[ On Saturday, September 21, 2002 at 12:06:29 (-0400), setient@vlaxus.cjb.net wrote: ]
> Subject: sun ipc's and netbsd 1.6
>
> i noticed that alot of things do not compile on a sun ipc with 48 megs ram
> and 2 gig hd running netbsd from the pkgsrc collection one time it would
> exit with a signal 10 then a signal 11 then just have some errors in the
> code....

The problem is not that things don't compile -- it's that larger and
longer running programs often die on Sun-4c machines without hardware
cache flush assist, such as the IPC, SS1, SS1+, etc.

As you no doubt already know by now there's been several long threads of
discussion about this issue here and in various GNATS PRs for the past
few months.....  :-)

I got bored last night in the mini-heat-wave we've been having here in
Toronto and with my main workstation still crawling real slow using the
work-around Manuel Boyer originally found and now I'm currently testing
a much more efficient patch.  So far it looks like I can get 1.6 to run
big long-running processes at least as reliably as I remember 1.3.2
running them, with a very heavily stress-tested xserver running for
about 4 and a half hours before crashing (normally under similar load
the xserver will crash within a few minutes, and that test run included
/etc/daily and /etc/weekly in the background too! -- under 1.3.2 I ran
for between a week or two under normal load before the xserver died).
I'm now into my second test run at 1.5 hours with 30 minutes of CPU time
on the XsunMono process.  If it makes it far enough beyond four hours
I'll post the patch sources and make a patched kernel available to
anyone who doesn't have another machine on which to build their own.  If
it doesn't then I've got one more change I'd like to throw at it.....

-- 
								Greg A. Woods

+1 416 218-0098;            <g.a.woods@ieee.org>;           <woods@robohack.ca>
Planix, Inc. <woods@planix.com>; VE3TCP; Secrets of the Weird <woods@weird.com>