The diff below helps, and will be in snaps until the situation is resolved.
> On Fri, Jan 05, 2018 at 12:38:27AM +0200, Lauri Tirkkonen wrote: > > Hi, > > > > On Mon, Nov 27 2017 20:22:30 +1100, Jonathan Gray wrote: > > > I'm interested in reports from people who saw corruption on Intel > > > graphics during the brief period when Mesa 17.1.6 was in the tree. > > > > Sorry for not testing this sooner; I had forgotten about seeing this > > corruption previously. Updating to a snapshot just now (from OpenBSD > > 6.2-current (GENERIC.MP) #305: Thu Dec 21 14:53:41 MST 2017 to OpenBSD > > 6.2-current (GENERIC.MP) #314: Wed Jan 3 13:14:26 MST 2018) did > > result in corruption on my X220 and jogged my memory that I've seen this > > problem in the past... dmesg follows. > > I wonder if this is exposing an inteldrm problem. > > Does switching to the intel driver with xorg.conf or the below > diff change anything? > > Index: xserver/hw/xfree86/common/xf86pciBus.c > =================================================================== > RCS file: /cvs/xenocara/xserver/hw/xfree86/common/xf86pciBus.c,v > retrieving revision 1.15 > diff -u -p -r1.15 xf86pciBus.c > --- xserver/hw/xfree86/common/xf86pciBus.c 8 Dec 2017 15:02:00 -0000 > 1.15 > +++ xserver/hw/xfree86/common/xf86pciBus.c 4 Jan 2018 22:57:20 -0000 > @@ -1192,11 +1192,26 @@ xf86VideoPtrToDriverList(struct pci_devi > case 0xa011: > case 0x0042: /* XXX Ironlake is unstable with modesetting */ > case 0x0046: > - /* Use intel driver on 2nd and 3rd generation */ > + /* Sandy Bridge */ > + case 0x0102: > + case 0x0112: > + case 0x0122: > + case 0x010a: > + case 0x0106: > + case 0x0116: > + case 0x0126: > + /* Ivy Bridge */ > + case 0x0156: > + case 0x0166: > + case 0x0152: > + case 0x0162: > + case 0x015a: > + case 0x016a: > + /* Use intel driver until Ivy Bridge */ > driverList = "intel"; > break; > default: > - /* Use modesetting driver on 4th generation and up */ > + /* Use modesetting driver on Haswell and up */ > driverList = "modesetting"; > break; > } >