On Mon, 2026-09-07 at 11:42 +0200, Philipp Stanner wrote: > Moreover, even if we did remove ops->release in drm_sched, it would > still be a race: a driver's contract is the hardware_fence, the rule > being that you have to signal those. So after signaling the last > hardware_fence, you could actually start releasing resources, but it > might be that finished_fences are still in-flight and are unsignaled. > > So we have some sort of fence -> fence race here, too.
Forget about that, signaling the hw_fence should always result in undelayed signaling of the finished-fence. So it would seem that making the signaled-state instead of ops == NULL the decoupling point in dma_fence is the solution to go for? P.
