Re: [PATCH] [20/28] x86: MCE: Switch x86 machine check handler to Monarch election. II

[Andi Kleen]

On Fri, Apr 17, 2009 at 03:09:44PM +0200, Andi Kleen wrote:
> > > +
> > > +	/*
> > > +	 * Monarch starts executing now, the others wait.
> > > +	 */
> > > +	if (*order == 1) {
> > > +		atomic_set(&global_nwo, 0);
> > 
> > Monarch should clear global_nwo after all Subjects have read it.
> 
> The subjects don't care about the global nwo state, it only matters to the 
> Monarch who does the panic in mce_end(). The only exception would be timeout, 
> but in this case all the decisions are local only anyways.
> We ensure that all the subjects have written it first.

I thought some more about this and I think I was a little fast to dismiss
it due to the bank sharing heuristics. It's safer to indeed wait
for the subjects here too before clearing. I'll fix that up, thanks.

-Andi

-- 
ak@xxxxxxxxxxxxxxx -- Speaking for myself only.
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/