xfs
[Top] [All Lists]

Re: am-utils xfs PATCH

To: Nathan Scott <nathans@xxxxxxxxxxxxxxxxxxxxxxxx>
Subject: Re: am-utils xfs PATCH
From: Chris Bednar <cjb@xxxxxxxxxxxxxxxxxxxxxxxxx>
Date: Wed, 18 Jul 2001 00:53:53 -0500 (CDT)
Cc: linux-xfs@xxxxxxxxxxx, amd-dev@xxxxxxxxxxxxxxx
In-reply-to: <10107181102.ZM195527@wobbly.melbourne.sgi.com>
Sender: owner-linux-xfs@xxxxxxxxxxx
On Wed, 18 Jul 2001, Nathan Scott wrote:

> This doesn't look quite right - the xfs_args struct in IRIX
> is used to pass the mount arguments across the mount syscall
> interface (as binary data), but in Linux each filesystem's
> options are passed in as a comma separated list (a string of
> characters).
> 
> xfs_da_args_t is something unrelated to mount(2), and is only
> for use within the kernel and libxfs.  In general, you don't
> want to be #include'ing files from /usr/include/xfs/* at all.
> 
> The right way to implement this would be to figure out how the
> other Linux filesystems are handled and do XFS (Linux/XFS, that
> is) the same way, then make the existing amd code for XFS only
> be used on IRIX.  I've only had a quick look at the am-utils
> code just now, so can't really give you more advice than that,
> I'm afraid.

    Yeah, it was really just a stab in the dark, to get it 
to compile on my ppc system. I have to confess, I only use
amd for NFS and local symlink mounts, so I don't need any
other FS-specific support. I ran out of energy trying to
figure out how to do it right. ``Right'', in my case, would
be to figure out how to disable XFS in the config scripts.


----
Chris J. Bednar   <http://optics.tamu.edu/~bednar/>
Director, Distributed Computing Product Group
http://AdvancedDataSolutions.com/


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