Re: PATCH: Raw device IO for 2.1.131

Colin Plumb (colin@nyx.net)
Sun, 13 Dec 1998 16:45:10 -0700 (MST)


Jamie Lokier wrote:
> Exactly when is a write operation "complete"?

Well, some people like to know when it has been comitted to stable storage.

But even without that, you still need to know when it's safe to modify the
buffer without corrupting the write in progress.

The latter is the guarantee that Unix makes on return from the write()
system call. Achieving it in the face of write-behind requires copying.
(Although for large buffers, the copy may be done using MMU cow tricks
if desired.)

If the async notify is cheaper than the copy, this offers a potential
cost savings.

-- 
	-Colin

- 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/