Messages in this thread | | | Date | Mon, 31 Jul 2000 18:03:34 -0400 (EDT) | From | Alexander Viro <> | Subject | Re: [RFC][Long][Horror story] Mount flags |
| |
On Mon, 31 Jul 2000, Andries Brouwer wrote:
> On Sun, Jul 30, 2000 at 07:09:35PM -0400, Alexander Viro wrote:
> > describe the rules in a manpage and see if you will be able to look at it > > without a feeling that it's an overcomplicated kludge. > > Description: "flags is a 31-bit flags word; the first bit must be zero".
Bzzzzert. I _hope_ that you will not follow that rule in mount(8), because otherwise... Guess what it will do under 2.2? Right.
> This is not the time to introduce new mount syscalls.
You've just done that _and_ kept the same syscall number. mount(2) documented as above is _NOT_ the same syscall as it used to be since '94. Different documented calling conventions. And anyone trying to follow them will get a program that behaves on any kernel prior to 2.4 other than expected. Worse yet, you will not get an error from mount(2).
Moreover, now we get an undocumented and strange-looking interface that _is_ compatible between 2.4 and earlier kernels. Fun, fun, looks like mount(8) uses it instead of what had been said in the manpage.
Sorry, but having the parameters used by mount(8) left _undefined_ in mount(2) manpage strikes me as extremely bad idea.
- 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/
| |