On Mon, 14 Jul 1997 11:04:44 +1000
"Rudi Bierach (Rudnuts)" <s2171802@cse.unsw.edu.au> writes:

>Jul  7 01:35:44 myname /netbsd: asc_intr: IN: dmalen 10 len 0 fifo 1
>What does this error msg mean please?!?

You can safely ignore this.

I *think* what it means is that the SCSI driver caused an accidental
collision between data coming in from your drive (possibly after a
disconnect/reconnect sequence) and DMAing a new SCSI command from host
memory.  The message is a warning.  Nothing bad has happened.  Michael
Hitch may have a better explanation, but we are sure that the message
is benign. 


(About 18 months ago, it wasn't always; but the message is still there.)

Maybe we should wrap this in an #ifdef DIAGNOSTIC or something.
Though I'd defer to anything Michael Hitch says. Michael?


>also sometimes when i boot the keyboard doesn't work but when i move
>the = >mouse it puts junk characters on the screen in place of the
>keyboard = >input. Its a dec mouse (you know that round no ball mouse)

I have no idea what might cause this. I don't have a Personal
Decstation to investigate this with.  Could it be a hardware problem,
or a poor connection to the DesktopBus mouse?  

Has anyone else seen this problem on a  Personal DECstation
or DECsystem (5000/20, /25, /33, or /150)?


.the cursor in xwindows doesn't seem to move pass a certain point in the =
>x direction, the pointer actually moves because the window focus and =
>menus appear where the graphical cursor should be... very strange
>
>i'm using snapshot 1997-05-25 and the latest x win

This is an old bug in the kernel driver dtop.c which is, I beleive,
now fixed.  (The 1.1 NetBSD/pmax framebuffer code had the horizontal
and vertical bounds switched on all or almost all the framebuffer
drivers).  It's possible that in starting to clean out that old code I
did something stupid and backed out the fix.

I'm building a new snapshot kernel, with Michael Hitch's fix for
polled console input from the Desktopbus and an LK-501.  Could you try
a new snapshot kernel tomorrow-ish?  The new kernel changes the size
of `struct proc' and *will* break /bin/ps and some other commands.

If everything goes well I should have a full snapshot of a -current
userland Tuesday-ish.  I hope not too much more changes between that
and the 1.3 freeze.