Excerpts from Peter Zijlstra's message of July 2, 2020 6:28 pm:

On Thu, Jul 02, 2020 at 05:48:33PM +1000, Nicholas Piggin wrote:

quoted

There is no need for rmb(), this allows faster lwsync here.

Since you determined this; I'm thinking you actually understand the
ordering here. How about recording this understanding in a comment?

Also, should the lock->slock load not use READ_ONCE() ?

Yeah, good point. Maybe I'll drop it from this series, doesn't really 
belong I just saw the cleanup and didn't want to forget it.

We we just ordering the two loads in this function, and !SMP isn't a 
concern (i.e., no issues of !SMP guest on SMP HV), but yeah fixing
the lack of comment is warranted, thanks.

Thanks,
Nick