When there is a PCI-X mode 2 capable device behind the HT<->PCI-X
bridge, the pci core decides that the device has the extended 4K
config space, even though the bus is not operating in mode 2. This is
because the u3_ht pci ops silently accept offsets greater than 255 but
use only the 8 least significant bits, which means reading at offset
0x100 gets the data at offset 0x0, and causes confusion for lspci.
Reject accesses to configuration space offsets greater than 255.
Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
if (hose == NULL)
return PCIBIOS_DEVICE_NOT_FOUND;
+ if (offset > 0xff)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
+
addr = u3_ht_cfg_access(hose, bus->number, devfn, offset);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;
if (hose == NULL)
return PCIBIOS_DEVICE_NOT_FOUND;
+ if (offset > 0xff)
+ return PCIBIOS_BAD_REGISTER_NUMBER;
+
addr = u3_ht_cfg_access(hose, bus->number, devfn, offset);
if (!addr)
return PCIBIOS_DEVICE_NOT_FOUND;