Re: [PATCH net-next 2/2] net: phy: sfp: enable i2c-bus detection on ACPI based systems

Andrew Lunn writes:

On Sun, May 05, 2019 at 03:05:23PM -0700, Ruslan Babayev wrote:

quoted

Lookup I2C adapter using the "i2c-bus" device property on ACPI based
systems similar to how it's done with DT.

An example DSD describing an SFP on an ACPI based system:

Device (SFP0)
{
    Name (_HID, "PRP0001")
    Name (_DSD, Package ()
    {
        ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
        Package () {
            Package () { "compatible", "sff,sfp" },
            Package () { "i2c-bus", \_SB.PCI0.RP01.I2C.MUX.CH0 },
        },
    })
}

Hi Ruslan

So this gives you the I2C bus. But what about the 6 GPIOs? And the
maximum power property? You are defining the ACPI interface which from
now on everybody has to follow. So it would be good to make it
complete. ACPI also seems to be poorly documented. There does not
appear to be anything like Documentation/devicetree. So having one
patch, with a good commit message, which implements everything makes
it easier for those that follow.

Hi Andrew,

I had the GPIOs and the "maximum-power" property in my ACPI snippet initially,
but then decided to take it out thinking it was not relevant for the
current patch. I can add the missing pieces back in V2.
This is what it would like:

Device (SFP0)
{
    Name (_HID, "PRP0001")
    Name (_CRS, ResourceTemplate()
    {
        GpioIo(Exclusive, PullDefault, 0, 0, IoRestrictionNone,
               "\\_SB.PCI0.RP01.GPIO", 0, ResourceConsumer)
            { 0, 1, 2, 3, 4 }
    })
    Name (_DSD, Package ()
    {
        ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
        Package () {
            Package () { "compatible", "sff,sfp" },
            Package () { "i2c-bus", \_SB.PCI0.RP01.I2C.MUX.CH0 },
            Package () { "maximum-power-milliwatt", 1000 },
            Package () { "tx-disable-gpios", Package () { ^SFP0, 0, 0, 1} },
            Package () { "reset-gpio",       Package () { ^SFP0, 0, 1, 1} },
            Package () { "mod-def0-gpios",   Package () { ^SFP0, 0, 2, 1} },
            Package () { "tx-fault-gpios",   Package () { ^SFP0, 0, 3, 0} },
            Package () { "los-gpios",        Package () { ^SFP0, 0, 4, 1} },
        },
    })
}

This appears to be enough to get a very minimal SFP instantiated. But
then what?  How are you using it? How do you instantiate a Phylink
instance for the MAC? How do you link the SFP to the Phylink?

Before accepting this patch, i would like to know more about the
complete solution.

Thanks
	Andrew

I haven't gotten that far yet, but for the Phylink I was thinking something along the
lines of:

Device (PHY0)
{
    Name (_HID, "PRP0001")
    Name (_DSD, Package ()
    {
        ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"),
        Package () {
            Package () { "compatible", "ethernet-phy-ieee802.3-c45" },
            Package () { "sfp", \_SB.PCI0.RP01.SFP0 },
        },
    })
}

Phylink is already using the fwnode_property_get_reference_args(fwnode,
"sfp", ...), so it should work with ACPI.

I don't have a complete solution working yet. With these patches
I was hoping to get some early feedback.

Thanks,
Ruslan