Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Message-ID: <5d2ffeca-b41c-20b4-454b-703c39bbb29e@linaro.org>
Date:   Wed, 26 Oct 2022 10:02:44 +0200
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:102.0) Gecko/20100101
 Thunderbird/102.3.1
From:   Neil Armstrong <neil.armstrong@linaro.org>
Reply-To: neil.armstrong@linaro.org
Subject: Re: [PATCHv3] arm64: dts: meson: Enable active coling using gpio-fan
 on Odroid N2/N2+
To:     Anand Moon <linux.amoon@gmail.com>,
        Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Cc:     Rob Herring <robh+dt@kernel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
        Kevin Hilman <khilman@baylibre.com>,
        Jerome Brunet <jbrunet@baylibre.com>,
        Dan Johansen <strit@manjaro.org>, devicetree@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org,
        linux-amlogic@lists.infradead.org, linux-kernel@vger.kernel.org
References: <20221022084737.1028-1-linux.amoon@gmail.com>
 <CAFBinCBu_-0m9JeAr5kd-v7Z5LTi7w0WmmUYybW_kL4KJXXpOQ@mail.gmail.com>
 <CANAwSgRzdD0FWg+z6hTFs7KvpsD64bChX-k0dPXJfACXZH2zbQ@mail.gmail.com>
 <CAFBinCCNJiL-ZKRYesQAwys6bBMpYHJbUDK-Zi_VhGDVSvF7uQ@mail.gmail.com>
 <CANAwSgTLTCA9=WvWXcR-40baauN3kAk2qx9k4FQOPCa1+_O0Yw@mail.gmail.com>
Content-Language: en-US
Organization: Linaro Developer Services
In-Reply-To: <CANAwSgTLTCA9=WvWXcR-40baauN3kAk2qx9k4FQOPCa1+_O0Yw@mail.gmail.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

Hi,

On 25/10/2022 20:06, Anand Moon wrote:
> Hi Martin,
> 
> On Sat, 22 Oct 2022 at 17:22, Martin Blumenstingl
> <martin.blumenstingl@googlemail.com> wrote:
>>
>> Hi Anand,
>>
>> On Sat, Oct 22, 2022 at 1:27 PM Anand Moon <linux.amoon@gmail.com> wrote:
>> [...]
>>>>> @@ -1982,7 +1982,6 @@ pwm_ao_d_10_pins: pwm-ao-d-10 {
>>>>>                                                  mux {
>>>>>                                                          groups = "pwm_ao_d_10";
>>>>>                                                          function = "pwm_ao_d";
>>>>> -                                                       bias-disable;
>>>> &pwm_ao_d_10_pins is not referenced anywhere so it seems that this
>>>> change has no impact on controlling the fan on Odroid-N2(+).
>>>> How did you test this change?
>>>>
>>> Ok I felt these changes affect the behavior of the pinctrl
>>>
>>>    * @PIN_CONFIG_BIAS_DISABLE: disable any pin bias on the pin, a
>>>   *  transition from say pull-up to pull-down implies that you disable
>>>   *  pull-up in the process, this setting disables all biasing.
>>>
>>> I mapped this is linked in pinctrl driver, pwm_ao_d_10_pins GPIOAO_10 see below
>> Yes, I understand this part.
>> My concern is: &pwm_ao_d_10_pins settings only become active when this
>> node is actively referenced. You can even see it in your output
>> below...
>>
>> [...]
>>> pin 10 (GPIOAO_10): (MUX UNCLAIMED) aobus-banks:1958
>> This shows that it's used as a GPIO. If the &pwm_ao_d_10_pins setting
>> was used then it would show "function pwm_ao_d group pwm_ao_d_10"
>> (similar to what GPIOE_1 shows in your output)
>>
>> If you want to know if a pull-up/down is enabled you can look at the output of:
>> $ cat /sys/kernel/debug/pinctrl/ff800000.sys-ctrl\:pinctrl@14-pinctrl-meson/pinconf-pins
>> (I'm sure this can also be retrieved from some userspace tools, but I
>> don't know how)
>>
> 
> I now switch using pwm-fan with the local changes I am able to link
> pwm_ao_d_10_pins
> but now the issue is fan keeps on spinning on boot-up and stays on.
> 
> I can manually turn on off by using
> $ sudo gpioset gpiochip1 10=1   // fan on
> $ sudo gpioset gpiochip1 10=0   // fan off

By doing that actually override the PWM function of the pin and set it as a GPIO.

> 
> It is not controlled by the thermal tip as expected.
> I feel some configuration is missing in pwm-meson driver.
> Any input for me?
> 
> $ sudo cat /sys/kernel/debug/pinctrl/ff800000.sys-ctrl\:pinctrl@14-pinctrl-meson/pinmux-pins
> [sudo] password for alarm:
> Pinmux settings per pin
> Format: pin (name): mux_owner gpio_owner hog?
> pin 0 (GPIOAO_0): ff803000.serial (GPIO UNCLAIMED) function uart_ao_a
> group uart_ao_a_tx
> pin 1 (GPIOAO_1): ff803000.serial (GPIO UNCLAIMED) function uart_ao_a
> group uart_ao_a_rx
> pin 2 (GPIOAO_2): (MUX UNCLAIMED) aobus-banks:1950
> pin 3 (GPIOAO_3): (MUX UNCLAIMED) (GPIO UNCLAIMED)
> pin 4 (GPIOAO_4): (MUX UNCLAIMED) (GPIO UNCLAIMED)
> pin 5 (GPIOAO_5): ff808000.ir (GPIO UNCLAIMED) function
> remote_ao_input group remote_ao_input
> pin 6 (GPIOAO_6): (MUX UNCLAIMED) (GPIO UNCLAIMED)
> pin 7 (GPIOAO_7): (MUX UNCLAIMED) (GPIO UNCLAIMED)
> pin 8 (GPIOAO_8): (MUX UNCLAIMED) aobus-banks:1956
> pin 9 (GPIOAO_9): (MUX UNCLAIMED) aobus-banks:1957
> pin 10 (GPIOAO_10): ff807000.pwm (GPIO UNCLAIMED) function pwm_ao_d
> group pwm_ao_d_10
> pin 11 (GPIOAO_11): (MUX UNCLAIMED) aobus-banks:1959
> pin 12 (GPIOE_0): (MUX UNCLAIMED) (GPIO UNCLAIMED)
> pin 13 (GPIOE_1): ff802000.pwm (GPIO UNCLAIMED) function pwm_ao_d
> group pwm_ao_d_e
> pin 14 (GPIOE_2): ffd1b000.pwm (GPIO UNCLAIMED) function pwm_a_e group pwm_a_e
> 
> $ sudo cat /sys/kernel/debug/pwm
> platform/ffd1b000.pwm, 2 PWM devices
>   pwm-0   (regulator-vddcpu-a  ): requested enabled period: 1250 ns
> duty: 838 ns polarity: normal
>   pwm-1   ((null)              ): period: 0 ns duty: 0 ns polarity: normal
> 
> platform/ff807000.pwm, 2 PWM devices
>   pwm-0   (pwm-fan             ): requested period: 1250 ns duty: 0 ns
> polarity: normal
>   pwm-1   ((null)              ): period: 0 ns duty: 0 ns polarity: normal

This should be on the pwm-1, hence the "pwm_AO_cd" name, "c" and "d" and the
names of the outputs.

So you need to use 1 as first PWM phandle argument instead of 0.

> 
> platform/ff802000.pwm, 2 PWM devices
>   pwm-0   ((null)              ): period: 0 ns duty: 0 ns polarity: normal
>   pwm-1   (regulator-vddcpu-b  ): requested enabled period: 1250 ns
> duty: 1213 ns polarity: normal
> 
> I could observe a change in duty when we have stress testing the CPU.

Can you share the complete change you did here ?

> 
> Thanks
> 
> -Anand

Neil