[PATCH 2/3] PCI: ARM: add support for virtual PCI host controller

[Will Deacon]

On Wed, Feb 12, 2014 at 06:19:13PM +0000, Jason Gunthorpe wrote:
> On Wed, Feb 12, 2014 at 06:10:15PM +0000, Will Deacon wrote:
> 
> > > AFAIK, the job is fairly simple, when you call pci_add_resource_offset
> > > for memory compute the offset from 
> > >   of_pci_range.pci_addr - of_pci_range.cpu_addr
> > > 
> > > (or is it the other way around ?)
> > 
> > I think it's the other way round: bus = cpu - offset, then Arnd's example of
> > PCI bus 0 works out as: 0 = cpu - pci->mem_start.
> 
> That looks right to me
> 
> > I added that to my driver, but I get some weird looking bus addresses in
> > dmesg:
> > 
> > [    0.307585] pci-arm-generic 40000000.pci: PCI host bridge to bus 0000:00
> > [    0.307601] pci_bus 0000:00: root bus resource [io  0x0000-0xffff]
> > [    0.307615] pci_bus 0000:00: root bus resource [mem 0x00000000-0x3effffff] (bus address [0xffffffffbf000000-0xfffffffffdffffff])
> > 
> > Looking at drivers/pci/probe.c, it seems to think that res->start - offset
> > gives a bus address, which implies that the resources are indeed *CPU*
> > addresses.
> > 
> > Are you sure pci_add_resource_offset wants bus addresses?
> 
> Sorry, I wasn't clear: It accepts a cpu address in the struct
> resource and an offset to convert back to a bus address.
> 
> You should compute 0 as the offset in the normal case, ie
> of_pci_range.pci_addr and of_pci_range.cpu_addr should be identical,
> which depends on the DT ranges being correct..

Aha! That explains all of the confusion. I'll remove my homebrew
resource translation code then :)

Thanks,

Will