xfs
[Top] [All Lists]

Re: file corruption during emacs build on XFS logical volume

To: Steve Lord <lord@xxxxxxx>, Sean Neakums <sneakums@xxxxxxxx>
Subject: Re: file corruption during emacs build on XFS logical volume
From: Seth Mos <knuffie@xxxxxxxxx>
Date: Fri, 04 Jan 2002 21:26:06 +0100
Cc: Linux XFS <linux-xfs@xxxxxxxxxxx>
In-reply-to: <1010174871.30053.6.camel@xxxxxxxxxxxxxxxxxxxx>
References: <6ug05mj6ah.fsf@xxxxxxxxxxxxx> <1010141444.1992.3.camel@pyewacket> <1010164935.1945.8.camel@UberGeek> <6u666ikn5n.fsf@xxxxxxxxxxxxx> <6usn9mj7id.fsf@xxxxxxxxxxxxx> <6uofkaj6zj.fsf@xxxxxxxxxxxxx> <6ug05mj6ah.fsf@xxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
At 14:07 4-1-2002 -0600, Steve Lord wrote:
On Fri, 2002-01-04 at 12:54, Sean Neakums wrote:
> begin  Sean Neakums quotation:
>
> > I am now going to do a build of the upstream source and see if I can
> > make the dump break on that too.  I'm hopeful that it will, as the
> > unexec code is the same in upstream as in the Debian emacs21
> > package.
>
> Yep, I just got me a dumped emacs binary full of NULs with the
> original GNU source.

OK, bingo, I can replicate this now, I too have a bad binary, it looks
like the memory is getting reclaimed without getting flushed out to
disk - which is not good at all.

This actually gives me something to go on, hopefully it won't take too
long to find now. Thanks for your persistance in generating a test
case here.

This might explain the fs corruption that some other people saw as well on 2.4.17-xfs.

I can't remember who it was, i'll look it up.

Cheers

--
Seth
Every program has two purposes one for which
it was written and another for which it wasn't
I use the last kind.


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