wmb vs mmiowb

[Benjamin Herrenschmidt]

On Wed, 2007-08-22 at 06:57 +0200, Nick Piggin wrote:

> It doesn't seem like this primary function of mmiowb has anything to do
> with a write barrier that we are used to (it may have a seconary semantic
> of a wmb as well, but let's ignore that for now). A write barrier will
> never provide you with those semantics (writes from 2 CPUs seen in the
> same order by a 3rd party). If anything, I think it is closer to being
> a read barrier issued on behalf of the target device.  But even that I
> think is not much better, because the target is not participating in the
> synchronisation that the CPUs are, so the "read barrier request" could
> still arrive at the device out of order WRT the other CPU's writes.
> 
> It really seems like it is some completely different concept from a
> barrier. And it shows, on the platform where it really matters (sn2), where
> the thing actually spins.

The way mmiowb was actually defined to me by the ia64 folks who came up
with it is essentially to order an MMIO write with a spin_unlock.

Ben.