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

[Jamie Lokier]

Linus Torvalds wrote:
> So what we do is to introduce a _new_ system call 
> (system call number NNN), which takes a different form of timeout, namely 
> "absolute value of end time".

An "absolute value of end time" variant of
select/poll/epoll/io_getevents would be good anyway from userspace, to
avoid the gettimeofday+poll race condition.

So, perhaps the solution here is to simply provide absolute time
variants of the system calls which currently take time delays, and
have the relative-time variants rewrite themselves into absolute form?

That's architecture neutral _and_ fixes a long-standing race
condition..

-- Jamie