Re: [PATCH v4 5/6] connector/cn_proc: Performance improvements

[Anjali Kulkarni]

> On Jun 1, 2023, at 10:15 AM, Jakub Kicinski <kuba@xxxxxxxxxx> wrote:
> 
> On Thu, 1 Jun 2023 16:53:07 +0000 Anjali Kulkarni wrote:
>>> Is it possible to recode the sample so the format can be decided based
>>> on cmd line argument? To be honest samples are kinda dead, it'd be best
>>> if the code was rewritten to act as a selftest.  
>> 
>> Yes, I can recode to use a cmd line argument. Where would a selftest
>> be committed?
> 
> The path flow is the same as for the sample, the file just goes to
> tools/testing/selftests rather than samples/.
> 
>> This is kind of a self test in the sense that this is
>> working code  to test the other kernel code. What else is needed to
>> make it a selftest?
> 
> Not much, really. I think the requirement is to exit with a non-zero
> return code on failure, which you already do. 0 means success; 1 means
> failure; 2 means skip, IIRC.
> 
> The main work in your case would be that the selftest needs to do its
> checking and exit, so the stimuli must be triggered automatically.
> (You can use a bash script to drive the events.)

Thanks! So this will be part of the kselftest infra right? 
https://docs.kernel.org/dev-tools/kselftest.html ?