Not that I should chime in here BUT..

>From a management standpoint, it is a PITA whether it is technically
correct or not. You literally have to reboot the machine to stop and start
the client on Solaris.

Just out of curiousity, how many non-threaded kernel systems are we
currently supporting now? Is this just more legacy code that could be
deprecated or tailored for specific non-kernel threaded systems?



On Mon, 11 Feb 2008, Jeffrey Hutzelman wrote:

> --On Monday, February 11, 2008 07:50:10 PM +0100 "Frank Batschulat (Home)"
> <[EMAIL PROTECTED]> wrote:
>
> > if thats the intent, ie. block all signal over the AFS syscall kernel
> > execution, the afsd could possibly use sigfillset(3C) &
> > thr_sigsetmask(3C), e.g
>
> (1) That's not "all signals".  SIGKILL is a signal.  So is SIGSTOP.
> (2) It's not afsd's job to set up the operation of the cache manager's
> kernel threads.  It's afsd's job to fork and make a syscall to donate its
> context.  This is a perfectly legitimate way to create kernel threads, and
> as with a number of cases, the reason AFS does it this way is because we've
> been using kernel threads since before most operating systems had them.
>
> -- Jeffrey T. Hutzelman (N3NHS) <[EMAIL PROTECTED]>
>    Carnegie Mellon University - Pittsburgh, PA
>
> _______________________________________________
> OpenAFS-devel mailing list
> [email protected]
> https://lists.openafs.org/mailman/listinfo/openafs-devel
>

--------------------------------------
  Sean O'Malley, Information Technologist
  Michigan State University
-------------------------------------

_______________________________________________
OpenAFS-devel mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-devel

Reply via email to