On 13/08/24 09:19, Dhruva Gole wrote:

On Aug 12, 2024 at 14:16:45 -0700, Kevin Hilman wrote:
[...]

quoted

quoted

quoted

+/**
+ * struct ti_sci_msg_resp_lpm_wake_reason - Response for TI_SCI_MSG_LPM_WAKE_REASON.
+ *
+ * @hdr:		Generic header.
+ * @wake_source:	The wake up source that woke soc from LPM.
+ * @wake_timestamp:	Timestamp at which soc woke.
+ *
+ * Response to a generic message with message type TI_SCI_MSG_LPM_WAKE_REASON,
+ * used to query the wake up source from low power mode.
+ */
+struct ti_sci_msg_resp_lpm_wake_reason {
+	struct ti_sci_msg_hdr hdr;
+	u32 wake_source;
+	u64 wake_timestamp;
+} __packed;
+

It looks like we forgot to update the parameters in this API.
See [1]:

struct tisci_msg_lpm_wake_reason_req

We're missing here the wake_pin, mode and 2 rsvd fields as well.

[1] https://software-dl.ti.com/tisci/esd/latest/2_tisci_msgs/pm/lpm.html#tisci-msg-lpm-wake-reason

The docs mention how to interpret the pin, but not the mode.  This
driver should translate this mode number to some human-readable state
for better debug messages.

Kevin

Agreed, we would all benefit from a logic that prints the human-readable
state. The way to interpret the mode would be just as:
https://software-dl.ti.com/tisci/esd/latest/2_tisci_msgs/pm/lpm.html#supported-low-power-modes

So for eg. 0x0 = TISCI_MSG_VALUE_SLEEP_MODE_DEEP_SLEEP

Missing the call to ti_sci_msg_cmd_lpm_wake_reason in resume path.

Regards,
Akashdeep Kaur