Subject: 2 problems with i386 version of V1.3 (not 1.3.1)
To: None <port-i386@NetBSD.ORG>
From: Martin Grossman <grossman@BBN.COM>
List: port-i386
Date: 03/23/1998 08:49:44
First, This is my first msg on this list so please be forgiving!

I have two problems.....

FIRST,

I've tried 3 times to do a fresh install of v1.3!
First time was via ftp (over ethernet) from a host 2 feet away.
Everything went fine untill the reboot (ie to boot off the hard disk)
well, it didn't boot!  After 2 or three reboots from the boot floppy
I figured out that boot sectors were not setup correctly via installboot!
The disk boot sector had its original code in it (from BSDI OS) but the
partition tables were setup correctly (ie 1-blank 2-blank 3-blank 4-netbsd

This was accessed via wd0d which covers the whole disk!
The correct code (boot code) was placed in wd0a's boot sector which
starts 1 head down (ie cyl 0 head 1 sector 1)!

SO, in each of the three tries (on different size drives) each via a different
method (ie 1-ftp 2-NFS 3-floppies), I fixed things via...
   boot from boot floppy, cd /tmp, dd if=/dev/rwd0a, dd if=/dev/rwd0d,
   dd correct partition table part of 1 sector file into correct code in the
   other sector file (ie combine parts of both), then dd fixed boot sector
   to both rwd0a & rwd0d.....then system boots fine!

SECOND PROBLEM,
I have a adaptek 2940UW board (on pci bus).
When a exabyte 4200 (4mm no compression) is hooked up everything works fine.
When a exabyte 8200 (8mm no compression) is hooked up everything works (ie
I can write to the tape with tar, and can read everything back in via tar)
BUT I'm getting lots of "Target Busy" messages on the console.
a) file dev/ic/aic7xxx.c
   subroutine ahc_handle_seqint()
	switch (intstat&SEQINT_MASK)
	case BAD_STATUS
		switch(AHC_INB(ahc,SCB_TARGET_STATUS)
		case SCSI_BUSY
			printf("Target Busy");
b) I've tried a second Exabyte 8200 and still get the same printouts!
c) I've tried LINUX v5.0 (Redhat) and guess what....I get the same printouts!
d) I have not tried putting BSDI (2.0 or 3.0 or 3.1) yet!

SO, its realy not a netbsd problem!  BUT, can anyone explain why this is
happening (or better yet) what it means?  As far as I follow the netbsd code
it looks like the tape drive is busy, and we can't even give the scsi board
another queued command since its busy too!