[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <06548ca9-8bc9-f0ec-7e75-1dd675972d13@roeck-us.net>
Date: Thu, 19 Jul 2018 06:19:53 -0700
From: Guenter Roeck <linux@...ck-us.net>
To: Stewart Smith <stewart@...ux.ibm.com>,
Shilpasri G Bhat <shilpa.bhat@...ux.vnet.ibm.com>,
mpe@...erman.id.au
Cc: linuxppc-dev@...ts.ozlabs.org, linux-hwmon@...r.kernel.org,
linux-kernel@...r.kernel.org, ego@...ux.vnet.ibm.com
Subject: Re: [PATCH v6 2/2] hwmon: ibmpowernv: Add attributes to
enable/disable sensor groups
On 07/18/2018 11:59 PM, Stewart Smith wrote:
> Shilpasri G Bhat <shilpa.bhat@...ux.vnet.ibm.com> writes:
>> On-Chip-Controller(OCC) is an embedded micro-processor in POWER9 chip
>> which measures various system and chip level sensors. These sensors
>> comprises of environmental sensors (like power, temperature, current
>> and voltage) and performance sensors (like utilization, frequency).
>> All these sensors are copied to main memory at a regular interval of
>> 100ms. OCC provides a way to select a group of sensors that is copied
>> to the main memory to increase the update frequency of selected sensor
>> groups. When a sensor-group is disabled, OCC will not copy it to main
>> memory and those sensors read 0 values.
>
> OCC is an implementation detail rather than a core part of this firmware
> API.
>
> Why not something like this:
>
> OPAL firmware provides the facility for some groups of sensors to be
> enabled/disabled at runtime to give the user the option of using the
> system resources for collecting these sensors or not.
>
> For example, on POWER9 systems, the On Chip Controller (OCC) gathers
> various system and chip level sensors and maintains their values in main
> memory.
>
>
>> +static int init_sensor_group_data(struct platform_device *pdev,
>> + struct platform_data *pdata)
>> +{
>> + struct sensor_group_data *sgrp_data;
>> + struct device_node *groups, *sgrp;
>> + enum sensors type;
>> + int count = 0, ret = 0;
>> +
>> + groups = of_find_node_by_path("/ibm,opal/sensor-groups");
>> + if (!groups)
>> + return ret;
>
> Why not look for the compatible property?
>
>
For both, I don't really care either way. Can you folks get to an agreement
and let me know after you decided ?
Thanks,
Guenter
Powered by blists - more mailing lists