Erik Fair wrote:
> NetBSD 3.1_STABLE works for me without IDE complaint on the Sun Netra T1
> AC200, and its successor, the Sun Fire V120, with the caveat that I have
> not used optical media on these systems yet, just boot/probe/autoconfig.
> I can send dmesg.txt from either system if asked.
> 
> So, what changed between 3.1 and 4?

I'm not sure what changed, but the contents of the bug claims that on
machines where it doesn't work, it doesn't work as far back as 1.6.  I'm
going to try 3.1_STABLE on my AC200 and will report what I see.

--rafal

For GNATS benefit, here's the original comments I added, that somehow
didn't get into the PR (either I sent to the wrong address or GNATS
somehow ate my original attempts):

I'm seeing this same issue on a "new" (to me) Netra T1 AC200.  A couple
of interesting additions:

1. aceride0 (the on-board IDE controller supporting the CD-ROM) is also
complaining of lost-interrupt messages during the bus probe.  I presume
this and the SCSI timeout are caused by some system-wide interrupt
routing issue.

2. The hardware is fine; OpenBSD 4.2 installed on it A-OK (I didn't have
Solaris media and wanted to verify that all system components were happy
before my eBay non-DOA warranty ran out ;)).

This was with a INSTALL kernel from a 2007-11-23 HEAD snapshot:

NetBSD 4.99.37 (INSTALL) #0: Fri Nov 23 07:40:58 PST 2007
builds@wb44:/home/builds/ab/HEAD/sparc64/200711230002Z-obj/home/builds/ab/HEAD/src/sys/arch/sparc64/compile/INSTALL