Re: [PATCH 04/22] tools lib traceevent: Add jbd2 plugin

[Namhyung Kim]

2013-11-21 (ë), 12:01 +0100, Jiri Olsa:
> Backporting jbd2 plugin.
> 
> Backported from Steven Rostedt's trace-cmd repo (HEAD 0f2c2fb):
> git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/trace-cmd.git
> 
> This plugin adds field resolving functions for following
> tracepoint events:
>   jbd2:jbd2_checkpoint_stats
>   jbd2:jbd2_run_stats
> 
> The diff of 'perf script' output generated by old and new code:
> (data was generated by 'perf record -e 'jbd2:jbd2_run_stats,jbd2:jbd2_checkpoint_stats' -a')
> 
> --- script.jbd2.old
> +++ script.jbd2.new
> -     jbd2/dm-3-8   576 [000]  2983.748423: jbd2:jbd2_checkpoint_stats: [FAILED TO PARSE] dev=265289731 tid=0x3f2bbb chp_time=0x0 forced_to_close=0 written=0 dropped=2
> +     jbd2/dm-3-8   576 [000]  2983.748423: jbd2:jbd2_checkpoint_stats: dev 253,3 tid 4139963 chp_time 0 forced_to_close 0 written 0 dropped 2
> -     jbd2/dm-3-8   576 [000]  2983.852789: jbd2:jbd2_run_stats: [FAILED TO PARSE] dev=265289731 tid=0x3f2bc0 wait=0x0 request_delay=0x0 running=0x138d locked=0x0 flushing=0x0 logging=0x68 handle_count=2995 blocks=17 blocks_logged=18
> +     jbd2/dm-3-8   576 [000]  2983.852789: jbd2:jbd2_run_stats: dev 253,3 tid 4139968 wait 0 request_delay 0 running 5005 locked 0 flushing 0 logging 104 handle_count 2995 blocks 17 blocks_logged 18
> 
> This plugin also adds jbd2_dev_to_name function, which
> was removed from jdb2 tracepoints recently. Keeping it
> in for backward compatibility reasons.

[SNIP]
> +#define MINORBITS	20
> +#define MINORMASK	((1U << MINORBITS) - 1)
> +
> +#define MAJOR(dev)	((unsigned int) ((dev) >> MINORBITS))
> +#define MINOR(dev)	((unsigned int) ((dev) & MINORMASK))
> +
> +unsigned long long process_jbd2_dev_to_name(struct trace_seq *s,
> +					    unsigned long long *args)

Isn't it better to make these functions static?


> +{
> +	unsigned int dev = args[0];
> +
> +	trace_seq_printf(s, "%d:%d", MAJOR(dev), MINOR(dev));
> +	return 0;
> +}
> +
> +unsigned long long process_jiffies_to_msecs(struct trace_seq *s,
> +					    unsigned long long *args)
> +{
> +	unsigned long long jiffies = args[0];
> +
> +	trace_seq_printf(s, "%lld", jiffies);
> +	return jiffies;
> +}
> +
> +int PEVENT_PLUGIN_LOADER(struct pevent *pevent)
> +{
> +	pevent_register_print_function(pevent,
> +				       process_jbd2_dev_to_name,
> +				       PEVENT_FUNC_ARG_STRING,

Actually the function returns long long not string.  But it seems the
current code doesn't care about the return type.


> +				       "jbd2_dev_to_name",
> +				       PEVENT_FUNC_ARG_INT,
> +				       PEVENT_FUNC_ARG_VOID);
> +
> +	pevent_register_print_function(pevent,
> +				       process_jiffies_to_msecs,
> +				       PEVENT_FUNC_ARG_LONG,
> +				       "jiffies_to_msecs",
> +				       PEVENT_FUNC_ARG_LONG,
> +				       PEVENT_FUNC_ARG_VOID);
> +	return 0;
> +}

Shouldn't it unregister these functions when unloaded?

Thanks,
Namhyung


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