Messages in this thread | | | Subject | Re: VFS event hooks | Date | Sun, 27 Jun 1999 23:35:05 -0400 | From | Paul Barton-Davis <> |
| |
>> the semantics that were being discussed were at the VFS level, and >> poll is part of the implementation of a specific file system. > >If you're using a networked filesystem, and a file that you're interested in >is changed by some other host, then the polling is definitely >filesystem-specific, surely?
yes, and network filesystems in general are the best argument i can think of for using poll. In fact, its hard to see how this could be supported across NFS (or any networked file system) without using poll(). The other mechanism (which I prefer for reasons that have been pointed out by others) would have the feature that it did not require that an application holds an fd for the files its interested in. If we don't, no current network filesystem server in its right mind will propagate any changes to us (where "us" is a host with no processes holding the relevant fd).
so - i don't know.
--p
- 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/
| |