Thanks for the quick help. I built a kernel with your fix. The system is booting up with a warning, now.
... scsibus3 at softraid0: 256 targets sd2 at scsibus3 targ 1 lun 0: <OPENBSD, SR RAID 1, 006> sd2: 122103MB, 512 bytes/sector, 250067198 sectors root on sd2a (da12fadb67cf7a4d.a) swap on sd2b dump on sd2b drm : drm_WARN_ON(d->wake_count)drm : drm_WARN_ON(d->wake_count)Device initialization failed (-22) Automatic boot in progress: starting file system checks. /dev/sd2a (da12fadb67cf7a4d.a): file system is clean; not checking ... On 2021-05-04 12:20, Jonathan Gray wrote: > > the vma setup is somehow not occuring > > gen6_ppgtt_create ppgtt->vma = pd_vma_create(... > __ppgtt_create > i915_ppgtt_create > > this should help but isn't a proper fix > > Index: sys/dev/pci/drm/i915/i915_vma.c > =================================================================== > RCS file: /cvs/src/sys/dev/pci/drm/i915/i915_vma.c,v > retrieving revision 1.5 > diff -u -p -r1.5 i915_vma.c > --- sys/dev/pci/drm/i915/i915_vma.c 11 Oct 2020 05:45:33 -0000 1.5 > +++ sys/dev/pci/drm/i915/i915_vma.c 4 May 2021 10:13:37 -0000 > @@ -1029,8 +1029,13 @@ static void flush_idle_contexts(struct i > > int i915_ggtt_pin(struct i915_vma *vma, u32 align, unsigned int flags) > { > - struct i915_address_space *vm = vma->vm; > + struct i915_address_space *vm; > int err; > + > + if (vma == NULL) > + return -EINVAL; > + > + vm = vma->vm; > > GEM_BUG_ON(!i915_vma_is_ggtt(vma)); > >