Quoting Chris Wilson (2018-05-17 21:06:10) > We make a decision at module load whether to use the GuC backend or not, > but lose that setup across set-wedge. Currently, the guc doesn't > override the engine->set_default_submission hook letting execlists sneak > back in temporarily on unwedging leading to an unbalanced park/unpark. > > v2: Remove comment about switching back temporarily to execlists on > guc_submission_disable(). We currently only call disable on shutdown, > and plan to also call disable before suspend and reset, in which case we > will either restore guc submission or mark the driver as wedged, making > the reset back to execlists pointless. > > Testcase: igt/gem_eio > Signed-off-by: Chris Wilson <[email protected]> > Cc: Michał Winiarski <[email protected]> > Cc: Michal Wajdeczko <[email protected]>
I think this makes sense given the current limitations; and fixes the crash we have in gem_eio on guc. Worth pulling into your series Michal and giving it another spin? -Chris _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
