Bug 130441 - Network packet from other machine reproduceably freezes my PC (sk98lin)
Summary: Network packet from other machine reproduceably freezes my PC (sk98lin)
Status: RESOLVED WORKSFORME
Alias: None
Product: SUSE LINUX 10.0
Classification: openSUSE
Component: Kernel (show other bugs)
Version: Final
Hardware: x86-64 SuSE Linux 10.0
: P5 - None : Major
Target Milestone: ---
Assignee: Karsten Keil
QA Contact: E-mail List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2005-10-25 08:28 UTC by Manfred Hollstein
Modified: 2005-10-25 11:22 UTC (History)
0 users

See Also:
Found By: Other
Services Priority:
Business Priority:
Blocker: ---
Marketing QA Status: ---
IT Deployment: ---

Attachments
Output from running "lspci -v" on PC1 (6.17 KB, text/plain)
2005-10-25 08:29 UTC, Manfred Hollstein 		Details
Screenshot when error occured (1.24 MB, image/jpeg)
2005-10-25 08:32 UTC, Manfred Hollstein 		Details
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Manfred Hollstein 2005-10-25 08:28:32 UTC 
This is the scenario:

1. Start PC1 with all nfsserver related services enabled.
2. When PC1 has finished booting, type the following on PC2 (autofs enabled
   using /etc/auto.net to access remote filesystems):

    cd /net/PC1

3. PC1 freezes immediately without any message.

Adding "nmi_watchdog=1" to the boot parameters reveals a problem in the sk98lin
driver used for the GigE device.

The funny thing is, that doing the same with the roles of PC1 and PC2 swapped
makes the problem go away; PC1 remains fully useable afterwards. There must be
some missing/wrong initialization of the sk98lin network driver.

I'll add output from lspci and a screenshot of the crash to this report in a
few seconds.
Comment 1 Manfred Hollstein 2005-10-25 08:29:29 UTC 
Created attachment 55299 [details]
Output from running "lspci -v" on PC1
Comment 2 Manfred Hollstein 2005-10-25 08:32:22 UTC 
Created attachment 55300 [details]
Screenshot when error occured
Comment 3 Andreas Kleen 2005-10-25 10:18:10 UTC 
A single broken driver is not a blocker.
It likely comes from the sk98lin-update.
Comment 4 Olaf Kirch 2005-10-25 10:26:43 UTC 
We cannot support the sk98lin driver. It's out of tree and SK hasn't
really been forthcoming in addressing problems.

Did you try using skge instead of sk98lin?
Comment 5 Manfred Hollstein 2005-10-25 10:38:38 UTC 
(In reply to comment #3)
> A single broken driver is not a blocker.
> It likely comes from the sk98lin-update.
> 

For me it is. If somebody from within the network can crash my PC...
Comment 6 Manfred Hollstein 2005-10-25 10:39:22 UTC 
(In reply to comment #4)
> We cannot support the sk98lin driver. It's out of tree and SK hasn't
> really been forthcoming in addressing problems.
> 
> Did you try using skge instead of sk98lin?
> 

Not yet, will do that in a minute (didn't know there is an alternative one).
Comment 7 Manfred Hollstein 2005-10-25 10:55:26 UTC 
Update: skge works, the system survives the first packet (and further ones,
too ;-) ) coming from the network. FWIW, modinfo doesn't show skge as being
supported either, /var/log/messages contains this:

kernel: skge: module not supported by Novell, setting U taint flag.

But, I can live with that. Thanks a lot for your help! Maybe we should think
about, replacing usage of sk98lin with skge as a default if we cannot support
it and SK isn't responsive at all?
Comment 8 Olaf Kirch 2005-10-25 11:22:37 UTC 
That is indeed the plan for 10.1 and SLES10.

Thanks!