Krzysztof Kozlowski writes:

On Sun, Sep 25, 2016 at 11:10:10PM +0200, Wolfgang Wiedmeyer wrote:

quoted

The capacity property uses the RepSOC register to report the current state
of charge. This register did not provide a reliable SOC value during my
testing with the max17047 variant on a Galaxy S3 (Trats2/GT-I9300). The
reported value did not change or even stayed zero in some cases.
However, the VF SOC register provided an accurate SOC value at all times.
It uses the voltage fuel gauge to determine the SOC.

Signed-off-by: Wolfgang Wiedmeyer <redacted>
---
 drivers/power/max17042_battery.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/power/max17042_battery.c b/drivers/power/max17042_battery.c
index da7a75f..20cb1fd 100644
--- a/drivers/power/max17042_battery.c
+++ b/drivers/power/max17042_battery.c

@@ -246,7 +246,7 @@ static int max17042_get_property(struct power_supply *psy,
 		val->intval = data * 625 / 8;
 		break;
 	case POWER_SUPPLY_PROP_CAPACITY:
-		ret = regmap_read(map, MAX17042_RepSOC, &data);
+		ret = regmap_read(map, MAX17042_VFSOC, &data);
 		if (ret < 0)
 			return ret;

The RepSOC is for ModelGauge m3 which requires current sense resistor. I
don't remember whether the resistor is present on Trats2. If not, then
m1 is used. However in both cases (m1 and m3) the battery
characteristics (cell information) should be loaded which in case of DT
driver is not supported.

Overall, I am not sure whether your change is correct. It might fix this
particular scenario because:
1. We are not providing the cell information,
2. We mre not providing the SNS resistor value so we are in m1 mode (if
there is no SNS resistor).  but it might break other applications where
SNS is present and cell configuration is provided. Unless you tested it
in such?

I'm not able to test other applications than the Galaxy S3.

Probably this should be based on Device Tree property describing what is
configured (e.g. missing model data). Maybe existing maxim,rsns-microohm
could be used - in case of lack of it, fall back to reading VFSOC?

Ok, I'll include a check if maxim,rsns-microohm exists and do the
fallback to VFSOC if it's not there.

Thanks,
Wolfgang

-- 
Website: https://fossencdi.org
Jabber: wolfgang@wiedmeyer.de
OpenPGP: 0F30 D1A0 2F73 F70A 6FEE  048E 5816 A24C 1075 7FC4
Key download: https://wiedmeyer.de/keys/ww.asc