Re: [PATCH 8/9] bq27x00: Add miscdevice for each battery with ioctl for reading registers

[Lars-Peter Clausen <lars@metafoo.de>]

On 11/01/2011 01:23 PM, Pali Rohár wrote:
> 2011/11/1 Lars-Peter Clausen <lars@metafoo.de>:
>> On 11/01/2011 01:43 AM, Pali Rohár wrote:
>>> * When bq27x00_battery module is loaded it take control of bq27200 i2c battery chip and then it is not possible to use i2cget program from user space (returns -EBUSY)
>>> * This patch adds new miscdevice for each battery which has ioctl BQ27X00_READ_REG for reading battery registers from user space
>>>
>>> * ioctl cmd:
>>>   #define BQ27X00_READ_REG _IO(MISC_MAJOR, 0)
>>>
>>> * ioctl arg:
>>>   struct bq27x00_reg_parms {
>>>     int reg; /* battery register (in) */
>>>     int single; /* 1 - 8bit register, 0 - 16bit register (in) */
>>>     int ret; /* register status, negative indicate error (out) */
>>>   };
>>>
>>> Signed-off-by: Pali Rohár <pali.rohar@gmail.com>
>>
>> I'm not sure if this is such a good idea. Why do you need to read the registers
>> from userspace? Shouldn't all relevant information be available in a
>> standardized format through the power_supply userspace API? If you want to use
>> it for debugging it might be a better idea to provide a debugfs file.
>>
>> - Lars
>>
> 
> bq27200 chip has a lot of registers (for diagnostic) which is usefull
> on Nokia N900. power_supply interface is not very abstract to handle
> all of them. 

Which registers are we talking about here. And if this generally useful
information does it make sense to add a new property to the power_supply interface?

> Also on Nokia N900 is needed proprietary pogram BME for
> battery charging which need to access to bq27200 registers. A simple
> LD_PRELOAD library which wrap i2c-dev access to this device provided
> by this patch is here:
> https://code.launchpad.net/~pali/+junk/maemo_libbqioctl

Yes, but that really is a hack to work around the closed source nature of the
BME program. I don't think we should put such hacks into the kernel.