Re: Unlock-lock questions and the Linux Kernel Memory Model

From: Peter Zijlstra
Date: Wed Nov 29 2017 - 14:59:01 EST


On Wed, Nov 29, 2017 at 11:04:53AM -0800, Daniel Lustig wrote:
> "MP+wmb+xchg-acq" (or some such)
>
> {}
>
> P0(int *x, int *y)
> {
> WRITE_ONCE(*x, 1);
> smp_wmb();
> WRITE_ONCE(*y, 1);
> }
>
> P1(int *x, int *y)
> {
> r1 = atomic_xchg_relaxed(y, 2);
> r2 = smp_load_acquire(y);

Oh, I wasn't careful enough reading; these are both y.

And then Alan raises a good point in that RmW have dependencies.

Much tricker than I initially thought.

> r3 = READ_ONCE(*x);
> }
>
> exists (1:r1=1 /\ 1:r2=2 /\ 1:r3=0)