Re: [PATCH 2/2] hwmon: (brcmstb) Add driver for Broadcom STB DPFE

From: Florian Fainelli
Date: Tue Apr 18 2017 - 20:28:16 EST


On 04/18/2017 05:15 PM, Markus Mayer wrote:
> On 18 April 2017 at 15:47, Guenter Roeck <linux@xxxxxxxxxxxx> wrote:
>> Hi Florian,
>>
>> On Tue, Apr 18, 2017 at 03:29:55PM -0700, Florian Fainelli wrote:
>>> Hey Guenter,
>>>
>>> On 04/18/2017 01:58 PM, Guenter Roeck wrote:
>>>> Hi Markus,
>>>>
>>>> On Tue, Apr 18, 2017 at 01:17:02PM -0700, Markus Mayer wrote:
>>>>> From: Markus Mayer <mmayer@xxxxxxxxxxxx>
>>>>>
>>>>> This driver allows access to DRAM properties, such as the refresh rate,
>>>>> via the Broadcom STB DDR PHY Front End (DPFE). The refresh rate can be
>>>>> used as indirect indicator of the DRAM temperature.
>>>>>
>>>>> The driver also allows setting of the sampling interval.
>>>>>
>>>>> Signed-off-by: Markus Mayer <mmayer@xxxxxxxxxxxx>
>>>>> ---
>>>> [ ... ]
>>>>
>>>>> +
>>>>> +static SENSOR_DEVICE_ATTR(dpfe_info, 0444, show_info, NULL, 1000);
>>>>> +static SENSOR_DEVICE_ATTR(dpfe_refresh, 0644, show_refresh, store_refresh,
>>>>> + 1000);
>>>>> +static SENSOR_DEVICE_ATTR(dpfe_vendor, 0444, show_vendor, NULL, 1000);
>>>>> +static struct attribute *dpfe_attrs[] = {
>>>>> + &sensor_dev_attr_dpfe_info.dev_attr.attr,
>>>>> + &sensor_dev_attr_dpfe_refresh.dev_attr.attr,
>>>>> + &sensor_dev_attr_dpfe_vendor.dev_attr.attr,
>>>>> + NULL
>>>>> +};
>>>>> +ATTRIBUTE_GROUPS(dpfe);
>>>>> +
>>>>
>>>> There is not a single standard hwmon attribute. I don't know how
>>>> to classify this driver, and where it should reside, but it is not
>>>> a hwmon driver.
>>>
>>> This is a driver that talks to an embedded CPU running firmware which is
>>> capable of giving various informations about the DRAM chip being
>>> populated, including a temperature trend (hotter or cooler). We thought
>>> initially we would be able to expose the actual temperature, but this in
>>> turn required a lot more knowledge about the DRAM chip that we wish we
>>> knew about. That is sort of where and why this driver was proposed for
>>> hwmon.
>>>
>>> Which subsystem do you think would be best for this driver drivers/misc/
>>> or drivers/soc/bcm/brcmstb/ maybe?
>>
>> Both should work. I would probably try misc first and let Greg tell me
>> which way to go ;-).
>
> Thanks for the tip. As Florian said, it was not the idea to submit a
> completely non-standard hwmon driver. I was sure I would be able to
> report at least some standard hwmon data also.

Good thing with drivers/soc/bcm/brcmstb/ is that I can take such patches
directly through ARM SoC pull requests, and since this is inherently
Broadcom STB specific, that sounds like the most ideal location IMHO.
--
Florian