re PR target/61026 (sh-rtems4.11 build of 4.9.0 fails on FreeBSD 10 c++ (clang).)
[gcc.git] / gcc / ChangeLog
1 2014-05-03 Oleg Endo <olegendo@gcc.gnu.org>
2
3 PR target/61026
4 * config/sh/sh.c: Include stdlib headers before everything else.
5
6 2014-05-02 Jakub Jelinek <jakub@redhat.com>
7
8 * gimplify.c (gimplify_adjust_omp_clauses_1): Handle
9 GOVD_FIRSTPRIVATE | GOVD_LASTPRIVATE.
10 (gimplify_adjust_omp_clauses): Simd region is never
11 directly nested in combined parallel. Instead, for linear
12 with copyin/copyout, if in combined for simd loop, make decl
13 firstprivate/lastprivate on OMP_FOR.
14 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15 expand_omp_for_static_chunk): When setting endvar, also set
16 fd->loop.v to the same value.
17
18 2014-05-02 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19
20 * hwint.h (zext_hwi): Fix signed overflow for prec == 63.
21
22 2014-05-02 Alan Lawrence <alan.lawrence@arm.com>
23
24 * config/aarch64/aarch64.c (aarch64_expand_vec_perm_1): Tidy bit-flip
25 expression.
26
27 2014-05-02 Marek Polacek <polacek@redhat.com>
28
29 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
30
31 2014-02-26 Kito Cheng <kito@0xlab.org>
32
33 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
34 to a C expression marco.
35 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
36 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
37 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
38 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
39 HONOR_REG_ALLOC_ORDER.
40 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
41
42 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
43
44 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
45
46 2014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
47
48 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
49
50 2014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
51
52 * tree-if-conv.c (is_cond_scalar_reduction): New function.
53 (convert_scalar_cond_reduction): Likewise.
54 (predicate_scalar_phi): Add recognition and transformation
55 of simple conditioanl reduction to be vectorizable.
56
57 2014-05-01 Marek Polacek <polacek@redhat.com>
58
59 PR c/43245
60 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
61
62 2014-04-30 Alan Lawrence <alan.lawrence@arm.com>
63
64 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
65 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
66 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
67 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
68 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
69 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
70 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
71 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
72
73 2014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
74
75 * config/arc/arc.opt (mlra): Move comment above option name
76 to avoid mis-parsing as language options.
77
78 2014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
79
80 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
81 * config/sol2.h: ... here.
82 * config/sol2-10.h: Remove.
83
84 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
85 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
86 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
87 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
88 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
89 * config/sol2.h: ... here.
90 (SECTION_NAME_FORMAT): Don't redefine.
91 (STARTFILE_ARCH32_SPEC): Rename to ...
92 (STARTFILE_ARCH_SPEC): ... this.
93 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
94 * config/sparc/sol2.h: ... here.
95 (SECTION_NAME_FORMAT): Don't undef.
96 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
97 (SUBTARGET_EXTRA_SPECS): Remove.
98 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
99
100 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
101 (MD_STARTFILE_PREFIX): Remove.
102 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
103 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
104 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
105 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
106 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
107 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
108 * config/i386/sol2.h: ... here.
109 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
110 * config/i386/sol2-bi.h: Remove.
111 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
112 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
113
114 * config/i386/t-sol2-64: Rename to ...
115 * config/i386/t-sol2: ... this.
116 * config/sparc/t-sol2-64: Rename to ...
117 * config/sparc/t-sol2: ... this.
118
119 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
120 sol2_tm_file_head, sol2_tm_file_tail.
121 Include ${cpu_type}/sol2.h before sol2.h.
122 Remove sol2-10.h.
123 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
124 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
125 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
126 Reflect i386/t-sol2-64 renaming.
127 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
128 Reflect sparc/t-sol2-64 renaming.
129
130 2014-04-30 Richard Biener <rguenther@suse.de>
131
132 * passes.c (execute_function_todo): Move TODO_verify_stmts
133 and TODO_verify_ssa under the TODO_verify_il umbrella.
134 * tree-ssa.h (verify_ssa): Adjust prototype.
135 * tree-ssa.c (verify_ssa): Add parameter to tell whether
136 we should verify SSA operands.
137 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
138 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
139 whether we should verify whether not throwing stmts have EH info.
140 * graphite-scop-detection.c (create_sese_edges): Adjust.
141 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
142 * tree-eh.c (lower_try_finally_switch): Do not add the
143 default case label twice.
144
145 2014-04-30 Marek Polacek <polacek@redhat.com>
146
147 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
148 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
149 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
150 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
151
152 2014-04-29 Alan Lawrence <alan.lawrence@arm.com>
153
154 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
155 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
156 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
157 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
158 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
159 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
160 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
161 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
162
163 2014-04-29 David Malcolm <dmalcolm@redhat.com>
164
165 * tree-cfg.c (dump_function_to_file): Dump the return type of
166 functions, in a line to itself before the function body, mimicking
167 the layout of a C function.
168
169 2014-04-29 Jakub Jelinek <jakub@redhat.com>
170
171 PR tree-optimization/60971
172 * tree-tailcall.c (process_assignment): Reject conversions which
173 reduce precision.
174
175 2014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
176
177 * calls.c (initialize_argument_information): Always treat
178 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
179 (expand_call): Likewise.
180 (emit_library_call_calue_1): Likewise.
181 * expr.c (PUSH_ARGS_REVERSED): Do not define.
182 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
183 code accordingly.
184
185 2014-04-29 Nick Clifton <nickc@redhat.com>
186
187 * config/msp430/msp430.md (umulsidi): Fix typo.
188 (mulhisi3): Enable even inside interrupt handlers.
189 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
190 bigger return address pushed in large mode.
191
192 2014-04-29 Nick Clifton <nickc@redhat.com>
193
194 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
195 (arc_init_reg_tables): Use a machine_mode enum to iterate over
196 available modes.
197 * config/m32r/m32r.c (init_reg_tables): Likewise.
198 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
199 enum to hold the modes.
200
201 2014-04-29 Richard Biener <rguenther@suse.de>
202
203 * dominance.c (free_dominance_info): Add overload with
204 function parameter.
205 (dom_info_state): Likewise.
206 (dom_info_available_p): Likewise.
207 * basic-block.h (free_dominance_info, dom_info_state,
208 dom_info_available_p): Declare overloads.
209 * passes.c (execute_function_todo): Verify that verifiers
210 don't change dominator info state. Drop dominator info
211 for IPA pass invocations.
212 * cgraph.c (release_function_body): Restore asserts that
213 dominator information is released.
214
215 2014-04-29 Patrick Palka <patrick@parcs.ath.cx>
216
217 * doc/invoke.texi: Fix typo.
218 * tree-vrp.c: Fix typos.
219 * gimple.c (infer_nonnull_range): Reorder operands of an &&
220 condition.
221
222 2014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
223
224 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
225
226 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
227
228 * config/aarch64/aarch64-builtins.c
229 (aarch64_types_storestruct_lane_qualifiers): New.
230 (TYPES_STORESTRUCT_LANE): Likewise.
231 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
232 (st3_lane): Likewise.
233 (st4_lane): Likewise.
234 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
235 (vec_store_lanesci_lane<mode>): Likewise.
236 (vec_store_lanesxi_lane<mode>): Likewise.
237 (aarch64_st2_lane<VQ:mode>): Likewise.
238 (aarch64_st3_lane<VQ:mode>): Likewise.
239 (aarch64_st4_lane<VQ:mode>): Likewise.
240 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
241 * config/aarch64/arm_neon.h
242 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
243 use new macro arguments.
244 (__ST3_LANE_FUNC): Likewise.
245 (__ST4_LANE_FUNC): Likewise.
246 * config/aarch64/iterators.md (V_TWO_ELEM): New.
247 (V_THREE_ELEM): Likewise.
248 (V_FOUR_ELEM): Likewise.
249
250 2014-04-28 David Malcolm <dmalcolm@redhat.com>
251
252 * doc/gimple.texi: Replace the description of the now-defunct
253 union gimple_statement_d with a diagram showing the
254 gimple_statement_base class hierarchy and its relationships to
255 the GSS_ and GIMPLE_ enums.
256
257 2014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
258
259 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
260 * config/aarch64/aarch64.c
261 (aarch64_cannot_change_mode_class): Weaken conditions.
262 (aarch64_modes_tieable_p): New.
263 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
264
265 2014-04-28 Pat Haugen <pthaugen@us.ibm.com>
266
267 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
268 (loadsync_<mode>): Change mode.
269 (load_quadpti, store_quadpti): New.
270 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
271 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
272
273 2014-04-28 Martin Jambor <mjambor@suse.cz>
274
275 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
276 same alias type as the original statement.
277 (subreplacement_assignment_data): New type.
278 (handle_unscalarized_data_in_subtree): New type of parameter,
279 generate new memory accesses with same alias type as the original
280 statement.
281 (load_assign_lhs_subreplacements): Likewise.
282 (sra_modify_constructor_assign): Generate new memory accesses with
283 same alias type as the original statement.
284
285 2014-04-28 Richard Biener <rguenther@suse.de>
286
287 * tree-pass.h (TODO_verify_il): Define.
288 (TODO_verify_all): Complete properly.
289 * passes.c (execute_function_todo): Move existing loop-closed
290 SSA verification under TODO_verify_il.
291 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
292 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
293 Fix tree sharing issue.
294
295 2014-04-28 Richard Biener <rguenther@suse.de>
296
297 PR middle-end/60092
298 * builtins.def (DEF_C11_BUILTIN): Add.
299 (BUILT_IN_ALIGNED_ALLOC): Likewise.
300 * coretypes.h (enum function_class): Add function_c11_misc.
301 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
302 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
303 (call_may_clobber_ref_p_1): Likewise.
304 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
305 (mark_all_reaching_defs_necessary_1): Likewise.
306 (propagate_necessity): Likewise.
307 (eliminate_unnecessary_stmts): Likewise.
308 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
309
310 2014-04-28 Richard Biener <rguenther@suse.de>
311
312 * tree-vrp.c (vrp_var_may_overflow): Remove.
313 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
314 with overflow immediately bump to one before that value and
315 let iteration figure out overflow status.
316
317 2014-04-28 Richard Biener <rguenther@suse.de>
318
319 * configure.ac: Do valgrind header checks unconditionally.
320 Add --enable-valgrind-annotations.
321 * system.h: Guard valgrind header inclusion with
322 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
323 * alloc-pool.c (pool_alloc, pool_free): Use
324 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
325 to guard possibly dead code.
326 * config.in: Regenerated.
327 * configure: Likewise.
328
329 2014-04-28 Jeff Law <law@redhat.com>
330
331 PR tree-optimization/60902
332 * tree-ssa-threadedge.c
333 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
334 over real defs when invalidating outputs from statements that do not
335 produce useful outputs for threading.
336
337 2014-04-28 Richard Biener <rguenther@suse.de>
338
339 PR tree-optimization/60979
340 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
341 SCOPs that end in a block with a successor with abnormal
342 predecessors.
343
344 2014-04-28 Richard Biener <rguenther@suse.de>
345
346 * tree-pass.h (execute_pass_list): Adjust prototype.
347 * passes.c (pass_manager::execute_early_local_passes):
348 Adjust.
349 (do_per_function): Change callback signature, push all actual
350 work to the callbals.
351 (do_per_function_toporder): Likewise.
352 (execute_function_dump): Adjust.
353 (execute_function_todo): Likewise.
354 (clear_last_verified): Likewise.
355 (verify_curr_properties): Likewise.
356 (update_properties_after_pass): Likewise.
357 (execute_pass_list_1): Split out from ...
358 (execute_pass_list): ... here. Adjust.
359 (execute_ipa_pass_list): Likewise.
360 * cgraphunit.c (cgraph_add_new_function): Adjust.
361 (analyze_function): Likewise.
362 (expand_function): Likewise.
363 * cgraph.c (release_function_body): Free dominance info
364 here instead of asserting it was magically freed elsewhere.
365
366 2014-04-28 Eric Botcazou <ebotcazou@adacore.com>
367
368 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
369 * configure: Regenerate.
370 * config/sparc/sparc.opt (muser-mode): New option.
371 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
372 for LEON3.
373 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
374 * doc/invoke.texi (SPARC options): Document -muser-mode.
375
376 2014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
377
378 * cselib.c (find_slot_memmode): Delete.
379 (cselib_hasher): Change compare_type to a struct.
380 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
381 constants.
382 (preserve_constants_and_equivs): Adjust for new compare_type.
383 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
384 (wrap_constant): Delete.
385 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
386
387 2014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
388
389 * doc/install.texi (Building with profile feedback): Remove
390 outdated sentence.
391
392 2014-04-26 Tom de Vries <tom@codesourcery.com>
393
394 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
395 array accesses.
396
397 2014-04-25 Cary Coutant <ccoutant@google.com>
398
399 PR debug/60929
400 * dwarf2out.c (should_move_die_to_comdat): A type definition
401 can contain a subprogram definition, but don't move it to a
402 comdat unit.
403 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
404 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
405 from original DIE.
406 (clone_tree_hash): Rename to...
407 (clone_tree_partial): ...this; change callers. Copy
408 DW_TAG_subprogram DIEs as declarations.
409 (copy_decls_walk): Don't copy children of a declaration into a
410 type unit.
411
412 2014-04-25 H.J. Lu <hongjiu.lu@intel.com>
413
414 PR target/60969
415 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
416 alternative 12.
417
418 2014-04-25 Jiong Wang <jiong.wang@arm.com>
419
420 * config/arm/predicates.md (call_insn_operand): Add long_call check.
421 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
422 reg for long_call.
423 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
424 restriction.
425
426 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
427
428 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
429 T16-related fields.
430
431 2014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
432
433 PR tree-optimization/60930
434 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
435 creating a multiply candidate by folding two constant
436 multiplicands when the result overflows.
437
438 2014-04-25 Jakub Jelinek <jakub@redhat.com>
439
440 PR tree-optimization/60960
441 * tree-vect-generic.c (expand_vector_operation): Only call
442 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
443
444 2014-04-25 Tom de Vries <tom@codesourcery.com>
445
446 * expr.c (clobber_reg_mode): New function.
447 * expr.h (clobber_reg): New function.
448
449 2014-04-25 Tom de Vries <tom@codesourcery.com>
450
451 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
452 clobbers.
453
454 2014-04-25 Radovan Obradovic <robradovic@mips.com>
455 Tom de Vries <tom@codesourcery.com>
456
457 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
458 handle.
459 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
460 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
461 new argument to find_all_hard_reg_sets call.
462
463 2014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
464
465 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
466 Use HOST_WIDE_INT_C for mask literal.
467 (aarch_rev16_shleft_mask_imm_p): Likewise.
468
469 2014-04-25 Eric Botcazou <ebotcazou@adacore.com>
470
471 PR target/60941
472 * config/sparc/sparc.md (ashlsi3_extend): Delete.
473
474 2014-04-25 Marc Glisse <marc.glisse@inria.fr>
475
476 PR preprocessor/56540
477 * config/i386/i386-c.c (ix86_target_macros): Define
478 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
479
480 2014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
481
482 * configure.ac (tga_func): Remove.
483 (LIB_TLS_SPEC): Remove.
484 * configure: Regenerate.
485 * config.in: Regenerate.
486 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
487
488 2014-04-25 Richard Biener <rguenther@suse.de>
489
490 PR ipa/60912
491 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
492 call stmt use/clobber sets during stmt walk instead of
493 walking the possibly incomplete set of caller edges.
494
495 2014-04-25 Richard Biener <rguenther@suse.de>
496
497 PR ipa/60911
498 * passes.c (apply_ipa_transforms): Inline into only caller ...
499 (execute_one_pass): ... here. Properly bring in function
500 bodies for nodes we want to apply IPA transforms to.
501
502 2014-04-24 Cong Hou <congh@google.com>
503
504 PR tree-optimization/60896
505 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
506 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
507 (vect_mark_pattern_stmts): Set the def type of all statements in
508 PATTERN_DEF_SEQ as vect_internal_def.
509
510 2014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
511
512 * doc/extend.texi (PowerPC Built-in Functions): Document new
513 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
514 functions.
515 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
516
517 * config/rs6000/predicates.md (const_0_to_3_operand): New
518 predicate to match 0..3 integer constants.
519
520 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
521 to support adding miscellaneous builtin functions.
522 (BU_DFP_MISC_2): Likewise.
523 (BU_P7_MISC_1): Likewise.
524 (BU_P7_MISC_2): Likewise.
525 (BU_P8V_MISC_3): Likewise.
526 (BU_MISC_1): Likewise.
527 (BU_MISC_2): Likewise.
528 (DIVWE): Add extended divide builtin functions.
529 (DIVWEO): Likewise.
530 (DIVWEU): Likewise.
531 (DIVWEUO): Likewise.
532 (DIVDE): Likewise.
533 (DIVDEO): Likewise.
534 (DIVDEU): Likewise.
535 (DIVDEUO): Likewise.
536 (DXEX): Add decimal floating-point builtin functions.
537 (DXEXQ): Likewise.
538 (DDEDPD): Likewise.
539 (DDEDPDQ): Likewise.
540 (DENBCD): Likewise.
541 (DENBCDQ): Likewise.
542 (DIEX): Likewise.
543 (DIEXQ): Likewise.
544 (DSCLI): Likewise.
545 (DSCLIQ): Likewise.
546 (DSCRI): Likewise.
547 (DSCRIQ): Likewise.
548 (CDTBCD): Add new BCD builtin functions.
549 (CBCDTD): Likewise.
550 (ADDG6S): Likewise.
551 (BCDADD): Likewise.
552 (BCDADD_LT): Likewise.
553 (BCDADD_EQ): Likewise.
554 (BCDADD_GT): Likewise.
555 (BCDADD_OV): Likewise.
556 (BCDSUB): Likewise.
557 (BCDSUB_LT): Likewise.
558 (BCDSUB_EQ): Likewise.
559 (BCDSUB_GT): Likewise.
560 (BCDSUB_OV): Likewise.
561 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
562 (UNPACK_TD): Likewise.
563 (PACK_TF): Likewise.
564 (UNPACK_TF): Likewise.
565 (UNPACK_TF_0): Likewise.
566 (UNPACK_TF_1): Likewise.
567 (PACK_V1TI): Likewise.
568 (UNPACK_V1TI): Likewise.
569
570 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
571 support for decimal floating point builtin functions.
572 (rs6000_expand_ternop_builtin): Add checks for the new builtin
573 functions that take constant arguments.
574 (rs6000_invalid_builtin): Add decimal floating point builtin
575 support.
576 (rs6000_init_builtins): Setup long double, _Decimal64, and
577 _Decimal128 types for new builtin functions.
578 (builtin_function_type): Set the unsigned flags appropriately for
579 the new builtin functions.
580 (rs6000_opt_masks): Add support for decimal floating point builtin
581 functions.
582
583 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
584 floating point builtin functions.
585 (RS6000_BTM_COMMON): Likewise.
586 (RS6000_BTI_long_double): Likewise.
587 (RS6000_BTI_dfloat64): Likewise.
588 (RS6000_BTI_dfloat128): Likewise.
589 (long_double_type_internal_node): Likewise.
590 (dfloat64_type_internal_node): Likewise.
591 (dfloat128_type_internal_node): Likewise.
592
593 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
594 2.07 bcd arithmetic instructions.
595 (UNSPEC_BCDSUB): Likewise.
596 (UNSPEC_BCD_OVERFLOW): Likewise.
597 (UNSPEC_BCD_ADD_SUB): Likewise.
598 (bcd_add_sub): Likewise.
599 (BCD_TEST): Likewise.
600 (bcd<bcd_add_sub>): Likewise.
601 (bcd<bcd_add_sub>_test): Likewise.
602 (bcd<bcd_add_sub>_test2): Likewise.
603 (bcd<bcd_add_sub>_<code>): Likewise.
604 (peephole2 for combined bcd ops): Likewise.
605
606 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
607 decimal floating point builtin functions.
608 (UNSPEC_DENBCD): Likewise.
609 (UNSPEC_DXEX): Likewise.
610 (UNSPEC_DIEX): Likewise.
611 (UNSPEC_DSCLI): Likewise.
612 (UNSPEC_DSCRI): Likewise.
613 (D64_D128): Likewise.
614 (dfp_suffix): Likewise.
615 (dfp_ddedpd_<mode>): Likewise.
616 (dfp_denbcd_<mode>): Likewise.
617 (dfp_dxex_<mode>): Likewise.
618 (dfp_diex_<mode>): Likewise.
619 (dfp_dscli_<mode>): Likewise.
620 (dfp_dscri_<mode>): Likewise.
621
622 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
623 builtin functions.
624 (UNSPEC_CDTBCD): Likewise.
625 (UNSPEC_CBCDTD): Likewise.
626 (UNSPEC_DIVE): Add support for new extended divide builtin
627 functions.
628 (UNSPEC_DIVEO): Likewise.
629 (UNSPEC_DIVEU): Likewise.
630 (UNSPEC_DIVEUO): Likewise.
631 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
632 pack/unpack 128-bit types.
633 (UNSPEC_PACK_128BIT): Likewise.
634 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
635 (udiv<mode>3): Use idiv_ldiv mode attribute.
636 (div<mode>3): Likewise.
637 (addg6s): Add new BCD builtin functions.
638 (cdtbcd): Likewise.
639 (cbcdtd): Likewise.
640 (UNSPEC_DIV_EXTEND): Add support for new extended divide
641 instructions.
642 (div_extend): Likewise.
643 (div<div_extend>_<mode>"): Likewise.
644 (FP128_64): Add support for new builtin functions to pack/unpack
645 128-bit types.
646 (unpack<mode>): Likewise.
647 (unpacktf_0): Likewise.
648 (unpacktf_1): Likewise.
649 (unpack<mode>_dm): Likewise.
650 (unpack<mode>_nodm): Likewise.
651 (pack<mode>): Likewise.
652 (unpackv1ti): Likewise.
653 (packv1ti): Likewise.
654
655 2014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
656
657 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
658 is disabled.
659
660 2014-04-24 Jakub Jelinek <jakub@redhat.com>
661
662 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
663 * gimplify.c (omp_is_private): Change last argument's type to int.
664 Only diagnose lastprivate if the simd argument is 1, only diagnose
665 linear if the simd argument is 2.
666 (gimplify_omp_for): Adjust omp_is_private callers. When adding
667 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
668 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
669 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
670 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
671 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
672 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
673 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
674 * tree-nested.c (convert_nonlocal_omp_clauses,
675 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
676
677 2014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
678
679 PR target/60822
680 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
681 operand 1.
682
683 2014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
684
685 * flag-types.h (enum ivar_visibility): Add.
686
687 2014-04-24 Trevor Saunders <tsaunders@mozilla.com>
688
689 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
690 function * argument.
691
692 2014-04-24 Alan Lawrence <alan.lawrence@arm.com>
693
694 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
695
696 2014-04-24 Radovan Obradovic <robradovic@mips.com>
697 Tom de Vries <tom@codesourcery.com>
698
699 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
700 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
701 reg-note.
702 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
703 * emit-rtl.c (try_split): Same.
704
705 2014-04-24 Radovan Obradovic <robradovic@mips.com>
706 Tom de Vries <tom@codesourcery.com>
707
708 * common.opt (fuse-caller-save): New option.
709
710 2014-04-24 Tejas Belagod <tejas.belagod@arm.com>
711
712 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
713 for big-endian.
714
715 2014-04-24 Richard Biener <rguenther@suse.de>
716
717 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
718 during TER and instead use the sepops interface for expanding
719 non-GIMPLE_SINGLE_RHS.
720
721 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
722
723 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
724 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
725
726 2014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
727
728 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
729 assembler 64-bit option.
730 * configure: Regenerate.
731
732 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
733
734 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
735 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
736 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
737 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
738 (TARGET_CRYPTO): Take TARGET_SIMD into account.
739
740 2014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
741
742 * config/aarch64/aarch64-builtins.c
743 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
744 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
745 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
746 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
747 builtins.
748 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
749 (Vrevsuff): New mode attribute.
750
751 2014-04-24 Terry Guo <terry.guo@arm.com>
752
753 * config/arm/arm.h (machine_function): Define variable
754 after_arm_reorg here.
755 * config/arm/arm.c (after_arm_reorg): Remove the definition.
756 (arm_split_constant): Update the way to access variable
757 after_arm_reorg.
758 (arm_reorg): Ditto.
759 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
760
761 2014-04-23 Tom de Vries <tom@codesourcery.com>
762
763 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
764
765 2014-04-23 David Malcolm <dmalcolm@redhat.com>
766
767 * is-a.h: Update comments to reflect the following changes to the
768 "pointerness" of the API, making the template parameter match the
769 return type, allowing use of is-a.h with typedefs of pointers.
770 (is_a_helper::cast): Return a T rather then a pointer to a T, so
771 that the return type matches the parameter to the is_a_helper.
772 (as_a): Likewise.
773 (dyn_cast): Likewise.
774
775 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
776 pointer from the is-a.h API.
777
778 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
779 (is_a_helper <cgraph_node *>::test): ...this, matching change to
780 is-a.h API.
781 (is_a_helper <varpool_node>::test): Likewise, convert to...
782 (is_a_helper <varpool_node *>::test): ...this.
783
784 (varpool_first_variable): Update for removal of implicit pointer
785 from the is-a.h API.
786 (varpool_next_variable): Likewise.
787 (varpool_first_static_initializer): Likewise.
788 (varpool_next_static_initializer): Likewise.
789 (varpool_first_defined_variable): Likewise.
790 (varpool_next_defined_variable): Likewise.
791 (cgraph_first_defined_function): Likewise.
792 (cgraph_next_defined_function): Likewise.
793 (cgraph_first_function): Likewise.
794 (cgraph_next_function): Likewise.
795 (cgraph_first_function_with_gimple_body): Likewise.
796 (cgraph_next_function_with_gimple_body): Likewise.
797 (cgraph_alias_target): Likewise.
798 (varpool_alias_target): Likewise.
799 (cgraph_function_or_thunk_node): Likewise.
800 (varpool_variable_node): Likewise.
801 (symtab_real_symbol_p): Likewise.
802 * cgraphunit.c (referred_to_p): Likewise.
803 (analyze_functions): Likewise.
804 (handle_alias_pairs): Likewise.
805 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
806 * gimple-ssa.h (gimple_vuse_op): Likewise.
807 (gimple_vdef_op): Likewise.
808 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
809 * gimple.c (gimple_build_asm_1): Likewise.
810 (gimple_build_try): Likewise.
811 (gimple_build_resx): Likewise.
812 (gimple_build_eh_dispatch): Likewise.
813 (gimple_build_omp_for): Likewise.
814 (gimple_omp_for_set_clauses): Likewise.
815
816 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
817 (is_a_helper <gimple_statement_asm *>::test): ...this.
818 (is_a_helper <gimple_statement_bind>::test): Convert to...
819 (is_a_helper <gimple_statement_bind *>::test): ...this.
820 (is_a_helper <gimple_statement_call>::test): Convert to...
821 (is_a_helper <gimple_statement_call *>::test): ...this.
822 (is_a_helper <gimple_statement_catch>::test): Convert to...
823 (is_a_helper <gimple_statement_catch *>::test): ...this.
824 (is_a_helper <gimple_statement_resx>::test): Convert to...
825 (is_a_helper <gimple_statement_resx *>::test): ...this.
826 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
827 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
828 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
829 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
830 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
831 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
832 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
833 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
834 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
835 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
836 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
837 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
838 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
839 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
840 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
841 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
842 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
843 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
844 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
845 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
846 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
847 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
848 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
849 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
850 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
851 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
852 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
853 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
854 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
855 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
856 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
857 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
858 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
859 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
860 (is_a_helper <gimple_statement_phi>::test): Convert to...
861 (is_a_helper <gimple_statement_phi *>::test): ...this.
862 (is_a_helper <gimple_statement_transaction>::test): Convert to...
863 (is_a_helper <gimple_statement_transaction *>::test): ...this.
864 (is_a_helper <gimple_statement_try>::test): Convert to...
865 (is_a_helper <gimple_statement_try *>::test): ...this.
866 (is_a_helper <gimple_statement_wce>::test): Convert to...
867 (is_a_helper <gimple_statement_wce *>::test): ...this.
868 (is_a_helper <const gimple_statement_asm>::test): Convert to...
869 (is_a_helper <const gimple_statement_asm *>::test): ...this.
870 (is_a_helper <const gimple_statement_bind>::test): Convert to...
871 (is_a_helper <const gimple_statement_bind *>::test): ...this.
872 (is_a_helper <const gimple_statement_call>::test): Convert to...
873 (is_a_helper <const gimple_statement_call *>::test): ...this.
874 (is_a_helper <const gimple_statement_catch>::test): Convert to...
875 (is_a_helper <const gimple_statement_catch *>::test): ...this.
876 (is_a_helper <const gimple_statement_resx>::test): Convert to...
877 (is_a_helper <const gimple_statement_resx *>::test): ...this.
878 (is_a_helper <const gimple_statement_eh_dispatch>::test):
879 Convert to...
880 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
881 ...this.
882 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
883 to...
884 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
885 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
886 Convert to...
887 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
888 ...this.
889 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
890 Convert to...
891 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
892 ...this.
893 (is_a_helper <const gimple_statement_omp_return>::test): Convert
894 to...
895 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
896 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
897 to...
898 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
899 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
900 to...
901 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
902 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
903 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
904 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
905 to...
906 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
907 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
908 to...
909 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
910 (is_a_helper <const gimple_statement_omp_target>::test): Convert
911 to...
912 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
913 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
914 to...
915 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
916 (is_a_helper <const gimple_statement_omp_single>::test): Convert
917 to...
918 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
919 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
920 to...
921 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
922 (is_a_helper <const gimple_statement_omp_task>::test): Convert
923 to...
924 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
925 (is_a_helper <const gimple_statement_phi>::test): Convert to...
926 (is_a_helper <const gimple_statement_phi *>::test): ...this.
927 (is_a_helper <const gimple_statement_transaction>::test): Convert
928 to...
929 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
930 (is_a_helper <const gimple_statement_with_ops>::test): Convert
931 to...
932 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
933 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
934 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
935 (is_a_helper <const gimple_statement_with_memory_ops>::test):
936 Convert to...
937 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
938 ...this.
939 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
940 to...
941 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
942
943 (gimple_use_ops): Update for removal of implicit pointer from the
944 is-a.h API.
945 (gimple_set_use_ops): Likewise.
946 (gimple_vuse): Likewise.
947 (gimple_vdef): Likewise.
948 (gimple_vuse_ptr): Likewise.
949 (gimple_vdef_ptr): Likewise.
950 (gimple_set_vuse): Likewise.
951 (gimple_set_vdef): Likewise.
952 (gimple_omp_return_set_lhs): Likewise.
953 (gimple_omp_return_lhs): Likewise.
954 (gimple_omp_return_lhs_ptr): Likewise.
955 (gimple_call_fntype): Likewise.
956 (gimple_call_set_fntype): Likewise.
957 (gimple_call_set_internal_fn): Likewise.
958 (gimple_call_use_set): Likewise.
959 (gimple_call_clobber_set): Likewise.
960 (gimple_bind_vars): Likewise.
961 (gimple_bind_set_vars): Likewise.
962 (gimple_bind_body_ptr): Likewise.
963 (gimple_bind_set_body): Likewise.
964 (gimple_bind_add_stmt): Likewise.
965 (gimple_bind_block): Likewise.
966 (gimple_bind_set_block): Likewise.
967 (gimple_asm_ninputs): Likewise.
968 (gimple_asm_noutputs): Likewise.
969 (gimple_asm_nclobbers): Likewise.
970 (gimple_asm_nlabels): Likewise.
971 (gimple_asm_input_op): Likewise.
972 (gimple_asm_input_op_ptr): Likewise.
973 (gimple_asm_output_op): Likewise.
974 (gimple_asm_output_op_ptr): Likewise.
975 (gimple_asm_set_output_op): Likewise.
976 (gimple_asm_clobber_op): Likewise.
977 (gimple_asm_set_clobber_op): Likewise.
978 (gimple_asm_label_op): Likewise.
979 (gimple_asm_set_label_op): Likewise.
980 (gimple_asm_string): Likewise.
981 (gimple_catch_types): Likewise.
982 (gimple_catch_types_ptr): Likewise.
983 (gimple_catch_handler_ptr): Likewise.
984 (gimple_catch_set_types): Likewise.
985 (gimple_catch_set_handler): Likewise.
986 (gimple_eh_filter_types): Likewise.
987 (gimple_eh_filter_types_ptr): Likewise.
988 (gimple_eh_filter_failure_ptr): Likewise.
989 (gimple_eh_filter_set_types): Likewise.
990 (gimple_eh_filter_set_failure): Likewise.
991 (gimple_eh_must_not_throw_fndecl): Likewise.
992 (gimple_eh_must_not_throw_set_fndecl): Likewise.
993 (gimple_eh_else_n_body_ptr): Likewise.
994 (gimple_eh_else_e_body_ptr): Likewise.
995 (gimple_eh_else_set_n_body): Likewise.
996 (gimple_eh_else_set_e_body): Likewise.
997 (gimple_try_eval_ptr): Likewise.
998 (gimple_try_cleanup_ptr): Likewise.
999 (gimple_try_set_eval): Likewise.
1000 (gimple_try_set_cleanup): Likewise.
1001 (gimple_wce_cleanup_ptr): Likewise.
1002 (gimple_wce_set_cleanup): Likewise.
1003 (gimple_phi_capacity): Likewise.
1004 (gimple_phi_num_args): Likewise.
1005 (gimple_phi_result): Likewise.
1006 (gimple_phi_result_ptr): Likewise.
1007 (gimple_phi_set_result): Likewise.
1008 (gimple_phi_arg): Likewise.
1009 (gimple_phi_set_arg): Likewise.
1010 (gimple_resx_region): Likewise.
1011 (gimple_resx_set_region): Likewise.
1012 (gimple_eh_dispatch_region): Likewise.
1013 (gimple_eh_dispatch_set_region): Likewise.
1014 (gimple_omp_critical_name): Likewise.
1015 (gimple_omp_critical_name_ptr): Likewise.
1016 (gimple_omp_critical_set_name): Likewise.
1017 (gimple_omp_for_clauses): Likewise.
1018 (gimple_omp_for_clauses_ptr): Likewise.
1019 (gimple_omp_for_set_clauses): Likewise.
1020 (gimple_omp_for_collapse): Likewise.
1021 (gimple_omp_for_index): Likewise.
1022 (gimple_omp_for_index_ptr): Likewise.
1023 (gimple_omp_for_set_index): Likewise.
1024 (gimple_omp_for_initial): Likewise.
1025 (gimple_omp_for_initial_ptr): Likewise.
1026 (gimple_omp_for_set_initial): Likewise.
1027 (gimple_omp_for_final): Likewise.
1028 (gimple_omp_for_final_ptr): Likewise.
1029 (gimple_omp_for_set_final): Likewise.
1030 (gimple_omp_for_incr): Likewise.
1031 (gimple_omp_for_incr_ptr): Likewise.
1032 (gimple_omp_for_set_incr): Likewise.
1033 (gimple_omp_for_pre_body_ptr): Likewise.
1034 (gimple_omp_for_set_pre_body): Likewise.
1035 (gimple_omp_parallel_clauses): Likewise.
1036 (gimple_omp_parallel_clauses_ptr): Likewise.
1037 (gimple_omp_parallel_set_clauses): Likewise.
1038 (gimple_omp_parallel_child_fn): Likewise.
1039 (gimple_omp_parallel_child_fn_ptr): Likewise.
1040 (gimple_omp_parallel_set_child_fn): Likewise.
1041 (gimple_omp_parallel_data_arg): Likewise.
1042 (gimple_omp_parallel_data_arg_ptr): Likewise.
1043 (gimple_omp_parallel_set_data_arg): Likewise.
1044 (gimple_omp_task_clauses): Likewise.
1045 (gimple_omp_task_clauses_ptr): Likewise.
1046 (gimple_omp_task_set_clauses): Likewise.
1047 (gimple_omp_task_child_fn): Likewise.
1048 (gimple_omp_task_child_fn_ptr): Likewise.
1049 (gimple_omp_task_set_child_fn): Likewise.
1050 (gimple_omp_task_data_arg): Likewise.
1051 (gimple_omp_task_data_arg_ptr): Likewise.
1052 (gimple_omp_task_set_data_arg): Likewise.
1053 (gimple_omp_taskreg_clauses): Likewise.
1054 (gimple_omp_taskreg_clauses_ptr): Likewise.
1055 (gimple_omp_taskreg_set_clauses): Likewise.
1056 (gimple_omp_taskreg_child_fn): Likewise.
1057 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1058 (gimple_omp_taskreg_set_child_fn): Likewise.
1059 (gimple_omp_taskreg_data_arg): Likewise.
1060 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1061 (gimple_omp_taskreg_set_data_arg): Likewise.
1062 (gimple_omp_task_copy_fn): Likewise.
1063 (gimple_omp_task_copy_fn_ptr): Likewise.
1064 (gimple_omp_task_set_copy_fn): Likewise.
1065 (gimple_omp_task_arg_size): Likewise.
1066 (gimple_omp_task_arg_size_ptr): Likewise.
1067 (gimple_omp_task_set_arg_size): Likewise.
1068 (gimple_omp_task_arg_align): Likewise.
1069 (gimple_omp_task_arg_align_ptr): Likewise.
1070 (gimple_omp_task_set_arg_align): Likewise.
1071 (gimple_omp_single_clauses): Likewise.
1072 (gimple_omp_single_clauses_ptr): Likewise.
1073 (gimple_omp_single_set_clauses): Likewise.
1074 (gimple_omp_target_clauses): Likewise.
1075 (gimple_omp_target_clauses_ptr): Likewise.
1076 (gimple_omp_target_set_clauses): Likewise.
1077 (gimple_omp_target_child_fn): Likewise.
1078 (gimple_omp_target_child_fn_ptr): Likewise.
1079 (gimple_omp_target_set_child_fn): Likewise.
1080 (gimple_omp_target_data_arg): Likewise.
1081 (gimple_omp_target_data_arg_ptr): Likewise.
1082 (gimple_omp_target_set_data_arg): Likewise.
1083 (gimple_omp_teams_clauses): Likewise.
1084 (gimple_omp_teams_clauses_ptr): Likewise.
1085 (gimple_omp_teams_set_clauses): Likewise.
1086 (gimple_omp_sections_clauses): Likewise.
1087 (gimple_omp_sections_clauses_ptr): Likewise.
1088 (gimple_omp_sections_set_clauses): Likewise.
1089 (gimple_omp_sections_control): Likewise.
1090 (gimple_omp_sections_control_ptr): Likewise.
1091 (gimple_omp_sections_set_control): Likewise.
1092 (gimple_omp_for_set_cond): Likewise.
1093 (gimple_omp_for_cond): Likewise.
1094 (gimple_omp_atomic_store_set_val): Likewise.
1095 (gimple_omp_atomic_store_val): Likewise.
1096 (gimple_omp_atomic_store_val_ptr): Likewise.
1097 (gimple_omp_atomic_load_set_lhs): Likewise.
1098 (gimple_omp_atomic_load_lhs): Likewise.
1099 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1100 (gimple_omp_atomic_load_set_rhs): Likewise.
1101 (gimple_omp_atomic_load_rhs): Likewise.
1102 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1103 (gimple_omp_continue_control_def): Likewise.
1104 (gimple_omp_continue_control_def_ptr): Likewise.
1105 (gimple_omp_continue_set_control_def): Likewise.
1106 (gimple_omp_continue_control_use): Likewise.
1107 (gimple_omp_continue_control_use_ptr): Likewise.
1108 (gimple_omp_continue_set_control_use): Likewise.
1109 (gimple_transaction_body_ptr): Likewise.
1110 (gimple_transaction_label): Likewise.
1111 (gimple_transaction_label_ptr): Likewise.
1112 (gimple_transaction_set_body): Likewise.
1113 (gimple_transaction_set_label): Likewise.
1114
1115 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1116 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1117 * ipa-ref.c (ipa_record_reference): Likewise.
1118 * ipa-reference.c (analyze_function): Likewise.
1119 (ipa_reference_write_optimization_summary): Likewise.
1120 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1121 (address_taken_from_non_vtable_p): Likewise.
1122 (comdat_can_be_unshared_p_1): Likewise.
1123 * lto-cgraph.c (lto_output_ref): Likewise.
1124 (add_references): Likewise.
1125 (compute_ltrans_boundary): Likewise.
1126 (output_symtab): Likewise.
1127 (input_ref): Likewise.
1128 (input_cgraph_1): Likewise.
1129 (output_cgraph_opt_summary): Likewise.
1130 * lto-streamer-out.c (lto_output): Likewise.
1131 (output_symbol_p): Likewise.
1132 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1133 (lsei_start_function_in_partition): Likewise.
1134 (lsei_next_variable_in_partition): Likewise.
1135 (lsei_start_variable_in_partition): Likewise.
1136 * symtab.c (insert_to_assembler_name_hash): Likewise.
1137 (unlink_from_assembler_name_hash): Likewise.
1138 (symtab_unregister_node): Likewise.
1139 (symtab_remove_node): Likewise.
1140 (dump_symtab_node): Likewise.
1141 (verify_symtab_base): Likewise.
1142 (verify_symtab_node): Likewise.
1143 (symtab_make_decl_local): Likewise.
1144 (symtab_alias_ultimate_target): Likewise.
1145 (symtab_resolve_alias): Likewise.
1146 (symtab_get_symbol_partitioning_class): Likewise.
1147 * tree-phinodes.c (allocate_phi_node): Likewise.
1148 (reserve_phi_args_for_new_edge): Likewise.
1149 (remove_phi_args): Likewise.
1150 * varpool.c (varpool_node_for_asm): Likewise.
1151 (varpool_remove_unreferenced_decls): Likewise.
1152
1153 2014-04-23 Jeff Law <law@redhat.com>
1154
1155 PR tree-optimization/60902
1156 * tree-ssa-threadedge.c
1157 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1158 invalidate outputs from statements that do not produce useful
1159 outputs for threading.
1160
1161 2014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1162
1163 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1164 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1165 machine descriptions for Stack Smashing Protector.
1166
1167 2014-04-23 Richard Earnshaw <rearnsha@arm.com>
1168
1169 * aarch64.md (<optab>_rol<mode>3): New pattern.
1170 (<optab>_rolsi3_uxtw): Likewise.
1171 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1172
1173 2014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1174
1175 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1176 (arm_cortex_a12_tune): Likewise.
1177
1178 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1179
1180 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1181
1182 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1183
1184 * config/arm/arm.md (arm_rev16si2): New pattern.
1185 (arm_rev16si2_alt): Likewise.
1186 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1187
1188 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1189
1190 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1191 (rev16<mode>2_alt): Likewise.
1192 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1193 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1194 (aarch_rev16_shleft_mask_imm_p): Likewise.
1195 (aarch_rev16_p_1): Likewise.
1196 (aarch_rev16_p): Likewise.
1197 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1198 (aarch_rev16_shright_mask_imm_p): Likewise.
1199 (aarch_rev16_shleft_mask_imm_p): Likewise.
1200
1201 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1202
1203 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1204 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1205 rev cost.
1206 (cortex_a53_extra_costs): Likewise.
1207 (cortex_a57_extra_costs): Likewise.
1208 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1209 (cortexa7_extra_costs): Likewise.
1210 (cortexa8_extra_costs): Likewise.
1211 (cortexa12_extra_costs): Likewise.
1212 (cortexa15_extra_costs): Likewise.
1213 (v7m_extra_costs): Likewise.
1214 (arm_new_rtx_costs): Handle BSWAP.
1215
1216 2013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1217
1218 * config/arm/arm.c (cortexa8_extra_costs): New table.
1219 (arm_cortex_a8_tune): New tuning struct.
1220 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1221
1222 2014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1223
1224 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1225
1226 2014-04-23 Richard Biener <rguenther@suse.de>
1227
1228 * Makefile.in (OBJS): Remove loop-unswitch.o.
1229 * loop-unswitch.c: Delete.
1230 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1231 * passes.def (pass_rtl_unswitch): Likewise.
1232 * loop-init.c (gate_rtl_unswitch): Likewise.
1233 (rtl_unswitch): Likewise.
1234 (pass_data_rtl_unswitch): Likewise.
1235 (pass_rtl_unswitch): Likewise.
1236 (make_pass_rtl_unswitch): Likewise.
1237 * rtl.h (reversed_condition): Likewise.
1238 (compare_and_jump_seq): Likewise.
1239 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1240 and make static.
1241 * loop-unroll.c (compare_and_jump_seq): Likewise.
1242
1243 2014-04-23 Richard Biener <rguenther@suse.de>
1244
1245 PR tree-optimization/60903
1246 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1247 commented code block.
1248 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1249 loop flags to newly created BBs and edges.
1250
1251 2014-04-23 Nick Clifton <nickc@redhat.com>
1252
1253 * config/msp430/msp430.c (msp430_handle_option): Move function
1254 to msp430-common.c
1255 (msp430_option_override): Simplify mcu and mcpu option handling.
1256 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1257 support for -mhwmult command line option.
1258 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1259 -mhwmult command line option.
1260 (msp430_hwmult_enabled): Delete.
1261 (msp43o_output_labelref): Add support for -mhwmult command line
1262 option.
1263 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1264 (umulsidi3): Likewise.
1265 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1266 (mcpu, mlarge, msmall): Likewise.
1267 (mhwmult): New option.
1268 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1269 prototype.
1270 (msp430_is_f5_mcu): Remove prototype.
1271 (msp430_use_f5_series_hwmult): Add prototype.
1272 * config/msp430/msp430-opts.h: New file.
1273 * common/config/msp430: New directory.
1274 * common/config/msp430/msp430-common.c: New file.
1275 * config.gcc (msp430): Remove target_has_targetm_common.
1276 * doc/invoke.texi: Document -mhwmult command line option.
1277
1278 2014-04-23 Nick Clifton <nickc@redhat.com>
1279
1280 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1281 default-manifest.o if it can be found in the search path.
1282 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1283
1284 2014-04-23 Terry Guo <terry.guo@arm.com>
1285
1286 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1287
1288 2014-04-23 Richard Biener <rguenther@suse.de>
1289
1290 PR middle-end/60895
1291 * tree-inline.c (declare_return_variable): Use mark_addressable.
1292
1293 2014-04-23 Richard Biener <rguenther@suse.de>
1294
1295 PR middle-end/60891
1296 * loop-init.c (loop_optimizer_init): Make sure to apply
1297 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1298
1299 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1300
1301 PR sanitizer/60275
1302 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1303 New options.
1304 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1305 if flag_sanitize_undefined_trap_on_error.
1306 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1307 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1308 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1309 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1310 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1311 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1312 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1313 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1314 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1315 * ubsan.c (ubsan_instrument_unreachable): Return
1316 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1317 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1318 if flag_sanitize_undefined_trap_on_error and
1319 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1320 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1321 instrument_bool_enum_load): Emit __builtin_trap () if
1322 flag_sanitize_undefined_trap_on_error and
1323 __builtin_handle_*_abort () if !flag_sanitize_recover.
1324 * doc/invoke.texi (-fsanitize-recover,
1325 -fsanitize-undefined-trap-on-error): Document.
1326
1327 2014-04-22 Christian Bruel <christian.bruel@st.com>
1328
1329 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1330 Force immediates to SImode.
1331
1332 2014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1333
1334 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1335 (lroundsfsi2): New.
1336 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1337 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1338 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1339 (nios2_fpu_insn): Add entry for round.
1340 (N2FPU_NO_ERRNO_P): Define.
1341 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1342 flag_errno_math.
1343 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1344
1345 2014-04-22 Richard Henderson <rth@redhat.com>
1346
1347 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1348 (add<GPI>3_compare0): Remove leading * from name.
1349 (add<GPI>3_carryin): Likewise.
1350 (sub<GPI>3_compare0): Likewise.
1351 (sub<GPI>3_carryin): Likewise.
1352 (<su_optab>mulditi3): New expander.
1353 (multi3): New expander.
1354 (madd<GPI>): Remove leading * from name.
1355
1356 2014-04-22 Martin Jambor <mjambor@suse.cz>
1357
1358 * cgraphclones.c (cgraph_function_versioning): Copy
1359 ipa_transforms_to_apply instead of asserting it is empty.
1360
1361 2014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1362
1363 PR target/60868
1364 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
1365 on count_exp to get mode.
1366
1367 2014-04-22 Andrew Pinski <apinski@cavium.com>
1368
1369 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1370 Handle TLS for ILP32.
1371 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1372 (tlsie_small_<mode>): this and handle PTR.
1373 (tlsie_small_sidi): New pattern.
1374 (tlsle_small): Change to an expand to handle ILP32.
1375 (tlsle_small_<mode>): New pattern.
1376 (tlsdesc_small): Rename to ...
1377 (tlsdesc_small_<mode>): this and handle PTR.
1378
1379 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1380
1381 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1382
1383 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1384
1385 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1386 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1387 (aarch64_types_signed_poly_qualifiers): Likewise.
1388 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1389 (aarch64_types_poly_signed_qualifiers): Likewise.
1390 (TYPES_REINTERP_SS): Type macro added.
1391 (TYPES_REINTERP_SU): Likewise.
1392 (TYPES_REINTERP_SP): Likewise.
1393 (TYPES_REINTERP_US): Likewise.
1394 (TYPES_REINTERP_PS): Likewise.
1395 (aarch64_fold_builtin): New expression folding added.
1396 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1397 Declarations removed.
1398 (REINTERP_SS): Declarations added.
1399 (REINTERP_US): Likewise.
1400 (REINTERP_PS): Likewise.
1401 (REINTERP_SU): Likewise.
1402 (REINTERP_SP): Likewise.
1403 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1404 (vreinterpretq_p8_f64): Likewise.
1405 (vreinterpret_p16_f64): Likewise.
1406 (vreinterpretq_p16_f64): Likewise.
1407 (vreinterpret_f32_f64): Likewise.
1408 (vreinterpretq_f32_f64): Likewise.
1409 (vreinterpret_f64_f32): Likewise.
1410 (vreinterpret_f64_p8): Likewise.
1411 (vreinterpret_f64_p16): Likewise.
1412 (vreinterpret_f64_s8): Likewise.
1413 (vreinterpret_f64_s16): Likewise.
1414 (vreinterpret_f64_s32): Likewise.
1415 (vreinterpret_f64_s64): Likewise.
1416 (vreinterpret_f64_u8): Likewise.
1417 (vreinterpret_f64_u16): Likewise.
1418 (vreinterpret_f64_u32): Likewise.
1419 (vreinterpret_f64_u64): Likewise.
1420 (vreinterpretq_f64_f32): Likewise.
1421 (vreinterpretq_f64_p8): Likewise.
1422 (vreinterpretq_f64_p16): Likewise.
1423 (vreinterpretq_f64_s8): Likewise.
1424 (vreinterpretq_f64_s16): Likewise.
1425 (vreinterpretq_f64_s32): Likewise.
1426 (vreinterpretq_f64_s64): Likewise.
1427 (vreinterpretq_f64_u8): Likewise.
1428 (vreinterpretq_f64_u16): Likewise.
1429 (vreinterpretq_f64_u32): Likewise.
1430 (vreinterpretq_f64_u64): Likewise.
1431 (vreinterpret_s64_f64): Likewise.
1432 (vreinterpretq_s64_f64): Likewise.
1433 (vreinterpret_u64_f64): Likewise.
1434 (vreinterpretq_u64_f64): Likewise.
1435 (vreinterpret_s8_f64): Likewise.
1436 (vreinterpretq_s8_f64): Likewise.
1437 (vreinterpret_s16_f64): Likewise.
1438 (vreinterpretq_s16_f64): Likewise.
1439 (vreinterpret_s32_f64): Likewise.
1440 (vreinterpretq_s32_f64): Likewise.
1441 (vreinterpret_u8_f64): Likewise.
1442 (vreinterpretq_u8_f64): Likewise.
1443 (vreinterpret_u16_f64): Likewise.
1444 (vreinterpretq_u16_f64): Likewise.
1445 (vreinterpret_u32_f64): Likewise.
1446 (vreinterpretq_u32_f64): Likewise.
1447
1448 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1449
1450 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1451 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1452 (vreinterpret_p8_s8): Likewise.
1453 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1454 (vreinterpret_p8_s16): Likewise.
1455 (vreinterpret_p8_s32): Likewise.
1456 (vreinterpret_p8_s64): Likewise.
1457 (vreinterpret_p8_f32): Likewise.
1458 (vreinterpret_p8_u8): Likewise.
1459 (vreinterpret_p8_u16): Likewise.
1460 (vreinterpret_p8_u32): Likewise.
1461 (vreinterpret_p8_u64): Likewise.
1462 (vreinterpret_p8_p16): Likewise.
1463 (vreinterpretq_p8_s8): Likewise.
1464 (vreinterpretq_p8_s16): Likewise.
1465 (vreinterpretq_p8_s32): Likewise.
1466 (vreinterpretq_p8_s64): Likewise.
1467 (vreinterpretq_p8_f32): Likewise.
1468 (vreinterpretq_p8_u8): Likewise.
1469 (vreinterpretq_p8_u16): Likewise.
1470 (vreinterpretq_p8_u32): Likewise.
1471 (vreinterpretq_p8_u64): Likewise.
1472 (vreinterpretq_p8_p16): Likewise.
1473 (vreinterpret_p16_s8): Likewise.
1474 (vreinterpret_p16_s16): Likewise.
1475 (vreinterpret_p16_s32): Likewise.
1476 (vreinterpret_p16_s64): Likewise.
1477 (vreinterpret_p16_f32): Likewise.
1478 (vreinterpret_p16_u8): Likewise.
1479 (vreinterpret_p16_u16): Likewise.
1480 (vreinterpret_p16_u32): Likewise.
1481 (vreinterpret_p16_u64): Likewise.
1482 (vreinterpret_p16_p8): Likewise.
1483 (vreinterpretq_p16_s8): Likewise.
1484 (vreinterpretq_p16_s16): Likewise.
1485 (vreinterpretq_p16_s32): Likewise.
1486 (vreinterpretq_p16_s64): Likewise.
1487 (vreinterpretq_p16_f32): Likewise.
1488 (vreinterpretq_p16_u8): Likewise.
1489 (vreinterpretq_p16_u16): Likewise.
1490 (vreinterpretq_p16_u32): Likewise.
1491 (vreinterpretq_p16_u64): Likewise.
1492 (vreinterpretq_p16_p8): Likewise.
1493 (vreinterpret_f32_s8): Likewise.
1494 (vreinterpret_f32_s16): Likewise.
1495 (vreinterpret_f32_s32): Likewise.
1496 (vreinterpret_f32_s64): Likewise.
1497 (vreinterpret_f32_u8): Likewise.
1498 (vreinterpret_f32_u16): Likewise.
1499 (vreinterpret_f32_u32): Likewise.
1500 (vreinterpret_f32_u64): Likewise.
1501 (vreinterpret_f32_p8): Likewise.
1502 (vreinterpret_f32_p16): Likewise.
1503 (vreinterpretq_f32_s8): Likewise.
1504 (vreinterpretq_f32_s16): Likewise.
1505 (vreinterpretq_f32_s32): Likewise.
1506 (vreinterpretq_f32_s64): Likewise.
1507 (vreinterpretq_f32_u8): Likewise.
1508 (vreinterpretq_f32_u16): Likewise.
1509 (vreinterpretq_f32_u32): Likewise.
1510 (vreinterpretq_f32_u64): Likewise.
1511 (vreinterpretq_f32_p8): Likewise.
1512 (vreinterpretq_f32_p16): Likewise.
1513 (vreinterpret_s64_s8): Likewise.
1514 (vreinterpret_s64_s16): Likewise.
1515 (vreinterpret_s64_s32): Likewise.
1516 (vreinterpret_s64_f32): Likewise.
1517 (vreinterpret_s64_u8): Likewise.
1518 (vreinterpret_s64_u16): Likewise.
1519 (vreinterpret_s64_u32): Likewise.
1520 (vreinterpret_s64_u64): Likewise.
1521 (vreinterpret_s64_p8): Likewise.
1522 (vreinterpret_s64_p16): Likewise.
1523 (vreinterpretq_s64_s8): Likewise.
1524 (vreinterpretq_s64_s16): Likewise.
1525 (vreinterpretq_s64_s32): Likewise.
1526 (vreinterpretq_s64_f32): Likewise.
1527 (vreinterpretq_s64_u8): Likewise.
1528 (vreinterpretq_s64_u16): Likewise.
1529 (vreinterpretq_s64_u32): Likewise.
1530 (vreinterpretq_s64_u64): Likewise.
1531 (vreinterpretq_s64_p8): Likewise.
1532 (vreinterpretq_s64_p16): Likewise.
1533 (vreinterpret_u64_s8): Likewise.
1534 (vreinterpret_u64_s16): Likewise.
1535 (vreinterpret_u64_s32): Likewise.
1536 (vreinterpret_u64_s64): Likewise.
1537 (vreinterpret_u64_f32): Likewise.
1538 (vreinterpret_u64_u8): Likewise.
1539 (vreinterpret_u64_u16): Likewise.
1540 (vreinterpret_u64_u32): Likewise.
1541 (vreinterpret_u64_p8): Likewise.
1542 (vreinterpret_u64_p16): Likewise.
1543 (vreinterpretq_u64_s8): Likewise.
1544 (vreinterpretq_u64_s16): Likewise.
1545 (vreinterpretq_u64_s32): Likewise.
1546 (vreinterpretq_u64_s64): Likewise.
1547 (vreinterpretq_u64_f32): Likewise.
1548 (vreinterpretq_u64_u8): Likewise.
1549 (vreinterpretq_u64_u16): Likewise.
1550 (vreinterpretq_u64_u32): Likewise.
1551 (vreinterpretq_u64_p8): Likewise.
1552 (vreinterpretq_u64_p16): Likewise.
1553 (vreinterpret_s8_s16): Likewise.
1554 (vreinterpret_s8_s32): Likewise.
1555 (vreinterpret_s8_s64): Likewise.
1556 (vreinterpret_s8_f32): Likewise.
1557 (vreinterpret_s8_u8): Likewise.
1558 (vreinterpret_s8_u16): Likewise.
1559 (vreinterpret_s8_u32): Likewise.
1560 (vreinterpret_s8_u64): Likewise.
1561 (vreinterpret_s8_p8): Likewise.
1562 (vreinterpret_s8_p16): Likewise.
1563 (vreinterpretq_s8_s16): Likewise.
1564 (vreinterpretq_s8_s32): Likewise.
1565 (vreinterpretq_s8_s64): Likewise.
1566 (vreinterpretq_s8_f32): Likewise.
1567 (vreinterpretq_s8_u8): Likewise.
1568 (vreinterpretq_s8_u16): Likewise.
1569 (vreinterpretq_s8_u32): Likewise.
1570 (vreinterpretq_s8_u64): Likewise.
1571 (vreinterpretq_s8_p8): Likewise.
1572 (vreinterpretq_s8_p16): Likewise.
1573 (vreinterpret_s16_s8): Likewise.
1574 (vreinterpret_s16_s32): Likewise.
1575 (vreinterpret_s16_s64): Likewise.
1576 (vreinterpret_s16_f32): Likewise.
1577 (vreinterpret_s16_u8): Likewise.
1578 (vreinterpret_s16_u16): Likewise.
1579 (vreinterpret_s16_u32): Likewise.
1580 (vreinterpret_s16_u64): Likewise.
1581 (vreinterpret_s16_p8): Likewise.
1582 (vreinterpret_s16_p16): Likewise.
1583 (vreinterpretq_s16_s8): Likewise.
1584 (vreinterpretq_s16_s32): Likewise.
1585 (vreinterpretq_s16_s64): Likewise.
1586 (vreinterpretq_s16_f32): Likewise.
1587 (vreinterpretq_s16_u8): Likewise.
1588 (vreinterpretq_s16_u16): Likewise.
1589 (vreinterpretq_s16_u32): Likewise.
1590 (vreinterpretq_s16_u64): Likewise.
1591 (vreinterpretq_s16_p8): Likewise.
1592 (vreinterpretq_s16_p16): Likewise.
1593 (vreinterpret_s32_s8): Likewise.
1594 (vreinterpret_s32_s16): Likewise.
1595 (vreinterpret_s32_s64): Likewise.
1596 (vreinterpret_s32_f32): Likewise.
1597 (vreinterpret_s32_u8): Likewise.
1598 (vreinterpret_s32_u16): Likewise.
1599 (vreinterpret_s32_u32): Likewise.
1600 (vreinterpret_s32_u64): Likewise.
1601 (vreinterpret_s32_p8): Likewise.
1602 (vreinterpret_s32_p16): Likewise.
1603 (vreinterpretq_s32_s8): Likewise.
1604 (vreinterpretq_s32_s16): Likewise.
1605 (vreinterpretq_s32_s64): Likewise.
1606 (vreinterpretq_s32_f32): Likewise.
1607 (vreinterpretq_s32_u8): Likewise.
1608 (vreinterpretq_s32_u16): Likewise.
1609 (vreinterpretq_s32_u32): Likewise.
1610 (vreinterpretq_s32_u64): Likewise.
1611 (vreinterpretq_s32_p8): Likewise.
1612 (vreinterpretq_s32_p16): Likewise.
1613 (vreinterpret_u8_s8): Likewise.
1614 (vreinterpret_u8_s16): Likewise.
1615 (vreinterpret_u8_s32): Likewise.
1616 (vreinterpret_u8_s64): Likewise.
1617 (vreinterpret_u8_f32): Likewise.
1618 (vreinterpret_u8_u16): Likewise.
1619 (vreinterpret_u8_u32): Likewise.
1620 (vreinterpret_u8_u64): Likewise.
1621 (vreinterpret_u8_p8): Likewise.
1622 (vreinterpret_u8_p16): Likewise.
1623 (vreinterpretq_u8_s8): Likewise.
1624 (vreinterpretq_u8_s16): Likewise.
1625 (vreinterpretq_u8_s32): Likewise.
1626 (vreinterpretq_u8_s64): Likewise.
1627 (vreinterpretq_u8_f32): Likewise.
1628 (vreinterpretq_u8_u16): Likewise.
1629 (vreinterpretq_u8_u32): Likewise.
1630 (vreinterpretq_u8_u64): Likewise.
1631 (vreinterpretq_u8_p8): Likewise.
1632 (vreinterpretq_u8_p16): Likewise.
1633 (vreinterpret_u16_s8): Likewise.
1634 (vreinterpret_u16_s16): Likewise.
1635 (vreinterpret_u16_s32): Likewise.
1636 (vreinterpret_u16_s64): Likewise.
1637 (vreinterpret_u16_f32): Likewise.
1638 (vreinterpret_u16_u8): Likewise.
1639 (vreinterpret_u16_u32): Likewise.
1640 (vreinterpret_u16_u64): Likewise.
1641 (vreinterpret_u16_p8): Likewise.
1642 (vreinterpret_u16_p16): Likewise.
1643 (vreinterpretq_u16_s8): Likewise.
1644 (vreinterpretq_u16_s16): Likewise.
1645 (vreinterpretq_u16_s32): Likewise.
1646 (vreinterpretq_u16_s64): Likewise.
1647 (vreinterpretq_u16_f32): Likewise.
1648 (vreinterpretq_u16_u8): Likewise.
1649 (vreinterpretq_u16_u32): Likewise.
1650 (vreinterpretq_u16_u64): Likewise.
1651 (vreinterpretq_u16_p8): Likewise.
1652 (vreinterpretq_u16_p16): Likewise.
1653 (vreinterpret_u32_s8): Likewise.
1654 (vreinterpret_u32_s16): Likewise.
1655 (vreinterpret_u32_s32): Likewise.
1656 (vreinterpret_u32_s64): Likewise.
1657 (vreinterpret_u32_f32): Likewise.
1658 (vreinterpret_u32_u8): Likewise.
1659 (vreinterpret_u32_u16): Likewise.
1660 (vreinterpret_u32_u64): Likewise.
1661 (vreinterpret_u32_p8): Likewise.
1662 (vreinterpret_u32_p16): Likewise.
1663 (vreinterpretq_u32_s8): Likewise.
1664 (vreinterpretq_u32_s16): Likewise.
1665 (vreinterpretq_u32_s32): Likewise.
1666 (vreinterpretq_u32_s64): Likewise.
1667 (vreinterpretq_u32_f32): Likewise.
1668 (vreinterpretq_u32_u8): Likewise.
1669 (vreinterpretq_u32_u16): Likewise.
1670 (vreinterpretq_u32_u64): Likewise.
1671 (vreinterpretq_u32_p8): Likewise.
1672 (vreinterpretq_u32_p16): Likewise.
1673
1674 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1675
1676 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1677 Pattern extended.
1678 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1679 extended.
1680 (sqabs): Likewise.
1681 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1682 (vqnegd_s64): Likewise.
1683 (vqabs_s64): Likewise.
1684 (vqabsd_s64): Likewise.
1685
1686 2014-04-22 Richard Henderson <rth@redhat.com>
1687
1688 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1689 computation to the top of the loop.
1690
1691 2014-04-22 Renlin <renlin.li@arm.com>
1692 Jiong Wang <jiong.wang@arm.com>
1693
1694 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1695 * config/aarch64/aarch64.c (aarch64_layout_frame)
1696 (aarch64_initial_elimination_offset): Likewise.
1697
1698 2014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1699
1700 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1701 Fix indentation.
1702
1703 2014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1704
1705 * machmode.h (bitwise_mode_for_mode): Declare.
1706 * stor-layout.h (bitwise_type_for_mode): Likewise.
1707 * stor-layout.c (bitwise_mode_for_mode): New function.
1708 (bitwise_type_for_mode): Likewise.
1709 * builtins.c (fold_builtin_memory_op): Use it instead of
1710 int_mode_for_mode and build_nonstandard_integer_type.
1711
1712 2014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1713
1714 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1715 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1716 (*-*-solaris2*): Simplify.
1717 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1718 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1719 *-*-solaris2.9* handling.
1720
1721 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1722 as bug.
1723 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1724 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1725 handling, simplify.
1726 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1727 * configure: Regenerate.
1728
1729 * config/i386/sol2-9.h: Remove.
1730
1731 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1732 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1733 Remove Solaris 9 references.
1734
1735 2014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1736
1737 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1738 (floatuns<GPI:mode><GPF:mode>2): Remove.
1739 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1740 and floatuns conversions.
1741 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1742 and floatuns conversions.
1743 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1744 (w1,w2): New mode attributes for inequal width conversions.
1745
1746 2014-04-22 Renlin Li <Renlin.Li@arm.com>
1747
1748 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1749 the output asm format.
1750
1751 2014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1752
1753 * config/aarch64/aarch64-simd.md
1754 (aarch64_cm<optab>di): Always split.
1755 (*aarch64_cm<optab>di): New.
1756 (aarch64_cmtstdi): Always split.
1757 (*aarch64_cmtstdi): New.
1758
1759 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1760
1761 PR tree-optimization/60823
1762 * omp-low.c (ipa_simd_modify_function_body): Go through
1763 all SSA_NAMEs and for those refering to vector arguments
1764 which are going to be replaced adjust SSA_NAME_VAR and,
1765 if it is a default definition, change it into a non-default
1766 definition assigned at the beginning of function from new_decl.
1767 (ipa_simd_modify_stmt_ops): Rewritten.
1768 * tree-dfa.c (set_ssa_default_def): When removing default def,
1769 check for NULL loc instead of NULL *loc.
1770
1771 2014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1772
1773 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1774 restrictions on core registers for DImode values in Thumb2.
1775
1776 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1777
1778 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1779 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1780
1781 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1782
1783 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1784 (*iordi_notzesidi_di): Likewise.
1785 (*iordi_notsesidi_di): Likewise.
1786
1787 2014-04-22 Ian Bolton <ian.bolton@arm.com>
1788
1789 * config/arm/arm-protos.h (tune_params): New struct members.
1790 * config/arm/arm.c: Initialise tune_params per processor.
1791 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1792 for speed, based on new tune_params.
1793
1794 2014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1795
1796 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1797 added.
1798 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1799 macro.
1800 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1801 corrected.
1802 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1803 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1804 (vrnda_f64): Likewise.
1805 (vrndi_f64): Likewise.
1806 (vrndm_f64): Likewise.
1807 (vrndn_f64): Likewise.
1808 (vrndp_f64): Likewise.
1809 (vrndx_f64): Likewise.
1810
1811 2014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1812
1813 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1814 GET_MODE_SIZE argument is enum machine_mode.
1815
1816 2014-04-22 Jakub Jelinek <jakub@redhat.com>
1817
1818 PR target/60910
1819 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1820 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1821
1822 2014-04-22 Lin Zuojian <manjian2006@gmail.com>
1823
1824 PR middle-end/60281
1825 * asan.c (asan_emit_stack_protection): Force the base to align to
1826 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1827 appropriate bits if STRICT_ALIGNMENT.
1828 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1829 when asan is on.
1830 (expand_used_vars): Leave a space in the stack frame for alignment
1831 if STRICT_ALIGNMENT.
1832
1833 2014-04-21 David Malcolm <dmalcolm@redhat.com>
1834
1835 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1836 than a gimple.
1837 (gimple_store_p): Likewise.
1838 (gimple_assign_load_p): Likewise.
1839 (gimple_assign_cast_p): Likewise.
1840 (gimple_clobber_p): Likewise.
1841
1842 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1843 rather than a gimple.
1844 (gimple_assign_cast_p): Likewise.
1845
1846 2014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1847
1848 PR target/60735
1849 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1850 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1851
1852 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1853 more debug information for E500 if -mdebug=reg.
1854
1855 2014-04-21 Uros Bizjak <ubizjak@gmail.com>
1856
1857 PR target/60909
1858 * config/i386/i386.c (ix86_expand_builtin)
1859 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1860 register for target RTX.
1861 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1862
1863 2014-04-18 Cong Hou <congh@google.com>
1864
1865 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
1866 the widen-mult pattern by handling two operands with different sizes,
1867 and operands whose size is smaller than half of the result type.
1868
1869 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1870
1871 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1872 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1873 (do_estimate_edge_time): Compute it.
1874 * ipa-inline.c (want_inline_small_function_p): Bypass
1875 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
1876
1877 2014-04-18 Jan Hubicka <hubicka@ucw.cz>
1878
1879 * ipa-inline.c (spec_rem): New static variable.
1880 (dump_overall_stats): New function.
1881 (dump_inline_stats): New function.
1882
1883 2014-04-18 Richard Henderson <rth@redhat.com>
1884
1885 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1886 to GET_MODE_SIZE, not a reg_class_t.
1887
1888 2014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1889
1890 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
1891 (vsx_xxmrglw_<mode>): Likewise.
1892
1893 2014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1894
1895 PR target/60876
1896 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
1897 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
1898 (rs6000_init_hard_regno_mode_ok): Likewise.
1899
1900 2014-04-17 Jan Hubicka <hubicka@ucw.cz>
1901
1902 * ipa-inline.c (inline_small_functions): Account only non-cold
1903 functions.
1904 * doc/invoke.texi (inline-unit-growth): Update documentation.
1905
1906 2014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1907
1908 * config/rs6000/rs6000.md (addti3, subti3): New.
1909
1910 2014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1911
1912 PR target/60863
1913 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1914 comment. Check optimize_insn_for_size_p instead of
1915 optimize_insn_for_speed_p.
1916
1917 2014-04-17 Martin Jambor <mjambor@suse.cz>
1918
1919 * gimple-iterator.c (gsi_start_edge): New function.
1920 * gimple-iterator.h (gsi_start_edge): Declare.
1921 * tree-sra.c (single_non_eh_succ): New function.
1922 (disqualify_ops_if_throwing_stmt): Renamed to
1923 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1924 having one non-EH successor BB.
1925 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1926 generate loads into replacements.
1927 (sra_modify_assign): Likewise and and also use the simple path for
1928 such statements.
1929 (sra_modify_function_body): Commit statements on edges.
1930
1931 2014-04-17 Richard Biener <rguenther@suse.de>
1932
1933 PR middle-end/60849
1934 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1935 comparison results and add clarifying comment.
1936
1937 2014-04-17 Jakub Jelinek <jakub@redhat.com>
1938
1939 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1940 (blank_mode): Initialize it.
1941 (emit_mode_size_inline, emit_mode_nunits_inline,
1942 emit_mode_inner_inline): New functions.
1943 (emit_insn_modes_h): Call them and surround their output with
1944 #if GCC_VERSION >= 4001 ... #endif.
1945 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1946 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1947 mode_* arrays if the argument is __builtin_constant_p.
1948 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1949 is enum machine_mode.
1950
1951 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1952
1953 * passes.c (opt_pass::execute): Adjust.
1954 (pass_manager::execute_pass_mode_switching): Likewise.
1955 (early_local_passes::execute): Likewise.
1956 (execute_one_pass): Pass cfun to the pass's execute method.
1957 * tree-pass.h (opt_pass::execute): Add function * argument.
1958 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1959 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1960 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1961 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1962 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1963 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1964 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1965 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1966 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1967 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1968 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1969 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1970 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1971 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1972 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1973 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1974 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1975 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1976 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1977 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1978 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1979 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1980 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1981 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1982 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1983 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1984 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1985 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1986 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1987 Adjust.
1988
1989 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1990
1991 * passes.c (opt_pass::gate): Take function * argument.
1992 (gate_all_early_local_passes): Merge into
1993 (early_local_passes::gate): this.
1994 (gate_all_early_optimizations): Merge into
1995 (all_early_optimizations::gate): this.
1996 (gate_all_optimizations): Mege into
1997 (all_optimizations::gate): this.
1998 (gate_all_optimizations_g): Merge into
1999 (all_optimizations_g::gate): this.
2000 (gate_rest_of_compilation): Mege into
2001 (rest_of_compilation::gate): this.
2002 (gate_postreload): Merge into
2003 (postreload::gate): this.
2004 (dump_one_pass): Pass cfun to the pass's gate method.
2005 (execute_ipa_summary_passes): Likewise.
2006 (execute_one_pass): Likewise.
2007 (ipa_write_summaries_2): Likewise.
2008 (ipa_write_optimization_summaries_1): Likewise.
2009 (ipa_read_summaries_1): Likewise.
2010 (ipa_read_optimization_summaries_1): Likewise.
2011 (execute_ipa_stmt_fixups): Likewise.
2012 * tree-pass.h (opt_pass::gate): Add function * argument.
2013 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
2014 combine-stack-adj.c, combine.c, compare-elim.c,
2015 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2016 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
2017 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
2018 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
2019 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
2020 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
2021 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2022 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
2023 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
2024 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
2025 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2026 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2027 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2028 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2029 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2030 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2031 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2032 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2033 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2034 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2035 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2036 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2037 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2038 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2039 var-tracking.c, vtable-verify.c, web.c: Adjust.
2040
2041 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2042
2043 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2044 * configure: Regenerate.
2045
2046 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2047
2048 * passes.c (dump_one_pass): don't check pass->has_gate.
2049 (execute_ipa_summary_passes): Likewise.
2050 (execute_one_pass): Likewise.
2051 (ipa_write_summaries_2): Likewise.
2052 (ipa_write_optimization_summaries_1): Likewise.
2053 (ipa_read_optimization_summaries_1): Likewise.
2054 (execute_ipa_stmt_fixups): Likewise.
2055 * tree-pass.h (pass_data::has_gate): Remove.
2056 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2057 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2058 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2059 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2060 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2061 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2062 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2063 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2064 gimple-low.c, gimple-ssa-isolate-paths.c,
2065 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2066 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2067 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2068 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2069 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2070 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2071 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2072 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2073 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2074 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2075 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2076 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2077 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2078 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2079 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2080 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2081 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2082 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2083 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2084 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2085 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2086 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2087 Adjust.
2088
2089 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2090
2091 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2092 declaration.
2093 * passes.c (pass_manager::register_dump_files_1): Merge into
2094 (pass_manager::register_dump_files): this, and remove its handling of
2095 properties since the pass always has the properties anyway.
2096 (pass_manager::pass_manager): Adjust.
2097
2098 2014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2099
2100 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2101 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2102 dealing with properties.
2103 (pass_manager::register_dump_files): Adjust.
2104
2105 2014-03-20 Mark Wielaard <mjw@redhat.com>
2106
2107 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2108 then represent the bound as normal constant value.
2109
2110 2014-04-17 Jakub Jelinek <jakub@redhat.com>
2111
2112 PR target/60847
2113 Forward port from 4.8 branch
2114 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2115
2116 * config/i386/bmiintrin.h (_blsi_u32): New.
2117 (_blsi_u64): Ditto.
2118 (_blsr_u32): Ditto.
2119 (_blsr_u64): Ditto.
2120 (_blsmsk_u32): Ditto.
2121 (_blsmsk_u64): Ditto.
2122 (_tzcnt_u32): Ditto.
2123 (_tzcnt_u64): Ditto.
2124
2125 2014-04-17 Kito Cheng <kito@0xlab.org>
2126
2127 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2128
2129 2014-04-17 Richard Biener <rguenther@suse.de>
2130
2131 PR middle-end/60849
2132 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2133 boolean results for comparisons.
2134
2135 2014-04-17 Richard Biener <rguenther@suse.de>
2136
2137 PR tree-optimization/60836
2138 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2139 initial PHI args to be gimple values.
2140
2141 2014-04-17 Richard Biener <rguenther@suse.de>
2142
2143 PR tree-optimization/60841
2144 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2145 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2146 of stmts to SLP build.
2147 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2148 (vect_analyze_slp): Likewise.
2149 (vect_analyze_slp_instance): Likewise.
2150 (vect_build_slp_tree): Limit overall SLP tree growth.
2151 * tree-vectorizer.h (vect_analyze_data_refs,
2152 vect_analyze_slp): Adjust prototypes.
2153
2154 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2155
2156 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2157 Silvermont.
2158
2159 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2160
2161 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2162 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2163 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2164 for TARGET_SLOW_PSHUFB
2165
2166 2014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2167
2168 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2169 * config/i386/i386.c (intel_cost): Ditto.
2170
2171 2014-04-17 Joey Ye <joey.ye@arm.com>
2172
2173 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2174
2175 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2176
2177 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2178 with -fuse-profile.
2179
2180 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2181
2182 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2183 (type_all_derivations_known_p): New predicate.
2184 (type_all_ctors_visible_p): New predicate.
2185 (type_possibly_instantiated_p): New predicate.
2186 (get_odr_type): Compute all_derivations_known.
2187 (dump_odr_type): Dump the flag.
2188 (maybe_record_type): Cleanup.
2189 (record_target_from_binfo): Add bases_to_consider array;
2190 record bases for types w/o instances and skip CXX destructor.
2191 (possible_polymorphic_call_targets_1): Add bases_to_consider
2192 and consider_construction parameters; check if type may have instance.
2193 (get_polymorphic_call_info): Set maybe_in_construction to true
2194 when we know nothing.
2195 (record_targets_from_bases): Skip CXX destructors; they are
2196 never called for types in construction.
2197 (possible_polymorphic_call_targets): Do not record target when
2198 type may not have instance.
2199
2200 2014-04-16 Jan Hubicka <hubicka@ucw.cz>
2201
2202 PR ipa/60854
2203 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2204 external aliases alive, too.
2205
2206 2014-04-16 Andrew Pinski <apinski@cavium.com>
2207
2208 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
2209 definition.
2210
2211 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2212
2213 * final.c (compute_alignments): Do not apply loop alignment to a block
2214 falling through to the exit.
2215
2216 2014-04-16 Catherine Moore <clm@codesourcery.com>
2217
2218 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2219 Adjust constraints for microMIPS store patterns.
2220
2221 2014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2222
2223 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2224
2225 2014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2226
2227 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2228 (append_use): Run at -O0.
2229 (append_vdef): Likewise.
2230 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2231 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2232
2233 2014-04-16 Jakub Jelinek <jakub@redhat.com>
2234
2235 PR tree-optimization/60844
2236 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2237 (propagate_op_to_single_use, remove_visited_stmt_chain,
2238 linearize_expr, repropagate_negates, reassociate_bb): Use it
2239 instead of gsi_remove.
2240
2241 2014-04-16 Martin Jambor <mjambor@suse.cz>
2242
2243 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2244 ipa_transforms_to_apply.
2245 (cgraph_function_versioning): Assert that old_node has empty
2246 ipa_transforms_to_apply.
2247 * trans-mem.c (ipa_tm_create_version): Likewise.
2248 * tree-inline.c (tree_function_versioning): Do not duplicate
2249 ipa_transforms_to_apply.
2250
2251 2014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2252
2253 PR target/60817
2254 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2255 x86_64-*-* cases.
2256 Pass necessary as flags on 64-bit Solaris/x86.
2257 Use lowercase relocs for x86_64-*-*.
2258 * configure: Regenerate.
2259
2260 2014-04-15 Jan Hubicka <jh@suse.cz>
2261
2262 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2263 (maybe_record_node, likely_target_p): Use it.
2264
2265 2014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2266
2267 PR target/60839
2268 Revert following patch
2269
2270 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2271
2272 PR target/60735
2273 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2274 software floating point or no floating point registers, do not
2275 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2276 in GPRs that occurs after we tested for GPRs that would never be
2277 true.
2278
2279 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2280 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2281 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2282 specifically allow DDmode, since that does not use the SPE SIMD
2283 instructions.
2284
2285 2014-03-21 Mark Wielaard <mjw@redhat.com>
2286
2287 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2288 as unsigned or int depending on type and value used.
2289
2290 2014-04-15 Richard Biener <rguenther@suse.de>
2291
2292 PR rtl-optimization/56965
2293 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2294 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2295 ... here.
2296 * alias.c (true_dependence_1): Do not call
2297 nonoverlapping_component_refs_p.
2298 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2299 nonoverlapping_component_refs_p.
2300 (indirect_refs_may_alias_p): Likewise.
2301
2302 2014-04-15 Teresa Johnson <tejohnson@google.com>
2303
2304 * cfg.c (dump_bb_info): Fix flags check.
2305 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2306
2307 2014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2308
2309 PR rtl-optimization/60663
2310 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2311 avoid 0 cost.
2312
2313 2014-04-15 Richard Biener <rguenther@suse.de>
2314
2315 * lto-streamer.h (LTO_major_version): Bump to 4.
2316
2317 2014-04-15 Richard Biener <rguenther@suse.de>
2318
2319 * common.opt (lto_partition_model): New enum.
2320 (flto-partition=): Merge separate options with a single with argument,
2321 add -flto-partition=one support.
2322 * flag-types.h (enum lto_partition_model): Declare.
2323 * opts.c (finish_options): Remove duplicate -flto-partition=
2324 option check.
2325 * lto-wrapper.c (run_gcc): Adjust.
2326
2327 2014-04-15 Richard Biener <rguenther@suse.de>
2328
2329 * alias.c (ncr_compar): New function.
2330 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2331
2332 2014-04-15 Richard Biener <rguenther@suse.de>
2333
2334 * alias.c (record_component_aliases): Do not walk BINFOs.
2335
2336 2014-04-15 Richard Biener <rguenther@suse.de>
2337
2338 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2339 Add struct function argument and adjust.
2340 (find_func_aliases_for_call): Likewise.
2341 (find_func_aliases): Likewise.
2342 (find_func_clobbers): Likewise.
2343 (intra_create_variable_infos): Likewise.
2344 (compute_points_to_sets): Likewise.
2345 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2346
2347 2014-04-15 Richard Biener <rguenther@suse.de>
2348
2349 * tree.c (iterative_hash_expr): Use enum tree_code_class
2350 to store TREE_CODE_CLASS.
2351 (tree_block): Likewise.
2352 (tree_set_block): Likewise.
2353 * tree.h (fold_build_pointer_plus_loc): Use
2354 convert_to_ptrofftype_loc.
2355
2356 2014-04-15 Jakub Jelinek <jakub@redhat.com>
2357
2358 PR plugins/59335
2359 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2360 added in 4.9.
2361
2362 2014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2363
2364 * cfgloop.h (struct loop): Move force_vectorize down.
2365 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2366 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2367 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2368 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2369 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2370 * tree-core.h (enum annot_expr_kind): Add new kind values.
2371 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2372 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2373 kinds.
2374 * tree.def (ANNOTATE_EXPR): Tweak comment.
2375
2376 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2377
2378 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2379 cxa_pure_virtual).
2380
2381 2014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2382
2383 * tree.h (TYPE_IDENTIFIER): Declare.
2384 * tree.c (subrange_type_for_debug_p): Use it.
2385 * godump.c (go_format_type): Likewise.
2386 * dwarf2out.c (is_cxx_auto, modified_type_die,
2387 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2388 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2389
2390 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2391
2392 PR lto/60820
2393 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2394
2395 2014-04-14 Uros Bizjak <ubizjak@gmail.com>
2396
2397 * config/i386/i386.c (examine_argument): Return bool. Return true if
2398 parameter should be passed in memory.
2399 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2400 (construct_container): Update calls to examine_argument.
2401 (function_arg_advance_64): Ditto.
2402 (return_in_memory_32): Merge with ix86_return_in_memory.
2403 (return_in_memory_64): Ditto.
2404 (return_in_memory_ms_64): Ditto.
2405
2406 2014-04-14 Jan Hubicka <hubicka@ucw.cz>
2407
2408 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2409 * coverage.c (coverage_compute_profile_id): Handle externally visible
2410 symbols.
2411
2412 2014-04-14 Martin Jambor <mjambor@suse.cz>
2413
2414 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2415 DECL_DISREGARD_INLINE_LIMITS functions.
2416
2417 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2418
2419 PR target/60827
2420 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2421
2422 2014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2423
2424 PR target/60827
2425 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2426 optimize_insn_for_speed_p instead of
2427 optimize_function_for_speed_p.
2428
2429 2014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2430
2431 * doc/invoke.texi (free): Document AArch64.
2432
2433 2014-04-14 Richard Biener <rguenther@suse.de>
2434
2435 PR tree-optimization/60042
2436 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2437 (insert_into_preds_of_block): Do not prevent PHI insertion
2438 for REFERENCE exprs here ...
2439 (eliminate_dom_walker::before_dom_children): ... but prevent
2440 their use here under similar conditions when applied to the
2441 IL after PRE optimizations.
2442
2443 2014-04-14 Richard Biener <rguenther@suse.de>
2444
2445 * passes.def: Move early points-to after early SRA.
2446
2447 2014-04-14 Richard Biener <rguenther@suse.de>
2448
2449 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2450 check for which sign-changes we allow when forwarding
2451 a converted value into a switch.
2452
2453 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2454
2455 * stor-layout.c (place_field): Finalize non-constant offset for the
2456 field, if any.
2457
2458 2014-04-14 Richard Biener <rguenther@suse.de>
2459
2460 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2461 as argument.
2462 (expand_switch_using_bit_tests_p): Likewise.
2463 (process_switch): Compute and pass on speed_p based on the
2464 switch stmt.
2465 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2466 optimize_bb_for_speed_p.
2467
2468 2014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2469
2470 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2471 * function.h (struct function): Rename has_force_vect_loops into
2472 has_force_vectorize_loops.
2473 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2474 (input_struct_function_base): Likewise.
2475 * lto-streamer-out.c (output_cfg): Likewise.
2476 (output_struct_function_base): Likewise.
2477 * omp-low.c (expand_omp_simd): Likewise.
2478 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2479 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2480 (version_loop_for_if_conversion): Likewise.
2481 (tree_if_conversion): Likewise.
2482 (main_tree_if_conversion): Likewise.
2483 (gate_tree_if_conversion): Likewise.
2484 * tree-inline.c (copy_loops): Likewise.
2485 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2486 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2487 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2488 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2489 * tree-vectorizer.c (vectorize_loops): Likewise.
2490 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2491
2492 2014-04-14 Richard Biener <rguenther@suse.de>
2493
2494 PR lto/60720
2495 * lto-streamer-out.c (wrap_refs): New function.
2496 (lto_output): Wrap symbol references in global initializes in
2497 type-preserving MEM_REFs.
2498
2499 2014-04-14 Christian Bruel <christian.bruel@st.com>
2500
2501 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
2502
2503 2014-04-14 Christian Bruel <christian.bruel@st.com>
2504
2505 * config/sh/sh.md (setmemqi): New expand pattern.
2506 * config/sh/sh.h (CLEAR_RATIO): Define.
2507 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2508 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2509
2510 2014-04-14 Richard Biener <rguenther@suse.de>
2511
2512 PR middle-end/55022
2513 * fold-const.c (negate_expr_p): Don't negate directional rounding
2514 division.
2515 (fold_negate_expr): Likewise.
2516
2517 2014-04-14 Richard Biener <rguenther@suse.de>
2518
2519 PR tree-optimization/59817
2520 PR tree-optimization/60453
2521 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2522 recursion to catch all CHRECs in the scalar evolution and restrict
2523 the predicate for the remains appropriately.
2524
2525 2014-04-12 Catherine Moore <clm@codesourcery.com>
2526
2527 * config/mips/constraints.md: Add new register constraint "kb".
2528 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2529 (*movhi_internal): Likewise.
2530 (*movqi_internal): Likewise.
2531 * config/mips/mips.h (M16_STORE_REGS): New register class.
2532 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2533 (REG_CLASS_CONTENTS): Likewise.
2534 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2535
2536 2014-04-11 Tobias Burnus <burnus@net-b.de>
2537
2538 PR c/60194
2539 * doc/invoke.texi (-Wformat-signedness): Document it.
2540 (Wformat=2): Mention that this enables -Wformat-signedness.
2541
2542 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2543
2544 * common/config/epiphany/epiphany-common.c
2545 (epiphany_option_optimization_table): Enable section anchors by
2546 default at -O1 or higher.
2547 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2548 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2549 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2550 carries no extra cost.
2551 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2552 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2553 * config/epiphany/predicates.md (memclob_operand): New predicate.
2554 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2555 Use memclob_operand predicate and X constraint for operand 3.
2556
2557 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2558
2559 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2560 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2561 its operands.
2562
2563 2014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2564
2565 PR rtl-optimization/60651
2566 * mode-switching.c (optimize_mode_switching): Make sure to emit
2567 sets of a lower numbered entity before sets of a higher numbered
2568 entity to a mode of the same or lower priority.
2569 When creating a seginfo for a basic block that starts with a code
2570 label, move the insertion point past the code label.
2571 (new_seginfo): Document and enforce requirement that
2572 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2573 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2574 * doc/tm.texi: Regenerate.
2575
2576 2014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2577
2578 PR target/60811
2579 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2580
2581 2013-04-11 Jakub Jelinek <jakub@redhat.com>
2582
2583 * BASE-VER: Set to 4.10.0.
2584
2585 2014-04-11 Tobias Burnus <burnus@net-b.de>
2586
2587 PR other/59055
2588 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
2589 * doc/gcc.texi (Service): Update description in the @menu
2590 * doc/invoke.texi (Option Summary): Remove misplaced and
2591 duplicated @menu.
2592
2593 2014-04-11 Steve Ellcey <sellcey@mips.com>
2594 Jakub Jelinek <jakub@redhat.com>
2595
2596 PR middle-end/60556
2597 * expr.c (convert_move): Use emit_store_flag_force instead of
2598 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2599 argument to it.
2600
2601 2014-04-11 Richard Biener <rguenther@suse.de>
2602
2603 PR middle-end/60797
2604 * varasm.c (assemble_alias): Avoid endless error reporting
2605 recursion by setting TREE_ASM_WRITTEN.
2606
2607 2014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2608
2609 * config/s390/s390.md: Add a splitter for NOT rtx.
2610
2611 2014-04-11 Jakub Jelinek <jakub@redhat.com>
2612
2613 PR rtl-optimization/60663
2614 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
2615
2616 2014-04-10 Jan Hubicka <hubicka@ucw.cz>
2617 Jakub Jelinek <jakub@redhat.com>
2618
2619 PR lto/60567
2620 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2621 flag from decl_node to node.
2622
2623 2014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2624
2625 PR debug/60655
2626 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2627 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2628 ameliorating the cases where it can be.
2629
2630 2014-04-09 David Edelsohn <dje.gcc@gmail.com>
2631
2632 Revert
2633 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2634
2635 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2636 (loadsync_<mode>): Change mode.
2637 (load_quadpti, store_quadpti): New.
2638 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2639 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2640 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2641
2642 2014-04-09 Cong Hou <congh@google.com>
2643
2644 PR testsuite/60773
2645 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2646 documentation.
2647
2648 2014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2649
2650 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2651 instead of vnor to exploit possible fusion opportunity in the
2652 future.
2653 (altivec_expand_vec_perm_const_le): Likewise.
2654
2655 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2656
2657 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2658 (loadsync_<mode>): Change mode.
2659 (load_quadpti, store_quadpti): New.
2660 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2661 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2662
2663 2014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2664
2665 PR target/60763
2666 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2667 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2668 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2669
2670 2014-04-08 Richard Biener <rguenther@suse.de>
2671
2672 PR middle-end/60706
2673 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2674 a 64bit widest int print double-int similar to on HWI64 hosts.
2675
2676 2014-04-08 Richard Biener <rguenther@suse.de>
2677
2678 PR tree-optimization/60785
2679 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2680 default defs properly.
2681
2682 2014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2683
2684 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2685 (Weffc++): Likewise.
2686
2687 2014-04-07 Jan Hubicka <hubcika@ucw.cz>
2688
2689 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2690 set completep to false rather than true.
2691
2692 2014-04-07 Douglas B Rupp <rupp@adacore.com>
2693
2694 PR target/60504
2695 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2696 ARM_TARGET2_DWARF_FORMAT.
2697
2698 2014-04-07 Charles Baylis <charles.baylis@linaro.org>
2699
2700 PR target/60609
2701 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2702 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2703 ADDR_DIFF_VEC.
2704
2705 2014-04-07 Richard Biener <rguenther@suse.de>
2706
2707 PR tree-optimization/60766
2708 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2709 (may_eliminate_iv): Convert cand_value_at result to desired type.
2710
2711 2014-04-07 Jason Merrill <jason@redhat.com>
2712
2713 PR c++/60731
2714 * common.opt (-fno-gnu-unique): Add.
2715 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2716
2717 2014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2718
2719 * haifa-sched.c: Fix outdated function reference and minor
2720 grammar errors in introductory comment.
2721
2722 2014-04-07 Richard Biener <rguenther@suse.de>
2723
2724 PR middle-end/60750
2725 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2726 for noreturn calls.
2727 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2728
2729 2014-04-06 John David Anglin <danglin@gcc.gnu.org>
2730
2731 PR debug/55794
2732 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2733 size accounting for thunks.
2734 (pa_asm_output_mi_thunk): Use final_start_function() and
2735 final_end_function() to output function start and end directives.
2736
2737 2014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2738
2739 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2740 device specific ISA/ feature information. Remove short_sp and
2741 errata_skip ds. Add avr_device_specific_features enum to have device
2742 specific info.
2743 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2744 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2745 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2746 updated device specific info.
2747 * config/avr/avr-mcus.def: Merge device specific details to
2748 dev_attribute field.
2749 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2750 errata_skip.
2751 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2752 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2753 assembler if RMW isa supported by current device.
2754 * config/avr/genmultilib.awk: Update as device info structure changed.
2755 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2756
2757 2014-04-04 Cong Hou <congh@google.com>
2758
2759 PR tree-optimization/60656
2760 * tree-vect-stmts.c (supportable_widening_operation):
2761 Fix a bug that elements in a vector with vect_used_by_reduction
2762 property are incorrectly reordered when the operation on it is not
2763 consistant with the one in reduction operation.
2764
2765 2014-04-04 John David Anglin <danglin@gcc.gnu.org>
2766
2767 PR rtl-optimization/60155
2768 * gcse.c (record_set_data): New function.
2769 (single_set_gcse): New function.
2770 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2771 (hoist_code): Likewise.
2772 (get_pressure_class_and_nregs): Likewise.
2773
2774 2014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2775
2776 * explow.c (probe_stack_range): Emit a final optimization blockage.
2777
2778 2014-04-04 Anthony Green <green@moxielogic.com>
2779
2780 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2781 typos.
2782
2783 2014-04-04 Jan Hubicka <hubicka@ucw.cz>
2784
2785 PR ipa/59626
2786 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2787 flags are set only during streaming.
2788 * ipa.c (process_references, walk_polymorphic_call_targets,
2789 symtab_remove_unreachable_nodes): Drop bodies of always inline
2790 after early inlining.
2791 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
2792
2793 2014-04-04 Jakub Jelinek <jakub@redhat.com>
2794 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2795
2796 PR debug/60655
2797 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2798 containing a NOT.
2799
2800 2014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2801
2802 PR bootstrap/60743
2803 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2804 duration.
2805 (cortex_a53_fdivd): Likewise.
2806
2807 2014-04-04 Martin Jambor <mjambor@suse.cz>
2808
2809 PR ipa/60640
2810 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2811 Adjust all callers.
2812 * cgraph.c (clone_of_p): Also return true if thunks match.
2813 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2814 cgraph_function_or_thunk_node and an obsolete comment.
2815 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2816 file.
2817 (build_function_decl_skip_args): Likewise.
2818 (set_new_clone_decl_and_node_flags): New function.
2819 (duplicate_thunk_for_node): Likewise.
2820 (redirect_edge_duplicating_thunks): Likewise.
2821 (cgraph_clone_node): New parameter args_to_skip, pass it to
2822 redirect_edge_duplicating_thunks which is called instead of
2823 cgraph_redirect_edge_callee.
2824 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
2825 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2826
2827 2014-04-04 Jeff Law <law@redhat.com>
2828
2829 PR target/60657
2830 * config/arm/predicates.md (const_int_I_operand): New predicate.
2831 (const_int_M_operand): Similarly.
2832 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2833 const_int_operand.
2834 (insv_t2, extv_reg, extzv_t2): Likewise.
2835 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2836 (pop_multiple_with_writeback_and_return): Likewise.
2837 (vfp_pop_multiple_with_writeback): Likewise
2838
2839 2014-04-04 Richard Biener <rguenther@suse.de>
2840
2841 PR ipa/60746
2842 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2843 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2844 non-GIMPLE_LABELs.
2845 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2846 * gimplify.c (gimple_add_tmp_var_fn): New function.
2847 * gimple-expr.h (create_tmp_reg_fn): Declare.
2848 * gimple-expr.c (create_tmp_reg_fn): New function.
2849 * gimple-low.c (record_vars_into): Don't change cfun.
2850 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2851 code generation without cfun.
2852
2853 2014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2854
2855 PR bootstrap/60719
2856 * Makefile.in (install-driver): Fix shell scripting.
2857
2858 2014-04-03 Cong Hou <congh@google.com>
2859
2860 PR tree-optimization/60505
2861 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
2862 threshold of number of iterations below which no vectorization
2863 will be done.
2864 * tree-vect-loop.c (new_loop_vec_info):
2865 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2866 * tree-vect-loop.c (vect_analyze_loop_operations):
2867 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2868 * tree-vect-loop.c (vect_transform_loop):
2869 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2870 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2871 of iterations of the loop and see if we should build the epilogue.
2872
2873 2014-04-03 Richard Biener <rguenther@suse.de>
2874
2875 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
2876 (streamer_tree_cache_create): Adjust.
2877 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2878 to allow optional nodes array.
2879 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2880 (streamer_tree_cache_append): Likewise.
2881 (streamer_tree_cache_create): Create nodes array optionally
2882 as specified by parameter.
2883 * lto-streamer-out.c (create_output_block): Avoid maintaining
2884 the node array in the writer cache.
2885 (DFS_write_tree): Remove assertion.
2886 (produce_asm_for_decls): Free the out decl state hash table early.
2887 * lto-streamer-in.c (lto_data_in_create): Adjust for
2888 streamer_tree_cache_create prototype change.
2889
2890 2014-04-03 Richard Biener <rguenther@suse.de>
2891
2892 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2893 set TREE_CHAIN to NULL_TREE.
2894
2895 2014-04-03 Richard Biener <rguenther@suse.de>
2896
2897 PR tree-optimization/60740
2898 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2899 over all GIMPLE_COND operands.
2900
2901 2014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2902
2903 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2904 (Weffc++): Remove Scott's numbering, merge lists and reference
2905 Wnon-virtual-dtor.
2906
2907 2014-04-03 Nick Clifton <nickc@redhat.com>
2908
2909 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2910 properly.
2911
2912 2014-04-03 Martin Jambor <mjambor@suse.cz>
2913
2914 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2915 mention gcc_unreachable before failing.
2916 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2917 removed symbols.
2918
2919 2014-04-02 Jan Hubicka <hubicka@ucw.cz>
2920
2921 PR ipa/60659
2922 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2923 inconsistent code and instead mark the context inconsistent.
2924 (possible_polymorphic_call_targets): For inconsistent contexts
2925 return empty complete list.
2926
2927 2014-04-02 Anthony Green <green@moxielogic.com>
2928
2929 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2930 (extendqisi2, extendhisi2): Define.
2931 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2932 (WCHAR_TYPE): Change to unsigned int.
2933
2934 2014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2935
2936 PR tree-optimization/60733
2937 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2938 insertion point for PHI candidates to be the end of the feeding
2939 block for the PHI argument.
2940
2941 2014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2942
2943 PR rtl-optimization/60650
2944 * lra-constraints.c (process_alt_operands): Decrease reject for
2945 earlyclobber matching.
2946
2947 2014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2948
2949 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2950
2951 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2952
2953 * config/spu/spu.c (pad_bb): Do not crash when the last
2954 insn is CODE_FOR_blockage.
2955
2956 2014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2957
2958 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2959 lies outside the target mode.
2960
2961 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2962
2963 PR target/60735
2964 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2965 software floating point or no floating point registers, do not
2966 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2967 in GPRs that occurs after we tested for GPRs that would never be
2968 true.
2969
2970 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2971 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2972 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2973 specifically allow DDmode, since that does not use the SPE SIMD
2974 instructions.
2975
2976 2014-04-02 Richard Biener <rguenther@suse.de>
2977
2978 PR middle-end/60729
2979 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2980 MODE_INTs. Properly use negv_optab.
2981 (expand_abs): Likewise.
2982
2983 2014-04-02 Richard Biener <rguenther@suse.de>
2984
2985 PR bootstrap/60719
2986 * Makefile.in (install-driver): Guard extra installs with special
2987 names properly.
2988
2989 2014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2990
2991 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2992 Document vec_vgbbd.
2993
2994 2014-04-01 Richard Henderson <rth@redhat.com>
2995
2996 PR target/60704
2997 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2998 alternative enabled before register allocation.
2999
3000 2014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
3001
3002 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
3003 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
3004 typo.
3005 (nios2_large_got_address): Remove unneeded 'sym' parameter.
3006 (nios2_got_address): Update nios2_large_got_address call site.
3007 (nios2_delegitimize_address): New function.
3008 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
3009 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
3010 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
3011
3012 2014-04-01 Martin Husemann <martin@duskware.de>
3013
3014 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
3015 for -mabi=32.
3016
3017 2014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
3018
3019 PR rtl-optimization/60604
3020 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
3021 check from register_operand.
3022 (register_operand): Redefine in terms of general_operand.
3023 (nonmemory_operand): Use register_operand for the non-constant cases.
3024
3025 2014-04-01 Richard Biener <rguenther@suse.de>
3026
3027 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
3028
3029 2014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3030
3031 * doc/invoke.texi (mapp-regs): Clarify.
3032
3033 2014-03-31 Ulrich Drepper <drepper@gmail.com>
3034
3035 * config/i386/avx512fintrin.h (__v32hi): Define type.
3036 (__v64qi): Likewise.
3037 (_mm512_set1_epi8): Define.
3038 (_mm512_set1_epi16): Define.
3039 (_mm512_set4_epi32): Define.
3040 (_mm512_set4_epi64): Define.
3041 (_mm512_set4_pd): Define.
3042 (_mm512_set4_ps): Define.
3043 (_mm512_setr4_epi64): Define.
3044 (_mm512_setr4_epi32): Define.
3045 (_mm512_setr4_pd): Define.
3046 (_mm512_setr4_ps): Define.
3047 (_mm512_setzero_epi32): Define.
3048
3049 2014-03-31 Martin Jambor <mjambor@suse.cz>
3050
3051 PR middle-end/60647
3052 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3053 callsite_arguments_match_p. Updated all callers. Also check types of
3054 corresponding formal parameters and actual arguments.
3055 (not_all_callers_have_enough_arguments_p) Renamed to
3056 some_callers_have_mismatched_arguments_p.
3057
3058 2014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3059
3060 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3061
3062 2014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3063
3064 PR target/60034
3065 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3066 section anchor.
3067
3068 2014-03-30 Uros Bizjak <ubizjak@gmail.com>
3069
3070 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3071 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3072 Split out
3073 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3074 Use FMAMODE_NOVF512 mode iterator.
3075 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3076 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3077 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3078 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3079 Split out
3080 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3081 Use VF_128_256 mode iterator.
3082 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3083 Ditto.
3084
3085 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3086
3087 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3088 static chain if needed.
3089
3090 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3091
3092 PR target/60697
3093 * lra-constraints.c (index_part_to_reg): New.
3094 (process_address): Use it.
3095
3096 2014-03-27 Jeff Law <law@redhat.com>
3097 Jakub Jelinek <jakub@redhat.com>
3098
3099 PR target/60648
3100 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3101 gen_rtx_{PLUS,MULT} to build up the address expression.
3102
3103 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3104 creating non-canonical RTL.
3105
3106 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3107
3108 PR ipa/60243
3109 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3110 functions; reorganize to make cheap checks first.
3111 (inline_small_functions): Do not estimate growth when dumping;
3112 it is expensive.
3113 * ipa-inline.h (inline_summary): Add min_size.
3114 (growth_likely_positive): New function.
3115 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3116 (set_cond_stmt_execution_predicate): Cleanup.
3117 (estimate_edge_size_and_time): Compute min_size.
3118 (estimate_calls_size_and_time): Likewise.
3119 (estimate_node_size_and_time): Likewise.
3120 (inline_update_overall_summary): Update min_size.
3121 (do_estimate_edge_time): Likewise.
3122 (do_estimate_edge_size): Update.
3123 (do_estimate_edge_hints): Update.
3124 (growth_likely_positive): New function.
3125
3126 2014-03-28 Jakub Jelinek <jakub@redhat.com>
3127
3128 PR target/60693
3129 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3130 also if addr has VOIDmode.
3131
3132 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3133
3134 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3135 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3136 Declare extern.
3137 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3138 instructions as well as AdvancedSIMD loads.
3139
3140 2014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3141
3142 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3143 Use crypto_aese type.
3144 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3145 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3146 crypto_aese, crypto_aesmc. Move to types.md.
3147 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3148 crypto_aesmc.
3149 * config/arm/iterators.md (crypto_type): Likewise.
3150
3151 2014-03-28 Jan Hubicka <hubicka@ucw.cz>
3152
3153 * cgraph.c: Include expr.h and tree-dfa.h.
3154 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3155 remove LHS.
3156
3157 2014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3158
3159 PR target/60675
3160 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3161 regs from checking multi-reg pseudos.
3162
3163 2014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3164
3165 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3166
3167 2014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3168
3169 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3170 if it would clobber the stack pointer, even temporarily.
3171
3172 2014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3173
3174 * mode-switching.c: Make small adjustments to the top comment.
3175
3176 2014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3177
3178 * config/rs6000/constraints.md (wD constraint): New constraint to
3179 match the constant integer to get the top DImode/DFmode out of a
3180 vector in a VSX register.
3181
3182 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3183 match the constant integer to get the top DImode/DFmode out of a
3184 vector in a VSX register.
3185
3186 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3187 for ISA 2.07.
3188
3189 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3190 vbpermq builtins.
3191
3192 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3193 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3194
3195 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3196 Optimize vec_extract of 64-bit values, where the value being
3197 extracted is in the top word, where we can use scalar
3198 instructions. Add direct move and store support. Combine the big
3199 endian/little endian vector select load support into a single insn.
3200 (vsx_extract_<mode>_internal1): Likewise.
3201 (vsx_extract_<mode>_internal2): Likewise.
3202 (vsx_extract_<mode>_load): Likewise.
3203 (vsx_extract_<mode>_store): Likewise.
3204 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3205 combined into vsx_extract_<mode>_load.
3206 (vsx_extract_<mode>_one_le): Likewise.
3207
3208 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3209 define the top 64-bit vector element.
3210
3211 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3212 constraint.
3213
3214 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3215 Document vec_vbpermq builtin.
3216
3217 PR target/60672
3218 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3219 enable use of xxsldwi and xxpermdi builtin functions.
3220 (vec_xxpermdi): Likewise.
3221
3222 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3223 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3224
3225 2014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3226
3227 PR rtl-optimization/60650
3228 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
3229 first_p. Use it.
3230 (find_spills_for): New.
3231 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3232 Spill all pseudos on the second iteration.
3233
3234 2014-03-27 Marek Polacek <polacek@redhat.com>
3235
3236 PR c/50347
3237 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3238 types.
3239
3240 2014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3241
3242 * config/s390/s390.c (s390_can_use_return_insn): Check for
3243 call-saved FPRs on 31 bit.
3244
3245 2014-03-27 Jakub Jelinek <jakub@redhat.com>
3246
3247 PR middle-end/60682
3248 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3249 if they need regimplification, just drop them instead of
3250 calling gimple_regimplify_operands on them.
3251
3252 2014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3253
3254 PR target/60580
3255 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3256 (aarch64_frame_pointer_required): Adjust logic.
3257 (aarch64_can_eliminate): Adjust logic.
3258 (aarch64_override_options_after_change): Adjust logic.
3259
3260 2014-03-27 Dehao Chen <dehao@google.com>
3261
3262 * ipa-inline.c (early_inliner): Update node's inline info.
3263
3264 2014-03-26 Dehao Chen <dehao@google.com>
3265
3266 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3267 compiler inserted conditional jumps for NAN float check.
3268
3269 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3270
3271 * ubsan.h (ubsan_create_data): Change second argument's type
3272 to const location_t *.
3273 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3274 _("<unknown>").
3275 (ubsan_create_data): Change second argument to const location_t *PLOC.
3276 Create Loc field whenever PLOC is non-NULL.
3277 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3278 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3279 callers.
3280
3281 PR other/59545
3282 * real.c (real_to_integer2): Change type of low to UHWI.
3283
3284 2014-03-26 Tobias Burnus <burnus@net-b.de>
3285
3286 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3287 (CILK_SELF_SPECS): New define.
3288 (driver_self_specs): Use it.
3289
3290 2014-03-26 Richard Biener <rguenther@suse.de>
3291
3292 * tree-pretty-print.c (percent_K_format): Implement special
3293 case for LTO and its stripped down BLOCK tree.
3294
3295 2014-03-26 Jakub Jelinek <jakub@redhat.com>
3296
3297 PR sanitizer/60636
3298 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3299
3300 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3301 one range is range_int_cst_p, but not both, at least optimize
3302 addition/subtraction of 0 and multiplication by 0 or 1.
3303 * gimple-fold.c (gimple_fold_call): Fold
3304 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3305 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3306 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3307
3308 2014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3309
3310 PR rtl-optimization/60452
3311 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3312 <case REG>: Return 1 for invalid offsets from the frame pointer.
3313
3314 2014-03-26 Marek Polacek <polacek@redhat.com>
3315
3316 PR c/37428
3317 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3318 a structure/union.
3319
3320 2014-03-26 Marek Polacek <polacek@redhat.com>
3321
3322 PR c/39525
3323 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3324 field members.
3325
3326 2014-03-26 Marek Polacek <polacek@redhat.com>
3327
3328 PR other/59545
3329 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3330 multiplication in unsigned type.
3331
3332 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3333
3334 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3335
3336 2014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3337
3338 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3339
3340 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3341
3342 PR ipa/60315
3343 * cif-code.def (UNREACHABLE) New code.
3344 * ipa-inline.c (inline_small_functions): Skip edges to
3345 __builtlin_unreachable.
3346 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3347 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3348 predicate to __bulitin_unreachable.
3349 (set_cond_stmt_execution_predicate): Fix issue when
3350 invert_tree_comparison returns ERROR_MARK.
3351 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3352 propagate to inline clones.
3353 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3354 to unreachable.
3355 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3356 * cgraphclones.c (cgraph_clone_node): If call destination is already
3357 ureachable, do not redirect it back.
3358 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3359 unreachable.
3360
3361 2014-03-25 Jan Hubicka <hubicka@ucw.cz>
3362
3363 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3364 Do not modify inline clones.
3365
3366 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3367
3368 * config/i386/i386.md (general_sext_operand): New mode attr.
3369 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3370 don't generate (sign_extend (const_int)).
3371 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
3372 operands[2]. Use We constraint instead of <i> and
3373 <general_sext_operand> predicate instead of <general_operand>.
3374 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3375 * config/i386/constraints.md (We): New constraint.
3376 * config/i386/predicates.md (x86_64_sext_operand,
3377 sext_operand): New predicates.
3378
3379 2014-03-25 Martin Jambor <mjambor@suse.cz>
3380
3381 PR ipa/60600
3382 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3383 inconsistent devirtualizations to __builtin_unreachable.
3384
3385 2014-03-25 Marek Polacek <polacek@redhat.com>
3386
3387 PR c/35449
3388 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3389
3390 2014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3391
3392 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3393 order of elements for big-endian.
3394
3395 2014-03-25 Richard Biener <rguenther@suse.de>
3396
3397 PR middle-end/60635
3398 * gimplify-me.c (gimple_regimplify_operands): Update the
3399 re-gimplifed stmt.
3400
3401 2014-03-25 Martin Jambor <mjambor@suse.cz>
3402
3403 PR ipa/59176
3404 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3405 (lto_output_varpool_node): Likewise.
3406 (input_overwrite_node): Likewise.
3407 (input_varpool_node): Likewise.
3408
3409 2014-03-25 Richard Biener <rguenther@suse.de>
3410
3411 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
3412 (run_gcc): Likewise.
3413
3414 2014-03-25 Jakub Jelinek <jakub@redhat.com>
3415
3416 * combine.c (simplify_compare_const): Add MODE argument.
3417 Handle mode_width 0 as very large mode_width.
3418 (try_combine, simplify_comparison): Adjust callers.
3419
3420 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3421 type to avoid signed integer overflow.
3422 * explow.c (plus_constant): Likewise.
3423
3424 2014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3425
3426 * doc/generic.texi: Correct typos.
3427
3428 2014-03-24 Tobias Burnus <burnus@net-b.de>
3429
3430 * doc/invoke.texi (-flto): Expand section about
3431 using static libraries with LTO.
3432
3433 2014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3434
3435 PR rtl-optimization/60501
3436 * optabs.def (addptr3_optab): New optab.
3437 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3438 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3439 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3440
3441 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3442
3443 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3444
3445 2014-03-24 Ulrich Drepper <drepper@gmail.com>
3446
3447 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3448 _mm512_set1_pd.
3449
3450 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3451 (_mm256_undefined_ps): Define.
3452 (_mm256_undefined_pd): Define.
3453 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3454 (_mm_undefined_pd): Define.
3455 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3456 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3457 (_mm512_undefined_ps): Define.
3458 (_mm512_undefined_pd): Define.
3459 Use _mm*_undefined_*.
3460 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3461
3462 2014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3463
3464 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3465 (lshr_simd): DI mode added.
3466 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3467 (aarch64_ushr_simddi): Likewise.
3468 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3469 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3470 (vshrd_n_u64): Likewise.
3471
3472 2014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3473
3474 * Makefile.in (s-macro_list): Depend on cc1.
3475
3476 2014-03-23 Teresa Johnson <tejohnson@google.com>
3477
3478 * ipa-utils.c (ipa_print_order): Use specified dump file.
3479
3480 2014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3481
3482 PR rtl-optimization/60601
3483 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3484
3485 * gcc.c (eval_spec_function): Initialize save_growing_value.
3486
3487 2014-03-22 Jakub Jelinek <jakub@redhat.com>
3488
3489 PR sanitizer/60613
3490 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3491 code == MINUS_EXPR, never swap op0 with op1.
3492
3493 * toplev.c (init_local_tick): Avoid signed integer multiplication
3494 overflow.
3495 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3496 shift by first operand's bitsize.
3497
3498 2014-03-21 Jakub Jelinek <jakub@redhat.com>
3499
3500 PR target/60610
3501 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3502 redefine to 1 or 0.
3503 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3504 TARGET_ISA_64BIT_P(x).
3505
3506 2014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3507
3508 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3509 pattern for vector nor instead of subtract from splat(-1).
3510 (altivec_expand_vec_perm_const_le): Likewise.
3511
3512 2014-03-21 Richard Henderson <rth@twiddle.net>
3513
3514 PR target/60598
3515 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3516 related insns after epilogue_completed.
3517
3518 2014-03-21 Martin Jambor <mjambor@suse.cz>
3519
3520 PR ipa/59176
3521 * cgraph.h (symtab_node): New flag body_removed.
3522 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3523 when removing bodies.
3524 * symtab.c (dump_symtab_base): Dump body_removed flag.
3525 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3526 had their bodies removed.
3527
3528 2014-03-21 Martin Jambor <mjambor@suse.cz>
3529
3530 PR ipa/60419
3531 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3532 in the border.
3533
3534 2014-03-21 Richard Biener <rguenther@suse.de>
3535
3536 PR tree-optimization/60577
3537 * tree-core.h (struct tree_base): Document nothrow_flag use
3538 in VAR_DECL_NONALIASED.
3539 * tree.h (VAR_DECL_NONALIASED): New.
3540 (may_be_aliased): Adjust.
3541 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3542
3543 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3544
3545 * expr.c (expand_expr_real_1): Remove outdated comment.
3546
3547 2014-03-20 Jakub Jelinek <jakub@redhat.com>
3548
3549 PR middle-end/60597
3550 * ira.c (adjust_cleared_regs): Call copy_rtx on
3551 *reg_equiv[REGNO (loc)].src_p before passing it to
3552 simplify_replace_fn_rtx.
3553
3554 PR target/60568
3555 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3556 into CONST, put pic register as first operand of PLUS. Use
3557 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3558
3559 2014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3560
3561 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3562
3563 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3564
3565 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3566 around for store forwarding issue in the FPU on the UT699.
3567 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3568 loads and operations if -mfix-ut699 is specified.
3569 (divtf3_hq): Tweak attribute.
3570 (sqrttf2_hq): Likewise.
3571
3572 2014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3573
3574 * calls.c (store_one_arg): Remove incorrect const qualification on the
3575 type of the temporary.
3576 * cfgexpand.c (expand_return): Likewise.
3577 * expr.c (expand_constructor): Likewise.
3578 (expand_expr_real_1): Likewise.
3579
3580 2014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3581
3582 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3583 of parts.
3584
3585 2014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3586
3587 PR target/60039
3588 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3589
3590 2014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3591
3592 * config/arm/aarch-common-protos.h
3593 (alu_cost_table): Fix spelling of "extend".
3594 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3595
3596 2014-03-19 Richard Biener <rguenther@suse.de>
3597
3598 PR middle-end/60553
3599 * tree-core.h (tree_type_common): Re-order pointer members
3600 to reduce recursion depth during GC walks.
3601
3602 2014-03-19 Marek Polacek <polacek@redhat.com>
3603
3604 PR sanitizer/60569
3605 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3606 before accessing it.
3607
3608 2014-03-19 Richard Biener <rguenther@suse.de>
3609
3610 PR lto/59543
3611 * lto-streamer-in.c (input_function): In WPA stage do not drop
3612 debug stmts.
3613
3614 2014-03-19 Jakub Jelinek <jakub@redhat.com>
3615
3616 PR tree-optimization/60559
3617 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3618 with build_zero_cst assignment.
3619
3620 2014-03-18 Kai Tietz <ktietz@redhat.com>
3621
3622 PR rtl-optimization/56356
3623 * sdbout.c (sdbout_parms): Verify that parms'
3624 incoming argument is valid.
3625 (sdbout_reg_parms): Likewise.
3626
3627 2014-03-18 Richard Henderson <rth@redhat.com>
3628
3629 PR target/60562
3630 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3631 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
3632 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
3633
3634 2014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3635
3636 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
3637 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
3638 Italicize plugin event names in description. Explain that
3639 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3640 Remind that no GCC functions should be called after PLUGIN_FINISH.
3641 Explain what pragmas with expansion are.
3642
3643 2014-03-18 Martin Liska <mliska@suse.cz>
3644
3645 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3646 gimple call statement is update.
3647 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3648 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3649
3650 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3651
3652 PR sanitizer/60557
3653 * ubsan.c (ubsan_instrument_unreachable): Call
3654 initialize_sanitizer_builtins.
3655 (ubsan_pass): Likewise.
3656
3657 PR sanitizer/60535
3658 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3659 varpool_finalize_decl instead of rest_of_decl_compilation.
3660
3661 2014-03-18 Richard Biener <rguenther@suse.de>
3662
3663 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3664 by using bitmap_and_compl instead of bitmap_and_compl_into.
3665 (df_rd_transfer_function): Likewise.
3666
3667 2014-03-18 Richard Biener <rguenther@suse.de>
3668
3669 * doc/lto.texi (fresolution): Fix typo.
3670
3671 2014-03-18 Richard Biener <rguenther@suse.de>
3672
3673 * doc/invoke.texi (flto): Update for changes in 4.9.
3674
3675 2014-03-18 Richard Biener <rguenther@suse.de>
3676
3677 * doc/loop.texi: Remove section on the removed lambda framework.
3678 Update loop docs with recent changes in preserving loop structure.
3679
3680 2014-03-18 Richard Biener <rguenther@suse.de>
3681
3682 * doc/lto.texi (-fresolution): Document.
3683
3684 2014-03-18 Richard Biener <rguenther@suse.de>
3685
3686 * doc/contrib.texi: Adjust my name.
3687
3688 2014-03-18 Jakub Jelinek <jakub@redhat.com>
3689
3690 PR ipa/58721
3691 * internal-fn.c: Include diagnostic-core.h.
3692 (expand_BUILTIN_EXPECT): New function.
3693 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3694 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3695 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3696 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3697 IFN_BUILTIN_EXPECT.
3698 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3699 Revert 3 argument __builtin_expect code.
3700 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3701 * gimple-fold.c (gimple_fold_call): Likewise.
3702 * tree.h (fold_builtin_expect): New prototype.
3703 * builtins.c (build_builtin_expect_predicate): Add predictor
3704 argument, if non-NULL, create 3 argument __builtin_expect.
3705 (fold_builtin_expect): No longer static. Add ARG2 argument,
3706 pass it through to build_builtin_expect_predicate.
3707 (fold_builtin_2): Adjust caller.
3708 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3709 * internal-fn.def (BUILTIN_EXPECT): New.
3710
3711 2014-03-18 Tobias Burnus <burnus@net-b.de>
3712
3713 PR ipa/58721
3714 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3715 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3716 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3717
3718 2014-03-18 Jan Hubicka <hubicka@ucw.cz>
3719
3720 PR ipa/58721
3721 * predict.c (combine_predictions_for_bb): Fix up formatting.
3722 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3723 fill what it points to if non-NULL.
3724 (tree_predict_by_opcode): Adjust caller, use the predictor.
3725 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3726
3727 2014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3728
3729 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3730 proper constant for the store mode.
3731
3732 2014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3733
3734 * symtab.c (change_decl_assembler_name): Fix transparent alias
3735 chain construction.
3736
3737 2014-03-16 Renlin Li <Renlin.Li@arm.com>
3738
3739 * config/aarch64/aarch64.c: Correct the comments about the
3740 aarch64 stack layout.
3741
3742 2014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3743
3744 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3745 check for GF_OMP_FOR_KIND_FOR.
3746
3747 2013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3748
3749 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3750 ymm and zmm register names.
3751
3752 2014-03-17 Jakub Jelinek <jakub@redhat.com>
3753
3754 PR target/60516
3755 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3756 note creation for the 2010-08-31 changes.
3757
3758 2014-03-17 Marek Polacek <polacek@redhat.com>
3759
3760 PR middle-end/60534
3761 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3762 as -fno-tree-loop-vectorize.
3763 (expand_omp_simd): Likewise.
3764
3765 2014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3766
3767 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3768 (eligible_for_call_delay): New prototype.
3769 * config/sparc/sparc.c (tls_call_delay): Rename into...
3770 (eligible_for_call_delay): ...this. Return false if the instruction
3771 cannot be put in the delay slot of a branch.
3772 (eligible_for_restore_insn): Simplify.
3773 (eligible_for_return_delay): Return false if the instruction cannot be
3774 put in the delay slot of a branch and simplify.
3775 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3776 put in the delay slot of a branch.
3777 * config/sparc/sparc.md (fix_ut699): New attribute.
3778 (tls_call_delay): Delete.
3779 (in_call_delay): Reimplement.
3780 (eligible_for_sibcall_delay): Rename into...
3781 (in_sibcall_delay): ...this.
3782 (eligible_for_return_delay): Rename into...
3783 (in_return_delay): ...this.
3784 (in_branch_delay): Reimplement.
3785 (in_uncond_branch_delay): Delete.
3786 (in_annul_branch_delay): Delete.
3787
3788 2014-03-14 Richard Henderson <rth@redhat.com>
3789
3790 PR target/60525
3791 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3792 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3793 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3794 (floathi splitters): Remove.
3795 (float<SWI48x>xf2): New pattern.
3796 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3797 code that tried to handle DImode for 32-bit, but which was excluded
3798 by the pattern's condition. Drop allocation of stack temporary.
3799 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3800 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3801 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3802 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3803 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3804 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3805 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3806 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3807 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3808 (*float<SWI48x><X87MODEF>2_i387): Remove.
3809 (all float _with_temp splitters): Remove.
3810 (*float<SWI48x><MODEF>2_i387): New pattern.
3811 (*float<SWI48><MODEF>2_sse): New pattern.
3812 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3813 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3814
3815 2014-03-14 Jakub Jelinek <jakub@redhat.com>
3816 Marek Polacek <polacek@redhat.com>
3817
3818 PR middle-end/60484
3819 * common.opt (dump_base_name_prefixed): New Variable.
3820 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3821 if x_dump_base_name_prefixed is already set, set it at the end.
3822
3823 2014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3824
3825 PR rtl-optimization/60508
3826 * lra-constraints.c (get_reload_reg): Add new parameter
3827 in_subreg_p.
3828 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3829 Pass the new parameter values.
3830
3831 2014-03-14 Richard Biener <rguenther@suse.de>
3832
3833 * common.opt: Revert unintented changes from r205065.
3834 * opts.c: Likewise.
3835
3836 2014-03-14 Richard Biener <rguenther@suse.de>
3837
3838 PR middle-end/60518
3839 * cfghooks.c (split_block): Properly adjust all loops the
3840 block was a latch of.
3841
3842 2014-03-14 Martin Jambor <mjambor@suse.cz>
3843
3844 PR lto/60461
3845 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3846 and simplify it.
3847
3848 2014-03-14 Georg-Johann Lay <avr@gjlay.de>
3849
3850 PR target/59396
3851 * config/avr/avr.c (avr_set_current_function): Pass function name
3852 through default_strip_name_encoding before sanity checking instead
3853 of skipping the first char of the assembler name.
3854
3855 2014-03-13 Richard Henderson <rth@redhat.com>
3856
3857 PR debug/60438
3858 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3859 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3860 * config/i386/i386-protos.h: Likewise.
3861 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3862 in the expander instead of a splitter.
3863 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3864 any possibility of requiring a memory.
3865 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3866 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3867 (fp branch splitters): Update for ix86_split_fp_branch.
3868 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3869 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3870 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3871 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3872 (*fop_<MODEF>_3_i387): Likewise.
3873 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3874 (splitters for the fop_* register patterns): Remove.
3875 (fscalexf4_i387): Rename from *fscalexf4_i387.
3876 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
3877
3878 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3879
3880 PR tree-optimization/59779
3881 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3882 type for bitsize and maxsize instead of HOST_WIDE_INT.
3883
3884 2014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3885
3886 PR rtl-optimization/57320
3887 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3888 the CFG after thread_prologue_and_epilogue_insns.
3889
3890 2014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3891
3892 PR rtl-optimization/57189
3893 * lra-constraints.c (process_alt_operands): Disfavor spilling
3894 vector pseudos.
3895
3896 2014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3897
3898 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
3899
3900 2014-03-13 Jakub Jelinek <jakub@redhat.com>
3901
3902 PR tree-optimization/59025
3903 PR middle-end/60418
3904 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3905 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3906
3907 2014-03-13 Georg-Johann Lay <avr@gjlay.de>
3908
3909 PR target/60486
3910 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3911 calls of avr_out_plus_1.
3912
3913 2014-03-13 Bin Cheng <bin.cheng@arm.com>
3914
3915 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3916 BB's single pred and update the father loop's latch info later.
3917
3918 2014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3919
3920 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3921 (VEC_M): Likewise.
3922 (VEC_N): Likewise.
3923 (VEC_R): Likewise.
3924 (VEC_base): Likewise.
3925 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3926 registers, we need to swap double words in little endian mode.
3927
3928 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3929 to be a container mode for 128-bit integer operations added in ISA
3930 2.07. Unlike TImode and PTImode, the preferred register set is
3931 the Altivec/VMX registers for the 128-bit operations.
3932
3933 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3934 declarations.
3935 (rs6000_split_128bit_ok_p): Likewise.
3936
3937 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3938 macros for creating ISA 2.07 normal and overloaded builtin
3939 functions with 3 arguments.
3940 (BU_P8V_OVERLOAD_3): Likewise.
3941 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3942 for use as overloaded functions.
3943 (VPERM_1TI_UNS): Likewise.
3944 (VSEL_1TI): Likewise.
3945 (VSEL_1TI_UNS): Likewise.
3946 (ST_INTERNAL_1ti): Likewise.
3947 (LD_INTERNAL_1ti): Likewise.
3948 (XXSEL_1TI): Likewise.
3949 (XXSEL_1TI_UNS): Likewise.
3950 (VPERM_1TI): Likewise.
3951 (VPERM_1TI_UNS): Likewise.
3952 (XXPERMDI_1TI): Likewise.
3953 (SET_1TI): Likewise.
3954 (LXVD2X_V1TI): Likewise.
3955 (STXVD2X_V1TI): Likewise.
3956 (VEC_INIT_V1TI): Likewise.
3957 (VEC_SET_V1TI): Likewise.
3958 (VEC_EXT_V1TI): Likewise.
3959 (EQV_V1TI): Likewise.
3960 (NAND_V1TI): Likewise.
3961 (ORC_V1TI): Likewise.
3962 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3963 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3964 overloaded builtin.
3965 (VADDUQM): Likewise.
3966 (VSUBCUQ): Likewise.
3967 (VADDEUQM): Likewise.
3968 (VADDECUQ): Likewise.
3969 (VSUBEUQM): Likewise.
3970 (VSUBECUQ): Likewise.
3971
3972 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3973 __int128_t and __uint128_t types.
3974 (__uint128_type): Likewise.
3975 (altivec_categorize_keyword): Add support for vector __int128_t,
3976 vector __uint128_t, vector __int128, and vector unsigned __int128
3977 as a container type for TImode operations that need to be done in
3978 VSX/Altivec registers.
3979 (rs6000_macro_to_expand): Likewise.
3980 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3981 to support 128-bit integer instructions vaddcuq, vadduqm,
3982 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3983 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3984
3985 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
3986 for V1TImode, and set up preferences to use VSX/Altivec registers.
3987 Setup VSX reload handlers.
3988 (rs6000_debug_reg_global): Likewise.
3989 (rs6000_init_hard_regno_mode_ok): Likewise.
3990 (rs6000_preferred_simd_mode): Likewise.
3991 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
3992 (easy_altivec_constant): Likewise.
3993 (output_vec_const_move): Likewise.
3994 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3995 simple move.
3996 (rs6000_expand_vector_extract): Likewise.
3997 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3998 addressing.
3999 (rs6000_const_vec): Add support for V1TImode.
4000 (rs6000_emit_le_vsx_load): Swap double words when loading or
4001 storing TImode/V1TImode.
4002 (rs6000_emit_le_vsx_store): Likewise.
4003 (rs6000_emit_le_vsx_move): Likewise.
4004 (rs6000_emit_move): Add support for V1TImode.
4005 (altivec_expand_ld_builtin): Likewise.
4006 (altivec_expand_st_builtin): Likewise.
4007 (altivec_expand_vec_init_builtin): Likewise.
4008 (altivec_expand_builtin): Likewise.
4009 (rs6000_init_builtins): Add support for V1TImode type. Add
4010 support for ISA 2.07 128-bit integer builtins. Define type names
4011 for the VSX/Altivec vector types.
4012 (altivec_init_builtins): Add support for overloaded vector
4013 functions with V1TImode type.
4014 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
4015 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
4016 external function.
4017 (rs6000_split_128bit_ok_p): Likewise.
4018 (rs6000_handle_altivec_attribute): Create V1TImode from vector
4019 __int128_t and vector __uint128_t.
4020
4021 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
4022 and mode attributes.
4023 (VSX_M): Likewise.
4024 (VSX_M2): Likewise.
4025 (VSm): Likewise.
4026 (VSs): Likewise.
4027 (VSr): Likewise.
4028 (VSv): Likewise.
4029 (VS_scalar): Likewise.
4030 (VS_double): Likewise.
4031 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
4032
4033 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4034 we support the ISA 2.07 128-bit integer arithmetic instructions.
4035 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4036 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4037 and TImode types for use with the builtin functions.
4038 (V1TI_type_node): Likewise.
4039 (unsigned_V1TI_type_node): Likewise.
4040 (intTI_type_internal_node): Likewise.
4041 (uintTI_type_internal_node): Likewise.
4042
4043 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4044 128-bit builtin functions.
4045 (UNSPEC_VADDEUQM): Likewise.
4046 (UNSPEC_VADDECUQ): Likewise.
4047 (UNSPEC_VSUBCUQ): Likewise.
4048 (UNSPEC_VSUBEUQM): Likewise.
4049 (UNSPEC_VSUBECUQ): Likewise.
4050 (VM): Add V1TImode to vector mode iterators.
4051 (VM2): Likewise.
4052 (VI_unit): Likewise.
4053 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4054 (altivec_vaddcuq): Likewise.
4055 (altivec_vsubuqm): Likewise.
4056 (altivec_vsubcuq): Likewise.
4057 (altivec_vaddeuqm): Likewise.
4058 (altivec_vaddecuq): Likewise.
4059 (altivec_vsubeuqm): Likewise.
4060 (altivec_vsubecuq): Likewise.
4061
4062 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4063 mode iterators.
4064 (BOOL_128): Likewise.
4065 (BOOL_REGS_OUTPUT): Likewise.
4066 (BOOL_REGS_OP1): Likewise.
4067 (BOOL_REGS_OP2): Likewise.
4068 (BOOL_REGS_UNARY): Likewise.
4069 (BOOL_REGS_AND_CR0): Likewise.
4070
4071 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4072 128-bit integer builtin support.
4073 (vec_vadduqm): Likewise.
4074 (vec_vaddecuq): Likewise.
4075 (vec_vaddeuqm): Likewise.
4076 (vec_vsubecuq): Likewise.
4077 (vec_vsubeuqm): Likewise.
4078 (vec_vsubcuq): Likewise.
4079 (vec_vsubuqm): Likewise.
4080
4081 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4082 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4083 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4084 128-bit integer add/subtract to ISA 2.07.
4085
4086 2014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4087
4088 * config/arc/arc.c (arc_predicate_delay_insns):
4089 Fix third argument passed to conditionalize_nonjump.
4090
4091 2014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4092
4093 * config/aarch64/aarch64-builtins.c
4094 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4095 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4096 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4097 instead of __builtin_lfloor.
4098 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4099
4100 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4101
4102 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4103 (tree_ssa_ifcombine_bb_1): New function.
4104 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4105 is an empty forwarder block to then_bb or vice versa and then_bb
4106 and else_bb are effectively swapped.
4107
4108 2014-03-12 Christian Bruel <christian.bruel@st.com>
4109
4110 PR target/60264
4111 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4112 REG_CFA_DEF_CFA note.
4113 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4114 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4115
4116 2014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4117
4118 PR tree-optimization/60454
4119 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4120
4121 2014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4122
4123 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4124 Do not define target_cpu_default2 to generic.
4125 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4126 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4127 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4128
4129 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4130 Marc Glisse <marc.glisse@inria.fr>
4131
4132 PR tree-optimization/60502
4133 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4134 instead of build_low_bits_mask.
4135
4136 2014-03-12 Jakub Jelinek <jakub@redhat.com>
4137
4138 PR middle-end/60482
4139 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4140 if there are multiple uses, but op doesn't live on E edge.
4141 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4142 clobber stmts before __builtin_unreachable.
4143
4144 2014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4145
4146 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4147 hard_frame_pointer_rtx.
4148 * cse.c (cse_insn): Remove volatile check.
4149 * cselib.c (cselib_process_insn): Likewise.
4150 * dse.c (scan_insn): Likewise.
4151
4152 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4153
4154 * config/arc/arc.c (conditionalize_nonjump): New function,
4155 broken out of ...
4156 (arc_ifcvt): ... this.
4157 (arc_predicate_delay_insns): Use it.
4158
4159 2014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4160
4161 * config/arc/predicates.md (extend_operand): During/after reload,
4162 allow const_int_operand.
4163 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4164 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4165 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4166 to "i".
4167 (umulsi3_highpart_i): Likewise.
4168
4169 2014-03-11 Richard Biener <rguenther@suse.de>
4170
4171 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4172 Add asserts to guard possible wrong-code bugs.
4173
4174 2014-03-11 Richard Biener <rguenther@suse.de>
4175
4176 PR tree-optimization/60429
4177 PR tree-optimization/60485
4178 * tree-ssa-structalias.c (set_union_with_increment): Properly
4179 take into account all fields that overlap the shifted vars.
4180 (do_sd_constraint): Likewise.
4181 (do_ds_constraint): Likewise.
4182 (get_constraint_for_ptr_offset): Likewise.
4183
4184 2014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4185
4186 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4187 (nios2_compute_frame_layout):
4188 Add calculation of cfun->machine->fp_save_offset.
4189 (nios2_expand_prologue): Correct setting of frame pointer register
4190 in prologue.
4191 (nios2_expand_epilogue): Update recovery of stack pointer from
4192 frame pointer accordingly.
4193 (nios2_initial_elimination_offset): Update calculation of offset
4194 for eliminating to HARD_FRAME_POINTER_REGNUM.
4195
4196 2014-03-10 Jakub Jelinek <jakub@redhat.com>
4197
4198 PR ipa/60457
4199 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4200 cgraph_get_create_node on VAR_DECLs.
4201
4202 2014-03-10 Richard Biener <rguenther@suse.de>
4203
4204 PR middle-end/60474
4205 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4206
4207 2014-03-08 Douglas B Rupp <rupp@gnat.com>
4208
4209 * config/vms/vms.opt (vms_float_format): New variable.
4210
4211 2014-03-08 Tobias Burnus <burnus@net-b.de>
4212
4213 * doc/invoke.texi (-fcilkplus): Update implementation status.
4214
4215 2014-03-08 Paulo Matos <paulo@matos-sorge.com>
4216 Richard Biener <rguenther@suse.de>
4217
4218 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4219 consistently accross all TUs.
4220 (run_gcc): Enable -fshort-double automatically at link at link-time
4221 and disallow override.
4222
4223 2014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4224
4225 PR target/58271
4226 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4227 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4228 if they can't be used.
4229
4230 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4231
4232 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4233 for Solaris 11/x86 ld.
4234 * configure: Regenerate.
4235
4236 2014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4237
4238 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4239 (LIB_TLS_SPEC): Save as ld_tls_libs.
4240 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4241 (HAVE_AS_IX86_TLSLDM): New test.
4242 * configure, config.in: Regenerate.
4243 * config/i386/i386.c (legitimize_tls_address): Fall back to
4244 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4245 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4246 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4247 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4248
4249 2014-03-07 Paulo Matos <paulo@matos-sorge.com>
4250
4251 * common.opt (fira-loop-pressure): Mark as optimization.
4252
4253 2014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4254
4255 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4256 an OpenMP mappable type.
4257
4258 2014-03-06 Matthias Klose <doko@ubuntu.com>
4259
4260 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4261 MULTILIB_OSDIRNAMES is not defined.
4262
4263 2014-03-06 Jakub Jelinek <jakub@redhat.com>
4264 Meador Inge <meadori@codesourcery.com>
4265
4266 PR target/58595
4267 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4268 (arm_legitimize_address): Call legitimize_tls_address for any
4269 arm_tls_referenced_p expression, handle constant addend. Call it
4270 before testing for !TARGET_ARM.
4271 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4272
4273 2014-03-06 Richard Biener <rguenther@suse.de>
4274
4275 PR middle-end/60445
4276 PR lto/60424
4277 PR lto/60427
4278 Revert
4279 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4280
4281 * tree-streamer.c (record_common_node): Assert we don't record
4282 nodes with type double.
4283 (preload_common_node): Skip type double, complex double and double
4284 pointer since it is now frontend dependent due to fshort-double option.
4285
4286 2014-03-06 Richard Biener <rguenther@suse.de>
4287
4288 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4289 or -fno-lto is specified and the linker has full plugin support.
4290 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4291 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
4292 * lto-wrapper.c (merge_and_complain): Merge compile-time
4293 optimization levels.
4294 (run_gcc): And pass it through to the link options.
4295
4296 2014-03-06 Alexandre Oliva <aoliva@redhat.com>
4297
4298 PR debug/60381
4299 Revert:
4300 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4301 PR debug/59992
4302 * cselib.c (remove_useless_values): Skip to avoid quadratic
4303 behavior if the condition moved from...
4304 (cselib_process_insn): ... here holds.
4305
4306 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4307
4308 PR plugins/59335
4309 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4310 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4311
4312 PR plugins/59335
4313 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4314 (TM_H): Add x86-tune.def.
4315
4316 2014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4317
4318 * config/aarch64/aarch64.c (generic_tunings):
4319 Use cortexa57_extra_costs.
4320
4321 2014-03-05 Jakub Jelinek <jakub@redhat.com>
4322
4323 PR lto/60404
4324 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4325 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4326 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4327 cost for in_lto_p.
4328
4329 2014-03-04 Heiher <r@hev.cc>
4330
4331 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4332 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4333
4334 2014-03-04 Uros Bizjak <ubizjak@gmail.com>
4335
4336 * config/i386/predicates.md (const2356_operand): Change to ...
4337 (const2367_operand): ... this.
4338 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4339 const2367_operand.
4340 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4341 (*avx512pf_scatterpf<mode>sf): Ditto.
4342 (avx512pf_scatterpf<mode>df): Ditto.
4343 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4344 (*avx512pf_scatterpf<mode>df): Ditto.
4345 * config/i386/i386.c (ix86_expand_builtin): Update
4346 incorrect hint operand error message.
4347
4348 2014-03-04 Richard Biener <rguenther@suse.de>
4349
4350 * lto-section-in.c (lto_get_section_data): Fix const cast.
4351
4352 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4353
4354 * tree-streamer.c (record_common_node): Assert we don't record
4355 nodes with type double.
4356 (preload_common_node): Skip type double, complex double and double
4357 pointer since it is now frontend dependent due to fshort-double option.
4358
4359 2014-03-04 Richard Biener <rguenther@suse.de>
4360
4361 PR lto/60405
4362 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
4363 (lto_input_toplevel_asms): Likewise.
4364 * lto-section-in.c (lto_get_section_data): Instead do it here
4365 for every section.
4366
4367 2014-03-04 Richard Biener <rguenther@suse.de>
4368
4369 PR tree-optimization/60382
4370 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4371 dead PHIs a reduction.
4372
4373 2014-03-03 Uros Bizjak <ubizjak@gmail.com>
4374
4375 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4376 hint value.
4377 (_mm_prefetch): Move out of GCC target("sse") pragma.
4378 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4379 GCC target("prfchw") pragma.
4380 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4381 for locality <= 2.
4382 * config/i386/i386.c (ix86_option_override_internal): Enable
4383 -mprfchw with -mprefetchwt1.
4384
4385 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4386
4387 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4388 Mark as varying.
4389
4390 2014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4391
4392 * opts.h (CL_PCH_IGNORE): Define.
4393 * targhooks.c (option_affects_pch_p):
4394 Return false for options that have CL_PCH_IGNORE set.
4395 * opt-functions.awk: Process PchIgnore.
4396 * doc/options.texi: Document PchIgnore.
4397
4398 * config/arc/arc.opt (misize): Add PchIgnore property.
4399
4400 2014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4401
4402 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4403 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4404 constraint on constants to permit them being loaded into
4405 GENERAL_REGS or BASE_REGS.
4406
4407 2014-03-03 Nick Clifton <nickc@redhat.com>
4408
4409 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4410 anti-cacnonical alternatives.
4411 (negandhi3_real): New pattern.
4412 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4413
4414 2014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4415
4416 * config/avr/avr-mcus.def: Remove atxmega16x1.
4417 * config/avr/avr-tables.opt: Regenerate.
4418 * config/avr/t-multilib: Regenerate.
4419 * doc/avr-mmcu.texi: Regenerate.
4420
4421 2014-03-03 Tobias Grosser <tobias@grosser.es>
4422 Mircea Namolaru <mircea.namolaru@inria.fr>
4423
4424 PR tree-optimization/58028
4425 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4426 scalar dimensions.
4427
4428 2014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4429
4430 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4431 not handled by recognizers.
4432
4433 2014-03-03 Jakub Jelinek <jakub@redhat.com>
4434
4435 PR middle-end/60175
4436 * function.c (expand_function_end): Don't emit
4437 clobber_return_register sequence if clobber_after is a BARRIER.
4438 * cfgexpand.c (construct_exit_block): Append instructions before
4439 return_label to prev_bb.
4440
4441 2014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4442
4443 * config/rs6000/constraints.md: Document reserved use of "wc".
4444
4445 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4446
4447 PR ipa/60150
4448 * ipa.c (function_and_variable_visibility): When dissolving comdat
4449 group, also set all symbols to local.
4450
4451 2014-03-02 Jan Hubicka <hubicka@ucw.cz>
4452
4453 PR ipa/60306
4454
4455 Revert:
4456 2013-12-14 Jan Hubicka <jh@suse.cz>
4457 PR middle-end/58477
4458 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
4459
4460 2014-03-02 Jon Beniston <jon@beniston.com>
4461
4462 PR bootstrap/48230
4463 PR bootstrap/50927
4464 PR bootstrap/52466
4465 PR target/46898
4466 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
4467 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
4468 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
4469 (simple_return, *simple_return): New patterns
4470 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4471 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4472
4473 2014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4474
4475 * dwarf2out.c (gen_subprogram_die): Tidy.
4476
4477 2014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4478
4479 PR target/60071
4480 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4481 (*mov_t_msb_neg_negc): ... this new insn.
4482
4483 2014-02-28 Jason Merrill <jason@redhat.com>
4484
4485 PR c++/58678
4486 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4487 function.
4488
4489 2014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4490
4491 PR c++/60314
4492 * dwarf2out.c (decltype_auto_die): New static.
4493 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4494 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4495 (is_cxx_auto): Likewise.
4496
4497 2014-02-28 Ian Bolton <ian.bolton@arm.com>
4498
4499 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4500 we are not using general regs only.
4501
4502 2014-02-28 Richard Biener <rguenther@suse.de>
4503
4504 PR target/60280
4505 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4506 previous fix and only allow to remove trivial pre-headers
4507 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
4508 (remove_forwarder_block): Properly update the latch of a loop.
4509
4510 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4511
4512 PR debug/59992
4513 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4514 (cselib_preserved_hash_table): New.
4515 (preserve_constants_and_equivs): Move preserved vals to it.
4516 (cselib_find_slot): Look it up first.
4517 (cselib_init): Initialize it.
4518 (cselib_finish): Release it.
4519 (dump_cselib_table): Dump it.
4520
4521 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4522
4523 PR debug/59992
4524 * cselib.c (remove_useless_values): Skip to avoid quadratic
4525 behavior if the condition moved from...
4526 (cselib_process_insn): ... here holds.
4527
4528 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
4529
4530 PR debug/57232
4531 * var-tracking.c (vt_initialize): Apply the same condition to
4532 preserve the CFA base value.
4533
4534 2014-02-28 Joey Ye <joey.ye@arm.com>
4535
4536 PR target/PR60169
4537 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4538 if reload in progress or completed.
4539
4540 2014-02-28 Tobias Burnus <burnus@net-b.de>
4541
4542 PR middle-end/60147
4543 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4544 NAMELIST_DECL.
4545
4546 2014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4547
4548 * doc/tm.texi.in (Condition Code Status): Update documention for
4549 relative locations of cc0-setter and cc0-user.
4550
4551 2014-02-27 Jeff Law <law@redhat.com>
4552
4553 PR rtl-optimization/52714
4554 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4555 into two independent simple sets, if I3 is a jump, ensure the
4556 pattern we place into I3 is a (set (pc) ...).
4557
4558 2014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4559 Jeff Law <law@redhat.com>
4560
4561 PR rtl-optimization/49847
4562 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4563 are in different blocks.
4564 * doc/tm.texi (Condition Code Status): Update documention for
4565 relative locations of cc0-setter and cc0-user.
4566
4567 2014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4568
4569 PR target/59222
4570 * lra.c (lra_emit_add): Check SUBREG too.
4571
4572 2014-02-27 Andreas Schwab <schwab@suse.de>
4573
4574 * config/m68k/m68k.c (m68k_option_override): Disable
4575 -flive-range-shrinkage for classic m68k.
4576 (m68k_override_options_after_change): Likewise.
4577
4578 2014-02-27 Marek Polacek <polacek@redhat.com>
4579
4580 PR middle-end/59223
4581 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4582 -Wmaybe-uninitialized.
4583
4584 2014-02-27 Alan Modra <amodra@gmail.com>
4585
4586 PR target/57936
4587 * reload1.c (emit_input_reload_insns): When reload_override_in,
4588 set old to rl->in_reg when rl->in_reg is a subreg.
4589
4590 2014-02-26 Richard Biener <rguenther@suse.de>
4591
4592 PR bootstrap/60343
4593 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4594
4595 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4596
4597 * common/config/i386/predicates.md (const1256_operand): Remove.
4598 (const2356_operand): New.
4599 (const_1_to_2_operand): Remove.
4600 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4601 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4602 (*avx512pf_gatherpf<mode>sf): Ditto.
4603 (avx512pf_gatherpf<mode>df): Ditto.
4604 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4605 (*avx512pf_gatherpf<mode>df): Ditto.
4606 (avx512pf_scatterpf<mode>sf): Ditto.
4607 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4608 (*avx512pf_scatterpf<mode>sf): Ditto.
4609 (avx512pf_scatterpf<mode>df): Ditto.
4610 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4611 (*avx512pf_scatterpf<mode>df): Ditto.
4612 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4613
4614 2014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4615
4616 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4617 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4618 (_mm512_mask_testn_epi64_mask): Move to ...
4619 * config/i386/avx512cdintrin.h: Here.
4620 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4621 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4622 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4623 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4624 TARGET_AVX512F from TARGET_AVX512CD.
4625
4626 2014-02-26 Richard Biener <rguenther@suse.de>
4627
4628 PR ipa/60327
4629 * ipa.c (walk_polymorphic_call_targets): Properly guard
4630 call to inline_update_overall_summary.
4631
4632 2014-02-26 Bin Cheng <bin.cheng@arm.com>
4633
4634 PR target/60280
4635 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4636 and latches only if requested. Fix latch if it is removed.
4637 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4638 LOOPS_HAVE_PREHEADERS.
4639
4640 2014-02-25 Andrew Pinski <apinski@cavium.com>
4641
4642 * builtins.c (expand_builtin_thread_pointer): Create a new target
4643 when the target is NULL.
4644
4645 2014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4646
4647 PR rtl-optimization/60317
4648 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4649 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4650 * lra-assigns.c: Include params.h.
4651 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4652 other reload pseudos considerations.
4653
4654 2014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4655
4656 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4657 to use canonical form for nor<mode>3.
4658
4659 2014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4660
4661 PR target/55426
4662 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4663 conversions.
4664
4665 2014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4666
4667 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4668 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4669 (ix86_handle_option): Handle OPT_mprefetchwt1.
4670 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4671 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4672 PREFETCHWT1 CPUID.
4673 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4674 OPTION_MASK_ISA_PREFETCHWT1.
4675 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4676 (PTA_PREFETCHWT1): New.
4677 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4678 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
4679 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
4680 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4681 (*prefetch_avx512pf_<mode>_: Change into ...
4682 (*prefetch_prefetchwt1_<mode>: This.
4683 * config/i386/i386.opt (mprefetchwt1): New.
4684 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4685 (_mm_prefetch): Handle intent to write.
4686 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4687
4688 2014-02-25 Richard Biener <rguenther@suse.de>
4689
4690 PR middle-end/60291
4691 * emit-rtl.c (mem_attrs_htab): Remove.
4692 (mem_attrs_htab_hash): Likewise.
4693 (mem_attrs_htab_eq): Likewise.
4694 (set_mem_attrs): Always allocate new mem-attrs when something changed.
4695 (init_emit_once): Do not allocate mem_attrs_htab.
4696
4697 2014-02-25 Richard Biener <rguenther@suse.de>
4698
4699 PR lto/60319
4700 * lto-opts.c (lto_write_options): Output non-explicit conservative
4701 -fwrapv, -fno-trapv and -fno-strict-overflow.
4702 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4703 (run_gcc): And pass them through.
4704
4705 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4706
4707 * sel-sched.c (calculate_new_fences): New parameter ptime.
4708 Calculate it as a maximum over all fence cycles.
4709 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4710 Print the final schedule timing when sched_verbose.
4711
4712 2014-02-25 Andrey Belevantsev <abel@ispras.ru>
4713
4714 PR rtl-optimization/60292
4715 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4716 bit fot the fence instruction.
4717
4718 2014-02-24 Alangi Derick <alangiderick@gmail.com>
4719
4720 * calls.h: Fix typo in comment.
4721
4722 2014-02-24 John David Anglin <danglin@gcc.gnu.org>
4723
4724 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4725 adjusting offsetable addresses.
4726
4727 2014-02-24 Guozhi Wei <carrot@google.com>
4728
4729 * sparseset.h (sparseset_pop): Fix the wrong index.
4730
4731 2014-02-24 Walter Lee <walt@tilera.com>
4732
4733 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4734 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4735 triplet.
4736 * common/config/tilegx/tilegx-common.c
4737 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4738 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4739 (LINK_SPEC): Ditto.
4740 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4741 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4742 (tilegx_gimplify_va_arg_expr): Handle big endian.
4743 (tilegx_expand_unaligned_load): Ditto.
4744 (tilegx_expand_unaligned_store): Ditto.
4745 (TARGET_RETURN_IN_MSB): New.
4746 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4747 (TARGET_ENDIAN_DEFAULT): New.
4748 (TARGET_BIG_ENDIAN): Handle big endian.
4749 (BYTES_BIG_ENDIAN): Ditto.
4750 (WORDS_BIG_ENDIAN): Ditto.
4751 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4752 (ENDIAN_SPEC): New.
4753 (EXTRA_SPECS): New.
4754 * config/tilegx/tilegx.md (extv): Handle big endian.
4755 (extzv): Ditto.
4756 (insn_st<n>): Ditto.
4757 (insn_st<n>_add<bitsuffix>): Ditto.
4758 (insn_stnt<n>): Ditto.
4759 (insn_stnt<n>_add<bitsuffix>):Ditto.
4760 (vec_interleave_highv8qi): Handle big endian.
4761 (vec_interleave_highv8qi_be): New.
4762 (vec_interleave_highv8qi_le): New.
4763 (insn_v1int_h): Handle big endian.
4764 (vec_interleave_lowv8qi): Handle big endian.
4765 (vec_interleave_lowv8qi_be): New.
4766 (vec_interleave_lowv8qi_le): New.
4767 (insn_v1int_l): Handle big endian.
4768 (vec_interleave_highv4hi): Handle big endian.
4769 (vec_interleave_highv4hi_be): New.
4770 (vec_interleave_highv4hi_le): New.
4771 (insn_v2int_h): Handle big endian.
4772 (vec_interleave_lowv4hi): Handle big endian.
4773 (vec_interleave_lowv4hi_be): New.
4774 (vec_interleave_lowv4hi_le): New.
4775 (insn_v2int_l): Handle big endian.
4776 (vec_interleave_highv2si): Handle big endian.
4777 (vec_interleave_highv2si_be): New.
4778 (vec_interleave_highv2si_le): New.
4779 (insn_v4int_h): Handle big endian.
4780 (vec_interleave_lowv2si): Handle big endian.
4781 (vec_interleave_lowv2si_be): New.
4782 (vec_interleave_lowv2si_le): New.
4783 (insn_v4int_l): Handle big endian.
4784 * config/tilegx/tilegx.opt (mbig-endian): New option.
4785 (mlittle-endian): New option.
4786 * doc/install.texi: Document tilegxbe-linux.
4787 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4788
4789 2014-02-24 Martin Jambor <mjambor@suse.cz>
4790
4791 PR ipa/60266
4792 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4793 there are no parameter descriptors.
4794
4795 2014-02-24 Andrey Belevantsev <abel@ispras.ru>
4796
4797 PR rtl-optimization/60268
4798 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4799 initialization to ...
4800 (sched_rgn_init): ... here.
4801 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4802
4803 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4804
4805 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4806 names.
4807
4808 2014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
4809
4810 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4811 definition.
4812
4813 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4814
4815 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4816 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
4817
4818 2014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
4819
4820 * config/microblaze/predicates.md: Add cmp_op predicate.
4821 * config/microblaze/microblaze.md: Add branch_compare instruction
4822 which uses cmp_op predicate and emits cmp insn before branch.
4823 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
4824 to microblaze_expand_conditional_branch and consolidate logic.
4825 (microblaze_expand_conditional_branch): emit branch_compare
4826 insn instead of handling cmp op separate from branch insn.
4827
4828 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4829
4830 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4831 to permit subregs.
4832
4833 2014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4834
4835 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4836 define_insn with define_expand and new define_insn
4837 *altivec_lve<VI_char>x_internal.
4838 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4839 and new define_insn *altivec_stve<VI_char>x_internal.
4840 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4841 prototype.
4842 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4843 lve*x built-ins.
4844 (altivec_expand_stvex_be): New function.
4845
4846 2014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4847
4848 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
4849 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
4850 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4851 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4852
4853 2014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4854
4855 PR target/60298
4856 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4857 instead of emit_move_insn.
4858
4859 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4860
4861 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4862 vspltw with vsldoi.
4863 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4864 gen_altivec_vsumsws.
4865
4866 2014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4867
4868 * config/rs6000/altivec.md (altivec_lvxl): Rename as
4869 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
4870 (altivec_lvxl_<mode>): New define_expand incorporating
4871 -maltivec=be semantics where needed.
4872 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4873 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4874 semantics where needed.
4875 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4876 (altivec_stvx_<mode>): New define_expand incorporating
4877 -maltivec=be semantics where needed.
4878 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4879 VM2 iterator instead of V4SI.
4880 (altivec_stvxl_<mode>): New define_expand incorporating
4881 -maltivec=be semantics where needed.
4882 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4883 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
4884 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4885 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4886 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
4887 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4888 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4889 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4890 ALTIVEC_BUILTIN_STVXL.
4891 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
4892 (altivec_expand_stvx_be): Likewise.
4893 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4894 (altivec_expand_lvx_be): Likewise.
4895 (altivec_expand_stvx_be): Likewise.
4896 (altivec_expand_builtin): Add cases for
4897 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4898 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4899 (altivec_init_builtins): Add definitions for
4900 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
4901 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
4902
4903 2014-02-21 Catherine Moore <clm@codesourcery.com>
4904
4905 * doc/invoke.texi (mvirt, mno-virt): Document.
4906 * config/mips/mips.opt (mvirt): New option.
4907 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4908
4909 2014-02-21 Richard Biener <rguenther@suse.de>
4910
4911 PR tree-optimization/60276
4912 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4913 (STMT_VINFO_MIN_NEG_DIST): New macro.
4914 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4915 STMT_VINFO_MIN_NEG_DIST.
4916 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4917 made for negative dependence distances still hold.
4918
4919 2014-02-21 Richard Biener <rguenther@suse.de>
4920
4921 PR middle-end/60291
4922 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4923 DECL_INITIAL for globals not in the current function context.
4924
4925 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4926
4927 PR tree-optimization/56490
4928 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4929 * tree-ssa-uninit.c: Include params.h.
4930 (compute_control_dep_chain): Add num_calls argument, return false
4931 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4932 num_calls to recursive call.
4933 (find_predicates): Change dep_chain into normal array,
4934 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4935 variable and adjust compute_control_dep_chain caller.
4936 (find_def_preds): Likewise.
4937
4938 2014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4939
4940 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4941 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4942
4943 2014-02-21 Nick Clifton <nickc@redhat.com>
4944
4945 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4946 (pushhi1): Likewise.
4947 (popqi1): Add mode to pre_dec.
4948 (pophi1): Likewise.
4949
4950 2014-02-21 Jakub Jelinek <jakub@redhat.com>
4951
4952 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4953 mode for mask of V8SFmode permutation.
4954
4955 2014-02-20 Richard Henderson <rth@redhat.com>
4956
4957 PR c++/60272
4958 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4959 a new pseudo for OLDVAL.
4960
4961 2014-02-20 Jakub Jelinek <jakub@redhat.com>
4962
4963 PR target/57896
4964 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4965 gen_reg_rtx if d->testing_p.
4966 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4967 if d->testing_p and we will certainly return true.
4968 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4969 if d->testing_p.
4970
4971 2014-02-20 Uros Bizjak <ubizjak@gmail.com>
4972
4973 * emit-rtl.c (gen_reg_rtx): Assert that
4974 crtl->emit.regno_pointer_align_length is non-zero.
4975
4976 2014-02-20 Richard Henderson <rth@redhat.com>
4977
4978 PR c++/60272
4979 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4980 on failure the store back into EXPECT.
4981
4982 2014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4983 Sandra Loosemore <sandra@codesourcery.com>
4984
4985 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
4986 * config/nios2/nios2.c (nios2_function_profiler): Add
4987 -fPIC (flag_pic == 2) support.
4988 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4989 (nios2_large_offset_p): New function.
4990 (nios2_unspec_reloc_p): Move up position, update to use
4991 nios2_large_offset_p.
4992 (nios2_unspec_address): Remove function.
4993 (nios2_unspec_offset): New function.
4994 (nios2_large_got_address): New function.
4995 (nios2_got_address): Add large offset support.
4996 (nios2_legitimize_tls_address): Update usage of removed and new
4997 functions.
4998 (nios2_symbol_binds_local_p): New function.
4999 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
5000 (nios2_legitimize_address): Update to use nios2_large_offset_p.
5001 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
5002 (nios2_print_operand): Merge H/L processing, add hiadj/lo
5003 processing for (const (unspec ...)).
5004 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
5005
5006 2014-02-20 Richard Biener <rguenther@suse.de>
5007
5008 * tree-cfg.c (replace_uses_by): Mark altered BBs before
5009 doing the substitution.
5010 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
5011
5012 2014-02-20 Martin Jambor <mjambor@suse.cz>
5013
5014 PR ipa/55260
5015 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
5016 info when checking whether lattices are bottom.
5017
5018 2014-02-20 Richard Biener <rguenther@suse.de>
5019
5020 PR middle-end/60221
5021 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
5022 regions at -O0.
5023
5024 2014-02-20 Jan Hubicka <hubicka@ucw.cz>
5025
5026 PR ipa/58555
5027 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5028 parameter specifying the scaling.
5029 (inline_call): Update.
5030 (want_inline_recursively): Guard division by zero.
5031 (recursive_inlining): Update.
5032 * ipa-inline.h (clone_inlined_nodes): Update.
5033
5034 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5035
5036 PR target/60204
5037 * config/i386/i386.c (classify_argument): Pass structures of size
5038 64 bytes or less in register.
5039
5040 2014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5041 Kirill Yukhin <kirill.yukhin@intel.com>
5042
5043 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5044 (_mm_rcp28_round_ss): Ditto.
5045 (_mm_rsqrt28_round_sd): Ditto.
5046 (_mm_rsqrt28_round_ss): Ditto.
5047 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5048 (_mm_rcp14_round_ss): Ditto.
5049 (_mm_rsqrt14_round_sd): Ditto.
5050 (_mm_rsqrt14_round_ss): Ditto.
5051 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5052 the first input operand, get rid of match_dup.
5053 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5054 attribute to sse.
5055 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5056 Ditto.
5057 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5058 operand as the first input operand, set type attribute.
5059 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5060 Set type attribute.
5061 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5062 operand as the first input operand, set type attribute.
5063
5064 2014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5065
5066 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5067 bit of zero.
5068
5069 2014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5070
5071 PR target/60207
5072 * config/i386/i386.c (construct_container): Remove TFmode check
5073 for X86_64_INTEGER_CLASS.
5074
5075 2014-02-19 Uros Bizjak <ubizjak@gmail.com>
5076
5077 PR target/59794
5078 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5079 only when -Wpsabi is enabled.
5080
5081 2014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5082
5083 PR target/59799
5084 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5085 passing arrays in registers are the same as for structs, so remove the
5086 special case for them.
5087
5088 2014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5089
5090 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5091 destination type, extract only the valid bits if the source type is not
5092 integral and has a different mode.
5093
5094 2014-02-19 Richard Biener <rguenther@suse.de>
5095
5096 PR ipa/60243
5097 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5098 for all calls.
5099
5100 2014-02-19 Richard Biener <rguenther@suse.de>
5101
5102 PR ipa/60243
5103 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5104 (ipa_modify_call_arguments): Emit an argument load explicitely and
5105 preserve virtual SSA form there and for the replacement call.
5106 Do not update SSA form nor free dominance info.
5107
5108 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5109
5110 * ipa.c (function_and_variable_visibility): Also clear WEAK
5111 flag when disolving COMDAT_GROUP.
5112
5113 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5114
5115 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5116 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5117 not devirtualizing.
5118 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5119 do more sanity checks.
5120 (detect_type_change): Return true when giving up early.
5121 (compute_complex_assign_jump_func): Fix type parameter of
5122 ipa_set_ancestor_jf.
5123 (compute_complex_ancestor_jump_func): Likewise.
5124 (update_jump_functions_after_inlining): Fix updating of
5125 ancestor function.
5126 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
5127
5128 2014-02-18 Jan Hubicka <hubicka@ucw.cz>
5129
5130 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5131 inline clones when edge disappears.
5132
5133 2014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5134
5135 PR target/60203
5136 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5137 Split 64-bit moves into 2 patterns. Do not allow the use of
5138 direct move for TDmode in little endian, since the decimal value
5139 has little endian bytes within a word, but the 64-bit pieces are
5140 ordered in a big endian fashion, and normal subreg's of TDmode are
5141 not allowed.
5142 (mov<mode>_64bit_dm): Likewise.
5143 (movtd_64bit_nodm): Likewise.
5144
5145 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5146
5147 PR tree-optimization/60174
5148 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5149 statement of an SSA_NAME that occurs in an abnormal PHI node.
5150
5151 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5152
5153 PR sanitizer/60142
5154 * final.c (SEEN_BB): Remove.
5155 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5156 (final_scan_insn): Don't force_source_line on second
5157 NOTE_INSN_BASIC_BLOCK.
5158
5159 2014-02-18 Uros Bizjak <ubizjak@gmail.com>
5160
5161 PR target/60205
5162 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5163 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5164 (type_natural_mode): Warn ABI change when %zmm register is not
5165 available for AVX512F vector value passing.
5166
5167 2014-02-18 Kai Tietz <ktietz@redhat.com>
5168
5169 PR target/60193
5170 * config/i386/i386.c (ix86_expand_prologue): Use value in
5171 rax register as displacement when restoring %r10 or %rax.
5172 Fix wrong offset when restoring both registers.
5173
5174 2014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5175
5176 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5177 assertion with conditional return.
5178
5179 2014-02-18 Jakub Jelinek <jakub@redhat.com>
5180 Uros Bizjak <ubizjak@gmail.com>
5181
5182 PR driver/60233
5183 * config/i386/driver-i386.c (host_detect_local_cpu): If
5184 YMM state is not saved by the OS, also clear has_f16c. Move
5185 CPUID 0x80000001 handling before YMM state saving checking.
5186
5187 2014-02-18 Andrey Belevantsev <abel@ispras.ru>
5188
5189 PR rtl-optimization/58960
5190 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5191 factored out from ...
5192 (sched_init): ... here.
5193 (free_global_sched_pressure_data): New, factored out from ...
5194 (sched_finish): ... here.
5195 * sched-int.h (free_global_sched_pressure_data): Declare.
5196 * sched-rgn.c (nr_regions_initial): New static global.
5197 (haifa_find_rgns): Initialize it.
5198 (schedule_region): Disable sched-pressure for the newly
5199 generated regions.
5200
5201 2014-02-17 Richard Biener <rguenther@suse.de>
5202
5203 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5204 release SSA defs of pattern stmts.
5205
5206 2014-02-17 Richard Biener <rguenther@suse.de>
5207
5208 * tree-inline.c (expand_call_inline): Release the virtual
5209 operand defined by the call we are about to inline.
5210
5211 2014-02-17 Richard Biener <rguenther@suse.de>
5212
5213 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5214
5215 2014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5216 Ilya Tocar <ilya.tocar@intel.com>
5217
5218 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5219 arguments order in builtin.
5220 (_mm512_permutexvar_epi64): Ditto.
5221 (_mm512_mask_permutexvar_epi64): Ditto
5222 (_mm512_maskz_permutexvar_epi32): Ditto
5223 (_mm512_permutexvar_epi32): Ditto
5224 (_mm512_mask_permutexvar_epi32): Ditto
5225
5226 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5227
5228 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
5229 (p8_vmrgow): Likewise.
5230
5231 2014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5232
5233 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5234 endian targets.
5235
5236 2014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5237
5238 PR target/60203
5239 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5240 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5241 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5242 using direct move instructions on ISA 2.07. Also adjust
5243 instruction length for 64-bit.
5244 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5245 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5246
5247 2014-02-15 Alan Modra <amodra@gmail.com>
5248
5249 PR target/58675
5250 PR target/57935
5251 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5252 find_replacement on parts of insn rtl that might be reloaded.
5253
5254 2014-02-15 Richard Biener <rguenther@suse.de>
5255
5256 PR tree-optimization/60183
5257 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
5258 (tree_ssa_phiprop): Calculate and free post-dominators.
5259
5260 2014-02-14 Jeff Law <law@redhat.com>
5261
5262 PR rtl-optimization/60131
5263 * ree.c (get_extended_src_reg): New function.
5264 (combine_reaching_defs): Use it rather than assuming location of REG.
5265 (find_and_remove_re): Verify first operand of extension is
5266 a REG before adding the insns to the copy list.
5267
5268 2014-02-14 Roland McGrath <mcgrathr@google.com>
5269
5270 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5271 * configure: Regenerated.
5272 * config.in: Regenerated.
5273 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5274 instead of ASM_SHORT.
5275
5276 2014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5277 Richard Earnshaw <rearnsha@arm.com>
5278
5279 PR rtl-optimization/59535
5280 * lra-constraints.c (process_alt_operands): Encourage alternative
5281 when unassigned pseudo class is superset of the alternative class.
5282 (inherit_reload_reg): Don't inherit when optimizing for code size.
5283 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5284 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5285 modes not less than 4 for Thumb1.
5286
5287 2014-02-14 Kyle McMartin <kyle@redhat.com>
5288
5289 PR pch/60010
5290 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5291
5292 2014-02-14 Richard Biener <rguenther@suse.de>
5293
5294 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5295 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5296 Do not strip INDIRECT_REFs.
5297
5298 2014-02-14 Richard Biener <rguenther@suse.de>
5299
5300 PR lto/60179
5301 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5302 DECL_FUNCTION_SPECIFIC_TARGET.
5303 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5304 * tree-streamer-out.c (pack_ts_target_option): Remove.
5305 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
5306 (write_ts_function_decl_tree_pointers): Do not stream
5307 DECL_FUNCTION_SPECIFIC_TARGET.
5308 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5309 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5310 (lto_input_ts_function_decl_tree_pointers): Do not stream
5311 DECL_FUNCTION_SPECIFIC_TARGET.
5312
5313 2014-02-14 Jakub Jelinek <jakub@redhat.com>
5314
5315 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
5316 (get_initial_def_for_induction, vectorizable_induction): Ignore
5317 debug stmts when looking for exit_phi.
5318 (vectorizable_live_operation): Fix up condition.
5319
5320 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5321
5322 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5323 nreverse() because it changes the content of original tree list.
5324
5325 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5326
5327 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5328 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5329
5330 2014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5331
5332 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5333 GNU coding standards.
5334
5335 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5336
5337 PR debug/60152
5338 * dwarf2out.c (gen_subprogram_die): Don't call
5339 add_calling_convention_attribute if subr_die is old_die.
5340
5341 2014-02-13 Sharad Singhai <singhai@google.com>
5342
5343 * doc/optinfo.texi: Fix order of nodes.
5344
5345 2014-02-13 Uros Bizjak <ubizjak@gmail.com>
5346
5347 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5348 operands[2], not operands[3].
5349
5350 2014-02-13 Richard Biener <rguenther@suse.de>
5351
5352 PR bootstrap/59878
5353 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5354 mention the possibility of an in-tree build.
5355 (CLooG): Update recommended version to 0.18.1, mention the
5356 possibility of an in-tree build and clarify that the ISL
5357 bundled with CLooG does not work.
5358
5359 2014-02-13 Jakub Jelinek <jakub@redhat.com>
5360
5361 PR target/43546
5362 * expr.c (compress_float_constant): If x is a hard register,
5363 extend into a pseudo and then move to x.
5364
5365 2014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5366
5367 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5368 caused by bad second argument to warning_at() with -mhotpatch and
5369 nested functions (e.g. with gfortran).
5370
5371 2014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5372
5373 * opts.c (option_name): Remove "enabled by default" rider.
5374
5375 2014-02-12 John David Anglin <danglin@gcc.gnu.org>
5376
5377 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5378
5379 2014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5380 Uros Bizjak <ubizjak@gmail.com>
5381
5382 PR target/60151
5383 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
5384 * configure: Regenerated.
5385
5386 2014-02-12 Richard Biener <rguenther@suse.de>
5387
5388 * vec.c (vec_prefix::calculate_allocation): Move as
5389 inline variant to vec.h.
5390 (vec_prefix::calculate_allocation_1): New out-of-line version.
5391 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5392 (vec_prefix::m_has_auto_buf): Rename to ...
5393 (vec_prefix::m_using_auto_storage): ... this.
5394 (vec_prefix::calculate_allocation): Inline the easy cases
5395 and dispatch to calculate_allocation_1 which doesn't need the
5396 prefix address.
5397 (va_heap::reserve): Use gcc_checking_assert.
5398 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5399 m_using_auto_storage.
5400 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5401 member and adjust.
5402 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5403 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5404 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5405
5406 2014-02-12 Richard Biener <rguenther@suse.de>
5407
5408 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5409 when we found a dependence.
5410
5411 2014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5412
5413 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5414 common code...
5415 (maybe_fold_stmt): ... into this new function.
5416 * omp-low.c (lower_omp): Update comment.
5417
5418 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5419 last use.
5420
5421 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5422 dereference.
5423
5424 2014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5425
5426 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5427 identifiers in comments.
5428 (cortexa53_extra_costs): Likewise.
5429 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
5430 (cortexa7_extra_costs): Likewise.
5431 (cortexa12_extra_costs): Likewise.
5432 (cortexa15_extra_costs): Likewise.
5433 (v7m_extra_costs): Likewise.
5434
5435 2014-02-12 Richard Biener <rguenther@suse.de>
5436
5437 PR middle-end/60092
5438 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5439 of posix_memalign being successful.
5440 (lower_stmt): Restrict lowering of posix_memalign to when
5441 -ftree-bit-ccp is enabled.
5442
5443 2014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5444
5445 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5446 arg_loc.
5447 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5448
5449 2014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5450
5451 PR rtl-optimization/60116
5452 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5453 other_insn once the combination has been validated.
5454
5455 2014-02-11 Jan Hubicka <hubicka@ucw.cz>
5456
5457 PR lto/59468
5458 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5459 and wrapper.
5460 * ipa-devirt.c: Include demangle.h
5461 (odr_violation_reported): New static variable.
5462 (add_type_duplicate): Update odr_violations.
5463 (maybe_record_node): Add completep parameter; update it.
5464 (record_target_from_binfo): Add COMPLETEP parameter;
5465 update it as needed.
5466 (possible_polymorphic_call_targets_1): Likewise.
5467 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5468 rename FINAL to COMPLETE.
5469 (record_targets_from_bases): Sanity check we found the binfo;
5470 fix COMPLETEP updating.
5471 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5472 parameter, fix computing of COMPLETEP.
5473 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5474 at LTO time do demangling.
5475 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5476 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5477 parameter.
5478 (gimple_get_virt_method_for_binfo): Likewise.
5479 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5480 gimple_get_virt_method_for_vtable): Update prototypes.
5481
5482 2014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5483
5484 PR target/49008
5485 * genautomata.c (add_presence_absence): Fix typo with
5486 {final_}presence_list.
5487
5488 2014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5489
5490 PR target/60137
5491 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5492 for VSX/Altivec vectors that land in GPR registers.
5493
5494 2014-02-11 Richard Henderson <rth@redhat.com>
5495 Jakub Jelinek <jakub@redhat.com>
5496
5497 PR debug/59776
5498 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5499 around drhs if type conversion to lacc->type is not useless.
5500
5501 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5502
5503 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5504 tuning struct.
5505 (cortex-a57.cortex-a53): Likewise.
5506 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5507
5508 2014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5509
5510 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5511 arm_restrict_it.
5512
5513 2014-02-11 Renlin Li <Renlin.Li@arm.com>
5514
5515 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5516 add_options_for_arm_vfp3.
5517
5518 2014-02-11 Jeff Law <law@redhat.com>
5519
5520 PR middle-end/54041
5521 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5522 object with an undesirable mode.
5523
5524 2014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5525
5526 PR libgomp/60107
5527 * config/i386/sol2-9.h: New file.
5528 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5529 *-*-solaris2.9*): Use it.
5530
5531 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5532
5533 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5534 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5535
5536 2014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
5537
5538 * config/microblaze/microblaze.c: Extend mcpu version format
5539
5540 2014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
5541
5542 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5543
5544 2014-02-10 Richard Henderson <rth@redhat.com>
5545
5546 PR target/59927
5547 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5548 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5549 ms-abi vs -mno-accumulate-outgoing-args.
5550 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5551 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5552 respect to ms-abi.
5553
5554 2014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5555
5556 PR middle-end/60080
5557 * cfgexpand.c (expand_asm_operands): Attach source location to
5558 ASM_INPUT rtx objects.
5559 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5560
5561 2014-02-10 Nick Clifton <nickc@redhat.com>
5562
5563 * config/mn10300/mn10300.c (popcount): New function.
5564 (mn10300_expand_prologue): Include saved registers in stack usage
5565 count.
5566
5567 2014-02-10 Jeff Law <law@redhat.com>
5568
5569 PR middle-end/52306
5570 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
5571 when changing the SET_DEST of a prior insn to avoid an input reload.
5572
5573 2014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5574
5575 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5576 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5577 -mcall-openbsd, or -mcall-linux.
5578 (CC1_ENDIAN_BIG_SPEC): Remove.
5579 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5580 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5581 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5582 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5583 and %cc1_endian_default.
5584 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5585
5586 2014-02-10 Richard Biener <rguenther@suse.de>
5587
5588 PR tree-optimization/60115
5589 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5590 MEM_REF handling. Properly verify that the accesses are not
5591 out of the objects bound.
5592
5593 2014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5594
5595 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5596 coretex to cortex.
5597
5598 2014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5599
5600 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5601 proper constants and fix formatting.
5602 (possible_polymorphic_call_targets): Fix formatting.
5603
5604 2014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5605 Ilya Tocar <ilya.tocar@intel.com>
5606
5607 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5608 (_mm512_loadu_epi32): Renamed into...
5609 (_mm512_loadu_si512): This.
5610 (_mm512_storeu_epi32): Renamed into...
5611 (_mm512_storeu_si512): This.
5612 (_mm512_maskz_ceil_ps): Removed.
5613 (_mm512_maskz_ceil_pd): Ditto.
5614 (_mm512_maskz_floor_ps): Ditto.
5615 (_mm512_maskz_floor_pd): Ditto.
5616 (_mm512_floor_round_ps): Ditto.
5617 (_mm512_floor_round_pd): Ditto.
5618 (_mm512_ceil_round_ps): Ditto.
5619 (_mm512_ceil_round_pd): Ditto.
5620 (_mm512_mask_floor_round_ps): Ditto.
5621 (_mm512_mask_floor_round_pd): Ditto.
5622 (_mm512_mask_ceil_round_ps): Ditto.
5623 (_mm512_mask_ceil_round_pd): Ditto.
5624 (_mm512_maskz_floor_round_ps): Ditto.
5625 (_mm512_maskz_floor_round_pd): Ditto.
5626 (_mm512_maskz_ceil_round_ps): Ditto.
5627 (_mm512_maskz_ceil_round_pd): Ditto.
5628 (_mm512_expand_pd): Ditto.
5629 (_mm512_expand_ps): Ditto.
5630 * config/i386/i386.c (ix86_builtins): Remove
5631 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5632 (bdesc_args): Ditto.
5633 * config/i386/predicates.md (const1256_operand): New.
5634 (const_1_to_2_operand): Ditto.
5635 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5636 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5637 (*avx512pf_gatherpf<mode>sf): Ditto.
5638 (avx512pf_gatherpf<mode>df): Ditto.
5639 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5640 (*avx512pf_gatherpf<mode>df): Ditto.
5641 (avx512pf_scatterpf<mode>sf): Ditto.
5642 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5643 (*avx512pf_scatterpf<mode>sf): Ditto.
5644 (avx512pf_scatterpf<mode>df): Ditto.
5645 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5646 (*avx512pf_scatterpf<mode>df): Ditto.
5647 (avx512f_expand<mode>): Removed.
5648 (<shift_insn><mode>3<mask_name>): Change predicate type.
5649
5650 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5651
5652 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5653 not at the end of datarefs vector use ordered_remove to avoid
5654 reordering datarefs vector.
5655
5656 PR c/59984
5657 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5658 mark local addressable non-static vars as GOVD_PRIVATE
5659 instead of GOVD_LOCAL.
5660 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5661 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5662 than copying them.
5663
5664 PR middle-end/60092
5665 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5666 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5667 assume_aligned or alloc_align attributes.
5668 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5669 arguments. Handle also assume_aligned and alloc_align attributes.
5670 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5671 calls to functions with assume_aligned or alloc_align attributes.
5672 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
5673
5674 2014-02-08 Terry Guo <terry.guo@arm.com>
5675
5676 * doc/invoke.texi: Document ARM -march=armv7e-m.
5677
5678 2014-02-08 Jakub Jelinek <jakub@redhat.com>
5679
5680 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5681 flag on __cilkrts_rethrow builtin.
5682
5683 PR ipa/60026
5684 * ipa-cp.c (determine_versionability): Fail at -O0
5685 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5686 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5687
5688 Revert:
5689 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5690
5691 PR ipa/60026
5692 * tree-inline.c (copy_forbidden): Fail for
5693 __attribute__((optimize (0))) functions.
5694
5695 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5696
5697 * varpool.c: Include pointer-set.h.
5698 (varpool_remove_unreferenced_decls): Variables in other partitions
5699 will not be output; be however careful to not lose information
5700 about partitioning.
5701
5702 2014-02-07 Jan Hubicka <hubicka@ucw.cz>
5703
5704 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5705 lookup in the vtable constructor.
5706
5707 2014-02-07 Jeff Law <law@redhat.com>
5708
5709 PR target/40977
5710 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5711 define_insn_and_split.
5712
5713 * ipa-inline.c (inline_small_functions): Fix typos.
5714
5715 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5716
5717 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5718 (s390_can_use_return_insn): Declare.
5719 * config/s390/s390.h (EPILOGUE_USES): Define.
5720 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5721 instructions.
5722 (s390_chunkify_start): Handle return JUMP_LABELs.
5723 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5724 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5725 (s390_can_use_return_insn): New functions.
5726 (s390_fix_long_loop_prediction): Handle conditional returns.
5727 (TARGET_SET_UP_BY_PROLOGUE): Define.
5728 * config/s390/s390.md (ANY_RETURN): New code iterator.
5729 (*creturn, *csimple_return, return, simple_return): New patterns.
5730
5731 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5732
5733 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5734 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5735 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5736 REG_CFA_RESTORE list when deciding not to restore a register.
5737
5738 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5739
5740 * config/s390/s390.c: Include tree-pass.h and context.h.
5741 (s390_early_mach): New function, split out from...
5742 (s390_emit_prologue): ...here.
5743 (pass_data_s390_early_mach): New pass structure.
5744 (pass_s390_early_mach): New class.
5745 (s390_option_override): Create and register early_mach pass.
5746 Move to end of file.
5747
5748 2014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5749
5750 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5751 to match for the exit block.
5752
5753 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5754
5755 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5756 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5757 Reject misaligned operands.
5758
5759 2014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5760
5761 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
5762
5763 2014-02-07 Richard Biener <rguenther@suse.de>
5764
5765 PR middle-end/60092
5766 * gimple-low.c (lower_builtin_posix_memalign): New function.
5767 (lower_stmt): Call it to lower posix_memalign in a way
5768 to make alignment info accessible.
5769
5770 2014-02-07 Jakub Jelinek <jakub@redhat.com>
5771
5772 PR c++/60082
5773 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5774 __builtin_setjmp_receiver.
5775
5776 2014-02-07 Richard Biener <rguenther@suse.de>
5777
5778 PR middle-end/60092
5779 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5780 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5781 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5782 Handle BUILT_IN_POSIX_MEMALIGN.
5783 (find_func_clobbers): Likewise.
5784 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5785 (call_may_clobber_ref_p_1): Likewise.
5786
5787 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5788
5789 PR ipa/59918
5790 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5791 sanity check.
5792
5793 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5794
5795 PR ipa/59469
5796 * lto-cgraph.c (lto_output_node): Use
5797 symtab_get_symbol_partitioning_class.
5798 (lto_output_varpool_node): likewise.
5799 (symtab_get_symbol_partitioning_class): Move here from
5800 lto/lto-partition.c
5801 * cgraph.h (symbol_partitioning_class): Likewise.
5802 (symtab_get_symbol_partitioning_class): Declare.
5803
5804 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5805
5806 * ggc.h (ggc_internal_cleared_alloc): New macro.
5807 * vec.h (vec_safe_copy): Handle memory stats.
5808 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5809 * target-globals.c (save_target_globals): Likewise.
5810
5811 2014-02-06 Jan Hubicka <hubicka@ucw.cz>
5812
5813 PR target/60077
5814 * expr.c (emit_move_resolve_push): Export; be bit more selective
5815 on when to clear alias set.
5816 * expr.h (emit_move_resolve_push): Declare.
5817 * function.h (struct function): Add tail_call_marked.
5818 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5819 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5820 * config/i386/i386.md (TImode move expander): De not call
5821 ix86_expand_push.
5822 (FP push expanders): Preserve memory attributes.
5823 * config/i386/sse.md (push<mode>1): Remove.
5824 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
5825 (ix86_expand_push): Remove.
5826 * config/i386/mmx.md (push<mode>1): Remove.
5827
5828 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5829
5830 PR rtl-optimization/60030
5831 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5832 lopart with paradoxical subreg before shifting it up by hprec.
5833
5834 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5835
5836 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5837 Remove extra newline at end of file.
5838 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5839 (arm_issue_rate): Handle cortexa57.
5840 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5841 (cortex-a57.cortex-a53): Likewise.
5842
5843 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5844
5845 PR target/59575
5846 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5847 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5848 bitmask.
5849 (arm_expand_prologue): Adjust all callers.
5850 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5851 info, registers also at the lowest numbered registers side. Use
5852 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5853 XEXP.
5854
5855 PR debug/59992
5856 * var-tracking.c (adjust_mems): Before adding a SET to
5857 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
5858
5859 2014-02-06 Alan Modra <amodra@gmail.com>
5860
5861 PR target/60032
5862 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5863 change SDmode to DDmode when lra_in_progress.
5864
5865 2014-02-06 Jakub Jelinek <jakub@redhat.com>
5866
5867 PR middle-end/59150
5868 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5869 free_data_ref on the dr first, and before goto again also set dr
5870 to the next dr. For simd_lane_access, free old datarefs[i] before
5871 overwriting it. For get_vectype_for_scalar_type failure, don't
5872 free_data_ref if simd_lane_access.
5873
5874 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5875
5876 PR target/60062
5877 * tree.h (opts_for_fn): New inline function.
5878 (opt_for_fn): Define.
5879 * config/i386/i386.c (ix86_function_regparm): Use
5880 opt_for_fn (decl, optimize) instead of optimize.
5881
5882 2014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5883
5884 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5885 for SYMBOL_REF in large memory model.
5886
5887 2014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5888
5889 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5890 and crypto support.
5891 (cortex-a57): Likewise.
5892 (cortex-a57.cortex-a53): Likewise.
5893
5894 2014-02-06 Yury Gribov <y.gribov@samsung.com>
5895 Kugan Vivekanandarajah <kuganv@linaro.org>
5896
5897 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5898 unaligned_access.
5899 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5900
5901 2014-02-06 Richard Biener <rguenther@suse.de>
5902
5903 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5904 set_loop_copy and initialize_original_copy_tables.
5905
5906 2014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5907
5908 * config/aarch64/aarch64-simd.md
5909 (aarch64_ashr_simddi): Change QI to SI.
5910
5911 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5912 Jakub Jelinek <jakub@redhat.com>
5913
5914 PR middle-end/60013
5915 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5916 of the dataflow.
5917
5918 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5919
5920 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5921 CODE_FOR_altivec_vpku[hw]um to
5922 CODE_FOR_altivec_vpku[hw]um_direct.
5923 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5924 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5925 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5926 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5927
5928 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5929
5930 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5931 generation for -maltivec=be.
5932 (altivec_vsumsws): Simplify redundant test.
5933
5934 2014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5935
5936 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5937 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5938 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5939 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5940 gen_altivec_vpkuwum.
5941 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5942 BYTES_BIG_ENDIAN.
5943 (altivec_vpks<VI_char>ss): Likewise.
5944 (altivec_vpks<VI_char>us): Likewise.
5945 (altivec_vpku<VI_char>us): Likewise.
5946 (altivec_vpku<VI_char>um): Likewise.
5947 (altivec_vpku<VI_char>um_direct): New (copy of
5948 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5949 internal use).
5950 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5951 target is little endian and -maltivec=be is not specified.
5952 (*altivec_vupkhs<VU_char>_direct): New (copy of
5953 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
5954 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5955 target is little endian and -maltivec=be is not specified.
5956 (*altivec_vupkls<VU_char>_direct): New (copy of
5957 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
5958 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5959 little endian and -maltivec=be is not specified.
5960 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5961 little endian and -maltivec=be is not specified.
5962
5963 2014-02-05 Richard Henderson <rth@redhat.com>
5964
5965 PR debug/52727
5966 * combine-stack-adj.c: Revert r206943.
5967 * sched-int.h (struct deps_desc): Add last_args_size.
5968 * sched-deps.c (init_deps): Initialize it.
5969 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5970 contain REG_ARGS_SIZE notes.
5971
5972 2014-02-05 Jan Hubicka <hubicka@ucw.cz>
5973
5974 * lto-cgraph.c (asm_nodes_output): Make global.
5975 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
5976 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
5977 (driver_handle_option): Handle OPT_fwpa.
5978
5979 2014-02-05 Jakub Jelinek <jakub@redhat.com>
5980
5981 PR ipa/59947
5982 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5983 a comment typo and formatting issue. If odr_hash hasn't been
5984 created, return vNULL and set *completep to false.
5985
5986 PR middle-end/57499
5987 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5988 bb with no successors.
5989
5990 2014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5991
5992 PR target/59718
5993 * doc/invoke.texi (-march): Clarify documentation for ARM.
5994 (-mtune): Likewise.
5995 (-mcpu): Likewise.
5996
5997 2014-02-05 Richard Biener <rguenther@suse.de>
5998
5999 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
6000 when not vectorizing because of too many alias checks.
6001 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
6002 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
6003
6004 2014-02-05 Nick Clifton <nickc@redhat.com>
6005
6006 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
6007 accept extended registers in any mode when compiling for the MN10300.
6008
6009 2014-02-05 Yury Gribov <y.gribov@samsung.com>
6010
6011 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
6012 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
6013 sanitization attributes.
6014 (can_inline_edge_p): Likewise.
6015 (sanitize_attrs_match_for_inline_p): New function.
6016
6017 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6018
6019 * ipa-prop.c (detect_type_change): Shor circuit testing of
6020 type changes on THIS pointer.
6021
6022 2014-02-04 John David Anglin <danglin@gcc.gnu.org>
6023
6024 PR target/59777
6025 * config/pa/pa.c (legitimize_tls_address): Return original address
6026 if not passed a SYMBOL_REF rtx.
6027 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6028 addresses.
6029 (pa_emit_move_sequence): Simplify TLS source operands.
6030 (pa_legitimate_constant_p): Reject all TLS constants.
6031 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6032 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6033
6034 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6035
6036 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6037 groups when we know they are controlled by LTO.
6038 * varasm.c (default_binds_local_p_1): If object is in other partition,
6039 it will be resolved locally.
6040
6041 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6042
6043 * config/host-linux.c (linux_gt_pch_use_address): Don't
6044 use SSIZE_MAX because it is not always defined.
6045
6046 2014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6047
6048 PR bootstrap/59913
6049 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6050 threshold for pseudo splitting.
6051 (update_ebb_live_info): Process call argument hard registers and
6052 hard registers from insn definition too.
6053 (max_small_class_regs_num): New constant.
6054 (inherit_in_ebb): Update live hard regs through EBBs. Update
6055 reloads_num only for small register classes. Don't split for
6056 outputs of jumps.
6057
6058 2014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6059
6060 PR ipa/60058
6061 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6062 is non-null.
6063
6064 2014-02-04 Jan Hubicka <hubicka@ucw.cz>
6065
6066 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6067 visibility is safe.
6068
6069 2014-02-04 Marek Polacek <polacek@redhat.com>
6070
6071 * gdbinit.in (pel): Define.
6072
6073 2014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6074
6075 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6076 behavior.
6077
6078 2014-02-04 Richard Biener <rguenther@suse.de>
6079
6080 PR lto/59723
6081 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6082 in function context local.
6083 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6084 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6085 similar to LTO_imported_decl_ref.
6086
6087 2014-02-04 Jakub Jelinek <jakub@redhat.com>
6088
6089 PR tree-optimization/60002
6090 * cgraphclones.c (build_function_decl_skip_args): Clear
6091 DECL_LANG_SPECIFIC.
6092
6093 PR tree-optimization/60023
6094 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6095 false to gsi_replace.
6096 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6097 has been in some EH region and vec_stmt could throw, add
6098 vec_stmt into the same EH region.
6099 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6100 has no lhs, ignore it.
6101 * internal-fn.c (expand_MASK_LOAD): Likewise.
6102
6103 PR ipa/60026
6104 * tree-inline.c (copy_forbidden): Fail for
6105 __attribute__((optimize (0))) functions.
6106
6107 PR other/58712
6108 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6109 is set, copy one less argument.
6110 (expand_simd_clones): Don't subtract clone_info->inbranch
6111 from simd_clone_struct_alloc argument.
6112
6113 PR rtl-optimization/57915
6114 * recog.c (simplify_while_replacing): If all unary/binary/relational
6115 operation arguments are constant, attempt to simplify those.
6116
6117 PR middle-end/59261
6118 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6119 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6120
6121 2014-02-04 Richard Biener <rguenther@suse.de>
6122
6123 PR tree-optimization/60012
6124 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6125 TBAA disambiguation to all DDRs.
6126
6127 2014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6128
6129 PR target/59788
6130 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6131 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6132
6133 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6134
6135 PR ipa/59882
6136 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6137
6138 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6139
6140 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6141 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6142
6143 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6144
6145 PR ipa/59831
6146 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6147 to figure out targets of polymorphic calls with known decl.
6148 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6149 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6150 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6151 (get_polymorphic_call_info): ... here.
6152 (get_polymorphic_call_info_from_invariant): New function.
6153
6154 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6155
6156 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6157 lookup via vtable pointer; check for type consistency
6158 and turn inconsitent facts into UNREACHABLE.
6159 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6160 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6161 type inconsistent querries; return UNREACHABLE instead.
6162
6163 2014-02-03 Richard Henderson <rth@twiddle.net>
6164
6165 PR tree-opt/59924
6166 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6167 already processed this node.
6168 (normalize_one_pred_1): Pass along mark_set.
6169 (normalize_one_pred): Create and destroy a pointer_set_t.
6170 (normalize_one_pred_chain): Likewise.
6171
6172 2014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6173
6174 PR gcov-profile/58602
6175 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
6176
6177 2014-02-03 Jan Hubicka <hubicka@ucw.cz>
6178
6179 PR ipa/59831
6180 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6181 -fno-devirtualize; try to devirtualize by the knowledge of
6182 virtual table pointer given by aggregate propagation.
6183 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6184 (ipa_print_node_jump_functions): Dump also offset that
6185 is relevant for polymorphic calls.
6186 (determine_known_aggregate_parts): Add arg_type parameter; use it
6187 instead of determining the type from pointer type.
6188 (ipa_compute_jump_functions_for_edge): Update call of
6189 determine_known_aggregate_parts.
6190 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6191 (gimple_get_virt_method_for_binfo): ... here; simplify using
6192 vtable_pointer_value_to_vtable.
6193 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6194 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
6195 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
6196 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6197 POINTER_PLUS_EXPR.
6198 (vtable_pointer_value_to_binfo): ... here.
6199 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6200
6201 2014-02-03 Teresa Johnson <tejohnson@google.com>
6202
6203 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6204 redef of outer loop index variable.
6205
6206 2014-02-03 Marc Glisse <marc.glisse@inria.fr>
6207
6208 PR c++/53017
6209 PR c++/59211
6210 * doc/extend.texi (Function Attributes): Typo.
6211
6212 2014-02-03 Cong Hou <congh@google.com>
6213
6214 PR tree-optimization/60000
6215 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6216 if the vectorized statement is a store. A store statement can only
6217 appear at the end of pattern statements.
6218
6219 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6220
6221 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6222 (ix86_option_override_internal): Default long double to 64-bit for
6223 32-bit Bionic and to 128-bit for 64-bit Bionic.
6224
6225 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6226 TARGET_LONG_DOUBLE_128 is true.
6227 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6228
6229 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6230 (mlong-double-64): Negate -mlong-double-128.
6231 (mlong-double-128): New option.
6232
6233 * config/i386/i386-c.c (ix86_target_macros): Define
6234 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6235
6236 * doc/invoke.texi: Document -mlong-double-128.
6237
6238 2014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6239
6240 PR rtl-optimization/60024
6241 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6242
6243 2014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6244
6245 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
6246
6247 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6248
6249 PR rtl-optimization/57662
6250 * sel-sched.c (code_motion_path_driver): Do not mark already not
6251 existing blocks in the visiting bitmap.
6252
6253 2014-02-03 Andrey Belevantsev <abel@ispras.ru>
6254
6255 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6256 on the insn being emitted.
6257
6258 2014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6259 Will Deacon <will.deacon@arm.com>
6260
6261 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6262
6263 2014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6264
6265 * config/arm/arm-tables.opt: Regenerate.
6266
6267 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6268
6269 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6270 for vector types other than V16QImode.
6271 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6272 define_expand, and call altivec_expand_vec_perm_le when producing
6273 code with little endian element order.
6274 (*altivec_vperm_<mode>_internal): New insn having previous
6275 behavior of altivec_vperm_<mode>.
6276 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6277 altivec_expand_vec_perm_le when producing code with little endian
6278 element order.
6279 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6280 behavior of altivec_vperm_<mode>_uns.
6281
6282 2014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6283
6284 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6285 (altivec_vsumsws): Add handling for -maltivec=be with a little
6286 endian target.
6287 (altivec_vsumsws_direct): New.
6288 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6289 gen_altivec_vsumsws.
6290
6291 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6292
6293 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6294 vtable_pointer_value_to_binfo): New functions.
6295 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6296 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6297
6298 2014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6299
6300 * config/nios2/nios2.md (load_got_register): Initialize GOT
6301 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6302 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6303
6304 2014-02-02 Jan Hubicka <hubicka@ucw.cz>
6305
6306 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6307 preserverd by passthrough, do not propagate the type.
6308
6309 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6310
6311 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6312 (mips_atomic_assign_expand_fenv): New function.
6313 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6314
6315 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6316
6317 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6318 (__builtin_mips_set_fcsr): Likewise.
6319 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6320 MIPS_USI_FTYPE_VOID.
6321 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6322 (mips16_expand_set_fcsr): Likewise.
6323 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6324 (mips16_set_fcsr_stub): Likewise.
6325 (mips16_get_fcsr_one_only_stub): New class.
6326 (mips16_set_fcsr_one_only_stub): Likewise.
6327 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6328 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6329 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6330 (hard_float): New availability predicate.
6331 (mips_builtins): Add get_fcsr and set_fcsr.
6332 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6333 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6334 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6335 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6336 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6337 patterns.
6338
6339 2014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6340
6341 * config/mips/mips.c (mips_one_only_stub): New class.
6342 (mips_need_mips16_rdhwr_p): Replace with...
6343 (mips16_rdhwr_stub): ...this new variable.
6344 (mips16_stub_call_address): New function.
6345 (mips16_rdhwr_one_only_stub): New class.
6346 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6347 (mips_output_mips16_rdhwr): Delete.
6348 (mips_finish_stub): New function.
6349 (mips_code_end): Use it to handle rdhwr stubs.
6350
6351 2014-02-02 Uros Bizjak <ubizjak@gmail.com>
6352
6353 PR target/60017
6354 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6355 when calculating size of integer atomic types.
6356
6357 2014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6358
6359 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6360
6361 2014-02-01 Jakub Jelinek <jakub@redhat.com>
6362
6363 PR tree-optimization/60003
6364 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6365 * profile.c (branch_prob): Use gimple_call_builtin_p
6366 to check for BUILT_IN_SETJMP_RECEIVER.
6367 * tree-inline.c (copy_bb): Call notice_special_calls.
6368
6369 2014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6370
6371 PR bootstrap/59985
6372 * lra-constraints.c (process_alt_operands): Update reload_sum only
6373 on the first pass.
6374
6375 2014-01-31 Richard Henderson <rth@redhat.com>
6376
6377 PR middle-end/60004
6378 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6379 until after else_eh is processed.
6380
6381 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6382
6383 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6384 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6385 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6386 in smmintrin.h, remove them.
6387 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6388 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6389 * config/i386/i386.md (ROUND_SAE): Fix value.
6390 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6391 (const48_operand): New.
6392 * config/i386/subst.md (round), (round_expand): Use
6393 const_4_or_8_to_11_operand.
6394 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6395
6396 2014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6397
6398 * config/i386/constraints.md (Yk): Swap meaning with k.
6399 * config/i386/i386.md (movhi_internal): Change Yk to k.
6400 (movqi_internal): Ditto.
6401 (*k<logic><mode>): Ditto.
6402 (*andhi_1): Ditto.
6403 (*andqi_1): Ditto.
6404 (kandn<mode>): Ditto.
6405 (*<code>hi_1): Ditto.
6406 (*<code>qi_1): Ditto.
6407 (kxnor<mode>): Ditto.
6408 (kortestzhi): Ditto.
6409 (kortestchi): Ditto.
6410 (kunpckhi): Ditto.
6411 (*one_cmplhi2_1): Ditto.
6412 (*one_cmplqi2_1): Ditto.
6413 * config/i386/sse.md (): Change k to Yk.
6414 (avx512f_load<mode>_mask): Ditto.
6415 (avx512f_blendm<mode>): Ditto.
6416 (avx512f_store<mode>_mask): Ditto.
6417 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6418 (avx512f_storedqu<mode>_mask): Ditto.
6419 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6420 Ditto.
6421 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6422 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6423 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6424 (avx512f_maskcmp<mode>3): Ditto.
6425 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6426 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6427 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6428 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6429 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6430 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6431 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6432 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6433 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6434 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6435 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6436 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6437 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6438 (vec_extract_lo_<mode>_maskm): Ditto.
6439 (vec_extract_hi_<mode>_maskm): Ditto.
6440 (avx512f_vternlog<mode>_mask): Ditto.
6441 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6442 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6443 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6444 (avx512f_<code>v8div16qi2_mask): Ditto.
6445 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6446 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6447 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6448 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6449 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6450 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6451 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6452 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6453 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6454 (avx512cd_maskb_vec_dupv8di): Ditto.
6455 (avx512cd_maskw_vec_dupv16si): Ditto.
6456 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6457 (avx512f_vpermi2var<mode>3_mask): Ditto.
6458 (avx512f_vpermi2var<mode>3_mask): Ditto.
6459 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6460 (*avx512f_gathersi<mode>): Ditto.
6461 (*avx512f_gathersi<mode>_2): Ditto.
6462 (*avx512f_gatherdi<mode>): Ditto.
6463 (*avx512f_gatherdi<mode>_2): Ditto.
6464 (*avx512f_scattersi<mode>): Ditto.
6465 (*avx512f_scatterdi<mode>): Ditto.
6466 (avx512f_compress<mode>_mask): Ditto.
6467 (avx512f_compressstore<mode>_mask): Ditto.
6468 (avx512f_expand<mode>_mask): Ditto.
6469 * config/i386/subst.md (mask): Change k to Yk.
6470 (mask_scalar_merge): Ditto.
6471 (sd): Ditto.
6472
6473 2014-01-31 Marc Glisse <marc.glisse@inria.fr>
6474
6475 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6476
6477 2014-01-31 Richard Biener <rguenther@suse.de>
6478
6479 PR middle-end/59990
6480 * builtins.c (fold_builtin_memory_op): Make sure to not
6481 use a floating-point mode or a boolean or enumeral type for
6482 the copy operation.
6483
6484 2014-01-30 DJ Delorie <dj@redhat.com>
6485
6486 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6487 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6488 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6489 whenever main() has an epilogue.
6490
6491 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6492
6493 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6494 unused variable "field".
6495 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6496 (vsx_mergeh_<mode>): Likewise.
6497 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6498 (altivec_vmrghh): Likewise.
6499 (altivec_vmrghw): Likewise.
6500 (altivec_vmrglb): Likewise.
6501 (altivec_vmrglh): Likewise.
6502 (altivec_vmrglw): Likewise.
6503 (altivec_vspltb): Add missing uses.
6504 (altivec_vsplth): Likewise.
6505 (altivec_vspltw): Likewise.
6506 (altivec_vspltsf): Likewise.
6507
6508 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6509
6510 PR target/59923
6511 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6512 frame related instructions.
6513
6514 2014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6515
6516 PR rtl-optimization/59959
6517 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6518 any reload of register whose subreg is invalid.
6519
6520 2014-01-30 Jakub Jelinek <jakub@redhat.com>
6521
6522 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6523 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6524 Add missing return type - void.
6525
6526 2014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6527
6528 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6529 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6530 remove element index adjustment for endian (now handled in vsx.md
6531 and altivec.md).
6532 (altivec_expand_vec_perm_const): Use
6533 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6534 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6535 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6536 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6537 define_expand and a new define_insn *altivec_vspltb_internal;
6538 adjust for -maltivec=be on a little endian target.
6539 (altivec_vspltb_direct): New.
6540 (altivec_vsplth): Divide into a define_expand and a new
6541 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6542 little endian target.
6543 (altivec_vsplth_direct): New.
6544 (altivec_vspltw): Divide into a define_expand and a new
6545 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6546 little endian target.
6547 (altivec_vspltw_direct): New.
6548 (altivec_vspltsf): Divide into a define_expand and a new
6549 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6550 a little endian target.
6551
6552 2014-01-30 Richard Biener <rguenther@suse.de>
6553
6554 PR tree-optimization/59993
6555 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6556 can propagate form the earlier stmt and avoid the transform
6557 when the intermediate result is needed.
6558
6559 2014-01-30 Alangi Derick <alangiderick@gmail.com>
6560
6561 * README.Portability: Fix typo.
6562
6563 2014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
6564
6565 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6566 comparison_operator with ordered_comparison_operator.
6567
6568 2014-01-30 Nick Clifton <nickc@redhat.com>
6569
6570 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6571 Rename to mn10300_store_multiple_regs.
6572 * config/mn10300/mn10300.c: Likewise.
6573 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6574 store_multiple_regs.
6575 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6576 Call mn10300_store_multiple_regs.
6577
6578 2014-01-30 Nick Clifton <nickc@redhat.com>
6579 DJ Delorie <dj@redhat.com>
6580
6581 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6582 %fp 2 to keep registers after it properly word-aligned.
6583 (rl78_alloc_physical_registers_umul): Handle the case where both
6584 input operands are the same.
6585
6586 2014-01-30 Richard Biener <rguenther@suse.de>
6587
6588 PR tree-optimization/59903
6589 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6590 check properly.
6591
6592 2014-01-30 Jason Merrill <jason@redhat.com>
6593
6594 PR c++/59633
6595 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6596
6597 PR c++/59645
6598 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6599
6600 2014-01-30 Richard Biener <rguenther@suse.de>
6601
6602 PR tree-optimization/59951
6603 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
6604
6605 2014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6606
6607 PR target/59784
6608 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6609 SFmode to DFmode case.
6610
6611 2014-01-29 DJ Delorie <dj@redhat.com>
6612
6613 * config/msp430/msp430.opt (-minrt): New.
6614 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6615 if -minrt given.
6616 (ENDFILE_SPEC): Likewise.
6617
6618 2014-01-29 Jan Hubicka <hubicka@ucw.cz>
6619
6620 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6621 (estimate_function_body_sizes): Use it.
6622
6623 2014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6624
6625 PR c++/58561
6626 * dwarf2out.c (is_cxx_auto): New.
6627 (is_base_type): Use it.
6628 (gen_type_die_with_usage): Likewise.
6629
6630 2014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6631
6632 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6633 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6634 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6635 -maltivec=be with LE targets.
6636 (vsx_mergeh_<mode>): Likewise.
6637 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
6638 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6639 (altivec_vmrghb): Replace with define_expand and new
6640 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
6641 (altivec_vmrghb_direct): New define_insn.
6642 (altivec_vmrghh): Replace with define_expand and new
6643 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
6644 (altivec_vmrghh_direct): New define_insn.
6645 (altivec_vmrghw): Replace with define_expand and new
6646 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
6647 (altivec_vmrghw_direct): New define_insn.
6648 (*altivec_vmrghsf): Adjust for endianness.
6649 (altivec_vmrglb): Replace with define_expand and new
6650 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
6651 (altivec_vmrglb_direct): New define_insn.
6652 (altivec_vmrglh): Replace with define_expand and new
6653 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
6654 (altivec_vmrglh_direct): New define_insn.
6655 (altivec_vmrglw): Replace with define_expand and new
6656 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
6657 (altivec_vmrglw_direct): New define_insn.
6658 (*altivec_vmrglsf): Adjust for endianness.
6659 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6660 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6661 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6662 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6663 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6664 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6665 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6666 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6667
6668 2014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6669
6670 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6671 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6672 whitespace.
6673
6674 2014-01-29 Richard Biener <rguenther@suse.de>
6675
6676 PR tree-optimization/58742
6677 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6678 associate_pointerplus_align.
6679 (associate_pointerplus_diff): New function.
6680 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6681 and associate_pointerplus_diff.
6682
6683 2014-01-29 Richard Biener <rguenther@suse.de>
6684
6685 * lto-streamer.h (LTO_major_version): Bump to 3.
6686 (LTO_minor_version): Reset to 0.
6687
6688 2014-01-29 Renlin Li <Renlin.Li@arm.com>
6689
6690 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6691 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6692 (arm_file_start): Generate correct asm header for armv7ve.
6693 * config/arm/bpabi.h: Add multilib support for armv7ve.
6694 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6695 and cortex-a15 to armv7ve.
6696 * config/arm/t-aprofile: Add multilib support for armv7ve.
6697 * doc/invoke.texi: Document -march=armv7ve.
6698
6699 2014-01-29 Richard Biener <rguenther@suse.de>
6700
6701 PR tree-optimization/58742
6702 * tree-ssa-forwprop.c (associate_plusminus): Return true
6703 if we changed sth, defer EH cleanup to ...
6704 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6705 (simplify_mult): New function.
6706
6707 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6708
6709 PR middle-end/59917
6710 PR tree-optimization/59920
6711 * tree.c (build_common_builtin_nodes): Remove
6712 __builtin_setjmp_dispatcher initialization.
6713 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6714 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6715 instead of gsi_after_labels + manually skipping debug stmts.
6716 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6717 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6718 * tree-inline.c (copy_edges_for_bb): Remove
6719 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6720 argument. Ignore computed_goto_p stmts. Don't call
6721 make_abnormal_goto_edges. If a call might need abnormal edges
6722 for non-local gotos, see if it already has an edge to
6723 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6724 with true argument, don't do anything then, otherwise add
6725 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6726 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6727 caller.
6728 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6729 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6730 (lower_stmt): Don't set data->calls_builtin_setjmp.
6731 (lower_builtin_setjmp): Adjust comment.
6732 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6733 * tree-cfg.c (found_computed_goto): Remove.
6734 (factor_computed_gotos): Remove.
6735 (make_goto_expr_edges): Return bool, true for computed gotos.
6736 Don't call make_abnormal_goto_edges.
6737 (build_gimple_cfg): Don't set found_computed_goto, don't call
6738 factor_computed_gotos.
6739 (computed_goto_p): No longer static.
6740 (make_blocks): Don't set found_computed_goto.
6741 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6742 (make_edges): If make_goto_expr_edges returns true, push bb
6743 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6744 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6745 vector. Record mapping between bbs and OpenMP regions if there
6746 are any, adjust make_gimple_omp_edges caller. Call
6747 handle_abnormal_edges.
6748 (make_abnormal_goto_edges): Remove.
6749 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6750 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6751 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
6752 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
6753 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6754 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6755 filling *region also set *region_idx to (*region)->entry->index.
6756
6757 PR other/58712
6758 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6759 For REGs set ORIGINAL_REGNO.
6760
6761 2014-01-29 Bingfeng Mei <bmei@broadcom.com>
6762
6763 * doc/md.texi: Mention that a target shouldn't implement
6764 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6765 than hi/lo pair.
6766
6767 2014-01-29 Jakub Jelinek <jakub@redhat.com>
6768
6769 PR tree-optimization/59594
6770 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6771 a copy of the datarefs vector rather than the vector itself.
6772
6773 2014-01-28 Jason Merrill <jason@redhat.com>
6774
6775 PR c++/53756
6776 * dwarf2out.c (auto_die): New static.
6777 (gen_type_die_with_usage): Handle C++1y 'auto'.
6778 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6779 on definition.
6780
6781 2014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6782
6783 PR target/59672
6784 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6785 (SPEC_X32): Likewise.
6786 (SPEC_64): Likewise.
6787 * config/i386/i386.c (ix86_option_override_internal): Turn off
6788 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6789 for TARGET_16BIT.
6790 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6791 * config/i386/i386.h (TARGET_16BIT): New macro.
6792 (TARGET_16BIT_P): Likewise.
6793 * config/i386/i386.opt: Add m16.
6794 * doc/invoke.texi: Document -m16.
6795
6796 2014-01-28 Jakub Jelinek <jakub@redhat.com>
6797
6798 PR preprocessor/59935
6799 * input.c (location_get_source_line): Bail out on when line number
6800 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
6801
6802 2014-01-28 Richard Biener <rguenther@suse.de>
6803
6804 PR tree-optimization/58742
6805 * tree-ssa-forwprop.c (associate_plusminus): Handle
6806 pointer subtraction of the form (T)(P + A) - (T)P.
6807
6808 2014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6809
6810 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6811 at const_int_cost.
6812
6813 2014-01-28 Richard Biener <rguenther@suse.de>
6814
6815 Revert
6816 2014-01-28 Richard Biener <rguenther@suse.de>
6817
6818 PR rtl-optimization/45364
6819 PR rtl-optimization/59890
6820 * var-tracking.c (local_get_addr_clear_given_value): Handle
6821 already cleared slot.
6822 (val_reset): Handle not allocated local_get_addr_cache.
6823 (vt_find_locations): Use post-order on the inverted CFG.
6824
6825 2014-01-28 Richard Biener <rguenther@suse.de>
6826
6827 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6828
6829 2014-01-28 Richard Biener <rguenther@suse.de>
6830
6831 PR rtl-optimization/45364
6832 PR rtl-optimization/59890
6833 * var-tracking.c (local_get_addr_clear_given_value): Handle
6834 already cleared slot.
6835 (val_reset): Handle not allocated local_get_addr_cache.
6836 (vt_find_locations): Use post-order on the inverted CFG.
6837
6838 2014-01-28 Alan Modra <amodra@gmail.com>
6839
6840 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6841 * configure.ac <recursive call for build != host>: Define
6842 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6843 and LD_FOR_BUILD too.
6844 * configure: Regenerate.
6845
6846 2014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6847
6848 * config/i386/i386.c (get_builtin_code_for_version): Separate
6849 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6850 Broadwell from Haswell.
6851
6852 2014-01-27 Steve Ellcey <sellcey@mips.com>
6853
6854 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6855 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6856 * config/mips/mips.c (mips_option_override): Change setting
6857 of TARGET_DSP.
6858 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
6859 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6860 Change from Mask to Var.
6861
6862 2014-01-27 Jeff Law <law@redhat.com>
6863
6864 * ipa-inline.c (inline_small_functions): Fix typo.
6865
6866 2014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6867
6868 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6869 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6870 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6871 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6872 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6873 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6874 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6875 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6876 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6877 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6878 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6879 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6880 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6881 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6882 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6883 (_mm512_storeu_epi64): Ditto.
6884 (_mm512_cmpge_epi32_mask): Ditto.
6885 (_mm512_cmpge_epu32_mask): Ditto.
6886 (_mm512_cmpge_epi64_mask): Ditto.
6887 (_mm512_cmpge_epu64_mask): Ditto.
6888 (_mm512_cmple_epi32_mask): Ditto.
6889 (_mm512_cmple_epu32_mask): Ditto.
6890 (_mm512_cmple_epi64_mask): Ditto.
6891 (_mm512_cmple_epu64_mask): Ditto.
6892 (_mm512_cmplt_epi32_mask): Ditto.
6893 (_mm512_cmplt_epu32_mask): Ditto.
6894 (_mm512_cmplt_epi64_mask): Ditto.
6895 (_mm512_cmplt_epu64_mask): Ditto.
6896 (_mm512_cmpneq_epi32_mask): Ditto.
6897 (_mm512_cmpneq_epu32_mask): Ditto.
6898 (_mm512_cmpneq_epi64_mask): Ditto.
6899 (_mm512_cmpneq_epu64_mask): Ditto.
6900 (_mm512_expand_pd): Ditto.
6901 (_mm512_expand_ps): Ditto.
6902 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6903 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6904 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6905 * config/i386/i386.c (ix86_builtins): Add
6906 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6907 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6908 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6909 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6910 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6911 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6912 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6913 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6914 IX86_BUILTIN_PMOVUSQW512_MEM.
6915 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6916 __builtin_ia32_pmovsqd512mem_mask,
6917 __builtin_ia32_pmovqd512mem_mask,
6918 __builtin_ia32_pmovusqw512mem_mask,
6919 __builtin_ia32_pmovsqw512mem_mask,
6920 __builtin_ia32_pmovqw512mem_mask,
6921 __builtin_ia32_pmovusdw512mem_mask,
6922 __builtin_ia32_pmovsdw512mem_mask,
6923 __builtin_ia32_pmovdw512mem_mask,
6924 __builtin_ia32_pmovqb512mem_mask,
6925 __builtin_ia32_pmovusqb512mem_mask,
6926 __builtin_ia32_pmovsqb512mem_mask,
6927 __builtin_ia32_pmovusdb512mem_mask,
6928 __builtin_ia32_pmovsdb512mem_mask,
6929 __builtin_ia32_pmovdb512mem_mask.
6930 (bdesc_args): Add __builtin_ia32_expanddf512,
6931 __builtin_ia32_expandsf512.
6932 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6933 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6934 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6935 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6936 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6937 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6938 (avx512f_<code>v8div16qi2_mask_store): This.
6939 (avx512f_expand<mode>): New.
6940
6941 2014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6942
6943 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
6944 New.
6945 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6946 (_mm512_prefetch_i32scatter_pd): Ditto.
6947 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6948 (_mm512_prefetch_i64scatter_pd): Ditto.
6949 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6950 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6951 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6952 (_mm512_prefetch_i32scatter_ps): Ditto.
6953 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6954 (_mm512_prefetch_i64scatter_ps): Ditto.
6955 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6956 * config/i386/i386-builtin-types.def: Define
6957 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6958 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
6959 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6960 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6961 IX86_BUILTIN_SCATTERPFQPD.
6962 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6963 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6964 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6965 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6966 __builtin_ia32_scatterpfqps.
6967 (ix86_expand_builtin): Expand new built-ins.
6968 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6969 fix memory access data type.
6970 (*avx512pf_gatherpf<mode>_mask): Ditto.
6971 (*avx512pf_gatherpf<mode>): Ditto.
6972 (avx512pf_scatterpf<mode>): Ditto.
6973 (*avx512pf_scatterpf<mode>_mask): Ditto.
6974 (*avx512pf_scatterpf<mode>): Ditto.
6975 (GATHER_SCATTER_SF_MEM_MODE): New.
6976 (avx512pf_gatherpf<mode>df): Ditto.
6977 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6978 (*avx512pf_scatterpf<mode>df): Ditto.
6979
6980 2014-01-27 Jakub Jelinek <jakub@redhat.com>
6981
6982 PR bootstrap/59934
6983 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6984 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6985 reached.
6986
6987 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6988
6989 * common/config/arm/arm-common.c
6990 (arm_rewrite_mcpu): Handle multiple names.
6991 * config/arm/arm.h
6992 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6993
6994 2014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6995
6996 * gimple-builder.h (create_gimple_tmp): Delete.
6997
6998 2014-01-27 Christian Bruel <christian.bruel@st.com>
6999
7000 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
7001 words comparisons.
7002
7003 2014-01-26 John David Anglin <danglin@gcc.gnu.org>
7004
7005 * config/pa/pa.md (call): Generate indirect long calls to non-local
7006 functions when outputing 32-bit code.
7007 (call_value): Likewise except for special call to buggy powf function.
7008
7009 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
7010 portable runtime and PIC indirect calls.
7011 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
7012 and PIC call sequences. Use ldo instead of blr to set return register
7013 in PIC call sequence.
7014
7015 2014-01-25 Walter Lee <walt@tilera.com>
7016
7017 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
7018 avoid clobbering a live register.
7019
7020 2014-01-25 Walter Lee <walt@tilera.com>
7021
7022 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
7023 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
7024 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
7025 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7026
7027 2014-01-25 Walter Lee <walt@tilera.com>
7028
7029 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7030 arguments on even registers.
7031 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7032 STACK_BOUNDARY.
7033 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7034 (BIGGEST_ALIGNMENT): Ditto.
7035 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7036
7037 2014-01-25 Walter Lee <walt@tilera.com>
7038
7039 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7040 insns before bundling.
7041 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
7042
7043 2014-01-25 Walter Lee <walt@tilera.com>
7044
7045 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7046 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7047 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
7048
7049 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7050
7051 * config/mips/constraints.md (kl): Delete.
7052 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7053 define expands, using...
7054 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7055 instructions for MIPS16.
7056 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7057 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7058
7059 2014-01-25 Walter Lee <walt@tilera.com>
7060
7061 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
7062 (clzdi2): Ditto.
7063 (ffsdi2): Ditto.
7064
7065 2014-01-25 Walter Lee <walt@tilera.com>
7066
7067 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7068 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7069
7070 2014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7071
7072 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7073 Handle XOR.
7074
7075 2014-01-25 Jakub Jelinek <jakub@redhat.com>
7076
7077 * print-rtl.c (in_call_function_usage): New var.
7078 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7079 EXPR_LIST mode as mode and not as reg note name.
7080
7081 PR middle-end/59561
7082 * cfgloopmanip.c (copy_loop_info): If
7083 loop->warned_aggressive_loop_optimizations, make sure
7084 the flag is set in target loop too.
7085
7086 2014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7087
7088 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7089 flag_cilkplus.
7090 * builtins.def: Likewise.
7091 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7092 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7093 * ira.c (ira_setup_eliminable_regset): Likewise.
7094 * omp-low.c (gate_expand_omp): Likewise.
7095 (execute_lower_omp): Likewise.
7096 (diagnose_sb_0): Likewise.
7097 (gate_diagnose_omp_blocks): Likewise.
7098 (simd_clone_clauses_extract): Likewise.
7099 (gate): Likewise.
7100
7101 2014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7102
7103 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7104 correction for little endian...
7105 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7106 here.
7107
7108 2014-01-24 Jeff Law <law@redhat.com>
7109
7110 PR tree-optimization/59919
7111 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7112 for non-returning calls.
7113
7114 2014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7115
7116 * common/config/aarch64/aarch64-common.c
7117 (aarch64_rewrite_mcpu): Handle multiple names.
7118 * config/aarch64/aarch64.h
7119 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7120
7121 2014-01-24 Dodji Seketeli <dodji@redhat.com>
7122
7123 * input.c (add_file_to_cache_tab): Handle the case where fopen
7124 returns NULL.
7125
7126 2014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7127
7128 PR target/59929
7129 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7130 from push operand if code of push isn't PRE_DEC.
7131
7132 2014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7133
7134 PR target/59909
7135 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7136 -mquad-memory-atomic. Update -mquad-memory documentation to say
7137 it is only used for non-atomic loads/stores.
7138
7139 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7140 -mquad-memory or -mquad-memory-atomic switches.
7141
7142 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7143 -mquad-memory-atomic to ISA 2.07 support.
7144
7145 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
7146 to separate support of normal quad word memory operations (ldq, stq)
7147 from the atomic quad word memory operations.
7148
7149 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7150 support to separate non-atomic quad word operations from atomic
7151 quad word operations. Disable non-atomic quad word operations in
7152 little endian mode so that we don't have to swap words after the
7153 load and before the store.
7154 (quad_load_store_p): Add comment about atomic quad word support.
7155 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7156 options printed with -mdebug=reg.
7157
7158 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7159 -mquad-memory-atomic as the test for whether we have quad word
7160 atomic instructions.
7161 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7162 or -mp8-vector are used, allow byte/half-word atomic operations.
7163
7164 * config/rs6000/sync.md (load_lockedti): Insure that the address
7165 is a proper indexed or indirect address for the lqarx instruction.
7166 On little endian systems, swap the hi/lo registers after the lqarx
7167 instruction.
7168 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7169 insure the address is valid for the lqarx instruction.
7170 (store_conditionalti): Insure that the address is a proper indexed
7171 or indirect address for the stqcrx. instruction. On little endian
7172 systems, swap the hi/lo registers before doing the stqcrx.
7173 instruction.
7174 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7175 insure the address is valid for the stqcrx. instruction.
7176
7177 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7178 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7179 type of quad memory support is available.
7180
7181 2014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7182
7183 PR regression/59915
7184 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7185 there is a danger of looping.
7186
7187 2014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7188
7189 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7190 force flag_ira_loop_pressure if set via command line.
7191
7192 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7193
7194 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7195 (ashr_simd): New builtin handling DI mode.
7196 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7197 (aarch64_sshr_simddi): New match pattern.
7198 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7199 (vshrd_n_s64): Likewise.
7200 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7201
7202 2014-01-23 Nick Clifton <nickc@redhat.com>
7203
7204 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7205 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7206 favour of mcu specific scripts.
7207 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7208 430x multilibs.
7209
7210 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7211 Alex Velenko <Alex.Velenko@arm.com>
7212
7213 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7214 (vaddv_s16): Likewise.
7215 (vaddv_s32): Likewise.
7216 (vaddv_u8): Likewise.
7217 (vaddv_u16): Likewise.
7218 (vaddv_u32): Likewise.
7219 (vaddvq_s8): Likewise.
7220 (vaddvq_s16): Likewise.
7221 (vaddvq_s32): Likewise.
7222 (vaddvq_s64): Likewise.
7223 (vaddvq_u8): Likewise.
7224 (vaddvq_u16): Likewise.
7225 (vaddvq_u32): Likewise.
7226 (vaddvq_u64): Likewise.
7227 (vaddv_f32): Likewise.
7228 (vaddvq_f32): Likewise.
7229 (vaddvq_f64): Likewise.
7230 (vmaxv_f32): Likewise.
7231 (vmaxv_s8): Likewise.
7232 (vmaxv_s16): Likewise.
7233 (vmaxv_s32): Likewise.
7234 (vmaxv_u8): Likewise.
7235 (vmaxv_u16): Likewise.
7236 (vmaxv_u32): Likewise.
7237 (vmaxvq_f32): Likewise.
7238 (vmaxvq_f64): Likewise.
7239 (vmaxvq_s8): Likewise.
7240 (vmaxvq_s16): Likewise.
7241 (vmaxvq_s32): Likewise.
7242 (vmaxvq_u8): Likewise.
7243 (vmaxvq_u16): Likewise.
7244 (vmaxvq_u32): Likewise.
7245 (vmaxnmv_f32): Likewise.
7246 (vmaxnmvq_f32): Likewise.
7247 (vmaxnmvq_f64): Likewise.
7248 (vminv_f32): Likewise.
7249 (vminv_s8): Likewise.
7250 (vminv_s16): Likewise.
7251 (vminv_s32): Likewise.
7252 (vminv_u8): Likewise.
7253 (vminv_u16): Likewise.
7254 (vminv_u32): Likewise.
7255 (vminvq_f32): Likewise.
7256 (vminvq_f64): Likewise.
7257 (vminvq_s8): Likewise.
7258 (vminvq_s16): Likewise.
7259 (vminvq_s32): Likewise.
7260 (vminvq_u8): Likewise.
7261 (vminvq_u16): Likewise.
7262 (vminvq_u32): Likewise.
7263 (vminnmv_f32): Likewise.
7264 (vminnmvq_f32): Likewise.
7265 (vminnmvq_f64): Likewise.
7266
7267 2014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7268
7269 * config/aarch64/aarch64-simd.md
7270 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7271 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7272 (*aarch64_mul3_elt<mode>): Likewise.
7273 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7274 (*aarch64_mul3_elt_to_64v2df): Likewise.
7275 (*aarch64_mla_elt<mode>): Likewise.
7276 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7277 (*aarch64_mls_elt<mode>): Likewise.
7278 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7279 (*aarch64_fma4_elt<mode>): Likewise.
7280 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7281 (*aarch64_fma4_elt_to_64v2df): Likewise.
7282 (*aarch64_fnma4_elt<mode>): Likewise.
7283 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7284 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7285 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7286 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7287 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7288 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7289 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7290 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7291 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7292
7293 2013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7294
7295 * config/aarch64/aarch64-simd.md
7296 (aarch64_be_checked_get_lane<mode>): New define_expand.
7297 * config/aarch64/aarch64-simd-builtins.def
7298 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
7299 New builtin definition.
7300 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7301 Use new safe be builtin.
7302
7303 2014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7304
7305 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7306 New define_insn.
7307 (aarch64_be_st1<mode>): Likewise.
7308 (aarch_ld1<VALL:mode>): Define_expand modified.
7309 (aarch_st1<VALL:mode>): Likewise.
7310 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7311 (UNSPEC_ST1): Likewise.
7312
7313 2014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
7314
7315 * config/microblaze/microblaze.md: Add trap insn and attribute
7316
7317 2014-01-23 Dodji Seketeli <dodji@redhat.com>
7318
7319 PR preprocessor/58580
7320 * input.h (location_get_source_line): Take an additional line_size
7321 parameter.
7322 (void diagnostics_file_cache_fini): Declare new function.
7323 * input.c (struct fcache): New type.
7324 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7325 New static constants.
7326 (diagnostic_file_cache_init, total_lines_num)
7327 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7328 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7329 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7330 (get_next_line, read_next_line, goto_next_line, read_line_num):
7331 New static function definitions.
7332 (diagnostic_file_cache_fini): New function.
7333 (location_get_source_line): Take an additional output line_len
7334 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7335 read_line_num.
7336 * diagnostic.c (diagnostic_finish): Call
7337 diagnostic_file_cache_fini.
7338 (adjust_line): Take an additional input parameter for the length
7339 of the line, rather than calculating it with strlen.
7340 (diagnostic_show_locus): Adjust the use of
7341 location_get_source_line and adjust_line with respect to their new
7342 signature. While displaying a line now, do not stop at the first
7343 null byte. Rather, display the zero byte as a space and keep
7344 going until we reach the size of the line.
7345 * Makefile.in: Add vec.o to OBJS-libcommon
7346
7347 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7348 Ilya Tocar <ilya.tocar@intel.com>
7349
7350 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7351 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7352 (__builtin_ia32_kmov16): Ditto.
7353 * config/i386/i386.md (UNSPEC_KMOV): New.
7354 (kmovw): Ditto.
7355
7356 2014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7357
7358 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7359 (_mm512_storeu_si512): Ditto.
7360
7361 2014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7362
7363 PR target/52125
7364 * rtl.h (get_referenced_operands): Declare.
7365 * recog.c (get_referenced_operands): New function.
7366 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7367 operands have been referenced when recording LO_SUM references.
7368
7369 2014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
7370
7371 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7372
7373 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7374
7375 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7376 Enable for generic and recent AMD targets.
7377
7378 2014-01-22 Jan Hubicka <hubicka@ucw.cz>
7379
7380 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7381 ARG_SIZE note when adjustment was eliminated.
7382
7383 2014-01-22 Jeff Law <law@redhat.com>
7384
7385 PR tree-optimization/59597
7386 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7387 in file. Accept new argument REGISTERING and use it to modify
7388 dump output appropriately.
7389 (register_jump_thread): Corresponding changes.
7390 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7391 thread paths involving joiner blocks. Add code to dump cancelled
7392 jump threading paths.
7393
7394 2014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7395
7396 PR rtl-optimization/59477
7397 * lra-constraints.c (inherit_in_ebb): Process call for living hard
7398 regs. Update reloads_num and potential_reload_hard_regs for all insns.
7399
7400 2014-01-22 Tom Tromey <tromey@redhat.com>
7401
7402 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7403 PARAMS.
7404 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7405
7406 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7407
7408 PR rtl-optimization/59896
7409 * lra-constraints.c (process_alt_operands): Check unused note for
7410 matched operands of insn with no output reloads.
7411
7412 2014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7413
7414 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7415 (mips_move_from_gpr_cost): Likewise.
7416
7417 2014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7418
7419 PR rtl-optimization/59858
7420 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7421 ira_class_hard_regs_num.
7422 (process_alt_operands): Increase reject for dying matched operand.
7423
7424 2014-01-21 Jakub Jelinek <jakub@redhat.com>
7425
7426 PR target/59003
7427 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7428 smaller than size, perform several stores or loads and stores
7429 at dst + count - size to store or copy all of size bytes, rather
7430 than just last modesize bytes.
7431
7432 2014-01-20 DJ Delorie <dj@redhat.com>
7433
7434 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7435 that CLOBBERs are REGs before propogating their values.
7436
7437 2014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7438
7439 PR middle-end/59789
7440 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7441 (cgraph_inline_failed_type): New function.
7442 * cgraph.h (DEFCIFCODE): Add type.
7443 (cgraph_inline_failed_type_t): New enum.
7444 (cgraph_inline_failed_type): New prototype.
7445 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7446 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7447 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7448 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7449 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7450 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7451 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
7452 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
7453 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7454 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7455 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7456 OPTIMIZATION_MISMATCH.
7457 * tree-inline.c (expand_call_inline): Emit errors during
7458 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
7459
7460 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7461
7462 PR target/59685
7463 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7464 mode attribute in insn output.
7465
7466 2014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7467
7468 * output.h (output_constant): Delete.
7469 * varasm.c (output_constant): Make private.
7470
7471 2014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7472
7473 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7474
7475 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7476
7477 PR middle-end/59860
7478 * tree.h (fold_builtin_strcat): New prototype.
7479 * builtins.c (fold_builtin_strcat): No longer static. Add len
7480 argument, if non-NULL, don't call c_strlen. Optimize
7481 directly into __builtin_memcpy instead of __builtin_strcpy.
7482 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7483 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7484
7485 2014-01-20 Uros Bizjak <ubizjak@gmail.com>
7486
7487 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7488 for SImode_address_operand operands, having only a REG argument.
7489
7490 2014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7491
7492 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7493 loader name using mbig-endian.
7494 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7495
7496 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7497
7498 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7499 (-mtune): Likewise.
7500 (-mcpu): Likewise.
7501
7502 2014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7503
7504 * config/aarch64/aarch64-protos.h
7505 (aarch64_cannot_change_mode_class_ptr): Declare.
7506 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7507 aarch64_cannot_change_mode_class_ptr): New.
7508 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7509 backend hook aarch64_cannot_change_mode_class.
7510
7511 2014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7512
7513 * common/config/aarch64/aarch64-common.c
7514 (aarch64_handle_option): Don't handle any option order logic here.
7515 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7516 selected_cpu, warn on architecture version mismatch.
7517 (aarch64_override_options): Fix parsing order for option strings.
7518
7519 2014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7520 Iain Sandoe <iain@codesourcery.com>
7521
7522 PR bootstrap/59496
7523 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7524 warning. Amend comment to reflect current functionality.
7525
7526 2014-01-20 Richard Biener <rguenther@suse.de>
7527
7528 PR middle-end/59860
7529 * builtins.c (fold_builtin_strcat): Remove case better handled
7530 by tree-ssa-strlen.c.
7531
7532 2014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7533
7534 * config/aarch64/aarch64.opt
7535 (mcpu, march, mtune): Make case-insensitive.
7536
7537 2014-01-20 Jakub Jelinek <jakub@redhat.com>
7538
7539 PR target/59880
7540 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7541 if operands[1] is a REG or ZERO_EXTEND of a REG.
7542
7543 2014-01-19 Jan Hubicka <hubicka@ucw.cz>
7544
7545 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7546
7547 2014-01-19 John David Anglin <danglin@gcc.gnu.org>
7548
7549 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7550 long non-pic millicode calls.
7551
7552 2014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7553
7554 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7555
7556 2014-01-19 Kito Cheng <kito@0xlab.org>
7557
7558 * builtins.c (expand_movstr): Check movstr expand done or fail.
7559
7560 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7561 H.J. Lu <hongjiu.lu@intel.com>
7562
7563 PR target/59379
7564 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7565 to DImode for zero-extended addresses.
7566
7567 2014-01-19 Jakub Jelinek <jakub@redhat.com>
7568
7569 PR rtl-optimization/57763
7570 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7571 on the new indirect jump_insn and increment LABEL_NUSES (label).
7572
7573 2014-01-18 H.J. Lu <hongjiu.lu@intel.com>
7574
7575 PR bootstrap/59580
7576 PR bootstrap/59583
7577 * config.gcc (x86_archs): New variable.
7578 (x86_64_archs): Likewise.
7579 (x86_cpus): Likewise.
7580 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7581 --with-arch/--with-cpu= options.
7582 Support --with-arch=/--with-cpu={nehalem,westmere,
7583 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7584
7585 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7586
7587 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7588 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7589
7590 2014-01-18 Uros Bizjak <ubizjak@gmail.com>
7591
7592 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7593
7594 2014-01-18 Jakub Jelinek <jakub@redhat.com>
7595
7596 PR target/58944
7597 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7598 clear cpp_get_options (parse_in)->warn_unused_macros for
7599 ix86_target_macros_internal with cpp_define.
7600
7601 2014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7602
7603 * jump.c (delete_related_insns): Keep (use (insn))s.
7604 * reorg.c (redundant_insn): Check for barriers too.
7605
7606 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7607
7608 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
7609
7610 2014-01-17 John David Anglin <danglin@gcc.gnu.org>
7611
7612 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7613 call to $$dyncall when TARGET_LONG_CALLS is true.
7614
7615 2014-01-17 Jeff Law <law@redhat.com>
7616
7617 * ree.c (combine_set_extension): Temporarily disable test for
7618 changing number of hard registers.
7619
7620 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7621
7622 PR middle-end/58125
7623 * ipa-inline-analysis.c (inline_free_summary):
7624 Do not free summary of aliases.
7625
7626 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7627
7628 PR middle-end/59706
7629 * gimplify.c (gimplify_expr): Use create_tmp_var
7630 instead of create_tmp_var_raw. If cond doesn't have
7631 integral type, don't add the IFN_ANNOTATE builtin at all.
7632
7633 2014-01-17 Martin Jambor <mjambor@suse.cz>
7634
7635 PR ipa/59736
7636 * ipa-cp.c (prev_edge_clone): New variable.
7637 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7638 Also resize prev_edge_clone vector.
7639 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7640 (ipcp_edge_removal_hook): New function.
7641 (ipcp_driver): Register ipcp_edge_removal_hook.
7642
7643 2014-01-17 Andrew Pinski <apinski@cavium.com>
7644 Steve Ellcey <sellcey@mips.com>
7645
7646 PR target/59462
7647 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7648 of operator mode.
7649
7650 2014-01-17 Jeff Law <law@redhat.com>
7651
7652 PR middle-end/57904
7653 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7654 so that pass_ccp runs first.
7655
7656 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7657
7658 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7659 (ix86_adjust_cost): Use !TARGET_XXX.
7660 (do_reorder_for_imul): Likewise.
7661 (swap_top_of_ready_list): Likewise.
7662 (ix86_sched_reorder): Likewise.
7663
7664 2014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7665
7666 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7667 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7668 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7669 (intel_memset): New. Duplicate slm_memset.
7670 (intel_cost): New. Duplicate slm_cost.
7671 (m_INTEL): New macro.
7672 (processor_target_table): Add "intel".
7673 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7674 with PROCESSOR_INTEL for "intel".
7675 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7676 PROCESSOR_SILVERMONT.
7677 (ix86_issue_rate): Likewise.
7678 (ix86_adjust_cost): Likewise.
7679 (ia32_multipass_dfa_lookahead): Likewise.
7680 (swap_top_of_ready_list): Likewise.
7681 (ix86_sched_reorder): Likewise.
7682 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7683 instead of TARGET_OPT_AGU.
7684 * config/i386/i386.h (TARGET_INTEL): New.
7685 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7686 (processor_type): Add PROCESSOR_INTEL.
7687 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7688 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
7689
7690 2014-01-17 Marek Polacek <polacek@redhat.com>
7691
7692 PR c/58346
7693 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7694 size is zero.
7695
7696 2014-01-17 Richard Biener <rguenther@suse.de>
7697
7698 PR tree-optimization/46590
7699 * opts.c (default_options_table): Add entries for
7700 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7701 all enabled at -O1 but not for -Og.
7702 * common.opt (fbranch-count-reg): Remove Init(1).
7703 (fmove-loop-invariants): Likewise.
7704 (ftree-pta): Likewise.
7705
7706 2014-01-17 Jakub Jelinek <jakub@redhat.com>
7707
7708 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7709 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7710 decls to at least the GCC 4.8 used alignments.
7711
7712 PR fortran/59440
7713 * tree-nested.c (convert_nonlocal_reference_stmt,
7714 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7715 of GIMPLE_BIND stmts, adjust associated decls.
7716
7717 2014-01-17 Richard Biener <rguenther@suse.de>
7718
7719 PR tree-optimization/46590
7720 * vec.h (vec<>::bseach): New member function implementing
7721 binary search according to C89 bsearch.
7722 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7723 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7724 bitmap pointer again. Make accesses_in_loop a flat array.
7725 (mem_ref_obstack): New global.
7726 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7727 (mark_ref_stored): Likewise.
7728 (ref_indep_loop_p_2): Likewise.
7729 (set_ref_stored_in_loop): New helper function.
7730 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7731 (memref_free): Adjust.
7732 (record_mem_ref_loc): Simplify.
7733 (gather_mem_refs_stmt): Adjust.
7734 (sort_locs_in_loop_postorder_cmp): New function.
7735 (analyze_memory_references): Sort accesses_in_loop after
7736 loop postorder number.
7737 (find_ref_loc_in_loop_cmp): New function.
7738 (for_all_locs_in_loop): Find relevant cluster of locs in
7739 accesses_in_loop and iterate without recursion.
7740 (execute_sm): Avoid uninit warning.
7741 (struct ref_always_accessed): Simplify.
7742 (ref_always_accessed::operator ()): Likewise.
7743 (ref_always_accessed_p): Likewise.
7744 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7745 loop postorder numbers here.
7746 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7747 numbers.
7748
7749 2014-01-17 Jan Hubicka <hubicka@ucw.cz>
7750
7751 PR c++/57945
7752 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7753 on decls for which assemble_alias has been called.
7754
7755 2014-01-17 Nick Clifton <nickc@redhat.com>
7756
7757 * config/msp430/msp430.opt: (mcpu): New option.
7758 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7759 (msp430_option_override): Parse target_cpu. If the MCU name
7760 matches a generic string, clear target_mcu.
7761 (msp430_attr): Allow numeric interrupt values up to 63.
7762 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7763 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7764 option.
7765 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7766 Add mcpu matches.
7767 * config/msp430/msp430.md (popm): Use %J rather than %I.
7768 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7769 (addhi_cy_i): Use immediate_operand for operand 2.
7770 * doc/invoke.texi: Document -mcpu option.
7771
7772 2014-01-17 Richard Biener <rguenther@suse.de>
7773
7774 PR rtl-optimization/38518
7775 * df.h (df_analyze_loop): Declare.
7776 * df-core.c: Include cfgloop.h.
7777 (df_analyze_1): Split out main part of df_analyze.
7778 (df_analyze): Adjust.
7779 (loop_inverted_post_order_compute): New function.
7780 (loop_post_order_compute): Likewise.
7781 (df_analyze_loop): New function avoiding whole-function
7782 postorder computes.
7783 * loop-invariant.c (find_defs): Use df_analyze_loop.
7784 (find_invariants): Adjust.
7785 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7786
7787 2014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7788
7789 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7790 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7791
7792 2014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7793
7794 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7795 traversal when removing references.
7796
7797 2014-01-16 Jan Hubicka <hubicka@ucw.cz>
7798
7799 PR ipa/59775
7800 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7801
7802 2014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7803
7804 PR middle-end/56791
7805 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7806 pushing a reload for an autoinc when we had previously reloaded an
7807 inner part of the address.
7808
7809 2014-01-16 Jakub Jelinek <jakub@redhat.com>
7810
7811 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7812 field.
7813 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7814 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7815 when not giving up or versioning for alias only because of
7816 loop->safelen.
7817 (vect_analyze_data_ref_dependences): Set to true.
7818 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7819 is a GIMPLE_PHI.
7820 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7821 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7822 to the condition.
7823
7824 PR middle-end/58344
7825 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7826
7827 PR target/59839
7828 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7829 operand 0 predicate for gathers, use a new pseudo as subtarget.
7830
7831 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7832
7833 PR middle-end/59609
7834 * lra-constraints.c (process_alt_operands): Add printing debug info.
7835 Check absence of input/output reloads for matched operands too.
7836
7837 2014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7838
7839 PR rtl-optimization/59835
7840 * ira.c (ira_init_register_move_cost): Increase cost for
7841 impossible modes.
7842
7843 2014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7844
7845 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
7846
7847 2014-01-16 Richard Earnshaw <rearnsha@arm.com>
7848
7849 PR target/59780
7850 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7851 non-register objects. Use gen_(high/low)part more consistently.
7852 Fix assertions.
7853
7854 2014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7855
7856 PR target/59844
7857 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7858 endian support, remove tests for WORDS_BIG_ENDIAN.
7859 (p8_mfvsrd_3_<mode>): Likewise.
7860 (reload_gpr_from_vsx<mode>): Likewise.
7861 (reload_gpr_from_vsxsf): Likewise.
7862 (p8_mfvsrd_4_disf): Likewise.
7863
7864 2014-01-16 Richard Biener <rguenther@suse.de>
7865
7866 PR rtl-optimization/46590
7867 * lcm.c (compute_antinout_edge): Use postorder iteration.
7868 (compute_laterin): Use inverted postorder iteration.
7869
7870 2014-01-16 Nick Clifton <nickc@redhat.com>
7871
7872 PR middle-end/28865
7873 * varasm.c (output_constant): Return the number of bytes actually
7874 emitted.
7875 (output_constructor_array_range): Update the field size with the
7876 number of bytes emitted by output_constant.
7877 (output_constructor_regular_field): Likewise. Also do not
7878 complain if the total number of bytes emitted is now greater
7879 than the expected fieldpos.
7880 * output.h (output_constant): Update prototype and descriptive comment.
7881
7882 2014-01-16 Marek Polacek <polacek@redhat.com>
7883
7884 PR middle-end/59827
7885 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7886 it is error_mark_node.
7887
7888 2014-01-15 Uros Bizjak <ubizjak@gmail.com>
7889
7890 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7891 VALID_AVX256_REG_OR_OI_MODE.
7892
7893 2014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7894
7895 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7896 current procedure should be profiled.
7897
7898 2014-01-15 Andrew Pinski <apinski@cavium.com>
7899
7900 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7901 of moving from/to the STACK_REG register class.
7902
7903 2014-01-15 Richard Henderson <rth@redhat.com>
7904
7905 PR debug/54694
7906 * reginfo.c (global_regs_decl): Globalize.
7907 * rtl.h (global_regs_decl): Declare.
7908 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7909 reserved via global_regs.
7910
7911 2014-01-15 Teresa Johnson <tejohnson@google.com>
7912
7913 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7914
7915 2014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7916
7917 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7918 and vmulosh rather than call gen_vec_widen_smult_*.
7919 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7920 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7921 (vec_widen_smult_even_v16qi): Likewise.
7922 (vec_widen_umult_even_v8hi): Likewise.
7923 (vec_widen_smult_even_v8hi): Likewise.
7924 (vec_widen_umult_odd_v16qi): Likewise.
7925 (vec_widen_smult_odd_v16qi): Likewise.
7926 (vec_widen_umult_odd_v8hi): Likewise.
7927 (vec_widen_smult_odd_v8hi): Likewise.
7928 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7929 vmuloub rather than call gen_vec_widen_umult_*.
7930 (vec_widen_umult_lo_v16qi): Likewise.
7931 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7932 vmulosb rather than call gen_vec_widen_smult_*.
7933 (vec_widen_smult_lo_v16qi): Likewise.
7934 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7935 rather than call gen_vec_widen_umult_*.
7936 (vec_widen_umult_lo_v8hi): Likewise.
7937 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7938 rather than call gen_vec_widen_smult_*.
7939 (vec_widen_smult_lo_v8hi): Likewise.
7940
7941 2014-01-15 Jeff Law <law@redhat.com>
7942
7943 PR tree-optimization/59747
7944 * ree.c (find_and_remove_re): Properly handle case where a second
7945 eliminated extension requires widening a copy created for elimination
7946 of a prior extension.
7947 (combine_set_extension): Ensure that the number of hard regs needed
7948 for a destination register does not change when we widen it.
7949
7950 2014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7951
7952 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7953 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7954 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7955 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7956 (avr-*-rtems*): Likewise.
7957 (bfin*-rtems*): Likewise.
7958 (moxie-*-rtems*): Likewise.
7959 (h8300-*-rtems*): Likewise.
7960 (i[34567]86-*-rtems*): Likewise.
7961 (lm32-*-rtems*): Likewise.
7962 (m32r-*-rtems*): Likewise.
7963 (m68k-*-rtems*): Likewise.
7964 (microblaze*-*-rtems*): Likewise.
7965 (mips*-*-rtems*): Likewise.
7966 (powerpc-*-rtems*): Likewise.
7967 (sh-*-rtems*): Likewise.
7968 (sparc-*-rtems*): Likewise.
7969 (sparc64-*-rtems*): Likewise.
7970 (v850-*-rtems*): Likewise.
7971 (m32c-*-rtems*): Likewise.
7972
7973 2014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7974
7975 PR rtl-optimization/59511
7976 * ira.c (ira_init_register_move_cost): Use memory costs for some
7977 cases of register move cost calculations.
7978 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7979 instead of BB frequency.
7980 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7981 * lra-assigns.c (find_hard_regno_for): Ditto.
7982
7983 2014-01-15 Richard Biener <rguenther@suse.de>
7984
7985 PR tree-optimization/59822
7986 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7987 (vectorizable_load): Use it to hoist defs of uses of invariant
7988 loads out of the loop.
7989
7990 2014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
7991 Kugan Vivekanandarajah <kuganv@linaro.org>
7992
7993 PR target/59695
7994 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7995 truncation.
7996
7997 2014-01-15 Richard Biener <rguenther@suse.de>
7998
7999 PR rtl-optimization/59802
8000 * lcm.c (compute_available): Use inverted postorder to seed
8001 the initial worklist.
8002
8003 2014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8004
8005 PR target/59803
8006 * config/s390/s390.c (s390_preferred_reload_class): Don't return
8007 ADDR_REGS for invalid symrefs in non-PIC code.
8008
8009 2014-01-15 Jakub Jelinek <jakub@redhat.com>
8010
8011 PR other/58712
8012 * builtins.c (determine_block_size): Initialize *probable_max_size
8013 even if len_rtx is CONST_INT.
8014
8015 2014-01-14 Andrew Pinski <apinski@cavium.com>
8016
8017 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
8018 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
8019 (cortexa53_tunings): Likewise.
8020 (aarch64_sched_issue_rate): New function.
8021 (TARGET_SCHED_ISSUE_RATE): Define.
8022
8023 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8024
8025 * ira-costs.c (find_costs_and_classes): Add missed
8026 ira_init_register_move_cost_if_necessary.
8027
8028 2014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8029
8030 PR target/59787
8031 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8032
8033 2014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8034
8035 PR target/59794
8036 * config/i386/i386.c (type_natural_mode): Add a bool parameter
8037 to indicate if type is used for function return value. Warn ABI
8038 change if the vector mode isn't available for function return value.
8039 (ix86_function_arg_advance): Pass false to type_natural_mode.
8040 (ix86_function_arg): Likewise.
8041 (ix86_gimplify_va_arg): Likewise.
8042 (function_arg_32): Don't warn ABI change.
8043 (ix86_function_value): Pass true to type_natural_mode.
8044 (ix86_return_in_memory): Likewise.
8045 (ix86_struct_value_rtx): Removed.
8046 (TARGET_STRUCT_VALUE_RTX): Likewise.
8047
8048 2014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8049
8050 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8051 converting a conditional jump into a conditional return.
8052
8053 2014-01-14 Richard Biener <rguenther@suse.de>
8054
8055 PR tree-optimization/58921
8056 PR tree-optimization/59006
8057 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8058 hoisting invariant stmts.
8059 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8060 invariant loads on the preheader edge if possible.
8061
8062 2014-01-14 Joey Ye <joey.ye@arm.com>
8063
8064 * doc/plugin.texi (Building GCC plugins): Update to C++.
8065
8066 2014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8067
8068 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
8069 (_mm_rcp28_round_ss): Ditto.
8070 (_mm_rsqrt28_round_sd): Ditto.
8071 (_mm_rsqrt28_round_ss): Ditto.
8072 (_mm_rcp28_sd): Ditto.
8073 (_mm_rcp28_ss): Ditto.
8074 (_mm_rsqrt28_sd): Ditto.
8075 (_mm_rsqrt28_ss): Ditto.
8076 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8077 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8078 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8079 (IX86_BUILTIN_RCP28SD): Ditto.
8080 (IX86_BUILTIN_RCP28SS): Ditto.
8081 (IX86_BUILTIN_RSQRT28SD): Ditto.
8082 (IX86_BUILTIN_RSQRT28SS): Ditto.
8083 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8084 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8085 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8086 (ix86_expand_special_args_builtin): Expand new FTYPE.
8087 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8088 (srcp14<mode>): Make insn unary.
8089 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8090 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8091 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8092 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8093 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8094 Fix rounding: make it SAE only.
8095 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8096 Ditto.
8097 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8098 Ditto.
8099 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8100 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8101 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8102 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8103 (round_saeonly_mask_scalar_operand4): Ditto.
8104 (round_saeonly_mask_scalar_op3): Ditto.
8105 (round_saeonly_mask_scalar_op4): Ditto.
8106
8107 2014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8108
8109 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8110 Implement -maltivec=be for vec_insert and vec_extract.
8111
8112 2014-01-10 DJ Delorie <dj@redhat.com>
8113
8114 * config/msp430/msp430.md (call_internal): Don't allow memory
8115 references with SP as the base register.
8116 (call_value_internal): Likewise.
8117 * config/msp430/constraints.md (Yc): New. For memory references
8118 that don't use SP as a base register.
8119
8120 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8121 "an integer without a # prefix"
8122 * config/msp430/msp430.md (epilogue_helper): Use it.
8123
8124 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8125
8126 PR target/59617
8127 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8128 AVX512F gather builtins.
8129 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8130 on gather decls with INTEGER_TYPE masktype.
8131 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8132 directly into the builtin rather than hoisting it before loop.
8133
8134 PR tree-optimization/59387
8135 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8136 (scev_const_prop): If folded_casts and type has undefined overflow,
8137 use force_gimple_operand instead of force_gimple_operand_gsi and
8138 for each added stmt if it is assign with
8139 arith_code_with_undefined_signed_overflow, call
8140 rewrite_to_defined_overflow.
8141 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8142 gimple-fold.h instead.
8143 (arith_code_with_undefined_signed_overflow,
8144 rewrite_to_defined_overflow): Moved to ...
8145 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8146 rewrite_to_defined_overflow): ... here. No longer static.
8147 Include gimplify-me.h.
8148 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8149 rewrite_to_defined_overflow): New prototypes.
8150
8151 2014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8152
8153 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8154
8155 2014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8156
8157 * builtins.c (get_object_alignment_2): Minor tweak.
8158 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8159
8160 2014-01-13 Christian Bruel <christian.bruel@st.com>
8161
8162 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
8163 optimized non constant lengths.
8164
8165 2014-01-13 Jakub Jelinek <jakub@redhat.com>
8166
8167 PR libgomp/59194
8168 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8169 load as __atomic_load_N if possible.
8170
8171 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8172
8173 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8174 target parameter.
8175 (rs6000_expand_builtin): Adjust call.
8176
8177 2014-01-11 David Edelsohn <dje.gcc@gmail.com>
8178
8179 PR target/58115
8180 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8181 * config/rs6000/rs6000.c: Include target-globals.h.
8182 (rs6000_set_current_function): Instead of doing target_reinit
8183 unconditionally, use save_target_globals_default_opts and
8184 restore_target_globals.
8185
8186 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8187 FPSCR.
8188 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8189 (rs6000_expand_builtin): Handle mffs and mtfsf.
8190 (rs6000_init_builtins): Define mffs and mtfsf.
8191 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8192 (rs6000_mffs): New pattern.
8193 (rs6000_mtfsf): New pattern.
8194
8195 2014-01-11 Bin Cheng <bin.cheng@arm.com>
8196
8197 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8198 Start narrowing with START. Apply candidate-use pair
8199 and check overall cost in narrowing.
8200 (iv_ca_prune): Pass new argument.
8201
8202 2014-01-10 Jeff Law <law@redhat.com>
8203
8204 PR middle-end/59743
8205 * ree.c (combine_reaching_defs): Ensure the defining statement
8206 occurs before the extension when optimizing extensions with
8207 different source and destination hard registers.
8208
8209 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8210
8211 PR ipa/58585
8212 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8213 vtables into the type inheritance graph.
8214
8215 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8216
8217 PR rtl-optimization/59754
8218 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8219 modes in the REGNO != REGNO case.
8220
8221 2014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8222
8223 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8224
8225 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8226
8227 PR tree-optimization/59745
8228 * tree-predcom.c (tree_predictive_commoning_loop): Call
8229 free_affine_expand_cache if giving up because components is NULL.
8230
8231 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8232 GC in payload of target_globals struct instead of allocating them on
8233 the heap and the larger structs separately using GC.
8234 * target-globals.h (struct target_globals): Make regs, hard_regs,
8235 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8236 of GTY((skip)) and change type to void *.
8237 (reset_target_globals): Cast loads from those fields to corresponding
8238 types.
8239
8240 2014-01-10 Steve Ellcey <sellcey@mips.com>
8241
8242 PR plugins/59335
8243 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8244 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8245 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8246
8247 2014-01-10 Richard Earnshaw <rearnsha@arm.com>
8248
8249 PR target/59744
8250 * aarch64-modes.def (CC_Zmode): New flags mode.
8251 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8252 represents an equality.
8253 (aarch64_get_condition_code): Handle CC_Zmode.
8254 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8255
8256 2014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8257
8258 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8259 extraction in good case.
8260
8261 2014-01-10 Richard Biener <rguenther@suse.de>
8262
8263 PR tree-optimization/59374
8264 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8265 checking after SLP discovery. Mark stmts not participating
8266 in any SLP instance properly.
8267
8268 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8269
8270 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8271 when handling a SET rtx.
8272
8273 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8274
8275 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8276 (cortex-a57): Likewise.
8277 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8278
8279 2014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8280
8281 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8282 non-iwmmxt builtins.
8283
8284 2014-01-10 Jan Hubicka <hubicka@ucw.cz>
8285
8286 PR ipa/58252
8287 PR ipa/59226
8288 * ipa-devirt.c record_target_from_binfo): Take as argument
8289 stack of binfos and lookup matching one for virtual inheritance.
8290 (possible_polymorphic_call_targets_1): Update.
8291
8292 2014-01-10 Huacai Chen <chenhc@lemote.com>
8293
8294 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8295 kernel strings for Loongson-2E/2F/3A.
8296
8297 2014-01-10 Jakub Jelinek <jakub@redhat.com>
8298
8299 PR middle-end/59670
8300 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8301 is_gimple_call before calling gimple_call_internal_p.
8302
8303 2014-01-09 Steve Ellcey <sellcey@mips.com>
8304
8305 * Makefile.in (TREE_FLOW_H): Remove.
8306 (TREE_SSA_H): Add file names from tree-flow.h.
8307 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8308 * tree.h: Remove tree-flow.h reference.
8309 * hash-table.h: Remove tree-flow.h reference.
8310 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8311 reference with tree-ssa-loop.h.
8312
8313 2014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8314
8315 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8316 default element-order behavior for -maltivec.
8317 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8318 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8319 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8320 when targeting big endian, at least for now.
8321 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8322
8323 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8324
8325 PR middle-end/47735
8326 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8327 var satisfies use_register_for_decl, just take into account type
8328 alignment, rather than decl alignment.
8329
8330 PR tree-optimization/59622
8331 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8332 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8333 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8334 Don't devirtualize for inplace at all. For targets.length () == 1,
8335 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8336
8337 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8338
8339 * config/i386/i386.md (cpu): Remove the unused btver1.
8340
8341 2014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8342
8343 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8344
8345 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8346
8347 PR target/58115
8348 * tree-core.h (struct target_globals): New forward declaration.
8349 (struct tree_target_option): Add globals field.
8350 * tree.h (TREE_TARGET_GLOBALS): Define.
8351 (prepare_target_option_nodes_for_pch): New prototype.
8352 * target-globals.h (struct target_globals): Define even if
8353 !SWITCHABLE_TARGET.
8354 * tree.c (prepare_target_option_node_for_pch,
8355 prepare_target_option_nodes_for_pch): New functions.
8356 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8357 * config/i386/i386.c: Include target-globals.h.
8358 (ix86_set_current_function): Instead of doing target_reinit
8359 unconditionally, use save_target_globals_default_opts and
8360 restore_target_globals.
8361
8362 2014-01-09 Richard Biener <rguenther@suse.de>
8363
8364 PR tree-optimization/59715
8365 * tree-cfg.h (split_critical_edges): Declare.
8366 * tree-cfg.c (split_critical_edges): Export.
8367 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8368
8369 2014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8370
8371 * cfgexpand.c (expand_stack_vars): Optionally disable
8372 asan stack protection.
8373 (expand_used_vars): Likewise.
8374 (partition_stack_vars): Likewise.
8375 * asan.c (asan_emit_stack_protection): Optionally disable
8376 after return stack usage.
8377 (instrument_derefs): Optionally disable memory access instrumentation.
8378 (instrument_builtin_call): Likewise.
8379 (instrument_strlen_call): Likewise.
8380 (asan_protect_global): Optionally disable global variables protection.
8381 * doc/invoke.texi: Added doc for new options.
8382 * params.def: Added new options.
8383 * params.h: Likewise.
8384
8385 2014-01-09 Jakub Jelinek <jakub@redhat.com>
8386
8387 PR rtl-optimization/59724
8388 * ifcvt.c (cond_exec_process_if_block): Don't call
8389 flow_find_head_matching_sequence with 0 longest_match.
8390 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8391 non-active insns if !stop_after.
8392 (try_head_merge_bb): Revert 2014-01-07 changes.
8393
8394 2014-01-08 Jeff Law <law@redhat.com>
8395
8396 * ree.c (get_sub_rtx): New function, extracted from...
8397 (merge_def_and_ext): Here.
8398 (combine_reaching_defs): Use get_sub_rtx.
8399
8400 2014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8401
8402 * cgraph.h (varpool_variable_node): Do not choke on null node.
8403
8404 2014-01-08 Catherine Moore <clm@codesourcery.com>
8405
8406 * config/mips/mips.md (simple_return): Attempt to use JRC
8407 for microMIPS.
8408 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8409
8410 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8411
8412 PR rtl-optimization/59137
8413 * reorg.c (steal_delay_list_from_target): Call update_block for
8414 elided insns.
8415 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8416
8417 2014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8418
8419 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8420 two duplicate entries.
8421
8422 2014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8423
8424 Revert:
8425 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8426
8427 * config/mips/mips.c (mips_truncated_op_cost): New function.
8428 (mips_rtx_costs): Adjust test for BADDU.
8429 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8430
8431 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8432
8433 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8434 (*baddu_si): ...this new pattern.
8435
8436 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8437
8438 PR ipa/59722
8439 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8440
8441 2014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8442
8443 PR middle-end/57748
8444 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8445 inner_reference_p.
8446 (expand_expr, expand_normal): Adjust.
8447 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8448 inner_reference_p. Use inner_reference_p to expand inner references.
8449 (store_expr): Adjust.
8450 * cfgexpand.c (expand_call_stmt): Adjust.
8451
8452 2014-01-08 Rong Xu <xur@google.com>
8453
8454 * gcov-io.c (gcov_var): Move from gcov-io.h.
8455 (gcov_position): Ditto.
8456 (gcov_is_error): Ditto.
8457 (gcov_rewrite): Ditto.
8458 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8459 only part to libgcc/libgcov.h.
8460
8461 2014-01-08 Marek Polacek <polacek@redhat.com>
8462
8463 PR middle-end/59669
8464 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8465
8466 2014-01-08 Marek Polacek <polacek@redhat.com>
8467
8468 PR sanitizer/59667
8469 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8470
8471 2014-01-08 Jakub Jelinek <jakub@redhat.com>
8472
8473 PR rtl-optimization/59649
8474 * stor-layout.c (get_mode_bounds): For BImode return
8475 0 and STORE_FLAG_VALUE.
8476
8477 2014-01-08 Richard Biener <rguenther@suse.de>
8478
8479 PR middle-end/59630
8480 * gimple.h (is_gimple_builtin_call): Remove.
8481 (gimple_builtin_call_types_compatible_p): New.
8482 (gimple_call_builtin_p): New overload.
8483 * gimple.c (is_gimple_builtin_call): Remove.
8484 (validate_call): Rename to ...
8485 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8486 check return types.
8487 (validate_type): New static function.
8488 (gimple_call_builtin_p): New overload and adjust.
8489 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8490 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8491 (gimple_fold_stmt_to_constant_1): Likewise.
8492 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8493
8494 2014-01-08 Richard Biener <rguenther@suse.de>
8495
8496 PR middle-end/59471
8497 * gimplify.c (gimplify_expr): Gimplify register-register type
8498 VIEW_CONVERT_EXPRs to separate stmts.
8499
8500 2014-01-07 Jeff Law <law@redhat.com>
8501
8502 PR middle-end/53623
8503 * ree.c (combine_set_extension): Handle case where source
8504 and destination registers in an extension insn are different.
8505 (combine_reaching_defs): Allow source and destination registers
8506 in extension to be different under limited circumstances.
8507 (add_removable_extension): Remove restriction that the
8508 source and destination registers in the extension are the same.
8509 (find_and_remove_re): Emit a copy from the extension's
8510 destination to its source after the defining insn if
8511 the source and destination registers are different.
8512
8513 PR middle-end/59285
8514 * ifcvt.c (merge_if_block): If we are merging a block with more than
8515 one successor with a block with no successors, remove any BARRIER
8516 after the second block.
8517
8518 2014-01-07 Dan Xio Qiang <ziyan01@163.com>
8519
8520 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8521
8522 2014-01-07 John David Anglin <danglin@gcc.gnu.org>
8523
8524 PR target/59652
8525 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8526 for 14-bit register offsets when INT14_OK_STRICT is false.
8527
8528 2014-01-07 Roland Stigge <stigge@antcom.de>
8529 Michael Meissner <meissner@linux.vnet.ibm.com>
8530
8531 PR 57386/target
8532 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
8533 Only check TFmode for SPE constants. Don't check TImode or TDmode.
8534
8535 2014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8536
8537 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8538 -mcpu.
8539
8540 2014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8541
8542 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8543 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8544 rtx is const0_rtx or not.
8545
8546 2014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8547
8548 PR target/58115
8549 * target-globals.c (save_target_globals): Remove this_fn_optab
8550 handling.
8551 * toplev.c: Include optabs.h.
8552 (target_reinit): Temporarily restore the global options if another
8553 set of options are in force.
8554
8555 2014-01-07 Jakub Jelinek <jakub@redhat.com>
8556
8557 PR rtl-optimization/58668
8558 * cfgcleanup.c (flow_find_cross_jump): Don't count
8559 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8560 to determine what is counted.
8561 (flow_find_head_matching_sequence): Use active_insn_p to determine
8562 what is counted.
8563 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8564 counting change.
8565 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8566 determine what is counted.
8567
8568 PR tree-optimization/59643
8569 * tree-predcom.c (split_data_refs_to_components): If one dr is
8570 read and one write, determine_offset fails and the write isn't
8571 in the bad component, just put the read into the bad component.
8572
8573 2014-01-07 Mike Stump <mikestump@comcast.net>
8574 Jakub Jelinek <jakub@redhat.com>
8575
8576 PR pch/59436
8577 * tree-core.h (struct tree_optimization_option): Change optabs
8578 type from unsigned char * to void *.
8579 * optabs.c (init_tree_optimization_optabs): Adjust
8580 TREE_OPTIMIZATION_OPTABS initialization.
8581
8582 2014-01-06 Jakub Jelinek <jakub@redhat.com>
8583
8584 PR target/59644
8585 * config/i386/i386.h (struct machine_function): Add
8586 no_drap_save_restore field.
8587 * config/i386/i386.c (ix86_save_reg): Use
8588 !cfun->machine->no_drap_save_restore instead of
8589 crtl->stack_realign_needed.
8590 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8591 this function clears frame_pointer_needed. Set
8592 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8593 and DRAP reg is needed.
8594
8595 2014-01-06 Marek Polacek <polacek@redhat.com>
8596
8597 PR c/57773
8598 * doc/implement-c.texi: Mention that other integer types are
8599 permitted as bit-field types in strictly conforming mode.
8600
8601 2014-01-06 Felix Yang <fei.yang0953@gmail.com>
8602
8603 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8604 is newly allocated.
8605
8606 2014-01-06 Richard Earnshaw <rearnsha@arm.com>
8607
8608 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8609
8610 2014-01-06 Martin Jambor <mjambor@suse.cz>
8611
8612 PR ipa/59008
8613 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8614 to int.
8615 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8616
8617 2014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8618
8619 PR debug/59350
8620 PR debug/59510
8621 * var-tracking.c (add_stores): Preserve the value of the source even if
8622 we don't record the store.
8623
8624 2014-01-06 Terry Guo <terry.guo@arm.com>
8625
8626 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8627
8628 2014-01-05 Iain Sandoe <iain@codesourcery.com>
8629
8630 PR bootstrap/59541
8631 * config/darwin.c (darwin_function_section): Adjust return values to
8632 correspond to optimisation changes made in r206070.
8633
8634 2014-01-05 Uros Bizjak <ubizjak@gmail.com>
8635
8636 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8637 from prefetch_block tune setting.
8638 (nocona_cost): Correct size of prefetch block to 64.
8639
8640 2014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8641
8642 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8643 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8644 used to save the static chain register in the computation of the offset
8645 from which the FP registers need to be restored.
8646
8647 2014-01-04 Jakub Jelinek <jakub@redhat.com>
8648
8649 PR tree-optimization/59519
8650 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8651 ICE if get_current_def (current_new_name) is already non-NULL, as long
8652 as it is a phi result of some other phi in *new_exit_bb that has
8653 the same argument.
8654
8655 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8656 or vmovdqu* for misaligned_operand.
8657 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8658 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8659 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8660 aligned_mem for AVX512F masked aligned load and store builtins and for
8661 non-temporal moves.
8662
8663 2014-01-03 Bingfeng Mei <bmei@broadcom.com>
8664
8665 PR tree-optimization/59651
8666 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
8667 Address range for negative step should be added by TYPE_SIZE_UNIT.
8668
8669 2014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8670
8671 * config/m68k/m68k.c (handle_move_double): Handle pushes with
8672 overlapping registers also for registers other than the stack pointer.
8673
8674 2014-01-03 Marek Polacek <polacek@redhat.com>
8675
8676 PR other/59661
8677 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8678 __builtin_FILE.
8679
8680 2014-01-03 Jakub Jelinek <jakub@redhat.com>
8681
8682 PR target/59625
8683 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8684 asm goto as jump.
8685
8686 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8687 (push splitter): Use <P:MODE_SIZE> instead of
8688 GET_MODE_SIZE (<P:MODE>mode).
8689 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8690 (mov -1, reg peephole2): Likewise.
8691 * config/i386/sse.md (*mov<mode>_internal,
8692 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8693 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8694 *<code><mode>3, *andnot<mode>3<mask_name>,
8695 <mask_codefor><code><mode>3<mask_name>): Likewise.
8696 * config/i386/subst.md (mask_mode512bit_condition,
8697 sd_mask_mode512bit_condition): Likewise.
8698
8699 2014-01-02 Xinliang David Li <davidxl@google.com>
8700
8701 PR tree-optimization/59303
8702 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
8703 (dump_predicates): Better output format.
8704 (pred_equal_p): New function.
8705 (is_neq_relop_p): Ditto.
8706 (is_neq_zero_form_p): Ditto.
8707 (pred_expr_equal_p): Ditto.
8708 (pred_neg_p): Ditto.
8709 (simplify_pred): Ditto.
8710 (simplify_preds_2): Ditto.
8711 (simplify_preds_3): Ditto.
8712 (simplify_preds_4): Ditto.
8713 (simplify_preds): Ditto.
8714 (push_pred): Ditto.
8715 (push_to_worklist): Ditto.
8716 (get_pred_info_from_cmp): Ditto.
8717 (is_degenerated_phi): Ditto.
8718 (normalize_one_pred_1): Ditto.
8719 (normalize_one_pred): Ditto.
8720 (normalize_one_pred_chain): Ditto.
8721 (normalize_preds): Ditto.
8722 (normalize_cond_1): Remove function.
8723 (normalize_cond): Ditto.
8724 (is_gcond_subset_of): Ditto.
8725 (is_subset_of_any): Ditto.
8726 (is_or_set_subset_of): Ditto.
8727 (is_and_set_subset_of): Ditto.
8728 (is_norm_cond_subset_of): Ditto.
8729 (pred_chain_length_cmp): Ditto.
8730 (convert_control_dep_chain_into_preds): Type change.
8731 (find_predicates): Ditto.
8732 (find_def_preds): Ditto.
8733 (destroy_predicates_vecs): Ditto.
8734 (find_matching_predicates_in_rest_chains): Ditto.
8735 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8736 (is_pred_expr_subset): Ditto.
8737 (is_pred_chain_subset_of): Ditto.
8738 (is_included_in): Ditto.
8739 (is_superset_of): Ditto.
8740
8741 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8742
8743 Update copyright years.
8744
8745 2014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8746
8747 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8748 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
8749 config/arc/arc.md, config/arc/arc.opt,
8750 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8751 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8752 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8753 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8754 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8755 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8756 vtable-verify.c, vtable-verify.h: Use the standard form for the
8757 copyright notice.
8758
8759 2014-01-02 Tobias Burnus <burnus@net-b.de>
8760
8761 * gcc.c (process_command): Update copyright notice dates.
8762 * gcov-dump.c: Ditto.
8763 * gcov.c: Ditto.
8764 * doc/cpp.texi: Bump @copying's copyright year.
8765 * doc/cppinternals.texi: Ditto.
8766 * doc/gcc.texi: Ditto.
8767 * doc/gccint.texi: Ditto.
8768 * doc/gcov.texi: Ditto.
8769 * doc/install.texi: Ditto.
8770 * doc/invoke.texi: Ditto.
8771
8772 2014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8773
8774 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8775
8776 2014-01-01 Jakub Jelinek <jakub@redhat.com>
8777
8778 * config/i386/sse.md (*mov<mode>_internal): Guard
8779 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8780
8781 PR rtl-optimization/59647
8782 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8783 new_rtx into UNSIGNED_FLOAT rtxes.
8784 \f
8785 Copyright (C) 2014 Free Software Foundation, Inc.
8786
8787 Copying and distribution of this file, with or without modification,
8788 are permitted in any medium without royalty provided the copyright
8789 notice and this notice are preserved.