linux-kernel - Re: [PATCH] iio: tmag5273: fix temperature offset

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Fri, 27 Oct 2023 17:24:23 +0200
From:   Javier Carrasco <javier.carrasco@...fvision.net>
To:     Jonathan Cameron <jic23@...nel.org>
Cc:     Gerald Loacker <gerald.loacker@...fvision.net>,
        Lars-Peter Clausen <lars@...afoo.de>,
        linux-iio@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] iio: tmag5273: fix temperature offset



On 27.10.23 16:15, Jonathan Cameron wrote:
> On Mon, 23 Oct 2023 11:50:02 +0200
> Javier Carrasco <javier.carrasco@...fvision.net> wrote:
> 
>> The current offset has the scale already applied to it. The ABI
>> documentation defines the offset parameter as "offset to be added
>> to <type>[Y]_raw prior to scaling by <type>[Y]_scale in order to
>> obtain value in the <type> units as specified in <type>[Y]_raw
>> documentation"
>>
>> The right value is obtained at 0 degrees Celsius by the formula provided
>> in the datasheet:
>>
>> T = Tsens_t0 + (Tadc_t - Tadc_t0) / Tadc_res
> 
> So base units for temperature are milli degrees celsius. 
> T = 1000 * (25 + (adc - 17508) / 60.1)
> 
> T =  1000/60.1 * (25 * 60.1 + adc - 17508)
> T = 10000/601 * (-16005.5 + adc)  
> So I think the maths is a little off..
>>
>> where:
>> T = 0 degrees Celsius
>> Tsens_t0 (reference temperature) = 25 degrees Celsius
>> Tadc_t0 (16-bit format for Tsens_t0) = 17508
>> Tadc_res = 60.1 LSB/degree Celsius
>>
>> The resulting offset is 16605.5, which has been truncated to 16005 to
> Interesting - the truncated value you have looks good to be but that's
> not matching with the resulting offset or the value below...
> 
You are right, there is a typo in the resulting offset and value below.

The right value is the truncated one I mentioned, which is the one
obtained with the formula.
In the case of milli degrees Celsius for T = 0 (milli or not does not
matter then):

0 = 1000 * (25 + (adc - 17508) / 60.1) =>
=> adc = -25*60.1 + 17508 = 16005.5

I will fix the value to use the one I actually obtained (16005)

Thank you.
>> provide an integer value with a precision loss smaller than the 1-LSB
>> measurement precision.
>>
>> Fix the offset to apply its value prior to scaling.
>>
>> Signed-off-by: Javier Carrasco <javier.carrasco@...fvision.net>
>> ---
>>  drivers/iio/magnetometer/tmag5273.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/iio/magnetometer/tmag5273.c b/drivers/iio/magnetometer/tmag5273.c
>> index c5e5c4ad681e..d22ca39007b6 100644
>> --- a/drivers/iio/magnetometer/tmag5273.c
>> +++ b/drivers/iio/magnetometer/tmag5273.c
>> @@ -356,7 +356,7 @@ static int tmag5273_read_raw(struct iio_dev *indio_dev,
>>  	case IIO_CHAN_INFO_OFFSET:
>>  		switch (chan->type) {
>>  		case IIO_TEMP:
>> -			*val = -266314;
>> +			*val = -16605;
>>  			return IIO_VAL_INT;
>>  		default:
>>  			return -EINVAL;
>>
>> ---
>> base-commit: 05d3ef8bba77c1b5f98d941d8b2d4aeab8118ef1
>> change-id: 20231023-topic-tmag5273x1_temp_offset-17774cbce961
>>
>> Best regards,
>