tweaking mounted filesystems by fsid

Alfred Perlstein alfred at freebsd.org
Fri Jul 2 11:42:39 PDT 2004


Hi.

The last year or so I spent bringing file system improvements into
OS X.

The mechanism I used to tweak filesystems was sysctl.

I created a node that would route a request to a filesystem based
on a fsid.  This would allow tweaking of filesystems without entering
the namespace.

I realize we have the nmount syscall.  I have several questions
about it.

1) can I muck with it so that it functions like unmount(2) by
taking the "FSID:val0:val1" parameter in order to properly
route requests?

2) what if i want to pass binary data?  I can do that right?
I assume by just passing the binary gook via the value of the
key value pair.

Any comments?

On of the issues I have is that I need the call to be callable
from both inside and outside of the kernel, I'm guessing this can
be taken care of by the internal options...

Ideas people?

Use nmount or sysctl?

FYI, the stuff that depends on this is filesystem mobility, filesystem
notifications and autofs.

-- 
- Alfred Perlstein
- Research Engineering Development Inc.
- email: bright at mu.org cell: 408-480-4684


More information about the freebsd-fs mailing list