Re: [PATCH 1/4] tools/power/cpupower: Read energy_perf_bias from sysfs

[Shuah Khan <skhan@linuxfoundation.org>]

On 10/16/20 2:37 AM, Borislav Petkov wrote:
> On Thu, Oct 15, 2020 at 11:49:32AM -0600, Shuah Khan wrote:
>> Is there a reason to move "int fd"?
> 
Sorry for a late response.

Okay. Looked odd since it didn't need changing.

> Habit from tip - we sort function-local variables in a reverse fir tree
> order. And since I'm adding cpupower_write_sysfs(), I made them look
> consistent.
> 
>>> +	numwritten = write(fd, buf, buflen - 1);
>>> +	if (numwritten < 1) {
>>> +		perror("write failed");
>>
>> Please add filename to the error message
> 
> 	perror(path);
> 
> or do you want me to build a string with an error message and filename?

Right. It will be great if you can add filename to the message.
> 
>> Please add return check for snprintf, please add a define for
>> "cpu%u/power/energy_perf_bias" since it is hardcoded in
>> read/write functions.
> 
> None of the other snprintf() calls in cpupower do that. Nothing checks
> snprintf() retval and the last part of the sysfs path is a naked string.
> 
> Why is this different?
> 

All of the other ones should be changed as such. Why add more?

thanks,
-- Shuah