(In reply to Pierre Moreau from comment #10)

> What exactly is reproducible? The artefacts I would assume, but which error
> exactly? The NULL pointer dereference, or the “trapped read at 0080000000 on
> channel 1 [0fbb0000 DRM] engine 00 [PGRAPH] client 03 [DISPATCH] subclient
> 04 [M2M_IN] reason 00000006 [NULL_DMAOBJ]” one?
> I did manage to reproduce the second one with Wayland, but still no luck
> with the first one.

This error:
[  372.954548] nouveau 0000:03:00.0: imem: OOM: 00100000 00001000 -28
[  372.954702] nouveau 0000:03:00.0: gr: TRAP_M2MF 00000002 [IN]
[  372.954712] nouveau 0000:03:00.0: gr: TRAP_M2MF 00320951 c0001fc0 00000000
04000430
[  372.954722] nouveau 0000:03:00.0: gr: 00200000 [] ch 1 [000fbb0000 DRM] subc
4 class 5039 mthd 0100 data 00000000
[  372.954752] nouveau 0000:03:00.0: fb: trapped read at 00c0000000 on channel
1 [0fbb0000 DRM] engine 00 [PGRAPH] client 03 [DISPATCH] subclient 04 [M2M_IN]
reason 00000000 [PT_NOT_PRESENT]

kernel-4.15.6-300.vanilla.knurd.1.fc27.x86_64
mesa-17.3.6

Screenshot:
https://s9.postimg.org/hajle1skf/Screenshot_from_2018-03-01_10-12-28.png

> (In reply to Nick Lee from comment #8)
> > Fedora-Workstation-Live-x86_64-Rawhide-20180220.n.0.iso
> > wayland session
> > 
> > [ 1035.437016] nouveau 0000:03:00.0: swiotlb buffer is full (sz: 2097152
> 
> I’ll try the iso. Did you get this error with supertuxkart as well, or
> simply by launching the session?
> 

by launching the session