power: supply: qcom_battmgr: improve charge control threshold handling

The sysfs API documentation says that drivers "round written values to the nearest supported value" for charge_control_end_threshold.

Let's do this for both thresholds, as userspace (e.g. upower) generally does not expect these writes to fail at all.

Existing userspace (in particular, upower) disables charge control by setting the start threshold to 0 and the stop threshold to 100.

Handle that by actually setting the enable bit to 0 when a start threshold of 0 was requested.

Fixes cc3e883a0625 ("power: supply: qcom_battmgr: Add charge control support")

Signed-off-by Val Packett <[email protected]>

---
  drivers/power/supply/qcom_battmgr.c | 12 ++++++++----
  1 file changed, 8 insertions(+), 4 deletions(-)

drivers/power/supply/qcom_battmgr.c +6 -3

index c8028606bba0..e6f01e0122e1 100644
@@ -257,6 +257,7 @@ struct qcom_battmgr_info {
  	unsigned int capacity_warning;
  	unsigned int cycle_count;
  	unsigned int charge_count;
+	bool charge_ctrl_enable;
  	unsigned int charge_ctrl_start;
  	unsigned int charge_ctrl_end;
  	char model_number[BATTMGR_STRING_LEN];

@@ -659,13 +660,13 @@ static int qcom_battmgr_bat_get_property(struct power_supply *psy,
  }
 
  static int qcom_battmgr_set_charge_control(struct qcom_battmgr *battmgr,
-					   u32 target_soc, u32 delta_soc)
+					   bool enable, u32 target_soc, u32 delta_soc)
  {
  	struct qcom_battmgr_charge_ctrl_request request = {
  		.hdr.owner = cpu_to_le32(PMIC_GLINK_OWNER_BATTMGR),
  		.hdr.type = cpu_to_le32(PMIC_GLINK_REQ_RESP),
  		.hdr.opcode = cpu_to_le32(BATTMGR_CHG_CTRL_LIMIT_EN),
-		.enable = cpu_to_le32(1),
+		.enable = cpu_to_le32(enable),
  		.target_soc = cpu_to_le32(target_soc),
  		.delta_soc = cpu_to_le32(delta_soc),
  	};

@@ -677,6 +678,7 @@ static int qcom_battmgr_set_charge_start_threshold(struct qcom_battmgr *battmgr,
  {
  	u32 target_soc, delta_soc;
  	int ret;
+	bool enable = start_soc != 0;
 
  	start_soc = clamp(start_soc, CHARGE_CTRL_START_THR_MIN, CHARGE_CTRL_START_THR_MAX);
 

@@ -696,9 +698,10 @@ static int qcom_battmgr_set_charge_start_threshold(struct qcom_battmgr *battmgr,
  	}
 
  	mutex_lock(&battmgr->lock);
-	ret = qcom_battmgr_set_charge_control(battmgr, target_soc, delta_soc);
+	ret = qcom_battmgr_set_charge_control(battmgr, enable, target_soc, delta_soc);
  	mutex_unlock(&battmgr->lock);
  	if (!ret) {
+		battmgr->info.charge_ctrl_enable = enable;
  		battmgr->info.charge_ctrl_start = start_soc;
  		battmgr->info.charge_ctrl_end = target_soc;
  	}

drivers/power/supply/qcom_battmgr.c +1 -1

  	end_soc = clamp(end_soc, CHARGE_CTRL_END_THR_MIN, CHARGE_CTRL_END_THR_MAX);
 

@@ -717,7 +721,7 @@ static int qcom_battmgr_set_charge_end_threshold(struct qcom_battmgr *battmgr, i
  		delta_soc = end_soc - battmgr->info.charge_ctrl_start;
 
  	mutex_lock(&battmgr->lock);
-	ret = qcom_battmgr_set_charge_control(battmgr, end_soc, delta_soc);
+	ret = qcom_battmgr_set_charge_control(battmgr, enable, end_soc, delta_soc);
  	mutex_unlock(&battmgr->lock);
  	if (!ret) {
  		battmgr->info.charge_ctrl_start = end_soc - delta_soc;

Currently, upowerd is unable to turn off the battery preservation mode[1] on Qualcomm laptops, because it does that by setting the start threshold to zero and the driver returns an error:

pmic_glink.power-supply.0: charge control start threshold exceed range: [50 - 95]

Kernel documentation says the end threshold must be clamped[2] but does not say anything about the start threshold.

[...]

Quoted 9 lines · Val Packett

Currently, upowerd is unable to turn off the battery preservation mode[1] on Qualcomm laptops, because it does that by setting the start threshold to zero and the driver returns an error:

pmic_glink.power-supply.0: charge control start threshold exceed range: [50 - 95]

Kernel documentation says the end threshold must be clamped[2] but does not say anything about the start threshold.

Quoted 1 line

[...]

Applied, thanks!

[1/2] power: supply: qcom_battmgr: clamp charge control thresholds

       commit: 8809980fdc8a86070667032fa4005ee83f1c62f3
[2/2] power: supply: qcom_battmgr: support disabling charge control
       commit: 446fcf494691da4e685923e5fad02b163955fc0e

Quoted 10 lines · Sebastian Reichel

Quoted 9 lines · Val Packett

Currently, upowerd is unable to turn off the battery preservation mode[1] on Qualcomm laptops, because it does that by setting the start threshold to zero and the driver returns an error:

pmic_glink.power-supply.0: charge control start threshold exceed range: [50 - 95]

Kernel documentation says the end threshold must be clamped[2] but does not say anything about the start threshold.

Quoted 7 lines

Quoted 1 line

[...]

Applied, thanks!

[1/2] power: supply: qcom_battmgr: clamp charge control thresholds

       commit: 8809980fdc8a86070667032fa4005ee83f1c62f3
[2/2] power: supply: qcom_battmgr: support disabling charge control
       commit: 446fcf494691da4e685923e5fad02b163955fc0e

Woahh.. please revert the second one.

I'm sorry, I thought this was discussed here but apparently it was only on IRC and I must've assumed that the patches weren't going anywhere because of the lack of R-b..

The disable bit was acting rather strange after all, we'd need more work to figure out if that's even possible. Let's leave it at the clamp only.

Currently, upowerd is unable to turn off the battery preservation mode[1] on Qualcomm laptops, because it does that by setting the start threshold to zero and the driver returns an error:

pmic_glink.power-supply.0: charge control start threshold exceed range: [50 - 95]

Kernel documentation says the end threshold must be clamped[2] but does not say anything about the start threshold.

In this proposal I've special-cased start==0 to actually disable the functionality via the enable bit, and otherwise made both start and end thresholds be clamped to the acceptable range. Hopefully that's fine?

Or should the [1 - 49] range for start actually be rejected?

[1]: https://gitlab.freedesktop.org/upower/upower/-/issues/327
[2]: https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-class-power

Thanks,
~val

Val Packett (2):
power: supply: qcom_battmgr: clamp charge control thresholds
power: supply: qcom_battmgr: support disabling charge control

  drivers/power/supply/qcom_battmgr.c | 26 ++++++++++----------------
  1 file changed, 10 insertions(+), 16 deletions(-)

Quoted 9 lines · Val Packett

Currently, upowerd is unable to turn off the battery preservation mode[1] on Qualcomm laptops, because it does that by setting the start threshold to zero and the driver returns an error:

pmic_glink.power-supply.0: charge control start threshold exceed range: [50 - 95]

Kernel documentation says the end threshold must be clamped[2] but does not say anything about the start threshold.

Quoted 4 lines

In this proposal I've special-cased start==0 to actually disable the functionality via the enable bit, and otherwise made both start and end thresholds be clamped to the acceptable range. Hopefully that's fine?

It is fine to clamping the threshold to the acceptable range. Thank you for making the changes.

Quoted 1 line

Or should the [1 - 49] range for start actually be rejected?

The minimum charging start threshold was set to 50 to improve user experience. If the threshold is too low and the system keeps drawing power from the battery frequently due to a large system load and a weak charger, the laptop will only begin charging when the battery level falls below that threshold. If the user disconnects the charger at that time, then the device would be only having a battery below 50%. Setting the threshold at 50 ensures the battery always stays above 50%.

Quoted 15 lines · Fenglin Wu

Quoted 6 lines · Val Packett

Currently, upowerd is unable to turn off the battery preservation mode[1] on Qualcomm laptops, because it does that by setting the start threshold to zero and the driver returns an error:

pmic_glink.power-supply.0: charge control start threshold exceed range: [50 - 95]

Quoted 7 lines

Kernel documentation says the end threshold must be clamped[2] but does not say anything about the start threshold.

In this proposal I've special-cased start==0 to actually disable the functionality via the enable bit, and otherwise made both start and end thresholds be clamped to the acceptable range. Hopefully that's fine?

It is fine to clamping the threshold to the acceptable range. Thank you for making the changes.

Quoted 2 lines

Quoted 1 line

Or should the [1 - 49] range for start actually be rejected?

The minimum charging start threshold was set to 50 to improve user experience. If the threshold is too low and the system keeps drawing power from the battery frequently due to a large system load and a weak charger, the laptop will only begin charging when the battery level falls below that threshold. If the user disconnects the charger at that time, then the device would be only having a battery below 50%. Setting the threshold at 50 ensures the battery always stays above 50%.

So can we set it lower?

Such decisions are best deferred to userspace and/or the user, which can limit what the kernel exposes as necessary/deemed useful

Konrad

Quoted 23 lines · Konrad Dybcio

Quoted 10 lines · Fenglin Wu

Quoted 9 lines · Val Packett

Currently, upowerd is unable to turn off the battery preservation mode[1] on Qualcomm laptops, because it does that by setting the start threshold to zero and the driver returns an error:

pmic_glink.power-supply.0: charge control start threshold exceed range: [50 - 95]

Kernel documentation says the end threshold must be clamped[2] but does not say anything about the start threshold.

Quoted 7 lines

Quoted 4 lines

In this proposal I've special-cased start==0 to actually disable the functionality via the enable bit, and otherwise made both start and end thresholds be clamped to the acceptable range. Hopefully that's fine?

It is fine to clamping the threshold to the acceptable range. Thank you for making the changes.

Quoted 1 line

Or should the [1 - 49] range for start actually be rejected?

The minimum charging start threshold was set to 50 to improve user experience. If the threshold is too low and the system keeps drawing power from the battery frequently due to a large system load and a weak charger, the laptop will only begin charging when the battery level falls below that threshold. If the user disconnects the charger at that time, then the device would be only having a battery below 50%. Setting the threshold at 50 ensures the battery always stays above 50%.

So can we set it lower?

Such decisions are best deferred to userspace and/or the user, which can limit what the kernel exposes as necessary/deemed useful

Quoted 1 line

Konrad

Yes, it can be set to a lower value.

However, I am still having concerns that the inappropriate start and end threshold settings would cause a very bad user experience if they are misused, since these thresholds are stored in nvmem and they won't be reset until battery is unplugged or completely drained. For example, if someone intentionally sets the start threshold to 1 and end threshold to 6, and if the laptop was shutdown with a battery SoC less than the end threshold, I am not sure if <6% percent battery level would be good enough to boot up the laptop successfully, if it is not, then the laptop may not have chance to charge up until you hot plug the battery.

Also, from battery management firmware point of view, the charge control feature was mainly designed for battery health management, to slow the aging of Li-ion battery by preventing it from being frequently charged to full state. Having a too low minimum start threshold setting won't help anything on that.