Created attachment 130083 [details] /sys/class/drm/card0/error (gz) When activating intel_iommu=on GPU hangs can happen (here in GNOME Shell Wayland shortly after wakeup from suspend). With intel_iommu=on,igfx_off it seems to be fine. CPU: Intel Core i3-5010U CPU @ 2.1GHz GPU: Mesa DRI Intel(R) HD Graphics 5500 (Broadwell GT2) Linux 4.9.0-2-amd64 #1 SMP Debian 4.9.13-1 (2017-02-27) x86_64 GNU/Linux dmesg: [ 470.659847] DMAR: DRHD: handling fault status reg 3 [ 470.659854] DMAR: [DMA Write] Request device [00:02.0] fault addr fb149000 [fault reason 23] Unknown [ 480.871397] [drm] GPU HANG: ecode 8:0:0x85dffffb, in gnome-shell [2879], reason: Hang on render ring, action: reset [ 480.871398] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 480.871399] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 480.871399] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 480.871400] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [ 480.871400] [drm] GPU crash dump saved to /sys/class/drm/card0/error [ 480.876006] drm/i915: Resetting chip after gpu hang [ 492.838851] drm/i915: Resetting chip after gpu hang [ 502.823063] drm/i915: Resetting chip after gpu hang
When testing tboot on an Intel NUC5i5MYHE, graphics display nothing but a black screen after system boot up. There's also GPU hang found in kernel log: [ 68.752002] [drm] GPU HANG: ecode 8:0:0x85dffffb, in X [1616], reason: Hang on render ring, action: reset [ 68.752005] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. [ 68.752007] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel [ 68.752008] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 68.752008] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. [ 68.752009] [drm] GPU crash dump saved to /sys/class/drm/card0/error Such GPU hang is seen with RHEL 7.4 SP1/SLES12 SP3 Beta3/Ubuntu 17.04. Attached RHEL 7.4 kernel log and error log from /sys/class/drm/card0/error. After adding "intel_iommu=on,igfx_off" to boot cmdline, no GPU hang found and display shows desktop normally after system boot up.
Created attachment 131761 [details] gpu hang kernel log of RHEL 7.4 SP1 with tboot
Created attachment 131762 [details] /sys/class/drm/card0/error from RHEL7.4 SP1 with tboot
Changing this bug to duplicate of 100203 *** This bug has been marked as a duplicate of bug 100203 ***
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.