[PATCH v6 1/4] arm: introduce psci_smp_ops

[Stefano Stabellini]

On Mon, 22 Apr 2013, Russell King - ARM Linux wrote:
> > > Now, if the psci stuff can't be relied upon to provide the correct
> > > functionality, then that's a separate problem which needs addressing
> > > differently.
> > > 
> > > This should allow the Xen problem to be resolved, because Xen will
> > > provide the PSCI operations, and it's correct in that case to override
> > > the platform's SMP operations.
> > 
> > Yes, increasing the priority of PSCI helps Xen a lot.
> > In order to completely solve the issue for Xen though, another patch is
> > needed (http://marc.info/?l=linux-kernel&m=136630106201968&w=2) because
> > of the introduction of smp_init.
> 
> Given the way the code in setup.c will work:
> 
> +static bool __init xen_smp_init(void)
> +{
> +#ifdef CONFIG_SMP
> +	/* If we are running on Xen, use PSCI if available.
> +	 * In any case do not try to use the native smp_ops. */
> +	if (psci_smp_available())
> +		smp_set_ops(&psci_smp_ops);
> +#endif
> +	return true;
> +}
> 
> Doesn't this just need to return false, and then we'll drop down to
> using PSCI if those operations are available?

If we are running on Xen, we are sure that the only available mechanism
to boot secondary cpus is PSCI; there is no need to try anything else if
that one is not available. This is way xen_smp_init is returning true
unconditionally.

However I am going to drop the "xen/arm: introduce xen_early_init, use
PSCI on xen" patch linked above entirely, because I'll just assume that
if we are running on Xen smp_init is going to fail and return false. The
next option on the list is PSCI so in theory we won't have to do
anything special to make it work.