[PATCH V6 1/4] firmware: ti_sci: Introduce Power Management Ops

[Andrew Davis]

On 8/3/23 10:42 AM, Dhruva Gole wrote:
> On Aug 03, 2023 at 10:14:03 -0500, Andrew Davis wrote:
> [..snip..]
>>>    #define TI_SCI_IRQ_SECONDARY_HOST_INVALID	0xff
>>>    /**
>>> diff --git a/include/linux/soc/ti/ti_sci_protocol.h b/include/linux/soc/ti/ti_sci_protocol.h
>>> index bd0d11af76c5..f2d1d74ab8fc 100644
>>> --- a/include/linux/soc/ti/ti_sci_protocol.h
>>> +++ b/include/linux/soc/ti/ti_sci_protocol.h
>>> @@ -195,6 +195,49 @@ struct ti_sci_clk_ops {
>>>    			u64 *current_freq);
>>>    };
>>> +/* TISCI LPM wake up sources */
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_WKUP_I2C0	0x00
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_WKUP_UART0	0x10
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_MCU_GPIO0	0x20
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_WKUP_ICEMELTER0	0x30
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_WKUP_TIMER0	0x40
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_WKUP_TIMER1	0x41
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_WKUP_RTC0	0x50
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_RESET		0x60
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_USB0		0x70
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_USB1		0x71
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_MAIN_IO		0x80
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_MCU_IO		0x81
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_CAN_IO		0x82
>>> +#define TISCI_MSG_VALUE_LPM_WAKE_SOURCE_INVALID		0xFF
>>
>> I don't see these used in this series, do you need them? lpm_wake_reason()
> 
> True, we are not currently using these macros. They _maybe required in
> future.
> I can remove them if required?
> 
>> doesn't seem used either and I'm not sure how you plan to use it, could
>> you detail that?
> 
> When the system wakes from suspend-to-mem we can check which
> subsystem has woken us up with the TISCI LPM_WAKEUP_REASON message.
> There's no hardware event generated and we have to ask the firmware
> for the actual wake reason.
> 
> We may want to add support for a wake up interrupt controller that will
> generate an interrupt for other subsystems. This might end up using this
> lpm_wake_reason API, hence even though the function maybe unused today
> it will be required for above described scenario.
> 
> However if you prefer that I remove it and then add it in future when we
> finally write a working interrupt controller driver, then do let me
> know.
> 

It's easier to review code that is used. We may go with a
completely different approach by then, will be easier to unwind
this if you don't start with it, only add it when used.

Andrew

>>
>> Andrew
>>
>>> +
>>> +/* TISCI LPM IO isolation control values */
>>> +#define TISCI_MSG_VALUE_IO_ENABLE			1
>>> +#define TISCI_MSG_VALUE_IO_DISABLE			0
>>> +
> [..snip..]
>