headers
David Cannings | 1 Nov 2004 01:00

Re: [uml-user] generic ebtables script ?

Sunny Dubey wrote:
> Or should I skip ebtables altogether and use the brnf patch ?

Are you still using 2.4?  The functionality of the brnf patch is
included in 2.6 netfilter by default.

David

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Permalink | Reply | 1 comment |
headers
Sunny Dubey | 1 Nov 2004 02:01

Re: [uml-user] generic ebtables script ?

On Sunday 31 October 2004 19:00, David Cannings wrote:
> Sunny Dubey wrote:
> > Or should I skip ebtables altogether and use the brnf patch ?
>
> Are you still using 2.4?  The functionality of the brnf patch is
> included in 2.6 netfilter by default.

ah, I see it

thanks

I've largely ignored filtering on the bridge level till now.  I can't see 
myself doing much other than figuring how to prevent users from stealing IPs 
and such.

thanks again

Sunny Dubey

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Permalink | Reply | 0 comments |
headers
Tupshin Harper | 1 Nov 2004 04:33

[uml-user] skas for 2.6 x86_64???

Is there a working skas patch for a recent 2.6 x86_64 kernel? I see one 
2.4 patch for x86_64, and a whole bunch of skas options for x86 2.6, but 
nothing for 2.6 x86_64. Does anybody have this working yet?

FWIW, I'm running a debian pure64 distro, and running a regular x86 2.4 
uml kernel inside of a x86 chroot but the uml performance seems pretty 
bad without skas.

I would also be quite interested in any working x86_64 uml kernels as 
well. I've seen the 2.6.9 x86_64 patches at 
http://user-mode-linux.sourceforge.net/patches.html, but it looks to be 
"not quite there yet". Does this accurately reflect the current situation?

-Tupshin

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
Permalink | Reply | 2 comments |
headers
Lei Yang | 1 Nov 2004 05:54

[uml-user] get stack traces for kernel panic

Hello,

I am debugging a block device driver with uml 2.4.26 under skas mode. 
After I have a kernel panic, how can I get the stack traces other than 
the nanosleep stuff? They don't help to find what caused the panic.

TIA and good night
Lei
-------------------------------------------------------------------------------------------------------------------------
Here is the trace I get after I interrupt UML by sending a SIGINT signal 
from another shell.

linux-uml:/home/lei/testblk # mke2fs /dev/testblk/1
mke2fs 1.34 (25-Jul-2003)
Filesystem label=
OS type: Linux
Block size=1024 (log=0)
Fragment size=1024 (log=0)
128 inodes, 1024 blocks
51 blocks (4.98%) reserved for the super user
First data block=1
1 block group
8192 blocks per group, 8192 fragments per group
128 inodes per group

Writing inode tables: done
Writing superblocks and filesystem accounting information: Kernel panic: 
Kernel mode fault at addr 0x400, ip 0x4011b271

Message from syslogd <at> linux-uml at Mon Nov  1 04:33:07 2004 ...
(Continue reading)

Permalink | Reply | 0 comments |
headers
Itai Tavor | 1 Nov 2004 08:17
Picon
Picon

[uml-user] UML not starting up... host 2.6.10-rc1, UML 2.6.8.1

Hi,

I'm using the following:

- Host Fedora Core 2 with kernel 2.6.10-rc1
- UML kernel 2.6.8.1 with uml-patch-2.6.8.1-1
- root_fs.rh-9-full.pristine.20030724.bz2

UML kernel built with:
     make mrproper ARCH=um
     make menuconfig ARCH=um
     make linux ARCH=um

On running ./linux, the kernel hangs at

     VFS: Mounted root (ext2 filesystem) readonly.
     Mounted devfs on /dev

And takes %50 CPU forever after that...

Any tips?

TIA, Itai

-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click
(Continue reading)

Permalink | Reply | 3 comments |
headers
Erik de Bruijn - BudgetDedicated.com | 1 Nov 2004 12:25

Re: [uml-user] skas for 2.6 x86_64???

We're very interested too...

I'm also stuck with an old kernel on my system if I want SKAS, and I 
really do! I also don't want the 2.4 kernel because it hardly supports 
any 64 bits hardware (not mine at least).

An update on this would be greatly appreciated!

-- Erik

P.s. Don't vote Bush, someone else should have the possibility to do 
what is best for _their_ world (which IS the US, right?!). </sarcastic>

Tupshin Harper wrote:

> Is there a working skas patch for a recent 2.6 x86_64 kernel? I see 
> one 2.4 patch for x86_64, and a whole bunch of skas options for x86 
> 2.6, but nothing for 2.6 x86_64. Does anybody have this working yet?
>
> FWIW, I'm running a debian pure64 distro, and running a regular x86 
> 2.4 uml kernel inside of a x86 chroot but the uml performance seems 
> pretty bad without skas.
>
> I would also be quite interested in any working x86_64 uml kernels as 
> well. I've seen the 2.6.9 x86_64 patches at 
> http://user-mode-linux.sourceforge.net/patches.html, but it looks to 
> be "not quite there yet". Does this accurately reflect the current 
> situation?
>
> -Tupshin
(Continue reading)

Permalink | Reply | 1 comment |
headers
Kristian Sørensen | 1 Nov 2004 16:13
Picon
Picon
Favicon

[uml-user] Adding system call in UML and compiling program for it

Hi!

We have a small problem when adding a new system call to Linux .. It works 
perfect in the plain kernel, but UML yields some trouble when compiling a 
program for it.

The system call is named umb_set_child_restrictions and is implemeted in our 
Umbrella LSM module.

When compiling a user space program using the system call (and the 
C_INCLUDE_PATH is pointing correctly at linux-2.6.8.1-um/include) we get a 
huge trace of errors (attached).

We found no problems in linux-2.6.3. For all versions of Linux it works 
smoothly for i386.

Can anyone give a pointer of how to go from here?

Best regards,
Kristian Sørensen.

PS. Is UML migrated into the vanilla kernel? I saw a few posts on LKML about 
UM...

--

-- 
Kristian Sørensen
- The Umbrella Project
  http://umbrella.sourceforge.net

E-mail: ipqw <at> users.sf.net, Phone: +45 29723816
(Continue reading)

Permalink | Reply | 1 comment |
headers
Erik de Bruijn - BudgetDedicated.com | 2 Nov 2004 00:08

Re: [uml-user] UML not starting up... host 2.6.10-rc1, UML 2.6.8.1

Hello Itai,

Have you taken a look at the thread: [uml-user] 2.4.24-1um hang at 
"Mounted devfs on /dev"?

Are you sure that this system works with different host/guest kernel 
configurations? Probably it's a filesystem failure. Have you perhaps 
crashed/killed the UML and then fsck'ed it? This is a process I can 
repeat that breaks the working "root_fs.rh-9-full.pristine.20030724". 
I'm very happy I've found this out after weeks of experimenting with 
ever growing pressure by people that were waiting for me to get this 
redhat system ready...

Please check on if whether any version of the same redhat pristine 
rootfs works for you. If you have a working version (back it up and), 
see what is does if you kill it while running (just kill -15 for 
example). Then run it, it should still work. Then kill it and repair it 
from the host. I did this with fsck.ext3 -p ("yes to all"). After this 
operation it STOPPED bootting correctly. Doing the same but reparing 
from inside the UML, keeps it working/booting.

A general question: Should one never fix such a UML root_fs from the 
host, or is it a problem that the killed UML process?

--

-- 
Kind regards,

Erik de Bruijn
BudgetDedicated
(Continue reading)

Permalink | Reply | 1 comment |
headers
Itai Tavor | 2 Nov 2004 02:56
Picon
Picon

Re: [uml-user] UML not starting up... host 2.6.10-rc1, UML 2.6.8.1

On 02/11/2004, at 10:08 AM, Erik de Bruijn - BudgetDedicated.com wrote:

> Hello Itai,
>
> Have you taken a look at the thread: [uml-user] 2.4.24-1um hang at 
> "Mounted devfs on /dev"?

I tried to change O2 to O1 in the kernel Makefile as that thread 
suggests, but it had no effect.
>
> Are you sure that this system works with different host/guest kernel 
> configurations? Probably it's a filesystem failure. Have you perhaps 
> crashed/killed the UML and then fsck'ed it? This is a process I can 
> repeat that breaks the working "root_fs.rh-9-full.pristine.20030724". 
> I'm very happy I've found this out after weeks of experimenting with 
> ever growing pressure by people that were waiting for me to get this 
> redhat system ready...
>
> Please check on if whether any version of the same redhat pristine 
> rootfs works for you. If you have a working version (back it up and), 
> see what is does if you kill it while running (just kill -15 for 
> example). Then run it, it should still work. Then kill it and repair 
> it from the host. I did this with fsck.ext3 -p ("yes to all"). After 
> this operation it STOPPED bootting correctly. Doing the same but 
> reparing from inside the UML, keeps it working/booting.

I'm a new UML user... I had no working configuration. I downloaded the 
root fs, built the kernel and tried to start it. I've done nothing to 
the root fs other than mounting it once to see what's in it. I never 
crashed the UML as it never actually started up...
(Continue reading)

Permalink | Reply | 0 comments |
headers
Dan Lund | 2 Nov 2004 06:35
Picon

Re: [uml-user] How to get error log without x-term

use "con0=fd:0,fd:1 con=pty" to make it pure console.  It'll work inside 
of an xterm without a problem.  Use "screen" or "conserver" to make it 
backgroundable (not dependant on that particular console staying up, and 
redirectable), and your starting on the path to professional uml usage.

Also, in your /etc/inittab, make sure you have a console pointing to the 
tty0.
In my Gentoo UML, I have this (for one console):
c0:12345:respawn:/sbin/agetty 38400 tty0 linux

However, my RH8 UML has this (for one console):
0:2345:respawn:/sbin/mingetty ttys/0

That'll fix your problem if it stopping at "Initializing stdio console 
driver".  It's not really stopping there, you're just not seeing the 
rest because at that point it initializes the console and redirects all 
output there.

--Dan

Jimmy Pan wrote:

> Hi,
> i use xterm=none because i don't want the extra 2 x-terms come up at 
> my server desktop...
> and one more thing, when i run UML and it came to
> .............
>
> ubdb: unknown partition table
> ubdc: unknown partition table
(Continue reading)

Permalink | Reply | 0 comments |

Gmane