Re: [PATCH v2] tools/memory-model: Add extra ordering for locks and remove it for ordinary release/acquire

From: Peter Zijlstra
Date: Mon Jul 16 2018 - 15:02:15 EST


On Tue, Jul 17, 2018 at 12:40:19AM +1000, Michael Ellerman wrote:
>
> I guess arguably it's not a very macro benchmark, but we have a
> context_switch benchmark in the tree[1] which we often use to tune
> things, and it degrades badly. It just spins up two threads and has them
> ping-pong using yield.

The one advantage you'd get from putting it in lock() is that you could
do away with smp_mb__after_spinlock().

But yes, I completely forgot about your IO thingy.. those bench results
make me sad :/ a well.