On 22/01/18 01:19 PM, Tom St Denis wrote: > On 22/01/18 09:03 AM, Tom St Denis wrote: >> I've rolled back quite a ways from the tip of drm-next (to >> 196f74897ba79f6d586894519f09796447d95be5) which is ~1700 commits back >> from the tip and still get the attached KASAN report every time. >> >> I'm running piglit with DRI_PRIME=1 with the max size memory tests >> disabled. >> >> Tom >> > > Should add I don't get the KASAN if I don't use DRI_PRIME=1. As soon as > I run it with DRI_PRIME=1 I get a KASAN error in a manner of seconds > (within the first 100 or so tests). I've found an interesting way of recreating this. Apparently it is triggered when a GL application goes full screen. For instance, DRI_PRIME=1 glxgears works fine but DRI_PRIME=1 glxgears -fullscreen will trigger the KASAN report. Tom