This patch adds support for the Statistical Profiling Extension (SPE) on
AArch64. Even though the compiler will not generate code any differently
given this extension, it will need to pass it on to the assembler in
order to let it correctly assemble inline asm containing accesses to the
extension's system registers. The same applies when using the
preprocessor on an assembly file as this first must pass through cc1.
I left the hwcaps string for SPE empty as the kernel does not define a
feature string for this extension. The current effect of this is that
driver will disable profile feature bit in GCC. This is OK though
because we don't, nor do we ever, enable this feature bit, as codegen is
not affect by the SPE support and more importantly the driver will still
pass the extension down to the assembler regardless.
gcc/ChangeLog
2018-07-19 Andre Vieira <andre.simoesdiasvieira@arm.com>
* config/aarch64/aarch64-option-extensions.def: New entry for profile
extension.
* config/aarch64/aarch64.h (AARCH64_FL_PROFILE): New.
* doc/invoke.texi (aarch64-feature-modifiers): New entry for profile
extension.
gcc/testsuite/ChangeLog
2018-07-19 Andre Vieira <andre.simoesdiasvieira@arm.com>
* gcc.target/aarch64/profile.c: New test.
From-SVN: r262882
gcc/ChangeLog
2018-07-19 Andre Vieira <andre.simoesdiasvieira@arm.com>
* config/aarch64/aarch64-simd.md (aarch64_simd_mov<VQ:mode>): Replace
Umq with Umn.
(store_pair_lanes<mode>): Likewise.
* config/aarch64/aarch64-protos.h (aarch64_addr_query_type): Add new
enum value 'ADDR_QUERY_LDP_STP_N'.
* config/aarch64/aarch64.c (aarch64_addr_query_type): Likewise.
(aarch64_print_address_internal): Add declaration.
(aarch64_print_ldpstp_address): Remove.
(aarch64_classify_address): Adapt mode for 'ADDR_QUERY_LDP_STP_N'.
(aarch64_print_operand): Change printing of 'y'.
* config/aarch64/predicates.md (aarch64_mem_pair_lanes_operand): Use
new enum value 'ADDR_QUERY_LDP_STP_N', don't hardcode mode and use
'true' rather than '1'.
* gcc/config/aarch64/constraints.md (Uml): Likewise.
(Uml): Rename to Umn.
(Umq): Remove.
From-SVN: r262880
2018-07-19 Richard Biener <rguenther@suse.de>
* tree-ssa-sccvn.h (struct vn_phi_s): Make phiargs member
a trailing array.
* tree-ssa-sccvn.c: Remove alloc-pool.h use.
(vn_phi_hasher): Derive from nofree_ptr_hash and remove remove method.
(vn_reference_hasher): Likewise.
(struct vn_tables_s): Remove obstack and alloc-pool members.
(vn_tables_obstack, vn_tables_insert_obstack): New global obstacks.
(vn_nary_build_or_lookup_1): Manually build in vn_tables_insert_obstack.
(vn_reference_insert): Allocate from obstack instead of from alloc-pool.
(vn_reference_insert_pieces): Likewise.
(alloc_vn_nary_op_noinit): Adjust.
(vn_nary_op_insert_stmt): Allocate phiargs in-place.
(vn_phi_eq): Adjust.
(shared_lookup_phiargs): Remove.
(vn_phi_lookup): Allocate temporary vn_phi_s on the stack.
(vn_phi_insert): Allocate from obstack instead of from alloc-pool.
(visit_reference_op_call): Likewise.
(copy_nary, copy_phi, copy_reference): Remove.
(process_scc): Rewind the obstack when iterating. Do not
copy the elements to valid_info but just move them from one
hashtable to the other.
(allocate_vn_table): Adjust.
(free_vn_table): Likewise.
(init_scc_vn): Likewise.
(free_scc_vn): Likewise.
From-SVN: r262879
In
struct ucontext;
typedef struct ucontext ucontext_t;
extern int (*bar) (ucontext_t *__restrict __oucp,
const ucontext_t *__restrict __ucp)
__attribute__((__indirect_return__));
extern int res;
void
foo (ucontext_t *oucp, ucontext_t *ucp)
{
res = bar (oucp, ucp);
}
bar() may return via indirect branch. This patch changes indirect_return
to type attribute to allow indirect_return attribute on variable or type
of function pointer so that ENDBR can be inserted after call to bar().
gcc/
PR target/86560
* config/i386/i386.c (rest_of_insert_endbranch): Lookup
indirect_return as function type attribute.
(ix86_attribute_table): Change indirect_return to function
type attribute.
* doc/extend.texi: Update indirect_return attribute.
gcc/testsuite/
PR target/86560
* gcc.target/i386/pr86560-1.c: New test.
* gcc.target/i386/pr86560-2.c: Likewise.
* gcc.target/i386/pr86560-3.c: Likewise.
From-SVN: r262877
gcc/ChangeLog:
2018-07-18 Kelvin Nilsen <kelvin@gcc.gnu.org>
* doc/extend.texi (PowerPC AltiVec Built-in Functions): Rename
this subsection to "PowerPC AltiVec/VSX Built-in Functions".
(PowerPC AltiVec/VSX Built-in Functions): New name for subsection
previously known as "PowerPC AltiVec Built-in Functions". Move
some material to new subsubsections "PowerPC AltiVec Built-in
Functions on ISA 2.06" and "PowerPC AltiVec Built-in Functions on
ISA 2.07".
(PowerPC Altivec Built-in Functions on ISA 2.05): New subsubsection.
(PowerPC Altivec Built-in Functions on ISA 2.06): Likewise.
(PowerPC Altivec Built-in Functions on ISA 2.07): Likewise.
(PowerPC Altivec Built-in Functions on ISA 3.0): Likewise.
From-SVN: r262863
libcpp:
2018-07-18 Bernd Edlinger <bernd.edlinger@hotmail.de>
PR 69558
* macro.c (enter_macro_context): Change the location info for builtin
macros and _Pragma from location of the closing parenthesis to location
of the macro expansion point.
testsuite:
2018-07-18 Bernd Edlinger <bernd.edlinger@hotmail.de>
PR 69558
* c-c++-common/cpp/diagnostic-pragma-2.c: New test.
* c-c++-common/pr69558.c: Remove xfail.
* gcc.dg/cpp/builtin-macro-1.c: Adjust test expectations.
* gcc.dg/pr61817-1.c: Likewise.
* gcc.dg/pr61817-2.c: Likewise.
* g++.dg/plugin/pragma_plugin.c: Warn at expansion_point_location.
From-SVN: r262861
2018-07-18 Janus Weil <janus@gcc.gnu.org>
Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/85599
* dump-parse-tree.c (show_attr): Add handling of implicit_pure.
* frontend-passes.c (do_warn_function_elimination): Do not warn for
pure functions.
* gfortran.h: Add prototypes for gfc_pure_function and
gfc_implicit_pure_function.
* gfortran.texi: Add chapter on evaluation of logical expressions.
* invoke.texi: Mention that -Wfunction-elimination is implied
by -Wextra.
* lang.opt: Make -Wextra imply -Wfunction-elimination.
* resolve.c (pure_function): Rename to gfc_pure_function.
(gfc_implicit_pure_function): New function.
(check_pure_function): Use it here.
(impure_function_callback): New function.
(resolve_operator): Call it via gfc_expr_walker.
2018-07-18 Janus Weil <janus@gcc.gnu.org>
PR fortran/85599
* gfortran.dg/function_optimize_5.f90: Add option
'-faggressive-function-elimination' and update dg-warning clauses.
* gfortran.dg/short_circuiting.f90: New test.
Co-Authored-By: Thomas Koenig <tkoenig@gcc.gnu.org>
From-SVN: r262860
CL 123362 introduced a bug in creating alias type's backend
representation. A type's btype_ should not be set before
named types are converted if it is a placeholder. For alias
type, it is set too early. This may result in unresolved
placeholders. This CL fixes it.
Reviewed-on: https://go-review.googlesource.com/123975
From-SVN: r262833
* lex.c (_cpp_lex_direct): Use CPP_DL_NOTE instead of CPP_DL_PEDWARN,
CPP_DL_WARNING or CPP_DL_ERROR for note that diagnostics for C++ style
comments is reported only once per file and guard those calls on the
preceding cpp_error returning true.
* gcc.dg/cpp/pr61854-c90.c (foo): Expect a note, rather than error.
* gcc.dg/cpp/pr61854-c94.c (foo): Likewise.
* gcc.dg/cpp/pr61854-4.c (foo): Likewise.
* gcc.dg/cpp/pr61854-8.c: New test.
From-SVN: r262832
Previously, when creating the backend representation of a
circular type, we resolve the placeholder to a
circular_pointer_type. The backend doesn't know what the
concrete type is.
This CL changes it to resolve the placeholder to the concrete
type instead, so the backend may have better knowledge of the
concrete type.
Reviewed-on: https://go-review.googlesource.com/123738
From-SVN: r262830
Use -Wabi=2 to fix warnings about -Wabi having no effect on its own.
This requires suppressing two warnings in src/c++11/debug.cc which do
not affect the library ABI.
Previously libstdc++ defaulted to --enable-werror but the -Werror flag
was not actually added unless --enable-maintainer-mode was used. This is
not documented and not the expected behaviour. This removes any special
treatment for maintainer-mode, makes -Werror depend directly on
--enable-werror, and changes the default to --enable-werror=no.
PR libstdc++/86450
* acinclude.m4 (GLIBCXX_CHECK_COMPILER_FEATURES): Don't define WERROR.
(GLIBCXX_EXPORT_FLAGS): Use -Wabi=2 instead of -Wabi.
* configure: Regenerate.
* configure.ac: Change GLIBCXX_ENABLE_WERROR default to "no".
* doc/Makefile.in: Regenerate.
* fragment.am: Set WERROR_FLAG to -Werror instead of $(WERROR).
* include/Makefile.in: Regenerate.
* libsupc++/Makefile.in: Regenerate.
* po/Makefile.in: Regenerate.
* python/Makefile.in: Regenerate.
* src/Makefile.in: Regenerate.
* src/c++11/Makefile.in: Regenerate.
* src/c++11/debug.cc: Use diagnostic pragmas to suppress warnings
from -Wabi=2 that don't affect exported symbols.
* src/c++98/Makefile.in: Regenerate.
* src/filesystem/Makefile.in: Regenerate.
* testsuite/Makefile.in: Regenerate.
From-SVN: r262824
The explicit instantiation of std::call_once used to require an
instantiation of __bind_simple, but call_once was changed by r241031 to
not use __bind_simple. The instantiation of __bind_simple (and the
definitions it uses) are not needed. They should have been removed
instead of doing the changes in r241111 that kept them compiling.
The use of std::call_once by _Async_state_common::_M_join can be
simplified to use a pointer instead of reference wrapper. The call_once
symbol isn't exported so the change isn't visible outside the library.
* src/c++11/compatibility-thread-c++0x.cc [_GLIBCXX_SHARED]
(_Async_state_common::_M_join): Simplify use of std::call_once and
corresponding explicit instantiation.
(_Maybe_wrap_member_pointer, _Bind_simple, _Bind_simple_helper)
(__bind_simple): Remove definitions and explicit instantiation that
are not required by exported symbols.
From-SVN: r262823
2018-07-17 Richard Biener <rguenther@suse.de>
PR lto/86456
* dwarf2out.c (init_sections_and_labels): Always generate
a debug_line_str_section for early LTO debug.
(dwarf2out_finish): Reset debug_line_str_hash output early.
Bump counter for extra dwarf5 .debug_loc labels to not conflict
with fat LTO part.
(dwarf2out_early_finish): Output debug_line_str.
* g++.dg/debug/dwarf2/pr86456.C: New testcase.
From-SVN: r262819
gcc/ChangeLog:
2018-07-17 Robin Dapp <rdapp@linux.ibm.com>
* config/s390/s390.c (preferred_la_operand_p): Do not use LA with
index register on z196 or later.
From-SVN: r262818
PR middle-end/86542
* omp-low.c (create_task_copyfn): Copy over also fields corresponding
to _looptemp_ clauses, other than the first two.
* testsuite/libgomp.c++/pr86542.C: New test.
From-SVN: r262815
* gcc-interface/decl.c (choices_to_gnu): Rename parameters. Deal with
an operand of Character type. Factor out range generation to the end.
Check that the bounds are literals and convert them to the type of the
operand before building the ranges.
* gcc-interface/utils.c (make_dummy_type): Minor tweak.
(make_packable_type): Propagate TYPE_DEBUG_TYPE.
(maybe_pad_type): Likewise.
From-SVN: r262812
