Ian Lance Taylor [Wed, 13 Apr 2011 21:00:59 +0000 (21:00 +0000)]
Unify handling of runtime support functions.
This introduces the new approach, and rewrites the lowering
code which uses runtime functions. The code which calls
runtime functions at GENERIC conversion time is not yet
rewritten.
tree-flow.h (struct gimple_df): Make free_ssanames a VEC.
* tree-flow.h (struct gimple_df): Make free_ssanames a VEC.
* tree-ssanames.c (fini_ssanames): VEC_free it.
(make_ssa_name_fn): Update for VECness of free_ssanames.
(release_ssa_name, release_dead_ssa_names): Likewise.
* tree.h (struct tree_ssa_name): Include tree_typed instead of
tree_common.
* tree.c (initialize_tree_contains_struct): Mark TS_SSA_NAME as
TS_TYPED instead of TS_COMMON.
Jan Hubicka [Wed, 13 Apr 2011 17:26:50 +0000 (19:26 +0200)]
ipa-inline.h: New file.
* ipa-inline.h: New file.
* ipa-inline-analysis.c: New file. Broken out of ...
* ipa-inline.c: ... this file; update toplevel comment;
include ipa-inline.h
(inline_summary): Move to ipa-inline.h
(cgraph_estimate_edge_time): Rename to estimate_edge_time; move to
ipa-inline-analysis.c.
(cgraph_estimate_time_after_inlining): Rename to estiamte_time_after_inlining;
move to ipa-inline-analysis.c
(cgraph_estimate_edge_growth): Move to ipa-inline-analysis.c; rename
to estimate_edge_growth.
(cgraph_estimate_size_after_inlining): Move to ipa-inline-analysis.c;
rename to estimate_size_after_inlining.
(cgraph_mark_inline_edge): Update for new naming convention.
(cgraph_check_inline_limits): Likewise.
(cgraph_edge_badness): Likewise.
(cgraph_decide_recursive_inlining): Likewise.
(cgraph_decide_inlining_of_small_functions): Likewise.
(cgraph_decide_inlining_incrementally): Likewise.
(cgraph_estimate_growth): Rename to estimate_growth; move to ipa-inline-analysis.c.
(eliminated_by_inlining_prob): Move to ipa-inline-analysis.c.
(estimate_function_body_sizes): Move to ipa-inline-analysis.c.
(compute_inline_parameters): Likewise.
(compute_inline_parameters_for_current): Likewise.
(pass_inline_parameters): Likewise.
(inline_indirect_intraprocedural_analysis): Likewise.
(analyze_function): Rename to inline_analyze_function; likewise.
(add_new_function): Move to ipa-inline-analysis.c.
(inline_generate_summary): Likewise.
(inline_read_summary): Likewise.
(inline_write_summary): Likewise.
* Makefile.in (ipa-inline-analysis.c): New file.
tree.h (struct tree_int_cst, [...]): Include tree_typed instead of tree_common.
* tree.h (struct tree_int_cst, struct real_value): Include tree_typed
instead of tree_common.
(struct tree_fixed_cst, struct tree_string, struct tree_complex):
Likewise.
* tree.c (initialize_tree_contains_struct): Mark such nodes as being
TS_TYPED rather than TS_COMMON.
* print-tree.c (print_node) [STRING_CST]: Don't print TREE_CHAIN.
Jakub Jelinek [Wed, 13 Apr 2011 15:50:02 +0000 (17:50 +0200)]
re PR middle-end/48591 (OpenMP ICE with atomics on __float128 in 32-bit only i686 cc1)
PR middle-end/48591
* omp-low.c (expand_omp_atomic_fetch_op): Return false if decl is
NULL.
(expand_omp_atomic_pipeline): Return false if cmpxchg is NULL.
dwarf2out.c (struct dw_cfi_struct): Remove member dw_cfi_next.
* dwarf2out.c (struct dw_cfi_struct): Remove member dw_cfi_next.
(dw_cfi_ref): Add DEF_VEC_P and some DEF_VEC_ALLOC_Ps.
(cfi_vec): New typedef.
(struct dw_fde_struct): Make dw_fde_cfi a cfi_vec. Replace
dw_fde_switch_cfi with an integer dw_fde_switch_cfi_index.
(cie_cfi_vec): New static variable.
(cie_cfi_head): Delete.
(add_cfi): Accept a cfi_vec * as first argument. All callers and
declaration changed. Use vector rather than list operations.
(new_cfi): Don't initialize the dw_cfi_next field.
(add_fde_cfi): Allocate cie_cfi_vec if necessary. Use vector
rather than list operations.
(lookup_cfa): Use vector rather than list operations.
(output_cfis): New argument upto. Accept a cfi_vec rather than
a dw_cfi_ref list head as argument. All callers changed.
Iterate over the vector using upto as a maximum index.
(output_all_cfis): New static function.
(output_fde): Use vector rather than list operations. Use the
new upto argument for output_cfis rather than manipulating a
list.
(dwarf2out_begin_prologue): Change initializations to match
new struct members.
(dwarf2out_switch_text_section): Initialize dw_fde_switch_cfi_index
from the vector length rather than searching for the end of a list.
Use output_all_cfis.
(convert_cfa_to_fb_loc_list): Use vector rather than list operations.
* config/i386/sse.md (pinsrbits): Remove.
(sse2_packsswb): Merge with *avx_packsswb.
(sse2_packssdw): Merge with *avx_packssdw.
(sse2_packuswb): Merge with *avx_packuswb.
(vec_interleave_highv16qi): Merge with *avx_interleave_highv16qi.
(vec_interleave_lowv16qi): Merge with *avx_interleave_lowv16qi.
(vec_interleave_highv8hi): Merge with *avx_interleave_highv8hi.
(vec_interleave_lowv8hi): Merge with *avx_interleave_lowv8hi.
(vec_interleave_highv4si): Merge with *avx_interleave_highv4si.
(vec_interleave_lowv4si): Merge with *avx_interleave_lowv4si.
(*sse4_1_pinsrb): Merge with *avx_pinsr<ssevecsize>.
(*sse2_pinsrw): Merge with *avx_pinsr<ssevecsize>.
(*sse4_1_pinsrd): Merge with *avx_pinsr<ssevecsize>.
(*sse4_1_pinsrq): Merge with *avx_pinsrq.
(sse2_loadld): Merge with *avx_loadld.
(*vec_extractv2di_1_rex64): Merge with *vec_extractv2di_1_rex64_avx.
(*vec_extractv2di_1_sse2): Merge with *vec_extractv2di_1_avx.
(*vec_concatv2si_sse4_1): Merge with *vec_concatv2si_avx.
(*vec_concatv2di_rex64_sse4_1): Merge with *vec_concatv2di_rex64_avx.
(vec_concatv2di): Merge with *vec_concatv2di_avx.
In gcc/objc/:
2011-04-12 Nicola Pero <nicola.pero@meta-innovation.com>
* objc-act.c (objc_build_message_expr): Accept two arguments
instead of one so that callers can simply pass the arguments
without having to create a temporary chain to hold them.
In gcc/cp/:
2011-04-12 Nicola Pero <nicola.pero@meta-innovation.com>
* parser.c (cp_parser_objc_message_expression): Updated call
to objc_build_message_expr.
Nicola Pero [Tue, 12 Apr 2011 18:53:08 +0000 (18:53 +0000)]
In gcc/objc/: 2011-04-12 Nicola Pero <nicola.pero@meta-innovation.com>
In gcc/objc/:
2011-04-12 Nicola Pero <nicola.pero@meta-innovation.com>
* objc-act.c (comp_proto_with_proto): Do not create and use
inefficient temporary argument lists. Compare the arguments
directly. (match_proto_with_proto): Removed; incorporated into
comp_proto_with_proto ().
Jakub Jelinek [Tue, 12 Apr 2011 17:49:11 +0000 (19:49 +0200)]
combine.c (update_cfg_for_uncondjump): Instead of testing at_end assert it is always true.
* combine.c (update_cfg_for_uncondjump): Instead of testing at_end
assert it is always true.
(try_combine): Don't call update_cfg_for_uncondjump for noop non-jump
moves.
Nicola Pero [Tue, 12 Apr 2011 17:37:39 +0000 (17:37 +0000)]
In gcc/objc/: 2011-04-12 Nicola Pero <nicola.pero@meta-innovation.com>
In gcc/objc/:
2011-04-12 Nicola Pero <nicola.pero@meta-innovation.com>
* objc-act.c (printable_ivar_name): New.
(add_instance_variable): Call printable_ivar_name() when an error
message needs to be printed. Do not prepare the instance variable
for printing unless there is an actual error.
recog.h (insn_operand_data): Add an "allows_mem" field.
gcc/
* recog.h (insn_operand_data): Add an "allows_mem" field.
* genoutput.c (output_operand_data): Initialize it.
* optabs.c (maybe_legitimize_operand_same_code): New function.
(maybe_legitimize_operand): Use it when matching the original
op->value.
genpreds.c (process_define_predicate): Move most processing to gensupport.c.
gcc/
* genpreds.c (process_define_predicate): Move most processing
to gensupport.c. Continue to validate the expression.
* genrecog.c (did_you_mean_codes, compute_predicate_codes)
(process_define_predicate): Move processing to gensupport.c.
(main): Remove DEFINE_PREDICATE and DEFINE_SPECIAL_PREDICATE cases.
* gensupport.c (did_you_mean_codes): Moved from genrecog.c.
(compute_predicate_codes): Moved from genrecog.c. Add lineno
argument.
(valid_predicate_name_p): New function, split out from old
genpreds.c:process_define_predicate.
(process_define_predicate): New function, combining code from
old genpreds.c and genrecog.c functions.
(process_rtx): Call it for DEFINE_PREDICATE and
DEFINE_SPECIAL_PREDICATE.
arm.c (arm_print_operand): Use MEM_SIZE to get the size of a '%A' memory reference.
gcc/
* config/arm/arm.c (arm_print_operand): Use MEM_SIZE to get the
size of a '%A' memory reference.
(T_DREG, T_QREG): New neon_builtin_type_bits.
(arm_init_neon_builtins): Assert that the load and store operands
are neon_struct_operands.
(locate_neon_builtin_icode): Provide the neon_builtin_type_bits.
(NEON_ARG_MEMORY): New builtin_arg.
(neon_dereference_pointer): New function.
(arm_expand_neon_args): Add a neon_builtin_type_bits argument.
Handle NEON_ARG_MEMORY.
(arm_expand_neon_builtin): Update after above interface changes.
Use NEON_ARG_MEMORY for loads and stores.
* config/arm/predicates.md (neon_struct_operand): New predicate.
* config/arm/iterators.md (V_two_elem): Tweak formatting.
(V_three_elem): Use BLKmode for accesses that have no associated mode.
(V_four_elem): Tweak formatting.
* config/arm/neon.md (neon_vld1<mode>, neon_vld1_dup<mode>)
(neon_vst1_lane<mode>, neon_vst1<mode>, neon_vld2<mode>)
(neon_vld2_lane<mode>, neon_vld2_dup<mode>, neon_vst2<mode>)
(neon_vst2_lane<mode>, neon_vld3<mode>, neon_vld3_lane<mode>)
(neon_vld3_dup<mode>, neon_vst3<mode>, neon_vst3_lane<mode>)
(neon_vld4<mode>, neon_vld4_lane<mode>, neon_vld4_dup<mode>)
(neon_vst4<mode>): Replace pointer operand with a memory operand.
Use %A in the output template.
(neon_vld3qa<mode>, neon_vld3qb<mode>, neon_vst3qa<mode>)
(neon_vst3qb<mode>, neon_vld4qa<mode>, neon_vld4qb<mode>)
(neon_vst4qa<mode>, neon_vst4qb<mode>): Likewise, but halve
the width of the memory access. Remove post-increment.
* config/arm/neon-testgen.ml: Allow addresses to have an alignment.
gcc/testsuite/
* gcc.target/arm/neon-vld3-1.c: New test.
* gcc.target/arm/neon-vst3-1.c: New test.
* gcc.target/arm/neon/v*.c: Regenerate.
Nick Clifton [Tue, 12 Apr 2011 11:34:26 +0000 (11:34 +0000)]
v850.c (expand_prologue): Do not use the CALLT instruction for interrupt handlers if...
* config/v850/v850.c (expand_prologue): Do not use the CALLT
instruction for interrupt handlers if the target is the basic V850
architecture.
(expand_epilogue): Likewise.
Jakub Jelinek [Tue, 12 Apr 2011 10:53:47 +0000 (12:53 +0200)]
re PR rtl-optimization/48549 (Combiner ICE with -g)
PR rtl-optimization/48549
* combine.c (propagate_for_debug): Also stop after BB_END of
this_basic_block. Process LAST and just stop processing after it.
(combine_instructions): If last_combined_insn has been deleted,
set last_combined_insn to its PREV_INSN.
Richard Guenther [Tue, 12 Apr 2011 10:44:15 +0000 (10:44 +0000)]
re PR tree-optimization/46076 (constant propagation and compile-time math no longer happening versus 4.4 and 4.5)
2011-04-12 Richard Guenther <rguenther@suse.de>
PR tree-optimization/46076
* gimple.h (struct gimple_statement_call): Add fntype field.
(gimple_call_fntype): Adjust.
(gimple_call_set_fntype): New function.
* gimple.c (gimple_build_call_1): Set the call function type.
* gimplify.c (gimplify_call_expr): Preserve the function
type the frontend used for the call.
(gimplify_modify_expr): Likewise.
* lto-streamer-in.c (input_gimple_stmt): Input the call stmts
function type.
* lto-streamer-out.c (output_gimple_stmt): Output the call stmts
function type.
* tree-ssa.c (useless_type_conversion_p): Function pointer
conversions are useless.
Jakub Jelinek [Tue, 12 Apr 2011 06:27:23 +0000 (08:27 +0200)]
re PR c/48552 (ICE with void type expressions in asm inputs/outputs)
PR c/48552
* c-typeck.c (build_asm_expr): Error out on attempts to use
void type outputs or inputs for constraints that allow reg or
don't allow memory.
* gcc.dg/pr48552-1.c: New test.
* gcc.dg/pr48552-2.c: New test.
re PR target/48250 (ICE in reload_cse_simplify_operands, at postreload.c:403)
2011-04-11 Chung-Lin Tang <cltang@codesourcery.com>
Richard Earnshaw <rearnsha@arm.com>
PR target/48250
* config/arm/arm.c (arm_legitimize_reload_address): Update cases
to use sign-magnitude offsets. Reject unsupported unaligned
cases. Add detailed description in comments.
* config/arm/arm.md (reload_outdf): Disable for ARM mode; change
condition from TARGET_32BIT to TARGET_ARM.
Co-Authored-By: Richard Earnshaw <rearnsha@arm.com>
From-SVN: r172297
* tree.h (struct typed_tree): New.
(struct tree_common): Include it instead of tree_base.
(TREE_TYPE): Update for new location of type field.
(TYPE_USER_ALIGN, TYPE_PACKED): Refer to base field directly.
(DECL_USER_ALIGN, DECL_PACKED): Likewise.
(union tree_node): Add typed field.
* treestruct.def (TS_TYPED): New.
* lto-streamer.c (check_handled_ts_structures): Handle it.
* tree.c (MARK_TS_TYPED): New macro.
(MARK_TS_COMMON): Call it instead of MARK_TS_BASE.
* config/i386/sse.md (VI): New mode iterator.
(SSEMODEI): Remove.
(AVX256MODEI): Ditto.
(AVXMODEF4P): Ditto.
(avxvecpsmode): Ditto.
(one_cmpl<mode>2): Enable for TARGET_SSE. Use VI mode iterator.
(sse2_andnot<mode>3): New expander.
(*andnot<mode>3): Merge with *sse2_andnot<mode>3 and
*avx_andnot<mode>3. Enable for TARGET_SSE. Use VI mode iterator.
(<any_logic:code><mode>3): Use VI mode iterator.
(*<any_logic:code><mode>3): Merge with *sse2_<any_logic:code><mode>3
and *avx_<any_logic:code><mode>3. Use VI mode iterator.
(*andnottf3): Handle AVX three-operand constraints.
(*<any_logic:code>tf3): Handle AVX three-operand constraints.
Paolo Carlini [Mon, 11 Apr 2011 17:50:39 +0000 (17:50 +0000)]
moveable.cc: Actually run it in parallel-mode for check-parallel.
2011-04-11 Paolo Carlini <paolo.carlini@oracle.com>
* testsuite/25_algorithms/inplace_merge/moveable.cc: Actually run
it in parallel-mode for check-parallel.
* testsuite/25_algorithms/rotate/moveable.cc: Likewise.
* testsuite/25_algorithms/stable_partition/moveable.cc: Likewise.
* testsuite/25_algorithms/remove/moveable.cc: Likewise.
* testsuite/25_algorithms/sort_heap/check_compare_by_value.cc:
Likewise.
* testsuite/25_algorithms/unique/moveable.cc: Likewise.
* testsuite/25_algorithms/remove_if/moveable.cc: Likewise.
* testsuite/25_algorithms/heap/moveable.cc: Likewise.
* testsuite/25_algorithms/next_permutation/moveable.cc: Only skip
the tests based on std::lexicographical_compare for check-parallel.
* testsuite/25_algorithms/prev_permutation/moveable.cc: Likewise.
* config/i386/sse.md (VI_128): New mode iterator.
(VI12_128): Rename from SSEMODE12.
(VI14_128): Rename from SSEMODE14.
(VI124_128): New mode iterator.
(VI24_128): Rename from SSEMODE248.
(VI248_128): Rename from SSEMODE248.
(SSEMODE124C8): Remove.
(*<plusminus_insn><mode>3): Merge with *avx_<plusminus_insn><mode>3.
(*sse2_<plusminus_insn><mode>3): Merge with
*avx_<plusminus_insn><mode>3.
(*mulv8hi3): Merge with *avx_mulv8hi3.
(*<s>mulv8hi3_highpart): Merge with *avx_<s>mulv8hi3_highpart.
(*sse2_umulv2siv2di3): Merge with *avx_umulv2siv2di3.
(*sse2_pmaddwd): Merge with *avx_pmaddwd.
(*sse4_1_mulv4si3): Merge with *avx_mulv4si3.
(ashr<mode>3): Merge with *avx_ashr<mode>3.
(lshr<mode>3): Merge with *avx_lshr<mode>3.
(ashl<mode>3): Merge with *avx_ashl<mode>3.
(sse2_ashlv1ti3): Merge with *avx_ashlv1ti3.
(sse2_lshrv1ti3): Merge with *avx_lshrv1ti3.
(*sse4_1_<smaxmin:code><mode>3): Merge with *avx_<smaxmin:code><mode>3.
(*<smaxmin:code>v8hi3): Ditto.
(*sse4_1_<umaxmin:code><mode>3): Merge with *avx_<umaxmin:code><mode>3.
(*<smaxmin:code>v16qi3): Ditto.
(*sse4_1_eqv2di3): Merge with *avx_eq<mode>3.
(*sse2_eq<mode>3): Ditto.
(*sse4_2_eqv2di3): Merge with *avx_gt<mode>3.
(*sse2_gt<mode>3): Ditto.
(vcondv2di): Split out of vcond<mode>.
(vconduv2di): Split out of vcondu<mode>.
re PR fortran/18918 (Eventually support Fortran 2008's coarrays [co-arrays])
2011-04-11 Tobias Burnus <burnus@net-b.de>
PR fortran/18918
* simplify.c (simplify_bound_dim): Exit for
ucobound's last dimension unless -fcoarray=single.
* trans-array (gfc_conv_descriptor_size_1): Renamed from
gfc_conv_descriptor_size, made static, has now from_dim and
to_dim arguments.
(gfc_conv_descriptor_size): Call gfc_conv_descriptor_size.
(gfc_conv_descriptor_cosize): New function.
* trans-array.h (gfc_conv_descriptor_cosize): New prototype.
* trans-intrinsic.c (conv_intrinsic_cobound): Add input_location
and handle last codim of ucobound for when -fcoarray is not "single".
Richard Guenther [Mon, 11 Apr 2011 15:47:30 +0000 (15:47 +0000)]
stor-layout.c (layout_type): Compute all array index size operations in the original type.
2011-04-11 Richard Guenther <rguenther@suse.de>
* stor-layout.c (layout_type): Compute all array index size operations
in the original type.
(initialize_sizetypes): Add comment.
(set_sizetype): Do not set TREE_TYPE of a TREE_VEC.
c-family/
* c-common.c (complete_array_type): Build a range type of
proper type.
Martin Jambor [Mon, 11 Apr 2011 14:25:55 +0000 (16:25 +0200)]
re PR tree-optimization/48195 (ICE: vector VEC(ipa_node_params_t,base) index domain error, in ipa_analyze_node at ipa-prop.c:1525 with -flto --param partial-inlining-entry-probability=101)
2011-04-11 Martin Jambor <mjambor@suse.cz>
PR tree-optimization/48195
* ipa-cp.c (ipcp_driver): Call ipa_check_create_node_params and
ipa_check_create_edge_args.
(ipcp_generate_summary): Do not call ipa_check_create_node_params and
ipa_check_create_edge_args.
* ipa-inline.c (inline_generate_summary): Do not call
ipa_check_create_node_params and ipa_check_create_edge_args.
* ipa-prop.c (ipa_analyze_node): Call ipa_check_create_node_params and
ipa_check_create_edge_args.
Jim Meyering [Sun, 10 Apr 2011 18:13:23 +0000 (18:13 +0000)]
Avoid memory overrun in a test leading to potential double-free.
2011-04-10 Jim Meyering <meyering@redhat.com>
Avoid memory overrun in a test leading to potential double-free.
* testsuite/test-expandargv.c (writeout_test): Fix off-by-one error:
i.e., do copy the trailing NUL byte.