Re: [PATCH v11 2/5] ACPI/PCI: Ignore _OSC negotiation result if pcie_ports_native is set.

[Kuppuswamy, Sathyanarayanan]

Hi Bjorn,

Thanks for the review.

On 11/25/20 12:12 PM, Bjorn Helgaas wrote:
> On Mon, Oct 26, 2020 at 07:57:05PM -0700, Kuppuswamy Sathyanarayanan wrote:
> > pcie_ports_native is set only if user requests native handling
> > of PCIe capabilities via pcie_port_setup command line option.
> > User input takes precedence over _OSC based control negotiation
> > result. So consider the _OSC negotiated result only if
> > pcie_ports_native is unset.
> >
> > Also, since struct pci_host_bridge ->native_* members caches the
> > ownership status of various PCIe capabilities, use them instead
> > of distributed checks for pcie_ports_native.
> >
> > Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@xxxxxxxxxxxxxxx>
> > ---
> >   drivers/acpi/pci_root.c           | 35 ++++++++++++++++++++++---------
> >   drivers/pci/hotplug/pciehp_core.c |  2 +-
> >   drivers/pci/pci-acpi.c            |  3 ---
> >   drivers/pci/pcie/aer.c            |  2 +-
> >   drivers/pci/pcie/portdrv_core.c   |  9 +++-----
> >   include/linux/acpi.h              |  2 ++
> >   6 files changed, 32 insertions(+), 21 deletions(-)
> >
> > diff --git a/drivers/acpi/pci_root.c b/drivers/acpi/pci_root.c
> > index c12b5fb3e8fb..a9e6b782622d 100644
> > --- a/drivers/acpi/pci_root.c
> > +++ b/drivers/acpi/pci_root.c
> > @@ -41,6 +41,12 @@ static int acpi_pci_root_scan_dependent(struct acpi_device *adev)
> >   				| OSC_PCI_CLOCK_PM_SUPPORT \
> >   				| OSC_PCI_MSI_SUPPORT)

> > +
> > +	if (pcie_ports_native) {
> > +		decode_osc_control(root, "OS forcibly taking over",
> > +				   OSC_PCI_EXPRESS_CONTROL_MASKS);
>
> The only place OSC_PCI_EXPRESS_CONTROL_MASKS is used is right here, so
> it's kind of pointless.
>
> I think I'd rather have this:
>
>    dev_info(&root->device->dev, "Ignoring PCIe-related _OSC results because \"pcie_ports=native\" specified\n");
I was trying to keep the same print format. In pci_root.c,
decode_os_control() is repeatedly used to print info related to
PCIe capability ownership.

But either way is fine with me. I can use the format you mentioned.
> followed by something like this after we're done fiddling with all the
> host_bridge->native* bits:

>    #define FLAG(x) ((x) ? '+' : '-')
>
>    dev_info(&root->device->dev, "OS native features: SHPCHotplug%c PCIeCapability%c PCIeHotplug%c PME%c AER%c DPC%c LTR%c\n",
>             FLAG(host_bridge->native_shpc_hotplug),
> 	   ?,
>             FLAG(host_bridge->native_pcie_hotplug),
> 	   ...);
>
> But I don't know how to handle OSC_PCI_EXPRESS_CAPABILITY_CONTROL
> since we don't track it the same way.  Maybe we'd have to omit it from
> this message for now?
I will add it in next version. But for now, its not worry about
OSC_PCI_EXPRESS_CAPABILITY_CONTROL.
>

> >   	/*
> >   	 * Evaluate the "PCI Boot Configuration" _DSM Function.  If it

> > --
> > 2.17.1

--
Sathyanarayanan Kuppuswamy
Linux Kernel Developer