RE: [PATCH] hwmon: max31790: support to config PWM as TACH

[Delphine_CC_Chiu/WYHQ/Wiwynn]

> -----Original Message-----
> From: Guenter Roeck <groeck7@xxxxxxxxx> On Behalf Of Guenter Roeck
> Sent: Thursday, September 7, 2023 12:20 AM
> To: Delphine_CC_Chiu/WYHQ/Wiwynn <Delphine_CC_Chiu@xxxxxxxxxx>;
> patrick@xxxxxxxxx; Jean Delvare <jdelvare@xxxxxxxx>
> Cc: linux-hwmon@xxxxxxxxxxxxxxx; linux-kernel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH] hwmon: max31790: support to config PWM as TACH
> 
>   Security Reminder: Please be aware that this email is sent by an external
> sender.
> 
> On 9/6/23 01:48, Delphine CC Chiu wrote:
> > The PWM outputs of max31790 could be used as tachometer inputs by
> > setting the fan configuration register, but the driver doesn't support
> > to config the PWM outputs as tachometer inputs currently.
> >
> > Add a function to get properties of the setting of max31790 to config
> > PWM outputs as tachometer inputs before initializing max31790.
> > For example: set `pwm-as-tach = /bits/ 8 <2 5>` in DTS for max31790
> > and the driver will config PWMOUT2 and PWMOUT5 as TACH8 and
> TACH11.
> >
> 
> Devicetree properties have to be documented in a property file and have to
> be approved by a devicetree maintainer.
> 
> Personally I don't think this is the proper way of configuring this, but I'll let
> devicetree maintainers decide.
> 
Thanks for your reply.
We will add document of max31790 in v2 patch.

> > Signed-off-by: Delphine CC Chiu <Delphine_CC_Chiu@xxxxxxxxxx>
> > ---
> >   drivers/hwmon/max31790.c | 50
> ++++++++++++++++++++++++++++++++++++++++
> >   1 file changed, 50 insertions(+)
> >
> > diff --git a/drivers/hwmon/max31790.c b/drivers/hwmon/max31790.c
> index
> > 0cd44c1e998a..0f8fe911539b 100644
> > --- a/drivers/hwmon/max31790.c
> > +++ b/drivers/hwmon/max31790.c
> > @@ -480,6 +480,52 @@ static const struct hwmon_chip_info
> max31790_chip_info = {
> >       .info = max31790_info,
> >   };
> >
> > +static int max31790_config_pwm_as_tach(struct device *dev,
> > +                                    struct i2c_client *client) {
> > +     struct device_node *np = dev->of_node;
> > +     int i, ret = 0, size, channel;
> > +     u8 pwm_index[NR_CHANNEL] = { 0 };
> > +     u8 fan_config;
> > +
> > +     size = of_property_count_u8_elems(np, "pwm-as-tach");
> > +
> > +     if ((size > 0) && (size <= NR_CHANNEL)) {
> 
> Please refrain from unnecessary ( ).
> 
Will remove in the v2 patch.

> > +             ret = of_property_read_u8_array(np, "pwm-as-tach",
> pwm_index,
> > +                                             size);
> > +             if (ret) {
> > +                     dev_err(dev,
> > +                             "Property 'pwm-as-tach' cannot be
> read.\n");
> > +                     return ret;
> > +             }
> > +
> > +             for (i = 0; i < size; i++) {
> > +                     if ((pwm_index[i] == 0) ||
> > +                         (pwm_index[i] > NR_CHANNEL)) {
> > +                             continue;
> > +                     }
> 
> Silently accepting bad data seems like a bad idea to me.
> 
Add error handling in v2 patch.

> > +
> > +                     channel = pwm_index[i] - 1;
> > +                     fan_config = i2c_smbus_read_byte_data(
> > +                             client,
> MAX31790_REG_FAN_CONFIG(channel));
> > +                     if (fan_config < 0) {
> 
> An u8 is never < 0
> 
Use integer to get the return value first and set to fan_config in v2 patch.

> > +                             dev_err(dev,
> > +                                     "Read fan config for channel
> %d failed.\n",
> > +                                     channel);
> > +                             return fan_config;
> > +                     }
> > +
> > +                     fan_config |=
> (MAX31790_FAN_CFG_CTRL_MON |
> > +
> MAX31790_FAN_CFG_TACH_INPUT);
> 
> This assumes that the channel is configured as pwm.
> What if the BIOS / ROMMON configured another channel which you want as
> pwm channel as fan input channel ?
> 
This will config the channel as TACH.
Could you provide more information about the scenario you mentioned?
In our system, there is only BMC that will set the config of fan device.
> > +                     i2c_smbus_write_byte_data(
> > +                             client,
> MAX31790_REG_FAN_CONFIG(channel),
> > +                             fan_config);
> > +             }
> > +     }
> 
> Silently ignoring errors seems like a bad idea.
> 
Add error handling in v2 patch.

> > +
> > +     return 0;
> > +}
> > +
> >   static int max31790_init_client(struct i2c_client *client,
> >                               struct max31790_data *data)
> >   {
> > @@ -521,6 +567,10 @@ static int max31790_probe(struct i2c_client
> *client)
> >       data->client = client;
> >       mutex_init(&data->update_lock);
> >
> > +     err = max31790_config_pwm_as_tach(dev, client);
> > +     if (err)
> > +             dev_crit(dev, "Config PWM as TACH failed.\n");
> > +
> >       /*
> >        * Initialize the max31790 chip
> >        */