public inbox for gcc-bugs@sourceware.org help / color / mirror / Atom feed
* [Bug c++/102177] New: Implement C++17 P0418R2
@ 2021-09-02 14:28 jakub at gcc dot gnu.org
2021-09-02 14:28 ` [Bug c++/102177] " jakub at gcc dot gnu.org
` (4 more replies)
0 siblings, 5 replies; 6+ messages in thread
From: jakub at gcc dot gnu.org @ 2021-09-02 14:28 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102177
Bug ID: 102177
Summary: Implement C++17 P0418R2
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: c++
Assignee: unassigned at gcc dot gnu.org
Reporter: jakub at gcc dot gnu.org
Target Milestone: ---
I think we aren't implementing https://wg21.link/p0418r2
which dropped the requirement that failure mode can't be stronger than success
mode.
For
long
bar (long *x, long y, long z)
{
__atomic_compare_exchange_n (x, &y, z, 0, __ATOMIC_RELAXED,
__ATOMIC_SEQ_CST);
return y;
}
we warn:
warning: failure memory model cannot be stronger than success memory model for
‘__atomic_compare_exchange’ [-Winvalid-memory-model]
and document it in doc/extend.texi:
Otherwise, @code{false} is returned and memory is affected according
to @var{failure_memorder}. This memory order cannot be
@code{__ATOMIC_RELEASE} nor @code{__ATOMIC_ACQ_REL}. It also cannot be a
stronger order than that specified by @var{success_memorder}.
and libstdc++ has:
__glibcxx_assert(__b2 <= __b1);
in various spots in bits/atomic_base.h.
Do we emit right code on aarch64 and other weak ordering targets for this case
despite the warning?
^ permalink raw reply [flat|nested] 6+ messages in thread* [Bug c++/102177] Implement C++17 P0418R2 2021-09-02 14:28 [Bug c++/102177] New: Implement C++17 P0418R2 jakub at gcc dot gnu.org @ 2021-09-02 14:28 ` jakub at gcc dot gnu.org 2021-09-02 14:40 ` jakub at gcc dot gnu.org ` (3 subsequent siblings) 4 siblings, 0 replies; 6+ messages in thread From: jakub at gcc dot gnu.org @ 2021-09-02 14:28 UTC (permalink / raw) To: gcc-bugs https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102177 Jakub Jelinek <jakub at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- Last reconfirmed| |2021-09-02 Ever confirmed|0 |1 CC| |amacleod at redhat dot com, | |redi at gcc dot gnu.org, | |rsandifo at gcc dot gnu.org Status|UNCONFIRMED |NEW ^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug c++/102177] Implement C++17 P0418R2 2021-09-02 14:28 [Bug c++/102177] New: Implement C++17 P0418R2 jakub at gcc dot gnu.org 2021-09-02 14:28 ` [Bug c++/102177] " jakub at gcc dot gnu.org @ 2021-09-02 14:40 ` jakub at gcc dot gnu.org 2021-09-02 15:14 ` jakub at gcc dot gnu.org ` (2 subsequent siblings) 4 siblings, 0 replies; 6+ messages in thread From: jakub at gcc dot gnu.org @ 2021-09-02 14:40 UTC (permalink / raw) To: gcc-bugs https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102177 Jakub Jelinek <jakub at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |dje at gcc dot gnu.org --- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> --- E.g. aarch64 has comment like: /* Normally the succ memory model must be stronger than fail, but in the unlikely event of fail being ACQUIRE and succ being RELEASE we need to promote succ to ACQ_REL so that we don't lose the acquire semantics. */ which makes me think that maybe it needs some work when the stronger requirement isn't honored. What about powerpc? ^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug c++/102177] Implement C++17 P0418R2 2021-09-02 14:28 [Bug c++/102177] New: Implement C++17 P0418R2 jakub at gcc dot gnu.org 2021-09-02 14:28 ` [Bug c++/102177] " jakub at gcc dot gnu.org 2021-09-02 14:40 ` jakub at gcc dot gnu.org @ 2021-09-02 15:14 ` jakub at gcc dot gnu.org 2021-09-02 17:51 ` cvs-commit at gcc dot gnu.org 2022-04-21 15:33 ` cvs-commit at gcc dot gnu.org 4 siblings, 0 replies; 6+ messages in thread From: jakub at gcc dot gnu.org @ 2021-09-02 15:14 UTC (permalink / raw) To: gcc-bugs https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102177 --- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> --- Created attachment 51406 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51406&action=edit gcc12-pr102177.patch Untested patch to remove the warning and just use maximum for success if weaker than failure instead of always seq_cst. On the other side, even current C2X draft has that: The failure argument shall be no stronger than the success argument. requirement. So should we keep the restriction for __atomic_* APIs and just pass maximum from libstdc++ headers? ^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug c++/102177] Implement C++17 P0418R2 2021-09-02 14:28 [Bug c++/102177] New: Implement C++17 P0418R2 jakub at gcc dot gnu.org ` (2 preceding siblings ...) 2021-09-02 15:14 ` jakub at gcc dot gnu.org @ 2021-09-02 17:51 ` cvs-commit at gcc dot gnu.org 2022-04-21 15:33 ` cvs-commit at gcc dot gnu.org 4 siblings, 0 replies; 6+ messages in thread From: cvs-commit at gcc dot gnu.org @ 2021-09-02 17:51 UTC (permalink / raw) To: gcc-bugs https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102177 --- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> --- The master branch has been updated by Jonathan Wakely <redi@gcc.gnu.org>: https://gcc.gnu.org/g:dba1ab212292839572fda60df00965e094a11252 commit r12-3317-gdba1ab212292839572fda60df00965e094a11252 Author: Jonathan Wakely <jwakely@redhat.com> Date: Thu Sep 2 15:29:22 2021 +0100 libstdc++: Remove "no stronger" assertion in compare exchange [PR102177] P0418R2 removed some preconditions from std::atomic::compare_exchange_* but we still enforce them via __glibcxx_assert. This removes those assertions. Signed-off-by: Jonathan Wakely <jwakely@redhat.com> libstdc++-v3/ChangeLog: PR c++/102177 * include/bits/atomic_base.h (__is_valid_cmpexch_failure_order): New function to check if a memory order is valid for the failure case of compare exchange operations. (__atomic_base<I>::compare_exchange_weak): Simplify assertions by using __is_valid_cmpexch_failure_order. (__atomic_base<I>::compare_exchange_strong): Likewise. (__atomic_base<P*>::compare_exchange_weak): Likewise. (__atomic_base<P*>::compare_exchange_strong): Likewise. (__atomic_impl::compare_exchange_weak): Add assertion. (__atomic_impl::compare_exchange_strong): Likewise. * include/std/atomic (atomic::compare_exchange_weak): Likewise. (atomic::compare_exchange_strong): Likewise. ^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug c++/102177] Implement C++17 P0418R2 2021-09-02 14:28 [Bug c++/102177] New: Implement C++17 P0418R2 jakub at gcc dot gnu.org ` (3 preceding siblings ...) 2021-09-02 17:51 ` cvs-commit at gcc dot gnu.org @ 2022-04-21 15:33 ` cvs-commit at gcc dot gnu.org 4 siblings, 0 replies; 6+ messages in thread From: cvs-commit at gcc dot gnu.org @ 2022-04-21 15:33 UTC (permalink / raw) To: gcc-bugs https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102177 --- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> --- The releases/gcc-11 branch has been updated by Jonathan Wakely <redi@gcc.gnu.org>: https://gcc.gnu.org/g:fef3d31bd35f53cbcacaec63b279d38c38b78abe commit r11-9926-gfef3d31bd35f53cbcacaec63b279d38c38b78abe Author: Jonathan Wakely <jwakely@redhat.com> Date: Thu Sep 2 15:29:22 2021 +0100 libstdc++: Remove "no stronger" assertion in compare exchange [PR102177] P0418R2 removed some preconditions from std::atomic::compare_exchange_* but we still enforce them via __glibcxx_assert. This removes those assertions. Signed-off-by: Jonathan Wakely <jwakely@redhat.com> libstdc++-v3/ChangeLog: PR c++/102177 * include/bits/atomic_base.h (__is_valid_cmpexch_failure_order): New function to check if a memory order is valid for the failure case of compare exchange operations. (__atomic_base<I>::compare_exchange_weak): Simplify assertions by using __is_valid_cmpexch_failure_order. (__atomic_base<I>::compare_exchange_strong): Likewise. (__atomic_base<P*>::compare_exchange_weak): Likewise. (__atomic_base<P*>::compare_exchange_strong): Likewise. (__atomic_impl::compare_exchange_weak): Add assertion. (__atomic_impl::compare_exchange_strong): Likewise. * include/std/atomic (atomic::compare_exchange_weak): Likewise. (atomic::compare_exchange_strong): Likewise. (cherry picked from commit dba1ab212292839572fda60df00965e094a11252) ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2022-04-21 15:33 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2021-09-02 14:28 [Bug c++/102177] New: Implement C++17 P0418R2 jakub at gcc dot gnu.org 2021-09-02 14:28 ` [Bug c++/102177] " jakub at gcc dot gnu.org 2021-09-02 14:40 ` jakub at gcc dot gnu.org 2021-09-02 15:14 ` jakub at gcc dot gnu.org 2021-09-02 17:51 ` cvs-commit at gcc dot gnu.org 2022-04-21 15:33 ` cvs-commit at gcc dot gnu.org
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).