On Thu, Jul 3, 2025 at 9:34 PM Andrew Pinski wrote:
>
> These 2 testcases were originally designed for the default -march= of
> x86_64 so if you pass -march=native (on a target with AVX512 enabled),
> they will fail. It fix this, we add `-mno-sse3 -mprefer-vector-width=512`
Did you mean to use -m
On Thu, Jul 3, 2025 at 9:08 PM Andrew Pinski wrote:
>
> This is an obvious fix for this small regression. Basically after
> r15-328-g5726de79e2154a,
> there is a call to tree_expr_nonnegative_warnv_p where the type of the
> expression is now
> error_mark_node. Though there was only a check if th
"H.J. Lu" writes:
> On Thu, Jul 3, 2025 at 11:02 PM Richard Sandiford
> wrote:
>>
>> "H.J. Lu" writes:
>> > Since a backend may ignore user type alignment for arguments passed on
>> > stack, update alignment for arguments passed on stack when copying MEM's
>> > memory attributes.
>> >
>> > gcc/
Konstantinos Eleftheriou writes:
> On Wed, May 7, 2025 at 11:29 AM Richard Sandiford
> wrote:
>> But I thought the code was allowing multiple stores to be forwarded to
>> a single (wider) load. E.g. 4 individual byte stores at address X, X+1,
>> X+2 and X+3 could be forwarded to a 4-byte load at
This is a snapshot of work on P2714 "Bind front and back to NTTP
callables", posted for reference.
Questions:
1. Jonathan asks if __type_forward_like_t does the same job as __like_t
in bits/move.h.
2. Could the "if constexpr" statements be better expressed as requires
clauses via the A=>B == !A||B
On 7/3/25 5:19 PM, Vineet Gupta wrote:
Provide a fallback alternaive register contraint for LRA in the light of
the tightened "Q" constraint. Cures the following ICE ...
| gcc/testsuite/gcc.target/riscv/pr118241-b.cc:31:19: error: unable to generate
reloads for:
| 31 | void m() { a.l(); }
On 7/3/25 5:19 PM, Vineet Gupta wrote:
Spotted this by chance as I saw a similar fixup in comments.
From comments, I think this is needed, but I've not hit any issues due
to this.
gcc/ChangeLog:
* config/riscv/predicates.md (prefetch_operand): mack 5 bits.
Signed-off-by: Vineet Gup
在 2025/7/4 上午11:14, Xi Ruoyao 写道:
On Fri, 2025-07-04 at 09:47 +0800, Lulu Cheng wrote:
在 2025/7/2 下午3:31, Xi Ruoyao 写道:
The register_operand predicate can match subreg, then we'd have a subreg
of subreg and it's invalid. Use lowpart_subreg to avoid the nested
subreg.
gcc/ChangeLog:
在 2025/7/4 上午11:25, Xi Ruoyao 写道:
On Fri, 2025-07-04 at 11:14 +0800, Xi Ruoyao wrote:
On Fri, 2025-07-04 at 09:47 +0800, Lulu Cheng wrote:
在 2025/7/2 下午3:31, Xi Ruoyao 写道:
The register_operand predicate can match subreg, then we'd have a
subreg
of subreg and it's invalid. Use lowpart_subreg
On Fri, 2025-07-04 at 11:14 +0800, Xi Ruoyao wrote:
> On Fri, 2025-07-04 at 09:47 +0800, Lulu Cheng wrote:
> >
> > 在 2025/7/2 下午3:31, Xi Ruoyao 写道:
> > > The register_operand predicate can match subreg, then we'd have a
> > > subreg
> > > of subreg and it's invalid. Use lowpart_subreg to avoid th
On Fri, 2025-07-04 at 09:47 +0800, Lulu Cheng wrote:
>
> 在 2025/7/2 下午3:31, Xi Ruoyao 写道:
> > The register_operand predicate can match subreg, then we'd have a subreg
> > of subreg and it's invalid. Use lowpart_subreg to avoid the nested
> > subreg.
> >
> > gcc/ChangeLog:
> >
> > * config
On Thu, 3 Jul 2025, Pietro Monteiro wrote:
> Use TImode instead of the default DImode. Fixes ICE when building libstc++.
I'll have to look into this. There might be a delay.
Thanks for the patch though!
brgds, H-P
>
> Additionally, this change fixes:
>
> c-c++-common/pr111309-1.c -Wc++-com
In GCC 16 the compiler is smarter and it optimizes away the unneeded
zero-extension during the expand pass. Thus we can no longer match
and_alsl_reversed.
Drop the scan-rtl-dump for and_alsl_reversed and add scan-assembler-not
against bstrpick.d to detect the unneeded zero-extension in case it
re
在 2025/7/2 下午3:31, Xi Ruoyao 写道:
The register_operand predicate can match subreg, then we'd have a subreg
of subreg and it's invalid. Use lowpart_subreg to avoid the nested
subreg.
gcc/ChangeLog:
* config/loongarch/loongarch.md (crc_combine): Avoid nested
subreg.
gcc/tests
From: Pan Li
Add asm dump check test for vec_duplicate + vsadd.vv combine to
vsadd.vx, with the GR2VR cost is 0, 1 and 2.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/vx_vf/vx-4-i16.c: Add asm check.
* gcc.target/riscv/rvv/autovec/vx_vf/vx-4-i32.c: Ditto.
* gc
From: Pan Li
Add asm dump check and run test for vec_duplicate + vsadd.vv
combine to vsadd.vx, with the GR2VR cost is 0, 2 and 15.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/rvv/autovec/vx_vf/vx-1-i16.c: Add asm check.
* gcc.target/riscv/rvv/autovec/vx_vf/vx-1-i32.c: Ditto.
From: Pan Li
This patch would like to combine the vec_duplicate + vsadd.vv to the
vsadd.vx. From example as below code. The related pattern will depend
on the cost of vec_duplicate from GR2VR. Then the late-combine will
take action if the cost of GR2VR is zero, and reject the combination
if th
From: Pan Li
This patch would like to introduce the combine of vec_dup + vsadd.vv
into vsadd.vx on the cost value of GR2VR. The late-combine will take
place if the cost of GR2VR is zero, or reject the combine if non-zero
like 1, 2, 15 in test. There will be two cases for the combine:
Case 0:
Use TImode instead of the default DImode. Fixes ICE when building libstc++.
Additionally, this change fixes:
c-c++-common/pr111309-1.c -Wc++-compat (test for excess errors)
c-c++-common/pr111309-1.c -Wc++-compat execution test
gcc.dg/pr105094.c (test for excess errors)
gcc.dg/torture/pr11648
Spotted this by chance as I saw a similar fixup in comments.
>From comments, I think this is needed, but I've not hit any issues due
to this.
gcc/ChangeLog:
* config/riscv/predicates.md (prefetch_operand): mack 5 bits.
Signed-off-by: Vineet Gupta
---
gcc/config/riscv/predicates.md | 4
Provide a fallback alternaive register contraint for LRA in the light of
the tightened "Q" constraint. Cures the following ICE ...
| gcc/testsuite/gcc.target/riscv/pr118241-b.cc:31:19: error: unable to generate
reloads for:
| 31 | void m() { a.l(); }
| | ^
|(insn 26 25 27
Tested x86_64-pc-linux-gnu, applying to trunk.
-- 8< --
In this testcase there is nothing in the lambda except a static_assert which
mentions a variable from the enclosing scope but does not odr-use it, so we
want prune_lambda_captures to remove its capture. Since the lambda is so
empty, there's
Tested x86_64-pc-linux-gnu, applying to trunk.
-- 8< --
This testcase was crashing from infinite recursion in the diagnostic
machinery, trying to print the lambda signature, which referred to the
__this capture field in the lambda, which wanted to print the lambda again.
But we don't want the si
On Thu, 3 Jul 2025 at 23:14, Nathan Myers wrote:
>
> This is a snapshot of work on P2714 "Bind front and back to NTTP
> callables", posted for reference. Not tested.
>
> libstdc++-v3/ChangeLog:
> PR libstdc++/119744
> * include/bits/version.def: Redefine __cpp_lib_bind_front etc.
>
This is a snapshot of work on P2714 "Bind front and back to NTTP
callables", posted for reference. Not tested.
libstdc++-v3/ChangeLog:
PR libstdc++/119744
* include/bits/version.def: Redefine __cpp_lib_bind_front etc.
* include/bits/version.h: Ditto.
* include/std/f
On Thu, Jul 3, 2025 at 11:02 PM Richard Sandiford
wrote:
>
> "H.J. Lu" writes:
> > Since a backend may ignore user type alignment for arguments passed on
> > stack, update alignment for arguments passed on stack when copying MEM's
> > memory attributes.
> >
> > gcc/
> >
> > PR target/120839
> > *
On Thu, 3 Jul 2025 at 23:14, Nathan Myers wrote:
>
> This is a snapshot of work on P2714 "Bind front and back to NTTP
> callables", posted for reference. Not tested.
>
> libstdc++-v3/ChangeLog:
> PR libstdc++/119744
> * include/bits/version.def: Redefine __cpp_lib_bind_front etc.
>
This is a snapshot of work on P2714 "Bind front and back to NTTP
callables", posted for reference. Not tested.
libstdc++-v3/ChangeLog:
PR libstdc++/119744
* include/bits/version.def: Redefine __cpp_lib_bind_front etc.
* include/bits/version.h: Ditto.
* include/std/f
When we split an argument between the stack and a registers we might end
up with a misaligned access, so use this newly implemented hook to
instead bias the codegen towards the registers rather than the stack.
PR/82106
gcc/ChangeLog:
* config/riscv/riscv.cc (struct riscv_arg_info
We currently handle arguments that are split between the stack and
registers by storing the registers to the stack and then treating the
argument as if it was entirely passed on the stack. Allow targets to
override this behavior and instead treat the argument as if it was
passed entirely in regist
This is really Jim's code, but it's been sitting around in Bugzilla for a while
so I've picked it up. All I really did here is add a target hook and mangle
some comments, but I think I understand enough about what's going on to try and
get things moving forward. So I'm writing up a pretty big cov
Hi!
No idea how this slipped in, I'm terribly sorry.
Strangely nothing in the testsuite has caught this, so I've added
a new test for that.
Bootstrapped/regtested on x86_64-linux and i686-linux, committed to trunk
and for 15.2, 14.4, 13.4 and 12.5.
2025-07-03 Jakub Jelinek
PR c++/120
On Thu, Jul 03, 2025 at 10:12:52PM +0200, Mikael Morin wrote:
> From: Mikael Morin
>
> Regression-tested on aarch64-unknown-linux-gnu.
> OK for master?
>
Yes. Almost looks obvious once someone finds and fixes the issue.
Thanks for the patch.
--
Steve
From: Mikael Morin
Regression-tested on aarch64-unknown-linux-gnu.
OK for master?
-- >8 --
Add the preliminary code produced for the evaluation of the FROM and TO
arguments of the MOVE_ALLOC intrinsic before using their values.
Before this change, the preliminary code was ignored and dropped,
l
These 2 testcases were originally designed for the default -march= of
x86_64 so if you pass -march=native (on a target with AVX512 enabled),
they will fail. It fix this, we add `-mno-sse3 -mprefer-vector-width=512`
to the options to force a specific arch to the testcase.
Tested on a skylake-avx512
On 02/07/25 21:44, Jose E. Marchesi wrote:
One can also pass a precompiled BPF object with the desired
optimization options to the script to check it with the verifier.
Yes, but AFAIK building objects that can be actually loaded in the
kernel and verified requires in practice including kernel h
This is an obvious fix for this small regression. Basically after
r15-328-g5726de79e2154a,
there is a call to tree_expr_nonnegative_warnv_p where the type of the
expression is now
error_mark_node. Though there was only a check if the expression was
error_mark_node.
Bootstrapped and tested on x8
Hello!
I've prepared a patch, which adds all members missing from
std::numeric_limits<> specializations for integer-class types.
Jonathan, please let me know whether you like these changes
and do not see any bugs or issues with them. From my side, I just want to
say that:
- Since all std::num
As mentioned in
https://inbox.sourceware.org/gcc/ea828262-8f8f-4362-9ca8-312f7c20e...@nvidia.com/T/#m6e7e8e11656189598c759157d5d49cbd0ac9ba7c.
Adding myself as an aarch64 port reviewer.
ChangeLog:
* MAINTAINERS: Add myself as an aarch64 port reviewer.
Signed-off-by: Andrew Pinski
---
On Thu, Jul 03, 2025 at 08:12:22PM +0200, Richard Biener wrote:
> > So, instead at least for now, the following patch keeps doing the
> > optimization, just doesn't perform it in pointer arithmetics.
> > pointer_int_sum itself actually adds the multiplication by size_exp,
> > so ptr + expr is turne
> Am 03.07.2025 um 16:11 schrieb Jakub Jelinek :
>
> Hi!
>
> The following testcase is miscompiled with -fsanitize=undefined but we
> introduce UB into the IL even without that flag.
>
> The optimization ptr +- (expr +- cst) when expr/cst have undefined
> overflow into (ptr +- cst) +- expr i
Tested on x86-64/Linux, applied on the mainline.
2025-07-03 Eric Botcazou
* gcc-interface/Makefile.in (gnatlib-sjlj): Delete.
(gnatlib-zcx): Do not modify Frontend_Exceptions constant.
* libgnat/system-linux-loongarch.ads (Frontend_Exceptions): Delete.
--
Eric Botcaz
On Tue, May 20, 2025 at 6:44 PM Andrew Pinski wrote:
>
> This was introduced by r15-1797-gd8fe4f05ef448e . I had missed that
> the LHS of the cabs call could be NULL. This seems to only happen at -O0,
> I tried to produce one that happens at -O1 but needed many different
> options to prevent the r
This is primarily Shreya's work from a few months back. I just fixed
the formatting, cobbled together the cover letter/ChangeLog.
We were looking to evaluate some changes from Artemiy that improve GCC's
ability to discover fusible instruction pairs. There was no good way to
get any static d
On 7/2/25 00:42, Richard Biener wrote:
> On Tue, Jun 10, 2025 at 11:40 PM David Faust wrote:
>>
>> Add two new c-family attributes, "btf_type_tag" and "btf_decl_tag"
>> along with a simple shared handler for them.
>>
>> gcc/c-family/
>> * c-attribs.cc (c_common_attribute_table): Add btf
On Thu, Jul 03, 2025 at 02:43:43PM +0200, Michael Matz wrote:
> Hello,
>
> On Thu, 3 Jul 2025, Yuao Ma wrote:
>
> > This patch adds the required function for Fortran trigonometric functions to
> > work with glibc versions prior to 2.26. It's based on glibc source commit
> > 632d895f3e5d98162f77b9
On Thu, Jul 3, 2025 at 12:01 PM H.J. Lu wrote:
>
> When profiling is enabled with shrink wrapping, the mcount call may not
> be placed at the function entry after
>
> pushq %rbp
> movq %rsp,%rbp
>
> As the result, the profile data may be skewed which makes PGO less
> effective.
>
> Add --enable-x8
Hi all,
On 7/3/2025 9:21 PM, Jakub Jelinek wrote:
On Thu, Jul 03, 2025 at 02:43:43PM +0200, Michael Matz wrote:
Hello,
On Thu, 3 Jul 2025, Yuao Ma wrote:
This patch adds the required function for Fortran trigonometric functions to
work with glibc versions prior to 2.26. It's based on glibc s
Another update on this:
> On Jun 30, 2025, at 11:51, Qing Zhao wrote:
>>
>>> For each single predecessor block, locate the conditional statement
>>> in the end of the block. determine whether the STMT is on the taken
>>> path of the condition. Add these two information to each event of
>>>
Alfie Richards writes:
> +/* string_slice inherits from array_slice, specifically to refer to a
> substring
> + of a character array.
> + It includes some string like helpers. */
> +class string_slice : public array_slice
> +{
> +public:
> + string_slice () : array_slice () {}
> + string_s
On 7/3/25 4:24 PM, Juergen Christ wrote:
On 6/27/25 8:09 PM, Juergen Christ wrote:
s390 missed constant vector permutation cases based on the vector pack
instruction or changing the size of the vector elements during vector
merge. This enables some more patterns that do not need to load a
con
"H.J. Lu" writes:
> Since a backend may ignore user type alignment for arguments passed on
> stack, update alignment for arguments passed on stack when copying MEM's
> memory attributes.
>
> gcc/
>
> PR target/120839
> * emit-rtl.cc (set_mem_attrs): Update alignment for argument on
> stack.
>
> gc
On Thu, 3 Jul 2025, Jason Merrill wrote:
> On 7/2/25 7:58 PM, Patrick Palka wrote:
> > Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
> > for trunk?
> >
> > -- >8 --
> >
> > Here the flag -fno-delete-null-pointer-checks causes the trivial address
> > comparison in
> >
> >
Generalize existing scalar gimple_fold rules to apply the same
bitwise comparison simplifications to vector types. Previously, an
expression like
(x < y) && (x > y)
would fold to `false` if x and y are scalars, but equivalent vector
comparisons were left untouched. T
New in V4:
Check whether the vector is of boolean type in specific comparisons.
If it is, determine whether the operation can be expanded using the selected
expression. If so, proceed with the optimization; otherwise, skip the
optimization.
--
Merge simple_comparison patterns under a single vec_cond_expr for bit_and,
bit_ior, and bit_xor in the simplify pass.
Ensure that when both operands of a bit_and, bit_or, or bit_xor are
simple_comparison
results, they reside within the same vec_cond_expr rather than separate ones.
This prepares t
Richard Biener writes:
> The following fixes bad alignment computaton for epilog vectorization
> when as in this case for 510.parest_r and masked epilog vectorization
> with AVX512 we end up choosing AVX to vectorize the main loop and
> masked AVX512 (sic!) to vectorize the epilog. In that case a
> On 6/27/25 8:09 PM, Juergen Christ wrote:
> > s390 missed constant vector permutation cases based on the vector pack
> > instruction or changing the size of the vector elements during vector
> > merge. This enables some more patterns that do not need to load a
> > constant vector for permutation
On Thu, 3 Jul 2025 at 15:19, Patrick Palka wrote:
>
> Tested on x86_64-pc-linux-gnu, does this look OK for trunk/15?
yes for both, thanks.
>
> -- >8 --
>
> In r15-4555-gf191c830154565 we proactively implemented the initial
> proposed resolution for LWG 4166 which was later revealed to be
> insuf
Tested on x86_64-pc-linux-gnu, does this look OK for trunk/15?
-- >8 --
In r15-4555-gf191c830154565 we proactively implemented the initial
proposed resolution for LWG 4166 which was later revealed to be
insufficient, since we must also require equality_comparable of the
underlying iterators befor
Thank you for the nice review! I've locally implemented everything and
I'll send a v3 later today or tomorrow; after squashing the commits
correctly; and retesting everything.
Meanwhile a couple of comments below.
On 7/1/25 16:42, Tomasz Kaminski wrote:
On Fri, Jun 27, 2025 at 11:37 AM Luc Gros
Hi!
The following testcase is miscompiled with -fsanitize=undefined but we
introduce UB into the IL even without that flag.
The optimization ptr +- (expr +- cst) when expr/cst have undefined
overflow into (ptr +- cst) +- expr is sometimes simply not valid,
without careful analysis on what ptr poi
On Jul 03 2025, Michael Matz wrote:
> Yes. And then the above is multiplied by PI, passed to cos/sin and that
> one then tries to figure out the multiple of PI (i.e. the 'x' above) again
> via range reduction (not a _terribly_ slow one anymore in a good
> implementation, because of the limited
On Thu, 3 Jul 2025, Michael Matz wrote:
> Yes. And then the above is multiplied by PI, passed to cos/sin and that
> one then tries to figure out the multiple of PI (i.e. the 'x' above) again
> via range reduction (not a _terribly_ slow one anymore in a good
> implementation, because of the lim
The following fixes bad alignment computaton for epilog vectorization
when as in this case for 510.parest_r and masked epilog vectorization
with AVX512 we end up choosing AVX to vectorize the main loop and
masked AVX512 (sic!) to vectorize the epilog. In that case alignment
analysis for the epilog
Hello,
On Thu, 3 Jul 2025, Joseph Myers wrote:
> > > Isn't the whole raison d'etre for the trig-pi functions that the internal
> > > argument reduction against multiples of pi becomes trivial and hence (a)
> > > performant, and (b) doesn't introduce rounding artifacts? Expressing the
> > > tr
On Thu, 3 Jul 2025, Jakub Jelinek wrote:
> > Isn't the whole raison d'etre for the trig-pi functions that the internal
> > argument reduction against multiples of pi becomes trivial and hence (a)
> > performant, and (b) doesn't introduce rounding artifacts? Expressing the
> > trig-pi functions
On 7/2/25 7:58 PM, Patrick Palka wrote:
Bootstrapped and regtested on x86_64-pc-linux-gnu, does this look OK
for trunk?
-- >8 --
Here the flag -fno-delete-null-pointer-checks causes the trivial address
comparison in
inline int a, b;
static_assert(&a != &b);
to be rejected as non-constan
This patch adds the bpf-vmtest-tool subdirectory under contrib which tests
BPF programs under a live kernel using a QEMU VM. It automatically
builds the specified kernel version with eBPF support enabled
and stores it under "~/.bpf-vmtest-tool", which is reused for future
invocations.
It can also
This patch adds initial version of bpf-vmtest-tool script to test BPF programs
on live kernel
For now, the tool is standalone, but it is intended to be integrated with the
DejaGnu testsuite to run BPF testcases in future patches.
Current Limitations:
- Only x86_64 is supported. Support for addit
On Thu, Jul 03, 2025 at 02:43:43PM +0200, Michael Matz wrote:
> Hello,
>
> On Thu, 3 Jul 2025, Yuao Ma wrote:
>
> > This patch adds the required function for Fortran trigonometric functions to
> > work with glibc versions prior to 2.26. It's based on glibc source commit
> > 632d895f3e5d98162f77b9
With this fix-up for commit 387209938d2c476a67966c6ddbdbf817626f24a2
"OpenMP: Add omp_get_initial_device/omp_get_num_devices builtins", we progress:
PASS: c-c++-common/gomp/omp_get_num_devices_initial_device.c (test for
excess errors)
PASS: c-c++-common/gomp/omp_get_num_devices_initial_
On Thu, Jul 03, 2025 at 02:55:37PM +0200, Stefan Schulze Frielinghaus wrote:
> Ok for mainline?
ChangeLog is missing.
And I think I'd appreciate another pair of eyes, Rainer/Mike, what do you
think about this?
> If a check-function-bodies test is compiled using -fstack-protector*,
> -fhardened, -
ping^2 ?
On Wed, 18 Jun 2025 at 12:11, Christophe Lyon
wrote:
>
> ping?
>
> On Mon, 26 May 2025 at 17:26, Christophe Lyon
> wrote:
> >
> > On Mon, 26 May 2025 at 17:14, Christophe Lyon
> > wrote:
> > >
> > > Commit r15-7152-g57b706d141b87c removed
> > > /* { dg-do run { target*-*-linux* *-*-gnu
From: Stefan Schulze Frielinghaus
My understand is that during check_compile compiler_flags contains all
the options passed to gcc and current_compiler_flags contains options
passed via dg-options and dg-additional-options. I did a couple of
experiments and printf-style debugging which endorsed
Hello,
On Thu, 3 Jul 2025, Yuao Ma wrote:
> This patch adds the required function for Fortran trigonometric functions to
> work with glibc versions prior to 2.26. It's based on glibc source commit
> 632d895f3e5d98162f77b9c3c1da4ec19968b671.
>
> I've built it successfully on my end. Documentation
gcc.dg/ipa/pr120295.c FAILs on Solaris:
FAIL: gcc.dg/ipa/pr120295.c (test for excess errors)
Excess errors:
ld: warning: symbol 'glob' has differing types:
(file /var/tmp//ccsDR59c.o type=OBJT; file /lib/libc.so type=FUNC);
/var/tmp//ccsDR59c.o definition taken
Fixed by renaming
On 6/27/25 8:09 PM, Juergen Christ wrote:
s390 missed constant vector permutation cases based on the vector pack
instruction or changing the size of the vector elements during vector
merge. This enables some more patterns that do not need to load a
constant vector for permutation.
Bootstrapped
Karl Meakin writes:
> This patch series adds support for the CMPBR extension. It includes the
> new `+cmpbr` option and rules to generate the new instructions when
> lowering conditional branches.
Thanks for the update, LGTM. I've pushed the series to trunk.
Richard
> Changelog:
> * v9:
> -
On 7/3/25 12:45, Jonathan Wakely wrote:
On Thu, 3 Jul 2025 at 11:12, Tomasz Kaminski wrote:
On Thu, Jul 3, 2025 at 12:08 PM Luc Grosheintz wrote:
The reasoning for this approach was:
1. The mapping::operator() and mdspan::operator[] have the same
precondition; and mdspan::operator[
On 2025-07-03 03:13, Jakub Jelinek wrote:
On Thu, Jul 03, 2025 at 08:33:45AM +0200, Richard Biener wrote:
On Wed, Jul 2, 2025 at 11:32 PM Siddhesh Poyarekar wrote:
MEM_REF cast of a subobject to its containing object has negative
offsets, which objsz sees as an invalid access. Support this u
Those methods's implementation is relying on duck-typing at compile
time.
The structure corresponding to the node of a doubly linked list needs
to define attributes 'prev' and 'next' which are pointers on the type
of a node.
The structure wrapping the nodes and others metadata (first, last, size)
n
This patch was originally part of [1]. Merging it in GCC is a prerequisite of
merging it inside binutils.
Those methods's implementation is relying on duck-typing at compile time. The
structure corresponding to the node of a doubly linked list needs to define
attributes 'prev' and 'next' which
On Thu, 3 Jul 2025 at 11:12, Tomasz Kaminski wrote:
>
>
>
> On Thu, Jul 3, 2025 at 12:08 PM Luc Grosheintz
> wrote:
>>
>>
>>
>> On 7/1/25 22:56, Jonathan Wakely wrote:
>> > On Tue, 1 Jul 2025 at 11:32, Tomasz Kaminski wrote:
>> >>
>> >> Hi,
>> >> More of the review will be later, but I have not
This commit completes the implementation of P2897R7 by implementing and
testing the template class aligned_accessor.
libstdc++-v3/ChangeLog:
* include/bits/version.def (aligned_accessor): Add.
* include/bits/version.h: Regenerate.
* include/std/mdspan (aligned_accessor): N
This commit implements and tests the function is_sufficiently_aligned
from P2897R7.
libstdc++-v3/ChangeLog:
* include/bits/align.h (is_sufficiently_aligned): New function.
* include/bits/version.def (is_sufficiently_aligned): Add.
* include/bits/version.h: Regenerate.
This patch series implements the aligned_accessor paper P2897R7 in three
parts:
- Implement `is_sufficiently_aligned` which is part of .
- Prepare the accessor tests for reuse.
- Implement aligned_accessor.
A couple of remarks:
- The paper P2897R7 and spec N5008 don't specify that the al
All test code of default_accessor can be reused. This commit moves
the reuseable code into a file generic.cc and prepares the tests for
reuse with aligned_accessor.
The AllocatorTrait creates a unified interface for creating both
default_accessor and aligned_accessor typenames.
libstdc++-v3/Chang
On Thu, Jul 3, 2025 at 6:07 PM Uros Bizjak wrote:
>
> On Thu, Jul 3, 2025 at 11:54 AM H.J. Lu wrote:
> >
> > commit ecc81e33123d7ac9c11742161e128858d844b99d (HEAD)
> > Author: Andi Kleen
> > Date: Fri Sep 26 04:06:40 2014 +
> >
> > Add direct support for Linux kernel __fentry__ patchin
For TLS calls:
1. UNSPEC_TLS_GD:
(parallel [
(set (reg:DI 0 ax)
(call:DI (mem:QI (symbol_ref:DI ("__tls_get_addr")))
(const_int 0 [0])))
(unspec:DI [(symbol_ref:DI ("e") [flags 0x50])
(reg/f:DI 7 sp)] UNSPEC_TLS_GD)
(clobber (reg:DI 5 di))]
On Thu, Jul 3, 2025 at 12:08 PM Luc Grosheintz
wrote:
>
>
> On 7/1/25 22:56, Jonathan Wakely wrote:
> > On Tue, 1 Jul 2025 at 11:32, Tomasz Kaminski
> wrote:
> >>
> >> Hi,
> >> More of the review will be later, but I have noticed that you have
> added preconditions checks
> >> to the layouts, an
On 7/1/25 22:56, Jonathan Wakely wrote:
On Tue, 1 Jul 2025 at 11:32, Tomasz Kaminski wrote:
Hi,
More of the review will be later, but I have noticed that you have added
preconditions checks
to the layouts, and then avoid checking them inside the operator[] of the
mdspan. This is general
s
On Thu, Jul 3, 2025 at 11:54 AM H.J. Lu wrote:
>
> commit ecc81e33123d7ac9c11742161e128858d844b99d (HEAD)
> Author: Andi Kleen
> Date: Fri Sep 26 04:06:40 2014 +
>
> Add direct support for Linux kernel __fentry__ patching
>
> emitted a label, 1, for __mcount_loc section:
>
> 1: call mco
Hi,
ipa-cp converts sreal times to int, while point of sreal is to accomodate very
large values that can happen for loops with large number of iteraitons and also
when profile is inconsistent. This happens with afdo in testsuite where loop
preheader is estimated to have 0 excutions while loop body
Hi,
ipa-cp cloning disables itself for all functions not passing opt_for_fn
(node->decl, optimize_size) which disables it for cold wrappers of hot
functions where we want to propagate. Since we later want to time saved
to be considered hot, we do not need to make this early test.
The patch also f
When profiling is enabled with shrink wrapping, the mcount call may not
be placed at the function entry after
pushq %rbp
movq %rsp,%rbp
As the result, the profile data may be skewed which makes PGO less
effective.
Add --enable-x86-64-mfentry to enable -mfentry by default to use
__fentry__, added
On 02/07/2025 18:45, Karl Meakin wrote:
This patch series adds support for the CMPBR extension. It includes the
new `+cmpbr` option and rules to generate the new instructions when
lowering conditional branches.
Changelog:
* v8:
- Support far branches for the `CBB` and `CBH` instructions, an
commit ecc81e33123d7ac9c11742161e128858d844b99d (HEAD)
Author: Andi Kleen
Date: Fri Sep 26 04:06:40 2014 +
Add direct support for Linux kernel __fentry__ patching
emitted a label, 1, for __mcount_loc section:
1: call mcount
.section __mcount_loc, "a",@progbits
.quad 1b
.previous
If _
Hi Jerry,
thanks for the review and the ok. Committed as gcc-16-1967-g15413e05eb9.
And special thanks for the kind words in the private mail you send me. It's
very much appreciated that you even applied a translator to translate it to
German. Thank you very much.
I have set myself a reminder to
A typo in r15-4473-g3abe751ea86e34 made it ill-formed to use
std::uninitialized_fill with iterators that aren't pointers (or pointers
wrapped in our __normal_iterator) if the value type is a narrow
character type.
libstdc++-v3/ChangeLog:
PR libstdc++/120931
* include/bits/stl_unin
1 - 100 of 145 matches
Mail list logo
