David Howells wrote:

Chris Snook [off-list ref] wrote:

quoted

To head off the criticism, I admit this is an oversimplification, and true
busy-waiters should be using cpu_relax(), which contains a barrier.

Why would you want to use cpu_relax()?  That's there to waste time efficiently,
isn't it?  Shouldn't you be using smp_rmb() or something like that?

David

cpu_relax() contains a barrier, so it should do the right thing.  For 
non-smp architectures, I'm concerned about interacting with interrupt 
handlers.  Some drivers do use atomic_* operations.

	-- Chris