[erlang-questions] Tracking memory leaks in NIF/Linkedin drivers with Valgrind (or anything else)

[Zabrane Mickael]

Hi Jon,

On Mar 18, 2013, at 5:53 PM, Jon Schneider wrote:

>> /opt/r16b/usr/lib/erlang/erts-5.10.1/bin/beam.valgrind.smp)
>> ==59400==    by 0x10000191F: main (in
>> /opt/r16b/usr/lib/erlang/erts-5.10.1/bin/beam.valgrind.smp)
>> ==59400==
>> ==59400== ERROR SUMMARY: 460 errors from 15 contexts (suppressed: 0 from
>> 0)
>> How can I distinguish between these falsy reported numbers and the real
>> ones?
> 
> Isn't the information you require in the return
> address/function/executable reported even if this needs a little
> post-processing ?

Yet it is. But it's mixed with a lot of false positive (see crypto example).
And if I can avoid post-processing, that should be great.

>> Finally, is there any other way than Valgrind?
> 
> Yes by instrumenting your code.

DTrace? SystemTap? How?
Can you elaborate with a concret example?

Regards,
Zabrane



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://erlang.org/pipermail/erlang-questions/attachments/20130318/6f158237/attachment.htm>