xfs
[Top] [All Lists]

Re: 2.4.22-pre patches

To: Net Llama! <netllama@xxxxxxxxxxxxx>
Subject: Re: 2.4.22-pre patches
From: Matteo Centonza <matteo@xxxxxx>
Date: Thu, 26 Jun 2003 15:57:01 +0200 (CEST)
Cc: linux-xfs@xxxxxxxxxxx
In-reply-to: <Pine.LNX.4.55.0306260926010.14754@linux-sxs.org>
Sender: linux-xfs-bounce@xxxxxxxxxxx
Hi,

> > On Wed, 25 Jun 2003, Net Llama! wrote:
> >
> > > On 06/25/03 14:22, Christoph Hellwig wrote:
> > > > On Wed, Jun 25, 2003 at 04:19:08PM -0500, Steve Lord wrote:
> > > >> On Wed, 2003-06-25 at 15:56, Net Llama! wrote:
> > > >> > Anyone know if/when the cvs snapshots will cleanly apply against the
> > > >> > 2.4.22-pre1 kernel?  I'd really like to use it, as the acpi support 
> > > >> > is
> > > >> > significantly better than what was in 2.4.21(-final).
> > > >> >
> > > >> >
> > > >>
> > > >> Your best bet is to take the xfs cvs kernel, and apply the 2.4.22-pre
> > > >> patch to it. Apart from kdb, it is usually pretty clean.
> > > >
> > > > I don't think it's that easy.  2.4.22-pre has the O_DIRECT changes from 
> > > > sct.
> > > > Fixing XFS so that it compiles isn't that difficult (just some conflicts
> > > > in filemap.c) but it's much more work to not cripple XFS O_DIRECT 
> > > > support
> > > > due to these new locks.  Without any changes all O_DIRECT I/O would 
> > > > happen
> > > > with i_sem held now.
> > >
> > > Ya, unfortunately, its not that easy.  I gave it a run, and there were a
> > > handful of failed hunks, _alot_ of fuzz & offsets.  thanks though.
> >
> >
> > why not applying the acpi patch only?
> >
> > They have a pretty acpi-20030619-2.4.21.diff.gz on www.sf.net/projects/acpi
> 
> What do you mean by 'only'?  That is the patch that i tried to apply,
> however i don't even think its neccesary in 2.4.22-pre1, because its
> already been updated with the same patches.

as i've understood it, if you're interested only in new acpi
features of 2.4.22-pre1 (and of course in XFS), you can download the 
current XFS cvs kernel (2.4.21 based) and apply the latest acpi 
patch (against 2.4.21).

This merge should have the acpi bits you want.

You have to resolve 1 trivial reject (induced by KDB) in  
arch/i386/kernel/io_apic.c

Ciao,

-m


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