regulator: Display in use regulators in sysfs
RFC 1/1 regulator: Display in use regulators in sysfs
Please excuse any poor formatting that may occur, as this is copy-paste into my mailer, but what do you think of this idea?
To locate a consumer that currently holds a regulattor, a new sysfs entry is created. The consumer are published in /sys/class/regulator/regulator.<#>/use.
--- drivers/regulator/core.c | 36 ++++++++++++++++++++++++++++++++++++ 1 file changed, 36 insertions(+)
drivers/regulator/core.c +1 -0
@@ -81,6 +81,7 @@ struct regulator {struct device_attribute dev_attr;struct regulator_dev *rdev;struct dentry *debugfs; +int use; };static int _regulator_is_enabled (struct regulator_dev *rdev);
@@ -602,6 +603,32 @@ static ssize_t
regulator_suspend_standby_state_show(struct device *dev,
static DEVICE_ATTR(suspend_standby_state, 0444,
regulator_suspend_standby_state_show,
NULL);
+static ssize_t regulator_use_show(struct device *dev,
+ struct device_attribute *attr, char *buf)
+{
+ struct regulator_dev *rdev = dev_get_drvdata(dev);
+ struct regulator *reg;
+ size_t size = 0;
+
+ if (rdev->use_count == 0)
+ return sprintf(buf, "no users\n");
+
+ list_for_each_entry(reg, &rdev->consumer_list, list) {
+ if (!reg->use)
+ continue;
+
+ if (reg->dev != NULL)
+ size += sprintf((buf + size), "%s (%d) ",
+ dev_name(reg->dev), reg->use);
+ else
+ size += sprintf((buf + size), "unknown (%d) ",
+ reg->use);
+ }
+ size += sprintf((buf + size), "\n");
+
+ return size;
+}
+static DEVICE_ATTR(use, 0444, regulator_use_show, NULL);
/*
* These are the only attributes are present for all regulators.
@@ -1540,6 +1567,8 @@ int regulator_enable(struct regulator *regulator)
if (ret != 0 && rdev->supply)
regulator_disable(rdev->supply);
+ else
+ regulator->use++;
return ret; } @@ -1613,6 +1642,9 @@ int regulator_disable(struct regulator *regulator) if (ret == 0 && rdev->supply)
regulator_disable(rdev->supply);
+ if (ret == 0)
+ regulator->use--;
+
return ret;
}
EXPORT_SYMBOL_GPL(regulator_disable);
@@ -2699,6 +2731,10 @@ static int add_regulator_attributes(struct
regulator_dev *rdev)
struct regulator_ops *ops = rdev->desc->ops;
int status = 0;
+ status = device_create_file(dev, &dev_attr_use);
+ if (status < 0)
+ dev_warn(dev, "Create sysfs file \"use\" failed");
+
/* some attributes need specific methods to be displayed */
if ((ops->get_voltage && ops->get_voltage(rdev) >= 0) ||
(ops->get_voltage_sel && ops->get_voltage_sel(rdev) >= 0))
{Signature · 4 lines
Lee Jones Linaro ST-Ericsson Landing Team Lead Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Quoted 3 lines · Lee Jones
To locate a consumer that currently holds a regulattor, a new sysfs entry is created. The consumer are published in /sys/class/regulator/regulator.<#>/use.
We call our reference counts use count but this was in fact the original design, the local reference counts got dropped due to David Brownell's forceful insistance :/
Quoted 9 lines · Mark Brown
Quoted 3 lines · Lee Jones
To locate a consumer that currently holds a regulattor, a new sysfs entry is created. The consumer are published in /sys/class/regulator/regulator.<#>/use.
We call our reference counts use count but this was in fact the original design, the local reference counts got dropped due to David Brownell's forceful insistance :/
So I should pull the patch?
How do you feel about being able to access it via sysfs? Good or bad idea? If it's a good idea, I can always change the name of the variable?
Signature · 4 lines
Lee Jones Linaro ST-Ericsson Landing Team Lead Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Quoted 7 lines · Lee Jones
Quoted 3 lines · Mark Brown
We call our reference counts use count but this was in fact the original design, the local reference counts got dropped due to David Brownell's forceful insistance :/
So I should pull the patch?
Not sure what you mean by pull; a stright revert won't work.
Quoted 3 lines
How do you feel about being able to access it via sysfs? Good or bad idea? If it's a good idea, I can always change the name of the variable?
That or debugfs.
Quoted 10 lines · Mark Brown
Quoted 7 lines · Lee Jones
Quoted 3 lines · Mark Brown
We call our reference counts use count but this was in fact the original design, the local reference counts got dropped due to David Brownell's forceful insistance :/
So I should pull the patch?
Not sure what you mean by pull; a stright revert won't work.
Sorry, perhaps I should have been clearer.
Is the patch any good, or should I throw it in the bin (pull it)?
Quoted 5 lines
Quoted 3 lines
How do you feel about being able to access it via sysfs? Good or bad idea? If it's a good idea, I can always change the name of the variable?
That or debugfs.
So the premise of the patch is good and is worth updating and upstreaming?
Signature · 4 lines
Lee Jones Linaro ST-Ericsson Landing Team Lead Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog
Quoted 1 line · Lee Jones
So the premise of the patch is good and is worth updating and upstreaming?
Yes.