Add G762/G763 PWM fan controller

Signed-off-by Arnaud Ebalard <[email protected]>

---
 drivers/hwmon/Kconfig              |   10 +
 drivers/hwmon/Makefile             |    1 +
 drivers/hwmon/g762.c               | 1012 ++++++++++++++++++++++++++++++++++++
 include/linux/platform_data/g762.h |   54 ++
 4 files changed, 1077 insertions(+)
 create mode 100644 drivers/hwmon/g762.c
 create mode 100644 include/linux/platform_data/g762.h

drivers/hwmon/Kconfig +0 -0

index 0428e8a..142bdf8 100644

include/linux/platform_data/g762.h new file +30 -0

new file mode 100644
index 0000000..29e3934
@@ -0,0 +1,54 @@
+/*
+ * Platform data structure for g762 fan controller driver
+ *
+ * Copyright (C) 2013, Arnaud EBALARD <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#ifndef __LINUX_PLATFORM_DATA_G762_H__
+#define __LINUX_PLATFORM_DATA_G762_H__
+
+#define G762_DEFAULT_NO_OVERLOAD 0xffffffff
+
+struct g762_platform_data {
+	u32 fan_startv;
+	u32 fan_gear_mode;
+	u32 fan_div;
+	u32 fan_pulses;
+	u32 fan_target;

Quoted 10 lines · Arnaud Ebalard

Signed-off-by Arnaud Ebalard <[email protected]>

---
 drivers/hwmon/Kconfig              |   10 +
 drivers/hwmon/Makefile             |    1 +
 drivers/hwmon/g762.c               | 1012 ++++++++++++++++++++++++++++++++++++
 include/linux/platform_data/g762.h |   54 ++
 4 files changed, 1077 insertions(+)
 create mode 100644 drivers/hwmon/g762.c
 create mode 100644 include/linux/platform_data/g762.h

Hi Arnaud,

After more tests on my 2Big Network v2 board, it appears that the fan doesn't rotate when PWM mode (the preferred operating mode for this board) is selected. Nevertheless, DC mode is usable (even if not ideal given the hardware). After some investigations I noticed that an extra initialization is needed to enable PWM mode on my board: the set_cnt register must be set to 0 while the default value is 0xff. Is that specific to my hardware ? Is PWM mode working on your ReadyNAS with the default set_cnt value ?

I haven't noticed this issue while testing your v1 patch series because at the time I was using a board with an U-Boot modified by LaCie. This last sets the set_cnt register to 0 while U-Boot mainline don't. Actually, the 2Big fan is not configured nor even started by U-Boot mainline. I have to fix this but maybe that something can be done in the g762 Linux driver too ?

Quoted 4 lines

+/* When filling a g762_platform_data structure to be passed during platform
+ * init to the driver, it needs to be initialized to the following default
+ * value before changing specific fields. This is needed to avoid a sparse
+ * initialization to have current values replaced by 0 */

In other places, you use this multi-line comment format:

/*
* ...
* ...
*/

include/linux/platform_data/g762.h line 52

Quoted 11 lines

include/linux/platform_data/g762.h new file +11 -0

+static const struct g762_platform_data G762_DEFAULT_PDATA = {
+	.fan_startv = G762_DEFAULT_NO_OVERLOAD,
+	.fan_gear_mode = G762_DEFAULT_NO_OVERLOAD,
+	.fan_div = G762_DEFAULT_NO_OVERLOAD,
+	.fan_pulses = G762_DEFAULT_NO_OVERLOAD,
+	.fan_target = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_polarity = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_enable = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_freq = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_mode = G762_DEFAULT_NO_OVERLOAD,
+};

Are you sure that all this settings are needed ? IMHO the g762 platform data should only holds the platform specific configuration: fan_startv, fan_gear_mode, fan_div and fan_pulses. The other settings are not really related with the platform but more with the operating fan configuration. For my part, I am happy enough with the sysfs hwmon interface.

The same comment applies to the g762 DT binding properties.

Quoted 6 lines

I haven't noticed this issue while testing your v1 patch series because at the time I was using a board with an U-Boot modified by LaCie. This last sets the set_cnt register to 0 while U-Boot mainline don't. Actually, the 2Big fan is not configured nor even started by U-Boot mainline. I have to fix this but maybe that something can be done in the g762 Linux driver too ?

Hi,

Quoted 6 lines

Quoted 6 lines

I haven't noticed this issue while testing your v1 patch series because at the time I was using a board with an U-Boot modified by LaCie. This last sets the set_cnt register to 0 while U-Boot mainline don't. Actually, the 2Big fan is not configured nor even started by U-Boot mainline. I have to fix this but maybe that something can be done in the g762 Linux driver too ?

At the moment, DT bindings and platform_data structure both allow setting an initial fan_target value. I could add a 'pwm' binding for DT and an assoicated field in platform_data to allow passing an initial value so that maintainers of platforms with a 2 wire fan (open loop) can still make the fan somewhat rotate at boot.

Quoted 9 lines · Arnaud Ebalard

Hi,

Quoted 6 lines

Quoted 6 lines

I haven't noticed this issue while testing your v1 patch series because at the time I was using a board with an U-Boot modified by LaCie. This last sets the set_cnt register to 0 while U-Boot mainline don't. Actually, the 2Big fan is not configured nor even started by U-Boot mainline. I have to fix this but maybe that something can be done in the g762 Linux driver too ?

Quoted 6 lines

At the moment, DT bindings and platform_data structure both allow setting an initial fan_target value. I could add a 'pwm' binding for DT and an assoicated field in platform_data to allow passing an initial value so that maintainers of platforms with a 2 wire fan (open loop) can still make the fan somewhat rotate at boot.

Devicetree properties are supposed to describe the system, not its configuration. There are exceptions, but I don't recall seeing a case where a parameter can be set both with devicetree properties and sysfs attributes.

Before going further on this, we will need to have an ack from devicetree maintainers which specifically permits your dual-configuration approach. Even with that, you'll still have convince me that you need both (while so far no other driver needed it).

Hi Simon and Guenter,

Quoted 14 lines · Simon Guinot

Quoted 10 lines · Arnaud Ebalard

Signed-off-by Arnaud Ebalard <[email protected]>

---
 drivers/hwmon/Kconfig              |   10 +
 drivers/hwmon/Makefile             |    1 +
 drivers/hwmon/g762.c               | 1012 ++++++++++++++++++++++++++++++++++++
 include/linux/platform_data/g762.h |   54 ++
 4 files changed, 1077 insertions(+)
 create mode 100644 drivers/hwmon/g762.c
 create mode 100644 include/linux/platform_data/g762.h

Hi Arnaud,

Quoted 8 lines

After more tests on my 2Big Network v2 board, it appears that the fan doesn't rotate when PWM mode (the preferred operating mode for this board) is selected. Nevertheless, DC mode is usable (even if not ideal given the hardware). After some investigations I noticed that an extra initialization is needed to enable PWM mode on my board: the set_cnt register must be set to 0 while the default value is 0xff. Is that specific to my hardware ? Is PWM mode working on your ReadyNAS with the default set_cnt value ?

First, thanks for testing this!

Regarding your problem, I first started by booting current version of my driver on the Duo v2 w/o touching chip registers (only clock reference value). This way, I inherit the values installed by (NETGEAR's) u-boot:

$ for k in fan* pwm* ; do echo -n "$k:" ; echo `cat $k `; done
fan1_alarm:0
fan1_div:2
fan1_fault:0
fan1_input:1807
fan1_target:1807
pwm1:221
pwm1_enable:2        /* closed-loop, i.e. config is done via set_cnt */
pwm1_mode:1          /* PWM mode */ 

If G762 datasheet is correct, NETGEAR's u-boot alters fan divisor ('01' i.e. 2 instead of '00' i.e. 1). It also changes the default mode from linear to pwm and sets set_cnt to 34 (255-221).

Then, I grabbed the GPL sources (5.3.7) of NETGEAR's u-boot and found the following in a boot function:

/* Turn G76x(FAN controller, i2c address 0x3e) on.
* The FAN_SET_CNT register's offset is 0x0.
* Set [1300(rpm) == 65% == 5a(FAN_SET_CNT)] as default. */
unsigned char byte=0x5a;
if(i2c_write(0, 0x3e, 0x0, 1, &byte, 1) != 0)
puts ("Error writing the i2c chip : G76x(Fan controller).\n");

AFAICT, this is the only location in the whole source tarball where 0x3e reg is modified in an i2c function.

Then, I grabbed the u-boot source code from NETGEAR for ReadyNAS 102 (Armada-370 based NAS, also using a G762) and it has the following:

/* Turn G76x(FAN controller, i2c address 0x3e) on.
* The FAN_SET_CNT register's offset is 0x0.
* Set [1300(rpm) == 65% == 5a(FAN_SET_CNT)] as default. */
MV_U8 byte=0x5a;
if(i2c_write(0x3e, 0x0, 1, &byte, 1) != 0)

(quoted diff) +0 -0

 	puts ("Error writing the i2c chip : G76x(Fan controller).\n");
 /* Tune the FAN_STARTV */
 if (i2c_read(0x3e, 0x5, 1, &byte, 1) != 0)
 	puts ("Error reading the i2c chip : G76x(Fan controller).\n");
 byte |= 0x3;
 if(i2c_write(0x3e, 0x5, 1, &byte, 1) != 0)
 	puts ("Error writing the i2c chip : G76x(Fan controller).\n");

In the end, it does not seem NETGEAR's u-boot does some specific things except setting a non-zero value to SET_CNT register for the fan to start rotating.

BTW, if you wonder why some attributes reported at the beginning have different values than the ones in the datasheet (0x5a set in u-boot, reading 0x22) although NETGEAR's u-boot code does not seem to tamper those, I don't have the answer.

Quoted 6 lines

I haven't noticed this issue while testing your v1 patch series because at the time I was using a board with an U-Boot modified by LaCie. This last sets the set_cnt register to 0 while U-Boot mainline don't. Actually, the 2Big fan is not configured nor even started by U-Boot mainline. I have to fix this but maybe that something can be done in the g762 Linux driver too ?

You mean like reading set_cnt, setting it to 0 and then setting it back to previous value? I guess it would be good to confirm that the G762 needs that to start operating on all platforms. The datasheet does not provide info and I cannot confirm/infirm that on ReadyNAS Duo v2 and 102 platforms.

Quoted 7 lines

Quoted 6 lines

+/* When filling a g762_platform_data structure to be passed during platform
+ * init to the driver, it needs to be initialized to the following default
+ * value before changing specific fields. This is needed to avoid a sparse
+ * initialization to have current values replaced by 0 */

Quoted 6 lines

In other places, you use this multi-line comment format:

/*
* ...
* ...
*/

I missed than one, thanks.

Quoted 18 lines

include/linux/platform_data/g762.h line 52

Quoted 11 lines

include/linux/platform_data/g762.h new file +11 -0

+static const struct g762_platform_data G762_DEFAULT_PDATA = {
+	.fan_startv = G762_DEFAULT_NO_OVERLOAD,
+	.fan_gear_mode = G762_DEFAULT_NO_OVERLOAD,
+	.fan_div = G762_DEFAULT_NO_OVERLOAD,
+	.fan_pulses = G762_DEFAULT_NO_OVERLOAD,
+	.fan_target = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_polarity = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_enable = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_freq = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_mode = G762_DEFAULT_NO_OVERLOAD,
+};

Are you sure that all this settings are needed ? IMHO the g762 platform data should only holds the platform specific configuration: fan_startv, fan_gear_mode, fan_div and fan_pulses. The other settings are not really related with the platform but more with the operating fan configuration. For my part, I am happy enough with the sysfs hwmon interface.

Quoted 1 line

The same comment applies to the g762 DT binding properties.

Reading Guenter's mail, it seems he has the same opinion. He also agrees with you that the array is not a good idea in a header.

Regarding the former, I will comply to what Guenter decides but here are the reasons why I wanted to expose all those attributes.

I think choosing PWM mode and closed-loop is also platform related. For instance, on my ReadyNAS Duo v2, the fan is a 3 wire one with a tachometer and it makes sense to set the chip in closed-loop mode using previous knobs. Setting a minimal fan_target is also useful to prevent someone booting a kernel to just have no cooling. One could argue that the u-boot provided by the manufacturer has already taken care of those but:

- the user may install a stock u-boot

- the manufacturer may have performed this config in the kernel and not using u-boot

Regarding the latter, I will try and come with an other approach, probably using a function instead of an array if it is acceptable. My initial idea was that the platform maintainer would be able to to do sth like:

  struct g762_platform_data pdata = G762_DEFAULT_PDATA;
  pdata.pwm_freq = 8192;

and then pass the structure to the i2c init routine if he only wants to give a reference clock frequency. Other attributes would not be modified.

Additionally, the problem with using 0 as a "use default value" is that 0 is used for linear mode (sysfs interface for pwm_mode), 0 is also a valid value for fan_startv. Same for pwm_polarity and fan_gear_mode. Hence my decision to use something else.

Guenter, I will work on a v3 this evening based on your comments.

Cheers,

a+

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Quoted 33 lines · Arnaud Ebalard

Hi Simon and Guenter,

Quoted 14 lines · Simon Guinot

Quoted 10 lines · Arnaud Ebalard

Signed-off-by Arnaud Ebalard <[email protected]>

---
 drivers/hwmon/Kconfig              |   10 +
 drivers/hwmon/Makefile             |    1 +
 drivers/hwmon/g762.c               | 1012 ++++++++++++++++++++++++++++++++++++
 include/linux/platform_data/g762.h |   54 ++
 4 files changed, 1077 insertions(+)
 create mode 100644 drivers/hwmon/g762.c
 create mode 100644 include/linux/platform_data/g762.h

Hi Arnaud,

Quoted 8 lines

After more tests on my 2Big Network v2 board, it appears that the fan doesn't rotate when PWM mode (the preferred operating mode for this board) is selected. Nevertheless, DC mode is usable (even if not ideal given the hardware). After some investigations I noticed that an extra initialization is needed to enable PWM mode on my board: the set_cnt register must be set to 0 while the default value is 0xff. Is that specific to my hardware ? Is PWM mode working on your ReadyNAS with the default set_cnt value ?

First, thanks for testing this!

Regarding your problem, I first started by booting current version of my driver on the Duo v2 w/o touching chip registers (only clock reference value). This way, I inherit the values installed by (NETGEAR's) u-boot:

Quoted 9 lines

$ for k in fan* pwm* ; do echo -n "$k:" ; echo `cat $k `; done
fan1_alarm:0
fan1_div:2
fan1_fault:0
fan1_input:1807
fan1_target:1807
pwm1:221
pwm1_enable:2        /* closed-loop, i.e. config is done via set_cnt */
pwm1_mode:1          /* PWM mode */ 

Sorry, I realize that I have not been very accurate in my description of the problem: The fan doesn't rotate when PWM _and_ open-loop control are selected. On the 2Big2 board, 2 wires are used to drive the fan.

Then with pwm1_mode=1, pwm1_enable=1 and set_cnt=0xff (default value), nothing happen whatever the value I write in pwm1.

By testing, I have discovered that writing 0 into set_cnt allows to work around the issue. I can do this by using the closed-loop control:
pwm1_mode=1, pwm1_enable=2 and pwm=255. Now, set_cnt worths 0 and if I switch back to open-loop control, all works as expected.

Can you try open-loop control and PWM mode with your board ? I wonder if this issue is specific to the 2Big2 hardware.

Hi Simon,

Quoted 44 lines · Simon Guinot

Quoted 33 lines · Arnaud Ebalard

Hi Simon and Guenter,

Quoted 14 lines · Simon Guinot

Quoted 10 lines · Arnaud Ebalard

Signed-off-by Arnaud Ebalard <[email protected]>

---
 drivers/hwmon/Kconfig              |   10 +
 drivers/hwmon/Makefile             |    1 +
 drivers/hwmon/g762.c               | 1012 ++++++++++++++++++++++++++++++++++++
 include/linux/platform_data/g762.h |   54 ++
 4 files changed, 1077 insertions(+)
 create mode 100644 drivers/hwmon/g762.c
 create mode 100644 include/linux/platform_data/g762.h

Hi Arnaud,

Quoted 8 lines

After more tests on my 2Big Network v2 board, it appears that the fan doesn't rotate when PWM mode (the preferred operating mode for this board) is selected. Nevertheless, DC mode is usable (even if not ideal given the hardware). After some investigations I noticed that an extra initialization is needed to enable PWM mode on my board: the set_cnt register must be set to 0 while the default value is 0xff. Is that specific to my hardware ? Is PWM mode working on your ReadyNAS with the default set_cnt value ?

First, thanks for testing this!

Regarding your problem, I first started by booting current version of my driver on the Duo v2 w/o touching chip registers (only clock reference value). This way, I inherit the values installed by (NETGEAR's) u-boot:

Quoted 9 lines

$ for k in fan* pwm* ; do echo -n "$k:" ; echo `cat $k `; done
fan1_alarm:0
fan1_div:2
fan1_fault:0
fan1_input:1807
fan1_target:1807
pwm1:221
pwm1_enable:2        /* closed-loop, i.e. config is done via set_cnt */
pwm1_mode:1          /* PWM mode */ 

Quoted 6 lines

Sorry, I realize that I have not been very accurate in my description of the problem: The fan doesn't rotate when PWM _and_ open-loop control are selected. On the 2Big2 board, 2 wires are used to drive the fan.

Then with pwm1_mode=1, pwm1_enable=1 and set_cnt=0xff (default value), nothing happen whatever the value I write in pwm1.

After boot, with set_cnt and set_out untouched by me (set_cnt set to 0x5a by u-boot, i.e. 255-0x5a read on pwm1):

fan1_alarm:0
fan1_div:1
fan1_fault:0
fan1_input:1365
fan1_target:1365
pwm1:0 /* set_out is 0 (considering fan1_target,

                           set_cnt is 0x5a)*/

pwm1_enable:2 /* closed-loop */
pwm1_mode:0 /* DC mode */

# echo 1 > pwm1_mode    /* PWM mode */       # fan rotates 
# echo 1 > pwm1_enable  /* open-loop */      # fan stops rotating

At that point we have:

fan1_alarm:0
fan1_div:1
fan1_fault:0
fan1_input:0
fan1_target:1365
pwm1:0 /* set_out is 0, set_cnt is still 0x5a */
pwm1_enable:1
pwm1_mode:1

Then:

# echo 100 > pwm1                            # fan rotates

Quoted 7 lines

By testing, I have discovered that writing 0 into set_cnt allows to work around the issue. I can do this by using the closed-loop control:
pwm1_mode=1, pwm1_enable=2 and pwm=255. Now, set_cnt worths 0 and if I switch back to open-loop control, all works as expected.

Quoted 2 lines

Can you try open-loop control and PWM mode with your board ? I wonder if this issue is specific to the 2Big2 hardware.

AFAICT, set_cnt has never been set to 0 in my test and PWM+open-loop works as expected, i.e. just by setting a non-zero value in set_out. Tell me if I missed something or if you want me to perform another test.

Cheers,

a+

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Quoted 82 lines · Arnaud Ebalard

Hi Simon,

Quoted 44 lines · Simon Guinot

Quoted 33 lines · Arnaud Ebalard

Hi Simon and Guenter,

Quoted 14 lines · Simon Guinot

Quoted 10 lines · Arnaud Ebalard

Signed-off-by Arnaud Ebalard <[email protected]>

---
 drivers/hwmon/Kconfig              |   10 +
 drivers/hwmon/Makefile             |    1 +
 drivers/hwmon/g762.c               | 1012 ++++++++++++++++++++++++++++++++++++
 include/linux/platform_data/g762.h |   54 ++
 4 files changed, 1077 insertions(+)
 create mode 100644 drivers/hwmon/g762.c
 create mode 100644 include/linux/platform_data/g762.h

Hi Arnaud,

Quoted 8 lines

After more tests on my 2Big Network v2 board, it appears that the fan doesn't rotate when PWM mode (the preferred operating mode for this board) is selected. Nevertheless, DC mode is usable (even if not ideal given the hardware). After some investigations I noticed that an extra initialization is needed to enable PWM mode on my board: the set_cnt register must be set to 0 while the default value is 0xff. Is that specific to my hardware ? Is PWM mode working on your ReadyNAS with the default set_cnt value ?

First, thanks for testing this!

Regarding your problem, I first started by booting current version of my driver on the Duo v2 w/o touching chip registers (only clock reference value). This way, I inherit the values installed by (NETGEAR's) u-boot:

Quoted 9 lines

$ for k in fan* pwm* ; do echo -n "$k:" ; echo `cat $k `; done
fan1_alarm:0
fan1_div:2
fan1_fault:0
fan1_input:1807
fan1_target:1807
pwm1:221
pwm1_enable:2        /* closed-loop, i.e. config is done via set_cnt */
pwm1_mode:1          /* PWM mode */ 

Quoted 6 lines

Sorry, I realize that I have not been very accurate in my description of the problem: The fan doesn't rotate when PWM _and_ open-loop control are selected. On the 2Big2 board, 2 wires are used to drive the fan.

Then with pwm1_mode=1, pwm1_enable=1 and set_cnt=0xff (default value), nothing happen whatever the value I write in pwm1.

After boot, with set_cnt and set_out untouched by me (set_cnt set to 0x5a by u-boot, i.e. 255-0x5a read on pwm1):

fan1_alarm:0
fan1_div:1
fan1_fault:0
fan1_input:1365
fan1_target:1365
pwm1:0 /* set_out is 0 (considering fan1_target,

                           set_cnt is 0x5a)*/

pwm1_enable:2 /* closed-loop */
pwm1_mode:0 /* DC mode */

# echo 1 > pwm1_mode    /* PWM mode */       # fan rotates 
# echo 1 > pwm1_enable  /* open-loop */      # fan stops rotating

At that point we have:

fan1_alarm:0
fan1_div:1
fan1_fault:0
fan1_input:0
fan1_target:1365
pwm1:0 /* set_out is 0, set_cnt is still 0x5a */
pwm1_enable:1
pwm1_mode:1

Quoted 3 lines

Then:

# echo 100 > pwm1                            # fan rotates

OK you are lucky. Your bootloader initialize set_cnt with 0x5a. Mine don't and it is precisely the issue: set_cnt is still 0xff (the default value) when the driver controls the fan.

Please, could you try open-loop mode with set_cnt set to 0xff ? Maybe you can enforce this value for the test purpose ?

If you observe the same behaviour than me, then could modify the driver to ensure that set_cnt is not 0xff when open-loop mode is selected ? Maybe by systematically setting a different value (as 0) ?

Quoted 7 lines

Quoted 7 lines

By testing, I have discovered that writing 0 into set_cnt allows to work around the issue. I can do this by using the closed-loop control:
pwm1_mode=1, pwm1_enable=2 and pwm=255. Now, set_cnt worths 0 and if I switch back to open-loop control, all works as expected.

Quoted 7 lines

Quoted 2 lines

Can you try open-loop control and PWM mode with your board ? I wonder if this issue is specific to the 2Big2 hardware.

AFAICT, set_cnt has never been set to 0 in my test and PWM+open-loop works as expected, i.e. just by setting a non-zero value in set_out. Tell me if I missed something or if you want me to perform another test.

Actually, I have only problems when set_cnt worths 0xff. Please, could you try the tests mentioned above ?

Thanks in advance,

Simon

Quoted 9 lines

Cheers,

a+

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel at lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

OK you are lucky. Your bootloader initialize set_cnt with 0x5a. Mine don't and it is precisely the issue: set_cnt is still 0xff (the default value) when the driver controls the fan.

Please, could you try open-loop mode with set_cnt set to 0xff ? Maybe you can enforce this value for the test purpose ?

If you observe the same behaviour than me, then could modify the driver to ensure that set_cnt is not 0xff when open-loop mode is selected ? Maybe by systematically setting a different value (as 0) ?

Regarding your problem, I first started by booting current version of my driver on the Duo v2 w/o touching chip registers (only clock reference value). This way, I inherit the values installed by (NETGEAR's) u-boot:

$ for k in fan* pwm* ; do echo -n "$k:" ; echo `cat $k `; done
fan1_alarm:0
fan1_div:2
fan1_fault:0
fan1_input:1807
fan1_target:1807
pwm1:221
pwm1_enable:2        /* closed-loop, i.e. config is done via set_cnt */
pwm1_mode:1          /* PWM mode */ 

If G762 datasheet is correct, NETGEAR's u-boot alters fan divisor ('01' i.e. 2 instead of '00' i.e. 1). It also changes the default mode from linear to pwm and sets set_cnt to 34 (255-221).

Then, I grabbed the GPL sources (5.3.7) of NETGEAR's u-boot and found the following in a boot function:

/* Turn G76x(FAN controller, i2c address 0x3e) on.
* The FAN_SET_CNT register's offset is 0x0.
* Set [1300(rpm) == 65% == 5a(FAN_SET_CNT)] as default. */
unsigned char byte=0x5a;
if(i2c_write(0, 0x3e, 0x0, 1, &byte, 1) != 0)
puts ("Error writing the i2c chip : G76x(Fan controller).\n");

AFAICT, this is the only location in the whole source tarball where 0x3e reg is modified in an i2c function.

Then, I grabbed the u-boot source code from NETGEAR for ReadyNAS 102 (Armada-370 based NAS, also using a G762) and it has the following:

/* Turn G76x(FAN controller, i2c address 0x3e) on.
* The FAN_SET_CNT register's offset is 0x0.
* Set [1300(rpm) == 65% == 5a(FAN_SET_CNT)] as default. */
MV_U8 byte=0x5a;
if(i2c_write(0x3e, 0x0, 1, &byte, 1) != 0)

(quoted diff) +0 -0

 	puts ("Error writing the i2c chip : G76x(Fan controller).\n");
 /* Tune the FAN_STARTV */
 if (i2c_read(0x3e, 0x5, 1, &byte, 1) != 0)
 	puts ("Error reading the i2c chip : G76x(Fan controller).\n");
 byte |= 0x3;
 if(i2c_write(0x3e, 0x5, 1, &byte, 1) != 0)
 	puts ("Error writing the i2c chip : G76x(Fan controller).\n");

In the end, it does not seem NETGEAR's u-boot does some specific things except setting a non-zero value to SET_CNT register for the fan to start rotating.

BTW, if you wonder why some attributes reported at the beginning have different values than the ones in the datasheet (0x5a set in u-boot, reading 0x22) although NETGEAR's u-boot code does not seem to tamper those, I don't have the answer.

Signed-off-by Arnaud Ebalard <[email protected]>

---
 drivers/hwmon/Kconfig              |   10 +
 drivers/hwmon/Makefile             |    1 +
 drivers/hwmon/g762.c               | 1012 ++++++++++++++++++++++++++++++++++++
 include/linux/platform_data/g762.h |   54 ++
 4 files changed, 1077 insertions(+)
 create mode 100644 drivers/hwmon/g762.c
 create mode 100644 include/linux/platform_data/g762.h

drivers/hwmon/Kconfig +10 -0

index 0428e8a..142bdf8 100644
@@ -456,6 +456,16 @@ config SENSORS_G760A
 	  This driver can also be built as a module.  If so, the module
 	  will be called g760a.
 
+config SENSORS_G762
+	tristate "GMT G762 and G763"
+	depends on I2C
+	help
+	  If you say yes here you get support for Global Mixed-mode
+	  Technology Inc G762 and G763 fan speed PWM controller chips.
+
+	  This driver can also be built as a module.  If so, the module
+	  will be called g762.
+
 config SENSORS_GL518SM
 	tristate "Genesys Logic GL518SM"
 	depends on I2C

drivers/hwmon/Makefile +1 -0

index d17d3e6..4f0fb52 100644
@@ -60,6 +60,7 @@ obj-$(CONFIG_SENSORS_F75375S)	+= f75375s.o
 obj-$(CONFIG_SENSORS_FAM15H_POWER) += fam15h_power.o
 obj-$(CONFIG_SENSORS_FSCHMD)	+= fschmd.o
 obj-$(CONFIG_SENSORS_G760A)	+= g760a.o
+obj-$(CONFIG_SENSORS_G762)	+= g762.o
 obj-$(CONFIG_SENSORS_GL518SM)	+= gl518sm.o
 obj-$(CONFIG_SENSORS_GL520SM)	+= gl520sm.o
 obj-$(CONFIG_SENSORS_GPIO_FAN)	+= gpio-fan.o

drivers/hwmon/g762.c new file +182 -0

new file mode 100644
index 0000000..d21ec24
@@ -0,0 +1,1012 @@
+/*
+ * g762 - Driver for the Global Mixed-mode Technology Inc. fan speed
+ *        PWM controller chips from G762 family, i.e. G762 and G763
+ *
+ * Copyright (C) 2013, Arnaud EBALARD <[email protected]>
+ *
+ * This work is based on a basic version for 2.6.31 kernel developed
+ * by Olivier Mouchet for LaCie. Updates and correction have been
+ * performed to run on recent kernels. Additional features, like the
+ * ability to configure various characteristics via .dts file, have
+ * been added. Detailed datasheet on which this development is based
+ * is available here:
+ *
+ *  http://natisbad.org/NAS/refs/GMT_EDS-762_763-080710-0.2.pdf
+ *
+ * Headers from previous developments have been kept below:
+ *
+ * Copyright (c) 2009 LaCie
+ *
+ * Author: Olivier Mouchet <[email protected]>
+ *
+ * based on g760a code written by Herbert Valerio Riedel <[email protected]>
+ * Copyright (C) 2007  Herbert Valerio Riedel <[email protected]>
+ *
+ * g762: minimal datasheet available at:
+ *       http://www.gmt.com.tw/product/datasheet/EDS-762_3.pdf
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation.
+ */
+
+#include <linux/device.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <linux/jiffies.h>
+#include <linux/i2c.h>
+#include <linux/hwmon.h>
+#include <linux/hwmon-sysfs.h>
+#include <linux/err.h>
+#include <linux/mutex.h>
+#include <linux/kernel.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_data/g762.h>
+
+#define DRVNAME "g762"
+
+static const struct i2c_device_id g762_id[] = {
+	{ "g762", 0 },
+	{ "g763", 0 },
+	{ }
+};
+MODULE_DEVICE_TABLE(i2c, g762_id);
+
+enum g762_regs {
+	G762_REG_SET_CNT  = 0x00,
+	G762_REG_ACT_CNT  = 0x01,
+	G762_REG_FAN_STA  = 0x02,
+	G762_REG_SET_OUT  = 0x03,
+	G762_REG_FAN_CMD1 = 0x04,
+	G762_REG_FAN_CMD2 = 0x05,
+};
+
+/* Config register bits */
+#define G762_REG_FAN_CMD1_DET_FAN_FAIL  0x80 /* enable fan_fail signal */
+#define G762_REG_FAN_CMD1_DET_FAN_OOC   0x40 /* enable fan_out_of_control */
+#define G762_REG_FAN_CMD1_OUT_MODE      0x20 /* out mode, PWM or DC */
+#define G762_REG_FAN_CMD1_FAN_MODE      0x10 /* fan mode: closed/open loop */
+#define G762_REG_FAN_CMD1_CLK_DIV_ID1   0x08 /* clock divisor value */
+#define G762_REG_FAN_CMD1_CLK_DIV_ID0   0x04
+#define G762_REG_FAN_CMD1_PWM_POLARITY  0x02 /* PWM polarity */
+#define G762_REG_FAN_CMD1_PULSE_PER_REV 0x01 /* pulse per fan revolution */
+
+#define G762_REG_FAN_CMD2_GEAR_MODE_1   0x08 /* fan gear mode */
+#define G762_REG_FAN_CMD2_GEAR_MODE_0   0x04
+#define G762_REG_FAN_CMD2_FAN_STARTV_1  0x02 /* fan startup voltage */
+#define G762_REG_FAN_CMD2_FAN_STARTV_0  0x01
+
+#define G762_REG_FAN_STA_FAIL           0x02 /* fan fail */
+#define G762_REG_FAN_STA_OOC            0x01 /* fan out of control */
+
+/* config register values */
+#define OUT_MODE_PWM            1
+#define OUT_MODE_DC             0
+
+#define FAN_MODE_CLOSED_LOOP    2
+#define FAN_MODE_OPEN_LOOP      1
+
+/* register data is read (and cached) at most once per second */
+#define G762_UPDATE_INTERVAL    HZ
+
+/*
+ * extract pulse count per fan revolution value (2 or 4) from given
+ * FAN_CMD1 register value
+ */
+#define PULSE_FROM_REG(reg) \
+	((((reg) & G762_REG_FAN_CMD1_PULSE_PER_REV)+1) << 1)
+
+/*
+ * extract fan clock divisor (1, 2, 4 or 8) from given FAN_CMD1
+ * register value
+ */
+#define CLKDIV_FROM_REG(reg) \
+	(1 << (((reg) & (G762_REG_FAN_CMD1_CLK_DIV_ID0 |	\
+			 G762_REG_FAN_CMD1_CLK_DIV_ID1)) >> 2))
+
+/*
+ * extract fan gear mode multiplier value (0, 2 or 4) from given
+ * FAN_CMD2 register value
+ */
+#define GEARMULT_FROM_REG(reg) \
+	(1 << (((reg) & (G762_REG_FAN_CMD2_GEAR_MODE_0 |	\
+			 G762_REG_FAN_CMD2_GEAR_MODE_1)) >> 2))
+
+struct g762_data {
+	struct i2c_client *client;
+	struct device *hwmon_dev;
+
+	/* update mutex */
+	struct mutex update_lock;
+
+	/* board specific parameters. */
+	u32 clk;
+
+	/* g762 register cache */
+	bool valid;
+	unsigned long last_updated; /* in jiffies */
+
+	u8 set_cnt;  /* controls fan rotation speed in closed-loop mode */
+	u8 act_cnt;  /* provides access to current fan RPM value */
+	u8 fan_sta;  /* bit 0: set when actual fan speed is more than
+		      *        25% outside requested fan speed
+		      * bit 1: set when no transition occurs on fan
+		      *        pin for 0.7s
+		      */
+	u8 set_out;  /* controls fan rotation speed in open-loop mode */
+	u8 fan_cmd1; /*   0: FG_PLS_ID0 FG pulses count per revolution
+		      *      0: 2 counts per revolution
+		      *      1: 4 counts per revolution
+		      *   1: PWM_POLARITY 1: negative_duty
+		      *                   0: positive_duty
+		      * 2,3: [FG_CLOCK_ID0, FG_CLK_ID1]
+		      *         00: Divide fan clock by 1
+		      *         01: Divide fan clock by 2
+		      *         10: Divide fan clock by 4
+		      *         11: Divide fan clock by 8
+		      *   4: FAN_MODE 1:closed-loop, 0:open-loop
+		      *   5: OUT_MODE 1:PWM, 0:DC
+		      *   6: DET_FAN_OOC enable "fan ooc" status
+		      *   7: DET_FAN_FAIL enable "fan fail" status
+		      */
+	u8 fan_cmd2; /* 0,1: FAN_STARTV 0,1,2,3 -> 0,32,64,96 dac_code
+		      * 2,3: FG_GEAR_MODE
+		      *         00: multiplier = 1
+		      *         01: multiplier = 2
+		      *         10: multiplier = 4
+		      *   4: Mask ALERT# (g763 only)
+		      */
+};
+
+/*
+ * sysfs PWM interface uses value from 0 to 255 when g762 FAN_SET_CNT register
+ * uses values from 255 (off) to 0 (full speed). Note that FAN_SET_OUT register
+ * uses values from 0 (off) to 255 (full speed), i.e. does not need translation.
+ */
+#define PWM_FROM_CNT(cnt)       (0xff - (cnt))
+#define PWM_TO_CNT(pwm)         (0xff - (pwm))
+
+/*
+ * Convert count value from fan controller registers (FAN_SET_CNT/FAN_ACT_CNT)
+ * into fan speed RPM value. Note that the datasheet documents a basic formula.

drivers/hwmon/g762.c new file +21 -0

+}
+
+/* Convert fan RPM value from sysfs into count value for fan controller register
+ * (FAN_SET_CNT). */
+static inline unsigned char cnt_from_rpm(u32 rpm, u32 clk, u16 p,
+					 u8 clk_div, u8 gear_mult)
+{
+	if (!rpm)         /* to stop the fan, set cnt to 255 */
+		return 0xff;
+
+	return clamp_val(((clk*30*gear_mult)/(rpm*p*clk_div)), 0, 255);
+}
+
+/* helper to grab and cache data, at most one time per second */
+static struct g762_data *g762_update_client(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = i2c_get_clientdata(client);
+
+	mutex_lock(&data->update_lock);
+	if (time_before(jiffies, data->last_updated + G762_UPDATE_INTERVAL) &&

drivers/hwmon/g762.c new file +246 -0

+	data->last_updated = jiffies;
+	data->valid = true;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return data;
+}
+
+/*
+ * helpers for passing hardware characteristics via DT. Some of those
+ * are also used by sysfs handlers (write functions) later in the file.
+ */
+
+/* Set pwm mode. Accepts either 0 (PWM mode) or 1 (DC mode) */
+static int do_set_pwm_mode(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (val) {
+	case OUT_MODE_PWM:
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_OUT_MODE;
+		break;
+	case OUT_MODE_DC:
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_OUT_MODE;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD1,
+					data->fan_cmd1);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/*
+ * Set reference clock. Accepted values are between 0 and 0xffffff.
+ * Note that this is a characteristics of the system but an internal
+ * parameter, i.e. value is not passed to the device.
+ */
+static int do_set_pwm_freq(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = i2c_get_clientdata(client);
+
+	if (val > 0xffffff)
+		return -EINVAL;
+
+	data->clk = val;
+
+	return 0;
+}
+
+/* Set fan clock divisor. Accepted values are 1, 2, 4 and 8. */
+static int do_set_fan_div(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (val) {
+	case 1:
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_CLK_DIV_ID0;
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_CLK_DIV_ID1;
+		break;
+	case 2:
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_CLK_DIV_ID0;
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_CLK_DIV_ID1;
+		break;
+	case 4:
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_CLK_DIV_ID0;
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_CLK_DIV_ID1;
+		break;
+	case 8:
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_CLK_DIV_ID0;
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_CLK_DIV_ID1;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD1,
+					data->fan_cmd1);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Set fan gear mode. Accepted values are either 0, 1 or 2. */
+static int do_set_fan_gear_mode(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (val) {
+	case 0:
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_GEAR_MODE_0;
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_GEAR_MODE_1;
+		break;
+	case 1:
+		data->fan_cmd2 |=  G762_REG_FAN_CMD2_GEAR_MODE_0;
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_GEAR_MODE_1;
+		break;
+	case 2:
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_GEAR_MODE_0;
+		data->fan_cmd2 |=  G762_REG_FAN_CMD2_GEAR_MODE_1;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD2,
+					data->fan_cmd2);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Set pulse per revolution value. Accepts either 2 or 4. */
+static int do_set_fan_pulses(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (val) {
+	case 2:
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_PULSE_PER_REV;
+		break;
+	case 4:
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_PULSE_PER_REV;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD1,
+					data->fan_cmd1);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Set fan mode. Accepts either 1 (open-loop) or 2 (closed-loop). */
+static int do_set_pwm_enable(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (val) {
+	case FAN_MODE_OPEN_LOOP:
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_FAN_MODE;
+		break;
+	case FAN_MODE_CLOSED_LOOP:
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_FAN_MODE;
+		break;
+	default:
+		goto out;
+	}
+
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD1,
+					data->fan_cmd1);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Set PWM polarity (0 for negative duty, 1 for positive duty) */
+static int do_set_pwm_polarity(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (val) {
+	case 0: /* i.e. negative duty */
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_PWM_POLARITY;
+		break;
+	case 1: /* i.e. positive duty */
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_PWM_POLARITY;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD1,
+					data->fan_cmd1);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/*
+ * Set pwm value. Accepted values are between 0 (stops the fan) and 255 (full
+ * speed). Note that the chip register used for setting the value depends on
+ * current fan mode of the device (closed or open-loop).
+ */
+static int do_set_pwm(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret;
+	u8 reg;
+
+	if (val > 255)
+		return -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	if (data->fan_cmd1 & G762_REG_FAN_CMD1_FAN_MODE) { /* closed-loop */
+		val = PWM_TO_CNT(val);
+		data->set_cnt = val;
+		reg = G762_REG_SET_CNT;
+	} else {                                           /* open-loop */
+		data->set_out = val;
+		reg = G762_REG_SET_OUT;
+	}
+	ret = i2c_smbus_write_byte_data(client, reg, (u8)val);
+	data->valid = false;
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Set fan RPM value. This only makes sense in closed-loop mode. */
+static int do_set_fan_target(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);

drivers/hwmon/g762.c new file +186 -0

+		data->valid = false;
+	}
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Enable/disable fan failure detection. Accepted values are 1 and 0. */
+static int do_fan_failure_detection_toggle(struct device *dev,
+					   unsigned long enable)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (enable) {
+	case 0:
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_DET_FAN_FAIL;
+		break;
+	case 1:
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_DET_FAN_FAIL;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD1,
+					data->fan_cmd1);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Enable/disable fan out of control detection. Accepted values are 1 and 0 */
+static int do_fan_ooc_detection_toggle(struct device *dev, unsigned int enable)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (enable) {
+	case 0:
+		data->fan_cmd1 &= ~G762_REG_FAN_CMD1_DET_FAN_OOC;
+		break;
+	case 1:
+		data->fan_cmd1 |=  G762_REG_FAN_CMD1_DET_FAN_OOC;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD1,
+					data->fan_cmd1);
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/* Set fan startup voltage. Accepted values are either 0, 1, 2 or 3. */
+static int do_set_fan_startv(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	switch (val) {
+	case 0:
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_FAN_STARTV_0;
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_FAN_STARTV_1;
+		data->valid = false;
+		break;
+	case 1:
+		data->fan_cmd2 |=  G762_REG_FAN_CMD2_FAN_STARTV_0;
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_FAN_STARTV_1;
+		data->valid = false;
+		break;
+	case 2:
+		data->fan_cmd2 &= ~G762_REG_FAN_CMD2_FAN_STARTV_0;
+		data->fan_cmd2 |=  G762_REG_FAN_CMD2_FAN_STARTV_1;
+		data->valid = false;
+		break;
+	case 3:
+		data->fan_cmd2 |=  G762_REG_FAN_CMD2_FAN_STARTV_0;
+		data->fan_cmd2 |=  G762_REG_FAN_CMD2_FAN_STARTV_1;
+		data->valid = false;
+		break;
+	default:
+		goto out;
+	}
+	ret = i2c_smbus_write_byte_data(client, G762_REG_FAN_CMD2,
+					data->fan_cmd2);
+	data->valid = false;
+ out:
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+/*
+ * Configuration-related definitions
+ */
+
+static inline void g762_platform_data_init(struct g762_platform_data *pdata)
+{
+	*pdata = G762_DEFAULT_PDATA;
+}
+
+static inline int g762_one_prop_commit(struct i2c_client *client,
+				       u32 pval, const char *pname,
+				       int (*psetter)(struct device *dev,
+						      unsigned long val))
+{
+	int ret = 0;
+
+	if ((pval != G762_DEFAULT_NO_OVERLOAD) &&
+	    (*psetter)(&client->dev, pval)) {
+		dev_err(&client->dev, "unable to set %s (%d)\n", pname, pval);
+		ret = -EINVAL;
+	}
+
+	return ret;
+}
+
+static int g762_platform_data_commit(struct i2c_client *client,
+				     struct g762_platform_data *pdata)
+{
+	int ret = 0;
+
+	if (g762_one_prop_commit(client, pdata->fan_startv,
+				 "fan_startv", do_set_fan_startv) ||
+	    g762_one_prop_commit(client, pdata->fan_div,
+				 "fan_div", do_set_fan_div) ||
+	    g762_one_prop_commit(client, pdata->fan_pulses,
+				 "fan_pulses", do_set_fan_pulses) ||
+	    g762_one_prop_commit(client, pdata->pwm_freq,
+				 "pwm_freq", do_set_pwm_freq) ||
+	    g762_one_prop_commit(client, pdata->fan_gear_mode,
+				 "fan_gear_mode", do_set_fan_gear_mode) ||
+	    g762_one_prop_commit(client, pdata->pwm_polarity,
+				 "pwm_polarity", do_set_pwm_polarity) ||
+	    g762_one_prop_commit(client, pdata->pwm_mode,
+				 "pwm_mode", do_set_pwm_mode) ||
+	    g762_one_prop_commit(client, pdata->pwm_enable,
+				 "pwm_enable", do_set_pwm_enable) ||
+	    g762_one_prop_commit(client, pdata->fan_target,
+				 "fan_target", do_set_fan_target)) {
+		ret = -EINVAL;
+	}
+
+	return ret;
+}
+
+/*
+ * Helpers to import hardware characteristics from .dts file and overload
+ * default platform data values.
+ */
+
+#ifdef CONFIG_OF
+static struct of_device_id g762_dt_match[] = {
+	{ .compatible = "gmt,g762" },
+	{ .compatible = "gmt,g763" },
+	{ },
+};
+
+static inline void g762_of_import_one_prop(struct i2c_client *client,
+					   u32 *dest, const char *pname)
+{
+	const __be32 *prop;
+	int len;
+
+	prop = of_get_property(client->dev.of_node, pname, &len);
+	if (prop && len == sizeof(u32)) {
+		*dest = be32_to_cpu(prop[0]);
+		dev_dbg(&client->dev, "found %s (%d)\n", pname, *dest);
+	}
+}
+
+static void g762_platform_data_of_overload(struct i2c_client *client,
+					   struct g762_platform_data *pdata)
+{
+	if (!client->dev.of_node)
+		return;

drivers/hwmon/g762.c new file +139 -0

+}
+#else
+static void g762_platform_data_of_overload(struct i2c_client *client,
+				    struct g762_platform_data *pdata) { }
+#endif
+
+/*
+ * helper to overload driver parameters from board init file for those
+ * not already converted to device tree
+ */
+static void g762_platform_data_overload(struct i2c_client *client,
+					struct g762_platform_data *pdata)
+{
+	struct g762_platform_data *opdata = client->dev.platform_data;
+
+	if (opdata)
+		*pdata = *opdata;
+}
+
+/*
+ * sysfs attributes
+ */
+
+/*
+ * Read function for fan1_input sysfs file. Return current fan RPM value, or
+ * 0 if fan is out of control. The function is available in both closed and
+ * open-loop mode.
+ */
+static ssize_t get_fan_rpm(struct device *dev, struct device_attribute *da,
+			   char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+	unsigned int rpm = 0;
+
+	mutex_lock(&data->update_lock);
+	/* reverse logic: fan out of control reporting is enabled low */
+	if (data->fan_sta & G762_REG_FAN_STA_OOC) {
+		rpm = rpm_from_cnt(data->act_cnt, data->clk,
+				   PULSE_FROM_REG(data->fan_cmd1),
+				   CLKDIV_FROM_REG(data->fan_cmd1),
+				   GEARMULT_FROM_REG(data->fan_cmd2));
+	}
+	mutex_unlock(&data->update_lock);
+
+	return sprintf(buf, "%u\n", rpm);
+}
+
+/*
+ * Read and write functions for pwm1_mode sysfs file. Get and set fan speed
+ * control mode i.e. PWM (1) or DC (0).
+ */
+static ssize_t get_pwm_mode(struct device *dev, struct device_attribute *da,
+			    char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+
+	return sprintf(buf, "%d\n",
+		       !!(data->fan_cmd1 & G762_REG_FAN_CMD1_OUT_MODE));
+}
+
+static ssize_t set_pwm_mode(struct device *dev, struct device_attribute *da,
+			    const char *buf, size_t count)
+{
+	unsigned long val;
+
+	if (kstrtoul(buf, 10, &val) || do_set_pwm_mode(dev, val))
+		return -EINVAL;
+
+	return count;
+}
+
+/*
+ * Read and write functions for fan1_div sysfs file. Get and set fan
+ * controller prescaler value
+ */
+static ssize_t get_fan_div(struct device *dev,
+			       struct device_attribute *da, char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+
+	return sprintf(buf, "%d\n", CLKDIV_FROM_REG(data->fan_cmd1));
+}
+
+static ssize_t set_fan_div(struct device *dev,
+			       struct device_attribute *da,
+			       const char *buf, size_t count)
+{
+	unsigned long val;
+
+	if (kstrtoul(buf, 10, &val) || do_set_fan_div(dev, val))
+		return -EINVAL;
+
+	return count;
+}
+
+/*
+ * Read and write functions for pwm1_enable. Get and set the fan speed control
+ * mode (i.e. closed or open-loop).
+ *
+ * Following documentation about hwmon's sysfs interface, a pwm1_enable node
+ * should accept followings:
+ *
+ *  0 : no fan speed control (i.e. fan at full speed)
+ *  1 : manual fan speed control enabled (use pwm[1-*]) (open-loop)
+ *  2+: automatic fan speed control enabled (use fan[1-*]_target) (closed-loop)
+ *
+ * but we do not accept 0 as "no-control" mode is not supported by g762,
+ * -EINVAL is returned in this case.
+ */
+static ssize_t get_pwm_enable(struct device *dev,
+			      struct device_attribute *da, char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+
+	return sprintf(buf, "%d\n",
+		       (!!(data->fan_cmd1 & G762_REG_FAN_CMD1_FAN_MODE)) + 1);
+}
+
+static ssize_t set_pwm_enable(struct device *dev,
+			      struct device_attribute *da,
+			      const char *buf, size_t count)
+{
+	unsigned long val;
+
+	if (kstrtoul(buf, 10, &val) || do_set_pwm_enable(dev, val))
+		return -EINVAL;
+
+	return count;
+}
+
+/*
+ * Read and write functions for pwm1 sysfs file. Get and set the fan speed
+ * in both open and closed-loop mode. Speed is given as a relative value
+ * between 0 and 255, where 255 is maximum speed. Note that due to rounding
+ * errors it is possible that you don't read back exactly the value you have
+ * set.
+ */
+static ssize_t get_pwm(struct device *dev, struct device_attribute *da,
+		       char *buf)

drivers/hwmon/g762.c new file +188 -0

+	mutex_unlock(&data->update_lock);
+
+	return sprintf(buf, "%d\n", val);
+}
+
+static ssize_t set_pwm(struct device *dev, struct device_attribute *da,
+		       const char *buf, size_t count)
+{
+	unsigned long val;
+
+	if (kstrtoul(buf, 10, &val) || do_set_pwm(dev, val))
+		return -EINVAL;
+
+	return count;
+}
+
+/*
+ * Read and write function for fan1_target sysfs file. Get/set the fan speed in
+ * closed-loop mode. Speed is given as a RPM value; then the chip will regulate
+ * the fan speed using pulses from fan tachometer. The functions both return
+ * -EINVAL when called in open-loop mode.
+ *
+ * Refer to rpm_from_cnt() implementation above to get info about count number
+ * calculation.
+ *
+ * Also note that due to rounding errors it is possible that you don't read
+ * back exactly the value you have set.
+ */
+static ssize_t get_fan_target(struct device *dev, struct device_attribute *da,
+			      char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+	unsigned int rpm;
+	ssize_t ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	if (data->fan_cmd1 & G762_REG_FAN_CMD1_FAN_MODE) { /* closed-loop */
+		rpm = rpm_from_cnt(data->set_cnt, data->clk,
+				   PULSE_FROM_REG(data->fan_cmd1),
+				   CLKDIV_FROM_REG(data->fan_cmd1),
+				   GEARMULT_FROM_REG(data->fan_cmd2));
+		ret = sprintf(buf, "%u\n", rpm);
+	}
+	mutex_unlock(&data->update_lock);
+
+	return ret;
+}
+
+static ssize_t set_fan_target(struct device *dev, struct device_attribute *da,
+			      const char *buf, size_t count)
+{
+	unsigned long val;
+
+	if (kstrtoul(buf, 10, &val) || do_set_fan_target(dev, val))
+		return -EINVAL;
+
+	return count;
+}
+
+/* read function for fan1_fault sysfs file. */
+static ssize_t get_fan_failure(struct device *dev, struct device_attribute *da,
+			       char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+
+	return sprintf(buf, "%u\n", !!(data->fan_sta & G762_REG_FAN_STA_FAIL));
+}
+
+/*
+ * read function for fan1_alarm sysfs file. Note that OOC confition is
+ * enabled low
+ */
+static ssize_t get_fan_ooc(struct device *dev, struct device_attribute *da,
+			   char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+
+	return sprintf(buf, "%u\n", !(data->fan_sta & G762_REG_FAN_STA_OOC));
+}
+
+static DEVICE_ATTR(pwm1, S_IWUSR | S_IRUGO, get_pwm, set_pwm);
+static DEVICE_ATTR(pwm1_mode, S_IWUSR | S_IRUGO, get_pwm_mode, set_pwm_mode);
+static DEVICE_ATTR(pwm1_enable, S_IWUSR | S_IRUGO,
+		   get_pwm_enable, set_pwm_enable);
+static DEVICE_ATTR(fan1_input, S_IRUGO, get_fan_rpm, NULL);
+static DEVICE_ATTR(fan1_alarm, S_IRUGO, get_fan_ooc, NULL);
+static DEVICE_ATTR(fan1_fault, S_IRUGO, get_fan_failure, NULL);
+static DEVICE_ATTR(fan1_target, S_IWUSR | S_IRUGO,
+		   get_fan_target, set_fan_target);
+static DEVICE_ATTR(fan1_div, S_IWUSR | S_IRUGO, get_fan_div, set_fan_div);
+
+/* Driver data */
+static struct attribute *g762_attributes[] = {
+	&dev_attr_fan1_input.attr,
+	&dev_attr_fan1_alarm.attr,
+	&dev_attr_fan1_fault.attr,
+	&dev_attr_fan1_target.attr,
+	&dev_attr_fan1_div.attr,
+	&dev_attr_pwm1.attr,
+	&dev_attr_pwm1_mode.attr,
+	&dev_attr_pwm1_enable.attr,
+	NULL
+};
+
+static const struct attribute_group g762_group = {
+	.attrs = g762_attributes,
+};
+
+static int g762_probe(struct i2c_client *client, const struct i2c_device_id *id)
+{
+	struct g762_data *data;
+	struct g762_platform_data pdata;
+	int err;
+
+	if (!i2c_check_functionality(client->adapter,
+				     I2C_FUNC_SMBUS_BYTE_DATA))
+		return -ENODEV;
+
+	data = devm_kzalloc(&client->dev, sizeof(struct g762_data), GFP_KERNEL);
+	if (!data)
+		return -ENOMEM;
+
+	i2c_set_clientdata(client, data);
+
+	data->client = client;
+	mutex_init(&data->update_lock);
+
+	/* Enable fan protection and fan fail detection by default */
+	do_fan_ooc_detection_toggle(&client->dev, 1);
+	do_fan_failure_detection_toggle(&client->dev, 1);
+
+	/*
+	 * Set default configuration values before passing the structure
+	 * first to overload routine used by boards non converted to DT and
+	 * then to OF helper (to overload them using those provided by .dts
+	 * file, if any). Final result is then commited.
+	 */
+	g762_platform_data_init(&pdata);
+	g762_platform_data_overload(client, &pdata);
+	g762_platform_data_of_overload(client, &pdata);
+	err = g762_platform_data_commit(client, &pdata);
+	if (err)
+		return err;
+
+	/* Register sysfs hooks */
+	err = sysfs_create_group(&client->dev.kobj, &g762_group);
+	if (err)
+		return err;
+
+	data->hwmon_dev = (struct device *)hwmon_device_register(&client->dev);
+	if (IS_ERR(data->hwmon_dev)) {
+		err = PTR_ERR(data->hwmon_dev);
+		goto err_out;
+	}
+
+	return 0;
+
+ err_out:
+	sysfs_remove_group(&client->dev.kobj, &g762_group);
+	return err;
+}
+
+static int g762_remove(struct i2c_client *client)
+{
+	struct g762_data *data = i2c_get_clientdata(client);
+
+	hwmon_device_unregister(data->hwmon_dev);
+	sysfs_remove_group(&client->dev.kobj, &g762_group);
+
+	return 0;
+}
+
+static struct i2c_driver g762_driver = {
+	.driver = {
+		.name = DRVNAME,
+		.owner = THIS_MODULE,
+		.of_match_table = of_match_ptr(g762_dt_match),
+	},
+	.probe	  = g762_probe,
+	.remove	  = g762_remove,
+	.id_table = g762_id,
+};
+
+module_i2c_driver(g762_driver);
+
+MODULE_AUTHOR("Arnaud EBALARD <[email protected]>");
+MODULE_DESCRIPTION("GMT G762/G763 driver");
+MODULE_LICENSE("GPL");

include/linux/platform_data/g762.h new file +31 -0

new file mode 100644
index 0000000..29e3934
@@ -0,0 +1,54 @@
+/*
+ * Platform data structure for g762 fan controller driver
+ *
+ * Copyright (C) 2013, Arnaud EBALARD <[email protected]>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ */
+#ifndef __LINUX_PLATFORM_DATA_G762_H__
+#define __LINUX_PLATFORM_DATA_G762_H__
+
+#define G762_DEFAULT_NO_OVERLOAD 0xffffffff
+
+struct g762_platform_data {
+	u32 fan_startv;
+	u32 fan_gear_mode;
+	u32 fan_div;
+	u32 fan_pulses;
+	u32 fan_target;
+	u32 pwm_polarity;

include/linux/platform_data/g762.h new file +1 -0

+#endif /* __LINUX_PLATFORM_DATA_G762_H__ */

Signature · 1 line

1.7.10.4

drivers/hwmon/g762.c line 192

Quoted 1 line

drivers/hwmon/g762.c new file +1 -0

+	return (clk*30*gear_mult)/((cnt ? cnt : 1)*p*clk_div);

Please follow CodingStyle and add spaces between operators (not just here).

Quoted 12 lines

drivers/hwmon/g762.c new file +12 -0

+static struct g762_data *g762_update_client(struct device *dev)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = i2c_get_clientdata(client);
+
+	mutex_lock(&data->update_lock);
+	if (time_before(jiffies, data->last_updated + G762_UPDATE_INTERVAL) &&
+	    likely(data->valid))
+		goto out;
+
+	data->set_cnt =  i2c_smbus_read_byte_data(client, G762_REG_SET_CNT);
+	data->act_cnt =  i2c_smbus_read_byte_data(client, G762_REG_ACT_CNT);

drivers/hwmon/g762.c line 223

Quoted 5 lines

drivers/hwmon/g762.c new file +5 -0

+	data->fan_sta =  i2c_smbus_read_byte_data(client, G762_REG_FAN_STA);
+	data->set_out =  i2c_smbus_read_byte_data(client, G762_REG_SET_OUT);
+	data->fan_cmd1 = i2c_smbus_read_byte_data(client, G762_REG_FAN_CMD1);
+	data->fan_cmd2 = i2c_smbus_read_byte_data(client, G762_REG_FAN_CMD2);
+

Failed i2c reads (returning negative values) not a problem ? Driver operation will be more or less random in that case, and may result in erratic behavior.

Quoted 11 lines

drivers/hwmon/g762.c new file +11 -0

+/* Set fan RPM value. This only makes sense in closed-loop mode. */
+static int do_set_fan_target(struct device *dev, unsigned long val)
+{
+	struct i2c_client *client = to_i2c_client(dev);
+	struct g762_data *data = g762_update_client(dev);
+	int ret = -EINVAL;
+
+	mutex_lock(&data->update_lock);
+	if (data->fan_cmd1 & G762_REG_FAN_CMD1_FAN_MODE) { /* closed-loop */
+		data->set_cnt = cnt_from_rpm(val, data->clk,
+					     PULSE_FROM_REG(data->fan_cmd1),

drivers/hwmon/g762.c line 479

Quoted 4 lines

drivers/hwmon/g762.c new file +4 -0

+					     CLKDIV_FROM_REG(data->fan_cmd1),
+					     GEARMULT_FROM_REG(data->fan_cmd2));
+		ret = i2c_smbus_write_byte_data(client, G762_REG_SET_CNT,
+						data->set_cnt);

In closed loop mode you have two paths to set the same register. Given that, it doesn't make sense to me to set the fan target with set_pwm as well.

If you use one attribute (set_pwm) for setting set_out, and the other for set_cnt, your problems with conditional command acceptance should be solved.

Quoted 10 lines

drivers/hwmon/g762.c new file +10 -0

+static void g762_platform_data_of_overload(struct i2c_client *client,
+					   struct g762_platform_data *pdata)
+{
+	if (!client->dev.of_node)
+		return;
+
+	g762_of_import_one_prop(client, &pdata->fan_gear_mode, "fan_gear_mode");
+	g762_of_import_one_prop(client, &pdata->pwm_polarity, "pwm_polarity");
+	g762_of_import_one_prop(client, &pdata->fan_startv, "fan_startv");
+	g762_of_import_one_prop(client, &pdata->pwm_freq, "pwm_freq");

drivers/hwmon/g762.c line 675

Quoted 5 lines

drivers/hwmon/g762.c new file +5 -0

+	g762_of_import_one_prop(client, &pdata->fan_div, "fan_div");
+	g762_of_import_one_prop(client, &pdata->fan_pulses, "fan_pulses");
+	g762_of_import_one_prop(client, &pdata->pwm_mode, "pwm_mode");
+	g762_of_import_one_prop(client, &pdata->fan_target, "fan_target");
+	g762_of_import_one_prop(client, &pdata->pwm_enable, "pwm_enable");

I second the other comments ... why not use the sysfs interface to set the values it supports ?

Quoted 14 lines

drivers/hwmon/g762.c new file +14 -0

+/*
+ * Read and write functions for pwm1 sysfs file. Get and set the fan speed
+ * in both open and closed-loop mode. Speed is given as a relative value
+ * between 0 and 255, where 255 is maximum speed. Note that due to rounding
+ * errors it is possible that you don't read back exactly the value you have
+ * set.
+ */
+static ssize_t get_pwm(struct device *dev, struct device_attribute *da,
+		       char *buf)
+{
+	struct g762_data *data = g762_update_client(dev);
+	int val;
+
+	mutex_lock(&data->update_lock);

drivers/hwmon/g762.c line 824

Quoted 5 lines

drivers/hwmon/g762.c new file +5 -0

+	if (data->fan_cmd1 & G762_REG_FAN_CMD1_FAN_MODE) { /* closed-loop */
+		val = PWM_FROM_CNT(data->set_cnt);
+	} else {                                           /* open-loop */
+		val = data->set_out;
+	}

Again, this is not very consistent, as set_cnt is pretty much reported twice (as pwm and as target speed) in closed loop mode. Just report set_out here.

Side note: { } is not needed in single-statement conditionals.

include/linux/platform_data/g762.h line 41

Quoted 5 lines

include/linux/platform_data/g762.h new file +5 -0

+/* When filling a g762_platform_data structure to be passed during platform
+ * init to the driver, it needs to be initialized to the following default
+ * value before changing specific fields. This is needed to avoid a sparse
+ * initialization to have current values replaced by 0 */
+

comment style

Quoted 10 lines

include/linux/platform_data/g762.h new file +10 -0

+static const struct g762_platform_data G762_DEFAULT_PDATA = {
+	.fan_startv = G762_DEFAULT_NO_OVERLOAD,
+	.fan_gear_mode = G762_DEFAULT_NO_OVERLOAD,
+	.fan_div = G762_DEFAULT_NO_OVERLOAD,
+	.fan_pulses = G762_DEFAULT_NO_OVERLOAD,
+	.fan_target = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_polarity = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_enable = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_freq = G762_DEFAULT_NO_OVERLOAD,
+	.pwm_mode = G762_DEFAULT_NO_OVERLOAD,

Review

Quoted 2 lines

(quoted diff) +2 -0

+};
+

A constant array in an include file ? Please not; this is completely unacceptable. You can pass platform data to the driver like every other driver does. The entire approach with G762_DEFAULT_NO_OVERLOAD is odd anyway; it would be great if you could come up with another mechanism to pass platform data. It is quite common to use '0' for 'default' to avoid having to initialize all default fields.

This series adds support for GMT G762/G763. This work is based on a basic version for 2.6.31 kernel developed Olivier Mouchet for LaCie NAS. Updates have been performed to run on recent kernels. Support has been completed and additional features added: ability to configure various characteristics from .dts file, better initialization, alarms and error reporting support, gear mode, polarity, fan pulse per revolution, fan startup voltage control.

Quoted 8 lines · On Tuesday 28 May 2013 00:02:29 Arnaud Ebalard wrote:

This series adds support for GMT G762/G763. This work is based on a basic version for 2.6.31 kernel developed Olivier Mouchet for LaCie NAS. Updates have been performed to run on recent kernels. Support has been completed and additional features added: ability to configure various characteristics from .dts file, better initialization, alarms and error reporting support, gear mode, polarity, fan pulse per revolution, fan startup voltage control.

I wonder if this could be split into two separate drivers, one for the pwm subsystem, and one for a hardware-independent fan controller based on the pwm interfaces.

Hi Arnd,

Quoted 10 lines · Arnd Bergmann

Quoted 8 lines · On Tuesday 28 May 2013 00:02:29 Arnaud Ebalard wrote:

This series adds support for GMT G762/G763. This work is based on a basic version for 2.6.31 kernel developed Olivier Mouchet for LaCie NAS. Updates have been performed to run on recent kernels. Support has been completed and additional features added: ability to configure various characteristics from .dts file, better initialization, alarms and error reporting support, gear mode, polarity, fan pulse per revolution, fan startup voltage control.

Quoted 3 lines

I wonder if this could be split into two separate drivers, one for the pwm subsystem, and one for a hardware-independent fan controller based on the pwm interfaces.

To be honest, I wouldn't even know how to start in order to do that. Additionally, it would be worth the effort if other drivers could be easily refactored using this idea. I'll let hwmon people provide some feedback and try and comply with their directions (based on the cpu cycle I can spend on this).

Quoted 18 lines · Arnaud Ebalard

Hi Arnd,

Quoted 10 lines · Arnd Bergmann

Quoted 8 lines · On Tuesday 28 May 2013 00:02:29 Arnaud Ebalard wrote:

This series adds support for GMT G762/G763. This work is based on a basic version for 2.6.31 kernel developed Olivier Mouchet for LaCie NAS. Updates have been performed to run on recent kernels. Support has been completed and additional features added: ability to configure various characteristics from .dts file, better initialization, alarms and error reporting support, gear mode, polarity, fan pulse per revolution, fan startup voltage control.

Quoted 3 lines

I wonder if this could be split into two separate drivers, one for the pwm subsystem, and one for a hardware-independent fan controller based on the pwm interfaces.

Quoted 5 lines

To be honest, I wouldn't even know how to start in order to do that. Additionally, it would be worth the effort if other drivers could be easily refactored using this idea. I'll let hwmon people provide some feedback and try and comply with their directions (based on the cpu cycle I can spend on this).

What Arnd proposes doesn't sounds like a very good idea. PWM-controlled fans should be able to work just fine using a generic driver that uses the PWM framework, similar to the PWM backlight and LED drivers.

Quoted 25 lines · Thierry Reding

Quoted 23 lines · Arnaud Ebalard

Hi Arnd,

Quoted 10 lines · Arnd Bergmann

Quoted 8 lines · On Tuesday 28 May 2013 00:02:29 Arnaud Ebalard wrote:

This series adds support for GMT G762/G763. This work is based on a basic version for 2.6.31 kernel developed Olivier Mouchet for LaCie NAS. Updates have been performed to run on recent kernels. Support has been completed and additional features added: ability to configure various characteristics from .dts file, better initialization, alarms and error reporting support, gear mode, polarity, fan pulse per revolution, fan startup voltage control.

Quoted 3 lines

I wonder if this could be split into two separate drivers, one for the pwm subsystem, and one for a hardware-independent fan controller based on the pwm interfaces.

To be honest, I wouldn't even know how to start in order to do that. Additionally, it would be worth the effort if other drivers could be easily refactored using this idea. I'll let hwmon people provide some feedback and try and comply with their directions (based on the cpu cycle I can spend on this).

Quoted 4 lines

What Arnd proposes doesn't sounds like a very good idea. PWM-controlled fans should be able to work just fine using a generic driver that uses the PWM framework, similar to the PWM backlight and LED drivers.

Isn't that exactly what Arnd proposed ?

Quoted 26 lines · Guenter Roeck

Quoted 25 lines · Thierry Reding

Quoted 23 lines · Arnaud Ebalard

Hi Arnd,

Quoted 10 lines · Arnd Bergmann

Quoted 8 lines · On Tuesday 28 May 2013 00:02:29 Arnaud Ebalard wrote:

This series adds support for GMT G762/G763. This work is based on a basic version for 2.6.31 kernel developed Olivier Mouchet for LaCie NAS. Updates have been performed to run on recent kernels. Support has been completed and additional features added: ability to configure various characteristics from .dts file, better initialization, alarms and error reporting support, gear mode, polarity, fan pulse per revolution, fan startup voltage control.

Quoted 3 lines

I wonder if this could be split into two separate drivers, one for the pwm subsystem, and one for a hardware-independent fan controller based on the pwm interfaces.

To be honest, I wouldn't even know how to start in order to do that. Additionally, it would be worth the effort if other drivers could be easily refactored using this idea. I'll let hwmon people provide some feedback and try and comply with their directions (based on the cpu cycle I can spend on this).

Quoted 5 lines

Quoted 4 lines

What Arnd proposes doesn't sounds like a very good idea. PWM-controlled fans should be able to work just fine using a generic driver that uses the PWM framework, similar to the PWM backlight and LED drivers.

Isn't that exactly what Arnd proposed ?

Oh, right. I meant to say: "What Arnd proposes sounds like a very good idea." =) Sorry for the confusion.