Re: [patch 2.5.31] transparent PCI-to-PCI bridges

[Benjamin Herrenschmidt]

>> Why not
>> 	if ((dev->class & 0xff) == 0x01)
>> 
>> Is the lowest bit an indicator of subtractive decoding, or is 
>> Progif==0x01 the indicator of subtractive decoding?
>
>The latter.
>
>> The code and the comment should match.
>
>Ok. Updated patch appended.

While we are at it, I still think the loop copying parent resource
pointers in the case of a transparent bridge should copy the 4
resource pointers of the parent and not only 3.
The structure pci_bus has 4 slots, let's copy them all, we really
don't need to care about the fact that the parent is a PCI<->PCI bridge
(using 3 slots), a Cardbus bridge, or a host bridge or whatever wants to
define a slighly different layout for those resources at this point, we
just want _all_ of the parent resources to be copied.

There are archs where host bridges may define 4 resources, I don't
see how it would break anything to take care of copying them all
and not only the first 3 ones ;) I know some code in setup-bus.c
won't cope well with such a layout, but it typically happens on arch
like PPC that don't use setup-bus.

Ben.