shinys000114/linux
1
0
Fork 0
mirror of https://github.com/hardkernel/linux.git synced 2026-06-07 03:15:31 +09:00
Code Issues Packages Projects Releases Wiki Activity

drm/nouveau/tmr: fully separate alarm execution/pending lists

Browse Source 
commit b4e382ca75 upstream.

Reusing the list_head for both is a bad idea.  Callback execution is done
with the lock dropped so that alarms can be rescheduled from the callback,
which means that with some unfortunate timing, lists can get corrupted.

The execution list should not require its own locking, the single function
that uses it can only be called from a single context.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This commit is contained in:
Ben Skeggs
2017-06-05 17:23:32 +10:00
committed by Greg Kroah-Hartman
parent 74276868b4
commit 5dffc1be65
2 changed files with 5 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
drivers/gpu/drm/nouveau/include/nvkm/subdev/timer.h
View File

@@ -4,6 +4,7 @@
struct nvkm_alarm {
struct list_head head;
struct list_head exec;
u64 timestamp;
void (*func)(struct nvkm_alarm *);
};

7
drivers/gpu/drm/nouveau/nvkm/subdev/timer/base.c
View File

@@ -50,7 +50,8 @@ nvkm_timer_alarm_trigger(struct nvkm_timer *tmr)
/* Move to completed list. We'll drop the lock before
* executing the callback so it can reschedule itself.
*/
list_move_tail(&alarm->head, &exec);
list_del_init(&alarm->head);
list_add(&alarm->exec, &exec);
}
/* Shut down interrupt if no more pending alarms. */
@@ -59,8 +60,8 @@ nvkm_timer_alarm_trigger(struct nvkm_timer *tmr)
spin_unlock_irqrestore(&tmr->lock, flags);
/* Execute completed callbacks. */
list_for_each_entry_safe(alarm, atemp, &exec, head) {
list_del_init(&alarm->head);
list_for_each_entry_safe(alarm, atemp, &exec, exec) {
list_del(&alarm->exec);
alarm->func(alarm);
}
}