Andrew de Quincey wrote:
I just spent a while trying to figure out how I would go about using the 2.6 request_firmware() calls for the tda1004x driver. As far as I can see, its impossible right now.
The problem is you have to pass a struct device* into request_firmware() to tell the kernel which device you are. Frontend drivers have no idea of which physical device they are attached to.
we could move the 'struct dvb_device *dvbdev' member from the struct dvb_frontend_data to struct dvb_frontend, then it would get passed directly to the ioctl.
oops - sorry, I misunderstood the problem.
To solve it cleanly this would need to rewrite the entire dvbdev code (to register a device/interface/bus/system class 'DVB' so that we can also load MPEG decoder and/or DSP firmwares cleanly -- in those cases we also have no attached predefined bus or device classes, at least not in STBs) but this would again make the entire code unportable and 2.6-only.
I'd vote instead for a politically less correct solution and leave the firmware code as-is and delay this for the v4 code -- this will probably be 2.6-only anyway.
Holger
-- Info: To unsubscribe send a mail to [EMAIL PROTECTED] with "unsubscribe linux-dvb" as subject.
