xfs-masters
[Top] [All Lists]

[xfs-masters] Re: 2.6.19-rc6 : Spontaneous reboots, stack overflows - se

To: David Miller <davem@xxxxxxxxxxxxx>
Subject: [xfs-masters] Re: 2.6.19-rc6 : Spontaneous reboots, stack overflows - seems to implicate xfs, scsi, networking, SMP
From: Al Viro <viro@xxxxxxxxxxxxxxxx>
Date: Thu, 23 Nov 2006 04:35:43 +0000
Cc: dgc@xxxxxxx, jesper.juhl@xxxxxxxxx, chatz@xxxxxxxxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx, xfs-masters@xxxxxxxxxxx, netdev@xxxxxxxxxxxxxxx, linux-scsi@xxxxxxxxxxxxxxx
In-reply-to: <20061122.201013.112290046.davem@xxxxxxxxxxxxx>
References: <9a8748490611211551v2ebe88fel2bcf25af004c338a@xxxxxxxxxxxxxx> <9a8748490611220458w4d94d953v21f7a29a9f1bdb72@xxxxxxxxxxxxxx> <20061123011809.GY37654165@xxxxxxxxxxxxxxxxx> <20061122.201013.112290046.davem@xxxxxxxxxxxxx>
Reply-to: xfs-masters@xxxxxxxxxxx
Sender: xfs-masters-bounce@xxxxxxxxxxx
User-agent: Mutt/1.4.1i
On Wed, Nov 22, 2006 at 08:10:13PM -0800, David Miller wrote:
> From: David Chinner <dgc@xxxxxxx>
> Date: Thu, 23 Nov 2006 12:18:09 +1100
> 
> > So, assuming the stacks less than 32 bytes are 32 bytes, we've got
> > 1380 bytes in the XFS stack there, 
> 
> On sparc64 just the XFS parts of the backtrace would be a minimum of
> 2816 bytes (each function has a minimum 8 * 16 byte stack frame, and
> there are about 22 calls in that trace).  It's probably a lot more
> with local variables and such.
> 
> It's way too much.  You guys have to fix this stuff.
> 
> If TCP's full send and receive path can be done in less function
> calls, XFS can allocate blocks in less too.
> 
> I would even say 10 function calls deep to allocate file blocks
> is overkill, but 22 it just astronomically bad.

Especially since a large part is due to cxfs...


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