linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Masami Hiramatsu <mhiramat@kernel.org>
To: Taeung Song <treeze.taeung@gmail.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>,
	linux-kernel@vger.kernel.org, Jiri Olsa <jolsa@kernel.org>,
	Namhyung Kim <namhyung@kernel.org>,
	Ingo Molnar <mingo@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Masami Hiramatsu <mhiramat@kernel.org>
Subject: Re: [PATCH v6 4/4] perf config: Initialize perf_config_set with all default configs
Date: Mon, 4 Apr 2016 22:28:23 +0900	[thread overview]
Message-ID: <20160404222823.2d5e7add53eabf95ee0c4273@kernel.org> (raw)
In-Reply-To: <1459761427-16214-5-git-send-email-treeze.taeung@gmail.com>

On Mon,  4 Apr 2016 18:17:07 +0900
Taeung Song <treeze.taeung@gmail.com> wrote:

> To avoid duplicated config variables and
> use perf_config_set classifying between standard
> perf config variables and unknown or new config
> variables other than them, initialize perf_config_set
> with all default configs.
> 
> And this will be needed when showing all configs with
> default value or checking correct type of a config
> variable in the near future.
> 
> Cc: Namhyung Kim <namhyung@kernel.org>
> Cc: Jiri Olsa <jolsa@kernel.org>
> Signed-off-by: Taeung Song <treeze.taeung@gmail.com>
> ---
>  tools/perf/builtin-config.c | 11 +++++++----
>  tools/perf/util/config.c    | 33 ++++++++++++++++++++++++++++++---
>  tools/perf/util/config.h    |  7 +++++++
>  3 files changed, 44 insertions(+), 7 deletions(-)
> 
> diff --git a/tools/perf/builtin-config.c b/tools/perf/builtin-config.c
> index 1133224..66a269e 100644
> --- a/tools/perf/builtin-config.c
> +++ b/tools/perf/builtin-config.c
> @@ -35,23 +35,26 @@ static struct option config_options[] = {
>  
>  static int show_config(struct perf_config_set *set)
>  {
> +	bool has_value = false;
>  	struct perf_config_section *section;
>  	struct perf_config_item *item;
>  	struct list_head *sections = &set->sections;
>  
> -	if (list_empty(sections))
> -		return -1;
> -
>  	list_for_each_entry(section, sections, node) {
>  		list_for_each_entry(item, &section->items, node) {
>  			char *value = item->value;
>  
> -			if (value)
> +			if (value) {
>  				printf("%s.%s=%s\n", section->name,
>  				       item->name, value);
> +				has_value = true;
> +			}
>  		}
>  	}
>  
> +	if (!has_value)
> +		return -1;

Could this path be really executed? It seems that the set is
always initialized with default values. This means not empty.
If it is for asserting bugs, we also have to check set != NULL
beforehand. 

> +
>  	return 0;
>  }
>  
> diff --git a/tools/perf/util/config.c b/tools/perf/util/config.c
> index f937124..1855560 100644
> --- a/tools/perf/util/config.c
> +++ b/tools/perf/util/config.c
> @@ -663,6 +663,7 @@ static struct perf_config_section *add_section(struct list_head *sections,
>  	if (!section)
>  		return NULL;
>  
> +	section->is_custom = true;
>  	INIT_LIST_HEAD(&section->items);
>  	section->name = strdup(section_name);
>  	if (!section->name) {
> @@ -683,6 +684,7 @@ static struct perf_config_item *add_config_item(struct perf_config_section *sect
>  	if (!item)
>  		return NULL;
>  
> +	item->is_custom = true;
>  	item->name = strdup(name);
>  	if (!item->name) {
>  		pr_debug("%s: strdup failed\n", __func__);
> @@ -751,12 +753,35 @@ out_free:
>  	return -1;
>  }
>  
> +static struct perf_config_set *perf_config_set__init(struct perf_config_set *set)
> +{
> +	int i, j;
> +	struct perf_config_section *section;
> +	struct perf_config_item *items;
> +	struct list_head *sections = &set->sections;
> +
> +	INIT_LIST_HEAD(&set->sections);
> +
> +	for (i = 0; i != CONFIG_END; i++) {
> +		section = &default_sections[i];
> +		INIT_LIST_HEAD(&section->items);
> +
> +		items = default_config_items[i];
> +		for (j = 0; items[j].name != NULL; j++)
> +			list_add_tail(&items[j].node, &section->items);
> +
> +		list_add_tail(&section->node, sections);
> +	}
> +
> +	return set;
> +}
> +
>  struct perf_config_set *perf_config_set__new(void)
>  {
>  	struct perf_config_set *set = zalloc(sizeof(*set));
>  
>  	if (set) {
> -		INIT_LIST_HEAD(&set->sections);
> +		perf_config_set__init(set);
>  		perf_config(collect_config, set);
>  	}
>  
> @@ -776,7 +801,8 @@ static void perf_config_section__purge(struct perf_config_section *section)
>  
>  	list_for_each_entry_safe(item, tmp, &section->items, node) {
>  		list_del_init(&item->node);
> -		perf_config_item__delete(item);
> +		if (item->is_custom)
> +			perf_config_item__delete(item);
>  	}
>  }
>  
> @@ -793,7 +819,8 @@ static void perf_config_set__purge(struct perf_config_set *set)
>  
>  	list_for_each_entry_safe(section, tmp, &set->sections, node) {
>  		list_del_init(&section->node);
> -		perf_config_section__delete(section);
> +		if (section->is_custom)
> +			perf_config_section__delete(section);

Here, unless perf_config_section__delete() is called, the items
under the section is never be checked. However, if a user changes
existing item value, its section->is_custom is false, but the
item itself is_custom be true.
This means that such custom item values may not be freed.
So, the section->is_custom flag should be checked in *__delete
method as below,

static void perf_config_section__delete(struct perf_config_section *section)
{
	perf_config_section__purge(section);
	if (section->is_custom) {
		zfree(&section->name);
		free(section);
	}
}

Or, you can just call perf_config_section__purge(section) in
perf_config_set__purge().

Also, I like ->allocated instead of ->is_custom.

Thank you,

>  	}
>  }
>  
> diff --git a/tools/perf/util/config.h b/tools/perf/util/config.h
> index 84dcc1d..1df96b7 100644
> --- a/tools/perf/util/config.h
> +++ b/tools/perf/util/config.h
> @@ -14,6 +14,11 @@ enum perf_config_type {
>  	CONFIG_TYPE_STRING
>  };
>  
> +/**
> + * struct perf_config_item - element of perf's configs
> + *
> + * @is_custom: unknown or new config other than default config
> + */
>  struct perf_config_item {
>  	const char *name;
>  	char *value;
> @@ -27,11 +32,13 @@ struct perf_config_item {
>  		const char *s;
>  	} default_value;
>  	enum perf_config_type type;
> +	bool is_custom;
>  	struct list_head node;
>  };
>  
>  struct perf_config_section {
>  	const char *name;
> +	bool is_custom;
>  	struct list_head items;
>  	struct list_head node;
>  };
> -- 
> 2.5.0
> 


-- 
Masami Hiramatsu <mhiramat@kernel.org>

  reply	other threads:[~2016-04-04 13:28 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-04-04  9:17 [PATCH v6 0/4] Infrastructure code for perf-config Taeung Song
2016-04-04  9:17 ` [PATCH v6 1/4] perf config: Introduce perf_config_set class Taeung Song
2016-04-04 12:41   ` Masami Hiramatsu
2016-04-04 12:42     ` Taeung Song
2016-04-04  9:17 ` [PATCH v6 2/4] perf config: Let show_config() work with perf_config_set Taeung Song
2016-04-04 12:49   ` Masami Hiramatsu
2016-04-05  1:54     ` Taeung Song
2016-04-04  9:17 ` [PATCH v6 3/4] perf config: Prepare all default configs Taeung Song
2016-04-04  9:17 ` [PATCH v6 4/4] perf config: Initialize perf_config_set with " Taeung Song
2016-04-04 13:28   ` Masami Hiramatsu [this message]
2016-04-05  6:06     ` Taeung Song
2016-04-05 11:23       ` Masami Hiramatsu
2016-04-05 11:47         ` Taeung Song
2016-04-07 10:35         ` Taeung Song

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20160404222823.2d5e7add53eabf95ee0c4273@kernel.org \
    --to=mhiramat@kernel.org \
    --cc=acme@kernel.org \
    --cc=jolsa@kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@kernel.org \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=treeze.taeung@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).