Re: [PATCH v2] cpupower : Handle set and info subcommands correctly

[shuah <shuah@kernel.org>]

On 10/14/19 3:38 AM, Abhishek wrote:
> Hi Shuah,
> 
> 
> On 10/04/2019 03:45 AM, shuah wrote:
>> Hi Abhishek,
>>
>> On 10/3/19 8:38 AM, shuah wrote:
>>> On 9/13/19 2:07 AM, Abhishek Goel wrote:
>>>> Cpupower tool has set and info options which are being used only by
>>>> x86 machines. This patch removes support for these two subcommands
>>>> from generic cpupower utility. Thus, these two subcommands will now be
>>>> available only for intel.
>>>> This removes the ambiguous error message while using set option in case
>>>> of using non-intel systems.
>>>>
>>>> Without this patch on a non-intel box:
>>>>
>>>> root@ubuntu:~# cpupower info
>>>> System does not support Intel's performance bias setting
>>>>
>>>> root@ubuntu:~# cpupower set -b 10
>>>> Error setting perf-bias value on CPU
>>>>
>>>> With this patch on a non-intel box:
>>>>
>>>> root@ubuntu:~# cpupower info
>>>> Supported commands are:
>>>>          frequency-info
>>>>          frequency-set
>>>>          idle-info
>>>>          idle-set
>>>>          monitor
>>>>          help
>>>>
>>>> Same result for set subcommand.
>>>>
>>>> This patch does not affect results on a intel box.
>>>>
>>>> Signed-off-by: Abhishek Goel <huntbag@linux.vnet.ibm.com>
>>>> Acked-by: Thomas Renninger <trenn@suse.de>
>>>> ---
>>>>
>>>> changes from v1:
>>>>     Instead of bailing out early in set and info commands, in V2, we
>>>>     are cutting out support for these two commands for non-intel
>>>>     systems.
>>>
>>> thanks. I will get this in for 5.4-rc3 veru likely. Definitely in 5.4
>>>
>>
>> Okay I almost applied this and decided it needs improvements.
>>
>> I don't like using #if defined(__x86_64__) || defined(__i386__)
>>
>> tools/power/cpupower/utils/cpupower.c main() already does this
>> dynamically using uname(). Please use the same logic do this,
>> instead of adding compile time code.
>>
>> thanks,
>> -- Shuah
> 
> Do you want the decision to be taken in cpupower-set and cpupower-info
> file as was done in v1 but using uname() by identifying the architecture
> there itself?
> 

As I recall my objection to the v1 was that it is making the call very
early and bails out. Ideally, I would like to see the change as part
of set/info handling and then print out appropriate message.

thanks,
-- Shuah