Hi!
Just a short question: is there a patch for 2.5.26 to update ext3 to
ext3-0.9.18? There's still ext3-0.9.16 from Dec. 2001 present in 2.5.26.
At ../people/sct on ftp.kernel.org there are only updates for kernel
2.2 and 2.4.
Greetings, Heinz.
--
# Heinz Diehl, 68259 Mannheim, Germany
Heinz Diehl wrote:
>
> Hi!
>
> Just a short question: is there a patch for 2.5.26 to update ext3 to
> ext3-0.9.18? There's still ext3-0.9.16 from Dec. 2001 present in 2.5.26.
> At ../people/sct on ftp.kernel.org there are only updates for kernel
> 2.2 and 2.4.
2.5 is uptodate wrt the current ext3-for-2.4 development tree.
That means that it's more uptodate than 2.4 is...
Some recent changes to ext3 have exposed a data=journal bug
in 2.5 which is also present in 2.4, but is much harder to hit
there. I'm not sure what Stephen's intentions are on a 2.4
upgrade, but I'd be inclined to sit tight until 2.4.20-pre.
-
Hi,
On Sun, Jul 21, 2002 at 09:53:18PM -0700, Andrew Morton wrote:
> > Just a short question: is there a patch for 2.5.26 to update ext3 to
> > ext3-0.9.18? There's still ext3-0.9.16 from Dec. 2001 present in 2.5.26.
> > At ../people/sct on ftp.kernel.org there are only updates for kernel
> > 2.2 and 2.4.
>
> 2.5 is uptodate wrt the current ext3-for-2.4 development tree.
> That means that it's more uptodate than 2.4 is...
Yes --- I've been holding back on the changes in the ext3 CVS because
of one nagging bug which I've been hunting for, and which I think I
just found two weeks ago, the day before I left for a holiday. (It's
another possible cause for a "buffer_jdirty()" assert failure in
commit.c on SMP machines.)
I'll get that checked in shortly.
> Some recent changes to ext3 have exposed a data=journal bug
> in 2.5 which is also present in 2.4, but is much harder to hit
> there. I'm not sure what Stephen's intentions are on a 2.4
> upgrade, but I'd be inclined to sit tight until 2.4.20-pre.
I'm just back from holiday so I haven't been able to do much on this
recently, but I've got a fix mostly coded --- it just doesn't
actually work particularly well, yet. :-)
Cheers,
Stephen