2005-05-19 12:04:48 UTC
Ok, it seems that the driver uses the Universe register definitions when talking to the PCI config space, rather than the linux defines in linux/pci.h. Also, it appears that the MVME2400 boards only implement the first baseaddress register space (memory mapped i/o), because the second base address register (i/o mapped) doesn't seem to work. Also, the driver tries to write 0xA000000 to the first baseaddress register, but the linux boot-up system has already entered a value that won't conflict with any resources (so it should leave the contents alone!).
I can now load the driver. Now I need to test it.