Re: [PATCH v2 3/4] hwmon: (pmbus/core): Add interrupt support

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

 



Hi

On 11-02-2023 09:07 pm, Guenter Roeck wrote:
On Tue, Feb 07, 2023 at 01:02:40PM +0100, Naresh Solanki wrote:
From: Patrick Rudolph <patrick.rudolph@xxxxxxxxxxxxx>

Implement PMBUS irq handler.

Signed-off-by: Patrick Rudolph <patrick.rudolph@xxxxxxxxxxxxx>
Signed-off-by: Naresh Solanki <Naresh.Solanki@xxxxxxxxxxxxx>
---
  drivers/hwmon/pmbus/pmbus.h      |  2 +-
  drivers/hwmon/pmbus/pmbus_core.c | 85 ++++++++++++++++++++++++++++++++
  2 files changed, 86 insertions(+), 1 deletion(-)

diff --git a/drivers/hwmon/pmbus/pmbus.h b/drivers/hwmon/pmbus/pmbus.h
index 713ea7915425..11e84e141126 100644
--- a/drivers/hwmon/pmbus/pmbus.h
+++ b/drivers/hwmon/pmbus/pmbus.h
@@ -26,7 +26,7 @@ enum pmbus_regs {
PMBUS_CAPABILITY = 0x19,
  	PMBUS_QUERY			= 0x1A,
-
+	PMBUS_SMBALERT_MASK		= 0x1B,
  	PMBUS_VOUT_MODE			= 0x20,
  	PMBUS_VOUT_COMMAND		= 0x21,
  	PMBUS_VOUT_TRIM			= 0x22,
diff --git a/drivers/hwmon/pmbus/pmbus_core.c b/drivers/hwmon/pmbus/pmbus_core.c
index 5ccae8126a56..d5403baad60a 100644
--- a/drivers/hwmon/pmbus/pmbus_core.c
+++ b/drivers/hwmon/pmbus/pmbus_core.c
@@ -3093,6 +3093,85 @@ static int pmbus_regulator_register(struct pmbus_data *data)
  }
  #endif
+static int pmbus_write_smbalert_mask(struct i2c_client *client, u8 page, u8 reg, u8 val)
+{
+	int err;
+
+	err = pmbus_check_word_register(client, page, reg | (val << 8));
+	if (err)
+		return err;

I am not convinced that this is necessary. The next command will return an
error anyway if the register or the specific mask is not supported, so what
is the point ?

Sure. will remove.
+
+	return pmbus_write_word_data(client, page, PMBUS_SMBALERT_MASK, reg | (val << 8));
+}
+
+static irqreturn_t pmbus_fault_handler(int irq, void *pdata)
+{
+	struct pmbus_data *data = pdata;
+	struct i2c_client *client = to_i2c_client(data->dev);
+	int i, status;
+
+	mutex_lock(&data->update_lock);
+	for (i = 0; i < data->info->pages; i++) {
+		status = pmbus_read_status_word(client, i);
+		if (status < 0) {
+			mutex_unlock(&data->update_lock);
+			return status;
+		}
+
+		if (status & ~(PB_STATUS_OFF | PB_STATUS_BUSY | PB_STATUS_POWER_GOOD_N))
+			pmbus_clear_fault_page(client, i);
+	}
+	mutex_unlock(&data->update_lock);
+
+	return IRQ_HANDLED;
+}
+
+static int pmbus_irq_setup(struct i2c_client *client, struct pmbus_data *data)
+{
+	struct device *dev = &client->dev;
+	const struct pmbus_status_category *cat;
+	const struct pmbus_status_assoc *bit;
+	int i, j, err, ret, func;
+	u8 mask;
+	u8 misc_status[] = {PMBUS_STATUS_CML, PMBUS_STATUS_OTHER, PMBUS_STATUS_MFR_SPECIFIC,
+			    PMBUS_STATUS_FAN_12, PMBUS_STATUS_FAN_34};

static const u8 ...

Done
+
+	for (i = 0; i < data->info->pages; i++) {
+		func = data->info->func[i];
+
+		for (j = 0; j < ARRAY_SIZE(pmbus_status_flag_map); j++) {
+			cat = &pmbus_status_flag_map[j];
+			if (!(func & cat->func))
+				continue;
+			mask = 0;
+			for (bit = cat->bits; bit->pflag; bit++)
+				mask |= bit->pflag;
+
+			err = pmbus_write_smbalert_mask(client, i, cat->reg, ~mask);
+			if (err)
+				dev_err_once(dev, "Failed to set smbalert for reg 0x%02x\n",
+					     cat->reg);

dev_err implies an error. This is ignored and thus not an error. On top of that,
not all chips support PMBUS_SMBALERT_MASK. All of those would see this message.
We can't have that. At best make it a dev_dbg.

Sure. Will make it dev_dbg_once.
+		}
+
+		for (j = 0; j < ARRAY_SIZE(misc_status); j++) {
+			err = pmbus_write_smbalert_mask(client, i, misc_status[j], 0xff);
+			if (err)
+				dev_err_once(dev, "Failed to set smbalert for reg 0x%02x\n",
+					     misc_status[j]);

We definitely can't have a message here; that would fire for almost
every chip.

Sure. Will remove printing of error here.
+		}
+	}
+
+	/* Register notifiers - can fail if IRQ is not given */

If there is no irq, what is the point of executing this code in the first
place ? No, wait, in that case the function isn't called in the first place.
I think the comment doesn't add any value and is just confusing.

Will clean this comment.
+	ret = devm_request_threaded_irq(dev, client->irq, NULL, pmbus_fault_handler, 0,
+					"pmbus-irq", data);
+	if (ret) {

Why both "err" and "ret" ?

Will replace ret with err.
+		dev_warn(dev, "IRQ disabled %d\n", ret);

The calling code aborts, so this should be dev_err() and say something
better than "IRQ disabled"; It should be something like "failed to
request irq".

Sure. Will update to "failed to request an irq"
+		return ret;
+	}
+
+	return 0;
+}
+
  static struct dentry *pmbus_debugfs_dir;	/* pmbus debugfs directory */
#if IS_ENABLED(CONFIG_DEBUG_FS)
@@ -3455,6 +3534,12 @@ int pmbus_do_probe(struct i2c_client *client, struct pmbus_driver_info *info)
  	if (ret)
  		return ret;
+ if (client->irq > 0) {
+		ret = pmbus_irq_setup(client, data);
+		if (ret)
+			return ret;
+	}
+

I think it would be better to have the check in pmbus_irq_setup():

pmbus_irq_setup()
{
	if (!client->irq)
		return 0;
	
	...
}

and here
	ret = pmbus_irq_setup(client, data);
	if (ret)
		return ret;


Sure
  	ret = pmbus_init_debugfs(client, data);
  	if (ret)
  		dev_warn(dev, "Failed to register debugfs\n");



[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