Hello,

Now that I've got NetBSD 1.3.3 running on two SS1+'s, I seem to get these
messages every now and then:

Jan 20 12:53:54 pyy /netbsd: (esp0:0:0): select; 1 left in DMA buffer [intr
18, stat 11, step 4]

I asked about them before and got an answer that they may not be dangerous.
If this is the case, maybe I should just comment out the printf in kernel
and forget about them? ;)

Or is there a fix to this "problem"?

One of these machines has one SCSI disk (quantum viking), the other has two
(quantum fireball and seagate medalist (?)).  I first thought the problem
was only with the system with two disks, but that doesn't seem to be the
case.

BTW, the 1.3.3 SCSI driver is much better than the 1.2 one, at least when
it comes to reselection.  With the 1.2 driver, the system paniced when I
tried to access two disks simultaneously, but under 1.3.3, I haven't seen
such a problem so far.

  -jm