linux-kernel - Re: [PATCH v8 2/2] Add PWM fan controller driver for LGM SoC

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] [thread-next>] [day] [month] [year] [list]
Date:   Fri, 21 Aug 2020 14:07:49 +0800
From:   "Tanwar, Rahul" <rahul.tanwar@...ux.intel.com>
To:     Andy Shevchenko <andriy.shevchenko@...el.com>
Cc:     u.kleine-koenig@...gutronix.de, linux-pwm@...r.kernel.org,
        lee.jones@...aro.org, thierry.reding@...il.com,
        p.zabel@...gutronix.de, robh+dt@...nel.org,
        linux-kernel@...r.kernel.org, devicetree@...r.kernel.org,
        songjun.Wu@...el.com, cheol.yong.kim@...el.com,
        qi-ming.wu@...el.com, rahul.tanwar.linux@...il.com,
        rtanwar@...linear.com
Subject: Re: [PATCH v8 2/2] Add PWM fan controller driver for LGM SoC


Hi Andy,

On 20/8/2020 6:52 pm, Andy Shevchenko wrote:
> On Thu, Aug 20, 2020 at 12:50:46PM +0800, Rahul Tanwar wrote:
>> Intel Lightning Mountain(LGM) SoC contains a PWM fan controller.
>> This PWM controller does not have any other consumer, it is a
>> dedicated PWM controller for fan attached to the system. Add
>> driver for this PWM fan controller.
> ...
>
>> +config PWM_INTEL_LGM
>> +	tristate "Intel LGM PWM support"
>> +	depends on OF && HAS_IOMEM
>> +	depends on X86 || COMPILE_TEST
> For better test coverage you may rewrite this
>
> 	depends on HAS_IOMEM
> 	depends on (OF && X86) || COMPILE_TEST

Sure, will update.

>> +	select REGMAP_MMIO
>> +	help
>> +	  Generic PWM fan controller driver for LGM SoC.
>> +
>> +	  To compile this driver as a module, choose M here: the module
>> +	  will be called pwm-intel-lgm.
> ...
>
>> +#include <linux/bitfield.h>
>> +#include <linux/clk.h>
>> +#include <linux/module.h>
>> +#include <linux/of_device.h>
> This should be mod_devicetable.h.

Well noted.

>> +#include <linux/pwm.h>
>> +#include <linux/regmap.h>
>> +#include <linux/reset.h>
> ...
>
>> +#define LGM_PWM_PERIOD_2WIRE_NSECS	40000000
> NSECS -> NS
> 40000000 -> (40 * NSEC_PER_MSEC)

Well noted.

> ...
>
>> +	if (state->polarity != PWM_POLARITY_NORMAL ||
>> +	    state->period < pc->period)
> It can be one line.

Ok.

>> +		return -EINVAL;
> ...
>
>> +	if (!state->enabled) {
>> +		ret = lgm_pwm_enable(chip, 0);
>> +		return ret;
> What is the point?

I guess you mean to change it to return lgm_pwm_enable(chip, 0);
Will do, thanks.

>> +	}
> ...
>
>> +	ret = lgm_pwm_enable(chip, 1);
>> +
>> +	return ret;
> Ditto.
>
> ...
>
>> +	state->duty_cycle = DIV_ROUND_UP(duty * pc->period,
>> +					 LGM_PWM_MAX_DUTY_CYCLE);
> One line?

Ok.

> ...
>
>> +	struct lgm_pwm_chip *pc;
>> +	struct device *dev = &pdev->dev;
> Use reversed xmas tree order.

Sure, will update.

>> +	void __iomem *io_base;
>> +	int ret;
> ...
>
>> +	pc->regmap = devm_regmap_init_mmio(dev, io_base, &lgm_pwm_regmap_config);
>> +	if (IS_ERR(pc->regmap)) {
>> +		ret = PTR_ERR(pc->regmap);
>> +		if (ret != -EPROBE_DEFER)
>> +			dev_err(dev, "failed to init register map: %pe\n",
>> +				pc->regmap);
>> +		return ret;
> dev_err_probe()

Will update. Thanks.

>> +	}
> ...
>
>> +	pc->clk = devm_clk_get(dev, NULL);
>> +	if (IS_ERR(pc->clk)) {
>> +		ret = PTR_ERR(pc->clk);
>> +		if (ret != -EPROBE_DEFER)
>> +			dev_err(dev, "failed to get clock: %pe\n", pc->clk);
>> +		return ret;
> Ditto.
>
>> +	}
>> +
>> +	pc->rst = devm_reset_control_get_exclusive(dev, NULL);
>> +	if (IS_ERR(pc->rst)) {
>> +		ret = PTR_ERR(pc->rst);
>> +		if (ret != -EPROBE_DEFER)
>> +			dev_err(dev, "failed to get reset control: %pe\n",
>> +				pc->rst);
>> +		return ret;
> Ditto.
>
>> +	}
>> +
>> +	ret = reset_control_deassert(pc->rst);
>> +	if (ret) {
>> +		if (ret != -EPROBE_DEFER)
>> +			dev_err(dev, "cannot deassert reset control: %pe\n",
>> +				ERR_PTR(ret));
>> +		return ret;
> Ditto.
>
>> +	}
> ...
>
>> +	ret = clk_prepare_enable(pc->clk);
> Wrap it with devm_add_action_or_reset(). Same for reset_control_deassert().
> You probably can even put them under one function.

I did some study and research for using devm_add_action_or_reset(). But
still i have some doubts. Below steps is what i intend to do in order to
switch to using this API. Please do review and let me know it is ok and
i am not missing anything else. Thanks.

1. Call reset_control_assert()
2. Call clk_prepare_enable()
3. Call pwmchip_add()
4. Call devm_add_action_or_reset(dev, my_action, pc)
5. Remove explicit calls to unprepare/reset_control_assert from probe in
failure cases.
6. static void my_action(void *pc)
   {
      pwmchip_remove(&pc->chip);
      clk_disable_upprepare(pc->clk);
      reset_control_assert(pc->rst);
   }
7. Remove platform_driver.remove() entirely.

>> +	if (ret) {
>> +		dev_err(dev, "failed to enable clock\n");
>> +		reset_control_assert(pc->rst);
>> +		return ret;
>> +	}
> ...
>
>> +	ret = pwmchip_add(&pc->chip);
>> +	if (ret < 0) {
> Does ' < 0' have any meaning?

I use < 0 because this API's return code is mentioned as below:
Returns: 0 on success or a negative error code on failure.
Also, all other PWM drivers check for <0 for this call.

>> +		dev_err(dev, "failed to add PWM chip: %pe\n", ERR_PTR(ret));
>> +		clk_disable_unprepare(pc->clk);
>> +		reset_control_assert(pc->rst);
>> +		return ret;
>> +	}
> ...
>
>> +	ret = pwmchip_remove(&pc->chip);
>> +	if (ret < 0)
> Ditto.

Same as above.

>> +		return ret;
Thanks.

Regards,
Rahul