Doug Gilmore [Tue, 4 Oct 2016 15:28:23 +0000 (15:28 +0000)]
re PR tree-optimization/77808 (ICE in duplicate_ssa_name_ptr_info, at tree-ssanames.c:630 starting with r240439)
Fix PR tree-optimization/77808
gcc/
PR tree-optimization/77808
* tree-ssa-loop-prefetch.c (issue_prefetch_ref): Check base_addr
and addr are different before copying points-to information.
gcc/testsuite/
PR tree-optimization/77808
* gcc.dg/tree-ssa/pr77808.c: New testcase.
Jeff Law [Mon, 3 Oct 2016 19:28:24 +0000 (13:28 -0600)]
re PR tree-optimization/71550 (wrong code at -O3 on x86_64-linux-gnu)
PR tree-optimization/71550
PR tree-optimization/71403
* tree-ssa-threadbackward.c: Include tree-vectorizer.h
(profitable_jump_thread_path): Also return boolean indicating if
the realized path will create an irreducible loop.
Remove loop depth tests from 71403.
(fsm_find_control_statement_thread_paths): Remove loop depth tests
from 71403. If threading will create an irreducible loop, then
throw away loop iteration and related information.
PR tree-optimization/71550
PR tree-optimization/71403
* gcc.c-torture/execute/pr71550.c: New test.
The Go runtime package is picking up C macros from runtime_sysinfo.go
and then re-exporting them to runtime.inc. This can cause name
conflicts. Change the Makefile so that we only put the macros we need
into runtime.inc. These are the constants that are actually defined by
Go code, not runtime_sysinfo.go. There are only a few, so we can
pattern match.
This is an additional hack on runtime.inc. The long term goal is to
convert the runtime package to Go and eliminate runtime.inc entirely, so
a few hacks seem acceptable.
PR preprocessor/77699
* input.c (maybe_grow): Don't allocate one byte extra headroom.
(get_next_line): Return false on error.
(read_next_line): Removed, use get_next_line instead.
(read_line_num): Don't copy the line.
(location_get_source_line): Don't use static data.
(selftest::test_reading_source_line): Add more test cases.
The compiler defaults to thin LTO objects if built with linker plugin
support, and in that case, the error expected by the testcase appears
only at link time, if at all. Forcing fat LTO object file creation
generates the error consistently at compile time, as expected.
Jonathan Wakely [Mon, 3 Oct 2016 11:50:46 +0000 (12:50 +0100)]
Ensure "C++" language linkage for std::abs overloads
PR libstdc++/77814
* include/bits/std_abs.h: Use "C++" language linkage.
* testsuite/17_intro/headers/c++2011/linkage.cc: Move <complex.h> to
the end. Add <stdalign.h>.
re PR libfortran/77663 (libgfortran/caf/single.c: three minor problems and a lost token)
gcc/testsuite/ChangeLog:
2016-10-01 Andre Vehreschild <vehre@gcc.gnu.org>
PR fortran/77663
* gfortran.dg/coarray_send_by_ref_1.f08: New test.
libgfortran/ChangeLog:
2016-10-01 Andre Vehreschild <vehre@gcc.gnu.org>
PR fortran/77663
* caf/single.c (caf_internal_error): Fix not terminating va-list.
(_gfortran_caf_register): Free memory also when other allocs failed.
(_gfortran_caf_get_by_ref): Fixed style.
(send_by_ref): Token is now stored at the correct position preventing
inaccessible tokens, memory loss and possibly crashes.
Jakub Jelinek [Fri, 30 Sep 2016 22:46:39 +0000 (00:46 +0200)]
configure.ac: Add target-libffi to target_libraries.
* configure.ac: Add target-libffi to target_libraries.
Readd libgcj target disablings, modified to only target-libffi.
Readd target addition of go to unsupported languages.
* configure: Regenerated.
check.tpl: Convert line endings to unix on test outputs
2016-09-30 Tadek Kijkowski <tkijkowski@gmail.com>
* check.tpl: Convert line endings to unix on test outputs
* fixfixes.c: Fixed passing file name to apply_fix when
SEPARATE_FIX_PROC is defined
* fixincl.c: Use system_with_shell, fixes for MinGW and DJGPP
* fixlib.c, fixlib.h: Added system_with_shell and fix_path_separators
Ian Lance Taylor [Fri, 30 Sep 2016 13:45:08 +0000 (13:45 +0000)]
runtime: copy internal locking code from Go 1.7 runtime
Remove the old locking code written in C.
Add a shell script mkrsysinfo.sh to generate the runtime_sysinfo.go
file, so that we can get Go copies of the system time structures and
other types.
Tweak the compiler so that when compiling the runtime package the
address operator does not cause local variables to escape. When the gc
compiler compiles the runtime, an escaping local variable is treated as
an error. We should implement that, instead of this change, when escape
analysis is turned on.
Tweak the compiler so that the generated C header does not include names
that start with an underscore followed by a non-upper-case letter,
except for the special cases of _defer and _panic. Otherwise we
translate C types to Go in runtime_sysinfo.go and then generate those Go
types back as C types in runtime.inc, which is useless and painful for
the C code.
Change entersyscall and friends to take a dummy argument, as the gc
versions do, to simplify calls from the shared code.
Matthew Wahab [Fri, 30 Sep 2016 13:25:04 +0000 (13:25 +0000)]
[Testsuite] Use correct effective-target settings for ARM fp16-aapcs tests.
2016-09-30 Matthew Wawhab <matthew.wahab@arm.com>
* gcc.target/arm/fp16-aapcs-3.c: Replace arm_fp16_ok with
arm_fp16_alternative_ok as the required effective target.
* gcc.target/arm/fp16-aapcs-4.c: Likewise. Also add
-mfp16-format=alternative to the dg-options directive and remove
the dg-add-options directive.
trans-array.c (gfc_array_allocate): Use the token from coarray's .token member.
gcc/fortran/ChangeLog:
2016-09-30 Andre Vehreschild <vehre@gcc.gnu.org>
* trans-array.c (gfc_array_allocate): Use the token from coarray's
.token member.
* trans-intrinsic.c (conv_expr_ref_to_caf_ref): Only generate
caf-reference chains from the first coarray references on.
* trans-types.c (gfc_get_derived_type): Switch on mandatory .token
member generation for allocatable arrays in coarrays in derived types.
gcc/testsuite/ChangeLog:
2016-09-30 Andre Vehreschild <vehre@gcc.gnu.org>
* gfortran.dg/coarray_allocate_10.f08: New test.
* gfortran.dg/coindexed_1.f90: Above fixes allow execution.
Jakub Jelinek [Thu, 29 Sep 2016 20:54:56 +0000 (22:54 +0200)]
Implement P0001R1 - C++17 removal of register storage class specifier c-family/
Implement P0001R1 - C++17 removal of register storage class specifier
c-family/
* c.opt (Wregister): New warning.
* c-opts.c (c_common_post_options): Enable -Wregister by
default for C++17.
cp/
* decl.c (cp_finish_decl): Diagnose register storage class
on vars except when used in GNU global or local register variable
extension.
(grokdeclarator): Diagnose register storage class on parameters.
* except.c (expand_start_catch_block): Set DECL_REGISTER only
after cp_finish_decl call.
testsuite/
* c-c++-common/Wvarargs-2.c (foo1): Except new warning for C++17.
* c-c++-common/vector-subscript-2.c (vf): Expect new error for
C++17.
* c-c++-common/vector-subscript-5.c (foo): Don't use register
keyword if not __SSE2__.
* c-c++-common/Wvarargs.c (foo1, foo3): Expect new warnings for
C++17.
* g++.dg/compat/struct-layout-1_generate.c (iterative_hash): Remove
register keywords.
* g++.dg/eh/pr29166.C: Add -Wno-register option.
* g++.dg/warn/register-parm-1.C (erroneous_warning,
no_erroneous_warning): Expect new warnings for C++17.
* g++.dg/warn/register-var-2.C (f): Likewise.
* g++.dg/parse/register1.C (f): Expect new error for C++17.
* g++.dg/parse/linkage2.C (foo): Likewise.
* g++.dg/torture/pr36826.C (CoinMin, CoinMax): Avoid register
keyword on parameters for C++17.
* g++.dg/cpp1z/register1.C: New test.
* g++.dg/cpp1z/register2.C: New test.
* g++.dg/cpp1z/register3.C: New test.
Make optional::reset noexcept, make optional::value work in constant expressions.
Make optional::reset noexcept, make optional::value
work in constant expressions.
* include/std/optional (_M_get): Make constexpr.
(reset): Make noexcept.
* testsuite/20_util/optional/assignment/7.cc: New.
* testsuite/20_util/optional/observers/6.cc: New.
Martin Liska [Thu, 29 Sep 2016 12:58:28 +0000 (14:58 +0200)]
objc: update documetation and add test-case of
* doc/extend.texi: Remove limitation of Objective C for
__attribute__((constructor)) and __attribute__((destructor)).
* objc/execute/construct1.m: New test.
James Greenhalgh [Thu, 29 Sep 2016 08:41:34 +0000 (08:41 +0000)]
[Patch] Remove all uses of TARGET_FLT_EVAL_METHOD_NON_DEFAULT and poison it
I'm going to hookize TARGET_FLT_EVAL_METHOD, so the current definition
of TARGET_FLT_EVAL_METHOD_NON_DEFAULT will stop working.
The relaxation enabled by TARGET_FLT_METHOD_NON_DEFAULT isn't portable,
and keeping it in place is tricky, so this patch removes it, and poisons
TARGET_FLT_EVAL_METHOD_NON_DEFAULT in system.h to prevent future use.
Ian Lance Taylor [Thu, 29 Sep 2016 00:56:44 +0000 (00:56 +0000)]
runtime: copy runtime.go and runtime1.go from Go 1.7
Also copy over cputicks.go, env_posix.go, vdso_none.go, stubs2.go, and a
part of os_linux.go. Remove the corresponding functions from the C code
in libgo/go/runtime. Add some transitional support functions to
stubs.go. This converts several minor functions from C to Go.
Alan Modra [Wed, 28 Sep 2016 22:03:17 +0000 (07:33 +0930)]
Disable .gnu.attributes tags in compatibility-ldbl.o
compatibility-ldbl.o is compiled with -mlong-double-64. When
long double .gnu.attributes tags are checked by the linker, it
complains about the mismatch between this file and others in
libstdc++.
Alan Modra [Wed, 28 Sep 2016 22:01:48 +0000 (07:31 +0930)]
[RS6000] .gnu.attributes Tag_GNU_Power_ABI_FP
Extend this attribute to cover long double ABIs, for 64-bit too.
This patch also corrects an error that crept in to code setting
rs6000_passes_float. See the added comment. Passing IEEE128 values
in vsx regs ought to set both Tag_GNU_Power_ABI_FP and
Tag_GNU_Power_ABI_Vector. Also adds a new option, default on, that
disables output of .gnu_attribute assembly directives.
* config/rs6000/sysv4.opt (mgnu-attribute): New option.
* doc/invoke.texi: Document it.
* config/rs6000/rs6000.c (HAVE_LD_PPC_GNU_ATTR_LONG_DOUBLE): Define.
(rs6000_passes_float): Comment.
(rs6000_passes_long_double): New static var.
(call_ABI_of_interest): Return false unless rs6000_gnu_attr is set.
(init_cumulative_args): Set up to emit fp .gnu_attribute for
ELF 64-bit ABIs as well as 32-bit ELF. Correct rs6000_passes_float
to include fp values returned in vectors.
Set rs6000_passes_long_double.
(rs6000_function_arg_advance_1): Likewise for function args.
(rs6000_elf_file_end): Emit fp .gnu_attribute for ELF 64-bit ABIs,
and SPE. Emit long double tag value too.
(rs6000_opt_vars): Add gnu-attr.
* configure.ac (HAVE_LD_PPC_GNU_ATTR_LONG_DOUBLE): New ppc32 test.
* configure: Regenerate.
* config.in: Regenerate.
Jakub Jelinek [Wed, 28 Sep 2016 19:47:04 +0000 (21:47 +0200)]
gimple-ssa-sprintf.c: Fix comment formatting.
* gimple-ssa-sprintf.c: Fix comment formatting.
(format_integer): Use is_gimple_assign.
(pass_sprintf_length::handle_gimple_call): Use gimple_call_builtin_p
and gimple_call_fndecl. Reorder case BUILT_IN_SPRINTF_CHK. Fix up
BUILT_IN_SNPRINTF_CHK comment. Replace "to to" with "to" in comment.
(pass_sprintf_length::execute): Use is_gimple_call.
Jakub Jelinek [Wed, 28 Sep 2016 19:21:47 +0000 (21:21 +0200)]
re PR c++/77467 (Segmentation fault with switch statement in constexpr function)
PR c++/77467
* constexpr.c (enum constexpr_switch_state): New.
(struct constexpr_ctx): Add css_state field.
(label_matches): Add CTX and STMT arguments, remove I and
DEFAULT_LABEL. For CASE_LABEL_EXPR assert ctx->css_state != NULL,
handle default labels according to css_state.
(cxx_eval_statement_list): Remove statement skipping, label_matches
and default_label handling code.
(cxx_eval_loop_expr): Exit after first iteration even if
switches (jump_target).
(cxx_eval_switch_expr): Set up css_state field in ctx, if default
label has been seen in the body, but no cases matched, evaluate
the body second time.
(cxx_eval_constant_expression): Handle stmt skipping and label_matches
here. Handle PREDICT_EXPR. For MODIFY_EXPR or INIT_EXPR, assert
statement is not skipped. For COND_EXPR during skipping, don't
evaluate condition, just the then block and if still skipping at the
end also the else block.
(cxx_eval_outermost_constant_expr): Adjust constexpr_ctx initializer.
(is_sub_constant_expr): Likewise.
Ian Lance Taylor [Wed, 28 Sep 2016 18:02:40 +0000 (18:02 +0000)]
libgo: fix for runtime/check failure with "-O0 -g"
Tweak the makefile rules for the runtime/check test to
insure that the runtime package is compiled with
"-fgo-compiling-runtime". This resolves a test failure
(unsat on runtime.getcallerpc) when in a build directory where the
compiler flags have been configured to disable optimization.
projects / thirdparty / gcc.git / log
