HeapTrack Query - Stripped Binary

[sindhu krishnan]

Attaching the backtrace of the crash:

Program terminated with signal SIGSEGV, Segmentation fault.
#0  0xb3466b3e in _ULarm_step () from /usr/lib/libunwind.so.8
[Current thread is 1 (LWP 30383)]
(gdb) bt
#0  0xb3466b3e in _ULarm_step () from /usr/lib/libunwind.so.8
#1  0xb3466110 in unw_backtrace () from /usr/lib/libunwind.so.8
#2  0xb3f68f32 in heaptrack_malloc () from
/usr/lib/heaptrack/libheaptrack_preload.so
#3  0xb3f67582 in posix_memalign () from
/usr/lib/heaptrack/libheaptrack_preload.so
#4  0xb393b1d2 in ?? () from /usr/lib/libglib-2.0.so.0
Backtrace stopped: previous frame identical to this frame (corrupt stack?)
(gdb)

Thanks

On Tue, Jun 13, 2023 at 11:29 AM sindhu krishnan
<sindhu.muthu25 at gmail.com> wrote:
>
> Hi,
> Recently, I started using heaptrack to detect memory leaks in my
> application. However, I encountered a problem when running heaptrack
> on my program. It crashed, and when I debugged it using gdb, I found
> that it was failing in the resolveAddress function in
> heaptrack_interpret.cpp.
>
> I see that heaptrack requires an unstripped binary to properly print
> the frames. Unfortunately, adding debug symbols would significantly
> increase the size of the binary, which is not desirable. As a
> workaround, I thought of copying the debug symbols to a USB drive and
> mounting it, allowing heaptrack to fetch the debug symbols from there.
> I am using the Yocto build system.
>
> I would appreciate your guidance or suggestions on whether it is
> possible to use heaptrack with a stripped binary. This would greatly
> assist me in resolving the issue.
>
> Thanks,
> Sindhu