Teresa Johnson [Fri, 25 Sep 2015 16:53:40 +0000 (16:53 +0000)]
This patch unsets -freorder-blocks-and-partition when -fprofile-use is not specified.
This patch unsets -freorder-blocks-and-partition when -fprofile-use
is not specified. Function splitting was not actually being performed
in that case, as probably_never_executed_bb_p does not distinguish
any basic blocks as being cold vs hot when there is no profile data.
Leaving it enabled, however, causes the assembly code generator to create
(empty) cold sections and labels, leading to unnecessary size overhead.
2015-09-25 Teresa Johnson <tejohnson@google.com>
* opts.c (finish_options): Unset -freorder-blocks-and-partition
if not using profile.
Avoid creating dangling references in case of nested tuples
for tuple constructors that construct from other tuples.
2015-09-25 Ville Voutilainen <ville.voutilainen@gmail.com>
Avoid creating dangling references in case of nested tuples
for tuple constructors that construct from other tuples.
* include/std/tuple (_TC::_NonNestedTuple): New.
* include/std/tuple (tuple::_TNTC): New.
* include/std/tuple (tuple(const tuple<_UElements...>&),
tuple(tuple<_UElements...>&&): Use _TNTC.
* testsuite/20_util/tuple/cons/nested_tuple_construct.cc: New.
PR pretty-print/67567 do not pass NULL as a string
Fortran passes NULL where a non-null string is expected by the pretty-printer,
which causes a sanitizer warning. This could have been found earlier by using
gcc_checking_assert. Even if the assertion is false, the result is just an
incomplete diagnostic, thus it seems more user-friendly to assert only when
checking. I do not have any idea how to properly fix the Fortran bug, thus this
patch simply works-around it.
gcc/fortran/ChangeLog:
2015-09-25 Manuel López-Ibáñez <manu@gcc.gnu.org>
PR pretty-print/67567
* resolve.c (resolve_fl_procedure): Work-around when iface->module
== NULL.
re PR target/67675 ([SH] Improve __builtin_strcmp alignment test)
gcc/
PR target/67675
* config/sh/sh-mem.cc (sh_expand_cmpstr): Check alignment of addr1 and
addr2 individually. Don't emit logical or insn if one is known to
be aligned approriately.
(sh_expand_cmpnstr): Likewise.
[AArch64] Force __builtin_aarch64_fp[sc]r argument into a REG
The testcase triggered an ICE because the builtin expansion
code passed the output of expand_normal directly to the SET_FP[SC]R
generator, without forcing it into a register first.
gcc/
* config/aarch64/aarch64-builtins.c (aarch64_expand_builtin): Force
__builtin_aarch64_fp[sc]r arguments into a register.
gcc/testsuite/
* gcc.target/aarch64/fpcr_fpsr_1.c: New file.
fdiagnostics-color=never does not disable color for some diagnostics
Actually, I was trying to reject non-warning options as argument to
-Werror=. However, the new test fails because -fdiagnostics-color=never is
always placed by the driver after the warning options when calling the compiler
proper. This patch prunes all -fdiagnostics-color from the command-line but the
last one, which is moved to the first position.
gcc/ChangeLog:
2015-09-24 Manuel López-Ibáñez <manu@gcc.gnu.org>
PR driver/67640
* opts-common.c (prune_options): Discard all -fdiagnostics-color
but the last one, which is moved to the front to be processed
first.
* opts.c (enable_warning_as_error): Reject options that do not
control warnings.
gcc/testsuite:
* lib/target-supports.exp (check_effective_target_pie): Check for
PIE support on Solaris 11.x and 12.
libgcc:
* config.host (*-*-solaris2*): Add t-crtstuff-pic to tmake_file.
Add crtbeginS.o, crtendS.o to extra_parts if libgcc_cv_solaris_crts.
* config/sol2/gmon.c: (monstartup): Don't write trailing NUL of
messages.
(internal_mcount): Likewise.
* config/sol2/t-sol2 (crtp.o, crtpg.o, gmon.o): Compile with
crt_compile, add CRTSTUFF_T_CFLAGS_S.
gcc:
* configure.ac (gcc_cv_ld_pie): Check for gld >= 2.26 on Solaris.
Check for ld -type pie on Solaris 11.x and 12.
* configure: Regenerate.
* config.in: Regenerate.
* gcc.c (LD_PIE_SPEC): Allow redefinition.
* config/sol2.h (STARTFILE_CRTBEGIN_SPEC): Define.
(STARTFILE_SPEC): Use it.
(ENDFILE_CRTEND_SPEC): Define.
(ENDFILE_SPEC): Use it and ENDFILE_ARCH_SPEC.
(SUBTARGET_EXTRA_SPECS): Add STARTFILE_CRTBEGIN_SPEC,
ENDFILE_ARCH_SPEC, ENDFILE_CRTEND_SPEC.
[HAVE_LD_PIE && HAVE_SOLARIS_CRTS] (LD_PIE_SPEC): Define.
(!(HAVE_LD_PIE && HAVE_SOLARIS_CRTS)] (LINK_PIE_SPEC): Define.
* config/i386/sol2.h (ENDFILE_SPEC): Remove.
(ENDFILE_ARCH_SPEC): Define.
* config/sparc/sol2.h (ENDFILE_ARCH_SPEC): Define.
Richard Biener [Thu, 24 Sep 2015 07:35:55 +0000 (07:35 +0000)]
tree-ssa-sccvn.h (vn_reference_op_struct): Add clique and base members.
2015-09-24 Richard Biener <rguenther@suse.de>
* tree-ssa-sccvn.h (vn_reference_op_struct): Add clique and base
members.
* tree-ssa-sccvn.c (copy_reference_ops_from_ref): Record clique
and base for MEM_REF and TARGET_MEM_REF. Handle BIT_FIELD_REF
offset.
(ao_ref_init_from_vn_reference): Record clique and base in the
built base.
* tree-ssa-pre.c (create_component_ref_by_pieces_1): Likewise
Patrick Palka [Thu, 24 Sep 2015 01:55:59 +0000 (01:55 +0000)]
Fix a -Wmisleading-indentation false-negative
gcc/c-family/ChangeLog:
* c-indentation.c (should_warn_for_misleading_indentation):
Compare next_stmt_vis_column with guard_line_first_nws instead
of with guard_line_vis_column.
* config/pa/linux-atomic.c (__kernel_cmpxchg2): Reorder error checks.
(__sync_fetch_and_##OP##_##WIDTH): Change result to match type of
__kernel_cmpxchg2.
(__sync_##OP##_and_fetch_##WIDTH): Likewise.
(__sync_val_compare_and_swap_##WIDTH): Likewise.
(__sync_bool_compare_and_swap_##WIDTH): Likewise.
(__sync_lock_test_and_set_##WIDTH): Likewise.
(__sync_lock_release_##WIDTH): Likewise.
(__sync_fetch_and_##OP##_4): Change result to match type of
__kernel_cmpxchg.
(__sync_##OP##_and_fetch_4): Likewise.
(__sync_val_compare_and_swap_4): Likewise.
(__sync_bool_compare_and_swap_4): likewise.
(__sync_lock_test_and_set_4): Likewise.
(__sync_lock_release_4): Likewise.
(FETCH_AND_OP_2): Add long long variants.
(OP_AND_FETCH_2): Likewise.
(COMPARE_AND_SWAP_2 ): Likewise.
(SYNC_LOCK_TEST_AND_SET_2): Likewise.
(SYNC_LOCK_RELEASE_2): Likewise.
(__sync_bool_compare_and_swap_##WIDTH): Correct return.
Thomas Schwinge [Wed, 23 Sep 2015 14:52:50 +0000 (16:52 +0200)]
Fix --enable-offload-targets/-foffload handling, pt. 1
gcc/
* configure.ac (offload_targets, OFFLOAD_TARGETS): Separate
offload targets by commas, not colons.
* config.in: Regenerate.
* configure: Likewise.
* gcc.c (driver::maybe_putenv_COLLECT_LTO_WRAPPER): Due to that,
instead of setting up the default offload targets here...
(process_command): ..., do it here.
libgomp/
* plugin/configfrag.ac (OFFLOAD_TARGETS): Clarify that offload
targets are separated by commas.
* config.h.in: Regenerate.
[c-family/49654/49655] reject invalid options in pragma diagnostic
Use find_opt instead of linear search through options in
handle_pragma_diagnostic (PR 49654) and reject non-warning options and
options not valid for the current language (PR 49655).
gcc/testsuite/ChangeLog:
2015-09-23 Manuel López-Ibáñez <manu@gcc.gnu.org>
PR c/49655
* gcc.dg/pragma-diag-6.c: New test.
gcc/ChangeLog:
2015-09-23 Manuel López-Ibáñez <manu@gcc.gnu.org>
PR c/49655
* opts.h (write_langs): Declare.
* opts-global.c (write_langs): Make it extern.
gcc/c-family/ChangeLog:
2015-09-23 Manuel López-Ibáñez <manu@gcc.gnu.org>
PR c/49654
PR c/49655
* c-pragma.c (handle_pragma_diagnostic): Detect non-warning
options and options not valid for the current language.
Jonathan Wakely [Wed, 23 Sep 2015 11:26:45 +0000 (12:26 +0100)]
Fix semantics of Filesystem TS directory iterators
[class.directory_iterator] p4 and [directory_iterator.members] p4
require that only the default constructor and ignored permission denied
errors can create the end iterator.
* acinclude.m4 (GLIBCXX_CHECK_FILESYSTEM_DEPS): Remove _GLIBCXX_
prefix from HAVE_STRUCT_DIRENT_D_TYPE.
* config.h.in: Regenerate.
* configure: Regenerate.
* include/experimental/fs_dir.h (operator==, operator==):
Use owner_before instead of pointer equality.
(directory_iterator(std::shared_ptr<_Dir>, error_code*)): Remove.
* src/filesystem/dir.cc (ErrorCode): Remove.
(_Dir::advance): Change ErrorCode parameter to error_code*, add
directory_options parameter and check it on error.
(opendir): Rename to open_dir to avoid clashing with macro. Change
ErrorCode parameter to error_code*.
(make_shared_dir): Remove.
(native_readdir) [_GLIBCXX_FILESYSTEM_IS_WINDOWS]: Don't set errno.
(directory_iterator(std::shared_ptr<_Dir>, error_code*)): Remove.
(directory_iterator(const path&, directory_options, error_code*)):
Pass options to _Dir::advance and create non-end iterator on error.
(recursive_directory_iterator(const path&, directory_options,
error_code*)): Clear error_code on ignored error, create non-end
iterator otherwise.
(recursive_directory_iterator::increment): Pass _M_options to
_Dir::advance.
(recursive_directory_iterator::pop): Likewise.
* testsuite/experimental/filesystem/iterators/directory_iterator.cc:
New.
* testsuite/experimental/filesystem/iterators/
recursive_directory_iterator.cc: New.
Patrick Palka [Tue, 22 Sep 2015 23:06:31 +0000 (23:06 +0000)]
Fix -Wmisleading-indentation false-positive
gcc/c-family/ChangeLog:
* c-indentation.c (should_warn_for_misleading_indentation):
Float out and consolidate the calls to get_visual_column that
are passed guard_exploc as an argument. Compare
next_stmt_vis_column with guard_line_first_nws instead of with
body_line_first_nws.
shrink-wrap: Handle multiple predecessors of prologue
The caller of try_shrink_wrapping wants to be returned a single edge to
put the prologue on. To make that work even if there are multiple edges
(all pointing to the PRO block) that need the prologue, add a new block
that becomes the destination of all such edges, and then jumps to PRO.
In the general case, some edges to PRO will need to be redirected, and
not all edges *can* be redirected. This adds a can_get_prologue function
that detects such cases. This then happily can also handle the "prologue
clobbers some reg that is live on the edge we want to insert it on" case.
Not all EDGE_CROSSING edges can be redirected, so handle those the same
as EDGE_COMPLEX edges.
* function.c (thread_prologue_and_epilogue_insns): Delete
orig_entry_edge argument to try_shrink_wrapping.
* shrink-wrap.c (can_get_prologue): New function.
(can_dup_for_shrink_wrapping): Also handle EDGE_CROSSING.
(try_shrink_wrapping): Delete orig_entry_edge argument. Use
can_get_prologue where needed. Remove code that finds a single
edge for the prologue. Remove code that tests if any reg clobbered
by the prologue is live on the prologue edge. Remove code that finds
the new prologue edge after duplicating blocks. Make a new prologue
block and edge.
* shrink-wrap.h (try_shrink_wrapping): Delete orig_entry_edge argument.
Jason Merrill [Tue, 22 Sep 2015 15:18:26 +0000 (11:18 -0400)]
Disconnect -Wabi=<n> from -fabi-compat-version=<n>.
gcc/c-family/
* c-common.h (abi_compat_version_crosses): New.
(warn_abi_version): Declare.
* c-common.c: Define it.
* c-opts.c (c_common_post_options): Handle it.
flag_abi_compat_version defaults to 8.
gcc/cp/
* mangle.c (abi_warn_or_compat_version_crosses): New.
Use it instead of abi_version_crosses.
(mangle_decl): Deal with -fabi-compat-version and -Wabi separately.
* config/nios2/nios2.c (nios2_legitimize_address): When handling
'reg + reloc' cases, allow first operand to be non-REG, and use
force_reg() to enforce address pattern.
[PATCH tree-inline] Do not say "called from here" with UNKNOWN_LOCATION
In https://sourceware.org/ml/libc-alpha/2014-12/msg00300.html, we give a
"called from here" note without actually having a location, which looks
strange. I haven't been able to generate such a testcase. If this happens,
try to at least point to the current function being called. If that location is
also unknown, skip the extra note.
gcc/ChangeLog:
2015-09-22 Manuel López-Ibáñez <manu@gcc.gnu.org>
* tree-inline.c (expand_call_inline): Use inform for extra note.
Do not give a note with UNKNOWN_LOCATION.
Replace input_location with gimple_location (stmt).
Use true/false instead of TRUE/FALSE.
David S. Miller [Tue, 22 Sep 2015 03:52:45 +0000 (03:52 +0000)]
Revert LRA SPARC changes for now.
gcc/
PR/67622
Revert:
2015-09-11 David S. Miller <davem@davemloft.net>
* config/sparc/constraints.md: Make "U" constraint a real register
constraint.
* config/sparc/sparc.c (TARGET_LRA_P): Define.
(D_MODES, DF_MODES): Add missing cast.
(TF_MODES, TF_MODES_NO_S): Include T_MODE.
(OF_MODES, OF_MODES_NO_S): Include O_MODE.
(sparc_register_move_cost): Decrease Niagara/UltrsSPARC memory
cost to 8.
* config/sparc/sparc.h (PROMOTE_MODE): Define.
* config/sparc/sparc.md (*movsi_lo_sum, *movsi_high): Do not
provide these insn when flag_pic.
2015-09-17 David S. Miller <davem@davemloft.net>
* config/sparc/sparc-protos.h (sparc_secondary_memory_needed):
Declare.
* config/sparc/sparc.c (sparc_secondary_memory_needed): New
function.
* config/sparc/sparc.h (SECONDARY_MEMORY_NEEDED): Use it.
(HARD_REGNO_CALLER_SAVE_MODE): Define.
* config/sparc/sparc.md (sethi_di_medlow, losum_di_medlow, seth44)
(setm44, setl44, sethh, setlm, sethm, setlo, embmedany_sethi)
(embmedany_losum, embmedany_brsum, embmedany_textuhi)
(embmedany_texthi, embmedany_textulo, embmedany_textlo): Do not
provide when flag_pic.
James Bowman [Mon, 21 Sep 2015 22:31:53 +0000 (22:31 +0000)]
[FT32] fixes for the default hardware startup
libgcc/
* config/ft32/crti-hw.S: Use __PMSIZE to allow configurable
memory layout. Deal correctly with BSS region larger than 32K.
Handle a watchdog reset like a power-on reset. Clean up unused
code.
Complete the implementation of N4230, Nested namespace definition.
/c-family
2015-09-21 Ville Voutilainen <ville.voutilainen@gmail.com>
Complete the implementation of N4230, Nested namespace definition.
* c-cppbuiltin.c: Add __cpp_namespace_attributes and
__cpp_nested_namespace_definitions.
/cp
2015-09-21 Ville Voutilainen <ville.voutilainen@gmail.com>
Complete the implementation of N4230, Nested namespace definition.
* parser.c (cp_parser_namespace_definition): Support namespace
attributes both before and after the namespace identifier.
/testsuite
2015-09-21 Ville Voutilainen <ville.voutilainen@gmail.com>
Complete the implementation of N4230, Nested namespace definition.
* g++.dg/cpp1y/feat-cxx11-neg.C: Add tests for C++17 namespace
attributes and nested namespace definitions.
* g++.dg/cpp1y/feat-cxx98-neg.C: Likewise.
* g++.dg/cpp1z/feat-cxx1z.C: Likewise.
* g++.dg/cpp1y/feat-cxx14-neg.C: New.
* g++.dg/cpp1z/namespace-attribs.C: Likewise.
* g++.dg/cpp1z/nested-namespace-def1.C: Add tests for attributes
appearing before the namespace identifier.
Handle lines encoded into several maps in linemap_position_for_loc_and_offset
linemap_position_for_loc_and_offset() tries to generate a location_t
encoding a column offset from the current location, for example, point
to a certain character inside a string. This is trivial to do when the
new location "fits within" the map of the original location. However,
it may happen that the (long) line corresponding to the original
location is encoded in several maps, thus the new location should
actually be encoded in a subsequent map from the original location.
This patch detects this case and adjusts the map correspondingly.
(This shows that the line-map representation is quite wasteful in this
case, because line-maps always start at column 0. That is, map[0]
highest location may encode up to line 8 column 80, then
map[1]->start_location starts encoding at line 8 column 0. Thus, there
are two location_t values that point to the same source location.)
libcpp/ChangeLog:
2015-09-21 Manuel López-Ibáñez <manu@gcc.gnu.org>
PR c/66415
* line-map.c (linemap_position_for_loc_and_offset): Handle the
case of long lines encoded in multiple maps.
gcc/testsuite/ChangeLog:
2015-09-21 Manuel López-Ibáñez <manu@gcc.gnu.org>
PR c/66415
* gcc.dg/cpp/pr66415-1.c: Test column number.
Richard Biener [Mon, 21 Sep 2015 14:02:27 +0000 (14:02 +0000)]
re PR debug/67664 (Duplicate AT_const_value attributes in some c++ cases.)
2015-09-21 Richard Biener <rguenther@suse.de>
PR debug/67664
* dwarf2out.c (add_location_or_const_value_attribute): Remove
attribute parameter. Early exit if either DW_AT_const_value
or DW_AT_location are present already.
(gen_variable_die): Adjust caller.
(dwarf2out_late_global_decl): Likewise.
re PR target/67657 ([SH][5/6 Regression]: internal compiler error: in cselib_record_set, at cselib.c:2396 when compiling libjpeg-turbo)
gcc/
PR target/67657
* config/sh/sh.c (sh_remove_overlapping_post_inc,
sh_peephole_emit_move_insn): Add new functions.
* config/sh/sh-protos.h (sh_remove_overlapping_post_inc,
sh_peephole_emit_move_insn): Declere them.
* config/sh/sh.md: Use them in various peephole2 patterns.
projects / thirdparty / gcc.git / log
