Re: [PATCH v2 2/4] perf config: Finally write changed configs on config file at a time

[Arnaldo Carvalho de Melo]

Em Mon, May 08, 2017 at 08:07:40PM +0900, Taeung Song escreveu:
> Currently set_config() can be repeatedly called for each
> input config on the below case:
> 
>   $ perf config kmem.default=slab report.children=false ...
> 
> But it's a waste, so finally write changed configs at a time.
> 
> Cc: Jiri Olsa <jolsa@xxxxxxxxxx>
> Cc: Namhyung Kim <namhyung@xxxxxxxxxx>
> Signed-off-by: Taeung Song <treeze.taeung@xxxxxxxxx>
> ---
>  tools/perf/builtin-config.c | 15 ++++++++++-----
>  1 file changed, 10 insertions(+), 5 deletions(-)
> 
> diff --git a/tools/perf/builtin-config.c b/tools/perf/builtin-config.c
> index 7545966..0d38599 100644
> --- a/tools/perf/builtin-config.c
> +++ b/tools/perf/builtin-config.c
> @@ -33,8 +33,7 @@ static struct option config_options[] = {
>  	OPT_END()
>  };
>  
> -static int set_config(struct perf_config_set *set, const char *file_name,
> -		      const char *var, const char *value)
> +static int set_config(struct perf_config_set *set, const char *file_name)
>  {
>  	struct perf_config_section *section = NULL;
>  	struct perf_config_item *item = NULL;
> @@ -48,7 +47,6 @@ static int set_config(struct perf_config_set *set, const char *file_name,
>  	if (!fp)
>  		return -1;
>  
> -	perf_config_set__collect(set, file_name, var, value);
>  	fprintf(fp, "%s\n", first_line);
>  
>  	/* overwrite configvariables */
> @@ -160,6 +158,7 @@ int cmd_config(int argc, const char **argv)
>  	struct perf_config_set *set;
>  	char *user_config = mkpath("%s/.perfconfig", getenv("HOME"));
>  	const char *config_filename;
> +	bool changed = false;
>  
>  	argc = parse_options(argc, argv, config_options, config_usage,
>  			     PARSE_OPT_STOP_AT_NON_OPTION);
> @@ -227,10 +226,16 @@ int cmd_config(int argc, const char **argv)
>  
>  			if (value == NULL)
>  				ret = show_spec_config(set, var);
> -			else
> -				ret = set_config(set, config_filename, var, value);
> +			else {
> +				perf_config_set__collect(set, config_filename,
> +							 var, value);
> +				changed = true;
> +			}
>  			free(arg);
>  		}
> +
> +		if (changed)
> +			ret = set_config(set, config_filename);

Ok, this improves the situation, even fixing what seems like a bug, i.e.
that ret variable set in the if (value == NULL) case is not being
checked, the look just goes on to completion and whatever is the last
value left in 'ret' is what will be returned by cmd_config()...

I think we should fix that bug first, making it exit the loop and
perhaps show some error message, then do this 'call set_config()' only
at the end', no?

I applied the first patch, the one just renesting things.

- Arnaldo

>  	}
>  
>  	perf_config_set__delete(set);
> -- 
> 2.7.4