1 2009-01-11 Markus Schoepflin <markus.schoepflin@comsoft.de>
4 * gcc/mips-tfile.c (parse_def): Fix parsing of def strings
7 2009-01-10 Jakub Jelinek <jakub@redhat.com>
10 * config/arm/arm.c (arm_is_long_call_p): Don't call
11 arm_function_in_section_p if decl isn't a FUNCTION_DECL.
13 2009-01-09 Steven Bosscher <steven@gcc.gnu.org>
15 * regrename.c (regrename_optimize): Fix dumping.
16 (find_oldest_value_reg): Preserve REG_POINTER.
17 (copy_hardreg_forward_1): Likewise.
19 2009-01-09 Diego Novillo <dnovillo@google.com>
21 * gimple.h (struct gimple_statement_base)<uid>: Document
22 the restrictions on its use.
24 (gimple_set_uid): Tidy.
26 2009-01-09 Jakub Jelinek <jakub@redhat.com>
28 * config/i386/i386.c (ix86_expand_movmem, ix86_expand_setmem): Add
29 zero guard even if align_bytes != 0 and count is smaller than
32 2008-01-09 Vladimir Makarov <vmakarov@redhat.com>
34 PR rtl-optimization/38495
35 * ira-emit.c (print_move_list, ira_debug_move_list): New
37 (add_range_and_copies_from_move_list): Print all added ranges.
38 Add ranges to memory optimized destination.
40 2009-01-09 Jakub Jelinek <jakub@redhat.com>
44 * config/i386/i386.c (override_options): Reject
45 -mstringop-strategy=rep_8byte with -m32.
46 (ix86_expand_movmem): For size_needed == 1 set epilogue_size_needed
47 to 1. Do count comparison against epilogue_size_needed at compile
48 time even when count_exp was constant forced into register. For
49 size_needed don't jump to epilogue, instead just avoid aligning
50 and invoke the body algorithm. If need_zero_guard, add zero guard
51 even if count is non-zero, but smaller than size_needed + number of
52 bytes that could be stored for alignment.
53 (ix86_expand_setmem): For size_needed == 1 set epilogue_size_needed
54 to 1. If need_zero_guard, add zero guard even if count is non-zero,
55 but smaller than size_needed + number of bytes that could be stored
56 for alignment. Compare size_needed with epilogue_size_needed instead
57 of desired_align - align, don't adjust size_needed, pass
58 epilogue_size_needed to the epilogue expanders.
61 * c-pretty-print.c (pp_c_expression): Handle GOTO_EXPR like
64 2009-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
66 * pa.c (last_address): Change to unsigned.
67 (update_total_code_bytes): Change argument to unsigned. Don't
68 check if insn addresses are set.
69 (pa_output_function_epilogue): Set last_address to UINT_MAX if insn
70 addresses are not set.
71 (pa_asm_output_mi_thunk): Handle wrap when updating last_address.
73 2009-01-09 Nick Clifton <nickc@redhat.com>
75 * config/sh/symbian.c: Replace uses of DECL_INLINE with
76 DECL_DECLARED_INLINE_P.
78 2009-01-09 Jakub Jelinek <jakub@redhat.com>
81 * dojump.c (do_jump_by_parts_zero_rtx): Use mode instead of
82 GET_MODE (op0) in operand_subword_force calls.
85 * fold-const.c (fold_unary): For COMPOUND_EXPR and COND_EXPR,
86 fold_convert arg0 operands to TREE_TYPE (op0) first.
88 2009-01-08 Vladimir Makarov <vmakarov@redhat.com>
90 * params.def (ira-max-conflict-table-size): Decrease default value
93 2009-01-08 Jakub Jelinek <jakub@redhat.com>
95 PR tree-optimization/37031
96 * lambda-code.c (lambda_collect_parameters): Call pointer_set_destroy
98 (build_access_matrix): Reserve correct size for AM_MATRIX vector,
99 allocate it using gc instead of heap, use VEC_quick_push instead of
101 * graphite.c (build_access_matrix): Allocate AM_MATRIX vector using gc
102 instead of heap, use VEC_quick_push instead of VEC_safe_push.
103 * tree-data-ref.h (struct access_matrix): Change matrix to gc
104 allocated vector from heap allocated.
105 * lambda.h: Add DEF_VEC_ALLOC_P for gc allocated lambda_vector.
106 * tree-loop-linear.c (linear_transform_loops): Allocate nest
107 vector only after perfect_loop_nest_depth call.
109 2009-01-08 Sebastian Pop <sebastian.pop@amd.com>
110 Jan Sjodin <jan.sjodin@amd.com>
112 PR tree-optimization/38559
113 * graphite.c (debug_value, copy_constraint,
114 swap_constraint_variables, scale_constraint_variable, ): New.
115 (get_lower_bound, get_upper_bound): Removed.
116 (graphite_trans_bb_strip_mine): Clean up this code that works
117 only for constant number of iterations. Fully copy upper and
118 lower bound constraints, not only the constant part of them.
119 * graphite.h (debug_value): Declared.
121 2009-01-08 Ira Rosen <irar@il.ibm.com>
123 PR tree-optimization/37194
124 * tree-vect-transform.c (vect_estimate_min_profitable_iters):
125 Don't add the cost of cost model guard in prologue to scalar
126 outside cost in case of known number of iterations.
128 2009-01-07 Nathan Froyd <froydnj@codesourcery.com>
129 Alan Modra <amodra@bigpond.net.au>
131 * config/rs6000/rs6000.c (rs6000_legitimize_address): Check for
132 non-word-aligned REG+CONST addressing.
134 2009-01-07 Uros Bizjak <ubizjak@gmail.com>
137 * config/alpha/alpha.c (alpha_end_function): For TARGET_ABI_OSF, call
138 free_after_compilation when outputting a thunk.
139 (alpha_output_mi_thunk_osf): Assert that we are processing a thunk.
140 Do not call free_after_compilation here.
142 2009-01-07 Uros Bizjak <ubizjak@gmail.com>
144 * config/i386/i386.c (ix86_target_string): Use ARRAY_SIZE.
145 (ix86_valid_target_attribute_inner_p): Ditto.
147 2009-01-07 Jan Sjodin <jan.sjodin@amd.com>
149 PR tree-optimization/38492
150 PR tree-optimization/38498
151 * tree-check.c (operator_is_linear, scev_is_linear_expression): New.
152 * tree-chrec.h (scev_is_linear_expression): Declared.
153 * graphite.c (graphite_cannot_represent_loop_niter): New.
154 (scopdet_basic_block_info): Call graphite_cannot_represent_loop_niter.
155 (graphite_loop_normal_form): Use gcc_assert.
156 (scan_tree_for_params): Use CASE_CONVERT.
157 (phi_node_is_iv, bb_contains_non_iv_scalar_phi_nodes): New.
158 (build_scop_conditions_1): Call bb_contains_non_iv_scalar_phi_nodes.
159 Use gcc_assert. Discard scops that contain unhandled cases.
160 (build_scop_conditions): Return a boolean status for unhandled cases.
161 (strip_mine_profitable_p): Print the loop number, not its depth.
162 (is_interchange_valid): Pass the depth of the loop nest, don't
163 recompute it wrongly.
164 (graphite_trans_bb_block): Same.
165 (graphite_trans_bb_block): Print tentative of loop blocking.
166 (graphite_trans_scop_block): Do not print that the loop has been
168 (graphite_transform_loops): Do not handle scops that contain condition
171 2009-01-07 H.J. Lu <hongjiu.lu@intel.com>
173 AVX Programming Reference (December, 2008)
174 * config/i386/avxintrin.h (_mm256_stream_si256): New.
175 (_mm256_stream_pd): Likewise.
176 (_mm256_stream_ps): Likewise.
178 * config/i386/i386.c (ix86_builtins): Add IX86_BUILTIN_MOVNTDQ256,
179 IX86_BUILTIN_MOVNTPD256 and IX86_BUILTIN_MOVNTPS256.
180 (ix86_special_builtin_type): Add VOID_FTYPE_PV4DI_V4DI.
181 (bdesc_special_args): Add __builtin_ia32_movntdq256,
182 __builtin_ia32_movntpd256 and __builtin_ia32_movntps256.
183 (ix86_init_mmx_sse_builtins): Handle VOID_FTYPE_PV4DI_V4DI.
184 (ix86_expand_special_args_builtin): Likewise.
186 * config/i386/sse.md (AVXMODEDI): New.
187 (avx_movnt<mode>): Likewise.
188 (avx_movnt<mode>): Likewise.
189 (<sse>_movnt<mode>): Remove AVX support.
190 (sse2_movntv2di): Likewise.
192 2009-01-07 Richard Guenther <rguenther@suse.de>
195 * fold-const.c (extract_muldiv): Remove obsolete comment.
196 (fold_plusminus_mult_expr): Undo MINUS_EXPR
197 to PLUS_EXPR canonicalization for the canonicalization.
199 2009-01-07 Gerald Pfeifer <gerald@pfeifer.com>
201 * doc/install.texi (alpha*-dec-osf*): Remove note on 32-bit
202 hosted cross-compilers generating less efficient code.
204 2009-01-06 Richard Sandiford <rdsandiford@googlemail.com>
206 * function.h (rtl_data): Add a dbr_scheduled_p field.
207 * reorg.c (dbr_schedule): Set it.
208 (gate_handle_delay_slots): Check it.
209 * config/mips/mips.c (mips_base_delayed_branch): Delete.
210 (mips_reorg): Check flag_delayed_branch instead of
211 mips_base_delayed_branch.
212 (mips_override_options): Don't set mips_base_delayed_branch
213 or flag_delayed_branch.
215 2009-01-06 Richard Sandiford <rdsandiford@googlemail.com>
217 PR rtl-optimization/38426.
218 * ira.c (ira): Set current_function_is_leaf earlier.
220 2009-01-06 Jakub Jelinek <jakub@redhat.com>
222 PR rtl-optimization/38722
223 * combine.c (try_combine): Don't modify PATTERN (i3) and notes
224 too early, only set a flag and modify after last possible
227 2009-01-06 Janis Johnson <janis187@us.ibm.com>
230 * ginclude/float.h: Rename DECnn_DEN to DECnn_SUBNORMAL_MIN.
231 * real.c (decimal_single_format): Correct values of emin and emax.
232 (decimal_double_format): Ditto.
233 (decimal_quad_format): Ditto.
234 * c-cppbuiltin.c (builtin_define_decimal_float_constants): Adjust
235 computation of DECnn_MIN and DECnn_MAX for corrected values of
236 emin and emax. Define __DECnn_SUBNORMAL_MIN__ instead of
237 __DECnn_MIN__, and adjust its computation for the corrected value
240 2009-01-06 Jan Hubicka <jh@suse.cz>
243 * i386.c (ix86_expand_call): Use ARRAY_SIZE.
245 2009-01-06 Gerald Pfeifer <gerald@pfeifer.com>
247 * doc/contrib.texi (Contributors): Slightly adjust the end note.
248 Add Robert Clark to the list of testers.
250 2009-01-06 Jan Hubicka <jh@suse.cz>
251 Kai Tietz <kai.tietz@onevision.com>
253 * i386.md (*msabi_syvabi): Add SSE regs clobbers.
254 * i386.c (ix86_expand_call): Add clobbers.
256 2009-01-06 Jan Hubicka <jh@suse.cz>
257 Kai Tietz <kai.tietz@onevision.com>
259 * i386.h (CONDITIONAL_CALL_USAGE): SSE regs are not used for w64 ABI.
260 * i386.c (struct ix86_frame): Add padding0 and nsseregs.
261 (ix86_nsaved_regs): Count only general purpose regs.
262 (ix86_nsaved_sseregs): New.
263 (ix86_compute_frame_layout): Update nsseregs; set preferred alignment
264 to 16 for w64; compute padding and size of sse reg save area.
265 (ix86_emit_save_regs, ix86_emit_save_regs_using_mov): Save only
266 general purpose regs.
267 (ix86_emit_save_sse_regs_using_mov): New.
268 (ix86_expand_prologue): Save SSE regs if needed.
269 (ix86_emit_restore_regs_using_mov): Use only general purpose regs.
270 (ix86_emit_restore_sse_regs_using_mov): New.
271 (ix86_expand_epilogue): Save SSE regs if needed.
273 2009-01-06 Jan Hubicka <jh@suse.cz>
274 Kai Tietz <kai.tietz@onevision.com>
276 * i386.h (ACCUMULATE_OUTGOING_ARGS): Enable for MSABI
277 * i386.c (init_cumulative_args): Disallow calls of MSABI functions
278 when accumulate outgoing args is off.
280 2009-01-06 H.J. Lu <hongjiu.lu@intel.com>
283 * ira-color.c (ira_reuse_stack_slot): Check ENABLE_IRA_CHECKING
284 before using pseudos_have_intersected_live_ranges_p.
286 * ira-int.h (ira_assert): Always define.
288 2009-01-06 H.J. Lu <hongjiu.lu@intel.com>
290 AVX Programming Reference (December, 2008)
291 * config/i386/avxintrin.h (_mm_permute2_pd): Removed.
292 (_mm256_permute2_pd): Likewise.
293 (_mm_permute2_ps): Likewise.
294 (_mm256_permute2_ps): Likewise.
295 * config/i386/i386.md (UNSPEC_VPERMIL2): Likewise.
296 * config/i386/sse.md (avx_vpermil2<mode>3): Likewise.
298 * config/i386/i386.c (ix86_builtins): Remove
299 IX86_BUILTIN_VPERMIL2PD, IX86_BUILTIN_VPERMIL2PS,
300 IX86_BUILTIN_VPERMIL2PD256 and IX86_BUILTIN_VPERMIL2PS256.
301 (ix86_builtin_type): Remove V8SF_FTYPE_V8SF_V8SF_V8SI_INT,
302 V4DF_FTYPE_V4DF_V4DF_V4DI_INT, V4SF_FTYPE_V4SF_V4SF_V4SI_INT
303 and V2DF_FTYPE_V2DF_V2DF_V2DI_INT.
304 (bdesc_args): Remove __builtin_ia32_vpermil2pd,
305 __builtin_ia32_vpermil2ps, __builtin_ia32_vpermil2pd256 and
306 __builtin_ia32_vpermil2ps256.
307 (ix86_init_mmx_sse_builtins): Updated.
308 (ix86_expand_args_builtin): Likewise.
310 2009-01-05 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
312 * pa.c (output_call): Relocate non-jump insns in the delay slot of
313 long absolute calls when generating PA 2.0 code.
315 2009-01-05 Vladimir Makarov <vmakarov@redhat.com>
317 PR rtl-optimization/38583
318 * params.h (IRA_MAX_CONFLICT_TABLE_SIZE): New macro.
320 * params.def (ira-max-conflict-table-size): New.
322 * doc/invoke.texi (ira-max-conflict-table-size): Decribe.
324 * ira.h (ira_conflicts_p): New external definition.
326 * ira-conflicts.c (build_conflict_bit_table): Do not build too big
327 table. Report this. Return result of building.
328 (ira_build_conflicts): Use ira_conflicts_p. Check result of
329 building conflict table.
331 * ira-color.c (fast_allocation): Use num instead of ira_allocnos_num.
332 (ira_color): Use ira_conflicts_p.
334 * global.c: Include ira.h.
335 (pseudo_for_reload_consideration_p, build_insn_chain): Use
338 * Makefile.in (global.o): Add ira.h.
340 * ira-build.c (mark_all_loops_for_removal,
341 propagate_some_info_from_allocno): New.
342 (remove_unnecessary_allocnos): Call
343 propagate_some_info_from_allocno.
344 (remove_low_level_allocnos): New.
345 (remove_unnecessary_regions): Add parameter. Call
346 mark_all_loops_for_removal and remove_low_level_allocnos. Pass
347 parameter to remove_unnecessary_regions.
348 (ira_build): Remove all regions but root if the conflict table was
349 not built. Update conflict hard regs for allocnos crossing calls.
351 * ira.c (ira_conflicts_p): New global.
352 (ira): Define and use ira_conflicts_p.
354 * reload1.c (compute_use_by_pseudos, reload, count_pseudo,
355 count_spilled_pseudo, find_reg, alter_reg, finish_spills,
356 emit_input_reload_insns, delete_output_reload): Use ira_conflicts_p.
358 2009-01-06 Ben Elliston <bje@au.ibm.com>
360 * gengtype-lex.l (YY_NO_INPUT): Define.
362 2009-01-05 Andrew Pinski <andrew_pinski@playstation.sony.com>
365 * c-common.c (handle_vector_size_attribute): Also reject
368 2009-01-05 Sebastian Pop <sebastian.pop@amd.com>
370 PR tree-optimization/38492
371 * graphite.c (rename_map_elt, debug_rename_elt,
372 debug_rename_map_1, debug_rename_map, new_rename_map_elt,
373 rename_map_elt_info, eq_rename_map_elts,
374 get_new_name_from_old_name, bb_in_sese_p): Moved around.
375 (sese_find_uses_to_rename_use): Renamed sese_build_livein_liveouts_use.
376 (sese_find_uses_to_rename_bb): Renamed sese_build_livein_liveouts_bb.
377 (sese_build_livein_liveouts): New.
378 (new_sese, free_sese): New.
379 (new_scop): Call new_sese.
380 (free_scop): Call free_sese.
381 (rename_variables_from_edge, rename_phis_end_scop): Removed.
382 (register_old_new_names): Renamed register_old_and_new_names.
383 (register_scop_liveout_renames, add_loop_exit_phis,
384 insert_loop_close_phis, struct igp,
385 default_liveout_before_guard, add_guard_exit_phis,
386 insert_guard_phis, copy_renames): New.
387 (translate_clast): Call insert_loop_close_phis and insert_guard_phis.
388 (sese_add_exit_phis_edge): Renamed scop_add_exit_phis_edge.
389 (rewrite_into_sese_closed_ssa): Renamed scop_insert_phis_for_liveouts.
390 (scop_adjust_phis_for_liveouts): New.
391 (gloog): Call scop_adjust_phis_for_liveouts.
393 * graphite.h (struct sese): Documented. Added fields liveout,
395 (SESE_LIVEOUT, SESE_LIVEIN, SESE_LIVEIN_VER, SESE_NUM_VER): New.
396 (new_sese, free_sese, sese_build_livein_liveouts): Declared.
397 (struct scop): Added field liveout_renames.
398 (SCOP_LIVEOUT_RENAMES): New.
400 2009-01-05 Harsha Jagasia <harsha.jagasia@amd.com>
402 PR tree-optimization/38510
403 * graphite.c (recompute_all_dominators): Call mark_irreducible_loops.
404 (translate_clast): Call recompute_all_dominators before
406 (gloog): Call recompute_all_dominators before graphite_verify.
408 2009-01-05 Harsha Jagasia <harsha.jagasia@amd.com>
409 Jan Sjodin <jan.sjodin@amd.com>
411 PR tree-optimization/38500
412 * graphite.c (create_sese_edges): Call fix_loop_structure after
415 2009-01-05 Joel Sherrill <joel.sherrill@oarcorp.com>
417 * config.gcc: Add m32r*-*-rtems*.
418 * config/m32r/rtems.h: New file.
420 2009-01-05 Ben Elliston <bje@au.ibm.com>
422 * Makefile.in (.po.gmo): Use mkinstalldirs, not test -d || mkdir.
424 (po/gcc.pot): Likewise.
426 2009-01-04 David S. Miller <davem@davemloft.net>
428 * config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED_RTX): Delete.
429 (STARTING_FRAME_OFFSET): Always set to zero.
431 2009-01-04 Richard Sandiford <rdsandiford@googlemail.com>
433 * tree.def (LSHIFT_EXPR, RSHIFT_EXPR): Add commentary.
434 * tree-cfg.c (verify_gimple_assign_binary): Allow shifts of
435 fixed-point types, and vectors of the same.
437 2009-01-04 Richard Sandiford <rdsandiford@googlemail.com>
439 * config/mips/sync.md (*mb_barrier): Rename to...
440 (*memory_barrier): ...this.
442 2009-01-04 Jonathan Wakely <jwakely.gcc@gmail.com>
444 * doc/extend.texi (Function Attributes): Move @cindex after @item
445 for 'artificial' and 'flatten'. Fix grammar for 'externally_visible'
446 and put in alphabetical order. Fix 'target' name and put in order.
447 * doc/invoke.texi (-Wstrict-null-sentinel, -fipa-matrix-reorg): Fix
450 2009-01-04 Uros Bizjak <ubizjak@gmail.com>
452 * config/s390/s390.md (UNSPEC_MB): Rename from UNSPECV_MB.
453 (memory_barrier): Expand as unspec instead of unspec_volatile.
454 Remove mem:BLK from insn operands. Use Pmode scratch register.
455 (*memory_barrier): Define as unspec instead of unspec_volatile.
456 Use (match_dup 0) as input operand.
458 * config/sparc/sparc.md (UNSPEC_MEMBAR): Rename from UNSPECV_MEMBAR.
459 * config/sparc/sync.md (memory_barrier): Expand as unspec instead of
460 unspec_volatile. Remove mem:BLK from insn operands. Use Pmode
461 scratch register. Remove operand 1.
462 (*stbar): Define as unspec instead of unspec_volatile.
463 Use (match_dup 0) as input operand, remove (const_int 8).
464 (*membar): Define as unspec instead of unspec_volatile.
465 Use (match_dup 0) as input operand, remove input operand 2.
467 * config/xtensa/xtensa.md (UNSPEC_MEMW): Rename from UNSPECV_MEMW.
468 (memory_barrier): Expand as unspec instead of unspec_volatile.
469 Remove mem:BLK from insn operands. Use Pmode scratch register.
470 (*memory_barrier): Define as unspec instead of unspec_volatile.
471 Use (match_dup 0) as input operand.
473 * config/ia64/sync.md (memory_barrier): Redefine as expander pattern.
474 Remove mem:BLK from insn operands. Use Pmode scratch register.
475 Set volatile flag on operand 0.
476 (*memory_barrier): New insn pattern.
478 * config/rs6000/sync.md (memory_barrier): Remove mem:BLK from
480 (*memory_barrier): Use (match_dup 0) as input operand.
482 * config/mips/sync.md (memory_barrier): Redefine as expander pattern.
483 Remove mem:BLK from insn operands. Use Pmode scratch register.
484 Set volatile flag on operand 0.
485 (*mb_internal): New insn pattern.
487 * config/alpha/sync.md (*memory_barrier): Rename from *mb_internal.
489 2009-01-04 Steven Bosscher <steven@gcc.gnu.org>
492 * function.c (struct temp_slot): Move to the section of the file
493 that deals with temp slots. Remove field 'address'.
494 (temp_slot_address_table): New hash table of address -> temp slot.
495 (struct temp_slot_address_entry): New struct, items for the table.
496 (temp_slot_address_compute_hash, temp_slot_address_hash,
497 temp_slot_address_eq, insert_temp_slot_address): Support functions
499 (find_temp_slot_from_address): Rewrite to use the new hash table.
500 (remove_unused_temp_slot_addresses): Remove addresses of temp
501 slots that have been made available.
502 (remove_unused_temp_slot_addresses_1): Call-back for htab_traverse,
503 worker function for remove_unused_temp_slot_addresses.
504 (assign_stack_temp_for_type): Don't clear the temp slot address list.
505 Add the temp slot address to the address -> temp slot map.
506 (update_temp_slot_address): Update via insert_temp_slot_address.
507 (free_temp_slots): Call remove_unused_temp_slot_addresses.
508 (pop_temp_slots): Likewise.
509 (init_temp_slots): Allocate the address -> temp slot map, or empty
510 the map if it is already allocated.
511 (prepare_function_start): Initialize temp slot processing.
513 2009-01-04 Steven Bosscher <steven@gcc.gnu.org>
516 * cfgexpand.c (estimate_stack_frame_size): Simplify the estimate:
517 Calculate the size of all stack vars assuming no packing of stack
518 vars will happen, replacing a quadratic algorithm with a linear one.
520 2009-01-03 Jakub Jelinek <jakub@redhat.com>
523 * expmed.c (store_bit_field_1): Don't modify op0 if movstrict insn
526 2009-01-03 Diego Novillo <dnovillo@google.com>
528 * doc/contrib.texi: Update contributions.
530 2009-01-03 Jakub Jelinek <jakub@redhat.com>
533 * builtins.c (fold_builtin_memory_op): Give up if either operand
534 is volatile. Set srctype or desttype to non-qualified version
538 * builtins.c (fold_builtin_expect): Only check DECL_WEAK for VAR_DECLs
541 2009-01-02 Kenneth Zadeck <zadeck@naturalbridge.com>
543 PR rtl-optimization/35805
544 * df-problems.c (df_lr_finalize): Add recursive call to resolve lr
545 problem if fast dce is able to remove any instructions.
546 * dce.c (dce_process_block): Fix dump message.
548 2009-01-02 Mark Mitchell <mark@codesourcery.com>
551 * tree-ssa-pre.c (compute_antic): Correct loop bounds.
553 2009-01-02 Jakub Jelinek <jakub@redhat.com>
556 * tree-flow.h (op_code_prio, op_prio): New prototypes.
557 * tree-pretty-print.c (op_code_prio): New function.
558 (op_prio): No longer static. Use op_code_prio.
559 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs):
560 Use op_prio and op_code_prio to determine if () should be
561 printed around operand(s) or not.
563 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
564 dump_gimple_call, dump_gimple_switch, dump_gimple_cond,
565 dump_gimple_label, dump_gimple_try, dump_symbols, dump_gimple_phi,
566 dump_gimple_mem_ops, dump_bb_header, dump_bb_end, pp_cfg_jump): Use
567 pp_character instead of pp_string for single letter printing.
569 2009-01-02 Richard Sandiford <rdsandiford@googlemail.com>
571 * doc/extend.texi: Fix '#pragma GCC option' typo.
573 2009-01-02 Richard Guenther <rguenther@suse.de>
575 * doc/install.texi (--enable-checking): Mention different
577 (--enable-stage1-checking): Document.
579 2009-01-01 Andrew Pinski <pinskia@gmail.com>
582 * tree-cfg.c (verify_expr): Add INDIRECT_REF case. Change MODIFY_EXPR
585 2009-01-02 Ben Elliston <bje@au.ibm.com>
587 * config/fp-bit.h (pack_d): Constify argument.
588 * config/fp-bit.c (makenan): Constify return type. Remove casts.
589 (isnan): Constify argument.
593 (_fpadd_parts): Constify return type.
594 (_fpmul_parts): Likewise.
595 (_fpdiv_parts): Likewise.
597 2009-01-01 Jakub Jelinek <jakub@redhat.com>
600 * c-typeck.c (add_pending_init): Add IMPLICIT argument. Only
601 warn about overwriting initializer with side-effects or
602 -Woverride-init if !IMPLICIT.
603 (output_init_element): Likewise. Pass IMPLICIT down to
605 (process_init_element): Add IMPLICIT argument. Pass it down
606 to output_init_element.
607 (push_init_element, pop_init_level, set_designator): Adjust
608 process_init_element callers.
609 (set_nonincremental_init, set_nonincremental_init_from_string):
610 Adjust add_pending_init callers.
611 (output_pending_init_elements): Adjust output_init_element callers.
612 * c-tree.h (process_init_element): Adjust prototype.
613 * c-parser.c (c_parser_initelt, c_parser_initval): Adjust
614 process_init_element callers.