2003-06-28 10:25:55

by Margit Schubert-While

[permalink] [raw]
Subject: 2.4.22-pre2 unresolved proc_get_inode

if [ -r System.map ]; then /sbin/depmod -ae -F System.map 2.4.22-pre2; fi
depmod: *** Unresolved symbols in
/lib/modules/2.4.22-pre2/kernel/drivers/net/wan/comx.o
depmod: proc_get_inode

I suppose we let Christoph and Marc fight it out.

Margit


2003-06-28 11:07:36

by Al Viro

[permalink] [raw]
Subject: Re: 2.4.22-pre2 unresolved proc_get_inode

On Sat, Jun 28, 2003 at 12:40:36PM +0200, Margit Schubert-While wrote:
> if [ -r System.map ]; then /sbin/depmod -ae -F System.map 2.4.22-pre2; fi
> depmod: *** Unresolved symbols in
> /lib/modules/2.4.22-pre2/kernel/drivers/net/wan/comx.o
> depmod: proc_get_inode
>
> I suppose we let Christoph and Marc fight it out.

You know what? I'm so fed up with that crap, that today Marcelo will
get a patch killing proc_get_inode(), making proc_lookup() static and
eliminating ->proc_iops completely.

Enough is enough. comx is the only user of that crap and all procfs
code in comx is broken by design and trivially exploitable. It's
unsalvagable and any attempt to fix it will amount to rewrite from
scratch anyway.

It's not a new problem, it had been discussed to hell and back and
comx folks could not have been arsed to do anything about it in what,
3 years?

2003-06-28 12:01:05

by Margit Schubert-While

[permalink] [raw]
Subject: Re: 2.4.22-pre2 unresolved proc_get_inode

Fine by me :-)
Don't use it, don't need it.
As a matter of course, I throw a standard Suse config in and answer
"m"/"y" to the new stuff. Serves to shake out a few things.

Margit

At 12:21 28.06.2003 +0100, Viro wrote:
>On Sat, Jun 28, 2003 at 12:40:36PM +0200, Margit Schubert-While wrote:
> > if [ -r System.map ]; then /sbin/depmod -ae -F System.map 2.4.22-pre2; fi
> > depmod: *** Unresolved symbols in
> > /lib/modules/2.4.22-pre2/kernel/drivers/net/wan/comx.o
> > depmod: proc_get_inode
> >
> > I suppose we let Christoph and Marc fight it out.
>
>You know what? I'm so fed up with that crap, that today Marcelo will
>get a patch killing proc_get_inode(), making proc_lookup() static and
>eliminating ->proc_iops completely.
>
>Enough is enough. comx is the only user of that crap and all procfs
>code in comx is broken by design and trivially exploitable. It's
>unsalvagable and any attempt to fix it will amount to rewrite from
>scratch anyway.
>
>It's not a new problem, it had been discussed to hell and back and
>comx folks could not have been arsed to do anything about it in what,
>3 years?