[Top] [All Lists]

Re: Bug in sgi-xfs?

To: Krzysztof Rusocki <kszysiu@xxxxxxxxxxxxxxxxx>
Subject: Re: Bug in sgi-xfs?
From: Eric Sandeen <sandeen@xxxxxxx>
Date: Tue, 09 Oct 2001 09:34:41 -0500
Cc: Seth Mos <knuffie@xxxxxxxxx>, Willi.Langenberger@xxxxxxxxxxxxx, linux-xfs@xxxxxxxxxxx
References: <15298.49018.477634.619452@xxxxxxxxxxxxxxxxxxx> <15298.48625.831518.499561@xxxxxxxxxxxxxxxxxxx> <15298.49018.477634.619452@xxxxxxxxxxxxxxxxxxx> <20011009131721.A9570@xxxxxxxxxxxxxxxxx> <> <20011009143414.B9570@xxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
Hi Krzysztof -

A couple of points - "XFS 1.0.1 for Linux 2.4.10" is almost impossible,
by definition - the VM has changed so much in 2.4.10 that XFS/pagebuf
had to be rearranged around the kernel - and parts of the XFS code for
2.4.10 is fundamentally different code than that released for 2.4.5.  So
I wouldn't want to call any XFS code running under Linux 2.4.10 "XFS

Regarding testing of 2.4.10, there has not been extensive testing of
that snapshot at SGI (at least not on the level of a release kernel). 
However, I have found the CVS tree to be quite stable, at least when
it's at one of Linus' major releases (not -preX).  In fact, Mandrake
took the 2.4.8 snapshot for their release kernel, and they were very
impressed with the stability.

There will continue to be stable, heavily tested XFS releases, but not
for every kernel version, simply due to how much time and resources are


Krzysztof Rusocki wrote:
> On Tue, Oct 09, 2001 at 01:54:17PM +0200, Seth Mos wrote:
> Hi Seth,
> > That would mean a kernel needs to be heavily tested before it can be used
> > for an update disk.
> So, you mean that testing 2.4.10 received is still insufficient ?
> > Since the development is past 1.0.1 already and fitting 1.0.1 from 2.4.5 on
> > a newer kernel wouldn't be a really good idea either.
> I really can't see the point why it shouldn't be good idea. What speaks
> against doing so? any known problems?
> Actually I tried to speak in the name of few people that either already 
> reported
> issues concerning 2.4.5 itself, or - undoubtfully - will report such problems
> in future.
> I really doubt if 2.4.5 is more ,,stable'' (according to new stability
> definition that came with 2.4 series ;P ) than 2.4.10.
> Personally, I use cvs kernels so it does not affect me, what I said is just
> my suggestion/opinion, choice is yours ;-)
> > It is very likely that by the time a redhat 7.2 update disk comes along we
> > will probably have a newer version included anyways.
> Those words doesn't speak against upgrading rpms ;-)
> Cheers,
> Krzysztof

Eric Sandeen      XFS for Linux     http://oss.sgi.com/projects/xfs
sandeen@xxxxxxx   SGI, Inc.

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