| Summary: | crash with xorg-server-1.16.4 - xf86-video-intel-2.99.917 and kernel 3.10 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | DRI | Reporter: | Agostino Sarubbo <ago> | ||||||||
| Component: | DRM/Intel | Assignee: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||||
| Status: | CLOSED FIXED | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||||
| Severity: | critical | ||||||||||
| Priority: | high | CC: | intel-gfx-bugs | ||||||||
| Version: | XOrg git | ||||||||||
| Hardware: | Other | ||||||||||
| OS: | All | ||||||||||
| Whiteboard: | |||||||||||
| i915 platform: | i915 features: | ||||||||||
| Attachments: |
|
||||||||||
|
Description
Agostino Sarubbo
2015-03-03 08:29:46 UTC
Xorg.0.log? If possible, an Xorg.0.log with --enable-debug=full would be very useful (USE=full-debug) And on the other side, a drm.debug=7 dmesg (capturing the error message) would also be useful. Created attachment 114313 [details] Xorg.0.log (In reply to Chris Wilson from comment #1) > Xorg.0.log? That's it. (In reply to Chris Wilson from comment #2) > If possible, an Xorg.0.log with --enable-debug=full would be very useful > (USE=full-debug) Sorry, but if I compile with debug=full, xorg does not start and the machine hangs. I don't have a shell and I'm able to poweroff just because of acpid. Created attachment 114314 [details] drm.debug=7 dmesg (In reply to Chris Wilson from comment #3) > And on the other side, a drm.debug=7 dmesg (capturing the error message) > would also be useful. that's it. Now we have a slightly different issue. The GPU hangs very early. Can you please attach /sys/class/drm/card0/error? (In reply to Chris Wilson from comment #6) > Now we have a slightly different issue. The GPU hangs very early. Can you > please attach /sys/class/drm/card0/error? There is no file called error after the hang. BTW, do you guess that update to a newer kernel would solve the problem? After the hang, before the next reboot. In the latest Xorg/dmesg attached, there should be an error state. If it is not there, you have multiple bugs and we may need to diagnose each separately (i.e. attach Xorg/dmesg for each different failure mode). (In reply to Chris Wilson from comment #8) > After the hang, before the next reboot. I know, but there isn't From dmesg I see that it points to /sys/kernel/debug/dri/0/i915_error_state instead of something else.. Ok, your kernel is that old! Please attach /sys/kernel/debug/dri/0/i915_error_state then! Created attachment 114379 [details]
i915_error_state
I confirm that the problem is just the kernel 3.10. I updated to 3.18 and all work as expected. Do you guess this bug should remains as NEEDINFO? The relocation routine in the kernel failed, so the first BLT overwrote the CS data structures and not unsurprisingly the GPU died. If you really, really want to stick with 3.10, you need to work with the gentoo kernel team to see which patch they are missing. But as for upstream, we can mark this as resolved. (In reply to Chris Wilson from comment #13) > If you really, really want to stick with 3.10, you need to work with the > gentoo kernel team to see which patch they are missing. But as for upstream, > we can mark this as resolved. The problem is not gentoo here. We are using the vanilla sources plus some patches. So the fact is that xorg-server-1.16.4 and xf86-video-intel-2.99.917 does not work with the 3.10 upstream kernel which is an LTS. |
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.