Re: [PATCH 3/4] x86: save user rsp in pt_regs->sp on SYSCALL64 fastpath

[Ingo Molnar]

* Andy Lutomirski <luto@xxxxxxxxxxxxxx> wrote:

> > Since this patch does add two extra MOVs,
> > I did benchmark these patches. They add exactly one cycle
> > to system call code path on my Sandy Bridge CPU.
> 
> Personally, I'm willing to pay that cycle.  It could be a bigger 
> savings on context switch, and the simplification it enables is 
> pretty good.

But, but ... context switches are a relative slow path, compared to 
system calls. And I say this with the scheduler maintainer hat on as 
well.

So this is not a good bargain IMHO, assuming it's not some _huge_ 
difference in maintainability - but having an extra percpu field
isn't really much of a problem.

I don't claim that we couldn't in some other situation decide that a 
certain type of speedup isn't worth it - but what's the big problem 
here? A bit of arithmetics shouldn't be a problem?

Thanks,

	Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/