Re: [PATCH] compatibility syscall layer (lets try again)

[Linus Torvalds]

On Mon, 9 Dec 2002, Martin Schwidefsky wrote:
>
> I had been looking at 2.5.50, we had a different meaning of current.
> If you are saying that for any implementation of nanosleep I have to implement
> the -ERESTART_RESTARTBLOCK thingy anyway, then I better start with it.

You don't _have_ to. An architecture for which restarting is just too
painful can just always choose to return -EINTR, that should be ok. That's
how nanosleep() used to work before - it may not be 100% SuS compliant,
but it's not as if anybody really cares, I suspect.

		Linus