NetBSD Problem Report #35704

From www@NetBSD.org  Mon Feb 19 11:00:50 2007
Return-Path: <www@NetBSD.org>
Received: by narn.NetBSD.org (Postfix, from userid 31301)
	id B4AB263B9FB; Mon, 19 Feb 2007 11:00:50 +0000 (UTC)
Message-Id: <20070219110050.B4AB263B9FB@narn.NetBSD.org>
Date: Mon, 19 Feb 2007 11:00:50 +0000 (UTC)
From: p.jordan@teles.de
Reply-To: p.jordan@teles.de
To: gnats-bugs@NetBSD.org
Subject: problems in a full ffs file system
X-Send-Pr-Version: www-1.0

>Number:         35704
>Category:       kern
>Synopsis:       problems in a full ffs file system
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Feb 19 11:05:00 +0000 2007
>Closed-Date:    Tue Jun 29 15:49:47 +0000 2010
>Last-Modified:  Tue Jun 29 15:49:47 +0000 2010
>Originator:     Peter
>Release:        NetBSD 3.0
>Organization:
>Environment:
System: NetBSD performance_c1 3.0 NetBSD 3.0 (MYKERNEL) #1: Fri Feb 16 14:00:57 2007 admin@NETBSD:/usr/NetBSD-3.0/cross/i386/obj/sys/arch/i386/compile/MYKERNEL
i386
>Description:
Hello,

we have a Problem in NetBSD 3.0 and ffs filesystem.
If we fill as a root user a ffs file system until no space is available,
we get corrupted files.
        /home: write failed, file system is full
        mount:
        /dev/wd0g     970M     970M      -48M   105%    /home
After that, the root user is able to create files like:
        echo 'aaaaaaaaa' > ffs_test
this file (ffs_test) is ok.
After making this file bigger with:
        cat ffs_test >> ffs_test
this file (ffs_test) is corrupt.
This error occurs if we mounted this filesystem in softdep-mode and
default-mode. With mounting this filesystem in synchronous mode,
this error DOES NOT occure. Also by filling the filesystem as a non
root user, who is not able to write more then 100 %,
the error DOES NOT occure.
On other systems we get even coredumps after panic
and some other problems occured (e.g. binaries did not start anymore).

>How-To-Repeat:
* login as root user
* make your filesystem full until:
       /home: write failed, file system is full
* create a new file:
       echo 'aaaaaaaaa' > ffs_test
* make this file bigger with:
       cat ffs_test >> ffs_test

>Fix:

>Release-Note:

>Audit-Trail:
From: Manuel Bouyer <bouyer@antioche.eu.org>
To: gnats-bugs@NetBSD.org
Cc: kern-bug-people@NetBSD.org, gnats-admin@NetBSD.org, netbsd-bugs@NetBSD.org
Subject: Re: kern/35704: problems in a full ffs file system
Date: Tue, 2 Feb 2010 23:01:52 +0100

 On Mon, Feb 19, 2007 at 11:05:01AM +0000, p.jordan@teles.de wrote:
 > >How-To-Repeat:
 > * login as root user
 > * make your filesystem full until:
 >        /home: write failed, file system is full
 > * create a new file:
 >        echo 'aaaaaaaaa' > ffs_test
 > * make this file bigger with:
 >        cat ffs_test >> ffs_test

 Still reproductible on netbsd-5.
 The patch proposed in
 http://mail-index.netbsd.org/tech-kern/2010/02/02/msg007156.html
 fixes it for me ...

 -- 
 Manuel Bouyer <bouyer@antioche.eu.org>
      NetBSD: 26 ans d'experience feront toujours la difference
 --

From: Manuel Bouyer <bouyer@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/35704 CVS commit: src/sys/ufs
Date: Sun, 7 Feb 2010 17:12:40 +0000

 Module Name:	src
 Committed By:	bouyer
 Date:		Sun Feb  7 17:12:40 UTC 2010

 Modified Files:
 	src/sys/ufs/ext2fs: ext2fs_inode.c
 	src/sys/ufs/ffs: ffs_inode.c
 	src/sys/ufs/lfs: lfs_inode.c
 	src/sys/ufs/ufs: ufs_inode.c

 Log Message:
 - ufs_balloc_range(): on error, only PG_RELEASED the pages that were
   allocated to extend the file to the new size. Releasing all pages
   may release pages that contains previously-written data not yet flushed
   to disk. Should fix PR kern/35704
 - {ffs,lfs,ext2fs}_truncate(): Even if the inode's size is the same as
   the new length, call uvm_vnp_setsize(). *_truncate() may have been
   called by *_write() in the error path (e.g. block allocation failure
   because of quota of file system full), and at this point v_writesize
   has been set to the desired size of the file and not reverted to the
   old size. Not adjusting v_writesize to the real size cause
   genfs_do_io() to write to disk past the real end of the file.


 To generate a diff of this commit:
 cvs rdiff -u -r1.70 -r1.71 src/sys/ufs/ext2fs/ext2fs_inode.c
 cvs rdiff -u -r1.103 -r1.104 src/sys/ufs/ffs/ffs_inode.c
 cvs rdiff -u -r1.120 -r1.121 src/sys/ufs/lfs/lfs_inode.c
 cvs rdiff -u -r1.78 -r1.79 src/sys/ufs/ufs/ufs_inode.c

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

From: Soren Jacobsen <snj@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/35704 CVS commit: [netbsd-5] src/sys/ufs
Date: Mon, 22 Feb 2010 04:43:47 +0000

 Module Name:	src
 Committed By:	snj
 Date:		Mon Feb 22 04:43:47 UTC 2010

 Modified Files:
 	src/sys/ufs/ext2fs [netbsd-5]: ext2fs_inode.c
 	src/sys/ufs/ffs [netbsd-5]: ffs_inode.c
 	src/sys/ufs/lfs [netbsd-5]: lfs_inode.c
 	src/sys/ufs/ufs [netbsd-5]: ufs_inode.c

 Log Message:
 Pull up following revision(s) (requested by bouyer in ticket #1302):
 	sys/ufs/ext2fs/ext2fs_inode.c: revision 1.71
 	sys/ufs/ffs/ffs_inode.c: revision 1.104
 	sys/ufs/lfs/lfs_inode.c: revision 1.121
 	sys/ufs/ufs/ufs_inode.c: revision 1.79
 - ufs_balloc_range(): on error, only PG_RELEASED the pages that were
   allocated to extend the file to the new size. Releasing all pages
   may release pages that contains previously-written data not yet flushed
   to disk. Should fix PR kern/35704
 - {ffs,lfs,ext2fs}_truncate(): Even if the inode's size is the same as
   the new length, call uvm_vnp_setsize(). *_truncate() may have been
   called by *_write() in the error path (e.g. block allocation failure
   because of quota of file system full), and at this point v_writesize
   has been set to the desired size of the file and not reverted to the
   old size. Not adjusting v_writesize to the real size cause
   genfs_do_io() to write to disk past the real end of the file.


 To generate a diff of this commit:
 cvs rdiff -u -r1.66 -r1.66.8.1 src/sys/ufs/ext2fs/ext2fs_inode.c
 cvs rdiff -u -r1.99 -r1.99.4.1 src/sys/ufs/ffs/ffs_inode.c
 cvs rdiff -u -r1.120 -r1.120.10.1 src/sys/ufs/lfs/lfs_inode.c
 cvs rdiff -u -r1.76.4.1 -r1.76.4.2 src/sys/ufs/ufs/ufs_inode.c

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

State-Changed-From-To: open->feedback
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Sun, 28 Feb 2010 23:39:58 +0000
State-Changed-Why:
This should be fixed in 5.0_STABLE and in HEAD; can you confirm?


State-Changed-From-To: feedback->closed
State-Changed-By: pooka@NetBSD.org
State-Changed-When: Tue, 29 Jun 2010 18:49:47 +0300
State-Changed-Why:
feedback timeout


>Unformatted:

Home	PR Database Search