Subject: hmmm, debug weirdness
To: None <port-sun3@NetBSD.ORG>
From: Jim Mercer <jim@reptiles.org>
List: port-sun3
Date: 12/05/1995 06:37:22
in an attempt to streamline my kernel for news processing, i removed the
following definitions from my config file and built a new kernel:

#options        KTRACE              # system call tracing support
#options    DDB 

#options    DIAGNOSTIC  
#options    SCSIDEBUG   

and changed:
pseudo-device   pty 64      # pseudo-terminals
to:
pseudo-device   pty 32      # pseudo-terminals

when i booted the new kernel, i got something like:
MNT RPC failed
(sorry, i scrolled off the screen when i rebooted with the original kernel)
the boot hung, after starting nfsd and mountd and such.

a ^C allowed the boot to proceed, but the NFS mounts failed.

any idea's what in the above might have caused that?

it could also be the way i built the kernel.

i did a config TUATARA, then cd'd to compile/... and did make depend ; make.

this ended with some link errors, so i basically nuked the .o's of the
files it was complaining about, and did a make.

this produced a kernel without errors.

is it generally accepted that one should nuke a previous kernel build
before building another one?

is it possible that i complete kernel build would not have caused the
above error?




-- 
[ Jim Mercer   Reptilian Research  merce@iguana.reptiles.org  +1 416 410-5633 ]
[  when you eat hot chiles you should follow them with some ice cream.  Then, ]
[  when you're sitting on the toilet the next day and the burn commences, you ]
[  can holler "Come onnnnnn, ice cream!"-- from the chile-heads mailing list  ]