| Message ID | 20240620211310.820579-3-sean.anderson@linux.dev (mailing list archive) |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | hwmon: iio: Add labels | expand |
On 6/20/24 14:13, Sean Anderson wrote: > Add labels from IIO channels to our channels. This allows userspace to > display more meaningful names instead of "in0" or "temp5". > > Signed-off-by: Sean Anderson <sean.anderson@linux.dev> > --- > > drivers/hwmon/iio_hwmon.c | 33 ++++++++++++++++++++++++++++++--- > 1 file changed, 30 insertions(+), 3 deletions(-) > > diff --git a/drivers/hwmon/iio_hwmon.c b/drivers/hwmon/iio_hwmon.c > index 4c8a80847891..588b64c18e63 100644 > --- a/drivers/hwmon/iio_hwmon.c > +++ b/drivers/hwmon/iio_hwmon.c > @@ -33,6 +33,17 @@ struct iio_hwmon_state { > struct attribute **attrs; > }; > > +static ssize_t iio_hwmon_read_label(struct device *dev, > + struct device_attribute *attr, > + char *buf) > +{ > + struct sensor_device_attribute *sattr = to_sensor_dev_attr(attr); > + struct iio_hwmon_state *state = dev_get_drvdata(dev); > + struct iio_channel *chan = &state->channels[sattr->index]; > + > + return iio_read_channel_label(chan, buf); This can return -EINVAL if there is no label. Since the label attribute is created unconditionally, every affected system would end up with lots of error messages when running the "sensors" command. This is not acceptable. Guenter > +} > + > /* > * Assumes that IIO and hwmon operate in the same base units. > * This is supposed to be true, but needs verification for > @@ -90,12 +101,12 @@ static int iio_hwmon_probe(struct platform_device *pdev) > > st->channels = channels; > > - /* count how many attributes we have */ > + /* count how many channels we have */ > while (st->channels[st->num_channels].indio_dev) > st->num_channels++; > > st->attrs = devm_kcalloc(dev, > - st->num_channels + 1, sizeof(*st->attrs), > + 2 * st->num_channels + 1, sizeof(*st->attrs), > GFP_KERNEL); > if (st->attrs == NULL) > return -ENOMEM; > @@ -147,7 +158,23 @@ static int iio_hwmon_probe(struct platform_device *pdev) > a->dev_attr.show = iio_hwmon_read_val; > a->dev_attr.attr.mode = 0444; > a->index = i; > - st->attrs[i] = &a->dev_attr.attr; > + st->attrs[2 * i] = &a->dev_attr.attr; > + > + a = devm_kzalloc(dev, sizeof(*a), GFP_KERNEL); > + if (a == NULL) > + return -ENOMEM; > + > + sysfs_attr_init(&a->dev_attr.attr); > + a->dev_attr.attr.name = devm_kasprintf(dev, GFP_KERNEL, > + "%s%d_label", > + prefix, n); > + if (!a->dev_attr.attr.name) > + return -ENOMEM; > + > + a->dev_attr.show = iio_hwmon_read_label; > + a->dev_attr.attr.mode = 0444; > + a->index = i; > + st->attrs[2 * i + 1] = &a->dev_attr.attr; > } > > st->attr_group.attrs = st->attrs;
On 6/21/24 11:08, Guenter Roeck wrote: > On 6/20/24 14:13, Sean Anderson wrote: >> Add labels from IIO channels to our channels. This allows userspace to >> display more meaningful names instead of "in0" or "temp5". >> >> Signed-off-by: Sean Anderson <sean.anderson@linux.dev> >> --- >> >> drivers/hwmon/iio_hwmon.c | 33 ++++++++++++++++++++++++++++++--- >> 1 file changed, 30 insertions(+), 3 deletions(-) >> >> diff --git a/drivers/hwmon/iio_hwmon.c b/drivers/hwmon/iio_hwmon.c >> index 4c8a80847891..588b64c18e63 100644 >> --- a/drivers/hwmon/iio_hwmon.c >> +++ b/drivers/hwmon/iio_hwmon.c >> @@ -33,6 +33,17 @@ struct iio_hwmon_state { >> struct attribute **attrs; >> }; >> +static ssize_t iio_hwmon_read_label(struct device *dev, >> + struct device_attribute *attr, >> + char *buf) >> +{ >> + struct sensor_device_attribute *sattr = to_sensor_dev_attr(attr); >> + struct iio_hwmon_state *state = dev_get_drvdata(dev); >> + struct iio_channel *chan = &state->channels[sattr->index]; >> + >> + return iio_read_channel_label(chan, buf); > > This can return -EINVAL if there is no label. Since the label attribute > is created unconditionally, every affected system would end up with > lots of error messages when running the "sensors" command. > This is not acceptable. The sensors command gracefully handles this. There are no errors, and the label is unused. --Sean
On 6/21/24 11:22, Sean Anderson wrote: > On 6/21/24 11:08, Guenter Roeck wrote: >> On 6/20/24 14:13, Sean Anderson wrote: >>> Add labels from IIO channels to our channels. This allows userspace to >>> display more meaningful names instead of "in0" or "temp5". >>> >>> Signed-off-by: Sean Anderson <sean.anderson@linux.dev> >>> --- >>> >>> drivers/hwmon/iio_hwmon.c | 33 ++++++++++++++++++++++++++++++--- >>> 1 file changed, 30 insertions(+), 3 deletions(-) >>> >>> diff --git a/drivers/hwmon/iio_hwmon.c b/drivers/hwmon/iio_hwmon.c >>> index 4c8a80847891..588b64c18e63 100644 >>> --- a/drivers/hwmon/iio_hwmon.c >>> +++ b/drivers/hwmon/iio_hwmon.c >>> @@ -33,6 +33,17 @@ struct iio_hwmon_state { >>> struct attribute **attrs; >>> }; >>> +static ssize_t iio_hwmon_read_label(struct device *dev, >>> + struct device_attribute *attr, >>> + char *buf) >>> +{ >>> + struct sensor_device_attribute *sattr = to_sensor_dev_attr(attr); >>> + struct iio_hwmon_state *state = dev_get_drvdata(dev); >>> + struct iio_channel *chan = &state->channels[sattr->index]; >>> + >>> + return iio_read_channel_label(chan, buf); >> >> This can return -EINVAL if there is no label. Since the label attribute >> is created unconditionally, every affected system would end up with >> lots of error messages when running the "sensors" command. >> This is not acceptable. > > The sensors command gracefully handles this. There are no errors, and the label is unused. For example, without IIO labels I get: $ sensors hwmon_ams_ps-isa-0000 hwmon_ams_ps-isa-0000 Adapter: ISA adapter in1: 825.00 mV in2: 826.00 mV in3: 1.81 V in4: 1.18 V in5: 1.80 V in6: 1.80 V in7: 3.27 V in8: 1.81 V in9: 825.00 mV in10: 1.81 V in11: 1.80 V temp1: +79.8 C temp2: +80.9 C and with labels I get $ sensors hwmon_ams_ps-isa-0000 hwmon_ams_ps-isa-0000 Adapter: ISA adapter VCC_PSINTLP: 824.00 mV VCC_PSINTFP: 822.00 mV VCC_PSAUX: 1.81 V VCC_PSDDR: 1.18 V VCC_PSIO3: 1.80 V VCC_PSIO0: 1.80 V VCC_PSIO1: 3.27 V VCC_PSIO2: 1.80 V PS_MGTRAVCC: 822.00 mV PS_MGTRAVTT: 1.81 V VCC_PSADC: 1.80 V Temp_LPD: +79.5 C Temp_FPD: +80.2 C I also was concerned about the same thing, but lm-sensors handles things gracefully, allowing us to simplify the implementation. --Sean
On 6/21/24 08:31, Sean Anderson wrote: > On 6/21/24 11:22, Sean Anderson wrote: >> On 6/21/24 11:08, Guenter Roeck wrote: >>> On 6/20/24 14:13, Sean Anderson wrote: >>>> Add labels from IIO channels to our channels. This allows userspace to >>>> display more meaningful names instead of "in0" or "temp5". >>>> >>>> Signed-off-by: Sean Anderson <sean.anderson@linux.dev> >>>> --- >>>> >>>> drivers/hwmon/iio_hwmon.c | 33 ++++++++++++++++++++++++++++++--- >>>> 1 file changed, 30 insertions(+), 3 deletions(-) >>>> >>>> diff --git a/drivers/hwmon/iio_hwmon.c b/drivers/hwmon/iio_hwmon.c >>>> index 4c8a80847891..588b64c18e63 100644 >>>> --- a/drivers/hwmon/iio_hwmon.c >>>> +++ b/drivers/hwmon/iio_hwmon.c >>>> @@ -33,6 +33,17 @@ struct iio_hwmon_state { >>>> struct attribute **attrs; >>>> }; >>>> +static ssize_t iio_hwmon_read_label(struct device *dev, >>>> + struct device_attribute *attr, >>>> + char *buf) >>>> +{ >>>> + struct sensor_device_attribute *sattr = to_sensor_dev_attr(attr); >>>> + struct iio_hwmon_state *state = dev_get_drvdata(dev); >>>> + struct iio_channel *chan = &state->channels[sattr->index]; >>>> + >>>> + return iio_read_channel_label(chan, buf); >>> >>> This can return -EINVAL if there is no label. Since the label attribute >>> is created unconditionally, every affected system would end up with >>> lots of error messages when running the "sensors" command. >>> This is not acceptable. >> >> The sensors command gracefully handles this. There are no errors, and the label is unused. > > For example, without IIO labels I get: > > $ sensors hwmon_ams_ps-isa-0000 > hwmon_ams_ps-isa-0000 > Adapter: ISA adapter > in1: 825.00 mV > in2: 826.00 mV > in3: 1.81 V > in4: 1.18 V > in5: 1.80 V > in6: 1.80 V > in7: 3.27 V > in8: 1.81 V > in9: 825.00 mV > in10: 1.81 V > in11: 1.80 V > temp1: +79.8 C > temp2: +80.9 C > > and with labels I get > > $ sensors hwmon_ams_ps-isa-0000 > hwmon_ams_ps-isa-0000 > Adapter: ISA adapter > VCC_PSINTLP: 824.00 mV > VCC_PSINTFP: 822.00 mV > VCC_PSAUX: 1.81 V > VCC_PSDDR: 1.18 V > VCC_PSIO3: 1.80 V > VCC_PSIO0: 1.80 V > VCC_PSIO1: 3.27 V > VCC_PSIO2: 1.80 V > PS_MGTRAVCC: 822.00 mV > PS_MGTRAVTT: 1.81 V > VCC_PSADC: 1.80 V > Temp_LPD: +79.5 C > Temp_FPD: +80.2 C > You are correct, the sensors code ignores access errors on label attributes. Still, I am not even going there. If there is no label, don't create the attribute, period. The ABI says Should only be created if the driver has hints about what this voltage channel is being used for, and user-space doesn't. In all other cases, the label is provided by user-space. It should be obvious that this doesn't mean that you can create random attributes and then just return an error if it isn't actually supported. Guenter
diff --git a/drivers/hwmon/iio_hwmon.c b/drivers/hwmon/iio_hwmon.c index 4c8a80847891..588b64c18e63 100644 --- a/drivers/hwmon/iio_hwmon.c +++ b/drivers/hwmon/iio_hwmon.c @@ -33,6 +33,17 @@ struct iio_hwmon_state { struct attribute **attrs; }; +static ssize_t iio_hwmon_read_label(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct sensor_device_attribute *sattr = to_sensor_dev_attr(attr); + struct iio_hwmon_state *state = dev_get_drvdata(dev); + struct iio_channel *chan = &state->channels[sattr->index]; + + return iio_read_channel_label(chan, buf); +} + /* * Assumes that IIO and hwmon operate in the same base units. * This is supposed to be true, but needs verification for @@ -90,12 +101,12 @@ static int iio_hwmon_probe(struct platform_device *pdev) st->channels = channels; - /* count how many attributes we have */ + /* count how many channels we have */ while (st->channels[st->num_channels].indio_dev) st->num_channels++; st->attrs = devm_kcalloc(dev, - st->num_channels + 1, sizeof(*st->attrs), + 2 * st->num_channels + 1, sizeof(*st->attrs), GFP_KERNEL); if (st->attrs == NULL) return -ENOMEM; @@ -147,7 +158,23 @@ static int iio_hwmon_probe(struct platform_device *pdev) a->dev_attr.show = iio_hwmon_read_val; a->dev_attr.attr.mode = 0444; a->index = i; - st->attrs[i] = &a->dev_attr.attr; + st->attrs[2 * i] = &a->dev_attr.attr; + + a = devm_kzalloc(dev, sizeof(*a), GFP_KERNEL); + if (a == NULL) + return -ENOMEM; + + sysfs_attr_init(&a->dev_attr.attr); + a->dev_attr.attr.name = devm_kasprintf(dev, GFP_KERNEL, + "%s%d_label", + prefix, n); + if (!a->dev_attr.attr.name) + return -ENOMEM; + + a->dev_attr.show = iio_hwmon_read_label; + a->dev_attr.attr.mode = 0444; + a->index = i; + st->attrs[2 * i + 1] = &a->dev_attr.attr; } st->attr_group.attrs = st->attrs;
Add labels from IIO channels to our channels. This allows userspace to display more meaningful names instead of "in0" or "temp5". Signed-off-by: Sean Anderson <sean.anderson@linux.dev> --- drivers/hwmon/iio_hwmon.c | 33 ++++++++++++++++++++++++++++++--- 1 file changed, 30 insertions(+), 3 deletions(-)