> I agree with him.  If we didn't have the disintegrated buffer and VM
> caches, we wouldn't need it either, because writes to the memory area
> would change (and dirty) the buffer cache pages, thus (a) affecting
> future filesystem reads immediately and (b) effecting a file write when
> the buffer cache page eventually gets cleaned.

I can think of OS architectures that have integrated VM/buffer cache that
would still require msync() under certain circumstances.

Is the problem just an msync()?  Surely to God that must be less work to put
in than some of the other things that must be done for compatibility!

-- 
David Jones, M.A.Sc student, Electronics Group (VLSI), University of Toronto
           email: dej@eecg.toronto.edu, finger for PGP public key
         For a good time, telnet torfree.net and log in as `guest'.
          Click me!