Coda File System

Re: Coda on OS X

From: Greg Troxel <gdt_at_ir.bbn.com>
Date: 14 Feb 2004 10:40:26 -0500
  I wanted to informally poll this group in an attempt to gauge the
  amount of interest in Coda on OS X and perhaps better understand what
  potential "resources" are available in the open source community to
  help with any porting effort.

I think there is significant interest.  I wonder sometimes if I should
get a mac notebook with OS X instead of an i386 arch with NetBSD, and
not having coda is one of the reasons in the 'no' column.

Basically, I think that coda on OS X requires mostly porting of the
code in sys/coda from NetBSD or FreeBSD to Darwin, whichever is closer
to the vnode interface in Darwin.  Then the userland programs will
have to be compiled, but that shouldn't be hard.

Coda is not yet 64-bit clean.  So the issues I raised on the list in
the last week about NetBSD/sparc64 will apply to 64-bit macs, but I
suspect that running the userland code in 32-bit mode will be a
workaround for now.

As for resources in the open source community, someone will do the
above when they are in the mood or want it bad enough, and so far this
hasn't happened.  I don't mean to be difficult, but there aren't a lot
of spare coda hackers looking for stuff to do.  That said, I would
expect that someone who is competent with the kernel vnode internals
could port the kernel part in a few months at most, probably faster,
and the userland port (to 32-bit systems) should be a few weeks.

I realize that for Apple 'making commitments' is something that must
be done carefully, but assigning someone for a few months to do this
work and contribute it back would be highly useful and appreciated and
not a commitment.  (Actually, the Darwin kernel support doesn't need
'contributing' back to the coda project, and the minor venus porting
that may or may not be needed will be of no consequence from an IP
point of view.)

-- 
        Greg Troxel <gdt_at_ir.bbn.com>
Received on 2004-02-14 10:45:49