I have been very pleased lately to get my Beige G3 (OF 2.4) up
and running reliably with NetBSD 1.6.2.  This new machine is slated
to replace my Power Center 132 as my main web server (woo hoo!)
As expected, 1.6.2 is
reliable and efficient.  I now need to get some essential services
working, and I have been running into roadblocks.

1. bind appears to be busted.  For some reason, I get lots of log
messages complaining about "Malformed response".  When I download
and build a newer version of bind, these messages go away.  The
volume of these messages makes me nervous about using the stock bind on
a machine that will be a web server, so I plan to run 8.4.4.

2. The cyclades 8-port serial card driver is still busted.  In its virgin
state, it still has a problem where it gets "stuck", having "forgotten"
an interrupt.  I have put in an ugly kludge, (and I do mean *ugly*)
This is the same kludge I put in a couple of years ago when I
ran into the same problem on 1.5.2.  The problem appears to be some
sort of race condition in the bowels of the interrupt machinery...

3. Having put the kludge in for the cyclades driver, I find that pppd
is also busted.  Again, I went to the 1.5.2 pppd that I patched, and built
it with my fixes.  This got me logging in with ppp.

4. Having gotten login to work, I find that pppd does not set up routing
when it sets up the ppp interface.  I worked around that with an explicit
"route add dialup1 -interface mymachine".

My question is: What is proper procedure to get these problems - for which
I have kludges, if not fixes - addressed?  My enthusiasm for digging in and
properly fixing these problems is proportional to the likelihood that the
fixes will be "committed", and inversely proportional to the amount of effort
I will have to expend to get that commitment.

Help?

Thanks,

-dgl-