On Mon, 21 Feb 2005, Ivor Hewitt wrote: > Hi, > > Compilation of the current cvs dvb fails with:- > > drivers/media/dvb/frontends/or51211.h:27:19: bt878.h: No such file or > directory > > for me without the attached patch to the include dirs. Hi, Your patch fixes the compilation problem, but the question is, why does the frontend has the bt878.h included? Normally the frontend should not hardwire the device. I don't know if it is possible at all, but what if another vendor decides to use the or51211-frontend on a device which is not driven by a bt-chip. There won't be any bttv_gpio_* functions then. Maybe these parts should be written in a way like its done for the fe_sleep-callback in other drivers (skystar2, dvb-dibusb). Here the actual fe_sleep set by the frontend_attach function is stored in the device-specific-struct and re-set to a device-function. So when the frontend is going to sleep, the device-function is called. Then the device-specific calls are in the device-source, but the fe_sleep can still be reached. I know easier said than done (I saw more complex functions in or51211), but necessary when other devices appear with this frontend. best regards, Patrick.
