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