Date:   Tue, 21 Nov 2000 02:30:04 -0800 (PST)
Message-Id: <200011211030.CAA95634@feature.engr.sgi.com>
Reply-To: sgi.bugs.xfs@fido.engr.sgi.com
From:   pv@fddi-odin.corp.sgi.com (lord@sgi.com)
Subject: TAKE 808315 - Full file systems can become corrupted with XFS Linux.
To:     nb@sgi.com
Cc:     linux-xfs@oss.sgi.com
Sender: owner-linux-xfs@oss.sgi.com
Precedence: bulk

 Submitter : cattelan                 *Status : closed                      
 Assigned Engineer : nb               *Fixed By : lord                      
*Fixed By Domain : sgi.com            *Closed Date : 11/21/00               
 Priority : 2                         *Modified Date : 11/21/00             
*Modified User : lord                 *Modified User Domain : sgi.com       
*Fix Description :
From: steve lord <lord@sgi.com> (TAKE)
Date: Nov 21 2000 02:30:03AM
[pvnews version: 1.71]
----------------------------

Date:  Tue Nov 21 02:24:08 PST 2000
Workarea:  jen.americas.sgi.com:/src/lord/xfs-linux.2.4

The following file(s) were checked into:
  bonnie.engr.sgi.com:/isms/slinx/2.4.x-xfs


Modid:  2.4.x-xfs:slinx:78638a
linux/fs/pagebuf/page_buf_io.c - 1.36
	- Final version of the fix for the corruption problem, do the in place
	  extent conversion during the write call, but do not overwrite the bmap
	  being used by the top half of the write path.
Description :
Linux XFS may corrupt file system when disk si filled.

Very large write requests causes portions of the user data IO path
to continue writing beyond the end of an allocted extent thus over-writting
the start of the next AG. 

This usally show up as trashed secondary super blocks and probably
other data structures in the AG.