From 7db0ba242b3a7ddcfc1450bc50eb3102c68f0244 Mon Sep 17 00:00:00 2001 From: Daniel Vetter Date: Thu, 6 Dec 2012 16:23:37 +0100 Subject: [PATCH] drm/i915: clarify concurrent hang detect/gpu reset consistency Damien Lespiau wondered how race the gpu reset/hang detection code is against concurrent gpu resets/hang detections or combinations thereof. Luckily the single work item is guranteed to never run concurrently, so reset handling is already single-threaded. Hence we only have to worry about concurrent hang detections, or a hang detection firing off while we're still processing an older gpu reset request. Due to the new mechanism of setting the reset in progress flag and the ordering guaranteed by the schedule_work function there's nothing to do but add a comment explaining why we're safe. The only thing I've noticed is that we still try to reset the gpu now, even when it is declared terminally wedged. Add a check for that to avoid continous warnings about failed resets, in case the hangcheck timer ever gets stuck. Reviewed-by: Damien Lespiau Signed-off-by: Daniel Vetter --- drivers/gpu/drm/i915/i915_irq.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index f833f2c155f..943db1001cf 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -875,9 +875,20 @@ static void i915_error_work_func(struct work_struct *work) kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, error_event); - if (i915_reset_in_progress(error)) { + /* + * Note that there's only one work item which does gpu resets, so we + * need not worry about concurrent gpu resets potentially incrementing + * error->reset_counter twice. We only need to take care of another + * racing irq/hangcheck declaring the gpu dead for a second time. A + * quick check for that is good enough: schedule_work ensures the + * correct ordering between hang detection and this work item, and since + * the reset in-progress bit is only ever set by code outside of this + * work we don't need to worry about any other races. + */ + if (i915_reset_in_progress(error) && !i915_terminally_wedged(error)) { DRM_DEBUG_DRIVER("resetting chip\n"); - kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, reset_event); + kobject_uevent_env(&dev->primary->kdev.kobj, KOBJ_CHANGE, + reset_event); ret = i915_reset(dev); -- 2.46.0