Re: Spinlock implementation on x86_64 (was Re: Better LWLocks with compare-and-swap (9.4))

Heikki Linnakangas <hlinnakangas@vmware.com> writes:
> So, my plan is to apply the attached non-locked-tas-spin-x86_64.patch to 
> master. But I would love to get feedback from people running different 
> x86_64 hardware.

Surely this patch should update the existing comment at line 209?  Or at
least point out that a non-locked test in TAS_SPIN is not the same as a
non-locked test in tas() itself.

Other than the commenting, I have no objection to this.  I think you're
probably right that the old tests in which this looked like a bad idea
were adding the unlocked test to tas() not only the spin case.
        regards, tom lane