Re: [Valgrind-users] Empty stack shown by Valgrind

[Tom H. <to...@co...>]

On 02/01/15 13:36, Philippe Waroquiers wrote:
> On Fri, 2015-01-02 at 17:14 +0530, Austin Einter wrote:
>
>
>> ==28501== 7 bytes in 1 blocks are indirectly lost in loss record 33 of
>> 152
>> ==28501==    at 0x402CFC7: malloc (vg_replace_malloc.c:296)
>> ==28501==    by 0x40AF985: ???
>> ==28501==    by 0x40AE2A1: ???
>> ==28501==    by 0x40ADB1B: ???
>> ==28501==    by 0x40ADAAC: ???
>> ==28501==    by 0x40C3B24: ???
>> ==28501==    by 0x40BDAE2: ???
>> ==28501==    by 0x40BDFA7: ???
>> ==28501==    by 0x40BE6F2: ???
>> ==28501==    by 0x40BF290: ???
>> ==28501==    by 0x40BF483: ???
>> ==28501==    by 0x40E63C0: ???
>> ==28501==
>>
>>
>> My query is, why stack trace is empty. In such situations how do I
>> catch where from problem origins.
> Usually, when Valgrind cannot show function name, file name and linenr,
> it is because there is no debug information for this part
> of the code.
>
> So, probably you did not compile with debug info (-g option).

Given that this is a leak report the other likely cause is that the 
addresses are in a shared object that has been unloaded, though it's 
surprising that none at all are translatable.

Otherwise I would at least expect to see the object name.

Tom

-- 
Tom Hughes (to...@co...)
http://compton.nu/