On Sun, 12 Mar 2000, Peter J. Braam wrote:
>
> I think that's ok with the Coda directories, if not it's really easy to
> fix in userland for us. Assume it.
That's fine. Because you already assume that underlying filesystem has
pagecache working... Translation: I'm going to convert coda_venus_readdir()
to ->readpage() from underlying fs. No extra copies, no setfs() mess... Along
with trivial modification of binfmt stuff it will allow to kill read_exec().
And when fhandle_to_dentry() patch will go in we'll be able to get rid of
init_private_file(). Which is one of the things I wanted to do since _long_.