Re: [tip:perf/core] perf probe: Clean up probe_point_lazy_walker()return value

From: Masami Hiramatsu
Date: Wed Mar 16 2011 - 06:38:34 EST


(2011/03/16 16:59), Peter Zijlstra wrote:
> On Tue, 2011-03-15 at 23:18 +0000, tip-bot for Ingo Molnar wrote:
>
>> perf probe: Clean up probe_point_lazy_walker() return value
>
>> --- a/tools/perf/util/probe-finder.c
>> +++ b/tools/perf/util/probe-finder.c
>> @@ -1328,7 +1328,7 @@ static int probe_point_lazy_walker(const char *fname, int lineno,
>> * Continue if no error, because the lazy pattern will match
>> * to other lines
>> */
>> - return ret < 0 ?: 0;
>> + return ret < 0 ? ret : 0;
>> }
>
> It is a slight change in semantics though, the return value will now be
> negative instead of 1. If its used as a boolean that's fine, but still.
>
> I'd have changed it to:
>
> return ret < 0;
>
> Which is identical to the previous statement.

Sorry this was my mistake, that statement actually aims to
what Ingo has done.

If probe_point_lazy_walker() returns non-zero value,
die_walk_lines() returns immediately, because a positive return
value means the handler finds line and wanna stop walking on
the lines, and a negative return value means some errors
occurred in the handler and must stop.

However, probe_point_lazy_walker() has to continue to walk,
because given lazy-pattern can match several lines. Thus, this
must return 0 if ret is a positive value, and return an error code
if ret is a negative value.

Thank you,


--
Masami HIRAMATSU
2nd Dept. Linux Technology Center
Hitachi, Ltd., Systems Development Laboratory
E-mail: masami.hiramatsu.pt@xxxxxxxxxxx
--
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/