Re: [Xen-users] Badness in softirq.c / no modules loaded / related to network interface

[Sebastian Gutweiler <sebastian@xxxxxxxxxxxxx>]

On Monday 09 May 2005 17:55, Nick Craig-Wood wrote:

Hi,

> On Fri, May 06, 2005 at 06:40:33PM +0200, sebastian@xxxxxxxxxxxxx wrote:
> > I have a porblem with the unprivileged kernel. When I start my domain
> > using the unprivileged kernel I get the message:
> >
> > Badness in local_bh_enable at kernel/softirq.c:140
> >  [<c011f3c0>] local_bh_enable+0x80/0x90
> >  [<c0223089>] skb_checksum+0x129/0x2a0
> >  [<c026736c>] udp_poll+0x9c/0x150
> >  [<c021e409>] sock_poll+0x29/0x40
> >  [<c016ba3e>] do_select+0x25e/0x2d0
> >  [<c016b630>] __pollwait+0x0/0xd0
> >  [<c016bd9f>] sys_select+0x2bf/0x4d0
> >  [<c01093f4>] syscall_call+0x7/0xb
>
> We are still having this problem too.  We only have it on some hosts,
> not others so its related to some activity in the domUs - we haven't
> figured out what though.  The backtrace suggests UDP traffic

I experience this, too. Actually the reappearing of the message in the 
dom0-logs is caused by my dhcp-client regularly sending UDP-Packets.

> I'm prety sure we didn't miscompile our modules - I disassembled them
> to check for cli/sti.  The kernel was compiled using

I'm almost sure, too. 

>   make-kpkg --arch xen --append_to_version -xen --revision=2.6.11
> kernel_image
>
> We see this with the e1000 driver, using kernel 2.6.11.7 + debian
> patches + xen stable 2.0.5, on dell poweredge 750 hardware.

Same kernel here, but xen-testing which is about 5-6 days old.

My kernel who has this problem was compiled with Debian's standard 
kernel .config but ISA-Bus-Support disabled. 

I compiled a standard xen0 kernel for testing with only some extra drivers 
enabled. It turned out that this kernel did not have this problem.

I will compile the standard Debian-kernel without e1000 support and retry this 
kernel.

> No functionality appears to be affected, other than 1000s of these
> messages in the log.

I experience notable higer packet packet drops when pinging dom0-host from an 
domU-host with hughe packet sizes, e.g. ping -s 8008 dom0-host. It's about 
30% when using the kernel with the problem, about 0% when using the standard 
dom0-kernel-configuration.

Sebastian 

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users