If we are planning to perform a lite-restore, we want to replace the running context with itself. We can therefore exempt this from the preemption timeout.
Doing so does however increase the risk of a higher priority client arriving later finding itself unable to proceed because of an unchecked earlier failed preemption. Signed-off-by: Chris Wilson <[email protected]> --- drivers/gpu/drm/i915/gt/intel_lrc.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index 9a0d0282f3ca..eb83c87c8b4e 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -1737,6 +1737,11 @@ static unsigned long active_preempt_timeout(struct intel_engine_cs *engine) if (!rq) return 0; + /* Don't enforce a lite-restore, the client is already running */ + GEM_BUG_ON(!engine->execlists.pending[0]); + if (rq->context == engine->execlists.pending[0]->context) + return 0; + /* Force a fast reset for terminated contexts (ignoring sysfs!) */ if (unlikely(intel_context_is_banned(rq->context))) return 1; -- 2.25.0 _______________________________________________ Intel-gfx mailing list [email protected] https://lists.freedesktop.org/mailman/listinfo/intel-gfx
