Re: [PATCH] Improve spinlock inline assembly for x86.

On 2016-01-18 16:56:22 -0600, Kevin Grittner wrote:
> On Mon, Jan 18, 2016 at 4:50 PM, Andres Freund <andres@anarazel.de> wrote:
> 
> > Now I'm equally unconvinced that it's worthwhile to do anything
> > here. I just don't think benchmarking plays a role either way.
> 
> Well, that would be the crucial point on which we differ -- the
> rest is all agreement.  I don't think we should accept the patch
> *in the absence* of benchmarking to show a result that is neutral
> or better.  Spinlocks are just too performance-critical and too
> fussy to accept a change on the basis that "the source code looks
> fine".  IMO, anyway.

By that justification we need to start benchmarking adding new variables
on the stack, that'll most of the time have a bigger performance impact
than this.  Benchmarking minor source code cleanups is just not worth
the time.