driver for mic184 chip

Frank Beesley frank.beesley at aeroflex.com
Fri Jan 28 23:47:19 CET 2005


>Hi Frank,
>
>  
>
>>Please find attached the source file (gzip=4k, native=14.3k) for a 
>>mic184 sensor. We have an older driver for the 2.4.2x kernel but it 
>>doesn't completely match what your are asking for. I ported it to
>>kernel  2.6.9 using your guides, modified i/o to match what you have
>>requested,  and have tested to work on our hardware configuration.
>>Please let me  know of any changes that need to be made and when it
>>might incorporated  into the standard kernel distribution as we will
>>be moving forward.
>>    
>>
>
>Never heard of that chip before, but it is so similar to the lm75 that
>it doesn't make any sense to me to add a different driver to the kernel
>tree to handle it.
>
>  
>


This chip is from Micrel (http://www.micrel.com/_PDF/mic184.pdf) and is 
a local/remote thermal supervisor. The data sheet states "This device is 
a pin-for-pin and and software compatible upgrade for the industry 
standard LM75." While this is true, the LM75 driver failed to work as we 
needed when I tested with it. Since we had an older (kernel 2.4.25) 
driver that did work, I chose to bring it forward instead of 
debugging/modifying the LM75.



>The only difference your driver really makes (that I can see) is that it
>exposes the configuration register to userspace. I don't see much
>interest in doing so, since the value isn't anything standard so it
>won't help the regular user. Randomly changing the configuration of the
>fly isn't exactly clever either, and I can see that you probably never
>did that yourself since this part of your driver is obviously broken.
>
>So I don't see much added value to your driver when compared with the
>existing, well tested lm75 driver.
>
>  
>

I would like to know what is broke about the config register (unless I 
sent you the wrong version) because I have in fact used it. By default 
the config register gets loaded with 0x60 (I don't recall what all this 
configures but it is the default we have used forever) and when our 
application prepares to use it we set it to 0x40.

Additional features include (beyond the LM75): remote temp measurement 
and interrupt status and mask bits for sw polling.

If no one needs it, that is fine. Just trying to contribute back.


>>I am also working on drivers for ad5321, pca9540, adm1030, and x24165 
>>(all for kernel 2.6.9). We also use DS18B20 and DS2401 one wire
>>devices  that I am trying to get working on a ISA bus (through a
>>CPLD). If you  are interested in these let me know and I will forward
>>them to you when  they are working.
>>    
>>
>
>Please note that we already have a driver for the ADM1030 in the 2.6
>kernel tree.
>
>There is a preliminary driver for the PCA9540 chip in lm_sensor (for 2.4
>kernels) but such multiplexer chips would need to integrate deeped into
>the i2c bus architecture instead of being simple chip drivers like I
>did. We have plans to do that, part of which are already implemented in
>lm_sensors, but nothing functional yet as some points of implementation
>still need to be discussed (wish I had more time...)
>
>For one wire devices, you should take a look at the w1 subsystem which
>was recently introduced in the 2.6 kernels by Evgeniy Polyakov.
>
>  
>

I am looking at the ADM1030/1031 driver currently and having some issues 
with it that I need to resolve. It appears that no tested it or only had 
limited configurations for it.

I have briefly looked at w1 from Evgeniy but not to deep yet because of 
the other drivers.




More information about the lm-sensors mailing list