Re: [PATCHv5 2/5] ARM: dts: Odroid-XU3 Enable TMU at Exynos5422 base

[Krzysztof Kozlowski <k.kozlowski@samsung.com>]

On 20.05.2015 23:59, Anand Moon wrote:
> On 20 May 2015 at 08:57, Dongjin Kim <tobetter@gmail.com> wrote:
>> Hello Kryzsztof,
>>
>> Are you able to check if TMU is under VDDI power domain from Exynos5422
>> datasheet?
>> If it is, XU3 use BUCK3 for TMU and more internal blocks.
>>
>> Thank you,
>> Dongjin.
>>
>> On Tue, May 19, 2015 at 4:42 PM, Krzysztof Kozlowski
>> <k.kozlowski@samsung.com> wrote:
>>>
>>> 2015-05-19 16:28 GMT+09:00 Anand Moon <linux.amoon@gmail.com>:
>>>> On 15 May 2015 at 05:42, Krzysztof Kozlowski <k.kozlowski@samsung.com>
>>>> wrote:
>>>>> 2015-05-15 1:16 GMT+09:00 Anand Moon <linux.amoon@gmail.com>:
>>>>>> On 13 May 2015 at 14:02, Krzysztof Kozlowski <k.kozlowski@samsung.com>
>>>>>> wrote:
>>>>>>> 2015-05-13 17:21 GMT+09:00 Anand Moon <linux.amoon@gmail.com>:
>>>>>>>> On 13 May 2015 at 12:51, Krzysztof Kozlowski
>>>>>>>> <k.kozlowski@samsung.com> wrote:
>>>>>>>>> 2015-05-13 15:36 GMT+09:00 Anand Moon <linux.amoon@gmail.com>:
>>>>>>>>>> This changes enables TMU IP block on the Exynos5422 Odroid-XU3
>>>>>>>>>> device.
>>>>>>>>>>
>>>>>>>>>> Tested-by: Markus Reichl <m.reichl@fivetechno.de>
>>>>>>>>>> Acked-by: Lukasz Majewski <l.majewski@samsung.com>
>>>>>>>>>> Signed-off-by: Anand Moon <linux.amoon@gmail.com>
>>>>>>>>>> ---
>>>>>>>>>>  arch/arm/boot/dts/exynos5422-odroidxu3.dts | 25
>>>>>>>>>> +++++++++++++++++++++++++
>>>>>>>>>>  1 file changed, 25 insertions(+)
>>>>>>>>>>
>>>>>>>>>> diff --git a/arch/arm/boot/dts/exynos5422-odroidxu3.dts
>>>>>>>>>> b/arch/arm/boot/dts/exynos5422-odroidxu3.dts
>>>>>>>>>> index 9446e28..cd78816 100644
>>>>>>>>>> --- a/arch/arm/boot/dts/exynos5422-odroidxu3.dts
>>>>>>>>>> +++ b/arch/arm/boot/dts/exynos5422-odroidxu3.dts
>>>>>>>>>> @@ -319,6 +319,31 @@
>>>>>>>>>>                 #cooling-cells = <2>;
>>>>>>>>>>                 cooling-levels = <0 130 170 230>;
>>>>>>>>>>         };
>>>>>>>>>> +
>>>>>>>>>> +       tmu@10060000 {
>>>>>>>>>
>>>>>>>>> Here and for other overrides please use label notation, like:
>>>>>>>>>
>>>>>>>>> &tmu_cpu0 {
>>>>>>>>>               ...
>>>>>>>>> };
>>>>>>>>>
>>>>>>>>>> +               vtmu-supply = <&ldo10_reg>;
>>>>>>>>>
>>>>>>>>> I am curious, how did you find that LDO10 supplies TMU unit?
>>>>>>>>>
>>>>>>>>> Best regards,
>>>>>>>>> Krzysztof
>>>>>>>>
>>>>>>>> Hi Krzysztof,
>>>>>>>>
>>>>>>>> I have re-base my work on earlier Lukasz Majewski patches.
>>>>>>>>
>>>>>>>> https://patchwork.kernel.org/patch/5693201/
>>>>>>>
>>>>>>> NAK.
>>>>>>> I am sorry, but this is not sufficient explanation. Actually such
>>>>>>> explanation could mean that you did just blindly copied everything
>>>>>>> instead of developing it.
>>>>>>>
>>>>>>> You cannot use some regulator here just because some Exynos4 boards
>>>>>>> use it. You have to be sure that this regulator supplies this part of
>>>>>>> SoC or device.
>>>>>>
>>>>>> Hi Krzysztof,
>>>>>>
>>>>>> After going through the schematics, I came to understanding that their
>>>>>> is
>>>>>> missing regulator related to TEMP SE in the exynos5422-odroidxu3.dts.
>>>>>>
>>>>>> Below is the schematic of the board.
>>>>>>
>>>>>> http://dn.odroid.com/5422/ODROID-XU3/Schematics/XU3_MAIN_REV0.2.PDF
>>>>>>
>>>>>>   ldo18_reg: LDO18 {
>>>>>>                                          regulator-name = "vdd_ldo18";
>>>>>>                                          regulator-min-microvolt =
>>>>>> <1800000>;
>>>>>>                                          regulator-max-microvolt =
>>>>>> <1800000>;
>>>>>>                                          regulator-always-on;
>>>>>>                                };
>>>>>
>>>>> The output of LDO18 goes to VDD_EMMC_1V8. This is not regulator for
>>>>> TMU.
>>>>>
>>>>> I think the schematics are missing some of details but it can be
>>>>> deducted that:
>>>>> 1. TEMP SE is supplied by VDD18_TS power domain. It consists of 5
>>>>> pairs of pins (XTSTEST_OUT[0-4], XTSEXT_RES[0-4]).
>>>>> 2. The VDD18_TS01, VDD18_TS23 and VDD18_TS4 are wired to theL DO7 of
>>>>> S2MPS11 PMIC.
>>>>> 3. I confirmed with the Exynos5422 datasheet that these
>>>>> VDD18_TS{01,23,4} supply the XTSTEST pins (OUT and RES).
>>>>>
>>>>> So the LDO7 it is... but before using it there is a caveat. The LDO7
>>>>> is also connected to VDD of MIPI, HDMI and few more. So when you use
>>>>> this regulator in TMU it may be turned off by TMU driver (e.g. during
>>>>> unbind). In such case these other blocks also should be tested and
>>>>> checked whether they take this regulator and enable it.
>>>>
>>>> hi Krzysztof,
>>>>
>>>> I tried to use the LDO7 regulator for TMU but it failed to register.
>>>>
>>>> [    3.231329] ina2xx 0-0045: power monitor ina231 (Rshunt = 10000 uOhm)
>>>> [    3.237691] thermal thermal_zone0: failed to read out thermal zone
>>>> (-22)
>>>> [    3.243033] exynos-tmu 10060000.tmu: Looking up vtmu-supply from
>>>> device tree
>>>> [    3.243936] thermal thermal_zone1: failed to read out thermal zone
>>>> (-22)
>>>> [    3.249791] exynos-tmu 10064000.tmu: Looking up vtmu-supply from
>>>> device tree
>>>> [    3.250677] thermal thermal_zone2: failed to read out thermal zone
>>>> (-22)
>>>> [    3.256410] exynos-tmu 10068000.tmu: Looking up vtmu-supply from
>>>> device tree
>>>> [    3.257345] thermal thermal_zone3: failed to read out thermal zone
>>>> (-22)
>>>> [    3.263050] exynos-tmu 1006c000.tmu: Looking up vtmu-supply from
>>>> device tree
>>>> [    3.263984] thermal thermal_zone4: failed to read out thermal zone
>>>> (-22)
>>>> [    3.269769] exynos-tmu 100a0000.tmu: Looking up vtmu-supply from
>>>> device tree
>>>> [    3.270363] usb 5-1: New USB device found, idVendor=0424,
>>>> idProduct=9514
>>>> [    3.276389] usb 5-1: New USB device strings: Mfr=0, Product=0,
>>>> SerialNumber=0
>>>
>>> Indeed.
>>>
> 
> Hi Krzysztof/Dongjin
> 
> BUCK3 is option for TMU as suggested by Dongjin

What do you mean by that? VDD_INT is one of important regulators. It
supplies many parts of chip and should not be disabled during exynos-tmu
driver removal.

Of course it wouldn't because it is "always-on"... so what is the
benefit of using it in exynos-tmu?

What about regulator supplying TMU sensors? Shouldn't it be enabled? Is
it the same?

Before posting a new solution please be sure that you have sufficient
answer for each of these questions. Anwser that "someone told me so"
unfortunately is not sufficient :).



> Earlier I have some missing CONFIG option's hence It was not working.
> Now its registering with TMU.
> 
> Bellow is the output device tree.
> 
> root@odroidxu3: cd /sys/firmware/devicetree/base/
> root@odroidxu3:/sys/firmware/devicetree/base# cat tmu@10060000/status
> okay
> root@odroidxu3:/sys/firmware/devicetree/base#
> root@odroidxu3:/sys/firmware/devicetree/base# cat tmu@10064000/status
> okay
> root@odroidxu3:/sys/firmware/devicetree/base# cat tmu@10068000/status
> okay
> root@odroidxu3:/sys/firmware/devicetree/base# cat tmu@1006c000/status
> okay
> root@odroidxu3:/sys/firmware/devicetree/base# cat tmu@100a0000/status
> okay
> root@odroidxu3:/sys/firmware/devicetree/base#

If you looked at the driver then you would know, that above status does
not mean anything for this discussion about regulator. You could use
EMMC regulator (which you proposed) and the results would be the same.

Best regards,
Krzysztof