[OpenAFS] Re: Continuing problems with aklog on Mac OS X (rc10), internal state confused?

Ken Hornstein kenh@cmf.nrl.navy.mil
Tue, 04 Apr 2006 10:43:12 -0400

>You may instead be wanting to instrument
>afs/afs_dynroot.c or afs/afs_cell.c .  This run as part of
>the cache manager, nearly always inside the kernel (except
>for ms windows).  Inside the kernel, printf or uprintf usually work.
>printf would go to the machine console (probably nowhere on macos,
>but might show up in syslog); uprintf would go to the user's tty
>(same place as writing /dev/tty from userland).  You do Not want
>to send much chit-chat this way.

When I looked at this code last time Adam talked about this problem, I
wasn't able to convince myself that anything useful was happening in
there (by "useful" I mean, "likely to fail").  It looked like all of
the interesting stuff was in the afsd upcall code that actually was
making the DNS queries.  I'm not saying that instrumenting that code
wouldn't be useful, but I think starting with the userland code is
probably easier (if printf() isn't useful inside of afsd, you could
always use syslog()).