After installing 1.4 on a few different Sun-3 units, here are some odd
problems I'm seeing:

   1) There appears to be something going on in the TCP/IP stack. When
connected to certain other systems, incoming data will sometimes simply
stall, forever. For those systems for which this occurs at all, it is
easy to reproduce. Sending a packet on the connection will restart the
flow of data. As far as I've been able to tell, select() is doing the
right thing, and there is no data waiting in buffers, it just stops
coming in from the network. I have not had time to examine the problem
with a packet sniffer yet. It occurs on both a 3/60 and a 3/4xx, so it's
independent of the ethernet driver (le vs. ie) and not specific to the
kernel architecture either. Has anybody else seen this?

   2) I'm getting occasional SCSI timeouts on the 3/4xx with a si board.
This wouldn't be so bad except that the machine locks up for about five
minutes before recovering, and sometimes it doesn't recover at all.

   3) I'm also getting occasional "ie0: DMA underflow" messages on the
3/4xx. Haven't seen it on the 3/50s and 3/60s, so it may be in the ie
driver. This appears to be harmless, since the data always seems to be
intact. It shows up when transfering very large (dozens of megabytes) files
via FTP.

   4) Is the ECC on 3/4xx memory supported? I came home one day to find an
unending and uninterruptible stream of messages about ipl_autovec 7 being
unclaimed scrolling up the console. I suspect the ECC memory here because I
notice that in the boot messages it is listed as having ipl 7.

   Thanks in advance for any light that anybody can shed on these subjects.

   --James