): Likewise.
511 (BCDADD_LT): Likewise.
512 (BCDADD_EQ): Likewise.
513 (BCDADD_GT): Likewise.
514 (BCDADD_OV): Likewise.
515 (BCDSUB): Likewise.
516 (BCDSUB_LT): Likewise.
517 (BCDSUB_EQ): Likewise.
518 (BCDSUB_GT): Likewise.
519 (BCDSUB_OV): Likewise.
520 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
521 (UNPACK_TD): Likewise.
522 (PACK_TF): Likewise.
523 (UNPACK_TF): Likewise.
524 (UNPACK_TF_0): Likewise.
525 (UNPACK_TF_1): Likewise.
526 (PACK_V1TI): Likewise.
527 (UNPACK_V1TI): Likewise.
528
529 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
530 support for decimal floating point builtin functions.
531 (rs6000_expand_ternop_builtin): Add checks for the new builtin
532 functions that take constant arguments.
533 (rs6000_invalid_builtin): Add decimal floating point builtin
534 support.
535 (rs6000_init_builtins): Setup long double, _Decimal64, and
536 _Decimal128 types for new builtin functions.
537 (builtin_function_type): Set the unsigned flags appropriately for
538 the new builtin functions.
539 (rs6000_opt_masks): Add support for decimal floating point builtin
540 functions.
541
542 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
543 floating point builtin functions.
544 (RS6000_BTM_COMMON): Likewise.
545 (RS6000_BTI_long_double): Likewise.
546 (RS6000_BTI_dfloat64): Likewise.
547 (RS6000_BTI_dfloat128): Likewise.
548 (long_double_type_internal_node): Likewise.
549 (dfloat64_type_internal_node): Likewise.
550 (dfloat128_type_internal_node): Likewise.
551
552 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
553 2.07 bcd arithmetic instructions.
554 (UNSPEC_BCDSUB): Likewise.
555 (UNSPEC_BCD_OVERFLOW): Likewise.
556 (UNSPEC_BCD_ADD_SUB): Likewise.
557 (bcd_add_sub): Likewise.
558 (BCD_TEST): Likewise.
559 (bcd<bcd_add_sub>): Likewise.
560 (bcd<bcd_add_sub>_test): Likewise.
561 (bcd<bcd_add_sub>_test2): Likewise.
562 (bcd<bcd_add_sub>_<code>): Likewise.
563 (peephole2 for combined bcd ops): Likewise.
564
565 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
566 decimal floating point builtin functions.
567 (UNSPEC_DENBCD): Likewise.
568 (UNSPEC_DXEX): Likewise.
569 (UNSPEC_DIEX): Likewise.
570 (UNSPEC_DSCLI): Likewise.
571 (UNSPEC_DSCRI): Likewise.
572 (D64_D128): Likewise.
573 (dfp_suffix): Likewise.
574 (dfp_ddedpd_<mode>): Likewise.
575 (dfp_denbcd_<mode>): Likewise.
576 (dfp_dxex_<mode>): Likewise.
577 (dfp_diex_<mode>): Likewise.
578 (dfp_dscli_<mode>): Likewise.
579 (dfp_dscri_<mode>): Likewise.
580
581 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
582 builtin functions.
583 (UNSPEC_CDTBCD): Likewise.
584 (UNSPEC_CBCDTD): Likewise.
585 (UNSPEC_DIVE): Add support for new extended divide builtin
586 functions.
587 (UNSPEC_DIVEO): Likewise.
588 (UNSPEC_DIVEU): Likewise.
589 (UNSPEC_DIVEUO): Likewise.
590 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
591 pack/unpack 128-bit types.
592 (UNSPEC_PACK_128BIT): Likewise.
593 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
594 (udiv<mode>3): Use idiv_ldiv mode attribute.
595 (div<mode>3): Likewise.
596 (addg6s): Add new BCD builtin functions.
597 (cdtbcd): Likewise.
598 (cbcdtd): Likewise.
599 (UNSPEC_DIV_EXTEND): Add support for new extended divide
600 instructions.
601 (div_extend): Likewise.
602 (div<div_extend>_<mode>"): Likewise.
603 (FP128_64): Add support for new builtin functions to pack/unpack
604 128-bit types.
605 (unpack<mode>): Likewise.
606 (unpacktf_0): Likewise.
607 (unpacktf_1): Likewise.
608 (unpack<mode>_dm): Likewise.
609 (unpack<mode>_nodm): Likewise.
610 (pack<mode>): Likewise.
611 (unpackv1ti): Likewise.
612 (packv1ti): Likewise.
613
614 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
615
616 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
617 is disabled.
618
619 2014-04-24 Jakub Jelinek <jakub@redhat.com>
620
621 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
622 * gimplify.c (omp_is_private): Change last argument's type to int.
623 Only diagnose lastprivate if the simd argument is 1, only diagnose
624 linear if the simd argument is 2.
625 (gimplify_omp_for): Adjust omp_is_private callers. When adding
626 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
627 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
628 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
629 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
630 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
631 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
632 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
633 * tree-nested.c (convert_nonlocal_omp_clauses,
634 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
635
636 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
637
638 PR target/60822
639 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
640 operand 1.
641
642 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
643
644 * flag-types.h (enum ivar_visibility): Add.
645
646 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
647
648 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
649 function * argument.
650
651 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
652
653 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
654
655 2014-04-24 Radovan Obradovic <robradovic@mips.com>
656 Tom de Vries <tom@codesourcery.com>
657
658 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
659 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
660 reg-note.
661 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
662 * emit-rtl.c (try_split): Same.
663
664 2014-04-24 Radovan Obradovic <robradovic@mips.com>
665 Tom de Vries <tom@codesourcery.com>
666
667 * common.opt (fuse-caller-save): New option.
668
669 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
670
671 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
672 for big-endian.
673
674 2014-04-24 Richard Biener <rguenther@suse.de>
675
676 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
677 during TER and instead use the sepops interface for expanding
678 non-GIMPLE_SINGLE_RHS.
679
680 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
681
682 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
683 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
684
685 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
686
687 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
688 assembler 64-bit option.
689 * configure: Regenerate.
690
691 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
692
693 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
694 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
695 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
696 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
697 (TARGET_CRYPTO): Take TARGET_SIMD into account.
698
699 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
700
701 * config/aarch64/aarch64-builtins.c
702 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
703 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
704 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
705 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
706 builtins.
707 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
708 (Vrevsuff): New mode attribute.
709
710 2014-04-24 Terry Guo <terry.guo@arm.com>
711
712 * config/arm/arm.h (machine_function): Define variable
713 after_arm_reorg here.
714 * config/arm/arm.c (after_arm_reorg): Remove the definition.
715 (arm_split_constant): Update the way to access variable
716 after_arm_reorg.
717 (arm_reorg): Ditto.
718 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
719
720 2014-04-23 Tom de Vries <tom@codesourcery.com>
721
722 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
723
724 2014-04-23 David Malcolm <dmalcolm@redhat.com>
725
726 * is-a.h: Update comments to reflect the following changes to the
727 "pointerness" of the API, making the template parameter match the
728 return type, allowing use of is-a.h with typedefs of pointers.
729 (is_a_helper::cast): Return a T rather then a pointer to a T, so
730 that the return type matches the parameter to the is_a_helper.
731 (as_a): Likewise.
732 (dyn_cast): Likewise.
733
734 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
735 pointer from the is-a.h API.
736
737 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
738 (is_a_helper <cgraph_node *>::test): ...this, matching change to
739 is-a.h API.
740 (is_a_helper <varpool_node>::test): Likewise, convert to...
741 (is_a_helper <varpool_node *>::test): ...this.
742
743 (varpool_first_variable): Update for removal of implicit pointer
744 from the is-a.h API.
745 (varpool_next_variable): Likewise.
746 (varpool_first_static_initializer): Likewise.
747 (varpool_next_static_initializer): Likewise.
748 (varpool_first_defined_variable): Likewise.
749 (varpool_next_defined_variable): Likewise.
750 (cgraph_first_defined_function): Likewise.
751 (cgraph_next_defined_function): Likewise.
752 (cgraph_first_function): Likewise.
753 (cgraph_next_function): Likewise.
754 (cgraph_first_function_with_gimple_body): Likewise.
755 (cgraph_next_function_with_gimple_body): Likewise.
756 (cgraph_alias_target): Likewise.
757 (varpool_alias_target): Likewise.
758 (cgraph_function_or_thunk_node): Likewise.
759 (varpool_variable_node): Likewise.
760 (symtab_real_symbol_p): Likewise.
761 * cgraphunit.c (referred_to_p): Likewise.
762 (analyze_functions): Likewise.
763 (handle_alias_pairs): Likewise.
764 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
765 * gimple-ssa.h (gimple_vuse_op): Likewise.
766 (gimple_vdef_op): Likewise.
767 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
768 * gimple.c (gimple_build_asm_1): Likewise.
769 (gimple_build_try): Likewise.
770 (gimple_build_resx): Likewise.
771 (gimple_build_eh_dispatch): Likewise.
772 (gimple_build_omp_for): Likewise.
773 (gimple_omp_for_set_clauses): Likewise.
774
775 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
776 (is_a_helper <gimple_statement_asm *>::test): ...this.
777 (is_a_helper <gimple_statement_bind>::test): Convert to...
778 (is_a_helper <gimple_statement_bind *>::test): ...this.
779 (is_a_helper <gimple_statement_call>::test): Convert to...
780 (is_a_helper <gimple_statement_call *>::test): ...this.
781 (is_a_helper <gimple_statement_catch>::test): Convert to...
782 (is_a_helper <gimple_statement_catch *>::test): ...this.
783 (is_a_helper <gimple_statement_resx>::test): Convert to...
784 (is_a_helper <gimple_statement_resx *>::test): ...this.
785 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
786 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
787 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
788 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
789 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
790 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
791 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
792 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
793 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
794 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
795 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
796 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
797 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
798 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
799 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
800 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
801 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
802 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
803 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
804 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
805 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
806 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
807 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
808 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
809 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
810 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
811 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
812 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
813 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
814 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
815 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
816 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
817 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
818 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
819 (is_a_helper <gimple_statement_phi>::test): Convert to...
820 (is_a_helper <gimple_statement_phi *>::test): ...this.
821 (is_a_helper <gimple_statement_transaction>::test): Convert to...
822 (is_a_helper <gimple_statement_transaction *>::test): ...this.
823 (is_a_helper <gimple_statement_try>::test): Convert to...
824 (is_a_helper <gimple_statement_try *>::test): ...this.
825 (is_a_helper <gimple_statement_wce>::test): Convert to...
826 (is_a_helper <gimple_statement_wce *>::test): ...this.
827 (is_a_helper <const gimple_statement_asm>::test): Convert to...
828 (is_a_helper <const gimple_statement_asm *>::test): ...this.
829 (is_a_helper <const gimple_statement_bind>::test): Convert to...
830 (is_a_helper <const gimple_statement_bind *>::test): ...this.
831 (is_a_helper <const gimple_statement_call>::test): Convert to...
832 (is_a_helper <const gimple_statement_call *>::test): ...this.
833 (is_a_helper <const gimple_statement_catch>::test): Convert to...
834 (is_a_helper <const gimple_statement_catch *>::test): ...this.
835 (is_a_helper <const gimple_statement_resx>::test): Convert to...
836 (is_a_helper <const gimple_statement_resx *>::test): ...this.
837 (is_a_helper <const gimple_statement_eh_dispatch>::test):
838 Convert to...
839 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
840 ...this.
841 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
842 to...
843 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
844 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
845 Convert to...
846 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
847 ...this.
848 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
849 Convert to...
850 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
851 ...this.
852 (is_a_helper <const gimple_statement_omp_return>::test): Convert
853 to...
854 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
855 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
856 to...
857 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
858 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
859 to...
860 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
861 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
862 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
863 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
864 to...
865 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
866 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
867 to...
868 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
869 (is_a_helper <const gimple_statement_omp_target>::test): Convert
870 to...
871 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
872 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
873 to...
874 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
875 (is_a_helper <const gimple_statement_omp_single>::test): Convert
876 to...
877 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
878 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
879 to...
880 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
881 (is_a_helper <const gimple_statement_omp_task>::test): Convert
882 to...
883 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
884 (is_a_helper <const gimple_statement_phi>::test): Convert to...
885 (is_a_helper <const gimple_statement_phi *>::test): ...this.
886 (is_a_helper <const gimple_statement_transaction>::test): Convert
887 to...
888 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
889 (is_a_helper <const gimple_statement_with_ops>::test): Convert
890 to...
891 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
892 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
893 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
894 (is_a_helper <const gimple_statement_with_memory_ops>::test):
895 Convert to...
896 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
897 ...this.
898 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
899 to...
900 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
901
902 (gimple_use_ops): Update for removal of implicit pointer from the
903 is-a.h API.
904 (gimple_set_use_ops): Likewise.
905 (gimple_vuse): Likewise.
906 (gimple_vdef): Likewise.
907 (gimple_vuse_ptr): Likewise.
908 (gimple_vdef_ptr): Likewise.
909 (gimple_set_vuse): Likewise.
910 (gimple_set_vdef): Likewise.
911 (gimple_omp_return_set_lhs): Likewise.
912 (gimple_omp_return_lhs): Likewise.
913 (gimple_omp_return_lhs_ptr): Likewise.
914 (gimple_call_fntype): Likewise.
915 (gimple_call_set_fntype): Likewise.
916 (gimple_call_set_internal_fn): Likewise.
917 (gimple_call_use_set): Likewise.
918 (gimple_call_clobber_set): Likewise.
919 (gimple_bind_vars): Likewise.
920 (gimple_bind_set_vars): Likewise.
921 (gimple_bind_body_ptr): Likewise.
922 (gimple_bind_set_body): Likewise.
923 (gimple_bind_add_stmt): Likewise.
924 (gimple_bind_block): Likewise.
925 (gimple_bind_set_block): Likewise.
926 (gimple_asm_ninputs): Likewise.
927 (gimple_asm_noutputs): Likewise.
928 (gimple_asm_nclobbers): Likewise.
929 (gimple_asm_nlabels): Likewise.
930 (gimple_asm_input_op): Likewise.
931 (gimple_asm_input_op_ptr): Likewise.
932 (gimple_asm_output_op): Likewise.
933 (gimple_asm_output_op_ptr): Likewise.
934 (gimple_asm_set_output_op): Likewise.
935 (gimple_asm_clobber_op): Likewise.
936 (gimple_asm_set_clobber_op): Likewise.
937 (gimple_asm_label_op): Likewise.
938 (gimple_asm_set_label_op): Likewise.
939 (gimple_asm_string): Likewise.
940 (gimple_catch_types): Likewise.
941 (gimple_catch_types_ptr): Likewise.
942 (gimple_catch_handler_ptr): Likewise.
943 (gimple_catch_set_types): Likewise.
944 (gimple_catch_set_handler): Likewise.
945 (gimple_eh_filter_types): Likewise.
946 (gimple_eh_filter_types_ptr): Likewise.
947 (gimple_eh_filter_failure_ptr): Likewise.
948 (gimple_eh_filter_set_types): Likewise.
949 (gimple_eh_filter_set_failure): Likewise.
950 (gimple_eh_must_not_throw_fndecl): Likewise.
951 (gimple_eh_must_not_throw_set_fndecl): Likewise.
952 (gimple_eh_else_n_body_ptr): Likewise.
953 (gimple_eh_else_e_body_ptr): Likewise.
954 (gimple_eh_else_set_n_body): Likewise.
955 (gimple_eh_else_set_e_body): Likewise.
956 (gimple_try_eval_ptr): Likewise.
957 (gimple_try_cleanup_ptr): Likewise.
958 (gimple_try_set_eval): Likewise.
959 (gimple_try_set_cleanup): Likewise.
960 (gimple_wce_cleanup_ptr): Likewise.
961 (gimple_wce_set_cleanup): Likewise.
962 (gimple_phi_capacity): Likewise.
963 (gimple_phi_num_args): Likewise.
964 (gimple_phi_result): Likewise.
965 (gimple_phi_result_ptr): Likewise.
966 (gimple_phi_set_result): Likewise.
967 (gimple_phi_arg): Likewise.
968 (gimple_phi_set_arg): Likewise.
969 (gimple_resx_region): Likewise.
970 (gimple_resx_set_region): Likewise.
971 (gimple_eh_dispatch_region): Likewise.
972 (gimple_eh_dispatch_set_region): Likewise.
973 (gimple_omp_critical_name): Likewise.
974 (gimple_omp_critical_name_ptr): Likewise.
975 (gimple_omp_critical_set_name): Likewise.
976 (gimple_omp_for_clauses): Likewise.
977 (gimple_omp_for_clauses_ptr): Likewise.
978 (gimple_omp_for_set_clauses): Likewise.
979 (gimple_omp_for_collapse): Likewise.
980 (gimple_omp_for_index): Likewise.
981 (gimple_omp_for_index_ptr): Likewise.
982 (gimple_omp_for_set_index): Likewise.
983 (gimple_omp_for_initial): Likewise.
984 (gimple_omp_for_initial_ptr): Likewise.
985 (gimple_omp_for_set_initial): Likewise.
986 (gimple_omp_for_final): Likewise.
987 (gimple_omp_for_final_ptr): Likewise.
988 (gimple_omp_for_set_final): Likewise.
989 (gimple_omp_for_incr): Likewise.
990 (gimple_omp_for_incr_ptr): Likewise.
991 (gimple_omp_for_set_incr): Likewise.
992 (gimple_omp_for_pre_body_ptr): Likewise.
993 (gimple_omp_for_set_pre_body): Likewise.
994 (gimple_omp_parallel_clauses): Likewise.
995 (gimple_omp_parallel_clauses_ptr): Likewise.
996 (gimple_omp_parallel_set_clauses): Likewise.
997 (gimple_omp_parallel_child_fn): Likewise.
998 (gimple_omp_parallel_child_fn_ptr): Likewise.
999 (gimple_omp_parallel_set_child_fn): Likewise.
1000 (gimple_omp_parallel_data_arg): Likewise.
1001 (gimple_omp_parallel_data_arg_ptr): Likewise.
1002 (gimple_omp_parallel_set_data_arg): Likewise.
1003 (gimple_omp_task_clauses): Likewise.
1004 (gimple_omp_task_clauses_ptr): Likewise.
1005 (gimple_omp_task_set_clauses): Likewise.
1006 (gimple_omp_task_child_fn): Likewise.
1007 (gimple_omp_task_child_fn_ptr): Likewise.
1008 (gimple_omp_task_set_child_fn): Likewise.
1009 (gimple_omp_task_data_arg): Likewise.
1010 (gimple_omp_task_data_arg_ptr): Likewise.
1011 (gimple_omp_task_set_data_arg): Likewise.
1012 (gimple_omp_taskreg_clauses): Likewise.
1013 (gimple_omp_taskreg_clauses_ptr): Likewise.
1014 (gimple_omp_taskreg_set_clauses): Likewise.
1015 (gimple_omp_taskreg_child_fn): Likewise.
1016 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1017 (gimple_omp_taskreg_set_child_fn): Likewise.
1018 (gimple_omp_taskreg_data_arg): Likewise.
1019 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1020 (gimple_omp_taskreg_set_data_arg): Likewise.
1021 (gimple_omp_task_copy_fn): Likewise.
1022 (gimple_omp_task_copy_fn_ptr): Likewise.
1023 (gimple_omp_task_set_copy_fn): Likewise.
1024 (gimple_omp_task_arg_size): Likewise.
1025 (gimple_omp_task_arg_size_ptr): Likewise.
1026 (gimple_omp_task_set_arg_size): Likewise.
1027 (gimple_omp_task_arg_align): Likewise.
1028 (gimple_omp_task_arg_align_ptr): Likewise.
1029 (gimple_omp_task_set_arg_align): Likewise.
1030 (gimple_omp_single_clauses): Likewise.
1031 (gimple_omp_single_clauses_ptr): Likewise.
1032 (gimple_omp_single_set_clauses): Likewise.
1033 (gimple_omp_target_clauses): Likewise.
1034 (gimple_omp_target_clauses_ptr): Likewise.
1035 (gimple_omp_target_set_clauses): Likewise.
1036 (gimple_omp_target_child_fn): Likewise.
1037 (gimple_omp_target_child_fn_ptr): Likewise.
1038 (gimple_omp_target_set_child_fn): Likewise.
1039 (gimple_omp_target_data_arg): Likewise.
1040 (gimple_omp_target_data_arg_ptr): Likewise.
1041 (gimple_omp_target_set_data_arg): Likewise.
1042 (gimple_omp_teams_clauses): Likewise.
1043 (gimple_omp_teams_clauses_ptr): Likewise.
1044 (gimple_omp_teams_set_clauses): Likewise.
1045 (gimple_omp_sections_clauses): Likewise.
1046 (gimple_omp_sections_clauses_ptr): Likewise.
1047 (gimple_omp_sections_set_clauses): Likewise.
1048 (gimple_omp_sections_control): Likewise.
1049 (gimple_omp_sections_control_ptr): Likewise.
1050 (gimple_omp_sections_set_control): Likewise.
1051 (gimple_omp_for_set_cond): Likewise.
1052 (gimple_omp_for_cond): Likewise.
1053 (gimple_omp_atomic_store_set_val): Likewise.
1054 (gimple_omp_atomic_store_val): Likewise.
1055 (gimple_omp_atomic_store_val_ptr): Likewise.
1056 (gimple_omp_atomic_load_set_lhs): Likewise.
1057 (gimple_omp_atomic_load_lhs): Likewise.
1058 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1059 (gimple_omp_atomic_load_set_rhs): Likewise.
1060 (gimple_omp_atomic_load_rhs): Likewise.
1061 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1062 (gimple_omp_continue_control_def): Likewise.
1063 (gimple_omp_continue_control_def_ptr): Likewise.
1064 (gimple_omp_continue_set_control_def): Likewise.
1065 (gimple_omp_continue_control_use): Likewise.
1066 (gimple_omp_continue_control_use_ptr): Likewise.
1067 (gimple_omp_continue_set_control_use): Likewise.
1068 (gimple_transaction_body_ptr): Likewise.
1069 (gimple_transaction_label): Likewise.
1070 (gimple_transaction_label_ptr): Likewise.
1071 (gimple_transaction_set_body): Likewise.
1072 (gimple_transaction_set_label): Likewise.
1073
1074 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1075 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1076 * ipa-ref.c (ipa_record_reference): Likewise.
1077 * ipa-reference.c (analyze_function): Likewise.
1078 (ipa_reference_write_optimization_summary): Likewise.
1079 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1080 (address_taken_from_non_vtable_p): Likewise.
1081 (comdat_can_be_unshared_p_1): Likewise.
1082 * lto-cgraph.c (lto_output_ref): Likewise.
1083 (add_references): Likewise.
1084 (compute_ltrans_boundary): Likewise.
1085 (output_symtab): Likewise.
1086 (input_ref): Likewise.
1087 (input_cgraph_1): Likewise.
1088 (output_cgraph_opt_summary): Likewise.
1089 * lto-streamer-out.c (lto_output): Likewise.
1090 (output_symbol_p): Likewise.
1091 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1092 (lsei_start_function_in_partition): Likewise.
1093 (lsei_next_variable_in_partition): Likewise.
1094 (lsei_start_variable_in_partition): Likewise.
1095 * symtab.c (insert_to_assembler_name_hash): Likewise.
1096 (unlink_from_assembler_name_hash): Likewise.
1097 (symtab_unregister_node): Likewise.
1098 (symtab_remove_node): Likewise.
1099 (dump_symtab_node): Likewise.
1100 (verify_symtab_base): Likewise.
1101 (verify_symtab_node): Likewise.
1102 (symtab_make_decl_local): Likewise.
1103 (symtab_alias_ultimate_target): Likewise.
1104 (symtab_resolve_alias): Likewise.
1105 (symtab_get_symbol_partitioning_class): Likewise.
1106 * tree-phinodes.c (allocate_phi_node): Likewise.
1107 (reserve_phi_args_for_new_edge): Likewise.
1108 (remove_phi_args): Likewise.
1109 * varpool.c (varpool_node_for_asm): Likewise.
1110 (varpool_remove_unreferenced_decls): Likewise.
1111
1112 2014-04-23 Jeff Law <law@redhat.com>
1113
1114 PR tree-optimization/60902
1115 * tree-ssa-threadedge.c
1116 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1117 invalidate outputs from statements that do not produce useful
1118 outputs for threading.
1119
1120 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1121
1122 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1123 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1124 machine descriptions for Stack Smashing Protector.
1125
1126 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1127
1128 * aarch64.md (<optab>_rol<mode>3): New pattern.
1129 (<optab>_rolsi3_uxtw): Likewise.
1130 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1131
1132 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1133
1134 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1135 (arm_cortex_a12_tune): Likewise.
1136
1137 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1138
1139 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1140
1141 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1142
1143 * config/arm/arm.md (arm_rev16si2): New pattern.
1144 (arm_rev16si2_alt): Likewise.
1145 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1146
1147 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1148
1149 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1150 (rev16<mode>2_alt): Likewise.
1151 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1152 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1153 (aarch_rev16_shleft_mask_imm_p): Likewise.
1154 (aarch_rev16_p_1): Likewise.
1155 (aarch_rev16_p): Likewise.
1156 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1157 (aarch_rev16_shright_mask_imm_p): Likewise.
1158 (aarch_rev16_shleft_mask_imm_p): Likewise.
1159
1160 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1161
1162 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1163 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1164 rev cost.
1165 (cortex_a53_extra_costs): Likewise.
1166 (cortex_a57_extra_costs): Likewise.
1167 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1168 (cortexa7_extra_costs): Likewise.
1169 (cortexa8_extra_costs): Likewise.
1170 (cortexa12_extra_costs): Likewise.
1171 (cortexa15_extra_costs): Likewise.
1172 (v7m_extra_costs): Likewise.
1173 (arm_new_rtx_costs): Handle BSWAP.
1174
1175 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1176
1177 * config/arm/arm.c (cortexa8_extra_costs): New table.
1178 (arm_cortex_a8_tune): New tuning struct.
1179 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1180
1181 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1182
1183 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1184
1185 2014-04-23 Richard Biener <rguenther@suse.de>
1186
1187 * Makefile.in (OBJS): Remove loop-unswitch.o.
1188 * loop-unswitch.c: Delete.
1189 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1190 * passes.def (pass_rtl_unswitch): Likewise.
1191 * loop-init.c (gate_rtl_unswitch): Likewise.
1192 (rtl_unswitch): Likewise.
1193 (pass_data_rtl_unswitch): Likewise.
1194 (pass_rtl_unswitch): Likewise.
1195 (make_pass_rtl_unswitch): Likewise.
1196 * rtl.h (reversed_condition): Likewise.
1197 (compare_and_jump_seq): Likewise.
1198 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1199 and make static.
1200 * loop-unroll.c (compare_and_jump_seq): Likewise.
1201
1202 2014-04-23 Richard Biener <rguenther@suse.de>
1203
1204 PR tree-optimization/60903
1205 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1206 commented code block.
1207 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1208 loop flags to newly created BBs and edges.
1209
1210 2014-04-23 Nick Clifton <nickc@redhat.com>
1211
1212 * config/msp430/msp430.c (msp430_handle_option): Move function
1213 to msp430-common.c
1214 (msp430_option_override): Simplify mcu and mcpu option handling.
1215 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1216 support for -mhwmult command line option.
1217 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1218 -mhwmult command line option.
1219 (msp430_hwmult_enabled): Delete.
1220 (msp43o_output_labelref): Add support for -mhwmult command line
1221 option.
1222 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1223 (umulsidi3): Likewise.
1224 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1225 (mcpu, mlarge, msmall): Likewise.
1226 (mhwmult): New option.
1227 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1228 prototype.
1229 (msp430_is_f5_mcu): Remove prototype.
1230 (msp430_use_f5_series_hwmult): Add prototype.
1231 * config/msp430/msp430-opts.h: New file.
1232 * common/config/msp430: New directory.
1233 * common/config/msp430/msp430-common.c: New file.
1234 * config.gcc (msp430): Remove target_has_targetm_common.
1235 * doc/invoke.texi: Document -mhwmult command line option.
1236
1237 2014-04-23 Nick Clifton <nickc@redhat.com>
1238
1239 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1240 default-manifest.o if it can be found in the search path.
1241 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1242
1243 2014-04-23 Terry Guo <terry.guo@arm.com>
1244
1245 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1246
1247 2014-04-23 Richard Biener <rguenther@suse.de>
1248
1249 PR middle-end/60895
1250 * tree-inline.c (declare_return_variable): Use mark_addressable.
1251
1252 2014-04-23 Richard Biener <rguenther@suse.de>
1253
1254 PR middle-end/60891
1255 * loop-init.c (loop_optimizer_init): Make sure to apply
1256 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1257
1258 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1259
1260 PR sanitizer/60275
1261 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1262 New options.
1263 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1264 if flag_sanitize_undefined_trap_on_error.
1265 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1266 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1267 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1268 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1269 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1270 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1271 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1272 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1273 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1274 * ubsan.c (ubsan_instrument_unreachable): Return
1275 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1276 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1277 if flag_sanitize_undefined_trap_on_error and
1278 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1279 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1280 instrument_bool_enum_load): Emit __builtin_trap () if
1281 flag_sanitize_undefined_trap_on_error and
1282 __builtin_handle_*_abort () if !flag_sanitize_recover.
1283 * doc/invoke.texi (-fsanitize-recover,
1284 -fsanitize-undefined-trap-on-error): Document.
1285
1286 2014-04-22 Christian Bruel <christian.bruel@st.com>
1287
1288 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1289 Force immediates to SImode.
1290
1291 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1292
1293 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1294 (lroundsfsi2): New.
1295 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1296 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1297 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1298 (nios2_fpu_insn): Add entry for round.
1299 (N2FPU_NO_ERRNO_P): Define.
1300 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1301 flag_errno_math.
1302 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1303
1304 2014-04-22 Richard Henderson <rth@redhat.com>
1305
1306 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1307 (add<GPI>3_compare0): Remove leading * from name.
1308 (add<GPI>3_carryin): Likewise.
1309 (sub<GPI>3_compare0): Likewise.
1310 (sub<GPI>3_carryin): Likewise.
1311 (<su_optab>mulditi3): New expander.
1312 (multi3): New expander.
1313 (madd<GPI>): Remove leading * from name.
1314
1315 2014-04-22 Martin Jambor <mjambor@suse.cz>
1316
1317 * cgraphclones.c (cgraph_function_versioning): Copy
1318 ipa_transforms_to_apply instead of asserting it is empty.
1319
1320 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1321
1322 PR target/60868
1323 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1324 on count_exp to get mode.
1325
1326 2014-04-22 Andrew Pinski <apinski@cavium.com>
1327
1328 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1329 Handle TLS for ILP32.
1330 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1331 (tlsie_small_<mode>): this and handle PTR.
1332 (tlsie_small_sidi): New pattern.
1333 (tlsle_small): Change to an expand to handle ILP32.
1334 (tlsle_small_<mode>): New pattern.
1335 (tlsdesc_small): Rename to ...
1336 (tlsdesc_small_<mode>): this and handle PTR.
1337
1338 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1339
1340 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1341
1342 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1343
1344 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1345 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1346 (aarch64_types_signed_poly_qualifiers): Likewise.
1347 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1348 (aarch64_types_poly_signed_qualifiers): Likewise.
1349 (TYPES_REINTERP_SS): Type macro added.
1350 (TYPES_REINTERP_SU): Likewise.
1351 (TYPES_REINTERP_SP): Likewise.
1352 (TYPES_REINTERP_US): Likewise.
1353 (TYPES_REINTERP_PS): Likewise.
1354 (aarch64_fold_builtin): New expression folding added.
1355 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1356 Declarations removed.
1357 (REINTERP_SS): Declarations added.
1358 (REINTERP_US): Likewise.
1359 (REINTERP_PS): Likewise.
1360 (REINTERP_SU): Likewise.
1361 (REINTERP_SP): Likewise.
1362 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1363 (vreinterpretq_p8_f64): Likewise.
1364 (vreinterpret_p16_f64): Likewise.
1365 (vreinterpretq_p16_f64): Likewise.
1366 (vreinterpret_f32_f64): Likewise.
1367 (vreinterpretq_f32_f64): Likewise.
1368 (vreinterpret_f64_f32): Likewise.
1369 (vreinterpret_f64_p8): Likewise.
1370 (vreinterpret_f64_p16): Likewise.
1371 (vreinterpret_f64_s8): Likewise.
1372 (vreinterpret_f64_s16): Likewise.
1373 (vreinterpret_f64_s32): Likewise.
1374 (vreinterpret_f64_s64): Likewise.
1375 (vreinterpret_f64_u8): Likewise.
1376 (vreinterpret_f64_u16): Likewise.
1377 (vreinterpret_f64_u32): Likewise.
1378 (vreinterpret_f64_u64): Likewise.
1379 (vreinterpretq_f64_f32): Likewise.
1380 (vreinterpretq_f64_p8): Likewise.
1381 (vreinterpretq_f64_p16): Likewise.
1382 (vreinterpretq_f64_s8): Likewise.
1383 (vreinterpretq_f64_s16): Likewise.
1384 (vreinterpretq_f64_s32): Likewise.
1385 (vreinterpretq_f64_s64): Likewise.
1386 (vreinterpretq_f64_u8): Likewise.
1387 (vreinterpretq_f64_u16): Likewise.
1388 (vreinterpretq_f64_u32): Likewise.
1389 (vreinterpretq_f64_u64): Likewise.
1390 (vreinterpret_s64_f64): Likewise.
1391 (vreinterpretq_s64_f64): Likewise.
1392 (vreinterpret_u64_f64): Likewise.
1393 (vreinterpretq_u64_f64): Likewise.
1394 (vreinterpret_s8_f64): Likewise.
1395 (vreinterpretq_s8_f64): Likewise.
1396 (vreinterpret_s16_f64): Likewise.
1397 (vreinterpretq_s16_f64): Likewise.
1398 (vreinterpret_s32_f64): Likewise.
1399 (vreinterpretq_s32_f64): Likewise.
1400 (vreinterpret_u8_f64): Likewise.
1401 (vreinterpretq_u8_f64): Likewise.
1402 (vreinterpret_u16_f64): Likewise.
1403 (vreinterpretq_u16_f64): Likewise.
1404 (vreinterpret_u32_f64): Likewise.
1405 (vreinterpretq_u32_f64): Likewise.
1406
1407 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1408
1409 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1410 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1411 (vreinterpret_p8_s8): Likewise.
1412 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1413 (vreinterpret_p8_s16): Likewise.
1414 (vreinterpret_p8_s32): Likewise.
1415 (vreinterpret_p8_s64): Likewise.
1416 (vreinterpret_p8_f32): Likewise.
1417 (vreinterpret_p8_u8): Likewise.
1418 (vreinterpret_p8_u16): Likewise.
1419 (vreinterpret_p8_u32): Likewise.
1420 (vreinterpret_p8_u64): Likewise.
1421 (vreinterpret_p8_p16): Likewise.
1422 (vreinterpretq_p8_s8): Likewise.
1423 (vreinterpretq_p8_s16): Likewise.
1424 (vreinterpretq_p8_s32): Likewise.
1425 (vreinterpretq_p8_s64): Likewise.
1426 (vreinterpretq_p8_f32): Likewise.
1427 (vreinterpretq_p8_u8): Likewise.
1428 (vreinterpretq_p8_u16): Likewise.
1429 (vreinterpretq_p8_u32): Likewise.
1430 (vreinterpretq_p8_u64): Likewise.
1431 (vreinterpretq_p8_p16): Likewise.
1432 (vreinterpret_p16_s8): Likewise.
1433 (vreinterpret_p16_s16): Likewise.
1434 (vreinterpret_p16_s32): Likewise.
1435 (vreinterpret_p16_s64): Likewise.
1436 (vreinterpret_p16_f32): Likewise.
1437 (vreinterpret_p16_u8): Likewise.
1438 (vreinterpret_p16_u16): Likewise.
1439 (vreinterpret_p16_u32): Likewise.
1440 (vreinterpret_p16_u64): Likewise.
1441 (vreinterpret_p16_p8): Likewise.
1442 (vreinterpretq_p16_s8): Likewise.
1443 (vreinterpretq_p16_s16): Likewise.
1444 (vreinterpretq_p16_s32): Likewise.
1445 (vreinterpretq_p16_s64): Likewise.
1446 (vreinterpretq_p16_f32): Likewise.
1447 (vreinterpretq_p16_u8): Likewise.
1448 (vreinterpretq_p16_u16): Likewise.
1449 (vreinterpretq_p16_u32): Likewise.
1450 (vreinterpretq_p16_u64): Likewise.
1451 (vreinterpretq_p16_p8): Likewise.
1452 (vreinterpret_f32_s8): Likewise.
1453 (vreinterpret_f32_s16): Likewise.
1454 (vreinterpret_f32_s32): Likewise.
1455 (vreinterpret_f32_s64): Likewise.
1456 (vreinterpret_f32_u8): Likewise.
1457 (vreinterpret_f32_u16): Likewise.
1458 (vreinterpret_f32_u32): Likewise.
1459 (vreinterpret_f32_u64): Likewise.
1460 (vreinterpret_f32_p8): Likewise.
1461 (vreinterpret_f32_p16): Likewise.
1462 (vreinterpretq_f32_s8): Likewise.
1463 (vreinterpretq_f32_s16): Likewise.
1464 (vreinterpretq_f32_s32): Likewise.
1465 (vreinterpretq_f32_s64): Likewise.
1466 (vreinterpretq_f32_u8): Likewise.
1467 (vreinterpretq_f32_u16): Likewise.
1468 (vreinterpretq_f32_u32): Likewise.
1469 (vreinterpretq_f32_u64): Likewise.
1470 (vreinterpretq_f32_p8): Likewise.
1471 (vreinterpretq_f32_p16): Likewise.
1472 (vreinterpret_s64_s8): Likewise.
1473 (vreinterpret_s64_s16): Likewise.
1474 (vreinterpret_s64_s32): Likewise.
1475 (vreinterpret_s64_f32): Likewise.
1476 (vreinterpret_s64_u8): Likewise.
1477 (vreinterpret_s64_u16): Likewise.
1478 (vreinterpret_s64_u32): Likewise.
1479 (vreinterpret_s64_u64): Likewise.
1480 (vreinterpret_s64_p8): Likewise.
1481 (vreinterpret_s64_p16): Likewise.
1482 (vreinterpretq_s64_s8): Likewise.
1483 (vreinterpretq_s64_s16): Likewise.
1484 (vreinterpretq_s64_s32): Likewise.
1485 (vreinterpretq_s64_f32): Likewise.
1486 (vreinterpretq_s64_u8): Likewise.
1487 (vreinterpretq_s64_u16): Likewise.
1488 (vreinterpretq_s64_u32): Likewise.
1489 (vreinterpretq_s64_u64): Likewise.
1490 (vreinterpretq_s64_p8): Likewise.
1491 (vreinterpretq_s64_p16): Likewise.
1492 (vreinterpret_u64_s8): Likewise.
1493 (vreinterpret_u64_s16): Likewise.
1494 (vreinterpret_u64_s32): Likewise.
1495 (vreinterpret_u64_s64): Likewise.
1496 (vreinterpret_u64_f32): Likewise.
1497 (vreinterpret_u64_u8): Likewise.
1498 (vreinterpret_u64_u16): Likewise.
1499 (vreinterpret_u64_u32): Likewise.
1500 (vreinterpret_u64_p8): Likewise.
1501 (vreinterpret_u64_p16): Likewise.
1502 (vreinterpretq_u64_s8): Likewise.
1503 (vreinterpretq_u64_s16): Likewise.
1504 (vreinterpretq_u64_s32): Likewise.
1505 (vreinterpretq_u64_s64): Likewise.
1506 (vreinterpretq_u64_f32): Likewise.
1507 (vreinterpretq_u64_u8): Likewise.
1508 (vreinterpretq_u64_u16): Likewise.
1509 (vreinterpretq_u64_u32): Likewise.
1510 (vreinterpretq_u64_p8): Likewise.
1511 (vreinterpretq_u64_p16): Likewise.
1512 (vreinterpret_s8_s16): Likewise.
1513 (vreinterpret_s8_s32): Likewise.
1514 (vreinterpret_s8_s64): Likewise.
1515 (vreinterpret_s8_f32): Likewise.
1516 (vreinterpret_s8_u8): Likewise.
1517 (vreinterpret_s8_u16): Likewise.
1518 (vreinterpret_s8_u32): Likewise.
1519 (vreinterpret_s8_u64): Likewise.
1520 (vreinterpret_s8_p8): Likewise.
1521 (vreinterpret_s8_p16): Likewise.
1522 (vreinterpretq_s8_s16): Likewise.
1523 (vreinterpretq_s8_s32): Likewise.
1524 (vreinterpretq_s8_s64): Likewise.
1525 (vreinterpretq_s8_f32): Likewise.
1526 (vreinterpretq_s8_u8): Likewise.
1527 (vreinterpretq_s8_u16): Likewise.
1528 (vreinterpretq_s8_u32): Likewise.
1529 (vreinterpretq_s8_u64): Likewise.
1530 (vreinterpretq_s8_p8): Likewise.
1531 (vreinterpretq_s8_p16): Likewise.
1532 (vreinterpret_s16_s8): Likewise.
1533 (vreinterpret_s16_s32): Likewise.
1534 (vreinterpret_s16_s64): Likewise.
1535 (vreinterpret_s16_f32): Likewise.
1536 (vreinterpret_s16_u8): Likewise.
1537 (vreinterpret_s16_u16): Likewise.
1538 (vreinterpret_s16_u32): Likewise.
1539 (vreinterpret_s16_u64): Likewise.
1540 (vreinterpret_s16_p8): Likewise.
1541 (vreinterpret_s16_p16): Likewise.
1542 (vreinterpretq_s16_s8): Likewise.
1543 (vreinterpretq_s16_s32): Likewise.
1544 (vreinterpretq_s16_s64): Likewise.
1545 (vreinterpretq_s16_f32): Likewise.
1546 (vreinterpretq_s16_u8): Likewise.
1547 (vreinterpretq_s16_u16): Likewise.
1548 (vreinterpretq_s16_u32): Likewise.
1549 (vreinterpretq_s16_u64): Likewise.
1550 (vreinterpretq_s16_p8): Likewise.
1551 (vreinterpretq_s16_p16): Likewise.
1552 (vreinterpret_s32_s8): Likewise.
1553 (vreinterpret_s32_s16): Likewise.
1554 (vreinterpret_s32_s64): Likewise.
1555 (vreinterpret_s32_f32): Likewise.
1556 (vreinterpret_s32_u8): Likewise.
1557 (vreinterpret_s32_u16): Likewise.
1558 (vreinterpret_s32_u32): Likewise.
1559 (vreinterpret_s32_u64): Likewise.
1560 (vreinterpret_s32_p8): Likewise.
1561 (vreinterpret_s32_p16): Likewise.
1562 (vreinterpretq_s32_s8): Likewise.
1563 (vreinterpretq_s32_s16): Likewise.
1564 (vreinterpretq_s32_s64): Likewise.
1565 (vreinterpretq_s32_f32): Likewise.
1566 (vreinterpretq_s32_u8): Likewise.
1567 (vreinterpretq_s32_u16): Likewise.
1568 (vreinterpretq_s32_u32): Likewise.
1569 (vreinterpretq_s32_u64): Likewise.
1570 (vreinterpretq_s32_p8): Likewise.
1571 (vreinterpretq_s32_p16): Likewise.
1572 (vreinterpret_u8_s8): Likewise.
1573 (vreinterpret_u8_s16): Likewise.
1574 (vreinterpret_u8_s32): Likewise.
1575 (vreinterpret_u8_s64): Likewise.
1576 (vreinterpret_u8_f32): Likewise.
1577 (vreinterpret_u8_u16): Likewise.
1578 (vreinterpret_u8_u32): Likewise.
1579 (vreinterpret_u8_u64): Likewise.
1580 (vreinterpret_u8_p8): Likewise.
1581 (vreinterpret_u8_p16): Likewise.
1582 (vreinterpretq_u8_s8): Likewise.
1583 (vreinterpretq_u8_s16): Likewise.
1584 (vreinterpretq_u8_s32): Likewise.
1585 (vreinterpretq_u8_s64): Likewise.
1586 (vreinterpretq_u8_f32): Likewise.
1587 (vreinterpretq_u8_u16): Likewise.
1588 (vreinterpretq_u8_u32): Likewise.
1589 (vreinterpretq_u8_u64): Likewise.
1590 (vreinterpretq_u8_p8): Likewise.
1591 (vreinterpretq_u8_p16): Likewise.
1592 (vreinterpret_u16_s8): Likewise.
1593 (vreinterpret_u16_s16): Likewise.
1594 (vreinterpret_u16_s32): Likewise.
1595 (vreinterpret_u16_s64): Likewise.
1596 (vreinterpret_u16_f32): Likewise.
1597 (vreinterpret_u16_u8): Likewise.
1598 (vreinterpret_u16_u32): Likewise.
1599 (vreinterpret_u16_u64): Likewise.
1600 (vreinterpret_u16_p8): Likewise.
1601 (vreinterpret_u16_p16): Likewise.
1602 (vreinterpretq_u16_s8): Likewise.
1603 (vreinterpretq_u16_s16): Likewise.
1604 (vreinterpretq_u16_s32): Likewise.
1605 (vreinterpretq_u16_s64): Likewise.
1606 (vreinterpretq_u16_f32): Likewise.
1607 (vreinterpretq_u16_u8): Likewise.
1608 (vreinterpretq_u16_u32): Likewise.
1609 (vreinterpretq_u16_u64): Likewise.
1610 (vreinterpretq_u16_p8): Likewise.
1611 (vreinterpretq_u16_p16): Likewise.
1612 (vreinterpret_u32_s8): Likewise.
1613 (vreinterpret_u32_s16): Likewise.
1614 (vreinterpret_u32_s32): Likewise.
1615 (vreinterpret_u32_s64): Likewise.
1616 (vreinterpret_u32_f32): Likewise.
1617 (vreinterpret_u32_u8): Likewise.
1618 (vreinterpret_u32_u16): Likewise.
1619 (vreinterpret_u32_u64): Likewise.
1620 (vreinterpret_u32_p8): Likewise.
1621 (vreinterpret_u32_p16): Likewise.
1622 (vreinterpretq_u32_s8): Likewise.
1623 (vreinterpretq_u32_s16): Likewise.
1624 (vreinterpretq_u32_s32): Likewise.
1625 (vreinterpretq_u32_s64): Likewise.
1626 (vreinterpretq_u32_f32): Likewise.
1627 (vreinterpretq_u32_u8): Likewise.
1628 (vreinterpretq_u32_u16): Likewise.
1629 (vreinterpretq_u32_u64): Likewise.
1630 (vreinterpretq_u32_p8): Likewise.
1631 (vreinterpretq_u32_p16): Likewise.
1632
1633 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1634
1635 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1636 Pattern extended.
1637 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1638 extended.
1639 (sqabs): Likewise.
1640 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1641 (vqnegd_s64): Likewise.
1642 (vqabs_s64): Likewise.
1643 (vqabsd_s64): Likewise.
1644
1645 2014-04-22 Richard Henderson <rth@redhat.com>
1646
1647 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1648 computation to the top of the loop.
1649
1650 2014-04-22 Renlin <renlin.li@arm.com>
1651 Jiong Wang <jiong.wang@arm.com>
1652
1653 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1654 * config/aarch64/aarch64.c (aarch64_layout_frame)
1655 (aarch64_initial_elimination_offset): Likewise.
1656
1657 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1658
1659 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1660 Fix indentation.
1661
1662 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1663
1664 * machmode.h (bitwise_mode_for_mode): Declare.
1665 * stor-layout.h (bitwise_type_for_mode): Likewise.
1666 * stor-layout.c (bitwise_mode_for_mode): New function.
1667 (bitwise_type_for_mode): Likewise.
1668 * builtins.c (fold_builtin_memory_op): Use it instead of
1669 int_mode_for_mode and build_nonstandard_integer_type.
1670
1671 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1672
1673 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1674 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1675 (*-*-solaris2*): Simplify.
1676 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1677 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1678 *-*-solaris2.9* handling.
1679
1680 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1681 as bug.
1682 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1683 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1684 handling, simplify.
1685 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1686 * configure: Regenerate.
1687
1688 * config/i386/sol2-9.h: Remove.
1689
1690 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1691 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1692 Remove Solaris 9 references.
1693
1694 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1695
1696 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1697 (floatuns<GPI:mode><GPF:mode>2): Remove.
1698 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1699 and floatuns conversions.
1700 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1701 and floatuns conversions.
1702 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1703 (w1,w2): New mode attributes for inequal width conversions.
1704
1705 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1706
1707 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1708 the output asm format.
1709
1710 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1711
1712 * config/aarch64/aarch64-simd.md
1713 (aarch64_cm<optab>di): Always split.
1714 (*aarch64_cm<optab>di): New.
1715 (aarch64_cmtstdi): Always split.
1716 (*aarch64_cmtstdi): New.
1717
1718 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1719
1720 PR tree-optimization/60823
1721 * omp-low.c (ipa_simd_modify_function_body): Go through
1722 all SSA_NAMEs and for those refering to vector arguments
1723 which are going to be replaced adjust SSA_NAME_VAR and,
1724 if it is a default definition, change it into a non-default
1725 definition assigned at the beginning of function from new_decl.
1726 (ipa_simd_modify_stmt_ops): Rewritten.
1727 * tree-dfa.c (set_ssa_default_def): When removing default def,
1728 check for NULL loc instead of NULL *loc.
1729
1730 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1731
1732 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1733 restrictions on core registers for DImode values in Thumb2.
1734
1735 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1736
1737 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1738 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1739
1740 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1741
1742 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1743 (*iordi_notzesidi_di): Likewise.
1744 (*iordi_notsesidi_di): Likewise.
1745
1746 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1747
1748 * config/arm/arm-protos.h (tune_params): New struct members.
1749 * config/arm/arm.c: Initialise tune_params per processor.
1750 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1751 for speed, based on new tune_params.
1752
1753 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1754
1755 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1756 added.
1757 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1758 macro.
1759 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1760 corrected.
1761 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1762 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1763 (vrnda_f64): Likewise.
1764 (vrndi_f64): Likewise.
1765 (vrndm_f64): Likewise.
1766 (vrndn_f64): Likewise.
1767 (vrndp_f64): Likewise.
1768 (vrndx_f64): Likewise.
1769
1770 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1771
1772 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1773 GET_MODE_SIZE argument is enum machine_mode.
1774
1775 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1776
1777 PR target/60910
1778 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1779 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1780
1781 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1782
1783 PR middle-end/60281
1784 * asan.c (asan_emit_stack_protection): Force the base to align to
1785 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1786 appropriate bits if STRICT_ALIGNMENT.
1787 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1788 when asan is on.
1789 (expand_used_vars): Leave a space in the stack frame for alignment
1790 if STRICT_ALIGNMENT.
1791
1792 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1793
1794 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1795 than a gimple.
1796 (gimple_store_p): Likewise.
1797 (gimple_assign_load_p): Likewise.
1798 (gimple_assign_cast_p): Likewise.
1799 (gimple_clobber_p): Likewise.
1800
1801 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1802 rather than a gimple.
1803 (gimple_assign_cast_p): Likewise.
1804
1805 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1806
1807 PR target/60735
1808 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1809 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1810
1811 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1812 more debug information for E500 if -mdebug=reg.
1813
1814 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1815
1816 PR target/60909
1817 * config/i386/i386.c (ix86_expand_builtin)
1818 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1819 register for target RTX.
1820 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1821
1822 2014-04-18 Cong Hou <congh@google.com>
1823
1824 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1825 the widen-mult pattern by handling two operands with different sizes,
1826 and operands whose size is smaller than half of the result type.
1827
1828 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1829
1830 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1831 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1832 (do_estimate_edge_time): Compute it.
1833 * ipa-inline.c (want_inline_small_function_p): Bypass
1834 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1835
1836 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1837
1838 * ipa-inline.c (spec_rem): New static variable.
1839 (dump_overall_stats): New function.
1840 (dump_inline_stats): New function.
1841
1842 2014-04-18 Richard Henderson <rth@redhat.com>
1843
1844 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1845 to GET_MODE_SIZE, not a reg_class_t.
1846
1847 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1848
1849 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1850 (vsx_xxmrglw_<mode>): Likewise.
1851
1852 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1853
1854 PR target/60876
1855 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1856 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1857 (rs6000_init_hard_regno_mode_ok): Likewise.
1858
1859 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1860
1861 * ipa-inline.c (inline_small_functions): Account only non-cold
1862 functions.
1863 * doc/invoke.texi (inline-unit-growth): Update documentation.
1864
1865 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1866
1867 * config/rs6000/rs6000.md (addti3, subti3): New.
1868
1869 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1870
1871 PR target/60863
1872 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1873 comment. Check optimize_insn_for_size_p instead of
1874 optimize_insn_for_speed_p.
1875
1876 2014-04-17 Martin Jambor <mjambor@suse.cz>
1877
1878 * gimple-iterator.c (gsi_start_edge): New function.
1879 * gimple-iterator.h (gsi_start_edge): Declare.
1880 * tree-sra.c (single_non_eh_succ): New function.
1881 (disqualify_ops_if_throwing_stmt): Renamed to
1882 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1883 having one non-EH successor BB.
1884 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1885 generate loads into replacements.
1886 (sra_modify_assign): Likewise and and also use the simple path for
1887 such statements.
1888 (sra_modify_function_body): Commit statements on edges.
1889
1890 2014-04-17 Richard Biener <rguenther@suse.de>
1891
1892 PR middle-end/60849
1893 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1894 comparison results and add clarifying comment.
1895
1896 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1897
1898 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1899 (blank_mode): Initialize it.
1900 (emit_mode_size_inline, emit_mode_nunits_inline,
1901 emit_mode_inner_inline): New functions.
1902 (emit_insn_modes_h): Call them and surround their output with
1903 #if GCC_VERSION >= 4001 ... #endif.
1904 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1905 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1906 mode_* arrays if the argument is __builtin_constant_p.
1907 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1908 is enum machine_mode.
1909
1910 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1911
1912 * passes.c (opt_pass::execute): Adjust.
1913 (pass_manager::execute_pass_mode_switching): Likewise.
1914 (early_local_passes::execute): Likewise.
1915 (execute_one_pass): Pass cfun to the pass's execute method.
1916 * tree-pass.h (opt_pass::execute): Add function * argument.
1917 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1918 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1919 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1920 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1921 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1922 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1923 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1924 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1925 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1926 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1927 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1928 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1929 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1930 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1931 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1932 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1933 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1934 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1935 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1936 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1937 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1938 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1939 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1940 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1941 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1942 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1943 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1944 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1945 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1946 Adjust.
1947
1948 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1949
1950 * passes.c (opt_pass::gate): Take function * argument.
1951 (gate_all_early_local_passes): Merge into
1952 (early_local_passes::gate): this.
1953 (gate_all_early_optimizations): Merge into
1954 (all_early_optimizations::gate): this.
1955 (gate_all_optimizations): Mege into
1956 (all_optimizations::gate): this.
1957 (gate_all_optimizations_g): Merge into
1958 (all_optimizations_g::gate): this.
1959 (gate_rest_of_compilation): Mege into
1960 (rest_of_compilation::gate): this.
1961 (gate_postreload): Merge into
1962 (postreload::gate): this.
1963 (dump_one_pass): Pass cfun to the pass's gate method.
1964 (execute_ipa_summary_passes): Likewise.
1965 (execute_one_pass): Likewise.
1966 (ipa_write_summaries_2): Likewise.
1967 (ipa_write_optimization_summaries_1): Likewise.
1968 (ipa_read_summaries_1): Likewise.
1969 (ipa_read_optimization_summaries_1): Likewise.
1970 (execute_ipa_stmt_fixups): Likewise.
1971 * tree-pass.h (opt_pass::gate): Add function * argument.
1972 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1973 combine-stack-adj.c, combine.c, compare-elim.c,
1974 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1975 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1976 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1977 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1978 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1979 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1980 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1981 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1982 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1983 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1984 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1985 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1986 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
1987 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
1988 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
1989 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1990 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1991 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1992 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1993 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1994 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1995 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1996 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
1997 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
1998 var-tracking.c, vtable-verify.c, web.c: Adjust.
1999
2000 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2001
2002 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2003 * configure: Regenerate.
2004
2005 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2006
2007 * passes.c (dump_one_pass): don't check pass->has_gate.
2008 (execute_ipa_summary_passes): Likewise.
2009 (execute_one_pass): Likewise.
2010 (ipa_write_summaries_2): Likewise.
2011 (ipa_write_optimization_summaries_1): Likewise.
2012 (ipa_read_optimization_summaries_1): Likewise.
2013 (execute_ipa_stmt_fixups): Likewise.
2014 * tree-pass.h (pass_data::has_gate): Remove.
2015 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2016 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2017 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2018 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2019 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2020 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2021 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2022 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2023 gimple-low.c, gimple-ssa-isolate-paths.c,
2024 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2025 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2026 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2027 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2028 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2029 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2030 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2031 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2032 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2033 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2034 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2035 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2036 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2037 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2038 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2039 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2040 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2041 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2042 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2043 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2044 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2045 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2046 Adjust.
2047
2048 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2049
2050 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2051 declaration.
2052 * passes.c (pass_manager::register_dump_files_1): Merge into
2053 (pass_manager::register_dump_files): this, and remove its handling of
2054 properties since the pass always has the properties anyway.
2055 (pass_manager::pass_manager): Adjust.
2056
2057 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2058
2059 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2060 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2061 dealing with properties.
2062 (pass_manager::register_dump_files): Adjust.
2063
2064 2014-03-20 Mark Wielaard <mjw@redhat.com>
2065
2066 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2067 then represent the bound as normal constant value.
2068
2069 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2070
2071 PR target/60847
2072 Forward port from 4.8 branch
2073 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2074
2075 * config/i386/bmiintrin.h (_blsi_u32): New.
2076 (_blsi_u64): Ditto.
2077 (_blsr_u32): Ditto.
2078 (_blsr_u64): Ditto.
2079 (_blsmsk_u32): Ditto.
2080 (_blsmsk_u64): Ditto.
2081 (_tzcnt_u32): Ditto.
2082 (_tzcnt_u64): Ditto.
2083
2084 2014-04-17 Kito Cheng <kito@0xlab.org>
2085
2086 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2087
2088 2014-04-17 Richard Biener <rguenther@suse.de>
2089
2090 PR middle-end/60849
2091 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2092 boolean results for comparisons.
2093
2094 2014-04-17 Richard Biener <rguenther@suse.de>
2095
2096 PR tree-optimization/60836
2097 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2098 initial PHI args to be gimple values.
2099
2100 2014-04-17 Richard Biener <rguenther@suse.de>
2101
2102 PR tree-optimization/60841
2103 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2104 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2105 of stmts to SLP build.
2106 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2107 (vect_analyze_slp): Likewise.
2108 (vect_analyze_slp_instance): Likewise.
2109 (vect_build_slp_tree): Limit overall SLP tree growth.
2110 * tree-vectorizer.h (vect_analyze_data_refs,
2111 vect_analyze_slp): Adjust prototypes.
2112
2113 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2114
2115 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2116 Silvermont.
2117
2118 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2119
2120 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2121 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2122 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2123 for TARGET_SLOW_PSHUFB
2124
2125 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2126
2127 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2128 * config/i386/i386.c (intel_cost): Ditto.
2129
2130 2014-04-17 Joey Ye <joey.ye@arm.com>
2131
2132 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2133
2134 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2135
2136 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2137 with -fuse-profile.
2138
2139 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2140
2141 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2142 (type_all_derivations_known_p): New predicate.
2143 (type_all_ctors_visible_p): New predicate.
2144 (type_possibly_instantiated_p): New predicate.
2145 (get_odr_type): Compute all_derivations_known.
2146 (dump_odr_type): Dump the flag.
2147 (maybe_record_type): Cleanup.
2148 (record_target_from_binfo): Add bases_to_consider array;
2149 record bases for types w/o instances and skip CXX destructor.
2150 (possible_polymorphic_call_targets_1): Add bases_to_consider
2151 and consider_construction parameters; check if type may have instance.
2152 (get_polymorphic_call_info): Set maybe_in_construction to true
2153 when we know nothing.
2154 (record_targets_from_bases): Skip CXX destructors; they are
2155 never called for types in construction.
2156 (possible_polymorphic_call_targets): Do not record target when
2157 type may not have instance.
2158
2159 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2160
2161 PR ipa/60854
2162 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2163 external aliases alive, too.
2164
2165 2014-04-16 Andrew Pinski <apinski@cavium.com>
2166
2167 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2168 definition.
2169
2170 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2171
2172 * final.c (compute_alignments): Do not apply loop alignment to a block
2173 falling through to the exit.
2174
2175 2014-04-16 Catherine Moore <clm@codesourcery.com>
2176
2177 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2178 Adjust constraints for microMIPS store patterns.
2179
2180 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2181
2182 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2183
2184 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2185
2186 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2187 (append_use): Run at -O0.
2188 (append_vdef): Likewise.
2189 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2190 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2191
2192 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2193
2194 PR tree-optimization/60844
2195 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2196 (propagate_op_to_single_use, remove_visited_stmt_chain,
2197 linearize_expr, repropagate_negates, reassociate_bb): Use it
2198 instead of gsi_remove.
2199
2200 2014-04-16 Martin Jambor <mjambor@suse.cz>
2201
2202 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2203 ipa_transforms_to_apply.
2204 (cgraph_function_versioning): Assert that old_node has empty
2205 ipa_transforms_to_apply.
2206 * trans-mem.c (ipa_tm_create_version): Likewise.
2207 * tree-inline.c (tree_function_versioning): Do not duplicate
2208 ipa_transforms_to_apply.
2209
2210 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2211
2212 PR target/60817
2213 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2214 x86_64-*-* cases.
2215 Pass necessary as flags on 64-bit Solaris/x86.
2216 Use lowercase relocs for x86_64-*-*.
2217 * configure: Regenerate.
2218
2219 2014-04-15 Jan Hubicka <jh@suse.cz>
2220
2221 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2222 (maybe_record_node, likely_target_p): Use it.
2223
2224 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2225
2226 PR target/60839
2227 Revert following patch
2228
2229 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2230
2231 PR target/60735
2232 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2233 software floating point or no floating point registers, do not
2234 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2235 in GPRs that occurs after we tested for GPRs that would never be
2236 true.
2237
2238 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2239 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2240 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2241 specifically allow DDmode, since that does not use the SPE SIMD
2242 instructions.
2243
2244 2014-03-21 Mark Wielaard <mjw@redhat.com>
2245
2246 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2247 as unsigned or int depending on type and value used.
2248
2249 2014-04-15 Richard Biener <rguenther@suse.de>
2250
2251 PR rtl-optimization/56965
2252 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2253 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2254 ... here.
2255 * alias.c (true_dependence_1): Do not call
2256 nonoverlapping_component_refs_p.
2257 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2258 nonoverlapping_component_refs_p.
2259 (indirect_refs_may_alias_p): Likewise.
2260
2261 2014-04-15 Teresa Johnson <tejohnson@google.com>
2262
2263 * cfg.c (dump_bb_info): Fix flags check.
2264 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2265
2266 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2267
2268 PR rtl-optimization/60663
2269 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2270 avoid 0 cost.
2271
2272 2014-04-15 Richard Biener <rguenther@suse.de>
2273
2274 * lto-streamer.h (LTO_major_version): Bump to 4.
2275
2276 2014-04-15 Richard Biener <rguenther@suse.de>
2277
2278 * common.opt (lto_partition_model): New enum.
2279 (flto-partition=): Merge separate options with a single with argument,
2280 add -flto-partition=one support.
2281 * flag-types.h (enum lto_partition_model): Declare.
2282 * opts.c (finish_options): Remove duplicate -flto-partition=
2283 option check.
2284 * lto-wrapper.c (run_gcc): Adjust.
2285
2286 2014-04-15 Richard Biener <rguenther@suse.de>
2287
2288 * alias.c (ncr_compar): New function.
2289 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2290
2291 2014-04-15 Richard Biener <rguenther@suse.de>
2292
2293 * alias.c (record_component_aliases): Do not walk BINFOs.
2294
2295 2014-04-15 Richard Biener <rguenther@suse.de>
2296
2297 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2298 Add struct function argument and adjust.
2299 (find_func_aliases_for_call): Likewise.
2300 (find_func_aliases): Likewise.
2301 (find_func_clobbers): Likewise.
2302 (intra_create_variable_infos): Likewise.
2303 (compute_points_to_sets): Likewise.
2304 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2305
2306 2014-04-15 Richard Biener <rguenther@suse.de>
2307
2308 * tree.c (iterative_hash_expr): Use enum tree_code_class
2309 to store TREE_CODE_CLASS.
2310 (tree_block): Likewise.
2311 (tree_set_block): Likewise.
2312 * tree.h (fold_build_pointer_plus_loc): Use
2313 convert_to_ptrofftype_loc.
2314
2315 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2316
2317 PR plugins/59335
2318 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2319 added in 4.9.
2320
2321 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2322
2323 * cfgloop.h (struct loop): Move force_vectorize down.
2324 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2325 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2326 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2327 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2328 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2329 * tree-core.h (enum annot_expr_kind): Add new kind values.
2330 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2331 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2332 kinds.
2333 * tree.def (ANNOTATE_EXPR): Tweak comment.
2334
2335 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2336
2337 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2338 cxa_pure_virtual).
2339
2340 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2341
2342 * tree.h (TYPE_IDENTIFIER): Declare.
2343 * tree.c (subrange_type_for_debug_p): Use it.
2344 * godump.c (go_format_type): Likewise.
2345 * dwarf2out.c (is_cxx_auto, modified_type_die,
2346 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2347 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2348
2349 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2350
2351 PR lto/60820
2352 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2353
2354 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2355
2356 * config/i386/i386.c (examine_argument): Return bool. Return true if
2357 parameter should be passed in memory.
2358 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2359 (construct_container): Update calls to examine_argument.
2360 (function_arg_advance_64): Ditto.
2361 (return_in_memory_32): Merge with ix86_return_in_memory.
2362 (return_in_memory_64): Ditto.
2363 (return_in_memory_ms_64): Ditto.
2364
2365 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2366
2367 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2368 * coverage.c (coverage_compute_profile_id): Handle externally visible
2369 symbols.
2370
2371 2014-04-14 Martin Jambor <mjambor@suse.cz>
2372
2373 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2374 DECL_DISREGARD_INLINE_LIMITS functions.
2375
2376 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2377
2378 PR target/60827
2379 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2380
2381 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2382
2383 PR target/60827
2384 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2385 optimize_insn_for_speed_p instead of
2386 optimize_function_for_speed_p.
2387
2388 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2389
2390 * doc/invoke.texi (free): Document AArch64.
2391
2392 2014-04-14 Richard Biener <rguenther@suse.de>
2393
2394 PR tree-optimization/60042
2395 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2396 (insert_into_preds_of_block): Do not prevent PHI insertion
2397 for REFERENCE exprs here ...
2398 (eliminate_dom_walker::before_dom_children): ... but prevent
2399 their use here under similar conditions when applied to the
2400 IL after PRE optimizations.
2401
2402 2014-04-14 Richard Biener <rguenther@suse.de>
2403
2404 * passes.def: Move early points-to after early SRA.
2405
2406 2014-04-14 Richard Biener <rguenther@suse.de>
2407
2408 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2409 check for which sign-changes we allow when forwarding
2410 a converted value into a switch.
2411
2412 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2413
2414 * stor-layout.c (place_field): Finalize non-constant offset for the
2415 field, if any.
2416
2417 2014-04-14 Richard Biener <rguenther@suse.de>
2418
2419 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2420 as argument.
2421 (expand_switch_using_bit_tests_p): Likewise.
2422 (process_switch): Compute and pass on speed_p based on the
2423 switch stmt.
2424 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2425 optimize_bb_for_speed_p.
2426
2427 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2428
2429 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2430 * function.h (struct function): Rename has_force_vect_loops into
2431 has_force_vectorize_loops.
2432 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2433 (input_struct_function_base): Likewise.
2434 * lto-streamer-out.c (output_cfg): Likewise.
2435 (output_struct_function_base): Likewise.
2436 * omp-low.c (expand_omp_simd): Likewise.
2437 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2438 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2439 (version_loop_for_if_conversion): Likewise.
2440 (tree_if_conversion): Likewise.
2441 (main_tree_if_conversion): Likewise.
2442 (gate_tree_if_conversion): Likewise.
2443 * tree-inline.c (copy_loops): Likewise.
2444 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2445 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2446 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2447 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2448 * tree-vectorizer.c (vectorize_loops): Likewise.
2449 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2450
2451 2014-04-14 Richard Biener <rguenther@suse.de>
2452
2453 PR lto/60720
2454 * lto-streamer-out.c (wrap_refs): New function.
2455 (lto_output): Wrap symbol references in global initializes in
2456 type-preserving MEM_REFs.
2457
2458 2014-04-14 Christian Bruel <christian.bruel@st.com>
2459
2460 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2461
2462 2014-04-14 Christian Bruel <christian.bruel@st.com>
2463
2464 * config/sh/sh.md (setmemqi): New expand pattern.
2465 * config/sh/sh.h (CLEAR_RATIO): Define.
2466 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2467 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2468
2469 2014-04-14 Richard Biener <rguenther@suse.de>
2470
2471 PR middle-end/55022
2472 * fold-const.c (negate_expr_p): Don't negate directional rounding
2473 division.
2474 (fold_negate_expr): Likewise.
2475
2476 2014-04-14 Richard Biener <rguenther@suse.de>
2477
2478 PR tree-optimization/59817
2479 PR tree-optimization/60453
2480 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2481 recursion to catch all CHRECs in the scalar evolution and restrict
2482 the predicate for the remains appropriately.
2483
2484 2014-04-12 Catherine Moore <clm@codesourcery.com>
2485
2486 * config/mips/constraints.md: Add new register constraint "kb".
2487 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2488 (*movhi_internal): Likewise.
2489 (*movqi_internal): Likewise.
2490 * config/mips/mips.h (M16_STORE_REGS): New register class.
2491 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2492 (REG_CLASS_CONTENTS): Likewise.
2493 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2494
2495 2014-04-11 Tobias Burnus <burnus@net-b.de>
2496
2497 PR c/60194
2498 * doc/invoke.texi (-Wformat-signedness): Document it.
2499 (Wformat=2): Mention that this enables -Wformat-signedness.
2500
2501 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2502
2503 * common/config/epiphany/epiphany-common.c
2504 (epiphany_option_optimization_table): Enable section anchors by
2505 default at -O1 or higher.
2506 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2507 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2508 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2509 carries no extra cost.
2510 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2511 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2512 * config/epiphany/predicates.md (memclob_operand): New predicate.
2513 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2514 Use memclob_operand predicate and X constraint for operand 3.
2515
2516 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2517
2518 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2519 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2520 its operands.
2521
2522 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2523
2524 PR rtl-optimization/60651
2525 * mode-switching.c (optimize_mode_switching): Make sure to emit
2526 sets of a lower numbered entity before sets of a higher numbered
2527 entity to a mode of the same or lower priority.
2528 When creating a seginfo for a basic block that starts with a code
2529 label, move the insertion point past the code label.
2530 (new_seginfo): Document and enforce requirement that
2531 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2532 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2533 * doc/tm.texi: Regenerate.
2534
2535 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2536
2537 PR target/60811
2538 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2539
2540 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2541
2542 * BASE-VER: Set to 4.10.0.
2543
2544 2014-04-11 Tobias Burnus <burnus@net-b.de>
2545
2546 PR other/59055
2547 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2548 * doc/gcc.texi (Service): Update description in the @menu
2549 * doc/invoke.texi (Option Summary): Remove misplaced and
2550 duplicated @menu.
2551
2552 2014-04-11 Steve Ellcey <sellcey@mips.com>
2553 Jakub Jelinek <jakub@redhat.com>
2554
2555 PR middle-end/60556
2556 * expr.c (convert_move): Use emit_store_flag_force instead of
2557 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2558 argument to it.
2559
2560 2014-04-11 Richard Biener <rguenther@suse.de>
2561
2562 PR middle-end/60797
2563 * varasm.c (assemble_alias): Avoid endless error reporting
2564 recursion by setting TREE_ASM_WRITTEN.
2565
2566 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2567
2568 * config/s390/s390.md: Add a splitter for NOT rtx.
2569
2570 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2571
2572 PR rtl-optimization/60663
2573 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2574
2575 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2576 Jakub Jelinek <jakub@redhat.com>
2577
2578 PR lto/60567
2579 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2580 flag from decl_node to node.
2581
2582 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2583
2584 PR debug/60655
2585 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2586 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2587 ameliorating the cases where it can be.
2588
2589 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2590
2591 Revert
2592 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2593
2594 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2595 (loadsync_<mode>): Change mode.
2596 (load_quadpti, store_quadpti): New.
2597 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2598 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2599 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2600
2601 2014-04-09 Cong Hou <congh@google.com>
2602
2603 PR testsuite/60773
2604 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2605 documentation.
2606
2607 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2608
2609 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2610 instead of vnor to exploit possible fusion opportunity in the
2611 future.
2612 (altivec_expand_vec_perm_const_le): Likewise.
2613
2614 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2615
2616 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2617 (loadsync_<mode>): Change mode.
2618 (load_quadpti, store_quadpti): New.
2619 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2620 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2621
2622 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2623
2624 PR target/60763
2625 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2626 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2627 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2628
2629 2014-04-08 Richard Biener <rguenther@suse.de>
2630
2631 PR middle-end/60706
2632 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2633 a 64bit widest int print double-int similar to on HWI64 hosts.
2634
2635 2014-04-08 Richard Biener <rguenther@suse.de>
2636
2637 PR tree-optimization/60785
2638 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2639 default defs properly.
2640
2641 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2642
2643 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2644 (Weffc++): Likewise.
2645
2646 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2647
2648 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2649 set completep to false rather than true.
2650
2651 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2652
2653 PR target/60504
2654 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2655 ARM_TARGET2_DWARF_FORMAT.
2656
2657 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2658
2659 PR target/60609
2660 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2661 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2662 ADDR_DIFF_VEC.
2663
2664 2014-04-07 Richard Biener <rguenther@suse.de>
2665
2666 PR tree-optimization/60766
2667 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2668 (may_eliminate_iv): Convert cand_value_at result to desired type.
2669
2670 2014-04-07 Jason Merrill <jason@redhat.com>
2671
2672 PR c++/60731
2673 * common.opt (-fno-gnu-unique): Add.
2674 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2675
2676 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2677
2678 * haifa-sched.c: Fix outdated function reference and minor
2679 grammar errors in introductory comment.
2680
2681 2014-04-07 Richard Biener <rguenther@suse.de>
2682
2683 PR middle-end/60750
2684 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2685 for noreturn calls.
2686 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2687
2688 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2689
2690 PR debug/55794
2691 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2692 size accounting for thunks.
2693 (pa_asm_output_mi_thunk): Use final_start_function() and
2694 final_end_function() to output function start and end directives.
2695
2696 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2697
2698 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2699 device specific ISA/ feature information. Remove short_sp and
2700 errata_skip ds. Add avr_device_specific_features enum to have device
2701 specific info.
2702 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2703 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2704 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2705 updated device specific info.
2706 * config/avr/avr-mcus.def: Merge device specific details to
2707 dev_attribute field.
2708 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2709 errata_skip.
2710 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2711 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2712 assembler if RMW isa supported by current device.
2713 * config/avr/genmultilib.awk: Update as device info structure changed.
2714 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2715
2716 2014-04-04 Cong Hou <congh@google.com>
2717
2718 PR tree-optimization/60656
2719 * tree-vect-stmts.c (supportable_widening_operation):
2720 Fix a bug that elements in a vector with vect_used_by_reduction
2721 property are incorrectly reordered when the operation on it is not
2722 consistant with the one in reduction operation.
2723
2724 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2725
2726 PR rtl-optimization/60155
2727 * gcse.c (record_set_data): New function.
2728 (single_set_gcse): New function.
2729 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2730 (hoist_code): Likewise.
2731 (get_pressure_class_and_nregs): Likewise.
2732
2733 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2734
2735 * explow.c (probe_stack_range): Emit a final optimization blockage.
2736
2737 2014-04-04 Anthony Green <green@moxielogic.com>
2738
2739 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2740 typos.
2741
2742 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2743
2744 PR ipa/59626
2745 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2746 flags are set only during streaming.
2747 * ipa.c (process_references, walk_polymorphic_call_targets,
2748 symtab_remove_unreachable_nodes): Drop bodies of always inline
2749 after early inlining.
2750 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2751
2752 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2753 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2754
2755 PR debug/60655
2756 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2757 containing a NOT.
2758
2759 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2760
2761 PR bootstrap/60743
2762 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2763 duration.
2764 (cortex_a53_fdivd): Likewise.
2765
2766 2014-04-04 Martin Jambor <mjambor@suse.cz>
2767
2768 PR ipa/60640
2769 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2770 Adjust all callers.
2771 * cgraph.c (clone_of_p): Also return true if thunks match.
2772 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2773 cgraph_function_or_thunk_node and an obsolete comment.
2774 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2775 file.
2776 (build_function_decl_skip_args): Likewise.
2777 (set_new_clone_decl_and_node_flags): New function.
2778 (duplicate_thunk_for_node): Likewise.
2779 (redirect_edge_duplicating_thunks): Likewise.
2780 (cgraph_clone_node): New parameter args_to_skip, pass it to
2781 redirect_edge_duplicating_thunks which is called instead of
2782 cgraph_redirect_edge_callee.
2783 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2784 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2785
2786 2014-04-04 Jeff Law <law@redhat.com>
2787
2788 PR target/60657
2789 * config/arm/predicates.md (const_int_I_operand): New predicate.
2790 (const_int_M_operand): Similarly.
2791 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2792 const_int_operand.
2793 (insv_t2, extv_reg, extzv_t2): Likewise.
2794 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2795 (pop_multiple_with_writeback_and_return): Likewise.
2796 (vfp_pop_multiple_with_writeback): Likewise
2797
2798 2014-04-04 Richard Biener <rguenther@suse.de>
2799
2800 PR ipa/60746
2801 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2802 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2803 non-GIMPLE_LABELs.
2804 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2805 * gimplify.c (gimple_add_tmp_var_fn): New function.
2806 * gimple-expr.h (create_tmp_reg_fn): Declare.
2807 * gimple-expr.c (create_tmp_reg_fn): New function.
2808 * gimple-low.c (record_vars_into): Don't change cfun.
2809 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2810 code generation without cfun.
2811
2812 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2813
2814 PR bootstrap/60719
2815 * Makefile.in (install-driver): Fix shell scripting.
2816
2817 2014-04-03 Cong Hou <congh@google.com>
2818
2819 PR tree-optimization/60505
2820 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2821 threshold of number of iterations below which no vectorization
2822 will be done.
2823 * tree-vect-loop.c (new_loop_vec_info):
2824 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2825 * tree-vect-loop.c (vect_analyze_loop_operations):
2826 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2827 * tree-vect-loop.c (vect_transform_loop):
2828 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2829 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2830 of iterations of the loop and see if we should build the epilogue.
2831
2832 2014-04-03 Richard Biener <rguenther@suse.de>
2833
2834 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2835 (streamer_tree_cache_create): Adjust.
2836 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2837 to allow optional nodes array.
2838 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2839 (streamer_tree_cache_append): Likewise.
2840 (streamer_tree_cache_create): Create nodes array optionally
2841 as specified by parameter.
2842 * lto-streamer-out.c (create_output_block): Avoid maintaining
2843 the node array in the writer cache.
2844 (DFS_write_tree): Remove assertion.
2845 (produce_asm_for_decls): Free the out decl state hash table early.
2846 * lto-streamer-in.c (lto_data_in_create): Adjust for
2847 streamer_tree_cache_create prototype change.
2848
2849 2014-04-03 Richard Biener <rguenther@suse.de>
2850
2851 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2852 set TREE_CHAIN to NULL_TREE.
2853
2854 2014-04-03 Richard Biener <rguenther@suse.de>
2855
2856 PR tree-optimization/60740
2857 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2858 over all GIMPLE_COND operands.
2859
2860 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2861
2862 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2863 (Weffc++): Remove Scott's numbering, merge lists and reference
2864 Wnon-virtual-dtor.
2865
2866 2014-04-03 Nick Clifton <nickc@redhat.com>
2867
2868 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2869 properly.
2870
2871 2014-04-03 Martin Jambor <mjambor@suse.cz>
2872
2873 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2874 mention gcc_unreachable before failing.
2875 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2876 removed symbols.
2877
2878 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2879
2880 PR ipa/60659
2881 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2882 inconsistent code and instead mark the context inconsistent.
2883 (possible_polymorphic_call_targets): For inconsistent contexts
2884 return empty complete list.
2885
2886 2014-04-02 Anthony Green <green@moxielogic.com>
2887
2888 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2889 (extendqisi2, extendhisi2): Define.
2890 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2891 (WCHAR_TYPE): Change to unsigned int.
2892
2893 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2894
2895 PR tree-optimization/60733
2896 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2897 insertion point for PHI candidates to be the end of the feeding
2898 block for the PHI argument.
2899
2900 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2901
2902 PR rtl-optimization/60650
2903 * lra-constraints.c (process_alt_operands): Decrease reject for
2904 earlyclobber matching.
2905
2906 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2907
2908 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2909
2910 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2911
2912 * config/spu/spu.c (pad_bb): Do not crash when the last
2913 insn is CODE_FOR_blockage.
2914
2915 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2916
2917 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2918 lies outside the target mode.
2919
2920 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2921
2922 PR target/60735
2923 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2924 software floating point or no floating point registers, do not
2925 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2926 in GPRs that occurs after we tested for GPRs that would never be
2927 true.
2928
2929 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2930 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2931 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2932 specifically allow DDmode, since that does not use the SPE SIMD
2933 instructions.
2934
2935 2014-04-02 Richard Biener <rguenther@suse.de>
2936
2937 PR middle-end/60729
2938 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2939 MODE_INTs. Properly use negv_optab.
2940 (expand_abs): Likewise.
2941
2942 2014-04-02 Richard Biener <rguenther@suse.de>
2943
2944 PR bootstrap/60719
2945 * Makefile.in (install-driver): Guard extra installs with special
2946 names properly.
2947
2948 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2949
2950 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2951 Document vec_vgbbd.
2952
2953 2014-04-01 Richard Henderson <rth@redhat.com>
2954
2955 PR target/60704
2956 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2957 alternative enabled before register allocation.
2958
2959 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2960
2961 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2962 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2963 typo.
2964 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2965 (nios2_got_address): Update nios2_large_got_address call site.
2966 (nios2_delegitimize_address): New function.
2967 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2968 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2969 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2970
2971 2014-04-01 Martin Husemann <martin@duskware.de>
2972
2973 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2974 for -mabi=32.
2975
2976 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2977
2978 PR rtl-optimization/60604
2979 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2980 check from register_operand.
2981 (register_operand): Redefine in terms of general_operand.
2982 (nonmemory_operand): Use register_operand for the non-constant cases.
2983
2984 2014-04-01 Richard Biener <rguenther@suse.de>
2985
2986 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
2987
2988 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
2989
2990 * doc/invoke.texi (mapp-regs): Clarify.
2991
2992 2014-03-31 Ulrich Drepper <drepper@gmail.com>
2993
2994 * config/i386/avx512fintrin.h (__v32hi): Define type.
2995 (__v64qi): Likewise.
2996 (_mm512_set1_epi8): Define.
2997 (_mm512_set1_epi16): Define.
2998 (_mm512_set4_epi32): Define.
2999 (_mm512_set4_epi64): Define.
3000 (_mm512_set4_pd): Define.
3001 (_mm512_set4_ps): Define.
3002 (_mm512_setr4_epi64): Define.
3003 (_mm512_setr4_epi32): Define.
3004 (_mm512_setr4_pd): Define.
3005 (_mm512_setr4_ps): Define.
3006 (_mm512_setzero_epi32): Define.
3007
3008 2014-03-31 Martin Jambor <mjambor@suse.cz>
3009
3010 PR middle-end/60647
3011 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3012 callsite_arguments_match_p. Updated all callers. Also check types of
3013 corresponding formal parameters and actual arguments.
3014 (not_all_callers_have_enough_arguments_p) Renamed to
3015 some_callers_have_mismatched_arguments_p.
3016
3017 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3018
3019 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3020
3021 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3022
3023 PR target/60034
3024 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3025 section anchor.
3026
3027 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3028
3029 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3030 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3031 Split out
3032 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3033 Use FMAMODE_NOVF512 mode iterator.
3034 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3035 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3036 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3037 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3038 Split out
3039 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3040 Use VF_128_256 mode iterator.
3041 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3042 Ditto.
3043
3044 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3045
3046 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3047 static chain if needed.
3048
3049 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3050
3051 PR target/60697
3052 * lra-constraints.c (index_part_to_reg): New.
3053 (process_address): Use it.
3054
3055 2014-03-27 Jeff Law <law@redhat.com>
3056 Jakub Jelinek <jakub@redhat.com>
3057
3058 PR target/60648
3059 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3060 gen_rtx_{PLUS,MULT} to build up the address expression.
3061
3062 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3063 creating non-canonical RTL.
3064
3065 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3066
3067 PR ipa/60243
3068 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3069 functions; reorganize to make cheap checks first.
3070 (inline_small_functions): Do not estimate growth when dumping;
3071 it is expensive.
3072 * ipa-inline.h (inline_summary): Add min_size.
3073 (growth_likely_positive): New function.
3074 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3075 (set_cond_stmt_execution_predicate): Cleanup.
3076 (estimate_edge_size_and_time): Compute min_size.
3077 (estimate_calls_size_and_time): Likewise.
3078 (estimate_node_size_and_time): Likewise.
3079 (inline_update_overall_summary): Update min_size.
3080 (do_estimate_edge_time): Likewise.
3081 (do_estimate_edge_size): Update.
3082 (do_estimate_edge_hints): Update.
3083 (growth_likely_positive): New function.
3084
3085 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3086
3087 PR target/60693
3088 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3089 also if addr has VOIDmode.
3090
3091 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3092
3093 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3094 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3095 Declare extern.
3096 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3097 instructions as well as AdvancedSIMD loads.
3098
3099 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3100
3101 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3102 Use crypto_aese type.
3103 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3104 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3105 crypto_aese, crypto_aesmc. Move to types.md.
3106 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3107 crypto_aesmc.
3108 * config/arm/iterators.md (crypto_type): Likewise.
3109
3110 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3111
3112 * cgraph.c: Include expr.h and tree-dfa.h.
3113 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3114 remove LHS.
3115
3116 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3117
3118 PR target/60675
3119 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3120 regs from checking multi-reg pseudos.
3121
3122 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3123
3124 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3125
3126 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3127
3128 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3129 if it would clobber the stack pointer, even temporarily.
3130
3131 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3132
3133 * mode-switching.c: Make small adjustments to the top comment.
3134
3135 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3136
3137 * config/rs6000/constraints.md (wD constraint): New constraint to
3138 match the constant integer to get the top DImode/DFmode out of a
3139 vector in a VSX register.
3140
3141 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3142 match the constant integer to get the top DImode/DFmode out of a
3143 vector in a VSX register.
3144
3145 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3146 for ISA 2.07.
3147
3148 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3149 vbpermq builtins.
3150
3151 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3152 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3153
3154 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3155 Optimize vec_extract of 64-bit values, where the value being
3156 extracted is in the top word, where we can use scalar
3157 instructions. Add direct move and store support. Combine the big
3158 endian/little endian vector select load support into a single insn.
3159 (vsx_extract_<mode>_internal1): Likewise.
3160 (vsx_extract_<mode>_internal2): Likewise.
3161 (vsx_extract_<mode>_load): Likewise.
3162 (vsx_extract_<mode>_store): Likewise.
3163 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3164 combined into vsx_extract_<mode>_load.
3165 (vsx_extract_<mode>_one_le): Likewise.
3166
3167 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3168 define the top 64-bit vector element.
3169
3170 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3171 constraint.
3172
3173 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3174 Document vec_vbpermq builtin.
3175
3176 PR target/60672
3177 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3178 enable use of xxsldwi and xxpermdi builtin functions.
3179 (vec_xxpermdi): Likewise.
3180
3181 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3182 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3183
3184 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3185
3186 PR rtl-optimization/60650
3187 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3188 first_p. Use it.
3189 (find_spills_for): New.
3190 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3191 Spill all pseudos on the second iteration.
3192
3193 2014-03-27 Marek Polacek <polacek@redhat.com>
3194
3195 PR c/50347
3196 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3197 types.
3198
3199 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3200
3201 * config/s390/s390.c (s390_can_use_return_insn): Check for
3202 call-saved FPRs on 31 bit.
3203
3204 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3205
3206 PR middle-end/60682
3207 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3208 if they need regimplification, just drop them instead of
3209 calling gimple_regimplify_operands on them.
3210
3211 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3212
3213 PR target/60580
3214 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3215 (aarch64_frame_pointer_required): Adjust logic.
3216 (aarch64_can_eliminate): Adjust logic.
3217 (aarch64_override_options_after_change): Adjust logic.
3218
3219 2014-03-27 Dehao Chen <dehao@google.com>
3220
3221 * ipa-inline.c (early_inliner): Update node's inline info.
3222
3223 2014-03-26 Dehao Chen <dehao@google.com>
3224
3225 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3226 compiler inserted conditional jumps for NAN float check.
3227
3228 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3229
3230 * ubsan.h (ubsan_create_data): Change second argument's type
3231 to const location_t *.
3232 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3233 _("<unknown>").
3234 (ubsan_create_data): Change second argument to const location_t *PLOC.
3235 Create Loc field whenever PLOC is non-NULL.
3236 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3237 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3238 callers.
3239
3240 PR other/59545
3241 * real.c (real_to_integer2): Change type of low to UHWI.
3242
3243 2014-03-26 Tobias Burnus <burnus@net-b.de>
3244
3245 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3246 (CILK_SELF_SPECS): New define.
3247 (driver_self_specs): Use it.
3248
3249 2014-03-26 Richard Biener <rguenther@suse.de>
3250
3251 * tree-pretty-print.c (percent_K_format): Implement special
3252 case for LTO and its stripped down BLOCK tree.
3253
3254 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3255
3256 PR sanitizer/60636
3257 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3258
3259 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3260 one range is range_int_cst_p, but not both, at least optimize
3261 addition/subtraction of 0 and multiplication by 0 or 1.
3262 * gimple-fold.c (gimple_fold_call): Fold
3263 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3264 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3265 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3266
3267 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3268
3269 PR rtl-optimization/60452
3270 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3271 <case REG>: Return 1 for invalid offsets from the frame pointer.
3272
3273 2014-03-26 Marek Polacek <polacek@redhat.com>
3274
3275 PR c/37428
3276 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3277 a structure/union.
3278
3279 2014-03-26 Marek Polacek <polacek@redhat.com>
3280
3281 PR c/39525
3282 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3283 field members.
3284
3285 2014-03-26 Marek Polacek <polacek@redhat.com>
3286
3287 PR other/59545
3288 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3289 multiplication in unsigned type.
3290
3291 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3292
3293 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3294
3295 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3296
3297 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3298
3299 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3300
3301 PR ipa/60315
3302 * cif-code.def (UNREACHABLE) New code.
3303 * ipa-inline.c (inline_small_functions): Skip edges to
3304 __builtlin_unreachable.
3305 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3306 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3307 predicate to __bulitin_unreachable.
3308 (set_cond_stmt_execution_predicate): Fix issue when
3309 invert_tree_comparison returns ERROR_MARK.
3310 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3311 propagate to inline clones.
3312 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3313 to unreachable.
3314 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3315 * cgraphclones.c (cgraph_clone_node): If call destination is already
3316 ureachable, do not redirect it back.
3317 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3318 unreachable.
3319
3320 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3321
3322 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3323 Do not modify inline clones.
3324
3325 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3326
3327 * config/i386/i386.md (general_sext_operand): New mode attr.
3328 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3329 don't generate (sign_extend (const_int)).
3330 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3331 operands[2]. Use We constraint instead of <i> and
3332 <general_sext_operand> predicate instead of <general_operand>.
3333 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3334 * config/i386/constraints.md (We): New constraint.
3335 * config/i386/predicates.md (x86_64_sext_operand,
3336 sext_operand): New predicates.
3337
3338 2014-03-25 Martin Jambor <mjambor@suse.cz>
3339
3340 PR ipa/60600
3341 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3342 inconsistent devirtualizations to __builtin_unreachable.
3343
3344 2014-03-25 Marek Polacek <polacek@redhat.com>
3345
3346 PR c/35449
3347 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3348
3349 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3350
3351 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3352 order of elements for big-endian.
3353
3354 2014-03-25 Richard Biener <rguenther@suse.de>
3355
3356 PR middle-end/60635
3357 * gimplify-me.c (gimple_regimplify_operands): Update the
3358 re-gimplifed stmt.
3359
3360 2014-03-25 Martin Jambor <mjambor@suse.cz>
3361
3362 PR ipa/59176
3363 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3364 (lto_output_varpool_node): Likewise.
3365 (input_overwrite_node): Likewise.
3366 (input_varpool_node): Likewise.
3367
3368 2014-03-25 Richard Biener <rguenther@suse.de>
3369
3370 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3371 (run_gcc): Likewise.
3372
3373 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3374
3375 * combine.c (simplify_compare_const): Add MODE argument.
3376 Handle mode_width 0 as very large mode_width.
3377 (try_combine, simplify_comparison): Adjust callers.
3378
3379 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3380 type to avoid signed integer overflow.
3381 * explow.c (plus_constant): Likewise.
3382
3383 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3384
3385 * doc/generic.texi: Correct typos.
3386
3387 2014-03-24 Tobias Burnus <burnus@net-b.de>
3388
3389 * doc/invoke.texi (-flto): Expand section about
3390 using static libraries with LTO.
3391
3392 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3393
3394 PR rtl-optimization/60501
3395 * optabs.def (addptr3_optab): New optab.
3396 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3397 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3398 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3399
3400 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3401
3402 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3403
3404 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3405
3406 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3407 _mm512_set1_pd.
3408
3409 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3410 (_mm256_undefined_ps): Define.
3411 (_mm256_undefined_pd): Define.
3412 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3413 (_mm_undefined_pd): Define.
3414 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3415 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3416 (_mm512_undefined_ps): Define.
3417 (_mm512_undefined_pd): Define.
3418 Use _mm*_undefined_*.
3419 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3420
3421 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3422
3423 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3424 (lshr_simd): DI mode added.
3425 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3426 (aarch64_ushr_simddi): Likewise.
3427 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3428 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3429 (vshrd_n_u64): Likewise.
3430
3431 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3432
3433 * Makefile.in (s-macro_list): Depend on cc1.
3434
3435 2014-03-23 Teresa Johnson <tejohnson@google.com>
3436
3437 * ipa-utils.c (ipa_print_order): Use specified dump file.
3438
3439 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3440
3441 PR rtl-optimization/60601
3442 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3443
3444 * gcc.c (eval_spec_function): Initialize save_growing_value.
3445
3446 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3447
3448 PR sanitizer/60613
3449 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3450 code == MINUS_EXPR, never swap op0 with op1.
3451
3452 * toplev.c (init_local_tick): Avoid signed integer multiplication
3453 overflow.
3454 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3455 shift by first operand's bitsize.
3456
3457 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3458
3459 PR target/60610
3460 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3461 redefine to 1 or 0.
3462 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3463 TARGET_ISA_64BIT_P(x).
3464
3465 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3466
3467 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3468 pattern for vector nor instead of subtract from splat(-1).
3469 (altivec_expand_vec_perm_const_le): Likewise.
3470
3471 2014-03-21 Richard Henderson <rth@twiddle.net>
3472
3473 PR target/60598
3474 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3475 related insns after epilogue_completed.
3476
3477 2014-03-21 Martin Jambor <mjambor@suse.cz>
3478
3479 PR ipa/59176
3480 * cgraph.h (symtab_node): New flag body_removed.
3481 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3482 when removing bodies.
3483 * symtab.c (dump_symtab_base): Dump body_removed flag.
3484 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3485 had their bodies removed.
3486
3487 2014-03-21 Martin Jambor <mjambor@suse.cz>
3488
3489 PR ipa/60419
3490 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3491 in the border.
3492
3493 2014-03-21 Richard Biener <rguenther@suse.de>
3494
3495 PR tree-optimization/60577
3496 * tree-core.h (struct tree_base): Document nothrow_flag use
3497 in VAR_DECL_NONALIASED.
3498 * tree.h (VAR_DECL_NONALIASED): New.
3499 (may_be_aliased): Adjust.
3500 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3501
3502 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3503
3504 * expr.c (expand_expr_real_1): Remove outdated comment.
3505
3506 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3507
3508 PR middle-end/60597
3509 * ira.c (adjust_cleared_regs): Call copy_rtx on
3510 *reg_equiv[REGNO (loc)].src_p before passing it to
3511 simplify_replace_fn_rtx.
3512
3513 PR target/60568
3514 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3515 into CONST, put pic register as first operand of PLUS. Use
3516 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3517
3518 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3519
3520 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3521
3522 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3523
3524 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3525 around for store forwarding issue in the FPU on the UT699.
3526 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3527 loads and operations if -mfix-ut699 is specified.
3528 (divtf3_hq): Tweak attribute.
3529 (sqrttf2_hq): Likewise.
3530
3531 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3532
3533 * calls.c (store_one_arg): Remove incorrect const qualification on the
3534 type of the temporary.
3535 * cfgexpand.c (expand_return): Likewise.
3536 * expr.c (expand_constructor): Likewise.
3537 (expand_expr_real_1): Likewise.
3538
3539 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3540
3541 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3542 of parts.
3543
3544 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3545
3546 PR target/60039
3547 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3548
3549 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3550
3551 * config/arm/aarch-common-protos.h
3552 (alu_cost_table): Fix spelling of "extend".
3553 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3554
3555 2014-03-19 Richard Biener <rguenther@suse.de>
3556
3557 PR middle-end/60553
3558 * tree-core.h (tree_type_common): Re-order pointer members
3559 to reduce recursion depth during GC walks.
3560
3561 2014-03-19 Marek Polacek <polacek@redhat.com>
3562
3563 PR sanitizer/60569
3564 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3565 before accessing it.
3566
3567 2014-03-19 Richard Biener <rguenther@suse.de>
3568
3569 PR lto/59543
3570 * lto-streamer-in.c (input_function): In WPA stage do not drop
3571 debug stmts.
3572
3573 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3574
3575 PR tree-optimization/60559
3576 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3577 with build_zero_cst assignment.
3578
3579 2014-03-18 Kai Tietz <ktietz@redhat.com>
3580
3581 PR rtl-optimization/56356
3582 * sdbout.c (sdbout_parms): Verify that parms'
3583 incoming argument is valid.
3584 (sdbout_reg_parms): Likewise.
3585
3586 2014-03-18 Richard Henderson <rth@redhat.com>
3587
3588 PR target/60562
3589 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3590 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3591 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3592
3593 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3594
3595 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3596 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3597 Italicize plugin event names in description. Explain that
3598 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3599 Remind that no GCC functions should be called after PLUGIN_FINISH.
3600 Explain what pragmas with expansion are.
3601
3602 2014-03-18 Martin Liska <mliska@suse.cz>
3603
3604 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3605 gimple call statement is update.
3606 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3607 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3608
3609 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3610
3611 PR sanitizer/60557
3612 * ubsan.c (ubsan_instrument_unreachable): Call
3613 initialize_sanitizer_builtins.
3614 (ubsan_pass): Likewise.
3615
3616 PR sanitizer/60535
3617 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3618 varpool_finalize_decl instead of rest_of_decl_compilation.
3619
3620 2014-03-18 Richard Biener <rguenther@suse.de>
3621
3622 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3623 by using bitmap_and_compl instead of bitmap_and_compl_into.
3624 (df_rd_transfer_function): Likewise.
3625
3626 2014-03-18 Richard Biener <rguenther@suse.de>
3627
3628 * doc/lto.texi (fresolution): Fix typo.
3629
3630 2014-03-18 Richard Biener <rguenther@suse.de>
3631
3632 * doc/invoke.texi (flto): Update for changes in 4.9.
3633
3634 2014-03-18 Richard Biener <rguenther@suse.de>
3635
3636 * doc/loop.texi: Remove section on the removed lambda framework.
3637 Update loop docs with recent changes in preserving loop structure.
3638
3639 2014-03-18 Richard Biener <rguenther@suse.de>
3640
3641 * doc/lto.texi (-fresolution): Document.
3642
3643 2014-03-18 Richard Biener <rguenther@suse.de>
3644
3645 * doc/contrib.texi: Adjust my name.
3646
3647 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3648
3649 PR ipa/58721
3650 * internal-fn.c: Include diagnostic-core.h.
3651 (expand_BUILTIN_EXPECT): New function.
3652 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3653 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3654 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3655 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3656 IFN_BUILTIN_EXPECT.
3657 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3658 Revert 3 argument __builtin_expect code.
3659 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3660 * gimple-fold.c (gimple_fold_call): Likewise.
3661 * tree.h (fold_builtin_expect): New prototype.
3662 * builtins.c (build_builtin_expect_predicate): Add predictor
3663 argument, if non-NULL, create 3 argument __builtin_expect.
3664 (fold_builtin_expect): No longer static. Add ARG2 argument,
3665 pass it through to build_builtin_expect_predicate.
3666 (fold_builtin_2): Adjust caller.
3667 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3668 * internal-fn.def (BUILTIN_EXPECT): New.
3669
3670 2014-03-18 Tobias Burnus <burnus@net-b.de>
3671
3672 PR ipa/58721
3673 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3674 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3675 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3676
3677 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3678
3679 PR ipa/58721
3680 * predict.c (combine_predictions_for_bb): Fix up formatting.
3681 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3682 fill what it points to if non-NULL.
3683 (tree_predict_by_opcode): Adjust caller, use the predictor.
3684 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3685
3686 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3687
3688 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3689 proper constant for the store mode.
3690
3691 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3692
3693 * symtab.c (change_decl_assembler_name): Fix transparent alias
3694 chain construction.
3695
3696 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3697
3698 * config/aarch64/aarch64.c: Correct the comments about the
3699 aarch64 stack layout.
3700
3701 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3702
3703 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3704 check for GF_OMP_FOR_KIND_FOR.
3705
3706 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3707
3708 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3709 ymm and zmm register names.
3710
3711 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3712
3713 PR target/60516
3714 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3715 note creation for the 2010-08-31 changes.
3716
3717 2014-03-17 Marek Polacek <polacek@redhat.com>
3718
3719 PR middle-end/60534
3720 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3721 as -fno-tree-loop-vectorize.
3722 (expand_omp_simd): Likewise.
3723
3724 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3725
3726 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3727 (eligible_for_call_delay): New prototype.
3728 * config/sparc/sparc.c (tls_call_delay): Rename into...
3729 (eligible_for_call_delay): ...this. Return false if the instruction
3730 cannot be put in the delay slot of a branch.
3731 (eligible_for_restore_insn): Simplify.
3732 (eligible_for_return_delay): Return false if the instruction cannot be
3733 put in the delay slot of a branch and simplify.
3734 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3735 put in the delay slot of a branch.
3736 * config/sparc/sparc.md (fix_ut699): New attribute.
3737 (tls_call_delay): Delete.
3738 (in_call_delay): Reimplement.
3739 (eligible_for_sibcall_delay): Rename into...
3740 (in_sibcall_delay): ...this.
3741 (eligible_for_return_delay): Rename into...
3742 (in_return_delay): ...this.
3743 (in_branch_delay): Reimplement.
3744 (in_uncond_branch_delay): Delete.
3745 (in_annul_branch_delay): Delete.
3746
3747 2014-03-14 Richard Henderson <rth@redhat.com>
3748
3749 PR target/60525
3750 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3751 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3752 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3753 (floathi splitters): Remove.
3754 (float<SWI48x>xf2): New pattern.
3755 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3756 code that tried to handle DImode for 32-bit, but which was excluded
3757 by the pattern's condition. Drop allocation of stack temporary.
3758 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3759 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3760 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3761 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3762 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3763 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3764 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3765 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3766 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3767 (*float<SWI48x><X87MODEF>2_i387): Remove.
3768 (all float _with_temp splitters): Remove.
3769 (*float<SWI48x><MODEF>2_i387): New pattern.
3770 (*float<SWI48><MODEF>2_sse): New pattern.
3771 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3772 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3773
3774 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3775 Marek Polacek <polacek@redhat.com>
3776
3777 PR middle-end/60484
3778 * common.opt (dump_base_name_prefixed): New Variable.
3779 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3780 if x_dump_base_name_prefixed is already set, set it at the end.
3781
3782 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3783
3784 PR rtl-optimization/60508
3785 * lra-constraints.c (get_reload_reg): Add new parameter
3786 in_subreg_p.
3787 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3788 Pass the new parameter values.
3789
3790 2014-03-14 Richard Biener <rguenther@suse.de>
3791
3792 * common.opt: Revert unintented changes from r205065.
3793 * opts.c: Likewise.
3794
3795 2014-03-14 Richard Biener <rguenther@suse.de>
3796
3797 PR middle-end/60518
3798 * cfghooks.c (split_block): Properly adjust all loops the
3799 block was a latch of.
3800
3801 2014-03-14 Martin Jambor <mjambor@suse.cz>
3802
3803 PR lto/60461
3804 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3805 and simplify it.
3806
3807 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3808
3809 PR target/59396
3810 * config/avr/avr.c (avr_set_current_function): Pass function name
3811 through default_strip_name_encoding before sanity checking instead
3812 of skipping the first char of the assembler name.
3813
3814 2014-03-13 Richard Henderson <rth@redhat.com>
3815
3816 PR debug/60438
3817 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3818 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3819 * config/i386/i386-protos.h: Likewise.
3820 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3821 in the expander instead of a splitter.
3822 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3823 any possibility of requiring a memory.
3824 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3825 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3826 (fp branch splitters): Update for ix86_split_fp_branch.
3827 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3828 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3829 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3830 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3831 (*fop_<MODEF>_3_i387): Likewise.
3832 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3833 (splitters for the fop_* register patterns): Remove.
3834 (fscalexf4_i387): Rename from *fscalexf4_i387.
3835 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3836
3837 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3838
3839 PR tree-optimization/59779
3840 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3841 type for bitsize and maxsize instead of HOST_WIDE_INT.
3842
3843 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3844
3845 PR rtl-optimization/57320
3846 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3847 the CFG after thread_prologue_and_epilogue_insns.
3848
3849 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3850
3851 PR rtl-optimization/57189
3852 * lra-constraints.c (process_alt_operands): Disfavor spilling
3853 vector pseudos.
3854
3855 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3856
3857 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3858
3859 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3860
3861 PR tree-optimization/59025
3862 PR middle-end/60418
3863 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3864 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3865
3866 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3867
3868 PR target/60486
3869 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3870 calls of avr_out_plus_1.
3871
3872 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3873
3874 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3875 BB's single pred and update the father loop's latch info later.
3876
3877 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3878
3879 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3880 (VEC_M): Likewise.
3881 (VEC_N): Likewise.
3882 (VEC_R): Likewise.
3883 (VEC_base): Likewise.
3884 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3885 registers, we need to swap double words in little endian mode.
3886
3887 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3888 to be a container mode for 128-bit integer operations added in ISA
3889 2.07. Unlike TImode and PTImode, the preferred register set is
3890 the Altivec/VMX registers for the 128-bit operations.
3891
3892 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3893 declarations.
3894 (rs6000_split_128bit_ok_p): Likewise.
3895
3896 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3897 macros for creating ISA 2.07 normal and overloaded builtin
3898 functions with 3 arguments.
3899 (BU_P8V_OVERLOAD_3): Likewise.
3900 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3901 for use as overloaded functions.
3902 (VPERM_1TI_UNS): Likewise.
3903 (VSEL_1TI): Likewise.
3904 (VSEL_1TI_UNS): Likewise.
3905 (ST_INTERNAL_1ti): Likewise.
3906 (LD_INTERNAL_1ti): Likewise.
3907 (XXSEL_1TI): Likewise.
3908 (XXSEL_1TI_UNS): Likewise.
3909 (VPERM_1TI): Likewise.
3910 (VPERM_1TI_UNS): Likewise.
3911 (XXPERMDI_1TI): Likewise.
3912 (SET_1TI): Likewise.
3913 (LXVD2X_V1TI): Likewise.
3914 (STXVD2X_V1TI): Likewise.
3915 (VEC_INIT_V1TI): Likewise.
3916 (VEC_SET_V1TI): Likewise.
3917 (VEC_EXT_V1TI): Likewise.
3918 (EQV_V1TI): Likewise.
3919 (NAND_V1TI): Likewise.
3920 (ORC_V1TI): Likewise.
3921 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3922 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3923 overloaded builtin.
3924 (VADDUQM): Likewise.
3925 (VSUBCUQ): Likewise.
3926 (VADDEUQM): Likewise.
3927 (VADDECUQ): Likewise.
3928 (VSUBEUQM): Likewise.
3929 (VSUBECUQ): Likewise.
3930
3931 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3932 __int128_t and __uint128_t types.
3933 (__uint128_type): Likewise.
3934 (altivec_categorize_keyword): Add support for vector __int128_t,
3935 vector __uint128_t, vector __int128, and vector unsigned __int128
3936 as a container type for TImode operations that need to be done in
3937 VSX/Altivec registers.
3938 (rs6000_macro_to_expand): Likewise.
3939 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3940 to support 128-bit integer instructions vaddcuq, vadduqm,
3941 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3942 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3943
3944 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3945 for V1TImode, and set up preferences to use VSX/Altivec registers.
3946 Setup VSX reload handlers.
3947 (rs6000_debug_reg_global): Likewise.
3948 (rs6000_init_hard_regno_mode_ok): Likewise.
3949 (rs6000_preferred_simd_mode): Likewise.
3950 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3951 (easy_altivec_constant): Likewise.
3952 (output_vec_const_move): Likewise.
3953 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3954 simple move.
3955 (rs6000_expand_vector_extract): Likewise.
3956 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3957 addressing.
3958 (rs6000_const_vec): Add support for V1TImode.
3959 (rs6000_emit_le_vsx_load): Swap double words when loading or
3960 storing TImode/V1TImode.
3961 (rs6000_emit_le_vsx_store): Likewise.
3962 (rs6000_emit_le_vsx_move): Likewise.
3963 (rs6000_emit_move): Add support for V1TImode.
3964 (altivec_expand_ld_builtin): Likewise.
3965 (altivec_expand_st_builtin): Likewise.
3966 (altivec_expand_vec_init_builtin): Likewise.
3967 (altivec_expand_builtin): Likewise.
3968 (rs6000_init_builtins): Add support for V1TImode type. Add
3969 support for ISA 2.07 128-bit integer builtins. Define type names
3970 for the VSX/Altivec vector types.
3971 (altivec_init_builtins): Add support for overloaded vector
3972 functions with V1TImode type.
3973 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
3974 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3975 external function.
3976 (rs6000_split_128bit_ok_p): Likewise.
3977 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3978 __int128_t and vector __uint128_t.
3979
3980 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3981 and mode attributes.
3982 (VSX_M): Likewise.
3983 (VSX_M2): Likewise.
3984 (VSm): Likewise.
3985 (VSs): Likewise.
3986 (VSr): Likewise.
3987 (VSv): Likewise.
3988 (VS_scalar): Likewise.
3989 (VS_double): Likewise.
3990 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
3991
3992 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
3993 we support the ISA 2.07 128-bit integer arithmetic instructions.
3994 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
3995 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
3996 and TImode types for use with the builtin functions.
3997 (V1TI_type_node): Likewise.
3998 (unsigned_V1TI_type_node): Likewise.
3999 (intTI_type_internal_node): Likewise.
4000 (uintTI_type_internal_node): Likewise.
4001
4002 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4003 128-bit builtin functions.
4004 (UNSPEC_VADDEUQM): Likewise.
4005 (UNSPEC_VADDECUQ): Likewise.
4006 (UNSPEC_VSUBCUQ): Likewise.
4007 (UNSPEC_VSUBEUQM): Likewise.
4008 (UNSPEC_VSUBECUQ): Likewise.
4009 (VM): Add V1TImode to vector mode iterators.
4010 (VM2): Likewise.
4011 (VI_unit): Likewise.
4012 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4013 (altivec_vaddcuq): Likewise.
4014 (altivec_vsubuqm): Likewise.
4015 (altivec_vsubcuq): Likewise.
4016 (altivec_vaddeuqm): Likewise.
4017 (altivec_vaddecuq): Likewise.
4018 (altivec_vsubeuqm): Likewise.
4019 (altivec_vsubecuq): Likewise.
4020
4021 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4022 mode iterators.
4023 (BOOL_128): Likewise.
4024 (BOOL_REGS_OUTPUT): Likewise.
4025 (BOOL_REGS_OP1): Likewise.
4026 (BOOL_REGS_OP2): Likewise.
4027 (BOOL_REGS_UNARY): Likewise.
4028 (BOOL_REGS_AND_CR0): Likewise.
4029
4030 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4031 128-bit integer builtin support.
4032 (vec_vadduqm): Likewise.
4033 (vec_vaddecuq): Likewise.
4034 (vec_vaddeuqm): Likewise.
4035 (vec_vsubecuq): Likewise.
4036 (vec_vsubeuqm): Likewise.
4037 (vec_vsubcuq): Likewise.
4038 (vec_vsubuqm): Likewise.
4039
4040 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4041 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4042 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4043 128-bit integer add/subtract to ISA 2.07.
4044
4045 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4046
4047 * config/arc/arc.c (arc_predicate_delay_insns):
4048 Fix third argument passed to conditionalize_nonjump.
4049
4050 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4051
4052 * config/aarch64/aarch64-builtins.c
4053 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4054 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4055 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4056 instead of __builtin_lfloor.
4057 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4058
4059 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4060
4061 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4062 (tree_ssa_ifcombine_bb_1): New function.
4063 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4064 is an empty forwarder block to then_bb or vice versa and then_bb
4065 and else_bb are effectively swapped.
4066
4067 2014-03-12 Christian Bruel <christian.bruel@st.com>
4068
4069 PR target/60264
4070 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4071 REG_CFA_DEF_CFA note.
4072 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4073 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4074
4075 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4076
4077 PR tree-optimization/60454
4078 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4079
4080 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4081
4082 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4083 Do not define target_cpu_default2 to generic.
4084 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4085 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4086 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4087
4088 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4089 Marc Glisse <marc.glisse@inria.fr>
4090
4091 PR tree-optimization/60502
4092 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4093 instead of build_low_bits_mask.
4094
4095 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4096
4097 PR middle-end/60482
4098 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4099 if there are multiple uses, but op doesn't live on E edge.
4100 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4101 clobber stmts before __builtin_unreachable.
4102
4103 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4104
4105 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4106 hard_frame_pointer_rtx.
4107 * cse.c (cse_insn): Remove volatile check.
4108 * cselib.c (cselib_process_insn): Likewise.
4109 * dse.c (scan_insn): Likewise.
4110
4111 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4112
4113 * config/arc/arc.c (conditionalize_nonjump): New function,
4114 broken out of ...
4115 (arc_ifcvt): ... this.
4116 (arc_predicate_delay_insns): Use it.
4117
4118 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4119
4120 * config/arc/predicates.md (extend_operand): During/after reload,
4121 allow const_int_operand.
4122 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4123 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4124 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4125 to "i".
4126 (umulsi3_highpart_i): Likewise.
4127
4128 2014-03-11 Richard Biener <rguenther@suse.de>
4129
4130 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4131 Add asserts to guard possible wrong-code bugs.
4132
4133 2014-03-11 Richard Biener <rguenther@suse.de>
4134
4135 PR tree-optimization/60429
4136 PR tree-optimization/60485
4137 * tree-ssa-structalias.c (set_union_with_increment): Properly
4138 take into account all fields that overlap the shifted vars.
4139 (do_sd_constraint): Likewise.
4140 (do_ds_constraint): Likewise.
4141 (get_constraint_for_ptr_offset): Likewise.
4142
4143 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4144
4145 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4146 (nios2_compute_frame_layout):
4147 Add calculation of cfun->machine->fp_save_offset.
4148 (nios2_expand_prologue): Correct setting of frame pointer register
4149 in prologue.
4150 (nios2_expand_epilogue): Update recovery of stack pointer from
4151 frame pointer accordingly.
4152 (nios2_initial_elimination_offset): Update calculation of offset
4153 for eliminating to HARD_FRAME_POINTER_REGNUM.
4154
4155 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4156
4157 PR ipa/60457
4158 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4159 cgraph_get_create_node on VAR_DECLs.
4160
4161 2014-03-10 Richard Biener <rguenther@suse.de>
4162
4163 PR middle-end/60474
4164 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4165
4166 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4167
4168 * config/vms/vms.opt (vms_float_format): New variable.
4169
4170 2014-03-08 Tobias Burnus <burnus@net-b.de>
4171
4172 * doc/invoke.texi (-fcilkplus): Update implementation status.
4173
4174 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4175 Richard Biener <rguenther@suse.de>
4176
4177 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4178 consistently accross all TUs.
4179 (run_gcc): Enable -fshort-double automatically at link at link-time
4180 and disallow override.
4181
4182 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4183
4184 PR target/58271
4185 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4186 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4187 if they can't be used.
4188
4189 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4190
4191 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4192 for Solaris 11/x86 ld.
4193 * configure: Regenerate.
4194
4195 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4196
4197 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4198 (LIB_TLS_SPEC): Save as ld_tls_libs.
4199 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4200 (HAVE_AS_IX86_TLSLDM): New test.
4201 * configure, config.in: Regenerate.
4202 * config/i386/i386.c (legitimize_tls_address): Fall back to
4203 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4204 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4205 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4206 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4207
4208 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4209
4210 * common.opt (fira-loop-pressure): Mark as optimization.
4211
4212 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4213
4214 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4215 an OpenMP mappable type.
4216
4217 2014-03-06 Matthias Klose <doko@ubuntu.com>
4218
4219 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4220 MULTILIB_OSDIRNAMES is not defined.
4221
4222 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4223 Meador Inge <meadori@codesourcery.com>
4224
4225 PR target/58595
4226 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4227 (arm_legitimize_address): Call legitimize_tls_address for any
4228 arm_tls_referenced_p expression, handle constant addend. Call it
4229 before testing for !TARGET_ARM.
4230 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4231
4232 2014-03-06 Richard Biener <rguenther@suse.de>
4233
4234 PR middle-end/60445
4235 PR lto/60424
4236 PR lto/60427
4237 Revert
4238 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4239
4240 * tree-streamer.c (record_common_node): Assert we don't record
4241 nodes with type double.
4242 (preload_common_node): Skip type double, complex double and double
4243 pointer since it is now frontend dependent due to fshort-double option.
4244
4245 2014-03-06 Richard Biener <rguenther@suse.de>
4246
4247 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4248 or -fno-lto is specified and the linker has full plugin support.
4249 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4250 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4251 * lto-wrapper.c (merge_and_complain): Merge compile-time
4252 optimization levels.
4253 (run_gcc): And pass it through to the link options.
4254
4255 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4256
4257 PR debug/60381
4258 Revert:
4259 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4260 PR debug/59992
4261 * cselib.c (remove_useless_values): Skip to avoid quadratic
4262 behavior if the condition moved from...
4263 (cselib_process_insn): ... here holds.
4264
4265 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4266
4267 PR plugins/59335
4268 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4269 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4270
4271 PR plugins/59335
4272 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4273 (TM_H): Add x86-tune.def.
4274
4275 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4276
4277 * config/aarch64/aarch64.c (generic_tunings):
4278 Use cortexa57_extra_costs.
4279
4280 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4281
4282 PR lto/60404
4283 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4284 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4285 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4286 cost for in_lto_p.
4287
4288 2014-03-04 Heiher <r@hev.cc>
4289
4290 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4291 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4292
4293 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4294
4295 * config/i386/predicates.md (const2356_operand): Change to ...
4296 (const2367_operand): ... this.
4297 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4298 const2367_operand.
4299 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4300 (*avx512pf_scatterpf<mode>sf): Ditto.
4301 (avx512pf_scatterpf<mode>df): Ditto.
4302 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4303 (*avx512pf_scatterpf<mode>df): Ditto.
4304 * config/i386/i386.c (ix86_expand_builtin): Update
4305 incorrect hint operand error message.
4306
4307 2014-03-04 Richard Biener <rguenther@suse.de>
4308
4309 * lto-section-in.c (lto_get_section_data): Fix const cast.
4310
4311 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4312
4313 * tree-streamer.c (record_common_node): Assert we don't record
4314 nodes with type double.
4315 (preload_common_node): Skip type double, complex double and double
4316 pointer since it is now frontend dependent due to fshort-double option.
4317
4318 2014-03-04 Richard Biener <rguenther@suse.de>
4319
4320 PR lto/60405
4321 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4322 (lto_input_toplevel_asms): Likewise.
4323 * lto-section-in.c (lto_get_section_data): Instead do it here
4324 for every section.
4325
4326 2014-03-04 Richard Biener <rguenther@suse.de>
4327
4328 PR tree-optimization/60382
4329 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4330 dead PHIs a reduction.
4331
4332 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4333
4334 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4335 hint value.
4336 (_mm_prefetch): Move out of GCC target("sse") pragma.
4337 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4338 GCC target("prfchw") pragma.
4339 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4340 for locality <= 2.
4341 * config/i386/i386.c (ix86_option_override_internal): Enable
4342 -mprfchw with -mprefetchwt1.
4343
4344 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4345
4346 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4347 Mark as varying.
4348
4349 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4350
4351 * opts.h (CL_PCH_IGNORE): Define.
4352 * targhooks.c (option_affects_pch_p):
4353 Return false for options that have CL_PCH_IGNORE set.
4354 * opt-functions.awk: Process PchIgnore.
4355 * doc/options.texi: Document PchIgnore.
4356
4357 * config/arc/arc.opt (misize): Add PchIgnore property.
4358
4359 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4360
4361 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4362 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4363 constraint on constants to permit them being loaded into
4364 GENERAL_REGS or BASE_REGS.
4365
4366 2014-03-03 Nick Clifton <nickc@redhat.com>
4367
4368 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4369 anti-cacnonical alternatives.
4370 (negandhi3_real): New pattern.
4371 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4372
4373 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4374
4375 * config/avr/avr-mcus.def: Remove atxmega16x1.
4376 * config/avr/avr-tables.opt: Regenerate.
4377 * config/avr/t-multilib: Regenerate.
4378 * doc/avr-mmcu.texi: Regenerate.
4379
4380 2014-03-03 Tobias Grosser <tobias@grosser.es>
4381 Mircea Namolaru <mircea.namolaru@inria.fr>
4382
4383 PR tree-optimization/58028
4384 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4385 scalar dimensions.
4386
4387 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4388
4389 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4390 not handled by recognizers.
4391
4392 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4393
4394 PR middle-end/60175
4395 * function.c (expand_function_end): Don't emit
4396 clobber_return_register sequence if clobber_after is a BARRIER.
4397 * cfgexpand.c (construct_exit_block): Append instructions before
4398 return_label to prev_bb.
4399
4400 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4401
4402 * config/rs6000/constraints.md: Document reserved use of "wc".
4403
4404 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4405
4406 PR ipa/60150
4407 * ipa.c (function_and_variable_visibility): When dissolving comdat
4408 group, also set all symbols to local.
4409
4410 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4411
4412 PR ipa/60306
4413
4414 Revert:
4415 2013-12-14 Jan Hubicka <jh@suse.cz>
4416 PR middle-end/58477
4417 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4418
4419 2014-03-02 Jon Beniston <jon@beniston.com>
4420
4421 PR bootstrap/48230
4422 PR bootstrap/50927
4423 PR bootstrap/52466
4424 PR target/46898
4425 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4426 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4427 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4428 (simple_return, *simple_return): New patterns
4429 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4430 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4431
4432 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4433
4434 * dwarf2out.c (gen_subprogram_die): Tidy.
4435
4436 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4437
4438 PR target/60071
4439 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4440 (*mov_t_msb_neg_negc): ... this new insn.
4441
4442 2014-02-28 Jason Merrill <jason@redhat.com>
4443
4444 PR c++/58678
4445 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4446 function.
4447
4448 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4449
4450 PR c++/60314
4451 * dwarf2out.c (decltype_auto_die): New static.
4452 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4453 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4454 (is_cxx_auto): Likewise.
4455
4456 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4457
4458 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4459 we are not using general regs only.
4460
4461 2014-02-28 Richard Biener <rguenther@suse.de>
4462
4463 PR target/60280
4464 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4465 previous fix and only allow to remove trivial pre-headers
4466 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4467 (remove_forwarder_block): Properly update the latch of a loop.
4468
4469 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4470
4471 PR debug/59992
4472 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4473 (cselib_preserved_hash_table): New.
4474 (preserve_constants_and_equivs): Move preserved vals to it.
4475 (cselib_find_slot): Look it up first.
4476 (cselib_init): Initialize it.
4477 (cselib_finish): Release it.
4478 (dump_cselib_table): Dump it.
4479
4480 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4481
4482 PR debug/59992
4483 * cselib.c (remove_useless_values): Skip to avoid quadratic
4484 behavior if the condition moved from...
4485 (cselib_process_insn): ... here holds.
4486
4487 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4488
4489 PR debug/57232
4490 * var-tracking.c (vt_initialize): Apply the same condition to
4491 preserve the CFA base value.
4492
4493 2014-02-28 Joey Ye <joey.ye@arm.com>
4494
4495 PR target/PR60169
4496 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4497 if reload in progress or completed.
4498
4499 2014-02-28 Tobias Burnus <burnus@net-b.de>
4500
4501 PR middle-end/60147
4502 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4503 NAMELIST_DECL.
4504
4505 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4506
4507 * doc/tm.texi.in (Condition Code Status): Update documention for
4508 relative locations of cc0-setter and cc0-user.
4509
4510 2014-02-27 Jeff Law <law@redhat.com>
4511
4512 PR rtl-optimization/52714
4513 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4514 into two independent simple sets, if I3 is a jump, ensure the
4515 pattern we place into I3 is a (set (pc) ...).
4516
4517 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4518 Jeff Law <law@redhat.com>
4519
4520 PR rtl-optimization/49847
4521 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4522 are in different blocks.
4523 * doc/tm.texi (Condition Code Status): Update documention for
4524 relative locations of cc0-setter and cc0-user.
4525
4526 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4527
4528 PR target/59222
4529 * lra.c (lra_emit_add): Check SUBREG too.
4530
4531 2014-02-27 Andreas Schwab <schwab@suse.de>
4532
4533 * config/m68k/m68k.c (m68k_option_override): Disable
4534 -flive-range-shrinkage for classic m68k.
4535 (m68k_override_options_after_change): Likewise.
4536
4537 2014-02-27 Marek Polacek <polacek@redhat.com>
4538
4539 PR middle-end/59223
4540 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4541 -Wmaybe-uninitialized.
4542
4543 2014-02-27 Alan Modra <amodra@gmail.com>
4544
4545 PR target/57936
4546 * reload1.c (emit_input_reload_insns): When reload_override_in,
4547 set old to rl->in_reg when rl->in_reg is a subreg.
4548
4549 2014-02-26 Richard Biener <rguenther@suse.de>
4550
4551 PR bootstrap/60343
4552 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4553
4554 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4555
4556 * common/config/i386/predicates.md (const1256_operand): Remove.
4557 (const2356_operand): New.
4558 (const_1_to_2_operand): Remove.
4559 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4560 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4561 (*avx512pf_gatherpf<mode>sf): Ditto.
4562 (avx512pf_gatherpf<mode>df): Ditto.
4563 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4564 (*avx512pf_gatherpf<mode>df): Ditto.
4565 (avx512pf_scatterpf<mode>sf): Ditto.
4566 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4567 (*avx512pf_scatterpf<mode>sf): Ditto.
4568 (avx512pf_scatterpf<mode>df): Ditto.
4569 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4570 (*avx512pf_scatterpf<mode>df): Ditto.
4571 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4572
4573 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4574
4575 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4576 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4577 (_mm512_mask_testn_epi64_mask): Move to ...
4578 * config/i386/avx512cdintrin.h: Here.
4579 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4580 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4581 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4582 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4583 TARGET_AVX512F from TARGET_AVX512CD.
4584
4585 2014-02-26 Richard Biener <rguenther@suse.de>
4586
4587 PR ipa/60327
4588 * ipa.c (walk_polymorphic_call_targets): Properly guard
4589 call to inline_update_overall_summary.
4590
4591 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4592
4593 PR target/60280
4594 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4595 and latches only if requested. Fix latch if it is removed.
4596 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4597 LOOPS_HAVE_PREHEADERS.
4598
4599 2014-02-25 Andrew Pinski <apinski@cavium.com>
4600
4601 * builtins.c (expand_builtin_thread_pointer): Create a new target
4602 when the target is NULL.
4603
4604 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4605
4606 PR rtl-optimization/60317
4607 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4608 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4609 * lra-assigns.c: Include params.h.
4610 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4611 other reload pseudos considerations.
4612
4613 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4614
4615 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4616 to use canonical form for nor<mode>3.
4617
4618 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4619
4620 PR target/55426
4621 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4622 conversions.
4623
4624 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4625
4626 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4627 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4628 (ix86_handle_option): Handle OPT_mprefetchwt1.
4629 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4630 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4631 PREFETCHWT1 CPUID.
4632 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4633 OPTION_MASK_ISA_PREFETCHWT1.
4634 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4635 (PTA_PREFETCHWT1): New.
4636 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4637 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4638 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4639 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4640 (*prefetch_avx512pf_<mode>_: Change into ...
4641 (*prefetch_prefetchwt1_<mode>: This.
4642 * config/i386/i386.opt (mprefetchwt1): New.
4643 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4644 (_mm_prefetch): Handle intent to write.
4645 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4646
4647 2014-02-25 Richard Biener <rguenther@suse.de>
4648
4649 PR middle-end/60291
4650 * emit-rtl.c (mem_attrs_htab): Remove.
4651 (mem_attrs_htab_hash): Likewise.
4652 (mem_attrs_htab_eq): Likewise.
4653 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4654 (init_emit_once): Do not allocate mem_attrs_htab.
4655
4656 2014-02-25 Richard Biener <rguenther@suse.de>
4657
4658 PR lto/60319
4659 * lto-opts.c (lto_write_options): Output non-explicit conservative
4660 -fwrapv, -fno-trapv and -fno-strict-overflow.
4661 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4662 (run_gcc): And pass them through.
4663
4664 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4665
4666 * sel-sched.c (calculate_new_fences): New parameter ptime.
4667 Calculate it as a maximum over all fence cycles.
4668 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4669 Print the final schedule timing when sched_verbose.
4670
4671 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4672
4673 PR rtl-optimization/60292
4674 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4675 bit fot the fence instruction.
4676
4677 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4678
4679 * calls.h: Fix typo in comment.
4680
4681 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4682
4683 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4684 adjusting offsetable addresses.
4685
4686 2014-02-24 Guozhi Wei <carrot@google.com>
4687
4688 * sparseset.h (sparseset_pop): Fix the wrong index.
4689
4690 2014-02-24 Walter Lee <walt@tilera.com>
4691
4692 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4693 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4694 triplet.
4695 * common/config/tilegx/tilegx-common.c
4696 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4697 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4698 (LINK_SPEC): Ditto.
4699 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4700 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4701 (tilegx_gimplify_va_arg_expr): Handle big endian.
4702 (tilegx_expand_unaligned_load): Ditto.
4703 (tilegx_expand_unaligned_store): Ditto.
4704 (TARGET_RETURN_IN_MSB): New.
4705 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4706 (TARGET_ENDIAN_DEFAULT): New.
4707 (TARGET_BIG_ENDIAN): Handle big endian.
4708 (BYTES_BIG_ENDIAN): Ditto.
4709 (WORDS_BIG_ENDIAN): Ditto.
4710 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4711 (ENDIAN_SPEC): New.
4712 (EXTRA_SPECS): New.
4713 * config/tilegx/tilegx.md (extv): Handle big endian.
4714 (extzv): Ditto.
4715 (insn_st<n>): Ditto.
4716 (insn_st<n>_add<bitsuffix>): Ditto.
4717 (insn_stnt<n>): Ditto.
4718 (insn_stnt<n>_add<bitsuffix>):Ditto.
4719 (vec_interleave_highv8qi): Handle big endian.
4720 (vec_interleave_highv8qi_be): New.
4721 (vec_interleave_highv8qi_le): New.
4722 (insn_v1int_h): Handle big endian.
4723 (vec_interleave_lowv8qi): Handle big endian.
4724 (vec_interleave_lowv8qi_be): New.
4725 (vec_interleave_lowv8qi_le): New.
4726 (insn_v1int_l): Handle big endian.
4727 (vec_interleave_highv4hi): Handle big endian.
4728 (vec_interleave_highv4hi_be): New.
4729 (vec_interleave_highv4hi_le): New.
4730 (insn_v2int_h): Handle big endian.
4731 (vec_interleave_lowv4hi): Handle big endian.
4732 (vec_interleave_lowv4hi_be): New.
4733 (vec_interleave_lowv4hi_le): New.
4734 (insn_v2int_l): Handle big endian.
4735 (vec_interleave_highv2si): Handle big endian.
4736 (vec_interleave_highv2si_be): New.
4737 (vec_interleave_highv2si_le): New.
4738 (insn_v4int_h): Handle big endian.
4739 (vec_interleave_lowv2si): Handle big endian.
4740 (vec_interleave_lowv2si_be): New.
4741 (vec_interleave_lowv2si_le): New.
4742 (insn_v4int_l): Handle big endian.
4743 * config/tilegx/tilegx.opt (mbig-endian): New option.
4744 (mlittle-endian): New option.
4745 * doc/install.texi: Document tilegxbe-linux.
4746 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4747
4748 2014-02-24 Martin Jambor <mjambor@suse.cz>
4749
4750 PR ipa/60266
4751 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4752 there are no parameter descriptors.
4753
4754 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4755
4756 PR rtl-optimization/60268
4757 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4758 initialization to ...
4759 (sched_rgn_init): ... here.
4760 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4761
4762 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4763
4764 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4765 names.
4766
4767 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4768
4769 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4770 definition.
4771
4772 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4773
4774 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4775 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4776
4777 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4778
4779 * config/microblaze/predicates.md: Add cmp_op predicate.
4780 * config/microblaze/microblaze.md: Add branch_compare instruction
4781 which uses cmp_op predicate and emits cmp insn before branch.
4782 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4783 to microblaze_expand_conditional_branch and consolidate logic.
4784 (microblaze_expand_conditional_branch): emit branch_compare
4785 insn instead of handling cmp op separate from branch insn.
4786
4787 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4788
4789 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4790 to permit subregs.
4791
4792 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4793
4794 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4795 define_insn with define_expand and new define_insn
4796 *altivec_lve<VI_char>x_internal.
4797 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4798 and new define_insn *altivec_stve<VI_char>x_internal.
4799 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4800 prototype.
4801 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4802 lve*x built-ins.
4803 (altivec_expand_stvex_be): New function.
4804
4805 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4806
4807 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4808 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4809 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4810 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4811
4812 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4813
4814 PR target/60298
4815 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4816 instead of emit_move_insn.
4817
4818 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4819
4820 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4821 vspltw with vsldoi.
4822 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4823 gen_altivec_vsumsws.
4824
4825 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4826
4827 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4828 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4829 (altivec_lvxl_<mode>): New define_expand incorporating
4830 -maltivec=be semantics where needed.
4831 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4832 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4833 semantics where needed.
4834 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4835 (altivec_stvx_<mode>): New define_expand incorporating
4836 -maltivec=be semantics where needed.
4837 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4838 VM2 iterator instead of V4SI.
4839 (altivec_stvxl_<mode>): New define_expand incorporating
4840 -maltivec=be semantics where needed.
4841 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4842 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4843 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4844 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4845 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4846 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4847 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4848 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4849 ALTIVEC_BUILTIN_STVXL.
4850 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4851 (altivec_expand_stvx_be): Likewise.
4852 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4853 (altivec_expand_lvx_be): Likewise.
4854 (altivec_expand_stvx_be): Likewise.
4855 (altivec_expand_builtin): Add cases for
4856 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4857 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4858 (altivec_init_builtins): Add definitions for
4859 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4860 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4861
4862 2014-02-21 Catherine Moore <clm@codesourcery.com>
4863
4864 * doc/invoke.texi (mvirt, mno-virt): Document.
4865 * config/mips/mips.opt (mvirt): New option.
4866 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4867
4868 2014-02-21 Richard Biener <rguenther@suse.de>
4869
4870 PR tree-optimization/60276
4871 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4872 (STMT_VINFO_MIN_NEG_DIST): New macro.
4873 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4874 STMT_VINFO_MIN_NEG_DIST.
4875 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4876 made for negative dependence distances still hold.
4877
4878 2014-02-21 Richard Biener <rguenther@suse.de>
4879
4880 PR middle-end/60291
4881 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4882 DECL_INITIAL for globals not in the current function context.
4883
4884 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4885
4886 PR tree-optimization/56490
4887 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4888 * tree-ssa-uninit.c: Include params.h.
4889 (compute_control_dep_chain): Add num_calls argument, return false
4890 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4891 num_calls to recursive call.
4892 (find_predicates): Change dep_chain into normal array,
4893 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4894 variable and adjust compute_control_dep_chain caller.
4895 (find_def_preds): Likewise.
4896
4897 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4898
4899 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4900 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4901
4902 2014-02-21 Nick Clifton <nickc@redhat.com>
4903
4904 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4905 (pushhi1): Likewise.
4906 (popqi1): Add mode to pre_dec.
4907 (pophi1): Likewise.
4908
4909 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4910
4911 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4912 mode for mask of V8SFmode permutation.
4913
4914 2014-02-20 Richard Henderson <rth@redhat.com>
4915
4916 PR c++/60272
4917 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4918 a new pseudo for OLDVAL.
4919
4920 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4921
4922 PR target/57896
4923 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4924 gen_reg_rtx if d->testing_p.
4925 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4926 if d->testing_p and we will certainly return true.
4927 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4928 if d->testing_p.
4929
4930 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4931
4932 * emit-rtl.c (gen_reg_rtx): Assert that
4933 crtl->emit.regno_pointer_align_length is non-zero.
4934
4935 2014-02-20 Richard Henderson <rth@redhat.com>
4936
4937 PR c++/60272
4938 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4939 on failure the store back into EXPECT.
4940
4941 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4942 Sandra Loosemore <sandra@codesourcery.com>
4943
4944 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4945 * config/nios2/nios2.c (nios2_function_profiler): Add
4946 -fPIC (flag_pic == 2) support.
4947 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4948 (nios2_large_offset_p): New function.
4949 (nios2_unspec_reloc_p): Move up position, update to use
4950 nios2_large_offset_p.
4951 (nios2_unspec_address): Remove function.
4952 (nios2_unspec_offset): New function.
4953 (nios2_large_got_address): New function.
4954 (nios2_got_address): Add large offset support.
4955 (nios2_legitimize_tls_address): Update usage of removed and new
4956 functions.
4957 (nios2_symbol_binds_local_p): New function.
4958 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4959 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4960 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4961 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4962 processing for (const (unspec ...)).
4963 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4964
4965 2014-02-20 Richard Biener <rguenther@suse.de>
4966
4967 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4968 doing the substitution.
4969 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
4970
4971 2014-02-20 Martin Jambor <mjambor@suse.cz>
4972
4973 PR ipa/55260
4974 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4975 info when checking whether lattices are bottom.
4976
4977 2014-02-20 Richard Biener <rguenther@suse.de>
4978
4979 PR middle-end/60221
4980 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4981 regions at -O0.
4982
4983 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
4984
4985 PR ipa/58555
4986 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
4987 parameter specifying the scaling.
4988 (inline_call): Update.
4989 (want_inline_recursively): Guard division by zero.
4990 (recursive_inlining): Update.
4991 * ipa-inline.h (clone_inlined_nodes): Update.
4992
4993 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
4994
4995 PR target/60204
4996 * config/i386/i386.c (classify_argument): Pass structures of size
4997 64 bytes or less in register.
4998
4999 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5000 Kirill Yukhin <kirill.yukhin@intel.com>
5001
5002 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5003 (_mm_rcp28_round_ss): Ditto.
5004 (_mm_rsqrt28_round_sd): Ditto.
5005 (_mm_rsqrt28_round_ss): Ditto.
5006 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5007 (_mm_rcp14_round_ss): Ditto.
5008 (_mm_rsqrt14_round_sd): Ditto.
5009 (_mm_rsqrt14_round_ss): Ditto.
5010 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5011 the first input operand, get rid of match_dup.
5012 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5013 attribute to sse.
5014 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5015 Ditto.
5016 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5017 operand as the first input operand, set type attribute.
5018 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5019 Set type attribute.
5020 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5021 operand as the first input operand, set type attribute.
5022
5023 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5024
5025 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5026 bit of zero.
5027
5028 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5029
5030 PR target/60207
5031 * config/i386/i386.c (construct_container): Remove TFmode check
5032 for X86_64_INTEGER_CLASS.
5033
5034 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5035
5036 PR target/59794
5037 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5038 only when -Wpsabi is enabled.
5039
5040 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5041
5042 PR target/59799
5043 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5044 passing arrays in registers are the same as for structs, so remove the
5045 special case for them.
5046
5047 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5048
5049 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5050 destination type, extract only the valid bits if the source type is not
5051 integral and has a different mode.
5052
5053 2014-02-19 Richard Biener <rguenther@suse.de>
5054
5055 PR ipa/60243
5056 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5057 for all calls.
5058
5059 2014-02-19 Richard Biener <rguenther@suse.de>
5060
5061 PR ipa/60243
5062 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5063 (ipa_modify_call_arguments): Emit an argument load explicitely and
5064 preserve virtual SSA form there and for the replacement call.
5065 Do not update SSA form nor free dominance info.
5066
5067 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5068
5069 * ipa.c (function_and_variable_visibility): Also clear WEAK
5070 flag when disolving COMDAT_GROUP.
5071
5072 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5073
5074 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5075 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5076 not devirtualizing.
5077 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5078 do more sanity checks.
5079 (detect_type_change): Return true when giving up early.
5080 (compute_complex_assign_jump_func): Fix type parameter of
5081 ipa_set_ancestor_jf.
5082 (compute_complex_ancestor_jump_func): Likewise.
5083 (update_jump_functions_after_inlining): Fix updating of
5084 ancestor function.
5085 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5086
5087 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5088
5089 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5090 inline clones when edge disappears.
5091
5092 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5093
5094 PR target/60203
5095 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5096 Split 64-bit moves into 2 patterns. Do not allow the use of
5097 direct move for TDmode in little endian, since the decimal value
5098 has little endian bytes within a word, but the 64-bit pieces are
5099 ordered in a big endian fashion, and normal subreg's of TDmode are
5100 not allowed.
5101 (mov<mode>_64bit_dm): Likewise.
5102 (movtd_64bit_nodm): Likewise.
5103
5104 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5105
5106 PR tree-optimization/60174
5107 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5108 statement of an SSA_NAME that occurs in an abnormal PHI node.
5109
5110 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5111
5112 PR sanitizer/60142
5113 * final.c (SEEN_BB): Remove.
5114 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5115 (final_scan_insn): Don't force_source_line on second
5116 NOTE_INSN_BASIC_BLOCK.
5117
5118 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5119
5120 PR target/60205
5121 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5122 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5123 (type_natural_mode): Warn ABI change when %zmm register is not
5124 available for AVX512F vector value passing.
5125
5126 2014-02-18 Kai Tietz <ktietz@redhat.com>
5127
5128 PR target/60193
5129 * config/i386/i386.c (ix86_expand_prologue): Use value in
5130 rax register as displacement when restoring %r10 or %rax.
5131 Fix wrong offset when restoring both registers.
5132
5133 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5134
5135 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5136 assertion with conditional return.
5137
5138 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5139 Uros Bizjak <ubizjak@gmail.com>
5140
5141 PR driver/60233
5142 * config/i386/driver-i386.c (host_detect_local_cpu): If
5143 YMM state is not saved by the OS, also clear has_f16c. Move
5144 CPUID 0x80000001 handling before YMM state saving checking.
5145
5146 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5147
5148 PR rtl-optimization/58960
5149 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5150 factored out from ...
5151 (sched_init): ... here.
5152 (free_global_sched_pressure_data): New, factored out from ...
5153 (sched_finish): ... here.
5154 * sched-int.h (free_global_sched_pressure_data): Declare.
5155 * sched-rgn.c (nr_regions_initial): New static global.
5156 (haifa_find_rgns): Initialize it.
5157 (schedule_region): Disable sched-pressure for the newly
5158 generated regions.
5159
5160 2014-02-17 Richard Biener <rguenther@suse.de>
5161
5162 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5163 release SSA defs of pattern stmts.
5164
5165 2014-02-17 Richard Biener <rguenther@suse.de>
5166
5167 * tree-inline.c (expand_call_inline): Release the virtual
5168 operand defined by the call we are about to inline.
5169
5170 2014-02-17 Richard Biener <rguenther@suse.de>
5171
5172 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5173
5174 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5175 Ilya Tocar <ilya.tocar@intel.com>
5176
5177 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5178 arguments order in builtin.
5179 (_mm512_permutexvar_epi64): Ditto.
5180 (_mm512_mask_permutexvar_epi64): Ditto
5181 (_mm512_maskz_permutexvar_epi32): Ditto
5182 (_mm512_permutexvar_epi32): Ditto
5183 (_mm512_mask_permutexvar_epi32): Ditto
5184
5185 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5186
5187 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5188 (p8_vmrgow): Likewise.
5189
5190 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5191
5192 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5193 endian targets.
5194
5195 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5196
5197 PR target/60203
5198 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5199 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5200 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5201 using direct move instructions on ISA 2.07. Also adjust
5202 instruction length for 64-bit.
5203 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5204 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5205
5206 2014-02-15 Alan Modra <amodra@gmail.com>
5207
5208 PR target/58675
5209 PR target/57935
5210 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5211 find_replacement on parts of insn rtl that might be reloaded.
5212
5213 2014-02-15 Richard Biener <rguenther@suse.de>
5214
5215 PR tree-optimization/60183
5216 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5217 (tree_ssa_phiprop): Calculate and free post-dominators.
5218
5219 2014-02-14 Jeff Law <law@redhat.com>
5220
5221 PR rtl-optimization/60131
5222 * ree.c (get_extended_src_reg): New function.
5223 (combine_reaching_defs): Use it rather than assuming location of REG.
5224 (find_and_remove_re): Verify first operand of extension is
5225 a REG before adding the insns to the copy list.
5226
5227 2014-02-14 Roland McGrath <mcgrathr@google.com>
5228
5229 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5230 * configure: Regenerated.
5231 * config.in: Regenerated.
5232 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5233 instead of ASM_SHORT.
5234
5235 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5236 Richard Earnshaw <rearnsha@arm.com>
5237
5238 PR rtl-optimization/59535
5239 * lra-constraints.c (process_alt_operands): Encourage alternative
5240 when unassigned pseudo class is superset of the alternative class.
5241 (inherit_reload_reg): Don't inherit when optimizing for code size.
5242 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5243 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5244 modes not less than 4 for Thumb1.
5245
5246 2014-02-14 Kyle McMartin <kyle@redhat.com>
5247
5248 PR pch/60010
5249 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5250
5251 2014-02-14 Richard Biener <rguenther@suse.de>
5252
5253 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5254 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5255 Do not strip INDIRECT_REFs.
5256
5257 2014-02-14 Richard Biener <rguenther@suse.de>
5258
5259 PR lto/60179
5260 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5261 DECL_FUNCTION_SPECIFIC_TARGET.
5262 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5263 * tree-streamer-out.c (pack_ts_target_option): Remove.
5264 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5265 (write_ts_function_decl_tree_pointers): Do not stream
5266 DECL_FUNCTION_SPECIFIC_TARGET.
5267 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5268 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5269 (lto_input_ts_function_decl_tree_pointers): Do not stream
5270 DECL_FUNCTION_SPECIFIC_TARGET.
5271
5272 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5273
5274 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5275 (get_initial_def_for_induction, vectorizable_induction): Ignore
5276 debug stmts when looking for exit_phi.
5277 (vectorizable_live_operation): Fix up condition.
5278
5279 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5280
5281 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5282 nreverse() because it changes the content of original tree list.
5283
5284 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5285
5286 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5287 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5288
5289 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5290
5291 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5292 GNU coding standards.
5293
5294 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5295
5296 PR debug/60152
5297 * dwarf2out.c (gen_subprogram_die): Don't call
5298 add_calling_convention_attribute if subr_die is old_die.
5299
5300 2014-02-13 Sharad Singhai <singhai@google.com>
5301
5302 * doc/optinfo.texi: Fix order of nodes.
5303
5304 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5305
5306 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5307 operands[2], not operands[3].
5308
5309 2014-02-13 Richard Biener <rguenther@suse.de>
5310
5311 PR bootstrap/59878
5312 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5313 mention the possibility of an in-tree build.
5314 (CLooG): Update recommended version to 0.18.1, mention the
5315 possibility of an in-tree build and clarify that the ISL
5316 bundled with CLooG does not work.
5317
5318 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5319
5320 PR target/43546
5321 * expr.c (compress_float_constant): If x is a hard register,
5322 extend into a pseudo and then move to x.
5323
5324 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5325
5326 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5327 caused by bad second argument to warning_at() with -mhotpatch and
5328 nested functions (e.g. with gfortran).
5329
5330 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5331
5332 * opts.c (option_name): Remove "enabled by default" rider.
5333
5334 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5335
5336 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5337
5338 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5339 Uros Bizjak <ubizjak@gmail.com>
5340
5341 PR target/60151
5342 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5343 * configure: Regenerated.
5344
5345 2014-02-12 Richard Biener <rguenther@suse.de>
5346
5347 * vec.c (vec_prefix::calculate_allocation): Move as
5348 inline variant to vec.h.
5349 (vec_prefix::calculate_allocation_1): New out-of-line version.
5350 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5351 (vec_prefix::m_has_auto_buf): Rename to ...
5352 (vec_prefix::m_using_auto_storage): ... this.
5353 (vec_prefix::calculate_allocation): Inline the easy cases
5354 and dispatch to calculate_allocation_1 which doesn't need the
5355 prefix address.
5356 (va_heap::reserve): Use gcc_checking_assert.
5357 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5358 m_using_auto_storage.
5359 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5360 member and adjust.
5361 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5362 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5363 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5364
5365 2014-02-12 Richard Biener <rguenther@suse.de>
5366
5367 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5368 when we found a dependence.
5369
5370 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5371
5372 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5373 common code...
5374 (maybe_fold_stmt): ... into this new function.
5375 * omp-low.c (lower_omp): Update comment.
5376
5377 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5378 last use.
5379
5380 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5381 dereference.
5382
5383 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5384
5385 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5386 identifiers in comments.
5387 (cortexa53_extra_costs): Likewise.
5388 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5389 (cortexa7_extra_costs): Likewise.
5390 (cortexa12_extra_costs): Likewise.
5391 (cortexa15_extra_costs): Likewise.
5392 (v7m_extra_costs): Likewise.
5393
5394 2014-02-12 Richard Biener <rguenther@suse.de>
5395
5396 PR middle-end/60092
5397 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5398 of posix_memalign being successful.
5399 (lower_stmt): Restrict lowering of posix_memalign to when
5400 -ftree-bit-ccp is enabled.
5401
5402 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5403
5404 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5405 arg_loc.
5406 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5407
5408 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5409
5410 PR rtl-optimization/60116
5411 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5412 other_insn once the combination has been validated.
5413
5414 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5415
5416 PR lto/59468
5417 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5418 and wrapper.
5419 * ipa-devirt.c: Include demangle.h
5420 (odr_violation_reported): New static variable.
5421 (add_type_duplicate): Update odr_violations.
5422 (maybe_record_node): Add completep parameter; update it.
5423 (record_target_from_binfo): Add COMPLETEP parameter;
5424 update it as needed.
5425 (possible_polymorphic_call_targets_1): Likewise.
5426 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5427 rename FINAL to COMPLETE.
5428 (record_targets_from_bases): Sanity check we found the binfo;
5429 fix COMPLETEP updating.
5430 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5431 parameter, fix computing of COMPLETEP.
5432 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5433 at LTO time do demangling.
5434 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5435 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5436 parameter.
5437 (gimple_get_virt_method_for_binfo): Likewise.
5438 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5439 gimple_get_virt_method_for_vtable): Update prototypes.
5440
5441 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5442
5443 PR target/49008
5444 * genautomata.c (add_presence_absence): Fix typo with
5445 {final_}presence_list.
5446
5447 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5448
5449 PR target/60137
5450 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5451 for VSX/Altivec vectors that land in GPR registers.
5452
5453 2014-02-11 Richard Henderson <rth@redhat.com>
5454 Jakub Jelinek <jakub@redhat.com>
5455
5456 PR debug/59776
5457 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5458 around drhs if type conversion to lacc->type is not useless.
5459
5460 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5461
5462 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5463 tuning struct.
5464 (cortex-a57.cortex-a53): Likewise.
5465 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5466
5467 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5468
5469 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5470 arm_restrict_it.
5471
5472 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5473
5474 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5475 add_options_for_arm_vfp3.
5476
5477 2014-02-11 Jeff Law <law@redhat.com>
5478
5479 PR middle-end/54041
5480 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5481 object with an undesirable mode.
5482
5483 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5484
5485 PR libgomp/60107
5486 * config/i386/sol2-9.h: New file.
5487 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5488 *-*-solaris2.9*): Use it.
5489
5490 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5491
5492 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5493 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5494
5495 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5496
5497 * config/microblaze/microblaze.c: Extend mcpu version format
5498
5499 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5500
5501 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5502
5503 2014-02-10 Richard Henderson <rth@redhat.com>
5504
5505 PR target/59927
5506 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5507 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5508 ms-abi vs -mno-accumulate-outgoing-args.
5509 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5510 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5511 respect to ms-abi.
5512
5513 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5514
5515 PR middle-end/60080
5516 * cfgexpand.c (expand_asm_operands): Attach source location to
5517 ASM_INPUT rtx objects.
5518 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5519
5520 2014-02-10 Nick Clifton <nickc@redhat.com>
5521
5522 * config/mn10300/mn10300.c (popcount): New function.
5523 (mn10300_expand_prologue): Include saved registers in stack usage
5524 count.
5525
5526 2014-02-10 Jeff Law <law@redhat.com>
5527
5528 PR middle-end/52306
5529 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5530 when changing the SET_DEST of a prior insn to avoid an input reload.
5531
5532 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5533
5534 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5535 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5536 -mcall-openbsd, or -mcall-linux.
5537 (CC1_ENDIAN_BIG_SPEC): Remove.
5538 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5539 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5540 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5541 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5542 and %cc1_endian_default.
5543 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5544
5545 2014-02-10 Richard Biener <rguenther@suse.de>
5546
5547 PR tree-optimization/60115
5548 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5549 MEM_REF handling. Properly verify that the accesses are not
5550 out of the objects bound.
5551
5552 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5553
5554 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5555 coretex to cortex.
5556
5557 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5558
5559 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5560 proper constants and fix formatting.
5561 (possible_polymorphic_call_targets): Fix formatting.
5562
5563 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5564 Ilya Tocar <ilya.tocar@intel.com>
5565
5566 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5567 (_mm512_loadu_epi32): Renamed into...
5568 (_mm512_loadu_si512): This.
5569 (_mm512_storeu_epi32): Renamed into...
5570 (_mm512_storeu_si512): This.
5571 (_mm512_maskz_ceil_ps): Removed.
5572 (_mm512_maskz_ceil_pd): Ditto.
5573 (_mm512_maskz_floor_ps): Ditto.
5574 (_mm512_maskz_floor_pd): Ditto.
5575 (_mm512_floor_round_ps): Ditto.
5576 (_mm512_floor_round_pd): Ditto.
5577 (_mm512_ceil_round_ps): Ditto.
5578 (_mm512_ceil_round_pd): Ditto.
5579 (_mm512_mask_floor_round_ps): Ditto.
5580 (_mm512_mask_floor_round_pd): Ditto.
5581 (_mm512_mask_ceil_round_ps): Ditto.
5582 (_mm512_mask_ceil_round_pd): Ditto.
5583 (_mm512_maskz_floor_round_ps): Ditto.
5584 (_mm512_maskz_floor_round_pd): Ditto.
5585 (_mm512_maskz_ceil_round_ps): Ditto.
5586 (_mm512_maskz_ceil_round_pd): Ditto.
5587 (_mm512_expand_pd): Ditto.
5588 (_mm512_expand_ps): Ditto.
5589 * config/i386/i386.c (ix86_builtins): Remove
5590 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5591 (bdesc_args): Ditto.
5592 * config/i386/predicates.md (const1256_operand): New.
5593 (const_1_to_2_operand): Ditto.
5594 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5595 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5596 (*avx512pf_gatherpf<mode>sf): Ditto.
5597 (avx512pf_gatherpf<mode>df): Ditto.
5598 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5599 (*avx512pf_gatherpf<mode>df): Ditto.
5600 (avx512pf_scatterpf<mode>sf): Ditto.
5601 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5602 (*avx512pf_scatterpf<mode>sf): Ditto.
5603 (avx512pf_scatterpf<mode>df): Ditto.
5604 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5605 (*avx512pf_scatterpf<mode>df): Ditto.
5606 (avx512f_expand<mode>): Removed.
5607 (<shift_insn><mode>3<mask_name>): Change predicate type.
5608
5609 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5610
5611 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5612 not at the end of datarefs vector use ordered_remove to avoid
5613 reordering datarefs vector.
5614
5615 PR c/59984
5616 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5617 mark local addressable non-static vars as GOVD_PRIVATE
5618 instead of GOVD_LOCAL.
5619 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5620 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5621 than copying them.
5622
5623 PR middle-end/60092
5624 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5625 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5626 assume_aligned or alloc_align attributes.
5627 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5628 arguments. Handle also assume_aligned and alloc_align attributes.
5629 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5630 calls to functions with assume_aligned or alloc_align attributes.
5631 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5632
5633 2014-02-08 Terry Guo <terry.guo@arm.com>
5634
5635 * doc/invoke.texi: Document ARM -march=armv7e-m.
5636
5637 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5638
5639 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5640 flag on __cilkrts_rethrow builtin.
5641
5642 PR ipa/60026
5643 * ipa-cp.c (determine_versionability): Fail at -O0
5644 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5645 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5646
5647 Revert:
5648 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5649
5650 PR ipa/60026
5651 * tree-inline.c (copy_forbidden): Fail for
5652 __attribute__((optimize (0))) functions.
5653
5654 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5655
5656 * varpool.c: Include pointer-set.h.
5657 (varpool_remove_unreferenced_decls): Variables in other partitions
5658 will not be output; be however careful to not lose information
5659 about partitioning.
5660
5661 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5662
5663 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5664 lookup in the vtable constructor.
5665
5666 2014-02-07 Jeff Law <law@redhat.com>
5667
5668 PR target/40977
5669 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5670 define_insn_and_split.
5671
5672 * ipa-inline.c (inline_small_functions): Fix typos.
5673
5674 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5675
5676 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5677 (s390_can_use_return_insn): Declare.
5678 * config/s390/s390.h (EPILOGUE_USES): Define.
5679 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5680 instructions.
5681 (s390_chunkify_start): Handle return JUMP_LABELs.
5682 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5683 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5684 (s390_can_use_return_insn): New functions.
5685 (s390_fix_long_loop_prediction): Handle conditional returns.
5686 (TARGET_SET_UP_BY_PROLOGUE): Define.
5687 * config/s390/s390.md (ANY_RETURN): New code iterator.
5688 (*creturn, *csimple_return, return, simple_return): New patterns.
5689
5690 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5691
5692 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5693 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5694 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5695 REG_CFA_RESTORE list when deciding not to restore a register.
5696
5697 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5698
5699 * config/s390/s390.c: Include tree-pass.h and context.h.
5700 (s390_early_mach): New function, split out from...
5701 (s390_emit_prologue): ...here.
5702 (pass_data_s390_early_mach): New pass structure.
5703 (pass_s390_early_mach): New class.
5704 (s390_option_override): Create and register early_mach pass.
5705 Move to end of file.
5706
5707 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5708
5709 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5710 to match for the exit block.
5711
5712 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5713
5714 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5715 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5716 Reject misaligned operands.
5717
5718 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5719
5720 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5721
5722 2014-02-07 Richard Biener <rguenther@suse.de>
5723
5724 PR middle-end/60092
5725 * gimple-low.c (lower_builtin_posix_memalign): New function.
5726 (lower_stmt): Call it to lower posix_memalign in a way
5727 to make alignment info accessible.
5728
5729 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5730
5731 PR c++/60082
5732 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5733 __builtin_setjmp_receiver.
5734
5735 2014-02-07 Richard Biener <rguenther@suse.de>
5736
5737 PR middle-end/60092
5738 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5739 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5740 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5741 Handle BUILT_IN_POSIX_MEMALIGN.
5742 (find_func_clobbers): Likewise.
5743 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5744 (call_may_clobber_ref_p_1): Likewise.
5745
5746 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5747
5748 PR ipa/59918
5749 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5750 sanity check.
5751
5752 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5753
5754 PR ipa/59469
5755 * lto-cgraph.c (lto_output_node): Use
5756 symtab_get_symbol_partitioning_class.
5757 (lto_output_varpool_node): likewise.
5758 (symtab_get_symbol_partitioning_class): Move here from
5759 lto/lto-partition.c
5760 * cgraph.h (symbol_partitioning_class): Likewise.
5761 (symtab_get_symbol_partitioning_class): Declare.
5762
5763 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5764
5765 * ggc.h (ggc_internal_cleared_alloc): New macro.
5766 * vec.h (vec_safe_copy): Handle memory stats.
5767 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5768 * target-globals.c (save_target_globals): Likewise.
5769
5770 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5771
5772 PR target/60077
5773 * expr.c (emit_move_resolve_push): Export; be bit more selective
5774 on when to clear alias set.
5775 * expr.h (emit_move_resolve_push): Declare.
5776 * function.h (struct function): Add tail_call_marked.
5777 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5778 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5779 * config/i386/i386.md (TImode move expander): De not call
5780 ix86_expand_push.
5781 (FP push expanders): Preserve memory attributes.
5782 * config/i386/sse.md (push<mode>1): Remove.
5783 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5784 (ix86_expand_push): Remove.
5785 * config/i386/mmx.md (push<mode>1): Remove.
5786
5787 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5788
5789 PR rtl-optimization/60030
5790 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5791 lopart with paradoxical subreg before shifting it up by hprec.
5792
5793 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5794
5795 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5796 Remove extra newline at end of file.
5797 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5798 (arm_issue_rate): Handle cortexa57.
5799 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5800 (cortex-a57.cortex-a53): Likewise.
5801
5802 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5803
5804 PR target/59575
5805 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5806 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5807 bitmask.
5808 (arm_expand_prologue): Adjust all callers.
5809 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5810 info, registers also at the lowest numbered registers side. Use
5811 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5812 XEXP.
5813
5814 PR debug/59992
5815 * var-tracking.c (adjust_mems): Before adding a SET to
5816 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5817
5818 2014-02-06 Alan Modra <amodra@gmail.com>
5819
5820 PR target/60032
5821 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5822 change SDmode to DDmode when lra_in_progress.
5823
5824 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5825
5826 PR middle-end/59150
5827 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5828 free_data_ref on the dr first, and before goto again also set dr
5829 to the next dr. For simd_lane_access, free old datarefs[i] before
5830 overwriting it. For get_vectype_for_scalar_type failure, don't
5831 free_data_ref if simd_lane_access.
5832
5833 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5834
5835 PR target/60062
5836 * tree.h (opts_for_fn): New inline function.
5837 (opt_for_fn): Define.
5838 * config/i386/i386.c (ix86_function_regparm): Use
5839 opt_for_fn (decl, optimize) instead of optimize.
5840
5841 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5842
5843 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5844 for SYMBOL_REF in large memory model.
5845
5846 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5847
5848 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5849 and crypto support.
5850 (cortex-a57): Likewise.
5851 (cortex-a57.cortex-a53): Likewise.
5852
5853 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5854 Kugan Vivekanandarajah <kuganv@linaro.org>
5855
5856 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5857 unaligned_access.
5858 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5859
5860 2014-02-06 Richard Biener <rguenther@suse.de>
5861
5862 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5863 set_loop_copy and initialize_original_copy_tables.
5864
5865 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5866
5867 * config/aarch64/aarch64-simd.md
5868 (aarch64_ashr_simddi): Change QI to SI.
5869
5870 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5871 Jakub Jelinek <jakub@redhat.com>
5872
5873 PR middle-end/60013
5874 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5875 of the dataflow.
5876
5877 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5878
5879 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5880 CODE_FOR_altivec_vpku[hw]um to
5881 CODE_FOR_altivec_vpku[hw]um_direct.
5882 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5883 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5884 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5885 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5886
5887 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5888
5889 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5890 generation for -maltivec=be.
5891 (altivec_vsumsws): Simplify redundant test.
5892
5893 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5894
5895 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5896 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5897 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5898 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5899 gen_altivec_vpkuwum.
5900 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5901 BYTES_BIG_ENDIAN.
5902 (altivec_vpks<VI_char>ss): Likewise.
5903 (altivec_vpks<VI_char>us): Likewise.
5904 (altivec_vpku<VI_char>us): Likewise.
5905 (altivec_vpku<VI_char>um): Likewise.
5906 (altivec_vpku<VI_char>um_direct): New (copy of
5907 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5908 internal use).
5909 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5910 target is little endian and -maltivec=be is not specified.
5911 (*altivec_vupkhs<VU_char>_direct): New (copy of
5912 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5913 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5914 target is little endian and -maltivec=be is not specified.
5915 (*altivec_vupkls<VU_char>_direct): New (copy of
5916 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5917 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5918 little endian and -maltivec=be is not specified.
5919 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5920 little endian and -maltivec=be is not specified.
5921
5922 2014-02-05 Richard Henderson <rth@redhat.com>
5923
5924 PR debug/52727
5925 * combine-stack-adj.c: Revert r206943.
5926 * sched-int.h (struct deps_desc): Add last_args_size.
5927 * sched-deps.c (init_deps): Initialize it.
5928 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5929 contain REG_ARGS_SIZE notes.
5930
5931 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5932
5933 * lto-cgraph.c (asm_nodes_output): Make global.
5934 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5935 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5936 (driver_handle_option): Handle OPT_fwpa.
5937
5938 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5939
5940 PR ipa/59947
5941 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5942 a comment typo and formatting issue. If odr_hash hasn't been
5943 created, return vNULL and set *completep to false.
5944
5945 PR middle-end/57499
5946 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5947 bb with no successors.
5948
5949 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5950
5951 PR target/59718
5952 * doc/invoke.texi (-march): Clarify documentation for ARM.
5953 (-mtune): Likewise.
5954 (-mcpu): Likewise.
5955
5956 2014-02-05 Richard Biener <rguenther@suse.de>
5957
5958 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5959 when not vectorizing because of too many alias checks.
5960 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5961 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5962
5963 2014-02-05 Nick Clifton <nickc@redhat.com>
5964
5965 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
5966 accept extended registers in any mode when compiling for the MN10300.
5967
5968 2014-02-05 Yury Gribov <y.gribov@samsung.com>
5969
5970 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5971 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5972 sanitization attributes.
5973 (can_inline_edge_p): Likewise.
5974 (sanitize_attrs_match_for_inline_p): New function.
5975
5976 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5977
5978 * ipa-prop.c (detect_type_change): Shor circuit testing of
5979 type changes on THIS pointer.
5980
5981 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
5982
5983 PR target/59777
5984 * config/pa/pa.c (legitimize_tls_address): Return original address
5985 if not passed a SYMBOL_REF rtx.
5986 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
5987 addresses.
5988 (pa_emit_move_sequence): Simplify TLS source operands.
5989 (pa_legitimate_constant_p): Reject all TLS constants.
5990 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
5991 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
5992
5993 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
5994
5995 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
5996 groups when we know they are controlled by LTO.
5997 * varasm.c (default_binds_local_p_1): If object is in other partition,
5998 it will be resolved locally.
5999
6000 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6001
6002 * config/host-linux.c (linux_gt_pch_use_address): Don't
6003 use SSIZE_MAX because it is not always defined.
6004
6005 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6006
6007 PR bootstrap/59913
6008 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6009 threshold for pseudo splitting.
6010 (update_ebb_live_info): Process call argument hard registers and
6011 hard registers from insn definition too.
6012 (max_small_class_regs_num): New constant.
6013 (inherit_in_ebb): Update live hard regs through EBBs. Update
6014 reloads_num only for small register classes. Don't split for
6015 outputs of jumps.
6016
6017 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6018
6019 PR ipa/60058
6020 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6021 is non-null.
6022
6023 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6024
6025 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6026 visibility is safe.
6027
6028 2014-02-04 Marek Polacek <polacek@redhat.com>
6029
6030 * gdbinit.in (pel): Define.
6031
6032 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6033
6034 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6035 behavior.
6036
6037 2014-02-04 Richard Biener <rguenther@suse.de>
6038
6039 PR lto/59723
6040 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6041 in function context local.
6042 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6043 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6044 similar to LTO_imported_decl_ref.
6045
6046 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6047
6048 PR tree-optimization/60002
6049 * cgraphclones.c (build_function_decl_skip_args): Clear
6050 DECL_LANG_SPECIFIC.
6051
6052 PR tree-optimization/60023
6053 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6054 false to gsi_replace.
6055 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6056 has been in some EH region and vec_stmt could throw, add
6057 vec_stmt into the same EH region.
6058 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6059 has no lhs, ignore it.
6060 * internal-fn.c (expand_MASK_LOAD): Likewise.
6061
6062 PR ipa/60026
6063 * tree-inline.c (copy_forbidden): Fail for
6064 __attribute__((optimize (0))) functions.
6065
6066 PR other/58712
6067 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6068 is set, copy one less argument.
6069 (expand_simd_clones): Don't subtract clone_info->inbranch
6070 from simd_clone_struct_alloc argument.
6071
6072 PR rtl-optimization/57915
6073 * recog.c (simplify_while_replacing): If all unary/binary/relational
6074 operation arguments are constant, attempt to simplify those.
6075
6076 PR middle-end/59261
6077 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6078 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6079
6080 2014-02-04 Richard Biener <rguenther@suse.de>
6081
6082 PR tree-optimization/60012
6083 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6084 TBAA disambiguation to all DDRs.
6085
6086 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6087
6088 PR target/59788
6089 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6090 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6091
6092 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6093
6094 PR ipa/59882
6095 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6096
6097 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6098
6099 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6100 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6101
6102 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6103
6104 PR ipa/59831
6105 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6106 to figure out targets of polymorphic calls with known decl.
6107 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6108 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6109 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6110 (get_polymorphic_call_info): ... here.
6111 (get_polymorphic_call_info_from_invariant): New function.
6112
6113 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6114
6115 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6116 lookup via vtable pointer; check for type consistency
6117 and turn inconsitent facts into UNREACHABLE.
6118 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6119 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6120 type inconsistent querries; return UNREACHABLE instead.
6121
6122 2014-02-03 Richard Henderson <rth@twiddle.net>
6123
6124 PR tree-opt/59924
6125 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6126 already processed this node.
6127 (normalize_one_pred_1): Pass along mark_set.
6128 (normalize_one_pred): Create and destroy a pointer_set_t.
6129 (normalize_one_pred_chain): Likewise.
6130
6131 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6132
6133 PR gcov-profile/58602
6134 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6135
6136 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6137
6138 PR ipa/59831
6139 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6140 -fno-devirtualize; try to devirtualize by the knowledge of
6141 virtual table pointer given by aggregate propagation.
6142 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6143 (ipa_print_node_jump_functions): Dump also offset that
6144 is relevant for polymorphic calls.
6145 (determine_known_aggregate_parts): Add arg_type parameter; use it
6146 instead of determining the type from pointer type.
6147 (ipa_compute_jump_functions_for_edge): Update call of
6148 determine_known_aggregate_parts.
6149 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6150 (gimple_get_virt_method_for_binfo): ... here; simplify using
6151 vtable_pointer_value_to_vtable.
6152 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6153 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6154 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6155 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6156 POINTER_PLUS_EXPR.
6157 (vtable_pointer_value_to_binfo): ... here.
6158 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6159
6160 2014-02-03 Teresa Johnson <tejohnson@google.com>
6161
6162 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6163 redef of outer loop index variable.
6164
6165 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6166
6167 PR c++/53017
6168 PR c++/59211
6169 * doc/extend.texi (Function Attributes): Typo.
6170
6171 2014-02-03 Cong Hou <congh@google.com>
6172
6173 PR tree-optimization/60000
6174 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6175 if the vectorized statement is a store. A store statement can only
6176 appear at the end of pattern statements.
6177
6178 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6179
6180 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6181 (ix86_option_override_internal): Default long double to 64-bit for
6182 32-bit Bionic and to 128-bit for 64-bit Bionic.
6183
6184 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6185 TARGET_LONG_DOUBLE_128 is true.
6186 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6187
6188 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6189 (mlong-double-64): Negate -mlong-double-128.
6190 (mlong-double-128): New option.
6191
6192 * config/i386/i386-c.c (ix86_target_macros): Define
6193 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6194
6195 * doc/invoke.texi: Document -mlong-double-128.
6196
6197 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6198
6199 PR rtl-optimization/60024
6200 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6201
6202 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6203
6204 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6205
6206 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6207
6208 PR rtl-optimization/57662
6209 * sel-sched.c (code_motion_path_driver): Do not mark already not
6210 existing blocks in the visiting bitmap.
6211
6212 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6213
6214 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6215 on the insn being emitted.
6216
6217 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6218 Will Deacon <will.deacon@arm.com>
6219
6220 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6221
6222 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6223
6224 * config/arm/arm-tables.opt: Regenerate.
6225
6226 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6227
6228 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6229 for vector types other than V16QImode.
6230 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6231 define_expand, and call altivec_expand_vec_perm_le when producing
6232 code with little endian element order.
6233 (*altivec_vperm_<mode>_internal): New insn having previous
6234 behavior of altivec_vperm_<mode>.
6235 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6236 altivec_expand_vec_perm_le when producing code with little endian
6237 element order.
6238 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6239 behavior of altivec_vperm_<mode>_uns.
6240
6241 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6242
6243 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6244 (altivec_vsumsws): Add handling for -maltivec=be with a little
6245 endian target.
6246 (altivec_vsumsws_direct): New.
6247 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6248 gen_altivec_vsumsws.
6249
6250 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6251
6252 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6253 vtable_pointer_value_to_binfo): New functions.
6254 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6255 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6256
6257 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6258
6259 * config/nios2/nios2.md (load_got_register): Initialize GOT
6260 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6261 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6262
6263 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6264
6265 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6266 preserverd by passthrough, do not propagate the type.
6267
6268 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6269
6270 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6271 (mips_atomic_assign_expand_fenv): New function.
6272 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6273
6274 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6275
6276 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6277 (__builtin_mips_set_fcsr): Likewise.
6278 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6279 MIPS_USI_FTYPE_VOID.
6280 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6281 (mips16_expand_set_fcsr): Likewise.
6282 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6283 (mips16_set_fcsr_stub): Likewise.
6284 (mips16_get_fcsr_one_only_stub): New class.
6285 (mips16_set_fcsr_one_only_stub): Likewise.
6286 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6287 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6288 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6289 (hard_float): New availability predicate.
6290 (mips_builtins): Add get_fcsr and set_fcsr.
6291 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6292 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6293 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6294 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6295 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6296 patterns.
6297
6298 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6299
6300 * config/mips/mips.c (mips_one_only_stub): New class.
6301 (mips_need_mips16_rdhwr_p): Replace with...
6302 (mips16_rdhwr_stub): ...this new variable.
6303 (mips16_stub_call_address): New function.
6304 (mips16_rdhwr_one_only_stub): New class.
6305 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6306 (mips_output_mips16_rdhwr): Delete.
6307 (mips_finish_stub): New function.
6308 (mips_code_end): Use it to handle rdhwr stubs.
6309
6310 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6311
6312 PR target/60017
6313 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6314 when calculating size of integer atomic types.
6315
6316 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6317
6318 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6319
6320 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6321
6322 PR tree-optimization/60003
6323 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6324 * profile.c (branch_prob): Use gimple_call_builtin_p
6325 to check for BUILT_IN_SETJMP_RECEIVER.
6326 * tree-inline.c (copy_bb): Call notice_special_calls.
6327
6328 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6329
6330 PR bootstrap/59985
6331 * lra-constraints.c (process_alt_operands): Update reload_sum only
6332 on the first pass.
6333
6334 2014-01-31 Richard Henderson <rth@redhat.com>
6335
6336 PR middle-end/60004
6337 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6338 until after else_eh is processed.
6339
6340 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6341
6342 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6343 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6344 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6345 in smmintrin.h, remove them.
6346 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6347 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6348 * config/i386/i386.md (ROUND_SAE): Fix value.
6349 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6350 (const48_operand): New.
6351 * config/i386/subst.md (round), (round_expand): Use
6352 const_4_or_8_to_11_operand.
6353 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6354
6355 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6356
6357 * config/i386/constraints.md (Yk): Swap meaning with k.
6358 * config/i386/i386.md (movhi_internal): Change Yk to k.
6359 (movqi_internal): Ditto.
6360 (*k<logic><mode>): Ditto.
6361 (*andhi_1): Ditto.
6362 (*andqi_1): Ditto.
6363 (kandn<mode>): Ditto.
6364 (*<code>hi_1): Ditto.
6365 (*<code>qi_1): Ditto.
6366 (kxnor<mode>): Ditto.
6367 (kortestzhi): Ditto.
6368 (kortestchi): Ditto.
6369 (kunpckhi): Ditto.
6370 (*one_cmplhi2_1): Ditto.
6371 (*one_cmplqi2_1): Ditto.
6372 * config/i386/sse.md (): Change k to Yk.
6373 (avx512f_load<mode>_mask): Ditto.
6374 (avx512f_blendm<mode>): Ditto.
6375 (avx512f_store<mode>_mask): Ditto.
6376 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6377 (avx512f_storedqu<mode>_mask): Ditto.
6378 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6379 Ditto.
6380 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6381 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6382 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6383 (avx512f_maskcmp<mode>3): Ditto.
6384 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6385 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6386 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6387 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6388 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6389 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6390 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6391 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6392 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6393 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6394 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6395 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6396 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6397 (vec_extract_lo_<mode>_maskm): Ditto.
6398 (vec_extract_hi_<mode>_maskm): Ditto.
6399 (avx512f_vternlog<mode>_mask): Ditto.
6400 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6401 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6402 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6403 (avx512f_<code>v8div16qi2_mask): Ditto.
6404 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6405 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6406 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6407 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6408 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6409 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6410 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6411 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6412 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6413 (avx512cd_maskb_vec_dupv8di): Ditto.
6414 (avx512cd_maskw_vec_dupv16si): Ditto.
6415 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6416 (avx512f_vpermi2var<mode>3_mask): Ditto.
6417 (avx512f_vpermi2var<mode>3_mask): Ditto.
6418 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6419 (*avx512f_gathersi<mode>): Ditto.
6420 (*avx512f_gathersi<mode>_2): Ditto.
6421 (*avx512f_gatherdi<mode>): Ditto.
6422 (*avx512f_gatherdi<mode>_2): Ditto.
6423 (*avx512f_scattersi<mode>): Ditto.
6424 (*avx512f_scatterdi<mode>): Ditto.
6425 (avx512f_compress<mode>_mask): Ditto.
6426 (avx512f_compressstore<mode>_mask): Ditto.
6427 (avx512f_expand<mode>_mask): Ditto.
6428 * config/i386/subst.md (mask): Change k to Yk.
6429 (mask_scalar_merge): Ditto.
6430 (sd): Ditto.
6431
6432 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6433
6434 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6435
6436 2014-01-31 Richard Biener <rguenther@suse.de>
6437
6438 PR middle-end/59990
6439 * builtins.c (fold_builtin_memory_op): Make sure to not
6440 use a floating-point mode or a boolean or enumeral type for
6441 the copy operation.
6442
6443 2014-01-30 DJ Delorie <dj@redhat.com>
6444
6445 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6446 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6447 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6448 whenever main() has an epilogue.
6449
6450 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6451
6452 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6453 unused variable "field".
6454 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6455 (vsx_mergeh_<mode>): Likewise.
6456 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6457 (altivec_vmrghh): Likewise.
6458 (altivec_vmrghw): Likewise.
6459 (altivec_vmrglb): Likewise.
6460 (altivec_vmrglh): Likewise.
6461 (altivec_vmrglw): Likewise.
6462 (altivec_vspltb): Add missing uses.
6463 (altivec_vsplth): Likewise.
6464 (altivec_vspltw): Likewise.
6465 (altivec_vspltsf): Likewise.
6466
6467 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6468
6469 PR target/59923
6470 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6471 frame related instructions.
6472
6473 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6474
6475 PR rtl-optimization/59959
6476 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6477 any reload of register whose subreg is invalid.
6478
6479 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6480
6481 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6482 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6483 Add missing return type - void.
6484
6485 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6486
6487 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6488 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6489 remove element index adjustment for endian (now handled in vsx.md
6490 and altivec.md).
6491 (altivec_expand_vec_perm_const): Use
6492 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6493 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6494 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6495 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6496 define_expand and a new define_insn *altivec_vspltb_internal;
6497 adjust for -maltivec=be on a little endian target.
6498 (altivec_vspltb_direct): New.
6499 (altivec_vsplth): Divide into a define_expand and a new
6500 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6501 little endian target.
6502 (altivec_vsplth_direct): New.
6503 (altivec_vspltw): Divide into a define_expand and a new
6504 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6505 little endian target.
6506 (altivec_vspltw_direct): New.
6507 (altivec_vspltsf): Divide into a define_expand and a new
6508 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6509 a little endian target.
6510
6511 2014-01-30 Richard Biener <rguenther@suse.de>
6512
6513 PR tree-optimization/59993
6514 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6515 can propagate form the earlier stmt and avoid the transform
6516 when the intermediate result is needed.
6517
6518 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6519
6520 * README.Portability: Fix typo.
6521
6522 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6523
6524 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6525 comparison_operator with ordered_comparison_operator.
6526
6527 2014-01-30 Nick Clifton <nickc@redhat.com>
6528
6529 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6530 Rename to mn10300_store_multiple_regs.
6531 * config/mn10300/mn10300.c: Likewise.
6532 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6533 store_multiple_regs.
6534 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6535 Call mn10300_store_multiple_regs.
6536
6537 2014-01-30 Nick Clifton <nickc@redhat.com>
6538 DJ Delorie <dj@redhat.com>
6539
6540 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6541 %fp 2 to keep registers after it properly word-aligned.
6542 (rl78_alloc_physical_registers_umul): Handle the case where both
6543 input operands are the same.
6544
6545 2014-01-30 Richard Biener <rguenther@suse.de>
6546
6547 PR tree-optimization/59903
6548 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6549 check properly.
6550
6551 2014-01-30 Jason Merrill <jason@redhat.com>
6552
6553 PR c++/59633
6554 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6555
6556 PR c++/59645
6557 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6558
6559 2014-01-30 Richard Biener <rguenther@suse.de>
6560
6561 PR tree-optimization/59951
6562 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6563
6564 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6565
6566 PR target/59784
6567 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6568 SFmode to DFmode case.
6569
6570 2014-01-29 DJ Delorie <dj@redhat.com>
6571
6572 * config/msp430/msp430.opt (-minrt): New.
6573 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6574 if -minrt given.
6575 (ENDFILE_SPEC): Likewise.
6576
6577 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6578
6579 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6580 (estimate_function_body_sizes): Use it.
6581
6582 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6583
6584 PR c++/58561
6585 * dwarf2out.c (is_cxx_auto): New.
6586 (is_base_type): Use it.
6587 (gen_type_die_with_usage): Likewise.
6588
6589 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6590
6591 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6592 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6593 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6594 -maltivec=be with LE targets.
6595 (vsx_mergeh_<mode>): Likewise.
6596 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6597 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6598 (altivec_vmrghb): Replace with define_expand and new
6599 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6600 (altivec_vmrghb_direct): New define_insn.
6601 (altivec_vmrghh): Replace with define_expand and new
6602 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6603 (altivec_vmrghh_direct): New define_insn.
6604 (altivec_vmrghw): Replace with define_expand and new
6605 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6606 (altivec_vmrghw_direct): New define_insn.
6607 (*altivec_vmrghsf): Adjust for endianness.
6608 (altivec_vmrglb): Replace with define_expand and new
6609 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6610 (altivec_vmrglb_direct): New define_insn.
6611 (altivec_vmrglh): Replace with define_expand and new
6612 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6613 (altivec_vmrglh_direct): New define_insn.
6614 (altivec_vmrglw): Replace with define_expand and new
6615 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6616 (altivec_vmrglw_direct): New define_insn.
6617 (*altivec_vmrglsf): Adjust for endianness.
6618 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6619 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6620 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6621 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6622 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6623 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6624 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6625 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6626
6627 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6628
6629 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6630 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6631 whitespace.
6632
6633 2014-01-29 Richard Biener <rguenther@suse.de>
6634
6635 PR tree-optimization/58742
6636 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6637 associate_pointerplus_align.
6638 (associate_pointerplus_diff): New function.
6639 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6640 and associate_pointerplus_diff.
6641
6642 2014-01-29 Richard Biener <rguenther@suse.de>
6643
6644 * lto-streamer.h (LTO_major_version): Bump to 3.
6645 (LTO_minor_version): Reset to 0.
6646
6647 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6648
6649 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6650 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6651 (arm_file_start): Generate correct asm header for armv7ve.
6652 * config/arm/bpabi.h: Add multilib support for armv7ve.
6653 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6654 and cortex-a15 to armv7ve.
6655 * config/arm/t-aprofile: Add multilib support for armv7ve.
6656 * doc/invoke.texi: Document -march=armv7ve.
6657
6658 2014-01-29 Richard Biener <rguenther@suse.de>
6659
6660 PR tree-optimization/58742
6661 * tree-ssa-forwprop.c (associate_plusminus): Return true
6662 if we changed sth, defer EH cleanup to ...
6663 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6664 (simplify_mult): New function.
6665
6666 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6667
6668 PR middle-end/59917
6669 PR tree-optimization/59920
6670 * tree.c (build_common_builtin_nodes): Remove
6671 __builtin_setjmp_dispatcher initialization.
6672 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6673 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6674 instead of gsi_after_labels + manually skipping debug stmts.
6675 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6676 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6677 * tree-inline.c (copy_edges_for_bb): Remove
6678 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6679 argument. Ignore computed_goto_p stmts. Don't call
6680 make_abnormal_goto_edges. If a call might need abnormal edges
6681 for non-local gotos, see if it already has an edge to
6682 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6683 with true argument, don't do anything then, otherwise add
6684 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6685 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6686 caller.
6687 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6688 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6689 (lower_stmt): Don't set data->calls_builtin_setjmp.
6690 (lower_builtin_setjmp): Adjust comment.
6691 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6692 * tree-cfg.c (found_computed_goto): Remove.
6693 (factor_computed_gotos): Remove.
6694 (make_goto_expr_edges): Return bool, true for computed gotos.
6695 Don't call make_abnormal_goto_edges.
6696 (build_gimple_cfg): Don't set found_computed_goto, don't call
6697 factor_computed_gotos.
6698 (computed_goto_p): No longer static.
6699 (make_blocks): Don't set found_computed_goto.
6700 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6701 (make_edges): If make_goto_expr_edges returns true, push bb
6702 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6703 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6704 vector. Record mapping between bbs and OpenMP regions if there
6705 are any, adjust make_gimple_omp_edges caller. Call
6706 handle_abnormal_edges.
6707 (make_abnormal_goto_edges): Remove.
6708 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6709 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6710 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6711 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6712 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6713 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6714 filling *region also set *region_idx to (*region)->entry->index.
6715
6716 PR other/58712
6717 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6718 For REGs set ORIGINAL_REGNO.
6719
6720 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6721
6722 * doc/md.texi: Mention that a target shouldn't implement
6723 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6724 than hi/lo pair.
6725
6726 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6727
6728 PR tree-optimization/59594
6729 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6730 a copy of the datarefs vector rather than the vector itself.
6731
6732 2014-01-28 Jason Merrill <jason@redhat.com>
6733
6734 PR c++/53756
6735 * dwarf2out.c (auto_die): New static.
6736 (gen_type_die_with_usage): Handle C++1y 'auto'.
6737 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6738 on definition.
6739
6740 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6741
6742 PR target/59672
6743 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6744 (SPEC_X32): Likewise.
6745 (SPEC_64): Likewise.
6746 * config/i386/i386.c (ix86_option_override_internal): Turn off
6747 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6748 for TARGET_16BIT.
6749 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6750 * config/i386/i386.h (TARGET_16BIT): New macro.
6751 (TARGET_16BIT_P): Likewise.
6752 * config/i386/i386.opt: Add m16.
6753 * doc/invoke.texi: Document -m16.
6754
6755 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6756
6757 PR preprocessor/59935
6758 * input.c (location_get_source_line): Bail out on when line number
6759 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6760
6761 2014-01-28 Richard Biener <rguenther@suse.de>
6762
6763 PR tree-optimization/58742
6764 * tree-ssa-forwprop.c (associate_plusminus): Handle
6765 pointer subtraction of the form (T)(P + A) - (T)P.
6766
6767 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6768
6769 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6770 at const_int_cost.
6771
6772 2014-01-28 Richard Biener <rguenther@suse.de>
6773
6774 Revert
6775 2014-01-28 Richard Biener <rguenther@suse.de>
6776
6777 PR rtl-optimization/45364
6778 PR rtl-optimization/59890
6779 * var-tracking.c (local_get_addr_clear_given_value): Handle
6780 already cleared slot.
6781 (val_reset): Handle not allocated local_get_addr_cache.
6782 (vt_find_locations): Use post-order on the inverted CFG.
6783
6784 2014-01-28 Richard Biener <rguenther@suse.de>
6785
6786 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6787
6788 2014-01-28 Richard Biener <rguenther@suse.de>
6789
6790 PR rtl-optimization/45364
6791 PR rtl-optimization/59890
6792 * var-tracking.c (local_get_addr_clear_given_value): Handle
6793 already cleared slot.
6794 (val_reset): Handle not allocated local_get_addr_cache.
6795 (vt_find_locations): Use post-order on the inverted CFG.
6796
6797 2014-01-28 Alan Modra <amodra@gmail.com>
6798
6799 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6800 * configure.ac <recursive call for build != host>: Define
6801 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6802 and LD_FOR_BUILD too.
6803 * configure: Regenerate.
6804
6805 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6806
6807 * config/i386/i386.c (get_builtin_code_for_version): Separate
6808 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6809 Broadwell from Haswell.
6810
6811 2014-01-27 Steve Ellcey <sellcey@mips.com>
6812
6813 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6814 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6815 * config/mips/mips.c (mips_option_override): Change setting
6816 of TARGET_DSP.
6817 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6818 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6819 Change from Mask to Var.
6820
6821 2014-01-27 Jeff Law <law@redhat.com>
6822
6823 * ipa-inline.c (inline_small_functions): Fix typo.
6824
6825 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6826
6827 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6828 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6829 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6830 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6831 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6832 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6833 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6834 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6835 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6836 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6837 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6838 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6839 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6840 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6841 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6842 (_mm512_storeu_epi64): Ditto.
6843 (_mm512_cmpge_epi32_mask): Ditto.
6844 (_mm512_cmpge_epu32_mask): Ditto.
6845 (_mm512_cmpge_epi64_mask): Ditto.
6846 (_mm512_cmpge_epu64_mask): Ditto.
6847 (_mm512_cmple_epi32_mask): Ditto.
6848 (_mm512_cmple_epu32_mask): Ditto.
6849 (_mm512_cmple_epi64_mask): Ditto.
6850 (_mm512_cmple_epu64_mask): Ditto.
6851 (_mm512_cmplt_epi32_mask): Ditto.
6852 (_mm512_cmplt_epu32_mask): Ditto.
6853 (_mm512_cmplt_epi64_mask): Ditto.
6854 (_mm512_cmplt_epu64_mask): Ditto.
6855 (_mm512_cmpneq_epi32_mask): Ditto.
6856 (_mm512_cmpneq_epu32_mask): Ditto.
6857 (_mm512_cmpneq_epi64_mask): Ditto.
6858 (_mm512_cmpneq_epu64_mask): Ditto.
6859 (_mm512_expand_pd): Ditto.
6860 (_mm512_expand_ps): Ditto.
6861 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6862 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6863 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6864 * config/i386/i386.c (ix86_builtins): Add
6865 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6866 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6867 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6868 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6869 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6870 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6871 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6872 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6873 IX86_BUILTIN_PMOVUSQW512_MEM.
6874 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6875 __builtin_ia32_pmovsqd512mem_mask,
6876 __builtin_ia32_pmovqd512mem_mask,
6877 __builtin_ia32_pmovusqw512mem_mask,
6878 __builtin_ia32_pmovsqw512mem_mask,
6879 __builtin_ia32_pmovqw512mem_mask,
6880 __builtin_ia32_pmovusdw512mem_mask,
6881 __builtin_ia32_pmovsdw512mem_mask,
6882 __builtin_ia32_pmovdw512mem_mask,
6883 __builtin_ia32_pmovqb512mem_mask,
6884 __builtin_ia32_pmovusqb512mem_mask,
6885 __builtin_ia32_pmovsqb512mem_mask,
6886 __builtin_ia32_pmovusdb512mem_mask,
6887 __builtin_ia32_pmovsdb512mem_mask,
6888 __builtin_ia32_pmovdb512mem_mask.
6889 (bdesc_args): Add __builtin_ia32_expanddf512,
6890 __builtin_ia32_expandsf512.
6891 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6892 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6893 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6894 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6895 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6896 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6897 (avx512f_<code>v8div16qi2_mask_store): This.
6898 (avx512f_expand<mode>): New.
6899
6900 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6901
6902 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6903 New.
6904 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6905 (_mm512_prefetch_i32scatter_pd): Ditto.
6906 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6907 (_mm512_prefetch_i64scatter_pd): Ditto.
6908 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6909 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6910 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6911 (_mm512_prefetch_i32scatter_ps): Ditto.
6912 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6913 (_mm512_prefetch_i64scatter_ps): Ditto.
6914 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6915 * config/i386/i386-builtin-types.def: Define
6916 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6917 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6918 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6919 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6920 IX86_BUILTIN_SCATTERPFQPD.
6921 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6922 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6923 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6924 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6925 __builtin_ia32_scatterpfqps.
6926 (ix86_expand_builtin): Expand new built-ins.
6927 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6928 fix memory access data type.
6929 (*avx512pf_gatherpf<mode>_mask): Ditto.
6930 (*avx512pf_gatherpf<mode>): Ditto.
6931 (avx512pf_scatterpf<mode>): Ditto.
6932 (*avx512pf_scatterpf<mode>_mask): Ditto.
6933 (*avx512pf_scatterpf<mode>): Ditto.
6934 (GATHER_SCATTER_SF_MEM_MODE): New.
6935 (avx512pf_gatherpf<mode>df): Ditto.
6936 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6937 (*avx512pf_scatterpf<mode>df): Ditto.
6938
6939 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6940
6941 PR bootstrap/59934
6942 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6943 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6944 reached.
6945
6946 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6947
6948 * common/config/arm/arm-common.c
6949 (arm_rewrite_mcpu): Handle multiple names.
6950 * config/arm/arm.h
6951 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6952
6953 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6954
6955 * gimple-builder.h (create_gimple_tmp): Delete.
6956
6957 2014-01-27 Christian Bruel <christian.bruel@st.com>
6958
6959 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6960 words comparisons.
6961
6962 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
6963
6964 * config/pa/pa.md (call): Generate indirect long calls to non-local
6965 functions when outputing 32-bit code.
6966 (call_value): Likewise except for special call to buggy powf function.
6967
6968 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6969 portable runtime and PIC indirect calls.
6970 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6971 and PIC call sequences. Use ldo instead of blr to set return register
6972 in PIC call sequence.
6973
6974 2014-01-25 Walter Lee <walt@tilera.com>
6975
6976 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6977 avoid clobbering a live register.
6978
6979 2014-01-25 Walter Lee <walt@tilera.com>
6980
6981 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
6982 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
6983 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
6984 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
6985
6986 2014-01-25 Walter Lee <walt@tilera.com>
6987
6988 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
6989 arguments on even registers.
6990 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
6991 STACK_BOUNDARY.
6992 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
6993 (BIGGEST_ALIGNMENT): Ditto.
6994 (BIGGEST_FIELD_ALIGNMENT): Ditto.
6995
6996 2014-01-25 Walter Lee <walt@tilera.com>
6997
6998 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
6999 insns before bundling.
7000 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7001
7002 2014-01-25 Walter Lee <walt@tilera.com>
7003
7004 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7005 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7006 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7007
7008 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7009
7010 * config/mips/constraints.md (kl): Delete.
7011 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7012 define expands, using...
7013 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7014 instructions for MIPS16.
7015 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7016 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7017
7018 2014-01-25 Walter Lee <walt@tilera.com>
7019
7020 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7021 (clzdi2): Ditto.
7022 (ffsdi2): Ditto.
7023
7024 2014-01-25 Walter Lee <walt@tilera.com>
7025
7026 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7027 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7028
7029 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7030
7031 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7032 Handle XOR.
7033
7034 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7035
7036 * print-rtl.c (in_call_function_usage): New var.
7037 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7038 EXPR_LIST mode as mode and not as reg note name.
7039
7040 PR middle-end/59561
7041 * cfgloopmanip.c (copy_loop_info): If
7042 loop->warned_aggressive_loop_optimizations, make sure
7043 the flag is set in target loop too.
7044
7045 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7046
7047 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7048 flag_cilkplus.
7049 * builtins.def: Likewise.
7050 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7051 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7052 * ira.c (ira_setup_eliminable_regset): Likewise.
7053 * omp-low.c (gate_expand_omp): Likewise.
7054 (execute_lower_omp): Likewise.
7055 (diagnose_sb_0): Likewise.
7056 (gate_diagnose_omp_blocks): Likewise.
7057 (simd_clone_clauses_extract): Likewise.
7058 (gate): Likewise.
7059
7060 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7061
7062 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7063 correction for little endian...
7064 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7065 here.
7066
7067 2014-01-24 Jeff Law <law@redhat.com>
7068
7069 PR tree-optimization/59919
7070 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7071 for non-returning calls.
7072
7073 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7074
7075 * common/config/aarch64/aarch64-common.c
7076 (aarch64_rewrite_mcpu): Handle multiple names.
7077 * config/aarch64/aarch64.h
7078 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7079
7080 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7081
7082 * input.c (add_file_to_cache_tab): Handle the case where fopen
7083 returns NULL.
7084
7085 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7086
7087 PR target/59929
7088 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7089 from push operand if code of push isn't PRE_DEC.
7090
7091 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7092
7093 PR target/59909
7094 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7095 -mquad-memory-atomic. Update -mquad-memory documentation to say
7096 it is only used for non-atomic loads/stores.
7097
7098 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7099 -mquad-memory or -mquad-memory-atomic switches.
7100
7101 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7102 -mquad-memory-atomic to ISA 2.07 support.
7103
7104 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7105 to separate support of normal quad word memory operations (ldq, stq)
7106 from the atomic quad word memory operations.
7107
7108 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7109 support to separate non-atomic quad word operations from atomic
7110 quad word operations. Disable non-atomic quad word operations in
7111 little endian mode so that we don't have to swap words after the
7112 load and before the store.
7113 (quad_load_store_p): Add comment about atomic quad word support.
7114 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7115 options printed with -mdebug=reg.
7116
7117 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7118 -mquad-memory-atomic as the test for whether we have quad word
7119 atomic instructions.
7120 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7121 or -mp8-vector are used, allow byte/half-word atomic operations.
7122
7123 * config/rs6000/sync.md (load_lockedti): Insure that the address
7124 is a proper indexed or indirect address for the lqarx instruction.
7125 On little endian systems, swap the hi/lo registers after the lqarx
7126 instruction.
7127 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7128 insure the address is valid for the lqarx instruction.
7129 (store_conditionalti): Insure that the address is a proper indexed
7130 or indirect address for the stqcrx. instruction. On little endian
7131 systems, swap the hi/lo registers before doing the stqcrx.
7132 instruction.
7133 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7134 insure the address is valid for the stqcrx. instruction.
7135
7136 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7137 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7138 type of quad memory support is available.
7139
7140 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7141
7142 PR regression/59915
7143 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7144 there is a danger of looping.
7145
7146 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7147
7148 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7149 force flag_ira_loop_pressure if set via command line.
7150
7151 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7152
7153 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7154 (ashr_simd): New builtin handling DI mode.
7155 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7156 (aarch64_sshr_simddi): New match pattern.
7157 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7158 (vshrd_n_s64): Likewise.
7159 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7160
7161 2014-01-23 Nick Clifton <nickc@redhat.com>
7162
7163 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7164 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7165 favour of mcu specific scripts.
7166 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7167 430x multilibs.
7168
7169 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7170 Alex Velenko <Alex.Velenko@arm.com>
7171
7172 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7173 (vaddv_s16): Likewise.
7174 (vaddv_s32): Likewise.
7175 (vaddv_u8): Likewise.
7176 (vaddv_u16): Likewise.
7177 (vaddv_u32): Likewise.
7178 (vaddvq_s8): Likewise.
7179 (vaddvq_s16): Likewise.
7180 (vaddvq_s32): Likewise.
7181 (vaddvq_s64): Likewise.
7182 (vaddvq_u8): Likewise.
7183 (vaddvq_u16): Likewise.
7184 (vaddvq_u32): Likewise.
7185 (vaddvq_u64): Likewise.
7186 (vaddv_f32): Likewise.
7187 (vaddvq_f32): Likewise.
7188 (vaddvq_f64): Likewise.
7189 (vmaxv_f32): Likewise.
7190 (vmaxv_s8): Likewise.
7191 (vmaxv_s16): Likewise.
7192 (vmaxv_s32): Likewise.
7193 (vmaxv_u8): Likewise.
7194 (vmaxv_u16): Likewise.
7195 (vmaxv_u32): Likewise.
7196 (vmaxvq_f32): Likewise.
7197 (vmaxvq_f64): Likewise.
7198 (vmaxvq_s8): Likewise.
7199 (vmaxvq_s16): Likewise.
7200 (vmaxvq_s32): Likewise.
7201 (vmaxvq_u8): Likewise.
7202 (vmaxvq_u16): Likewise.
7203 (vmaxvq_u32): Likewise.
7204 (vmaxnmv_f32): Likewise.
7205 (vmaxnmvq_f32): Likewise.
7206 (vmaxnmvq_f64): Likewise.
7207 (vminv_f32): Likewise.
7208 (vminv_s8): Likewise.
7209 (vminv_s16): Likewise.
7210 (vminv_s32): Likewise.
7211 (vminv_u8): Likewise.
7212 (vminv_u16): Likewise.
7213 (vminv_u32): Likewise.
7214 (vminvq_f32): Likewise.
7215 (vminvq_f64): Likewise.
7216 (vminvq_s8): Likewise.
7217 (vminvq_s16): Likewise.
7218 (vminvq_s32): Likewise.
7219 (vminvq_u8): Likewise.
7220 (vminvq_u16): Likewise.
7221 (vminvq_u32): Likewise.
7222 (vminnmv_f32): Likewise.
7223 (vminnmvq_f32): Likewise.
7224 (vminnmvq_f64): Likewise.
7225
7226 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7227
7228 * config/aarch64/aarch64-simd.md
7229 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7230 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7231 (*aarch64_mul3_elt<mode>): Likewise.
7232 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7233 (*aarch64_mul3_elt_to_64v2df): Likewise.
7234 (*aarch64_mla_elt<mode>): Likewise.
7235 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7236 (*aarch64_mls_elt<mode>): Likewise.
7237 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7238 (*aarch64_fma4_elt<mode>): Likewise.
7239 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7240 (*aarch64_fma4_elt_to_64v2df): Likewise.
7241 (*aarch64_fnma4_elt<mode>): Likewise.
7242 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7243 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7244 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7245 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7246 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7247 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7248 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7249 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7250 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7251
7252 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7253
7254 * config/aarch64/aarch64-simd.md
7255 (aarch64_be_checked_get_lane<mode>): New define_expand.
7256 * config/aarch64/aarch64-simd-builtins.def
7257 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7258 New builtin definition.
7259 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7260 Use new safe be builtin.
7261
7262 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7263
7264 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7265 New define_insn.
7266 (aarch64_be_st1<mode>): Likewise.
7267 (aarch_ld1<VALL:mode>): Define_expand modified.
7268 (aarch_st1<VALL:mode>): Likewise.
7269 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7270 (UNSPEC_ST1): Likewise.
7271
7272 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7273
7274 * config/microblaze/microblaze.md: Add trap insn and attribute
7275
7276 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7277
7278 PR preprocessor/58580
7279 * input.h (location_get_source_line): Take an additional line_size
7280 parameter.
7281 (void diagnostics_file_cache_fini): Declare new function.
7282 * input.c (struct fcache): New type.
7283 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7284 New static constants.
7285 (diagnostic_file_cache_init, total_lines_num)
7286 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7287 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7288 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7289 (get_next_line, read_next_line, goto_next_line, read_line_num):
7290 New static function definitions.
7291 (diagnostic_file_cache_fini): New function.
7292 (location_get_source_line): Take an additional output line_len
7293 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7294 read_line_num.
7295 * diagnostic.c (diagnostic_finish): Call
7296 diagnostic_file_cache_fini.
7297 (adjust_line): Take an additional input parameter for the length
7298 of the line, rather than calculating it with strlen.
7299 (diagnostic_show_locus): Adjust the use of
7300 location_get_source_line and adjust_line with respect to their new
7301 signature. While displaying a line now, do not stop at the first
7302 null byte. Rather, display the zero byte as a space and keep
7303 going until we reach the size of the line.
7304 * Makefile.in: Add vec.o to OBJS-libcommon
7305
7306 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7307 Ilya Tocar <ilya.tocar@intel.com>
7308
7309 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7310 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7311 (__builtin_ia32_kmov16): Ditto.
7312 * config/i386/i386.md (UNSPEC_KMOV): New.
7313 (kmovw): Ditto.
7314
7315 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7316
7317 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7318 (_mm512_storeu_si512): Ditto.
7319
7320 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7321
7322 PR target/52125
7323 * rtl.h (get_referenced_operands): Declare.
7324 * recog.c (get_referenced_operands): New function.
7325 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7326 operands have been referenced when recording LO_SUM references.
7327
7328 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7329
7330 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7331
7332 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7333
7334 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7335 Enable for generic and recent AMD targets.
7336
7337 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7338
7339 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7340 ARG_SIZE note when adjustment was eliminated.
7341
7342 2014-01-22 Jeff Law <law@redhat.com>
7343
7344 PR tree-optimization/59597
7345 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7346 in file. Accept new argument REGISTERING and use it to modify
7347 dump output appropriately.
7348 (register_jump_thread): Corresponding changes.
7349 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7350 thread paths involving joiner blocks. Add code to dump cancelled
7351 jump threading paths.
7352
7353 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7354
7355 PR rtl-optimization/59477
7356 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7357 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7358
7359 2014-01-22 Tom Tromey <tromey@redhat.com>
7360
7361 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7362 PARAMS.
7363 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7364
7365 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7366
7367 PR rtl-optimization/59896
7368 * lra-constraints.c (process_alt_operands): Check unused note for
7369 matched operands of insn with no output reloads.
7370
7371 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7372
7373 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7374 (mips_move_from_gpr_cost): Likewise.
7375
7376 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7377
7378 PR rtl-optimization/59858
7379 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7380 ira_class_hard_regs_num.
7381 (process_alt_operands): Increase reject for dying matched operand.
7382
7383 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7384
7385 PR target/59003
7386 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7387 smaller than size, perform several stores or loads and stores
7388 at dst + count - size to store or copy all of size bytes, rather
7389 than just last modesize bytes.
7390
7391 2014-01-20 DJ Delorie <dj@redhat.com>
7392
7393 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7394 that CLOBBERs are REGs before propogating their values.
7395
7396 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7397
7398 PR middle-end/59789
7399 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7400 (cgraph_inline_failed_type): New function.
7401 * cgraph.h (DEFCIFCODE): Add type.
7402 (cgraph_inline_failed_type_t): New enum.
7403 (cgraph_inline_failed_type): New prototype.
7404 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7405 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7406 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7407 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7408 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7409 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7410 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7411 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7412 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7413 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7414 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7415 OPTIMIZATION_MISMATCH.
7416 * tree-inline.c (expand_call_inline): Emit errors during
7417 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7418
7419 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7420
7421 PR target/59685
7422 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7423 mode attribute in insn output.
7424
7425 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7426
7427 * output.h (output_constant): Delete.
7428 * varasm.c (output_constant): Make private.
7429
7430 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7431
7432 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7433
7434 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7435
7436 PR middle-end/59860
7437 * tree.h (fold_builtin_strcat): New prototype.
7438 * builtins.c (fold_builtin_strcat): No longer static. Add len
7439 argument, if non-NULL, don't call c_strlen. Optimize
7440 directly into __builtin_memcpy instead of __builtin_strcpy.
7441 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7442 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7443
7444 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7445
7446 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7447 for SImode_address_operand operands, having only a REG argument.
7448
7449 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7450
7451 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7452 loader name using mbig-endian.
7453 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7454
7455 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7456
7457 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7458 (-mtune): Likewise.
7459 (-mcpu): Likewise.
7460
7461 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7462
7463 * config/aarch64/aarch64-protos.h
7464 (aarch64_cannot_change_mode_class_ptr): Declare.
7465 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7466 aarch64_cannot_change_mode_class_ptr): New.
7467 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7468 backend hook aarch64_cannot_change_mode_class.
7469
7470 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7471
7472 * common/config/aarch64/aarch64-common.c
7473 (aarch64_handle_option): Don't handle any option order logic here.
7474 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7475 selected_cpu, warn on architecture version mismatch.
7476 (aarch64_override_options): Fix parsing order for option strings.
7477
7478 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7479 Iain Sandoe <iain@codesourcery.com>
7480
7481 PR bootstrap/59496
7482 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7483 warning. Amend comment to reflect current functionality.
7484
7485 2014-01-20 Richard Biener <rguenther@suse.de>
7486
7487 PR middle-end/59860
7488 * builtins.c (fold_builtin_strcat): Remove case better handled
7489 by tree-ssa-strlen.c.
7490
7491 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7492
7493 * config/aarch64/aarch64.opt
7494 (mcpu, march, mtune): Make case-insensitive.
7495
7496 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7497
7498 PR target/59880
7499 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7500 if operands[1] is a REG or ZERO_EXTEND of a REG.
7501
7502 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7503
7504 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7505
7506 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7507
7508 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7509 long non-pic millicode calls.
7510
7511 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7512
7513 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7514
7515 2014-01-19 Kito Cheng <kito@0xlab.org>
7516
7517 * builtins.c (expand_movstr): Check movstr expand done or fail.
7518
7519 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7520 H.J. Lu <hongjiu.lu@intel.com>
7521
7522 PR target/59379
7523 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7524 to DImode for zero-extended addresses.
7525
7526 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7527
7528 PR rtl-optimization/57763
7529 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7530 on the new indirect jump_insn and increment LABEL_NUSES (label).
7531
7532 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7533
7534 PR bootstrap/59580
7535 PR bootstrap/59583
7536 * config.gcc (x86_archs): New variable.
7537 (x86_64_archs): Likewise.
7538 (x86_cpus): Likewise.
7539 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7540 --with-arch/--with-cpu= options.
7541 Support --with-arch=/--with-cpu={nehalem,westmere,
7542 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7543
7544 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7545
7546 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7547 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7548
7549 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7550
7551 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7552
7553 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7554
7555 PR target/58944
7556 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7557 clear cpp_get_options (parse_in)->warn_unused_macros for
7558 ix86_target_macros_internal with cpp_define.
7559
7560 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7561
7562 * jump.c (delete_related_insns): Keep (use (insn))s.
7563 * reorg.c (redundant_insn): Check for barriers too.
7564
7565 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7566
7567 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7568
7569 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7570
7571 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7572 call to $$dyncall when TARGET_LONG_CALLS is true.
7573
7574 2014-01-17 Jeff Law <law@redhat.com>
7575
7576 * ree.c (combine_set_extension): Temporarily disable test for
7577 changing number of hard registers.
7578
7579 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7580
7581 PR middle-end/58125
7582 * ipa-inline-analysis.c (inline_free_summary):
7583 Do not free summary of aliases.
7584
7585 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7586
7587 PR middle-end/59706
7588 * gimplify.c (gimplify_expr): Use create_tmp_var
7589 instead of create_tmp_var_raw. If cond doesn't have
7590 integral type, don't add the IFN_ANNOTATE builtin at all.
7591
7592 2014-01-17 Martin Jambor <mjambor@suse.cz>
7593
7594 PR ipa/59736
7595 * ipa-cp.c (prev_edge_clone): New variable.
7596 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7597 Also resize prev_edge_clone vector.
7598 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7599 (ipcp_edge_removal_hook): New function.
7600 (ipcp_driver): Register ipcp_edge_removal_hook.
7601
7602 2014-01-17 Andrew Pinski <apinski@cavium.com>
7603 Steve Ellcey <sellcey@mips.com>
7604
7605 PR target/59462
7606 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7607 of operator mode.
7608
7609 2014-01-17 Jeff Law <law@redhat.com>
7610
7611 PR middle-end/57904
7612 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7613 so that pass_ccp runs first.
7614
7615 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7616
7617 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7618 (ix86_adjust_cost): Use !TARGET_XXX.
7619 (do_reorder_for_imul): Likewise.
7620 (swap_top_of_ready_list): Likewise.
7621 (ix86_sched_reorder): Likewise.
7622
7623 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7624
7625 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7626 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7627 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7628 (intel_memset): New. Duplicate slm_memset.
7629 (intel_cost): New. Duplicate slm_cost.
7630 (m_INTEL): New macro.
7631 (processor_target_table): Add "intel".
7632 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7633 with PROCESSOR_INTEL for "intel".
7634 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7635 PROCESSOR_SILVERMONT.
7636 (ix86_issue_rate): Likewise.
7637 (ix86_adjust_cost): Likewise.
7638 (ia32_multipass_dfa_lookahead): Likewise.
7639 (swap_top_of_ready_list): Likewise.
7640 (ix86_sched_reorder): Likewise.
7641 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7642 instead of TARGET_OPT_AGU.
7643 * config/i386/i386.h (TARGET_INTEL): New.
7644 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7645 (processor_type): Add PROCESSOR_INTEL.
7646 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7647 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7648
7649 2014-01-17 Marek Polacek <polacek@redhat.com>
7650
7651 PR c/58346
7652 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7653 size is zero.
7654
7655 2014-01-17 Richard Biener <rguenther@suse.de>
7656
7657 PR tree-optimization/46590
7658 * opts.c (default_options_table): Add entries for
7659 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7660 all enabled at -O1 but not for -Og.
7661 * common.opt (fbranch-count-reg): Remove Init(1).
7662 (fmove-loop-invariants): Likewise.
7663 (ftree-pta): Likewise.
7664
7665 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7666
7667 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7668 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7669 decls to at least the GCC 4.8 used alignments.
7670
7671 PR fortran/59440
7672 * tree-nested.c (convert_nonlocal_reference_stmt,
7673 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7674 of GIMPLE_BIND stmts, adjust associated decls.
7675
7676 2014-01-17 Richard Biener <rguenther@suse.de>
7677
7678 PR tree-optimization/46590
7679 * vec.h (vec<>::bseach): New member function implementing
7680 binary search according to C89 bsearch.
7681 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7682 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7683 bitmap pointer again. Make accesses_in_loop a flat array.
7684 (mem_ref_obstack): New global.
7685 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7686 (mark_ref_stored): Likewise.
7687 (ref_indep_loop_p_2): Likewise.
7688 (set_ref_stored_in_loop): New helper function.
7689 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7690 (memref_free): Adjust.
7691 (record_mem_ref_loc): Simplify.
7692 (gather_mem_refs_stmt): Adjust.
7693 (sort_locs_in_loop_postorder_cmp): New function.
7694 (analyze_memory_references): Sort accesses_in_loop after
7695 loop postorder number.
7696 (find_ref_loc_in_loop_cmp): New function.
7697 (for_all_locs_in_loop): Find relevant cluster of locs in
7698 accesses_in_loop and iterate without recursion.
7699 (execute_sm): Avoid uninit warning.
7700 (struct ref_always_accessed): Simplify.
7701 (ref_always_accessed::operator ()): Likewise.
7702 (ref_always_accessed_p): Likewise.
7703 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7704 loop postorder numbers here.
7705 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7706 numbers.
7707
7708 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7709
7710 PR c++/57945
7711 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7712 on decls for which assemble_alias has been called.
7713
7714 2014-01-17 Nick Clifton <nickc@redhat.com>
7715
7716 * config/msp430/msp430.opt: (mcpu): New option.
7717 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7718 (msp430_option_override): Parse target_cpu. If the MCU name
7719 matches a generic string, clear target_mcu.
7720 (msp430_attr): Allow numeric interrupt values up to 63.
7721 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7722 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7723 option.
7724 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7725 Add mcpu matches.
7726 * config/msp430/msp430.md (popm): Use %J rather than %I.
7727 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7728 (addhi_cy_i): Use immediate_operand for operand 2.
7729 * doc/invoke.texi: Document -mcpu option.
7730
7731 2014-01-17 Richard Biener <rguenther@suse.de>
7732
7733 PR rtl-optimization/38518
7734 * df.h (df_analyze_loop): Declare.
7735 * df-core.c: Include cfgloop.h.
7736 (df_analyze_1): Split out main part of df_analyze.
7737 (df_analyze): Adjust.
7738 (loop_inverted_post_order_compute): New function.
7739 (loop_post_order_compute): Likewise.
7740 (df_analyze_loop): New function avoiding whole-function
7741 postorder computes.
7742 * loop-invariant.c (find_defs): Use df_analyze_loop.
7743 (find_invariants): Adjust.
7744 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7745
7746 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7747
7748 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7749 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7750
7751 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7752
7753 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7754 traversal when removing references.
7755
7756 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7757
7758 PR ipa/59775
7759 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7760
7761 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7762
7763 PR middle-end/56791
7764 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7765 pushing a reload for an autoinc when we had previously reloaded an
7766 inner part of the address.
7767
7768 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7769
7770 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7771 field.
7772 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7773 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7774 when not giving up or versioning for alias only because of
7775 loop->safelen.
7776 (vect_analyze_data_ref_dependences): Set to true.
7777 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7778 is a GIMPLE_PHI.
7779 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7780 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7781 to the condition.
7782
7783 PR middle-end/58344
7784 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7785
7786 PR target/59839
7787 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7788 operand 0 predicate for gathers, use a new pseudo as subtarget.
7789
7790 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7791
7792 PR middle-end/59609
7793 * lra-constraints.c (process_alt_operands): Add printing debug info.
7794 Check absence of input/output reloads for matched operands too.
7795
7796 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7797
7798 PR rtl-optimization/59835
7799 * ira.c (ira_init_register_move_cost): Increase cost for
7800 impossible modes.
7801
7802 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7803
7804 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7805
7806 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7807
7808 PR target/59780
7809 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7810 non-register objects. Use gen_(high/low)part more consistently.
7811 Fix assertions.
7812
7813 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7814
7815 PR target/59844
7816 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7817 endian support, remove tests for WORDS_BIG_ENDIAN.
7818 (p8_mfvsrd_3_<mode>): Likewise.
7819 (reload_gpr_from_vsx<mode>): Likewise.
7820 (reload_gpr_from_vsxsf): Likewise.
7821 (p8_mfvsrd_4_disf): Likewise.
7822
7823 2014-01-16 Richard Biener <rguenther@suse.de>
7824
7825 PR rtl-optimization/46590
7826 * lcm.c (compute_antinout_edge): Use postorder iteration.
7827 (compute_laterin): Use inverted postorder iteration.
7828
7829 2014-01-16 Nick Clifton <nickc@redhat.com>
7830
7831 PR middle-end/28865
7832 * varasm.c (output_constant): Return the number of bytes actually
7833 emitted.
7834 (output_constructor_array_range): Update the field size with the
7835 number of bytes emitted by output_constant.
7836 (output_constructor_regular_field): Likewise. Also do not
7837 complain if the total number of bytes emitted is now greater
7838 than the expected fieldpos.
7839 * output.h (output_constant): Update prototype and descriptive comment.
7840
7841 2014-01-16 Marek Polacek <polacek@redhat.com>
7842
7843 PR middle-end/59827
7844 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7845 it is error_mark_node.
7846
7847 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7848
7849 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7850 VALID_AVX256_REG_OR_OI_MODE.
7851
7852 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7853
7854 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7855 current procedure should be profiled.
7856
7857 2014-01-15 Andrew Pinski <apinski@cavium.com>
7858
7859 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7860 of moving from/to the STACK_REG register class.
7861
7862 2014-01-15 Richard Henderson <rth@redhat.com>
7863
7864 PR debug/54694
7865 * reginfo.c (global_regs_decl): Globalize.
7866 * rtl.h (global_regs_decl): Declare.
7867 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7868 reserved via global_regs.
7869
7870 2014-01-15 Teresa Johnson <tejohnson@google.com>
7871
7872 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7873
7874 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7875
7876 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7877 and vmulosh rather than call gen_vec_widen_smult_*.
7878 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7879 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7880 (vec_widen_smult_even_v16qi): Likewise.
7881 (vec_widen_umult_even_v8hi): Likewise.
7882 (vec_widen_smult_even_v8hi): Likewise.
7883 (vec_widen_umult_odd_v16qi): Likewise.
7884 (vec_widen_smult_odd_v16qi): Likewise.
7885 (vec_widen_umult_odd_v8hi): Likewise.
7886 (vec_widen_smult_odd_v8hi): Likewise.
7887 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7888 vmuloub rather than call gen_vec_widen_umult_*.
7889 (vec_widen_umult_lo_v16qi): Likewise.
7890 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7891 vmulosb rather than call gen_vec_widen_smult_*.
7892 (vec_widen_smult_lo_v16qi): Likewise.
7893 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7894 rather than call gen_vec_widen_umult_*.
7895 (vec_widen_umult_lo_v8hi): Likewise.
7896 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7897 rather than call gen_vec_widen_smult_*.
7898 (vec_widen_smult_lo_v8hi): Likewise.
7899
7900 2014-01-15 Jeff Law <law@redhat.com>
7901
7902 PR tree-optimization/59747
7903 * ree.c (find_and_remove_re): Properly handle case where a second
7904 eliminated extension requires widening a copy created for elimination
7905 of a prior extension.
7906 (combine_set_extension): Ensure that the number of hard regs needed
7907 for a destination register does not change when we widen it.
7908
7909 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7910
7911 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7912 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7913 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7914 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7915 (avr-*-rtems*): Likewise.
7916 (bfin*-rtems*): Likewise.
7917 (moxie-*-rtems*): Likewise.
7918 (h8300-*-rtems*): Likewise.
7919 (i[34567]86-*-rtems*): Likewise.
7920 (lm32-*-rtems*): Likewise.
7921 (m32r-*-rtems*): Likewise.
7922 (m68k-*-rtems*): Likewise.
7923 (microblaze*-*-rtems*): Likewise.
7924 (mips*-*-rtems*): Likewise.
7925 (powerpc-*-rtems*): Likewise.
7926 (sh-*-rtems*): Likewise.
7927 (sparc-*-rtems*): Likewise.
7928 (sparc64-*-rtems*): Likewise.
7929 (v850-*-rtems*): Likewise.
7930 (m32c-*-rtems*): Likewise.
7931
7932 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7933
7934 PR rtl-optimization/59511
7935 * ira.c (ira_init_register_move_cost): Use memory costs for some
7936 cases of register move cost calculations.
7937 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7938 instead of BB frequency.
7939 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7940 * lra-assigns.c (find_hard_regno_for): Ditto.
7941
7942 2014-01-15 Richard Biener <rguenther@suse.de>
7943
7944 PR tree-optimization/59822
7945 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7946 (vectorizable_load): Use it to hoist defs of uses of invariant
7947 loads out of the loop.
7948
7949 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7950 Kugan Vivekanandarajah <kuganv@linaro.org>
7951
7952 PR target/59695
7953 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7954 truncation.
7955
7956 2014-01-15 Richard Biener <rguenther@suse.de>
7957
7958 PR rtl-optimization/59802
7959 * lcm.c (compute_available): Use inverted postorder to seed
7960 the initial worklist.
7961
7962 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7963
7964 PR target/59803
7965 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7966 ADDR_REGS for invalid symrefs in non-PIC code.
7967
7968 2014-01-15 Jakub Jelinek <jakub@redhat.com>
7969
7970 PR other/58712
7971 * builtins.c (determine_block_size): Initialize *probable_max_size
7972 even if len_rtx is CONST_INT.
7973
7974 2014-01-14 Andrew Pinski <apinski@cavium.com>
7975
7976 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7977 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7978 (cortexa53_tunings): Likewise.
7979 (aarch64_sched_issue_rate): New function.
7980 (TARGET_SCHED_ISSUE_RATE): Define.
7981
7982 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7983
7984 * ira-costs.c (find_costs_and_classes): Add missed
7985 ira_init_register_move_cost_if_necessary.
7986
7987 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7988
7989 PR target/59787
7990 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
7991
7992 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
7993
7994 PR target/59794
7995 * config/i386/i386.c (type_natural_mode): Add a bool parameter
7996 to indicate if type is used for function return value. Warn ABI
7997 change if the vector mode isn't available for function return value.
7998 (ix86_function_arg_advance): Pass false to type_natural_mode.
7999 (ix86_function_arg): Likewise.
8000 (ix86_gimplify_va_arg): Likewise.
8001 (function_arg_32): Don't warn ABI change.
8002 (ix86_function_value): Pass true to type_natural_mode.
8003 (ix86_return_in_memory): Likewise.
8004 (ix86_struct_value_rtx): Removed.
8005 (TARGET_STRUCT_VALUE_RTX): Likewise.
8006
8007 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8008
8009 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8010 converting a conditional jump into a conditional return.
8011
8012 2014-01-14 Richard Biener <rguenther@suse.de>
8013
8014 PR tree-optimization/58921
8015 PR tree-optimization/59006
8016 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8017 hoisting invariant stmts.
8018 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8019 invariant loads on the preheader edge if possible.
8020
8021 2014-01-14 Joey Ye <joey.ye@arm.com>
8022
8023 * doc/plugin.texi (Building GCC plugins): Update to C++.
8024
8025 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8026
8027 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8028 (_mm_rcp28_round_ss): Ditto.
8029 (_mm_rsqrt28_round_sd): Ditto.
8030 (_mm_rsqrt28_round_ss): Ditto.
8031 (_mm_rcp28_sd): Ditto.
8032 (_mm_rcp28_ss): Ditto.
8033 (_mm_rsqrt28_sd): Ditto.
8034 (_mm_rsqrt28_ss): Ditto.
8035 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8036 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8037 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8038 (IX86_BUILTIN_RCP28SD): Ditto.
8039 (IX86_BUILTIN_RCP28SS): Ditto.
8040 (IX86_BUILTIN_RSQRT28SD): Ditto.
8041 (IX86_BUILTIN_RSQRT28SS): Ditto.
8042 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8043 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8044 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8045 (ix86_expand_special_args_builtin): Expand new FTYPE.
8046 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8047 (srcp14<mode>): Make insn unary.
8048 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8049 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8050 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8051 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8052 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8053 Fix rounding: make it SAE only.
8054 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8055 Ditto.
8056 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8057 Ditto.
8058 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8059 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8060 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8061 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8062 (round_saeonly_mask_scalar_operand4): Ditto.
8063 (round_saeonly_mask_scalar_op3): Ditto.
8064 (round_saeonly_mask_scalar_op4): Ditto.
8065
8066 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8067
8068 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8069 Implement -maltivec=be for vec_insert and vec_extract.
8070
8071 2014-01-10 DJ Delorie <dj@redhat.com>
8072
8073 * config/msp430/msp430.md (call_internal): Don't allow memory
8074 references with SP as the base register.
8075 (call_value_internal): Likewise.
8076 * config/msp430/constraints.md (Yc): New. For memory references
8077 that don't use SP as a base register.
8078
8079 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8080 "an integer without a # prefix"
8081 * config/msp430/msp430.md (epilogue_helper): Use it.
8082
8083 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8084
8085 PR target/59617
8086 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8087 AVX512F gather builtins.
8088 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8089 on gather decls with INTEGER_TYPE masktype.
8090 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8091 directly into the builtin rather than hoisting it before loop.
8092
8093 PR tree-optimization/59387
8094 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8095 (scev_const_prop): If folded_casts and type has undefined overflow,
8096 use force_gimple_operand instead of force_gimple_operand_gsi and
8097 for each added stmt if it is assign with
8098 arith_code_with_undefined_signed_overflow, call
8099 rewrite_to_defined_overflow.
8100 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8101 gimple-fold.h instead.
8102 (arith_code_with_undefined_signed_overflow,
8103 rewrite_to_defined_overflow): Moved to ...
8104 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8105 rewrite_to_defined_overflow): ... here. No longer static.
8106 Include gimplify-me.h.
8107 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8108 rewrite_to_defined_overflow): New prototypes.
8109
8110 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8111
8112 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8113
8114 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8115
8116 * builtins.c (get_object_alignment_2): Minor tweak.
8117 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8118
8119 2014-01-13 Christian Bruel <christian.bruel@st.com>
8120
8121 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8122 optimized non constant lengths.
8123
8124 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8125
8126 PR libgomp/59194
8127 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8128 load as __atomic_load_N if possible.
8129
8130 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8131
8132 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8133 target parameter.
8134 (rs6000_expand_builtin): Adjust call.
8135
8136 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8137
8138 PR target/58115
8139 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8140 * config/rs6000/rs6000.c: Include target-globals.h.
8141 (rs6000_set_current_function): Instead of doing target_reinit
8142 unconditionally, use save_target_globals_default_opts and
8143 restore_target_globals.
8144
8145 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8146 FPSCR.
8147 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8148 (rs6000_expand_builtin): Handle mffs and mtfsf.
8149 (rs6000_init_builtins): Define mffs and mtfsf.
8150 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8151 (rs6000_mffs): New pattern.
8152 (rs6000_mtfsf): New pattern.
8153
8154 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8155
8156 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8157 Start narrowing with START. Apply candidate-use pair
8158 and check overall cost in narrowing.
8159 (iv_ca_prune): Pass new argument.
8160
8161 2014-01-10 Jeff Law <law@redhat.com>
8162
8163 PR middle-end/59743
8164 * ree.c (combine_reaching_defs): Ensure the defining statement
8165 occurs before the extension when optimizing extensions with
8166 different source and destination hard registers.
8167
8168 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8169
8170 PR ipa/58585
8171 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8172 vtables into the type inheritance graph.
8173
8174 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8175
8176 PR rtl-optimization/59754
8177 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8178 modes in the REGNO != REGNO case.
8179
8180 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8181
8182 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8183
8184 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8185
8186 PR tree-optimization/59745
8187 * tree-predcom.c (tree_predictive_commoning_loop): Call
8188 free_affine_expand_cache if giving up because components is NULL.
8189
8190 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8191 GC in payload of target_globals struct instead of allocating them on
8192 the heap and the larger structs separately using GC.
8193 * target-globals.h (struct target_globals): Make regs, hard_regs,
8194 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8195 of GTY((skip)) and change type to void *.
8196 (reset_target_globals): Cast loads from those fields to corresponding
8197 types.
8198
8199 2014-01-10 Steve Ellcey <sellcey@mips.com>
8200
8201 PR plugins/59335
8202 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8203 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8204 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8205
8206 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8207
8208 PR target/59744
8209 * aarch64-modes.def (CC_Zmode): New flags mode.
8210 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8211 represents an equality.
8212 (aarch64_get_condition_code): Handle CC_Zmode.
8213 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8214
8215 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8216
8217 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8218 extraction in good case.
8219
8220 2014-01-10 Richard Biener <rguenther@suse.de>
8221
8222 PR tree-optimization/59374
8223 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8224 checking after SLP discovery. Mark stmts not participating
8225 in any SLP instance properly.
8226
8227 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8228
8229 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8230 when handling a SET rtx.
8231
8232 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8233
8234 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8235 (cortex-a57): Likewise.
8236 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8237
8238 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8239
8240 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8241 non-iwmmxt builtins.
8242
8243 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8244
8245 PR ipa/58252
8246 PR ipa/59226
8247 * ipa-devirt.c record_target_from_binfo): Take as argument
8248 stack of binfos and lookup matching one for virtual inheritance.
8249 (possible_polymorphic_call_targets_1): Update.
8250
8251 2014-01-10 Huacai Chen <chenhc@lemote.com>
8252
8253 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8254 kernel strings for Loongson-2E/2F/3A.
8255
8256 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8257
8258 PR middle-end/59670
8259 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8260 is_gimple_call before calling gimple_call_internal_p.
8261
8262 2014-01-09 Steve Ellcey <sellcey@mips.com>
8263
8264 * Makefile.in (TREE_FLOW_H): Remove.
8265 (TREE_SSA_H): Add file names from tree-flow.h.
8266 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8267 * tree.h: Remove tree-flow.h reference.
8268 * hash-table.h: Remove tree-flow.h reference.
8269 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8270 reference with tree-ssa-loop.h.
8271
8272 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8273
8274 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8275 default element-order behavior for -maltivec.
8276 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8277 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8278 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8279 when targeting big endian, at least for now.
8280 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8281
8282 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8283
8284 PR middle-end/47735
8285 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8286 var satisfies use_register_for_decl, just take into account type
8287 alignment, rather than decl alignment.
8288
8289 PR tree-optimization/59622
8290 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8291 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8292 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8293 Don't devirtualize for inplace at all. For targets.length () == 1,
8294 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8295
8296 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8297
8298 * config/i386/i386.md (cpu): Remove the unused btver1.
8299
8300 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8301
8302 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8303
8304 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8305
8306 PR target/58115
8307 * tree-core.h (struct target_globals): New forward declaration.
8308 (struct tree_target_option): Add globals field.
8309 * tree.h (TREE_TARGET_GLOBALS): Define.
8310 (prepare_target_option_nodes_for_pch): New prototype.
8311 * target-globals.h (struct target_globals): Define even if
8312 !SWITCHABLE_TARGET.
8313 * tree.c (prepare_target_option_node_for_pch,
8314 prepare_target_option_nodes_for_pch): New functions.
8315 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8316 * config/i386/i386.c: Include target-globals.h.
8317 (ix86_set_current_function): Instead of doing target_reinit
8318 unconditionally, use save_target_globals_default_opts and
8319 restore_target_globals.
8320
8321 2014-01-09 Richard Biener <rguenther@suse.de>
8322
8323 PR tree-optimization/59715
8324 * tree-cfg.h (split_critical_edges): Declare.
8325 * tree-cfg.c (split_critical_edges): Export.
8326 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8327
8328 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8329
8330 * cfgexpand.c (expand_stack_vars): Optionally disable
8331 asan stack protection.
8332 (expand_used_vars): Likewise.
8333 (partition_stack_vars): Likewise.
8334 * asan.c (asan_emit_stack_protection): Optionally disable
8335 after return stack usage.
8336 (instrument_derefs): Optionally disable memory access instrumentation.
8337 (instrument_builtin_call): Likewise.
8338 (instrument_strlen_call): Likewise.
8339 (asan_protect_global): Optionally disable global variables protection.
8340 * doc/invoke.texi: Added doc for new options.
8341 * params.def: Added new options.
8342 * params.h: Likewise.
8343
8344 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8345
8346 PR rtl-optimization/59724
8347 * ifcvt.c (cond_exec_process_if_block): Don't call
8348 flow_find_head_matching_sequence with 0 longest_match.
8349 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8350 non-active insns if !stop_after.
8351 (try_head_merge_bb): Revert 2014-01-07 changes.
8352
8353 2014-01-08 Jeff Law <law@redhat.com>
8354
8355 * ree.c (get_sub_rtx): New function, extracted from...
8356 (merge_def_and_ext): Here.
8357 (combine_reaching_defs): Use get_sub_rtx.
8358
8359 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8360
8361 * cgraph.h (varpool_variable_node): Do not choke on null node.
8362
8363 2014-01-08 Catherine Moore <clm@codesourcery.com>
8364
8365 * config/mips/mips.md (simple_return): Attempt to use JRC
8366 for microMIPS.
8367 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8368
8369 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8370
8371 PR rtl-optimization/59137
8372 * reorg.c (steal_delay_list_from_target): Call update_block for
8373 elided insns.
8374 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8375
8376 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8377
8378 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8379 two duplicate entries.
8380
8381 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8382
8383 Revert:
8384 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8385
8386 * config/mips/mips.c (mips_truncated_op_cost): New function.
8387 (mips_rtx_costs): Adjust test for BADDU.
8388 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8389
8390 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8391
8392 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8393 (*baddu_si): ...this new pattern.
8394
8395 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8396
8397 PR ipa/59722
8398 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8399
8400 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8401
8402 PR middle-end/57748
8403 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8404 inner_reference_p.
8405 (expand_expr, expand_normal): Adjust.
8406 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8407 inner_reference_p. Use inner_reference_p to expand inner references.
8408 (store_expr): Adjust.
8409 * cfgexpand.c (expand_call_stmt): Adjust.
8410
8411 2014-01-08 Rong Xu <xur@google.com>
8412
8413 * gcov-io.c (gcov_var): Move from gcov-io.h.
8414 (gcov_position): Ditto.
8415 (gcov_is_error): Ditto.
8416 (gcov_rewrite): Ditto.
8417 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8418 only part to libgcc/libgcov.h.
8419
8420 2014-01-08 Marek Polacek <polacek@redhat.com>
8421
8422 PR middle-end/59669
8423 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8424
8425 2014-01-08 Marek Polacek <polacek@redhat.com>
8426
8427 PR sanitizer/59667
8428 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8429
8430 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8431
8432 PR rtl-optimization/59649
8433 * stor-layout.c (get_mode_bounds): For BImode return
8434 0 and STORE_FLAG_VALUE.
8435
8436 2014-01-08 Richard Biener <rguenther@suse.de>
8437
8438 PR middle-end/59630
8439 * gimple.h (is_gimple_builtin_call): Remove.
8440 (gimple_builtin_call_types_compatible_p): New.
8441 (gimple_call_builtin_p): New overload.
8442 * gimple.c (is_gimple_builtin_call): Remove.
8443 (validate_call): Rename to ...
8444 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8445 check return types.
8446 (validate_type): New static function.
8447 (gimple_call_builtin_p): New overload and adjust.
8448 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8449 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8450 (gimple_fold_stmt_to_constant_1): Likewise.
8451 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8452
8453 2014-01-08 Richard Biener <rguenther@suse.de>
8454
8455 PR middle-end/59471
8456 * gimplify.c (gimplify_expr): Gimplify register-register type
8457 VIEW_CONVERT_EXPRs to separate stmts.
8458
8459 2014-01-07 Jeff Law <law@redhat.com>
8460
8461 PR middle-end/53623
8462 * ree.c (combine_set_extension): Handle case where source
8463 and destination registers in an extension insn are different.
8464 (combine_reaching_defs): Allow source and destination registers
8465 in extension to be different under limited circumstances.
8466 (add_removable_extension): Remove restriction that the
8467 source and destination registers in the extension are the same.
8468 (find_and_remove_re): Emit a copy from the extension's
8469 destination to its source after the defining insn if
8470 the source and destination registers are different.
8471
8472 PR middle-end/59285
8473 * ifcvt.c (merge_if_block): If we are merging a block with more than
8474 one successor with a block with no successors, remove any BARRIER
8475 after the second block.
8476
8477 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8478
8479 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8480
8481 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8482
8483 PR target/59652
8484 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8485 for 14-bit register offsets when INT14_OK_STRICT is false.
8486
8487 2014-01-07 Roland Stigge <stigge@antcom.de>
8488 Michael Meissner <meissner@linux.vnet.ibm.com>
8489
8490 PR 57386/target
8491 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8492 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8493
8494 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8495
8496 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8497 -mcpu.
8498
8499 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8500
8501 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8502 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8503 rtx is const0_rtx or not.
8504
8505 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8506
8507 PR target/58115
8508 * target-globals.c (save_target_globals): Remove this_fn_optab
8509 handling.
8510 * toplev.c: Include optabs.h.
8511 (target_reinit): Temporarily restore the global options if another
8512 set of options are in force.
8513
8514 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8515
8516 PR rtl-optimization/58668
8517 * cfgcleanup.c (flow_find_cross_jump): Don't count
8518 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8519 to determine what is counted.
8520 (flow_find_head_matching_sequence): Use active_insn_p to determine
8521 what is counted.
8522 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8523 counting change.
8524 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8525 determine what is counted.
8526
8527 PR tree-optimization/59643
8528 * tree-predcom.c (split_data_refs_to_components): If one dr is
8529 read and one write, determine_offset fails and the write isn't
8530 in the bad component, just put the read into the bad component.
8531
8532 2014-01-07 Mike Stump <mikestump@comcast.net>
8533 Jakub Jelinek <jakub@redhat.com>
8534
8535 PR pch/59436
8536 * tree-core.h (struct tree_optimization_option): Change optabs
8537 type from unsigned char * to void *.
8538 * optabs.c (init_tree_optimization_optabs): Adjust
8539 TREE_OPTIMIZATION_OPTABS initialization.
8540
8541 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8542
8543 PR target/59644
8544 * config/i386/i386.h (struct machine_function): Add
8545 no_drap_save_restore field.
8546 * config/i386/i386.c (ix86_save_reg): Use
8547 !cfun->machine->no_drap_save_restore instead of
8548 crtl->stack_realign_needed.
8549 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8550 this function clears frame_pointer_needed. Set
8551 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8552 and DRAP reg is needed.
8553
8554 2014-01-06 Marek Polacek <polacek@redhat.com>
8555
8556 PR c/57773
8557 * doc/implement-c.texi: Mention that other integer types are
8558 permitted as bit-field types in strictly conforming mode.
8559
8560 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8561
8562 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8563 is newly allocated.
8564
8565 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8566
8567 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8568
8569 2014-01-06 Martin Jambor <mjambor@suse.cz>
8570
8571 PR ipa/59008
8572 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8573 to int.
8574 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8575
8576 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8577
8578 PR debug/59350
8579 PR debug/59510
8580 * var-tracking.c (add_stores): Preserve the value of the source even if
8581 we don't record the store.
8582
8583 2014-01-06 Terry Guo <terry.guo@arm.com>
8584
8585 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8586
8587 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8588
8589 PR bootstrap/59541
8590 * config/darwin.c (darwin_function_section): Adjust return values to
8591 correspond to optimisation changes made in r206070.
8592
8593 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8594
8595 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8596 from prefetch_block tune setting.
8597 (nocona_cost): Correct size of prefetch block to 64.
8598
8599 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8600
8601 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8602 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8603 used to save the static chain register in the computation of the offset
8604 from which the FP registers need to be restored.
8605
8606 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8607
8608 PR tree-optimization/59519
8609 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8610 ICE if get_current_def (current_new_name) is already non-NULL, as long
8611 as it is a phi result of some other phi in *new_exit_bb that has
8612 the same argument.
8613
8614 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8615 or vmovdqu* for misaligned_operand.
8616 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8617 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8618 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8619 aligned_mem for AVX512F masked aligned load and store builtins and for
8620 non-temporal moves.
8621
8622 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8623
8624 PR tree-optimization/59651
8625 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8626 Address range for negative step should be added by TYPE_SIZE_UNIT.
8627
8628 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8629
8630 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8631 overlapping registers also for registers other than the stack pointer.
8632
8633 2014-01-03 Marek Polacek <polacek@redhat.com>
8634
8635 PR other/59661
8636 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8637 __builtin_FILE.
8638
8639 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8640
8641 PR target/59625
8642 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8643 asm goto as jump.
8644
8645 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8646 (push splitter): Use <P:MODE_SIZE> instead of
8647 GET_MODE_SIZE (<P:MODE>mode).
8648 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8649 (mov -1, reg peephole2): Likewise.
8650 * config/i386/sse.md (*mov<mode>_internal,
8651 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8652 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8653 *<code><mode>3, *andnot<mode>3<mask_name>,
8654 <mask_codefor><code><mode>3<mask_name>): Likewise.
8655 * config/i386/subst.md (mask_mode512bit_condition,
8656 sd_mask_mode512bit_condition): Likewise.
8657
8658 2014-01-02 Xinliang David Li <davidxl@google.com>
8659
8660 PR tree-optimization/59303
8661 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8662 (dump_predicates): Better output format.
8663 (pred_equal_p): New function.
8664 (is_neq_relop_p): Ditto.
8665 (is_neq_zero_form_p): Ditto.
8666 (pred_expr_equal_p): Ditto.
8667 (pred_neg_p): Ditto.
8668 (simplify_pred): Ditto.
8669 (simplify_preds_2): Ditto.
8670 (simplify_preds_3): Ditto.
8671 (simplify_preds_4): Ditto.
8672 (simplify_preds): Ditto.
8673 (push_pred): Ditto.
8674 (push_to_worklist): Ditto.
8675 (get_pred_info_from_cmp): Ditto.
8676 (is_degenerated_phi): Ditto.
8677 (normalize_one_pred_1): Ditto.
8678 (normalize_one_pred): Ditto.
8679 (normalize_one_pred_chain): Ditto.
8680 (normalize_preds): Ditto.
8681 (normalize_cond_1): Remove function.
8682 (normalize_cond): Ditto.
8683 (is_gcond_subset_of): Ditto.
8684 (is_subset_of_any): Ditto.
8685 (is_or_set_subset_of): Ditto.
8686 (is_and_set_subset_of): Ditto.
8687 (is_norm_cond_subset_of): Ditto.
8688 (pred_chain_length_cmp): Ditto.
8689 (convert_control_dep_chain_into_preds): Type change.
8690 (find_predicates): Ditto.
8691 (find_def_preds): Ditto.
8692 (destroy_predicates_vecs): Ditto.
8693 (find_matching_predicates_in_rest_chains): Ditto.
8694 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8695 (is_pred_expr_subset): Ditto.
8696 (is_pred_chain_subset_of): Ditto.
8697 (is_included_in): Ditto.
8698 (is_superset_of): Ditto.
8699
8700 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8701
8702 Update copyright years.
8703
8704 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8705
8706 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8707 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8708 config/arc/arc.md, config/arc/arc.opt,
8709 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8710 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8711 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8712 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8713 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8714 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8715 vtable-verify.c, vtable-verify.h: Use the standard form for the
8716 copyright notice.
8717
8718 2014-01-02 Tobias Burnus <burnus@net-b.de>
8719
8720 * gcc.c (process_command): Update copyright notice dates.
8721 * gcov-dump.c: Ditto.
8722 * gcov.c: Ditto.
8723 * doc/cpp.texi: Bump @copying's copyright year.
8724 * doc/cppinternals.texi: Ditto.
8725 * doc/gcc.texi: Ditto.
8726 * doc/gccint.texi: Ditto.
8727 * doc/gcov.texi: Ditto.
8728 * doc/install.texi: Ditto.
8729 * doc/invoke.texi: Ditto.
8730
8731 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8732
8733 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8734
8735 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8736
8737 * config/i386/sse.md (*mov<mode>_internal): Guard
8738 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8739
8740 PR rtl-optimization/59647
8741 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8742 new_rtx into UNSIGNED_FLOAT rtxes.
8743 \f
8744 Copyright (C) 2014 Free Software Foundation, Inc.
8745
8746 Copying and distribution of this file, with or without modification,
8747 are permitted in any medium without royalty provided the copyright
8748 notice and this notice are preserved.