From: Paolo Bonzini <pbonzini@redhat.com>
To: Marc Morcos <marcmorcos@google.com>,
Richard Henderson <richard.henderson@linaro.org>,
Eduardo Habkost <eduardo@habkost.net>,
"Dr . David Alan Gilbert" <dave@treblig.org>
Cc: "Michael S . Tsirkin" <mst@redhat.com>,
Marcel Apfelbaum <marcel.apfelbaum@gmail.com>,
Markus Armbruster <armbru@redhat.com>,
Marcelo Tosatti <mtosatti@redhat.com>,
qemu-devel@nongnu.org, kvm@vger.kernel.org
Subject: Re: [PATCH 2/4] thread-pool: Fix thread race
Date: Mon, 15 Dec 2025 15:43:51 +0100 [thread overview]
Message-ID: <159ced33-46e4-4b86-85e7-eda01406f768@redhat.com> (raw)
In-Reply-To: <20251213001443.2041258-3-marcmorcos@google.com>
On 12/13/25 01:14, Marc Morcos wrote:
>
> req->ret = ret;
Better use qatomic_set here---will fix it myself, thanks!
Paolo
> - /* Write ret before state. */
> - smp_wmb();
> - req->state = THREAD_DONE;
> + /* _release to write ret before state. */
> + qatomic_store_release(&req->state, THREAD_DONE);
>
> qemu_bh_schedule(pool->completion_bh);
> qemu_mutex_lock(&pool->lock);
> @@ -180,7 +184,8 @@ static void thread_pool_completion_bh(void *opaque)
>
> restart:
> QLIST_FOREACH_SAFE(elem, &pool->head, all, next) {
> - if (elem->state != THREAD_DONE) {
> + /* _acquire to read state before ret. */
> + if (qatomic_load_acquire(&elem->state) != THREAD_DONE) {
> continue;
> }
>
> @@ -189,9 +194,6 @@ restart:
> QLIST_REMOVE(elem, all);
>
> if (elem->common.cb) {
> - /* Read state before ret. */
> - smp_rmb();
> -
> /* Schedule ourselves in case elem->common.cb() calls aio_poll() to
> * wait for another request that completed at the same time.
> */
> @@ -223,12 +225,12 @@ static void thread_pool_cancel(BlockAIOCB *acb)
> trace_thread_pool_cancel_aio(elem, elem->common.opaque);
>
> QEMU_LOCK_GUARD(&pool->lock);
> - if (elem->state == THREAD_QUEUED) {
> + if (qatomic_read(&elem->state) == THREAD_QUEUED) {
> QTAILQ_REMOVE(&pool->request_list, elem, reqs);
> qemu_bh_schedule(pool->completion_bh);
>
> - elem->state = THREAD_DONE;
> - elem->ret = -ECANCELED;
> + qatomic_set(&elem->ret, -ECANCELED);
> + qatomic_store_release(&elem->state, THREAD_DONE);
> }
>
> }
next prev parent reply other threads:[~2025-12-15 14:44 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-13 0:14 [PATCH 0/4] Clean up TSAN warnings Marc Morcos
2025-12-13 0:14 ` [PATCH 1/4] apic: Resize APICBASE Marc Morcos
2025-12-15 9:48 ` Alex Bennée
2025-12-13 0:14 ` [PATCH 2/4] thread-pool: Fix thread race Marc Morcos
2025-12-15 14:43 ` Paolo Bonzini [this message]
2025-12-15 17:51 ` Marc Morcos
2025-12-13 0:14 ` [PATCH 3/4] qmp: " Marc Morcos
2025-12-15 14:52 ` Paolo Bonzini
2025-12-15 17:48 ` Marc Morcos
2025-12-13 0:14 ` [PATCH 4/4] apic: Make apicbase accesses atomic to fix data race Marc Morcos
-- strict thread matches above, loose matches on Subject: below --
2025-12-13 0:12 [PATCH 0/4] Clean up TSAN warnings Marc Morcos
2025-12-13 0:12 ` [PATCH 2/4] thread-pool: Fix thread race Marc Morcos
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=159ced33-46e4-4b86-85e7-eda01406f768@redhat.com \
--to=pbonzini@redhat.com \
--cc=armbru@redhat.com \
--cc=dave@treblig.org \
--cc=eduardo@habkost.net \
--cc=kvm@vger.kernel.org \
--cc=marcel.apfelbaum@gmail.com \
--cc=marcmorcos@google.com \
--cc=mst@redhat.com \
--cc=mtosatti@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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 NNTP newsgroup(s).