Skip to contentSkip to conversation

mfd: MAX6650/6651 support

PATCH Superseded Laszlo Papp 4 participants 15 messages Feb 12, 2014 – Feb 17, 2014 Multifunction devices (MFD)

v6 of 7 Superseded Latest is v7 →

lore
Share & tools
lore
#0 lore raw

PATCH v6 1/1 mfd: MAX6650/6651 support

MAX6650/MAX6651 chip is a multi-function device with I2C busses. The chip includes fan-speed regulators and monitors, GPIO, and alarm.

This patch is an initial release of a MAX6650/6651 MFD driver that supports to enable the chip with its primary I2C bus that will connect the hwmon, and then the gpio devices for now.

---
 drivers/mfd/Kconfig                 | 11 +++++
 drivers/mfd/Makefile                |  1 +
 drivers/mfd/max665x.c               | 87 +++++++++++++++++++++++++++++++++++++
 include/linux/mfd/max665x-private.h | 34 +++++++++++++++
 4 files changed, 133 insertions(+)
 create mode 100644 drivers/mfd/max665x.c
 create mode 100644 include/linux/mfd/max665x-private.h
drivers/mfd/Kconfig +11 -0
index 49bb445..a6c3152 100644
@@ -368,6 +368,17 @@ config MFD_MAX8907
 	  accessing the device; additional drivers must be enabled in order
 	  to use the functionality of the device.
 
+config MFD_MAX665X
+	bool "Maxim Semiconductor MAX6650/MAX6651 Support"
+	depends on I2C
+	select MFD_CORE
+	select REGMAP_I2C
+	help
+	  Say yes here to add support for Maxim Semiconductor MAX6650/MAX6651. This
+	  is a fan speed regulator and monitor IC. This driver provides common
+	  support for accessing the device, additional drivers must be enabled in
+	  order to use the functionality of the device.
+
 config MFD_MAX8925
 	bool "Maxim Semiconductor MAX8925 PMIC Support"
 	depends on I2C=y
drivers/mfd/Makefile +1 -0
index 5aea5ef..63668c5 100644
@@ -111,6 +111,7 @@ obj-$(CONFIG_MFD_DA9055)	+= da9055.o
 da9063-objs			:= da9063-core.o da9063-irq.o da9063-i2c.o
 obj-$(CONFIG_MFD_DA9063)	+= da9063.o
 
+obj-$(CONFIG_MFD_MAX665X)	+= max665x.o
 obj-$(CONFIG_MFD_MAX14577)	+= max14577.o
 obj-$(CONFIG_MFD_MAX77686)	+= max77686.o max77686-irq.o
 obj-$(CONFIG_MFD_MAX77693)	+= max77693.o max77693-irq.o
drivers/mfd/max665x.c new file +87 -0

87 lines not on this page · raw message

include/linux/mfd/max665x-private.h new file +34 -0
new file mode 100644
index 0000000..293db4b
@@ -0,0 +1,34 @@
+/*
+ * max665x-private.h - Driver for the Maxim 6650/6651
+ *
+ * Copyright 2013 Polatis Ltd.
+ *
+ * Author: Laszlo Papp <[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.
+ *
+ */
+
+#ifndef __LINUX_MFD_MAX665X_PRIVATE_H
+#define __LINUX_MFD_MAX665X_PRIVATE_H
+
+#include <linux/i2c.h>
+#include <linux/regmap.h>
+
+struct max665x_dev {
+	struct device *dev;
+	struct mutex iolock;
+	struct i2c_client *i2c;
+	struct regmap     *map;
+	int type;
+};
+
+#endif
Signature · 1 line
1.8.5.4
#1 lore raw

Hi Laszlo,

Sorry for missing out on a couple of points during my earlier review. Please see inline.

Quoted 9 lines · Laszlo Papp

MAX6650/MAX6651 chip is a multi-function device with I2C busses. The chip includes fan-speed regulators and monitors, GPIO, and alarm.

This patch is an initial release of a MAX6650/6651 MFD driver that supports to enable the chip with its primary I2C bus that will connect the hwmon, and then the gpio devices for now.

---

[snip]

Quoted 22 lines
drivers/mfd/max665x.c new file +19 -0
@@ -0,0 +1,87 @@
+/*
+ * Device access for MAX6650-MAX6651
+ *
+ * Copyright(c) 2013 Polatis Ltd.
+ *
+ * Author: Laszlo Papp <[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.
+ */
+
+#include <linux/device.h>
+#include <linux/mfd/core.h>
+#include <linux/module.h>
+#include <linux/i2c.h>
+
+#include <linux/mfd/max665x-private.h>

drivers/mfd/max665x.c line 29

Quoted 10 lines
drivers/mfd/max665x.c new file +10 -0
+
+static struct mfd_cell max665x_devs[] = {
+       { .name = "max6651-gpio", },
+       { .name = "max6650", }, /* hwmon driver */
+       {},
+};
+
+const struct regmap_config max665x_regmap_config = {
+       .reg_bits = 5,
+};

This should be static.

Quoted 4 lines
drivers/mfd/max665x.c new file +4 -0
+static int max665x_probe(struct i2c_client *i2c,
+                           const struct i2c_device_id *id)
+{
+       struct max665x_dev *max665x;

drivers/mfd/max665x.c line 44

Quoted 10 lines
drivers/mfd/max665x.c new file +10 -0
+       int ret;
+
+       max665x = devm_kzalloc(&i2c->dev, sizeof(*max665x), GFP_KERNEL);
+       if (!max665x)
+               return -ENOMEM;
+
+       i2c_set_clientdata(i2c, max665x);
+       max665x->dev = &i2c->dev;
+       max665x->i2c = i2c;
+       max665x->map = devm_regmap_init_i2c(i2c, &max665x_regmap_config);

Don't you need to check the return value of devm_regmap_init_i2c?

Signature · 2 lines
With warm regards,
Sachin
#2 lore raw
Quoted 4 lines · Sachin Kamat

Hi Laszlo,

Sorry for missing out on a couple of points during my earlier review. Please see inline.

Np.

Quoted 38 lines · Laszlo Papp
Quoted 9 lines

MAX6650/MAX6651 chip is a multi-function device with I2C busses. The chip includes fan-speed regulators and monitors, GPIO, and alarm.

This patch is an initial release of a MAX6650/6651 MFD driver that supports to enable the chip with its primary I2C bus that will connect the hwmon, and then the gpio devices for now.

---

[snip]

Quoted 26 lines
drivers/mfd/max665x.c new file +23 -0
@@ -0,0 +1,87 @@
+/*
+ * Device access for MAX6650-MAX6651
+ *
+ * Copyright(c) 2013 Polatis Ltd.
+ *
+ * Author: Laszlo Papp <[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.
+ */
+
+#include <linux/device.h>
+#include <linux/mfd/core.h>
+#include <linux/module.h>
+#include <linux/i2c.h>
+
+#include <linux/mfd/max665x-private.h>
+
+static struct mfd_cell max665x_devs[] = {
+       { .name = "max6651-gpio", },
+       { .name = "max6650", }, /* hwmon driver */
Quoted 8 lines

drivers/mfd/max665x.c line 29

Quoted 6 lines
drivers/mfd/max665x.c new file +6 -0
+       {},
+};
+
+const struct regmap_config max665x_regmap_config = {
+       .reg_bits = 5,
+};

This should be static.

Agree.

Quoted 8 lines
Quoted 8 lines
drivers/mfd/max665x.c new file +8 -0
+static int max665x_probe(struct i2c_client *i2c,
+                           const struct i2c_device_id *id)
+{
+       struct max665x_dev *max665x;
+       int ret;
+
+       max665x = devm_kzalloc(&i2c->dev, sizeof(*max665x), GFP_KERNEL);
+       if (!max665x)
Quoted 8 lines

drivers/mfd/max665x.c line 44

Quoted 6 lines
drivers/mfd/max665x.c new file +6 -0
+               return -ENOMEM;
+
+       i2c_set_clientdata(i2c, max665x);
+       max665x->dev = &i2c->dev;
+       max665x->i2c = i2c;
+       max665x->map = devm_regmap_init_i2c(i2c, &max665x_regmap_config);

Don't you need to check the return value of devm_regmap_init_i2c?

I personally think I should. I strived for consistency though with other similar drivers. After this many reviews about such things, it seems that consistency matters here less than other projects which I can cope with, thanks.

#3 lore raw
Quoted 8 lines
Quoted 3 lines
Quoted 1 line

+ max665x->map = devm_regmap_init_i2c(i2c, &max665x_regmap_config);

Don't you need to check the return value of devm_regmap_init_i2c?

I personally think I should. I strived for consistency though with other similar drivers. After this many reviews about such things, it seems that consistency matters here less than other projects which I can cope with, thanks.

The right thing to do is normally preferred.

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#4 lore raw
Quoted 10 lines · Lee Jones
Quoted 8 lines
Quoted 3 lines
Quoted 1 line

+ max665x->map = devm_regmap_init_i2c(i2c, &max665x_regmap_config);

Don't you need to check the return value of devm_regmap_init_i2c?

I personally think I should. I strived for consistency though with other similar drivers. After this many reviews about such things, it seems that consistency matters here less than other projects which I can cope with, thanks.

The right thing to do is normally preferred.

Yes, sure.

#5 lore raw
In reply to Laszlo Papp

drivers/mfd/max665x.c line 29

Quoted 3 lines · Laszlo Papp
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+	.reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

drivers/mfd/max665x.c line 46

Quoted 1 line
drivers/mfd/max665x.c new file +1 -0
+	mutex_init(&max665x->iolock);

What is this needed for?

#6 lore raw

Laszlo,

Quoted 6 lines

drivers/mfd/max665x.c line 29

Quoted 3 lines
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+	.reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

Mark (privately) mentioned to me that this patch can't possibly work given the current Regmap configuration. Patches are accepted into Mainline based on the premise that they are fully tested and working, _prior_ to submitting [1].

It's also pretty pointless having an MFD driver without any children, so unless (at least one of) the child device drivers have been accepted by pull-time, your work will not be part of the pull-request headed for Mainline.

Please inform me of your plans as you with to proceed.

[] Documentation/SubmitChecklist

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#7 lore raw
Quoted 11 lines · Lee Jones

Laszlo,

Quoted 6 lines

drivers/mfd/max665x.c line 29

Quoted 3 lines
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+   .reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

Mark (privately) mentioned to me that this patch can't possibly work given the current Regmap configuration.

Strange because I have tested the change, although not for days and weeks. What exactly cannot possible work?

Quoted 3 lines

Patches are accepted into
Mainline based on the premise that they are fully tested and working, _prior_ to submitting [1].

Yes, I am aware of it, hence the quick testing done.

Quoted 2 lines

It's also pretty pointless having an MFD driver without any children,

Agree.

Quoted 3 lines

so unless (at least one of) the child device drivers have been accepted by pull-time, your work will not be part of the pull-request headed for Mainline.

Sure, I think that is reasonable.

Quoted 1 line

Please inform me of your plans as you with to proceed.

I already sent the hwmon changes to the maintainers - you included or linked -, but got no response so far. I am all for making it work ASAP. I will make a more thorough testing today or tomorrow. After that, it is up to them...

#8 lore raw
Quoted 10 lines · Laszlo Papp
Quoted 8 lines · Lee Jones

Laszlo,

Quoted 6 lines

drivers/mfd/max665x.c line 29

Quoted 3 lines
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+   .reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

Quoted 5 lines
Quoted 2 lines

Mark (privately) mentioned to me that this patch can't possibly work given the current Regmap configuration.

Strange because I have tested the change, although not for days and weeks. What exactly cannot possible work?

The fact that it's using 5 bit registers should cause it to be rejected when trying to initialise as the regmap code doesn't support 5 bit registers, and I'd not expect us to have selected sensible value formatting code either.

#9 lore raw
In reply to Laszlo Papp
Quoted 10 lines · Laszlo Papp
Quoted 9 lines · Lee Jones

Laszlo,

Quoted 6 lines

drivers/mfd/max665x.c line 29

Quoted 3 lines
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+   .reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

Quoted 5 lines
Quoted 2 lines

Mark (privately) mentioned to me that this patch can't possibly work given the current Regmap configuration.

Strange because I have tested the change, although not for days and weeks. What exactly cannot possible work?

Hmm, yes, as it seems, I have not actually tested the patch set _after_ all the modifications. I deemed some of them trivial enough, but apparently I was wrong. In this case, sending trivial updates will slow down, but I guess that is the way to follow. Well, I apologize.

#10 lore raw
In reply to Mark Brown
Quoted 8 lines · Mark Brown

drivers/mfd/max665x.c line 29

Quoted 3 lines · Laszlo Papp
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+     .reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

Makes sense, thanks. Shall I also set the .max_register or any other mandatory variables in here?

Quoted 3 lines
Quoted 1 line

+ mutex_init(&max665x->iolock);

What is this needed for?

It was done for consistency with the other mfd drivers (maxim), e.g. 8997 or 8998 as the closest resemblence in this family. Would you prefer me removing this mutex locker?

#11 lore raw
Quoted 7 lines
Quoted 3 lines
Quoted 1 line

+ mutex_init(&max665x->iolock);

What is this needed for?

It was done for consistency with the other mfd drivers (maxim), e.g. 8997 or 8998 as the closest resemblence in this family. Would you prefer me removing this mutex locker?

If you're not using mutexes, why would you need to initialise it?

Signature · 4 lines
Lee Jones
Linaro STMicroelectronics Landing Team Lead
Linaro.org │ Open source software for ARM SoCs
Follow Linaro: Facebook | Twitter | Blog
#12 lore raw
Quoted 9 lines · Lee Jones
Quoted 7 lines
Quoted 3 lines
Quoted 1 line

+ mutex_init(&max665x->iolock);

What is this needed for?

It was done for consistency with the other mfd drivers (maxim), e.g. 8997 or 8998 as the closest resemblence in this family. Would you prefer me removing this mutex locker?

If you're not using mutexes, why would you need to initialise it?

Yeah, I had the same thought about an hour ago that this sneaked in because I was not originally using regmap, but now with the regmap usage, it is just a left-over. So, it is not even for consistency anymore. I was wrong claiming that.

#13 lore raw
In reply to Laszlo Papp
Quoted 10 lines · Laszlo Papp
Quoted 8 lines · Mark Brown

drivers/mfd/max665x.c line 29

Quoted 3 lines · Laszlo Papp
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+     .reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

Quoted 2 lines

Makes sense, thanks. Shall I also set the .max_register or any other mandatory variables in here?

It's better to set max_register since that makes the debug features of the regmap API more useful but it's not required, only reg_bits and val_bits are absolutely required. In general the more complete the description of the register map in the driver the better.

#14 lore raw
Quoted 11 lines · Mark Brown
Quoted 9 lines · Laszlo Papp
Quoted 8 lines · Mark Brown

drivers/mfd/max665x.c line 29

Quoted 3 lines · Laszlo Papp
drivers/mfd/max665x.c new file +3 -0
+const struct regmap_config max665x_regmap_config = {
+     .reg_bits = 5,
+};

This would normally be static too, and I'd *really* expect to see a val_bits set here. I'm a bit surprised this works without one.

Quoted 7 lines
Quoted 2 lines

Makes sense, thanks. Shall I also set the .max_register or any other mandatory variables in here?

It's better to set max_register since that makes the debug features of the regmap API more useful but it's not required, only reg_bits and val_bits are absolutely required. In general the more complete the description of the register map in the driver the better.

OK, thanks.