Re: [PATCH 2/2] extcon intel-cht-wc: Enable external charger

[Yauhen Kharuzhy]

On Thu, Feb 14, 2019 at 05:31:48PM +0100, Hans de Goede wrote:
> Hi,
> 
> On 10-02-19 21:36, Yauhen Kharuzhy wrote:
> > In some configuration external charge "#charge enable" signal is
> > connected to PMIC. Enable it at device probing to allow charging.
> > 
> > Tested at Lenovo Yoga Book (YB1-X91L).
> > 
> > Signed-off-by: Yauhen Kharuzhy <jekhor@xxxxxxxxx>
> > ---
> >   drivers/extcon/extcon-intel-cht-wc.c | 33 ++++++++++++++++++++++++++++
> >   1 file changed, 33 insertions(+)
> > 
> > diff --git a/drivers/extcon/extcon-intel-cht-wc.c b/drivers/extcon/extcon-intel-cht-wc.c
> > index 4f6ba249bc30..00cb3084955e 100644
> > --- a/drivers/extcon/extcon-intel-cht-wc.c
> > +++ b/drivers/extcon/extcon-intel-cht-wc.c
> > @@ -57,6 +57,11 @@
> >   #define CHT_WC_USBSRC_TYPE_OTHER	8
> >   #define CHT_WC_USBSRC_TYPE_DCP_EXTPHY	9
> > +#define CHT_WC_CHGDISCTRL		0x5e2f
> > +#define CHT_WC_CHGDISCTRL_CCSM_DIS	0x11
> > +#define CHT_WC_CHGDISCTRL_CCSM_EN	0x00
> 
> Hmm, the enable mask here does not match the enable mask from:
> 
> https://github.com/01org/ProductionKernelQuilts/blob/master/uefi/cht-m1stable/patches/EM-Charger-Disable-battery-charging-in-S3-and-enable.patch
> 
> Which has:
> 
> #define CHGDISFN_EN_CCSM_VAL           0x50
> #define CHGDISFN_DIS_CCSM_VAL          0x11
> #define CHGDISFN_CCSM_MASK             0x51
> 
> Where as on my hardware, the PMIC comes up with 0x50
> in the 0x5e2f register, exactly matching the values
> from that patch.
> 
> Why did you change this value ?

Good question... I found this values in Lenovo's sources and use them
'as is':
https://github.com/jekhor/yogabook-linux-android-kernel/blob/b7aa015ab794b516da7b6cb76e5e2d427e3b8b0c/drivers/power/intel_pmic_ccsm.h#L255

I don't remember if charger worked with Intel's value, I need to
re-check this.

(...why does Intel close documentation for this piece of HW?...)

> 
> It would be interesting to get a dump of the
> charger's i2c registers using i2cdump before
> and after writing this new value to register
> 0x5e2f. Note you can simply leave the patch adding
> the write out of the kernel, and then do:
> 
> i2cdump <charger>
> i2cget -y -f 6 0x5e 0x2f
> # check 0x00 is the correct val
> i2cset -y -f 6 0x5e 0x2f 0x00
> i2cdump <charger>
> 
> To see if any reg values of the charger-ic have changed.
> 
> Note bus 6 is the right bus for the PMIC on my systems,
> but it might be different, to find the right bus do:
> 
> ls -l /sys/bus/devices/i2c-INT34D3:00
> 
> And look at what the i2c bus-number is in that symlink.
> 
> 
> Also can you please provide i2cdump output for the
> 0x5e and 0x6e devices of the PMIC, preferably from a
> clean boot, without your patches, so that we can
> see what the value of the 0x5e2f register is before
> your code modifies it.
> 
> i2cdump -y -f 6 0x5e
> i2cdump -y -f 6 0x6e

OK, I will send this dumps soon. I used i2get/i2cset for checking this
values and PMIC behaviour already but don't remember exactly results.


-- 
Yauhen Kharuzhy