Re: [PATCH v10 01/22] perf tools: Add 'perf-config' command

[Arnaldo Carvalho de Melo <acme@kernel.org>]

Em Mon, Nov 09, 2015 at 10:58:06AM -0300, Arnaldo Carvalho de Melo escreveu:
> Em Mon, Nov 09, 2015 at 11:32:54AM +0900, Taeung Song escreveu:
> > The perf configuration file contains many variables to change various
> > aspects of each of its tools, including output, disk usage, etc.
> > But looking through state of configuration is difficult and
> > there's no knowing what kind of other variables except variables
> > in perfconfig.example exist. So This patch adds 'perf-config'
> > command with '--list' option.
> > 
> >     perf config [options]
> > 
> >     display current perf config variables.
> >     # perf config
> >     or
> >     # perf config -l | --list
> 
> So, trying to use it:
> 
>   $ perf config
>   $
> 
> It does nothing, it should either list the configuration values, I think

And if I go ahead and do:

  $ perf config -h

   Usage: perf config [options]

      -l, --list            show current config variables

  $ 

Cool, tells me what is available, but then:

  $ perf config -l
  $

I suggest you print instead:

  $ perf config -l
  # Nothing configured, please check your ~/.perfconfig file
  $

This way you can redirect the output and it will always produce
something that can be used as a ~/.perfconfig file.

It will furthermore inform the user that ~/.perfconfig is the place to
add its configs.

I haven't checked, but do you have something like a :

  $ perf config --skel > ~/.perfconfig

That would produce an skeleton with the possible config variables so
that one could then use $EDITOR ~/.perfconfig and go on setting the
knobs?

Ideally that would produce something with comments over the variables
explaining what each of those options configures.

> probably showing just what is in the ~/.perfconfig files, i.e. things
> the user explicitely changed, or show an usage message, like all the
> other tools, for instance:
> 
>   $ perf stat
> 
>    Usage: perf stat [<options>] [<command>]
> 
>      -a, --all-cpus        system-wide collection from all CPUs
>        -A, --no-aggr         disable CPU count aggregation
>       -B, --big-num         print large numbers with thousands' separators
>       -C, --cpu <cpu>       list of cpus to monitor in system-wide
>  <SNIP>
> 
> 'perf evlist' alternatively does, like perf list, show what a default
> action:
> 
>   $ perf evlist
>   cycles:pp
>   $
> 
> And:
> 
>   $ perf list | head -5
>   branch-instructions OR branches                    [Hardware event]
>   branch-misses                                      [Hardware event]
>   bus-cycles                                         [Hardware event]
>   cache-misses                                       [Hardware event]
>   cache-references                                   [Hardware event]
>   $ 
> 
> While 'perf config' does nothing :-\
> 
> I suggest you do just like 'perf evlist' and 'perf list'.
> 
> - Arnaldo