Re: [PATCH v4 4/8] platform/x86: lenovo-wmi-other: Limit adding attributes to supported devices

[Rong Zhang <i@rong.moe>]

Hi Derek,

On Sat, 2026-03-14 at 20:47 -0700, Derek J. Clark wrote:
> On March 14, 2026 6:26:08 PM PDT, Rong Zhang <i@rong.moe> wrote:
> > Hi Derek,
> > 
> > On Thu, 2026-03-12 at 03:10 +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).
> > > - 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.
> > > 
> > > 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>
> > > ---
> > > v4:
> > >   - Use for loop instead of backtrace gotos for checking if an attribute
> > >     is supported.
> > >   - Add include for dev_printk.
> > >   - Wrap dev_dbg in lwmi_is_attr_01_supported earlier.
> > >   - Don't use symmetric cleanup of attributes in error states.
> > > ---
> > >  drivers/platform/x86/lenovo/wmi-other.c | 76 ++++++++++++++++++++++++-
> > >  1 file changed, 74 insertions(+), 2 deletions(-)
> > > 
> > > diff --git a/drivers/platform/x86/lenovo/wmi-other.c b/drivers/platform/x86/lenovo/wmi-other.c
> > > index 9fff9c1f768c..55a26e5617d4 100644
> > > --- a/drivers/platform/x86/lenovo/wmi-other.c
> > > +++ b/drivers/platform/x86/lenovo/wmi-other.c
> > > @@ -32,6 +32,7 @@
> > >  #include <linux/component.h>
> > >  #include <linux/container_of.h>
> > >  #include <linux/device.h>
> > > +#include <linux/dev_printk.h>
> > >  #include <linux/export.h>
> > >  #include <linux/gfp_types.h>
> > >  #include <linux/hwmon.h>
> > > @@ -871,6 +872,76 @@ 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)
> > > +{
> > > +	u8 modes[2] = { LWMI_GZ_THERMAL_MODE_CUSTOM, LWMI_GZ_THERMAL_MODE_NONE };
> > > +	struct lwmi_om_priv *priv = dev_get_drvdata(tunable_attr->dev);
> > > +	struct wmi_method_args_32 args;
> > > +	bool cd_mode_found = false;
> > > +	bool cv_mode_found = false;
> > > +	struct capdata01 capdata;
> > > +	int retval, ret, i;
> > > +
> > > +	/* Determine tunable_attr->cd_mode_id*/
> > > +	for (i = 0; i < ARRAY_SIZE(modes); i++) {
> > 
> > #include <linux/array_size.h>
> > 
> > > +		args.arg0 = lwmi_attr_id(tunable_attr->device_id, tunable_attr->feature_id,
> > > +					 modes[i], tunable_attr->type_id);
> > > +
> > > +		ret = lwmi_cd01_get_data(priv->cd01_list, args.arg0, &capdata);
> > > +		if (ret || !capdata.supported)
> > > +			continue;
> > > +		tunable_attr->cd_mode_id = modes[i];
> > > +		cd_mode_found = true;
> > > +		break;
> > > +	}
> > > +
> > > +	if (!cd_mode_found)
> > > +		return -EOPNOTSUPP;
> > > +
> > > +	/* Determine tunable_attr->cv_mode_id, returns 1 if supported*/
> > > +	for (i = 0; i < ARRAY_SIZE(modes); i++) {
> > > +		args.arg0 = lwmi_attr_id(tunable_attr->device_id, tunable_attr->feature_id,
> > > +					 modes[i], tunable_attr->type_id);
> > > +
> > > +		ret = lwmi_dev_evaluate_int(priv->wdev, 0x0, LWMI_FEATURE_VALUE_GET,
> > > +					    (unsigned char *)&args, sizeof(args),
> > > +					    &retval);
> > > +		if (ret || !retval)
> > > +			continue;
> > > +		tunable_attr->cv_mode_id = modes[i];
> > > +		cv_mode_found = true;
> > > +		break;
> > > +	}
> > > +
> > > +	if (!cv_mode_found)
> > > +		return -EOPNOTSUPP;
> > > +
> > > +	dev_dbg(tunable_attr->dev,
> > > +		"cd_mode_id: %02x%02x%02x%02x, cv_mode_id: %#08x attribute support level: %x\n",
> >                             ^                               ^  ^                          ^
> > > +		tunable_attr->device_id, tunable_attr->feature_id, tunable_attr->cd_mode_id,
> > > +		tunable_attr->type_id, args.arg0, capdata.supported);
> >                ^
> > 
> > 	dev_dbg(tunable_attr->dev,
> > 		"cd_mode_id: %#10x, cv_mode_id: %#10x, attribute support level: %#10x\n",
> > 		lwmi_attr_id(...), args.arg0, capdata.supported);
> > 
> 
> I hadn't thought about using that here, good idea.
> 
> > > +
> > > +	return capdata.supported;
> > 
> > You are casting u32 to int. Return it in a pointer argument if you need
> > it.
> 
> We don't do anything with the specific value, we just check that the level is above 0. We also don't do anything with the error codes specifically. Perhaps this should return a bool instead and we can return true if this is above 0 and false in all other paths? That would also improve the syntax a bit since the function name is a question and that would allow for !.._is_supported checks.

Yeah, that's a good idea.

Thanks,
Rong

> 
> Thanks,
> Derek
> 
> > > +}
> > > +
> > >  /* Lenovo WMI Other Mode Attribute macros */
> > >  #define __LWMI_ATTR_RO(_func, _name)                                  \
> > >  	{                                                             \
> > > @@ -994,12 +1065,13 @@ static int lwmi_om_fw_attr_add(struct lwmi_om_priv *priv)
> > >  	}
> > >  
> > >  	for (i = 0; i < ARRAY_SIZE(cd01_attr_groups) - 1; i++) {
> > > +		cd01_attr_groups[i].tunable_attr->dev = &priv->wdev->dev;
> > > +		if (lwmi_attr_01_is_supported(cd01_attr_groups[i].tunable_attr)  <= 0)
> > 
> > Extra whitespace.
> > 
> > > +			continue;
> > 
> > Add an empty line in between.
> > 
> > Thanks,
> > Rong
> > 
> > >  		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;
> > >  	}
> > >  	return 0;
> > >