Re: [PowerPC] Next May 8 boot failure: OOPS during ibmveth moduleinit

Hi all,

On Mon, 11 May 2009 22:37:09 +0530 Sachin Sant [off-list ref] wrote:

Jiri Pirko wrote:

quoted

Fri, May 08, 2009 at 09:57:22PM CEST, davem@davemloft.net wrote:
  

quoted

From: Sachin Sant <redacted>
Date: Fri, 08 May 2009 18:22:48 +0530

    

quoted

Todays Next failed to boot on a Power6 JS22 blade with following oops.
      

Jiri, I suspect this might be your address list changes.

Although that's just a guess.  But please take a look.
    

Hmm, only thing I see might cause the problem would be if calling __hw_addr_add
in dev_addr_init fails, then dev->dev_addr would contain zeroes (which looks
this is not the case). But in this case the oops would appear earlier (in
ibmveth_probe dev_addr memcpy).

Will do the patch which checks the result to behave correctly in case of oom
but this imho this wouldn't help. Strange, I will dig into this more tomorrow.

Bisected down to:

commit f001fde5eadd915f4858d22ed70d7040f48767cf
Author: Jiri Pirko [off-list ref]
Date:   Tue May 5 02:48:28 2009 +0000

    net: introduce a list of device addresses dev_addr_list (v6)

I can reproduce this at will on the last three linux-next trees.
-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/