[Top] [All Lists]

Re: Tomorrow

To: linux-xfs@xxxxxxxxxxx
Subject: Re: Tomorrow
From: Ethan Benson <erbenson@xxxxxxxxxx>
Date: Tue, 27 May 2003 22:49:12 -0800
In-reply-to: <1054048771.1939.73.camel@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Mail-copies-to: nobody
Mail-followup-to: linux-xfs@xxxxxxxxxxx
References: <1053694002.2887.1.camel@xxxxxxxxxxxxxxxxxxxxx> <1053697162.21472.51.camel@xxxxxxxxxxxxxxxxxxxx> <20030523134438.GC30288@xxxxxxxxxxxxx> <20030523150530.A31022@xxxxxxxxxxxxx> <20030524071709.GK27626@xxxxxxxxxxxxxxx> <20030524095245.A24074@xxxxxxxxxxxxx> <Pine.LNX.4.53.0305270026390.17955@xxxxxxxxxxxxxxxxxxxxxxx> <1054048771.1939.73.camel@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mutt/1.3.28i
On Tue, May 27, 2003 at 11:19:32AM -0400, Jer Jackson wrote:
> What about putting this into VFS and letting VFS use trusted extended
> attributes to store in on XFS, NTFS, ext2/3, whatever?

because they are just bits, its really overkill to use a xattr for
that.  the VFS already handles all the enforcment and checking for
immutable/appendonly the filesystem only needs to set them in the vfs
inode structure, and provide some means to set/clear them (usually an
ioctl).  xfs has several unused bits in the inode which could be used
for this, and it won't break backwards compatibility since old kernels
will simply ignore and leave alone reserved bits.

Ethan Benson

Attachment: pgp45rKyrH8d8.pgp
Description: PGP signature

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