Bug 718232

Summary:	[xfs] mis-sized O_DIRECT I/O results in hung task timeouts
Product:	Red Hat Enterprise Linux 5	Reporter:	Jeff Moyer <jmoyer>
Component:	kernel	Assignee:	Jeff Moyer <jmoyer>
Status:	CLOSED ERRATA	QA Contact:	Petr Beňas <pbenas>
Severity:	urgent	Docs Contact:
Priority:	urgent
Version:	5.6	CC:	andrei, branto, bstein, cww, dchinner, dhoward, eguan, esandeen, jmoyer, mgahagan, pbenas, pstehlik, rwheeler, syeghiay
Target Milestone:	rc	Keywords:	ZStream
Target Release:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Fixed In Version:		Doc Type:	Bug Fix
Doc Text:	A problem with the XFS dio error handling was discovered. If a misaligned write I/O operation was issued, XFS would return -EINVAL without unlocking the inode's mutex. This caused any further operations on the inode to become unresponsive. This update adds a missing mutex_unlock operation to the dio error path, solving this issue.	Story Points:	---
Clone Of:	716991	Environment:
Last Closed:	2012-02-21 03:42:48 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:	695827, 716991
Bug Blocks:	727590

Comment 1 RHEL Program Management 2011-07-01 14:19:49 UTC

This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux maintenance release.  Product Management has requested
further review of this request by Red Hat Engineering, for potential
inclusion in a Red Hat Enterprise Linux Update release for currently deployed
products.  This request is not yet committed for inclusion in an Update
release.

Comment 12 Jarod Wilson 2011-08-23 14:04:55 UTC

Patch(es) available in kernel-2.6.18-282.el5
You can download this test kernel (or newer) from http://people.redhat.com/jwilson/el5
Detailed testing feedback is always welcomed.

Comment 13 Andrei Maslennikov 2011-08-31 16:35:52 UTC

I had an XFS-related issue 00511623 with RedHat which I just closed after trying the test kernel (-283.el5xen) as suggested by Jared. The error I was observing was 100% reproducible with kernel (-274.el5xen). To repeat it just reboot into kernel -274.el5xen, then place /usr/src/redhat into an XFS filesystem and try to rebuild the coreutils-5.97-34 package with "rpmbuild --rebuild coreutils-5.97-34.el5.src.rpm". When the build arrives to the post-compile tests, it would hang on the "shred" test. The same build ends up correctly when /usr/src/redhat is inside ext3. 

This does not happen with "-238.19.1.el5xen" kernel, and the problem was apparently taken care of with XFS-related patches in "-283.el5xen".

Andrei.

Comment 14 Martin Prpič 2011-09-08 14:59:00 UTC

    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
A problem with the XFS dio error handling was discovered. If a misaligned write I/O operation was issued, XFS would return -EINVAL without unlocking the inode's mutex. This caused any further operations on the inode to become unresponsive. This update adds a missing mutex_unlock operation to the dio error path, solving this issue.

Comment 16 Petr Beňas 2011-10-17 07:36:40 UTC

Reproduced in 2.6.18-275.el5 and verified in 2.6.18-276.el5.

Comment 17 errata-xmlrpc 2012-02-21 03:42:48 UTC

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory, and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

http://rhn.redhat.com/errata/RHSA-2012-0150.html