Update uart_mcux.c

[Yusmed]

adding pm support

[ceolin]

you are defining it but not binding to the driver instance. Define it inside the UART_MCUX_INIT loop and bind it to the instance. Basically, on DEVICE_DT_INST_DEFINE instead of NULL you pass PM_DEVICE_DT_INST_GET(n)

[ceolin]

Thanks a lot to send it ! Please address the comments and remove the merge commit. You just need the commit with your change.

[ceolin]

@mmahadevan108 can you check it please ?

[mmahadevan108]

Minor issue, please fix the indentation of the break commands

[mmahadevan108]

Looks like this bit keeps the UART clock running during low power modes. Should this be something the user gets to decide through a device tree property?

[mmahadevan108]

Another point is that we don't need to clear and enable this bit in suspend and resume. This is a one time operation where the user decides if the UART clock should be active in low power modes.

[ceolin]

that is not correct, what you have to do is;

#define UART_MCUX_INIT(n)						\
	PINCTRL_DEFINE(n)						\
									\
	static struct uart_mcux_data uart_mcux_##n##_data = {		\
		.uart_cfg = {						\
			.stop_bits = UART_CFG_STOP_BITS_1,		\
			.data_bits = UART_CFG_DATA_BITS_8,		\
			.baudrate  = DT_INST_PROP(n, current_speed),	\
			.parity    = UART_CFG_PARITY_NONE,		\
			.flow_ctrl = DT_INST_PROP(n, hw_flow_control) ?	\
				UART_CFG_FLOW_CTRL_RTS_CTS : UART_CFG_FLOW_CTRL_NONE,\
		},							\
	};								\
									\
	static const struct uart_mcux_config uart_mcux_##n##_config;	\
									\
        PM_DEVICE_DT_INST_DEFINE(n, uart_mcux_pm_action);\
	DEVICE_DT_INST_DEFINE(n,					\
			    &uart_mcux_init,				\
			    PM_DEVICE_DT_INST_GET(n),					\
			    &uart_mcux_##n##_data,			\
			    &uart_mcux_##n##_config,			\
			    PRE_KERNEL_1,				\
			    CONFIG_SERIAL_INIT_PRIORITY,		\
			    &uart_mcux_driver_api);			\
									\
	UART_MCUX_CONFIG_FUNC(n)					\
									\
	UART_MCUX_INIT_CFG(n);

DT_INST_FOREACH_STATUS_OKAY(UART_MCUX_INIT)

[edersondisouza]

The commit message doesn't make much sense. The bit " Using bit mask instead of individual bit
manipulation" appears to belong to an update of the PR. Just state what the patch does and any noteworthy stuff, if needed (like, stating that PM for this device means gating the clock for it).