Re: [PATCH v3 2/6] platform/x86: lenovo-wmi-other: Limit adding attributes to supported devices

[Rong Zhang <i@rong.moe>]

Hi Derek,

On Wed, 2026-02-25 at 10:09 -0800, Derek J. Clark wrote:
> On February 25, 2026 9:14:12 AM PST, Rong Zhang <i@rong.moe> wrote:
> > Hi Derek,
> > 
> > On Tue, 2026-02-24 at 04:31 +0000, Derek J. Clark wrote:
> > > Adds lwmi_is_attr_01_supported, and only creates the attribute subfolder
> > > if the attribute is supported by the hardware. Due to some poorly
> > > implemented BIOS, this is a multi-step sequence of events. This is
> > > because:
> > > - Some BIOS support getting the capability data from custom mode (0xff),
> > >   while others only support it in no-mode (0x00).
> > > - Similarly, some BIOS support get/set for the current value from custom
> > >   mode (0xff), while others only support it in no-mode (0x00).
> > > - Some BIOS report capability data for a method that is not fully
> > >   implemented.
> > > - Some BIOS have methods fully implemented, but no complimentary
> > >   capability data.
> > > 
> > > To ensure we only expose fully implemented methods with corresponding
> > > capability data, we check each outcome before reporting that an
> > > attribute can be supported.
> > > 
> > > Checking for lwmi_is_attr_01_supported during remove is not done to
> > > ensure that we don't attempt to call cd01 or send WMI events if one of
> > > the interfaces being removed was the cause of the driver unloading.
> > > 
> > > While adding members to tunable_attr_01, remove unused capdata pointer
> > > and limit size of all ID's to the appropriate size.
> > > 
> > > Reviewed-by: Mark Pearson <mpearson-lenovo@squebb.ca>
> > > Reported-by: Kurt Borja <kuurtb@gmail.com>
> > > Closes: https://lore.kernel.org/platform-driver-x86/DG60P3SHXR8H.3NSEHMZ6J7XRC@gmail.com/
> > > Signed-off-by: Derek J. Clark <derekjohn.clark@gmail.com>
> > > ---
> > >  drivers/platform/x86/lenovo/wmi-other.c | 117 +++++++++++++++++++++---
> > >  1 file changed, 102 insertions(+), 15 deletions(-)
> > > 
> > > diff --git a/drivers/platform/x86/lenovo/wmi-other.c b/drivers/platform/x86/lenovo/wmi-other.c
> > > index 95886df39c8d..f3f12303e379 100644
> > > --- a/drivers/platform/x86/lenovo/wmi-other.c
> > > +++ b/drivers/platform/x86/lenovo/wmi-other.c
> > > @@ -545,11 +545,12 @@ static void lwmi_om_fan_info_collect_cd_fan(struct device *dev, struct cd_list *
> > >  /* ======== fw_attributes (component: lenovo-wmi-capdata 01) ======== */
> > >  
> > >  struct tunable_attr_01 {
> > > -	struct capdata01 *capdata;
> > >  	struct device *dev;
> > > -	u32 feature_id;
> > > -	u32 device_id;
> > > -	u32 type_id;
> > > +	u8 feature_id;
> > > +	u8 device_id;
> > > +	u8 type_id;
> > > +	u8 cd_mode_id; /* mode arg for searching capdata */
> > > +	u8 cv_mode_id; /* mode arg for set/get current_value */
> > >  };
> > >  
> > >  static struct tunable_attr_01 ppt_pl1_spl = {
> > > @@ -716,7 +717,7 @@ static ssize_t attr_capdata01_show(struct kobject *kobj,
> > >  	int value, ret;
> > >  
> > >  	attribute_id = LWMI_ATTR_ID(tunable_attr->device_id, tunable_attr->feature_id,
> > > -				    LWMI_GZ_THERMAL_MODE_CUSTOM, tunable_attr->type_id);
> > > +				    tunable_attr->cd_mode_id, tunable_attr->type_id);
> > >  
> > >  	ret = lwmi_cd01_get_data(priv->cd01_list, attribute_id, &capdata);
> > >  	if (ret)
> > > @@ -782,7 +783,7 @@ static ssize_t attr_current_value_store(struct kobject *kobj,
> > >  		return -EBUSY;
> > >  
> > >  	args.arg0 = LWMI_ATTR_ID(tunable_attr->device_id, tunable_attr->feature_id,
> > > -				 mode, tunable_attr->type_id);
> > > +				 tunable_attr->cd_mode_id, tunable_attr->type_id);
> > >  
> > >  	ret = lwmi_cd01_get_data(priv->cd01_list, args.arg0, &capdata);
> > >  	if (ret)
> > > @@ -795,6 +796,8 @@ static ssize_t attr_current_value_store(struct kobject *kobj,
> > >  	if (value < capdata.min_value || value > capdata.max_value)
> > >  		return -EINVAL;
> > >  
> > > +	args.arg0 = LWMI_ATTR_ID(tunable_attr->device_id, tunable_attr->feature_id,
> > > +				 tunable_attr->cv_mode_id, tunable_attr->type_id);
> > >  	args.arg1 = value;
> > >  
> > >  	ret = lwmi_dev_evaluate_int(priv->wdev, 0x0, LWMI_FEATURE_VALUE_SET,
> > > @@ -828,13 +831,16 @@ static ssize_t attr_current_value_show(struct kobject *kobj,
> > >  	struct lwmi_om_priv *priv = dev_get_drvdata(tunable_attr->dev);
> > >  	struct wmi_method_args_32 args;
> > >  	enum thermal_mode mode;
> > > -	int retval;
> > > -	int ret;
> > > +	int retval, ret;
> > >  
> > >  	ret = lwmi_om_notifier_call(&mode);
> > >  	if (ret)
> > >  		return ret;
> > >  
> > > +	/* If "no-mode" is the supported mode, ensure we never send current mode */
> > > +	if (tunable_attr->cv_mode_id == LWMI_GZ_THERMAL_MODE_NONE)
> > > +		mode = tunable_attr->cv_mode_id;
> > > +
> > >  	args.arg0 = LWMI_ATTR_ID(tunable_attr->device_id, tunable_attr->feature_id,
> > >  				 mode, tunable_attr->type_id);
> > >  
> > > @@ -847,6 +853,85 @@ static ssize_t attr_current_value_show(struct kobject *kobj,
> > >  	return sysfs_emit(buf, "%d\n", retval);
> > >  }
> > >  
> > > +/**
> > > + * lwmi_attr_01_is_supported() - Determine if the given attribute is supported.
> > > + * @tunable_attr: The attribute to verify.
> > > + *
> > > + * First check if the attribute has a corresponding capdata01 table in the cd01
> > > + * module under the "custom" mode (0xff). If that is not present then check if
> > > + * there is a corresponding "no-mode" (0x00) entry. If either of those passes,
> > > + * check capdata->supported for values > 0. If capdata is available, attempt to
> > > + * determine the set/get mode for the current value property using a similar
> > > + * pattern. If the value returned by either custom or no-mode is 0, or we get
> > > + * an error, we assume that mode is not supported. If any of the above checks
> > > + * fail then the attribute is not fully supported.
> > > + *
> > > + * The probed cd_mode_id/cv_mode_id are stored on the tunable_attr for later
> > > + * reference.
> > > + *
> > > + * Return: Support level, or an error code.
> > > + */
> > > +static int lwmi_attr_01_is_supported(struct tunable_attr_01 *tunable_attr)
> > > +{
> > > +	struct lwmi_om_priv *priv = dev_get_drvdata(tunable_attr->dev);
> > > +	u8 mode = LWMI_GZ_THERMAL_MODE_CUSTOM;
> > > +	struct wmi_method_args_32 args;
> > > +	struct capdata01 capdata;
> > > +	int retval, ret;
> > > +
> > > +	/* Determine tunable_attr->cd_mode_id */
> > > +no_mode_fallback_1:
> > > +	args.arg0 = LWMI_ATTR_ID(tunable_attr->device_id, tunable_attr->feature_id,
> > > +				 mode, tunable_attr->type_id);
> > > +
> > > +	ret = lwmi_cd01_get_data(priv->cd01_list, args.arg0, &capdata);
> > > +	if (ret && mode) {
> > > +		dev_dbg(tunable_attr->dev, "Attribute id %x not supported\n", args.arg0);
> > > +		mode = LWMI_GZ_THERMAL_MODE_NONE;
> > > +		goto no_mode_fallback_1;
> > > +	}
> > > +	if (ret)
> > > +		goto not_supported;
> > > +	if (!capdata.supported) {
> > > +		ret = -EOPNOTSUPP;
> > > +		goto not_supported;
> > > +	}
> > > +
> > > +	tunable_attr->cd_mode_id = mode;
> > > +
> > > +	/* Determine tunable_attr->cv_mode_id */
> > > +	mode = LWMI_GZ_THERMAL_MODE_CUSTOM;
> > > +no_mode_fallback_2:
> > > +	args.arg0 = LWMI_ATTR_ID(tunable_attr->device_id, tunable_attr->feature_id,
> > > +				 mode, tunable_attr->type_id);
> > > +
> > > +	ret = lwmi_dev_evaluate_int(priv->wdev, 0x0, LWMI_FEATURE_VALUE_GET,
> > > +				    (unsigned char *)&args, sizeof(args),
> > > +				    &retval);
> > > +	if ((ret && mode) || (!retval && mode)) {
> > > +		dev_dbg(tunable_attr->dev, "Attribute id %x not supported\n", args.arg0);
> > > +		mode = LWMI_GZ_THERMAL_MODE_NONE;
> > > +		goto no_mode_fallback_2;
> > > +	}
> > > +	if (ret)
> > > +		goto not_supported;
> > > +	if (retval == 0) {
> > > +		ret = -EOPNOTSUPP;
> > > +		goto not_supported;
> > > +	}
> > > +
> > 
> > I agree with Ilpo that backward gotos are hard to read and should be
> > avoided if possible.
> > 
> > > +	tunable_attr->cv_mode_id = mode;
> > > +	dev_dbg(tunable_attr->dev, "cd_mode_id: %02x%02x%02x%02x, cv_mode_id: %#08x attribute support level: %x\n",
> > 
> > It's a long line. Wrap at the first comma?
> > 
> > > +		tunable_attr->device_id, tunable_attr->feature_id, tunable_attr->cd_mode_id,
> > > +		tunable_attr->type_id, args.arg0, capdata.supported);
> > > +
> > > +	return capdata.supported;
> > > +
> > > +not_supported:
> > > +	dev_dbg(tunable_attr->dev, "Attribute id %x not supported\n", args.arg0);
> > > +	return ret;
> > > +}
> > > +
> > >  /* Lenovo WMI Other Mode Attribute macros */
> > >  #define __LWMI_ATTR_RO(_func, _name)                                  \
> > >  	{                                                             \
> > > @@ -970,19 +1055,21 @@ static int lwmi_om_fw_attr_add(struct lwmi_om_priv *priv)
> > >  	}
> > >  
> > >  	for (i = 0; i < ARRAY_SIZE(cd01_attr_groups) - 1; i++) {
> > > -		err = sysfs_create_group(&priv->fw_attr_kset->kobj,
> > > -					 cd01_attr_groups[i].attr_group);
> > > -		if (err)
> > > -			goto err_remove_groups;
> > > -
> > >  		cd01_attr_groups[i].tunable_attr->dev = &priv->wdev->dev;
> > > +		if (lwmi_attr_01_is_supported(cd01_attr_groups[i].tunable_attr) > 0) {
> > > +			err = sysfs_create_group(&priv->fw_attr_kset->kobj,
> > > +						 cd01_attr_groups[i].attr_group);
> > > +			if (err)
> > > +				goto err_remove_groups;
> > > +		}
> > >  	}
> > >  	return 0;
> > >  
> > >  err_remove_groups:
> > >  	while (i--)
> > > -		sysfs_remove_group(&priv->fw_attr_kset->kobj,
> > > -				   cd01_attr_groups[i].attr_group);
> > > +		if (lwmi_attr_01_is_supported(cd01_attr_groups[i].tunable_attr) > 0)
> > 
> > I don't see lwmi_om_fw_attr_remove() doing the same. If we care about
> > symmetric cleanup, I would imagine a bitmap or so recording what have
> > been created is used there.
> > 
> > Also it's doubtful if the firmware always return the same result on the
> > same parameters. If a broken firmware somehow returns different results
> > here, some attributes may be left uncleaned.
> > 
> > That being said, I am not sure if we really need symmetric cleanup.
> > IIUC, sysfs_remove_group() is a no-op if we pass an attr_group that has
> > not been created.
> 
> Hi Rong,
> We don't want to query the interface if it was removed, and the no-op is harmless. It doesn't seem worth it to me to retain that in memory to achieve the same results.

Makes sense. Then I would suggest drop the lwmi_attr_01_is_supported()
check here so that we don't need to worry about broken firmware.

Thanks,
Rong

> > > +			sysfs_remove_group(&priv->fw_attr_kset->kobj,
> > > +					   cd01_attr_groups[i].attr_group);
> > >  
> > >  	kset_unregister(priv->fw_attr_kset);
> > >  
> > 
> > Thanks,
> > Rong