From: Emanuele Giuseppe Esposito <eesposit@redhat.com>
To: qemu-block@nongnu.org
Cc: Kevin Wolf <kwolf@redhat.com>, Hanna Reitz <hreitz@redhat.com>,
Paolo Bonzini <pbonzini@redhat.com>, John Snow <jsnow@redhat.com>,
Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>,
Wen Congyang <wencongyang2@huawei.com>,
Xie Changlong <xiechanglong.d@gmail.com>,
Markus Armbruster <armbru@redhat.com>,
Stefan Hajnoczi <stefanha@redhat.com>, Fam Zheng <fam@euphon.net>,
qemu-devel@nongnu.org
Subject: [PATCH v7 15/18] job: detect change of aiocontext within job coroutine
Date: Thu, 16 Jun 2022 09:18:32 -0400 [thread overview]
Message-ID: <20220616131835.2004262-16-eesposit@redhat.com> (raw)
In-Reply-To: <20220616131835.2004262-1-eesposit@redhat.com>
From: Paolo Bonzini <pbonzini@redhat.com>
We want to make sure access of job->aio_context is always done
under either BQL or job_mutex. The problem is that using
aio_co_enter(job->aiocontext, job->co) in job_start and job_enter_cond
makes the coroutine immediately resume, so we can't hold the job lock.
And caching it is not safe either, as it might change.
job_start is under BQL, so it can freely read job->aiocontext, but
job_enter_cond is not. In order to fix this, use aio_co_wake():
the advantage is that it won't use job->aiocontext, but the
main disadvantage is that it won't be able to detect a change of
job AioContext.
Calling bdrv_try_set_aio_context() will issue the following calls
(simplified):
* in terms of bdrv callbacks:
.drained_begin -> .set_aio_context -> .drained_end
* in terms of child_job functions:
child_job_drained_begin -> child_job_set_aio_context -> child_job_drained_end
* in terms of job functions:
job_pause_locked -> job_set_aio_context -> job_resume_locked
We can see that after setting the new aio_context, job_resume_locked
calls again job_enter_cond, which then invokes aio_co_wake(). But
while job->aiocontext has been set in job_set_aio_context,
job->co->ctx has not changed, so the coroutine would be entering in
the wrong aiocontext.
Using aio_co_schedule in job_resume_locked() might seem as a valid
alternative, but the problem is that the bh resuming the coroutine
is not scheduled immediately, and if in the meanwhile another
bdrv_try_set_aio_context() is run (see test_propagate_mirror() in
test-block-iothread.c), we would have the first schedule in the
wrong aiocontext, and the second set of drains won't even manage
to schedule the coroutine, as job->busy would still be true from
the previous job_resume_locked().
The solution is to stick with aio_co_wake(), but then detect every time
the coroutine resumes back from yielding if job->aio_context
has changed. If so, we can reschedule it to the new context.
Check for the aiocontext change in job_do_yield_locked because:
1) aio_co_reschedule_self requires to be in the running coroutine
2) since child_job_set_aio_context allows changing the aiocontext only
while the job is paused, this is the exact place where the coroutine
resumes, before running JobDriver's code.
Reviewed-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
job.c | 22 ++++++++++++++++++++--
1 file changed, 20 insertions(+), 2 deletions(-)
diff --git a/job.c b/job.c
index b6b9431b2d..389c134a90 100644
--- a/job.c
+++ b/job.c
@@ -543,11 +543,12 @@ void job_enter_cond_locked(Job *job, bool(*fn)(Job *job))
return;
}
- assert(!job->deferred_to_main_loop);
timer_del(&job->sleep_timer);
job->busy = true;
real_job_unlock();
- aio_co_enter(job->aio_context, job->co);
+ job_unlock();
+ aio_co_wake(job->co);
+ job_lock();
}
void job_enter(Job *job)
@@ -568,6 +569,8 @@ void job_enter(Job *job)
*/
static void coroutine_fn job_do_yield_locked(Job *job, uint64_t ns)
{
+ AioContext *next_aio_context;
+
real_job_lock();
if (ns != -1) {
timer_mod(&job->sleep_timer, ns);
@@ -579,6 +582,19 @@ static void coroutine_fn job_do_yield_locked(Job *job, uint64_t ns)
qemu_coroutine_yield();
job_lock();
+ next_aio_context = job->aio_context;
+ /*
+ * Coroutine has resumed, but in the meanwhile the job AioContext
+ * might have changed via bdrv_try_set_aio_context(), so we need to move
+ * the coroutine too in the new aiocontext.
+ */
+ while (qemu_get_current_aio_context() != next_aio_context) {
+ job_unlock();
+ aio_co_reschedule_self(next_aio_context);
+ job_lock();
+ next_aio_context = job->aio_context;
+ }
+
/* Set by job_enter_cond_locked() before re-entering the coroutine. */
assert(job->busy);
}
@@ -1122,6 +1138,8 @@ static void coroutine_fn job_co_entry(void *opaque)
void job_start(Job *job)
{
+ assert(qemu_in_main_thread());
+
WITH_JOB_LOCK_GUARD() {
assert(job && !job_started(job) && job->paused &&
job->driver && job->driver->run);
--
2.31.1
next prev parent reply other threads:[~2022-06-16 13:45 UTC|newest]
Thread overview: 48+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-16 13:18 [PATCH v7 00/18] job: replace AioContext lock with job_mutex Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 01/18] job.c: make job_mutex and job_lock/unlock() public Emanuele Giuseppe Esposito
2022-06-21 13:47 ` Vladimir Sementsov-Ogievskiy
2022-06-24 18:22 ` Vladimir Sementsov-Ogievskiy
2022-06-28 13:08 ` Emanuele Giuseppe Esposito
2022-06-28 15:20 ` Vladimir Sementsov-Ogievskiy
2022-06-16 13:18 ` [PATCH v7 02/18] job.h: categorize fields in struct Job Emanuele Giuseppe Esposito
2022-06-21 14:29 ` Vladimir Sementsov-Ogievskiy
2022-06-16 13:18 ` [PATCH v7 03/18] job.c: API functions not used outside should be static Emanuele Giuseppe Esposito
2022-06-21 14:34 ` Vladimir Sementsov-Ogievskiy
2022-06-16 13:18 ` [PATCH v7 04/18] aio-wait.h: introduce AIO_WAIT_WHILE_UNLOCKED Emanuele Giuseppe Esposito
2022-06-21 14:40 ` Vladimir Sementsov-Ogievskiy
2022-06-16 13:18 ` [PATCH v7 05/18] job.h: add _locked duplicates for job API functions called with and without job_mutex Emanuele Giuseppe Esposito
2022-06-21 15:03 ` Vladimir Sementsov-Ogievskiy
2022-06-22 14:26 ` Emanuele Giuseppe Esposito
2022-06-22 18:12 ` Vladimir Sementsov-Ogievskiy
2022-06-16 13:18 ` [PATCH v7 06/18] jobs: protect jobs with job_lock/unlock Emanuele Giuseppe Esposito
2022-06-21 16:47 ` Vladimir Sementsov-Ogievskiy
2022-06-21 17:09 ` Vladimir Sementsov-Ogievskiy
2022-06-16 13:18 ` [PATCH v7 07/18] jobs: add job lock in find_* functions Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 08/18] jobs: use job locks also in the unit tests Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 09/18] block/mirror.c: use of job helpers in drivers to avoid TOC/TOU Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 10/18] jobs: rename static functions called with job_mutex held Emanuele Giuseppe Esposito
2022-06-21 17:26 ` Vladimir Sementsov-Ogievskiy
2022-06-22 14:26 ` Emanuele Giuseppe Esposito
2022-06-22 18:38 ` Vladimir Sementsov-Ogievskiy
2022-06-23 9:08 ` Emanuele Giuseppe Esposito
2022-06-23 11:10 ` Vladimir Sementsov-Ogievskiy
2022-06-23 11:19 ` Emanuele Giuseppe Esposito
2022-06-23 11:58 ` Vladimir Sementsov-Ogievskiy
2022-06-24 14:29 ` Kevin Wolf
2022-06-24 15:28 ` Paolo Bonzini
2022-06-24 17:20 ` Emanuele Giuseppe Esposito
2022-06-28 7:40 ` Emanuele Giuseppe Esposito
2022-06-28 10:47 ` Vladimir Sementsov-Ogievskiy
2022-06-28 13:04 ` Emanuele Giuseppe Esposito
2022-06-28 15:22 ` Vladimir Sementsov-Ogievskiy
2022-06-28 15:26 ` Vladimir Sementsov-Ogievskiy
2022-06-28 17:28 ` Emanuele Giuseppe Esposito
2022-06-28 19:42 ` Vladimir Sementsov-Ogievskiy
2022-06-16 13:18 ` [PATCH v7 11/18] job.h: rename job API " Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 12/18] block_job: rename block_job " Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 13/18] job.h: define unlocked functions Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 14/18] commit and mirror: create new nodes using bdrv_get_aio_context, and not the job aiocontext Emanuele Giuseppe Esposito
2022-06-16 13:18 ` Emanuele Giuseppe Esposito [this message]
2022-06-16 13:18 ` [PATCH v7 16/18] jobs: protect job.aio_context with BQL and job_mutex Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 17/18] job.c: enable job lock/unlock and remove Aiocontext locks Emanuele Giuseppe Esposito
2022-06-16 13:18 ` [PATCH v7 18/18] block_job_query: remove atomic read Emanuele Giuseppe Esposito
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=20220616131835.2004262-16-eesposit@redhat.com \
--to=eesposit@redhat.com \
--cc=armbru@redhat.com \
--cc=fam@euphon.net \
--cc=hreitz@redhat.com \
--cc=jsnow@redhat.com \
--cc=kwolf@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-block@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=stefanha@redhat.com \
--cc=vsementsov@virtuozzo.com \
--cc=wencongyang2@huawei.com \
--cc=xiechanglong.d@gmail.com \
/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).