On Wed, Apr 28, 2021 at 10:21:12AM -0700, Kuppuswamy, Sathyanarayanan wrote: > > > On 4/28/21 1:18 AM, Joerg Roedel wrote: > > From: Joerg Roedel <jroedel@xxxxxxx> > > > > The acpi_pci_osc_support() does an _OSC query with _OSC supported set > > to what the OS supports but a zero _OSC control value. This is > > problematic on some platforms where the firmware allows to configure > > whether DPC is under OS or Firmware control. > > Do we run acpi_pci_osc_support() only to check whether _OSC is > supported ? Or does it serve any other purpose. I am not 100% sure, but to me it looks like the pure purpose of the acpi_pci_osc_support() call was indeed to check whether the firmware is willing to grant the OS control over some PCIe features. > > When DPC is configured to be under OS control these platforms will > > issue a warning in the firmware log that the OS does not support DPC. > > Also, is there any other benefit from this patch other than fixing > a warning message in firmware? Not much other benefit, besides some removed code. But those messages can confuse the system owner and are worth getting rid of imho. Regards, Joerg
