In message <200201281442.g0SEgJK67524@mrynet.com>, you write: 

-> I've been struggling for quite a long time to get a snapshot built.
-> 
-> At first, I was struggling with this on an R5K INDY.  This machine
-> would regularly hang--no pattern to when or where.  Only power
-> cycling the machine would get it going again.  
-> 
-> So, this weekend I moved to an R4400, incase I had a hardware
-> problem.  I should mention that the R5K works fine using one
-> of my snapshots from late December.

Hmm, ENOR5k, so I can't help you here 8-/

-> On the R4400 now, the following race condition occurs:
-> 
-> install ===> libc
[...]
-> STRIP=/sgimips/src/tools/obj/tools.NetBSD-1.5Z-mipseb/bin/mipseb--netbsd-str
-> ip
-> /sgimips/src/tools/obj/tools.NetBSD-1.5Z-mipseb/bin/nbinstall  -l s -r libc.
-> so.12.82  /usr/lib/libc.so
-> Segmentation fault - core dumped
-> *** Error code 139
-> 
-> Once that libc.so link is created, the majority of dynamic executables
-> on the machine are rendered inoperable, resulting in the same
-> Segfault.
-> 
-> I'm at a loss as to what I should do at this point.  Any suggestions?
-> Even more importantly, has anyone been able to build a snapshot this
-> year?

MAKE SURE TO BUILD & INSTALL ld.elf_so FIRST!!! (sorry to yell, but it is
*really* important).  Once that's done, you should be OK.

Probably the correct method to build is:
	build.sh -t
	(maybe do a make includes)
	(now with new tools) build & install ld.elf_so
	build.sh to build rest of system.

I built a complete snapshot shortly before the new year (either right after
or right before Jason switched to NEW_TOOLCHAIN); I haven't had much time to
play with my SGI gear since, but I did start looking at your serial console
crash last week -- the good news is that it's reproducible here, the bad is
that it's baffling (though I only had a little time to look at it last week).

--rafal

----
Rafal Boni                                                  rafal@mediaone.net