[Top] [All Lists]

Re: vfs_altfsid & dm_fsid

To: linux-xfs@xxxxxxxxxxx
Subject: Re: vfs_altfsid & dm_fsid
From: Aurelien Degremont - Stagiaire <degremont@xxxxxxxxxxx>
Date: Fri, 13 May 2005 17:54:54 +0200
Cc: Dean Roehrich <roehrich@xxxxxxx>
In-reply-to: <42829491.2080502@sgi.com>
References: <20050428021133.A885C4FE57@chewtoy.americas.sgi.com> <4278CDB8.3060309@ocre.cea.fr> <42829491.2080502@sgi.com>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; fr-FR; rv:1.2.1) Gecko/20030225
Let's see if I can get the attachment right....

There should be an attachment here that has a draft of the fsid
changes.  This is lightly tested.

This looks ok but as you said, it need testing, i can do some of them, but i haven't a huge HSM system with dmapi support to be sure all is ok.

The only problem i'm concerned is about the dm_send_mount_event code.
(It's not a very important problem, just details :-))
You're currently adding more and more stuff to this function. Particularly concerning the filesystem registration code (dmapi_register() removal). I don't like having a function called "Send a mount message" like many others (send_namesp_event, ...) doing a job it isn't intended to do.
I spent some time to understand the dm_send_mount_event() and dm_send_unmount_event() are the functions that registered and unregistered the filesystem.
Moreover, it all these changes are done, the dm_send_mount_event() will have a lot of parameters, and I don't think a code with functions with 11 or 12 parameter is a good code. Maybe something cleaner could be done here...

'dm_register_and_send_mount_event' (like dm_find_fsreg_and_lock ? :))
Not really short but at least, very clear :).

It is seems a good idea to register the filesystem in the same time it is mounted, but this implies a huge and ugly dmapi send mount event function ;)


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