Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.kernel > #1567221 > unrolled thread

[PATCH 2/3] perf ftrace: Add ftrace.tracer config option

Started byTaeung Song <treeze.taeung@gmail.com>
First post2017-01-26 10:40 +0100
Last post2017-01-27 05:00 +0100
Articles 4 — 2 participants

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  [PATCH 2/3] perf ftrace: Add ftrace.tracer config option Taeung Song <treeze.taeung@gmail.com> - 2017-01-26 10:40 +0100
    Re: [PATCH 2/3] perf ftrace: Add ftrace.tracer config option Arnaldo Carvalho de Melo <acme@kernel.org> - 2017-01-26 20:00 +0100
      Re: [PATCH 2/3] perf ftrace: Add ftrace.tracer config option Taeung Song <treeze.taeung@gmail.com> - 2017-01-27 04:10 +0100
      Re: [PATCH 2/3] perf ftrace: Add ftrace.tracer config option Taeung Song <treeze.taeung@gmail.com> - 2017-01-27 05:00 +0100

#1567221 — [PATCH 2/3] perf ftrace: Add ftrace.tracer config option

FromTaeung Song <treeze.taeung@gmail.com>
Date2017-01-26 10:40 +0100
Subject[PATCH 2/3] perf ftrace: Add ftrace.tracer config option
Message-ID<t3OH8-5Cz-11@gated-at.bofh.it>
Currently perf ftrace command will select 'function_graph' or 'function'.
So add ftrace.tracer config option to select tracer

    # cat ~/.perfconfig
    [ftrace]
        tracer = function

    # perf ftrace usleep 123456 | head -10
      <...>-14450 [002] d... 10089.284231: finish_task_switch <-__schedule
      <...>-14450 [002] .... 10089.284232: finish_wait <-pipe_wait
      <...>-14450 [002] .... 10089.284232: mutex_lock <-pipe_wait
      <...>-14450 [002] .... 10089.284232: _cond_resched <-mutex_lock
      <...>-14450 [002] .... 10089.284233: generic_pipe_buf_confirm <-pipe_read

Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Taeung Song <treeze.taeung@gmail.com>
---
 tools/perf/builtin-ftrace.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
index 414444d..8df5416 100644
--- a/tools/perf/builtin-ftrace.c
+++ b/tools/perf/builtin-ftrace.c
@@ -17,6 +17,7 @@
 #include "evlist.h"
 #include "target.h"
 #include "thread_map.h"
+#include "util/config.h"
 
 
 #define DEFAULT_TRACER  "function_graph"
@@ -198,6 +199,23 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
 	return done ? 0 : -1;
 }
 
+static int perf_ftrace_config(const char *var, const char *value, void *cb)
+{
+	struct perf_ftrace *ftrace = cb;
+
+	if (!strcmp(var, "ftrace.tracer")) {
+		if (!strcmp(value, "function_graph"))
+			ftrace->tracer = DEFAULT_TRACER;
+		else if (!strcmp(value, "function"))
+			ftrace->tracer = "function";
+		else
+			return -1;
+		return 0;
+	}
+
+	return 0;
+}
+
 int cmd_ftrace(int argc, const char **argv, const char *prefix __maybe_unused)
 {
 	int ret;
@@ -218,6 +236,8 @@ int cmd_ftrace(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_END()
 	};
 
+	perf_config(perf_ftrace_config, &ftrace);
+
 	argc = parse_options(argc, argv, ftrace_options, ftrace_usage,
 			    PARSE_OPT_STOP_AT_NON_OPTION);
 	if (!argc)
-- 
2.7.4

[toc] | [next] | [standalone]


#1567654

FromArnaldo Carvalho de Melo <acme@kernel.org>
Date2017-01-26 20:00 +0100
Message-ID<t3Xr4-2rI-29@gated-at.bofh.it>
In reply to#1567221
Em Thu, Jan 26, 2017 at 06:35:38PM +0900, Taeung Song escreveu:
> Currently perf ftrace command will select 'function_graph' or 'function'.
> So add ftrace.tracer config option to select tracer
> 
>     # cat ~/.perfconfig
>     [ftrace]
>         tracer = function
> 
>     # perf ftrace usleep 123456 | head -10
>       <...>-14450 [002] d... 10089.284231: finish_task_switch <-__schedule
>       <...>-14450 [002] .... 10089.284232: finish_wait <-pipe_wait
>       <...>-14450 [002] .... 10089.284232: mutex_lock <-pipe_wait
>       <...>-14450 [002] .... 10089.284232: _cond_resched <-mutex_lock
>       <...>-14450 [002] .... 10089.284233: generic_pipe_buf_confirm <-pipe_read
> 
> Cc: Jiri Olsa <jolsa@kernel.org>
> Cc: Namhyung Kim <namhyung@kernel.org>
> Signed-off-by: Taeung Song <treeze.taeung@gmail.com>
> ---
>  tools/perf/builtin-ftrace.c | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)
> 
> diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
> index 414444d..8df5416 100644
> --- a/tools/perf/builtin-ftrace.c
> +++ b/tools/perf/builtin-ftrace.c
> @@ -17,6 +17,7 @@
>  #include "evlist.h"
>  #include "target.h"
>  #include "thread_map.h"
> +#include "util/config.h"
>  
>  
>  #define DEFAULT_TRACER  "function_graph"
> @@ -198,6 +199,23 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
>  	return done ? 0 : -1;
>  }
>  
> +static int perf_ftrace_config(const char *var, const char *value, void *cb)
> +{
> +	struct perf_ftrace *ftrace = cb;
> +
> +	if (!strcmp(var, "ftrace.tracer")) {
> +		if (!strcmp(value, "function_graph"))
> +			ftrace->tracer = DEFAULT_TRACER;
> +		else if (!strcmp(value, "function"))
> +			ftrace->tracer = "function";
> +		else
> +			return -1;

Please warn the user for invalid values and either just say that it is
ignoring that setting or return -1 to make perf_config() return that,
then make cmd_ftrace() check the return of perf_config() and bail out if
you're not ignoring the config value.

I think its better to do a: "pr_err()" in that "return -1" branch and
make 'perf ftrace' fail, so that the user can fix its config before
proceeding.

- Arnaldo

> +		return 0;
> +	}
> +
> +	return 0;
> +}
> +
>  int cmd_ftrace(int argc, const char **argv, const char *prefix __maybe_unused)
>  {
>  	int ret;
> @@ -218,6 +236,8 @@ int cmd_ftrace(int argc, const char **argv, const char *prefix __maybe_unused)
>  	OPT_END()
>  	};
>  
> +	perf_config(perf_ftrace_config, &ftrace);
> +
>  	argc = parse_options(argc, argv, ftrace_options, ftrace_usage,
>  			    PARSE_OPT_STOP_AT_NON_OPTION);
>  	if (!argc)
> -- 
> 2.7.4

[toc] | [prev] | [next] | [standalone]


#1567875

FromTaeung Song <treeze.taeung@gmail.com>
Date2017-01-27 04:10 +0100
Message-ID<t455f-7eB-1@gated-at.bofh.it>
In reply to#1567654
Hi, Arnaldo :)

On 01/27/2017 03:58 AM, Arnaldo Carvalho de Melo wrote:
> Em Thu, Jan 26, 2017 at 06:35:38PM +0900, Taeung Song escreveu:
>> Currently perf ftrace command will select 'function_graph' or 'function'.
>> So add ftrace.tracer config option to select tracer
>>
>>     # cat ~/.perfconfig
>>     [ftrace]
>>         tracer = function
>>
>>     # perf ftrace usleep 123456 | head -10
>>       <...>-14450 [002] d... 10089.284231: finish_task_switch <-__schedule
>>       <...>-14450 [002] .... 10089.284232: finish_wait <-pipe_wait
>>       <...>-14450 [002] .... 10089.284232: mutex_lock <-pipe_wait
>>       <...>-14450 [002] .... 10089.284232: _cond_resched <-mutex_lock
>>       <...>-14450 [002] .... 10089.284233: generic_pipe_buf_confirm <-pipe_read
>>
>> Cc: Jiri Olsa <jolsa@kernel.org>
>> Cc: Namhyung Kim <namhyung@kernel.org>
>> Signed-off-by: Taeung Song <treeze.taeung@gmail.com>
>> ---
>>  tools/perf/builtin-ftrace.c | 20 ++++++++++++++++++++
>>  1 file changed, 20 insertions(+)
>>
>> diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
>> index 414444d..8df5416 100644
>> --- a/tools/perf/builtin-ftrace.c
>> +++ b/tools/perf/builtin-ftrace.c
>> @@ -17,6 +17,7 @@
>>  #include "evlist.h"
>>  #include "target.h"
>>  #include "thread_map.h"
>> +#include "util/config.h"
>>
>>
>>  #define DEFAULT_TRACER  "function_graph"
>> @@ -198,6 +199,23 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
>>  	return done ? 0 : -1;
>>  }
>>
>> +static int perf_ftrace_config(const char *var, const char *value, void *cb)
>> +{
>> +	struct perf_ftrace *ftrace = cb;
>> +
>> +	if (!strcmp(var, "ftrace.tracer")) {
>> +		if (!strcmp(value, "function_graph"))
>> +			ftrace->tracer = DEFAULT_TRACER;
>> +		else if (!strcmp(value, "function"))
>> +			ftrace->tracer = "function";
>> +		else
>> +			return -1;
>
> Please warn the user for invalid values and either just say that it is
> ignoring that setting or return -1 to make perf_config() return that,
> then make cmd_ftrace() check the return of perf_config() and bail out if
> you're not ignoring the config value.
>
> I think its better to do a: "pr_err()" in that "return -1" branch and
> make 'perf ftrace' fail, so that the user can fix its config before
> proceeding.
>

I understood it!
I'll modify this patch and resend it. :)

Thanks,
Taeung

[toc] | [prev] | [next] | [standalone]


#1567887

FromTaeung Song <treeze.taeung@gmail.com>
Date2017-01-27 05:00 +0100
Message-ID<t45RD-7A4-11@gated-at.bofh.it>
In reply to#1567654

On 01/27/2017 03:58 AM, Arnaldo Carvalho de Melo wrote:
> Em Thu, Jan 26, 2017 at 06:35:38PM +0900, Taeung Song escreveu:
>> Currently perf ftrace command will select 'function_graph' or 'function'.
>> So add ftrace.tracer config option to select tracer
>>
>>     # cat ~/.perfconfig
>>     [ftrace]
>>         tracer = function
>>
>>     # perf ftrace usleep 123456 | head -10
>>       <...>-14450 [002] d... 10089.284231: finish_task_switch <-__schedule
>>       <...>-14450 [002] .... 10089.284232: finish_wait <-pipe_wait
>>       <...>-14450 [002] .... 10089.284232: mutex_lock <-pipe_wait
>>       <...>-14450 [002] .... 10089.284232: _cond_resched <-mutex_lock
>>       <...>-14450 [002] .... 10089.284233: generic_pipe_buf_confirm <-pipe_read
>>
>> Cc: Jiri Olsa <jolsa@kernel.org>
>> Cc: Namhyung Kim <namhyung@kernel.org>
>> Signed-off-by: Taeung Song <treeze.taeung@gmail.com>
>> ---
>>  tools/perf/builtin-ftrace.c | 20 ++++++++++++++++++++
>>  1 file changed, 20 insertions(+)
>>
>> diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
>> index 414444d..8df5416 100644
>> --- a/tools/perf/builtin-ftrace.c
>> +++ b/tools/perf/builtin-ftrace.c
>> @@ -17,6 +17,7 @@
>>  #include "evlist.h"
>>  #include "target.h"
>>  #include "thread_map.h"
>> +#include "util/config.h"
>>
>>
>>  #define DEFAULT_TRACER  "function_graph"
>> @@ -198,6 +199,23 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
>>  	return done ? 0 : -1;
>>  }
>>
>> +static int perf_ftrace_config(const char *var, const char *value, void *cb)
>> +{
>> +	struct perf_ftrace *ftrace = cb;
>> +
>> +	if (!strcmp(var, "ftrace.tracer")) {
>> +		if (!strcmp(value, "function_graph"))
>> +			ftrace->tracer = DEFAULT_TRACER;
>> +		else if (!strcmp(value, "function"))
>> +			ftrace->tracer = "function";
>> +		else
>> +			return -1;
>
> Please warn the user for invalid values and either just say that it is
> ignoring that setting or return -1 to make perf_config() return that,
> then make cmd_ftrace() check the return of perf_config() and bail out if
> you're not ignoring the config value.
>
> I think its better to do a: "pr_err()" in that "return -1" branch and
> make 'perf ftrace' fail, so that the user can fix its config before
> proceeding.
>


Arnaldo,

I modified this patch as below.

diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
index 8df5416..00e228f 100644
--- a/tools/perf/builtin-ftrace.c
+++ b/tools/perf/builtin-ftrace.c
@@ -208,8 +208,11 @@ static int perf_ftrace_config(const char *var, 
const char *value, void *cb)
                          ftrace->tracer = DEFAULT_TRACER;
                  else if (!strcmp(value, "function"))
                          ftrace->tracer = "function";
-                else
+                else {
+                        pr_err("Please select function_graph(default)"
+                               "or function to use tracer.\n");
                          return -1;
+                }
                  return 0;
          }

@@ -236,7 +239,9 @@ int cmd_ftrace(int argc, const char **argv, const 
char *prefix __maybe_unused)
          OPT_END()
          };

-        perf_config(perf_ftrace_config, &ftrace);
+        ret = perf_config(perf_ftrace_config, &ftrace);
+        if (ret < 0)
+                return -1;

          argc = parse_options(argc, argv, ftrace_options, ftrace_usage,
                              PARSE_OPT_STOP_AT_NON_OPTION);


But if setting wrong config value,
existing error message is printed in perf_config().

So output can be printed as follows

1) Two error messages

     # perf ftrace usleep 123456
     Please select function_graph(default)or function to use tracer.
     Error: wrong config key-value pair ftrace.tracer=functionds

2) Or only one error message in perf_config()
(if we don't "pr_err()" in that "return -1" branch)

     # perf ftrace usleep 123456
     Error: wrong config key-value pair ftrace.tracer=functionds


Which do you like better, 1) or 2) ?

Thanks,
Taeung

[toc] | [prev] | [standalone]


Back to top | Article view | linux.kernel


csiph-web