>
> I've seen two possible explanations:
> 
> a.) a bad batch of WRAPS was supposedly produce around Sept 2004.
>     (this seems a bit unlikely as 1.2b2 seems to work fine on the
>      affected WRAPs)

since I said this, I will clarify as much as I can, some WRAP 128mb/3 
lan boards manufactured between july and august 04, were recalled late 
august due to a wrong resistor being wired to one of the nics, as I 
can't seem to find the mail at the minute, I can't say precisely what 
the error was nor how it manifested itself, although I will speculate 
that it had an adverse effect on that nic's performance under certain 
conditions ;)

if you bought directly from pcengines then you got an email asking to 
send the boards back in, if you bought from a third-party or off ebay or 
something then your mileage may vary.

I think this was a small batch and the problem was rapidly diagnosed, 
however, there may still be a few floating around somewhere. If in 
doubt, I would suggest asking pcengines directly with your hardware 
revision numbers.

bridging also tends to bring out problems on some buggy nics/drivers, I 
remember fustration trying to set up my first openbsd filtering bridge 
using a TI card which was known NOT to work bridged, as always, RTFM first.

Fyi, I have a WRAP board running fine in bridge mode under OpenBSD since 
November. So the DP8316 can at least be bridged...

> b.) perhaps the kernel HZ bump between 1.2b2 and 1.2b3 to 1000hz
>     is bringing out flaws in some WRAP older/buggy hardware/firmware?
> 
> It might help if we could collect the BIOS revision and hardware
> revision numbers of the affected boards.

I unfortunately don't have access to this info right now.

cheers,

mike