On Wed, Sep 10, 2014 at 05:53:47PM +0100, Liviu Dudau wrote:

[...]

quoted

quoted

quoted

quoted

quoted

quoted

+		if (resource_type(res) == IORESOURCE_IO) {
+			if (*io_base)

You do not zero io_base in the first place so you should ask the API
user to do that. Is 0 a valid value BTW ? If it is you've got to resort
to something else to detect multiple IO resources.

No, zero is not a valid value. It is the cpu_addr value from the IO range, I'm
hopying that no one is crazy enough to map PCI address space at CPU address zero.
Thanks for spotting the lack of initialisation though, I need to fix it.

Mmm...wasn't a trick question sorry :D

PCI host bridge /pci ranges:
   IO 0x00000000..0x0000ffff -> 0x00000000
   More than one I/O resource converted. CPU offset for old range lost!
     MEM 0x41000000..0x7fffffff -> 0x41000000
     pci-host-generic 40000000.pci: PCI host bridge to bus 0000:00
     pci_bus 0000:00: root bus resource [bu-01]
     pci_bus 0000:00: root bus resource [io  0x0000-0xffff]
     pci_bus 0000:00: root bus resource [mem 0x41000000-0x7fffffff]

Your DT puts the IO space at CPU address zero? Could you copy the relevant
host bridge node from DT here?

It is kvmtools generated dts, may not be representative of real HW but
that's what it is. For the whole story:

http://comments.gmane.org/gmane.linux.kernel.pci/29199

Lorenzo