Quoting Bloomfield, Jon (2019-08-12 15:39:33) > > -----Original Message----- > > From: Chris Wilson <[email protected]> > > Sent: Friday, August 9, 2019 4:35 PM > > To: [email protected] > > Cc: Joonas Lahtinen <[email protected]>; Winiarski, Michal > > <[email protected]>; Bloomfield, Jon <[email protected]> > > Subject: Re: [PATCH 5/5] drm/i915: Cancel non-persistent contexts on close > > > > Quoting Chris Wilson (2019-08-06 14:47:25) > > > Normally, we rely on our hangcheck to prevent persistent batches from > > > hogging the GPU. However, if the user disables hangcheck, this mechanism > > > breaks down. Despite our insistence that this is unsafe, the users are > > > equally insistent that they want to use endless batches and will disable > > > the hangcheck mechanism. We are looking are perhaps replacing hangcheck > > > with a softer mechanism, that sends a pulse down the engine to check if > > > it is well. We can use the same preemptive pulse to flush an active > > > persistent context off the GPU upon context close, preventing resources > > > being lost and unkillable requests remaining on the GPU, after process > > > termination. To avoid changing the ABI and accidentally breaking > > > existing userspace, we make the persistence of a context explicit and > > > enable it by default. Userspace can opt out of persistent mode (forcing > > > requests to be cancelled when the context is closed by process > > > termination or explicitly) by a context parameter, or to facilitate > > > existing use-cases by disabling hangcheck (i915.enable_hangcheck=0). > > > (Note, one of the outcomes for supporting endless mode will be the > > > removal of hangchecking, at which point opting into persistent mode will > > > be mandatory, or maybe the default.) > > > > For the record, I've finally run into examples of desktop clients > > exiting before their rendering is shown. No longer hypothetical. > > -Chris > > Can you share any details - Might be useful for testing.
In cinnamon, the atl-tab switcher seems to be one. Another one seems to be around firefox, but I haven't established the trigger. I should actually log who owned the context! -Chris _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
