Bug 46078

Summary: [ivb] Call Trace in 3.2.4
Product: DRI Reporter: cc <chao.a.chen>
Component: DRM/IntelAssignee: Daniel Vetter <daniel>
Status: CLOSED FIXED QA Contact:
Severity: major    
Priority: high CC: ben, chris, daniel, jbarnes
Version: unspecifiedKeywords: regression
Hardware: All   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
call-error-dmesg.log none

Description cc 2012-02-14 19:22:07 UTC 
System Environment:
--------------------------
Arch:           i386
Platform:       IvyBridge
Libdrm:		(master)2.4.31-3-gced219ebbd3b266ac8326223bad62f994907ae6b
Mesa:		(8.0)65526d54aa2599b069bd443d3e6e9762e613042d
Xserver:	(server-1.11-branch)xorg-server-1.11.4
Xf86_video_intel: (master)2.17.0-650-gc0376b7f7b083ab2e87edc36e56fd8eb99c3cd05
Kernel:    3.2.4

Bug detailed description:
-------------------------
Call Trace:
 <IRQ>  [<ffffffff8103c33c>] warn_slowpath_common+0x80/0x98
 [<ffffffff8103c369>] warn_slowpath_null+0x15/0x17
 [<ffffffffa00947ad>] gen6_gt_force_wake_put+0x27/0x43 [i915]
 [<ffffffffa00948ef>] i915_read32+0x49/0x98 [i915]
 [<ffffffffa00c7026>] intel_ring_get_active_head+0x2b/0x2d [i915]
 [<ffffffffa00c704d>] gen6_ring_get_seqno+0x25/0x36 [i915]
 [<ffffffffa0098c88>] notify_ring+0x29/0xb3 [i915]
 [<ffffffffa009968c>] ivybridge_irq_handler+0x100/0x2c3 [i915]
 [<ffffffff8108a8f6>] handle_irq_event_percpu+0x5a/0x187
 [<ffffffff81042202>] ? __do_softirq+0x16c/0x182
 [<ffffffff8108aa55>] handle_irq_event+0x32/0x4e
 [<ffffffff8101a017>] ? ack_apic_edge+0x31/0x35
 [<ffffffff8108cfb2>] handle_edge_irq+0xcb/0xed
 [<ffffffff81003a1c>] handle_irq+0x8c/0x94
 [<ffffffff810031ef>] do_IRQ+0x48/0xaf
 [<ffffffff813b932b>] common_interrupt+0x6b/0x6b
 <EOI>  [<ffffffff813bf417>] ? system_call_after_swapgs+0x17/0x65
 [<ffffffff813bf598>] ? sysret_audit+0x16/0x20
---[ end trace aa2d938b7d8ba7eb ]---


Reproduce steps:
-----------------------
1.start x
2..opengl/xeglgears
Comment 1 cc 2012-02-14 22:38:05 UTC 
This demo (opengl/eglgears_x11) has the same issue too.

Additionally, I also tested the Alpha version and did not find this issue. So It's a regression issue.
Comment 2 libo 2012-02-14 23:56:33 UTC 
There are also such error messages at the same time when appear this bug:
"repeated kernel crash".
And it will make machine hang sometimes when run 3DMMES/uimark.
Comment 3 Gordon Jin 2012-02-15 00:43:32 UTC 
(In reply to comment #1)
> This demo (opengl/eglgears_x11) has the same issue too.
> Additionally, I also tested the Alpha version and did not find this issue. So It's a regression issue.

You mean 3.2 kernel by "Alpha version"?
Comment 4 cc 2012-02-15 00:49:53 UTC 
Yes. The alpha version is 3.2.0:

Kernel: (drm-intel-fixes)805a6af8dba5dfdd35ec35dc52ec0122400b2610
Some additional commit info:
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Wed Jan 4 15:55:44 2012 -0800
Comment 5 cc 2012-02-15 01:03:17 UTC 
Add some info maybe important.
1. Follow the lastest check. we found "Call Trace" happens as long as start gnome.
2. If we start X only and, run intel-gpu-tools, the "Call Trace" happens too.
Comment 6 Chris Wilson 2012-02-15 02:06:20 UTC 
There should be a few lines before the Call Trace that identify the error, can you please also always attach the full dmesg when quoting from it?
Comment 7 cc 2012-02-15 02:18:35 UTC 
Created attachment 57073 [details]
call-error-dmesg.log
Comment 8 cc 2012-02-15 02:21:05 UTC 
(In reply to comment #6)
> There should be a few lines before the Call Trace that identify the error, can you please also always attach the full dmesg when quoting from it?

This dmesg is just after start gnome without doing any other operate.
Comment 9 Daniel Vetter 2012-02-15 02:44:26 UTC 
This is the not-so-correct forcewake locking. This should be fixed in latest -fixes (or 3.3.-rc3, which contains it). Please reopen if I've misdiagnosed this and you still see this. Backports of these fixes are heading towards 3.2, but I'm not sure whether they've landed already.

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.