Re: [PATCH 1/5] hwmon: pmbus: Implement generic bus access delay

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 7/15/24 00:30, Patrick Rudolph wrote:
Some drivers, like the max15301 or zl6100, are intentionally delaying
SMBus communications, to prevent transmission errors. As this is necessary
on additional PMBus compatible devices, implement a generic delay mechanism
in the pmbus core.

Introduces two delay settings in the pmbus_driver_info struct, one applies
to every SMBus transaction and the other is for write transaction only.
Once set by the driver the SMBus traffic, using the generic pmbus access
helpers, is automatically delayed when necessary.

The two settings are:
access_delay:
   - Unit in microseconds
   - Stores the accessed timestamp after every SMBus access
   - Delays when necessary before the next SMBus access

write_delay:
   - Unit in microseconds
   - Stores the written timestamp after a write SMBus access
   - Delays when necessary before the next SMBus access

This allows to drop the custom delay code from the drivers and easily
introduce this feature in additional pmbus drivers.

Signed-off-by: Patrick Rudolph <patrick.rudolph@xxxxxxxxxxxxx>
---
  drivers/hwmon/pmbus/pmbus.h      | 10 ++++
  drivers/hwmon/pmbus/pmbus_core.c | 92 +++++++++++++++++++++++++++++---
  2 files changed, 96 insertions(+), 6 deletions(-)

diff --git a/drivers/hwmon/pmbus/pmbus.h b/drivers/hwmon/pmbus/pmbus.h
index fb442fae7b3e..5d5dc774187b 100644
--- a/drivers/hwmon/pmbus/pmbus.h
+++ b/drivers/hwmon/pmbus/pmbus.h
@@ -466,6 +466,16 @@ struct pmbus_driver_info {
/* custom attributes */
  	const struct attribute_group **groups;
+
+	/*
+	 * Some chips need a little delay between SMBus communication. When
+	 * set, the generic PMBus helper functions will wait if necessary
+	 * to meet this requirement. The access delay is honored after
+	 * every SMBus operation. The write delay is only honored after
+	 * SMBus write operations.
+	 */
+	int access_delay;		/* in microseconds */
+	int write_delay;		/* in microseconds */
  };
/* Regulator ops */
diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
index cb4c65a7f288..5cb093c898a1 100644
--- a/drivers/hwmon/pmbus/pmbus_core.c
+++ b/drivers/hwmon/pmbus/pmbus_core.c
@@ -7,6 +7,7 @@
   */
#include <linux/debugfs.h>
+#include <linux/delay.h>
  #include <linux/kernel.h>
  #include <linux/math64.h>
  #include <linux/module.h>
@@ -108,6 +109,8 @@ struct pmbus_data {
int vout_low[PMBUS_PAGES]; /* voltage low margin */
  	int vout_high[PMBUS_PAGES];	/* voltage high margin */
+	ktime_t write_time;		/* Last SMBUS write timestamp */
+	ktime_t access_time;		/* Last SMBUS access timestamp */
  };
struct pmbus_debugfs_entry {
@@ -158,6 +161,39 @@ void pmbus_set_update(struct i2c_client *client, u8 reg, bool update)
  }
  EXPORT_SYMBOL_NS_GPL(pmbus_set_update, PMBUS);
+/* Some chips need a delay between accesses. */
+static inline void pmbus_optional_wait(struct i2c_client *client)

I'd suggest to name the function either pmbus_access_wait() or even simply pmbus_wait().

+{
+	struct pmbus_data *data = i2c_get_clientdata(client);
+	const struct pmbus_driver_info *info = data->info;
+	s64 delta;
+
+	if (info->access_delay) {
+		delta = ktime_us_delta(ktime_get(), data->access_time);
+
+		if (delta < info->access_delay)
+			udelay(info->access_delay - delta);

I am not too happy with this one; the delay can get large.
I'd suggest to use fsleep().

+	} else if (info->write_delay) {
+		delta = ktime_us_delta(ktime_get(), data->write_time);
+
+		if (delta < info->write_delay)
+			udelay(info->write_delay - delta);

Same here.

Thanks,
Guenter





[Index of Archives]     [LM Sensors]     [Linux Sound]     [ALSA Users]     [ALSA Devel]     [Linux Audio Users]     [Linux Media]     [Kernel]     [Gimp]     [Yosemite News]     [Linux Media]

  Powered by Linux