http://bugzilla.opensuse.org/show_bug.cgi?id=1059302 Bug ID: 1059302 Summary: Intel HD 616/kabylake crashes after hibernate, not after suspend-to-ram Classification: openSUSE Product: openSUSE Distribution Version: Leap 42.3 Hardware: x86-64 OS: openSUSE 42.3 Status: NEW Severity: Normal Priority: P5 - None Component: X.Org Assignee: xorg-maintainer-bugs@forge.provo.novell.com Reporter: steiner-reg@bio.ifi.lmu.de QA Contact: xorg-maintainer-bugs@forge.provo.novell.com Found By: --- Blocker: --- Created attachment 741046 --> http://bugzilla.opensuse.org/attachment.cgi?id=741046&action=edit /sys/class/drm/card0/error Hi, our laptop is Lenovo T570 with an i5-7200 CPU with HD 610 graphics: 00:02.0 VGA compatible controller: Intel Corporation Device 5916 (rev 02) It comes back without problems after suspend-to-ram, but fails after suspend-to-disk/hibernate. X comes back but immediately freezes and /var/log/messages shows: 2017-09-19T10:27:48.261833+02:00 helsinki kernel: [ 642.797969] [drm] GPU HANG: ecode 9:0:0xa43ff68b, in X [14823], reason: Hang on render ring, action: reset 2017-09-19T10:27:48.261866+02:00 helsinki kernel: [ 642.797975] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace. 2017-09-19T10:27:48.261871+02:00 helsinki kernel: [ 642.797976] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel 2017-09-19T10:27:48.261875+02:00 helsinki kernel: [ 642.797977] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. 2017-09-19T10:27:48.261878+02:00 helsinki kernel: [ 642.797978] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it. 2017-09-19T10:27:48.261882+02:00 helsinki kernel: [ 642.797980] [drm] GPU crash dump saved to /sys/class/drm/card0/error 2017-09-19T10:27:48.261885+02:00 helsinki kernel: [ 642.798092] drm/i915: Resetting chip after gpu hang 2017-09-19T10:27:48.261888+02:00 helsinki kernel: [ 642.798199] [drm] RC6 on 2017-09-19T10:27:48.273803+02:00 helsinki kernel: [ 642.811661] [drm] GuC firmware load skipped 2017-09-19T10:27:59.297801+02:00 helsinki kernel: [ 653.833722] drm/i915: Resetting chip after gpu hang 2017-09-19T10:27:59.297832+02:00 helsinki kernel: [ 653.833830] [drm] RC6 on 2017-09-19T10:27:59.309814+02:00 helsinki kernel: [ 653.848065] [drm] GuC firmware load skipped And after about 10 seconds, kdm restarts. dmesg output from boot and the bug situation: [ 3.708133] i915 0000:00:02.0: enabling device (0006 -> 0007) [ 3.722593] [drm] Finished loading i915/kbl_dmc_ver1_01.bin (v1.1) [ 3.742234] [drm] Initialized i915 1.6.0 20160919 for 0000:00:02.0 on minor 0 [ 5.595734] i915 0000:00:02.0: fb0: inteldrmfb frame buffer device [ 629.870249] [drm:gen9_set_dc_state [i915]] *ERROR* DC state mismatch (0x0 -> 0x2) [ 642.797977] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue. [ 642.798092] drm/i915: Resetting chip after gpu hang [ 653.833722] drm/i915: Resetting chip after gpu hang It works fine after deinstalling drm-kmp-default, but as far as I've read, the kabylake chips somehow need the newer code? We couldn't see any drawbacks without drm-kmp-default at the first glance. cu, Frank -- You are receiving this mail because: You are on the CC list for the bug.