Re: [B.A.T.M.A.N.] "wrong count at exit" after spin_lock/_unlock pair

[Sven Eckelmann <sven.eckelmann@xxxxxx>]

Sven Eckelmann wrote:
> Hi,
> 
> I use sparse (cgcc) with different kernel versions to do some quick checks
> on the batman-adv kernel module. I noticed that my change from UP to SMP
> kernel generated some 'weird' warnings regarding a 'context imbalance'
> which I couldn't find, but went away when switching from spin_(un)lock to
> spin_(un)lock_bh - something which doesn't make sense in my mind.
> 
> I am not quite sure if this is linux-headers or sparse related - so I try
> to post it here first. I choose 2.6.32 because it is the latest long
> stable kernel and will be used in by Debian Squeeze.

I looked a little more at the problem and it is not a real sparse problem. The 
problem came with bb7bed082500179519c7caf0678ba3bed9752658 and and was 
corrected with b97c4bc16734a2e597dac7f91ee9eb78f4aeef9a. So patching 2.6.32 
and v2.6.33 by reverting bb7bed082500179519c7caf0678ba3bed9752658 is enough 
for me.

Best regards,
	Sven

Attachment: signature.asc
Description: This is a digitally signed message part.