Re: knfsd and system crashes

Andrew E. Mileski (aem@netcom.ca)
Sun, 16 Nov 1997 16:35:12 -0500 (EST)


> If dentries are doomed to fail, then it would be good to
> just admit it. There is no shame in reverting to inodes.
> The dentries are a great idea, but perhaps better suited
> for a non-POSIX system.
>
> So, what do the VFS hackers think? Can dentries work?
> Would throwing them out be harder at this point?

I think that dentries should stay, but I also think NFS
should go instead :-)

Other *nixes use a dcache of sorts (SCO does), so I don't
see it as being a concept problem. It looks like it's an
implementation problem.

>From here on in is complete jibberish...

I think the problems with the present dcache stems from it
being the master of all inodes. Instead, it should be a
slave to them.

I would suggest that inodes be restored to master status.
NFS can deal with this - this is a proven fact. Of course
this would mean restoring VFS to its pre-dcache state.

Each inode would get a pointer to a list of it's dentries.
This gives the file system fine control over what appears
in the dcache. Right now, there is no reverse look-up
of inode to dentries either.

For flexibility, all dcache look-ups should also optionally
go through the file system for any last minute adjustments.
A new dcache() VFS routine perhaps.

--
Andrew E. Mileski   mailto:aem@netcom.ca