> -原始邮件-
> 发件人: "Jeff Law"
> 发送时间: 2025-06-03 03:37:10 (星期二)
> 收件人: Jiawei , gcc-patches@gcc.gnu.org
> 抄送: kito.ch...@gmail.com, pal...@rivosinc.com, christoph.muell...@vrull.eu
> 主题: Re: [PATCH] RISC-V: Add Shlcofideleg extension.
>
>
>
> On 5/27/25 1:32 AM, Jiawei wrote:
> > This p
On Tue, Jun 3, 2025 at 10:43 PM H.J. Lu wrote:
>
> On Tue, Jun 3, 2025 at 10:51 AM Andrew Pinski
> wrote:
> >
> > When we have a smallish CSWTCH, it could be placed in the rodata.cst16
> > section so it can be merged with other constants across TUs.
> >
> > The fix is simple; just mark the decl
On Tue, Jun 3, 2025 at 10:51 AM Andrew Pinski wrote:
>
> When we have a smallish CSWTCH, it could be placed in the rodata.cst16
> section so it can be merged with other constants across TUs.
>
> The fix is simple; just mark the decl as mergable (DECL_MERGEABLE).
> DECL_MERGEABLE was added with r14
On s390x, for input:
(call_insn/u 7 6 11 2 (parallel [
(set (reg:SI 2 %r2)
(call (subreg:QI (symbol_ref:SI ("__tls_get_offset")
[flags 0x1]) 3)
(const_int 0 [0])))
(clobber (reg:SI 14 %r14))
(use (unspec:SI [
On Tue, Jun 3, 2025 at 11:36 PM Richard Sandiford
wrote:
>
> "H.J. Lu" writes:
> > On Tue, Jun 3, 2025 at 6:06 AM H.J. Lu wrote:
> >> Always add REG_CALL_DECL note for CALL so that get_call_fndecl works
> >> without -fipa-ra.
> >>
> >> PR other/120494
> >> * calls.cc (expand_call): Always add RE
On Tue, Jun 3, 2025 at 12:44 AM Robin Dapp wrote:
>
> > 1. riscv64-linux-gcc -march=rv64gc -march=foo-cpu -mtune=foo-cpu
> > 2. riscv64-linux-gcc -march=rv64gc -march=foo-cpu
> > 3. riscv64-linux-gcc -march=rv64gc -march=unset -mtune=unset -mcpu=foo-cpu
> >
> > Preference to me:
> > - Prefer optio
LGTM
於 2025年6月4日 週三，11:23寫道：
> From: Pan Li
>
> Some similar code could be wrapped to func get_vector_binary_rtx_cost,
> thus leverage this function to avoid code duplication.
>
> The below test suites are passed for this patch series.
> * The rv64gcv fully regression test.
>
> gcc/ChangeLog:
>
From: Pan Li
Some similar code could be wrapped to func get_vector_binary_rtx_cost,
thus leverage this function to avoid code duplication.
The below test suites are passed for this patch series.
* The rv64gcv fully regression test.
gcc/ChangeLog:
* config/riscv/riscv.cc (get_vector_bin
On Thu, May 29, 2025 at 4:56 PM Hu, Lin1 wrote:
>
> Hi,
>
> The patch aims to optimize
> movb(%rdi), %al
> movq%rdi, %rbx
> xorl%esi, %eax, %edx
> movb%dl, (%rdi)
> cmpb%sil, %al
> jne
> to
> xorb%sil, (%rdi)
>
On Mon, May 26, 2025 at 4:55 PM Hu, Lin1 wrote:
>
> Hi, all
>
> Enable -mapxf will change some patterns about adc/sbb.
>
> Hence gcc will raise an extra mov like
> movq8(%rdi), %rax
> adcq%rax, 8(%rsi), %rax
> movq%rax, 8(%rdi)
> rather than
> movq
On 5/21/25 11:07, Jeff Law wrote:
> On 5/20/25 4:05 PM, Edwin Lu wrote:
>> The instruction scheduler appears to be speculatively hoisting vsetvl
>> insns outside of their basic block without checking for data
>> dependencies. This resulted in a situation where the following occurs
>>
>> vs
On Sat, 2025-05-31 at 23:25 -0400, Jason Merrill wrote:
> From f57505c7c6fa05a14e285c14a81021655a43ccbc Mon Sep 17 00:00:00
> 2001
> From: Jason Merrill
> Date: Wed, 20 Nov 2024 16:20:52 +0100
> Subject: [PATCH] c++: modules and #pragma diagnostic
> To: gcc-patches@gcc.gnu.org
>
> To respect the
On 6/3/25 4:11 PM, Joseph Myers wrote:
On Mon, 2 Jun 2025, Jeff Law wrote:
On 6/1/25 1:12 AM, H.J. Lu wrote:
Move get_call_rtx_from to final.c and call call_from_call_insn.
PR other/120493
* final.cc (call_from_call_insn): Change the argument type to
const rtx_call_insn *.
(get_call_rtx_fr
On Mon, 2 Jun 2025, Jeff Law wrote:
> On 6/1/25 1:12 AM, H.J. Lu wrote:
> > Move get_call_rtx_from to final.c and call call_from_call_insn.
> >
> > PR other/120493
> > * final.cc (call_from_call_insn): Change the argument type to
> > const rtx_call_insn *.
> > (get_call_rtx_from): New.
> > * rtl.
On 5/29/25 2:30 PM, Iain Sandoe wrote:
Tested on x86_64-darwin, powerpc64le-linux, OK for trunk?
OK.
thanks
Iain
--- 8< ---
This continues code cleanups and migration to encapsulation of the
whole coroutine transform.
gcc/cp/ChangeLog:
* coroutines.cc (analyze_fn_parms): Move from
Hi Harald,
I don't think it is necessarily the save attribute, but rather it's
representation in "assembler". To my understanding the type's size is
needed to allocate space in the .RSS or .data segment of the binary. To
manage this the size needs to be compile time computable, which it is not
On 6/2/25 5:44 PM, Andrew Pinski wrote:
From: Gwenole Beauchesne
Handle '#pragma GCC optimize' earlier as the __OPTIMIZE__ macro may need
to be defined as well for certain usages. Add additional tests for the
'#pragma GCC target' case with auto-vectorization enabled and multiple
combinations of
Hi Andre,
On 6/3/25 13:31, Andre Vehreschild wrote:
Hi all,
thanks for the explanations, Christophe. This is very much appreciated. And
sorry, I can't follow all presentations, conferences and publications. There is
meanwhile way too much for me to process out there.
Anyway, the regression I p
Hi Harald,
Lgtm. That patch is nearly obvious. Ok for trunk and backport.
Thanks for the patch,
Andre
Andre Vehreschild * ve...@gmx.de
Am 3. Juni 2025 20:43:52 schrieb Harald Anlauf :
Dear all,
here's a fix for another one of Gerhard's "torture tests" that triggers
an ICE with -Warray-tempor
Hi Andre,
On 6/3/25 21:03, Andre Vehreschild wrote:
Hi Harald,
Lgtm. That patch is nearly obvious. Ok for trunk and backport.
Thanks for the patch,
Andre
thanks for the swift review!
Pushed as r16-1090-g0768ec0d32f570.
Thanks,
Harald
Andre Vehreschild * ve...@gmx.de
Am 3. Juni 2025 20:43
On Tue, 3 Jun 2025, Andrew Pinski wrote:
> This is just a simple check to see if the flags like LangEnabledBy
> have the correct case. By putting everything into upper case and
> seeing if there is a match (if previously there was not a match).
> This would have caught PR 120078 much earlier.
> Te
Dear all,
here's a fix for another one of Gerhard's "torture tests" that triggers
an ICE with -Warray-temporaries. The cause was that the locus was not
set for the initializer of the symbol for the testcase. The obvious
fix is to then use the locus from the array constructor generated for
the d
On 6/3/25 10:11, Richard Sandiford wrote:
> Vineet Gupta writes:
>> On 6/3/25 08:24, Richard Sandiford wrote:
>>> I think the issue is that:
>>>
>>> (insn 9 8 27 2 (parallel [
>>> (asm_operands/v ("fsrm %0") ("") 0 [
>>> (reg:SI 15 a5 [139])
>>> ]
This is just a simple check to see if the flags like LangEnabledBy
have the correct case. By putting everything into upper case and
seeing if there is a match (if previously there was not a match).
This would have caught PR 120078 much earlier.
Tested both with and without the fix for PR 120078.
g
Am Dienstag, dem 03.06.2025 um 11:22 -0700 schrieb Andrew Pinski:
> On Tue, Jun 3, 2025 at 11:18 AM Martin Uecker wrote:
> >
> > Am Dienstag, dem 03.06.2025 um 10:56 -0700 schrieb Andrew Pinski:
> > > On Tue, Jun 3, 2025 at 10:45 AM Martin Uecker wrote:
> > > >
> > > >
> > > > This version onl
On Tue, Jun 3, 2025 at 11:18 AM Martin Uecker wrote:
>
> Am Dienstag, dem 03.06.2025 um 10:56 -0700 schrieb Andrew Pinski:
> > On Tue, Jun 3, 2025 at 10:45 AM Martin Uecker wrote:
> > >
> > >
> > > This version only contains the fix for -Wc++-compat.
> >
> > Note this was recorded as PR 120078. I
Am Dienstag, dem 03.06.2025 um 10:56 -0700 schrieb Andrew Pinski:
> On Tue, Jun 3, 2025 at 10:45 AM Martin Uecker wrote:
> >
> >
> > This version only contains the fix for -Wc++-compat.
>
> Note this was recorded as PR 120078. I didn't notice the typo when I
> filed the bug report; otherwise I
On Tue, Jun 3, 2025 at 11:06 AM Joseph Myers wrote:
>
> On Tue, 3 Jun 2025, Martin Uecker wrote:
>
> > This version only contains the fix for -Wc++-compat.
> >
> > Bootstrapped and regression tested for x86_64.
> >
> > Martin
> >
> >
> > c: Enable -Wjump-misses-init for -Wc++-compat
> >
> >
On Tue, 3 Jun 2025, Martin Uecker wrote:
> This version only contains the fix for -Wc++-compat.
>
> Bootstrapped and regression tested for x86_64.
>
> Martin
>
>
> c: Enable -Wjump-misses-init for -Wc++-compat
>
> Fix a typo that prevented the warning from being activated with
>
On Wed, Jun 04, 2025 at 01:28:39AM +0800, yxj-github-437 wrote:
> Here an unexpanded parameter pack pass into asm_operand which doesn't
> expect to see an operand without type. So use check_for_bare_parameter_packs
> to remedy that.
>
> gcc/cp/ChangeLog:
>
> * parse.cc (cp_parser_asm_operan
On Tue, Jun 3, 2025 at 10:45 AM Martin Uecker wrote:
>
>
> This version only contains the fix for -Wc++-compat.
Note this was recorded as PR 120078. I didn't notice the typo when I
filed the bug report; otherwise I would have fixed it.
Note I think the typo fix is obvious and should be backported
On Tue, 3 Jun 2025, 16:07 Tomasz Kaminski, wrote:
>
>
> On Tue, Jun 3, 2025 at 4:40 PM Jonathan Wakely wrote:
>
>> On Tue, 3 Jun 2025 at 14:46, Jonathan Wakely wrote:
>> >
>> > For some 32-bit targets Glibc supports changing the size of time_t to be
>> > 64 bits by defining _TIME_BITS=64. That
This version only contains the fix for -Wc++-compat.
Bootstrapped and regression tested for x86_64.
Martin
c: Enable -Wjump-misses-init for -Wc++-compat
Fix a typo that prevented the warning from being activated with -Wc++compat.
gcc/c-family/ChangeLog:
* c.o
Here an unexpanded parameter pack pass into asm_operand which doesn't
expect to see an operand without type. So use check_for_bare_parameter_packs
to remedy that.
gcc/cp/ChangeLog:
* parse.cc (cp_parser_asm_operand_list): Check for unexpanded
parameter packs.
gcc/testsuite/Change
Vineet Gupta writes:
> On 6/3/25 08:24, Richard Sandiford wrote:
>> I think the issue is that:
>>
>> (insn 9 8 27 2 (parallel [
>> (asm_operands/v ("fsrm %0") ("") 0 [
>> (reg:SI 15 a5 [139])
>> ]
>> [
>> (asm_inp
Today I found a file called `gcc/cp/cxxapi-data.csv` that contains
# unimplemented ,extents,1,no
# unimplemented ,dextents,1,no
# unimplemented ,layout_left,1,no
# unimplemented ,layout_right,1,no
# unimplemented ,layout_stride,1,no
# unimplemented ,default_accessor,1,no
# unimplemented ,mdspan,1
Thank you Tomasz for the round of review. IIUC we have several
cases of if there's a v6, please change this, and no mandatory
changes.
Would you like me to submit a v6?
On 5/30/25 18:42, Luc Grosheintz wrote:
The discussion for v4 is here:
https://gcc.gnu.org/pipermail/libstdc++/2025-May/061665
> On 3 Jun 2025, at 17:03, Jason Merrill wrote:
>
> On 5/29/25 8:29 AM, Iain Sandoe wrote:
>> To trigger this involves somewhat tortuous pathways through the
>> c++ requires code. I did consider the alternative of putting in
>> an assert and then checking every call-site, but that seemed to
>
On Tue, 3 Jun 2025 at 09:46, Jonathan Wakely wrote:
>
> This adds a new implementation of std::counting_semaphore for the
> case where Max == 1, i.e. the std::binary_semaphore typedef. When the
> maximum counter value is 1 we don't need to load the current counter
> value before doing a compare-exc
On 6/3/25 08:24, Richard Sandiford wrote:
> I think the issue is that:
>
> (insn 9 8 27 2 (parallel [
> (asm_operands/v ("fsrm %0") ("") 0 [
> (reg:SI 15 a5 [139])
> ]
> [
> (asm_input:SI ("r") frm-run-1.c:33)
>
Hi Jason,
>>a bootstrap + testsuite (but without Ada or D).
>Hmm, it looks like make_typename_type wants to call lookup_template_class with
>a class as CONTEXT.
> But it first does the lookup separately.
So I read that to indicate this does not need addressing (at this point) - we
could exten
Tested x86_64-pc-linux-gnu, applying to trunk.
-- 8< --
These tests in the modules/ directory specifically want to test
behavior without modules enabled, so let's make that explicit.
gcc/testsuite/ChangeLog:
* g++.dg/modules/cpp-1.C
* g++.dg/modules/cpp-3.C
* g++.dg/modu
> On 3 Jun 2025, at 17:56, Richard Sandiford wrote:
>
> Tamar Christina writes:
>> As requested in my patch for -mmax-vectorization this promotes the parameter
>> --param aarch64-autovec-preference to a first class top target flag.
>>
>> If both the parameter and the flag is specified the par
"H.J. Lu" writes:
> On Tue, Jun 3, 2025 at 6:06 AM H.J. Lu wrote:
>> Always add REG_CALL_DECL note for CALL so that get_call_fndecl works
>> without -fipa-ra.
>>
>> PR other/120494
>> * calls.cc (expand_call): Always add REG_CALL_DECL note.
>> (emit_library_call_value_1): Likewise.
>>
>> Thanks.
On 5/29/25 8:29 AM, Iain Sandoe wrote:
To trigger this involves somewhat tortuous pathways through the
c++ requires code. I did consider the alternative of putting in
an assert and then checking every call-site, but that seemed to
be a much larger change.
tested on x86_64-darwin and powerpc64le-
Dimitar Dimitrov writes:
> On Thu, May 29, 2025 at 05:59:44PM +0100, Richard Sandiford wrote:
>> Sorry for the slow reply.
>>
>> Dimitar Dimitrov writes:
>> > On Fri, May 16, 2025 at 06:14:30PM +0100, Richard Sandiford wrote:
>> >> Dimitar Dimitrov writes:
> ...
>> >> It might still be worth tr
On 5/29/25 2:31 PM, Iain Sandoe wrote:
Tested on x86_64-darwin, powerpc64le-linux, OK for trunk?
thanks
Iain
--- 8< ---
At present, we can issue diagnostics about missing or malformed
awaiter or promise methods when we encounter their uses in the
body of a users function. We might then re-issu
Tamar Christina writes:
> As requested in my patch for -mmax-vectorization this promotes the parameter
> --param aarch64-autovec-preference to a first class top target flag.
>
> If both the parameter and the flag is specified the parameter takes precedence
> with the reasoning that it may already
On Tue, Jun 3, 2025 at 3:07 AM Tamar Christina wrote:
>
> Hi All,
>
> With the middle-end providing a way to make vectorization more profitable by
> scaling vect-scalar-cost-multiplier this makes a more user friendly option
> to make it easier to use.
>
> I propose making it an actual -m option th
Vineet Gupta writes:
>> ...is from late-combine2, so after RA has completed, whereas the earlier
>> dump is from mode switching, so it's hard to tell what late-combine2 is
>> operating on. Could you give the RTL as late-combine2 sees it?
>> (That would normally be the result of pass_postreload_cs
> -Original Message-
> From: Richard Biener
> Sent: Tuesday, June 3, 2025 2:12 PM
> To: Tamar Christina
> Cc: Richard Biener ; Richard Sandiford
> ; Pengfei Li ; gcc-
> patc...@gcc.gnu.org; ktkac...@nvidia.com
> Subject: Re: [PATCH] vect: Improve vectorization for small-trip-count loops
On 5/29/25 11:16 AM, Iain Sandoe wrote:
Tested on x86_64-darwin, powerpc64le-linux, OK for trunk?
OK.
thanks
Iain
--- 8< ---
We do not need to generate this code early, since it does not affect
any of the analysis. Lowering it later takes less code, and avoids
modifying the initial await e
On 6/2/25 4:54 AM, Jakub Jelinek wrote:
Hi!
The following patch implements the C++26 P3068R5 - constexpr exceptions
paper.
As the IL cxx_eval_constant* functions see already contains the low
level calls like __cxa_{allocate,free}_exception, __cxa_{,re}throw etc.,
the patch just makes 10 extern
On Tue, Jun 3, 2025 at 4:40 PM Jonathan Wakely wrote:
> On Tue, 3 Jun 2025 at 14:46, Jonathan Wakely wrote:
> >
> > For some 32-bit targets Glibc supports changing the size of time_t to be
> > 64 bits by defining _TIME_BITS=64. That causes an ABI change which
> > would affect std::chrono::system
In GIMPLE we can always use SSA names, so do that as it is cheaper.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
So now we can remove create_tmp_reg_or_ssa_name.
Richard.
* gimple-fold.cc (create_tmp_reg_or_ssa_name): Always
create a SSA name.
---
gcc/gimple-fol
On 6/3/25 10:02 AM, Iain Sandoe wrote:
On 3 Jun 2025, at 02:36, Jason Merrill wrote:
On 6/2/25 5:27 PM, Iain Sandoe wrote:
On 2 Jun 2025, at 21:17, Jason Merrill wrote:
On 6/1/25 3:30 AM, Iain Sandoe wrote:
Updated. I realised we no longer need to refer to
initial_await_resume_called in
On Tue, 3 Jun 2025 at 14:46, Jonathan Wakely wrote:
>
> For some 32-bit targets Glibc supports changing the size of time_t to be
> 64 bits by defining _TIME_BITS=64. That causes an ABI change which
> would affect std::chrono::system_clock::to_time_t. Because to_time_t is
> not a function template,
On Tue, Jun 3, 2025 at 3:46 PM Jonathan Wakely wrote:
> For some 32-bit targets Glibc supports changing the size of time_t to be
> 64 bits by defining _TIME_BITS=64. That causes an ABI change which
> would affect std::chrono::system_clock::to_time_t. Because to_time_t is
> not a function template
The current implementation of gcov_merge works only when the topn is
present in both gcov_info. Otherwise, the function
topn_to_memory_representation is not called, and the memory
representation remains in a disk-formatted state, which leads to
unpredictable behavior in the write_topn_counters func
On Tue, Jun 03, 2025 at 03:26:40PM +0200, Richard Biener wrote:
> On Tue, Jun 3, 2025 at 3:09 PM Spencer Abson wrote:
> >
> > Floating-point to integer conversions can be inexact or invalid (e.g., due
> > to
> > overflow or NaN). However, since users of operation_could_trap_p infer the
> > bool
Tested x86_64-pc-linux-gnu, applying to trunk.
-- 8< --
There are a lot of uses of | on WMB_Flags, that currently need to then be
cast back to WMB_Flags. Let's avoid the need for that.
gcc/cp/ChangeLog:
* name-lookup.h (operator|, operator|=): Define for WMB_Flags.
---
gcc/cp/name-loo
> On 3 Jun 2025, at 02:36, Jason Merrill wrote:
>
> On 6/2/25 5:27 PM, Iain Sandoe wrote:
>>> On 2 Jun 2025, at 21:17, Jason Merrill wrote:
>>>
>>> On 6/1/25 3:30 AM, Iain Sandoe wrote:
Updated. I realised we no longer need to refer to
initial_await_resume_called in the ramp at all
For some 32-bit targets Glibc supports changing the size of time_t to be
64 bits by defining _TIME_BITS=64. That causes an ABI change which
would affect std::chrono::system_clock::to_time_t. Because to_time_t is
not a function template, its mangled name does not depend on the return
type, so it has
On 6/3/25 4:53 AM, Iain Sandoe wrote:
gcc/cp/ChangeLog:
* pt.cc (lookup_template_class): Honour provided namespace contexts
when looking up class templates.
This is good, but I'd also expect namespace context to take priority over
innermost_non_namespace_value just above?
So
On Tue, Jun 3, 2025 at 3:09 PM Spencer Abson wrote:
>
> Floating-point to integer conversions can be inexact or invalid (e.g., due to
> overflow or NaN). However, since users of operation_could_trap_p infer the
> bool FP_OPERATION argument from the expression's type, the FIX_TRUNC family
> are co
On Fri, May 30, 2025 at 6:48 PM Luc Grosheintz
wrote:
> [mdspan.layout.left.cons] of N4950 states that this ctor is not
> noexcept. Since, all other ctors of layout_left, layout_right or
> layout_stride are noexcept, the choice was made, based on
> [res.on.exception.handling], to make this ctor n
On Fri, May 30, 2025 at 6:44 PM Luc Grosheintz
wrote:
> Implements the tests for layout_stride and for the features of the other
> two layouts that depend on layout_stride.
>
> libstdc++-v3/ChangeLog:
>
> * testsuite/23_containers/mdspan/layouts/class_mandate_neg.cc: Add
> tests f
DR_INIT is already measured in bytes, so there's no need to multiply
the DR_INIT difference of two DRs by the size of one of the DRs when
comparing that difference against MAX_BITSIZE_MODE_ANY_MODE.
Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed.
PR tree-optimization/120517
On Fri, May 9, 2025 at 4:05 PM Tamar Christina wrote:
>
> > -Original Message-
> > From: Richard Biener
> > Sent: Friday, May 9, 2025 2:44 PM
> > To: Tamar Christina
> > Cc: Richard Sandiford ; Pengfei Li
> > ; gcc-patches@gcc.gnu.org; ktkac...@nvidia.com
> > Subject: RE: [PATCH] vect: I
On Tue, Jun 3, 2025 at 2:50 PM Luc Grosheintz
wrote:
>
>
> On 6/3/25 14:31, Tomasz Kaminski wrote:
> > On Mon, Jun 2, 2025 at 9:07 AM Luc Grosheintz
> > wrote:
> >
> >>
> >>
> >> On 5/30/25 18:42, Luc Grosheintz wrote:
> >>> Implements a suite of tests for the currently implemented parts of
> >>
On Sun, Jun 1, 2025 at 6:57 PM Andrew Pinski wrote:
>
> On Sun, Jun 1, 2025 at 3:54 AM Richard Biener
> wrote:
> >
> > On Sat, May 31, 2025 at 8:41 PM Andrew Pinski
> > wrote:
> > >
> > > Currently gimple_folder::convert_and_fold calls create_tmp_var; that
> > > means while in ssa form,
> > >
On Tue, 3 Jun 2025 at 13:08, Tomasz Kaminski wrote:
>
>
>
> On Fri, May 30, 2025 at 6:45 PM Luc Grosheintz
> wrote:
>>
>> libstdc++-v3/ChangeLog:
>>
>> * include/std/mdspan(__mdspan::_ExtentsStorage): Change name
>> of private member _M_dynamic_extens to _M_dyn_exts.
>> *
Floating-point to integer conversions can be inexact or invalid (e.g., due to
overflow or NaN). However, since users of operation_could_trap_p infer the
bool FP_OPERATION argument from the expression's type, the FIX_TRUNC family
are considered non-trapping here.
This patch handles them explicitly
Hi,
This is a V2 of https://gcc.gnu.org/pipermail/gcc-patches/2025-May/683650.html.
Apologies for the delay - I was hesistant to push without preserving
sve/pr96357.c and regtesting x86. The change to sve/pr96357.c means that it
can still be vectorized, and will still trigger the ICE without Prz
On Tue, 3 Jun 2025 at 13:13, Tomasz Kaminski wrote:
>
>
>
> On Fri, May 30, 2025 at 6:47 PM Luc Grosheintz
> wrote:
>>
>> Implements the parts of layout_left that don't depend on any of the
>> other layouts.
>>
>> libstdc++-v3/ChangeLog:
>>
>> * include/std/mdspan (layout_left): New clas
On 6/3/25 14:31, Tomasz Kaminski wrote:
On Mon, Jun 2, 2025 at 9:07 AM Luc Grosheintz
wrote:
On 5/30/25 18:42, Luc Grosheintz wrote:
Implements a suite of tests for the currently implemented parts of
layout_left. The individual tests are templated over the layout type, to
allow reuse as
On Fri, May 30, 2025 at 6:47 PM Luc Grosheintz
wrote:
> Implements the remaining parts of layout_left and layout_right; and all
> of layout_stride.
>
> The implementation of layout_stride::mapping::is_exhaustive applies
> the following change to the standard:
>
> 4266. layout_stride::mapping sh
The __detail::__wait_until function has a comment that should have been
removed when r16-1000-g225622398a9631 changed the return type from a
std::pair to a struct with three members.
Fix _M_spin_until_impl so that the _M_has_val member of the result is
accurate. If the deadline has passed then it
On Fri, May 30, 2025 at 6:50 PM Luc Grosheintz
wrote:
> Adds tests for layout_right and for the parts of layout_left that depend
> on layout_right.
>
> libstdc++-v3/ChangeLog:
>
> * testsuite/23_containers/mdspan/layouts/class_mandate_neg.cc: Add
> tests for layout_right.
>
On Fri, May 30, 2025 at 6:49 PM Luc Grosheintz
wrote:
> Implement the parts of layout_left that depend on layout_right; and the
> parts of layout_right that don't depend on layout_stride.
>
> libstdc++-v3/ChangeLog:
>
> * include/std/mdspan (layout_right): New class.
> * src/c++23
On Mon, Jun 2, 2025 at 9:07 AM Luc Grosheintz
wrote:
>
>
> On 5/30/25 18:42, Luc Grosheintz wrote:
> > Implements a suite of tests for the currently implemented parts of
> > layout_left. The individual tests are templated over the layout type, to
> > allow reuse as more layouts are added.
> >
> >
On Mon, Jun 2, 2025 at 10:56 PM Jonathan Wakely wrote:
> libstdc++-v3/ChangeLog:
>
> * include/std/type_traits (is_destructible, is_destructible_v):
> Define using new built-in.
> (is_nothrow_destructible, is_nothrow_destructible_v): Likewise.
> (is_trivially_destr
On Fri, May 30, 2025 at 6:52 PM Luc Grosheintz
wrote:
> Implements a suite of tests for the currently implemented parts of
> layout_left. The individual tests are templated over the layout type, to
> allow reuse as more layouts are added.
>
> libstdc++-v3/ChangeLog:
>
> * testsuite/23_con
On Fri, May 30, 2025 at 6:47 PM Luc Grosheintz
wrote:
> Implements the parts of layout_left that don't depend on any of the
> other layouts.
>
> libstdc++-v3/ChangeLog:
>
> * include/std/mdspan (layout_left): New class.
> * src/c++23/std.cc.in: Add layout_left.
>
> Signed-off-by:
On Fri, May 30, 2025 at 6:45 PM Luc Grosheintz
wrote:
> libstdc++-v3/ChangeLog:
>
> * include/std/mdspan(__mdspan::_ExtentsStorage): Change name
> of private member _M_dynamic_extens to _M_dyn_exts.
> * include/std/mdspan(extents): Change name of private member
> f
Thanks, Alfie. I agree that having a table with just one entry looks a
little odd, but the rest of the file follows this pattern. For example:
;; -
;; [FP] Absolute difference
;;
On Tue, 3 Jun 2025 at 11:04, Tomasz Kaminski wrote:
>
>
>
> On Tue, Jun 3, 2025 at 12:02 PM Jonathan Wakely wrote:
>>
>> On Tue, 3 Jun 2025 at 09:38, Tomasz Kaminski wrote:
>> >
>> >
>> >
>> > On Mon, Jun 2, 2025 at 10:59 PM Jonathan Wakely wrote:
>> >>
>> >> These function templates are apparen
Hi Sandra, hello world,
Sandra Loosemore wrote:
On 6/2/25 12:15, Tobias Burnus wrote:
The problem is that 'int'/'int*' became 'omp_interop_rc_t ret_code'
and 'omp_interop_rc_t *ret_code'.
...
I think the patch just confuses readers, as-is. Plus it has a grammar
bug. How about instead of
+
Hi all,
thanks for the explanations, Christophe. This is very much appreciated. And
sorry, I can't follow all presentations, conferences and publications. There is
meanwhile way too much for me to process out there.
Anyway, the regression I produced in gomp should be fixed by the new version of
p
Hi All,
With the middle-end providing a way to make vectorization more profitable by
scaling vect-scalar-cost-multiplier this makes a more user friendly option
to make it easier to use.
I propose making it an actual -m option that we document and retain vs using
the parameter name. In the future
Adding a tests for behavior of the ostream operator and the formatting
with empty chronio-spec for the chrono types. Current coverage is:
* time point, zoned_time and local_time_format in this commit,
* calendar types in r16-1016-g28a17985dd34b7.
libstdc++-v3/ChangeLog:
* testsuite/std/
Gentle ping for https://gcc.gnu.org/pipermail/gcc-patches/2025-March/676875.html
This patch would like to introduce the combine of vec_dup + vdiv.vv into
vdiv.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, 15
in test. There will be two cases for the combine:
The series is OK, thanks.
These function templates are apparently never used or instantiated,
because they don't compile. This fixes them, but they're still unused.
I plan to make use of them in a later commit.
In __atomic_wait_address_until_v, the &__args should have bee changed to
just __args by r16-988-g219bb905a60d95,
Check the appropriate feature test macro instead of checking that
__cplusplus >= 201703L.
libstdc++-v3/ChangeLog:
* include/std/stop_token: Check __glibcxx_jthread instead of
__cplusplus.
---
Tested x86_64-linux. Pushed to trunk.
libstdc++-v3/include/std/stop_token | 7 +++
On Tue, 3 Jun 2025 at 10:54, Tomasz Kaminski wrote:
>
>
>
> On Mon, Jun 2, 2025 at 10:56 PM Jonathan Wakely wrote:
>>
>> libstdc++-v3/ChangeLog:
>>
>> * include/std/type_traits (is_destructible, is_destructible_v):
>> Define using new built-in.
>> (is_nothrow_destructible,
The documentation for outline atomics is missing the entry for
-mno-outline-atomics which this patch adds.
Bootstrapped Regtested on aarch64-none-linux-gnu and no issues.
Ok for master?
Thanks,
Tamar
gcc/ChangeLog:
* doc/extend.texi (outline-atomics): Document the inverse -mno flag.
-
The __detail::__wait_until function has a comment that should have been
removed when r16-1000-g225622398a9631 changed the return type from a
std::pair to a struct with three members.
The __atomic_wait_address_until_v and __atomic_wait_address_for_v
function templates are apparently never used or i
With the change to the title, this looks good to me.
On Tue, Jun 3, 2025 at 12:02 PM Tomasz Kaminski wrote:
> The tile says that you are doing only comment fixes, while there is code
> change
> in __spin_until_impl. Could you please adjust it:
> Fix incorrect returns and comments on atomic timed
On Tue, Jun 3, 2025 at 12:02 PM Jonathan Wakely wrote:
> On Tue, 3 Jun 2025 at 09:38, Tomasz Kaminski wrote:
> >
> >
> >
> > On Mon, Jun 2, 2025 at 10:59 PM Jonathan Wakely
> wrote:
> >>
> >> These function templates are apparently never used or instantiated,
> >> because they don't compile. Thi
1 - 100 of 124 matches
Mail list logo
