On Mon, Jun 13, 2016 at 11:33 AM, Daniel Vetter <daniel@xxxxxxxx> wrote: > On Mon, Jun 13, 2016 at 11:19:20AM -0400, Alex Deucher wrote: >> On Thu, Jun 9, 2016 at 3:29 PM, Rob Clark <robdclark@xxxxxxxxx> wrote: >> > There were a couple messed up things about this fail path. >> > (1) it would drop object_name_lock twice >> > (2) drm_gem_handle_delete() (in drm_gem_remove_prime_handles()) >> > needs to grab prime_lock >> > >> > Reported-by: Alex Deucher <alexdeucher@xxxxxxxxx> >> > Signed-off-by: Rob Clark <robdclark@xxxxxxxxx> >> > --- >> > Untested, but I think this should fix the potential deadlock that Alex >> > reported. Would be nice if someone with setup to reproduce could test >> > this. >> >> Sorry for the confusion. There were actually two deadlocks. The >> first one (https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1579610 >> fixed by https://github.com/torvalds/linux/commit/6984128d01cf935820a0563f3a00c6623ba58109) >> was that one we hit in testing. This one was just one that Qiang >> noticed by inspection while debugging the first. That said, the error >> path is obviously wrong and the patch looks correct to me. >> >> Reviewed-by: Alex Deucher <alexander.deucher@xxxxxxx> > > Ok, I didn't add the bugzilla to this one. Applied to drm-misc, thanks. There actually is a bugzilla for the potential deadlock fixed by this patch: https://bugzilla.kernel.org/show_bug.cgi?id=117861 Alex _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
