Hi Geert,
b/drivers/net/atari_91C111.c | 191 ++++++++++++++++++++++++++++++++++++++++---Euh, the patch doesn't provide the whole drivers/net/atari_91C111.c; that part seems to be incremental?
Incremental to smc91Cx.c - I probably made you patch that one instead of creating atari_91C111.c as I intendend to. Sorry. patch sometimes gets ahead of me.
Can you please resend the full patch?
Coming soon.
Apparently, the EtherNAT has no station addres (MAC) PROM on board, so you will have to set the MAC address manually, using ifconfig eth%d hw ether 00:00:de:ad:be:efI think there's infrastructure for that. Ah, random_ether_addr() in include/linux/etherdevice.h
Sweet!
Sending and receiving data via EtherNAT works to some degree, that is, I can ping remote hosts (but receive duplicate reply packets from the local net when I have the EtherNEC up at the same time). I cannot use slogin (which may be due to the dups, or because I use the same local net for both interfaces without properly separating the networks).Do EtherNAT and EtherNEC share the same interrupt?
Yep. I think we still have one MFP timer interrupt left that I could use instead of sharing - I'd have to find the specs for MFP timer programming to actually use it (short of brute force guessing the values).
Do you have both of them connected to the same local net?
Mostly. I tried hooking up my powerbook directly using a crossover cable (and using a different network), didn't change anything. I noticed that the destination MAC address as shown on the wire is broken (either byte swapped or null byte inserted), most likely there's some endianness issue with word access in my ISA IO patch. I'll work on that some more as soon as I can take hobbes' buildd down again.
Michael