Re: [RFC PATCH v1 1/4] arm/pci: PCI setup and PCI host bridge discovery within XEN on ARM.

From: "Roger Pau Monné" <roger.pau@citrix.com>
To: Julien Grall <julien@xen.org>
Cc: Rahul Singh <rahul.singh@arm.com>,
	Bertrand.Marquis@arm.com,
	Stefano Stabellini <sstabellini@kernel.org>,
	xen-devel@lists.xenproject.org, nd@arm.com,
	Volodymyr Babchuk <Volodymyr_Babchuk@epam.com>
Subject: Re: [RFC PATCH v1 1/4] arm/pci: PCI setup and PCI host bridge discovery within XEN on ARM.
Date: Fri, 24 Jul 2020 17:42:58 +0200	[thread overview]
Message-ID: <20200724154258.GN7191@Air-de-Roger> (raw)
In-Reply-To: <20200724152905.GM7191@Air-de-Roger>

On Fri, Jul 24, 2020 at 05:29:05PM +0200, Roger Pau Monné wrote:
> On Fri, Jul 24, 2020 at 04:15:47PM +0100, Julien Grall wrote:
> > 
> > 
> > On 24/07/2020 15:44, Roger Pau Monné wrote:
> > > > +
> > > > +    if ( acpi_disabled )
> > > > +        dt_pci_init();
> > > > +    else
> > > > +        acpi_pci_init();
> > > 
> > > Isn't there an enum or something that tells you whether the system
> > > description is coming from ACPI or from DT?
> > > 
> > > This if .. else seems fragile.
> > > 
> > 
> > This is the common way we do it on Arm.... I would welcome any improvement,
> > but I don't think this should be part of this work.
> 
> Ack. In any case I think for ACPI PCI init will get called by
> acpi_mmcfg_init as part of acpi_boot_init, so I'm not sure there's
> much point in having something about ACPI added here, as it seems this
> will be DT only?

Sorry I got confused and acpi_boot_init is a x86 specific function,
wrong context. If Arm is not using that path then maybe it makes sense
to init PCI here for ACPI also.

Roger.