Re: [PATCH v2 1/2] PCI/APCI: Move acpi_pci_osc_support() check to negotiation phase

[Joerg Roedel]

Hi Rafael,

On Wed, Jul 14, 2021 at 02:04:17PM +0200, Rafael J. Wysocki wrote:
> >         decode_osc_support(root, "OS supports", support);
> > -       status = acpi_pci_osc_support(root, support);
> > -       if (ACPI_FAILURE(status)) {
> > -               *no_aspm = 1;
> >
> > -               /* _OSC is optional for PCI host bridges */
> > -               if ((status == AE_NOT_FOUND) && !is_pcie)
> > +       if (!pcie_ports_disabled) {
> 
> If pcie_ports_disabled is set, we don't want to request any control
> from the platform firmware at all and, specifically, we don't want to
> evaluate _OSC with the OSC_QUERY_ENABLE clear in
> capbuf[OSC_QUERY_DWORD].
> 
> I'm not sure how this is achieved after your changes.

Yeah, it isn't. The acpi_pci_osc_control_set() function will always do
an _OSC call with OSC_QUERY_ENABLE clear. I will come up with a new
approach.

Thanks,

	Joerg