On Monday, August 19th, 2024 at 16:07, Sebastien Marie <sema...@kapouay.eu.org> 
wrote:

> 
> 
> Dannis 't Hart dannisth...@protonmail.ch writes:
> 
> > Hello,
> > It seems I'm not the only one having some trouble with Xorg/xenocara in 
> > recent snapshots. Perhaps there is a common cause.
> > 
> > On my laptop (Thinkpad T490S) I use mostly snapshots, troubles are seldom. 
> > The snapshot from August 13th worked fine, updating to the snapshot of 
> > August 17th seemed to break the startup of Xorg.
> > 
> > I tested also with the snapshots of August 16th and 18th, same problem. So 
> > I suspect some change between 13th and 16th must be the cause. So for now I 
> > use the snapshot of August 13th, to which I reverted back, which works just 
> > fine.
> > 
> > Xorg/xenodm doesn't come up. The screen stays black. The backlight of the 
> > laptop screen is on, though. And the cpu is kept quite busy, soon after the 
> > attempted start of X the fan is switched on and the temperature increase 
> > can easily be felt. After a few minutes, Xorg apparently segfaults and I 
> > just get to the regular command line.
> > 
> > I seems the segfault is at or after the loading of the DRI2 module. There 
> > is an earlier error (EE in log) about a missing 
> > /usr/X11R6/lib/modules/dri/i965_dri.so failed, which has been noted by 
> > someone else with the same problem of not being able to start X, but I 
> > don't think this is it (although this EE seems funny, I don't think above 
> > mentioned file is in base?).
> > 
> > I hope the cause can be identified and fixed!
> 
> 
> It should be already fixed (but the commit is recent, so no snapshots
> has it for now).
> 
> see sys/dev/pci/drm/i915/gem/i915_gem_mman.c rev 1.20
> https://cvsweb.openbsd.org/cgi-bin/cvsweb/src/sys/dev/pci/drm/i915/gem/i915_gem_mman.c
> 
> you could rebuild just the kernel if you can't wait for snapshot.
> 
> Thanks.
> --
> Sebastien Marie

Thank you, Sebastien! And pointing me to the fix by jsg.
I just did a cvs update on my /usr/src with the intention of building an 
utterly recent kernel, but noticed after a sysupgrade that the August 19th 
snapshot indeed already contains the fix.
Entirely happy now :-)

Thanks, kind regards,
Dannis.

Reply via email to