Radeon FireMV 2400 PCI

[Roland Scheidegger]

Erwin Rol wrote:
> On Fri, 2006-01-27 at 04:21 +0100, Roland Scheidegger wrote:
>> Erwin Rol wrote:
>>> My question is are there any databooks, documentation etc. on the
>>> M9 chips available without a NDA?
>> Well those are just standard rv280 chips (in the pci versions of
>> the firemv cards, in contrast to the rv370 on the pci-e versions).
>> At least
> 
> Actually i think the M9 is a RV250 chip and not a RV280, at least
> that is what lspci tells me.
lspci is likely wrong in this case. I've noticed that when trying to
clean up that pci id list, the same pci id is used for two different 
devices (i.e. 4c66/4c6e) :-(. And in contrast to some phantom devices, 
both actually seem to exist. Only the subsystem id is different I 
believe between the firemv 2400 pci and some radeon 9000 mobility M9. A 
bit strange that two different chips would show up with the same id, 
unless that M9 is wrong and it's actually a M9+ (catalyst driver, 
neither normal nor mobility, does not support this chip so I don't know 
how it treats it), or the catalyst .inf file could be wrong too. This 
actually is a problem, while the chips are very similar they are treated 
slightly different in both ddx and dri, and both those components 
differentiate only by pci id.

            +-0c.0-[0000:02]--+-01.0
            |                 +-01.1
            |                 +-05.0
            |                 \-05.1

Ah, shows up with 4 ids then.

Roland