[PATCH SLOF] pci-properties: Remove redundant call to device-type

Thomas Huth thuth at linux.vnet.ibm.com
Wed Mar 11 19:30:09 AEDT 2015


On Wed, 11 Mar 2015 19:21:02 +1100
Benjamin Herrenschmidt <benh at kernel.crashing.org> wrote:

> On Wed, 2015-03-11 at 08:50 +0100, Thomas Huth wrote:
> > 
> > - Graphic cards should get the "display" device_type ... I guess
> > that's
> >   missing now?
> 
> At least the code for qemu-vga does it

Ah, right, I was only grep'ing for "device-type" (with "-" instead of
"_", the Forth word to set the property), that's why I didn't notice
it ... the code in board-qemu sets the property directly instead.
And so does the code in board-js2x, so this should be fine.

> > - pci bus nodes (like bridges) should get the "pci" device_type
> >   according to the Open Firmware PCI binding, is this still the case?
> 
> Ah yes we need to check that, specifically P2P bridges

There seems to be some code in pci-bridge-generic-setup at last ...
Might be still worth checking this on a real running system, but at a
first glance it looks ok.

> > - Seems like there are some other devices with device_type property
> >   where SLOF does not provide an implemented interface, like in
> >   slof/fs/usb/dev-mouse.fs ... should these be revised, too?
> 
> not a *huge* issue but we can look into it when needed.

Right.

So the patch sounds now fine to me as it is.

 Thomas



More information about the Linuxppc-dev mailing list