Attention is currently required from: Felix Singer, Subrata Banik, Michael Niewöhner, Patrick Rudolph.

View Change

1 comment:

• File src/soc/intel/tigerlake/pcie_rp.c:

  • Patch Set #1, Line 21:

    static bool is_part_of_group(const struct device *dev,
    			     const struct pcie_rp_group *groups)
    {
    	if (dev->path.type != DEVICE_PATH_PCI)
    		return false;
    
    	const unsigned int slot_to_find = PCI_SLOT(dev->path.pci.devfn);
    	const unsigned int fn_to_find = PCI_FUNC(dev->path.pci.devfn);
    	const struct pcie_rp_group *group;
    	unsigned int i;
    	unsigned int fn;
    
    	for (group = groups; group->count; ++group) {
    		for (i = 0, fn = rp_start_fn(group); i < group->count; i++, fn++) {
    			if (slot_to_find == group->slot && fn_to_find == fn)
    				return true;
    		}
    	}
    
    	return false;
    }
    
    

    This seems pretty generic and could be common code. The inner loop […]

    It could be. The only consumer of it for now is the PMC IPC driver, which IIUC is only applicable for TGL onwards. I believe prior to TGL the RTD3 sequence for PCIe RPs was quite different and did not involve sending any PMC IPC messages.

To view, visit change 60943. To unsubscribe, or for help writing mail filters, visit settings.