* sem_ch6.adb (Analyze_Procedure_Call) If CodePeer_Mode is True,
do not analyze the expression of a machine code insertion.
* sem_ch13.adb (Analyze_Code_Statement) If CodePeer_Mode is True,
mark the N_Code_Statement node as analyzed and do nothing else.
From-SVN: r230225
2015-11-12 Ed Schonberg <schonberg@adacore.com>
* sem_ch6.adb (Check_Limited_Return): Make global to package
for use elsewhere.
(Analyze_Expression_Function): Remove duplicated code, pre-analyze
expression to capture names and call Check_Limited_Return so
that semantic checks are identical to those for regular functions
returning limited types.
2015-11-12 Gary Dismukes <dismukes@adacore.com>
* bindgen.adb: Fix typo.
* sem_ch6.adb: Minor reformatting.
2015-11-12 Emmanuel Briot <briot@adacore.com>
* s-os_lib.adb (Argument_String_To_List): fix handling of
windows separators
From-SVN: r230224
gcc/
2015-11-12 Renlin Li <renlin.li@arm.com>
* config/arm/arm.md (addsi3_compare_op2): Make the order of
assembly pattern consistent with constraint order.
From-SVN: r230222
2015-11-11 Claudiu Zissulescu <claziss@synopsys.com>
* config/arc/arc.c (gen_compare_reg): Swap operands also when we
do not expand to rtl.
From-SVN: r230217
* config/i386/i386.c (ix86_legitimate_combined_insn): Reject
combined insn if the alignment of vector mode memory operand
is less than ssememalign.
testsuite/ChangeLog:
* gcc.target/i386/sse-1.c (swizzle): Assume that a is
aligned to 64 bits.
From-SVN: r230215
2015-11-12 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
* fold-const.c (fold_binary_loc) : Move Convert A/B/C to A/(B*C)
to match.pd.
Move Convert A/(B/C) to (A/B)*C to match.pd.
Move Convert C1/(X*C2) into (C1/C2)/X to match.pd.
Move Optimize (X & (-A)) / A where A is a power of 2, to
X >> log2(A) to match.pd.
* match.pd (rdiv (rdiv:s @0 @1) @2): New simplifier.
(rdiv @0 (rdiv:s @1 @2)): New simplifier.
(div (convert? (bit_and @0 INTEGER_CST@1)) INTEGER_CST@2):
New simplifier.
(rdiv REAL_CST@0 (mult @1 REAL_CST@2)): New simplifier.
From-SVN: r230204
Earlier, graphite used to translate portions of the original program after
scop-detection in order to represent the SCoP into polyhedral model. This was
required because each basic block was represented as independent basic block in
the polyhedral model. So all the cross-basic-block dependencies were translated
out-of-ssa.
With this patch those dependencies are also exposed to the ISL, so there is no
need to modify the original structure of the program.
After this patch we should be able to enable graphite at some default
optimization level.
Highlights:
Remove cross bb scalar to array translation
For reductions, add support for more than just INT_CST
Early bailout on codegen.
Verify loop-closed ssa structure during copy of renames
The uses of exprs should come from bb which dominates the bb
Collect the init value of close phi in loop-guard
Do not follow vuses for close-phi, postpone loop-close phi until the
corresponding loop-phi is processed
Bail out if no bb found to place cond/loop -phis
Move insertion of liveouts at the end of codegen
Insert loop-phis in the loop-header.
This patch passes regtest and bootstrap with BOOT_CFLAGS='-O2 -fgraphite-identity -floop-nest-optimize'
2015-11-11 Aditya Kumar <aditya.k7@samsung.com>
Sebastian Pop <s.pop@samsung.com>
* graphite-isl-ast-to-gimple.c (class translate_isl_ast_to_gimple):
New member codegen_error
(translate_isl_ast_for_loop): Remove call to single_succ_edge and early return.
(translate_isl_ast_node_user): Early return in case of error.
(translate_isl_ast_to_gimple::translate_isl_ast): Same.
(translate_isl_ast_to_gimple::translate_pending_phi_nodes): New.
(add_parameters_to_ivs_params): Remove macro.
(graphite_regenerate_ast_isl): Add if_region pointer to region.
* graphite-poly.c (new_poly_dr): Remove macro.
(print_pdr): Same.
(new_gimple_poly_bb): Same.
(free_gimple_poly_bb): Same.
(print_scop_params): Same.
* graphite-poly.h (struct poly_dr): Same.
(struct poly_bb): Add new_bb.
(gbb_from_bb): Remove dead code.
(pbb_from_bb): Same.
* graphite-scop-detection.c (parameter_index_in_region_1): Same.
(parameter_index_in_region): Same.
(find_scop_parameters): Same.
(build_cross_bb_scalars_def): New.
(build_cross_bb_scalars_use): New.
(graphite_find_cross_bb_scalar_vars): New
(try_generate_gimple_bb): Reads and Writes.
(build_alias_set): Move.
(gather_bbs::before_dom_children): Gather bbs visited.
(build_scops): call build_alias_set.
* graphite-sese-to-poly.c (phi_arg_in_outermost_loop): Delete.
(remove_simple_copy_phi): Delete.
(remove_invariant_phi): Delete.
(simple_copy_phi_p): Delete.
(reduction_phi_p): Delete.
(isl_id_for_dr): Remove unused param.
(parameter_index_in_region_1): Remove macro usage.
(set_scop_parameter_dim): Same.
(add_param_constraints): Same.
(add_conditions_to_constraints): Same
(build_scop_iteration_domain): Same.
(pdr_add_alias_set): Comment.
(add_scalar_version_numbers): New.
(build_poly_dr): ISL id.
(build_scop_drs): Move.
(build_poly_sr_1): Same.
(insert_stmts): Remove.
(build_poly_sr): New.
(new_pbb_from_pbb): Delete.
(insert_out_of_ssa_copy_on_edge): Delete.
(create_zero_dim_array): Delete.
(scalar_close_phi_node_p): Delete.
(propagate_expr_outside_region): Delete.
(rewrite_close_phi_out_of_ssa): Delete.
(rewrite_phi_out_of_ssa): Delete.
(rewrite_degenerate_phi): Delete.
(rewrite_reductions_out_of_ssa): Delete.
(rewrite_cross_bb_scalar_dependence): Delete.
(handle_scalar_deps_crossing_scop_limits):
(rewrite_cross_bb_scalar_deps): Delete.
(build_poly_scop): Remove calls to out-of-ssa functions.
* graphite.c (graphite_transform_loops): Early return in case of codegen error.
* sese.c (debug_rename_map_1): Delete.
(debug_rename_map): Delete.
(sese_record_loop): Remove macro.
(build_sese_loop_nests): Same.
(new_sese_info): Same.
(free_sese_info): Same.
(sese_insert_phis_for_liveouts):
(is_loop_closed_ssa_use): New.
(number_of_phi_nodes): New.
(bb_contains_loop_close_phi_nodes): New.
(bb_contains_loop_phi_nodes): New.
(phi_uses_name): New.
(is_valid_rename):
(get_rename): Add old_bb and loop_phi for more precise matching of
exprs.
(set_rename): Pass region.
(later_of_the_two): New.
(gsi_insert_earliest): New.
(collect_all_ssa_names): New.
(substitute_ssa_name): New.
(rename_all_uses): New.
(get_rename_from_scev): New.
(rename_uses): Pass old_bb for more precise matching of exprs.
(get_def_bb_for_const): New.
(get_new_name): New.
(get_loc): New.
(get_edges): New.
(copy_loop_phi_args): New.
(copy_loop_phi_nodes): New.
(get_loop_init_value): New.
(find_init_value): New.
(find_init_value_close_phi): New.
(copy_loop_close_phi_args): New.
(copy_loop_close_phi_nodes): New.
(add_phi_arg_for_new_expr): New.
(copy_cond_phi_args): New.
(copy_cond_phi_nodes): New.
(copy_phi_nodes): New.
(should_copy_to_new_region): New.
(set_rename_for_each_def): New.
(graphite_copy_stmts_from_block): Early return in case of error.
(copy_bb_and_scalar_dependences): Same.
* sese.h (vec_find): New.
(SESE_PARAMS): Delete.
(SESE_LOOPS): Delete.
(SESE_LOOP_NEST): Delete.
(sese_contains_loop): Remove macro usage.
(sese_nb_params): Same.
(struct gimple_poly_bb): Added read_scalar_refs, write_scalar_refs.
Co-Authored-By: Sebastian Pop <s.pop@samsung.com>
From-SVN: r230200
The patch builds the original schedule based on the now optimized scattering
dimension instead of building one based on the loop index only.
The implementation is simpler and catches more cases where the original schedule
and the transformed schedule are the same, such as the one below:
for (i = 0; i < 1000; i++)
{
Temp = F[i];
for (j = 0; j < 1000; j++)
{
D[j] = E[j] * Temp;
A[i][j] = A[i][j] + B[i][j] * C[i][j] - D[j] ;
}
D[i] = E[i] * F[i];
}
* graphite-sese-to-poly.c (build_scop_original_schedule): Call
isl_union_map_add_map on every pbb->schedule.
From-SVN: r230191
2015-11-11 Tom de Vries <tom@codesourcery.com>
* tree-parloops.c (transform_to_exit_first_loop_alt): Insert new exit
block only when needed.
From-SVN: r230188
PR libstdc++/60421
* include/std/thread (this_thread::sleep_for): Retry on EINTR.
(this_thread::sleep_until): Retry if time not reached.
* src/c++11/thread.cc (__sleep_for): Retry on EINTR.
* testsuite/30_threads/this_thread/60421.cc: Test interruption and
non-steady clocks.
From-SVN: r230183
2015-11-11 Richard Biener <rguenther@suse.de>
* tree-vectorizer.h (vect_slp_analyze_and_verify_instance_alignment):
Declare.
(vect_analyze_data_refs_alignment): Make loop vect specific.
(vect_verify_datarefs_alignment): Likewise.
* tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependences):
Add missing continue.
(vect_compute_data_ref_alignment): Export.
(vect_compute_data_refs_alignment): Merge into...
(vect_analyze_data_refs_alignment): ... this.
(verify_data_ref_alignment): Split out from ...
(vect_verify_datarefs_alignment): ... here.
(vect_slp_analyze_and_verify_node_alignment): New function.
(vect_slp_analyze_and_verify_instance_alignment): Likewise.
* tree-vect-slp.c (vect_supported_load_permutation_p): Remove
misplaced checks on alignment.
(vect_slp_analyze_bb_1): Add fatal output parameter. Do
alignment analysis after SLP discovery and do it per instance.
(vect_slp_bb): When vect_slp_analyze_bb_1 fatally failed do not
bother to re-try using different vector sizes.
From-SVN: r230173
The testcase used to fail on 64-bit, but it was disabled there.
This patch makes it run there, and beefs up the checking of the
generated code a bit.
gcc/testsuite/
* gcc.target/powerpc/20050603-3.c: Don't restrict to ilp32. Do more
tests for the expected generated code.
From-SVN: r230167
If we have
(truncate:M1 (and:M2 (lshiftrt:M2 (x:M2) C) C2))
we can write it instead as
(and:M1 (lshiftrt:M1 (truncate:M1 (x:M2)) C) C2)
(if that is valid, of course), which has smaller modes for the
binary ops, and the truncate can often simplify further (if "x"
is a register, for example).
* gcc/simplify-rtx.c (simplify_truncation): Simplify TRUNCATE
of AND of [LA]SHIFTRT.
From-SVN: r230164
Fix PR rtl-optimization/68287
PR rtl-optimization/68287
* lra-lives.c (lra_create_live_ranges_1): Reserve the right
number of elements.
Co-Authored-By: Richard Biener <rguenther@suse.de>
From-SVN: r230163
gcc/
* config/mips/mips.c (mips_breakable_sequence_p): New function.
(mips_break_sequence): New function.
(mips_reorg_process_insns) Use them. Use compact branches in selected
situations.
gcc/testsuite/
* gcc.target/mips/split-ds-sequence.c: New test.
From-SVN: r230160
2015-11-11 Jiong Wang <jiong.wang@arm.com>
Jim Wilson <wilson@gcc.gnu.org>
PR target/67305
* config/arm/arm.md (neon_vector_mem_operand): Return FALSE if strict
be true and eliminable registers mentioned.
Co-Authored-By: Jim Wilson <wilson@gcc.gnu.org>
From-SVN: r230158
