]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
Daily bump.
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
403a3fb7
JDA
12013-09-14 John David Anglin <danglin@gcc.gnu.org>
2
3 PR target/58382
4 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5 calls to word_mode.
6
fb5610fb
IS
72013-09-14 Iain Sandoe <iain@codesourcery.com>
8
9 PR target/48094
10 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
11 seen. (darwin_objc1_section): Likewise.
12 (darwin_file_end): Emit Image Info section when required.
13
afdec9bd
JH
142013-09-14 Jan Hubicka <jh@suse.cz>
15
16 * tree-into-ssa.c (gate_into_ssa): New.
17 (pass_data_build_ssa): Use it.
18 * cgraph.h (expand_thunk): Update prototype.
19 * cgraphunit.c (analyze_function): Expand thunks early.
20 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
21 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
22 set lowered flag; do not add new function.
23 (assemble_thunks_and_aliases): Update.
24 * tree-ssa.c (gate_init_datastructures): New gate.
25 (pass_data_init_datastructures): Use it.
26
492d1e0a
IS
272013-09-14 Iain Sandoe <iain@codesourcery.com>
28
29 PR target/58269
30 config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
31 xmm register set described in the psABI.
32
d13c6b70
EG
332013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
34
35 * dwarf2out.c (should_emit_struct_debug): Add check
36 for type_decl variable is not NULL.
37
88b0e79e
JC
382013-09-13 Jacek Caban <jacek@codeweavers.com>
39
40 * config.gcc: Use new winnt-c.c target hooks
41 * config/t-winnt: New file
42 * config/winnt-c.c: New file
43 * doc/tm.texi.in: Document new hook
44 * doc/tm.texi: Regenerated
45
19ba6aab
JH
462013-09-13 Jan Hubicka <jh@suse.cz>
47
48 PR middle-end/58094
49 * ipa-inline.c (check_callers): New function.
50 (check_caller_edge): Remove.
51 (want_inline_function_to_all_callers_p): Also permit alises that are
52 called dirrectly.
53 (inline_to_all_callers): Terminate the walk when devirtualization
54 introduce new calls.
55
a93c18c8
JH
562013-09-13 Jan Hubicka <jh@suse.cz>
57
58 * ipa-inline-analysis.c (struct growth_data): Add node.
59 (do_estimate_growth_1): Fix detection of recursion.
60
f3b331d1
JJ
612013-09-13 Jakub Jelinek <jakub@redhat.com>
62
63 PR tree-optimization/58392
64 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
65 to avoid shadowing of outer loop variable. If
66 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
67 replace_by_duplicate_decl simduid of loops that have it set and
68 set dest_cfun->has_simduid_loops and/or
69 dest_cfun->has_force_vect_loops.
70 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
71 instead of maybe_lookup_decl.
72 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
73 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
74 call. Copy over force_vect and copy and remap simduid. Set
75 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
76 (copy_cfg_body): Remove blocks_to_copy argument. Use
77 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
78 caller. Don't set cfun->has_simduid_loops and/or
79 cfun->has_force_vect_loops here.
80 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
81 caller.
82 (expand_call_inline, tree_function_versioning): Adjust copy_body
83 callers.
84
ae6d0907
MJ
852013-09-13 Martin Jambor <mjambor@suse.cz>
86
87 PR bootstrap/58388
88 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
89 the assert if the edge was a speculative one.
90
2fd5894f
RB
912013-09-13 Richard Biener <rguenther@suse.de>
92
93 * tree-data-ref.h (known_dependences_p): Move here ...
94 * tree-loop-distribution.c (known_dependences_p): ... from here.
95 (dump_rdg_component, debug_rdg_component): Remove.
96 (dump_rdg): Adjust.
97 (generate_loops_for_partition): Use gimple_uid instead of
98 relying on matching stmt visit order.
99 (rdg_build_partitions): Take starting stmt vector.
100 (ldist_gen): Merge into ...
101 (distribute_loop): ... this function. Do not compute starting
102 vertices vector.
103 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
104
86eb4bd7
KT
1052013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
106
107 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
108 Set type attribute correctly. Set predicable_short_it attribute.
109 (cmpsi_shiftsi): Remove %? from output template.
110
83a95546
RB
1112013-09-13 Richard Biener <rguenther@suse.de>
112
113 * tree-loop-distribution.c (struct rdg_component,
114 rdg_defs_used_in_other_loops_p, free_rdg_components,
115 rdg_build_components): Remove.
116 (stmts_from_loop): Do not record virtual PHIs.
117 (generate_loops_for_partition): Skip virtual PHIs.
118 (build_rdg_partition_for_component): Rename to ...
119 (build_rdg_partition_for_vertex): ... this and adjust.
120 (rdg_build_partitions): Take a vector of starting vertices
121 instead of components. Remove unnecessary leftover handling.
122 (ldist_gen): Do not build components or record other stores.
123 (distribute_loop): Do not distribute loops containing stmts
124 with side-effects.
125
c0ecf7f1
CB
1262013-09-13 Christian Bruel <christian.bruel@st.com>
127
128 PR target/58314
129 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
130
34f56efc
KT
1312013-09-13 Kai Tietz <ktietz@redhat.com>
132
740f2bf3
KT
133 * config.gcc: Separate cases for mingw and cygwin targets,
134 and add 64-bit cygwin target case.
135
34f56efc
KT
136 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
137 dll-export inline-functions.
138 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
139
140
1412013-09-13 Jeff Law <law@redhat.com>
823b5d6b
JL
142
143 PR middle-end/58387
144 Revert:
145 2013-09-06 Jeff Law <law@redhat.com>
146
147 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
148 edge implied equivalences into successor phis.
149
aadb52ad
DD
1502013-09-12 DJ Delorie <dj@redhat.com>
151
152 * config/rl78/rl78-virt.md: Change from | to \; for asm line
153 separators.
154
90ae7019
BM
1552013-09-12 Brooks Moses <bmoses@google.com>
156
157 PR driver/42955
158 * Makefile.in: Do not install driver binaries in $(target)/bin.
159
17b2557c
DD
1602013-09-12 DJ Delorie <dj@redhat.com>
161
c5215a9b
DD
162 * config/rl78/rl78.opt (mrelax): New.
163 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
164 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
165
17b2557c
DD
166 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
167 between SP and FP.
168 (rl78_expand_epilogue): Likewise.
169
ecafcf05
VM
1702013-09-12 Vladimir Makarov <vmakarov@redhat.com>
171
172 PR middle-end/58335
84f6d9fa 173 * lra-eliminations.c (remove_reg_equal_offset_note): New.
ecafcf05
VM
174 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
175 pointer elimination with using remove_reg_equal_offset_note.
176
8369f38a
DD
1772013-09-12 DJ Delorie <dj@redhat.com>
178
f6a83b4a
DD
179 * config/msp430/: New port.
180 * config.gcc (msp430): Added.
181 * doc/invoke.texi: Document MSP430 options.
182 * doc/install.texi: Document msp430-elf
183 * doc/md.texi: Document msp430-elf
184 * doc/contrib.texi: Document msp430-elf
185
8369f38a
DD
186 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
187 PSImode.
188
a854f856
MJ
1892013-09-12 Martin Jambor <mjambor@suse.cz>
190
191 PR ipa/58389
192 * ipa-prop.c (remove_described_reference): Give up if the edge in the
193 reference descriptor is NULL.
194 (ipa_edge_removal_hook): If owning a reference descriptor, set its
195 edge to NULL.
196
b12ebd96
AM
1972013-09-12 Andrew MacLeod <amacleod@redhat.com>
198
199 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
200 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
201 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
202 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
203 make_temp_ssa_name): move to tree-ssanames.h
204 * tree-ssa-alias.h: Move prototype.
205 * tree-ssa.h: Include tree-ssanames.h.
206 * tree-ssanames.c (FREE_SSANAMES): Move to here.
207 * tree-ssanames.h: New. Move items from tree-flow*.h
208 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
209
e1fb4ad3
RB
2102013-09-12 Richard Biener <rguenther@suse.de>
211
212 PR tree-optimization/58404
213 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
214 propagate non-invariant addresses into dereferences wrapped
215 in component references.
216
154c9188
RB
2172013-09-12 Richard Biener <rguenther@suse.de>
218
219 PR tree-optimization/58402
220 * passes.def: Move pass_late_warn_uninitialized later.
221
7a300452
AM
2222013-09-12 Andrew MacLeod <amacleod@redhat.com>
223
224 * tree-ssa.h: New. Move content from tree-flow.h and
225 tree-flow-inline.h.
226 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
227 Move prototypes belonging to tree-ssa.c.
228 * tree-flow-inline.h (redirect_edge_var_map_def,
229 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
230 tree-ssa.h.
231 * gimple.h: Adjust prototypes.
232 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move to...
233 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
234 * tree.h: Move prototype to tree-ssa.h.
235 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
236 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
237 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
238 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
239 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
240 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
241 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
242 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
243 graphite-blocking.c, graphite-clast-to-gimple.c, graphite-dependences.c,
244 graphite-interchange.c, graphite-optimize-isl.c, graphite-poly.c,
245 graphite-scop-detection.c, graphite-sese-to-poly.c, graphite.c,
246 ipa-cp.c, ipa-inline-analysis.c, ipa-inline-transform.c, ipa-inline.c,
247 ipa-prop.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c, ipa-utils.c,
248 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
249 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
250 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
251 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
252 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
253 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
254 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
255 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
256 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
257 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
258 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
259 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
260 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
261 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
262 value-prof.c, var-tracking.c,
263 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
264
97463b2b
RB
2652013-09-12 Richard Biener <rguenther@suse.de>
266
267 PR tree-optimization/58396
268 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
269 (build_rdg): Take a loop-nest parameter, fix memleaks.
270 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
271
80ecd89d
YR
2722013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
273
274 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
275 for SLM.
276
893e85fa
CM
2772013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
278
279 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
280 The FMA instruction names should have a 'v' prefix.
281
174ec470
RB
2822013-09-12 Richard Biener <rguenther@suse.de>
283
284 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
285 (dot_rdg): Use popen instead of system in optional code.
286 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
287 (already_processed_vertex_p): Adjust.
288 (has_anti_or_output_dependence, predecessor_has_mem_write,
289 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
290 rdg_flag_uses): Remove.
291 (rdg_flag_vertex): Simplify.
292 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
293 remove recursion.
294 (build_rdg_partition_for_component): Process the first vertex
295 of a component only.
296 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
297
38c89763
AM
2982013-09-12 Alan Modra <amodra@gmail.com>
299
300 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
301
3e3d9d17
DD
3022013-09-11 DJ Delorie <dj@redhat.com>
303 Nick Clifton <nickc@redhat.com>
304
305 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
306 (rl78_stack_based_mem): New.
307 * config/rl78/constraints.md (Iv08): New.
308 (Iv16): New.
309 (Iv24): New.
310 (Is09): New.
311 (Is17): New.
312 (Is25): New.
313 (ISsi): New.
314 (IShi): New.
315 (ISqi): New.
316 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
317 (movhi): Likewise.
318 (movsi): Change from expand to insn-and-split.
319 (ashrsi3): Clobber AX.
320 (lshrsi3): New.
321 (ashlsi3): New.
322 (cbranchsi4): New.
323 * config/rl78/rl78.md (CC_REG): Fix.
324 (addsi3): Allow memory and immediate operands.
325 (addsi3_internal): Split into...
326 (addsi3_internal_virt): ...new, and ...
327 (addsi3_internal_real): ...new.
328 (subsi): New.
329 (subsi3_internal_virt): New.
330 (subsi3_internal_real): New.
331 (mulsi3): Add memory operand.
332 (mulsi3_rl78): Likewise.
333 (mulsi3_g13): Likewise.
334 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
335 (cbranchqi4_real): Add more constraint options.
336 (cbranchhi4_real): Expand pattern.
337 (cbranchhi4_real_signed): New.
338 (cbranchhi4_real_inverted): New.
339 (cbranchsi4_real_lt): New.
340 (cbranchsi4_real_ge): New.
341 (cbranchsi4_real_signed): New.
342 (cbranchsi4_real): New.
343 (peephole2): New.
893e85fa
CM
344 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
345 constant shifts.
3e3d9d17
DD
346 (lshrsi3_virt): Likewise.
347 (ashlsi3_virt): Likewise.
348 (cbranchqi4_virt_signed): New.
349 (cbranchhi4_virt_signed): New.
350 (cbranchsi4_virt): New.
351 * config/rl78/rl78.c: Whitespace fixes throughout.
352 (move_elim_pass): New.
353 (pass_data_rl78_move_elim): New.
354 (pass_rl78_move_elim): New.
355 (make_pass_rl78_move_elim): New.
356 (rl78_devirt_info): Run devirt earlier.
357 (rl78_move_elim_info): New.
358 (rl78_asm_file_start): Register it.
359 (rl78_split_movsi): New.
360 (rl78_as_legitimate_address): Allow virtual base registers when
361 appropriate.
362 (rl78_addr_space_convert): Remove spurious debug stuff.
363 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
364 (rl78_print_operand): More cases for not printing '#'.
365 (rl78_expand_compare): Remove most of the logic.
366 (content_memory): New.
367 (clear_content_memory): New.
368 (get_content_index): New.
369 (get_content_name): New.
370 (display_content_memory): New.
371 (update_content): New.
372 (record_content): New.
373 (already_contains): New.
374 (insn_ok_now): Re-recog insns with virtual registers.
375 (add_postponed_content_update): New.
376 (process_postponed_content_update): New.
377 (gen_and_emit_move): New.
893e85fa
CM
378 (transcode_memory_rtx): Record new location content.
379 Use gen_and_emit_move.
3e3d9d17
DD
380 (force_into_acc): New.
381 (move_to_acc): Use gen_and_emit_move.
382 (move_from_acc): Likewise.
383 (move_acc_to_reg): Likewise.
384 (move_to_x): Likewise.
385 (move_to_hl): Likewise.
386 (move_to_de): Likewise.
387 (rl78_alloc_physical_registers_op1): Record location content.
388 (has_constraint): New.
389 (rl78_alloc_physical_registers_op2): Record location content.
390 Optimize use of HL.
391 (rl78_alloc_physical_registers_ro1): Likewise.
392 (rl78_alloc_physical_registers_cmp): Likewise.
393 (rl78_alloc_physical_registers_umul): Likewise.
394 (rl78_alloc_address_registers_macax): New.
395 (rl78_alloc_physical_registers): Initialize and set location
396 content memory as needed.
397 (rl78_reorg): Make sure split2 is called.
398 (rl78_rtx_costs): New.
399
088845a5
RS
4002013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
401
402 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
403 for (not (neg ...)) and (neg (not ...)) cases.
404
5c2961cf
RB
4052013-09-11 Richard Biener <rguenther@suse.de>
406
407 PR middle-end/58377
408 * passes.def: Split critical edges before late uninit warning passes.
409 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
410
427e6a14
JJ
4112013-09-11 Jakub Jelinek <jakub@redhat.com>
412
413 PR tree-optimization/58385
414 * fold-const.c (build_range_check): If both low and high are NULL,
415 use omit_one_operand_loc to preserve exp side-effects.
416
df45c0cc
KT
4172013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
418
419 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
420
80ab0b19
RB
4212013-09-11 Richard Biener <rguenther@suse.de>
422
423 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
424 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
425 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
426 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
427 stmts_from_loop, known_dependences_p, build_empty_rdg,
428 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
429 * tree-loop-distribution.c: ... here.
430 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
431 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
432 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
433 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
434 * tree-loop-distribution.c: ... here.
435 * tree-loop-distribution.c: Include gimple-pretty-print.h.
436 (struct partition_s): Add loops member.
437 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
438 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
439 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
440
85a77221
AI
4412013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
442 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
443 Sergey Lega <sergey.s.lega@intel.com>
444 Anna Tikhonova <anna.tikhonova@intel.com>
445 Ilya Tocar <ilya.tocar@intel.com>
446 Andrey Turetskiy <andrey.turetskiy@intel.com>
447 Ilya Verbin <ilya.verbin@intel.com>
448 Kirill Yukhin <kirill.yukhin@intel.com>
449 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
450
451 * config/i386/constraints.md (k): New.
452 (Yk): Ditto.
453 * config/i386/i386.c (const regclass_map): Add new mask registers.
454 (dbx_register_map): Ditto.
455 (dbx64_register_map): Ditto.
456 (svr4_dbx_register_map): Ditto.
457 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
458 disabled.
459 (ix86_preferred_reload_class): Disable constants for mask registers.
460 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
461 (ix86_hard_regno_mode_ok): Support new mask registers.
462 (x86_order_regs_for_local_alloc): Ditto.
463 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
464 (FIXED_REGISTERS): Add new mask registers.
465 (CALL_USED_REGISTERS): Ditto.
466 (REG_ALLOC_ORDER): Ditto.
467 (VALID_MASK_REG_MODE): New.
468 (FIRST_MASK_REG): Ditto.
469 (LAST_MASK_REG): Ditto.
470 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
471 (MAYBE_MASK_CLASS_P): New.
472 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
473 (REG_CLASS_CONTENTS): Ditto.
474 (MASK_REGNO_P): New.
475 (ANY_MASK_REG_P): Ditto.
476 (HI_REGISTER_NAMES): Add new mask registers.
893e85fa
CM
477 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
478 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
479 mask registers.
85a77221
AI
480 (attribute "type"): Add mskmov, msklog.
481 (attribute "length_immediate"): Support them.
482 (attribute "memory"): Ditto.
483 (attribute "prefix_0f"): Ditto.
484 (*movhi_internal): Support new mask registers.
485 (*movqi_internal): Ditto.
486 (define_split): Split out clobber pattern is a logic
487 insn on mask registers.
488 (*k<logic><mode>): New.
489 (*andhi_1): Extend to support mask regs.
490 (*andqi_1): Extend to support mask regs.
491 (kandn<mode>): New.
492 (define_split): Split and-not to and and not if operands
493 are not mask regs.
494 (*<code><mode>_1): Separate HI mode to new pattern...
495 (*<code>hi_1): This.
496 (*<code>qi_1): Extend to support mask regs.
497 (kxnor<mode>): New.
498 (kortestzhi): Ditto.
499 (kortestchi): Ditto.
500 (kunpckhi): Ditto.
501 (*one_cmpl<mode>2_1): Remove HImode and handle it...
502 (*one_cmplhi2_1): ...Here, now with mask registers support.
503 (*one_cmplqi2_1): Support new mask registers.
893e85fa
CM
504 (HI/QImode arithmetics splitter): Don't split if mask registers
505 are used.
85a77221
AI
506 (HI/QImode not splitter): Ditto.
507 * config/i386/predicated.md (mask_reg_operand): New.
508 (general_reg_operand): Ditto.
509
7e570821
AI
5102013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
511
512 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
513 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
514
9e1376e9
JL
5152013-09-10 Jeff Law <law@redhat.com>
516
517 PR tree-optimization/58380
518 * tree-ssa-threadupdate.c (thread_block): Recognize another case
519 of threading through a buried loop header.
520
521 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
522 return value for single successor case.
523
5bbcb888
JH
5242013-09-10 Jan Hubicka <jh@suse.cz>
525
526 * ipa-devirt.c (ipa_devirt): Enable with LTO.
527
d742ff4b
RE
5282013-09-10 Richard Earnshaw <rearnsha@arm.com>
529
530 PR target/58361
531 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
532 support conditional execution.
533 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
534
b0681c9e
VM
5352013-09-10 Vladimir Makarov <vmakarov@redhat.com>
536
537 * lra.c (lra): Clear lra_optional_reload_pseudos before every
538 constraint pass.
539 * lra-constraints.c (curr_insn_transform): Switch on optional
540 reloads. Check destination too to check move insn.
541 (undo_optional_reloads): Add check that the original peudo did not
542 changed its allocation and the optional reload was inherited on
543 last inheritance pass. Break loop after deciding to keep optional
544 reload.
545 (lra_undo_inheritance): Add check that inherited pseudo still in
546 memory.
547
c9b1c957
JG
5482013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
549
550 * config/aarch64/aarch64.md (generic_sched): New.
551 * config/aarch64/aarch64-generic.md (load): Make conditional
552 on generic_sched attribute.
553 (nonload): Likewise.
554
82d618d3
JH
5552013-09-10 Jan Hubicka <jh@suse.cz>
556
557 * lto-cgraph.c: Include ipa-utils.h.
558 (compute_ltrans_boundary): Also add possible targets into the boundary.
559
668b6894
JH
5602013-09-10 Jan Hubicka <jh@suse.cz>
561
562 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
563 VAR_DECL of vtable rather than full expression.
564
4f90d3e0
JH
5652013-09-10 Jan Hubicka <jh@suse.cz>
566 Paolo Carlini <paolo.carlini@oracle.com>
567
568 * cgraphunit.c (analyze_functions): Save input_location, set it
569 to UNKNOWN_LOCATION and restore it at the end.
570
39e87baf
MJ
5712013-09-10 Martin Jambor <mjambor@suse.cz>
572
573 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
574 represented by a thunk.
575
581aedec
JL
5762013-09-10 Jeff Law <law@redhat.com>
577
578 PR tree-optimization/58343
579 * tree-ssa-threadupdate.c (thread_block): Identify and disable
580 jump threading requests through loop headers buried in the middle
581 of a jump threading path.
582
583 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
584 in return value/type.
585
84cf4ab6
JJ
5862013-09-10 Jakub Jelinek <jakub@redhat.com>
587
588 PR rtl-optimization/58365
589 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
590 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
591 it differs.
592
447dd906
RB
5932013-09-10 Richard Biener <rguenther@suse.de>
594
595 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
596 * tree-data-ref.c (create_rdg_vertices): Collect all data
597 references, signal failure to the caller, use data-ref API.
893e85fa
CM
598 (build_rdg): Compute data references only once. Maintain lifetime
599 of data references and data dependences from within RDG.
447dd906
RB
600 (free_rdg): Free dependence relations.
601 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
602 inventing extra dependences.
603 (distribute_loop): Update for RDG API changes.
604
f14a7324
KT
6052013-09-10 Kai Tietz <ktietz@redhat.com>
606
607 * doc/invoke.texi (fms-extensions): Document changed
608 behavior for ms-abi targets.
609 * config/i386/i386.c (ix86_option_override_internal):
610 Set default value of option -fms-extension for ms-abi targets.
611
ff131f26
MZ
6122013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
613
614 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
615
33115a2c
AM
6162013-09-10 Alan Modra <amodra@gmail.com>
617
618 PR target/58330
619 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
620
b6872a13
AM
6212013-09-10 Alan Modra <amodra@gmail.com>
622
623 * config/rs6000/predicates.md (add_cint_operand): New.
624 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
625 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
626 using add_cint_operand.
627 (largetoc_high_plus_aix): Likewise.
628
28fd0ba2
JJ
6292013-09-09 Jakub Jelinek <jakub@redhat.com>
630
631 PR tree-optimization/58364
632 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
633 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
634 the current range can't be an unconditional true or false.
635
552d99fe
JG
6362013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
637
638 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
639
f43245d1
UB
6402013-09-09 Uros Bizjak <ubizjak@gmail.com>
641
642 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
643
c9b0866a
PC
6442013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
645
646 PR c++/43452
647 * doc/invoke.texi (-Wdelete-incomplete): Document it.
648
78d8b9f0
IB
6492013-09-09 Ian Bolton <ian.bolton@arm.com>
650
651 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
652 NO_REGS for immediate that can't be moved directly into FP_REGS.
653
274b2532
KT
6542013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
655
656 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
657 comparison with negated operand.
c9b0866a
PC
658 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
659 RTL form.
274b2532 660
467a3558
RB
6612013-09-09 Richard Biener <rguenther@suse.de>
662
663 PR middle-end/58326
664 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
665 of a subloop record all its block as affecting loop-closed SSA form.
666
088c5368
RS
6672013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
668
669 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
670 of an rtx/bitpos pair.
671 (store_fixed_bit_field): Update accordingly.
672
69db2d57
RS
6732013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
674
675 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
676 GEN_INT.
677 * builtins.c (expand_errno_check): Likewise.
678 * dwarf2cfi.c (init_return_column_size): Likewise.
679 * except.c (sjlj_mark_call_sites): Likewise.
680 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
681 * lra-constraints.c (emit_inc): Likewise.
682 * ree.c (combine_set_extension): Likewise.
683 * regmove.c (fixup_match_2): Likewise.
684 * reload1.c (inc_for_reload): Likewise.
685
69a59f0f
RS
6862013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
687
688 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
689 (force_to_mode, simplify_shift_const_1, simplify_comparison):
690 Use gen_int_mode with the mode of the associated simplify_* call.
691 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
692 * expmed.c (expand_shift_1): Likewise.
693 * function.c (instantiate_virtual_regs_in_insn): Likewise.
694 * loop-iv.c (iv_number_of_iterations): Likewise.
695 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
696 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
697
2f1cd2eb
RS
6982013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
699
700 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
701 of the associated expand_* call.
702 (asan_emit_stack_protection): Likewise.
703 * builtins.c (round_trampoline_addr): Likewise.
704 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
705 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
706 (emit_store_flag): Likewise.
707 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
708 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
709 Likewise.
710 * function.c (instantiate_virtual_regs_in_insn): Likewise.
711 * ifcvt.c (noce_try_store_flag_constants): Likewise.
712 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
713 * modulo-sched.c (generate_prolog_epilog): Likewise.
714 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
715 (expand_ctz, expand_ffs, expand_unop): Likewise.
716
4789c0ce
RS
7172013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
718
719 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
720 of the associated gen_rtx_* call.
721 * caller-save.c (init_caller_save): Likewise.
722 * combine.c (find_split_point, make_extraction): Likewise.
723 (make_compound_operation): Likewise.
724 * dwarf2out.c (mem_loc_descriptor): Likewise.
725 * explow.c (plus_constant, probe_stack_range): Likewise.
726 * expmed.c (expand_mult_const): Likewise.
727 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
728 * reload1.c (init_reload): Likewise.
729 * valtrack.c (cleanup_auto_inc_dec): Likewise.
730 * var-tracking.c (adjust_mems): Likewise.
731 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
732 rather than gen_rtx_fmt_ee.
733
a811f7cb
JH
7342013-09-09 Jan Hubicka <jh@suse.cz>
735
736 PR middle-end/58294
737 * value-prof.c (gimple_ic): Copy also abnormal edges.
738
dcad1dd3
RS
7392013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
740
741 * asan.c (asan_shadow_cst): Use gen_int_mode.
742
daf5c770
JH
7432013-09-08 Jan Hubicka <jh@suse.cz>
744
745 * ipa-profile.c: Add toplevel comment.
746 (ipa_propagate_frequency_1): Be more conservative when profile is read.
747 (contains_hot_call_p): New function.
748 (ipa_propagate_frequency): Set frequencies based on counts when
749 profile is read.
750 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
751 profile; do not tamper with profile after inlining if it is read.
752
a12cd2db
JH
7532013-09-08 Jan Hubicka <jh@suse.cz>
754
755 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
756 speculative edges.
757
856a1588
JH
7582013-09-08 Jan Hubicka <jh@suse.cz>
759
760 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
761 summary generation.
762
4f4b0b73
JL
7632013-09-08 Jeff Law <law@redhat.com>
764
765 PR bootstrap/58340
766 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
767 of 'found'.
768
cc4029ee
AK
7692013-09-08 Andi Kleen <ak@linux.intel.com>
770
771 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
772
7b395ddd
JH
7732013-09-08 Jan Hubicka <jh@suse.cz>
774
775 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
776
c34c46dd
RS
7772013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
778
779 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
780 for non-debug insns.
781 * lra.c (new_insn_reg): Take the containing insn as a parameter.
782 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
783 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
784 accordingly.
785
3462aa02
JH
7862013-09-08 Jan Hubicka <jh@suse.cz>
787
788 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
789 targets and devirtualize to BUILT_IN_UNREACHABLE.
790 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
791 * ipa.c (walk_polymorphic_call_targets): New function.
792 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
793 functions; use the new timevar.
794 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
795 was removed from the program.
796 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
797 not consider it in the walk when its vtable is dead.
798 (possible_polymorphic_call_targets_1): Pass anonymous flag to
799 record_binfo.
800 (devirt_variable_node_removal_hook): New function.
801 (possible_polymorphic_call_targets): Also register
802 devirt_variable_node_removal_hook.
803 (ipa_devirt): Do not do non-speculative devirtualization.
804 (gate_ipa_devirt): One execute if devirtualizing speculatively.
805
26e5b0fd
JH
8062013-09-08 Jan Hubicka <jh@suse.cz>
807
808 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
809 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
810 varpool_remove_variable_insertion_hook): Declare.
811 * varpool.c (varpool_node_hook_list): New structure.
812 (first_varpool_node_removal_hook,
813 first_varpool_variable_insertion_hook): New variables.
814 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
815 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
816 varpool_remove_variable_insertion_hook,
817 varpool_call_variable_insertion_hooks): New functions.
818 (varpool_remove_node): Use it.
819
97f1f314 8202013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 821
97f1f314
PC
822 PR c++/54941
823 * diagnostic.c (diagnostic_build_prefix): When s.file is
824 "<built-in>" don't output line and column numbers.
6df2ca09 825
bcb650cb
JH
8262013-09-06 Jan Hubicka <jh@suse.cz>
827
828 * cgraphunit.c (expand_thunk): Get body before touching arguments.
829 * lto-streamer-out.c: Stream thunks, too.
830 * lto-streamer-in.c (input_function): Pop cfun here
831 (lto_read_body): Instead of here.
832
87f3fea8
CT
8332013-09-06 Caroline Tice <cmtice@google.com>
834
835 * doc/install.texi: Add documentation for the --enable-vtable-verify
836 and the --disable-libvtv configure options.
837
5e8d9b45
JL
8382013-09-06 Jeff Law <law@redhat.com>
839
840 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
841 edge implied equivalences into successor phis.
842
3a001aff
JR
8432013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
844
845 * resource.c (mark_referenced_resources): Handle COND_EXEC.
846
8472013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ
848
849 * resource.c (mark_target_live_regs): Compute resources taking
850 into account if a call is predicated or not.
851
2c641f8b
EB
8522013-09-06 Eric Botcazou <ebotcazou@adacore.com>
853
854 * toplev.c (output_stack_usage): Be prepared for suffixes created by
855 the compiler in the function names.
856
a81b0a3d
JH
8572013-09-06 Jan Hubicka <jh@suse.cz>
858
859 PR middle-end/58094
860 * ipa-inline.c (has_caller_p): New function.
861 (want_inline_function_to_all_callers_p): Use it.
862 (sum_callers, inline_to_all_callers): Break out from ...
863 (ipa_inline): ... here.
864
68e4ca7e
JH
8652013-09-06 Jan Hubicka <jh@suse.cz>
866
2c641f8b
EB
867 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
868 only when AVX is enabled.
68e4ca7e 869
3c73e2fe
JG
8702013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
871
872 * config/aarch64/aarch64.md
873 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
874 is fpsimd_<load/store>2.
875 (load_pair<mode>): Likewise.
876 (store_pair<mode>): Likewise.
877
8c48eecd
JG
8782013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
879
880 * config/arm/types.md (type): Add "mrs" type.
881 * config/aarch64/aarch64.md
882 (aarch64_load_tp_hard): Make type "mrs".
883 * config/arm/arm.md
884 (load_tp_hard): Make type "mrs".
885 * config/arm/cortex-a15.md: Update with new attributes.
886 * config/arm/cortex-a5.md: Update with new attributes.
887 * config/arm/cortex-a53.md: Update with new attributes.
888 * config/arm/cortex-a7.md: Update with new attributes.
889 * config/arm/cortex-a8.md: Update with new attributes.
890 * config/arm/cortex-a9.md: Update with new attributes.
891 * config/arm/cortex-m4.md: Update with new attributes.
892 * config/arm/cortex-r4.md: Update with new attributes.
893 * config/arm/fa526.md: Update with new attributes.
894 * config/arm/fa606te.md: Update with new attributes.
895 * config/arm/fa626te.md: Update with new attributes.
896 * config/arm/fa726te.md: Update with new attributes.
897
9ea187a5
JG
8982013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
899
900 * config/aarch64/aarch64.md
901 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
902 (*movtf_aarch64): Likewise.
903 * config/arm/arm.md
904 (thumb1_movdi_insn): Use "multiple" for type where more than one
905 instruction is used for a move.
906 (*arm32_movhf): Likewise.
907 (*thumb_movdf_insn): Likewise.
908
292b89b3
JG
9092013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
910
911 * config/arm/types.md (type): Rename fcpys to fmov.
912 * config/arm/vfp.md
913 (*arm_movsi_vfp): Rename type fcpys as fmov.
914 (*thumb2_movsi_vfp): Likewise
915 (*movhf_vfp_neon): Likewise
916 (*movhf_vfp): Likewise
917 (*movsf_vfp): Likewise
918 (*thumb2_movsf_vfp): Likewise
919 (*movsfcc_vfp): Likewise
920 (*thumb2_movsfcc_vfp): Likewise
921 * config/aarch64/aarch64-simd.md
922 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
923 * config/aarch64/aarch64.md
924 (*movsi_aarch64): Replace type mov_reg with fmovs.
925 (*movdi_aarch64): Likewise
926 (*movsf_aarch64): Likewise
927 (*movdf_aarch64): Likewise
928 * config/arm/arm.c
929 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
930 * config/arm/iwmmxt.md
931 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
932 * config/arm/arm1020e.md: Update with new attributes.
933 * config/arm/cortex-a15-neon.md: Update with new attributes.
934 * config/arm/cortex-a5.md: Update with new attributes.
935 * config/arm/cortex-a53.md: Update with new attributes.
936 * config/arm/cortex-a7.md: Update with new attributes.
937 * config/arm/cortex-a8-neon.md: Update with new attributes.
938 * config/arm/cortex-a9.md: Update with new attributes.
939 * config/arm/cortex-m4-fpu.md: Update with new attributes.
940 * config/arm/cortex-r4f.md: Update with new attributes.
941 * config/arm/marvell-pj4.md: Update with new attributes.
942 * config/arm/vfp11.md: Update with new attributes.
943
6a24a83e
JG
9442013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
945
946 * config/aarch64/aarch64.md
947 (*madd<mode>): Fix type attribute.
948 (*maddsi_uxtw): Likewise.
949 (*msub<mode>): Likewise.
950 (*msubsi_uxtw): Likewise.
951 (<su_optab>maddsidi4): Likewise.
952 (<su_optab>msubsidi4): Likewise.
953
b86923f0
JG
9542013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
955
956 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
957 * config/arm/arm.md (core_cycles): Remove fdiv.
958 * config/arm/vfp.md:
959 (*sqrtsf2_vfp): Update for attribute changes.
960 (*sqrtdf2_vfp): Likewise.
961 * config/aarch64/aarch64.md:
962 (sqrt<mode>2): Update for attribute changes.
963 * config/arm/arm1020e.md: Update with new attributes.
964 * config/arm/cortex-a15-neon.md: Update with new attributes.
965 * config/arm/cortex-a5.md: Update with new attributes.
966 * config/arm/cortex-a53.md: Update with new attributes.
967 * config/arm/cortex-a7.md: Update with new attributes.
968 * config/arm/cortex-a8-neon.md: Update with new attributes.
969 * config/arm/cortex-a9.md: Update with new attributes.
970 * config/arm/cortex-m4-fpu.md: Update with new attributes.
971 * config/arm/cortex-r4f.md: Update with new attributes.
972 * config/arm/marvell-pj4.md: Update with new attributes.
973 * config/arm/vfp11.md: Update with new attributes.
974
7b49c9e1
JG
9752013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
976
977 * config/arm/types.md
978 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
979 * config/aarch64/aarch64.md
980 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
981 new attributes.
982 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
983 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
984 (float<GPI:mode><GPF:mode>2): Likewise.
985 * config/arm/vfp.md
986 (*truncsisf2_vfp): Update with new attributes.
987 (*truncsidf2_vfp): Likewise.
988 (fixuns_truncsfsi2): Likewise.
989 (fixuns_truncdfsi2): Likewise.
990 (*floatsisf2_vfp): Likewise.
991 (*floatsidf2_vfp): Likewise.
992 (floatunssisf2): Likewise.
993 (floatunssidf2): Likewise.
994 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
995 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
996 * config/arm/arm1020e.md: Update with new attributes.
997 * config/arm/cortex-a15-neon.md: Update with new attributes.
998 * config/arm/cortex-a5.md: Update with new attributes.
999 * config/arm/cortex-a53.md: Update with new attributes.
1000 * config/arm/cortex-a7.md: Update with new attributes.
1001 * config/arm/cortex-a8-neon.md: Update with new attributes.
1002 * config/arm/cortex-a9.md: Update with new attributes.
1003 * config/arm/cortex-m4-fpu.md: Update with new attributes.
1004 * config/arm/cortex-r4f.md: Update with new attributes.
1005 * config/arm/marvell-pj4.md: Update with new attributes.
1006 * config/arm/vfp11.md: Update with new attributes.
1007
e7df8af8
JG
10082013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1009
1010 * config/aarch64/arm_neon.h
1011 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
1012 (vqtbx<1,2,3,4><q>_s8): Likewise.
1013
594726e4
JG
10142013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1015
f43245d1
UB
1016 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
1017 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG
1018 (add<mode>3): Add type attribute.
1019 (add<mode>3): Likewise.
1020 (usadd<mode>3): Likewise.
1021 (ssadd<mode>3): Likewise.
1022 (sub<mode>3): Likewise.
1023 (sub<mode>3): Likewise.
1024 (ussub<mode>3): Likewise.
1025 (sssub<mode>3): Likewise.
1026 (ssmulsa3): Likewise.
1027 (usmulusa3): Likewise.
1028 (arm_usatsihi): Likewise.
1029 * config/arm/vfp.md
1030 (*movdi_vfp): Add types for all instructions.
1031 (*movdi_vfp_cortexa8): Likewise.
1032 (*movhf_vfp_neon): Likewise.
1033 (*movhf_vfp): Likewise.
1034 (*movdf_vfp): Likewise.
1035 (*thumb2_movdf_vfp): Likewise.
1036 (*thumb2_movdfcc_vfp): Likewise.
1037 * config/arm/arm.md: Add type attribute to all insn patterns.
1038 (*thumb1_adddi3): Add type attribute.
1039 (*arm_adddi3): Likewise.
1040 (*adddi_sesidi_di): Likewise.
1041 (*adddi_zesidi_di): Likewise.
1042 (*thumb1_addsi3): Likewise.
1043 (addsi3_compare0): Likewise.
1044 (*addsi3_compare0_scratch): Likewise.
1045 (*compare_negsi_si): Likewise.
1046 (cmpsi2_addneg): Likewise.
1047 (*addsi3_carryin_<optab>): Likewise.
1048 (*addsi3_carryin_alt2_<optab>): Likewise.
1049 (*addsi3_carryin_clobercc_<optab>): Likewise.
1050 (*subsi3_carryin): Likewise.
1051 (*subsi3_carryin_const): Likewise.
1052 (*subsi3_carryin_compare): Likewise.
1053 (*subsi3_carryin_compare_const): Likewise.
1054 (*arm_subdi3): Likewise.
1055 (*thumb_subdi3): Likewise.
1056 (*subdi_di_zesidi): Likewise.
1057 (*subdi_di_sesidi): Likewise.
1058 (*subdi_zesidi_di): Likewise.
1059 (*subdi_sesidi_di): Likewise.
1060 (*subdi_zesidi_ze): Likewise.
1061 (thumb1_subsi3_insn): Likewise.
1062 (*arm_subsi3_insn): Likewise.
1063 (*anddi3_insn): Likewise.
1064 (*anddi_zesidi_di): Likewise.
1065 (*anddi_sesdi_di): Likewise.
1066 (*ne_zeroextracts): Likewise.
1067 (*ne_zeroextracts): Likewise.
1068 (*ite_ne_zeroextr): Likewise.
1069 (*ite_ne_zeroextr): Likewise.
1070 (*anddi_notdi_di): Likewise.
1071 (*anddi_notzesidi): Likewise.
1072 (*anddi_notsesidi): Likewise.
1073 (andsi_notsi_si): Likewise.
1074 (thumb1_bicsi3): Likewise.
1075 (*iordi3_insn): Likewise.
1076 (*iordi_zesidi_di): Likewise.
1077 (*iordi_sesidi_di): Likewise.
1078 (*thumb1_iorsi3_insn): Likewise.
1079 (*xordi3_insn): Likewise.
1080 (*xordi_zesidi_di): Likewise.
1081 (*xordi_sesidi_di): Likewise.
1082 (*arm_xorsi3): Likewise.
1083 (*andsi_iorsi3_no): Likewise.
1084 (*smax_0): Likewise.
1085 (*smax_m1): Likewise.
1086 (*arm_smax_insn): Likewise.
1087 (*smin_0): Likewise.
1088 (*arm_smin_insn): Likewise.
1089 (*arm_umaxsi3): Likewise.
1090 (*arm_uminsi3): Likewise.
1091 (*minmax_arithsi): Likewise.
1092 (*minmax_arithsi_): Likewise.
1093 (*satsi_<SAT:code>): Likewise.
1094 (arm_ashldi3_1bit): Likewise.
1095 (arm_ashrdi3_1bit): Likewise.
1096 (arm_lshrdi3_1bit): Likewise.
1097 (*arm_negdi2): Likewise.
1098 (*thumb1_negdi2): Likewise.
1099 (*arm_negsi2): Likewise.
1100 (*thumb1_negsi2): Likewise.
1101 (*negdi_extendsid): Likewise.
1102 (*negdi_zero_extend): Likewise.
1103 (*arm_abssi2): Likewise.
1104 (*thumb1_abssi2): Likewise.
1105 (*arm_neg_abssi2): Likewise.
1106 (*thumb1_neg_abss): Likewise.
1107 (one_cmpldi2): Likewise.
1108 (extend<mode>di2): Likewise.
1109 (*compareqi_eq0): Likewise.
1110 (*arm_extendhisi2addsi): Likewise.
1111 (*arm_movdi): Likewise.
1112 (*thumb1_movdi_insn): Likewise.
1113 (*arm_movt): Likewise.
1114 (*thumb1_movsi_insn): Likewise.
1115 (pic_add_dot_plus_four): Likewise.
1116 (pic_add_dot_plus_eight): Likewise.
1117 (tls_load_dot_plus_eight): Likewise.
1118 (*thumb1_movhi_insn): Likewise.
1119 (*thumb1_movsf_insn): Likewise.
1120 (*movdf_soft_insn): Likewise.
1121 (*thumb_movdf_insn): Likewise.
1122 (cbranchsi4_insn): Likewise.
1123 (cbranchsi4_scratch): Likewise.
1124 (*negated_cbranchsi4): Likewise.
1125 (*tbit_cbranch): Likewise.
1126 (*tlobits_cbranch): Likewise.
1127 (*tstsi3_cbranch): Likewise.
1128 (*cbranchne_decr1): Likewise.
1129 (*addsi3_cbranch): Likewise.
1130 (*addsi3_cbranch_scratch): Likewise.
1131 (*arm_cmpdi_insn): Likewise.
1132 (*arm_cmpdi_unsig): Likewise.
1133 (*arm_cmpdi_zero): Likewise.
1134 (*thumb_cmpdi_zero): Likewise.
1135 (*deleted_compare): Likewise.
1136 (*mov_scc): Likewise.
1137 (*mov_negscc): Likewise.
1138 (*mov_notscc): Likewise.
1139 (*cstoresi_eq0_thumb1_insn): Likewise.
1140 (cstoresi_nltu_thumb1): Likewise.
1141 (cstoresi_ltu_thu): Likewise.
1142 (thumb1_addsi3_addgeu): Likewise.
1143 (*arm_jump): Likewise.
1144 (*thumb_jump): Likewise.
1145 (*check_arch2): Likewise.
1146 (arm_casesi_internal): Likewise.
1147 (thumb1_casesi_dispatch): Likewise.
1148 (*arm_indirect_jump): Likewise.
1149 (*thumb1_indirect_jump): Likewise.
1150 (nop): Likewise.
1151 (*and_scc): Likewise.
1152 (*ior_scc): Likewise.
1153 (*compare_scc): Likewise.
1154 (*cond_move): Likewise.
1155 (*cond_arith): Likewise.
1156 (*cond_sub): Likewise.
1157 (*cmp_ite0): Likewise.
1158 (*cmp_ite1): Likewise.
1159 (*cmp_and): Likewise.
1160 (*cmp_ior): Likewise.
1161 (*ior_scc_scc): Likewise.
1162 (*ior_scc_scc_cmp): Likewise.
1163 (*and_scc_scc): Likewise.
1164 (*and_scc_scc_cmp): Likewise.
1165 (*and_scc_scc_nod): Likewise.
1166 (*negscc): Likewise.
1167 (movcond_addsi): Likewise.
1168 (movcond): Likewise.
1169 (*ifcompare_plus_move): Likewise.
1170 (*if_plus_move): Likewise.
1171 (*ifcompare_move_plus): Likewise.
1172 (*if_move_plus): Likewise.
1173 (*ifcompare_arith_arith): Likewise.
1174 (*if_arith_arith): Likewise.
1175 (*ifcompare_arith_move): Likewise.
1176 (*if_arith_move): Likewise.
1177 (*ifcompare_move_arith): Likewise.
1178 (*if_move_arith): Likewise.
1179 (*ifcompare_move_not): Likewise.
1180 (*if_move_not): Likewise.
1181 (*ifcompare_not_move): Likewise.
1182 (*if_not_move): Likewise.
1183 (*ifcompare_shift_move): Likewise.
1184 (*if_shift_move): Likewise.
1185 (*ifcompare_move_shift): Likewise.
1186 (*if_move_shift): Likewise.
1187 (*ifcompare_shift_shift): Likewise.
1188 (*ifcompare_not_arith): Likewise.
1189 (*ifcompare_arith_not): Likewise.
1190 (*if_arith_not): Likewise.
1191 (*ifcompare_neg_move): Likewise.
1192 (*if_neg_move): Likewise.
1193 (*ifcompare_move_neg): Likewise.
1194 (*if_move_neg): Likewise.
1195 (prologue_thumb1_interwork): Likewise.
1196 (*cond_move_not): Likewise.
1197 (*sign_extract_onebit): Likewise.
1198 (*not_signextract_onebit): Likewise.
1199 (stack_tie): Likewise.
1200 (align_4): Likewise.
1201 (align_8): Likewise.
1202 (consttable_end): Likewise.
1203 (consttable_1): Likewise.
1204 (consttable_2): Likewise.
1205 (consttable_4): Likewise.
1206 (consttable_8): Likewise.
1207 (consttable_16): Likewise.
1208 (*thumb1_tablejump): Likewise.
1209 (prefetch): Likewise.
1210 (force_register_use): Likewise.
1211 (thumb_eh_return): Likewise.
1212 (load_tp_hard): Likewise.
1213 (load_tp_soft): Likewise.
1214 (tlscall): Likewise.
1215 (*arm_movtas_ze): Likewise.
1216 (*arm_rev): Likewise.
1217 (*arm_revsh): Likewise.
1218 (*arm_rev16): Likewise.
1219 * config/arm/thumb2.md
1220 (*thumb2_smaxsi3): Likewise.
1221 (*thumb2_sminsi3): Likewise.
1222 (*thumb32_umaxsi3): Likewise.
1223 (*thumb2_uminsi3): Likewise.
1224 (*thumb2_negdi2): Likewise.
1225 (*thumb2_abssi2): Likewise.
1226 (*thumb2_neg_abss): Likewise.
1227 (*thumb2_movsi_insn): Likewise.
1228 (tls_load_dot_plus_four): Likewise.
1229 (*thumb2_movhi_insn): Likewise.
1230 (*thumb2_mov_scc): Likewise.
1231 (*thumb2_mov_negs): Likewise.
1232 (*thumb2_mov_negs): Likewise.
1233 (*thumb2_mov_nots): Likewise.
1234 (*thumb2_mov_nots): Likewise.
1235 (*thumb2_movsicc_): Likewise.
1236 (*thumb2_movsfcc_soft_insn): Likewise.
1237 (*thumb2_indirect_jump): Likewise.
1238 (*thumb2_and_scc): Likewise.
1239 (*thumb2_ior_scc): Likewise.
1240 (*thumb2_ior_scc_strict_it): Likewise.
1241 (*thumb2_cond_move): Likewise.
1242 (*thumb2_cond_arith): Likewise.
1243 (*thumb2_cond_ari): Likewise.
1244 (*thumb2_cond_sub): Likewise.
1245 (*thumb2_negscc): Likewise.
1246 (*thumb2_movcond): Likewise.
1247 (thumb2_casesi_internal): Likewise.
1248 (thumb2_casesi_internal_pic): Likewise.
1249 (*thumb2_alusi3_short): Likewise.
1250 (*thumb2_mov<mode>_shortim): Likewise.
1251 (*thumb2_addsi_short): Likewise.
1252 (*thumb2_subsi_short): Likewise.
1253 (thumb2_addsi3_compare0): Likewise.
1254 (*thumb2_cbz): Likewise.
1255 (*thumb2_cbnz): Likewise.
1256 (*thumb2_one_cmplsi2_short): Likewise.
1257 (*thumb2_negsi2_short): Likewise.
1258 (*orsi_notsi_si): Likewise.
1259 * config/arm/arm1020e.md: Update with new attributes.
1260 * config/arm/arm1026ejs.md: Update with new attributes.
1261 * config/arm/arm1136jfs.md: Update with new attributes.
1262 * config/arm/arm926ejs.md: Update with new attributes.
1263 * config/arm/cortex-a15.md: Update with new attributes.
1264 * config/arm/cortex-a5.md: Update with new attributes.
1265 * config/arm/cortex-a53.md: Update with new attributes.
1266 * config/arm/cortex-a7.md: Update with new attributes.
1267 * config/arm/cortex-a8.md: Update with new attributes.
1268 * config/arm/cortex-a9.md: Update with new attributes.
1269 * config/arm/cortex-m4.md: Update with new attributes.
1270 * config/arm/cortex-r4.md: Update with new attributes.
1271 * config/arm/fa526.md: Update with new attributes.
1272 * config/arm/fa606te.md: Update with new attributes.
1273 * config/arm/fa626te.md: Update with new attributes.
1274 * config/arm/fa726te.md: Update with new attributes.
1275
1c83b673
JG
12762013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
1277
1278 * config/aarch64/aarch64-simd.md
1279 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
1280 <vwx> iterator to ensure correct register choice.
1281 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1282 (aarch64_sqdmull_n<mode>): Likewise.
1283 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1284 * config/aarch64/arm_neon.h
1285 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
1286 (vml<as><q>_n_<su>16): Likewise.
1287 (vml<as>l_high_lane<q>_<su>16): Likewise.
1288 (vml<as>l_high_n_<su>16): Likewise.
1289 (vml<as>l_lane<q>_<su>16): Likewise.
1290 (vml<as>l_n_<su>16): Likewise.
1291 (vmul<q>_lane<q>_<su>16): Likewise.
1292 (vmul<q>_n_<su>16): Likewise.
1293 (vmull_lane<q>_<su>16): Likewise.
1294 (vmull_n_<su>16): Likewise.
1295 (vmull_high_lane<q>_<su>16): Likewise.
1296 (vmull_high_n_<su>16): Likewise.
1297 (vqrdmulh<q>_n_s16): Likewise.
1298
bb1ae543
TB
12992013-09-06 Tejas Belagod <tejas.belagod@arm.com>
1300
1301 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
1302 have the correct lane parameter.
1303
f23c0742
RB
13042013-09-06 Richard Biener <rguenther@suse.de>
1305
1306 * cfganal.c (control_dependences::~control_dependences):
1307 Properly free all of the vector.
1308
c1d49770
KY
13092013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
1310
1311 PR target/58269
1312 * config/i386/i386.c (ix86_conditional_register_usage):
1313 Proper initialize extended SSE registers.
1314
6da47f52
JH
13152013-09-06 Jan Hubicka <jh@suse.cz>
1316
1317 PR tree-optimization/58311
1318 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
1319
13202013-09-06 Jan Hubicka <jh@suse.cz>
1321
1322 * Makefile.in (tree-sra.o): Update dependencies.
1323 * tree-sra.c: Include ipa-utils.h
1324 (scan_function): Use recursive_call_p.
1325 (has_caller_p): New function.
1326 (cgraph_for_node_and_aliases): Count also callers of aliases.
1327
fc11f321
JH
13282013-09-06 Jan Hubicka <jh@suse.cz>
1329
1330 PR middle-end/58094
1331 * cgraph.h (symtab_semantically_equivalent_p): Declare.
1332 * tree-tailcall.c: Include ipa-utils.h.
1333 (find_tail_calls): Use it.
1334 * ipa-pure-const.c (check_call): Likewise.
1335 * ipa-utils.c (recursive_call_p): New function.
1336 * ipa-utils.h (recursive_call_p): Dclare.
1337 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
1338 (symtab_semantically_equivalent_p): New function.
1339 * Makefile.in (tree-tailcall.o): Update dependencies.
1340
9a6af450
EB
13412013-09-06 Eric Botcazou <ebotcazou@adacore.com>
1342
1343 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
1344 non-inlinable part.
1345
35f32ad4
RB
13462013-09-06 Richard Biener <rguenther@suse.de>
1347
1348 * lto-streamer.h (lto_global_var_decls): Remove.
1349 * Makefile.in (OBJS): Remove lto-symtab.o.
1350 (lto-symtab.o): Remove.
1351 (GTFILES): Remove lto-symtab.c
1352 * lto-symtab.c: Move to lto/
1353
d12a76f3
AK
13542013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1355
1356 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
1357 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
1358 (UNSPEC_FPINT_RINT): New constant definitions.
1359 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
1360 definition with 2 attributes.
1361 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
1362 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
1363 definitions.
1364
13652013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1366
f43245d1 1367 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK
1368 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
1369 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
1370 attribute to "z196_alone".
1371 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
1372 "zEC12_simple".
1373
c8e9d8c3
RB
13742013-09-06 Richard Biener <rguenther@suse.de>
1375
1376 * basic-block.h (class control_dependences): New.
1377 * tree-ssa-dce.c (control_dependence_map): Remove.
1378 (cd): New global.
1379 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
1380 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
1381 find_pdom, find_control_dependence, find_all_control_dependences):
1382 Move to cfganal.c.
f43245d1
UB
1383 (mark_control_dependent_edges_necessary,
1384 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
1385 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB
1386 * cfganal.c (set_control_dependence_map_bit,
1387 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
1388 find_all_control_dependences): Move from tree-ssa-dce.c and
1389 implement as methods of control_dependences class.
1390 (control_dependences::control_dependences): New.
1391 (control_dependences::~control_dependences): Likewise.
1392 (control_dependences::get_edges_dependent_on): Likewise.
1393 (control_dependences::get_edge): Likewise.
1394
7d2ba471
JH
13952013-09-04 Jan Hubicka <jh@suse.cz>
1396
1397 * tree.c (types_same_for_odr): Drop overactive check.
1398 * ipa-devirt.c (hash_type_name): Likewise.
1399
e18412fc
JH
14002013-09-04 Jan Hubicka <jh@suse.cz>
1401
1402 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
1403 (analyze_functions): ... here.
1404
d352b245
JH
14052013-09-04 Jan Hubicka <jh@suse.cz>
1406
1407 PR middle-end/58201
1408 * cgraphunit.c (analyze_functions): Clear AUX fields
1409 after processing; initialize assembler name has.
1410
770da076
JL
14112013-09-05 Jeff Law <law@redhat.com>
1412
1413 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
1414 from thread_around_empty_block. Record threading path into PATH.
1415 Recurse if threading through the initial block is successful.
1416 (thread_across_edge): Corresponding changes to slightly simplify.
1417
fe6f68e2
JG
14182013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1419
1420 * config/aarch64/aarch64.md
1421 (type): Remove frecpe, frecps, frecpx.
1422 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
1423 fix to be a TARGET_SIMD instruction.
1424 (aarch64_frecps): Remove.
1425 * config/aarch64/aarch64-simd.md
1426 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
1427 (aarch64_frecps<mode>): Handle all float/vector of float modes.
1428
6e4150e1
JG
14292013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1430 Sofiane Naci <sofiane.naci@arm.com>
1431
f43245d1 1432 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 1433 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB
1434 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
1435 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
1436 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
1437 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
1438 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
1439 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
1440 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 1441 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 1442 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG
1443 shift* types.
1444 * config/arm/arm-fixed.md: Update for attribute changes
1445 all occurrences of arlo_* types.
1446 * config/arm/thumb2.md: Update for attribute changes all occurrences
1447 of arlo_* types.
1448 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
1449 (cortexa7_older_only): Likewise.
1450 (cortexa7_younger): Likewise.
1451 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
1452 (1020alu_shift_op): Likewise.
1453 (1020alu_shift_reg_op): Likewise.
1454 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
1455 (alu_shift_op): Likewise.
1456 (alu_shift_reg_op): Likewise.
f43245d1 1457 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG
1458 (11_alu_shift_op): Likewise.
1459 (11_alu_shift_reg_op): Likewise.
1460 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
1461 (9_alu_shift_reg_op): Likewise.
1462 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
1463 attribute changes.
1464 (cortex_a15_alu_shift): Likewise.
1465 (cortex_a15_alu_shift_reg): Likewise.
1466 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
1467 attribute changes.
1468 (cortex_a5_alu_shift): Likewise.
f43245d1
UB
1469 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
1470 attribute changes.
6e4150e1 1471 (cortex_a53_alu_shift): Likewise.
f43245d1
UB
1472 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
1473 attribute changes.
6e4150e1
JG
1474 (cortex_a7_alu_reg): Likewise.
1475 (cortex_a7_alu_shift): Likewise.
f43245d1
UB
1476 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
1477 attribute changes.
6e4150e1
JG
1478 (cortex_a8_alu_shift): Likewise.
1479 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 1480 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 1481 (cortex_a9_dp_shift): Likewise.
f43245d1
UB
1482 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
1483 attribute changes.
6e4150e1
JG
1484 * config/arm/cortex-r4.md
1485 (cortex_r4_alu): Update for attribute changes.
1486 (cortex_r4_mov): Likewise.
1487 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 1488 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 1489 (526_alu_shift_op): Likewise.
f43245d1
UB
1490 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
1491 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 1492 (626te_alu_shift_op): Likewise.
f43245d1 1493 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG
1494 (726te_alu_shift_op): Likewise.
1495 (726te_alu_shift_reg_op): Likewise.
1496 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
1497 (mp626_alu_shift_op): Likewise.
1498 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
1499 (pj4_alu_conds): Likewise.
1500 (pj4_shift): Likewise.
1501 (pj4_shift_conds): Likewise.
1502 (pj4_alu_shift): Likewise.
1503 (pj4_alu_shift_conds): Likewise.
1504 * config/aarch64/aarch64.md: Update for attribute change
1505 all occurrences of arlo_* and shift* types.
1506
e6523306
MS
15072013-09-05 Mike Stump <mikestump@comcast.net>
1508
1509 * tree.h: Move documentation for tree_function_decl to tree-core.h
1510 with the declaration.
1511
80a18161
PB
15122013-09-05 Peter Bergner <bergner@vnet.ibm.com>
1513
1514 PR target/58139
1515 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
1516 looking for widest mode.
1517
6994430a
EB
15182013-09-05 Eric Botcazou <ebotcazou@adacore.com>
1519
1520 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
1521
e67f39f7
RB
15222013-09-05 Richard Biener <rguenther@suse.de>
1523
1524 PR tree-optimization/58137
1525 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
1526 Do not create vectors of pointers.
1527 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
1528 types for the components of the vector initializer.
1529 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
1530 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
1531
568cda29
MJ
15322013-09-05 Martin Jambor <mjambor@suse.cz>
1533
1534 * ipa-prop.c (remove_described_reference): Accept missing references,
1535 return false if that hppens, otherwise return true.
1536 (cgraph_node_for_jfunc): New function.
1537 (try_decrement_rdesc_refcount): Likewise.
1538 (try_make_edge_direct_simple_call): Use them.
1539 (ipa_edge_removal_hook): Remove references from rdescs.
1540 (ipa_edge_duplication_hook): Clone rdescs and their references
1541 when the new edge has the same caller as the old one.
1542 * cgraph.c (cgraph_resolve_speculation): Remove speculative
1543 reference before removing any edges.
1544
5a200acb
RE
15452013-09-05 Richard Earnshaw <rearnsha@arm.com>
1546
1547 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
1548 initial store.
1549 * thumb2.md (thumb2_storewb_parisi): New pattern.
1550
5922847b
YZ
15512013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
1552
1553 * config/aarch64/aarch64-option-extensions.def: Add
1554 AARCH64_OPT_EXTENSION of 'crc'.
1555 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
1556 (AARCH64_ISA_CRC): Ditto.
1557 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
1558 description of the CRC extension.
1559
996746aa
AI
15602013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
1561
1562 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
1563 * config/rs6000/linux.h: Ditto.
1564 * alpha/linux.h: Ditto.
1565 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
1566 no_c99_libc_has_function.
1567 * config/c6x/uclinux-elf.h: Ditto.
1568 * config/lm32/uclinux-elf.h: Ditto.
1569 * config/m68k/uclinux.h: Ditto.
1570 * config/moxie/uclinux.h: Ditto.
1571 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
1572 (crisv32-*-linux*, cris-*-linux*): Ditto.
1573 * config/bfin/bfin.c: Include "tm_p.h".
1574
c370d99f
RB
15752013-09-05 Richard Biener <rguenther@suse.de>
1576
1577 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
1578 check for a definition without a basic-block.
1579
003bb7f3
JG
15802013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
1581 Sofiane Naci <sofiane.naci@arm.com>
1582
1583 * config/aarch64/aarch64.md
1584 (*movti_aarch64): Rename r_2_f and f_2_r.
1585 (*movsf_aarch64): Likewise.
1586 (*movdf_aarch64): Likewise.
1587 (*movtf_aarch64): Likewise.
1588 (aarch64_movdi_<mode>low): Likewise.
1589 (aarch64_movdi_<mode>high): Likewise.
1590 (aarch64_mov<mode>high_di): Likewise.
1591 (aarch64_mov<mode>low_di): Likewise.
1592 (aarch64_movtilow_tilow): Likewise.
1593 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
1594 values to config/arm/types.md
1595 (attribute "conds"): Update for attribute change.
1596 (anddi3_insn): Likewise.
1597 (iordi3_insn): Likewise.
1598 (xordi3_insn): Likewise.
1599 (one_cmpldi2): Likewise.
1600 * config/arm/types.md (type): Add Neon types.
1601 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
1602 use "type" attribute.
1603 (movmisalign<mode>_neon_store): Likewise.
1604 (movmisalign<mode>_neon_load): Likewise.
1605 (vec_set<mode>_internal): Likewise.
1606 (vec_setv2di_internal): Likewise.
1607 (vec_extract<mode>): Likewise.
1608 (vec_extractv2di): Likewise.
1609 (add<mode>3_neon): Likewise.
1610 (adddi3_neon): Likewise.
1611 (sub<mode>3_neon): Likewise.
1612 (subdi3_neon): Likewise.
1613 (mul<mode>3_neon): Likewise.
1614 (mul<mode>3add<mode>_neon): Likewise.
1615 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
1616 (fma<VCVTF:mode>4)): Likewise.
1617 (fma<VCVTF:mode>4_intrinsic): Likewise.
1618 (fmsub<VCVTF:mode>4)): Likewise.
1619 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1620 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1621 (ior<mode>3): Likewise.
1622 (and<mode>3): Likewise.
1623 (anddi3_neon): Likewise.
1624 (orn<mode>3_neon): Likewise.
1625 (orndi3_neon): Likewise.
1626 (bic<mode>3_neon): Likewise.
1627 (bicdi3_neon): Likewise.
1628 (xor<mode>3): Likewise.
1629 (one_cmpl<mode>2): Likewise.
1630 (abs<mode>2): Likewise.
1631 (neg<mode>2): Likewise.
1632 (umin<mode>3_neon): Likewise.
1633 (umax<mode>3_neon): Likewise.
1634 (smin<mode>3_neon): Likewise.
1635 (smax<mode>3_neon): Likewise.
1636 (vashl<mode>3): Likewise.
1637 (vashr<mode>3_imm): Likewise.
1638 (vlshr<mode>3_imm): Likewise.
1639 (ashl<mode>3_signed): Likewise.
1640 (ashl<mode>3_unsigned): Likewise.
1641 (neon_load_count): Likewise.
1642 (ashldi3_neon_noclobber): Likewise.
1643 (signed_shift_di3_neon): Likewise.
1644 (unsigned_shift_di3_neon): Likewise.
1645 (ashrdi3_neon_imm_noclobber): Likewise.
1646 (lshrdi3_neon_imm_noclobber): Likewise.
1647 (widen_ssum<mode>3): Likewise.
1648 (widen_usum<mode>3): Likewise.
1649 (quad_halves_<code>v4si): Likewise.
1650 (quad_halves_<code>v4sf): Likewise.
1651 (quad_halves_<code>v8hi): Likewise.
1652 (quad_halves_<code>v16qi): Likewise.
1653 (reduc_splus_v2di): Likewise.
1654 (neon_vpadd_internal<mode>): Likewise.
1655 (neon_vpsmin<mode>): Likewise.
1656 (neon_vpsmax<mode>): Likewise.
1657 (neon_vpumin<mode>): Likewise.
1658 (neon_vpumax<mode>): Likewise.
1659 (ss_add<mode>_neon): Likewise.
1660 (us_add<mode>_neon): Likewise.
1661 (ss_sub<mode>_neon): Likewise.
1662 (us_sub<mode>_neon): Likewise.
1663 (neon_vadd<mode>_unspec): Likewise.
1664 (neon_vaddl<mode>): Likewise.
1665 (neon_vaddw<mode>): Likewise.
1666 (neon_vhadd<mode>): Likewise.
1667 (neon_vqadd<mode>): Likewise.
1668 (neon_vaddhn<mode>): Likewise.
1669 (neon_vmul<mode>): Likewise.
1670 (neon_vmla<mode>): Likewise.
1671 (neon_vmlal<mode>): Likewise.
1672 (neon_vmls<mode>): Likewise.
1673 (neon_vmlsl<mode>): Likewise.
1674 (neon_vqdmulh<mode>): Likewise.
1675 (neon_vqdmlal<mode>): Likewise.
1676 (neon_vqdmlsl<mode>): Likewise.
1677 (neon_vmull<mode>): Likewise.
1678 (neon_vqdmull<mode>): Likewise.
1679 (neon_vsub<mode>_unspec): Likewise.
1680 (neon_vsubl<mode>): Likewise.
1681 (neon_vsubw<mode>): Likewise.
1682 (neon_vqsub<mode>): Likewise.
1683 (neon_vhsub<mode>): Likewise.
1684 (neon_vsubhn<mode>): Likewise.
1685 (neon_vceq<mode>): Likewise.
1686 (neon_vcge<mode>): Likewise.
1687 (neon_vcgeu<mode>): Likewise.
1688 (neon_vcgt<mode>): Likewise.
1689 (neon_vcgtu<mode>): Likewise.
1690 (neon_vcle<mode>): Likewise.
1691 (neon_vclt<mode>): Likewise.
1692 (neon_vcage<mode>): Likewise.
1693 (neon_vcagt<mode>): Likewise.
1694 (neon_vtst<mode>): Likewise.
1695 (neon_vabd<mode>): Likewise.
1696 (neon_vabdl<mode>): Likewise.
1697 (neon_vaba<mode>): Likewise.
1698 (neon_vabal<mode>): Likewise.
1699 (neon_vmax<mode>): Likewise.
1700 (neon_vmin<mode>): Likewise.
1701 (neon_vpaddl<mode>): Likewise.
1702 (neon_vpadal<mode>): Likewise.
1703 (neon_vpmax<mode>): Likewise.
1704 (neon_vpmin<mode>): Likewise.
1705 (neon_vrecps<mode>): Likewise.
1706 (neon_vrsqrts<mode>): Likewise.
1707 (neon_vqabs<mode>): Likewise.
1708 (neon_vqneg<mode>): Likewise.
1709 (neon_vcls<mode>): Likewise.
1710 (clz<mode>2): Likewise.
1711 (popcount<mode>2): Likewise.
1712 (neon_vrecpe): Likewise.
1713 (neon_vrsqrte): Likewise.
1714 (neon_vget_lane<mode>_sext_internal): Likewise.
1715 (neon_vget_lane<mode>_zext_internal): Likewise.
1716 (neon_vdup_n<mode>): Likewise.
1717 (neon_vdup_nv2di): Likewise.
1718 (neon_vdpu_lane<mode>_internal): Likewise.
1719 (neon_vswp<mode>): Likewise.
1720 (float<mode><V_cvtto>2): Likewise.
1721 (floatuns<mode><V_cvtto>2): Likewise.
1722 (fix_trunc<mode><V_cvtto>)2): Likewise
1723 (fixuns_trunc<mode><V_cvtto)2): Likewise.
1724 (neon_vcvt<mode>): Likewise.
1725 (neon_vcvtv4sfv4hf): Likewise.
1726 (neon_vcvtv4hfv4sf): Likewise.
1727 (neon_vcvt_n<mode>): Likewise.
1728 (neon_vmovn<mode>): Likewise.
1729 (neon_vqmovn<mode>): Likewise.
1730 (neon_vqmovun<mode>): Likewise.
1731 (neon_vmovl<mode>): Likewise.
1732 (neon_vmul_lane<mode>): Likewise.
1733 (neon_vmull_lane<mode>): Likewise.
1734 (neon_vqdmull_lane<mode>): Likewise.
1735 (neon_vqdmulh_lane<mode>): Likewise.
1736 (neon_vmla_lane<mode>): Likewise.
1737 (neon_vmlal_lane<mode>): Likewise.
1738 (neon_vqdmlal_lane<mode>): Likewise.
1739 (neon_vmls_lane<mode>): Likewise.
1740 (neon_vmlsl_lane<mode>): Likewise.
1741 (neon_vqdmlsl_lane<mode>): Likewise.
1742 (neon_vext<mode>): Likewise.
1743 (neon_vrev64<mode>): Likewise.
1744 (neon_vrev32<mode>): Likewise.
1745 (neon_vrev16<mode>): Likewise.
1746 (neon_vbsl<mode>_internal): Likewise.
1747 (neon_vshl<mode>): Likewise.
1748 (neon_vqshl<mode>): Likewise.
1749 (neon_vshr_n<mode>): Likewise.
1750 (neon_vshrn_n<mode>): Likewise.
1751 (neon_vqshrn_n<mode>): Likewise.
1752 (neon_vqshrun_n<mode>): Likewise.
1753 (neon_vshl_n<mode>): Likewise.
1754 (neon_vqshl_n<mode>): Likewise.
1755 (neon_vqshlu_n<mode>): Likewise.
1756 (neon_vshll_n<mode>): Likewise.
1757 (neon_vsra_n<mode>): Likewise.
1758 (neon_vsri_n<mode>): Likewise.
1759 (neon_vsli_n<mode>): Likewise.
1760 (neon_vtbl1v8qi): Likewise.
1761 (neon_vtbl2v8qi): Likewise.
1762 (neon_vtbl3v8qi): Likewise.
1763 (neon_vtbl4v8qi): Likewise.
1764 (neon_vtbx1v8qi): Likewise.
1765 (neon_vtbx2v8qi): Likewise.
1766 (neon_vtbx3v8qi): Likewise.
1767 (neon_vtbx4v8qi): Likewise.
1768 (neon_vtrn<mode>_internal): Likewise.
1769 (neon_vzip<mode>_internal): Likewise.
1770 (neon_vuzp<mode>_internal): Likewise.
1771 (neon_vld1<mode>): Likewise.
1772 (neon_vld1_lane<mode>): Likewise.
1773 (neon_vld1_dup<mode>): Likewise.
1774 (neon_vld1_dupv2di): Likewise.
1775 (neon_vst1<mode>): Likewise.
1776 (neon_vst1_lane<mode>): Likewise.
1777 (neon_vld2<mode>): Likewise.
1778 (neon_vld2_lane<mode>): Likewise.
1779 (neon_vld2_dup<mode>): Likewise.
1780 (neon_vst2<mode>): Likewise.
1781 (neon_vst2_lane<mode>): Likewise.
1782 (neon_vld3<mode>): Likewise.
1783 (neon_vld3qa<mode>): Likewise.
1784 (neon_vld3qb<mode>): Likewise.
1785 (neon_vld3_lane<mode>): Likewise.
1786 (neon_vld3_dup<mode>): Likewise.
1787 (neon_vst3<mode>): Likewise.
1788 (neon_vst3qa<mode>): Likewise.
1789 (neon_vst3qb<mode>): Likewise.
1790 (neon_vst3_lane<mode>): Likewise.
1791 (neon_vld4<mode>): Likewise.
1792 (neon_vld4qa<mode>): Likewise.
1793 (neon_vld4qb<mode>): Likewise.
1794 (neon_vld4_lane<mode>): Likewise.
1795 (neon_vld4_dup<mode>): Likewise.
1796 (neon_vst4<mode>): Likewise.
1797 (neon_vst4qa<mode>): Likewise.
1798 (neon_vst4qb<mode>): Likewise.
1799 (neon_vst4_lane<mode>): Likewise.
1800 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1801 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1802 (neon_vec_<US>mult_lo_<mode>): Likewise.
1803 (neon_vec_<US>mult_hi_<mode>): Likewise.
1804 (neon_vec_<US>shiftl_<mode>): Likewise.
1805 (neon_unpack<US>_<mode>): Likewise.
1806 (neon_vec_<US>mult_<mode>): Likewise.
1807 (vec_pack_trunc_<mode>): Likewise.
1808 (neon_vec_pack_trunk_<mode>): Likewise.
1809 (neon_vabd<mode>_2): Likewise.
1810 (neon_vabd<mode>_3): Likewise.
1811 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
1812 (thumb2_movsi_vfp): Likewise.
1813 (movdi_vfp): Likewise.
1814 (movdi_vfp_cortexa8): Likewise.
1815 (movhf_vfp_neon): Likewise.
1816 (movhf_vfp): Likewiwse.
1817 (movsf_vfp): Likewiwse.
1818 (thumb2_movsf_vfp): Likewiwse.
1819 (movdf_vfp): Likewise.
1820 (thumb2_movdf_vfp): Likewise.
1821 (movsfcc_vfp): Likewise.
1822 (thumb2_movsfcc_vfp): Likewise.
1823 (movdfcc_vfp): Likewise.
1824 (thumb2_movdfcc_vfp): Likewise.
1825 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
1826 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
1827 (v10_v2c): Likewise.
1828 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
1829 attribute change.
1830 (cortex_a15_neon_int_2): Likewise.
1831 (cortex_a15_neon_int_3): Likewise.
1832 (cortex_a15_neon_int_4): Likewise.
1833 (cortex_a15_neon_int_5): Likewise.
1834 (cortex_a15_neon_vqneg_vqabs): Likewise.
1835 (cortex_a15_neon_vmov): Likewise.
1836 (cortex_a15_neon_vaba): Likewise.
1837 (cortex_a15_neon_vaba_qqq): Likewise.
1838 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1839 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1840 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
1841 scalar_64_32_long_scalar): Likewise.
1842 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1843 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1844 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1845 lotype_qdd_64_32_long): Likewise.
1846 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1847 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1848 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1849 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1850 (cortex_a15_neon_shift_1): Likewise.
1851 (cortex_a15_neon_shift_2): Likewise.
1852 (cortex_a15_neon_shift_3): Likewise.
1853 (cortex_a15_neon_vshl_ddd): Likewise.
1854 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1855 (cortex_a15_neon_vsra_vrsra): Likewise.
1856 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1857 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1858 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1859 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1860 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1861 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1862 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1863 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1864 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1865 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1866 (cortex_a15_neon_bp_simple): Likewise.
1867 (cortex_a15_neon_bp_2cycle): Likewise.
1868 (cortex_a15_neon_bp_3cycle): Likewise.
1869 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1870 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1871 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1872 (cortex_a15_neon_vld2_4_regs): Likewise.
1873 (cortex_a15_neon_vld3_vld4): Likewise.
1874 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1875 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1876 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
1877 (cortex_a15_neon_vst3_vst4): Likewise.
1878 (cortex_a15_neon_vld1_vld2_lane): Likewise.
1879 (cortex_a15_neon_vld3_vld4_lane" 10
1880 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1881 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1882 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1883 (cortex_a15_neon_ldm_2): Likewise.0
1884 (cortex_a15_neon_stm_2): Likewise.
1885 (cortex_a15_neon_mcr): Likewise.
1886 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1887 (cortex_a15_neon_mrc): Likewise.
1888 (cortex_a15_neon_mrrc): Likewise.
1889 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
1890 change.
1891 (cortex_a15_alu_shift): Likewise.
1892 (cortex_a15_alu_shift_reg): Likewise.
1893 (cortex_a15_mult32): Likewise.
1894 (cortex_a15_mult64): Likewise.
1895 (cortex_a15_block): Likewise.
1896 (cortex_a15_branch): Likewise.
1897 (cortex_a15_load1): Likewise.
1898 (cortex_a15_load3): Likewise.
1899 (cortex_a15_store1): Likewise.
1900 (cortex_a15_store3): Likewise.
1901 (cortex_a15_call): Likewise.
f43245d1 1902 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG
1903 (cortex_a5_f2r): Likewise.
1904 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
1905 change.
1906 (cortex_a53_f2r): Likewise.
1907 * config/arm/cortex-a7.md
1908 (cortex_a7_branch): Update for attribute change.
1909 (cortex_a7_call): Likewise.
1910 (cortex_a7_alu_imm): Likewise.
1911 (cortex_a7_alu_reg): Likewise.
1912 (cortex_a7_alu_shift): Likewise.
1913 (cortex_a7_mul): Likewise.
1914 (cortex_a7_load1): Likewise.
1915 (cortex_a7_store1): Likewise.
1916 (cortex_a7_load2): Likewise.
1917 (cortex_a7_store2): Likewise.
1918 (cortex_a7_load3): Likewise.
1919 (cortex_a7_store3): Likewise.
1920 (cortex_a7_load4): Likewise.
1921 (cortex_a7_store4): Likewise.
1922 (cortex_a7_fpalu): Likewise.
1923 (cortex_a7_fconst): Likewise.
1924 (cortex_a7_fpmuls): Likewise.
1925 (cortex_a7_neon_mul): Likewise.
1926 (cortex_a7_fpmacs): Likewise.
1927 (cortex_a7_neon_mla: Likewise.
1928 (cortex_a7_fpmuld: Likewise.
1929 (cortex_a7_fpmacd: Likewise.
1930 (cortex_a7_fpfmad: Likewise.
1931 (cortex_a7_fdivs: Likewise.
1932 (cortex_a7_fdivd: Likewise.
1933 (cortex_a7_r2f: Likewise.
1934 (cortex_a7_f2r: Likewise.
1935 (cortex_a7_f_flags: Likewise.
1936 (cortex_a7_f_loads: Likewise.
1937 (cortex_a7_f_loadd: Likewise.
1938 (cortex_a7_f_stores: Likewise.
1939 (cortex_a7_f_stored: Likewise.
1940 (cortex_a7_neon): Likewise.
1941 * config/arm/cortex-a8-neon.md
1942 (cortex_a8_neon_mrc): Update for attribute change.
1943 (cortex_a8_neon_mrrc): Likewise.
1944 (cortex_a8_neon_int_1): Likewise.
1945 (cortex_a8_neon_int_2): Likewise.
1946 (cortex_a8_neon_int_3): Likewise.
1947 (cortex_a8_neon_int_4): Likewise.
1948 (cortex_a8_neon_int_5): Likewise.
1949 (cortex_a8_neon_vqneg_vqabs): Likewise.
1950 (cortex_a8_neon_vmov): Likewise.
1951 (cortex_a8_neon_vaba): Likewise.
1952 (cortex_a8_neon_vaba_qqq): Likewise.
1953 (cortex_a8_neon_vsma): Likewise.
1954 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1955 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB
1956 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
1957 long_scalar): Likewise.
003bb7f3
JG
1958 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1959 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1960 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
1961 long_scalar_qdd_64_32_long): Likewise.
1962 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1963 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1964 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1965 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1966 (cortex_a8_neon_shift_1): Likewise.
1967 (cortex_a8_neon_shift_2): Likewise.
1968 (cortex_a8_neon_shift_3): Likewise.
1969 (cortex_a8_neon_vshl_ddd): Likewise.
1970 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1971 (cortex_a8_neon_vsra_vrsra): Likewise.
1972 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1973 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1974 (cortex_a8_neon_fp_vsum): Likewise.
1975 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1976 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1977 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1978 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1979 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1980 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1981 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1982 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1983 (cortex_a8_neon_bp_simple): Likewise.
1984 (cortex_a8_neon_bp_2cycle): Likewise.
1985 (cortex_a8_neon_bp_3cycle): Likewise.
1986 (cortex_a8_neon_ldr): Likewise.
1987 (cortex_a8_neon_str): Likewise.
1988 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1989 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1990 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1991 (cortex_a8_neon_vld2_4_regs): Likewise.
1992 (cortex_a8_neon_vld3_vld4): Likewise.
1993 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1994 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1995 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1996 (cortex_a8_neon_vst3_vst4): Likewise.
1997 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1998 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1999 (cortex_a8_neon_vst1_vst2_lane): Likewise.
2000 (cortex_a8_neon_vst3_vst4_lane): Likewise.
2001 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
2002 (cortex_a8_neon_mcr): Likewise.
2003 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 2004 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG
2005 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
2006 change.
2007 (ca9_neon_mrrc): Likewise.
2008 (cortex_a9_neon_int_1): Likewise.
2009 (cortex_a9_neon_int_2): Likewise.
2010 (cortex_a9_neon_int_3): Likewise.
2011 (cortex_a9_neon_int_4): Likewise.
2012 (cortex_a9_neon_int_5): Likewise.
2013 (cortex_a9_neon_vqneg_vqabs): Likewise.
2014 (cortex_a9_neon_vmov): Likewise.
2015 (cortex_a9_neon_vaba): Likewise.
2016 (cortex_a9_neon_vaba_qqq): Likewise.
2017 (cortex_a9_neon_vsma): Likewise.
2018 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2019 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB
2020 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
2021 long_scalar): Likewise.
003bb7f3
JG
2022 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
2023 (cortex_a9_neon_mla_qqq_8_16): Likewise.
2024 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
2025 long_scalar_qdd_64_32_long): Likewise.
2026 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
2027 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
2028 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
2029 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
2030 (cortex_a9_neon_shift_1): Likewise.
2031 (cortex_a9_neon_shift_2): Likewise.
2032 (cortex_a9_neon_shift_3): Likewise.
2033 (cortex_a9_neon_vshl_ddd): Likewise.
2034 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
2035 (cortex_a9_neon_vsra_vrsra): Likewise.
2036 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
2037 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
2038 (cortex_a9_neon_fp_vsum): Likewise.
2039 (cortex_a9_neon_fp_vmul_ddd): Likewise.
2040 (cortex_a9_neon_fp_vmul_qqd): Likewise.
2041 (cortex_a9_neon_fp_vmla_ddd): Likewise.
2042 (cortex_a9_neon_fp_vmla_qqq): Likewise.
2043 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
2044 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
2045 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
2046 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
2047 (cortex_a9_neon_bp_simple): Likewise.
2048 (cortex_a9_neon_bp_2cycle): Likewise.
2049 (cortex_a9_neon_bp_3cycle): Likewise.
2050 (cortex_a9_neon_ldr): Likewise.
2051 (cortex_a9_neon_str): Likewise.
2052 (cortex_a9_neon_vld1_1_2_regs): Likewise.
2053 (cortex_a9_neon_vld1_3_4_regs): Likewise.
2054 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
2055 (cortex_a9_neon_vld2_4_regs): Likewise.
2056 (cortex_a9_neon_vld3_vld4): Likewise.
2057 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
2058 (cortex_a9_neon_vst1_3_4_regs): Likewise.
2059 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
2060 (cortex_a9_neon_vst3_vst4): Likewise.
2061 (cortex_a9_neon_vld1_vld2_lane): Likewise.
2062 (cortex_a9_neon_vld3_vld4_lane): Likewise.
2063 (cortex_a9_neon_vst1_vst2_lane): Likewise.
2064 (cortex_a9_neon_vst3_vst4_lane): Likewise.
2065 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
2066 (cortex_a9_neon_mcr): Likewise.
2067 (cortex_a9_neon_mcr_2_mcrr): Likewise.
2068 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
2069 (cortex_a9_fps): Likewise.
2070 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
2071 change.
2072 (cortex_m4_fmuls): Likewise.
2073 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
2074 change.
2075 (cortex_r4_mrc): Likewise.
2076 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 2077 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 2078 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB
2079 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
2080 attribute change.
003bb7f3
JG
2081 (pj4_core_to_vfp): Likewise.
2082 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
2083 attribute change.
2084 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
2085 (vfp_fstore): Likewise.
2086 * doc/md.texi: Change references to neon_type to refer to type.
2087
fbe0dc78
DS
20882013-09-04 Dodji Seketeli <dodji@redhat.com>
2089
2090 * tree.h (DECL_BUILT_IN): Fix typo in comment.
2091
c167bc5b
DE
20922013-09-04 David Edelsohn <dje.gcc@gmail.com>
2093
2094 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
2095 lglobl if not weak.
2096
035cb59f
ER
20972013-09-04 Easwaran Raman <eraman@google.com>
2098
2099 PR middle-end/57370
2100 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
2101 (build_and_add_sum): Use it.
2102 (appears_later_in_bb): Simplify code.
2103
71191083
TJ
21042013-09-04 Teresa Johnson <tejohnson@google.com>
2105
2106 * dumpfile.c (dump_finish): Don't close stderr/stdout.
2107
0c488213
JG
21082013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
2109
2110 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
2111
61a74079
JH
21122013-09-04 Jan Hubicka <jh@suse.cz>
2113
2114 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
2115 * ipa-devirt.c: Include diganostic.h
2116 (odr_type_d): Add types and types_set.
2117 (hash_type_name): Work for types with vtables during LTO.
2118 (odr_hasher::remove): Fix comment; destroy types_set.
2119 (add_type_duplicate): New function,
2120 (get_odr_type): Use it.
2121 (dump_type_inheritance_graph): Dump type duplicates.
2122 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
2123 graph.
2124 * tree.c (types_same_for_odr): Give exact answers on types with
2125 virtual tables.
2126
24ba8195
DS
21272013-09-04 Dodji Seketeli <dodji@redhat.com>
2128
2129 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
2130 explaining their differences.
2131
0a04869e
SKS
21322013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
2133
2134 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
2135
3b18bc42
JL
21362013-09-03 Jeff Law <law@redhat.com>
2137
2138 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
2139 when not threading through a joiner block. Pass joiner/no joiner
2140 state to register_jump_thread.
2141 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
2142 state from argument rather than implying on path length.
2143 Dump the entire jump thread path into debugging dump.
2144 * tree-flow.h (register_jump_thread): Update prototype.
2145
ae84e151
XDL
21462013-08-29 Xinliang David Li <davidxl@google.com>
2147
2148 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2149 Remove a trivial gcc_assert.
2150
c716e67f
XDL
21512013-08-29 Xinliang David Li <davidxl@google.com>
2152
2153 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
2154 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
2155 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
2156 Delay base decl alignment adjustment.
2157 * tree-vectorizer.c (vect_destroy_datarefs): New function.
2158 * tree-vectorizer.h: New data structure.
2159 (set_dr_misalignment): New function.
2160 (dr_misalignment): Ditto.
2161 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
2162 (vectorizable_load): Ditto.
2163 (ensure_base_align): New function.
2164 (vectorize_loops): Add dbg_cnt support.
2165 (execute_vect_slp): Ditto.
2166 * dbgcnt.def: New debug counter.
2167 * Makefile: New dependency.
2168
db8800bc
MI
21692013-09-03 Meador Inge <meadori@codesourcery.com>
2170
2171 Revert:
2172
2173 2013-08-30 Meador Inge <meadori@codesourcery.com>
2174
2175 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2176
0d9b4f5c
DE
21772013-09-03 David Edelsohn <dje.gcc@gmail.com>
2178
2179 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
2180 function descriptor.
2181
29e810b9
RB
21822013-09-03 Richard Biener <rguenther@suse.de>
2183
2184 * tree-affine.c (add_elt_to_tree): Fix association issue,
2185 avoid useless converts and make sure to always return a
2186 properly typed result.
2187
dbfc2894
RB
21882013-09-03 Richard Biener <rguenther@suse.de>
2189
2190 PR middle-end/57656
2191 * fold-const.c (negate_expr_p): Fix division case.
2192 (negate_expr): Likewise.
2193
fa3bf4ce
RB
21942013-09-03 Richard Biener <rguenther@suse.de>
2195
2196 PR lto/58285
2197 * tree-streamer-out.c: Include tm.h.
2198 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
2199
a64bbb3f
JH
22002013-09-03 Jan Hubicka <jh@suse.cz>
2201
2202 * tree-profile.c (tree_profiling): Cleanup CFG when done.
2203
93ed07e2
AM
22042013-09-03 Alan Modra <amodra@gmail.com>
2205
2206 * config.gcc (powerpc*-*-linux*): Add support for little-endian
2207 multilibs to big-endian target and vice versa.
2208 * config/rs6000/t-linux64: Use := assignment on all vars.
2209 (MULTILIB_EXTRA_OPTS): Remove fPIC.
2210 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
2211 * config/rs6000/t-linux64le: New file.
2212 * config/rs6000/t-linux64bele: New file.
2213 * config/rs6000/t-linux64lebe: New file.
2214
07d6cd64
JH
22152013-09-02 Jan Hubicka <jh@suse.cz>
2216
2217 * ipa-inline-transform.c (inline_transform): Do not
2218 optimize_inline_calls when not optimizing.
2219
8e4c9a10
JH
22202013-09-02 Jan Hubicka <jh@suse.cz>
2221
2222 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
2223 duplicated nodes for assembler names.
2224 * symtab.c (symtab_unregister_node): Do not attempt to unlink
2225 hard registers from assembler name hash.
2226
c91061e6
JH
22272013-09-02 Jan Hubicka <jh@suse.cz>
2228
2229 * ipa-split.c (execute_split_functions): Split externally visible
2230 functions called once.
2231
2fd0985c
MJ
22322013-09-02 Martin Jambor <mjambor@suse.cz>
2233
2234 PR ipa/58106
2235 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
2236 linked list. When finding the correct duplicate, also consider also
2237 the caller in additon to its inlined_to node.
2238
91bd4114
JG
22392013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
2240
2241 * config/aarch64/aarch64-simd-builtins.def
2242 (dup_lane_scalar): Remove.
2243 * config/aarch64/aarch64-simd.md
2244 (aarch64_simd_dup): Add 'w->w' alternative.
2245 (aarch64_dup_lane<mode>): Allow for VALL.
2246 (aarch64_dup_lane_scalar<mode>): Remove.
2247 (aarch64_dup_lane_<vswap_width_name><mode>): New.
2248 (aarch64_get_lane_signed<mode>): Add w->w altenative.
2249 (aarch64_get_lane_unsigned<mode>): Likewise.
2250 (aarch64_get_lane<mode>): Likewise.
2251 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
2252 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
2253 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
2254 (VCON): Change container of V2SF.
2255 (vswap_width_name): Likewise.
2256 * config/aarch64/arm_neon.h
2257 (__aarch64_vdup_lane_any): New.
2258 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2259 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
2260 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
2261
d617d2d8
EB
22622013-09-02 Eric Botcazou <ebotcazou@adacore.com>
2263
2264 PR middle-end/56382
2265 * expr.c (emit_move_complex): Do not move complex FP values as parts if
2266 the source or the destination is a single hard register.
2267
63fdb7be
RB
22682013-09-02 Richard Biener <rguenther@suse.de>
2269
2270 PR middle-end/57511
2271 * tree-scalar-evolution.c (instantiate_scev_name): Allow
2272 non-linear SCEVs.
2273
78de2333
RB
22742013-09-02 Richard Biener <rguenther@suse.de>
2275
2276 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
2277 arithmetic to sizetype.
2278
85ff4ec6
BC
22792013-09-02 Bin Cheng <bin.cheng@arm.com>
2280
2281 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
2282 Find auto-increment use both before and after candidate.
2283
576016fe
MP
22842013-09-02 Marek Polacek <polacek@redhat.com>
2285
2286 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
2287
08f835dc
JH
22882013-09-01 Jan Hubicka <jh@suse.cz>
2289
2290 * Makefile.in: Add ipa-profile.o
2291 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
2292 * cgraph.c (struct cgraph_propagate_frequency_data,
2293 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
2294 ipa-profile.c; replace cgraph_ by ipa_ prefix.
2295 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB
2296 * ipa-inline-analysis.c: Include ipa-utils.h;
2297 drop duplicated cfgloop.h.
08f835dc
JH
2298 (inline_update_callee_summaries): Update.
2299 * ipa-profile.c: New file.
2300 * ipa-utils.h (ipa_propagate_frequency): Declare.
2301 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 2302 data-streamer.h, value-prof.h.
08f835dc
JH
2303 (symtab_remove_unreachable_nodes): Update profile.
2304 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
2305 account_time_size, cmp_counts, dump_histogram,
2306 ipa_profile_generate_summary, ipa_profile_write_summary,
2307 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
2308 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 2309 Move to ipa-profile.c.
08f835dc 2310
cec8583c
JDA
23112013-09-01 John David Anglin <danglin@gcc.gnu.org>
2312
2313 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
2314
bbc9396b
JH
23152013-09-01 Jan Hubicka <jh@suse.cz>
2316
2317 * common.opt (fdevirtualize-speculatively): New function.
2318 * invoke.texi (fdevirtualize-speculatively): Document.
2319 * ipa-devirt.c: Include ipa-inline.h
2320 (likely_target_p): New function.
2321 (ipa_devirt): New function.
2322 (gate_ipa_devirt): New function.
2323 (pass_data_ipa_devirt): New static var.
2324 (pass_ipa_devirt): Likewise.
2325 (make_pass_ipa_devirt): New function.
2326 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
2327 (common_handle_option): Disable devirtualization when
2328 value range profiling is available.
2329 * passes.def (pass_ipa_devirt): Add.
2330 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 2331 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 2332
0cea211e
IS
23332013-09-01 Iain Sandoe <iain@codesourcery.com>
2334
2335 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
2336 include sanitize(undefined).
2337
5e351e96
DN
23382013-08-31 Diego Novillo <dnovillo@google.com>
2339
2340 * Makefile.in (TREE_CORE_H): Define.
2341 (TREE_H): Use.
2342 (GTFILES): Add tree-core.h.
2343 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
2344 size the array.
2345 * tree-core.h: New file.
2346 Move all data structures, enum, typedefs, global
2347 declarations and constants from ...
2348 * tree.h: ... here.
2349
63bf9a90
JH
23502013-08-31 Jan Hubicka <jh@suse.cz>
2351
2352 * bulitins.c (expand_builtin): Do not early exit for gcov
2353 instrumented functions.
2354
5c0abd6a
MP
23552013-08-31 Marek Polacek <polacek@redhat.com>
2356
2357 * ubsan.c: Include tm_p.h.
2358
4843f032
JH
23592013-08-31 Jan Hubicka <jh@suse.cz>
2360
8356c89c
JH
2361 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
2362 warning.
2363
4843f032
JH
2364 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
2365 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 2366 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH
2367 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
2368 (ipa_merge_profiles): New function.
2369 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
2370 (lto_input_function_body): Likewise.
5e581212
JH
2371 * ipa-utils.h (ipa_merge_profiles): Declare.
2372 * lto-streamer.h (lto_input_function_body): Update prototype.
2373 (emit_label_in_global_context_p): Remove.
2374 * lto-symtab.c: Include ipa-utils.h
2375 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 2376
57292ce9
JH
23772013-08-31 Jan Hubicka <jh@suse.cz>
2378
2379 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
2380
b9cbfeeb
JH
23812013-08-31 Jan Hubicka <jh@suse.cz>
2382
2383 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
2384
41dedebd
UB
23852013-08-31 Uros Bizjak <ubizjak@gmail.com>
2386
2387 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
2388 "cmp" RTX before signed_comparison_operator check to account
2389 for "code" changes.
2390
06d65050
JH
23912013-08-30 Jan Hubicka <jh@suse.cz>
2392
41dedebd 2393 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH
2394 (detect_type_change_1): Rename to ...
2395 (detect_type_change): ... this one; early return on non-polymorphic
2396 types.
41dedebd 2397 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH
2398 use of detect_type_change.
2399 (compute_complex_assign_jump_func): Add param_type parameter;
2400 update use of detect_type_change_ssa.
2401 (compute_complex_ancestor_jump_func): Likewise.
2402 (ipa_get_callee_param_type): New function.
2403 (ipa_compute_jump_functions_for_edge): Compute parameter type;
2404 update calls to the jump function computation functions.
2405
600b5b1d
TJ
24062013-08-30 Teresa Johnson <tejohnson@google.com>
2407 Steven Bosscher <steven@gcc.gnu.org>
2408
2409 * cfgrtl.c (fixup_new_cold_bb): New routine.
2410 (commit_edge_insertions): Invoke fixup_partitions.
2411 (find_partition_fixes): New routine.
2412 (fixup_partitions): Ditto.
2413 (verify_hot_cold_block_grouping): Update comments.
2414 (rtl_verify_edges): Invoke find_partition_fixes.
2415 (rtl_verify_bb_pointers): Update comments.
2416 (rtl_verify_bb_layout): Ditto.
2417 * basic-block.h (probably_never_executed_edge_p): Declare.
2418 (fixup_partitions): Ditto.
2419 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
2420 * bb-reorder.c (sanitize_hot_paths): New function.
2421 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
2422 sanitize_hot_paths.
2423 * predict.c (probably_never_executed_edge_p): New routine.
2424 * cfg.c (check_bb_profile): Add partition insanity warnings.
2425
0a6ea5c9
MI
24262013-08-30 Meador Inge <meadori@codesourcery.com>
2427
2428 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
2429
de5a5fa1
MP
24302013-08-30 Marek Polacek <polacek@redhat.com>
2431
2432 * Makefile.in (ubsan.o): Add.
2433 (c-family/c-ubsan.o): Add.
2434 (builtins.o): Add ubsan.h dependency.
2435 * ubsan.h: New file.
2436 * ubsan.c: New file.
2437 * common.opt: Add -fsanitize=undefined option.
2438 (flag_sanitize): Add variable.
2439 (fsanitize=): Add option. Add Driver.
2440 (fsanitize=thread): Remove option.
2441 (fsanitize=address): Likewise.
2442 (static-libubsan): New option.
2443 * doc/invoke.texi: Document the new flag and -static-libubsan.
2444 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
2445 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
2446 * builtin-attrs.def (ATTR_COLD): Define.
2447 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2448 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
2449 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
2450 * flag-types.h (sanitize_code): New enum.
2451 * opts.c (common_handle_option): Parse command line arguments
2452 of -fsanitize=. Add -fsanitize=unreachable option.
2453 * varasm.c (get_variable_section): Adjust.
2454 (assemble_noswitch_variable): Likewise.
2455 (assemble_variable): Likewise.
2456 (output_constant_def_contents): Likewise.
2457 (categorize_decl_for_section): Likewise.
2458 (place_block_symbol): Likewise.
2459 (output_object_block): Likewise.
2460 * builtins.def: Likewise.
2461 * toplev.c (compile_file): Likewise.
2462 (process_options): Likewise.
2463 * cppbuiltin.c: Likewise.
2464 * tsan.c (tsan_pass): Likewise.
2465 (tsan_gate): Likewise.
2466 (tsan_gate_O0): Likewise.
2467 * cfgexpand.c (partition_stack_vars): Likewise.
2468 (expand_stack_vars): Likewise.
2469 (defer_stack_allocation): Likewise.
2470 (expand_used_vars): Likewise.
2471 * cfgcleanup.c (old_insns_match_p): Likewise.
2472 * asan.c (asan_finish_file): Likewise.
2473 (asan_instrument): Likewise.
2474 (gate_asan): Likewise.
2475 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
2476 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
2477 (asan_global_struct): Use pointer_sized_int_node instead
2478 calling build_nonstandard_integer_type.
2479 (initialize_sanitizer_builtins): Likewise.
2480 (asan_finish_file): Likewise.
2481 * gcc.c: Document %{%:function(args):X}.
2482 (static_spec_functions): Add sanitize.
2483 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
2484 store funcval != NULL there.
2485 (do_spec_1): Adjust handle_spec_function caller.
2486 (handle_braces): Allow %:function(args) as condition.
2487 (sanitize_spec_function): New function.
2488 (ADD_STATIC_LIBUBSAN_LIBS): Define.
2489 (LIBUBSAN_SPEC): Likewise.
2490 (LIBUBSAN_EARLY_SPEC): Likewise.
2491 (SANITIZER_SPEC): Handle libubsan.
2492 (SANITIZER_EARLY_SPEC): Likewise.
2493 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
2494 instead of fsanitize=address.
2495 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
2496 instead of fsanitize=address*.
2497 * builtins.c: Include ubsan.h.
2498 (fold_builtin_0): Instrument __builtin_unreachable.
2499 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
2500 instead of flag_asan.
2501 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
2502 (pointer_sized_int_node): Define.
41dedebd 2503 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 2504
f07f30cf
MS
25052013-08-30 Mike Stump <mikestump@comcast.net>
2506
2507 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
2508 with RE patterns.
2509
8a41354f
JH
25102013-08-29 Jan Hubicka <jh@suse.cz>
2511
2512 * cgraph.c (cgraph_function_body_availability): Handle weakref
2513 correctly.
2514 * passes.def: Remove pass_fixup_cfg.
2515 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
2516 track when we need to remove functions.
2517 (gate_ipa_inline): Execute inlining always; add comment why.
2518 (pass_data_ipa_inline): Remove TODO_remove_functions.
2519 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
2520 do not produce summaries.
2521 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
2522 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
2523 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
2524 constant pool and vtable.
2525
dc8d7a0f
TB
25262013-08-30 Tejas Belagod <tejas.belagod@arm.com>
2527
41dedebd
UB
2528 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
2529 New arm_neon.h's internal macros to specify 64-bit constants.
2530 Avoid using stdint.h's macros.
dc8d7a0f 2531
ac7eacd2
JR
25322013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
2533
2534 * recog.c (verify_changes): Verify that changes[i].old is non-zero
2535 before applying REG_P.
2536
8b29fd4e
JJ
25372013-08-30 Jakub Jelinek <jakub@redhat.com>
2538
2539 PR tree-optimization/58277
2540 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
2541 after seeing too many stmts with vdef in between dombb and current
2542 bb, invalidate everything.
2543
26d75703
RB
25442013-08-30 Richard Biener <rguenther@suse.de>
2545
2546 * fold-const.c (fold_single_bit_test): Fix overflow test.
2547
f7e088e7
EB
25482013-08-30 Eric Botcazou <ebotcazou@adacore.com>
2549
2550 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
2551 and which can live in a register, always retrieve the value on entry.
2552 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
2553 passed by invisible reference specially.
2554 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
2555 (vt_add_function_parameter): Correctly deal with a parameter passed by
2556 invisible reference.
2557
25582013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH
2559
2560 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
2561
6e8dad05
RB
25622013-08-30 Richard Biener <rguenther@suse.de>
2563
2564 PR tree-optimization/58228
2565 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
2566 allow invariant loads in nested loop vectorization.
2567
062ef2c8
RB
25682013-08-30 Richard Biener <rguenther@suse.de>
2569
2570 PR tree-optimization/58223
2571 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
2572 (has_anti_or_output_dependence): ... this and adjust to also
2573 look for output dependences.
2574 (mark_nodes_having_upstream_mem_writes): Adjust.
2575 (rdg_flag_uses): Likewise.
2576
7a764c60
RB
25772013-08-30 Richard Biener <rguenther@suse.de>
2578
2579 PR tree-optimization/58010
2580 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
2581 assert that we have a loop-closed PHI.
2582
82e9d642
JH
25832013-08-29 Jan Hubicka <jh@suse.cz>
2584
2585 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
2586 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
2587 * lto-section-in.c (lto_free_function_in_decl_state): New function.
2588 (lto_free_function_in_decl_state_for_node): New function.
2589
9cc1fb4b
XDL
25902013-08-29 Xinliang David Li <davidxl@google.com>
2591
41dedebd 2592 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL
2593 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
2594 Emit alignment peeling message with default -fopt-info.
2595 (vect_loop_versioning): Emit loop version info message.
41dedebd 2596 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL
2597 (execute_vect_slp): Ditto.
2598
440a5082
EB
25992013-08-29 Eric Botcazou <ebotcazou@adacore.com>
2600
2601 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
2602 of the clone from the DECL_NAME of the original function.
2603
3fa3690d
OE
26042013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
2605
2606 * passes.c (register_pass): Add overload.
2607 * tree-pass.h (register_pass): Forward declare it. Add comment.
2608
0170f33c
JH
26092013-08-29 Jan Hubicka <jh@suse.cz>
2610
41dedebd
UB
2611 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
2612 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
2613 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
2614 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
2615 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
2616 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 2617 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
2618 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
2619
103ff0d6
TJ
26202013-08-29 Teresa Johnson <tejohnson@google.com>
2621
2622 * dumpfile.c (dump_loc): Output column number.
2623 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
2624 * doc/invoke.texi: Document optall -fopt-info flag.
2625 * profile.c (read_profile_edge_counts): Use new dump framework.
2626 (compute_branch_probabilities): Ditto.
2627 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
2628 when pass not in any opt group.
2629 * pass_manager.h (pass_manager::get_pass_profile): New method.
2630 * value-prof.c (check_counter): Use new dump framework.
2631 (check_ic_target): Ditto.
2632 * coverage.c (get_coverage_counts): Ditto.
2633 (coverage_init): Setup new dump framework.
2634
301bbc16
RB
26352013-08-29 Richard Biener <rguenther@suse.de>
2636
2637 PR tree-optimization/58246
2638 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
2639 handle the dominance check inside a basic-block.
2640
1b275000
RB
26412013-08-29 Richard Biener <rguenther@suse.de>
2642
2643 PR middle-end/57287
2644 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
2645 of default defs that appear in abnormal PHI nodes.
2646
6b1184ba
RB
26472013-08-29 Richard Biener <rguenther@suse.de>
2648
2649 PR tree-optimization/57685
2650 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
2651 single-use operands to avoid exponential complexity.
2652
1b08b734
DC
26532013-08-28 Dehao Chen <dehao@google.com>
2654
2655 * ipa-inline.c (edge_badness): Fix integer underflow.
2656
48a3fa69
UB
26572013-08-28 Uros Bizjak <ubizjak@gmail.com>
2658
2659 * gtm-builtins.def (_ITM_free): Declare leaf.
2660
215f73e6
JJ
26612013-08-28 Jakub Jelinek <jakub@redhat.com>
2662
2663 PR target/58067
2664 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
2665 (*tls_local_dynamic_base_64_largepic): Likewise.
2666 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
2667 Remove predicate from call operand.
2668 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
2669 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
2670
34554d1a
JL
26712013-08-28 Jeff Law <law@redhat.com>
2672
2673 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
2674 checks for the number of predecessors and successors allowed.
2675 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
2676 which require copying a joiner block if there is a request which
2677 is a subpath that requires no joiner block copying.
2678
c01c111b
JH
26792013-08-28 Jan Hubicka <jh@suse.cz>
2680
2681 * lto-streamer-out.c (DFS_write_tree_body): Drop
2682 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
2683 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
2684 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
2685 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
2686 Do not read DECL_ERROR_ISSUED.
2687 (unpack_ts_decl_with_vis_value_fields): Do not read
2688 DECL_DEFER_OUTPUT.
48a3fa69
UB
2689 (lto_input_ts_binfo_tree_pointers): Do not read
2690 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH
2691 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
2692 write DECL_ERROR_ISSUED..
2693 (pack_ts_decl_with_vis_value_fields): Do not write
2694 DECL_DEFER_OUTPUT.
2695 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 2696 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH
2697 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
2698 * tree.h (tree_decl_common): Update comment.
2699 (DECL_ERROR_ISSUED): Remove.
2700
4b128ece
JJ
27012013-08-28 Jakub Jelinek <jakub@redhat.com>
2702
2703 PR middle-end/58257
2704 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
2705
12211b99 27062013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH
2707
2708 * builtins.def (free): Declare leaf.
2709
8dce4dbc
DM
27102013-08-27 David Malcolm <dmalcolm@redhat.com>
2711
2712 * gdbhooks.py: New.
2713 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
2714 * configure: Regenerate.
2715
b8f6e610
MJ
27162013-08-27 Martin Jambor <mjambor@suse.cz>
2717
2718 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
2719 (ipa_ancestor_jf_data): Likewise.
2720 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
2721 (ipa_get_jf_pass_through_type_preserved): New function.
2722 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
2723 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 2724 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ
2725 (ipa_get_jf_ancestor_result): Likewise.
2726 (propagate_vals_accross_pass_through): Use
2727 ipa_get_jf_pass_through_result to do all the value mappings.
2728 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
2729 type_preserved flag.
2730 (ipa_set_jf_cst_copy): New function.
2731 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
2732 (ipa_set_jf_arith_pass_through): Likewise.
2733 (ipa_set_ancestor_jf): Likewise.
2734 (compute_complex_assign_jump_func): Set type_preserved instead of
2735 punting.
2736 (ipa_compute_jump_functions_for_edge): Likewise.
2737 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
2738 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 2739 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ
2740 (ipa_write_jump_function): Stream the type_preserved flags.
2741 (ipa_read_jump_function): Likewise.
2742
74bf76ed
JJ
27432013-08-27 Jakub Jelinek <jakub@redhat.com>
2744 Aldy Hernandez <aldyh@redhat.com>
2745
2746 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
2747 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
2748 * function.h (struct function): Add has_force_vect_loops and
2749 has_simduid_loops.
12211b99 2750 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ
2751 * gimple.c (gimple_build_omp_critical): Add KIND argument and
2752 handle it.
2753 * gimple.def: Update CLAUSES comments.
2754 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
2755 (gimple_build_omp_for): Add argument to prototype.
2756 (gimple_omp_for_kind): New.
2757 (gimple_omp_for_set_kind): New.
2758 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
2759 GOVD_DATA_SHARE_CLASS.
2760 (enum omp_region_type): Add ORT_SIMD.
2761 (gimple_add_tmp_var): Handle ORT_SIMD.
2762 (gimplify_var_or_parm_decl): Same.
2763 (is_gimple_stmt): Same.
2764 (omp_firstprivatize_variable): Same.
2765 (omp_add_variable): Only use splay_tree_insert if lookup failed.
2766 (omp_notice_variable): Handle ORT_SIMD.
12211b99 2767 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ
2768 (omp_check_private): Handle ORT_SIMD.
2769 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2770 OMP_CLAUSE_SAFELEN.
2771 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
2772 Handle OMP_CLAUSE_LASTPRIVATE.
2773 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
2774 OMP_CLAUSE_SAFELEN.
2775 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
2776 (gimplify_expr): Handle OMP_SIMD.
2777 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
2778 (expand_GOMP_SIMD_VF): New.
2779 (expand_GOMP_SIMD_LAST_LANE): New.
2780 * internal-fn.def (GOMP_SIMD_LANE): New.
2781 (GOMP_SIMD_VF): New.
2782 (GOMP_SIMD_LAST_LANE): New.
2783 * omp-low.c: Include target.h.
2784 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
2785 OMP_CLAUSE_SAFELEN.
2786 (check_omp_nesting_restrictions): Same.
2787 (omp_max_vf): New.
2788 (lower_rec_simd_input_clauses): New.
2789 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
2790 OMP_CLAUSE_LINEAR.
2791 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
2792 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
2793 (expand_omp_build_assign): New.
2794 (expand_omp_for_init_counts): New.
2795 (expand_omp_for_init_vars): New.
2796 (extract_omp_for_update_vars): New.
2797 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
2798 and rewrite accordingly.
2799 (expand_omp_simd): New.
2800 (expand_omp_for): Use expand_omp_simd.
2801 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
2802 (lower_omp_for): Do not lower the body.
2803 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
2804 in their own loops.
2805 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB
2806 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
2807 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ
2808 (gate_tree_if_conversion): Similarly.
2809 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
2810 gimple_build_omp_for.
2811 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
2812 * tree-parloops (create_parallel_loop): Pass kind argument to
2813 gimple_build_omp_for.
2814 * tree-pretty-print.c (dump_omp_clause): Add cases for
2815 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
2816 OMP_CLAUSE__SIMDUID_.
2817 (dump_generic_node): Handle OMP_SIMD.
2818 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
2819 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
2820 unroll OMP_SIMD loops here.
12211b99 2821 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 2822 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 2823 loop->safelen.
74bf76ed
JJ
2824 (vect_analyze_data_refs): Handle simd loops.
2825 * tree-vect-loop.c (vectorizable_live_operation): Handle
2826 IFN_GOMP_SIMD*.
12211b99 2827 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ
2828 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
2829 (vectorizable_load): Same.
12211b99 2830 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ
2831 (struct simduid_to_vf): New.
2832 (simduid_to_vf::hash): New.
2833 (simduid_to-vf::equal): New.
2834 (struct simd_array_to_simduid): New.
2835 (simd_array_to_simduid::hash): New.
2836 (simd_array_to_simduid::equal): New.
2837 (adjust_simduid_builtins): New.
2838 (struct note_simd_array_uses_struct): New.
2839 (note_simd_array_uses_cb): New.
2840 (note_simd_array_uses): New.
2841 (vectorize_loops): Handle simd hints and adjust simd builtins
2842 accordingly.
2843 * tree-vectorizer.h (struct _stmt_vec_info): Add
2844 simd_lane_access_p field.
2845 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
2846 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
2847 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
2848 (omp_clause_code_name): Same.
2849 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
2850 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
2851 * tree.def (OMP_SIMD): New entry.
12211b99
UB
2852 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
2853 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ
2854 (OMP_CLAUSE_DECL): Adjust range for new clauses.
2855 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
2856 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
2857 (OMP_CLAUSE_LINEAR_STEP): New.
2858 (OMP_CLAUSE_SAFELEN_EXPR): New.
2859 (OMP_CLAUSE__SIMDUID__DECL): New.
2860 (find_omp_clause): New prototype.
2861
d0cf4e84
L
28622013-08-27 H.J. Lu <hongjiu.lu@intel.com>
2863
2864 * config/i386/driver-i386.c (host_detect_local_cpu): Update
2865 Haswell processor detection.
2866
992592ec
CW
28672013-08-27 Christian Widmer <shadow@umbrox.de>
2868
2869 PR target/57927
2870 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
2871 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
2872 AVX2 capable processors.
2873
9116eb22
TB
28742013-08-27 Tejas Belagod <tejas.belagod@arm.com>
2875
2876 * config/aarch64/arm_neon.h: Replace all inline asm implementations
2877 of vget_low_* with implementations in terms of other intrinsics.
2878
1c5abb53
MG
28792013-08-27 Marc Glisse <marc.glisse@inria.fr>
2880
2881 PR middle-end/57219
2882 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
2883 values to -1, 0 and 1.
2884
2e100703
VP
28852013-08-27 Vidya Praveen <vidyapraveen@arm.com>
2886
2887 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
2888 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
2889 (<optab><mode>3_insn): Remove.
2890 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
2891 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
2892 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
2893 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
2894 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
2895 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
2896 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
2897 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
2898 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
2899 (ror<mode>3_insn): Likewise.
2900 * config/aarch64/predicates.md (aarch64_simd_register): New.
2901
4ded8276
RB
29022013-08-27 Richard Biener <rguenther@suse.de>
2903
2904 PR tree-optimization/57521
2905 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
2906 one edge is non-critical.
2907 (find_phi_replacement_condition): Make sure to use a non-critical
2908 edge. Cleanup and remove old bug workarounds.
2909 (bb_postdominates_preds): Remove.
2910 (if_convertible_loop_p_1): Do not compute post-dominators.
2911 (combine_blocks): Do not free post-dominators.
2912 (main_tree_if_conversion): Likewise.
2913 (pass_data_if_conversion): Add TODO_verify_ssa.
2914
5aa11061
DD
29152013-08-27 DJ Delorie <dj@redhat.com>
2916
2917 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
2918
fd91cfe3
YZ
29192013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
2920
2921 * function.c (assign_parm_find_data_types): Set passed_mode and
2922 nominal_mode to the TYPE_MODE of nominal_type for the built
2923 pointer type in case of the struct-pass-by-reference.
2924
907555ce
JR
29252013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2926
2927 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
2928 (UINT16_TYPE): Change default to "unsigned int".
2929
67518c93
JR
2930 * config/avr/avr.opt (mfract-convert-truncate): New option.
2931 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
2932 is set, round negative fractional integers according to n1169
2933 when converting to integer types.
2934
d6d989fc
JH
29352013-08-26 Jan Hubicka <jh@suse.cz>
2936
2937 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
2938 methods can not be called indirectly when their address is not taken.
2939
2aa3da06
JH
29402013-08-26 Jan Hubicka <jh@suse.cz>
2941
12211b99
UB
2942 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
2943 ctor_for_folding.
2aa3da06 2944
0987ffe7
JH
29452013-08-26 Jan Hubicka <jh@suse.cz>
2946
2947 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
2948 can be unshared.
2949
befe8647
JR
29502013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2951
2952 * reload.c (find_valid_class): Allow classes that do not include
2953 FIRST_PSEUDO_REGISTER - 1.
2954
a21e735e
JH
29552013-08-26 Jan Hubicka <jh@suse.cz>
2956
2957 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
2958 fix edge count/frequency when speculation failed; fix type check
2959 for the direct call.
2960
e067bd43
JH
29612013-08-26 Jan Hubicka <jh@suse.cz>
2962
2963 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
2964
e8aec975
JH
29652013-08-26 Jan Hubicka <jh@suse.cz>
2966
2967 * ipa-inline-transform.c (inline_transform): Be ready for basic block
2968 to be changed by edge redirection.
2969
d0b66480
JH
29702013-08-26 Jan Hubicka <jh@suse.cz>
2971
12211b99
UB
2972 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
2973 formating; add sanity check.
d0b66480
JH
2974 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
2975 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
2976 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
2977 (resolve_noninline_speculation): Update callee keys, too.
2978
0f9aaac7
JH
29792013-08-26 Jan Hubicka <jh@suse.cz>
2980
2981 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
2982 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
2983
4b37444e
JR
29842013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
2985
2986 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
2987 into proper place.
2988
5339fc0c
UB
29892013-08-26 Uros Bizjak <ubizjak@gmail.com>
2990
2991 * config/i386/i386.c (ix86_debug_options): Remove prototype.
2992 (x86_64_elf_select_section): Ditto.
2993 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
2994 arguments.
2995 (ix86_pass_by_reference): Ditto.
5339fc0c
UB
2996 (output_set_got): Ditto.
2997 (ix86_unary_operator_ok): Ditto.
2998 (ix86_expand_builtin): Ditto.
2999
d5c3d3ef
JH
30002013-08-23 Jan Hubicka <jh@suse.cz>
3001
5339fc0c 3002 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 3003
5b1e7435
JH
30042013-08-23 Jan Hubicka <jh@suse.cz>
3005
3006 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
3007 (tree_decl_with_vis): Add FINAL field.
3008
ead69dac
JL
30092013-08-23 Jeff Law <law@redhat.com>
3010
3011 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
3012 the debugging dump when the expression is fully redundant.
3013
025311c4
GDR
30142013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3015
3016 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
3017 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
3018 * pretty-print.c (pp_formatted_text_data): Likewise.
3019 (pp_write_text_to_stream): Likewise.
3020 (pp_write_text_as_dot_label_to_stream): Likewise.
3021 (pp_append_r): Likewise.
3022 (pp_format): Likewise.
3023 (pp_flush): Likewise.
3024 (pp_clear_output_area): Likewise.
3025 (pp_append_text): Likewise.
3026 (pp_formatted_text): Likewise.
3027 (pp_remaining_character_count_for_line): Likewise.
3028 (pp_newline): Likewise.
3029 (pp_character): Likewise.
3030 (output_buffer::~output_buffer): Define.
3031 (pretty_printer::~pretty_printer): Destruct output buffer.
3032 * pretty-print.h (output_buffer::~output_buffer): Declare.
3033 (pretty_printer::~pretty_printer): Declare virtual.
3034
0cadbfaa
MG
30352013-08-24 Marc Glisse <marc.glisse@inria.fr>
3036
3037 PR other/57324
3038 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
3039 HOST_WIDE_INT_M1U): New macros.
3040 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
3041 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
3042 unsigned -1 for lshift.
3043 * cse.c (cse_insn): Likewise.
3044 * double-int.c (rshift_double, lshift_double): Likewise.
3045 * builtins.c (fold_builtin_bitop): Likewise.
3046 * combine.c (force_to_mode): Likewise.
3047 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
3048 * simplify-rtx.c (simplify_const_unary_operation,
3049 simplify_const_binary_operation): Likewise.
3050 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
3051 check_va_list_escapes): Likewise.
3052 * rtlanal.c (nonzero_bits1): Likewise.
3053 * expmed.c (expand_smod_pow2): Likewise.
3054 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
3055
009e5353
JH
30562013-08-23 Jan Hubicka <jh@suse.cz>
3057
3058 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
3059 as having address taken.
3060
c4be6568
JH
30612013-08-23 Jan Hubicka <jh@suse.cz>
3062
64cbf23d
JH
3063 * ipa-utils.h (method_class_type): Declare.
3064 * ipa-devirt.c (method_class_type): Export.
3065
c4be6568
JH
3066 * cgraphunit.c (analyze_functions): Do basic devirtualization;
3067 do not walk base classes of anonymous types.
3068
97aba8e9
KK
30692013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
3070
3071 PR rtl-optimization/58220
3072 PR regression/58221
3073 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
3074 handle SEQUENCE insns properly.
3075
0fc80001
GDR
30762013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
3077
3078 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
3079 definition.
3080 (pp_newline_and_indent): Likewise.
3081 (pp_separate_with): Likewise.
3082 * pretty-print.c (pp_newline_and_flush): Define.
3083 (pp_newline_and_indent): Likewise.
3084 (pp_separate_with): Likewise.
3085
520a5868
JJ
30862013-08-23 Jakub Jelinek <jakub@redhat.com>
3087
3088 PR target/58218
3089 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
3090 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
3091
e500c62a
KY
30922013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
3093
f43245d1
UB
3094 * config/i386/predicates.md (ext_sse_reg_operand): New.
3095 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY
3096 predicate to determine if EVEX is needed.
3097 (*movsi_internal): Ditto.
3098 (*movdf_internal): Ditto.
3099 (*movsf_internal): Ditto.
f43245d1 3100 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 3101
6b00d7dd
JJ
31022013-08-23 Jakub Jelinek <jakub@redhat.com>
3103
3104 PR tree-optimization/58209
3105 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
3106 (find_tail_calls): Give up for pointer result types if m is non-NULL.
3107 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
3108 emit POINTER_PLUS_EXPR.
3109 (create_tailcall_accumulator): For pointer result type accumulate in
3110 sizetype type.
3111
4f2a9f90
PC
31122013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
3113
3114 * configure.ac: Add backslashes missing from the last change.
3115 * configure: Regenerate.
3116
7bf4274e
JH
31172013-08-22 Jan Hubicka <jh@susue.cz>
3118
4f2a9f90
PC
3119 * ipa.c (function_and_variable_visibility): First remember function
3120 was global and then make it local.
7bf4274e 3121
1bea243d
JB
31222013-08-22 Julian Brown <julian@codesourcery.com>
3123
3124 * configure.ac: Add aarch64 to list of arches which use "nop" in
3125 debug_line test.
3126 * configure: Regenerate.
3127
bcba7ecc
AK
31282013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3129
3130 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
3131 gnu_libc_has_function.
3132 * config/s390/tpf.h: Likewise.
3133
4c495b0d
JH
31342013-08-22 Jan Hubicka <jh@susue.cz>
3135
3136 * timevar.c (validate_phases): Add cast.
3137
5e302bca
JH
31382013-08-22 Jan Hubicka <jh@susue.cz>
3139
3140 * timevar.c (validate_phases): Use size_t for memory.
3141 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
3142
da6ca2b5
GDR
31432013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
3144
3145 * pretty-print.h (output_buffer::output_buffer): Declare.
3146 (pretty_printer::pretty_printer): Likewise.
3147 (pp_construct): Remove.
3148 * pretty-print.c (output_buffer::output_buffer): Define.
3149 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
3150 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
3151 (print_gimple_expr): Likewise.
3152 (print_gimple_seq): Likewise.
3153 (gimple_dump_bb): Likewise.
3154 * sched-vis.c (dump_value_slim): Likewise.
3155 (dump_insn_slim): Likewise.
3156 (dump_rtl_slim): Likewise.
3157 (str_pattern_slim): Likewise.
3158 * tree-mudflap.c (mf_varname_tree): Likewise.
3159 * graph.c (print_graph_cfg): Likewise.
3160 (start_graph_dump): Likewise.
3161 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
3162 placement-new.
3163 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
3164 pretty printer initialization.
3165 * coretypes.h (diagnostic_context): Remove superflous type alias
3166 declaration.
3167 (pretty_printer): Likewise. Declare directly as a class.
3168 (pretty_print_info): Remove declaration as class.
3169 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
3170 and pp_clear_output_area.
3171 (asan_add_global): Likewise.
3172
0e1474e5
JH
31732013-08-22 Jan Hubicka <jh@suse.cz>
3174
3175 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
3176 * ipa-utils.h (update_type_inheritance_graph): Declare.
3177 (possible_polymorphic_call_target_p): Declare.
3178 (possible_polymorphic_call_target_p): New.
3179 * ipa-devirt.c: Update toplevel comments.
3180 (cached_polymorphic_call_targets): Move up.
3181 (odr_type_d): Move ID down.
3182 (polymorphic_type_binfo_p): Update comment.
3183 (odr_hasher::remove): Likewise;
3184 (get_odr_type): Set anonymous_namespace.
3185 (dump_odr_type): Dump it.
3186 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
3187 (maybe_record_node): Record node in cached_polymorphic_call_targets.
3188 (record_binfo): Add comment.
5339fc0c
UB
3189 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
3190 built.
0e1474e5
JH
3191 (devirt_node_removal_hook): Do not iCE when cache is freed.
3192 (possible_polymorphic_call_target_p): New predicate.
3193 (update_type_inheritance_graph): New function.
3194
3f97cb0b
AI
31952013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
3196 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3197 Sergey Lega <sergey.s.lega@intel.com>
3198 Anna Tikhonova <anna.tikhonova@intel.com>
3199 Ilya Tocar <ilya.tocar@intel.com>
3200 Andrey Turetskiy <andrey.turetskiy@intel.com>
3201 Ilya Verbin <ilya.verbin@intel.com>
3202 Kirill Yukhin <kirill.yukhin@intel.com>
3203 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3204
3205 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
3206 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
3207 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
3208 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
3209 (OPTION_MASK_ISA_AVX2_UNSET): Update.
3210 (OPTION_MASK_ISA_AVX512F_UNSET): New.
3211 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
3212 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
3213 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
3214 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
3215 OPT_mavx512pf, OPT_mavx512er cases.
3216 * config/i386/constraints.md (v): New constraint.
3217 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
3218 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
3219 (bit_AVX512CD): New.
3220 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
3221 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
3222 * config/i386/i386-c.c (ix86_target_macros_internal):
3223 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
3224 __AVX512PF__.
3225 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
3226 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
3227 * config/i386/i386.c (regclass_map, dbx_register_map)
3228 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
3229 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
3230 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
3231 -mavx512pf options.
3232 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
3233 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
3234 -mavx512pf options. Fix formatting.
3235 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
3236 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
3237 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
3238 -mavx512cd, -mavx512pf options.
3239 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
3240 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
3241 (ix86_preferred_output_reload_class): Replace SSE_REGS with
3242 ALL_SSE_REGS.
3243 (ix86_hard_regno_mode_ok): Support 512-bit registers.
3244 (ix86_set_reg_reg_cost): Ditto.
3245 (x86_order_regs_for_local_alloc): Ditto.
3246 (MAX_VECT_LEN): Extend to 64-byte.
3247 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
3248 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
3249 (TARGET_AVX512ER, TARGET_AVX512CD): New.
3250 (BIGGEST_ALIGNMENT): Extend to 512-bits.
3251 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
3252 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
3253 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
3254 (SSE_REG_MODE_P): Support new modes.
3255 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
3256 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
3257 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
3258 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
3259 (REG_CLASS_CONTENTS): Add new registers.
3260 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
3261 (EXT_REX_SSE_REGNO_P): New.
3262 (HI_REGISTER_NAMES): Add new registers.
3263 * config/i386/i386.md: Define constants for new registers.
3264 (mode): Add new 512-bit modes.
3265 (prefix): Support evex prefix.
3266 (isa): Support avx512f, noavx512f, fma_avx512f.
3267 (ssemodesuffix): Add new 512-bit modes.
3268 (movxi): New.
3269 (*movxi_internal_avx512f): Ditto.
3270 (*movdi_internal): Replace constraint "x" with the new constraint "v".
3271 Support MODE_XI.
3272 (*movsi_internal): Likewise.
3273 (*movdf_internal): Likewise.
3274 (*movsf_internal): Likewise.
3275 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
3276 (<code><mode>3): Likewise.
5339fc0c
UB
3277 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
3278 New.
3f97cb0b
AI
3279 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
3280 with the new constraint "v".
3281 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
3282 modes.
3283 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
3284 with the new constraint "v".
3285 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
3286 (<sse2>_storedqu<avxsizesuffix>): Likewise.
3287 (*<plusminus_insn><mode>3): Likewise.
3288 (<sse>_vm<plusminus_insn><mode>3): Likewise.
3289 (*mul<mode>3): Likewise.
3290 (<sse>_vmmul<mode>3): Likewise.
3291 (<sse>_div<mode>3): Likewise.
3292 (<sse>_vmdiv<mode>3): Likewise.
3293 (<sse>_sqrt<mode>2): Likewise.
3294 (<sse>_vmsqrt<mode>2): Likewise.
3295 (*<code><mode>3_finite): Likewise.
3296 (*<code><mode>3) <smaxmin>: Likewise.
3297 (<sse>_vm<code><mode>3): Likewise.
3298 (*<code><mode>3) <any_logic>: Likewise.
3299 (*fma_fmadd_<mode>): Likewise.
3300 (*fma_fmsub_<mode>): Likewise.
3301 (*fma_fnmadd_<mode>): Likewise.
3302 (*fma_fnmsub_<mode>): Likewise.
3303 (*fma_fmaddsub_<mode>): Likewise.
3304 (*fma_fmsubadd_<mode>): Likewise.
3305 (*fmai_fmadd_<mode>): Likewise.
3306 (*fmai_fmsub_<mode>): Likewise.
3307 (*fmai_fnmadd_<mode>): Likewise.
3308 (*fmai_fnmsub_<mode>): Likewise.
3309 (sse_cvtsi2ss): Likewise.
3310 (sse_cvtsi2ssq): Likewise.
3311 (sse_cvtss2si): Likewise.
3312 (sse_cvtss2si_2): Likewise.
3313 (sse_cvtss2siq): Likewise.
3314 (sse_cvtss2siq_2): Likewise.
3315 (sse_cvttss2si): Likewise.
3316 (sse_cvtss2siq_2): Likewise.
3317 (float<sseintvecmodelower><mode>2): Likewise.
3318 (sse2_cvtsd2si_2): Likewise.
3319 (sse2_cvtsd2siq_2): Likewise.
3320 (*<plusminus_insn><mode>3): Likewise.
3321 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
3322 (*<sse4_1_avx2>_mul<mode>3): Likewise.
3323 (ashr<mode>3): Likewise.
3324 (<shift_insn><mode>3): Likewise.
3325 (avx2_<code><mode>3): Likewise.
3326 (*avx2_<code><mode>3): Likewise.
3327 (*andnot<mode>3): Likewise.
3328 (*<code><mode>3) <any_logic>: Likewise.
3329 (abs<mode>2): Likewise.
3330 (avx2_permvar<mode>): Likewise.
3331 (avx2_perm<mode>_1): Likewise.
3332 (*avx_vpermilp<mode>): Likewise.
3333 (avx_vpermilvar<mode>3): Likewise.
3334 (avx2_ashrv<mode>): Likewise.
3335 (avx2_<shift_insn>v<mode>): Likewise.
3336 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
3337 -mavx512cd.
3338 * doc/rtl.texi: Document XImode.
3339
dd5e8423
JL
33402013-08-21 Jeff Law <law@redhat.com>
3341
b9ebee5d
JL
3342 * tree-flow.h (register_jump_thread): Pass vector of edges
3343 instead of each important edge.
3344 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
3345 thread path into a vector and pass that to register_jump_thread.
3346 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
3347 passed in edge vector to the current 3-edge form.
3348
dd5e8423
JL
3349 Revert:
3350 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3351
9bb6628e 3352 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL
3353 delete_unmarked_insns removed anything.
3354
445dc8df
JR
33552013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
3356
3357 * reload.h (struct reg_equivs): Rename to ..
3358 (struct reg_equivs_s): .. this.
3359
e0df53dd
ML
33602013-08-20 Martin Liska <marxin.liska@gmail.com>
3361
3362 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
3363
e7d1d3eb
RO
33642013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3365
5339fc0c 3366 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 3367
5562e26e
JL
33682013-08-21 Jeff Law <law@redhat.com>
3369
3370 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
3371 simplify assignments too. If the RHS collapses to a singleton
3372 range, then return the value for the range.
3373
c7ecdec6
KY
33742013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
3375
3376 * config/i386/sse.md (V16): Rename to...
3377 (VMOVE): this.
3378 (mov<mode>): Update iterator name.
3379 (*mov<mode>_internal): Ditto.
3380 (push<mode>1): Ditto.
3381 (movmisalign<mode>): Ditto.
3382
bfa3b50a
JH
33832013-08-20 Jan Hubicka <jh@suse.cz>
3384
3385 PR bootstrap/58186
3386 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
3387 entry for direct edges.
3388 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
3389
6cd4d135
DM
33902013-08-20 David Malcolm <dmalcolm@redhat.com>
3391
3392 Revert my last two changes, r201865 and r201864:
3393
3394 Revert r201865:
3395 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3396
3397 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3398 instances can own GC refs.
3399
3400 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3401 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3402 (gcc::context::gt_pch_nx): Likewise.
3403 (gcc::context::gt_pch_nx): Likewise.
3404 * ggc.h (gt_ggc_mx <T>): New.
3405 (gt_pch_nx_with_op <T>): New.
3406 (gt_pch_nx <T>): New.
3407 * passes.c (opt_pass::gt_ggc_mx): New.
3408 (opt_pass::gt_pch_nx): New.
3409 (opt_pass::gt_pch_nx_with_op): New.
3410 (pass_manager::gt_ggc_mx): New.
3411 (pass_manager::gt_pch_nx): New.
3412 (pass_manager::gt_pch_nx_with_op): New.
3413 (pass_manager::operator new): Use
3414 ggc_internal_cleared_alloc_stat rather than xcalloc.
3415 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3416 (pass_manager::gt_ggc_mx): New.
3417 (pass_manager::gt_pch_nx): New.
3418 (pass_manager::gt_pch_nx_with_op): New.
3419 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3420 (opt_pass::operator new): New.
3421 (opt_pass::gt_ggc_mx): New.
3422 (opt_pass::gt_pch_nx): New.
3423 (opt_pass::gt_pch_nx_with_op): New.
3424
3425 Revert r201864:
3426 2013-08-20 David Malcolm <dmalcolm@redhat.com>
3427
3428 * Makefile.in (GTFILES): Add context.h.
3429 * context.c (gcc::context::operator new): New.
3430 (gcc::context::gt_ggc_mx): New.
3431 (gcc::context::gt_pch_nx): New.
3432 (gcc::context::gt_pch_nx): New.
3433 * context.h (gcc::context): Add GTY((user)) marking.
3434 (gcc::context::operator new): New.
3435 (gcc::context::gt_ggc_mx): New.
3436 (gcc::context::gt_pch_nx): New.
3437 (gcc::context::gt_pch_nx): New.
3438 (g): Add GTY marking.
3439 (gt_ggc_mx (gcc::context *)): New.
3440 (gt_pch_nx (gcc::context *)): New.
3441 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3442 void *cookie)): New.
3443 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3444
a046136a
AM
34452013-08-20 Alexey Makhalov <makhaloff@gmail.com>
3446
3447 * dce.c (fini_dce): Call df_analyze again just in case
3448 delete_unmarked_insns removed anything.
3449
67598720
TJ
34502013-08-20 Teresa Johnson <tejohnson@google.com>
3451
3452 PR rtl-optimizations/57451
3453 * final.c (reemit_insn_block_notes): Prevent lexical blocks
3454 from crossing split section boundaries.
3455
f8693faf
MGD
34562013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
3457
3458 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
3459 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
3460 with MULTLIB_DEFAULTS.
3461
7d0b9a9c
NC
34622013-08-20 Nick Clifton <nickc@redhat.com>
3463
3464 * target.def (narrow_volatile_bitfield): Note that the default
3465 value is false, not !TARGET_STRICT_ALIGN.
3466 * doc/tm.texi: Regenerate.
3467
8edb8dc8
PC
34682013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
3469
3470 Fix LIB_SPEC for systems without libpthread.
3471
3472 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
3473 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
3474 for Android.
3475 * config/i386/linux-common.h: Likewise.
3476 * config/mips/linux-common.h: Likewise.
3477
12211b99 34782013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ
3479
3480 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
3481 checks.
3482
e42c64cb
DM
34832013-08-20 David Malcolm <dmalcolm@redhat.com>
3484
3485 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
3486 instances can own GC refs.
3487
3488 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
3489 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
3490 (gcc::context::gt_pch_nx): Likewise.
3491 (gcc::context::gt_pch_nx): Likewise.
3492 * ggc.h (gt_ggc_mx <T>): New.
3493 (gt_pch_nx_with_op <T>): New.
3494 (gt_pch_nx <T>): New.
3495 * passes.c (opt_pass::gt_ggc_mx): New.
3496 (opt_pass::gt_pch_nx): New.
3497 (opt_pass::gt_pch_nx_with_op): New.
3498 (pass_manager::gt_ggc_mx): New.
3499 (pass_manager::gt_pch_nx): New.
3500 (pass_manager::gt_pch_nx_with_op): New.
3501 (pass_manager::operator new): Use
3502 ggc_internal_cleared_alloc_stat rather than xcalloc.
3503 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
3504 (pass_manager::gt_ggc_mx): New.
3505 (pass_manager::gt_pch_nx): New.
3506 (pass_manager::gt_pch_nx_with_op): New.
3507 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
3508 (opt_pass::operator new): New.
3509 (opt_pass::gt_ggc_mx): New.
3510 (opt_pass::gt_pch_nx): New.
3511 (opt_pass::gt_pch_nx_with_op): New.
3512
910c02a0
DM
35132013-08-20 David Malcolm <dmalcolm@redhat.com>
3514
3515 * Makefile.in (GTFILES): Add context.h.
3516 * context.c (gcc::context::operator new): New.
3517 (gcc::context::gt_ggc_mx): New.
3518 (gcc::context::gt_pch_nx): New.
3519 (gcc::context::gt_pch_nx): New.
3520 * context.h (gcc::context): Add GTY((user)) marking.
3521 (gcc::context::operator new): New.
3522 (gcc::context::gt_ggc_mx): New.
3523 (gcc::context::gt_pch_nx): New.
3524 (gcc::context::gt_pch_nx): New.
3525 (g): Add GTY marking.
3526 (gt_ggc_mx (gcc::context *)): New.
3527 (gt_pch_nx (gcc::context *)): New.
3528 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
3529 void *cookie)): New.
3530 * gengtype.c (open_base_files) <ifiles>: Add context.h.
3531
af4b4236
AM
35322013-08-20 Alan Modra <amodra@gmail.com>
3533
3534 PR target/57865
3535 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
3536 (rs6000_emit_epilogue): Likewise.
3537
12211b99 35382013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC
3539
3540 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
3541
e2323f5b
PB
35422013-08-19 Peter Bergner <bergner@vnet.ibm.com>
3543 Jakub Jelinek <jakub@redhat.com>
3544
3545 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
3546 (BUILT_IN_FABSD64): Likewise.
3547 (BUILT_IN_FABSD128): Likewise.
3548 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
3549 (fold_builtin_1): Likewise.
5339fc0c
UB
3550 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
3551 destination and source operands.
e2323f5b
PB
3552 (*abstd2_fpr): Likewise.
3553 (*nabstd2_fpr): Likewise.
3554
cfbf3ee8
RS
35552013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
3556
3557 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
3558 JUMP_P and INSN_P.
3559
d3136aeb
AH
35602013-08-19 Aldy Hernandez <aldyh@redhat.com>
3561
3562 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
3563 incomplete.
3564
d33d9e47
AI
35652013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
3566
3567 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
3568 * builtins.c (default_libc_has_function): New.
3569 (gnu_libc_has_function): Ditto.
3570 (no_c99_libc_has_function): Ditto.
3571 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
3572 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
3573 (fold_builtin_sincos): Likewise.
3574 (fold_builtin_cexp): Likewise.
3575 * builtins.def (DEF_C94_BUILTIN): Likewise.
3576 (DEF_C99_BUILTIN): Likewise.
3577 (DEF_C99_C90RES_BUILTIN): Likewise.
3578 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
3579 definitions to using this define.
3580 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 3581 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI
3582 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
3583 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
3584 * config/darwin.h: Ditto.
3585 * config/elfos.h: Ditto.
3586 * config/freebsd.h: Ditto.
3587 * config/i386/cygming.h: Ditto.
3588 * config/i386/djgpp.h: Ditto.
3589 * config/i386/i386-interix.h: Ditto.
3590 * config/microblaze/microblaze.h: Ditto.
3591 * config/mmix/mmix.h: Ditto.
3592 * config/gnu-user.h: Ditto.
3593 * config/ia64/hpux.h: Ditto.
3594 * config/pa/pa-hpux.h: Ditto.
3595 * config/pdp11/pdp11.h: Ditto.
3596 * config/picochip/picochip.h: Ditto.
3597 * config/linux.h: Ditto.
3598 * config/netbsd.h: Ditto.
3599 * config/openbsd.h: Ditto.
3600 * config/rs6000/aix43.h: Ditto.
3601 * config/rs6000/aix51.h: Ditto.
3602 * config/rs6000/aix52.h: Ditto.
3603 * config/rs6000/aix53.h: Ditto.
3604 * config/rs6000/aix61.h: Ditto.
3605 * config/rs6000/darwin.h: Ditto.
3606 * config/rs6000/linux.h: Ditto.
3607 * config/rs6000/linux64.h: Ditto.
3608 * config/s390/tpf.h: Ditto.
3609 * config/sol2-10.h: Ditto.
3610 * config/sol2.h: Ditto.
3611 * config/vms/vms.h: Ditto.
3612 * config/vxworks.h: Ditto.
3613 * config/linux-android.c (linux_android_libc_has_function):
3614 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
3615 * config/linux-protos.h (linux_android_libc_has_function):
3616 New declaration.
3617 * config/i386/i386.c (ix86_libc_has_function): New.
3618 * config/i386/i386-protos.h
3619 (ix86_libc_has_function): New declaration.
3620 * config/i386/i386.md
3621 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
3622 ("isinf<mode>2): Likewise.
3623 * convert.c (convert_to_integer): Using new target hook
3624 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
3625 TARGET_C99_FUNCTIONS.
3626 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
3627 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
3628 * coretypes.h (function_class): New enum for different
3629 classes of functions.
3630 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
3631 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
3632 (TARGET_HAS_SINCOS): Likewise.
3633 (TARGET_LIBC_HAS_FUNCTION): New.
3634 * doc/tm.texi: Regenerated.
3635 * targhooks.h (default_libc_has_function): New declaration.
3636 (no_c99_libc_has_function): Ditto.
3637 (gnu_libc_has_function): Ditto.
3638 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
3639 and TARGET_HAS_SINCOS.
3640
eefe9a99
JH
36412013-08-18 Jan Hubicka <jh@suse.cz>
3642
3643 * Makeifle-in (ipa-devirt.o): New.
3644 (GTFILES): Add ipa-utils.h and ipa-devirt.c
3645 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
3646 (analyze_functions): Look into possible targets of polymorphic call.
3647 * dumpfile.c (dump_files): Add type-inheritance dump.
3648 * dumpfile.h (TDI_inheritance): New.
3649 * ipa-devirt.c: New file.
3650 * ipa-utils.h (odr_type_d): Forward declare.
3651 (odr_type): New type.
3652 (build_type_inheritance_graph): Declare.
3653 (possible_polymorphic_call_targets): Declare and introduce inline
3654 variant when only edge is pased.
3655 (dump_possible_polymorphic_call_targets): Likewise.
3656 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
3657 * tree.c (type_in_anonymous_namespace_p): Break out from ...
3658 (types_same_for_odr): ... here.
3659 * tree.h (type_in_anonymous_namespace_p): Declare.
3660
29b89442
JJ
36612013-08-18 Jakub Jelinek <jakub@redhat.com>
3662
3663 PR tree-optimization/58006
3664 * tree-parloops.c (take_address_of): Don't ICE if get_name
3665 returns NULL.
3666 (eliminate_local_variables_stmt): Remove clobber stmts.
3667
4f219961
EB
36682013-08-18 Eric Botcazou <ebotcazou@adacore.com>
3669
3670 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
3671 error message is issued for an alias to undefined symbol.
3672
1d5755ef
JH
36732013-08-18 Jan Hubicka <jh@suse.cz>
3674
3675 * cgraph.c (cgraph_create_indirect_edge): Discover
3676 polymorphic calls and record basic info into indirect_info.
3677 * gimple-fold.c (gimple_fold_call): When doing BINFO based
3678 devirtualization, ignore objc function calls.
3679 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
3680 call with no parm index info.
3681 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
3682 * tree.c (virtual_method_call_p): New function.
3683 * tree.h (virtual_method_call_p): Declare.
3684
5945bebf
JH
36852013-08-16 Jan Hubicka <jh@suse.cz>
3686
3687 PR middle-end/58179
3688 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
3689
47fa5d23
DE
36902013-08-16 David Edelsohn <dje.gcc@gmail.com>
3691
3692 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
3693 attribute.
3694
c74f54a0
DM
36952013-08-16 David Malcolm <dmalcolm@redhat.com>
3696
3697 * gengtype.c (type_for_name): Add special-case support for
3698 locating types within the "gcc::" namespace.
3699 (open_base_files): Emit a "using namespace gcc" directive.
3700
f758f299
MM
37012013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3702
3703 PR target/58160
3704 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
3705 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
3706
3707 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
3708 array instead of each individual operand as a separate argument.
3709 (emit_fusion_gpr_load): Likewise.
3710 (expand_fusion_gpr_load): Add new function declaration.
3711
3712 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
3713 signature to have the operands passed as an array, instead of as
3714 separate arguments. Allow ZERO_EXTEND to be in the memory
3715 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
3716 depend on the register live/dead flags when peepholes are run.
3717 (expand_fusion_gpr_load): New function to be called from the
3718 peephole2 pass, to change the register that addis sets to be the
3719 target register.
3720 (emit_fusion_gpr_load): Change the calling signature to have the
3721 operands passed as an array, instead of as separate arguments.
3722 Allow ZERO_EXTEND to be in the memory address, and also
3723 SIGN_EXTEND if -mpower8-fusion-sign.
3724
3725 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
3726 unspec enumeration.
3727 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
3728 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 3729 depend on the register live/dead state when the peephole pass is done.
f758f299 3730
158f4e4f
DM
37312013-08-16 David Malcolm <dmalcolm@redhat.com>
3732
3733 * gengtype.c (create_user_defined_type): Ensure that the kind
3734 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
3735 declaration is seen before the GTY((user)) marking.
3736
22f8cea5
BE
37372013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
3738
3739 PR target/58105
3740 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
3741
c49bdb2e
JH
37422013-08-16 Jan Hubicka <jh@suse.cz>
3743
3744 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
3745 arugment expected_type.
3746 (gimple_fold_call): Use it.
3747 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
3748 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 3749 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH
3750 (try_make_edge_direct_virtual_call): Likewise.
3751 * tree.c (obj_type_ref_class): New.
3752 * tree.h (obj_type_ref_class): Use it.
3753
4042dca9
GDR
37542013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
3755
3756 * sched-vis.c (rtl_slim_pp_initialized): Remove.
3757 (rtl_slim_pp): Likewise.
3758 (init_rtl_slim_pretty_print): Likewise.
3759 (dump_value_slim): Don't call it. Use local pretty printer.
3760 (dump_insn_slim): Likewise.
3761 (dump_rtl_slim): Likewise.
3762 (str_pattern_slim): Likewise.
3763 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
3764 Simplify.
3765
7a460594
JJ
37662013-08-16 Jakub Jelinek <jakub@redhat.com>
3767
639dc669
JJ
3768 PR tree-optimization/58164
3769 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
3770 walk gimple_goto_dest of GIMPLE_GOTO.
3771
7a460594
JJ
3772 PR tree-optimization/58165
3773 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
3774 bi_call must be the last stmt in a bb, don't split_block, instead
3775 use fallthru edge from it and give up if there is none.
3776 Release conds vector when returning early.
3777
e3c02cc7
XDL
37782013-08-14 Xinliang David Li <davidxl@google.com>
3779
3780 * config/i386/i386.c (ix86_option_override_internal):
3781 Remove unused variable and field.
3782
6469da2c
BS
37832013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3784
3785 PR target/57949
22f8cea5 3786 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS
3787 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
3788 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
3789 and Linux, correct BLKmode alignment when 128-bit alignment is
3790 required and compatibility flag is not set.
22f8cea5
BE
3791 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
3792 for zero-size arguments when compatibility flag is not set.
6469da2c 3793
4ca890e2
JJ
37942013-08-14 Jakub Jelinek <jakub@redhat.com>
3795
3796 PR tree-optimization/58145
3797 * tree-sra.c (build_ref_for_offset): If prev_base has
3798 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
3799
daec1759 38002013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 3801
daec1759 3802 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 3803 Fix uninitialized variable error.
daec1759 3804
3ad20bd4
XDL
38052013-08-14 Xinliang David Li <davidxl@google.com>
3806
3807 * config/i386/i386.opt: Define two new options.
3808 * config/i386/x86-tune.def: Add arch selector field in macros.
3809 * config/i386/i386.h: Adjust macro definition.
3810 * config/i386/i386.c (ix86_option_override_internal):
3811 Refactor the code.
3812 (parse_mtune_ctrl_str): New function.
3813 (set_ix86_tune_features): New function.
3814 (ix86_function_specific_restore): Call the new helper function.
3815
e839e2a9
AB
38162013-08-14 Andrey Belevantsev <abel@ispras.ru>
3817
3818 PR rtl-optimization/57662
3819 * sel-sched.c (code_motion_process_successors): When the current insn
3820 is removed after the recursive traversal, break from the loop.
3821 Add comments and debug printouts.
3822
27ed665f
JJ
38232013-08-14 Jakub Jelinek <jakub@redhat.com>
3824 Alexandre Oliva <aoliva@redhat.com>
3825
3826 PR target/58067
3827 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
3828 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
3829 there also UNSPEC_PLTOFF.
3830
61c1a609
MP
38312013-08-14 Marek Polacek <polacek@redhat.com>
3832
3833 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
3834 NUM_CONDITIONS bit positions.
3835
0e901c67
CC
38362013-08-13 Cary Coutant <ccoutant@google.com>
3837
3838 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
3839 (attr_checksum): Hash vector contents instead of pointer.
3840 (attr_checksum_ordered): Likewise.
3841
061eff6d
UB
38422013-08-13 Uros Bizjak <ubizjak@gmail.com>
3843
3844 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
3845 when Pmode != word_mode. Add length_address attribute.
3846 (sse3_monitor_<mode>): Merge from sse3_monitor and
3847 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
3848 Pmode != word_mode. Update insn length attribute.
3849 * config/i386/i386.c (ix86_option_override_internal): Update
3850 ix86_gen_monitor selection for merged sse3_monitor insn.
3851
8ec9249b
JB
38522013-08-13 Julian Brown <julian@codesourcery.com>
3853
3854 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
3855 perform invalid legitimization on greater-than-word-size modes for
3856 TARGET_E500_DOUBLE.
3857
761a8eb7
VM
38582013-08-13 Vladimir Makarov <vmakarov@redhat.com>
3859
3860 * ira.c (setup_class_translate_array): Use aclass instead of cl
3861 for classes not fully covered by allocno classes.
3862
99c2bd54
JJ
38632013-08-13 Jakub Jelinek <jakub@redhat.com>
3864
4029a5e0
JJ
3865 PR tree-optimization/57661
3866 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
3867 * tree-inline.c (tree_function_versioning): Initialize it.
3868 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
3869 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
3870 that is not being copied.
3871
99c2bd54
JJ
3872 PR sanitizer/56417
3873 * asan.c (instrument_strlen_call): Fix typo in comment.
3874 Use char * type even for the lhs of POINTER_PLUS_EXPR.
3875
b5f8f063
SE
38762013-08-13 Steve Ellcey <sellcey@mips.com>
3877
3878 * config/mips/mips.md (prefetch): Use lw instead of ld on
3879 loongson in 32bit mode.
3880
4de80584
NC
38812013-08-13 Nick Clifton <nickc@redhat.com>
3882
3883 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
3884
09ce3660
JH
38852013-08-13 Jan Hubicka <jh@suse.cz>
3886
3887 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
3888 introduced edge; fix typo in sanity check.
3889 (cgraph_resolve_speculation): Export; improve diagnostic.
3890 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
3891 speculation at type mismatch.
3892 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
3893 (cgraph_resolve_speculation): Declare.
3894 (symtab_can_be_discarded): New function.
3895 * value-prof.c (gimple_ic_transform): Remove actual transform code.
3896 * ipa-inline-transform.c (speculation_removed): New global var.
3897 (clone_inlined_nodes): See if speculation can be removed.
3898 (inline_call): If speculations was removed, we growths may not match.
3899 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
3900 (speculation_useful_p): New function.
3901 (resolve_noninline_speculation): New function.
3902 (inline_small_functions): Resolve useless speculations.
3903 * ipa-inline.h (speculation_useful_p): Declare
3904 * ipa.c (can_replace_by_local_alias): Simplify.
3905 (ipa_profile): Produce speculative calls in non-lto, too;
3906 add simple cost model; produce local aliases.
3907
f971dc24
DM
39082013-08-13 David Malcolm <dmalcolm@redhat.com>
3909
3910 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 3911 PASS_MANAGER_H.
f971dc24 3912
2e507b9a
PC
39132013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
3914
3915 * config/i386/i386.c (ix86_function_versions): Use error + inform.
3916
c97fd4b8
UB
39172013-08-12 Uros Bizjak <ubizjak@gmail.com>
3918
3919 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
3920 iterator instead of X87MODEF.
3921
39222013-08-12 Perez Read <netfirewall@gmail.com>
3923
3924 PR target/58132
3925 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
3926 operand 0 for intel asm alternative.
3927 (*movabs<mode>_2): Ditto for operand 1.
3928
f4f4204c
JG
39292013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
3930
3931 * config/aarch64/arm_none.h
3932 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
3933
133b1a8e
NC
39342013-08-12 Nick Clifton <nickc@redhat.com>
3935
3936 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
3937
15dd8b3a
YR
39382013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
3939
3940 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
3941 expand for QI/HImode operand to produce more effictive code for
3942 unsigned char(short) --> float(double) conversion.
3943
94109a6a
AM
39442013-08-12 Alexander Monakov <amonakov@ispras.ru>
3945
3946 * doc/invoke.texi: Mention that -ftls-model does not force the final
3947 model.
3948
4e115102
MP
39492013-08-12 Marek Polacek <polacek@redhat.com>
3950 Marc Glisse <marc.glisse@inria.fr>
3951
3952 PR tree-optimization/57980
3953 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
3954 when creating -1 constant.
3955
748d71f3
JH
39562013-08-10 Jan Hubicka <jh@suse.cz>
3957
1f97e193 3958 Workaround binutils PR14342.
748d71f3
JH
3959 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
3960 (gimple_init_edge_profiler): Likewise.
3961 (gimple_gen_ic_func_profiler): Likewise.
3962
5979aa54
JH
39632013-08-09 Jan Hubicka <jh@suse.cz>
3964
3965 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
3966
ad83025e
XDL
39672013-08-09 Xinliang David Li <davidxl@google.com>
3968
3969 * config/i386/stringop.def: New file.
3970 * config/i386/stringop.opt: New file.
3971 * config/i386/i386-opts.h: Include stringopt.def.
3972 * config/i386/i386.opt: Include stringopt.opt.
3973 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 3974 Override default size based stringop inline strategies with options.
ad83025e
XDL
3975 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
3976 New function.
3977
e59df5fd
JH
39782013-08-09 Jan Hubicka <jh@suse.cz>
3979
c97fd4b8 3980 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 3981
634ab819
JH
39822013-08-09 Jan Hubicka <jh@suse.cz>
3983
3984 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
3985 CGRAPH_FREQ_MAX.
3986 (dump_cgraph_node): Dump profile-id.
3987 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
3988 and common_target_probability.
3989 * lto-cgraph.c (lto_output_edge): Stream common targets.
3990 (lto_output_node): Stream profile ids.
3991 (input_node): Stream profile ids.
3992 (input_edge): Stream common targets.
3993 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
3994 * ipa.c: Include value-prof.h
3995 (ipa_profile_generate_summary): Turn indirect call statement histograms
3996 into common targets.
3997 (ipa_profile): Turn common targets into speculative edges.
3998
2fa3d31b
JH
39992013-08-09 Jan Hubicka <jh@suse.cz>
4000
4001 * cgraph.h (cgraph_node): Add profile_id.
4002 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 4003 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH
4004 (del_node_map): Update.
4005 (find_func_by_funcdef_no): Replace by ...
4006 (find_func_by_profile_id): ... this one.
4007 (gimple_ic_transform): Do not remove useful histograms when
4008 speculation is not done; dump info when indirect call removal
4009 can happen at LTO.
4010 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
4011 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
4012 (__gcov_indirect_call_profiler_v2): .. this one.
4013 * profile.h (init_node_map): Update.
4014 * coverage.c (coverage_compute_profile_id): New function.
4015 * coverage.h (coverage_compute_profile_id): Declare.
4016 * tree-profile.c (init_ic_make_global_vars): Make
4017 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
4018 (gimple_init_edge_profiler): Update prototype of
4019 __gcov_indirect_call_profiler.
4020 (gimple_gen_ic_func_profiler): Simplify.
4021 (tree_profiling): Use init_node_map
4022
042ae7d2
JH
40232013-08-09 Jan Hubicka <jh@suse.cz>
4024
c97fd4b8
UB
4025 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
4026 non-speculative refs.
042ae7d2
JH
4027 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
4028 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
4029 (cgraph_set_call_stmt): Likewise.
4030 (cgraph_create_edge_1): Fix release checking compilatoin;
4031 clear lto_stmt_uid.
4032 (cgraph_free_edge): Free indirect info.
4033 (cgraph_turn_edge_to_speculative): New function.
4034 (cgraph_speculative_call_info): New function.
4035 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 4036 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH
4037 (dump_cgraph_node): Dump speculation.
4038 (verify_edge_count_and_frequency): Accept speculative edges.
4039 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
4040 (verify_cgraph_node): Handle speculation.
4041 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
4042 (cgraph_set_call_stmt): Update prototype.
4043 (cgraph_make_edge_direct): Update prototype.
4044 (cgraph_speculative_call_info): Declare.
4045 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
4046 to change; update call of ipa_find_references.
4047 * ipa-ref.c (ipa_record_reference): Fix return value; clear
4048 lto_stmt_uid and speculative flags.
4049 (ipa_dump_references): Dump speculation.
4050 (ipa_clone_references): Clone speculative flag.
4051 (ipa_clone_referring): Likewise.
4052 (ipa_clone_ref): New function.
4053 (ipa_find_reference): Look into lto_stmt_uids
4054 (ipa_clear_stmts_in_references): Do not clear speculative calls.
4055 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
4056 (ipa_find_reference): Update declaration.
4057 (ipa_clone_ref): Declare.
4058 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
4059 stream speculative flag.
4060 (lto_output_ref): Stream statements uids and speculation.
4061 (input_ref): Likewise.
4062 (input_edge): Stream speuclation.
4063 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
4064 (cgraph_set_call_stmt_including_clones): Handle speculation.
4065 * ipa-inline.c (heap_edge_removal_hook): New function.
4066 (inline_small_functions): Register it.
4067 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
4068 also initialize refs.
4069 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
4070 edge to change.
4071 (try_make_edge_direct_simple_call): Likewise.
4072 (try_make_edge_direct_simple_call): Likewise.
4073 (update_indirect_edges_after_inlining): Likewise.
4074 (remove_described_reference): Look proper lto_stmt_uid.
4075 (propagate_controlled_uses): Likewise.
4076 (propagate_controlled_uses): Liekwise.
4077 * tree-inline.c (copy_bb): Copy speculative edges.
4078 (redirect_all_calls): New function.
c97fd4b8
UB
4079 (copy_cfg_body): Do redirection after loop info is updated.
4080 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 4081
aa1e10cc
JH
40822013-08-09 Jan Hubicka <jh@suse.cz>
4083
4084 * lto-streamer-out.c (output_function): Renumber PHIs.
4085 * lto-streamer-in.c (input_function): Likewise.
4086
66adb8eb
JG
40872013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
4088
4089 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
4090 (get_lane_unsigned): Likewise.
4091 (dup_lane_scalar): Likewise.
4092 (get_lane): enable for VALL.
4093 * config/aarch64/aarch64-simd.md
4094 (aarch64_dup_lane_scalar<mode>): Remove.
4095 (aarch64_get_lane_signed<mode>): Likewise.
4096 (aarch64_get_lane_unsigned<mode>): Likewise.
4097 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
4098 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
4099 (aarch64_get_lane<mode>): Enable for all vector modes.
4100 (aarch64_get_lanedi): Remove misleading constraints.
4101 * config/aarch64/arm_neon.h
4102 (__aarch64_vget_lane_any): Define.
4103 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
4104 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
4105 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
4106 * config/aarch64/iterators.md (VDQQH): New.
4107 (VDQQHS): Likewise.
4108 (vwcore): Likewise.
4109
1f65ae7a
EB
41102013-08-09 Eric Botcazou <ebotcazou@adacore.com>
4111
4112 * configure.ac: Add GAS check for LEON instructions on SPARC.
4113 * configure: Regenerate.
4114 * config.in: Likewise.
4115 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
4116 sparc*-*-* block.
4117 * config/sparc/sparc.opt (LEON, LEON3): New masks.
4118 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
4119 for LEON or LEON3.
4120 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
4121 (AS_LEON_FLAG): New macro.
4122 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
4123 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
4124 Deal with LEON and LEON3 for the memory model.
23a6cb78 4125 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB
4126 (atomic_compare_and_swap<mode>_1): Likewise.
4127 (*atomic_compare_and_swap<mode>_1): Likewise.
4128
ff522f7f
ZC
41292013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
4130
4131 * config/arm/neon.md (vcond): Fix floating-point vector
4132 comparisons against 0.
4133
cb1cca12
VM
41342013-08-08 Vladimir Makarov <vmakarov@redhat.com>
4135
4136 * lra-constraints.c (emit_spill_move): Remove assert.
4137 (process_alt_operands): Add more debugging
4138 output. Increase reject for spilling into memory. Decrease
4139 reject for reloading scratch.
4140 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
4141
a3719e31
SE
41422013-08-08 Steve Ellcey <sellcey@mips.com>
4143
4144 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
4145 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
4146 micromips incompatible. Add nan2008.
4147 (MULTILIB_DIRNAMES): Add nan2008.
4148 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4149 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
4150 and micromips incompatible. Add nan2008.
4151 (MULTILIB_DIRNAMES): Add nan2008.
4152 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
4153
f40423e2
RS
41542013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
4155
4156 PR rtl-optimization/58079
4157 * combine.c (combine_simplify_rtx): Avoid using SUBST if
4158 simplify_comparison has widened a comparison with an integer.
4159
869b9125
KT
41602013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4161
4162 * config/arm/neon.md (movmisalign<mode>): Disable when we
4163 don't allow unaligned accesses.
4164 (*movmisalign<mode>_neon_store): Likewise.
4165 (*movmisalign<mode>_neon_load): Likewise.
4166 (*movmisalign<mode>_neon_store): Likewise.
4167 (*movmisalign<mode>_neon_load): Likewise.
4168
71cafea9
JH
41692013-08-08 Jan Hubicka <jh@suse.cz>
4170
4171 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
4172 (make_pass_rebuild_cgraph_edges): Also clear references.
4173 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
4174 * ipa-inline-transform.c (inline_transform): Remove all references
4175 after inlining.
c97fd4b8
UB
4176 * cgraphunit.c (expand_function): Remove all references after
4177 expansion.
71cafea9
JH
4178 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
4179 (ipa_find_reference): Rewrite to iterator.
4180 (remove_stmt_references): Likewise.
4181 (ipa_clear_stmts_in_references): New function.
4182 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB
4183 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
4184 clear references.
71cafea9
JH
4185 * ipa-split.c (split_function): Remove references in split function.
4186
27d2e612
RE
41872013-08-08 Richard Earnshaw <rearnsha@arm.com>
4188
4189 PR target/57431
1f65ae7a 4190 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE
4191 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
4192
5efc661c
RE
41932013-08-08 Richard Earnshaw <rearnsha@arm.com>
4194
4195 PR target/56979
1f65ae7a 4196 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE
4197 suggested mode for the assignment isn't compatible with the
4198 registers required.
4199
f276d31d
BE
42002013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
4201
4202 PR target/58065
4203 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
4204
4b8bc035
XDL
42052013-08-07 Xinliang David Li <davidxl@google.com>
4206
4207 * config/i386/i386.opt: New option -mtune-ctrl=.
4208 * config/i386/x86-tune.def: New file.
4209 * config/i386/i386.h: include x86-tune.def.
4210 * config/i386/i386.c (ix86_option_override_internal):
4211 Parsing -mtune-ctrl= option and set tune features.
4212
12211b99 42132013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE
4214
4215 PR other/12081
03b0ee0a 4216 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE
4217 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
4218 to gen_2arg_fn_t.
4219
5cf6635b
EB
42202013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4221
4222 * rtl.h (update_alignments): Declare.
4223 * final.c (grow_label_align): New function extracted from...
4224 (shorten_branches): ...here. Call it.
4225 (update_alignments): New function.
4226 * reorg.c (sibling_labels): New variable.
4227 (get_label_before): Add SIBLING parameter. If it is non-zero, push
4228 the new label along with it onto the sibling_labels vector.
4229 (fill_simple_delay_slots): Adjust call to get_label_before.
4230 (fill_slots_from_thread): Likewise.
4231 (relax_delay_slots): Likewise.
4232 (make_return_insns): Likewise.
4233 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
4234
3c8ca1ab
EB
42352013-08-07 Eric Botcazou <ebotcazou@adacore.com>
4236
4237 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
4238 document its semantics.
4239 (diagnostic_report_diagnostic): Adjust accordingly.
4240
5ee5b32c
DM
42412013-08-07 David Malcolm <dmalcolm@redhat.com>
4242
03b0ee0a 4243 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM
4244 (sparc_option_override): ...and port to new C++ pass API.
4245 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
4246
abd566fa
PB
42472013-08-07 Peter Bergner <bergner@vnet.ibm.com>
4248
4249 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
4250
2077db1b
CT
42512013-08-06 Caroline Tice <cmtice@google.com>
4252
4253 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
4254 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
4255 * tree-pass.h: Add pass_vtable_verify.
4256 * varasm.c (assemble_variable): Add code to properly set the comdat
4257 section and name for the .vtable_map_vars section.
4258 (assemble_vtyv_preinit_initializer): New function.
4259 (default_sectin_type_flags): Make sure .vtable_map_vars section has
4260 LINK_ONCE flag.
4261 * output.h: Add function decl for assemble_vtv_preinit_initializer.
4262 * vtable-verify.c: New file.
4263 * vtable-verify.h: New file.
4264 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
4265 initialiation levels.
4266 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
4267 * passes.def: Insert pass_vtable_verify.
4268 * aclocal.m4: Reorder includes.
4269 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
4270 -fvtv-counts options.
4271 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
4272 as appropriate, if -fvtable-verify=... is used.
4273 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
4274 -fvtable-verify=... is used.
4275 * Makefile.in (OBJS): Add vtable-verify.o to list.
4276 (vtable-verify.o): Add new build rule.
4277 (GTFILES): Add vtable-verify.c to list.
4278 * common.opt (fvtable-verify=): New flag.
4279 (vtv_priority): Values for fvtable-verify= flag.
4280 (fvtv-counts): New flag.
4281 (fvtv-debug): New flag.
4282 * tree.h (save_vtable_map_decl): New extern function decl.
4283
03085d1c
DM
42842013-08-07 David Malcolm <dmalcolm@redhat.com>
4285
4286 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
4287 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
4288 (pass_data_rl78_devirt): ...new pass_data instance and...
4289 (make_pass_rl78_devirt): ...new function.
4290 (rl78_asm_file_start): Port pass registration to new C++ API.
4291
05555c4a
DM
42922013-08-07 David Malcolm <dmalcolm@redhat.com>
4293
4294 * coretypes.h (rtl_opt_pass): Add.
4295 (gcc::context): Add.
4296 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
4297 (epiphany_init): Port to new C++ pass API.
4298 (epiphany_optimize_mode_switching): Likewise.
4299 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
4300 (pass_manager::get_pass_mode_switching): New.
4301 (pass_manager::get_pass_peephole2): New.
4302 * mode-switching.c (pass_mode_switching): Add clone method.
4303 * recog.c (pass_peephole2): Add clone method.
4304 (pass_split_all_insns): Add clone method.
4305
c5a12e26
DM
43062013-08-06 David Malcolm <dmalcolm@redhat.com>
4307
03b0ee0a
UB
4308 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
4309 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 4310
a2e2a668
JH
43112013-08-06 Jan Hubicka <jh@suse.cz>
4312
4313 * cgraph.c (cgraph_get_body): New function based on lto.c
4314 implementation.
4315 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB
4316 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
4317 LTO paths.
a2e2a668
JH
4318 * cgraphunit.c (expand_function): Get body prior expanding.
4319 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB
4320 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
4321 really need.
a2e2a668
JH
4322 * passes.c (do_per_function_toporder): Get body.
4323 * tree-inline.c (expand_call_inline): Get body prior inlining it.
4324 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
4325
15682f24
MJ
43262013-08-06 Martin Jambor <mjambor@suse.cz>
4327
4328 PR fortran/57987
4329 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
4330 is not re-finalized. Rename second parameter to no_collect.
4331
78f6dd68
MJ
43322013-08-06 Martin Jambor <mjambor@suse.cz>
4333
4334 PR middle-end/58041
4335 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
4336 MEM_REF has proper alignment information.
4337
12211b99 43382013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE
4339
4340 PR other/12081
4341 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
4342 class insn_gen_fn.
4343 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
4344 rtx (*) (rtx, ...) with insn_gen_fn.
4345 * genoutput.c (output_insn_data): Cast gen_? function pointers to
4346 insn_gen_fn::stored_funcptr. Add initializer braces.
4347
8ac69a6c
DM
43482013-08-05 David Malcolm <dmalcolm@redhat.com>
4349
4350 Rewrite how instances of passes are cloned to remove assumptions
4351 about their sizes (thus allowing pass subclasses to have
4352 additional data fields, albeit non-GC-managed ones at this point).
4353
4354 * passes.c (make_pass_instance): Now that passes have clone
4355 methods, rewrite this function to eliminate XNEW and memcpy
4356 calls that used hardcoded sizes. Since this function no longer
4357 creates pass instances, rename it to...
03b0ee0a
UB
4358 (add_pass_instance): ...this. Document the old way that passes were
4359 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM
4360 (next_pass_1): Add an initial_pass argument for use by
4361 add_pass_instance.
4362 (position_pass): When adding multiple instances of a pass, use
4363 the pass's clone method, rather than relying on the XNEW/memcpy
4364 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB
4365 (pass_manager::pass_manager): When invoking next_pass_1, also supply
4366 the initial instance of the current pass within the pass manager.
8ac69a6c 4367
27a4cd48
DM
43682013-08-05 David Malcolm <dmalcolm@redhat.com>
4369
4370 This is the automated part of the conversion of passes from C
4371 structs to C++ classes.
4372
4373 Patch autogenerated by refactor_passes.py from
4374 https://github.com/davidmalcolm/gcc-refactoring-scripts
4375 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
4376
4377 * asan.c (pass_asan): Convert from a global struct to a subclass of
4378 gimple_opt_pass along with...
4379 (pass_data_asan): ...new pass_data instance and...
4380 (make_pass_asan): ...new function.
4381 (pass_asan_O0): Convert from a global struct to a subclass of
4382 gimple_opt_pass along with...
4383 (pass_data_asan_O0): ...new pass_data instance and...
4384 (make_pass_asan_O0): ...new function.
4385 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
4386 subclass of rtl_opt_pass along with...
4387 (pass_data_inc_dec): ...new pass_data instance and...
4388 (make_pass_inc_dec): ...new function.
4389 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
4390 a subclass of rtl_opt_pass along with...
4391 (pass_data_reorder_blocks): ...new pass_data instance and...
4392 (make_pass_reorder_blocks): ...new function.
4393 (pass_duplicate_computed_gotos): Convert from a global struct to a
4394 subclass of rtl_opt_pass along with...
4395 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
4396 (make_pass_duplicate_computed_gotos): ...new function.
4397 (pass_partition_blocks): Convert from a global struct to a subclass of
4398 rtl_opt_pass along with...
4399 (pass_data_partition_blocks): ...new pass_data instance and...
4400 (make_pass_partition_blocks): ...new function.
4401 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
4402 struct to a subclass of rtl_opt_pass along with...
4403 (pass_data_branch_target_load_optimize1): ...new pass_data instance
4404 and...
4405 (make_pass_branch_target_load_optimize1): ...new function.
4406 (pass_branch_target_load_optimize2): Convert from a global struct to a
4407 subclass of rtl_opt_pass along with...
4408 (pass_data_branch_target_load_optimize2): ...new pass_data instance
4409 and...
4410 (make_pass_branch_target_load_optimize2): ...new function.
4411 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
4412 of rtl_opt_pass along with...
4413 (pass_data_jump): ...new pass_data instance and...
4414 (make_pass_jump): ...new function.
4415 (pass_jump2): Convert from a global struct to a subclass of
4416 rtl_opt_pass along with...
4417 (pass_data_jump2): ...new pass_data instance and...
4418 (make_pass_jump2): ...new function.
4419 * cfgexpand.c (pass_expand): Convert from a global struct to a
4420 subclass of rtl_opt_pass along with...
4421 (pass_data_expand): ...new pass_data instance and...
4422 (make_pass_expand): ...new function.
4423 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
4424 of rtl_opt_pass along with...
4425 (pass_data_free_cfg): ...new pass_data instance and...
4426 (make_pass_free_cfg): ...new function.
4427 (pass_into_cfg_layout_mode): Convert from a global struct to a
4428 subclass of rtl_opt_pass along with...
4429 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
4430 (make_pass_into_cfg_layout_mode): ...new function.
4431 (pass_outof_cfg_layout_mode): Convert from a global struct to a
4432 subclass of rtl_opt_pass along with...
4433 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
4434 (make_pass_outof_cfg_layout_mode): ...new function.
4435 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
4436 struct to a subclass of gimple_opt_pass along with...
4437 (pass_data_build_cgraph_edges): ...new pass_data instance and...
4438 (make_pass_build_cgraph_edges): ...new function.
4439 (pass_rebuild_cgraph_edges): Convert from a global struct to a
4440 subclass of gimple_opt_pass along with...
4441 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
4442 (make_pass_rebuild_cgraph_edges): ...new function.
4443 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
4444 subclass of gimple_opt_pass along with...
4445 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
4446 and...
4447 (make_pass_remove_cgraph_callee_edges): ...new function.
4448 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
4449 struct to a subclass of rtl_opt_pass along with...
4450 (pass_data_stack_adjustments): ...new pass_data instance and...
4451 (make_pass_stack_adjustments): ...new function.
4452 * combine.c (pass_combine): Convert from a global struct to a subclass
4453 of rtl_opt_pass along with...
4454 (pass_data_combine): ...new pass_data instance and...
4455 (make_pass_combine): ...new function.
4456 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
4457 global struct to a subclass of rtl_opt_pass along with...
4458 (pass_data_compare_elim_after_reload): ...new pass_data instance
4459 and...
4460 (make_pass_compare_elim_after_reload): ...new function.
4461 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
4462 of rtl_opt_pass along with...
4463 (pass_data_rtl_cprop): ...new pass_data instance and...
4464 (make_pass_rtl_cprop): ...new function.
4465 * cse.c (pass_cse): Convert from a global struct to a subclass of
4466 rtl_opt_pass along with...
4467 (pass_data_cse): ...new pass_data instance and...
4468 (make_pass_cse): ...new function.
4469 (pass_cse2): Convert from a global struct to a subclass of
4470 rtl_opt_pass along with...
4471 (pass_data_cse2): ...new pass_data instance and...
4472 (make_pass_cse2): ...new function.
4473 (pass_cse_after_global_opts): Convert from a global struct to a
4474 subclass of rtl_opt_pass along with...
4475 (pass_data_cse_after_global_opts): ...new pass_data instance and...
4476 (make_pass_cse_after_global_opts): ...new function.
4477 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
4478 of rtl_opt_pass along with...
4479 (pass_data_ud_rtl_dce): ...new pass_data instance and...
4480 (make_pass_ud_rtl_dce): ...new function.
4481 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
4482 rtl_opt_pass along with...
4483 (pass_data_fast_rtl_dce): ...new pass_data instance and...
4484 (make_pass_fast_rtl_dce): ...new function.
4485 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
4486 a subclass of rtl_opt_pass along with...
4487 (pass_data_df_initialize_opt): ...new pass_data instance and...
4488 (make_pass_df_initialize_opt): ...new function.
4489 (pass_df_initialize_no_opt): Convert from a global struct to a
4490 subclass of rtl_opt_pass along with...
4491 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
4492 (make_pass_df_initialize_no_opt): ...new function.
4493 (pass_df_finish): Convert from a global struct to a subclass of
4494 rtl_opt_pass along with...
4495 (pass_data_df_finish): ...new pass_data instance and...
4496 (make_pass_df_finish): ...new function.
4497 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
4498 rtl_opt_pass along with...
4499 (pass_data_rtl_dse1): ...new pass_data instance and...
4500 (make_pass_rtl_dse1): ...new function.
4501 (pass_rtl_dse2): Convert from a global struct to a subclass of
4502 rtl_opt_pass along with...
4503 (pass_data_rtl_dse2): ...new pass_data instance and...
4504 (make_pass_rtl_dse2): ...new function.
4505 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
4506 subclass of rtl_opt_pass along with...
4507 (pass_data_dwarf2_frame): ...new pass_data instance and...
4508 (make_pass_dwarf2_frame): ...new function.
4509 * except.c (pass_set_nothrow_function_flags): Convert from a global
4510 struct to a subclass of rtl_opt_pass along with...
4511 (pass_data_set_nothrow_function_flags): ...new pass_data instance
4512 and...
4513 (make_pass_set_nothrow_function_flags): ...new function.
4514 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
4515 subclass of rtl_opt_pass along with...
4516 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
4517 and...
4518 (make_pass_convert_to_eh_region_ranges): ...new function.
4519 * final.c (pass_compute_alignments): Convert from a global struct to a
4520 subclass of rtl_opt_pass along with...
4521 (pass_data_compute_alignments): ...new pass_data instance and...
4522 (make_pass_compute_alignments): ...new function.
4523 (pass_final): Convert from a global struct to a subclass of
4524 rtl_opt_pass along with...
4525 (pass_data_final): ...new pass_data instance and...
4526 (make_pass_final): ...new function.
4527 (pass_shorten_branches): Convert from a global struct to a subclass of
4528 rtl_opt_pass along with...
4529 (pass_data_shorten_branches): ...new pass_data instance and...
4530 (make_pass_shorten_branches): ...new function.
4531 (pass_clean_state): Convert from a global struct to a subclass of
4532 rtl_opt_pass along with...
4533 (pass_data_clean_state): ...new pass_data instance and...
4534 (make_pass_clean_state): ...new function.
4535 * function.c (pass_instantiate_virtual_regs): Convert from a global
4536 struct to a subclass of rtl_opt_pass along with...
4537 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
4538 (make_pass_instantiate_virtual_regs): ...new function.
4539 (pass_leaf_regs): Convert from a global struct to a subclass of
4540 rtl_opt_pass along with...
4541 (pass_data_leaf_regs): ...new pass_data instance and...
4542 (make_pass_leaf_regs): ...new function.
4543 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
4544 subclass of rtl_opt_pass along with...
4545 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
4546 and...
4547 (make_pass_thread_prologue_and_epilogue): ...new function.
4548 (pass_match_asm_constraints): Convert from a global struct to a
4549 subclass of rtl_opt_pass along with...
4550 (pass_data_match_asm_constraints): ...new pass_data instance and...
4551 (make_pass_match_asm_constraints): ...new function.
4552 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
4553 subclass of rtl_opt_pass along with...
4554 (pass_data_rtl_fwprop): ...new pass_data instance and...
4555 (make_pass_rtl_fwprop): ...new function.
4556 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
4557 rtl_opt_pass along with...
4558 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
4559 (make_pass_rtl_fwprop_addr): ...new function.
4560 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
4561 rtl_opt_pass along with...
4562 (pass_data_rtl_pre): ...new pass_data instance and...
4563 (make_pass_rtl_pre): ...new function.
4564 (pass_rtl_hoist): Convert from a global struct to a subclass of
4565 rtl_opt_pass along with...
4566 (pass_data_rtl_hoist): ...new pass_data instance and...
4567 (make_pass_rtl_hoist): ...new function.
4568 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
4569 subclass of gimple_opt_pass along with...
4570 (pass_data_lower_cf): ...new pass_data instance and...
4571 (make_pass_lower_cf): ...new function.
4572 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
4573 from a global struct to a subclass of gimple_opt_pass along with...
4574 (pass_data_strength_reduction): ...new pass_data instance and...
4575 (make_pass_strength_reduction): ...new function.
4576 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
4577 of rtl_opt_pass along with...
4578 (pass_data_rtl_ifcvt): ...new pass_data instance and...
4579 (make_pass_rtl_ifcvt): ...new function.
4580 (pass_if_after_combine): Convert from a global struct to a subclass of
4581 rtl_opt_pass along with...
4582 (pass_data_if_after_combine): ...new pass_data instance and...
4583 (make_pass_if_after_combine): ...new function.
4584 (pass_if_after_reload): Convert from a global struct to a subclass of
4585 rtl_opt_pass along with...
4586 (pass_data_if_after_reload): ...new pass_data instance and...
4587 (make_pass_if_after_reload): ...new function.
4588 * init-regs.c (pass_initialize_regs): Convert from a global struct to
4589 a subclass of rtl_opt_pass along with...
4590 (pass_data_initialize_regs): ...new pass_data instance and...
4591 (make_pass_initialize_regs): ...new function.
4592 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
4593 of ipa_opt_pass_d along with...
4594 (pass_data_ipa_cp): ...new pass_data instance and...
4595 (make_pass_ipa_cp): ...new function.
4596 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
4597 global struct to a subclass of gimple_opt_pass along with...
4598 (pass_data_inline_parameters): ...new pass_data instance and...
4599 (make_pass_inline_parameters): ...new function.
4600 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
4601 subclass of gimple_opt_pass along with...
4602 (pass_data_early_inline): ...new pass_data instance and...
4603 (make_pass_early_inline): ...new function.
4604 (pass_ipa_inline): Convert from a global struct to a subclass of
4605 ipa_opt_pass_d along with...
4606 (pass_data_ipa_inline): ...new pass_data instance and...
4607 (make_pass_ipa_inline): ...new function.
4608 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
4609 struct to a subclass of gimple_opt_pass along with...
4610 (pass_data_local_pure_const): ...new pass_data instance and...
4611 (make_pass_local_pure_const): ...new function.
4612 (pass_ipa_pure_const): Convert from a global struct to a subclass of
4613 ipa_opt_pass_d along with...
4614 (pass_data_ipa_pure_const): ...new pass_data instance and...
4615 (make_pass_ipa_pure_const): ...new function.
4616 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
4617 to a subclass of ipa_opt_pass_d along with...
4618 (pass_data_ipa_reference): ...new pass_data instance and...
4619 (make_pass_ipa_reference): ...new function.
4620 * ipa-split.c (pass_split_functions): Convert from a global struct to
4621 a subclass of gimple_opt_pass along with...
4622 (pass_data_split_functions): ...new pass_data instance and...
4623 (make_pass_split_functions): ...new function.
4624 (pass_feedback_split_functions): Convert from a global struct to a
4625 subclass of gimple_opt_pass along with...
4626 (pass_data_feedback_split_functions): ...new pass_data instance and...
4627 (make_pass_feedback_split_functions): ...new function.
4628 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
4629 global struct to a subclass of simple_ipa_opt_pass along with...
4630 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
4631 instance and...
4632 (make_pass_ipa_function_and_variable_visibility): ...new function.
4633 (pass_ipa_free_inline_summary): Convert from a global struct to a
4634 subclass of simple_ipa_opt_pass along with...
4635 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
4636 (make_pass_ipa_free_inline_summary): ...new function.
4637 (pass_ipa_whole_program_visibility): Convert from a global struct to a
4638 subclass of ipa_opt_pass_d along with...
4639 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
4640 and...
4641 (make_pass_ipa_whole_program_visibility): ...new function.
4642 (pass_ipa_profile): Convert from a global struct to a subclass of
4643 ipa_opt_pass_d along with...
4644 (pass_data_ipa_profile): ...new pass_data instance and...
4645 (make_pass_ipa_profile): ...new function.
4646 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
4647 ipa_opt_pass_d along with...
4648 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
4649 (make_pass_ipa_cdtor_merge): ...new function.
4650 * ira.c (pass_ira): Convert from a global struct to a subclass of
4651 rtl_opt_pass along with...
4652 (pass_data_ira): ...new pass_data instance and...
4653 (make_pass_ira): ...new function.
4654 (pass_reload): Convert from a global struct to a subclass of
4655 rtl_opt_pass along with...
4656 (pass_data_reload): ...new pass_data instance and...
4657 (make_pass_reload): ...new function.
4658 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
4659 subclass of rtl_opt_pass along with...
4660 (pass_data_cleanup_barriers): ...new pass_data instance and...
4661 (make_pass_cleanup_barriers): ...new function.
4662 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
4663 of rtl_opt_pass along with...
4664 (pass_data_loop2): ...new pass_data instance and...
4665 (make_pass_loop2): ...new function.
4666 (pass_rtl_loop_init): Convert from a global struct to a subclass of
4667 rtl_opt_pass along with...
4668 (pass_data_rtl_loop_init): ...new pass_data instance and...
4669 (make_pass_rtl_loop_init): ...new function.
4670 (pass_rtl_loop_done): Convert from a global struct to a subclass of
4671 rtl_opt_pass along with...
4672 (pass_data_rtl_loop_done): ...new pass_data instance and...
4673 (make_pass_rtl_loop_done): ...new function.
4674 (pass_rtl_move_loop_invariants): Convert from a global struct to a
4675 subclass of rtl_opt_pass along with...
4676 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
4677 (make_pass_rtl_move_loop_invariants): ...new function.
4678 (pass_rtl_unswitch): Convert from a global struct to a subclass of
4679 rtl_opt_pass along with...
4680 (pass_data_rtl_unswitch): ...new pass_data instance and...
4681 (make_pass_rtl_unswitch): ...new function.
4682 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
4683 subclass of rtl_opt_pass along with...
4684 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
4685 and...
4686 (make_pass_rtl_unroll_and_peel_loops): ...new function.
4687 (pass_rtl_doloop): Convert from a global struct to a subclass of
4688 rtl_opt_pass along with...
4689 (pass_data_rtl_doloop): ...new pass_data instance and...
4690 (make_pass_rtl_doloop): ...new function.
4691 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
4692 a subclass of rtl_opt_pass along with...
4693 (pass_data_lower_subreg): ...new pass_data instance and...
4694 (make_pass_lower_subreg): ...new function.
4695 (pass_lower_subreg2): Convert from a global struct to a subclass of
4696 rtl_opt_pass along with...
4697 (pass_data_lower_subreg2): ...new pass_data instance and...
4698 (make_pass_lower_subreg2): ...new function.
4699 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
4700 struct to a subclass of ipa_opt_pass_d along with...
4701 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
4702 (make_pass_ipa_lto_gimple_out): ...new function.
4703 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
4704 of ipa_opt_pass_d along with...
4705 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
4706 (make_pass_ipa_lto_finish_out): ...new function.
4707 * mode-switching.c (pass_mode_switching): Convert from a global struct
4708 to a subclass of rtl_opt_pass along with...
4709 (pass_data_mode_switching): ...new pass_data instance and...
4710 (make_pass_mode_switching): ...new function.
4711 * modulo-sched.c (pass_sms): Convert from a global struct to a
4712 subclass of rtl_opt_pass along with...
4713 (pass_data_sms): ...new pass_data instance and...
4714 (make_pass_sms): ...new function.
4715 * omp-low.c (pass_expand_omp): Convert from a global struct to a
4716 subclass of gimple_opt_pass along with...
4717 (pass_data_expand_omp): ...new pass_data instance and...
4718 (make_pass_expand_omp): ...new function.
4719 (pass_lower_omp): Convert from a global struct to a subclass of
4720 gimple_opt_pass along with...
4721 (pass_data_lower_omp): ...new pass_data instance and...
4722 (make_pass_lower_omp): ...new function.
4723 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
4724 of gimple_opt_pass along with...
4725 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
4726 (make_pass_diagnose_omp_blocks): ...new function.
4727 * passes.c (pass_early_local_passes): Convert from a global struct to
4728 a subclass of simple_ipa_opt_pass along with...
4729 (pass_data_early_local_passes): ...new pass_data instance and...
4730 (make_pass_early_local_passes): ...new function.
4731 (pass_all_early_optimizations): Convert from a global struct to a
4732 subclass of gimple_opt_pass along with...
4733 (pass_data_all_early_optimizations): ...new pass_data instance and...
4734 (make_pass_all_early_optimizations): ...new function.
4735 (pass_all_optimizations): Convert from a global struct to a subclass
4736 of gimple_opt_pass along with...
4737 (pass_data_all_optimizations): ...new pass_data instance and...
4738 (make_pass_all_optimizations): ...new function.
4739 (pass_all_optimizations_g): Convert from a global struct to a subclass
4740 of gimple_opt_pass along with...
4741 (pass_data_all_optimizations_g): ...new pass_data instance and...
4742 (make_pass_all_optimizations_g): ...new function.
4743 (pass_rest_of_compilation): Convert from a global struct to a subclass
4744 of rtl_opt_pass along with...
4745 (pass_data_rest_of_compilation): ...new pass_data instance and...
4746 (make_pass_rest_of_compilation): ...new function.
4747 (pass_postreload): Convert from a global struct to a subclass of
4748 rtl_opt_pass along with...
4749 (pass_data_postreload): ...new pass_data instance and...
4750 (make_pass_postreload): ...new function.
4751 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
4752 subclass of rtl_opt_pass along with...
4753 (pass_data_gcse2): ...new pass_data instance and...
4754 (make_pass_gcse2): ...new function.
4755 * postreload.c (pass_postreload_cse): Convert from a global struct to
4756 a subclass of rtl_opt_pass along with...
4757 (pass_data_postreload_cse): ...new pass_data instance and...
4758 (make_pass_postreload_cse): ...new function.
4759 * predict.c (pass_profile): Convert from a global struct to a subclass
4760 of gimple_opt_pass along with...
4761 (pass_data_profile): ...new pass_data instance and...
4762 (make_pass_profile): ...new function.
4763 (pass_strip_predict_hints): Convert from a global struct to a subclass
4764 of gimple_opt_pass along with...
4765 (pass_data_strip_predict_hints): ...new pass_data instance and...
4766 (make_pass_strip_predict_hints): ...new function.
4767 * recog.c (pass_peephole2): Convert from a global struct to a subclass
4768 of rtl_opt_pass along with...
4769 (pass_data_peephole2): ...new pass_data instance and...
4770 (make_pass_peephole2): ...new function.
4771 (pass_split_all_insns): Convert from a global struct to a subclass of
4772 rtl_opt_pass along with...
4773 (pass_data_split_all_insns): ...new pass_data instance and...
4774 (make_pass_split_all_insns): ...new function.
4775 (pass_split_after_reload): Convert from a global struct to a subclass
4776 of rtl_opt_pass along with...
4777 (pass_data_split_after_reload): ...new pass_data instance and...
4778 (make_pass_split_after_reload): ...new function.
4779 (pass_split_before_regstack): Convert from a global struct to a
4780 subclass of rtl_opt_pass along with...
4781 (pass_data_split_before_regstack): ...new pass_data instance and...
4782 (make_pass_split_before_regstack): ...new function.
4783 (pass_split_before_sched2): Convert from a global struct to a subclass
4784 of rtl_opt_pass along with...
4785 (pass_data_split_before_sched2): ...new pass_data instance and...
4786 (make_pass_split_before_sched2): ...new function.
4787 (pass_split_for_shorten_branches): Convert from a global struct to a
4788 subclass of rtl_opt_pass along with...
4789 (pass_data_split_for_shorten_branches): ...new pass_data instance
4790 and...
4791 (make_pass_split_for_shorten_branches): ...new function.
4792 * ree.c (pass_ree): Convert from a global struct to a subclass of
4793 rtl_opt_pass along with...
4794 (pass_data_ree): ...new pass_data instance and...
4795 (make_pass_ree): ...new function.
4796 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
4797 subclass of rtl_opt_pass along with...
4798 (pass_data_stack_regs): ...new pass_data instance and...
4799 (make_pass_stack_regs): ...new function.
4800 (pass_stack_regs_run): Convert from a global struct to a subclass of
4801 rtl_opt_pass along with...
4802 (pass_data_stack_regs_run): ...new pass_data instance and...
4803 (make_pass_stack_regs_run): ...new function.
4804 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
4805 subclass of rtl_opt_pass along with...
4806 (pass_data_cprop_hardreg): ...new pass_data instance and...
4807 (make_pass_cprop_hardreg): ...new function.
4808 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
4809 subclass of rtl_opt_pass along with...
4810 (pass_data_reginfo_init): ...new pass_data instance and...
4811 (make_pass_reginfo_init): ...new function.
4812 * regmove.c (pass_regmove): Convert from a global struct to a subclass
4813 of rtl_opt_pass along with...
4814 (pass_data_regmove): ...new pass_data instance and...
4815 (make_pass_regmove): ...new function.
4816 * regrename.c (pass_regrename): Convert from a global struct to a
4817 subclass of rtl_opt_pass along with...
4818 (pass_data_regrename): ...new pass_data instance and...
4819 (make_pass_regrename): ...new function.
4820 * reorg.c (pass_delay_slots): Convert from a global struct to a
4821 subclass of rtl_opt_pass along with...
4822 (pass_data_delay_slots): ...new pass_data instance and...
4823 (make_pass_delay_slots): ...new function.
4824 (pass_machine_reorg): Convert from a global struct to a subclass of
4825 rtl_opt_pass along with...
4826 (pass_data_machine_reorg): ...new pass_data instance and...
4827 (make_pass_machine_reorg): ...new function.
4828 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
4829 of rtl_opt_pass along with...
4830 (pass_data_sched): ...new pass_data instance and...
4831 (make_pass_sched): ...new function.
4832 (pass_sched2): Convert from a global struct to a subclass of
4833 rtl_opt_pass along with...
4834 (pass_data_sched2): ...new pass_data instance and...
4835 (make_pass_sched2): ...new function.
4836 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
4837 to a subclass of rtl_opt_pass along with...
4838 (pass_data_stack_ptr_mod): ...new pass_data instance and...
4839 (make_pass_stack_ptr_mod): ...new function.
4840 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
4841 to a subclass of rtl_opt_pass along with...
4842 (pass_data_rtl_store_motion): ...new pass_data instance and...
4843 (make_pass_rtl_store_motion): ...new function.
4844 * tracer.c (pass_tracer): Convert from a global struct to a subclass
4845 of gimple_opt_pass along with...
4846 (pass_data_tracer): ...new pass_data instance and...
4847 (make_pass_tracer): ...new function.
4848 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
4849 to a subclass of gimple_opt_pass along with...
4850 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
4851 (make_pass_diagnose_tm_blocks): ...new function.
4852 (pass_lower_tm): Convert from a global struct to a subclass of
4853 gimple_opt_pass along with...
4854 (pass_data_lower_tm): ...new pass_data instance and...
4855 (make_pass_lower_tm): ...new function.
4856 (pass_tm_init): Convert from a global struct to a subclass of
4857 gimple_opt_pass along with...
4858 (pass_data_tm_init): ...new pass_data instance and...
4859 (make_pass_tm_init): ...new function.
4860 (pass_tm_mark): Convert from a global struct to a subclass of
4861 gimple_opt_pass along with...
4862 (pass_data_tm_mark): ...new pass_data instance and...
4863 (make_pass_tm_mark): ...new function.
4864 (pass_tm_edges): Convert from a global struct to a subclass of
4865 gimple_opt_pass along with...
4866 (pass_data_tm_edges): ...new pass_data instance and...
4867 (make_pass_tm_edges): ...new function.
4868 (pass_tm_memopt): Convert from a global struct to a subclass of
4869 gimple_opt_pass along with...
4870 (pass_data_tm_memopt): ...new pass_data instance and...
4871 (make_pass_tm_memopt): ...new function.
4872 (pass_ipa_tm): Convert from a global struct to a subclass of
4873 simple_ipa_opt_pass along with...
4874 (pass_data_ipa_tm): ...new pass_data instance and...
4875 (make_pass_ipa_tm): ...new function.
4876 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
4877 subclass of gimple_opt_pass along with...
4878 (pass_data_call_cdce): ...new pass_data instance and...
4879 (make_pass_call_cdce): ...new function.
4880 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
4881 subclass of gimple_opt_pass along with...
4882 (pass_data_build_cfg): ...new pass_data instance and...
4883 (make_pass_build_cfg): ...new function.
4884 (pass_split_crit_edges): Convert from a global struct to a subclass of
4885 gimple_opt_pass along with...
4886 (pass_data_split_crit_edges): ...new pass_data instance and...
4887 (make_pass_split_crit_edges): ...new function.
4888 (pass_warn_function_return): Convert from a global struct to a
4889 subclass of gimple_opt_pass along with...
4890 (pass_data_warn_function_return): ...new pass_data instance and...
4891 (make_pass_warn_function_return): ...new function.
4892 (pass_warn_function_noreturn): Convert from a global struct to a
4893 subclass of gimple_opt_pass along with...
4894 (pass_data_warn_function_noreturn): ...new pass_data instance and...
4895 (make_pass_warn_function_noreturn): ...new function.
4896 (pass_warn_unused_result): Convert from a global struct to a subclass
4897 of gimple_opt_pass along with...
4898 (pass_data_warn_unused_result): ...new pass_data instance and...
4899 (make_pass_warn_unused_result): ...new function.
4900 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
4901 a subclass of gimple_opt_pass along with...
4902 (pass_data_merge_phi): ...new pass_data instance and...
4903 (make_pass_merge_phi): ...new function.
4904 * tree-complex.c (pass_lower_complex): Convert from a global struct to
4905 a subclass of gimple_opt_pass along with...
4906 (pass_data_lower_complex): ...new pass_data instance and...
4907 (make_pass_lower_complex): ...new function.
4908 (pass_lower_complex_O0): Convert from a global struct to a subclass of
4909 gimple_opt_pass along with...
4910 (pass_data_lower_complex_O0): ...new pass_data instance and...
4911 (make_pass_lower_complex_O0): ...new function.
4912 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
4913 subclass of gimple_opt_pass along with...
4914 (pass_data_lower_eh): ...new pass_data instance and...
4915 (make_pass_lower_eh): ...new function.
4916 (pass_refactor_eh): Convert from a global struct to a subclass of
4917 gimple_opt_pass along with...
4918 (pass_data_refactor_eh): ...new pass_data instance and...
4919 (make_pass_refactor_eh): ...new function.
4920 (pass_lower_resx): Convert from a global struct to a subclass of
4921 gimple_opt_pass along with...
4922 (pass_data_lower_resx): ...new pass_data instance and...
4923 (make_pass_lower_resx): ...new function.
4924 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
4925 of gimple_opt_pass along with...
4926 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
4927 (make_pass_lower_eh_dispatch): ...new function.
4928 (pass_cleanup_eh): Convert from a global struct to a subclass of
4929 gimple_opt_pass along with...
4930 (pass_data_cleanup_eh): ...new pass_data instance and...
4931 (make_pass_cleanup_eh): ...new function.
4932 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
4933 to a subclass of simple_ipa_opt_pass along with...
4934 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
4935 (make_pass_ipa_lower_emutls): ...new function.
4936 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
4937 a subclass of gimple_opt_pass along with...
4938 (pass_data_if_conversion): ...new pass_data instance and...
4939 (make_pass_if_conversion): ...new function.
4940 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
4941 subclass of gimple_opt_pass along with...
4942 (pass_data_build_ssa): ...new pass_data instance and...
4943 (make_pass_build_ssa): ...new function.
4944 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
4945 global struct to a subclass of gimple_opt_pass along with...
4946 (pass_data_loop_distribution): ...new pass_data instance and...
4947 (make_pass_loop_distribution): ...new function.
4948 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
4949 subclass of gimple_opt_pass along with...
4950 (pass_data_mudflap_1): ...new pass_data instance and...
4951 (make_pass_mudflap_1): ...new function.
4952 (pass_mudflap_2): Convert from a global struct to a subclass of
4953 gimple_opt_pass along with...
4954 (pass_data_mudflap_2): ...new pass_data instance and...
4955 (make_pass_mudflap_2): ...new function.
4956 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
4957 subclass of gimple_opt_pass along with...
4958 (pass_data_mudflap_1): ...new pass_data instance and...
4959 (make_pass_mudflap_1): ...new function.
4960 (pass_mudflap_2): Convert from a global struct to a subclass of
4961 gimple_opt_pass along with...
4962 (pass_data_mudflap_2): ...new pass_data instance and...
4963 (make_pass_mudflap_2): ...new function.
4964 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
4965 gimple_opt_pass along with...
4966 (pass_data_nrv): ...new pass_data instance and...
4967 (make_pass_nrv): ...new function.
4968 (pass_return_slot): Convert from a global struct to a subclass of
4969 gimple_opt_pass along with...
4970 (pass_data_return_slot): ...new pass_data instance and...
4971 (make_pass_return_slot): ...new function.
4972 * tree-object-size.c (pass_object_sizes): Convert from a global struct
4973 to a subclass of gimple_opt_pass along with...
4974 (pass_data_object_sizes): ...new pass_data instance and...
4975 (make_pass_object_sizes): ...new function.
4976 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
4977 global struct to a subclass of gimple_opt_pass along with...
4978 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
4979 and...
4980 (make_pass_cleanup_cfg_post_optimizing): ...new function.
4981 (pass_fixup_cfg): Convert from a global struct to a subclass of
4982 gimple_opt_pass along with...
4983 (pass_data_fixup_cfg): ...new pass_data instance and...
4984 (make_pass_fixup_cfg): ...new function.
4985 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
4986 (make_pass_mudflap_1): ...new function.
4987 (pass_mudflap_2): Replace declaration with that of...
4988 (make_pass_mudflap_2): ...new function.
4989 (pass_asan): Replace declaration with that of...
4990 (make_pass_asan): ...new function.
4991 (pass_asan_O0): Replace declaration with that of...
4992 (make_pass_asan_O0): ...new function.
4993 (pass_tsan): Replace declaration with that of...
4994 (make_pass_tsan): ...new function.
4995 (pass_tsan_O0): Replace declaration with that of...
4996 (make_pass_tsan_O0): ...new function.
4997 (pass_lower_cf): Replace declaration with that of...
4998 (make_pass_lower_cf): ...new function.
4999 (pass_refactor_eh): Replace declaration with that of...
5000 (make_pass_refactor_eh): ...new function.
5001 (pass_lower_eh): Replace declaration with that of...
5002 (make_pass_lower_eh): ...new function.
5003 (pass_lower_eh_dispatch): Replace declaration with that of...
5004 (make_pass_lower_eh_dispatch): ...new function.
5005 (pass_lower_resx): Replace declaration with that of...
5006 (make_pass_lower_resx): ...new function.
5007 (pass_build_cfg): Replace declaration with that of...
5008 (make_pass_build_cfg): ...new function.
5009 (pass_early_tree_profile): Replace declaration with that of...
5010 (make_pass_early_tree_profile): ...new function.
5011 (pass_cleanup_eh): Replace declaration with that of...
5012 (make_pass_cleanup_eh): ...new function.
5013 (pass_sra): Replace declaration with that of...
5014 (make_pass_sra): ...new function.
5015 (pass_sra_early): Replace declaration with that of...
5016 (make_pass_sra_early): ...new function.
5017 (pass_early_ipa_sra): Replace declaration with that of...
5018 (make_pass_early_ipa_sra): ...new function.
5019 (pass_tail_recursion): Replace declaration with that of...
5020 (make_pass_tail_recursion): ...new function.
5021 (pass_tail_calls): Replace declaration with that of...
5022 (make_pass_tail_calls): ...new function.
5023 (pass_tree_loop): Replace declaration with that of...
5024 (make_pass_tree_loop): ...new function.
5025 (pass_tree_loop_init): Replace declaration with that of...
5026 (make_pass_tree_loop_init): ...new function.
5027 (pass_lim): Replace declaration with that of...
5028 (make_pass_lim): ...new function.
5029 (pass_tree_unswitch): Replace declaration with that of...
5030 (make_pass_tree_unswitch): ...new function.
5031 (pass_predcom): Replace declaration with that of...
5032 (make_pass_predcom): ...new function.
5033 (pass_iv_canon): Replace declaration with that of...
5034 (make_pass_iv_canon): ...new function.
5035 (pass_scev_cprop): Replace declaration with that of...
5036 (make_pass_scev_cprop): ...new function.
5037 (pass_empty_loop): Replace declaration with that of...
5038 (make_pass_empty_loop): ...new function.
5039 (pass_record_bounds): Replace declaration with that of...
5040 (make_pass_record_bounds): ...new function.
5041 (pass_graphite): Replace declaration with that of...
5042 (make_pass_graphite): ...new function.
5043 (pass_graphite_transforms): Replace declaration with that of...
5044 (make_pass_graphite_transforms): ...new function.
5045 (pass_if_conversion): Replace declaration with that of...
5046 (make_pass_if_conversion): ...new function.
5047 (pass_loop_distribution): Replace declaration with that of...
5048 (make_pass_loop_distribution): ...new function.
5049 (pass_vectorize): Replace declaration with that of...
5050 (make_pass_vectorize): ...new function.
5051 (pass_slp_vectorize): Replace declaration with that of...
5052 (make_pass_slp_vectorize): ...new function.
5053 (pass_complete_unroll): Replace declaration with that of...
5054 (make_pass_complete_unroll): ...new function.
5055 (pass_complete_unrolli): Replace declaration with that of...
5056 (make_pass_complete_unrolli): ...new function.
5057 (pass_parallelize_loops): Replace declaration with that of...
5058 (make_pass_parallelize_loops): ...new function.
5059 (pass_loop_prefetch): Replace declaration with that of...
5060 (make_pass_loop_prefetch): ...new function.
5061 (pass_iv_optimize): Replace declaration with that of...
5062 (make_pass_iv_optimize): ...new function.
5063 (pass_tree_loop_done): Replace declaration with that of...
5064 (make_pass_tree_loop_done): ...new function.
5065 (pass_ch): Replace declaration with that of...
5066 (make_pass_ch): ...new function.
5067 (pass_ccp): Replace declaration with that of...
5068 (make_pass_ccp): ...new function.
5069 (pass_phi_only_cprop): Replace declaration with that of...
5070 (make_pass_phi_only_cprop): ...new function.
5071 (pass_build_ssa): Replace declaration with that of...
5072 (make_pass_build_ssa): ...new function.
5073 (pass_build_alias): Replace declaration with that of...
5074 (make_pass_build_alias): ...new function.
5075 (pass_build_ealias): Replace declaration with that of...
5076 (make_pass_build_ealias): ...new function.
5077 (pass_dominator): Replace declaration with that of...
5078 (make_pass_dominator): ...new function.
5079 (pass_dce): Replace declaration with that of...
5080 (make_pass_dce): ...new function.
5081 (pass_dce_loop): Replace declaration with that of...
5082 (make_pass_dce_loop): ...new function.
5083 (pass_cd_dce): Replace declaration with that of...
5084 (make_pass_cd_dce): ...new function.
5085 (pass_call_cdce): Replace declaration with that of...
5086 (make_pass_call_cdce): ...new function.
5087 (pass_merge_phi): Replace declaration with that of...
5088 (make_pass_merge_phi): ...new function.
5089 (pass_split_crit_edges): Replace declaration with that of...
5090 (make_pass_split_crit_edges): ...new function.
5091 (pass_pre): Replace declaration with that of...
5092 (make_pass_pre): ...new function.
5093 (pass_profile): Replace declaration with that of...
5094 (make_pass_profile): ...new function.
5095 (pass_strip_predict_hints): Replace declaration with that of...
5096 (make_pass_strip_predict_hints): ...new function.
5097 (pass_lower_complex_O0): Replace declaration with that of...
5098 (make_pass_lower_complex_O0): ...new function.
5099 (pass_lower_complex): Replace declaration with that of...
5100 (make_pass_lower_complex): ...new function.
5101 (pass_lower_vector): Replace declaration with that of...
5102 (make_pass_lower_vector): ...new function.
5103 (pass_lower_vector_ssa): Replace declaration with that of...
5104 (make_pass_lower_vector_ssa): ...new function.
5105 (pass_lower_omp): Replace declaration with that of...
5106 (make_pass_lower_omp): ...new function.
5107 (pass_diagnose_omp_blocks): Replace declaration with that of...
5108 (make_pass_diagnose_omp_blocks): ...new function.
5109 (pass_expand_omp): Replace declaration with that of...
5110 (make_pass_expand_omp): ...new function.
5111 (pass_expand_omp_ssa): Replace declaration with that of...
5112 (make_pass_expand_omp_ssa): ...new function.
5113 (pass_object_sizes): Replace declaration with that of...
5114 (make_pass_object_sizes): ...new function.
5115 (pass_strlen): Replace declaration with that of...
5116 (make_pass_strlen): ...new function.
5117 (pass_fold_builtins): Replace declaration with that of...
5118 (make_pass_fold_builtins): ...new function.
5119 (pass_stdarg): Replace declaration with that of...
5120 (make_pass_stdarg): ...new function.
5121 (pass_early_warn_uninitialized): Replace declaration with that of...
5122 (make_pass_early_warn_uninitialized): ...new function.
5123 (pass_late_warn_uninitialized): Replace declaration with that of...
5124 (make_pass_late_warn_uninitialized): ...new function.
5125 (pass_cse_reciprocals): Replace declaration with that of...
5126 (make_pass_cse_reciprocals): ...new function.
5127 (pass_cse_sincos): Replace declaration with that of...
5128 (make_pass_cse_sincos): ...new function.
5129 (pass_optimize_bswap): Replace declaration with that of...
5130 (make_pass_optimize_bswap): ...new function.
5131 (pass_optimize_widening_mul): Replace declaration with that of...
5132 (make_pass_optimize_widening_mul): ...new function.
5133 (pass_warn_function_return): Replace declaration with that of...
5134 (make_pass_warn_function_return): ...new function.
5135 (pass_warn_function_noreturn): Replace declaration with that of...
5136 (make_pass_warn_function_noreturn): ...new function.
5137 (pass_cselim): Replace declaration with that of...
5138 (make_pass_cselim): ...new function.
5139 (pass_phiopt): Replace declaration with that of...
5140 (make_pass_phiopt): ...new function.
5141 (pass_forwprop): Replace declaration with that of...
5142 (make_pass_forwprop): ...new function.
5143 (pass_phiprop): Replace declaration with that of...
5144 (make_pass_phiprop): ...new function.
5145 (pass_tree_ifcombine): Replace declaration with that of...
5146 (make_pass_tree_ifcombine): ...new function.
5147 (pass_dse): Replace declaration with that of...
5148 (make_pass_dse): ...new function.
5149 (pass_nrv): Replace declaration with that of...
5150 (make_pass_nrv): ...new function.
5151 (pass_rename_ssa_copies): Replace declaration with that of...
5152 (make_pass_rename_ssa_copies): ...new function.
5153 (pass_sink_code): Replace declaration with that of...
5154 (make_pass_sink_code): ...new function.
5155 (pass_fre): Replace declaration with that of...
5156 (make_pass_fre): ...new function.
5157 (pass_check_data_deps): Replace declaration with that of...
5158 (make_pass_check_data_deps): ...new function.
5159 (pass_copy_prop): Replace declaration with that of...
5160 (make_pass_copy_prop): ...new function.
5161 (pass_vrp): Replace declaration with that of...
5162 (make_pass_vrp): ...new function.
5163 (pass_uncprop): Replace declaration with that of...
5164 (make_pass_uncprop): ...new function.
5165 (pass_return_slot): Replace declaration with that of...
5166 (make_pass_return_slot): ...new function.
5167 (pass_reassoc): Replace declaration with that of...
5168 (make_pass_reassoc): ...new function.
5169 (pass_rebuild_cgraph_edges): Replace declaration with that of...
5170 (make_pass_rebuild_cgraph_edges): ...new function.
5171 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
5172 (make_pass_remove_cgraph_callee_edges): ...new function.
5173 (pass_build_cgraph_edges): Replace declaration with that of...
5174 (make_pass_build_cgraph_edges): ...new function.
5175 (pass_local_pure_const): Replace declaration with that of...
5176 (make_pass_local_pure_const): ...new function.
5177 (pass_tracer): Replace declaration with that of...
5178 (make_pass_tracer): ...new function.
5179 (pass_warn_unused_result): Replace declaration with that of...
5180 (make_pass_warn_unused_result): ...new function.
5181 (pass_diagnose_tm_blocks): Replace declaration with that of...
5182 (make_pass_diagnose_tm_blocks): ...new function.
5183 (pass_lower_tm): Replace declaration with that of...
5184 (make_pass_lower_tm): ...new function.
5185 (pass_tm_init): Replace declaration with that of...
5186 (make_pass_tm_init): ...new function.
5187 (pass_tm_mark): Replace declaration with that of...
5188 (make_pass_tm_mark): ...new function.
5189 (pass_tm_memopt): Replace declaration with that of...
5190 (make_pass_tm_memopt): ...new function.
5191 (pass_tm_edges): Replace declaration with that of...
5192 (make_pass_tm_edges): ...new function.
5193 (pass_split_functions): Replace declaration with that of...
5194 (make_pass_split_functions): ...new function.
5195 (pass_feedback_split_functions): Replace declaration with that of...
5196 (make_pass_feedback_split_functions): ...new function.
5197 (pass_strength_reduction): Replace declaration with that of...
5198 (make_pass_strength_reduction): ...new function.
5199 (pass_ipa_lower_emutls): Replace declaration with that of...
5200 (make_pass_ipa_lower_emutls): ...new function.
5201 (pass_ipa_function_and_variable_visibility): Replace declaration with
5202 that of...
5203 (make_pass_ipa_function_and_variable_visibility): ...new function.
5204 (pass_ipa_tree_profile): Replace declaration with that of...
5205 (make_pass_ipa_tree_profile): ...new function.
5206 (pass_early_local_passes): Replace declaration with that of...
5207 (make_pass_early_local_passes): ...new function.
5208 (pass_ipa_whole_program_visibility): Replace declaration with that
5209 of...
5210 (make_pass_ipa_whole_program_visibility): ...new function.
5211 (pass_ipa_lto_gimple_out): Replace declaration with that of...
5212 (make_pass_ipa_lto_gimple_out): ...new function.
5213 (pass_ipa_increase_alignment): Replace declaration with that of...
5214 (make_pass_ipa_increase_alignment): ...new function.
5215 (pass_ipa_inline): Replace declaration with that of...
5216 (make_pass_ipa_inline): ...new function.
5217 (pass_ipa_free_lang_data): Replace declaration with that of...
5218 (make_pass_ipa_free_lang_data): ...new function.
5219 (pass_ipa_free_inline_summary): Replace declaration with that of...
5220 (make_pass_ipa_free_inline_summary): ...new function.
5221 (pass_ipa_cp): Replace declaration with that of...
5222 (make_pass_ipa_cp): ...new function.
5223 (pass_ipa_reference): Replace declaration with that of...
5224 (make_pass_ipa_reference): ...new function.
5225 (pass_ipa_pure_const): Replace declaration with that of...
5226 (make_pass_ipa_pure_const): ...new function.
5227 (pass_ipa_pta): Replace declaration with that of...
5228 (make_pass_ipa_pta): ...new function.
5229 (pass_ipa_lto_finish_out): Replace declaration with that of...
5230 (make_pass_ipa_lto_finish_out): ...new function.
5231 (pass_ipa_tm): Replace declaration with that of...
5232 (make_pass_ipa_tm): ...new function.
5233 (pass_ipa_profile): Replace declaration with that of...
5234 (make_pass_ipa_profile): ...new function.
5235 (pass_ipa_cdtor_merge): Replace declaration with that of...
5236 (make_pass_ipa_cdtor_merge): ...new function.
5237 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
5238 of...
5239 (make_pass_cleanup_cfg_post_optimizing): ...new function.
5240 (pass_init_datastructures): Replace declaration with that of...
5241 (make_pass_init_datastructures): ...new function.
5242 (pass_fixup_cfg): Replace declaration with that of...
5243 (make_pass_fixup_cfg): ...new function.
5244 (pass_expand): Replace declaration with that of...
5245 (make_pass_expand): ...new function.
5246 (pass_instantiate_virtual_regs): Replace declaration with that of...
5247 (make_pass_instantiate_virtual_regs): ...new function.
5248 (pass_rtl_fwprop): Replace declaration with that of...
5249 (make_pass_rtl_fwprop): ...new function.
5250 (pass_rtl_fwprop_addr): Replace declaration with that of...
5251 (make_pass_rtl_fwprop_addr): ...new function.
5252 (pass_jump): Replace declaration with that of...
5253 (make_pass_jump): ...new function.
5254 (pass_jump2): Replace declaration with that of...
5255 (make_pass_jump2): ...new function.
5256 (pass_lower_subreg): Replace declaration with that of...
5257 (make_pass_lower_subreg): ...new function.
5258 (pass_cse): Replace declaration with that of...
5259 (make_pass_cse): ...new function.
5260 (pass_fast_rtl_dce): Replace declaration with that of...
5261 (make_pass_fast_rtl_dce): ...new function.
5262 (pass_ud_rtl_dce): Replace declaration with that of...
5263 (make_pass_ud_rtl_dce): ...new function.
5264 (pass_rtl_dce): Replace declaration with that of...
5265 (make_pass_rtl_dce): ...new function.
5266 (pass_rtl_dse1): Replace declaration with that of...
5267 (make_pass_rtl_dse1): ...new function.
5268 (pass_rtl_dse2): Replace declaration with that of...
5269 (make_pass_rtl_dse2): ...new function.
5270 (pass_rtl_dse3): Replace declaration with that of...
5271 (make_pass_rtl_dse3): ...new function.
5272 (pass_rtl_cprop): Replace declaration with that of...
5273 (make_pass_rtl_cprop): ...new function.
5274 (pass_rtl_pre): Replace declaration with that of...
5275 (make_pass_rtl_pre): ...new function.
5276 (pass_rtl_hoist): Replace declaration with that of...
5277 (make_pass_rtl_hoist): ...new function.
5278 (pass_rtl_store_motion): Replace declaration with that of...
5279 (make_pass_rtl_store_motion): ...new function.
5280 (pass_cse_after_global_opts): Replace declaration with that of...
5281 (make_pass_cse_after_global_opts): ...new function.
5282 (pass_rtl_ifcvt): Replace declaration with that of...
5283 (make_pass_rtl_ifcvt): ...new function.
5284 (pass_into_cfg_layout_mode): Replace declaration with that of...
5285 (make_pass_into_cfg_layout_mode): ...new function.
5286 (pass_outof_cfg_layout_mode): Replace declaration with that of...
5287 (make_pass_outof_cfg_layout_mode): ...new function.
5288 (pass_loop2): Replace declaration with that of...
5289 (make_pass_loop2): ...new function.
5290 (pass_rtl_loop_init): Replace declaration with that of...
5291 (make_pass_rtl_loop_init): ...new function.
5292 (pass_rtl_move_loop_invariants): Replace declaration with that of...
5293 (make_pass_rtl_move_loop_invariants): ...new function.
5294 (pass_rtl_unswitch): Replace declaration with that of...
5295 (make_pass_rtl_unswitch): ...new function.
5296 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
5297 (make_pass_rtl_unroll_and_peel_loops): ...new function.
5298 (pass_rtl_doloop): Replace declaration with that of...
5299 (make_pass_rtl_doloop): ...new function.
5300 (pass_rtl_loop_done): Replace declaration with that of...
5301 (make_pass_rtl_loop_done): ...new function.
5302 (pass_web): Replace declaration with that of...
5303 (make_pass_web): ...new function.
5304 (pass_cse2): Replace declaration with that of...
5305 (make_pass_cse2): ...new function.
5306 (pass_df_initialize_opt): Replace declaration with that of...
5307 (make_pass_df_initialize_opt): ...new function.
5308 (pass_df_initialize_no_opt): Replace declaration with that of...
5309 (make_pass_df_initialize_no_opt): ...new function.
5310 (pass_reginfo_init): Replace declaration with that of...
5311 (make_pass_reginfo_init): ...new function.
5312 (pass_inc_dec): Replace declaration with that of...
5313 (make_pass_inc_dec): ...new function.
5314 (pass_stack_ptr_mod): Replace declaration with that of...
5315 (make_pass_stack_ptr_mod): ...new function.
5316 (pass_initialize_regs): Replace declaration with that of...
5317 (make_pass_initialize_regs): ...new function.
5318 (pass_combine): Replace declaration with that of...
5319 (make_pass_combine): ...new function.
5320 (pass_if_after_combine): Replace declaration with that of...
5321 (make_pass_if_after_combine): ...new function.
5322 (pass_ree): Replace declaration with that of...
5323 (make_pass_ree): ...new function.
5324 (pass_partition_blocks): Replace declaration with that of...
5325 (make_pass_partition_blocks): ...new function.
5326 (pass_match_asm_constraints): Replace declaration with that of...
5327 (make_pass_match_asm_constraints): ...new function.
5328 (pass_regmove): Replace declaration with that of...
5329 (make_pass_regmove): ...new function.
5330 (pass_split_all_insns): Replace declaration with that of...
5331 (make_pass_split_all_insns): ...new function.
5332 (pass_fast_rtl_byte_dce): Replace declaration with that of...
5333 (make_pass_fast_rtl_byte_dce): ...new function.
5334 (pass_lower_subreg2): Replace declaration with that of...
5335 (make_pass_lower_subreg2): ...new function.
5336 (pass_mode_switching): Replace declaration with that of...
5337 (make_pass_mode_switching): ...new function.
5338 (pass_sms): Replace declaration with that of...
5339 (make_pass_sms): ...new function.
5340 (pass_sched): Replace declaration with that of...
5341 (make_pass_sched): ...new function.
5342 (pass_ira): Replace declaration with that of...
5343 (make_pass_ira): ...new function.
5344 (pass_reload): Replace declaration with that of...
5345 (make_pass_reload): ...new function.
5346 (pass_clean_state): Replace declaration with that of...
5347 (make_pass_clean_state): ...new function.
5348 (pass_branch_prob): Replace declaration with that of...
5349 (make_pass_branch_prob): ...new function.
5350 (pass_value_profile_transformations): Replace declaration with that
5351 of...
5352 (make_pass_value_profile_transformations): ...new function.
5353 (pass_postreload_cse): Replace declaration with that of...
5354 (make_pass_postreload_cse): ...new function.
5355 (pass_gcse2): Replace declaration with that of...
5356 (make_pass_gcse2): ...new function.
5357 (pass_split_after_reload): Replace declaration with that of...
5358 (make_pass_split_after_reload): ...new function.
5359 (pass_branch_target_load_optimize1): Replace declaration with that
5360 of...
5361 (make_pass_branch_target_load_optimize1): ...new function.
5362 (pass_thread_prologue_and_epilogue): Replace declaration with that
5363 of...
5364 (make_pass_thread_prologue_and_epilogue): ...new function.
5365 (pass_stack_adjustments): Replace declaration with that of...
5366 (make_pass_stack_adjustments): ...new function.
5367 (pass_peephole2): Replace declaration with that of...
5368 (make_pass_peephole2): ...new function.
5369 (pass_if_after_reload): Replace declaration with that of...
5370 (make_pass_if_after_reload): ...new function.
5371 (pass_regrename): Replace declaration with that of...
5372 (make_pass_regrename): ...new function.
5373 (pass_cprop_hardreg): Replace declaration with that of...
5374 (make_pass_cprop_hardreg): ...new function.
5375 (pass_reorder_blocks): Replace declaration with that of...
5376 (make_pass_reorder_blocks): ...new function.
5377 (pass_branch_target_load_optimize2): Replace declaration with that
5378 of...
5379 (make_pass_branch_target_load_optimize2): ...new function.
5380 (pass_leaf_regs): Replace declaration with that of...
5381 (make_pass_leaf_regs): ...new function.
5382 (pass_split_before_sched2): Replace declaration with that of...
5383 (make_pass_split_before_sched2): ...new function.
5384 (pass_compare_elim_after_reload): Replace declaration with that of...
5385 (make_pass_compare_elim_after_reload): ...new function.
5386 (pass_sched2): Replace declaration with that of...
5387 (make_pass_sched2): ...new function.
5388 (pass_stack_regs): Replace declaration with that of...
5389 (make_pass_stack_regs): ...new function.
5390 (pass_stack_regs_run): Replace declaration with that of...
5391 (make_pass_stack_regs_run): ...new function.
5392 (pass_df_finish): Replace declaration with that of...
5393 (make_pass_df_finish): ...new function.
5394 (pass_compute_alignments): Replace declaration with that of...
5395 (make_pass_compute_alignments): ...new function.
5396 (pass_duplicate_computed_gotos): Replace declaration with that of...
5397 (make_pass_duplicate_computed_gotos): ...new function.
5398 (pass_variable_tracking): Replace declaration with that of...
5399 (make_pass_variable_tracking): ...new function.
5400 (pass_free_cfg): Replace declaration with that of...
5401 (make_pass_free_cfg): ...new function.
5402 (pass_machine_reorg): Replace declaration with that of...
5403 (make_pass_machine_reorg): ...new function.
5404 (pass_cleanup_barriers): Replace declaration with that of...
5405 (make_pass_cleanup_barriers): ...new function.
5406 (pass_delay_slots): Replace declaration with that of...
5407 (make_pass_delay_slots): ...new function.
5408 (pass_split_for_shorten_branches): Replace declaration with that of...
5409 (make_pass_split_for_shorten_branches): ...new function.
5410 (pass_split_before_regstack): Replace declaration with that of...
5411 (make_pass_split_before_regstack): ...new function.
5412 (pass_convert_to_eh_region_ranges): Replace declaration with that
5413 of...
5414 (make_pass_convert_to_eh_region_ranges): ...new function.
5415 (pass_shorten_branches): Replace declaration with that of...
5416 (make_pass_shorten_branches): ...new function.
5417 (pass_set_nothrow_function_flags): Replace declaration with that of...
5418 (make_pass_set_nothrow_function_flags): ...new function.
5419 (pass_dwarf2_frame): Replace declaration with that of...
5420 (make_pass_dwarf2_frame): ...new function.
5421 (pass_final): Replace declaration with that of...
5422 (make_pass_final): ...new function.
5423 (pass_rtl_seqabstr): Replace declaration with that of...
5424 (make_pass_rtl_seqabstr): ...new function.
5425 (pass_release_ssa_names): Replace declaration with that of...
5426 (make_pass_release_ssa_names): ...new function.
5427 (pass_early_inline): Replace declaration with that of...
5428 (make_pass_early_inline): ...new function.
5429 (pass_inline_parameters): Replace declaration with that of...
5430 (make_pass_inline_parameters): ...new function.
5431 (pass_update_address_taken): Replace declaration with that of...
5432 (make_pass_update_address_taken): ...new function.
5433 (pass_convert_switch): Replace declaration with that of...
5434 (make_pass_convert_switch): ...new function.
5435 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
5436 to a subclass of simple_ipa_opt_pass along with...
5437 (pass_data_ipa_tree_profile): ...new pass_data instance and...
5438 (make_pass_ipa_tree_profile): ...new function.
5439 * tree-sra.c (pass_sra_early): Convert from a global struct to a
5440 subclass of gimple_opt_pass along with...
5441 (pass_data_sra_early): ...new pass_data instance and...
5442 (make_pass_sra_early): ...new function.
5443 (pass_sra): Convert from a global struct to a subclass of
5444 gimple_opt_pass along with...
5445 (pass_data_sra): ...new pass_data instance and...
5446 (make_pass_sra): ...new function.
5447 (pass_early_ipa_sra): Convert from a global struct to a subclass of
5448 gimple_opt_pass along with...
5449 (pass_data_early_ipa_sra): ...new pass_data instance and...
5450 (make_pass_early_ipa_sra): ...new function.
5451 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
5452 subclass of gimple_opt_pass along with...
5453 (pass_data_ccp): ...new pass_data instance and...
5454 (make_pass_ccp): ...new function.
5455 (pass_fold_builtins): Convert from a global struct to a subclass of
5456 gimple_opt_pass along with...
5457 (pass_data_fold_builtins): ...new pass_data instance and...
5458 (make_pass_fold_builtins): ...new function.
5459 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
5460 subclass of gimple_opt_pass along with...
5461 (pass_data_copy_prop): ...new pass_data instance and...
5462 (make_pass_copy_prop): ...new function.
5463 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
5464 global struct to a subclass of gimple_opt_pass along with...
5465 (pass_data_rename_ssa_copies): ...new pass_data instance and...
5466 (make_pass_rename_ssa_copies): ...new function.
5467 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
5468 subclass of gimple_opt_pass along with...
5469 (pass_data_dce): ...new pass_data instance and...
5470 (make_pass_dce): ...new function.
5471 (pass_dce_loop): Convert from a global struct to a subclass of
5472 gimple_opt_pass along with...
5473 (pass_data_dce_loop): ...new pass_data instance and...
5474 (make_pass_dce_loop): ...new function.
5475 (pass_cd_dce): Convert from a global struct to a subclass of
5476 gimple_opt_pass along with...
5477 (pass_data_cd_dce): ...new pass_data instance and...
5478 (make_pass_cd_dce): ...new function.
5479 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
5480 subclass of gimple_opt_pass along with...
5481 (pass_data_dominator): ...new pass_data instance and...
5482 (make_pass_dominator): ...new function.
5483 (pass_phi_only_cprop): Convert from a global struct to a subclass of
5484 gimple_opt_pass along with...
5485 (pass_data_phi_only_cprop): ...new pass_data instance and...
5486 (make_pass_phi_only_cprop): ...new function.
5487 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
5488 subclass of gimple_opt_pass along with...
5489 (pass_data_dse): ...new pass_data instance and...
5490 (make_pass_dse): ...new function.
5491 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
5492 a subclass of gimple_opt_pass along with...
5493 (pass_data_forwprop): ...new pass_data instance and...
5494 (make_pass_forwprop): ...new function.
5495 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
5496 struct to a subclass of gimple_opt_pass along with...
5497 (pass_data_tree_ifcombine): ...new pass_data instance and...
5498 (make_pass_tree_ifcombine): ...new function.
5499 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
5500 subclass of gimple_opt_pass along with...
5501 (pass_data_ch): ...new pass_data instance and...
5502 (make_pass_ch): ...new function.
5503 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
5504 subclass of gimple_opt_pass along with...
5505 (pass_data_tree_loop): ...new pass_data instance and...
5506 (make_pass_tree_loop): ...new function.
5507 (pass_tree_loop_init): Convert from a global struct to a subclass of
5508 gimple_opt_pass along with...
5509 (pass_data_tree_loop_init): ...new pass_data instance and...
5510 (make_pass_tree_loop_init): ...new function.
5511 (pass_lim): Convert from a global struct to a subclass of
5512 gimple_opt_pass along with...
5513 (pass_data_lim): ...new pass_data instance and...
5514 (make_pass_lim): ...new function.
5515 (pass_tree_unswitch): Convert from a global struct to a subclass of
5516 gimple_opt_pass along with...
5517 (pass_data_tree_unswitch): ...new pass_data instance and...
5518 (make_pass_tree_unswitch): ...new function.
5519 (pass_predcom): Convert from a global struct to a subclass of
5520 gimple_opt_pass along with...
5521 (pass_data_predcom): ...new pass_data instance and...
5522 (make_pass_predcom): ...new function.
5523 (pass_vectorize): Convert from a global struct to a subclass of
5524 gimple_opt_pass along with...
5525 (pass_data_vectorize): ...new pass_data instance and...
5526 (make_pass_vectorize): ...new function.
5527 (pass_graphite): Convert from a global struct to a subclass of
5528 gimple_opt_pass along with...
5529 (pass_data_graphite): ...new pass_data instance and...
5530 (make_pass_graphite): ...new function.
5531 (pass_graphite_transforms): Convert from a global struct to a subclass
5532 of gimple_opt_pass along with...
5533 (pass_data_graphite_transforms): ...new pass_data instance and...
5534 (make_pass_graphite_transforms): ...new function.
5535 (pass_check_data_deps): Convert from a global struct to a subclass of
5536 gimple_opt_pass along with...
5537 (pass_data_check_data_deps): ...new pass_data instance and...
5538 (make_pass_check_data_deps): ...new function.
5539 (pass_iv_canon): Convert from a global struct to a subclass of
5540 gimple_opt_pass along with...
5541 (pass_data_iv_canon): ...new pass_data instance and...
5542 (make_pass_iv_canon): ...new function.
5543 (pass_scev_cprop): Convert from a global struct to a subclass of
5544 gimple_opt_pass along with...
5545 (pass_data_scev_cprop): ...new pass_data instance and...
5546 (make_pass_scev_cprop): ...new function.
5547 (pass_record_bounds): Convert from a global struct to a subclass of
5548 gimple_opt_pass along with...
5549 (pass_data_record_bounds): ...new pass_data instance and...
5550 (make_pass_record_bounds): ...new function.
5551 (pass_complete_unroll): Convert from a global struct to a subclass of
5552 gimple_opt_pass along with...
5553 (pass_data_complete_unroll): ...new pass_data instance and...
5554 (make_pass_complete_unroll): ...new function.
5555 (pass_complete_unrolli): Convert from a global struct to a subclass of
5556 gimple_opt_pass along with...
5557 (pass_data_complete_unrolli): ...new pass_data instance and...
5558 (make_pass_complete_unrolli): ...new function.
5559 (pass_parallelize_loops): Convert from a global struct to a subclass
5560 of gimple_opt_pass along with...
5561 (pass_data_parallelize_loops): ...new pass_data instance and...
5562 (make_pass_parallelize_loops): ...new function.
5563 (pass_loop_prefetch): Convert from a global struct to a subclass of
5564 gimple_opt_pass along with...
5565 (pass_data_loop_prefetch): ...new pass_data instance and...
5566 (make_pass_loop_prefetch): ...new function.
5567 (pass_iv_optimize): Convert from a global struct to a subclass of
5568 gimple_opt_pass along with...
5569 (pass_data_iv_optimize): ...new pass_data instance and...
5570 (make_pass_iv_optimize): ...new function.
5571 (pass_tree_loop_done): Convert from a global struct to a subclass of
5572 gimple_opt_pass along with...
5573 (pass_data_tree_loop_done): ...new pass_data instance and...
5574 (make_pass_tree_loop_done): ...new function.
5575 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
5576 struct to a subclass of gimple_opt_pass along with...
5577 (pass_data_cse_reciprocals): ...new pass_data instance and...
5578 (make_pass_cse_reciprocals): ...new function.
5579 (pass_cse_sincos): Convert from a global struct to a subclass of
5580 gimple_opt_pass along with...
5581 (pass_data_cse_sincos): ...new pass_data instance and...
5582 (make_pass_cse_sincos): ...new function.
5583 (pass_optimize_bswap): Convert from a global struct to a subclass of
5584 gimple_opt_pass along with...
5585 (pass_data_optimize_bswap): ...new pass_data instance and...
5586 (make_pass_optimize_bswap): ...new function.
5587 (pass_optimize_widening_mul): Convert from a global struct to a
5588 subclass of gimple_opt_pass along with...
5589 (pass_data_optimize_widening_mul): ...new pass_data instance and...
5590 (make_pass_optimize_widening_mul): ...new function.
5591 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
5592 subclass of gimple_opt_pass along with...
5593 (pass_data_phiopt): ...new pass_data instance and...
5594 (make_pass_phiopt): ...new function.
5595 (pass_cselim): Convert from a global struct to a subclass of
5596 gimple_opt_pass along with...
5597 (pass_data_cselim): ...new pass_data instance and...
5598 (make_pass_cselim): ...new function.
5599 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
5600 subclass of gimple_opt_pass along with...
5601 (pass_data_phiprop): ...new pass_data instance and...
5602 (make_pass_phiprop): ...new function.
5603 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
5604 subclass of gimple_opt_pass along with...
5605 (pass_data_pre): ...new pass_data instance and...
5606 (make_pass_pre): ...new function.
5607 (pass_fre): Convert from a global struct to a subclass of
5608 gimple_opt_pass along with...
5609 (pass_data_fre): ...new pass_data instance and...
5610 (make_pass_fre): ...new function.
5611 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
5612 subclass of gimple_opt_pass along with...
5613 (pass_data_reassoc): ...new pass_data instance and...
5614 (make_pass_reassoc): ...new function.
5615 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
5616 subclass of gimple_opt_pass along with...
5617 (pass_data_sink_code): ...new pass_data instance and...
5618 (make_pass_sink_code): ...new function.
5619 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
5620 subclass of gimple_opt_pass along with...
5621 (pass_data_strlen): ...new pass_data instance and...
5622 (make_pass_strlen): ...new function.
5623 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
5624 struct to a subclass of gimple_opt_pass along with...
5625 (pass_data_build_alias): ...new pass_data instance and...
5626 (make_pass_build_alias): ...new function.
5627 (pass_build_ealias): Convert from a global struct to a subclass of
5628 gimple_opt_pass along with...
5629 (pass_data_build_ealias): ...new pass_data instance and...
5630 (make_pass_build_ealias): ...new function.
5631 (pass_ipa_pta): Convert from a global struct to a subclass of
5632 simple_ipa_opt_pass along with...
5633 (pass_data_ipa_pta): ...new pass_data instance and...
5634 (make_pass_ipa_pta): ...new function.
5635 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
5636 subclass of gimple_opt_pass along with...
5637 (pass_data_uncprop): ...new pass_data instance and...
5638 (make_pass_uncprop): ...new function.
5639 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
5640 global struct to a subclass of gimple_opt_pass along with...
5641 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
5642 (make_pass_late_warn_uninitialized): ...new function.
5643 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
5644 to a subclass of gimple_opt_pass along with...
5645 (pass_data_init_datastructures): ...new pass_data instance and...
5646 (make_pass_init_datastructures): ...new function.
5647 (pass_early_warn_uninitialized): Convert from a global struct to a
5648 subclass of gimple_opt_pass along with...
5649 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
5650 (make_pass_early_warn_uninitialized): ...new function.
5651 (pass_update_address_taken): Convert from a global struct to a
5652 subclass of gimple_opt_pass along with...
5653 (pass_data_update_address_taken): ...new pass_data instance and...
5654 (make_pass_update_address_taken): ...new function.
5655 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
5656 struct to a subclass of gimple_opt_pass along with...
5657 (pass_data_release_ssa_names): ...new pass_data instance and...
5658 (make_pass_release_ssa_names): ...new function.
5659 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
5660 subclass of gimple_opt_pass along with...
5661 (pass_data_stdarg): ...new pass_data instance and...
5662 (make_pass_stdarg): ...new function.
5663 * tree-switch-conversion.c (pass_convert_switch): Convert from a
5664 global struct to a subclass of gimple_opt_pass along with...
5665 (pass_data_convert_switch): ...new pass_data instance and...
5666 (make_pass_convert_switch): ...new function.
5667 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
5668 to a subclass of gimple_opt_pass along with...
5669 (pass_data_tail_recursion): ...new pass_data instance and...
5670 (make_pass_tail_recursion): ...new function.
5671 (pass_tail_calls): Convert from a global struct to a subclass of
5672 gimple_opt_pass along with...
5673 (pass_data_tail_calls): ...new pass_data instance and...
5674 (make_pass_tail_calls): ...new function.
5675 * tree-vect-generic.c (pass_lower_vector): Convert from a global
5676 struct to a subclass of gimple_opt_pass along with...
5677 (pass_data_lower_vector): ...new pass_data instance and...
5678 (make_pass_lower_vector): ...new function.
5679 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
5680 gimple_opt_pass along with...
5681 (pass_data_lower_vector_ssa): ...new pass_data instance and...
5682 (make_pass_lower_vector_ssa): ...new function.
5683 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
5684 to a subclass of gimple_opt_pass along with...
5685 (pass_data_slp_vectorize): ...new pass_data instance and...
5686 (make_pass_slp_vectorize): ...new function.
5687 (pass_ipa_increase_alignment): Convert from a global struct to a
5688 subclass of simple_ipa_opt_pass along with...
5689 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
5690 (make_pass_ipa_increase_alignment): ...new function.
5691 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
5692 gimple_opt_pass along with...
5693 (pass_data_vrp): ...new pass_data instance and...
5694 (make_pass_vrp): ...new function.
5695 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
5696 subclass of simple_ipa_opt_pass along with...
5697 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
5698 (make_pass_ipa_free_lang_data): ...new function.
5699 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
5700 gimple_opt_pass along with...
5701 (pass_data_tsan): ...new pass_data instance and...
5702 (make_pass_tsan): ...new function.
5703 (pass_tsan_O0): Convert from a global struct to a subclass of
5704 gimple_opt_pass along with...
5705 (pass_data_tsan_O0): ...new pass_data instance and...
5706 (make_pass_tsan_O0): ...new function.
5707 * var-tracking.c (pass_variable_tracking): Convert from a global
5708 struct to a subclass of rtl_opt_pass along with...
5709 (pass_data_variable_tracking): ...new pass_data instance and...
5710 (make_pass_variable_tracking): ...new function.
5711 * web.c (pass_web): Convert from a global struct to a subclass of
5712 rtl_opt_pass along with...
5713 (pass_data_web): ...new pass_data instance and...
5714 (make_pass_web): ...new function.
5715 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
5716 declaration with that of...
5717 (make_pass_mode_switch_use): ...new function.
5718 (pass_resolve_sw_modes): Replace declaration with that of...
5719 (make_pass_resolve_sw_modes): ...new function.
5720 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
5721 from a global struct to a subclass of rtl_opt_pass along with...
5722 (pass_data_mode_switch_use): ...new pass_data instance and...
5723 (make_pass_mode_switch_use): ...new function.
5724 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
5725 from a global struct to a subclass of rtl_opt_pass along with...
5726 (pass_data_resolve_sw_modes): ...new pass_data instance and...
5727 (make_pass_resolve_sw_modes): ...new function.
5728 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
5729 struct to a subclass of rtl_opt_pass along with...
5730 (pass_data_insert_vzeroupper): ...new pass_data instance and...
5731 (make_pass_insert_vzeroupper): ...new function.
5732 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
5733 global struct to a subclass of rtl_opt_pass along with...
5734 (pass_data_work_around_errata): ...new pass_data instance and...
5735 (make_pass_work_around_errata): ...new function.
5736 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
5737 struct to a subclass of rtl_opt_pass along with...
5738 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
5739 (make_pass_mips_machine_reorg2): ...new function.
5740
6a389ed5
DM
57412013-08-05 David Malcolm <dmalcolm@redhat.com>
5742
5743 * passes.c (pass_manager::operator new): New.
5744
f7695dbf
DM
57452013-08-05 David Malcolm <dmalcolm@redhat.com>
5746
5747 Handwritten part of conversion of passes to C++ classes.
5748
5749 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
5750 (toplev.o): Add dep on PASS_MANAGER_H.
5751 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
5752 of early local pases to reflect this moving from a global to a
5753 member of gcc::pass_manager.
5754 (cgraph_add_new_function): Likewise.
5755 * lto-cgraph.c (lto_output_node): Update for conversion of
5756 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
5757 * passes.c (opt_pass::clone): New.
5758 (opt_pass::gate): New.
5759 (opt_pass::execute): New.
5760 (opt_pass::opt_pass): New.
5761 (pass_manager::execute_early_local_passes): New.
5762 (pass_manager::execute_pass_mode_switching): new.
5763 (finish_optimization_passes): Convert to...
5764 (pass_manager::finish_optimization_passes): ...this.
5765 (finish_optimization_passes): Update for conversion of passes to
5766 C++ classes.
5767 (register_dump_files_1): Use has_gate since we cannot portably
5768 check a vtable entry against NULL.
5769 (dump_one_pass): Likewise.
5770 (ipa_write_summaries_2): Likewise.
5771 (ipa_write_optimization_summaries_1): Likewise.
5772 (ipa_read_summaries_1): Likewise.
5773 (ipa_read_optimization_summaries_1): Likewise.
5774 (execute_ipa_stmt_fixups): Likewise.
5775 (pass_manager::pass_manager): Rewrite pass-creation, invoking
5776 pass-creation functions rather than wiring up globals, and
5777 storing the results in fields of pass_manager generated using
5778 pass-instances.def.
5779 (pass_manager::dump_profile_report): Update for conversion of
5780 passes to C++ classes.
5781 (pass_manager::execute_ipa_summary_passes): Likewise.
5782 (execute_one_ipa_transform_pass): Likewise.
5783 (execute_one_pass): Use has_gate and has_execute since we cannot
5784 portably check a vtable entry against NULL.
5785 * pass_manager.h (pass_manager::finish_optimization_passes): New.
5786 (pass_manager): Use pass-instances.def to add fields for the
5787 various pass instances.
5788 * toplev.c (finalize): Update for move of
5789 finish_optimization_passes to a method of gcc::pass_manager.
5790 * toplev.h (finish_optimization_passes): Move to method of class
5791 pass_manager.
5792 * tree-pass.h (struct pass_data): New.
03b0ee0a 5793 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM
5794 (opt_pass::gate): Convert to virtual function.
5795 (opt_pass::~opt_pass): New.
5796 (opt_pass::clone): New.
5797 (opt_pass::execute): Convert to virtual function.
5798 (opt_pass::opt_pass): New.
5799 (opt_pass::ctxt_): new.
5800 (gimple_opt_pass): Convert to subclass of opt_pass.
5801 (gimple_opt_pass::gimple_opt_pass): New.
5802 (rtl_opt_pass): Convert to subclass of opt_pass.
5803 (rtl_opt_pass::rtl_opt_pass): New.
5804 (ipa_opt_pass_d): Convert to subclass of opt_pass.
5805 (ipa_opt_pass_d::ipa_opt_pass_d): New.
5806 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
5807 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
5808 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
5809 invocation of pass_mode_switching to reflect this moving from a
5810 global to a member of gcc::pass_manager.
5811 (ix86_option_override): Rework how pass_insert_vzeroupper is
5812 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 5813 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 5814
d0b2f831
RE
58152013-08-05 Richard Earnshaw <rearnsha@arm.com>
5816
5817 PR rtl-optimization/57708
5818 * recog.c (peep2_find_free_register): Validate all regs in a
5819 multi-reg mode.
5820
51a5c0c2
JH
58212013-08-05 Jan Hubicka <jh@suse.cz>
5822
5823 PR lto/57602
03b0ee0a
UB
5824 * cgraph.c (verify_cgraph_node): Accept local flags from other
5825 partitions.
51a5c0c2
JH
5826 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
5827 (function_and_variable_visibility): Likewise.
5828 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
5829
65f0a120
GDR
58302013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5831
5832 * graph.c (init_graph_slim_pretty_print): Remove.
5833 (print_graph_cfg): Do not call it. Use local pretty printer.
5834 (start_graph_dump): Likewise.
5835
b3f80694
GDR
58362013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5837
5838 * gimple-pretty-print.c (buffer): Remove.
5839 (initialized): Likewise.
5840 (maybe_init_pretty_print): Likewise.
5841 (print_gimple_stmt): Do not call it. Use non-static local
5842 pretty_printer variable.
5843 (print_gimple_expr): Likewise.
5844 (print_gimple_seq): Likewise.
5845 (gimple_dump_bb): Likewise.
5846
11a877b3
GDR
58472013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
5848
5849 * asan.c (asan_pp): Remove.
5850 (asan_pp_initialized): Likewise.
5851 (asan_pp_initialize): Likewise.
5852 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
5853 (asan_emit_stack_protection): Tidy. Use local pretty printer.
5854 (asan_add_global): Likewise.
5855
b066401f
GDR
58562013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
5857
5858 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
5859 * diagnostic.h (diagnostic_flush_buffer): Adjust.
5860 * pretty-print.c (pp_formatted_text_data): Likewise.
5861 (pp_indent): Rename from pp_base_indent.
5862 (pp_format): Rename from pp_base_format.
5863 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
5864 (pp_format_verbatim): Rename from pp_base_format_verbatim.
5865 (pp_flush): Rename from pp_base_flush.
5866 (pp_set_line_maximum_length): Rename from
5867 pp_base_set_line_maximum_length.
5868 (pp_clear_output_area): Rename from pp_base_clear_output_area.
5869 (pp_set_prefix): Rename from pp_base_set_prefix.
5870 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
5871 (pp_emit_prefix): Rename from pp_base_emit_prefix.
5872 (pp_append_text): Rename from pp_base_append_text.
5873 (pp_formatted_text): Rename from pp_base_formatted_text.
5874 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
5875 (pp_remaining_character_count_for_line): Rename from
5876 pp_base_remaining_character_count_for_line.
5877 (pp_newline): Rename from pp_base_newline.
5878 (pp_character): Rename from pp_base_character.
5879 (pp_string): Rename from pp_base_string.
5880 (pp_maybe_space): Rename from pp_base_maybe_space.
5881 * asan.c (asan_pp_string): Adjust.
5882 (asan_emit_stack_protection): Likewise.
5883 (asan_add_global): Likewise.
5884 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
5885 * tree-mudflap.c (mf_varname_tree): Likewise.
5886 * tree-pretty-print.c (pp_tree_identifier): Rename from
5887 pp_base_tree_identifier.
5888 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
5889 Declare as function.
5890
137a1a27
GDR
58912013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5892
5893 * pretty-print.h (pp_bar_bar): New.
5894 (pp_ampersand_ampersand): Likewise.
5895 (pp_less_equal): Likewise.
5896 (pp_greater_equal): Likewise.
5897 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
5898 printer functions instead of pp_string or operators and punctuators.
5899 (dump_gimple_call): Likewise.
5900 (dump_gimple_omp_for): Likewise.
5901 (dump_gimple_transaction): Likewise.
5902 (dump_gimple_phi): Likewise.
5903 (pp_gimple_stmt_1): Likewise.
5904 * sched-vis.c (print_insn): Likewise.
5905 * tree-mudflap.c (mf_varname_tree): Likewise.
5906 * tree-pretty-print.c (dump_block_node): Likewise.
5907 (dump_generic_node): Likewise.
5908
815effe1
JH
59092013-08-02 Jan Hubicka <jh@suse.cz>
5910
5911 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
5912 boundaries.
5913 * lto-streamer-out.c (tree_is_indexable): Results decls and
5914 parm decls are not indexable.
5915 (DFS_write_tree_body): Do not follow args and results.
5916 (hash_tree): Likewise.
5917 (output_functions): Rearrange so struct function is needed
5918 only when real body is output; be able to also ouptut abstract
5919 functions; output DECL_ARGUMENTS and DECL_RESULT.
5920 (lto_output): When not in WPA, ale store abstract functions.
5921 (write_symbol): Do not care about RESULT_DECL.
5922 (output_symbol_p): Handle correctly sbtract decls.
5923 * lto-streamer-in.c (input_function): Rearrange so struct
5924 function can be NULL at entry; allow streaming of
5925 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
5926 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
5927 sanity check during LTO.
5928 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
5929 RESULT_DECl and DECL_ARGUMENTS.
5930 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
5931 Likewise.
5932
07838b13
GDR
59332013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
5934
5935 * pretty-print.h (pp_underscore): New.
5936 (pp_comma): Tidy.
5937 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
5938 printer functions instead of pp_character.
5939 (dump_binary_rhs): Likewise.
5940 (dump_ternary_rhs): Likewise.
5941 (dump_gimple_call_args): Likewise.
5942 (pp_points_to_solution): Likewise.
5943 (dump_gimple_call): Likewise.
5944 (dump_gimple_switch): Likewise.
5945 (dump_gimple_cond): Likewise.
5946 (dump_gimple_bind): Likewise.
5947 (dump_gimple_try): Likewise.
5948 (dump_gimple_omp_for): Likewise.
5949 (dump_gimple_omp_continue): Likewise.
5950 (dump_gimple_omp_single): Likewise.
5951 (dump_gimple_omp_sections): Likewise.
5952 (dump_gimple_omp_block): Likewise.
5953 (dump_gimple_omp_critical): Likewise.
5954 (dump_gimple_transaction): Likewise.
5955 (dump_gimple_asm): Likewise.
5956 (dump_gimple_phi): Likewise.
5957 (dump_gimple_omp_parallel): Likewise.
5958 (dump_gimple_omp_task): Likewise.
5959 (dump_gimple_omp_atomic_load): Likewise.
5960 (dump_gimple_omp_atomic_store): Likewise.
5961 (dump_gimple_mem_ops): Likewise.
5962 (pp_gimple_stmt_1): Likewise.
5963 (pp_cfg_jump): Likewise.
5964 (dump_implicit_edges): Likewise.
5965 (gimple_dump_bb_for_graph): Likewise.
5966 * graph.c (draw_cfg_node): Likewise.
5967 * langhooks.c (lhd_print_error_function): Likewise.
5968 * sched-vis.c (print_exp): Likewise.
5969 (print_value): Likewise.
5970 (print_pattern): Likewise.
5971 (print_insn): Likewise.
5972 (rtl_dump_bb_for_graph): Likewise.
5973 * tree-pretty-print.c (dump_function_declaration): Likewise.
5974 (dump_array_domain): Likewise.
5975 (dump_omp_clause): Likewise.
5976 (dump_location): Likewise.
5977 (dump_generic_node): Likewise.
5978 (print_struct_decl): Likewise.
5979 * diagnostic.c (diagnostic_show_locus): Use pp_space.
5980
bb0d2039
BS
59812013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
5982
5983 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
5984 candidate table when replacing a candidate statement.
5985 (replace_rhs_if_not_dup): Likewise.
5986 (replace_one_candidate): Likewise.
5987
bc0ec027 59882013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 5989 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH
5990
5991 * cgraphunit.c (add_new_function): Fix logic when adding from
5992 late IPA pass.
5993 (assemble_thunk): Rename to ...
03b0ee0a
UB
5994 (expand_thunk); .. this one; export; get it working with
5995 general functions; make produced gimple valid.
bc0ec027
JH
5996 * cgraph.h (expand_thunk): Declare.
5997
0e8853ee
JH
59982013-08-02 Jan Hubicka <jh@suse.cz>
5999
03b0ee0a
UB
6000 * ipa-cp.c (gather_context_independent_values): Use
6001 ipa_get_param_move_cost.
6002 (get_replacement_map): Remove PARAM; move parameter folding
6003 into tree-inline.c
0e8853ee
JH
6004 (create_specialized_node): Update.
6005 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
6006 assert that we have gimple body; update move_cost.
6007 (count_formal_params): Assert that we have gimple body.
6008 (ipa_dump_param): New function.
6009 (ipa_alloc_node_params): Break out from ...
6010 (ipa_initialize_node_params): ... here.
6011 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
6012 (ipa_write_node_info): Stream move costs.
6013 (ipa_read_node_info): Read move costs.
6014 (ipa_update_after_lto_read): Do not recompute node params.
6015 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
6016 (ipa_get_param): Check we are not in WPA.
6017 (ipa_get_param_move_cost): New.
6018 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
6019 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
6020 parm numbers to be present.
6021
ecee672b
VM
60222013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6023
6024 PR rtl-optimization/58048
6025 * lra-constraints.c (process_alt_operands): Don't check asm
6026 operand on register.
6027
13ec0527
EB
60282013-08-02 Eric Botcazou <ebotcazou@adacore.com>
6029
6030 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
6031 the implied StoreLoad barrier for atomic operations if before.
6032
af15184a
JH
60332013-08-02 Jan Hubicka <jh@suse.cz>
6034 Martin Liska <marxin.liska@gmail.com>
6035
03b0ee0a
UB
6036 * cgraph.c (cgraph_function_body_availability): Do not check
6037 cgraph flags.
af15184a
JH
6038 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
6039 symtab_node_availability): Declare.
6040 * ipa.c (can_replace_by_local_alias): New.
6041 (function_and_variable_visibility): Use it.
03b0ee0a
UB
6042 * symtab.c (symtab_for_node_and_aliases,
6043 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 6044
01e54ef8
VM
60452013-08-02 Vladimir Makarov <vmakarov@redhat.com>
6046
6047 PR rtl-optimization/57963
03b0ee0a 6048 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM
6049 (lra_constraints): Use them.
6050
e2fd7ca7
SN
60512013-08-02 Sofiane Naci <sofiane.naci@arm.com>
6052
03b0ee0a
UB
6053 * config/arm/types.md (define_attr "type"): Add "load_acq"
6054 and "store_rel".
e2fd7ca7
SN
6055 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
6056 changes.
6057 (cortex_a53_store1): Likewise.
6058
0dcc26c3
JH
60592013-08-01 Jan Hubicka <jh@suse.cz>
6060
03b0ee0a
UB
6061 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
6062 partitions are not needed.
0dcc26c3 6063
0bd72901
UB
60642013-08-01 Uros Bizjak <ubizjak@gmail.com>
6065
6066 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
6067 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
6068 MAYBE_NON_Q_CLASS_P where appropriate.
6069
12123452
JH
60702013-08-01 Jan Hubicka <jh@suse.cz>
6071
6072 * cgraph.h (release_function_body): Declare.
6073 * tree.c (free_lang_data_in_decl): Free, parameters and return values
6074 of unused delcarations.
6075
ce852f9c
KT
60762013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6077
03b0ee0a
UB
6078 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
6079 RTL form when subtracting a constant.
ce852f9c 6080
c743b246
KT
60812013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6082
6083 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
6084 Generate canonical plus rtx with negated immediate instead of minus
6085 where appropriate.
6086 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
6087
c0c123ef
JH
60882013-08-01 Jan Hubicka <jh@suse.cz>
6089
6090 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
6091 (cgraph_release_function_body): Likewise.
6092 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
6093 * cgraph.h (cgrpah_node): Rename abstract_and_needed
6094 to used_as_abstract_origin.
6095 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
6096 symbols used as abstract origins.
6097 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB
6098 * ipa.c (symtab_remove_unreachable_nodes): Recompute
6099 used_as_abstract_origin.
c0c123ef 6100 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB
6101 used_as_abstract_origin; be ready for DECL_RESULT and
6102 DECL_ARGUMENTS to be NULL.
c0c123ef 6103
0bd72901
UB
6104 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
6105 for abstract functions.
6106 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
6107 real symbols.
c0c123ef 6108
1f029433
JH
61092013-08-01 Jan Hubicka <jh@suse.cz>
6110
6111 * profile.c (compute_value_histograms): Fix thinko.
6112
c451f4d6
SN
61132013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6114
6115 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
6116 aarch-common-protos.h to extra_headers.
6117 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
6118 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
6119 * config/aarch64/t-aarch64 (aarch-common.o): Define.
6120
34b512f5
SN
61212013-08-01 Sofiane Naci <sofiane.naci@arm.com>
6122
6123 * config/aarch64/aarch64.md (define_attr "type"): Delete.
6124 Include "../arm/types.md". Define "type" attribute for all patterns.
6125 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
6126 attribute changes.
6127
d86e633a
MM
61282013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
6129
6130 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
6131 to support power8 load fusion.
6132 (fusion_gpr_mem_load): Likewise.
6133
6134 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
6135
6136 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
6137 declarations for power8 load fusion.
6138 (emit_fusion_gpr_load): Likewise.
6139
6140 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
6141 tuning for power8, turn on fusion mode by default. Turn on sign
6142 extending fusion mode if normal fusion mode is on, and we are at
6143 -O2 or -O3.
6144 (fusion_gpr_load_p): New function, return true if we can fuse an
6145 addis instruction with a dependent load to a GPR.
6146 (emit_fusion_gpr_load): Emit the instructions for power8 load
6147 fusion to GPRs.
6148
0bd72901 6149 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM
6150 (VSX load fusion peepholes): New peepholes to fuse together an
6151 addi instruction with a VSX load instruction.
6152
6153 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
6154 peepholes to fuse an addis instruction with a load to a GPR base
6155 register. If we are supporting sign extending fusions, convert
6156 sign extending loads to zero extending loads and add an explicit
6157 sign extension.
6158
c3f35647
SN
61592013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6160
6161 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
6162 aarch-common-protos.h to extra_headers.
6163 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
6164 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
6165 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
6166 (arm_no_early_alu_shift_dep): Likewise.
6167 (arm_no_early_alu_shift_value_dep): Likewise.
6168 (arm_no_early_mul_dep): Likewise.
6169 (arm_no_early_store_addr_dep): Likewise.
6170 (arm_mac_accumulator_is_mul_result): Likewise.
6171 (arm_mac_accumulator_is_result): Likewise.
c3f35647 6172 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB
6173 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
6174 here to ...
c3f35647 6175 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
6176 (arm_no_early_alu_shift_dep): Likewise.
6177 (arm_no_early_alu_shift_value_dep): Likewise.
6178 (arm_no_early_mul_dep): Likewise.
6179 (arm_no_early_store_addr_dep): Likewise.
6180 (arm_mac_accumulator_is_mul_result): Likewise.
6181 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN
6182 * config/arm/aarch-common-protos.h: ... here. New file.
6183 * config/arm/t-arm (aarch-common.o): Define.
6184
61852013-07-31 Sofiane Naci <sofiane.naci@arm.com>
6186
6187 * config/arm/arm.md: Include new file "types.md".
6188 (define_attr "type"): Move from here to ...
6189 (define_attr "mul32"): Likewise.
6190 (define_attr "mul64"): Likewise.
6191 * config/arm/types.md: ... here. New file.
6192
228c1313
SH
61932013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
6194
6195 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
6196 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
6197
d6e9046d 61982013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG
6199
6200 * gen-pass-instances.awk: Fix offset of substr().
6201
a167b052
DM
62022013-07-31 David Malcolm <dmalcolm@redhat.com>
6203
6204 * Makefile.in (pass-instances.def): New.
6205 (passes.o): Replace dependency on passes.def with one on
6206 pass-instances.def
6207
6208 * gen-pass-instances.awk: New.
6209
6210 * passes.c (pass_manager::pass_manager): Use pass-instances.def
6211 rather than passes.def, updating local definition of NEXT_PASS
6212 macro to add an extra NUM parameter (currently unused).
6213
315f8c0e
DM
62142013-07-30 David Malcolm <dmalcolm@redhat.com>
6215
6216 * Makefile.in (PASS_MANAGER_H): New.
6217 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
6218 (passes.o): Likewise.
6219 (statistics.o): Likewise.
6220 (cgraphunit.o): Likewise.
6221 (context.o): Depend on PASS_MANAGER_H.
6222
6223 * pass_manager.h: New.
6224
6225 * cgraphunit.c (cgraph_add_new_function): Update for moves
6226 of globals to fields of pass_manager.
6227 (analyze_function): Likewise.
6228 (expand_function): Likewise.
6229 (ipa_passes): Likewise.
6230 (compile): Likewise.
6231
6232 * context.c (context::context): New.
6233 * context.h (context::context): New.
6234 (context::get_passes): New.
6235 (context::passes_): New.
6236
6237 * lto-cgraph.c (input_node): Update for moves of globals to
6238 fields of pass_manager.
6239
6240 * passes.c (all_passes): Remove, in favor of a field of the
6241 same name within the new class pass_manager.
6242 (all_small_ipa_passes): Likewise.
6243 (all_lowering_passes): Likewise.
6244 (all_regular_ipa_passes): Likewise.
6245 (all_late_ipa_passes): Likewise.
6246 (all_lto_gen_passes): Likewise.
6247 (passes_by_id): Likewise.
6248 (passes_by_id_size): Likewise.
6249 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6250 the new class pass_manager.
6251 (set_pass_for_id): Convert to...
6252 (pass_manager::set_pass_for_id): ...method.
6253 (get_pass_for_id): Convert to...
6254 (pass_manager::get_pass_for_id): ...method.
6255 (register_one_dump_file): Move body of implementation into...
6256 (pass_manager::register_one_dump_file): ...here.
6257 (register_dump_files_1): Convert to...
6258 (pass_manager::register_dump_files_1): ...method.
6259 (register_dump_files): Convert to...
6260 (pass_manager::register_dump_files): ...method.
6261 (create_pass_tab): Update for moves of globals to fields of
6262 pass_manager.
6263 (dump_passes): Move body of implementation into...
6264 (pass_manager::dump_passes): ...here.
6265 (register_pass): Move body of implementation into...
6266 (pass_manager::register_pass): ...here.
6267 (init_optimization_passes): Convert into...
6268 (pass_manager::pass_manager): ...constructor for new
6269 pass_manager class, and initialize the pass_lists array.
6270 (check_profile_consistency): Update for moves of globals to
6271 fields of pass_manager.
6272 (dump_profile_report): Move body of implementation into...
6273 (pass_manager::dump_profile_report): ...here.
6274 (ipa_write_summaries_1): Update for moves of pass lists from
6275 being globals to fields of pass_manager.
6276 (ipa_write_optimization_summaries): Likewise.
6277 (ipa_read_summaries): Likewise.
6278 (ipa_read_optimization_summaries): Likewise.
6279 (execute_all_ipa_stmt_fixups): Likewise.
6280
6281 * statistics.c (statistics_fini): Update for moves of globals to
6282 fields of pass_manager.
6283
6284 * toplev.c (general_init): Replace call to
6285 init_optimization_passes with construction of the pass_manager
6286 instance.
6287
6288 * tree-pass.h (all_passes): Remove, in favor of a field of the
6289 same name within the new class pass_manager.
6290 (all_small_ipa_passes): Likewise.
6291 (all_lowering_passes): Likewise.
6292 (all_regular_ipa_passes): Likewise.
6293 (all_lto_gen_passes): Likewise.
6294 (all_late_ipa_passes): Likewise.
6295 (passes_by_id): Likewise.
6296 (passes_by_id_size): Likewise.
6297 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
6298 the new class pass_manager.
6299 (get_pass_for_id): Remove.
6300
02cab1c1
RE
63012013-07-30 Richard Earnshaw <rearnsha@arm.com>
6302
6303 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
6304 configs.
6305
b2d3f886
RE
63062013-07-30 Richard Earnshaw <rearnsha@arm.com>
6307
6308 * arm.md (mulhi3): New expand pattern.
6309
605e86fa
JH
63102013-07-30 Jan Hubicka <jh@suse.cz>
6311 Martin Liska <marxin.liska@gmail.com>
6312
6313 * profile.c (compute_value_histograms): Do not ICE when
6314 there is mismatch only on some counters.
6315
14a87636
ZC
63162013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6317
6318 PR rtl-optimization/57637
6319 * function.c (move_insn_for_shrink_wrap): Also check the
6320 GEN set of the LIVE problem for the liveness analysis
6321 if it exists, otherwise give up.
6322
0100cd3f
BS
63232013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
6324
6325 PR tree-optimization/57993
6326 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
6327 replaced statement in the candidate table.
6328 (phi_add_costs): Return infinite cost when the hidden basis does
6329 not dominate all phis on which the candidate is dependent.
6330 (replace_one_candidate): Record replaced statement in the
6331 candidate table.
6332
f0d811f7
JR
63332013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
6334
6335 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
6336 (ashlv2si3): New expander.
6337 (*ashlv2si3_i): New define_insn_and_split.
6338 * predicates.md (float_operation): Allow patterns with three
6339 basic sub-patterns.
6340
bba33211
JR
6341 PR rtl-optimization/58021
6342 * mode-switching.c (create_pre_exit): Always split off preceding
6343 insns if we are not at the basic block head.
6344
ff3f3951
MR
63452013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
6346
6347 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
6348 (UCLIBC_DYNAMIC_LINKER): New macro.
6349 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
6350 `-mnan=2008'.
6351 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
6352 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
6353 `-mnan=2008'.
6354 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
6355 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
6356 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
6357 for SF and DF modes. Use ieee_quad_format for TF mode.
6358 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
6359 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
6360 (mips_option_override): Handle `-mnan=legacy'.
6361 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
6362 `-mabs=2008' and `-mnan=2008'.
6363 (OPTION_DEFAULT_SPECS): Add "nan" default.
6364 (ASM_SPEC): Handle `-mnan='.
6365 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
6366 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
6367 comment accordingly.
6368 (neg<mode>2): Likewise.
6369 * config/mips/mips.opt (mabs, mnan): New options.
6370 * doc/install.texi (Configuration): Document `--with-nan=' option.
6371 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
6372 `-mnan=' options.
6373 (MIPS Options): Document them.
6374 * config.gcc <mips*-*-*>: Handle `--with-nan='.
6375 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
6376 * configure: Regenerate.
6377 * config.in: Regenerate.
6378
576e2f37
UB
63792013-07-29 Uros Bizjak <ubizjak@gmail.com>
6380
6381 * config/i386/i386.md (float post-reload splitters): Do not check
6382 for subregs of SSE registers.
6383
63842013-07-29 Uros Bizjak <ubizjak@gmail.com>
6385 H.J. Lu <hongjiu.lu@intel.com>
6386
6387 PR target/57954
6388 PR target/57988
6389 * config/i386/i386.md (post-reload splitter
6390 to avoid partial SSE reg dependency stalls): New pattern.
6391
a71f0749
DV
63922013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
6393
6394 * config/s390/s390.md ("movcc"): Swap load and store instructions.
6395
4c97f1cc
JR
63962013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
6397
6398 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
6399 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
6400
b0c733d8
CC
64012013-07-26 Cary Coutant <ccoutant@google.com>
6402
6403 * dwarf2out.c (die_checksum_ordered): Don't include template
6404 instantiations in signature.
6405 (is_template_parameter): New function.
6406 (is_template_instantiation): New function.
6407 (generate_skeleton_bottom_up): Don't include template instantiations
6408 in type unit DIE.
6409 (generate_skeleton): Likewise.
6410 (break_out_comdat_types): Move recursive call to break out nested
6411 types earlier.
6412 (prune_unused_types_mark_generic_parms_dies): Call
6413 is_template_parameter.
6414
040d8a1c
IB
64152013-07-26 Ian Bolton <ian.bolton@arm.com>
6416
6417 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
6418 uses vector registers.
89fdc743 6419 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 6420
85bd4ac6 64212013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 6422 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT
6423
6424 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 6425 where appropriate.
85bd4ac6
KT
6426 * config/arm/arm.md (movcond_addsi): New splitter.
6427
b6051207
SA
64282013-07-25 Sterling Augustine <saugustine@google.com>
6429
6430 * dwarf2out.c (size_of_pubnames): Move code to...
6431 (include_pubname_in_output): ...here. New.
6432 (want_pubnames): Rearrange.
6433 (output_pubnames): Call include_pubname_in_output. Move assertion.
6434
41a7c215
CM
64352013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6436
6437 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
6438
64392013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
6440
6441 PR target/38836
6442 * doc/extend.texi: Remove obsolete builtins. Fix
6443 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
6444
49bde175
JH
64452013-07-25 Jan Hubicka <jh@suse.cz>
6446
6447 * cgraph.c (release_function_body): Break out from ...
6448 (cgraph_release_function_body): ... this one; also release DECL_RESULT
6449 and DECL_ARGUMENTS.
6450 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
6451 old_tree in the map.
6452 (create_specialized_node): Update.
6453 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
6454 into index.
41a7c215
CM
6455 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
6456 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH
6457 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
6458 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
6459 DECL_RESULT.
6460
4870352d
KT
64612013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6462
6463 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
6464 addsi3_carryin_alt2_<optab>): Correct output template.
6465
1d6aee1c
KT
64662013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6467
6468 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
6469 Adjust for arm_restrict_it.
6470 Remove trailing whitespace.
6471
c7cafd75
MK
64722013-07-25  Mark Kettenis  <kettenis@openbsd.org>
6473
017d38f5
MK
6474 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
6475 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
6476
c7cafd75
MK
6477 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
6478
00b7527b
VM
64792013-07-25 Vladimir Makarov <vmakarov@redhat.com>
6480
6481 PR rtl-optimization/57960
6482 * lra-constraints.c (process_alt_operands): Use the right mode
6483 when checking strict_low.
6484
4f63dfc6
JH
64852013-07-25 Jan Hubicka <jh@suse.cz>
6486
6487 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
6488 * cgraph.c (cgraph_remove_node): Do not release function body
6489 when in cgraph streaming.
6490 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM
6491 in other partitions are not considered reachable; fix handling of
6492 clones.
4f63dfc6 6493
e6dcfa18
RR
64942013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6495
6496 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
6497
0b93d3b6
RR
64982013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
6499
6500 PR target/19599
6501 PR target/57731
bb80c2eb 6502 PR target/57837
e6dcfa18 6503 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 6504 Ss with US. Adjust output for v5 and v4t.
41a7c215 6505 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 6506
0b93d3b6
RR
6507 * config/arm/constraints.md ("Ss"): Rename to US.
6508
690688b3
TG
65092013-07-25 Terry Guo <terry.guo@arm.com>
6510
6511 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
6512 shift_add/shift_sub0/shift_sub1 RTXs.
6513
a4ad093b 65142013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 6515 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS
6516
6517 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
6518 (altivec_vpks<VI_char>ss): Likewise.
6519 (altivec_vpks<VI_char>us): Likewise.
6520 (altivec_vpku<VI_char>us): Likewise.
6521 (altivec_vpku<VI_char>um): Likewise.
6522
8c5005ce
DM
65232013-07-24 David Malcolm <dmalcolm@redhat.com>
6524
6525 Introduce context class.
6526
6527 * Makefile.in (CONTEXT_H): New.
6528 (OBJS): Add context.o.
6529 (toplev.o): Add CONTEXT_H to dependencies.
6530 (context.o): New.
6531
41a7c215 6532 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM
6533
6534 * context.c: New.
6535
6536 * context.h: New.
6537
ce4a9422
JR
65382013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
6539
6540 PR rtl-optimization/57968
6541 * mode-switching.c (create_pre_exit): Allow instructions that
6542 don't set a return register to need a non-exit mode.
6543
493f4c9e 65442013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 6545 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 6546
b5860fd3
WS
6547 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
6548 operands to vperm for little endian.
6549 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
6550 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 6551
73792b92 65522013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 6553 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS
6554
6555 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
6556 two operands for little-endian.
6557
14c21302
SE
65582013-07-23 Steve Ellcey <sellcey@mips.com>
6559
6560 * config/mips/mips.c (mips_case_values_threshold): New.
6561 (TARGET_CASE_VALUES_THRESHOLD): Define.
6562
a6056198 65632013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 6564 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS
6565
6566 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
6567 selection of field for vector splat in little endian mode.
6568
dd7a40e1
MM
65692013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6570
6571 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
6572 expanders to rs6000.md.
6573 (ior<mode>3): Likewise.
6574 (and<mode>3): Likewise.
6575 (one_cmpl<mode>2): Likewise.
6576 (nor<mode>3): Likewise.
6577 (andc<mode>3): Likewise.
6578 (eqv<mode>3): Likewise.
6579 (nand<mode>3): Likewise.
6580 (orc<mode>3): Likewise.
6581
6582 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
6583 declaration.
6584
6585 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
6586 to split multi-word logical operations.
6587 (rs6000_split_logical_di): Likewise.
6588 (rs6000_split_logical): Likewise.
6589
6590 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
6591 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
6592 and allow TImode operations in 32-bit.
6593 (vsx_and<mode>3_64bit): Likewise.
6594 (vsx_ior<mode>3_32bit): Likewise.
6595 (vsx_ior<mode>3_64bit): Likewise.
6596 (vsx_xor<mode>3_32bit): Likewise.
6597 (vsx_xor<mode>3_64bit): Likewise.
6598 (vsx_one_cmpl<mode>2_32bit): Likewise.
6599 (vsx_one_cmpl<mode>2_64bit): Likewise.
6600 (vsx_nor<mode>3_32bit): Likewise.
6601 (vsx_nor<mode>3_64bit): Likewise.
6602 (vsx_andc<mode>3_32bit): Likewise.
6603 (vsx_andc<mode>3_64bit): Likewise.
6604 (vsx_eqv<mode>3_32bit): Likewise.
6605 (vsx_eqv<mode>3_64bit): Likewise.
6606 (vsx_nand<mode>3_32bit): Likewise.
6607 (vsx_nand<mode>3_64bit): Likewise.
6608 (vsx_orc<mode>3_32bit): Likewise.
6609 (vsx_orc<mode>3_64bit): Likewise.
6610
6611 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
6612 logical types in GPRs.
6613
6614 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
6615 logical insns to rs6000.md, and allow TImode operations in
6616 32-bit.
6617 (altivec_ior<mode>3): Likewise.
6618 (altivec_xor<mode>3): Likewise.
6619 (altivec_one_cmpl<mode>2): Likewise.
6620 (altivec_nor<mode>3): Likewise.
6621 (altivec_andc<mode>3): Likewise.
6622
6623 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
6624 attributes for moving the 128-bit logical operations into
6625 rs6000.md.
6626 (BOOL_REGS_OUTPUT): Likewise.
6627 (BOOL_REGS_OP1): Likewise.
6628 (BOOL_REGS_OP2): Likewise.
6629 (BOOL_REGS_UNARY): Likewise.
6630 (BOOL_REGS_AND_CR0): Likewise.
6631 (one_cmpl<mode>2): Add support for DI logical operations on
6632 32-bit, splitting the operations to 32-bit.
6633 (anddi3): Likewise.
6634 (iordi3): Likewise.
6635 (xordi3): Likewise.
6636 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
6637 changes to combine the 32/64-bit code, allow logical operations on
6638 TI mode in 32-bit, and to use similar match_operator patterns like
6639 scalar mode uses. Combine the Altivec and VSX code for logical
6640 operations, and move it here.
6641 (ior<mode>3, 128-bit types): Likewise.
6642 (xor<mode>3, 128-bit types): Likewise.
6643 (one_cmpl<mode>3, 128-bit types): Likewise.
6644 (nor<mode>3, 128-bit types): Likewise.
6645 (andc<mode>3, 128-bit types): Likewise.
6646 (eqv<mode>3, 128-bit types): Likewise.
6647 (nand<mode>3, 128-bit types): Likewise.
6648 (orc<mode>3, 128-bit types): Likewise.
6649 (and<mode>3_internal): Likewise.
6650 (bool<mode>3_internal): Likewise.
6651 (boolc<mode>3_internal1): Likewise.
6652 (boolc<mode>3_internal2): Likewise.
6653 (boolcc<mode>3_internal1): Likewise.
6654 (boolcc<mode>3_internal2): Likewise.
6655 (eqv<mode>3_internal1): Likewise.
6656 (eqv<mode>3_internal2): Likewise.
6657 (one_cmpl1<mode>3_internal): Likewise.
6658
12211b99 66592013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH
6660
6661 * config/microblaze/microblaze.c (microblaze_expand_prologue):
6662 Rename flag_stack_usage to flag_stack_usage_info.
6663
12211b99 66642013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 6665
a5f257fd
ME
6666 * config/microblaze/sync.md: New file.
6667 * config/microblaze/microblaze.md: Include sync.md
6668 * config/microblaze/microblaze.c: Add print_operand 'y'.
6669 * config/microblaze/constraints.md: Add memory_contraint
6670 'Q' which is a single register.
450b0ebb 6671
49fba14f
EB
66722013-07-23 Eric Botcazou <ebotcazou@adacore.com>
6673
6674 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
6675
c7b3b99f
PCC
66762013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
6677
6678 * reload.c (find_reloads): Exit loop once we find this operand
6679 cannot be reloaded somehow for this alternative.
6680
6681 * reload.c (find_reloads): Exit loop once we find a hard register.
6682
6683 * rtlanal.c (computed_jump_p): Exit loop once we find label
6684 reference is used.
6685
6686 * i386.c (ix86_pad_returns): Exit loop after setting replace.
6687
6688 * cfgloopmanip.c (remove_path): Exit loop after setting
6689 irred_invalidated.
6690
6691 * gensupport.c (subst_dup): Avoid loop if code is not
6692 MATCH_DUP nor MATCH_OP_DUP.
6693
c8fbf1fa
NBJ
66942013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
6695
6696 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
6697
9259db42
YZ
66982013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6699
6700 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
6701 true for SP_REGNUM if mode == ptr_mode.
6702 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
6703 with value R0_REGNUM + 31.
6704
00edcfbe
YZ
67052013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6706
6707 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
6708 pad pointer-typed argument downward.
6709
43be9a95
YZ
67102013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6711
6712 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
6713 and __ILP32__ when the ILP32 model is in use.
6714
28514dda
YZ
67152013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6716
6717 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
6718 (aarch64_load_symref_appropriately): In the case of
6719 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
6720 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
6721 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
6722 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
6723 if 'mode' doesn't equal to 'ptr_mode'.
6724 (aarch64_output_mi_thunk): Add an assertion on the alignment of
6725 'vcall_offset'; change to call aarch64_emit_move differently depending
6726 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
6727 to calculate the upper bound of 'vcall_offset'.
6728 (aarch64_cannot_force_const_mem): Change to also return true if
6729 mode != ptr_mode.
6730 (aarch64_legitimize_reload_address): In the case of large
6731 displacements, add new local variable 'xmode' and an assertion
6732 based on it; change to use 'xmode' to generate the new rtx and
6733 reload.
6734 (aarch64_asm_trampoline_template): Change to generate the template
6735 differently depending on TARGET_ILP32 or not; change to use
6736 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
6737 (aarch64_trampoline_size): Removed.
6738 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
6739 and replace immediate literals with it. Change to use 'ptr_mode'
6740 instead of 'DImode' and call convert_memory_address if the mode
6741 of 'fnaddr' doesn't equal to 'ptr_mode'.
6742 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
6743 to output symbol.
6744 (aarch64_elf_asm_destructor): Likewise.
6745 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
6746 on TARGET_ILP32 instead of aarch64_trampoline_size.
6747 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
6748 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
6749 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
6750 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
6751 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
6752 (storewb_pair<GPI:mode>_<P:mode>): ... this.
6753 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
6754 depending on the value of 'mode'.
6755 (add_losym_<mode>): New.
6756 (ldr_got_small_<mode>): New, based on ldr_got_small.
6757 (ldr_got_small): Remove.
6758 (ldr_got_small_sidi): New.
6759 * config/aarch64/iterators.md (P): New.
6760 (PTR): Change to 'ptr_mode' in the condition.
6761
17a819cb
YZ
67622013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
6763
6764 * config.gcc (aarch64*-*-*): Support --with-abi.
6765 (aarch64*-*-elf): Support --with-multilib-list.
6766 (aarch64*-*-linux*): Likewise.
6767 (supported_defaults): Add abi to aarch64*-*-*.
6768 * configure.ac: Mention AArch64 for --with-multilib-list.
6769 * configure: Re-generated.
6770 * config/aarch64/biarchilp32.h: New file.
6771 * config/aarch64/biarchlp64.h: New file.
6772 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
6773 (ABI_SPEC): Ditto.
6774 (MULTILIB_DEFAULTS): Ditto.
6775 (DRIVER_SELF_SPECS): Ditto.
6776 (ASM_SPEC): Update to also substitute -mabi.
6777 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
6778 file whose name depends on -mabi= and -mbig-endian.
6779 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
6780 TARGET_ILP32.
6781 (POINTER_SIZE): New define.
6782 (POINTERS_EXTEND_UNSIGNED): Ditto.
6783 (enum aarch64_abi_type): New enumeration tag.
6784 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
6785 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
6786 (TARGET_ILP32): New define.
6787 * config/aarch64/aarch64.opt (mabi): New.
6788 (aarch64_abi): New.
6789 (ilp32, lp64): New values for -mabi.
6790 * config/aarch64/t-aarch64 (comma): New define.
6791 (MULTILIB_OPTIONS): Ditto.
6792 (MULTILIB_DIRNAMES): Ditto.
6793 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
6794 * doc/invoke.texi: Document -mabi for AArch64.
6795
9f9cbdce
GJL
67962013-07-23 Georg-Johann Lay <avr@gjlay.de>
6797
6798 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
6799
7061977b 68002013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 6801 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS
6802
6803 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
6804 endianness when selecting field to splat.
6805
75a41878
EC
68062013-07-22 Eric Christopher <echristo@gmail.com>
6807
6808 * dwarf2out.c (die_odr_checksum): New function to use
6809 CHECKSUM_ macros and ULEB128 for DIE tag.
6810 (generate_type_signature): Use.
41a7c215 6811
38ae58ca
EB
68122013-07-22 Eric Botcazou <ebotcazou@adacore.com>
6813
6814 * config.gcc (sparc*-*-*): Accept leon3 processor.
6815 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
6816 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
6817 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
6818 * config/sparc/sparc.opt (enum processor_type): Add leon3.
6819 (mfix-ut699): Adjust comment.
6820 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
6821 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
6822 (CPP_CPU_SPEC): Likewise.
6823 (ASM_CPU_SPEC): Likewise.
6824 * config/sparc/sparc.c (leon3_cost): New constant.
6825 (sparc_option_override): Add leon3 support.
6826 (mem_ref): New function.
6827 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
6828 (sparc_do_work_around_errata): Look into the instruction in the delay
6829 slot and adjust accordingly. Add fix for the data cache nullify issues
6830 of the UT699. Change insertion position for the NOP.
6831 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
6832 (leon3_load): New reservation.
6833 (leon_store): Bump latency to 2.
6834 (grfpu): New automaton.
6835 (grfpu_alu): New unit.
6836 (grfpu_ds): Likewise.
6837 (leon_fp_alu): Adjust.
6838 (leon_fp_mult): Delete.
6839 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
6840 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
6841 * config/sparc/sparc.md (cpu): Add leon3.
6842 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
6843 (swapsi): Likewise.
6844 (atomic_test_and_set): Likewise.
6845 (ldstub): Likewise.
6846
04dfc6df
JU
68472013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
6848
6849 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
6850 default for R5900 targets.
6851 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
6852 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
6853 * config/mips/mips.c (mips_option_override): Report an error for
6854 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
6855 for -march=r5900 -mhard-float.
6856
c7b3b99f 68572013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC
6858
6859 * df-problems.c (can_move_insns_across): Exit loop once we
6860 find a non-fixed, non-global register.
6861
6862 * ipa-pure-const.c (propagate_nothrow): Exit loop after
6863 setting can_throw.
6864
6865 * omega.c (omega_eliminate_red): Break after setting red_found.
6866 (omega_problem_has_red_equations): Similarly after setting found.
6867 (omega_query_variable): Similarly after setting coupled.
6868
9b6e6981
MP
68692013-07-22 Marek Polacek <polacek@redhat.com>
6870
6871 * gimplify.c: Don't include gimple.h twice.
6872
5e5f7673
KT
68732013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6874
6875 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
6876 instead of TARGET_THUMB1.
6877 (Pz): New constraint.
6878 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
6879 encodings.
6880 (compare_negsi_si): Likewise.
6881 (compare_addsi2_op0): Likewise.
6882 (compare_addsi2_op1): Likewise.
6883 (addsi3_carryin_<optab>): Likewise.
6884 (addsi3_carryin_alt2_<optab>): Likewise.
6885 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
6886 for arm_restrict_it.
6887 (subsi3_carryin): Likewise.
6888 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
6889 (minmax_arithsi): Disable for arm_restrict_it.
6890 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
6891 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
6892 (satsi_<SAT:code>_shift): Likewise.
6893 (arm_shiftsi3): Add alternative for 16-bit encoding.
6894 (arm32_movhf): Disable for arm_restrict_it.
6895 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
6896 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
6897
859abddd
SN
68982013-07-22 Sofiane Naci <sofiane.naci@arm.com>
6899
6900 * config/arm/arm.md (attribute "insn"): Delete.
6901 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
6902 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
6903 (not_shiftsi): Update for attribute change.
6904 (not_shiftsi_compare0): Likewise.
6905 (not_shiftsi_compare0_scratch): Likewise.
6906 (arm_one_cmplsi2): Likewise.
6907 (thumb1_one_cmplsi2): Likewise.
6908 (notsi_compare0): Likewise.
6909 (notsi_compare0_scratch): Likewise.
6910 (thumb1_movdi_insn): Likewise.
6911 (arm_movsi_insn): Likewise.
6912 (movhi_insn_arch4): Likewise.
6913 (movhi_bytes): Likewise.
6914 (arm_movqi_insn): Likewise.
6915 (thumb1_movqi_insn): Likewise.
6916 (arm32_movhf): Likewise.
6917 (thumb1_movhf): Likewise.
6918 (arm_movsf_soft_insn): Likewise.
6919 (thumb1_movsf_insn): Likewise.
6920 (thumb_movdf_insn): Likewise.
6921 (movsicc_insn): Likewise.
6922 (movsfcc_soft_insn): Likewise.
6923 (and_scc): Likewise.
6924 (cond_move): Likewise.
6925 (if_move_not): Likewise.
6926 (if_not_move): Likewise.
6927 (if_shift_move): Likewise.
6928 (if_move_shift): Likewise.
6929 (if_shift_shift): Likewise.
6930 (if_not_arith): Likewise.
6931 (if_arith_not): Likewise.
6932 (cond_move_not): Likewise.
6933 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
6934 (neon_mov<mode>): Likewise.
6935 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
6936 (thumb2_movsi_vfp): Likewise.
6937 (movsf_vfp): Likewise.
6938 (thumb2_movsf_vfp): Likewise.
371e77e3
UB
6939 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
6940 change.
859abddd
SN
6941 (cortexa7_older_only): Likewise.
6942 (cortexa7_younger): Likewise.
6943 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
6944 (1020alu_shift_op): Likewise.
6945 (1020alu_shift_reg_op): Likewise.
6946 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
6947 (alu_shift_op): Likewise.
6948 (alu_shift_reg_op): Likewise.
6949 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
6950 (11_alu_shift_op): Likewise.
6951 (11_alu_shift_reg_op): Likewise.
6952 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
6953 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
6954 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6955 change.
859abddd
SN
6956 (cortex_a15_alu_shift): Likewise.
6957 (cortex_a15_alu_shift_reg): Likewise.
6958 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
6959 (cortex_a5_alu_shift): Likewise.
371e77e3
UB
6960 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
6961 change.
859abddd 6962 (cortex_a53_alu_shift): Likewise.
371e77e3
UB
6963 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
6964 change.
859abddd
SN
6965 (cortex_a7_alu_reg): Likewise.
6966 (cortex_a7_alu_shift): Likewise.
6967 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6968 (cortex_a8_alu_shift): Likewise.
6969 (cortex_a8_alu_shift_reg): Likewise.
6970 (cortex_a8_mov): Likewise.
6971 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6972 (cortex_a9_dp_shift): Likewise.
6973 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
6974 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
6975 (cortex_r4_mov): Likewise.
6976 (cortex_r4_alu_shift): Likewise.
6977 (cortex_r4_alu_shift_reg): Likewise.
6978 * config/arm/fa526.md (526_alu_op): Update for attribute change.
6979 (526_alu_shift_op): Likewise.
6980 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
6981 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
6982 (626te_alu_shift_op): Likewise.
6983 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
6984 (726te_alu_op): Likewise.
6985 (726te_alu_shift_op): Likewise.
6986 (726te_alu_shift_reg_op): Likewise.
6987 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
6988 (mp626_alu_shift_op): Likewise.
6989 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
6990 (pj4_alu_e1_conds): Likewise.
6991 (pj4_alu): Likewise.
6992 (pj4_alu_conds): Likewise.
6993 (pj4_shift): Likewise.
6994 (pj4_shift_conds): Likewise.
6995 (pj4_alu_shift): Likewise.
6996 (pj4_alu_shift_conds): Likewise.
6997
95b97fac
KT
69982013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6999
7000 * config/arm/predicates.md (shiftable_operator_strict_it):
7001 New predicate.
7002 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
7003 Disable cond_exec version for arm_restrict_it.
7004 (thumb2_smaxsi3): Convert to generate cond_exec.
7005 (thumb2_sminsi3): Likewise.
7006 (thumb32_umaxsi3): Likewise.
7007 (thumb2_uminsi3): Likewise.
7008 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
7009 (thumb2_neg_abssi2): Likewise.
7010 (thumb2_mov_scc): Add alternative for 16-bit encoding.
7011 (thumb2_movsicc_insn): Adjust alternatives.
7012 (thumb2_mov_negscc): Disable for arm_restrict_it.
7013 (thumb2_mov_negscc_strict_it): New pattern.
7014 (thumb2_mov_notscc_strict_it): New pattern.
7015 (thumb2_mov_notscc): Disable for arm_restrict_it.
7016 (thumb2_ior_scc): Likewise.
7017 (thumb2_ior_scc_strict_it): New pattern.
7018 (thumb2_cond_move): Adjust for arm_restrict_it.
7019 (thumb2_cond_arith): Disable for arm_restrict_it.
7020 (thumb2_cond_arith_strict_it): New pattern.
7021 (thumb2_cond_sub): Adjust for arm_restrict_it.
7022 (thumb2_movcond): Likewise.
7023 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
7024 (thumb2_zero_extendhisi2_v6): Likewise.
7025 (thumb2_zero_extendqisi2_v6): Likewise.
7026 (orsi_notsi_si): Likewise.
7027 (orsi_not_shiftsi_si): Likewise.
7028
d5a6ef82
GJL
70292013-07-22 Georg-Johann Lay <avr@gjlay.de>
7030
7031 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
7032 instruction sequence is 1 byte shorter.
7033
d9e71154
UB
70342013-07-22 Uros Bizjak <ubizjak@gmail.com>
7035
7036 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
7037 it is not needed after split.
7038
8c6381ec
IS
70392013-07-20 Iain Sandoe <iain@codesourcery.com>
7040
7041 PR target/51784
7042 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
7043 second label for nonlocal goto receivers. Don't output pic base labels
7044 unless we're producing PIC; mark that action unreachable().
7045 (ix86_save_reg): If the function contains a nonlocal label, save the
7046 PIC base reg.
7047 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 7048 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS
7049 (update_pic_label_number_if_needed): New.
7050 (machopic_output_function_base_name): Adjust for nonlocal receiver
7051 case.
7052 (machopic_should_output_picbase_label): New.
7053 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
7054 (nonlocal_goto_receiver): New insn and split.
7055
d05d0709
JG
70562013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
7057
7058 * config/aarch64/aarch64-builtins.c
7059 (aarch64_fold_builtin): Fold abs in all modes.
7060 * config/aarch64/aarch64-simd-builtins.def
7061 (abs): Enable for all modes.
7062 * config/aarch64/arm_neon.h
7063 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
7064 (vabs_f64): Add missing intrinsic.
7065
ba590e63
IB
70662013-07-19 Ian Bolton <ian.bolton@arm.com>
7067
7068 * config/aarch64/arm_neon.h (vabs_s64): New function
7069
02371798
GJL
70702013-07-19 Georg-Johann Lay <avr@gjlay.de>
7071
7072 PR target/57516
7073 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
7074 * config/avr/avr.md (adjust_len): Add `round'.
7075 * config/avr/avr-protos.h (avr_out_round): New prototype.
7076 (avr_out_plus): Add `out_label' argument.
7077 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
7078 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
7079 Handle the case where `insn' is just a pattern.
7080 (avr_out_bitop): Handle the case where `insn' is just a pattern.
7081 (avr_out_round): New function.
7082 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
7083
12211b99 70842013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH
7085
7086 * config/microblaze/microblaze.c (microblaze_expand_prologue):
7087 Add check for flag_stack_usage to handle -fstack-usage support
7088
12211b99 70892013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH
7090
7091 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
7092 interaction for new Power8 flags and VSX.
7093
df9dda2d
ST
70942013-07-18 Sriraman Tallam <tmsriram@google.com>
7095
7096 PR middle-end/57698
7097 * tree-inline.c (expand_call_inline): Emit errors during
7098 early_inlining only if optimization is not turned on.
7099
0f9cdef4
DM
71002013-07-18 David Malcolm <dmalcolm@redhat.com>
7101
7102 * passes.def: New.
7103
7104 * passes.c (init_optimization_passes): Move the construction of
7105 the pass hierarchy into a new passes.def file.
7106
7107 * Makefile.in (passes.o): Add dependency on passes.def.
7108
2efa4087
DM
71092013-07-18 David Malcolm <dmalcolm@redhat.com>
7110
7111 * passes.c (init_optimization_passes): Introduce macros for
7112 constructing the tree of passes (INSERT_PASSES_AFTER,
7113 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
7114 TERMINATE_PASS_LIST).
7115
bc404e1b
VM
71162013-07-18 Vladimir Makarov <vmakarov@redhat.com>
7117 Wei Mi <wmi@google.com>
7118
7119 PR rtl-optimization/57878
7120 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
7121 top.
7122 (reload_pseudo_compare_func): Check nregs first for reload
7123 pseudos.
7124
d675843f
DM
71252013-07-18 David Malcolm <dmalcolm@redhat.com>
7126
7127 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
7128
c7b3b99f 71292013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC
7130
7131 * read-rtl.c (validate_const_int): Once an invalid character is
7132 seen, quit the loop.
7133
7134 * gengtype.c (write_roots): Similarly once we find the "deletable"
7135 or "if_marked" option.
7136
52fceb44
SN
71372013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7138
7139 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
7140 "xtab" and "sat". Move value "clz" from here to ...
7141 (attriubte "type"): ... here.
7142 (satsi_<SAT:code>): Delete "insn" attribute.
7143 (satsi_<SAT:code>_shift): Likewise.
7144 (arm_zero_extendqisi2addsi): Likewise.
7145 (arm_extendqisi2addsi): Likewise.
7146 (clzsi2): Update for attribute changes.
7147 (rbitsi2): Likewise.
371e77e3
UB
7148 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
7149 attribute.
52fceb44
SN
7150 (arm_usatsihi): Likewise.
7151 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7152
006bd006
SN
71532013-07-18 Sofiane Naci <sofiane.naci@arm.com>
7154
7155 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB
7156 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
7157 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN
7158 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
7159 in alphabetical order.
7160 (attribute "core_cycles"): Update for attribute changes.
7161 (arm_addsi3): Likewise.
7162 (addsi3_compare0): Likewise.
7163 (addsi3_compare0_scratch): Likewise.
7164 (addsi3_compare_op1): Likewise.
7165 (addsi3_compare_op2): Likewise.
7166 (compare_addsi2_op0): Likewise.
7167 (compare_addsi2_op1): Likewise.
7168 (addsi3_carryin_shift_<optab>): Likewise.
7169 (subsi3_carryin_shift): Likewise.
7170 (rsbsi3_carryin_shift): Likewise.
7171 (arm_subsi3_insn): Likewise.
7172 (subsi3_compare0): Likewise.
7173 (subsi3_compare): Likewise.
7174 (arm_andsi3_insn): Likewise.
7175 (thumb1_andsi3_insn): Likewise.
7176 (andsi3_compare0): Likewise.
7177 (andsi3_compare0_scratch): Likewise.
7178 (zeroextractsi_compare0_scratch
7179 (andsi_not_shiftsi_si): Likewise.
7180 (iorsi3_insn): Likewise.
7181 (iorsi3_compare0): Likewise.
7182 (iorsi3_compare0_scratch): Likewise.
7183 (arm_xorsi3): Likewise.
7184 (thumb1_xorsi3_insn): Likewise.
7185 (xorsi3_compare0): Likewise.
7186 (xorsi3_compare0_scratch): Likewise.
7187 (satsi_<SAT:code>_shift): Likewise.
7188 (rrx): Likewise.
7189 (arm_shiftsi3): Likewise.
7190 (shiftsi3_compare0): Likewise.
7191 (not_shiftsi): Likewise.
7192 (not_shiftsi_compare0): Likewise.
7193 (not_shiftsi_compare0_scratch): Likewise.
7194 (arm_one_cmplsi2): Likewise.
7195 (thumb_one_complsi2): Likewise.
7196 (notsi_compare0): Likewise.
7197 (notsi_compare0_scratch): Likewise.
7198 (thumb1_zero_extendhisi2): Likewise.
7199 (arm_zero_extendhisi2): Likewise.
7200 (arm_zero_extendhisi2_v6): Likewise.
7201 (arm_zero_extendhisi2addsi): Likewise.
7202 (thumb1_zero_extendqisi2): Likewise.
7203 (thumb1_zero_extendqisi2_v6): Likewise.
7204 (arm_zero_extendqisi2): Likewise.
7205 (arm_zero_extendqisi2_v6): Likewise.
7206 (arm_zero_extendqisi2addsi): Likewise.
7207 (thumb1_extendhisi2): Likewise.
7208 (arm_extendhisi2): Likewise.
7209 (arm_extendhisi2_v6): Likewise.
7210 (arm_extendqisi): Likewise.
7211 (arm_extendqisi_v6): Likewise.
7212 (arm_extendqisi2addsi): Likewise.
7213 (thumb1_extendqisi2): Likewise.
7214 (thumb1_movdi_insn): Likewise.
7215 (arm_movsi_insn): Likewise.
7216 (movsi_compare0): Likewise.
7217 (movhi_insn_arch4): Likewise.
7218 (movhi_bytes): Likewise.
7219 (arm_movqi_insn): Likewise.
7220 (thumb1_movqi_insn): Likewise.
7221 (arm32_movhf): Likewise.
7222 (thumb1_movhf): Likewise.
7223 (arm_movsf_soft_insn): Likewise.
7224 (thumb1_movsf_insn): Likewise.
7225 (movdf_soft_insn): Likewise.
7226 (thumb_movdf_insn): Likewise.
7227 (arm_cmpsi_insn): Likewise.
7228 (cmpsi_shiftsi): Likewise.
7229 (cmpsi_shiftsi_swp): Likewise.
7230 (arm_cmpsi_negshiftsi_si): Likewise.
7231 (movsicc_insn): Likewise.
7232 (movsfcc_soft_insn): Likewise.
7233 (arith_shiftsi): Likewise.
7234 (arith_shiftsi_compare0
7235 (arith_shiftsi_compare0_scratch
7236 (sub_shiftsi): Likewise.
7237 (sub_shiftsi_compare0
7238 (sub_shiftsi_compare0_scratch
7239 (and_scc): Likewise.
7240 (cond_move): Likewise.
7241 (if_plus_move): Likewise.
7242 (if_move_plus): Likewise.
7243 (if_move_not): Likewise.
7244 (if_not_move): Likewise.
7245 (if_shift_move): Likewise.
7246 (if_move_shift): Likewise.
7247 (if_shift_shift): Likewise.
7248 (if_not_arith): Likewise.
7249 (if_arith_not): Likewise.
7250 (cond_move_not): Likewise.
7251 (thumb1_ashlsi3): Set type attribute.
7252 (thumb1_ashrsi3): Likewise.
7253 (thumb1_lshrsi3): Likewise.
7254 (thumb1_rotrsi3): Likewise.
7255 (shiftsi3_compare0_scratch): Likewise.
7256 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
7257 (neon_mov<mode>): Likewise.
371e77e3
UB
7258 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
7259 attribute changes.
006bd006
SN
7260 (thumb2_movsi_insn): Likewise.
7261 (thumb2_cmpsi_neg_shiftsi): Likewise.
7262 (thumb2_extendqisi_v6): Likewise.
7263 (thumb2_zero_extendhisi2_v6): Likewise.
7264 (thumb2_zero_extendqisi2_v6): Likewise.
7265 (thumb2_shiftsi3_short): Likewise.
7266 (thumb2_addsi3_compare0_scratch): Likewise.
7267 (orsi_not_shiftsi_si): Likewise.
7268 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7269 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
7270 changes.
7271 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7272 (1020alu_shift_op): Likewise.
7273 (1020alu_shift_reg_op): Likewise.
7274 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7275 (alu_shift_op): Likewise.
7276 (alu_shift_reg_op): Likewise.
7277 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7278 (11_alu_shift_op): Likewise.
7279 (11_alu_shift_reg_op): Likewise.
7280 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7281 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
7282 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7283 changes.
006bd006
SN
7284 (cortex_a15_alu_shift): Likewise.
7285 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB
7286 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
7287 changes.
006bd006
SN
7288 (cortex_a5_alu_shift): Likewise.
7289 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
7290 changes.
7291 (cortex_a53_alu_shift): Likewise.
7292 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
7293 changes.
7294 (cortex_a7_alu_reg): Likewise.
7295 (cortex_a7_alu_shift): Likewise.
371e77e3
UB
7296 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
7297 changes.
006bd006
SN
7298 (cortex_a8_alu_shift): Likewise.
7299 (cortex_a8_alu_shift_reg): Likewise.
7300 (cortex_a8_mov): Likewise.
7301 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7302 (cortex_a9_dp_shift): Likewise.
371e77e3
UB
7303 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
7304 changes.
7305 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
7306 changes.
006bd006
SN
7307 (cortex_r4_mov): Likewise.
7308 (cortex_r4_alu_shift): Likewise.
7309 (cortex_r4_alu_shift_reg): Likewise.
7310 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7311 (526_alu_shift_op): Likewise.
7312 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7313 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7314 (626te_alu_shift_op): Likewise.
7315 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
7316 (726te_alu_op): Likewise.
7317 (726te_alu_shift_op): Likewise.
7318 (726te_alu_shift_reg_op): Likewise.
7319 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7320 (mp626_alu_shift_op): Likewise.
7321 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
7322 (pj4_alu_e1_conds): Likewise.
7323 (pj4_alu): Likewise.
7324 (pj4_alu_conds): Likewise.
7325 (pj4_shift): Likewise.
7326 (pj4_shift_conds): Likewise.
7327 (pj4_alu_shift): Likewise.
7328 (pj4_alu_shift_conds): Likewise.
371e77e3
UB
7329 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
7330 changes.
006bd006
SN
7331 (cortexa7_older_only): Likewise.
7332 (cortexa7_younger): Likewise.
7333
651df1b2
DM
73342013-07-18 David Malcolm <dmalcolm@redhat.com>
7335
7336 * ipa-pure-const.c (generate_summary): Rename to...
7337 (pure_const_generate_summary): ... this.
7338
ea717bd7
IS
73392013-07-17 Iain Sandoe <iain@codesourcery.com>
7340
7341 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
7342
a1259a13
YR
73432013-07-17 Yvan Roux <yvan.roux@linaro.org>
7344
7345 PR target/57909
7346 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
7347 usage in HI mode.
7348
167f68ed
AK
73492013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7350
7351 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
7352 enabled without -march=zEC12.
7353 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
7354 flags to be set.
7355
f900a982
MR
73562013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7357
7358 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
7359 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
7360 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
7361 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
7362 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
7363 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
7364 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
7365 ISA_HAS_FP4.
7366 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
7367 and ISA_HAS_NMADD3_NMSUB3.
7368 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
7369 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
7370 (nmsub4<mode>, nmsub3<mode>): Likewise.
7371 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
7372
d4ed27eb
MR
73732013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7374
7375 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
7376 TARGET_MIPS5400 checking.
7377
7a5add18
PB
73782013-07-16 Jakub Jelinek <jakub@redhat.com>
7379 Peter Bergner <bergner@vnet.ibm.com>
7380
7381 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
7382 registers in the comment.
7383 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
7384 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
7385 rather than FIRST_PSEUDO_REGISTERS.
7386
73872013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB
7388
7389 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
7390 enable extra ISA flags with TARGET_HTM.
7391
d2ab0929
MR
73922013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
7393
7394 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
7395 Fix comment typos.
7396
839c74bc
CH
73972013-07-15 Cong Hou <congh@google.com>
7398
7399 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
7400 in compare function for sorting.
7401
7a5add18 74022013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB
7403
7404 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
7405 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
7406 * config/rs6000/rs6000.opt: Add -mhtm option.
7407 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
7408 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
7409 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
7410 __HTM__ if the HTM instructions are available.
7411 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
7412 htm_spr_reg_operand): New define_predicates.
7413 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
7414 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
7415 Include htm.md.
7416 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
7417 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
7418 HTM builtin functions.
7419 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
7420 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
7421 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
7422 (rs6000_builtin_mask_calculate): Likewise.
7423 (rs6000_option_override_internal): Likewise.
7424 (bdesc_htm): Add new HTM builtin support.
7425 (htm_spr_num): New function.
7426 (htm_spr_regno): Likewise.
7427 (rs6000_htm_spr_icode): Likewise.
7428 (htm_expand_builtin): Likewise.
7429 (htm_init_builtins): Likewise.
7430 (rs6000_expand_builtin): Add support for HTM builtin functions.
7431 (rs6000_init_builtins): Likewise.
371e77e3
UB
7432 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
7433 option.
0258b6e4
PB
7434 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
7435 (TARGET_HTM, MASK_HTM): Define macros.
7436 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
7437 (FIXED_REGISTERS): Likewise.
7438 (CALL_USED_REGISTERS): Likewise.
7439 (CALL_REALLY_USED_REGISTERS): Likewise.
7440 (REG_ALLOC_ORDER): Likewise.
7441 (enum reg_class): Likewise.
7442 (REG_CLASS_NAMES): Likewise.
7443 (REG_CLASS_CONTENTS): Likewise.
7444 (REGISTER_NAMES): Likewise.
7445 (ADDITIONAL_REGISTER_NAMES): Likewise.
7446 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
7447 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
7448 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
7449 * config/rs6000/htm.md: New file.
7450 * config/rs6000/htmintrin.h: New file.
7451 * config/rs6000/htmxlintrin.h: New file.
7452
87dd8ab0
MS
74532013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
7454
7455 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
7456 Define SYMBOL_TINY_GOT, update comment.
7457 * config/aarch64/aarch64.c
7458 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
7459 (aarch64_expand_mov_immediate): Likewise.
7460 (aarch64_print_operand): Likewise.
7461 (aarch64_classify_symbol): Likewise.
7462 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
7463 (ldr_got_tiny): Define.
7464
52d676b6
TG
74652013-07-13 Tobias Grosser <tobias@grosser.es>
7466
7467 PR tree-optimization/54094
7468 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
7469 scheduling dimension for the parallelism check from the polyhedral
7470 information in the AST.
7471 * graphite-dependences.c (carries_deps): Do not assume the schedule is
7472 in 2D + 1 form.
7473
286e8fc1
JM
74742013-07-13 Jason Merrill <jason@redhat.com>
7475
7476 * print-tree.c (debug_vec_tree): Use debug_raw.
7477 (debug_raw (vec<tree, va_gc> &)): New.
7478 (debug_raw (vec<tree, va_gc> *)): New.
7479 * tree.h: Declare them.
7480
e43257e8
BC
74812013-07-13 Bin Cheng <bin.cheng@arm.com>
7482
7483 * ifcvt.c (ifcvt_after_combine): New static variable.
7484 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
7485 for size.
7486 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
7487 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
7488 rest_of_handle_if_after_reload): Pass new argument for if_convert.
7489
da582d46
MR
74902013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
7491
7492 * config/mips/mips.c (mips_expand_call): Remove empty statement.
7493
18b0ea8f
MM
74942013-07-12 Michael Matz <matz@suse.de>
7495
7496 PR middle-end/55771
7497 * convert.c (convert_to_real): Reject non-float inner types.
7498
e4f0f84d
TB
74992013-07-12 Tejas Belagod <tejas.belagod@arm.com>
7500
7501 * config/aarch64/aarch64-protos.h
7502 (aarch64_simd_immediate_valid_for_move): Remove.
7503 * config/aarch64/aarch64.c (simd_immediate_info): New member.
7504 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
7505 cases.
7506 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
7507
76500022
SE
75082013-07-11 Steve Ellcey <sellcey@mips.com>
7509
7510 * config/mips/mips.c (mips_conditional_register_usage): Do not
7511 use t[0-7] registers in MIPS16 mode when optimizing for size.
7512
78039734
ST
75132013-07-11 Sriraman Tallam <tmsriram@google.com>
7514
7515 * config/i386/i386.c (dispatch_function_versions): Fix array
7516 indexing of function_version_info to match actual_versions.
7517
7b24b675
TJ
75182013-07-11 Teresa Johnson <tejohnson@google.com>
7519
7520 * vec.h (struct va_gc): Move release out-of-line.
7521 (va_gc::release): Call ggc_free on released vec.
7522
85d4cbb8
UW
75232013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
7524
7525 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
7526 Require GOT register as additional operand in UNSPEC.
7527 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
7528 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7529 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
7530 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
7531 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
7532 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
7533 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
7534
52b3f9e2
GJL
75352013-07-11 Georg-Johann Lay <avr@gjlay.de>
7536
7537 PR target/57631
7538 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
7539 name seen by assembler/linker rather if available.
7540
3f3e266e
AS
75412013-07-11 Andreas Schwab <schwab@suse.de>
7542
7543 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
7544
21231ca6
VM
75452013-07-10 Vladimir Makarov <vmakarov@redhat.com>
7546
371e77e3 7547 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 7548
d331d374
JM
75492013-07-10 Joseph Myers <joseph@codesourcery.com>
7550
673c2f63
JM
7551 * doc/tm.texi.in: Move hook documentation to ....
7552 * target.def: ... here.
7553
d331d374
JM
7554 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
7555 text on @hook line.
7556 * doc/tm.texi: Regenerate.
7557
dfeadaa0
PC
75582013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
7559
7560 PR c++/57869
7561 * doc/invoke.texi: Document -Wconditionally-supported.
7562
bb6f2bac
GJL
75632013-07-10 Georg-Johann Lay <avr@gjlay.de>
7564
7565 PR target/57844
7566 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
7567 of my_fp.
7568
5219b232
GJL
75692013-07-10 Georg-Johann Lay <avr@gjlay.de>
7570
7571 PR target/57506
7572 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
7573 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
7574 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
7575 Remove duplicate devices.
7576 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
7577 * config/avr/t-multilib: Regenerate.
7578 * config/avr/avr-tables.opt: Regenerate.
7579 * doc/avr-mmcu.texi: Regenerate.
7580
283045c0
GJL
75812013-07-10 Georg-Johann Lay <avr@gjlay.de>
7582
7583 PR target/56987
7584 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
7585
ceb2d59f 75862013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 7587
ceb2d59f
GS
7588 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
7589 the cost of MULT when optimizing for size.
7590
055e0a99
JBG
75912013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7592
7593 * config/cr16/cr16-protos.h: Don't include target.h.
7594
ef43e868
JM
75952013-07-09 Joseph Myers <joseph@codesourcery.com>
7596
7597 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
7598 adjust register size for TDmode and TFmode for VSX registers.
7599
4a283fdf
KT
76002013-07-08 Kai Tietz <ktietz@redhat.com>
7601
7602 PR target/56892
7603 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
7604 hook_bool_const_tree_true.
7605
2cf4c39e
AK
76062013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7607
7608 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
7609 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
7610 * config/s390/s390.md: Define FPR*_REGNUM constants.
7611 Fix FPR2_REGNUM constant (18 -> 17).
7612 ("*trunc<BFP:mode><DFP_ALL:mode>2")
7613 ("*trunc<DFP_ALL:mode><BFP:mode>2")
7614 ("trunc<BFP:mode><DFP_ALL:mode>2")
7615 ("trunc<DFP_ALL:mode><BFP:mode>2")
7616 ("*extend<BFP:mode><DFP_ALL:mode>2")
7617 ("*extend<DFP_ALL:mode><BFP:mode>2")
7618 ("extend<BFP:mode><DFP_ALL:mode>2")
7619 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
7620 FPR4_REGNUM.
7621
e6ac0270 76222013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 7623
6a2fa4b2 7624 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 7625
b89b22fc
AK
76262013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7627
7628 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
7629 and cfun_fpr_bit_p to cfun_fpr_save_p.
7630 (s390_frame_area, s390_register_info, s390_frame_info)
7631 (s390_emit_prologue, s390_emit_epilogue)
7632 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
7633 register numbers.
7634 * config/s390/s390.h: Define *_REGNUM macros for floating point
7635 register numbers.
7636
07f398aa
EB
76372013-07-08 Eric Botcazou <ebotcazou@adacore.com>
7638
7639 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
7640
75fefa91
PCC
76412013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
7642
7643 PR rtl-optimization/57786
7644 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
7645 and break out of the loop when it is set to false.
7646
046f1eee
JJ
76472013-07-08 Jakub Jelinek <jakub@redhat.com>
7648
8140c065
JJ
7649 PR target/57819
7650 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
7651 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
7652 (const_int 63)) 0)).
7653 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
7654 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
7655 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
7656
046f1eee
JJ
7657 PR rtl-optimization/57829
7658 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
7659 mask bits outside of mode are just sign-extension from mode to HWI.
7660
930b700b
MZ
76612013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
7662
7663 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
7664 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
7665 adjust_address instead of change_address to keep info about alignment.
7666 (emit_strmov): Remove.
7667 (emit_memmov): New function.
7668 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
7669 (expand_movmem_epilogue): Likewise and return updated rtx for
7670 destination.
7671 (expand_constant_movmem_prologue): Likewise and return updated rtx for
7672 destination and source.
7673 (decide_alignment): Refactor, handle vector_loop.
7674 (ix86_expand_movmem): Likewise.
7675 (ix86_expand_setmem): Likewise.
7676 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
7677
7b9d1bd8
UB
76782013-07-07 Uros Bizjak <ubizjak@gmail.com>
7679
7680 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
7681 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
7682
eabb5f48
UB
76832013-07-06 Uros Bizjak <ubizjak@gmail.com>
7684
7685 * config/i386/sse.md (sse_movlhps): Change alternative 3
7686 of operand 2 to "m".
7687
76882013-07-06 Uros Bizjak <ubizjak@gmail.com>
7689
7690 PR target/57807
7691 * config/i386/sse.md (iptr): New mode attribute.
7692 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
7693 (<sse>_vm<plusminus_insn><mode>3): Ditto.
7694 (<sse>_vmmul<mode>3): Ditto.
7695 (<sse>_vmdiv<mode>3): Ditto.
7696 (sse_vmrcpv4sf2): Ditto.
7697 (<sse>_vmsqrt<mode>2): Ditto.
7698 (sse_vmrsqrtv4sf2): Ditto.
7699 (<sse>_vm<code><mode>3): Ditto.
7700 (avx_vmcmp<mode>3): Ditto.
7701 (<sse>_vmmaskcmp<mode>3): Ditto.
7702 (<sse>_comi): Ditto.
7703 (<sse>_ucomi): Ditto.
7704 (*xop_vmfrcz_<mode>): Ditto.
7705 (*fmai_fmadd_<mode>): Ditto.
7706 (*fmai_fmsub_<mode>): Ditto.
7707 (*fmai_fnmadd_<mode>): Ditto.
7708 (*fmai_fnmsub_<mode>): Ditto.
7709 (*fma4i_vmfmadd_<mode>): Ditto.
7710 (*fma4i_vmfmsub_<mode>): Ditto.
7711 (*fma4i_vmfnmadd_<mode>): Ditto.
7712 (*fma4i_vmfnmsub_<mode>): Ditto.
7713 (*xop_vmfrcz_<mode>): Ditto.
7714 (sse_cvtps2pi): Ditto.
7715 (sse_cvttps2pi): Ditto.
7716 (sse_cvtss2si): Ditto.
7717 (sse_cvtss2si_2): Ditto.
7718 (sse_cvtss2siq_2): Ditto.
7719 (sse_cvttss2si): Ditto.
7720 (sse_cvttss2siq): Ditto.
7721 (sse_cvtsd2si): Ditto.
7722 (sse_cvtsd2si_2): Ditto.
7723 (sse_cvtsd2siq_2): Ditto.
7724 (sse_cvttsd2si): Ditto.
7725 (sse_cvttsd2siq): Ditto.
7726 (sse_cvtsd2ss): Ditto.
7727 (sse_cvtss2sd): Ditto.
7728 (avx2_pbroadcast<mode>): Ditto.
7729 (avx2_pbroadcast<mode>_1): Ditto.
7730 (*avx_vperm_broadcast_v4sf): Ditto.
7731
7732 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
7733 (sse_movlhps): Ditto.
7734 (sse_storehps): Ditto.
7735 (sse_loadhps): Ditto.
7736 (sse_storelps): Ditto.
7737 (sse_loadlps): Ditto.
7738 (*vec_concatv4sf): Ditto.
7739 (*vec_interleave_highv2df): Ditto.
7740 (*vec_interleave_lowv2df): Ditto.
7741 (*vec_extractv2df_1_sse): Ditto.
7742 (*vec_extractv2df_0_sse): Ditto.
7743 (sse2_storelpd): Ditto.
7744 (sse2_loadlpd): Ditto.
7745 (sse2_movsd): Ditto.
7746 (*vec_concatv4si): Ditto.
7747 (vec_concatv2di): Ditto.
7748
7749 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
7750 for Intel asm dialect.
7751 (mmx_punpcklwd): Ditto.
7752 (mmx_punpckldq): Ditto.
7753
7754 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
7755 for intel assembler dialect.
7756
1f6eac90
JJ
77572013-07-06 Jakub Jelinek <jakub@redhat.com>
7758
7759 PR target/29776
7760 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
7761 for BUILT_IN_C{LZ,LRSB}*.
7762 * tree.h (CASE_INT_FN): Add FN##IMAX case.
7763 * tree-vrp.c (extract_range_basic): Handle
7764 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
7765 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
7766 fall thru to code calling set_value*.
7767 * builtins.c (expand_builtin): Remove *IMAX cases.
7768 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
7769 if width is bigger than 2*HWI.
7770
2b778c9d
VM
77712013-07-05 Vladimir Makarov <vmakarov@redhat.com>
7772
7773 PR rtl-optimization/55342
7774 * lra-int.h (lra_subreg_reload_pseudos): New.
7775 * lra.c: Add undoing optional reloads to the block diagram.
7776 (lra_subreg_reload_pseudos): New.
7777 (lra_optional_reload_pseudos): Change comments.
7778 (lra): Init and clear lra_subreg_reload_pseudos. Clear
7779 lra_optional_reload_pseudos after undo transformations.
7780 * lra-assigns.c (pseudo_prefix_title): New.
7781 (lra_setup_reg_renumber): Use it.
7782 (spill_for): Ditto. Check subreg reload pseudos too.
7783 (assign_by_spills): Consider subreg reload pseudos too.
7784 * lra-constraints.c (simplify_operand_subreg): Use
7785 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
7786 (curr_insn_transform): Recognize and do optional reloads.
7787 (undo_optional_reloads): New.
7788 (lra_undo_inheritance): Call undo_optional_reloads.
7789
207156e4
TQ
77902013-07-05 Thomas Quinot <quinot@adacore.com>
7791
7792 * tree-complex.c (expand_complex_operations_1): Fix typo.
7793
8990e73a
TB
77942013-07-04 Tejas Belagod <tejas.belagod@arm.com>
7795
7796 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
7797 (tune_params): New member 'const vec_costs'.
7798 * config/aarch64/aarch64.c (generic_vector_cost): New.
7799 (generic_tunings): New member 'generic_vector_cost'.
7800 (aarch64_builtin_vectorization_cost): New.
7801 (aarch64_add_stmt_cost): New.
7802 (TARGET_VECTORIZE_ADD_STMT_COST): New.
7803 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
7804
f7bc421d
JJ
78052013-07-03 Jakub Jelinek <jakub@redhat.com>
7806
7807 PR target/57777
7808 * config/i386/predicates.md (vsib_address_operand): Disallow
7809 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
7810
e90d1568
HPN
78112013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
7812
7813 PR middle-end/55030
7814 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
7815 expand_builtin_setjmp_receiver.
7816 (expand_label): Adjust, call expand_builtin_setjmp_receiver
7817 with NULL for the label parameter.
7818 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
7819 the frame-pointer. Adjust comments.
7820 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
7821 only if LABEL is non-NULL.
7822
cbd7413b
YZ
78232013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
7824
7825 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
7826 (ARM_ABI_AAPCS64): Ditto.
7827 (arm_abi): Ditto.
7828 (ARM_DEFAULT_ABI): Ditto.
7829
dec11868
JG
78302013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
7831
7832 * config/aarch64/aarch64-builtins.c
7833 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
7834 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
7835 (st1): Likewise.
7836 * config/aarch64/aarch64-simd.md
7837 (aarch64_ld1<VALL:mode>): New.
7838 (aarch64_st1<VALL:mode>): Likewise.
7839 * config/aarch64/arm_neon.h
7840 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
7841
786fb9b6
ST
78422013-07-02 Sriraman Tallam <tmsriram@google.com>
7843
0ba6aff3
UB
7844 * config/i386/i386.c (gate_insert_vzeroupper): Check if
7845 target ISA is AVX.
786fb9b6
ST
7846 (ix86_option_override_internal):Turn on all -mavx target flags by
7847 default as they are dependent on AVX anyway.
7848
4b043553
CC
78492013-07-02 Cary Coutant <ccoutant@google.com>
7850
7851 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
7852 deterministic hash.
7853 (loc_checksum_ordered): Likewise.
7854 (hash_loc_operands): Remove inline keyword.
7855
c4e87a13
JJ
78562013-07-02 Jakub Jelinek <jakub@redhat.com>
7857
7858 PR tree-optimization/57741
7859 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
7860 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
7861 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
7862 Allow REAL_CST step_exprs if flag_associative_math.
7863 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
7864
26366d28
IB
78652013-07-02 Ian Bolton <ian.bolton@arm.com>
7866
0ba6aff3 7867 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 7868
2879bb2b
IB
78692013-07-02 Ian Bolton <ian.bolton@arm.com>
7870
7871 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
7872
afc5e8a6
KT
78732013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7874
7875 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
7876 encoding.
7877 (iorsi3_insn): Likewise.
7878 (arm_xorsi3): Likewise.
7879
31f8442b
SN
78802013-07-01 Sofiane Naci <sofiane.naci@arm.com>
7881
7882 * arm.md (attribute "wtype"): Delete. Move attribute values from here
7883 to ...
7884 (attribute "type"): ... here, and prefix with "wmmx_".
7885 (attribute "core_cycles"): Update for attribute changes.
7886 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
7887 (tbcstv4hi): Likewise.
7888 (tbcstv2si): Likewise.
7889 (iwmmxt_iordi3): Likewise.
7890 (iwmmxt_xordi3): Likewise.
7891 (iwmmxt_anddi3): Likewise.
7892 (iwmmxt_nanddi3): Likewise.
7893 (iwmmxt_arm_movdi): Likewise.
7894 (iwmmxt_movsi_insn): Likewise.
7895 (mov<mode>_internal): Likewise.
7896 (and<mode>3_iwmmxt): Likewise.
7897 (ior<mode>3_iwmmxt): Likewise.
7898 (xor<mode>3_iwmmxt): Likewise.
7899 (add<mode>3_iwmmxt): Likewise.
7900 (ssaddv8qi3): Likewise.
7901 (ssaddv4hi3): Likewise.
7902 (ssaddv2si3): Likewise.
7903 (usaddv8qi3): Likewise.
7904 (usaddv4hi3): Likewise.
7905 (usaddv2si3): Likewise.
7906 (sub<mode>3_iwmmxt): Likewise.
7907 (sssubv8qi3): Likewise.
7908 (sssubv4hi3): Likewise.
7909 (sssubv2si3): Likewise.
7910 (ussubv8qi3): Likewise.
7911 (ussubv4hi3): Likewise.
7912 (ussubv2si3): Likewise.
7913 (mulv4hi3_iwmmxt): Likewise.
7914 (smulv4hi3_highpart): Likewise.
7915 (umulv4hi3_highpart): Likewise.
7916 (iwmmxt_wmacs): Likewise.
7917 (iwmmxt_wmacsz): Likewise.
7918 (iwmmxt_wmacu): Likewise.
7919 (iwmmxt_wmacuz): Likewise.
7920 (iwmmxt_clrdi): Likewise.
7921 (iwmmxt_clrv8qi): Likewise.
7922 (iwmmxt_clr4hi): Likewise.
7923 (iwmmxt_clr2si): Likewise.
7924 (iwmmxt_uavgrndv8qi3): Likewise.
7925 (iwmmxt_uavgrndv4hi3): Likewise.
7926 (iwmmxt_uavgv8qi3): Likewise.
7927 (iwmmxt_uavgv4hi3): Likewise.
7928 (iwmmxt_tinsrb): Likewise.
7929 (iwmmxt_tinsrh): Likewise.
7930 (iwmmxt_tinsrw): Likewise.
7931 (iwmmxt_textrmub): Likewise.
7932 (iwmmxt_textrmsb): Likewise.
7933 (iwmmxt_textrmuh): Likewise.
7934 (iwmmxt_textrmsh): Likewise.
7935 (iwmmxt_textrmw): Likewise.
7936 (iwmxxt_wshufh): Likewise.
7937 (eqv8qi3): Likewise.
7938 (eqv4hi3): Likewise.
7939 (eqv2si3): Likewise.
7940 (gtuv8qi3): Likewise.
7941 (gtuv4hi3): Likewise.
7942 (gtuv2si3): Likewise.
7943 (gtv8qi3): Likewise.
7944 (gtv4hi3): Likewise.
7945 (gtv2si3): Likewise.
7946 (smax<mode>3_iwmmxt): Likewise.
7947 (umax<mode>3_iwmmxt): Likewise.
7948 (smin<mode>3_iwmmxt): Likewise.
7949 (umin<mode>3_iwmmxt): Likewise.
7950 (iwmmxt_wpackhss): Likewise.
7951 (iwmmxt_wpackwss): Likewise.
7952 (iwmmxt_wpackdss): Likewise.
7953 (iwmmxt_wpackhus): Likewise.
7954 (iwmmxt_wpackwus): Likewise.
7955 (iwmmxt_wpackdus): Likewise.
7956 (iwmmxt_wunpckihb): Likewise.
7957 (iwmmxt_wunpckihh): Likewise.
7958 (iwmmxt_wunpckihw): Likewise.
7959 (iwmmxt_wunpckilb): Likewise.
7960 (iwmmxt_wunpckilh): Likewise.
7961 (iwmmxt_wunpckilw): Likewise.
7962 (iwmmxt_wunpckehub): Likewise.
7963 (iwmmxt_wunpckehuh): Likewise.
7964 (iwmmxt_wunpckehuw): Likewise.
7965 (iwmmxt_wunpckehsb): Likewise.
7966 (iwmmxt_wunpckehsh): Likewise.
7967 (iwmmxt_wunpckehsw): Likewise.
7968 (iwmmxt_wunpckelub): Likewise.
7969 (iwmmxt_wunpckeluh): Likewise.
7970 (iwmmxt_wunpckeluw): Likewise.
7971 (iwmmxt_wunpckelsb): Likewise.
7972 (iwmmxt_wunpckelsh): Likewise.
7973 (iwmmxt_wunpckelsw): Likewise.
7974 (ror<mode>3): Likewise.
7975 (ashr<mode>3_iwmmxt): Likewise.
7976 (lshr<mode>3_iwmmxt): Likewise.
7977 (ashl<mode>3_iwmmxt): Likewise.
7978 (ror<mode>3_di): Likewise.
7979 (ashr<mode>3_di): Likewise.
7980 (lshr<mode>3_di): Likewise.
7981 (ashl<mode>3_di): Likewise.
7982 (iwmmxt_wmadds): Likewise.
7983 (iwmmxt_wmaddu): Likewise.
7984 (iwmmxt_tmia): Likewise.
7985 (iwmmxt_tmiaph): Likewise.
7986 (iwmmxt_tmiabb): Likewise.
7987 (iwmmxt_tmiatb): Likewise.
7988 (iwmmxt_tmiabt): Likewise.
7989 (iwmmxt_tmiatt): Likewise.
7990 (iwmmxt_tmovmskb): Likewise.
7991 (iwmmxt_tmovmskh): Likewise.
7992 (iwmmxt_tmovmskw): Likewise.
7993 (iwmmxt_waccb): Likewise.
7994 (iwmmxt_wacch): Likewise.
7995 (iwmmxt_waccw): Likewise.
7996 (iwmmxt_waligni): Likewise.
7997 (iwmmxt_walignr): Likewise.
7998 (iwmmxt_walignr0): Likewise.
7999 (iwmmxt_walignr1): Likewise.
8000 (iwmmxt_walignr2): Likewise.
8001 (iwmmxt_walignr3): Likewise.
8002 (iwmmxt_wsadb): Likewise.
8003 (iwmmxt_wsadh): Likewise.
8004 (iwmmxt_wsadbz): Likewise.
8005 (iwmmxt_wsadhz): Likewise.
8006 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
8007 (iwmmxt_wabsdiffb): Likewise.
8008 (iwmmxt_wabsdiffh): Likewise.
8009 (iwmmxt_wabsdiffw): Likewise.
8010 (iwmmxt_waddsubhx): Likewise
8011 (iwmmxt_wsubaddhx): Likewise.
8012 (addc<mode>3): Likewise.
8013 (iwmmxt_avg4): Likewise.
8014 (iwmmxt_avg4r): Likewise.
8015 (iwmmxt_wmaddsx): Likewise.
8016 (iwmmxt_wmaddux): Likewise.
8017 (iwmmxt_wmaddsn): Likewise.
8018 (iwmmxt_wmaddun): Likewise.
8019 (iwmmxt_wmulwsm): Likewise.
8020 (iwmmxt_wmulwum): Likewise.
8021 (iwmmxt_wmulsmr): Likewise.
8022 (iwmmxt_wmulumr): Likewise.
8023 (iwmmxt_wmulwsmr): Likewise.
8024 (iwmmxt_wmulwumr): Likewise.
8025 (iwmmxt_wmulwl): Likewise.
8026 (iwmmxt_wqmulm): Likewise.
8027 (iwmmxt_wqmulwm): Likewise.
8028 (iwmmxt_wqmulmr): Likewise.
8029 (iwmmxt_wqmulwmr): Likewise.
8030 (iwmmxt_waddbhusm): Likewise.
8031 (iwmmxt_waddbhusl): Likewise.
8032 (iwmmxt_wqmiabb): Likewise.
8033 (iwmmxt_wqmiabt): Likewise.
8034 (iwmmxt_wqmiatb): Likewise.
8035 (iwmmxt_wqmiatt): Likewise.
8036 (iwmmxt_wqmiabbn): Likewise.
8037 (iwmmxt_wqmiabtn): Likewise.
8038 (iwmmxt_wqmiatbn): Likewise.
8039 (iwmmxt_wqmiattn): Likewise.
8040 (iwmmxt_wmiabb): Likewise.
8041 (iwmmxt_wmiabt): Likewise.
8042 (iwmmxt_wmiatb): Likewise.
8043 (iwmmxt_wmiatt): Likewise.
8044 (iwmmxt_wmiabbn): Likewise.
8045 (iwmmxt_wmiabtn): Likewise.
8046 (iwmmxt_wmiatbn): Likewise.
8047 (iwmmxt_wmiattn): Likewise.
8048 (iwmmxt_wmiawbb): Likewise.
8049 (iwmmxt_wmiawbt): Likewise.
8050 (iwmmxt_wmiawtb): Likewise.
8051 (iwmmxt_wmiawtt): Likewise.
8052 (iwmmxt_wmiawbbn): Likewise.
8053 (iwmmxt_wmiawbtn): Likewise.
8054 (iwmmxt_wmiawtbn): Likewise.
8055 (iwmmxt_wmiawttn): Likewise.
8056 (iwmmxt_wmerge): Likewise.
8057 (iwmmxt_tandc<mode>3): Likewise.
8058 (iwmmxt_torc<mode>3): Likewise.
8059 (iwmmxt_torvsc<mode>3): Likewise.
8060 (iwmmxt_textrc<mode>3): Likewise.
8061 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
8062 (wmmxt_pack): Likewise.
8063 (wmmxt_mult_c1): Likewise.
8064 (wmmxt_mult_c2): Likewise.
8065 (wmmxt_alu_c1): Likewise.
8066 (wmmxt_alu_c2): Likewise.
8067 (wmmxt_alu_c3): Likewise.
8068 (wmmxt_transfer_c1): Likewise.
8069 (wmmxt_transfer_c2): Likewise.
8070 (wmmxt_transfer_c3): Likewise.
8071 (marvell_f_iwmmxt_wstr): Likewise.
8072 (marvell_f_iwmmxt_wldr): Likewise.
8073
bdb7bf8a
YZ
80742013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
8075
0ba6aff3 8076 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 8077
e6631d38
VM
80782013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8079
8080 Revert:
8081 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8082 * lra-constraints.c (need_for_split_p): Check call used hard regs
8083 living through calls.
8084
8085 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
8086 call used regs for call insn.
8087
53c1275b
JJ
80882013-06-28 Jakub Jelinek <jakub@redhat.com>
8089
8090 PR target/57736
0ba6aff3
UB
8091 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
8092 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 8093
1141ed3f
BI
80942013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
8095
8096 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 8097
e32e4c4a
VM
80982013-06-28 Vladimir Makarov <vmakarov@redhat.com>
8099
8100 * lra-constraints.c (need_for_split_p): Check call used hard regs
8101 living through calls.
8102
f161bfd3
MM
81032013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
8104
8105 PR target/57744
8106 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
8107 to tie with any other modes. Eliminate Altivec vector mode tests,
8108 since these are a subset of ALTIVEC or VSX vector modes. Simplify
8109 code, to return 0 if testing MODE2 for a condition, if we've
8110 already tested MODE1 for the same condition.
8111
7eda14e1
MS
81122013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8113
8114 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
8115 layout.
8116
c822f852
MS
81172013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8118
8119 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
8120 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
8121
da4f13a4
MS
81222013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
8123
0ba6aff3
UB
8124 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
8125 Define.
da4f13a4
MS
8126 (aarch64_symbolic_constant_p): Remove.
8127 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
8128 static. Fix line length and white space.
8129 (aarch64_symbolic_constant_p): Remove.
8130 * config/aarch64/predicates.md (aarch64_valid_symref):
8131 Use aarch64_classify_symbol_expression.
8132
956a95a5
KT
81332013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8134
8135 * config/arm/constraints.md (Ts): New constraint.
8136 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
8137 16-bit encodings.
8138 (compare_scc): Use "Ts" constraint for operand 0.
8139 (ior_scc_scc): Likewise.
8140 (and_scc_scc): Likewise.
8141 (and_scc_scc_nodom): Likewise.
8142 (ior_scc_scc_cmp): Likewise for operand 7.
8143 (and_scc_scc_cmp): Likewise.
8144 * config/arm/thumb2.md (thumb2_movsi_insn):
8145 Add alternatives for 16-bit encodings.
8146 (thumb2_movhi_insn): Likewise.
8147 (thumb2_movsicc_insn): Likewise.
8148 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
8149 (thumb2_negscc): Use "Ts" constraint.
8150 Move mvn instruction outside cond_exec block.
8151 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
8152 for 16-bit encodings.
8153
31dfce10
KT
81542013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8155
8156 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
8157 encoding.
8158 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
8159 (mulsi3subsi): Likewise.
8160 (mulsidi3adddi): Likewise.
8161 (mulsidi3_v6): Likewise.
8162 (umulsidi3_v6): Likewise.
8163 (umulsidi3adddi_v6): Likewise.
8164 (smulsi3_highpart_v6): Likewise.
8165 (umulsi3_highpart_v6): Likewise.
8166 (mulhisi3tb): Likewise.
8167 (mulhisi3bt): Likewise.
8168 (mulhisi3tt): Likewise.
8169 (maddhisi4): Likewise.
8170 (maddhisi4tb): Likewise.
8171 (maddhisi4tt): Likewise.
8172 (maddhidi4): Likewise.
8173 (maddhidi4tb): Likewise.
8174 (maddhidi4tt): Likewise.
8175 (zeroextractsi_compare0_scratch): Likewise.
8176 (insv_zero): Likewise.
8177 (insv_t2): Likewise.
8178 (anddi_notzesidi_di): Likewise.
8179 (anddi_notsesidi_di): Likewise.
8180 (andsi_notsi_si): Likewise.
8181 (iordi_zesidi_di): Likewise.
8182 (xordi_zesidi_di): Likewise.
8183 (andsi_iorsi3_notsi): Likewise.
8184 (smax_0): Likewise.
8185 (smax_m1): Likewise.
8186 (smin_0): Likewise.
8187 (not_shiftsi): Likewise.
8188 (unaligned_loadsi): Likewise.
8189 (unaligned_loadhis): Likewise.
8190 (unaligned_loadhiu): Likewise.
8191 (unaligned_storesi): Likewise.
8192 (unaligned_storehi): Likewise.
8193 (extv_reg): Likewise.
8194 (extzv_t2): Likewise.
8195 (divsi3): Likewise.
8196 (udivsi3): Likewise.
8197 (arm_zero_extendhisi2addsi): Likewise.
8198 (arm_zero_extendqisi2addsi): Likewise.
8199 (compareqi_eq0): Likewise.
8200 (arm_extendhisi2_v6): Likewise.
8201 (arm_extendqisi2addsi): Likewise.
8202 (arm_movt): Likewise.
8203 (thumb2_ldrd): Likewise.
8204 (thumb2_ldrd_base): Likewise.
8205 (thumb2_ldrd_base_neg): Likewise.
8206 (thumb2_strd): Likewise.
8207 (thumb2_strd_base): Likewise.
8208 (thumb2_strd_base_neg): Likewise.
8209 (arm_negsi2): Add alternative for 16-bit encoding.
8210 (arm_one_cmplsi2): Likewise.
8211
1572e697
KT
82122013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8213
8214 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
8215 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
8216 (movdfcc): Likewise.
8217 * config/arm/vfp.md (*thumb2_movsf_vfp):
8218 Disable predication for arm_restrict_it.
8219 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
8220 (*thumb2_movdfcc_vfp): Likewise.
8221 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
8222 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
8223 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
8224 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
8225 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
8226 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
8227 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
8228 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
8229 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
8230 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
8231 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
8232 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
8233 Disable predication for arm_restrict_it.
8234
ebf8f0ea
KY
82352013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
8236
8237 * config/i386/bmiintrin.h (_bextr_u32): New.
8238 (_bextr_u64): Ditto.
8239
a290fcda
RS
82402013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
8241
8242 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
8243 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
8244 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
8245 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
8246 * config/mips/n32-elf.h: ...this new file.
8247
ea169996
MG
82482013-06-27 Marc Glisse <marc.glisse@inria.fr>
8249
8250 PR target/57224
8251 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
8252 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
8253
12211b99 82542013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM
8255
8256 * config/mips/mips-tables.opt: Regenerate.
8257 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 8258 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 8259 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 8260
2444244e
JJ
82612013-06-27 Jakub Jelinek <jakub@redhat.com>
8262
cd06d2a2
JJ
8263 PR target/57623
8264 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
8265 constraints of operand 1 and 2.
8266
2444244e
JJ
8267 PR target/57623
8268 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
8269 to match RTL canonicalization. Swap predicates and
8270 constraints of operand 1 and 2.
8271
2f259720
VM
82722013-06-27 Vladimir Makarov <vmakarov@redhat.com>
8273
0ba6aff3
UB
8274 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
8275 Process OP_INOUT regs for splitting too.
2f259720 8276
fef4d2b3
JJ
82772013-06-27 Jakub Jelinek <jakub@redhat.com>
8278
8279 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
8280 decl before the loop, initialize to NULL.
8281 (vectorizable_load): Initialize ptr_incr to NULL.
8282
79ee9826
MJ
82832013-06-27 Martin Jambor <mjambor@suse.cz>
8284
8285 PR lto/57208
8286 * ipa-ref.h (ipa_maybe_record_reference): Declare.
8287 * ipa-ref.c (ipa_maybe_record_reference): New function.
8288 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
8289 * ipa-cp.c (create_specialized_node): Record potential references from
8290 aggvals.
8291 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
8292
e18b4a81
YZ
82932013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8294
8295 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
8296 parameter 'mode' of type 'enum machine_mode mode'; change to pass
8297 'mode' to force_reg.
8298 (aarch64_add_offset): Update calls to aarch64_force_temporary.
8299 (aarch64_expand_mov_immediate): Likewise.
8300
9c023bf0
YZ
83012013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
8302
8303 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
8304 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
8305
5a3fe9b6
AK
83062013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8307
8308 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
8309 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8310 (struct machine_function): Add tbegin_p.
8311 (s390_canonicalize_comparison): Fold CC mode compares to
8312 conditional jump if possible.
8313 (s390_emit_jump): Return the emitted jump.
8314 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
8315 Handle CCRAWmode compares.
8316 (s390_option_override): Default to -mhtm if available.
8317 (s390_reg_clobbered_rtx): Handle floating point regs as well.
8318 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
8319 FPRs instead of df_regs_ever_live_p.
8320 (s390_optimize_nonescaping_tx): New function.
8321 (s390_init_frame_layout): Extend clobbered_regs array to cover
8322 FPRs as well.
8323 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
8324 (s390_expand_tbegin): New function.
8325 (enum s390_builtin): New enum definition.
8326 (code_for_builtin): New array definition.
8327 (s390_init_builtins): New function.
8328 (s390_expand_builtin): New function.
8329 (TARGET_INIT_BUILTINS): Define.
8330 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 8331 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK
8332 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
8333 (s390_alc_comparison): Likewise.
8334 * config/s390/s390-modes.def: Add CCRAWmode.
8335 * config/s390/s390.h (processor_flags): Add PF_TX.
8336 (TARGET_CPU_HTM): Define macro.
8337 (TARGET_HTM): Define macro.
8338 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
8339 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
8340 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
8341 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 8342 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK
8343 (TBEGIN_MASK, TBEGINC_MASK): New constants.
8344 ("*cc_to_int"): Move up.
8345 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
8346 constants other than 0.
8347 ("*ccraw_to_int"): New insn and splitter definition.
8348 ("tbegin", "tbegin_nofloat", "tbegin_retry")
8349 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
8350 ("tx_assist"): New expander.
8351 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
8352 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
8353 * config/s390/s390.opt: Add -mhtm option.
8354 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
8355 * config/s390/htmxlintrin.h: New file.
8356 * config/s390/htmintrin.h: New file.
8357 * config/s390/s390intrin.h: New file.
8358 * doc/extend.texi: Document htm builtins.
8359 * config.gcc: Add the new header files to extra_headers.
8360
9e216629
TS
83612013-06-26 Thomas Schwinge <thomas@codesourcery.com>
8362
8363 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 8364 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 8365
fd8c65e7 83662013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
8367 Pat Haugen <pthaugen@us.ibm.com>
8368 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM
8369
8370 * config/rs6000/power8.md: New.
8371 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
8372 setting for power8 entry.
8373 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
8374 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
8375 test for Power4/Power5 only.
8376 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
8377 support.
8378 (force_new_group): Adjust comment.
8379 * config/rs6000/rs6000.md: Include power8.md.
8380
b24a2ce5
GY
83812013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
8382
8383 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
8384 * config/arm/arm-protos.h (arm_max_conditional_execute): New
8385 declaration.
8386 (tune_params): Update comment.
8387 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
8388 (arm_max_conditional_execute): New function.
8389 (thumb2_final_prescan_insn): Use max_insn_skipped and
8390 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
8391
f6fdeb62
JJ
83922013-06-25 Jakub Jelinek <jakub@redhat.com>
8393
8394 PR tree-optimization/57705
8395 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
8396 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 8397 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ
8398 (get_initial_def_for_induction): Handle SSA_NAME IV step.
8399
c13bc3d9
MJ
84002013-06-25 Martin Jambor <mjambor@suse.cz>
8401
8402 PR middle-end/57670
8403 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
8404 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
8405 calls in the dump.
8406 (ipa_note_param_call): Initialize member_ptr flag.
8407 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
8408 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
8409 (ipa_write_indirect_edge_info): Stream member_ptr flag.
8410 (ipa_read_indirect_edge_info): Likewise.
8411
8a845901
RB
84122013-06-25 Richard Biener <rguenther@suse.de>
8413
8414 PR middle-end/56977
8415 * passes.c (init_optimization_passes): Move pass_fold_builtins
8416 and pass_dce earlier with -Og.
8417
33be0bec
EB
84182013-06-25 Eric Botcazou <ebotcazou@adacore.com>
8419
8420 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
8421 <BIT_FIELD_REF>: Remove trailing TAB.
8422 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
8423 remove blank line.
8424
5fe8e757
MJ
84252013-06-24 Martin Jambor <mjambor@suse.cz>
8426
8427 PR tree-optimization/57358
8428 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
8429 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
8430 (ipa_analyze_params_uses): Generate pessimistic info when true.
8431
44a60244
MJ
84322013-06-24 Martin Jambor <mjambor@suse.cz>
8433
8434 PR tree-optimization/57539
8435 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
8436 global.inlined_to of the new node to it. All callers changed.
8437 * ipa-inline-transform.c (clone_inlined_nodes): New variable
8438 inlining_into, pass it to cgraph_clone_node.
8439 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
8440 ipa_free_edge_args_substructures.
8441 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
8442 rdesc linked list. Do not assert rdesc edges have inlined caller.
8443 Assert we have found an rdesc in the rdesc list.
8444
7c5848b8
RB
84452013-06-24 Richard Biener <rguenther@suse.de>
8446
0ba6aff3 8447 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB
8448 (pointer_set_lookup): Declare.
8449 (class pointer_map): New template class implementing a
8450 generic pointer to T map.
8451 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
8452 pointer_map<T>::contains, pointer_map<T>::insert,
8453 pointer_map<T>::traverse): New functions.
8454 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
8455 (pointer_set_lookup): New function.
8456 (pointer_set_contains): Use pointer_set_lookup.
8457 (pointer_set_insert): Likewise.
8458 (insert_aux): Remove.
8459 (struct pointer_map_t): Embed a pointer_set_t.
8460 (pointer_map_create): Adjust.
8461 (pointer_map_destroy): Likewise.
8462 (pointer_map_contains): Likewise.
8463 (pointer_map_insert): Likewise.
8464 (pointer_map_traverse): Likewise.
8465 * tree-streamer.h (struct streamer_tree_cache_d): Use a
8466 pointer_map<unsigned> instead of a pointer_map_t.
8467 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
8468 (streamer_tree_cache_lookup): Likewise.
8469 (streamer_tree_cache_create): Likewise.
8470 (streamer_tree_cache_delete): Likewise.
8471 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
8472 pointer_map<unsigned> instead of a pointer_map_t.
8473 (lto_init_tree_ref_encoder): Adjust.
8474 (lto_destroy_tree_ref_encoder): Likewise.
8475 * lto-section-out.c (lto_output_decl_index): Likewise.
8476 (lto_record_function_out_decl_state): Likewise.
8477 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
8478
3dbc97a9
RB
84792013-06-24 Richard Biener <rguenther@suse.de>
8480
8481 PR tree-optimization/57488
8482 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
8483
8345965a
AM
84842013-06-24 Alan Modra <amodra@gmail.com>
8485
8486 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
8487 (gen_easy_altivec_constant): Likewise.
8488 * config/rs6000/predicates.md (easy_vector_constant_add_self,
8489 easy_vector_constant_msb): Likewise.
8490
b7df379f
JJ
84912013-06-23 Jakub Jelinek <jakub@redhat.com>
8492
8493 PR target/57688
8494 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
8495 add missing return true.
8496
12211b99 84972013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE
8498
8499 PR target/52483
8500 * config/sh/predicates.md (general_extend_operand): Invoke
8501 general_movsrc_operand for memory operands.
8502 (general_movsrc_operand): Allow reg+reg addressing, do not use
8503 general_operand for memory operands.
8504
97db2bf7
ST
85052013-06-23 Sriraman Tallam <tmsriram@google.com>
8506
8507 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
8508 when current target options does not apply.
8509 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
8510 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB
8511 * config/i386/bmiintrin.h: Pass appropriate target
8512 attributes to header.
97db2bf7
ST
8513 * config/i386/mmintrin.h: Ditto.
8514 * config/i386/nmmintrin.h: Ditto.
8515 * config/i386/avx2intrin.h: Ditto.
8516 * config/i386/fxsrintrin.h: Ditto.
8517 * config/i386/tbmintrin.h: Ditto.
8518 * config/i386/xsaveintrin.h: Ditto.
8519 * config/i386/f16cintrin.h: Ditto.
8520 * config/i386/xtestintrin.h: Ditto.
8521 * config/i386/xsaveoptintrin.h: Ditto.
8522 * config/i386/bmi2intrin.h: Ditto.
8523 * config/i386/lzcntintrin.h: Ditto.
8524 * config/i386/smmintrin.h: Ditto.
8525 * config/i386/wmmintrin.h: Ditto.
8526 * config/i386/x86intrin.h: Remove all header include guards.
8527 * config/i386/prfchwintrin.h: Ditto.
8528 * config/i386/pmmintrin.h: Ditto.
8529 * config/i386/tmmintrin.h: Ditto.
8530 * config/i386/xmmintrin.h: Ditto.
8531 * config/i386/popcntintrin.h: Ditto.
8532 * config/i386/rdseedintrin.h: Ditto.
8533 * config/i386/ammintrin.h: Ditto.
8534 * config/i386/emmintrin.h: Ditto.
8535 * config/i386/immintrin.h: Remove all header include guards.
8536 * config/i386/fma4intrin.h: Ditto.
8537 * config/i386/lwpintrin.h: Ditto.
8538 * config/i386/xopintrin.h: Ditto.
8539 * config/i386/ia32intrin.h: Ditto.
8540 * config/i386/avxintrin.h: Ditto.
8541 * config/i386/rtmintrin.h: Ditto.
8542 * config/i386/fmaintrin.h: Ditto.
8543 * config/i386/mm3dnow.h: Ditto.
8544
495e6879
ST
85452013-06-22 Sriraman Tallam <tmsriram@google.com>
8546
8547 * common/config/i386/i386-common.c: Handle LZCNT.
8548
11e8f687
AK
85492013-06-22 Andi Kleen <ak@linux.intel.com>
8550
8551 * doc/extend.texi: Use __atomic_store_n instead of
8552 __atomic_store in HLE example.
8553
12211b99 85542013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE
8555
8556 * config/sh/sh.c: Remove <cstdlib> workaround.
8557
927f908b
AK
85582013-06-21 Andi Kleen <ak@linux.intel.com>
8559
0ba6aff3 8560 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 8561
cca410b6
AK
85622013-06-21 Andi Kleen <ak@linux.intel.com>
8563
8564 * doc/extend.texi: Document that __atomic_clear and
8565 __atomic_test_and_set should only be used with bool.
8566
d84db1ed
JH
85672013-06-20 Jan Hubicka <jh@suse.cz>
8568
8569 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
8570 types_same_for_odr.
8571 * tree.c (decls_same_for_odr): New function.
8572 (same_for_edr): New function.
8573 (types_same_for_odr): New function.
8574 (get_binfo_at_offset): Use it.
8575 * tree.h (types_same_for_odr): Declare.
8576
12211b99 85772013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE
8578 Jason Merrill <jason@redhat.com>
8579
8580 * system.h: Include <cstdlib> as well as <stdlib.h>.
8581
49b4ff63
UB
85822013-06-20 Uros Bizjak <ubizjak@gmail.com>
8583
8584 PR target/57655
8585 * config/i386/i386.c (construct_container): Report error if
8586 long double is used with disabled x87 float returns.
8587
7a89ad00
JH
85882013-06-20 Jan Hubicka <jh@suse.cz>
8589
8590 * lto-cgraph.c (input_symtab): Do not set cgraph state.
8591
12211b99 85922013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR
8593
8594 PR rtl-optimization/57425
8595 PR rtl-optimization/57569
8596 * alias.c (write_dependence_p): Remove parameters mem_mode and
8597 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
8598 Changed all callers.
8599 (canon_anti_dependence): Get comments and semantics in sync.
8600 Add parameter mem_canonicalized. Changed all callers.
8601 * rtl.h (canon_anti_dependence): Update prototype.
8602
d16e9a99
RB
86032013-06-20 Richard Biener <rguenther@suse.de>
8604
8605 * data-streamer-in.c (streamer_read_uhwi): Optimize single
8606 byte case, inline streamer_read_uchar and defer section
8607 overrun check.
8608
c3a9b91b
RB
86092013-06-20 Richard Biener <rguenther@suse.de>
8610
8611 PR tree-optimization/57584
8612 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
8613 SSA names into the expanded expression that take part in
8614 abnormal coalescing.
8615
371e77e3 86162013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS
8617
8618 * gcov.c (print_usage): Handle new option.
8619 (process_args): Ditto.
8620 (get_gcov_intermediate_filename): New function.
8621 (output_intermediate_file): New function.
8622 (output_gcov_file): New function
8623 (generate_results): Handle new option.
8624 (release_function): Relase demangled name.
8625 (read_graph_file): Handle demangled name.
8626 (output_lines): Ditto.
8627 * doc/gcov.texi: Document gcov intermediate format.
8628
11f5b71a
VM
86292013-06-19 Vladimir Makarov <vmakarov@redhat.com>
8630
8631 PR bootstrap/57604
8632 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
8633 (lra_emit_add): Use the functions. Add comment about Y as an
8634 address segment.
8635
e1f9340b
DE
86362013-06-19 David Edelsohn <dje.gcc@gmail.com>
8637
8638 PR driver/57652
8639 * collect2.c (collect_atexit): New.
8640 (collect_exit): Delete.
8641 (main): Register collect_atexit with atexit.
8642 (collect_wait): Change collect_exit to exit.
8643 (do_wait): Same.
8644 * collect2.h (collect_exit): Delete.
8645 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
8646
e3f9e0ac
WM
86472013-06-19 Wei Mi <wmi@google.com>
8648
8649 PR rtl-optimization/57518
8650 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
8651 if regno is used in paradoxical subreg.
8652 (update_equiv_regs): Check pdx_subregs[regno] before
8653 set a reg to be equivalent with a mem.
8654
a09f1a76
MK
86552013-06-19 Matthias Klose <doko@ubuntu.com>
8656
8657 PR driver/57651
8658 * file-find.h (find_a_file): Add a mode parameter.
8659 * file-find.c (find_a_file): Likewise.
8660 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
8661 with X_OK for the executables.
8662 * collect2.c (main): Call find_a_file with X_OK.
8663
7d18b0ad
SE
86642013-06-19 Steve Ellcey <sellcey@mips.com>
8665
8666 PR target/56942
8667 * config/mips/mips.md (casesi_internal_mips16_<mode>):
8668 Use NEXT_INSN instead of next_real_insn.
8669
6a6dac52
JH
86702013-06-19 Jan Hubicka <jh@suse.cz>
8671
8672 * cgraph.h (const_value_known_p): Replace by ...
8673 (ctor_for_folding): .. this one.
8674 * cgraphunit.c (process_function_and_variable_attributes): Use it.
8675 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
8676 * expr.c (expand_expr_real_1): Likewise.
8677 (string_constant): Likewise.
8678 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
8679 * ipa.c (process_references): Likewise.
8680 (symtab_remove_unreachable_nodes): Likewise.
8681 * ipa-inline-analysis.c (param_change_prob): Likewise.
8682 * gimple-fold.c (canonicalize_constructor_val): Likewise.
8683 (get_base_constructor): Likwise.
8684 * varpool.c (varpool_remove_node): Likewise.
8685 (varpool_remove_initializer): LIkewise.
8686 (dump_varpool_node): LIkwise.
8687 (const_value_known_p): Rewrite to ...
8688 (ctor_for_folding): ... this one.
8689
216c12ab
JJ
86902013-06-19 Jakub Jelinek <jakub@redhat.com>
8691
8692 PR driver/57651
8693 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
8694 PERSONALITY in $PATH derived prefixes.
8695
d53e2f99
JL
86962013-06-19 Jeff Law <law@redhat.com>
8697
ecdbf306
JL
8698 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
8699 in comment.
8700
d53e2f99
JL
8701 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
8702 (simplify_bitwise_binary): Use it to simpify certain binary ops on
8703 booleans.
8704
89e7fc04
SN
87052013-06-19 Sofiane Naci <sofiane.naci@arm.com>
8706
8707 * config/arm/vfp.md: Move VFP instruction classification documentation
8708 to ...
8709 * config/arm/arm.md: ... here. Update instruction classification
8710 documentation.
8711
5e7f6aaa
RE
87122013-06-19 Richard Earnshaw <rearnsha@arm.com>
8713
8714 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
8715 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
8716 pattern. Use more efficient sequences on ARMv5 and Thumb2.
8717
5fdc1e5d
SB
87182013-06-19 Steven Bosscher <steven@gcc.gnu.org>
8719
8720 PR target/57609
8721 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8722 with NEXT_INSN. Use tablejump_p to check for jump table data
8723 insns.
8724
e952afba
PC
87252013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
8726
8727 PR c++/56544
8728 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
8729 that now in C++ the value is correct per the C++ standards.
8730
3fd0303a
RB
87312013-06-19 Richard Biener <rguenther@suse.de>
8732
8733 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
8734 for global context.
8735
b04e4b08
AK
87362013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8737
8738 Revert:
8739 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8740
8741 PR target/57609
8742 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8743 with next_active_insn.
8744
bef8491a
ST
87452013-06-18 Sriraman Tallam <tmsriram@google.com>
8746
8747 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
8748 functions are inlined during failures to flag an error.
8749 * tree-inline.c (expand_call_inline): Allow the error to be flagged
8750 in early inline pass.
22fc7d3d 8751
c21b257d
L
87522013-06-18 H.J. Lu <hongjiu.lu@intel.com>
8753
8754 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
8755 in comments.
8756
33255ae3
JB
87572013-06-18 Julian Brown <julian@codesourcery.com>
8758
8759 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
8760 Permit virtual register pre-reload if !strict.
8761 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
8762 change.
8763 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
8764 prototype.
8765 * config/arm/neon.md (movmisalign<mode>): Use
8766 neon_perm_struct_or_reg_operand instead of
8767 neon_struct_or_register_operand.
8768 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
8769 neon_permissive_struct_operand instead of neon_struct_operand.
8770 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
8771 neon_vector_mem_operand.
8772 * config/arm/predicates.md (neon_struct_operand): Adjust call to
8773 neon_vector_mem_operand.
8774 (neon_permissive_struct_operand): New.
8775 (neon_struct_or_register_operand): Rename to...
8776 (neon_perm_struct_or_reg_operand): This. Adjust call to
8777 neon_vector_mem_operand.
8778
d579fcda
RB
87792013-06-18 Richard Biener <rguenther@suse.de>
8780
8781 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
8782 * lto-streamer.h: Include pointer-set.h.
8783 (struct lto_decl_slot): Remove.
8784 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
8785 Remove next_index entry.
8786 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8787 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8788 (lto_init_tree_ref_encoder): Adjust.
8789 (lto_destroy_tree_ref_encoder): Likewise.
8790 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
8791 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
8792 (lto_output_decl_index): Adjust.
8793 (lto_new_out_decl_state): Likewise.
8794 (lto_record_function_out_decl_state): Likewise.
8795 * lto-streamer-out.c (copy_function): Likewise.
8796
31ee20ba
RB
87972013-06-18 Richard Biener <rguenther@suse.de>
8798
8799 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
8800 * cgraphunit.c: Include cfgloop.h.
8801 (init_lowered_empty_function): Initialize the loop tree.
8802 (assemble_thunk): Insert new BBs into loops.
8803
ac0511f2
RB
88042013-06-18 Richard Biener <rguenther@suse.de>
8805
8806 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
8807 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
8808 the map from cache entry to cache index optional.
8809 (streamer_tree_cache_replace_tree): Adjust accordingly.
8810 (streamer_tree_cache_append): Likewise.
8811 (streamer_tree_cache_delete): Likewise.
8812 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
8813 streamer cache map from cache entry to cache index.
8814 * lto-streamer-out.c (create_output_block): Adjust.
8815
09485a08
SN
88162013-06-18 Sofiane Naci <sofiane.naci@arm.com>
8817
22fc7d3d
UB
8818 * config/arm/arm.md (attribute "insn"): Move multiplication and
8819 division attributes to...
09485a08
SN
8820 (attribute "type"): ... here. Remove mult.
8821 (attribute "mul32"): New attribute.
8822 (attribute "mul64"): Add umaal.
8823 (*arm_mulsi3): Update attributes.
8824 (*arm_mulsi3_v6): Likewise.
8825 (*thumb_mulsi3): Likewise.
8826 (*thumb_mulsi3_v6): Likewise.
8827 (*mulsi3_compare0): Likewise.
8828 (*mulsi3_compare0_v6): Likewise.
8829 (*mulsi_compare0_scratch): Likewise.
8830 (*mulsi_compare0_scratch_v6): Likewise.
8831 (*mulsi3addsi): Likewise.
8832 (*mulsi3addsi_v6): Likewise.
8833 (*mulsi3addsi_compare0): Likewise.
8834 (*mulsi3addsi_compare0_v6): Likewise.
8835 (*mulsi3addsi_compare0_scratch): Likewise.
8836 (*mulsi3addsi_compare0_scratch_v6): Likewise.
8837 (*mulsi3subsi): Likewise.
8838 (*mulsidi3adddi): Likewise.
8839 (*mulsi3addsi_v6): Likewise.
8840 (*mulsidi3adddi_v6): Likewise.
8841 (*mulsidi3_nov6): Likewise.
8842 (*mulsidi3_v6): Likewise.
8843 (*umulsidi3_nov6): Likewise.
8844 (*umulsidi3_v6): Likewise.
8845 (*umulsidi3adddi): Likewise.
8846 (*umulsidi3adddi_v6): Likewise.
8847 (*smulsi3_highpart_nov6): Likewise.
8848 (*smulsi3_highpart_v6): Likewise.
8849 (*umulsi3_highpart_nov6): Likewise.
8850 (*umulsi3_highpart_v6): Likewise.
8851 (mulhisi3): Likewise.
8852 (*mulhisi3tb): Likewise.
8853 (*mulhisi3bt): Likewise.
8854 (*mulhisi3tt): Likewise.
8855 (maddhisi4): Likewise.
8856 (*maddhisi4tb): Likewise.
8857 (*maddhisi4tt): Likewise.
8858 (maddhidi4): Likewise.
8859 (*maddhidi4tb): Likewise.
8860 (*maddhidi4tt): Likewise.
8861 (divsi3): Likewise.
8862 (udivsi3): Likewise.
8863 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
8864 (thumb2_mulsi_short_compare0): Likewise.
8865 (thumb2_mulsi_short_compare0_scratch): Likewise.
8866 * config/arm/arm1020e.md (1020mult1): Update attribute change.
8867 (1020mult2): Likewise.
8868 (1020mult3): Likewise.
8869 (1020mult4): Likewise.
8870 (1020mult5): Likewise.
8871 (1020mult6): Likewise.
22fc7d3d
UB
8872 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
8873 change.
09485a08
SN
8874 (cortex_a15_mult64): Likewise.
8875 (cortex_a15_sdiv): Likewise.
8876 (cortex_a15_udiv): Likewise.
8877 * config/arm/arm1026ejs.md (mult1): Update attribute change.
8878 (mult2): Likewise.
8879 (mult3): Likewise.
8880 (mult4): Likewise.
8881 (mult5): Likewise.
8882 (mult6): Likewise.
8883 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
8884 (pj4_ir_div): Likewise.
8885 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
8886 (11_mult2): Likewise.
8887 (11_mult3): Likewise.
8888 (11_mult4): Likewise.
8889 (11_mult5): Likewise.
8890 (11_mult6): Likewise.
8891 (11_mult7): Likewise.
8892 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
8893 (cortex_a8_mla): Likewise.
8894 (cortex_a8_mull): Likewise.
8895 (cortex_a8_smulwy): Likewise.
8896 (cortex_a8_smlald): Likewise.
8897 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
8898 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
8899 (cortex_r4_mul_3): Likewise.
8900 (cortex_r4_mla_4): Likewise.
8901 (cortex_r4_mla_3): Likewise.
8902 (cortex_r4_smlald): Likewise.
8903 (cortex_r4_mull): Likewise.
8904 (cortex_r4_sdiv): Likewise.
8905 (cortex_r4_udiv): Likewise.
8906 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
8907 (cortex_a7_idiv): Likewise.
8908 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
8909 (9_mult2): Likewise.
8910 (9_mult3): Likewise.
8911 (9_mult4): Likewise.
8912 (9_mult5): Likewise.
8913 (9_mult6): Likewise.
8914 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
8915 (cortex_a53_sdiv): Likewise.
8916 (cortex_a53_udiv): Likewise.
8917 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
8918 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
8919 (mp626_mult2): Likewise.
8920 (mp626_mult3): Likewise.
8921 (mp626_mult4): Likewise.
8922 * config/arm/fa526.md (526_mult1): Update attribute change.
8923 (526_mult2): Likewise.
8924 * config/arm/arm-generic.md (mult): Update attribute change.
8925 (mult_ldsched_strongarm): Likewise.
8926 (mult_ldsched): Likewise.
8927 (multi_cycle): Likewise.
8928 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
8929 * config/arm/fa606te.md (606te_mult1): Update attribute change.
8930 (606te_mult2): Likewise.
8931 (606te_mult3): Likewise.
8932 (606te_mult4): Likewise.
8933 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
8934 (cortex_a9_mac16): Likewise.
8935 (cortex_a9_multiply): Likewise.
8936 (cortex_a9_mac): Likewise.
8937 (cortex_a9_multiply_long): Likewise.
8938 * config/arm/fa626te.md (626te_mult1): Update attribute change.
8939 (626te_mult2): Likewise.
8940 (626te_mult3): Likewise.
8941 (626te_mult4): Likewise.
8942
8951f345
RB
89432013-06-18 Richard Biener <rguenther@suse.de>
8944
8945 PR lto/57334
8946 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
8947
d8101d05
AK
89482013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8949
8950 PR target/57609
8951 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
8952 with next_active_insn.
8953
69eff9da
AM
89542013-06-18 Alan Modra <amodra@gmail.com>
8955
8956 * config/rs6000/rs6000.h (enum data_align): New.
8957 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
8958 (DATA_ABI_ALIGNMENT): Define.
8959 (CONSTANT_ALIGNMENT): Correct comment.
8960 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
8961 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
8962
8ed00d76
DM
89632013-06-17 David Malcolm <dmalcolm@redhat.com>
8964
8965 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
8966 ATTRIBUTE_UNUSED marking.
8967
3d33d06b
SN
89682013-06-17 Sofiane Naci <sofiane.naci@arm.com>
8969
8970 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
8971 alternative and update.
8972 (aarch64_dup_lanedi): Delete.
8973 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
8974 * config/aarch64/aarch64-simd-builtins.def: Update.
8975
ee03e71d
RB
89762013-06-17 Richard Biener <rguenther@suse.de>
8977
8978 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
8979 (lto_input_scc): Declare.
8980 (lto_input_tree_1): Likewise.
8981 (struct lto_stats_d): Add num_tree_bodies_output and
8982 num_pickle_refs_output.
8983 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
8984 (lto_read_tree_1): Split out from ...
8985 (lto_read_tree): ... this.
8986 (lto_input_scc): New function.
8987 (lto_input_tree_1): Split out from ...
8988 (lto_input_tree): ... this. Handle LTO_tree_scc.
8989 (lto_data_in_create): Create the streamer cache without hashes.
8990 * lto-streamer-out.c (create_output_block): Create the streamer
8991 cache with hashes when not doing WPA.
8992 (lto_write_tree_1): Split out from ...
8993 (lto_write_tree): ... this.
8994 (get_symbol_initial_value): New function.
8995 (lto_output_tree_1): Split out from ...
8996 (lto_output_tree): ... this. Write trees as series of SCCs
8997 using a DFS walk via DFS_write_tree.
8998 (struct sccs, struct scc_entry): New types.
8999 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
9000 (DFS_write_tree_body): New function.
9001 (DFS_write_tree): Likewise.
9002 (hash_tree): Likewise.
9003 (scc_entry_compare): Likewise.
9004 (hash_scc): Likewise.
9005 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
9006 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
9007 TREE_CHAIN as regular reference.
9008 (streamer_read_integer_cst): Remove.
9009 (streamer_get_pickled_tree): Adjust.
9010 * tree-streamer-out.c (streamer_write_chain): Disable streaming
9011 of DECL_EXTERNALs in BLOCK_VARS for now.
9012 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
9013 reference.
9014 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
9015 Add hash value argument and record that if hashes are recorded
9016 in the cache.
9017 (streamer_tree_cache_insert_1): Adjust.
9018 (streamer_tree_cache_insert): Likewise.
9019 (streamer_tree_cache_insert_at): Rename to ...
9020 (streamer_tree_cache_replace_tree): ... this and adjust.
9021 (streamer_tree_cache_append): Adjust.
9022 (record_common_node): Likewise.
9023 (streamer_tree_cache_create): Add argument whether to
9024 record hash values together with trees.
9025 (streamer_tree_cache_delete): Adjust.
9026 * tree-streamer.h (struct streamer_tree_cache_d): Add
9027 vector of hashes.
9028 (streamer_read_integer_cst): Remove.
9029 (streamer_tree_cache_insert): Adjust.
9030 (streamer_tree_cache_append): Likewise.
9031 (streamer_tree_cache_insert_at): Rename to ...
9032 (streamer_tree_cache_replace_tree): ... this and adjust.
9033 (streamer_tree_cache_create): Add argument whether to record hashes.
9034 (streamer_tree_cache_get): Rename to ...
9035 (streamer_tree_cache_get_tree): ... this.
9036 (streamer_tree_cache_get_hash): New function.
9037 * tree.c (cache_integer_cst): New function.
9038 * tree.h (cache_integer_cst): Declare.
9039 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
9040 * lto-symtab.c (lto_varpool_replace_node): Only release
9041 DECL_INITIAL of non-prevailing decls.
9042 * varpool.c (varpool_remove_initializer): Do not release
9043 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
9044
0de86a92
JU
90452013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9046
9047 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
9048 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
9049 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
9050 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
9051 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
9052 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
9053 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
9054 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
9055 instead of TARGET_64BIT.
9056 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
9057 Require ISA_HAS_<D>DIV.
9058
74240413
RS
90592013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
9060
9061 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
9062 (mips*-*-linux*): Move default with_llsc setting to where other
9063 defaults are set.
9064 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
9065 with_arch block.
9066 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
9067 Likewise. Remove default with_tune setting. Move default float
9068 setting to its own block. Handle with_llsc in the same block as above.
9069
12211b99 90702013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR
9071
9072 PR rtl-optimization/57425
9073 PR rtl-optimization/57569
299a5f6a 9074 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR
9075 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
9076 Changed all callers.
9077 (canon_anti_dependence): New function.
9078 * cse.c (check_dependence): Use canon_anti_dependence.
9079 * cselib.c (cselib_invalidate_mem): Likewise.
9080 * rtl.h (canon_anti_dependence): Declare.
9081
b259d352
JU
90822013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
9083
9084 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
9085 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
9086 ".set mips3" for 64-bit targets.
9087
4de09b85
DC
90882013-06-15 Dehao Chen <dehao@google.com>
9089
9090 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
9091 * gimple-low.c (gimple_check_call_matching_types): Likewise.
9092 (gimple_check_call_args): Likewise.
9093 * value-prof.c (check_ic_target): Likewise.
9094 * ipa-inline.c (early_inliner): Likewise.
9095 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
9096 * cgraph.c (cgraph_create_edge_1): Likewise.
9097 (cgraph_make_edge_direct): Likewise.
9098
d1aa4a76
MM
90992013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
9100
9101 PR target/57615
9102 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
9103 rs6000_output_move_128bit to handle emitting quad memory
9104 operations. Set attribute length to 8 bytes.
9105
12211b99 91062013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP
9107
9108 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
9109 New pattern.
9110 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
9111 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
9112 (aarch64_<su>mlsl<mode>): Likewise.
9113
f5273574
MS
91142013-06-14 Mike Stump <mikestump@comcast.net>
9115
9116 * Makefile.in (TARGET_H): Add insn-codes.h.
9117
ef7e7ab2
AM
91182013-06-14 Alan Modra <amodra@gmail.com>
9119
9120 PR middle-end/57134
9121 PR middle-end/57586
9122 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
9123 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
9124 bitfield expansion when EXPAND_MEMORY.
9125 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
9126
135faab6
MM
91272013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
9128
9129 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
9130 test for clearing quad memory on 32-bit later.
9131
948a1fd9
MG
91322013-06-13 Marc Glisse <marc.glisse@inria.fr>
9133
9134 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
9135 (fold_negate_expr): Likewise.
9136 (fold_real_zero_addition_p): Handle vectors.
9137 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
9138
0597fb9c
AM
91392013-06-14 Alan Modra <amodra@gmail.com>
9140
9141 * varasm.c (force_const_mem): Revert 2013-06-07 change.
9142
e5b962d0
JH
91432013-06-13 Jan Hubicka <jh@suse.cz>
9144
9145 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
9146 Local comdats are not externally visible.
9147 * symtab.c (dump_symtab_base): Dump externally visible.
9148 (verify_symtab_base): Verify back links in the symtab hash.
9149
a8355e51
BC
91502013-06-13 Bin Cheng <bin.cheng@arm.com>
9151
9152 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
9153 CONVERT_EXPR as equal nodes.
9154
8f7e6e33
BC
91552013-06-13 Bin Cheng <bin.cheng@arm.com>
9156
22fc7d3d 9157 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 9158
5d418483
MG
91592013-06-13 Marc Glisse <marc.glisse@inria.fr>
9160
9161 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
9162 Generalize to complex and vector.
9163 * tree.c (build_all_ones_cst): New function.
9164 * tree.h (build_all_ones_cst): Declare it.
9165
50751417
AM
91662013-06-13 Alan Modra <amodra@gmail.com>
9167
9168 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
9169 * config/rs6000/rs6000.md (signbittf2): New insn.
9170 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
9171 (abstf2_internal, cmptf_internal2): Likewise.
9172 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
9173
4b02c962 91742013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
9175 Pat Haugen <pthaugen@us.ibm.com>
9176 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM
9177
9178 * config/rs6000/rs6000.c (emit_load_locked): Add support for
9179 power8 byte, half-word, and quad-word atomic instructions.
9180 (emit_store_conditional): Likewise.
9181 (rs6000_expand_atomic_compare_and_swap): Likewise.
9182 (rs6000_expand_atomic_op): Likewise.
9183
9184 * config/rs6000/sync.md (larx): Add new modes for power8.
9185 (stcx): Likewise.
9186 (AINT): New mode iterator to include TImode as well as normal
9187 integer modes on power8.
9188 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
9189 that VSX registers are not considered. Use AINT mode iterator
9190 instead of INT1 to allow inclusion of quad word atomic operations
9191 on power8.
9192 (load_locked<mode>): Likewise.
9193 (store_conditional<mode>): Likewise.
9194 (atomic_compare_and_swap<mode>): Likewise.
9195 (atomic_exchange<mode>): Likewise.
9196 (atomic_nand<mode>): Likewise.
9197 (atomic_fetch_<fetchop_name><mode>): Likewise.
9198 (atomic_nand_fetch<mode>): Likewise.
9199 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
9200 each type.
9201 (ATOMIC): On power8, add QImode, HImode modes.
9202 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
9203 modes that promote to SImode.
9204 (load_lockedti): Convert TImode arguments to PTImode, so that we
9205 get a guaranteed even/odd register pair.
9206 (load_lockedpti): Likewise.
9207 (store_conditionalti): Likewise.
9208 (store_conditionalpti): Likewise.
9209
9210 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
9211 atomic load/store instructions.
9212 (HSI): Likewise.
9213
710b6cc1
RS
92142013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
9215
9216 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
9217 loads.
9218 (insn_count): New attribute, with most cases extracted from...
9219 (length): ...here. Redefine most cases in terms of insn_count.
9220 (single_insn): Delete.
9221 (can_delay): Use insn_count to check for single instructions.
9222 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
9223 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
9224 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
9225 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
9226 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
9227 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
9228 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
9229 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
9230 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
9231 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
9232 rather than "length".
9233 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
9234 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
9235 Use "insn_count" rather than "length".
9236 * config/mips/mips-dsp.md
9237 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
9238 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
9239 length attributes.
9240
1951f101
MG
92412013-06-12 Marc Glisse <marc.glisse@inria.fr>
9242
9243 PR tree-optimization/57361
9244 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
9245
8b033a8a
SN
92462013-06-12 Sofiane Naci <sofiane.naci@arm.com>
9247
22fc7d3d
UB
9248 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
9249 to split.
8b033a8a
SN
9250 (aarch64_simd_combine<mode>): New instruction expansion.
9251 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
9252 function prototype.
9253 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
9254 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
9255
ca0f62a8
JH
92562013-06-12 Jan Hubicka <jh@suse.cz>
9257
9258 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
9259 decl has when in streaming stage.
9260 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
9261 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
9262
12211b99 92632013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS
9264
9265 PR target/57578
9266 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
9267
bfdeda2c
JJ
92682013-06-12 Jakub Jelinek <jakub@redhat.com>
9269
9270 PR tree-optimization/57537
9271 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
9272 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
9273
a4fa02d1
RB
92742013-06-12 Richard Biener <rguenther@suse.de>
9275
9276 * data-streamer.h (streamer_write_char_stream): CSE
9277 obs->current_pointer.
9278 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
9279 streamer_write_char_stream manually and optimize the resulting loop.
9280 (streamer_write_hwi_stream): Likewise.
9281
bbf9ad07
JH
92822013-06-12 Jan Hubicka <jh@suse.cz>
9283
9284 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
9285 * cgraph.h (varpool_create_empty_node): Declare.
9286 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
9287 duplicated nodes.
9288 * symtab.c (symtab_unregister_node): Be lax about missin entries
9289 in node hash.
9290 (symtab_get_node): Update comment.
9291 * varpool.c (varpool_create_empty_node): Break out from ...
9292 (varpool_node_for_decl): ... here.
9293 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
9294
a53f90ad
EB
92952013-06-12 Eric Botcazou <ebotcazou@adacore.com>
9296
9297 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
9298 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
9299 part. Use straight-line flow at the end.
9300 <COMPONENT_REF>: Remove superfluous else.
9301 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
9302
946f9306
JJ
93032013-06-12 Jakub Jelinek <jakub@redhat.com>
9304
9305 PR target/56564
9306 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
9307 target hook even for !TREE_PUBLIC decls. If no resolution info
9308 is available, return false for common and external decls.
9309
a7fca6f0
KP
93102013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
9311
9312 * config/rl78/constraints.md (U): New constraint.
9313 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
9314 valloc attribute.
9315
c7d8f446
MM
93162013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
9317
9318 PR target/57589
9319 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
9320 to allow returning address to AT_PLATFORM name.
9321
08346abd
JH
93222013-06-11 Jan Hubicka <jh@suse.cz>
9323
9324 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
9325 * cgraph.h (symtab_node_base): Add weakref flag.
9326 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
9327 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
9328 (output_weakrefs): Use weakref flag.
9329 * fold-const.c (simple_operand_p): Handle WEAK.
9330 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
9331 * ipa.c (varpool_externally_visible_p): Drop weakref.
9332 (function_and_variable_visibility): Update comment; fix weakref
9333 sanity checks; do not clear DECL_WEAK on them.
9334 * lto-cgraph.c (lto_output_node): update.
9335 (lto_output_varpool_node): Update.
9336 (input_overwrite_node): Update.
9337 (input_node): Update.
9338 (input_varpool_node): Update.
9339 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
9340 (lto_symtab_merge_symbols): Add sanity check.
9341 (lto_symtab_prevailing_decl): Do not special case weakrefs.
9342 * passes.c (rest_of_decl_compilation): Set static flag, too.
9343 * symtab.c (dump_symtab_base): Dump weakref.
9344 (verify_symtab_base): Sanity check weakrefs.
9345 (symtab_make_decl_local): Remove duplicated code.
9346 (symtab_alias_ultimate_target): Simplify.
9347 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 9348
2e6a2f88
TV
93492013-06-11 Tom de Vries <tom@codesourcery.com>
9350
9351 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
9352 sequence_vect == NULL.
9353
d9af3652
DD
93542013-06-11 DJ Delorie <dj@redhat.com>
9355
9356 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
9357 (rl78_unwind_word_mode): New.
9358
ffa66012
DM
93592013-06-11 David Malcolm <dmalcolm@redhat.com>
9360
9361 * final.c (debug_prefix_maps): Make static.
9362
783f2d35
DM
93632013-06-11 David Malcolm <dmalcolm@redhat.com>
9364
9365 * function.c (initial_trampoline): Remove stray copy.
9366
7af26def
SN
93672013-06-11 Sofiane Naci <sofiane.naci@arm.com>
9368
9369 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
9370
97756c0e
MJ
93712013-06-11 Martin Jambor <mjambor@suse.cz>
9372
9373 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
9374 within bounds at the beginning of the function.
9375
2c7b8bf6
AM
93762013-06-11 Alan Modra <amodra@gmail.com>
9377
9378 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
9379 reporting.
9380 (get_named_section): Don't NULL !DECL_P decl.
9381
77cea46e
IZ
93822013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
9383
9384 * doc/invoke.texi (core-avx2): Document.
9385 (slm): Likewise.
9386 (atom): Updated with MOVBE.
9387
888f0920
RB
93882013-06-11 Richard Biener <rguenther@suse.de>
9389
22fc7d3d 9390 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 9391
6b438d58
AB
93922013-06-11 Anton Blanchard <anton@samba.org>
9393
9394 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
9395 correct shift value in little-endian mode.
9396
3bc66938
JJ
93972013-06-11 Jakub Jelinek <jakub@redhat.com>
9398
9399 PR target/56564
9400 * varasm.c (get_variable_align): Move #endif to the right place.
9401
a5dc7f88
CC
94022013-06-10 Cary Coutant <ccoutant@google.com>
9403
9404 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
9405 for hash so that hash table traversal order is deterministic.
9406
19be72ab 94072013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
9408 Pat Haugen <pthaugen@us.ibm.com>
9409 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM
9410
9411 * config/rs6000/vector.md (GPR move splitter): Do not split moves
9412 of vectors in GPRS if they are direct moves or quad word load or
9413 store moves.
9414
9415 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
9416 declaration.
9417 (direct_move_p): Likewise.
9418 (quad_load_store_p): Likewise.
9419
9420 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
9421 classes into bins based on the physical register type.
9422 (reg_class_to_reg_type): Likewise.
9423 (IS_STD_REG_TYPE): Likewise.
9424 (IS_FP_VECT_REG_TYPE): Likewise.
9425 (reload_fpr_gpr): Arrays to determine what insn to use if we can
9426 use direct move instructions.
9427 (reload_gpr_vsx): Likewise.
9428 (reload_vsx_gpr): Likewise.
9429 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
9430 information that is a simplification of register classes. Also
9431 precalculate direct move reload helpers.
9432 (direct_move_p): New function to return true if the operation can
9433 be done as a direct move instruciton.
9434 (quad_load_store_p): New function to return true if the operation
9435 is a quad memory operation.
9436 (rs6000_legitimize_address): If quad memory, only allow register
9437 indirect for TImode addresses.
9438 (rs6000_legitimate_address_p): Likewise.
9439 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
9440 (rs6000_reload_register_type): Likewise.
9441 (register_to_reg_type): Return register type.
9442 (rs6000_secondary_reload_simple_move): New helper function for
9443 secondary reload and secondary memory needed to identify anything
9444 that is a simple move, and does not need reloading.
9445 (rs6000_secondary_reload_direct_move): New helper function for
9446 secondary reload to identify cases that can be done with several
9447 instructions via the direct move instructions.
9448 (rs6000_secondary_reload_move): New helper function for secondary
9449 reload to identify moves between register types that can be done.
9450 (rs6000_secondary_reload): Add support for quad memory operations
9451 and for direct move.
9452 (rs6000_secondary_memory_needed): Likewise.
9453 (rs6000_debug_secondary_memory_needed): Change argument names.
9454 (rs6000_output_move_128bit): New function to return the move to
9455 use for 128-bit moves, including knowing about the various
9456 limitations of quad memory operations.
9457
9458 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
9459 memory operations. call rs6000_output_move_128bit for the actual
9460 instruciton(s) to generate.
9461 (vsx_movti_64bit): Likewise.
9462
9463 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
9464 (UNSPEC_P8V_MTVSRWZ): Likewise.
9465 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
9466 (UNSPEC_P8V_MTVSRD): Likewise.
9467 (UNSPEC_P8V_XXPERMDI): Likewise.
9468 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
9469 (UNSPEC_FUSION_GPR): Likewise.
9470 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 9471 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM
9472 (f32_sv): Likewise.
9473 (f32_dm): Likewise.
9474 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
9475 loads and direct move instructions.
9476 (zero_extendsidi2_lfiwzx): Likewise.
9477 (extendsidi2_lfiwax): Likewise.
9478 (extendsidi2_nocell): Likewise.
9479 (floatsi<mode>2_lfiwax): Likewise.
9480 (lfiwax): Likewise.
9481 (floatunssi<mode>2_lfiwzx): Likewise.
9482 (lfiwzx): Likewise.
9483 (fix_trunc<mode>_stfiwx): Likewise.
9484 (fixuns_trunc<mode>_stfiwx): Likewise.
9485 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
9486 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
9487 (parity<mode>2_cmpb): Set length/type attr.
9488 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
9489 for 'mr.' to fast_compare.
9490 (bpermd_<mode>): Change type attr to popcnt.
9491 (p8_fmrgow_<mode>): New insns for power8 direct move support.
9492 (p8_mtvsrwz_1): Likewise.
9493 (p8_mtvsrwz_2): Likewise.
9494 (reload_fpr_from_gpr<mode>): Likewise.
9495 (p8_mtvsrd_1): Likewise.
9496 (p8_mtvsrd_2): Likewise.
9497 (p8_xxpermdi_<mode>): Likewise.
9498 (reload_vsx_from_gpr<mode>): Likewise.
9499 (reload_vsx_from_gprsf): Likewise.
9500 (p8_mfvsrd_3_<mode>): LIkewise.
9501 (reload_gpr_from_vsx<mode>): Likewise.
9502 (reload_gpr_from_vsxsf): Likewise.
9503 (p8_mfvsrd_4_disf): Likewise.
9504 (multi-word GPR splits): Do not split direct moves or quad memory
9505 operations.
9506
16876bdc
DM
95072013-06-10 David Malcolm <dmalcolm@redhat.com>
9508
9509 * tree-into-ssa.c (interesting_blocks): Make static.
9510
df8a1d28
JJ
95112013-06-10 Jakub Jelinek <jakub@redhat.com>
9512
9513 PR target/56564
9514 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
9515 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
9516 Use DATA_ABI_ALIGNMENT for that case instead if defined.
9517 (get_variable_align): New function.
9518 (get_variable_section, emit_bss, emit_common,
9519 assemble_variable_contents, place_block_symbol): Use
9520 get_variable_align instead of DECL_ALIGN.
9521 (assemble_noswitch_variable): Add align argument, use it
9522 instead of DECL_ALIGN.
9523 (assemble_variable): Adjust caller. Use get_variable_align
9524 instead of DECL_ALIGN.
9525 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
9526 caller.
9527 (DATA_ABI_ALIGNMENT): Define.
9528 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
9529 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
9530 opt is false, only return the psABI mandated alignment increase.
9531 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
9532 (DATA_ABI_ALIGNMENT): ... this.
9533 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
9534 (DATA_ABI_ALIGNMENT): ... this.
9535 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
9536 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
9537 (DATA_ABI_ALIGNMENT): ... this.
9538 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
9539 * doc/tm.texi: Regenerated.
9540
57b29ca6
UB
95412013-06-10 Uros Bizjak <ubizjak@gmail.com>
9542
9543 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
9544 cmp_code to construct REG_EQUAL note.
9545
3e56ed50
JJ
95462013-06-09 Jakub Jelinek <jakub@redhat.com>
9547
9548 PR target/57568
9549 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
9550 that operands[2] doesn't overlap with operands[0].
9551
74fe2a1d
DE
95522013-06-09 David Edelsohn <dje.gcc@gmail.com>
9553 Jan Hubicka <jh@suse.cz>
9554
9555 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
9556 hack to mark symbols as used.
9557
ed52a84e
VM
95582013-06-08 Vladimir Makarov <vmakarov@redhat.com>
9559
9560 PR rtl-optimization/57559
9561 * lra-constraints.c (process_alt_operands): Don't discourage
9562 memory with known offset for offsetable memory constraint.
9563 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
9564
299316ed
EB
95652013-06-08 Eric Botcazou <ebotcazou@adacore.com>
9566
9567 * varasm.c (struct oc_local_state): Reorder fields.
9568 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
9569 and adjust accordingly.
9570 (output_constructor): Reorder initialization code and adjust call to
9571 output_constructor_bitfield.
9572
88ac60d3
JH
95732013-06-07 Jan Hubicka <jh@suse.cz>
9574
9575 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
9576
4cdce1a8
DM
95772013-06-07 David Malcolm <dmalcolm@redhat.com>
9578
9579 * tree-object-size.c (unknown): Make const.
9580
b5e0425c
AK
95812013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9582
9583 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
9584 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
9585 for last alternative in the cpu_facility attribute.
9586
79678d04
KT
95872013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9588
9589 PR target/56315
9590 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
9591 (xordi3): Change operand 2 constraint to arm_xordi_operand.
9592 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
9593 * config/arm/constraints.md (Dg): New constraint.
9594 * config/arm/neon.md (xordi3_neon): Remove.
9595 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
9596 * config/arm/predicates.md (arm_xordi_operand): New predicate.
9597
b31ddbdb
KT
95982013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9599
9600 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
9601 Clean up alternatives.
9602
8f90b7d4
AM
96032013-06-07 Alan Modra <amodra@gmail.com>
9604
9605 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
9606 va_list_gpr_size.
9607
961ce119
AM
96082013-06-07 Alan Modra <amodra@gmail.com>
9609
9610 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
9611
33e49835
KT
96122013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9613
9614 * config/arm/constraints.md (Df): New constraint.
9615 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
9616 Correct length attribute for last two alternatives.
9617
acd17ae6
AM
96182013-06-07 Alan Modra <amodra@gmail.com>
9619
9620 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
9621 override user -mfp-in-toc.
9622 (offsettable_ok_by_alignment): Consider just the current access
9623 rather than the whole object, unless BLKmode. Handle
9624 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
9625 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
9626 for -mcmodel=medium.
9627 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
9628 override user -mfp-in-toc or -msum-in-toc. Default to
9629 -mno-fp-in-toc for -mcmodel=medium.
9630
73310b0e
DD
96312013-06-06 DJ Delorie <dj@redhat.com>
9632
9633 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
9634 TARGET_VALID_POINTER_MODE.
9635
0bd62dca 96362013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
9637 Pat Haugen <pthaugen@us.ibm.com>
9638 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM
9639
9640 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
9641 Document new power8 builtins.
9642
9643 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
9644 condition code register, to allow 128-bit logical operations to be
9645 done in the VSX or GPR registers.
9646 (nor<mode>3): Use the canonical form for nor.
9647 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
9648 vclz*, and vpopcnt* vector instructions.
9649 (nand<mode>3): Likewise.
9650 (orc<mode>3): Likewise.
9651 (clz<mode>2): LIkewise.
9652 (popcount<mode>2): Likewise.
9653
9654 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
9655 that only the GPRs are recognized.
9656
9657 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
9658 support for new power8 builtins.
9659
9660 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
9661 builtin functions.
9662 (xscvdpspn): Likewise.
9663 (vclz): Likewise.
9664 (vclzb): Likewise.
9665 (vclzh): Likewise.
9666 (vclzw): Likewise.
9667 (vclzd): Likewise.
9668 (vpopcnt): Likewise.
9669 (vpopcntb): Likewise.
9670 (vpopcnth): Likewise.
9671 (vpopcntw): Likewise.
9672 (vpopcntd): Likewise.
9673 (vgbbd): Likewise.
9674 (vmrgew): Likewise.
9675 (vmrgow): Likewise.
9676 (eqv): Likewise.
9677 (eqv_v16qi3): Likewise.
9678 (eqv_v8hi3): Likewise.
9679 (eqv_v4si3): Likewise.
9680 (eqv_v2di3): Likewise.
9681 (eqv_v4sf3): Likewise.
9682 (eqv_v2df3): Likewise.
9683 (nand): Likewise.
9684 (nand_v16qi3): Likewise.
9685 (nand_v8hi3): Likewise.
9686 (nand_v4si3): Likewise.
9687 (nand_v2di3): Likewise.
9688 (nand_v4sf3): Likewise.
9689 (nand_v2df3): Likewise.
9690 (orc): Likewise.
9691 (orc_v16qi3): Likewise.
9692 (orc_v8hi3): Likewise.
9693 (orc_v4si3): Likewise.
9694 (orc_v2di3): Likewise.
9695 (orc_v4sf3): Likewise.
9696 (orc_v2df3): Likewise.
9697
9698 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
9699 allow power8 quad mode in 64-bit.
9700 (rs6000_builtin_vectorized_function): Add support to vectorize
9701 ISA 2.07 count leading zeros, population count builtins.
9702 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
9703 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
9704 (builtin_function_type): Add vgbbd builtin function which takes an
9705 unsigned argument.
9706 (altivec_expand_vec_perm_const): Add support for new power8 merge
9707 instructions.
9708
9709 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
9710 that does not include TImdoe for use with 32-bit.
9711 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
9712 instructions.
9713 (UNSPEC_VSX_CVDPSPN): Likewise.
9714 (vsx_xscvdpspn): Likewise.
9715 (vsx_xscvspdpn): Likewise.
9716 (vsx_xscvdpspn_scalar): Likewise.
9717 (vsx_xscvspdpn_directmove): Likewise.
9718 (vsx_and<mode>3): Split logical operations into 32-bit and
9719 64-bit. Add support to do logical operations on TImode as well as
9720 VSX vector types. Allow logical operations to be done in either
9721 VSX registers or in general purpose registers in 64-bit mode. Add
9722 splitters if GPRs were used. For AND, add clobber of CCmode to
9723 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
9724 encoding.
9725 (vsx_and<mode>3_32bit): Likewise.
9726 (vsx_and<mode>3_64bit): Likewise.
9727 (vsx_ior<mode>3): Likewise.
9728 (vsx_ior<mode>3_32bit): Likewise.
9729 (vsx_ior<mode>3_64bit): Likewise.
9730 (vsx_xor<mode>3): Likewise.
9731 (vsx_xor<mode>3_32bit): Likewise.
9732 (vsx_xor<mode>3_64bit): Likewise.
9733 (vsx_one_cmpl<mode>2): Likewise.
9734 (vsx_one_cmpl<mode>2_32bit): Likewise.
9735 (vsx_one_cmpl<mode>2_64bit): Likewise.
9736 (vsx_nor<mode>3): Likewise.
9737 (vsx_nor<mode>3_32bit): Likewise.
9738 (vsx_nor<mode>3_64bit): Likewise.
9739 (vsx_andc<mode>3): Likewise.
9740 (vsx_andc<mode>3_32bit): Likewise.
9741 (vsx_andc<mode>3_64bit): Likewise.
9742 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
9743 and xxlorc instructions.
9744 (vsx_eqv<mode>3_64bit): Likewise.
9745 (vsx_nand<mode>3_32bit): Likewise.
9746 (vsx_nand<mode>3_64bit): Likewise.
9747 (vsx_orc<mode>3_32bit): Likewise.
9748 (vsx_orc<mode>3_64bit): Likewise.
9749
9750 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
9751
9752 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
9753 instruction.
9754 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
9755 (p8_vmrgow): Likewise.
9756 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
9757 GPRs to be split under VSX.
9758 (p8v_clz<mode>2): Add power8 count leading zero support.
9759 (p8v_popcount<mode>2): Add power8 population count support.
9760 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
9761 support.
9762
9763 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
9764 instruction.
9765
9766 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
9767 builtin functions.
9768 (vec_nand): Likewise.
9769 (vec_vclz): Likewise.
9770 (vec_vclzb): Likewise.
9771 (vec_vclzd): Likewise.
9772 (vec_vclzh): Likewise.
9773 (vec_vclzw): Likewise.
9774 (vec_vgbbd): Likewise.
9775 (vec_vmrgew): Likewise.
9776 (vec_vmrgow): Likewise.
9777 (vec_vpopcnt): Likewise.
9778 (vec_vpopcntb): Likewise.
9779 (vec_vpopcntd): Likewise.
9780 (vec_vpopcnth): Likewise.
9781 (vec_vpopcntw): Likewise.
9782
37684c46
VM
97832013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9784
9785 PR rtl-optimization/57468
9786 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
9787 spilled pseudos.
9788
d89ae27c
VM
97892013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9790
9791 PR rtl-optimization/57459
9792 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
9793 type when setting live regs.
9794
3597e113
VM
97952013-06-06 Vladimir Makarov <vmakarov@redhat.com>
9796
9797 * config/s390/s390.opt (mlra): New option.
9798 * config/s390/s390.c (s390_decompose_address): Check displacement
9799 for all registers for LRA.
9800 (s390_secondary_reload): Don't used secondary reloads for LRA.
9801 (s390_lra_p): New function.
9802 (TARGET_LRA_P): Define.
9803 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
9804 of attribute cpu_facility to zarch for the last alternative.
9805 (*cmpmem_short): Ditto.
9806
01037aeb
EB
98072013-06-06 Eric Botcazou <ebotcazou@adacore.com>
9808
9809 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
9810 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
9811 (arm_expand_prologue): Likewise.
9812
3371a64f
TJ
98132013-06-06 Teresa Johnson <tejohnson@google.com>
9814
9815 PR c++/53743
9816 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
9817 as this is now done by redirect_edge_and_branch_force.
9818 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
9819 barriers, and fix interaction with splitting.
9820 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
9821 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
9822 reflect changes made in the routine.
9823 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
9824 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
9825 since this is called in cfglayout mode, and replace partition fixup
9826 with assert as that is now done by force_nonfallthru_and_redirect.
9827 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
9828 already be marked with region crossing note.
9829 (insert_section_boundary_note): Make non-static, gate on flag
9830 has_bb_partition, rewrite to also check for multiple partitions.
9831 (rest_of_handle_reorder_blocks): Remove call to
9832 insert_section_boundary_note, now done later during free_cfg.
9833 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
9834 * bb-reorder.h (insert_section_boundary_note): Declare.
9835 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
9836 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
9837 invoke insert_section_boundary_note.
9838 (try_redirect_by_replacing_jump): Remove unnecessary
9839 check for region crossing note.
9840 (fixup_partition_crossing): New function.
9841 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
9842 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
9843 in non-cfglayout mode.
9844 (force_nonfallthru_and_redirect): Fixup partition boundaries,
9845 remove old code that tried to do this. Emit barrier correctly
9846 when we are in cfglayout mode.
9847 (last_bb_in_partition): New function.
9848 (rtl_split_edge): Correctly fixup partition boundaries.
9849 (commit_one_edge_insertion): Remove old code that tried to
9850 fixup region crossing edge since this is now handled in
9851 split_block, and set up insertion point correctly since
9852 block may now end in a jump.
9853 (verify_hot_cold_block_grouping): Guard against checking when not in
9854 linearized RTL mode.
9855 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
9856 notes.
9857 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
9858 rtl_verify_flow_info, so not called in cfglayout mode.
9859 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
9860 (fixup_reorder_chain): Remove old code that attempted to fixup region
9861 crossing note as this is now handled in force_nonfallthru_and_redirect.
9862 (duplicate_insn_chain): Don't duplicate switch section notes.
9863 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
9864 note.
9865 * basic-block.h (emit_barrier_after_bb): Declare.
9866
66071e10
KT
98672013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9868
9869 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
9870 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
9871 arm_usatsihi): Adjust alternatives for arm_restrict_it.
9872
0e26bf3d
KT
98732013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9874
9875 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
9876 where appropriate.
9877 * config/arm/ldmstm.md: Regenerate.
9878
12b4e7ef
KT
98792013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9880
9881 * config/arm/sync.md (atomic_loaddi_1):
9882 Disable predication for arm_restrict_it.
9883 (arm_load_exclusive<mode>): Likewise.
9884 (arm_load_exclusivesi): Likewise.
9885 (arm_load_exclusivedi): Likewise.
9886 (arm_load_acquire_exclusive<mode>): Likewise.
9887 (arm_load_acquire_exclusivesi): Likewise.
9888 (arm_load_acquire_exclusivedi): Likewise.
9889 (arm_store_exclusive<mode>): Likewise.
9890 (arm_store_exclusive<mode>): Likewise.
9891 (arm_store_release_exclusivedi): Likewise.
9892 (arm_store_release_exclusive<mode>): Likewise.
9893
15d16c8a
RB
98942013-06-06 Richard Biener <rguenther@suse.de>
9895
9896 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
9897 after LTO_null.
9898 (lto_tag_is_tree_code_p): Adjust.
9899 (lto_tag_is_gimple_code_p): Likewise.
9900 (lto_gimple_code_to_tag): Likewise.
9901 (lto_tag_to_gimple_code): Likewise.
9902 (lto_tree_code_to_tag): Likewise.
9903 (lto_tag_to_tree_code): Likewise.
9904 * data-streamer.h (streamer_write_hwi_in_range): Use
9905 uhwi streaming to stream the normalized range.
9906 (streamer_read_hwi_in_range): Likewise.
9907
17a7fc37
KT
99082013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9909
9910 * config/arm/arm.md (enabled_for_depr_it): New attribute.
9911 (predicable_short_it): Likewise.
9912 (predicated): Likewise.
9913 (enabled): Handle above.
9914 (define_cond_exec): Set predicated attribute to yes.
9915
b57ca59b
MS
99162013-06-05 Mike Stump <mikestump@comcast.net>
9917
9918 * gdbinit.in (__FUNCTION__): Add.
9919
c1e183a9
UB
99202013-06-05 Uros Bizjak <ubizjak@gmail.com>
9921
9922 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
9923 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
9924
0378bddb
JH
99252013-06-05 Jan Hubicka <jh@suse.cz>
9926
9927 * varasm.c (mark_decl_referenced): Revert the removal until targets
9928 are fixed.
9929
8f8a2057
DE
99302013-06-05 David Edelsohn <dje.gcc@gmail.com>
9931
9932 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
9933 instead of mark_decl_referenced.
9934
edb983b2
JH
99352013-06-05 Jan Hubicka <jh@suse.cz>
9936
9937 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
9938 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
9939 and symtab_used_from_object_file_p.
9940 (cgraph_make_node_local_1): Clear forced_by_abi.
9941 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
9942 * cgraph.h (symtab_node_base): Add forced_by_abi.
9943 (decide_is_variable_needed): Remove.
9944 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
9945 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
9946 (decide_is_symbol_needed): ... this one; handle symbols in general;
9947 always analyze virtuals; honnor forced_by_abi.
9948 (cgraph_finalize_function): Update.
9949 (varpool_finalize_decl): Update.
9950 (symbol_defined_and_needed): Remove.
9951 (analyze_functions): Update.
9952 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
9953 output_refs, input_overwrite_node): Handle forced_by_abi.
9954 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
9955 (address_taken_from_non_vtable_p): ... this one.
9956 (comdat_can_be_unshared_p_1): New function.
9957 (cgraph_comdat_can_be_unshared_p): Rename to ...
9958 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
9959 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
9960 (function_and_variable_visibility): Clear forced_by_abi as needed.
9961 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
9962 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH
9963 * symtab.c (dump_symtab_base): Dump forced_by_abi.
9964 * varpool.c (decide_is_variable_needed): Remove.
9965
9912dbe5
KT
99662013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9967
9968 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
9969 (arm_option_override): Override arm_restrict_it where appropriate.
9970 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
9971 * config/arm/arm.opt (mrestrict-it): New command-line option.
9972 * doc/invoke.texi: Document -mrestrict-it.
9973
34c136b6
DM
99742013-06-05 David Malcolm <dmalcolm@redhat.com>
9975
9976 * tsan.c (tsan_atomic_table): Make const.
9977
a367df53
RB
99782013-06-05 Richard Biener <rguenther@suse.de>
9979
9980 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
9981 index associated with the tree we are supposed to replace.
9982 * tree-streamer-out.c (pack_ts_base_value_fields): Output
9983 TREE_ASM_WRITTEN as zero for everything but SSA names.
9984
70d51a19
DM
99852013-06-05 David Malcolm <dmalcolm@redhat.com>
9986
9987 * tree-ssa-structalias.c (call_stmt_vars): Make static.
9988
b9bd2075
JH
99892013-06-04 Jan Hubicka <jh@suse.cz>
9990
9991 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
9992 (input_node, input_varpool_node): Handle correctly external same
9993 body aliases.
9994 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
9995 nodes at ltrans stage.
9996
aaae719d
JH
99972013-06-04 Jan Hubicka <jh@suse.cz>
9998
9999 * ipa-inline.c (update_caller_keys): Fix availability test.
10000 (update_callee_keys): Likewise.
10001 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
10002 to follow ELF standard.
10003
107eea2c
JU
100042013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
10005
10006 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
10007 (mips64r5900el-*-elf*): New configurations.
10008 * config/mips/mips-cpus.def (r5900): New processor.
10009 * config/mips/mips-tables.opt: Regenerate.
10010 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
10011 (mips_issue_rate): Handle PROCESSOR_R5900.
10012 (mips_reorg_process_insns): Force reorder mode for the R5900.
10013 * config/mips/mips.h (TARGET_MIPS5900): Define.
10014 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
10015 TARGET_MIPS5900.
10016 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
10017 TARGET_MIPS5900.
10018 * config/mips/mips.md (processor): Add r5900.
10019 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
10020
b7342d25
IB
100212013-06-04 Ian Bolton <ian.bolton@arm.com>
10022
10023 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
10024 into function to generate MOVI instruction.
c1e183a9 10025 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB
10026 (aarch64_preferred_simd_mode): Turn into wrapper.
10027 (aarch64_output_scalar_simd_mov_immediate): New function.
10028 * config/aarch64/aarch64-protos.h: Add prototype for above.
10029
81c2dfb9
IB
100302013-06-04 Ian Bolton <ian.bolton@arm.com>
10031
10032 * config/aarch64/aarch64.c (simd_immediate_info): Remove
10033 element_char member.
10034 (sizetochar): Return signed char.
10035 (aarch64_simd_valid_immediate): Remove elchar and other
10036 unnecessary variables.
10037 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
10038 Calculate element_char as required.
10039 * config/aarch64/aarch64-protos.h: Update and move prototype
10040 for aarch64_output_simd_mov_immediate.
10041 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
10042 Update arguments.
10043
48063b9d
IB
100442013-06-04 Ian Bolton <ian.bolton@arm.com>
10045
10046 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
10047 information completed by aarch64_simd_valid_immediate.
10048 (aarch64_legitimate_constant_p): Update arguments.
10049 (aarch64_simd_valid_immediate): Work with struct rather than many
10050 pointers.
10051 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
10052 (aarch64_simd_make_constant): Update arguments.
10053 (aarch64_output_simd_mov_immediate): Work with struct rather than
10054 many pointers. Output immediate directly rather than as operand.
10055 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10056 Update prototype.
10057 * config/aarch64/constraints.md (Dn): Update arguments.
10058
3ea63f60
IB
100592013-06-04 Ian Bolton <ian.bolton@arm.com>
10060
10061 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
10062 longer static.
10063 (aarch64_simd_immediate_valid_for_move): Remove.
10064 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
10065 (aarch64_simd_make_constant): Update call.
10066 (aarch64_output_simd_mov_immediate): Update call.
10067 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
10068 Add prototype.
10069 * config/aarch64/constraints.md (Dn): Update call.
10070
d8edd899
IB
100712013-06-04 Ian Bolton <ian.bolton@arm.com>
10072
10073 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
10074 return type to bool for prototype.
10075 (aarch64_legitimate_constant_p): Check for true instead of not -1.
10076 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
10077 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
10078
44b20bb8
CM
100792013-06-04 Catherine Moore <clm@codesourcery.com>
10080
10081 * config/mips/mips.opt (meva): New.
c1e183a9 10082 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM
10083 (ASM_SPEC): Handle -meva.
10084 * doc/invoke.texi (meva): Document.
10085
52befbd8
AM
100862013-06-04 Alan Modra <amodra@gmail.com>
10087
10088 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
10089 constant output.
10090
aadaf24e
KT
100912013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10092
10093 * rtl.def: Add extra fourth optional field to define_cond_exec.
10094 * gensupport.c (process_one_cond_exec): Process attributes from
10095 define_cond_exec.
10096 * doc/md.texi: Document fourth field in define_cond_exec.
10097
0cd9e9ee
EB
100982013-06-04 Eric Botcazou <ebotcazou@adacore.com>
10099
10100 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
10101 out the processing order as in store_bit_field_1.
10102
cb0f665d
JH
101032013-06-04 Jan Hubicka <jh@suse.cz>
10104
10105 PR middle-end/57500
10106 * cgraphunit.c (cgraph_process_same_body_aliases): Create
10107 non-VAR_DECL node if it does not exist yet.
10108
53984b9b
RS
101092013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
10110
10111 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
10112 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
10113 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
10114 target_cpu_default setting.
10115
5d318fd4
TJ
101162013-06-03 Teresa Johnson <tejohnson@google.com>
10117
10118 * dumpfile.c (opt_info_switch_p): Change -fopt-info
10119 default to -fopt-info=optimized instead of all.
10120 * doc/invoke.texi: Ditto.
10121 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
10122 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
10123 (execute_vect_slp): Emit BB vectorization success under
10124 MSG_OPTIMIZED_LOCATIONS.
10125 * tree-vect-slp.c (vect_slp_transform_bb): Change
10126 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
10127 * tree-vect-loop.c (vect_transform_loop): Ditto.
10128
ec3af349
JM
101292013-06-03 Jason Merrill <jason@redhat.com>
10130
10131 PR c++/57415
10132 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
10133 Use TARGET_EXPR for C++.
10134
1baf9159
JJ
101352013-06-03 Jakub Jelinek <jakub@redhat.com>
10136
10137 PR rtl-optimization/57268
0cd9e9ee 10138 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ
10139 if DEBUG_INSN_P (insn).
10140
10141 Reapply
10142 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10143
10144 PR rtl-optimization/57268
10145 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10146 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10147
0f1d3965
YR
101482013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10149
10150 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
10151 (ix86_avoid_lea_for_addr): Likewise.
10152 (exact_dependency_1): Likewise.
10153 (ix86_adjust_cost): Likewise.
10154 (swap_top_of_ready_list): Fix formatting and !reload_completed check
10155 removed.
10156 (do_reorder_for_imul): Fix typo, formatting and
10157 !reload_completed check removed.
10158 (ix86_sched_reorder): Fix typo and formatting.
10159 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
10160 list.
10161
3bc30db4
SN
101622013-06-03 Sofiane Naci <sofiane.naci@arm.com>
10163
10164 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
10165
19cc6fac
EB
101662013-06-03 Eric Botcazou <ebotcazou@adacore.com>
10167
10168 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
10169 <STRING_CST>: Likewise.
10170 <VECTOR_CST>: Likewise.
10171
c68a6e08
JW
101722013-06-01 Janus Weil <janus@gcc.gnu.org>
10173 Mikael Morin <mikael@gcc.gnu.org>
10174
10175 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
10176 * config.in: Regenerated.
10177 * configure: Regenerated.
10178
38e55e5c
JH
101792013-06-01 Jan Hubicka <jh@suse.cz>
10180
10181 PR middle-end/57366
10182 * cgraphunit.c (compile): When weakref is not supported,
10183 set up transparent aliases before final output pass.
10184 * varasm.c (assemble_alias): Do not try to do it here.
10185
eb51d2ff
JH
101862013-06-01 Jan Hubicka <jh@suse.cz>
10187
10188 PR middle-end/57467
10189 * passes.c (for_per_function): Skip unanalyzed functions.
10190
40a7fe1e
JH
101912013-06-01 Jan Hubicka <jh@suse.cz>
10192
10193 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
10194 (lto_symtab_merge_symbols_1): ... this one.
10195 (lto_symtab_merge_cgraph_nodes): Rename to ...
10196 (lto_symtab_merge_symbols): ... this one; simplify.
10197 * cgraph.c (same_body_aliases_done): Rename to ...
10198 (cpp_implicit_aliases_done): ... this one.
10199 (cgraph_create_function_alias): Update.
10200 (cgraph_same_body_alias): Update.
c1e183a9 10201 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 10202 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 10203 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH
10204 (cgraph_node): Remove same_body_alias.
10205 (varpool_node): Remove alias_of and extra_name_alias.
10206 (same_body_aliases_done): Rename to ..
10207 (cpp_implicit_aliases_done): ... this one.
10208 (symtab_alias_ultimate_target): Add default parameter.
10209 (symtab_resolve_alias): New function.
10210 (fixup_same_cpp_alias_visibility): Declare.
10211 (cgraph_function_node): Add default parameter.
10212 (cgraph_node_asm_name): Likewise.
10213 (cgraph_function_or_thunk_node): Add default parameter; do
10214 not ICE when it is NULL.
10215 (varpool_variable_node): Likewise.
10216 * tree-emutls.c (create_emultls_var): Update.
10217 (ipa_lower_emutls): Update.
10218 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
10219 (cgraph_reset_node): Reset alias info.
10220 (cgraph_finalize_function): Update.
10221 (fixup_same_cpp_alias_visibility): Move to symtab.c.
10222 (analyze_function): Simplify.
10223 (cgraph_process_same_body_aliases): Simplify.
10224 (analyze_functions): Fixup same body aliases.
10225 (handle_alias_pairs): Simplify.
10226 (assemble_thunk): Update.
10227 (assemble_thunks_and_aliases): Update.
10228 (output_weakrefs): Rewrite.
10229 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
10230 (lto_output_varpool_node): Likewise.
10231 (compute_ltrans_boundary): Remve assert.
10232 (get_alias_symbol): New functoin.
10233 (input_node): Rewrite alias handling.
10234 (input_varpool_node): Likewise.
10235 * ipa-pure-const.c (propagate_pure_const): Fix formating.
10236 * ipa.c (process_references): Handle weakrefs correctly.
10237 (symtab_remove_unreachable_nodes): Likewise.
10238 * trans-mem.c (get_cg_data): Update.
10239 (ipa_tm_create_version_alias): Update.
10240 (ipa_tm_execute): Update.
10241 * symtab.c (dump_symtab_base): Dump aliases.
10242 (verify_symtab_base): Verify aliases.
10243 (symtab_node_availability): New function.
10244 (symtab_alias_ultimate_target): Simplify.
10245 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
10246 handle all the fixup cases.
10247 (symtab_resolve_alias): New function.
10248 * passes.c (ipa_write_summaries): Handle weakrefs.
10249 * varpool.c (varpool_analyze_node): Simplify.
10250 (assemble_aliases): Update.
10251 (varpool_create_variable_alias): Simplify.
10252 (varpool_extra_name_alias): Simplify.
10253 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
10254 (lto_symtab_merge_symbols): ... this one.
10255
21759881
DT
102562013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
10257
10258 Revert
10259 PR rtl-optimization/57268
10260 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10261 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10262
daf4e940
TB
102632013-06-01 Tobias Burnus <burnus@net-b.de>
10264
10265 Partially reverted:
10266 2013-05-31 Tobias Burnus <burnus@net-b.de>
10267
10268 PR middle-end/57073
10269 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
10270 further up.
10271
c359d8d0
DT
102722013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
10273
10274 PR rtl-optimization/57268
10275 * sched-deps.c (sched_analyze_2): Flush dependence lists if
10276 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
10277
4172215d
EB
102782013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10279
10280 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
10281 unordered comparison operators when -fno-trapping-math is in effect
10282 on the e500.
10283 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
10284 and implement unordered comparison operators properly on the e500.
10285
a8c50132
EB
102862013-05-31 Eric Botcazou <ebotcazou@adacore.com>
10287
10288 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
10289 for constant scalar integers.
10290 (simplify_relational_operation_1): Likewise.
10291
44626634
SB
102922013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
10293
10294 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
10295 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
10296 Fix comment.
10297
55805e54
YR
102982013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10299 Igor Zamyatin <igor.zamyatin@intel.com>
10300
10301 Silvermont (SLM) architecture performance tuning.
10302 * config/i386/i386.h (enum ix86_tune_indices): Add
10303 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
10304 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
10305
10306 * config/i386/i386.c (initial_ix86_tune_features)
10307 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
10308 (ix86_lea_outperforms): Handle Silvermont tuning.
10309 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
10310 call.
10311 (ix86_use_lea_for_mov): Likewise.
10312 (ix86_avoid_lea_for_addr): Likewise.
10313 (ix86_lea_for_add_ok): Likewise.
10314 (exact_dependency_1): New function.
10315 (exact_store_load_dependency): Likewise.
10316 (ix86_adjust_cost): Handle Silvermont tuning.
10317 (do_reoder_for_imul): Likewise.
10318 (swap_top_of_ready_list): New function.
10319 (ix86_sched_reorder): Changed to handle Silvermont tuning.
10320
10321 * config/i386/i386.md (peepholes that split memory operand in fp
10322 converts): New.
10323
e19c9de2
MS
103242013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10325
10326 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
10327 Remove un-necessary braces.
10328
38e6c9a6
MS
103292013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
10330
10331 * config/aarch64/aarch64.c (aarch64_classify_symbol):
10332 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
10333
e3530904
TB
103342013-05-31 Tobias Burnus <burnus@net-b.de>
10335
10336 PR middle-end/57073
c1e183a9 10337 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 10338
b6af05a9
KT
103392013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10340
10341 PR target/56315
10342 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
10343 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
10344 * config/arm/neon.md (iordi3_neon): Remove.
10345 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
10346 * config/arm/predicates.md (imm_for_neon_logic_operand):
10347 Move to earlier in the file.
10348 (neon_logic_op2): Likewise.
10349 (arm_iordi_operand_neon): New predicate.
10350
f800c166
RB
103512013-05-31 Richard Biener <rguenther@suse.de>
10352
10353 PR tree-optimization/57478
10354 PR tree-optimization/57453
10355 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
10356 are life as well.
10357
e571fa59
KP
103582013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
10359
10360 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
10361 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
10362
0fa6e0ef
TB
103632013-05-30 Tobias Burnus <burnus@net-b.de>
10364 Thomas Koenig <tkoenig@gcc.gnu.org>
10365
10366 PR middle-end/57073
10367 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
10368 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
10369
4688ddf5
SB
103702013-05-30 Steven Bosscher <steven@gcc.gnu.org>
10371
0fa6e0ef 10372 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 10373
3b9ceb4b
VM
103742013-05-30 Vladimir Makarov <vmakarov@redhat.com>
10375
10376 * target.def (register_usage_leveling_p): New hook.
10377 * targhooks.c (default_register_usage_leveling_p): New.
10378 * targhooks.h (default_register_usage_leveling_p): New prototype.
10379 * lra-assigns.c (register_usage_leveling_p): Use the hook.
10380 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
10381 * doc/tm.texi: Update.
10382 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
10383
65f9e789
IB
103842013-05-30 Ian Bolton <ian.bolton@arm.com>
10385
10386 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
10387 (*insv_reg<mode>): New define_insn.
10388
12211b99 103892013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR
10390
10391 PR rtl-optimization/57439
10392 * postreload.c (move2add_valid_value_p): Check that we have
10393 a zero subreg_regno_offset when accessing the register in
10394 the requested mode.
10395
0b871ccf
YR
103962013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
10397 Igor Zamyatin <igor.zamyatin@intel.com>
10398
10399 Silvermont (SLM) architecture pipeline model, tuning and
10400 insn selection.
10401 * config.gcc: Add slm config options and target.
10402
10403 * config/i386/slm.md: New.
10404
10405 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
10406
f43245d1 10407 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR
10408 PROCESSOR_SLM.
10409 (ix86_target_macros_internal): Likewise.
10410
f43245d1 10411 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR
10412 (m_SLM): New macro flag.
10413 (initial_ix86_tune_features): Set m_SLM.
10414 (x86_accumulate_outgoing_args): Likewise.
10415 (x86_arch_always_fancy_math_387): Likewise.
10416 (processor_target_table): Add slm cost.
10417 (cpu_names): Add slm cpu name.
10418 (x86_option_override_internal): Set SLM ISA.
10419 (ix86_issue_rate): New case PROCESSOR_SLM.
10420 (ia32_multipass_dfa_lookahead): Likewise.
10421 (fold_builtin_cpu): Add slm.
10422
10423 * config/i386/i386.h (TARGET_SLM): New target macro.
10424 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
10425 (processor_type): Add PROCESSOR_SLM.
10426
10427 * config/i386/i386.md (cpu): Add new value "slm".
10428 (slm.md): Include slm.md.
10429
24d5b097
XG
104302013-05-30 Bernd Schmidt <bernds@codesourcery.com>
10431 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10432
10433 * config/arm/arm-protos.h: Add and update function protos.
10434 * config/arm/arm.c (use_simple_return_p): New added.
10435 (thumb2_expand_return): Check simple_return flag.
10436 * config/arm/arm.md: Add simple_return and conditional simple_return.
10437 * config/arm/iterators.md: Add iterator for return and simple_return.
10438
c1cccc15
ZC
104392013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10440
10441 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
10442 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
10443 (arm_emit_vfp_multi_reg_pop): Likewise.
10444 (thumb2_emit_ldrd_pop): Likewise.
10445 (arm_expand_epilogue): Add misc REG_CFA notes.
10446 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
10447
3a4f280b
LC
104482013-05-29 Lawrence Crowl <crowl@google.com>
10449
10450 * config/arm/t-arm: Update for below.
10451
10452 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
10453 Change type to hash_table. Update dependent calls and types.
10454
10455 * config/i386/t-cygming: Update for below.
10456
10457 * config/i386/t-interix: Update for below.
10458
10459 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
10460 Change type to hash_table. Update dependent calls and types.
10461 (i386_find_on_wrapper_list::wrappers): Likewise.
10462
10463 * config/ia64/t-ia64: Update for below.
10464
10465 * config/ia64/ia64.c (bundle_state_table):
10466 Change type to hash_table. Update dependent calls and types.
10467
10468 * config/mips/mips.c (mips_reorg_process_insns::htab):
10469 Change type to hash_table. Update dependent calls and types.
10470
10471 * config/sol2.c (solaris_comdat_htab):
10472 Change type to hash_table. Update dependent calls and types.
10473
10474 * config/t-sol2: Update for above.
10475
1388a0e3
TJ
104762013-05-29 Teresa Johnson <tejohnson@google.com>
10477
10478 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
10479 functions are not yet marked as defined.
10480
a5965b52 104812013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
10482 Pat Haugen <pthaugen@us.ibm.com>
10483 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM
10484
10485 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
10486 instructions.
10487 (VEC_A): Likewise.
10488 (VEC_C): Likewise.
10489 (vrotl<mode>3): Likewise.
10490 (vashl<mode>3): Likewise.
10491 (vlshr<mode>3): Likewise.
10492 (vashr<mode>3): Likewise.
10493
10494 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
10495 support for power8 V2DI builtins.
10496
10497 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
10498 power8 V2DI builtins.
10499 (vupkhsw): Likewise.
10500 (vupklsw): Likewise.
10501 (vaddudm): Likewise.
10502 (vminsd): Likewise.
10503 (vmaxsd): Likewise.
10504 (vminud): Likewise.
10505 (vmaxud): Likewise.
10506 (vpkudum): Likewise.
10507 (vpksdss): Likewise.
10508 (vpkudus): Likewise.
10509 (vpksdus): Likewise.
10510 (vrld): Likewise.
10511 (vsld): Likewise.
10512 (vsrd): Likewise.
10513 (vsrad): Likewise.
10514 (vsubudm): Likewise.
10515 (vcmpequd): Likewise.
10516 (vcmpgtsd): Likewise.
10517 (vcmpgtud): Likewise.
10518 (vcmpequd_p): Likewise.
10519 (vcmpgtsd_p): Likewise.
10520 (vcmpgtud_p): Likewise.
10521 (vupkhsw): Likewise.
10522 (vupklsw): Likewise.
10523 (vaddudm): Likewise.
10524 (vmaxsd): Likewise.
10525 (vmaxud): Likewise.
10526 (vminsd): Likewise.
10527 (vminud): Likewise.
10528 (vpksdss): Likewise.
10529 (vpksdus): Likewise.
10530 (vpkudum): Likewise.
10531 (vpkudus): Likewise.
10532 (vrld): Likewise.
10533 (vsld): Likewise.
10534 (vsrad): Likewise.
10535 (vsrd): Likewise.
10536 (vsubudm): Likewise.
10537
10538 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
10539 support for power8 V2DI instructions.
10540
10541 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
10542 power8 V2DI instructions. Combine pack and unpack insns to use an
10543 iterator for each mode. Check whether a particular mode supports
10544 Altivec instructions instead of just checking TARGET_ALTIVEC.
10545 (UNSPEC_VPKUWUM): Likewise.
10546 (UNSPEC_VPKSHSS): Likewise.
10547 (UNSPEC_VPKSWSS): Likewise.
10548 (UNSPEC_VPKUHUS): Likewise.
10549 (UNSPEC_VPKSHUS): Likewise.
10550 (UNSPEC_VPKUWUS): Likewise.
10551 (UNSPEC_VPKSWUS): Likewise.
10552 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
10553 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
10554 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
10555 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
10556 (UNSPEC_VUPKHSB): Likewise.
10557 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
10558 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
10559 (UNSPEC_VUPKHSH): Likewise.
10560 (UNSPEC_VUPKLSB): Likewise.
10561 (UNSPEC_VUPKLSH): Likewise.
10562 (VI2): Likewise.
10563 (VI_char): Likewise.
10564 (VI_scalar): Likewise.
10565 (VI_unit): Likewise.
10566 (VP): Likewise.
10567 (VP_small): Likewise.
10568 (VP_small_lc): Likewise.
10569 (VU_char): Likewise.
10570 (add<mode>3): Likewise.
10571 (altivec_vaddcuw): Likewise.
10572 (altivec_vaddu<VI_char>s): Likewise.
10573 (altivec_vadds<VI_char>s): Likewise.
10574 (sub<mode>3): Likewise.
10575 (altivec_vsubcuw): Likewise.
10576 (altivec_vsubu<VI_char>s): Likewise.
10577 (altivec_vsubs<VI_char>s): Likewise.
10578 (altivec_vavgs<VI_char>): Likewise.
10579 (altivec_vcmpbfp): Likewise.
10580 (altivec_eq<mode>): Likewise.
10581 (altivec_gt<mode>): Likewise.
10582 (altivec_gtu<mode>): Likewise.
10583 (umax<mode>3): Likewise.
10584 (smax<mode>3): Likewise.
10585 (umin<mode>3): Likewise.
10586 (smin<mode>3): Likewise.
10587 (altivec_vpkuhum): Likewise.
10588 (altivec_vpkuwum): Likewise.
10589 (altivec_vpkshss): Likewise.
10590 (altivec_vpkswss): Likewise.
10591 (altivec_vpkuhus): Likewise.
10592 (altivec_vpkshus): Likewise.
10593 (altivec_vpkuwus): Likewise.
10594 (altivec_vpkswus): Likewise.
10595 (altivec_vpks<VI_char>ss): Likewise.
10596 (altivec_vpks<VI_char>us): Likewise.
10597 (altivec_vpku<VI_char>us): Likewise.
10598 (altivec_vpku<VI_char>um): Likewise.
10599 (altivec_vrl<VI_char>): Likewise.
10600 (altivec_vsl<VI_char>): Likewise.
10601 (altivec_vsr<VI_char>): Likewise.
10602 (altivec_vsra<VI_char>): Likewise.
10603 (altivec_vsldoi_<mode>): Likewise.
10604 (altivec_vupkhsb): Likewise.
10605 (altivec_vupkhs<VU_char>): Likewise.
10606 (altivec_vupkls<VU_char>): Likewise.
10607 (altivec_vupkhsh): Likewise.
10608 (altivec_vupklsb): Likewise.
10609 (altivec_vupklsh): Likewise.
10610 (altivec_vcmpequ<VI_char>_p): Likewise.
10611 (altivec_vcmpgts<VI_char>_p): Likewise.
10612 (altivec_vcmpgtu<VI_char>_p): Likewise.
10613 (abs<mode>2): Likewise.
10614 (vec_unpacks_hi_v16qi): Likewise.
10615 (vec_unpacks_hi_v8hi): Likewise.
10616 (vec_unpacks_lo_v16qi): Likewise.
10617 (vec_unpacks_hi_<VP_small_lc>): Likewise.
10618 (vec_unpacks_lo_v8hi): Likewise.
10619 (vec_unpacks_lo_<VP_small_lc>): Likewise.
10620 (vec_pack_trunc_v8h): Likewise.
10621 (vec_pack_trunc_v4si): Likewise.
10622 (vec_pack_trunc_<mode>): Likewise.
10623
10624 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
10625 V2DI builtins.
10626 (vec_vmaxsd): Likewise.
10627 (vec_vmaxud): Likewise.
10628 (vec_vminsd): Likewise.
10629 (vec_vminud): Likewise.
10630 (vec_vpksdss): Likewise.
10631 (vec_vpksdus): Likewise.
10632 (vec_vpkudum): Likewise.
10633 (vec_vpkudus): Likewise.
10634 (vec_vrld): Likewise.
10635 (vec_vsld): Likewise.
10636 (vec_vsrad): Likewise.
10637 (vec_vsrd): Likewise.
10638 (vec_vsubudm): Likewise.
10639 (vec_vupkhsw): Likewise.
10640 (vec_vupklsw): Likewise.
10641
e70670cf
JH
106422013-05-29 Jan Hubicka <jh@suse.cz>
10643
10644 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
10645 flags; reorder rest of fields in more consistent way.
10646 (varpool_node): Remove analyzed, finalized and alias.
10647 (cgraph_ndoe): Likewise.
10648 (symtab_alias_ultimate_target): New function.
10649 (cgraph_function_node): Move offline.
10650 (cgraph_reset_node): Declare.
10651 (cgraph_comdat_can_be_unshared_p): Remove.
10652 (varpool_remove_initializer): Declare.
10653 (varpool_first_defined_variable, varpool_next_defined_variable
10654 cgraph_first_defined_function, cgraph_next_defined_function): Update.
10655 (cgraph_function_with_gimple_body_p): Update.
10656 (varpool_all_refs_explicit_p): Update.
10657 (symtab_alias_target): New function.
10658 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
10659 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB
10660 (cgraph_function_or_thunk_node): Simplify using
10661 symtab_alias_ultimate_target.
e70670cf
JH
10662 (varpool_variable_node): Likewise.
10663 * cgraph.c (cgraph_create_function_alias): Update.
10664 (cgraph_add_thunk): Update.
10665 (cgraph_remove_node): Update.
10666 (dump_cgraph_node): Do not dump removed flags.
10667 (cgraph_function_body_availability): Update.
10668 (cgraph_propagate_frequency): Update.
10669 (verify_cgraph_node): Check sanity of local flag.
10670 (cgraph_function_node): Move here from cgraph.h; revamp for
10671 cgraph_function_or_thunk_node.
10672 * lto-symtab.c (lto_varpool_replace_node): Update.
10673 (lto_symtab_resolve_can_prevail_p): Update.
10674 (lto_symtab_merge_cgraph_nodes): Update.
10675 * ipa-cp.c (determine_versionability, initialize_node_lattices,
10676 propagate_constants_accross_call, devirtualization_time_bonus,
10677 ipcp_propagate_stage): Update.
10678 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB
10679 * ipa-inline-transform.c (clone_inlined_nodes,
10680 preserve_function_body_p): Update.
e70670cf
JH
10681 * ipa-reference.c (propagate): Update.
10682 (write_node_summary_p): Update.
10683 * toplev.c (wrapup_global_declaration_2): Update.
10684 * cgraphunit.c (cgraph_analyze_function): Rename to ...
10685 (analyze_function) ... this one.
10686 (cgraph_process_new_functions): Update.
10687 (cgraph_reset_node): Export.
10688 (cgraph_finalize_function): Update.
10689 (cgraph_add_new_function): Update.
10690 (process_function_and_variable_attributes): Update.
10691 (varpool_finalize_decl): Update.
10692 (symbol_finalized): Remove.
10693 (symbol_finalized_and_needed): Rename to ...
10694 (symbol_defined_and_needed): ... update.
10695 (cgraph_analyze_functions): Update.
10696 (handle_alias_pairs): Update.
10697 (mark_functions_to_output): Update.
10698 (assemble_thunk): Update.
10699 (output_in_order): Update.
10700 (output_weakrefs): Update.
10701 (finalize_compilation_unit): Update.
10702 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
10703 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
10704 input_node, input_varpool_node): Update.
10705 * dbxout.c (dbxout_expand_expr): Update.
10706 * cgraphclones.c (cgraph_clone_node): Update.
10707 (cgraph_copy_node_for_versioning): Update.
10708 (cgraph_materialize_clone): Update.
10709 (cgraph_materialize_all_clones): Update.
10710 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
10711 propagate_pure_const, propagate_nothrow): Update.
10712 * lto-streamer-out.c (lto_output, write_symbol): Update.
10713 * ipa-utils.c (ipa_reverse_postorder): Update.
10714 * ipa-inline.c (can_inline_edge_p): Update.
10715 (update_caller_keys, ipa_inline): Update.
10716 * dwarf2out.c (reference_to_unused,
10717 premark_types_used_by_global_vars_helper): Update.
10718 * tree-eh.c (tree_could_trap_p): Update.
10719 * ipa-split.c (consider_split, execute_split_functions): Update.
10720 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB
10721 has_addr_references_p): Update; move ahead in file for better
10722 readability.
e70670cf
JH
10723 (process_references): Simplify.
10724 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
10725 bodies are removed.
10726 (cgraph_comdat_can_be_unshared_p): Make static.
10727 (cgraph_externally_visible_p): Update.
10728 (varpool_externally_visible_p): Update.
10729 (function_and_variable_visibility): Update.
10730 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
10731 ipa_tm_mark_force_output_node): Update.
10732 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
10733 estimate_edge_devirt_benefit, inline_generate_summary,
10734 inline_write_summary): Update.
10735 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
10736 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB
10737 (ipa_print_node_params, ipa_prop_read_section,
10738 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH
10739 * varasm.c (mark_decl_referenced): Update.
10740 (assemble_alias, dump_tm_clone_pairs): Update.
10741 * tree-inline.c (copy_bb): Update.
10742 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
10743 Update.
10744 * symtab.c (dump_symtab_base): Print new flags.
10745 (verify_symtab_base): Verify new flags.
10746 (symtab_alias_ultimate_target): New function.
10747 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
10748 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
10749 Update.
c1e183a9
UB
10750 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
10751 Update.
e70670cf
JH
10752 * i386.c (ix86_get_function_versions_dispatcher,
10753 ix86_generate_version_dispatcher_body): Update.
10754 (fold_builtin_cpu): Use varpool_add_new_variable.
10755 * varpool.c (varpool_remove_initializer): Break out from ...
10756 (varpool_remove_node): ... this one.
10757 (dump_varpool_node, varpool_node_for_asm,
10758 cgraph_variable_initializer_availability, varpool_analyze_node,
10759 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB
10760 varpool_finalize_named_section_flags, varpool_create_variable_alias):
10761 Update.
e70670cf 10762
182802ad
JH
107632013-05-29 Jan Hubicka <jh@suse.cz>
10764
10765 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
10766
8190b609
ER
107672013-05-29 Easwaran Raman <eraman@google.com>
10768
10769 PR tree-optimization/57442
10770 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
10771 when control exits the main loop.
10772
69f5aa9b
SKS
107732013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
10774
10775 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
10776 and RX600.
c1e183a9 10777 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS
10778 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
10779 * rx/t-rx: Add rx100 under multi library matches option for nofpu
10780 option.
10781
4b847da9
BS
107822013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
10783
10784 PR tree-optimization/57441
10785 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
10786 Don't limit size of incr_vec to number of candidates.
10787
4fc43c39
SE
107882013-05-29 Steve Ellcey <sellcey@imgtec.com>
10789
10790 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
10791 and mips16 directories.
c1e183a9 10792 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE
10793 (MULTILIB_DIRNAMES): Ditto.
10794 (MULTILIB_EXCEPTIONS): Add new exceptions.
10795 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
10796 (MULTILIB_DIRNAMES): Ditto.
10797 (MULTILIB_EXCEPTIONS): Add new exceptions.
10798
12211b99 107992012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS
10800 Marcus Shawcroft <marcus.shawcroft@arm.com>
10801
10802 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
10803 SYMBOL_TINY_ABSOLUTE.
10804 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
10805 SYMBOL_TINY_ABSOLUTE.
10806 (aarch64_expand_mov_immediate): Likewise.
10807 (aarch64_classify_symbol): Likewise.
10808 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
10809 Permit SYMBOL_TINY_ABSOLUTE.
10810 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
10811
12211b99 108122013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS
10813 Marcus Shawcroft <marcus.shawcroft@arm.com>
10814
10815 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
10816 Refactor if/switch. Replace gcc_assert with if.
10817
c0186656
GG
108182013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
10819
10820 * config/i386/i386.c (initial_ix86_tune_features): Enable
10821 FP Reassociation for AMD bdver1 and bdver2.
10822
d20188f3
MJ
108232013-05-29 Martin Jambor <mjambor@suse.cz>
10824
10825 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
10826 and IMAGPART_EXPR do not occur within other handled_components.
10827
292cba13
RB
108282013-05-29 Richard Biener <rguenther@suse.de>
10829
10830 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
10831 access on whether the use is in the BB we currently try to
10832 vectorize.
10833 (vect_bb_vectorization_profitable_p): Pass the BB we currently
10834 vectorize to vect_bb_slp_scalar_cost.
10835
6eddf228
RB
108362013-05-29 Richard Biener <rguenther@suse.de>
10837
10838 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
10839 computing scalar cost offsetted by stmts that are kept live
10840 by scalar uses.
10841 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
10842 for computation of scalar cost.
10843
7df36117
SE
108442013-05-28 Steve Ellcey <sellcey@mips.com>
10845
10846 * config/mips/mips-cpus.def (mips32r2): Change processor type.
10847
36536d79
BI
108482013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
10849
10850 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
10851 array notation built-in reduction functions.
10852 * doc/passes.texi (Passes): Added documentation about changes done
10853 for Cilk Plus.
10854 * doc/invoke.texi (C Dialect Options): Added documentation about
10855 the -fcilkplus flag.
10856 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
10857 (BUILTINS_DEF): Depend on cilkplus.def.
10858 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
10859 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
10860 * cilkplus.def: New file.
10861
12211b99 108622013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR
10863
10864 PR rtl-optimization/57439
10865 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
10866
9daf6dbb
ER
108672013-05-28 Easwaran Raman <eraman@google.com>
10868
10869 PR tree-optimization/57337
10870 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
10871 (find_insert_point): Correctly identify the insertion point
10872 when two statements with the same UID is compared.
10873
fbd7e877
RB
108742013-05-28 Richard Biener <rguenther@suse.de>
10875
10876 PR tree-optimization/56787
10877 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
10878 from the list of data references.
10879 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
10880 clobbers.
10881 (vect_analyze_loop_operations): Likewise.
10882 (vect_transform_loop): Remove clobbers.
10883
bbba1117
MJ
108842013-05-28 Martin Jambor <mjambor@suse.cz>
10885
10886 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
10887 and REALPART_EXPRs have scalar type.
10888
bd388c2a
RB
108892013-05-28 Richard Biener <rguenther@suse.de>
10890
10891 PR tree-optimization/57411
10892 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
10893 virtual operands.
10894 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
10895 virtual operand propagation.
10896
2f56a311
EB
108972013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10898
10899 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
10900 destination register for bmasksi_vis.
10901 (vector_init_bshuffle): Likewise.
10902 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
10903
5c3eacbb
EB
109042013-05-28 Eric Botcazou <ebotcazou@adacore.com>
10905
10906 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
10907 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
10908 mode if the instruction isn't available in the original mode.
10909 * config/sparc/sparc.opt (mfix-ut699): New option.
10910 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
10911 (divdf3): Turn into expander.
10912 (divdf3_nofix): New insn.
10913 (divdf3_fix): Likewise.
10914 (divsf3): Disable if -mfix-ut699.
10915 (sqrtdf2): Turn into expander.
10916 (sqrtdf2_nofix): New insn.
10917 (sqrtdf2_fix): Likewise.
10918 (sqrtsf2): Disable if -mfix-ut699.
10919
a1756c0a
RB
109202013-05-27 Richard Biener <rguenther@suse.de>
10921
10922 PR middle-end/57412
10923 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
10924 block for the new loop.
10925
5a892248
RB
109262013-05-27 Richard Biener <rguenther@suse.de>
10927
10928 PR tree-optimization/57343
10929 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
10930 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
10931 (number_of_iterations_cond): Do not build the folded tree.
10932
d1de852b
RB
109332013-05-27 Richard Biener <rguenther@suse.de>
10934
10935 Revert
10936 PR middle-end/57381
10937 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
10938 OEP_CONSTANT_ADDRESS_OF retained.
10939
10940 PR tree-optimization/57417
10941 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
10942 for unchanged base.
10943 (set_ssa_val_to): Compare addresses using
10944 get_addr_base_and_unit_offset.
10945
12211b99 109462013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR
10947
10948 PR rtl-optimization/56833
10949 * postreload.c (move2add_record_mode): New function.
10950 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
10951 (move2add_use_add2_insn): Use move2add_record_sym_value.
10952 (move2add_use_add3_insn): Likewise.
10953 (reload_cse_move2add): Use move2add_valid_value_p and
10954 move2add_record_mode. Invalidate call-clobbered and REG_INC
10955 affected regs by setting reg_mode to VOIDmode.
10956 (move2add_note_store): Don't pretend the inside of a SUBREG is
10957 the actual destination. Invalidate single/leading registers by
10958 setting reg_mode to VOIDmode.
10959 Use move2add_record_sym_value, move2add_valid_value_p and
10960 move2add_record_mode.
10961
b03be25f
RB
109622013-05-27 Richard Biener <rguenther@suse.de>
10963
10964 PR tree-optimization/57396
10965 * tree-affine.c (double_int_constant_multiple_p): Properly
10966 return false for val == 0 and div != 0.
10967
44e88db2
RS
109682013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10969
10970 * config/mips/mips.h: Use #elif in preprocessor conditions.
10971
3b859704
RS
109722013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10973
10974 PR target/53916
10975 * config/mips/constraints.md (kl): New constraint.
10976 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
10977 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
10978 constraint for operand 0. Split after CSE for MIPS16. Emit a move
10979 from LO for MIPS16.
10980 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
10981
c979d5f5
RS
109822013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
10983
10984 PR target/55777
10985 * config/mips/mips.c (mips_can_inline_p): New function.
10986 (TARGET_CAN_INLINE_P): Define.
10987
8e90de43
SB
109882013-05-25 Steven Bosscher <steven@gcc.gnu.org>
10989
10990 * sched-int.h (ds_t, dw_t): Make unsigned int.
10991 Fix documentation that describes how all the ds_t bits are used.
10992 Reserve the last bit for delayed-branch scheduling.
10993 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
10994 (BITS_PER_DEP_WEAK): Fix definition and documentation.
10995 (gen_dep_weak_1): Remove prototype.
10996 * sched-deps.c (get_dep_weak_1): Make static.
10997 * target.def (speculate_insn, needs_block_p, gen_spec_check,
10998 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
10999 * doc/tm.texi: Regenerate.
11000 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
11001
cb5cb194
SB
110022013-05-24 Steven Bosscher <steven@gcc.gnu.org>
11003
11004 PR debug/56950
11005 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
11006
5bd93ff6
NS
110072013-05-24 Nathan Sidwell <nathan@codesourcery.com>
11008 Sandra Loosemore <sandra@codesourcery.com>
11009
11010 * config.gcc (powerpc-*): Allow native for with-cpu.
11011
2343af65
JL
110122013-05-24 Jeff Law <law@redhat.com>
11013
11014 PR tree-optimization/57124
11015 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
11016 conversion feeding a condition if the range has an overflow
11017 if -fstrict-overflow. Add warnings for when we do make the
11018 transformation.
11019
3ad695b9
DC
110202013-05-24 Dehao Chen <dehao@google.com>
11021
04960246 11022 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC
11023 (locus_discrim_hasher): Likewise.
11024 (locus_discrim_hasher::hash): Likewise.
11025 (locus_discrim_hasher::equal): Likewise.
11026
82338059
MJ
110272013-05-24 Martin Jambor <mjambor@suse.cz>
11028
11029 PR tree-optimization/57294
11030 * cgraph.h (ipa_record_stmt_references): Declare.
11031 * cgraphbuild.c (ipa_record_stmt_references): New function.
11032 (build_cgraph_edges): Use ipa_record_stmt_references.
11033 (rebuild_cgraph_edges): Likewise.
11034 (cgraph_rebuild_references): Likewise.
11035 * ipa-prop.c (ipa_modify_call_arguments): Discard references
11036 associated with the old statement and build references from the
11037 newly built statements.
11038 * ipa-ref.c (ipa_remove_stmt_references): New function.
11039 * ipa-ref.h (ipa_remove_stmt_references): Declare.
11040
1ccd4874
VM
110412013-05-24 Vladimir Makarov <vmakarov@redhat.com>
11042
55805e54 11043 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 11044 mem-mem moves.
55805e54 11045 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 11046 too.
55805e54 11047 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM
11048 secondary memory moves.
11049 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
11050 reg set up in the current insn.
11051
25e25c73
DC
110522013-05-24 Dehao Chen <dehao@google.com>
11053
04960246 11054 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC
11055 hash function.
11056 (locus_descrim_hasher::equal): Likewise.
04960246 11057 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC
11058 (make_edges): Likewise.
11059 (next_discriminator_for_locus): Likewise.
11060 (same_line_p): Likewise.
11061 (assign_discriminators): Likewise.
11062 (make_cond_expr_edges): Likewise.
11063 (make_gimple_switch_edges): Likewise.
11064 (make_goto_expr_edges): Likewise.
11065 (make_gimple_asm_edges): Likewise.
11066
50d38551
IB
110672013-05-24 Ian Bolton <ian.bolton@arm.com>
11068
11069 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
11070 X format specifier to only display bottom 16 bits.
11071 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
11072 immediate to match for operand 2, since it will be masked.
11073
aea0101d
RB
110742013-05-24 Richard Biener <rguenther@suse.de>
11075
11076 PR tree-optimization/57287
11077 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
11078 all SSA names that occur in abnormal PHIs.
11079
634e03d3
AI
110802013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
11081
11082 PR tree-ssa/57385
11083 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
11084 that index is not negative.
11085
b17c024f
EB
110862013-05-24 Eric Botcazou <ebotcazou@adacore.com>
11087
11088 PR rtl-optimization/55177
11089 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
11090 (simplify_byte_swapping_operation): New.
11091 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
11092 (simplify_relational_operation_1): Deal with BSWAP.
11093
46aeac1b
RH
110942013-05-23 Richard Henderson <rth@redhat.com>
11095
11096 PR target/56742
11097 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
11098 (ix86_reorg): Call it.
11099
70cc1536
UB
111002013-05-23 Uros Bizjak <ubizjak@gmail.com>
11101
11102 PR target/57379
11103 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
11104 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
11105 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
11106
b3851501
CB
111072013-05-23 Christian Bruel <christian.bruel@st.com>
11108
11109 PR debug/57351
11110 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
11111
12211b99 111122013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS
11113 Marcus Shawcroft <marcus.shawcroft@arm.com>
11114
11115 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
11116 * config/aarch64/constraints.md (Usa): Remove.
11117 * doc/md.texi (AArch64 Usa): Remove.
11118
12211b99 111192013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS
11120 Marcus Shawcroft <marcus.shawcroft@arm.com>
11121
11122 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
11123 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
11124 * config/aarch64/predicates.md (aarch64_const_address): Remove.
11125 (aarch64_mov_operand): Use aarch64_mov_operand_p.
11126
12211b99 111272013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP
11128
11129 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
11130 instruction (AdvSIMD).
11131 * config/aarch64/aarch64-builtins.c
11132 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
11133 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
11134
76c36cb1 111352013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ
11136
11137 PR middle-end/57347
11138 * tree.h (contains_bitfld_component_ref_p): Declare.
11139 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB
11140 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
11141 caller.
7d2fb524
MJ
11142 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
11143 not access a bit-field. Assert all final offsets are byte-aligned.
11144
ce521ae6
RB
111452013-05-23 Richard Biener <rguenther@suse.de>
11146
11147 PR tree-optimization/57380
11148 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
11149 least one invariant or re-used load.
11150 * passes.c (init_optimization_passes): Move pass_phiprop before
11151 pass_forwprop.
11152
75c7257f
JG
111532013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
11154
11155 * config/aarch64/aarch64-simd.md
11156 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
11157
af6d44b5
RB
111582013-05-23 Richard Biener <rguenther@suse.de>
11159
11160 PR middle-end/57381
11161 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
11162 OEP_CONSTANT_ADDRESS_OF retained.
11163
bd3647bf
JJ
111642013-05-23 Jakub Jelinek <jakub@redhat.com>
11165
11166 PR middle-end/57344
70cc1536
UB
11167 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
11168 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 11169
a55757ea
RB
111702013-05-23 Richard Biener <rguenther@suse.de>
11171
11172 PR rtl-optimization/57341
11173 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
11174 instead of true_dependence.
11175
cfb00b41
DM
111762013-05-22 David Malcolm <dmalcolm@redhat.com>
11177
dd1c676f
DM
11178 * bb-reorder.c (branch_threshold): Make const.
11179 (exec_threshold): Ditto.
cfb00b41 11180
f62511da 111812013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
11182 Pat Haugen <pthaugen@us.ibm.com>
11183 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
11184
11185 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
11186 documentation for the power8 crypto builtins.
11187
11188 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
11189
11190 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
11191 macros for defining power8 builtin functions.
11192 (BU_P8V_AV_2): Likewise.
11193 (BU_P8V_AV_P): Likewise.
11194 (BU_P8V_VSX_1): Likewise.
11195 (BU_P8V_OVERLOAD_1): Likewise.
11196 (BU_P8V_OVERLOAD_2): Likewise.
11197 (BU_CRYPTO_1): Likewise.
11198 (BU_CRYPTO_2): Likewise.
11199 (BU_CRYPTO_3): Likewise.
11200 (BU_CRYPTO_OVERLOAD_1): Likewise.
11201 (BU_CRYPTO_OVERLOAD_2): Likewise.
11202 (XSCVSPDP): Fix typo, point to the correct instruction.
11203 (VCIPHER): Add power8 crypto builtins.
11204 (VCIPHERLAST): Likewise.
11205 (VNCIPHER): Likewise.
11206 (VNCIPHERLAST): Likewise.
11207 (VPMSUMB): Likewise.
11208 (VPMSUMH): Likewise.
11209 (VPMSUMW): Likewise.
11210 (VPERMXOR_V2DI): Likewise.
11211 (VPERMXOR_V4SI: Likewise.
11212 (VPERMXOR_V8HI: Likewise.
11213 (VPERMXOR_V16QI: Likewise.
11214 (VSHASIGMAW): Likewise.
11215 (VSHASIGMAD): Likewise.
11216 (VPMSUM): Likewise.
11217 (VPERMXOR): Likewise.
11218 (VSHASIGMA): Likewise.
11219
11220 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
11221 __CRYPTO__ if the crypto instructions are available.
11222 (altivec_overloaded_builtins): Add support for overloaded power8
11223 builtins.
11224
11225 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
11226 support for power8 crypto builtins.
11227 (builtin_function_type): Likewise.
11228 (altivec_init_builtins): Add support for builtins that take vector
11229 long long (V2DI) arguments.
11230
11231 * config/rs6000/crypto.md: New file, define power8 crypto
11232 instructions.
11233
112342013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
11235 Pat Haugen <pthaugen@us.ibm.com>
11236 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
11237
11238 * doc/invoke.texi (Option Summary): Add power8 options.
11239 (RS/6000 and PowerPC Options): Likewise.
11240
11241 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
11242 constraints.md instead of rs6000.h. Reorder w* constraints. Add
11243 wm, wn, wr documentation.
11244
f43245d1 11245 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM
11246 registers if direct move instructions are enabled.
11247 (wn): New constraint for no registers.
11248 (wq): New constraint for quad word even GPR registers.
11249 (wr): New constraint if 64-bit instructions are enabled.
11250 (wv): New constraint if power8 vector instructions are enabled.
11251 (wQ): New constraint for quad word memory locations.
11252
f43245d1 11253 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM
11254 constraint for 0..15 for crypto instructions.
11255 (gpc_reg_operand): If VSX allow registers in VSX registers as well
11256 as GPR and floating point registers.
11257 (int_reg_operand): New predicate to match only GPR registers.
11258 (base_reg_operand): New predicate to match base registers.
11259 (quad_int_reg_operand): New predicate to match even GPR registers
11260 for quad memory operations.
11261 (vsx_reg_or_cint_operand): New predicate to allow vector logical
11262 operations in both GPR and VSX registers.
11263 (quad_memory_operand): New predicate for quad memory operations.
11264 (reg_or_indexed_operand): New predicate for direct move support.
11265
f43245d1 11266 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM
11267 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
11268 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
11269 (POWERPC_MASKS): Add power8 options.
11270 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
11271 various options.
11272
f43245d1 11273 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM
11274 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
11275
f43245d1 11276 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM
11277 (-mpower8-fusion): New power8 options.
11278 (-mpower8-fusion-sign): Likewise.
11279 (-mpower8-vector): Likewise.
11280 (-mcrypto): Likewise.
11281 (-mdirect-move): Likewise.
11282 (-mquad-memory): Likewise.
11283
f43245d1 11284 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM
11285 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
11286 registers.
70cc1536 11287 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM
11288 (rs6000_debug_vector_unit): Add p8_vector.
11289 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
11290 definitions. Also print fusion state.
11291 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
11292 (rs6000_builtin_mask_calculate): Add power8 builtin support.
11293 (rs6000_option_override_internal): Add support for power8.
11294 (rs6000_common_init_builtins): Add debugging for skipped builtins
11295 if -mdebug=builtin.
11296 (rs6000_adjust_cost): Add power8 support.
11297 (rs6000_issue_rate): Likewise.
11298 (insn_must_be_first_in_group): Likewise.
11299 (insn_must_be_last_in_group): Likewise.
11300 (force_new_group): Likewise.
11301 (rs6000_register_move_cost): Likewise.
11302 (rs6000_opt_masks): Likewise.
11303
11304 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
11305 power8 capable assembler, default to power7 options.
11306 (TARGET_DIRECT_MOVE): Likewise.
11307 (TARGET_CRYPTO): Likewise.
11308 (TARGET_P8_VECTOR): Likewise.
11309 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
11310 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
11311 (VECTOR_MEM_P8_VECTOR_P): Likewise.
11312 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
11313 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
11314 (TARGET_XSCVDPSPN): Likewise.
11315 (TARGET_XSCVSPDPN): Likewsie.
11316 (TARGET_SYNC_HI_QI): Likewise.
11317 (TARGET_SYNC_TI): Likewise.
11318 (MASK_CRYPTO): Likewise.
11319 (MASK_DIRECT_MOVE): Likewise.
11320 (MASK_P8_FUSION): Likewise.
11321 (MASK_P8_VECTOR): Likewise.
70cc1536
UB
11322 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
11323 temporary used by some of the direct move instructions to get two FP
11324 temporary registers does not force creation of a stack frame.
f62511da
MM
11325 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
11326 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
11327 that any VSX registers are tieable, even if they are also an
11328 Altivec vector mode.
11329 (r6000_reg_class_enum): Add wm, wr, wv constraints.
11330 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
11331 (RS6000_BTM_CRYPTO): Likewise.
11332 (RS6000_BTM_COMMON): Likewise.
11333
11334 * config/rs6000/rs6000.md (cpu attribute): Add power8.
11335 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
11336 (enum rs6000_vector): Add power8 vector support.
11337
73a1a707
RR
113382013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11339
11340 PR target/19599
11341 PR target/57340
11342 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
11343 (any_sibcall_could_use_r3): this and handle indirect calls.
11344 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
11345
d28073d4
BS
113462013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11347
11348 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
11349
bee0b10c
RB
113502013-05-22 Richard Biener <rguenther@suse.de>
11351
11352 PR middle-end/57349
11353 * profile.c (branch_prob): Do not split blocks that are
11354 abnormally receiving from ECF_RETURNS_TWICE functions.
11355
98409b51
RS
113562013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
11357
11358 * recog.c (offsettable_address_addr_space_p): Fix calculation of
11359 address mode. Move pointer mode initialization to the same place.
11360
c0602ab8
MZ
113612013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11362
11363 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
11364 while it has any effect.
11365
4f7a634e
ER
113662013-05-21 Easwaran Raman <eraman@google.com>
11367
11368 PR tree-optimization/57322
9daf6dbb
ER
11369 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
11370 UID of the statement added to the BB to be 1.
4f7a634e 11371
a4ce1258
JJ
113722013-05-21 Jakub Jelinek <jakub@redhat.com>
11373
11374 PR tree-optimization/57331
70cc1536
UB
11375 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
11376 of conversion from pointer type to integral type with integer.
a4ce1258 11377
1b14621a
MJ
113782013-05-21 Martin Jambor <mjambor@suse.cz>
11379
11380 PR lto/57289
11381 * ipa-prop.c (ipa_read_node_info): Process param_used and
11382 controlled_uses in the same order as when writing.
11383
e60661f0
MG
113842013-05-21 Magnus Granberg <baldrick@free.fr>
11385
11386 PR plugins/56754
ee49aa34 11387 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 11388
4a61be9a
RB
113892013-05-21 Richard Biener <rguenther@suse.de>
11390
11391 PR tree-optimization/57318
11392 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
11393 estimate stmts with side-effects as likely eliminated.
11394
c52da5f7
RB
113952013-05-21 Richard Biener <rguenther@suse.de>
11396
11397 PR tree-optimization/57330
11398 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
11399 preserve the call stmts fntype.
11400
7ec67e2a
RB
114012013-05-21 Richard Biener <rguenther@suse.de>
11402
11403 PR tree-optimization/57303
11404 * tree-ssa-sink.c (statement_sink_location): Improve killing
11405 stmt detection and properly handle self-assignments.
11406
b112d513
CB
114072013-05-21 Christian Bruel <christian.bruel@st.com>
11408
55805e54
YR
11409 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
11410 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB
11411 registers. Set register size out of the PARALLEL loop.
11412
14c2ec26
OE
114132013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
11414
11415 PR target/56547
11416 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
11417 (*fmasf4, *fmasf4_media): New insns.
11418
da734fa1
RS
114192013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11420
11421 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
11422 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
11423 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
11424 (mips_idiv_insns): Update the comments to say that the returned
11425 instruction counts are in units of BASE_INSN_LENGTH.
11426 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
11427 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
11428 using 2 rather than 4 as the length of indirect MIPS16 and
11429 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
11430 length of a NOP. Don't divide MIPS16 lengths by 2.
11431 (mips16_split_long_branches): Assume a branch is long if the
11432 length is greater than 4 rather than 8.
11433 * config/mips/mips.md (length): Give MIPS16 lengths directly,
11434 rather than multiplying them by 2. Multiply instruction counts
11435 by BASE_INSN_LENGTH rather than 4.
11436 (*jump_mips16, tls_get_tp_mips16_<mode>)
11437 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
11438
13719e8b
RS
114392013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11440
11441 * config/mips/mips.md (extended_mips16): Remove branch case.
11442 (length): Remove duplicated extended_mips16 test.
11443
c3850d14
RS
114442013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
11445
11446 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
11447
8da2e059
RS
114482013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
11449
11450 * recog.h (Recog_data): Rename to...
11451 (recog_data_d): ...this.
11452 (recog_data): Update accordingly.
11453 * recog.c (recog_data): Likewise.
11454 * reload.c (save_recog_data): Likewise.
11455 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
11456 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
11457
2df013f3
JB
114582013-05-17 Julian Brown <julian@codesourcery.com>
11459
11460 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
11461 found in a REG_EQUAL note, invalidate it.
11462
371e77e3 114632013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER
11464
11465 * tree-ssa-reassoc.c (find_insert_point): New function.
11466 (insert_stmt_after): Likewise.
11467 (get_def_stmt): Likewise.
11468 (ensure_ops_are_available): Likewise.
11469 (not_dominated_by): Likewise.
11470 (rewrite_expr_tree): Do not move statements beyond what is
11471 necessary. Remove call to swap_ops_for_binary_stmt...
11472 (reassociate_bb): ... and move it here.
11473 (build_and_add_sum): Assign UIDs for new statements.
11474 (linearize_expr): Likewise.
11475 (do_reassoc): Renumber gimple statement UIDs.
11476
e01c7cca
JH
114772013-05-17 Jan Hubicka <jh@suse.cz>
11478
11479 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
11480 weakrefs.
11481 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB
11482 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
11483 weakrefs.
e01c7cca
JH
11484 (output_weakrefs): Update.
11485
c3272a92
PCC
114862013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
11487 Martin Jambor <mjambor@suse.cz>
11488
11489 PR middle-end/57276
11490 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
11491 value that corresponds to the given aggval is found in values vector.
11492
11c2aa39
UB
114932013-05-17 Uros Bizjak <ubizjak@gmail.com>
11494
11495 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
11496 sse, sse2, sse3, ssse3 and sse4a flags to options.
11497
abecc8c6
DM
114982013-05-17 David Malcolm <dmalcolm@redhat.com>
11499
11500 * gengtype-state.c: (s_expr_writer): New class, to handle
11501 prettifying of output layout of s-expressions.
11502 (state_writer): New class, to write out gtype.state.
11503 (state_written_type_count): Move this variable into member data of
11504 state_writer.
11505 (s_expr_writer::s_expr_writer): New code: constructor for new class
11506 (state_writer::state_writer(): ditto
11507 (s_expr_writer::write_new_line): New function
11508 (s_expr_writer::write_any_indent): ditto
11509 (s_expr_writer::begin_s_expr): ditto
11510 (s_expr_writer::end_s_expr): ditto
11511 (write_state_fileloc): convert to method of state_writer...
11512 (state_writer:: write_state_fileloc): ...and use methods of
11513 s_expr_writer to write indentation into the gtype.state output file
11514 to visually represent the hierarchical structure of the list
11515 structures
11516 (write_state_fields): ditto, renaming to...
11517 (state_writer::write_state_fields)
11518 (write_state_a_string): ditto, renaming to...
11519 (state_writer::write_state_a_string)
11520 (write_state_string_option): ditto, renaming to...
11521 (state_writer::write_state_string_option)
11522 (write_state_type_option): ditto, renaming to...
11523 (state_writer::write_state_type_option)
11524 (write_state_nested_option): ditto, renaming to...
11525 (state_writer::write_state_nested_option)
11526 (write_state_option): ditto, renaming to...
11527 (state_writer::write_state_option)
11528 (write_state_options): ditto, renaming to...
11529 (state_writer::write_state_options)
11530 (write_state_lang_bitmap): ditto, renaming to...
11531 (state_writer::write_state_lang_bitmap)
11532 (write_state_version): ditto, renaming to...
11533 (state_writer::write_state_version)
11534 (write_state_scalar_type): ditto, renaming to...
11535 (state_writer::write_state_scalar_type)
11536 (write_state_string_type): ditto, renaming to...
11537 (state_writer::write_state_string_type)
11538 (write_state_undefined_type): ditto, renaming to...
11539 (state_writer::write_state_undefined_type)
11540 (write_state_struct_union_type): ditto, renaming to...
11541 (state_writer::write_state_struct_union_type)
11542 (write_state_struct_type): ditto, renaming to...
11543 (state_writer::write_state_struct_type)
11544 (write_state_user_struct_type): ditto, renaming to...
11545 (state_writer::write_state_user_struct_type)
11546 (write_state_lang_struct_type): ditto, renaming to...
11547 (state_writer::write_state_lang_struct_type)
11548 (write_state_param_struct_type): ditto, renaming to...
11549 (state_writer::write_state_param_struct_type)
11550 (write_state_pointer_type): ditto, renaming to...
11551 (state_writer::write_state_pointer_type)
11552 (write_state_array_type): ditto, renaming to...
11553 (state_writer::write_state_array_type)
11554 (write_state_gc_used): ditto, renaming to...
11555 (state_writer::write_state_gc_used)
11556 (write_state_common_type_content): ditto, renaming to...
11557 (state_writer::write_state_common_type_content)
11558 (write_state_type): ditto, renaming to...
11559 (state_writer::write_state_type)
11560 (write_state_pair_list): ditto, renaming to...
11561 (state_writer::write_state_pair_list)
11562 (write_state_pair): ditto, renaming to...
11563 (state_writer::write_state_pair)
11564 (write_state_typedefs): ditto, renaming to...
11565 (state_writer::write_state_typedefs)
11566 (write_state_structures): ditto, renaming to...
11567 (state_writer::write_state_structures)
11568 (write_state_param_structs): ditto, renaming to...
11569 (state_writer::write_state_param_structs)
11570 (write_state_variables): ditto, renaming to...
11571 (state_writer::write_state_variables)
11572 (write_state_srcdir): ditto, renaming to...
11573 (state_writer::write_state_srcdir)
11574 (write_state_files_list): ditto, renaming to...
11575 (state_writer::write_state_files_list)
11576 (write_state_languages): ditto, renaming to...
11577 (state_writer::write_state_languages)
11578 (write_state): create a state_writer instance and use it when
11579 writing out the state file
11580
d6545f29
MS
115812013-05-17 Mike Stump <mikestump@comcast.net>
11582
816a3f73 11583 PR rtl-optimization/57304
d6545f29
MS
11584 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
11585 accessing DF_REF_REAL_LOC.
11586
38c821cf
JJ
115872013-05-17 Jakub Jelinek <jakub@redhat.com>
11588
11589 PR rtl-optimization/57281
11590 PR rtl-optimization/57300
11591 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
11592 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
11593 what the other splitter did if the registers are dead.
11594
2a293391
RB
115952013-05-17 Richard Biener <rguenther@suse.de>
11596
11597 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
11598 MEM_REF offsets.
11599
17042d2b
JJ
116002013-05-17 Jakub Jelinek <jakub@redhat.com>
11601
11602 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
11603 linking.
11604
5b115c1f
MP
116052013-05-17 Marek Polacek <polacek@redhat.com>
11606
11c2aa39
UB
11607 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
11608 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 11609
52d84413
JJ
116102013-05-17 Jakub Jelinek <jakub@redhat.com>
11611
68119618
JJ
11612 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
11613 vect_external_def oprnd1 with loop_vinfo, try to emit
11614 optional cast, negation and and stmts on the loop preheader
11615 edge instead of into the pattern def seq.
11616
52d84413
JJ
11617 PR tree-optimization/57051
11618 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
11619 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
11620
17b962bd
NC
116212013-05-16 Nick Clifton <nickc@redhat.com>
11622
11623 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
11624 (rl78_is_naked_func): New function.
56aefbf7
UB
11625 (rl78_expand_prologue): Skip prologue generation for naked functions.
11626 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC
11627 * doc/extend.texi (naked): Add RL78 to the list of processors
11628 that supports this attribute.
11629
b1a0f84e
JL
116302013-05-16 Jeff Law <law@redhat.com>
11631
11632 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
11633
af0e415b
UB
116342013-05-16 Uros Bizjak <ubizjak@gmail.com>
11635
11636 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
11637 cache parameters using detect_caches_amd also for CYRIX,
11638 NSC and TM2 signatures.
11639
19db293a
UB
116402013-05-16 Uros Bizjak <ubizjak@gmail.com>
11641 Dzianis Kahanovich <mahatma@eu.by>
11642
11643 PR target/45359
11644 PR target/46396
11645 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11646 VIA/Centaur processors and determine their cache parameters
11647 using detect_caches_amd.
11648
251a41b9
TJ
116492013-05-16 Teresa Johnson <tejohnson@google.com>
11650
11651 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
11652 (rtl_verify_edges): New function.
11653 (rtl_verify_bb_insns): Ditto.
11654 (rtl_verify_bb_pointers): Ditto.
11655 (rtl_verify_bb_insn_chain): Ditto.
11656 (rtl_verify_fallthru): Ditto.
11657 (rtl_verify_bb_layout): Ditto.
11658 (rtl_verify_flow_info_1): Outline checks into new functions.
11659 (rtl_verify_flow_info): Ditto.
11660
f14540b6
SE
116612013-05-16 Steve Ellcey <sellcey@imgtec.com>
11662
11663 * cfghooks.c (copy_bbs): Add update_dominance argument.
11664 * cfghooks.h (copy_bbs): Update prototype.
11665 * tree-cfg.c (gimple_duplicate_sese_region):
11666 Add update_dominance argument.
11667 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
11668 * tree-ssa-loop-ch.c (copy_loop_headers): Update
11669 gimple_duplicate_sese_region call.
11670 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
11671 Update copy_bbs call.
11672 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
11673 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
11674
7e9a3abb
JJ
116752013-05-16 Jakub Jelinek <jakub@redhat.com>
11676
11677 * tree-vectorizer.h (NUM_PATTERNS): Increment.
11678 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
11679 vect_recog_rotate_pattern.
11680 (vect_recog_rotate_pattern): New function.
11681
427b248d
JM
116822013-05-16 Jason Merrill <jason@redhat.com>
11683
11684 * Makefile.in (LLINKER): New variable.
11685 (mostlyclean): Remove link mutex.
11686 * configure.ac: Handle --enable-link-mutex.
11687 * lock-and-run.sh: New script.
11688
b871e3d2
RR
116892013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11690
11691 PR target/19599
11692 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
11693 for NULL decl.
11694
ce858126
RO
116952013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11696
11697 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
11698
798d3d04
GY
116992013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
11700
11701 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
11702 * config/arm/arm.c (next_consecutive_mem): New function.
11703 (gen_movmem_ldrd_strd): Likewise.
11704 * config/arm/arm.md (movmemqi): Update condition and code.
11705 (unaligned_loaddi, unaligned_storedi): New patterns.
11706
0baddc45
RO
117072013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11708
11709 * config.gcc: Obsolete *-*-solaris2.9*.
11710 * doc/install.texi (Specific, *-*-solaris2*): Document it.
11711
aa9d5bcf
RB
117122013-05-16 Richard Biener <rguenther@suse.de>
11713
11714 * passes.c (init_optimization_passes): Move pass_parallelize_loops
11715 earlier, after GRAPHITE transforms and IV canonicalization.
11716
5a0f4dd3
JJ
117172013-05-16 Jakub Jelinek <jakub@redhat.com>
11718
11719 * omp-low.c (extract_omp_for_data): For collapsed loops,
11720 if at least one of the loops is known at compile time to
11721 iterate zero times, set count to 0.
11722 (expand_omp_regimplify_p): New function.
11723 (expand_omp_for_generic): For collapsed loops, if at least
11724 one of the loops isn't known to iterate at least once,
11725 add runtime check with setting count to 0.
11726 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
11727 For unsigned types if it isn't known at compile time that
11728 the loop will iterate at least once, add runtime check to bypass
11729 the whole loop if initial condition isn't true.
11730
e3753785
NS
117312013-05-16 Nathan Sidwell <nathan@codesourcery.com>
11732
11733 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
11734
43bb4dd1
MG
117352013-05-16 Marc Glisse <marc.glisse@inria.fr>
11736
11737 PR middle-end/57286
11738 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
11739 transformations to avoid an infinite loop.
11740
3571dde6
MP
117412013-05-16 Marek Polacek <polacek@redhat.com>
11742
11743 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
11744
b8b3f0ca
LE
117452013-05-15 Leif Ekblad <leif@rdos.net>
11746
11747 * config/i386/i386.c (ix86_decompose_address): Use
11748 DEFAULT_TLS_SEG_REG to access TLS segment register.
11749 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
11750 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
11751 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
11752
e299a383
RS
117532013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
11754
11755 PR target/57260
11756 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
11757 sibling calls to functions that would normally be lazily bound,
11758 unless $gp is call-clobbered.
11759
e7208ea3
UB
117602013-05-15 Uros Bizjak <ubizjak@gmail.com>
11761
19db293a 11762 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 11763 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB
11764 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
11765 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB
11766 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
11767 of TARGET_3DNOW.
11768 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
11769
e7413f3d
AS
117702013-05-15 Andreas Schwab <schwab@suse.de>
11771
11772 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
11773 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
11774 third operand.
11775
ad4db775
TJ
117762013-05-15 Teresa Johnson <tejohnson@google.com>
11777
11778 * loop-unroll.c (report_unroll_peel): Check decision before
11779 emitting unroll/peel message.
11780
af205f67
TJ
117812013-05-15 Teresa Johnson <tejohnson@google.com>
11782
11783 * function.h (has_bb_partition): New rtl_data flag.
11784 (bb_reorder_complete): Ditto.
11785 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
11786 instead of flag_reorder_blocks_and_partition.
11787 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
11788 with some enhancements.
11789 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
11790 * bb-reorder.c (connect_traces): Check for has_bb_partition
11791 instead of flag_reorder_blocks_and_partition.
11792 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
11793 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
11794 verify_hot_cold_block_grouping.
11795 (partition_hot_cold_basic_blocks): Set has_bb_partition.
11796
9adcfa3c
RR
117972013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11798
11799 PR target/19599
11800 * config/arm/predicates.md (call_insn_operand): New predicate.
11801 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
11802 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
11803 if insn is not a tail call.
11804 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
11805 registers.
11806 * config/arm/arm.h (enum reg_class): New caller save register class.
11807 (REG_CLASS_NAMES): Likewise.
11808 (REG_CLASS_CONTENTS): Likewise.
11809 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
11810 without decls.
11811
ccb3ad87
RB
118122013-05-15 Richard Biener <rguenther@suse.de>
11813
11814 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
11815 of MSG_OPTIMIZED_LOCATIONS.
11816 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
11817 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
11818 message.
11819 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
11820 of MSG_OPTIMIZED_LOCATIONS.
11821 (execute_vect_slp): Likewise.
11822 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
11823 (vect_create_cond_for_alias_checks): Likewise.
11824 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
11825 (vect_recog_widen_mult_pattern): Likewise.
11826 (vect_recog_widen_sum_pattern): Likewise.
11827 (vect_recog_over_widening_pattern): Likewise.
11828 (vect_recog_widen_shift_pattern): Likewise.
11829 (vect_recog_vector_vector_shift_pattern): Likewise.
11830 (vect_recog_divmod_pattern): Likewise.
11831 (vect_recog_mixed_size_cond_pattern): Likewise.
11832 (vect_recog_bool_pattern): Likewise.
11833 (vect_pattern_recog_1): Likewise.
11834
48b1474e
MJ
118352013-05-15 Martin Jambor <mjambor@suse.cz>
11836
11837 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
11838 non-functions to builtin_unreachable.
11839 * ipa-inline-transform.c (inline_call): Do not assert estimates were
11840 correct when new direct edges were discovered.
11841
9de04252
MJ
118422013-05-15 Martin Jambor <mjambor@suse.cz>
11843
11844 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
11845 header, print symbol order instead of node uid, print more information
11846 about indirect edge targets.
11847 (ipa_make_edge_direct_to_target): Print symbol order instead of node
11848 uids.
11849 (ipa_make_edge_direct_to_target): Likewise.
11850 (remove_described_reference): Likewise.
11851 (propagate_controlled_uses): Likewise.
11852 (ipa_print_node_params): Also print symbol order.
11853 (ipcp_transform_function): Print symbol order instead of node uids.
11854 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
11855 (cgraph_get_create_real_symbol_node): Likewise.
11856 * ipa-cp.c (print_lattice): Likewise.
11857 (print_all_lattices): Likewise.
11858 (determine_versionability): Likewise.
11859 (initialize_node_lattices): Likewise.
11860 (estimate_local_effects): Likewise.
11861 (update_profiling_info): Likewise.
11862 (create_specialized_node): Likewise.
11863 (perhaps_add_new_callers): Likewise.
11864 (decide_about_value): Likewise.
11865 (decide_whether_version_node): Likewise.
11866 (identify_dead_nodes): Likewise.
11867 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
11868 (dump_inline_summary): Likewise.
11869 (estimate_node_size_and_time): Likewise.
11870 (inline_analyze_function): Likewise.
11871 * ipa-inline.c (report_inline_failed_reason): Likewise.
11872 (want_early_inline_function_p): Likewise.
11873 (edge_badness): Likewise.
11874 (update_edge_key): Likewise.
11875 (inline_small_functions): Likewise. Add dumping of order to two other
11876 dumps.
11877 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
11878 instead of node uids.
11879 (propagate_pure_const): Likewise.
11880 (propagate_pure_const): Likewise.
11881 * ipa-utils.c (dump_cgraph_node_set): Likewise.
11882 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
11883 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
11884 of node uids.
11885 * tree-pretty-print.c (dump_function_header): Likewise.
11886 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
11887 Print symbol order instead of node uids.
11888
1dd03b91
AK
118892013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11890
11891 * config/s390/s390.c (s390_register_move_cost): Don't impose the
11892 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
11893
fdf6a7b9
RB
118942013-05-15 Richard Biener <rguenther@suse.de>
11895
11896 PR tree-optimization/57275
11897 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
11898 return value for fail to do runtime alias checks for gather loads.
11899
2d6e4603
JH
119002013-05-15 Jan Hubicka <jh@suse.cz>
11901
11902 PR lto/57038
11903 PR lto/47375
e7208ea3
UB
11904 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
11905 weakrefs are not external.
11906 (lto_symtab_merge_decls): Fix thinko when dealing with
11907 non-lto_symtab decls.
2d6e4603
JH
11908 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
11909 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
11910 * varpool.c (dump_varpool_node): Dump more flags.
11911
83f44b39
GG
119122013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
11913
11914 * config/i386/i386.c (processor_alias_table): Add instruction
11915 FSGSBASE for AMD bdver3 architecture.
11916
2e55d062
JJ
119172013-05-14 Jakub Jelinek <jakub@redhat.com>
11918
11919 * tree.c (warn_deprecated_use): Print file:line using locus color.
11920 * diagnostic.c (diagnostic_report_current_module): Print file:line
11921 and file:line:column using locus color.
11922
0bfdb81e
MS
119232013-05-14 Mike Stump <mikestump@comcast.net>
11924
11925 * gdbinit.in: Add __null.
11926
a508ef22
MS
119272013-05-14 Mike Stump <mikestump@comcast.net>
11928
11929 * recog.h: Rename struct recog_data to Recog_data.
11930 * recog.c: Likewise.
11931 * reload.c (can_reload_into): Likewise.
11932 * config/picochip/picochip.c: Likewise.
11933
e7180acb
MS
119342013-05-14 Mike Stump <mikestump@comcast.net>
11935
11936 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
11937
61216c26
SB
119382013-05-14 Steven Bosscher <steven@gcc.gnu.org>
11939
fd6beed4
SB
11940 * resource.h (struct resources): Remove unch_memory member.
11941 (CLEAR_RESOURCE): Don't clear unch_memory.
11942 * resource.c (mark_referenced_resources): Don't set it.
11943 (mark_set_resources): Likewise.
11944 (mark_target_live_regs): Don't clear it.
11945 (init_resource_info): Likewise.
11946 * reorg.c (resource_conflicts_p): Don't compare it.
11947 (redundant_insn): Don't set it.
11948
61216c26
SB
11949 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
11950 Remove prototypes.
11951 * emit-rtl.c (next_label): Remove unused function.
11952 (skip_consecutive_labels, link_cc0_insns): Move to ...
11953 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
11954 only place where these functions are used, and make them static.
11955
418d1b87
MG
119562013-05-14 Marc Glisse <marc.glisse@inria.fr>
11957
11958 * fold-const.c (fold_negate_expr): Handle vectors.
11959 (fold_truth_not_expr): Make it static.
11960 (fold_invert_truthvalue): New static function.
11961 (invert_truthvalue_loc): Handle vectors. Do not call
11962 fold_truth_not_expr directly.
11963 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
11964 <TRUTH_NOT_EXPR>: Do not cast to boolean.
11965 (fold_comparison): Handle vector constants.
11966 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
11967 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
11968 * tree.h (fold_truth_not_expr): Remove declaration.
11969
fc21784d
JG
119702013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
11971
11972 * config/aarch64/aarch64-simd.md
11973 (aarch64_vcond_internal<mode>): Rename to...
11974 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
11975 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
11976 float modes. Clarify all iterator modes.
11977 (vcond<mode><mode>): Use new name for vcond expanders.
11978 (vcond<v_cmp_result><mode>): Likewise.
11979 (vcondu<mode><mode>: Likewise.
11980 * config/aarch64/iterators.md (VDQF_COND): New.
11981
d4c52634
MG
119822013-05-14 Marc Glisse <marc.glisse@inria.fr>
11983
11984 PR bootstrap/57266
11985 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
11986 variable for the shift amount. Check that we shift by non-negative
11987 amounts.
11988
2b261262
CLT
119892013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
11990
11991 PR target/42017
11992 * config/arm/arm.h (EPILOGUE_USES): Only return true
11993 for LR_REGNUM after epilogue_completed.
11994
12211b99 119952013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR
11996
11997 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 11998 is error_mark_node.
6e022d7b 11999
56cf7859
RO
120002013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12001
12002 PR target/57261
12003 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
12004 and Solaris 11+/x86 with gld.
12005 * configure: Regenerate.
12006
75776c6d
JJ
120072013-05-14 Jakub Jelinek <jakub@redhat.com>
12008
12009 * expmed.c (expand_shift_1): Canonicalize rotates by
12010 constant bitsize / 2 to bitsize - 1.
7f998021 12011 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ
12012 case ROTATERT>: Likewise.
12013
12014 Revert:
12015 2013-05-10 Jakub Jelinek <jakub@redhat.com>
12016
12017 * config/i386/i386.md (rotateinv): New code attr.
12018 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12019 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12020 roll $31, %eax, etc.
12021
df35498a
RB
120222013-05-14 Richard Biener <rguenther@suse.de>
12023
12024 PR middle-end/57235
12025 * tree-eh.c (sink_clobbers): Give up for successors with
12026 multiple predecessors and no virtual uses.
12027
cc6e7ece
EB
120282013-05-14 Eric Botcazou <ebotcazou@adacore.com>
12029
12030 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
12031 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
12032
f778dd4d
JJ
120332013-05-14 Jakub Jelinek <jakub@redhat.com>
12034
12035 PR middle-end/57251
12036 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
12037 the case when both op0 and op1 have VOIDmode.
12038
bad4df9b
KP
120392013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
12040
12041 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
12042 in multiply-accumulate mode.
12043
56f3e9ac
GW
120442013-05-13 Guozhi Wei <carrot@google.com>
12045
12046 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
12047
a3d7ab92
KT
120482013-05-13 Kai Tietz <ktietz@redhat.com>
12049
12050 PR target/56975
12051 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 12052 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 12053 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 12054 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT
12055 (ix86_expand_prologue): Likewise.
12056 (ix86_expand_split_stack_prologue): Likewise.
12057 (legitimate_pic_address_disp_p): Likewise.
12058 (legitimize_pic_address): Likewise.
12059 (legitimize_tls_address): Likewise.
12060 (legitimize_pe_coff_symbol): Likewise.
12061 (output_pic_addr_const): Likewise.
12062 (construct_plt_address): Likewise.
12063 (ix86_expand_call): Likewise.
12064 (x86_output_mi_thunk): Likewise.
12065 (x86_function_profiler): Likewise.
12066
c59b7e28
SN
120672013-05-13 Sofiane Naci <sofiane.naci@arm.com>
12068
12069 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
12070 similar switch cases.
12071 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
12072 (aarch64_simd_mov_to_<mode>low): Delete.
12073 (aarch64_simd_mov_to_<mode>high): Delete.
12074 (move_lo_quad_<mode>): Add w<-r alternative.
12075 (aarch64_simd_move_hi_quad_<mode>): Likewise.
12076 (aarch64_simd_mov_from_*): Update type attribute.
12077 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
12078 statement.
12079
5f28524a
JH
120802013-05-13 Jan Hubicka <jh@suse.cz>
12081
12082 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
12083 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB
12084 ix86_expand_epilogue, emit_i387_cw_initialization,
12085 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
12086 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 12087
ae6fa899
JJ
120882013-05-13 Jakub Jelinek <jakub@redhat.com>
12089
12090 PR tree-optimization/45216
12091 PR tree-optimization/57157
12092 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
12093 the (-Y) & (B - 1) variant if OP is |.
12094 * expmed.c (expand_shift_1): For rotations by const0_rtx just
12095 return shifted. Use (-op1) & (prec - 1) as other_amount
12096 instead of prec - op1.
12097
4502fe8d
MJ
120982013-05-13 Martin Jambor <mjambor@suse.cz>
12099
12100 PR middle-end/42371
12101 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
12102 (ipa_constant_data): New type.
12103 (ipa_jump_func): Use ipa_constant_data to hold information about
12104 constant jump functions.
12105 (ipa_get_jf_constant): Adjust to jump function type changes.
12106 (ipa_get_jf_constant_rdesc): New function.
12107 (ipa_param_descriptor): New field controlled_uses.
12108 (ipa_get_controlled_uses): New function.
12109 (ipa_set_controlled_uses): Likewise.
12110 * ipa-ref.h (ipa_find_reference): Declare.
12111 * ipa-prop.c (ipa_cst_ref_desc): New type.
12112 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
12113 changes.
12114 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
12115 New parameter cs. Adjust all callers.
12116 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
12117 (remove_described_reference): New function.
12118 (jfunc_rdesc_usable): Likewise.
12119 (try_make_edge_direct_simple_call): Decrement controlled use count,
12120 attempt to remove reference if it hits zero.
12121 (combine_controlled_uses_counters): New function.
12122 (propagate_controlled_uses): Likewise.
12123 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
12124 (ipa_edge_duplication_hook): Duplicate reference descriptions.
12125 (ipa_print_node_params): Print described use counter.
12126 (ipa_write_jump_function): Adjust to jump function type changes.
12127 (ipa_read_jump_function): New parameter CS, pass it to
12128 ipa_set_jf_constant. Adjust caller.
12129 (ipa_write_node_info): Stream controlled use count
12130 (ipa_read_node_info): Likewise.
12131 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
12132 asserting.
12133 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
12134 count. Remove cloning-added reference if it reaches zero.
12135 * ipa-ref.c (ipa_find_reference): New function.
12136
0864bfc2
GG
121372013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
12138
e7208ea3 12139 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG
12140 alignment values for AMD BD and BT architectures.
12141
640bfeb2
MG
121422013-05-13 Marc Glisse <marc.glisse@inria.fr>
12143
12144 * tree-vect-generic.c (uniform_vector_p): Move ...
12145 * tree.c (uniform_vector_p): ... here.
12146 * tree.h (uniform_vector_p): Declare it.
12147 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
12148 into a scalar.
12149
3a60f32b
JJ
121502013-05-13 Jakub Jelinek <jakub@redhat.com>
12151
198fe1bf
JJ
12152 PR tree-optimization/57230
12153 * tree-ssa-strlen.c (handle_char_store): Record length for
12154 array store from STRING_CST.
12155
3a60f32b
JJ
12156 PR tree-optimization/57230
12157 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
12158 check.
12159
566be57c
JR
121602013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
12161
12162 * config/epiphany/epiphany.c (epiphany_init): Check size of
12163 NUM_MODES_FOR_MODE_SWITCHING.
12164 (epiphany_expand_prologue):
12165 Remove CONFIG_REGNUM initial value handling code.
12166 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
12167 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 12168 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR
12169 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
12170 Don't return 1 for FP_MODE_NONE.
12171 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
12172 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
12173 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
12174 * config/epiphany/epiphany.md (save_config): New pattern.
12175
0f2c2331
UB
121762013-05-12 Uros Bizjak <ubizjak@gmail.com>
12177
12178 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
12179
5b3f93c7
UB
121802013-05-10 Uros Bizjak <ubizjak@gmail.com>
12181
12182 * config/i386/i386.md (memory): Handle sseishft1.
12183 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
12184 (*vec_extractv2di_1): Ditto.
12185
1f873f0e
VM
121862013-05-10 Vladimir Makarov <vmakarov@redhat.com>
12187
12188 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
12189 saved registers.
12190
341427fa 121912013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH
12192
12193 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
12194 Add mthumb/march=armv7-a multilib.
12195 Add mthumb/march=armv7-r multilib.
12196 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
12197
9e69bdde
RC
121982013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
12199
12200 * config/v850/t-rtems: Add more multilibs.
12201
9ff09a22
RB
122022013-05-10 Richard Biener <rguenther@suse.de>
12203
12204 PR tree-optimization/57214
12205 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
12206 not propagate from SSA names that occur in abnormal PHI nodes.
12207
a5e0cd1d
MG
122082013-05-10 Marc Glisse <marc.glisse@inria.fr>
12209
12210 * stor-layout.c (element_precision): New function.
12211 * machmode.h (element_precision): Declare it.
12212 * tree.c (build_minus_one_cst): New function.
12213 (element_precision): Likewise.
12214 * tree.h (build_minus_one_cst): Declare new function.
12215 (element_precision): Likewise.
12216 * fold-const.c (operand_equal_p): Use element_precision.
12217 (fold_binary_loc): Handle vector types.
12218 * convert.c (convert_to_integer): Use element_precision.
12219 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
12220 separately.
12221
cb2558bc
RS
122222013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
12223
12224 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
12225 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
12226 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
12227 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12228 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
12229 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
12230 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
12231 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
12232 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
12233 (Uuw8): New constraints.
12234 (Usb4): Move into alphabetical order.
12235 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
12236 (sd8_operand, ub8_operand, uw8_operand): New predicates.
12237 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
12238 previously unnamed patterns.
12239 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
12240 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
12241 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
12242 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
12243 of set_attr_alternative/if_then_else. Use extended_mips16 instead
12244 of specific lengths.
12245
cb3b8d33
JJ
122462013-05-10 Jakub Jelinek <jakub@redhat.com>
12247
6f93c008
JJ
12248 * config/i386/i386.md (rotateinv): New code attr.
12249 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
12250 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
12251 roll $31, %eax, etc.
12252
cb3b8d33
JJ
12253 PR tree-optimization/45216
12254 PR tree-optimization/57157
12255 * tree-ssa-forwprop.c (simplify_rotate): New function.
12256 (ssa_forward_propagate_and_combine): Call it.
12257
afb119be
RB
122582013-05-10 Richard Biener <rguenther@suse.de>
12259
12260 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
12261 disable peeling when we version for aliasing.
12262 (vector_alignment_reachable_p): Honor explicit user alignment.
12263 (vect_supportable_dr_alignment): Likewise.
12264 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
12265 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
12266 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
12267 then peeling to arrange for the cost-model check to come first.
12268
01ae4861
AM
122692013-05-10 Alan Modra <amodra@gmail.com>
12270
12271 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
12272 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
12273 * configure: Regenerate.
12274
ebc9a431
AM
122752013-05-10 Alan Modra <amodra@gmail.com>
12276
12277 PR target/55033
12278 * varasm.c (default_elf_select_section): Move !DECL_P check..
12279 (get_named_section): ..to here before calling get_section_name.
12280 Adjust assertion.
12281 (default_section_type_flags): Add DECL_P check.
12282 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
12283 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
12284
d4bca93c
JR
122852013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
12286
12287 * config/epiphany/epiphany.c (epiphany_expand_prologue):
12288 When using gen_stack_adjust_str with a register offset, add a
12289 REG_FRAME_RELATED_EXPR note.
12290
60ca9a65
UB
122912013-05-09 Uros Bizjak <ubizjak@gmail.com>
12292
12293 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
12294 (*vec_extractv4si_zext_mem): Ditto.
12295 (*vec_extractv2di): Add 0->x and x->x alternatives.
12296 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
12297 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
12298
a3409c02
JM
122992013-05-09 Jason Merrill <jason@redhat.com>
12300
0138d6b2
JM
12301 N3639 C++1y VLA support
12302 * gimplify.c (gimplify_vla_decl): Don't touch an existing
12303 DECL_VALUE_EXPR.
12304
a3409c02
JM
12305 * tree.c (build_constructor_va): New.
12306 * tree.h: Declare it.
12307
66e6b990
MJ
123082013-05-09 Martin Jambor <mjambor@suse.cz>
12309
12310 PR lto/57084
12311 * gimple-fold.c (canonicalize_constructor_val): Call
12312 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
12313
64cfa6c0
JH
123142013-05-09 Jan Hubicka <jh@suse.cz>
12315 Richard Biener <rguenther@suse.de>
12316
12317 PR lto/54095
12318 * symtab.c (symtab_make_decl_local): Do not add private names.
12319
c3167b00
JH
123202013-05-09 Jan Hubicka <jh@suse.cz>
12321
12322 PR lto/54095
12323 * symtab.c (insert_to_assembler_name_hash): Handle clones.
12324 (unlink_from_assembler_name_hash): Likewise.
12325 (symtab_prevail_in_asm_name_hash, symtab_register_node,
12326 symtab_unregister_node, symtab_initialize_asm_name_hash,
12327 change_decl_assembler_name): Update.
12328
12dc6974
SN
123292013-05-09 Sofiane Naci <sofiane.naci@arm.com>
12330
12331 * config/aarch64/aarch64.md: New movtf split.
12332 (*movtf_aarch64): Update.
12333 (aarch64_movdi_tilow): Handle TF modes and rename to
12334 aarch64_movdi_<mode>low.
12335 (aarch64_movdi_tihigh): Handle TF modes and rename to
12336 aarch64_movdi_<mode>high
12337 (aarch64_movtihigh_di): Handle TF modes and rename to
12338 aarch64_mov<mode>high_di
12339 (aarch64_movtilow_di): Handle TF modes and rename to
12340 aarch64_mov<mode>low_di
12341 (aarch64_movtilow_tilow): Remove spurious whitespace.
12342 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
12343 splits.
12344 (aarch64_print_operand): Update.
12345
227eb343
AM
123462013-05-09 Alan Modra <amodra@gmail.com>
12347
12348 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
12349 powerpc64le.
12350 * configure: Regenerate.
12351
0b013847
UB
123522013-05-08 Uros Bizjak <ubizjak@gmail.com>
12353
12354 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
12355 splitter preparation statements.
12356 * config/i386/sse.md (*vec_extract* splitters): Ditto.
12357 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
12358 adjust_address_nv.
12359
1dc3d6e9
BS
123602013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12361
12362 * gimple-ssa-strength-reduction.c (count_candidates): Change
12363 return value to int.
12364 (analyze_candidates_and_replace): Change type of length to int.
12365
e61e7d28
UB
123662013-05-08 Uros Bizjak <ubizjak@gmail.com>
12367
12368 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
12369 (*vec_extract<mode>): Use VI12_128 mode iterator.
12370 (*vec_extract<mode>_mem): Ditto.
12371 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
12372 attribute.
12373
4fbfcf44
DN
123742013-05-08 Diego Novillo <dnovillo@google.com>
12375
12376 PR bootstrap/54659
12377
12378 Revert:
4fbfcf44
DN
12379 2012-08-17 Diego Novillo <dnovillo@google.com>
12380
e61e7d28
UB
12381 PR bootstrap/54281
12382 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
12383 * config.in: Regenerate.
12384 * configure: Regenerate.
12385 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 12386
702d8703
JH
123872013-05-08 Jan Hubicka <jh@suse.cz>
12388
12389 PR lto/54095
12390 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
12391 * cgraph.h (symtab_node_base): Add unique_name.
12392 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
12393 input_overwrite_node, input_varpool_node): Stream unique_name.
12394 * cgraphclones.c (cgraph_create_virtual_clone,
12395 cgraph_function_versioning): Set unique_name.
12396 * ipa.c (function_and_variable_visibility): Set unique_name.
12397
8b28cf47
BS
123982013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12399
12400 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
12401 (alloc_cand_and_find_basis): Restrict conditional candidate
12402 processing to CAND_MULTs.
12403
e86074fd
JH
124042013-05-08 Jan Hubicka <jh@suse.cz>
12405
12406 PR lto/54095
12407 lto-symtab.c (lto_symtab_symbol_p): New function.
12408 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
12409 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
12410 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
12411 Skip static symbols.
12412
44398cbe
PC
124132013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
12414
12415 PR tree-optimization/57200
12416 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
12417 Only call inform if the preceding warning_at returns true.
12418
f6bc1c4a
HS
124192013-05-07 Han Shen <shenhan@google.com>
12420
12421 * cfgexpand.c (record_or_union_type_has_array_p): New function.
12422 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
12423 * common.opt (fstack-protector-strong): New option.
12424 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
12425 * doc/invoke.texi (Optimization Options): Document
12426 "-fstack-protector-strong".
12427 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
12428
4ffecb1f
SB
124292013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12430
12431 * config/mips/mips.c (mips_machine_reorg2): Return 0.
12432
5a107a0f
VM
124332013-05-07 Vladimir Makarov <vmakarov@redhat.com>
12434
12435 * ira.c (update_equiv_regs): Add insn having equiv memory even if
12436 it is not lhs of the insn.
12437 (setup_reg_equiv): Remove insn having equiv memory which it is not
12438 lhs of the insn.
12439 * lra-constraints.c (process_address): Try to improve generation
12440 code for address base + disp.
12441 (lra_constraints): Make correct the code for checking insn setting
12442 up backward equivalence. Remove insn only if it is in the init
12443 insn list.
12444 * lra-eliminations.c (update_reg_eliminate): Change return value.
12445 (lra_eliminate): Use the result.
12446
3f5783ea
UB
124472013-05-07 Uros Bizjak <ubizjak@gmail.com>
12448
12449 * config/i386/sse.md (ssescalarnummask): New mode attribute.
12450 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
12451 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
12452 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
12453 register target operands.
12454 (*vec_extractv8hi_sse2): New pattern.
12455 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
12456 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
12457 (*vec_extract<mode>_mem): New insn and split pattern.
12458
8a5800b8
CL
124592013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
12460
12461 * config/arm/arm.c (arm_asan_shadow_offset): New function.
12462 (TARGET_ASAN_SHADOW_OFFSET): Define.
12463 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
12464 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
12465
7bf55a70
BS
124662013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12467
12468 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
12469 (incr_vec_index): Return -1 if increment not found.
12470 (create_add_on_incoming_edge): Assert if increment not found.
12471 (record_increment): Limit number of increments recorded.
12472 (all_phi_incrs_profitable): Return false if an increment not found.
12473 (replace_profitable_candidates): Don't process increments that were
12474 not recorded.
12475 (analyze_candidates_and_replace): Limit size of incr_vec.
12476
3f8825c0
RB
124772013-05-07 Richard Biener <rguenther@suse.de>
12478
12479 * calls.c (special_function_p): setjmp-like functions are leaf.
12480 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
12481 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
12482
fd4842cd
SN
124832013-05-07 Sofiane Naci <sofiane.naci@arm.com>
12484
12485 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
12486 (aarch64_simd_mov<mode>): New expander.
12487 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
12488 (aarch64_simd_mov_to_<mode>high): Likewise.
12489 (aarch64_simd_mov_from_<mode>low): Likewise.
12490 (aarch64_simd_mov_from_<mode>high): Likewise.
12491 (aarch64_dup_lane<mode>): Update.
12492 (aarch64_dup_lanedi): New instruction pattern.
12493 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
12494 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
12495
a7a7d10e
BS
124962013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12497
12498 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
12499 (replace_mult_candidate): Remove unnecessary argument; remove
12500 unnecessary parameter from call to introduce_cast_before_cand.
12501 (replace_unconditional_candidate): Remove unnecessary parameter
12502 from call to replace_mult_candidate.
12503 (replace_conditional_candidate): Likewise.
12504 (insert_initializers): Use make_temp_ssa_name.
12505 (introduce_cast_before_cand): Remove unnecessary argument; use
12506 make_temp_ssa_name.
12507 (replace_one_candidate): Remove unnecessary argument; remove
12508 unnecessary parameter from calls to introduce_cast_before_cand.
12509 (replace_profitable_candidates): Remove unnecessary parameters
12510 from calls to replace_one_candidate.
12511
29105868
BS
125122013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12513
12514 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
12515 phi def as possibly hiding a basis for a CAND_ADD whose operands
12516 have been commuted in the analysis.
12517 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
12518
4095f9fa
N
125192013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
12520
12521 * config/aarch64/aarch64.md
12522 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
12523 shift value between 0-4.
12524
39e843e8
RB
125252013-05-07 Richard Biener <rguenther@suse.de>
12526
12527 * double-int.h (rshift): New overload.
12528 * double-int.c (rshift): New function.
12529 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
12530 (create_reference_ops_from_ref): Remove.
12531 (vn_reference_insert): Use shared ops for constructing the
12532 reference and copy it.
12533
0a1a83cb
RB
125342013-05-07 Richard Biener <rguenther@suse.de>
12535
12536 PR middle-end/57190
12537 * tree-eh.c (sink_clobbers): Properly propagate
12538 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
12539
43303d6f
JJ
125402013-05-07 Jakub Jelinek <jakub@redhat.com>
12541
ba7e83f8
JJ
12542 PR tree-optimization/57149
12543 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
12544 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
12545 collect_phi_def_edges, execute_late_warn_uninitialized): Use
12546 uninit_undefined_value_p instead of ssa_undefined_value_p.
12547
43303d6f
JJ
12548 PR debug/57184
12549 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
12550 for modifier == EXPAND_INITIALIZER.
12551
14523c25
AB
125522013-05-07 Anton Blanchard <anton@samba.org>
12553
12554 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
12555 for powerpc64 little endian.
12556 * configure: Regenerate.
12557
cb7c8be9
GS
125582013-05-06 Graham Stott <grahams@btinternet.com>
12559
12560 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
12561 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
12562 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
12563 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
12564
125652013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS
12566
12567 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
12568 codes which allow non-lvalues.
12569
9a0ee7b0
MG
125702013-05-06 Marc Glisse <marc.glisse@inria.fr>
12571
12572 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
12573 components are all 1s.
12574 (integer_minus_onep): New function.
12575 * tree.h (integer_minus_onep): Declare it.
12576 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
12577 integer_minus_onep instead of integer_all_onesp.
12578
f2c17ea9
OE
125792013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12580
12581 PR target/52933
12582 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
12583 variations of these patterns.
12584
f75e6a51
UB
125852013-05-06 Uros Bizjak <ubizjak@gmail.com>
12586
12587 * config/i386/i386.md (isa): Add x64_sse4 member.
12588 (enabled): Handle x64_sse4.
12589 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
12590 instruction for 64bit SSE4_1 targets. Update insn attributes.
12591 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
12592 instruction for SSE4_1 targets. Update insn attributes.
12593 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12594 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
12595 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
12596 const_1 selector.
12597 (*vec_extractv4si): Rename from *sse4_1_pextrd.
12598 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
12599 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
12600
a986d468
OE
126012013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12602
12603 PR target/57108
12604 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
12605
382522cb
MK
126062013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
12607
12608 * final.c (do_assembler_dialects): Don't handle curly braces and
12609 vertical bar escaped by % as dialect delimiters.
12610 (output_asm_insn): Print curly braces and vertical bar if escaped
12611 by % and ASSEMBLER_DIALECT defined.
12612 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
12613 * doc/tm.texi: Regenerated.
12614
fb0d5c60
SB
126152013-05-06 Steven Bosscher <steven@gcc.gnu.org>
12616
fb0d5c60
SB
12617 * config/mips/mips.c: Include tree-pass.h.
12618 (mips_reorg): Split in pre- and post-dbr_schedule parts.
12619 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
12620 (pass_mips_machine_reorg2): New machine specific pass.
12621 (insert_pass_mips_machine_reorg2): New pass plugin definition.
12622 (mips_option_override): Register the new pass.
12623 * rtl.h (cleanup_barriers): Remove prototype.
12624 (dbr_schedule): Likewise.
12625 * jump.c (cleanup_barriers): Make static.
12626 * reorg.c (dbr_schedule): Likewise.
12627
aa06a978
RB
126282013-05-06 Richard Biener <rguenther@suse.de>
12629
12630 PR tree-optimization/57185
12631 * tree-parloops.c (add_field_for_reduction): Handle anonymous
12632 SSA names properly.
12633
0b953bec
UB
126342013-05-06 Uros Bizjak <ubizjak@gmail.com>
12635
12636 PR target/57106
12637 * config/i386/i386.c (add_parameter_dependencies): Add dependence
12638 between "first_arg" and "insn", not "last" and "insn".
12639
28708525
WS
126402013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12641
12642 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
12643 (find_candidates_in_block): Re-enable slsr_process_phi.
12644 (create_phi_basis): Fix double counting of candidate adjustment.
12645
0107dca2
RB
126462013-05-06 Richard Biener <rguenther@suse.de>
12647
12648 PR middle-end/57147
12649 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
12650 the edge is also fallthru, preserve it and just clear the
12651 abnormal flag.
12652 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
12653 also complex, preserve that and just clear the fallthru flag.
12654 * tree-inline.c (update_ssa_across_abnormal_edges): Also
12655 update virtual operands.
12656
470d4d13
AM
126572013-05-06 Alan Modra <amodra@gmail.com>
12658
12659 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
12660 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
12661 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
12662 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
12663 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12664 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
12665
8f1dbf8d
AM
126662013-05-06 Alan Modra <amodra@gmail.com>
12667
12668 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
12669 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
12670 (DEFAULT_ASM_ENDIAN): Define.
12671 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
12672 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
12673 Update -K PIC clause from sysv4.h.
12674 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
12675 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
12676
54890767
AM
126772013-05-06 Alan Modra <amodra@gmail.com>
12678
12679 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
12680 twice for little-endian.
12681 (ashrdi3_no_power, ashrdi3): Support little-endian.
12682
2353515d
OE
126832013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
12684
12685 PR target/55303
12686 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
12687 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
12688 related expanders.
12689 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
12690 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
12691 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
12692 New predicates.
12693
33e67557
SB
126942013-05-05 Steven Bosscher <steven@gcc.gnu.org>
12695 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
12696
12697 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
12698 * config/pa/pa.opt: Make mbig-switch a no-op.
12699 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
12700 (CASE_VECTOR_MODE): Always return SImode.
12701 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
12702 for the !TARGET_BIG_SWITCH case.
12703 * config/pa/pa-linux.h: Likewise.
12704 * config/pa/pa-openbsd.h: Likewise.
12705 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
12706 * config/pa/pa.md (short_jump): Remove define_insn.
12707 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
12708 (casesi0): Remove define_insn.
12709 (type): Remove btable_branch.
12710 (pa_combine_type): Likewise.
12711 (in_nullified_branch_delay): Likewise.
12712 (in_call_delay): Likewise.
12713 (define_delay): Likewise.
12714 (define_insn_reservation "Z3"): Likewise.
12715 (define_insn_reservation "Z4"): Likewise.
12716 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
12717 (pa_adjust_insn_length): Remove adjustment for btable branches.
12718 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
12719 and mno-big-switch
12720
3095685e
UB
127212013-05-05 Uros Bizjak <ubizjak@gmail.com>
12722
12723 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
12724 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
12725 Add m->r,x alternatives.
12726 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
12727 splitters using SWI48x mode iterator.
12728 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
12729 TARGET_64BIT. Add m->x alternative.
12730 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
12731 Add o->x alternative. Enable for TARGET_SSE.
12732 (sse_storeq): Remove expander.
12733 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
12734 with memory input operand.
12735 (*vec_extractv2di_1 splitter): New.
12736 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
12737 * config/i386/i386.md (ssevecmodelower): New mode attribute.
12738
4b36ae28
SB
127392013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12740
12741 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
12742 (INT_LOWPART): Delete.
12743 (extract_MB): Adjust.
12744 (extract_ME): Adjust.
12745 (print_operand): Adjust.
12746
da226db2
SB
127472013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
12748
12749 * config/rs6000/predicates.md (reg_or_add_cint_operand,
12750 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12751 (reg_or_logical_cint_operand, easy_fp_constant,
12752 logical_const_operand): Delete "CONST_DOUBLE" case.
12753 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
12754 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 12755 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB
12756 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
12757 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
12758 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
12759 test.
12760 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
12761 CONST_DOUBLE DImode/VOIDmode case.
12762 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
12763 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12764 CONST_DOUBLE VOIDmode case.
12765 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
12766 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
12767 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
12768 Delete CONST_DOUBLE case.
12769 (splitters for mov FMOVE64 const_double): Delete
12770 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
12771 "HOST_BITS_PER_WIDE_INT >= 64" test.
12772 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
12773 case.
12774 (mov DI const_double): Delete.
12775
40de22d6
JJ
127762013-05-04 Jakub Jelinek <jakub@redhat.com>
12777
12778 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
12779 on op shows all bits zero in mode of a lowpart subreg, return zero.
12780
5ec6aff2
MM
127812013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
12782
12783 PR target/57150
12784 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
12785 to save TFmode registers and DImode to save TImode registers for
12786 caller save operations.
12787 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
12788 mark being partially clobbered since they only use the first
12789 double word.
12790
12791 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
12792 and TDmode only use the upper 64-bits of each VSX register.
12793
2cefad90
BS
127942013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12795
12796 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
12797 (find_candidates_in_block): Disable slsr_process_phi.
12798
d6d7eee1
GW
127992013-05-03 Guozhi Wei <carrot@google.com>
12800
12801 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB
12802 constructor to ...
12803 (build_init_ctor): ... here.
d6d7eee1 12804
9b92d12b
BS
128052013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12806
12807 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
12808 (slsr_cand_d): Redefine def_phi.
12809 (stride_status, phi_adjust_status, count_phis_status): New enums.
12810 (find_phi_def): New.
12811 (find_basis_for_base_expr): New.
12812 (find_basis_for_candidate): Handle hidden bases.
12813 (alloc_cand_and_find_basis): Handle phi candidates.
12814 (slsr_process_phi): New.
12815 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
12816 (create_mul_imm_cand): Likewise.
12817 (create_add_ssa_cand): Exclude phi base candidates.
12818 (create_add_imm_cand): Likewise.
12819 (slsr_process_cast): Likewise.
12820 (slsr_process_copy): Likewise.
12821 (find_candidates_in_block): Handle phi candidates.
12822 (dump_candidate): Likewise.
12823 (unconditional_cands): Delete.
12824 (unconditional_cands_with_known_stride_p): Delete.
12825 (phi_dependent_cand_p): New.
12826 (cand_increment): Handle phi-dependent candidates.
12827 (replace_dependent): Delete.
12828 (replace_mult_candidate): New.
12829 (replace_unconditional_candidate): New.
12830 (incr_vec_index): Move to avoid forward reference.
12831 (create_add_on_incoming_edge): New.
12832 (create_phi_basis): New.
12833 (replace_dependents): Delete.
12834 (replace_conditional_candidate): New.
12835 (phi_add_costs): New.
12836 (replace_uncond_cands_and_profitable_phis): New.
12837 (record_increment): Handle phi adjustments.
12838 (record_phi_increments): New.
12839 (record_increments): Handle phi adjustments.
12840 (phi_incr_cost): New.
12841 (lowest_cost_path): Handle phis.
12842 (total_savings): Likewise.
12843 (analyze_increments): Likewise.
12844 (ncd_with_phi): New.
12845 (ncd_of_cand_and_phis): New.
12846 (nearest_common_dominator_for_cands): Handle phi increments.
12847 (all_phi_incrs_profitable): New.
12848 (replace_profitable_candidates): Handle phi-dependent candidates.
12849 (analyze_candidates_and_replace): Likewise.
12850
68f073d4
TJ
128512013-05-03 Teresa Johnson <tejohnson@google.com>
12852
12853 PR bootstrap/57154
12854 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
12855 do not exceed REG_BR_PROB_BASE.
12856
a4ee7cb9
JL
128572013-05-03 Jeff Law <law@redhat.com>
12858
ade67f70 12859 PR tree-optimization/57144
a4ee7cb9
JL
12860 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
12861 operand of the condition will bit into the new type when eliminating
12862 a cast feeding a condition.
12863
47954c4e
JJ
128642013-05-03 Jakub Jelinek <jakub@redhat.com>
12865
12866 PR rtl-optimization/57130
3095685e
UB
12867 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
12868 of COMPARE as in_code to the recursive call if needed.
47954c4e 12869
3c21604f
UB
128702013-05-03 Uros Bizjak <ubizjak@gmail.com>
12871
12872 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
12873 (enabled): Handle new members.
12874 * config/i386/sse.md (*vec_concatv2si): Merge from
12875 *vec_concatv2si_sse2 and vec_concatv2si_sse.
12876 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
12877
12211b99 128782013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR
12879
12880 PR tree-optimization/57027
12881 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
12882 for fnms opportunity, check we got the prerequisite kind
12883 of tree / gimple before using accessor functions.
12884
07bfc9ec
RB
128852013-05-03 Richard Biener <rguenther@suse.de>
12886
12887 * double-int.h (lshift): New overload without precision
12888 and arith argument.
12889 (operator *=, operator +=, operator -=): Move ...
12890 * double-int.c (operator *=, operator +=, operator -=): ... here
12891 and implement more efficiently.
12892 (mul_double_with_sign): Remove.
12893 (lshift_double): Adjust to take unsinged shift argument, push
12894 dispatching code to callers.
12895 (mul_double_wide_with_sign): Add early out for callers that
12896 are not interested in high parts or overflow.
12897 (lshift): New function.
12898 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
12899 dispatch code here.
12900 (lrotate, rrotate): Use logical shifts.
12901 * expr.c (get_inner_reference): Use lshift.
12902 * fixed-value.c (do_fixed_divide): Likewise.
12903 * tree-dfa.c (get_ref_base_and_extent): Likewise.
12904 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
12905 (indirect_refs_may_alias_p): Likewise.
12906 (stmt_kills_ref_p_1): Likewise.
12907
7769bb64
VP
129082013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12909
12910 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
12911
1d0c8e5c
VP
129122013-05-03 Vidya Praveen <vidyapraveen@arm.com>
12913
12914 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
12915 scalar form of FABD instruction.
12916
f15feaf9
VM
129172013-05-02 Vladimir Makarov <vmakarov@redhat.com>
12918
12919 * lra-constraints.c (process_alt_operands): Add checking alt
12920 number to choose the best alternative.
12921
d90e76d4
RB
129222013-05-02 Richard Biener <rguenther@suse.de>
12923
12924 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
12925 bitmap and its handling.
12926 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
12927
9f8e7a96
RB
129282013-05-02 Richard Biener <rguenther@suse.de>
12929
12930 PR middle-end/57140
12931 * tree-inline.c (copy_loops): Properly handle removed loops.
12932 (copy_cfg_body): Mark destination loops for fixup if source
12933 loops needed fixup.
12934
f3a81b39
GY
129352013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
12936
12937 PR target/56732
12938 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
12939 generating simple_return for naked functions.
12940
7b920a9a
MJ
129412013-05-02 Martin Jambor <mjambor@suse.cz>
12942
12943 PR middle-end/56988
12944 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
12945 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
12946 flags match.
12947 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
12948 ipa_agg_replacement_value structures.
12949 (known_aggs_to_agg_replacement_list): Likewise.
12950 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
12951 (read_agg_replacement_chain): Likewise.
12952 (ipcp_transform_function): Also check that by_ref flags match.
12953
2c41c19d
RB
129542013-05-02 Richard Biener <rguenther@suse.de>
12955
12956 * graphds.h (struct graph): Add obstack member.
12957 * graphds.c (new_graph): Initialize obstack and allocate
12958 vertices from it.
12959 (add_edge): Allocate edge from the obstack.
3c21604f 12960 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 12961
8b47039c
TJ
129622013-05-02 Teresa Johnson <tejohnson@google.com>
12963
12964 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
12965 divides.
12966 * cfg.c (update_bb_profile_for_threading): Ditto.
12967 * tree-inline.c (copy_bb): Ditto.
12968 (copy_edges_for_bb): Ditto.
12969 (initialize_cfun): Ditto.
12970 (copy_cfg_body): Ditto.
12971 (expand_call_inline): Ditto.
12972 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
12973 (estimate_node_size_and_time): Ditto.
12974 (inline_merge_summary): Ditto.
12975 * cgraphclones.c (cgraph_clone_edge): Ditto.
12976 (cgraph_clone_node): Ditto.
12977 * sched-rgn.c (compute_dom_prob_ps): Ditto.
12978 (compute_trg_info): Ditto.
12979
da65928c
IB
129802013-05-02 Ian Bolton <ian.bolton@arm.com>
12981
12982 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
12983 S reg when fp attribute set.
12984 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
12985
add3c965
IB
129862013-05-02 Ian Bolton <ian.bolton@arm.com>
12987
12988 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
12989 New pattern.
12990 (*and_one_cmplsi3_compare0_uxtw): Likewise.
12991 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
12992 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
12993
fdd43ac4
RB
129942013-05-02 Richard Biener <rguenther@suse.de>
12995
12996 * tree-scalar-evolution.c (scev_info_hasher): Remove.
12997 (struct instantiate_cache_entry): New type.
12998 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
12999 (struct instantiate_cache_type): New type.
13000 (set_instantiated_value, get_instantiated_value): Remove.
13001 (get_instantiated_value_entry): New function.
13002 (instantiate_scev_name): Use the new cache and adjust.
13003 (instantiate_scev_poly): Adjust.
13004 (instantiate_scev_binary): Likewise.
13005 (instantiate_array_ref): Likewise.
13006 (instantiate_scev_convert): Likewise.
13007 (instantiate_scev_not): Likewise.
13008 (instantiate_scev_3): Likewise.
13009 (instantiate_scev_2): Likewise.
13010 (instantiate_scev_r): Likewise.
13011 (instantiate_scev): Likewise.
13012 (resolve_mixers): Likewise.
13013
36ff9dfb
VM
130142013-05-01 Vladimir Makarov <vmakarov@redhat.com>
13015
13016 PR target/57091
13017 * lra-constraints.c (best_small_class_operands_num): Remove.
13018 (process_alt_operands): Remove small_class_operands_num. Take
13019 small classes operands into losers and only if the operand is not
13020 matched. Modify debugging output.
13021 (curr_insn_transform): Remove best_small_class_operands_num.
13022 Print insn name.
13023
36054fab
JG
130242013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13025
13026 * config/aarch64/aarch64-builtins.c
13027 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
13028 * config/aarch64/aarch64-simd-builtins.def
13029 (reduc_splus_): Add new modes.
13030 (reduc_uplus_): New.
13031 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
13032 (reduc_uplus_v4sf): Likewise.
13033 (reduc_splus_v4sf): Likewise.
13034 (aarch64_addv<mode>): Likewise.
13035 (reduc_uplus_<mode>): Likewise.
13036 (reduc_splus_<mode>): Likewise.
13037 (aarch64_addvv2di): Likewise.
13038 (reduc_uplus_v2di): Likewise.
13039 (reduc_splus_v2di): Likewise.
13040 (aarch64_addvv2si): Likewise.
13041 (reduc_uplus_v2si): Likewise.
13042 (reduc_splus_v2si): Likewise.
13043 (reduc_<sur>plus_<mode>): New.
13044 (reduc_<sur>plus_v2di): Likewise.
13045 (reduc_<sur>plus_v2si): Likewise.
13046 (reduc_<sur>plus_v4sf): Likewise.
13047 (aarch64_addpv4sf): Likewise.
13048 * config/aarch64/arm_neon.h
13049 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
13050 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
13051 add UNSPEC_SADDV, UNSPEC_UADDV.
13052 (SUADDV): New.
13053 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
13054
6dce23a8
JG
130552013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13056
13057 * config/aarch64/arm_neon.h
13058 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
13059
1598945b
JG
130602013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13061
13062 * config/aarch64/aarch64-builtins
13063 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
13064
998eaf97
JG
130652013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13066
13067 * config/aarch64/aarch64-simd-builtins.def
13068 (reduc_smax_): New.
13069 (reduc_smin_): Likewise.
13070 (reduc_umax_): Likewise.
13071 (reduc_umin_): Likewise.
13072 (reduc_smax_nan_): Likewise.
13073 (reduc_smin_nan_): Likewise.
13074 (fmax): Remove.
13075 (fmin): Likewise.
13076 (smax): Update for V2SF, V4SF and V2DF modes.
13077 (smin): Likewise.
13078 (smax_nan): New.
13079 (smin_nan): Likewise.
13080 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
13081 (<su><maxmin><mode>3): ...This, refactor.
13082 (s<maxmin><mode>3): New.
13083 (<maxmin_uns><mode>3): Likewise.
13084 (reduc_<maxmin_uns>_<mode>): Refactor.
13085 (reduc_<maxmin_uns>_v4sf): Likewise.
13086 (reduc_<maxmin_uns>_v2si): Likewise.
13087 (aarch64_<fmaxmin><mode>: Remove.
13088 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
13089 new builtin names.
13090 (vmin<q>_f<32,64>): Likewise.
13091 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13092 (FMAXMIN): New.
13093 (su): Add mappings for smax, smin, umax, umin.
13094 (maxmin): New.
13095 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
13096 (FMAXMIN): Rename as...
13097 (FMAXMIN_UNS): ...This.
13098 (maxminv): Remove.
13099 (fmaxminv): Likewise.
13100 (fmaxmin): Likewise.
13101 (maxmin_uns): New.
13102 (maxmin_uns_op): Likewise.
13103
bd11644e
JG
131042013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13105
13106 * config/aarch64/arm_neon.h
13107 (vac<ge, gt><sd>_f<32, 64>): Rename to...
13108 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
13109 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
13110
75dd5ace
JG
131112013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13112
13113 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
13114 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
13115
7c19979f
JG
131162013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13117
13118 * config/aarch64/aarch64-simd.md
13119 (vcond<mode>_internal): Handle special cases for constant masks.
13120 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
13121 (vcondu<mode><mode>): Likewise.
13122 (vcond<v_cmp_result><mode>): New.
13123
bb60efd9
JG
131242013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13125
13126 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
13127 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
13128 * config/aarch64/aarch64-simd-builtins.def
13129 (cmeq): Update to BUILTIN_VALLDI.
13130 (cmgt): Likewise.
13131 (cmge): Likewise.
13132 (cmle): Likewise.
13133 (cmlt): Likewise.
13134 * config/aarch64/arm_neon.h
13135 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
13136 to builtins or C as appropriate.
13137
889b9412
JG
131382013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
13139
13140 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
13141 (cmgeu): ...This.
13142 (cmhi): Rename to...
13143 (cmgtu): ...This.
13144 * config/aarch64/aarch64-simd.md
13145 (simd_mode): Add SF.
13146 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
13147 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
13148 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
13149 (cstore<mode>_neg): ...This.
13150 * config/aarch64/iterators.md
13151 (VALLF): new.
13152 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
13153 (COMPARISONS): New.
13154 (UCOMPARISONS): Likewise.
13155 (optab): Add missing comparisons.
13156 (n_optab): New.
13157 (cmp_1): Likewise.
13158 (cmp_2): Likewise.
13159 (CMP): Likewise.
13160 (cmp): Remove.
13161 (VCMP_S): Likewise.
13162 (VCMP_U): Likewise.
13163 (V_cmp_result): Add DF, SF modes.
13164 (v_cmp_result): Likewise.
13165 (v): Likewise.
13166 (vmtype): Likewise.
13167 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
13168
0a7dbb76
GY
131692013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
13170
13171 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
13172 define_insn to define_insn_and_split.
13173 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
13174 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
13175 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
13176 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
13177 (thumb2_negscc): Likewise.
13178
fb614ca6
GY
131792013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13180
13181 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
13182
9e64a0bf
GY
131832013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
13184
13185 * config/arm/thumb2.md: Remove trailing whitespaces.
13186
d6b28156
RS
131872013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
13188
13189 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
13190 Use gen_int_mode rather than GEN_INT.
13191
f91674c3
L
131922013-04-30 H.J. Lu <hongjiu.lu@intel.com>
13193
b0dec607 13194 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L
13195 debug_gimple_stmt.
13196
3551257c
RB
131972013-04-30 Richard Biener <rguenther@suse.de>
13198
13199 PR middle-end/57122
3c21604f 13200 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 13201
4c1aff1c
RB
132022013-04-30 Richard Biener <rguenther@suse.de>
13203
13204 PR middle-end/57107
13205 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
13206
54b8379a
AB
132072013-04-30 Andrey Belevantsev <abel@ispras.ru>
13208
13209 PR rtl-optimization/56957
13210 PR rtl-optimization/57105
54b8379a
AB
13211 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
13212 variable. Use just INSN_UID for determining whether an insn
13213 should be only disconnected from the insn stream.
13214 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
13215
abc27962
JJ
132162013-04-30 Jakub Jelinek <jakub@redhat.com>
13217
13218 PR tree-optimization/57104
13219 * tsan.c (instrument_expr): Don't instrument accesses to
13220 DECL_HARD_REGISTER VAR_DECLs.
13221
0fc822d0
RB
132222013-04-30 Richard Biener <rguenther@suse.de>
13223
13224 * function.h (loops_for_fn): New inline function.
13225 (set_loops_for_fn): Likewise.
13226 * cfgloop.h (place_new_loop): Add struct function parameter.
13227 (get_loop): Likewise.
13228 (get_loops): Likewise.
13229 (number_of_loops): Likewise.
13230 (fel_next): Adjust.
13231 (fel_init): Likewise.
13232 * cfg.c (get_loop_copy): Adjust.
13233 * cfgloop.c (flow_loops_dump): Likewise.
13234 (record_loop_exits): Likewise.
13235 (verify_loop_structure): Likewise.
13236 * cfgloopanal.c (mark_irreducible_loops): Likewise.
13237 (estimate_reg_pressure_cost): Likewise.
13238 (mark_loop_exit_edges): Likewise.
13239 * cfgloopmanip.c (place_new_loop): Likewise.
13240 (add_loop): Likewise.
13241 (duplicate_loop): Likewise.
13242 * graph.c (draw_cfg_nodes): Likewise.
13243 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
13244 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
13245 (extract_affine_chrec): Likewise.
13246 (build_scop_iteration_domain): Likewise.
13247 * graphite.c (graphite_initialize): Likewise.
13248 * ira-build.c (create_loop_tree_nodes): Likewise.
13249 (more_one_region_p): Likewise.
13250 (rebuild_regno_allocno_maps): Likewise.
13251 (mark_loops_for_removal): Likewise.
13252 (mark_all_loops_for_removal): Likewise.
13253 (remove_unnecessary_regions): Likewise.
13254 (ira_build): Likewise.
13255 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
13256 * loop-init.c (fix_loop_structure): Likewise.
13257 (gate_rtl_move_loop_invariants): Likewise.
13258 (gate_rtl_unswitch): Likewise.
13259 (gate_rtl_unroll_and_peel_loops): Likewise.
13260 (rtl_doloop): Likewise.
13261 * lto-streamer-in.c (input_cfg): Likewise.
13262 * lto-streamer-out.c (output_cfg): Likewise.
13263 * modulo-sched.c (sms_schedule): Likewise.
13264 * predict.c (tree_estimate_probability): Likewise.
13265 (tree_estimate_probability_driver): Likewise.
13266 (estimate_loops): Likewise.
13267 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
13268 (move_sese_region_to_fn): Likewise.
13269 (debug_loop_num): Likewise.
13270 * tree-chrec.c (chrec_evaluate): Likewise.
13271 (hide_evolution_in_other_loops_than_loop): Likewise.
13272 (chrec_component_in_loop_num): Likewise.
13273 (reset_evolution_in_loop): Likewise.
13274 (evolution_function_is_invariant_rec_p): Likewise.
13275 * tree-if-conv.c (main_tree_if_conversion): Likewise.
13276 * tree-inline.c (copy_loops): Likewise.
13277 (copy_cfg_body): Likewise.
13278 (tree_function_versioning): Likewise.
13279 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
13280 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
13281 Likewise.
13282 (add_to_evolution_1): Likewise.
13283 (scev_const_prop): Likewise.
13284 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
13285 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
13286 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
13287 (tree_ssa_lim_initialize): Likewise.
13288 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
13289 (verify_loop_closed_ssa): Likewise.
13290 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
13291 (tree_ssa_loop_im): Likewise.
13292 (tree_ssa_loop_unswitch): Likewise.
13293 (tree_vectorize): Likewise.
13294 (check_data_deps): Likewise.
13295 (tree_ssa_loop_ivcanon): Likewise.
13296 (tree_ssa_loop_bounds): Likewise.
13297 (tree_complete_unroll): Likewise.
13298 (tree_complete_unroll_inner): Likewise.
13299 (tree_parallelize_loops): Likewise.
13300 (tree_ssa_loop_prefetch): Likewise.
13301 (tree_ssa_loop_ivopts): Likewise.
13302 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
13303 * tree-vectorizer.c (vectorize_loops): Likewise.
13304
37953bd3
MF
133052013-04-29 Mike Frysinger <vapier@gentoo.org>
13306
13307 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
13308 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
13309 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
13310 with EABI_LINK_SPEC.
13311
f9ed28db
UB
133122013-04-29 Uros Bizjak <ubizjak@gmail.com>
13313
13314 PR target/44578
13315 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
13316 alternative.
13317
deca73f5
VM
133182013-04-29 Vladimir Makarov <vmakarov@redhat.com>
13319
13320 PR target/57097
37953bd3 13321 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM
13322 using memory for pseudos. Print cost dump for alternatives.
13323 Modify cost values for conflicts with early clobbers.
13324 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
13325
fdca7d03
UB
133262013-04-29 Uros Bizjak <ubizjak@gmail.com>
13327
13328 PR target/57098
13329 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
13330
0b064172
IB
133312013-04-29 Ian Bolton <ian.bolton@arm.com>
13332
13333 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
13334 from/to S register.
13335 (movdi_aarch64): Support LDR/STR from/to D register.
13336
473cec55
IB
133372013-04-29 Ian Bolton <ian.bolton@arm.com>
13338
13339 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
13340 or higher by default.
13341
a6f30e66
RB
133422013-04-29 Richard Biener <rguenther@suse.de>
13343
13344 PR middle-end/57075
13345 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
13346 even if not adding abnormal edges for calls that can make
13347 abnormal gotos.
13348
0c2b2040
RB
133492013-04-29 Richard Biener <rguenther@suse.de>
13350
13351 PR middle-end/57103
13352 * tree-cfg.c (move_stmt_op): Fix condition under which to update
13353 TREE_BLOCK.
13354 (move_stmt_r): Remove redundant checking.
13355
f41f80f9
TJ
133562013-04-29 Teresa Johnson <tejohnson@google.com>
13357
13358 PR bootstrap/57077
13359 * basic-block.h (apply_scale): New function.
13360 (apply_probability): Use apply_scale.
13361 * gimple-streamer-in.c (input_bb): Ditto.
13362 * lto-streamer-in.c (input_cfg): Ditto.
13363 * lto-cgraph.c (merge_profile_summaries): Ditto.
13364 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 13365 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ
13366 (copy_edges_for_bb): Ditto.
13367 (copy_cfg_body): Ditto.
13368
315bbd2e
TV
133692013-04-29 Tom de Vries <tom@codesourcery.com>
13370
13371 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
13372 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
13373 (tail_merge_optimize): Handle current_loops == NULL.
13374
ebbd90d8
JL
133752013-04-26 Jeff Law <law@redhat.com>
13376
13377 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
13378 (simplify_cond_using_ranges): Generalize code to simplify
13379 COND_EXPRs where one argument is a constant and the other
13380 is an SSA_NAME created by an integral type conversion.
13381
8b9b57eb
KT
133822013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13383
13384 * config/arm/arm.md (store_minmaxsi): Use only when
13385 optimize_insn_for_size_p.
13386
9498e5dc
CB
133872013-04-29 Christian Bruel <christian.bruel@st.com>
13388
13389 PR target/57108
13390 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
13391
6093bc06
RB
133922013-04-29 Richard Biener <rguenther@suse.de>
13393
13394 PR middle-end/57089
fdca7d03
UB
13395 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
13396 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB
13397 (expand_omp_for_generic): Properly add loops.
13398 (expand_omp_for_static_nochunk): Likewise.
13399 (expand_omp_for_static_chunk): Likewise.
13400 (expand_omp_for): For the degenerate case fixup loops.
13401 (expand_omp_sections): Fix default bb placement in loops.
13402 (expand_omp_atomic_pipeline): Properly add loops.
13403
84aacbfd
KT
134042013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13405
13406 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
13407
8dee4479
TV
134082013-04-29 Tom de Vries <tom@codesourcery.com>
13409
13410 * tree-ssa-tail-merge.c: Update header comment.
13411
47934dc4
JG
134122013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13413
13414 * config/aarch64/arm_neon.h
13415 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
13416 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
13417 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
13418 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
13419 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
13420 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
13421 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
13422 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
13423
384be29f
JG
134242013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13425
13426 * config/aarch64/aarch64-simd.md
13427 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
13428 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
13429 fix_trunc, fixuns_trunc.
13430 (ftrunc<VDQF:mode>2): New.
13431 * config/aarch64/iterators.md (optab): Add fix, fixuns.
13432 (fix_trunc_optab): New.
13433
0386b123
JG
134342013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13435
13436 * config/aarch64/aarch64-builtins.c
13437 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
13438 iceilf, lround, iroundf.
13439
00fcb892
UB
134402013-04-29 Uros Bizjak <ubizjak@gmail.com>
13441
13442 PR target/54349
13443 * config/i386/i386.h (enum ix86_tune_indices)
13444 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
13445 New, split from X86_TUNE_INTER_UNIT_MOVES.
13446 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
13447 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
13448 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
13449 (TARGET_INTER_UNIT_MOVES): Remove.
13450 * config/i386/i386.c (initial_ix86_tune_features): Update.
13451 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
13452 (ix86_expand_convert_uns_didf_sse): Use
13453 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13454 (ix86_expand_vector_init_one_nonzero): Ditto.
13455 (ix86_expand_vector_init_interleave): Ditto.
13456 (inline_secondary_memory_needed): Return true for moves from SSE class
13457 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
13458 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
13459 * config/i386/constraints.md (Yi, Ym): Depend on
13460 TARGET_INTER_UNIT_MOVES_TO_VEC.
13461 (Yj, Yn): New constraints.
13462 * config/i386/i386.md (*movdi_internal): Change constraints of
13463 operand 1 from Yi to Yj and from Ym to Yn.
13464 (*movsi_internal): Ditto.
13465 (*movdf_internal): Ditto.
13466 (*movsf_internal): Ditto.
13467 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
13468 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
13469 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
13470 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
13471 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
13472 * config/i386/sse.md (movdi_to_sse): Ditto.
13473 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
13474 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
13475 TARGET_INTER_UNIT_MOVES.
13476 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
13477 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
13478 instead of TARGET_INTER_UNIT_MOVES.
13479 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
13480 operand 1 from Yi to Yj and from Ym to Yn.
13481
4c871069
JG
134822013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13483
13484 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
13485 (float_truncate_hi_): Likewise.
13486 (float_extend_lo_): Likewise.
13487 (float_truncate_lo_): Likewise.
13488 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
13489 (aarch64_float_extend_lo_v2df): Likewise.
13490 (vec_unpacks_hi_v4sf): Likewise.
13491 (aarch64_float_truncate_lo_v2sf): Likewise.
13492 (aarch64_float_truncate_hi_v4sf): Likewise.
13493 (vec_pack_trunc_v2df): Likewise.
13494 (vec_pack_trunc_df): Likewise.
13495
1709ff9b
JG
134962013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13497
13498 * config/aarch64/aarch64-builtins.c
13499 (aarch64_fold_builtin): Fold float conversions.
13500 * config/aarch64/aarch64-simd-builtins.def
13501 (floatv2si, floatv4si, floatv2di): New.
13502 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
13503 * config/aarch64/aarch64-simd.md
13504 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
13505 * config/aarch64/iterators.md (FLOATUORS): New.
13506 (optab): Add float, floatuns.
13507 (su_optab): Likewise.
13508
ce966824
JG
135092013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13510
13511 * config/aarch64/aarch64-builtins.c
13512 (aarch64_builtin_vectorized_function): Use new names for
13513 fcvt builtins.
13514 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
13515 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
13516 (fcvtzu): Split as...
13517 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
13518 (fcvtas): Split as...
13519 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
13520 (fcvtau): Split as...
13521 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
13522 (fcvtps): Split as...
13523 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
13524 (fcvtpu): Split as...
13525 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
13526 (fcvtms): Split as...
13527 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
13528 (fcvtmu): Split as...
13529 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
13530 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
13531 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
13532 (lfrintnusf, lfrintnudf): Likewise.
13533 * config/aarch64/aarch64-simd.md
13534 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
13535 define_insn.
13536 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
13537 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
13538 (fcvt_pattern): Likewise.
13539
b9de24fe
JG
135402013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13541
13542 * config/aarch64/aarch64-simd.md
13543 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
13544 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
13545
77a205be
JG
135462013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13547
13548 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
13549 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
13550 (vrnd<a,m,n,p>_f32): Implement using builtins.
13551 (vrnd<i,x><q>_f<32, 64>): New.
13552
0659ce6f
JG
135532013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
13554
13555 * config/aarch64/aarch64-builtins.c
13556 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
13557 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
13558 (frintz): Rename to...
13559 (btrunc): ...this.
13560 (frintp): Rename to...
13561 (ceil): ...this.
13562 (frintm): Rename to...
13563 (floor): ...this.
13564 (frinti): Rename to...
13565 (nearbyint): ...this.
13566 (frintx): Rename to...
13567 (rint): ...this.
13568 (frinta): Rename to...
13569 (round): ...this.
13570 * config/aarch64/aarch64-simd.md
13571 (aarch64_frint<frint_suffix><mode>): Delete.
13572 (<frint_pattern><mode>2): Convert to insn.
13573 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
13574 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
13575 (frint_pattern): Likewise.
13576 (frint_suffix): Likewise.
13577
ea78906a
RB
135782013-04-29 Richard Biener <rguenther@suse.de>
13579
13580 PR tree-optimization/57081
13581 * loop-init.c: Include tree-flow.h.
13582 (loop_optimizer_finalize): Free number of iteration estimates.
13583 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
13584
baee1763
JJ
135852013-04-29 Jakub Jelinek <jakub@redhat.com>
13586
94dc5332
JJ
13587 PR tree-optimization/57083
13588 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
13589 non-singleton shift count range, zero extend low_bound for uns case.
13590
baee1763
JJ
13591 * config/i386/predicates.md (general_vector_operand): New predicate.
13592 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
13593 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
13594 if they aren't nonimmediate operands. If their original values
13595 satisfy const_vector_equal_evenodd_p, don't shift them.
13596 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
13597 predicates. For the SSE4.1 case force operands[{1,2}] into registers
13598 if not nonimmediate_operand.
13599 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
13600 instead of register_operand.
13601 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
13602
a9073727 136032013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB
13604
13605 * stor-layout.c (finalize_size_functions): Allocate a structure and
13606 reset cfun before dumping the functions.
13607
ba8011e6
JJ
136082013-04-27 Jakub Jelinek <jakub@redhat.com>
13609
d6fde69e
JJ
13610 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
13611
ba8011e6
JJ
13612 PR target/56866
13613 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
13614 use xop_pmacsdqh if uns_p.
13615 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
13616 the immediate rotate count.
13617
2c62cbaa
VM
136182013-04-26 Vladimir Makarov <vmakarov@redhat.com>
13619
13620 * rtl.h (struct rtx_def): Add comment for field jump.
13621 (LRA_SUBREG_P): New macro.
13622 * recog.c (register_operand): Check LRA_SUBREG_P.
13623 * lra.c (lra): Add note at the end of RTL code. Align non-empty
13624 stack frame.
13625 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
13626 (lra_final_code_change): Skip subreg change for operators.
13627 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
13628 if there are no operand changes.
13629 * lra-constraints.c (curr_insn_set): New.
13630 (match_reload): Set LRA_SUBREG_P.
13631 (emit_spill_move): Ditto.
13632 (check_and_process_move): Use curr_insn_set. Process only single
13633 set insns. Don't initialize sec_mem_p and change_p.
13634 (simplify_operand_subreg): Use LRA_SUBREG_P.
13635 (reg_in_class_p): New function.
13636 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
13637 of #ifdef. Add code to remove cycling.
13638 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
13639 non-null disp. Reload inner instead of disp when base and index
13640 are null. Try to put lo_sum into register.
13641 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
13642 (check_and_process_move): Move code for move cost check to
13643 simple_move_p. Remove equiv_substitution.
13644 (simple_move_p): New function.
13645 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
13646 curr_insn_set. Call check_and_process_move only for single set
13647 insns. Use the new function. Move call of check_and_process_move
13648 after operand equiv substitution and address process.
13649
e7d764f3
JJ
136502013-04-26 Jakub Jelinek <jakub@redhat.com>
13651
13652 PR go/57045
13653 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
13654 with nonlocal goto receivers or returns twice calls, ignore
13655 unininitialized values from abnormal edges to nl goto receiver
13656 or returns twice call.
13657
41e10689
JJ
136582013-04-26 Jakub Jelinek <jakub@redhat.com>
13659
13660 PR tree-optimization/57051
13661 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
13662 and VEC_RSHIFT_EXPR if shift count is a multiple of element
13663 bitsize.
13664
d7ed20db
RB
136652013-04-26 Richard Biener <rguenther@suse.de>
13666
13667 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
13668 (expand_omp_taskreg): Likewise. Mark loops for fixup.
13669 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
13670 (fixup_loop_arrays_after_move): New function.
13671 (move_sese_region_to_fn): Properly outline the loop tree parts
13672 of the SESE region.
13673
df93505e
UB
136742013-04-26 Uros Bizjak <ubizjak@gmail.com>
13675
13676 * config/i386/i386.md (type, unit): Fix long lines.
13677
dd366ec3
RB
136782013-04-26 Richard Biener <rguenther@suse.de>
13679
13680 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
13681 (lto-streamer-out.o): Likewise.
13682 * cfgloop.c (init_loops_structure): Export, add struct function
13683 argument and adjust.
13684 (flow_loops_find): Adjust.
13685 * cfgloop.h (enum loop_estimation): Add EST_LAST.
13686 (init_loops_structure): Declare.
13687 * lto-streamer-in.c: Include cfgloop.h.
13688 (input_cfg): Input the loop tree.
13689 * lto-streamer-out.c: Include cfgloop.h.
13690 (output_cfg): Output the loop tree.
13691 (output_struct_function_base): Do not drop PROP_loops.
13692
a9e0d843
RB
136932013-03-26 Richard Biener <rguenther@suse.de>
13694
13695 * tree-cfg.c (execute_build_cfg): Build the loop tree.
13696 (pass_build_cfg): Provide PROP_loops.
13697 (move_sese_region_to_fn): Remove loops that are outlined into fn
13698 for now.
13699 * tree-inline.c: Include cfgloop.h.
13700 (initialize_cfun): Do not drop PROP_loops.
13701 (copy_loops): New function.
13702 (copy_cfg_body): Copy loop structure.
13703 (tree_function_versioning): Initialize destination loop tree.
13704 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
13705 (pass_parallelize_loops): Do IL verification.
13706 * loop-init.c (loop_optimizer_init): Fixup loops if required.
13707 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
13708 the CFG make sure we fixup loops as well.
13709 * tree-ssa-tail-merge.c: Include cfgloop.h.
13710 (replace_block_by): When merging loop latches mark loops for fixup.
13711 * lto-streamer-out.c (output_struct_function_base): Drop
13712 PROP_loops for now.
13713 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
13714 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
13715 * ipa-split.c: Include cfgloop.h.
13716 (split_function): Add the new return block to the loop tree root.
13717 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
13718 whether we have removed the forwarder block.
13719 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
13720 * cfgloop.h (place_new_loop): Declare.
13721 * cfgloopmanip.c (place_new_loop): Export.
13722 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
13723 (tree-switch-conversion.o): Likewise.
13724 (tree-complex.o): Likewise.
13725 (tree-inline.o): Likewise.
13726 (tree-ssa-tailmerge.o): Likewise.
13727 (ipa-split.o): Likewise.
13728 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
13729 (tree-ssa-copy.o): Likewise.
13730 * tree-switch-conversion.c: Include cfgloop.h
13731 (process_switch): If we emit a bit-test cascade, schedule loops
13732 for fixup.
13733 * tree-complex.c: Include cfgloop.h.
13734 (expand_complex_div_wide): Properly add new basic-blocks to loops.
13735 * asan.c: Include cfgloop.h.
13736 (create_cond_insert_point): Properly add new basic-blocks to
13737 loops, schedule loop fixup.
13738 * cfgloop.c (verify_loop_structure): Check that looks are not
13739 marked for fixup.
13740 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
13741 to loops.
13742 (expand_omp_for_generic): Likewise.
13743 (expand_omp_sections): Likewise.
13744 (expand_omp_atomic_pipeline): Schedule loops for fixup.
13745 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
13746 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
13747 is initialized, not when loops are present.
13748 * tree-parloops.c (parallelize_loops): Remove checking here.
13749 * passes.c (init_optimization_passes): Schedule a copy-propagation
13750 pass before complete unrolling of inner loops.
13751
e78e8a0b
JJ
137522013-04-26 Jakub Jelinek <jakub@redhat.com>
13753
a2e836b2
JJ
13754 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
13755 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
13756 (colorize_init): Add argument to _WIN32 version.
13757 * toplev.c: Include diagnostic-color.h.
13758 (process_options): Default to -fdiagnostics-color=auto if
13759 GCC_COLORS env var is in the environment.
13760 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB
13761 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
13762 env var is in the environment, the default is auto rather than never.
a2e836b2 13763
e78e8a0b
JJ
13764 * diagnostic.h (file_name_as_prefix): Add context argument.
13765 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
13766 the string as locus.
13767 * langhooks.c (lhd_print_error_function): Adjust caller.
13768
013e5ef9
LC
137692013-04-25 Lawrence Crowl <crowl@google.com>
13770
13771 * var-tracking.c (shared_hash_def::htab):
13772 Change type to hash_table. Update dependent calls and types.
13773
4a8fb1a1
LC
137742013-04-25 Lawrence Crowl <crowl@google.com>
13775
13776 * Makefile.in: Update as needed below.
13777
13778 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
13779 Move declaration to after the type's method definitons.
13780
13781 * attribs.c (htab_t scoped_attributes::attribute_hash):
13782 Change type to hash_table. Update dependent calls and types.
13783
13784 * bitmap.c (htab_t bitmap_desc_hash):
13785 Change type to hash_table. Update dependent calls and types.
13786
13787 * cselib.c (htab_t cselib_hash_table):
13788 Change type to hash_table. Update dependent calls and types.
13789
13790 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
13791 (hash_string_slot_node): Move implementation into lto-streamer.h
13792 struct string_slot_hasher.
13793 (eq_string_slot_node): Likewise.
13794
13795 * data-streamer-out.c: Update output_block::string_hash_table
13796 dependent calls and types.
13797
13798 * dwarf2cfi.c (htab_t trace_index):
13799 Change type to hash_table. Update dependent calls and types.
13800
13801 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
13802 Change type to hash_table. Update dependent calls and types.
13803 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
13804 (htab_t optimize_external_refs::map): Likewise.
13805 (htab_t output_comp_unit::extern_map): Likewise.
13806 (htab_t output_comdat_type_unit::extern_map): Likewise.
13807 (htab_t output_macinfo::macinfo_htab): Likewise.
13808 (htab_t optimize_location_lists::htab): Likewise.
13809 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
13810
13811 * except.c (htab_t ehspec_hash_type):
13812 Change type to hash_table. Update dependent calls and types.
13813 (assign_filter_values::ttypes): Likewise.
13814 (assign_filter_values::ehspec): Likewise.
13815 (sjlj_assign_call_site_values::ar_hash): Likewise.
13816 (convert_to_eh_region_ranges::ar_hash): Likewise.
13817
13818 * gcse.c (htab_t pre_ldst_table):
13819 Change type to hash_table. Update dependent calls and types.
13820
13821 * ggc-common.c (htab_t saving_htab):
13822 Change type to hash_table. Update dependent calls and types.
13823 (htab_t loc_hash): Likewise.
13824 (htab_t ptr_hash): Likewise.
13825 (call_count): Rename ggc_call_count.
13826 (call_alloc): Rename ggc_call_alloc.
13827 (loc_descriptor): Rename make_loc_descriptor.
13828 (add_statistics): Rename ggc_add_statistics.
13829
13830 * ggc-common.c (saving_htab):
13831 Change type to hash_table. Update dependent calls and types.
13832
13833 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
13834 (push_gimplify_context): Likewise.
13835 (pop_gimplify_context): Likewise.
13836 (struct gimple_temp_hash_elt): Added.
13837 (struct gimplify_hasher): Likewise.
13838 (struct gimplify_ctx.temp_htab):
13839 Change type to hash_table. Update dependent calls and types.
13840
13841 * gimple-fold.c: Include gimplify-ctx.h.
13842
13843 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
13844 Change type to hash_table. Update dependent calls and types.
13845 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
13846 avoid potential global name collision.
13847
13848 * gimplify.c: Include gimplify-ctx.h.
13849 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
13850 (htab_t gimplify_ctx::temp_htab):
13851 Update dependent calls and types for new type hash_table.
13852 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
13853 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
13854
13855 * gimplify-ctx.h: New.
13856 (struct gimple_temp_hash_elt): Move from gimplify.c.
13857 (class gimplify_hasher): New.
13858 (struct gimplify_ctx): Move from gimple.h.
13859 (htab_t gimplify_ctx::temp_htab):
13860 Change type to hash_table. Update dependent calls and types.
13861
13862 * graphite-clast-to-gimple.c: Include graphite-htab.h.
13863 (htab_t ivs_params::newivs_index):
13864 Change type to hash_table. Update dependent calls and types.
13865 (htab_t ivs_params::params_index): Likewise.
13866 (htab_t print_generated_program::params_index): Likewise.
13867 (htab_t gloog::newivs_index): Likewise.
13868 (htab_t gloog::params_index): Likewise.
13869
13870 * graphite.c: Include graphite-htab.h.
13871 4htab_t graphite_transform_loops::bb_pbb_mapping):
13872 Change type to hash_table. Update dependent calls and types.
13873
13874 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
13875 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
13876 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
13877
13878 * graphite-dependences.c: Include graphite-htab.h.
13879 (loop_is_parallel_p): Change hash table type of parameter.
13880
13881 * graphite-htab.h: New.
13882 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
13883 (extern find_pbb_via_hash): Move from graphite-poly.h.
13884 (extern loop_is_parallel_p): Move from graphite-poly.h.
13885 (extern get_loop_body_pbbs): Move from graphite-poly.h.
13886
13887 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
13888 (extern loop_is_parallel_p): Move to graphite-htab.h.
13889 (extern get_loop_body_pbbs): Move to graphite-htab.h.
13890
13891 * haifa-sched.c (htab_t delay_htab):
13892 Change type to hash_table. Update dependent calls and types.
13893 (htab_t delay_htab_i2): Likewise.
13894
13895 * ira-color.c (htab_t allocno_hard_regs_htab):
13896 Change type to hash_table. Update dependent calls and types.
13897
13898 * ira-costs.c (htab_t cost_classes_htab):
13899 Change type to hash_table. Update dependent calls and types.
13900
13901 * loop-invariant.c (htab_t merge_identical_invariants::eq):
13902 Change type to hash_table. Update dependent calls and types.
13903
13904 * loop-iv.c (htab_t bivs):
13905 Change type to hash_table. Update dependent calls and types.
13906
13907 * loop-unroll.c (htab_t opt_info::insns_to_split):
13908 Change type to hash_table. Update dependent calls and types.
13909 (htab_t opt_info::insns_with_var_to_expand): Likewise.
13910
13911 * lto-streamer.h (struct string_slot): Move from data-streamer.h
13912 (struct string_slot_hasher): New.
13913 (htab_t output_block::string_hash_table):
13914 Change type to hash_table. Update dependent calls and types.
13915
13916 * lto-streamer-in.c (freeing_string_slot_hasher): New.
13917 (htab_t file_name_hash_table):
13918 Change type to hash_table. Update dependent calls and types.
13919
13920 * lto-streamer-out.c: Update output_block::string_hash_table dependent
13921 calls and types.
13922
13923 * lto-streamer.c (htab_t tree_htab):
13924 Change type to hash_table. Update dependent calls and types.
13925
13926 * omp-low.c: Include gimplify-ctx.h.
13927
13928 * passes.c (htab_t name_to_pass_map):
13929 Change type to hash_table. Update dependent calls and types.
13930 (pass_traverse): Rename to passes_pass_traverse.
13931
13932 * plugin.c (htab_t event_tab):
13933 Change type to hash_table. Update dependent calls and types.
13934
13935 * postreload-gcse.c (htab_t expr_table):
13936 Change type to hash_table. Update dependent calls and types.
13937 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
13938
13939 * sese.c (debug_rename_map_1): Make extern.
13940 (htab_t copy_bb_and_scalar_dependences::rename_map):
13941 Change type to hash_table. Update dependent calls and types.
13942
13943 * sese.h (extern debug_rename_map): Move to .c file.
13944
13945 * store-motion.c (htab_t store_motion_mems_table):
13946 Change type to hash_table. Update dependent calls and types.
13947
13948 * trans-mem.c (htab_t tm_new_mem_hash):
13949 Change type to hash_table. Update dependent calls and types.
13950
13951 * tree-browser.c (htab_t TB_up_ht):
13952 Change type to hash_table. Update dependent calls and types.
13953
13954 * tree-cfg.c (htab_t discriminator_per_locus):
13955 Change type to hash_table. Update dependent calls and types.
13956
13957 * tree-complex.c: Include tree-hasher.h
13958 (htab_t complex_variable_components):
13959 Change type to hash_table. Update dependent calls and types.
13960
13961 * tree-eh.c (htab_t finally_tree):
13962 Change type to hash_table. Update dependent calls and types.
13963
13964 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
13965 struct int_tree_hasher.
13966 (extern int_tree_map_eq): Likewise.
13967 (uid_decl_map_hash): Removed.
13968 (extern decl_tree_map_eq): Likewise.
13969
13970 * tree-hasher.h: New.
13971 (struct int_tree_hasher): New.
13972 (typedef int_tree_htab_type): New.
13973
13974 * tree-inline.c: Include gimplify-ctx.h.
13975
13976 * tree-mudflap.c: Include gimplify-ctx.h.
13977
13978 * tree-parloops.c: Include tree-hasher.h.
13979 (htab_t eliminate_local_variables_stmt::decl_address):
13980 Change type to hash_table. Update dependent calls and types.
13981 (htab_t separate_decls_in_region::decl_copies): Likewise.
13982
13983 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
13984 Change type to hash_table. Update dependent calls and types.
13985
13986 * tree-sra.c (candidates):
13987 Change type to hash_table. Update dependent calls and types.
13988
13989 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
13990 in tree-flow.h.
13991 (int_tree_map_hash): Likewise.
13992
13993 * tree-ssa-dom.c (htab_t avail_exprs):
13994 Change type to hash_table. Update dependent calls and types.
13995
13996 * tree-ssa-live.c (var_map_base_init::tree_to_index):
13997 Change type to hash_table. Update dependent calls and types.
13998
13999 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
14000 Change type to hash_table. Update dependent calls and types.
14001
14002 * tree-ssa-phiopt.c (seen_ssa_names):
14003 Change type to hash_table. Update dependent calls and types.
14004
14005 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
14006 Change type to hash_table. Update dependent calls and types.
14007
14008 * tree-ssa-uncprop.c (equiv):
14009 Change type to hash_table. Update dependent calls and types.
14010
c5a44004
JJ
140112013-04-25 Jakub Jelinek <jakub@redhat.com>
14012
14013 PR rtl-optimization/57003
14014 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
14015 call note_stores with kill_clobbered_value callback again after
14016 killing regs_invalidated_by_call.
14017
09962a4a
JG
140182013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14019
14020 * config/aarch64/aarch64-simd.md
14021 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
14022 (aarch64_simd_bsl<mode>): Likewise.
14023 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
14024
ea28bb0b
MP
140252013-04-25 Marek Polacek <polacek@redhat.com>
14026
14027 PR tree-optimization/57066
3c21604f 14028 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 14029
96659611
JG
140302013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
14031
14032 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
14033
9697e620
JG
140342013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14035
14036 * config/aarch64/aarch64-builtins.c
14037 (aarch64_fold_builtin): New.
14038 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
14039 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
14040 * config/aarch64/aarch64-simd-builtins.def (abs): New.
14041 * config/aarch64/arm_neon.h
14042 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
14043
0ac198d3
JG
140442013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14045 Tejas Belagod <tejas.belagod@arm.com>
14046
14047 * config/aarch64/aarch64-builtins.c
14048 (aarch64_gimple_fold_builtin): New.
14049 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
14050 * config/aarch64/aarch64-simd-builtins.def (addv): New.
14051 * config/aarch64/aarch64-simd.md (addpv4sf): New.
14052 (addvv4sf): Update.
14053 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
14054
58cff58c
N
140552013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14056
df93505e 14057 * config/aarch64/aarch64.md
58cff58c
N
14058 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
14059
7e0228bf
N
140602013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
14061
14062 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
14063 (*ngcsi_uxtw): New pattern.
14064
5819f96f 140652013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 14066 Julian Brown <julian@codesourcery.com>
5819f96f
KT
14067
14068 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
14069 (TB_DREG): Add T_V4HF.
14070 (v4hf_UP): New macro.
14071 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 14072 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
14073 Handle initialisation of V4HF. Adjust initialisation of reinterpret
14074 built-ins.
df93505e 14075 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
14076 (arm_vector_mode_supported_p): Handle V4HF.
14077 (arm_mangle_map): Handle V4HFmode.
14078 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
14079 * config/arm/arm_neon_builtins.def: Add entries for
14080 vcvtv4hfv4sf, vcvtv4sfv4hf.
14081 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
14082 (neon_vcvtv4hfv4sf): Likewise.
14083 * config/arm/neon-gen.ml: Handle half-precision floating point
14084 features.
14085 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
14086 * config/arm/arm_neon.h: Regenerate.
14087 * config/arm/neon.ml (type elts): Add F16.
14088 (type vectype): Add T_float16x4, T_floatHF.
14089 (type vecmode): Add V4HF.
14090 (type features): Add Requires_FP_bit feature.
14091 (elt_width): Handle F16.
14092 (elt_class): Likewise.
14093 (elt_of_class_width): Likewise.
14094 (mode_of_elt): Refactor.
14095 (type_for_elt): Handle F16, fix error messages.
14096 (vectype_size): Handle T_float16x4.
14097 (vcvt_sh): New function.
14098 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
14099 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
14100 (string_of_mode): Handle V4HF.
14101 * doc/arm-neon-intrinsics.texi: Regenerate.
14102
1ef395e4
JG
141032013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
14104
14105 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
14106 format specifier in 'X' case.
14107
41c34e94
AM
141082013-04-25 Alan Modra <amodra@gmail.com>
14109
14110 PR target/57052
14111 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
14112 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
14113 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
14114 Repeat for many other rotate/shift and mask patterns using subregs.
14115 Name lshiftrt insns.
14116 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
14117 on WORDS_BIG_ENDIAN.
14118
b9a7eb5d
AM
141192013-04-25 Alan Modra <amodra@gmail.com>
14120
14121 * config.gcc: Support little-endian powerpc-linux targets.
14122 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
14123 (LINK_OS_LINUX_SPEC): Define.
14124 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
14125 Preserve MASK_LITTLE_ENDIAN.
14126 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
14127 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
14128 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
14129 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
14130 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
14131 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
14132 Correct fp word order for little-endian. Don't shift toc entries
14133 smaller than a word for little-endian.
14134 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
14135 (bswapdi2 splits): Correct low-part subreg for little-endian.
14136 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
14137 low/high where such is correct only for be.
14138 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
14139 little-endian for -mcall-aixdesc.
14140
87f73374
AM
141412013-04-25 Alan Modra <amodra@gmail.com>
14142
14143 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
14144 replace_equiv_address_nv.
14145
cabf91cd
AM
141462013-04-25 Alan Modra <amodra@gmail.com>
14147
14148 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
14149
0ae24cc8
VM
141502013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14151
14152 Revert:
14153 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14154 * rtl.h (struct rtx_def): ...
cabf91cd 14155
77bce07c
VM
141562013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14157
14158 PR rtl-optimizations/57046
14159 * lra-constraints (split_reg): Set up lra_risky_transformations_p
14160 for multi-reg splits.
14161
0db63e7f
L
141622013-04-24 H.J. Lu <hongjiu.lu@intel.com>
14163
14164 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
14165
3362b6b6
SA
141662013-04-24 Sterling Augustine <saugustine@google.com>
14167
14168 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
14169 (comp_dir_string, debug_str_dwo_section): New.
14170 (DEBUG_STR_DWO_SECTION): Rename to ...
14171 (DEBUG_DWO_STR_SECTION): ... this.
14172 (DEBUG_NORM_STR_SECTION): Delete.
14173 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
14174 (DEBUG_STR_DWO_SECTION_FLAGS): New.
14175 (find_AT_string): Move most logic to ...
14176 (find_AT_string_in_table): ... here. New.
14177 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
14178 add_skeleton_AT_string. Delete logic.
14179 (output_skeleton_debug_sections): Remove call to
14180 add_top_level_skeleton_die_attrs.
14181 (add_comp_dir_attribute): Move logic to comp_dir_string.
14182 (dwarf2out_init): Initialize debug_str_dwo_section.
14183 (output_indirect_string): Call find_string_form.
14184 (output_indirect_strings): Rewrite.
14185 (prune_unused_types): Empty skeleton_debug_str_hash.
14186 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
14187 (dwarf2out_finish): Call output_indirect_strings.
14188
e93e18e9
PC
141892013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14190
14191 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
14192
f6ce35ac
VM
141932013-04-24 Vladimir Makarov <vmakarov@redhat.com>
14194
cabf91cd 14195 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM
14196 (LRA_SUBREG_P): New macro.
14197 * recog.c (register_operand): Check LRA_SUBREG_P.
14198 * lra.c (lra): Add note at the end of RTL code. Align non-empty
14199 stack frame.
14200 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
14201 (lra_final_code_change): Skip subreg change for operators.
14202 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
14203 if there are no operand changes.
14204 * lra-constraints.c (curr_insn_set): New.
14205 (match_reload): Set LRA_SUBREG_P.
14206 (emit_spill_move): Ditto.
14207 (check_and_process_move): Use curr_insn_set. Process only single
14208 set insns. Don't initialize sec_mem_p and change_p.
14209 (simplify_operand_subreg): Use LRA_SUBREG_P.
14210 (reg_in_class_p): New function.
14211 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
14212 of #ifdef. Add code to remove cycling.
14213 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
14214 non-null disp. Reload inner instead of disp when base and index
14215 are null. Try to put lo_sum into register.
14216 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 14217 (check_and_process_move): Move code for move cost check to
f6ce35ac 14218 simple_move_p. Remove equiv_substitution.
cabf91cd 14219 (simple_move_p): New function.
f6ce35ac
VM
14220 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
14221 curr_insn_set. Call check_and_process_move only for single set
14222 insns. Use the new function. Move call of check_and_process_move
14223 after operand equiv substitution and address process.
14224
38047d90
JG
142252013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
14226
14227 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
14228 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
14229 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
14230
13f39b2e
PC
142312013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
14232
14233 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
14234
97c116dc
MP
142352013-04-24 Marek Polacek <polacek@redhat.com>
14236
14237 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
14238 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
14239 (select_loops_exit_conditions): Likewise.
14240 (number_of_iterations_for_all_loops): Likewise.
14241 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
14242 (scev_analysis): Likewise.
14243
83082391 142442013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 14245 Chao-ying Fu <fu@mips.com>
83082391 14246
cabf91cd
AM
14247 * config/mips/micromips.md (jraddiusp): New pattern.
14248 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
14249 instruction if possible.
83082391 14250
19e34aa2
AM
142512013-04-24 Alan Modra <amodra@gmail.com>
14252
14253 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
14254
fdb6603c
JB
142552013-04-24 Julian Brown <julian@codesourcery.com>
14256 Chung-Lin Tang <cltang@codesourcery.com>
14257
14258 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB
14259 dependency behavior in enumeration type DIE generation. Add TODO note
14260 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 14261
bf190e8d
LC
142622013-04-23 Lawrence Crowl <crowl@google.com>
14263
14264 * Makefile.in: Update as needed below.
14265
14266 * hash-table.h (class hash_table):
14267 Correct many methods with parameter types compare_type to the correct
14268 value_type. (Correct code was unlikely to notice the change.)
14269 (hash_table::elements_with_deleted) New.
14270 (class hashtable::iterator): New.
14271 (hashtable::begin()): New.
14272 (hashtable::end()): New.
14273 (FOR_EACH_HASH_TABLE_ELEMENT): New.
14274
14275 * statistics.c (statistics_hashes):
14276 Change type to hash_table. Update dependent calls and types.
14277
14278 * tree-into-ssa.c (var_infos):
14279 Change type to hash_table. Update dependent calls and types.
14280
14281 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
14282 Change type to hash_table. Update dependent calls and types.
14283
14284 * tree-ssa-loop-im.c (struct mem_ref.refs):
14285 Change type to hash_table. Update dependent calls and types.
14286
14287 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
14288 Change type to hash_table. Update dependent calls and types.
14289
14290 * tree-ssa-sccvn.c (vn_tables_s::nary):
14291 Change type to hash_table. Update dependent calls and types.
14292 (vn_tables_s::phis): Likewise.
14293 (vn_tables_s::references): Likewise.
14294
14295 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
14296 (vn_reference_eq): Update parameter and return types.
14297
14298 * tree-ssa-structalias.c (pointer_equiv_class_table):
14299 Change type to hash_table. Update dependent calls and types.
14300 (location_equiv_class_table): Likewise.
14301
14302 * tree-vect-data-refs.c: Consequential changes for making
14303 peeling a hash_table.
14304
14305 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
14306 (destroy_loop_vec_info): Dependent hash_table update.
14307
14308 * tree-vectorizer.h (peeling_htab):
14309 Change type to hash_table. Update dependent calls and types.
14310
d70a81dd
SC
143112013-04-23 Shiva Chen <shiva0217@gmail.com>
14312
cabf91cd
AM
14313 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
14314 to check the register content is equal or not.
14315 * lra-constraints.c (match_reload): Use lra_assign_reg_val
14316 to assign register content record.
14317 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 14318 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM
14319 * lra-int.h (struct lra_reg): Add offset member.
14320 (lra_reg_val_equal_p): New static inline function.
14321 (lra_update_reg_val_offset): New static inline function.
14322 (lra_assign_reg_val): New static inline function.
14323 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
14324 to assign register content record.
14325 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 14326
b894a1f3
CM
143272013-04-23 Catherine Moore <clm@codesourcery.com>
14328
14329 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
14330 operands. Record compression.
14331
ecd14de9
XDL
143322013-04-23 Xinliang David Li <davidxl@google.com>
14333
14334 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
14335
92e776e9
RB
143362013-04-23 Richard Biener <rguenther@suse.de>
14337
14338 PR middle-end/57036
14339 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
14340 parameter, only add abnormal goto edges from the copied body
14341 if the call could perform abnormal gotos.
14342 (copy_cfg_body): Adjust.
14343
a15ee567
SN
143442013-04-23 Sofiane Naci <sofiane.naci@arm.com>
14345
14346 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
14347
08c52234
AS
143482013-04-23 Andreas Schwab <schwab@linux-m68k.org>
14349
14350 * coretypes.h (gimple_stmt_iterator): Add struct to make
14351 compatible with C.
14352
999c1171
RB
143532013-04-23 Richard Biener <rguenther@suse.de>
14354
14355 PR tree-optimization/57026
14356 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
14357 from SSA names occuring in abnormal PHI nodes.
14358
53669259
AK
143592013-04-22 Andi Kleen <ak@linux.intel.com>
14360
14361 * lto/lto.c (print_lto_report_1): Fix LTO report names.
14362
1a0ad150
AK
143632013-04-22 Andi Kleen <ak@linux.intel.com>
14364
14365 * lto/lto.c (print_lto_report_1): Declare early.
14366 (read_cgraph_and_symbols): Call print_lto_report_1 early.
14367
057f8f20
AK
143682013-04-22 Andi Kleen <ak@linux.intel.com>
14369
14370 * common.opt (-flto-report-wpa): Add.
14371 * doc/invoke.texi (-flto-report-wpa): Add.
14372 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
14373 (lto_main): dito.
14374
473b1e05
XDL
143752013-04-22 Xinliang David Li <davidxl@google.com>
14376
14377 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
14378 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
14379 * Makefile.in: New dependency
14380
cabf91cd 14381 David Daney <ddaney.cavm@gmail.com>
b1485a33 14382
cabf91cd
AM
14383 * configure.ac (gcc_cv_as_micromips_support): Use the
14384 --fatal-warnings option.
14385 * configure: Regenerate.
b1485a33 14386
829d0168
MP
143872013-04-22 Marek Polacek <polacek@redhat.com>
14388
14389 PR sanitizer/56990
14390 * tsan.c (instrument_expr): Don't instrument expression
14391 in case its size is zero.
14392
6d9b7208
UB
143932013-04-22 Uros Bizjak <ubizjak@gmail.com>
14394
14395 PR target/57032
14396 Revert:
14397 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
14398
14399 * config/alpha/alpha.c (TARGET_LRA_P): New define.
14400
ea679d55
JG
144012013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14402
14403 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
14404 (gimple_stmt_iterator): New typedef.
14405 * gimple.h (gimple_stmt_iterator): Rename to...
14406 (gimple_stmt_iterator_d): ... This.
14407 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
14408 trees be valid for GIMPLE and GENERIC.
14409 (TARGET_GIMPLE_FOLD_BUILTIN): New.
14410 * gimple-fold.c (gimple_fold_call): Call target hook
14411 gimple_fold_builtin.
14412 * hooks.c (hook_bool_gsiptr_false): New.
14413 * hooks.h (hook_bool_gsiptr_false): New.
14414 * target.def (fold_stmt): New.
14415 * doc/tm.texi: Regenerate.
14416
88a581da
VM
144172013-04-22 Vladimir Makarov <vmakarov@redhat.com>
14418
14419 PR target/57018
14420 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
14421 a set sp if no stack realignment.
14422
92be22dc
NC
144232013-04-22 Nick Clifton <nickc@redhat.com>
14424
14425 * config.gcc (tilegx-linux): Extend extra_objs rather than
14426 overwriting it.
14427 (tilepro-linux): Likewise.
14428
0ddec79f
JG
144292013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14430
14431 * config/aarch64/aarch64-builtins.c
14432 (CF): Remove.
14433 (CF0, CF1, CF2, CF3, CF4, CF10): New.
14434 (VAR<1-12>): Add MAP parameter.
14435 (BUILTIN_*): Likewise.
14436 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
14437 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
14438 (aarch64_ushl_n<mode>): Likewise.
14439 (aarch64_sshr_n<mode>): Likewise.
14440 (aarch64_ushr_n<mode>): Likewise.
14441 (aarch64_<maxmin><mode>): Likewise.
14442 (aarch64_sqrt<mode>): Likewise.
14443 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
14444 (vshr<q>_n_*): Likewise.
14445
0050faf8
JG
144462013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
14447
14448 * config/aarch64/aarch64-builtins.c
14449 (aarch64_simd_builtin_type_mode): Handle SF types.
14450 (sf_UP): Define.
14451 (BUILTIN_GPF): Define.
14452 (aarch64_init_simd_builtins): Handle SF types.
14453 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
14454 (frecps): Likewise.
14455 (frecpx): Likewise.
14456 * config/aarch64/aarch64-simd.md
14457 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
14458 (aarch64_frecpe<mode>): New.
14459 (aarch64_frecps<mode>): Likewise.
14460 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
14461 (v8type): Add frecp<esx>.
14462 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
14463 (aarch64_frecps<mode>): Likewise.
14464 * config/aarch64/iterators.md (FRECP): New.
14465 (frecp_suffix): Likewise.
14466 * config/aarch64/arm_neon.h
14467 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
14468
0fad3dbc 144692013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB
14470
14471 PR target/56995
14472 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
14473 (REG_CLASS_NAMES): Idem.
14474 (REG_CLASS_CONTENTS): Idem.
14475 (REGCLASS_HAS_FP_REG): Idem.
14476 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
14477 (sh_conditional_register_usage): Idem.
14478
3e8a33f9
JL
144792013-04-21 Jeff Law <law@redhat.com>
14480
14481 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
14482 (ssa_forward_propagate_and_combine): Use it.
14483
f38e1b0a
VM
144842013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14485
14486 * lra.c: Update the flow chart diagram.
14487
682303da
VM
144882013-04-19 Vladimir Makarov <vmakarov@redhat.com>
14489
14490 PR rtl-optimization/56847
14491 * lra-constraints.c (process_alt_operands): Discourage alternative
14492 with non-matche doffsettable memory constraint fro memory with
14493 known offset.
14494
f6b64c35
RB
144952013-04-19 Richard Biener <rguenther@suse.de>
14496
14497 PR tree-optimization/56982
14498 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
14499 function.
14500 * gimplify.c (gimplify_call_expr): Notice special calls.
14501 (gimplify_modify_expr): Likewise.
14502 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
14503 abnormal control flow receivers.
14504 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
14505 in the same way as cfun->has_nonlocal_labels.
14506 (gimple_purge_dead_abnormal_call_edges): Likewise.
14507 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
14508 receivers start a basic-block.
14509
01d8bf07
RB
145102013-04-19 Richard Biener <rguenther@suse.de>
14511
14512 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
14513 member ...
14514 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
14515 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
14516 (SLP_TREE_LOAD_PERMUTATION): Add.
14517 (vect_transform_slp_perm_load): Adjust prototype.
14518 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
14519 (vect_free_slp_instance): Likewise.
14520 (vect_create_new_slp_node): Likewise.
14521 (vect_supported_slp_permutation_p): Remove.
14522 (vect_slp_rearrange_stmts): Adjust.
14523 (vect_supported_load_permutation_p): Likewise. Inline
14524 vect_supported_slp_permutation_p here.
14525 (vect_analyze_slp_instance): Compute load permutations per
14526 slp node instead of per instance.
14527 (vect_get_slp_defs): Adjust.
14528 (vect_transform_slp_perm_load): Likewise.
14529 (vect_schedule_slp_instance): Remove redundant code.
14530 (vect_schedule_slp): Remove hack for PR56270, add it ...
14531 * tree-vect-stmts.c (vectorizable_load): ... here, do not
14532 CSE loads for SLP. Adjust.
14533
ede22fc3
GY
145342013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14535
14536 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
14537 spelling in two comments.
14538
67bc84fb
GY
145392013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
14540
14541 PR target/56797
14542 * config/arm/arm.c (load_multiple_sequence): Require SP
14543 as base register for loads if SP is in the register list.
14544
e248d83f
MJ
145452013-04-19 Martin Jambor <mjambor@suse.cz>
14546
14547 PR tree-optimization/56718
14548 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
14549 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
14550 and made public. Adjusted all callers.
14551 (ipa_intraprocedural_devirtualization): New function.
14552 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
14553 (ipa_intraprocedural_devirtualization): Likewise.
14554 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
14555
4891e8f8
RB
145562013-04-19 Richard Biener <rguenther@suse.de>
14557
14558 PR tree-optimization/57000
14559 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
14560
dad89f7c
TG
145612013-04-19 Terry Guo <terry.guo@arm.com>
14562
14563 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
14564 Replace with ...
14565 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
14566 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
14567 (cortex_m4_fmacs): Use new reservations.
14568 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
14569
72ea0d47
VM
145702013-04-18 Vladimir Makarov <vmakarov@redhat.com>
14571
f1e6512c 14572 PR rtl-optimization/56999
72ea0d47
VM
14573 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
14574 related code.
14575 (lra_coalesce): Remove split_origin_bitmap and related code.
14576 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
14577 ranges if necessary.
14578
780a5b71
UB
145792013-04-18 Uros Bizjak <ubizjak@gmail.com>
14580
14581 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
14582 New array.
14583 (ix86_expand_call): Remove clobbered_registers array and use
14584 x86_64_ms_sysv_extra_clobbered_registers instead.
14585 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
14586 Declare here.
14587 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
14588 predicate.
14589 * config/i386/i386.md (*call_rex64_ms_sysv): Use
14590 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
14591 (*call_value_rex64_ms_sysv): Ditto.
14592
6f5a366a
CC
145932013-04-18 Cary Coutant <ccoutant@google.com>
14594
14595 * dwarf2out.c (output_pubnames): Check die_perennial_p of
14596 parent instead of die_mark.
14597
475b8f37
DN
145982013-04-18 Diego Novillo <dnovillo@google.com>
14599
14600 * gimple.c (create_gimple_tmp): New.
14601 (get_expr_type): New.
14602 (build_assign): New.
14603 (build_type_cast): New.
14604 * gimple.h (enum ssa_mode): Define.
14605 (gimple_seq_set_location): New.
14606 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 14607 to use build_assign and build_type_cast.
475b8f37 14608
08940f33
RB
146092013-04-18 Richard Biener <rguenther@suse.de>
14610
14611 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
14612 handle negative step. Remove redundant checks.
14613 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
14614 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
14615 for negative step and grouped loads fail to vectorize.
14616
0e0f87d4
SB
146172013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14618
14619 * emit-rtl.c (reset_insn_used_flags): New function.
14620 (reset_all_used_flags): Use it.
14621 (verify_insn_sharing): New function.
14622 (verify_rtl_sharing): Fix verification for SEQUENCEs.
14623
4c445590
JJ
146242013-04-18 Jakub Jelinek <jakub@redhat.com>
14625
14626 PR tree-optimization/56984
14627 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
14628 and (x >> M) >= N don't register any assertion if N << M is the
14629 minimum value.
14630
6873ecab
SB
146312013-04-18 Steven Bosscher <steven@gcc.gnu.org>
14632
14633 * lower-subreg.c (resolve_simple_move): If called self-recursive,
14634 do not delete_insn insns that have not yet been emitted, only
14635 unlink them with remove_insn.
14636 * df-scan.c (df_insn_delete): Revert r197492.
14637
3ccb989e
SB
146382013-04-17 Steven Bosscher <steven@gcc.gnu.org>
14639
14640 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
14641 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
14642
146432013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY
14644
14645 * config/arm/arm.md (movsicc_insn): Convert define_insn into
14646 define_insn_and_split.
14647 (and_scc,ior_scc,negscc): Likewise.
14648 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
14649
3ccb989e 146502013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY
14651
14652 * config/arm/arm.c (use_return_insn): Return 0 for targets that
14653 can benefit from using a sequence of LDRD instructions in epilogue
14654 instead of a single LDM instruction.
14655
6d10a203
MLI
146562013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
14657
14658 PR 45688
14659 * doc/extend.texi: Fix typo.
14660
6983e6b5
RB
146612013-04-17 Richard Biener <rguenther@suse.de>
14662
14663 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
14664 (vect_build_slp_tree): ... here.
14665 (vect_build_slp_tree_1): Compute which stmts of the SLP group
14666 match. Remove special-casing of mismatched complex loads.
14667 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
14668 re-try the match with swapped commutative operands.
14669 (vect_supported_load_permutation_p): Remove special-casing of
14670 mismatched complex loads.
14671 (vect_analyze_slp_instance): Adjust.
14672
ef23e6a2
RB
146732013-04-17 Richard Biener <rguenther@suse.de>
14674
14675 PR rtl-optimization/56921
14676 * cfgloop.h (struct loop): Add simple_loop_desc member.
14677 (struct niter_desc): Mark with GTY(()).
14678 (simple_loop_desc): Do not use aux field but simple_loop_desc.
14679 * loop-iv.c (get_simple_loop_desc): Likewise.
14680 (free_simple_loop_desc): Likewise.
14681
14682 Revert
14683 2013-04-16 Richard Biener <rguenther@suse.de>
14684
14685 PR rtl-optimization/56921
14686 * loop-init.c (pass_rtl_move_loop_invariants): Add
14687 TODO_do_not_ggc_collect to todo_flags_finish.
14688 (pass_rtl_unswitch): Same.
14689 (pass_rtl_unroll_and_peel_loops): Same.
14690 (pass_rtl_doloop): Same.
14691
fc6f94f5
EB
146922013-04-17 Eric Botcazou <ebotcazou@adacore.com>
14693
14694 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
14695 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
14696 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
14697 references.
14698 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
14699 * tree-streamer.c (record_common_node): Adjust reference in comment.
14700
10a88311
TG
147012013-04-17 Terry Guo <terry.guo@arm.com>
14702
14703 * config/arm/cortex-m4.md: Add a new bypass.
14704
6d9b7208 147052013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N
14706
14707 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
14708 New pattern.
14709 (*subs_<optab><mode>_multp2): New pattern.
14710 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14711 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
14712
6d9b7208 147132013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N
14714
14715 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
14716 (*subs_mul_imm_<mode>): New pattern.
14717
18a6701e
DE
147182013-04-16 David Edelsohn <dje.gcc@gmail.com>
14719
14720 PR target/56948
14721 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
14722 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
14723 (vsx_movti_32bit): Change j->wa to O->wa.
14724
07c37b2f
RB
147252013-04-16 Richard Biener <rguenther@suse.de>
14726
14727 PR rtl-optimization/56921
14728 * loop-init.c (pass_rtl_move_loop_invariants): Add
14729 TODO_do_not_ggc_collect to todo_flags_finish.
14730 (pass_rtl_unswitch): Same.
14731 (pass_rtl_unroll_and_peel_loops): Same.
14732 (pass_rtl_doloop): Same.
14733
0e0f87d4 147342013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY
14735
14736 * config/arm/arm.c (emit_multi_reg_push): New declaration
14737 for an existing function.
14738 (arm_emit_strd_push): New function.
14739 (arm_expand_prologue): Used here.
14740 (arm_emit_ldrd_pop): New function.
14741 (arm_expand_epilogue): Used here.
14742 (arm_get_frame_offsets): Update condition.
14743 (arm_emit_multi_reg_pop): Add a special case for load of a single
14744 register with writeback.
14745
5e8e2af4
UB
147462013-04-16 Uros Bizjak <ubizjak@gmail.com>
14747
14748 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
14749 description.
14750
9fd9ccf7
RB
147512013-04-16 Richard Biener <rguenther@suse.de>
14752
14753 PR tree-optimization/56756
14754 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
14755 (first_mem_ref_loc): New.
14756 (execute_sm): Place the load temporarily before a previous
14757 access instead of in the latch edge to ensure its SSA dependencies
14758 are defined at points dominating the load.
14759
96fba521
SB
147602013-04-16 Steven Bosscher <steven@gcc.gnu.org>
14761
4c8af858
SB
14762 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
14763 correct fix by moving header and footer insn to the footer of
14764 the merged basic block. Clear BB_END of the merged-away block.
14765
96fba521
SB
14766 PR middle-end/43631
14767 * emit-rtl.c (make_note_raw): New function.
14768 (link_insn_into_chain): New static inline function.
14769 (add_insn): Use it.
14770 (add_insn_before, add_insn_after): Factor insn chain linking code...
14771 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
14772 using link_insn_into_chain.
14773 (note_outside_basic_block_p): New helper function for emit_note_after
14774 and emit_note_before.
14775 (emit_note_after): Use nobb variant of add_insn_after if the note
14776 should not be contained in a basic block.
14777 (emit_note_before): Use nobb variant of add_insn_before if the note
14778 should not be contained in a basic block.
14779 (emit_note_copy): Use make_note_raw.
14780 (emit_note): Likewise.
14781 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
14782 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
14783 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
14784 the moved barrier the tail of the basic block it follows.
14785 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
14786
7b8265ba
JJ
147872013-04-15 Jakub Jelinek <jakub@redhat.com>
14788
14789 PR tree-optimization/56962
14790 * gimple-ssa-strength-reduction.c (record_increment): Only set
14791 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
14792 either rhs1 or rhs2 is equal to c->base_expr.
14793
5185d248
RB
147942013-04-15 Richard Biener <rguenther@suse.de>
14795
14796 PR tree-optimization/56933
14797 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
14798 member.
14799 (GROUP_READ_WRITE_DEPENDENCE): Remove.
14800 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
14801 * tree-vect-data-refs.c (vect_analyze_group_access): Move
14802 dependence check ...
14803 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
14804 ... here.
14805 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
14806 GROUP_READ_WRITE_DEPENDENCE.
14807
a24243a0
AK
148082013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14809
14810 * emit-rtl.c (reset_all_used_flags): New function.
14811 (verify_rtl_sharing): Call reset_all_used_flags before and after
14812 performing the checks.
14813
1c50eada
KT
148142013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14815
14816 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
14817 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
14818 * config/arm/constraints.md (De): New constraint.
14819 * config/arm/neon.md (anddi3_neon): Delete.
14820 (neon_vand<mode>): Expand to standard anddi3 pattern.
14821 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
14822 Move earlier in the file.
14823 (neon_inv_logic_op2): Likewise.
14824 (arm_anddi_operand_neon): New predicate.
14825
e927b6ad
RO
148262013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
14827
14828 * configure.ac (gcc_cv_ld_as_needed): Set
14829 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
14830 Use -z ignore, -z record on *-*-solaris2*.
14831 (HAVE_LD_AS_NEEDED): Update comment.
14832 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
14833 * configure: Regenerate.
14834 * config.in: Regenerate.
14835 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
14836 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
14837 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
14838 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
14839 equivalents. Fix markup.
14840 * doc/tm.texi: Regenerate.
14841
e0ea8797
AH
148422013-04-15 Andrew Hsieh <andrewhsieh.google.com>
14843
14844 * config/i386/i386.opt: New option mstack-protector-guard=.
14845 * config/i386/i386-opts.h: Add enum stack_protector_guard.
14846 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
14847 TARGET_SSP_TLS_GUARD.
14848 * config/i386/i386.c (ix86_option_override_internal): Set
14849 ix86_stack_protector_guard.
14850 * config/i386/i386.md (stack_protect_set): Enable for
14851 TARGET_SSP_TLS_GUARD only.
14852 (stack_protect_set_<mode>): Ditto.
14853 (stack_protect_test): Ditto.
14854 (stack_protect_test_<mode>): Ditto.
14855 * doc/invoke.texi (i386 Option): Document.
14856
811b72f9
EB
148572013-04-15 Eric Botcazou <ebotcazou@adacore.com>
14858
14859 PR target/56890
14860 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
14861 (S_MODES): Set H_MODE bit.
14862 (SF_MODES): Set only S_MODE and SF_MODE bits.
14863 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
14864 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
14865 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
14866 <MODE_FLOAT>: Likewise.
14867
5529fdd6
JY
148682013-04-15 Joey Ye <joey.ye@arm.com>
14869
14870 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
14871
517b1da2
JY
148722013-04-15 Joey Ye <joey.ye@arm.com>
14873
14874 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
14875 for real far jump.
14876 (thumb_far_jump_used_p): Count instruction size and set
14877 far_jump_used.
14878
01007ae0
EB
148792013-04-14 Eric Botcazou <ebotcazou@adacore.com>
14880
14881 * reorg.c (fill_simple_delay_slots): Reindent block of code.
14882 * resource.c (mark_target_live_regs): Reformat conditional block.
14883
c46f6580
SB
148842013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14885
14886 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
14887 notes, they are emitted only just before final.
14888 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
14889
1f397f45
SB
148902013-04-13 Steven Bosscher <steven@gcc.gnu.org>
14891
14892 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
14893 * cfgrtl.c (delete_insn): Call it here instead.
14894 * lra-spills.c (lra_final_code_change): Use delete_insn.
14895 * haifa-sched.c (sched_remove_insn): Likewise.
14896 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
14897 returning to the nop pool.
14898 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
14899 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
14900
58a51369
SB
149012013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14902
14903 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
14904 * doc/tm.texi: Regenerated.
14905
33159866
UB
149062013-04-12 Uros Bizjak <ubizjak@gmail.com>
14907
14908 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
14909 QImode checks.
14910
226e378f
SB
149112013-04-12 Steven Bosscher <steven@gcc.gnu.org>
14912
14913 * df-core.c (df_find_def): Compare register numbers.
14914 (df_find_use): Likewise.
14915
fafb9b18
VM
149162013-04-12 Vladimir Makarov <vmakarov@redhat.com>
14917
14918 PR target/56903
14919 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
14920 lra_in_progress for return.
14921
9a946fd6
GY
149222013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14923
14924 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
14925 define_insn into define_insn_and_split and emit movsicc patterns.
14926
41b83758
GY
149272013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
14928
14929 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
14930
d3afd9aa
RB
149312013-04-12 Richard Biener <rguenther@suse.de>
14932
14933 * tree-pass.h (TODO_do_not_ggc_collect): New.
14934 * passes.c (execute_one_ipa_transform_pass): Honor
14935 TODO_do_not_ggc_collect.
14936 (execute_one_pass): Likewise.
14937
14938 Revert
14939 2013-04-10 Richard Biener <rguenther@suse.de>
14940
14941 * passes.c (init_optimization_passes): Remove reload pass.
14942 * ira.c (do_reload): Merge into ...
14943 (ira): ... this.
14944 (rest_of_handle_reload): Remove.
14945 (pass_reload): Likewise.
14946 * config/i386/i386.c (ix86_option_override): Refer to ira instead
14947 of reload for vzeroupper pass placement.
14948
06f9b387
JJ
149492013-04-12 Jakub Jelinek <jakub@redhat.com>
14950
14951 PR tree-optimization/56918
14952 PR tree-optimization/56920
14953 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
14954 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
14955 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
14956 use wide_mul_with_sign method.
14957
953094d2
RB
149582013-04-12 Richard Biener <rguenther@suse.de>
14959
14960 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
14961 not be considered a gimple constant.
14962
bb506982
MG
149632013-04-12 Marc Glisse <marc.glisse@inria.fr>
14964
14965 * fold-const.c (const_binop): Handle vector shifts by a scalar.
14966 (fold_binary_loc): Call const_binop also for mixed vector-scalar
14967 operations.
14968
4b84d650
JJ
149692013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
14970 Jakub Jelinek <jakub@redhat.com>
14971
14972 * opts.c: Include diagnostic-color.h.
14973 (common_handle_option): Handle OPT_fdiagnostics_color_.
14974 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
14975 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
14976 (diagnostic-color.o): New.
14977 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
14978 (diagnostic_color_rule): New enum.
14979 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
14980 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
14981 the location string.
14982 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
14983 either NULL, or color kind.
14984 * diagnostic-color.c: New file.
14985 * diagnostic-color.h: New file.
14986 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
14987 arguments.
14988 * doc/invoke.texi (-fdiagnostics-color): Document.
14989 * pretty-print.h (pp_show_color): Define.
14990 (struct pretty_print_info): Add show_color field.
14991 * diagnostic.c: Include diagnostic-color.h.
14992 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
14993 macros. Colorize error:, warning: etc. strings and also the location
14994 string.
14995 (diagnostic_show_locus): Colorize the caret line.
14996 * pretty-print.c: Include diagnostic-color.h.
14997 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
14998 inside of %< %> quotes or quoted through q format modifier.
14999
067a1e71
AK
150002013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15001
33159866 15002 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 15003
33159866 150042013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N
15005
15006 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
15007 code in CC_NZ mode.
15008 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
15009 pattern.
15010
7b55f98f
MP
150112013-04-11 Marek Polacek <polacek@redhat.com>
15012
15013 PR tree-optimization/48184
33159866 15014 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 15015
966b587e
EB
150162013-04-11 Eric Botcazou <ebotcazou@adacore.com>
15017
15018 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
15019 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
15020 (skip_simple_arithmetic): Tidy up.
15021 * tree.h (skip_simple_constant_arithmetic): Declare.
15022
33159866 150232013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N
15024
15025 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
15026
1d42c1ec
RB
150272013-04-11 Richard Biener <rguenther@suse.de>
15028
15029 * tree-vect-loop.c (get_initial_def_for_induction): Properly
15030 generate vector constants.
15031
4ba5ea11
RB
150322013-04-11 Richard Biener <rguenther@suse.de>
15033
15034 PR tree-optimization/56878
15035 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
15036 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
15037 New function.
15038 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
15039 Prefer to align the DR with the most invariant base address.
15040
f0defe58
SKS
150412013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
15042
15043 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
15044 comment.
15045
d07458be
JG
150462013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
15047
15048 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
15049 floating-point vector comparisons against 0.
15050
146b8692
JJ
150512013-04-11 Jakub Jelinek <jakub@redhat.com>
15052
15053 PR tree-optimization/56899
15054 * fold-const.c (extract_muldiv_1): Apply distributive law
15055 only if TYPE_OVERFLOW_WRAPS (ctype).
15056
b8578ff7
BC
150572013-04-11 Bin Cheng <bin.cheng@arm.com>
15058
15059 PR target/56124
15060 * ira-costs.c (scan_one_insn): Check whether the source rtx of
15061 loading has side effect.
15062
0ea8a6f9
SB
150632013-04-10 Steven Bosscher <steven@gcc.gnu.org>
15064
15065 * config/sparc/sparc.c: Include tree-pass.h.
15066 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
15067 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
15068 head of file. Change return type. Split off gate function.
15069 (sparc_gate_work_around_errata): New function.
15070 (pass_work_around_errata): New pass definition.
15071 (insert_pass_work_around_errata) New pass insert definition to
15072 insert pass_work_around_errata just after delayed-branch scheduling.
15073 (sparc_option_override): Insert the pass.
15074 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
15075
42e37616
DM
150762013-04-10 David S. Miller <davem@davemloft.net>
15077
89deeb3b
DM
15078 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
15079 or -mcpu=hypersparc.
15080
42e37616
DM
15081 * target.def (cstore_mode): New hook.
15082 * target.h: Include insn-codes.h
15083 * targhooks.c: Likewise.
15084 (default_cstore_mode): New function.
15085 * targhooks.h: Declare it.
15086 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
15087 * doc/tm.texi: Rebuild.
15088 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
15089 target hook, rather than inspecting the insn_data.
15090 * config/sparc/sparc.c (sparc_cstore_mode): New function.
15091 (TARGET_CSTORE_MODE): Redefine.
15092 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
15093 result patterns.
15094 * config/sparc/predicates.md (cstore_result_operand): New special
15095 predicate.
15096 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
15097 Use it for operand 0.
15098 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
15099 (*snesi_special): Likewise.
15100 (*snesi_zero): Likewise.
15101 (*seqsi_zero): Likewise.
15102 (*sltu_insn): Likewise.
15103 (*sgeu_insn): Likewise.
15104 (*seqdi_special): Make operand 0 and comparison operation be of
15105 DImode.
15106 (*snedi_special): Likewise.
15107 (*snedi_special_vis3): Likewise.
15108 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
15109 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
15110 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
15111 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
15112 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
15113 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
15114 (*sltu_extend_sp64): Likewise.
15115 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
15116 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
15117 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
15118 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
15119 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
15120
95ca411e
YZ
151212013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
15122
15123 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
15124 (aarch64_start_file): Use the new function.
15125
6782438d 151262013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 15127 Jason Merrill <jason@redhat.com>
6782438d
SKS
15128
15129 * common.opt: Add -gdwarf.
15130 * opts.c (common_handle_option): Handle it.
15131 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
15132
bb313b93
RB
151332013-04-10 Richard Biener <rguenther@suse.de>
15134
15135 * passes.c (execute_todo): Do not call ggc_collect conditional here.
15136 (execute_one_ipa_transform_pass): But unconditionally here.
15137 (execute_one_pass): And here.
15138 (init_optimization_passes): Remove reload pass.
15139 * tree-pass.h (TODO_ggc_collect): Remove.
15140 (pass_reload): Likewise.
15141 * ira.c (do_reload): Merge into ...
15142 (ira): ... this.
15143 (rest_of_handle_reload): Remove.
15144 (pass_reload): Likewise.
15145 * config/i386/i386.c (ix86_option_override): Refer to ira instead
15146 of reload for vzeroupper pass placement.
15147 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
15148 and todo_flags_finish of all passes.
15149
793d9a16
RB
151502013-04-10 Richard Biener <rguenther@suse.de>
15151
15152 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
15153 first_const_oprnd field, rename first_def_type to first_op_type.
15154 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
15155 (vect_get_and_check_slp_defs): Always use the type of the
15156 operand. Allow mixed vect_external_def, vect_constant_def types.
15157 (vect_get_constant_vectors): Handle mixed vect_external_def,
15158 vect_constant_def types.
15159
12211b99 151602013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR
15161
15162 PR tree-optimization/55524
15163 * tree-ssa-math-opts.c
15164 (convert_mult_to_fma): Don't use an fms construct
15165 when we don't have an fms operation, but fnma, and it looks
15166 likely that we'll be able to use the latter.
15167
12211b99 151682013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ
15169
15170 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
15171 function.
15172 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
15173 inline fail caused by overwritable functions.
15174
34ab4a5b
CJW
151752013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
15176
15177 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
15178 unnecessary bits in the constant power of two case.
15179
abf9bfbc
RB
151802013-04-10 Richard Biener <rguenther@suse.de>
15181
15182 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
15183 broken code swapping operands.
15184 (vect_build_slp_tree): Do not compute load permutations here.
15185 (vect_analyze_slp_instance): Compute load permutations here,
15186 after building the SLP tree.
15187
f408477e
CB
151882013-04-09 Christian Bruel <christian.bruel@st.com>
15189
15190 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
15191 of next/prev_real_insn.
15192
5ac42672
JH
151932013-04-09 Jan Hubicka <jh@suse.cz>
15194
abf9bfbc
RB
15195 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15196 Drop aliased parameter.
5ac42672
JH
15197 (function_and_variable_visibility): Do not handle alias pairs.
15198 * cgraph.c (varpool_externally_visible_p): Update prototype.
15199 * varpool.c (varpool_add_new_variable): Update.
15200
5017f1d2
KT
152012013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15202
15203 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
15204
48eecbee
SB
152052013-04-09 Steven Bosscher <steven@gcc.gnu.org>
15206
a949cf1c
SB
15207 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
15208
48eecbee
SB
15209 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
15210
75ef8e3d
MP
152112013-04-09 Marek Polacek <polacek@redhat.com>
15212
15213 PR tree-optimization/48762
33159866 15214 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 15215
23847df4
RB
152162013-04-09 Richard Biener <rguenther@suse.de>
15217
15218 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
15219 dealing with cost.
15220 (vect_build_slp_tree): Likewise.
15221 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
15222 calculating the cost of a SLP instance.
15223 (vect_analyze_slp_instance): Use it from here, after building
15224 the SLP tree.
15225
ea3a0fde
JJ
152262013-04-09 Jakub Jelinek <jakub@redhat.com>
15227
15228 PR middle-end/56883
15229 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
15230 expand_omp_for_static_chunk): Use simple_p = true in
15231 force_gimple_operand_gsi calls when assigning to addressable decls.
15232
a32dfe9d
JL
152332013-04-09 Jeff Law <law@redhat.com>
15234
15235 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
15236 when the boolean was created by converting a wider object which
15237 had a boolean range.
15238
d755c7ef
RB
152392013-04-09 Richard Biener <rguenther@suse.de>
15240
15241 * tree-vectorizer.h (slp_void_p): Remove.
15242 (slp_tree): Typedef before _slp_tree declaration.
15243 (struct _slp_tree): Use a vector of slp_tree as children.
15244 (vect_get_place_in_interleaving_chain): Remove.
15245 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
15246 Move ...
15247 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
15248 and make static.
15249 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
15250 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
15251 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
15252 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
15253 Use slp_node instead of slp_void_p and adjust.
15254
3d741091
RB
152552013-04-09 Richard Biener <rguenther@suse.de>
15256
15257 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
15258 work that is not necessary.
15259
39307ba7
JJ
152602013-04-09 Jakub Jelinek <jakub@redhat.com>
15261
15262 PR tree-optimization/56854
15263 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
15264 forward into clobber stmts if it would change MEM_REF lhs into
15265 non-MEM_REF.
15266
343881fd
MK
152672013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
15268
15269 * tree.c (type_hash_lookup, type_hash_add): Make static.
15270 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
15271
3922658a
RB
152722013-04-09 Richard Biener <rguenther@suse.de>
15273
15274 * tree.h (unsave_expr_now): Remove.
15275 * tree-inline.c (mark_local_for_remap_r): Remove.
15276 (unsave_expr_1): Likewise.
15277 (unsave_r): Likewise.
15278 (unsave_expr_now): Likewise.
15279 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
15280 (propagate_tree_value): Likewise.
15281
9fb6b620
SB
152822013-04-08 Steven Bosscher <steven@gcc.gnu.org>
15283
15284 * doc/rtl.texi (sequence): Rewrite documentation to match the
15285 current use of SEQUENCE rtl objects.
15286 * rtl.def (SEQUENCE): Likewise.
15287
15288 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
15289 Update documentation.
15290 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
15291 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
15292
15293 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
15294
8ddb5a29
TJ
152952013-04-08 Teresa Johnson <tejohnson@google.com>
15296
15297 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
15298 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 15299 methods.
8ddb5a29
TJ
15300 (estimate_edge_size_and_time): Add comment to suggest using rounding
15301 methods.
15302 (estimate_node_size_and_time): Ditto.
15303 (remap_edge_change_prob): Use helper rounding divide methods.
15304 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
15305 (gimple_mod_pow2_value_transform): Ditto.
15306 (gimple_mod_subtract_transform): Ditto.
15307 (gimple_ic_transform): Ditto.
15308 (gimple_stringops_transform): Ditto.
15309 * stmt.c (conditional_probability): Ditto.
15310 (emit_case_dispatch_table): Ditto.
15311 * lto-cgraph.c (merge_profile_summaries): Ditto.
15312 * tree-optimize.c (execute_fixup_cfg): Ditto.
15313 * cfgcleanup.c (try_forward_edges): Ditto.
15314 * cfgloopmanip.c (scale_loop_profile): Ditto.
15315 (loopify): Ditto.
15316 (duplicate_loop_to_header_edge): Ditto.
15317 (lv_adjust_loop_entry_edge): Ditto.
15318 * tree-vect-loop.c (vect_transform_loop): Ditto.
15319 * profile.c (compute_branch_probabilities): Ditto.
15320 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
15321 * lto-streamer-in.c (input_cfg): Ditto.
15322 * gimple-streamer-in.c (input_bb): Ditto.
15323 * ipa-cp.c (update_profiling_info): Ditto.
15324 (update_specialized_profile): Ditto.
15325 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
15326 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 15327 rounding methods.
8ddb5a29
TJ
15328 * sched-rgn.c (compute_dom_prob_ps): Ditto.
15329 (compute_trg_info): Ditto.
15330 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
15331 (purge_dead_edges): Ditto.
15332 * loop-unswitch.c (unswitch_loop): Ditto.
15333 * cgraphclones.c (cgraph_clone_edge): Ditto.
15334 (cgraph_clone_node): Ditto.
15335 * tree-inline.c (copy_bb): Ditto.
15336 (copy_edges_for_bb): Ditto.
15337 (initialize_cfun): Ditto.
15338 (copy_cfg_body): Ditto.
15339 (expand_call_inline): Ditto.
15340
661e6bd7
KT
153412013-04-08 Kai Tietz <ktietz@redhat.com>
15342
15343 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
15344 TARGET_CYGWIN64 by TARGET_64BIT.
15345
105766f3
JR
153462013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15347
15348 * config/epiphany/epiphany.md (GPR_1): New constant.
15349 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
15350 * config/epiphany/epiphany.c (gen_compare_reg):
15351 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
15352 is already in place.
15353 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
15354 Don't require being called during rtl expansion; If y operlaps r0,
15355 return 0.
15356 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
15357 (epiphany_expand_epilogue): Likewise.
15358
8afab237
JR
15359 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
15360 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 15361 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 15362
fa7707d6
JR
15363 * config/epiphany/constraints.md (CnL): New constraint.
15364 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
15365 * config/epiphany/predicates.md (add_operand): Allow 1024.
15366
5138e18d
JR
15367 * config/epiphany/epiphany.md (logical_op): New code iterator.
15368 (op_mnc): New code attribute.
15369 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
15370 (mov_f+1, mov_f+2): New peephole2 patterns.
15371
2ccc703d
JR
15372 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
15373 (cstoresi4): Also allow re-use of zero result when doing a NE
15374 comparison to a non-zero operand.
aefb0819 15375 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 15376
093ac0a5
JR
15377 * config/epiphany/epiphany.md (<insn_opname>v2si3):
15378 Use gen_addsi3_i / gen_subsi3_i.
15379
f223bb13
JJ
153802013-04-08 Jakub Jelinek <jakub@redhat.com>
15381
15382 PR c++/34949
15383 PR c++/50243
15384 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
15385 contain anything but clobbers, at most one __builtin_stack_restore,
15386 optionally debug stmts and final resx, and if it has at least one
15387 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
15388 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
15389 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
15390 which isn't defaut definition, remove them.
15391 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
15392 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
15393 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
15394 with MEM_REF LHS with SSA_NAME address.
15395
4481581f
JL
153962013-04-08 Jeff Law <law@redhat.com>
15397
15398 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 15399
451dabda
RB
154002013-04-08 Richard Biener <rguenther@suse.de>
15401
15402 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
15403 extra newline.
15404 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
15405 determined vector type.
15406 (vect_analyze_data_refs): Likewise.
15407 (vect_get_new_vect_var): Adjust.
15408 (vect_create_destination_var): Preserve SSA name versions.
15409 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
15410 not dump anything here.
15411
3b088b47
JR
154122013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
15413
15414 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
15415 Add member lr_slot_known.
15416 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
15417 if necessary.
15418 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15419 Remove code that sets lr_slot_offset according to what a previous
15420 version of epiphany_emit_save_restore used to do.
15421 (epiphany_emit_save_restore): When doing an lr save or restore,
15422 set/verify lr_slot_known and lr_slot_offset.
15423
d8484d41
XQ
154242013-04-08 Xinyu Qi <xyqi@marvell.com>
15425
33159866
UB
15426 PR target/54338
15427 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ
15428 in ALL_REGS.
15429
481be1c4
RB
154302013-04-08 Richard Biener <rguenther@suse.de>
15431
15432 * alias.c (find_base_term): Fix thinko in previous change.
15433
401f3a81
JJ
154342013-04-08 Jakub Jelinek <jakub@redhat.com>
15435
15436 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
15437 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
15438 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
15439 if possible to compute val.
15440 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
15441 For QImode integers don't require anything about precision. Use
15442 const_with_all_bytes_same to find out if the constant doesn't have
15443 repeated bytes in it.
15444
03ed99a8
AK
154452013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15446
15447 * config/s390/s390.c (s390_expand_insv): Only accept insertions
15448 within mode size.
15449
781b2e62
MP
154502013-04-08 Marek Polacek <polacek@redhat.com>
15451
15452 PR rtl-optimization/48182
15453 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
15454 value to 1.
15455
27e430a2
JDA
154562013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
15457
15458 PR target/55487
15459 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
15460 nuses, make sure we have a label.
15461
4902aa64
BS
154622013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15463
15464 PR target/56843
15465 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
15466 (rs6000_emit_swdiv_low_precision): Remove.
15467 (rs6000_emit_swdiv): Rewrite to handle between one and four
15468 iterations of Newton-Raphson generally; modify required number of
15469 iterations for some cases.
15470 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
15471
7bca81dc
SB
154722013-04-05 Steven Bosscher <steven@gcc.gnu.org>
15473
15474 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
15475 set-but-unused variable.
15476
15477 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
15478 basic blocks of released function bodies garbage-collectable.
15479
15480 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
15481 (struct rtl_opt_pass): Add TODO_df_finish.
15482
15483 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
15484
4542a38a
GY
154852013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15486
15487 * config/arm/constraints.md (q): New constraint.
15488 * config/arm/ldrdstrd.md: New file.
15489 * config/arm/arm.md (ldrdstrd.md) New include.
15490 (arm_movdi): Use "q" instead of "r" constraint
15491 for double-word memory access.
15492 (movdf_soft_insn): Likewise.
15493 * config/arm/vfp.md (movdi_vfp): Likewise.
15494 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 15495 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY
15496 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
15497 (mem_ok_for_ldrd_strd): Likewise.
15498 (output_move_double): Update assertion.
15499
2385b218
GY
155002013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15501
15502 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
15503
75fe1cb5
GY
155042013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15505
15506 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
15507 define_insn_and_split.
15508 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
15509
dd660e8e
GY
155102013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15511
15512 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
15513 define_insn_and_split.
33159866 15514 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY
15515 (shiftsi3_compare): New pattern.
15516 (rrx): New pattern.
15517 * config/arm/unspecs.md (UNSPEC_RRX): New.
15518
045e472c
GY
155192013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15520
15521 * config/arm/arm.md (negdi_extendsidi): New pattern.
15522 (negdi_zero_extendsidi): Likewise.
15523
3f3bf1a8
GY
155242013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15525
15526 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
15527 define_insn_and_split.
15528 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
15529 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
15530
b0b49556
GY
155312013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15532
15533 * config/arm/arm.md (arm_subdi3): Convert define_insn into
15534 define_insn_and_split.
15535 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
15536 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
15537
d633dd84
GY
155382013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15539
15540 * config/arm/arm.md (subsi3_carryin): New pattern.
15541 (subsi3_carryin_const): Likewise.
15542 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
15543 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
15544
f4499066
GY
155452013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15546
15547 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
15548
ceef6fd9
GY
155492013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
15550
15551 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 15552 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 15553
ddbdd8a7
KT
155542013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15555
15556 * config/arm/arm.c (arm_expand_builtin): Change fcode
15557 type to unsigned int.
15558
8456d78a
RR
155592013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15560
15561 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
15562
526c230f
ILT
155632013-04-04 Ian Lance Taylor <iant@google.com>
15564
33159866
UB
15565 * doc/standards.texi (Standards): The Go frontend supports the Go 1
15566 language standard.
526c230f 15567
3cfbe04d
SB
155682013-04-04 Steven Bosscher <steven@gcc.gnu.org>
15569
15570 PR middle-end/56729
15571 * df-scan.c (df_insn_delete): Disable failing assert.
15572
dfa3f8d0
KT
155732013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15574
15575 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
15576 New function prototype.
15577 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
15578 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
15579 (arm_builtin_vectorized_function): New function.
15580
f7837758
KT
155812013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15582
15583 * config/arm/arm_neon_builtins.def: New file.
15584 * config/arm/arm.c (neon_builtin_data): Move contents to
15585 arm_neon_builtins.def.
15586 (enum arm_builtins): Include neon builtin definitions.
15587 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 15588 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 15589
39c1b6db
MP
155902013-04-04 Marek Polacek <polacek@redhat.com>
15591
15592 PR tree-optimization/48186
15593 * predict.c (maybe_hot_frequency_p): Return false if
15594 HOT_BB_FREQUENCY_FRACTION is 0.
15595 (cgraph_maybe_hot_edge_p): Likewise.
15596
314f64eb
RB
155972013-04-04 Richard Biener <rguenther@suse.de>
15598
15599 PR tree-optimization/56826
15600 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
15601 more accurately.
15602
51a905b2
RB
156032013-04-04 Richard Biener <rguenther@suse.de>
15604
15605 PR tree-optimization/56213
15606 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 15607 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 15608
f1bf4f3a
RB
156092013-04-04 Richard Biener <rguenther@suse.de>
15610
15611 PR tree-optimization/56837
15612 * tree-loop-distribution.c (classify_partition): For non-zero
15613 values require that the value has the same precision as its
15614 mode to be useful as memset value.
15615
0bca7ded
NC
156162013-04-03 Nick Clifton <nickc@redhat.com>
15617
33159866 15618 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC
15619 (fmssf4): Use fmsf.s on E3V5 architectures.
15620 (fnmasf4): Use fnmaf.s on E3V5 architectures.
15621 (fnmssf4): Use fnmsf.s on E3V5 architectures.
15622
b4019227
JL
156232013-04-03 Jeff Law <law@redhat.com>
15624
15625 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
15626 (lra-eliminations.o): Likewise.
15627
f57ddb5b
TJ
156282013-04-03 Teresa Johnson <tejohnson@google.com>
15629
15630 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 15631 compute_working_sets here from profile.c.
f57ddb5b 15632 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB
15633 (gcov_working_set_t): Moved typedef here from basic-block.h
15634 (compute_working_set): Declare.
f57ddb5b
TJ
15635 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
15636 (get_working_sets): Renamed from compute_working_set,
33159866 15637 replace most of body with call to new compute_working_sets.
f57ddb5b 15638 (get_exec_counts): Replace call to compute_working_sets
33159866
UB
15639 to get_working_sets.
15640 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 15641 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 15642 to get_working_sets.
f57ddb5b
TJ
15643 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
15644 * gcov-dump.c (dump_working_sets): New function.
15645
12211b99 156462013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ
15647
15648 * hwint.c (sext_hwi, zext_hwi): New functions.
15649 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
15650 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
15651 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
15652 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
15653 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
15654 (sext_hwi, zext_hwi): New functions.
0bca7ded 15655
be672e08
JL
156562013-04-03 Jeff Law <law@redhat.com>
15657
15658 PR tree-optimization/56799
33159866
UB
15659 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
15660 back test for widening conversion erroneously dropped in prior change.
be672e08 15661
9d821fa5
KT
156622013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15663
15664 PR target/56809
15665 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
15666 instead of next_real_insn.
15667
71c581e7
MP
156682013-04-03 Marek Polacek <polacek@redhat.com>
15669
15670 PR sanitizer/55702
33159866 15671 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 15672
4a32ef80
KT
156732013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15674
15675 PR target/56809
15676 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
15677 next_real_insn.
15678 (thumb1_output_casesi): Likewise.
15679 (thumb2_output_casesi): Likewise.
15680
1b2253d4
RB
156812013-04-03 Richard Biener <rguenther@suse.de>
15682
15683 PR tree-optimization/56817
15684 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
15685 Split out ...
15686 (tree_unroll_loops_completely_1): ... new function to manually
15687 walk the loop tree, properly defering outer loops of unrolled
15688 loops to later iterations.
15689
38000232
MG
156902013-04-03 Marc Glisse <marc.glisse@inria.fr>
15691
15692 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
15693 (vectorizable_load): Likewise.
15694 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
15695 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
15696
3fa35298
MG
156972013-04-03 Marc Glisse <marc.glisse@inria.fr>
15698
15699 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
15700 BIT_FIELD_REF.
15701
b3d45ff0
UW
157022013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15703
15704 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
15705
ec9202a8
BC
157062013-04-03 Bin Cheng <bin.cheng@arm.com>
15707
15708 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
15709
6805bd36
MG
157102013-04-03 Marc Glisse <marc.glisse@inria.fr>
15711
15712 PR tree-optimization/56790
33159866
UB
15713 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
15714 folding.
6805bd36 15715
da694a77
MG
157162013-04-03 Marc Glisse <marc.glisse@inria.fr>
15717
15718 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
15719 Handle VEC_MERGE.
15720 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
15721 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
15722 equal arguments.
15723
4e7d7b3d
JJ
157242013-04-03 Jakub Jelinek <jakub@redhat.com>
15725
15726 PR c/19449
15727 * tree.h (force_folding_builtin_constant_p): New decl.
15728 * builtins.c (force_folding_builtin_constant_p): New variable.
15729 (fold_builtin_constant_p): Fold immediately also if
15730 force_folding_builtin_constant_p.
15731
e6c9d234
RB
157322013-04-03 Richard Biener <rguenther@suse.de>
15733
15734 PR tree-optimization/56812
15735 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
15736 DRs of the same interleaving chain are independent.
15737
984d07dd
JM
157382013-04-02 Jason Merrill <jason@redhat.com>
15739
15740 * gdbinit.in (pbb): Use debug fn.
15741
622849c9
LC
157422013-04-02 Lawrence Crowl <crowl@google.com>
15743
15744 * sese.h (struct ivtype_map_elt_s): Remove unused.
15745 (extern debug_ivtype_map): Remove unused.
15746 (extern eq_ivtype_map_elts): Remove unused.
15747 * sese.c (debug_ivtype_map): Removed unused.
15748 (debug_ivtype_map_1): Removed unused.
15749 (debug_ivtype_elt): Remove unused.
15750 (eq_ivtype_map_elts): Remove unused.
15751
15752
82c0e1a0
KT
157532013-04-02 Kai Tietz <ktietz@redhat.com>
15754
15755 PR target/52790
15756 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
15757 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB
15758 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
15759 function.
82c0e1a0
KT
15760 (legitimize_pe_coff_symbol): Likewise.
15761 (is_imported_p): New helper-function.
15762 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
15763 for Windows x64 targets.
15764 (ix86_expand_prologue): Optimize for pe-coff targets.
15765 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
15766 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
15767 medium/large code-model.
15768 (legitimize_pic_address): Likewise.
15769 (legitimize_tls_address): Likewise.
15770 (ix86_expand_call): Likewise.
15771 (x86_output_mi_thunk): Likewise.
15772 (get_dllimport_decl): Add new beimport argument.
15773 (construct_plt_address): Don't assert for x64 pe-coff targets.
15774 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
15775 targets.
15776 (SYMBOL_FLAG_STUBVAR): New macro.
15777 (SYMBOL_REF_STUBVAR_P): Likewise.
15778 * config/i386/winnt.c (stub_list): New structure.
15779 (stub_head): New local variable.
15780 (i386_pe_record_stub): New function.
15781 (i386_pe_file_end): Emit refptr-stubs.
15782
5d751b0c
JJ
157832013-04-02 Jakub Jelinek <jakub@redhat.com>
15784
09bb4c99
JJ
15785 PR rtl-optimization/56745
15786 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
15787 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
15788
a1d8947a
JJ
15789 PR c++/34949
15790 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
15791 and both of them are MEM_REFs, just compare first argument for
15792 equality and attempt to deal even with differing offsets.
15793
5d751b0c
JJ
15794 PR c++/34949
15795 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
15796 of gimple_clobber_p to be MEM_REF.
15797 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
15798 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
15799 after gimplification.
15800 * asan.c (get_mem_ref_of_assignment): Don't instrument
15801 gimple_clobber_p stmts.
15802 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
15803 gimple_clobber_p stmt if they have MEM_REF lhs and
15804 are dead because of another gimple_clobber_p stmt.
15805 * tree-ssa-live.c (clear_unused_block_pointer): Treat
15806 gimple_clobber_p stmts like debug stmts.
15807 (remove_unused_locals): Remove clobbers with MEM_REF lhs
15808 that refer to unused VAR_DECLs or uninitialized values.
15809 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
15810 gimple_clobber_p stmts if they refer to removed parameters.
15811 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
15812 formatting.
15813
e594716a
UB
158142013-04-02 Uros Bizjak <ubizjak@gmail.com>
15815
15816 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
15817 using SWI48 mode attribute.
15818
7a80735b
WM
158192013-04-02 Wei Mi <wmi@google.com>
15820
15821 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
15822 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
15823 *<rotate_insn><mode>3_mask in i386.md.
15824
f423a9e4
AI
158252013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
15826
15827 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
15828
90eb75f2
RB
158292013-04-02 Richard Biener <rguenther@suse.de>
15830
15831 PR tree-optimization/56778
15832 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
15833 Runtime alias tests are not supported for gather loads.
15834 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
15835 stmts referenced from SSA operands before updating SSA form.
15836
d8c69a92
IC
158372013-04-02 Ian Caulfield <ian.caulfield@arm.com>
15838 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15839
15840 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB
15841 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
15842 * config/arm/cortex-a53.md: New file.
15843 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
15844 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
15845 * config/arm/arm.c (arm_issue_rate): Likewise.
15846 * config/arm/arm-tune.md: Regenerate
15847 * config/arm/arm-tables.opt: Regenerate.
15848 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 15849
239eb04c
ZC
158502013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
15851
15852 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
15853 non-static link.
15854
c902d3c8
SN
158552013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15856
d8c69a92
IC
15857 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
15858 scalar load/store operations using B/H registers.
c902d3c8
SN
15859 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
15860
051d0e2f
SN
158612013-04-02 Sofiane Naci <sofiane.naci@arm.com>
15862
15863 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
15864 scalar move.
15865 * config/aarch64/aarch64.c
15866 (aarch64_simd_scalar_immediate_valid_for_move): New.
15867 * config/aarch64/aarch64-protos.h
15868 (aarch64_simd_scalar_immediate_valid_for_move): New.
15869 * config/aarch64/constraints.md (Dh, Dq): New.
15870 * config/aarch64/iterators.md (hq): New.
15871
0ee1e3d9
EB
158722013-04-02 Eric Botcazou <ebotcazou@adacore.com>
15873
15874 * reorg.c (get_branch_condition): Deal with conditional returns.
15875 (fill_simple_delay_slots): Remove dead code dealing with jumps.
15876
136fb3f7
WM
158772013-04-01 Wei Mi <wmi@google.com>
15878
15879 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
15880 Truncate operand 2 using %b asm operand modifier.
15881 (*<shift_insn><mode>3_mask): Ditto.
15882 (*<rotate_insn><mode>3_mask): Ditto.
15883
6388c738
SB
158842013-04-01 Steven Bosscher <steven@gcc.gnu.org>
15885
15886 PR middle-end/56798
15887 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
15888
f7a4d826
KK
158892013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
15890
15891 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
15892 of next_real_insn.
15893 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
15894
4b943a49
LC
158952013-03-30 Lawrence Crowl <crowl@google.com>
15896
15897 * dse.c (clear_alias_sets): Remove never set.
15898 (disqualified_clear_alias_sets): Remove never set.
15899 (clear_alias_mode_pool): Remove never set.
15900 (dse_step0): Remove condition that is never true.
15901 (canon_address): Remove condition that is never true.
15902 (dse_step7): Remove condition that is never true.
15903 (rest_of_handle_dse): Remove condition that is never true.
15904 (rest_of_handle_dse::did_global): Remove never read from above.
15905 (dse_step2_spill): Remove never called from above.
15906 (dse_step5_spill): Remove never called from above.
15907
39718607
SB
159082013-03-30 Steven Bosscher <steven@gcc.gnu.org>
15909
da5c6bde
SB
15910 * doc/md.texi (Standard Names) <casesi>: Update documentation for
15911 JUMP_TABLE_DATA changes.
15912 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
15913 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
15914 (Insns) <jump_table_data>: New entry.
15915 * doc/tm.texi: Regenerate.
15916
39718607
SB
15917 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
15918
15919 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
15920 for table jump at the end of a basic block using tablejump_p.
15921 * targhooks.c (default_invalid_within_doloop): Likewise.
15922 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
15923 target hook implementation that is identical to the default hook.
15924 (rs6000_invalid_within_doloop): Remove.
15925
15926 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
15927 unused variable from tablejump_p call.
15928
15929 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
15930 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
15931 (INSN_DELETED_P): Likewise.
15932 (emit_jump_table_data): New prototype.
15933 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
15934 after 4th as unused.
15935 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
15936 * sched-vis.c (print_insn): Likewise.
15937 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
15938 insn for compatibility with back ends that use next_active_insn to
15939 identify jump table data.
15940 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
15941 (remove_insn): Likewise.
15942 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
15943 to be emitted.
15944 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
15945 (emit_jump_table_data): New function.
15946
15947 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
15948 basic block, a JUMP_TABLE_DATA never is.
15949 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
15950 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
15951 off from code handling real insns.
15952 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
15953 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
15954 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
15955 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
15956 is not a NONDEBUG_INSN_P.
15957 * ira-costs.c (scan_one_insn): Likewise.
15958 * jump.c (mark_all_labels): Likewise.
15959 (mark_jump_label_1): Likewise.
15960 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
15961 * lra.c (get_insn_freq): Expect all insns reaching here to be in
15962 a basic block.
15963 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
15964 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
15965 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
15966 JUMP_TABLE_DATA_P insns.
15967 (calculate_elim_costs_all_insns): Likewise.
15968 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
15969 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
15970 (delete_output_reload): Code style fixups.
15971 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
15972 insn flags on this non-insn.
15973 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
15974 as scheduling barriers, for pre-change compatibility.
15975 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
15976 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
15977
15978 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
15979 redundant JUMP_TABLE_DATA_P test.
15980 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
15981 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
15982 (frv_for_each_packet): Likewise.
15983 * config/i386/i386.c (min_insn_size): Likewise.
15984 (ix86_avoid_jump_mispredicts): Likewise.
15985 * config/m32r/m32r.c (m32r_is_insn): Likewise.
15986 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
15987 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
15988 (mips16_insn_length): Robustify.
15989 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
15990 (mips16_split_long_branches): Likewise.
15991 * config/pa/pa.c (pa_combine_instructions): Likewise.
15992 * config/rs6000/rs6000.c (get_next_active_insn): Treat
15993 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
15994 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
15995 as contributing to pool range lengths.
15996 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
15997 Remove redundant JUMP_TABLE_DATA_P test.
15998 (sh_loop_align): Likewise.
15999 (split_branches): Likewise.
16000 (sh_insn_length_adjustment): Likewise.
16001 * config/spu/spu.c (get_branch_target): Likewise.
16002
0208f7da
JH
160032013-03-29 Jan Hubicka <jh@suse.cz>
16004
16005 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
16006 gcov streaming; stream hot bb threshold to ltrans.
16007 * predict.c (get_hot_bb_threshold): Break out from ....
16008 (maybe_hot_count_p): ... here.
16009 (set_hot_bb_threshold): New function.
16010 * lto-section-in.c (lto_section_name): Add profile.
16011 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
16012 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
16013 and data-streamer.h
16014 (histogram_entry): New structure.
16015 (histogram, histogram_pool): New global vars.
16016 (histogram_hash): New structure.
16017 (histogram_hash::hash): New method.
16018 (histogram_hash::equal): Likewise.
16019 (account_time_size): New function.
16020 (cmp_counts): New function.
16021 (dump_histogram): New function.
16022 (ipa_profile_generate_summary): New function.
16023 (ipa_profile_write_summary): New function.
16024 (ipa_profile_read_summary): New function.
16025 (ipa_profile): Decide on threshold.
e594716a
UB
16026 (pass_ipa_profile): Add ipa_profile_write_summary and
16027 ipa_profile_read_summary.
0208f7da
JH
16028 * Makefile.in (ipa.o): Update dependencies.
16029 * lto-streamer.h (LTO_section_ipa_profile): New section.
16030
5a6ccc94
GDR
160312013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
16032
16033 * tree.h (VAR_P): New.
16034
39385fa6
PC
160352013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
16036
16037 PR lto/56777
16038 * doc/invoke.texi ([-fwhole-program]): Fix typo.
16039
34f0d87a
SB
160402013-03-29 Steven Bosscher <steven@gcc.gnu.org>
16041
16042 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
16043 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16044 (control_flow_insn_p): Likewise.
16045 * cfgrtl.c (duplicate_insn_chain): Likewise.
16046 * final.c (get_attr_length_1): Likewise.
16047 (shorten_branches): Likewise.
16048 (final_scan_insn): Likewise.
16049 * function.c (instantiate_virtual_regs): Likewise.
16050 * gcse.c (insert_insn_end_basic_block): Likewise.
16051 * ira-costs.c (scan_one_insn): Likewise.
16052 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
16053 * lra.c (check_rtl): Likewise.
16054 * reload1.c (elimination_costs_in_insn): Likewise.
16055 * reorg.c (follow_jumps): Likewise.
16056
16057 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
16058 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
16059 (thumb_far_jump_used_p): Likewise.
16060 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
16061 (workaround_speculation): Likewise.
16062 (add_sched_insns_for_speculation): Likewise.
16063 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
16064 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
16065 (frv_for_each_packet): Likewise.
16066 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
16067 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
16068 (final_emit_insn_group_barriers): Likewise.
16069 * config/m32r/m32r.c (m32r_is_insn): Likewise.
16070 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
16071 (mips16_insn_length): Likewise.
16072 * config/pa/pa.c (pa_reorg): Likewise.
16073 (pa_combine_instructions): Likewise.
16074 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
16075 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
16076 (sh_reorg): Likewise.
16077 (split_branches): Likewise.
16078 * config/spu/spu.c (get_branch_target): Likewise.
16079
16080 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
16081 JUMP_TABLE_DATA_P.
16082
4ac761b0
KY
160832013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
16084
39385fa6 16085 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY
16086 Fix declaration name.
16087
58a49189
LC
160882013-03-28 Lawrence Crowl <crowl@google.com>
16089
16090 * graphds.h (struct graph.indicies): Remove unused.
16091 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
16092 (SCOP_ORIGINAL_PDDRS): Remove unused.
16093 * sese.h (extern insert_loop_close_phis): Removed unused.
16094 (extern insert_guard_phis): Removed unused.
16095 (extern ivtype_map_elt_info): Removed unused.
16096 (new_ivtype_map_elt): Removed unused.
16097 * sese.c (ivtype_map_elt_info): Removed unused.
16098
7b3b6ae4
LC
160992013-03-28 Lawrence Crowl <crowl@google.com>
16100
16101 * Makefile.in: Add several missing include dependences.
16102 (DUMPFILE_H): New.
16103 (test-dump.o): New. This object is not added to any executable,
16104 but is present for ad-hoc testing.
16105 * bitmap.c
16106 (debug (const bitmap_head_def &)): New.
16107 (debug (const bitmap_head_def *)): New.
16108 * bitmap.h
16109 (extern debug (const bitmap_head_def &)): New.
16110 (extern debug (const bitmap_head_def *)): New.
16111 * cfg.c
16112 (debug (edge_def &)): New.
16113 (debug (edge_def *)): New.
16114 * cfghooks.c
16115 (debug (basic_block_def &)): New.
16116 (debug (basic_block_def *)): New.
16117 * dumpfile.h
16118 (dump_node (const_tree, int, FILE *)): Correct source file.
16119 * dwarf2out.c
16120 (debug (die_struct &)): New.
16121 (debug (die_struct *)): New.
16122 * dwarf2out.h
16123 (extern debug (die_struct &)): New.
16124 (extern debug (die_struct *)): New.
16125 * gimple-pretty-print.c
16126 (debug (gimple_statement_d &)): New.
16127 (debug (gimple_statement_d *)): New.
16128 * gimple-pretty-print.h
16129 (extern debug (gimple_statement_d &)): New.
16130 (extern debug (gimple_statement_d *)): New.
16131 * ira-build.c
16132 (debug (ira_allocno_copy &)): New.
16133 (debug (ira_allocno_copy *)): New.
16134 (debug (ira_allocno &)): New.
16135 (debug (ira_allocno *)): New.
16136 * ira-int.h
16137 (extern debug (ira_allocno_copy &)): New.
16138 (extern debug (ira_allocno_copy *)): New.
16139 (extern debug (ira_allocno &)): New.
16140 (extern debug (ira_allocno *)): New.
16141 * ira-lives.c
16142 (debug (live_range &)): New.
16143 (debug (live_range *)): New.
16144 * lra-int.h
16145 (debug (lra_live_range &)): New.
16146 (debug (lra_live_range *)): New.
16147 * lra-lives.c
16148 (debug (lra_live_range &)): New.
16149 (debug (lra_live_range *)): New.
16150 * omega.c
16151 (debug (omega_pb_d &)): New.
16152 (debug (omega_pb_d *)): New.
16153 * omega.h
16154 (extern debug (omega_pb_d &)): New.
16155 (extern debug (omega_pb_d *)): New.
16156 * print-rtl.c
16157 (debug (const rtx_def &)): New.
16158 (debug (const rtx_def *)): New.
16159 * print-tree.c
16160 (debug_tree (tree): Move within file.
16161 (debug_raw (const tree_node &)): New.
16162 (debug_raw (const tree_node *)): New.
16163 (dump_tree_via_hooks (const tree_node *, int)): New.
16164 (debug (const tree_node &)): New.
16165 (debug (const tree_node *)): New.
16166 (debug_verbose (const tree_node &)): New.
16167 (debug_verbose (const tree_node *)): New.
16168 (debug_head (const tree_node &)): New.
16169 (debug_head (const tree_node *)): New.
16170 (debug_body (const tree_node &)): New.
16171 (debug_body (const tree_node *)): New.
16172 (debug_vec_tree (tree): Move and reimplement in terms of dump.
16173 (debug (vec<tree, va_gc> &)): New.
16174 (debug (vec<tree, va_gc> *)): New.
16175 * rtl.h
16176 (extern debug (const rtx_def &)): New.
16177 (extern debug (const rtx_def *)): New.
16178 * sbitmap.c
16179 (debug_raw (simple_bitmap_def &)): New.
16180 (debug_raw (simple_bitmap_def *)): New.
16181 (debug (simple_bitmap_def &)): New.
16182 (debug (simple_bitmap_def *)): New.
16183 * sbitmap.h
16184 (extern debug (simple_bitmap_def &)): New.
16185 (extern debug (simple_bitmap_def *)): New.
16186 (extern debug_raw (simple_bitmap_def &)): New.
16187 (extern debug_raw (simple_bitmap_def *)): New.
16188 * sel-sched-dump.c
16189 (debug (vinsn_def &)): New.
16190 (debug (vinsn_def *)): New.
16191 (debug_verbose (vinsn_def &)): New.
16192 (debug_verbose (vinsn_def *)): New.
16193 (debug (expr_def &)): New.
16194 (debug (expr_def *)): New.
16195 (debug_verbose (expr_def &)): New.
16196 (debug_verbose (expr_def *)): New.
16197 (debug (vec<rtx> &)): New.
16198 (debug (vec<rtx> *)): New.
16199 * sel-sched-dump.h
16200 (extern debug (vinsn_def &)): New.
16201 (extern debug (vinsn_def *)): New.
16202 (extern debug_verbose (vinsn_def &)): New.
16203 (extern debug_verbose (vinsn_def *)): New.
16204 (extern debug (expr_def &)): New.
16205 (extern debug (expr_def *)): New.
16206 (extern debug_verbose (expr_def &)): New.
16207 (extern debug_verbose (expr_def *)): New.
16208 (extern debug (vec<rtx> &)): New.
16209 (extern debug (vec<rtx> *)): New.
16210 * sel-sched-ir.h
16211 (_list_iter_cond_expr): Make inline instead of static.
16212 * sreal.c
16213 (debug (sreal &)): New.
16214 (debug (sreal *)): New.
16215 * sreal.h
16216 (extern debug (sreal &)): New.
16217 (extern debug (sreal *)): New.
16218 * tree.h
16219 (extern debug_raw (const tree_node &)): New.
16220 (extern debug_raw (const tree_node *)): New.
16221 (extern debug (const tree_node &)): New.
16222 (extern debug (const tree_node *)): New.
16223 (extern debug_verbose (const tree_node &)): New.
16224 (extern debug_verbose (const tree_node *)): New.
16225 (extern debug_head (const tree_node &)): New.
16226 (extern debug_head (const tree_node *)): New.
16227 (extern debug_body (const tree_node &)): New.
16228 (extern debug_body (const tree_node *)): New.
16229 (extern debug (vec<tree, va_gc> &)): New.
16230 (extern debug (vec<tree, va_gc> *)): New.
16231 * tree-cfg.c
16232 (debug (struct loop &)): New.
16233 (debug (struct loop *)): New.
16234 (debug_verbose (struct loop &)): New.
16235 (debug_verbose (struct loop *)): New.
16236 * tree-dump.c: Add header dependence.
16237 * tree-flow.h
16238 (extern debug (struct loop &)): New.
16239 (extern debug (struct loop *)): New.
16240 (extern debug_verbose (struct loop &)): New.
16241 (extern debug_verbose (struct loop *)): New.
16242 * tree-data-ref.c
16243 (debug (data_reference &)): New.
16244 (debug (data_reference *)): New.
16245 (debug (vec<data_reference_p> &)): New.
16246 (debug (vec<data_reference_p> *)): New.
16247 (debug (vec<ddr_p> &)): New.
16248 (debug (vec<ddr_p> *)): New.
16249 * tree-data-ref.h
16250 (extern debug (data_reference &)): New.
16251 (extern debug (data_reference *)): New.
16252 (extern debug (vec<data_reference_p> &)): New.
16253 (extern debug (vec<data_reference_p> *)): New.
16254 (extern debug (vec<ddr_p> &)): New.
16255 (extern debug (vec<ddr_p> *)): New.
16256 * tree-ssa-alias.c
16257 (debug (pt_solution &)): New.
16258 (debug (pt_solution *)): New.
16259 * tree-ssa-alias.h
16260 (extern debug (pt_solution &)): New.
16261 (extern debug (pt_solution *)): New.
16262 * tree-ssa-alias.c
16263 (debug (_var_map &)): New.
16264 (debug (_var_map *)): New.
16265 (debug (tree_live_info_d &)): New.
16266 (debug (tree_live_info_d *)): New.
16267 * tree-ssa-alias.h
16268 (extern debug (_var_map &)): New.
16269 (extern debug (_var_map *)): New.
16270 (extern debug (tree_live_info_d &)): New.
16271 (extern debug (tree_live_info_d *)): New.
16272
be77e1e5
JH
162732013-03-28 Jan Hubicka <jh@suse.cz>
16274
16275 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
16276
777e6976
IB
162772013-03-28 Ian Bolton <ian.bolton@arm.com>
16278
16279 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
16280 record only when desired or required.
16281
9cd347ae
UB
162822013-03-28 Uros Bizjak <ubizjak@gmail.com>
16283
16284 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
16285 *vec_extractv2di_1_rex64. Use x64 isa attribute.
16286
33159866 162872013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N
16288
16289 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
16290 (*andsi3_compare0_uxtw): New pattern.
16291 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
16292 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
16293
89ab31c1
JH
162942013-03-28 Jan Hubicka <jh@suse.cz>
16295
16296 * data-streamer-in.c (streamer_read_gcov_count): New function.
16297 * gimple-streamer-out.c: Include value-prof.h.
16298 (output_gimple_stmt): Output histogram.
16299 (output_bb): Use streamer_write_gcov_count.
16300 * value-prof.c: Include data-streamer.h
16301 (dump_histogram_value): Add HIST_TYPE_MAX.
16302 (stream_out_histogram_value): New function.
16303 (stream_in_histogram_value): New function.
16304 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
16305 (stream_out_histogram_value, stream_in_histogram_value): Declare.
16306 * data-streamer-out.c (streamer_write_gcov_count): New function.
16307 (streamer_write_gcov_count_stream): New function.
16308 * lto-cgraph.c (lto_output_edge): Update counter streaming.
16309 (lto_output_node): Likewise.
16310 (input_node, input_edge): Likewise.
16311 * lto-streamer-out.c (output_cfg): Update streaming.
16312 * lto-streamer-in.c (input_cfg): Likewise.
16313 * data-streamer.h (streamer_write_gcov_count,
16314 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
16315 * gimple-streamer-in.c: Include value-prof.h
16316 (input_gimple_stmt): Input histograms.
16317 (input_bb): Update profile streaming.
16318
e594716a 163192013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 16320
33159866 16321 * genmodes.c (emit_max_int): New function.
8697be17 16322 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC
16323 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
16324 Added doc.
8697be17 16325 * machmode.def: Fixed comment.
89ab31c1 16326
e594716a 163272013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ
16328
16329 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ
16330 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
16331
74946978
MP
163322013-03-28 Marek Polacek <polacek@redhat.com>
16333 Richard Biener <rguenther@suse.de>
16334
16335 PR tree-optimization/56695
16336 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
16337 build signed result of a vector comparison.
16338 * tree-cfg.c (verify_gimple_comparison): Check that a result
16339 of a vector comparison has signed type.
16340
a64b9c26
RB
163412013-03-28 Richard Biener <rguenther@suse.de>
16342
16343 PR tree-optimization/37021
16344 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
16345 do not restrict gaps between groups.
16346 * tree-vect-stmts.c (vectorizable_load): Properly account for
16347 a gap between groups.
16348
a9dc2a2f
EB
163492013-03-28 Eric Botcazou <ebotcazou@adacore.com>
16350
16351 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
16352 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
16353 is not enabled.
16354
53cb97f9
GP
163552013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
16356
16357 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
16358 * doc/extend.texi (Named Address Spaces): Ditto.
16359 (Variable Attributes): Ditto.
16360
b802ae5c
KT
163612013-03-27 Kai Tietz <ktietz@redhat.com>
16362
eddae10a
KT
16363 * config.build: Add support for cygwin x64 target.
16364 * config.gcc: Likewise.
16365 * config.host: Likewise.
16366 * configure.ac: Likewise
16367 * configure: Regenerated.
b802ae5c 16368
371e77e3 163692013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT
16370
16371 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
16372 * config/i386/t-cygwin-w64: New file.
16373 * config/i386/cygwin-w64.h: New file.
16374 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
16375 and add support for x64-cygwin target.
16376 (CPP_SPEC): Likewise.
16377 (CXX_WRAP_SPEC_LIST): Undefine before define.
16378 (LIBGCJ_SONAME): Use 15 as version.
16379
f49b33cb
RB
163802013-03-27 Richard Biener <rguenther@suse.de>
16381
16382 PR tree-optimization/56716
16383 * tree-ssa-structalias.c (perform_var_substitution): Adjust
16384 dumping for ref nodes.
16385
b37a6ce5
MJ
163862013-03-27 Martin Jambor <mjambor@suse.cz>
16387
16388 PR tree-optimization/55334
16389 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
16390 restricted pointers to arrays.
16391
9469b9b2
GDR
163922013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
16393
16394 * Makefile.in (.SUFFIXES): Add .cc.
16395 (.c.o): Apply same recipe for implicit rule .cc.o.
16396
7d24f650
RB
163972013-03-27 Richard Biener <rguenther@suse.de>
16398
16399 PR tree-optimization/37021
16400 * tree-vect-data-refs.c (vect_check_strided_load): Allow
16401 REALPART/IMAGPART_EXPRs around the supported refs.
16402 * tree-ssa-structalias.c (find_func_aliases): Assume that
16403 floating-point values are not used to transfer pointers.
16404
2f251a05
AI
164052013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
16406
f43245d1
UB
16407 * target.def (TARGET_HAS_IFUNC_P): New target hook.
16408 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
16409 * doc/tm.texi: Regenerate.
16410 * targhooks.h (default_has_ifunc_p): New.
16411 * targhooks.c (default_has_ifunc_p): Ditto.
16412 * config/linux-protos.h: New file.
16413 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
16414 hook for linux which disables support of indirect functions in android.
16415 * config/linux-android.c: New file.
16416 * config/t-linux-android.c: Ditto.
16417 * config.gcc: Added new object file linux-android.o.
16418 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 16419 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 16420 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI
16421 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
16422 doesn't support indirect functions.
16423 * configure: Regenerate.
16424
78b4e425
BC
164252013-03-27 Bin Cheng <bin.cheng@arm.com>
16426
16427 PR target/56102
16428 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
16429 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
16430 mult-word mode.
16431
a5ba7b92
AK
164322013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16433
16434 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
16435
be624986
TG
164362013-03-27 Terry Guo <terry.guo@arm.com>
16437
16438 * config/arm/arm-cores.def: Added core cortex-r7.
16439 * config/arm/arm-tune.md: Regenerated.
16440 * config/arm/arm-tables.opt: Regenerated.
16441 * doc/invoke.texi: Added entry for core cortex-r7.
16442
0a514f47
WL
164432013-03-27 Walter Lee <walt@tilera.com>
16444
16445 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
16446 double-decrement of next_scratch_regno.
16447
9332b0d2
WL
164482013-03-27 Walter Lee <walt@tilera.com>
16449
801d9b2a 16450 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL
16451 input operands.
16452 (insn_v1mulus): Ditto.
16453 (insn_v2muls): Ditto.
16454
f54ea5dd
WL
164552013-03-27 Walter Lee <walt@tilera.com>
16456
e594716a 16457 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL
16458 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
16459
bf60f4ca
WL
164602013-03-27 Walter Lee <walt@tilera.com>
16461
e594716a 16462 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL
16463 (*sibcall_value): Ditto.
16464
e3b51eeb
WL
164652013-03-27 Walter Lee <walt@tilera.com>
16466
16467 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
16468 (insn_mnz_v8qi): ... this ...
16469 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
16470 vector equivalent.
16471 (insn_v<n>mnz): Replaced by ...
16472 (insn_v1mnz): ... this ...
16473 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
16474 equivalent.
16475 (insn_mz_<mode>): Replaced by ...
16476 (insn_mz_v8qi): ... this ...
16477 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
16478 vector equivalent.
16479 (insn_v<n>mz): Replaced by ...
16480 (insn_v1mz): ... this ...
16481 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
16482 equivalent.
16483
4fc7b145
EB
164842013-03-26 Eric Botcazou <ebotcazou@adacore.com>
16485
16486 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
16487
6f33817e
RM
164882013-03-26 Roland McGrath <mcgrathr@google.com>
16489
16490 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
16491 than fprintf with a non-constant, non-format string.
16492
93a4145b
UB
164932013-03-26 Uros Bizjak <ubizjak@gmail.com>
16494
16495 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
16496 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
16497 operand 0 predicate.
16498 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
16499 attribute. Use general_x64nomem_operand as operand 1 predicate.
16500 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
16501 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16502 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
16503 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
16504 (mov<mode>_insv_1): Remove expander. Merge insn with
16505 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
16506 Use general_x64nomem_operand as operand 1 predicate.
16507 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
16508 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
16509 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16510 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
16511 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
16512 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
16513 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
16514 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
16515 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
16516 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
16517 (general_x64nomem_operand): Ditto.
16518
c6a9ed5a 165192013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH
16520
16521 * config/rtems.opt: Add -pthread option.
16522
31b0a960
RB
165232013-03-26 Richard Biener <rguenther@suse.de>
16524
93a4145b
UB
16525 * alias.c (find_base_term): Avoid redundant and not used recursion.
16526 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB
16527 (true_dependence_1): Compute and pass base terms to base_alias_check.
16528 (write_dependence_p): Likewise.
16529 (may_alias_p): Likewise.
16530
79517551
SN
165312013-03-26 Sofiane Naci <sofiane.naci@arm.com>
16532
16533 * config/aarch64/aarch64.c (aarch64_classify_address): Support
16534 PC-relative load in SI modes and above only.
16535
a76213b9
XQ
165362013-03-26 Xinyu Qi <xyqi@marvell.com>
16537
16538 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB
16539 * config/arm/iwmmxt.md (WCGR0): Update.
16540 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 16541
37ff9355
UB
165422013-03-26 Uros Bizjak <ubizjak@gmail.com>
16543
16544 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
16545 Use x64 and nox64 isa attributes.
16546
30b0317c
RB
165472013-03-26 Richard Biener <rguenther@suse.de>
16548
16549 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
16550 alignment computations and rely on get_object_alignment_1
16551 for the !TYPE_P case.
16552 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
16553
cbcdb140
WL
165542013-03-26 Walter Lee <walt@tilera.com>
16555
16556 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
16557 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
16558
dd884525 165592013-03-25 Jeff Law <law@redhat.com>
be672e08 16560
33159866
UB
16561 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
16562 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL
16563 wrong version of prior patch.
16564
0051d3ec
WL
165652013-03-25 Walter Lee <walt@tilera.com>
16566
16567 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
16568 TILEGX_INSN_SHUFFLEBYTES1.
16569 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
16570 shufflebytes1.
16571 (tilegx_builtins): Ditto.
16572 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
16573
065a3605
WL
165742013-03-25 Walter Lee <walt@tilera.com>
16575
16576 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL
16577 (floatunssisf2): New pattern.
16578 (floatsidf2): New pattern.
16579 (floatunssidf2): New pattern.
065a3605 16580
5b2a3c82
WL
165812013-03-25 Walter Lee <walt@tilera.com>
16582
16583 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
16584 tests for constraint J, K, N, P.
16585
192ea533
WL
165862013-03-25 Walter Lee <walt@tilera.com>
16587
16588 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
16589 Use indirect/pcrel encoding.
16590 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
16591 Ditto.
16592
855e0d0b
SE
165932013-03-25 Steve Ellcey <sellcey@mips.com>
16594
16595 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
16596 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
16597 * config/mips/mips.c (mips_option_override): Set IMADD default.
16598 * config/mips/mips.h (PTF_AVOID_IMADD): New.
16599 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
16600 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
16601 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
16602 * doc/invoke.texi (-mimadd/-mno-imadd): New.
16603
39e45653
JL
166042013-03-25 Jeff Law <law@redhat.com>
16605
16606 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
16607 slightly to avoid creating and folding useless trees. Simplify
16608 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
16609
7482c470
UB
166102013-03-25 Uros Bizjak <ubizjak@gmail.com>
16611
16612 * config/i386/i386.md (*zero_extendsidi2): Merge with
16613 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
16614 * config/i386/predicates.md (x86_64_zext_operand): Rename from
16615 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
16616 targets. Clarify comment.
16617
4a53743e
MJ
166182013-03-25 Martin Jambor <mjambor@suse.cz>
16619
16620 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
16621 pass-through jump functions differently.
16622 (ipa_read_jump_function): Likewise. Also use setter functions to set
16623 up jump functions.
16624
162712de
MJ
166252013-03-25 Martin Jambor <mjambor@suse.cz>
16626
16627 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
16628 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
16629 process it.
16630 (ipa_get_indirect_edge_target): New function.
16631 (devirtualization_time_bonus): New parameter known_aggs, pass it to
16632 ipa_get_indirect_edge_target. Update all callers.
16633 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
16634 ipa_get_indirect_edge_target_1 instead of calling
16635 ipa_get_indirect_edge_target.
16636 (create_specialized_node): Pass aggvlas to
16637 ipcp_discover_new_direct_edges.
16638
a5a4c20a
KT
166392013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16640
16641 * config/arm/arm.md (f_sels, f_seld): New types.
16642 (*cmov<mode>): New pattern.
16643 * config/arm/predicates.md (arm_vsel_comparison_operator): New
16644 predicate.
16645
f992b9fc
KT
166462013-03-25 Kai Tietz <ktietz@redhat.com>
16647
16648 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
16649 POSIX-printf for mingw-hosted builds.
16650
c16fd676
RB
166512013-03-25 Richard Biener <rguenther@suse.de>
16652
16653 PR middle-end/56694
16654 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
16655 must-not-throw stmt location.
16656
5ad29f12
KT
166572013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16658
16659 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
16660 Emit load-acquire versions when acq is true.
16661 (arm_emit_store_exclusive): Add rel parameter.
16662 Emit store-release versions when rel is true.
16663 (arm_split_compare_and_swap): Use acquire-release instructions
16664 instead.
16665 of barriers when appropriate.
16666 (arm_split_atomic_op): Likewise.
16667 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
16668 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
16669 (VUNSPEC_SLX): Likewise.
16670 (VUNSPEC_LDA): Likewise.
16671 (VUNSPEC_STL): Likewise.
16672 * config/arm/sync.md (atomic_load<mode>): New pattern.
16673 (atomic_store<mode>): Likewise.
16674 (arm_load_acquire_exclusive<mode>): Likewise.
16675 (arm_load_acquire_exclusivesi): Likewise.
16676 (arm_load_acquire_exclusivedi): Likewise.
16677 (arm_store_release_exclusive<mode>): Likewise.
16678
03a7dddb
CM
166792013-03-25 Catherine Moore <clm@codesourcery.com>
16680
16681 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
16682 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
16683 * config/mip/predicates.md (lwsp_swsp_operand,
16684 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
16685 sb16_operand, db4_operand, db7_operand, ib3_operand,
16686 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
16687 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
16688 andi16_operand): New predicates.
16689 * config/mips/mips.md (compression): New attribute.
16690 (enabled): New attribute.
16691 (length): Consider compression in computing length.
16692 (shift_compression): New code attribute.
16693 (*add<mode>3): New operands. Record compression.
16694 (sub<mode>3): Likewise.
16695 (one_cmpl<mode>2): Likewise.
16696 (*and<mode>3): Likewise.
16697 (*ior<mode>3): Likewise.
16698 (unnamed pattern for xor): Likewise.
16699 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
16700 (*<optab><mode>3): Likewise.
16701 (*mov<mode>_internal: Likewise.
16702 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
16703 (mips_unsigned_immediate_p): New.
16704 (umips_lwsp_swsp_address_p): New.
16705 (m16_based_address_p): New.
16706 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
16707 (mips_unsigned_immediate_p): New prototype.
16708 (lwsp_swsp_address_p): New prototype.
16709 (m16_based_address_p): New prototype.
16710 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
16711 (mips_signed_immediate_p): New function.
16712 (m16_based_address_p): New function.
16713 (lwsp_swsp_address_p): New function.
16714 (mips_print_operand_punctuation): Recognize short delay slot insns
16715 for microMIPS.add<mode>3"
16716
f35c297f 167172013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 16718
f35c297f
KT
16719 PR target/56720
16720 * config/arm/iterators.md (v_cmp_result): New mode attribute.
16721 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
16722
051b9446
RB
167232013-03-25 Richard Biener <rguenther@suse.de>
16724
16725 PR tree-optimization/56689
16726 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
16727 any edge.
16728
374001cb
RB
167292013-03-25 Richard Biener <rguenther@suse.de>
16730
16731 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
16732 of bitmap.
16733 (memory_references): Likewise.
16734 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
16735 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
16736 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
16737 (gather_mem_refs_in_loops): Fold into ...
16738 (analyze_memory_references): ... this. Move initialization
16739 to tree_ssa_lim_initialize.
16740 (fill_always_executed_in): Rename to ...
16741 (fill_always_executed_in_1): ... this.
16742 (fill_always_executed_in): Move contains_call computation to
16743 this new function from ...
16744 (tree_ssa_lim_initialize): ... here.
16745 (tree_ssa_lim): Call fill_always_executed_in.
16746
57895947
EB
167472013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16748
16749 * postreload.c (reload_combine): Fix code detecting returns.
16750
1e1b18c1
EB
167512013-03-25 Eric Botcazou <ebotcazou@adacore.com>
16752
16753 * function.c (emit_use_return_register_into_block): On cc0 targets,
16754 do not emit the sequence between cc0 setter and user.
16755
9216baf1
KT
167562013-03-25 Kai Tietz <ktietz@redhat.com>
16757
1e1b18c1
EB
16758 * config/i386/predicates.md (local_symbolic_operand): Interpret
16759 dll-imported symbols as none-local.
9216baf1 16760
76421b44
RB
167612013-03-25 Richard Biener <rguenther@suse.de>
16762
16763 * tree-ssa-loop-im.c (struct depend): Remove.
16764 (struct lim_aux_data): Make depends a vec of gimples.
16765 (free_lim_aux_data): Adjust.
16766 (add_dependency): Likewise.
16767 (set_level): Likewise.
16768
d154bfa2
RB
167692013-03-25 Richard Biener <rguenther@suse.de>
16770
16771 PR middle-end/56434
16772 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
16773 the pointer returned by calls with ECF_MALLOC set.
16774
2cd9804e
UB
167752013-03-24 Uros Bizjak <ubizjak@gmail.com>
16776
acef5fe0
UB
16777 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
16778
167792013-03-24 Uros Bizjak <ubizjak@gmail.com>
16780
16781 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB
16782 using MMXMODE mode iterator.
16783 (*move<mode>_internal): Merge with *movv2sf_internal and
16784 *movv2sf_internal_rex64 using MMXMODE mode iterator.
16785
225ccc68
SB
167862013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16787
85c0f02d
SB
16788 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
16789 (record_last_mem_set_info): Likewise.
16790
225ccc68
SB
16791 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
16792 of XNEWVEC followed by memset.
16793 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
16794
b64925dc
SB
167952013-03-23 Steven Bosscher <steven@gcc.gnu.org>
16796
16797 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
16798 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
16799 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
16800 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
16801 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
16802 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
16803 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
16804 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
16805 BARRIER_P instead of GET_CODE.
16806
b9aaf52e
EB
168072013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16808
16809 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
16810 inaccuracy in the probing code.
16811
16812 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
16813 (ctrapdi4): Likewise.
16814
66de4d7c
EB
168152013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16816
16817 * calls.c (expand_call): Add missing guard to code handling return
16818 of non-BLKmode structures in MSB.
16819 * function.c (expand_function_end): Likewise.
16820
4f25c8fb
EB
168212013-03-23 Eric Botcazou <ebotcazou@adacore.com>
16822
16823 * combine.c (try_combine): Adjust comment. Do not add the set of
16824 insn #0 if the destination indirectly is set or dies in insn #2.
16825 Tidy up code to distribute a new note.
16826
23b7850d
UB
168272013-03-22 Uros Bizjak <ubizjak@gmail.com>
16828
16829 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
16830 also for alternatives 16 and 17.
16831
bed852cf
UB
168322013-03-22 Uros Bizjak <ubizjak@gmail.com>
16833
16834 * config/i386/sse.md (*mov<mode>_internal): Merge with
16835 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
16836 Emit insn template depending on type attribute. Use
16837 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
16838 movd instead of movq mnemonic for interunit moves. Rewrite mode
16839 attribute calculation. Remove unit attribute calculation.
16840 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
16841 Set prefix_data16 attribute for DImode ssemov types.
16842 Use Ym instead of y for SSE-MMX conversion alternatives.
16843 Reorder operand constraints.
16844
1e8a7937
SB
168452013-03-22 Steven Bosscher <steven@gcc.gnu.org>
16846
16847 * df.h (df_insn_delete): Adjust prototype.
16848 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
16849 and let it decide whether mark the basic block dirty.
16850 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
16851 * df-scan.c (df_insn_info_delete): New helper function, split
16852 off from df_insn_delete.
16853 (df_scan_free_bb_info): Use it.
16854 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
16855 Likewise.
16856 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
16857 that the insn is actually an insn and it has a non-NULL basic block.
16858 Do not mark basic block dirty if only deleting a DEBUG_INSN.
16859
e14d094c
RB
168602013-03-22 Richard Biener <rguenther@suse.de>
16861
16862 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
16863 dep_ref members.
16864 (mem_ref_alloc): Do not allocate them.
16865 (refs_independent_p): Do not query or maintain a cache.
16866
e9cf7316
RB
168672013-03-22 Richard Biener <rguenther@suse.de>
16868
16869 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
16870 (gather_mem_refs_in_loops): Do not compute it.
16871 (analyze_memory_references): Do not allocate it.
16872 (tree_ssa_lim_finalize): Do not free it.
16873 (for_all_locs_in_loop): Do not query all_refs_in_loop.
16874
f046e81b
RB
168752013-03-22 Richard Biener <rguenther@suse.de>
16876
16877 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 16878
a05c0ddf
IB
168792013-03-22 Ian Bolton <ian.bolton@arm.com>
16880
16881 * config/aarch64/aarch64.c (aarch64_print_operand): New
16882 format specifier for printing a constant in hex.
16883 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
16884 format specifier for printing second operand.
16885
c00217fc
RB
168862013-03-22 Richard Biener <rguenther@suse.de>
16887
16888 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
16889 bitmaps.
16890 (gather_mem_refs_in_loops): Perform store accumulation here.
16891 (create_vop_ref_mapping_loop): Remove.
16892 (create_vop_ref_mapping): Likewise.
16893 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 16894 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB
16895 (record_indep_loop): Remove.
16896 (record_dep_loop): New function.
16897 (ref_indep_loop_p_1): Adjust to only walk over references
16898 in the loop, not its subloops.
16899 (ref_indep_loop_p): Rename to ...
16900 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
16901 maintaining a more fine-grained cache.
16902 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
16903 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
16904
15d19bf8
RB
169052013-03-22 Richard Biener <rguenther@suse.de>
16906
16907 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
16908 (struct mem_ref): Make accesses_in_loop a vec of a vec of
16909 aggregate mem_ref_loc.
16910 (free_mem_ref_locs): Inline into ...
16911 (memref_free): ... this and adjust.
16912 (mem_ref_alloc): Adjust.
16913 (mem_ref_locs_alloc): Remove.
16914 (record_mem_ref_loc): Adjust.
16915 (get_all_locs_in_loop): Rewrite into ...
16916 (for_all_locs_in_loop): ... this iterator.
16917 (rewrite_mem_ref_loc): New functor.
16918 (rewrite_mem_refs): Use for_all_locs_in_loop.
16919 (sm_set_flag_if_changed): New functor.
16920 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
16921 (ref_always_accessed): New functor.
16922 (ref_always_accessed_p): Use for_all_locs_in_loop.
16923
6f37411d
MG
169242013-03-21 Marc Glisse <marc.glisse@inria.fr>
16925
16926 * tree-pass.h (PROP_gimple_lvec): New.
16927 * passes.c (dump_properties): Handle PROP_gimple_lvec.
16928 (init_optimization_passes): Move pass_lower_vector.
16929 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
16930 PROP_gimple_lvec.
16931 (pass_lower_vector): Provide PROP_gimple_lvec.
16932 (pass_lower_vector_ssa): Likewise.
16933 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
16934
511ef689
MW
169352013-03-21 Mark Wielaard <mjw@redhat.com>
16936
16937 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
16938
cc1df30b
UB
169392013-03-21 Uros Bizjak <ubizjak@gmail.com>
16940
16941 * config/i386/i386.md (*movdi_internal): Disparage slightly
16942 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
16943 conversion alternatives.
16944
c845cfe1
JJ
169452013-03-21 Jakub Jelinek <jakub@redhat.com>
16946
37e99116
JJ
16947 PR middle-end/48087
16948 * diagnostic.def (DK_WERROR): New kind.
16949 * diagnostic.h (werrorcount): Define.
16950 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
16951 promoted to DK_ERROR, increment DK_WERROR counter instead of
16952 DK_ERROR counter.
16953 * toplev.c (toplev_main): Call print_ignored_options even if
16954 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
16955 even if just werrorcount is non-zero.
16956
c845cfe1
JJ
16957 PR debug/55608
16958 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
16959 on failure.
16960 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
16961 (string_cst_pool_decl): New function.
16962 (optimize_one_addr_into_implicit_ptr): New function.
16963 (resolve_addr_in_expr): Optimize DWARF location expression
16964 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
16965 which doesn't live in memory, but has DW_AT_location or
16966 DW_AT_const_value, or refers to a string literal, into
16967 DW_OP_GNU_implicit_pointer.
16968 (optimize_location_into_implicit_ptr): New function.
16969 (resolve_addr): If removing DW_AT_location of a variable because
16970 it was DW_OP_addr of address of the variable, but the variable doesn't
16971 live in memory, try to emit const value attribute for the initializer.
16972
08e0cda6
MG
169732013-03-21 Marc Glisse <marc.glisse@inria.fr>
16974
16975 * tree.h (VECTOR_TYPE_P): New macro.
16976 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
16977 TYPE_MODE): Use it.
16978 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
16979 VEC_COND_EXPR cannot be lvalues.
16980 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
16981
d08633b4
MG
169822013-03-21 Marc Glisse <marc.glisse@inria.fr>
16983
16984 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
16985 Restrict the transformation to equal modes.
16986
e6647190
RB
169872013-03-21 Richard Biener <rguenther@suse.de>
16988
16989 PR tree-optimization/39326
16990 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
16991 (MEM_ANALYZABLE): Adjust.
16992 (record_mem_ref_loc): Move bitmap ops ...
16993 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
16994 unanalyzable refs, do not record locations for it.
16995 (analyze_memory_references): Allocate ref zero as shared
16996 unanalyzable ref.
16997 (refs_independent_p): Do not test for unanalyzed mems here.
16998 (ref_indep_loop_p_1): Special-case disambiguation against
16999 the unanalyzed ref.
cc1df30b 17000 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 17001
65074f54
CL
170022013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
17003
17004 * config/arm/arm-protos.h (tune_params): Add
17005 prefer_neon_for_64bits field.
17006 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
17007 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
17008 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
17009 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
17010 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
17011 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
17012 (arm_option_override): Handle -mneon-for-64bits new option.
17013 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
17014 (prefer_neon_for_64bits): Declare new variable.
17015 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 17016 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL
17017 (arch_enabled): Handle new arch types. Remove support for onlya8
17018 and nota8.
17019 (one_cmpldi2): Use new arch names.
bc5faa5b 17020 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL
17021 * config/arm/arm.opt (mneon-for-64bits): Add option.
17022 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
17023 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
17024 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
17025 of onlya8.
17026 * doc/invoke.texi (-mneon-for-64bits): Document.
17027
5a2d2a79
RB
170282013-03-21 Richard Biener <rguenther@suse.de>
17029
17030 PR tree-optimization/39326
17031 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
17032 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 17033 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 17034
5abe1e05
RB
170352013-03-21 Richard Biener <rguenther@suse.de>
17036
17037 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
17038 (vect_insert_into_interleaving_chain): Likewise.
17039 (vect_drs_dependent_in_basic_block): Inline ...
17040 (vect_slp_analyze_data_ref_dependence): ... here. New function,
17041 split out from ...
17042 (vect_analyze_data_ref_dependence): ... here. Simplify.
17043 (vect_check_interleaving): Simplify.
17044 (vect_analyze_data_ref_dependences): Likewise. Split out ...
17045 (vect_slp_analyze_data_ref_dependences): ... this new function.
17046 (dr_group_sort_cmp): New function.
17047 (vect_analyze_data_ref_accesses): Compute data-reference groups
17048 here instead of in vect_analyze_data_ref_dependence. Use
17049 a more efficient algorithm.
17050 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
17051 vect_slp_analyze_data_ref_dependences. Call
17052 vect_analyze_data_ref_accesses earlier.
17053 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
17054 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
17055 (vect_slp_analyze_data_ref_dependences): New prototype.
17056
cad1735b
RB
170572013-03-21 Richard Biener <rguenther@suse.de>
17058
17059 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
17060 ref is stored in the loop.
17061 (find_refs_for_sm): Walk only over all stores.
17062 (store_motion_loop): Allocate from lim_bitmap_obstack.
17063 (store_motion): Likewise.
17064
141310ef
RB
170652013-03-21 Richard Biener <rguenther@suse.de>
17066
17067 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
17068 Update virtual SSA form.
17069
5022315a
RO
170702013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17071
17072 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
17073 * configure: Regenerate.
17074 * config.in: Regenerate.
17075 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
17076 if !HAVE_LD_EH_FRAME_CIEV3.
17077
50d4421c
RB
170782013-03-21 Richard Biener <rguenther@suse.de>
17079
17080 * tree-cfg.c (verify_expr_no_block): New function.
17081 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
17082 nor DECL_VALUE_EXPR have locations with associated blocks.
17083 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
17084 (clear_unused_block_pointer): Remove code dealing with
17085 blocks in DECL_DEBUG_EXPR locations.
17086
839b422f
RB
170872013-03-21 Richard Biener <rguenther@suse.de>
17088
17089 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
17090 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
17091 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
17092 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
17093 instead of DECL_DEBUG_EXPR_IS_FROM.
17094 * gimplify.c (gimplify_modify_expr): Likewise.
17095 * tree-cfg.c (verify_expr_location_1): Likewise.
17096 * tree-complex.c (create_one_component_var): Likewise.
17097 * tree-sra.c (create_access_replacement): Likewise.
17098 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
17099 (clear_unused_block_pointer): Likewise.
17100 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
17101 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
17102 * var-tracking.c (var_debug_decl): Likewise.
17103 (track_expr_p): Likewise.
17104 * tree-inline.c (add_local_variables): Likewise. Set
17105 DECL_HAS_DEBUG_EXPR_P after copying it.
17106 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
17107 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
17108
fe04878d
UB
171092013-03-21 Uros Bizjak <ubizjak@gmail.com>
17110
17111 PR bootstrap/56656
17112 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
17113 * configure: Regenerate.
17114 * config.in: Regenerate.
17115 * config/i386/i386.md (*movdf_internal): Use
17116 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
17117 movd instead of movq mnemonic for interunit moves.
17118 (*movdi_internal): Ditto.
17119
bd059b26 171202013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N
17121
17122 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
17123 (abd<mode>_3): New pattern.
17124 (aba<mode>_3): New pattern.
17125 (fabd<mode>_3): New pattern.
17126
bd059b26 171272013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N
17128
17129 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
17130 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
17131 occurrence of REGISTER_PREFIX as its empty string.
17132
4f1f78b9
JL
171332013-03-20 Jeff Law <law@redhat.com>
17134
17135 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
17136 addititional equivalences for equality comparisons between an SSA_NAME
17137 and a constant where the SSA_NAME was set from a widening conversion.
17138
327a1118
WL
171392013-03-20 Walter Lee <walt@tilera.com>
17140
17141 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
17142
813e0036
UB
171432013-03-20 Uros Bizjak <ubizjak@gmail.com>
17144
17145 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
17146 depending on type attribute.
17147 (*movti_internal): Ditto.
17148 (*movtf_internal): Ditto.
17149 (*movxf_internal): Ditto.
17150 (*movdf_internal): Ditto.
17151 (*movsf_internal): Ditto.
17152
7cf34aae
UB
171532013-03-20 Uros Bizjak <ubizjak@gmail.com>
17154
17155 * config/i386/i386.md (*movti_internal): Set prefix attribute to
17156 maybe_vex for sselog1 and ssemov types.
17157 (*movdi_internal): Reorder operand constraints.
17158 (*movsi_internal): Ditto. Set prefix attribute to
17159 maybe_vex for sselog1 and ssemov types.
17160 (*movtf_internal): Set prefix attribute to maybe_vex
17161 for sselog1 and ssemov types.
17162 (*movdf_internal): Ditto. Set prefix_data16 attribute for
17163 DImode ssemov types. Reorder operand constraints.
17164 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
17165 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
17166 attribute for SImode ssemov types. Reorder operand constraints.
17167
19321415
MJ
171682013-03-20 Martin Jambor <mjambor@suse.cz>
17169
17170 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
17171 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
17172
12211b99 171732013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM
17174
17175 * config/rs6000/predicates.md (indexed_address, update_address_mem
17176 update_indexed_address_mem): New predicates.
17177 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
17178 attribute for load/store instructions.
17179 * config/rs6000/dfp.md (movsd_store): Likewise.
17180 (movsd_load): Likewise.
17181 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
17182 (unnamed HI->DI extend define_insn): Likewise.
17183 (unnamed SI->DI extend define_insn): Likewise.
17184 (unnamed QI->SI extend define_insn): Likewise.
17185 (unnamed QI->HI extend define_insn): Likewise.
17186 (unnamed HI->SI extend define_insn): Likewise.
17187 (unnamed HI->SI extend define_insn): Likewise.
17188 (extendsfdf2_fpr): Likewise.
17189 (movsi_internal1): Likewise.
17190 (movsi_internal1_single): Likewise.
17191 (movhi_internal): Likewise.
17192 (movqi_internal): Likewise.
17193 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
17194 attribute for load/store instructions.
17195 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
17196 instructions.
17197 (mov<mode>_softfloat): Likewise.
17198 (mov<mode>_hardfloat32): Likewise.
17199 (mov<mode>_hardfloat64): Likewise.
17200 (mov<mode>_softfloat64): Likewise.
17201 (movdi_internal32): Likewise.
17202 (movdi_internal64): Likewise.
17203 (probe_stack_<mode>): Likewise.
17204
172052013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
17206
17207 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
17208 floating point, and decimal floating point to reload iterator.
17209
17210 * config/rs6000/constraints.md (wl constraint): New constraints to
17211 return FLOAT_REGS if certain options are used to reduce the number
17212 of separate patterns that exist in the file.
17213 (wx constraint): Likewise.
17214 (wz constraint): Likewise.
17215
17216 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17217 -mdebug=reg, print wg, wl, wx, and wz constraints.
17218 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17219 Initialize the reload functions for 64-bit binary/decimal floating
17220 point types.
17221 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17222 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17223 create the buffer on the stack to overcome not having a 32-bit
17224 load and store.
17225 (rs6000_emit_move): Likewise.
17226 (rs6000_secondary_memory_needed_rtx): Likewise.
17227 (rs6000_alloc_sdmode_stack_slot): Likewise.
17228 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17229 via xxlxor, just like DFmode 0.0.
17230
17231 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
17232 define as 1 if we are running on a power7 or newer.
17233 (enum r6000_reg_class_enum): Add new constraints.
17234
17235 * config/rs6000/dfp.md (movsd): Delete, combine with binary
17236 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17237 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 17238 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM
17239 (movsd splitter): Likewise.
17240 (movsd_hardfloat): Likewise.
17241 (movsd_softfloat): Likewise.
17242
17243 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
17244 binary and decimal floating point moves.
17245 (fmove_ok): New attributes to combine binary and decimal floating
17246 point moves, and to combine power6x (mfpgpr) moves along normal
17247 floating moves.
17248 (real_value_to_target): Likewise.
17249 (f32_lr): Likewise.
17250 (f32_lm): Likewise.
17251 (f32_li): Likewise.
17252 (f32_sr): Likewise.
17253 (f32_sm): Likewise.
17254 (f32_si): Likewise.
17255 (movsf): Combine binary and decimal floating point moves. Combine
17256 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 17257 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM
17258 (mov<mode> for SFmode/SDmode); Likewise.
17259 (SFmode/SDmode splitters): Likewise.
17260 (movsf_hardfloat): Likewise.
17261 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
17262 (movsf_softfloat): Likewise.
17263 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
17264
17265 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
17266 wx and wz constraints.
17267
17268 * config/rs6000/constraints.md (wg constraint): New constraint to
17269 return FLOAT_REGS if -mmfpgpr (power6x) was used.
17270
17271 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
17272 constraint.
17273
17274 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
17275 -mdebug=reg, print wg, wl, wx, and wz constraints.
17276 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
17277 Initialize the reload functions for 64-bit binary/decimal floating
17278 point types.
17279 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
17280 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
17281 create the buffer on the stack to overcome not having a 32-bit
17282 load and store.
17283 (rs6000_emit_move): Likewise.
17284 (rs6000_secondary_memory_needed_rtx): Likewise.
17285 (rs6000_alloc_sdmode_stack_slot): Likewise.
17286 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
17287 via xxlxor, just like DFmode 0.0.
17288
c6d5ff83
MM
17289 * config/rs6000/dfp.md (movdd): Delete, combine with binary
17290 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
17291 with other moves by using conditional constraits (wg). Use LFIWZX
17292 and STFIWX for loading SDmode on power7.
17293 (movdd splitters): Likewise.
17294 (movdd_hardfloat32): Likewise.
17295 (movdd_softfloat32): Likewise.
17296 (movdd_hardfloat64_mfpgpr): Likewise.
17297 (movdd_hardfloat64): Likewise.
17298 (movdd_softfloat64): Likewise.
17299
17300 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
17301 64-bit binary and decimal floating point moves.
17302 (FMOVE64X): Likewise.
17303 (movdf): Combine 64-bit binary and decimal floating point moves.
17304 Combine power6x (mfpgpr) moves with other moves by using
17305 conditional constraits (wg).
17306 (mov<mode> for DFmode/DDmode): Likewise.
17307 (DFmode/DDmode splitters): Likewise.
17308 (movdf_hardfloat32): Likewise.
17309 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
17310 (movdf_softfloat32): Likewise.
17311 (movdf_hardfloat64_mfpgpr): Likewise.
17312 (movdf_hardfloat64): Likewise.
17313 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
17314 (movdf_softfloat64): Likewise.
17315 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
17316 (reload_<mode>_load): Move to later in the file so they aren't in
17317 the middle of the floating point move insns.
17318 (reload_<mode>_store): Likewise.
17319
17320 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
17321 constraint.
17322
17323 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
17324 constraint if -mdebug=reg.
bd059b26
UB
17325 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
17326 Enable using dd reload support if needed.
c6d5ff83
MM
17327
17328 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
17329 binary and decimal floating point moves in rs6000.md.
17330 (movtd_internal): Likewise.
17331
17332 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
17333 decimal floating point moves.
17334 (movtf): Likewise.
17335 (movtf_internal): Likewise.
17336 (mov<mode>_internal, TDmode/TFmode): Likewise.
17337 (movtf_softfloat): Likewise.
17338 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
17339
17340 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
17341 movdi_internal64, using wg constraint for move direct operations.
17342 (movdi_internal64): Likewise.
17343
17344 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
17345 MODES_TIEABLE_P for selected modes. Print the numerical value of
17346 the various virtual registers. Use GPR/FPR first/last values,
17347 instead of hard coding the register numbers. Print which modes
17348 have reload functions registered.
bd059b26
UB
17349 (rs6000_option_override_internal): If -mdebug=reg, trace the options
17350 settings before/after setting cpu, target and subtarget settings.
17351 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
17352 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM
17353 (rs6000_secondary_reload_fail): Likewise.
17354 (rs6000_secondary_reload_inner): Likewise.
17355
17356 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
17357 macros for first/last GPR and FPR registers.
17358 (LAST_GPR_REGNO): Likewise.
17359 (FIRST_FPR_REGNO): Likewise.
17360 (LAST_FPR_REGNO): Likewise.
17361
17362 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
17363 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
17364 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
17365 (vcond<mode><mode>): Likewise.
17366 (vcondu<mode><mode>): Likewise.
17367 (vector_gtu<mode>): Likewise.
17368 (vector_gte<mode>): Likewise.
17369 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 17370 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM
17371 (ior<mode>3): Likewise.
17372 (and<mode>3): Likewise.
17373 (one_cmpl<mode>2): Likewise.
17374 (nor<mode>3): Likewise.
17375 (andc<mode>3): Likewise.
17376
17377 * config/rs6000/constraints.md (wt constraint): New constraint
17378 that returns VSX_REGS if TImode is allowed in VSX registers.
17379
17380 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
17381 constant under VSX.
17382
17383 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
17384 similar to TImode, but it is restricted to being in the GPRs.
17385
17386 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
17387 TImode to occupy a single VSX register.
17388
17389 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
17390 -mvsx-timode for power7/power8.
17391 (power7 cpu): Likewise.
17392 (power8 cpu): Likewise.
17393
17394 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
17395 sure that TFmode/TDmode take up two registers if they are ever
17396 allowed in the upper VSX registers.
17397 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
17398 registers.
17399 (rs6000_init_hard_regno_mode_ok): Likewise.
17400 (rs6000_debug_reg_global): Add debugging for PTImode and wt
17401 constraint. Print if LRA is turned on.
17402 (rs6000_option_override_internal): Give an error if -mvsx-timode
17403 and VSX is not enabled.
17404 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
17405 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
17406 to reg+offset addressing. Use PTImode when checking offset
17407 addresses for validity.
17408 (reg_offset_addressing_ok_p): Likewise.
17409 (rs6000_legitimate_offset_address_p): Likewise.
17410 (rs6000_legitimize_address): Likewise.
17411 (rs6000_legitimize_reload_address): Likewise.
17412 (rs6000_legitimate_address_p): Likewise.
17413 (rs6000_eliminate_indexed_memrefs): Likewise.
17414 (rs6000_emit_move): Likewise.
17415 (rs6000_secondary_reload): Likewise.
17416 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
17417 reloads to fpr registers to continue to use reg+offset addressing,
17418 but 64-bit reloads to altivec registers need reg+reg addressing.
17419 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
17420 it. Treat LO_SUM like a PLUS operation.
17421 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 17422 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM
17423 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
17424 registers to share a register with a smaller sized type, since VSX
17425 puts scalars in the upper 64-bits.
17426 (print_operand): Add support for PTImode.
17427 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
17428 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
17429 registers, but don't have arithmetic support.
17430 (rs6000_memory_move_cost): Add test for VSX.
17431 (rs6000_opt_masks): Add -mvsx-timode.
17432
17433 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
17434 for TImode.
17435 (VSs): Likewise.
17436 (VSr): Use wt constraint for TImode.
17437 (VSv): Drop TImode support.
17438 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
17439 (vsx_movti_64bit): Likewise.
17440 (vsx_movti_32bit): Likewise.
17441 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
17442 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
17443 one '?' on the appropriate output constraint. Do not allow TImode
17444 logical operations on 32-bit systems.
17445 (vsx_ior<mode>3): Likewise.
17446 (vsx_xor<mode>3): Likewise.
17447 (vsx_one_cmpl<mode>2): Likewise.
17448 (vsx_nor<mode>3): Likewise.
17449 (vsx_andc<mode>3): Likewise.
17450 (vsx_concat_<mode>): Likewise.
17451 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
17452
17453 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
17454 OPTION_MASK_VSX_TIMODE.
17455 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
17456 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
17457
17458 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
17459 (TI2 iterator): New iterator for TImode, PTImode.
17460 (wd mode attribute): Add values for vector types.
bd059b26
UB
17461 (movti_string): Replace TI move operations with operations for TImode
17462 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM
17463 (mov<mode>_string, TImode/PTImode): Likewise.
17464 (movti_ppc64): Likewise.
17465 (mov<mode>_ppc64, TImode/PTImode): Likewise.
17466 (TI mode splitters): Likewise.
17467
17468 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
17469 constraint.
17470
1fc5eced
MG
174712013-03-20 Marc Glisse <marc.glisse@inria.fr>
17472
17473 PR tree-optimization/56355
17474 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
17475 Also handle integers with undefined overflow.
17476
22c4c869
CM
174772013-03-20 Catherine Moore <clm@codesourcery.com>
17478 Maciej W. Rozycki <macro@codesourcery.com>
17479 Tom de Vries <tom@codesourcery.com>
12211b99 17480 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM
17481 Iain Sandoe <iain@codesourcery.com>
17482 Nathan Froyd <froydnj@codesourcery.com>
12211b99 17483 Chao-ying Fu <fu@mips.com>
22c4c869
CM
17484
17485 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 17486 Document new function attributes.
22c4c869
CM
17487 * doc/invoke.texi (minterlink-compressed, mmicromips,
17488 m14k, m14ke, m14kec): Document new options.
17489 (minterlink-mips16): Update documentation.
17490 * doc/md.texi (ZC, ZD): Document new constraints.
17491 * configure.ac (gcc_cv_as_micromips): Check if linker
17492 supports the .set micromips directive.
17493 * configure: Regenerate.
17494 * config.in: Regenerate.
17495 * config/mips/mips-tables.opt: Regenerate.
17496 * config/mips/micromips.md: New file.
17497 * constraints.md (ZC, ZD): New constraints.
17498 * config/mips/predicates.md (movep_src_register): New predicate.
17499 (movep_src_operand): New predicate.
17500 (non_volatile_mem_operand): New predicate.
17501 * config/mips/mips.md (multimem): New type.
17502 (length): Differentiate between 17-bit and 18-bit branch offsets.
17503 (MOVEP1, MOVEP2): New mode iterator.
33159866 17504 (mov_<load>l): Use ZC constraint.
22c4c869
CM
17505 (mov_<load>r): Likewise.
17506 (mov_<store>l): Likewise.
17507 (mov_<store>r): Likewise.
17508 (*branch_equality<mode>_inverted): Add microMIPS support.
17509 (*branch_equality<mode>): Likewise.
17510 (*jump_absolute): Likewise.
17511 (indirect_jump_<mode>): Likewise.
17512 (tablejump_<mode>): Likewise.
17513 (<optab>_internal): Likewise.
17514 (sibcall_internal): Likewise.
17515 (sibcall_value_internal): Likewise.
17516 (prefetch): Use constraint ZD.
17517 * config/mips/mips.opt (minterlink-compressed): New option.
17518 (minterlink-mips16): Now an alias for minterlink-compressed.
17519 (mmicromips): New option.
17520 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
17521 (compare_and_swap_12): Likewise.
17522 (sync_add<mode>): Likewise.
17523 (sync_<optab>_12): Likewise.
17524 (sync_old_<optab>_12): Likewise.
17525 (sync_new_<optab>_12): Likewise.
17526 (sync_nand_12): Likewise.
17527 (sync_old_nand_12): Likewise.
17528 (sync_new_nand_12): Likewise.
17529 (sync_sub<mode>): Likewise.
17530 (sync_old_add<mode>): Likewise.
17531 (sync_old_sub<mode>): Likewise.
17532 (sync_new_add<mode>): Likewise.
17533 (sync_new_sub<mode>): Likewise.
17534 (sync_<optab><mode>): Likewise.
17535 (sync_old_<optab><mode>): Likewise.
17536 (sync_new_<optab><mode>): Likewise.
17537 (sync_nand<mode>): Likewise.
17538 (sync_old_nand<mode>): Likewise.
17539 (sync_new_nand<mode>): Likewise.
17540 (sync_lock_test_and_set<mode>): Likewise.
17541 (test_and_set_12): Likewise.
17542 (atomic_compare_and_swap<mode>): Likewise.
17543 (atomic_exchange<mode>_llsc): Likewise.
17544 (atomic_fetch_add<mode>_llsc): Likewise.
17545 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
17546 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
17547 (umips_save_restore_pattern_p): Likewise.
17548 (umips_load_store_pair_p): Likewise.
17549 (umips_output_load_store_pair): Likewise.
17550 (umips_movep_target_p): Likewise.
17551 (umips_12bit_offset_address_p): Likewise.
17552 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
17553 (mips_base_mips16): Rename this...
17554 (mips_base_compression_flags): ...to this. Update all uses.
17555 (mips_attribute_table): Add micromips, nomicromips and nocompression.
17556 (mips_mips16_decl_p): Delete.
17557 (mips_nomips16_decl_p): Delete.
33159866 17558 (mips_get_compress_on_flags): New function.
22c4c869
CM
17559 (mips_get_compress_off_flags): New function.
17560 (mips_get_compress_mode): New function.
17561 (mips_get_compress_on_name): New function.
17562 (mips_get_compress_off_name): New function.
17563 (mips_insert_attributes): Support multiple compression types.
17564 (mips_merge_decl_attributes): Likewise.
17565 (umips_12bit_offset_address_p): New function.
17566 (mips_start_function_definition): Emit .set micromips directive.
17567 (mips_call_may_need_jalx_p): New function.
17568 (mips_function_ok_for_sibcall): Add microMIPS support.
17569 (mips_print_operand_punctuation): Support short delay slots and
17570 compact jumps.
17571 (umips_swm_mask, umips_swm_encoding): New.
17572 (umips_build_save_restore): New function.
17573 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
17574 (was_mips16_p): Remove.
17575 (old_compression_mode): New.
17576 (mips_set_compression_mode): New function.
17577 (mips_set_current_function): Add microMIPS support.
17578 (mips_option_override): Likewise.
17579 (umips_save_restore_pattern_p): New function.
17580 (umips_output_save_restore): New function.
17581 (umips_load_store_pair_p_1): New function.
17582 (umips_load_store_pair_p): New function.
17583 (umips_output_load_store_pair_1): New function.
17584 (umips_output_load_store_pair): New function.
17585 (umips_movep_target_p) New function.
17586 (mips_prepare_pch_save): Add microMIPS support.
17587 * config/mips/mips.h (TARGET_COMPRESSION): New.
17588 (TARGET_CPU_CPP_BUILTINS): Update macro
17589 to use new compression flags and to support microMIPS.
17590 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
17591 (MIPS_ARCH_FLOAT_SPEC): Likewise.
17592 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
17593 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
17594 (ASM_SPEC): Support mmicromips and mno-micromips.
17595 (M16STORE_REG_P): New macro.
17596 (MIPS_CALL): Support TARGET_MICROMIPS.
17597 (MICROMIPS_J): New macro.
17598 (mips_base_mips16): Rename this...
17599 (mips_base_compression_flags): ...to this.
17600 (UMIPS_12BIT_OFFSET_P): New macro.
17601 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
17602 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB
176032013-03-20 Richard Biener <rguenther@suse.de>
17604
17605 PR tree-optimization/56661
17606 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
17607 the result does not have to be distinct.
17608
54714c68
RB
176092013-03-20 Richard Biener <rguenther@suse.de>
17610
17611 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
17612 remap_gimple_op_r.
17613
cca1130d 176142013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 17615 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS
17616
17617 PR rtl-optimization/56605
17618 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
17619
2a930377
UB
176202013-03-20 Uros Bizjak <ubizjak@gmail.com>
17621
17622 PR bootstrap/56656
17623 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
17624 that require movd instead of movq.
17625
d6d305fe
RB
176262013-03-20 Richard Biener <rguenther@suse.de>
17627
17628 * tree-ssa-structalias.c (struct variable_info): Add pointer
17629 to the first field of an aggregate with sub-vars. Make
17630 this and the pointer to the next subfield its ID.
17631 (vi_next): New function.
17632 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
17633 storedanything_id, integer_id): Increment by one.
17634 (new_var_info, get_call_vi, lookup_call_clobber_vi,
17635 get_call_clobber_vi): Adjust.
17636 (solution_set_expand): Simplify and speedup.
17637 (solution_set_add): Inline into ...
17638 (set_union_with_increment): ... this. Adjust accordingly.
17639 (do_sd_constraint): Likewise.
17640 (do_ds_constraint): Likewise.
17641 (do_complex_constraint): Simplify.
17642 (build_pred_graph): Adjust.
17643 (solve_graph): Likewise. Simplify and speedup.
17644 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
17645 get_constraint_for_component_ref, get_constraint_for_1,
17646 first_vi_for_offset, first_or_preceding_vi_for_offset,
17647 create_function_info_for, create_variable_info_for_1,
17648 create_variable_info_for, intra_create_variable_infos): Adjust.
17649 (init_base_vars): Push NULL for ID zero.
17650 (compute_points_to_sets): Adjust.
17651
a271b42d
RB
176522013-03-20 Richard Biener <rguenther@suse.de>
17653
17654 * cfgloop.c (verify_loop_structure): Streamline and avoid
17655 ICEing on corrupt loop tree.
17656 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
17657 loop tree.
17658
ebd65954
RB
176592013-03-20 Richard Biener <rguenther@suse.de>
17660
17661 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
17662 check whether an SSA update is needed.
17663
4547b7ee
RS
176642013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
17665
17666 * config/mips/constraints.md (T): Rename to...
17667 (Yf): ...this.
17668 (U): Rename to...
17669 (Yd): ...this.
17670 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
17671 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
17672
0b8cdc58
IB
176732013-03-19 Ian Bolton <ian.bolton@arm.com>
17674
17675 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
17676 (*subsi3_carryin_uxtw): Likewise.
17677
5977a10d
IB
176782013-03-19 Ian Bolton <ian.bolton@arm.com>
17679
17680 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
17681 (*rorsi3_insn_uxtw): Likewise.
17682
bd83ff2c
IB
176832013-03-19 Ian Bolton <ian.bolton@arm.com>
17684
17685 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
17686 (*extrsi5_insn_uxtw): Likewise.
17687
ba65123c
RB
176882013-03-19 Richard Biener <rguenther@suse.de>
17689
17690 PR tree-optimization/56273
17691 * passes.c (init_optimization_passes): Move second VRP after DOM.
17692
2eac0476
UB
176932013-03-19 Uros Bizjak <ubizjak@gmail.com>
17694
17695 * config/i386/i386.md (*movti_internal): Merge from
17696 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
17697 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
17698 nox64 isa attributes.
17699
6e55eda7
RB
177002013-03-18 Richard Biener <rguenther@suse.de>
17701
17702 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
17703 (unite): Likewise.
17704 (merge_node_constraints): Likewise.
17705 (build_succ_graph): Likewise.
17706 (valid_graph_edge): Inline into single caller.
17707 (unify_nodes): Likewise. Use bitmap_set_bit return value
17708 and cache varinfo.
17709 (scc_visit): Fix formatting and variable use.
17710 (do_sd_constraint): Use gcc_checking_assert.
17711 (do_ds_constraint): Likewise.
17712 (do_complex_constraint): Likewise.
17713 (condense_visit): Likewise. Cleanup.
17714 (dump_pred_graph): New function.
17715 (perform_var_substitution): Dump the pred-graph before
17716 variable substitution.
17717 (find_equivalent_node): Use gcc_checking_assert.
17718 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
17719
4bdd44c4
RB
177202013-03-18 Richard Biener <rguenther@suse.de>
17721
17722 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
17723 Remove cond_expr_stmt_list argument and do not gimplify the
17724 built expression.
17725 (vect_loop_versioning): Adjust.
17726 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
17727 Cleanup to use less temporaries.
17728 (vect_create_data_ref_ptr): Cleanup.
17729
38c56a5b
JJ
177302013-03-18 Jakub Jelinek <jakub@redhat.com>
17731
17732 PR tree-optimization/56635
17733 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
17734 require types_compatible_p types.
17735
20b2e6a0
NC
177362013-03-18 Nick Clifton <nickc@redhat.com>
17737
a6178a25
NC
17738 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
17739 spurious backslash.
17740
20b2e6a0
NC
17741 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
17742 Add missing line to comment describing function.
17743
92608d0e
RB
177442013-03-18 Richard Biener <rguenther@suse.de>
17745
17746 PR tree-optimization/56210
17747 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
17748 Handle string / character search functions.
17749 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
17750
31348d52
RB
177512013-03-18 Richard Biener <rguenther@suse.de>
17752
17753 PR middle-end/56483
17754 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
17755 and implement properly.
17756 * gimple.h (gimple_cond_single_var_p): Remove.
17757
fcac74a1
RB
177582013-03-18 Richard Biener <rguenther@suse.de>
17759
17760 * tree-data-ref.h (find_data_references_in_loop): Declare.
17761 * tree-data-ref.c (get_references_in_stmt): Use a stack
17762 vector pre-allocated in the callers.
17763 (find_data_references_in_stmt): Adjust.
17764 (graphite_find_data_references_in_stmt): Likewise.
17765 (create_rdg_vertices): Likewise.
17766 (find_data_references_in_loop): Export.
17767 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
17768 Compute dependences here...
17769 (vect_analyze_data_refs): ...not here. When we encounter
17770 a non-vectorizable data reference in basic-block vectorization
17771 truncate the data reference vector. Do not bother to
17772 fixup data-dependence information for gather loads.
17773 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
17774 of data references, as reported.
17775
0d5a1b56
RB
177762013-03-18 Richard Biener <rguenther@suse.de>
17777
17778 PR tree-optimization/3713
17779 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
17780 has_constants and expr.
17781 (stmt_has_constants): Properly valueize SSA names when deciding
17782 whether the stmt has constants.
17783
789c34e3
RB
177842013-03-18 Richard Biener <rguenther@suse.de>
17785
17786 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
17787 whole function when there is nothing to do.
17788 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
17789 * tree-vectorizer.c (vectorize_loops): Update virtual and
17790 loop-closed SSA once.
17791 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
17792
076b4605
RB
177932013-03-18 Richard Biener <rguenther@suse.de>
17794
17795 PR middle-end/56113
17796 * domwalk.c (bb_postorder): New global static.
17797 (cmp_bb_postorder): New function.
17798 (walk_dominator_tree): Replace scheme imposing an order for
17799 visiting dominator sons by one sorting them at the time they
17800 are pushed on the stack.
17801
bdb01696
RB
178022013-03-18 Richard Biener <rguenther@suse.de>
17803
17804 PR tree-optimization/39326
17805 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
17806 (struct mem_ref): Replace mem member with ao_ref typed member.
17807 (MEM_ANALYZABLE): Adjust.
17808 (memref_eq): Likewise.
17809 (mem_ref_alloc): Likewise.
17810 (gather_mem_refs_stmt): Likewise.
17811 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
17812 (execute_sm_if_changed_flag_set): Adjust.
17813 (execute_sm): Likewise.
17814 (ref_always_accessed_p): Likewise.
17815 (refs_independent_p): Likewise.
17816 (can_sm_ref_p): Likewise.
17817
12d2dc5e
JJ
178182013-03-18 Jakub Jelinek <jakub@redhat.com>
17819
17820 PR c/56566
17821 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
17822 return 1 even for !unsignedp.
17823
286fb677
UB
178242013-03-17 Uros Bizjak <ubizjak@gmail.com>
17825
17826 * config/i386/i386.md (isa): Add x64 and nox64.
17827 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
17828 (*pushtf): Enable *roF alternative for x64 isa only.
17829 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
17830 mode attribute of integer alternatives to DImode for TARGET_64BIT.
17831 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
17832 (*movtf_internal): Merge from *movtf_internal_rex64 and
17833 *movtf_internal_sse. Use x64 and nox64 isa attributes.
17834 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
17835 nox64 isa attributes.
17836 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
17837 nox64 isa attributes.
17838 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
17839
88b97037
UB
178402013-03-17 Uros Bizjak <ubizjak@gmail.com>
17841
17842 * config/alpha/alpha.c (TARGET_LRA_P): New define.
17843
9f4f1735
JJ
178442013-03-17 Jakub Jelinek <jakub@redhat.com>
17845
17846 PR target/56640
17847 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
17848 class names. Remove trailing comma after "ALL_REGS".
17849
48f4a6fa
JH
178502013-03-16 Jan Hubicka <jh@suse.cz>
17851
17852 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
17853 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
17854 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
17855 of cgraph_get_create_node.
17856 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
17857
98e81123
JM
178582013-03-16 Jason Merrill <jason@redhat.com>
17859
17860 PR debug/49090
17861 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
17862 with DW_AT_default_value.
17863
8a298c04
JJ
178642013-03-16 Jakub Jelinek <jakub@redhat.com>
17865
17866 * BASE-VER: Set to 4.9.0.
17867
4323afa0
AK
178682013-03-14 Andi Kleen <ak@linux.intel.com>
17869
17870 PR target/56619
17871 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
17872 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
17873 Document _x* TSX intrinsics.
17874
b3c0d469
JJ
178752013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
17876 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI
17877
17878 * configure.ac: Add MicroBlaze TLS support detection.
17879 * configure: Regenerate.
88b97037
UB
17880 * config/microblaze/microblaze-protos.h
17881 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
17882 symbol_mentioned_p, label_mentioned_p): Add prototypes.
17883 * config/microblaze/microblaze.c (microblaze_address_type): Add
17884 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI
17885 (microblaze_address_info): Add tls_reloc.
17886 (TARGET_HAVE_TLS): Define.
17887 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
17888 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB
17889 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
17890 load_tls_operand, microblaze_call_tls_get_addr,
17891 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI
17892 (microblaze_classify_unspec): Handle UNSPEC_TLS.
17893 (get_base_reg): Use microblaze_tls_symbol_p.
17894 (microblaze_classify_address): Handle TLS.
88b97037
UB
17895 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
17896 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI
17897 (microblaze_legitimize_address): Handle TLS.
17898 (microblaze_address_insns): Handle ADDRESS_TLS.
17899 (pic_address_needs_scratch): Handle TLS.
17900 (print_operand_address): Handle TLS.
17901 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
17902 (microblaze_expand_move): Handle TLS.
88b97037
UB
17903 (microblaze_legitimate_constant_p): Check
17904 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI
17905 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
17906 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
17907 (PIC_OFFSET_TABLE_REGNUM): Set.
17908 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
17909 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
17910 (addsi3, movsi_internal2, movdf_internal): Update constraints
17911 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB
17912 (move_operand): Redefine as move_src_operand,
17913 check microblaze_tls_referenced_p.
8cc9a5a5 17914
d803a491
IB
179152013-03-14 Ian Bolton <ian.bolton@arm.com>
17916
17917 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
17918 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
17919
a8504f22
IB
179202013-03-14 Ian Bolton <ian.bolton@arm.com>
17921
17922 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
17923 CC mode for AND.
17924
df2dfaea
JJ
179252013-03-14 Jakub Jelinek <jakub@redhat.com>
17926
fbd28bc3
JJ
17927 PR tree-optimization/53265
17928 * common.opt (Waggressive-loop-optimizations): New option.
17929 * tree-ssa-loop-niter.c: Include tree-pass.h.
17930 (do_warn_aggressive_loop_optimizations): New function.
17931 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
17932 if number_of_latch_executions returned constant.
17933 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
17934 early. If number_of_latch_executions returned constant, set
17935 nb_iterations_upper_bound back to it.
17936 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
17937 field.
17938 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
17939 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
17940
df2dfaea
JJ
17941 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
17942 (MULTILIB_OSDIRNAMES): Set.
17943 * genmultilib: If defaultosdirname doesn't start with :: , set
17944 defaultosdirname2 instead, clear it and emit two . multilib_raw
17945 entries instead of just one.
17946
ee0d2b68
KK
179472013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
17948
17949 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17950 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
17951 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
17952 (SUBTARGET_OVERRIDE_OPTIONS): New.
17953
decc676e
OE
179542013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
17955
17956 PR target/49880
17957 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
17958 (musermode): Convert to Var(TARGET_USERMODE).
17959 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
17960 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
17961 * config/sh/sh.c (sh_option_override): Use
17962 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
17963 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
17964 condition.
17965 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
17966 TARGET_SH4.
17967 (udivsi3_i4_single, divsi3_i4_single): Use
17968 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
17969
f4b719c7
DK
179702013-03-13 Dave Korn <dave.korn.cygwin@....>
17971
17972 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
17973 default setting.
17974
c40eced0
RB
179752013-03-13 Richard Biener <rguenther@suse.de>
17976
17977 PR tree-optimization/56608
17978 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
17979 calls when vectorizing basic-blocks.
17980
1bfa5973
JJ
179812013-03-13 Jakub Jelinek <jakub@redhat.com>
17982
17983 PR plugins/45078
17984 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
17985 tm_file.
17986
6fcf5434
JJ
179872013-03-12 Jakub Jelinek <jakub@redhat.com>
17988
17989 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
17990
4f38fa8c
JH
179912013-03-11 Jan Hubicka <jh@suse.cz>
17992
17993 PR lto/56557
17994 * lto-streamer-out.c (output_symbol_p): Skip references from
17995 constructors of external variables.
17996
c5c5ba89
JH
179972013-03-11 Jan Hubicka <jh@suse.cz>
17998
17999 PR middle-end/56571
18000 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
18001 from pseudos.
18002 * emit-rtl.c (verify_rtx_sharing): Likewise.
18003 (copy_insn_1): Likewise.
18004 * rtl.c (copy_rtx): Likewise.
18005
c2a939b1
GJL
180062013-03-11 Georg-Johann Lay <avr@gjlay.de>
18007
18008 PR target/56591
18009 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
18010 output_operand_lossage message.
88b97037 18011
3c3279fb
RE
180122013-03-11 Richard Earnshaw <rearnsha@arm.com>
18013
18014 PR target/56470
18015 * arm.c (shift_op): Validate RTL pattern on the fly.
18016 (arm_print_operand, case 'S'): Don't use shift_operator to validate
18017 the RTL.
18018
aef5ef9d
JDA
180192013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18020
18021 PR target/56347
18022 * config/pa/pa.md (call_value): Check for calls to powf and direct to
18023 new call patterns that clobber %fr12.
18024 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
18025 split and postreload patterns.
18026 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
18027 registers %fr12 and %fr12R as call used.
18028
456610d3
SB
180292013-03-09 Steven Bosscher <steven@gcc.gnu.org>
18030
18031 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
18032 (canon_address, record_store, replace_read, check_mem_read_rtx,
18033 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
18034 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
18035 rest_of_handle_dse): Likewise.
18036
4b1baac8
RS
180372013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
18038
18039 PR middle-end/56524
18040 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
18041 Add base_optabs.
18042 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
18043 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
18044 (save_optabs_if_changed): Replace with...
18045 (init_tree_optimization_optabs): ...this.
18046 * optabs.c (save_optabs_if_changed): Rename to...
18047 (init_tree_optimization_optabs): ...this. Take the optimization node
18048 as argument. Do nothing if the base optabs are already correct.
18049 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
18050 to recompute optabs.
18051 * function.h (function): Remove optabs field.
18052 * function.c (invoke_set_current_function_hook): Call
18053 init_tree_optimization_optabs. Use the result to initialize
18054 this_fn_optabs.
18055
b7a78683
AH
180562013-02-27 Aldy Hernandez <aldyh@redhat.com>
18057
18058 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
18059 if GTMA_HAS_NO_INSTRUMENTATION.
18060 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
18061 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
18062 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
18063 * gimple-pretty-print.c (dump_gimple_transaction): Handle
18064 GTMA_HAS_NO_INSTRUMENTATION.
18065
6384c29b
JJ
180662013-03-08 Jakub Jelinek <jakub@redhat.com>
18067
18068 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
18069 libasan_preinit.o.
18070
ecd4f20a
MP
180712013-03-08 Marek Polacek <polacek@redhat.com>
18072 Jakub Jelinek <jakub@redhat.com>
18073
18074 PR tree-optimization/56478
18075 * predict.c (is_comparison_with_loop_invariant_p): Change the
18076 type of loop_step to tree.
18077 (predict_loops): Adjust.
18078 (predict_iv_comparison): Perform the computations on double_ints.
18079
64366d35
RB
180802013-03-08 Richard Biener <rguenther@suse.de>
18081
18082 PR tree-optimization/56570
18083 * tree-cfg.c (verify_expr_location_1): Verify locations for
18084 DECL_DEBUG_EXPR.
18085 * tree-sra.c (create_access_replacement): Strip locations
18086 from DECL_DEBUG_EXPRs.
18087
a9d5a059
RB
180882013-03-08 Richard Biener <rguenther@suse.de>
18089
18090 * tree-inline.c (expand_call_inline): Do not associate
18091 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
18092 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
18093
b7aa4e9a
RB
180942013-03-08 Richard Biener <rguenther@suse.de>
18095
18096 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
18097 or block changes with -Og. Fix for location / block encoding
18098 changes and PHI arguments with locations.
18099
c4c2f9fa
SB
181002013-03-07 Steven Bosscher <steven@gcc.gnu.org>
18101
18102 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
18103 for all counters.
18104 (struct output_info): Likewise.
18105 (register_overhead): Remove bad gcc_assert.
18106 (bitmap_find_bit): If there is only a single bitmap element, do not
18107 count a miss as a search.
18108 (print_statistics): Update for counter type changes.
18109 (dump_bitmap_statistics): Likewise. Print headers such that they
18110 are properly lined up with the printed counters.
18111
5bf6606a
JJ
181122013-03-07 Jakub Jelinek <jakub@redhat.com>
18113
18114 PR tree-optimization/56559
18115 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
18116 check that it has only a single use.
18117
2c653d46
RB
181182013-03-07 Richard Biener <rguenther@suse.de>
18119
18120 * doc/invoke.texi (fwhole-program): Discourage use in combination
18121 with -flto.
18122
a72d8780
JJ
181232013-03-06 Jakub Jelinek <jakub@redhat.com>
18124
01a454df
JJ
18125 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
18126
c1781047
JJ
18127 PR tree-optimization/56539
18128 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
18129 instead of GSI_CONTINUE_LINKING as last argument to
18130 force_gimple_operand_gsi. Adjust function comment.
18131
9772c47a
JJ
18132 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
18133 aarch64-cores.def.
18134
a72d8780
JJ
18135 PR middle-end/56548
18136 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
18137 promoted mode, convert the result back to the original mode.
18138
fa5556de
RB
181392013-03-06 Richard Biener <rguenther@suse.de>
18140
18141 PR middle-end/56294
18142 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
18143 (insert_updated_phi_nodes_compare_uids): New function.
18144 (update_ssa): Sort symbols_to_rename after UID before
18145 traversing it to insert PHI nodes.
18146
010403d1
RB
181472013-03-06 Richard Biener <rguenther@suse.de>
18148
18149 PR middle-end/50494
18150 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
18151 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
18152
18153 Revert
18154 2013-02-13 Richard Biener <rguenther@suse.de>
18155
18156 PR lto/50494
18157 * varasm.c (output_constant_def_1): Get the decl representing
18158 the constant as argument.
18159 (output_constant_def): Wrap output_constant_def_1.
18160 (make_decl_rtl): Use output_constant_def_1 with the decl
18161 representing the constant.
18162 (build_constant_desc): Optionally re-use a decl already
18163 representing the constant.
18164 (tree_output_constant_def): Adjust.
18165
3c27ce4c
JY
181662013-03-06 Joey Ye <joey.ye@arm.com>
18167
18168 PR lto/50293
18169 * gcc.c (convert_white_space): New function.
18170 (main): Handles white space in function name.
18171
8f6d1c86
OE
181722013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
18173
18174 PR target/56529
18175 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
18176 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
18177 to SH_DIV_CALL_TABLE for TARGET_SH2.
18178 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
18179 list.
18180 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
18181 call-table options.
18182
798a209f
SA
181832013-03-05 Sterling Augustine <saugustine@google.com>
18184 Cary Coutant <ccoutant@google.com>
18185
18186 PR debug/55364
18187 * dwarf2out.c (resolve_addr): Don't call
18188 remove_loc_list_addr_table_entries a second time for the same
18189 expression.
18190
6cfa417f 181912013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 18192
6cfa417f
JJ
18193 PR debug/56510
18194 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
18195 (avoid_complex_debug_insns): New function.
18196 (expand_debug_locations): Call it.
18197
18198 PR rtl-optimization/56484
18199 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
18200 lifetimes of hard registers on small register class machines.
18201
182022013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
18203
18204 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 18205 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 18206 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH
18207 fast_interrupt.
18208 (microblaze_fast_interrupt_function_p): New function.
18209 (microblaze_is_interrupt_handler): Rename to
18210 microblaze_is_interrupt_variant and add fast_interrupt check.
18211 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
18212 (save_restore_insns): Likewise.
18213 (compute_frame_size): Likewise.
18214 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
18215 (microblaze_globalize_label): Likewise.
6cfa417f
JJ
18216 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
18217 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH
18218 microblaze_is_interrupt_variant.
18219
731aefac
KT
182202013-03-05 Kai Tietz <ktietz@redhat.com>
18221
18222 * sdbout.c (sdbout_one_type): Switch to current function's section
18223 supporting cold/hot.
18224
a72d8780 182252013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH
18226
18227 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
18228 -mxl-reorder.
18229
334e71e8
JJ
182302013-03-05 Jakub Jelinek <jakub@redhat.com>
18231
0b50e654
JJ
18232 PR middle-end/56461
18233 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
18234 if VALGRIND_GET_VBITS is defined, temporarily make object
18235 memory all defined, and restore previous valgrind addressability
18236 and definability afterwards. Free this_object at the end.
18237
4ccf8f43
JJ
18238 PR middle-end/56461
18239 * lra.c (lra): Call lra_clear_live_ranges if live_p,
18240 right before calling lra_create_live_ranges, also call it
18241 when clearing live_p. Only call lra_clear_live_ranges
18242 at the end if live_p.
18243
334e71e8
JJ
18244 PR middle-end/56461
18245 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
18246
9ca966ca
RB
182472013-03-05 Richard Biener <rguenther@suse.de>
18248
18249 PR tree-optimization/56521
18250 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
18251 value-id.
18252
d273b176
SB
182532013-03-05 Steven Bosscher <steven@gcc.gnu.org>
18254
18255 PR c++/55135
18256 * except.h (remove_unreachable_eh_regions): New prototype.
18257 * except.c (remove_eh_handler_splicer): New function, split out
18258 of remove_eh_handler.
18259 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
18260 warning about running it on many EH regions one at a time.
18261 (remove_unreachable_eh_regions_worker): New function, walk the
18262 EH tree in depth-first order and remove non-marked regions.
18263 (remove_unreachable_eh_regions): New function.
18264 * tree-eh.c (mark_reachable_handlers): New function, split out
18265 from remove_unreachable_handlers.
18266 (remove_unreachable_handlers): Use mark_reachable_handlers and
18267 remove_unreachable_eh_regions.
18268 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
18269 and remove_unreachable_eh_regions.
18270
5e7f1aef
RB
182712013-03-05 Richard Biener <rguenther@suse.de>
18272
18273 PR middle-end/56525
18274 * loop-init.c (fix_loop_structure): Remove loops in two stages,
18275 not freeing them until the end.
18276
f276b762
AK
182772013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18278
18279 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
18280
12a43ab8
RB
182812013-03-05 Richard Biener <rguenther@suse.de>
18282
18283 PR tree-optimization/56270
18284 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
18285 of loads after scheduling an SLP instance.
18286
93675444
JJ
182872013-03-05 Jakub Jelinek <jakub@redhat.com>
18288
db4138e3
JJ
18289 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
18290 tic6x.exp.
18291 (check_gcc_parallelize): Run guality.exp as a separate job from
18292 vect.exp with unsorted.exp and $(dg_target_exps) separately from
18293 struct-layout-1.exp with stackalign.exp.
18294
dd3d1ec0
JJ
18295 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
18296
b8d381a3
JJ
18297 PR middle-end/56461
18298 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
18299 load_index sbitmap even if some bit in it isn't set.
18300
b4f9786b
JJ
18301 PR middle-end/56461
18302 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
18303 (discover_iteration_bound_by_body_walk): Change queues to
18304 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
18305 spelling in comment. Call safe_push on queues[bound_index] directly.
18306 Release queues[queue_index] in every iteration unconditionally.
18307 Release bounds vector.
18308
93675444
JJ
18309 PR middle-end/56461
18310 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
18311 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
18312 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
18313 inner_phis vector.
18314
3e492e9c
RB
183152013-03-05 Richard Biener <rguenther@suse.de>
18316
18317 PR lto/56515
18318 * tree-inline.c (remap_blocks_to_null): New function.
18319 (expand_call_inline): When expanding a call stmt without
18320 an associated block inline remap all callee blocks to NULL.
18321
a5d1569a
JJ
183222013-03-05 Jakub Jelinek <jakub@redhat.com>
18323
86efb5cd
JJ
18324 PR rtl-optimization/56494
18325 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
18326 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
18327 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
18328
85f5dbea
JJ
18329 PR middle-end/56461
18330 * sel-sched-ir.c (free_sched_pools): Release
18331 succs_info_pool.stack[succs_info_pool.max_top] vectors too
18332 if succs_info_pool.max_top isn't -1.
18333
a5d1569a
JJ
18334 PR bootstrap/56509
18335 * opts.c (opts_obstack, opts_concat): Moved to...
18336 * opts-common.c (opts_obstack, opts_concat): ... here.
18337
4432aa6c
JJ
183382013-03-04 Jakub Jelinek <jakub@redhat.com>
18339
18340 PR middle-end/56461
18341 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
18342
f3a1fb91
MJ
183432013-03-04 Martin Jambor <mjambor@suse.cz>
18344
18345 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
18346 all appropriate places.
18347
22110e6c
EB
183482013-01-04 Eric Botcazou <ebotcazou@adacore.com>
18349
18350 PR tree-optimization/56424
18351 * ipa-split.c (split_function): Do not set the RSO flag if result is
18352 not by reference and its type is a register type.
18353
a72d8780 183542013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 18355
88b97037 18356 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 18357 (microblaze_legitimate_pic_operand): Likewise
88b97037 18358 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH
18359 new function microblaze_legitimate_pic_operand
18360 * config/microblaze/microblaze-protos.h
18361 (microblaze_legitimate_pic_operand): Declare.
18362
a72d8780 183632013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 18364
a72d8780 18365 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 18366 New predicate for supported rtx code types.
a72d8780 18367 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI
18368 call_insn_simple_operand predicate.
18369
541d9ac8
JJ
183702013-03-04 Jakub Jelinek <jakub@redhat.com>
18371
5eb010bc
JJ
18372 PR middle-end/56461
18373 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
18374 partitions.ordered_remove.
18375
30862efc
JJ
18376 PR middle-end/56461
18377 * tree-vect-stmts.c (vectorizable_conversion): Don't call
18378 vec_oprnds0.create (1) for modifier == NONE.
18379
8930f723
JJ
18380 PR middle-end/56461
18381 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
18382 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
18383 vec_oprnds1 right before pushing anything to it for
18384 scalar_shift_arg.
18385
541d9ac8
JJ
18386 PR middle-end/56461
18387 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
18388 set nbbs to 0 instead of having separate code path.
18389 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
18390 instead of false as last argument if returning NULL.
18391
7aa7f2e3
SL
183922013-03-03 Sandra Loosemore <sandra@codesourcery.com>
18393
18394 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
18395 the attribute is now called "target" instead of "option".
18396 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
18397 * doc/tm.texi.in (Target Attributes): Likewise document the correct
18398 attribute/pragma name for TARGET_OPTION_VALID_P and
18399 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
18400 * doc/tm.texi: Regenerated.
18401
8930f723 184022013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 18403
8930f723 18404 * config/microblaze/microblaze.c:
cb8a1637 18405 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ
18406 * config/microblaze/microblaze.h: Add -mxl-reorder to
18407 DRIVER_SELF_SPECS.
18408 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
18409 instructions emitted if TARGET_REORDER.
88b97037
UB
18410 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
18411 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 18412 separately.
cb8a1637 18413
4dc1d68c
XDL
184142013-03-01 Xinliang David Li <davidxl@google.com>
18415
18416 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
18417 walk length.
18418
689eaba3
JJ
184192013-03-01 Jakub Jelinek <jakub@redhat.com>
18420
9d676bf6
JJ
18421 PR middle-end/56461
18422 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
18423 vector even when returning true. Fix up function comment formatting.
18424
caff7edf
JJ
18425 PR middle-end/56461
18426 * ira-build.c (ira_loop_nodes_count): New variable.
18427 (create_loop_tree_nodes): Initialize it.
18428 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
18429
b6b9227d
JJ
18430 PR middle-end/56461
18431 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
18432 method on dr_chain and result_chain.
18433 * tree-vect-stmts.c (vectorizable_store): Only call
18434 result_chain.create if j == 0.
18435
689eaba3
JJ
18436 PR middle-end/56461
18437 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
18438 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
18439 before overwriting it.
18440
40bf31ed
TB
184412013-03-01 Tobias Burnus <burnus@net-b.de>
18442
18443 * doc/extended.texi (C Extensions): Change order in @menu
18444 to match @node.
18445 (Other MIPS Built-in Functions): Move last MIPS entry before
18446 "picoChip Built-in Functions".
18447 (SH Built-in Functions): Move after RX Built-in Functions.
18448 * doc/gcc.texi (Introduction): Change order in @menu
18449 to match @node.
18450 * doc/md.texi (Constraints): Ditto.
18451 * gty.texi (Type Information): Ditto.
18452 (User-provided marking routines for template types): Make
18453 subsection.
18454 * doc/invoke.texi (AArch64 Options): Move before
18455 "Adapteva Epiphany Options".
18456
e664c61c
KS
184572013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
18458 Jakub Jelinek <jakub@redhat.com>
18459
18460 PR sanitizer/56454
18461 * asan.c (gate_asan): Lookup no_sanitize_address instead of
18462 no_address_safety_analysis attribute.
18463 * doc/extend.texi (no_address_safety_attribute): Rename to
18464 no_sanitize_address attribute, mention no_address_safety_analysis
18465 attribute as deprecated alias.
18466
37b5ec8f
JJ
184672013-02-28 Jakub Jelinek <jakub@redhat.com>
18468
18469 PR middle-end/56461
18470 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
18471 type to vec<vec<tree> > *.
18472 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
18473 to be vec<tree> instead of vec<tree> *, set vec_defs
18474 to vNULL and call vec_defs.create (number_of_vects), adjust other
18475 uses of vec_defs.
18476 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
18477 vectorizable_condition): Adjust vect_get_slp_defs callers.
18478
ba96cdfb
JG
184792013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18480
18481 * config/aarch64/aarch64.c
18482 (aarch64_float_const_representable): Remove unused variable.
18483
6f549691
JG
184842013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18485
18486 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
18487
af55e82d
JG
184882013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18489
18490 * config/aarch64/aarch64-builtins.c
18491 (aarch64_init_simd_builtins): Make static.
18492
1df3f464
JG
184932013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
18494
18495 * config/aarch64/aarch64.c
18496 (aarch64_simd_make_constant): Make static.
18497
f8f42513
MJ
184982013-02-28 Martin Jambor <mjambor@suse.cz>
18499
18500 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
18501 with no initialization to the RHS of debug statements.
18502
b48b3fc4
MJ
185032013-02-28 Martin Jambor <mjambor@suse.cz>
18504
18505 PR tree-optimization/56294
18506 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
18507 Adjust dumping.
18508 (get_access_replacement): Do not call create_access_replacement.
18509 Assert a replacement exists.
18510 (get_repl_default_def_ssa_name): Create the replacement declaration
18511 itself.
18512
c3ae224c
RR
185132013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
18514
18515 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
18516 final_end_function.
18517
45fa0eef
MP
185182013-02-28 Marek Polacek <polacek@redhat.com>
18519
18520 PR rtl-optimization/56466
18521 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
18522 if we're changing a loop.
18523 (peel_loops_completely): Likewise.
18524
502c067d
PC
185252013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
18526
18527 PR c++/55813
18528 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
18529
ff24afc8
GJL
185302013-02-28 Georg-Johann Lay <avr@gjlay.de>
18531
18532 PR target/56445
18533 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
18534 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
18535 INTX_FTYPE_FX, FX_FTYPE_INTX.
18536 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
18537
c7b1fc1b
GJL
185382013-02-28 Georg-Johann Lay <avr@gjlay.de>
18539
18540 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
18541 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
18542 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
18543 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
18544 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
18545 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
18546 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
18547 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
18548 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
18549 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
18550 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
18551 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
18552 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
18553 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
18554 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
18555 (avrxmega6): Increase max flash segments from 5 to 6.
18556 * config/avr/t-multilib: Regenerate.
18557 * config/avr/avr-tables.opt: Regenerate.
18558 * doc/avr-mmcu.texi: Regenerate.
18559
4a0e3cfe
GJL
185602013-02-28 Georg-Johann Lay <avr@gjlay.de>
18561
18562 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
18563 (avr_device_to_arch): Rename to avr_device_to_ld.
18564 (avr_device_to_as): New prototype.
18565 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
18566 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
18567 * config/avr/driver-avr.c (avr_device_to_as): New.
18568 (avr_device_to_arch): Rename to avr_device_to_ld.
18569
97785e52
JJ
185702013-02-27 Jakub Jelinek <jakub@redhat.com>
18571
3f292312
JJ
18572 PR middle-end/56461
18573 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
18574 method on dr_chain and result_chain.
18575
a344216b
JJ
18576 PR middle-end/56461
18577 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
18578 pointer_set_destroy on not_executed_last_iteration.
18579
f121ad02 18580 PR middle-end/56461
88b97037 18581 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 18582
307f83a3
JJ
18583 PR middle-end/56461
18584 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
18585 FOR_EACH_DEFINED_FUNCTION when freeing state.
18586
e19624ee
JJ
18587 PR middle-end/56461
18588 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
18589 pool_free.
18590 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
18591 overwriting it.
18592
90e709fd
JJ
18593 PR middle-end/56461
18594 * ipa-cp.c (decide_whether_version_node): Call vec_free on
18595 known_aggs[i].items and release known_aggs vector.
18596
97785e52
JJ
18597 PR middle-end/56461
18598 * ipa-reference.c (propagate): Free node_info even for alias nodes.
18599
186002013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 18601
97785e52
JJ
18602 * config/microblaze/microblaze.c (microblaze_emit_compare):
18603 Use xor for EQ/NE comparisions.
18604 * config/microblaze/microblaze.md (cstoresf4): Add constraints
18605 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 18606
6fa5e0ed
JJ
186072013-02-27 Jakub Jelinek <jakub@redhat.com>
18608
18609 PR middle-end/56461
18610 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
18611 vector.
18612 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
18613 vec_safe_push, always update *slot.
18614 (redirect_edge_var_map_clear): Use vec_free.
18615 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
18616 (free_var_map_entry): Use vec_free.
18617 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
18618 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
18619
436a956a
AB
186202013-02-27 Andrey Belevantsev <abel@ispras.ru>
18621
18622 PR middle-end/45472
436a956a
AB
18623 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
18624 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB
18625 Reorder tests for speculativeness in the logical and operator.
18626
f5c2caca
JJ
186272013-02-27 Jakub Jelinek <jakub@redhat.com>
18628
0fcb564b
JJ
18629 * incpath.c (add_standard_paths): Use reconcat instead of concat
18630 where appropriate and avoid leaking memory.
18631
dc357798
JJ
18632 * opts.h: Include obstack.h.
18633 (opts_concat): New prototype.
18634 (opts_obstack): New declaration.
18635 * opts.c (opts_concat): New function.
18636 (opts_obstack): New variable.
18637 (init_options_struct): Call gcc_init_obstack on opts_obstack.
18638 (finish_options): Use opts_concat instead of concat
18639 and XOBNEWVEC instead of XNEWVEC.
18640 * opts-common.c (generate_canonical_option, decode_cmdline_option,
18641 generate_option): Likewise.
18642 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
18643 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
18644
f5c2caca
JJ
18645 PR target/56455
18646 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
18647 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
18648
d0163673
JJ
186492013-02-26 Jakub Jelinek <jakub@redhat.com>
18650
18651 PR middle-end/56461
18652 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
18653
b5ad2b8e
JR
186542013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18655
18656 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
18657 (arm_block_move_unaligned_straight): Likewise.
18658 (arm_adjust_block_mem): Likewise.
18659
9b639e2c
JR
186602013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
18661
18662 PR target/48901
18663 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
18664 temp, cond and label.
18665 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
18666
e13a0ccb
JR
18667 PR target/52500
18668 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
18669 * config/c6x/c6x.h (dbx_register_map): Update declaration.
18670
fbe4f171
JR
18671 PR target/52501
18672 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
18673 of prologue/epilogue functions.
18674
ae006543
JR
18675 PR target/52550
18676 * config/tilegx/tilegx.c (tilegx_expand_prologue):
18677 Remove unused variable cfa_offset.
18678 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
18679
c81369fa
JR
18680 PR target/54639
18681 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
18682 type promotion to unsigned.
18683
f8a8fea7
JR
18684 PR target/54640
18685 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
18686 for HOST_WIDE_INT of 32 bit / same size as int.
18687 (arm_block_move_unaligned_straight): Likewise.
18688 (arm_adjust_block_mem): Likewise.
18689
f8be5169
JR
18690 PR target/54662
18691 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
18692 ALL_CFLAGS.
18693
f1ad3354
MP
186942013-02-26 Marek Polacek <polacek@redhat.com>
18695
18696 PR tree-optimization/56426
88b97037 18697 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 18698
a6af0f42
RB
186992013-02-26 Richard Biener <rguenther@suse.de>
18700
18701 PR target/56444
18702 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
18703 unused variable loops.
18704
aca43c6c
JJ
187052013-02-26 Jakub Jelinek <jakub@redhat.com>
18706
cecbe5d9
JJ
18707 PR tree-optimization/56448
18708 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
18709 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
18710 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
18711 later operands of the references, or even first operand for
18712 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
18713
aca43c6c
JJ
18714 PR tree-optimization/56443
18715 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
18716 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
18717 to type_for_mode langhook.
18718
53e2e141
MT
187192013-02-25 Matt Turner <mattst88@gmail.com>
18720
18721 * doc/invoke.texi: Document r4700.
18722
259ee451
RB
187232013-02-25 Richard Biener <rguenther@suse.de>
18724
18725 PR tree-optimization/56175
18726 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
18727 split out from ...
18728 (simplify_bitwise_binary): ... here. Also guard the conversion
18729 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
18730
6d65e8f1
CM
187312013-02-25 Catherine Moore <clm@codesourcery.com>
18732
18733 Revert:
18734 2013-02-24 Catherine Moore <clm@codesourcery.com>
18735 Maciej W. Rozycki <macro@codesourcery.com>
18736 Tom de Vries <tom@codesourcery.com>
a72d8780 18737 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM
18738 Iain Sandoe <iain@codesourcery.com>
18739 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 18740 Chao-ying Fu <fu@mips.com>
88b97037 18741
6d65e8f1 18742 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 18743 Document new function attributes.
6d65e8f1
CM
18744 * doc/invoke.texi (minterlink-compressed, mmicromips,
18745 m14k, m14ke, m14kec): Document new options.
18746 (minterlink-mips16): Update documentation.
18747 * doc/md.texi (ZC, ZD): Document new constraints.
18748 * configure.ac (gcc_cv_as_micromips): Check if linker
18749 supports the .set micromips directive.
18750 * configure: Regenerate.
18751 * config.in: Regenerate.
18752 * config/mips/mips-tables.opt: Regenerate.
18753 * config/mips/micromips.md: New file.
18754 * constraints.md (ZC, AD): New constraints.
18755 * config/mips/predicates.md (movep_src_register): New predicate.
18756 (movep_src_operand): New predicate.
18757 (non_volatile_mem_operand): New predicate.
18758 * config/mips/mips.md (multimem): New type.
18759 (length): Differentiate between 17-bit and 18-bit branch offsets.
18760 (MOVEP1, MOVEP2): New mode iterator.
18761 (mov_<load>l): Use ZC constraint.
18762 (mov_<load>r): Likewise.
18763 (mov_<store>l): Likewise.
18764 (mov_<store>r): Likewise.
18765 (*branch_equality<mode>_inverted): Add microMIPS support.
18766 (*branch_equality<mode>): Likewise.
18767 (*jump_absolute): Likewise.
18768 (indirect_jump_<mode>): Likewise.
18769 (tablejump_<mode>): Likewise.
18770 (<optab>_internal): Likewise.
18771 (sibcall_internal): Likewise.
18772 (sibcall_value_internal): Likewise.
18773 (prefetch): Use constraint ZD.
18774 * config/mips/mips.opt (minterlink-compressed): New option.
18775 (minterlink-mips16): Now an alias for minterlink-compressed.
18776 (mmicromips): New option.
18777 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18778 (compare_and_swap_12): Likewise.
18779 (sync_add<mode>): Likewise.
18780 (sync_<optab>_12): Likewise.
18781 (sync_old_<optab>_12): Likewise.
18782 (sync_new_<optab>_12): Likewise.
18783 (sync_nand_12): Likewise.
18784 (sync_old_nand_12): Likewise.
18785 (sync_new_nand_12): Likewise.
18786 (sync_sub<mode>): Likewise.
18787 (sync_old_add<mode>): Likewise.
18788 (sync_old_sub<mode>): Likewise.
18789 (sync_new_add<mode>): Likewise.
18790 (sync_new_sub<mode>): Likewise.
18791 (sync_<optab><mode>): Likewise.
18792 (sync_old_<optab><mode>): Likewise.
18793 (sync_new_<optab><mode>): Likewise.
18794 (sync_nand<mode>): Likewise.
18795 (sync_old_nand<mode>): Likewise.
18796 (sync_new_nand<mode>): Likewise.
18797 (sync_lock_test_and_set<mode>): Likewise.
18798 (test_and_set_12): Likewise.
18799 (atomic_compare_and_swap<mode>): Likewise.
18800 (atomic_exchange<mode>_llsc): Likewise.
18801 (atomic_fetch_add<mode>_llsc): Likewise.
18802 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18803 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18804 (umips_save_restore_pattern_p): Likewise.
18805 (umips_load_store_pair_p): Likewise.
18806 (umips_output_load_store_pair): Likewise.
18807 (umips_movep_target_p): Likewise.
18808 (umips_12bit_offset_address_p): Likewise.
18809 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18810 (mips_base_mips16): Rename this...
18811 (mips_base_compression_flags): ...to this. Update all uses.
18812 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18813 (mips_mips16_decl_p): Delete.
18814 (mips_nomips16_decl_p): Delete.
18815 (mips_get_compress_on_flags): New function.
18816 (mips_get_compress_off_flags): New function.
18817 (mips_get_compress_mode): New function.
18818 (mips_get_compress_on_name): New function.
18819 (mips_get_compress_off_name): New function.
18820 (mips_insert_attributes): Support multiple compression types.
18821 (mips_merge_decl_attributes): Likewise.
18822 (umips_12bit_offset_address_p): New function.
18823 (mips_start_function_definition): Emit .set micromips directive.
18824 (mips_call_may_need_jalx_p): New function.
18825 (mips_function_ok_for_sibcall): Add microMIPS support.
18826 (mips_print_operand_punctuation): Support short delay slots and
18827 compact jumps.
18828 (umips_swm_mask, umips_swm_encoding): New.
18829 (umips_build_save_restore): New function.
18830 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18831 (was_mips16_p): Remove.
18832 (old_compression_mode): New.
18833 (mips_set_compression_mode): New function.
18834 (mips_set_current_function): Add microMIPS support.
18835 (mips_option_override): Likewise.
18836 (umips_save_restore_pattern_p): New function.
18837 (umips_output_save_restore): New function.
18838 (umips_load_store_pair_p_1): New function.
18839 (umips_load_store_pair_p): New function.
18840 (umips_output_load_store_pair_1): New function.
18841 (umips_output_load_store_pair): New function.
18842 (umips_movep_target_p) New function.
18843 (mips_prepare_pch_save): Add microMIPS support.
18844 * config/mips/mips.h (TARGET_COMPRESSION): New.
18845 (TARGET_CPU_CPP_BUILTINS): Update macro
18846 to use new compression flags and to support microMIPS.
18847 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18848 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18849 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
18850 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
18851 (ASM_SPEC): Support mmicromips and mno-micromips.
18852 (M16STORE_REG_P): New macro.
18853 (MIPS_CALL): Support TARGET_MICROMIPS.
18854 (MICROMIPS_J): New macro.
18855 (mips_base_mips16): Rename this...
18856 (mips_base_compression_flags): ...to this.
18857 (UMIPS_12BIT_OFFSET_P): New macro.
18858 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
18859 (MULTILIB_DIRNAMES): Likewise.
18860
5e5df392
TV
188612013-02-25 Tom de Vries <tom@codesourcery.com>
18862
18863 PR rtl-optimization/56131
18864 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
18865 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
18866 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
18867
3fdb53c1
TB
188682013-02-25 Tobias Burnus <burnus@net-b.de>
18869
18870 * doc/invoke.texi (-fsanitize=): Move from optimization
18871 to debugging options.
18872
ed358aea
AB
188732013-02-25 Andrey Belevantsev <abel@ispras.ru>
18874
18875 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
18876
f45e9053
AB
188772013-02-25 Andrey Belevantsev <abel@ispras.ru>
18878 Alexander Monakov <amonakov@ispras.ru>
18879
18880 PR middle-end/56077
18881 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
18882 flush pending lists also on non-jumps. Adjust comment.
18883
6941b508
CM
188842013-02-24 Catherine Moore <clm@codesourcery.com>
18885 Maciej W. Rozycki <macro@codesourcery.com>
18886 Tom de Vries <tom@codesourcery.com>
a72d8780 18887 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM
18888 Iain Sandoe <iain@codesourcery.com>
18889 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 18890 Chao-ying Fu <fu@mips.com>
6941b508
CM
18891
18892 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 18893 Document new function attributes.
6941b508
CM
18894 * doc/invoke.texi (minterlink-compressed, mmicromips,
18895 m14k, m14ke, m14kec): Document new options.
18896 (minterlink-mips16): Update documentation.
18897 * doc/md.texi (ZC, ZD): Document new constraints.
18898 * configure.ac (gcc_cv_as_micromips): Check if linker
18899 supports the .set micromips directive.
18900 * configure: Regenerate.
18901 * config.in: Regenerate.
18902 * config/mips/mips-tables.opt: Regenerate.
18903 * config/mips/micromips.md: New file.
18904 * constraints.md (ZC, AD): New constraints.
18905 * config/mips/predicates.md (movep_src_register): New predicate.
18906 (movep_src_operand): New predicate.
18907 (non_volatile_mem_operand): New predicate.
18908 * config/mips/mips.md (multimem): New type.
18909 (length): Differentiate between 17-bit and 18-bit branch offsets.
18910 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 18911 (mov_<load>l): Use ZC constraint.
6941b508
CM
18912 (mov_<load>r): Likewise.
18913 (mov_<store>l): Likewise.
18914 (mov_<store>r): Likewise.
18915 (*branch_equality<mode>_inverted): Add microMIPS support.
18916 (*branch_equality<mode>): Likewise.
18917 (*jump_absolute): Likewise.
18918 (indirect_jump_<mode>): Likewise.
18919 (tablejump_<mode>): Likewise.
18920 (<optab>_internal): Likewise.
18921 (sibcall_internal): Likewise.
18922 (sibcall_value_internal): Likewise.
18923 (prefetch): Use constraint ZD.
18924 * config/mips/mips.opt (minterlink-compressed): New option.
18925 (minterlink-mips16): Now an alias for minterlink-compressed.
18926 (mmicromips): New option.
18927 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
18928 (compare_and_swap_12): Likewise.
18929 (sync_add<mode>): Likewise.
18930 (sync_<optab>_12): Likewise.
18931 (sync_old_<optab>_12): Likewise.
18932 (sync_new_<optab>_12): Likewise.
18933 (sync_nand_12): Likewise.
18934 (sync_old_nand_12): Likewise.
18935 (sync_new_nand_12): Likewise.
18936 (sync_sub<mode>): Likewise.
18937 (sync_old_add<mode>): Likewise.
18938 (sync_old_sub<mode>): Likewise.
18939 (sync_new_add<mode>): Likewise.
18940 (sync_new_sub<mode>): Likewise.
18941 (sync_<optab><mode>): Likewise.
18942 (sync_old_<optab><mode>): Likewise.
18943 (sync_new_<optab><mode>): Likewise.
18944 (sync_nand<mode>): Likewise.
18945 (sync_old_nand<mode>): Likewise.
18946 (sync_new_nand<mode>): Likewise.
18947 (sync_lock_test_and_set<mode>): Likewise.
18948 (test_and_set_12): Likewise.
18949 (atomic_compare_and_swap<mode>): Likewise.
18950 (atomic_exchange<mode>_llsc): Likewise.
18951 (atomic_fetch_add<mode>_llsc): Likewise.
18952 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
18953 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
18954 (umips_save_restore_pattern_p): Likewise.
18955 (umips_load_store_pair_p): Likewise.
18956 (umips_output_load_store_pair): Likewise.
18957 (umips_movep_target_p): Likewise.
18958 (umips_12bit_offset_address_p): Likewise.
18959 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
18960 (mips_base_mips16): Rename this...
18961 (mips_base_compression_flags): ...to this. Update all uses.
18962 (mips_attribute_table): Add micromips, nomicromips and nocompression.
18963 (mips_mips16_decl_p): Delete.
18964 (mips_nomips16_decl_p): Delete.
3fdb53c1 18965 (mips_get_compress_on_flags): New function.
6941b508
CM
18966 (mips_get_compress_off_flags): New function.
18967 (mips_get_compress_mode): New function.
18968 (mips_get_compress_on_name): New function.
18969 (mips_get_compress_off_name): New function.
18970 (mips_insert_attributes): Support multiple compression types.
18971 (mips_merge_decl_attributes): Likewise.
18972 (umips_12bit_offset_address_p): New function.
18973 (mips_start_function_definition): Emit .set micromips directive.
18974 (mips_call_may_need_jalx_p): New function.
18975 (mips_function_ok_for_sibcall): Add microMIPS support.
18976 (mips_print_operand_punctuation): Support short delay slots and
18977 compact jumps.
18978 (umips_swm_mask, umips_swm_encoding): New.
18979 (umips_build_save_restore): New function.
18980 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
18981 (was_mips16_p): Remove.
18982 (old_compression_mode): New.
18983 (mips_set_compression_mode): New function.
18984 (mips_set_current_function): Add microMIPS support.
18985 (mips_option_override): Likewise.
18986 (umips_save_restore_pattern_p): New function.
18987 (umips_output_save_restore): New function.
18988 (umips_load_store_pair_p_1): New function.
18989 (umips_load_store_pair_p): New function.
18990 (umips_output_load_store_pair_1): New function.
18991 (umips_output_load_store_pair): New function.
18992 (umips_movep_target_p) New function.
18993 (mips_prepare_pch_save): Add microMIPS support.
18994 * config/mips/mips.h (TARGET_COMPRESSION): New.
18995 (TARGET_CPU_CPP_BUILTINS): Update macro
18996 to use new compression flags and to support microMIPS.
18997 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
18998 (MIPS_ARCH_FLOAT_SPEC): Likewise.
18999 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
19000 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
19001 (ASM_SPEC): Support mmicromips and mno-micromips.
19002 (M16STORE_REG_P): New macro.
19003 (MIPS_CALL): Support TARGET_MICROMIPS.
19004 (MICROMIPS_J): New macro.
19005 (mips_base_mips16): Rename this...
19006 (mips_base_compression_flags): ...to this.
19007 (UMIPS_12BIT_OFFSET_P): New macro.
19008 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
19009 (MULTILIB_DIRNAMES): Likewise.
19010
cdebbc6d
JJ
190112013-02-24 Jakub Jelinek <jakub@redhat.com>
19012
19013 PR target/52555
19014 * target-globals.c (save_target_globals): For init_reg_sets and
19015 target_reinit remporarily set this_fn_optabs to this_target_optabs.
19016
18c63565
JG
190172013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
19018
19019 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
19020 * config/aarch64/t-aarch64
19021 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
19022
2194f7a2
VM
190232013-02-22 Vladimir Makarov <vmakarov@redhat.com>
19024
19025 PR inline-asm/56148
19026 * lra-constraints.c (process_alt_operands): Reload operand
19027 conflicting with earlier clobber only if no more other conflicting
19028 operands.
19029
7d613735
JJ
190302013-02-22 Jakub Jelinek <jakub@redhat.com>
19031
19032 PR sanitizer/56393
19033 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
19034 if not linking a shared library.
19035
ac8d93a7
SL
190362013-02-22 Seth LaForge <sethml@google.com>
19037
19038 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
19039
e0237780
GY
190402013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
19041
3fdb53c1
TB
19042 * config/arm/arm.md (split for extendsidi): Update condition.
19043 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
19044 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
19045 (qhs_zextenddi_cstr): Likewise.
e0237780 19046
d7fde18c
JJ
190472013-02-21 Jakub Jelinek <jakub@redhat.com>
19048
be63b77d
JJ
19049 PR middle-end/56420
19050 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
19051 avoid signed wrapping.
19052 (expand_mult): Handle properly multiplication by
19053 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
19054 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
19055 in the compiler if coeff is HOST_WIDE_INT_MIN.
19056 (expand_divmod): Don't make ext_op1 static, change it's type to
19057 uhwi. Avoid undefined behavior in -INTVAL (op1).
19058
d7fde18c
JJ
19059 PR rtl-optimization/50339
19060 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
19061 field.
19062 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
19063 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
19064 into splitting_ashiftrt field.
19065 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
19066 ASHIFTRT.
19067 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
19068 choices.
19069
6aad4455
AH
190702013-02-20 Aldy Hernandez <aldyh@redhat.com>
19071
19072 PR middle-end/56108
19073 * trans-mem.c (execute_tm_mark): Do not expand transactions that
19074 are sure to go irrevocable.
19075
38fe784d
HPN
190762013-02-21 Hans-Peter Nilsson <hp@axis.com>
19077
19078 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
19079 scalars are valid operands.
19080
0fd44da3
MJ
190812013-02-21 Martin Jambor <mjambor@suse.cz>
19082
19083 PR tree-optimization/56310
19084 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
19085 only matching indices and non-negative final offsets.
19086 (intersect_aggregates_with_edge): Pass src_idx to
19087 agg_replacements_to_vector. Pass src_idx insstead of index to
19088 intersect_with_agg_replacements.
19089
7a92038b
MJ
190902013-02-21 Martin Jambor <mjambor@suse.cz>
19091
19092 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
19093 instead of hard-wired defaults.
19094
c0da9c37
MR
190952013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
19096
19097 * doc/invoke.texi (MIPS Options): Update documentation of the
19098 floating-point multiply-accumulate instruction restrictions.
19099
d247ea0c 191002013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS
19101
19102 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
19103 asan_shadow_offset on x86_64 linux.
19104
22deefcb
RB
191052013-02-21 Richard Biener <rguenther@suse.de>
19106
19107 PR tree-optimization/56415
19108 Revert
19109 2013-02-11 Richard Biener <rguenther@suse.de>
19110
19111 PR tree-optimization/56273
19112 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19113 first VRP run.
19114
7bcc6e75
JJ
191152013-02-21 Jakub Jelinek <jakub@redhat.com>
19116
a0ad148f
JJ
19117 PR bootstrap/56258
19118 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
19119 instead of @itemx.
19120
7bcc6e75
JJ
19121 PR inline-asm/56405
19122 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
19123 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
19124
a0a7b611
JH
191252013-02-20 Jan Hubicka <jh@suse.cz>
19126
19127 PR tree-optimization/56265
88b97037
UB
19128 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
19129 when target is referenced for first time.
a0a7b611 19130
c0e50f72
RB
191312013-02-20 Richard Biener <rguenther@suse.de>
19132
19133 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
19134 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
19135 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
19136 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
19137 not return anything.
19138 (rename_ssa_copies): Do not remove unused locals.
19139 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 19140 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB
19141 * passes.c (execute_function_todo): Do not schedule unused locals
19142 removal if cleanup_tree_cfg did something.
19143 * tree-ssa-live.c (remove_unused_locals): Dump statistics
19144 about the number of removed locals.
19145
a52ca739
RB
191462013-02-20 Richard Biener <rguenther@suse.de>
19147
19148 PR tree-optimization/56398
88b97037 19149 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 19150
ec9c9d1b
MJ
191512013-02-20 Martin Jambor <mjambor@suse.cz>
19152
19153 PR tree-optimization/55334
19154 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
19155 restricted pointers to arrays.
19156
e91c8ed6 191572013-02-20 Richard Biener <rguenther@suse.de>
88b97037 19158 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB
19159
19160 PR tree-optimization/56396
19161 * tree-ssa-ccp.c (n_const_val): New static variable.
19162 (get_value): Return NULL for SSA names we don't have a lattice
19163 entry for.
19164 (ccp_initialize): Initialize n_const_val.
19165 * tree-ssa-copy.c (n_copy_of): New static variable.
19166 (init_copy_prop): Initialize n_copy_of.
19167 (get_value): Return NULL_TREE for SSA names we don't have a
19168 lattice entry for.
19169
3d916479
MJ
191702013-02-20 Martin Jambor <mjambor@suse.cz>
19171
19172 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
19173
71a86758
RB
191742013-02-20 Richard Biener <rguenther@suse.de>
19175
19176 * genpreds.c (write_lookup_constraint): Do not compare first
19177 letter of the constraint again.
19178
79836a12
RB
191792013-02-20 Richard Biener <rguenther@suse.de>
19180
19181 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
19182 and ceil_log2.
19183 (get_use_iv_cost): Terminate hashtable walk when coming across
19184 an empty entry.
19185
bbe4fb2c
IZ
191862013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
19187
19188 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
19189 reassociation for avx2 targets.
19190
a72d8780 191912012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 19192
c77f83d5
EI
19193 * config/microblaze/microblaze.c: microblaze_has_clz = 0
19194 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 19195 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI
19196 version and TARGET_PATTERN_COMPARE check
19197 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 19198
a72d8780 191992012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 19200
a72d8780 19201 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI
19202 function before branching.
19203
242387fa
AB
192042012-02-19 Andrey Belevantsev <abel@ispras.ru>
19205
19206 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
19207 DUMP_INSN_RTX_UID.
19208 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
19209
b4979ab9
AB
192102012-02-19 Andrey Belevantsev <abel@ispras.ru>
19211
19212 PR middle-end/55889
b4979ab9
AB
19213 * sel-sched.c: Include ira.h.
19214 (implicit_clobber_conflict_p): New function.
19215 (moveup_expr): Use it.
88b97037 19216 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 19217
24d63016
RB
192182013-02-19 Richard Biener <rguenther@suse.de>
19219
19220 PR tree-optimization/56384
19221 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
19222 (vn_hash_type): Split out from ...
19223 (vn_hash_constant_with_type): ... here.
19224 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
19225 (vn_phi_eq): Compare types from vn_phi_s structure.
19226 (vn_phi_lookup): Populate vn_phi_s type.
19227 (vn_phi_insert): Likewise.
19228
a475fd3d
JJ
192292013-02-19 Jakub Jelinek <jakub@redhat.com>
19230
47cc28f5
JJ
19231 PR tree-optimization/56350
19232 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
19233 if haven't found reduction or nested cycle operand, rather than
19234 asserting we must find it.
19235
a475fd3d
JJ
19236 PR tree-optimization/56381
19237 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
19238 to fold_build3.
19239
192402013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH
19241 Jakub Jelinek <jakub@redhat.com>
19242
19243 PR target/52555
19244 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
19245 (swap_optab_enable): Same.
19246 (init_all_optabs): Use argument instead of global.
88b97037 19247 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH
19248 * expr.h (init_all_optabs): Add argument to prototype.
19249 (TREE_OPTIMIZATION_OPTABS): New.
19250 (save_optabs_if_changed): Protoize.
19251 * optabs.h: Declare this_fn_optabs.
19252 * optabs.c (save_optabs_if_changed): New.
19253 Declare this_fn_optabs.
19254 (init_optabs): Add argument to init_all_optabs() call.
19255 * function.c (invoke_set_current_function_hook): Handle per
19256 function optabs.
19257 * function.h (struct function): New field optabs.
19258 * config/mips/mips.c (mips_set_mips16_mode): Handle when
19259 optimization_current_node has changed.
19260 * target-globals.h (save_target_globals_default_opts): Protoize.
19261 * target-globals.c (save_target_globals_default_opts): New.
19262
3f587ca3
JDA
192632013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19264
19265 PR target/56347
19266 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
19267 registers %fr12 and %fr12R as call used.
ceaca33e
JDA
19268
19269 PR target/56214
19270 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB
19271 and HImode, require all displacements to be an integer multiple of
19272 their mode size.
ceaca33e
JDA
19273 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
19274 only allow QImode and HImode when reload is in progress and strict is
19275 true. Likewise for symbolic addresses. Use base14_operand to check
19276 displacements in REG+BASE addresses.
19277
fe0b4796
RB
192782013-02-18 Richard Biener <rguenther@suse.de>
19279
19280 PR tree-optimization/56366
19281 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19282 handle sign-conversion of outer-loop initial induction value.
19283
6aaf596b
RB
192842013-02-18 Richard Biener <rguenther@suse.de>
19285
73db8ff1 19286 PR middle-end/56349
6aaf596b
RB
19287 * cfghooks.c (merge_blocks): If we merge a latch into another
19288 block adjust references to it.
19289 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
19290 (verify_loop_structure): Verify that a recorded latch is in fact
19291 a latch.
19292
5e97dfb6
RB
192932013-02-18 Richard Biener <rguenther@suse.de>
19294
19295 PR tree-optimization/56321
19296 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
19297 order SSA name release and virtual operand unlinking.
19298
825527e8
EI
192992013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19300
19301 * config/microblaze/microblaze.md (save_stack_block): Define.
19302 (restore_stack_block): Likewise.
19303
debd11d9
EI
193042013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19305
19306 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
19307 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
19308 * config/microblaze/microblaze.c (microblaze_option_override):
19309 Bail out early for PIC modes when target does not support PIC.
19310
8ec77be0
EI
193112013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19312
88b97037 19313 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI
19314 Replace with a microblaze version.
19315 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 19316 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI
19317 microblaze.
19318
b41288b3
JJ
193192013-02-16 Jakub Jelinek <jakub@redhat.com>
19320 Dodji Seketeli <dodji@redhat.com>
19321
19322 PR asan/56330
88b97037 19323 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ
19324 (instrument_mem_region_access): Do not forget to always put
19325 instrumentation of the of 'base' and 'base + len' in a "if (len !=
19326 0) statement, even for cases where either 'base' or 'base + len'
19327 are not instrumented -- because they have been previously
19328 instrumented. Simplify the logic by putting all the statements
19329 instrument 'base + len' inside a sequence, and then insert that
19330 sequence right before the current insertion point. Then, to
19331 instrument 'base + len', just get an iterator on that statement.
19332 And do not forget to update the pointer to iterator the function
19333 received as argument.
19334
47918951
VM
193352013-02-15 Vladimir Makarov <vmakarov@redhat.com>
19336
19337 PR rtl-optimization/56348
19338 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
19339
3fb7c699
SB
193402013-02-15 Steven Bosscher <steven@gcc.gnu.org>
19341
19342 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
19343 (clean_graph_dump_file): Pass base to start_graph_dump.
19344
576fe41a
RH
193452013-02-14 Richard Henderson <rth@redhat.com>
19346
19347 PR target/55941
19348 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
19349
bb3accfa
SB
193502013-02-14 Steven Bosscher <steven@gcc.gnu.org>
19351
19352 * collect2-aix.h: Define F_LOADONLY.
19353
03143140
RB
193542013-02-14 Richard Biener <rguenther@suse.de>
19355
19356 PR lto/50494
19357 * varasm.c (output_constant_def_1): Get the decl representing
19358 the constant as argument.
19359 (output_constant_def): Wrap output_constant_def_1.
19360 (make_decl_rtl): Use output_constant_def_1 with the decl
19361 representing the constant.
19362 (build_constant_desc): Optionally re-use a decl already
19363 representing the constant.
19364 (tree_output_constant_def): Adjust.
19365
8fb06726
DS
193662013-02-14 Dodji Seketeli <dodji@redhat.com>
19367
19368 Fix an asan crash
19369 * asan.c (instrument_builtin_call): Really put the length of the
19370 second source argument into src1_len.
19371
c4bfe8bf
JJ
193722013-02-13 Jakub Jelinek <jakub@redhat.com>
19373
19374 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
19375 argument. If it is false, don't create edge from then_bb to
19376 fallthru_bb.
19377 (insert_if_then_before_iter): Pass true to it.
19378 (build_check_stmt): Pass false to it.
19379 (transform_statements): Flush hash table only on extended basic
19380 block boundaries, rather than at the beginning of every bb.
19381 Don't flush hash table on nonfreeing_call_p calls.
19382 * tree-flow.h (nonfreeing_call_p): New prototype.
19383 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
19384
7afe2801
DM
193852013-02-13 David S. Miller <davem@davemloft.net>
19386
19387 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
19388
70cc3288
VM
193892013-02-13 Vladimir Makarov <vmakarov@redhat.com>
19390
19391 PR target/56184
19392 * ira.c (max_regno_before_ira): Move from ...
19393 (ira): ... here.
19394 (fix_reg_equiv_init): Use max_regno_before_ira instead of
19395 vec_safe_length.
19396
6422242b
JJ
193972013-02-13 Jakub Jelinek <jakub@redhat.com>
19398
19399 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
19400
976a81ee
RB
194012013-02-13 Richard Biener <rguenther@suse.de>
19402
19403 PR lto/56295
19404 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
19405 globals in MEM_REFs.
19406
c1874a87
RB
194072013-02-13 Richard Biener <rguenther@suse.de>
19408
19409 * loop-init.c (loop_optimizer_init): Clear loop state when
19410 re-initializing preserved loops.
19411 * loop-unswitch.c (unswitch_single_loop): Return whether
19412 we unswitched the loop. Do not verify loop state here.
88b97037 19413 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 19414
b4ab7d34
KS
194152013-02-13 Kostya Serebryany <kcc@google.com>
19416
88b97037
UB
19417 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
19418 on x86_64 linux.
b4ab7d34
KS
19419 * sanitizer.def: Rename __asan_init to __asan_init_v1.
19420
bdcbe80c
DS
194212013-02-12 Dodji Seketeli <dodji@redhat.com>
19422
19423 Avoid instrumenting duplicated memory access in the same basic block
19424 * Makefile.in (asan.o): Add new dependency on hash-table.h
19425 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
19426 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
19427 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
19428 (free_mem_ref_resources, has_mem_ref_been_instrumented)
19429 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
19430 (get_mem_ref_of_assignment): New functions.
19431 (get_mem_refs_of_builtin_call): Extract from
19432 instrument_builtin_call and tweak a little bit to make it fit with
19433 the new signature.
19434 (instrument_builtin_call): Use the new
19435 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
19436 of is_gimple_builtin_call.
19437 (instrument_derefs, instrument_mem_region_access): Insert the
19438 instrumented memory reference into the hash table.
19439 (maybe_instrument_assignment): Renamed instrument_assignment into
19440 this, and change it to advance the iterator when instrumentation
19441 actually happened and return true in that case. This makes it
19442 homogeneous with maybe_instrument_assignment, and thus give a
19443 chance to callers to be more 'regular'.
19444 (transform_statements): Clear the memory reference hash table
19445 whenever we enter a new BB, when we cross a function call, or when
19446 we are done transforming statements. Use
19447 maybe_instrument_assignment instead of instrumentation. No more
19448 need to special case maybe_instrument_assignment and advance the
19449 iterator after calling it; it's now handled just like
19450 maybe_instrument_call. Update comment.
19451
4861a1f7
RB
194522013-02-13 Richard Biener <rguenther@suse.de>
19453
19454 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
19455 Fix loop discovery code.
19456
1c86bd80
VM
194572013-02-12 Vladimir Makarov <vmakarov@redhat.com>
19458
19459 PR inline-asm/56148
19460 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 19461 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM
19462 the operand is not reloaded. Prefer to reload conflicting operand
19463 if earlyclobber and matching operands are the same.
19464
7cbda518
RB
194652013-02-12 Richard Biener <rguenther@suse.de>
19466
19467 PR lto/56297
19468 * lto-streamer-out.c (write_symbol): Do not output symbols
19469 for hard register variables.
19470
e68a4ef6
GJL
194712013-02-12 Georg-Johann Lay <avr@gjlay.de>
19472
19473 PR target/54222
19474 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
19475 (umulsidi3_insn, mulsidi3_insn): New insns.
19476
a72d8780 194772013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL
19478
19479 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
19480 (struct tune_params): Add vec_costs field.
19481 * config/arm/arm.c (arm_builtin_vectorization_cost)
19482 (arm_add_stmt_cost): New functions.
19483 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
19484 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
19485 (arm_default_vec_cost): New struct of type cpu_vec_costs.
19486 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
19487 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
19488 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
19489 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
19490
43320568
RB
194912013-02-12 Richard Biener <rguenther@suse.de>
19492
19493 PR lto/56295
19494 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
19495 decls again if possible.
19496
a011aa39
RB
194972013-02-12 Richard Biener <rguenther@suse.de>
19498
19499 PR middle-end/56288
19500 * tree-ssa.c (verify_ssa_name): Fix check, move
19501 SSA_NAME_IN_FREE_LIST check up.
19502
6da26889
JJ
195032013-02-12 Jakub Jelinek <jakub@redhat.com>
19504 Steven Bosscher <steven@gcc.gnu.org>
19505
19506 PR rtl-optimization/56151
19507 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
19508 equal to op0 or op1, and last_insn pattern is CODE operation
19509 with MEM dest and one of the operands matches that MEM.
19510
f80e0faf
ST
195112013-02-11 Sriraman Tallam <tmsriramgoogle.com>
19512
19513 * doc/extend.texi: Document Function Multiversioning and "default"
19514 parameter string to target attribute.
19515 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
19516 target attribute parameter is "default".
19517 (ix86_compare_version_priority): Remove checks for target attribute.
19518 (ix86_mangle_function_version_assembler_name): Change error to sorry.
19519 Remove check for target attribute equal to NULL. Add assert.
19520 (ix86_generate_version_dispatcher_body): Change error to sorry.
19521
6c59ffd1
IS
195222013-02-11 Iain Sandoe <iain@codesourcery.com>
19523 Jack Howarth <howarth@bromo.med.uc.edu>
19524 Patrick Marlier <patrick.marlier@gmail.com>
19525
19526 PR libitm/55693
19527 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
19528 define ENDFILE_SPEC as TM_DESTRUCTOR.
19529 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
19530
a72d8780 195312013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP
19532 Jack Howarth <howarth@bromo.med.uc.edu>
19533 Jakub Jelinek <jakub@redhat.com>
19534
19535 PR sanitizer/55617
19536 * config/darwin.c (cdtor_record): Rename ctor_record.
19537 (sort_cdtor_records): Rename sort_ctor_records.
19538 (finalize_dtors): New routine to sort destructors by
19539 priority before use in assemble_integer.
19540 (machopic_asm_out_destructor): Use finalize_dtors if needed.
19541
b63fe007
UB
195422013-02-11 Uros Bizjak <ubizjak@gmail.com>
19543
19544 PR rtl-optimization/56275
19545 * simplify-rtx.c (avoid_constant_pool_reference): Check that
19546 offset is non-negative and less than cmode size before
19547 calling simplify_subreg.
19548
8e89b5b5
RB
195492013-02-11 Richard Biener <rguenther@suse.de>
19550
19551 PR tree-optimization/56264
19552 * cfgloop.h (fix_loop_structure): Adjust prototype.
19553 * loop-init.c (fix_loop_structure): Return the number of
19554 newly discovered loops.
19555 * tree-cfgcleanup.c (repair_loop_structures): When new loops
19556 are discovered, do a full loop-closed SSA rewrite.
19557
b4a4b56d
RB
195582013-02-11 Richard Biener <rguenther@suse.de>
19559
19560 PR tree-optimization/56273
19561 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
19562 first VRP run.
19563 (check_array_ref): Fix missing newline in dumps.
19564 (search_for_addr_array): Likewise.
19565
0c885229
DE
195662013-02-09 David Edelsohn <dje.gcc@gmail.com>
19567
19568 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
19569
59ac9a55
JJ
195702013-02-09 Jakub Jelinek <jakub@redhat.com>
19571
19572 PR target/56256
19573 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
19574
25bb0bb5
VM
195752013-02-08 Vladimir Makarov <vmakarov@redhat.com>
19576
19577 PR rtl-optimization/56246
0c885229 19578 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM
19579 reload pseudo.
19580 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
19581 constraints are satisfied.
19582
a698cc03
JL
195832013-02-08 Jeff Law <law@redhat.com>
19584
19585 PR debug/53948
19586 * emit-rtl.c (reg_is_parm_p): New function.
19587 * regs.h (reg_is_parm_p): New prototype.
19588 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
19589 callee-clobbered registers.
19590
e1122ddd
MM
195912013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
19592
19593 PR target/56043
19594 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
19595 If there is no implicit builtin declaration, just return NULL.
19596
19c5f6e6
UB
195972013-02-08 Uros Bizjak <ubizjak@gmail.com>
19598
19599 * config/i386/sse.md (FMAMODEM): New mode iterator.
19600 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
19601 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
19602
2480f2ca 196032013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 19604
2480f2ca
UB
19605 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
19606 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
19607 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
19608
196092013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
19610
19611 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
19612 (microblaze*-*-elf): Likewise.
19613 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
19614 LINK_SPEC.
19615 * config/microblaze/microblaze-c.c: Add builtin defines for
19616 _LITTLE_ENDIAN and _BIG_ENDIAN.
19617 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
19618 add to TARGET_DEFAULT flags.
76ef61fb 19619 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB
19620 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
19621 * config/microblaze/microblaze.md: Update extendsidi2 and
19622 movdi_internal instructions to use low-order / high-order reg
19623 print_operands.
19624 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
19625 options and inversemask / mask of LITTLE_ENDIAN.
19626 * config/microblaze/t-microblaze: Expand multilib options to
19627 include mlittle-endian (le) and update exceptions patterns.
19628
600a5961
JJ
196292013-02-08 Jakub Jelinek <jakub@redhat.com>
19630
5df81313
JJ
19631 PR rtl-optimization/56195
19632 * lra-constraints.c (get_reload_reg): Don't reuse regs
19633 if they have smaller mode than requested, if they have
19634 wider mode than requested, try to return a SUBREG.
19635
600a5961
JJ
19636 PR tree-optimization/56250
19637 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
19638 if type is unsigned and code isn't MULT_EXPR.
19639
ff544649
GJL
196402013-02-08 Georg-Johann Lay <avr@gjlay.de>
19641
19642 PR tree-optimization/56064
19643 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
19644 bits according to mode.
19645 * fixed-value.h (fixed_from_double_int)
19646 (const_fixed_from_double_int): Adjust comments.
19647
e45cde98
RB
196482013-02-08 Richard Biener <rguenther@suse.de>
19649
19650 PR lto/56231
19651 * lto-streamer.h (struct data_in): Remove current_file, current_line
19652 and current_col members.
19653 * lto-streamer-out.c (lto_output_location): Stream changed bits
19654 en-block for efficiency.
19655 * lto-streamer-in.c (clear_line_info): Remove.
19656 (lto_input_location): Cache current file, line and column
19657 globally via local statics. Read changed bits en-block.
19658 (input_function): Do not call clear_line_info.
19659 (lto_read_body): Likewise.
19660 (lto_input_toplevel_asms): Likewise.
19661
c1ca73d8
MM
196622013-02-08 Michael Matz <matz@suse.de>
19663
19664 PR tree-optimization/52448
19665 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
19666 (nt_call_phase): New static.
19667 (add_or_mark_expr): Only mark accesses with newer phase than any
19668 call seen.
19669 (nonfreeing_call_p): New.
19670 (nt_init_block): Update nt_call_phase, mark blocks as visited.
19671 (nt_fini_block): Keep blocks marked as visited.
19672 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
19673
57548aa2
RB
196742013-02-08 Richard Biener <rguenther@suse.de>
19675
19676 * ira.c (ira): Free broken dominator information.
19677
8e10366f
UB
196782013-02-08 Uros Bizjak <ubizjak@gmail.com>
19679
19680 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
19681
f64fb0fa
MP
196822013-02-08 Marek Polacek <polacek@redhat.com>
19683
8e10366f 19684 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 19685
0375167b
RB
196862013-02-08 Richard Biener <rguenther@suse.de>
19687
19688 PR middle-end/56181
19689 * cfgloop.h (flow_loops_find): Adjust.
19690 (bb_loop_header_p): Declare.
19691 * cfgloop.c (bb_loop_header_p): New function split out from ...
19692 (flow_loops_find): ... here. Adjust function signature,
19693 support incremental loop structure update.
19694 (verify_loop_structure): Cleanup. Verify a loop is a loop.
19695 * cfgloopmanip.c (fix_loop_structure): Move ...
19696 * loop-init.c (fix_loop_structure): ... here.
19697 (apply_loop_flags): Split out from ...
19698 (loop_optimizer_init): ... here.
19699 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
19700 in incremental mode, only remove dead loops here.
19701
85d768f3
GJL
197022013-02-08 Georg-Johann Lay <avr@gjlay.de>
19703
19704 PR target/54222
19705 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
19706 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
19707 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
19708 (*round<mode>3.libgcc): New insns for fixed-modes.
19709 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
19710 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
19711 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
19712 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
19713 implementations. Define to __builtin_avr_absFX,
19714 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
19715 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
19716 __builtin_avr_countlsFX, respectively.
19717 * config/avr/avr-c.c (target.h): Include it.
19718 (enum avr_builtin_id): New enum.
19719 (avr_resolve_overloaded_builtin): New static function.
19720 (avr_register_target_pragmas): Use it to set
19721 targetm.resolve_overloaded_builtin.
19722 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
19723 tree nodes used by DEF_BUILTIN.
19724 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
19725 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
19726 <AVR_BUILTIN_xxBITS>: Same.
19727
661bc682
RB
197282013-02-08 Richard Biener <rguenther@suse.de>
19729
19730 * cfgloop.c (verify_loop_structure): Properly handle
19731 a loop exiting to another loop header.
19732 * ira-int.h (ira_loops): Remove.
19733 * ira.c (ira_loops): Remove.
19734 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
19735 (do_reload): Use loop_optimizer_finalize.
19736 * ira-build.c (create_loop_tree_nodes): Use get_loops and
19737 number_of_loops to access the loop tree.
19738 (more_one_region_p): Likewise.
19739 (finish_loop_tree_nodes): Likewise.
19740 (rebuild_regno_allocno_maps): Likewise.
19741 (mark_loops_for_removal): Likewise.
19742 (mark_all_loops_for_removal): Likewise.
19743 (remove_unnecessary_regions): Likewise.
19744 (ira_build): Likewise.
19745 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19746
0d5049b2
RB
197472013-02-08 Richard Biener <rguenther@suse.de>
19748
19749 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
19750 * ipa-pure-const.c (analyze_function): Avoid calling
19751 mark_irreducible_loops twice.
8e10366f 19752 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 19753
499f32e8
DM
197542013-02-07 David S. Miller <davem@davemloft.net>
19755
19756 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
19757 on 'reg'.
19758 * var-tracking.c (vt_add_function_parameter): Test the presence of
19759 HAVE_window_save properly and do not remap argument registers when
19760 we have a leaf function.
19761
6edc3e32
UB
197622013-02-07 Uros Bizjak <ubizjak@gmail.com>
19763
19764 PR bootstrap/56227
19765 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
19766 instead of "ll".
19767 * config/i386/i386.c (ix86_print_operand): Ditto.
19768
5306401f
VM
197692013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19770
6edc3e32 19771 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 19772
027ece11
VM
197732013-02-07 Vladimir Makarov <vmakarov@redhat.com>
19774
19775 PR rtl-optimization/56225
19776 * lra-constraints.c (process_alt_operands): Check that reload hard
19777 reg can hold value for strict_low_part.
19778
f980dfdb
JJ
197792013-02-07 Jakub Jelinek <jakub@redhat.com>
19780
19781 PR debug/56154
19782 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
19783 dwarf2out_end_function.
19784 (in_first_function_p, maybe_at_text_label_p,
19785 first_loclabel_num_not_at_text_label): New variables.
19786 (dwarf2out_var_location): In the first function find out
6edc3e32 19787 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ
19788 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
19789 functions.
19790
73dd3123
EB
197912013-02-07 Eric Botcazou <ebotcazou@adacore.com>
19792
19793 PR rtl-optimization/56178
19794 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
19795 SUBREG of a register. Tidy up related block of code.
19796 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
19797 note if the source is a register or a SUBREG of a register.
19798
e3936f47
JJ
197992013-02-07 Jakub Jelinek <jakub@redhat.com>
19800
19801 PR target/56228
19802 * config/rs6000/rs6000.md (ptrm): New mode attr.
19803 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
19804 call_value_indirect_aix<pttrsize>,
19805 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
19806 m in constraints.
19807
d96d674b
MH
198082013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
19809
19810 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
19811 if -bnortl. Convert to strcmp and strncmp.
19812
921f2dee
AM
198132013-02-07 Alan Modra <amodra@gmail.com>
19814
19815 PR target/54009
19816 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
19817 addresses won't wrap when offsetting.
19818 (rs6000_secondary_reload): Provide secondary reloads needed for
19819 wrapping LO_SUM addresses.
19820
d09c7dba
TS
198212013-02-06 Thomas Schwinge <thomas@codesourcery.com>
19822
19823 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
19824 MACH, just __MACH__.
19825
a44bbd48
RB
198262013-02-06 Richard Biener <rguenther@suse.de>
19827
19828 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
19829 instead of calling fix_loop_structure.
19830
6d840d99
JJ
198312013-02-06 Jakub Jelinek <jakub@redhat.com>
19832
19833 PR middle-end/56217
19834 * omp-low.c (use_pointer_for_field): Return false if
19835 lower_send_shared_vars doesn't generate any copy-out code.
19836
0f33baa9
TV
198372013-02-06 Tom de Vries <tom@codesourcery.com>
19838
19839 PR rtl-optimization/56131
19840 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
19841 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
19842 of the label is NULL. Add comment.
19843
11f1e3ab
JJ
198442013-02-05 Jakub Jelinek <jakub@redhat.com>
19845
31502f9f
JJ
19846 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
19847
a74db9bd
JJ
19848 PR sanitizer/55374
19849 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
19850 (STATIC_LIBTSAN_LIBS): Likewise.
19851 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
19852 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
19853 is defined, don't add anything else beyond that.
19854 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
19855 (LINK_COMMAND_SPEC): Use them.
19856
11f1e3ab
JJ
19857 PR tree-optimization/56205
19858 * tree-stdarg.c (check_all_va_list_escapes): Return true if
19859 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
19860 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
19861
14ac6aa2
RB
198622013-02-05 Richard Biener <rguenther@suse.de>
19863
19864 PR tree-optimization/53342
19865 PR tree-optimization/53185
19866 * tree-vectorizer.h (vect_check_strided_load): Remove.
19867 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
19868 not disallow peeling for vectorized strided loads.
19869 (vect_check_strided_load): Make static and simplify.
19870 (vect_analyze_data_refs): Adjust.
19871 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
19872 correctly when vectorizing strided loads.
19873
6f22445a
RB
198742013-02-05 Richard Biener <rguenther@suse.de>
19875
19876 * doc/install.texi: Refer to ISL, not PPL.
19877
39f9719e
JH
198782013-02-05 Jan Hubicka <jh@suse.cz>
19879
ec4224ac
JH
19880 PR tree-optimization/55789
19881 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
19882
198832013-02-05 Jan Hubicka <jh@suse.cz>
19884
19885 PR tree-optimization/55789
39f9719e
JH
19886 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
19887 the dead call anyway.
19888
956d3b33
EB
198892013-02-05 Eric Botcazou <ebotcazou@adacore.com>
19890
19891 PR sanitizer/55374
19892 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
19893
198942013-02-04 Alexander Potapenko <glider@google.com>
19895 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP
19896 Jakub Jelinek <jakub@redhat.com>
19897
19898 PR sanitizer/55617
19899 * config/darwin.c (sort_ctor_records): Stabilized qsort
19900 on constructor priority by using original position.
19901 (finalize_ctors): New routine to sort constructors by
19902 priority before use in assemble_integer.
19903 (machopic_asm_out_constructor): Use finalize_ctors if needed.
19904
7ac3af38
JJ
199052013-02-04 Jakub Jelinek <jakub@redhat.com>
19906
19907 PR libstdc++/54314
19908 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
19909 about visibility on artificial decls.
19910 * config/sol2.c (solaris_assemble_visibility): Likewise.
19911
152689dc
KT
199122013-02-04 Kai Tietz <ktietz@redhat.com>
19913
19914 PR target/56186
19915 * config/i386/i386.c (function_value_ms_64): Add additional valtype
19916 argument and improve checking of return-argument types for 16-byte
19917 modes.
19918 (ix86_function_value_1): Add additional valtype argument on call
19919 of function_value_64.
19920 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
19921 handling infunction_value_64 function.
19922
2b5987b5
MGD
199232013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
19924
7ac3af38 19925 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 19926
14597080
RB
199272013-02-04 Richard Biener <rguenther@suse.de>
19928
19929 PR tree-optimization/56188
19930 * tree-ssa-structalias.c (label_visit): Consider case with
19931 initially non-empty points-to set.
19932 (perform_var_substitution): Dump node mapping and clean up.
19933
ed73881e
RG
199342013-02-04 Richard Guenther <rguenther@suse.de>
19935
19936 PR lto/56168
19937 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
19938 node prevail as last resort.
19939 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 19940 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 19941
139a0707
RB
199422013-02-04 Richard Biener <rguenther@suse.de>
19943
19944 PR tree-optimization/56113
19945 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
19946 Merge into ...
19947 (equiv_class_lookup_or_add): ... this.
19948 (label_visit): Adjust and fix error in previous patch.
19949 (perform_var_substitution): Adjust.
19950
50fe8924
OE
199512013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
19952
19953 * config/sh/divtab.c: Fix formatting and comments throughout the file.
19954 * config/sh/sh4-300.md: Likewise.
19955 * config/sh/sh4a.md: Likewise.
19956 * config/sh/constraints.md: Likewise.
19957 * config/sh/sh.md: Likewise.
19958 * config/sh/netbsd-elf.h: Likewise.
19959 * config/sh/predicates.md: Likewise.
19960 * config/sh/sh-protos.h: Likewise.
19961 * config/sh/ushmedia.h: Likewise.
19962 * config/sh/linux.h: Likewise.
19963 * config/sh/sh.c: Likewise.
19964 * config/sh/superh.h: Likewise.
19965 * config/sh/elf.h: Likewise.
19966 * config/sh/sh4.md: Likewise.
19967 * config/sh/sh.h: Likewise.
19968
1a04ac2b
JDA
199692013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
19970
19971 * config/pa/constraints.md: Adjust unused letters. Change "T"
19972 constraint to match_test floating_point_store_memory_operand().
19973 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
19974 (base14_operand): New.
19975 (floating_point_store_memory_operand): New.
19976 (integer_store_memory_operand): Revise to use base14_operand and
19977 reg_plus_base_memory_operand.
19978 (move_dest_operand): Allow symbolic_memory_operands.
19979 (symbolic_memory_operand): Check for LO_SOM.
19980 (symbolic_operand): Change default case to break.
19981 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
19982 CONST_DOUBLE values to be reloaded by putting them into memory when
19983 the destination is a floating point register.
19984 (movdf): Remove code to handle CONST_DOUBLE.
19985 (movsf): Likewise.
19986 (reload_indf_r1): New.
19987 (reload_insf_r1): New.
19988 Consistently use "Q" and "T" constraints with integer and floating
19989 point move instructions, respectively.
19990 (movdi): Remove FAIL.
19991 Change predicate for source operand unamed DImode move from
19992 general_operand to move_src_operand.
19993 (umulsidi3): Change predicate for destination operand to
19994 register_operand.
19995 Likewise for similar unamed patterns.
19996 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
19997 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
19998 (hppa_legitimize_address): Simplify mask calculation.
19999 (pa_emit_move_sequence): Revised handling of secondary reloads from
20000 REG+D addresses for floating point loads and stores. Directly handle
20001 loading CONST0_RTX (mode) to a floating point register.
20002 (pa_secondary_reload): Handle reloading DF and SFmode constant values
20003 to floating point registers. Don't restrict secondary reloads to
20004 floating point registers to integer modes. Revise some comments and
20005 cleanup some code.
20006 (TARGET_LEGITIMATE_ADDRESS_P): Define.
20007 (pa_legitimate_address_p): New.
20008 (pa_legitimize_reload_address): New.
20009 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
20010 (STRICT_REG_OK_FOR_BASE_P): New.
20011 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
20012 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
20013
3bf65591
DE
200142013-02-03 David Edelsohn <dje.gcc@gmail.com>
20015 Andrew Dixie <andrewd@gentrack.com>
20016
20017 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
20018 flag set.
20019
c0a8a3e6
RS
200202013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
20021
20022 * expmed.c (extract_bit_field_1): Pass the full width of the
20023 structure to get_best_reg_extraction_insn.
20024
99113dff
DE
200252013-02-01 David Edelsohn <dje.gcc@gmail.com>
20026
20027 PR target/54601
20028 * configure.ac (use_cxa_atexit): Add AIX.
20029 * configure: Regenerate.
20030
20031 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
20032
ddd84654
JJ
200332013-02-01 Jakub Jelinek <jakub@redhat.com>
20034
20035 PR debug/54793
20036 * final.c (need_profile_function): New variable.
20037 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
20038 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
20039 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
20040 notes, targetm.asm_out.function_prologue doesn't emit anything,
20041 HAVE_prologue and profiler should be emitted before prologue,
20042 set need_profile_function instead of emitting it.
20043 (final_scan_insn): If need_profile_function, emit
20044 profile_function on the first NOTE_INSN_BASIC_BLOCK or
20045 NOTE_INSN_FUNCTION_BEG note.
20046
200472013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH
20048
20049 * config/rs6000/rs6000.md (smulditi3): New.
20050 (umulditi3): New.
20051
ff2a9d88
RH
20052 * config/alpha/alpha.md (umulditi3): New.
20053
14d52b90
DE
200542013-02-01 David Edelsohn <dje.gcc@gmail.com>
20055
20056 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
20057 (ASM_OUTPUT_ALIGNED_LOCAL): New.
20058
8c7ca45c
RB
200592013-02-01 Richard Biener <rguenther@suse.de>
20060
20061 PR tree-optimization/56113
20062 * tree-ssa-structalias.c (label_visit): Reduce work for
20063 single-predecessor nodes.
20064
9f419393
EB
200652013-02-01 Eric Botcazou <ebotcazou@adacore.com>
20066
20067 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
20068 range isn't testing for zero.
20069
c5f4be84
SB
200702013-01-31 Steven Bosscher <steven@gcc.gnu.org>
20071
20072 PR middle-end/56113
20073 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
20074
dbdbd982
NC
200752013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
20076 Nick Clifton <nickc@redhat.com>
20077
20078 * config/v850/constraints.md (Q): Define as a memory constraint.
20079 * config/v850/predicates.md (label_ref_operand): New predicate.
20080 (e3v5_shift_operand): New predicate.
20081 (ior_operator): New predicate.
20082 * config/v850/t-v850: Add e3v5 multilib.
20083 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
20084 (v850_gen_movdi): Prototype.
20085 * config/v850/v850.c: Add support for e3v5 architecture.
20086 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
20087 TARGET_V850E_UP.
20088 (construct_save_jarl): Add e3v5 long JARL support.
20089 (v850_adjust_insn_length): New function. Adjust length of call
20090 insns when using e3v5 instructions.
20091 (v850_gen_movdi): New function: Generate instructions to move a
20092 DImode value.
20093 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
20094 (CPP_SPEC): Define __v850e3v5__ as appropriate.
20095 (TARGET_USE_FPU): Enable for e3v5.
20096 (CONST_OK_FOR_W): New macro.
20097 (ADJUST_INSN_LENGTH): Define.
20098 * config/v850/v850.md (UNSPEC_LOOP): Define.
20099 (attr cpu): Add v850e3v5.
20100 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
20101 (movdi): New pattern.
20102 (movdi_internal): New pattern.
20103 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
20104 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
20105 (cstoresf4): Likewise.
20106 (cstoredf4): Likewise.
20107 (insv): New pattern.
20108 (rotlso3_a): New pattern.
20109 (rotlsi3_b): New pattern
20110 (rotlsi3_v850e3v5): New pattern.
20111 (doloop_begin): New pattern.
20112 (fix_loop_counter): New pattern.
20113 (doloop_end): New pattern.
20114 (branch_normal): Add e3v5 long branch support.
20115 (branch_invert): Likewise.
20116 (branch_z_normal): Likewise.
20117 (branch_z_invert): Likewise.
20118 (branch_nz_normal): Likewise.
20119 (branch_nz_invert): Likewise.
20120 (call_internal_short): Add e3v5 register-indirect JARL support.
20121 (call_internal_long): Likewise.
20122 (call_value_internal_short): Likewise.
20123 (call_value_internal_long): Likewise.
20124 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
20125 (mloop): New option.
20126 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 20127 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 20128
73861a41
PK
201292013-01-31 Paul Koning <ni1d@arrl.net>
20130
20131 PR debug/55059
20132 PR debug/54508
20133 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
20134 children if parent is a class.
20135 (prune_unused_types_prune): Don't add DW_AT_declaration.
20136
e44978dc
RB
201372013-01-31 Richard Biener <rguenther@suse.de>
20138
20139 PR tree-optimization/56157
20140 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
20141 match up operand with SLP child.
20142
90b10dec
JM
201432013-01-31 Jason Merrill <jason@redhat.com>
20144
04d2dadd 20145 PR debug/54410
90b10dec
JM
20146 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
20147 parameters the first time.
20148 (gen_scheduled_generic_parms_dies): Check completeness here.
20149
6e616110
RB
201502013-01-31 Richard Biener <rguenther@suse.de>
20151
20152 PR middle-end/53073
20153 * common.opt (faggressive-loop-optimizations): New flag,
20154 enabled by default.
20155 * doc/invoke.texi (faggressive-loop-optimizations): Document.
20156 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
20157 infer_loop_bounds_from_undefined by it.
20158
636f59cf
RB
201592013-01-31 Richard Biener <rguenther@suse.de>
20160
20161 PR tree-optimization/56150
20162 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
20163 visit virtual operands.
20164 (find_uses_to_rename_bb): Likewise.
20165
201662013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB
20167
20168 PR tree-optimization/56150
20169 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
20170 mixed store non-store stmts.
20171
32887460
JJ
201722013-01-30 Jakub Jelinek <jakub@redhat.com>
20173
e60e09a0
JJ
20174 PR sanitizer/55374
20175 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
20176 LIBASAN_EARLY_SPEC is defined.
20177 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
20178 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
20179 before %o.
20180 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
20181
32887460
JJ
20182 PR c++/55742
20183 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
20184 invalid args instead of ICEing on it.
20185 (ix86_valid_target_attribute_tree): Return error_mark_node if
20186 ix86_valid_target_attribute_inner_p failed.
20187 (ix86_valid_target_attribute_p): Return false only if
20188 ix86_valid_target_attribute_tree returned error_mark_node. Allow
20189 target("default") attribute.
20190 (sorted_attr_string): Change argument from const char * to tree,
20191 merge in all target attribute arguments rather than just one.
20192 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
20193 instead of free. Avoid using strcat.
20194 (ix86_mangle_function_version_assembler_name): Mangle
20195 target("default") as if no target attribute is present. Adjust
20196 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
20197 instead of xmalloc and XDELETEVEC instead of free.
20198 (ix86_function_versions): Don't return true if one of the decls
20199 doesn't have target attribute. If they don't and one of the decls
20200 is DECL_FUNCTION_VERSIONED, report an error. Adjust
20201 sorted_attr_string caller. Use XDELETEVEC instead of free.
20202 (ix86_supports_function_versions): Remove.
20203 (make_name): Fix up formatting.
20204 (make_dispatcher_decl): Remove resolver_name and its initialization.
20205 Avoid leaking memory.
20206 (is_function_default_version): Return true if there is
20207 target("default") attribute rather than no target attribute at all.
20208 (make_resolver_func): Avoid leaking memory.
20209 (ix86_generate_version_dispatcher_body): Likewise.
20210 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
20211 * target.def (supports_function_versions): Remove.
20212 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
20213 * doc/tm.texi: Regenerated.
20214
73cca0cc
VM
202152013-01-30 Vladimir Makarov <vmakarov@redhat.com>
20216
20217 PR rtl-optimization/56144
20218 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
20219 for values with side effects.
20220
11452e7b
RB
202212013-01-30 Richard Biener <rguenther@suse.de>
20222
20223 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
20224 (sparseset_pop): Likewise.
20225 * cfganal.c (compute_idf): Likewise. Increase work-stack size
20226 to be able to use quick_push in the worker loop.
20227
01cb1ef5
MP
202282013-01-30 Marek Polacek <polacek@redhat.com>
20229
20230 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
20231
e892936e
RB
202322013-01-30 Richard Biener <rguenther@suse.de>
20233
20234 PR lto/56147
6edc3e32 20235 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 20236
cc06c01d
GJL
202372013-01-30 Georg-Johann Lay <avr@gjlay.de>
20238
20239 PR tree-optimization/56064
20240 * fixed-value.c (fixed_from_double_int): New function.
20241 * fixed-value.h (fixed_from_double_int): New prototype.
20242 (const_fixed_from_double_int): New static inline function.
20243 * fold-const.c (native_interpret_fixed): New static function.
20244 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
20245 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
20246 (native_encode_fixed): New static function.
20247 (native_encode_expr) <FIXED_CST>: Use it.
20248 (native_interpret_int): Move double_int worker code to...
20249 * double-int.c (double_int::from_buffer): ...this new static method.
20250 * double-int.h (double_int::from_buffer): Prototype it.
20251
d394a308
RB
202522013-01-30 Richard Biener <rguenther@suse.de>
20253
20254 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
20255 New pointer-map and obstack.
20256 (init_alias_vars): Allocate pointer-map and obstack.
20257 (delete_points_to_sets): Free them.
20258 (find_what_var_points_to): Cache result.
20259 (find_what_p_points_to): Adjust for changed interface of
20260 find_what_var_points_to.
20261 (compute_points_to_sets): Likewise.
20262 (ipa_pta_execute): Likewise.
20263
20804d96
RO
202642013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20265
20266 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
20267 * configure: Regenerate.
20268 * config.in: Regenerate.
20269 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
20270 #nobits/#progbits if supported.
20271
a7ad88a2
OE
202722013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
20273
20274 PR target/56121
20275 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
20276 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
20277 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
20278
7af79f92
GY
202792013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20280
91bfca59
OE
20281 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
20282 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 20283
753bcf7b
GY
202842013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20285
91bfca59
OE
20286 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
20287 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 20288
8cbc2ea8
GY
202892013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20290
91bfca59
OE
20291 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
20292 declaration.
20293 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
20294 * config/arm/cortex-a7.md: New bypasses using
20295 arm_mac_accumulator_is_result.
8cbc2ea8 20296
697a3325
GY
202972013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20298
91bfca59 20299 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 20300 (cortex_a7_neon_mla): Likewise.
91bfca59
OE
20301 (cortex_a7_fpfmad): New reservation.
20302 (cortex_a7_fpmacs): Use ffmas and update required units.
20303 (cortex_a7_fpmuld): Update required units and latency.
20304 (cortex_a7_fpmacd): Likewise.
20305 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
20306 (cortex_a7_neon). Likewise.
20307 (bypass) Update participating units.
697a3325 20308
29637783
GY
203092013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
20310
91bfca59
OE
20311 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
20312 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
20313 from fmac to ffma.
20314 * config/arm/vfp11.md (vfp_farith): Use ffmas.
20315 (vfp_fmul): Use ffmad.
20316 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
20317 (cortex_r4_fmacd): Use ffmad.
20318 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
20319 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
20320 (cortex_a9_fmacd): Use ffmad.
20321 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
20322 (cortex_a8_vfp_macd): Use ffmad.
20323 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
20324 (cortex_a5_fpmacd): Use ffmad.
20325 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
20326 (cortex_a15_vfp_macd): Use ffmad.
20327 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 20328
2ee8a2d5
JM
203292013-01-29 Jason Merrill <jason@redhat.com>
20330
20331 PR libstdc++/54314
20332 * varasm.c (default_assemble_visibility): Don't warn about
20333 visibility on artificial decls.
20334
78d087bc
RB
203352013-01-29 Richard Biener <rguenther@suse.de>
20336
20337 PR tree-optimization/56113
20338 * tree-ssa-structalias.c (equiv_class_lookup): Also return
20339 the bitmap leader.
20340 (label_visit): Free duplicate bitmaps and record the leader instead.
20341 (perform_var_substitution): Adjust.
20342
83ba4d6f
RB
203432013-01-29 Richard Biener <rguenther@suse.de>
20344
20345 PR tree-optimization/55270
20346 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
20347 the CFG, schedule loops for fixup.
20348
66dfe13f
NC
203492013-01-29 Nick Clifton <nickc@redhat.com>
20350
20351 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
20352 SP_REG.
20353
5a579c3b
LE
203542013-01-28 Leif Ekblad <leif@rdos.net>
20355
20356 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
20357 * config/i386/i386.h (TARGET_RDOS): New macro.
20358 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
20359 * config/i386/i386.c (ix86_option_override_internal): For 64bit
20360 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
20361 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
20362 DEFAULT_LARGE_SECTION_THRESHOLD.
20363 * config/i386/i386.md (R14_REG, R15_REG): New constants.
20364 * config/i386/rdos.h: New file.
20365 * config/i386/rdos64.h: New file.
20366
51e44392
BS
203672013-01-28 Bernd Schmidt <bernds@codesourcery.com>
20368
20369 PR other/54814
20370 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
20371 TEST_HARD_REG_BIT.
20372
db1fb332
JJ
203732013-01-28 Jakub Jelinek <jakub@redhat.com>
20374
20375 PR rtl-optimization/56117
20376 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
20377 call cselib_lookup_from_insn on the MEM before calling
20378 add_insn_mem_dependence.
20379
16917761
RB
203802013-01-28 Richard Biener <rguenther@suse.de>
20381
20382 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
20383 to a stmt that didn't have one.
20384 (copy_phis_for_bb): Likewise for PHI arguments.
20385 (copy_debug_stmt): Likewise for debug stmts.
20386
b9fc0497
RB
203872013-01-28 Richard Biener <rguenther@suse.de>
20388
20389 PR tree-optimization/56034
6edc3e32 20390 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB
20391 (partition_builtin_p): Adjust.
20392 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
20393 it is the last partition.
20394 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
20395 up the vertex for the definition.
20396 (classify_partition): Classify whether a partition is a
20397 PKIND_REDUCTION, thus has uses outside of the loop.
20398 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
20399 Merge all PKIND_REDUCTION partitions into the last partition.
20400 (tree_loop_distribution): Seed partitions from reductions as well.
20401
aa710d25
JJ
204022013-01-28 Jakub Jelinek <jakub@redhat.com>
20403
0bfbca58
JJ
20404 PR tree-optimization/56125
20405 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
20406 pow(x,c) into sqrt(x) * powi(x, n/2) or
20407 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
20408 optimizing for size.
20409 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
20410 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
20411 integer.
20412
aa710d25
JJ
20413 PR tree-optimization/56094
20414 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
20415 to UNKNOWN_LOCATION while gimplifying expr.
20416
77dc5297
UB
204172013-01-27 Uros Bizjak <ubizjak@gmail.com>
20418
20419 PR target/56114
20420 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
20421 operand 0 in movabs insn template for -masm=intel asm alternative.
20422 (*movabs<mode>_2): Ditto for operand 1.
20423
0bfbca58 204242013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH
20425
20426 PR target/54663
20427 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
20428 of microblaze-c.o
20429
0bfbca58 204302013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI
20431
20432 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
20433 tm_file.
20434
0bfbca58 204352013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N
20436
20437 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
20438 Undef to avoid warning.
20439
478f60f9
MH
204402013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
20441
20442 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
20443 * configure: Regenerate.
20444
d7fa6ee2
JJ
204452013-01-25 Jakub Jelinek <jakub@redhat.com>
20446
20447 PR tree-optimization/56098
20448 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
20449 for stmts with volatile ops.
20450 (cond_store_replacement): Don't optimize if assign has volatile ops.
20451 (cond_if_else_store_replacement_1): Don't optimize if either
20452 then_assign or else_assign have volatile ops.
20453 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
20454 volatile ops.
20455
f8fe87bd
GJL
204562013-01-25 Georg-Johann Lay <avr@gjlay.de>
20457
20458 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
20459
28398d0d
GJL
204602013-01-25 Georg-Johann Lay <avr@gjlay.de>
20461
20462 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
20463 missing ':' in asm example.
20464
b7d7d917
TB
204652013-01-25 Tejas Belagod <tejas.belagod@arm.com>
20466
20467 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
20468 entries into lane and laneq entries.
77dc5297
UB
20469 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
20470 Remove AdvSIMD scalar modes.
b7d7d917
TB
20471 (aarch64_sq<r>dmulh_laneq<mode>): New.
20472 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
20473 modes.
20474 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
20475 builtin implementations to relfect changes in RTL in aarch64-simd.md.
20476 * config/aarch64/iterators.md (VCOND): New.
20477 (VCONQ): New.
20478
556f9906
GJL
204792013-01-25 Georg-Johann Lay <avr@gjlay.de>
20480
20481 PR target/54222
20482 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
20483 Add NULL LIBNAME argument to existing definitions.
20484 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
20485 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
20486 * config/avr/avr.c (DEF_BUILTIN): Same.
20487 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
20488 (avr_expand_builtin): Expand to a vanilla call if a libgcc
20489 implementation is available (DECL_ASSEMBLER_NAME is set).
20490 (avr_fold_absfx): New static function.
20491 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
20492 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
20493 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
20494 AVR_BUILTIN_ABSLLK.
20495 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
20496 (abshk, absk, abslk, absllk): Provide as static inline functions.
20497
1f546bbb
MP
204982013-01-25 Marek Polacek <polacek@redhat.com>
20499
20500 PR tree-optimization/56035
20501 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
20502
7e184bd7
UB
205032012-01-24 Uros Bizjak <ubizjak@gmail.com>
20504
20505 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
20506 (*movtf_internal_rex64): Add (!o,C) alternative
20507 (*movxf_internal_rex64): Ditto.
20508 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
20509
f5ba49ac
SM
205102013-01-24 Shenghou Ma <minux.ma@gmail.com>
20511
20512 * doc/invoke.texi: fix typo.
20513 * doc/objc.texi: fix typo.
20514
a6343728
RS
205152013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
20516
20517 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
20518 for the first two alternatives.
20519
cd030c07
DN
205202013-01-24 Diego Novillo <dnovillo@google.com>
20521
77dc5297 20522 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN
20523 (ggc-zone.o): Remove.
20524 * configure.ac: Remove option --with-gc.
20525 * configure: Re-generate.
20526 * doc/install.texi: Remove documentation for --with-gc.
20527 * gengtype.c (write_enum_defn): Remove. Update all users.
20528 (write_Types_process_field): Remove generation of gt_e_* argument.
20529 (output_type_enum): Remove. Update all users.
20530 (write_enum_defn): Remove. Update all users.
20531 (enum alloc_zone): Remove. Update all users.
77dc5297 20532 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN
20533 * ggc-common.c (ggc_splay_alloc): Remove first argument.
20534 Update all callers.
20535 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 20536 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN
20537 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
20538 Update all users.
aaf1e810 20539 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN
20540 (struct alloc_zone): Remove.
20541 (ggc_internal_alloc_zone_stat): Remove.
20542 (ggc_internal_cleared_alloc_zone_stat): Remove.
20543 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB
20544 (ggc_pch_count_object): Remove last argument. Update all users.
20545 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN
20546 (struct alloc_zone): Remove.
20547 * ggc-zone.c: Remove.
77dc5297 20548 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN
20549 (struct alloc_zone): Remove.
20550 (ggc_alloc_typed_stat): Remove.
20551 (ggc_alloc_typed): Remove.
20552 (ggc_splay_alloc): Remove first argument.
20553 (rtl_zone): Remove. Update all users.
20554 (tree_zone): Remove. Update all users.
20555 (tree_id_zone): Remove. Update all users.
20556 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 20557 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 20558 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 20559 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 20560
a861ffa4
GJL
205612013-01-24 Georg-Johann Lay <avr@gjlay.de>
20562
20563 * config/avr/avr.c (avr_out_fract): Make register numbers that
20564 might be outside of source operand signed.
20565
593c0ddd
UB
205662013-01-24 Uros Bizjak <ubizjak@gmail.com>
20567
20568 * config/i386/constraints.md (Yf): New constraint.
20569 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
20570 of f constraint to conditionaly disable x87 register preferences.
20571 (*movdf_internal): Ditto.
20572 (*movsf_internal): Ditto.
20573
e86c0101
SB
205742013-01-24 Steven Bosscher <steven@gcc.gnu.org>
20575
20576 PR inline-asm/55934
20577 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
20578 that have operands with impossible constraints.
20579 Add a FIXME for a speed-up opportunity.
20580 * lra-constraints.c (process_alt_operands): Verify that a class
20581 selected from constraints on asms is valid for the operand mode.
20582 (curr_insn_transform): Remove incorrect comment.
20583
f6fee35f
DE
205842013-01-23 David Edelsohn <dje.gcc@gmail.com>
20585
20586 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
20587 TOC operand is a valid symbol ref in the constant pool.
20588
205892013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 20590
aaf1e810 20591 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 20592
dc62d7d1
GJL
205932013-01-23 Georg-Johann Lay <avr@gjlay.de>
20594
20595 PR target/54222
20596 * config/avr/stdfix.h: New file.
20597 * t-avr (stdfix-gcc.h): New rule to build it.
20598 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
20599
2660d12d
KS
206002013-01-23 Kostya Serebryany <kcc@google.com>
20601
77dc5297
UB
20602 * config/darwin.h: remove dependency on
20603 CoreFoundation (asan on Mac OS).
2660d12d 20604
a70418fc
JJ
206052013-01-23 Jakub Jelinek <jakub@redhat.com>
20606
20607 PR target/49069
20608 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
20609 instead of cmpdi_operand for first comparison operand.
20610 Don't assert that comparison operands aren't both constants.
20611
47876a2a
JW
206122013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
20613
20614 * doc/install.texi (Downloading the Source): Update references to
20615 downloading separate components.
20616
206172013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW
20618
20619 * doc/extend.texi (__int128): Improve grammar.
20620
47876a2a 206212013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB
20622
20623 PR target/56028
20624 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
20625 alternative to (o,r).
20626 (*movdi_internal_rex64): Remove (!o,n) alternative.
20627 (DImode immediate->memory splitter): Remove.
20628 (DImode immediate->memory peephole2): Remove.
20629 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
20630 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
20631 alternative to (!o,*r).
20632 (*movtf_internal_sse): New pattern.
20633 (*movxf_internal_rex64): New pattern.
20634 (*movxf_internal): Disable for TARGET_64BIT.
20635 (*movdf_internal_rex64): Remove (!o,F) alternative.
20636
3a984f10
JJ
206372013-01-22 Jakub Jelinek <jakub@redhat.com>
20638
502498d5
JJ
20639 PR middle-end/56074
20640 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
20641 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
20642 * tree-vect-loop-manip.c (find_loop_location): Also ignore
20643 stmt locations where LOCATION_LOCUS of the stmt location is
20644 UNKNOWN_LOCATION or BUILTINS_LOCATION.
20645
3a984f10
JJ
20646 PR target/55686
20647 * config/i386/i386.md (UNSPEC_STOS): New.
20648 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
20649 *strsetqi_1): Add UNSPEC_STOS.
20650
fa817f7f
PC
206512013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
20652
20653 PR c++/56067
20654 * doc/invoke.texi: Remove left over -Wsynth example.
20655
8f498c1b
JJ
206562013-01-21 Jakub Jelinek <jakub@redhat.com>
20657
20658 PR tree-optimization/56051
20659 * fold-const.c (fold_binary_loc): Don't fold
20660 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
20661 a narrowing conversion, or widening conversion from signed
20662 to unsigned.
20663
47876a2a 206642013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB
20665
20666 PR rtl-optimization/56023
20667 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
20668 dependent on debug instruction.
20669
5b9db1bc
MJ
206702013-01-21 Martin Jambor <mjambor@suse.cz>
20671
20672 PR middle-end/56022
20673 * function.c (allocate_struct_function): Call
20674 invoke_set_current_function_hook earlier.
20675
e8bb7d68
JJ
206762013-01-21 Jakub Jelinek <jakub@redhat.com>
20677
20678 * reload1.c (init_reload): Only initialize reload_obstack
20679 during the first call.
20680
616a4e32
MP
206812013-01-21 Marek Polacek <polacek@redhat.com>
20682
20683 * cfgloop.c (verify_loop_structure): Fix up grammar.
20684
4401981b
YHH
206852013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
20686
20687 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
20688 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
20689
8e87740b
RR
206902013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20691
20692 PR target/56058
20693 * config/arm/marvell-pj4.md: Update copyright year.
20694 Fix up use of alu to alu_reg and simple_alu_imm.
20695
47876a2a 206962013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB
20697
20698 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
20699
89d56d79
VM
207002013-01-20 Vladimir Makarov <vmakarov@redhat.com>
20701
20702 PR target/55433
20703 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 20704 insn for secondary memory move when memory mode should be different.
89d56d79 20705
fe603553
JDA
207062013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20707
20708 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
20709 atomic_storedi_1): New patterns.
20710
01284895
VK
207112013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
20712
20713 btver2 pipeline descriptions.
20714 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
20715 descriptions.
20716 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 20717 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK
20718 type attributes.
20719 * config/i386/btver2.md: New file describing btver2 pipelines.
20720
5630e3e1
JL
207212013-01-19 Andrew Pinski <apinski@cavium.com>
20722
20723 PR tree-optimization/52631
20724 * tree-ssa-sccvn (visit_use): Before looking up the original
20725 statement, try looking up the simplified expression.
20726
650ae806
AG
207272013-01-19 Anthony Green <green@moxielogic.com>
20728
20729 * config/moxie/moxie.c (moxie_expand_prologue): Set
20730 current_function_static_stack_size.
20731
e300ec2d
JJ
207322013-01-18 Jakub Jelinek <jakub@redhat.com>
20733
20734 PR tree-optimization/56029
20735 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
20736 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
20737
a3d7af04
SS
207382013-01-18 Sharad Singhai <singhai@google.com>
20739
20740 PR tree-optimization/55995
20741 * dumpfile.c (dump_loc): Print location only if available.
20742 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
20743
66aa7879
VM
207442013-01-18 Vladimir Makarov <vmakarov@redhat.com>
20745
20746 PR target/55433
20747 * lra-constraints.c (curr_insn_transform): Reuse original insn for
20748 secondary memory move.
20749 (inherit_reload_reg): Use rclass instead of cl for
20750 check_secondary_memory_needed_p.
20751
3f0fee7b
JJ
207522013-01-18 Jakub Jelinek <jakub@redhat.com>
20753
20754 PR middle-end/56015
20755 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 20756 the case where writing real complex part of target modifies op1.
3f0fee7b 20757
70c67693
JG
207582013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
20759
20760 * config/aarch64/aarch64-simd.md
20761 (aarch64_vcond_internal<mode>): Handle unordered cases.
20762 * config/aarch64/iterators.md (v_cmp_result): New.
20763
df8de9b3
YHH
207642013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
20765 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20766
20767 * config/arm/marvell-pj4.md: New file.
20768 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
20769 * config/arm/arm.md (generic_sched): Add marvell_pj4.
20770 (generic_vfp): Likewise.
20771 * config/arm/arm-cores.def: Add marvell-pj4.
20772 * config/arm/arm-tune.md: Regenerate.
20773 * config/arm/arm-tables.opt: Regenerate.
20774 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
20775 * doc/invoke.texi: Document marvell-pj4.
20776
be30c356
TB
207772013-01-18 Tejas Belagod <tejas.belagod@arm.com>
20778
20779 * config/aarch64/arm_neon.h: Map scalar types to standard types.
20780
0bfbca58 207812013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO
20782
20783 PR debug/54114
20784 PR debug/54402
20785 PR debug/49888
20786 * var-tracking.c (negative_power_of_two_p): New.
20787 (global_get_addr_cache, local_get_addr_cache): New.
20788 (get_addr_from_global_cache, get_addr_from_local_cache): New.
20789 (vt_canonicalize_addr): Rewrite using the above. Adjust the
20790 heading comment.
20791 (vt_stack_offset_p): Remove.
20792 (vt_canon_true_dep): Always canonicalize loc's address.
20793 (clobber_overlapping_mems): Make sure we have a MEM.
20794 (local_get_addr_clear_given_value): New.
20795 (val_reset): Clear local cached entries.
20796 (compute_bb_dataflow): Create and release the local cache.
20797 Disable duplicate MEMs clobbering.
20798 (emit_notes_in_bb): Clobber MEMs likewise.
20799 (vt_emit_notes): Create and release the local cache.
20800 (vt_initialize, vt_finalize): Create and release the global
20801 cache, respectively.
1f6bc337 20802 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 20803
0bfbca58 208042013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO
20805
20806 PR libmudflap/53359
20807 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
20808 not found in the symtab.
20809
0bfbca58 208102013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 20811
c350ba53 20812 PR debug/56006
3aa03517
AO
20813 PR rtl-optimization/55547
20814 PR rtl-optimization/53827
20815 PR debug/53671
20816 PR debug/49888
20817 * alias.c (offset_overlap_p): New, factored out of...
20818 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
20819 the conservative special case for symbolic constants. Don't
20820 adjust zero sizes on alignment.
20821
c664546f
JL
208222013-01-18 Bernd Schmidt <bernds@codesourcery.com>
20823
20824 PR rtl-optimization/52573
20825 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
20826 REG_UNUSED for the same register.
20827
1bd3f750
MP
208282013-01-17 Richard Biener <rguenther@suse.de>
20829 Marek Polacek <polacek@redhat.com>
20830
20831 PR rtl-optimization/55833
20832 * loop-unswitch.c (unswitch_loops): Move loop verification...
20833 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
20834 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
20835 Set it to true when we're removing a loop from hierarchy tree in
20836 an irreducible region.
20837 (fix_bb_placements): Adjust caller.
20838 (fix_loop_placements): Likewise.
20839
e52a8b71
GJL
208402013-01-17 Georg-Johann Lay <avr@gjlay.de>
20841
20842 * config/avr/builtins.def (DEF_BUILTIN): Factor out
20843 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
20844 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
20845 Remove ID. Adjust comments.
20846 * config/avr/avr-c.c (avr_builtin_name): Remove.
20847 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
20848 * config/avr/avr.c (avr_tolower): New static function.
20849 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
20850 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
20851 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
20852 default expansion.
20853
8386a7ea
JH
208542013-01-17 Jan Hubicka <jh@suse.cz>
20855
610fb637 20856 PR tree-optimization/55273
8386a7ea
JH
20857 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
20858
47876a2a 208592013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB
20860
20861 PR target/55981
20862 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
20863 store through atomic_store<mode>_1.
20864 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
20865
8222c37e
MJ
208662013-01-17 Martin Jambor <mjambor@suse.cz>
20867
20868 PR tree-optimizations/55264
20869 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
20870 for virtual methods.
20871 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
20872 virtual methods before inlining is over.
20873 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
20874 virtual functions.
20875 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
20876 non-virtual.
20877
79f01c76
VM
208782013-01-16 Vladimir Makarov <vmakarov@redhat.com>
20879
20880 PR rtl-optimization/56005
20881 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
20882 pending reads for prefetch.
20883
d0b6bb1b
IB
208842013-01-16 Ian Bolton <ian.bolton@arm.com>
20885
aaf1e810 20886 * config/aarch64/aarch64.md
d0b6bb1b
IB
20887 (*cstoresi_neg_uxtw): New pattern.
20888 (*cmovsi_insn_uxtw): New pattern.
20889 (*<optab>si3_uxtw): New pattern.
20890 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
20891 (*<optab>si3_insn_uxtw): New pattern.
20892 (*bswapsi2_uxtw): New pattern.
20893
cb9cf03b
RB
208942013-01-16 Richard Biener <rguenther@suse.de>
20895
20896 * tree-inline.c (tree_function_versioning): Remove set but
20897 never used variable.
20898
2cfc56b9
RB
208992013-01-16 Richard Biener <rguenther@suse.de>
20900
20901 PR tree-optimization/55964
20902 * tree-flow.h (rename_variables_in_loop): Remove.
20903 (rename_variables_in_bb): Likewise.
20904 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
20905 (copy_loop_before): Adjust and delete update-ssa status.
20906 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 20907 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB
20908 (rename_variables_in_loop): Remove.
20909 (slpeel_update_phis_for_duplicate_loop): Likewise.
20910 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
20911 use available cfg machinery instead of duplicating it.
20912 Update PHI nodes and perform poor-mans SSA update here.
20913 (slpeel_tree_peel_loop_to_edge): Adjust.
20914
c25a0c60
RB
209152013-01-16 Richard Biener <rguenther@suse.de>
20916
20917 PR tree-optimization/54767
20918 PR tree-optimization/53465
20919 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
20920 (vrp_visit_phi_node): For PHI arguments coming via backedges
20921 drop all symbolical range information.
20922 (execute_vrp): Compute backedges.
20923
04b535af
RB
209242013-01-16 Richard Biener <rguenther@suse.de>
20925
20926 * doc/install.texi: Update CLooG and ISL requirements to
20927 0.18.0 and 0.11.1.
20928
8b0a1e0b
CB
209292013-01-16 Christian Bruel <christian.bruel@st.com>
20930
20931 PR target/55301
20932 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
20933 (broken_move): Handle UNSPECV_SP_SWITCH_B.
20934 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
20935
209362013-01-16 DJ Delorie <dj@redhat.com>
20937
20938 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
20939 (UNSPECV_SP_SWITCH_E): New.
20940 (sp_switch_1): Change to an unspec.
20941 (sp_switch_2): Change to an unspec. Don't use post-inc when we
20942 replace $r15.
20943
209442013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB
20945
20946 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
20947 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
20948 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
20949 (expand_mem_thread_fence): Ditto.
20950 (expand_mem_signal_fence): Ditto.
20951 (expand_atomic_load): Ditto.
20952 (expand_atomic_store): Ditto.
20953
0bfbca58 209542013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO
20955
20956 PR rtl-optimization/55547
20957 PR rtl-optimization/53827
20958 PR debug/53671
20959 PR debug/49888
20960 * alias.c (memrefs_conflict_p): Set sizes to negative after
20961 AND adjustments.
20962
305e3ac1
JJ
209632013-01-15 Jakub Jelinek <jakub@redhat.com>
20964
20965 PR target/55940
20966 * function.c (thread_prologue_and_epilogue_insns): Always
20967 add crtl->drap_reg to set_up_by_prologue.set, even if
20968 stack_realign_drap is false.
20969
f78ac4f2
JBG
209702013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
20971
20972 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
20973 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
20974 *call): Fix indention.
20975
a78a8cc4
TV
209762013-01-15 Tom de Vries <tom@codesourcery.com>
20977
20978 PR target/55876
20979 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
20980 Update comment.
20981
0e80383f
VM
209822013-01-15 Vladimir Makarov <vmakarov@redhat.com>
20983
305e3ac1 20984 PR rtl-optimization/55153
0e80383f
VM
20985 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
20986
207b5956
MJ
209872013-01-15 Martin Jambor <mjambor@suse.cz>
20988
20989 PR tree-optimization/55920
20990 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
20991 accesses as grp_to_be_debug_replaced.
20992
a7818b54
JJ
209932013-01-15 Jakub Jelinek <jakub@redhat.com>
20994
20995 PR tree-optimization/55920
20996 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
20997 there is non-useless type conversion needed from debug rhs to lhs,
20998 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
20999
b0fe107e
JM
210002013-01-15 Joseph Myers <joseph@codesourcery.com>
21001 Mikael Pettersson <mikpe@it.uu.se>
21002
21003 PR target/43961
21004 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
21005 Thumb.
21006 (ASM_OUTPUT_CASE_LABEL): Remove.
21007 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
21008 * final.c (shorten_branches): Update alignment of labels before
21009 jump tables if CASE_VECTOR_SHORTEN_MODE.
21010
34ab62ee
RB
210112013-01-15 Richard Biener <rguenther@suse.de>
21012
21013 PR bootstrap/55961
21014 * system.h: Do not include gmp.h for building host tools.
21015
783a3a05
RB
210162013-01-15 Richard Biener <rguenther@suse.de>
21017
21018 PR middle-end/55882
21019 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
21020 account for bitpos when computing alignment.
21021
3a579e09
VY
210222013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
21023
21024 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
21025 (ix86_target_macros_internal): Likewise.
21026
21027 * config/i386/i386.c (m_CORE2I7): Removed.
21028 (m_CORE_HASWELL): New macro.
21029 (m_CORE_ALL): Likewise.
21030 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
21031 (initial_ix86_arch_features): Likewise.
21032 (processor_target_table): Initializations for Core avx2.
21033 (cpu_names): New names "core-avx2".
21034 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
21035 PROCESSOR_CORE_HASWELL.
21036 (ix86_issue_rate): New case.
21037 (ia32_multipass_dfa_lookahead): Likewise.
21038 (ix86_sched_init_global): Likewise.
21039
21040 * config/i386/i386.h (TARGET_HASWELL): New macro.
21041 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
21042 (processor_type): New PROCESSOR_HASWELL.
21043
955f5a07
JJ
210442013-01-15 Jakub Jelinek <jakub@redhat.com>
21045
ff784829
JJ
21046 PR tree-optimization/55955
21047 * tree-vect-loop.c (vectorizable_reduction): Give up early on
21048 *SHIFT_EXPR and *ROTATE_EXPR codes.
21049
955f5a07
JJ
21050 PR tree-optimization/48766
21051 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
21052 -ftrapv disable -fwrapv.
21053
aeb8b4e9
GJL
210542013-01-14 Georg-Johann Lay <avr@gjlay.de>
21055
21056 PR target/55974
21057 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
21058 etc. to 1 and not to __flash.
21059 Use LL suffix for __INT24_MAX__ with -mint8.
21060 Use ULL suffix for __UINT24_MAX__ with -mint8.
21061
1c494c6a
GJL
210622013-01-14 Georg-Johann Lay <avr@gjlay.de>
21063
21064 * config/avr/avr-arch.h
21065 (struct base_arch_s): Use typedef avr_arch_t instead.
21066 (struct arch_info_s): Use typedef avr_arch_info_t instead.
21067 (struct mcu_type_s): Use typedef avr_mcu_t instead.
21068 * config/avr/avr.c: Same.
21069 * config/avr/avr-devices.c: Same.
21070 * config/avr/driver-avr.c: Same.
21071 * config/avr/gen-avr-mmcu-texi.c: Same.
21072 * config/avr/avr-mcus.def: Adjust comment.
21073
a50344cb
TB
210742013-01-14 Tejas Belagod <tejas.belagod@arm.com>
21075
88e784e6
UB
21076 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
21077 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 21078
47876a2a 210792013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB
21080 Andi Kleen <ak@linux.intel.com>
21081
21082 PR target/55948
21083 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
21084 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
21085 memmodel flag.
21086
00892272
GJL
210872013-01-14 Georg-Johann Lay <avr@gjlay.de>
21088
21089 * config/avr/avr-stdint.h: Remove trailing blanks.
21090 * config/avr/avr-log.h: Same.
21091 * config/avr/avr-arch.h: Same.
21092 * config/avr/avr-devices.c: Same.
21093 * config/avr/avr-dimode.md: Same.
21094 * config/avr/predicates.md: Same.
21095 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 21096
00892272
GJL
21097 * config/avr/avr-protos.h: Same. And:
21098 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
21099 (init_cumulative_args): Rename to avr_init_cumulative_args.
21100 (expand_prologue): Rename to avr_expand_prologue.
21101 (expand_epilogue): Rename to avr_expand_epilogue.
21102 (adjust_insn_length): Rename to avr_adjust_insn_length.
21103 (notice_update_cc): Rename to avr_notice_update_cc.
21104 (final_prescan_insn): Rename to avr_final_prescan_insn.
21105 * config/avr/avr.c: Same.
21106 * config/avr/avr.h: Same.
21107 * config/avr/avr.md: Remove trailing blanks.
21108 (prologue): Use avr_expand_prologue.
21109 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
21110
7bb01996
RB
211112013-01-14 Richard Biener <rguenther@suse.de>
21112
21113 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
21114 verify_location, collect_subblocks): New functions.
21115 (verify_gimple_in_cfg): Verify that locations only reference
21116 BLOCKs in the functions BLOCK tree.
21117
2724573f
RB
211182013-01-14 Richard Biener <rguenther@suse.de>
21119
21120 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
21121 PHI argument.
21122 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
21123 unshare reference.
21124 (insert_out_of_ssa_copy_on_edge): Likewise.
21125 (rewrite_close_phi_out_of_ssa): Likewise.
21126 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
21127 debug expressions.
21128 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
21129 propagated constants.
21130 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
21131 can not be shared.
21132
9a0bbab6
GJL
211332013-01-14 Georg-Johann Lay <avr@gjlay.de>
21134
21135 * config/avr/avr-modes.def: Add GPL copyright notice.
21136
45805f17
UB
211372013-01-13 Uros Bizjak <ubizjak@gmail.com>
21138
21139 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
21140 MEMMODEL_MASK to determine memory model.
21141 (atomic_store<mode>): Ditto from operands[2].
21142 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
21143
9d60be38
JJ
211442013-01-13 Jakub Jelinek <jakub@redhat.com>
21145
21146 PR fortran/55935
45805f17 21147 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ
21148 (fold_gimple_assign): Don't call unshare_expr here.
21149 (fold_ctor_reference): Call unshare_expr.
21150
e7f49d92
TG
211512013-01-13 Terry Guo <terry.guo@arm.com>
21152
aaf1e810
EB
21153 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
21154 * doc/fragments.texi: Document MULTILIB_REUSE.
21155 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 21156 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 21157 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG
21158 (tmpmultilib4): Ditto.
21159 (multilib_reuse): New multilib argument.
21160
fbd03a27
RS
211612013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
21162
21163 * Makefile.in: Update copyright.
21164
b3681f13
TV
211652013-01-12 Tom de Vries <tom@codesourcery.com>
21166
21167 PR middle-end/55890
21168 * calls.c (expand_call): Check if arg_nr is valid.
21169
3f287e4b
MM
211702013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
21171
21172 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
21173 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM
21174 documentation. Add missing '__' in front of
21175 __builtin_ia32_packssdw256.
3f287e4b 21176
1abcd5eb
AK
211772013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21178
21179 PR target/55719
21180 * config/s390/s390.c (s390_preferred_reload_class): Do not return
21181 NO_REGS for larl operands.
21182 (s390_reload_larl_operand): Use s390_load_address instead of
21183 emit_move_insn.
21184
980d0812
RB
211852013-01-11 Richard Biener <rguenther@suse.de>
21186
21187 * tree-cfg.c (verify_node_sharing_1): Split out from ...
21188 (verify_node_sharing): ... here.
21189 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
21190
7cb9fd07
EB
211912013-01-11 Eric Botcazou <ebotcazou@adacore.com>
21192
21193 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
21194 Substitute TREECHECKING.
21195 * configure: Regenerate.
21196 * Makefile.in (TREECHECKING): New.
21197
47876a2a 211982013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG
21199
21200 PR tree-optimization/44061
21201 * tree-vrp.c (extract_range_basic): Compute zero as
21202 value-range for __builtin_constant_p of function parameters.
21203
d1e082c2
RS
212042013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
21205
45805f17 21206 Update copyright years.
d1e082c2 21207
f85021b0
VM
212082013-01-10 Vladimir Makarov <vmakarov@redhat.com>
21209
9d60be38 21210 PR rtl-optimization/55672
0160442c
VM
21211 * lra-eliminations.c (mark_not_eliminable): Permit addition with
21212 const to be eliminable.
f85021b0 21213
7a8b1ec4
DE
212142013-01-10 David Edelsohn <dje.gcc@gmail.com>
21215
21216 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
21217 * configure: Regenerate.
21218
ad2e5b71
RB
212192013-01-10 Richard Biener <rguenther@suse.de>
21220
ddf9322d 21221 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 21222
ffc5b2bb
RB
212232013-01-10 Richard Biener <rguenther@suse.de>
21224
21225 PR bootstrap/55792
21226 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
21227 locations for virtual PHI arguments.
21228 (rewrite_update_phi_arguments): Likewise.
21229
e1f674e4
JS
212302013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
21231
21232 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
21233 on to assembler.
21234
a57fc743
JJ
212352013-01-10 Jakub Jelinek <jakub@redhat.com>
21236
21237 PR tree-optimization/55921
21238 * tree-complex.c (expand_complex_asm): New function.
21239 (expand_complex_operations_1): Call it for GIMPLE_ASM.
21240
0ff4390d
AK
212412013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21242
21243 PR target/55718
21244 * config/s390/s390.c (s390_symref_operand_p)
21245 (s390_loadrelative_operand_p): Merge the two functions.
21246 (s390_check_qrst_address, print_operand_address): Add parameters
21247 to s390_loadrelative_operand_p invokation.
21248 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
21249 (s390_reload_larl_operand, s390_secondary_reload): Use
21250 s390_loadrelative_operand_p instead of s390_symref_operand_p.
21251 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
21252
6f557e0e
MS
212532013-01-09 Mike Stump <mikestump@comcast.net>
21254
21255 * dse.c (record_store): Remove unnecessary assert.
21256
7770c9e9
JH
212572013-01-09 Jan Hubicka <jh@suse.cz>
21258
21259 PR tree-optimization/55569
21260 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
21261 * cfgloop.h (scale_loop_profile): Likewise.
21262
a19b1432
JH
212632013-01-09 Jan Hubicka <jh@suse.cz>
21264
21265 PR lto/45375
ddf9322d
UB
21266 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
21267 functions.
a19b1432
JH
21268 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
21269
9a002da8
RS
212702013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
21271
21272 PR middle-end/55114
21273 * expr.h (maybe_emit_group_store): Declare.
21274 * expr.c (maybe_emit_group_store): New function.
21275 * builtins.c (expand_builtin_int_roundingfn): Call it.
21276 (expand_builtin_int_roundingfn_2): Likewise.
21277
511dcace
VM
212782013-01-09 Vladimir Makarov <vmakarov@redhat.com>
21279
e1f2b729 21280 PR rtl-optimization/55829
511dcace
VM
21281 * lra-constraints.c (match_reload): Add code for absent output.
21282 (curr_insn_transform): Add code for reloads of matched inputs
21283 without output.
21284
7b0fe4f4
UB
212852013-01-09 Uros Bizjak <ubizjak@gmail.com>
21286
21287 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
21288 attribute of movddup insn to DF.
21289 (*vec_interleave_lowv2df): Ditto.
21290 (vec_dupv2df): Ditto.
21291
870ca331
JH
212922013-01-09 Jan Hubicka <jh@suse.cz>
21293
21294 PR tree-optimiation/55875
21295 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
21296 EVERY_ITERATION parameter.
7b0fe4f4 21297 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH
21298 (idx_infer_loop_bounds): Similarly here.
21299 (n_of_executions_at_most): Simplify
21300 to only test for cases where statement is dominated by the
7b0fe4f4 21301 particular bound; handle correctly the "postdominance" test.
870ca331
JH
21302 (scev_probably_wraps_p): Use max loop iterations info
21303 as a global bound first.
21304
6f575fe4 213052013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD
21306 Nick Clifton <nickc@redhat.com>
21307
21308 * config/v850/v850.md (cbranchsf4): New pattern.
21309 (cstoresf4): New pattern.
21310 (cbranchdf4): New pattern.
21311 (cstoredf4): New pattern.
21312 (movsicc): Disallow floating point comparisons.
21313 (cmpsf_le_insn): Fix order of operators.
21314 (cmpsf_lt_insn): Likewise.
21315 (cmpsf_eq_insn): Likewise.
21316 (cmpdf_le_insn): Likewise.
21317 (cmpdf_lt_insn): Likewise.
21318 (cmpdf_eq_insn): Likewise.
21319 (cmpsf_ge_insn): Use LE comparison.
21320 (cmpdf_ge_insn): Likewise.
21321 (cmpsf_gt_insn): Use LT comparison.
21322 (cmpdf_gt_insn): Likewise.
21323 (cmpsf_ne_insn): Delete pattern.
21324 (cmpdf_ne_insn): Delete pattern.
21325 * config/v850/v850.c (v850_gen_float_compare): Use
21326 gen_cmpdf_eq_insn for NE comparison.
21327 (v850_float_z_comparison_operator)
21328 (v850_float_nz_comparison_operator): Move from here ...
21329 * config/v850/predicates.md: ... to here. Move GT and GE
21330 comparisons into v850_float_z_comparison_operator.
21331 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
21332 Delete prototype.
21333 (v850_float_nz_comparison_operator): Likewise.
21334
f0d54148
JDA
213352013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21336
21337 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
21338 with calls to gen_insvsi/gen_insvdi.
21339
8f01beca
VK
213402013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
21341
21342 * config/i386/i386.c (initial_ix86_tune_features): Set up
21343 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
21344
2aa579ad
JJ
213452013-01-09 Steven Bosscher <steven@gcc.gnu.org>
21346 Jakub Jelinek <jakub@redhat.com>
21347
21348 PR tree-optimization/48189
21349 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
21350 If nitercst is 0, don't predict the exit edge.
21351
6edc3e32 213522013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N
21353
21354 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
21355 in asm_fprintf with reg_names.
21356 (aarch64_print_operand_address): Likewise.
21357 (aarch64_return_addr): Likewise.
21358 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
21359
f6f94d94
JDA
213602013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21361
21362 * config/pa/pa.h (VAL_U6_BITS_P): Define.
21363 (INT_U6_BITS): Likewise.
21364 * config/pa/predicates.md (uint6_operand): New predicate.
21365 (shift5_operand, shift6_operand): Likewise.
21366 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
21367 arith32_operand.
21368 (lshrdi3): Use shift6_operand.
21369 (shrpsi4, shrpdi4): New insn patterns.
21370 (extzv): Delete expander.
21371 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
21372 predicates in unamed zero extract patterns. Tighten common constraint.
21373 (extv): Delete expander.
21374 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
21375 predicates in unamed sign extract patterns. Tighten common constraint.
21376 (insv): Delete expander.
21377 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
21378 predicates in unamed insert patterns. Tighten common constraint.
21379 Change uint32_operand predicate to uint6_operand predicate in unamed
21380 DImode pattern to insert constant values of type 1...1xxxx.
21381
36b72910
JH
213822013-01-04 Jan Hubicka <jh@suse.cz>
21383
21384 PR tree-optimization/55823
7b0fe4f4
UB
21385 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
21386 issue.
36b72910 21387
47876a2a 213882013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ
21389 Uros Bizjak <ubizjak@gmail.com>
21390
21391 PR rtl-optimization/55845
21392 * df-problems.c (can_move_insns_across): Stop scanning at
21393 volatile_insn_p source instruction or give up if
21394 across_from .. across_to range contains any volatile_insn_p
21395 instructions.
21396
4369c11e
TB
213972013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21398
7b0fe4f4
UB
21399 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
21400 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
21401 Declare.
4369c11e 21402 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 21403 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 21404
aeb7e7c1
JJ
214052013-01-08 Jakub Jelinek <jakub@redhat.com>
21406
21407 PR fortran/55341
21408 * asan.c (asan_clear_shadow): New function.
21409 (asan_emit_stack_protection): Use it.
21410
a02ad1aa
TB
214112013-01-08 Tejas Belagod <tejas.belagod@arm.com>
21412
21413 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
21414 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
21415 with tab instead of space.
21416
f90d87f5
NC
214172013-01-08 Nick Clifton <nickc@redhat.com>
21418
21419 * config/rl78/rl78.c (rl78_expand_prologue): Always select
21420 register bank 0 at the start of an interrupt handler.
83ffd964
NC
21421 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
21422 MDBH registers.
f90d87f5 21423
385eb93d
JG
214242013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21425
21426 * config/aarch64/aarch64-simd.md
21427 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
21428 (aarch64_simd_bsl): Likewise.
21429 (aarch64_vcond_internal<mode>): Likewise.
21430 (vcond<mode><mode>): Likewise.
21431 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
21432 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
21433
4dcd1054
JG
214342013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
21435
21436 * config/aarch64/aarch64-builtins.c
21437 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
21438
4267a4a6
MJ
214392013-01-08 Martin Jambor <mjambor@suse.cz>
21440
21441 PR debug/55579
21442 * tree-sra.c (analyze_access_subtree): Return true also after
21443 potentially creating a debug-only replacement.
21444
5f4e6de3
JJ
214452013-01-08 Jakub Jelinek <jakub@redhat.com>
21446
3138f224
JJ
21447 PR middle-end/55890
21448 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
21449
5f4e6de3
JJ
21450 PR tree-optimization/54120
21451 * tree-vrp.c (range_fits_type_p): Don't allow
21452 src_precision < precision from signed vr to unsigned_p
21453 if vr->min or vr->max is negative.
21454 (simplify_float_conversion_using_ranges): Test can_float_p
21455 against CODE_FOR_nothing.
21456
f3ef18ff
JJ
214572013-01-08 Jakub Jelinek <jakub@redhat.com>
21458 Richard Biener <rguenther@suse.de>
21459
21460 PR middle-end/55851
21461 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
21462 types instead of just INTEGER_TYPE types.
21463
25c210f9
MK
214642013-01-07 Mark Kettenis <kettenis@openbsd.org>
21465
21466 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
21467 TF_SIZE): Define.
f90d87f5 21468
81826a7b
SE
214692013-01-07 Steve Ellcey <sellcey@mips.com>
21470
21471 PR target/42661
21472 * config/mips/mips.opt: Change mad to mmad to match documentation.
21473
35678514
GJL
214742013-01-07 Georg-Johann Lay <avr@gjlay.de>
21475
21476 PR target/55897
21477 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
21478 .progmemx.data now.
21479
e5669488
GJL
214802013-01-07 Georg-Johann Lay <avr@gjlay.de>
21481
21482 PR target/55897
21483 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
21484 (avr_addrspace_t): Add .section_name field.
21485 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
21486 array size.
21487 (avr_addrspace): Same. Initialize .section_name. Remove last
21488 NULL entry. Put __memx into .progmemx.data.
21489 (progmem_section_prefix): Remove.
21490 (avr_asm_init_sections): No need to initialize progmem_section.
21491 (avr_asm_named_section): Use avr_addrspace[].section_name to get
21492 section name prefix.
21493 (avr_asm_select_section): Ditto. And use get_unnamed_section to
21494 retrieve the progmem section.
21495 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
21496 boundary to run over avr_addrspace[].
21497 (avr_register_target_pragmas): Ditto.
21498
656e6f37
JJ
214992013-01-06 Jakub Jelinek <jakub@redhat.com>
21500
21501 * varasm.c (output_constant_def_contents): For asan_protect_global
21502 protected strings, adjust DECL_ALIGN if needed, before testing for
21503 anchored symbols.
21504 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
21505 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
21506 normal decls.
21507 (output_object_block): For asan protected decls, emit asan padding
21508 after their contents.
21509 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
21510 (asan_finish_file): Test it here instead.
21511
6edc3e32
UB
215122013-01-07 Nick Clifton <nickc@redhat.com>
21513 Matthias Klose <doko@debian.org>
21514 Doug Kwan <dougkwan@google.com>
21515 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC
21516
21517 PR driver/55470
21518 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
21519
21520 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
21521
21522 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
21523
7b0fe4f4 21524 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC
21525
21526 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
21527
2a095093
GJL
215282013-01-07 Georg-Johann Lay <avr@gjlay.de>
21529
f3b503f4 21530 PR target/54461
2a095093
GJL
21531 * doc/install.texi (Cross-Compiler-Specific Options): Document
21532 --with-avrlibc.
21533
383f9b34
TB
215342013-01-07 Tejas Belagod <tejas.belagod@arm.com>
21535
21536 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
21537 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
21538 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
21539 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
21540 vqmovun_high_s64): Fix source operand number and update copyright.
21541
3626621a
RB
215422013-01-07 Richard Biener <rguenther@suse.de>
21543
21544 PR middle-end/55890
21545 * gimple.h (gimple_call_builtin_p): New overload.
21546 * gimple.c (validate_call): New function.
21547 (gimple_call_builtin_p): Likewise.
21548 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
21549 Use gimple_call_builtin_p.
21550 (find_func_clobbers): Likewise.
21551 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
21552 (strlen_optimize_stmt): Likewise.
21553
8b2ea410
JG
215542013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21555
21556 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
21557 (vld1q_dup_*): Likewise.
21558 (vld1_*): Likewise.
21559 (vld1q_*): Likewise.
21560 (vld1_lane_*): Likewise.
21561 (vld1q_lane_*): Likewise.
21562
9713d329
RB
215632013-01-07 Richard Biener <rguenther@suse.de>
21564
21565 * lto-streamer.h (LTO_minor_version): Bump to 2.
21566
3520f7cc
JG
215672013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
21568
21569 * config/aarch64/aarch64-protos.h
21570 (aarch64_const_double_zero_rtx_p): Rename to...
21571 (aarch64_float_const_zero_rtx_p): ...this.
21572 (aarch64_float_const_representable_p): New.
21573 (aarch64_output_simd_mov_immediate): Likewise.
21574 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
21575 move immediate case.
21576 * config/aarch64/aarch64.c
21577 (aarch64_const_double_zero_rtx_p): Rename to...
21578 (aarch64_float_const_zero_rtx_p): ...this.
21579 (aarch64_print_operand): Allow printing of new constants.
21580 (aarch64_valid_floating_const): New.
21581 (aarch64_legitimate_constant_p): Check for valid floating-point
21582 constants.
21583 (aarch64_simd_valid_immediate): Likewise.
21584 (aarch64_vect_float_const_representable_p): New.
21585 (aarch64_float_const_representable_p): Likewise.
21586 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
21587 (aarch64_output_simd_mov_immediate): New.
21588 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
21589 (*movdf_aarch64): Likewise.
21590 * config/aarch64/constraints.md (Ufc): New.
21591 (Y): call aarch64_float_const_zero_rtx.
21592 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
21593
e167c04d
RB
215942013-01-07 Richard Biener <rguenther@suse.de>
21595
21596 PR tree-optimization/55888
21597 PR tree-optimization/55862
21598 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
21599 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
21600 not if it is contained therein.
21601
0139aaab
GJL
216022013-01-07 Georg-Johann Lay <avr@gjlay.de>
21603
21604 * config/avr/t-avr: Typo.
21605
4a176b23
GJL
216062013-01-07 Georg-Johann Lay <avr@gjlay.de>
21607
21608 PR55243
21609 * config/avr/t-avr: Don't automatically rebuild
21610 $(srcdir)/config/avr/t-multilib
21611 $(srcdir)/config/avr/avr-tables.opt
21612 $(srcdir)/doc/avr-mmcu.texi
21613 (avr-mcus): New phony target to build them on request.
21614 (s-avr-mlib, s-avr-mmcu-texi): Remove.
21615 * avr/avr-mcus.def: Adjust comments.
21616
c7afdc98
UB
216172013-01-07 Uros Bizjak <ubizjak@gmail.com>
21618
21619 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
21620
1ab05c31
RS
216212013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21622
21623 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
21624
488e3acc
RS
216252013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
21626
21627 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
21628
a89599d2
DE
216292013-01-05 David Edelsohn <dje.gcc@gmail.com>
21630
21631 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
21632 to generate profiling.
21633 * config/rs6000/aix64.h (LIB_SPEC): Same.
21634
70f09188
AP
216352013-01-04 Andrew Pinski <apinski@cavium.com>
21636
21637 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
21638 New function.
21639 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
21640
918d445f
UB
216412013-01-04 Uros Bizjak <ubizjak@gmail.com>
21642
21643 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
21644 unconditionally.
21645 (ix86_expand_move): Ditto.
21646 (ix86_zero_extend_to_Pmode): Ditto.
21647 (ix86_expand_call): Ditto.
21648 (ix86_expand_special_args_builtin): Ditto.
21649 (ix86_expand_builtin): Ditto.
21650
361618ec
RB
216512013-01-04 Richard Biener <rguenther@suse.de>
21652
21653 PR tree-optimization/55862
21654 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
21655 translating them through PHI nodes.
21656
87eab554
MJ
216572013-01-04 Martin Jambor <mjambor@suse.cz>
21658
21659 PR tree-optimization/55755
21660 * tree-sra.c (sra_modify_assign): Do not check that an access has no
21661 children when trying to avoid producing a VIEW_CONVERT_EXPR.
21662
33879b9f
MP
216632013-01-04 Marek Polacek <polacek@redhat.com>
21664
21665 PR middle-end/55859
21666 * opts.c (default_options_optimization): Clarify error message.
21667
3068819a
RB
216682013-01-04 Richard Biener <rguenther@suse.de>
21669
21670 PR middle-end/55863
21671 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
21672 reassociation.
21673
0e4ae794
JDA
216742013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21675
21676 PR target/53789
21677 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA
21678 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
21679 references.
0e4ae794 21680
65c88cef
RH
216812013-01-03 Richard Henderson <rth@redhat.com>
21682
21683 * config/i386/i386.c (ix86_expand_move): Always assign to op1
21684 after eliminating TLS symbols.
21685
8ac16127
MG
216862013-01-03 Marc Glisse <marc.glisse@inria.fr>
21687
21688 PR bootstrap/50167
21689 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
21690 * graphite-poly.c (debug_gmp_value): Likewise.
21691
bb664f09
UB
216922013-01-03 Uros Bizjak <ubizjak@gmail.com>
21693
21694 PR target/55712
21695 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
21696 selected code model, define __code_mode_small__, __code_model_medium__,
21697 __code_model_large__, __code_model_32__ or __code_model_kernel__.
21698 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
21699 xchg temporary register with %k. Declare temporary register as
21700 early clobbered.
21701 [__x86_64__]: For medium and large code models, preserve %rbx register.
21702
bcf1ef00
RB
217032013-01-03 Richard Biener <rguenther@suse.de>
21704
0506634a 21705 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB
21706 (dump_subscript): Adjust.
21707 (finalize_ddr_dependent): Do not dump redundant info.
21708 (analyze_siv_subscript): Adjust.
21709 (subscript_dependence_tester): Likewise.
21710 (compute_affine_dependence): Likewise.
21711
59fd17e3
RB
217122013-01-03 Richard Biener <rguenther@suse.de>
21713
21714 Revert
21715 2013-01-03 Richard Biener <rguenther@suse.de>
21716
21717 PR tree-optimization/55857
21718 * tree-vect-stmts.c (vectorizable_load): Do not setup
21719 re-alignment for invariant loads.
21720
21721 2013-01-02 Richard Biener <rguenther@suse.de>
21722
21723 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 21724 invariant load do not generate a vector load from the scalar location.
59fd17e3 21725
595c2679
RB
217262013-01-03 Richard Biener <rguenther@suse.de>
21727
21728 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
21729 for not vectorizing.
21730 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
21731 not build INDIRECT_REFs, call get_name once only.
21732 (vect_create_data_ref_ptr): Likewise. Dump base object kind
21733 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
21734
90bb7d7a
RB
217352013-01-03 Richard Biener <rguenther@suse.de>
21736
21737 PR tree-optimization/55857
21738 * tree-vect-stmts.c (vectorizable_load): Do not setup
21739 re-alignment for invariant loads.
21740
f09b77ca
RB
217412013-01-03 Richard Biener <rguenther@suse.de>
21742
21743 PR lto/55848
21744 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
21745 prefer a built-in decl.
21746
3461a16e
JJ
217472013-01-03 Jakub Jelinek <jakub@redhat.com>
21748
df8e2b4f
JJ
21749 * gcc.c (process_command): Update copyright notice dates.
21750 * gcov.c (print_version): Likewise.
21751 * gcov-dump.c (print_version): Likewise.
21752
3461a16e
JJ
21753 PR rtl-optimization/55838
21754 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
21755 iv0.step, iv1.step and step.
21756
8b5546d6
JJ
217572013-01-03 Jakub Jelinek <jakub@redhat.com>
21758 Marc Glisse <marc.glisse@inria.fr>
21759
21760 PR tree-optimization/55832
21761 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
21762 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
21763 integer_{one,zero}_node.
21764
8ab1d2e9
JJ
217652013-01-03 Jakub Jelinek <jakub@redhat.com>
21766
21767 PR debug/54402
21768 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
21769 * var-tracking.c (reverse_op): Don't add reverse ops to
21770 VALUEs that have already
0506634a 21771 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 21772
5b9ad1d4
GP
217732013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
21774
21775 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
21776
e25a6711
TJ
217772013-01-02 Teresa Johnson <tejohnson@google.com>
21778
21779 * dumpfile.c (dump_loc): Print filename with location.
21780 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
21781 new location_t parameter to emit complete unroll message with
21782 new dump framework.
21783 (canonicalize_loop_induction_variables): Compute loops location
21784 and pass to try_unroll_loop_completely.
21785 * loop-unroll.c (report_unroll_peel): New function.
21786 (peel_loops_completely): Use new dump format with location
21787 for main dumpfile message, and invoke report_unroll_peel on success.
21788 (decide_unrolling_and_peeling): Ditto.
21789 (decide_peel_once_rolling): Remove old dumpfile message subsumed
21790 by report_unroll_peel.
21791 (decide_peel_completely): Ditto.
21792 (decide_unroll_constant_iterations): Ditto.
21793 (decide_unroll_runtime_iterations): Ditto.
21794 (decide_peel_simple): Ditto.
21795 (decide_unroll_stupid): Ditto.
21796 * cfgloop.c (get_loop_location): New function.
21797 * cfgloop.h (get_loop_location): Declare.
21798
77878621
ST
217992013-01-02 Sriraman Tallam <tmsriram@google.com>
21800
21801 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
21802 NULL.
21803
9e65d03e
JDA
218042013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21805
21806 PR middle-end/55198
21807 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
21808 BLKmode objects when EXPAND_MEMORY is specified.
21809
6a7da30f
ST
218102013-01-02 Sriraman Tallam <tmsriram@google.com>
21811
21812 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
21813 in loop predicate.
21814 (fold_builtin_cpu): Do not share cpu model decls across statements.
21815
e78167f2
JM
218162013-01-02 Jason Merrill <jason@redhat.com>
21817
21818 PR c++/55804
21819 * tree.c (build_array_type_1): Revert earlier change.
21820
8c075fb4
YZ
218212013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
21822
21823 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
21824 "cortex-a57".
21825 * config/aarch64/aarch64-tune.md: Re-generate.
21826
0682ed3e
RB
218272013-01-02 Richard Biener <rguenther@suse.de>
21828
21829 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 21830 invariant load do not generate a vector load from the scalar location.
0682ed3e 21831
b8f6a302
RB
218322013-01-02 Richard Biener <rguenther@suse.de>
21833
21834 PR bootstrap/55784
21835 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
21836 * configure: Regenerate.
21837
04b80dbb
RS
218382013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21839
21840 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
21841 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
21842 (expand_builtin_int_roundingfn_2): Keep the original target around
21843 for the fallback case.
21844
635b0b0c
RS
218452013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
21846
21847 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
21848 to be clear for sign changes.
21849
42d57399
JH
218502013-01-01 Jan Hubicka <jh@suse.cz>
21851
21852 * ipa-inline-analysis.c: Fix formatting.
21853
5bb6669d
JJ
218542013-01-01 Jakub Jelinek <jakub@redhat.com>
21855
21856 PR tree-optimization/55831
21857 * tree-vect-loop.c (get_initial_def_for_induction): Use
21858 gsi_after_labels instead of gsi_start_bb.
ad41bd84 21859\f
86a2db33 21860Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM
21861
21862Copying and distribution of this file, with or without modification,
21863are permitted in any medium without royalty provided the copyright
21864notice and this notice are preserved.