]> gcc.gnu.org Git - gcc.git/blame - gcc/ChangeLog
* doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
[gcc.git] / gcc / ChangeLog
CommitLineData
fb90f9f3
MP
12014-05-02 Marek Polacek <polacek@redhat.com>
2
3 * doc/invoke.texi: Describe -fsanitize=float-divide-by-zero.
4
ed15c598
KC
52014-02-26 Kito Cheng <kito@0xlab.org>
6
7 * defaults.h (HONOR_REG_ALLOC_ORDER): Change HONOR_REG_ALLOC_ORDER
8 to a C expression marco.
9 * ira-color.c (HONOR_REG_ALLOC_ORDER) : Ditto.
10 * config/arm/arm.h (HONOR_REG_ALLOC_ORDER): Ditto.
11 * config/nds32/nds32.h (HONOR_REG_ALLOC_ORDER): Ditto.
12 * doc/tm.texi (HONOR_REG_ALLOC_ORDER): Update document for
13 HONOR_REG_ALLOC_ORDER.
14 * doc/tm.texi.in (HONOR_REG_ALLOC_ORDER): Ditto.
15
53c8d5a7
JBG
162014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
17
18 * config/arc/arc.c (TARGET_LRA_P): Undef before redefine.
19
486c559b
JBG
202014-05-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
21
22 * config/arc/arc.c (arc_select_cc_mode): Fix typo.
23
944052b9
YR
242014-05-01 Yuri Rumyantsev <ysrumyan@gmail.com>
25
26 * tree-if-conv.c (is_cond_scalar_reduction): New function.
27 (convert_scalar_cond_reduction): Likewise.
28 (predicate_scalar_phi): Add recognition and transformation
29 of simple conditioanl reduction to be vectorizable.
30
d00887e8
MP
312014-05-01 Marek Polacek <polacek@redhat.com>
32
33 PR c/43245
34 * doc/invoke.texi: Document -Wdiscarded-qualifiers.
35
7211512a
AL
362014-04-30 Alan Lawrence <alan.lawrence@arm.com>
37
38 * config/aarch64/arm_neon.h (vuzp1_f32, vuzp1_p8, vuzp1_p16, vuzp1_s8,
39 vuzp1_s16, vuzp1_s32, vuzp1_u8, vuzp1_u16, vuzp1_u32, vuzp1q_f32,
40 vuzp1q_f64, vuzp1q_p8, vuzp1q_p16, vuzp1q_s8, vuzp1q_s16, vuzp1q_s32,
41 vuzp1q_s64, vuzp1q_u8, vuzp1q_u16, vuzp1q_u32, vuzp1q_u64, vuzp2_f32,
42 vuzp2_p8, vuzp2_p16, vuzp2_s8, vuzp2_s16, vuzp2_s32, vuzp2_u8,
43 vuzp2_u16, vuzp2_u32, vuzp2q_f32, vuzp2q_f64, vuzp2q_p8, vuzp2q_p16,
44 vuzp2q_s8, vuzp2q_s16, vuzp2q_s32, vuzp2q_s64, vuzp2q_u8, vuzp2q_u16,
45 vuzp2q_u32, vuzp2q_u64): Replace temporary asm with __builtin_shuffle.
46
e5bd20a4
JR
472014-04-30 Joern Rennecke <joern.rennecke@embecosm.com>
48
49 * config/arc/arc.opt (mlra): Move comment above option name
50 to avoid mis-parsing as language options.
51
2cf03b11
RO
522014-04-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
53
54 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Move ...
55 * config/sol2.h: ... here.
56 * config/sol2-10.h: Remove.
57
58 * config/sol2-bi.h (WCHAR_TYPE, WCHAR_TYPE_SIZE, WINT_TYPE)
59 (WINT_TYPE_SIZE, MULTILIB_DEFAULTS, DEF_ARCH32_SPEC)
60 (DEF_ARCH64_SPEC, ASM_CPU_DEFAULT_SPEC, LINK_ARCH64_SPEC_BASE)
61 (LINK_ARCH64_SPEC, ARCH_DEFAULT_EMULATION, TARGET_LD_EMULATION)
62 (LINK_ARCH_SPEC, SUBTARGET_EXTRA_SPECS): Move ...
63 * config/sol2.h: ... here.
64 (SECTION_NAME_FORMAT): Don't redefine.
65 (STARTFILE_ARCH32_SPEC): Rename to ...
66 (STARTFILE_ARCH_SPEC): ... this.
67 (ASM_OUTPUT_ALIGNED_COMMON): Move ...
68 * config/sparc/sol2.h: ... here.
69 (SECTION_NAME_FORMAT): Don't undef.
70 * config/i386/sol2.h (ASM_CPU_DEFAULT_SPEC)
71 (SUBTARGET_EXTRA_SPECS): Remove.
72 * config/sparc/sol2.h (ASM_CPU_DEFAULT_SPEC): Remove.
73
74 * config/i386/sol2-bi.h (TARGET_SUBTARGET_DEFAULT)
75 (MD_STARTFILE_PREFIX): Remove.
76 (SUBTARGET_OPTIMIZATION_OPTIONS, ASM_CPU32_DEFAULT_SPEC)
77 (ASM_CPU64_DEFAULT_SPEC, ASM_CPU_SPEC, ASM_SPEC, DEFAULT_ARCH32_P)
78 (ARCH64_SUBDIR, ARCH32_EMULATION, ARCH64_EMULATION)
79 (ASM_COMMENT_START, JUMP_TABLES_IN_TEXT_SECTION)
80 (ASM_OUTPUT_DWARF_PCREL, ASM_OUTPUT_ALIGNED_COMMON)
81 (USE_IX86_FRAME_POINTER, USE_X86_64_FRAME_POINTER): Move ...
82 * config/i386/sol2.h: ... here.
83 (TARGET_SUBTARGET_DEFAULT, SIZE_TYPE, PTRDIFF_TYPE): Remove.
84 * config/i386/sol2-bi.h: Remove.
85 * config/sol2.h (MD_STARTFILE_PREFIX): Remove.
86 (LINK_ARCH32_SPEC_BASE): Remove /usr/ccs/lib/libp, /usr/ccs/lib.
87
88 * config/i386/t-sol2-64: Rename to ...
89 * config/i386/t-sol2: ... this.
90 * config/sparc/t-sol2-64: Rename to ...
91 * config/sparc/t-sol2: ... this.
92
93 * config.gcc (*-*-solaris2*): Split sol2_tm_file into
94 sol2_tm_file_head, sol2_tm_file_tail.
95 Include ${cpu_type}/sol2.h before sol2.h.
96 Remove sol2-10.h.
97 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Include
98 i386/x86-64.h between sol2_tm_file_head and sol2_tm_file_tail.
99 Remove i386/sol2-bi.h, sol2-bi.h from tm_file.
100 Reflect i386/t-sol2-64 renaming.
101 (sparc*-*-solaris2*): Remove sol2-bi.h from tm_file.
102 Reflect sparc/t-sol2-64 renaming.
103
e9ff9caf
RB
1042014-04-30 Richard Biener <rguenther@suse.de>
105
106 * passes.c (execute_function_todo): Move TODO_verify_stmts
107 and TODO_verify_ssa under the TODO_verify_il umbrella.
108 * tree-ssa.h (verify_ssa): Adjust prototype.
109 * tree-ssa.c (verify_ssa): Add parameter to tell whether
110 we should verify SSA operands.
111 * tree-cfg.h (verify_gimple_in_cfg): Adjust prototype.
112 * tree-cfg.c (verify_gimple_in_cfg): Add parameter to tell
113 whether we should verify whether not throwing stmts have EH info.
114 * graphite-scop-detection.c (create_sese_edges): Adjust.
115 * tree-ssa-loop-manip.c (verify_loop_closed_ssa): Likewise.
116 * tree-eh.c (lower_try_finally_switch): Do not add the
117 default case label twice.
118
f8ed5150
MP
1192014-04-30 Marek Polacek <polacek@redhat.com>
120
121 * gcc.c (sanitize_spec_function): Handle SANITIZE_FLOAT_DIVIDE.
122 * builtins.def: Initialize builtins even for SANITIZE_FLOAT_DIVIDE.
123 * flag-types.h (enum sanitize_code): Add SANITIZE_FLOAT_DIVIDE.
124 * opts.c (common_handle_option): Add -fsanitize=float-divide-by-zero.
125
7cb0403f
AL
1262014-04-29 Alan Lawrence <alan.lawrence@arm.com>
127
128 * config/aarch64/arm_neon.h (vzip1_f32, vzip1_p8, vzip1_p16, vzip1_s8,
129 vzip1_s16, vzip1_s32, vzip1_u8, vzip1_u16, vzip1_u32, vzip1q_f32,
130 vzip1q_f64, vzip1q_p8, vzip1q_p16, vzip1q_s8, vzip1q_s16, vzip1q_s32,
131 vzip1q_s64, vzip1q_u8, vzip1q_u16, vzip1q_u32, vzip1q_u64, vzip2_f32,
132 vzip2_p8, vzip2_p16, vzip2_s8, vzip2_s16, vzip2_s32, vzip2_u8,
133 vzip2_u16, vzip2_u32, vzip2q_f32, vzip2q_f64, vzip2q_p8, vzip2q_p16,
134 vzip2q_s8, vzip2q_s16, vzip2q_s32, vzip2q_s64, vzip2q_u8, vzip2q_u16,
135 vzip2q_u32, vzip2q_u64): Replace inline __asm__ with __builtin_shuffle.
136
3979649a
DM
1372014-04-29 David Malcolm <dmalcolm@redhat.com>
138
139 * tree-cfg.c (dump_function_to_file): Dump the return type of
140 functions, in a line to itself before the function body, mimicking
141 the layout of a C function.
142
62fb101e
JJ
1432014-04-29 Jakub Jelinek <jakub@redhat.com>
144
145 PR tree-optimization/60971
146 * tree-tailcall.c (process_assignment): Reject conversions which
147 reduce precision.
148
3d9684ae
JG
1492014-04-29 James Greenhalgh <james.greenhalgh@arm.com>
150
151 * calls.c (initialize_argument_information): Always treat
152 PUSH_ARGS_REVERSED as 1, simplify code accordingly.
153 (expand_call): Likewise.
154 (emit_library_call_calue_1): Likewise.
155 * expr.c (PUSH_ARGS_REVERSED): Do not define.
156 (emit_push_insn): Always treat PUSH_ARGS_REVERSED as 1, simplify
157 code accordingly.
158
d77f7b19
NC
1592014-04-29 Nick Clifton <nickc@redhat.com>
160
161 * config/msp430/msp430.md (umulsidi): Fix typo.
162 (mulhisi3): Enable even inside interrupt handlers.
163 * config/msp430/msp430.c (msp430_print_operand): %O: Allow for the
164 bigger return address pushed in large mode.
165
f8d91e80
NC
1662014-04-29 Nick Clifton <nickc@redhat.com>
167
168 * config/arc/arc.c (arc_select_cc_mode): Fix parentheses.
169 (arc_init_reg_tables): Use a machine_mode enum to iterate over
170 available modes.
171 * config/m32r/m32r.c (init_reg_tables): Likewise.
172 * config/m32c/m32c.c (m32c_illegal_subreg_p): Use a machine_mode
173 enum to hold the modes.
174
e3f613cb
RB
1752014-04-29 Richard Biener <rguenther@suse.de>
176
177 * dominance.c (free_dominance_info): Add overload with
178 function parameter.
179 (dom_info_state): Likewise.
180 (dom_info_available_p): Likewise.
181 * basic-block.h (free_dominance_info, dom_info_state,
182 dom_info_available_p): Declare overloads.
183 * passes.c (execute_function_todo): Verify that verifiers
184 don't change dominator info state. Drop dominator info
185 for IPA pass invocations.
186 * cgraph.c (release_function_body): Restore asserts that
187 dominator information is released.
188
36f291f7
PP
1892014-04-29 Patrick Palka <patrick@parcs.ath.cx>
190
191 * doc/invoke.texi: Fix typo.
192 * tree-vrp.c: Fix typos.
193 * gimple.c (infer_nonnull_range): Reorder operands of an &&
194 condition.
195
d3a3182b
ZC
1962014-04-29 Zhenqiang Chen <zhenqiang.chen@linaro.org>
197
198 * config/aarch64/aarch64.md (mov<mode>cc): New for GPF.
199
ba081b77
JG
2002014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
201
202 * config/aarch64/aarch64-builtins.c
203 (aarch64_types_storestruct_lane_qualifiers): New.
204 (TYPES_STORESTRUCT_LANE): Likewise.
205 * config/aarch64/aarch64-simd-builtins.def (st2_lane): New.
206 (st3_lane): Likewise.
207 (st4_lane): Likewise.
208 * config/aarch64/aarch64-simd.md (vec_store_lanesoi_lane<mode>): New.
209 (vec_store_lanesci_lane<mode>): Likewise.
210 (vec_store_lanesxi_lane<mode>): Likewise.
211 (aarch64_st2_lane<VQ:mode>): Likewise.
212 (aarch64_st3_lane<VQ:mode>): Likewise.
213 (aarch64_st4_lane<VQ:mode>): Likewise.
214 * config/aarch64/aarch64.md (unspec): Add UNSPEC_ST{2,3,4}_LANE.
215 * config/aarch64/arm_neon.h
216 (__ST2_LANE_FUNC): Rewrite using builtins, update use points to
217 use new macro arguments.
218 (__ST3_LANE_FUNC): Likewise.
219 (__ST4_LANE_FUNC): Likewise.
220 * config/aarch64/iterators.md (V_TWO_ELEM): New.
221 (V_THREE_ELEM): Likewise.
222 (V_FOUR_ELEM): Likewise.
223
1bc6d43c
DM
2242014-04-28 David Malcolm <dmalcolm@redhat.com>
225
226 * doc/gimple.texi: Replace the description of the now-defunct
227 union gimple_statement_d with a diagram showing the
228 gimple_statement_base class hierarchy and its relationships to
229 the GSS_ and GIMPLE_ enums.
230
97e1ad78
JG
2312014-04-28 James Greenhalgh <james.greenhalgh@arm.com>
232
233 * config/aarch64/aarch64-protos.h (aarch64_modes_tieable_p): New.
234 * config/aarch64/aarch64.c
235 (aarch64_cannot_change_mode_class): Weaken conditions.
236 (aarch64_modes_tieable_p): New.
237 * config/aarch64/aarch64.h (MODES_TIEABLE_P): Use it.
238
bf245bf4
PH
2392014-04-28 Pat Haugen <pthaugen@us.ibm.com>
240
241 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
242 (loadsync_<mode>): Change mode.
243 (load_quadpti, store_quadpti): New.
244 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
245 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
246
28151221
MJ
2472014-04-28 Martin Jambor <mjambor@suse.cz>
248
249 * tree-sra.c (sra_modify_expr): Generate new memory accesses with
250 same alias type as the original statement.
251 (subreplacement_assignment_data): New type.
252 (handle_unscalarized_data_in_subtree): New type of parameter,
253 generate new memory accesses with same alias type as the original
254 statement.
255 (load_assign_lhs_subreplacements): Likewise.
256 (sra_modify_constructor_assign): Generate new memory accesses with
257 same alias type as the original statement.
258
9ba5fb43
RB
2592014-04-28 Richard Biener <rguenther@suse.de>
260
261 * tree-pass.h (TODO_verify_il): Define.
262 (TODO_verify_all): Complete properly.
263 * passes.c (execute_function_todo): Move existing loop-closed
264 SSA verification under TODO_verify_il.
265 (execute_one_pass): Trigger TODO_verify_il at todo-after time.
266 * graphite-sese-to-poly.c (rewrite_cross_bb_scalar_deps):
267 Fix tree sharing issue.
268
22869a37
RB
2692014-04-28 Richard Biener <rguenther@suse.de>
270
271 PR middle-end/60092
272 * builtins.def (DEF_C11_BUILTIN): Add.
273 (BUILT_IN_ALIGNED_ALLOC): Likewise.
274 * coretypes.h (enum function_class): Add function_c11_misc.
275 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Handle
276 BUILT_IN_ALIGNED_ALLOC like BUILT_IN_MALLOC.
277 (call_may_clobber_ref_p_1): Likewise.
278 * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Likewise.
279 (mark_all_reaching_defs_necessary_1): Likewise.
280 (propagate_necessity): Likewise.
281 (eliminate_unnecessary_stmts): Likewise.
282 * tree-ssa-ccp.c (evaluate_stmt): Handle BUILT_IN_ALIGNED_ALLOC.
283
771c9501
RB
2842014-04-28 Richard Biener <rguenther@suse.de>
285
286 * tree-vrp.c (vrp_var_may_overflow): Remove.
287 (vrp_visit_phi_node): Rather than bumping to +-INF possibly
288 with overflow immediately bump to one before that value and
289 let iteration figure out overflow status.
290
279a935f
RB
2912014-04-28 Richard Biener <rguenther@suse.de>
292
293 * configure.ac: Do valgrind header checks unconditionally.
294 Add --enable-valgrind-annotations.
295 * system.h: Guard valgrind header inclusion with
296 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING.
297 * alloc-pool.c (pool_alloc, pool_free): Use
298 ENABLE_VALGRIND_ANNOTATIONS instead of ENABLE_VALGRIND_CHECKING
299 to guard possibly dead code.
300 * config.in: Regenerated.
301 * configure: Likewise.
302
07357f2f
JL
3032014-04-28 Jeff Law <law@redhat.com>
304
305 PR tree-optimization/60902
306 * tree-ssa-threadedge.c
307 (record_temporary_equivalences_from_stmts_at_dest): Only iterate
308 over real defs when invalidating outputs from statements that do not
309 produce useful outputs for threading.
310
6c6aa8e6
RB
3112014-04-28 Richard Biener <rguenther@suse.de>
312
313 PR tree-optimization/60979
314 * graphite-scop-detection.c (scopdet_basic_block_info): Reject
315 SCOPs that end in a block with a successor with abnormal
316 predecessors.
317
2cbf2d95
RB
3182014-04-28 Richard Biener <rguenther@suse.de>
319
320 * tree-pass.h (execute_pass_list): Adjust prototype.
321 * passes.c (pass_manager::execute_early_local_passes):
322 Adjust.
323 (do_per_function): Change callback signature, push all actual
324 work to the callbals.
325 (do_per_function_toporder): Likewise.
326 (execute_function_dump): Adjust.
327 (execute_function_todo): Likewise.
328 (clear_last_verified): Likewise.
329 (verify_curr_properties): Likewise.
330 (update_properties_after_pass): Likewise.
331 (execute_pass_list_1): Split out from ...
332 (execute_pass_list): ... here. Adjust.
333 (execute_ipa_pass_list): Likewise.
334 * cgraphunit.c (cgraph_add_new_function): Adjust.
335 (analyze_function): Likewise.
336 (expand_function): Likewise.
337 * cgraph.c (release_function_body): Free dominance info
338 here instead of asserting it was magically freed elsewhere.
339
ba21a04a
EB
3402014-04-28 Eric Botcazou <ebotcazou@adacore.com>
341
342 * configure.ac: Tweak GAS check for LEON instructions on SPARC.
343 * configure: Regenerate.
344 * config/sparc/sparc.opt (muser-mode): New option.
345 * config/sparc/sync.md (atomic_compare_and_swap<mode>_1): Do not enable
346 for LEON3.
347 (atomic_compare_and_swap_leon3_1): New instruction for LEON3.
348 * doc/invoke.texi (SPARC options): Document -muser-mode.
349
f956adb9
RS
3502014-04-27 Richard Sandiford <rdsandiford@googlemail.com>
351
352 * cselib.c (find_slot_memmode): Delete.
353 (cselib_hasher): Change compare_type to a struct.
354 (cselib_hasher::equal): Update accordingly. Don't expect wrapped
355 constants.
356 (preserve_constants_and_equivs): Adjust for new compare_type.
357 (cselib_find_slot): Likewise. Take the mode of the rtx as argument.
358 (wrap_constant): Delete.
359 (cselib_lookup_mem, cselib_lookup_1): Update calls to cselib_find_slot.
360
167c3e96
MT
3612014-04-26 Markus Trippelsdorf <markus@trippelsdorf.de>
362
363 * doc/install.texi (Building with profile feedback): Remove
364 outdated sentence.
365
4148ef5f
TV
3662014-04-26 Tom de Vries <tom@codesourcery.com>
367
368 * config/i386/i386.md (define_expand "ldexpxf3"): Fix out-of-bounds
369 array accesses.
370
8d8ffd86
CC
3712014-04-25 Cary Coutant <ccoutant@google.com>
372
373 PR debug/60929
374 * dwarf2out.c (should_move_die_to_comdat): A type definition
375 can contain a subprogram definition, but don't move it to a
376 comdat unit.
377 (clone_as_declaration): Copy DW_AT_abstract_origin attribute.
378 (generate_skeleton_bottom_up): Remove DW_AT_object_pointer attribute
379 from original DIE.
380 (clone_tree_hash): Rename to...
381 (clone_tree_partial): ...this; change callers. Copy
382 DW_TAG_subprogram DIEs as declarations.
383 (copy_decls_walk): Don't copy children of a declaration into a
384 type unit.
385
255671b0
L
3862014-04-25 H.J. Lu <hongjiu.lu@intel.com>
387
388 PR target/60969
389 * config/i386/i386.md (*movsf_internal): Set MODE to SI for
390 alternative 12.
391
10432733
JW
3922014-04-25 Jiong Wang <jiong.wang@arm.com>
393
394 * config/arm/predicates.md (call_insn_operand): Add long_call check.
395 * config/arm/arm.md (sibcall, sibcall_value): Force the address to
396 reg for long_call.
397 * config/arm/arm.c (arm_function_ok_for_sibcall): Remove long_call
398 restriction.
399
0e4c913f
KT
4002014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
401
402 * config/arm/arm.c (arm_cortex_a8_tune): Initialise
403 T16-related fields.
404
61ba7329
BS
4052014-04-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
406
407 PR tree-optimization/60930
408 * gimple-ssa-strength-reduction.c (create_mul_imm_cand): Reject
409 creating a multiply candidate by folding two constant
410 multiplicands when the result overflows.
411
2b332829
JJ
4122014-04-25 Jakub Jelinek <jakub@redhat.com>
413
414 PR tree-optimization/60960
415 * tree-vect-generic.c (expand_vector_operation): Only call
416 expand_vector_divmod if type's mode satisfies VECTOR_MODE_P.
417
97891c11
TV
4182014-04-25 Tom de Vries <tom@codesourcery.com>
419
420 * expr.c (clobber_reg_mode): New function.
421 * expr.h (clobber_reg): New function.
422
3ee634fd
TV
4232014-04-25 Tom de Vries <tom@codesourcery.com>
424
425 * rtlanal.c (find_all_hard_reg_sets): Note INSN_CALL_FUNCTION_USAGE
426 clobbers.
427
356bf593
RO
4282014-04-25 Radovan Obradovic <robradovic@mips.com>
429 Tom de Vries <tom@codesourcery.com>
430
431 * rtlanal.c (find_all_hard_reg_sets): Add bool implicit parameter and
432 handle.
433 * rtl.h (find_all_hard_reg_sets): Add bool parameter.
434 * haifa-sched.c (recompute_todo_spec, check_clobbered_conditions): Add
435 new argument to find_all_hard_reg_sets call.
436
fe65a7ed
KT
4372014-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
438
439 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p):
440 Use HOST_WIDE_INT_C for mask literal.
441 (aarch_rev16_shleft_mask_imm_p): Likewise.
442
b357d40d
EB
4432014-04-25 Eric Botcazou <ebotcazou@adacore.com>
444
445 PR target/60941
446 * config/sparc/sparc.md (ashlsi3_extend): Delete.
447
30c0a59a
MG
4482014-04-25 Marc Glisse <marc.glisse@inria.fr>
449
450 PR preprocessor/56540
451 * config/i386/i386-c.c (ix86_target_macros): Define
452 __SIZEOF_FLOAT80__ and __SIZEOF_FLOAT128__.
453
c02f59e2
RO
4542014-04-25 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
455
456 * configure.ac (tga_func): Remove.
457 (LIB_TLS_SPEC): Remove.
458 * configure: Regenerate.
459 * config.in: Regenerate.
460 * config/sol2.h (LIB_SPEC): Don't use LIB_TLS_SPEC.
461
92b3326b
RB
4622014-04-25 Richard Biener <rguenther@suse.de>
463
464 PR ipa/60912
465 * tree-ssa-structalias.c (ipa_pta_execute): Compute direct
466 call stmt use/clobber sets during stmt walk instead of
467 walking the possibly incomplete set of caller edges.
468
78422fb1
RB
4692014-04-25 Richard Biener <rguenther@suse.de>
470
471 PR ipa/60911
472 * passes.c (apply_ipa_transforms): Inline into only caller ...
473 (execute_one_pass): ... here. Properly bring in function
474 bodies for nodes we want to apply IPA transforms to.
475
56f8faae
CH
4762014-04-24 Cong Hou <congh@google.com>
477
478 PR tree-optimization/60896
479 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Pick up
480 all statements in PATTERN_DEF_SEQ in recognized widen-mult pattern.
481 (vect_mark_pattern_stmts): Set the def type of all statements in
482 PATTERN_DEF_SEQ as vect_internal_def.
483
06b39289
MM
4842014-04-24 Michael Meissner <meissner@linux.vnet.ibm.com>
485
486 * doc/extend.texi (PowerPC Built-in Functions): Document new
487 powerpc extended divide, bcd, pack/unpack 128-bit, builtin
488 functions.
489 (PowerPC AltiVec/VSX Built-in Functions): Likewise.
490
491 * config/rs6000/predicates.md (const_0_to_3_operand): New
492 predicate to match 0..3 integer constants.
493
494 * config/rs6000/rs6000-builtin.def (BU_DFP_MISC_1): Add new macros
495 to support adding miscellaneous builtin functions.
496 (BU_DFP_MISC_2): Likewise.
497 (BU_P7_MISC_1): Likewise.
498 (BU_P7_MISC_2): Likewise.
499 (BU_P8V_MISC_3): Likewise.
500 (BU_MISC_1): Likewise.
501 (BU_MISC_2): Likewise.
502 (DIVWE): Add extended divide builtin functions.
503 (DIVWEO): Likewise.
504 (DIVWEU): Likewise.
505 (DIVWEUO): Likewise.
506 (DIVDE): Likewise.
507 (DIVDEO): Likewise.
508 (DIVDEU): Likewise.
509 (DIVDEUO): Likewise.
510 (DXEX): Add decimal floating-point builtin functions.
511 (DXEXQ): Likewise.
512 (DDEDPD): Likewise.
513 (DDEDPDQ): Likewise.
514 (DENBCD): Likewise.
515 (DENBCDQ): Likewise.
516 (DIEX): Likewise.
517 (DIEXQ): Likewise.
518 (DSCLI): Likewise.
519 (DSCLIQ): Likewise.
520 (DSCRI): Likewise.
521 (DSCRIQ): Likewise.
522 (CDTBCD): Add new BCD builtin functions.
523 (CBCDTD): Likewise.
524 (ADDG6S): Likewise.
525 (BCDADD): Likewise.
526 (BCDADD_LT): Likewise.
527 (BCDADD_EQ): Likewise.
528 (BCDADD_GT): Likewise.
529 (BCDADD_OV): Likewise.
530 (BCDSUB): Likewise.
531 (BCDSUB_LT): Likewise.
532 (BCDSUB_EQ): Likewise.
533 (BCDSUB_GT): Likewise.
534 (BCDSUB_OV): Likewise.
535 (PACK_TD): Add new pack/unpack 128-bit type builtin functions.
536 (UNPACK_TD): Likewise.
537 (PACK_TF): Likewise.
538 (UNPACK_TF): Likewise.
539 (UNPACK_TF_0): Likewise.
540 (UNPACK_TF_1): Likewise.
541 (PACK_V1TI): Likewise.
542 (UNPACK_V1TI): Likewise.
543
544 * config/rs6000/rs6000.c (rs6000_builtin_mask_calculate): Add
545 support for decimal floating point builtin functions.
546 (rs6000_expand_ternop_builtin): Add checks for the new builtin
547 functions that take constant arguments.
548 (rs6000_invalid_builtin): Add decimal floating point builtin
549 support.
550 (rs6000_init_builtins): Setup long double, _Decimal64, and
551 _Decimal128 types for new builtin functions.
552 (builtin_function_type): Set the unsigned flags appropriately for
553 the new builtin functions.
554 (rs6000_opt_masks): Add support for decimal floating point builtin
555 functions.
556
557 * config/rs6000/rs6000.h (RS6000_BTM_DFP): Add support for decimal
558 floating point builtin functions.
559 (RS6000_BTM_COMMON): Likewise.
560 (RS6000_BTI_long_double): Likewise.
561 (RS6000_BTI_dfloat64): Likewise.
562 (RS6000_BTI_dfloat128): Likewise.
563 (long_double_type_internal_node): Likewise.
564 (dfloat64_type_internal_node): Likewise.
565 (dfloat128_type_internal_node): Likewise.
566
567 * config/rs6000/altivec.h (UNSPEC_BCDADD): Add support for ISA
568 2.07 bcd arithmetic instructions.
569 (UNSPEC_BCDSUB): Likewise.
570 (UNSPEC_BCD_OVERFLOW): Likewise.
571 (UNSPEC_BCD_ADD_SUB): Likewise.
572 (bcd_add_sub): Likewise.
573 (BCD_TEST): Likewise.
574 (bcd<bcd_add_sub>): Likewise.
575 (bcd<bcd_add_sub>_test): Likewise.
576 (bcd<bcd_add_sub>_test2): Likewise.
577 (bcd<bcd_add_sub>_<code>): Likewise.
578 (peephole2 for combined bcd ops): Likewise.
579
580 * config/rs6000/dfp.md (UNSPEC_DDEDPD): Add support for new
581 decimal floating point builtin functions.
582 (UNSPEC_DENBCD): Likewise.
583 (UNSPEC_DXEX): Likewise.
584 (UNSPEC_DIEX): Likewise.
585 (UNSPEC_DSCLI): Likewise.
586 (UNSPEC_DSCRI): Likewise.
587 (D64_D128): Likewise.
588 (dfp_suffix): Likewise.
589 (dfp_ddedpd_<mode>): Likewise.
590 (dfp_denbcd_<mode>): Likewise.
591 (dfp_dxex_<mode>): Likewise.
592 (dfp_diex_<mode>): Likewise.
593 (dfp_dscli_<mode>): Likewise.
594 (dfp_dscri_<mode>): Likewise.
595
596 * config/rs6000/rs6000.md (UNSPEC_ADDG6S): Add support for new BCD
597 builtin functions.
598 (UNSPEC_CDTBCD): Likewise.
599 (UNSPEC_CBCDTD): Likewise.
600 (UNSPEC_DIVE): Add support for new extended divide builtin
601 functions.
602 (UNSPEC_DIVEO): Likewise.
603 (UNSPEC_DIVEU): Likewise.
604 (UNSPEC_DIVEUO): Likewise.
605 (UNSPEC_UNPACK_128BIT): Add support for new builtin functions to
606 pack/unpack 128-bit types.
607 (UNSPEC_PACK_128BIT): Likewise.
608 (idiv_ldiv): New mode attribute to set the 32/64-bit divide type.
609 (udiv<mode>3): Use idiv_ldiv mode attribute.
610 (div<mode>3): Likewise.
611 (addg6s): Add new BCD builtin functions.
612 (cdtbcd): Likewise.
613 (cbcdtd): Likewise.
614 (UNSPEC_DIV_EXTEND): Add support for new extended divide
615 instructions.
616 (div_extend): Likewise.
617 (div<div_extend>_<mode>"): Likewise.
618 (FP128_64): Add support for new builtin functions to pack/unpack
619 128-bit types.
620 (unpack<mode>): Likewise.
621 (unpacktf_0): Likewise.
622 (unpacktf_1): Likewise.
623 (unpack<mode>_dm): Likewise.
624 (unpack<mode>_nodm): Likewise.
625 (pack<mode>): Likewise.
626 (unpackv1ti): Likewise.
627 (packv1ti): Likewise.
628
6e4f81db
V
6292014-04-24 Vishnu K S <Vishnu.k_s@atmel.com>
630
631 * gcc/config/avr/avr.c: Add comment on why -fdelete-null-pointer-checks
632 is disabled.
633
f7468577
JJ
6342014-04-24 Jakub Jelinek <jakub@redhat.com>
635
636 * tree.h (OMP_CLAUSE_LINEAR_GIMPLE_SEQ): Define.
637 * gimplify.c (omp_is_private): Change last argument's type to int.
638 Only diagnose lastprivate if the simd argument is 1, only diagnose
639 linear if the simd argument is 2.
640 (gimplify_omp_for): Adjust omp_is_private callers. When adding
641 lastprivate or private, add the clause to OMP_FOR_CLAUSES. Pass
642 GOVD_EXPLICIT to omp_add_variable. For simd with collapse == 1
643 create OMP_CLAUSE_LINEAR rather than OMP_CLAUSE_PRIVATE for var.
644 If var != decl and decl is in OMP_CLAUSE_LINEAR, gimplify decl
645 increment to OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
646 * omp-low.c (scan_sharing_clauses, lower_lastprivate_clauses): Handle
647 OMP_CLAUSE_LINEAR_GIMPLE_SEQ.
648 * tree-nested.c (convert_nonlocal_omp_clauses,
649 convert_local_omp_clauses): Handle OMP_CLAUSE_LINEAR.
650
2f7ac5ce
JL
6512014-04-24 Segher Boessenkool <segher@kernel.crashing.org>
652
653 PR target/60822
654 * config/m68k/m68k.md (extendplussidi): Don't allow memory for
655 operand 1.
656
dcaaa5a0
DP
6572014-04-24 Dimitris Papavasiliou <dpapavas@gmail.com>
658
659 * flag-types.h (enum ivar_visibility): Add.
660
78c7d18e
TS
6612014-04-24 Trevor Saunders <tsaunders@mozilla.com>
662
663 * config/sh/sh_treg_combine.c (sh_treg_combine::execute): Take
664 function * argument.
665
fbbde65e
AL
6662014-04-24 Alan Lawrence <alan.lawrence@arm.com>
667
668 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Enable for bigendian.
669
4f660b15
RO
6702014-04-24 Radovan Obradovic <robradovic@mips.com>
671 Tom de Vries <tom@codesourcery.com>
672
673 * reg-notes.def (REG_NOTE (CALL_DECL)): New reg-note REG_CALL_DECL.
674 * calls.c (expand_call, emit_library_call_value_1): Add REG_CALL_DECL
675 reg-note.
676 * combine.c (distribute_notes): Handle REG_CALL_DECL reg-note.
677 * emit-rtl.c (try_split): Same.
678
d996e61a
RO
6792014-04-24 Radovan Obradovic <robradovic@mips.com>
680 Tom de Vries <tom@codesourcery.com>
681
682 * common.opt (fuse-caller-save): New option.
683
bbcc9c00
TB
6842014-04-24 Tejas Belagod <tejas.belagod@arm.com>
685
686 * config/aarch64/aarch64.c (aarch64_evpc_tbl): Reverse order of elements
687 for big-endian.
688
d2d11532
RB
6892014-04-24 Richard Biener <rguenther@suse.de>
690
691 * expr.c (expand_expr_real_1): Avoid gimple_assign_rhs_to_tree
692 during TER and instead use the sepops interface for expanding
693 non-GIMPLE_SINGLE_RHS.
694
8117ef98
RO
6952014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
696
697 * config/i386/sol2.h (ASM_PREFERRED_EH_DATA_FORMAT): Only redefine
698 if not HAVE_AS_IX86_DIFF_SECT_DELTA.
699
ab0afae3
RO
7002014-04-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
701
702 * configure.ac (gcc_cv_as_cfi_directive): Support Solaris/x86
703 assembler 64-bit option.
704 * configure: Regenerate.
705
683e3333
KT
7062014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
707
708 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Check
709 TARGET_SIMD rather than TARGET_GENERAL_REGS_ONLY.
710 (TARGET_SIMD): Take AARCH64_ISA_SIMD into account.
711 (TARGET_FLOAT): Take AARCH64_ISA_FP into account.
712 (TARGET_CRYPTO): Take TARGET_SIMD into account.
713
c7f28cd5
KT
7142014-04-24 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
715
716 * config/aarch64/aarch64-builtins.c
717 (aarch64_builtin_vectorized_function): Handle BUILT_IN_BSWAP16,
718 BUILT_IN_BSWAP32, BUILT_IN_BSWAP64.
719 * config/aarch64/aarch64-simd.md (bswap<mode>): New pattern.
720 * config/aarch64/aarch64-simd-builtins.def: Define vector bswap
721 builtins.
722 * config/aarch64/iterator.md (VDQHSD): New mode iterator.
723 (Vrevsuff): New mode attribute.
724
b0419491
TG
7252014-04-24 Terry Guo <terry.guo@arm.com>
726
727 * config/arm/arm.h (machine_function): Define variable
728 after_arm_reorg here.
729 * config/arm/arm.c (after_arm_reorg): Remove the definition.
730 (arm_split_constant): Update the way to access variable
731 after_arm_reorg.
732 (arm_reorg): Ditto.
733 (arm_output_function_epilogue): Remove the reset of after_arm_reorg.
734
74e4d71a
TV
7352014-04-23 Tom de Vries <tom@codesourcery.com>
736
737 * target-hooks-macros.h: Fix DEFHOOKPOD argument order in comment.
738
7de90a6c
DM
7392014-04-23 David Malcolm <dmalcolm@redhat.com>
740
741 * is-a.h: Update comments to reflect the following changes to the
742 "pointerness" of the API, making the template parameter match the
743 return type, allowing use of is-a.h with typedefs of pointers.
744 (is_a_helper::cast): Return a T rather then a pointer to a T, so
745 that the return type matches the parameter to the is_a_helper.
746 (as_a): Likewise.
747 (dyn_cast): Likewise.
748
749 * cgraph.c (cgraph_node_for_asm): Update for removal of implicit
750 pointer from the is-a.h API.
751
752 * cgraph.h (is_a_helper <cgraph_node>::test): Convert to...
753 (is_a_helper <cgraph_node *>::test): ...this, matching change to
754 is-a.h API.
755 (is_a_helper <varpool_node>::test): Likewise, convert to...
756 (is_a_helper <varpool_node *>::test): ...this.
757
758 (varpool_first_variable): Update for removal of implicit pointer
759 from the is-a.h API.
760 (varpool_next_variable): Likewise.
761 (varpool_first_static_initializer): Likewise.
762 (varpool_next_static_initializer): Likewise.
763 (varpool_first_defined_variable): Likewise.
764 (varpool_next_defined_variable): Likewise.
765 (cgraph_first_defined_function): Likewise.
766 (cgraph_next_defined_function): Likewise.
767 (cgraph_first_function): Likewise.
768 (cgraph_next_function): Likewise.
769 (cgraph_first_function_with_gimple_body): Likewise.
770 (cgraph_next_function_with_gimple_body): Likewise.
771 (cgraph_alias_target): Likewise.
772 (varpool_alias_target): Likewise.
773 (cgraph_function_or_thunk_node): Likewise.
774 (varpool_variable_node): Likewise.
775 (symtab_real_symbol_p): Likewise.
776 * cgraphunit.c (referred_to_p): Likewise.
777 (analyze_functions): Likewise.
778 (handle_alias_pairs): Likewise.
779 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
780 * gimple-ssa.h (gimple_vuse_op): Likewise.
781 (gimple_vdef_op): Likewise.
782 * gimple-streamer-in.c (input_gimple_stmt): Likewise.
783 * gimple.c (gimple_build_asm_1): Likewise.
784 (gimple_build_try): Likewise.
785 (gimple_build_resx): Likewise.
786 (gimple_build_eh_dispatch): Likewise.
787 (gimple_build_omp_for): Likewise.
788 (gimple_omp_for_set_clauses): Likewise.
789
790 * gimple.h (is_a_helper <gimple_statement_asm>::test): Convert to...
791 (is_a_helper <gimple_statement_asm *>::test): ...this.
792 (is_a_helper <gimple_statement_bind>::test): Convert to...
793 (is_a_helper <gimple_statement_bind *>::test): ...this.
794 (is_a_helper <gimple_statement_call>::test): Convert to...
795 (is_a_helper <gimple_statement_call *>::test): ...this.
796 (is_a_helper <gimple_statement_catch>::test): Convert to...
797 (is_a_helper <gimple_statement_catch *>::test): ...this.
798 (is_a_helper <gimple_statement_resx>::test): Convert to...
799 (is_a_helper <gimple_statement_resx *>::test): ...this.
800 (is_a_helper <gimple_statement_eh_dispatch>::test): Convert to...
801 (is_a_helper <gimple_statement_eh_dispatch *>::test): ...this.
802 (is_a_helper <gimple_statement_eh_else>::test): Convert to...
803 (is_a_helper <gimple_statement_eh_else *>::test): ...this.
804 (is_a_helper <gimple_statement_eh_filter>::test): Convert to...
805 (is_a_helper <gimple_statement_eh_filter *>::test): ...this.
806 (is_a_helper <gimple_statement_eh_mnt>::test): Convert to...
807 (is_a_helper <gimple_statement_eh_mnt *>::test): ...this.
808 (is_a_helper <gimple_statement_omp_atomic_load>::test): Convert to...
809 (is_a_helper <gimple_statement_omp_atomic_load *>::test): ...this.
810 (is_a_helper <gimple_statement_omp_atomic_store>::test): Convert to...
811 (is_a_helper <gimple_statement_omp_atomic_store *>::test): ...this.
812 (is_a_helper <gimple_statement_omp_return>::test): Convert to...
813 (is_a_helper <gimple_statement_omp_return *>::test): ...this.
814 (is_a_helper <gimple_statement_omp_continue>::test): Convert to...
815 (is_a_helper <gimple_statement_omp_continue *>::test): ...this.
816 (is_a_helper <gimple_statement_omp_critical>::test): Convert to...
817 (is_a_helper <gimple_statement_omp_critical *>::test): ...this.
818 (is_a_helper <gimple_statement_omp_for>::test): Convert to...
819 (is_a_helper <gimple_statement_omp_for *>::test): ...this.
820 (is_a_helper <gimple_statement_omp_taskreg>::test): Convert to...
821 (is_a_helper <gimple_statement_omp_taskreg *>::test): ...this.
822 (is_a_helper <gimple_statement_omp_parallel>::test): Convert to...
823 (is_a_helper <gimple_statement_omp_parallel *>::test): ...this.
824 (is_a_helper <gimple_statement_omp_target>::test): Convert to...
825 (is_a_helper <gimple_statement_omp_target *>::test): ...this.
826 (is_a_helper <gimple_statement_omp_sections>::test): Convert to...
827 (is_a_helper <gimple_statement_omp_sections *>::test): ...this.
828 (is_a_helper <gimple_statement_omp_single>::test): Convert to...
829 (is_a_helper <gimple_statement_omp_single *>::test): ...this.
830 (is_a_helper <gimple_statement_omp_teams>::test): Convert to...
831 (is_a_helper <gimple_statement_omp_teams *>::test): ...this.
832 (is_a_helper <gimple_statement_omp_task>::test): Convert to...
833 (is_a_helper <gimple_statement_omp_task *>::test): ...this.
834 (is_a_helper <gimple_statement_phi>::test): Convert to...
835 (is_a_helper <gimple_statement_phi *>::test): ...this.
836 (is_a_helper <gimple_statement_transaction>::test): Convert to...
837 (is_a_helper <gimple_statement_transaction *>::test): ...this.
838 (is_a_helper <gimple_statement_try>::test): Convert to...
839 (is_a_helper <gimple_statement_try *>::test): ...this.
840 (is_a_helper <gimple_statement_wce>::test): Convert to...
841 (is_a_helper <gimple_statement_wce *>::test): ...this.
842 (is_a_helper <const gimple_statement_asm>::test): Convert to...
843 (is_a_helper <const gimple_statement_asm *>::test): ...this.
844 (is_a_helper <const gimple_statement_bind>::test): Convert to...
845 (is_a_helper <const gimple_statement_bind *>::test): ...this.
846 (is_a_helper <const gimple_statement_call>::test): Convert to...
847 (is_a_helper <const gimple_statement_call *>::test): ...this.
848 (is_a_helper <const gimple_statement_catch>::test): Convert to...
849 (is_a_helper <const gimple_statement_catch *>::test): ...this.
850 (is_a_helper <const gimple_statement_resx>::test): Convert to...
851 (is_a_helper <const gimple_statement_resx *>::test): ...this.
852 (is_a_helper <const gimple_statement_eh_dispatch>::test):
853 Convert to...
854 (is_a_helper <const gimple_statement_eh_dispatch *>::test):
855 ...this.
856 (is_a_helper <const gimple_statement_eh_filter>::test): Convert
857 to...
858 (is_a_helper <const gimple_statement_eh_filter *>::test): ...this.
859 (is_a_helper <const gimple_statement_omp_atomic_load>::test):
860 Convert to...
861 (is_a_helper <const gimple_statement_omp_atomic_load *>::test):
862 ...this.
863 (is_a_helper <const gimple_statement_omp_atomic_store>::test):
864 Convert to...
865 (is_a_helper <const gimple_statement_omp_atomic_store *>::test):
866 ...this.
867 (is_a_helper <const gimple_statement_omp_return>::test): Convert
868 to...
869 (is_a_helper <const gimple_statement_omp_return *>::test): ...this.
870 (is_a_helper <const gimple_statement_omp_continue>::test): Convert
871 to...
872 (is_a_helper <const gimple_statement_omp_continue *>::test): ...this.
873 (is_a_helper <const gimple_statement_omp_critical>::test): Convert
874 to...
875 (is_a_helper <const gimple_statement_omp_critical *>::test): ...this.
876 (is_a_helper <const gimple_statement_omp_for>::test): Convert to...
877 (is_a_helper <const gimple_statement_omp_for *>::test): ...this.
878 (is_a_helper <const gimple_statement_omp_taskreg>::test): Convert
879 to...
880 (is_a_helper <const gimple_statement_omp_taskreg *>::test): ...this.
881 (is_a_helper <const gimple_statement_omp_parallel>::test): Convert
882 to...
883 (is_a_helper <const gimple_statement_omp_parallel *>::test): ...this.
884 (is_a_helper <const gimple_statement_omp_target>::test): Convert
885 to...
886 (is_a_helper <const gimple_statement_omp_target *>::test): ...this.
887 (is_a_helper <const gimple_statement_omp_sections>::test): Convert
888 to...
889 (is_a_helper <const gimple_statement_omp_sections *>::test): ...this.
890 (is_a_helper <const gimple_statement_omp_single>::test): Convert
891 to...
892 (is_a_helper <const gimple_statement_omp_single *>::test): ...this.
893 (is_a_helper <const gimple_statement_omp_teams>::test): Convert
894 to...
895 (is_a_helper <const gimple_statement_omp_teams *>::test): ...this.
896 (is_a_helper <const gimple_statement_omp_task>::test): Convert
897 to...
898 (is_a_helper <const gimple_statement_omp_task *>::test): ...this.
899 (is_a_helper <const gimple_statement_phi>::test): Convert to...
900 (is_a_helper <const gimple_statement_phi *>::test): ...this.
901 (is_a_helper <const gimple_statement_transaction>::test): Convert
902 to...
903 (is_a_helper <const gimple_statement_transaction *>::test): ...this.
904 (is_a_helper <const gimple_statement_with_ops>::test): Convert
905 to...
906 (is_a_helper <const gimple_statement_with_ops *>::test): ...this.
907 (is_a_helper <gimple_statement_with_ops>::test): Convert to...
908 (is_a_helper <gimple_statement_with_ops *>::test): ...this.
909 (is_a_helper <const gimple_statement_with_memory_ops>::test):
910 Convert to...
911 (is_a_helper <const gimple_statement_with_memory_ops *>::test):
912 ...this.
913 (is_a_helper <gimple_statement_with_memory_ops>::test): Convert
914 to...
915 (is_a_helper <gimple_statement_with_memory_ops *>::test): ...this.
916
917 (gimple_use_ops): Update for removal of implicit pointer from the
918 is-a.h API.
919 (gimple_set_use_ops): Likewise.
920 (gimple_vuse): Likewise.
921 (gimple_vdef): Likewise.
922 (gimple_vuse_ptr): Likewise.
923 (gimple_vdef_ptr): Likewise.
924 (gimple_set_vuse): Likewise.
925 (gimple_set_vdef): Likewise.
926 (gimple_omp_return_set_lhs): Likewise.
927 (gimple_omp_return_lhs): Likewise.
928 (gimple_omp_return_lhs_ptr): Likewise.
929 (gimple_call_fntype): Likewise.
930 (gimple_call_set_fntype): Likewise.
931 (gimple_call_set_internal_fn): Likewise.
932 (gimple_call_use_set): Likewise.
933 (gimple_call_clobber_set): Likewise.
934 (gimple_bind_vars): Likewise.
935 (gimple_bind_set_vars): Likewise.
936 (gimple_bind_body_ptr): Likewise.
937 (gimple_bind_set_body): Likewise.
938 (gimple_bind_add_stmt): Likewise.
939 (gimple_bind_block): Likewise.
940 (gimple_bind_set_block): Likewise.
941 (gimple_asm_ninputs): Likewise.
942 (gimple_asm_noutputs): Likewise.
943 (gimple_asm_nclobbers): Likewise.
944 (gimple_asm_nlabels): Likewise.
945 (gimple_asm_input_op): Likewise.
946 (gimple_asm_input_op_ptr): Likewise.
947 (gimple_asm_output_op): Likewise.
948 (gimple_asm_output_op_ptr): Likewise.
949 (gimple_asm_set_output_op): Likewise.
950 (gimple_asm_clobber_op): Likewise.
951 (gimple_asm_set_clobber_op): Likewise.
952 (gimple_asm_label_op): Likewise.
953 (gimple_asm_set_label_op): Likewise.
954 (gimple_asm_string): Likewise.
955 (gimple_catch_types): Likewise.
956 (gimple_catch_types_ptr): Likewise.
957 (gimple_catch_handler_ptr): Likewise.
958 (gimple_catch_set_types): Likewise.
959 (gimple_catch_set_handler): Likewise.
960 (gimple_eh_filter_types): Likewise.
961 (gimple_eh_filter_types_ptr): Likewise.
962 (gimple_eh_filter_failure_ptr): Likewise.
963 (gimple_eh_filter_set_types): Likewise.
964 (gimple_eh_filter_set_failure): Likewise.
965 (gimple_eh_must_not_throw_fndecl): Likewise.
966 (gimple_eh_must_not_throw_set_fndecl): Likewise.
967 (gimple_eh_else_n_body_ptr): Likewise.
968 (gimple_eh_else_e_body_ptr): Likewise.
969 (gimple_eh_else_set_n_body): Likewise.
970 (gimple_eh_else_set_e_body): Likewise.
971 (gimple_try_eval_ptr): Likewise.
972 (gimple_try_cleanup_ptr): Likewise.
973 (gimple_try_set_eval): Likewise.
974 (gimple_try_set_cleanup): Likewise.
975 (gimple_wce_cleanup_ptr): Likewise.
976 (gimple_wce_set_cleanup): Likewise.
977 (gimple_phi_capacity): Likewise.
978 (gimple_phi_num_args): Likewise.
979 (gimple_phi_result): Likewise.
980 (gimple_phi_result_ptr): Likewise.
981 (gimple_phi_set_result): Likewise.
982 (gimple_phi_arg): Likewise.
983 (gimple_phi_set_arg): Likewise.
984 (gimple_resx_region): Likewise.
985 (gimple_resx_set_region): Likewise.
986 (gimple_eh_dispatch_region): Likewise.
987 (gimple_eh_dispatch_set_region): Likewise.
988 (gimple_omp_critical_name): Likewise.
989 (gimple_omp_critical_name_ptr): Likewise.
990 (gimple_omp_critical_set_name): Likewise.
991 (gimple_omp_for_clauses): Likewise.
992 (gimple_omp_for_clauses_ptr): Likewise.
993 (gimple_omp_for_set_clauses): Likewise.
994 (gimple_omp_for_collapse): Likewise.
995 (gimple_omp_for_index): Likewise.
996 (gimple_omp_for_index_ptr): Likewise.
997 (gimple_omp_for_set_index): Likewise.
998 (gimple_omp_for_initial): Likewise.
999 (gimple_omp_for_initial_ptr): Likewise.
1000 (gimple_omp_for_set_initial): Likewise.
1001 (gimple_omp_for_final): Likewise.
1002 (gimple_omp_for_final_ptr): Likewise.
1003 (gimple_omp_for_set_final): Likewise.
1004 (gimple_omp_for_incr): Likewise.
1005 (gimple_omp_for_incr_ptr): Likewise.
1006 (gimple_omp_for_set_incr): Likewise.
1007 (gimple_omp_for_pre_body_ptr): Likewise.
1008 (gimple_omp_for_set_pre_body): Likewise.
1009 (gimple_omp_parallel_clauses): Likewise.
1010 (gimple_omp_parallel_clauses_ptr): Likewise.
1011 (gimple_omp_parallel_set_clauses): Likewise.
1012 (gimple_omp_parallel_child_fn): Likewise.
1013 (gimple_omp_parallel_child_fn_ptr): Likewise.
1014 (gimple_omp_parallel_set_child_fn): Likewise.
1015 (gimple_omp_parallel_data_arg): Likewise.
1016 (gimple_omp_parallel_data_arg_ptr): Likewise.
1017 (gimple_omp_parallel_set_data_arg): Likewise.
1018 (gimple_omp_task_clauses): Likewise.
1019 (gimple_omp_task_clauses_ptr): Likewise.
1020 (gimple_omp_task_set_clauses): Likewise.
1021 (gimple_omp_task_child_fn): Likewise.
1022 (gimple_omp_task_child_fn_ptr): Likewise.
1023 (gimple_omp_task_set_child_fn): Likewise.
1024 (gimple_omp_task_data_arg): Likewise.
1025 (gimple_omp_task_data_arg_ptr): Likewise.
1026 (gimple_omp_task_set_data_arg): Likewise.
1027 (gimple_omp_taskreg_clauses): Likewise.
1028 (gimple_omp_taskreg_clauses_ptr): Likewise.
1029 (gimple_omp_taskreg_set_clauses): Likewise.
1030 (gimple_omp_taskreg_child_fn): Likewise.
1031 (gimple_omp_taskreg_child_fn_ptr): Likewise.
1032 (gimple_omp_taskreg_set_child_fn): Likewise.
1033 (gimple_omp_taskreg_data_arg): Likewise.
1034 (gimple_omp_taskreg_data_arg_ptr): Likewise.
1035 (gimple_omp_taskreg_set_data_arg): Likewise.
1036 (gimple_omp_task_copy_fn): Likewise.
1037 (gimple_omp_task_copy_fn_ptr): Likewise.
1038 (gimple_omp_task_set_copy_fn): Likewise.
1039 (gimple_omp_task_arg_size): Likewise.
1040 (gimple_omp_task_arg_size_ptr): Likewise.
1041 (gimple_omp_task_set_arg_size): Likewise.
1042 (gimple_omp_task_arg_align): Likewise.
1043 (gimple_omp_task_arg_align_ptr): Likewise.
1044 (gimple_omp_task_set_arg_align): Likewise.
1045 (gimple_omp_single_clauses): Likewise.
1046 (gimple_omp_single_clauses_ptr): Likewise.
1047 (gimple_omp_single_set_clauses): Likewise.
1048 (gimple_omp_target_clauses): Likewise.
1049 (gimple_omp_target_clauses_ptr): Likewise.
1050 (gimple_omp_target_set_clauses): Likewise.
1051 (gimple_omp_target_child_fn): Likewise.
1052 (gimple_omp_target_child_fn_ptr): Likewise.
1053 (gimple_omp_target_set_child_fn): Likewise.
1054 (gimple_omp_target_data_arg): Likewise.
1055 (gimple_omp_target_data_arg_ptr): Likewise.
1056 (gimple_omp_target_set_data_arg): Likewise.
1057 (gimple_omp_teams_clauses): Likewise.
1058 (gimple_omp_teams_clauses_ptr): Likewise.
1059 (gimple_omp_teams_set_clauses): Likewise.
1060 (gimple_omp_sections_clauses): Likewise.
1061 (gimple_omp_sections_clauses_ptr): Likewise.
1062 (gimple_omp_sections_set_clauses): Likewise.
1063 (gimple_omp_sections_control): Likewise.
1064 (gimple_omp_sections_control_ptr): Likewise.
1065 (gimple_omp_sections_set_control): Likewise.
1066 (gimple_omp_for_set_cond): Likewise.
1067 (gimple_omp_for_cond): Likewise.
1068 (gimple_omp_atomic_store_set_val): Likewise.
1069 (gimple_omp_atomic_store_val): Likewise.
1070 (gimple_omp_atomic_store_val_ptr): Likewise.
1071 (gimple_omp_atomic_load_set_lhs): Likewise.
1072 (gimple_omp_atomic_load_lhs): Likewise.
1073 (gimple_omp_atomic_load_lhs_ptr): Likewise.
1074 (gimple_omp_atomic_load_set_rhs): Likewise.
1075 (gimple_omp_atomic_load_rhs): Likewise.
1076 (gimple_omp_atomic_load_rhs_ptr): Likewise.
1077 (gimple_omp_continue_control_def): Likewise.
1078 (gimple_omp_continue_control_def_ptr): Likewise.
1079 (gimple_omp_continue_set_control_def): Likewise.
1080 (gimple_omp_continue_control_use): Likewise.
1081 (gimple_omp_continue_control_use_ptr): Likewise.
1082 (gimple_omp_continue_set_control_use): Likewise.
1083 (gimple_transaction_body_ptr): Likewise.
1084 (gimple_transaction_label): Likewise.
1085 (gimple_transaction_label_ptr): Likewise.
1086 (gimple_transaction_set_body): Likewise.
1087 (gimple_transaction_set_label): Likewise.
1088
1089 * ipa-devirt.c (build_type_inheritance_graph): Likewise.
1090 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1091 * ipa-ref.c (ipa_record_reference): Likewise.
1092 * ipa-reference.c (analyze_function): Likewise.
1093 (ipa_reference_write_optimization_summary): Likewise.
1094 * ipa.c (symtab_remove_unreachable_nodes): Likewise.
1095 (address_taken_from_non_vtable_p): Likewise.
1096 (comdat_can_be_unshared_p_1): Likewise.
1097 * lto-cgraph.c (lto_output_ref): Likewise.
1098 (add_references): Likewise.
1099 (compute_ltrans_boundary): Likewise.
1100 (output_symtab): Likewise.
1101 (input_ref): Likewise.
1102 (input_cgraph_1): Likewise.
1103 (output_cgraph_opt_summary): Likewise.
1104 * lto-streamer-out.c (lto_output): Likewise.
1105 (output_symbol_p): Likewise.
1106 * lto-streamer.h (lsei_next_function_in_partition): Likewise.
1107 (lsei_start_function_in_partition): Likewise.
1108 (lsei_next_variable_in_partition): Likewise.
1109 (lsei_start_variable_in_partition): Likewise.
1110 * symtab.c (insert_to_assembler_name_hash): Likewise.
1111 (unlink_from_assembler_name_hash): Likewise.
1112 (symtab_unregister_node): Likewise.
1113 (symtab_remove_node): Likewise.
1114 (dump_symtab_node): Likewise.
1115 (verify_symtab_base): Likewise.
1116 (verify_symtab_node): Likewise.
1117 (symtab_make_decl_local): Likewise.
1118 (symtab_alias_ultimate_target): Likewise.
1119 (symtab_resolve_alias): Likewise.
1120 (symtab_get_symbol_partitioning_class): Likewise.
1121 * tree-phinodes.c (allocate_phi_node): Likewise.
1122 (reserve_phi_args_for_new_edge): Likewise.
1123 (remove_phi_args): Likewise.
1124 * varpool.c (varpool_node_for_asm): Likewise.
1125 (varpool_remove_unreferenced_decls): Likewise.
1126
9219922e
JL
11272014-04-23 Jeff Law <law@redhat.com>
1128
1129 PR tree-optimization/60902
1130 * tree-ssa-threadedge.c
1131 (record_temporary_equivalences_from_stmts_at_dest): Make sure to
1132 invalidate outputs from statements that do not produce useful
1133 outputs for threading.
1134
36e17020
VK
11352014-04-23 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
1136
1137 * config/aarch64/aarch64.md (stack_protect_set, stack_protect_test)
1138 (stack_protect_set_<mode>, stack_protect_test_<mode>): Add
1139 machine descriptions for Stack Smashing Protector.
1140
57b77d46
RE
11412014-04-23 Richard Earnshaw <rearnsha@arm.com>
1142
1143 * aarch64.md (<optab>_rol<mode>3): New pattern.
1144 (<optab>_rolsi3_uxtw): Likewise.
1145 * aarch64.c (aarch64_strip_shift): Handle ROTATE and ROTATERT.
1146
984c2f30
JG
11472014-04-23 James Greenhalgh <james.greenhalgh@arm.com>
1148
1149 * config/arm/arm.c (arm_cortex_a57_tune): Initialize all fields.
1150 (arm_cortex_a12_tune): Likewise.
1151
18b42b2a
KT
11522014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1153
1154 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle BSWAP.
1155
ea1e9168
KT
11562014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1157
1158 * config/arm/arm.md (arm_rev16si2): New pattern.
1159 (arm_rev16si2_alt): Likewise.
1160 * config/arm/arm.c (arm_new_rtx_costs): Handle rev16 case.
1161
f7d5cf8d
KT
11622014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1163
1164 * config/aarch64/aarch64.md (rev16<mode>2): New pattern.
1165 (rev16<mode>2_alt): Likewise.
1166 * config/aarch64/aarch64.c (aarch64_rtx_costs): Handle rev16 case.
1167 * config/arm/aarch-common.c (aarch_rev16_shright_mask_imm_p): New.
1168 (aarch_rev16_shleft_mask_imm_p): Likewise.
1169 (aarch_rev16_p_1): Likewise.
1170 (aarch_rev16_p): Likewise.
1171 * config/arm/aarch-common-protos.h (aarch_rev16_p): Declare extern.
1172 (aarch_rev16_shright_mask_imm_p): Likewise.
1173 (aarch_rev16_shleft_mask_imm_p): Likewise.
1174
9ac05ae5
KT
11752014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1176
1177 * config/arm/aarch-common-protos.h (alu_cost_table): Add rev field.
1178 * config/arm/aarch-cost-tables.h (generic_extra_costs): Specify
1179 rev cost.
1180 (cortex_a53_extra_costs): Likewise.
1181 (cortex_a57_extra_costs): Likewise.
1182 * config/arm/arm.c (cortexa9_extra_costs): Likewise.
1183 (cortexa7_extra_costs): Likewise.
1184 (cortexa8_extra_costs): Likewise.
1185 (cortexa12_extra_costs): Likewise.
1186 (cortexa15_extra_costs): Likewise.
1187 (v7m_extra_costs): Likewise.
1188 (arm_new_rtx_costs): Handle BSWAP.
1189
e0d8c86c
KT
11902013-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1191
1192 * config/arm/arm.c (cortexa8_extra_costs): New table.
1193 (arm_cortex_a8_tune): New tuning struct.
1194 * config/arm/arm-cores.def (cortex-a8): Use cortex_a8 tuning struct.
1195
ba9b1a63
KT
11962014-04-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1197
1198 * config/arm/arm.c (arm_new_rtx_costs): Handle FMA.
1199
da4cfeac
RB
12002014-04-23 Richard Biener <rguenther@suse.de>
1201
1202 * Makefile.in (OBJS): Remove loop-unswitch.o.
1203 * loop-unswitch.c: Delete.
1204 * tree-pass.h (make_pass_rtl_unswitch): Remove.
1205 * passes.def (pass_rtl_unswitch): Likewise.
1206 * loop-init.c (gate_rtl_unswitch): Likewise.
1207 (rtl_unswitch): Likewise.
1208 (pass_data_rtl_unswitch): Likewise.
1209 (pass_rtl_unswitch): Likewise.
1210 (make_pass_rtl_unswitch): Likewise.
1211 * rtl.h (reversed_condition): Likewise.
1212 (compare_and_jump_seq): Likewise.
1213 * loop-iv.c (reversed_condition): Move here from loop-unswitch.c
1214 and make static.
1215 * loop-unroll.c (compare_and_jump_seq): Likewise.
1216
22718afe
RB
12172014-04-23 Richard Biener <rguenther@suse.de>
1218
1219 PR tree-optimization/60903
1220 * tree-ssa-loop-im.c (analyze_memory_references): Remove
1221 commented code block.
1222 (execute_sm_if_changed): Properly apply IRREDUCIBLE_LOOP
1223 loop flags to newly created BBs and edges.
1224
f7961364
NC
12252014-04-23 Nick Clifton <nickc@redhat.com>
1226
1227 * config/msp430/msp430.c (msp430_handle_option): Move function
1228 to msp430-common.c
1229 (msp430_option_override): Simplify mcu and mcpu option handling.
1230 (msp430_is_f5_mcu): Rename to msp430_use_f5_series_hwmult. Add
1231 support for -mhwmult command line option.
1232 (has_32bit_hwmult): Rename to use_32bit_hwmult. Add support for
1233 -mhwmult command line option.
1234 (msp430_hwmult_enabled): Delete.
1235 (msp43o_output_labelref): Add support for -mhwmult command line
1236 option.
1237 * config/msp430/msp430.md (mulhisi3, umulhisi3, mulsidi3)
1238 (umulsidi3): Likewise.
1239 * config/msp430/msp430.opt (mmcu): Add Report attribute.
1240 (mcpu, mlarge, msmall): Likewise.
1241 (mhwmult): New option.
1242 * config/msp430/msp430-protos.h (msp430_hwmult_enabled): Remove
1243 prototype.
1244 (msp430_is_f5_mcu): Remove prototype.
1245 (msp430_use_f5_series_hwmult): Add prototype.
1246 * config/msp430/msp430-opts.h: New file.
1247 * common/config/msp430: New directory.
1248 * common/config/msp430/msp430-common.c: New file.
1249 * config.gcc (msp430): Remove target_has_targetm_common.
1250 * doc/invoke.texi: Document -mhwmult command line option.
1251
f70df035
NC
12522014-04-23 Nick Clifton <nickc@redhat.com>
1253
1254 * config/i386/cygwin.h (ENDFILE_SPEC): Include
1255 default-manifest.o if it can be found in the search path.
1256 * config/i386/mingw32.h (ENDFILE_SPEC): Likewise.
1257
6a9accca
TG
12582014-04-23 Terry Guo <terry.guo@arm.com>
1259
1260 * config/arm/arm.h (ASM_APP_OFF): Re-define it in a cleaner way.
1261
62ba699e
RB
12622014-04-23 Richard Biener <rguenther@suse.de>
1263
1264 PR middle-end/60895
1265 * tree-inline.c (declare_return_variable): Use mark_addressable.
1266
92562f88
RB
12672014-04-23 Richard Biener <rguenther@suse.de>
1268
1269 PR middle-end/60891
1270 * loop-init.c (loop_optimizer_init): Make sure to apply
1271 LOOPS_MAY_HAVE_MULTIPLE_LATCHES before fixing up loops.
1272
1c33c9b7
JJ
12732014-04-22 Jakub Jelinek <jakub@redhat.com>
1274
1275 PR sanitizer/60275
1276 * common.opt (fsanitize-recover, fsanitize-undefined-trap-on-error):
1277 New options.
1278 * gcc.c (sanitize_spec_function): Don't return "" for "undefined"
1279 if flag_sanitize_undefined_trap_on_error.
1280 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW_ABORT,
1281 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS_ABORT,
1282 BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE_ABORT,
1283 BUILT_IN_UBSAN_HANDLE_TYPE_MISMATCH_ABORT,
1284 BUILT_IN_UBSAN_HANDLE_ADD_OVERFLOW_ABORT,
1285 BUILT_IN_UBSAN_HANDLE_SUB_OVERFLOW_ABORT,
1286 BUILT_IN_UBSAN_HANDLE_MUL_OVERFLOW_ABORT,
1287 BUILT_IN_UBSAN_HANDLE_NEGATE_OVERFLOW_ABORT,
1288 BUILT_IN_UBSAN_HANDLE_LOAD_INVALID_VALUE_ABORT): New builtins.
1289 * ubsan.c (ubsan_instrument_unreachable): Return
1290 __builtin_trap () if flag_sanitize_undefined_trap_on_error.
1291 (ubsan_expand_null_ifn): Emit __builtin_trap ()
1292 if flag_sanitize_undefined_trap_on_error and
1293 __ubsan_handle_type_mismatch_abort if !flag_sanitize_recover.
1294 (ubsan_expand_null_ifn, ubsan_build_overflow_builtin,
1295 instrument_bool_enum_load): Emit __builtin_trap () if
1296 flag_sanitize_undefined_trap_on_error and
1297 __builtin_handle_*_abort () if !flag_sanitize_recover.
1298 * doc/invoke.texi (-fsanitize-recover,
1299 -fsanitize-undefined-trap-on-error): Document.
1300
5e66b2e5
CB
13012014-04-22 Christian Bruel <christian.bruel@st.com>
1302
1303 * config/sh/sh.md (mov<mode>): Replace movQIHI.
1304 Force immediates to SImode.
1305
aa221564
SL
13062014-04-22 Sandra Loosemore <sandra@codesourcery.com>
1307
1308 * config/nios2/nios2.md (UNSPEC_ROUND): New.
1309 (lroundsfsi2): New.
1310 * config/nios2/nios2.opt (mno-custom-round, mcustom-round=): New.
1311 * config/nios2/nios2-opts.h (N2FPU_ALL_CODES): Add round.
1312 * config/nios2/nios2.c (N2F_NO_ERRNO): Define.
1313 (nios2_fpu_insn): Add entry for round.
1314 (N2FPU_NO_ERRNO_P): Define.
1315 (nios2_custom_check_insns): Add check for N2F_NO_ERRNO and
1316 flag_errno_math.
1317 * doc/invoke.texi (Nios II Options): Document -mcustom-round.
1318
2195867f
RH
13192014-04-22 Richard Henderson <rth@redhat.com>
1320
1321 * config/aarch64/aarch64 (addti3, subti3): New expanders.
1322 (add<GPI>3_compare0): Remove leading * from name.
1323 (add<GPI>3_carryin): Likewise.
1324 (sub<GPI>3_compare0): Likewise.
1325 (sub<GPI>3_carryin): Likewise.
1326 (<su_optab>mulditi3): New expander.
1327 (multi3): New expander.
1328 (madd<GPI>): Remove leading * from name.
1329
c6d43074
MJ
13302014-04-22 Martin Jambor <mjambor@suse.cz>
1331
1332 * cgraphclones.c (cgraph_function_versioning): Copy
1333 ipa_transforms_to_apply instead of asserting it is empty.
1334
e54c2dd3
L
13352014-04-22 H.J. Lu <hongjiu.lu@intel.com>
1336
1337 PR target/60868
f70df035 1338 * config/i386/i386.c (ix86_expand_set_or_movmem): Call counter_mode
e54c2dd3
L
1339 on count_exp to get mode.
1340
621ad2de
AP
13412014-04-22 Andrew Pinski <apinski@cavium.com>
1342
1343 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
1344 Handle TLS for ILP32.
1345 * config/aarch64/aarch64.md (tlsie_small): Rename to ...
1346 (tlsie_small_<mode>): this and handle PTR.
1347 (tlsie_small_sidi): New pattern.
1348 (tlsle_small): Change to an expand to handle ILP32.
1349 (tlsle_small_<mode>): New pattern.
1350 (tlsdesc_small): Rename to ...
1351 (tlsdesc_small_<mode>): this and handle PTR.
1352
5cb74e90
RR
13532014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1354
1355 * config/aarch64/aarch64.c (TARGET_FLAGS_REGNUM): Define.
1356
bcd48995
AV
13572014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1358
1359 * config/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1360 (aarch64_types_signed_unsigned_qualifiers): Qualifier added.
1361 (aarch64_types_signed_poly_qualifiers): Likewise.
1362 (aarch64_types_unsigned_signed_qualifiers): Likewise.
1363 (aarch64_types_poly_signed_qualifiers): Likewise.
1364 (TYPES_REINTERP_SS): Type macro added.
1365 (TYPES_REINTERP_SU): Likewise.
1366 (TYPES_REINTERP_SP): Likewise.
1367 (TYPES_REINTERP_US): Likewise.
1368 (TYPES_REINTERP_PS): Likewise.
1369 (aarch64_fold_builtin): New expression folding added.
1370 * config/aarch64/aarch64-simd-builtins.def (REINTERP):
1371 Declarations removed.
1372 (REINTERP_SS): Declarations added.
1373 (REINTERP_US): Likewise.
1374 (REINTERP_PS): Likewise.
1375 (REINTERP_SU): Likewise.
1376 (REINTERP_SP): Likewise.
1377 * config/aarch64/arm_neon.h (vreinterpret_p8_f64): Implemented.
1378 (vreinterpretq_p8_f64): Likewise.
1379 (vreinterpret_p16_f64): Likewise.
1380 (vreinterpretq_p16_f64): Likewise.
1381 (vreinterpret_f32_f64): Likewise.
1382 (vreinterpretq_f32_f64): Likewise.
1383 (vreinterpret_f64_f32): Likewise.
1384 (vreinterpret_f64_p8): Likewise.
1385 (vreinterpret_f64_p16): Likewise.
1386 (vreinterpret_f64_s8): Likewise.
1387 (vreinterpret_f64_s16): Likewise.
1388 (vreinterpret_f64_s32): Likewise.
1389 (vreinterpret_f64_s64): Likewise.
1390 (vreinterpret_f64_u8): Likewise.
1391 (vreinterpret_f64_u16): Likewise.
1392 (vreinterpret_f64_u32): Likewise.
1393 (vreinterpret_f64_u64): Likewise.
1394 (vreinterpretq_f64_f32): Likewise.
1395 (vreinterpretq_f64_p8): Likewise.
1396 (vreinterpretq_f64_p16): Likewise.
1397 (vreinterpretq_f64_s8): Likewise.
1398 (vreinterpretq_f64_s16): Likewise.
1399 (vreinterpretq_f64_s32): Likewise.
1400 (vreinterpretq_f64_s64): Likewise.
1401 (vreinterpretq_f64_u8): Likewise.
1402 (vreinterpretq_f64_u16): Likewise.
1403 (vreinterpretq_f64_u32): Likewise.
1404 (vreinterpretq_f64_u64): Likewise.
1405 (vreinterpret_s64_f64): Likewise.
1406 (vreinterpretq_s64_f64): Likewise.
1407 (vreinterpret_u64_f64): Likewise.
1408 (vreinterpretq_u64_f64): Likewise.
1409 (vreinterpret_s8_f64): Likewise.
1410 (vreinterpretq_s8_f64): Likewise.
1411 (vreinterpret_s16_f64): Likewise.
1412 (vreinterpretq_s16_f64): Likewise.
1413 (vreinterpret_s32_f64): Likewise.
1414 (vreinterpretq_s32_f64): Likewise.
1415 (vreinterpret_u8_f64): Likewise.
1416 (vreinterpretq_u8_f64): Likewise.
1417 (vreinterpret_u16_f64): Likewise.
1418 (vreinterpretq_u16_f64): Likewise.
1419 (vreinterpret_u32_f64): Likewise.
1420 (vreinterpretq_u32_f64): Likewise.
1421
0bf3afc1
AV
14222014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1423
1424 * config/aarch64/aarch64/aarch64-builtins.c (TYPES_REINTERP): Removed.
1425 * config/aarch64/aarch64/aarch64-simd-builtins.def (REINTERP): Removed.
1426 (vreinterpret_p8_s8): Likewise.
1427 * config/aarch64/aarch64/arm_neon.h (vreinterpret_p8_s8): Uses cast.
1428 (vreinterpret_p8_s16): Likewise.
1429 (vreinterpret_p8_s32): Likewise.
1430 (vreinterpret_p8_s64): Likewise.
1431 (vreinterpret_p8_f32): Likewise.
1432 (vreinterpret_p8_u8): Likewise.
1433 (vreinterpret_p8_u16): Likewise.
1434 (vreinterpret_p8_u32): Likewise.
1435 (vreinterpret_p8_u64): Likewise.
1436 (vreinterpret_p8_p16): Likewise.
1437 (vreinterpretq_p8_s8): Likewise.
1438 (vreinterpretq_p8_s16): Likewise.
1439 (vreinterpretq_p8_s32): Likewise.
1440 (vreinterpretq_p8_s64): Likewise.
1441 (vreinterpretq_p8_f32): Likewise.
1442 (vreinterpretq_p8_u8): Likewise.
1443 (vreinterpretq_p8_u16): Likewise.
1444 (vreinterpretq_p8_u32): Likewise.
1445 (vreinterpretq_p8_u64): Likewise.
1446 (vreinterpretq_p8_p16): Likewise.
1447 (vreinterpret_p16_s8): Likewise.
1448 (vreinterpret_p16_s16): Likewise.
1449 (vreinterpret_p16_s32): Likewise.
1450 (vreinterpret_p16_s64): Likewise.
1451 (vreinterpret_p16_f32): Likewise.
1452 (vreinterpret_p16_u8): Likewise.
1453 (vreinterpret_p16_u16): Likewise.
1454 (vreinterpret_p16_u32): Likewise.
1455 (vreinterpret_p16_u64): Likewise.
1456 (vreinterpret_p16_p8): Likewise.
1457 (vreinterpretq_p16_s8): Likewise.
1458 (vreinterpretq_p16_s16): Likewise.
1459 (vreinterpretq_p16_s32): Likewise.
1460 (vreinterpretq_p16_s64): Likewise.
1461 (vreinterpretq_p16_f32): Likewise.
1462 (vreinterpretq_p16_u8): Likewise.
1463 (vreinterpretq_p16_u16): Likewise.
1464 (vreinterpretq_p16_u32): Likewise.
1465 (vreinterpretq_p16_u64): Likewise.
1466 (vreinterpretq_p16_p8): Likewise.
1467 (vreinterpret_f32_s8): Likewise.
1468 (vreinterpret_f32_s16): Likewise.
1469 (vreinterpret_f32_s32): Likewise.
1470 (vreinterpret_f32_s64): Likewise.
1471 (vreinterpret_f32_u8): Likewise.
1472 (vreinterpret_f32_u16): Likewise.
1473 (vreinterpret_f32_u32): Likewise.
1474 (vreinterpret_f32_u64): Likewise.
1475 (vreinterpret_f32_p8): Likewise.
1476 (vreinterpret_f32_p16): Likewise.
1477 (vreinterpretq_f32_s8): Likewise.
1478 (vreinterpretq_f32_s16): Likewise.
1479 (vreinterpretq_f32_s32): Likewise.
1480 (vreinterpretq_f32_s64): Likewise.
1481 (vreinterpretq_f32_u8): Likewise.
1482 (vreinterpretq_f32_u16): Likewise.
1483 (vreinterpretq_f32_u32): Likewise.
1484 (vreinterpretq_f32_u64): Likewise.
1485 (vreinterpretq_f32_p8): Likewise.
1486 (vreinterpretq_f32_p16): Likewise.
1487 (vreinterpret_s64_s8): Likewise.
1488 (vreinterpret_s64_s16): Likewise.
1489 (vreinterpret_s64_s32): Likewise.
1490 (vreinterpret_s64_f32): Likewise.
1491 (vreinterpret_s64_u8): Likewise.
1492 (vreinterpret_s64_u16): Likewise.
1493 (vreinterpret_s64_u32): Likewise.
1494 (vreinterpret_s64_u64): Likewise.
1495 (vreinterpret_s64_p8): Likewise.
1496 (vreinterpret_s64_p16): Likewise.
1497 (vreinterpretq_s64_s8): Likewise.
1498 (vreinterpretq_s64_s16): Likewise.
1499 (vreinterpretq_s64_s32): Likewise.
1500 (vreinterpretq_s64_f32): Likewise.
1501 (vreinterpretq_s64_u8): Likewise.
1502 (vreinterpretq_s64_u16): Likewise.
1503 (vreinterpretq_s64_u32): Likewise.
1504 (vreinterpretq_s64_u64): Likewise.
1505 (vreinterpretq_s64_p8): Likewise.
1506 (vreinterpretq_s64_p16): Likewise.
1507 (vreinterpret_u64_s8): Likewise.
1508 (vreinterpret_u64_s16): Likewise.
1509 (vreinterpret_u64_s32): Likewise.
1510 (vreinterpret_u64_s64): Likewise.
1511 (vreinterpret_u64_f32): Likewise.
1512 (vreinterpret_u64_u8): Likewise.
1513 (vreinterpret_u64_u16): Likewise.
1514 (vreinterpret_u64_u32): Likewise.
1515 (vreinterpret_u64_p8): Likewise.
1516 (vreinterpret_u64_p16): Likewise.
1517 (vreinterpretq_u64_s8): Likewise.
1518 (vreinterpretq_u64_s16): Likewise.
1519 (vreinterpretq_u64_s32): Likewise.
1520 (vreinterpretq_u64_s64): Likewise.
1521 (vreinterpretq_u64_f32): Likewise.
1522 (vreinterpretq_u64_u8): Likewise.
1523 (vreinterpretq_u64_u16): Likewise.
1524 (vreinterpretq_u64_u32): Likewise.
1525 (vreinterpretq_u64_p8): Likewise.
1526 (vreinterpretq_u64_p16): Likewise.
1527 (vreinterpret_s8_s16): Likewise.
1528 (vreinterpret_s8_s32): Likewise.
1529 (vreinterpret_s8_s64): Likewise.
1530 (vreinterpret_s8_f32): Likewise.
1531 (vreinterpret_s8_u8): Likewise.
1532 (vreinterpret_s8_u16): Likewise.
1533 (vreinterpret_s8_u32): Likewise.
1534 (vreinterpret_s8_u64): Likewise.
1535 (vreinterpret_s8_p8): Likewise.
1536 (vreinterpret_s8_p16): Likewise.
1537 (vreinterpretq_s8_s16): Likewise.
1538 (vreinterpretq_s8_s32): Likewise.
1539 (vreinterpretq_s8_s64): Likewise.
1540 (vreinterpretq_s8_f32): Likewise.
1541 (vreinterpretq_s8_u8): Likewise.
1542 (vreinterpretq_s8_u16): Likewise.
1543 (vreinterpretq_s8_u32): Likewise.
1544 (vreinterpretq_s8_u64): Likewise.
1545 (vreinterpretq_s8_p8): Likewise.
1546 (vreinterpretq_s8_p16): Likewise.
1547 (vreinterpret_s16_s8): Likewise.
1548 (vreinterpret_s16_s32): Likewise.
1549 (vreinterpret_s16_s64): Likewise.
1550 (vreinterpret_s16_f32): Likewise.
1551 (vreinterpret_s16_u8): Likewise.
1552 (vreinterpret_s16_u16): Likewise.
1553 (vreinterpret_s16_u32): Likewise.
1554 (vreinterpret_s16_u64): Likewise.
1555 (vreinterpret_s16_p8): Likewise.
1556 (vreinterpret_s16_p16): Likewise.
1557 (vreinterpretq_s16_s8): Likewise.
1558 (vreinterpretq_s16_s32): Likewise.
1559 (vreinterpretq_s16_s64): Likewise.
1560 (vreinterpretq_s16_f32): Likewise.
1561 (vreinterpretq_s16_u8): Likewise.
1562 (vreinterpretq_s16_u16): Likewise.
1563 (vreinterpretq_s16_u32): Likewise.
1564 (vreinterpretq_s16_u64): Likewise.
1565 (vreinterpretq_s16_p8): Likewise.
1566 (vreinterpretq_s16_p16): Likewise.
1567 (vreinterpret_s32_s8): Likewise.
1568 (vreinterpret_s32_s16): Likewise.
1569 (vreinterpret_s32_s64): Likewise.
1570 (vreinterpret_s32_f32): Likewise.
1571 (vreinterpret_s32_u8): Likewise.
1572 (vreinterpret_s32_u16): Likewise.
1573 (vreinterpret_s32_u32): Likewise.
1574 (vreinterpret_s32_u64): Likewise.
1575 (vreinterpret_s32_p8): Likewise.
1576 (vreinterpret_s32_p16): Likewise.
1577 (vreinterpretq_s32_s8): Likewise.
1578 (vreinterpretq_s32_s16): Likewise.
1579 (vreinterpretq_s32_s64): Likewise.
1580 (vreinterpretq_s32_f32): Likewise.
1581 (vreinterpretq_s32_u8): Likewise.
1582 (vreinterpretq_s32_u16): Likewise.
1583 (vreinterpretq_s32_u32): Likewise.
1584 (vreinterpretq_s32_u64): Likewise.
1585 (vreinterpretq_s32_p8): Likewise.
1586 (vreinterpretq_s32_p16): Likewise.
1587 (vreinterpret_u8_s8): Likewise.
1588 (vreinterpret_u8_s16): Likewise.
1589 (vreinterpret_u8_s32): Likewise.
1590 (vreinterpret_u8_s64): Likewise.
1591 (vreinterpret_u8_f32): Likewise.
1592 (vreinterpret_u8_u16): Likewise.
1593 (vreinterpret_u8_u32): Likewise.
1594 (vreinterpret_u8_u64): Likewise.
1595 (vreinterpret_u8_p8): Likewise.
1596 (vreinterpret_u8_p16): Likewise.
1597 (vreinterpretq_u8_s8): Likewise.
1598 (vreinterpretq_u8_s16): Likewise.
1599 (vreinterpretq_u8_s32): Likewise.
1600 (vreinterpretq_u8_s64): Likewise.
1601 (vreinterpretq_u8_f32): Likewise.
1602 (vreinterpretq_u8_u16): Likewise.
1603 (vreinterpretq_u8_u32): Likewise.
1604 (vreinterpretq_u8_u64): Likewise.
1605 (vreinterpretq_u8_p8): Likewise.
1606 (vreinterpretq_u8_p16): Likewise.
1607 (vreinterpret_u16_s8): Likewise.
1608 (vreinterpret_u16_s16): Likewise.
1609 (vreinterpret_u16_s32): Likewise.
1610 (vreinterpret_u16_s64): Likewise.
1611 (vreinterpret_u16_f32): Likewise.
1612 (vreinterpret_u16_u8): Likewise.
1613 (vreinterpret_u16_u32): Likewise.
1614 (vreinterpret_u16_u64): Likewise.
1615 (vreinterpret_u16_p8): Likewise.
1616 (vreinterpret_u16_p16): Likewise.
1617 (vreinterpretq_u16_s8): Likewise.
1618 (vreinterpretq_u16_s16): Likewise.
1619 (vreinterpretq_u16_s32): Likewise.
1620 (vreinterpretq_u16_s64): Likewise.
1621 (vreinterpretq_u16_f32): Likewise.
1622 (vreinterpretq_u16_u8): Likewise.
1623 (vreinterpretq_u16_u32): Likewise.
1624 (vreinterpretq_u16_u64): Likewise.
1625 (vreinterpretq_u16_p8): Likewise.
1626 (vreinterpretq_u16_p16): Likewise.
1627 (vreinterpret_u32_s8): Likewise.
1628 (vreinterpret_u32_s16): Likewise.
1629 (vreinterpret_u32_s32): Likewise.
1630 (vreinterpret_u32_s64): Likewise.
1631 (vreinterpret_u32_f32): Likewise.
1632 (vreinterpret_u32_u8): Likewise.
1633 (vreinterpret_u32_u16): Likewise.
1634 (vreinterpret_u32_u64): Likewise.
1635 (vreinterpret_u32_p8): Likewise.
1636 (vreinterpret_u32_p16): Likewise.
1637 (vreinterpretq_u32_s8): Likewise.
1638 (vreinterpretq_u32_s16): Likewise.
1639 (vreinterpretq_u32_s32): Likewise.
1640 (vreinterpretq_u32_s64): Likewise.
1641 (vreinterpretq_u32_f32): Likewise.
1642 (vreinterpretq_u32_u8): Likewise.
1643 (vreinterpretq_u32_u16): Likewise.
1644 (vreinterpretq_u32_u64): Likewise.
1645 (vreinterpretq_u32_p8): Likewise.
1646 (vreinterpretq_u32_p16): Likewise.
1647
9551c7ec
AV
16482014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1649
1650 * gcc/config/aarch64/aarch64-simd.md (aarch64_s<optab><mode>):
1651 Pattern extended.
1652 * config/aarch64/aarch64-simd-builtins.def (sqneg): Iterator
1653 extended.
1654 (sqabs): Likewise.
1655 * config/aarch64/arm_neon.h (vqneg_s64): New intrinsic.
1656 (vqnegd_s64): Likewise.
1657 (vqabs_s64): Likewise.
1658 (vqabsd_s64): Likewise.
1659
f2a2c4b5
RH
16602014-04-22 Richard Henderson <rth@redhat.com>
1661
1662 * config/sparc/sparc.c (sparc_init_modes): Hoist GET_MODE_SIZE
1663 computation to the top of the loop.
1664
53e5ace2
RL
16652014-04-22 Renlin <renlin.li@arm.com>
1666 Jiong Wang <jiong.wang@arm.com>
1667
1668 * config/aarch64/aarch64.h (aarch64_frame): Delete "fp_lr_offset".
1669 * config/aarch64/aarch64.c (aarch64_layout_frame)
1670 (aarch64_initial_elimination_offset): Likewise.
1671
78c29983
MS
16722014-04-22 Marcus Shawcroft <marcus.shawcroft@arm.com>
1673
1674 * config/aarch64/aarch64.c (aarch64_initial_elimination_offset):
1675 Fix indentation.
1676
dbf23a79
RS
16772014-04-22 Richard Sandiford <rdsandiford@googlemail.com>
1678
1679 * machmode.h (bitwise_mode_for_mode): Declare.
1680 * stor-layout.h (bitwise_type_for_mode): Likewise.
1681 * stor-layout.c (bitwise_mode_for_mode): New function.
1682 (bitwise_type_for_mode): Likewise.
1683 * builtins.c (fold_builtin_memory_op): Use it instead of
1684 int_mode_for_mode and build_nonstandard_integer_type.
1685
d9f069ab
RO
16862014-04-22 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1687
1688 * config.gcc (enable_obsolete): Remove *-*-solaris2.9*.
1689 (*-*-solaris2.[0-9] | *-*-solaris2.[0-9].*): Mark unsupported.
1690 (*-*-solaris2*): Simplify.
1691 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Likewise.
1692 (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*): Remove
1693 *-*-solaris2.9* handling.
1694
1695 * configure.ac (gcc_cv_as_hidden): Remove test for Solaris 9/x86
1696 as bug.
1697 (gcc_cv_ld_hidden): Remove *-*-solaris2.9* handling.
1698 (ld_tls_support): Remove i?86-*-solaris2.9, sparc*-*-solaris2.9
1699 handling, simplify.
1700 (gcc_cv_as_gstabs_flag): Remove workaround for Solaris 9/x86 as bug.
1701 * configure: Regenerate.
1702
1703 * config/i386/sol2-9.h: Remove.
1704
1705 * doc/install.texi (Specific, i?86-*-solaris2.9): Remove.
1706 (Specific, *-*-solaris2*): Mention Solaris 9 support removal.
1707 Remove Solaris 9 references.
1708
0d35c5c2
VP
17092014-04-22 Vidya Praveen <vidyapraveen@arm.com>
1710
1711 * aarch64.md (float<GPI:mode><GPF:mode>2): Remove.
1712 (floatuns<GPI:mode><GPF:mode>2): Remove.
1713 (<optab><fcvt_target><GPF:mode>2): New pattern for equal width float
1714 and floatuns conversions.
1715 (<optab><fcvt_iesize><GPF:mode>2): New pattern for inequal width float
1716 and floatuns conversions.
1717 * iterators.md (fcvt_target, FCVT_TARGET): Support SF and DF modes.
1718 (w1,w2): New mode attributes for inequal width conversions.
1719
16a3246f
RL
17202014-04-22 Renlin Li <Renlin.Li@arm.com>
1721
1722 * config/aarch64/aarch64.c (aarch64_print_operand_address): Adjust
1723 the output asm format.
1724
110e1ccc
JG
17252014-04-22 James Greenhalgh <james.greenhalgh@arm.com>
1726
1727 * config/aarch64/aarch64-simd.md
1728 (aarch64_cm<optab>di): Always split.
1729 (*aarch64_cm<optab>di): New.
1730 (aarch64_cmtstdi): Always split.
1731 (*aarch64_cmtstdi): New.
1732
2044a4c3
JJ
17332014-04-22 Jakub Jelinek <jakub@redhat.com>
1734
1735 PR tree-optimization/60823
1736 * omp-low.c (ipa_simd_modify_function_body): Go through
1737 all SSA_NAMEs and for those refering to vector arguments
1738 which are going to be replaced adjust SSA_NAME_VAR and,
1739 if it is a default definition, change it into a non-default
1740 definition assigned at the beginning of function from new_decl.
1741 (ipa_simd_modify_stmt_ops): Rewritten.
1742 * tree-dfa.c (set_ssa_default_def): When removing default def,
1743 check for NULL loc instead of NULL *loc.
1744
9d8b4d1c
RR
17452014-04-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1746
1747 * config/arm/arm.c (arm_hard_regno_mode_ok): Loosen
1748 restrictions on core registers for DImode values in Thumb2.
1749
a01be1ae
IB
17502014-04-22 Ian Bolton <ian.bolton@arm.com>
1751
1752 * config/arm/arm.md (*anddi_notdi_zesidi): New pattern.
1753 * config/arm/thumb2.md (*iordi_notdi_zesidi): New pattern.
1754
80d3417b
IB
17552014-04-22 Ian Bolton <ian.bolton@arm.com>
1756
1757 * config/arm/thumb2.md (*iordi_notdi_di): New pattern.
1758 (*iordi_notzesidi_di): Likewise.
1759 (*iordi_notsesidi_di): Likewise.
1760
46fbb3eb
IB
17612014-04-22 Ian Bolton <ian.bolton@arm.com>
1762
1763 * config/arm/arm-protos.h (tune_params): New struct members.
1764 * config/arm/arm.c: Initialise tune_params per processor.
1765 (thumb2_reorg): Suppress conversion from t32 to t16 when optimizing
1766 for speed, based on new tune_params.
1767
74dc11ed
AV
17682014-04-22 Alex Velenko <Alex.Velenko@arm.com>
1769
1770 * config/aarch64/aarch64-builtins.c (BUILTIN_VDQF_DF): Macro
1771 added.
1772 * config/aarch64/aarch64-simd-builtins.def (frintn): Use added
1773 macro.
1774 * config/aarch64/aarch64-simd.md (<frint_pattern>): Comment
1775 corrected.
1776 * config/aarch64/aarch64.md (<frint_pattern>): Likewise.
1777 * config/aarch64/arm_neon.h (vrnd_f64): Added.
1778 (vrnda_f64): Likewise.
1779 (vrndi_f64): Likewise.
1780 (vrndm_f64): Likewise.
1781 (vrndn_f64): Likewise.
1782 (vrndp_f64): Likewise.
1783 (vrndx_f64): Likewise.
1784
35923e46
ZC
17852014-04-22 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1786
1787 * config/arm/arm.c (arm_print_operand, thumb_exit): Make sure
1788 GET_MODE_SIZE argument is enum machine_mode.
1789
4552e70c
JJ
17902014-04-22 Jakub Jelinek <jakub@redhat.com>
1791
1792 PR target/60910
1793 * config/sparc/sparc.c (sparc_init_modes): Pass enum machine_mode
1794 value instead of int to GET_MODE_CLASS and GET_MODE_SIZE macros.
1795
e5dcd695
LZ
17962014-04-22 Lin Zuojian <manjian2006@gmail.com>
1797
1798 PR middle-end/60281
1799 * asan.c (asan_emit_stack_protection): Force the base to align to
1800 appropriate bits if STRICT_ALIGNMENT. Set shadow_mem align to
1801 appropriate bits if STRICT_ALIGNMENT.
1802 * cfgexpand.c (expand_stack_vars): Set base_align appropriately
1803 when asan is on.
1804 (expand_used_vars): Leave a space in the stack frame for alignment
1805 if STRICT_ALIGNMENT.
1806
7455080c
DM
18072014-04-21 David Malcolm <dmalcolm@redhat.com>
1808
1809 * gimple.h (gimple_assign_single_p): Accept a const_gimple rather
1810 than a gimple.
1811 (gimple_store_p): Likewise.
1812 (gimple_assign_load_p): Likewise.
1813 (gimple_assign_cast_p): Likewise.
1814 (gimple_clobber_p): Likewise.
1815
1816 * doc/gimple.texi (gimple_assign_cast_p): Accept a const_gimple
1817 rather than a gimple.
1818 (gimple_assign_cast_p): Likewise.
1819
aeb9579a
MM
18202014-04-21 Michael Meissner <meissner@linux.vnet.ibm.com>
1821
1822 PR target/60735
1823 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64 case):
1824 If mode is DDmode and TARGET_E500_DOUBLE allow move.
1825
1826 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print some
1827 more debug information for E500 if -mdebug=reg.
1828
b15d92bf
UB
18292014-04-21 Uros Bizjak <ubizjak@gmail.com>
1830
1831 PR target/60909
1832 * config/i386/i386.c (ix86_expand_builtin)
1833 <case IX86_BUILTIN_RDRAND{16,32,64}_STEP>: Use temporary
1834 register for target RTX.
1835 <case IX86_BUILTIN_RDSEED{16,32,64}_STEP>: Ditto.
1836
d367387c
CH
18372014-04-18 Cong Hou <congh@google.com>
1838
1839 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): Enhance
b15d92bf
UB
1840 the widen-mult pattern by handling two operands with different sizes,
1841 and operands whose size is smaller than half of the result type.
d367387c 1842
b6d627e4
JH
18432014-04-18 Jan Hubicka <hubicka@ucw.cz>
1844
1845 * ipa-inline.h (INLINE_HINT_known_hot): New hint.
1846 * ipa-inline-analysis.c (dump_inline_hints): Dump it.
1847 (do_estimate_edge_time): Compute it.
1848 * ipa-inline.c (want_inline_small_function_p): Bypass
b15d92bf 1849 INLINE_INSNS_AUTO/SINGLE limits for calls that are known to be hot.
b6d627e4 1850
e86a910f
JH
18512014-04-18 Jan Hubicka <hubicka@ucw.cz>
1852
1853 * ipa-inline.c (spec_rem): New static variable.
1854 (dump_overall_stats): New function.
1855 (dump_inline_stats): New function.
1856
8a3a7e67
RH
18572014-04-18 Richard Henderson <rth@redhat.com>
1858
1859 * config/aarch64/aarch64.c (aarch64_register_move_cost): Pass a mode
1860 to GET_MODE_SIZE, not a reg_class_t.
1861
ed79f4d0
BS
18622014-04-18 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1863
b15d92bf 1864 * config/rs6000/vsx.md (vsx_xxmrghw_<mode>): Adjust for little-endian.
ed79f4d0
BS
1865 (vsx_xxmrglw_<mode>): Likewise.
1866
fe799eea
MM
18672014-04-17 Michael Meissner <meissner@linux.vnet.ibm.com>
1868
1869 PR target/60876
1870 * config/rs6000/rs6000.c (rs6000_setup_reg_addr_masks): Make sure
b15d92bf 1871 GET_MODE_SIZE gets passed an enum machine_mode type and not integer.
fe799eea
MM
1872 (rs6000_init_hard_regno_mode_ok): Likewise.
1873
5e750dc6
JH
18742014-04-17 Jan Hubicka <hubicka@ucw.cz>
1875
1876 * ipa-inline.c (inline_small_functions): Account only non-cold
1877 functions.
1878 * doc/invoke.texi (inline-unit-growth): Update documentation.
1879
32337f10
PH
18802014-04-17 Pat Haugen <pthaugen@us.ibm.com>
1881
1882 * config/rs6000/rs6000.md (addti3, subti3): New.
1883
40ed344a
L
18842014-04-17 H.J. Lu <hongjiu.lu@intel.com>
1885
1886 PR target/60863
1887 * config/i386/i386.c (ix86_expand_clear): Remove outdated
1888 comment. Check optimize_insn_for_size_p instead of
1889 optimize_insn_for_speed_p.
1890
104cb50b
MJ
18912014-04-17 Martin Jambor <mjambor@suse.cz>
1892
1893 * gimple-iterator.c (gsi_start_edge): New function.
1894 * gimple-iterator.h (gsi_start_edge): Declare.
1895 * tree-sra.c (single_non_eh_succ): New function.
1896 (disqualify_ops_if_throwing_stmt): Renamed to
1897 disqualify_if_bad_bb_terminating_stmt. Allow throwing statements
1898 having one non-EH successor BB.
1899 (sra_modify_expr): If stmt ends bb, use single non-EH successor to
1900 generate loads into replacements.
1901 (sra_modify_assign): Likewise and and also use the simple path for
1902 such statements.
1903 (sra_modify_function_body): Commit statements on edges.
1904
b94970bc
RB
19052014-04-17 Richard Biener <rguenther@suse.de>
1906
1907 PR middle-end/60849
1908 * tree-ssa-propagate.c (valid_gimple_rhs_p): Allow vector
1909 comparison results and add clarifying comment.
1910
92f0f3ec
JJ
19112014-04-17 Jakub Jelinek <jakub@redhat.com>
1912
1913 * genmodes.c (struct mode_data): Add need_bytesize_adj field.
1914 (blank_mode): Initialize it.
1915 (emit_mode_size_inline, emit_mode_nunits_inline,
1916 emit_mode_inner_inline): New functions.
1917 (emit_insn_modes_h): Call them and surround their output with
1918 #if GCC_VERSION >= 4001 ... #endif.
1919 * machmode.h (GET_MODE_SIZE, GET_MODE_NUNITS, GET_MODE_INNER):
1920 For GCC_VERSION >= 4001 use mode_*_inline routines instead of
1921 mode_* arrays if the argument is __builtin_constant_p.
1922 * lower-subreg.c (dump_choices): Make sure GET_MODE_SIZE argument
1923 is enum machine_mode.
1924
be55bfe6
TS
19252014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1926
1927 * passes.c (opt_pass::execute): Adjust.
1928 (pass_manager::execute_pass_mode_switching): Likewise.
1929 (early_local_passes::execute): Likewise.
1930 (execute_one_pass): Pass cfun to the pass's execute method.
1931 * tree-pass.h (opt_pass::execute): Add function * argument.
1932 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
1933 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
1934 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
1935 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1936 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
1937 config/sparc/sparc.c, cprop.c, dce.c, df-core.c, dse.c, dwarf2cfi.c,
1938 except.c, final.c, function.c, fwprop.c, gcse.c, gimple-low.c,
1939 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1940 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1941 ipa-inline-analysis.c, ipa-inline.c, ipa-profile.c, ipa-pure-const.c,
1942 ipa-reference.c, ipa-split.c, ipa.c, ira.c, jump.c, loop-init.c,
1943 lower-subreg.c, mode-switching.c, omp-low.c, postreload-gcse.c,
1944 postreload.c, predict.c, recog.c, ree.c, reg-stack.c, regcprop.c,
1945 reginfo.c, regrename.c, reorg.c, sched-rgn.c, stack-ptr-mod.c,
1946 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
1947 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
1948 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
1949 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-ssa-ccp.c,
1950 tree-ssa-copy.c, tree-ssa-copyrename.c, tree-ssa-dce.c,
1951 tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
1952 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
1953 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
1954 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
1955 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
1956 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
1957 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
1958 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
1959 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
1960 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
1961 Adjust.
1962
1a3d085c
TS
19632014-04-17 Trevor Saunders <tsaunders@mozilla.com>
1964
1965 * passes.c (opt_pass::gate): Take function * argument.
1966 (gate_all_early_local_passes): Merge into
1967 (early_local_passes::gate): this.
1968 (gate_all_early_optimizations): Merge into
1969 (all_early_optimizations::gate): this.
1970 (gate_all_optimizations): Mege into
1971 (all_optimizations::gate): this.
1972 (gate_all_optimizations_g): Merge into
1973 (all_optimizations_g::gate): this.
1974 (gate_rest_of_compilation): Mege into
1975 (rest_of_compilation::gate): this.
1976 (gate_postreload): Merge into
1977 (postreload::gate): this.
1978 (dump_one_pass): Pass cfun to the pass's gate method.
1979 (execute_ipa_summary_passes): Likewise.
1980 (execute_one_pass): Likewise.
1981 (ipa_write_summaries_2): Likewise.
1982 (ipa_write_optimization_summaries_1): Likewise.
1983 (ipa_read_summaries_1): Likewise.
1984 (ipa_read_optimization_summaries_1): Likewise.
1985 (execute_ipa_stmt_fixups): Likewise.
1986 * tree-pass.h (opt_pass::gate): Add function * argument.
1987 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c,
1988 combine-stack-adj.c, combine.c, compare-elim.c,
1989 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
1990 config/rl78/rl78.c, config/sh/sh_optimize_sett_clrt.cc,
1991 config/sh/sh_treg_combine.cc, config/sparc/sparc.c, cprop.c, cse.c,
1992 dce.c, df-core.c, dse.c, dwarf2cfi.c, except.c, fwprop.c, gcse.c,
1993 gimple-ssa-isolate-paths.c, gimple-ssa-strength-reduction.c,
1994 graphite.c, ifcvt.c, init-regs.c, ipa-cp.c, ipa-devirt.c,
1995 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
1996 loop-init.c, lower-subreg.c, mode-switching.c, modulo-sched.c,
1997 omp-low.c, postreload-gcse.c, postreload.c, predict.c, recog.c, ree.c,
1998 reg-stack.c, regcprop.c, regrename.c, reorg.c, sched-rgn.c,
1999 store-motion.c, tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c,
2000 tree-cfgcleanup.c, tree-complex.c, tree-eh.c, tree-emutls.c,
2001 tree-if-conv.c, tree-into-ssa.c, tree-loop-distribution.c,
2002 tree-nrv.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2003 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2004 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2005 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2006 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2007 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2008 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2009 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2010 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2011 tree-ssa.c, tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
2012 tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c, tsan.c, ubsan.c,
2013 var-tracking.c, vtable-verify.c, web.c: Adjust.
2014
45887209
TS
20152014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2016
2017 * configure.ac: Check for -Woverloaded-virtual and enable it if found.
2018 * configure: Regenerate.
2019
0481253d 20202014-04-17 Trevor Saunders <tsaunders@mozilla.com>
b15d92bf 2021
0481253d
TS
2022 * passes.c (dump_one_pass): don't check pass->has_gate.
2023 (execute_ipa_summary_passes): Likewise.
2024 (execute_one_pass): Likewise.
2025 (ipa_write_summaries_2): Likewise.
2026 (ipa_write_optimization_summaries_1): Likewise.
2027 (ipa_read_optimization_summaries_1): Likewise.
2028 (execute_ipa_stmt_fixups): Likewise.
2029 * tree-pass.h (pass_data::has_gate): Remove.
2030 * asan.c, auto-inc-dec.c, bb-reorder.c, bt-load.c, cfgcleanup.c,
2031 cfgexpand.c, cfgrtl.c, cgraphbuild.c, combine-stack-adj.c, combine.c,
2032 compare-elim.c, config/arc/arc.c, config/epiphany/mode-switch-use.c,
2033 config/epiphany/resolve-sw-modes.c, config/i386/i386.c,
2034 config/mips/mips.c, config/rl78/rl78.c, config/s390/s390.c,
2035 config/sh/sh_optimize_sett_clrt.cc, config/sh/sh_treg_combine.cc,
2036 config/sparc/sparc.c, cprop.c, cse.c, dce.c, df-core.c, dse.c,
2037 dwarf2cfi.c, except.c, final.c, function.c, fwprop.c, gcse.c,
2038 gimple-low.c, gimple-ssa-isolate-paths.c,
2039 gimple-ssa-strength-reduction.c, graphite.c, ifcvt.c, init-regs.c,
2040 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2041 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa.c,
2042 ira.c, jump.c, loop-init.c, lower-subreg.c, mode-switching.c,
2043 modulo-sched.c, omp-low.c, postreload-gcse.c, postreload.c, predict.c,
2044 recog.c, ree.c, reg-stack.c, regcprop.c, reginfo.c, regrename.c,
2045 reorg.c, sched-rgn.c, stack-ptr-mod.c, store-motion.c, tracer.c,
2046 trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
2047 tree-complex.c, tree-eh.c, tree-emutls.c, tree-if-conv.c,
2048 tree-into-ssa.c, tree-loop-distribution.c, tree-nrv.c,
2049 tree-object-size.c, tree-parloops.c, tree-predcom.c, tree-profile.c,
2050 tree-sra.c, tree-ssa-ccp.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
2051 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
2052 tree-ssa-ifcombine.c, tree-ssa-loop-ch.c, tree-ssa-loop-im.c,
2053 tree-ssa-loop-ivcanon.c, tree-ssa-loop-prefetch.c,
2054 tree-ssa-loop-unswitch.c, tree-ssa-loop.c, tree-ssa-math-opts.c,
2055 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
2056 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
2057 tree-ssa-structalias.c, tree-ssa-uncprop.c, tree-ssa-uninit.c,
2058 tree-ssa.c, tree-ssanames.c, tree-stdarg.c, tree-switch-conversion.c,
2059 tree-tailcall.c, tree-vect-generic.c, tree-vectorizer.c, tree-vrp.c,
2060 tree.c, tsan.c, ubsan.c, var-tracking.c, vtable-verify.c, web.c:
2061 Adjust.
2062
a23c217d
TS
20632014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2064
b15d92bf
UB
2065 * pass_manager.h (pass_manager::register_dump_files_1): Remove
2066 declaration.
a23c217d
TS
2067 * passes.c (pass_manager::register_dump_files_1): Merge into
2068 (pass_manager::register_dump_files): this, and remove its handling of
2069 properties since the pass always has the properties anyway.
2070 (pass_manager::pass_manager): Adjust.
2071
0cd11b40
TS
20722014-04-17 Trevor Saunders <tsaunders@mozilla.com>
2073
2074 * pass_manager.h (pass_manager::register_dump_files_1): Adjust.
2075 * passes.c (pass_manager::register_dump_files_1): Remove dead code
2076 dealing with properties.
2077 (pass_manager::register_dump_files): Adjust.
2078
79896351
MW
20792014-03-20 Mark Wielaard <mjw@redhat.com>
2080
2081 * dwarf2out.c (add_bound_info): If HOST_WIDE_INT is big enough,
2082 then represent the bound as normal constant value.
2083
5a65129e
JJ
20842014-04-17 Jakub Jelinek <jakub@redhat.com>
2085
2086 PR target/60847
2087 Forward port from 4.8 branch
2088 2013-07-19 Kirill Yukhin <kirill.yukhin@intel.com>
2089
2090 * config/i386/bmiintrin.h (_blsi_u32): New.
2091 (_blsi_u64): Ditto.
2092 (_blsr_u32): Ditto.
2093 (_blsr_u64): Ditto.
2094 (_blsmsk_u32): Ditto.
2095 (_blsmsk_u64): Ditto.
2096 (_tzcnt_u32): Ditto.
2097 (_tzcnt_u64): Ditto.
2098
59b266b1
KC
20992014-04-17 Kito Cheng <kito@0xlab.org>
2100
2101 * gcc.c (used_arg): Prevent out of bound access for multilib_options.
2102
83ad208e
RB
21032014-04-17 Richard Biener <rguenther@suse.de>
2104
2105 PR middle-end/60849
2106 * tree-ssa-propagate.c (valid_gimple_rhs_p): Only allow effective
2107 boolean results for comparisons.
2108
730e78b0
RB
21092014-04-17 Richard Biener <rguenther@suse.de>
2110
2111 PR tree-optimization/60836
2112 * tree-vect-loop.c (vect_create_epilog_for_reduction): Force
2113 initial PHI args to be gimple values.
2114
1428105c
RB
21152014-04-17 Richard Biener <rguenther@suse.de>
2116
2117 PR tree-optimization/60841
2118 * tree-vect-data-refs.c (vect_analyze_data_refs): Count stmts.
2119 * tree-vect-loop.c (vect_analyze_loop_2): Pass down number
2120 of stmts to SLP build.
2121 * tree-vect-slp.c (vect_slp_analyze_bb_1): Likewise.
2122 (vect_analyze_slp): Likewise.
2123 (vect_analyze_slp_instance): Likewise.
2124 (vect_build_slp_tree): Limit overall SLP tree growth.
2125 * tree-vectorizer.h (vect_analyze_data_refs,
2126 vect_analyze_slp): Adjust prototypes.
2127
170c1776
ES
21282014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2129
2130 * config/i386/i386.c (x86_add_stmt_cost): Fix vector cost model for
2131 Silvermont.
2132
a4ef7f3e
ES
21332014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2134
2135 * config/i386/x86-tune.def (TARGET_SLOW_PSHUFB): New tune definition.
2136 * config/i386/i386.h (TARGET_SLOW_PSHUFB): New tune flag.
2137 * config/i386/i386.c (expand_vec_perm_even_odd_1): Avoid byte shuffles
2138 for TARGET_SLOW_PSHUFB
2139
cf055f6e
ES
21402014-04-17 Evgeny Stupachenko <evstupac@gmail.com>
2141
2142 * config/i386/i386.c (slm_cost): Adjust vec_to_scalar_cost.
2143 * config/i386/i386.c (intel_cost): Ditto.
2144
eeee2277
JY
21452014-04-17 Joey Ye <joey.ye@arm.com>
2146
2147 * opts.c (OPT_fif_conversion, OPT_fif_conversion2): Disable for Og.
2148
1c58fe29
JH
21492014-04-16 Jan Hubicka <hubicka@ucw.cz>
2150
2151 * opts.c (common_handle_option): Disable -fipa-reference coorectly
2152 with -fuse-profile.
2153
2d1644bf
JH
21542014-04-16 Jan Hubicka <hubicka@ucw.cz>
2155
2156 * ipa-devirt.c (odr_type_d): Add field all_derivations_known.
2157 (type_all_derivations_known_p): New predicate.
2158 (type_all_ctors_visible_p): New predicate.
2159 (type_possibly_instantiated_p): New predicate.
2160 (get_odr_type): Compute all_derivations_known.
2161 (dump_odr_type): Dump the flag.
2162 (maybe_record_type): Cleanup.
2163 (record_target_from_binfo): Add bases_to_consider array;
2164 record bases for types w/o instances and skip CXX destructor.
2165 (possible_polymorphic_call_targets_1): Add bases_to_consider
b15d92bf 2166 and consider_construction parameters; check if type may have instance.
2d1644bf
JH
2167 (get_polymorphic_call_info): Set maybe_in_construction to true
2168 when we know nothing.
2169 (record_targets_from_bases): Skip CXX destructors; they are
2170 never called for types in construction.
2171 (possible_polymorphic_call_targets): Do not record target when
2172 type may not have instance.
2173
789c2741
JH
21742014-04-16 Jan Hubicka <hubicka@ucw.cz>
2175
2176 PR ipa/60854
2177 * ipa.c (symtab_remove_unreachable_nodes): Mark targets of
2178 external aliases alive, too.
2179
9950a4f2
AP
21802014-04-16 Andrew Pinski <apinski@cavium.com>
2181
2182 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Change aarch64 ilp32
b15d92bf 2183 definition.
9950a4f2 2184
82b9c015
EB
21852014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2186
2187 * final.c (compute_alignments): Do not apply loop alignment to a block
2188 falling through to the exit.
2189
d2dfaca3
CM
21902014-04-16 Catherine Moore <clm@codesourcery.com>
2191
2192 * mips.md (*mov<mode>_internal, *movhi_internal, *movqi_internal):
2193 Adjust constraints for microMIPS store patterns.
2194
10353a79
PS
21952014-04-16 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2196
2197 * config/avr/avr-mcus.def: Correct typo for atxmega256a3bu macro.
2198
61c7fb30
EB
21992014-04-16 Eric Botcazou <ebotcazou@adacore.com>
2200
2201 * tree-ssa-operands.c (create_vop_var): Set DECL_IGNORED_P.
2202 (append_use): Run at -O0.
2203 (append_vdef): Likewise.
2204 * tree-ssa-ter.c (ter_is_replaceable_p): Do not special-case -O0.
2205 * tree-ssa-uninit.c (warn_uninitialized_vars): Remove obsolete comment.
2206
42fae17c
JJ
22072014-04-16 Jakub Jelinek <jakub@redhat.com>
2208
2209 PR tree-optimization/60844
2210 * tree-ssa-reassoc.c (reassoc_remove_stmt): New function.
2211 (propagate_op_to_single_use, remove_visited_stmt_chain,
2212 linearize_expr, repropagate_negates, reassociate_bb): Use it
2213 instead of gsi_remove.
2214
ca860d03
MJ
22152014-04-16 Martin Jambor <mjambor@suse.cz>
2216
2217 * cgraphclones.c (cgraph_create_virtual_clone): Duplicate
2218 ipa_transforms_to_apply.
2219 (cgraph_function_versioning): Assert that old_node has empty
2220 ipa_transforms_to_apply.
2221 * trans-mem.c (ipa_tm_create_version): Likewise.
2222 * tree-inline.c (tree_function_versioning): Do not duplicate
2223 ipa_transforms_to_apply.
2224
fe4e71e4
RO
22252014-04-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2226
2227 PR target/60817
2228 * configure.ac (set_have_as_tls): Merge i[34567]86-*-* and
2229 x86_64-*-* cases.
2230 Pass necessary as flags on 64-bit Solaris/x86.
2231 Use lowercase relocs for x86_64-*-*.
2232 * configure: Regenerate.
2233
ccb05ef2
JH
22342014-04-15 Jan Hubicka <jh@suse.cz>
2235
2236 * ipa-devirt.c (referenced_from_vtable_p): New predicate.
2237 (maybe_record_node, likely_target_p): Use it.
2238
dfe449d1
BS
22392014-04-15 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2240
2241 PR target/60839
2242 Revert following patch
2243
2244 2014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2245
2246 PR target/60735
2247 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2248 software floating point or no floating point registers, do not
2249 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2250 in GPRs that occurs after we tested for GPRs that would never be
2251 true.
2252
2253 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2254 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2255 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2256 specifically allow DDmode, since that does not use the SPE SIMD
2257 instructions.
2258
25dce5c6
MW
22592014-03-21 Mark Wielaard <mjw@redhat.com>
2260
2261 * dwarf2out.c (gen_enumeration_type_die): Add DW_AT_const_value
2262 as unsigned or int depending on type and value used.
2263
8d3c076f
RB
22642014-04-15 Richard Biener <rguenther@suse.de>
2265
2266 PR rtl-optimization/56965
2267 * alias.c (ncr_compar, nonoverlapping_component_refs_p): Move ...
2268 * tree-ssa-alias.c (ncr_compar, nonoverlapping_component_refs_p):
2269 ... here.
2270 * alias.c (true_dependence_1): Do not call
2271 nonoverlapping_component_refs_p.
2272 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Call
2273 nonoverlapping_component_refs_p.
2274 (indirect_refs_may_alias_p): Likewise.
2275
d00e4baa
TJ
22762014-04-15 Teresa Johnson <tejohnson@google.com>
2277
2278 * cfg.c (dump_bb_info): Fix flags check.
2279 * tree-cfg.c (remove_bb): Only dump TDF_BLOCKS when removing.
2280
98a2fdfb
KT
22812014-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2282
2283 PR rtl-optimization/60663
2284 * config/arm/arm.c (arm_new_rtx_costs): Improve ASM_OPERANDS case,
2285 avoid 0 cost.
2286
b9ed2c2c
RB
22872014-04-15 Richard Biener <rguenther@suse.de>
2288
2289 * lto-streamer.h (LTO_major_version): Bump to 4.
2290
783dab6b
RB
22912014-04-15 Richard Biener <rguenther@suse.de>
2292
2293 * common.opt (lto_partition_model): New enum.
2294 (flto-partition=): Merge separate options with a single with argument,
2295 add -flto-partition=one support.
2296 * flag-types.h (enum lto_partition_model): Declare.
2297 * opts.c (finish_options): Remove duplicate -flto-partition=
2298 option check.
2299 * lto-wrapper.c (run_gcc): Adjust.
2300
d8e4ce6d
RB
23012014-04-15 Richard Biener <rguenther@suse.de>
2302
2303 * alias.c (ncr_compar): New function.
2304 (nonoverlapping_component_refs_p): Re-implement in O (n log n).
2305
1999fd7d
RB
23062014-04-15 Richard Biener <rguenther@suse.de>
2307
2308 * alias.c (record_component_aliases): Do not walk BINFOs.
2309
628169e0
RB
23102014-04-15 Richard Biener <rguenther@suse.de>
2311
2312 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
2313 Add struct function argument and adjust.
2314 (find_func_aliases_for_call): Likewise.
2315 (find_func_aliases): Likewise.
2316 (find_func_clobbers): Likewise.
2317 (intra_create_variable_infos): Likewise.
2318 (compute_points_to_sets): Likewise.
2319 (ipa_pta_execute): Adjust. Do not push/pop cfun.
2320
349e6210
RB
23212014-04-15 Richard Biener <rguenther@suse.de>
2322
2323 * tree.c (iterative_hash_expr): Use enum tree_code_class
2324 to store TREE_CODE_CLASS.
2325 (tree_block): Likewise.
2326 (tree_set_block): Likewise.
2327 * tree.h (fold_build_pointer_plus_loc): Use
2328 convert_to_ptrofftype_loc.
2329
4f0ddb6f
JJ
23302014-04-15 Jakub Jelinek <jakub@redhat.com>
2331
2332 PR plugins/59335
2333 * Makefile.in (PLUGIN_HEADERS): Add various headers that have been
2334 added in 4.9.
2335
718c4601
EB
23362014-04-15 Eric Botcazou <ebotcazou@adacore.com>
2337
2338 * cfgloop.h (struct loop): Move force_vectorize down.
2339 * gimplify.c (gimple_boolify) <ANNOTATE_EXPR>: Handle new kinds.
2340 (gimplify_expr) <ANNOTATE_EXPR>: Minor tweak.
2341 * lto-streamer-in.c (input_cfg): Read dont_vectorize field.
2342 * lto-streamer-out.c (output_cfg): Write dont_vectorize field.
2343 * tree-cfg.c (replace_loop_annotate): Revamp and handle new kinds.
2344 * tree-core.h (enum annot_expr_kind): Add new kind values.
2345 * tree-inline.c (copy_loops): Copy dont_vectorize field and reorder.
2346 * tree-pretty-print.c (dump_generic_node) <ANNOTATE_EXPR>: Handle new
2347 kinds.
2348 * tree.def (ANNOTATE_EXPR): Tweak comment.
2349
88f592e3
JH
23502014-04-14 Jan Hubicka <hubicka@ucw.cz>
2351
2352 * ipa-devirt.c (maybe_record_node): Ignore all non-methods (including
2353 cxa_pure_virtual).
2354
9dba4b55
PC
23552014-04-14 Paolo Carlini <paolo.carlini@oracle.com>
2356
2357 * tree.h (TYPE_IDENTIFIER): Declare.
2358 * tree.c (subrange_type_for_debug_p): Use it.
2359 * godump.c (go_format_type): Likewise.
2360 * dwarf2out.c (is_cxx_auto, modified_type_die,
2361 gen_type_die_with_usage, gen_type_die_with_usage): Likewise.
2362 * dbxout.c (dbxout_type, dbxout_symbol): Likewise.
2363
006202e8
JH
23642014-04-14 Jan Hubicka <hubicka@ucw.cz>
2365
2366 PR lto/60820
2367 * varpool.c (varpool_remove_node): Do not alter decls when streaming.
2368
e0a81db1
UB
23692014-04-14 Uros Bizjak <ubizjak@gmail.com>
2370
2371 * config/i386/i386.c (examine_argument): Return bool. Return true if
2372 parameter should be passed in memory.
2373 <case X86_64_COMPLEX_X87_CLASS>: Adjust.
2374 (construct_container): Update calls to examine_argument.
2375 (function_arg_advance_64): Ditto.
2376 (return_in_memory_32): Merge with ix86_return_in_memory.
2377 (return_in_memory_64): Ditto.
2378 (return_in_memory_ms_64): Ditto.
2379
cb90235d
JH
23802014-04-14 Jan Hubicka <hubicka@ucw.cz>
2381
2382 * ipa-utils.c (ipa_merge_profiles): Merge profile_id.
2383 * coverage.c (coverage_compute_profile_id): Handle externally visible
2384 symbols.
2385
7b3b340e
MJ
23862014-04-14 Martin Jambor <mjambor@suse.cz>
2387
2388 * tree-sra.c (ipa_sra_preliminary_function_checks): Skip
2389 DECL_DISREGARD_INLINE_LIMITS functions.
2390
2ed3b4ee
L
23912014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2392
2393 PR target/60827
e0a81db1 2394 * config/i386/i386.md (*fixuns_trunc<mode>_1): Revert the last change.
2ed3b4ee 2395
10fa463f
L
23962014-04-14 H.J. Lu <hongjiu.lu@intel.com>
2397
2398 PR target/60827
2399 * config/i386/i386.md (*fixuns_trunc<mode>_1): Check
2400 optimize_insn_for_speed_p instead of
2401 optimize_function_for_speed_p.
2402
b621e875
YZ
24032014-04-14 Yufeng Zhang <yufeng.zhang@arm.com>
2404
2405 * doc/invoke.texi (free): Document AArch64.
2406
cddaefa3
RB
24072014-04-14 Richard Biener <rguenther@suse.de>
2408
2409 PR tree-optimization/60042
2410 * tree-ssa-pre.c (inhibit_phi_insertion): Remove.
2411 (insert_into_preds_of_block): Do not prevent PHI insertion
2412 for REFERENCE exprs here ...
2413 (eliminate_dom_walker::before_dom_children): ... but prevent
2414 their use here under similar conditions when applied to the
2415 IL after PRE optimizations.
2416
a9e4c82f
RB
24172014-04-14 Richard Biener <rguenther@suse.de>
2418
2419 * passes.def: Move early points-to after early SRA.
2420
7b4cae1b
RB
24212014-04-14 Richard Biener <rguenther@suse.de>
2422
2423 * tree-ssa-forwprop.c (simplify_gimple_switch): Enhance
2424 check for which sign-changes we allow when forwarding
2425 a converted value into a switch.
2426
cb27986c
EB
24272014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2428
2429 * stor-layout.c (place_field): Finalize non-constant offset for the
2430 field, if any.
2431
72798784
RB
24322014-04-14 Richard Biener <rguenther@suse.de>
2433
2434 * tree-switch-conversion.c (lshift_cheap_p): Get speed_p
2435 as argument.
2436 (expand_switch_using_bit_tests_p): Likewise.
2437 (process_switch): Compute and pass on speed_p based on the
2438 switch stmt.
2439 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Use
2440 optimize_bb_for_speed_p.
2441
b15b5979
EB
24422014-04-14 Eric Botcazou <ebotcazou@adacore.com>
2443
2444 * cfgloop.h (struct loop): Rename force_vect into force_vectorize.
2445 * function.h (struct function): Rename has_force_vect_loops into
2446 has_force_vectorize_loops.
2447 * lto-streamer-in.c (input_cfg): Adjust for renaming.
2448 (input_struct_function_base): Likewise.
2449 * lto-streamer-out.c (output_cfg): Likewise.
2450 (output_struct_function_base): Likewise.
2451 * omp-low.c (expand_omp_simd): Likewise.
2452 * tree-cfg.c (move_sese_region_to_fn): Likewise.
2453 * tree-if-conv.c (ifcvt_can_use_mask_load_store): Likewise.
2454 (version_loop_for_if_conversion): Likewise.
2455 (tree_if_conversion): Likewise.
2456 (main_tree_if_conversion): Likewise.
2457 (gate_tree_if_conversion): Likewise.
2458 * tree-inline.c (copy_loops): Likewise.
2459 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Likewise.
2460 * tree-ssa-loop.c (tree_loop_vectorize): Likewise.
2461 * tree-ssa-phiopt.c (tree_ssa_phiopt_worker): Likewise.
2462 * tree-vect-loop.c (vect_estimate_min_profitable_iters): Likewise.
2463 * tree-vectorizer.c (vectorize_loops): Likewise.
2464 * tree-vectorizer.h (unlimited_cost_model): Likewise.
2465
8359c87e
RB
24662014-04-14 Richard Biener <rguenther@suse.de>
2467
2468 PR lto/60720
2469 * lto-streamer-out.c (wrap_refs): New function.
2470 (lto_output): Wrap symbol references in global initializes in
2471 type-preserving MEM_REFs.
2472
9447df74
CB
24732014-04-14 Christian Bruel <christian.bruel@st.com>
2474
b15d92bf 2475 * config/sh/sh-mem.cc (sh_expand_strlen): Unroll last word.
9447df74 2476
fa1aecc1
CB
24772014-04-14 Christian Bruel <christian.bruel@st.com>
2478
2479 * config/sh/sh.md (setmemqi): New expand pattern.
2480 * config/sh/sh.h (CLEAR_RATIO): Define.
2481 * config/sh/sh-mem.cc (sh_expand_setmem): Define.
2482 * config/sh/sh-protos.h (sh_expand_setmem): Declare.
2483
a3c77ce9
RB
24842014-04-14 Richard Biener <rguenther@suse.de>
2485
2486 PR middle-end/55022
2487 * fold-const.c (negate_expr_p): Don't negate directional rounding
2488 division.
2489 (fold_negate_expr): Likewise.
2490
033aa406
RB
24912014-04-14 Richard Biener <rguenther@suse.de>
2492
2493 PR tree-optimization/59817
2494 PR tree-optimization/60453
2495 * graphite-scop-detection.c (graphite_can_represent_scev): Complete
2496 recursion to catch all CHRECs in the scalar evolution and restrict
2497 the predicate for the remains appropriately.
2498
5e7d8b4c
RS
24992014-04-12 Catherine Moore <clm@codesourcery.com>
2500
2501 * config/mips/constraints.md: Add new register constraint "kb".
2502 * config/mips/mips.md (*mov<mode>_internal): Use constraint "kb".
2503 (*movhi_internal): Likewise.
2504 (*movqi_internal): Likewise.
2505 * config/mips/mips.h (M16_STORE_REGS): New register class.
2506 (REG_CLASS_NAMES): Add M16_STORE_REGS.
2507 (REG_CLASS_CONTENTS): Likewise.
2508 * config/mips/mips.c (mips_regno_to_class): Add M16_STORE_REGS.
2509
cbbd2b1c
TB
25102014-04-11 Tobias Burnus <burnus@net-b.de>
2511
2512 PR c/60194
2513 * doc/invoke.texi (-Wformat-signedness): Document it.
2514 (Wformat=2): Mention that this enables -Wformat-signedness.
2515
d9bb5800
JR
25162014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2517
2518 * common/config/epiphany/epiphany-common.c
2519 (epiphany_option_optimization_table): Enable section anchors by
2520 default at -O1 or higher.
2521 * config/epiphany/epiphany.c (TARGET_MAX_ANCHOR_OFFSET): Define.
2522 (TARGET_MIN_ANCHOR_OFFSET): Likewise.
2523 (epiphany_rtx_costs) <SET>: For binary operators, the set as such
2524 carries no extra cost.
2525 (epiphany_legitimate_address_p): For BLKmode, apply SImode check.
2526 * config/epiphany/epiphany.h (ASM_OUTPUT_DEF): Define.
2527 * config/epiphany/predicates.md (memclob_operand): New predicate.
2528 * config/epiphany/epiphany.md (stack_adjust_add, stack_adjust_str):
2529 Use memclob_operand predicate and X constraint for operand 3.
2530
b5d0ecea
JR
25312014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2532
2533 * config/epiphany/epiphany.c (epiphany_rtx_cost): Compare
2534 with CC_N_NE / CC_C_LTU / CC_C_GTU carries no extra cost for
2535 its operands.
2536
473fd99a
JR
25372014-04-11 Joern Rennecke <joern.rennecke@embecosm.com>
2538
2539 PR rtl-optimization/60651
2540 * mode-switching.c (optimize_mode_switching): Make sure to emit
2541 sets of a lower numbered entity before sets of a higher numbered
2542 entity to a mode of the same or lower priority.
2543 When creating a seginfo for a basic block that starts with a code
2544 label, move the insertion point past the code label.
2545 (new_seginfo): Document and enforce requirement that
2546 NOTE_INSN_BASIC_BLOCK only appears for empty blocks.
2547 * doc/tm.texi.in: Document ordering constraint for emitted mode sets.
2548 * doc/tm.texi: Regenerate.
2549
6ace1161
JR
25502014-01-11 Joern Rennecke <joern.rennecke@embecosm.com>
2551
2552 PR target/60811
2553 * config/arc/arc.c (arc_save_restore): Fix assert typo.
2554
89eb97de
JJ
25552013-04-11 Jakub Jelinek <jakub@redhat.com>
2556
6602e7fc 2557 * BASE-VER: Set to 4.10.0.
89eb97de 2558
9d0c7214
TB
25592014-04-11 Tobias Burnus <burnus@net-b.de>
2560
2561 PR other/59055
e0a81db1 2562 * doc/bugreport.texi (Bugs): Remove nodes pointing to the nirvana.
9d0c7214
TB
2563 * doc/gcc.texi (Service): Update description in the @menu
2564 * doc/invoke.texi (Option Summary): Remove misplaced and
2565 duplicated @menu.
2566
3662b277
SE
25672014-04-11 Steve Ellcey <sellcey@mips.com>
2568 Jakub Jelinek <jakub@redhat.com>
2569
2570 PR middle-end/60556
2571 * expr.c (convert_move): Use emit_store_flag_force instead of
2572 emit_store_flag. Pass lowpart_mode instead of VOIDmode as 5th
2573 argument to it.
2574
8330537b
RB
25752014-04-11 Richard Biener <rguenther@suse.de>
2576
2577 PR middle-end/60797
2578 * varasm.c (assemble_alias): Avoid endless error reporting
2579 recursion by setting TREE_ASM_WRITTEN.
2580
3c91f126
AK
25812014-04-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2582
2583 * config/s390/s390.md: Add a splitter for NOT rtx.
2584
d8d6ea53
JJ
25852014-04-11 Jakub Jelinek <jakub@redhat.com>
2586
2587 PR rtl-optimization/60663
e0a81db1 2588 * cse.c (cse_insn): Set src_volatile on ASM_OPERANDS in PARALLEL.
d8d6ea53 2589
fa5866c0
JJ
25902014-04-10 Jan Hubicka <hubicka@ucw.cz>
2591 Jakub Jelinek <jakub@redhat.com>
073afca6 2592
fa5866c0 2593 PR lto/60567
e0a81db1
UB
2594 * ipa.c (function_and_variable_visibility): Copy forced_by_abi
2595 flag from decl_node to node.
073afca6 2596
b322c36a
RR
25972014-04-10 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2598
2599 PR debug/60655
2600 * config/arm/arm.c (TARGET_CONST_NOT_OK_FOR_DEBUG_P): Define
2601 (arm_const_not_ok_for_debug_p): Reject MINUS with SYM_REF's
2602 ameliorating the cases where it can be.
2603
8267f6bb
DE
26042014-04-09 David Edelsohn <dje.gcc@gmail.com>
2605
2606 Revert
2607 2014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2608
2609 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2610 (loadsync_<mode>): Change mode.
2611 (load_quadpti, store_quadpti): New.
2612 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2613 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2614 * config/rs6000/predicates.md (quad_memory_operand): !TARGET_SYNC_TI.
2615
5d1a5a53
CH
26162014-04-09 Cong Hou <congh@google.com>
2617
2618 PR testsuite/60773
2619 * doc/sourcebuild.texi (vect_widen_mult_si_to_di_pattern): Add
2620 documentation.
2621
da90fa4c
BS
26222014-04-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2623
2624 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Use vnand
2625 instead of vnor to exploit possible fusion opportunity in the
2626 future.
2627 (altivec_expand_vec_perm_const_le): Likewise.
2628
98e6ecab
PH
26292014-04-08 Pat Haugen <pthaugen@us.ibm.com>
2630
2631 * config/rs6000/sync.md (AINT mode_iterator): Move definition.
2632 (loadsync_<mode>): Change mode.
2633 (load_quadpti, store_quadpti): New.
2634 (atomic_load<mode>, atomic_store<mode>): Add support for TI mode.
2635 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_LSQ.
2636
8418cf33
RS
26372014-04-08 Richard Sandiford <rdsandiford@googlemail.com>
2638
2639 PR target/60763
2640 * config/rs6000/vsx.md (vsx_xscvdpspn_scalar): Change input to DImode.
2641 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Update accordingly.
2642 Use gen_rtx_REG rather than simplify_gen_subreg for op0_di.
2643
3fa68ccb
RB
26442014-04-08 Richard Biener <rguenther@suse.de>
2645
2646 PR middle-end/60706
2647 * tree-pretty-print.c (pp_double_int): For HWI32 hosts with
2648 a 64bit widest int print double-int similar to on HWI64 hosts.
2649
9038ae48
RB
26502014-04-08 Richard Biener <rguenther@suse.de>
2651
2652 PR tree-optimization/60785
2653 * graphite-sese-to-poly.c (rewrite_phi_out_of_ssa): Treat
2654 default defs properly.
2655
38f09da3
NS
26562014-04-08 Nathan Sidwell <nathan@codesourcery.com>
2657
2658 * doc/invoke (Wnon-virtual-dtor): Update to match implementation.
2659 (Weffc++): Likewise.
2660
0439a947
JH
26612014-04-07 Jan Hubicka <hubcika@ucw.cz>
2662
2663 * ipa-devirt.c (maybe_record_node): When node is not recorded,
2664 set completep to false rather than true.
2665
3f2f838e
EB
26662014-04-07 Douglas B Rupp <rupp@adacore.com>
2667
2668 PR target/60504
2669 * config/arm/arm.h (ASM_PREFERRED_EH_DATA_FORMAT): Expose from
2670 ARM_TARGET2_DWARF_FORMAT.
2671
e75c1617
CB
26722014-04-07 Charles Baylis <charles.baylis@linaro.org>
2673
2674 PR target/60609
2675 * config/arm/arm.h (ASM_OUTPUT_CASE_END): Remove.
2676 (LABEL_ALIGN_AFTER_BARRIER): Align barriers which occur after
2677 ADDR_DIFF_VEC.
2678
d6adff07
RB
26792014-04-07 Richard Biener <rguenther@suse.de>
2680
2681 PR tree-optimization/60766
3f2f838e
EB
2682 * tree-ssa-loop-ivopts.c (cand_value_at): Compute in an unsigned type.
2683 (may_eliminate_iv): Convert cand_value_at result to desired type.
d6adff07 2684
da34ade5
JM
26852014-04-07 Jason Merrill <jason@redhat.com>
2686
2687 PR c++/60731
2688 * common.opt (-fno-gnu-unique): Add.
2689 * config/elfos.h (USE_GNU_UNIQUE_OBJECT): Check it.
2690
151e9aac
KT
26912014-04-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2692
2693 * haifa-sched.c: Fix outdated function reference and minor
2694 grammar errors in introductory comment.
2695
308173e3
RB
26962014-04-07 Richard Biener <rguenther@suse.de>
2697
2698 PR middle-end/60750
2699 * tree-ssa-operands.c (maybe_add_call_vops): Also add VDEFs
2700 for noreturn calls.
2701 * tree-cfgcleanup.c (fixup_noreturn_call): Do not remove VDEFs.
2702
a9a302d9
JDA
27032014-04-06 John David Anglin <danglin@gcc.gnu.org>
2704
2705 PR debug/55794
2706 * config/pa/pa.c (pa_output_function_epilogue): Skip address and code
2707 size accounting for thunks.
2708 (pa_asm_output_mi_thunk): Use final_start_function() and
2709 final_end_function() to output function start and end directives.
2710
576ba35c
PS
27112014-04-05 Pitchumani Sivanupandi <Pitchumani.S@atmel.com>
2712
e0a81db1
UB
2713 * config/avr/avr-arch.h (avr_mcu_t): Add dev_attribute field to have
2714 device specific ISA/ feature information. Remove short_sp and
2715 errata_skip ds. Add avr_device_specific_features enum to have device
2716 specific info.
576ba35c
PS
2717 * config/avr/avr-c.c (avr_cpu_cpp_builtins): use dev_attribute to check
2718 errata_skip. Add __AVR_ISA_RMW__ builtin macro if RMW ISA available.
2719 * config/avr/avr-devices.c (avr_mcu_types): Update AVR_MCU macro for
2720 updated device specific info.
2721 * config/avr/avr-mcus.def: Merge device specific details to
2722 dev_attribute field.
2723 * config/avr/avr.c (avr_2word_insn_p): use dev_attribute field to check
2724 errata_skip.
2725 * config/avr/avr.h (AVR_HAVE_8BIT_SP): same for short sp info.
2726 * config/avr/driver-avr.c (avr_device_to_as): Pass -mrmw option to
2727 assembler if RMW isa supported by current device.
2728 * config/avr/genmultilib.awk: Update as device info structure changed.
2729 * doc/invoke.texi: Add info for __AVR_ISA_RMW__ builtin macro
2730
ebc047a2
CH
27312014-04-04 Cong Hou <congh@google.com>
2732
2733 PR tree-optimization/60656
2734 * tree-vect-stmts.c (supportable_widening_operation):
2735 Fix a bug that elements in a vector with vect_used_by_reduction
2736 property are incorrectly reordered when the operation on it is not
2737 consistant with the one in reduction operation.
2738
b6808818
JDA
27392014-04-04 John David Anglin <danglin@gcc.gnu.org>
2740
2741 PR rtl-optimization/60155
2742 * gcse.c (record_set_data): New function.
2743 (single_set_gcse): New function.
2744 (gcse_emit_move_after): Use single_set_gcse instead of single_set.
2745 (hoist_code): Likewise.
2746 (get_pressure_class_and_nregs): Likewise.
2747
eabcc725
EB
27482014-04-04 Eric Botcazou <ebotcazou@adacore.com>
2749
2750 * explow.c (probe_stack_range): Emit a final optimization blockage.
2751
e005b753
AG
27522014-04-04 Anthony Green <green@moxielogic.com>
2753
2754 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2): Fix
2755 typos.
2756
8fe91ca8
JH
27572014-04-04 Jan Hubicka <hubicka@ucw.cz>
2758
2759 PR ipa/59626
2760 * lto-cgraph.c (input_overwrite_node): Check that partitioning
2761 flags are set only during streaming.
2762 * ipa.c (process_references, walk_polymorphic_call_targets,
2763 symtab_remove_unreachable_nodes): Drop bodies of always inline
2764 after early inlining.
2765 (symtab_remove_unreachable_nodes): Remove always_inline attribute.
576ba35c 2766
88d94632
RR
27672014-04-04 Jakub Jelinek <jakub@redhat.com>
2768 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2769
2770 PR debug/60655
2771 * dwarf2out.c (const_ok_for_output_1): Reject expressions
2772 containing a NOT.
2773
094bc219
KT
27742014-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2775
2776 PR bootstrap/60743
2777 * config/arm/cortex-a53.md (cortex_a53_fdivs): Reduce reservation
2778 duration.
2779 (cortex_a53_fdivd): Likewise.
2780
610c8ef0
MJ
27812014-04-04 Martin Jambor <mjambor@suse.cz>
2782
2783 PR ipa/60640
b15d92bf
UB
2784 * cgraph.h (cgraph_clone_node): New parameter added to declaration.
2785 Adjust all callers.
610c8ef0
MJ
2786 * cgraph.c (clone_of_p): Also return true if thunks match.
2787 (verify_edge_corresponds_to_fndecl): Removed extraneous call to
2788 cgraph_function_or_thunk_node and an obsolete comment.
b15d92bf
UB
2789 * cgraphclones.c (build_function_type_skip_args): Moved upwards in the
2790 file.
2791 (build_function_decl_skip_args): Likewise.
610c8ef0 2792 (set_new_clone_decl_and_node_flags): New function.
b15d92bf
UB
2793 (duplicate_thunk_for_node): Likewise.
2794 (redirect_edge_duplicating_thunks): Likewise.
2795 (cgraph_clone_node): New parameter args_to_skip, pass it to
2796 redirect_edge_duplicating_thunks which is called instead of
2797 cgraph_redirect_edge_callee.
2798 (cgraph_create_virtual_clone): Pass args_to_skip to cgraph_clone_node,
610c8ef0
MJ
2799 moved setting of a lot of flags to set_new_clone_decl_and_node_flags.
2800
fc02bcca
JL
28012014-04-04 Jeff Law <law@redhat.com>
2802
57ef133b 2803 PR target/60657
fc02bcca
JL
2804 * config/arm/predicates.md (const_int_I_operand): New predicate.
2805 (const_int_M_operand): Similarly.
2806 * config/arm/arm.md (insv_zero): Use const_int_M_operand instead of
2807 const_int_operand.
2808 (insv_t2, extv_reg, extzv_t2): Likewise.
2809 (load_multiple_with_writeback): Similarly for const_int_I_operand.
2810 (pop_multiple_with_writeback_and_return): Likewise.
2811 (vfp_pop_multiple_with_writeback): Likewise
2812
45b62594
RB
28132014-04-04 Richard Biener <rguenther@suse.de>
2814
2815 PR ipa/60746
2816 * tree-ssanames.c (make_ssa_name_fn): Fix assert.
2817 * gimple.c (gimple_set_bb): Avoid ICEing for NULL cfun for
2818 non-GIMPLE_LABELs.
2819 * gimplify.h (gimple_add_tmp_var_fn): Declare.
2820 * gimplify.c (gimple_add_tmp_var_fn): New function.
2821 * gimple-expr.h (create_tmp_reg_fn): Declare.
2822 * gimple-expr.c (create_tmp_reg_fn): New function.
2823 * gimple-low.c (record_vars_into): Don't change cfun.
2824 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix
2825 code generation without cfun.
2826
bb609c35
TS
28272014-04-04 Thomas Schwinge <thomas@codesourcery.com>
2828
2829 PR bootstrap/60719
2830 * Makefile.in (install-driver): Fix shell scripting.
2831
090cd8dc
CH
28322014-04-03 Cong Hou <congh@google.com>
2833
2834 PR tree-optimization/60505
2835 * tree-vectorizer.h (struct _stmt_vec_info): Add th field as the
e0a81db1
UB
2836 threshold of number of iterations below which no vectorization
2837 will be done.
090cd8dc
CH
2838 * tree-vect-loop.c (new_loop_vec_info):
2839 Initialize LOOP_VINFO_COST_MODEL_THRESHOLD.
2840 * tree-vect-loop.c (vect_analyze_loop_operations):
2841 Set LOOP_VINFO_COST_MODEL_THRESHOLD.
2842 * tree-vect-loop.c (vect_transform_loop):
2843 Use LOOP_VINFO_COST_MODEL_THRESHOLD.
2844 * tree-vect-loop.c (vect_analyze_loop_2): Check the maximum number
2845 of iterations of the loop and see if we should build the epilogue.
2846
bdc67fd6
RB
28472014-04-03 Richard Biener <rguenther@suse.de>
2848
e0a81db1 2849 * tree-streamer.h (struct streamer_tree_cache_d): Add next_idx member.
bdc67fd6
RB
2850 (streamer_tree_cache_create): Adjust.
2851 * tree-streamer.c (streamer_tree_cache_add_to_node_array): Adjust
2852 to allow optional nodes array.
2853 (streamer_tree_cache_insert_1): Use next_idx to assign idx.
2854 (streamer_tree_cache_append): Likewise.
2855 (streamer_tree_cache_create): Create nodes array optionally
2856 as specified by parameter.
2857 * lto-streamer-out.c (create_output_block): Avoid maintaining
2858 the node array in the writer cache.
2859 (DFS_write_tree): Remove assertion.
e0a81db1 2860 (produce_asm_for_decls): Free the out decl state hash table early.
bdc67fd6
RB
2861 * lto-streamer-in.c (lto_data_in_create): Adjust for
2862 streamer_tree_cache_create prototype change.
2863
a0daf659
RB
28642014-04-03 Richard Biener <rguenther@suse.de>
2865
2866 * tree-streamer-out.c (streamer_write_chain): Do not temporarily
2867 set TREE_CHAIN to NULL_TREE.
2868
f16c88d2
RB
28692014-04-03 Richard Biener <rguenther@suse.de>
2870
2871 PR tree-optimization/60740
2872 * graphite-scop-detection.c (stmt_simple_for_scop_p): Iterate
2873 over all GIMPLE_COND operands.
2874
880a467b
NS
28752014-04-03 Nathan Sidwell <nathan@codesourcery.com>
2876
2877 * doc/invoke.texi (Wnon-virtual-dtor): Adjust documentation.
2878 (Weffc++): Remove Scott's numbering, merge lists and reference
2879 Wnon-virtual-dtor.
2880
08866760
NC
28812014-04-03 Nick Clifton <nickc@redhat.com>
2882
2883 * config/rl78/rl78-expand.md (movqi): Handle (SUBREG (SYMBOL_REF))
2884 properly.
2885
5bed50e8
MJ
28862014-04-03 Martin Jambor <mjambor@suse.cz>
2887
2888 * ipa-cp.c (ipcp_verify_propagated_values): Also dump symtab and
2889 mention gcc_unreachable before failing.
2890 * ipa.c (symtab_remove_unreachable_nodes): Also print order of
2891 removed symbols.
2892
3e86c6a8
JH
28932014-04-02 Jan Hubicka <hubicka@ucw.cz>
2894
2895 PR ipa/60659
e0a81db1
UB
2896 * ipa-devirt.c (get_polymorphic_call_info): Do not ICE on type
2897 inconsistent code and instead mark the context inconsistent.
3e86c6a8
JH
2898 (possible_polymorphic_call_targets): For inconsistent contexts
2899 return empty complete list.
2900
2b9912aa
AG
29012014-04-02 Anthony Green <green@moxielogic.com>
2902
2903 * config/moxie/moxie.md (zero_extendqisi2, zero_extendhisi2)
2904 (extendqisi2, extendhisi2): Define.
2905 * config/moxie/moxie.h (DEFAULT_SIGNED_CHAR): Change to 0.
2906 (WCHAR_TYPE): Change to unsigned int.
2907
1e386bb8
BS
29082014-04-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2909
2910 PR tree-optimization/60733
2911 * gimple-ssa-strength-reduction.c (ncd_with_phi): Change required
2912 insertion point for PHI candidates to be the end of the feeding
2913 block for the PHI argument.
2914
f4581282
VM
29152014-04-02 Vladimir Makarov <vmakarov@redhat.com>
2916
2917 PR rtl-optimization/60650
2918 * lra-constraints.c (process_alt_operands): Decrease reject for
2919 earlyclobber matching.
2920
a9d46e32
AK
29212014-04-02 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2922
2923 * config/s390/s390.c (s390_expand_insv): Use GET_MODE_BITSIZE.
2924
7c40228a
UW
29252014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2926
2927 * config/spu/spu.c (pad_bb): Do not crash when the last
2928 insn is CODE_FOR_blockage.
2929
c6596193
UW
29302014-04-02 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
2931
2932 * config/spu/spu.md ("insv"): Fail if bitoffset+bitsize
2933 lies outside the target mode.
2934
690c229f
MM
29352014-04-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2936
2937 PR target/60735
2938 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): If we have
2939 software floating point or no floating point registers, do not
2940 allow any type in the FPRs. Eliminate a test for SPE SIMD types
2941 in GPRs that occurs after we tested for GPRs that would never be
2942 true.
2943
2944 * config/rs6000/rs6000.md (mov<mode>_softfloat32, FMOVE64):
2945 Rewrite tests to use TARGET_DOUBLE_FLOAT and TARGET_E500_DOUBLE,
2946 since the FMOVE64 type is DFmode/DDmode. If TARGET_E500_DOUBLE,
2947 specifically allow DDmode, since that does not use the SPE SIMD
2948 instructions.
2949
dad16761
RB
29502014-04-02 Richard Biener <rguenther@suse.de>
2951
2952 PR middle-end/60729
2953 * optabs.c (expand_abs_nojump): Honor flag_trapv only for
2954 MODE_INTs. Properly use negv_optab.
2955 (expand_abs): Likewise.
2956
a1ce39be
RB
29572014-04-02 Richard Biener <rguenther@suse.de>
2958
2959 PR bootstrap/60719
2960 * Makefile.in (install-driver): Guard extra installs with special
2961 names properly.
2962
6fb619e4
MM
29632014-04-01 Michael Meissner <meissner@linux.vnet.ibm.com>
2964
2965 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
2966 Document vec_vgbbd.
2967
2a1ffd6d
RH
29682014-04-01 Richard Henderson <rth@redhat.com>
2969
2970 PR target/60704
2971 * config/i386/i386.md (*float<SWI48><MODEF>2_sse): Leave the second
2972 alternative enabled before register allocation.
2973
98e8dd4d
CLT
29742014-04-01 Chung-Lin Tang <cltang@codesourcery.com>
2975
2976 * config/nios2/nios2.md (unspec): Remove UNSPEC_TLS, UNSPEC_TLS_LDM.
2977 * config/nios2/nios2.c (nios2_function_profiler): Fix addi operand
2978 typo.
2979 (nios2_large_got_address): Remove unneeded 'sym' parameter.
2980 (nios2_got_address): Update nios2_large_got_address call site.
2981 (nios2_delegitimize_address): New function.
2982 (TARGET_DELEGITIMIZE_ADDRESS): Define to nios2_delegitimize_address.
2983 * config/nios2/linux.h (GLIBC_DYNAMIC_LINKER): Define.
2984 (LINK_SPEC): Specify dynamic linker using GNU_USER_DYNAMIC_LINKER.
2985
8785c60a
MH
29862014-04-01 Martin Husemann <martin@duskware.de>
2987
2988 * config/mips/netbsd.h (TARGET_OS_CPP_BUILTINS): Define __mips_o32
2989 for -mabi=32.
2990
3343cf6a
RS
29912014-04-01 Richard Sandiford <rdsandiford@googlemail.com>
2992
2993 PR rtl-optimization/60604
2994 * recog.c (general_operand): Incorporate REG_CANNOT_CHANGE_MODE_P
2995 check from register_operand.
2996 (register_operand): Redefine in terms of general_operand.
2997 (nonmemory_operand): Use register_operand for the non-constant cases.
2998
1503cb6b
RB
29992014-04-01 Richard Biener <rguenther@suse.de>
3000
e0a81db1 3001 * gimple.h (struct gimple_statement_base): Align subcode to 16 bits.
1503cb6b 3002
cded08ea
SH
30032014-04-01 Sebastian Huber <sebastian.huber@embedded-brains.de>
3004
3005 * doc/invoke.texi (mapp-regs): Clarify.
3006
7d9088c2
UD
30072014-03-31 Ulrich Drepper <drepper@gmail.com>
3008
3009 * config/i386/avx512fintrin.h (__v32hi): Define type.
3010 (__v64qi): Likewise.
3011 (_mm512_set1_epi8): Define.
3012 (_mm512_set1_epi16): Define.
3013 (_mm512_set4_epi32): Define.
3014 (_mm512_set4_epi64): Define.
3015 (_mm512_set4_pd): Define.
3016 (_mm512_set4_ps): Define.
3017 (_mm512_setr4_epi64): Define.
3018 (_mm512_setr4_epi32): Define.
3019 (_mm512_setr4_pd): Define.
3020 (_mm512_setr4_ps): Define.
3021 (_mm512_setzero_epi32): Define.
3022
c18ff8a4
MJ
30232014-03-31 Martin Jambor <mjambor@suse.cz>
3024
3025 PR middle-end/60647
3026 * tree-sra.c (callsite_has_enough_arguments_p): Renamed to
3027 callsite_arguments_match_p. Updated all callers. Also check types of
3028 corresponding formal parameters and actual arguments.
3029 (not_all_callers_have_enough_arguments_p) Renamed to
3030 some_callers_have_mismatched_arguments_p.
3031
9a96d5b7
YR
30322014-03-31 Yuri Rumyantsev <ysrumyan@gmail.com>
3033
3034 * tree-inline.c (copy_loops): Add missed copy of 'safelen'.
3035
6c031d8d
KV
30362014-03-31 Kugan Vivekanandarajah <kuganv@linaro.org>
3037
3038 PR target/60034
3039 * aarch64/aarch64.c (aarch64_classify_address): Fix alignment for
3040 section anchor.
3041
429749e2
UB
30422014-03-30 Uros Bizjak <ubizjak@gmail.com>
3043
3044 * config/i386/sse.md (FMAMODE_NOVF512): New mode iterator.
3045 (<sd_mask_codefor>fma_fmadd_<mode><sd_maskz_name><round_name>):
3046 Split out
3047 <sd_mask_codefor>fma_fmadd_<VF_512:mode><sd_maskz_name><round_name>.
3048 Use FMAMODE_NOVF512 mode iterator.
3049 (<sd_mask_codefor>fma_fmsub_<mode><sd_maskz_name><round_name>): Ditto.
3050 (<sd_mask_codefor>fma_fnmadd_<mode><sd_maskz_name><round_name>): Ditto.
3051 (<sd_mask_codefor>fma_fnmsub_<mode><sd_maskz_name><round_name>): Ditto.
3052 (<sd_mask_codefor>fma_fmaddsub_<mode><sd_maskz_name><round_name>):
3053 Split out
3054 <sd_mask_codefor>fma_fmaddsub_<VF_512:mode><sd_maskz_name><round_name>.
3055 Use VF_128_256 mode iterator.
3056 (<sd_mask_codefor>fma_fmsubadd_<mode><sd_maskz_name><round_name>):
3057 Ditto.
3058
1f91035f
JH
30592014-03-28 Jan Hubicka <hubicka@ucw.cz>
3060
3061 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Clear
3062 static chain if needed.
3063
6e071b1e
VM
30642014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3065
3066 PR target/60697
3067 * lra-constraints.c (index_part_to_reg): New.
3068 (process_address): Use it.
3069
e7f31789 30702014-03-27 Jeff Law <law@redhat.com>
4bb66ef3 3071 Jakub Jelinek <jakub@redhat.com>
e7f31789
JL
3072
3073 PR target/60648
4bb66ef3
JJ
3074 * expr.c (do_tablejump): Use simplify_gen_binary rather than
3075 gen_rtx_{PLUS,MULT} to build up the address expression.
e7f31789 3076
4bb66ef3
JJ
3077 * i386/i386.c (ix86_legitimize_address): Use copy_addr_to_reg to avoid
3078 creating non-canonical RTL.
e7f31789 3079
4cd8957f
JH
30802014-03-28 Jan Hubicka <hubicka@ucw.cz>
3081
3082 PR ipa/60243
3083 * ipa-inline.c (want_inline_small_function_p): Short circuit large
3084 functions; reorganize to make cheap checks first.
3085 (inline_small_functions): Do not estimate growth when dumping;
3086 it is expensive.
3087 * ipa-inline.h (inline_summary): Add min_size.
3088 (growth_likely_positive): New function.
3089 * ipa-inline-analysis.c (dump_inline_summary): Add min_size.
3090 (set_cond_stmt_execution_predicate): Cleanup.
3091 (estimate_edge_size_and_time): Compute min_size.
3092 (estimate_calls_size_and_time): Likewise.
3093 (estimate_node_size_and_time): Likewise.
3094 (inline_update_overall_summary): Update min_size.
3095 (do_estimate_edge_time): Likewise.
3096 (do_estimate_edge_size): Update.
3097 (do_estimate_edge_hints): Update.
3098 (growth_likely_positive): New function.
3099
ff76f4e2
JJ
31002014-03-28 Jakub Jelinek <jakub@redhat.com>
3101
3102 PR target/60693
3103 * config/i386/i386.c (ix86_copy_addr_to_reg): Call copy_addr_to_reg
3104 also if addr has VOIDmode.
3105
ed6eb6dc
KT
31062014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3107
3108 * config/arm/aarch-common.c (aarch_crypto_can_dual_issue): New.
3109 * config/arm/aarch-common-protos.h (aarch_crypto_can_dual_issue):
3110 Declare extern.
3111 * config/arm/cortex-a53.md: Add reservations and bypass for crypto
3112 instructions as well as AdvancedSIMD loads.
3113
b10baa95
KT
31142014-03-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3115
3116 * config/aarch64/aarch64-simd.md (aarch64_crypto_aes<aes_op>v16qi):
3117 Use crypto_aese type.
3118 (aarch64_crypto_aes<aesmc_op>v16qi): Use crypto_aesmc type.
3119 * config/arm/arm.md (is_neon_type): Replace crypto_aes with
3120 crypto_aese, crypto_aesmc. Move to types.md.
3121 * config/arm/types.md (crypto_aes): Split into crypto_aese,
3122 crypto_aesmc.
3123 * config/arm/iterators.md (crypto_type): Likewise.
3124
23a04216
JH
31252014-03-28 Jan Hubicka <hubicka@ucw.cz>
3126
3127 * cgraph.c: Include expr.h and tree-dfa.h.
429749e2
UB
3128 (cgraph_redirect_edge_call_stmt_to_callee): If call in noreturn;
3129 remove LHS.
23a04216 3130
a4971e68
VM
31312014-03-28 Vladimir Makarov <vmakarov@redhat.com>
3132
3133 PR target/60675
3134 * lra-assigns.c (find_hard_regno_for): Remove unavailable hard
3135 regs from checking multi-reg pseudos.
3136
7ac84c02
RR
31372014-03-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3138
3139 * config/arm/t-aprofile (MULTILIB_MATCHES): Correct A12 rule.
3140
bd302a18
UW
31412014-03-28 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
3142
3143 * config/rs6000/rs6000.c (fusion_gpr_load_p): Refuse optimization
3144 if it would clobber the stack pointer, even temporarily.
3145
cee9defb
EB
31462014-03-28 Eric Botcazou <ebotcazou@adacore.com>
3147
3148 * mode-switching.c: Make small adjustments to the top comment.
3149
117f16fb
MM
31502014-03-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3151
3152 * config/rs6000/constraints.md (wD constraint): New constraint to
3153 match the constant integer to get the top DImode/DFmode out of a
3154 vector in a VSX register.
3155
3156 * config/rs6000/predicates.md (vsx_scalar_64bit): New predicate to
3157 match the constant integer to get the top DImode/DFmode out of a
3158 vector in a VSX register.
3159
3160 * config/rs6000/rs6000-builtins.def (VBPERMQ): Add vbpermq builtin
3161 for ISA 2.07.
3162
3163 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
3164 vbpermq builtins.
3165
3166 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
3167 -mdebug=reg, print value of VECTOR_ELEMENT_SCALAR_64BIT.
3168
3169 * config/rs6000/vsx.md (vsx_extract_<mode>, V2DI/V2DF modes):
3170 Optimize vec_extract of 64-bit values, where the value being
3171 extracted is in the top word, where we can use scalar
3172 instructions. Add direct move and store support. Combine the big
429749e2 3173 endian/little endian vector select load support into a single insn.
117f16fb
MM
3174 (vsx_extract_<mode>_internal1): Likewise.
3175 (vsx_extract_<mode>_internal2): Likewise.
3176 (vsx_extract_<mode>_load): Likewise.
3177 (vsx_extract_<mode>_store): Likewise.
3178 (vsx_extract_<mode>_zero): Delete, big and little endian insns are
3179 combined into vsx_extract_<mode>_load.
3180 (vsx_extract_<mode>_one_le): Likewise.
3181
3182 * config/rs6000/rs6000.h (VECTOR_ELEMENT_SCALAR_64BIT): Macro to
3183 define the top 64-bit vector element.
3184
3185 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wD
3186 constraint.
3187
6788164b
MM
3188 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3189 Document vec_vbpermq builtin.
3190
117f16fb
MM
3191 PR target/60672
3192 * config/rs6000/altivec.h (vec_xxsldwi): Add missing define to
3193 enable use of xxsldwi and xxpermdi builtin functions.
3194 (vec_xxpermdi): Likewise.
3195
3196 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
3197 Document use of vec_xxsldwi and vec_xxpermdi builtins.
3198
9e038952
VM
31992014-03-27 Vladimir Makarov <vmakarov@redhat.com>
3200
3201 PR rtl-optimization/60650
f4581282 3202 * lra-assign.c (find_hard_regno_for, spill_for): Add parameter
9e038952
VM
3203 first_p. Use it.
3204 (find_spills_for): New.
3205 (assign_by_spills): Pass the new parameter to find_hard_regno_for.
3206 Spill all pseudos on the second iteration.
3207
324dc401
MP
32082014-03-27 Marek Polacek <polacek@redhat.com>
3209
3210 PR c/50347
3211 * doc/extend.texi (ffs Builtins): Change unsigned types to signed
3212 types.
3213
6e77facf
AK
32142014-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3215
3216 * config/s390/s390.c (s390_can_use_return_insn): Check for
3217 call-saved FPRs on 31 bit.
3218
47519a14
JJ
32192014-03-27 Jakub Jelinek <jakub@redhat.com>
3220
3221 PR middle-end/60682
3222 * omp-low.c (lower_omp_1): For gimple_clobber_p stmts,
3223 if they need regimplification, just drop them instead of
3224 calling gimple_regimplify_operands on them.
3225
0b7f8166
MS
32262014-03-27 Marcus Shawcroft <marcus.shawcroft@arm.com>
3227
3228 PR target/60580
3229 * config/aarch64/aarch64.c (faked_omit_frame_pointer): Remove.
3230 (aarch64_frame_pointer_required): Adjust logic.
3231 (aarch64_can_eliminate): Adjust logic.
3232 (aarch64_override_options_after_change): Adjust logic.
3233
2f2a7d15
DC
32342014-03-27 Dehao Chen <dehao@google.com>
3235
3236 * ipa-inline.c (early_inliner): Update node's inline info.
3237
a0dbf285
DC
32382014-03-26 Dehao Chen <dehao@google.com>
3239
3240 * dojump.c (do_compare_rtx_and_jump): Sets correct probability for
3241 compiler inserted conditional jumps for NAN float check.
3242
ed0ca1e1
JJ
32432014-03-26 Jakub Jelinek <jakub@redhat.com>
3244
07d72e1d
JJ
3245 * ubsan.h (ubsan_create_data): Change second argument's type
3246 to const location_t *.
3247 * ubsan.c (ubsan_source_location): If xloc.file is NULL, set it to
3248 _("<unknown>").
3249 (ubsan_create_data): Change second argument to const location_t *PLOC.
3250 Create Loc field whenever PLOC is non-NULL.
3251 (ubsan_instrument_unreachable, ubsan_expand_null_ifn,
3252 ubsan_build_overflow_builtin, instrument_bool_enum_load): Adjust
3253 callers.
3254
ed0ca1e1
JJ
3255 PR other/59545
3256 * real.c (real_to_integer2): Change type of low to UHWI.
3257
b9487dd8
TB
32582014-03-26 Tobias Burnus <burnus@net-b.de>
3259
3260 * gcc.c (LINK_COMMAND_SPEC): Use libcilkrts.spec for -fcilkplus.
3261 (CILK_SELF_SPECS): New define.
3262 (driver_self_specs): Use it.
3263
3a2700f6
RB
32642014-03-26 Richard Biener <rguenther@suse.de>
3265
3266 * tree-pretty-print.c (percent_K_format): Implement special
3267 case for LTO and its stripped down BLOCK tree.
3268
368b454d
JJ
32692014-03-26 Jakub Jelinek <jakub@redhat.com>
3270
aa6b742f
JJ
3271 PR sanitizer/60636
3272 * ubsan.c (instrument_si_overflow): Instrument ABS_EXPR.
3273
368b454d
JJ
3274 * tree-vrp.c (simplify_internal_call_using_ranges): If only
3275 one range is range_int_cst_p, but not both, at least optimize
3276 addition/subtraction of 0 and multiplication by 0 or 1.
3277 * gimple-fold.c (gimple_fold_call): Fold
3278 IFN_UBSAN_CHECK_{ADD,SUB,MUL}.
3279 (gimple_fold_stmt_to_constant_1): If both op0 and op1 aren't
3280 INTEGER_CSTs, try to fold at least x * 0 and y - y.
3281
c7e30a96
EB
32822014-03-26 Eric Botcazou <ebotcazou@adacore.com>
3283
3284 PR rtl-optimization/60452
3285 * rtlanal.c (rtx_addr_can_trap_p_1): Fix head comment.
3286 <case REG>: Return 1 for invalid offsets from the frame pointer.
3287
d615d763
MP
32882014-03-26 Marek Polacek <polacek@redhat.com>
3289
3290 PR c/37428
3291 * doc/extend.texi (C Extensions): Mention variable-length arrays in
3292 a structure/union.
3293
1b226529
MP
32942014-03-26 Marek Polacek <polacek@redhat.com>
3295
3296 PR c/39525
3297 * doc/extend.texi (Designated Inits): Describe what happens to omitted
3298 field members.
3299
6686e0bc
MP
33002014-03-26 Marek Polacek <polacek@redhat.com>
3301
3302 PR other/59545
3303 * ira-color.c (update_conflict_hard_regno_costs): Perform the
3304 multiplication in unsigned type.
3305
d4fbc3ae
CJW
33062014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3307
3308 * doc/install.texi: Document nds32le-*-elf and nds32be-*-elf.
3309
dfa6c87b
CJW
33102014-03-26 Chung-Ju Wu <jasonwucj@gmail.com>
3311
3312 * doc/contrib.texi: Add myself as Andes nds32 port contributor.
3313
9de6f6c3
JH
33142014-03-25 Jan Hubicka <hubicka@ucw.cz>
3315
3316 PR ipa/60315
3317 * cif-code.def (UNREACHABLE) New code.
429749e2
UB
3318 * ipa-inline.c (inline_small_functions): Skip edges to
3319 __builtlin_unreachable.
9de6f6c3
JH
3320 (estimate_edge_growth): Allow edges to __builtlin_unreachable.
3321 * ipa-inline-analysis.c (edge_set_predicate): Redirect edges with false
3322 predicate to __bulitin_unreachable.
429749e2
UB
3323 (set_cond_stmt_execution_predicate): Fix issue when
3324 invert_tree_comparison returns ERROR_MARK.
9de6f6c3
JH
3325 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow): Do not
3326 propagate to inline clones.
3327 * cgraph.c (verify_edge_corresponds_to_fndecl): Allow redirection
3328 to unreachable.
3329 * ipa-cp.c (create_specialized_node): Be ready for new node to appear.
3330 * cgraphclones.c (cgraph_clone_node): If call destination is already
3331 ureachable, do not redirect it back.
3332 * tree-inline.c (fold_marked_statements): Hanlde calls becoming
3333 unreachable.
3334
d7636f56
JH
33352014-03-25 Jan Hubicka <hubicka@ucw.cz>
3336
3337 * ipa-pure-const.c (propagate_pure_const, propagate_nothrow):
3338 Do not modify inline clones.
3339
d1873c57
JJ
33402014-03-25 Jakub Jelinek <jakub@redhat.com>
3341
3342 * config/i386/i386.md (general_sext_operand): New mode attr.
3343 (addv<mode>4, subv<mode>4, mulv<mode>4): If operands[2] is CONST_INT,
3344 don't generate (sign_extend (const_int)).
3345 (*addv<mode>4, *subv<mode>4, *mulv<mode>4): Disallow CONST_INT_P
429749e2
UB
3346 operands[2]. Use We constraint instead of <i> and
3347 <general_sext_operand> predicate instead of <general_operand>.
d1873c57
JJ
3348 (*addv<mode>4_1, *subv<mode>4_1, *mulv<mode>4_1): New insns.
3349 * config/i386/constraints.md (We): New constraint.
3350 * config/i386/predicates.md (x86_64_sext_operand,
3351 sext_operand): New predicates.
3352
b5165eb0
MJ
33532014-03-25 Martin Jambor <mjambor@suse.cz>
3354
3355 PR ipa/60600
3356 * ipa-cp.c (ipa_get_indirect_edge_target_1): Redirect type
3357 inconsistent devirtualizations to __builtin_unreachable.
3358
3102858d
MP
33592014-03-25 Marek Polacek <polacek@redhat.com>
3360
3361 PR c/35449
3362 * doc/extend.texi (Example of asm with clobbered asm reg): Fix typo.
3363
4b1e108c
AL
33642014-03-25 Alan Lawrence <alan.lawrence@arm.com>
3365
3366 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Reverse
3367 order of elements for big-endian.
3368
40b0722f
RB
33692014-03-25 Richard Biener <rguenther@suse.de>
3370
3371 PR middle-end/60635
3372 * gimplify-me.c (gimple_regimplify_operands): Update the
3373 re-gimplifed stmt.
3374
4a5798de
MJ
33752014-03-25 Martin Jambor <mjambor@suse.cz>
3376
3377 PR ipa/59176
3378 * lto-cgraph.c (lto_output_node): Stream body_removed flag.
3379 (lto_output_varpool_node): Likewise.
3380 (input_overwrite_node): Likewise.
3381 (input_varpool_node): Likewise.
3382
585d0088
RB
33832014-03-25 Richard Biener <rguenther@suse.de>
3384
429749e2 3385 * lto-wrapper.c (merge_and_complain): Handle OPT_fPIE like OPT_fpie.
585d0088
RB
3386 (run_gcc): Likewise.
3387
a8acccdd
JJ
33882014-03-25 Jakub Jelinek <jakub@redhat.com>
3389
51ee05d0
JJ
3390 * combine.c (simplify_compare_const): Add MODE argument.
3391 Handle mode_width 0 as very large mode_width.
3392 (try_combine, simplify_comparison): Adjust callers.
3393
a8acccdd
JJ
3394 * cselib.c (cselib_hash_rtx): Perform addition in unsigned
3395 type to avoid signed integer overflow.
3396 * explow.c (plus_constant): Likewise.
3397
57fc74a4
DV
33982014-03-25 Dominik Vogt <vogt@linux.vnet.ibm.com>
3399
3400 * doc/generic.texi: Correct typos.
3401
09f15d1b
TB
34022014-03-24 Tobias Burnus <burnus@net-b.de>
3403
3404 * doc/invoke.texi (-flto): Expand section about
3405 using static libraries with LTO.
3406
72a4ddf2
AK
34072014-03-24 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3408
3409 PR rtl-optimization/60501
3410 * optabs.def (addptr3_optab): New optab.
3411 * optabs.c (gen_addptr3_insn, have_addptr3_insn): New function.
3412 * doc/md.texi ("addptrm3"): Document new RTL standard expander.
3413 * expr.h (gen_addptr3_insn, have_addptr3_insn): Add prototypes.
3414
3415 * lra.c (emit_add3_insn): Use the addptr pattern if available.
3416
3417 * config/s390/s390.md ("addptrdi3", "addptrsi3"): New expanders.
3418
0b192937
UD
34192014-03-24 Ulrich Drepper <drepper@gmail.com>
3420
2b2384e8
UD
3421 * config/i386/avx512fintrin.h: Define _mm512_set1_ps and
3422 _mm512_set1_pd.
3423
0b192937
UD
3424 * config/i386/avxintrin.h (_mm256_undefined_si256): Define.
3425 (_mm256_undefined_ps): Define.
3426 (_mm256_undefined_pd): Define.
3427 * config/i386/emmintrin.h (_mm_undefined_si128): Define.
3428 (_mm_undefined_pd): Define.
3429 * config/i386/xmmintrin.h (_mm_undefined_ps): Define.
3430 * config/i386/avx512fintrin.h (_mm512_undefined_si512): Define.
3431 (_mm512_undefined_ps): Define.
3432 (_mm512_undefined_pd): Define.
3433 Use _mm*_undefined_*.
3434 * config/i386/avx2intrin.h: Use _mm*_undefined_*.
3435
252c7556
AV
34362014-03-24 Alex Velenko <Alex.Velenko@arm.com>
3437
3438 * config/aarch64/aarch64-simd-builtins.def (lshr): DI mode excluded.
3439 (lshr_simd): DI mode added.
3440 * config/aarch64/aarch64-simd.md (aarch64_lshr_simddi): New pattern.
3441 (aarch64_ushr_simddi): Likewise.
3442 * config/aarch64/aarch64.md (UNSPEC_USHR64): New unspec.
3443 * config/aarch64/arm_neon.h (vshr_n_u64): Intrinsic fixed.
3444 (vshrd_n_u64): Likewise.
3445
15e3d4f1
RO
34462014-03-24 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3447
3448 * Makefile.in (s-macro_list): Depend on cc1.
3449
1858c2d3
TJ
34502014-03-23 Teresa Johnson <tejohnson@google.com>
3451
3452 * ipa-utils.c (ipa_print_order): Use specified dump file.
3453
f87604f8
EB
34542014-03-23 Eric Botcazou <ebotcazou@adacore.com>
3455
3456 PR rtl-optimization/60601
3457 * bb-reorder.c (fix_up_fall_thru_edges): Test EDGE_FALLTHRU everywhere.
3458
3459 * gcc.c (eval_spec_function): Initialize save_growing_value.
3460
3db31fd1
JJ
34612014-03-22 Jakub Jelinek <jakub@redhat.com>
3462
f451d3a8
JJ
3463 PR sanitizer/60613
3464 * internal-fn.c (ubsan_expand_si_overflow_addsub_check): For
3465 code == MINUS_EXPR, never swap op0 with op1.
3466
3db31fd1
JJ
3467 * toplev.c (init_local_tick): Avoid signed integer multiplication
3468 overflow.
3469 * genautomata.c (reserv_sets_hash_value): Fix rotate idiom, avoid
3470 shift by first operand's bitsize.
3471
e49080ec
JJ
34722014-03-21 Jakub Jelinek <jakub@redhat.com>
3473
3474 PR target/60610
3475 * config/i386/i386.h (TARGET_64BIT_P): If not TARGET_BI_ARCH,
3476 redefine to 1 or 0.
3477 * config/i386/darwin.h (TARGET_64BIT_P): Redefine to
3478 TARGET_ISA_64BIT_P(x).
3479
44eb6c40
BS
34802014-03-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3481
3482 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Generate a
3483 pattern for vector nor instead of subtract from splat(-1).
3484 (altivec_expand_vec_perm_const_le): Likewise.
3485
ec6f831a
RH
34862014-03-21 Richard Henderson <rth@twiddle.net>
3487
3488 PR target/60598
3489 * ifcvt.c (dead_or_predicable): Return FALSE if there are any frame
3490 related insns after epilogue_completed.
3491
3d8d0043
MJ
34922014-03-21 Martin Jambor <mjambor@suse.cz>
3493
3494 PR ipa/59176
3495 * cgraph.h (symtab_node): New flag body_removed.
3496 * ipa.c (symtab_remove_unreachable_nodes): Set body_removed flag
3497 when removing bodies.
3498 * symtab.c (dump_symtab_base): Dump body_removed flag.
3499 * cgraph.c (verify_edge_corresponds_to_fndecl): Skip nodes which
3500 had their bodies removed.
3501
d833415c
MJ
35022014-03-21 Martin Jambor <mjambor@suse.cz>
3503
3504 PR ipa/60419
3505 * ipa.c (symtab_remove_unreachable_nodes): Clear thunk flag of nodes
3506 in the border.
3507
400a4f6c
RB
35082014-03-21 Richard Biener <rguenther@suse.de>
3509
3510 PR tree-optimization/60577
3511 * tree-core.h (struct tree_base): Document nothrow_flag use
3512 in VAR_DECL_NONALIASED.
3513 * tree.h (VAR_DECL_NONALIASED): New.
3514 (may_be_aliased): Adjust.
3515 * coverage.c (build_var): Set VAR_DECL_NONALIASED.
3516
c6c2314f
EB
35172014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3518
3519 * expr.c (expand_expr_real_1): Remove outdated comment.
3520
d02b6ac7
JJ
35212014-03-20 Jakub Jelinek <jakub@redhat.com>
3522
b8f045e2
JJ
3523 PR middle-end/60597
3524 * ira.c (adjust_cleared_regs): Call copy_rtx on
3525 *reg_equiv[REGNO (loc)].src_p before passing it to
3526 simplify_replace_fn_rtx.
3527
d02b6ac7
JJ
3528 PR target/60568
3529 * config/i386/i386.c (x86_output_mi_thunk): Surround UNSPEC_GOT
3530 into CONST, put pic register as first operand of PLUS. Use
3531 gen_const_mem for both 32-bit and 64-bit PIC got loads.
3532
e2f98c01
RR
35332014-03-20 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
3534
3535 * config/aarch64/aarch64.c (MEMORY_MOVE_COST): Delete.
3536
7075c792
EB
35372014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3538
3539 * config/sparc/sparc.c (sparc_do_work_around_errata): Implement work
3540 around for store forwarding issue in the FPU on the UT699.
3541 * config/sparc/sparc.md (in_branch_delay): Return false for single FP
3542 loads and operations if -mfix-ut699 is specified.
3543 (divtf3_hq): Tweak attribute.
3544 (sqrttf2_hq): Likewise.
3545
9ee5337d
EB
35462014-03-20 Eric Botcazou <ebotcazou@adacore.com>
3547
3548 * calls.c (store_one_arg): Remove incorrect const qualification on the
3549 type of the temporary.
3550 * cfgexpand.c (expand_return): Likewise.
3551 * expr.c (expand_constructor): Likewise.
3552 (expand_expr_real_1): Likewise.
3553
a69e09c8
ZC
35542014-03-20 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3555
3556 * config/arm/arm.c (arm_dwarf_register_span): Update the element number
3557 of parts.
3558
8659d57b
KK
35592014-03-19 Kaz Kojima <kkojima@gcc.gnu.org>
3560
3561 PR target/60039
3562 * config/sh/sh.md (udivsi3_i1): Clobber R1 register.
3563
985e0d33
JG
35642014-03-19 James Greenhalgh <james.greenhalgh@arm.com>
3565
3566 * config/arm/aarch-common-protos.h
3567 (alu_cost_table): Fix spelling of "extend".
3568 * config/arm/arm.c (arm_new_rtx_costs): Fix spelling of "extend".
3569
ce2f49b3
RB
35702014-03-19 Richard Biener <rguenther@suse.de>
3571
3572 PR middle-end/60553
3573 * tree-core.h (tree_type_common): Re-order pointer members
3574 to reduce recursion depth during GC walks.
3575
6daa2d91
MP
35762014-03-19 Marek Polacek <polacek@redhat.com>
3577
3578 PR sanitizer/60569
3579 * ubsan.c (ubsan_type_descriptor): Check that DECL_NAME is nonnull
3580 before accessing it.
3581
8c9cfbe6
RB
35822014-03-19 Richard Biener <rguenther@suse.de>
3583
3584 PR lto/59543
3585 * lto-streamer-in.c (input_function): In WPA stage do not drop
3586 debug stmts.
3587
3efe2e2c
JJ
35882014-03-19 Jakub Jelinek <jakub@redhat.com>
3589
3590 PR tree-optimization/60559
3591 * vectorizable_mask_load_store): Replace scalar MASK_LOAD
3592 with build_zero_cst assignment.
3593
234f76cb
KT
35942014-03-18 Kai Tietz <ktietz@redhat.com>
3595
3596 PR rtl-optimization/56356
3597 * sdbout.c (sdbout_parms): Verify that parms'
3598 incoming argument is valid.
3599 (sdbout_reg_parms): Likewise.
3600
58ca6edf
RH
36012014-03-18 Richard Henderson <rth@redhat.com>
3602
3603 PR target/60562
3604 * config/i386/i386.md (*float<SWI48x><MODEF>2_i387): Move down to
3605 be shadowed by *float<SWI48><MODEF>2_sse. Test X87_ENABLE_FLOAT.
ead8233b 3606 (*float<SWI48><MODEF>2_sse): Check X87_ENABLE_FLOAT for alternative 0.
b318e404
BS
3607
36082014-03-18 Basile Starynkevitch <basile@starynkevitch.net>
3609
4bb66ef3 3610 * plugin.def: Improve comment for PLUGIN_INCLUDE_FILE.
429749e2 3611 * doc/plugins.texi (Plugin callbacks): Mention PLUGIN_INCLUDE_FILE.
4bb66ef3 3612 Italicize plugin event names in description. Explain that
429749e2
UB
3613 PLUGIN_PRAGMAS has no sense for lto1. Explain PLUGIN_INCLUDE_FILE.
3614 Remind that no GCC functions should be called after PLUGIN_FINISH.
4bb66ef3 3615 Explain what pragmas with expansion are.
b318e404 3616
eb14a79f
ML
36172014-03-18 Martin Liska <mliska@suse.cz>
3618
3619 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Added case when
3620 gimple call statement is update.
3621 * gimple-fold.c (gimple_fold_call): Changed order for GIMPLE_ASSIGN and
3622 GIMPLE_CALL, where gsi iterator still points to GIMPLE CALL.
3623
415e2e6b
JJ
36242014-03-18 Jakub Jelinek <jakub@redhat.com>
3625
ead96f13
JJ
3626 PR sanitizer/60557
3627 * ubsan.c (ubsan_instrument_unreachable): Call
3628 initialize_sanitizer_builtins.
3629 (ubsan_pass): Likewise.
3630
415e2e6b
JJ
3631 PR sanitizer/60535
3632 * ubsan.c (ubsan_type_descriptor, ubsan_create_data): Call
3633 varpool_finalize_decl instead of rest_of_decl_compilation.
3634
4ca40f52
RB
36352014-03-18 Richard Biener <rguenther@suse.de>
3636
3637 * df-problems.c (df_rd_confluence_n): Avoid bitmap_copy
3638 by using bitmap_and_compl instead of bitmap_and_compl_into.
3639 (df_rd_transfer_function): Likewise.
3640
6404e190
RB
36412014-03-18 Richard Biener <rguenther@suse.de>
3642
3643 * doc/lto.texi (fresolution): Fix typo.
3644
b9abf793
RB
36452014-03-18 Richard Biener <rguenther@suse.de>
3646
3647 * doc/invoke.texi (flto): Update for changes in 4.9.
3648
3f00cf18
RB
36492014-03-18 Richard Biener <rguenther@suse.de>
3650
3651 * doc/loop.texi: Remove section on the removed lambda framework.
3652 Update loop docs with recent changes in preserving loop structure.
3653
c2679d84
RB
36542014-03-18 Richard Biener <rguenther@suse.de>
3655
3656 * doc/lto.texi (-fresolution): Document.
3657
c07990cc
RB
36582014-03-18 Richard Biener <rguenther@suse.de>
3659
3660 * doc/contrib.texi: Adjust my name.
3661
ed9c79e1
JJ
36622014-03-18 Jakub Jelinek <jakub@redhat.com>
3663
3664 PR ipa/58721
3665 * internal-fn.c: Include diagnostic-core.h.
3666 (expand_BUILTIN_EXPECT): New function.
3667 * gimplify.c (gimplify_call_expr): Use false instead of FALSE.
3668 (gimplify_modify_expr): Gimplify 3 argument __builtin_expect into
3669 IFN_BUILTIN_EXPECT call instead of __builtin_expect builtin call.
3670 * ipa-inline-analysis.c (find_foldable_builtin_expect): Handle
3671 IFN_BUILTIN_EXPECT.
3672 * predict.c (expr_expected_value_1): Handle IFN_BUILTIN_EXPECT.
3673 Revert 3 argument __builtin_expect code.
3674 (strip_predict_hints): Handle IFN_BUILTIN_EXPECT.
3675 * gimple-fold.c (gimple_fold_call): Likewise.
3676 * tree.h (fold_builtin_expect): New prototype.
3677 * builtins.c (build_builtin_expect_predicate): Add predictor
3678 argument, if non-NULL, create 3 argument __builtin_expect.
3679 (fold_builtin_expect): No longer static. Add ARG2 argument,
3680 pass it through to build_builtin_expect_predicate.
3681 (fold_builtin_2): Adjust caller.
3682 (fold_builtin_3): Handle BUILT_IN_EXPECT.
3683 * internal-fn.def (BUILTIN_EXPECT): New.
3684
36852014-03-18 Tobias Burnus <burnus@net-b.de>
3686
3687 PR ipa/58721
3688 * predict.def (PRED_FORTRAN_OVERFLOW, PRED_FORTRAN_FAIL_ALLOC,
3689 PRED_FORTRAN_FAIL_IO, PRED_FORTRAN_WARN_ONCE, PRED_FORTRAN_SIZE_ZERO,
3690 PRED_FORTRAN_INVALID_BOUND, PRED_FORTRAN_ABSENT_DUMMY): Add.
3691
36922014-03-18 Jan Hubicka <hubicka@ucw.cz>
3693
3694 PR ipa/58721
3695 * predict.c (combine_predictions_for_bb): Fix up formatting.
3696 (expr_expected_value_1, expr_expected_value): Add predictor argument,
3697 fill what it points to if non-NULL.
3698 (tree_predict_by_opcode): Adjust caller, use the predictor.
3699 * predict.def (PRED_COMPARE_AND_SWAP): Add.
3700
e33da4a1
EB
37012014-03-18 Eric Botcazou <ebotcazou@adacore.com>
3702
3703 * config/sparc/sparc.c (sparc_do_work_around_errata): Speed up and use
3704 proper constant for the store mode.
3705
1fed15fc
IE
37062014-03-18 Ilya Enkovich <ilya.enkovich@intel.com>
3707
3708 * symtab.c (change_decl_assembler_name): Fix transparent alias
3709 chain construction.
3710
454fdba9
RL
37112014-03-16 Renlin Li <Renlin.Li@arm.com>
3712
3713 * config/aarch64/aarch64.c: Correct the comments about the
3714 aarch64 stack layout.
3715
e2110f8f
TS
37162014-03-18 Thomas Schwinge <thomas@codesourcery.com>
3717
3718 * omp-low.c (lower_rec_input_clauses) <build_omp_barrier>: Restore
3719 check for GF_OMP_FOR_KIND_FOR.
3720
7c831c4d
KY
37212013-03-18 Kirill Yukhin <kirill.yukhin@intel.com>
3722
3723 * config/i386/i386.h (ADDITIONAL_REGISTER_NAMES): Add
3724 ymm and zmm register names.
3725
e318c1e1
JJ
37262014-03-17 Jakub Jelinek <jakub@redhat.com>
3727
3728 PR target/60516
3729 * config/i386/i386.c (ix86_expand_epilogue): Adjust REG_CFA_ADJUST_CFA
3730 note creation for the 2010-08-31 changes.
3731
ab2ffbeb
MP
37322014-03-17 Marek Polacek <polacek@redhat.com>
3733
3734 PR middle-end/60534
3735 * omp-low.c (omp_max_vf): Treat -fno-tree-loop-optimize the same
3736 as -fno-tree-loop-vectorize.
3737 (expand_omp_simd): Likewise.
3738
890fa568
EB
37392014-03-15 Eric Botcazou <ebotcazou@adacore.com>
3740
3741 * config/sparc/sparc-protos.h (tls_call_delay): Delete.
3742 (eligible_for_call_delay): New prototype.
3743 * config/sparc/sparc.c (tls_call_delay): Rename into...
3744 (eligible_for_call_delay): ...this. Return false if the instruction
3745 cannot be put in the delay slot of a branch.
3746 (eligible_for_restore_insn): Simplify.
3747 (eligible_for_return_delay): Return false if the instruction cannot be
3748 put in the delay slot of a branch and simplify.
3749 (eligible_for_sibcall_delay): Return false if the instruction cannot be
3750 put in the delay slot of a branch.
3751 * config/sparc/sparc.md (fix_ut699): New attribute.
3752 (tls_call_delay): Delete.
3753 (in_call_delay): Reimplement.
3754 (eligible_for_sibcall_delay): Rename into...
3755 (in_sibcall_delay): ...this.
3756 (eligible_for_return_delay): Rename into...
3757 (in_return_delay): ...this.
3758 (in_branch_delay): Reimplement.
3759 (in_uncond_branch_delay): Delete.
3760 (in_annul_branch_delay): Delete.
3761
cfce90ac
RH
37622014-03-14 Richard Henderson <rth@redhat.com>
3763
3764 PR target/60525
3765 * config/i386/i386.md (floathi<X87MODEF>2): Delete expander; rename
3766 define_insn from *floathi<X87MODEF>2_i387; allow nonimmediate_operand.
3767 (*floathi<X87MODEF>2_i387_with_temp): Remove.
3768 (floathi splitters): Remove.
3769 (float<SWI48x>xf2): New pattern.
3770 (float<SWI48><MODEF>2): Rename from float<SWI48x><X87MODEF>2. Drop
3771 code that tried to handle DImode for 32-bit, but which was excluded
3772 by the pattern's condition. Drop allocation of stack temporary.
3773 (*floatsi<MODEF>2_vector_mixed_with_temp): Remove.
3774 (*float<SWI48><MODEF>2_mixed_with_temp): Remove.
3775 (*float<SWI48><MODEF>2_mixed_interunit): Remove.
3776 (*float<SWI48><MODEF>2_mixed_nointerunit): Remove.
3777 (*floatsi<MODEF>2_vector_sse_with_temp): Remove.
3778 (*float<SWI48><MODEF>2_sse_with_temp): Remove.
3779 (*float<SWI48><MODEF>2_sse_interunit): Remove.
3780 (*float<SWI48><MODEF>2_sse_nointerunit): Remove.
3781 (*float<SWI48x><X87MODEF>2_i387_with_temp): Remove.
3782 (*float<SWI48x><X87MODEF>2_i387): Remove.
3783 (all float _with_temp splitters): Remove.
3784 (*float<SWI48x><MODEF>2_i387): New pattern.
3785 (*float<SWI48><MODEF>2_sse): New pattern.
3786 (float TARGET_USE_VECTOR_CONVERTS splitters): Merge them.
3787 (float TARGET_SSE_PARTIAL_REG_DEPENDENCY splitters): Merge them.
3788
91035653
JJ
37892014-03-14 Jakub Jelinek <jakub@redhat.com>
3790 Marek Polacek <polacek@redhat.com>
3791
3792 PR middle-end/60484
3793 * common.opt (dump_base_name_prefixed): New Variable.
3794 * opts.c (finish_options): Don't prepend directory to x_dump_base_name
3795 if x_dump_base_name_prefixed is already set, set it at the end.
3796
95921002
VM
37972014-03-14 Vladimir Makarov <vmakarov@redhat.com>
3798
3799 PR rtl-optimization/60508
3800 * lra-constraints.c (get_reload_reg): Add new parameter
3801 in_subreg_p.
3802 (process_addr_reg, simplify_operand_subreg, curr_insn_transform):
3803 Pass the new parameter values.
3804
25e57622
RB
38052014-03-14 Richard Biener <rguenther@suse.de>
3806
3807 * common.opt: Revert unintented changes from r205065.
3808 * opts.c: Likewise.
3809
1f3388fe
RB
38102014-03-14 Richard Biener <rguenther@suse.de>
3811
3812 PR middle-end/60518
3813 * cfghooks.c (split_block): Properly adjust all loops the
3814 block was a latch of.
3815
38162014-03-14 Martin Jambor <mjambor@suse.cz>
3d354792
MJ
3817
3818 PR lto/60461
3819 * ipa-prop.c (ipa_modify_call_arguments): Fix iteration condition
3820 and simplify it.
3821
2832dc22
GJL
38222014-03-14 Georg-Johann Lay <avr@gjlay.de>
3823
3824 PR target/59396
3825 * config/avr/avr.c (avr_set_current_function): Pass function name
3826 through default_strip_name_encoding before sanity checking instead
3827 of skipping the first char of the assembler name.
3828
3386d77e
RH
38292014-03-13 Richard Henderson <rth@redhat.com>
3830
3831 PR debug/60438
3832 * config/i386/i386.c (ix86_split_fp_branch): Remove pushed argument.
3833 (ix86_force_to_memory, ix86_free_from_memory): Remove.
3834 * config/i386/i386-protos.h: Likewise.
3835 * config/i386/i386.md (floathi<X87MODEF>2): Use assign_386_stack_local
3836 in the expander instead of a splitter.
3837 (float<SWI48x><X87MODEF>2): Use assign_386_stack_local if there is
3838 any possibility of requiring a memory.
3839 (*floatsi<MODEF>2_vector_mixed): Remove, and the splitters.
3840 (*floatsi<MODEF>2_vector_sse): Remove, and the splitters.
3841 (fp branch splitters): Update for ix86_split_fp_branch.
3842 (*jcc<X87MODEF>_<SWI24>_i387): Remove r/f alternative.
3843 (*jcc<X87MODEF>_<SWI24>_r_i387): Likewise.
3844 (splitter for jcc<X87MODEF>_<SWI24>_i387 r/f): Remove.
3845 (*fop_<MODEF>_2_i387): Remove f/r alternative.
3846 (*fop_<MODEF>_3_i387): Likewise.
3847 (*fop_xf_2_i387, *fop_xf_3_i387): Likewise.
3848 (splitters for the fop_* register patterns): Remove.
3849 (fscalexf4_i387): Rename from *fscalexf4_i387.
3850 (ldexpxf3): Use gen_floatsixf2 and gen_fscalexf4_i387.
4bb66ef3 3851
5b5d7f31
JJ
38522014-03-13 Jakub Jelinek <jakub@redhat.com>
3853
3854 PR tree-optimization/59779
3855 * tree-dfa.c (get_ref_base_and_extent): Use double_int
3856 type for bitsize and maxsize instead of HOST_WIDE_INT.
3857
bdc6e1ae
SB
38582014-03-13 Steven Bosscher <steven@gcc.gnu.org>
3859
3860 PR rtl-optimization/57320
3861 * function.c (rest_of_handle_thread_prologue_and_epilogue): Cleanup
3862 the CFG after thread_prologue_and_epilogue_insns.
3863
7891065a
VM
38642014-03-13 Vladimir Makarov <vmakarov@redhat.com>
3865
3866 PR rtl-optimization/57189
3867 * lra-constraints.c (process_alt_operands): Disfavor spilling
3868 vector pseudos.
3869
eba14fca
CP
38702014-03-13 Cesar Philippidis <cesar@codesourcery.com>
3871
429749e2 3872 * lto-wrapper.c (maybe_unlink_file): Suppress diagnostic messages.
eba14fca 3873
f661b085
JJ
38742014-03-13 Jakub Jelinek <jakub@redhat.com>
3875
3876 PR tree-optimization/59025
3877 PR middle-end/60418
3878 * tree-ssa-reassoc.c (sort_by_operand_rank): For SSA_NAMEs with the
3879 same rank, sort by bb_rank and gimple_uid of SSA_NAME_DEF_STMT first.
3880
8f3a3138
GJL
38812014-03-13 Georg-Johann Lay <avr@gjlay.de>
3882
3883 PR target/60486
3884 * config/avr/avr.c (avr_out_plus): Swap cc_plus and cc_minus in
3885 calls of avr_out_plus_1.
3886
d731ee04
BC
38872014-03-13 Bin Cheng <bin.cheng@arm.com>
3888
3889 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Record
3890 BB's single pred and update the father loop's latch info later.
3891
a16a872d
MM
38922014-03-12 Michael Meissner <meissner@linux.vnet.ibm.com>
3893
3894 * config/rs6000/vector.md (VEC_L): Add V1TI mode to vector types.
3895 (VEC_M): Likewise.
3896 (VEC_N): Likewise.
3897 (VEC_R): Likewise.
3898 (VEC_base): Likewise.
3899 (mov<MODE>, VEC_M modes): If we are loading TImode into VSX
3900 registers, we need to swap double words in little endian mode.
3901
3902 * config/rs6000/rs6000-modes.def (V1TImode): Add new vector mode
3903 to be a container mode for 128-bit integer operations added in ISA
3904 2.07. Unlike TImode and PTImode, the preferred register set is
3905 the Altivec/VMX registers for the 128-bit operations.
3906
3907 * config/rs6000/rs6000-protos.h (rs6000_move_128bit_ok_p): Add
3908 declarations.
3909 (rs6000_split_128bit_ok_p): Likewise.
3910
3911 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_3): Add new support
3912 macros for creating ISA 2.07 normal and overloaded builtin
3913 functions with 3 arguments.
3914 (BU_P8V_OVERLOAD_3): Likewise.
3915 (VPERM_1T): Add support for V1TImode in 128-bit vector operations
3916 for use as overloaded functions.
3917 (VPERM_1TI_UNS): Likewise.
3918 (VSEL_1TI): Likewise.
3919 (VSEL_1TI_UNS): Likewise.
3920 (ST_INTERNAL_1ti): Likewise.
3921 (LD_INTERNAL_1ti): Likewise.
3922 (XXSEL_1TI): Likewise.
3923 (XXSEL_1TI_UNS): Likewise.
3924 (VPERM_1TI): Likewise.
3925 (VPERM_1TI_UNS): Likewise.
3926 (XXPERMDI_1TI): Likewise.
3927 (SET_1TI): Likewise.
3928 (LXVD2X_V1TI): Likewise.
3929 (STXVD2X_V1TI): Likewise.
3930 (VEC_INIT_V1TI): Likewise.
3931 (VEC_SET_V1TI): Likewise.
3932 (VEC_EXT_V1TI): Likewise.
3933 (EQV_V1TI): Likewise.
3934 (NAND_V1TI): Likewise.
3935 (ORC_V1TI): Likewise.
3936 (VADDCUQ): Add support for 128-bit integer arithmetic instructions
3937 added in ISA 2.07. Add both normal 'altivec' builtins, and the
3938 overloaded builtin.
3939 (VADDUQM): Likewise.
3940 (VSUBCUQ): Likewise.
3941 (VADDEUQM): Likewise.
3942 (VADDECUQ): Likewise.
3943 (VSUBEUQM): Likewise.
3944 (VSUBECUQ): Likewise.
3945
3946 * config/rs6000/rs6000-c.c (__int128_type): New static to hold
3947 __int128_t and __uint128_t types.
3948 (__uint128_type): Likewise.
3949 (altivec_categorize_keyword): Add support for vector __int128_t,
3950 vector __uint128_t, vector __int128, and vector unsigned __int128
3951 as a container type for TImode operations that need to be done in
3952 VSX/Altivec registers.
3953 (rs6000_macro_to_expand): Likewise.
3954 (altivec_overloaded_builtins): Add ISA 2.07 overloaded functions
3955 to support 128-bit integer instructions vaddcuq, vadduqm,
3956 vaddecuq, vaddeuqm, vsubcuq, vsubuqm, vsubecuq, vsubeuqm.
3957 (altivec_resolve_overloaded_builtin): Add support for V1TImode.
3958
3959 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Add support
0bb29a05
UB
3960 for V1TImode, and set up preferences to use VSX/Altivec registers.
3961 Setup VSX reload handlers.
a16a872d
MM
3962 (rs6000_debug_reg_global): Likewise.
3963 (rs6000_init_hard_regno_mode_ok): Likewise.
3964 (rs6000_preferred_simd_mode): Likewise.
0bb29a05 3965 (vspltis_constant): Do not allow V1TImode as easy altivec constants.
a16a872d
MM
3966 (easy_altivec_constant): Likewise.
3967 (output_vec_const_move): Likewise.
3968 (rs6000_expand_vector_set): Convert V1TImode set and extract to
3969 simple move.
3970 (rs6000_expand_vector_extract): Likewise.
3971 (reg_offset_addressing_ok_p): Setup V1TImode to use VSX reg+reg
3972 addressing.
3973 (rs6000_const_vec): Add support for V1TImode.
3974 (rs6000_emit_le_vsx_load): Swap double words when loading or
3975 storing TImode/V1TImode.
3976 (rs6000_emit_le_vsx_store): Likewise.
3977 (rs6000_emit_le_vsx_move): Likewise.
3978 (rs6000_emit_move): Add support for V1TImode.
3979 (altivec_expand_ld_builtin): Likewise.
3980 (altivec_expand_st_builtin): Likewise.
3981 (altivec_expand_vec_init_builtin): Likewise.
3982 (altivec_expand_builtin): Likewise.
3983 (rs6000_init_builtins): Add support for V1TImode type. Add
3984 support for ISA 2.07 128-bit integer builtins. Define type names
3985 for the VSX/Altivec vector types.
3986 (altivec_init_builtins): Add support for overloaded vector
3987 functions with V1TImode type.
0bb29a05 3988 (rs6000_preferred_reload_class): Prefer Altivec registers for V1TImode.
a16a872d
MM
3989 (rs6000_move_128bit_ok_p): Move 128-bit move/split validation to
3990 external function.
3991 (rs6000_split_128bit_ok_p): Likewise.
3992 (rs6000_handle_altivec_attribute): Create V1TImode from vector
3993 __int128_t and vector __uint128_t.
3994
3995 * config/rs6000/vsx.md (VSX_L): Add V1TImode to vector iterators
3996 and mode attributes.
3997 (VSX_M): Likewise.
3998 (VSX_M2): Likewise.
3999 (VSm): Likewise.
4000 (VSs): Likewise.
4001 (VSr): Likewise.
4002 (VSv): Likewise.
4003 (VS_scalar): Likewise.
4004 (VS_double): Likewise.
0bb29a05 4005 (vsx_set_v1ti): New builtin function to create V1TImode from TImode.
a16a872d 4006
0bb29a05
UB
4007 * config/rs6000/rs6000.h (TARGET_VADDUQM): New macro to say whether
4008 we support the ISA 2.07 128-bit integer arithmetic instructions.
a16a872d
MM
4009 (ALTIVEC_OR_VSX_VECTOR_MODE): Add V1TImode.
4010 (enum rs6000_builtin_type_index): Add fields to hold V1TImode
4011 and TImode types for use with the builtin functions.
4012 (V1TI_type_node): Likewise.
4013 (unsigned_V1TI_type_node): Likewise.
4014 (intTI_type_internal_node): Likewise.
4015 (uintTI_type_internal_node): Likewise.
4016
0bb29a05
UB
4017 * config/rs6000/altivec.md (UNSPEC_VADDCUQ): New unspecs for ISA 2.07
4018 128-bit builtin functions.
a16a872d
MM
4019 (UNSPEC_VADDEUQM): Likewise.
4020 (UNSPEC_VADDECUQ): Likewise.
4021 (UNSPEC_VSUBCUQ): Likewise.
4022 (UNSPEC_VSUBEUQM): Likewise.
4023 (UNSPEC_VSUBECUQ): Likewise.
4024 (VM): Add V1TImode to vector mode iterators.
4025 (VM2): Likewise.
4026 (VI_unit): Likewise.
4027 (altivec_vadduqm): Add ISA 2.07 128-bit binary builtins.
4028 (altivec_vaddcuq): Likewise.
4029 (altivec_vsubuqm): Likewise.
4030 (altivec_vsubcuq): Likewise.
4031 (altivec_vaddeuqm): Likewise.
4032 (altivec_vaddecuq): Likewise.
4033 (altivec_vsubeuqm): Likewise.
4034 (altivec_vsubecuq): Likewise.
4035
4036 * config/rs6000/rs6000.md (FMOVE128_GPR): Add V1TImode to vector
4037 mode iterators.
4038 (BOOL_128): Likewise.
4039 (BOOL_REGS_OUTPUT): Likewise.
4040 (BOOL_REGS_OP1): Likewise.
4041 (BOOL_REGS_OP2): Likewise.
4042 (BOOL_REGS_UNARY): Likewise.
4043 (BOOL_REGS_AND_CR0): Likewise.
4044
4045 * config/rs6000/altivec.h (vec_vaddcuq): Add support for ISA 2.07
4046 128-bit integer builtin support.
4047 (vec_vadduqm): Likewise.
4048 (vec_vaddecuq): Likewise.
4049 (vec_vaddeuqm): Likewise.
4050 (vec_vsubecuq): Likewise.
4051 (vec_vsubeuqm): Likewise.
4052 (vec_vsubcuq): Likewise.
4053 (vec_vsubuqm): Likewise.
4054
4055 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
4056 Document vec_vaddcuq, vec_vadduqm, vec_vaddecuq, vec_vaddeuqm,
4057 vec_subecuq, vec_subeuqm, vec_vsubcuq, vec_vsubeqm builtins adding
4058 128-bit integer add/subtract to ISA 2.07.
4059
eeac7d15
JR
40602014-03-12 Joern Rennecke <joern.rennecke@embecosm.com>
4061
4062 * config/arc/arc.c (arc_predicate_delay_insns):
4063 Fix third argument passed to conditionalize_nonjump.
4064
bf0f324e
YZ
40652014-03-12 Yufeng Zhang <yufeng.zhang@arm.com>
4066
4067 * config/aarch64/aarch64-builtins.c
4068 (aarch64_builtin_vectorized_function): Add BUILT_IN_LFLOORF,
4069 BUILT_IN_LLFLOOR, BUILT_IN_LCEILF and BUILT_IN_LLCEIL.
4070 * config/aarch64/arm_neon.h (vcvtaq_u64_f64): Call __builtin_llfloor
4071 instead of __builtin_lfloor.
4072 (vcvtnq_u64_f64): Call __builtin_llceil instead of __builtin_lceil.
4073
bf4787b2
JJ
40742014-03-12 Jakub Jelinek <jakub@redhat.com>
4075
4076 * tree-ssa-ifcombine.c (forwarder_block_to): New function.
4077 (tree_ssa_ifcombine_bb_1): New function.
4078 (tree_ssa_ifcombine_bb): Use it. Handle also cases where else_bb
4079 is an empty forwarder block to then_bb or vice versa and then_bb
4080 and else_bb are effectively swapped.
4081
7d55b948
CB
40822014-03-12 Christian Bruel <christian.bruel@st.com>
4083
4084 PR target/60264
0bb29a05
UB
4085 * config/arm/arm.c (arm_emit_vfp_multi_reg_pop): Emit a
4086 REG_CFA_DEF_CFA note.
7d55b948
CB
4087 (arm_expand_epilogue_apcs_frame): call arm_add_cfa_adjust_cfa_note.
4088 (arm_unwind_emit): Allow REG_CFA_DEF_CFA.
4089
882a5fbe
TP
40902014-03-12 Thomas Preud'homme <thomas.preudhomme@arm.com>
4091
4092 PR tree-optimization/60454
4093 * tree-ssa-math-opts.c (find_bswap_1): Fix bswap detection.
4094
a3cd0246
KT
40952014-03-12 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4096
4097 * config.gcc (aarch64*-*-*): Use ISA flags from aarch64-arches.def.
4098 Do not define target_cpu_default2 to generic.
4099 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Use generic cpu.
4100 * config/aarch64/aarch64.c (aarch64_override_options): Update comment.
4101 * config/aarch64/aarch64-arches.def (armv8-a): Use generic cpu.
4102
c888139c
JJ
41032014-03-12 Jakub Jelinek <jakub@redhat.com>
4104 Marc Glisse <marc.glisse@inria.fr>
4105
4106 PR tree-optimization/60502
4107 * tree-ssa-reassoc.c (eliminate_not_pairs): Use build_all_ones_cst
4108 instead of build_low_bits_mask.
4109
b24ca895
JJ
41102014-03-12 Jakub Jelinek <jakub@redhat.com>
4111
4112 PR middle-end/60482
4113 * tree-vrp.c (register_edge_assert_for_1): Don't add assert
4114 if there are multiple uses, but op doesn't live on E edge.
4115 * tree-cfg.c (assert_unreachable_fallthru_edge_p): Also ignore
4116 clobber stmts before __builtin_unreachable.
4117
f1257268
RS
41182014-03-11 Richard Sandiford <rdsandiford@googlemail.com>
4119
4120 * builtins.c (expand_builtin_setjmp_receiver): Use and clobber
4121 hard_frame_pointer_rtx.
4122 * cse.c (cse_insn): Remove volatile check.
4123 * cselib.c (cselib_process_insn): Likewise.
4124 * dse.c (scan_insn): Likewise.
4125
bae56bbb
JR
41262014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4127
0bb29a05
UB
4128 * config/arc/arc.c (conditionalize_nonjump): New function,
4129 broken out of ...
4130 (arc_ifcvt): ... this.
bae56bbb
JR
4131 (arc_predicate_delay_insns): Use it.
4132
167ba5b9
JR
41332014-03-11 Joern Rennecke <joern.rennecke@embecosm.com>
4134
4135 * config/arc/predicates.md (extend_operand): During/after reload,
4136 allow const_int_operand.
4137 * config/arc/arc.md (mulsidi3_700): Use extend_operand predicate.
4138 (umulsidi3_700): Likewise. Change operand 2 constraint back to "cL".
4139 (mulsi3_highpart): Change operand 2 constraint alternatives 2 and 3
4140 to "i".
4141 (umulsi3_highpart_i): Likewise.
4142
dfd7d2d6
RB
41432014-03-11 Richard Biener <rguenther@suse.de>
4144
4145 * tree-ssa-structalias.c (get_constraint_for_ptr_offset):
4146 Add asserts to guard possible wrong-code bugs.
4147
af1ab449
RB
41482014-03-11 Richard Biener <rguenther@suse.de>
4149
4150 PR tree-optimization/60429
4151 PR tree-optimization/60485
4152 * tree-ssa-structalias.c (set_union_with_increment): Properly
4153 take into account all fields that overlap the shifted vars.
4154 (do_sd_constraint): Likewise.
4155 (do_ds_constraint): Likewise.
4156 (get_constraint_for_ptr_offset): Likewise.
4157
747425d0
CLT
41582014-03-11 Chung-Lin Tang <cltang@codesourcery.com>
4159
4160 * config/nios2/nios2.c (machine_function): Add fp_save_offset field.
4161 (nios2_compute_frame_layout):
4162 Add calculation of cfun->machine->fp_save_offset.
4163 (nios2_expand_prologue): Correct setting of frame pointer register
4164 in prologue.
4165 (nios2_expand_epilogue): Update recovery of stack pointer from
4166 frame pointer accordingly.
4167 (nios2_initial_elimination_offset): Update calculation of offset
4168 for eliminating to HARD_FRAME_POINTER_REGNUM.
4169
31dad809
JJ
41702014-03-10 Jakub Jelinek <jakub@redhat.com>
4171
4172 PR ipa/60457
4173 * ipa.c (symtab_remove_unreachable_nodes): Don't call
4174 cgraph_get_create_node on VAR_DECLs.
4175
3c1c08d6
RB
41762014-03-10 Richard Biener <rguenther@suse.de>
4177
4178 PR middle-end/60474
4179 * tree.c (signed_or_unsigned_type_for): Handle OFFSET_TYPEs.
4180
21d89690
DR
41812014-03-08 Douglas B Rupp <rupp@gnat.com>
4182
4183 * config/vms/vms.opt (vms_float_format): New variable.
4184
634b8e9b
TB
41852014-03-08 Tobias Burnus <burnus@net-b.de>
4186
0bb29a05 4187 * doc/invoke.texi (-fcilkplus): Update implementation status.
634b8e9b 4188
f42c637e
PM
41892014-03-08 Paulo Matos <paulo@matos-sorge.com>
4190 Richard Biener <rguenther@suse.de>
4191
4192 * lto-wrapper.c (merge_and_complain): Ensure -fshort-double is used
4193 consistently accross all TUs.
4194 (run_gcc): Enable -fshort-double automatically at link at link-time
4195 and disallow override.
4196
bd65fc87
RS
41972014-03-08 Richard Sandiford <rdsandiford@googlemail.com>
4198
4199 PR target/58271
4200 * config/mips/mips.c (mips_option_override): Promote -mpaired-single
4201 warning to an error. Disable TARGET_PAIRED_SINGLE and TARGET_MIPS3D
4202 if they can't be used.
4203
040c446d
RO
42042014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4205
0bb29a05
UB
4206 * configure.ac (HAVE_AS_IX86_TLSLDMPLT): Improve test
4207 for Solaris 11/x86 ld.
040c446d
RO
4208 * configure: Regenerate.
4209
caa16d41
RO
42102014-03-07 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
4211
4212 * configure.ac (TLS_SECTION_ASM_FLAG): Save as tls_section_flag.
4213 (LIB_TLS_SPEC): Save as ld_tls_libs.
4214 (HAVE_AS_IX86_TLSLDMPLT): Define as 1/0.
4215 (HAVE_AS_IX86_TLSLDM): New test.
4216 * configure, config.in: Regenerate.
4217 * config/i386/i386.c (legitimize_tls_address): Fall back to
4218 TLS_MODEL_GLOBAL_DYNAMIC on 32-bit Solaris/x86 if tool chain
4219 cannot support TLS_MODEL_LOCAL_DYNAMIC.
4220 * config/i386/i386.md (*tls_local_dynamic_base_32_gnu): Use if
4221 instead of #ifdef in HAVE_AS_IX86_TLSLDMPLT test.
4222
f3d83ffe
PM
42232014-03-07 Paulo Matos <paulo@matos-sorge.com>
4224
4225 * common.opt (fira-loop-pressure): Mark as optimization.
4226
b17a8b07
TS
42272014-03-07 Thomas Schwinge <thomas@codesourcery.com>
4228
4229 * langhooks.c (lhd_omp_mappable_type): The error_mark_node is not
4230 an OpenMP mappable type.
4231
4973b0f9
MK
42322014-03-06 Matthias Klose <doko@ubuntu.com>
4233
4234 * Makefile.in (s-mlib): Only pass MULTIARCH_DIRNAME if
4235 MULTILIB_OSDIRNAMES is not defined.
4236
ab19cc6c
JJ
42372014-03-06 Jakub Jelinek <jakub@redhat.com>
4238 Meador Inge <meadori@codesourcery.com>
4239
4240 PR target/58595
4241 * config/arm/arm.c (arm_tls_symbol_p): Remove.
4242 (arm_legitimize_address): Call legitimize_tls_address for any
4243 arm_tls_referenced_p expression, handle constant addend. Call it
4244 before testing for !TARGET_ARM.
4245 (thumb_legitimize_address): Don't handle arm_tls_symbol_p here.
4246
a74158c7
RB
42472014-03-06 Richard Biener <rguenther@suse.de>
4248
4249 PR middle-end/60445
4250 PR lto/60424
4251 PR lto/60427
4252 Revert
4253 2014-03-04 Paulo Matos <paulo@matos-sorge.com>
4254
0bb29a05
UB
4255 * tree-streamer.c (record_common_node): Assert we don't record
4256 nodes with type double.
4257 (preload_common_node): Skip type double, complex double and double
4258 pointer since it is now frontend dependent due to fshort-double option.
a74158c7 4259
f3ba16d0
RB
42602014-03-06 Richard Biener <rguenther@suse.de>
4261
4262 * gcc.c (PLUGIN_COND): Always enable unless -fno-use-linker-plugin
4263 or -fno-lto is specified and the linker has full plugin support.
0bb29a05
UB
4264 * collect2.c (lto_mode): Default to LTO_MODE_WHOPR if LTO is enabled.
4265 (main): Remove -flto processing, adjust lto_mode using use_plugin late.
f3ba16d0
RB
4266 * lto-wrapper.c (merge_and_complain): Merge compile-time
4267 optimization levels.
4268 (run_gcc): And pass it through to the link options.
4269
4bb66ef3 42702014-03-06 Alexandre Oliva <aoliva@redhat.com>
80662856
AO
4271
4272 PR debug/60381
4273 Revert:
4bb66ef3 4274 2014-02-28 Alexandre Oliva <aoliva@redhat.com>
80662856
AO
4275 PR debug/59992
4276 * cselib.c (remove_useless_values): Skip to avoid quadratic
4277 behavior if the condition moved from...
4278 (cselib_process_insn): ... here holds.
4279
a05cca9b
JJ
42802014-03-05 Jakub Jelinek <jakub@redhat.com>
4281
487125e7
JJ
4282 PR plugins/59335
4283 * Makefile.in (PLUGIN_HEADERS): Add tree-phinodes.h, stor-layout.h,
4284 ssa-iterators.h, $(RESOURCE_H) and tree-cfgcleanup.h.
4285
a05cca9b
JJ
4286 PR plugins/59335
4287 * config/i386/t-i386 (OPTIONS_H_EXTRA): Add stringop.def.
4288 (TM_H): Add x86-tune.def.
4289
4e2cd668
KT
42902014-03-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4291
4292 * config/aarch64/aarch64.c (generic_tunings):
4293 Use cortexa57_extra_costs.
4294
5525ed38
JJ
42952014-03-05 Jakub Jelinek <jakub@redhat.com>
4296
4297 PR lto/60404
4298 * cfgexpand.c (expand_used_vars): Do not assume all SSA_NAMEs
4299 of PARM/RESULT_DECLs must be coalesced with optimize && in_lto_p.
4300 * tree-ssa-coalesce.c (coalesce_ssa_name): Use MUST_COALESCE_COST - 1
4301 cost for in_lto_p.
4302
0c72fa78
H
43032014-03-04 Heiher <r@hev.cc>
4304
4305 * config/mips/mips-cpus.def (loongson3a): Mark as a MIPS64r2 processor.
4306 * config/mips/mips.h (MIPS_ISA_LEVEL_SPEC): Adjust accordingly.
4307
66b03f81
UB
43082014-03-04 Uros Bizjak <ubizjak@gmail.com>
4309
4310 * config/i386/predicates.md (const2356_operand): Change to ...
4311 (const2367_operand): ... this.
4312 * config/i386/sse.md (avx512pf_scatterpf<mode>sf): Use
4313 const2367_operand.
4314 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4315 (*avx512pf_scatterpf<mode>sf): Ditto.
4316 (avx512pf_scatterpf<mode>df): Ditto.
4317 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4318 (*avx512pf_scatterpf<mode>df): Ditto.
4319 * config/i386/i386.c (ix86_expand_builtin): Update
4320 incorrect hint operand error message.
4321
3c24e842
RB
43222014-03-04 Richard Biener <rguenther@suse.de>
4323
4324 * lto-section-in.c (lto_get_section_data): Fix const cast.
66b03f81 4325
3740bda6
PM
43262014-03-04 Paulo Matos <paulo@matos-sorge.com>
4327
0bb29a05
UB
4328 * tree-streamer.c (record_common_node): Assert we don't record
4329 nodes with type double.
4330 (preload_common_node): Skip type double, complex double and double
4331 pointer since it is now frontend dependent due to fshort-double option.
3740bda6 4332
91856735
RB
43332014-03-04 Richard Biener <rguenther@suse.de>
4334
4335 PR lto/60405
66b03f81 4336 * lto-streamer-in.c (lto_read_body): Remove LTO bytecode version check.
91856735
RB
4337 (lto_input_toplevel_asms): Likewise.
4338 * lto-section-in.c (lto_get_section_data): Instead do it here
4339 for every section.
66b03f81 4340
b75f28e1
RB
43412014-03-04 Richard Biener <rguenther@suse.de>
4342
4343 PR tree-optimization/60382
4344 * tree-vect-loop.c (vect_is_simple_reduction_1): Do not consider
4345 dead PHIs a reduction.
4346
dd96e83a
UB
43472014-03-03 Uros Bizjak <ubizjak@gmail.com>
4348
4349 * config/i386/xmmintrin.h (enum _mm_hint) <_MM_HINT_ET0>: Correct
4350 hint value.
4351 (_mm_prefetch): Move out of GCC target("sse") pragma.
4352 * config/i386/prfchwintrin.h (_m_prefetchw): Move out of
4353 GCC target("prfchw") pragma.
4354 * config/i386/i386.md (prefetch): Emit prefetchwt1 only
4355 for locality <= 2.
4356 * config/i386/i386.c (ix86_option_override_internal): Enable
4357 -mprfchw with -mprefetchwt1.
4358
220c1a51
JR
43592014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4360
4361 * config/arc/arc.md (casesi_load) <length attribute alternative 0>:
4362 Mark as varying.
4363
212bfe71
JR
43642014-03-03 Joern Rennecke <joern.rennecke@embecosm.com>
4365
4366 * opts.h (CL_PCH_IGNORE): Define.
4367 * targhooks.c (option_affects_pch_p):
4368 Return false for options that have CL_PCH_IGNORE set.
4369 * opt-functions.awk: Process PchIgnore.
4370 * doc/options.texi: Document PchIgnore.
4371
7a3d0a39
JR
4372 * config/arc/arc.opt (misize): Add PchIgnore property.
4373
b46ed885
BS
43742014-03-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4375
4376 * config/rs6000/rs6000.c (rs6000_preferred_reload_class): Disallow
4377 reload of PLUS rtx's outside of GENERAL_REGS or BASE_REGS; relax
4378 constraint on constants to permit them being loaded into
4379 GENERAL_REGS or BASE_REGS.
4380
c6f709ec
NC
43812014-03-03 Nick Clifton <nickc@redhat.com>
4382
4383 * config/rl78/rl78-real.md (cbranchsi4_real_signed): Add
4384 anti-cacnonical alternatives.
4385 (negandhi3_real): New pattern.
4386 * config/rl78/rl78-virt.md (negandhi3_virt): New pattern.
4387
5752f78f
SKS
43882014-03-03 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
4389
4390 * config/avr/avr-mcus.def: Remove atxmega16x1.
4391 * config/avr/avr-tables.opt: Regenerate.
4392 * config/avr/t-multilib: Regenerate.
4393 * doc/avr-mmcu.texi: Regenerate.
4394
0afe7332 43952014-03-03 Tobias Grosser <tobias@grosser.es>
dd96e83a 4396 Mircea Namolaru <mircea.namolaru@inria.fr>
0afe7332
TG
4397
4398 PR tree-optimization/58028
4399 * graphite-clast-to-gimple.c (set_cloog_options): Don't remove
4400 scalar dimensions.
4401
3416dd87
RR
44022014-03-03 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4403
4404 * config/arm/neon.md (*movmisalign<mode>): Legitimize addresses
4405 not handled by recognizers.
4406
79c7fda6
JJ
44072014-03-03 Jakub Jelinek <jakub@redhat.com>
4408
4409 PR middle-end/60175
4410 * function.c (expand_function_end): Don't emit
4411 clobber_return_register sequence if clobber_after is a BARRIER.
4412 * cfgexpand.c (construct_exit_block): Append instructions before
4413 return_label to prev_bb.
4414
715a5c85
BS
44152014-03-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4416
4417 * config/rs6000/constraints.md: Document reserved use of "wc".
4418
9ffbf271
JH
44192014-03-02 Jan Hubicka <hubicka@ucw.cz>
4420
4421 PR ipa/60150
4422 * ipa.c (function_and_variable_visibility): When dissolving comdat
4423 group, also set all symbols to local.
4424
993df21e 44252014-03-02 Jan Hubicka <hubicka@ucw.cz>
c6f709ec 4426
993df21e
JH
4427 PR ipa/60306
4428
4429 Revert:
4430 2013-12-14 Jan Hubicka <jh@suse.cz>
0bb29a05
UB
4431 PR middle-end/58477
4432 * ipa-prop.c (stmt_may_be_vtbl_ptr_store): Skip clobbers.
993df21e 4433
9b0af790
JB
44342014-03-02 Jon Beniston <jon@beniston.com>
4435
4436 PR bootstrap/48230
4437 PR bootstrap/50927
4438 PR bootstrap/52466
4439 PR target/46898
4440 * config/lm32/lm32.c (lm32_legitimate_constant_p): Remove, as incorrect.
c6f709ec 4441 (TARGET_LEGITIMATE_CONSTANT_P): Undefine, as not needed.
9b0af790 4442 * config/lm32/lm32.md (movsi_insn): Add 32-bit immediate support.
c6f709ec 4443 (simple_return, *simple_return): New patterns
9b0af790
JB
4444 * config/lm32/predicates.md (movsi_rhs_operand): Remove as obsolete.
4445 * configure.ac (force_sjlj_exceptions): Force sjlj exceptions for lm32.
4446
c582aac9
PC
44472014-03-01 Paolo Carlini <paolo.carlini@oracle.com>
4448
4449 * dwarf2out.c (gen_subprogram_die): Tidy.
4450
19305875
OE
44512014-03-01 Oleg Endo <olegendo@gcc.gnu.org>
4452
4453 PR target/60071
4454 * config/sh/sh.md (*mov_t_msb_neg): Split into ...
4455 (*mov_t_msb_neg_negc): ... this new insn.
4456
570215f9
JM
44572014-02-28 Jason Merrill <jason@redhat.com>
4458
4459 PR c++/58678
4460 * ipa-devirt.c (ipa_devirt): Don't choose an implicitly-declared
4461 function.
4462
c5b001b5
PC
44632014-02-28 Paolo Carlini <paolo.carlini@oracle.com>
4464
4465 PR c++/60314
4466 * dwarf2out.c (decltype_auto_die): New static.
4467 (gen_subprogram_die): Handle 'decltype(auto)' like 'auto'.
4468 (gen_type_die_with_usage): Handle 'decltype(auto)'.
4469 (is_cxx_auto): Likewise.
4470
5750e120
IB
44712014-02-28 Ian Bolton <ian.bolton@arm.com>
4472
4473 * config/aarch64/aarch64.h: Define __ARM_NEON by default if
4474 we are not using general regs only.
4475
fba7c564
RB
44762014-02-28 Richard Biener <rguenther@suse.de>
4477
4478 PR target/60280
4479 * tree-cfgcleanup.c (tree_forwarder_block_p): Restrict
4480 previous fix and only allow to remove trivial pre-headers
4481 and latches. Also honor LOOPS_MAY_HAVE_MULTIPLE_LATCHES.
84d7e312 4482 (remove_forwarder_block): Properly update the latch of a loop.
fba7c564 4483
4bb66ef3 44842014-02-28 Alexandre Oliva <aoliva@redhat.com>
0618dee5
AO
4485
4486 PR debug/59992
4487 * cselib.c (cselib_hasher::equal): Special-case VALUE lookup.
4488 (cselib_preserved_hash_table): New.
4489 (preserve_constants_and_equivs): Move preserved vals to it.
4490 (cselib_find_slot): Look it up first.
4491 (cselib_init): Initialize it.
4492 (cselib_finish): Release it.
4493 (dump_cselib_table): Dump it.
4494
4bb66ef3 44952014-02-28 Alexandre Oliva <aoliva@redhat.com>
22414f3f
AO
4496
4497 PR debug/59992
4498 * cselib.c (remove_useless_values): Skip to avoid quadratic
4499 behavior if the condition moved from...
4500 (cselib_process_insn): ... here holds.
4501
4bb66ef3 45022014-02-28 Alexandre Oliva <aoliva@redhat.com>
2e084917
AO
4503
4504 PR debug/57232
4505 * var-tracking.c (vt_initialize): Apply the same condition to
4506 preserve the CFA base value.
4507
9910c53c
JY
45082014-02-28 Joey Ye <joey.ye@arm.com>
4509
4510 PR target/PR60169
4511 * config/arm/arm.c (thumb_far_jump_used_p): Don't change
4512 if reload in progress or completed.
4513
ef271341
TB
45142014-02-28 Tobias Burnus <burnus@net-b.de>
4515
4516 PR middle-end/60147
4517 * tree-pretty-print.c (dump_generic_node, print_declaration): Handle
4518 NAMELIST_DECL.
4519
c4f6267b
L
45202014-02-27 H.J. Lu <hongjiu.lu@intel.com>
4521
4522 * doc/tm.texi.in (Condition Code Status): Update documention for
4523 relative locations of cc0-setter and cc0-user.
4524
7a76df7f
JL
45252014-02-27 Jeff Law <law@redhat.com>
4526
4527 PR rtl-optimization/52714
4528 * combine.c (try_combine): When splitting an unrecognized PARALLEL
4529 into two independent simple sets, if I3 is a jump, ensure the
84d7e312 4530 pattern we place into I3 is a (set (pc) ...).
7a76df7f 4531
728acca0
MP
45322014-02-27 Mikael Pettersson <mikpe@it.uu.se>
4533 Jeff Law <law@redhat.com>
4534
4535 PR rtl-optimization/49847
4536 * cse.c (fold_rtx) Handle case where cc0 setter and cc0 user
4537 are in different blocks.
4538 * doc/tm.texi (Condition Code Status): Update documention for
4539 relative locations of cc0-setter and cc0-user.
4540
93aaa05d
VM
45412014-02-27 Vladimir Makarov <vmakarov@redhat.com>
4542
4543 PR target/59222
4544 * lra.c (lra_emit_add): Check SUBREG too.
4545
1ee6eb01
AS
45462014-02-27 Andreas Schwab <schwab@suse.de>
4547
4548 * config/m68k/m68k.c (m68k_option_override): Disable
4549 -flive-range-shrinkage for classic m68k.
4550 (m68k_override_options_after_change): Likewise.
4551
bb0d8e84
MP
45522014-02-27 Marek Polacek <polacek@redhat.com>
4553
4554 PR middle-end/59223
4555 * tree-ssa-uninit.c (gate_warn_uninitialized): Run the pass even for
4556 -Wmaybe-uninitialized.
4557
3157b879
AM
45582014-02-27 Alan Modra <amodra@gmail.com>
4559
3e17cad2 4560 PR target/57936
3157b879
AM
4561 * reload1.c (emit_input_reload_insns): When reload_override_in,
4562 set old to rl->in_reg when rl->in_reg is a subreg.
4563
bb750f4f
RB
45642014-02-26 Richard Biener <rguenther@suse.de>
4565
4566 PR bootstrap/60343
4567 * lra-assigns.c (spill_for): Avoid mixed-sign comparison.
4568
22c8aab3
IT
45692014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4570
4571 * common/config/i386/predicates.md (const1256_operand): Remove.
4572 (const2356_operand): New.
4573 (const_1_to_2_operand): Remove.
4574 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
4575 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
4576 (*avx512pf_gatherpf<mode>sf): Ditto.
4577 (avx512pf_gatherpf<mode>df): Ditto.
4578 (*avx512pf_gatherpf<mode>df_mask): Ditto.
4579 (*avx512pf_gatherpf<mode>df): Ditto.
4580 (avx512pf_scatterpf<mode>sf): Ditto.
4581 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
4582 (*avx512pf_scatterpf<mode>sf): Ditto.
4583 (avx512pf_scatterpf<mode>df): Ditto.
4584 (*avx512pf_scatterpf<mode>df_mask): Ditto.
4585 (*avx512pf_scatterpf<mode>df): Ditto.
4586 * common/config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET0.
4587
260d3642
IT
45882014-02-26 Ilya Tocar <ilya.tocar@intel.com>
4589
4590 * config/i386/avx512fintrin.h (_mm512_testn_epi32_mask),
4591 (_mm512_mask_testn_epi32_mask), (_mm512_testn_epi64_mask),
4592 (_mm512_mask_testn_epi64_mask): Move to ...
4593 * config/i386/avx512cdintrin.h: Here.
4594 * config/i386/i386.c (bdesc_args): Change MASK_ISA for testnm.
4595 * config/i386/sse.md (avx512f_vmscalef<mode><round_name>): Remove %.
4596 (avx512f_scalef<mode><mask_name><round_name>): Ditto.
4597 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Change conditon to
4598 TARGET_AVX512F from TARGET_AVX512CD.
4599
477145c8
RB
46002014-02-26 Richard Biener <rguenther@suse.de>
4601
4602 PR ipa/60327
4603 * ipa.c (walk_polymorphic_call_targets): Properly guard
4604 call to inline_update_overall_summary.
4605
a3afdbb8
BC
46062014-02-26 Bin Cheng <bin.cheng@arm.com>
4607
4608 PR target/60280
84d7e312
UB
4609 * tree-cfgcleanup.c (tree_forwarder_block_p): Protect loop preheaders
4610 and latches only if requested. Fix latch if it is removed.
a3afdbb8
BC
4611 * tree-ssa-dom.c (tree_ssa_dominator_optimize): Set
4612 LOOPS_HAVE_PREHEADERS.
4613
b8a542c6
AP
46142014-02-25 Andrew Pinski <apinski@cavium.com>
4615
4616 * builtins.c (expand_builtin_thread_pointer): Create a new target
4617 when the target is NULL.
4618
88def637
VM
46192014-02-25 Vladimir Makarov <vmakarov@redhat.com>
4620
4621 PR rtl-optimization/60317
4622 * params.def (PARAM_LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4623 * params.h (LRA_MAX_CONSIDERED_RELOAD_PSEUDOS): New.
4624 * lra-assigns.c: Include params.h.
4625 (spill_for): Use LRA_MAX_CONSIDERED_RELOAD_PSEUDOS as guard for
4626 other reload pseudos considerations.
4627
68908a06
BS
46282014-02-25 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4629
cf36d2cc
WS
4630 * config/rs6000/vector.md (*vector_unordered<mode>): Change split
4631 to use canonical form for nor<mode>3.
68908a06 4632
db57bbc9
KT
46332014-02-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4634
4635 PR target/55426
4636 * config/arm/arm.h (CANNOT_CHANGE_MODE_CLASS): Allow 128 to 64-bit
4637 conversions.
4638
43b3f52f
IT
46392014-02-25 Ilya Tocar <ilya.tocar@intel.com>
4640
4641 * common/config/i386/i386-common.c (OPTION_MASK_ISA_PREFETCHWT1_SET),
4642 (OPTION_MASK_ISA_PREFETCHWT1_UNSET): New.
4643 (ix86_handle_option): Handle OPT_mprefetchwt1.
4644 * config/i386/cpuid.h (bit_PREFETCHWT1): New.
4645 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
4646 PREFETCHWT1 CPUID.
4647 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
4648 OPTION_MASK_ISA_PREFETCHWT1.
4649 * config/i386/i386.c (ix86_target_string): Handle mprefetchwt1.
4650 (PTA_PREFETCHWT1): New.
4651 (ix86_option_override_internal): Handle PTA_PREFETCHWT1.
4652 (ix86_valid_target_attribute_inner_p): Handle OPT_mprefetchwt1.
84d7e312 4653 * config/i386/i386.h (TARGET_PREFETCHWT1, TARGET_PREFETCHWT1_P): New.
43b3f52f
IT
4654 * config/i386/i386.md (prefetch): Check TARGET_PREFETCHWT1
4655 (*prefetch_avx512pf_<mode>_: Change into ...
84d7e312 4656 (*prefetch_prefetchwt1_<mode>: This.
43b3f52f
IT
4657 * config/i386/i386.opt (mprefetchwt1): New.
4658 * config/i386/xmmintrin.h (_mm_hint): Add _MM_HINT_ET1.
4659 (_mm_prefetch): Handle intent to write.
4660 * doc/invoke.texi (mprefetchwt1), (mno-prefetchwt1): Doccument.
4661
84053e02
RB
46622014-02-25 Richard Biener <rguenther@suse.de>
4663
4664 PR middle-end/60291
4665 * emit-rtl.c (mem_attrs_htab): Remove.
4666 (mem_attrs_htab_hash): Likewise.
4667 (mem_attrs_htab_eq): Likewise.
84d7e312 4668 (set_mem_attrs): Always allocate new mem-attrs when something changed.
84053e02
RB
4669 (init_emit_once): Do not allocate mem_attrs_htab.
4670
4094757e
RB
46712014-02-25 Richard Biener <rguenther@suse.de>
4672
4673 PR lto/60319
4674 * lto-opts.c (lto_write_options): Output non-explicit conservative
4675 -fwrapv, -fno-trapv and -fno-strict-overflow.
4676 * lto-wrapper.c (merge_and_complain): Handle merging those options.
4677 (run_gcc): And pass them through.
4678
41b2d514 46792014-02-25 Andrey Belevantsev <abel@ispras.ru>
43b3f52f 4680
41b2d514
AB
4681 * sel-sched.c (calculate_new_fences): New parameter ptime.
4682 Calculate it as a maximum over all fence cycles.
4683 (sel_sched_region_2): Adjust the call to calculate_new_fences.
4684 Print the final schedule timing when sched_verbose.
4685
c1c99405
AB
46862014-02-25 Andrey Belevantsev <abel@ispras.ru>
4687
4688 PR rtl-optimization/60292
4689 * sel-sched.c (fill_vec_av_set): Do not reset target availability
4690 bit fot the fence instruction.
4691
1d9def42
AD
46922014-02-24 Alangi Derick <alangiderick@gmail.com>
4693
4694 * calls.h: Fix typo in comment.
4695
325fefe0
JDA
46962014-02-24 John David Anglin <danglin@gcc.gnu.org>
4697
4698 * config/pa/pa.c (pa_output_move_double): Don't valididate when
4699 adjusting offsetable addresses.
4700
ad43b47a
GW
47012014-02-24 Guozhi Wei <carrot@google.com>
4702
4703 * sparseset.h (sparseset_pop): Fix the wrong index.
4704
341c653c
WL
47052014-02-24 Walter Lee <walt@tilera.com>
4706
4707 * config.gcc (tilepro-*-*): Change to tilepro*-*-*.
4708 (tilegx-*-linux*): Change to tilegx*-*-linux*; Support tilegxbe
4709 triplet.
4710 * common/config/tilegx/tilegx-common.c
4711 (TARGET_DEFAULT_TARGET_FLAGS): Define.
4712 * config/tilegx/linux.h (ASM_SPEC): Add endian_spec.
4713 (LINK_SPEC): Ditto.
4714 * config/tilegx/sync.md (atomic_test_and_set): Handle big endian.
4715 * config/tilegx/tilegx.c (tilegx_return_in_msb): New.
4716 (tilegx_gimplify_va_arg_expr): Handle big endian.
4717 (tilegx_expand_unaligned_load): Ditto.
4718 (tilegx_expand_unaligned_store): Ditto.
4719 (TARGET_RETURN_IN_MSB): New.
4720 * config/tilegx/tilegx.h (TARGET_DEFAULT): New.
4721 (TARGET_ENDIAN_DEFAULT): New.
4722 (TARGET_BIG_ENDIAN): Handle big endian.
4723 (BYTES_BIG_ENDIAN): Ditto.
4724 (WORDS_BIG_ENDIAN): Ditto.
4725 (FLOAT_WORDS_BIG_ENDIAN): Ditto.
4726 (ENDIAN_SPEC): New.
4727 (EXTRA_SPECS): New.
4728 * config/tilegx/tilegx.md (extv): Handle big endian.
4729 (extzv): Ditto.
4730 (insn_st<n>): Ditto.
4731 (insn_st<n>_add<bitsuffix>): Ditto.
4732 (insn_stnt<n>): Ditto.
4733 (insn_stnt<n>_add<bitsuffix>):Ditto.
4734 (vec_interleave_highv8qi): Handle big endian.
4735 (vec_interleave_highv8qi_be): New.
4736 (vec_interleave_highv8qi_le): New.
4737 (insn_v1int_h): Handle big endian.
4738 (vec_interleave_lowv8qi): Handle big endian.
4739 (vec_interleave_lowv8qi_be): New.
4740 (vec_interleave_lowv8qi_le): New.
4741 (insn_v1int_l): Handle big endian.
4742 (vec_interleave_highv4hi): Handle big endian.
4743 (vec_interleave_highv4hi_be): New.
4744 (vec_interleave_highv4hi_le): New.
4745 (insn_v2int_h): Handle big endian.
4746 (vec_interleave_lowv4hi): Handle big endian.
4747 (vec_interleave_lowv4hi_be): New.
4748 (vec_interleave_lowv4hi_le): New.
4749 (insn_v2int_l): Handle big endian.
4750 (vec_interleave_highv2si): Handle big endian.
4751 (vec_interleave_highv2si_be): New.
4752 (vec_interleave_highv2si_le): New.
4753 (insn_v4int_h): Handle big endian.
4754 (vec_interleave_lowv2si): Handle big endian.
4755 (vec_interleave_lowv2si_be): New.
4756 (vec_interleave_lowv2si_le): New.
4757 (insn_v4int_l): Handle big endian.
4758 * config/tilegx/tilegx.opt (mbig-endian): New option.
4759 (mlittle-endian): New option.
4760 * doc/install.texi: Document tilegxbe-linux.
4761 * doc/invoke.texi: Document -mbig-endian and -mlittle-endian.
4762
f3fec19f
MJ
47632014-02-24 Martin Jambor <mjambor@suse.cz>
4764
4765 PR ipa/60266
4766 * ipa-cp.c (propagate_constants_accross_call): Bail out early if
4767 there are no parameter descriptors.
4768
9039622a
AB
47692014-02-24 Andrey Belevantsev <abel@ispras.ru>
4770
4771 PR rtl-optimization/60268
4772 * sched-rgn.c (haifa_find_rgns): Move the nr_regions_initial variable
4773 initialization to ...
4774 (sched_rgn_init): ... here.
4775 (schedule_region): Check for SCHED_PRESSURE_NONE earlier.
4776
4bb66ef3 47772014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
e7e7bc4b 4778
9039622a
AB
4779 * config/microblaze/microblaze.md: Correct ashrsi_reg / lshrsi_reg
4780 names.
e7e7bc4b 4781
4bb66ef3 47822014-02-23 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
8011ccca
ME
4783
4784 * config/microblaze/microblaze.h: Remove SECONDARY_MEMORY_NEEDED
4785 definition.
4786
4bb66ef3 47872014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
c332c7df 4788
84d7e312
UB
4789 * /config/microblaze/microblaze.c: Add microblaze_asm_output_mi_thunk,
4790 define TARGET_ASM_OUTPUT_MI_THUNK and TARGET_ASM_CAN_OUTPUT_MI_THUNK.
c332c7df 4791
4bb66ef3 47922014-02-23 David Holsgrove <david.holsgrove@xilinx.com>
d5a19af1
DH
4793
4794 * config/microblaze/predicates.md: Add cmp_op predicate.
84d7e312 4795 * config/microblaze/microblaze.md: Add branch_compare instruction
d5a19af1 4796 which uses cmp_op predicate and emits cmp insn before branch.
84d7e312 4797 * config/microblaze/microblaze.c (microblaze_emit_compare): Rename
d5a19af1
DH
4798 to microblaze_expand_conditional_branch and consolidate logic.
4799 (microblaze_expand_conditional_branch): emit branch_compare
4800 insn instead of handling cmp op separate from branch insn.
4801
34c25d23
BS
48022014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4803
4804 * config/rs6000/rs6000.c (rs6000_emit_le_vsx_move): Relax assert
4805 to permit subregs.
4806
a6eecdc1
BS
48072014-02-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4808
d5a19af1
DH
4809 * config/rs6000/altivec.md (altivec_lve<VI_char>x): Replace
4810 define_insn with define_expand and new define_insn
4811 *altivec_lve<VI_char>x_internal.
4812 (altivec_stve<VI_char>x): Replace define_insn with define_expand
4813 and new define_insn *altivec_stve<VI_char>x_internal.
4814 * config/rs6000/rs6000-protos.h (altivec_expand_stvex_be): New
4815 prototype.
4816 * config/rs6000/rs6000.c (altivec_expand_lvx_be): Document use by
4817 lve*x built-ins.
4818 (altivec_expand_stvex_be): New function.
a6eecdc1 4819
8aa7d1fc
JR
48202014-02-22 Joern Rennecke <joern.rennecke@embecosm.com>
4821
4822 * config/avr/avr.c (avr_can_eliminate): Allow elimination from
84d7e312 4823 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM if !frame_pointer_needed.
8aa7d1fc
JR
4824 * config/avr/avr.c (ELIMINABLE_REGS): Add elimination from
4825 ARG_POINTER_REGNUM to STACK_POINTER_REGNUM.
4826
a810ee82
VM
48272014-02-21 Vladimir Makarov <vmakarov@redhat.com>
4828
4829 PR target/60298
4830 * lra-constraints.c (inherit_reload_reg): Use lra_emit_move
4831 instead of emit_move_insn.
4832
b90ab1ba
BS
48332014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4834
4835 * config/rs6000/altivec.md (altivec_vsumsws): Replace second
4836 vspltw with vsldoi.
4837 (reduc_uplus_v16qi): Use gen_altivec_vsumsws_direct instead of
4838 gen_altivec_vsumsws.
4839
1af73690
WS
48402014-02-21 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4841
4842 * config/rs6000/altivec.md (altivec_lvxl): Rename as
84d7e312 4843 *altivec_lvxl_<mode>_internal and use VM2 iterator instead of V4SI.
1af73690
WS
4844 (altivec_lvxl_<mode>): New define_expand incorporating
4845 -maltivec=be semantics where needed.
4846 (altivec_lvx): Rename as *altivec_lvx_<mode>_internal.
4847 (altivec_lvx_<mode>): New define_expand incorporating -maltivec=be
4848 semantics where needed.
4849 (altivec_stvx): Rename as *altivec_stvx_<mode>_internal.
4850 (altivec_stvx_<mode>): New define_expand incorporating
4851 -maltivec=be semantics where needed.
4852 (altivec_stvxl): Rename as *altivec_stvxl_<mode>_internal and use
4853 VM2 iterator instead of V4SI.
4854 (altivec_stvxl_<mode>): New define_expand incorporating
4855 -maltivec=be semantics where needed.
4856 * config/rs6000/rs6000-builtin.def: Add new built-in definitions
4857 LVXL_V2DF, LVXL_V2DI, LVXL_V4SF, LVXL_V4SI, LVXL_V8HI, LVXL_V16QI,
84d7e312
UB
4858 LVX_V2DF, LVX_V2DI, LVX_V4SF, LVX_V4SI, LVX_V8HI, LVX_V16QI, STVX_V2DF,
4859 STVX_V2DI, STVX_V4SF, STVX_V4SI, STVX_V8HI, STVX_V16QI, STVXL_V2DF,
4860 STVXL_V2DI, STVXL_V4SF, STVXL_V4SI, STVXL_V8HI, STVXL_V16QI.
1af73690
WS
4861 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Replace
4862 ALTIVEC_BUILTIN_LVX with ALTIVEC_BUILTIN_LVX_<MODE> throughout;
4863 similarly for ALTIVEC_BUILTIN_LVXL, ALTIVEC_BUILTIN_STVX, and
4864 ALTIVEC_BUILTIN_STVXL.
84d7e312 4865 * config/rs6000/rs6000-protos.h (altivec_expand_lvx_be): New prototype.
1af73690
WS
4866 (altivec_expand_stvx_be): Likewise.
4867 * config/rs6000/rs6000.c (swap_selector_for_mode): New function.
4868 (altivec_expand_lvx_be): Likewise.
4869 (altivec_expand_stvx_be): Likewise.
4870 (altivec_expand_builtin): Add cases for
4871 ALTIVEC_BUILTIN_STVX_<MODE>, ALTIVEC_BUILTIN_STVXL_<MODE>,
4872 ALTIVEC_BUILTIN_LVXL_<MODE>, and ALTIVEC_BUILTIN_LVX_<MODE>.
4873 (altivec_init_builtins): Add definitions for
4874 __builtin_altivec_lvxl_<mode>, __builtin_altivec_lvx_<mode>,
84d7e312 4875 __builtin_altivec_stvx_<mode>, and __builtin_altivec_stvxl_<mode>.
1af73690 4876
0a39d07b
CM
48772014-02-21 Catherine Moore <clm@codesourcery.com>
4878
4879 * doc/invoke.texi (mvirt, mno-virt): Document.
4880 * config/mips/mips.opt (mvirt): New option.
4881 * config/mips/mips.h (ASM_SPEC): Pass mvirt to the assembler.
4882
f2556b68
RB
48832014-02-21 Richard Biener <rguenther@suse.de>
4884
4885 PR tree-optimization/60276
4886 * tree-vectorizer.h (struct _stmt_vec_info): Add min_neg_dist field.
4887 (STMT_VINFO_MIN_NEG_DIST): New macro.
4888 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Record
4889 STMT_VINFO_MIN_NEG_DIST.
4890 * tree-vect-stmts.c (vectorizable_load): Verify if assumptions
4891 made for negative dependence distances still hold.
4892
32417082
RB
48932014-02-21 Richard Biener <rguenther@suse.de>
4894
4895 PR middle-end/60291
4896 * tree-ssa-live.c (mark_all_vars_used_1): Do not walk
4897 DECL_INITIAL for globals not in the current function context.
4898
92261ce0
JJ
48992014-02-21 Jakub Jelinek <jakub@redhat.com>
4900
4901 PR tree-optimization/56490
4902 * params.def (PARAM_UNINIT_CONTROL_DEP_ATTEMPTS): New param.
4903 * tree-ssa-uninit.c: Include params.h.
4904 (compute_control_dep_chain): Add num_calls argument, return false
4905 if it exceed PARAM_UNINIT_CONTROL_DEP_ATTEMPTS param, pass
4906 num_calls to recursive call.
4907 (find_predicates): Change dep_chain into normal array,
4908 cur_chain into auto_vec<edge, MAX_CHAIN_LEN + 1>, add num_calls
4909 variable and adjust compute_control_dep_chain caller.
4910 (find_def_preds): Likewise.
4911
aa6ef874
TS
49122014-02-21 Thomas Schwinge <thomas@codesourcery.com>
4913
4914 * gimple-pretty-print.c (dump_gimple_omp_for) [flags & TDF_RAW]
4915 <case GF_OMP_FOR_KIND_CILKSIMD>: Add missing break statement.
4916
4b156fd0
NC
49172014-02-21 Nick Clifton <nickc@redhat.com>
4918
4919 * config/stormy16/stormy16.md (pushdqi1): Add mode to post_inc.
4920 (pushhi1): Likewise.
4921 (popqi1): Add mode to pre_dec.
4922 (pophi1): Likewise.
4923
dffd569e
JJ
49242014-02-21 Jakub Jelinek <jakub@redhat.com>
4925
4926 * config/i386/i386.c (ix86_expand_vec_perm): Use V8SImode
4927 mode for mask of V8SFmode permutation.
4928
2fdc29e8
RH
49292014-02-20 Richard Henderson <rth@redhat.com>
4930
4931 PR c++/60272
4932 * builtins.c (expand_builtin_atomic_compare_exchange): Always make
4933 a new pseudo for OLDVAL.
4934
aa637f66
JJ
49352014-02-20 Jakub Jelinek <jakub@redhat.com>
4936
4937 PR target/57896
4938 * config/i386/i386.c (expand_vec_perm_interleave2): Don't call
4939 gen_reg_rtx if d->testing_p.
4940 (expand_vec_perm_pshufb2, expand_vec_perm_broadcast_1): Return early
4941 if d->testing_p and we will certainly return true.
4942 (expand_vec_perm_even_odd_1): Likewise. Don't call gen_reg_rtx
4943 if d->testing_p.
4944
004a7e45
UB
49452014-02-20 Uros Bizjak <ubizjak@gmail.com>
4946
4947 * emit-rtl.c (gen_reg_rtx): Assert that
4948 crtl->emit.regno_pointer_align_length is non-zero.
4949
49502014-02-20 Richard Henderson <rth@redhat.com>
672ce939
RH
4951
4952 PR c++/60272
4953 * builtins.c (expand_builtin_atomic_compare_exchange): Conditionalize
4954 on failure the store back into EXPECT.
4955
95ce7613
CLT
49562014-02-20 Chung-Lin Tang <cltang@codesourcery.com>
4957 Sandra Loosemore <sandra@codesourcery.com>
4958
4959 * config/nios2/nios2.md (unspec): Add UNSPEC_PIC_GOTOFF_SYM enum.
84d7e312
UB
4960 * config/nios2/nios2.c (nios2_function_profiler): Add
4961 -fPIC (flag_pic == 2) support.
95ce7613
CLT
4962 (nios2_handle_custom_fpu_cfg): Fix warning parameter.
4963 (nios2_large_offset_p): New function.
4964 (nios2_unspec_reloc_p): Move up position, update to use
4965 nios2_large_offset_p.
4966 (nios2_unspec_address): Remove function.
4967 (nios2_unspec_offset): New function.
4968 (nios2_large_got_address): New function.
4969 (nios2_got_address): Add large offset support.
4970 (nios2_legitimize_tls_address): Update usage of removed and new
4971 functions.
4972 (nios2_symbol_binds_local_p): New function.
4973 (nios2_load_pic_address): Add -fPIC (flag_pic == 2) support.
4974 (nios2_legitimize_address): Update to use nios2_large_offset_p.
4975 (nios2_emit_move_sequence): Avoid legitimizing (const (unspec ...)).
4976 (nios2_print_operand): Merge H/L processing, add hiadj/lo
4977 processing for (const (unspec ...)).
4978 (nios2_unspec_reloc_name): Add UNSPEC_PIC_GOTOFF_SYM case.
4979
efd2d3c8
RB
49802014-02-20 Richard Biener <rguenther@suse.de>
4981
4982 * tree-cfg.c (replace_uses_by): Mark altered BBs before
4983 doing the substitution.
004a7e45 4984 (verify_gimple_assign_single): Also verify bare MEM_REFs on the lhs.
efd2d3c8 4985
9576e7b1
MJ
49862014-02-20 Martin Jambor <mjambor@suse.cz>
4987
4988 PR ipa/55260
4989 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Uce correct
4990 info when checking whether lattices are bottom.
4991
25fe40b0
RB
49922014-02-20 Richard Biener <rguenther@suse.de>
4993
4994 PR middle-end/60221
4995 * tree-eh.c (execute_cleanup_eh_1): Also cleanup empty EH
4996 regions at -O0.
4997
bd936951
JH
49982014-02-20 Jan Hubicka <hubicka@ucw.cz>
4999
5000 PR ipa/58555
004a7e45
UB
5001 * ipa-inline-transform.c (clone_inlined_nodes): Add freq_scale
5002 parameter specifying the scaling.
bd936951
JH
5003 (inline_call): Update.
5004 (want_inline_recursively): Guard division by zero.
5005 (recursive_inlining): Update.
5006 * ipa-inline.h (clone_inlined_nodes): Update.
5007
3c898e1a
IT
50082014-02-20 Ilya Tocar <ilya.tocar@intel.com>
5009
5010 PR target/60204
5011 * config/i386/i386.c (classify_argument): Pass structures of size
5012 64 bytes or less in register.
5013
df62b4af 50142014-02-20 Ilya Tocar <ilya.tocar@intel.com>
004a7e45 5015 Kirill Yukhin <kirill.yukhin@intel.com>
df62b4af
IT
5016
5017 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): Swap operands.
5018 (_mm_rcp28_round_ss): Ditto.
5019 (_mm_rsqrt28_round_sd): Ditto.
5020 (_mm_rsqrt28_round_ss): Ditto.
5021 * config/i386/avx512erintrin.h (_mm_rcp14_round_sd): Ditto.
5022 (_mm_rcp14_round_ss): Ditto.
5023 (_mm_rsqrt14_round_sd): Ditto.
5024 (_mm_rsqrt14_round_ss): Ditto.
5025 * config/i386/sse.md (rsqrt14<mode>): Put nonimmediate operand as
5026 the first input operand, get rid of match_dup.
5027 (avx512er_exp2<mode><mask_name><round_saeonly_name>): Set type
5028 attribute to sse.
5029 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
5030 Ditto.
5031 (avx512er_vmrcp28<mode><round_saeonly_name>): Put nonimmediate
5032 operand as the first input operand, set type attribute.
5033 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
5034 Set type attribute.
5035 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Put nonimmediate
5036 operand as the first input operand, set type attribute.
5037
9254148e
BS
50382014-02-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5039
5040 * config/rs6000/rs6000.c (vspltis_constant): Fix most significant
5041 bit of zero.
5042
fff91793
L
50432014-02-19 H.J. Lu <hongjiu.lu@intel.com>
5044
5045 PR target/60207
5046 * config/i386/i386.c (construct_container): Remove TFmode check
5047 for X86_64_INTEGER_CLASS.
5048
6aa5b4b8
UB
50492014-02-19 Uros Bizjak <ubizjak@gmail.com>
5050
5051 PR target/59794
5052 * config/i386/i386.c (type_natural_mode): Warn for ABI changes
5053 only when -Wpsabi is enabled.
5054
aadc1c43
MHD
50552014-02-19 Michael Hudson-Doyle <michael.hudson@linaro.org>
5056
5057 PR target/59799
5058 * config/aarch64/aarch64.c (aarch64_pass_by_reference): The rules for
5059 passing arrays in registers are the same as for structs, so remove the
5060 special case for them.
5061
322913f8
EB
50622014-02-19 Eric Botcazou <ebotcazou@adacore.com>
5063
5064 * expr.c (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: For a bit-field
5065 destination type, extract only the valid bits if the source type is not
5066 integral and has a different mode.
5067
fd9710dc
RB
50682014-02-19 Richard Biener <rguenther@suse.de>
5069
5070 PR ipa/60243
5071 * tree-inline.c (estimate_num_insns): Avoid calling cgraph_get_node
5072 for all calls.
5073
4df65a85
RB
50742014-02-19 Richard Biener <rguenther@suse.de>
5075
5076 PR ipa/60243
5077 * ipa-prop.c: Include stringpool.h and tree-ssanames.h.
5078 (ipa_modify_call_arguments): Emit an argument load explicitely and
5079 preserve virtual SSA form there and for the replacement call.
5080 Do not update SSA form nor free dominance info.
5081
7fea98d8
JH
50822014-02-18 Jan Hubicka <hubicka@ucw.cz>
5083
5084 * ipa.c (function_and_variable_visibility): Also clear WEAK
5085 flag when disolving COMDAT_GROUP.
5086
0a2550e7
JH
50872014-02-18 Jan Hubicka <hubicka@ucw.cz>
5088
5089 * ipa-prop.h (ipa_ancestor_jf_data): Update ocmment.
5090 * ipa-prop.c (ipa_set_jf_known_type): Return early when
5091 not devirtualizing.
5092 (ipa_set_ancestor_jf): Set type to NULL hwen it is not preserved;
5093 do more sanity checks.
5094 (detect_type_change): Return true when giving up early.
5095 (compute_complex_assign_jump_func): Fix type parameter of
5096 ipa_set_ancestor_jf.
5097 (compute_complex_ancestor_jump_func): Likewise.
5098 (update_jump_functions_after_inlining): Fix updating of
5099 ancestor function.
6aa5b4b8 5100 * ipa-cp.c (ipa_get_jf_ancestor_result): Be ready for type to be NULL.
0a2550e7 5101
2dbe8b70
JH
51022014-02-18 Jan Hubicka <hubicka@ucw.cz>
5103
5104 * cgraph.c (cgraph_update_edges_for_call_stmt_node): Also remove
5105 inline clones when edge disappears.
5106
b9809dc4
MM
51072014-02-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5108
5109 PR target/60203
5110 * config/rs6000/rs6000.md (mov<mode>_64bit, TF/TDmode moves):
5111 Split 64-bit moves into 2 patterns. Do not allow the use of
5112 direct move for TDmode in little endian, since the decimal value
5113 has little endian bytes within a word, but the 64-bit pieces are
5114 ordered in a big endian fashion, and normal subreg's of TDmode are
5115 not allowed.
5116 (mov<mode>_64bit_dm): Likewise.
5117 (movtd_64bit_nodm): Likewise.
5118
bababbfb
EB
51192014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5120
5121 PR tree-optimization/60174
5122 * tree-ssa-reassoc.c (init_range_entry): Do not look into the defining
5123 statement of an SSA_NAME that occurs in an abnormal PHI node.
5124
70aacc97
JJ
51252014-02-18 Jakub Jelinek <jakub@redhat.com>
5126
5127 PR sanitizer/60142
5128 * final.c (SEEN_BB): Remove.
5129 (SEEN_NOTE, SEEN_EMITTED): Renumber.
5130 (final_scan_insn): Don't force_source_line on second
5131 NOTE_INSN_BASIC_BLOCK.
5132
223cdd15
UB
51332014-02-18 Uros Bizjak <ubizjak@gmail.com>
5134
5135 PR target/60205
5136 * config/i386/i386.h (struct ix86_args): Add warn_avx512f.
5137 * config/i386/i386.c (init_cumulative_args): Initialize warn_avx512f.
5138 (type_natural_mode): Warn ABI change when %zmm register is not
5139 available for AVX512F vector value passing.
5140
48810be0
KT
51412014-02-18 Kai Tietz <ktietz@redhat.com>
5142
5143 PR target/60193
223cdd15
UB
5144 * config/i386/i386.c (ix86_expand_prologue): Use value in
5145 rax register as displacement when restoring %r10 or %rax.
5146 Fix wrong offset when restoring both registers.
48810be0 5147
20afe640
EB
51482014-02-18 Eric Botcazou <ebotcazou@adacore.com>
5149
5150 * ipa-prop.c (compute_complex_ancestor_jump_func): Replace overzealous
5151 assertion with conditional return.
5152
d0b50387
JJ
51532014-02-18 Jakub Jelinek <jakub@redhat.com>
5154 Uros Bizjak <ubizjak@gmail.com>
5155
5156 PR driver/60233
5157 * config/i386/driver-i386.c (host_detect_local_cpu): If
5158 YMM state is not saved by the OS, also clear has_f16c. Move
5159 CPUID 0x80000001 handling before YMM state saving checking.
5160
c4cd7435
AB
51612014-02-18 Andrey Belevantsev <abel@ispras.ru>
5162
5163 PR rtl-optimization/58960
5164 * haifa-sched.c (alloc_global_sched_pressure_data): New,
5165 factored out from ...
5166 (sched_init): ... here.
5167 (free_global_sched_pressure_data): New, factored out from ...
5168 (sched_finish): ... here.
5169 * sched-int.h (free_global_sched_pressure_data): Declare.
5170 * sched-rgn.c (nr_regions_initial): New static global.
5171 (haifa_find_rgns): Initialize it.
5172 (schedule_region): Disable sched-pressure for the newly
5173 generated regions.
5174
f0281fde
RB
51752014-02-17 Richard Biener <rguenther@suse.de>
5176
5177 * tree-vect-stmts.c (free_stmt_vec_info): Clear BB and
5178 release SSA defs of pattern stmts.
5179
c742772c
RB
51802014-02-17 Richard Biener <rguenther@suse.de>
5181
5182 * tree-inline.c (expand_call_inline): Release the virtual
5183 operand defined by the call we are about to inline.
5184
0492158e
RB
51852014-02-17 Richard Biener <rguenther@suse.de>
5186
5187 * tree-ssa.c (verify_ssa): If verify_def found an error, ICE.
5188
583a9919
KY
51892014-02-17 Kirill Yukhin <kirill.yukhin@intel.com>
5190 Ilya Tocar <ilya.tocar@intel.com>
5191
5192 * config/i386/avx512fintrin.h (_mm512_maskz_permutexvar_epi64): Swap
5193 arguments order in builtin.
5194 (_mm512_permutexvar_epi64): Ditto.
5195 (_mm512_mask_permutexvar_epi64): Ditto
5196 (_mm512_maskz_permutexvar_epi32): Ditto
5197 (_mm512_permutexvar_epi32): Ditto
5198 (_mm512_mask_permutexvar_epi32): Ditto
5199
d737743f
BS
52002014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5201
223cdd15 5202 * config/rs6000/altivec.md (p8_vmrgew): Handle little endian targets.
d737743f
BS
5203 (p8_vmrgow): Likewise.
5204
54c4bfd7
BS
52052014-02-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5206
5207 * config/rs6000/vsx.md (vsx_xxpermdi_<mode>): Handle little
5208 endian targets.
5209
518fea64
MM
52102014-02-15 Michael Meissner <meissner@linux.vnet.ibm.com>
5211
5212 PR target/60203
5213 * config/rs6000/rs6000.md (rreg): Add TFmode, TDmode constraints.
5214 (mov<mode>_internal, TFmode/TDmode): Split TFmode/TDmode moves
5215 into 64-bit and 32-bit moves. On 64-bit moves, add support for
5216 using direct move instructions on ISA 2.07. Also adjust
5217 instruction length for 64-bit.
5218 (mov<mode>_64bit, TFmode/TDmode): Likewise.
5219 (mov<mode>_32bit, TFmode/TDmode): Likewise.
5220
61640916
AM
52212014-02-15 Alan Modra <amodra@gmail.com>
5222
5223 PR target/58675
5224 PR target/57935
5225 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
5226 find_replacement on parts of insn rtl that might be reloaded.
5227
a2b33cc3
RB
52282014-02-15 Richard Biener <rguenther@suse.de>
5229
5230 PR tree-optimization/60183
223cdd15 5231 * tree-ssa-phiprop.c (propagate_with_phi): Avoid speculating loads.
a2b33cc3
RB
5232 (tree_ssa_phiprop): Calculate and free post-dominators.
5233
059742a4
JL
52342014-02-14 Jeff Law <law@redhat.com>
5235
5236 PR rtl-optimization/60131
5237 * ree.c (get_extended_src_reg): New function.
223cdd15 5238 (combine_reaching_defs): Use it rather than assuming location of REG.
059742a4
JL
5239 (find_and_remove_re): Verify first operand of extension is
5240 a REG before adding the insns to the copy list.
5241
88f7c49a
RM
52422014-02-14 Roland McGrath <mcgrathr@google.com>
5243
5244 * configure.ac (HAVE_AS_IX86_UD2): New test for 'ud2' mnemonic.
5245 * configure: Regenerated.
5246 * config.in: Regenerated.
5247 * config/i386/i386.md (trap) [HAVE_AS_IX86_UD2]: Use the mnemonic
5248 instead of ASM_SHORT.
5249
2ae577fd
VM
52502014-02-14 Vladimir Makarov <vmakarov@redhat.com>
5251 Richard Earnshaw <rearnsha@arm.com>
5252
5253 PR rtl-optimization/59535
5254 * lra-constraints.c (process_alt_operands): Encourage alternative
5255 when unassigned pseudo class is superset of the alternative class.
5256 (inherit_reload_reg): Don't inherit when optimizing for code size.
5257 * config/arm/arm.h (MODE_BASE_REG_CLASS): Add version for LRA
5258 returning CORE_REGS for anything but Thumb1 and BASE_REGS for
5259 modes not less than 4 for Thumb1.
5260
5d88af08
KM
52612014-02-14 Kyle McMartin <kyle@redhat.com>
5262
5263 PR pch/60010
5264 * config/host-linux.c (TRY_EMPTY_VM_SPACE): Define for AArch64.
5265
3069b421
RB
52662014-02-14 Richard Biener <rguenther@suse.de>
5267
5268 * cilk-common.c (cilk_arrow): Build a MEM_REF, not an INDIRECT_REF.
5269 (get_frame_arg): Drop the assert with langhook types_compatible_p.
5270 Do not strip INDIRECT_REFs.
5271
1966fd99
RB
52722014-02-14 Richard Biener <rguenther@suse.de>
5273
5274 PR lto/60179
5275 * lto-streamer-out.c (DFS_write_tree_body): Do not follow
5276 DECL_FUNCTION_SPECIFIC_TARGET.
5277 (hash_tree): Do not hash DECL_FUNCTION_SPECIFIC_TARGET.
5278 * tree-streamer-out.c (pack_ts_target_option): Remove.
223cdd15 5279 (streamer_pack_tree_bitfields): Do not stream TS_TARGET_OPTION.
1966fd99
RB
5280 (write_ts_function_decl_tree_pointers): Do not stream
5281 DECL_FUNCTION_SPECIFIC_TARGET.
5282 * tree-streamer-in.c (unpack_ts_target_option): Remove.
5283 (unpack_value_fields): Do not stream TS_TARGET_OPTION.
5284 (lto_input_ts_function_decl_tree_pointers): Do not stream
5285 DECL_FUNCTION_SPECIFIC_TARGET.
5286
b010d601
JJ
52872014-02-14 Jakub Jelinek <jakub@redhat.com>
5288
223cdd15 5289 * tree-vect-loop.c (vect_is_slp_reduction): Don't set use_stmt twice.
b010d601
JJ
5290 (get_initial_def_for_induction, vectorizable_induction): Ignore
5291 debug stmts when looking for exit_phi.
5292 (vectorizable_live_operation): Fix up condition.
5293
f2dafb91
CJW
52942014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5295
5296 * config/nds32/nds32.c (nds32_asm_function_prologue): Do not use
5297 nreverse() because it changes the content of original tree list.
5298
59043e75
CJW
52992014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5300
5301 * config/nds32/t-mlibs (MULTILIB_OPTIONS): Fix typo in comment.
5302 * config/nds32/nds32.c (nds32_merge_decl_attributes): Likewise.
5303
810f736f
CJW
53042014-02-14 Chung-Ju Wu <jasonwucj@gmail.com>
5305
5306 * config/nds32/nds32.c (nds32_naked_function_p): Follow the
5307 GNU coding standards.
5308
1a025bbf
JJ
53092014-02-13 Jakub Jelinek <jakub@redhat.com>
5310
5311 PR debug/60152
5312 * dwarf2out.c (gen_subprogram_die): Don't call
5313 add_calling_convention_attribute if subr_die is old_die.
5314
69479ebd
SS
53152014-02-13 Sharad Singhai <singhai@google.com>
5316
5317 * doc/optinfo.texi: Fix order of nodes.
5318
1287ae50
UB
53192014-02-13 Uros Bizjak <ubizjak@gmail.com>
5320
5321 * config/i386/sse.md (xop_vmfrcz<mode>2): Generate const0 in
5322 operands[2], not operands[3].
5323
43372236
RB
53242014-02-13 Richard Biener <rguenther@suse.de>
5325
5326 PR bootstrap/59878
5327 * doc/install.texi (ISL): Update recommended version to 0.12.2,
5328 mention the possibility of an in-tree build.
5329 (CLooG): Update recommended version to 0.18.1, mention the
5330 possibility of an in-tree build and clarify that the ISL
5331 bundled with CLooG does not work.
5332
a4d70cfa
JJ
53332014-02-13 Jakub Jelinek <jakub@redhat.com>
5334
5335 PR target/43546
5336 * expr.c (compress_float_constant): If x is a hard register,
5337 extend into a pseudo and then move to x.
5338
e697d119
DV
53392014-02-13 Dominik Vogt <vogt@linux.vnet.ibm.com>
5340
5341 * config/s390/s390.c (s390_asm_output_function_label): Fix crash
5342 caused by bad second argument to warning_at() with -mhotpatch and
5343 nested functions (e.g. with gfortran).
5344
9f8da907
RS
53452014-02-13 Richard Sandiford <rdsandiford@googlemail.com>
5346
5347 * opts.c (option_name): Remove "enabled by default" rider.
5348
0fdd1196
JDA
53492014-02-12 John David Anglin <danglin@gcc.gnu.org>
5350
5351 * config/pa/pa.c (pa_option_override): Remove auto increment FIXME.
5352
0544c448
L
53532014-02-12 H.J. Lu <hongjiu.lu@intel.com>
5354 Uros Bizjak <ubizjak@gmail.com>
5355
5356 PR target/60151
1287ae50 5357 * configure.ac (HAVE_AS_GOTOFF_IN_DATA): Pass --32 to GNU assembler.
0544c448
L
5358 * configure: Regenerated.
5359
3a938d75
RB
53602014-02-12 Richard Biener <rguenther@suse.de>
5361
5362 * vec.c (vec_prefix::calculate_allocation): Move as
5363 inline variant to vec.h.
5364 (vec_prefix::calculate_allocation_1): New out-of-line version.
5365 * vec.h (vec_prefix::calculate_allocation_1): Declare.
5366 (vec_prefix::m_has_auto_buf): Rename to ...
5367 (vec_prefix::m_using_auto_storage): ... this.
5368 (vec_prefix::calculate_allocation): Inline the easy cases
5369 and dispatch to calculate_allocation_1 which doesn't need the
5370 prefix address.
5371 (va_heap::reserve): Use gcc_checking_assert.
5372 (vec<T, A, vl_embed>::embedded_init): Add argument to initialize
5373 m_using_auto_storage.
5374 (auto_vec): Change m_vecpfx member to a vec<T, va_heap, vl_embed>
5375 member and adjust.
5376 (vec<T, va_heap, vl_ptr>::reserve): Remove redundant check.
5377 (vec<T, va_heap, vl_ptr>::release): Avoid casting.
5378 (vec<T, va_heap, vl_ptr>::using_auto_storage): Simplify.
5379
ad0188be
RB
53802014-02-12 Richard Biener <rguenther@suse.de>
5381
5382 * gcse.c (compute_transp): break from loop over canon_modify_mem_list
5383 when we found a dependence.
5384
64e5ace5
TS
53852014-02-12 Thomas Schwinge <thomas@codesourcery.com>
5386
88ac13da
TS
5387 * gimplify.c (gimplify_call_expr, gimplify_modify_expr): Move
5388 common code...
5389 (maybe_fold_stmt): ... into this new function.
5390 * omp-low.c (lower_omp): Update comment.
5391
bae729a2
TS
5392 * omp-low.c (lower_omp_target): Add clobber for sizes array, after
5393 last use.
5394
64e5ace5
TS
5395 * omp-low.c (diagnose_sb_0): Make sure label_ctx is valid to
5396 dereference.
5397
7b40f5cf
JG
53982014-02-12 James Greenhalgh <james.greenhalgh@arm.com>
5399
5400 * config/arm/aarch-cost-tables.h (generic_extra_costs): Fix
5401 identifiers in comments.
5402 (cortexa53_extra_costs): Likewise.
1287ae50 5403 * config/arm/arm.c (cortexa9_extra_costs): Fix identifiers in comments.
7b40f5cf
JG
5404 (cortexa7_extra_costs): Likewise.
5405 (cortexa12_extra_costs): Likewise.
5406 (cortexa15_extra_costs): Likewise.
5407 (v7m_extra_costs): Likewise.
5408
c4c8514e
RB
54092014-02-12 Richard Biener <rguenther@suse.de>
5410
5411 PR middle-end/60092
5412 * gimple-low.c (lower_builtin_posix_memalign): Lower conditional
5413 of posix_memalign being successful.
5414 (lower_stmt): Restrict lowering of posix_memalign to when
5415 -ftree-bit-ccp is enabled.
5416
8eb651bd
SKS
54172014-02-12 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
5418
5419 * config/avr/avr-c.c (avr_resolve_overloaded_builtin): Pass vNULL for
5420 arg_loc.
5421 * config/spu/spu-c.c (spu_resolve_overloaded_builtin): Likewise.
5422
62b03553
EB
54232014-02-12 Eric Botcazou <ebotcazou@adacore.com>
5424
5425 PR rtl-optimization/60116
5426 * combine.c (try_combine): Also remove dangling REG_DEAD notes on the
5427 other_insn once the combination has been validated.
5428
ec77d61f
JH
54292014-02-11 Jan Hubicka <hubicka@ucw.cz>
5430
5431 PR lto/59468
5432 * ipa-utils.h (possible_polymorphic_call_targets): Update prototype
5433 and wrapper.
5434 * ipa-devirt.c: Include demangle.h
5435 (odr_violation_reported): New static variable.
5436 (add_type_duplicate): Update odr_violations.
5437 (maybe_record_node): Add completep parameter; update it.
5438 (record_target_from_binfo): Add COMPLETEP parameter;
5439 update it as needed.
5440 (possible_polymorphic_call_targets_1): Likewise.
5441 (struct polymorphic_call_target_d): Add nonconstruction_targets;
5442 rename FINAL to COMPLETE.
5443 (record_targets_from_bases): Sanity check we found the binfo;
5444 fix COMPLETEP updating.
5445 (possible_polymorphic_call_targets): Add NONCONSTRUTION_TARGETSP
5446 parameter, fix computing of COMPLETEP.
1287ae50
UB
5447 (dump_possible_polymorphic_call_targets): Imrove readability of dump;
5448 at LTO time do demangling.
ec77d61f
JH
5449 (ipa_devirt): Use nonconstruction_targets; Improve dumps.
5450 * gimple-fold.c (gimple_get_virt_method_for_vtable): Add can_refer
5451 parameter.
5452 (gimple_get_virt_method_for_binfo): Likewise.
5453 * gimple-fold.h (gimple_get_virt_method_for_binfo,
5454 gimple_get_virt_method_for_vtable): Update prototypes.
5455
5a4dcd9b
VM
54562014-02-11 Vladimir Makarov <vmakarov@redhat.com>
5457
5458 PR target/49008
5459 * genautomata.c (add_presence_absence): Fix typo with
5460 {final_}presence_list.
5461
69b7afed
MM
54622014-02-11 Michael Meissner <meissner@linux.vnet.ibm.com>
5463
5464 PR target/60137
5465 * config/rs6000/rs6000.md (128-bit GPR splitter): Add a splitter
5466 for VSX/Altivec vectors that land in GPR registers.
5467
8268ad5c
JJ
54682014-02-11 Richard Henderson <rth@redhat.com>
5469 Jakub Jelinek <jakub@redhat.com>
5470
5471 PR debug/59776
5472 * tree-sra.c (load_assign_lhs_subreplacements): Add VIEW_CONVERT_EXPR
5473 around drhs if type conversion to lacc->type is not useless.
5474
4fd92af6
KT
54752014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5476
5477 * config/aarch64/aarch64-cores.def (cortex-a57): Use cortexa57
5478 tuning struct.
5479 (cortex-a57.cortex-a53): Likewise.
5480 * config/aarch64/aarch64.c (cortexa57_tunings): New tuning struct.
5481
7cb14cb8
KT
54822014-02-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5483
5484 * config/arm/thumb2.md (*thumb2_movhi_insn): Add alternatives for
5485 arm_restrict_it.
5486
6d3715b9
RL
54872014-02-11 Renlin Li <Renlin.Li@arm.com>
5488
5489 * doc/sourcebuild.texi: Document check_effective_target_arm_vfp3_ok and
5490 add_options_for_arm_vfp3.
5491
cddddfff
JL
54922014-02-11 Jeff Law <law@redhat.com>
5493
5494 PR middle-end/54041
5495 * expr.c (expand_expr_addr_expr_1): Handle expand_expr returning an
5496 object with an undesirable mode.
5497
7c1aef7e
RO
54982014-02-11 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5499
5500 PR libgomp/60107
5501 * config/i386/sol2-9.h: New file.
5502 * config.gcc (i[34567]86-*-solaris2* | x86_64-*-solaris2.1[0-9]*,
5503 *-*-solaris2.9*): Use it.
5504
4bb66ef3 55052014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
d63f1311
NM
5506
5507 * config/microblaze/microblaze.md: Add movsi4_rev insn pattern.
5508 * config/microblaze/predicates.md: Add reg_or_mem_operand predicate.
5509
4bb66ef3 55102014-02-10 Nagaraju Mekala <nagaraju.mekala@xilinx.com>
ed8b71cc
NM
5511
5512 * config/microblaze/microblaze.c: Extend mcpu version format
5513
4bb66ef3 55142014-02-10 David Holsgrove <david.holsgrove@xilinx.com>
9b5414c6
DH
5515
5516 * config/microblaze/microblaze.h: Define SIZE_TYPE and PTRDIFF_TYPE.
5517
004a7e45 55182014-02-10 Richard Henderson <rth@redhat.com>
f830ddc2
RH
5519
5520 PR target/59927
5521 * calls.c (expand_call): Don't double-push for reg_parm_stack_space.
5522 * config/i386/i386.c (init_cumulative_args): Remove sorry for 64-bit
5523 ms-abi vs -mno-accumulate-outgoing-args.
5524 (ix86_expand_prologue): Unconditionally call ix86_eax_live_at_start_p.
5525 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Fix comment with
5526 respect to ms-abi.
5527
e2fc3b4f
BE
55282014-02-10 Bernd Edlinger <bernd.edlinger@hotmail.de>
5529
5530 PR middle-end/60080
5531 * cfgexpand.c (expand_asm_operands): Attach source location to
5532 ASM_INPUT rtx objects.
5533 * print-rtl.c (print_rtx): Check for UNKNOWN_LOCATION.
5534
3a22ad89
NC
55352014-02-10 Nick Clifton <nickc@redhat.com>
5536
5537 * config/mn10300/mn10300.c (popcount): New function.
5538 (mn10300_expand_prologue): Include saved registers in stack usage
5539 count.
5540
f27be550
JL
55412014-02-10 Jeff Law <law@redhat.com>
5542
5543 PR middle-end/52306
5544 * reload1.c (emit_input_reload_insns): Do not create invalid RTL
1287ae50 5545 when changing the SET_DEST of a prior insn to avoid an input reload.
f27be550 5546
e15e3815
UW
55472014-02-10 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
5548
5549 * config/rs6000/sysv4.h (ENDIAN_SELECT): Do not attempt to enforce
5550 big-endian mode for -mcall-aixdesc, -mcall-freebsd, -mcall-netbsd,
5551 -mcall-openbsd, or -mcall-linux.
5552 (CC1_ENDIAN_BIG_SPEC): Remove.
5553 (CC1_ENDIAN_LITTLE_SPEC): Remove.
5554 (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5555 (CC1_SPEC): Remove (always empty) %cc1_endian_... spec.
5556 (SUBTARGET_EXTRA_SPECS): Remove %cc1_endian_big, %cc1_endian_little,
5557 and %cc1_endian_default.
5558 * config/rs6000/sysv4le.h (CC1_ENDIAN_DEFAULT_SPEC): Remove.
5559
4e1f39e4
RB
55602014-02-10 Richard Biener <rguenther@suse.de>
5561
5562 PR tree-optimization/60115
5563 * tree-eh.c (tree_could_trap_p): Unify TARGET_MEM_REF and
5564 MEM_REF handling. Properly verify that the accesses are not
5565 out of the objects bound.
5566
e7af1c22
KT
55672014-02-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5568
5569 * config/aarch64/aarch64.c (aarch64_override_options): Fix typo from
5570 coretex to cortex.
5571
79c7de84
EB
55722014-02-10 Eric Botcazou <ebotcazou@adacore.com>
5573
5574 * ipa-devirt.c (get_polymorphic_call_info_from_invariant): Return
5575 proper constants and fix formatting.
5576 (possible_polymorphic_call_targets): Fix formatting.
5577
cf73ee60
KY
55782014-02-10 Kirill Yukhin <kirill.yukhin@intel.com>
5579 Ilya Tocar <ilya.tocar@intel.com>
5580
5581 * config/i386/avx512fintrin.h (_mm512_storeu_epi64): Removed.
5582 (_mm512_loadu_epi32): Renamed into...
5583 (_mm512_loadu_si512): This.
5584 (_mm512_storeu_epi32): Renamed into...
5585 (_mm512_storeu_si512): This.
5586 (_mm512_maskz_ceil_ps): Removed.
5587 (_mm512_maskz_ceil_pd): Ditto.
5588 (_mm512_maskz_floor_ps): Ditto.
5589 (_mm512_maskz_floor_pd): Ditto.
5590 (_mm512_floor_round_ps): Ditto.
5591 (_mm512_floor_round_pd): Ditto.
5592 (_mm512_ceil_round_ps): Ditto.
5593 (_mm512_ceil_round_pd): Ditto.
5594 (_mm512_mask_floor_round_ps): Ditto.
5595 (_mm512_mask_floor_round_pd): Ditto.
5596 (_mm512_mask_ceil_round_ps): Ditto.
5597 (_mm512_mask_ceil_round_pd): Ditto.
5598 (_mm512_maskz_floor_round_ps): Ditto.
5599 (_mm512_maskz_floor_round_pd): Ditto.
5600 (_mm512_maskz_ceil_round_ps): Ditto.
5601 (_mm512_maskz_ceil_round_pd): Ditto.
5602 (_mm512_expand_pd): Ditto.
5603 (_mm512_expand_ps): Ditto.
5604 * config/i386/i386.c (ix86_builtins): Remove
5605 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK.
5606 (bdesc_args): Ditto.
5607 * config/i386/predicates.md (const1256_operand): New.
5608 (const_1_to_2_operand): Ditto.
5609 * config/i386/sse.md (avx512pf_gatherpf<mode>sf): Change hint value.
5610 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
5611 (*avx512pf_gatherpf<mode>sf): Ditto.
5612 (avx512pf_gatherpf<mode>df): Ditto.
5613 (*avx512pf_gatherpf<mode>df_mask): Ditto.
5614 (*avx512pf_gatherpf<mode>df): Ditto.
5615 (avx512pf_scatterpf<mode>sf): Ditto.
5616 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
5617 (*avx512pf_scatterpf<mode>sf): Ditto.
5618 (avx512pf_scatterpf<mode>df): Ditto.
5619 (*avx512pf_scatterpf<mode>df_mask): Ditto.
5620 (*avx512pf_scatterpf<mode>df): Ditto.
5621 (avx512f_expand<mode>): Removed.
5622 (<shift_insn><mode>3<mask_name>): Change predicate type.
5623
8fcbce72
JJ
56242014-02-08 Jakub Jelinek <jakub@redhat.com>
5625
41475e96
JJ
5626 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers
5627 not at the end of datarefs vector use ordered_remove to avoid
5628 reordering datarefs vector.
5629
c74559df
JJ
5630 PR c/59984
5631 * gimplify.c (gimplify_bind_expr): In ORT_SIMD region
5632 mark local addressable non-static vars as GOVD_PRIVATE
5633 instead of GOVD_LOCAL.
5634 * omp-low.c (lower_omp_for): Move gimple_bind_vars
5635 and BLOCK_VARS of gimple_bind_block to new_stmt rather
5636 than copying them.
5637
8fcbce72
JJ
5638 PR middle-end/60092
5639 * tree-ssa-ccp.c (surely_varying_stmt_p): Don't return true
5640 if TYPE_ATTRIBUTES (gimple_call_fntype ()) contain
5641 assume_aligned or alloc_align attributes.
5642 (bit_value_assume_aligned): Add ATTR, PTRVAL and ALLOC_ALIGN
5643 arguments. Handle also assume_aligned and alloc_align attributes.
1287ae50
UB
5644 (evaluate_stmt): Adjust bit_value_assume_aligned caller. Handle
5645 calls to functions with assume_aligned or alloc_align attributes.
5646 * doc/extend.texi: Document assume_aligned and alloc_align attributes.
8fcbce72 5647
451bdd23
TG
56482014-02-08 Terry Guo <terry.guo@arm.com>
5649
5650 * doc/invoke.texi: Document ARM -march=armv7e-m.
5651
d31d42c7
JJ
56522014-02-08 Jakub Jelinek <jakub@redhat.com>
5653
d71dfeb7
JJ
5654 * cilk-common.c (cilk_init_builtins): Clear TREE_NOTHROW
5655 flag on __cilkrts_rethrow builtin.
5656
d31d42c7
JJ
5657 PR ipa/60026
5658 * ipa-cp.c (determine_versionability): Fail at -O0
5659 or __attribute__((optimize (0))) or -fno-ipa-cp functions.
5660 * tree-sra.c (ipa_sra_preliminary_function_checks): Similarly.
5661
5662 Revert:
5663 2014-02-04 Jakub Jelinek <jakub@redhat.com>
5664
5665 PR ipa/60026
5666 * tree-inline.c (copy_forbidden): Fail for
5667 __attribute__((optimize (0))) functions.
5668
a0a98fef
JH
56692014-02-07 Jan Hubicka <hubicka@ucw.cz>
5670
5671 * varpool.c: Include pointer-set.h.
5672 (varpool_remove_unreferenced_decls): Variables in other partitions
5673 will not be output; be however careful to not lose information
5674 about partitioning.
5675
8c311b50
JH
56762014-02-07 Jan Hubicka <hubicka@ucw.cz>
5677
5678 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do O(1)
5679 lookup in the vtable constructor.
5680
7ba03e5e
JL
56812014-02-07 Jeff Law <law@redhat.com>
5682
63e6247d
JL
5683 PR target/40977
5684 * config/m68k/m68k.md (ashldi_extsi): Turn into a
5685 define_insn_and_split.
5686
7ba03e5e
JL
5687 * ipa-inline.c (inline_small_functions): Fix typos.
5688
177bc204
RS
56892014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5690
5691 * config/s390/s390-protos.h (s390_can_use_simple_return_insn)
5692 (s390_can_use_return_insn): Declare.
5693 * config/s390/s390.h (EPILOGUE_USES): Define.
5694 * config/s390/s390.c (s390_mainpool_start): Allow two main_pool
5695 instructions.
5696 (s390_chunkify_start): Handle return JUMP_LABELs.
5697 (s390_early_mach): Emit a main_pool instruction on the entry edge.
5698 (s300_set_up_by_prologue, s390_can_use_simple_return_insn)
5699 (s390_can_use_return_insn): New functions.
5700 (s390_fix_long_loop_prediction): Handle conditional returns.
5701 (TARGET_SET_UP_BY_PROLOGUE): Define.
5702 * config/s390/s390.md (ANY_RETURN): New code iterator.
5703 (*creturn, *csimple_return, return, simple_return): New patterns.
5704
0621cf3c
RS
57052014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5706
5707 * config/s390/s390.c (s390_restore_gprs_from_fprs): Add REG_CFA_RESTORE
5708 notes to each restore. Also add REG_CFA_DEF_CFA when restoring %r15.
5709 (s390_optimize_prologue): Don't clear RTX_FRAME_RELATED_P. Update the
5710 REG_CFA_RESTORE list when deciding not to restore a register.
5711
4099494d
RS
57122014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5713
5714 * config/s390/s390.c: Include tree-pass.h and context.h.
5715 (s390_early_mach): New function, split out from...
5716 (s390_emit_prologue): ...here.
5717 (pass_data_s390_early_mach): New pass structure.
5718 (pass_s390_early_mach): New class.
5719 (s390_option_override): Create and register early_mach pass.
5720 Move to end of file.
5721
3489cc33
RS
57222014-02-07 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
5723
5724 * var-tracking.c (vt_stack_adjustments): Don't require stack_adjusts
5725 to match for the exit block.
5726
75cc21e2
AK
57272014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5728
5729 * config/s390/s390.md ("atomic_load<mode>", "atomic_store<mode>")
5730 ("atomic_compare_and_swap<mode>", "atomic_fetch_<atomic><mode>"):
5731 Reject misaligned operands.
5732
8bd7070a
AK
57332014-02-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5734
1287ae50 5735 * optabs.c (expand_atomic_compare_and_swap): Allow expander to fail.
8bd7070a 5736
831806cb
RB
57372014-02-07 Richard Biener <rguenther@suse.de>
5738
5739 PR middle-end/60092
5740 * gimple-low.c (lower_builtin_posix_memalign): New function.
5741 (lower_stmt): Call it to lower posix_memalign in a way
5742 to make alignment info accessible.
5743
7ee9c16f
JJ
57442014-02-07 Jakub Jelinek <jakub@redhat.com>
5745
5746 PR c++/60082
5747 * tree.c (build_common_builtin_nodes): Set ECF_LEAF for
5748 __builtin_setjmp_receiver.
5749
32cab212
RB
57502014-02-07 Richard Biener <rguenther@suse.de>
5751
5752 PR middle-end/60092
5753 * builtin-types.def (BT_FN_INT_PTRPTR_SIZE_SIZE): Add.
5754 * builtins.def (BUILT_IN_POSIX_MEMALIGN): Likewise.
5755 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
5756 Handle BUILT_IN_POSIX_MEMALIGN.
5757 (find_func_clobbers): Likewise.
5758 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
5759 (call_may_clobber_ref_p_1): Likewise.
5760
6d6af792
JH
57612014-02-06 Jan Hubicka <hubicka@ucw.cz>
5762
5763 PR ipa/59918
1287ae50
UB
5764 * ipa-devirt.c (record_target_from_binfo): Remove overactive
5765 sanity check.
6d6af792 5766
3c0f1105
JH
57672014-02-06 Jan Hubicka <hubicka@ucw.cz>
5768
5769 PR ipa/59469
5770 * lto-cgraph.c (lto_output_node): Use
5771 symtab_get_symbol_partitioning_class.
5772 (lto_output_varpool_node): likewise.
5773 (symtab_get_symbol_partitioning_class): Move here from
5774 lto/lto-partition.c
5775 * cgraph.h (symbol_partitioning_class): Likewise.
5776 (symtab_get_symbol_partitioning_class): Declare.
5777
b3bb0eb9
JH
57782014-02-06 Jan Hubicka <hubicka@ucw.cz>
5779
5780 * ggc.h (ggc_internal_cleared_alloc): New macro.
5781 * vec.h (vec_safe_copy): Handle memory stats.
5782 * omp-low.c (simd_clone_struct_alloc): Use ggc_internal_cleared_alloc.
5783 * target-globals.c (save_target_globals): Likewise.
5784
57852014-02-06 Jan Hubicka <hubicka@ucw.cz>
e2f0f5f7
JJ
5786
5787 PR target/60077
5788 * expr.c (emit_move_resolve_push): Export; be bit more selective
5789 on when to clear alias set.
5790 * expr.h (emit_move_resolve_push): Declare.
5791 * function.h (struct function): Add tail_call_marked.
5792 * tree-tailcall.c (optimize_tail_call): Set tail_call_marked.
5793 * config/i386/i386-protos.h (ix86_expand_push): Remove.
5794 * config/i386/i386.md (TImode move expander): De not call
5795 ix86_expand_push.
5796 (FP push expanders): Preserve memory attributes.
5797 * config/i386/sse.md (push<mode>1): Remove.
1287ae50 5798 * config/i386/i386.c (ix86_expand_vector_move): Handle push operation.
e2f0f5f7
JJ
5799 (ix86_expand_push): Remove.
5800 * config/i386/mmx.md (push<mode>1): Remove.
5801
47d552eb
JJ
58022014-02-06 Jakub Jelinek <jakub@redhat.com>
5803
5804 PR rtl-optimization/60030
5805 * internal-fn.c (ubsan_expand_si_overflow_mul_check): Surround
5806 lopart with paradoxical subreg before shifting it up by hprec.
5807
45c75ea7
KT
58082014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5809
5810 * config/arm/aarch-cost-tables.h (cortexa57_extra_costs): New table.
5811 Remove extra newline at end of file.
5812 * config/arm/arm.c (arm_cortex_a57_tune): New tuning struct.
5813 (arm_issue_rate): Handle cortexa57.
5814 * config/arm/arm-cores.def (cortex-a57): Use cortex_a57 tuning.
5815 (cortex-a57.cortex-a53): Likewise.
5816
af116cae
JJ
58172014-02-06 Jakub Jelinek <jakub@redhat.com>
5818
652a3e3a
JJ
5819 PR target/59575
5820 * config/arm/arm.c (emit_multi_reg_push): Add dwarf_regs_mask argument,
5821 don't record in REG_FRAME_RELATED_EXPR registers not set in that
5822 bitmask.
5823 (arm_expand_prologue): Adjust all callers.
5824 (arm_unwind_emit_sequence): Allow saved, but not important for unwind
5825 info, registers also at the lowest numbered registers side. Use
5826 gcc_assert instead of abort, and SET_SRC/SET_DEST macros instead of
5827 XEXP.
5828
af116cae 5829 PR debug/59992
1287ae50
UB
5830 * var-tracking.c (adjust_mems): Before adding a SET to
5831 amd->side_effects, adjust it's SET_SRC using simplify_replace_fn_rtx.
af116cae 5832
fdcee33b
AM
58332014-02-06 Alan Modra <amodra@gmail.com>
5834
5835 PR target/60032
5836 * config/rs6000/rs6000.c (rs6000_secondary_memory_needed_mode): Only
5837 change SDmode to DDmode when lra_in_progress.
5838
251901a0
JJ
58392014-02-06 Jakub Jelinek <jakub@redhat.com>
5840
d3ef8c53
JJ
5841 PR middle-end/59150
5842 * tree-vect-data-refs.c (vect_analyze_data_refs): For clobbers, call
5843 free_data_ref on the dr first, and before goto again also set dr
5844 to the next dr. For simd_lane_access, free old datarefs[i] before
5845 overwriting it. For get_vectype_for_scalar_type failure, don't
5846 free_data_ref if simd_lane_access.
5847
2754b38f
JJ
5848 * Makefile.in (prefix.o, cppbuiltin.o): Depend on $(BASEVER).
5849
251901a0
JJ
5850 PR target/60062
5851 * tree.h (opts_for_fn): New inline function.
5852 (opt_for_fn): Define.
5853 * config/i386/i386.c (ix86_function_regparm): Use
5854 opt_for_fn (decl, optimize) instead of optimize.
5855
4a985a37
MS
58562014-02-06 Marcus Shawcroft <marcus.shawcroft@arm.com>
5857
5858 * config/aarch64/aarch64.c (aarch64_classify_symbol): Fix logic
5859 for SYMBOL_REF in large memory model.
5860
c366d38c
KT
58612014-02-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5862
5863 * config/aarch64/aarch64-cores.def (cortex-a53): Specify CRC32
5864 and crypto support.
5865 (cortex-a57): Likewise.
5866 (cortex-a57.cortex-a53): Likewise.
5867
c801e246
YG
58682014-02-06 Yury Gribov <y.gribov@samsung.com>
5869 Kugan Vivekanandarajah <kuganv@linaro.org>
5870
5871 * config/arm/arm.c (arm_vector_alignment_reachable): Check
5872 unaligned_access.
5873 * config/arm/arm.c (arm_builtin_support_vector_misalignment): Likewise.
5874
d50f7b84
RB
58752014-02-06 Richard Biener <rguenther@suse.de>
5876
5877 * tree-cfg.c (gimple_duplicate_sese_region): Fix ordering of
5878 set_loop_copy and initialize_original_copy_tables.
5879
179d2941
AV
58802014-02-06 Alex Velenko <Alex.Velenko@arm.com>
5881
5882 * config/aarch64/aarch64-simd.md
5883 (aarch64_ashr_simddi): Change QI to SI.
5884
78b1469d
JH
58852014-02-05 Jan Hubicka <hubicka@ucw.cz>
5886 Jakub Jelinek <jakub@redhat.com>
5887
5888 PR middle-end/60013
5889 * ipa-inline-analysis.c (compute_bb_predicates): Ensure monotonicity
5890 of the dataflow.
5891
d85f364c
BS
58922014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5893
5894 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Change
5895 CODE_FOR_altivec_vpku[hw]um to
5896 CODE_FOR_altivec_vpku[hw]um_direct.
5897 * config/rs6000/altivec.md (vec_unpacks_hi_<VP_small_lc>): Change
5898 UNSPEC_VUNPACK_HI_SIGN to UNSPEC_VUNPACK_HI_SIGN_DIRECT.
5899 (vec_unpacks_lo_<VP_small_lc>): Change UNSPEC_VUNPACK_LO_SIGN to
5900 UNSPEC_VUNPACK_LO_SIGN_DIRECT.
5901
7b1cd427
BS
59022014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5903
5904 * config/rs6000/altivec.md (altivec_vsum2sws): Adjust code
5905 generation for -maltivec=be.
5906 (altivec_vsumsws): Simplify redundant test.
5907
52a93551
BS
59082014-02-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5909
5910 * altivec.md (UNSPEC_VPACK_UNS_UNS_MOD_DIRECT): New unspec.
5911 (UNSPEC_VUNPACK_HI_SIGN_DIRECT): Likewise.
5912 (UNSPEC_VUNPACK_LO_SIGN_DIRECT): Likewise.
5913 (mulv8hi3): Use gen_altivec_vpkuwum_direct instead of
5914 gen_altivec_vpkuwum.
5915 (altivec_vpkpx): Test for VECTOR_ELT_ORDER_BIG instead of for
5916 BYTES_BIG_ENDIAN.
5917 (altivec_vpks<VI_char>ss): Likewise.
5918 (altivec_vpks<VI_char>us): Likewise.
5919 (altivec_vpku<VI_char>us): Likewise.
5920 (altivec_vpku<VI_char>um): Likewise.
5921 (altivec_vpku<VI_char>um_direct): New (copy of
5922 altivec_vpku<VI_char>um that still relies on BYTES_BIG_ENDIAN, for
5923 internal use).
5924 (altivec_vupkhs<VU_char>): Emit vupkls* instead of vupkhs* when
5925 target is little endian and -maltivec=be is not specified.
5926 (*altivec_vupkhs<VU_char>_direct): New (copy of
1287ae50 5927 altivec_vupkhs<VU_char> that always emits vupkhs*, for internal use).
52a93551
BS
5928 (altivec_vupkls<VU_char>): Emit vupkhs* instead of vupkls* when
5929 target is little endian and -maltivec=be is not specified.
5930 (*altivec_vupkls<VU_char>_direct): New (copy of
1287ae50 5931 altivec_vupkls<VU_char> that always emits vupkls*, for internal use).
52a93551
BS
5932 (altivec_vupkhpx): Emit vupklpx instead of vupkhpx when target is
5933 little endian and -maltivec=be is not specified.
5934 (altivec_vupklpx): Emit vupkhpx instead of vupklpx when target is
5935 little endian and -maltivec=be is not specified.
5936
004a7e45 59372014-02-05 Richard Henderson <rth@redhat.com>
76df831f
RH
5938
5939 PR debug/52727
5940 * combine-stack-adj.c: Revert r206943.
5941 * sched-int.h (struct deps_desc): Add last_args_size.
5942 * sched-deps.c (init_deps): Initialize it.
5943 (sched_analyze_insn): Add OUTPUT dependencies between insns that
5944 contain REG_ARGS_SIZE notes.
5945
f300e7b8
JH
59462014-02-05 Jan Hubicka <hubicka@ucw.cz>
5947
5948 * lto-cgraph.c (asm_nodes_output): Make global.
5949 * lto-wrapper.c (run_gcc): Pass down paralelizm to WPA.
1287ae50 5950 * gcc.c (AS_NEEDS_DASH_FOR_PIPED_INPUT): Allow WPA parameter
f300e7b8
JH
5951 (driver_handle_option): Handle OPT_fwpa.
5952
609524d2
JJ
59532014-02-05 Jakub Jelinek <jakub@redhat.com>
5954
add5c763
JJ
5955 PR ipa/59947
5956 * ipa-devirt.c (possible_polymorphic_call_targets): Fix
5957 a comment typo and formatting issue. If odr_hash hasn't been
5958 created, return vNULL and set *completep to false.
5959
609524d2
JJ
5960 PR middle-end/57499
5961 * tree-eh.c (cleanup_empty_eh): Bail out on totally empty
5962 bb with no successors.
5963
05ab6e21
JG
59642014-02-05 James Greenhalgh <james.greenhalgh@arm.com>
5965
5966 PR target/59718
5967 * doc/invoke.texi (-march): Clarify documentation for ARM.
5968 (-mtune): Likewise.
5969 (-mcpu): Likewise.
5970
d55d9ed0
RB
59712014-02-05 Richard Biener <rguenther@suse.de>
5972
5973 * tree-vect-loop.c (vect_analyze_loop_2): Be more informative
5974 when not vectorizing because of too many alias checks.
5975 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
5976 Add more verboseness, avoid duplicate MSG_MISSED_OPTIMIZATION.
5977
c70da878
NC
59782014-02-05 Nick Clifton <nickc@redhat.com>
5979
5980 * config/mn10300/mn10300.c (mn10300_hard_regno_mode_ok): Do not
1287ae50 5981 accept extended registers in any mode when compiling for the MN10300.
c70da878 5982
25a07c7e
YG
59832014-02-05 Yury Gribov <y.gribov@samsung.com>
5984
5985 * cif-code.def (ATTRIBUTE_MISMATCH): New CIF code.
5986 * ipa-inline.c (report_inline_failed_reason): Handle mismatched
5987 sanitization attributes.
5988 (can_inline_edge_p): Likewise.
5989 (sanitize_attrs_match_for_inline_p): New function.
5990
4bf2a588
JH
59912014-02-04 Jan Hubicka <hubicka@ucw.cz>
5992
5993 * ipa-prop.c (detect_type_change): Shor circuit testing of
5994 type changes on THIS pointer.
5995
d92f4df0
JDA
59962014-02-04 John David Anglin <danglin@gcc.gnu.org>
5997
5998 PR target/59777
5999 * config/pa/pa.c (legitimize_tls_address): Return original address
6000 if not passed a SYMBOL_REF rtx.
6001 (hppa_legitimize_address): Call legitimize_tls_address for all TLS
6002 addresses.
6003 (pa_emit_move_sequence): Simplify TLS source operands.
6004 (pa_legitimate_constant_p): Reject all TLS constants.
6005 * config/pa/pa.h (PA_SYMBOL_REF_TLS_P): Correct comment.
6006 (CONSTANT_ADDRESS_P): Reject TLS CONST addresses.
6007
d665f8dd
JH
60082014-02-04 Jan Hubicka <hubicka@ucw.cz>
6009
6010 * ipa.c (function_and_variable_visibility): Decompose DECL_ONE_ONLY
6011 groups when we know they are controlled by LTO.
6012 * varasm.c (default_binds_local_p_1): If object is in other partition,
6013 it will be resolved locally.
6014
6a071860
BE
60152014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6016
e2fc3b4f 6017 * config/host-linux.c (linux_gt_pch_use_address): Don't
1287ae50 6018 use SSIZE_MAX because it is not always defined.
6a071860 6019
8e9d68a9
VM
60202014-02-04 Vladimir Makarov <vmakarov@redhat.com>
6021
6022 PR bootstrap/59913
6023 * lra-constraints.c (need_for_split_p): Use more 3 reloads as
6024 threshold for pseudo splitting.
6025 (update_ebb_live_info): Process call argument hard registers and
6026 hard registers from insn definition too.
6027 (max_small_class_regs_num): New constant.
6028 (inherit_in_ebb): Update live hard regs through EBBs. Update
6029 reloads_num only for small register classes. Don't split for
6030 outputs of jumps.
6031
8472fa80
MT
60322014-02-04 Markus Trippelsdorf <markus@trippelsdorf.de>
6033
6034 PR ipa/60058
6035 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that target
6036 is non-null.
6037
df98e37f
JH
60382014-02-04 Jan Hubicka <hubicka@ucw.cz>
6039
1287ae50
UB
6040 * gimple-fold.c (can_refer_decl_in_current_unit_p): Default
6041 visibility is safe.
df98e37f 6042
71e55f04
MP
60432014-02-04 Marek Polacek <polacek@redhat.com>
6044
6045 * gdbinit.in (pel): Define.
6046
229e56f9
BE
60472014-02-04 Bernd Edlinger <bernd.edlinger@hotmail.de>
6048
6049 * doc/invoke.texi (fstrict-volatile-bitfields): Clarify current
6050 behavior.
6051
c39276b8
RB
60522014-02-04 Richard Biener <rguenther@suse.de>
6053
6054 PR lto/59723
6055 * lto-streamer-out.c (tree_is_indexable): Force NAMELIST_DECLs
6056 in function context local.
6057 (lto_output_tree_ref): Do not write trees from lto_output_tree_ref.
6058 * lto-streamer-in.c (lto_input_tree_ref): Handle LTO_namelist_decl_ref
6059 similar to LTO_imported_decl_ref.
6060
66b3ed5f
JJ
60612014-02-04 Jakub Jelinek <jakub@redhat.com>
6062
029ce7a2
JJ
6063 PR tree-optimization/60002
6064 * cgraphclones.c (build_function_decl_skip_args): Clear
6065 DECL_LANG_SPECIFIC.
6066
8e91d222
JJ
6067 PR tree-optimization/60023
6068 * tree-if-conv.c (predicate_mem_writes): Pass true instead of
6069 false to gsi_replace.
6070 * tree-vect-stmts.c (vect_finish_stmt_generation): If stmt
6071 has been in some EH region and vec_stmt could throw, add
6072 vec_stmt into the same EH region.
6073 * tree-data-ref.c (get_references_in_stmt): If IFN_MASK_LOAD
6074 has no lhs, ignore it.
6075 * internal-fn.c (expand_MASK_LOAD): Likewise.
6076
eb259c4a
JJ
6077 PR ipa/60026
6078 * tree-inline.c (copy_forbidden): Fail for
6079 __attribute__((optimize (0))) functions.
6080
92d05580
JJ
6081 PR other/58712
6082 * omp-low.c (simd_clone_struct_copy): If from->inbranch
6083 is set, copy one less argument.
30540e79
JJ
6084 (expand_simd_clones): Don't subtract clone_info->inbranch
6085 from simd_clone_struct_alloc argument.
92d05580 6086
be3afd67
JJ
6087 PR rtl-optimization/57915
6088 * recog.c (simplify_while_replacing): If all unary/binary/relational
6089 operation arguments are constant, attempt to simplify those.
6090
66b3ed5f
JJ
6091 PR middle-end/59261
6092 * expmed.c (expand_mult): For MODE_VECTOR_INT multiplication
6093 if there is no vashl<mode>3 or ashl<mode>3 insn, skip_synth.
6094
5961d779
RB
60952014-02-04 Richard Biener <rguenther@suse.de>
6096
6097 PR tree-optimization/60012
6098 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Apply
6099 TBAA disambiguation to all DDRs.
6100
91da0481
RO
61012014-02-04 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6102
6103 PR target/59788
6104 * config/sol2.h (LINK_LIBGCC_MAPFILE_SPEC): Define.
6105 (LINK_SPEC): Use it for -shared, -shared-libgcc.
6106
39960d1c
JH
61072014-02-03 Jan Hubicka <hubicka@ucw.cz>
6108
6109 PR ipa/59882
6110 * tree.c (get_binfo_at_offset): Do not get confused by empty classes;
6111
61122014-02-03 Jan Hubicka <hubicka@ucw.cz>
6f746413
JH
6113
6114 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Remove.
6115 * gimple-fold.h (gimple_extract_devirt_binfo_from_cst): Remove.
6116
39960d1c 61172014-02-03 Jan Hubicka <hubicka@ucw.cz>
5bccb77a
JH
6118
6119 PR ipa/59831
6120 * ipa-cp.c (ipa_get_indirect_edge_target_1): Use ipa-devirt
6121 to figure out targets of polymorphic calls with known decl.
6122 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
6123 * ipa-utils.h (get_polymorphic_call_info_from_invariant): Declare.
6124 * ipa-devirt.c (get_polymorphic_call_info_for_decl): Break out from ...
6125 (get_polymorphic_call_info): ... here.
6126 (get_polymorphic_call_info_from_invariant): New function.
6127
39960d1c 61282014-02-03 Jan Hubicka <hubicka@ucw.cz>
9de2f554
JH
6129
6130 * ipa-cp.c (ipa_get_indirect_edge_target_1): Do direct
6131 lookup via vtable pointer; check for type consistency
6132 and turn inconsitent facts into UNREACHABLE.
6133 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
1287ae50
UB
6134 * gimple-fold.c (gimple_get_virt_method_for_vtable): Do not ICE on
6135 type inconsistent querries; return UNREACHABLE instead.
9de2f554 6136
ade3ff24
RH
61372014-02-03 Richard Henderson <rth@twiddle.net>
6138
6139 PR tree-opt/59924
6140 * tree-ssa-uninit.c (push_to_worklist): Don't re-push if we've
6141 already processed this node.
6142 (normalize_one_pred_1): Pass along mark_set.
6143 (normalize_one_pred): Create and destroy a pointer_set_t.
6144 (normalize_one_pred_chain): Likewise.
6145
55428cc3
LA
61462014-02-03 Laurent Aflonsi <laurent.alfonsi@st.com>
6147
6148 PR gcov-profile/58602
1287ae50 6149 * gcc/gcov-io.c (gcov_open): Open with truncation when mode < 0.
55428cc3 6150
85942f45
JH
61512014-02-03 Jan Hubicka <hubicka@ucw.cz>
6152
6153 PR ipa/59831
1287ae50
UB
6154 * ipa-cp.c (ipa_get_indirect_edge_target_1): Give up on
6155 -fno-devirtualize; try to devirtualize by the knowledge of
6156 virtual table pointer given by aggregate propagation.
85942f45 6157 * ipa-prop.c (try_make_edge_direct_virtual_call): Likewise.
1287ae50 6158 (ipa_print_node_jump_functions): Dump also offset that
85942f45
JH
6159 is relevant for polymorphic calls.
6160 (determine_known_aggregate_parts): Add arg_type parameter; use it
6161 instead of determining the type from pointer type.
6162 (ipa_compute_jump_functions_for_edge): Update call of
6163 determine_known_aggregate_parts.
6164 * gimple-fold.c (gimple_get_virt_method_for_vtable): Break out from ...
6165 (gimple_get_virt_method_for_binfo): ... here; simplify using
6166 vtable_pointer_value_to_vtable.
6167 * gimple-fold.h (gimple_get_virt_method_for_vtable): Declare.
6168 * ipa-devirt.c (subbinfo_with_vtable_at_offset): Turn OFFSET parameter
1287ae50 6169 to unsigned HOST_WIDE_INT; use vtable_pointer_value_to_vtable.
85942f45
JH
6170 (vtable_pointer_value_to_vtable): Break out from ...; handle also
6171 POINTER_PLUS_EXPR.
6172 (vtable_pointer_value_to_binfo): ... here.
6173 * ipa-utils.h (vtable_pointer_value_to_vtable): Declare.
6174
bddc974e
TJ
61752014-02-03 Teresa Johnson <tejohnson@google.com>
6176
6177 * tree-vect-slp.c (vect_supported_load_permutation_p): Avoid
6178 redef of outer loop index variable.
6179
5d77fb19
MG
61802014-02-03 Marc Glisse <marc.glisse@inria.fr>
6181
6182 PR c++/53017
6183 PR c++/59211
6184 * doc/extend.texi (Function Attributes): Typo.
6185
cf5b2be2
CH
61862014-02-03 Cong Hou <congh@google.com>
6187
6188 PR tree-optimization/60000
6189 * tree-vect-loop.c (vect_transform_loop): Set pattern_def_seq to NULL
6190 if the vectorized statement is a store. A store statement can only
6191 appear at the end of pattern statements.
6192
a2a1ddb5
L
61932014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6194
6195 * config/i386/i386.c (flag_opts): Add -mlong-double-128.
6196 (ix86_option_override_internal): Default long double to 64-bit for
6197 32-bit Bionic and to 128-bit for 64-bit Bionic.
6198
6199 * config/i386/i386.h (LONG_DOUBLE_TYPE_SIZE): Use 128 if
6200 TARGET_LONG_DOUBLE_128 is true.
6201 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Likewise.
6202
6203 * config/i386/i386.opt (mlong-double-80): Negate -mlong-double-64.
6204 (mlong-double-64): Negate -mlong-double-128.
6205 (mlong-double-128): New option.
6206
6207 * config/i386/i386-c.c (ix86_target_macros): Define
6208 __LONG_DOUBLE_128__ for TARGET_LONG_DOUBLE_128.
6209
6210 * doc/invoke.texi: Document -mlong-double-128.
6211
f742cf90
L
62122014-02-03 H.J. Lu <hongjiu.lu@intel.com>
6213
6214 PR rtl-optimization/60024
6215 * sel-sched.c (init_regs_for_mode): Check if mode is OK first.
6216
8f36fd30
MT
62172014-02-03 Markus Trippelsdorf <markus@trippelsdorf.de>
6218
1287ae50 6219 * doc/invoke.texi (fprofile-reorder-functions): Fix typo.
8f36fd30 6220
861ec4f3
AB
62212014-02-03 Andrey Belevantsev <abel@ispras.ru>
6222
6223 PR rtl-optimization/57662
6224 * sel-sched.c (code_motion_path_driver): Do not mark already not
6225 existing blocks in the visiting bitmap.
6226
fe08255d
AB
62272014-02-03 Andrey Belevantsev <abel@ispras.ru>
6228
6229 * sel-sched-ir.c (sel_gen_insn_from_expr_after): Reset INSN_DELETED_P
6230 on the insn being emitted.
6231
96d3a240
JG
62322014-02-03 James Greenhalgh <james.greenhalgh@arm.com>
6233 Will Deacon <will.deacon@arm.com>
6234
6235 * doc/gimple.texi (gimple_asm_clear_volatile): Remove.
6236
9dd6c9f7
KT
62372014-02-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6238
6239 * config/arm/arm-tables.opt: Regenerate.
6240
60331d00
BS
62412014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6242
6243 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): Generalize
6244 for vector types other than V16QImode.
6245 * config/rs6000/altivec.md (altivec_vperm_<mode>): Change to a
6246 define_expand, and call altivec_expand_vec_perm_le when producing
6247 code with little endian element order.
6248 (*altivec_vperm_<mode>_internal): New insn having previous
6249 behavior of altivec_vperm_<mode>.
6250 (altivec_vperm_<mode>_uns): Change to a define_expand, and call
6251 altivec_expand_vec_perm_le when producing code with little endian
6252 element order.
6253 (*altivec_vperm_<mode>_uns_internal): New insn having previous
6254 behavior of altivec_vperm_<mode>_uns.
6255
b80afde9
BS
62562014-02-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6257
6258 * config/rs6000/altivec.md (UNSPEC_VSUMSWS_DIRECT): New unspec.
6259 (altivec_vsumsws): Add handling for -maltivec=be with a little
6260 endian target.
6261 (altivec_vsumsws_direct): New.
6262 (reduc_splus_<mode>): Call gen_altivec_vsumsws_direct instead of
6263 gen_altivec_vsumsws.
6264
39960d1c 62652014-02-02 Jan Hubicka <hubicka@ucw.cz>
390675c8
JH
6266
6267 * ipa-devirt.c (subbinfo_with_vtable_at_offset,
6268 vtable_pointer_value_to_binfo): New functions.
6269 * ipa-utils.h (vtable_pointer_value_to_binfo): Declare.
6270 * ipa-prop.c (extr_type_from_vtbl_ptr_store): Use it.
6271
02b67731
SL
62722014-02-02 Sandra Loosemore <sandra@codesourcery.com>
6273
6274 * config/nios2/nios2.md (load_got_register): Initialize GOT
6275 pointer from _gp_got instead of _GLOBAL_OFFSET_TABLE_.
6276 * config/nios2/nios2.c (nios2_function_profiler): Likewise.
6277
2ace77c2
JH
62782014-02-02 Jan Hubicka <hubicka@ucw.cz>
6279
6280 * ipa-prop.c (update_jump_functions_after_inlining): When type is not
6281 preserverd by passthrough, do not propagate the type.
6282
70b2d364
RS
62832014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6284
6285 * config/mips/mips.c (MIPS_GET_FCSR, MIPS_SET_FCSR): New macros.
6286 (mips_atomic_assign_expand_fenv): New function.
6287 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): Define.
6288
a90c0245
RS
62892014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6290
6291 * doc/extend.texi (__builtin_mips_get_fcsr): Document.
6292 (__builtin_mips_set_fcsr): Likewise.
6293 * config/mips/mips-ftypes.def: Add MIPS_VOID_FTYPE_USI and
6294 MIPS_USI_FTYPE_VOID.
6295 * config/mips/mips-protos.h (mips16_expand_get_fcsr): Declare
6296 (mips16_expand_set_fcsr): Likewise.
6297 * config/mips/mips.c (mips16_get_fcsr_stub): New variable.
6298 (mips16_set_fcsr_stub): Likewise.
6299 (mips16_get_fcsr_one_only_stub): New class.
6300 (mips16_set_fcsr_one_only_stub): Likewise.
6301 (mips16_expand_get_fcsr, mips16_expand_set_fcsr): New functions.
6302 (mips_code_end): Output the get_fcsr and set_fcsr stubs, if needed.
6303 (BUILTIN_AVAIL_MIPS16, AVAIL_ALL): New macros.
6304 (hard_float): New availability predicate.
6305 (mips_builtins): Add get_fcsr and set_fcsr.
6306 (mips_expand_builtin): Check BUILTIN_AVAIL_MIPS16.
6307 * config/mips/mips.md (UNSPEC_GET_FCSR, UNSPEC_SET_FCSR): New unspecs.
6308 (GET_FCSR_REGNUM, SET_FCSR_REGNUM): New constants.
6309 (mips_get_fcsr, *mips_get_fcsr, mips_get_fcsr_mips16_<mode>)
6310 (mips_set_fcsr, *mips_set_fcsr, mips_set_fcsr_mips16_<mode>): New
6311 patterns.
6312
6d51cc90
RS
63132014-02-02 Richard Sandiford <rdsandiford@googlemail.com>
6314
6315 * config/mips/mips.c (mips_one_only_stub): New class.
6316 (mips_need_mips16_rdhwr_p): Replace with...
6317 (mips16_rdhwr_stub): ...this new variable.
6318 (mips16_stub_call_address): New function.
6319 (mips16_rdhwr_one_only_stub): New class.
6320 (mips_expand_thread_pointer): Use mips16_stub_call_address.
6321 (mips_output_mips16_rdhwr): Delete.
6322 (mips_finish_stub): New function.
6323 (mips_code_end): Use it to handle rdhwr stubs.
6324
6c90f137
UB
63252014-02-02 Uros Bizjak <ubizjak@gmail.com>
6326
6327 PR target/60017
6328 * config/i386/i386.c (classify_argument): Fix handling of bit_offset
6329 when calculating size of integer atomic types.
6330
5e64bbbb
L
63312014-02-02 H.J. Lu <hongjiu.lu@intel.com>
6332
6333 * ipa-inline-analysis.c (true_predicate_p): Fix a typo in comments.
6334
021293cb
JJ
63352014-02-01 Jakub Jelinek <jakub@redhat.com>
6336
6337 PR tree-optimization/60003
6338 * gimple-low.c (lower_builtin_setjmp): Set cfun->has_nonlocal_label.
6339 * profile.c (branch_prob): Use gimple_call_builtin_p
6340 to check for BUILT_IN_SETJMP_RECEIVER.
6341 * tree-inline.c (copy_bb): Call notice_special_calls.
6342
6334f3e9
VM
63432014-01-31 Vladimir Makarov <vmakarov@redhat.com>
6344
6345 PR bootstrap/59985
6346 * lra-constraints.c (process_alt_operands): Update reload_sum only
6347 on the first pass.
6348
efa7882f
RH
63492014-01-31 Richard Henderson <rth@redhat.com>
6350
6351 PR middle-end/60004
6352 * tree-eh.c (lower_try_finally_switch): Delay lowering finally block
6353 until after else_eh is processed.
6354
de72ea02
IT
63552014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6356
6357 * config/i386/avx512fintrin.h (_MM_FROUND_TO_NEAREST_INT),
6358 (_MM_FROUND_TO_NEG_INF), (_MM_FROUND_TO_POS_INF),
6359 (_MM_FROUND_TO_ZERO), (_MM_FROUND_CUR_DIRECTION): Are already defined
6360 in smmintrin.h, remove them.
6361 (_MM_FROUND_NO_EXC): Same as above, bit also wrong value.
6362 * config/i386/i386.c (ix86_print_operand): Split sae and rounding.
6363 * config/i386/i386.md (ROUND_SAE): Fix value.
6364 * config/i386/predicates.md (const_4_or_8_to_11_operand): New.
6365 (const48_operand): New.
6366 * config/i386/subst.md (round), (round_expand): Use
6367 const_4_or_8_to_11_operand.
6368 (round_saeonly), (round_saeonly_expand): Use const48_operand.
6369
be792bce
IT
63702014-01-31 Ilya Tocar <ilya.tocar@intel.com>
6371
6372 * config/i386/constraints.md (Yk): Swap meaning with k.
6373 * config/i386/i386.md (movhi_internal): Change Yk to k.
6374 (movqi_internal): Ditto.
6375 (*k<logic><mode>): Ditto.
6376 (*andhi_1): Ditto.
6377 (*andqi_1): Ditto.
6378 (kandn<mode>): Ditto.
6379 (*<code>hi_1): Ditto.
6380 (*<code>qi_1): Ditto.
6381 (kxnor<mode>): Ditto.
6382 (kortestzhi): Ditto.
6383 (kortestchi): Ditto.
6384 (kunpckhi): Ditto.
6385 (*one_cmplhi2_1): Ditto.
6386 (*one_cmplqi2_1): Ditto.
6387 * config/i386/sse.md (): Change k to Yk.
6388 (avx512f_load<mode>_mask): Ditto.
6389 (avx512f_blendm<mode>): Ditto.
6390 (avx512f_store<mode>_mask): Ditto.
6391 (avx512f_storeu<ssemodesuffix>512_mask): Ditto.
6392 (avx512f_storedqu<mode>_mask): Ditto.
33425d6c
UB
6393 (avx512f_cmp<mode>3<mask_scalar_merge_name><round_saeonly_name>):
6394 Ditto.
be792bce
IT
6395 (avx512f_ucmp<mode>3<mask_scalar_merge_name>): Ditto.
6396 (avx512f_vmcmp<mode>3<round_saeonly_name>): Ditto.
6397 (avx512f_vmcmp<mode>3_mask<round_saeonly_name>): Ditto.
6398 (avx512f_maskcmp<mode>3): Ditto.
6399 (avx512f_fmadd_<mode>_mask<round_name>): Ditto.
6400 (avx512f_fmadd_<mode>_mask3<round_name>): Ditto.
6401 (avx512f_fmsub_<mode>_mask<round_name>): Ditto.
6402 (avx512f_fmsub_<mode>_mask3<round_name>): Ditto.
6403 (avx512f_fnmadd_<mode>_mask<round_name>): Ditto.
6404 (avx512f_fnmadd_<mode>_mask3<round_name>): Ditto.
6405 (avx512f_fnmsub_<mode>_mask<round_name>): Ditto.
6406 (avx512f_fnmsub_<mode>_mask3<round_name>): Ditto.
6407 (avx512f_fmaddsub_<mode>_mask<round_name>): Ditto.
6408 (avx512f_fmaddsub_<mode>_mask3<round_name>): Ditto.
6409 (avx512f_fmsubadd_<mode>_mask<round_name>): Ditto.
6410 (avx512f_fmsubadd_<mode>_mask3<round_name>): Ditto.
6411 (avx512f_vextract<shuffletype>32x4_1_maskm): Ditto.
6412 (vec_extract_lo_<mode>_maskm): Ditto.
6413 (vec_extract_hi_<mode>_maskm): Ditto.
6414 (avx512f_vternlog<mode>_mask): Ditto.
6415 (avx512f_fixupimm<mode>_mask<round_saeonly_name>): Ditto.
6416 (avx512f_sfixupimm<mode>_mask<round_saeonly_name>): Ditto.
6417 (avx512f_<code><pmov_src_lower><mode>2_mask): Ditto.
6418 (avx512f_<code>v8div16qi2_mask): Ditto.
6419 (avx512f_<code>v8div16qi2_mask_store): Ditto.
6420 (avx512f_eq<mode>3<mask_scalar_merge_name>_1): Ditto.
6421 (avx512f_gt<mode>3<mask_scalar_merge_name>): Ditto.
6422 (avx512f_testm<mode>3<mask_scalar_merge_name>): Ditto.
6423 (avx512f_testnm<mode>3<mask_scalar_merge_name>): Ditto.
6424 (*avx512pf_gatherpf<mode>sf_mask): Ditto.
6425 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6426 (*avx512pf_scatterpf<mode>sf_mask): Ditto.
6427 (*avx512pf_scatterpf<mode>df_mask): Ditto.
6428 (avx512cd_maskb_vec_dupv8di): Ditto.
6429 (avx512cd_maskw_vec_dupv16si): Ditto.
6430 (avx512f_vpermi2var<mode>3_maskz): Ditto.
6431 (avx512f_vpermi2var<mode>3_mask): Ditto.
6432 (avx512f_vpermi2var<mode>3_mask): Ditto.
6433 (avx512f_vpermt2var<mode>3_maskz): Ditto.
6434 (*avx512f_gathersi<mode>): Ditto.
6435 (*avx512f_gathersi<mode>_2): Ditto.
6436 (*avx512f_gatherdi<mode>): Ditto.
6437 (*avx512f_gatherdi<mode>_2): Ditto.
6438 (*avx512f_scattersi<mode>): Ditto.
6439 (*avx512f_scatterdi<mode>): Ditto.
6440 (avx512f_compress<mode>_mask): Ditto.
6441 (avx512f_compressstore<mode>_mask): Ditto.
6442 (avx512f_expand<mode>_mask): Ditto.
6443 * config/i386/subst.md (mask): Change k to Yk.
6444 (mask_scalar_merge): Ditto.
6445 (sd): Ditto.
6446
0878d68a
MG
64472014-01-31 Marc Glisse <marc.glisse@inria.fr>
6448
6449 * doc/extend.texi (Vector Extensions): Document ?: in C++.
6450
fdfd537b
RB
64512014-01-31 Richard Biener <rguenther@suse.de>
6452
6453 PR middle-end/59990
6454 * builtins.c (fold_builtin_memory_op): Make sure to not
6455 use a floating-point mode or a boolean or enumeral type for
6456 the copy operation.
6457
4f50b9ff
DD
64582014-01-30 DJ Delorie <dj@redhat.com>
6459
6460 * config/msp430/msp430.h (LIB_SPEC): Add -lcrt
6461 * config/msp430/msp430.md (msp430_refsym_need_exit): New.
6462 * config/msp430/msp430.c (msp430_expand_epilogue): Call it
6463 whenever main() has an epilogue.
6464
c3e96073
BS
64652014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6466
6467 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Remove
6468 unused variable "field".
6469 * config/rs6000/vsx.md (vsx_mergel_<mode>): Add missing DONE.
6470 (vsx_mergeh_<mode>): Likewise.
6471 * config/rs6000/altivec.md (altivec_vmrghb): Likewise.
6472 (altivec_vmrghh): Likewise.
6473 (altivec_vmrghw): Likewise.
6474 (altivec_vmrglb): Likewise.
6475 (altivec_vmrglh): Likewise.
6476 (altivec_vmrglw): Likewise.
6477 (altivec_vspltb): Add missing uses.
6478 (altivec_vsplth): Likewise.
6479 (altivec_vspltw): Likewise.
6480 (altivec_vspltsf): Likewise.
6481
4bb9c32d
JJ
64822014-01-30 Jakub Jelinek <jakub@redhat.com>
6483
6484 PR target/59923
6485 * ifcvt.c (cond_exec_process_insns): Don't conditionalize
6486 frame related instructions.
6487
7613fa50
VM
64882014-01-30 Vladimir Makarov <vmakarov@redhat.com>
6489
6490 PR rtl-optimization/59959
6491 * lra-constrains.c (simplify_operand_subreg): Assign NO_REGS to
6492 any reload of register whose subreg is invalid.
6493
6fb82517
JJ
64942014-01-30 Jakub Jelinek <jakub@redhat.com>
6495
33425d6c 6496 * config/i386/f16cintrin.h (_cvtsh_ss): Avoid -Wnarrowing warning.
6fb82517
JJ
6497 * config/i386/avx512fintrin.h (_mm512_mask_cvtusepi64_storeu_epi32):
6498 Add missing return type - void.
6499
bf53d4b8
BS
65002014-01-30 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6501
6502 * gcc/config/rs6000/rs6000.c (rs6000_expand_vector_init): Use
6503 gen_vsx_xxspltw_v4sf_direct instead of gen_vsx_xxspltw_v4sf;
6504 remove element index adjustment for endian (now handled in vsx.md
6505 and altivec.md).
6506 (altivec_expand_vec_perm_const): Use
6507 gen_altivec_vsplt[bhw]_direct instead of gen_altivec_vsplt[bhw].
6508 * gcc/config/rs6000/vsx.md (UNSPEC_VSX_XXSPLTW): New unspec.
6509 (vsx_xxspltw_<mode>): Adjust element index for little endian.
6510 * gcc/config/rs6000/altivec.md (altivec_vspltb): Divide into a
6511 define_expand and a new define_insn *altivec_vspltb_internal;
6512 adjust for -maltivec=be on a little endian target.
6513 (altivec_vspltb_direct): New.
6514 (altivec_vsplth): Divide into a define_expand and a new
6515 define_insn *altivec_vsplth_internal; adjust for -maltivec=be on a
6516 little endian target.
6517 (altivec_vsplth_direct): New.
6518 (altivec_vspltw): Divide into a define_expand and a new
6519 define_insn *altivec_vspltw_internal; adjust for -maltivec=be on a
6520 little endian target.
6521 (altivec_vspltw_direct): New.
6522 (altivec_vspltsf): Divide into a define_expand and a new
6523 define_insn *altivec_vspltsf_internal; adjust for -maltivec=be on
6524 a little endian target.
6525
c965e187
RB
65262014-01-30 Richard Biener <rguenther@suse.de>
6527
6528 PR tree-optimization/59993
6529 * tree-ssa-forwprop.c (associate_pointerplus): Check we
6530 can propagate form the earlier stmt and avoid the transform
6531 when the intermediate result is needed.
6532
ba117645
AD
65332014-01-30 Alangi Derick <alangiderick@gmail.com>
6534
6535 * README.Portability: Fix typo.
6536
4bb66ef3 65372014-01-30 David Holsgrove <david.holsgrove@xilinx.com>
29bd5728
DH
6538
6539 * config/microblaze/microblaze.md(cstoresf4, cbranchsf4): Replace
6540 comparison_operator with ordered_comparison_operator.
6541
c345a0b1
NC
65422014-01-30 Nick Clifton <nickc@redhat.com>
6543
6544 * config/mn10300/mn10300-protos.h (mn10300_store_multiple_operation_p):
6545 Rename to mn10300_store_multiple_regs.
6546 * config/mn10300/mn10300.c: Likewise.
6547 * config/mn10300/mn10300.md (store_movm): Fix typo: call
6548 store_multiple_regs.
6549 * config/mn10300/predicates.md (mn10300_store_multiple_operation):
6550 Call mn10300_store_multiple_regs.
6551
2d70f6d4
NC
65522014-01-30 Nick Clifton <nickc@redhat.com>
6553 DJ Delorie <dj@redhat.com>
6554
6555 * config/rl78/rl78.c (register_sizes): Make the "upper half" of
6556 %fp 2 to keep registers after it properly word-aligned.
6557 (rl78_alloc_physical_registers_umul): Handle the case where both
6558 input operands are the same.
6559
c972624e
RB
65602014-01-30 Richard Biener <rguenther@suse.de>
6561
6562 PR tree-optimization/59903
6563 * tree-vect-loop.c (vect_transform_loop): Guard multiple-types
6564 check properly.
6565
fc044323
JM
65662014-01-30 Jason Merrill <jason@redhat.com>
6567
404c2aea
JM
6568 PR c++/59633
6569 * tree.c (walk_type_fields): Handle VECTOR_TYPE.
6570
fc044323
JM
6571 PR c++/59645
6572 * cgraphunit.c (expand_thunk): Copy volatile arg to a temporary.
6573
f30a0ba5
RB
65742014-01-30 Richard Biener <rguenther@suse.de>
6575
6576 PR tree-optimization/59951
33425d6c 6577 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Skip uses in debug insns.
f30a0ba5 6578
aad8816f
SZ
65792014-01-30 Savin Zlobec <savin.zlobec@gmail.com>
6580
6581 PR target/59784
6582 * config/nios2/nios2.c (nios2_fpu_insn_asm): Fix asm output of
6583 SFmode to DFmode case.
6584
3b16363e
DD
65852014-01-29 DJ Delorie <dj@redhat.com>
6586
6587 * config/msp430/msp430.opt (-minrt): New.
6588 * config/msp430/msp430.h (STARTFILE_SPEC): Link alternate runtime
6589 if -minrt given.
6590 (ENDFILE_SPEC): Likewise.
6591
39960d1c 65922014-01-29 Jan Hubicka <hubicka@ucw.cz>
3100142a
JH
6593
6594 * ipa-inline-analysis.c (clobber_only_eh_bb_p): New function.
6595 (estimate_function_body_sizes): Use it.
6596
1200933c
PC
65972014-01-29 Paolo Carlini <paolo.carlini@oracle.com>
6598
6599 PR c++/58561
6600 * dwarf2out.c (is_cxx_auto): New.
6601 (is_base_type): Use it.
6602 (gen_type_die_with_usage): Likewise.
6603
68d3bacf
BS
66042014-01-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6605
6606 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Use
6607 CODE_FOR_altivec_vmrg*_direct rather than CODE_FOR_altivec_vmrg*.
6608 * config/rs6000/vsx.md (vsx_mergel_<mode>): Adjust for
6609 -maltivec=be with LE targets.
6610 (vsx_mergeh_<mode>): Likewise.
33425d6c 6611 * config/rs6000/altivec.md (UNSPEC_VMRG[HL]_DIRECT): New unspecs.
68d3bacf
BS
6612 (mulv8hi3): Use gen_altivec_vmrg[hl]w_direct.
6613 (altivec_vmrghb): Replace with define_expand and new
33425d6c 6614 *altivec_vmrghb_internal insn; adjust for -maltivec=be with LE targets.
68d3bacf
BS
6615 (altivec_vmrghb_direct): New define_insn.
6616 (altivec_vmrghh): Replace with define_expand and new
33425d6c 6617 *altivec_vmrghh_internal insn; adjust for -maltivec=be with LE targets.
68d3bacf
BS
6618 (altivec_vmrghh_direct): New define_insn.
6619 (altivec_vmrghw): Replace with define_expand and new
33425d6c 6620 *altivec_vmrghw_internal insn; adjust for -maltivec=be with LE targets.
68d3bacf
BS
6621 (altivec_vmrghw_direct): New define_insn.
6622 (*altivec_vmrghsf): Adjust for endianness.
6623 (altivec_vmrglb): Replace with define_expand and new
33425d6c 6624 *altivec_vmrglb_internal insn; adjust for -maltivec=be with LE targets.
68d3bacf
BS
6625 (altivec_vmrglb_direct): New define_insn.
6626 (altivec_vmrglh): Replace with define_expand and new
33425d6c 6627 *altivec_vmrglh_internal insn; adjust for -maltivec=be with LE targets.
68d3bacf
BS
6628 (altivec_vmrglh_direct): New define_insn.
6629 (altivec_vmrglw): Replace with define_expand and new
33425d6c 6630 *altivec_vmrglw_internal insn; adjust for -maltivec=be with LE targets.
68d3bacf
BS
6631 (altivec_vmrglw_direct): New define_insn.
6632 (*altivec_vmrglsf): Adjust for endianness.
6633 (vec_widen_umult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6634 (vec_widen_umult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6635 (vec_widen_smult_hi_v16qi): Use gen_altivec_vmrghh_direct.
6636 (vec_widen_smult_lo_v16qi): Use gen_altivec_vmrglh_direct.
6637 (vec_widen_umult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6638 (vec_widen_umult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6639 (vec_widen_smult_hi_v8hi): Use gen_altivec_vmrghw_direct.
6640 (vec_widen_smult_lo_v8hi): Use gen_altivec_vmrglw_direct.
6641
aef66c94
MS
66422014-01-29 Marcus Shawcroft <marcus.shawcroft@arm.com>
6643
6644 * config/aarch64/aarch64.c (aarch64_expand_mov_immediate)
6645 (aarch64_legitimate_address_p, aarch64_class_max_nregs): Adjust
6646 whitespace.
6647
77574c35
RB
66482014-01-29 Richard Biener <rguenther@suse.de>
6649
6650 PR tree-optimization/58742
6651 * tree-ssa-forwprop.c (associate_pointerplus): Rename to
6652 associate_pointerplus_align.
6653 (associate_pointerplus_diff): New function.
6654 (associate_pointerplus): Likewise. Call associate_pointerplus_align
6655 and associate_pointerplus_diff.
6656
15b25b24
RB
66572014-01-29 Richard Biener <rguenther@suse.de>
6658
6659 * lto-streamer.h (LTO_major_version): Bump to 3.
6660 (LTO_minor_version): Reset to 0.
6661
eb6006ad
RL
66622014-01-29 Renlin Li <Renlin.Li@arm.com>
6663
6664 * config/arm/arm-arches.def (ARM_ARCH): Add armv7ve arch.
6665 * config/arm/arm.c (FL_FOR_ARCH7VE): New.
6666 (arm_file_start): Generate correct asm header for armv7ve.
6667 * config/arm/bpabi.h: Add multilib support for armv7ve.
6668 * config/arm/driver-arm.c: Change the architectures of cortex-a7
6669 and cortex-a15 to armv7ve.
6670 * config/arm/t-aprofile: Add multilib support for armv7ve.
6671 * doc/invoke.texi: Document -march=armv7ve.
6672
4bfb2fa2
RB
66732014-01-29 Richard Biener <rguenther@suse.de>
6674
6675 PR tree-optimization/58742
6676 * tree-ssa-forwprop.c (associate_plusminus): Return true
6677 if we changed sth, defer EH cleanup to ...
6678 (ssa_forward_propagate_and_combine): ... here. Call simplify_mult.
6679 (simplify_mult): New function.
6680
42eb8bd1
JJ
66812014-01-29 Jakub Jelinek <jakub@redhat.com>
6682
09b22f48
JJ
6683 PR middle-end/59917
6684 PR tree-optimization/59920
6685 * tree.c (build_common_builtin_nodes): Remove
6686 __builtin_setjmp_dispatcher initialization.
6687 * omp-low.h (make_gimple_omp_edges): Add a new int * argument.
6688 * profile.c (branch_prob): Use gsi_start_nondebug_after_labels_bb
6689 instead of gsi_after_labels + manually skipping debug stmts.
6690 Don't ignore bbs with BUILT_IN_SETJMP_DISPATCHER, instead
6691 ignore bbs with IFN_ABNORMAL_DISPATCHER.
6692 * tree-inline.c (copy_edges_for_bb): Remove
6693 can_make_abnormal_goto argument, instead add abnormal_goto_dest
6694 argument. Ignore computed_goto_p stmts. Don't call
6695 make_abnormal_goto_edges. If a call might need abnormal edges
6696 for non-local gotos, see if it already has an edge to
6697 IFN_ABNORMAL_DISPATCHER or if it is IFN_ABNORMAL_DISPATCHER
6698 with true argument, don't do anything then, otherwise add
6699 EDGE_ABNORMAL from the call's bb to abnormal_goto_dest.
6700 (copy_cfg_body): Compute abnormal_goto_dest, adjust copy_edges_for_bb
6701 caller.
6702 * gimple-low.c (struct lower_data): Remove calls_builtin_setjmp.
6703 (lower_function_body): Don't emit __builtin_setjmp_dispatcher.
6704 (lower_stmt): Don't set data->calls_builtin_setjmp.
6705 (lower_builtin_setjmp): Adjust comment.
6706 * builtins.def (BUILT_IN_SETJMP_DISPATCHER): Remove.
6707 * tree-cfg.c (found_computed_goto): Remove.
6708 (factor_computed_gotos): Remove.
6709 (make_goto_expr_edges): Return bool, true for computed gotos.
6710 Don't call make_abnormal_goto_edges.
6711 (build_gimple_cfg): Don't set found_computed_goto, don't call
6712 factor_computed_gotos.
6713 (computed_goto_p): No longer static.
6714 (make_blocks): Don't set found_computed_goto.
6715 (get_abnormal_succ_dispatcher, handle_abnormal_edges): New functions.
6716 (make_edges): If make_goto_expr_edges returns true, push bb
6717 into ab_edge_goto vector, for stmt_can_make_abnormal_goto calls
6718 instead of calling make_abnormal_goto_edges push bb into ab_edge_call
6719 vector. Record mapping between bbs and OpenMP regions if there
6720 are any, adjust make_gimple_omp_edges caller. Call
6721 handle_abnormal_edges.
6722 (make_abnormal_goto_edges): Remove.
6723 * tree-cfg.h (make_abnormal_goto_edges): Remove.
6724 (computed_goto_p, get_abnormal_succ_dispatcher): New prototypes.
6725 * internal-fn.c (expand_ABNORMAL_DISPATCHER): New function.
33425d6c 6726 * builtins.c (expand_builtin): Don't handle BUILT_IN_SETJMP_DISPATCHER.
09b22f48
JJ
6727 * internal-fn.def (ABNORMAL_DISPATCHER): New.
6728 * omp-low.c (make_gimple_omp_edges): Add region_idx argument, when
6729 filling *region also set *region_idx to (*region)->entry->index.
6730
42eb8bd1
JJ
6731 PR other/58712
6732 * read-rtl.c (read_rtx_code): Clear all of RTX_CODE_SIZE (code).
6733 For REGs set ORIGINAL_REGNO.
6734
4a271b7e
BM
67352014-01-29 Bingfeng Mei <bmei@broadcom.com>
6736
33425d6c 6737 * doc/md.texi: Mention that a target shouldn't implement
4a271b7e
BM
6738 vec_widen_(s|u)mul_even/odd pair if it is less efficient
6739 than hi/lo pair.
6740
3d54b29d
JJ
67412014-01-29 Jakub Jelinek <jakub@redhat.com>
6742
6743 PR tree-optimization/59594
6744 * tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Sort
6745 a copy of the datarefs vector rather than the vector itself.
6746
2e5e7103
JM
67472014-01-28 Jason Merrill <jason@redhat.com>
6748
6749 PR c++/53756
6750 * dwarf2out.c (auto_die): New static.
6751 (gen_type_die_with_usage): Handle C++1y 'auto'.
6752 (gen_subprogram_die): If in-class DIE had 'auto', emit type again
6753 on definition.
6754
d5d618b5
L
67552014-01-28 H.J. Lu <hongjiu.lu@intel.com>
6756
6757 PR target/59672
6758 * config/i386/gnu-user64.h (SPEC_32): Add "m16|" to "m32".
6759 (SPEC_X32): Likewise.
6760 (SPEC_64): Likewise.
6761 * config/i386/i386.c (ix86_option_override_internal): Turn off
6762 OPTION_MASK_ISA_64BIT, OPTION_MASK_ABI_X32 and OPTION_MASK_ABI_64
6763 for TARGET_16BIT.
6764 (x86_file_start): Output .code16gcc for TARGET_16BIT.
6765 * config/i386/i386.h (TARGET_16BIT): New macro.
6766 (TARGET_16BIT_P): Likewise.
6767 * config/i386/i386.opt: Add m16.
6768 * doc/invoke.texi: Document -m16.
6769
367c8286
DS
67702014-01-28 Jakub Jelinek <jakub@redhat.com>
6771
6772 PR preprocessor/59935
6773 * input.c (location_get_source_line): Bail out on when line number
33425d6c 6774 is zero, and test the return value of lookup_or_add_file_to_cache_tab.
367c8286 6775
07ec81f9
RB
67762014-01-28 Richard Biener <rguenther@suse.de>
6777
6778 PR tree-optimization/58742
6779 * tree-ssa-forwprop.c (associate_plusminus): Handle
6780 pointer subtraction of the form (T)(P + A) - (T)P.
6781
5facb998
KT
67822014-01-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6783
6784 * config/arm/arm.c (arm_new_rtx_costs): Remove useless statement
6785 at const_int_cost.
6786
e04faf24
RB
67872014-01-28 Richard Biener <rguenther@suse.de>
6788
6789 Revert
6790 2014-01-28 Richard Biener <rguenther@suse.de>
6791
6792 PR rtl-optimization/45364
6793 PR rtl-optimization/59890
6794 * var-tracking.c (local_get_addr_clear_given_value): Handle
6795 already cleared slot.
6796 (val_reset): Handle not allocated local_get_addr_cache.
6797 (vt_find_locations): Use post-order on the inverted CFG.
6798
6593260b
RB
67992014-01-28 Richard Biener <rguenther@suse.de>
6800
33425d6c 6801 * tree-data-ref.h (ddr_is_anti_dependent, ddrs_have_anti_deps): Remove.
6593260b 6802
9ec2d2c1
RB
68032014-01-28 Richard Biener <rguenther@suse.de>
6804
6805 PR rtl-optimization/45364
6806 PR rtl-optimization/59890
6807 * var-tracking.c (local_get_addr_clear_given_value): Handle
6808 already cleared slot.
6809 (val_reset): Handle not allocated local_get_addr_cache.
6810 (vt_find_locations): Use post-order on the inverted CFG.
6811
2ceb362d
AM
68122014-01-28 Alan Modra <amodra@gmail.com>
6813
6814 * Makefile.in (BUILD_CPPFLAGS): Do not use ALL_CPPFLAGS.
6815 * configure.ac <recursive call for build != host>: Define
6816 GENERATOR_FILE. Comment. Use CXX_FOR_BUILD, CXXFLAGS_FOR_BUILD
6817 and LD_FOR_BUILD too.
6818 * configure: Regenerate.
6819
0ebe2584
ASJ
68202014-01-27 Allan Sandfeld Jensen <sandfeld@kde.org>
6821
6822 * config/i386/i386.c (get_builtin_code_for_version): Separate
6823 Westmere from Nehalem, Ivy Bridge from Sandy Bridge and
6824 Broadwell from Haswell.
6825
a33fc7fe
SE
68262014-01-27 Steve Ellcey <sellcey@mips.com>
6827
6828 * common/config/mips/mips-common.c (TARGET_DEFAULT_TARGET_FLAGS):
6829 Remove TARGET_FP_EXCEPTIONS_DEFAULT and MASK_FUSED_MADD.
6830 * config/mips/mips.c (mips_option_override): Change setting
6831 of TARGET_DSP.
6832 * config/mips/mips.h (TARGET_FP_EXCEPTIONS_DEFAULT): Remove.
0ebe2584
ASJ
6833 * config/mips/mips.opt (DSP, DSPR2, FP_EXCEPTIONS, FUSED_MADD, MIPS3D):
6834 Change from Mask to Var.
a33fc7fe 6835
a99be3c9
JL
68362014-01-27 Jeff Law <law@redhat.com>
6837
6838 * ipa-inline.c (inline_small_functions): Fix typo.
6839
d256b866
IT
68402014-01-27 Ilya Tocar <ilya.tocar@intel.com>
6841
6842 * config/i386/avx512fintrin.h (_mm512_mask_cvtepi32_storeu_epi8): New.
6843 (_mm512_mask_cvtsepi32_storeu_epi8): Ditto.
6844 (_mm512_mask_cvtusepi32_storeu_epi8): Ditto.
6845 (_mm512_mask_cvtepi32_storeu_epi16): Ditto.
6846 (_mm512_mask_cvtsepi32_storeu_epi16): Ditto.
6847 (_mm512_mask_cvtusepi32_storeu_epi16): Ditto.
6848 (_mm512_mask_cvtepi64_storeu_epi32): Ditto.
6849 (_mm512_mask_cvtsepi64_storeu_epi32): Ditto.
6850 (_mm512_mask_cvtusepi64_storeu_epi32): Ditto.
6851 (_mm512_mask_cvtepi64_storeu_epi16): Ditto.
6852 (_mm512_mask_cvtsepi64_storeu_epi16): Ditto.
6853 (_mm512_mask_cvtusepi64_storeu_epi16): Ditto.
6854 (_mm512_mask_cvtepi64_storeu_epi8): Ditto.
6855 (_mm512_mask_cvtsepi64_storeu_epi8): Ditto.
6856 (_mm512_mask_cvtusepi64_storeu_epi8): Ditto.
6857 (_mm512_storeu_epi64): Ditto.
6858 (_mm512_cmpge_epi32_mask): Ditto.
6859 (_mm512_cmpge_epu32_mask): Ditto.
6860 (_mm512_cmpge_epi64_mask): Ditto.
6861 (_mm512_cmpge_epu64_mask): Ditto.
6862 (_mm512_cmple_epi32_mask): Ditto.
6863 (_mm512_cmple_epu32_mask): Ditto.
6864 (_mm512_cmple_epi64_mask): Ditto.
6865 (_mm512_cmple_epu64_mask): Ditto.
6866 (_mm512_cmplt_epi32_mask): Ditto.
6867 (_mm512_cmplt_epu32_mask): Ditto.
6868 (_mm512_cmplt_epi64_mask): Ditto.
6869 (_mm512_cmplt_epu64_mask): Ditto.
6870 (_mm512_cmpneq_epi32_mask): Ditto.
6871 (_mm512_cmpneq_epu32_mask): Ditto.
6872 (_mm512_cmpneq_epi64_mask): Ditto.
6873 (_mm512_cmpneq_epu64_mask): Ditto.
6874 (_mm512_expand_pd): Ditto.
6875 (_mm512_expand_ps): Ditto.
6876 * config/i386/i386-builtin-types.def: Add PV16QI, PV16QI, PV16HI,
6877 VOID_PV8SI_V8DI_QI, VOID_PV8HI_V8DI_QI, VOID_PV16QI_V8DI_QI,
6878 VOID_PV16QI_V16SI_HI, VOID_PV16HI_V16SI_HI.
6879 * config/i386/i386.c (ix86_builtins): Add
6880 IX86_BUILTIN_EXPANDPD512_NOMASK, IX86_BUILTIN_EXPANDPS512_NOMASK,
6881 IX86_BUILTIN_PMOVDB512_MEM, IX86_BUILTIN_PMOVDW512_MEM,
6882 IX86_BUILTIN_PMOVQB512_MEM, IX86_BUILTIN_PMOVQD512_MEM,
6883 IX86_BUILTIN_PMOVQW512_MEM, IX86_BUILTIN_PMOVSDB512_MEM,
6884 IX86_BUILTIN_PMOVSDW512_MEM, IX86_BUILTIN_PMOVSQB512_MEM,
6885 IX86_BUILTIN_PMOVSQD512_MEM, IX86_BUILTIN_PMOVSQW512_MEM,
6886 IX86_BUILTIN_PMOVUSDB512_MEM, IX86_BUILTIN_PMOVUSDW512_MEM,
6887 IX86_BUILTIN_PMOVUSQB512_MEM, IX86_BUILTIN_PMOVUSQD512_MEM,
6888 IX86_BUILTIN_PMOVUSQW512_MEM.
6889 (bdesc_special_args): Add __builtin_ia32_pmovusqd512mem_mask,
6890 __builtin_ia32_pmovsqd512mem_mask,
6891 __builtin_ia32_pmovqd512mem_mask,
6892 __builtin_ia32_pmovusqw512mem_mask,
6893 __builtin_ia32_pmovsqw512mem_mask,
6894 __builtin_ia32_pmovqw512mem_mask,
6895 __builtin_ia32_pmovusdw512mem_mask,
6896 __builtin_ia32_pmovsdw512mem_mask,
6897 __builtin_ia32_pmovdw512mem_mask,
6898 __builtin_ia32_pmovqb512mem_mask,
6899 __builtin_ia32_pmovusqb512mem_mask,
6900 __builtin_ia32_pmovsqb512mem_mask,
6901 __builtin_ia32_pmovusdb512mem_mask,
6902 __builtin_ia32_pmovsdb512mem_mask,
6903 __builtin_ia32_pmovdb512mem_mask.
6904 (bdesc_args): Add __builtin_ia32_expanddf512,
6905 __builtin_ia32_expandsf512.
6906 (ix86_expand_special_args_builtin): Handle VOID_FTYPE_PV8SI_V8DI_QI,
6907 VOID_FTYPE_PV8HI_V8DI_QI, VOID_FTYPE_PV16HI_V16SI_HI,
6908 VOID_FTYPE_PV16QI_V8DI_QI, VOID_FTYPE_PV16QI_V16SI_HI.
6909 * config/i386/sse.md (unspec): Add UNSPEC_EXPAND_NOMASK.
6910 (avx512f_<code><pmov_src_lower><mode>2_mask_store): New.
6911 (*avx512f_<code>v8div16qi2_store_mask): Renamed to ...
6912 (avx512f_<code>v8div16qi2_mask_store): This.
6913 (avx512f_expand<mode>): New.
6914
e711dffd
KY
69152014-01-27 Kirill Yukhin <kirill.yukhin@intel.com>
6916
1287ae50 6917 * config/i386/avx512pfintrin.h (_mm512_mask_prefetch_i32gather_pd):
0ebe2584 6918 New.
e711dffd
KY
6919 (_mm512_mask_prefetch_i64gather_pd): Ditto.
6920 (_mm512_prefetch_i32scatter_pd): Ditto.
6921 (_mm512_mask_prefetch_i32scatter_pd): Ditto.
6922 (_mm512_prefetch_i64scatter_pd): Ditto.
6923 (_mm512_mask_prefetch_i64scatter_pd): Ditto.
6924 (_mm512_mask_prefetch_i32gather_ps): Fix operand type.
6925 (_mm512_mask_prefetch_i64gather_ps): Ditto.
6926 (_mm512_prefetch_i32scatter_ps): Ditto.
6927 (_mm512_mask_prefetch_i32scatter_ps): Ditto.
6928 (_mm512_prefetch_i64scatter_ps): Ditto.
6929 (_mm512_mask_prefetch_i64scatter_ps): Ditto.
6930 * config/i386/i386-builtin-types.def: Define
0ebe2584
ASJ
6931 VOID_FTYPE_QI_V8SI_PCINT64_INT_INT
6932 and VOID_FTYPE_QI_V8DI_PCINT64_INT_INT.
e711dffd
KY
6933 * config/i386/i386.c (ix86_builtins): Define IX86_BUILTIN_GATHERPFQPD,
6934 IX86_BUILTIN_GATHERPFDPD, IX86_BUILTIN_SCATTERPFDPD,
6935 IX86_BUILTIN_SCATTERPFQPD.
6936 (ix86_init_mmx_sse_builtins): Define __builtin_ia32_gatherpfdpd,
6937 __builtin_ia32_gatherpfdps, __builtin_ia32_gatherpfqpd,
6938 __builtin_ia32_gatherpfqps, __builtin_ia32_scatterpfdpd,
6939 __builtin_ia32_scatterpfdps, __builtin_ia32_scatterpfqpd,
6940 __builtin_ia32_scatterpfqps.
6941 (ix86_expand_builtin): Expand new built-ins.
6942 * config/i386/sse.md (avx512pf_gatherpf<mode>): Add SF suffix,
6943 fix memory access data type.
6944 (*avx512pf_gatherpf<mode>_mask): Ditto.
6945 (*avx512pf_gatherpf<mode>): Ditto.
6946 (avx512pf_scatterpf<mode>): Ditto.
6947 (*avx512pf_scatterpf<mode>_mask): Ditto.
6948 (*avx512pf_scatterpf<mode>): Ditto.
6949 (GATHER_SCATTER_SF_MEM_MODE): New.
6950 (avx512pf_gatherpf<mode>df): Ditto.
6951 (*avx512pf_gatherpf<mode>df_mask): Ditto.
6952 (*avx512pf_scatterpf<mode>df): Ditto.
6953
904e5ccd
JJ
69542014-01-27 Jakub Jelinek <jakub@redhat.com>
6955
6956 PR bootstrap/59934
6957 * expmed.h (expmed_mode_index): Rework so that analysis and optimziers
6958 know when the MODE_PARTIAL_INT and MODE_VECTOR_INT cases can never be
6959 reached.
6960
84e90123
JG
69612014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6962
6963 * common/config/arm/arm-common.c
6964 (arm_rewrite_mcpu): Handle multiple names.
6965 * config/arm/arm.h
6966 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
6967
c6f6157a
JG
69682014-01-27 James Greenhalgh <james.greenhalgh@arm.com>
6969
6970 * gimple-builder.h (create_gimple_tmp): Delete.
6971
770516c9
CB
69722014-01-27 Christian Bruel <christian.bruel@st.com>
6973
6974 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Fix remaining bytes after
6975 words comparisons.
6976
cc5cec10
JDA
69772014-01-26 John David Anglin <danglin@gcc.gnu.org>
6978
5aa3c762
JDA
6979 * config/pa/pa.md (call): Generate indirect long calls to non-local
6980 functions when outputing 32-bit code.
6981 (call_value): Likewise except for special call to buggy powf function.
6982
cc5cec10
JDA
6983 * config/pa/pa.c (pa_attr_length_indirect_call): Adjust length of
6984 portable runtime and PIC indirect calls.
6985 (pa_output_indirect_call): Remove unnecessary nop from portable runtime
6986 and PIC call sequences. Use ldo instead of blr to set return register
6987 in PIC call sequence.
6988
6bb0e248
WL
69892014-01-25 Walter Lee <walt@tilera.com>
6990
6991 * config/tilegx/sync.md (atomic_fetch_sub): Fix negation and
6992 avoid clobbering a live register.
6993
dab03fe3
WL
69942014-01-25 Walter Lee <walt@tilera.com>
6995
770516c9 6996 * config/tilegx/tilegx-c.c (tilegx_cpu_cpp_builtins):
dab03fe3 6997 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2}.
770516c9 6998 * config/tilegx/tilepro-c.c (tilepro_cpu_cpp_builtins):
dab03fe3
WL
6999 Define __GCC_HAVE_SYNC_COMPARE_AND_SWAP_{1,2,4,8}.
7000
905c20c1
WL
70012014-01-25 Walter Lee <walt@tilera.com>
7002
7003 * config/tilegx/tilegx.c (tilegx_function_arg): Start 16-byte
7004 arguments on even registers.
7005 (tilegx_gimplify_va_arg_expr): Align 16-byte var args to
7006 STACK_BOUNDARY.
7007 * config/tilegx/tilegx.h (STACK_BOUNDARY): Change to 16 bytes.
7008 (BIGGEST_ALIGNMENT): Ditto.
7009 (BIGGEST_FIELD_ALIGNMENT): Ditto.
7010
8e90a625
WL
70112014-01-25 Walter Lee <walt@tilera.com>
7012
7013 * config/tilegx/tilegx.c (tilegx_gen_bundles): Delete barrier
7014 insns before bundling.
0ebe2584 7015 * config/tilegx/tilegx.md (tile_network_barrier): Update comment.
8e90a625 7016
450c1ffe
WL
70172014-01-25 Walter Lee <walt@tilera.com>
7018
7019 * config/tilegx/tilegx.c (tilegx_expand_builtin): Set
7020 PREFETCH_SCHEDULE_BARRIER_P to true for prefetches.
7021 * config/tilepro/tilepro.c (tilepro_expand_builtin): Ditto.
317a951f 7022
450c1ffe 70232014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
770516c9 7024
317a951f
RS
7025 * config/mips/constraints.md (kl): Delete.
7026 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Turn into
7027 define expands, using...
7028 (divmod<mode>4_mips16, udivmod<mode>4_mips16): ...these new
7029 instructions for MIPS16.
7030 (*divmod<mode>4, *udivmod<mode>4): New patterns, taken from the
7031 non-MIPS16 version of the old divmod<mode>4 and udivmod<mode>4.
7032
3abe9053
WL
70332014-01-25 Walter Lee <walt@tilera.com>
7034
0ebe2584 7035 * config/tilepro/tilepro.md (ctzdi2): Use register_operand predicate.
3abe9053
WL
7036 (clzdi2): Ditto.
7037 (ffsdi2): Ditto.
7038
b0e0fe41
WL
70392014-01-25 Walter Lee <walt@tilera.com>
7040
7041 * config/tilegx/tilegx.c (tilegx_expand_to_rtl_hook): New.
7042 (TARGET_EXPAND_TO_RTL_HOOK): Define.
7043
2c8798a2
RS
70442014-01-25 Richard Sandiford <rdsandiford@googlemail.com>
7045
7046 * rtlanal.c (canonicalize_condition): Split out duplicated mode check.
7047 Handle XOR.
7048
2105be5a
JJ
70492014-01-25 Jakub Jelinek <jakub@redhat.com>
7050
4def6060
JJ
7051 * print-rtl.c (in_call_function_usage): New var.
7052 (print_rtx): When in CALL_INSN_FUNCTION_USAGE, always print
7053 EXPR_LIST mode as mode and not as reg note name.
7054
2105be5a
JJ
7055 PR middle-end/59561
7056 * cfgloopmanip.c (copy_loop_info): If
7057 loop->warned_aggressive_loop_optimizations, make sure
7058 the flag is set in target loop too.
7059
b72271b9
BI
70602014-01-24 Balaji V. Iyer <balaji.v.iyer@intel.com>
7061
7062 * builtins.c (is_builtin_name): Renamed flag_enable_cilkplus to
7063 flag_cilkplus.
7064 * builtins.def: Likewise.
7065 * cilk.h (fn_contains_cilk_spawn_p): Likewise.
7066 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Likewise.
7067 * ira.c (ira_setup_eliminable_regset): Likewise.
7068 * omp-low.c (gate_expand_omp): Likewise.
7069 (execute_lower_omp): Likewise.
7070 (diagnose_sb_0): Likewise.
7071 (gate_diagnose_omp_blocks): Likewise.
7072 (simd_clone_clauses_extract): Likewise.
7073 (gate): Likewise.
7074
8adcc78b
BS
70752014-01-24 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
7076
7077 * config/rs6000/rs6000.c (rs6000_expand_vec_perm_const_1): Remove
7078 correction for little endian...
7079 * config/rs6000/vsx.md (vsx_xxpermdi2_<mode>_1): ...and move it to
7080 here.
7081
3d750496
JL
70822014-01-24 Jeff Law <law@redhat.com>
7083
7084 PR tree-optimization/59919
7085 * tree-vrp.c (find_assert_locations_1): Do not register asserts
7086 for non-returning calls.
7087
1c05df59
JG
70882014-01-24 James Greenhalgh <james.greenhalgh@arm.com>
7089
7090 * common/config/aarch64/aarch64-common.c
7091 (aarch64_rewrite_mcpu): Handle multiple names.
7092 * config/aarch64/aarch64.h
7093 (BIG_LITTLE_SPEC): Do not discard mcpu switches.
7094
317363b4
DS
70952014-01-24 Dodji Seketeli <dodji@redhat.com>
7096
7097 * input.c (add_file_to_cache_tab): Handle the case where fopen
7098 returns NULL.
7099
16370fa7
L
71002014-01-23 H.J. Lu <hongjiu.lu@intel.com>
7101
7102 PR target/59929
7103 * config/i386/i386.md (pushsf splitter): Get stack adjustment
7104 from push operand if code of push isn't PRE_DEC.
7105
b846c948
MM
71062014-01-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7107
7108 PR target/59909
7109 * doc/invoke.texi (RS/6000 and PowerPC Options): Document
7110 -mquad-memory-atomic. Update -mquad-memory documentation to say
7111 it is only used for non-atomic loads/stores.
7112
7113 * config/rs6000/predicates.md (quad_int_reg_operand): Allow either
7114 -mquad-memory or -mquad-memory-atomic switches.
7115
7116 * config/rs6000/rs6000-cpus.def (ISA_2_7_MASKS_SERVER): Add
7117 -mquad-memory-atomic to ISA 2.07 support.
7118
7119 * config/rs6000/rs6000.opt (-mquad-memory-atomic): Add new switch
0ebe2584
ASJ
7120 to separate support of normal quad word memory operations (ldq, stq)
7121 from the atomic quad word memory operations.
b846c948
MM
7122
7123 * config/rs6000/rs6000.c (rs6000_option_override_internal): Add
7124 support to separate non-atomic quad word operations from atomic
7125 quad word operations. Disable non-atomic quad word operations in
7126 little endian mode so that we don't have to swap words after the
7127 load and before the store.
7128 (quad_load_store_p): Add comment about atomic quad word support.
7129 (rs6000_opt_masks): Add -mquad-memory-atomic to the list of
7130 options printed with -mdebug=reg.
7131
7132 * config/rs6000/rs6000.h (TARGET_SYNC_TI): Use
7133 -mquad-memory-atomic as the test for whether we have quad word
7134 atomic instructions.
0ebe2584
ASJ
7135 (TARGET_SYNC_HI_QI): If either -mquad-memory-atomic, -mquad-memory,
7136 or -mp8-vector are used, allow byte/half-word atomic operations.
b846c948
MM
7137
7138 * config/rs6000/sync.md (load_lockedti): Insure that the address
7139 is a proper indexed or indirect address for the lqarx instruction.
7140 On little endian systems, swap the hi/lo registers after the lqarx
7141 instruction.
7142 (load_lockedpti): Use indexed_or_indirect_operand predicate to
7143 insure the address is valid for the lqarx instruction.
7144 (store_conditionalti): Insure that the address is a proper indexed
7145 or indirect address for the stqcrx. instruction. On little endian
7146 systems, swap the hi/lo registers before doing the stqcrx.
7147 instruction.
7148 (store_conditionalpti): Use indexed_or_indirect_operand predicate to
7149 insure the address is valid for the stqcrx. instruction.
7150
7151 * gcc/config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
7152 Define __QUAD_MEMORY__ and __QUAD_MEMORY_ATOMIC__ based on what
7153 type of quad memory support is available.
7154
6e23f296
VM
71552014-01-23 Vladimir Makarov <vmakarov@redhat.com>
7156
7157 PR regression/59915
7158 * lra-constraints.c (simplify_operand_subreg): Spill pseudo if
7159 there is a danger of looping.
7160
0a4f04e5
PH
71612014-01-23 Pat Haugen <pthaugen@us.ibm.com>
7162
7163 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
7164 force flag_ira_loop_pressure if set via command line.
7165
f9a4c9a6
AV
71662014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7167
7168 * config/aarch64/aarch64-simd-builtins.def (ashr): DI mode removed.
7169 (ashr_simd): New builtin handling DI mode.
7170 * config/aarch64/aarch64-simd.md (aarch64_ashr_simddi): New pattern.
7171 (aarch64_sshr_simddi): New match pattern.
7172 * config/aarch64/arm_neon.h (vshr_n_s32): Builtin call modified.
7173 (vshrd_n_s64): Likewise.
7174 * config/aarch64/predicates.md (aarch64_shift_imm64_di): New predicate.
7175
949ad971
NC
71762014-01-23 Nick Clifton <nickc@redhat.com>
7177
7178 * config/msp430/msp430.h (ASM_SPEC): Pass the -mcpu as -mcpu.
7179 (LIB_SPEC): Drop use of memory.ld and peripherals.ld scripts in
7180 favour of mcu specific scripts.
7181 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more matches for
7182 430x multilibs.
7183
d5ecead9
JG
71842014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7185 Alex Velenko <Alex.Velenko@arm.com>
7186
7187 * config/aarch64/arm_neon.h (vaddv_s8): __LANE0 cleanup.
7188 (vaddv_s16): Likewise.
7189 (vaddv_s32): Likewise.
7190 (vaddv_u8): Likewise.
7191 (vaddv_u16): Likewise.
7192 (vaddv_u32): Likewise.
7193 (vaddvq_s8): Likewise.
7194 (vaddvq_s16): Likewise.
7195 (vaddvq_s32): Likewise.
7196 (vaddvq_s64): Likewise.
7197 (vaddvq_u8): Likewise.
7198 (vaddvq_u16): Likewise.
7199 (vaddvq_u32): Likewise.
7200 (vaddvq_u64): Likewise.
7201 (vaddv_f32): Likewise.
7202 (vaddvq_f32): Likewise.
7203 (vaddvq_f64): Likewise.
7204 (vmaxv_f32): Likewise.
7205 (vmaxv_s8): Likewise.
7206 (vmaxv_s16): Likewise.
7207 (vmaxv_s32): Likewise.
7208 (vmaxv_u8): Likewise.
7209 (vmaxv_u16): Likewise.
7210 (vmaxv_u32): Likewise.
7211 (vmaxvq_f32): Likewise.
7212 (vmaxvq_f64): Likewise.
7213 (vmaxvq_s8): Likewise.
7214 (vmaxvq_s16): Likewise.
7215 (vmaxvq_s32): Likewise.
7216 (vmaxvq_u8): Likewise.
7217 (vmaxvq_u16): Likewise.
7218 (vmaxvq_u32): Likewise.
7219 (vmaxnmv_f32): Likewise.
7220 (vmaxnmvq_f32): Likewise.
7221 (vmaxnmvq_f64): Likewise.
7222 (vminv_f32): Likewise.
7223 (vminv_s8): Likewise.
7224 (vminv_s16): Likewise.
7225 (vminv_s32): Likewise.
7226 (vminv_u8): Likewise.
7227 (vminv_u16): Likewise.
7228 (vminv_u32): Likewise.
7229 (vminvq_f32): Likewise.
7230 (vminvq_f64): Likewise.
7231 (vminvq_s8): Likewise.
7232 (vminvq_s16): Likewise.
7233 (vminvq_s32): Likewise.
7234 (vminvq_u8): Likewise.
7235 (vminvq_u16): Likewise.
7236 (vminvq_u32): Likewise.
7237 (vminnmv_f32): Likewise.
7238 (vminnmvq_f32): Likewise.
7239 (vminnmvq_f64): Likewise.
7240
1dd055a2
JG
72412014-01-23 James Greenhalgh <james.greenhalgh@arm.com>
7242
7243 * config/aarch64/aarch64-simd.md
7244 (aarch64_dup_lane<mode>): Correct lane number on big-endian.
7245 (aarch64_dup_lane_<vswap_widthi_name><mode>): Likewise.
7246 (*aarch64_mul3_elt<mode>): Likewise.
7247 (*aarch64_mul3_elt<vswap_width_name><mode>): Likewise.
7248 (*aarch64_mul3_elt_to_64v2df): Likewise.
7249 (*aarch64_mla_elt<mode>): Likewise.
7250 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7251 (*aarch64_mls_elt<mode>): Likewise.
7252 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7253 (*aarch64_fma4_elt<mode>): Likewise.
7254 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7255 (*aarch64_fma4_elt_to_64v2df): Likewise.
7256 (*aarch64_fnma4_elt<mode>): Likewise.
7257 (*aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7258 (*aarch64_fnma4_elt_to_64v2df): Likewise.
7259 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
7260 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
7261 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7262 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
7263 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
7264 (aarch64_sqdmull_lane<mode>_internal): Likewise.
7265 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
7266
dafb9b64
AV
72672013-01-23 Alex Velenko <Alex.Velenko@arm.com>
7268
7269 * config/aarch64/aarch64-simd.md
7270 (aarch64_be_checked_get_lane<mode>): New define_expand.
7271 * config/aarch64/aarch64-simd-builtins.def
0ebe2584 7272 (BUILTIN_VALL (GETLANE, be_checked_get_lane, 0)):
dafb9b64
AV
7273 New builtin definition.
7274 * config/aarch64/arm_neon.h: (__aarch64_vget_lane_any):
7275 Use new safe be builtin.
7276
89b4515c
AV
72772014-01-23 Alex Velenko <Alex.Velenko@arm.com>
7278
7279 * config/aarch64/aarch64-simd.md (aarch64_be_ld1<mode>):
7280 New define_insn.
7281 (aarch64_be_st1<mode>): Likewise.
7282 (aarch_ld1<VALL:mode>): Define_expand modified.
7283 (aarch_st1<VALL:mode>): Likewise.
7284 * config/aarch64/aarch64.md (UNSPEC_LD1): New unspec definition.
7285 (UNSPEC_ST1): Likewise.
7286
4bb66ef3 72872014-01-23 David Holsgrove <david.holsgrove@xilinx.com>
f3eeb82c
DH
7288
7289 * config/microblaze/microblaze.md: Add trap insn and attribute
7290
7ecc3eb9
DS
72912014-01-23 Dodji Seketeli <dodji@redhat.com>
7292
7293 PR preprocessor/58580
7294 * input.h (location_get_source_line): Take an additional line_size
7295 parameter.
7296 (void diagnostics_file_cache_fini): Declare new function.
7297 * input.c (struct fcache): New type.
7298 (fcache_tab_size, fcache_buffer_size, fcache_line_record_size):
7299 New static constants.
7300 (diagnostic_file_cache_init, total_lines_num)
7301 (lookup_file_in_cache_tab, evicted_cache_tab_entry)
7302 (add_file_to_cache_tab, lookup_or_add_file_to_cache_tab)
7303 (needs_read, needs_grow, maybe_grow, read_data, maybe_read_data)
7304 (get_next_line, read_next_line, goto_next_line, read_line_num):
7305 New static function definitions.
7306 (diagnostic_file_cache_fini): New function.
7307 (location_get_source_line): Take an additional output line_len
7308 parameter. Re-write using lookup_or_add_file_to_cache_tab and
7309 read_line_num.
7310 * diagnostic.c (diagnostic_finish): Call
7311 diagnostic_file_cache_fini.
7312 (adjust_line): Take an additional input parameter for the length
7313 of the line, rather than calculating it with strlen.
7314 (diagnostic_show_locus): Adjust the use of
7315 location_get_source_line and adjust_line with respect to their new
7316 signature. While displaying a line now, do not stop at the first
7317 null byte. Rather, display the zero byte as a space and keep
7318 going until we reach the size of the line.
7319 * Makefile.in: Add vec.o to OBJS-libcommon
7320
2196a885
KY
73212014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7322 Ilya Tocar <ilya.tocar@intel.com>
7323
7324 * config/i386/avx512fintrin.h (_mm512_kmov): New.
7325 * config/i386/i386.c (IX86_BUILTIN_KMOV16): Ditto.
7326 (__builtin_ia32_kmov16): Ditto.
7327 * config/i386/i386.md (UNSPEC_KMOV): New.
7328 (kmovw): Ditto.
7329
70473c63
KY
73302014-01-23 Kirill Yukhin <kirill.yukhin@intel.com>
7331
7332 * config/i386/avx512fintrin.h (_mm512_loadu_si512): Rename.
7333 (_mm512_storeu_si512): Ditto.
7334
f9ae4df8
RS
73352014-01-23 Richard Sandiford <rdsandiford@googlemail.com>
7336
7337 PR target/52125
7338 * rtl.h (get_referenced_operands): Declare.
7339 * recog.c (get_referenced_operands): New function.
7340 * config/mips/mips.c (mips_reorg_process_insns): Check which asm
7341 operands have been referenced when recording LO_SUM references.
7342
4bb66ef3 73432014-01-22 David Holsgrove <david.holsgrove@xilinx.com>
e50f5f2e
DH
7344
7345 * config/microblaze/microblaze.md: Correct bswaphi2 insn.
7346
39960d1c 73472014-01-22 Jan Hubicka <hubicka@ucw.cz>
33b64438
JH
7348
7349 * config/i386/x86-tune.def (X86_TUNE_ACCUMULATE_OUTGOING_ARGS):
7350 Enable for generic and recent AMD targets.
7351
39960d1c 73522014-01-22 Jan Hubicka <hubicka@ucw.cz>
77b7a218
JH
7353
7354 * combine-stack-adj.c (combine_stack_adjustments_for_block): Remove
7355 ARG_SIZE note when adjustment was eliminated.
7356
bb50b870
JL
73572014-01-22 Jeff Law <law@redhat.com>
7358
7359 PR tree-optimization/59597
7360 * tree-ssa-threadupdate.c (dump_jump_thread_path): Move to earlier
7361 in file. Accept new argument REGISTERING and use it to modify
7362 dump output appropriately.
7363 (register_jump_thread): Corresponding changes.
7364 (mark_threaded_blocks): Reinstate code to cancel unprofitable
7365 thread paths involving joiner blocks. Add code to dump cancelled
7366 jump threading paths.
7367
df2980be
VM
73682014-01-22 Vladimir Makarov <vmakarov@redhat.com>
7369
7370 PR rtl-optimization/59477
7371 * lra-constraints.c (inherit_in_ebb): Process call for living hard
0ebe2584 7372 regs. Update reloads_num and potential_reload_hard_regs for all insns.
df2980be 7373
1bb99900
TT
73742014-01-22 Tom Tromey <tromey@redhat.com>
7375
7376 * config/i386/i386-interix.h (i386_pe_unique_section): Don't use
7377 PARAMS.
7378 * config/cr16/cr16-protos.h (notice_update_cc): Don't use PARAMS.
7379
9102dadd
VM
73802014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7381
7382 PR rtl-optimization/59896
7383 * lra-constraints.c (process_alt_operands): Check unused note for
7384 matched operands of insn with no output reloads.
7385
ca376eb8
RS
73862014-01-21 Richard Sandiford <rdsandiford@googlemail.com>
7387
7388 * config/mips/mips.c (mips_move_to_gpr_cost): Add M16_REGS case.
7389 (mips_move_from_gpr_cost): Likewise.
7390
a9711f36
VM
73912014-01-21 Vladimir Makarov <vmakarov@redhat.com>
7392
7393 PR rtl-optimization/59858
7394 * lra-constraints.c (SMALL_REGISTER_CLASS_P): Use
7395 ira_class_hard_regs_num.
7396 (process_alt_operands): Increase reject for dying matched operand.
7397
f8ea7cb0
JJ
73982014-01-21 Jakub Jelinek <jakub@redhat.com>
7399
7400 PR target/59003
7401 * config/i386/i386.c (expand_small_movmem_or_setmem): If mode is
7402 smaller than size, perform several stores or loads and stores
7403 at dst + count - size to store or copy all of size bytes, rather
7404 than just last modesize bytes.
7405
84db09e3
DD
74062014-01-20 DJ Delorie <dj@redhat.com>
7407
7408 * config/rl78/rl78.c (rl78_propogate_register_origins): Verify
7409 that CLOBBERs are REGs before propogating their values.
7410
1cf11770
L
74112014-01-20 H.J. Lu <hongjiu.lu@intel.com>
7412
7413 PR middle-end/59789
7414 * cgraph.c (cgraph_inline_failed_string): Add type to DEFCIFCODE.
7415 (cgraph_inline_failed_type): New function.
7416 * cgraph.h (DEFCIFCODE): Add type.
7417 (cgraph_inline_failed_type_t): New enum.
7418 (cgraph_inline_failed_type): New prototype.
7419 * cif-code.def: Add CIF_FINAL_NORMAL to OK, FUNCTION_NOT_CONSIDERED,
7420 FUNCTION_NOT_OPTIMIZED, REDEFINED_EXTERN_INLINE,
7421 FUNCTION_NOT_INLINE_CANDIDATE, LARGE_FUNCTION_GROWTH_LIMIT,
7422 LARGE_STACK_FRAME_GROWTH_LIMIT, MAX_INLINE_INSNS_SINGLE_LIMIT,
7423 MAX_INLINE_INSNS_AUTO_LIMIT, INLINE_UNIT_GROWTH_LIMIT,
7424 RECURSIVE_INLINING, UNLIKELY_CALL, NOT_DECLARED_INLINED,
7425 OPTIMIZING_FOR_SIZE, ORIGINALLY_INDIRECT_CALL,
0ebe2584 7426 INDIRECT_UNKNOWN_CALL, USES_COMDAT_LOCAL.
1cf11770
L
7427 Add CIF_FINAL_ERROR to UNSPECIFIED, BODY_NOT_AVAILABLE,
7428 FUNCTION_NOT_INLINABLE, OVERWRITABLE, MISMATCHED_ARGUMENTS,
7429 EH_PERSONALITY, NON_CALL_EXCEPTIONS, TARGET_OPTION_MISMATCH,
7430 OPTIMIZATION_MISMATCH.
7431 * tree-inline.c (expand_call_inline): Emit errors during
0ebe2584 7432 early_inlining if cgraph_inline_failed_type returns CIF_FINAL_ERROR.
1cf11770 7433
8586e4bd
UB
74342014-01-20 Uros Bizjak <ubizjak@gmail.com>
7435
7436 PR target/59685
7437 * config/i386/sse.md (*andnot<mode>3<mask_name>): Handle MODE_V16SF
7438 mode attribute in insn output.
7439
99482090
EB
74402014-01-20 Eric Botcazou <ebotcazou@adacore.com>
7441
7442 * output.h (output_constant): Delete.
7443 * varasm.c (output_constant): Make private.
7444
ede23272
AV
74452014-01-20 Alex Velenko <Alex.Velenko@arm.com>
7446
7447 * config/aarch64/aarch64-simd.md (vec_perm<mode>): Add BE check.
7448
9a7eefec
JJ
74492014-01-20 Jakub Jelinek <jakub@redhat.com>
7450
7451 PR middle-end/59860
7452 * tree.h (fold_builtin_strcat): New prototype.
7453 * builtins.c (fold_builtin_strcat): No longer static. Add len
7454 argument, if non-NULL, don't call c_strlen. Optimize
7455 directly into __builtin_memcpy instead of __builtin_strcpy.
7456 (fold_builtin_2): Adjust fold_builtin_strcat caller.
7457 * gimple-fold.c (gimple_fold_builtin): Handle BUILT_IN_STRCAT.
7458
3e729145
UB
74592014-01-20 Uros Bizjak <ubizjak@gmail.com>
7460
7461 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7462 for SImode_address_operand operands, having only a REG argument.
7463
eee0e487
MS
74642014-01-20 Marcus Shawcroft <marcus.shawcroft@arm.com>
7465
7466 * config/aarch64/aarch64-linux.h (GLIBC_DYNAMIC_LINKER): Expand
7467 loader name using mbig-endian.
7468 (LINUX_TARGET_LINK_SPEC): Pass linker -m flag.
7469
9e540e37
JG
74702014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7471
7472 * doc/invoke.texi (-march): Clarify documentation for AArch64.
7473 (-mtune): Likewise.
7474 (-mcpu): Likewise.
7475
69675d50
TB
74762014-01-20 Tejas Belagod <tejas.belagod@arm.com>
7477
7478 * config/aarch64/aarch64-protos.h
7479 (aarch64_cannot_change_mode_class_ptr): Declare.
7480 * config/aarch64/aarch64.c (aarch64_cannot_change_mode_class,
7481 aarch64_cannot_change_mode_class_ptr): New.
7482 * config/aarch64/aarch64.h (CANNOT_CHANGE_MODE_CLASS): Change to call
7483 backend hook aarch64_cannot_change_mode_class.
7484
ffee7aa9
JG
74852014-01-20 James Greenhalgh <james.greenhalgh@arm.com>
7486
7487 * common/config/aarch64/aarch64-common.c
7488 (aarch64_handle_option): Don't handle any option order logic here.
7489 * config/aarch64/aarch64.c (aarch64_parse_arch): Do not override
7490 selected_cpu, warn on architecture version mismatch.
7491 (aarch64_override_options): Fix parsing order for option strings.
7492
c7169779
JBG
74932014-01-20 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7494 Iain Sandoe <iain@codesourcery.com>
7495
7496 PR bootstrap/59496
7497 * config/rs6000/darwin.h (ADJUST_FIELD_ALIGN): Fix unused variable
7498 warning. Amend comment to reflect current functionality.
7499
409b6ac1
RB
75002014-01-20 Richard Biener <rguenther@suse.de>
7501
7502 PR middle-end/59860
7503 * builtins.c (fold_builtin_strcat): Remove case better handled
7504 by tree-ssa-strlen.c.
7505
608df31f
AL
75062014-01-20 Alan Lawrence <alan.lawrence@arm.com>
7507
7508 * config/aarch64/aarch64.opt
7509 (mcpu, march, mtune): Make case-insensitive.
7510
0fabe5f3
JJ
75112014-01-20 Jakub Jelinek <jakub@redhat.com>
7512
7513 PR target/59880
7514 * config/i386/i386.c (ix86_avoid_lea_for_addr): Return false
7515 if operands[1] is a REG or ZERO_EXTEND of a REG.
7516
39960d1c 75172014-01-19 Jan Hubicka <hubicka@ucw.cz>
cdafab3d
JH
7518
7519 * varasm.c (compute_reloc_for_constant): Use targetm.binds_local_p.
7520
925cb97d
JDA
75212014-01-19 John David Anglin <danglin@gcc.gnu.org>
7522
7523 * config/pa/pa.c (pa_attr_length_millicode_call): Correct length of
7524 long non-pic millicode calls.
7525
f43856db
JBG
75262014-01-19 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7527
7528 * config/vax/vax.h (FUNCTION_ARG_REGNO_P): Fix unused variable warning.
7529
d2be9965
ILT
75302014-01-19 Kito Cheng <kito@0xlab.org>
7531
7532 * builtins.c (expand_movstr): Check movstr expand done or fail.
7533
efc90043
UB
75342014-01-18 Uros Bizjak <ubizjak@gmail.com>
7535 H.J. Lu <hongjiu.lu@intel.com>
7536
7537 PR target/59379
7538 * config/i386/i386.md (*lea<mode>): Zero-extend return register
7539 to DImode for zero-extended addresses.
7540
8fce217e
JJ
75412014-01-19 Jakub Jelinek <jakub@redhat.com>
7542
7543 PR rtl-optimization/57763
7544 * bb-reorder.c (fix_crossing_unconditional_branches): Set JUMP_LABEL
7545 on the new indirect jump_insn and increment LABEL_NUSES (label).
7546
efc90043 75472014-01-18 H.J. Lu <hongjiu.lu@intel.com>
646bdeab
L
7548
7549 PR bootstrap/59580
7550 PR bootstrap/59583
7551 * config.gcc (x86_archs): New variable.
7552 (x86_64_archs): Likewise.
7553 (x86_cpus): Likewise.
7554 Use $x86_archs, $x86_64_archs and $x86_cpus to check valid
7555 --with-arch/--with-cpu= options.
7556 Support --with-arch=/--with-cpu={nehalem,westmere,
7557 sandybridge,ivybridge,haswell,broadwell,bonnell,silvermont}.
7558
7f3af6d3
UB
75592014-01-18 Uros Bizjak <ubizjak@gmail.com>
7560
7561 * config/i386/i386.c (ix86_adjust_cost): Reorder PROCESSOR_K8
7562 and PROCESSOR_ATHLON to simplify code. Move "memory" calculation.
7563
75642014-01-18 Uros Bizjak <ubizjak@gmail.com>
7565
7566 * config/i386/i386.md (*swap<mode>): Rename from swap<mode>.
7567
fa5d6c75
JJ
75682014-01-18 Jakub Jelinek <jakub@redhat.com>
7569
7570 PR target/58944
7571 * config/i386/i386-c.c (ix86_pragma_target_parse): Temporarily
7572 clear cpp_get_options (parse_in)->warn_unused_macros for
7573 ix86_target_macros_internal with cpp_define.
7574
04da5680
RS
75752014-01-18 Richard Sandiford <rdsandiford@googlemail.com>
7576
7577 * jump.c (delete_related_insns): Keep (use (insn))s.
7578 * reorg.c (redundant_insn): Check for barriers too.
7579
1e99bee5
L
75802014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7581
7f3af6d3 7582 * config/i386/i386.c (ix86_split_lea_for_addr): Fix a comment typo.
1e99bee5 7583
fa7d0c60
JDA
75842014-01-17 John David Anglin <danglin@gcc.gnu.org>
7585
7586 * config/pa/pa.c (pa_attr_length_indirect_call): Don't output a short
7587 call to $$dyncall when TARGET_LONG_CALLS is true.
7588
2043135a
JL
75892014-01-17 Jeff Law <law@redhat.com>
7590
7591 * ree.c (combine_set_extension): Temporarily disable test for
7592 changing number of hard registers.
7593
39960d1c 75942014-01-17 Jan Hubicka <hubicka@ucw.cz>
f3e11e05
JH
7595
7596 PR middle-end/58125
7597 * ipa-inline-analysis.c (inline_free_summary):
7598 Do not free summary of aliases.
7599
664ceb1e
JJ
76002014-01-17 Jakub Jelinek <jakub@redhat.com>
7601
7602 PR middle-end/59706
7603 * gimplify.c (gimplify_expr): Use create_tmp_var
7604 instead of create_tmp_var_raw. If cond doesn't have
7605 integral type, don't add the IFN_ANNOTATE builtin at all.
7606
aef83682
MJ
76072014-01-17 Martin Jambor <mjambor@suse.cz>
7608
7609 PR ipa/59736
7610 * ipa-cp.c (prev_edge_clone): New variable.
7611 (grow_next_edge_clone_vector): Renamed to grow_edge_clone_vectors.
7612 Also resize prev_edge_clone vector.
7613 (ipcp_edge_duplication_hook): Also update prev_edge_clone.
7614 (ipcp_edge_removal_hook): New function.
7615 (ipcp_driver): Register ipcp_edge_removal_hook.
7616
4bb66ef3 76172014-01-17 Andrew Pinski <apinski@cavium.com>
3b9c787b
AP
7618 Steve Ellcey <sellcey@mips.com>
7619
7620 PR target/59462
7621 * config/mips/mips.c (mips_print_operand): Check operand mode instead
7622 of operator mode.
7623
4ee5c752
JL
76242014-01-17 Jeff Law <law@redhat.com>
7625
7626 PR middle-end/57904
7627 * passes.def: Reorder pass_copy_prop, pass_unrolli, pass_ccp sequence
7628 so that pass_ccp runs first.
7629
40cfac7c
L
76302014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7631
7632 * config/i386/i386.c (ix86_lea_outperforms): Use TARGET_XXX.
7633 (ix86_adjust_cost): Use !TARGET_XXX.
7634 (do_reorder_for_imul): Likewise.
7635 (swap_top_of_ready_list): Likewise.
7636 (ix86_sched_reorder): Likewise.
7637
9a7f94d7
L
76382014-01-17 H.J. Lu <hongjiu.lu@intel.com>
7639
7640 * config/i386/i386-c.c (ix86_target_macros_internal): Handle
7641 PROCESSOR_INTEL. Treat like PROCESSOR_GENERIC.
7642 * config/i386/i386.c (intel_memcpy): New. Duplicate slm_memcpy.
7643 (intel_memset): New. Duplicate slm_memset.
7644 (intel_cost): New. Duplicate slm_cost.
7645 (m_INTEL): New macro.
7646 (processor_target_table): Add "intel".
7647 (ix86_option_override_internal): Replace PROCESSOR_SILVERMONT
7648 with PROCESSOR_INTEL for "intel".
7649 (ix86_lea_outperforms): Support PROCESSOR_INTEL. Duplicate
7650 PROCESSOR_SILVERMONT.
9a7f94d7
L
7651 (ix86_issue_rate): Likewise.
7652 (ix86_adjust_cost): Likewise.
7653 (ia32_multipass_dfa_lookahead): Likewise.
7654 (swap_top_of_ready_list): Likewise.
7655 (ix86_sched_reorder): Likewise.
8cdcf750
L
7656 (ix86_avoid_lea_for_addr): Check TARGET_AVOID_LEA_FOR_ADDR
7657 instead of TARGET_OPT_AGU.
9a7f94d7
L
7658 * config/i386/i386.h (TARGET_INTEL): New.
7659 (TARGET_AVOID_LEA_FOR_ADDR): Likewise.
7660 (processor_type): Add PROCESSOR_INTEL.
7f3af6d3
UB
7661 * config/i386/x86-tune.def: Support m_INTEL. Duplicate m_SILVERMONT.
7662 Add X86_TUNE_AVOID_LEA_FOR_ADDR.
9a7f94d7 7663
30078c0a
MP
76642014-01-17 Marek Polacek <polacek@redhat.com>
7665
7666 PR c/58346
7667 * gimple-fold.c (fold_array_ctor_reference): Don't fold if element
7668 size is zero.
7669
14379e66
RB
76702014-01-17 Richard Biener <rguenther@suse.de>
7671
7672 PR tree-optimization/46590
7673 * opts.c (default_options_table): Add entries for
7674 OPT_fbranch_count_reg, OPT_fmove_loop_invariants and OPT_ftree_pta,
7675 all enabled at -O1 but not for -Og.
7676 * common.opt (fbranch-count-reg): Remove Init(1).
7677 (fmove-loop-invariants): Likewise.
7678 (ftree-pta): Likewise.
7679
f7d594d2
JJ
76802014-01-17 Jakub Jelinek <jakub@redhat.com>
7681
88e18bd5
JJ
7682 * config/i386/i386.c (ix86_data_alignment): For compatibility with
7683 (incorrect) GCC 4.8 and earlier alignment assumptions ensure we align
7684 decls to at least the GCC 4.8 used alignments.
7685
f7d594d2
JJ
7686 PR fortran/59440
7687 * tree-nested.c (convert_nonlocal_reference_stmt,
7688 convert_local_reference_stmt): For NAMELIST_DECLs in gimple_bind_vars
7689 of GIMPLE_BIND stmts, adjust associated decls.
7690
32500433
RB
76912014-01-17 Richard Biener <rguenther@suse.de>
7692
7693 PR tree-optimization/46590
7694 * vec.h (vec<>::bseach): New member function implementing
7695 binary search according to C89 bsearch.
7696 (vec<>::qsort): Avoid calling ::qsort for vectors with sizes 0 or 1.
7697 * tree-ssa-loop-im.c (struct mem_ref): Make stored member a
7698 bitmap pointer again. Make accesses_in_loop a flat array.
7699 (mem_ref_obstack): New global.
7700 (outermost_indep_loop): Adjust for mem_ref->stored changes.
7701 (mark_ref_stored): Likewise.
7702 (ref_indep_loop_p_2): Likewise.
7703 (set_ref_stored_in_loop): New helper function.
7704 (mem_ref_alloc): Allocate mem_refs on the mem_ref_obstack obstack.
7705 (memref_free): Adjust.
7706 (record_mem_ref_loc): Simplify.
7707 (gather_mem_refs_stmt): Adjust.
7708 (sort_locs_in_loop_postorder_cmp): New function.
7709 (analyze_memory_references): Sort accesses_in_loop after
7710 loop postorder number.
7711 (find_ref_loc_in_loop_cmp): New function.
7712 (for_all_locs_in_loop): Find relevant cluster of locs in
7713 accesses_in_loop and iterate without recursion.
7714 (execute_sm): Avoid uninit warning.
7715 (struct ref_always_accessed): Simplify.
7716 (ref_always_accessed::operator ()): Likewise.
7717 (ref_always_accessed_p): Likewise.
7718 (tree_ssa_lim_initialize): Initialize mem_ref_obstack, compute
7719 loop postorder numbers here.
7720 (tree_ssa_lim_finalize): Free mem_ref_obstack and loop postorder
7721 numbers.
7722
24fcf4bc
JJ
77232014-01-17 Jan Hubicka <hubicka@ucw.cz>
7724
7725 PR c++/57945
7726 * passes.c (rest_of_decl_compilation): Don't call varpool_finalize_decl
7727 on decls for which assemble_alias has been called.
7728
d4f283a1
NC
77292014-01-17 Nick Clifton <nickc@redhat.com>
7730
7731 * config/msp430/msp430.opt: (mcpu): New option.
7732 * config/msp430/msp430.c (msp430_mcu_name): Use target_mcu.
7733 (msp430_option_override): Parse target_cpu. If the MCU name
7734 matches a generic string, clear target_mcu.
7735 (msp430_attr): Allow numeric interrupt values up to 63.
7736 (msp430_expand_epilogue): No longer invert operand 1 of gen_popm.
7737 * config/msp430/msp430.h (ASM_SPEC): Convert -mcpu into a -mmcu
7738 option.
7739 * config/msp430/t-msp430: (MULTILIB_MATCHES): Remove mcu matches.
7740 Add mcpu matches.
7741 * config/msp430/msp430.md (popm): Use %J rather than %I.
7742 (addsi3): Use msp430_nonimmediate_operand for operand 2.
7743 (addhi_cy_i): Use immediate_operand for operand 2.
7744 * doc/invoke.texi: Document -mcpu option.
7745
7be64667
RB
77462014-01-17 Richard Biener <rguenther@suse.de>
7747
7748 PR rtl-optimization/38518
7749 * df.h (df_analyze_loop): Declare.
7750 * df-core.c: Include cfgloop.h.
7751 (df_analyze_1): Split out main part of df_analyze.
7752 (df_analyze): Adjust.
7753 (loop_inverted_post_order_compute): New function.
7754 (loop_post_order_compute): Likewise.
7755 (df_analyze_loop): New function avoiding whole-function
7756 postorder computes.
7757 * loop-invariant.c (find_defs): Use df_analyze_loop.
7758 (find_invariants): Adjust.
7759 * loop-iv.c (iv_analysis_loop_init): Use df_analyze_loop.
7760
82a197a2
ZC
77612014-01-17 Zhenqiang Chen <zhenqiang.chen@arm.com>
7762
7763 * config/arm/arm.c (arm_v7m_tune): Set max_insns_skipped to 2.
7764 (thumb2_final_prescan_insn): Set max to MAX_INSN_PER_IT_BLOCK.
7765
c68b3f52
IE
77662014-01-16 Ilya Enkovich <ilya.enkovich@intel.com>
7767
7768 * ipa-ref.c (ipa_remove_stmt_references): Fix references
7769 traversal when removing references.
7770
39960d1c 77712014-01-16 Jan Hubicka <hubicka@ucw.cz>
f910753d
JH
7772
7773 PR ipa/59775
7774 * tree.c (get_binfo_at_offset): Look harder for virtual bases.
7775
9e6f9ad6
BS
77762014-01-16 Bernd Schmidt <bernds@codesourcery.com>
7777
7778 PR middle-end/56791
7779 * reload.c (find_reloads_address_1): Do not use RELOAD_OTHER when
7780 pushing a reload for an autoinc when we had previously reloaded an
7781 inner part of the address.
7782
a611d7cb
JJ
77832014-01-16 Jakub Jelinek <jakub@redhat.com>
7784
d1417442
JJ
7785 * tree-vectorizer.h (struct _loop_vec_info): Add no_data_dependencies
7786 field.
7787 (LOOP_VINFO_NO_DATA_DEPENDENCIES): Define.
7788 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Clear it
7789 when not giving up or versioning for alias only because of
7790 loop->safelen.
7791 (vect_analyze_data_ref_dependences): Set to true.
7792 * tree-vect-stmts.c (hoist_defs_of_uses): Return false if def_stmt
7793 is a GIMPLE_PHI.
7794 (vectorizable_load): Use LOOP_VINFO_NO_DATA_DEPENDENCIES instead of
7795 LOOP_REQUIRES_VERSIONING_FOR_ALIAS, add && !nested_in_vect_loop
7796 to the condition.
7797
42ed6cde
JJ
7798 PR middle-end/58344
7799 * expr.c (expand_expr_real_1): Handle init == NULL_TREE.
7800
a611d7cb 7801 PR target/59839
7f3af6d3
UB
7802 * config/i386/i386.c (ix86_expand_builtin): If target doesn't satisfy
7803 operand 0 predicate for gathers, use a new pseudo as subtarget.
a611d7cb 7804
8f21260c
VM
78052014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7806
7807 PR middle-end/59609
7f3af6d3
UB
7808 * lra-constraints.c (process_alt_operands): Add printing debug info.
7809 Check absence of input/output reloads for matched operands too.
8f21260c 7810
1a788c05
VM
78112014-01-16 Vladimir Makarov <vmakarov@redhat.com>
7812
7813 PR rtl-optimization/59835
7814 * ira.c (ira_init_register_move_cost): Increase cost for
7815 impossible modes.
7816
4cf24d27
AL
78172014-01-16 Alan Lawrence <alan.lawrence@arm.com>
7818
cc3a9f0d 7819 * config/arm/arm.opt (mcpu, march, mtune): Make case-insensitive.
4cf24d27 7820
030d03b8
RE
78212014-01-16 Richard Earnshaw <rearnsha@arm.com>
7822
7823 PR target/59780
7824 * aarch64.c (aarch64_split_128bit_move): Don't lookup REGNO on
7825 non-register objects. Use gen_(high/low)part more consistently.
7826 Fix assertions.
7827
e78f06a8
MM
78282014-01-16 Michael Meissner <meissner@linux.vnet.ibm.com>
7829
7830 PR target/59844
7831 * config/rs6000/rs6000.md (reload_vsx_from_gprsf): Add little
7832 endian support, remove tests for WORDS_BIG_ENDIAN.
7833 (p8_mfvsrd_3_<mode>): Likewise.
7834 (reload_gpr_from_vsx<mode>): Likewise.
7835 (reload_gpr_from_vsxsf): Likewise.
7836 (p8_mfvsrd_4_disf): Likewise.
7837
9d1ae52c
RB
78382014-01-16 Richard Biener <rguenther@suse.de>
7839
7840 PR rtl-optimization/46590
7841 * lcm.c (compute_antinout_edge): Use postorder iteration.
7842 (compute_laterin): Use inverted postorder iteration.
7843
54c7a7f3
NC
78442014-01-16 Nick Clifton <nickc@redhat.com>
7845
7846 PR middle-end/28865
7847 * varasm.c (output_constant): Return the number of bytes actually
7848 emitted.
7849 (output_constructor_array_range): Update the field size with the
7850 number of bytes emitted by output_constant.
7851 (output_constructor_regular_field): Likewise. Also do not
7852 complain if the total number of bytes emitted is now greater
7853 than the expected fieldpos.
7f3af6d3 7854 * output.h (output_constant): Update prototype and descriptive comment.
54c7a7f3 7855
5147d10a
MP
78562014-01-16 Marek Polacek <polacek@redhat.com>
7857
7858 PR middle-end/59827
7859 * cgraph.c (gimple_check_call_args): Don't use DECL_ARG_TYPE if
7860 it is error_mark_node.
7861
4ac005ba
UB
78622014-01-15 Uros Bizjak <ubizjak@gmail.com>
7863
7864 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use
7865 VALID_AVX256_REG_OR_OI_MODE.
7866
5d7574fa
PH
78672014-01-15 Pat Haugen <pthaugen@us.ibm.com>
7868
7869 * config/rs6000/rs6000.c (rs6000_output_function_prologue): Check if
7870 current procedure should be profiled.
7871
6ee70f81
AP
78722014-01-15 Andrew Pinski <apinski@cavium.com>
7873
7874 * config/aarch64/aarch64.c (aarch64_register_move_cost): Correct cost
7875 of moving from/to the STACK_REG register class.
7876
004a7e45 78772014-01-15 Richard Henderson <rth@redhat.com>
af6e8467
RH
7878
7879 PR debug/54694
7880 * reginfo.c (global_regs_decl): Globalize.
7881 * rtl.h (global_regs_decl): Declare.
7882 * ira.c (do_reload): Diagnose frame_pointer_needed and it
7883 reserved via global_regs.
7884
4583fada
TJ
78852014-01-15 Teresa Johnson <tejohnson@google.com>
7886
7887 * tree-ssa-sccvn.c (visit_reference_op_call): Handle NULL vdef.
7888
d13dfec8
BS
78892014-01-15 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
7890
7891 * config/rs6000/altivec.md (mulv8hi3): Explicitly generate vmulesh
7892 and vmulosh rather than call gen_vec_widen_smult_*.
7893 (vec_widen_umult_even_v16qi): Test VECTOR_ELT_ORDER_BIG rather
7894 than BYTES_BIG_ENDIAN to determine use of even or odd instruction.
7895 (vec_widen_smult_even_v16qi): Likewise.
7896 (vec_widen_umult_even_v8hi): Likewise.
7897 (vec_widen_smult_even_v8hi): Likewise.
7898 (vec_widen_umult_odd_v16qi): Likewise.
7899 (vec_widen_smult_odd_v16qi): Likewise.
7900 (vec_widen_umult_odd_v8hi): Likewise.
7901 (vec_widen_smult_odd_v8hi): Likewise.
7902 (vec_widen_umult_hi_v16qi): Explicitly generate vmuleub and
7903 vmuloub rather than call gen_vec_widen_umult_*.
7904 (vec_widen_umult_lo_v16qi): Likewise.
7905 (vec_widen_smult_hi_v16qi): Explicitly generate vmulesb and
7906 vmulosb rather than call gen_vec_widen_smult_*.
7907 (vec_widen_smult_lo_v16qi): Likewise.
7908 (vec_widen_umult_hi_v8hi): Explicitly generate vmuleuh and vmulouh
7909 rather than call gen_vec_widen_umult_*.
7910 (vec_widen_umult_lo_v8hi): Likewise.
7911 (vec_widen_smult_hi_v8hi): Explicitly gnerate vmulesh and vmulosh
7912 rather than call gen_vec_widen_smult_*.
7913 (vec_widen_smult_lo_v8hi): Likewise.
7914
a6a2d67b
JL
79152014-01-15 Jeff Law <law@redhat.com>
7916
7917 PR tree-optimization/59747
7918 * ree.c (find_and_remove_re): Properly handle case where a second
7919 eliminated extension requires widening a copy created for elimination
7920 of a prior extension.
7921 (combine_set_extension): Ensure that the number of hard regs needed
7922 for a destination register does not change when we widen it.
7923
aefe4056
SH
79242014-01-15 Sebastian Huber <sebastian.huber@embedded-brains.de>
7925
7926 * config.gcc (*-*-rtems*): Add t-rtems to tmake_file.
7927 (arm*-*-uclinux*eabi*): Do not override an existing tmake_file.
7928 (arm*-*-eabi* | arm*-*-symbianelf* | arm*-*-rtems*): Likwise.
7929 (arm*-*-rtems*): Use t-rtems from existing tmake_file.
7930 (avr-*-rtems*): Likewise.
7931 (bfin*-rtems*): Likewise.
7932 (moxie-*-rtems*): Likewise.
7933 (h8300-*-rtems*): Likewise.
7934 (i[34567]86-*-rtems*): Likewise.
7935 (lm32-*-rtems*): Likewise.
7936 (m32r-*-rtems*): Likewise.
7937 (m68k-*-rtems*): Likewise.
7938 (microblaze*-*-rtems*): Likewise.
7939 (mips*-*-rtems*): Likewise.
7940 (powerpc-*-rtems*): Likewise.
7941 (sh-*-rtems*): Likewise.
7942 (sparc-*-rtems*): Likewise.
7943 (sparc64-*-rtems*): Likewise.
7944 (v850-*-rtems*): Likewise.
7945 (m32c-*-rtems*): Likewise.
7946
fef37404
VM
79472014-01-15 Vladimir Makarov <vmakarov@redhat.com>
7948
7949 PR rtl-optimization/59511
7950 * ira.c (ira_init_register_move_cost): Use memory costs for some
7951 cases of register move cost calculations.
7952 * lra-constraints.c (lra_constraints): Use REG_FREQ_FROM_BB
7953 instead of BB frequency.
7954 * lra-coalesce.c (move_freq_compare_func, lra_coalesce): Ditto.
7955 * lra-assigns.c (find_hard_regno_for): Ditto.
7956
6b916b36
RB
79572014-01-15 Richard Biener <rguenther@suse.de>
7958
7959 PR tree-optimization/59822
7960 * tree-vect-stmts.c (hoist_defs_of_uses): New function.
7961 (vectorizable_load): Use it to hoist defs of uses of invariant
7962 loads out of the loop.
7963
d103f29b 79642014-01-15 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
1287ae50 7965 Kugan Vivekanandarajah <kuganv@linaro.org>
d103f29b
MGD
7966
7967 PR target/59695
7968 * config/aarch64/aarch64.c (aarch64_build_constant): Fix incorrect
7969 truncation.
7970
95cb8697
RB
79712014-01-15 Richard Biener <rguenther@suse.de>
7972
7973 PR rtl-optimization/59802
7974 * lcm.c (compute_available): Use inverted postorder to seed
7975 the initial worklist.
7976
cb4b6d17
AK
79772014-01-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7978
1c8b5303 7979 PR target/59803
cb4b6d17
AK
7980 * config/s390/s390.c (s390_preferred_reload_class): Don't return
7981 ADDR_REGS for invalid symrefs in non-PIC code.
7982
2738b4c7
JJ
79832014-01-15 Jakub Jelinek <jakub@redhat.com>
7984
7985 PR other/58712
7986 * builtins.c (determine_block_size): Initialize *probable_max_size
7987 even if len_rtx is CONST_INT.
7988
d126a4ae
AP
79892014-01-14 Andrew Pinski <apinski@cavium.com>
7990
7991 * config/aarch64/aarch64-protos.h (tune_params): Add issue_rate.
7992 * config/aarch64/aarch64.c (generic_tunings): Add issue rate of 2.
7993 (cortexa53_tunings): Likewise.
7994 (aarch64_sched_issue_rate): New function.
7995 (TARGET_SCHED_ISSUE_RATE): Define.
7996
dc687582
VM
79972014-01-14 Vladimir Makarov <vmakarov@redhat.com>
7998
7999 * ira-costs.c (find_costs_and_classes): Add missed
8000 ira_init_register_move_cost_if_necessary.
8001
e940b2ec
VM
80022014-01-14 Vladimir Makarov <vmakarov@redhat.com>
8003
8004 PR target/59787
8005 * config/arm/arm.c (arm_coproc_mem_operand): Add lra_in_progress.
8006
3292e376
L
80072014-01-14 H.J. Lu <hongjiu.lu@intel.com>
8008
8009 PR target/59794
8010 * config/i386/i386.c (type_natural_mode): Add a bool parameter
4ac005ba
UB
8011 to indicate if type is used for function return value. Warn ABI
8012 change if the vector mode isn't available for function return value.
3292e376
L
8013 (ix86_function_arg_advance): Pass false to type_natural_mode.
8014 (ix86_function_arg): Likewise.
8015 (ix86_gimplify_va_arg): Likewise.
8016 (function_arg_32): Don't warn ABI change.
8017 (ix86_function_value): Pass true to type_natural_mode.
8018 (ix86_return_in_memory): Likewise.
8019 (ix86_struct_value_rtx): Removed.
8020 (TARGET_STRUCT_VALUE_RTX): Likewise.
8021
db930875
RS
80222014-01-14 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
8023
8024 * jump.c (redirect_jump_2): Remove REG_CROSSING_JUMP notes when
8025 converting a conditional jump into a conditional return.
8026
a0e35eb0
RB
80272014-01-14 Richard Biener <rguenther@suse.de>
8028
8029 PR tree-optimization/58921
8030 PR tree-optimization/59006
8031 * tree-vect-loop-manip.c (vect_loop_versioning): Remove code
8032 hoisting invariant stmts.
8033 * tree-vect-stmts.c (vectorizable_load): Insert the splat of
8034 invariant loads on the preheader edge if possible.
8035
a984e92e
JY
80362014-01-14 Joey Ye <joey.ye@arm.com>
8037
8038 * doc/plugin.texi (Building GCC plugins): Update to C++.
8039
c56a42b9
KY
80402014-01-14 Kirill Yukhin <kirill.yukhin@intel.com>
8041
3292e376 8042 * config/i386/avx512erintrin.h (_mm_rcp28_round_sd): New.
c56a42b9
KY
8043 (_mm_rcp28_round_ss): Ditto.
8044 (_mm_rsqrt28_round_sd): Ditto.
8045 (_mm_rsqrt28_round_ss): Ditto.
8046 (_mm_rcp28_sd): Ditto.
8047 (_mm_rcp28_ss): Ditto.
8048 (_mm_rsqrt28_sd): Ditto.
8049 (_mm_rsqrt28_ss): Ditto.
8050 * config/i386/avx512fintrin.h (_mm512_stream_load_si512): Ditto.
8051 * config/i386/i386-builtin-types.def (V8DI_FTYPE_PV8DI): Ditto.
8052 * config/i386/i386.c (IX86_BUILTIN_MOVNTDQA512): Ditto.
8053 (IX86_BUILTIN_RCP28SD): Ditto.
8054 (IX86_BUILTIN_RCP28SS): Ditto.
8055 (IX86_BUILTIN_RSQRT28SD): Ditto.
8056 (IX86_BUILTIN_RSQRT28SS): Ditto.
8057 (bdesc_special_args): Define __builtin_ia32_movntdqa512,
8058 __builtin_ia32_rcp28sd_round, __builtin_ia32_rcp28ss_round,
8059 __builtin_ia32_rsqrt28sd_round, __builtin_ia32_rsqrt28ss_round.
8060 (ix86_expand_special_args_builtin): Expand new FTYPE.
8061 * config/i386/sse.md (define_mode_attr "sse4_1_avx2"): Expand to V8DI.
8062 (srcp14<mode>): Make insn unary.
8063 (avx512f_vmscalef<mode><round_name>): Use substed predicate.
8064 (avx512f_sgetexp<mode><round_saeonly_name>): Ditto.
8065 (avx512f_rndscale<mode><round_saeonly_name>): Ditto.
8066 (<sse4_1_avx2>_movntdqa): Extend to 512 bits.
8067 (avx512er_exp2<mode><mask_name><round_saeonly_name>):
8068 Fix rounding: make it SAE only.
4ac005ba
UB
8069 (<mask_codefor>avx512er_rcp28<mode><mask_name><round_saeonly_name>):
8070 Ditto.
8071 (<mask_codefor>avx512er_rsqrt28<mode><mask_name><round_saeonly_name>):
8072 Ditto.
c56a42b9
KY
8073 (avx512er_vmrcp28<mode><round_saeonly_name>): Ditto.
8074 (avx512er_vmrsqrt28<mode><round_saeonly_name>): Ditto.
8075 (avx512f_getmant<mode><mask_name><round_saeonly_name>): Ditto.
8076 * config/i386/subst.md (round_saeonly_mask_scalar_operand3): Remove.
8077 (round_saeonly_mask_scalar_operand4): Ditto.
8078 (round_saeonly_mask_scalar_op3): Ditto.
8079 (round_saeonly_mask_scalar_op4): Ditto.
8080
09fccb62
BS
80812014-01-13 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8082
8083 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
8084 Implement -maltivec=be for vec_insert and vec_extract.
8085
fb28dac0
DD
80862014-01-10 DJ Delorie <dj@redhat.com>
8087
8088 * config/msp430/msp430.md (call_internal): Don't allow memory
8089 references with SP as the base register.
8090 (call_value_internal): Likewise.
8091 * config/msp430/constraints.md (Yc): New. For memory references
8092 that don't use SP as a base register.
8093
8094 * config/msp430/msp430.c (msp430_print_operand): Add 'J' to mean
8095 "an integer without a # prefix"
8096 * config/msp430/msp430.md (epilogue_helper): Use it.
8097
19e51b40
JJ
80982014-01-13 Jakub Jelinek <jakub@redhat.com>
8099
03b9e8e4
JJ
8100 PR target/59617
8101 * config/i386/i386.c (ix86_vectorize_builtin_gather): Uncomment
8102 AVX512F gather builtins.
8103 * tree-vect-stmts.c (vectorizable_mask_load_store): For now punt
8104 on gather decls with INTEGER_TYPE masktype.
8105 (vectorizable_load): For INTEGER_TYPE masktype, put the INTEGER_CST
8106 directly into the builtin rather than hoisting it before loop.
8107
19e51b40
JJ
8108 PR tree-optimization/59387
8109 * tree-scalar-evolution.c: Include gimple-fold.h and gimplify-me.h.
8110 (scev_const_prop): If folded_casts and type has undefined overflow,
8111 use force_gimple_operand instead of force_gimple_operand_gsi and
8112 for each added stmt if it is assign with
8113 arith_code_with_undefined_signed_overflow, call
8114 rewrite_to_defined_overflow.
8115 * tree-ssa-loop-im.c: Don't include gimplify-me.h, include
8116 gimple-fold.h instead.
8117 (arith_code_with_undefined_signed_overflow,
8118 rewrite_to_defined_overflow): Moved to ...
8119 * gimple-fold.c (arith_code_with_undefined_signed_overflow,
8120 rewrite_to_defined_overflow): ... here. No longer static.
8121 Include gimplify-me.h.
8122 * gimple-fold.h (arith_code_with_undefined_signed_overflow,
8123 rewrite_to_defined_overflow): New prototypes.
8124
cd794ed4
KT
81252014-01-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8126
8127 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Fix typo in description.
8128
e75fde1a
EB
81292014-01-13 Eric Botcazou <ebotcazou@adacore.com>
8130
8131 * builtins.c (get_object_alignment_2): Minor tweak.
8132 * tree-ssa-loop-ivopts.c (may_be_unaligned_p): Rewrite.
8133
eae298d6
CB
81342014-01-13 Christian Bruel <christian.bruel@st.com>
8135
8136 * config/sh/sh-mem.cc (sh_expand_cmpnstr): Unroll small sizes and
e75fde1a 8137 optimized non constant lengths.
eae298d6 8138
8175be9a
JJ
81392014-01-13 Jakub Jelinek <jakub@redhat.com>
8140
8141 PR libgomp/59194
8142 * omp-low.c (expand_omp_atomic_pipeline): Expand the initial
8143 load as __atomic_load_N if possible.
8144
9c111368
DE
81452014-01-11 David Edelsohn <dje.gcc@gmail.com>
8146
8147 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): Remove
8148 target parameter.
8149 (rs6000_expand_builtin): Adjust call.
8150
c82846bc
DE
81512014-01-11 David Edelsohn <dje.gcc@gmail.com>
8152
8153 PR target/58115
8154 * config/rs6000/rs6000.h (SWITCHABLE_TARGET): Define.
8155 * config/rs6000/rs6000.c: Include target-globals.h.
8156 (rs6000_set_current_function): Instead of doing target_reinit
8157 unconditionally, use save_target_globals_default_opts and
8158 restore_target_globals.
8159
8160 * config/rs6000/rs6000-builtin.def (mffs, mtfsf): Add builtins for
8161 FPSCR.
8162 * config/rs6000/rs6000.c (rs6000_expand_mtfsf_builtin): New.
8163 (rs6000_expand_builtin): Handle mffs and mtfsf.
8164 (rs6000_init_builtins): Define mffs and mtfsf.
8165 * config/rs6000/rs6000.md (UNSPECV_MFFS, UNSPECV_MTFSF): New constants.
8166 (rs6000_mffs): New pattern.
8167 (rs6000_mtfsf): New pattern.
8168
2c407426
BC
81692014-01-11 Bin Cheng <bin.cheng@arm.com>
8170
8171 * tree-ssa-loop-ivopts.c (iv_ca_narrow): New parameter.
8172 Start narrowing with START. Apply candidate-use pair
8173 and check overall cost in narrowing.
8174 (iv_ca_prune): Pass new argument.
8175
7e41c852
JL
81762014-01-10 Jeff Law <law@redhat.com>
8177
8178 PR middle-end/59743
8179 * ree.c (combine_reaching_defs): Ensure the defining statement
8180 occurs before the extension when optimizing extensions with
8181 different source and destination hard registers.
8182
39960d1c 81832014-01-10 Jan Hubicka <hubicka@ucw.cz>
b270b096
JH
8184
8185 PR ipa/58585
4ac005ba
UB
8186 * ipa-devirt.c (build_type_inheritance_graph): Also add types of
8187 vtables into the type inheritance graph.
b270b096 8188
0d6d7b9a
JJ
81892014-01-10 Jakub Jelinek <jakub@redhat.com>
8190
8191 PR rtl-optimization/59754
8192 * ree.c (combine_reaching_defs): Disallow !SCALAR_INT_MODE_P
8193 modes in the REGNO != REGNO case.
8194
110132c1
BS
81952014-01-10 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8196
8197 * config/rs6000/rs6000-builtin.def: Fix pasto for VPKSDUS.
8198
5da96960
JJ
81992014-01-10 Jakub Jelinek <jakub@redhat.com>
8200
4f87d581
JJ
8201 PR tree-optimization/59745
8202 * tree-predcom.c (tree_predictive_commoning_loop): Call
8203 free_affine_expand_cache if giving up because components is NULL.
8204
5da96960
JJ
8205 * target-globals.c (save_target_globals): Allocate < 4KB structs using
8206 GC in payload of target_globals struct instead of allocating them on
8207 the heap and the larger structs separately using GC.
8208 * target-globals.h (struct target_globals): Make regs, hard_regs,
8209 reload, expmed, ira, ira_int and lra_fields GTY((atomic)) instead
8210 of GTY((skip)) and change type to void *.
8211 (reset_target_globals): Cast loads from those fields to corresponding
8212 types.
8213
1aa26aac
SE
82142014-01-10 Steve Ellcey <sellcey@mips.com>
8215
8216 PR plugins/59335
8217 * Makefile.in (PLUGIN_HEADERS): Add gimplify.h, gimple-iterator.h,
8218 gimple-ssa.h, fold-const.h, tree-cfg.h, tree-into-ssa.h,
8219 tree-ssanames.h, print-tree.h, varasm.h, and context.h.
8220
1c992d1e
RE
82212014-01-10 Richard Earnshaw <rearnsha@arm.com>
8222
7783a246 8223 PR target/59744
1c992d1e
RE
8224 * aarch64-modes.def (CC_Zmode): New flags mode.
8225 * aarch64.c (aarch64_select_cc_mode): Only allow NEG when the condition
8226 represents an equality.
7783a246 8227 (aarch64_get_condition_code): Handle CC_Zmode.
1c992d1e
RE
8228 * aarch64.md (compare_neg<mode>): Restrict to equality operations.
8229
86464cbd
AK
82302014-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8231
8232 * config/s390/s390.c (s390_expand_tbegin): Remove jump over CC
8233 extraction in good case.
8234
5e6667b2
RB
82352014-01-10 Richard Biener <rguenther@suse.de>
8236
8237 PR tree-optimization/59374
8238 * tree-vect-slp.c (vect_slp_analyze_bb_1): Move dependence
8239 checking after SLP discovery. Mark stmts not participating
8240 in any SLP instance properly.
8241
5619162c
KT
82422014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8243
8244 * config/arm/arm.c (arm_new_rtx_costs): Use destination mode
8245 when handling a SET rtx.
8246
2d17b99f
KT
82472014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8248
8249 * config/arm/arm-cores.def (cortex-a53): Specify FL_CRC32.
8250 (cortex-a57): Likewise.
8251 (cortex-a57.cortex-a53): Likewise. Remove redundant flags.
8252
bcca645c
KT
82532014-01-10 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8254
8255 * config/arm/arm.c (arm_init_iwmmxt_builtins): Skip
8256 non-iwmmxt builtins.
8257
a3788dde
JH
82582014-01-10 Jan Hubicka <hubicka@ucw.cz>
8259
8260 PR ipa/58252
8261 PR ipa/59226
8262 * ipa-devirt.c record_target_from_binfo): Take as argument
8263 stack of binfos and lookup matching one for virtual inheritance.
8264 (possible_polymorphic_call_targets_1): Update.
8265
87ed883e
HC
82662014-01-10 Huacai Chen <chenhc@lemote.com>
8267
8268 * config/mips/driver-native.c (host_detect_local_cpu): Handle new
8269 kernel strings for Loongson-2E/2F/3A.
8270
cd4447e2
JJ
82712014-01-10 Jakub Jelinek <jakub@redhat.com>
8272
8273 PR middle-end/59670
8274 * tree-vect-data-refs.c (vect_analyze_data_refs): Check
8275 is_gimple_call before calling gimple_call_internal_p.
8276
3fadf78a
SE
82772014-01-09 Steve Ellcey <sellcey@mips.com>
8278
8279 * Makefile.in (TREE_FLOW_H): Remove.
8280 (TREE_SSA_H): Add file names from tree-flow.h.
8281 * doc/tree-ssa.texi (Annotations): Remove reference to tree-flow.h
8282 * tree.h: Remove tree-flow.h reference.
8283 * hash-table.h: Remove tree-flow.h reference.
8284 * tree-ssa-loop-niter.c (dump_affine_iv): Replace tree-flow.h
8285 reference with tree-ssa-loop.h.
8286
6edc217d
BS
82872014-01-09 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8288
8289 * doc/invoke.texi: Add -maltivec={be,le} options, and document
8290 default element-order behavior for -maltivec.
8291 * config/rs6000/rs6000.opt: Add -maltivec={be,le} options.
8292 * config/rs6000/rs6000.c (rs6000_option_override_internal): Ensure
8293 that -maltivec={le,be} implies -maltivec; disallow -maltivec=le
8294 when targeting big endian, at least for now.
8295 * config/rs6000/rs6000.h: Add #define of VECTOR_ELT_ORDER_BIG.
8296
a70e9985
JJ
82972014-01-09 Jakub Jelinek <jakub@redhat.com>
8298
3396aba5
JJ
8299 PR middle-end/47735
8300 * cfgexpand.c (expand_one_var): For SSA_NAMEs, if the underlying
8301 var satisfies use_register_for_decl, just take into account type
8302 alignment, rather than decl alignment.
8303
a70e9985
JJ
8304 PR tree-optimization/59622
8305 * gimple-fold.c (gimple_fold_call): Fix a typo in message. For
8306 __builtin_unreachable replace the OBJ_TYPE_REF call with a call to
8307 __builtin_unreachable and add if needed a setter of the lhs SSA_NAME.
8308 Don't devirtualize for inplace at all. For targets.length () == 1,
8309 if the call is noreturn and cfun isn't in SSA form yet, clear lhs.
8310
6e9cc565
L
83112014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8312
8313 * config/i386/i386.md (cpu): Remove the unused btver1.
8314
d1c0e4ac
L
83152014-01-09 H.J. Lu <hongjiu.lu@intel.com>
8316
8317 * gdbasan.in: Put a breakpoint on __sanitizer::Report.
8318
e83b8e2e
JJ
83192014-01-09 Jakub Jelinek <jakub@redhat.com>
8320
8321 PR target/58115
8322 * tree-core.h (struct target_globals): New forward declaration.
8323 (struct tree_target_option): Add globals field.
8324 * tree.h (TREE_TARGET_GLOBALS): Define.
8325 (prepare_target_option_nodes_for_pch): New prototype.
8326 * target-globals.h (struct target_globals): Define even if
8327 !SWITCHABLE_TARGET.
8328 * tree.c (prepare_target_option_node_for_pch,
8329 prepare_target_option_nodes_for_pch): New functions.
8330 * config/i386/i386.h (SWITCHABLE_TARGET): Define.
8331 * config/i386/i386.c: Include target-globals.h.
8332 (ix86_set_current_function): Instead of doing target_reinit
8333 unconditionally, use save_target_globals_default_opts and
8334 restore_target_globals.
8335
2aaed0f3
RB
83362014-01-09 Richard Biener <rguenther@suse.de>
8337
8338 PR tree-optimization/59715
8339 * tree-cfg.h (split_critical_edges): Declare.
8340 * tree-cfg.c (split_critical_edges): Export.
8341 * tree-ssa-sink.c (execute_sink_code): Split critical edges.
8342
b5ebc991
MO
83432014-01-09 Max Ostapenko <m.ostapenko@partner.samsung.com>
8344
4ac005ba 8345 * cfgexpand.c (expand_stack_vars): Optionally disable
b5ebc991
MO
8346 asan stack protection.
8347 (expand_used_vars): Likewise.
8348 (partition_stack_vars): Likewise.
4ac005ba 8349 * asan.c (asan_emit_stack_protection): Optionally disable
b5ebc991 8350 after return stack usage.
7f3af6d3 8351 (instrument_derefs): Optionally disable memory access instrumentation.
b5ebc991
MO
8352 (instrument_builtin_call): Likewise.
8353 (instrument_strlen_call): Likewise.
7f3af6d3 8354 (asan_protect_global): Optionally disable global variables protection.
b5ebc991
MO
8355 * doc/invoke.texi: Added doc for new options.
8356 * params.def: Added new options.
8357 * params.h: Likewise.
8358
b59e0455
JJ
83592014-01-09 Jakub Jelinek <jakub@redhat.com>
8360
8361 PR rtl-optimization/59724
8362 * ifcvt.c (cond_exec_process_if_block): Don't call
8363 flow_find_head_matching_sequence with 0 longest_match.
8364 * cfgcleanup.c (flow_find_head_matching_sequence): Count even
8365 non-active insns if !stop_after.
8366 (try_head_merge_bb): Revert 2014-01-07 changes.
8367
650c4c85
JL
83682014-01-08 Jeff Law <law@redhat.com>
8369
8370 * ree.c (get_sub_rtx): New function, extracted from...
8371 (merge_def_and_ext): Here.
8372 (combine_reaching_defs): Use get_sub_rtx.
8373
ff36fcbe
EB
83742014-01-08 Eric Botcazou <ebotcazou@adacore.com>
8375
8376 * cgraph.h (varpool_variable_node): Do not choke on null node.
8377
9f9d82aa
CM
83782014-01-08 Catherine Moore <clm@codesourcery.com>
8379
4ac005ba
UB
8380 * config/mips/mips.md (simple_return): Attempt to use JRC
8381 for microMIPS.
9f9d82aa
CM
8382 * config/mips/mips.h (MIPS_CALL): Attempt to use JALS for microMIPS.
8383
df6c8808
RS
83842014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8385
8386 PR rtl-optimization/59137
8387 * reorg.c (steal_delay_list_from_target): Call update_block for
8388 elided insns.
8389 (steal_delay_list_from_fallthrough, relax_delay_slots): Likewise.
8390
6170ceff
BS
83912014-01-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8392
8393 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Remove
8394 two duplicate entries.
8395
3f140f32
RS
83962014-01-08 Richard Sandiford <rdsandiford@googlemail.com>
8397
8398 Revert:
8399 2012-10-07 Richard Sandiford <rdsandiford@googlemail.com>
8400
8401 * config/mips/mips.c (mips_truncated_op_cost): New function.
8402 (mips_rtx_costs): Adjust test for BADDU.
8403 * config/mips/mips.md (*baddu_di<mode>): Push truncates to operands.
8404
8405 2012-10-02 Richard Sandiford <rdsandiford@googlemail.com>
8406
8407 * config/mips/mips.md (*baddu_si_eb, *baddu_si_el): Merge into...
8408 (*baddu_si): ...this new pattern.
8409
c6de6665
JJ
84102014-01-08 Jakub Jelinek <jakub@redhat.com>
8411
8412 PR ipa/59722
8413 * ipa-prop.c (ipa_analyze_params_uses): Ignore uses in debug stmts.
8414
4c437f02
BE
84152014-01-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8416
8417 PR middle-end/57748
8418 * expr.h (expand_expr_real, expand_expr_real_1): Add new parameter
8419 inner_reference_p.
8420 (expand_expr, expand_normal): Adjust.
8421 * expr.c (expand_expr_real, expand_expr_real_1): Add new parameter
8422 inner_reference_p. Use inner_reference_p to expand inner references.
8423 (store_expr): Adjust.
8424 * cfgexpand.c (expand_call_stmt): Adjust.
8425
4ac005ba 84262014-01-08 Rong Xu <xur@google.com>
40d6b753
RX
8427
8428 * gcov-io.c (gcov_var): Move from gcov-io.h.
8429 (gcov_position): Ditto.
8430 (gcov_is_error): Ditto.
8431 (gcov_rewrite): Ditto.
8432 * gcov-io.h: Refactor. Move gcov_var to gcov-io.h, and libgcov
8433 only part to libgcc/libgcov.h.
8434
ab04b46e
MP
84352014-01-08 Marek Polacek <polacek@redhat.com>
8436
8437 PR middle-end/59669
8438 * omp-low.c (simd_clone_adjust): Don't crash if def is NULL.
8439
06636b32
MP
84402014-01-08 Marek Polacek <polacek@redhat.com>
8441
8442 PR sanitizer/59667
8443 * ubsan.c (ubsan_type_descriptor): Call strip_array_types on type2.
8444
c15677b6
JJ
84452014-01-08 Jakub Jelinek <jakub@redhat.com>
8446
8447 PR rtl-optimization/59649
8448 * stor-layout.c (get_mode_bounds): For BImode return
8449 0 and STORE_FLAG_VALUE.
8450
5c944c6c
RB
84512014-01-08 Richard Biener <rguenther@suse.de>
8452
8453 PR middle-end/59630
8454 * gimple.h (is_gimple_builtin_call): Remove.
8455 (gimple_builtin_call_types_compatible_p): New.
8456 (gimple_call_builtin_p): New overload.
8457 * gimple.c (is_gimple_builtin_call): Remove.
8458 (validate_call): Rename to ...
8459 (gimple_builtin_call_types_compatible_p): ... this and export. Also
8460 check return types.
8461 (validate_type): New static function.
8462 (gimple_call_builtin_p): New overload and adjust.
8463 * gimple-fold.c (gimple_fold_builtin): Fold the return value.
8464 (gimple_fold_call): Likewise. Use gimple_call_builtin_p.
8465 (gimple_fold_stmt_to_constant_1): Likewise.
8466 * tsan.c (instrument_gimple): Use gimple_call_builtin_p.
8467
0bd34ae4
RB
84682014-01-08 Richard Biener <rguenther@suse.de>
8469
8470 PR middle-end/59471
8471 * gimplify.c (gimplify_expr): Gimplify register-register type
8472 VIEW_CONVERT_EXPRs to separate stmts.
8473
04af8ab6
JL
84742014-01-07 Jeff Law <law@redhat.com>
8475
3c92da90
JL
8476 PR middle-end/53623
8477 * ree.c (combine_set_extension): Handle case where source
8478 and destination registers in an extension insn are different.
4ac005ba
UB
8479 (combine_reaching_defs): Allow source and destination registers
8480 in extension to be different under limited circumstances.
3c92da90 8481 (add_removable_extension): Remove restriction that the
4ac005ba 8482 source and destination registers in the extension are the same.
3c92da90
JL
8483 (find_and_remove_re): Emit a copy from the extension's
8484 destination to its source after the defining insn if
8485 the source and destination registers are different.
8486
04af8ab6
JL
8487 PR middle-end/59285
8488 * ifcvt.c (merge_if_block): If we are merging a block with more than
8489 one successor with a block with no successors, remove any BARRIER
8490 after the second block.
8491
4bb66ef3 84922014-01-07 Dan Xio Qiang <ziyan01@163.com>
955b33ed
JL
8493
8494 * hw-doloop.c (reorg_loops): Release the bitmap obstack.
8495
55ada375
JDA
84962014-01-07 John David Anglin <danglin@gcc.gnu.org>
8497
8498 PR target/59652
8499 * config/pa/pa.c (pa_legitimate_address_p): Return false before reload
8500 for 14-bit register offsets when INT14_OK_STRICT is false.
8501
877c72e7
RS
85022014-01-07 Roland Stigge <stigge@antcom.de>
8503 Michael Meissner <meissner@linux.vnet.ibm.com>
8504
8505 PR 57386/target
8506 * config/rs6000/rs6000.c (rs6000_legitimate_offset_address_p):
4ac005ba 8507 Only check TFmode for SPE constants. Don't check TImode or TDmode.
877c72e7 8508
8b2721da
JG
85092014-01-07 James Greenhalgh <james.greenhalgh@arm.com>
8510
8511 * config/aarch64/aarch64-elf.h (ASM_SPEC): Remove identity spec for
8512 -mcpu.
8513
5d72b79f
YZ
85142014-01-07 Yufeng Zhang <yufeng.zhang@arm.com>
8515
8516 * config/arm/arm.c (arm_expand_neon_args): Call expand_expr
8517 with EXPAND_MEMORY for NEON_ARG_MEMORY; check if the returned
8518 rtx is const0_rtx or not.
8519
48d53439
RS
85202014-01-07 Richard Sandiford <rdsandiford@googlemail.com>
8521
8522 PR target/58115
8523 * target-globals.c (save_target_globals): Remove this_fn_optab
8524 handling.
8525 * toplev.c: Include optabs.h.
8526 (target_reinit): Temporarily restore the global options if another
8527 set of options are in force.
8528
41626746
JJ
85292014-01-07 Jakub Jelinek <jakub@redhat.com>
8530
a0cbe71e
JJ
8531 PR rtl-optimization/58668
8532 * cfgcleanup.c (flow_find_cross_jump): Don't count
8533 any jumps if dir_p is NULL. Remove p1 variable, use active_insn_p
8534 to determine what is counted.
8535 (flow_find_head_matching_sequence): Use active_insn_p to determine
8536 what is counted.
8537 (try_head_merge_bb): Adjust for the flow_find_head_matching_sequence
8538 counting change.
8539 * ifcvt.c (count_bb_insns): Use active_insn_p && !JUMP_P to
8540 determine what is counted.
8541
41626746
JJ
8542 PR tree-optimization/59643
8543 * tree-predcom.c (split_data_refs_to_components): If one dr is
8544 read and one write, determine_offset fails and the write isn't
8545 in the bad component, just put the read into the bad component.
8546
cc349a39
MS
85472014-01-07 Mike Stump <mikestump@comcast.net>
8548 Jakub Jelinek <jakub@redhat.com>
8549
8550 PR pch/59436
8551 * tree-core.h (struct tree_optimization_option): Change optabs
8552 type from unsigned char * to void *.
8553 * optabs.c (init_tree_optimization_optabs): Adjust
8554 TREE_OPTIMIZATION_OPTABS initialization.
8555
529a6471
JJ
85562014-01-06 Jakub Jelinek <jakub@redhat.com>
8557
8558 PR target/59644
8559 * config/i386/i386.h (struct machine_function): Add
8560 no_drap_save_restore field.
8561 * config/i386/i386.c (ix86_save_reg): Use
8562 !cfun->machine->no_drap_save_restore instead of
8563 crtl->stack_realign_needed.
8564 (ix86_finalize_stack_realign_flags): Don't clear drap_reg unless
8565 this function clears frame_pointer_needed. Set
8566 cfun->machine->no_drap_save_restore if clearing frame_pointer_needed
8567 and DRAP reg is needed.
8568
96066ce1
MP
85692014-01-06 Marek Polacek <polacek@redhat.com>
8570
8571 PR c/57773
8572 * doc/implement-c.texi: Mention that other integer types are
8573 permitted as bit-field types in strictly conforming mode.
8574
955b33ed 85752014-01-06 Felix Yang <fei.yang0953@gmail.com>
33444996
FY
8576
8577 * modulo-sched.c (schedule_reg_moves): Clear distance1_uses if it
8578 is newly allocated.
8579
328402a9
RE
85802014-01-06 Richard Earnshaw <rearnsha@arm.com>
8581
8582 * aarch64.c (aarch64_rtx_costs): Fix cost calculation for MADD.
8583
a4e33812
MJ
85842014-01-06 Martin Jambor <mjambor@suse.cz>
8585
8586 PR ipa/59008
8587 * ipa-cp.c (ipcp_discover_new_direct_edges): Changed param_index type
8588 to int.
8589 * ipa-prop.c (ipa_print_node_params): Fix indentation.
8590
b5cd2a02
EB
85912014-01-06 Eric Botcazou <ebotcazou@adacore.com>
8592
8593 PR debug/59350
8594 PR debug/59510
8595 * var-tracking.c (add_stores): Preserve the value of the source even if
8596 we don't record the store.
8597
c7d19c0b
TG
85982014-01-06 Terry Guo <terry.guo@arm.com>
8599
8600 * config.gcc (arm*-*-*): Check --with-arch against arm-arches.def.
8601
e0c70bd7
IS
86022014-01-05 Iain Sandoe <iain@codesourcery.com>
8603
8604 PR bootstrap/59541
8605 * config/darwin.c (darwin_function_section): Adjust return values to
8606 correspond to optimisation changes made in r206070.
8607
c30f016c
UB
86082014-01-05 Uros Bizjak <ubizjak@gmail.com>
8609
8610 * config/i386/i386.c (ix86_data_alignment): Calculate max_align
8611 from prefetch_block tune setting.
8612 (nocona_cost): Correct size of prefetch block to 64.
8613
f2bc252b
EB
86142014-01-04 Eric Botcazou <ebotcazou@adacore.com>
8615
8616 * config/arm/arm.c (arm_get_frame_offsets): Revamp long lines.
8617 (arm_expand_epilogue_apcs_frame): Take into account the number of bytes
8618 used to save the static chain register in the computation of the offset
8619 from which the FP registers need to be restored.
8620
dad5ed2e
JJ
86212014-01-04 Jakub Jelinek <jakub@redhat.com>
8622
39719c84
JJ
8623 PR tree-optimization/59519
8624 * tree-vect-loop-manip.c (slpeel_update_phi_nodes_for_guard1): Don't
8625 ICE if get_current_def (current_new_name) is already non-NULL, as long
8626 as it is a phi result of some other phi in *new_exit_bb that has
8627 the same argument.
8628
dad5ed2e
JJ
8629 * config/i386/sse.md (avx512f_load<mode>_mask): Emit vmovup{s,d}
8630 or vmovdqu* for misaligned_operand.
8631 (<sse>_loadu<ssemodesuffix><avxsizesuffix><mask_name>,
8632 <sse2_avx_avx512f>_loaddqu<mode><mask_name>): Handle <mask_applied>.
8633 * config/i386/i386.c (ix86_expand_special_args_builtin): Set
8634 aligned_mem for AVX512F masked aligned load and store builtins and for
8635 non-temporal moves.
8636
86372014-01-03 Bingfeng Mei <bmei@broadcom.com>
82d89471
BM
8638
8639 PR tree-optimization/59651
c30f016c 8640 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
82d89471
BM
8641 Address range for negative step should be added by TYPE_SIZE_UNIT.
8642
e864837a
AS
86432014-01-03 Andreas Schwab <schwab@linux-m68k.org>
8644
8645 * config/m68k/m68k.c (handle_move_double): Handle pushes with
7f3af6d3 8646 overlapping registers also for registers other than the stack pointer.
e864837a 8647
56a34230
MP
86482014-01-03 Marek Polacek <polacek@redhat.com>
8649
8650 PR other/59661
8651 * doc/extend.texi: Fix the return value of __builtin_FUNCTION and
8652 __builtin_FILE.
8653
039eee3f
JJ
86542014-01-03 Jakub Jelinek <jakub@redhat.com>
8655
3bcdbd50
JJ
8656 PR target/59625
8657 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Don't consider
8658 asm goto as jump.
8659
039eee3f
JJ
8660 * config/i386/i386.md (MODE_SIZE): New mode attribute.
8661 (push splitter): Use <P:MODE_SIZE> instead of
8662 GET_MODE_SIZE (<P:MODE>mode).
8663 (lea splitter): Use <MODE_SIZE> instead of GET_MODE_SIZE (<MODE>mode).
8664 (mov -1, reg peephole2): Likewise.
8665 * config/i386/sse.md (*mov<mode>_internal,
8666 <sse>_storeu<ssemodesuffix><avxsizesuffix>,
8667 <sse2_avx_avx512f>_storedqu<mode>, <sse>_andnot<mode>3,
8668 *<code><mode>3, *andnot<mode>3<mask_name>,
8669 <mask_codefor><code><mode>3<mask_name>): Likewise.
8670 * config/i386/subst.md (mask_mode512bit_condition,
8671 sd_mask_mode512bit_condition): Likewise.
8672
927734cf
XDL
86732014-01-02 Xinliang David Li <davidxl@google.com>
8674
8675 PR tree-optimization/59303
039eee3f 8676 * tree-ssa-uninit.c (is_use_properly_guarded): Main cleanup.
927734cf
XDL
8677 (dump_predicates): Better output format.
8678 (pred_equal_p): New function.
8679 (is_neq_relop_p): Ditto.
8680 (is_neq_zero_form_p): Ditto.
8681 (pred_expr_equal_p): Ditto.
8682 (pred_neg_p): Ditto.
8683 (simplify_pred): Ditto.
8684 (simplify_preds_2): Ditto.
8685 (simplify_preds_3): Ditto.
8686 (simplify_preds_4): Ditto.
8687 (simplify_preds): Ditto.
8688 (push_pred): Ditto.
8689 (push_to_worklist): Ditto.
8690 (get_pred_info_from_cmp): Ditto.
8691 (is_degenerated_phi): Ditto.
8692 (normalize_one_pred_1): Ditto.
8693 (normalize_one_pred): Ditto.
8694 (normalize_one_pred_chain): Ditto.
8695 (normalize_preds): Ditto.
8696 (normalize_cond_1): Remove function.
8697 (normalize_cond): Ditto.
8698 (is_gcond_subset_of): Ditto.
8699 (is_subset_of_any): Ditto.
8700 (is_or_set_subset_of): Ditto.
8701 (is_and_set_subset_of): Ditto.
8702 (is_norm_cond_subset_of): Ditto.
8703 (pred_chain_length_cmp): Ditto.
8704 (convert_control_dep_chain_into_preds): Type change.
8705 (find_predicates): Ditto.
8706 (find_def_preds): Ditto.
8707 (destroy_predicates_vecs): Ditto.
8708 (find_matching_predicates_in_rest_chains): Ditto.
8709 (use_pred_not_overlap_with_undef_path_pred): Ditto.
8710 (is_pred_expr_subset): Ditto.
8711 (is_pred_chain_subset_of): Ditto.
8712 (is_included_in): Ditto.
8713 (is_superset_of): Ditto.
8714
23a5b65a
RS
87152014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8716
c30f016c 8717 Update copyright years.
23a5b65a 8718
f9030485
RS
87192014-01-02 Richard Sandiford <rdsandiford@googlemail.com>
8720
8721 * common/config/arc/arc-common.c, config/arc/arc-modes.def,
8722 config/arc/arc-protos.h, config/arc/arc.c, config/arc/arc.h,
c30f016c
UB
8723 config/arc/arc.md, config/arc/arc.opt,
8724 config/arm/arm_neon_builtins.def, config/arm/crypto.def,
8725 config/i386/avx512cdintrin.h, config/i386/avx512erintrin.h,
8726 config/i386/avx512fintrin.h, config/i386/avx512pfintrin.h,
8727 config/i386/btver2.md, config/i386/shaintrin.h, config/i386/slm.md,
8728 config/linux-protos.h, config/linux.c, config/winnt-c.c,
8729 diagnostic-color.c, diagnostic-color.h, gimple-ssa-isolate-paths.c,
8730 vtable-verify.c, vtable-verify.h: Use the standard form for the
8731 copyright notice.
f9030485 8732
98db73df
TB
87332014-01-02 Tobias Burnus <burnus@net-b.de>
8734
8735 * gcc.c (process_command): Update copyright notice dates.
8736 * gcov-dump.c: Ditto.
8737 * gcov.c: Ditto.
8738 * doc/cpp.texi: Bump @copying's copyright year.
8739 * doc/cppinternals.texi: Ditto.
8740 * doc/gcc.texi: Ditto.
8741 * doc/gccint.texi: Ditto.
8742 * doc/gcov.texi: Ditto.
8743 * doc/install.texi: Ditto.
8744 * doc/invoke.texi: Ditto.
8745
2898d204 87462014-01-01 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ca77de0
JBG
8747
8748 * config/nios2/nios2.h (BITS_PER_UNIT): Don't define it.
8749
dfebbdc6 87502014-01-01 Jakub Jelinek <jakub@redhat.com>
cf3e5a89 8751
726c3546
JJ
8752 * config/i386/sse.md (*mov<mode>_internal): Guard
8753 EXT_REX_SSE_REGNO_P (REGNO ()) uses with REG_P.
8754
dfebbdc6
JJ
8755 PR rtl-optimization/59647
8756 * cse.c (cse_process_notes_1): Don't substitute negative VOIDmode
8757 new_rtx into UNSIGNED_FLOAT rtxes.
ad41bd84 8758\f
970c3b33 8759Copyright (C) 2014 Free Software Foundation, Inc.
ad41bd84
JM
8760
8761Copying and distribution of this file, with or without modification,
8762are permitted in any medium without royalty provided the copyright
8763notice and this notice are preserved.
This page took 6.821696 seconds and 5 git commands to generate.