Re: [PATCH v4 04/10] s390/pci: Add architecture specific resource/bus address translation

[Farhan Ali]

On 10/1/2025 9:04 AM, Benjamin Block wrote:
> On Thu, Sep 25, 2025 at 12:54:07PM +0200, Niklas Schnelle wrote:
> > On Wed, 2025-09-24 at 10:16 -0700, Farhan Ali wrote:
> > > +void pcibios_resource_to_bus(struct pci_bus *bus, struct pci_bus_region *region,
> > > +			     struct resource *res)
> > > +{
> > > +	struct zpci_bus *zbus = bus->sysdata;
> > > +	struct zpci_bar_struct *zbar;
> > > +	struct zpci_dev *zdev;
> > > +
> > > +	region->start = res->start;
> > > +	region->end = res->end;
> > When we don't find a BAR matching the resource this would become the
> > region used. I'm not sure what a better value would be if we don't find
> > a match though and that should hopefully not happen in sensible uses.
> > Also this would keep the existing behavior so seems fine.
> I was wondering the same things, but I guess it matches what happens elsewhere
> as well, if no match is found
>
> 	void __weak pcibios_resource_to_bus(struct pci_bus *bus, struct pci_bus_region *region,
> 				     struct resource *res)
> 	{
> 		...
> 		resource_size_t offset = 0;
>
> 		resource_list_for_each_entry(window, &bridge->windows) {
> 			if (resource_contains(window->res, res)) {
> 				offset = window->offset;
> 				break;
> 			}
> 		}
>
> 		region->start = res->start - offset;
> 		region->end = res->end - offset;
> 	}
>
> So I guess that is fine.
>
> The thing I'm also unclear on is whether it is OK to "throw out" this whole
> logic about `resource_contains(window->res, res)` here and
> `region_contains(&bus_region, region)` in the other original function?
> I mean, the original function don't search for perfect matches, but also
> matches where are contained in a given resource/region, which is different
> from what we do here. Are we OK with not doing that at all?

I had thought about doing the range check, similar to 
resource_contains/region_contains rather than doing exact checks. But I 
think the way we expose the topology of the devices, the offset in our 
(s390x) case is zero. So I thought it should be safe to just doing exact 
match and might help us catch any issues if its not exact.

Thanks

Farhan