On Tue, 26 Aug 2014 03:54:53 +0200, Samuel Thibault said: > This changeset defers the second led_trigger_event call into a > workqueue, which avoids the nested locking altogether. This does > not prevent the user from shooting himself in the foot by creating a > vt::capsl <-> vt-capsl loop, but the only consequence is the workqueue > threads eating some CPU until the user breaks the loop, which is not too > bad. > > Signed-off-by: Samuel Thibault <samuel.thibault@xxxxxxxxxxxx> > > --- a/drivers/input/leds.c > +++ b/drivers/input/leds.c > @@ -100,13 +100,25 @@ static unsigned long vt_led_registered[B I admit having zero understanding of the code, but I can confirm that next-20140825 still throws the lockdep whine I was seeing, but the same tree with this patch on top of it boots without warning. Soo... Tested-By: Valdis Kletnieks <valdis.kletnieks@xxxxxx>
Attachment:
pgp_yd_HYmTYK.pgp
Description: PGP signature
