Bug 591709

Summary:	udev takes almost 100% CPU due to xorg (intel) continuously re-initializing displays
Product:	Red Hat Enterprise Linux 6	Reporter:	Matěj Cepl <mcepl>
Component:	kernel	Assignee:	Adam Jackson <ajax>
Status:	CLOSED CURRENTRELEASE	QA Contact:	desktop-bugs <desktop-bugs>
Severity:	high	Docs Contact:
Priority:	medium
Version:	6.0	CC:	airlied, cmeadors, gbarros, jlaska, jrb, kzhang, lkocman, sergio, tburke, xgl-maint
Target Milestone:	rc	Keywords:	Patch, Reopened, Triaged
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:	card_GM45
Fixed In Version:	2.6.32-52.el6	Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:	udev-intel-flood	Environment:
Last Closed:	2010-11-15 14:17:05 UTC	Type:	---
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:
Bug Depends On:	528312
Bug Blocks:

Comment 2 RHEL Program Management 2010-05-12 23:18:57 UTC

This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux major release.  Product Management has requested further
review of this request by Red Hat Engineering, for potential inclusion in a Red
Hat Enterprise Linux Major release.  This request is not yet committed for
inclusion.

Comment 3 Adam Jackson 2010-05-24 14:18:18 UTC

So the deal here is that we're enabling output hotplug interrupts on more outputs than we perhaps should, because Intel hardware doesn't give us an especially reliable connector table.  So we get status change events for outputs that aren't actually there.

There may also be a secondary bug where we _keep_ getting these events, triggered by the output rescan itself.

Getting the hardware would be an important first step here.

Comment 4 Adam Jackson 2010-05-24 16:58:25 UTC

alternatively, anyone who can reproduce this, please boot with drm.debug=0x0f and attach dmesg from the resulting udev storm.

Comment 5 Adam Jackson 2010-07-01 14:28:26 UTC

There's been some motion on this upstream, some patches that resolve it for some people but not for others.  I've got a couple in mind that I can backport, but it's hard to say whether they'll fix the issue completely.

Comment 7 Matěj Cepl 2010-07-15 09:37:34 UTC

Just to emphasize I cannot reproduce this bug anymore, I haven't seen this for couple of months with RHEL-6. However, given its heavy hardware dependency I wouldn't close this as fixed.

Comment 9 Christopher Aillon 2010-07-15 20:51:47 UTC

Closing based on reporter's comments in comment 7.  Reopen if still exists.

Comment 10 Adam Jackson 2010-07-16 15:27:52 UTC

Reopening, I still have changes pending for this.

Comment 11 Adam Jackson 2010-07-16 19:02:29 UTC

Patch posted

Comment 12 Cameron Meadors 2010-07-19 19:07:55 UTC

What hardware is actually affected by this?  I only found mention of the MSI-GT725 in the fedora.  Are there other laptops that exhibit this bug? Is this maybe somthing that only depends on what graphics card you have?  or is it the whole package and how the graphics is integrated into the system?

Comment 13 Dave Airlie 2010-07-21 01:32:57 UTC

its a whole package type of thing, though I think any of the T500/W500 should be able to reproduce. I think I reproduced on a W500 at one point.

Comment 16 Aristeu Rozanski 2010-07-26 15:16:43 UTC

Patch(es) available on kernel-2.6.32-52.el6

Comment 21 James Laska 2010-09-01 18:44:14 UTC

I tested this bug on a t500 borrowed from helpdesk.  I'm unable to reproduce any udev issues when coming out of suspend.  Both suspend+resume work beautifully on this laptop.  Considering I wasn't able to reproduce the reported problem, I'm marking this issue as verified:SanityOnly.

Comment 22 releng-rhel@redhat.com 2010-11-15 14:17:05 UTC

Red Hat Enterprise Linux 6.0 is now available and should resolve
the problem described in this bug report. This report is therefore being closed
with a resolution of CURRENTRELEASE. You may reopen this bug report if the
solution does not work for you.