linux-kernel - Re: [PATCH 2/4] perf tools: Add functions which can get or set perf config variables.

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <555010B9.905@gmail.com>
Date:	Mon, 11 May 2015 11:15:21 +0900
From:	taeung <treeze.taeung@...il.com>
To:	Jiri Olsa <jolsa@...hat.com>
CC:	Arnaldo Carvalho de Melo <acme@...nel.org>,
	Namhyung Kim <namhyung@...nel.org>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH 2/4] perf tools: Add functions which can get or set perf
 config variables.

Hi, Jiri Olsa

Thanks for your feedbacks on my patches.

There are one thing I don't understand very well.
I wrote a number 1) in the middle of the your feedbacks to mark it.

I don't follow, could you elaborate it little more?


On 05/05/2015 04:16 AM, Jiri Olsa wrote:
> On Mon, Apr 27, 2015 at 03:34:24PM +0900, Taeung Song wrote:
>> This patch consists of functions
>> which can get, set specific config variables.
>> For the syntax examples,
>>
>>     perf config [options] [section.subkey[=value] ...]
>>
>>     display key-value pairs of specific config variables
>>     # perf config report.queue-size report.children
> [jolsa@...va perf]$ ./perf config krava.krava
> krava.krava=true
>
> ?
>
> some comments below
>
>>     set specific config variables
>>     # perf config report.queue-size=100M report.children=true
>>
>> Signed-off-by: Taeung Song <treeze.taeung@...il.com>
>> ---
>>   tools/perf/Documentation/perf-config.txt |   2 +
>>   tools/perf/builtin-config.c              | 276 ++++++++++++++++++++++++++++++-
>>   tools/perf/util/cache.h                  |  17 ++
>>   tools/perf/util/config.c                 |  30 +++-
>>   4 files changed, 320 insertions(+), 5 deletions(-)
> SNIP
>
>> +static int set_spec_config(const char *section_name, const char *subkey,
>> +			   const char *value)
>>   {
>>   	int ret = 0;
>> +	ret += set_config(section_name, subkey, value);
>> +	ret += perf_configset_write_in_full();
>> +
>> +	return ret;
>> +}
>> +
>> +static void parse_key(const char *var, const char **section_name, const char **subkey)
>> +{
>> +	char *key = strdup(var);
>> +
>> +	if (!key)
>> +		die("%s: strdup failed\n", __func__);
>> +
>> +	*section_name = strsep(&key, ".");
>> +	*subkey = strsep(&key, ".");
> should this check the config syntax? could be used for command line check as well
>> +}
>> +
>> +static int collect_config(const char *var, const char *value,
>> +			  void *cb __maybe_unused)
>> +{
>> +	struct config_section *section_node;
>> +	const char *section_name, *subkey;
> SNIP
>
>> +	}
>> +	for (i = 0; key[i]; i++) {
>> +		if (i == 0 && !isalpha(key[i++]))
>> +			goto out_err;
>> +
>> +		switch (key[i]) {
>> +		case '.':
>> +			num_dot += 1;
>> +			if (!isalpha(key[++i]))
>> +				goto out_err;
>> +			break;
>> +		case '=':
>> +			num_equals += 1;
>> +			break;
>> +		default:
>> +			if (!isalpha(key[i]) && !isalnum(key[i]))
>> +				goto out_err;
> you dont allow '-' in the key report.queue-size, I think we should support also _
>
> also please put the name checks into separated function
>
>> +		}
>> +	}
>> +
>> +	if (num_equals > 1 || num_dot > 1)
>> +		goto out_err;
>> +
>> +	given_value = strchr(key, '=');
>> +	if (given_value == NULL || given_value == key)
>> +		given_value = NULL;
> SNIP
>
>>   	argc = parse_options(argc, argv, config_options, config_usage,
>>   			     PARSE_OPT_STOP_AT_NON_OPTION);
>> +	if (origin_argc > argc)
>> +		is_option = true;
>> +	else
>> +		is_option = false;
>> +
>> +	if (!is_option && argc >= 0) {
>> +		switch (argc) {
>> +		case 0:
>> +			break;
>> +		default:
>> +			for (i = 0; argv[i]; i++) {
>> +				value = strrchr(argv[i], '=');
>> +				if (value == NULL || value == argv[i])
> hum, so you let go in args like '=krava' ?
>
> why dont you completely check the name (assignment string) first
> and decide later about the callback
>
1) I understood that the name must be completely checked first.
    But I don't know the callback. What does it mean ?
    Could you elaborate it little more?

Thanks,
Taeung

>> +					ret = perf_configset_with_option(show_spec_config, argv[i]);
>> +				else
>> +					ret = perf_configset_with_option(set_spec_config, argv[i]);
>> +				if (ret < 0)
>> +					break;
>> +			}
>> +			goto out;
>> +		}
>> +	}
> SNIP
>
>> @@ -502,6 +501,31 @@ out:
>>   	return ret;
>>   }
>>   
>> +int perf_configset_write_in_full(void)
>> +{
>> +	struct config_section *section_node;
>> +	struct config_element *element_node;
>> +	const char *first_line = "# this file is auto-generated.";
> so you parse whole config, change it and write back..
> hum, I dont see better way.. and I like the first line ;-)
>
>> +	FILE *fp = fopen(config_file_name, "w");
>> +
>> +	if (!fp)
>> +		return -1;
>> +
>> +	fprintf(fp, "%s\n", first_line);
>> +	/* overwrite configvariables */
>> +	list_for_each_entry(section_node, sections, list) {
>> +		fprintf(fp, "[%s]\n", section_node->name);
>> +		list_for_each_entry(element_node, &section_node->element_head, list) {
>> +			if (element_node->value)
>> +				fprintf(fp, "\t%s = %s\n",
>> +					element_node->subkey, element_node->value);
>> +		}
>> +	}
>> +	fclose(fp);
>> +
>> +	return 0;
>> +}
>> +
>>   /*
>>    * Call this to report error for your variable that should not
>>    * get a boolean value (i.e. "[my] var" means "true").
>> -- 
>> 1.9.1
>>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/