Bug 135171

Summary:	Access to /proc/bus/usb/devices hangs the kernel
Product:	[Fedora] Fedora	Reporter:	Alan Cox <alan>
Component:	kernel	Assignee:	Pete Zaitcev <zaitcev>
Status:	CLOSED RAWHIDE	QA Contact:	Brian Brock <bbrock>
Severity:	high	Docs Contact:
Priority:	medium
Version:	3	CC:	piskozub, wtogami
Target Milestone:	---
Target Release:	---
Hardware:	All
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2004-10-16 21:45:16 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:
Bug Blocks:	123268, 130887

Description Alan Cox 2004-10-09 18:13:13 UTC

Description of problem:

At any point after a USB timeout occurs (as they tend to do plugging
devices) the 
USB devices file access hang the system deadlocked on a semaphore.
This has horrible consequences

One test setup shows this during boot up so Anaconda hangs and you
cannot install.
A second shows it when the monitor (with USB hub) powers down. After
that anything touching the USB devices file hangs the box. 

Thirdly you can cause it to order as a user by inserting and quickly
removing a device, killing the hal layer and rendering the box
unusable without even having login access to it.


This bug appears to have arrived between 2.6.6 and 2.6.7 somewhere.

Comment 1 Alan Cox 2004-10-09 18:15:16 UTC

This also btw seems to cause some of the kudzu hangs people see.

Comment 2 Bill Nottingham 2004-10-11 14:43:44 UTC

Can you reproduce with .603, or even BK head?

Comment 3 Alan Cox 2004-10-11 14:49:57 UTC

Its present in current 2.6.9-rc snapshots too. Came in between 2.6.6
and 2.6.7 somewhere.

Comment 4 Alan Cox 2004-10-15 15:26:11 UTC

Done some more digging

I cannot reproduce this with 2.6.8.1 vanilla, only various RH builds
so it looks like something that is specific to our tree. WIll test
.603 for completeness

Comment 5 Alan Cox 2004-10-15 16:01:50 UTC

*** Bug 134026 has been marked as a duplicate of this bug. ***

Comment 6 Alan Cox 2004-10-15 16:04:37 UTC

Propogating block from dup

Comment 7 Alan Cox 2004-10-16 21:45:16 UTC

541 has the buggy change -  624 is ok