> Michael van Elst writes:
>
>...
>
>
> To me that does not qualify as crap.
>
> Saying that: I did replace a vr0 card in another computer because
> of many problems. The _same_ problems showed up on the EPIA-M
> machine (aka pussyfoot). The problems went away after I applied
> the renegotiation patch. I haven't tested the old vr0 card since
> then.
>
> I have seen reports about other problems, mostly related to PCI-DMA
> on some cards that also use a VIA Rhine chip. But on the EPIA-M
> board the chip seems to work perfectly.
>
>

Actually I do not have a deep knowledge of the chip and the device driver.

But I've made a test on similar hardware:
 1) CPU VIA EDEN 533MHz, 256 MB RAM, 2 Realtek;
 2) CPU VIA EDEN 600MHz, 256 MB RAM, 2 Via Rhine (a VT6102 Rhine II, and
VT6105 Rhine III chip);

The same O.S. was used: FreeBSD 4.9 (on NetBSD 1.6.1, the Rhine III was
not usable); the box was configured as a bridge between two LAN.

Result:
 - on low/medium LAN traffic no difference noticeable;
 - on hight LAN traffic (several download between the 2 LAN) the Via Rhine
cards show to me a delay in the interactive sessions (ex. terminal
server, ssh);

So it seems to me even the infamous realtek cards perform better than Via
Rhine.

Roberto