Jakub Jelinek [Thu, 4 Apr 2024 08:47:52 +0000 (10:47 +0200)]
fold-const: Handle NON_LVALUE_EXPR in native_encode_initializer [PR114537]
The following testcase is incorrectly rejected. The problem is that
for bit-fields native_encode_initializer expects the corresponding
CONSTRUCTOR elt value must be INTEGER_CST, but that isn't the case
here, it is wrapped into NON_LVALUE_EXPR by maybe_wrap_with_location.
We could STRIP_ANY_LOCATION_WRAPPER as well, but as all we are looking for
is INTEGER_CST inside, just looking through NON_LVALUE_EXPR seems easier.
2024-04-04 Jakub Jelinek <jakub@redhat.com>
PR c++/114537
* fold-const.cc (native_encode_initializer): Look through
NON_LVALUE_EXPR if val is INTEGER_CST.
Jakub Jelinek [Thu, 4 Apr 2024 08:47:00 +0000 (10:47 +0200)]
bitint: Handle m_bitfld_load cast in outer m_cast_conditional [PR114555]
We ICE on the following testcase, because we use result of a PHI node
which is only conditional because of a m_cast_conditional on the outermost
loops PHI node argument and so is invalid SSA form.
The following patch fixes it like similar cases elsewhere by adding
needed intervening PHI(s).
2024-04-04 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/114555
* gimple-lower-bitint.cc (bitint_large_huge::handle_cast): For
m_bitfld_load and save_cast_conditional add any needed PHIs
and adjust t4 accordingly.
* gcc.dg/bitint-103.c: New test.
* gcc.dg/bitint-104.c: New test.
Richard Biener [Wed, 3 Apr 2024 12:53:30 +0000 (14:53 +0200)]
tree-optimization/114551 - loop splitting and undefined overflow
When loop splitting hoists a guard computation it needs to make sure
that can be safely evaluated at this place when it was previously
only conditionally evaluated. The following fixes this for the
case of undefined overflow.
PR tree-optimization/114551
* tree-ssa-loop-split.cc (split_loop): If the guard is
only conditionally evaluated rewrite computations with
possibly undefined overflow to unsigned arithmetic.
Eugene Rozenfeld [Tue, 26 Mar 2024 23:28:08 +0000 (16:28 -0700)]
Don't set full_profile in auto-profile [PR113765]
auto-profile currently doesn't guarantee that it will set probabilities
on all edges because of zero basic block counts. Normally those edges
just have probabilities set by the preceding profile_estimate pass but
under -O0 profile_estimate pass doesn't run. The patch removes setting
of full_profile to true in auto-profile.
Tested on x86_64-pc-linux-gnu.
gcc/ChangeLog:
PR gcov-profile/113765
* auto-profile.cc (afdo_annotate_cfg): Don't set full_profile to true
H.J. Lu [Mon, 26 Feb 2024 16:38:58 +0000 (08:38 -0800)]
tree-profile: Disable indirect call profiling for IFUNC resolvers
We can't profile indirect calls to IFUNC resolvers nor their callees as
it requires TLS which hasn't been set up yet when the dynamic linker is
resolving IFUNC symbols.
Add an IFUNC resolver caller marker to cgraph_node and set it if the
function is called by an IFUNC resolver. Disable indirect call profiling
for IFUNC resolvers and their callees.
Tested with profiledbootstrap on Fedora 39/x86-64.
lto-wrapper.cc: Add offload target name to 'offload_args' suffix
lto-wrapper.cc's compile_offload_image calls mkoffload with
an @./a.offload_args argument ('a.' in case of, e.g., 'a.out'). However,
when generating code for both nvptx and gcn, they use the same name
with -save-temps. Hence, this commit adds a <target> + '.' before
'offload_args' in line with other offload-target-specific files.
gcc/ChangeLog:
* lto-wrapper.cc (compile_offload_image): Prefix 'offload_args'
suffix by the target name.
GCN: Fix --with-arch= handling in mkoffload [PR111966]
The default -march= setting used in mkoffload did not reflect the modified
default set by GCC's configure-time --with-arch=, causing issues when
generating debug code.
gcc/ChangeLog:
PR other/111966
* config/gcn/mkoffload.cc (get_arch): New; moved -march= flag
handling from ...
(main): ... here; call it to handle --with-arch config option
and -march= commandline.
Jakub Jelinek [Wed, 3 Apr 2024 08:02:35 +0000 (10:02 +0200)]
libquadmath: Don't assume the storage for __float128 arguments is aligned [PR114533]
With the register_printf_type/register_printf_modifier/register_printf_specifier
APIs the C library is just told the size of the argument and is provided with
a callback to fetch the argument from va_list using va_arg into C library provided
memory. The C library isn't told what alignment requirement it has, but we were
using direct load of a __float128 value from that memory which assumes
__alignof (__float128) alignment.
The following patch fixes that by using memcpy instead.
I haven't been able to reproduce an actual crash, tried
#include <quadmath.h>
#include <stdlib.h>
#include <stdio.h>
int main ()
{
__float128 r;
int prec = 20;
int width = 46;
char buf[128];
r = 2.0q;
r = sqrtq (r);
int n = quadmath_snprintf (buf, sizeof buf, "%+-#*.20Qe", width, r);
if ((size_t) n < sizeof buf)
printf ("%s\n", buf);
/* Prints: +1.41421356237309504880e+00 */
quadmath_snprintf (buf, sizeof buf, "%Qa", r);
if ((size_t) n < sizeof buf)
printf ("%s\n", buf);
/* Prints: 0x1.6a09e667f3bcc908b2fb1366ea96p+0 */
n = quadmath_snprintf (NULL, 0, "%+-#46.*Qe", prec, r);
if (n > -1)
{
char *str = malloc (n + 1);
if (str)
{
quadmath_snprintf (str, n + 1, "%+-#46.*Qe", prec, r);
printf ("%s\n", str);
/* Prints: +1.41421356237309504880e+00 */
}
free (str);
}
printf ("%+-#*.20Qe\n", width, r);
printf ("%Qa\n", r);
printf ("%+-#46.*Qe\n", prec, r);
printf ("%d %Qe %d %Qe %d %Qe\n", 1, r, 2, r, 3, r);
return 0;
}
In any case, I think memcpy for loading from it is right.
2024-04-03 Simon Chopin <simon.chopin@canonical.com>
Jakub Jelinek <jakub@redhat.com>
PR libquadmath/114533
* printf/printf_fp.c (__quadmath_printf_fp): Use memcpy to copy
__float128 out of args.
* printf/printf_fphex.c (__quadmath_printf_fphex): Likewise.
Signed-off-by: Simon Chopin <simon.chopin@canonical.com>
Jakub Jelinek [Wed, 3 Apr 2024 07:59:45 +0000 (09:59 +0200)]
expr: Fix up emit_push_insn [PR114552]
r13-990 added optimizations in multiple spots to optimize during
expansion storing of constant initializers into targets.
In the load_register_parameters and expand_expr_real_1 cases,
it checks it has a tree as the source and so knows we are reading
that whole decl's value, so the code is fine as is, but in the
emit_push_insn case it checks for a MEM from which something
is pushed and checks for SYMBOL_REF as the MEM's address, but
still assumes the whole object is copied, which as the following
testcase shows might not always be the case. In the testcase,
k is 6 bytes, then 2 bytes of padding, then another 4 bytes,
while the emit_push_insn wants to store just the 6 bytes.
The following patch simply verifies it is the whole initializer
that is being stored, I think that is best thing to do so late
in GCC 14 cycle as well for backporting.
For GCC 15, perhaps the code could stop requiring it must be at offset zero,
nor that the size is equal, but could use
get_symbol_constant_value/fold_ctor_reference gimple-fold APIs to actually
extract just part of the initializer if we e.g. push just some subset
(of course, still verify that it is a subset). For sizes which are power
of two bytes and we have some integer modes, we could use as type for
fold_ctor_reference corresponding integral types, otherwise dunno, punt
or use some structure (e.g. try to find one in the initializer?), whatever.
But even in the other spots it could perhaps handle loading of
COMPONENT_REFs or MEM_REFs from the .rodata vars.
2024-04-03 Jakub Jelinek <jakub@redhat.com>
PR middle-end/114552
* expr.cc (emit_push_insn): Only use store_constructor for
immediate_const_ctor_p if int_expr_size matches size.
Richard Biener [Tue, 2 Apr 2024 10:31:04 +0000 (12:31 +0200)]
tree-optimization/114557 - reduce ehcleanup peak memory use
The following reduces peak memory use for the PR114480 testcase at -O1
which is almost exclusively spent by the ehcleanup pass in allocating
PHI nodes. The free_phinodes cache we maintain isn't very effective
since it has effectively two slots, one for 4 and one for 9 argument
PHIs and it is only ever used for allocations up to 9 arguments but
we put all larger PHIs in the 9 argument bucket. This proves
uneffective resulting in much garbage to be kept when incrementally
growing PHI nodes by edge redirection.
The mitigation is to rely on the GC freelist for larger sizes and
thus immediately return all larger bucket sized PHIs to it via ggc_free.
This reduces the peak memory use from 19.8GB to 11.3GB and compile-time
from 359s to 168s.
PR tree-optimization/114557
PR tree-optimization/114480
* tree-phinodes.cc (release_phi_node): Return PHIs from
allocation buckets not covered by free_phinodes to GC.
(remove_phi_node): Release the PHI LHS before freeing the
PHI node.
* tree-vect-loop.cc (vectorizable_live_operation): Get PHI lhs
before releasing it.
Gaius Mulley [Tue, 2 Apr 2024 22:47:42 +0000 (23:47 +0100)]
PR modula2/114565 progress trace would be useful to isolate ICE for users
This patch introduces the internal option -fm2-debug-trace= which can
be given a comma separated list of filter terms. Currently it allows:
all,line,token,quad. The patch allows users to trace the progress of
cc1gm2 so that source which causes an ICE can be reduced. Once
PR113836 is complete it is expected that the trace information will be
written to file.
gcc/m2/ChangeLog:
PR modula2/114565
* gm2-compiler/M2GenGCC.mod (CodeStatement): Test
GetDebugTraceQuad before calling DisplayQuad.
* gm2-compiler/M2LexBuf.mod (NumberIO): Import CardToStr.
(GetToken): Test GetDebugTraceToken before writing the
token number or token line.
* gm2-compiler/M2Options.def (SetDebugTraceQuad): Rename to
(SetM2DebugTraceFilter): ...this.
(SetDebugTraceAPI): Remove.
(GetDebugTraceQuad): New procedure function.
(GetDebugTraceTree): Ditto.
(GetDebugTraceToken): Ditto.
(GetDebugTraceLine): Ditto.
(GetDebugFunctionLineNumbers): Ditto.
* gm2-compiler/M2Options.mod (DebugFunctionLineNumbers): New
boolean variable.
(DebugTraceQuad): Ditto.
(DebugTraceTree): Ditto.
(DebugTraceLine): Ditto.
(DebugTraceToken): Ditto.
(errors1): New procedure.
(SetDebugTraceQuad): Remove.
(SetM2DebugTraceFilter): New procedure implemented.
(SetM2DebugTrace): Ditto.
(GetDebugTraceQuad): Ditto.
(GetDebugTraceToken ): Ditto.
(GetDebugTraceLine): Ditto.
(SetDebugTraceLine): Remove.
* gm2-compiler/M2Quads.mod (GenQuadOTrash): Test
GetDebugTraceQuad and call DisplayQuad.
(GenQuadOTypetok): Ditto.
* gm2-compiler/SymbolTable.mod: Replace
DebugFunctionLineNumbers with GetDebugFunctionLineNumbers.
* gm2-gcc/init.cc (_M2_M2LangDump_init): Add prototype.
(init_PerCompilationInit): Add call.
* gm2-gcc/m2misc.cc (m2misc_cerror): New function.
(m2misc_error): Ditto.
* gm2-gcc/m2misc.def (error): New procedure.
(cerror): Ditto.
* gm2-gcc/m2misc.h (m2misc_cerror): New prototype.
(m2misc_error): Ditto.
* gm2-gcc/m2options.h (M2Options_SetDebugTraceQuad): New
prototype.
(M2Options_SetDebugTraceAPI): Remove.
(M2Options_GetDebugTraceToken): New prototype.
(M2Options_GetDebugTraceLine): Ditto.
(M2Options_SetDebugFunctionLineNumbers): Ditto.
(M2Options_GetDebugFunctionLineNumbers): Ditto.
(M2Options_SetM2DebugTraceFilter): Ditto.
* gm2-lang.cc (gm2_langhook_init_options): Remove
OPT_fdebug_trace_quad case.
Remove OPT_fdebug_trace_api case.
Add OPT_fm2_debug_trace_ case.
* lang.opt (fm2-debug-trace): New option.
(fdebug-trace-api): Remove.
(fdebug-trace-quad): Remove.
* m2.flex (m2flex_M2Error): Check s for NULL.
(skipnewline): New function.
(consumeLine): Call traceline.
Jonathan Wakely [Thu, 28 Mar 2024 21:29:17 +0000 (21:29 +0000)]
libstdc++: Guard uses of char8_t with __cpp_char8_t [PR114519]
libstdc++-v3/ChangeLog:
PR libstdc++/114519
* include/bits/unicode.h (_Utf8_view): Guard with check for
char8_t being enabled.
(__literal_encoding_is_unicode): Guard use of char8_t with check
for it being enabled.
* testsuite/std/format/functions/114519.cc: New test.
Tom Tromey [Sat, 30 Mar 2024 19:48:30 +0000 (13:48 -0600)]
libiberty: Invoke D demangler when --format=auto
Investigating GDB PR d/31580 showed that the libiberty demangler
doesn't automatically demangle D mangled names. However, I think it
should -- like C++ and Rust (new-style), D mangled names are readily
distinguished by the leading "_D", and so the likelihood of confusion
is low. The other non-"auto" cases in this code are Ada (where the
encoded form could more easily be confused by ordinary programs) and
Java (which is long gone, but which also shared the C++ mangling and
thus was just an output style preference).
This patch also fixed another GDB bug, though of course that part
won't apply to the GCC repository.
Marek Polacek [Mon, 1 Apr 2024 16:55:46 +0000 (12:55 -0400)]
c++: make __is_array return false for T[0] [PR114479]
When we switched to using the __is_array built-in trait to implement
std::is_array in r14-6623-g7fd9c349e45534, we started saying that
T[0] is an array. There are various opinions as to whether that is
the best answer, but it seems prudent to keep the GCC 13 result.
PR c++/114479
gcc/cp/ChangeLog:
* semantics.cc (trait_expr_value) <case CPTK_IS_ARRAY>: Return false
for zero-sized arrays.
Marek Polacek [Fri, 29 Mar 2024 20:59:37 +0000 (16:59 -0400)]
c++: ICE with scoped enum in switch condition [PR103825]
Here we ICE when gimplifying
enum class Type { Pawn };
struct Piece {
Type type : 4;
};
void foo() {
switch (Piece().type)
case Type::Pawn:;
}
because we ended up with TYPE_PRECISION (cond) < TYPE_PRECISION (case).
That's because the case expr type here is the unlowered type Type,
whereas the conditional's type is the lowered <unnamed-signed:4>. This
is not supposed to happen: see the comment in pop_switch around the
is_bitfield_expr_with_lowered_type check.
But here we did not revert to the lowered SWITCH_STMT_TYPE, because
the conditional contains a TARGET_EXPR, which has side-effects, which
means that finish_switch_cond -> maybe_cleanup_point_expr wraps it
in a CLEANUP_POINT_EXPR. And is_bitfield_expr_with_lowered_type does
not see through those.
Jason Merrill [Tue, 2 Apr 2024 14:52:28 +0000 (10:52 -0400)]
c++: binding reference to comma expr [PR114561]
We represent a reference binding where the referent type is more qualified
by a ck_ref_bind around a ck_qual. We performed the ck_qual and then tried
to undo it with STRIP_NOPS, but that doesn't work if the conversion is
buried in COMPOUND_EXPR. So instead let's avoid performing that fake
conversion in the first place.
PR c++/114561
PR c++/114562
gcc/cp/ChangeLog:
* call.cc (convert_like_internal): Avoid adding qualification
conversion in direct reference binding.
gcc/testsuite/ChangeLog:
* g++.dg/conversion/ref10.C: New test.
* g++.dg/conversion/ref11.C: New test.
Patrick Palka [Tue, 2 Apr 2024 17:07:07 +0000 (13:07 -0400)]
libstdc++: Allow adjacent __maybe_present_t<false, ...> fields to overlap
Currently __maybe_present_t<false, T> maps to the same empty class
type independent of T. This is suboptimal because it means adjacent
__maybe_present_t<false, ...> members with the [[no_unique_address]]
attribute can't overlap even if the conditionally present types are
different.
This patch turns this empty class type into a template parameterized by
the conditionally present type, so that
This patch goes a step further and also adds an optional integer
discriminator parameter to allow for overlapping when T and U are
the same.
libstdc++-v3/ChangeLog:
* include/std/ranges (ranges::__detail::_Empty): Rename to ...
(ranges::__detail::_Absent): ... this. Turn into a template
parameterized by the absent type _Tp and discriminator _Disc.
(ranges::__detail::__maybe_present_t): Add an optional
discriminator parameter.
(slide_view::_M_cached_begin): Pass a discriminator argument to
__maybe_present_t.
(slide_view::_M_cached_end): Likewise.
* testsuite/std/ranges/adaptors/sizeof.cc: Verify the size of
slide_view<V> is 3 instead 4 pointers.
I think those ", 0" should not appear on their own lines. The issue
seems to be that print_dw_val should not generally emit a newline,
except when recursing.
gcc/ChangeLog
* dwarf2out.cc (print_dw_val) <dw_val_class_loc>: Don't
print newline when not recursing.
Paul Thomas [Tue, 2 Apr 2024 14:53:29 +0000 (15:53 +0100)]
Fortran: Add error for subroutine passed to a variable dummy [PR106999]
2024-04-02 Paul Thomas <pault@gcc.gnu.org>
gcc/fortran
PR fortran/106999
* interface.cc (gfc_compare_interfaces): Add error for a
subroutine proc pointer passed to a variable formal.
(compare_parameter): If a procedure pointer is being passed to
a non-procedure formal arg, and there is an an interface, use
gfc_compare_interfaces to check and provide a more useful error
message.
gcc/testsuite/
PR fortran/106999
* gfortran.dg/pr106999.f90: New test.
Paul Thomas [Tue, 2 Apr 2024 13:19:09 +0000 (14:19 +0100)]
Fortran: Fix wrong recursive errors and class initialization [PR112407]
2024-04-02 Paul Thomas <pault@gcc.gnu.org>
gcc/fortran
PR fortran/112407
* resolve.cc (resolve_procedure_expression): Change the test for
for recursion in the case of hidden procedures from modules.
(resolve_typebound_static): Add warning for possible recursive
calls to typebound procedures.
* trans-expr.cc (gfc_trans_class_init_assign): Do not apply
default initializer to class dummy where component initializers
are all null.
gcc/testsuite/
PR fortran/112407
* gfortran.dg/pr112407a.f90: New test.
* gfortran.dg/pr112407b.f90: New test.
Jakub Jelinek [Tue, 2 Apr 2024 11:40:27 +0000 (13:40 +0200)]
Fix up postboot dependencies [PR106472]
On Wed, Mar 13, 2024 at 10:13:37AM +0100, Jakub Jelinek wrote:
> While the first Makefile.tpl hunk looks obviously ok, the others look
> completely wrong to me.
> There is nothing special about libgo vs. libbacktrace/libatomic
> compared to any other target library which is not bootstrapped vs. any
> of its dependencies which are in the bootstrapped set.
> So, Makefile.tpl shouldn't hardcode such dependencies.
Here is my version of the fix.
The dependencies in the toplevel Makefile simply didn't take into account
that some target modules could be in a bootstrapped build built in some
configurations as bootstrap modules (typically as dependencies of other
target bootstrap modules), while in other configurations just as
dependencies of non-bootstrap target modules and so not built during the
bootstrap, but after it.
Makefile.tpl arranges for those postboot target module -> target module
dependencies to be emitted only inside of an @unless gcc-bootstrap block,
while for @if gcc-bootstrap it just emits
configure-target-whatever: stage_last
dependencies which ensure those postbootstrap target modules are only built
after everything that is bootstrapped has been.
Now, the libbacktrace/libatomic target modules have bootstrap=true
target_modules = { module= libbacktrace; bootstrap=true; };
target_modules = { module= libatomic; bootstrap=true; lib_path=.libs; };
because those modules are dependencies of libphobos target module, so
when d is included among bootstrapped languages, those are all bootstrapped
and everything works correctly.
While if d is not included, libphobos target module is disabled,
libbacktrace/libatomic target modules aren't bootstrapped, nothing during
bootstrap needs them, but post bootstrap libgo target module depends on
the libatomic and libbacktrace target modules, libgfortran target module
depends on the libbacktrace target module and libgm2 target module depends
on the libatomic target module, but those dependencies were emitted only
@unless gcc-bootstrap. There is a similar theoretical problem for zlib
target module if GCJ would be ressurected, libphobos as bootstrap target
module depends on the zlib target module, but if d is not configured,
fastjar also depends on it.
The following patch arranges for the @if gcc-bootstrap case to emit also
target module -> target module dependencies, but conditionally on the
on dependency not being bootstrapped.
In the generated Makefile.in you can see what the Makefile.tpl change
produces and that it just adds extra dependencies which weren't there
before in the @if gcc-bootstrap case.
I've bootstrapped without this patch with
../configure --enable-languages=c,c++,go; make
on x86_64-linux (note, make -j2 or higher usually worked) which failed
as described in the PR, then with this patch with the same command which
built fine and the Makefile difference between the two builds being
diff -up obj40{a,b}/Makefile
--- obj40a/Makefile 2024-03-31 00:35:22.243791499 +0100
+++ obj40b/Makefile 2024-03-31 22:40:38.143299144 +0200
@@ -29376,6 +29376,14 @@ configure-bison: stage_last
configure-flex: stage_last
configure-m4: stage_last
# Dependencies for target modules on other target modules are
# described by lang_env_dependencies; the defaults apply to anything
which I believe are exactly the extra dependencies we want.
Plus I've done normal x86_64-linux and i686-linux bootstraps/regtests
which in my case include --enable-languages=default,ada,obj-c++,lto,go,d,rust,m2
for x86_64 and the same except ada for i686; those with my usual make -j32.
The Makefile difference in those builds vs. unpatched case
is just an extra empty line.
2024-04-02 Jakub Jelinek <jakub@redhat.com>
PR bootstrap/106472
* Makefile.tpl (make-postboot-target-dep): New lambda.
Use it to add --enable-bootstrap dependencies of target modules
on other target modules if the latter aren't bootstrapped.
* Makefile.in: Regenerate.
This is just part of the changes, mostly for non-gcc directories.
I'll try to get to the rest soon. Obviously, the above command also
finds cases which are correct as is and shouldn't be changed, so one
needs to manually inspect everything.
I'd hope most of it is pretty obvious, but the config/ and libstdc++-v3/
hunks include a tweak in a license wording, though other copies of the
similar license have the wording right.
2024-04-02 Jakub Jelinek <jakub@redhat.com>
* Makefile.tpl: Fix duplicated words; returns returns ->
returns.
config/
* lcmessage.m4: Fix duplicated words; can can -> can,
package package -> package.
libdecnumber/
* decCommon.c (decFinalize): Fix duplicated words in
comment; the the -> the.
libgcc/
* unwind-dw2-fde.c (struct fde_accumulator): Fix duplicated
words in comment; is is -> is.
libgfortran/
* configure.host: Fix duplicated words; the the -> the.
libgm2/
* configure.host: Fix duplicated words; the the -> the.
libgomp/
* libgomp.texi (OpenMP 5.2): Fix duplicated words; with with ->
with.
(omp_target_associate_ptr): Fix duplicated words; either either ->
either.
(omp_init_allocator): Fix duplicated words; be be -> be.
(omp_realloc): Fix duplicated words; is is -> is.
(OMP_ALLOCATOR): Fix duplicated words; other other -> other.
* priority_queue.h (priority_queue_multi_p): Fix duplicated words;
to to -> to.
libiberty/
* regex.c (byte_re_match_2_internal): Fix duplicated words in comment;
next next -> next.
* dyn-string.c (dyn_string_init): Fix duplicated words in comment;
of of -> of.
libitm/
* beginend.cc (GTM::gtm_thread::begin_transaction): Fix duplicated
words in comment; not not -> not to.
libobjc/
* init.c (duplicate_classes): Fix duplicated words in comment; in in
-> in.
* sendmsg.c (__objc_prepare_dtable_for_class): Fix duplicated words
in comment; the the -> the.
* encoding.c (objc_layout_structure): Likewise.
libstdc++-v3/
* acinclude.m4: Fix duplicated words; file file -> file can.
* configure.host: Fix duplicated words; the the -> the.
libvtv/
* vtv_rts.cc (vtv_fail): Fix duplicated words; to to -> to.
* vtv_fail.cc (vtv_fail): Likewise.
Iain Sandoe [Sat, 13 Jan 2024 17:20:47 +0000 (17:20 +0000)]
jit, Darwin: Implement library exports list.
Currently, we have no exports list for libgccjit, which means that
all symbols are exported, including those from libstdc++ which is
linked statically into the lib. This causes failures when the
shared libstdc++ is used but some c++ symbols are satisfied from
libgccjit.
This implements an export file for Darwin (which is currently
manually created by cross-checking libgccjit.map). Ideally we'd
script this, at some point. Update libtool current and age to
reflect the current ABI version (we are not bumping the SO name
at this stage).
This fixes a number of new failures in jit testing.
gcc/jit/ChangeLog:
* Make-lang.in: Implement exports list, and use a shared
libgcc.
* libgccjit.exports: New file.
Iain Sandoe [Sun, 31 Mar 2024 10:22:58 +0000 (11:22 +0100)]
testsuite: Remove duplicate -lgcov [PR114034]
Duplicate library entries now cause linker warnings with newer linker
versions on Darwin which leads to these tests regressing. The library
is already added by the test flags so there is no need to put an extra
one in the options.
PR testsuite/114034
gcc/testsuite/ChangeLog:
* g++.dg/gcov/gcov-dump-1.C: Remove extra -lgcov.
* g++.dg/gcov/gcov-dump-2.C: Likewise.
When the version for dsymutil comes from a clang build, it is
of the form NNmm.pp.qq where NN and mm are the major and minor
LLVM version components. We need to check for a major version
greater than or equal to 7 - so use 700 in the check.
gcc/ChangeLog:
* config/darwin.cc (darwin_override_options): Update the
clang major version value in the dsymutil check.
Iain Sandoe [Sun, 31 Mar 2024 22:25:31 +0000 (23:25 +0100)]
Darwin: Do not emit .macinfo when dsymutil cannot consume it.
Some verions of dsymutil do not ignore .macinfo sections, but instead
ignore the entire debug in the file.
To avoid this total loss of debug, when we detect that the debug level
is g3 and the dsymutil version cannot support it, we reduce the level
to g2 and issue a note.
This behaviour can be overidden by -gstrict-dwarf (although the objects
will contain macinfo; dsymutil will not produce a .dSYM with it).
gcc/ChangeLog:
* config/darwin.cc (darwin_override_options): Reduce the debug
level to 2 if dsymutil cannot handle .macinfo sections.
Iain Sandoe [Fri, 8 Mar 2024 09:21:39 +0000 (09:21 +0000)]
testsuite, Darwin: Update bad-mapper-1 after libiberty changes.
A recent change to libiberty has improved the process spawning on
older Darwin platforms. This patch updates the expected test output
after the changes.
gcc/testsuite/ChangeLog:
* g++.dg/modules/bad-mapper-1.C: Update expected test output
for earlier Darwin.
The default is still traditional TLS model, but can be configured with
--with-tls={trad,desc}. The default can change to TLS descriptors once
libc and LLVM support this.
gcc/ChangeLog:
* config.gcc: Add --with-tls option to change TLS flavor.
* config/loongarch/genopts/loongarch.opt.in: Add -mtls-dialect to
configure TLS flavor.
* config/loongarch/loongarch-def.h (struct loongarch_target): Add
tls_dialect.
* config/loongarch/loongarch-driver.cc (la_driver_init): Add tls
flavor.
* config/loongarch/loongarch-opts.cc (loongarch_init_target): Add
tls_dialect.
(loongarch_config_target): Ditto.
(loongarch_update_gcc_opt_status): Ditto.
* config/loongarch/loongarch-opts.h (loongarch_init_target): Ditto.
(TARGET_TLS_DESC): New define.
* config/loongarch/loongarch.cc (loongarch_symbol_insns): Add TLS
DESC instructions sequence length.
(loongarch_legitimize_tls_address): New TLS DESC instruction sequence.
(loongarch_option_override_internal): Add la_opt_tls_dialect.
(loongarch_option_restore): Add la_target.tls_dialect.
* config/loongarch/loongarch.md (@got_load_tls_desc<mode>): Normal
code model for TLS DESC.
(got_load_tls_desc_off64): Extreme cmode model for TLS DESC.
* config/loongarch/loongarch.opt: Regenerate.
* config/loongarch/loongarch.opt.urls: Ditto.
* doc/invoke.texi: Add a description of the compilation option
'-mtls-dialect={trad,desc}'.
Gaius Mulley [Mon, 1 Apr 2024 18:18:36 +0000 (19:18 +0100)]
PR modula2/114548 gm2 fails to identify variable in a const expression
This patch introduces stricter checking within standard procedure
functions which detect whether paramaters are variable when used
in a const expression.
PR modula2/114548
* gm2/iso/const/fail/expression.mod: New test.
* gm2/iso/const/fail/iso-const-fail.exp: New test.
* gm2/iso/const/fail/testabs.mod: New test.
* gm2/iso/const/fail/testaddadr.mod: New test.
* gm2/iso/const/fail/testcap.mod: New test.
* gm2/iso/const/fail/testcap2.mod: New test.
* gm2/iso/const/fail/testchr.mod: New test.
* gm2/iso/const/fail/testchr2.mod: New test.
* gm2/iso/const/fail/testcmplx.mod: New test.
* gm2/iso/const/fail/testfloat.mod: New test.
* gm2/iso/const/fail/testim.mod: New test.
* gm2/iso/const/fail/testint.mod: New test.
* gm2/iso/const/fail/testlength.mod: New test.
* gm2/iso/const/fail/testodd.mod: New test.
* gm2/iso/const/fail/testord.mod: New test.
* gm2/iso/const/fail/testre.mod: New test.
* gm2/iso/const/fail/testtrunc.mod: New test.
* gm2/iso/const/fail/testval.mod: New test.
* gm2/iso/const/pass/constbool.mod: New test.
* gm2/iso/const/pass/constbool2.mod: New test.
* gm2/iso/const/pass/constbool3.mod: New test.
Jason Merrill [Fri, 29 Mar 2024 01:33:57 +0000 (21:33 -0400)]
c++: C++26 returning reference to temporary
P2748R5 makes it ill-formed to return a reference to temporary in C++26;
implementing this is a simple matter of changing the existing warning to a
permerror.
For most of the tests I just changed dg-warning to dg-message to accept
both; I test the specific diagnostic type in Wreturn-local-addr-5.C.
gcc/cp/ChangeLog:
* typeck.cc (maybe_warn_about_returning_address_of_local):
Permerror in C++26.
gcc/testsuite/ChangeLog:
* g++.dg/conversion/pr16333.C: Change dg-warning to dg-message.
* g++.dg/cpp0x/constexpr-48324.C
* g++.dg/other/pr94326.C
* g++.dg/warn/Wreturn-local-addr-2.C
* g++.old-deja/g++.jason/warning8.C: Likewise.
* g++.dg/cpp1y/auto-fn6.C: Check that others don't complain.
* g++.dg/warn/Wreturn-local-addr-5.C: Expect error in C++26.
Yang Yujie [Sat, 30 Mar 2024 08:43:14 +0000 (16:43 +0800)]
LoongArch: Split loongarch_option_override_internal into smaller procedures
gcc/ChangeLog:
* config/loongarch/genopts/loongarch.opt.in: Mark -m[no-]recip as
aliases to -mrecip={all,none}, respectively.
* config/loongarch/loongarch.opt: Regenerate.
* config/loongarch/loongarch-def.h (ABI_FPU_64): Rename to...
(ABI_FPU64_P): ...this.
(ABI_FPU_32): Rename to...
(ABI_FPU32_P): ...this.
(ABI_FPU_NONE): Rename to...
(ABI_NOFPU_P): ...this.
(ABI_LP64_P): Define.
* config/loongarch/loongarch.cc (loongarch_init_print_operand_punct):
Merged into loongarch_global_init.
(loongarch_cpu_option_override): Renamed to
loongarch_target_option_override.
(loongarch_option_override_internal): Move the work after
loongarch_config_target into loongarch_target_option_override.
(loongarch_global_init): Define.
(INIT_TARGET_FLAG): Move to loongarch-opts.cc.
(loongarch_option_override): Call loongarch_global_init
separately.
* config/loongarch/loongarch-opts.cc (loongarch_parse_mrecip_scheme):
Split the parsing of -mrecip=<string> from
loongarch_option_override_internal.
(loongarch_generate_mrecip_scheme): Define. Split from
loongarch_option_override_internal.
(loongarch_target_option_override): Define. Renamed from
loongarch_cpu_option_override.
(loongarch_init_misc_options): Define. Split from
loongarch_option_override_internal.
(INIT_TARGET_FLAG): Move from loongarch.cc.
* config/loongarch/loongarch-opts.h (loongarch_target_option_override):
New prototype.
(loongarch_parse_mrecip_scheme): New prototype.
(loongarch_init_misc_options): New prototype.
(TARGET_ABI_LP64): Simplify with ABI_LP64_P.
* config/loongarch/loongarch.h (TARGET_RECIP_DIV): Simplify.
Do not reference specific CPU architecture (LA664).
(TARGET_RECIP_SQRT): Same.
(TARGET_RECIP_RSQRT): Same.
(TARGET_RECIP_VEC_DIV): Same.
(TARGET_RECIP_VEC_SQRT): Same.
(TARGET_RECIP_VEC_RSQRT): Same.
Jeff Law [Sun, 31 Mar 2024 16:51:17 +0000 (10:51 -0600)]
[committed] RISC-V: Add missing insn types to XiangShan Nanhu scheduler model
The test for the recently added XiangShan Nanhu microarchitecture is failing
because the scheduler description does not have entries for certain insn types.
I'm adding branch, jalr, ret and sfb_alu to the scheduler description, that's
enough to get the trivial test to pass. However, I strongly suspect running
any significant code through the compiler when scheduling for this
microarchitecture will trigger faults.
Basically we have checking now that will fault if we have an insn in the IL
without an associated type or if we have an insn in the IL that does not map to
an insn reservation in the scheduler model. We were tripping the latter
assertion for one of those branch types. My suspicion is many insn types
aren't handled by that DFA.
The branch insns were pretty obvious and easy to fix. But someone with more
experience with the uarch needs to do an audit to ensure that all insn types
map to an insn reservation.
gcc/
* config/riscv/xiangshan.md (xiangshan_jump): Add branch, jalr, ret
and sfb_alu.
Pan Li [Sat, 30 Mar 2024 12:03:18 +0000 (20:03 +0800)]
RISC-V: Fix misspelled term builtin in error message
This patch would like to fix below misspelled term in error message.
../../gcc/config/riscv/riscv-vector-builtins.cc:4592:16: error:
misspelled term 'builtin function' in format; use 'built-in function' instead [-Werror=format-diag]
4592 | "builtin function %qE requires the V ISA extension", exp);
The below tests are passed for this patch.
* The riscv regression test on rvv.exp and riscv.exp.
gcc/ChangeLog:
* config/riscv/riscv-vector-builtins.cc (expand_builtin): Take
the term built-in over builtin.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/base/target_attribute_v_with_intrinsic-7.c:
Adjust test dg-error.
* gcc.target/riscv/rvv/base/target_attribute_v_with_intrinsic-8.c:
Ditto.
Xi Ruoyao [Wed, 20 Mar 2024 07:09:21 +0000 (15:09 +0800)]
mips: Fix C23 (...) functions returning large aggregates [PR114175]
We were assuming TYPE_NO_NAMED_ARGS_STDARG_P don't have any named
arguments and there is nothing to advance, but that is not the case
for (...) functions returning by hidden reference which have one such
artificial argument. This is causing gcc.dg/c23-stdarg-{6,8,9}.c to
fail.
Fix the issue by checking if arg.type is NULL, as r14-9503 explains.
gcc/ChangeLog:
PR target/114175
* config/mips/mips.cc (mips_setup_incoming_varargs): Only skip
mips_function_arg_advance for TYPE_NO_NAMED_ARGS_STDARG_P
functions if arg.type is NULL.
Arsen Arsenović [Fri, 29 Mar 2024 11:42:19 +0000 (12:42 +0100)]
libstdc++-v3: drop GCC Runtime Library Exception from gen tests
It was mistakenly added to these files.
libstdc++-v3/ChangeLog:
* testsuite/24_iterators/range_generators/01.cc: Drop GCC
Runtime Library Exception.
* testsuite/24_iterators/range_generators/02.cc: Drop GCC
Runtime Library Exception.
* testsuite/24_iterators/range_generators/copy.cc: Drop GCC
Runtime Library Exception.
* testsuite/24_iterators/range_generators/except.cc: Drop GCC
Runtime Library Exception.
* testsuite/24_iterators/range_generators/subrange.cc: Drop GCC
Runtime Library Exception.
* testsuite/24_iterators/range_generators/synopsis.cc: Drop GCC
Runtime Library Exception.
* testsuite/24_iterators/range_generators/iter_deref_return.cc:
Drop GCC Runtime Library Exception from the "You should have
received a copy" paragraph.
Thomas Schwinge [Wed, 20 Mar 2024 22:52:26 +0000 (23:52 +0100)]
GCN: Enable effective-target 'vect_hw_misalign'
... as made apparent by commit 4e1fcf44bdc582e71408175d75e025f5be8b0e55
"testsuite: vect: Require vect_hw_misalign in gcc.dg/vect/vect-cost-model-1.c etc. [PR98238]"
causing:
PASS: gcc.dg/vect/pr113431.c (test for excess errors)
PASS: gcc.dg/vect/pr113431.c execution test
-PASS: gcc.dg/vect/pr113431.c scan-tree-dump-times slp1 "optimized: basic block part vectorized" 2
..., which this commit all restores, and also enables a good number of further
FAIL -> PASS, UNSUPPORTED -> PASS, etc. progressions. There are also a small
number of regressions, mostly in the SLP area apparently:
..., so some further conditionalizing etc. seems necessary. These seem to
mostly appear next to pre-existing similar FAILs in related test cases.
(Overall, way more PASS than FAIL.)
gcc/testsuite/
* lib/target-supports.exp
(check_effective_target_vect_hw_misalign): Enable for GCN.
(check_effective_target_vect_element_align): Adjust.
PASS: gcc.dg/vect/tsvc/vect-tsvc-s332.c (test for excess errors)
PASS: gcc.dg/vect/tsvc/vect-tsvc-s332.c execution test
[-XFAIL:-]{+XPASS:+} gcc.dg/vect/tsvc/vect-tsvc-s332.c scan-tree-dump vect "vectorized 1 loops"
PASS: gcc.dg/vect/tsvc/vect-tsvc-s481.c (test for excess errors)
PASS: gcc.dg/vect/tsvc/vect-tsvc-s481.c execution test
[-XFAIL:-]{+XPASS:+} gcc.dg/vect/tsvc/vect-tsvc-s481.c scan-tree-dump vect "vectorized 1 loops"
PASS: gcc.dg/vect/tsvc/vect-tsvc-s482.c (test for excess errors)
PASS: gcc.dg/vect/tsvc/vect-tsvc-s482.c execution test
[-XFAIL:-]{+XPASS:+} gcc.dg/vect/tsvc/vect-tsvc-s482.c scan-tree-dump vect "vectorized 1 loops"
..., it became apparent that GCN, too, does support vectorization of loops with
early breaks. The relevant test cases are all-PASS with just the following
exceptions, to be looked into individually, later on:
PASS: gcc.dg/vect/vect-early-break_25.c (test for excess errors)
PASS: gcc.dg/vect/vect-early-break_25.c scan-tree-dump-times vect "vectorized 1 loops" 1
FAIL: gcc.dg/vect/vect-early-break_25.c scan-tree-dump-times vect "Alignment of access forced using peeling" 1
PASS: gcc.dg/vect/vect-early-break_56.c (test for excess errors)
PASS: gcc.dg/vect/vect-early-break_56.c execution test
XPASS: gcc.dg/vect/vect-early-break_56.c scan-tree-dump-times vect "vectorized 2 loops" 2
gcc/testsuite/
* lib/target-supports.exp
(check_effective_target_vect_early_break)
(check_effective_target_vect_early_break_hw): Enable for GCN.
Paul Thomas [Fri, 29 Mar 2024 07:23:19 +0000 (07:23 +0000)]
Fortran: Fix a gimplifier ICE/wrong result with finalization [PR36337]
2024-03-29 Paul Thomas <pault@gcc.gnu.org>
gcc/fortran
PR fortran/36337
PR fortran/110987
PR fortran/113885
* trans-expr.cc (gfc_trans_assignment_1): Place finalization
block before rhs post block for elemental rhs.
* trans.cc (gfc_finalize_tree_expr): Check directly if a type
has no components, rather than the zero components attribute.
Treat elemental zero component expressions in the same way as
scalars.
gcc/testsuite/
PR fortran/113885
* gfortran.dg/finalize_54.f90: New test.
* gfortran.dg/finalize_55.f90: New test.
gcc/testsuite/
PR fortran/110987
* gfortran.dg/finalize_56.f90: New test.
Xi Ruoyao [Tue, 6 Feb 2024 09:49:50 +0000 (17:49 +0800)]
testsuite: Add a test case for negating FP vectors containing zeros
Recently I've fixed two wrong FP vector negate implementation which
caused wrong sign bits in zeros in targets (r14-8786 and r14-8801). To
prevent a similar issue from happening again, add a test case.
Tested on x86_64 (with SSE2, AVX, AVX2, and AVX512F), AArch64, MIPS
(with MSA), LoongArch (with LSX and LASX).
Jeff Law [Thu, 28 Mar 2024 22:56:53 +0000 (16:56 -0600)]
[committed] Provide suitable output template for zero_extendqihi2 on H8
Segher's recent combine change, quite unexpectedly, triggered a regression on
the H8 port. It failed to build newlib.
The zero_extendqihi2 pattern provided two alternatives. One where the source
and destination matched. That turns into a suitable instruction trivially.
The second alternative was actually meant to capture cases where the value is
coming from memory.
What was missing here was the reg->reg case where the source and destination do
not match. That fell into the second case which was requested to be split by
the pattern's output template.
The splitter had a suitable condition to make sure it only triggered in the
right cases. Unfortunately with the pattern requiring a split in a case where
the splitter was going to fail led to the fault.
So regardless of what's going on in the combiner, this code was just wrong.
Fixed thusly by providing a suitable output template for the reg->reg case.
Regression tested on h8300-elf. Pushing to the trunk.
gcc/
* config/h8300/extensions.md (zero_extendqihi*): Add output
template for reg->reg case where the regs don't match.
Jason Merrill [Wed, 27 Mar 2024 20:14:01 +0000 (16:14 -0400)]
c++: __is_constructible ref binding [PR100667]
The requirement that a type argument be complete is excessive in the case of
direct reference binding to the same type, which does not rely on any
properties of the type. This is LWG 2939.
PR c++/100667
gcc/cp/ChangeLog:
* semantics.cc (same_type_ref_bind_p): New.
(finish_trait_expr): Use it.
Harald Anlauf [Wed, 27 Mar 2024 20:18:04 +0000 (21:18 +0100)]
Fortran: fix DATA and derived types with pointer components [PR114474]
When matching actual arguments in match_actual_arg, these are initially
treated as a possible dummy procedure, assuming that the correct type is
determined later. This resolution could fail when the procedure is a
derived type constructor with a pointer component and appears in a DATA
statement, where the pointer shall be associated with an initial data
target. Check for those cases where the type obviously has not been
resolved yet, and which were missed because there was no component
reference.
gcc/fortran/ChangeLog:
PR fortran/114474
* primary.cc (gfc_variable_attr): Catch variables used in structure
constructors within DATA statements that are still tagged with a
temporary type BT_PROCEDURE from match_actual_arg and which have the
target attribute, and fix their typespec.
gcc/testsuite/ChangeLog:
PR fortran/114474
* gfortran.dg/data_pointer_3.f90: New test.
Vineet Gupta [Wed, 27 Mar 2024 21:55:04 +0000 (14:55 -0700)]
RISC-V: testsuite: ensure vtype is call clobbered
Per classic Vector calling convention ABI, vtype is call clobbered,
so ensure gcc regenerates a VSETVLI in following cases:
- after a function call.
- after an inline asm fragment which clobbers vtype.
ATM gcc seems to be doing the right thing, but a test can never hurt.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/vtype-call-clobbered.c: New Test.
Gaius Mulley [Thu, 28 Mar 2024 16:49:44 +0000 (16:49 +0000)]
PR modula2/114520 Incorrect ordering of import/export statements cause confusion
The error recovery causes misleading error messages to appear if an
EXPORT and IMPORT statement are in the wrong order. This patch
detects the incorrect order and issues an error message and prevents
error recovery. The fix should be improved and made more general if
another similar case is required.
gcc/m2/ChangeLog:
PR modula2/114520
* gm2-compiler/P0SyntaxCheck.bnf (DetectImport): New
procedure.
(EnableImportCheck): New boolean.
(Expect): Call DetectImport.
(Export): Set EnableImportCheck TRUE before ';' and FALSE
afterwards.
Gaius Mulley [Thu, 28 Mar 2024 14:57:49 +0000 (14:57 +0000)]
PR modula2/114517 gm2 does not allow comparison operator hash in column one
This patch allows -fno-cpp to be supplied to gm2. Without this patch
it causes an ICE. The patch allows -fno-cpp to turn off cpp flags.
These are tested in m2.flex to decide whether a change of state is
allowed (enabling handling of #line directives).
gcc/ChangeLog:
PR modula2/114517
* doc/gm2.texi: Mention gm2 treats a # in the first column
as a preprocessor directive unless -fno-cpp is supplied.
gcc/m2/ChangeLog:
PR modula2/114517
* gm2-compiler/M2Options.def (SetCpp): Add comment.
(GetCpp): Move after SetCpp.
(GetLineDirectives): New procedure function.
* gm2-compiler/M2Options.mod (GetLineDirectives): New
procedure function.
* gm2-gcc/m2options.h (M2Options_GetLineDirectives): New
prototype.
* gm2-lang.cc (gm2_langhook_init_options): OPT_fcpp only
assert if !value.
* m2.flex: Test GetLineDirectives before changing to LINE0
state.
gcc/testsuite/ChangeLog:
PR modula2/114517
* gm2/cpp/fail/hashfirstcolumn2.mod: New test.
* gm2/imports/fail/imports-fail.exp: New test.
* gm2/imports/fail/localmodule2.mod: New test.
* gm2/imports/run/pass/localmodule.mod: New test.
Jakub Jelinek [Thu, 28 Mar 2024 14:00:44 +0000 (15:00 +0100)]
profile-count: Avoid overflows into uninitialized [PR112303]
The testcase in the patch ICEs with
--- gcc/tree-scalar-evolution.cc
+++ gcc/tree-scalar-evolution.cc
@@ -3881,7 +3881,7 @@ final_value_replacement_loop (class loop *loop)
/* Propagate constants immediately, but leave an unused initialization
around to avoid invalidating the SCEV cache. */
- if (CONSTANT_CLASS_P (def) && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (rslt))
+ if (0 && CONSTANT_CLASS_P (def) && !SSA_NAME_OCCURS_IN_ABNORMAL_PHI (rslt))
replace_uses_by (rslt, def);
/* Create the replacement statements. */
(the addition of the above made the ICE latent), because profile_count
addition doesn't check for overflows and if unlucky, we can even overflow
into the uninitialized value.
Getting really huge profile counts is very easy even when not using
recursive inlining in loops, e.g.
__attribute__((noipa)) void
bar (void)
{
__builtin_exit (0);
}
__attribute__((noipa)) void
foo (void)
{
for (int i = 0; i < 1000; ++i)
for (int j = 0; j < 1000; ++j)
for (int k = 0; k < 1000; ++k)
for (int l = 0; l < 1000; ++l)
for (int m = 0; m < 1000; ++m)
for (int n = 0; n < 1000; ++n)
for (int o = 0; o < 1000; ++o)
for (int p = 0; p < 1000; ++p)
for (int q = 0; q < 1000; ++q)
for (int r = 0; r < 1000; ++r)
for (int s = 0; s < 1000; ++s)
for (int t = 0; t < 1000; ++t)
for (int u = 0; u < 1000; ++u)
for (int v = 0; v < 1000; ++v)
for (int w = 0; w < 1000; ++w)
for (int x = 0; x < 1000; ++x)
for (int y = 0; y < 1000; ++y)
for (int z = 0; z < 1000; ++z)
for (int a = 0; a < 1000; ++a)
for (int b = 0; b < 1000; ++b)
bar ();
}
int
main ()
{
foo ();
}
reaches the maximum count already on the 11th loop.
Some other methods of profile_count like apply_scale already
do use MIN (val, max_count) before assignment to m_val, this patch
just extends that to operator{+,+=} methods.
Furthermore, one overload of apply_probability wasn't using
safe_scale_64bit and so could very easily overflow as well
- prob is required to be [0, 10000] and if m_val is near the max_count,
it can overflow even with multiplications by 8.
2024-03-28 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/112303
* profile-count.h (profile_count::operator+): Perform
addition in uint64_t variable and set m_val to MIN of that
val and max_count.
(profile_count::operator+=): Likewise.
(profile_count::operator-=): Formatting fix.
(profile_count::apply_probability): Use safe_scale_64bit
even in the int overload.
Maxim Kuvyrkov [Wed, 13 Mar 2024 06:48:47 +0000 (06:48 +0000)]
[testsuite] Fixup dg-options in {gcc,g++,gfortran}.dg/vect.exp tests
Testsuites driven by vect.exp rely on check_vect_support_and_set_flags
to set appropriate DEFAULT_VECTFLAGS for a given target (e.g., add
-mfpu=neon for arm-linux-gnueabi). Unfortunately, these flags are
overwritten by dg-options directive, which can cause tests to fail.
Behavior of dg-options is documented in vect.exp files, but not
all developers look at the .exp file when adding a new testcase.
This caused a few dg-options directives to be used instead of
the more appropriate dg-additional-options.
This patch changes target-independent dg-options into
dg-additional-options. This patch does not touch target-specific
dg-options and target-specific tests to avoid disturbing the gentle
balance of target-specific vectorization.
This patch also removes a couple of unneeded "dg-do run" directives
to avoid failures on compile-only targets. Default action is, again,
set by check_vect_support_and_set_flags.
Lastly, I avoided renaming tests that use -O<n> options to O<n>-*
filename format because this support is not consistent between
gcc.dg/vect/, g++.dg/vect/, and gfortran.dg/vect/ testsuites.
It seems dg-additional-options is cleaner.
This patch does the following,
1. do not change target-specific tests, e.g., gcc.dg/vect/costmodel/riscv/*;
2. do not change { dg-options FOO { target { target-*-pattern } } };
3. do not remove { dg-do run { target { target-*-pattern } } };
4. change { dg-options FOO } to { dg-additional-options FOO };
5. remove { dg-do run } in several tests, where it is clearly not needed.
Mikael Morin [Wed, 27 Mar 2024 15:30:42 +0000 (16:30 +0100)]
fortran: Fix specification expression check in submodules [PR114475]
The patch fixing PR111781 made the check of specification expressions more
restrictive, disallowing local variables in specification expressions of
dummy arguments. PR114475 showed an example where that change regressed,
disallowing in submodules expressions that had been allowed in the parent
module. In submodules indeed, the hierarchy of namespaces inherited from
the parent module is not reproduced so the host-association of symbols
can't be recognized by checking the nesting of namespaces.
This change fixes the problem by allowing in specification expressions
all the symbols in a submodule that are inherited from the parent module.
PR fortran/111781
PR fortran/114475
gcc/fortran/ChangeLog:
* expr.cc (check_restricted): In submodules, allow variables host-
associated from the parent module.
Palmer Dabbelt [Wed, 27 Mar 2024 19:54:04 +0000 (12:54 -0700)]
RISC-V: Add vxsat as a register
We aren't doing anything with vxsat right now, but I'd like to add it as
an accepted register to the clobber list. If we get this into GCC-14
then we'll avoid some preprocessor-based twiddling if we ever start
using vxsat in the future.
David Malcolm [Wed, 27 Mar 2024 22:26:51 +0000 (18:26 -0400)]
analyzer: fix ICE due to type mismatch when replaying call summary [PR114473]
gcc/analyzer/ChangeLog:
PR analyzer/114473
* call-summary.cc
(call_summary_replay::convert_svalue_from_summary): Assert that
the types match.
(call_summary_replay::convert_region_from_summary): Likewise.
(call_summary_replay::convert_region_from_summary_1): Add missing
cast for the deref of RK_SYMBOLIC case.
gcc/testsuite/ChangeLog:
PR analyzer/114473
* gcc.dg/analyzer/call-summaries-pr114473.c: New test.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
Jakub Jelinek [Wed, 27 Mar 2024 19:22:02 +0000 (20:22 +0100)]
btf: Fix up btf-datasec-1.c test on x86
> -/* The offset entry for each variable in a DATSEC should be 0 at compile time. */
> -/* { dg-final { scan-assembler-times "0\[\t \]+\[^\n\]*bts_offset" 7 } } */
> +/* The offset entry for each variable in a DATSEC should contain a label. */
> +/* { dg-final { scan-assembler-times ".4byte\[\t \]\[a-e\]\[\t \]+\[^\n\]*bts_offset" 5 } } */
4byte is used only on some targets, what exact assembler directive is used
for 4byte unaligned data is heavily target dependent.
2024-03-27 Jakub Jelinek <jakub@redhat.com>
* gcc.dg/debug/btf/btf-cvr-quals-1.c: Use dg-additional-options
instead of multiple dg-options.
* gcc.dg/debug/btf/btf-datasec-1.c: Likewise. Accept all supported
unaligned 4 byte assembler directives rather than assuming it must
be .4byte.
Jakub Jelinek [Wed, 27 Mar 2024 18:38:06 +0000 (19:38 +0100)]
c-family: Cast __atomic_load_*/__atomic_exchange_* result to _BitInt rather then VCE it [PR114469]
As written in the PR, torture/bitint-64.c test fails with -O2 -flto
and the reason is that on _BitInt arches where the padding bits
are undefined, the padding bits in the _Atomic vars are also undefined,
but when __atomic_load or __atomic_exchange on a _BitInt _Atomic variable
with some padding bits is lowered into __atomic_load_{1,2,4,8,16} or
__atomic_exchange_*, the mode precision unsigned result is VIEW_CONVERT_EXPR
converted to _BitInt and because of the VCE nothing actually sign/zero
extends it as needed for later uses - the var is no longer addressable and
expansion assumes such automatic vars are properly extended.
The following patch fixes that by using NOP_EXPR on it (the
VIEW_CONVERT_EXPR after it will then be optimized away during
gimplification, didn't want to repeat it in the code as else result = build1
(VIEW_CONVERT_EXPR, ...); twice.
2024-03-27 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/114469
* c-common.cc (resolve_overloaded_builtin): For _BitInt result
on !extended targets convert result to the _BitInt type before
using VIEW_CONVERT_EXPR.
In some cases combine will "combine" an I2 and I3, but end up putting
exactly the same thing back as I2 as was there before. This is never
progress, so we shouldn't do it, it will lead to oscillating behaviour
and the like.
If we want to canonicalise things, that's fine, but this is not the
way to do it.
Jakub Jelinek [Wed, 27 Mar 2024 14:41:59 +0000 (15:41 +0100)]
docs: Use @var{S} etc. in Spec File invoke.texi documentation
We got internally a question about the Spec File syntax, misunderstanding
what is the literal syntax and what are the placeholder variables in
the syntax descriptions.
The following patch attempts to use @var{S} etc. instead of just S
to clarify it stands for any option (or start of option etc.) rather
than literal S, say in %{S:X}. At least in HTML documentation it
then uses italics.
2024-03-27 Jakub Jelinek <jakub@redhat.com>
* doc/invoke.texi (Spec Files): Use @var{S} instead of S,
@var{X} instead of X etc. for other placeholders.
* include/experimental/bits/simd_x86.h (_S_masked_unary):
Cast inputs < 16 bytes to 16 byte vectors before calling the
right subtraction builtin. Before returning, truncate to the
return vector type.
* include/experimental/bits/simd_x86.h (_S_masked_unary): Call
the 4- and 8-byte variants of __builtin_ia32_subp[ds] without
rounding direction argument.
libstdc++: add ARM SVE support to std::experimental::simd
libstdc++-v3/ChangeLog:
* include/Makefile.am: Add simd_sve.h.
* include/Makefile.in: Add simd_sve.h.
* include/experimental/bits/simd.h: Add new SveAbi.
* include/experimental/bits/simd_builtin.h: Use
__no_sve_deduce_t to support existing Neon Abi.
* include/experimental/bits/simd_converter.h: Convert
sequentially when sve is available.
* include/experimental/bits/simd_detail.h: Define sve
specific macro.
* include/experimental/bits/simd_math.h: Fallback frexp
to execute sequntially when sve is available, to handle
fixed_size_simd return type that always uses sve.
* include/experimental/simd: Include bits/simd_sve.h.
* testsuite/experimental/simd/tests/bits/main.h: Enable
testing for sve128, sve256, sve512.
* include/experimental/bits/simd_sve.h: New file.
Richard Biener [Wed, 27 Mar 2024 10:37:16 +0000 (11:37 +0100)]
tree-optimization/114057 - handle BB reduction remain defs as LIVE
The following makes sure to record the scalars we add to the BB
reduction vectorization result as scalar uses for the purpose of
computing live lanes. This restores vectorization in the
bondfree.c TU of 435.gromacs.
PR tree-optimization/114057
* tree-vect-slp.cc (vect_bb_slp_mark_live_stmts): Mark
BB reduction remain defs as scalar uses.
Jakub Jelinek [Wed, 27 Mar 2024 11:00:58 +0000 (12:00 +0100)]
testsuite: Fix up ext-floating{3,12}.C on i686-linux
These tests FAIL for quite a while on i686-linux since July last year,
likely r14-2628 . Since that patch gcc claims _Float16 and __bf16
support even without -msse2 because some functions could be using
target attribute.
Later r14-2691 added -msse2 to add_options_for_float16, but didn't do that
for bfloat16, plus ext-floating{3,12}.C tests need the added dg-add-options,
so that float16 and bfloat16 effective targets match the __STDCPP_FLOAT16_T__
or __STDCPP_BFLOAT16_T__ macros.
Fixes
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++23 (test for errors, line 144)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++23 (test for errors, line 146)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++23 (test for errors, line 148)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++23 (test for errors, line 150)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++23 (test for errors, line 152)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++23 (test for errors, line 154)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++26 (test for errors, line 144)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++26 (test for errors, line 146)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++26 (test for errors, line 148)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++26 (test for errors, line 150)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++26 (test for errors, line 152)
-FAIL: g++.dg/cpp23/ext-floating12.C -std=gnu++26 (test for errors, line 154)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for errors, line 107)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for errors, line 114)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for errors, line 126)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for errors, line 79)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for errors, line 86)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for errors, line 98)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for warnings, line 22)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for warnings, line 23)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for warnings, line 24)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++23 (test for warnings, line 25)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for errors, line 107)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for errors, line 114)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for errors, line 126)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for errors, line 79)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for errors, line 86)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for errors, line 98)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for warnings, line 22)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for warnings, line 23)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for warnings, line 24)
-FAIL: g++.dg/cpp23/ext-floating3.C -std=gnu++26 (test for warnings, line 25)
on the latter and changes nothing on the former.
2024-03-27 Jakub Jelinek <jakub@redhat.com>
* lib/target-supports.exp (add_options_for_bfloat16): Add -msse2 on
i?86/x86_64.
* g++.dg/cpp23/ext-floating3.C: Add dg-add-options float16.
* g++.dg/cpp23/ext-floating12.C: Add dg-add-options float16 and
bfloat16.
projects / thirdparty / gcc.git / log
