NSLU2 running 5.0_RC not stable

[Andy Ruhl <acruhl%gmail.com@localhost>]

Hello all. I'm having some troubles with my NSLU2 which is evbarm-eb.

I built 5.0_RC from my 4.0_STABLE i386 system, and I'm netbooting it
as per these instructions (well, for the most part):

http://wiki.netbsd.se/How_to_install_NetBSD_on_the_Linksys_NSLU2_(Slug)_without_a_serial_port,_using_NFS_and_telnet

I'm trying to build stuff from a current pkgsrc, and this is where I'm
having my problems.

I'm getting checksum failures on downloaded source. sha1 seems to do
fine, but rmd160 will fail. For the heck of it, I mounted the same
source tree that I built the system on from my i386 box, and re-built
all the checksum tools for the NSLU2 natively (make USETOOLS=no
install), and it still fails.

The compiler will hang or fail in seemingly random places. I'll try to
get some output. When it fails it says "internal compiler error". I'm
trying to rebuild the world using the NSLU2, but I'm not having any
luck. I'm having more compiler errors.

It's not a load related problem I think. It will run happily for hours
at a 2.00 load average according to top, doing uncompressing of large
files and other stuff.

I noticed that I started having strange problems with the compiler
after I de-underclocked it, from here:

http://www.nslu2-linux.org/wiki/HowTo/OverClockTheSlug

The resistor came off cleanly and I don't expect that I had any
problems. I have some background with hands on electronics, so I think
I did it right.

I'd be willing to get the 5.0_RC world to someone to test the same
stuff I'm doing if they'd like. Maybe I just have a bad NSLU2, or I
hurt it by de-underclocking it.

Andy