Erik E. Fair said:

>
> Among the other things that I did when I encountered the problem
> reported in that PR is make a commit (with Jason Thorpe's approval)
> to the AMD PNA PHY code that correctly marks the HomePNA PHY as
> HomePNA rather than 10baseT as reported above. It's in current.
>
> Please note that the AMD 79c901 chip has two PHYs on it - the second
> one is a standard 10baseT and is on the second "instance" as we call
> it, or the higher PHY number.
>

No more panic:

vr0 at pci0 dev 7 function 0: VIA VT6102 (Rhine II) 10/100 Ethernet
vr0: interrupting at kn20aa irq 8
vr0: Ethernet address: 00:e0:98:9d:e2:1c
pnaphy0 at vr0 phy 2: Am79C901 HomePNA 1.0 interface, rev. 4
pnaphy0: HomePNA1
ukphy0 at vr0 phy 3: Generic IEEE 802.3u media interface
ukphy0: Am79C901 10BASE-T interface (OUI 0x000058, model 0x0037), rev. 1
ukphy0: 10baseT, 10baseT-FDX, auto

I'm not sure about this:
vr0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        address: 00:e0:98:9d:e2:1c
        media: Ethernet autoselect instance 1 (none)
        status: no carrier
        inet6 fe80::2e0:98ff:fe9d:e21c%vr0 prefixlen 64 scopeid 0x2

Does the instance calculation start from 0 or 1 ? Am I looking at
pnaphy or ukphy at the moment ? (or back then.. I have to rip the
card from my pppoe-gateway everytime I want to try it on the alpha)

OK, assuming that instance 1 is pnaphy (otherwise ignore the rest of the
email)

I couldn't get vr0 status to change from "no carrier" no matter what
I did. Naturally this hindered my chances of getting a working pppoe
configuration over it.

You did mention in the PR that you were not able to get the link up,
is that still the case ?

Thanks,
-- 
jht