On Thu, 12 Mar 2020 at 23:45, Florian Fainelli <f.fainelli@xxxxxxxxx> wrote:
>
> On 3/12/20 2:35 PM, Vladimir Oltean wrote:
> > On Thu, 12 Mar 2020 at 18:44, Michael Walle <michael@xxxxxxxx> wrote:
> >>
> >> If there is no specific configuration of the felix switch in the device
> >> tree, but only the default configuration (ie. given by the SoCs dtsi
> >> file), the probe fails because no CPU port has been set. On the other
> >> hand you cannot set a default CPU port because that depends on the
> >> actual board using the switch.
> >>
> >> [ 2.701300] DSA: tree 0 has no CPU port
> >> [ 2.705167] mscc_felix 0000:00:00.5: Failed to register DSA switch: -22
> >> [ 2.711844] mscc_felix: probe of 0000:00:00.5 failed with error -22
> >>
> >> Thus let the device tree disable this device entirely, like it is also
> >> done with the enetc driver of the same SoC.
> >>
> >> Signed-off-by: Michael Walle <michael@xxxxxxxx>
> >> ---
> >> drivers/net/dsa/ocelot/felix.c | 5 +++++
> >> 1 file changed, 5 insertions(+)
> >>
> >> diff --git a/drivers/net/dsa/ocelot/felix.c b/drivers/net/dsa/ocelot/felix.c
> >> index 69546383a382..531c7710063f 100644
> >> --- a/drivers/net/dsa/ocelot/felix.c
> >> +++ b/drivers/net/dsa/ocelot/felix.c
> >> @@ -699,6 +699,11 @@ static int felix_pci_probe(struct pci_dev *pdev,
> >> struct felix *felix;
> >> int err;
> >>
> >> + if (pdev->dev.of_node && !of_device_is_available(pdev->dev.of_node)) {
> >> + dev_info(&pdev->dev, "device is disabled, skipping\n");
> >> + return -ENODEV;
> >> + }
> >> +
> >
> > IMHO since DSA is already dependent on device tree for PHY bindings,
> > it would make more sense to move this there:
>
> Michael's solution makes more sense, as this is a driver specific
> problem whereby you have a pci_dev instance that is created and does not
> honor the status property provided in Device Tree. If you were to look
> for a proper solution it would likely be within the PCI core actually.
>
> No other DSA switch has that problem because they use the
> I2C/SPI/platform_device/GPIO/whatever entry points into the driver model.
True, my problem with doing it in the PCI core is that "the book" [0]
doesn't actually say anything about the "status" property, so this
patch might get some pushback from the PCI maintainers (but I don't
actually know):
diff --git a/drivers/pci/probe.c b/drivers/pci/probe.c
index 512cb4312ddd..50c2b3da134a 100644
--- a/drivers/pci/probe.c
+++ b/drivers/pci/probe.c
@@ -2281,6 +2281,12 @@ static struct pci_dev *pci_scan_device(struct
pci_bus *bus, int devfn)
pci_set_of_node(dev);
+ if (dev->dev.of_node && !of_device_is_available(dev->dev.of_node)) {
+ pci_bus_put(dev->bus);
+ kfree(dev);
+ return NULL;
+ }
+
if (pci_setup_device(dev)) {
pci_bus_put(dev->bus);
kfree(dev);
OF bindings are completely optional with PCI, as you probably already
know. But there's nothing "driver" specific in disabling (i.e. not
probing) a device.
> --
>
[0] https://www.openfirmware.info/data/docs/bus.pci.pdf
Thanks,
-Vladimir
