I currently have three problems with NetBSD plus one small problem.

Problem 1: serial ports.
I am having problems with reliable transfers.
I have used 386 machine running SCO with two ports locked at 38400
doing downloads with high load and no dropped characters with 16550.
But under NetBSD with a 486-33dx locked at 38400, I am getting LOTS
(hundreds in a few minutes) of silo overflows, and I presume lost characters,
since transfers are very unreliable.  A friend uses FreeBSD on a 40 meg disk
as a "SLIP router" locked at 115200, without problems.
I looked at porting FreeBSD sio.c, but it used different kernel structures
and I could not port it in about the 30-45 minutes I tried.
Is there a better com.c; is it a problem only I am seeing?

Problem 2: ifconfig ed0 ns 123: = crash.
If I do an ifconfig ed0 ns 123:, I get a kernel debug prompt
of "Warning, spl not lowered after system call" message on a NE2000.
I have traced it down as much as I can.
The NS code looks just like the IP code, and the IP code works,
but the NS codes does not. 
I have put #defs for spl*() to do printf's.
All the numbers "look" right.
It gets splimp()'s twice and splx()'s twice.
If anyone else with more spl*() and kernel network driver experience
would like to help me, I would help as much as I can.  If you have pointers,
I can look into as I have exhausted my resources.

Problem 3: ESDI 600 meg with 10 meg of bad sectors on UltraStor controller.
I have a drive with 16xxc, 15h, 3xs that nothing but DOS and OS/2 like,
and I believe it is because of bad sectors on the disk.
It will hit read errors on sectors like (1321c 8h 12s) and die during mkfs.
Linux, FreeBSD, NetBSD nothing unix will run.
track mapping on or off.
using 1500c 12h 28s (smaller than real size values) will not run.
I have tried to lock out the bad sectors but they stay.
anyone know any suggestions?





I have not used any of the *BSD or linux since 386bsd-0.1 and NetBSD-0.8
Linux had major networking problems and *BSD had problems staying up without
panics.

But 1.0 has been up since released and only once crashed under extreme swap
load.  It swapped until it died and the disk quit moving.  It would respond 
to pings, but either init was dead, inetd was not working, or something 
because it would not accept remote connections and Xfree86 was hung on the
console.  This is my small problem.


Risner