Re: Re: [PATCH -tip RESEND 2/2] ftrace: Introduce nr_saved_cmdlines I/F

[Yoshihiro YUNOMAE <yoshihiro.yunomae.ez@hitachi.com>]

Hi Namhyung,

(2014/02/17 16:13), Namhyung Kim wrote:
> Hi Yoshihiro,
> 
> On Mon, 17 Feb 2014 13:57:13 +0900, Yoshihiro YUNOMAE wrote:
>> Hi Namhyung,
>>
>> (2014/02/14 13:50), Namhyung Kim wrote:
>>> Hi Yoshihiro,
>>>
>>> On Thu, 13 Feb 2014 10:28:58 +0900, Yoshihiro YUNOMAE wrote:
>>>> Introduce nr_saved_cmdlines I/F for changing the number of pid-comm list.
>>>> saved_cmdlines can store 128 command names using SAVED_CMDLINES now, but
>>>> 'no-existing processes' names are often lost in saved_cmdlines when we
>>>> read trace data. So, by introducing nr_saved_cmdlines I/F, the rule storing
>>>> 128 command names is changed to the command numbers defined users.
>>>>
>>>> When we write a value to nr_saved_cmdlines, the number of the value will
>>>> be stored in pid-comm list:
>>>>
>>>> 	# echo 1024 > /sys/kernel/debug/tracing/nr_saved_cmdlines
>>>>
>>>> Here, 1024 command names are stored. The default number is 128 and the maximum
>>>> number is PID_MAX_DEFAULT (=32768 if CONFIG_BASE_SMALL is not set). So, if we
>>>> want to avoid to lose command names, we need to set 32768 to nr_saved_cmdlines.
>>>>
>>>> We can read the maximum number of the list:
>>>>
>>>> 	# cat /sys/kernel/debug/tracing/nr_saved_cmdlines
>>>> 	128
>>>
>>> [SNIP]
>>>> @@ -3685,7 +3760,8 @@ static void *saved_cmdlines_next(struct seq_file *m, void *v, loff_t *pos)
>>>>    
>>>>    	(*pos)++;
>>>>    
>>>> -	for (; ptr < &map_cmdline_to_pid[SAVED_CMDLINES]; ptr++) {
>>>> +	for (; ptr < &savedcmd->map_cmdline_to_pid[savedcmd->cmdline_num];
>>>> +	     ptr++) {
>>>>    		if (*ptr == -1 || *ptr == NO_CMDLINE_MAP)
>>>>    			continue;
>>>>    
>>>> @@ -3700,7 +3776,7 @@ static void *saved_cmdlines_start(struct seq_file *m, loff_t *pos)
>>>>    	void *v;
>>>>    	loff_t l = 0;
>>>>    
>>>> -	v = &map_cmdline_to_pid[0];
>>>> +	v = &savedcmd->map_cmdline_to_pid[0];
>>>>    	while (l <= *pos) {
>>>>    		v = saved_cmdlines_next(m, v, &l);
>>>>    		if (!v)
>>>
>>> Are you accessing the savecmd without trace_cmdline_lock?
>>
>> It does not need to get trace_cmdline_lock here.
>> The elements of map_pid_to_cmdline[] and saved_cmdlines[] are protected
>> by trace_cmdline_lock in trace_find_cmdline(), but on the other hand
>> map_cmdline_to_pid[] are not protected. There are no problems in
>> particular. This is because map_cmdline_to_pid[] always refers to a
>> valid process name or "<...>".
> 
> I don't get it.  What does protect the savedcmd from being changed
> during reading "saved_cmdlines" file with changing "nr_saved_cmdlines"?

Ah, that's true. As you say, this implementation can cause the
competition problem. I'll fix it in V2.

Thank you,
Yoshihiro YUNOMAE

-- 
Yoshihiro YUNOMAE
Software Platform Research Dept. Linux Technology Center
Hitachi, Ltd., Yokohama Research Laboratory
E-mail: yoshihiro.yunomae.ez@hitachi.com