linux-kernel - Re: [PATCH] regulator:s5m8767 Fully convert to GPIO descriptors

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <4fc82b83-d312-4878-96ea-4c6ce7e19631@189.cn>
Date: Mon, 6 Jan 2025 17:23:55 +0800
From: Song Chen <chensong_2000@....cn>
To: Krzysztof Kozlowski <krzk@...nel.org>, lgirdwood@...il.com,
 broonie@...nel.org, lee@...nel.org, linus.walleij@...aro.org, brgl@...ev.pl
Cc: linux-kernel@...r.kernel.org, linux-gpio@...r.kernel.org,
 linux-samsung-soc@...r.kernel.org
Subject: Re: [PATCH] regulator:s5m8767 Fully convert to GPIO descriptors

Hi Krzysztof,

Bart and I are talking about using gpiod array in s5m8767 by adding a 
new variant in gpio core, what do you think about that?

what's more, i also have another question about 
s5m8767_set_high/s5m8767_set_low, see detail below, i would appreciate 
it if you could give any comments about that.

Best regards,

Song


在 2024/12/13 18:30, Song Chen 写道:
> Hi Krzysztof,
> 
> I noticed that in s5m8767_set_high and s5m8767_set_low, the code looks 
> identical to each other, only order is different. Is there any problem 
> here or this way is on purpose correctly.
> 
> static inline int s5m8767_set_high(struct s5m8767_info *s5m8767)
> {
>      int temp_index = s5m8767->buck_gpioindex;
> 
>      gpio_set_value(s5m8767->buck_gpios[0], (temp_index >> 2) & 0x1);
>      gpio_set_value(s5m8767->buck_gpios[1], (temp_index >> 1) & 0x1);
>      gpio_set_value(s5m8767->buck_gpios[2], temp_index & 0x1);
> 
>      return 0;
> }
> 
> static inline int s5m8767_set_low(struct s5m8767_info *s5m8767)
> {
>      int temp_index = s5m8767->buck_gpioindex;
> 
>      gpio_set_value(s5m8767->buck_gpios[2], temp_index & 0x1);
>      gpio_set_value(s5m8767->buck_gpios[1], (temp_index >> 1) & 0x1);
>      gpio_set_value(s5m8767->buck_gpios[0], (temp_index >> 2) & 0x1);
> 
>      return 0;
> }
> 
> Song
> 
> 
> 
>> On 07/12/2024 07:16, Song Chen wrote:
>>>>>            }
>>>>> -        pdata->buck_gpios[i] = gpio;
>>>>> +
>>>>> +        /* SET GPIO*/
>>>>
>>>> What is a SET GPIO?
>>>>
>>>>> +        snprintf(label, sizeof(label), "%s%d", "S5M8767 SET", i + 1);
>>>>
>>>> Why using "SET" as name, not the actual name it is used for? Buck DVS?
>>>
>>> from below snippets:
>>> s5m8767_pmic_probe of drivers/regulator/s5m8767.c
>>>           ret = devm_gpio_request(&pdev->dev, pdata->buck_gpios[0],
>>>                       "S5M8767 SET1");
>>>           if (ret)
>>>               return ret;
>>>
>>>           ret = devm_gpio_request(&pdev->dev, pdata->buck_gpios[1],
>>>                       "S5M8767 SET2");
>>>           if (ret)
>>>               return ret;
>>>
>>>           ret = devm_gpio_request(&pdev->dev, pdata->buck_gpios[2],
>>>                       "S5M8767 SET3");
>>
>>
>> Yeah, your code is fine.
>>
>>>
>>> and arch/arm/boot/dts/samsung/exynos5250-spring.dts
>>>
>>>           s5m8767,pmic-buck-dvs-gpios = <&gpd1 0 GPIO_ACTIVE_LOW>, /* 
>>> DVS1 */
>>>                             <&gpd1 1 GPIO_ACTIVE_LOW>, /* DVS2 */
>>>                             <&gpd1 2 GPIO_ACTIVE_LOW>; /* DVS3 */
>>>
>>>           s5m8767,pmic-buck-ds-gpios = <&gpx2 3 GPIO_ACTIVE_LOW>, /* 
>>> SET1 */
>>>                            <&gpx2 4 GPIO_ACTIVE_LOW>, /* SET2 */
>>>                            <&gpx2 5 GPIO_ACTIVE_LOW>; /* SET3 */
>>>
>>>>
>>>>> +        gpiod_set_consumer_name(pdata->buck_gpios[i], label);
>>>>> +        gpiod_direction_output(pdata->buck_gpios[i],
>>>>> +                    (pdata->buck_default_idx >> (2 - i)) & 0x1);
>>>>
>>>> This is not an equivalent code. You set values for GPIOs 0-1 even if
>>>> requesting GPIO 2 fails.
>>>>
>>>> On which board did you test it?
>>>
>>> You are right ,it's not equivalent with original code, i will fix it.
>>> but i have a question here:
>>>
>>>           ret = devm_gpio_request(&pdev->dev, pdata->buck_gpios[0],
>>>                       "S5M8767 SET1");
>>>           if (ret)
>>>               return ret;
>>>
>>>           ret = devm_gpio_request(&pdev->dev, pdata->buck_gpios[1],
>>>                       "S5M8767 SET2");
>>>           if (ret)
>>>               return ret;
>>>
>>>           ret = devm_gpio_request(&pdev->dev, pdata->buck_gpios[2],
>>>                       "S5M8767 SET3");
>>>           if (ret)
>>>               return ret;
>>>
>>> if it fails to request buck_gpios[2] after successfully requests
>>> buck_gpios[0] and buck_gpios[1], the probe fails as well, should it call
>>> gpiod_put to return gpio resource?
>>
>>
>> Aren't you using devm interface? Please read the API. You do not need to
>> put anything, unless you use some other interface and I missed the point
>> of the question.
>>
>> Best regards,
>> Krzysztof
>>
>>