Skip to contentSkip to conversation

hwmon: (gpio-fan) allow to use alarm support alone from DT

PATCH New Simon Guinot 3 participants 12 messages Feb 23, 2015 – Feb 25, 2015

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

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

On some boards, such as the LaCie 2Big Network v2 or 2Big NAS (based on Marvell Kirkwood SoCs), an I2C fan controller is used but the alarm signal is wired to a separate GPIO. Unfortunately, the gpio-fan driver can't be used to handle GPIO alarm alone from DT: an error is returned if the "gpios" DT property is missing.

This patch allows to use the gpio-fan driver even if the "alarm-gpios" DT property is defined alone.

---
 .../devicetree/bindings/gpio/gpio-fan.txt          |  6 ++-
 drivers/hwmon/gpio-fan.c                           | 44 +++++++++++-----------
 2 files changed, 27 insertions(+), 23 deletions(-)
Documentation/devicetree/bindings/gpio/gpio-fan.txt +4 -2
index 2dd457a3469a..a4c5d15b72e6 100644
@@ -2,16 +2,18 @@ Bindings for fan connected to GPIO lines
 
 Required properties:
 - compatible : "gpio-fan"
+
+Optional properties:
 - gpios: Specifies the pins that map to bits in the control value,
   ordered MSB-->LSB.
 - gpio-fan,speed-map: A mapping of possible fan RPM speeds and the
   control value that should be set to achieve them. This array
   must have the RPM values in ascending order.
-
-Optional properties:
 - alarm-gpios: This pin going active indicates something is wrong with
   the fan, and a udev event will be fired.
 
+Note: At least one the "gpios" and "alarm-gpios" properties should be set.
+
 Examples:
 
 	gpio_fan {
drivers/hwmon/gpio-fan.c +23 -21

59 lines not on this page · raw message

Signature · 1 line
2.1.1
#1 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 26 lines · Simon Guinot

On some boards, such as the LaCie 2Big Network v2 or 2Big NAS (based on Marvell Kirkwood SoCs), an I2C fan controller is used but the alarm signal is wired to a separate GPIO. Unfortunately, the gpio-fan driver can't be used to handle GPIO alarm alone from DT: an error is returned if the "gpios" DT property is missing.

This patch allows to use the gpio-fan driver even if the "alarm-gpios" DT property is defined alone.

---
 .../devicetree/bindings/gpio/gpio-fan.txt          |  6 ++-
 drivers/hwmon/gpio-fan.c                           | 44 +++++++++++-----------
 2 files changed, 27 insertions(+), 23 deletions(-)
Documentation/devicetree/bindings/gpio/gpio-fan.txt +2 -0
index 2dd457a3469a..a4c5d15b72e6 100644
@@ -2,16 +2,18 @@ Bindings for fan connected to GPIO lines
 
 Required properties:
 - compatible : "gpio-fan"
+
+Optional properties:
 - gpios: Specifies the pins that map to bits in the control value,

Documentation/devicetree/bindings/gpio/gpio-fan.txt Bindings for fan connected to GPIO lines · line 15

Quoted 10 lines
Documentation/devicetree/bindings/gpio/gpio-fan.txt +1 -2
   ordered MSB-->LSB.
 - gpio-fan,speed-map: A mapping of possible fan RPM speeds and the
   control value that should be set to achieve them. This array
   must have the RPM values in ascending order.
-
-Optional properties:
 - alarm-gpios: This pin going active indicates something is wrong with
   the fan, and a udev event will be fired.
 
+Note: At least one the "gpios" and "alarm-gpios" properties should be set.

I think you have this sentence wrong. What i think you mean is

Note: At least one of "gpios" or "alarm-gpios" properties should be set.

Andrew

maintainer
#2 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

In reply to Simon Guinot
Quoted 9 lines · Simon Guinot

On some boards, such as the LaCie 2Big Network v2 or 2Big NAS (based on Marvell Kirkwood SoCs), an I2C fan controller is used but the alarm signal is wired to a separate GPIO. Unfortunately, the gpio-fan driver can't be used to handle GPIO alarm alone from DT: an error is returned if the "gpios" DT property is missing.

This patch allows to use the gpio-fan driver even if the "alarm-gpios" DT property is defined alone.

That is the wrong solution. The gpio alarm signal should be handled by the fan controller driver.

Guenter

#3 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 7 lines · Guenter Roeck
Quoted 6 lines · Simon Guinot

On some boards, such as the LaCie 2Big Network v2 or 2Big NAS (based on Marvell Kirkwood SoCs), an I2C fan controller is used but the alarm signal is wired to a separate GPIO. Unfortunately, the gpio-fan driver can't be used to handle GPIO alarm alone from DT: an error is returned if the "gpios" DT property is missing.

Quoted 6 lines
Quoted 3 lines

This patch allows to use the gpio-fan driver even if the "alarm-gpios" DT property is defined alone.

That is the wrong solution. The gpio alarm signal should be handled by the fan controller driver.

Hi Guenter,

Sure it should, but unfortunately it is not the case. I have several boards using this mechanism (ie: a separate fan alarm GPIO). I think the idea was to reduce the board cost...

Then this means I need a way to support this alarm signal and I can't find a better one than using gpio-fan. Note that this was possible with the original gpio-fan implementation (before the DT binding addition).

Let me know what you think.

Simon

Quoted 8 lines

Guenter

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

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150223/d5db44a8/attachment.sig>

maintainer
#4 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 12 lines · Simon Guinot
Quoted 11 lines · Guenter Roeck
Quoted 9 lines · Simon Guinot

On some boards, such as the LaCie 2Big Network v2 or 2Big NAS (based on Marvell Kirkwood SoCs), an I2C fan controller is used but the alarm signal is wired to a separate GPIO. Unfortunately, the gpio-fan driver can't be used to handle GPIO alarm alone from DT: an error is returned if the "gpios" DT property is missing.

This patch allows to use the gpio-fan driver even if the "alarm-gpios" DT property is defined alone.

Quoted 9 lines
Quoted 2 lines

That is the wrong solution. The gpio alarm signal should be handled by the fan controller driver.

Hi Guenter,

Sure it should, but unfortunately it is not the case. I have several boards using this mechanism (ie: a separate fan alarm GPIO). I think the idea was to reduce the board cost...

Well, yes, the driver for the fan controller chip needs to be updated to support interrupts.

Quoted 4 lines

Then this means I need a way to support this alarm signal and I can't find a better one than using gpio-fan. Note that this was possible with the original gpio-fan implementation (before the DT binding addition).

That doesn't help.

Guenter

#5 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 16 lines · Guenter Roeck
Quoted 15 lines · Simon Guinot
Quoted 13 lines · Guenter Roeck
Quoted 9 lines · Simon Guinot

On some boards, such as the LaCie 2Big Network v2 or 2Big NAS (based on Marvell Kirkwood SoCs), an I2C fan controller is used but the alarm signal is wired to a separate GPIO. Unfortunately, the gpio-fan driver can't be used to handle GPIO alarm alone from DT: an error is returned if the "gpios" DT property is missing.

This patch allows to use the gpio-fan driver even if the "alarm-gpios" DT property is defined alone.

That is the wrong solution. The gpio alarm signal should be handled by the fan controller driver.

Quoted 8 lines
Quoted 6 lines

Hi Guenter,

Sure it should, but unfortunately it is not the case. I have several boards using this mechanism (ie: a separate fan alarm GPIO). I think the idea was to reduce the board cost...

Well, yes, the driver for the fan controller chip needs to be updated to support interrupts.

This will not help for the boards I mentioned. As an attempt to reduce the board cost, a fan with 2 wires has been used. This means we don't have any tachymeter feedback and then the controller alarm mechanism can't be used. Instead a kind of hardware hack allows to have a fan alarm on a separate GPIO.

Quoted 7 lines
Quoted 5 lines

Then this means I need a way to support this alarm signal and I can't find a better one than using gpio-fan. Note that this was possible with the original gpio-fan implementation (before the DT binding addition).

That doesn't help.

Handle the GPIO fan alarm feature from the fan controller driver don't look good either to me. This alarm mechanism is not a part of the fan controller itself but rather something apart. Also I am afraid that the result would really look like a hack.

Simon

Quoted 6 lines

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

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150225/7fcd800a/attachment.sig>

#6 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 4 lines

Handle the GPIO fan alarm feature from the fan controller driver don't look good either to me. This alarm mechanism is not a part of the fan controller itself but rather something apart. Also I am afraid that the result would really look like a hack.

Hi Simon

It sounds like you need to extract the alarm code from gpio-fan into a little library. Then put a wrapper around it to make a gpio-fan-alarm driver.

Andrew

#7 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 10 lines · Andrew Lunn
Quoted 4 lines

Handle the GPIO fan alarm feature from the fan controller driver don't look good either to me. This alarm mechanism is not a part of the fan controller itself but rather something apart. Also I am afraid that the result would really look like a hack.

Hi Simon

It sounds like you need to extract the alarm code from gpio-fan into a little library. Then put a wrapper around it to make a gpio-fan-alarm driver.

Hi Andrew,

Yes, I am working on that. I'll come back with some patches based on this approach.

Simon
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 181 bytes
Desc: Digital signature
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20150225/794db90a/attachment.sig>

maintainer
#8 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 8 lines · Simon Guinot
Quoted 7 lines · Andrew Lunn
Quoted 4 lines

Handle the GPIO fan alarm feature from the fan controller driver don't look good either to me. This alarm mechanism is not a part of the fan controller itself but rather something apart. Also I am afraid that the result would really look like a hack.

Hi Simon

Quoted 9 lines
Quoted 3 lines

It sounds like you need to extract the alarm code from gpio-fan into a little library. Then put a wrapper around it to make a gpio-fan-alarm driver.

Hi Andrew,

Yes, I am working on that. I'll come back with some patches based on this approach.

I have not seen your patches, but from the context it sounds like a NACK.

Guenter

maintainer
#9 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

In reply to Andrew Lunn
Quoted 11 lines · Andrew Lunn
Quoted 4 lines

Handle the GPIO fan alarm feature from the fan controller driver don't look good either to me. This alarm mechanism is not a part of the fan controller itself but rather something apart. Also I am afraid that the result would really look like a hack.

Hi Simon

It sounds like you need to extract the alarm code from gpio-fan into a little library. Then put a wrapper around it to make a gpio-fan-alarm driver.

Please, the intend is to do the right thing, not to cause code bloat.

If it is in fact correct that the alarm mechanism in this case is not tied to the fan controller, using the gpio-fan driver is ok. However, we need to state and check in the code that _some_ property is mandatory. A driver with only optional properties doesn't make sense.

Guenter

#10 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 3 lines

However, we need to state and check in the code that _some_ property is mandatory. A driver with only optional properties doesn't make sense.

The patch to the binding documentation said:

Note: At least one the "gpios" and "alarm-gpios" properties should be set.

which i then suggested should be:

Note: At least one of "gpios" or "alarm-gpios" properties should be set.

What we could do is move this sentence into the Required Properties of the documentation.

The patch Simon submitted also does check that this condition is fulfilled.

Andrew

maintainer
#11 lore raw

PATCH 1/1 hwmon: (gpio-fan) allow to use alarm support alone from DT

Quoted 4 lines · Andrew Lunn
Quoted 3 lines

However, we need to state and check in the code that _some_ property is mandatory. A driver with only optional properties doesn't make sense.

Quoted 8 lines

The patch to the binding documentation said:

Note: At least one the "gpios" and "alarm-gpios" properties should be set.

which i then suggested should be:

Note: At least one of "gpios" or "alarm-gpios" properties should be set.

"should" is not "must".

Guenter

Quoted 10 lines

What we could do is move this sentence into the Required Properties of the documentation.

The patch Simon submitted also does check that this condition is fulfilled.

Andrew