The "Odd IO request" message is more for my benefit than for anyone else. You're right that it means the "lines" field in the CIS is incorrect. But fixing it won't do any good, because the drivers already noticed it was odd and took appropriate action. At this point, since you've got another serial device that works fine in exactly the same context (same IO, same irq), and this card also claims to be a perfectly ordinary serial device, I'm not sure what could be wrong. The kernel serial driver is not happy with this card. Maybe there's some mild incompatibility that isn't noticed by the Windows serial driver, but screws up the Linux serial device probe. Or maybe it is a timing problem in the initialization code. But I think the "Odd IO request" is a red herring. -- Dave |
Messages