>Thats wy i change in the etherboot pci.h the Winbond id to 0x10bd and >the device id to 0xe34. And now the floppyrom detect the card as an >Winbond card and found the address. But >Probing...[NE*000/PCI]< hangs. Those pesky NE2000 clones... %*^&*! They seem sensitive to I/O timing. You could try replacing the outb calls in the probe routine with outb_p but I don't guarantee it. Another reason may be this card requires some other sequence of I/O handling, i.e. you have to use the mfr's driver and not a generic NE2000 driver. I cannot tell without any specs on the card. You could try a packet driver with the Netboot package instead. >I ask Ken and he told me that the address from the pci detection is >written in an array and passed to the driver; thats alright. >Now i compared the ne.c from Donald Becker with the ns8390.c respecting >of initialisation for PCI clones. There are many differences. I'm >meaning the source from the kernel is better because working and of >course bigger but i think not to big for an eprom? The sources are different because they are not related to each other. Etherboot is derived from the FreeBSD netboot and shares nothing in common with the Linux drivers. I am however working on a project to use Linux drivers for netbooting. The project is called NILO and the Web page is http://www.slug.org.au/etherboot/NILO/ There is no stuff yet for users and I can't promise any dates. I am waiting for the next release of OSKit (which has some libraries for encapsulating Linux drivers). But there will be in the next week or so, a Linux user level program for simulating a network boot, exercising bootp and tftpd.
For requests or suggestions regarding this mailing list archive please write to netboot@gkminix.han.de.