xfs
[Top] [All Lists]

Re: 2.4.24-pre1 hangs with XFS on LVM filesystem full

To: Eric Sandeen <sandeen@xxxxxxx>, Emilio Gargiulo <emiliogargiulo@xxxxxxxxxxx>
Subject: Re: 2.4.24-pre1 hangs with XFS on LVM filesystem full
From: Nathan Scott <nathans@xxxxxxx>
Date: Tue, 16 Dec 2003 12:21:14 +1100
Cc: linux-kernel@xxxxxxxxxxxxxxx, linux-xfs@xxxxxxxxxxx
In-reply-to: <200312152200.08411.emiliogargiulo@xxxxxxxxxxx>
References: <Pine.LNX.4.44.0312132302520.766-100000@xxxxxxxxxxxxxxxxxxxxxx> <200312152200.08411.emiliogargiulo@xxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.5.3i
On Mon, Dec 15, 2003 at 10:00:08PM +0100, Emilio Gargiulo wrote:
> Yes, a simple overfill of a XFS filesystem on LVM hangs the kernel.
> It is easy to reproduce with dd if=/dev/zero of=file_on_dest_xfs_filesystem 
> bs=1024k.
> ...
> I compiled the kernel with gcc 3.2.3. (Slackware 9.1)
> I am unable to reproduce the problem if I compile the kernel with gcc 3.3.1.

Hmm, that would seem to be an important piece of information -
could be a gcc-3.2.3 compiler bug.  Since noone here is able to
trigger this problem with other compilers too (and under a lot
more varied conditions around the ENOSPC boundaries), I suppose
a gcc bug is a real possibility here.

cheers.

-- 
Nathan


<Prev in Thread] Current Thread [Next in Thread>