Re: [uml-devel] skas patch 2.6.0 -A4 mistake

[David B H. <dbh...@ee...>]

On Sun, 21 Dec 2003 00:56:15 +0100
Ingo Molnar <mi...@el...> wrote:
> > you have changed the ret=0 to ret=count in the PTRACE_SWITCH_MM case
> > from file kernel/ptrace.c instead of MM_COPY_SEGMENTS case from
> > mm/proc_mm.c
> 
> doh - you are right, -A5 uploaded.
> 
> /me should quit fixing bugs by hand-editing patches ...

Hey there :) I'm going through 2.6.0, testing that the functionality I
use is available and working (in preparation for an eventual rollout to
my machines), and I'm getting some odd behaviour with stock 2.6.0 from
kernel.org and your -combo-A5 patch applied. Debian Sid host, Debian
Woody guest:

$ linux umid=umltest ubd0=disks/umltest root=/dev/ubd0 mem=128M con0=fd:0,fd:1 con=pts eth0=tuntap,tap0 eth1=daemon init=/bin/sh
Checking for the skas3 patch in the host...found
Checking for /proc/mm...found
<snip>
VFS: Mounted root (ext3 filesystem) readonly.
init-2.05a# /etc/rcS.d/S10checkroot.sh
Activating swap.
Checking root file system...
fsck 1.27 (8-Mar-2002)
fsck.ext3: error while loading shared libraries: /lib/libcom_err.so.2: cannot read file data: Error 38
<snip>
init-2.05a# ls -l /lib/libcom_err.so.2 /lib/libcom_err.so.2.0 
lrwxrwxrwx    1 root     root           17 Dec 22 01:56 /lib/libcom_err.so.2 -> libcom_err.so.2.0
-rw-r--r--    1 root     root         5016 Mar 21  2002 /lib/libcom_err.so.2.0

On another boot, I got things like:

[ root@(none): ~/ ]# fsck.ext3 /dev/ubd0
e2fsck 1.27 (8-Mar-2002)
/dev/ubd0 contains a file system with errors, check forced.
Pass 1: Checking inodes, blocks, and sizes
Ignore error<y>? yes

Ignore error<y>? yes

Error reading block 9299 (Function not implemented) while reading indirect blocks of inode 4073.  Ignore error<y>? yes

Any thoughts, or any more information I can provide?