i2c-mxs: Problems with PIO mode on iMX28

i2c-mxs: Problems with PIO mode on iMX28

[Torsten Fleischer]

Hi,

on my board there is a temperature sensor TCN75A (compatible to LM75) 
connected to the iMX283.
In the user space the temperature is periodically polled by reading the file 
/sys/bus/i2c/devices/0-0048/temp1_input.
When triggering this file the LM75 driver consecutively reads 3 registers from 
the sensor and outputs the current temperature read from register 0.

With linux 3.9.7 I sometimes get a communication timeout when reading the 
sensor.

Since several patches have been added to the driver, I tried linux 3.10-rc7. 
With this version no timeout will be reported, but sometimes the value 
returned seems to be wrong. Normally I get a value of round about 30000. But 
sometimes it takes some seconds to read the temperature and the returned value 
is 75000. This value is not the current temperature but the temperature 
hysteresis that is the content of register 2 of the sensor. 

After some investigation I found that the i2c-mxs driver sometimes times out 
while waiting for the MXS_I2C_DEBUG0_DMAREQ bit when writing data (the 
I2C_M_RD flag is not set).
Unfortunately this error is never reported to the I2C core and thus to the 
LM75 driver. So the LM75 driver assumes that it has received the correct 
values.

This problem seems to be only in the PIO mode. After changing the driver so 
that the DMA is always used it works as expected.

Best regards,
Torsten Fleischer

Re: i2c-mxs: Problems with PIO mode on iMX28

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 1825 bytes --]


CCing Marek Vasut and Pengutronix...

On Thu, Jun 27, 2013 at 07:39:45PM +0200, Torsten Fleischer wrote:
> Hi,
> 
> on my board there is a temperature sensor TCN75A (compatible to LM75) 
> connected to the iMX283.
> In the user space the temperature is periodically polled by reading the file 
> /sys/bus/i2c/devices/0-0048/temp1_input.
> When triggering this file the LM75 driver consecutively reads 3 registers from 
> the sensor and outputs the current temperature read from register 0.
> 
> With linux 3.9.7 I sometimes get a communication timeout when reading the 
> sensor.
> 
> Since several patches have been added to the driver, I tried linux 3.10-rc7. 
> With this version no timeout will be reported, but sometimes the value 
> returned seems to be wrong. Normally I get a value of round about 30000. But 
> sometimes it takes some seconds to read the temperature and the returned value 
> is 75000. This value is not the current temperature but the temperature 
> hysteresis that is the content of register 2 of the sensor. 
> 
> After some investigation I found that the i2c-mxs driver sometimes times out 
> while waiting for the MXS_I2C_DEBUG0_DMAREQ bit when writing data (the 
> I2C_M_RD flag is not set).
> Unfortunately this error is never reported to the I2C core and thus to the 
> LM75 driver. So the LM75 driver assumes that it has received the correct 
> values.
> 
> This problem seems to be only in the PIO mode. After changing the driver so 
> that the DMA is always used it works as expected.
> 
> Best regards,
> Torsten Fleischer
> --
> To unsubscribe from this list: send the line "unsubscribe linux-i2c" in
> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 836 bytes --]

Re: i2c-mxs: Problems with PIO mode on iMX28

[Alexandre Belloni]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

Hi,

It really seems like what I observed in
http://www.spinics.net/lists/linux-i2c/msg11985.html
I still didn't have a deeper look at it.

Also, ccing Fabio that seemed interested in the resoultion of that issue

On 28/06/2013 11:13, Wolfram Sang wrote:
>
> CCing Marek Vasut and Pengutronix...
>
> On Thu, Jun 27, 2013 at 07:39:45PM +0200, Torsten Fleischer wrote:
>> Hi,
>>
>> on my board there is a temperature sensor TCN75A (compatible to LM75)
>> connected to the iMX283.
>> In the user space the temperature is periodically polled by reading
the file
>> /sys/bus/i2c/devices/0-0048/temp1_input.
>> When triggering this file the LM75 driver consecutively reads 3
registers from
>> the sensor and outputs the current temperature read from register 0.
>>
>> With linux 3.9.7 I sometimes get a communication timeout when reading
the
>> sensor.
>>
>> Since several patches have been added to the driver, I tried linux
3.10-rc7.
>> With this version no timeout will be reported, but sometimes the value
>> returned seems to be wrong. Normally I get a value of round about
30000. But
>> sometimes it takes some seconds to read the temperature and the
returned value
>> is 75000. This value is not the current temperature but the temperature
>> hysteresis that is the content of register 2 of the sensor.
>>
>> After some investigation I found that the i2c-mxs driver sometimes
times out
>> while waiting for the MXS_I2C_DEBUG0_DMAREQ bit when writing data (the
>> I2C_M_RD flag is not set).
>> Unfortunately this error is never reported to the I2C core and thus
to the
>> LM75 driver. So the LM75 driver assumes that it has received the correct
>> values.
>>
>> This problem seems to be only in the PIO mode. After changing the
driver so
>> that the DMA is always used it works as expected.
>>
>> Best regards,
>> Torsten Fleischer
>> --
>> To unsubscribe from this list: send the line "unsubscribe linux-i2c" in
>> the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
>> More majordomo info at  http://vger.kernel.org/majordomo-info.html


- -- 
Alexandre Belloni, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.12 (GNU/Linux)
Comment: Using GnuPG with undefined - http://www.enigmail.net/
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=aY1b
-----END PGP SIGNATURE-----