[Intel-gfx] [patch] PCI ID review

[Ville Syrjälä]

On Mon, Apr 27, 2020 at 10:07:06AM -0700, Rodrigo Vivi wrote:
> On Mon, Apr 27, 2020 at 11:18:46AM -0400, Alexei Podtelezhnikov wrote:
> > On Mon, Apr 27, 2020 at 10:45 AM Chris Wilson <chris at chris-wilson.co.uk> wrote:
> > >
> > > Quoting Alexei Podtelezhnikov (2020-04-27 15:40:42)
> > > > >
> > > > > These do not exist. They are fake PCI-ID for Windows95 multi monitor.
> > > > > The single device appears twice on the bus as a second function. We
> > > > > never had that restriction and could do multiple monitors on the single
> > > > > device.
> > > >
> > > > Windows 10 drivers list them, they do show up on lspci and I'll quote
> > > > from Atom datasheet.
> > > > "This register is unique in Function 1 (the Function 0 DID is
> > > > separate). This difference in Device ID is necessary for allowing
> > > > distinct Plug and Play enumeration of function 1 when both function 0
> > > > and function 1 have the same class code."
> > > > Whatever this means.
> > >
> > > It means it's a hack for the Window's driver. There is no HW behind it.
> > 
> > Intel talks about two separate engines (threads?)
> > https://www.intel.com/content/www/us/en/support/articles/000005880/graphics-drivers/legacy-graphics.html
> > https://www.intel.com/content/www/us/en/support/articles/000005747/graphics-drivers.html
> > Still not buying?
> 
> Hi Alexei, thanks for these.
> 
> With this information in mind it looks very wrong to simply add the PCI ID.
> 
> ADD2 device is not supported by i915. I don't believe it will be as simple
> as just adding the PCI ID here.

ADD2==sdvo, which is supported just fine. It has nothing to do
with the second fake PCI function though.

In practice adding the fake PCI IDs shouldn't do any harm as we reject
anything but function 0 in the driver probe. But it also means that
there is no point in adding them either.

-- 
Ville Syrjälä
Intel