Issue with TPS6598x driver on ThinkPad T14 Gen1

["Michał Kopeć" <michal@nozomi.space>]

Greetings,

I've been looking into fixing this error in dmesg relating to the 
tps6598x driver on my Lenovo ThinkPad T14 Gen1 AMD:

[ 12.808694] Serial bus multi instantiate pseudo device driver 
INT3515:00: error -ENXIO: IRQ index 1 not found
[ 12.808706] Serial bus multi instantiate pseudo device driver 
INT3515:00: error -ENXIO: Error requesting irq at index 1

Here's what the device resources look like in the ACPI tables:

Method (_CRS, 0, NotSerialized) // _CRS: Current Resource Settings
{
    Name (SBFX, ResourceTemplate ()
    {
        I2cSerialBusV2 (0x0023, ControllerInitiated, 0x00061A80,
            AddressingMode7Bit, "\\_SB.I2CA",
            0x00, ResourceConsumer, , Exclusive,
            )
        I2cSerialBusV2 (0x0027, ControllerInitiated, 0x00061A80,
            AddressingMode7Bit, "\\_SB.I2CA",
            0x00, ResourceConsumer, , Exclusive,
            )
        GpioInt (Edge, ActiveLow, ExclusiveAndWake, PullUp, 0x0000,
            "\\_SB.GPIO", 0x00, ResourceConsumer, ,
            )
            { // Pin list
                0x0045
            }
    })
    Return (SBFX) /* \_SB_.I2CA.UCMX._CRS.SBFX */
}

I believe the problem is that the serial-multi-instantiate driver is 
assuming there are separate irqs defined for each of the i2c device 
instances:

static const struct smi_node int3515_data = {
    .instances = {
        { "tps6598x", IRQ_RESOURCE_APIC, 0 },
        { "tps6598x", IRQ_RESOURCE_APIC, 1 },
        { "tps6598x", IRQ_RESOURCE_APIC, 2 },
        { "tps6598x", IRQ_RESOURCE_APIC, 3 },
        {}
    },
    .bus_type = SMI_I2C,
};

However as far as I know the tps6598x family of PD controllers expose 
multiple i2c slaves (per USB-C port) with one shared irq. I modified 
the smi driver as follows:

 static const struct smi_node int3515_data = {
     .instances = {
         { "tps6598x", IRQ_RESOURCE_AUTO, 0 },
         { "tps6598x", IRQ_RESOURCE_AUTO, 0 },
         { "tps6598x", IRQ_RESOURCE_AUTO, 0 },
         { "tps6598x", IRQ_RESOURCE_AUTO, 0 },
         {}
     },
     .bus_type = SMI_I2C,
 };

which resulted in a successful probe of the driver and creation of 
port{0,1} entries in /sys/class/typec:

[ 49.742618] Serial bus multi instantiate pseudo device driver 
INT3515:00: Using gpio irq
[ 49.747584] Serial bus multi instantiate pseudo device driver 
INT3515:00: Using gpio irq
[ 49.752350] Serial bus multi instantiate pseudo device driver 
INT3515:00: Instantiated 2 I2C devices.

So I guess my question for y'all is this: would this be an acceptable 
change to make in the driver? If so, I'd be happy to push out the 
patch. Or is it a case of the ACPI tables being bugged and I need to 
pester Lenovo to fix them? Or, do we need to add logic to the driver to 
handle both cases?

It would be nice if the authors of the driver could chime in.
Kind regards / Pozdrawiam,
Michał Kopeć