arm64: dts: imx93: enable UART wakeup on Variscite boards

arch/arm64/boot/dts/freescale/imx93-var-dart-sonata.dts can0: can@0 { · line 388

[Severity: Medium]
Does adding wakeup-source to &lpuart1 pass dtbs_check or take effect at runtime?

The lpuart1 node in arch/arm64/boot/dts/freescale/imx91_93_common.dtsi uses the "fsl,imx93-lpuart", "fsl,imx8ulp-lpuart", "fsl,imx7ulp-lpuart" compatibles, which are matched by Documentation/devicetree/bindings/serial/fsl-lpuart.yaml:

required:
- compatible
- reg
- interrupts
- clocks
- clock-names

unevaluatedProperties: false

Neither fsl-lpuart.yaml nor the schemas it references (serial.yaml and rs485.yaml) define wakeup-source, so this causes an unevaluated property validation failure on serial@44380000.

In addition, neither lpuart_probe() in drivers/tty/serial/fsl_lpuart.c nor serial_core_add_one_port() in drivers/tty/serial/serial_core.c reads wakeup-source from the controller device node. Instead, serial_core_add_one_port() unconditionally marks the child tty device as wakeup-capable:

  tty_dev = tty_port_register_device_attr_serdev(port, drv->tty_driver,
                  uport->line, uport->dev, &uport->port_dev->dev, port,
                  uport->tty_groups);
  if (!IS_ERR(tty_dev)) {
          device_set_wakeup_capable(tty_dev, 1);
  }

And lpuart_uport_is_active() in drivers/tty/serial/fsl_lpuart.c checks wakeup state on that child tty_dev (configured via sysfs) rather than the platform device:

  tty = tty_port_tty_get(port);
  if (tty) {
          tty_dev = tty->dev;
          may_wake = tty_dev && device_may_wakeup(tty_dev);
          tty_kref_put(tty);
  }

Should fsl-lpuart.yaml and the driver be updated to support wakeup-source before adding this property to the devicetree?

Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1

arch/arm64/boot/dts/freescale/imx93-var-som-symphony.dts typec1_dr_sw: endpoint { · line 305

[Severity: Medium]
Is wakeup-source supported by the fsl-lpuart binding and driver here?

In arch/arm64/boot/dts/freescale/imx91_93_common.dtsi, lpuart1 uses the compatible strings fsl,imx93-lpuart, fsl,imx8ulp-lpuart, and fsl,imx7ulp-lpuart, which bind against Documentation/devicetree/bindings/serial/fsl-lpuart.yaml. That schema only references serial.yaml and rs485.yaml and specifies:

unevaluatedProperties: false

Neither fsl-lpuart.yaml, serial.yaml, nor rs485.yaml defines wakeup-source, so adding it to &lpuart1 causes a dtbs_check validation failure.

In addition, neither lpuart_probe() in drivers/tty/serial/fsl_lpuart.c nor serial_core_add_one_port() in drivers/tty/serial/serial_core.c parses wakeup-source from the UART DT node. Instead, serial_core_add_one_port() unconditionally marks the tty device wakeup-capable (leaving wakeup disabled until enabled via sysfs), and lpuart_uport_is_active() checks that tty device:

drivers/tty/serial/serial_core.c:serial_core_add_one_port() {
    ...
	if (!IS_ERR(tty_dev)) {
		device_set_wakeup_capable(tty_dev, 1);
	}
    ...
}

drivers/tty/serial/fsl_lpuart.c:lpuart_uport_is_active() {
    ...
		may_wake = tty_dev && device_may_wakeup(tty_dev);
    ...
}

Does this property actually have any effect without corresponding binding and driver support?

Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=2