Hi,
This patch adds support for the PWM frequencies from 22.5 to 30 kHz
available on the LM96000.
It looks like this chip has been supported for a long time, but wasn't
mentioned in the docs (which I have updated).
Also, it has been using the generic 'lm85' prefix, which I have not
changed to avoid breaking userspace or device tree files. AFAICT,
lm85_detect() will only return 'lm85' for LM96000 chips, so it doesn't
look like you can get a bare 'lm85' prefix with any other chips.
I stumbled on to a 10 year old thread discussing a patch which looks
like an early attempt to add support for this chip, which may be of
interest:
https://lm-sensors.lm-sensors.narkive.com/1SIwaMDT/patch-hwmon-lm96000-support
Thanks for reviewing,
Jeremy
Jeremy Gebben (3):
hwmon: (lm85) remove freq_map size hardcodes
hwmon: (lm85) Document the LM96000 as supported
hwmon: (lm85) add support for LM96000 high frequencies
Documentation/hwmon/lm85 | 9 ++++++++-
drivers/hwmon/lm85.c | 32 ++++++++++++++++++++++----------
2 files changed, 30 insertions(+), 11 deletions(-)
--
2.17.1
On Tue, Jan 29, 2019 at 02:29:17PM -0700, Jeremy Gebben wrote:
> Hi,
>
> This patch adds support for the PWM frequencies from 22.5 to 30 kHz
> available on the LM96000.
>
> It looks like this chip has been supported for a long time, but wasn't
> mentioned in the docs (which I have updated).
>
> Also, it has been using the generic 'lm85' prefix, which I have not
> changed to avoid breaking userspace or device tree files. AFAICT,
> lm85_detect() will only return 'lm85' for LM96000 chips, so it doesn't
> look like you can get a bare 'lm85' prefix with any other chips.
>
> I stumbled on to a 10 year old thread discussing a patch which looks
> like an early attempt to add support for this chip, which may be of
> interest:
> https://lm-sensors.lm-sensors.narkive.com/1SIwaMDT/patch-hwmon-lm96000-support
>
> Thanks for reviewing,
>
Where are the actual patches ?
Guenter
Allow support for chips that support more than 8 frequencies.
The size still must be a power of two.
Signed-off-by: Jeremy Gebben <[email protected]>
---
drivers/hwmon/lm85.c | 25 +++++++++++++++----------
1 file changed, 15 insertions(+), 10 deletions(-)
diff --git a/drivers/hwmon/lm85.c b/drivers/hwmon/lm85.c
index 0a325878e8f5..48f59667a88c 100644
--- a/drivers/hwmon/lm85.c
+++ b/drivers/hwmon/lm85.c
@@ -198,13 +198,13 @@ static int RANGE_TO_REG(long range)
#define RANGE_FROM_REG(val) lm85_range_map[(val) & 0x0f]
/* These are the PWM frequency encodings */
-static const int lm85_freq_map[8] = { /* 1 Hz */
+static const int lm85_freq_map[] = { /* 1 Hz */
10, 15, 23, 30, 38, 47, 61, 94
};
-static const int adm1027_freq_map[8] = { /* 1 Hz */
+
+static const int adm1027_freq_map[] = { /* 1 Hz */
11, 15, 22, 29, 35, 44, 59, 88
};
-#define FREQ_MAP_LEN 8
static int FREQ_TO_REG(const int *map,
unsigned int map_size, unsigned long freq)
@@ -212,9 +212,9 @@ static int FREQ_TO_REG(const int *map,
return find_closest(freq, map, map_size);
}
-static int FREQ_FROM_REG(const int *map, u8 reg)
+static int FREQ_FROM_REG(const int *map, unsigned int map_size, u8 reg)
{
- return map[reg & 0x07];
+ return map[reg & (map_size - 1)];
}
/*
@@ -296,6 +296,8 @@ struct lm85_data {
struct i2c_client *client;
const struct attribute_group *groups[6];
const int *freq_map;
+ unsigned int freq_map_mask;
+
enum chips type;
bool has_vid5; /* true if VID5 is configured for ADT7463 or ADT7468 */
@@ -514,7 +516,7 @@ static struct lm85_data *lm85_update_device(struct device *dev)
data->autofan[i].config =
lm85_read_value(client, LM85_REG_AFAN_CONFIG(i));
val = lm85_read_value(client, LM85_REG_AFAN_RANGE(i));
- data->pwm_freq[i] = val & 0x07;
+ data->pwm_freq[i] = val & data->freq_map_mask;
data->zone[i].range = val >> 4;
data->autofan[i].min_pwm =
lm85_read_value(client, LM85_REG_AFAN_MINPWM(i));
@@ -791,7 +793,8 @@ static ssize_t show_pwm_freq(struct device *dev,
if (IS_ADT7468_HFPWM(data))
freq = 22500;
else
- freq = FREQ_FROM_REG(data->freq_map, data->pwm_freq[nr]);
+ freq = FREQ_FROM_REG(data->freq_map, data->freq_map_mask + 1,
+ data->pwm_freq[nr]);
return sprintf(buf, "%d\n", freq);
}
@@ -820,7 +823,7 @@ static ssize_t set_pwm_freq(struct device *dev,
lm85_write_value(client, ADT7468_REG_CFG5, data->cfg5);
} else { /* Low freq. mode */
data->pwm_freq[nr] = FREQ_TO_REG(data->freq_map,
- FREQ_MAP_LEN, val);
+ data->freq_map_mask + 1, val);
lm85_write_value(client, LM85_REG_AFAN_RANGE(nr),
(data->zone[nr].range << 4)
| data->pwm_freq[nr]);
@@ -1196,7 +1199,7 @@ static ssize_t set_temp_auto_temp_min(struct device *dev,
TEMP_FROM_REG(data->zone[nr].limit));
lm85_write_value(client, LM85_REG_AFAN_RANGE(nr),
((data->zone[nr].range & 0x0f) << 4)
- | (data->pwm_freq[nr] & 0x07));
+ | data->pwm_freq[nr]);
mutex_unlock(&data->update_lock);
return count;
@@ -1232,7 +1235,7 @@ static ssize_t set_temp_auto_temp_max(struct device *dev,
val - min);
lm85_write_value(client, LM85_REG_AFAN_RANGE(nr),
((data->zone[nr].range & 0x0f) << 4)
- | (data->pwm_freq[nr] & 0x07));
+ | data->pwm_freq[nr]);
mutex_unlock(&data->update_lock);
return count;
}
@@ -1569,9 +1572,11 @@ static int lm85_probe(struct i2c_client *client, const struct i2c_device_id *id)
case emc6d103:
case emc6d103s:
data->freq_map = adm1027_freq_map;
+ data->freq_map_mask = ARRAY_SIZE(adm1027_freq_map) - 1;
break;
default:
data->freq_map = lm85_freq_map;
+ data->freq_map_mask = ARRAY_SIZE(lm85_freq_map) - 1;
}
/* Set the VRM version */
--
2.17.1
This chip expands the freqency field from 3 to 4 bits, to
add more frequencies in the 22.5 to 30 kHz ranges.
Signed-off-by: Jeremy Gebben <[email protected]>
---
Documentation/hwmon/lm85 | 3 +++
drivers/hwmon/lm85.c | 9 +++++++++
2 files changed, 12 insertions(+)
diff --git a/Documentation/hwmon/lm85 b/Documentation/hwmon/lm85
index 9f3a945d1b80..264874f09196 100644
--- a/Documentation/hwmon/lm85
+++ b/Documentation/hwmon/lm85
@@ -140,6 +140,9 @@ of voltage and temperature channels.
SMSC EMC6D103S is similar to EMC6D103, but does not support pwm#_auto_pwm_minctl
and temp#_auto_temp_off.
+The LM96000 supports additional high frequency PWM modes (22.5 kHz, 24 kHz,
+25.7 kHz, 27.7 kHz and 30 kHz), which can be configured on a per-PWM basis.
+
Hardware Configurations
-----------------------
diff --git a/drivers/hwmon/lm85.c b/drivers/hwmon/lm85.c
index 48f59667a88c..37bfa7d105a2 100644
--- a/drivers/hwmon/lm85.c
+++ b/drivers/hwmon/lm85.c
@@ -202,6 +202,11 @@ static const int lm85_freq_map[] = { /* 1 Hz */
10, 15, 23, 30, 38, 47, 61, 94
};
+static const int lm85_high_freq_map[] = { /* 1 Hz */
+ 10, 15, 23, 30, 38, 47, 61, 94,
+ 22500, 24000, 25700, 25700, 27700, 27700, 30000, 30000
+};
+
static const int adm1027_freq_map[] = { /* 1 Hz */
11, 15, 22, 29, 35, 44, 59, 88
};
@@ -1574,6 +1579,10 @@ static int lm85_probe(struct i2c_client *client, const struct i2c_device_id *id)
data->freq_map = adm1027_freq_map;
data->freq_map_mask = ARRAY_SIZE(adm1027_freq_map) - 1;
break;
+ case lm85:
+ data->freq_map = lm85_high_freq_map;
+ data->freq_map_mask = ARRAY_SIZE(lm85_high_freq_map) - 1;
+ break;
default:
data->freq_map = lm85_freq_map;
data->freq_map_mask = ARRAY_SIZE(lm85_freq_map) - 1;
--
2.17.1
It has been supported as a generic lm85 for quite some time.
Signed-off-by: Jeremy Gebben <[email protected]>
---
Documentation/hwmon/lm85 | 6 +++++-
1 file changed, 5 insertions(+), 1 deletion(-)
diff --git a/Documentation/hwmon/lm85 b/Documentation/hwmon/lm85
index 7c49feaa79d2..9f3a945d1b80 100644
--- a/Documentation/hwmon/lm85
+++ b/Documentation/hwmon/lm85
@@ -3,9 +3,13 @@ Kernel driver lm85
Supported chips:
* National Semiconductor LM85 (B and C versions)
- Prefix: 'lm85'
+ Prefix: 'lm85b' or 'lm85c'
Addresses scanned: I2C 0x2c, 0x2d, 0x2e
Datasheet: http://www.national.com/pf/LM/LM85.html
+ * Texas Instruments LM96000
+ Prefix: 'lm85'
+ Addresses scanned: I2C 0x2c, 0x2d, 0x2e
+ Datasheet: http://www.ti.com/lit/ds/symlink/lm96000.pdf
* Analog Devices ADM1027
Prefix: 'adm1027'
Addresses scanned: I2C 0x2c, 0x2d, 0x2e
--
2.17.1
On Tue, Jan 29, 2019 at 4:38 PM Guenter Roeck <[email protected]> wrote:
>
> On Tue, Jan 29, 2019 at 02:29:17PM -0700, Jeremy Gebben wrote:
> > Hi,
> >
> > This patch adds support for the PWM frequencies from 22.5 to 30 kHz
> > available on the LM96000.
> >
> > It looks like this chip has been supported for a long time, but wasn't
> > mentioned in the docs (which I have updated).
> >
> > Also, it has been using the generic 'lm85' prefix, which I have not
> > changed to avoid breaking userspace or device tree files. AFAICT,
> > lm85_detect() will only return 'lm85' for LM96000 chips, so it doesn't
> > look like you can get a bare 'lm85' prefix with any other chips.
> >
> > I stumbled on to a 10 year old thread discussing a patch which looks
> > like an early attempt to add support for this chip, which may be of
> > interest:
> > https://lm-sensors.lm-sensors.narkive.com/1SIwaMDT/patch-hwmon-lm96000-support
> >
> > Thanks for reviewing,
> >
> Where are the actual patches ?
Oops! It looks like they didn't send correctly. Hopefully they should
be showing up shortly.
Jeremy
>
> Guenter
On Wed, Jan 30, 2019 at 09:14:45AM -0700, Jeremy Gebben wrote:
>
> Allow support for chips that support more than 8 frequencies.
> The size still must be a power of two.
>
This is risky. Someone later may not know/remember and
expand the table without maintaining that restriction.
I would suggest to use map_size instead of map_mask.
While that adds some divide operations, none of those are
in a critical path, and the code would be much more robust.
> Signed-off-by: Jeremy Gebben <[email protected]>
> ---
> drivers/hwmon/lm85.c | 25 +++++++++++++++----------
> 1 file changed, 15 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/hwmon/lm85.c b/drivers/hwmon/lm85.c
> index 0a325878e8f5..48f59667a88c 100644
> --- a/drivers/hwmon/lm85.c
> +++ b/drivers/hwmon/lm85.c
> @@ -198,13 +198,13 @@ static int RANGE_TO_REG(long range)
> #define RANGE_FROM_REG(val) lm85_range_map[(val) & 0x0f]
>
> /* These are the PWM frequency encodings */
> -static const int lm85_freq_map[8] = { /* 1 Hz */
> +static const int lm85_freq_map[] = { /* 1 Hz */
> 10, 15, 23, 30, 38, 47, 61, 94
> };
> -static const int adm1027_freq_map[8] = { /* 1 Hz */
> +
> +static const int adm1027_freq_map[] = { /* 1 Hz */
> 11, 15, 22, 29, 35, 44, 59, 88
> };
> -#define FREQ_MAP_LEN 8
>
> static int FREQ_TO_REG(const int *map,
> unsigned int map_size, unsigned long freq)
> @@ -212,9 +212,9 @@ static int FREQ_TO_REG(const int *map,
> return find_closest(freq, map, map_size);
> }
>
> -static int FREQ_FROM_REG(const int *map, u8 reg)
> +static int FREQ_FROM_REG(const int *map, unsigned int map_size, u8 reg)
> {
> - return map[reg & 0x07];
> + return map[reg & (map_size - 1)];
Passing in map_size (calculated as map_mask + 1) just to subtract one
doesn't make sense.
> }
>
> /*
> @@ -296,6 +296,8 @@ struct lm85_data {
> struct i2c_client *client;
> const struct attribute_group *groups[6];
> const int *freq_map;
> + unsigned int freq_map_mask;
> +
> enum chips type;
>
> bool has_vid5; /* true if VID5 is configured for ADT7463 or ADT7468 */
> @@ -514,7 +516,7 @@ static struct lm85_data *lm85_update_device(struct device *dev)
> data->autofan[i].config =
> lm85_read_value(client, LM85_REG_AFAN_CONFIG(i));
> val = lm85_read_value(client, LM85_REG_AFAN_RANGE(i));
> - data->pwm_freq[i] = val & 0x07;
> + data->pwm_freq[i] = val & data->freq_map_mask;
> data->zone[i].range = val >> 4;
> data->autofan[i].min_pwm =
> lm85_read_value(client, LM85_REG_AFAN_MINPWM(i));
> @@ -791,7 +793,8 @@ static ssize_t show_pwm_freq(struct device *dev,
> if (IS_ADT7468_HFPWM(data))
> freq = 22500;
> else
> - freq = FREQ_FROM_REG(data->freq_map, data->pwm_freq[nr]);
> + freq = FREQ_FROM_REG(data->freq_map, data->freq_map_mask + 1,
> + data->pwm_freq[nr]);
>
> return sprintf(buf, "%d\n", freq);
> }
> @@ -820,7 +823,7 @@ static ssize_t set_pwm_freq(struct device *dev,
> lm85_write_value(client, ADT7468_REG_CFG5, data->cfg5);
> } else { /* Low freq. mode */
> data->pwm_freq[nr] = FREQ_TO_REG(data->freq_map,
> - FREQ_MAP_LEN, val);
> + data->freq_map_mask + 1, val);
> lm85_write_value(client, LM85_REG_AFAN_RANGE(nr),
> (data->zone[nr].range << 4)
> | data->pwm_freq[nr]);
> @@ -1196,7 +1199,7 @@ static ssize_t set_temp_auto_temp_min(struct device *dev,
> TEMP_FROM_REG(data->zone[nr].limit));
> lm85_write_value(client, LM85_REG_AFAN_RANGE(nr),
> ((data->zone[nr].range & 0x0f) << 4)
> - | (data->pwm_freq[nr] & 0x07));
> + | data->pwm_freq[nr]);
>
> mutex_unlock(&data->update_lock);
> return count;
> @@ -1232,7 +1235,7 @@ static ssize_t set_temp_auto_temp_max(struct device *dev,
> val - min);
> lm85_write_value(client, LM85_REG_AFAN_RANGE(nr),
> ((data->zone[nr].range & 0x0f) << 4)
> - | (data->pwm_freq[nr] & 0x07));
> + | data->pwm_freq[nr]);
> mutex_unlock(&data->update_lock);
> return count;
> }
> @@ -1569,9 +1572,11 @@ static int lm85_probe(struct i2c_client *client, const struct i2c_device_id *id)
> case emc6d103:
> case emc6d103s:
> data->freq_map = adm1027_freq_map;
> + data->freq_map_mask = ARRAY_SIZE(adm1027_freq_map) - 1;
> break;
> default:
> data->freq_map = lm85_freq_map;
> + data->freq_map_mask = ARRAY_SIZE(lm85_freq_map) - 1;
> }
You'd have to make sure that freq_map_mask is indeed a mask, ie that
ARRAY_SIZE() is a power of two. I would suggest to use freq_map_size
instead.
>
> /* Set the VRM version */
> --
> 2.17.1
>