Re: [PATCH v6 4/4] xen/arm: introduce xen_early_init, use PSCI on xen

From: Stefano Stabellini <stefano.stabellini@eu.citrix.com>
To: Nicolas Pitre <nicolas.pitre@linaro.org>
Cc: Rob Herring <robherring2@gmail.com>,
	Stefano Stabellini <Stefano.Stabellini@eu.citrix.com>,
	"xen-devel@lists.xensource.com" <xen-devel@lists.xensource.com>,
	"arnd@arndb.de" <arnd@arndb.de>,
	"will.deacon@arm.com" <will.deacon@arm.com>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
	"rob.herring@calxeda.com" <rob.herring@calxeda.com>,
	"linux-arm-kernel@lists.infradead.org" 
	<linux-arm-kernel@lists.infradead.org>
Subject: Re: [PATCH v6 4/4] xen/arm: introduce xen_early_init, use PSCI on xen
Date: Sat, 6 Apr 2013 00:20:58 +0100	[thread overview]
Message-ID: <alpine.DEB.2.02.1304060015490.5353@kaball.uk.xensource.com> (raw)
In-Reply-To: <alpine.LFD.2.03.1304051718020.1171@syhkavp.arg>

On Fri, 5 Apr 2013, Nicolas Pitre wrote:
> On Fri, 5 Apr 2013, Rob Herring wrote:
> 
> > On 04/05/2013 02:36 PM, Nicolas Pitre wrote:
> > > On Fri, 5 Apr 2013, Stefano Stabellini wrote:
> > > 
> > >> This is what happens:
> > >>
> > >> - No Xen
> > >> Xen is not running on the platform and a Xen hypervisor node is not
> > >> available on device tree.
> > >> Everything keeps working seamlessly, this patch doesn't change anything.
> > >>
> > >> - we are running on Xen
> > >> Xen is running on the platform, we are running as a guest on Xen and an
> > >> hypervisor node is available on device tree.
> > >> Let's also assume that there aren't any "arm,cci" compatible nodes on
> > >> device tree because Xen wouldn't export this kind of information to any
> > >> guests right now. Therefore PSCI should be used to boot secondary cpus.
> > >> Because the versatile express machine sets smp_init to
> > >> vexpress_smp_init_ops, vexpress_smp_init_ops will be called.
> > >> vexpress_smp_init_ops sets smp_ops to vexpress_smp_ops, that *break*
> > >> Xen.
> > > 
> > > OK I see.
> > > 
> > >> With this patch, xen_smp_init will be called instead of
> > >> vexpress_smp_init_ops, and smp_ops will be set to psci_smp_ops,
> > >> therefore *unbreaking* Xen.
> > > 
> > > However that breaks MCPM.
> > 
> > You mean on bare metal, right? For the bare metal, "xen,xen" property
> > would not be present and xen_smp_init is not used. So the vexpress MCPM
> > ops will be used. Aren't Dom0 cpu's basically virtual cpus? If Xen ever
> > needs the MCPM support, the Xen hook itself can figure out whether to
> > use MCPM support.
> 
> Well, if Xen has its own mdesc distinct from the VExpress one then 
> things
> are indeed fine.

It's not about the mdesc: Xen has its own hypervisor node on device tree
if and only if Xen is running on the platform, therefore the Xen early
hook is never going to do anything at all on native.

In other words, this patch should NOT change the behaviour of Linux on
native, and if it did, it would be a bug.