> 
> 	I'm running -current as of about 2 weeks ago & trying to build netbsd
> 	onto a 4/300 (using an ELC as a server). Its using the vanilla 'SUN4'
> 	kernel.
> 
> 	Can boot it up ok, but when I tried newfs'ing the disk (a wren IV,
> 	two of whick work fine in the ELC) it blew up when I tried 
> 	'newfs sd0g' - every time (sd0a & sd0h are fine).
> 
> 	It hands for a couple of seconds then:
> 
> CSR = 1118
> panic: DMAWAIT1
> syncing disks... data fault: pc=f8033050 addr=200024 ser=80<INVAL>
> panic: kernel fault
> Frame pointer is at 0xf80bb3f0
> Call traceback:
> 	<I'll type it all in if needed :>


   I had a lot of problems with this also.
   In fact, I think I mentioned my problem about a month ago..

   Theo Deraadt and Chuck Cranor have been rewriting the esp driver to 
   fix this, and other serious problems.  I've tested their new code
   as of a few days ago and it had (apparently) fixed many of the 
   problems that I was experiencing. 

   I have a 4/330 with 5 scsi devices on it, and let me tell you, without
   those fixes there's no way you're going to run for long.

   I would like to suggest that we get Theo's driver rewrite on a couple
   more people's machines, and then see if we can't replace the old one.  
   Theo's been working on SCSI disconnect/reconnect, as well as sync 
   and some other features that the old driver NEVER handled. 

   Theo's driver is more or less a rewrite, there are so many changes
   involved. 

   I'm working on some framebuffer drivers with Theo to drive
   P4 based cg4 and cg8 cards, and until they are working I'm still
   switching back and forth between NetBSD and SunOS.  We're close on 
   the drivers, need to update rconsole to handle 24 bit framebuffers,
   and some other things that Theo mentioned, and I'm debugging a 
   problem with X retrying on the mouse indefinitely..  But we have got
   images up on the screens :-)  Things like 'screenload' work fine,
   but there are ioctl() calls that need to be added to the cg8 driver
   before a lot of things will run. 

   As before, the driver code that Theo had last week seemed to fix my DMAWAIT1 
   CSR=1118 panics.  Maybe Theo can fill the group in on what he's been
   working on, I'm sure I've left out details..

> 	On reboot it keeps panicing - it gets to the esp check & fails, and
> 	will keep on failing until I power off & back on.
> 
> esp0 at obio0 addr fa000000 pri 4: ESP100  <- note no newline, it 
> 					   hangs for about 3 secs & then
> 					   from CSR above
> 
> 	a) Is this a known problem?

   Known to myself, Chuck Cranor, and Theo, don't know if anyone else is
   running 4/300 machines.  Anyone else out there????
  
   Hey, do you have any P4 video boards on your machine(s)? 
   

> 	b) Can I do anything to help track it down?

   I think Theo and Chuck have figured it out.

> 	c) Do I just go away & leave my 4/300 running sunos until the next
> 	   round of sparc changes come out? :)

   No :-)   We need to get Theo's scsi driver in the tree.  It should REALLY
   outperform the old driver, as well as fix this and other problems. 

   Besides the SCSI, and framebuffer drivers, Theo and Chuck and I have
   been working on 4/100 updates.  I have a 4/110 running NetBSD with
   some changes beyond what is in -current.   

> 			David
> 
>   D.K.Brownlee@city.ac.uk (MIME).  Network Analyst  {post,host}master  abs.
>               (>=-=<)   Telephone: (+44) 171 477 8186.   (>=-=<)
>   Computing Services, City University, Northampton Square, London EC1V 0HB.
>        .-------------------------------------------------------------.
>       / Monochrome - biggest UK internet BBS - telnet mono.city.ac.uk \
>       `---------------------------------------------------------------'

   John Stone
   johns@cs.umr.edu