This is a note to let you know that I've just added the patch titled drm/nouveau/tmr: handle races with hw when updating the next alarm time to the 4.9-stable tree which can be found at: http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary The filename of the patch is: drm-nouveau-tmr-handle-races-with-hw-when-updating-the-next-alarm-time.patch and it can be found in the queue-4.9 subdirectory. If you, or anyone else, feels it should not be added to the stable tree, please let <stable@xxxxxxxxxxxxxxx> know about it. >From 1b0f84380b10ee97f7d2dd191294de9017e94d1d Mon Sep 17 00:00:00 2001 From: Ben Skeggs <bskeggs@xxxxxxxxxx> Date: Thu, 11 May 2017 17:19:48 +1000 Subject: drm/nouveau/tmr: handle races with hw when updating the next alarm time From: Ben Skeggs <bskeggs@xxxxxxxxxx> commit 1b0f84380b10ee97f7d2dd191294de9017e94d1d upstream. If the time to the next alarm is short enough, we could race with HW and end up with an ~4 second delay until it triggers. Fix this by checking again after we update HW. Signed-off-by: Ben Skeggs <bskeggs@xxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c | 26 ++++++++++++++--------- 1 file changed, 16 insertions(+), 10 deletions(-) --- a/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c @@ -36,23 +36,29 @@ nvkm_timer_alarm_trigger(struct nvkm_tim unsigned long flags; LIST_HEAD(exec); - /* move any due alarms off the pending list */ + /* Process pending alarms. */ spin_lock_irqsave(&tmr->lock, flags); list_for_each_entry_safe(alarm, atemp, &tmr->alarms, head) { - if (alarm->timestamp <= nvkm_timer_read(tmr)) - list_move_tail(&alarm->head, &exec); + /* Have we hit the earliest alarm that hasn't gone off? */ + if (alarm->timestamp > nvkm_timer_read(tmr)) { + /* Schedule it. If we didn't race, we're done. */ + tmr->func->alarm_init(tmr, alarm->timestamp); + if (alarm->timestamp > nvkm_timer_read(tmr)) + break; + } + + /* Move to completed list. We'll drop the lock before + * executing the callback so it can reschedule itself. + */ + list_move_tail(&alarm->head, &exec); } - /* reschedule interrupt for next alarm time */ - if (!list_empty(&tmr->alarms)) { - alarm = list_first_entry(&tmr->alarms, typeof(*alarm), head); - tmr->func->alarm_init(tmr, alarm->timestamp); - } else { + /* Shut down interrupt if no more pending alarms. */ + if (list_empty(&tmr->alarms)) tmr->func->alarm_fini(tmr); - } spin_unlock_irqrestore(&tmr->lock, flags); - /* execute any pending alarm handlers */ + /* Execute completed callbacks. */ list_for_each_entry_safe(alarm, atemp, &exec, head) { list_del_init(&alarm->head); alarm->func(alarm); Patches currently in stable-queue which might be from bskeggs@xxxxxxxxxx are queue-4.9/drm-nouveau-tmr-ack-interrupt-before-processing-alarms.patch queue-4.9/drm-nouveau-tmr-fix-corruption-of-the-pending-list-when-rescheduling-an-alarm.patch queue-4.9/drm-nouveau-tmr-avoid-processing-completed-alarms-when-adding-a-new-one.patch queue-4.9/drm-nouveau-mmu-nv4a-use-nv04-mmu-rather-than-the-nv44-one.patch queue-4.9/drm-nouveau-therm-remove-ineffective-workarounds-for-alarm-bugs.patch queue-4.9/drm-nouveau-tmr-handle-races-with-hw-when-updating-the-next-alarm-time.patch