Re: [PATCH v3] drm: Don't free jobs in wait_event_interruptible()

From: Steven Price <steven.price@arm.com>
To: "Koenig, Christian" <Christian.Koenig@amd.com>,
	Daniel Vetter <daniel@ffwll.ch>, David Airlie <airlied@linux.ie>
Cc: Sharat Masetty <smasetty@codeaurora.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"dri-devel@lists.freedesktop.org"
	<dri-devel@lists.freedesktop.org>,
	Nayan Deshmukh <nayan26deshmukh@gmail.com>,
	"Deucher, Alexander" <Alexander.Deucher@amd.com>
Subject: Re: [PATCH v3] drm: Don't free jobs in wait_event_interruptible()
Date: Thu, 26 Sep 2019 14:43:23 +0100	[thread overview]
Message-ID: <f045e1d6-5263-58c2-ce28-2cd215067bdb@arm.com> (raw)
In-Reply-To: <73f058bb-8ba8-611b-e3a0-f6282445179c@amd.com>

On 26/09/2019 14:37, Koenig, Christian wrote:
> Am 26.09.19 um 14:31 schrieb Steven Price:
>> drm_sched_cleanup_jobs() attempts to free finished jobs, however because
>> it is called as the condition of wait_event_interruptible() it must not
>> sleep. Unfortuantly some free callbacks (notibly for Panfrost) do sleep.
>>
>> Instead let's rename drm_sched_cleanup_jobs() to
>> drm_sched_get_cleanup_job() and simply return a job for processing if
>> there is one. The caller can then call the free_job() callback outside
>> the wait_event_interruptible() where sleeping is possible before
>> re-checking and returning to sleep if necessary.
>>
>> Signed-off-by: Steven Price <steven.price@arm.com>
>> ---
>>
>> Changes from v2:
>>   * Actually move list_first_entry_or_null() within the lock
>>
>>   drivers/gpu/drm/scheduler/sched_main.c | 42 ++++++++++++++------------
>>   1 file changed, 23 insertions(+), 19 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/scheduler/sched_main.c b/drivers/gpu/drm/scheduler/sched_main.c
>> index 9a0ee74d82dc..e4bd792f2b29 100644
>> --- a/drivers/gpu/drm/scheduler/sched_main.c
>> +++ b/drivers/gpu/drm/scheduler/sched_main.c
>> @@ -622,43 +622,41 @@ static void drm_sched_process_job(struct dma_fence *f, struct dma_fence_cb *cb)
>>   }
>>   
>>   /**
>> - * drm_sched_cleanup_jobs - destroy finished jobs
>> + * drm_sched_get_cleanup_job - fetch the next finished job to be destroyed
>>    *
>>    * @sched: scheduler instance
>>    *
>> - * Remove all finished jobs from the mirror list and destroy them.
>> + * Returns the next finished job from the mirror list (if there is one)
>> + * ready for it to be destroyed.
>>    */
>> -static void drm_sched_cleanup_jobs(struct drm_gpu_scheduler *sched)
>> +static struct drm_sched_job *
>> +drm_sched_get_cleanup_job(struct drm_gpu_scheduler *sched)
>>   {
>> +	struct drm_sched_job *job = NULL;
>>   	unsigned long flags;
>>   
>>   	/* Don't destroy jobs while the timeout worker is running */
>>   	if (sched->timeout != MAX_SCHEDULE_TIMEOUT &&
>>   	    !cancel_delayed_work(&sched->work_tdr))
>> -		return;
>> -
>> +		return NULL;
>>   
>> -	while (!list_empty(&sched->ring_mirror_list)) {
>> -		struct drm_sched_job *job;
>> +	spin_lock_irqsave(&sched->job_list_lock, flags);
>>   
>> -		job = list_first_entry(&sched->ring_mirror_list,
>> +	job = list_first_entry_or_null(&sched->ring_mirror_list,
>>   				       struct drm_sched_job, node);
>> -		if (!dma_fence_is_signaled(&job->s_fence->finished))
>> -			break;
>>   
>> -		spin_lock_irqsave(&sched->job_list_lock, flags);
>> +	if (job && dma_fence_is_signaled(&job->s_fence->finished)) {
>>   		/* remove job from ring_mirror_list */
>>   		list_del_init(&job->node);
>> -		spin_unlock_irqrestore(&sched->job_list_lock, flags);
>> -
>> -		sched->ops->free_job(job);
>> +	} else {
>> +		job = NULL;
>> +		/* queue timeout for next job */
>> +		drm_sched_start_timeout(sched);
> 
> Sorry I've only seen this right now: This won't work.
> 
> See you always need to start the timeout, or cancel_delayed_work() will 
> abort the whole thing after returning the first job.

Good spot!

My logic was that drm_sched_main() would call again and that would
re-arm the timeout after the job has been completed.

But looking at this more closely I see that isn't actually going to work
reliably - if the work has been cancelled the first time then
cancel_delayed_work() will return false causing an early out and the
timeout isn't re-armed.

Is it as simple as moving drm_sched_start_timeout() out of the else case
and unconditionally re-arming the timeout? I'm worried about races with
the job that is being returned, although it has already been removed
from ring_mirror_list so perhaps it's safe to do?

Alternatively the caller could manually re-arm the timeout after
handling the job free.

Steve

> Christian.
> 
>>   	}
>>   
>> -	/* queue timeout for next job */
>> -	spin_lock_irqsave(&sched->job_list_lock, flags);
>> -	drm_sched_start_timeout(sched);
>>   	spin_unlock_irqrestore(&sched->job_list_lock, flags);
>>   
>> +	return job;
>>   }
>>   
>>   /**
>> @@ -698,12 +696,18 @@ static int drm_sched_main(void *param)
>>   		struct drm_sched_fence *s_fence;
>>   		struct drm_sched_job *sched_job;
>>   		struct dma_fence *fence;
>> +		struct drm_sched_job *cleanup_job = NULL;
>>   
>>   		wait_event_interruptible(sched->wake_up_worker,
>> -					 (drm_sched_cleanup_jobs(sched),
>> +					 (cleanup_job = drm_sched_get_cleanup_job(sched)) ||
>>   					 (!drm_sched_blocked(sched) &&
>>   					  (entity = drm_sched_select_entity(sched))) ||
>> -					 kthread_should_stop()));
>> +					 kthread_should_stop());
>> +
>> +		while (cleanup_job) {
>> +			sched->ops->free_job(cleanup_job);
>> +			cleanup_job = drm_sched_get_cleanup_job(sched);
>> +		}
>>   
>>   		if (!entity)
>>   			continue;
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/dri-devel
> 