public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
* [PATCH] rtc: abx80x: Don't warn about oscillator failure after PoR
@ 2022-10-07 16:37 Sean Anderson
  2022-10-07 18:57 ` Alexandre Belloni
  0 siblings, 1 reply; 4+ messages in thread
From: Sean Anderson @ 2022-10-07 16:37 UTC (permalink / raw)
  To: Alessandro Zummo, Alexandre Belloni, linux-rtc
  Cc: linux-kernel, Sean Anderson

According to the datasheet, the "oscillator failure" bit is set

> ...on a power on reset, when both the system and battery voltages have
> dropped below acceptable levels. It is also set if an Oscillator Failure
> occurs....

From testing, this bit is also set if a software reset is initiated.

This bit has a confusing name; it really tells us whether the time data
is valid. We clear it when writing the time. If it is still set, that
means there is a persistent issue (such as an oscillator failure),
instead of a transient one (such as power loss).

Because there are several other reasons which might cause this bit
to be set (including booting for the first time or a battery failure),
do not warn about oscillator failures willy-nilly. This may cause system
integrators to waste time looking into the wrong line of investigation.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
---

 drivers/rtc/rtc-abx80x.c | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/drivers/rtc/rtc-abx80x.c b/drivers/rtc/rtc-abx80x.c
index 9b0138d07232..1eb752e4e39d 100644
--- a/drivers/rtc/rtc-abx80x.c
+++ b/drivers/rtc/rtc-abx80x.c
@@ -115,6 +115,7 @@ struct abx80x_priv {
 	struct rtc_device *rtc;
 	struct i2c_client *client;
 	struct watchdog_device wdog;
+	bool wrote_time;
 };
 
 static int abx80x_write_config_key(struct i2c_client *client, u8 key)
@@ -167,6 +168,7 @@ static int abx80x_enable_trickle_charger(struct i2c_client *client,
 static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
 {
 	struct i2c_client *client = to_i2c_client(dev);
+	struct abx80x_priv *priv = i2c_get_clientdata(client);
 	unsigned char buf[8];
 	int err, flags, rc_mode = 0;
 
@@ -181,7 +183,18 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
 			return flags;
 
 		if (flags & ABX8XX_OSS_OF) {
-			dev_err(dev, "Oscillator failure, data is invalid.\n");
+			/*
+			 * The OF bit can be set either because of a reset
+			 * (PoR/Software reset) or because of an oscillator
+			 * failure. Effectively, it indicates that the stored
+			 * time is invalid. When we write the time, we clear
+			 * this bit. If it stays set, then this indicates an
+			 * oscillator failure.
+			 */
+			if (priv->wrote_time)
+				dev_err(dev, "Oscillator failure\n");
+			else
+				dev_info(dev, "Time data invalid\n");
 			return -EINVAL;
 		}
 	}
@@ -207,6 +220,7 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
 static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
 {
 	struct i2c_client *client = to_i2c_client(dev);
+	struct abx80x_priv *priv = i2c_get_clientdata(client);
 	unsigned char buf[8];
 	int err, flags;
 
@@ -240,6 +254,7 @@ static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
 		dev_err(&client->dev, "Unable to write oscillator status register\n");
 		return err;
 	}
+	priv->wrote_time = true;
 
 	return 0;
 }
-- 
2.35.1.1320.gc452695387.dirty


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH] rtc: abx80x: Don't warn about oscillator failure after PoR
  2022-10-07 16:37 [PATCH] rtc: abx80x: Don't warn about oscillator failure after PoR Sean Anderson
@ 2022-10-07 18:57 ` Alexandre Belloni
  2022-10-07 19:05   ` Sean Anderson
  0 siblings, 1 reply; 4+ messages in thread
From: Alexandre Belloni @ 2022-10-07 18:57 UTC (permalink / raw)
  To: Sean Anderson; +Cc: Alessandro Zummo, linux-rtc, linux-kernel

Hi,

On 07/10/2022 12:37:12-0400, Sean Anderson wrote:
> According to the datasheet, the "oscillator failure" bit is set
> 
> > ...on a power on reset, when both the system and battery voltages have
> > dropped below acceptable levels. It is also set if an Oscillator Failure
> > occurs....
> 
> From testing, this bit is also set if a software reset is initiated.
> 
> This bit has a confusing name; it really tells us whether the time data
> is valid. We clear it when writing the time. If it is still set, that
> means there is a persistent issue (such as an oscillator failure),
> instead of a transient one (such as power loss).
> 
> Because there are several other reasons which might cause this bit
> to be set (including booting for the first time or a battery failure),
> do not warn about oscillator failures willy-nilly. This may cause system
> integrators to waste time looking into the wrong line of investigation.
> 
> Signed-off-by: Sean Anderson <sean.anderson@seco.com>
> ---
> 
>  drivers/rtc/rtc-abx80x.c | 17 ++++++++++++++++-
>  1 file changed, 16 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/rtc/rtc-abx80x.c b/drivers/rtc/rtc-abx80x.c
> index 9b0138d07232..1eb752e4e39d 100644
> --- a/drivers/rtc/rtc-abx80x.c
> +++ b/drivers/rtc/rtc-abx80x.c
> @@ -115,6 +115,7 @@ struct abx80x_priv {
>  	struct rtc_device *rtc;
>  	struct i2c_client *client;
>  	struct watchdog_device wdog;
> +	bool wrote_time;
>  };
>  
>  static int abx80x_write_config_key(struct i2c_client *client, u8 key)
> @@ -167,6 +168,7 @@ static int abx80x_enable_trickle_charger(struct i2c_client *client,
>  static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>  {
>  	struct i2c_client *client = to_i2c_client(dev);
> +	struct abx80x_priv *priv = i2c_get_clientdata(client);
>  	unsigned char buf[8];
>  	int err, flags, rc_mode = 0;
>  
> @@ -181,7 +183,18 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>  			return flags;
>  
>  		if (flags & ABX8XX_OSS_OF) {
> -			dev_err(dev, "Oscillator failure, data is invalid.\n");

Simply remove the line.

> +			/*
> +			 * The OF bit can be set either because of a reset
> +			 * (PoR/Software reset) or because of an oscillator
> +			 * failure. Effectively, it indicates that the stored
> +			 * time is invalid. When we write the time, we clear
> +			 * this bit. If it stays set, then this indicates an
> +			 * oscillator failure.
> +			 */
> +			if (priv->wrote_time)
> +				dev_err(dev, "Oscillator failure\n");
> +			else
> +				dev_info(dev, "Time data invalid\n");
>  			return -EINVAL;
>  		}
>  	}
> @@ -207,6 +220,7 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>  static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
>  {
>  	struct i2c_client *client = to_i2c_client(dev);
> +	struct abx80x_priv *priv = i2c_get_clientdata(client);
>  	unsigned char buf[8];
>  	int err, flags;
>  
> @@ -240,6 +254,7 @@ static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
>  		dev_err(&client->dev, "Unable to write oscillator status register\n");
>  		return err;
>  	}
> +	priv->wrote_time = true;
>  
>  	return 0;
>  }
> -- 
> 2.35.1.1320.gc452695387.dirty
> 

-- 
Alexandre Belloni, co-owner and COO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] rtc: abx80x: Don't warn about oscillator failure after PoR
  2022-10-07 18:57 ` Alexandre Belloni
@ 2022-10-07 19:05   ` Sean Anderson
  2022-10-26 21:17     ` Sean Anderson
  0 siblings, 1 reply; 4+ messages in thread
From: Sean Anderson @ 2022-10-07 19:05 UTC (permalink / raw)
  To: Alexandre Belloni; +Cc: Alessandro Zummo, linux-rtc, linux-kernel



On 10/7/22 2:57 PM, Alexandre Belloni wrote:
> Hi,
> 
> On 07/10/2022 12:37:12-0400, Sean Anderson wrote:
>> According to the datasheet, the "oscillator failure" bit is set
>> 
>> > ...on a power on reset, when both the system and battery voltages have
>> > dropped below acceptable levels. It is also set if an Oscillator Failure
>> > occurs....
>> 
>> From testing, this bit is also set if a software reset is initiated.
>> 
>> This bit has a confusing name; it really tells us whether the time data
>> is valid. We clear it when writing the time. If it is still set, that
>> means there is a persistent issue (such as an oscillator failure),
>> instead of a transient one (such as power loss).
>> 
>> Because there are several other reasons which might cause this bit
>> to be set (including booting for the first time or a battery failure),
>> do not warn about oscillator failures willy-nilly. This may cause system
>> integrators to waste time looking into the wrong line of investigation.
>> 
>> Signed-off-by: Sean Anderson <sean.anderson@seco.com>
>> ---
>> 
>>  drivers/rtc/rtc-abx80x.c | 17 ++++++++++++++++-
>>  1 file changed, 16 insertions(+), 1 deletion(-)
>> 
>> diff --git a/drivers/rtc/rtc-abx80x.c b/drivers/rtc/rtc-abx80x.c
>> index 9b0138d07232..1eb752e4e39d 100644
>> --- a/drivers/rtc/rtc-abx80x.c
>> +++ b/drivers/rtc/rtc-abx80x.c
>> @@ -115,6 +115,7 @@ struct abx80x_priv {
>>  	struct rtc_device *rtc;
>>  	struct i2c_client *client;
>>  	struct watchdog_device wdog;
>> +	bool wrote_time;
>>  };
>>  
>>  static int abx80x_write_config_key(struct i2c_client *client, u8 key)
>> @@ -167,6 +168,7 @@ static int abx80x_enable_trickle_charger(struct i2c_client *client,
>>  static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>>  {
>>  	struct i2c_client *client = to_i2c_client(dev);
>> +	struct abx80x_priv *priv = i2c_get_clientdata(client);
>>  	unsigned char buf[8];
>>  	int err, flags, rc_mode = 0;
>>  
>> @@ -181,7 +183,18 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>>  			return flags;
>>  
>>  		if (flags & ABX8XX_OSS_OF) {
>> -			dev_err(dev, "Oscillator failure, data is invalid.\n");
> 
> Simply remove the line.

I think it's important to warn the user if the oscillator actually fails
so they can e.g. replace the crystal. Additionally, this can help debug
failed batteries, since you will see "Time data invalid" in the boot log.

--Sean

>> +			/*
>> +			 * The OF bit can be set either because of a reset
>> +			 * (PoR/Software reset) or because of an oscillator
>> +			 * failure. Effectively, it indicates that the stored
>> +			 * time is invalid. When we write the time, we clear
>> +			 * this bit. If it stays set, then this indicates an
>> +			 * oscillator failure.
>> +			 */
>> +			if (priv->wrote_time)
>> +				dev_err(dev, "Oscillator failure\n");
>> +			else
>> +				dev_info(dev, "Time data invalid\n");
>>  			return -EINVAL;
>>  		}
>>  	}
>> @@ -207,6 +220,7 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>>  static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
>>  {
>>  	struct i2c_client *client = to_i2c_client(dev);
>> +	struct abx80x_priv *priv = i2c_get_clientdata(client);
>>  	unsigned char buf[8];
>>  	int err, flags;
>>  
>> @@ -240,6 +254,7 @@ static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
>>  		dev_err(&client->dev, "Unable to write oscillator status register\n");
>>  		return err;
>>  	}
>> +	priv->wrote_time = true;
>>  
>>  	return 0;
>>  }
>> -- 
>> 2.35.1.1320.gc452695387.dirty
>> 
> 

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: [PATCH] rtc: abx80x: Don't warn about oscillator failure after PoR
  2022-10-07 19:05   ` Sean Anderson
@ 2022-10-26 21:17     ` Sean Anderson
  0 siblings, 0 replies; 4+ messages in thread
From: Sean Anderson @ 2022-10-26 21:17 UTC (permalink / raw)
  To: Alexandre Belloni; +Cc: Alessandro Zummo, linux-rtc, linux-kernel

Hi Alexandre,

On 10/7/22 15:05, Sean Anderson wrote:
> 
> 
> On 10/7/22 2:57 PM, Alexandre Belloni wrote:
>> Hi,
>> 
>> On 07/10/2022 12:37:12-0400, Sean Anderson wrote:
>>> According to the datasheet, the "oscillator failure" bit is set
>>> 
>>> > ...on a power on reset, when both the system and battery voltages have
>>> > dropped below acceptable levels. It is also set if an Oscillator Failure
>>> > occurs....
>>> 
>>> From testing, this bit is also set if a software reset is initiated.
>>> 
>>> This bit has a confusing name; it really tells us whether the time data
>>> is valid. We clear it when writing the time. If it is still set, that
>>> means there is a persistent issue (such as an oscillator failure),
>>> instead of a transient one (such as power loss).
>>> 
>>> Because there are several other reasons which might cause this bit
>>> to be set (including booting for the first time or a battery failure),
>>> do not warn about oscillator failures willy-nilly. This may cause system
>>> integrators to waste time looking into the wrong line of investigation.
>>> 
>>> Signed-off-by: Sean Anderson <sean.anderson@seco.com>
>>> ---
>>> 
>>>  drivers/rtc/rtc-abx80x.c | 17 ++++++++++++++++-
>>>  1 file changed, 16 insertions(+), 1 deletion(-)
>>> 
>>> diff --git a/drivers/rtc/rtc-abx80x.c b/drivers/rtc/rtc-abx80x.c
>>> index 9b0138d07232..1eb752e4e39d 100644
>>> --- a/drivers/rtc/rtc-abx80x.c
>>> +++ b/drivers/rtc/rtc-abx80x.c
>>> @@ -115,6 +115,7 @@ struct abx80x_priv {
>>>  	struct rtc_device *rtc;
>>>  	struct i2c_client *client;
>>>  	struct watchdog_device wdog;
>>> +	bool wrote_time;
>>>  };
>>>  
>>>  static int abx80x_write_config_key(struct i2c_client *client, u8 key)
>>> @@ -167,6 +168,7 @@ static int abx80x_enable_trickle_charger(struct i2c_client *client,
>>>  static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>>>  {
>>>  	struct i2c_client *client = to_i2c_client(dev);
>>> +	struct abx80x_priv *priv = i2c_get_clientdata(client);
>>>  	unsigned char buf[8];
>>>  	int err, flags, rc_mode = 0;
>>>  
>>> @@ -181,7 +183,18 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>>>  			return flags;
>>>  
>>>  		if (flags & ABX8XX_OSS_OF) {
>>> -			dev_err(dev, "Oscillator failure, data is invalid.\n");
>> 
>> Simply remove the line.
> 
> I think it's important to warn the user if the oscillator actually fails
> so they can e.g. replace the crystal. Additionally, this can help debug
> failed batteries, since you will see "Time data invalid" in the boot log.

Have you considered my reply?

I'd also like to note that 

drivers/rtc/rtc-ds1672.c
drivers/rtc/rtc-pcf*.c
drivers/rtc/rtc-rs5c*.c
drivers/rtc/rtc-sc27xx.c

all produce some kind of message if they detect a power loss or oscillator failure.

--Sean


>>> +			/*
>>> +			 * The OF bit can be set either because of a reset
>>> +			 * (PoR/Software reset) or because of an oscillator
>>> +			 * failure. Effectively, it indicates that the stored
>>> +			 * time is invalid. When we write the time, we clear
>>> +			 * this bit. If it stays set, then this indicates an
>>> +			 * oscillator failure.
>>> +			 */
>>> +			if (priv->wrote_time)
>>> +				dev_err(dev, "Oscillator failure\n");
>>> +			else
>>> +				dev_info(dev, "Time data invalid\n");
>>>  			return -EINVAL;
>>>  		}
>>>  	}
>>> @@ -207,6 +220,7 @@ static int abx80x_rtc_read_time(struct device *dev, struct rtc_time *tm)
>>>  static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
>>>  {
>>>  	struct i2c_client *client = to_i2c_client(dev);
>>> +	struct abx80x_priv *priv = i2c_get_clientdata(client);
>>>  	unsigned char buf[8];
>>>  	int err, flags;
>>>  
>>> @@ -240,6 +254,7 @@ static int abx80x_rtc_set_time(struct device *dev, struct rtc_time *tm)
>>>  		dev_err(&client->dev, "Unable to write oscillator status register\n");
>>>  		return err;
>>>  	}
>>> +	priv->wrote_time = true;
>>>  
>>>  	return 0;
>>>  }
>>> -- 
>>> 2.35.1.1320.gc452695387.dirty
>>> 
>> 


^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2022-10-26 21:17 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2022-10-07 16:37 [PATCH] rtc: abx80x: Don't warn about oscillator failure after PoR Sean Anderson
2022-10-07 18:57 ` Alexandre Belloni
2022-10-07 19:05   ` Sean Anderson
2022-10-26 21:17     ` Sean Anderson

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox