Re: [patch] removes casting of (void*) private structure members

["Philipp Marek" <philipp@marek.priv.at>]

Hello Sam,

you wrote:
> Sample from your patch:
...
> Therefore the cast is legitimate.
>
> And I expect gcc to emit:
> warning: assignment makes pointer from integer without a cast
Possibly mine didn't ... I'm getting some build errors related to static
variables being initialized via htonl(), but I did a compile run with this
patch.

I think I used a clean kernel for testing - but maybe I was in the tree
where I tested converting some "unsigned long priv;" to "void*", looking
for problems - as I wrote:
> > There are some private structure variables that are not defined as
> > (void*), (but int or long) and cannot be done without a
> > cast; examples in
> >   drivers/char/hw_random/intel-rng.c
...
> > Should these be converted to (void*), and an additional
> > patch supplied?
If I do a patch for them, can that be a trivial patch, or should I send
each type conversion to the maintainer?


Regards,

Phil