| Summary: | [i945][KMS] Intel driver assert failure cause X to restart during resume | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Product: | xorg | Reporter: | Jie Luo <clotho67> | ||||||
| Component: | Driver/intel | Assignee: | Jesse Barnes <jbarnes> | ||||||
| Status: | RESOLVED FIXED | QA Contact: | Xorg Project Team <xorg-team> | ||||||
| Severity: | critical | ||||||||
| Priority: | medium | Keywords: | NEEDINFO | ||||||
| Version: | 7.4 (2008.09) | ||||||||
| Hardware: | x86 (IA32) | ||||||||
| OS: | Linux (All) | ||||||||
| Whiteboard: | |||||||||
| i915 platform: | i915 features: | ||||||||
| Attachments: |
|
||||||||
Created attachment 25248 [details]
gdm.0.log
This time pI830->batch_ptr is NULL when doing i830_uxa_copy.
X: i830_exa.c:377: i830_uxa_copy: Assertion `pI830->batch_ptr != ((void *)0)' failed.
So the problem should be pI830->batch_ptr or dri_bo_map() which initialized pI830->batch_ptr.
A possible way to reproduce this bug. Suspend to memory for a whole day and then resume, it seams that each time I hit this bug is after a almost whole day suspend. Adjusting severity: crashes & hangs should be marked critical. Do you still see this with current bits? AFAICT, the only way this could happen is if the intel_batch_init at EnterVT time fails, since it's responsible for allocating & mapping the new batch object (in intel_next_batch). It might help if you could get a backtrace from the failure... Ah the old rendering while VT switched bug, should be fixed: commit e54a23bff068416ccbdb75d538dc7dcd40a6c95c Author: Keith Packard <keithp@keithp.com> Date: Thu May 7 16:35:19 2009 -0700 Fallback when VT inactive Oh that's a commit to xf86-video-intel btw. (In reply to comment #6) > Oh that's a commit to xf86-video-intel btw. > Yes, this bug should be fixed. I can't reproduce it any more. Thanks very much. |
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.
Created attachment 25020 [details] Xorg.0.log When resume from memory, some times X will be reseted with a assert failed message in gdm log. X: i830_batchbuffer.h:78: intel_batch_emit_dword: Assertion `pI830->batch_ptr != ((void *)0)' failed. Since intel_batch_emit_dword is only called from OUT_BATCH macro, this is not really show the problem. So I add another assertion before calling intel_batch_emit_dword in OUT_BATCH. I wait the X to be reset again during resume, and then I get this. X: i830_exa.c:387: I830EXACopy: Assertion `pI830->batch_ptr != ((void *)0)' failed. And this bug didn't have a reliable way to reproduce, it only occurs randomly during resume. System environment: -- chipset: 945GM -- system architecture: x86 -- xf86-video-intel/xserver/mesa/drm version: xf86-video-intel: git commit 2e3b95ed0197971e81ab7509245c899e96859d5b xserver: 1.6.1 mesa: 7.4 libdrm: git commit -- kernel version: 2.6.30-rc2 -- Linux distribution: Gentoo -- Machine model: ThinkPad X60