I've been having three problems with NetBSD/sparc -current (kernel from 
July 6, binaries up-to-date as of July 12). This is on an SS1 with 16M 
RAM, and a 540M Quantum Lightning on sd0 and a Maxtor 105 on sd1 (sd0 is 
netbsd, sd1 is 4.4BSD).

1) I keep getting console messages that arplookup is unable to enter 
addresses, and that it can't allocate llinfo. Often, when rlogin'd, the 
machine will appear to freeze; it cannot be pinged at this point. Within 
a minute or less, though, it returns to life. There are other users and 
an essentially 0 load, so that's not it. While its frozen remotely, the 
console is fine. Here's the syslog excerpt:

Jul 10 12:16:10 margo /netbsd: arplookup: unable to enter address for 8cf73c14
Jul 10 12:19:10 margo /netbsd: arplookup: unable to enter address for 8cf73c29
Jul 10 12:26:10 margo /netbsd: arplookup: unable to enter address for 8cf73c14
Jul 10 12:29:11 margo /netbsd: arplookup: unable to enter address for 8cf73c29
Jul 10 12:32:45 margo /netbsd: arpresolve: can't allocate llinfo
Jul 10 12:32:52 margo last message repeated 13 times
Jul 10 12:32:53 margo /netbsd: arpresolve: can't allocate 
llinfo<7>arpresolve: c

This problem started when I rebuilt the binaries; the kernel worked fine 
with the Jun 19 snapshot binaries. Note that this problem does not exist 
under 4.4BSD on the same machine.

2) When I try to shutdown, halt, or reboot the machine, it panics on 
"DMAWAIT1" when it tries to sync the disks. This happens every time, even 
if I boot and immediately halt. Here's the panic:

Jul 12 12:01:41 margo /netbsd: syncing disks... CSR = 80001118
Jul 12 12:01:41 margo /netbsd: panic: DMAWAIT1
Jul 12 12:01:41 margo /netbsd: Frame pointer is at 0xf80d7938
Jul 12 12:01:41 margo /netbsd: Call traceback:
Jul 12 12:01:41 margo /netbsd:   pc = f80bc974  args = (300, 400fe1, f80e7000, f80d7a58, ffff, 0) fp = f80d79a0
Jul 12 12:01:41 margo /netbsd:   pc = f801c660  args = (100, f80de400, 1, f80d7abc, f80f3c00, f80f4c00) fp = f80d7a08
Jul 12 12:01:42 margo /netbsd:   pc = f80ab8b4  args = (f80ab680, 100, 1, f8e08000, 0, f8e08538) fp = f80d7a70
Jul 12 12:01:42 margo /netbsd:   pc = f80adf3c  args = (f8592b80, 0, 800, 0, f80ab798, 1ff) fp = f80d7ae0
Jul 12 12:01:42 margo /netbsd:   pc = f8006c50  args = (f8596800, f80ad8a8, ffffffff, 1, 0, 0) fp = f80d7b50
Jul 12 12:01:42 margo /netbsd:   pc = f801a984  args = (0, 0, f80086ac, 0, ffffffff, 4) fp = f97639b8
Jul 12 12:01:42 margo /netbsd: 
Jul 12 12:01:42 margo /netbsd: dumping to dev 701, offset 8

3) If the machine has panic'd and dumped, when I reboot savecore locates 
the core file, and then immediately locks hard when it tries to save it. 
Occasionally I can L1-A out. 

Are these known problems? Are there any fixes?

Thanks for any info.

--Aaron

--
Aaron B. Brown
Harvard University
abrown@eecs.harvard.edu