Re: One chart on pci bridge and its bus and their children

[Ram Pai <linuxram@us.ibm.com>]

On Wed, Jun 06, 2012 at 12:40:19AM +0800, Ram Pai wrote:
> On Tue, Jun 05, 2012 at 10:41:39PM +0800, Jiang Liu wrote:
> > On 06/05/2012 02:35 PM, Ram Pai wrote:
> > > On Mon, Jun 04, 2012 at 10:20:51PM -0700, Yinghai Lu wrote:
> > >> On Mon, Jun 4, 2012 at 10:18 PM, Yinghai Lu <yinghai@kernel.org> wrote:
> > >>> On Mon, Jun 4, 2012 at 9:37 PM, Bjorn Helgaas <bhelgaas@google.com> wrote:
> > >>>> That's true, but I don't think it answers the question.  If we have:
> > >>>>
> > >>>>    pci_bus->bridge == pci_bus->self->dev
> > >>>>
> > >>>> why would we need both "self" and "bridge"?  It would be interesting
> > >>>> to try to remove "bridge" and replace uses of it with "self->dev".
> > >>>
> > >>> then how about root bus?
> > >>>
> > >>> root bus ->self should be NULL.
> > >>>
> > >>> root bus ->bridge is to the hostbridge->dev.
> > > 
> > > Yinghai,
> > > 
> > > is there a reason why this can't be fixed in
> > > pci_create_root_bus() by having
> > >    bus->self = bridge; /* 'bridge' is the pci_device of the corresponding host bridge */
> > Hi Ram,
> > 	A host bridge has no associated pci_device because it's out of the
> > PCI domain. PCI domain only covers the hierarchy starting from the host bridge,
> > but doesn't include the host bridge itself.
> 
> That makes it even easier to kill pci_bus->bridge.
> 
> for non-root buses, we can access the bridge's device using
> bus->self->dev and for root bridges we can silently ignore
> the bridge's dev since we know that it does not exist; because
> pci_bus->self is null.

hmm..ok. I realize the struct device of the host bridge exists and needs
to be accessed, even though it is not a pci_device. So killing 
bus->bridge is not trivial.

RP