Created attachment 104588 [details] Xorg.0.log with error messages Bug description: Graphics display frozen (mouse pointer still moves) after resume from hibernation. Xorg.0.log filled with: (EE) intel(0): Failed to submit batch buffer, expect rendering corruption or even a frozen display: Invalid argument. This is a regression, hiberate worked fine up to Linux kernel 3.11. System environment: -- chipset: GM45 -- system architecture: i686 -- xf86-video-intel: 2.19.0 -- xserver: 1.12.4 -- mesa: 2.1 Mesa 8.0.5 -- libdrm: 2.4.40 -- kernel: 3.12.23, 3.14.16, 3.16 -- Linux distribution: Debian Wheezy -- Machine or mobo model: Lenovo Thinkpad X200 Tablet -- Display connector: Integrated panel Reproducing steps: * Boot system * On login screen (I use lightdm, but it's also reproducible from within an Xfce session), click "Hibernate" * Wait a few minutes (10 min always seems to be enough, immediate resume sometimes doesn't show the problem) * Power on * Wait for resume from hibernation * Old screen shown again, but cursor not flashing and no interaction possible * Switch to VT1 using Ctrl-Alt-F1 * Inspect /var/log/Xorg.0.log Additional info: git-bisect showed that commit 17fec8a08698bcab98788e1e89f5b8e7502ababd (drm/i915: Use Graphics Base of Stolen Memory on all gen3+) from https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git causes the problem. Reverting to the previous commit f63a484c2f606b8267eb4d1dbfce5d1d3416e0bb fixes the problem.
Created attachment 104589 [details] dmesg output, probably not very interesting
CC'ing Chris to comment on the bisect result.
Ville's fixes are still outstanding.
(In reply to comment #3) > Ville's fixes are still outstanding. Do you have more information about when they might be ready? Should the commit that caused the problem be reverted in the meantime, or would that cause even more breakage?
(In reply to comment #3) > Ville's fixes are still outstanding. Which fixes are these?
(In reply to comment #5) > (In reply to comment #3) > > Ville's fixes are still outstanding. > > Which fixes are these? Stolen address fixes.
Is this still happening on latest drm-intel-nightly?
No longer reproducible with 2ea23cd593ba60ead60e2f796fae675aa4475b1a from drm-intel-nightly. Thanks!
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.