Trond Myklebust wrote:
> >> If the change of 'symlink' causes cache invalidation then it
> >> should work if you turn off caching.
>
> > Can you turn off cacheing for the contents? man nfs(5)
> > explains "noac" but that only forces attribute verification on
> > every operation. It doesn't disable cacheing of the data, does
> > it?
>
> Sorry I was unneccessarily unclear. I meant if you use 'noac'. That
> was why I said it was broken: you're relying on attribute checking to
> invalidate the cache for you. If the attributes don't change, you're
> essentially screwed.
Changing the attributes is easy: there's a microsecond field in NFS
mtime. A microsecond delay here or there is not a big deal for the
server. But Linux ignores this. (This could easily be changed).
Bum! Is it common for NFS implementations to ignore microsecond
changes for cache decisions?
> No you cannot turn off caching. For symlinks (and possibly readdir)
> you could do it by the simple expedient of invalidating the page cache
> once the VFS is done with reading it. For files this isn't practical.
I don't think you have to explicitly invalidate the page cache after
using it -- and the races that implies. Instead, just pretend the
attribute check always says "the server has newer data". Conceptually
simpler IMO.
-- Jamie
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon May 15 2000 - 21:00:11 EST