Re: [PATCH] syscall latency improvement #1

[Jamie Lokier]

Alan Cox wrote:
> > NOTE! There are potentially other ways to do all of this, _without_ losing
> > atomicity. For example, you can move the "flags" value into the slot saved
> > for the CS segment (which, modulo vm86, will always be at a constant
> > offset on the stack), and make CS=0 be the work flag. That will cause the
> > CPU to trap atomically at the "iret".
> 
> Is the test even needed. Suppose we instead patch the return stack if we
> set need_resched/sigpending, and do it on the rare occassion we set the
> value.

Yes, that's what you'll find in one of Ingo Molnar's low latency patches.

-- Jamie