[Top] [All Lists]

Re: xfs and openafs problem

To: Willi.Langenberger@xxxxxxxxxxxxx
Subject: Re: xfs and openafs problem
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Tue, 23 Jul 2002 16:05:49 +0100
Cc: linux-xfs@xxxxxxxxxxx, linux-fsdevel@xxxxxxxxxxxxxxx
In-reply-to: <15677.27610.440858.682675@xxxxxxxxxxxxxxxxxxx>; from wlang@xxxxxxxxxxxxx on Tue, Jul 23, 2002 at 04:44:42PM +0200
References: <E49E46570933E44A9E4EE03B4B09363A853E@xxxxxxxxxxxxxxxxxxxxxxxx> <15677.25943.919190.769477@xxxxxxxxxxxxxxxxxxx> <20020723151936.A23121@xxxxxxxxxxxxx> <15677.27610.440858.682675@xxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
User-agent: Mutt/
On Tue, Jul 23, 2002 at 04:44:42PM +0200, Willi Langenberger wrote:
> According to Christoph Hellwig:
> > OpenAFS is broken.  No serious kernel deveoper (XFS or not) will care
> > for your reports if you have this piece of crap loaded.
> Have you any pointers?


> Anyway, what makes you sure, that it's OpenAFS' fault (besides that it
> is crap)?

lack of understanding and thus abuse of the Linux API.  Generally broken
code in all respects.

If you need specific examples look at how they patch the syscall table,
the way the ifdefs to support 64bit systems are organized or their struct
vnode that is smaller than a linux inode but casted to it all the time

One could also look at the way they hack into the Linux VM internals or
other broken abuses of the API.

Just DON't use OpenAFS if you want a sane machine.  (The same is true for
Transarc's propritary fork, btw - it's not better at all).

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