[lucid, maverick] SRU: apparmor_parser triggers a kernel panic

[John Johansen]

On 03/10/2011 09:20 AM, Stefan Bader wrote:
> On 03/10/2011 05:11 PM, Paolo Pisati wrote:
>> SRU Justification:
>>
>>     Impact: kernel panic when loading a malformed apparmor profile.
>>     Fix: see attached patch.
>>     Testcase: /etc/init.d/apparmor restart
>>
>>
>> Buglink:
>> https://bugs.launchpad.net/ubuntu/+source/linux-mvl-dove/+bug/732700
>>
>> This affetcs lucid/master, lucid/mvl-dove and maverick/mvl-dove.
>>
>> This fix a regression in the lucid/mvl-dove -proposed kernel.
>>
>> bye,
>> p
> Look reasonable. Personally I would add a bit more description into the commit
> message. Also as apparmor is upstream now and apparently is not affected I would
> wonder (without too much research on my own) why this is and whether there would
> be a equivalent upstream patch...
> 
> Ackish, but interested in more info.
> 
This brings the code in line with what is upstream.  However this appears to have
been fixed in the series of commits, cleaning up the code, making sure vars
were zeroed out etc, and was never actually identified as a bug.

It prevents the oops by ensuring the new_profile is null when it passes through
the cleanup block at the end of the function which does
	aa_put_profile(new_profile);

This will put the kfree and call free on the object unless it is null.

Acked-by: John Johansen <john.johansen at canonical.com>