Uh, replying to myself... :)

After having stared at the FreeBSD CVS history of the re driver a good =20
part of today, I tried to apply a change they made somewhere in =20
september 2004 and that fixes the problem for me. Hooray! I'll send a =20
pr tomorrow.

thanks for listening and goodnight
dieter


On 06/16/06 16:27:17, dieter roelants wrote:
>=20
> Hi,
>=20
> I have this new motherboard (yay!) with onboard realtek 8168B nic. =20
> After adding the PCI ids to the re driver(*), the card starts to =20
> attach but then prints the following:
>=20
> re0: WARNING, DMA FAILURE!
> re0: expected TX data: 00:68:65:6c:6c:6f/00:77:6f:72:6c:64/0x800
> re0: received RX data: 65:6c:6c:6f/00:77:6f:72:6c:64:08:00/0x0
> re0: You may have a defective 32-bit NIC plugged into a 64-bit PCI =20
> slot.
> re0: Please re-install the NIC in a 32-bit slot for proper operation.
> re0: Read the re(4) man page for more details.
>=20
> I don't think that's what going on here because the nic seems to work =20
> fine with a FreeBSD 7-something-current bootcd. It also looks odd to =20
> me that the received data is the same as the sent data, only 2 bytes =20
> shifted. That's what I see with tcpdump too(**); packets sent from =20
> the system arrive fine, but it "sees" the reply packets with the =20
> first 2 bytes off.
>=20
> Can anyone please help with bright ideas of what could be causing =20
> this, or what doesn't cause this in FreeBSD or maybe even a patch to =20
> try?
>=20
> Thanks,
> dieter
>=20
>=20
> (*) That was easy enough, but I had to look at the FreeBSD driver to =20
> find the hardware revision. How does one find out?
> (**) after having disabled the test, so the card would finish =20
> attaching
>=20