Skip to contentSkip to conversation

hwmon: driver for TI tmp102 temperature sensor

PATCH In review Steven King 3 participants 11 messages Feb 4, 2010 – Mar 14, 2010

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 hwmon: driver for TI tmp102 temperature sensor

The TI TMP102 is similar to the lm75. It differs from the lm75 by having a 16 bit conf register and the temp registers have a minimum resolution of 12bits; the extended conf register can select 13 bit resolution (which this driver does) and also change the update rate (which this driver currently doesn't use).

---

Driver for the TI TMP102.

 Documentation/hwmon/tmp102 |   27 ++++
 drivers/hwmon/Kconfig      |   10 ++
 drivers/hwmon/Makefile     |    1 +
 drivers/hwmon/tmp102.c     |  345 ++++++++++++++++++++++++++++++++++++++++++++
 4 files changed, 383 insertions(+), 0 deletions(-)
Documentation/hwmon/tmp102 new file +27 -0
new file mode 100644
index 0000000..239dded
@@ -0,0 +1,27 @@
+Kernel driver tmp102
+==========
+
+Supported chips:
+  * Texas Instruments TMP102
+    Prefix: 'tmp102'
+    Addresses scanned: I2C 0x48 0x49 0x4a 0x4b
+    Datasheet: http://focus.ti.com/docs/prod/folders/print/tmp102.html
+
+Author:
+	Steven King <[email protected]>
+
+Description
+-----------
+
+The Texas Instruments TMP102 implements one temperature sensor.  Limits can be
+set through the Overtemperature Shutdown register and Hysteresis register.  The
+sensor is accurate to 0.5 degrees over the range of -25 to +85 C, and to 1.0
+degrees from -40 to +125 C. Resolution of the sensor is 0.0625 degree.  The
+operating temperature has a minimum of -55 C and a maximum of +150 C.
+
+The TMP102 has a programmable update rate that can select between 8, 4, 1, and
+0.5 Hz. (Currently the driver only supports the default of 4 Hz).
+
+The driver provides the common sysfs-interface for temperatures (see
+/Documentation/hwmon/sysfs-interface under Temperatures).
+
drivers/hwmon/Kconfig +10 -0
index 68cf877..cf9d7d3 100644
@@ -812,6 +812,16 @@ config SENSORS_THMC50
 	  This driver can also be built as a module.  If so, the module
 	  will be called thmc50.
 
+config SENSORS_TMP102
+	tristate "Texas Instruments TMP102 and compatibles"
+	depends on I2C && EXPERIMENTAL
+	help
+	  If you say yes here you get support for Texas Instruments TMP102
+	  sensor chips.
+
+	  This driver can also be built as a module.  If so, the module
+	  will be called tmp102.
+
 config SENSORS_TMP401
 	tristate "Texas Instruments TMP401 and compatibles"
 	depends on I2C && EXPERIMENTAL
drivers/hwmon/Makefile +1 -0
index 4bc215c..0e5cf73 100644
@@ -88,6 +88,7 @@ obj-$(CONFIG_SENSORS_SMSC47M1)	+= smsc47m1.o
 obj-$(CONFIG_SENSORS_SMSC47M192)+= smsc47m192.o
 obj-$(CONFIG_SENSORS_AMC6821)	+= amc6821.o
 obj-$(CONFIG_SENSORS_THMC50)	+= thmc50.o
+obj-$(CONFIG_SENSORS_TMP102)	+= tmp102.o
 obj-$(CONFIG_SENSORS_TMP401)	+= tmp401.o
 obj-$(CONFIG_SENSORS_TMP421)	+= tmp421.o
 obj-$(CONFIG_SENSORS_VIA_CPUTEMP)+= via-cputemp.o
drivers/hwmon/tmp102.c new file +345 -0

345 lines not on this page · raw message

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

#1 lore raw
Quoted 5 lines · On Wed, 3 Feb 2010 17:23:49 -0800 Steven King <[email protected]> wrote:

The TI TMP102 is similar to the lm75. It differs from the lm75 by having a 16 bit conf register and the temp registers have a minimum resolution of 12bits; the extended conf register can select 13 bit resolution (which this driver does) and also change the update rate (which this driver currently doesn't use).

A neat little driver.

checkpatch spits this warning:

    WARNING: struct dev_pm_ops should normally be const
    #387: FILE: drivers/hwmon/tmp102.c:300:
    +static struct dev_pm_ops tmp102_dev_pm_ops = {

which seems truthful enough.

drivers/hwmon/tmp102.c renamed from drivers/hwmon/tmp102.c~hwmon-driver-for-ti-tmp102-temperature-sensor-checkpatch-fixes +1 -1
@@ -297,7 +297,7 @@ static int tmp102_resume(struct device *
 	return 0;
 }
 
-static struct dev_pm_ops tmp102_dev_pm_ops = {
+static const struct dev_pm_ops tmp102_dev_pm_ops = {
 	.suspend	= tmp102_suspend,
 	.resume		= tmp102_resume,
 };

_

And doing this will hurt readers' brains less:

Use conventional array-walk loop.

drivers/hwmon/tmp102.c renamed from drivers/hwmon/tmp102.c~hwmon-driver-for-ti-tmp102-temperature-sensor-fix +4 -3
@@ -91,13 +91,14 @@ static struct tmp102 *tmp102_update_devi
 
 	mutex_lock(&tmp102->lock);
 	if (time_after(jiffies, tmp102->last_update + HZ / 4)) {
-		int i = 0;
-		do {
+		int i;
+		
+		for (i = 0; i < ARRAY_SIZE(tmp102->temp); i++) {
 			int status = tmp102_read_reg(client, tmp102_reg[i]);
 			if (status > -1)
 				tmp102->temp[i] = tmp102_reg_to_mC(status);
 
-		} while (++i < ARRAY_SIZE(tmp102->temp));
+		}
 		tmp102->last_update = jiffies;
 	}
 	mutex_unlock(&tmp102->lock);

_

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

#2 lore raw
Quoted 10 lines · On Thursday 04 February 2010 10:22:03 Andrew Morton wrote:

On Wed, 3 Feb 2010 17:23:49 -0800

Quoted 5 lines · Steven King

The TI TMP102 is similar to the lm75. It differs from the lm75 by having a 16 bit conf register and the temp registers have a minimum resolution of 12bits; the extended conf register can select 13 bit resolution (which this driver does) and also change the update rate (which this driver currently doesn't use).

A neat little driver.

Thanks.

Quoted 7 lines

checkpatch spits this warning:

    WARNING: struct dev_pm_ops should normally be const
    #387: FILE: drivers/hwmon/tmp102.c:300:
    +static struct dev_pm_ops tmp102_dev_pm_ops = {

which seems truthful enough.

Indeed. I am, however, somewhat surprised since I ran the patch thru
checkpatch before posting it and no errors or warnings were reported. Is there a version of checkpatch other than the one included in the tree that I should be using?

Quoted 6 lines

And doing this will hurt readers' brains less:

Use conventional array-walk loop.

Ah yes, an idiosyncrasy of mine in preferring do while over for loops especially when I 'know' the initial test will pass. Whatever is the preferred idiom for the kernel is fine with me.

Signature · 6 lines
Steven King -- sfking at fdwdc dot com

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#3 lore raw
Quoted 6 lines · On Thu, 4 Feb 2010 13:17:37 -0800 Steven King <[email protected]> wrote:
Quoted 6 lines

checkpatch spits this warning:

    WARNING: struct dev_pm_ops should normally be const
    #387: FILE: drivers/hwmon/tmp102.c:300:
    +static struct dev_pm_ops tmp102_dev_pm_ops = {
Quoted 6 lines
Quoted 1 line

which seems truthful enough.

Indeed. I am, however, somewhat surprised since I ran the patch thru
checkpatch before posting it and no errors or warnings were reported. Is there a version of checkpatch other than the one included in the tree that I should be using?

It was -mm's checkpatchpl-extend-list-of-expected-to-be-const-structures.patch which found this.

From: Emese Revfy <[email protected]>

Based on Arjan's suggestion, extend the list of ops structures that should be const.

Cc Andy Whitcroft <[email protected]>
Cc Arjan van de Ven <[email protected]>

---

 scripts/checkpatch.pl |   41 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 39 insertions(+), 2 deletions(-)
scripts/checkpatch.pl renamed from scripts/checkpatch.pl~checkpatchpl-extend-list-of-expected-to-be-const-structures +39 -2
diff -puN scripts/checkpatch.pl~checkpatchpl-extend-list-of-expected-to-be-const-structures scripts/checkpatch.pl
@@ -2654,9 +2654,46 @@ sub process {
 		if ($line =~ /^.\s*__initcall\s*\(/) {
 			WARN("please use device_initcall() instead of __initcall()\n" . $herecurr);
 		}
-# check for struct file_operations, ensure they are const.
+# check for various ops structs, ensure they are const.
+		my $struct_ops = qr{acpi_dock_ops|
+				address_space_operations|
+				backlight_ops|
+				block_device_operations|
+				dentry_operations|
+				dev_pm_ops|
+				dma_map_ops|
+				extent_io_ops|
+				file_lock_operations|
+				file_operations|
+				hv_ops|
+				ide_dma_ops|
+				intel_dvo_dev_ops|
+				item_operations|
+				iwl_ops|
+				kgdb_arch|
+				kgdb_io|
+				kset_uevent_ops|
+				lock_manager_operations|
+				microcode_ops|
+				mtrr_ops|
+				neigh_ops|
+				nlmsvc_binding|
+				pci_raw_ops|
+				pipe_buf_operations|
+				platform_hibernation_ops|
+				platform_suspend_ops|
+				proto_ops|
+				rpc_pipe_ops|
+				seq_operations|
+				snd_ac97_build_ops|
+				soc_pcmcia_socket_ops|
+				stacktrace_ops|
+				sysfs_ops|
+				tty_operations|
+				usb_mon_operations|
+				wd_ops}x;
 		if ($line !~ /\bconst\b/ &&
-		    $line =~ /\bstruct\s+(file_operations|seq_operations)\b/) {
+		    $line =~ /\bstruct\s+($struct_ops)\b/) {
 			WARN("struct $1 should normally be const\n" .
 				$herecurr);
 		}

_

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors

#4 lore raw
In reply to Steven King

Hi Steven,

Quoted 9 lines · Steven King

The TI TMP102 is similar to the lm75. It differs from the lm75 by having a 16 bit conf register and the temp registers have a minimum resolution of 12bits; the extended conf register can select 13 bit resolution (which this driver does) and also change the update rate (which this driver currently doesn't use).

---

Driver for the TI TMP102.

Could you please send a dump of your TMP102 chip, using i2cdump in word mode?

Signature · 6 lines
Jean Delvare

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#5 lore raw
Quoted 9 lines · On Thursday 04 February 2010 10:37:12 Jean Delvare wrote:

Hi Steven,

Quoted 6 lines · Steven King

The TI TMP102 is similar to the lm75. It differs from the lm75 by having a 16 bit conf register and the temp registers have a minimum resolution of 12bits; the extended conf register can select 13 bit resolution (which this driver does) and also change the update rate (which this driver currently doesn't use).

Quoted 7 lines
Quoted 4 lines

---

Driver for the TI TMP102.

Could you please send a dump of your TMP102 chip, using i2cdump in word mode?

sure, like so?

# ./i2cdump 0 0x4a w
WARNING! This program can confuse your I2C bus, cause data loss and worse!
I will probe file /dev/i2c-0, address 0x4a, mode word
Continue? [Y/n] y
     0,8  1,9  2,a  3,b  4,c  5,d  6,e  7,f
00: d90c b062 002d 004b c01a d006 0000 0000
08: d90c b062 002d 004b c01a d006 0000 0000
10: d90c b062 002d 004b c01a d006 0000 0000
18: d90c b062 002d 004b c01a d006 0000 0000
20: d90c b062 002d 004b c01a d006 0000 0000
28: d90c b062 002d 004b c01a d006 0000 0000
30: d90c b062 002d 004b c01a d006 0000 0000
38: d90c b062 002d 004b c01a d006 0000 0000
40: d90c b062 002d 004b c01a d006 0000 0000
48: d90c b062 002d 004b c01a d006 0000 0000
50: d90c b062 002d 004b c01a d006 0000 0000
58: d90c b062 002d 004b c01a d006 0000 0000
60: d90c b062 002d 004b c01a d006 0000 0000
68: d90c b062 002d 004b c01a d006 0000 0000
70: d90c b062 002d 004b c01a d006 0000 0000
78: d90c b062 002d 004b c01a d006 0000 0000
80: d90c b062 002d 004b c01a d006 0000 0000
88: d90c b062 002d 004b c01a d006 0000 0000
90: d90c b062 002d 004b c01a d006 0000 0000
98: d90c b062 002d 004b c01a d006 0000 0000
a0: d90c b062 002d 004b c01a d006 0000 0000
a8: d90c b062 002d 004b c01a d006 0000 0000
b0: d90c b062 002d 004b c01a d006 0000 0000
b8: d90c b062 002d 004b c01a d006 0000 0000
c0: d90c b062 002d 004b c01a d006 0000 0000
c8: d90c b062 002d 004b c01a d006 0000 0000
d0: d90c b062 002d 004b c01a d006 0000 0000
d8: d90c b062 002d 004b c01a d006 0000 0000
e0: d10c b062 002d 004b c01a d006 0000 0000
e8: d10c b062 002d 004b c01a d006 0000 0000
f0: d10c b062 002d 004b c01a d006 0000 0000
f8: d10c b062 002d 004b c01a d006 0000 0000
# 
Signature · 7 lines
Steven King -- sfking at fdwdc dot com


_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#6 lore raw
Quoted 36 lines · Steven King
Quoted 2 lines · On Thursday 04 February 2010 10:37:12 Jean Delvare wrote:

Could you please send a dump of your TMP102 chip, using i2cdump in word mode?

sure, like so?

# ./i2cdump 0 0x4a w
WARNING! This program can confuse your I2C bus, cause data loss and worse!
I will probe file /dev/i2c-0, address 0x4a, mode word
Continue? [Y/n] y
     0,8  1,9  2,a  3,b  4,c  5,d  6,e  7,f
00: d90c b062 002d 004b c01a d006 0000 0000
08: d90c b062 002d 004b c01a d006 0000 0000
10: d90c b062 002d 004b c01a d006 0000 0000
18: d90c b062 002d 004b c01a d006 0000 0000
20: d90c b062 002d 004b c01a d006 0000 0000
28: d90c b062 002d 004b c01a d006 0000 0000
30: d90c b062 002d 004b c01a d006 0000 0000
38: d90c b062 002d 004b c01a d006 0000 0000
40: d90c b062 002d 004b c01a d006 0000 0000
48: d90c b062 002d 004b c01a d006 0000 0000
50: d90c b062 002d 004b c01a d006 0000 0000
58: d90c b062 002d 004b c01a d006 0000 0000
60: d90c b062 002d 004b c01a d006 0000 0000
68: d90c b062 002d 004b c01a d006 0000 0000
70: d90c b062 002d 004b c01a d006 0000 0000
78: d90c b062 002d 004b c01a d006 0000 0000
80: d90c b062 002d 004b c01a d006 0000 0000
88: d90c b062 002d 004b c01a d006 0000 0000
90: d90c b062 002d 004b c01a d006 0000 0000
98: d90c b062 002d 004b c01a d006 0000 0000
a0: d90c b062 002d 004b c01a d006 0000 0000
a8: d90c b062 002d 004b c01a d006 0000 0000
b0: d90c b062 002d 004b c01a d006 0000 0000
b8: d90c b062 002d 004b c01a d006 0000 0000
c0: d90c b062 002d 004b c01a d006 0000 0000
Quoted 8 lines
c8: d90c b062 002d 004b c01a d006 0000 0000
d0: d90c b062 002d 004b c01a d006 0000 0000
d8: d90c b062 002d 004b c01a d006 0000 0000
e0: d10c b062 002d 004b c01a d006 0000 0000
e8: d10c b062 002d 004b c01a d006 0000 0000
f0: d10c b062 002d 004b c01a d006 0000 0000
f8: d10c b062 002d 004b c01a d006 0000 0000
# 

Exactly, thank you!

Signature · 6 lines
Jean Delvare

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#7 lore raw

Re: [lm-sensors] [PATCH] hwmon: driver for TI tmp102 temperature

In reply to Steven King

Hi Andrew,

Adding back Steven and the lm-sensors list...

Quoted 1 line · Andrew Morton

So you're OK with the patch as-is?

I did not review it. And testing it doesn't seem positive. First of all, I get the following in my logs after loading the driver:

(null): not a tmp102

Which suggests the use of an uninitialized device struct. Also, the detection fails early, I don't think the detection routine works. Looking at the code, it doesn't seem to correspond to the TMP102 register map at all (assuming the dump I got from Steven is really from a TMP102) Steven, did you ever test it? Honestly, I don't think it makes sense to have a detect routine for this chip, given that it lacks identification registers. We relied on ugly detection quirks for the LM75 only because that chip was very popular on PC motherboards at one point in time. For devices used on embedded designs and which are always enumerated, we don't need detection routines.

Then, using the tmp102 driver on the dump sent by Steven produces the following "sensors" output:

tmp102-i2c-3-48
Adapter: SMBus stub driver
temp1: +2.6°C (high = +15.0°C, hyst = +9.0°C)

These values are suspiciously low and smell like the wrong base unit is used (1/100 °C instead of 1/1000 °C). Quick code examination seems to confirm this.

So, no, I am not OK with the patch as-is, it needs more work.

Signature · 6 lines
Jean Delvare

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#8 lore raw

Re: [lm-sensors] [PATCH] hwmon: driver for TI tmp102 temperature

In reply to Steven King
Quoted 12 lines · On Saturday 06 March 2010 12:35:21 Jean Delvare wrote:

Hi Andrew,

Adding back Steven and the lm-sensors list...

Quoted 1 line · Andrew Morton

So you're OK with the patch as-is?

I did not review it. And testing it doesn't seem positive. First of all, I get the following in my logs after loading the driver:

(null): not a tmp102

Quoted 10 lines

Which suggests the use of an uninitialized device struct. Also, the detection fails early, I don't think the detection routine works. Looking at the code, it doesn't seem to correspond to the TMP102 register map at all (assuming the dump I got from Steven is really from a TMP102) Steven, did you ever test it? Honestly, I don't think it makes sense to have a detect routine for this chip, given that it lacks identification registers. We relied on ugly detection quirks for the LM75 only because that chip was very popular on PC motherboards at one point in time. For devices used on embedded designs and which are always enumerated, we don't need detection routines.

For my specific use the part is connected to an embedded system and is always enumerated; as lm-sensors doesnt build for this target I wasnt aware of any way to test the detection code. I certainly wouldnt have any problem with removing the detection routine entirely...

Quoted 10 lines

Then, using the tmp102 driver on the dump sent by Steven produces the following "sensors" output:

tmp102-i2c-3-48
Adapter: SMBus stub driver
temp1: +2.6°C (high = +15.0°C, hyst = +9.0°C)

These values are suspiciously low and smell like the wrong base unit is used (1/100 °C instead of 1/1000 °C). Quick code examination seems to confirm this.

Doh! My bad.

Quoted 1 line

So, no, I am not OK with the patch as-is, it needs more work.

Okay, so besides removing the detect routine, incorporating Andrew's patch and fixing the base for the temperature conversion, was there anything else I need to do for v2?

Signature · 6 lines
Steven King -- sfking at fdwdc dot com

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#9 lore raw

Re: [lm-sensors] [PATCH] hwmon: driver for TI tmp102 temperature

In reply to Steven King

Hi Steven,

Quoted 17 lines · Steven King
Quoted 5 lines · On Saturday 06 March 2010 12:35:21 Jean Delvare wrote:

I did not review it. And testing it doesn't seem positive. First of all, I get the following in my logs after loading the driver:

(null): not a tmp102

Quoted 10 lines

Which suggests the use of an uninitialized device struct. Also, the detection fails early, I don't think the detection routine works. Looking at the code, it doesn't seem to correspond to the TMP102 register map at all (assuming the dump I got from Steven is really from a TMP102) Steven, did you ever test it? Honestly, I don't think it makes sense to have a detect routine for this chip, given that it lacks identification registers. We relied on ugly detection quirks for the LM75 only because that chip was very popular on PC motherboards at one point in time. For devices used on embedded designs and which are always enumerated, we don't need detection routines.

Quoted 2 lines

For my specific use the part is connected to an embedded system and is always enumerated; as lm-sensors doesnt build for this target I wasnt aware of any

Out of curiosity, why is lm-sensors not building? What is missing?

Quoted 2 lines

way to test the detection code. I certainly wouldnt have any problem with removing the detection routine entirely...

I'm fine with this.

Quoted 11 lines
Quoted 10 lines

Then, using the tmp102 driver on the dump sent by Steven produces the following "sensors" output:

tmp102-i2c-3-48
Adapter: SMBus stub driver
temp1: +2.6°C (high = +15.0°C, hyst = +9.0°C)

These values are suspiciously low and smell like the wrong base unit is used (1/100 °C instead of 1/1000 °C). Quick code examination seems to confirm this.

Quoted 7 lines

Doh! My bad.

Quoted 1 line

So, no, I am not OK with the patch as-is, it needs more work.

Okay, so besides removing the detect routine, incorporating Andrew's patch and fixing the base for the temperature conversion, was there anything else I need to do for v2?

Nothing spotted by my quick look, but I didn't actually review your code. If you post an updated version of your patch, I could review it.

Signature · 6 lines
Jean Delvare

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors
#10 lore raw

Re: [lm-sensors] [PATCH] hwmon: driver for TI tmp102 temperature

In reply to Steven King
Quoted 1 line · On Sunday 07 March 2010 04:23:38 Jean Delvare wrote:

Hi Steven,

Hi Jean,

Quoted 6 lines
Quoted 4 lines

For my specific use the part is connected to an embedded system and is always enumerated; as lm-sensors doesnt build for this target I wasnt aware of any

Out of curiosity, why is lm-sensors not building? What is missing?

The target is nommu (coldfire) and the build failed with a bunch of 'ERROR: reloc type R_68K_32 is not supported for PIC'. Probably trivial to fix but I havent had the chance to look at it yet.

Quoted 3 lines

Nothing spotted by my quick look, but I didn't actually review your code. If you post an updated version of your patch, I could review it.

Great! I'll get an updated patch out right away.

Signature · 6 lines
Steven King -- sfking at fdwdc dot com

_______________________________________________
lm-sensors mailing list
[email protected]
http://lists.lm-sensors.org/mailman/listinfo/lm-sensors