On Sun, 16 Feb 2003, Gerd Knorr wrote: > > I agree, which is why I asked. I will stick the card back into my computer > > and see if the drivers also OOPSes at my end. > > Take care. While browsing the code I noticed that they do floating > point math in kernel mode. There are some functions which indicate > that the authors didn't figure what the btand / btor / btaor macros > are good for. They also didn't figure what wait queues are. They use > global variables. Compiling it yields plenty of warnings. Broken > beyond repair :-/ Do you think? I have been getting rid of some of the warnings. I figure that if we can get this code to work, we can start bringing it up to standard. I'm putting in some error checking, maybe I can see what the floating point math is good for and see if there's an easy way around it. If you really think it's useless, I won't invest any time. I really don't mind doing the dirty work and cleaning up the code, so if you think there is any hope, I'll just work on. > The header file with all the register #defines is nice through. There is a lot of stuff missing though. In many places the PCI adresses of the different parts of the chip are hardcoded (audio subsystem, gphp subsystem, etcetera). If think I found the source of the OOPS though. The code is written for 23880 devices which have a feature that the 23881 don't (general purpose hardware port), add to that the fact that the don't do any error checking and you've got yourself an oops. > > Gerd > > -- > Weil die späten Diskussionen nicht mal mehr den Rotwein lohnen. > -- Wacholder in “Melanie” > > > > -- > video4linux-list mailing list > Unsubscribe mailto:video4linux-list-request@xxxxxxxxxx?subject=unsubscribe > https://listman.redhat.com/mailman/listinfo/video4linux-list > Groeten, Dries
