Re: Cnet 611 support
Joerg Albert <[email protected]>
| Newsgroups | gmane.linux.drivers.at76c503a.user |
|---|---|
| Message-ID | <Pine.LNX.4.56.0401212311220.512@invb> |
Hi, On Tue, 20 Jan 2004, sfg wrote: > OK I got it! 1371/0001 Device is now pinging back after lots of time and > hair pulling. > > Problem was that Mandrake 9.2 puts and expects its drivers for this at > series in a directory called /3rdParty/at76c503 not /usb. This meant > that the usbdfu.c routine knew it needed to load the driver but could > not find it (I think). More debug messages would have saved me lots of > time here as I am not a competent driver person at all. > > I made this change to the CVS Makefile and it loaded the modules where I > needed them: > > Makefile: > > # SFG altered from usb to 3rdParty/at76c503 for Mandrake 9.2 > #OLD ONE MODULE_DIR = > $(DESTDIR)/lib/modules/$(KERNEL_VERSION)/kernel/drivers/usb > new one >>>MODULE_DIR = > $(DESTDIR)/lib/modules/$(KERNEL_VERSION)/kernel/drivers/3rdParty/at76c503 I'd rather asssume that the old driver version (delivered as part of Mandrake 9.2) resides in 3rdParty/at76c503 and the Makefile of the Berlios sources put the new version into ../usb. If you get two drivers for the same USB id pair, it depends on depmod's scan order, which one will be put first in the modules.usbmap and get activated later. Maybe "make install" should first remove any driver files found under /lib/modules/`umask`/ before installing itself? /Jörg