On Mon, Mar 7, 2011 at 7:09 PM, Venkatesh Srinivas <m...@acm.jhu.edu> wrote:
>
> In 9P, if I wish to list a directory, I need to TWalk to the directory,
> TOpen the directory fid from the walk, and then TRead till I have all of the
> contents of the directory.
> If the directory's contents do not fit in a single read, I imagine I need to
> loop around TOpen / Tread / .... / Tread / TOpen, till I get the whole
> contents and see the same QIDs in each open, to get a coherent listing. (is
> this accurate?)
>
Offsets track where you are in the metadata, servers have multiple
approaches to dealing with this, usually there is some form of cookie
in the fid which tracks the last offset red and maybe a cache of where
it was in the directory listing. I think the QID matching and
coalescing is more of a union bind mechanism and happens if I am not
mistaken by a higher level entity (like ls)
> Unfortunately, a TOpened FID to a directory cannot be used for walking it
> (why?); so the only use of a TOpened fid to a directory is to list it or to
> get a QID. Would it be reasonable for a TOpened fid to a directory to create
> a copy of the directory listing at the time of the RPC?
That may be a potential back-end implementation on the server, I
believe it to be the only way to avoid the sort of race conditions you
might expect otherwise. But I haven't implemented the server side of
that code, only the client which is tricky, but easier -- so I'm sure
someone else may have a better opinion.
-eric
