msi: Avoid a potential NULL pointer dereference (coccicheck).

[Michael Stefaniuc]

Hans Leidekker wrote:
> On Thu, 2010-05-20 at 10:03 +0200, Paul Vriens wrote:
> 
>>>> The main reason I'm asking is because we have several of these lingering
>>>> around (and marked as 'FALSE' in Coverity).
>>> In this particular case we should just get rid of the goto.
>>>
>>>
>> But is it a (potential) issue? I guess not. If it's not an issue wee can 
>> leave the code as is, not?
> 
> It's a header construction where you typically exploit the fact that the
> header and extended structure start at the same address. So yes, we
> could leave the code as is.
"exploit" doesn't sound like a clean approach to me. The compiler does
the right thing here but there is quite some effort involved for a human
reader to validate that. The reports of good static analyze tools should
more like a hint "there is a problem around that code" and the problem
doesn't have to be the exact same thing the tool actually reported.

bye
	michael