Created attachment 117927 [details]
I get the following when load the i915 module on a broadwell-u:
[ 891.121322] [drm] Initialized drm 1.1.0 20060810
[ 891.127056] [drm] Memory usable by graphics device = 4096M
[ 891.127151] [drm] VT-d active for gfx access
[ 891.127251] [drm] Replacing VGA console driver
[ 891.127957] Console: switching to colour dummy device 80x25
[ 891.134284] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 891.134290] [drm] Driver supports precise vblank timestamp query.
[ 891.134435] vgaarb: device changed decodes: PCI:0000:00:02.0,olddecodes=io+mem,decodes=io+mem:owns=io+mem
[ 891.147350] ACPI: Video Device [GFX0] (multi-head: yes rom: no post: no)
[ 891.147690] input: Video Bus as /devices/LNXSYSTM:00/LNXSYBUS:00/PNP0A08:00/LNXVIDEO:00/input/input7
[ 891.147803] [drm] Initialized i915 1.6.0 20150327 for 0000:00:02.0 on minor 0
[ 891.185303] fbcon: inteldrmfb (fb0) is primary device
[ 891.313813] Console: switching to colour frame buffer device 240x67
[ 891.318478] i915 0000:00:02.0: fb0: inteldrmfb frame buffer device
[ 891.318504] i915 0000:00:02.0: registered panic notifier
[ 899.180411] [drm] stuck on render ring
[ 899.181241] [drm] GPU HANG: ecode 8:0:0x00201001, reason: Ring hung, action: reset
[ 899.181272] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[ 899.181307] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[ 899.181341] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[ 899.181376] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[ 899.181423] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[ 899.183561] drm/i915: Resetting chip after gpu hang
[ 905.179524] [drm] stuck on render ring
[ 905.180348] [drm] GPU HANG: ecode 8:0:0x00201001, reason: Ring hung, action: reset
[ 905.180435] [drm:i915_set_reset_status [i915]] *ERROR* gpu hanging too fast, banning!
[ 905.182602] drm/i915: Resetting chip after gpu hang
The error seems consistent with the other iommu-related snafu.
*** This bug has been marked as a duplicate of bug 89360 ***