Re: AMD PCnet/PCI Fast Ethernet driver issues in 5.3RC1 / 5.3-STABLE

quoting "Ketrien I. Saihr-Kesenchedra" <ketrien_at_error404.nls.net>:

> Daniel Engberg wrote:
>
> >Hi, the dmesg output is the following.
> >pcn0: <AMD PCnet/PCI 10/100BaseTX> port 0xe400-0xe41f mem
> >0xe2001000-0xe200101f irq 11 at device 10.0 on pci0
> >miibus0: <MII bus> on pcn0
> >lxtphy0: <LXT970 10/100 media interface> on miibus0
> >lxtphy0:  100baseFX, 100baseFX-FDX, 10baseT, 10baseT-FDX, 100baseTX,
> >100baseTX-FDX, audio
> >pcn0: Ethernet address: 00:60:b0:eb:c3:c6
> >pcn0: if_start running deferred for Giant
> >pcn0: [GIANT-LOCKED]
> >
> I must have missed earlier posts in this thread. (Not a surprise, I
> usually only have time to skip over -current.) The problem you are
> describing here is nearly identical to pcm-as-module here, with a
> generic/unknown PHY. I haven't had time to properly debug, since the pcn
> on the system in question is a) onboard and b) the only working NIC with
> driver currently in the system. The symptoms I saw were specifically
> that if pcn was compiled as module and loaded, it seemed not to attach
> PHY properly. Can you try compiling pcn as a module, and seeing if the
> behavior is identical on load? Then try unloading and reloading the
driver?
>
> -ksaihr
> 

Hmm, unfortunately your tip didn't solve my issue but it was worth a
try. I recompiled the kernel without pcn and lnc (since lnc will attach
if pcn is not found) and here's how it went:

kldload if_pcn
ifconfig pcn0 inet 192.168.1.40 netmask 255.255.255.0
ping 192.168.1.1 - 100% loss
kldunload if_pcn
kldload if_pcn
ifconfig pcn0 inet 192.168.1.40 netmask 255.255.255.0
ping 192.168.1.1 - 100% loss
kldunload pcn
kldload if_lnc
ifconfig lnc1 inet 192.168.1.40 netmask 255.255.255.0 - lnc1:
Initialisation failed

If you reboot and load lnc without touching the pcn driver the nic works
fine (including ping etc).
//Danne