Subject: Re: Anyone running a Sparc5/170 ??
To: None <port-sparc@NetBSD.ORG>
From: Greg Earle <earle@isolar.Tujunga.CA.US>
Date: 07/01/1997 15:43:31
Since Dave and Neil both asked me:
> On June 30, you wrote:
>> (Who continues to be amazed at how rock-solid his SS20/71 is under
>> 1.2.1 w/ only a MAXTSIZ bump + mb_map fix and pmap fixes added)
> Hmm? Which fixes are these?
I bumped up MAXTSIZ from 8 Mb (1.2.1) to 64 Mb (-current) so I could run
Netscape 4.0beta5. (Actually, I just updated sys/arch/sparc/include/vmparam.h
to whatever was -current at the time)
I bumped up NMBCLUSTERS in my kernel config file because I was getting
a rash of "mb_map full" errors after being up for awhile.
And finally, I applied a pmap.c patch from Aaron Brown that he posted here
a while back (check the archives) after people reported getting random SIGSEGV
crashes. I would presume that one is in -current by now. I've certainly
never seen any random program crashes since applying it.
(The only other change I've made to the stock 1.2.1 system is the bwtwo.c fix
from pk to cure compilation errors on non-Sun4's if RASTERCONSOLE is defined.)
I'm really happy with the stability of this system. I wish I had some of the
newer SVR4 emulation stuff in, and I sure wish we had an audio driver for the
sun4m's, but otherwise, I'm as happy as a clam. I used to think that the i386
port was the paragon of stability (in the official release versions), but my
Pentium 133 under 1.2 is doing Weird Things now, like trashing /var/cron/log
and having "amd" hang when it tries to do automounts and having the whole box
freeze up (unpingable!) while I tried to close my X windows and log off, etc.
None of this stuff happens on my SS20 ... (Too often on the mailing lists
we get bogged down in the endless cycle of new code/"This is broken!" posts.
I just feel the need to insert the occasional "Attaboy!" every so often (-: )