[RFC PATCH] watchdog: da9062: Correct the timeout values

Andrej Picej andrej.picej at norik.com
Sun Dec 5 23:29:04 PST 2021



On 3. 12. 21 18:31, Christoph Niedermaier wrote:
> From: Guenter Roeck
> Sent: Friday, December 3, 2021 5:52 PM
>> On 12/3/21 8:35 AM, Christoph Niedermaier wrote:
>>> I measured the timeout values of my DA9061 chip. According to the
>>> information in the data sheet the formula should be:
>>>
>>> timeout = 2.048 * 2^(regval - 1)
>>>
>>> But my measured values differ from that.
>>> Accoring to my measured values the formula must be:
>>>
>>> timeout = 3.2 * 2^(regval - 1)
>>>
>>> Is there something wrong with my chip, or has anyone else noticed this as well?
>>
>> The driver assumes a static and well defined clock rate. Maybe that rate
>> is different in your system (if that is possible) ?
>>
>> Guenter
> 
> @Andrej
> Do the values in the driver match what your chip does?
> 

Just did a quick test. The values in the driver match what the chip 
does. I checked multiple timeouts 16, 32, 65 and 131 seconds. The 
timeout triggers quite accurately.

> I have not changed anything. After power on, the chip behaves like this.
> So I guess it either come from an OTP value or the wiring outside the chip.
> Does anyone know what needs to be checked?

Can't help you here, sorry.

Best regards,
Andrej



More information about the linux-arm-kernel mailing list