Hi Marcel, On Fri, 2006-03-31 at 21:26 +0200, Marcel Siegert wrote: > added new FE_[GET|SET]_FRONTEND_NEW ioctls for new applications. as time goes by the new stuff will become current or old, but the name won't be able to reflect that. What will be the name in case another new ioctl will be introduced in the future? FE_SET_FRONTEND_EVEN_NEWER? I think this is much worse than FE_SET_FRONTEND2 - of course meaning the second version of the ioctl, not the new ioctl for DVB-S2. ;-) Sending your patch inline would make it easier to comment on it. + int act_standard_set; + fe_type_t act_standard; What does act mean here? If it's supposed to mean actual, then I think its a wrong translation from German, and should mean "current". What's the difference between those two variables? + err = fe->ops->get_frontend_new(fe, (struct dvb_frontend_parameters_new*) parg); In case parg is of type void *, then the cast is unneeded. Casts should always be avoided from and to void *, because they possibly modify annotations used by the "sparse" source code checker. > just wondering if the size member ( or a fe_type_t member ) or could be automatically assigned after FE_SET_STANDARD + /* be sure struct union pointer is set - even if it is NULL */ Why not just return -EINVAL and do nothing instead? + fetunesettings.parameters_new.u = (struct frontend_parameters_union *) kmalloc(castedparg->size, GFP_KERNEL); What happens, if a malicious TV consumer sets size to a value which kmalloc can't handle? Regards, Andreas _______________________________________________ linux-dvb@xxxxxxxxxxx http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
