Re: 2.6.20 ext3 fix backported to Debian?
On Sat, Jan 13, 2007 at 08:47:33AM +0100, Jean-Luc Coulon (f5ibh) wrote:
> Le 12.01.2007 23:14:48, Lennart Sorensen a écrit :
> >On Fri, Jan 12, 2007 at 10:32:23PM +0100, Jean-Luc Coulon (f5ibh)
> >wrote:
> >> Is there a 2.6.19 in Debian?
> >
> >I don't think so yet. I beleive the decision is that etch will use
> >2.6.18, which means unstable isn't too likely to move past 2.6.18
> >until
> >etch is released.
>
> My question was because the original question (see title) asked about a
> backport of a 2.6.20 fix to the _Debian_ 2.6.19.
>
> It seems that 2.6.19 has some bugs not fixed for the moment and
> probably harf to fix.
>
> Maybe 2.6.19 will be skipped?
To start, I think you're asking this on the wrong list. There is
probably is nothing amd64 or even 64-bit about this, and it's probably a
general bug. debian-kernel@lists.debian.org would be a better list to
ask such questions.
Anyway, I have no idea what bug you're talking about. I know about
corruption caused by msync(), which doesn't have anything to do with
ext3, it affects all filesystems.
This bug is actually present in Debian's 2.6.18-3 kernels because
of a backport of 2.6.19 and should get fixed before the release.
Debian's kernel team is actually the one that tracked which change
caused it.
See http://bugs.debian.org/401006 and related bugs.
If you're talking about another bug, it might be a good idea to file a
bug against the kernel package so that they know about it.
It's unlikely that Debian will have an 2.6.19 kernel in unstable since
2.6.20 is probably already out by that time. In any case, if that's the
bug you're talking about, and we get a 2.6.19 kernel in unstable, it
should be fixed. If it's not that bug, and you file a bug about it, it
probably gets fixed too.
Kurt
fixed too.
Reply to: