drv->drained_poll() is only implemented in mirror, and allows it to drain from within the coroutine. The mirror implementation uses in_drain flag to recognize when it is draining from coroutine, and consequently avoid deadlocking (wait the poll condition in child_job_drained_poll to wait for itself).
The problem is that this flag is dangerous, because it breaks bdrv_drained_begin() invariants: once drained_begin ends, all jobs, in_flight requests, and anything running in the iothread are blocked. This can be broken in such way: iothread(mirror): s->in_drain = true; // mirror.c:1112 main loop: bdrv_drained_begin(mirror_bs); /* * drained_begin wait for bdrv_drain_poll_top_level() condition, * that translates in child_job_drained_poll() for jobs, but * mirror implements drv->drained_poll() so it returns * !!in_flight_requests, which his 0 (assertion in mirror.c:1105). */ main loop: thinks iothread is stopped and is modifying the graph... iothread(mirror): *continues*, as nothing is stopping it iothread(mirror): bdrv_drained_begin(bs); /* draining reads the graph while it is modified!! */ main loop: done modifying the graph... In order to fix this, we can simply allow drv->drained_poll() to be called only by the iothread, and not the main loop. We distinguish it by using in_aio_context_home_thread(), that returns false if @ctx is not the same as the thread that runs it. Co-Developed-by: Paolo Bonzini <pbonz...@redhat.com> Signed-off-by: Emanuele Giuseppe Esposito <eespo...@redhat.com> --- blockjob.c | 13 ++++++++++--- 1 file changed, 10 insertions(+), 3 deletions(-) diff --git a/blockjob.c b/blockjob.c index 4868453d74..14a919b3cc 100644 --- a/blockjob.c +++ b/blockjob.c @@ -110,7 +110,9 @@ static bool child_job_drained_poll(BdrvChild *c) BlockJob *bjob = c->opaque; Job *job = &bjob->job; const BlockJobDriver *drv = block_job_driver(bjob); + AioContext *ctx; + ctx = job->aio_context; /* An inactive or completed job doesn't have any pending requests. Jobs * with !job->busy are either already paused or have a pause point after * being reentered, so no job driver code will run before they pause. */ @@ -118,9 +120,14 @@ static bool child_job_drained_poll(BdrvChild *c) return false; } - /* Otherwise, assume that it isn't fully stopped yet, but allow the job to - * override this assumption. */ - if (drv->drained_poll) { + /* + * Otherwise, assume that it isn't fully stopped yet, but allow the job to + * override this assumption, if the drain is being performed in the + * iothread. We need to check that the caller is the home thread because + * it could otherwise lead the main loop to exit polling while the job + * has not paused yet. + */ + if (in_aio_context_home_thread(ctx) && drv->drained_poll) { return drv->drained_poll(bjob); } else { return true; -- 2.31.1