Re: [RFC v5 8/8] PCI: hv: Turn on the host bridge probing on ARM64

From: Marc Zyngier <maz@kernel.org>
To: Boqun Feng <boqun.feng@gmail.com>
Cc: "Bjorn Helgaas" <bhelgaas@google.com>,
	"Arnd Bergmann" <arnd@arndb.de>,
	"Catalin Marinas" <catalin.marinas@arm.com>,
	"Will Deacon" <will@kernel.org>,
	"K. Y. Srinivasan" <kys@microsoft.com>,
	"Haiyang Zhang" <haiyangz@microsoft.com>,
	"Stephen Hemminger" <sthemmin@microsoft.com>,
	"Wei Liu" <wei.liu@kernel.org>,
	"Dexuan Cui" <decui@microsoft.com>,
	"Lorenzo Pieralisi" <lorenzo.pieralisi@arm.com>,
	"Rob Herring" <robh@kernel.org>,
	"Krzysztof Wilczyński" <kw@linux.com>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org, linux-hyperv@vger.kernel.org,
	linux-pci@vger.kernel.org,
	"Sunil Muthuswamy" <sunilmut@microsoft.com>,
	"Mike Rapoport" <rppt@kernel.org>
Subject: Re: [RFC v5 8/8] PCI: hv: Turn on the host bridge probing on ARM64
Date: Tue, 20 Jul 2021 16:53:08 +0100	[thread overview]
Message-ID: <87tukp7ziz.wl-maz@kernel.org> (raw)
In-Reply-To: <YPbk3Ya20z1PDn2H@boqun-archlinux>

On Tue, 20 Jul 2021 15:59:41 +0100,
Boqun Feng <boqun.feng@gmail.com> wrote:
> 
> On Tue, Jul 20, 2021 at 03:38:26PM +0100, Marc Zyngier wrote:
> > On Tue, 20 Jul 2021 14:44:29 +0100,
> > Boqun Feng <boqun.feng@gmail.com> wrote:
> > > 
> > > Now we have everything we need, just provide a proper sysdata type for
> > > the bus to use on ARM64 and everything else works.
> > > 
> > > Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
> > > ---
> > >  drivers/pci/controller/pci-hyperv.c | 7 +++++++
> > >  1 file changed, 7 insertions(+)
> > > 
> > > diff --git a/drivers/pci/controller/pci-hyperv.c b/drivers/pci/controller/pci-hyperv.c
> > > index e6276aaa4659..62dbe98d1fe1 100644
> > > --- a/drivers/pci/controller/pci-hyperv.c
> > > +++ b/drivers/pci/controller/pci-hyperv.c
> > > @@ -40,6 +40,7 @@
> > >  #include <linux/kernel.h>
> > >  #include <linux/module.h>
> > >  #include <linux/pci.h>
> > > +#include <linux/pci-ecam.h>
> > >  #include <linux/delay.h>
> > >  #include <linux/semaphore.h>
> > >  #include <linux/irqdomain.h>
> > > @@ -448,7 +449,11 @@ enum hv_pcibus_state {
> > >  };
> > >  
> > >  struct hv_pcibus_device {
> > > +#ifdef CONFIG_X86
> > >  	struct pci_sysdata sysdata;
> > > +#elif defined(CONFIG_ARM64)
> > > +	struct pci_config_window sysdata;
> > > +#endif
> > 
> > Am I the only one who find this rather odd? Nothing ever populates
> > this data structure on arm64, and its only purpose seems to serve as
> > an anchor to retrieve the hbus via container_of().
> > 
> 
> This field will also be used as the ->sysdata of pci_bus and
> pci_host_bridge, and some of the PCI core code touches. Although I made
> this field as all zeroed and make sure PCI core can handle (patch #4).

Huh, I see. I missed this particular nugget. This is so convoluted...

> > If that's indeed the case, I'd rather see an arch-specific to_hbus()
> > helper that uses another (preexisting) field as the anchor for arm64.
> > 
> 
> I did a quick look, but I didn't find another field works: the field
> needs to be placed inside hv_pcibus_device and the address can be
> retrieved via pci_bus. I'm open to any suggestion in case that I missed
> something.

No, the above pretty much kills my suggestion.

Thanks for the explanation,

	M.

-- 
Without deviation from the norm, progress is not possible.