Dear M3100 Vax People,
   Ragge gave me a different patch which adds in the offset to two places.  
With this patch I have been able to have a M3100 with drives on either the
SCSI A bus (internal) or SCSI B bus (external) or both without problems.
   Maybe we can put this problem behind us now.
johnh....   

On Wed, 4 Oct 2000, John Hayward wrote:

> Date: Wed, 4 Oct 2000 14:57:24 -0500 (CDT)
> From: John Hayward <John.C.Hayward@wheaton.edu>
> To: David Brownlee <abs@netbsd.org>
> Cc: Chuck McManis <cmcmanis@mcmanis.com>, port-vax@netbsd.org
> Subject: Re: boot blocks - (really DMA M3100/M10 problem)
> 
> Dear M3100 Vax People,
> 
> Sorry to report that when I attached the B chain to my computer with the
> internal drive on the A chain that I experienced problems.  I copied
> the kernel from the A chain to the B chain then did diff and chksum on
> these files.  They differed.  I could copy a file on the A chain to
> another place and do a diff and things appeared to be ok.  Copying files
> appeared not to work properly on the B chain to B chain.
> 
> Taking the B chain drive back to the original M3100 and that system seemed
> to boot and work fine.
> 
> The machine with the A chain drive would not boot off of the
> drive.  Net-booting then attempting to do a fsck resulted in lots of
> errors.
> 
> The machine with the A chain drives was a working system but it seems to
> loose it's date and boot flags when powered down.  There were several
> files which I thought I had copied when I set up the machine which did not
> appear - this was before attached the B chain drives - so I don't have
> 100% confidence in this machine.
> 
> Bottom line - The fix I have seems to work best with only having drives on
> the B chain - NOTE: A chain drives may become corrupted.
> 
> I'm doing a low level format at this point and attempt to observe behavior
> after re-installing.
> 
> johnh...
> 
> 
>  On Tue, 3 Oct 2000, John Hayward wrote:
> 
> > Date: Tue, 3 Oct 2000 23:44:22 -0500 (CDT)
> > From: John Hayward <johnh@ImapStu.wheaton.edu>
> > To: David Brownlee <abs@netbsd.org>
> > Cc: Chuck McManis <cmcmanis@mcmanis.com>, port-vax@netbsd.org
> > Subject: Re: boot blocks - (really DMA M3100/M10 problem)
> > 
> > Today I set up a second M3100/M10 which has an internal scsi drive on the
> > A chain.  The scsi drive is working without problems.  Maybe tomorrow I
> > can attach the external drive and try to see if there are no problems with
> > drives on both chains.
> > 
> > I did note some messages relative to a CD drive about sensing when
> > booting (which was on the B chain).
> > 
> > johnh...
> > 
> > On Tue, 3 Oct 2000, David Brownlee wrote:
> > 
> > > Date: Tue, 3 Oct 2000 15:17:55 +0100 (BST)
> > > From: David Brownlee <abs@netbsd.org>
> > > To: John Hayward <John.C.Hayward@wheaton.edu>
> > > Cc: Chuck McManis <cmcmanis@mcmanis.com>, port-vax@netbsd.org
> > > Subject: Re: boot blocks
> > > 
> > > On Tue, 3 Oct 2000, John Hayward wrote:
> > > 
> > > > I had a couple of problems - I had unpacked the i386 ALPHA2 sets on my vax
> > > > (-{ (big problem) and in switching between net booting and attempting to
> > > > boot off the disk entered netbsd for the boot file when the boot flag was
> > > > on to prompt for the boot file.  I have since unpacked the vax sets, a
> > > > kernel which has the dma offset removed and turned off the boot flag and I
> > > > can boot ALPHA2 on my M3100.  It seems faster than running of the network
> > > > (maybe because I could hear the disks doing stuff).
> > > > 
> > > 	Excellent news :)
> > > 
> > > > I'd still like to know if any one has docs on how the DMA should work on
> > > > the 3100 - I'm concerned that the tkxx drive hanging off the scsiA chain
> > > > might have dma problems with disks doing dma on the scsiB chain.  I need
> > > > to clean my tkxx drives to see if I can get them to work but this is a
> > > > lower priority item for me.
> > > 
> > > 	Do you have a spare disk you can hang off the scsiA chain to
> > > 	test them both in parallel?
> > > 
> > > 
> > 
> > 
> 
>