Martin Liska [Thu, 25 Feb 2016 16:57:39 +0000 (17:57 +0100)]
Do not gather mem stats in run_exit_handles (PR
PR middle-end/69919
* alloc-pool.c (after_memory_report): New variable.
* alloc-pool.h (base_pool_allocator ::release): Do not use
the infrastructure if after_memory_report.
* toplev.c (toplev::main): Mark after memory report.
Ilya Verbin [Thu, 25 Feb 2016 12:23:52 +0000 (12:23 +0000)]
re PR driver/68463 (Offloading fails when some objects are compiled with LTO and some without)
gcc/
PR driver/68463
* config/gnu-user.h (CRTOFFLOADBEGIN): Define. Add crtoffloadbegin.o if
offloading is enabled and -fopenacc or -fopenmp is specified.
(CRTOFFLOADEND): Likewise.
(GNU_USER_TARGET_STARTFILE_SPEC): Add CRTOFFLOADBEGIN.
(GNU_USER_TARGET_ENDFILE_SPEC): Add CRTOFFLOADEND.
* lto-wrapper.c (offloadbegin, offloadend): Remove static vars.
(offload_objects_file_name): New static var.
(tool_cleanup): Remove offload_objects_file_name file.
(find_offloadbeginend): Replace with ...
(find_crtoffloadtable): ... this.
(run_gcc): Remove offload_argc and offload_argv.
Get offload_objects_file_name from -foffload-objects=... option.
Read names of object files with offload from this file, pass them to
compile_images_for_offload_targets. Don't call find_offloadbeginend and
don't pass offloadbegin and offloadend to the linker. Don't pass
offload non-LTO files to the linker, because now they're not claimed.
libgcc/
PR driver/68463
* Makefile.in (crtoffloadtable$(objext)): New rule.
* configure.ac (extra_parts): Add crtoffloadtable$(objext) if
enable_offload_targets is not empty.
* configure: Regenerate.
* offloadstuff.c: Move __OFFLOAD_TABLE__ from crtoffloadend to
crtoffloadtable.
libgomp/
PR driver/68463
* testsuite/libgomp.oacc-c-c++-common/parallel-dims-2.c: Remove.
lto-plugin/
PR driver/68463
* lto-plugin.c (struct plugin_offload_file): New.
(offload_files): Change type.
(offload_files_last, offload_files_last_obj): New.
(offload_files_last_lto): New.
(free_2): Adjust accordingly.
(all_symbols_read_handler): Don't add offload files to lto_arg_ptr.
Don't call free_1 for offload_files. Write names of object files with
offloading to the temporary file. Add new option to lto_arg_ptr.
(claim_file_handler): Don't claim file if it contains offload sections
without LTO sections. If it contains offload sections, add to the list.
Jakub Jelinek [Thu, 25 Feb 2016 08:09:02 +0000 (09:09 +0100)]
re PR rtl-optimization/69896 (wrong code with -frename-registers @ x64_64)
PR rtl-optimization/69896
* regcprop.c: Include cfgrtl.h.
(copyprop_hardreg_forward_1): If noop_p insn uses narrower
than remembered mode, either delete it (if noop_move_p), or
treat like copy_p but not noop_p instruction.
Patrick Palka [Thu, 25 Feb 2016 01:14:27 +0000 (01:14 +0000)]
re PR c++/69736 ("error: too few arguments to function" in c++14 but not c++11)
Fix PR c++/69736
gcc/cp/ChangeLog:
PR c++/69736
* cp-tree.h (REF_PARENTHESIZED_P): Adjust documentation.
(maybe_undo_parenthesized_ref): Declare.
* semantics.c (maybe_undo_parenthesized_ref): Split out from
check_return_expr.
(finish_call_expr): Use it.
* typeck.c (check_return_expr): Use it.
* pt.c (tsubst_copy_and_build) [INDIRECT_REF]: Retain the
REF_PARENTHESIZED_P flag.
PR c++/69922
* class.c (build_base_path): Set TREE_NO_WARNING on the null_test.
Avoid folding it.
* init.c (build_vec_delete_1, build_delete): Don't fold the non-NULL
tests.
* cp-gimplify.c (cp_fold): For TREE_NO_WARNING comparisons with NULL,
unless they are folded into INTEGER_CST, error_mark_node or some
comparison with NULL, avoid folding them and use either the original
comparison or non-folded comparison of folded arguments.
* cp-ubsan.c (cp_ubsan_instrument_vptr): Set TREE_NO_WARNING on the
comparison, don't fold the comparison right away.
* g++.dg/warn/Wnonnull-compare-6.C: New test.
* g++.dg/warn/Wnonnull-compare-7.C: New test.
* g++.dg/ubsan/pr69922.C: New test.
Marek Polacek [Wed, 24 Feb 2016 21:47:59 +0000 (21:47 +0000)]
re PR c/69819 (ICE on invalid code on x86_64-linux-gnu in tree check: expected function_type or method_type, have array_type in function_args_iter_init, at tree.h:4536)
PR c/69819
* c-decl.c (finish_decl): Don't update the copy of the type of a
different decl type.
Martin Sebor [Wed, 24 Feb 2016 17:04:03 +0000 (17:04 +0000)]
Avoid making unportable assumptions about the relationship between SIZE_MAX
and UINT_MAX.
gcc/testsuite/ChangeLog:
* gcc/testsuite/gcc.dg/builtins-68.c: Avoid making unportable
assumptions about the relationship between SIZE_MAX and UINT_MAX.
* gcc/testsuite/g++.dg/ext/builtin_alloca.C: Same.
Richard Biener [Wed, 24 Feb 2016 14:54:01 +0000 (14:54 +0000)]
re PR tree-optimization/69760 (Wrong 64-bit memory address caused by an unneeded overflowing 32-bit integer multiplication on x86_64 under -O2 and -O3 code optimization)
2016-02-24 Richard Biener <rguenther@suse.de>
Jakub Jelinek <jakub@redhat.com>
Richard Biener [Wed, 24 Feb 2016 12:03:27 +0000 (12:03 +0000)]
re PR tree-optimization/68963 (O3 vs. O2 discards part of loop and terminates early)
2016-02-24 Richard Biener <rguenther@suse.de>
PR middle-end/68963
* tree-ssa-loop-niter.c (derive_constant_upper_bound_ops): Fix
bogus check.
(record_nonwrapping_iv): Do not fall back to the low/high bound
for non-constant IV bases if the stmt is not always executed.
Evandro Menezes [Tue, 23 Feb 2016 21:31:00 +0000 (21:31 +0000)]
Tweak the pipeline model for Exynos M1
gcc/
* config/aarch64/aarch64.c (exynosm1_tunings): Enable fusion of AES{D,E}
and AESMC pairs.
* config/arm/exynos-m1.md: Change cost of STP, fix bypass for stores
and add bypass for AES{D,E} and AESMC pairs.
Jerry DeLisle [Tue, 23 Feb 2016 18:38:31 +0000 (18:38 +0000)]
re PR fortran/69456 (Namelist value with trailing sign is ignored without error)
2016-02-23 Jerry DeLisle <jvdelisle@gcc.gnu.org>
PR libgfortran/69456
* io/list_read.c (read_real): If digit is missing from exponent issue
an error. (parse_real): Likewise and adjusted error message to clarify
it is part of a complex number.
(nml_read_obj): Bump item count and add comment that this is used to
identify which item in a namelist read has a problem.
PR libgfortran/69456
* gfortran.dg/namelist_89.f90: New test.
* gfortran.dg/pr59700.f90: Update test..
Martin Sebor [Tue, 23 Feb 2016 18:09:37 +0000 (18:09 +0000)]
PR middle-end/69780 - [4.9/5/6 Regression] ICE on __builtin_alloca_with_align
PR middle-end/69780 - [4.9/5/6 Regression] ICE on __builtin_alloca_with_align
with small alignment
PR c/69759 - __builtin_alloca and __builtin_alloca_with_align undocumented
gcc/c-family/ChangeLog:
* c-common.c (check_builtin_function_arguments): Validate and reject
invalid arguments to __builtin_alloca_with_align.
gcc/ChangeLog:
* doc/extend.texi (Other Builtins): Document __builtin_alloca and
__builtin_alloca_with_align.
gcc/testsuite/ChangeLog:
* g++.dg/ext/builtin_alloca.C: New test.
* gcc.dg/builtins-68.c: New test.
David Malcolm [Tue, 23 Feb 2016 17:44:28 +0000 (17:44 +0000)]
PR preprocessor/69126: avoid comparing ad-hoc and non-ad-hoc locations
gcc/testsuite/ChangeLog:
PR preprocessor/69126
PR preprocessor/69543
* c-c++-common/pr69126-2-long.c: New test.
* c-c++-common/pr69126-2-short.c: New test.
* c-c++-common/pr69543-1.c: Remove xfail.
libcpp/ChangeLog:
PR preprocessor/69126
PR preprocessor/69543
* line-map.c (linemap_compare_locations): At the function top,
replace inlined bodies of get_location_from_adhoc_loc with calls
to get_location_from_adhoc_loc. Add a pair of calls to
get_location_from_adhoc_loc at the bottom of the function, to
avoid meaningless comparisons of ad-hoc and non-ad-hoc locations.
David Malcolm [Tue, 23 Feb 2016 17:39:16 +0000 (17:39 +0000)]
Add test coverage for _Pragma (PR preprocessor 69126, 69543, 69558)
We had some regressions in the ability for _Pragma to disable a warning
(PR preprocessor/69126, PR preprocessor/69543, PR preprocessor/69558).
This patch attempts to add more test coverage for this, for the
various combinations of:
- various warnings:
-Wunused-variable
-Wuninitialized
-Wdeprecated-declarations
- various combinations of location of _Pragma relative to location of
the warning:
- _Pragma is in a macro, warning isn't a macro
- neither is in a macro
- _Pragma isnt't in a macro, warning is in a macro
- in different macros
- both in the same macro
- C vs C++ frontend.
It adds some XFAILs:
- pr69543-1.c for C++ (fixed in the followup patch)
- pr69543-3.c for both C and C++
- pr69543-4.c for both C and C++
- pr69558.c for C++ (moving it from gcc.dg to c-c++-common,
marking it as xfail for C++ for now)
Mark Wielaard [Mon, 22 Feb 2016 22:42:19 +0000 (22:42 +0000)]
PR28901 Add two levels for -Wunused-const-variable.
There is some controversy about enabling -Wunused-const-variable for all
unused static const variables because some feel there are too many errors
exposed in header files. Create two levels for -Wunused-const-variable.
One level to only check for unused static const variables in the main
compilation file. Which is enabled by -Wunused-variable. And a second
level that also checks for unused static const variables in included
files. Which must be explicitly enabled.
gcc/ChangeLog
PR c/28901
* cgraphunit.c (check_global_declaration): Check level of
warn_unused_const_variable and main_input_filename.
* doc/invoke.texi (Warning Options): Add -Wunused-const-variable=.
(-Wunused-variable): For C implies -Wunused-const-variable=1.
(-Wunused-const-variable): Explain levels 1 and 2.
gcc/c-family/ChangeLog
PR c/28901
* c.opt (Wunused-const-variable): Turn into Alias for...
(Wunused-const-variable=): New option.
gcc/testsuite/ChangeLog
PR c/28901
* gcc.dg/unused-variable-3.c: New test.
Jakub Jelinek [Mon, 22 Feb 2016 21:34:49 +0000 (22:34 +0100)]
re PR target/69888 (ICE: SIGSEGV in decide_alg (i386.c:26169) due to infinite (?) recursion with -minline-all-stringops -mmemset-strategy=no_stringop:-1:noalign)
PR target/69888
* config/i386/i386.c (decide_alg): Ensure we don't recurse with
identical arguments. Formatting and spelling fixes.
Jakub Jelinek [Mon, 22 Feb 2016 21:34:07 +0000 (22:34 +0100)]
re PR target/69888 (ICE: SIGSEGV in decide_alg (i386.c:26169) due to infinite (?) recursion with -minline-all-stringops -mmemset-strategy=no_stringop:-1:noalign)
PR target/69888
* config/i386/i386.c (decide_alg): Ensure we don't recurse with
identical arguments. Formatting and spelling fixes.
Richard Biener [Mon, 22 Feb 2016 09:32:35 +0000 (09:32 +0000)]
re PR middle-end/37448 (cannot compile big function)
2016-02-22 Richard Biener <rguenther@suse.de>
PR ipa/37448
* ipa-inline-transform.c (inline_call): When not updating
overall summaries adjust self size by the growth estimate.
* ipa-inline.c (inline_to_all_callers_1): Add to the callers
hash-set, do not update overall summaries here. Renamed from ...
(inline_to_all_callers): ... this which is now wrapping the
above and performing delayed overall summary update.
(early_inline_small_functions): Delay updating of the overall
summary.
Paul Thomas [Sat, 20 Feb 2016 18:26:59 +0000 (18:26 +0000)]
re PR fortran/69423 (Invalid optimization with deferred-length character)
2016-02-20 Paul Thomas <pault@gcc.gnu.org>
PR fortran/69423
* trans-decl.c (create_function_arglist): Deferred character
length functions, with and without declared results, address
the passed reference type as '.result' and the local string
length as '..result'.
(gfc_null_and_pass_deferred_len): Helper function to null and
return deferred string lengths, as needed.
(gfc_trans_deferred_vars): Call it, thereby reducing repeated
code, add call for deferred arrays and reroute pointer function
results. Avoid using 'tmp' for anything other that a temporary
tree by introducing 'type_of_array' for the arrayspec type.
2016-02-20 Paul Thomas <pault@gcc.gnu.org>
PR fortran/69423
* gfortran.dg/deferred_character_15.f90 : New test.
Jakub Jelinek [Fri, 19 Feb 2016 22:18:38 +0000 (23:18 +0100)]
re PR driver/69805 (ICE in greater_than_spec_func, at gcc.c:9722)
PR driver/69805
* gcc.c (LINK_COMMAND_SPEC, GOMP_SELF_SPECS): Use
:%* in %:gt() argument.
(greater_than_spec_func): Adjust for expecting only numbers,
if there are more than two numbers, compare the last two.
projects / thirdparty / gcc.git / log
