In message <199602250840.DAA05217@glacier.MIT.EDU>
	so spake Greg Hudson (ghudson):

> > In a nutshell, using 1024 for rsize and wsize causes NFS-mounts to
> > hang.
> 
> The problem seems to be that directory reads are still being done in
> 8k blocks, which are probably not reliably transported across your
> router.  You can set the directory read size to 1k at mount time.

Actually, this is between machines on the same subnet to the
router is not the cause.  As Charles and others pointed out,
the problem is caused by a mismatch between rsize and readdirsize.
[this is an NFSv2 mount btw]

 - todd