Re: [PATCH v2 2/6] drm/sched: Prevent teardown waitque from blocking too long

[Danilo Krummrich <dakr@kernel.org>]

On Fri, May 16, 2025 at 10:33:30AM +0100, Tvrtko Ursulin wrote:
> On 24/04/2025 10:55, Philipp Stanner wrote:
> > +	 * @kill_fence_context: kill the fence context belonging to this scheduler
> 
> Which fence context would that be? ;)

There's one one per ring and a scheduler instance represents a single ring. So,
what should be specified here?

> Also, "fence context" would be a new terminology in gpu_scheduler.h API
> level. You could call it ->sched_fini() or similar to signify at which point
> in the API it gets called and then the fact it takes sched as parameter
> would be natural.

The driver should tear down the fence context in this callback, not the while
scheduler. ->sched_fini() would hence be misleading.

> We also probably want some commentary on the topic of indefinite (or very
> long at least) blocking a thread exit / SIGINT/TERM/KILL time.

You mean in case the driver does implement the callback, but does *not* properly
tear down the fence context? So, you ask for describing potential consequences
of drivers having bugs in the implementation of the callback? Or something else?

> Is the idea to let drivers shoot themselves in the foot or what?

Please abstain from such rhetorical questions, that's not a good way of having
technical discussions.

- Danilo