More v3 quesions

[Corey Osgood]

I've got a few more questions that I could probably figure out answers to by
testing, but hopefully it's easier just to ask:

* If a device is defined in the dts, but disabled during
phase4_enable_disable which renders the device hidden, what will/should v3
do?

* If another device appears where the disabled device should have been (in
this case, the ide controller is now where sata used to be), what should
happen? Will the ide controller's config (that was intended for another
location) be used? In this case, it's D10F0 and F1.

* phase4_enable_disable passes the actual device it's enabling/disabling,
right?

* What's the format of devfn (as in dev->path.pci.devfn)? dev << 4 |
function?

Thanks!
Corey

[ron minnich]

On Thu, Oct 23, 2008 at 5:30 PM, Corey Osgood  wrote:
> I've got a few more questions that I could probably figure out answers to by
> testing, but hopefully it's easier just to ask:
> If a device is defined in the dts, but disabled during
>
> phase4_enable_disable which renders the device hidden, what will/should v3
> do?

the device should end up not even being found. marc felt that
phase4_enable_disable was ill-defined and we are working on making
this clearer.
> If another device appears where the disabled device should have been (in
>
> this case, the ide controller is now where sata used to be), what should
> happen? Will the ide controller's config (that was intended for another
> location) be used? In this case, it's D10F0 and F1.

It should not because the devid and vendor id should not match.
> phase4_enable_disable passes the actual device it's enabling/disabling,
>
> right?

yes.
> What's the format of devfn (as in dev->path.pci.devfn)? dev << 4 |
>
> function?

dev <<3 | fn

ron

[Corey Osgood]

On Thu, Oct 23, 2008 at 9:04 PM, ron minnich  wrote:
> On Thu, Oct 23, 2008 at 5:30 PM, Corey Osgood 
> wrote:
> > I've got a few more questions that I could probably figure out answers to
>
> by
> > testing, but hopefully it's easier just to ask:
> > If a device is defined in the dts, but disabled during
> >
> > phase4_enable_disable which renders the device hidden, what will/should
>
> v3
> > do?
>
> the device should end up not even being found. marc felt that
> phase4_enable_disable was ill-defined and we are working on making
> this clearer.

Alright, but what happens after that? coreboot dies with loud warning
messages, or moves on to the next device?
> > If another device appears where the disabled device should have been
>
> (in
> > this case, the ide controller is now where sata used to be), what should
> > happen? Will the ide controller's config (that was intended for another
> > location) be used? In this case, it's D10F0 and F1.
>
> It should not because the devid and vendor id should not match.

Let me try to clarify that: if the sata controller at d10f0 disappears, and
the ide controller moves from d10f1 to d10f0 (and maintains its original
vendor/dev ids), will the dts configuration that was intended for the ide
controller at d10f1 be used? If not, could the ide controller and sata
controller both be configured for d10f0, with the sata controller set to
disabled?

Thanks!
Corey
> > phase4_enable_disable passes the actual device it's enabling/disabling,
> >
> > right?
>
> yes.
> > What's the format of devfn (as in dev->path.pci.devfn)? dev << 4 |
> >
> > function?
>
> dev <<3 | fn
>
> ron

[Peter Stuge]

Corey Osgood wrote:
> if the sata controller at d10f0 disappears, and the ide controller
> moves from d10f1 to d10f0 (and maintains its original vendor/dev ids)

Should we now look at PCI hotplug?

//Peter

[ron minnich]

On Thu, Oct 23, 2008 at 6:13 PM, Corey Osgood  wrote:
> Let me try to clarify that: if the sata controller at d10f0 disappears, and
> the ide controller moves from d10f1 to d10f0 (and maintains its original
> vendor/dev ids), will the dts configuration that was intended for the ide
> controller at d10f1 be used? If not, could the ide controller and sata
> controller both be configured for d10f0, with the sata controller set to
> disabled?

If you are doing things that change what device will be visible at a
given devfn, they should actually probably be done in phase 2.

ron

[Marc Jones]

ron minnich wrote:
> On Thu, Oct 23, 2008 at 5:30 PM, Corey Osgood  wrote:
> > I've got a few more questions that I could probably figure out answers to by
> > testing, but hopefully it's easier just to ask:
> > If a device is defined in the dts, but disabled during
> >
> > phase4_enable_disable which renders the device hidden, what will/should v3
> > do?
>
> the device should end up not even being found. marc felt that
> phase4_enable_disable was ill-defined and we are working on making
> this clearer.

It will not be enabled or given resources because the device was not 
found. The scan code scans the hardware and then matches against the 
device tree to get the correct enable and resource allocation functions. 
In the dts it will not have dev->enabled set. (at least that is how I 
think it goes....)

Marc

[Corey Osgood]

Thanks for all the answers! One more problem, I've got to "know" how much
vga memory to allocate from two different device drivers. If I have the vga
memory size in the dts for one device, and then from the other device, do
dev_find_device() to find the first one, can I access its dts info? Or, is
there any reason not to move vga memory size to Kconfig?

[Peter Stuge]

Corey Osgood wrote:
> Or, is there any reason not to move vga memory size to Kconfig?

I agree, graphics memory size when sizable should be in Kconfig!

//Peter

[Corey Osgood]

On Sun, Oct 26, 2008 at 1:35 AM, Peter Stuge  wrote:
> Corey Osgood wrote:
> > Or, is there any reason not to move vga memory size to Kconfig?
>
> I agree, graphics memory size when sizable should be in Kconfig!

Will do!

-Corey