Re: [PATCH for-6.0? 1/3] job: Add job_wait_unpaused() for block-job-complete

[Max Reitz <mreitz@redhat.com>]

On 08.04.21 18:58, Vladimir Sementsov-Ogievskiy wrote:
> 08.04.2021 19:20, Max Reitz wrote:
>> block-job-complete can only be applied when the job is READY, not when
>> it is on STANDBY (ready, but paused).  Draining a job technically pauses
>> it (which makes a READY job enter STANDBY), and ending the drained
>> section does not synchronously resume it, but only schedules the job,
>> which will then be resumed.  So attempting to complete a job immediately
>> after a drained section may sometimes fail.
>>
>> That is bad at least because users cannot really work nicely around
>> this: A job may be paused and resumed at any time, so waiting for the
>> job to be in the READY state and then issuing a block-job-complete poses
>> a TOCTTOU problem.  The only way around it would be to issue
>> block-job-complete until it no longer fails due to the job being in the
>> STANDBY state, but that would not be nice.
>>
>> We can solve the problem by allowing block-job-complete to be invoked on
>> jobs that are on STANDBY, if that status is the result of a drained
>> section (not because the user has paused the job), and that section has
>> ended.  That is, if the job is on STANDBY, but scheduled to be resumed.
>>
>> Perhaps we could actually just directly allow this, seeing that mirror
>> is the only user of ready/complete, and that mirror_complete() could
>> probably work under the given circumstances, but there may be many side
>> effects to consider.
>>
>> It is simpler to add a function job_wait_unpaused() that waits for the
>> job to be resumed (under said circumstances), and to make
>> qmp_block_job_complete() use it to delay job_complete() until then.
>>
>> Buglink: https://bugzilla.redhat.com/show_bug.cgi?id=1945635
>> Signed-off-by: Max Reitz <mreitz@redhat.com>
>> ---
>>   include/qemu/job.h | 15 +++++++++++++++
>>   blockdev.c         |  3 +++
>>   job.c              | 42 ++++++++++++++++++++++++++++++++++++++++++
>>   3 files changed, 60 insertions(+)
>>
>> diff --git a/include/qemu/job.h b/include/qemu/job.h
>> index efc6fa7544..cf3082b6d7 100644
>> --- a/include/qemu/job.h
>> +++ b/include/qemu/job.h
>> @@ -563,4 +563,19 @@ void job_dismiss(Job **job, Error **errp);
>>    */
>>   int job_finish_sync(Job *job, void (*finish)(Job *, Error **errp), 
>> Error **errp);
>> +/**
>> + * If the job has been paused because of a drained section, and that
>> + * section has ended, wait until the job is resumed.
>> + *
>> + * Return 0 if the job is not paused, or if it has been successfully
>> + * resumed.
>> + * Return an error if the job has been paused in such a way that
>> + * waiting will not resume it, i.e. if it has been paused by the user,
>> + * or if it is still drained.
>> + *
>> + * Callers must be in the home AioContext and hold the AioContext lock
>> + * of job->aio_context.
>> + */
>> +int job_wait_unpaused(Job *job, Error **errp);
>> +
>>   #endif
>> diff --git a/blockdev.c b/blockdev.c
>> index a57590aae4..c0cc2fa364 100644
>> --- a/blockdev.c
>> +++ b/blockdev.c
>> @@ -3414,6 +3414,9 @@ void qmp_block_job_complete(const char *device, 
>> Error **errp)
>>           return;
>>       }
>> +    if (job_wait_unpaused(&job->job, errp) < 0) {
>> +        return;
>> +    }
>>       trace_qmp_block_job_complete(job);
>>       job_complete(&job->job, errp);
>>       aio_context_release(aio_context);
>> diff --git a/job.c b/job.c
>> index 289edee143..1ea30fd294 100644
>> --- a/job.c
>> +++ b/job.c
>> @@ -1023,3 +1023,45 @@ int job_finish_sync(Job *job, void 
>> (*finish)(Job *, Error **errp), Error **errp)
>>       job_unref(job);
>>       return ret;
>>   }
>> +
>> +int job_wait_unpaused(Job *job, Error **errp)
>> +{
>> +    /*
>> +     * Only run this function from the main context, because this is
>> +     * what we need, and this way we do not have to think about what
>> +     * happens if the user concurrently pauses the job from the main
>> +     * monitor.
>> +     */
>> +    assert(qemu_get_current_aio_context() == qemu_get_aio_context());
>> +
>> +    /*
>> +     * Quick path (e.g. so we do not get an error if pause_count > 0
>> +     * but the job is not even paused)
>> +     */
>> +    if (!job->paused) {
>> +        return 0;
>> +    }
>> +
>> +    /* If the user has paused the job, waiting will not help */
>> +    if (job->user_paused) {
>> +        error_setg(errp, "Job '%s' has been paused by the user", 
>> job->id);
>> +        return -EBUSY;
>> +    }
>> +
>> +    /* Similarly, if the job is still drained, waiting will not help 
>> either */
>> +    if (job->pause_count > 0) {
>> +        error_setg(errp, "Job '%s' is blocked and cannot be 
>> unpaused", job->id);
>> +        return -EBUSY;
>> +    }
>> +
>> +    /*
>> +     * This function is specifically for waiting for a job to be
>> +     * resumed after a drained section.  Ending the drained section
>> +     * includes a job_enter(), which schedules the job loop to be run,
>> +     * and once it does, job->paused will be cleared.  Therefore, we
>> +     * do not need to invoke job_enter() here.
>> +     */
>> +    AIO_WAIT_WHILE(job->aio_context, job->paused);
>> +
>> +    return 0;
>> +}
>>
> 
> Hmm.. It seems that when job->pause_count becomes 0, job_enter is 
> called, and the period when pause_count is 0 but paused is still true 
> should be relatively shot.

Yes.

> And patch doesn't help if user call 
> job-complete during drained section.

Is that possible?

(If the drained section is in some IO thread, I suppose we could drop 
the pause_count condition as I wrote in my reply to John.  Because if 
some IO thread has drained the job, waiting will help.)

> So it looks like the patch will 
> help relatively seldom.. Or I'm missing something?

No, you’re right.  The condition is rather rare.  Which I think is good 
because it means that block-job-complete having to wait for a job will 
be a rare occurrence.

I have a Ruby test script that can reproduce it sometimes, but I’m not 
sure that it will reproduce the problem on others’ machines at all. 
I’ll try to rewrite it as a Python iotest so you can try for yourself.

> job-complete command is async. Can we instead just add a boolean like 
> job->completion_requested, and set it if job-complete called in STANDBY 
> state, and on job_resume job_complete will be called automatically if 
> this boolean is true?

job-complete is not async, it can return errors, and mirror_complete() 
has several paths that do so.  What would you do with those errors?  I 
can only imagine sending an event (which would be ignored by existing 
management software) or failing the block job (which seems wrong, but 
perhaps would be the right thing to do).

The problem is that when block-job-complete returns success, a 
BLOCK_JOB_COMPLETE is bound to follow at some point (unless the job runs 
into an error, or it cannot converge).  This contract would be broken 
(unless we let mirror_complete() fail the whole block job).

Max