sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
[gcc.git] / gcc / ChangeLog
1 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2
3 * config/sh/sh.h (SH_ASM_SPEC): Handle m1, m2*, m3* and m4* cases.
4
5 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
6
7 * config/sh/sh.h (ROUND_ADVANCE): Delete macro.
8 (ROUND_REG, PASS_IN_REG_P): Move and rename macros to ...
9 * config/sh/sh.c (sh_round_reg, sh_pass_in_reg_p): ... these new
10 functions.
11 (sh_arg_partial_bytes, sh_function_arg, sh_function_arg_advance,
12 sh_setup_incoming_varargs): Replace usage of PASS_IN_REG_P with
13 sh_pass_in_reg_p.
14 Replace usage of ROUND_REG with sh_round_reg.
15 Use CEIL instead of ROUND_ADVANCE.
16
17 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
18
19 PR target/61026
20 * config/sh/sh.c: Include stdlib headers before everything else.
21
22 2014-05-02 Jakub Jelinek <jakub@redhat.com>
23
24 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
25 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
26 (gimplify_adjust_omp_clauses): Simd region is never
27 directly nested in combined parallel. Instead, for linear
28 with copyin/copyout, if in combined for simd loop, make decl
29 firstprivate/lastprivate on OMP_FOR.
30 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
31 expand_omp_for_static_chunk): When setting endvar, also set
32 fd->loop.v to the same value.
33
34 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
35
36 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
37
38 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
39
40 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
41 expression.
42
43 2014-05-02 Marek Polacek <polacek@redhat.com>
44
45 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
46
47 2014-02-26 Kito Cheng <kito@0xlab.org>
48
49 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
50 to a C expression marco.
51 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
52 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
53 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
54 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
55 HONOR_REG_ALLOC_ORDER.
56 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
57
58 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
59
60 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
61
62 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
63
64 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
65
66 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
67
68 * tree-if-conv.c (is_cond_scalar_reduction): New function.
69 (convert_scalar_cond_reduction): Likewise.
70 (predicate_scalar_phi): Add recognition and transformation
71 of simple conditioanl reduction to be vectorizable.
72
73 2014-05-01 Marek Polacek <polacek@redhat.com>
74
75 PR c/43245
76 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
77
78 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
79
80 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
81 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
82 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
83 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
84 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
85 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
86 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
87 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
88
89 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
90
91 * config/arc/arc.opt (mlra): Move comment above option name
92 to avoid mis-parsing as language options.
93
94 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
95
96 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
97 * config/sol2.h: ... here.
98 * config/sol2-10.h: Remove.
99
100 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
101 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
102 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
103 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
104 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
105 * config/sol2.h: ... here.
106 (SECTION_NAME_FORMAT): Don't redefine.
107 (STARTFILE_ARCH32_SPEC): Rename to ...
108 (STARTFILE_ARCH_SPEC): ... this.
109 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
110 * config/sparc/sol2.h: ... here.
111 (SECTION_NAME_FORMAT): Don't undef.
112 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
113 (SUBTARGET_EXTRA_SPECS): Remove.
114 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
115
116 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
117 (MD_STARTFILE_PREFIX): Remove.
118 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
119 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
120 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
121 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
122 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
123 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
124 * config/i386/sol2.h: ... here.
125 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
126 * config/i386/sol2-bi.h: Remove.
127 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
128 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
129
130 * config/i386/t-sol2-64: Rename to ...
131 * config/i386/t-sol2: ... this.
132 * config/sparc/t-sol2-64: Rename to ...
133 * config/sparc/t-sol2: ... this.
134
135 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
136 sol2_tm_file_head, sol2_tm_file_tail.
137 Include ${cpu_type}/sol2.h before sol2.h.
138 Remove sol2-10.h.
139 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
140 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
141 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
142 Reflect i386/t-sol2-64 renaming.
143 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
144 Reflect sparc/t-sol2-64 renaming.
145
146 2014-04-30 Richard Biener <rguenther@suse.de>
147
148 * passes.c (execute_function_todo): Move TODO_verify_stmts
149 and TODO_verify_ssa under the TODO_verify_il umbrella.
150 * tree-ssa.h (verify_ssa): Adjust prototype.
151 * tree-ssa.c (verify_ssa): Add parameter to tell whether
152 we should verify SSA operands.
153 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
154 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
155 whether we should verify whether not throwing stmts have EH info.
156 * graphite-scop-detection.c (create_sese_edges): Adjust.
157 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
158 * tree-eh.c (lower_try_finally_switch): Do not add the
159 default case label twice.
160
161 2014-04-30 Marek Polacek <polacek@redhat.com>
162
163 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
164 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
165 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
166 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
167
168 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
169
170 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
171 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
172 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
173 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
174 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
175 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
176 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
177 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
178
179 2014-04-29 David Malcolm <dmalcolm@redhat.com>
180
181 * tree-cfg.c (dump_function_to_file): Dump the return type of
182 functions, in a line to itself before the function body, mimicking
183 the layout of a C function.
184
185 2014-04-29 Jakub Jelinek <jakub@redhat.com>
186
187 PR tree-optimization/60971
188 * tree-tailcall.c (process_assignment): Reject conversions which
189 reduce precision.
190
191 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
192
193 * calls.c (initialize_argument_information): Always treat
194 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
195 (expand_call): Likewise.
196 (emit_library_call_calue_1): Likewise.
197 * expr.c (PUSH_ARGS_REVERSED): Do not define.
198 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
199 code accordingly.
200
201 2014-04-29 Nick Clifton <nickc@redhat.com>
202
203 * config/msp430/msp430.md (umulsidi): Fix typo.
204 (mulhisi3): Enable even inside interrupt handlers.
205 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
206 bigger return address pushed in large mode.
207
208 2014-04-29 Nick Clifton <nickc@redhat.com>
209
210 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
211 (arc_init_reg_tables): Use a machine_mode enum to iterate over
212 available modes.
213 * config/m32r/m32r.c (init_reg_tables): Likewise.
214 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
215 enum to hold the modes.
216
217 2014-04-29 Richard Biener <rguenther@suse.de>
218
219 * dominance.c (free_dominance_info): Add overload with
220 function parameter.
221 (dom_info_state): Likewise.
222 (dom_info_available_p): Likewise.
223 * basic-block.h (free_dominance_info, dom_info_state,
224 dom_info_available_p): Declare overloads.
225 * passes.c (execute_function_todo): Verify that verifiers
226 don't change dominator info state. Drop dominator info
227 for IPA pass invocations.
228 * cgraph.c (release_function_body): Restore asserts that
229 dominator information is released.
230
231 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
232
233 * doc/invoke.texi: Fix typo.
234 * tree-vrp.c: Fix typos.
235 * gimple.c (infer_nonnull_range): Reorder operands of an &&
236 condition.
237
238 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
239
240 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
241
242 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
243
244 * config/aarch64/aarch64-builtins.c
245 (aarch64_types_storestruct_lane_qualifiers): New.
246 (TYPES_STORESTRUCT_LANE): Likewise.
247 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
248 (st3_lane): Likewise.
249 (st4_lane): Likewise.
250 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
251 (vec_store_lanesci_lane<mode>): Likewise.
252 (vec_store_lanesxi_lane<mode>): Likewise.
253 (aarch64_st2_lane<VQ:mode>): Likewise.
254 (aarch64_st3_lane<VQ:mode>): Likewise.
255 (aarch64_st4_lane<VQ:mode>): Likewise.
256 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
257 * config/aarch64/arm_neon.h
258 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
259 use new macro arguments.
260 (__ST3_LANE_FUNC): Likewise.
261 (__ST4_LANE_FUNC): Likewise.
262 * config/aarch64/iterators.md (V_TWO_ELEM): New.
263 (V_THREE_ELEM): Likewise.
264 (V_FOUR_ELEM): Likewise.
265
266 2014-04-28 David Malcolm <dmalcolm@redhat.com>
267
268 * doc/gimple.texi: Replace the description of the now-defunct
269 union gimple_statement_d with a diagram showing the
270 gimple_statement_base class hierarchy and its relationships to
271 the GSS_ and GIMPLE_ enums.
272
273 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
274
275 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
276 * config/aarch64/aarch64.c
277 (aarch64_cannot_change_mode_class): Weaken conditions.
278 (aarch64_modes_tieable_p): New.
279 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
280
281 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
282
283 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
284 (loadsync_<mode>): Change mode.
285 (load_quadpti, store_quadpti): New.
286 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
287 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
288
289 2014-04-28 Martin Jambor <mjambor@suse.cz>
290
291 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
292 same alias type as the original statement.
293 (subreplacement_assignment_data): New type.
294 (handle_unscalarized_data_in_subtree): New type of parameter,
295 generate new memory accesses with same alias type as the original
296 statement.
297 (load_assign_lhs_subreplacements): Likewise.
298 (sra_modify_constructor_assign): Generate new memory accesses with
299 same alias type as the original statement.
300
301 2014-04-28 Richard Biener <rguenther@suse.de>
302
303 * tree-pass.h (TODO_verify_il): Define.
304 (TODO_verify_all): Complete properly.
305 * passes.c (execute_function_todo): Move existing loop-closed
306 SSA verification under TODO_verify_il.
307 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
308 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
309 Fix tree sharing issue.
310
311 2014-04-28 Richard Biener <rguenther@suse.de>
312
313 PR middle-end/60092
314 * builtins.def (DEF_C11_BUILTIN): Add.
315 (BUILT_IN_ALIGNED_ALLOC): Likewise.
316 * coretypes.h (enum function_class): Add function_c11_misc.
317 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
318 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
319 (call_may_clobber_ref_p_1): Likewise.
320 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
321 (mark_all_reaching_defs_necessary_1): Likewise.
322 (propagate_necessity): Likewise.
323 (eliminate_unnecessary_stmts): Likewise.
324 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
325
326 2014-04-28 Richard Biener <rguenther@suse.de>
327
328 * tree-vrp.c (vrp_var_may_overflow): Remove.
329 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
330 with overflow immediately bump to one before that value and
331 let iteration figure out overflow status.
332
333 2014-04-28 Richard Biener <rguenther@suse.de>
334
335 * configure.ac: Do valgrind header checks unconditionally.
336 Add --enable-valgrind-annotations.
337 * system.h: Guard valgrind header inclusion with
338 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
339 * alloc-pool.c (pool_alloc, pool_free): Use
340 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
341 to guard possibly dead code.
342 * config.in: Regenerated.
343 * configure: Likewise.
344
345 2014-04-28 Jeff Law <law@redhat.com>
346
347 PR tree-optimization/60902
348 * tree-ssa-threadedge.c
349 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
350 over real defs when invalidating outputs from statements that do not
351 produce useful outputs for threading.
352
353 2014-04-28 Richard Biener <rguenther@suse.de>
354
355 PR tree-optimization/60979
356 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
357 SCOPs that end in a block with a successor with abnormal
358 predecessors.
359
360 2014-04-28 Richard Biener <rguenther@suse.de>
361
362 * tree-pass.h (execute_pass_list): Adjust prototype.
363 * passes.c (pass_manager::execute_early_local_passes):
364 Adjust.
365 (do_per_function): Change callback signature, push all actual
366 work to the callbals.
367 (do_per_function_toporder): Likewise.
368 (execute_function_dump): Adjust.
369 (execute_function_todo): Likewise.
370 (clear_last_verified): Likewise.
371 (verify_curr_properties): Likewise.
372 (update_properties_after_pass): Likewise.
373 (execute_pass_list_1): Split out from ...
374 (execute_pass_list): ... here. Adjust.
375 (execute_ipa_pass_list): Likewise.
376 * cgraphunit.c (cgraph_add_new_function): Adjust.
377 (analyze_function): Likewise.
378 (expand_function): Likewise.
379 * cgraph.c (release_function_body): Free dominance info
380 here instead of asserting it was magically freed elsewhere.
381
382 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
383
384 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
385 * configure: Regenerate.
386 * config/sparc/sparc.opt (muser-mode): New option.
387 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
388 for LEON3.
389 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
390 * doc/invoke.texi (SPARC options): Document -muser-mode.
391
392 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
393
394 * cselib.c (find_slot_memmode): Delete.
395 (cselib_hasher): Change compare_type to a struct.
396 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
397 constants.
398 (preserve_constants_and_equivs): Adjust for new compare_type.
399 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
400 (wrap_constant): Delete.
401 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
402
403 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
404
405 * doc/install.texi (Building with profile feedback): Remove
406 outdated sentence.
407
408 2014-04-26 Tom de Vries <tom@codesourcery.com>
409
410 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
411 array accesses.
412
413 2014-04-25 Cary Coutant <ccoutant@google.com>
414
415 PR debug/60929
416 * dwarf2out.c (should_move_die_to_comdat): A type definition
417 can contain a subprogram definition, but don't move it to a
418 comdat unit.
419 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
420 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
421 from original DIE.
422 (clone_tree_hash): Rename to...
423 (clone_tree_partial): ...this; change callers. Copy
424 DW_TAG_subprogram DIEs as declarations.
425 (copy_decls_walk): Don't copy children of a declaration into a
426 type unit.
427
428 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
429
430 PR target/60969
431 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
432 alternative 12.
433
434 2014-04-25 Jiong Wang <jiong.wang@arm.com>
435
436 * config/arm/predicates.md (call_insn_operand): Add long_call check.
437 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
438 reg for long_call.
439 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
440 restriction.
441
442 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
443
444 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
445 T16-related fields.
446
447 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
448
449 PR tree-optimization/60930
450 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
451 creating a multiply candidate by folding two constant
452 multiplicands when the result overflows.
453
454 2014-04-25 Jakub Jelinek <jakub@redhat.com>
455
456 PR tree-optimization/60960
457 * tree-vect-generic.c (expand_vector_operation): Only call
458 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
459
460 2014-04-25 Tom de Vries <tom@codesourcery.com>
461
462 * expr.c (clobber_reg_mode): New function.
463 * expr.h (clobber_reg): New function.
464
465 2014-04-25 Tom de Vries <tom@codesourcery.com>
466
467 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
468 clobbers.
469
470 2014-04-25 Radovan Obradovic <robradovic@mips.com>
471 Tom de Vries <tom@codesourcery.com>
472
473 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
474 handle.
475 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
476 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
477 new argument to find_all_hard_reg_sets call.
478
479 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
480
481 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
482 Use HOST_WIDE_INT_C for mask literal.
483 (aarch_rev16_shleft_mask_imm_p): Likewise.
484
485 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
486
487 PR target/60941
488 * config/sparc/sparc.md (ashlsi3_extend): Delete.
489
490 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
491
492 PR preprocessor/56540
493 * config/i386/i386-c.c (ix86_target_macros): Define
494 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
495
496 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
497
498 * configure.ac (tga_func): Remove.
499 (LIB_TLS_SPEC): Remove.
500 * configure: Regenerate.
501 * config.in: Regenerate.
502 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
503
504 2014-04-25 Richard Biener <rguenther@suse.de>
505
506 PR ipa/60912
507 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
508 call stmt use/clobber sets during stmt walk instead of
509 walking the possibly incomplete set of caller edges.
510
511 2014-04-25 Richard Biener <rguenther@suse.de>
512
513 PR ipa/60911
514 * passes.c (apply_ipa_transforms): Inline into only caller ...
515 (execute_one_pass): ... here. Properly bring in function
516 bodies for nodes we want to apply IPA transforms to.
517
518 2014-04-24 Cong Hou <congh@google.com>
519
520 PR tree-optimization/60896
521 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
522 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
523 (vect_mark_pattern_stmts): Set the def type of all statements in
524 PATTERN_DEF_SEQ as vect_internal_def.
525
526 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
527
528 * doc/extend.texi (PowerPC Built-in Functions): Document new
529 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
530 functions.
531 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
532
533 * config/rs6000/predicates.md (const_0_to_3_operand): New
534 predicate to match 0..3 integer constants.
535
536 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
537 to support adding miscellaneous builtin functions.
538 (BU_DFP_MISC_2): Likewise.
539 (BU_P7_MISC_1): Likewise.
540 (BU_P7_MISC_2): Likewise.
541 (BU_P8V_MISC_3): Likewise.
542 (BU_MISC_1): Likewise.
543 (BU_MISC_2): Likewise.
544 (DIVWE): Add extended divide builtin functions.
545 (DIVWEO): Likewise.
546 (DIVWEU): Likewise.
547 (DIVWEUO): Likewise.
548 (DIVDE): Likewise.
549 (DIVDEO): Likewise.
550 (DIVDEU): Likewise.
551 (DIVDEUO): Likewise.
552 (DXEX): Add decimal floating-point builtin functions.
553 (DXEXQ): Likewise.
554 (DDEDPD): Likewise.
555 (DDEDPDQ): Likewise.
556 (DENBCD): Likewise.
557 (DENBCDQ): Likewise.
558 (DIEX): Likewise.
559 (DIEXQ): Likewise.
560 (DSCLI): Likewise.
561 (DSCLIQ): Likewise.
562 (DSCRI): Likewise.
563 (DSCRIQ): Likewise.
564 (CDTBCD): Add new BCD builtin functions.
565 (CBCDTD): Likewise.
566 (ADDG6S): Likewise.
567 (BCDADD): Likewise.
568 (BCDADD_LT): Likewise.
569 (BCDADD_EQ): Likewise.
570 (BCDADD_GT): Likewise.
571 (BCDADD_OV): Likewise.
572 (BCDSUB): Likewise.
573 (BCDSUB_LT): Likewise.
574 (BCDSUB_EQ): Likewise.
575 (BCDSUB_GT): Likewise.
576 (BCDSUB_OV): Likewise.
577 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
578 (UNPACK_TD): Likewise.
579 (PACK_TF): Likewise.
580 (UNPACK_TF): Likewise.
581 (UNPACK_TF_0): Likewise.
582 (UNPACK_TF_1): Likewise.
583 (PACK_V1TI): Likewise.
584 (UNPACK_V1TI): Likewise.
585
586 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
587 support for decimal floating point builtin functions.
588 (rs6000_expand_ternop_builtin): Add checks for the new builtin
589 functions that take constant arguments.
590 (rs6000_invalid_builtin): Add decimal floating point builtin
591 support.
592 (rs6000_init_builtins): Setup long double, _Decimal64, and
593 _Decimal128 types for new builtin functions.
594 (builtin_function_type): Set the unsigned flags appropriately for
595 the new builtin functions.
596 (rs6000_opt_masks): Add support for decimal floating point builtin
597 functions.
598
599 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
600 floating point builtin functions.
601 (RS6000_BTM_COMMON): Likewise.
602 (RS6000_BTI_long_double): Likewise.
603 (RS6000_BTI_dfloat64): Likewise.
604 (RS6000_BTI_dfloat128): Likewise.
605 (long_double_type_internal_node): Likewise.
606 (dfloat64_type_internal_node): Likewise.
607 (dfloat128_type_internal_node): Likewise.
608
609 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
610 2.07 bcd arithmetic instructions.
611 (UNSPEC_BCDSUB): Likewise.
612 (UNSPEC_BCD_OVERFLOW): Likewise.
613 (UNSPEC_BCD_ADD_SUB): Likewise.
614 (bcd_add_sub): Likewise.
615 (BCD_TEST): Likewise.
616 (bcd<bcd_add_sub>): Likewise.
617 (bcd<bcd_add_sub>_test): Likewise.
618 (bcd<bcd_add_sub>_test2): Likewise.
619 (bcd<bcd_add_sub>_<code>): Likewise.
620 (peephole2 for combined bcd ops): Likewise.
621
622 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
623 decimal floating point builtin functions.
624 (UNSPEC_DENBCD): Likewise.
625 (UNSPEC_DXEX): Likewise.
626 (UNSPEC_DIEX): Likewise.
627 (UNSPEC_DSCLI): Likewise.
628 (UNSPEC_DSCRI): Likewise.
629 (D64_D128): Likewise.
630 (dfp_suffix): Likewise.
631 (dfp_ddedpd_<mode>): Likewise.
632 (dfp_denbcd_<mode>): Likewise.
633 (dfp_dxex_<mode>): Likewise.
634 (dfp_diex_<mode>): Likewise.
635 (dfp_dscli_<mode>): Likewise.
636 (dfp_dscri_<mode>): Likewise.
637
638 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
639 builtin functions.
640 (UNSPEC_CDTBCD): Likewise.
641 (UNSPEC_CBCDTD): Likewise.
642 (UNSPEC_DIVE): Add support for new extended divide builtin
643 functions.
644 (UNSPEC_DIVEO): Likewise.
645 (UNSPEC_DIVEU): Likewise.
646 (UNSPEC_DIVEUO): Likewise.
647 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
648 pack/unpack 128-bit types.
649 (UNSPEC_PACK_128BIT): Likewise.
650 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
651 (udiv<mode>3): Use idiv_ldiv mode attribute.
652 (div<mode>3): Likewise.
653 (addg6s): Add new BCD builtin functions.
654 (cdtbcd): Likewise.
655 (cbcdtd): Likewise.
656 (UNSPEC_DIV_EXTEND): Add support for new extended divide
657 instructions.
658 (div_extend): Likewise.
659 (div<div_extend>_<mode>"): Likewise.
660 (FP128_64): Add support for new builtin functions to pack/unpack
661 128-bit types.
662 (unpack<mode>): Likewise.
663 (unpacktf_0): Likewise.
664 (unpacktf_1): Likewise.
665 (unpack<mode>_dm): Likewise.
666 (unpack<mode>_nodm): Likewise.
667 (pack<mode>): Likewise.
668 (unpackv1ti): Likewise.
669 (packv1ti): Likewise.
670
671 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
672
673 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
674 is disabled.
675
676 2014-04-24 Jakub Jelinek <jakub@redhat.com>
677
678 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
679 * gimplify.c (omp_is_private): Change last argument's type to int.
680 Only diagnose lastprivate if the simd argument is 1, only diagnose
681 linear if the simd argument is 2.
682 (gimplify_omp_for): Adjust omp_is_private callers. When adding
683 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
684 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
685 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
686 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
687 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
688 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
689 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
690 * tree-nested.c (convert_nonlocal_omp_clauses,
691 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
692
693 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
694
695 PR target/60822
696 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
697 operand 1.
698
699 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
700
701 * flag-types.h (enum ivar_visibility): Add.
702
703 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
704
705 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
706 function * argument.
707
708 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
709
710 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
711
712 2014-04-24 Radovan Obradovic <robradovic@mips.com>
713 Tom de Vries <tom@codesourcery.com>
714
715 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
716 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
717 reg-note.
718 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
719 * emit-rtl.c (try_split): Same.
720
721 2014-04-24 Radovan Obradovic <robradovic@mips.com>
722 Tom de Vries <tom@codesourcery.com>
723
724 * common.opt (fuse-caller-save): New option.
725
726 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
727
728 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
729 for big-endian.
730
731 2014-04-24 Richard Biener <rguenther@suse.de>
732
733 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
734 during TER and instead use the sepops interface for expanding
735 non-GIMPLE_SINGLE_RHS.
736
737 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
738
739 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
740 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
741
742 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
743
744 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
745 assembler 64-bit option.
746 * configure: Regenerate.
747
748 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
749
750 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
751 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
752 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
753 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
754 (TARGET_CRYPTO): Take TARGET_SIMD into account.
755
756 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
757
758 * config/aarch64/aarch64-builtins.c
759 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
760 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
761 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
762 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
763 builtins.
764 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
765 (Vrevsuff): New mode attribute.
766
767 2014-04-24 Terry Guo <terry.guo@arm.com>
768
769 * config/arm/arm.h (machine_function): Define variable
770 after_arm_reorg here.
771 * config/arm/arm.c (after_arm_reorg): Remove the definition.
772 (arm_split_constant): Update the way to access variable
773 after_arm_reorg.
774 (arm_reorg): Ditto.
775 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
776
777 2014-04-23 Tom de Vries <tom@codesourcery.com>
778
779 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
780
781 2014-04-23 David Malcolm <dmalcolm@redhat.com>
782
783 * is-a.h: Update comments to reflect the following changes to the
784 "pointerness" of the API, making the template parameter match the
785 return type, allowing use of is-a.h with typedefs of pointers.
786 (is_a_helper::cast): Return a T rather then a pointer to a T, so
787 that the return type matches the parameter to the is_a_helper.
788 (as_a): Likewise.
789 (dyn_cast): Likewise.
790
791 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
792 pointer from the is-a.h API.
793
794 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
795 (is_a_helper <cgraph_node *>::test): ...this, matching change to
796 is-a.h API.
797 (is_a_helper <varpool_node>::test): Likewise, convert to...
798 (is_a_helper <varpool_node *>::test): ...this.
799
800 (varpool_first_variable): Update for removal of implicit pointer
801 from the is-a.h API.
802 (varpool_next_variable): Likewise.
803 (varpool_first_static_initializer): Likewise.
804 (varpool_next_static_initializer): Likewise.
805 (varpool_first_defined_variable): Likewise.
806 (varpool_next_defined_variable): Likewise.
807 (cgraph_first_defined_function): Likewise.
808 (cgraph_next_defined_function): Likewise.
809 (cgraph_first_function): Likewise.
810 (cgraph_next_function): Likewise.
811 (cgraph_first_function_with_gimple_body): Likewise.
812 (cgraph_next_function_with_gimple_body): Likewise.
813 (cgraph_alias_target): Likewise.
814 (varpool_alias_target): Likewise.
815 (cgraph_function_or_thunk_node): Likewise.
816 (varpool_variable_node): Likewise.
817 (symtab_real_symbol_p): Likewise.
818 * cgraphunit.c (referred_to_p): Likewise.
819 (analyze_functions): Likewise.
820 (handle_alias_pairs): Likewise.
821 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
822 * gimple-ssa.h (gimple_vuse_op): Likewise.
823 (gimple_vdef_op): Likewise.
824 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
825 * gimple.c (gimple_build_asm_1): Likewise.
826 (gimple_build_try): Likewise.
827 (gimple_build_resx): Likewise.
828 (gimple_build_eh_dispatch): Likewise.
829 (gimple_build_omp_for): Likewise.
830 (gimple_omp_for_set_clauses): Likewise.
831
832 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
833 (is_a_helper <gimple_statement_asm *>::test): ...this.
834 (is_a_helper <gimple_statement_bind>::test): Convert to...
835 (is_a_helper <gimple_statement_bind *>::test): ...this.
836 (is_a_helper <gimple_statement_call>::test): Convert to...
837 (is_a_helper <gimple_statement_call *>::test): ...this.
838 (is_a_helper <gimple_statement_catch>::test): Convert to...
839 (is_a_helper <gimple_statement_catch *>::test): ...this.
840 (is_a_helper <gimple_statement_resx>::test): Convert to...
841 (is_a_helper <gimple_statement_resx *>::test): ...this.
842 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
843 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
844 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
845 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
846 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
847 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
848 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
849 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
850 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
851 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
852 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
853 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
854 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
855 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
856 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
857 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
858 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
859 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
860 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
861 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
862 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
863 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
864 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
865 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
866 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
867 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
868 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
869 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
870 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
871 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
872 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
873 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
874 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
875 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
876 (is_a_helper <gimple_statement_phi>::test): Convert to...
877 (is_a_helper <gimple_statement_phi *>::test): ...this.
878 (is_a_helper <gimple_statement_transaction>::test): Convert to...
879 (is_a_helper <gimple_statement_transaction *>::test): ...this.
880 (is_a_helper <gimple_statement_try>::test): Convert to...
881 (is_a_helper <gimple_statement_try *>::test): ...this.
882 (is_a_helper <gimple_statement_wce>::test): Convert to...
883 (is_a_helper <gimple_statement_wce *>::test): ...this.
884 (is_a_helper <const gimple_statement_asm>::test): Convert to...
885 (is_a_helper <const gimple_statement_asm *>::test): ...this.
886 (is_a_helper <const gimple_statement_bind>::test): Convert to...
887 (is_a_helper <const gimple_statement_bind *>::test): ...this.
888 (is_a_helper <const gimple_statement_call>::test): Convert to...
889 (is_a_helper <const gimple_statement_call *>::test): ...this.
890 (is_a_helper <const gimple_statement_catch>::test): Convert to...
891 (is_a_helper <const gimple_statement_catch *>::test): ...this.
892 (is_a_helper <const gimple_statement_resx>::test): Convert to...
893 (is_a_helper <const gimple_statement_resx *>::test): ...this.
894 (is_a_helper <const gimple_statement_eh_dispatch>::test):
895 Convert to...
896 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
897 ...this.
898 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
899 to...
900 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
901 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
902 Convert to...
903 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
904 ...this.
905 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
906 Convert to...
907 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
908 ...this.
909 (is_a_helper <const gimple_statement_omp_return>::test): Convert
910 to...
911 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
912 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
913 to...
914 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
915 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
916 to...
917 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
918 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
919 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
920 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
921 to...
922 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
923 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
924 to...
925 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
926 (is_a_helper <const gimple_statement_omp_target>::test): Convert
927 to...
928 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
929 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
930 to...
931 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
932 (is_a_helper <const gimple_statement_omp_single>::test): Convert
933 to...
934 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
935 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
936 to...
937 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
938 (is_a_helper <const gimple_statement_omp_task>::test): Convert
939 to...
940 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
941 (is_a_helper <const gimple_statement_phi>::test): Convert to...
942 (is_a_helper <const gimple_statement_phi *>::test): ...this.
943 (is_a_helper <const gimple_statement_transaction>::test): Convert
944 to...
945 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
946 (is_a_helper <const gimple_statement_with_ops>::test): Convert
947 to...
948 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
949 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
950 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
951 (is_a_helper <const gimple_statement_with_memory_ops>::test):
952 Convert to...
953 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
954 ...this.
955 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
956 to...
957 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
958
959 (gimple_use_ops): Update for removal of implicit pointer from the
960 is-a.h API.
961 (gimple_set_use_ops): Likewise.
962 (gimple_vuse): Likewise.
963 (gimple_vdef): Likewise.
964 (gimple_vuse_ptr): Likewise.
965 (gimple_vdef_ptr): Likewise.
966 (gimple_set_vuse): Likewise.
967 (gimple_set_vdef): Likewise.
968 (gimple_omp_return_set_lhs): Likewise.
969 (gimple_omp_return_lhs): Likewise.
970 (gimple_omp_return_lhs_ptr): Likewise.
971 (gimple_call_fntype): Likewise.
972 (gimple_call_set_fntype): Likewise.
973 (gimple_call_set_internal_fn): Likewise.
974 (gimple_call_use_set): Likewise.
975 (gimple_call_clobber_set): Likewise.
976 (gimple_bind_vars): Likewise.
977 (gimple_bind_set_vars): Likewise.
978 (gimple_bind_body_ptr): Likewise.
979 (gimple_bind_set_body): Likewise.
980 (gimple_bind_add_stmt): Likewise.
981 (gimple_bind_block): Likewise.
982 (gimple_bind_set_block): Likewise.
983 (gimple_asm_ninputs): Likewise.
984 (gimple_asm_noutputs): Likewise.
985 (gimple_asm_nclobbers): Likewise.
986 (gimple_asm_nlabels): Likewise.
987 (gimple_asm_input_op): Likewise.
988 (gimple_asm_input_op_ptr): Likewise.
989 (gimple_asm_output_op): Likewise.
990 (gimple_asm_output_op_ptr): Likewise.
991 (gimple_asm_set_output_op): Likewise.
992 (gimple_asm_clobber_op): Likewise.
993 (gimple_asm_set_clobber_op): Likewise.
994 (gimple_asm_label_op): Likewise.
995 (gimple_asm_set_label_op): Likewise.
996 (gimple_asm_string): Likewise.
997 (gimple_catch_types): Likewise.
998 (gimple_catch_types_ptr): Likewise.
999 (gimple_catch_handler_ptr): Likewise.
1000 (gimple_catch_set_types): Likewise.
1001 (gimple_catch_set_handler): Likewise.
1002 (gimple_eh_filter_types): Likewise.
1003 (gimple_eh_filter_types_ptr): Likewise.
1004 (gimple_eh_filter_failure_ptr): Likewise.
1005 (gimple_eh_filter_set_types): Likewise.
1006 (gimple_eh_filter_set_failure): Likewise.
1007 (gimple_eh_must_not_throw_fndecl): Likewise.
1008 (gimple_eh_must_not_throw_set_fndecl): Likewise.
1009 (gimple_eh_else_n_body_ptr): Likewise.
1010 (gimple_eh_else_e_body_ptr): Likewise.
1011 (gimple_eh_else_set_n_body): Likewise.
1012 (gimple_eh_else_set_e_body): Likewise.
1013 (gimple_try_eval_ptr): Likewise.
1014 (gimple_try_cleanup_ptr): Likewise.
1015 (gimple_try_set_eval): Likewise.
1016 (gimple_try_set_cleanup): Likewise.
1017 (gimple_wce_cleanup_ptr): Likewise.
1018 (gimple_wce_set_cleanup): Likewise.
1019 (gimple_phi_capacity): Likewise.
1020 (gimple_phi_num_args): Likewise.
1021 (gimple_phi_result): Likewise.
1022 (gimple_phi_result_ptr): Likewise.
1023 (gimple_phi_set_result): Likewise.
1024 (gimple_phi_arg): Likewise.
1025 (gimple_phi_set_arg): Likewise.
1026 (gimple_resx_region): Likewise.
1027 (gimple_resx_set_region): Likewise.
1028 (gimple_eh_dispatch_region): Likewise.
1029 (gimple_eh_dispatch_set_region): Likewise.
1030 (gimple_omp_critical_name): Likewise.
1031 (gimple_omp_critical_name_ptr): Likewise.
1032 (gimple_omp_critical_set_name): Likewise.
1033 (gimple_omp_for_clauses): Likewise.
1034 (gimple_omp_for_clauses_ptr): Likewise.
1035 (gimple_omp_for_set_clauses): Likewise.
1036 (gimple_omp_for_collapse): Likewise.
1037 (gimple_omp_for_index): Likewise.
1038 (gimple_omp_for_index_ptr): Likewise.
1039 (gimple_omp_for_set_index): Likewise.
1040 (gimple_omp_for_initial): Likewise.
1041 (gimple_omp_for_initial_ptr): Likewise.
1042 (gimple_omp_for_set_initial): Likewise.
1043 (gimple_omp_for_final): Likewise.
1044 (gimple_omp_for_final_ptr): Likewise.
1045 (gimple_omp_for_set_final): Likewise.
1046 (gimple_omp_for_incr): Likewise.
1047 (gimple_omp_for_incr_ptr): Likewise.
1048 (gimple_omp_for_set_incr): Likewise.
1049 (gimple_omp_for_pre_body_ptr): Likewise.
1050 (gimple_omp_for_set_pre_body): Likewise.
1051 (gimple_omp_parallel_clauses): Likewise.
1052 (gimple_omp_parallel_clauses_ptr): Likewise.
1053 (gimple_omp_parallel_set_clauses): Likewise.
1054 (gimple_omp_parallel_child_fn): Likewise.
1055 (gimple_omp_parallel_child_fn_ptr): Likewise.
1056 (gimple_omp_parallel_set_child_fn): Likewise.
1057 (gimple_omp_parallel_data_arg): Likewise.
1058 (gimple_omp_parallel_data_arg_ptr): Likewise.
1059 (gimple_omp_parallel_set_data_arg): Likewise.
1060 (gimple_omp_task_clauses): Likewise.
1061 (gimple_omp_task_clauses_ptr): Likewise.
1062 (gimple_omp_task_set_clauses): Likewise.
1063 (gimple_omp_task_child_fn): Likewise.
1064 (gimple_omp_task_child_fn_ptr): Likewise.
1065 (gimple_omp_task_set_child_fn): Likewise.
1066 (gimple_omp_task_data_arg): Likewise.
1067 (gimple_omp_task_data_arg_ptr): Likewise.
1068 (gimple_omp_task_set_data_arg): Likewise.
1069 (gimple_omp_taskreg_clauses): Likewise.
1070 (gimple_omp_taskreg_clauses_ptr): Likewise.
1071 (gimple_omp_taskreg_set_clauses): Likewise.
1072 (gimple_omp_taskreg_child_fn): Likewise.
1073 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1074 (gimple_omp_taskreg_set_child_fn): Likewise.
1075 (gimple_omp_taskreg_data_arg): Likewise.
1076 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1077 (gimple_omp_taskreg_set_data_arg): Likewise.
1078 (gimple_omp_task_copy_fn): Likewise.
1079 (gimple_omp_task_copy_fn_ptr): Likewise.
1080 (gimple_omp_task_set_copy_fn): Likewise.
1081 (gimple_omp_task_arg_size): Likewise.
1082 (gimple_omp_task_arg_size_ptr): Likewise.
1083 (gimple_omp_task_set_arg_size): Likewise.
1084 (gimple_omp_task_arg_align): Likewise.
1085 (gimple_omp_task_arg_align_ptr): Likewise.
1086 (gimple_omp_task_set_arg_align): Likewise.
1087 (gimple_omp_single_clauses): Likewise.
1088 (gimple_omp_single_clauses_ptr): Likewise.
1089 (gimple_omp_single_set_clauses): Likewise.
1090 (gimple_omp_target_clauses): Likewise.
1091 (gimple_omp_target_clauses_ptr): Likewise.
1092 (gimple_omp_target_set_clauses): Likewise.
1093 (gimple_omp_target_child_fn): Likewise.
1094 (gimple_omp_target_child_fn_ptr): Likewise.
1095 (gimple_omp_target_set_child_fn): Likewise.
1096 (gimple_omp_target_data_arg): Likewise.
1097 (gimple_omp_target_data_arg_ptr): Likewise.
1098 (gimple_omp_target_set_data_arg): Likewise.
1099 (gimple_omp_teams_clauses): Likewise.
1100 (gimple_omp_teams_clauses_ptr): Likewise.
1101 (gimple_omp_teams_set_clauses): Likewise.
1102 (gimple_omp_sections_clauses): Likewise.
1103 (gimple_omp_sections_clauses_ptr): Likewise.
1104 (gimple_omp_sections_set_clauses): Likewise.
1105 (gimple_omp_sections_control): Likewise.
1106 (gimple_omp_sections_control_ptr): Likewise.
1107 (gimple_omp_sections_set_control): Likewise.
1108 (gimple_omp_for_set_cond): Likewise.
1109 (gimple_omp_for_cond): Likewise.
1110 (gimple_omp_atomic_store_set_val): Likewise.
1111 (gimple_omp_atomic_store_val): Likewise.
1112 (gimple_omp_atomic_store_val_ptr): Likewise.
1113 (gimple_omp_atomic_load_set_lhs): Likewise.
1114 (gimple_omp_atomic_load_lhs): Likewise.
1115 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1116 (gimple_omp_atomic_load_set_rhs): Likewise.
1117 (gimple_omp_atomic_load_rhs): Likewise.
1118 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1119 (gimple_omp_continue_control_def): Likewise.
1120 (gimple_omp_continue_control_def_ptr): Likewise.
1121 (gimple_omp_continue_set_control_def): Likewise.
1122 (gimple_omp_continue_control_use): Likewise.
1123 (gimple_omp_continue_control_use_ptr): Likewise.
1124 (gimple_omp_continue_set_control_use): Likewise.
1125 (gimple_transaction_body_ptr): Likewise.
1126 (gimple_transaction_label): Likewise.
1127 (gimple_transaction_label_ptr): Likewise.
1128 (gimple_transaction_set_body): Likewise.
1129 (gimple_transaction_set_label): Likewise.
1130
1131 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1132 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1133 * ipa-ref.c (ipa_record_reference): Likewise.
1134 * ipa-reference.c (analyze_function): Likewise.
1135 (ipa_reference_write_optimization_summary): Likewise.
1136 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1137 (address_taken_from_non_vtable_p): Likewise.
1138 (comdat_can_be_unshared_p_1): Likewise.
1139 * lto-cgraph.c (lto_output_ref): Likewise.
1140 (add_references): Likewise.
1141 (compute_ltrans_boundary): Likewise.
1142 (output_symtab): Likewise.
1143 (input_ref): Likewise.
1144 (input_cgraph_1): Likewise.
1145 (output_cgraph_opt_summary): Likewise.
1146 * lto-streamer-out.c (lto_output): Likewise.
1147 (output_symbol_p): Likewise.
1148 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1149 (lsei_start_function_in_partition): Likewise.
1150 (lsei_next_variable_in_partition): Likewise.
1151 (lsei_start_variable_in_partition): Likewise.
1152 * symtab.c (insert_to_assembler_name_hash): Likewise.
1153 (unlink_from_assembler_name_hash): Likewise.
1154 (symtab_unregister_node): Likewise.
1155 (symtab_remove_node): Likewise.
1156 (dump_symtab_node): Likewise.
1157 (verify_symtab_base): Likewise.
1158 (verify_symtab_node): Likewise.
1159 (symtab_make_decl_local): Likewise.
1160 (symtab_alias_ultimate_target): Likewise.
1161 (symtab_resolve_alias): Likewise.
1162 (symtab_get_symbol_partitioning_class): Likewise.
1163 * tree-phinodes.c (allocate_phi_node): Likewise.
1164 (reserve_phi_args_for_new_edge): Likewise.
1165 (remove_phi_args): Likewise.
1166 * varpool.c (varpool_node_for_asm): Likewise.
1167 (varpool_remove_unreferenced_decls): Likewise.
1168
1169 2014-04-23 Jeff Law <law@redhat.com>
1170
1171 PR tree-optimization/60902
1172 * tree-ssa-threadedge.c
1173 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1174 invalidate outputs from statements that do not produce useful
1175 outputs for threading.
1176
1177 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1178
1179 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1180 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1181 machine descriptions for Stack Smashing Protector.
1182
1183 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1184
1185 * aarch64.md (<optab>_rol<mode>3): New pattern.
1186 (<optab>_rolsi3_uxtw): Likewise.
1187 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1188
1189 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1190
1191 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1192 (arm_cortex_a12_tune): Likewise.
1193
1194 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1195
1196 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1197
1198 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1199
1200 * config/arm/arm.md (arm_rev16si2): New pattern.
1201 (arm_rev16si2_alt): Likewise.
1202 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1203
1204 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1205
1206 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1207 (rev16<mode>2_alt): Likewise.
1208 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1209 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1210 (aarch_rev16_shleft_mask_imm_p): Likewise.
1211 (aarch_rev16_p_1): Likewise.
1212 (aarch_rev16_p): Likewise.
1213 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1214 (aarch_rev16_shright_mask_imm_p): Likewise.
1215 (aarch_rev16_shleft_mask_imm_p): Likewise.
1216
1217 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1218
1219 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1220 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1221 rev cost.
1222 (cortex_a53_extra_costs): Likewise.
1223 (cortex_a57_extra_costs): Likewise.
1224 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1225 (cortexa7_extra_costs): Likewise.
1226 (cortexa8_extra_costs): Likewise.
1227 (cortexa12_extra_costs): Likewise.
1228 (cortexa15_extra_costs): Likewise.
1229 (v7m_extra_costs): Likewise.
1230 (arm_new_rtx_costs): Handle BSWAP.
1231
1232 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1233
1234 * config/arm/arm.c (cortexa8_extra_costs): New table.
1235 (arm_cortex_a8_tune): New tuning struct.
1236 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1237
1238 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1239
1240 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1241
1242 2014-04-23 Richard Biener <rguenther@suse.de>
1243
1244 * Makefile.in (OBJS): Remove loop-unswitch.o.
1245 * loop-unswitch.c: Delete.
1246 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1247 * passes.def (pass_rtl_unswitch): Likewise.
1248 * loop-init.c (gate_rtl_unswitch): Likewise.
1249 (rtl_unswitch): Likewise.
1250 (pass_data_rtl_unswitch): Likewise.
1251 (pass_rtl_unswitch): Likewise.
1252 (make_pass_rtl_unswitch): Likewise.
1253 * rtl.h (reversed_condition): Likewise.
1254 (compare_and_jump_seq): Likewise.
1255 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1256 and make static.
1257 * loop-unroll.c (compare_and_jump_seq): Likewise.
1258
1259 2014-04-23 Richard Biener <rguenther@suse.de>
1260
1261 PR tree-optimization/60903
1262 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1263 commented code block.
1264 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1265 loop flags to newly created BBs and edges.
1266
1267 2014-04-23 Nick Clifton <nickc@redhat.com>
1268
1269 * config/msp430/msp430.c (msp430_handle_option): Move function
1270 to msp430-common.c
1271 (msp430_option_override): Simplify mcu and mcpu option handling.
1272 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1273 support for -mhwmult command line option.
1274 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1275 -mhwmult command line option.
1276 (msp430_hwmult_enabled): Delete.
1277 (msp43o_output_labelref): Add support for -mhwmult command line
1278 option.
1279 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1280 (umulsidi3): Likewise.
1281 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1282 (mcpu, mlarge, msmall): Likewise.
1283 (mhwmult): New option.
1284 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1285 prototype.
1286 (msp430_is_f5_mcu): Remove prototype.
1287 (msp430_use_f5_series_hwmult): Add prototype.
1288 * config/msp430/msp430-opts.h: New file.
1289 * common/config/msp430: New directory.
1290 * common/config/msp430/msp430-common.c: New file.
1291 * config.gcc (msp430): Remove target_has_targetm_common.
1292 * doc/invoke.texi: Document -mhwmult command line option.
1293
1294 2014-04-23 Nick Clifton <nickc@redhat.com>
1295
1296 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1297 default-manifest.o if it can be found in the search path.
1298 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1299
1300 2014-04-23 Terry Guo <terry.guo@arm.com>
1301
1302 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1303
1304 2014-04-23 Richard Biener <rguenther@suse.de>
1305
1306 PR middle-end/60895
1307 * tree-inline.c (declare_return_variable): Use mark_addressable.
1308
1309 2014-04-23 Richard Biener <rguenther@suse.de>
1310
1311 PR middle-end/60891
1312 * loop-init.c (loop_optimizer_init): Make sure to apply
1313 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1314
1315 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1316
1317 PR sanitizer/60275
1318 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1319 New options.
1320 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1321 if flag_sanitize_undefined_trap_on_error.
1322 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1323 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1324 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1325 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1326 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1327 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1328 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1329 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1330 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1331 * ubsan.c (ubsan_instrument_unreachable): Return
1332 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1333 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1334 if flag_sanitize_undefined_trap_on_error and
1335 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1336 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1337 instrument_bool_enum_load): Emit __builtin_trap () if
1338 flag_sanitize_undefined_trap_on_error and
1339 __builtin_handle_*_abort () if !flag_sanitize_recover.
1340 * doc/invoke.texi (-fsanitize-recover,
1341 -fsanitize-undefined-trap-on-error): Document.
1342
1343 2014-04-22 Christian Bruel <christian.bruel@st.com>
1344
1345 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1346 Force immediates to SImode.
1347
1348 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1349
1350 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1351 (lroundsfsi2): New.
1352 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1353 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1354 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1355 (nios2_fpu_insn): Add entry for round.
1356 (N2FPU_NO_ERRNO_P): Define.
1357 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1358 flag_errno_math.
1359 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1360
1361 2014-04-22 Richard Henderson <rth@redhat.com>
1362
1363 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1364 (add<GPI>3_compare0): Remove leading * from name.
1365 (add<GPI>3_carryin): Likewise.
1366 (sub<GPI>3_compare0): Likewise.
1367 (sub<GPI>3_carryin): Likewise.
1368 (<su_optab>mulditi3): New expander.
1369 (multi3): New expander.
1370 (madd<GPI>): Remove leading * from name.
1371
1372 2014-04-22 Martin Jambor <mjambor@suse.cz>
1373
1374 * cgraphclones.c (cgraph_function_versioning): Copy
1375 ipa_transforms_to_apply instead of asserting it is empty.
1376
1377 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1378
1379 PR target/60868
1380 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1381 on count_exp to get mode.
1382
1383 2014-04-22 Andrew Pinski <apinski@cavium.com>
1384
1385 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1386 Handle TLS for ILP32.
1387 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1388 (tlsie_small_<mode>): this and handle PTR.
1389 (tlsie_small_sidi): New pattern.
1390 (tlsle_small): Change to an expand to handle ILP32.
1391 (tlsle_small_<mode>): New pattern.
1392 (tlsdesc_small): Rename to ...
1393 (tlsdesc_small_<mode>): this and handle PTR.
1394
1395 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1396
1397 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1398
1399 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1400
1401 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1402 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1403 (aarch64_types_signed_poly_qualifiers): Likewise.
1404 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1405 (aarch64_types_poly_signed_qualifiers): Likewise.
1406 (TYPES_REINTERP_SS): Type macro added.
1407 (TYPES_REINTERP_SU): Likewise.
1408 (TYPES_REINTERP_SP): Likewise.
1409 (TYPES_REINTERP_US): Likewise.
1410 (TYPES_REINTERP_PS): Likewise.
1411 (aarch64_fold_builtin): New expression folding added.
1412 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1413 Declarations removed.
1414 (REINTERP_SS): Declarations added.
1415 (REINTERP_US): Likewise.
1416 (REINTERP_PS): Likewise.
1417 (REINTERP_SU): Likewise.
1418 (REINTERP_SP): Likewise.
1419 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1420 (vreinterpretq_p8_f64): Likewise.
1421 (vreinterpret_p16_f64): Likewise.
1422 (vreinterpretq_p16_f64): Likewise.
1423 (vreinterpret_f32_f64): Likewise.
1424 (vreinterpretq_f32_f64): Likewise.
1425 (vreinterpret_f64_f32): Likewise.
1426 (vreinterpret_f64_p8): Likewise.
1427 (vreinterpret_f64_p16): Likewise.
1428 (vreinterpret_f64_s8): Likewise.
1429 (vreinterpret_f64_s16): Likewise.
1430 (vreinterpret_f64_s32): Likewise.
1431 (vreinterpret_f64_s64): Likewise.
1432 (vreinterpret_f64_u8): Likewise.
1433 (vreinterpret_f64_u16): Likewise.
1434 (vreinterpret_f64_u32): Likewise.
1435 (vreinterpret_f64_u64): Likewise.
1436 (vreinterpretq_f64_f32): Likewise.
1437 (vreinterpretq_f64_p8): Likewise.
1438 (vreinterpretq_f64_p16): Likewise.
1439 (vreinterpretq_f64_s8): Likewise.
1440 (vreinterpretq_f64_s16): Likewise.
1441 (vreinterpretq_f64_s32): Likewise.
1442 (vreinterpretq_f64_s64): Likewise.
1443 (vreinterpretq_f64_u8): Likewise.
1444 (vreinterpretq_f64_u16): Likewise.
1445 (vreinterpretq_f64_u32): Likewise.
1446 (vreinterpretq_f64_u64): Likewise.
1447 (vreinterpret_s64_f64): Likewise.
1448 (vreinterpretq_s64_f64): Likewise.
1449 (vreinterpret_u64_f64): Likewise.
1450 (vreinterpretq_u64_f64): Likewise.
1451 (vreinterpret_s8_f64): Likewise.
1452 (vreinterpretq_s8_f64): Likewise.
1453 (vreinterpret_s16_f64): Likewise.
1454 (vreinterpretq_s16_f64): Likewise.
1455 (vreinterpret_s32_f64): Likewise.
1456 (vreinterpretq_s32_f64): Likewise.
1457 (vreinterpret_u8_f64): Likewise.
1458 (vreinterpretq_u8_f64): Likewise.
1459 (vreinterpret_u16_f64): Likewise.
1460 (vreinterpretq_u16_f64): Likewise.
1461 (vreinterpret_u32_f64): Likewise.
1462 (vreinterpretq_u32_f64): Likewise.
1463
1464 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1465
1466 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1467 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1468 (vreinterpret_p8_s8): Likewise.
1469 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1470 (vreinterpret_p8_s16): Likewise.
1471 (vreinterpret_p8_s32): Likewise.
1472 (vreinterpret_p8_s64): Likewise.
1473 (vreinterpret_p8_f32): Likewise.
1474 (vreinterpret_p8_u8): Likewise.
1475 (vreinterpret_p8_u16): Likewise.
1476 (vreinterpret_p8_u32): Likewise.
1477 (vreinterpret_p8_u64): Likewise.
1478 (vreinterpret_p8_p16): Likewise.
1479 (vreinterpretq_p8_s8): Likewise.
1480 (vreinterpretq_p8_s16): Likewise.
1481 (vreinterpretq_p8_s32): Likewise.
1482 (vreinterpretq_p8_s64): Likewise.
1483 (vreinterpretq_p8_f32): Likewise.
1484 (vreinterpretq_p8_u8): Likewise.
1485 (vreinterpretq_p8_u16): Likewise.
1486 (vreinterpretq_p8_u32): Likewise.
1487 (vreinterpretq_p8_u64): Likewise.
1488 (vreinterpretq_p8_p16): Likewise.
1489 (vreinterpret_p16_s8): Likewise.
1490 (vreinterpret_p16_s16): Likewise.
1491 (vreinterpret_p16_s32): Likewise.
1492 (vreinterpret_p16_s64): Likewise.
1493 (vreinterpret_p16_f32): Likewise.
1494 (vreinterpret_p16_u8): Likewise.
1495 (vreinterpret_p16_u16): Likewise.
1496 (vreinterpret_p16_u32): Likewise.
1497 (vreinterpret_p16_u64): Likewise.
1498 (vreinterpret_p16_p8): Likewise.
1499 (vreinterpretq_p16_s8): Likewise.
1500 (vreinterpretq_p16_s16): Likewise.
1501 (vreinterpretq_p16_s32): Likewise.
1502 (vreinterpretq_p16_s64): Likewise.
1503 (vreinterpretq_p16_f32): Likewise.
1504 (vreinterpretq_p16_u8): Likewise.
1505 (vreinterpretq_p16_u16): Likewise.
1506 (vreinterpretq_p16_u32): Likewise.
1507 (vreinterpretq_p16_u64): Likewise.
1508 (vreinterpretq_p16_p8): Likewise.
1509 (vreinterpret_f32_s8): Likewise.
1510 (vreinterpret_f32_s16): Likewise.
1511 (vreinterpret_f32_s32): Likewise.
1512 (vreinterpret_f32_s64): Likewise.
1513 (vreinterpret_f32_u8): Likewise.
1514 (vreinterpret_f32_u16): Likewise.
1515 (vreinterpret_f32_u32): Likewise.
1516 (vreinterpret_f32_u64): Likewise.
1517 (vreinterpret_f32_p8): Likewise.
1518 (vreinterpret_f32_p16): Likewise.
1519 (vreinterpretq_f32_s8): Likewise.
1520 (vreinterpretq_f32_s16): Likewise.
1521 (vreinterpretq_f32_s32): Likewise.
1522 (vreinterpretq_f32_s64): Likewise.
1523 (vreinterpretq_f32_u8): Likewise.
1524 (vreinterpretq_f32_u16): Likewise.
1525 (vreinterpretq_f32_u32): Likewise.
1526 (vreinterpretq_f32_u64): Likewise.
1527 (vreinterpretq_f32_p8): Likewise.
1528 (vreinterpretq_f32_p16): Likewise.
1529 (vreinterpret_s64_s8): Likewise.
1530 (vreinterpret_s64_s16): Likewise.
1531 (vreinterpret_s64_s32): Likewise.
1532 (vreinterpret_s64_f32): Likewise.
1533 (vreinterpret_s64_u8): Likewise.
1534 (vreinterpret_s64_u16): Likewise.
1535 (vreinterpret_s64_u32): Likewise.
1536 (vreinterpret_s64_u64): Likewise.
1537 (vreinterpret_s64_p8): Likewise.
1538 (vreinterpret_s64_p16): Likewise.
1539 (vreinterpretq_s64_s8): Likewise.
1540 (vreinterpretq_s64_s16): Likewise.
1541 (vreinterpretq_s64_s32): Likewise.
1542 (vreinterpretq_s64_f32): Likewise.
1543 (vreinterpretq_s64_u8): Likewise.
1544 (vreinterpretq_s64_u16): Likewise.
1545 (vreinterpretq_s64_u32): Likewise.
1546 (vreinterpretq_s64_u64): Likewise.
1547 (vreinterpretq_s64_p8): Likewise.
1548 (vreinterpretq_s64_p16): Likewise.
1549 (vreinterpret_u64_s8): Likewise.
1550 (vreinterpret_u64_s16): Likewise.
1551 (vreinterpret_u64_s32): Likewise.
1552 (vreinterpret_u64_s64): Likewise.
1553 (vreinterpret_u64_f32): Likewise.
1554 (vreinterpret_u64_u8): Likewise.
1555 (vreinterpret_u64_u16): Likewise.
1556 (vreinterpret_u64_u32): Likewise.
1557 (vreinterpret_u64_p8): Likewise.
1558 (vreinterpret_u64_p16): Likewise.
1559 (vreinterpretq_u64_s8): Likewise.
1560 (vreinterpretq_u64_s16): Likewise.
1561 (vreinterpretq_u64_s32): Likewise.
1562 (vreinterpretq_u64_s64): Likewise.
1563 (vreinterpretq_u64_f32): Likewise.
1564 (vreinterpretq_u64_u8): Likewise.
1565 (vreinterpretq_u64_u16): Likewise.
1566 (vreinterpretq_u64_u32): Likewise.
1567 (vreinterpretq_u64_p8): Likewise.
1568 (vreinterpretq_u64_p16): Likewise.
1569 (vreinterpret_s8_s16): Likewise.
1570 (vreinterpret_s8_s32): Likewise.
1571 (vreinterpret_s8_s64): Likewise.
1572 (vreinterpret_s8_f32): Likewise.
1573 (vreinterpret_s8_u8): Likewise.
1574 (vreinterpret_s8_u16): Likewise.
1575 (vreinterpret_s8_u32): Likewise.
1576 (vreinterpret_s8_u64): Likewise.
1577 (vreinterpret_s8_p8): Likewise.
1578 (vreinterpret_s8_p16): Likewise.
1579 (vreinterpretq_s8_s16): Likewise.
1580 (vreinterpretq_s8_s32): Likewise.
1581 (vreinterpretq_s8_s64): Likewise.
1582 (vreinterpretq_s8_f32): Likewise.
1583 (vreinterpretq_s8_u8): Likewise.
1584 (vreinterpretq_s8_u16): Likewise.
1585 (vreinterpretq_s8_u32): Likewise.
1586 (vreinterpretq_s8_u64): Likewise.
1587 (vreinterpretq_s8_p8): Likewise.
1588 (vreinterpretq_s8_p16): Likewise.
1589 (vreinterpret_s16_s8): Likewise.
1590 (vreinterpret_s16_s32): Likewise.
1591 (vreinterpret_s16_s64): Likewise.
1592 (vreinterpret_s16_f32): Likewise.
1593 (vreinterpret_s16_u8): Likewise.
1594 (vreinterpret_s16_u16): Likewise.
1595 (vreinterpret_s16_u32): Likewise.
1596 (vreinterpret_s16_u64): Likewise.
1597 (vreinterpret_s16_p8): Likewise.
1598 (vreinterpret_s16_p16): Likewise.
1599 (vreinterpretq_s16_s8): Likewise.
1600 (vreinterpretq_s16_s32): Likewise.
1601 (vreinterpretq_s16_s64): Likewise.
1602 (vreinterpretq_s16_f32): Likewise.
1603 (vreinterpretq_s16_u8): Likewise.
1604 (vreinterpretq_s16_u16): Likewise.
1605 (vreinterpretq_s16_u32): Likewise.
1606 (vreinterpretq_s16_u64): Likewise.
1607 (vreinterpretq_s16_p8): Likewise.
1608 (vreinterpretq_s16_p16): Likewise.
1609 (vreinterpret_s32_s8): Likewise.
1610 (vreinterpret_s32_s16): Likewise.
1611 (vreinterpret_s32_s64): Likewise.
1612 (vreinterpret_s32_f32): Likewise.
1613 (vreinterpret_s32_u8): Likewise.
1614 (vreinterpret_s32_u16): Likewise.
1615 (vreinterpret_s32_u32): Likewise.
1616 (vreinterpret_s32_u64): Likewise.
1617 (vreinterpret_s32_p8): Likewise.
1618 (vreinterpret_s32_p16): Likewise.
1619 (vreinterpretq_s32_s8): Likewise.
1620 (vreinterpretq_s32_s16): Likewise.
1621 (vreinterpretq_s32_s64): Likewise.
1622 (vreinterpretq_s32_f32): Likewise.
1623 (vreinterpretq_s32_u8): Likewise.
1624 (vreinterpretq_s32_u16): Likewise.
1625 (vreinterpretq_s32_u32): Likewise.
1626 (vreinterpretq_s32_u64): Likewise.
1627 (vreinterpretq_s32_p8): Likewise.
1628 (vreinterpretq_s32_p16): Likewise.
1629 (vreinterpret_u8_s8): Likewise.
1630 (vreinterpret_u8_s16): Likewise.
1631 (vreinterpret_u8_s32): Likewise.
1632 (vreinterpret_u8_s64): Likewise.
1633 (vreinterpret_u8_f32): Likewise.
1634 (vreinterpret_u8_u16): Likewise.
1635 (vreinterpret_u8_u32): Likewise.
1636 (vreinterpret_u8_u64): Likewise.
1637 (vreinterpret_u8_p8): Likewise.
1638 (vreinterpret_u8_p16): Likewise.
1639 (vreinterpretq_u8_s8): Likewise.
1640 (vreinterpretq_u8_s16): Likewise.
1641 (vreinterpretq_u8_s32): Likewise.
1642 (vreinterpretq_u8_s64): Likewise.
1643 (vreinterpretq_u8_f32): Likewise.
1644 (vreinterpretq_u8_u16): Likewise.
1645 (vreinterpretq_u8_u32): Likewise.
1646 (vreinterpretq_u8_u64): Likewise.
1647 (vreinterpretq_u8_p8): Likewise.
1648 (vreinterpretq_u8_p16): Likewise.
1649 (vreinterpret_u16_s8): Likewise.
1650 (vreinterpret_u16_s16): Likewise.
1651 (vreinterpret_u16_s32): Likewise.
1652 (vreinterpret_u16_s64): Likewise.
1653 (vreinterpret_u16_f32): Likewise.
1654 (vreinterpret_u16_u8): Likewise.
1655 (vreinterpret_u16_u32): Likewise.
1656 (vreinterpret_u16_u64): Likewise.
1657 (vreinterpret_u16_p8): Likewise.
1658 (vreinterpret_u16_p16): Likewise.
1659 (vreinterpretq_u16_s8): Likewise.
1660 (vreinterpretq_u16_s16): Likewise.
1661 (vreinterpretq_u16_s32): Likewise.
1662 (vreinterpretq_u16_s64): Likewise.
1663 (vreinterpretq_u16_f32): Likewise.
1664 (vreinterpretq_u16_u8): Likewise.
1665 (vreinterpretq_u16_u32): Likewise.
1666 (vreinterpretq_u16_u64): Likewise.
1667 (vreinterpretq_u16_p8): Likewise.
1668 (vreinterpretq_u16_p16): Likewise.
1669 (vreinterpret_u32_s8): Likewise.
1670 (vreinterpret_u32_s16): Likewise.
1671 (vreinterpret_u32_s32): Likewise.
1672 (vreinterpret_u32_s64): Likewise.
1673 (vreinterpret_u32_f32): Likewise.
1674 (vreinterpret_u32_u8): Likewise.
1675 (vreinterpret_u32_u16): Likewise.
1676 (vreinterpret_u32_u64): Likewise.
1677 (vreinterpret_u32_p8): Likewise.
1678 (vreinterpret_u32_p16): Likewise.
1679 (vreinterpretq_u32_s8): Likewise.
1680 (vreinterpretq_u32_s16): Likewise.
1681 (vreinterpretq_u32_s32): Likewise.
1682 (vreinterpretq_u32_s64): Likewise.
1683 (vreinterpretq_u32_f32): Likewise.
1684 (vreinterpretq_u32_u8): Likewise.
1685 (vreinterpretq_u32_u16): Likewise.
1686 (vreinterpretq_u32_u64): Likewise.
1687 (vreinterpretq_u32_p8): Likewise.
1688 (vreinterpretq_u32_p16): Likewise.
1689
1690 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1691
1692 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1693 Pattern extended.
1694 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1695 extended.
1696 (sqabs): Likewise.
1697 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1698 (vqnegd_s64): Likewise.
1699 (vqabs_s64): Likewise.
1700 (vqabsd_s64): Likewise.
1701
1702 2014-04-22 Richard Henderson <rth@redhat.com>
1703
1704 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1705 computation to the top of the loop.
1706
1707 2014-04-22 Renlin <renlin.li@arm.com>
1708 Jiong Wang <jiong.wang@arm.com>
1709
1710 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1711 * config/aarch64/aarch64.c (aarch64_layout_frame)
1712 (aarch64_initial_elimination_offset): Likewise.
1713
1714 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1715
1716 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1717 Fix indentation.
1718
1719 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1720
1721 * machmode.h (bitwise_mode_for_mode): Declare.
1722 * stor-layout.h (bitwise_type_for_mode): Likewise.
1723 * stor-layout.c (bitwise_mode_for_mode): New function.
1724 (bitwise_type_for_mode): Likewise.
1725 * builtins.c (fold_builtin_memory_op): Use it instead of
1726 int_mode_for_mode and build_nonstandard_integer_type.
1727
1728 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1729
1730 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1731 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1732 (*-*-solaris2*): Simplify.
1733 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1734 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1735 *-*-solaris2.9* handling.
1736
1737 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1738 as bug.
1739 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1740 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1741 handling, simplify.
1742 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1743 * configure: Regenerate.
1744
1745 * config/i386/sol2-9.h: Remove.
1746
1747 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1748 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1749 Remove Solaris 9 references.
1750
1751 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1752
1753 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1754 (floatuns<GPI:mode><GPF:mode>2): Remove.
1755 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1756 and floatuns conversions.
1757 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1758 and floatuns conversions.
1759 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1760 (w1,w2): New mode attributes for inequal width conversions.
1761
1762 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1763
1764 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1765 the output asm format.
1766
1767 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1768
1769 * config/aarch64/aarch64-simd.md
1770 (aarch64_cm<optab>di): Always split.
1771 (*aarch64_cm<optab>di): New.
1772 (aarch64_cmtstdi): Always split.
1773 (*aarch64_cmtstdi): New.
1774
1775 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1776
1777 PR tree-optimization/60823
1778 * omp-low.c (ipa_simd_modify_function_body): Go through
1779 all SSA_NAMEs and for those refering to vector arguments
1780 which are going to be replaced adjust SSA_NAME_VAR and,
1781 if it is a default definition, change it into a non-default
1782 definition assigned at the beginning of function from new_decl.
1783 (ipa_simd_modify_stmt_ops): Rewritten.
1784 * tree-dfa.c (set_ssa_default_def): When removing default def,
1785 check for NULL loc instead of NULL *loc.
1786
1787 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1788
1789 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1790 restrictions on core registers for DImode values in Thumb2.
1791
1792 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1793
1794 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1795 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1796
1797 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1798
1799 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1800 (*iordi_notzesidi_di): Likewise.
1801 (*iordi_notsesidi_di): Likewise.
1802
1803 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1804
1805 * config/arm/arm-protos.h (tune_params): New struct members.
1806 * config/arm/arm.c: Initialise tune_params per processor.
1807 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1808 for speed, based on new tune_params.
1809
1810 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1811
1812 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1813 added.
1814 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1815 macro.
1816 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1817 corrected.
1818 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1819 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1820 (vrnda_f64): Likewise.
1821 (vrndi_f64): Likewise.
1822 (vrndm_f64): Likewise.
1823 (vrndn_f64): Likewise.
1824 (vrndp_f64): Likewise.
1825 (vrndx_f64): Likewise.
1826
1827 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1828
1829 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1830 GET_MODE_SIZE argument is enum machine_mode.
1831
1832 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1833
1834 PR target/60910
1835 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1836 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1837
1838 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1839
1840 PR middle-end/60281
1841 * asan.c (asan_emit_stack_protection): Force the base to align to
1842 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1843 appropriate bits if STRICT_ALIGNMENT.
1844 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1845 when asan is on.
1846 (expand_used_vars): Leave a space in the stack frame for alignment
1847 if STRICT_ALIGNMENT.
1848
1849 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1850
1851 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1852 than a gimple.
1853 (gimple_store_p): Likewise.
1854 (gimple_assign_load_p): Likewise.
1855 (gimple_assign_cast_p): Likewise.
1856 (gimple_clobber_p): Likewise.
1857
1858 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1859 rather than a gimple.
1860 (gimple_assign_cast_p): Likewise.
1861
1862 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1863
1864 PR target/60735
1865 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1866 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1867
1868 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1869 more debug information for E500 if -mdebug=reg.
1870
1871 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1872
1873 PR target/60909
1874 * config/i386/i386.c (ix86_expand_builtin)
1875 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1876 register for target RTX.
1877 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1878
1879 2014-04-18 Cong Hou <congh@google.com>
1880
1881 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1882 the widen-mult pattern by handling two operands with different sizes,
1883 and operands whose size is smaller than half of the result type.
1884
1885 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1886
1887 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1888 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1889 (do_estimate_edge_time): Compute it.
1890 * ipa-inline.c (want_inline_small_function_p): Bypass
1891 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1892
1893 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1894
1895 * ipa-inline.c (spec_rem): New static variable.
1896 (dump_overall_stats): New function.
1897 (dump_inline_stats): New function.
1898
1899 2014-04-18 Richard Henderson <rth@redhat.com>
1900
1901 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1902 to GET_MODE_SIZE, not a reg_class_t.
1903
1904 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1905
1906 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1907 (vsx_xxmrglw_<mode>): Likewise.
1908
1909 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1910
1911 PR target/60876
1912 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1913 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1914 (rs6000_init_hard_regno_mode_ok): Likewise.
1915
1916 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1917
1918 * ipa-inline.c (inline_small_functions): Account only non-cold
1919 functions.
1920 * doc/invoke.texi (inline-unit-growth): Update documentation.
1921
1922 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1923
1924 * config/rs6000/rs6000.md (addti3, subti3): New.
1925
1926 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1927
1928 PR target/60863
1929 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1930 comment. Check optimize_insn_for_size_p instead of
1931 optimize_insn_for_speed_p.
1932
1933 2014-04-17 Martin Jambor <mjambor@suse.cz>
1934
1935 * gimple-iterator.c (gsi_start_edge): New function.
1936 * gimple-iterator.h (gsi_start_edge): Declare.
1937 * tree-sra.c (single_non_eh_succ): New function.
1938 (disqualify_ops_if_throwing_stmt): Renamed to
1939 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1940 having one non-EH successor BB.
1941 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1942 generate loads into replacements.
1943 (sra_modify_assign): Likewise and and also use the simple path for
1944 such statements.
1945 (sra_modify_function_body): Commit statements on edges.
1946
1947 2014-04-17 Richard Biener <rguenther@suse.de>
1948
1949 PR middle-end/60849
1950 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1951 comparison results and add clarifying comment.
1952
1953 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1954
1955 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1956 (blank_mode): Initialize it.
1957 (emit_mode_size_inline, emit_mode_nunits_inline,
1958 emit_mode_inner_inline): New functions.
1959 (emit_insn_modes_h): Call them and surround their output with
1960 #if GCC_VERSION >= 4001 ... #endif.
1961 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1962 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1963 mode_* arrays if the argument is __builtin_constant_p.
1964 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1965 is enum machine_mode.
1966
1967 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1968
1969 * passes.c (opt_pass::execute): Adjust.
1970 (pass_manager::execute_pass_mode_switching): Likewise.
1971 (early_local_passes::execute): Likewise.
1972 (execute_one_pass): Pass cfun to the pass's execute method.
1973 * tree-pass.h (opt_pass::execute): Add function * argument.
1974 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1975 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1976 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1977 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1978 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1979 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1980 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1981 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1982 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1983 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1984 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1985 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1986 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1987 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1988 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1989 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1990 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1991 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1992 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1993 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1994 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1995 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1996 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1997 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1998 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1999 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2000 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2001 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2002 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2003 Adjust.
2004
2005 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2006
2007 * passes.c (opt_pass::gate): Take function * argument.
2008 (gate_all_early_local_passes): Merge into
2009 (early_local_passes::gate): this.
2010 (gate_all_early_optimizations): Merge into
2011 (all_early_optimizations::gate): this.
2012 (gate_all_optimizations): Mege into
2013 (all_optimizations::gate): this.
2014 (gate_all_optimizations_g): Merge into
2015 (all_optimizations_g::gate): this.
2016 (gate_rest_of_compilation): Mege into
2017 (rest_of_compilation::gate): this.
2018 (gate_postreload): Merge into
2019 (postreload::gate): this.
2020 (dump_one_pass): Pass cfun to the pass's gate method.
2021 (execute_ipa_summary_passes): Likewise.
2022 (execute_one_pass): Likewise.
2023 (ipa_write_summaries_2): Likewise.
2024 (ipa_write_optimization_summaries_1): Likewise.
2025 (ipa_read_summaries_1): Likewise.
2026 (ipa_read_optimization_summaries_1): Likewise.
2027 (execute_ipa_stmt_fixups): Likewise.
2028 * tree-pass.h (opt_pass::gate): Add function * argument.
2029 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2030 combine-stack-adj.c, combine.c, compare-elim.c,
2031 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2032 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2033 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2034 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2035 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2036 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2037 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2038 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2039 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2040 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2041 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2042 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2043 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2044 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2045 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2046 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2047 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2048 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2049 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2050 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2051 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2052 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2053 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2054 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2055 var-tracking.c, vtable-verify.c, web.c: Adjust.
2056
2057 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2058
2059 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2060 * configure: Regenerate.
2061
2062 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2063
2064 * passes.c (dump_one_pass): don't check pass->has_gate.
2065 (execute_ipa_summary_passes): Likewise.
2066 (execute_one_pass): Likewise.
2067 (ipa_write_summaries_2): Likewise.
2068 (ipa_write_optimization_summaries_1): Likewise.
2069 (ipa_read_optimization_summaries_1): Likewise.
2070 (execute_ipa_stmt_fixups): Likewise.
2071 * tree-pass.h (pass_data::has_gate): Remove.
2072 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2073 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2074 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2075 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2076 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2077 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2078 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2079 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2080 gimple-low.c, gimple-ssa-isolate-paths.c,
2081 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2082 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2083 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2084 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2085 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2086 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2087 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2088 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2089 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2090 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2091 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2092 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2093 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2094 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2095 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2096 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2097 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2098 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2099 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2100 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2101 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2102 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2103 Adjust.
2104
2105 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2106
2107 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2108 declaration.
2109 * passes.c (pass_manager::register_dump_files_1): Merge into
2110 (pass_manager::register_dump_files): this, and remove its handling of
2111 properties since the pass always has the properties anyway.
2112 (pass_manager::pass_manager): Adjust.
2113
2114 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2115
2116 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2117 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2118 dealing with properties.
2119 (pass_manager::register_dump_files): Adjust.
2120
2121 2014-03-20 Mark Wielaard <mjw@redhat.com>
2122
2123 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2124 then represent the bound as normal constant value.
2125
2126 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2127
2128 PR target/60847
2129 Forward port from 4.8 branch
2130 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2131
2132 * config/i386/bmiintrin.h (_blsi_u32): New.
2133 (_blsi_u64): Ditto.
2134 (_blsr_u32): Ditto.
2135 (_blsr_u64): Ditto.
2136 (_blsmsk_u32): Ditto.
2137 (_blsmsk_u64): Ditto.
2138 (_tzcnt_u32): Ditto.
2139 (_tzcnt_u64): Ditto.
2140
2141 2014-04-17 Kito Cheng <kito@0xlab.org>
2142
2143 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2144
2145 2014-04-17 Richard Biener <rguenther@suse.de>
2146
2147 PR middle-end/60849
2148 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2149 boolean results for comparisons.
2150
2151 2014-04-17 Richard Biener <rguenther@suse.de>
2152
2153 PR tree-optimization/60836
2154 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2155 initial PHI args to be gimple values.
2156
2157 2014-04-17 Richard Biener <rguenther@suse.de>
2158
2159 PR tree-optimization/60841
2160 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2161 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2162 of stmts to SLP build.
2163 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2164 (vect_analyze_slp): Likewise.
2165 (vect_analyze_slp_instance): Likewise.
2166 (vect_build_slp_tree): Limit overall SLP tree growth.
2167 * tree-vectorizer.h (vect_analyze_data_refs,
2168 vect_analyze_slp): Adjust prototypes.
2169
2170 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2171
2172 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2173 Silvermont.
2174
2175 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2176
2177 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2178 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2179 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2180 for TARGET_SLOW_PSHUFB
2181
2182 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2183
2184 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2185 * config/i386/i386.c (intel_cost): Ditto.
2186
2187 2014-04-17 Joey Ye <joey.ye@arm.com>
2188
2189 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2190
2191 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2192
2193 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2194 with -fuse-profile.
2195
2196 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2197
2198 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2199 (type_all_derivations_known_p): New predicate.
2200 (type_all_ctors_visible_p): New predicate.
2201 (type_possibly_instantiated_p): New predicate.
2202 (get_odr_type): Compute all_derivations_known.
2203 (dump_odr_type): Dump the flag.
2204 (maybe_record_type): Cleanup.
2205 (record_target_from_binfo): Add bases_to_consider array;
2206 record bases for types w/o instances and skip CXX destructor.
2207 (possible_polymorphic_call_targets_1): Add bases_to_consider
2208 and consider_construction parameters; check if type may have instance.
2209 (get_polymorphic_call_info): Set maybe_in_construction to true
2210 when we know nothing.
2211 (record_targets_from_bases): Skip CXX destructors; they are
2212 never called for types in construction.
2213 (possible_polymorphic_call_targets): Do not record target when
2214 type may not have instance.
2215
2216 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2217
2218 PR ipa/60854
2219 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2220 external aliases alive, too.
2221
2222 2014-04-16 Andrew Pinski <apinski@cavium.com>
2223
2224 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2225 definition.
2226
2227 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2228
2229 * final.c (compute_alignments): Do not apply loop alignment to a block
2230 falling through to the exit.
2231
2232 2014-04-16 Catherine Moore <clm@codesourcery.com>
2233
2234 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2235 Adjust constraints for microMIPS store patterns.
2236
2237 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2238
2239 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2240
2241 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2242
2243 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2244 (append_use): Run at -O0.
2245 (append_vdef): Likewise.
2246 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2247 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2248
2249 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2250
2251 PR tree-optimization/60844
2252 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2253 (propagate_op_to_single_use, remove_visited_stmt_chain,
2254 linearize_expr, repropagate_negates, reassociate_bb): Use it
2255 instead of gsi_remove.
2256
2257 2014-04-16 Martin Jambor <mjambor@suse.cz>
2258
2259 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2260 ipa_transforms_to_apply.
2261 (cgraph_function_versioning): Assert that old_node has empty
2262 ipa_transforms_to_apply.
2263 * trans-mem.c (ipa_tm_create_version): Likewise.
2264 * tree-inline.c (tree_function_versioning): Do not duplicate
2265 ipa_transforms_to_apply.
2266
2267 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2268
2269 PR target/60817
2270 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2271 x86_64-*-* cases.
2272 Pass necessary as flags on 64-bit Solaris/x86.
2273 Use lowercase relocs for x86_64-*-*.
2274 * configure: Regenerate.
2275
2276 2014-04-15 Jan Hubicka <jh@suse.cz>
2277
2278 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2279 (maybe_record_node, likely_target_p): Use it.
2280
2281 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2282
2283 PR target/60839
2284 Revert following patch
2285
2286 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2287
2288 PR target/60735
2289 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2290 software floating point or no floating point registers, do not
2291 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2292 in GPRs that occurs after we tested for GPRs that would never be
2293 true.
2294
2295 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2296 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2297 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2298 specifically allow DDmode, since that does not use the SPE SIMD
2299 instructions.
2300
2301 2014-03-21 Mark Wielaard <mjw@redhat.com>
2302
2303 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2304 as unsigned or int depending on type and value used.
2305
2306 2014-04-15 Richard Biener <rguenther@suse.de>
2307
2308 PR rtl-optimization/56965
2309 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2310 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2311 ... here.
2312 * alias.c (true_dependence_1): Do not call
2313 nonoverlapping_component_refs_p.
2314 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2315 nonoverlapping_component_refs_p.
2316 (indirect_refs_may_alias_p): Likewise.
2317
2318 2014-04-15 Teresa Johnson <tejohnson@google.com>
2319
2320 * cfg.c (dump_bb_info): Fix flags check.
2321 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2322
2323 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2324
2325 PR rtl-optimization/60663
2326 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2327 avoid 0 cost.
2328
2329 2014-04-15 Richard Biener <rguenther@suse.de>
2330
2331 * lto-streamer.h (LTO_major_version): Bump to 4.
2332
2333 2014-04-15 Richard Biener <rguenther@suse.de>
2334
2335 * common.opt (lto_partition_model): New enum.
2336 (flto-partition=): Merge separate options with a single with argument,
2337 add -flto-partition=one support.
2338 * flag-types.h (enum lto_partition_model): Declare.
2339 * opts.c (finish_options): Remove duplicate -flto-partition=
2340 option check.
2341 * lto-wrapper.c (run_gcc): Adjust.
2342
2343 2014-04-15 Richard Biener <rguenther@suse.de>
2344
2345 * alias.c (ncr_compar): New function.
2346 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2347
2348 2014-04-15 Richard Biener <rguenther@suse.de>
2349
2350 * alias.c (record_component_aliases): Do not walk BINFOs.
2351
2352 2014-04-15 Richard Biener <rguenther@suse.de>
2353
2354 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2355 Add struct function argument and adjust.
2356 (find_func_aliases_for_call): Likewise.
2357 (find_func_aliases): Likewise.
2358 (find_func_clobbers): Likewise.
2359 (intra_create_variable_infos): Likewise.
2360 (compute_points_to_sets): Likewise.
2361 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2362
2363 2014-04-15 Richard Biener <rguenther@suse.de>
2364
2365 * tree.c (iterative_hash_expr): Use enum tree_code_class
2366 to store TREE_CODE_CLASS.
2367 (tree_block): Likewise.
2368 (tree_set_block): Likewise.
2369 * tree.h (fold_build_pointer_plus_loc): Use
2370 convert_to_ptrofftype_loc.
2371
2372 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2373
2374 PR plugins/59335
2375 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2376 added in 4.9.
2377
2378 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2379
2380 * cfgloop.h (struct loop): Move force_vectorize down.
2381 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2382 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2383 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2384 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2385 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2386 * tree-core.h (enum annot_expr_kind): Add new kind values.
2387 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2388 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2389 kinds.
2390 * tree.def (ANNOTATE_EXPR): Tweak comment.
2391
2392 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2393
2394 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2395 cxa_pure_virtual).
2396
2397 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2398
2399 * tree.h (TYPE_IDENTIFIER): Declare.
2400 * tree.c (subrange_type_for_debug_p): Use it.
2401 * godump.c (go_format_type): Likewise.
2402 * dwarf2out.c (is_cxx_auto, modified_type_die,
2403 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2404 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2405
2406 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2407
2408 PR lto/60820
2409 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2410
2411 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2412
2413 * config/i386/i386.c (examine_argument): Return bool. Return true if
2414 parameter should be passed in memory.
2415 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2416 (construct_container): Update calls to examine_argument.
2417 (function_arg_advance_64): Ditto.
2418 (return_in_memory_32): Merge with ix86_return_in_memory.
2419 (return_in_memory_64): Ditto.
2420 (return_in_memory_ms_64): Ditto.
2421
2422 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2423
2424 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2425 * coverage.c (coverage_compute_profile_id): Handle externally visible
2426 symbols.
2427
2428 2014-04-14 Martin Jambor <mjambor@suse.cz>
2429
2430 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2431 DECL_DISREGARD_INLINE_LIMITS functions.
2432
2433 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2434
2435 PR target/60827
2436 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2437
2438 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2439
2440 PR target/60827
2441 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2442 optimize_insn_for_speed_p instead of
2443 optimize_function_for_speed_p.
2444
2445 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2446
2447 * doc/invoke.texi (free): Document AArch64.
2448
2449 2014-04-14 Richard Biener <rguenther@suse.de>
2450
2451 PR tree-optimization/60042
2452 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2453 (insert_into_preds_of_block): Do not prevent PHI insertion
2454 for REFERENCE exprs here ...
2455 (eliminate_dom_walker::before_dom_children): ... but prevent
2456 their use here under similar conditions when applied to the
2457 IL after PRE optimizations.
2458
2459 2014-04-14 Richard Biener <rguenther@suse.de>
2460
2461 * passes.def: Move early points-to after early SRA.
2462
2463 2014-04-14 Richard Biener <rguenther@suse.de>
2464
2465 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2466 check for which sign-changes we allow when forwarding
2467 a converted value into a switch.
2468
2469 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2470
2471 * stor-layout.c (place_field): Finalize non-constant offset for the
2472 field, if any.
2473
2474 2014-04-14 Richard Biener <rguenther@suse.de>
2475
2476 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2477 as argument.
2478 (expand_switch_using_bit_tests_p): Likewise.
2479 (process_switch): Compute and pass on speed_p based on the
2480 switch stmt.
2481 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2482 optimize_bb_for_speed_p.
2483
2484 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2485
2486 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2487 * function.h (struct function): Rename has_force_vect_loops into
2488 has_force_vectorize_loops.
2489 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2490 (input_struct_function_base): Likewise.
2491 * lto-streamer-out.c (output_cfg): Likewise.
2492 (output_struct_function_base): Likewise.
2493 * omp-low.c (expand_omp_simd): Likewise.
2494 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2495 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2496 (version_loop_for_if_conversion): Likewise.
2497 (tree_if_conversion): Likewise.
2498 (main_tree_if_conversion): Likewise.
2499 (gate_tree_if_conversion): Likewise.
2500 * tree-inline.c (copy_loops): Likewise.
2501 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2502 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2503 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2504 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2505 * tree-vectorizer.c (vectorize_loops): Likewise.
2506 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2507
2508 2014-04-14 Richard Biener <rguenther@suse.de>
2509
2510 PR lto/60720
2511 * lto-streamer-out.c (wrap_refs): New function.
2512 (lto_output): Wrap symbol references in global initializes in
2513 type-preserving MEM_REFs.
2514
2515 2014-04-14 Christian Bruel <christian.bruel@st.com>
2516
2517 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2518
2519 2014-04-14 Christian Bruel <christian.bruel@st.com>
2520
2521 * config/sh/sh.md (setmemqi): New expand pattern.
2522 * config/sh/sh.h (CLEAR_RATIO): Define.
2523 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2524 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2525
2526 2014-04-14 Richard Biener <rguenther@suse.de>
2527
2528 PR middle-end/55022
2529 * fold-const.c (negate_expr_p): Don't negate directional rounding
2530 division.
2531 (fold_negate_expr): Likewise.
2532
2533 2014-04-14 Richard Biener <rguenther@suse.de>
2534
2535 PR tree-optimization/59817
2536 PR tree-optimization/60453
2537 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2538 recursion to catch all CHRECs in the scalar evolution and restrict
2539 the predicate for the remains appropriately.
2540
2541 2014-04-12 Catherine Moore <clm@codesourcery.com>
2542
2543 * config/mips/constraints.md: Add new register constraint "kb".
2544 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2545 (*movhi_internal): Likewise.
2546 (*movqi_internal): Likewise.
2547 * config/mips/mips.h (M16_STORE_REGS): New register class.
2548 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2549 (REG_CLASS_CONTENTS): Likewise.
2550 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2551
2552 2014-04-11 Tobias Burnus <burnus@net-b.de>
2553
2554 PR c/60194
2555 * doc/invoke.texi (-Wformat-signedness): Document it.
2556 (Wformat=2): Mention that this enables -Wformat-signedness.
2557
2558 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2559
2560 * common/config/epiphany/epiphany-common.c
2561 (epiphany_option_optimization_table): Enable section anchors by
2562 default at -O1 or higher.
2563 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2564 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2565 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2566 carries no extra cost.
2567 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2568 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2569 * config/epiphany/predicates.md (memclob_operand): New predicate.
2570 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2571 Use memclob_operand predicate and X constraint for operand 3.
2572
2573 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2574
2575 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2576 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2577 its operands.
2578
2579 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2580
2581 PR rtl-optimization/60651
2582 * mode-switching.c (optimize_mode_switching): Make sure to emit
2583 sets of a lower numbered entity before sets of a higher numbered
2584 entity to a mode of the same or lower priority.
2585 When creating a seginfo for a basic block that starts with a code
2586 label, move the insertion point past the code label.
2587 (new_seginfo): Document and enforce requirement that
2588 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2589 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2590 * doc/tm.texi: Regenerate.
2591
2592 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2593
2594 PR target/60811
2595 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2596
2597 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2598
2599 * BASE-VER: Set to 4.10.0.
2600
2601 2014-04-11 Tobias Burnus <burnus@net-b.de>
2602
2603 PR other/59055
2604 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2605 * doc/gcc.texi (Service): Update description in the @menu
2606 * doc/invoke.texi (Option Summary): Remove misplaced and
2607 duplicated @menu.
2608
2609 2014-04-11 Steve Ellcey <sellcey@mips.com>
2610 Jakub Jelinek <jakub@redhat.com>
2611
2612 PR middle-end/60556
2613 * expr.c (convert_move): Use emit_store_flag_force instead of
2614 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2615 argument to it.
2616
2617 2014-04-11 Richard Biener <rguenther@suse.de>
2618
2619 PR middle-end/60797
2620 * varasm.c (assemble_alias): Avoid endless error reporting
2621 recursion by setting TREE_ASM_WRITTEN.
2622
2623 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2624
2625 * config/s390/s390.md: Add a splitter for NOT rtx.
2626
2627 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2628
2629 PR rtl-optimization/60663
2630 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2631
2632 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2633 Jakub Jelinek <jakub@redhat.com>
2634
2635 PR lto/60567
2636 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2637 flag from decl_node to node.
2638
2639 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2640
2641 PR debug/60655
2642 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2643 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2644 ameliorating the cases where it can be.
2645
2646 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2647
2648 Revert
2649 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2650
2651 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2652 (loadsync_<mode>): Change mode.
2653 (load_quadpti, store_quadpti): New.
2654 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2655 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2656 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2657
2658 2014-04-09 Cong Hou <congh@google.com>
2659
2660 PR testsuite/60773
2661 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2662 documentation.
2663
2664 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2665
2666 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2667 instead of vnor to exploit possible fusion opportunity in the
2668 future.
2669 (altivec_expand_vec_perm_const_le): Likewise.
2670
2671 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2672
2673 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2674 (loadsync_<mode>): Change mode.
2675 (load_quadpti, store_quadpti): New.
2676 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2677 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2678
2679 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2680
2681 PR target/60763
2682 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2683 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2684 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2685
2686 2014-04-08 Richard Biener <rguenther@suse.de>
2687
2688 PR middle-end/60706
2689 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2690 a 64bit widest int print double-int similar to on HWI64 hosts.
2691
2692 2014-04-08 Richard Biener <rguenther@suse.de>
2693
2694 PR tree-optimization/60785
2695 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2696 default defs properly.
2697
2698 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2699
2700 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2701 (Weffc++): Likewise.
2702
2703 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2704
2705 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2706 set completep to false rather than true.
2707
2708 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2709
2710 PR target/60504
2711 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2712 ARM_TARGET2_DWARF_FORMAT.
2713
2714 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2715
2716 PR target/60609
2717 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2718 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2719 ADDR_DIFF_VEC.
2720
2721 2014-04-07 Richard Biener <rguenther@suse.de>
2722
2723 PR tree-optimization/60766
2724 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2725 (may_eliminate_iv): Convert cand_value_at result to desired type.
2726
2727 2014-04-07 Jason Merrill <jason@redhat.com>
2728
2729 PR c++/60731
2730 * common.opt (-fno-gnu-unique): Add.
2731 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2732
2733 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2734
2735 * haifa-sched.c: Fix outdated function reference and minor
2736 grammar errors in introductory comment.
2737
2738 2014-04-07 Richard Biener <rguenther@suse.de>
2739
2740 PR middle-end/60750
2741 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2742 for noreturn calls.
2743 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2744
2745 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2746
2747 PR debug/55794
2748 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2749 size accounting for thunks.
2750 (pa_asm_output_mi_thunk): Use final_start_function() and
2751 final_end_function() to output function start and end directives.
2752
2753 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2754
2755 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2756 device specific ISA/ feature information. Remove short_sp and
2757 errata_skip ds. Add avr_device_specific_features enum to have device
2758 specific info.
2759 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2760 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2761 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2762 updated device specific info.
2763 * config/avr/avr-mcus.def: Merge device specific details to
2764 dev_attribute field.
2765 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2766 errata_skip.
2767 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2768 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2769 assembler if RMW isa supported by current device.
2770 * config/avr/genmultilib.awk: Update as device info structure changed.
2771 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2772
2773 2014-04-04 Cong Hou <congh@google.com>
2774
2775 PR tree-optimization/60656
2776 * tree-vect-stmts.c (supportable_widening_operation):
2777 Fix a bug that elements in a vector with vect_used_by_reduction
2778 property are incorrectly reordered when the operation on it is not
2779 consistant with the one in reduction operation.
2780
2781 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2782
2783 PR rtl-optimization/60155
2784 * gcse.c (record_set_data): New function.
2785 (single_set_gcse): New function.
2786 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2787 (hoist_code): Likewise.
2788 (get_pressure_class_and_nregs): Likewise.
2789
2790 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2791
2792 * explow.c (probe_stack_range): Emit a final optimization blockage.
2793
2794 2014-04-04 Anthony Green <green@moxielogic.com>
2795
2796 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2797 typos.
2798
2799 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2800
2801 PR ipa/59626
2802 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2803 flags are set only during streaming.
2804 * ipa.c (process_references, walk_polymorphic_call_targets,
2805 symtab_remove_unreachable_nodes): Drop bodies of always inline
2806 after early inlining.
2807 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2808
2809 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2810 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2811
2812 PR debug/60655
2813 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2814 containing a NOT.
2815
2816 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2817
2818 PR bootstrap/60743
2819 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2820 duration.
2821 (cortex_a53_fdivd): Likewise.
2822
2823 2014-04-04 Martin Jambor <mjambor@suse.cz>
2824
2825 PR ipa/60640
2826 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2827 Adjust all callers.
2828 * cgraph.c (clone_of_p): Also return true if thunks match.
2829 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2830 cgraph_function_or_thunk_node and an obsolete comment.
2831 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2832 file.
2833 (build_function_decl_skip_args): Likewise.
2834 (set_new_clone_decl_and_node_flags): New function.
2835 (duplicate_thunk_for_node): Likewise.
2836 (redirect_edge_duplicating_thunks): Likewise.
2837 (cgraph_clone_node): New parameter args_to_skip, pass it to
2838 redirect_edge_duplicating_thunks which is called instead of
2839 cgraph_redirect_edge_callee.
2840 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2841 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2842
2843 2014-04-04 Jeff Law <law@redhat.com>
2844
2845 PR target/60657
2846 * config/arm/predicates.md (const_int_I_operand): New predicate.
2847 (const_int_M_operand): Similarly.
2848 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2849 const_int_operand.
2850 (insv_t2, extv_reg, extzv_t2): Likewise.
2851 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2852 (pop_multiple_with_writeback_and_return): Likewise.
2853 (vfp_pop_multiple_with_writeback): Likewise
2854
2855 2014-04-04 Richard Biener <rguenther@suse.de>
2856
2857 PR ipa/60746
2858 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2859 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2860 non-GIMPLE_LABELs.
2861 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2862 * gimplify.c (gimple_add_tmp_var_fn): New function.
2863 * gimple-expr.h (create_tmp_reg_fn): Declare.
2864 * gimple-expr.c (create_tmp_reg_fn): New function.
2865 * gimple-low.c (record_vars_into): Don't change cfun.
2866 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2867 code generation without cfun.
2868
2869 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2870
2871 PR bootstrap/60719
2872 * Makefile.in (install-driver): Fix shell scripting.
2873
2874 2014-04-03 Cong Hou <congh@google.com>
2875
2876 PR tree-optimization/60505
2877 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2878 threshold of number of iterations below which no vectorization
2879 will be done.
2880 * tree-vect-loop.c (new_loop_vec_info):
2881 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2882 * tree-vect-loop.c (vect_analyze_loop_operations):
2883 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2884 * tree-vect-loop.c (vect_transform_loop):
2885 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2886 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2887 of iterations of the loop and see if we should build the epilogue.
2888
2889 2014-04-03 Richard Biener <rguenther@suse.de>
2890
2891 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2892 (streamer_tree_cache_create): Adjust.
2893 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2894 to allow optional nodes array.
2895 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2896 (streamer_tree_cache_append): Likewise.
2897 (streamer_tree_cache_create): Create nodes array optionally
2898 as specified by parameter.
2899 * lto-streamer-out.c (create_output_block): Avoid maintaining
2900 the node array in the writer cache.
2901 (DFS_write_tree): Remove assertion.
2902 (produce_asm_for_decls): Free the out decl state hash table early.
2903 * lto-streamer-in.c (lto_data_in_create): Adjust for
2904 streamer_tree_cache_create prototype change.
2905
2906 2014-04-03 Richard Biener <rguenther@suse.de>
2907
2908 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2909 set TREE_CHAIN to NULL_TREE.
2910
2911 2014-04-03 Richard Biener <rguenther@suse.de>
2912
2913 PR tree-optimization/60740
2914 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2915 over all GIMPLE_COND operands.
2916
2917 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2918
2919 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2920 (Weffc++): Remove Scott's numbering, merge lists and reference
2921 Wnon-virtual-dtor.
2922
2923 2014-04-03 Nick Clifton <nickc@redhat.com>
2924
2925 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2926 properly.
2927
2928 2014-04-03 Martin Jambor <mjambor@suse.cz>
2929
2930 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2931 mention gcc_unreachable before failing.
2932 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2933 removed symbols.
2934
2935 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2936
2937 PR ipa/60659
2938 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2939 inconsistent code and instead mark the context inconsistent.
2940 (possible_polymorphic_call_targets): For inconsistent contexts
2941 return empty complete list.
2942
2943 2014-04-02 Anthony Green <green@moxielogic.com>
2944
2945 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2946 (extendqisi2, extendhisi2): Define.
2947 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2948 (WCHAR_TYPE): Change to unsigned int.
2949
2950 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2951
2952 PR tree-optimization/60733
2953 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2954 insertion point for PHI candidates to be the end of the feeding
2955 block for the PHI argument.
2956
2957 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2958
2959 PR rtl-optimization/60650
2960 * lra-constraints.c (process_alt_operands): Decrease reject for
2961 earlyclobber matching.
2962
2963 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2964
2965 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2966
2967 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2968
2969 * config/spu/spu.c (pad_bb): Do not crash when the last
2970 insn is CODE_FOR_blockage.
2971
2972 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2973
2974 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2975 lies outside the target mode.
2976
2977 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2978
2979 PR target/60735
2980 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2981 software floating point or no floating point registers, do not
2982 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2983 in GPRs that occurs after we tested for GPRs that would never be
2984 true.
2985
2986 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2987 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2988 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2989 specifically allow DDmode, since that does not use the SPE SIMD
2990 instructions.
2991
2992 2014-04-02 Richard Biener <rguenther@suse.de>
2993
2994 PR middle-end/60729
2995 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2996 MODE_INTs. Properly use negv_optab.
2997 (expand_abs): Likewise.
2998
2999 2014-04-02 Richard Biener <rguenther@suse.de>
3000
3001 PR bootstrap/60719
3002 * Makefile.in (install-driver): Guard extra installs with special
3003 names properly.
3004
3005 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
3006
3007 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3008 Document vec_vgbbd.
3009
3010 2014-04-01 Richard Henderson <rth@redhat.com>
3011
3012 PR target/60704
3013 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
3014 alternative enabled before register allocation.
3015
3016 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3017
3018 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3019 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3020 typo.
3021 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3022 (nios2_got_address): Update nios2_large_got_address call site.
3023 (nios2_delegitimize_address): New function.
3024 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3025 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3026 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3027
3028 2014-04-01 Martin Husemann <martin@duskware.de>
3029
3030 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3031 for -mabi=32.
3032
3033 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3034
3035 PR rtl-optimization/60604
3036 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3037 check from register_operand.
3038 (register_operand): Redefine in terms of general_operand.
3039 (nonmemory_operand): Use register_operand for the non-constant cases.
3040
3041 2014-04-01 Richard Biener <rguenther@suse.de>
3042
3043 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3044
3045 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3046
3047 * doc/invoke.texi (mapp-regs): Clarify.
3048
3049 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3050
3051 * config/i386/avx512fintrin.h (__v32hi): Define type.
3052 (__v64qi): Likewise.
3053 (_mm512_set1_epi8): Define.
3054 (_mm512_set1_epi16): Define.
3055 (_mm512_set4_epi32): Define.
3056 (_mm512_set4_epi64): Define.
3057 (_mm512_set4_pd): Define.
3058 (_mm512_set4_ps): Define.
3059 (_mm512_setr4_epi64): Define.
3060 (_mm512_setr4_epi32): Define.
3061 (_mm512_setr4_pd): Define.
3062 (_mm512_setr4_ps): Define.
3063 (_mm512_setzero_epi32): Define.
3064
3065 2014-03-31 Martin Jambor <mjambor@suse.cz>
3066
3067 PR middle-end/60647
3068 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3069 callsite_arguments_match_p. Updated all callers. Also check types of
3070 corresponding formal parameters and actual arguments.
3071 (not_all_callers_have_enough_arguments_p) Renamed to
3072 some_callers_have_mismatched_arguments_p.
3073
3074 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3075
3076 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3077
3078 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3079
3080 PR target/60034
3081 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3082 section anchor.
3083
3084 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3085
3086 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3087 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3088 Split out
3089 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3090 Use FMAMODE_NOVF512 mode iterator.
3091 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3092 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3093 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3094 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3095 Split out
3096 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3097 Use VF_128_256 mode iterator.
3098 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3099 Ditto.
3100
3101 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3102
3103 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3104 static chain if needed.
3105
3106 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3107
3108 PR target/60697
3109 * lra-constraints.c (index_part_to_reg): New.
3110 (process_address): Use it.
3111
3112 2014-03-27 Jeff Law <law@redhat.com>
3113 Jakub Jelinek <jakub@redhat.com>
3114
3115 PR target/60648
3116 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3117 gen_rtx_{PLUS,MULT} to build up the address expression.
3118
3119 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3120 creating non-canonical RTL.
3121
3122 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3123
3124 PR ipa/60243
3125 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3126 functions; reorganize to make cheap checks first.
3127 (inline_small_functions): Do not estimate growth when dumping;
3128 it is expensive.
3129 * ipa-inline.h (inline_summary): Add min_size.
3130 (growth_likely_positive): New function.
3131 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3132 (set_cond_stmt_execution_predicate): Cleanup.
3133 (estimate_edge_size_and_time): Compute min_size.
3134 (estimate_calls_size_and_time): Likewise.
3135 (estimate_node_size_and_time): Likewise.
3136 (inline_update_overall_summary): Update min_size.
3137 (do_estimate_edge_time): Likewise.
3138 (do_estimate_edge_size): Update.
3139 (do_estimate_edge_hints): Update.
3140 (growth_likely_positive): New function.
3141
3142 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3143
3144 PR target/60693
3145 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3146 also if addr has VOIDmode.
3147
3148 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3149
3150 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3151 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3152 Declare extern.
3153 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3154 instructions as well as AdvancedSIMD loads.
3155
3156 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3157
3158 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3159 Use crypto_aese type.
3160 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3161 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3162 crypto_aese, crypto_aesmc. Move to types.md.
3163 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3164 crypto_aesmc.
3165 * config/arm/iterators.md (crypto_type): Likewise.
3166
3167 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3168
3169 * cgraph.c: Include expr.h and tree-dfa.h.
3170 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3171 remove LHS.
3172
3173 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3174
3175 PR target/60675
3176 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3177 regs from checking multi-reg pseudos.
3178
3179 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3180
3181 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3182
3183 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3184
3185 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3186 if it would clobber the stack pointer, even temporarily.
3187
3188 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3189
3190 * mode-switching.c: Make small adjustments to the top comment.
3191
3192 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3193
3194 * config/rs6000/constraints.md (wD constraint): New constraint to
3195 match the constant integer to get the top DImode/DFmode out of a
3196 vector in a VSX register.
3197
3198 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3199 match the constant integer to get the top DImode/DFmode out of a
3200 vector in a VSX register.
3201
3202 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3203 for ISA 2.07.
3204
3205 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3206 vbpermq builtins.
3207
3208 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3209 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3210
3211 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3212 Optimize vec_extract of 64-bit values, where the value being
3213 extracted is in the top word, where we can use scalar
3214 instructions. Add direct move and store support. Combine the big
3215 endian/little endian vector select load support into a single insn.
3216 (vsx_extract_<mode>_internal1): Likewise.
3217 (vsx_extract_<mode>_internal2): Likewise.
3218 (vsx_extract_<mode>_load): Likewise.
3219 (vsx_extract_<mode>_store): Likewise.
3220 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3221 combined into vsx_extract_<mode>_load.
3222 (vsx_extract_<mode>_one_le): Likewise.
3223
3224 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3225 define the top 64-bit vector element.
3226
3227 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3228 constraint.
3229
3230 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3231 Document vec_vbpermq builtin.
3232
3233 PR target/60672
3234 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3235 enable use of xxsldwi and xxpermdi builtin functions.
3236 (vec_xxpermdi): Likewise.
3237
3238 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3239 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3240
3241 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3242
3243 PR rtl-optimization/60650
3244 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3245 first_p. Use it.
3246 (find_spills_for): New.
3247 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3248 Spill all pseudos on the second iteration.
3249
3250 2014-03-27 Marek Polacek <polacek@redhat.com>
3251
3252 PR c/50347
3253 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3254 types.
3255
3256 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3257
3258 * config/s390/s390.c (s390_can_use_return_insn): Check for
3259 call-saved FPRs on 31 bit.
3260
3261 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3262
3263 PR middle-end/60682
3264 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3265 if they need regimplification, just drop them instead of
3266 calling gimple_regimplify_operands on them.
3267
3268 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3269
3270 PR target/60580
3271 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3272 (aarch64_frame_pointer_required): Adjust logic.
3273 (aarch64_can_eliminate): Adjust logic.
3274 (aarch64_override_options_after_change): Adjust logic.
3275
3276 2014-03-27 Dehao Chen <dehao@google.com>
3277
3278 * ipa-inline.c (early_inliner): Update node's inline info.
3279
3280 2014-03-26 Dehao Chen <dehao@google.com>
3281
3282 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3283 compiler inserted conditional jumps for NAN float check.
3284
3285 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3286
3287 * ubsan.h (ubsan_create_data): Change second argument's type
3288 to const location_t *.
3289 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3290 _("<unknown>").
3291 (ubsan_create_data): Change second argument to const location_t *PLOC.
3292 Create Loc field whenever PLOC is non-NULL.
3293 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3294 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3295 callers.
3296
3297 PR other/59545
3298 * real.c (real_to_integer2): Change type of low to UHWI.
3299
3300 2014-03-26 Tobias Burnus <burnus@net-b.de>
3301
3302 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3303 (CILK_SELF_SPECS): New define.
3304 (driver_self_specs): Use it.
3305
3306 2014-03-26 Richard Biener <rguenther@suse.de>
3307
3308 * tree-pretty-print.c (percent_K_format): Implement special
3309 case for LTO and its stripped down BLOCK tree.
3310
3311 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3312
3313 PR sanitizer/60636
3314 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3315
3316 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3317 one range is range_int_cst_p, but not both, at least optimize
3318 addition/subtraction of 0 and multiplication by 0 or 1.
3319 * gimple-fold.c (gimple_fold_call): Fold
3320 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3321 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3322 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3323
3324 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3325
3326 PR rtl-optimization/60452
3327 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3328 <case REG>: Return 1 for invalid offsets from the frame pointer.
3329
3330 2014-03-26 Marek Polacek <polacek@redhat.com>
3331
3332 PR c/37428
3333 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3334 a structure/union.
3335
3336 2014-03-26 Marek Polacek <polacek@redhat.com>
3337
3338 PR c/39525
3339 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3340 field members.
3341
3342 2014-03-26 Marek Polacek <polacek@redhat.com>
3343
3344 PR other/59545
3345 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3346 multiplication in unsigned type.
3347
3348 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3349
3350 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3351
3352 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3353
3354 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3355
3356 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3357
3358 PR ipa/60315
3359 * cif-code.def (UNREACHABLE) New code.
3360 * ipa-inline.c (inline_small_functions): Skip edges to
3361 __builtlin_unreachable.
3362 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3363 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3364 predicate to __bulitin_unreachable.
3365 (set_cond_stmt_execution_predicate): Fix issue when
3366 invert_tree_comparison returns ERROR_MARK.
3367 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3368 propagate to inline clones.
3369 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3370 to unreachable.
3371 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3372 * cgraphclones.c (cgraph_clone_node): If call destination is already
3373 ureachable, do not redirect it back.
3374 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3375 unreachable.
3376
3377 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3378
3379 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3380 Do not modify inline clones.
3381
3382 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3383
3384 * config/i386/i386.md (general_sext_operand): New mode attr.
3385 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3386 don't generate (sign_extend (const_int)).
3387 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3388 operands[2]. Use We constraint instead of <i> and
3389 <general_sext_operand> predicate instead of <general_operand>.
3390 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3391 * config/i386/constraints.md (We): New constraint.
3392 * config/i386/predicates.md (x86_64_sext_operand,
3393 sext_operand): New predicates.
3394
3395 2014-03-25 Martin Jambor <mjambor@suse.cz>
3396
3397 PR ipa/60600
3398 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3399 inconsistent devirtualizations to __builtin_unreachable.
3400
3401 2014-03-25 Marek Polacek <polacek@redhat.com>
3402
3403 PR c/35449
3404 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3405
3406 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3407
3408 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3409 order of elements for big-endian.
3410
3411 2014-03-25 Richard Biener <rguenther@suse.de>
3412
3413 PR middle-end/60635
3414 * gimplify-me.c (gimple_regimplify_operands): Update the
3415 re-gimplifed stmt.
3416
3417 2014-03-25 Martin Jambor <mjambor@suse.cz>
3418
3419 PR ipa/59176
3420 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3421 (lto_output_varpool_node): Likewise.
3422 (input_overwrite_node): Likewise.
3423 (input_varpool_node): Likewise.
3424
3425 2014-03-25 Richard Biener <rguenther@suse.de>
3426
3427 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3428 (run_gcc): Likewise.
3429
3430 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3431
3432 * combine.c (simplify_compare_const): Add MODE argument.
3433 Handle mode_width 0 as very large mode_width.
3434 (try_combine, simplify_comparison): Adjust callers.
3435
3436 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3437 type to avoid signed integer overflow.
3438 * explow.c (plus_constant): Likewise.
3439
3440 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3441
3442 * doc/generic.texi: Correct typos.
3443
3444 2014-03-24 Tobias Burnus <burnus@net-b.de>
3445
3446 * doc/invoke.texi (-flto): Expand section about
3447 using static libraries with LTO.
3448
3449 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3450
3451 PR rtl-optimization/60501
3452 * optabs.def (addptr3_optab): New optab.
3453 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3454 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3455 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3456
3457 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3458
3459 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3460
3461 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3462
3463 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3464 _mm512_set1_pd.
3465
3466 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3467 (_mm256_undefined_ps): Define.
3468 (_mm256_undefined_pd): Define.
3469 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3470 (_mm_undefined_pd): Define.
3471 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3472 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3473 (_mm512_undefined_ps): Define.
3474 (_mm512_undefined_pd): Define.
3475 Use _mm*_undefined_*.
3476 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3477
3478 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3479
3480 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3481 (lshr_simd): DI mode added.
3482 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3483 (aarch64_ushr_simddi): Likewise.
3484 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3485 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3486 (vshrd_n_u64): Likewise.
3487
3488 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3489
3490 * Makefile.in (s-macro_list): Depend on cc1.
3491
3492 2014-03-23 Teresa Johnson <tejohnson@google.com>
3493
3494 * ipa-utils.c (ipa_print_order): Use specified dump file.
3495
3496 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3497
3498 PR rtl-optimization/60601
3499 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3500
3501 * gcc.c (eval_spec_function): Initialize save_growing_value.
3502
3503 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3504
3505 PR sanitizer/60613
3506 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3507 code == MINUS_EXPR, never swap op0 with op1.
3508
3509 * toplev.c (init_local_tick): Avoid signed integer multiplication
3510 overflow.
3511 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3512 shift by first operand's bitsize.
3513
3514 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3515
3516 PR target/60610
3517 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3518 redefine to 1 or 0.
3519 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3520 TARGET_ISA_64BIT_P(x).
3521
3522 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3523
3524 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3525 pattern for vector nor instead of subtract from splat(-1).
3526 (altivec_expand_vec_perm_const_le): Likewise.
3527
3528 2014-03-21 Richard Henderson <rth@twiddle.net>
3529
3530 PR target/60598
3531 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3532 related insns after epilogue_completed.
3533
3534 2014-03-21 Martin Jambor <mjambor@suse.cz>
3535
3536 PR ipa/59176
3537 * cgraph.h (symtab_node): New flag body_removed.
3538 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3539 when removing bodies.
3540 * symtab.c (dump_symtab_base): Dump body_removed flag.
3541 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3542 had their bodies removed.
3543
3544 2014-03-21 Martin Jambor <mjambor@suse.cz>
3545
3546 PR ipa/60419
3547 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3548 in the border.
3549
3550 2014-03-21 Richard Biener <rguenther@suse.de>
3551
3552 PR tree-optimization/60577
3553 * tree-core.h (struct tree_base): Document nothrow_flag use
3554 in VAR_DECL_NONALIASED.
3555 * tree.h (VAR_DECL_NONALIASED): New.
3556 (may_be_aliased): Adjust.
3557 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3558
3559 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3560
3561 * expr.c (expand_expr_real_1): Remove outdated comment.
3562
3563 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3564
3565 PR middle-end/60597
3566 * ira.c (adjust_cleared_regs): Call copy_rtx on
3567 *reg_equiv[REGNO (loc)].src_p before passing it to
3568 simplify_replace_fn_rtx.
3569
3570 PR target/60568
3571 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3572 into CONST, put pic register as first operand of PLUS. Use
3573 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3574
3575 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3576
3577 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3578
3579 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3580
3581 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3582 around for store forwarding issue in the FPU on the UT699.
3583 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3584 loads and operations if -mfix-ut699 is specified.
3585 (divtf3_hq): Tweak attribute.
3586 (sqrttf2_hq): Likewise.
3587
3588 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3589
3590 * calls.c (store_one_arg): Remove incorrect const qualification on the
3591 type of the temporary.
3592 * cfgexpand.c (expand_return): Likewise.
3593 * expr.c (expand_constructor): Likewise.
3594 (expand_expr_real_1): Likewise.
3595
3596 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3597
3598 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3599 of parts.
3600
3601 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3602
3603 PR target/60039
3604 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3605
3606 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3607
3608 * config/arm/aarch-common-protos.h
3609 (alu_cost_table): Fix spelling of "extend".
3610 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3611
3612 2014-03-19 Richard Biener <rguenther@suse.de>
3613
3614 PR middle-end/60553
3615 * tree-core.h (tree_type_common): Re-order pointer members
3616 to reduce recursion depth during GC walks.
3617
3618 2014-03-19 Marek Polacek <polacek@redhat.com>
3619
3620 PR sanitizer/60569
3621 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3622 before accessing it.
3623
3624 2014-03-19 Richard Biener <rguenther@suse.de>
3625
3626 PR lto/59543
3627 * lto-streamer-in.c (input_function): In WPA stage do not drop
3628 debug stmts.
3629
3630 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3631
3632 PR tree-optimization/60559
3633 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3634 with build_zero_cst assignment.
3635
3636 2014-03-18 Kai Tietz <ktietz@redhat.com>
3637
3638 PR rtl-optimization/56356
3639 * sdbout.c (sdbout_parms): Verify that parms'
3640 incoming argument is valid.
3641 (sdbout_reg_parms): Likewise.
3642
3643 2014-03-18 Richard Henderson <rth@redhat.com>
3644
3645 PR target/60562
3646 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3647 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3648 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3649
3650 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3651
3652 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3653 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3654 Italicize plugin event names in description. Explain that
3655 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3656 Remind that no GCC functions should be called after PLUGIN_FINISH.
3657 Explain what pragmas with expansion are.
3658
3659 2014-03-18 Martin Liska <mliska@suse.cz>
3660
3661 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3662 gimple call statement is update.
3663 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3664 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3665
3666 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3667
3668 PR sanitizer/60557
3669 * ubsan.c (ubsan_instrument_unreachable): Call
3670 initialize_sanitizer_builtins.
3671 (ubsan_pass): Likewise.
3672
3673 PR sanitizer/60535
3674 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3675 varpool_finalize_decl instead of rest_of_decl_compilation.
3676
3677 2014-03-18 Richard Biener <rguenther@suse.de>
3678
3679 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3680 by using bitmap_and_compl instead of bitmap_and_compl_into.
3681 (df_rd_transfer_function): Likewise.
3682
3683 2014-03-18 Richard Biener <rguenther@suse.de>
3684
3685 * doc/lto.texi (fresolution): Fix typo.
3686
3687 2014-03-18 Richard Biener <rguenther@suse.de>
3688
3689 * doc/invoke.texi (flto): Update for changes in 4.9.
3690
3691 2014-03-18 Richard Biener <rguenther@suse.de>
3692
3693 * doc/loop.texi: Remove section on the removed lambda framework.
3694 Update loop docs with recent changes in preserving loop structure.
3695
3696 2014-03-18 Richard Biener <rguenther@suse.de>
3697
3698 * doc/lto.texi (-fresolution): Document.
3699
3700 2014-03-18 Richard Biener <rguenther@suse.de>
3701
3702 * doc/contrib.texi: Adjust my name.
3703
3704 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3705
3706 PR ipa/58721
3707 * internal-fn.c: Include diagnostic-core.h.
3708 (expand_BUILTIN_EXPECT): New function.
3709 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3710 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3711 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3712 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3713 IFN_BUILTIN_EXPECT.
3714 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3715 Revert 3 argument __builtin_expect code.
3716 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3717 * gimple-fold.c (gimple_fold_call): Likewise.
3718 * tree.h (fold_builtin_expect): New prototype.
3719 * builtins.c (build_builtin_expect_predicate): Add predictor
3720 argument, if non-NULL, create 3 argument __builtin_expect.
3721 (fold_builtin_expect): No longer static. Add ARG2 argument,
3722 pass it through to build_builtin_expect_predicate.
3723 (fold_builtin_2): Adjust caller.
3724 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3725 * internal-fn.def (BUILTIN_EXPECT): New.
3726
3727 2014-03-18 Tobias Burnus <burnus@net-b.de>
3728
3729 PR ipa/58721
3730 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3731 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3732 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3733
3734 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3735
3736 PR ipa/58721
3737 * predict.c (combine_predictions_for_bb): Fix up formatting.
3738 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3739 fill what it points to if non-NULL.
3740 (tree_predict_by_opcode): Adjust caller, use the predictor.
3741 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3742
3743 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3744
3745 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3746 proper constant for the store mode.
3747
3748 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3749
3750 * symtab.c (change_decl_assembler_name): Fix transparent alias
3751 chain construction.
3752
3753 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3754
3755 * config/aarch64/aarch64.c: Correct the comments about the
3756 aarch64 stack layout.
3757
3758 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3759
3760 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3761 check for GF_OMP_FOR_KIND_FOR.
3762
3763 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3764
3765 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3766 ymm and zmm register names.
3767
3768 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3769
3770 PR target/60516
3771 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3772 note creation for the 2010-08-31 changes.
3773
3774 2014-03-17 Marek Polacek <polacek@redhat.com>
3775
3776 PR middle-end/60534
3777 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3778 as -fno-tree-loop-vectorize.
3779 (expand_omp_simd): Likewise.
3780
3781 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3782
3783 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3784 (eligible_for_call_delay): New prototype.
3785 * config/sparc/sparc.c (tls_call_delay): Rename into...
3786 (eligible_for_call_delay): ...this. Return false if the instruction
3787 cannot be put in the delay slot of a branch.
3788 (eligible_for_restore_insn): Simplify.
3789 (eligible_for_return_delay): Return false if the instruction cannot be
3790 put in the delay slot of a branch and simplify.
3791 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3792 put in the delay slot of a branch.
3793 * config/sparc/sparc.md (fix_ut699): New attribute.
3794 (tls_call_delay): Delete.
3795 (in_call_delay): Reimplement.
3796 (eligible_for_sibcall_delay): Rename into...
3797 (in_sibcall_delay): ...this.
3798 (eligible_for_return_delay): Rename into...
3799 (in_return_delay): ...this.
3800 (in_branch_delay): Reimplement.
3801 (in_uncond_branch_delay): Delete.
3802 (in_annul_branch_delay): Delete.
3803
3804 2014-03-14 Richard Henderson <rth@redhat.com>
3805
3806 PR target/60525
3807 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3808 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3809 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3810 (floathi splitters): Remove.
3811 (float<SWI48x>xf2): New pattern.
3812 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3813 code that tried to handle DImode for 32-bit, but which was excluded
3814 by the pattern's condition. Drop allocation of stack temporary.
3815 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3816 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3817 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3818 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3819 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3820 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3821 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3822 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3823 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3824 (*float<SWI48x><X87MODEF>2_i387): Remove.
3825 (all float _with_temp splitters): Remove.
3826 (*float<SWI48x><MODEF>2_i387): New pattern.
3827 (*float<SWI48><MODEF>2_sse): New pattern.
3828 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3829 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3830
3831 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3832 Marek Polacek <polacek@redhat.com>
3833
3834 PR middle-end/60484
3835 * common.opt (dump_base_name_prefixed): New Variable.
3836 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3837 if x_dump_base_name_prefixed is already set, set it at the end.
3838
3839 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3840
3841 PR rtl-optimization/60508
3842 * lra-constraints.c (get_reload_reg): Add new parameter
3843 in_subreg_p.
3844 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3845 Pass the new parameter values.
3846
3847 2014-03-14 Richard Biener <rguenther@suse.de>
3848
3849 * common.opt: Revert unintented changes from r205065.
3850 * opts.c: Likewise.
3851
3852 2014-03-14 Richard Biener <rguenther@suse.de>
3853
3854 PR middle-end/60518
3855 * cfghooks.c (split_block): Properly adjust all loops the
3856 block was a latch of.
3857
3858 2014-03-14 Martin Jambor <mjambor@suse.cz>
3859
3860 PR lto/60461
3861 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3862 and simplify it.
3863
3864 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3865
3866 PR target/59396
3867 * config/avr/avr.c (avr_set_current_function): Pass function name
3868 through default_strip_name_encoding before sanity checking instead
3869 of skipping the first char of the assembler name.
3870
3871 2014-03-13 Richard Henderson <rth@redhat.com>
3872
3873 PR debug/60438
3874 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3875 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3876 * config/i386/i386-protos.h: Likewise.
3877 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3878 in the expander instead of a splitter.
3879 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3880 any possibility of requiring a memory.
3881 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3882 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3883 (fp branch splitters): Update for ix86_split_fp_branch.
3884 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3885 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3886 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3887 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3888 (*fop_<MODEF>_3_i387): Likewise.
3889 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3890 (splitters for the fop_* register patterns): Remove.
3891 (fscalexf4_i387): Rename from *fscalexf4_i387.
3892 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3893
3894 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3895
3896 PR tree-optimization/59779
3897 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3898 type for bitsize and maxsize instead of HOST_WIDE_INT.
3899
3900 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3901
3902 PR rtl-optimization/57320
3903 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3904 the CFG after thread_prologue_and_epilogue_insns.
3905
3906 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3907
3908 PR rtl-optimization/57189
3909 * lra-constraints.c (process_alt_operands): Disfavor spilling
3910 vector pseudos.
3911
3912 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3913
3914 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3915
3916 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3917
3918 PR tree-optimization/59025
3919 PR middle-end/60418
3920 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3921 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3922
3923 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3924
3925 PR target/60486
3926 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3927 calls of avr_out_plus_1.
3928
3929 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3930
3931 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3932 BB's single pred and update the father loop's latch info later.
3933
3934 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3935
3936 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3937 (VEC_M): Likewise.
3938 (VEC_N): Likewise.
3939 (VEC_R): Likewise.
3940 (VEC_base): Likewise.
3941 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3942 registers, we need to swap double words in little endian mode.
3943
3944 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3945 to be a container mode for 128-bit integer operations added in ISA
3946 2.07. Unlike TImode and PTImode, the preferred register set is
3947 the Altivec/VMX registers for the 128-bit operations.
3948
3949 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3950 declarations.
3951 (rs6000_split_128bit_ok_p): Likewise.
3952
3953 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3954 macros for creating ISA 2.07 normal and overloaded builtin
3955 functions with 3 arguments.
3956 (BU_P8V_OVERLOAD_3): Likewise.
3957 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3958 for use as overloaded functions.
3959 (VPERM_1TI_UNS): Likewise.
3960 (VSEL_1TI): Likewise.
3961 (VSEL_1TI_UNS): Likewise.
3962 (ST_INTERNAL_1ti): Likewise.
3963 (LD_INTERNAL_1ti): Likewise.
3964 (XXSEL_1TI): Likewise.
3965 (XXSEL_1TI_UNS): Likewise.
3966 (VPERM_1TI): Likewise.
3967 (VPERM_1TI_UNS): Likewise.
3968 (XXPERMDI_1TI): Likewise.
3969 (SET_1TI): Likewise.
3970 (LXVD2X_V1TI): Likewise.
3971 (STXVD2X_V1TI): Likewise.
3972 (VEC_INIT_V1TI): Likewise.
3973 (VEC_SET_V1TI): Likewise.
3974 (VEC_EXT_V1TI): Likewise.
3975 (EQV_V1TI): Likewise.
3976 (NAND_V1TI): Likewise.
3977 (ORC_V1TI): Likewise.
3978 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3979 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3980 overloaded builtin.
3981 (VADDUQM): Likewise.
3982 (VSUBCUQ): Likewise.
3983 (VADDEUQM): Likewise.
3984 (VADDECUQ): Likewise.
3985 (VSUBEUQM): Likewise.
3986 (VSUBECUQ): Likewise.
3987
3988 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3989 __int128_t and __uint128_t types.
3990 (__uint128_type): Likewise.
3991 (altivec_categorize_keyword): Add support for vector __int128_t,
3992 vector __uint128_t, vector __int128, and vector unsigned __int128
3993 as a container type for TImode operations that need to be done in
3994 VSX/Altivec registers.
3995 (rs6000_macro_to_expand): Likewise.
3996 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3997 to support 128-bit integer instructions vaddcuq, vadduqm,
3998 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3999 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
4000
4001 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
4002 for V1TImode, and set up preferences to use VSX/Altivec registers.
4003 Setup VSX reload handlers.
4004 (rs6000_debug_reg_global): Likewise.
4005 (rs6000_init_hard_regno_mode_ok): Likewise.
4006 (rs6000_preferred_simd_mode): Likewise.
4007 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
4008 (easy_altivec_constant): Likewise.
4009 (output_vec_const_move): Likewise.
4010 (rs6000_expand_vector_set): Convert V1TImode set and extract to
4011 simple move.
4012 (rs6000_expand_vector_extract): Likewise.
4013 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
4014 addressing.
4015 (rs6000_const_vec): Add support for V1TImode.
4016 (rs6000_emit_le_vsx_load): Swap double words when loading or
4017 storing TImode/V1TImode.
4018 (rs6000_emit_le_vsx_store): Likewise.
4019 (rs6000_emit_le_vsx_move): Likewise.
4020 (rs6000_emit_move): Add support for V1TImode.
4021 (altivec_expand_ld_builtin): Likewise.
4022 (altivec_expand_st_builtin): Likewise.
4023 (altivec_expand_vec_init_builtin): Likewise.
4024 (altivec_expand_builtin): Likewise.
4025 (rs6000_init_builtins): Add support for V1TImode type. Add
4026 support for ISA 2.07 128-bit integer builtins. Define type names
4027 for the VSX/Altivec vector types.
4028 (altivec_init_builtins): Add support for overloaded vector
4029 functions with V1TImode type.
4030 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4031 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4032 external function.
4033 (rs6000_split_128bit_ok_p): Likewise.
4034 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4035 __int128_t and vector __uint128_t.
4036
4037 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4038 and mode attributes.
4039 (VSX_M): Likewise.
4040 (VSX_M2): Likewise.
4041 (VSm): Likewise.
4042 (VSs): Likewise.
4043 (VSr): Likewise.
4044 (VSv): Likewise.
4045 (VS_scalar): Likewise.
4046 (VS_double): Likewise.
4047 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4048
4049 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4050 we support the ISA 2.07 128-bit integer arithmetic instructions.
4051 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4052 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4053 and TImode types for use with the builtin functions.
4054 (V1TI_type_node): Likewise.
4055 (unsigned_V1TI_type_node): Likewise.
4056 (intTI_type_internal_node): Likewise.
4057 (uintTI_type_internal_node): Likewise.
4058
4059 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4060 128-bit builtin functions.
4061 (UNSPEC_VADDEUQM): Likewise.
4062 (UNSPEC_VADDECUQ): Likewise.
4063 (UNSPEC_VSUBCUQ): Likewise.
4064 (UNSPEC_VSUBEUQM): Likewise.
4065 (UNSPEC_VSUBECUQ): Likewise.
4066 (VM): Add V1TImode to vector mode iterators.
4067 (VM2): Likewise.
4068 (VI_unit): Likewise.
4069 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4070 (altivec_vaddcuq): Likewise.
4071 (altivec_vsubuqm): Likewise.
4072 (altivec_vsubcuq): Likewise.
4073 (altivec_vaddeuqm): Likewise.
4074 (altivec_vaddecuq): Likewise.
4075 (altivec_vsubeuqm): Likewise.
4076 (altivec_vsubecuq): Likewise.
4077
4078 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4079 mode iterators.
4080 (BOOL_128): Likewise.
4081 (BOOL_REGS_OUTPUT): Likewise.
4082 (BOOL_REGS_OP1): Likewise.
4083 (BOOL_REGS_OP2): Likewise.
4084 (BOOL_REGS_UNARY): Likewise.
4085 (BOOL_REGS_AND_CR0): Likewise.
4086
4087 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4088 128-bit integer builtin support.
4089 (vec_vadduqm): Likewise.
4090 (vec_vaddecuq): Likewise.
4091 (vec_vaddeuqm): Likewise.
4092 (vec_vsubecuq): Likewise.
4093 (vec_vsubeuqm): Likewise.
4094 (vec_vsubcuq): Likewise.
4095 (vec_vsubuqm): Likewise.
4096
4097 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4098 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4099 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4100 128-bit integer add/subtract to ISA 2.07.
4101
4102 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4103
4104 * config/arc/arc.c (arc_predicate_delay_insns):
4105 Fix third argument passed to conditionalize_nonjump.
4106
4107 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4108
4109 * config/aarch64/aarch64-builtins.c
4110 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4111 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4112 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4113 instead of __builtin_lfloor.
4114 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4115
4116 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4117
4118 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4119 (tree_ssa_ifcombine_bb_1): New function.
4120 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4121 is an empty forwarder block to then_bb or vice versa and then_bb
4122 and else_bb are effectively swapped.
4123
4124 2014-03-12 Christian Bruel <christian.bruel@st.com>
4125
4126 PR target/60264
4127 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4128 REG_CFA_DEF_CFA note.
4129 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4130 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4131
4132 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4133
4134 PR tree-optimization/60454
4135 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4136
4137 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4138
4139 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4140 Do not define target_cpu_default2 to generic.
4141 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4142 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4143 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4144
4145 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4146 Marc Glisse <marc.glisse@inria.fr>
4147
4148 PR tree-optimization/60502
4149 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4150 instead of build_low_bits_mask.
4151
4152 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4153
4154 PR middle-end/60482
4155 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4156 if there are multiple uses, but op doesn't live on E edge.
4157 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4158 clobber stmts before __builtin_unreachable.
4159
4160 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4161
4162 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4163 hard_frame_pointer_rtx.
4164 * cse.c (cse_insn): Remove volatile check.
4165 * cselib.c (cselib_process_insn): Likewise.
4166 * dse.c (scan_insn): Likewise.
4167
4168 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4169
4170 * config/arc/arc.c (conditionalize_nonjump): New function,
4171 broken out of ...
4172 (arc_ifcvt): ... this.
4173 (arc_predicate_delay_insns): Use it.
4174
4175 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4176
4177 * config/arc/predicates.md (extend_operand): During/after reload,
4178 allow const_int_operand.
4179 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4180 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4181 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4182 to "i".
4183 (umulsi3_highpart_i): Likewise.
4184
4185 2014-03-11 Richard Biener <rguenther@suse.de>
4186
4187 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4188 Add asserts to guard possible wrong-code bugs.
4189
4190 2014-03-11 Richard Biener <rguenther@suse.de>
4191
4192 PR tree-optimization/60429
4193 PR tree-optimization/60485
4194 * tree-ssa-structalias.c (set_union_with_increment): Properly
4195 take into account all fields that overlap the shifted vars.
4196 (do_sd_constraint): Likewise.
4197 (do_ds_constraint): Likewise.
4198 (get_constraint_for_ptr_offset): Likewise.
4199
4200 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4201
4202 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4203 (nios2_compute_frame_layout):
4204 Add calculation of cfun->machine->fp_save_offset.
4205 (nios2_expand_prologue): Correct setting of frame pointer register
4206 in prologue.
4207 (nios2_expand_epilogue): Update recovery of stack pointer from
4208 frame pointer accordingly.
4209 (nios2_initial_elimination_offset): Update calculation of offset
4210 for eliminating to HARD_FRAME_POINTER_REGNUM.
4211
4212 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4213
4214 PR ipa/60457
4215 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4216 cgraph_get_create_node on VAR_DECLs.
4217
4218 2014-03-10 Richard Biener <rguenther@suse.de>
4219
4220 PR middle-end/60474
4221 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4222
4223 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4224
4225 * config/vms/vms.opt (vms_float_format): New variable.
4226
4227 2014-03-08 Tobias Burnus <burnus@net-b.de>
4228
4229 * doc/invoke.texi (-fcilkplus): Update implementation status.
4230
4231 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4232 Richard Biener <rguenther@suse.de>
4233
4234 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4235 consistently accross all TUs.
4236 (run_gcc): Enable -fshort-double automatically at link at link-time
4237 and disallow override.
4238
4239 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4240
4241 PR target/58271
4242 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4243 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4244 if they can't be used.
4245
4246 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4247
4248 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4249 for Solaris 11/x86 ld.
4250 * configure: Regenerate.
4251
4252 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4253
4254 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4255 (LIB_TLS_SPEC): Save as ld_tls_libs.
4256 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4257 (HAVE_AS_IX86_TLSLDM): New test.
4258 * configure, config.in: Regenerate.
4259 * config/i386/i386.c (legitimize_tls_address): Fall back to
4260 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4261 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4262 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4263 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4264
4265 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4266
4267 * common.opt (fira-loop-pressure): Mark as optimization.
4268
4269 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4270
4271 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4272 an OpenMP mappable type.
4273
4274 2014-03-06 Matthias Klose <doko@ubuntu.com>
4275
4276 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4277 MULTILIB_OSDIRNAMES is not defined.
4278
4279 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4280 Meador Inge <meadori@codesourcery.com>
4281
4282 PR target/58595
4283 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4284 (arm_legitimize_address): Call legitimize_tls_address for any
4285 arm_tls_referenced_p expression, handle constant addend. Call it
4286 before testing for !TARGET_ARM.
4287 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4288
4289 2014-03-06 Richard Biener <rguenther@suse.de>
4290
4291 PR middle-end/60445
4292 PR lto/60424
4293 PR lto/60427
4294 Revert
4295 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4296
4297 * tree-streamer.c (record_common_node): Assert we don't record
4298 nodes with type double.
4299 (preload_common_node): Skip type double, complex double and double
4300 pointer since it is now frontend dependent due to fshort-double option.
4301
4302 2014-03-06 Richard Biener <rguenther@suse.de>
4303
4304 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4305 or -fno-lto is specified and the linker has full plugin support.
4306 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4307 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4308 * lto-wrapper.c (merge_and_complain): Merge compile-time
4309 optimization levels.
4310 (run_gcc): And pass it through to the link options.
4311
4312 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4313
4314 PR debug/60381
4315 Revert:
4316 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4317 PR debug/59992
4318 * cselib.c (remove_useless_values): Skip to avoid quadratic
4319 behavior if the condition moved from...
4320 (cselib_process_insn): ... here holds.
4321
4322 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4323
4324 PR plugins/59335
4325 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4326 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4327
4328 PR plugins/59335
4329 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4330 (TM_H): Add x86-tune.def.
4331
4332 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4333
4334 * config/aarch64/aarch64.c (generic_tunings):
4335 Use cortexa57_extra_costs.
4336
4337 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4338
4339 PR lto/60404
4340 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4341 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4342 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4343 cost for in_lto_p.
4344
4345 2014-03-04 Heiher <r@hev.cc>
4346
4347 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4348 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4349
4350 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4351
4352 * config/i386/predicates.md (const2356_operand): Change to ...
4353 (const2367_operand): ... this.
4354 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4355 const2367_operand.
4356 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4357 (*avx512pf_scatterpf<mode>sf): Ditto.
4358 (avx512pf_scatterpf<mode>df): Ditto.
4359 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4360 (*avx512pf_scatterpf<mode>df): Ditto.
4361 * config/i386/i386.c (ix86_expand_builtin): Update
4362 incorrect hint operand error message.
4363
4364 2014-03-04 Richard Biener <rguenther@suse.de>
4365
4366 * lto-section-in.c (lto_get_section_data): Fix const cast.
4367
4368 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4369
4370 * tree-streamer.c (record_common_node): Assert we don't record
4371 nodes with type double.
4372 (preload_common_node): Skip type double, complex double and double
4373 pointer since it is now frontend dependent due to fshort-double option.
4374
4375 2014-03-04 Richard Biener <rguenther@suse.de>
4376
4377 PR lto/60405
4378 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4379 (lto_input_toplevel_asms): Likewise.
4380 * lto-section-in.c (lto_get_section_data): Instead do it here
4381 for every section.
4382
4383 2014-03-04 Richard Biener <rguenther@suse.de>
4384
4385 PR tree-optimization/60382
4386 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4387 dead PHIs a reduction.
4388
4389 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4390
4391 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4392 hint value.
4393 (_mm_prefetch): Move out of GCC target("sse") pragma.
4394 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4395 GCC target("prfchw") pragma.
4396 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4397 for locality <= 2.
4398 * config/i386/i386.c (ix86_option_override_internal): Enable
4399 -mprfchw with -mprefetchwt1.
4400
4401 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4402
4403 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4404 Mark as varying.
4405
4406 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4407
4408 * opts.h (CL_PCH_IGNORE): Define.
4409 * targhooks.c (option_affects_pch_p):
4410 Return false for options that have CL_PCH_IGNORE set.
4411 * opt-functions.awk: Process PchIgnore.
4412 * doc/options.texi: Document PchIgnore.
4413
4414 * config/arc/arc.opt (misize): Add PchIgnore property.
4415
4416 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4417
4418 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4419 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4420 constraint on constants to permit them being loaded into
4421 GENERAL_REGS or BASE_REGS.
4422
4423 2014-03-03 Nick Clifton <nickc@redhat.com>
4424
4425 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4426 anti-cacnonical alternatives.
4427 (negandhi3_real): New pattern.
4428 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4429
4430 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4431
4432 * config/avr/avr-mcus.def: Remove atxmega16x1.
4433 * config/avr/avr-tables.opt: Regenerate.
4434 * config/avr/t-multilib: Regenerate.
4435 * doc/avr-mmcu.texi: Regenerate.
4436
4437 2014-03-03 Tobias Grosser <tobias@grosser.es>
4438 Mircea Namolaru <mircea.namolaru@inria.fr>
4439
4440 PR tree-optimization/58028
4441 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4442 scalar dimensions.
4443
4444 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4445
4446 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4447 not handled by recognizers.
4448
4449 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4450
4451 PR middle-end/60175
4452 * function.c (expand_function_end): Don't emit
4453 clobber_return_register sequence if clobber_after is a BARRIER.
4454 * cfgexpand.c (construct_exit_block): Append instructions before
4455 return_label to prev_bb.
4456
4457 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4458
4459 * config/rs6000/constraints.md: Document reserved use of "wc".
4460
4461 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4462
4463 PR ipa/60150
4464 * ipa.c (function_and_variable_visibility): When dissolving comdat
4465 group, also set all symbols to local.
4466
4467 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4468
4469 PR ipa/60306
4470
4471 Revert:
4472 2013-12-14 Jan Hubicka <jh@suse.cz>
4473 PR middle-end/58477
4474 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4475
4476 2014-03-02 Jon Beniston <jon@beniston.com>
4477
4478 PR bootstrap/48230
4479 PR bootstrap/50927
4480 PR bootstrap/52466
4481 PR target/46898
4482 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4483 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4484 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4485 (simple_return, *simple_return): New patterns
4486 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4487 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4488
4489 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4490
4491 * dwarf2out.c (gen_subprogram_die): Tidy.
4492
4493 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4494
4495 PR target/60071
4496 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4497 (*mov_t_msb_neg_negc): ... this new insn.
4498
4499 2014-02-28 Jason Merrill <jason@redhat.com>
4500
4501 PR c++/58678
4502 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4503 function.
4504
4505 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4506
4507 PR c++/60314
4508 * dwarf2out.c (decltype_auto_die): New static.
4509 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4510 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4511 (is_cxx_auto): Likewise.
4512
4513 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4514
4515 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4516 we are not using general regs only.
4517
4518 2014-02-28 Richard Biener <rguenther@suse.de>
4519
4520 PR target/60280
4521 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4522 previous fix and only allow to remove trivial pre-headers
4523 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4524 (remove_forwarder_block): Properly update the latch of a loop.
4525
4526 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4527
4528 PR debug/59992
4529 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4530 (cselib_preserved_hash_table): New.
4531 (preserve_constants_and_equivs): Move preserved vals to it.
4532 (cselib_find_slot): Look it up first.
4533 (cselib_init): Initialize it.
4534 (cselib_finish): Release it.
4535 (dump_cselib_table): Dump it.
4536
4537 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4538
4539 PR debug/59992
4540 * cselib.c (remove_useless_values): Skip to avoid quadratic
4541 behavior if the condition moved from...
4542 (cselib_process_insn): ... here holds.
4543
4544 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4545
4546 PR debug/57232
4547 * var-tracking.c (vt_initialize): Apply the same condition to
4548 preserve the CFA base value.
4549
4550 2014-02-28 Joey Ye <joey.ye@arm.com>
4551
4552 PR target/PR60169
4553 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4554 if reload in progress or completed.
4555
4556 2014-02-28 Tobias Burnus <burnus@net-b.de>
4557
4558 PR middle-end/60147
4559 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4560 NAMELIST_DECL.
4561
4562 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4563
4564 * doc/tm.texi.in (Condition Code Status): Update documention for
4565 relative locations of cc0-setter and cc0-user.
4566
4567 2014-02-27 Jeff Law <law@redhat.com>
4568
4569 PR rtl-optimization/52714
4570 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4571 into two independent simple sets, if I3 is a jump, ensure the
4572 pattern we place into I3 is a (set (pc) ...).
4573
4574 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4575 Jeff Law <law@redhat.com>
4576
4577 PR rtl-optimization/49847
4578 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4579 are in different blocks.
4580 * doc/tm.texi (Condition Code Status): Update documention for
4581 relative locations of cc0-setter and cc0-user.
4582
4583 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4584
4585 PR target/59222
4586 * lra.c (lra_emit_add): Check SUBREG too.
4587
4588 2014-02-27 Andreas Schwab <schwab@suse.de>
4589
4590 * config/m68k/m68k.c (m68k_option_override): Disable
4591 -flive-range-shrinkage for classic m68k.
4592 (m68k_override_options_after_change): Likewise.
4593
4594 2014-02-27 Marek Polacek <polacek@redhat.com>
4595
4596 PR middle-end/59223
4597 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4598 -Wmaybe-uninitialized.
4599
4600 2014-02-27 Alan Modra <amodra@gmail.com>
4601
4602 PR target/57936
4603 * reload1.c (emit_input_reload_insns): When reload_override_in,
4604 set old to rl->in_reg when rl->in_reg is a subreg.
4605
4606 2014-02-26 Richard Biener <rguenther@suse.de>
4607
4608 PR bootstrap/60343
4609 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4610
4611 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4612
4613 * common/config/i386/predicates.md (const1256_operand): Remove.
4614 (const2356_operand): New.
4615 (const_1_to_2_operand): Remove.
4616 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4617 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4618 (*avx512pf_gatherpf<mode>sf): Ditto.
4619 (avx512pf_gatherpf<mode>df): Ditto.
4620 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4621 (*avx512pf_gatherpf<mode>df): Ditto.
4622 (avx512pf_scatterpf<mode>sf): Ditto.
4623 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4624 (*avx512pf_scatterpf<mode>sf): Ditto.
4625 (avx512pf_scatterpf<mode>df): Ditto.
4626 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4627 (*avx512pf_scatterpf<mode>df): Ditto.
4628 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4629
4630 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4631
4632 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4633 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4634 (_mm512_mask_testn_epi64_mask): Move to ...
4635 * config/i386/avx512cdintrin.h: Here.
4636 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4637 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4638 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4639 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4640 TARGET_AVX512F from TARGET_AVX512CD.
4641
4642 2014-02-26 Richard Biener <rguenther@suse.de>
4643
4644 PR ipa/60327
4645 * ipa.c (walk_polymorphic_call_targets): Properly guard
4646 call to inline_update_overall_summary.
4647
4648 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4649
4650 PR target/60280
4651 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4652 and latches only if requested. Fix latch if it is removed.
4653 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4654 LOOPS_HAVE_PREHEADERS.
4655
4656 2014-02-25 Andrew Pinski <apinski@cavium.com>
4657
4658 * builtins.c (expand_builtin_thread_pointer): Create a new target
4659 when the target is NULL.
4660
4661 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4662
4663 PR rtl-optimization/60317
4664 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4665 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4666 * lra-assigns.c: Include params.h.
4667 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4668 other reload pseudos considerations.
4669
4670 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4671
4672 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4673 to use canonical form for nor<mode>3.
4674
4675 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4676
4677 PR target/55426
4678 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4679 conversions.
4680
4681 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4682
4683 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4684 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4685 (ix86_handle_option): Handle OPT_mprefetchwt1.
4686 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4687 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4688 PREFETCHWT1 CPUID.
4689 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4690 OPTION_MASK_ISA_PREFETCHWT1.
4691 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4692 (PTA_PREFETCHWT1): New.
4693 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4694 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4695 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4696 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4697 (*prefetch_avx512pf_<mode>_: Change into ...
4698 (*prefetch_prefetchwt1_<mode>: This.
4699 * config/i386/i386.opt (mprefetchwt1): New.
4700 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4701 (_mm_prefetch): Handle intent to write.
4702 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4703
4704 2014-02-25 Richard Biener <rguenther@suse.de>
4705
4706 PR middle-end/60291
4707 * emit-rtl.c (mem_attrs_htab): Remove.
4708 (mem_attrs_htab_hash): Likewise.
4709 (mem_attrs_htab_eq): Likewise.
4710 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4711 (init_emit_once): Do not allocate mem_attrs_htab.
4712
4713 2014-02-25 Richard Biener <rguenther@suse.de>
4714
4715 PR lto/60319
4716 * lto-opts.c (lto_write_options): Output non-explicit conservative
4717 -fwrapv, -fno-trapv and -fno-strict-overflow.
4718 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4719 (run_gcc): And pass them through.
4720
4721 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4722
4723 * sel-sched.c (calculate_new_fences): New parameter ptime.
4724 Calculate it as a maximum over all fence cycles.
4725 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4726 Print the final schedule timing when sched_verbose.
4727
4728 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4729
4730 PR rtl-optimization/60292
4731 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4732 bit fot the fence instruction.
4733
4734 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4735
4736 * calls.h: Fix typo in comment.
4737
4738 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4739
4740 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4741 adjusting offsetable addresses.
4742
4743 2014-02-24 Guozhi Wei <carrot@google.com>
4744
4745 * sparseset.h (sparseset_pop): Fix the wrong index.
4746
4747 2014-02-24 Walter Lee <walt@tilera.com>
4748
4749 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4750 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4751 triplet.
4752 * common/config/tilegx/tilegx-common.c
4753 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4754 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4755 (LINK_SPEC): Ditto.
4756 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4757 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4758 (tilegx_gimplify_va_arg_expr): Handle big endian.
4759 (tilegx_expand_unaligned_load): Ditto.
4760 (tilegx_expand_unaligned_store): Ditto.
4761 (TARGET_RETURN_IN_MSB): New.
4762 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4763 (TARGET_ENDIAN_DEFAULT): New.
4764 (TARGET_BIG_ENDIAN): Handle big endian.
4765 (BYTES_BIG_ENDIAN): Ditto.
4766 (WORDS_BIG_ENDIAN): Ditto.
4767 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4768 (ENDIAN_SPEC): New.
4769 (EXTRA_SPECS): New.
4770 * config/tilegx/tilegx.md (extv): Handle big endian.
4771 (extzv): Ditto.
4772 (insn_st<n>): Ditto.
4773 (insn_st<n>_add<bitsuffix>): Ditto.
4774 (insn_stnt<n>): Ditto.
4775 (insn_stnt<n>_add<bitsuffix>):Ditto.
4776 (vec_interleave_highv8qi): Handle big endian.
4777 (vec_interleave_highv8qi_be): New.
4778 (vec_interleave_highv8qi_le): New.
4779 (insn_v1int_h): Handle big endian.
4780 (vec_interleave_lowv8qi): Handle big endian.
4781 (vec_interleave_lowv8qi_be): New.
4782 (vec_interleave_lowv8qi_le): New.
4783 (insn_v1int_l): Handle big endian.
4784 (vec_interleave_highv4hi): Handle big endian.
4785 (vec_interleave_highv4hi_be): New.
4786 (vec_interleave_highv4hi_le): New.
4787 (insn_v2int_h): Handle big endian.
4788 (vec_interleave_lowv4hi): Handle big endian.
4789 (vec_interleave_lowv4hi_be): New.
4790 (vec_interleave_lowv4hi_le): New.
4791 (insn_v2int_l): Handle big endian.
4792 (vec_interleave_highv2si): Handle big endian.
4793 (vec_interleave_highv2si_be): New.
4794 (vec_interleave_highv2si_le): New.
4795 (insn_v4int_h): Handle big endian.
4796 (vec_interleave_lowv2si): Handle big endian.
4797 (vec_interleave_lowv2si_be): New.
4798 (vec_interleave_lowv2si_le): New.
4799 (insn_v4int_l): Handle big endian.
4800 * config/tilegx/tilegx.opt (mbig-endian): New option.
4801 (mlittle-endian): New option.
4802 * doc/install.texi: Document tilegxbe-linux.
4803 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4804
4805 2014-02-24 Martin Jambor <mjambor@suse.cz>
4806
4807 PR ipa/60266
4808 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4809 there are no parameter descriptors.
4810
4811 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4812
4813 PR rtl-optimization/60268
4814 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4815 initialization to ...
4816 (sched_rgn_init): ... here.
4817 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4818
4819 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4820
4821 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4822 names.
4823
4824 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4825
4826 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4827 definition.
4828
4829 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4830
4831 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4832 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4833
4834 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4835
4836 * config/microblaze/predicates.md: Add cmp_op predicate.
4837 * config/microblaze/microblaze.md: Add branch_compare instruction
4838 which uses cmp_op predicate and emits cmp insn before branch.
4839 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4840 to microblaze_expand_conditional_branch and consolidate logic.
4841 (microblaze_expand_conditional_branch): emit branch_compare
4842 insn instead of handling cmp op separate from branch insn.
4843
4844 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4845
4846 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4847 to permit subregs.
4848
4849 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4850
4851 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4852 define_insn with define_expand and new define_insn
4853 *altivec_lve<VI_char>x_internal.
4854 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4855 and new define_insn *altivec_stve<VI_char>x_internal.
4856 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4857 prototype.
4858 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4859 lve*x built-ins.
4860 (altivec_expand_stvex_be): New function.
4861
4862 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4863
4864 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4865 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4866 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4867 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4868
4869 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4870
4871 PR target/60298
4872 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4873 instead of emit_move_insn.
4874
4875 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4876
4877 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4878 vspltw with vsldoi.
4879 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4880 gen_altivec_vsumsws.
4881
4882 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4883
4884 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4885 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4886 (altivec_lvxl_<mode>): New define_expand incorporating
4887 -maltivec=be semantics where needed.
4888 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4889 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4890 semantics where needed.
4891 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4892 (altivec_stvx_<mode>): New define_expand incorporating
4893 -maltivec=be semantics where needed.
4894 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4895 VM2 iterator instead of V4SI.
4896 (altivec_stvxl_<mode>): New define_expand incorporating
4897 -maltivec=be semantics where needed.
4898 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4899 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4900 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4901 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4902 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4903 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4904 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4905 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4906 ALTIVEC_BUILTIN_STVXL.
4907 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4908 (altivec_expand_stvx_be): Likewise.
4909 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4910 (altivec_expand_lvx_be): Likewise.
4911 (altivec_expand_stvx_be): Likewise.
4912 (altivec_expand_builtin): Add cases for
4913 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4914 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4915 (altivec_init_builtins): Add definitions for
4916 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4917 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4918
4919 2014-02-21 Catherine Moore <clm@codesourcery.com>
4920
4921 * doc/invoke.texi (mvirt, mno-virt): Document.
4922 * config/mips/mips.opt (mvirt): New option.
4923 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4924
4925 2014-02-21 Richard Biener <rguenther@suse.de>
4926
4927 PR tree-optimization/60276
4928 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4929 (STMT_VINFO_MIN_NEG_DIST): New macro.
4930 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4931 STMT_VINFO_MIN_NEG_DIST.
4932 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4933 made for negative dependence distances still hold.
4934
4935 2014-02-21 Richard Biener <rguenther@suse.de>
4936
4937 PR middle-end/60291
4938 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4939 DECL_INITIAL for globals not in the current function context.
4940
4941 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4942
4943 PR tree-optimization/56490
4944 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4945 * tree-ssa-uninit.c: Include params.h.
4946 (compute_control_dep_chain): Add num_calls argument, return false
4947 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4948 num_calls to recursive call.
4949 (find_predicates): Change dep_chain into normal array,
4950 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4951 variable and adjust compute_control_dep_chain caller.
4952 (find_def_preds): Likewise.
4953
4954 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4955
4956 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4957 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4958
4959 2014-02-21 Nick Clifton <nickc@redhat.com>
4960
4961 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4962 (pushhi1): Likewise.
4963 (popqi1): Add mode to pre_dec.
4964 (pophi1): Likewise.
4965
4966 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4967
4968 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4969 mode for mask of V8SFmode permutation.
4970
4971 2014-02-20 Richard Henderson <rth@redhat.com>
4972
4973 PR c++/60272
4974 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4975 a new pseudo for OLDVAL.
4976
4977 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4978
4979 PR target/57896
4980 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4981 gen_reg_rtx if d->testing_p.
4982 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4983 if d->testing_p and we will certainly return true.
4984 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4985 if d->testing_p.
4986
4987 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4988
4989 * emit-rtl.c (gen_reg_rtx): Assert that
4990 crtl->emit.regno_pointer_align_length is non-zero.
4991
4992 2014-02-20 Richard Henderson <rth@redhat.com>
4993
4994 PR c++/60272
4995 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4996 on failure the store back into EXPECT.
4997
4998 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4999 Sandra Loosemore <sandra@codesourcery.com>
5000
5001 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
5002 * config/nios2/nios2.c (nios2_function_profiler): Add
5003 -fPIC (flag_pic == 2) support.
5004 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
5005 (nios2_large_offset_p): New function.
5006 (nios2_unspec_reloc_p): Move up position, update to use
5007 nios2_large_offset_p.
5008 (nios2_unspec_address): Remove function.
5009 (nios2_unspec_offset): New function.
5010 (nios2_large_got_address): New function.
5011 (nios2_got_address): Add large offset support.
5012 (nios2_legitimize_tls_address): Update usage of removed and new
5013 functions.
5014 (nios2_symbol_binds_local_p): New function.
5015 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5016 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5017 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5018 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5019 processing for (const (unspec ...)).
5020 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5021
5022 2014-02-20 Richard Biener <rguenther@suse.de>
5023
5024 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5025 doing the substitution.
5026 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5027
5028 2014-02-20 Martin Jambor <mjambor@suse.cz>
5029
5030 PR ipa/55260
5031 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5032 info when checking whether lattices are bottom.
5033
5034 2014-02-20 Richard Biener <rguenther@suse.de>
5035
5036 PR middle-end/60221
5037 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5038 regions at -O0.
5039
5040 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5041
5042 PR ipa/58555
5043 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5044 parameter specifying the scaling.
5045 (inline_call): Update.
5046 (want_inline_recursively): Guard division by zero.
5047 (recursive_inlining): Update.
5048 * ipa-inline.h (clone_inlined_nodes): Update.
5049
5050 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5051
5052 PR target/60204
5053 * config/i386/i386.c (classify_argument): Pass structures of size
5054 64 bytes or less in register.
5055
5056 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5057 Kirill Yukhin <kirill.yukhin@intel.com>
5058
5059 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5060 (_mm_rcp28_round_ss): Ditto.
5061 (_mm_rsqrt28_round_sd): Ditto.
5062 (_mm_rsqrt28_round_ss): Ditto.
5063 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5064 (_mm_rcp14_round_ss): Ditto.
5065 (_mm_rsqrt14_round_sd): Ditto.
5066 (_mm_rsqrt14_round_ss): Ditto.
5067 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5068 the first input operand, get rid of match_dup.
5069 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5070 attribute to sse.
5071 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5072 Ditto.
5073 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5074 operand as the first input operand, set type attribute.
5075 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5076 Set type attribute.
5077 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5078 operand as the first input operand, set type attribute.
5079
5080 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5081
5082 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5083 bit of zero.
5084
5085 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5086
5087 PR target/60207
5088 * config/i386/i386.c (construct_container): Remove TFmode check
5089 for X86_64_INTEGER_CLASS.
5090
5091 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5092
5093 PR target/59794
5094 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5095 only when -Wpsabi is enabled.
5096
5097 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5098
5099 PR target/59799
5100 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5101 passing arrays in registers are the same as for structs, so remove the
5102 special case for them.
5103
5104 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5105
5106 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5107 destination type, extract only the valid bits if the source type is not
5108 integral and has a different mode.
5109
5110 2014-02-19 Richard Biener <rguenther@suse.de>
5111
5112 PR ipa/60243
5113 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5114 for all calls.
5115
5116 2014-02-19 Richard Biener <rguenther@suse.de>
5117
5118 PR ipa/60243
5119 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5120 (ipa_modify_call_arguments): Emit an argument load explicitely and
5121 preserve virtual SSA form there and for the replacement call.
5122 Do not update SSA form nor free dominance info.
5123
5124 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5125
5126 * ipa.c (function_and_variable_visibility): Also clear WEAK
5127 flag when disolving COMDAT_GROUP.
5128
5129 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5130
5131 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5132 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5133 not devirtualizing.
5134 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5135 do more sanity checks.
5136 (detect_type_change): Return true when giving up early.
5137 (compute_complex_assign_jump_func): Fix type parameter of
5138 ipa_set_ancestor_jf.
5139 (compute_complex_ancestor_jump_func): Likewise.
5140 (update_jump_functions_after_inlining): Fix updating of
5141 ancestor function.
5142 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5143
5144 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5145
5146 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5147 inline clones when edge disappears.
5148
5149 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5150
5151 PR target/60203
5152 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5153 Split 64-bit moves into 2 patterns. Do not allow the use of
5154 direct move for TDmode in little endian, since the decimal value
5155 has little endian bytes within a word, but the 64-bit pieces are
5156 ordered in a big endian fashion, and normal subreg's of TDmode are
5157 not allowed.
5158 (mov<mode>_64bit_dm): Likewise.
5159 (movtd_64bit_nodm): Likewise.
5160
5161 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5162
5163 PR tree-optimization/60174
5164 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5165 statement of an SSA_NAME that occurs in an abnormal PHI node.
5166
5167 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5168
5169 PR sanitizer/60142
5170 * final.c (SEEN_BB): Remove.
5171 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5172 (final_scan_insn): Don't force_source_line on second
5173 NOTE_INSN_BASIC_BLOCK.
5174
5175 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5176
5177 PR target/60205
5178 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5179 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5180 (type_natural_mode): Warn ABI change when %zmm register is not
5181 available for AVX512F vector value passing.
5182
5183 2014-02-18 Kai Tietz <ktietz@redhat.com>
5184
5185 PR target/60193
5186 * config/i386/i386.c (ix86_expand_prologue): Use value in
5187 rax register as displacement when restoring %r10 or %rax.
5188 Fix wrong offset when restoring both registers.
5189
5190 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5191
5192 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5193 assertion with conditional return.
5194
5195 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5196 Uros Bizjak <ubizjak@gmail.com>
5197
5198 PR driver/60233
5199 * config/i386/driver-i386.c (host_detect_local_cpu): If
5200 YMM state is not saved by the OS, also clear has_f16c. Move
5201 CPUID 0x80000001 handling before YMM state saving checking.
5202
5203 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5204
5205 PR rtl-optimization/58960
5206 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5207 factored out from ...
5208 (sched_init): ... here.
5209 (free_global_sched_pressure_data): New, factored out from ...
5210 (sched_finish): ... here.
5211 * sched-int.h (free_global_sched_pressure_data): Declare.
5212 * sched-rgn.c (nr_regions_initial): New static global.
5213 (haifa_find_rgns): Initialize it.
5214 (schedule_region): Disable sched-pressure for the newly
5215 generated regions.
5216
5217 2014-02-17 Richard Biener <rguenther@suse.de>
5218
5219 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5220 release SSA defs of pattern stmts.
5221
5222 2014-02-17 Richard Biener <rguenther@suse.de>
5223
5224 * tree-inline.c (expand_call_inline): Release the virtual
5225 operand defined by the call we are about to inline.
5226
5227 2014-02-17 Richard Biener <rguenther@suse.de>
5228
5229 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5230
5231 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5232 Ilya Tocar <ilya.tocar@intel.com>
5233
5234 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5235 arguments order in builtin.
5236 (_mm512_permutexvar_epi64): Ditto.
5237 (_mm512_mask_permutexvar_epi64): Ditto
5238 (_mm512_maskz_permutexvar_epi32): Ditto
5239 (_mm512_permutexvar_epi32): Ditto
5240 (_mm512_mask_permutexvar_epi32): Ditto
5241
5242 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5243
5244 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5245 (p8_vmrgow): Likewise.
5246
5247 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5248
5249 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5250 endian targets.
5251
5252 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5253
5254 PR target/60203
5255 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5256 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5257 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5258 using direct move instructions on ISA 2.07. Also adjust
5259 instruction length for 64-bit.
5260 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5261 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5262
5263 2014-02-15 Alan Modra <amodra@gmail.com>
5264
5265 PR target/58675
5266 PR target/57935
5267 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5268 find_replacement on parts of insn rtl that might be reloaded.
5269
5270 2014-02-15 Richard Biener <rguenther@suse.de>
5271
5272 PR tree-optimization/60183
5273 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5274 (tree_ssa_phiprop): Calculate and free post-dominators.
5275
5276 2014-02-14 Jeff Law <law@redhat.com>
5277
5278 PR rtl-optimization/60131
5279 * ree.c (get_extended_src_reg): New function.
5280 (combine_reaching_defs): Use it rather than assuming location of REG.
5281 (find_and_remove_re): Verify first operand of extension is
5282 a REG before adding the insns to the copy list.
5283
5284 2014-02-14 Roland McGrath <mcgrathr@google.com>
5285
5286 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5287 * configure: Regenerated.
5288 * config.in: Regenerated.
5289 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5290 instead of ASM_SHORT.
5291
5292 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5293 Richard Earnshaw <rearnsha@arm.com>
5294
5295 PR rtl-optimization/59535
5296 * lra-constraints.c (process_alt_operands): Encourage alternative
5297 when unassigned pseudo class is superset of the alternative class.
5298 (inherit_reload_reg): Don't inherit when optimizing for code size.
5299 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5300 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5301 modes not less than 4 for Thumb1.
5302
5303 2014-02-14 Kyle McMartin <kyle@redhat.com>
5304
5305 PR pch/60010
5306 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5307
5308 2014-02-14 Richard Biener <rguenther@suse.de>
5309
5310 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5311 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5312 Do not strip INDIRECT_REFs.
5313
5314 2014-02-14 Richard Biener <rguenther@suse.de>
5315
5316 PR lto/60179
5317 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5318 DECL_FUNCTION_SPECIFIC_TARGET.
5319 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5320 * tree-streamer-out.c (pack_ts_target_option): Remove.
5321 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5322 (write_ts_function_decl_tree_pointers): Do not stream
5323 DECL_FUNCTION_SPECIFIC_TARGET.
5324 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5325 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5326 (lto_input_ts_function_decl_tree_pointers): Do not stream
5327 DECL_FUNCTION_SPECIFIC_TARGET.
5328
5329 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5330
5331 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5332 (get_initial_def_for_induction, vectorizable_induction): Ignore
5333 debug stmts when looking for exit_phi.
5334 (vectorizable_live_operation): Fix up condition.
5335
5336 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5337
5338 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5339 nreverse() because it changes the content of original tree list.
5340
5341 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5342
5343 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5344 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5345
5346 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5347
5348 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5349 GNU coding standards.
5350
5351 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5352
5353 PR debug/60152
5354 * dwarf2out.c (gen_subprogram_die): Don't call
5355 add_calling_convention_attribute if subr_die is old_die.
5356
5357 2014-02-13 Sharad Singhai <singhai@google.com>
5358
5359 * doc/optinfo.texi: Fix order of nodes.
5360
5361 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5362
5363 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5364 operands[2], not operands[3].
5365
5366 2014-02-13 Richard Biener <rguenther@suse.de>
5367
5368 PR bootstrap/59878
5369 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5370 mention the possibility of an in-tree build.
5371 (CLooG): Update recommended version to 0.18.1, mention the
5372 possibility of an in-tree build and clarify that the ISL
5373 bundled with CLooG does not work.
5374
5375 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5376
5377 PR target/43546
5378 * expr.c (compress_float_constant): If x is a hard register,
5379 extend into a pseudo and then move to x.
5380
5381 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5382
5383 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5384 caused by bad second argument to warning_at() with -mhotpatch and
5385 nested functions (e.g. with gfortran).
5386
5387 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5388
5389 * opts.c (option_name): Remove "enabled by default" rider.
5390
5391 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5392
5393 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5394
5395 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5396 Uros Bizjak <ubizjak@gmail.com>
5397
5398 PR target/60151
5399 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5400 * configure: Regenerated.
5401
5402 2014-02-12 Richard Biener <rguenther@suse.de>
5403
5404 * vec.c (vec_prefix::calculate_allocation): Move as
5405 inline variant to vec.h.
5406 (vec_prefix::calculate_allocation_1): New out-of-line version.
5407 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5408 (vec_prefix::m_has_auto_buf): Rename to ...
5409 (vec_prefix::m_using_auto_storage): ... this.
5410 (vec_prefix::calculate_allocation): Inline the easy cases
5411 and dispatch to calculate_allocation_1 which doesn't need the
5412 prefix address.
5413 (va_heap::reserve): Use gcc_checking_assert.
5414 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5415 m_using_auto_storage.
5416 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5417 member and adjust.
5418 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5419 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5420 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5421
5422 2014-02-12 Richard Biener <rguenther@suse.de>
5423
5424 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5425 when we found a dependence.
5426
5427 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5428
5429 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5430 common code...
5431 (maybe_fold_stmt): ... into this new function.
5432 * omp-low.c (lower_omp): Update comment.
5433
5434 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5435 last use.
5436
5437 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5438 dereference.
5439
5440 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5441
5442 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5443 identifiers in comments.
5444 (cortexa53_extra_costs): Likewise.
5445 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5446 (cortexa7_extra_costs): Likewise.
5447 (cortexa12_extra_costs): Likewise.
5448 (cortexa15_extra_costs): Likewise.
5449 (v7m_extra_costs): Likewise.
5450
5451 2014-02-12 Richard Biener <rguenther@suse.de>
5452
5453 PR middle-end/60092
5454 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5455 of posix_memalign being successful.
5456 (lower_stmt): Restrict lowering of posix_memalign to when
5457 -ftree-bit-ccp is enabled.
5458
5459 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5460
5461 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5462 arg_loc.
5463 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5464
5465 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5466
5467 PR rtl-optimization/60116
5468 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5469 other_insn once the combination has been validated.
5470
5471 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5472
5473 PR lto/59468
5474 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5475 and wrapper.
5476 * ipa-devirt.c: Include demangle.h
5477 (odr_violation_reported): New static variable.
5478 (add_type_duplicate): Update odr_violations.
5479 (maybe_record_node): Add completep parameter; update it.
5480 (record_target_from_binfo): Add COMPLETEP parameter;
5481 update it as needed.
5482 (possible_polymorphic_call_targets_1): Likewise.
5483 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5484 rename FINAL to COMPLETE.
5485 (record_targets_from_bases): Sanity check we found the binfo;
5486 fix COMPLETEP updating.
5487 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5488 parameter, fix computing of COMPLETEP.
5489 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5490 at LTO time do demangling.
5491 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5492 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5493 parameter.
5494 (gimple_get_virt_method_for_binfo): Likewise.
5495 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5496 gimple_get_virt_method_for_vtable): Update prototypes.
5497
5498 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5499
5500 PR target/49008
5501 * genautomata.c (add_presence_absence): Fix typo with
5502 {final_}presence_list.
5503
5504 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5505
5506 PR target/60137
5507 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5508 for VSX/Altivec vectors that land in GPR registers.
5509
5510 2014-02-11 Richard Henderson <rth@redhat.com>
5511 Jakub Jelinek <jakub@redhat.com>
5512
5513 PR debug/59776
5514 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5515 around drhs if type conversion to lacc->type is not useless.
5516
5517 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5518
5519 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5520 tuning struct.
5521 (cortex-a57.cortex-a53): Likewise.
5522 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5523
5524 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5525
5526 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5527 arm_restrict_it.
5528
5529 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5530
5531 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5532 add_options_for_arm_vfp3.
5533
5534 2014-02-11 Jeff Law <law@redhat.com>
5535
5536 PR middle-end/54041
5537 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5538 object with an undesirable mode.
5539
5540 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5541
5542 PR libgomp/60107
5543 * config/i386/sol2-9.h: New file.
5544 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5545 *-*-solaris2.9*): Use it.
5546
5547 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5548
5549 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5550 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5551
5552 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5553
5554 * config/microblaze/microblaze.c: Extend mcpu version format
5555
5556 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5557
5558 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5559
5560 2014-02-10 Richard Henderson <rth@redhat.com>
5561
5562 PR target/59927
5563 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5564 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5565 ms-abi vs -mno-accumulate-outgoing-args.
5566 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5567 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5568 respect to ms-abi.
5569
5570 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5571
5572 PR middle-end/60080
5573 * cfgexpand.c (expand_asm_operands): Attach source location to
5574 ASM_INPUT rtx objects.
5575 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5576
5577 2014-02-10 Nick Clifton <nickc@redhat.com>
5578
5579 * config/mn10300/mn10300.c (popcount): New function.
5580 (mn10300_expand_prologue): Include saved registers in stack usage
5581 count.
5582
5583 2014-02-10 Jeff Law <law@redhat.com>
5584
5585 PR middle-end/52306
5586 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5587 when changing the SET_DEST of a prior insn to avoid an input reload.
5588
5589 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5590
5591 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5592 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5593 -mcall-openbsd, or -mcall-linux.
5594 (CC1_ENDIAN_BIG_SPEC): Remove.
5595 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5596 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5597 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5598 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5599 and %cc1_endian_default.
5600 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5601
5602 2014-02-10 Richard Biener <rguenther@suse.de>
5603
5604 PR tree-optimization/60115
5605 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5606 MEM_REF handling. Properly verify that the accesses are not
5607 out of the objects bound.
5608
5609 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5610
5611 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5612 coretex to cortex.
5613
5614 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5615
5616 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5617 proper constants and fix formatting.
5618 (possible_polymorphic_call_targets): Fix formatting.
5619
5620 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5621 Ilya Tocar <ilya.tocar@intel.com>
5622
5623 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5624 (_mm512_loadu_epi32): Renamed into...
5625 (_mm512_loadu_si512): This.
5626 (_mm512_storeu_epi32): Renamed into...
5627 (_mm512_storeu_si512): This.
5628 (_mm512_maskz_ceil_ps): Removed.
5629 (_mm512_maskz_ceil_pd): Ditto.
5630 (_mm512_maskz_floor_ps): Ditto.
5631 (_mm512_maskz_floor_pd): Ditto.
5632 (_mm512_floor_round_ps): Ditto.
5633 (_mm512_floor_round_pd): Ditto.
5634 (_mm512_ceil_round_ps): Ditto.
5635 (_mm512_ceil_round_pd): Ditto.
5636 (_mm512_mask_floor_round_ps): Ditto.
5637 (_mm512_mask_floor_round_pd): Ditto.
5638 (_mm512_mask_ceil_round_ps): Ditto.
5639 (_mm512_mask_ceil_round_pd): Ditto.
5640 (_mm512_maskz_floor_round_ps): Ditto.
5641 (_mm512_maskz_floor_round_pd): Ditto.
5642 (_mm512_maskz_ceil_round_ps): Ditto.
5643 (_mm512_maskz_ceil_round_pd): Ditto.
5644 (_mm512_expand_pd): Ditto.
5645 (_mm512_expand_ps): Ditto.
5646 * config/i386/i386.c (ix86_builtins): Remove
5647 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5648 (bdesc_args): Ditto.
5649 * config/i386/predicates.md (const1256_operand): New.
5650 (const_1_to_2_operand): Ditto.
5651 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5652 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5653 (*avx512pf_gatherpf<mode>sf): Ditto.
5654 (avx512pf_gatherpf<mode>df): Ditto.
5655 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5656 (*avx512pf_gatherpf<mode>df): Ditto.
5657 (avx512pf_scatterpf<mode>sf): Ditto.
5658 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5659 (*avx512pf_scatterpf<mode>sf): Ditto.
5660 (avx512pf_scatterpf<mode>df): Ditto.
5661 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5662 (*avx512pf_scatterpf<mode>df): Ditto.
5663 (avx512f_expand<mode>): Removed.
5664 (<shift_insn><mode>3<mask_name>): Change predicate type.
5665
5666 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5667
5668 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5669 not at the end of datarefs vector use ordered_remove to avoid
5670 reordering datarefs vector.
5671
5672 PR c/59984
5673 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5674 mark local addressable non-static vars as GOVD_PRIVATE
5675 instead of GOVD_LOCAL.
5676 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5677 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5678 than copying them.
5679
5680 PR middle-end/60092
5681 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5682 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5683 assume_aligned or alloc_align attributes.
5684 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5685 arguments. Handle also assume_aligned and alloc_align attributes.
5686 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5687 calls to functions with assume_aligned or alloc_align attributes.
5688 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5689
5690 2014-02-08 Terry Guo <terry.guo@arm.com>
5691
5692 * doc/invoke.texi: Document ARM -march=armv7e-m.
5693
5694 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5695
5696 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5697 flag on __cilkrts_rethrow builtin.
5698
5699 PR ipa/60026
5700 * ipa-cp.c (determine_versionability): Fail at -O0
5701 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5702 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5703
5704 Revert:
5705 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5706
5707 PR ipa/60026
5708 * tree-inline.c (copy_forbidden): Fail for
5709 __attribute__((optimize (0))) functions.
5710
5711 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5712
5713 * varpool.c: Include pointer-set.h.
5714 (varpool_remove_unreferenced_decls): Variables in other partitions
5715 will not be output; be however careful to not lose information
5716 about partitioning.
5717
5718 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5719
5720 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5721 lookup in the vtable constructor.
5722
5723 2014-02-07 Jeff Law <law@redhat.com>
5724
5725 PR target/40977
5726 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5727 define_insn_and_split.
5728
5729 * ipa-inline.c (inline_small_functions): Fix typos.
5730
5731 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5732
5733 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5734 (s390_can_use_return_insn): Declare.
5735 * config/s390/s390.h (EPILOGUE_USES): Define.
5736 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5737 instructions.
5738 (s390_chunkify_start): Handle return JUMP_LABELs.
5739 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5740 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5741 (s390_can_use_return_insn): New functions.
5742 (s390_fix_long_loop_prediction): Handle conditional returns.
5743 (TARGET_SET_UP_BY_PROLOGUE): Define.
5744 * config/s390/s390.md (ANY_RETURN): New code iterator.
5745 (*creturn, *csimple_return, return, simple_return): New patterns.
5746
5747 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5748
5749 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5750 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5751 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5752 REG_CFA_RESTORE list when deciding not to restore a register.
5753
5754 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5755
5756 * config/s390/s390.c: Include tree-pass.h and context.h.
5757 (s390_early_mach): New function, split out from...
5758 (s390_emit_prologue): ...here.
5759 (pass_data_s390_early_mach): New pass structure.
5760 (pass_s390_early_mach): New class.
5761 (s390_option_override): Create and register early_mach pass.
5762 Move to end of file.
5763
5764 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5765
5766 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5767 to match for the exit block.
5768
5769 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5770
5771 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5772 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5773 Reject misaligned operands.
5774
5775 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5776
5777 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5778
5779 2014-02-07 Richard Biener <rguenther@suse.de>
5780
5781 PR middle-end/60092
5782 * gimple-low.c (lower_builtin_posix_memalign): New function.
5783 (lower_stmt): Call it to lower posix_memalign in a way
5784 to make alignment info accessible.
5785
5786 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5787
5788 PR c++/60082
5789 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5790 __builtin_setjmp_receiver.
5791
5792 2014-02-07 Richard Biener <rguenther@suse.de>
5793
5794 PR middle-end/60092
5795 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5796 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5797 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5798 Handle BUILT_IN_POSIX_MEMALIGN.
5799 (find_func_clobbers): Likewise.
5800 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5801 (call_may_clobber_ref_p_1): Likewise.
5802
5803 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5804
5805 PR ipa/59918
5806 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5807 sanity check.
5808
5809 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5810
5811 PR ipa/59469
5812 * lto-cgraph.c (lto_output_node): Use
5813 symtab_get_symbol_partitioning_class.
5814 (lto_output_varpool_node): likewise.
5815 (symtab_get_symbol_partitioning_class): Move here from
5816 lto/lto-partition.c
5817 * cgraph.h (symbol_partitioning_class): Likewise.
5818 (symtab_get_symbol_partitioning_class): Declare.
5819
5820 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5821
5822 * ggc.h (ggc_internal_cleared_alloc): New macro.
5823 * vec.h (vec_safe_copy): Handle memory stats.
5824 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5825 * target-globals.c (save_target_globals): Likewise.
5826
5827 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5828
5829 PR target/60077
5830 * expr.c (emit_move_resolve_push): Export; be bit more selective
5831 on when to clear alias set.
5832 * expr.h (emit_move_resolve_push): Declare.
5833 * function.h (struct function): Add tail_call_marked.
5834 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5835 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5836 * config/i386/i386.md (TImode move expander): De not call
5837 ix86_expand_push.
5838 (FP push expanders): Preserve memory attributes.
5839 * config/i386/sse.md (push<mode>1): Remove.
5840 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5841 (ix86_expand_push): Remove.
5842 * config/i386/mmx.md (push<mode>1): Remove.
5843
5844 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5845
5846 PR rtl-optimization/60030
5847 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5848 lopart with paradoxical subreg before shifting it up by hprec.
5849
5850 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5851
5852 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5853 Remove extra newline at end of file.
5854 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5855 (arm_issue_rate): Handle cortexa57.
5856 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5857 (cortex-a57.cortex-a53): Likewise.
5858
5859 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5860
5861 PR target/59575
5862 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5863 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5864 bitmask.
5865 (arm_expand_prologue): Adjust all callers.
5866 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5867 info, registers also at the lowest numbered registers side. Use
5868 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5869 XEXP.
5870
5871 PR debug/59992
5872 * var-tracking.c (adjust_mems): Before adding a SET to
5873 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5874
5875 2014-02-06 Alan Modra <amodra@gmail.com>
5876
5877 PR target/60032
5878 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5879 change SDmode to DDmode when lra_in_progress.
5880
5881 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5882
5883 PR middle-end/59150
5884 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5885 free_data_ref on the dr first, and before goto again also set dr
5886 to the next dr. For simd_lane_access, free old datarefs[i] before
5887 overwriting it. For get_vectype_for_scalar_type failure, don't
5888 free_data_ref if simd_lane_access.
5889
5890 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5891
5892 PR target/60062
5893 * tree.h (opts_for_fn): New inline function.
5894 (opt_for_fn): Define.
5895 * config/i386/i386.c (ix86_function_regparm): Use
5896 opt_for_fn (decl, optimize) instead of optimize.
5897
5898 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5899
5900 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5901 for SYMBOL_REF in large memory model.
5902
5903 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5904
5905 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5906 and crypto support.
5907 (cortex-a57): Likewise.
5908 (cortex-a57.cortex-a53): Likewise.
5909
5910 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5911 Kugan Vivekanandarajah <kuganv@linaro.org>
5912
5913 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5914 unaligned_access.
5915 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5916
5917 2014-02-06 Richard Biener <rguenther@suse.de>
5918
5919 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5920 set_loop_copy and initialize_original_copy_tables.
5921
5922 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5923
5924 * config/aarch64/aarch64-simd.md
5925 (aarch64_ashr_simddi): Change QI to SI.
5926
5927 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5928 Jakub Jelinek <jakub@redhat.com>
5929
5930 PR middle-end/60013
5931 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5932 of the dataflow.
5933
5934 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5935
5936 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5937 CODE_FOR_altivec_vpku[hw]um to
5938 CODE_FOR_altivec_vpku[hw]um_direct.
5939 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5940 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5941 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5942 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5943
5944 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5945
5946 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5947 generation for -maltivec=be.
5948 (altivec_vsumsws): Simplify redundant test.
5949
5950 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5951
5952 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5953 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5954 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5955 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5956 gen_altivec_vpkuwum.
5957 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5958 BYTES_BIG_ENDIAN.
5959 (altivec_vpks<VI_char>ss): Likewise.
5960 (altivec_vpks<VI_char>us): Likewise.
5961 (altivec_vpku<VI_char>us): Likewise.
5962 (altivec_vpku<VI_char>um): Likewise.
5963 (altivec_vpku<VI_char>um_direct): New (copy of
5964 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5965 internal use).
5966 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5967 target is little endian and -maltivec=be is not specified.
5968 (*altivec_vupkhs<VU_char>_direct): New (copy of
5969 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5970 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5971 target is little endian and -maltivec=be is not specified.
5972 (*altivec_vupkls<VU_char>_direct): New (copy of
5973 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5974 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5975 little endian and -maltivec=be is not specified.
5976 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5977 little endian and -maltivec=be is not specified.
5978
5979 2014-02-05 Richard Henderson <rth@redhat.com>
5980
5981 PR debug/52727
5982 * combine-stack-adj.c: Revert r206943.
5983 * sched-int.h (struct deps_desc): Add last_args_size.
5984 * sched-deps.c (init_deps): Initialize it.
5985 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5986 contain REG_ARGS_SIZE notes.
5987
5988 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5989
5990 * lto-cgraph.c (asm_nodes_output): Make global.
5991 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5992 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5993 (driver_handle_option): Handle OPT_fwpa.
5994
5995 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5996
5997 PR ipa/59947
5998 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5999 a comment typo and formatting issue. If odr_hash hasn't been
6000 created, return vNULL and set *completep to false.
6001
6002 PR middle-end/57499
6003 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
6004 bb with no successors.
6005
6006 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
6007
6008 PR target/59718
6009 * doc/invoke.texi (-march): Clarify documentation for ARM.
6010 (-mtune): Likewise.
6011 (-mcpu): Likewise.
6012
6013 2014-02-05 Richard Biener <rguenther@suse.de>
6014
6015 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6016 when not vectorizing because of too many alias checks.
6017 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6018 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6019
6020 2014-02-05 Nick Clifton <nickc@redhat.com>
6021
6022 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6023 accept extended registers in any mode when compiling for the MN10300.
6024
6025 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6026
6027 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6028 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6029 sanitization attributes.
6030 (can_inline_edge_p): Likewise.
6031 (sanitize_attrs_match_for_inline_p): New function.
6032
6033 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6034
6035 * ipa-prop.c (detect_type_change): Shor circuit testing of
6036 type changes on THIS pointer.
6037
6038 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6039
6040 PR target/59777
6041 * config/pa/pa.c (legitimize_tls_address): Return original address
6042 if not passed a SYMBOL_REF rtx.
6043 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6044 addresses.
6045 (pa_emit_move_sequence): Simplify TLS source operands.
6046 (pa_legitimate_constant_p): Reject all TLS constants.
6047 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6048 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6049
6050 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6051
6052 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6053 groups when we know they are controlled by LTO.
6054 * varasm.c (default_binds_local_p_1): If object is in other partition,
6055 it will be resolved locally.
6056
6057 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6058
6059 * config/host-linux.c (linux_gt_pch_use_address): Don't
6060 use SSIZE_MAX because it is not always defined.
6061
6062 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6063
6064 PR bootstrap/59913
6065 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6066 threshold for pseudo splitting.
6067 (update_ebb_live_info): Process call argument hard registers and
6068 hard registers from insn definition too.
6069 (max_small_class_regs_num): New constant.
6070 (inherit_in_ebb): Update live hard regs through EBBs. Update
6071 reloads_num only for small register classes. Don't split for
6072 outputs of jumps.
6073
6074 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6075
6076 PR ipa/60058
6077 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6078 is non-null.
6079
6080 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6081
6082 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6083 visibility is safe.
6084
6085 2014-02-04 Marek Polacek <polacek@redhat.com>
6086
6087 * gdbinit.in (pel): Define.
6088
6089 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6090
6091 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6092 behavior.
6093
6094 2014-02-04 Richard Biener <rguenther@suse.de>
6095
6096 PR lto/59723
6097 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6098 in function context local.
6099 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6100 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6101 similar to LTO_imported_decl_ref.
6102
6103 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6104
6105 PR tree-optimization/60002
6106 * cgraphclones.c (build_function_decl_skip_args): Clear
6107 DECL_LANG_SPECIFIC.
6108
6109 PR tree-optimization/60023
6110 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6111 false to gsi_replace.
6112 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6113 has been in some EH region and vec_stmt could throw, add
6114 vec_stmt into the same EH region.
6115 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6116 has no lhs, ignore it.
6117 * internal-fn.c (expand_MASK_LOAD): Likewise.
6118
6119 PR ipa/60026
6120 * tree-inline.c (copy_forbidden): Fail for
6121 __attribute__((optimize (0))) functions.
6122
6123 PR other/58712
6124 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6125 is set, copy one less argument.
6126 (expand_simd_clones): Don't subtract clone_info->inbranch
6127 from simd_clone_struct_alloc argument.
6128
6129 PR rtl-optimization/57915
6130 * recog.c (simplify_while_replacing): If all unary/binary/relational
6131 operation arguments are constant, attempt to simplify those.
6132
6133 PR middle-end/59261
6134 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6135 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6136
6137 2014-02-04 Richard Biener <rguenther@suse.de>
6138
6139 PR tree-optimization/60012
6140 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6141 TBAA disambiguation to all DDRs.
6142
6143 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6144
6145 PR target/59788
6146 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6147 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6148
6149 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6150
6151 PR ipa/59882
6152 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6153
6154 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6155
6156 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6157 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6158
6159 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6160
6161 PR ipa/59831
6162 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6163 to figure out targets of polymorphic calls with known decl.
6164 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6165 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6166 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6167 (get_polymorphic_call_info): ... here.
6168 (get_polymorphic_call_info_from_invariant): New function.
6169
6170 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6171
6172 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6173 lookup via vtable pointer; check for type consistency
6174 and turn inconsitent facts into UNREACHABLE.
6175 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6176 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6177 type inconsistent querries; return UNREACHABLE instead.
6178
6179 2014-02-03 Richard Henderson <rth@twiddle.net>
6180
6181 PR tree-opt/59924
6182 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6183 already processed this node.
6184 (normalize_one_pred_1): Pass along mark_set.
6185 (normalize_one_pred): Create and destroy a pointer_set_t.
6186 (normalize_one_pred_chain): Likewise.
6187
6188 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6189
6190 PR gcov-profile/58602
6191 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6192
6193 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6194
6195 PR ipa/59831
6196 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6197 -fno-devirtualize; try to devirtualize by the knowledge of
6198 virtual table pointer given by aggregate propagation.
6199 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6200 (ipa_print_node_jump_functions): Dump also offset that
6201 is relevant for polymorphic calls.
6202 (determine_known_aggregate_parts): Add arg_type parameter; use it
6203 instead of determining the type from pointer type.
6204 (ipa_compute_jump_functions_for_edge): Update call of
6205 determine_known_aggregate_parts.
6206 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6207 (gimple_get_virt_method_for_binfo): ... here; simplify using
6208 vtable_pointer_value_to_vtable.
6209 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6210 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6211 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6212 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6213 POINTER_PLUS_EXPR.
6214 (vtable_pointer_value_to_binfo): ... here.
6215 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6216
6217 2014-02-03 Teresa Johnson <tejohnson@google.com>
6218
6219 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6220 redef of outer loop index variable.
6221
6222 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6223
6224 PR c++/53017
6225 PR c++/59211
6226 * doc/extend.texi (Function Attributes): Typo.
6227
6228 2014-02-03 Cong Hou <congh@google.com>
6229
6230 PR tree-optimization/60000
6231 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6232 if the vectorized statement is a store. A store statement can only
6233 appear at the end of pattern statements.
6234
6235 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6236
6237 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6238 (ix86_option_override_internal): Default long double to 64-bit for
6239 32-bit Bionic and to 128-bit for 64-bit Bionic.
6240
6241 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6242 TARGET_LONG_DOUBLE_128 is true.
6243 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6244
6245 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6246 (mlong-double-64): Negate -mlong-double-128.
6247 (mlong-double-128): New option.
6248
6249 * config/i386/i386-c.c (ix86_target_macros): Define
6250 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6251
6252 * doc/invoke.texi: Document -mlong-double-128.
6253
6254 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6255
6256 PR rtl-optimization/60024
6257 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6258
6259 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6260
6261 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6262
6263 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6264
6265 PR rtl-optimization/57662
6266 * sel-sched.c (code_motion_path_driver): Do not mark already not
6267 existing blocks in the visiting bitmap.
6268
6269 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6270
6271 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6272 on the insn being emitted.
6273
6274 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6275 Will Deacon <will.deacon@arm.com>
6276
6277 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6278
6279 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6280
6281 * config/arm/arm-tables.opt: Regenerate.
6282
6283 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6284
6285 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6286 for vector types other than V16QImode.
6287 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6288 define_expand, and call altivec_expand_vec_perm_le when producing
6289 code with little endian element order.
6290 (*altivec_vperm_<mode>_internal): New insn having previous
6291 behavior of altivec_vperm_<mode>.
6292 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6293 altivec_expand_vec_perm_le when producing code with little endian
6294 element order.
6295 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6296 behavior of altivec_vperm_<mode>_uns.
6297
6298 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6299
6300 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6301 (altivec_vsumsws): Add handling for -maltivec=be with a little
6302 endian target.
6303 (altivec_vsumsws_direct): New.
6304 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6305 gen_altivec_vsumsws.
6306
6307 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6308
6309 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6310 vtable_pointer_value_to_binfo): New functions.
6311 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6312 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6313
6314 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6315
6316 * config/nios2/nios2.md (load_got_register): Initialize GOT
6317 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6318 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6319
6320 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6321
6322 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6323 preserverd by passthrough, do not propagate the type.
6324
6325 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6326
6327 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6328 (mips_atomic_assign_expand_fenv): New function.
6329 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6330
6331 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6332
6333 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6334 (__builtin_mips_set_fcsr): Likewise.
6335 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6336 MIPS_USI_FTYPE_VOID.
6337 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6338 (mips16_expand_set_fcsr): Likewise.
6339 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6340 (mips16_set_fcsr_stub): Likewise.
6341 (mips16_get_fcsr_one_only_stub): New class.
6342 (mips16_set_fcsr_one_only_stub): Likewise.
6343 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6344 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6345 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6346 (hard_float): New availability predicate.
6347 (mips_builtins): Add get_fcsr and set_fcsr.
6348 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6349 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6350 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6351 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6352 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6353 patterns.
6354
6355 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6356
6357 * config/mips/mips.c (mips_one_only_stub): New class.
6358 (mips_need_mips16_rdhwr_p): Replace with...
6359 (mips16_rdhwr_stub): ...this new variable.
6360 (mips16_stub_call_address): New function.
6361 (mips16_rdhwr_one_only_stub): New class.
6362 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6363 (mips_output_mips16_rdhwr): Delete.
6364 (mips_finish_stub): New function.
6365 (mips_code_end): Use it to handle rdhwr stubs.
6366
6367 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6368
6369 PR target/60017
6370 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6371 when calculating size of integer atomic types.
6372
6373 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6374
6375 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6376
6377 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6378
6379 PR tree-optimization/60003
6380 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6381 * profile.c (branch_prob): Use gimple_call_builtin_p
6382 to check for BUILT_IN_SETJMP_RECEIVER.
6383 * tree-inline.c (copy_bb): Call notice_special_calls.
6384
6385 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6386
6387 PR bootstrap/59985
6388 * lra-constraints.c (process_alt_operands): Update reload_sum only
6389 on the first pass.
6390
6391 2014-01-31 Richard Henderson <rth@redhat.com>
6392
6393 PR middle-end/60004
6394 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6395 until after else_eh is processed.
6396
6397 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6398
6399 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6400 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6401 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6402 in smmintrin.h, remove them.
6403 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6404 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6405 * config/i386/i386.md (ROUND_SAE): Fix value.
6406 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6407 (const48_operand): New.
6408 * config/i386/subst.md (round), (round_expand): Use
6409 const_4_or_8_to_11_operand.
6410 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6411
6412 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6413
6414 * config/i386/constraints.md (Yk): Swap meaning with k.
6415 * config/i386/i386.md (movhi_internal): Change Yk to k.
6416 (movqi_internal): Ditto.
6417 (*k<logic><mode>): Ditto.
6418 (*andhi_1): Ditto.
6419 (*andqi_1): Ditto.
6420 (kandn<mode>): Ditto.
6421 (*<code>hi_1): Ditto.
6422 (*<code>qi_1): Ditto.
6423 (kxnor<mode>): Ditto.
6424 (kortestzhi): Ditto.
6425 (kortestchi): Ditto.
6426 (kunpckhi): Ditto.
6427 (*one_cmplhi2_1): Ditto.
6428 (*one_cmplqi2_1): Ditto.
6429 * config/i386/sse.md (): Change k to Yk.
6430 (avx512f_load<mode>_mask): Ditto.
6431 (avx512f_blendm<mode>): Ditto.
6432 (avx512f_store<mode>_mask): Ditto.
6433 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6434 (avx512f_storedqu<mode>_mask): Ditto.
6435 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6436 Ditto.
6437 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6438 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6439 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6440 (avx512f_maskcmp<mode>3): Ditto.
6441 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6442 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6443 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6444 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6445 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6446 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6447 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6448 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6449 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6450 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6451 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6452 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6453 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6454 (vec_extract_lo_<mode>_maskm): Ditto.
6455 (vec_extract_hi_<mode>_maskm): Ditto.
6456 (avx512f_vternlog<mode>_mask): Ditto.
6457 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6458 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6459 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6460 (avx512f_<code>v8div16qi2_mask): Ditto.
6461 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6462 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6463 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6464 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6465 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6466 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6467 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6468 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6469 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6470 (avx512cd_maskb_vec_dupv8di): Ditto.
6471 (avx512cd_maskw_vec_dupv16si): Ditto.
6472 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6473 (avx512f_vpermi2var<mode>3_mask): Ditto.
6474 (avx512f_vpermi2var<mode>3_mask): Ditto.
6475 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6476 (*avx512f_gathersi<mode>): Ditto.
6477 (*avx512f_gathersi<mode>_2): Ditto.
6478 (*avx512f_gatherdi<mode>): Ditto.
6479 (*avx512f_gatherdi<mode>_2): Ditto.
6480 (*avx512f_scattersi<mode>): Ditto.
6481 (*avx512f_scatterdi<mode>): Ditto.
6482 (avx512f_compress<mode>_mask): Ditto.
6483 (avx512f_compressstore<mode>_mask): Ditto.
6484 (avx512f_expand<mode>_mask): Ditto.
6485 * config/i386/subst.md (mask): Change k to Yk.
6486 (mask_scalar_merge): Ditto.
6487 (sd): Ditto.
6488
6489 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6490
6491 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6492
6493 2014-01-31 Richard Biener <rguenther@suse.de>
6494
6495 PR middle-end/59990
6496 * builtins.c (fold_builtin_memory_op): Make sure to not
6497 use a floating-point mode or a boolean or enumeral type for
6498 the copy operation.
6499
6500 2014-01-30 DJ Delorie <dj@redhat.com>
6501
6502 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6503 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6504 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6505 whenever main() has an epilogue.
6506
6507 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6508
6509 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6510 unused variable "field".
6511 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6512 (vsx_mergeh_<mode>): Likewise.
6513 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6514 (altivec_vmrghh): Likewise.
6515 (altivec_vmrghw): Likewise.
6516 (altivec_vmrglb): Likewise.
6517 (altivec_vmrglh): Likewise.
6518 (altivec_vmrglw): Likewise.
6519 (altivec_vspltb): Add missing uses.
6520 (altivec_vsplth): Likewise.
6521 (altivec_vspltw): Likewise.
6522 (altivec_vspltsf): Likewise.
6523
6524 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6525
6526 PR target/59923
6527 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6528 frame related instructions.
6529
6530 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6531
6532 PR rtl-optimization/59959
6533 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6534 any reload of register whose subreg is invalid.
6535
6536 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6537
6538 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6539 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6540 Add missing return type - void.
6541
6542 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6543
6544 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6545 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6546 remove element index adjustment for endian (now handled in vsx.md
6547 and altivec.md).
6548 (altivec_expand_vec_perm_const): Use
6549 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6550 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6551 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6552 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6553 define_expand and a new define_insn *altivec_vspltb_internal;
6554 adjust for -maltivec=be on a little endian target.
6555 (altivec_vspltb_direct): New.
6556 (altivec_vsplth): Divide into a define_expand and a new
6557 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6558 little endian target.
6559 (altivec_vsplth_direct): New.
6560 (altivec_vspltw): Divide into a define_expand and a new
6561 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6562 little endian target.
6563 (altivec_vspltw_direct): New.
6564 (altivec_vspltsf): Divide into a define_expand and a new
6565 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6566 a little endian target.
6567
6568 2014-01-30 Richard Biener <rguenther@suse.de>
6569
6570 PR tree-optimization/59993
6571 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6572 can propagate form the earlier stmt and avoid the transform
6573 when the intermediate result is needed.
6574
6575 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6576
6577 * README.Portability: Fix typo.
6578
6579 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6580
6581 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6582 comparison_operator with ordered_comparison_operator.
6583
6584 2014-01-30 Nick Clifton <nickc@redhat.com>
6585
6586 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6587 Rename to mn10300_store_multiple_regs.
6588 * config/mn10300/mn10300.c: Likewise.
6589 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6590 store_multiple_regs.
6591 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6592 Call mn10300_store_multiple_regs.
6593
6594 2014-01-30 Nick Clifton <nickc@redhat.com>
6595 DJ Delorie <dj@redhat.com>
6596
6597 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6598 %fp 2 to keep registers after it properly word-aligned.
6599 (rl78_alloc_physical_registers_umul): Handle the case where both
6600 input operands are the same.
6601
6602 2014-01-30 Richard Biener <rguenther@suse.de>
6603
6604 PR tree-optimization/59903
6605 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6606 check properly.
6607
6608 2014-01-30 Jason Merrill <jason@redhat.com>
6609
6610 PR c++/59633
6611 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6612
6613 PR c++/59645
6614 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6615
6616 2014-01-30 Richard Biener <rguenther@suse.de>
6617
6618 PR tree-optimization/59951
6619 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6620
6621 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6622
6623 PR target/59784
6624 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6625 SFmode to DFmode case.
6626
6627 2014-01-29 DJ Delorie <dj@redhat.com>
6628
6629 * config/msp430/msp430.opt (-minrt): New.
6630 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6631 if -minrt given.
6632 (ENDFILE_SPEC): Likewise.
6633
6634 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6635
6636 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6637 (estimate_function_body_sizes): Use it.
6638
6639 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6640
6641 PR c++/58561
6642 * dwarf2out.c (is_cxx_auto): New.
6643 (is_base_type): Use it.
6644 (gen_type_die_with_usage): Likewise.
6645
6646 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6647
6648 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6649 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6650 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6651 -maltivec=be with LE targets.
6652 (vsx_mergeh_<mode>): Likewise.
6653 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6654 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6655 (altivec_vmrghb): Replace with define_expand and new
6656 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6657 (altivec_vmrghb_direct): New define_insn.
6658 (altivec_vmrghh): Replace with define_expand and new
6659 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6660 (altivec_vmrghh_direct): New define_insn.
6661 (altivec_vmrghw): Replace with define_expand and new
6662 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6663 (altivec_vmrghw_direct): New define_insn.
6664 (*altivec_vmrghsf): Adjust for endianness.
6665 (altivec_vmrglb): Replace with define_expand and new
6666 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6667 (altivec_vmrglb_direct): New define_insn.
6668 (altivec_vmrglh): Replace with define_expand and new
6669 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6670 (altivec_vmrglh_direct): New define_insn.
6671 (altivec_vmrglw): Replace with define_expand and new
6672 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6673 (altivec_vmrglw_direct): New define_insn.
6674 (*altivec_vmrglsf): Adjust for endianness.
6675 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6676 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6677 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6678 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6679 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6680 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6681 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6682 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6683
6684 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6685
6686 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6687 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6688 whitespace.
6689
6690 2014-01-29 Richard Biener <rguenther@suse.de>
6691
6692 PR tree-optimization/58742
6693 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6694 associate_pointerplus_align.
6695 (associate_pointerplus_diff): New function.
6696 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6697 and associate_pointerplus_diff.
6698
6699 2014-01-29 Richard Biener <rguenther@suse.de>
6700
6701 * lto-streamer.h (LTO_major_version): Bump to 3.
6702 (LTO_minor_version): Reset to 0.
6703
6704 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6705
6706 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6707 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6708 (arm_file_start): Generate correct asm header for armv7ve.
6709 * config/arm/bpabi.h: Add multilib support for armv7ve.
6710 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6711 and cortex-a15 to armv7ve.
6712 * config/arm/t-aprofile: Add multilib support for armv7ve.
6713 * doc/invoke.texi: Document -march=armv7ve.
6714
6715 2014-01-29 Richard Biener <rguenther@suse.de>
6716
6717 PR tree-optimization/58742
6718 * tree-ssa-forwprop.c (associate_plusminus): Return true
6719 if we changed sth, defer EH cleanup to ...
6720 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6721 (simplify_mult): New function.
6722
6723 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6724
6725 PR middle-end/59917
6726 PR tree-optimization/59920
6727 * tree.c (build_common_builtin_nodes): Remove
6728 __builtin_setjmp_dispatcher initialization.
6729 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6730 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6731 instead of gsi_after_labels + manually skipping debug stmts.
6732 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6733 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6734 * tree-inline.c (copy_edges_for_bb): Remove
6735 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6736 argument. Ignore computed_goto_p stmts. Don't call
6737 make_abnormal_goto_edges. If a call might need abnormal edges
6738 for non-local gotos, see if it already has an edge to
6739 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6740 with true argument, don't do anything then, otherwise add
6741 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6742 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6743 caller.
6744 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6745 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6746 (lower_stmt): Don't set data->calls_builtin_setjmp.
6747 (lower_builtin_setjmp): Adjust comment.
6748 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6749 * tree-cfg.c (found_computed_goto): Remove.
6750 (factor_computed_gotos): Remove.
6751 (make_goto_expr_edges): Return bool, true for computed gotos.
6752 Don't call make_abnormal_goto_edges.
6753 (build_gimple_cfg): Don't set found_computed_goto, don't call
6754 factor_computed_gotos.
6755 (computed_goto_p): No longer static.
6756 (make_blocks): Don't set found_computed_goto.
6757 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6758 (make_edges): If make_goto_expr_edges returns true, push bb
6759 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6760 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6761 vector. Record mapping between bbs and OpenMP regions if there
6762 are any, adjust make_gimple_omp_edges caller. Call
6763 handle_abnormal_edges.
6764 (make_abnormal_goto_edges): Remove.
6765 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6766 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6767 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6768 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6769 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6770 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6771 filling *region also set *region_idx to (*region)->entry->index.
6772
6773 PR other/58712
6774 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6775 For REGs set ORIGINAL_REGNO.
6776
6777 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6778
6779 * doc/md.texi: Mention that a target shouldn't implement
6780 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6781 than hi/lo pair.
6782
6783 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6784
6785 PR tree-optimization/59594
6786 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6787 a copy of the datarefs vector rather than the vector itself.
6788
6789 2014-01-28 Jason Merrill <jason@redhat.com>
6790
6791 PR c++/53756
6792 * dwarf2out.c (auto_die): New static.
6793 (gen_type_die_with_usage): Handle C++1y 'auto'.
6794 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6795 on definition.
6796
6797 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6798
6799 PR target/59672
6800 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6801 (SPEC_X32): Likewise.
6802 (SPEC_64): Likewise.
6803 * config/i386/i386.c (ix86_option_override_internal): Turn off
6804 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6805 for TARGET_16BIT.
6806 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6807 * config/i386/i386.h (TARGET_16BIT): New macro.
6808 (TARGET_16BIT_P): Likewise.
6809 * config/i386/i386.opt: Add m16.
6810 * doc/invoke.texi: Document -m16.
6811
6812 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6813
6814 PR preprocessor/59935
6815 * input.c (location_get_source_line): Bail out on when line number
6816 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6817
6818 2014-01-28 Richard Biener <rguenther@suse.de>
6819
6820 PR tree-optimization/58742
6821 * tree-ssa-forwprop.c (associate_plusminus): Handle
6822 pointer subtraction of the form (T)(P + A) - (T)P.
6823
6824 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6825
6826 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6827 at const_int_cost.
6828
6829 2014-01-28 Richard Biener <rguenther@suse.de>
6830
6831 Revert
6832 2014-01-28 Richard Biener <rguenther@suse.de>
6833
6834 PR rtl-optimization/45364
6835 PR rtl-optimization/59890
6836 * var-tracking.c (local_get_addr_clear_given_value): Handle
6837 already cleared slot.
6838 (val_reset): Handle not allocated local_get_addr_cache.
6839 (vt_find_locations): Use post-order on the inverted CFG.
6840
6841 2014-01-28 Richard Biener <rguenther@suse.de>
6842
6843 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6844
6845 2014-01-28 Richard Biener <rguenther@suse.de>
6846
6847 PR rtl-optimization/45364
6848 PR rtl-optimization/59890
6849 * var-tracking.c (local_get_addr_clear_given_value): Handle
6850 already cleared slot.
6851 (val_reset): Handle not allocated local_get_addr_cache.
6852 (vt_find_locations): Use post-order on the inverted CFG.
6853
6854 2014-01-28 Alan Modra <amodra@gmail.com>
6855
6856 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6857 * configure.ac <recursive call for build != host>: Define
6858 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6859 and LD_FOR_BUILD too.
6860 * configure: Regenerate.
6861
6862 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6863
6864 * config/i386/i386.c (get_builtin_code_for_version): Separate
6865 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6866 Broadwell from Haswell.
6867
6868 2014-01-27 Steve Ellcey <sellcey@mips.com>
6869
6870 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6871 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6872 * config/mips/mips.c (mips_option_override): Change setting
6873 of TARGET_DSP.
6874 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6875 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6876 Change from Mask to Var.
6877
6878 2014-01-27 Jeff Law <law@redhat.com>
6879
6880 * ipa-inline.c (inline_small_functions): Fix typo.
6881
6882 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6883
6884 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6885 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6886 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6887 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6888 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6889 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6890 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6891 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6892 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6893 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6894 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6895 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6896 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6897 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6898 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6899 (_mm512_storeu_epi64): Ditto.
6900 (_mm512_cmpge_epi32_mask): Ditto.
6901 (_mm512_cmpge_epu32_mask): Ditto.
6902 (_mm512_cmpge_epi64_mask): Ditto.
6903 (_mm512_cmpge_epu64_mask): Ditto.
6904 (_mm512_cmple_epi32_mask): Ditto.
6905 (_mm512_cmple_epu32_mask): Ditto.
6906 (_mm512_cmple_epi64_mask): Ditto.
6907 (_mm512_cmple_epu64_mask): Ditto.
6908 (_mm512_cmplt_epi32_mask): Ditto.
6909 (_mm512_cmplt_epu32_mask): Ditto.
6910 (_mm512_cmplt_epi64_mask): Ditto.
6911 (_mm512_cmplt_epu64_mask): Ditto.
6912 (_mm512_cmpneq_epi32_mask): Ditto.
6913 (_mm512_cmpneq_epu32_mask): Ditto.
6914 (_mm512_cmpneq_epi64_mask): Ditto.
6915 (_mm512_cmpneq_epu64_mask): Ditto.
6916 (_mm512_expand_pd): Ditto.
6917 (_mm512_expand_ps): Ditto.
6918 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6919 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6920 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6921 * config/i386/i386.c (ix86_builtins): Add
6922 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6923 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6924 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6925 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6926 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6927 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6928 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6929 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6930 IX86_BUILTIN_PMOVUSQW512_MEM.
6931 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6932 __builtin_ia32_pmovsqd512mem_mask,
6933 __builtin_ia32_pmovqd512mem_mask,
6934 __builtin_ia32_pmovusqw512mem_mask,
6935 __builtin_ia32_pmovsqw512mem_mask,
6936 __builtin_ia32_pmovqw512mem_mask,
6937 __builtin_ia32_pmovusdw512mem_mask,
6938 __builtin_ia32_pmovsdw512mem_mask,
6939 __builtin_ia32_pmovdw512mem_mask,
6940 __builtin_ia32_pmovqb512mem_mask,
6941 __builtin_ia32_pmovusqb512mem_mask,
6942 __builtin_ia32_pmovsqb512mem_mask,
6943 __builtin_ia32_pmovusdb512mem_mask,
6944 __builtin_ia32_pmovsdb512mem_mask,
6945 __builtin_ia32_pmovdb512mem_mask.
6946 (bdesc_args): Add __builtin_ia32_expanddf512,
6947 __builtin_ia32_expandsf512.
6948 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6949 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6950 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6951 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6952 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6953 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6954 (avx512f_<code>v8div16qi2_mask_store): This.
6955 (avx512f_expand<mode>): New.
6956
6957 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6958
6959 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6960 New.
6961 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6962 (_mm512_prefetch_i32scatter_pd): Ditto.
6963 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6964 (_mm512_prefetch_i64scatter_pd): Ditto.
6965 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6966 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6967 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6968 (_mm512_prefetch_i32scatter_ps): Ditto.
6969 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6970 (_mm512_prefetch_i64scatter_ps): Ditto.
6971 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6972 * config/i386/i386-builtin-types.def: Define
6973 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6974 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6975 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6976 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6977 IX86_BUILTIN_SCATTERPFQPD.
6978 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6979 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6980 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6981 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6982 __builtin_ia32_scatterpfqps.
6983 (ix86_expand_builtin): Expand new built-ins.
6984 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6985 fix memory access data type.
6986 (*avx512pf_gatherpf<mode>_mask): Ditto.
6987 (*avx512pf_gatherpf<mode>): Ditto.
6988 (avx512pf_scatterpf<mode>): Ditto.
6989 (*avx512pf_scatterpf<mode>_mask): Ditto.
6990 (*avx512pf_scatterpf<mode>): Ditto.
6991 (GATHER_SCATTER_SF_MEM_MODE): New.
6992 (avx512pf_gatherpf<mode>df): Ditto.
6993 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6994 (*avx512pf_scatterpf<mode>df): Ditto.
6995
6996 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6997
6998 PR bootstrap/59934
6999 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
7000 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
7001 reached.
7002
7003 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7004
7005 * common/config/arm/arm-common.c
7006 (arm_rewrite_mcpu): Handle multiple names.
7007 * config/arm/arm.h
7008 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7009
7010 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
7011
7012 * gimple-builder.h (create_gimple_tmp): Delete.
7013
7014 2014-01-27 Christian Bruel <christian.bruel@st.com>
7015
7016 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7017 words comparisons.
7018
7019 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7020
7021 * config/pa/pa.md (call): Generate indirect long calls to non-local
7022 functions when outputing 32-bit code.
7023 (call_value): Likewise except for special call to buggy powf function.
7024
7025 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7026 portable runtime and PIC indirect calls.
7027 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7028 and PIC call sequences. Use ldo instead of blr to set return register
7029 in PIC call sequence.
7030
7031 2014-01-25 Walter Lee <walt@tilera.com>
7032
7033 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7034 avoid clobbering a live register.
7035
7036 2014-01-25 Walter Lee <walt@tilera.com>
7037
7038 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7039 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7040 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7041 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7042
7043 2014-01-25 Walter Lee <walt@tilera.com>
7044
7045 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7046 arguments on even registers.
7047 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7048 STACK_BOUNDARY.
7049 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7050 (BIGGEST_ALIGNMENT): Ditto.
7051 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7052
7053 2014-01-25 Walter Lee <walt@tilera.com>
7054
7055 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7056 insns before bundling.
7057 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7058
7059 2014-01-25 Walter Lee <walt@tilera.com>
7060
7061 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7062 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7063 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7064
7065 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7066
7067 * config/mips/constraints.md (kl): Delete.
7068 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7069 define expands, using...
7070 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7071 instructions for MIPS16.
7072 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7073 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7074
7075 2014-01-25 Walter Lee <walt@tilera.com>
7076
7077 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7078 (clzdi2): Ditto.
7079 (ffsdi2): Ditto.
7080
7081 2014-01-25 Walter Lee <walt@tilera.com>
7082
7083 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7084 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7085
7086 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7087
7088 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7089 Handle XOR.
7090
7091 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7092
7093 * print-rtl.c (in_call_function_usage): New var.
7094 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7095 EXPR_LIST mode as mode and not as reg note name.
7096
7097 PR middle-end/59561
7098 * cfgloopmanip.c (copy_loop_info): If
7099 loop->warned_aggressive_loop_optimizations, make sure
7100 the flag is set in target loop too.
7101
7102 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7103
7104 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7105 flag_cilkplus.
7106 * builtins.def: Likewise.
7107 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7108 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7109 * ira.c (ira_setup_eliminable_regset): Likewise.
7110 * omp-low.c (gate_expand_omp): Likewise.
7111 (execute_lower_omp): Likewise.
7112 (diagnose_sb_0): Likewise.
7113 (gate_diagnose_omp_blocks): Likewise.
7114 (simd_clone_clauses_extract): Likewise.
7115 (gate): Likewise.
7116
7117 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7118
7119 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7120 correction for little endian...
7121 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7122 here.
7123
7124 2014-01-24 Jeff Law <law@redhat.com>
7125
7126 PR tree-optimization/59919
7127 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7128 for non-returning calls.
7129
7130 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7131
7132 * common/config/aarch64/aarch64-common.c
7133 (aarch64_rewrite_mcpu): Handle multiple names.
7134 * config/aarch64/aarch64.h
7135 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7136
7137 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7138
7139 * input.c (add_file_to_cache_tab): Handle the case where fopen
7140 returns NULL.
7141
7142 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7143
7144 PR target/59929
7145 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7146 from push operand if code of push isn't PRE_DEC.
7147
7148 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7149
7150 PR target/59909
7151 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7152 -mquad-memory-atomic. Update -mquad-memory documentation to say
7153 it is only used for non-atomic loads/stores.
7154
7155 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7156 -mquad-memory or -mquad-memory-atomic switches.
7157
7158 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7159 -mquad-memory-atomic to ISA 2.07 support.
7160
7161 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7162 to separate support of normal quad word memory operations (ldq, stq)
7163 from the atomic quad word memory operations.
7164
7165 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7166 support to separate non-atomic quad word operations from atomic
7167 quad word operations. Disable non-atomic quad word operations in
7168 little endian mode so that we don't have to swap words after the
7169 load and before the store.
7170 (quad_load_store_p): Add comment about atomic quad word support.
7171 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7172 options printed with -mdebug=reg.
7173
7174 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7175 -mquad-memory-atomic as the test for whether we have quad word
7176 atomic instructions.
7177 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7178 or -mp8-vector are used, allow byte/half-word atomic operations.
7179
7180 * config/rs6000/sync.md (load_lockedti): Insure that the address
7181 is a proper indexed or indirect address for the lqarx instruction.
7182 On little endian systems, swap the hi/lo registers after the lqarx
7183 instruction.
7184 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7185 insure the address is valid for the lqarx instruction.
7186 (store_conditionalti): Insure that the address is a proper indexed
7187 or indirect address for the stqcrx. instruction. On little endian
7188 systems, swap the hi/lo registers before doing the stqcrx.
7189 instruction.
7190 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7191 insure the address is valid for the stqcrx. instruction.
7192
7193 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7194 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7195 type of quad memory support is available.
7196
7197 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7198
7199 PR regression/59915
7200 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7201 there is a danger of looping.
7202
7203 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7204
7205 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7206 force flag_ira_loop_pressure if set via command line.
7207
7208 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7209
7210 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7211 (ashr_simd): New builtin handling DI mode.
7212 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7213 (aarch64_sshr_simddi): New match pattern.
7214 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7215 (vshrd_n_s64): Likewise.
7216 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7217
7218 2014-01-23 Nick Clifton <nickc@redhat.com>
7219
7220 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7221 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7222 favour of mcu specific scripts.
7223 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7224 430x multilibs.
7225
7226 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7227 Alex Velenko <Alex.Velenko@arm.com>
7228
7229 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7230 (vaddv_s16): Likewise.
7231 (vaddv_s32): Likewise.
7232 (vaddv_u8): Likewise.
7233 (vaddv_u16): Likewise.
7234 (vaddv_u32): Likewise.
7235 (vaddvq_s8): Likewise.
7236 (vaddvq_s16): Likewise.
7237 (vaddvq_s32): Likewise.
7238 (vaddvq_s64): Likewise.
7239 (vaddvq_u8): Likewise.
7240 (vaddvq_u16): Likewise.
7241 (vaddvq_u32): Likewise.
7242 (vaddvq_u64): Likewise.
7243 (vaddv_f32): Likewise.
7244 (vaddvq_f32): Likewise.
7245 (vaddvq_f64): Likewise.
7246 (vmaxv_f32): Likewise.
7247 (vmaxv_s8): Likewise.
7248 (vmaxv_s16): Likewise.
7249 (vmaxv_s32): Likewise.
7250 (vmaxv_u8): Likewise.
7251 (vmaxv_u16): Likewise.
7252 (vmaxv_u32): Likewise.
7253 (vmaxvq_f32): Likewise.
7254 (vmaxvq_f64): Likewise.
7255 (vmaxvq_s8): Likewise.
7256 (vmaxvq_s16): Likewise.
7257 (vmaxvq_s32): Likewise.
7258 (vmaxvq_u8): Likewise.
7259 (vmaxvq_u16): Likewise.
7260 (vmaxvq_u32): Likewise.
7261 (vmaxnmv_f32): Likewise.
7262 (vmaxnmvq_f32): Likewise.
7263 (vmaxnmvq_f64): Likewise.
7264 (vminv_f32): Likewise.
7265 (vminv_s8): Likewise.
7266 (vminv_s16): Likewise.
7267 (vminv_s32): Likewise.
7268 (vminv_u8): Likewise.
7269 (vminv_u16): Likewise.
7270 (vminv_u32): Likewise.
7271 (vminvq_f32): Likewise.
7272 (vminvq_f64): Likewise.
7273 (vminvq_s8): Likewise.
7274 (vminvq_s16): Likewise.
7275 (vminvq_s32): Likewise.
7276 (vminvq_u8): Likewise.
7277 (vminvq_u16): Likewise.
7278 (vminvq_u32): Likewise.
7279 (vminnmv_f32): Likewise.
7280 (vminnmvq_f32): Likewise.
7281 (vminnmvq_f64): Likewise.
7282
7283 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7284
7285 * config/aarch64/aarch64-simd.md
7286 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7287 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7288 (*aarch64_mul3_elt<mode>): Likewise.
7289 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7290 (*aarch64_mul3_elt_to_64v2df): Likewise.
7291 (*aarch64_mla_elt<mode>): Likewise.
7292 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7293 (*aarch64_mls_elt<mode>): Likewise.
7294 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7295 (*aarch64_fma4_elt<mode>): Likewise.
7296 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7297 (*aarch64_fma4_elt_to_64v2df): Likewise.
7298 (*aarch64_fnma4_elt<mode>): Likewise.
7299 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7300 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7301 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7302 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7303 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7304 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7305 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7306 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7307 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7308
7309 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7310
7311 * config/aarch64/aarch64-simd.md
7312 (aarch64_be_checked_get_lane<mode>): New define_expand.
7313 * config/aarch64/aarch64-simd-builtins.def
7314 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7315 New builtin definition.
7316 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7317 Use new safe be builtin.
7318
7319 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7320
7321 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7322 New define_insn.
7323 (aarch64_be_st1<mode>): Likewise.
7324 (aarch_ld1<VALL:mode>): Define_expand modified.
7325 (aarch_st1<VALL:mode>): Likewise.
7326 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7327 (UNSPEC_ST1): Likewise.
7328
7329 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7330
7331 * config/microblaze/microblaze.md: Add trap insn and attribute
7332
7333 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7334
7335 PR preprocessor/58580
7336 * input.h (location_get_source_line): Take an additional line_size
7337 parameter.
7338 (void diagnostics_file_cache_fini): Declare new function.
7339 * input.c (struct fcache): New type.
7340 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7341 New static constants.
7342 (diagnostic_file_cache_init, total_lines_num)
7343 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7344 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7345 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7346 (get_next_line, read_next_line, goto_next_line, read_line_num):
7347 New static function definitions.
7348 (diagnostic_file_cache_fini): New function.
7349 (location_get_source_line): Take an additional output line_len
7350 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7351 read_line_num.
7352 * diagnostic.c (diagnostic_finish): Call
7353 diagnostic_file_cache_fini.
7354 (adjust_line): Take an additional input parameter for the length
7355 of the line, rather than calculating it with strlen.
7356 (diagnostic_show_locus): Adjust the use of
7357 location_get_source_line and adjust_line with respect to their new
7358 signature. While displaying a line now, do not stop at the first
7359 null byte. Rather, display the zero byte as a space and keep
7360 going until we reach the size of the line.
7361 * Makefile.in: Add vec.o to OBJS-libcommon
7362
7363 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7364 Ilya Tocar <ilya.tocar@intel.com>
7365
7366 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7367 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7368 (__builtin_ia32_kmov16): Ditto.
7369 * config/i386/i386.md (UNSPEC_KMOV): New.
7370 (kmovw): Ditto.
7371
7372 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7373
7374 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7375 (_mm512_storeu_si512): Ditto.
7376
7377 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7378
7379 PR target/52125
7380 * rtl.h (get_referenced_operands): Declare.
7381 * recog.c (get_referenced_operands): New function.
7382 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7383 operands have been referenced when recording LO_SUM references.
7384
7385 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7386
7387 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7388
7389 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7390
7391 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7392 Enable for generic and recent AMD targets.
7393
7394 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7395
7396 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7397 ARG_SIZE note when adjustment was eliminated.
7398
7399 2014-01-22 Jeff Law <law@redhat.com>
7400
7401 PR tree-optimization/59597
7402 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7403 in file. Accept new argument REGISTERING and use it to modify
7404 dump output appropriately.
7405 (register_jump_thread): Corresponding changes.
7406 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7407 thread paths involving joiner blocks. Add code to dump cancelled
7408 jump threading paths.
7409
7410 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7411
7412 PR rtl-optimization/59477
7413 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7414 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7415
7416 2014-01-22 Tom Tromey <tromey@redhat.com>
7417
7418 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7419 PARAMS.
7420 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7421
7422 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7423
7424 PR rtl-optimization/59896
7425 * lra-constraints.c (process_alt_operands): Check unused note for
7426 matched operands of insn with no output reloads.
7427
7428 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7429
7430 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7431 (mips_move_from_gpr_cost): Likewise.
7432
7433 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7434
7435 PR rtl-optimization/59858
7436 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7437 ira_class_hard_regs_num.
7438 (process_alt_operands): Increase reject for dying matched operand.
7439
7440 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7441
7442 PR target/59003
7443 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7444 smaller than size, perform several stores or loads and stores
7445 at dst + count - size to store or copy all of size bytes, rather
7446 than just last modesize bytes.
7447
7448 2014-01-20 DJ Delorie <dj@redhat.com>
7449
7450 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7451 that CLOBBERs are REGs before propogating their values.
7452
7453 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7454
7455 PR middle-end/59789
7456 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7457 (cgraph_inline_failed_type): New function.
7458 * cgraph.h (DEFCIFCODE): Add type.
7459 (cgraph_inline_failed_type_t): New enum.
7460 (cgraph_inline_failed_type): New prototype.
7461 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7462 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7463 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7464 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7465 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7466 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7467 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7468 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7469 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7470 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7471 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7472 OPTIMIZATION_MISMATCH.
7473 * tree-inline.c (expand_call_inline): Emit errors during
7474 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7475
7476 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7477
7478 PR target/59685
7479 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7480 mode attribute in insn output.
7481
7482 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7483
7484 * output.h (output_constant): Delete.
7485 * varasm.c (output_constant): Make private.
7486
7487 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7488
7489 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7490
7491 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7492
7493 PR middle-end/59860
7494 * tree.h (fold_builtin_strcat): New prototype.
7495 * builtins.c (fold_builtin_strcat): No longer static. Add len
7496 argument, if non-NULL, don't call c_strlen. Optimize
7497 directly into __builtin_memcpy instead of __builtin_strcpy.
7498 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7499 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7500
7501 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7502
7503 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7504 for SImode_address_operand operands, having only a REG argument.
7505
7506 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7507
7508 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7509 loader name using mbig-endian.
7510 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7511
7512 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7513
7514 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7515 (-mtune): Likewise.
7516 (-mcpu): Likewise.
7517
7518 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7519
7520 * config/aarch64/aarch64-protos.h
7521 (aarch64_cannot_change_mode_class_ptr): Declare.
7522 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7523 aarch64_cannot_change_mode_class_ptr): New.
7524 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7525 backend hook aarch64_cannot_change_mode_class.
7526
7527 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7528
7529 * common/config/aarch64/aarch64-common.c
7530 (aarch64_handle_option): Don't handle any option order logic here.
7531 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7532 selected_cpu, warn on architecture version mismatch.
7533 (aarch64_override_options): Fix parsing order for option strings.
7534
7535 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7536 Iain Sandoe <iain@codesourcery.com>
7537
7538 PR bootstrap/59496
7539 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7540 warning. Amend comment to reflect current functionality.
7541
7542 2014-01-20 Richard Biener <rguenther@suse.de>
7543
7544 PR middle-end/59860
7545 * builtins.c (fold_builtin_strcat): Remove case better handled
7546 by tree-ssa-strlen.c.
7547
7548 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7549
7550 * config/aarch64/aarch64.opt
7551 (mcpu, march, mtune): Make case-insensitive.
7552
7553 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7554
7555 PR target/59880
7556 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7557 if operands[1] is a REG or ZERO_EXTEND of a REG.
7558
7559 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7560
7561 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7562
7563 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7564
7565 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7566 long non-pic millicode calls.
7567
7568 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7569
7570 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7571
7572 2014-01-19 Kito Cheng <kito@0xlab.org>
7573
7574 * builtins.c (expand_movstr): Check movstr expand done or fail.
7575
7576 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7577 H.J. Lu <hongjiu.lu@intel.com>
7578
7579 PR target/59379
7580 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7581 to DImode for zero-extended addresses.
7582
7583 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7584
7585 PR rtl-optimization/57763
7586 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7587 on the new indirect jump_insn and increment LABEL_NUSES (label).
7588
7589 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7590
7591 PR bootstrap/59580
7592 PR bootstrap/59583
7593 * config.gcc (x86_archs): New variable.
7594 (x86_64_archs): Likewise.
7595 (x86_cpus): Likewise.
7596 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7597 --with-arch/--with-cpu= options.
7598 Support --with-arch=/--with-cpu={nehalem,westmere,
7599 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7600
7601 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7602
7603 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7604 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7605
7606 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7607
7608 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7609
7610 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7611
7612 PR target/58944
7613 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7614 clear cpp_get_options (parse_in)->warn_unused_macros for
7615 ix86_target_macros_internal with cpp_define.
7616
7617 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7618
7619 * jump.c (delete_related_insns): Keep (use (insn))s.
7620 * reorg.c (redundant_insn): Check for barriers too.
7621
7622 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7623
7624 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7625
7626 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7627
7628 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7629 call to $$dyncall when TARGET_LONG_CALLS is true.
7630
7631 2014-01-17 Jeff Law <law@redhat.com>
7632
7633 * ree.c (combine_set_extension): Temporarily disable test for
7634 changing number of hard registers.
7635
7636 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7637
7638 PR middle-end/58125
7639 * ipa-inline-analysis.c (inline_free_summary):
7640 Do not free summary of aliases.
7641
7642 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7643
7644 PR middle-end/59706
7645 * gimplify.c (gimplify_expr): Use create_tmp_var
7646 instead of create_tmp_var_raw. If cond doesn't have
7647 integral type, don't add the IFN_ANNOTATE builtin at all.
7648
7649 2014-01-17 Martin Jambor <mjambor@suse.cz>
7650
7651 PR ipa/59736
7652 * ipa-cp.c (prev_edge_clone): New variable.
7653 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7654 Also resize prev_edge_clone vector.
7655 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7656 (ipcp_edge_removal_hook): New function.
7657 (ipcp_driver): Register ipcp_edge_removal_hook.
7658
7659 2014-01-17 Andrew Pinski <apinski@cavium.com>
7660 Steve Ellcey <sellcey@mips.com>
7661
7662 PR target/59462
7663 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7664 of operator mode.
7665
7666 2014-01-17 Jeff Law <law@redhat.com>
7667
7668 PR middle-end/57904
7669 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7670 so that pass_ccp runs first.
7671
7672 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7673
7674 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7675 (ix86_adjust_cost): Use !TARGET_XXX.
7676 (do_reorder_for_imul): Likewise.
7677 (swap_top_of_ready_list): Likewise.
7678 (ix86_sched_reorder): Likewise.
7679
7680 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7681
7682 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7683 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7684 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7685 (intel_memset): New. Duplicate slm_memset.
7686 (intel_cost): New. Duplicate slm_cost.
7687 (m_INTEL): New macro.
7688 (processor_target_table): Add "intel".
7689 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7690 with PROCESSOR_INTEL for "intel".
7691 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7692 PROCESSOR_SILVERMONT.
7693 (ix86_issue_rate): Likewise.
7694 (ix86_adjust_cost): Likewise.
7695 (ia32_multipass_dfa_lookahead): Likewise.
7696 (swap_top_of_ready_list): Likewise.
7697 (ix86_sched_reorder): Likewise.
7698 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7699 instead of TARGET_OPT_AGU.
7700 * config/i386/i386.h (TARGET_INTEL): New.
7701 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7702 (processor_type): Add PROCESSOR_INTEL.
7703 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7704 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7705
7706 2014-01-17 Marek Polacek <polacek@redhat.com>
7707
7708 PR c/58346
7709 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7710 size is zero.
7711
7712 2014-01-17 Richard Biener <rguenther@suse.de>
7713
7714 PR tree-optimization/46590
7715 * opts.c (default_options_table): Add entries for
7716 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7717 all enabled at -O1 but not for -Og.
7718 * common.opt (fbranch-count-reg): Remove Init(1).
7719 (fmove-loop-invariants): Likewise.
7720 (ftree-pta): Likewise.
7721
7722 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7723
7724 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7725 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7726 decls to at least the GCC 4.8 used alignments.
7727
7728 PR fortran/59440
7729 * tree-nested.c (convert_nonlocal_reference_stmt,
7730 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7731 of GIMPLE_BIND stmts, adjust associated decls.
7732
7733 2014-01-17 Richard Biener <rguenther@suse.de>
7734
7735 PR tree-optimization/46590
7736 * vec.h (vec<>::bseach): New member function implementing
7737 binary search according to C89 bsearch.
7738 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7739 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7740 bitmap pointer again. Make accesses_in_loop a flat array.
7741 (mem_ref_obstack): New global.
7742 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7743 (mark_ref_stored): Likewise.
7744 (ref_indep_loop_p_2): Likewise.
7745 (set_ref_stored_in_loop): New helper function.
7746 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7747 (memref_free): Adjust.
7748 (record_mem_ref_loc): Simplify.
7749 (gather_mem_refs_stmt): Adjust.
7750 (sort_locs_in_loop_postorder_cmp): New function.
7751 (analyze_memory_references): Sort accesses_in_loop after
7752 loop postorder number.
7753 (find_ref_loc_in_loop_cmp): New function.
7754 (for_all_locs_in_loop): Find relevant cluster of locs in
7755 accesses_in_loop and iterate without recursion.
7756 (execute_sm): Avoid uninit warning.
7757 (struct ref_always_accessed): Simplify.
7758 (ref_always_accessed::operator ()): Likewise.
7759 (ref_always_accessed_p): Likewise.
7760 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7761 loop postorder numbers here.
7762 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7763 numbers.
7764
7765 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7766
7767 PR c++/57945
7768 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7769 on decls for which assemble_alias has been called.
7770
7771 2014-01-17 Nick Clifton <nickc@redhat.com>
7772
7773 * config/msp430/msp430.opt: (mcpu): New option.
7774 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7775 (msp430_option_override): Parse target_cpu. If the MCU name
7776 matches a generic string, clear target_mcu.
7777 (msp430_attr): Allow numeric interrupt values up to 63.
7778 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7779 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7780 option.
7781 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7782 Add mcpu matches.
7783 * config/msp430/msp430.md (popm): Use %J rather than %I.
7784 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7785 (addhi_cy_i): Use immediate_operand for operand 2.
7786 * doc/invoke.texi: Document -mcpu option.
7787
7788 2014-01-17 Richard Biener <rguenther@suse.de>
7789
7790 PR rtl-optimization/38518
7791 * df.h (df_analyze_loop): Declare.
7792 * df-core.c: Include cfgloop.h.
7793 (df_analyze_1): Split out main part of df_analyze.
7794 (df_analyze): Adjust.
7795 (loop_inverted_post_order_compute): New function.
7796 (loop_post_order_compute): Likewise.
7797 (df_analyze_loop): New function avoiding whole-function
7798 postorder computes.
7799 * loop-invariant.c (find_defs): Use df_analyze_loop.
7800 (find_invariants): Adjust.
7801 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7802
7803 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7804
7805 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7806 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7807
7808 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7809
7810 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7811 traversal when removing references.
7812
7813 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7814
7815 PR ipa/59775
7816 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7817
7818 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7819
7820 PR middle-end/56791
7821 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7822 pushing a reload for an autoinc when we had previously reloaded an
7823 inner part of the address.
7824
7825 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7826
7827 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7828 field.
7829 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7830 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7831 when not giving up or versioning for alias only because of
7832 loop->safelen.
7833 (vect_analyze_data_ref_dependences): Set to true.
7834 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7835 is a GIMPLE_PHI.
7836 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7837 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7838 to the condition.
7839
7840 PR middle-end/58344
7841 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7842
7843 PR target/59839
7844 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7845 operand 0 predicate for gathers, use a new pseudo as subtarget.
7846
7847 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7848
7849 PR middle-end/59609
7850 * lra-constraints.c (process_alt_operands): Add printing debug info.
7851 Check absence of input/output reloads for matched operands too.
7852
7853 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7854
7855 PR rtl-optimization/59835
7856 * ira.c (ira_init_register_move_cost): Increase cost for
7857 impossible modes.
7858
7859 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7860
7861 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7862
7863 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7864
7865 PR target/59780
7866 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7867 non-register objects. Use gen_(high/low)part more consistently.
7868 Fix assertions.
7869
7870 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7871
7872 PR target/59844
7873 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7874 endian support, remove tests for WORDS_BIG_ENDIAN.
7875 (p8_mfvsrd_3_<mode>): Likewise.
7876 (reload_gpr_from_vsx<mode>): Likewise.
7877 (reload_gpr_from_vsxsf): Likewise.
7878 (p8_mfvsrd_4_disf): Likewise.
7879
7880 2014-01-16 Richard Biener <rguenther@suse.de>
7881
7882 PR rtl-optimization/46590
7883 * lcm.c (compute_antinout_edge): Use postorder iteration.
7884 (compute_laterin): Use inverted postorder iteration.
7885
7886 2014-01-16 Nick Clifton <nickc@redhat.com>
7887
7888 PR middle-end/28865
7889 * varasm.c (output_constant): Return the number of bytes actually
7890 emitted.
7891 (output_constructor_array_range): Update the field size with the
7892 number of bytes emitted by output_constant.
7893 (output_constructor_regular_field): Likewise. Also do not
7894 complain if the total number of bytes emitted is now greater
7895 than the expected fieldpos.
7896 * output.h (output_constant): Update prototype and descriptive comment.
7897
7898 2014-01-16 Marek Polacek <polacek@redhat.com>
7899
7900 PR middle-end/59827
7901 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7902 it is error_mark_node.
7903
7904 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7905
7906 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7907 VALID_AVX256_REG_OR_OI_MODE.
7908
7909 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7910
7911 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7912 current procedure should be profiled.
7913
7914 2014-01-15 Andrew Pinski <apinski@cavium.com>
7915
7916 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7917 of moving from/to the STACK_REG register class.
7918
7919 2014-01-15 Richard Henderson <rth@redhat.com>
7920
7921 PR debug/54694
7922 * reginfo.c (global_regs_decl): Globalize.
7923 * rtl.h (global_regs_decl): Declare.
7924 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7925 reserved via global_regs.
7926
7927 2014-01-15 Teresa Johnson <tejohnson@google.com>
7928
7929 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7930
7931 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7932
7933 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7934 and vmulosh rather than call gen_vec_widen_smult_*.
7935 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7936 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7937 (vec_widen_smult_even_v16qi): Likewise.
7938 (vec_widen_umult_even_v8hi): Likewise.
7939 (vec_widen_smult_even_v8hi): Likewise.
7940 (vec_widen_umult_odd_v16qi): Likewise.
7941 (vec_widen_smult_odd_v16qi): Likewise.
7942 (vec_widen_umult_odd_v8hi): Likewise.
7943 (vec_widen_smult_odd_v8hi): Likewise.
7944 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7945 vmuloub rather than call gen_vec_widen_umult_*.
7946 (vec_widen_umult_lo_v16qi): Likewise.
7947 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7948 vmulosb rather than call gen_vec_widen_smult_*.
7949 (vec_widen_smult_lo_v16qi): Likewise.
7950 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7951 rather than call gen_vec_widen_umult_*.
7952 (vec_widen_umult_lo_v8hi): Likewise.
7953 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7954 rather than call gen_vec_widen_smult_*.
7955 (vec_widen_smult_lo_v8hi): Likewise.
7956
7957 2014-01-15 Jeff Law <law@redhat.com>
7958
7959 PR tree-optimization/59747
7960 * ree.c (find_and_remove_re): Properly handle case where a second
7961 eliminated extension requires widening a copy created for elimination
7962 of a prior extension.
7963 (combine_set_extension): Ensure that the number of hard regs needed
7964 for a destination register does not change when we widen it.
7965
7966 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7967
7968 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7969 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7970 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7971 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7972 (avr-*-rtems*): Likewise.
7973 (bfin*-rtems*): Likewise.
7974 (moxie-*-rtems*): Likewise.
7975 (h8300-*-rtems*): Likewise.
7976 (i[34567]86-*-rtems*): Likewise.
7977 (lm32-*-rtems*): Likewise.
7978 (m32r-*-rtems*): Likewise.
7979 (m68k-*-rtems*): Likewise.
7980 (microblaze*-*-rtems*): Likewise.
7981 (mips*-*-rtems*): Likewise.
7982 (powerpc-*-rtems*): Likewise.
7983 (sh-*-rtems*): Likewise.
7984 (sparc-*-rtems*): Likewise.
7985 (sparc64-*-rtems*): Likewise.
7986 (v850-*-rtems*): Likewise.
7987 (m32c-*-rtems*): Likewise.
7988
7989 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7990
7991 PR rtl-optimization/59511
7992 * ira.c (ira_init_register_move_cost): Use memory costs for some
7993 cases of register move cost calculations.
7994 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7995 instead of BB frequency.
7996 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7997 * lra-assigns.c (find_hard_regno_for): Ditto.
7998
7999 2014-01-15 Richard Biener <rguenther@suse.de>
8000
8001 PR tree-optimization/59822
8002 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
8003 (vectorizable_load): Use it to hoist defs of uses of invariant
8004 loads out of the loop.
8005
8006 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8007 Kugan Vivekanandarajah <kuganv@linaro.org>
8008
8009 PR target/59695
8010 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
8011 truncation.
8012
8013 2014-01-15 Richard Biener <rguenther@suse.de>
8014
8015 PR rtl-optimization/59802
8016 * lcm.c (compute_available): Use inverted postorder to seed
8017 the initial worklist.
8018
8019 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8020
8021 PR target/59803
8022 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8023 ADDR_REGS for invalid symrefs in non-PIC code.
8024
8025 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8026
8027 PR other/58712
8028 * builtins.c (determine_block_size): Initialize *probable_max_size
8029 even if len_rtx is CONST_INT.
8030
8031 2014-01-14 Andrew Pinski <apinski@cavium.com>
8032
8033 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8034 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8035 (cortexa53_tunings): Likewise.
8036 (aarch64_sched_issue_rate): New function.
8037 (TARGET_SCHED_ISSUE_RATE): Define.
8038
8039 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8040
8041 * ira-costs.c (find_costs_and_classes): Add missed
8042 ira_init_register_move_cost_if_necessary.
8043
8044 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8045
8046 PR target/59787
8047 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8048
8049 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8050
8051 PR target/59794
8052 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8053 to indicate if type is used for function return value. Warn ABI
8054 change if the vector mode isn't available for function return value.
8055 (ix86_function_arg_advance): Pass false to type_natural_mode.
8056 (ix86_function_arg): Likewise.
8057 (ix86_gimplify_va_arg): Likewise.
8058 (function_arg_32): Don't warn ABI change.
8059 (ix86_function_value): Pass true to type_natural_mode.
8060 (ix86_return_in_memory): Likewise.
8061 (ix86_struct_value_rtx): Removed.
8062 (TARGET_STRUCT_VALUE_RTX): Likewise.
8063
8064 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8065
8066 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8067 converting a conditional jump into a conditional return.
8068
8069 2014-01-14 Richard Biener <rguenther@suse.de>
8070
8071 PR tree-optimization/58921
8072 PR tree-optimization/59006
8073 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8074 hoisting invariant stmts.
8075 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8076 invariant loads on the preheader edge if possible.
8077
8078 2014-01-14 Joey Ye <joey.ye@arm.com>
8079
8080 * doc/plugin.texi (Building GCC plugins): Update to C++.
8081
8082 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8083
8084 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8085 (_mm_rcp28_round_ss): Ditto.
8086 (_mm_rsqrt28_round_sd): Ditto.
8087 (_mm_rsqrt28_round_ss): Ditto.
8088 (_mm_rcp28_sd): Ditto.
8089 (_mm_rcp28_ss): Ditto.
8090 (_mm_rsqrt28_sd): Ditto.
8091 (_mm_rsqrt28_ss): Ditto.
8092 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8093 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8094 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8095 (IX86_BUILTIN_RCP28SD): Ditto.
8096 (IX86_BUILTIN_RCP28SS): Ditto.
8097 (IX86_BUILTIN_RSQRT28SD): Ditto.
8098 (IX86_BUILTIN_RSQRT28SS): Ditto.
8099 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8100 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8101 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8102 (ix86_expand_special_args_builtin): Expand new FTYPE.
8103 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8104 (srcp14<mode>): Make insn unary.
8105 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8106 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8107 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8108 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8109 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8110 Fix rounding: make it SAE only.
8111 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8112 Ditto.
8113 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8114 Ditto.
8115 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8116 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8117 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8118 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8119 (round_saeonly_mask_scalar_operand4): Ditto.
8120 (round_saeonly_mask_scalar_op3): Ditto.
8121 (round_saeonly_mask_scalar_op4): Ditto.
8122
8123 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8124
8125 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8126 Implement -maltivec=be for vec_insert and vec_extract.
8127
8128 2014-01-10 DJ Delorie <dj@redhat.com>
8129
8130 * config/msp430/msp430.md (call_internal): Don't allow memory
8131 references with SP as the base register.
8132 (call_value_internal): Likewise.
8133 * config/msp430/constraints.md (Yc): New. For memory references
8134 that don't use SP as a base register.
8135
8136 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8137 "an integer without a # prefix"
8138 * config/msp430/msp430.md (epilogue_helper): Use it.
8139
8140 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8141
8142 PR target/59617
8143 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8144 AVX512F gather builtins.
8145 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8146 on gather decls with INTEGER_TYPE masktype.
8147 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8148 directly into the builtin rather than hoisting it before loop.
8149
8150 PR tree-optimization/59387
8151 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8152 (scev_const_prop): If folded_casts and type has undefined overflow,
8153 use force_gimple_operand instead of force_gimple_operand_gsi and
8154 for each added stmt if it is assign with
8155 arith_code_with_undefined_signed_overflow, call
8156 rewrite_to_defined_overflow.
8157 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8158 gimple-fold.h instead.
8159 (arith_code_with_undefined_signed_overflow,
8160 rewrite_to_defined_overflow): Moved to ...
8161 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8162 rewrite_to_defined_overflow): ... here. No longer static.
8163 Include gimplify-me.h.
8164 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8165 rewrite_to_defined_overflow): New prototypes.
8166
8167 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8168
8169 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8170
8171 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8172
8173 * builtins.c (get_object_alignment_2): Minor tweak.
8174 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8175
8176 2014-01-13 Christian Bruel <christian.bruel@st.com>
8177
8178 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8179 optimized non constant lengths.
8180
8181 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8182
8183 PR libgomp/59194
8184 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8185 load as __atomic_load_N if possible.
8186
8187 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8188
8189 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8190 target parameter.
8191 (rs6000_expand_builtin): Adjust call.
8192
8193 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8194
8195 PR target/58115
8196 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8197 * config/rs6000/rs6000.c: Include target-globals.h.
8198 (rs6000_set_current_function): Instead of doing target_reinit
8199 unconditionally, use save_target_globals_default_opts and
8200 restore_target_globals.
8201
8202 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8203 FPSCR.
8204 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8205 (rs6000_expand_builtin): Handle mffs and mtfsf.
8206 (rs6000_init_builtins): Define mffs and mtfsf.
8207 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8208 (rs6000_mffs): New pattern.
8209 (rs6000_mtfsf): New pattern.
8210
8211 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8212
8213 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8214 Start narrowing with START. Apply candidate-use pair
8215 and check overall cost in narrowing.
8216 (iv_ca_prune): Pass new argument.
8217
8218 2014-01-10 Jeff Law <law@redhat.com>
8219
8220 PR middle-end/59743
8221 * ree.c (combine_reaching_defs): Ensure the defining statement
8222 occurs before the extension when optimizing extensions with
8223 different source and destination hard registers.
8224
8225 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8226
8227 PR ipa/58585
8228 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8229 vtables into the type inheritance graph.
8230
8231 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8232
8233 PR rtl-optimization/59754
8234 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8235 modes in the REGNO != REGNO case.
8236
8237 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8238
8239 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8240
8241 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8242
8243 PR tree-optimization/59745
8244 * tree-predcom.c (tree_predictive_commoning_loop): Call
8245 free_affine_expand_cache if giving up because components is NULL.
8246
8247 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8248 GC in payload of target_globals struct instead of allocating them on
8249 the heap and the larger structs separately using GC.
8250 * target-globals.h (struct target_globals): Make regs, hard_regs,
8251 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8252 of GTY((skip)) and change type to void *.
8253 (reset_target_globals): Cast loads from those fields to corresponding
8254 types.
8255
8256 2014-01-10 Steve Ellcey <sellcey@mips.com>
8257
8258 PR plugins/59335
8259 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8260 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8261 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8262
8263 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8264
8265 PR target/59744
8266 * aarch64-modes.def (CC_Zmode): New flags mode.
8267 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8268 represents an equality.
8269 (aarch64_get_condition_code): Handle CC_Zmode.
8270 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8271
8272 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8273
8274 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8275 extraction in good case.
8276
8277 2014-01-10 Richard Biener <rguenther@suse.de>
8278
8279 PR tree-optimization/59374
8280 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8281 checking after SLP discovery. Mark stmts not participating
8282 in any SLP instance properly.
8283
8284 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8285
8286 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8287 when handling a SET rtx.
8288
8289 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8290
8291 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8292 (cortex-a57): Likewise.
8293 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8294
8295 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8296
8297 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8298 non-iwmmxt builtins.
8299
8300 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8301
8302 PR ipa/58252
8303 PR ipa/59226
8304 * ipa-devirt.c record_target_from_binfo): Take as argument
8305 stack of binfos and lookup matching one for virtual inheritance.
8306 (possible_polymorphic_call_targets_1): Update.
8307
8308 2014-01-10 Huacai Chen <chenhc@lemote.com>
8309
8310 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8311 kernel strings for Loongson-2E/2F/3A.
8312
8313 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8314
8315 PR middle-end/59670
8316 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8317 is_gimple_call before calling gimple_call_internal_p.
8318
8319 2014-01-09 Steve Ellcey <sellcey@mips.com>
8320
8321 * Makefile.in (TREE_FLOW_H): Remove.
8322 (TREE_SSA_H): Add file names from tree-flow.h.
8323 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8324 * tree.h: Remove tree-flow.h reference.
8325 * hash-table.h: Remove tree-flow.h reference.
8326 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8327 reference with tree-ssa-loop.h.
8328
8329 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8330
8331 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8332 default element-order behavior for -maltivec.
8333 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8334 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8335 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8336 when targeting big endian, at least for now.
8337 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8338
8339 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8340
8341 PR middle-end/47735
8342 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8343 var satisfies use_register_for_decl, just take into account type
8344 alignment, rather than decl alignment.
8345
8346 PR tree-optimization/59622
8347 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8348 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8349 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8350 Don't devirtualize for inplace at all. For targets.length () == 1,
8351 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8352
8353 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8354
8355 * config/i386/i386.md (cpu): Remove the unused btver1.
8356
8357 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8358
8359 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8360
8361 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8362
8363 PR target/58115
8364 * tree-core.h (struct target_globals): New forward declaration.
8365 (struct tree_target_option): Add globals field.
8366 * tree.h (TREE_TARGET_GLOBALS): Define.
8367 (prepare_target_option_nodes_for_pch): New prototype.
8368 * target-globals.h (struct target_globals): Define even if
8369 !SWITCHABLE_TARGET.
8370 * tree.c (prepare_target_option_node_for_pch,
8371 prepare_target_option_nodes_for_pch): New functions.
8372 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8373 * config/i386/i386.c: Include target-globals.h.
8374 (ix86_set_current_function): Instead of doing target_reinit
8375 unconditionally, use save_target_globals_default_opts and
8376 restore_target_globals.
8377
8378 2014-01-09 Richard Biener <rguenther@suse.de>
8379
8380 PR tree-optimization/59715
8381 * tree-cfg.h (split_critical_edges): Declare.
8382 * tree-cfg.c (split_critical_edges): Export.
8383 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8384
8385 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8386
8387 * cfgexpand.c (expand_stack_vars): Optionally disable
8388 asan stack protection.
8389 (expand_used_vars): Likewise.
8390 (partition_stack_vars): Likewise.
8391 * asan.c (asan_emit_stack_protection): Optionally disable
8392 after return stack usage.
8393 (instrument_derefs): Optionally disable memory access instrumentation.
8394 (instrument_builtin_call): Likewise.
8395 (instrument_strlen_call): Likewise.
8396 (asan_protect_global): Optionally disable global variables protection.
8397 * doc/invoke.texi: Added doc for new options.
8398 * params.def: Added new options.
8399 * params.h: Likewise.
8400
8401 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8402
8403 PR rtl-optimization/59724
8404 * ifcvt.c (cond_exec_process_if_block): Don't call
8405 flow_find_head_matching_sequence with 0 longest_match.
8406 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8407 non-active insns if !stop_after.
8408 (try_head_merge_bb): Revert 2014-01-07 changes.
8409
8410 2014-01-08 Jeff Law <law@redhat.com>
8411
8412 * ree.c (get_sub_rtx): New function, extracted from...
8413 (merge_def_and_ext): Here.
8414 (combine_reaching_defs): Use get_sub_rtx.
8415
8416 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8417
8418 * cgraph.h (varpool_variable_node): Do not choke on null node.
8419
8420 2014-01-08 Catherine Moore <clm@codesourcery.com>
8421
8422 * config/mips/mips.md (simple_return): Attempt to use JRC
8423 for microMIPS.
8424 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8425
8426 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8427
8428 PR rtl-optimization/59137
8429 * reorg.c (steal_delay_list_from_target): Call update_block for
8430 elided insns.
8431 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8432
8433 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8434
8435 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8436 two duplicate entries.
8437
8438 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8439
8440 Revert:
8441 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8442
8443 * config/mips/mips.c (mips_truncated_op_cost): New function.
8444 (mips_rtx_costs): Adjust test for BADDU.
8445 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8446
8447 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8448
8449 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8450 (*baddu_si): ...this new pattern.
8451
8452 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8453
8454 PR ipa/59722
8455 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8456
8457 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8458
8459 PR middle-end/57748
8460 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8461 inner_reference_p.
8462 (expand_expr, expand_normal): Adjust.
8463 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8464 inner_reference_p. Use inner_reference_p to expand inner references.
8465 (store_expr): Adjust.
8466 * cfgexpand.c (expand_call_stmt): Adjust.
8467
8468 2014-01-08 Rong Xu <xur@google.com>
8469
8470 * gcov-io.c (gcov_var): Move from gcov-io.h.
8471 (gcov_position): Ditto.
8472 (gcov_is_error): Ditto.
8473 (gcov_rewrite): Ditto.
8474 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8475 only part to libgcc/libgcov.h.
8476
8477 2014-01-08 Marek Polacek <polacek@redhat.com>
8478
8479 PR middle-end/59669
8480 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8481
8482 2014-01-08 Marek Polacek <polacek@redhat.com>
8483
8484 PR sanitizer/59667
8485 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8486
8487 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8488
8489 PR rtl-optimization/59649
8490 * stor-layout.c (get_mode_bounds): For BImode return
8491 0 and STORE_FLAG_VALUE.
8492
8493 2014-01-08 Richard Biener <rguenther@suse.de>
8494
8495 PR middle-end/59630
8496 * gimple.h (is_gimple_builtin_call): Remove.
8497 (gimple_builtin_call_types_compatible_p): New.
8498 (gimple_call_builtin_p): New overload.
8499 * gimple.c (is_gimple_builtin_call): Remove.
8500 (validate_call): Rename to ...
8501 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8502 check return types.
8503 (validate_type): New static function.
8504 (gimple_call_builtin_p): New overload and adjust.
8505 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8506 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8507 (gimple_fold_stmt_to_constant_1): Likewise.
8508 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8509
8510 2014-01-08 Richard Biener <rguenther@suse.de>
8511
8512 PR middle-end/59471
8513 * gimplify.c (gimplify_expr): Gimplify register-register type
8514 VIEW_CONVERT_EXPRs to separate stmts.
8515
8516 2014-01-07 Jeff Law <law@redhat.com>
8517
8518 PR middle-end/53623
8519 * ree.c (combine_set_extension): Handle case where source
8520 and destination registers in an extension insn are different.
8521 (combine_reaching_defs): Allow source and destination registers
8522 in extension to be different under limited circumstances.
8523 (add_removable_extension): Remove restriction that the
8524 source and destination registers in the extension are the same.
8525 (find_and_remove_re): Emit a copy from the extension's
8526 destination to its source after the defining insn if
8527 the source and destination registers are different.
8528
8529 PR middle-end/59285
8530 * ifcvt.c (merge_if_block): If we are merging a block with more than
8531 one successor with a block with no successors, remove any BARRIER
8532 after the second block.
8533
8534 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8535
8536 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8537
8538 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8539
8540 PR target/59652
8541 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8542 for 14-bit register offsets when INT14_OK_STRICT is false.
8543
8544 2014-01-07 Roland Stigge <stigge@antcom.de>
8545 Michael Meissner <meissner@linux.vnet.ibm.com>
8546
8547 PR 57386/target
8548 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8549 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8550
8551 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8552
8553 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8554 -mcpu.
8555
8556 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8557
8558 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8559 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8560 rtx is const0_rtx or not.
8561
8562 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8563
8564 PR target/58115
8565 * target-globals.c (save_target_globals): Remove this_fn_optab
8566 handling.
8567 * toplev.c: Include optabs.h.
8568 (target_reinit): Temporarily restore the global options if another
8569 set of options are in force.
8570
8571 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8572
8573 PR rtl-optimization/58668
8574 * cfgcleanup.c (flow_find_cross_jump): Don't count
8575 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8576 to determine what is counted.
8577 (flow_find_head_matching_sequence): Use active_insn_p to determine
8578 what is counted.
8579 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8580 counting change.
8581 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8582 determine what is counted.
8583
8584 PR tree-optimization/59643
8585 * tree-predcom.c (split_data_refs_to_components): If one dr is
8586 read and one write, determine_offset fails and the write isn't
8587 in the bad component, just put the read into the bad component.
8588
8589 2014-01-07 Mike Stump <mikestump@comcast.net>
8590 Jakub Jelinek <jakub@redhat.com>
8591
8592 PR pch/59436
8593 * tree-core.h (struct tree_optimization_option): Change optabs
8594 type from unsigned char * to void *.
8595 * optabs.c (init_tree_optimization_optabs): Adjust
8596 TREE_OPTIMIZATION_OPTABS initialization.
8597
8598 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8599
8600 PR target/59644
8601 * config/i386/i386.h (struct machine_function): Add
8602 no_drap_save_restore field.
8603 * config/i386/i386.c (ix86_save_reg): Use
8604 !cfun->machine->no_drap_save_restore instead of
8605 crtl->stack_realign_needed.
8606 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8607 this function clears frame_pointer_needed. Set
8608 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8609 and DRAP reg is needed.
8610
8611 2014-01-06 Marek Polacek <polacek@redhat.com>
8612
8613 PR c/57773
8614 * doc/implement-c.texi: Mention that other integer types are
8615 permitted as bit-field types in strictly conforming mode.
8616
8617 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8618
8619 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8620 is newly allocated.
8621
8622 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8623
8624 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8625
8626 2014-01-06 Martin Jambor <mjambor@suse.cz>
8627
8628 PR ipa/59008
8629 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8630 to int.
8631 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8632
8633 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8634
8635 PR debug/59350
8636 PR debug/59510
8637 * var-tracking.c (add_stores): Preserve the value of the source even if
8638 we don't record the store.
8639
8640 2014-01-06 Terry Guo <terry.guo@arm.com>
8641
8642 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8643
8644 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8645
8646 PR bootstrap/59541
8647 * config/darwin.c (darwin_function_section): Adjust return values to
8648 correspond to optimisation changes made in r206070.
8649
8650 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8651
8652 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8653 from prefetch_block tune setting.
8654 (nocona_cost): Correct size of prefetch block to 64.
8655
8656 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8657
8658 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8659 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8660 used to save the static chain register in the computation of the offset
8661 from which the FP registers need to be restored.
8662
8663 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8664
8665 PR tree-optimization/59519
8666 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8667 ICE if get_current_def (current_new_name) is already non-NULL, as long
8668 as it is a phi result of some other phi in *new_exit_bb that has
8669 the same argument.
8670
8671 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8672 or vmovdqu* for misaligned_operand.
8673 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8674 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8675 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8676 aligned_mem for AVX512F masked aligned load and store builtins and for
8677 non-temporal moves.
8678
8679 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8680
8681 PR tree-optimization/59651
8682 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8683 Address range for negative step should be added by TYPE_SIZE_UNIT.
8684
8685 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8686
8687 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8688 overlapping registers also for registers other than the stack pointer.
8689
8690 2014-01-03 Marek Polacek <polacek@redhat.com>
8691
8692 PR other/59661
8693 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8694 __builtin_FILE.
8695
8696 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8697
8698 PR target/59625
8699 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8700 asm goto as jump.
8701
8702 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8703 (push splitter): Use <P:MODE_SIZE> instead of
8704 GET_MODE_SIZE (<P:MODE>mode).
8705 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8706 (mov -1, reg peephole2): Likewise.
8707 * config/i386/sse.md (*mov<mode>_internal,
8708 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8709 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8710 *<code><mode>3, *andnot<mode>3<mask_name>,
8711 <mask_codefor><code><mode>3<mask_name>): Likewise.
8712 * config/i386/subst.md (mask_mode512bit_condition,
8713 sd_mask_mode512bit_condition): Likewise.
8714
8715 2014-01-02 Xinliang David Li <davidxl@google.com>
8716
8717 PR tree-optimization/59303
8718 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8719 (dump_predicates): Better output format.
8720 (pred_equal_p): New function.
8721 (is_neq_relop_p): Ditto.
8722 (is_neq_zero_form_p): Ditto.
8723 (pred_expr_equal_p): Ditto.
8724 (pred_neg_p): Ditto.
8725 (simplify_pred): Ditto.
8726 (simplify_preds_2): Ditto.
8727 (simplify_preds_3): Ditto.
8728 (simplify_preds_4): Ditto.
8729 (simplify_preds): Ditto.
8730 (push_pred): Ditto.
8731 (push_to_worklist): Ditto.
8732 (get_pred_info_from_cmp): Ditto.
8733 (is_degenerated_phi): Ditto.
8734 (normalize_one_pred_1): Ditto.
8735 (normalize_one_pred): Ditto.
8736 (normalize_one_pred_chain): Ditto.
8737 (normalize_preds): Ditto.
8738 (normalize_cond_1): Remove function.
8739 (normalize_cond): Ditto.
8740 (is_gcond_subset_of): Ditto.
8741 (is_subset_of_any): Ditto.
8742 (is_or_set_subset_of): Ditto.
8743 (is_and_set_subset_of): Ditto.
8744 (is_norm_cond_subset_of): Ditto.
8745 (pred_chain_length_cmp): Ditto.
8746 (convert_control_dep_chain_into_preds): Type change.
8747 (find_predicates): Ditto.
8748 (find_def_preds): Ditto.
8749 (destroy_predicates_vecs): Ditto.
8750 (find_matching_predicates_in_rest_chains): Ditto.
8751 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8752 (is_pred_expr_subset): Ditto.
8753 (is_pred_chain_subset_of): Ditto.
8754 (is_included_in): Ditto.
8755 (is_superset_of): Ditto.
8756
8757 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8758
8759 Update copyright years.
8760
8761 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8762
8763 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8764 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8765 config/arc/arc.md, config/arc/arc.opt,
8766 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8767 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8768 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8769 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8770 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8771 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8772 vtable-verify.c, vtable-verify.h: Use the standard form for the
8773 copyright notice.
8774
8775 2014-01-02 Tobias Burnus <burnus@net-b.de>
8776
8777 * gcc.c (process_command): Update copyright notice dates.
8778 * gcov-dump.c: Ditto.
8779 * gcov.c: Ditto.
8780 * doc/cpp.texi: Bump @copying's copyright year.
8781 * doc/cppinternals.texi: Ditto.
8782 * doc/gcc.texi: Ditto.
8783 * doc/gccint.texi: Ditto.
8784 * doc/gcov.texi: Ditto.
8785 * doc/install.texi: Ditto.
8786 * doc/invoke.texi: Ditto.
8787
8788 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8789
8790 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8791
8792 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8793
8794 * config/i386/sse.md (*mov<mode>_internal): Guard
8795 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8796
8797 PR rtl-optimization/59647
8798 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8799 new_rtx into UNSIGNED_FLOAT rtxes.
8800 \f
8801 Copyright (C) 2014 Free Software Foundation, Inc.
8802
8803 Copying and distribution of this file, with or without modification,
8804 are permitted in any medium without royalty provided the copyright
8805 notice and this notice are preserved.