xfs
[Top] [All Lists]

Re: Q: Filesystem block sizes available?

To: Tux mailing list <tux-list@xxxxxxxxxx>
Subject: Re: Q: Filesystem block sizes available?
From: Ingo Molnar <mingo@xxxxxxx>
Date: Mon, 19 Nov 2001 11:27:21 +0100 (CET)
Cc: Steve Lord <lord@xxxxxxx>, XFS Mailing list <linux-xfs@xxxxxxxxxxx>
In-reply-to: <Pine.LNX.4.30.0111131556430.1160-100000@mustard.heime.net>
Reply-to: <mingo@xxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
On Tue, 13 Nov 2001, Roy Sigurd Karlsbakk wrote:

> Do you or anyone else have some idea of when the TUX/XFS potch
> incompatibility problem will be solved? I need Tux (the in-kernel web
> server from RedHat) in this project as well...

there is no fundamental incompatibility, it's just that both the XFS and
the TUX patches change similar areas of code and thus they wont apply
after each other. Both TUX and XFS can be blamed equally well for this
situation, both have changes (of varying complexity) that are not yet in
the Linus kernel.

you'll have to do extra work to 'merge' the current TUX code to SGI's XFS
source tree. Do something like this:

 - create a recent XFS tree, as directory linux-xfs
 - create the latest (XFS-less) TUX tree, as directory linux-tux

and execute these commands to 'upgrade' the XFS tree to the latest TUX
code:

  cp linux-tux/net/tux/* linux-xfs/net/tux/
  cp linux-tux/include/net/tux* linux-xfs/include/net/tux*

this *should* produce a linux-xfs tree that *should* compile & work just
as well as the normal linux-xfs tree would work. Merging of two such
orthogonal subsystems such as XFS and TUX usually doesnt produce hard to
fix problems, usually any merging problem either shows up at compile time,
or shows up as some fundamental inability of TUX to start up or function.
The basic rule, since XFS is a filesystem (and is responsible for your
files and data) and as such is the more intrusive patch, always merge TUX
to XFS, not the other way around.

but please always state in any possible bugreports that you are not using
my tree (this obviously holds for -aa patch users as well), just to be
able to filter out weird bugreports that no-one else is seeing.

        Ingo


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