The r16-142-g01e5ef3e8b9128 chagned return type of _Str_sink::view()
to basic_string_view<_CharT>. The mutable access is provided by _M_span
function, that is now used for mingw path.
libstdc++-v3/ChangeLog:
* include/std/ostream (vprint_unicode) [_WIN32 && !__CYGWIN__]: Call
_Str
This is a response to https://gcc.gnu.org/bugzilla/show_bug.cgi?id=14940#c57
The patch was submitted to MSYS2 for testing in 2022-5. No issue reports have
been received so far:
*
https://github.com/msys2/MINGW-packages/blob/455762a45250642b90e9ff34020d001c1be09015/mingw-w64-gcc/0021-PR14940-Al
Hello, gentle maintainer.
This is a message from the Translation Project robot.
A revised PO file for textual domain 'gcc' has been submitted
by the Swedish team of translators. The file is available at:
https://translationproject.org/latest/gcc/sv.po
(This file, 'gcc-15.1.0.sv.po', has ju
On Fri, Mar 28, 2025 at 12:20:18PM +0100, Jakub Jelinek wrote:
> Here is the incremental patch I was talking about.
> For noop sets, we don't need to test much, they can go to i2
> unless that would violate i3 JUMP condition.
>
> With this the try_combine on the pr119291.c testcase doesn't fail,
>
On Fri, Apr 25, 2025 at 09:12:45PM -0700, Andrew Pinski wrote:
> --- a/gcc/tree-tailcall.cc
> +++ b/gcc/tree-tailcall.cc
> @@ -1083,57 +1083,74 @@ find_tail_calls (basic_block bb, struct tailcall
> **ret, bool only_musttail,
> {
>bool ok = false;
>value_range val;
> - tre
Make sense to me, it looks like the combine will always take place if GR2VR
is 0, 1 or 2 for now.
I am try to customize the cost here to make it fail to combine but get failed
with below change.
+ if (rcode == VEC_DUPLICATE && SCALAR_INT_MODE_P (GET_MODE (XEXP (x, 0 {
+cost_val = 1;
+
The following rewords the documentation for -Og which over-promises
the ability to debug the generated code. While -Og enables
var-tracking and thus improves debugging in some areas the experience
is usually worse than -O0 for standard C code.
Any other comments/clarifications? OK?
Thanks,
Rich
Hi All,
The following testcase shows an incorrect masked codegen:
#define N 512
#define START 1
#define END 505
int x[N] __attribute__((aligned(32)));
int __attribute__((noipa))
foo (void)
{
int z = 0;
for (unsigned int i = START; i < END; ++i)
{
z++;
if (x[i] > 0)
Hi All,
When the input is already a subreg and we try to make a paradoxical
subreg out of it for copysign this can fail if it violates the subreg
relationship.
Use force_lowpart_subreg instead of lowpart_subreg to then force the
results to a register instead of ICEing.
Bootstrapped Regtested on
Hi Jeevitha,
Looks like Peter's review comments have not been incorporated.
Please update the patch and, as Peter suggested, send the next version of the
patch as it's own email thread.
Regards,
Surya
On 03/04/25 8:17 pm, jeevitha wrote:
>
> Ping!
>
> please review.
>
> Thanks & Regards
> Jee
From: Owen Avery
gcc/testsuite/ChangeLog:
* rust/compile/derive-debug1.rs: Adjust a path.
* rust/compile/nr2/exclude: Remove derive-debug1.rs.
Signed-off-by: Owen Avery
---
gcc/testsuite/rust/compile/derive-debug1.rs | 2 +-
gcc/testsuite/rust/compile/nr2/exclude | 1 -
2
From: Owen Avery
gcc/rust/ChangeLog:
* resolve/rust-forever-stack.hxx
(ForeverStack::resolve_path): Pass instance of Node to lambda by
reference instead of by value.
Signed-off-by: Owen Avery
---
gcc/rust/resolve/rust-forever-stack.hxx | 2 +-
1 file changed, 1 inserti
From: Pierre-Emmanuel Patry
This commit introduce a new public function to visit function parameters
in the default visitor. It allows visitors derived from DefaultVisitor
to override only a small part of the default visitor.
gcc/rust/ChangeLog:
* ast/rust-ast-visitor.cc (DefaultASTVisi
On Mon, 28 Apr 2025 at 16:22, Patrick Palka wrote:
>
> On Mon, 28 Apr 2025, Jonathan Wakely wrote:
>
> > On Mon, 28 Apr 2025 at 14:59, Patrick Palka wrote:
> > >
> > > Tested on x86_64-pc-linux-gnu, does this look OK for trunk/15?
> > >
> > > -- >8 --
> > >
> > > These std::erase_if overloads wer
From: Owen Avery
Fixes PR#119641
gcc/rust/ChangeLog:
* checks/errors/borrowck/rust-bir-place.h
(IndexVec::size_type): Add.
(IndexVec::MAX_INDEX): Add.
(IndexVec::size): Change the return type to the type of the
internal value used by the index type.
From: Pierre-Emmanuel Patry
gcc/testsuite/ChangeLog:
* rust/compile/multiple_bindings1.rs: Add missing lang items.
Signed-off-by: Pierre-Emmanuel Patry
---
.../rust/compile/multiple_bindings1.rs| 29 ---
1 file changed, 19 insertions(+), 10 deletions(-)
diff -
From: Pierre-Emmanuel Patry
gcc/rust/ChangeLog:
* ast/rust-ast.h: Add hash function.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/rust/ast/rust-ast.h | 13 +
1 file changed, 13 insertions(+)
diff --git a/gcc/rust/ast/rust-ast.h b/gcc/rust/ast/rust-ast.h
index 94585dc1344.
On Mon, Apr 28, 2025 at 9:05 AM Andrew Pinski wrote:
>
> On Mon, Apr 28, 2025 at 8:48 AM wrote:
> >
> > From: Pierre-Emmanuel Patry
> >
> > InlineAsm node does not support memory clobbers.
>
> A few review on this.
> I think this should just be called extended rather than referencing LLVM here.
From: Pierre-Emmanuel Patry
gcc/rust/ChangeLog:
* resolve/rust-late-name-resolver-2.0.cc (Late::visit): Add binding
creation in visitor.
* resolve/rust-late-name-resolver-2.0.h: Add function prototypes.
* resolve/rust-name-resolution-context.h: Add binding context
Dne 28. 04. 25 v 10:58 Jakub Jelinek napsal(a):
On Sun, Apr 27, 2025 at 11:56:21AM +0200, Josef Melcr wrote:
* builtin-attrs.def (0): New int list.
This is just weird. Write
* builtin-attrs.def: Add DEF_LIST_INT_INT (0,2).
?
+DEF_CALLBACK_ATTRIBUTE(GOMP, 1, 0)
+DEF_CALLBACK_
From: Philip Herron
We were wrongly adding the assertion that this must not be an enum but
this is a pointless assertion we only care that there are variant in the
ADT and if the field exists in the first variant.
Fixes Rust-GCC#3581
gcc/rust/ChangeLog:
* typecheck/rust-hir-type-check-
From: Pierre-Emmanuel Patry
gcc/testsuite/ChangeLog:
* rust/compile/black_box.rs: New test.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/testsuite/rust/compile/black_box.rs | 28 +
1 file changed, 28 insertions(+)
create mode 100644 gcc/testsuite/rust/compile/
From: Philip Herron
Fixes Rust-GCC#3652
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude: nr2 does not error on the T it should
require Self::T
* rust/compile/issue-3652.rs: New test.
Signed-off-by: Philip Herron
---
gcc/testsuite/rust/compile/issue-3652.rs | 7 +++
g
From: Pierre-Emmanuel Patry
gcc/rust/ChangeLog:
* expand/rust-macro-builtins-asm.cc (parse_asm_arg): Emit error
message.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/rust/expand/rust-macro-builtins-asm.cc | 8
1 file changed, 8 insertions(+)
diff --git a/gcc/rust/exp
From: Philip Herron
The error handling here was done long ago when we didnt know how to do
any error handling very well. This removed bad fatal_errors and adds in
some nice rich_location error diagnostics instead.
Fixes Rust-GCC#3628
gcc/rust/ChangeLog:
* typecheck/rust-hir-type-check-
From: Pierre-Emmanuel Patry
Both nodes had the same id, this led to a resolution conflict.
gcc/rust/ChangeLog:
* expand/rust-derive-clone.cc (DeriveClone::clone_enum_struct): Clone
path to avoid using the same nodeid.
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude
From: Pierre-Emmanuel Patry
Default visitor should visit all it's children.
gcc/rust/ChangeLog:
* ast/rust-ast-visitor.cc (DefaultASTVisitor::visit): Visit visibility.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/rust/ast/rust-ast-visitor.cc | 1 +
1 file changed, 1 insertion(+)
dif
From: Pierre-Emmanuel Patry
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude: Remove passing test from exclusion list.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/testsuite/rust/compile/nr2/exclude | 2 --
1 file changed, 2 deletions(-)
diff --git a/gcc/testsuite/rust/compile/nr2
From: Pierre-Emmanuel Patry
gcc/rust/ChangeLog:
* resolve/rust-default-resolver.cc (DefaultResolver::visit): Add visit
function for TypeParam.
* resolve/rust-default-resolver.h: Add function prototype.
* resolve/rust-forever-stack.h: Add function to check for forw
From: Owen Avery
gcc/rust/ChangeLog:
* ast/rust-ast-visitor.cc
(DefaultASTVisitor::visit): Visit the loop labels of
WhileLetLoopExpr instances before visiting their scrutinee
expressions.
* resolve/rust-early-name-resolver-2.0.cc
(Early::resolve_gl
On Mon, Apr 28, 2025 at 8:48 AM wrote:
>
> From: Pierre-Emmanuel Patry
>
> InlineAsm node does not support memory clobbers.
A few review on this.
I think this should just be called extended rather than referencing LLVM here.
>
> gcc/rust/ChangeLog:
>
> * ast/rust-ast-collector.cc (Token
From: Pierre-Emmanuel Patry
gcc/rust/ChangeLog:
* ast/rust-ast-collector.cc (TokenCollector::visit): Remove error kind
and change function call.
* ast/rust-ast-visitor.cc (DefaultASTVisitor::visit): Change call name.
* ast/rust-path.cc (ConstGenericParam::as_strin
Rebased to GCC 15.1
This patch introduces inline definitions for the __fma and __fmaf
functions in arm_acle.h for arm targets. These definitions rely on
__builtin_fma and __builtin_fmaf to ensure proper inlining and to meet
the ACLE requirements [1].
The patch has been tested locally using a cros
Hi
Pretty straightforward except for the aligned_storage one for which I
just avoid the check when gnu-versioned-namespace mode is being used.
libstdc++: [_GLIBCXX_INLINE_VERSION] Fix several tests failures
Several tests are failing when libstdc++ is configured with:
--enable-symv
On Mon, Apr 28, 2025 at 1:13 AM Richard Biener
wrote:
>
> On Sun, Apr 27, 2025 at 12:51 AM Andrew Pinski
> wrote:
> >
> > While looking into adding __ROTATE_LEFT and __ROTATE_RIGHT, I noticed
> > this code is just a bunch of if statments repeated. Instead we could just
> > use a simple lookup ar
Hi Andrew,
On 4/28/25 6:05 PM, Andrew Pinski wrote:
On Mon, Apr 28, 2025 at 8:48 AM wrote:
From: Pierre-Emmanuel Patry
InlineAsm node does not support memory clobbers.
A few review on this.
I think this should just be called extended rather than referencing LLVM here.
This node exists s
Hi,
On 4/28/25 6:09 PM, Andrew Pinski wrote:
On Mon, Apr 28, 2025 at 9:05 AM Andrew Pinski wrote:
On Mon, Apr 28, 2025 at 8:48 AM wrote:
From: Pierre-Emmanuel Patry
InlineAsm node does not support memory clobbers.
A few review on this.
I think this should just be called extended rather
From: Philip Herron
Trait constants were missing type resolution step, this adds that
as if it was a normal constant. The unsafe checker was missing a
null check.
Fixes Rust-GCC#3612
gcc/rust/ChangeLog:
* checks/errors/rust-unsafe-checker.cc (UnsafeChecker::visit): add null
check
From: Pierre-Emmanuel Patry
InlineAsm node does not support memory clobbers.
gcc/rust/ChangeLog:
* ast/rust-ast-collector.cc (TokenCollector::visit): Make visitor
unreachable.
* ast/rust-ast-collector.h: Add visit for LlvmInlineAsmNode.
* ast/rust-ast-visitor.cc
From: Zhi Heng
gcc/rust/ChangeLog:
* typecheck/rust-tyty.h: Add new `ReprKind` enum to
`ReprOptions`.
* typecheck/rust-hir-type-check-base.cc: Handle setting of
`repr_kind`.
* typecheck/rust-hir-type-check-item.cc: New check for
zero-variant enums.
From: Philip Herron
I copied a bad form of this check from the c front-end this updates it
to ensure the rhs is an INTEGER_CST and the lhs needs checked in the first
place.
Fixes Rust-GCC#3664
gcc/rust/ChangeLog:
* rust-gcc.cc (arithmetic_or_logical_expression): Ensure this is an
inte
From: Pierre-Emmanuel Patry
Add a new HIR LlvmInlineAsm HIR node as well as some structures to
represent it's options and operands. Lower AST::LlvmInlineAsm node to it
and then create a tree from that node.
gcc/rust/ChangeLog:
* ast/rust-ast-collector.cc (TokenCollector::visit): Remove
From: Owen Avery
gcc/rust/ChangeLog:
* resolve/rust-ast-resolve-path.cc
(ResolvePath::resolve_path): Adjust error messages.
* resolve/rust-ast-resolve-type.cc
(ResolveRelativeTypePath::go): Likewise.
* resolve/rust-forever-stack.hxx
(check_leading_
From: Pierre-Emmanuel Patry
gcc/rust/ChangeLog:
* ast/rust-ast.h: Add equality operator.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/rust/ast/rust-ast.h | 5 +
1 file changed, 5 insertions(+)
diff --git a/gcc/rust/ast/rust-ast.h b/gcc/rust/ast/rust-ast.h
index 55f178d782c..94585
From: Owen Avery
gcc/rust/ChangeLog:
* resolve/rust-early-name-resolver-2.0.cc
(Early::build_import_mapping): Avoid outputting an "unresolved
import" error if other errors are outputted during resolution.
* resolve/rust-early-name-resolver-2.0.h
(Early::re
On Mon, 28 Apr 2025, Jonathan Wakely wrote:
> On Mon, 28 Apr 2025 at 14:59, Patrick Palka wrote:
> >
> > Tested on x86_64-pc-linux-gnu, does this look OK for trunk/15?
> >
> > -- >8 --
> >
> > These std::erase_if overloads were wrongly implemented as hidden
> > friends, visible only via ADL, so e
On 15/04/2025 19:48, Jan Hubicka wrote:
> Hi,
> > gcc/ChangeLog:
> >
> > PR tree-optimization/117790
> > * tree-vect-loop.cc (scale_profile_for_vect_loop): Use
> > scale_loop_profile_hold_exit_counts instead of scale_loop_profile. Drop
> > the exit edge parameter, since the code n
From: Owen Avery
gcc/rust/ChangeLog:
* resolve/rust-late-name-resolver-2.0.cc
(Late::visit): Handle StructPatternFieldIdent.
* resolve/rust-late-name-resolver-2.0.h
(Late::visit): Likewise.
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude: Remove entr
From: Pierre-Emmanuel Patry
gcc/rust/ChangeLog:
* ast/rust-ast-collector.cc (TokenCollector::visit): Dump llvm inline
asm tokens.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/rust/ast/rust-ast-collector.cc | 39 +-
1 file changed, 38 insertions(+),
From: Philip Herron
gcc/rust/ChangeLog:
* hir/rust-hir-dump.cc (Dump::visit): add guard for optional label
Signed-off-by: Philip Herron
---
gcc/rust/hir/rust-hir-dump.cc | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/gcc/rust/hir/rust-hir-dump.cc b/gcc/rust/hir
From: Zhi Heng
gcc/rust/ChangeLog:
* typecheck/rust-hir-type-check-base.cc: Set enum representing
type properly if repr is an integer type.
* typecheck/rust-hir-type-check-item.cc: Update comments.
Signed-off-by: Yap Zhi Heng
---
gcc/rust/typecheck/rust-hir-type-check-
From: Pierre-Emmanuel Patry
gcc/testsuite/ChangeLog:
* rust/execute/black_box.rs: New test.
Signed-off-by: Pierre-Emmanuel Patry
---
gcc/testsuite/rust/execute/black_box.rs | 30 +
1 file changed, 30 insertions(+)
create mode 100644 gcc/testsuite/rust/execute/
From: Philip Herron
Fixes Rust-GCC#3662
gcc/testsuite/ChangeLog:
* rust/compile/issue-3662.rs: New test.
Signed-off-by: Philip Herron
---
gcc/testsuite/rust/compile/issue-3662.rs | 8
1 file changed, 8 insertions(+)
create mode 100644 gcc/testsuite/rust/compile/issue-3662.r
From: Philip Herron
It is not allowed to have a declared inference variable in the return
position of a function as this may never get infered you need good points
of truth.
Ideally if we get a student for GSoC 25 we will get the Default Hir Visitor
so that we can grab the HIR::InferredType locu
From: Philip Herron
This was already fixed in: bb01719f0e1 but we require fn_once lang item
to be defined as we are working on libcore support still.
Fixes Rust-GCC#3711
gcc/testsuite/ChangeLog:
* rust/compile/issue-3711.rs: New test.
Signed-off-by: Philip Herron
---
gcc/testsuite/r
From: Pierre-Emmanuel Patry
Name resolution 2.0 message describes the context around the unresolved
items and should therefore be kept.
gcc/testsuite/ChangeLog:
* rust/compile/nr2/exclude: Remove test from exclusion list.
* rust/compile/use_1.rs: Change expected output and remov
From: Pierre-Emmanuel Patry
Binding context may be stacked when a new binding group is introduced
within a const expression.
gcc/rust/ChangeLog:
* resolve/rust-name-resolution-context.h: Use BindingLayer instead.
* resolve/rust-name-resolution-context.cc (BindingLayer::BindingLa
From: Philip Herron
There are two cases when initilizing an array, this is the
const context which means we need to build the array ctor,
which means using lots of memory, its super inefficient
because we are using a big wrapper over the GCC internals here
but preallocating the vectors here cause
On Sun, Apr 27, 2025 at 11:56:21AM +0200, Josef Melcr wrote:
> * builtin-attrs.def (0): New int list.
This is just weird. Write
* builtin-attrs.def: Add DEF_LIST_INT_INT (0,2).
?
> +DEF_CALLBACK_ATTRIBUTE(GOMP, 1, 0)
> +DEF_CALLBACK_ATTRIBUTE(GOMP, 1, 2)
> +DEF_CALLBACK_ATTRIBUTE(O
On Mon, Apr 28, 2025 at 4:26 PM H.J. Lu wrote:
>
> > > This is what my patch does:
> > But it iterates through vector_insns, using a def-ref chain to find
> > those insns. I think we can just record those single_set with src as
> > const_m1/zero, and replace src for them.
>
> Will fix it.
Fixed
On Thu, Apr 10, 2025 at 4:09 PM Richard Biener wrote:
>
> The following removes the ability to switch back to non SLP-only
> operation of the vectorizer - a requirement to start cleaning out
> non-SLP paths without risk of regressing that case.
I have pushed this series now.
Richard.
>
On Thu, 3 Apr 2025 at 13:58, Rasmus Villemoes wrote:
>
> In many setups, especially when CI and/or some meta-build system like
> Yocto or buildroot, is involved, gcc ends up being invoked using
> absolute path names, which are often long and uninteresting.
>
> That amounts to a lot of noise both w
On Tue, Feb 18, 2025 at 1:44 PM Richard Biener wrote:
>
> The following changes how flag_complex_method is managed towards
> being able to record that in the optimization set so we can stream
> and restore it per function. Currently -fcx-fortran-rules and
> -fcx-limited-range are separate recorde
On Sat, Dec 14, 2024 at 9:22 PM Jeff Law wrote:
>
>
>
> On 12/3/24 10:07 AM, Richard Biener wrote:
> > Turns out SLP discovery can end up doing a lot of vector type
> > builds from scalar types. Those are all ultimatively cached but
> > end up built and layouted first. The latter is particularly
On Mon, 28 Apr 2025, Alexander Monakov wrote:
>
> On Mon, 28 Apr 2025, Richard Biener wrote:
>
> > The following rewords the documentation for -Og which over-promises
> > the ability to debug the generated code. While -Og enables
> > var-tracking and thus improves debugging in some areas the ex
From: Pan Li
This patch will add testcase for unsigned integer SAT_ADD form 7:
#define DEF_VEC_SAT_U_ADD_FMT_9(WT, T) \
void __attribute__((noinline)) \
vec_sat_u_add_##WT##_##T##_fmt_9 (T *out, T *op_1, T *o
From: Pan Li
This patch would like to support the form 7 of the unsigned
integer SAT_ADD, aka below example.
#define DEF_SAT_U_ADD_FMT_7(WT, T) \
T __attribute__((noinline))\
sat_u_add_##WT##_##T##_fmt_7(T x, T y) \
{ \
T max = -1;
Kyrylo Tkachov writes:
>> On 25 Apr 2025, at 19:55, Richard Sandiford
>> wrote:
>>
>> Jennifer Schmitz writes:
>>> If -msve-vector-bits=128, SVE loads and stores (LD1 and ST1) with a
>>> ptrue predicate can be replaced by neon instructions (LDR and STR),
>>> thus avoiding the predicate altoget
From: Pan Li
This patch serices would like to support form 7 of the unsigned integer
SAT_ADD. Different to another forms of SAT_ADD, the form 7 will
leverage a wider type to tell overflow or not, aka:
#define DEF_SAT_U_ADD_FMT_7(WT, T) \
T __attribute__((noinline))\
sat_u_
From: Pan Li
This patch will add testcase for unsigned integer SAT_ADD form 7:
#define DEF_SAT_U_ADD_FMT_7(WT, T) \
T __attribute__((noinline))\
sat_u_add_##WT##_##T##_fmt_7(T x, T y) \
{ \
T max = -1; \
On Mon, Apr 28, 2025 at 2:38 PM Richard Sandiford
wrote:
>
> native_decode_rtx handles integer modes by building up a wide_int
> and then converting it to an rtx. This patch splits out the
> wide_int part, so that callers who don't want an rtx can avoid
> creating garbage rtl.
>
> Bootstrapped &
> On 25 Apr 2025, at 19:55, Richard Sandiford wrote:
>
> Jennifer Schmitz writes:
>> If -msve-vector-bits=128, SVE loads and stores (LD1 and ST1) with a
>> ptrue predicate can be replaced by neon instructions (LDR and STR),
>> thus avoiding the predicate altogether. This also enables formation
For global initializers with IPA PTA we initialize them from the
IPA reference data but that lacks references to the constant pool.
The following conservatively considers the whole initializer.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
Richard.
PR ipa/119973
*
For floating point, !flag_trapping_math is needed for the pattern which
transforms 2 conversions to 1 conversion, and may lose 1 potential trap.
There shouldn't be any accuracy issue.
It also handles real_cst if it can be represented in different floating point
types without loss of precision.
Bo
On Mon, Apr 28, 2025 at 2:56 PM Qing Zhao wrote:
>
>
>
> > On Apr 28, 2025, at 06:49, Richard Biener
> > wrote:
> >
> > On Wed, Apr 23, 2025 at 6:09 PM Qing Zhao wrote:
> >>
> >> Richard,
> >>
> >> Thanks a lot for the hint.
> >>
> >>> On Apr 23, 2025, at 04:17, Richard Biener
> >>> wrote:
>
> -Original Message-
> From: Jennifer Schmitz
> Sent: Monday, April 28, 2025 11:40 AM
> To: gcc-patches@gcc.gnu.org
> Cc: Richard Sandiford ; Tamar Christina
>
> Subject: Re: [PATCH] aarch64: Optimize SVE extract last to Neon lane extract
> for
> 128-bit VLS.
>
>
>
> > On 27 Apr 202
> On Apr 28, 2025, at 09:40, Richard Biener wrote:
>
> On Mon, Apr 28, 2025 at 2:56 PM Qing Zhao wrote:
>>
>>
>>
>>> On Apr 28, 2025, at 06:49, Richard Biener
>>> wrote:
>>>
>>> On Wed, Apr 23, 2025 at 6:09 PM Qing Zhao wrote:
Richard,
Thanks a lot for the hint.
>>
Tested on x86_64-pc-linux-gnu, does this look OK for trunk/15?
-- >8 --
These std::erase_if overloads were wrongly implemented as hidden
friends, visible only via ADL, so erase_if(x) would work but not
std::erase_if(x).
PR libstdc++/119427
libstdc++-v3/ChangeLog:
* include/std/
From: "hongtao.liu"
> I think the comment is a bit off, it should be "For an empty BB ..." since
> we should not change behavior on whether there are debug stmts or not.
Changed.
For an empty BB with all debug_stmt, it will be ignored by
afdo_set_bb_count, but it can be set with count of single
I have committed this patch to trunk after bootstrap/regression testing
again on trunk.
I'll get to gcc14/15 once I flush the current queue.
Andrew
On 1/23/25 04:39, Richard Biener wrote:
On Wed, Jan 22, 2025 at 12:49 AM Andrew MacLeod wrote:
This patch simply adds an op2_range to operator_
When auditing the code, I noticed that
gimple_range_fold::range_of_range_op was reusing a local variable when
it shouldn't.
It was picking up a relation between op1 and op2, and using that
relation to assist in determining if there was a new relation being
formed bet6ween LHS and op1/op2.
I
When prange was split from irange, the functionality of lhs_op1_relation
() did not get ported. This means when we see:
ptr_2 = ptr_1 + 1
we do not get the relation ptr2 > ptr1 from the statement.
This patch adds that functionality back, and is also good example of how
to add new dispatch
The analyzer was issuing false warnings about uninitialized variables
in C++ in places where NRVO was marking DECL_RESULT with
DECL_BY_REFERENCE.
Fixed thusly.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-266-ga1922f0252b3b0.
gcc/analyzer/ChangeLog:
Implement ana::translation_unit for the C++ frontend with a
no-op placeholder implementation, for now.
No functional change intended; a follow-up may implement
things further.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-263-g13f4b12393fa98.
gcc/cp/Change
This patch adds a relation iterator to query the oracle to list either
all the relations on exit to a block, or just ones involving a specified
SSA_NAME.
The oracle then uses this iterator internally for printing as proof of
concept. It is also of use in a follow on patch for finding "nearest
Last fall, infer range processing was adjusted to allow a query to be
specified for something Jakub was working on. During VRP folding there
was an oversight, and ranger was not providing a query. This results
in contextual ranges being missed.
This patch corrects the oversight and passes
I have committed this patch to trunk after bootstrap/regression testing
again on trunk.
I'll get to gcc14/15 once I flush the current queue.
Andrew
On 4/17/25 06:44, Richard Biener wrote:
On Wed, Apr 16, 2025 at 10:55 PM Andrew MacLeod wrote:
This was a fun one! An actual bug, and it too
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-253-g9c4336cbf05228.
gcc/c-family/ChangeLog:
* name-hint.h (name_hint::name_hint): Use std::unique_ptr for
param.
gcc/c/ChangeLog:
* c-decl.cc: Include "make-unique.h".
(lookup_na
No functional change intended.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-251-g0c5b98be0806e2.
gcc/ChangeLog:
* diagnostic.h (diagnostic_context::set_abort_on_error): New.
(diagnostic_context::m_abort_on_error): Make private.
(dia
Make it clearer when we're accessing member data of pass_manager by
adding an "m"_ prefix to the fields handled via passes.def macros.
No functional change intended.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-254-g1dc8916bcbb14c.
gcc/ChangeLog:
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-250-ga25818cd302a77.
gcc/ChangeLog:
* diagnostic-format-sarif.cc (sarif_builder::get_opts): New
accessor.
(sarif_builder::get_version): Update for...
(sarif_builder::m_version): Re
Our headers are a major pain to work with: many require certain other
headers to be included in a particular (undocumented) order in order
to be includable.
Simplify includes in the analyzer by renaming analyzer/analyzer.h to
analyzer/common.h and have it include all the common headers needed
thro
Add a new function to help debugging -fanalyzer.
No functional change intended.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-255-g0ef00f71969f20.
gcc/analyzer/ChangeLog:
* call-details.cc (call_details::dump): New overload.
(call_details::
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-252-gbc39b0072f5660.
gcc/ChangeLog:
* diagnostic.h (diagnostic_context::m_opt_permissive): Convert
from int to diagnostic_option_id. Update comment.
Signed-off-by: David Malcolm
---
gcc/diagno
No functional change intended.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-259-gc2f8c7ad2db225.
gcc/analyzer/ChangeLog:
* analyzer.cc: Convert gcall * to gcall & where we know the
pointer must be non-null.
* call-details.cc: Likewi
No functional change intended.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-260-g4e1f545df9da1d.
gcc/analyzer/ChangeLog:
* call-summary.cc (call_summary_replay::call_summary_replay):
Convert "summary" from call_summary * to call_summary &.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-261-g6f9764571d2dd1.
gcc/analyzer/ChangeLog:
* engine.cc (class plugin_analyzer_init_impl): Convert
"m_checkers" to use std::vector of std::unique_ptr. Convert
"m_known_fn_mgr" to a refer
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-256-g5ecea59621c63a.
gcc/analyzer/ChangeLog:
* region-model.cc (region_model::on_stmt_pre): Use internal_error
if we see an unexpected gimple stmt code.
Signed-off-by: David Malcolm
---
gcc/ana
Modernization; no functional change intended.
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-258-g978e9dfc0a1331.
gcc/analyzer/ChangeLog:
* access-diagram.cc: Convert enum access_direction to
"enum class".
* bounds-checking.cc: Likewi
Successfully bootstrapped & regrtested on x86_64-pc-linux-gnu.
Pushed to trunk as r16-265-g2a63dc8c65d469.
gcc/analyzer/ChangeLog:
PR analyzer/109366
* region-model-manager.cc
(region_model_manager::maybe_fold_sub_svalue): Sub-values of zero
constants are zero.
gcc
1 - 100 of 169 matches
Mail list logo
