2009/9/17 Ulrich Drepper wrote:
On 09/17/2009 07:26 AM, Rakesh Pandit wrote:
> May this be reviewed ?
Why should it be needed? The program is about to finish. Any freeing
of resources will unnecessarily create work.
Even if it fixes memory leak ? In case it is not worth leave it.
I ran valgrind to check (valgrind --leak-check=full
--show-reachable=yes ./src/addr2line -S -e
/home/rakesh/ntop-3.3.10/ntop.o 0x0000000000000930) on one example
object file.
Earlier:
==24498== LEAK SUMMARY:
==24498== definitely lost: 104 bytes in 1 blocks
==24498== indirectly lost: 63,964 bytes in 17 blocks
==24498== possibly lost: 0 bytes in 0 blocks
==24498== still reachable: 1,657 bytes in 6 blocks
==24498== suppressed: 0 bytes in 0 blocks
After:
==24723== LEAK SUMMARY:
==24723== definitely lost: 0 bytes in 0 blocks
==24723== indirectly lost: 0 bytes in 0 blocks
==24723== possibly lost: 0 bytes in 0 blocks
==24723== still reachable: 32 bytes in 1 blocks
==24723== suppressed: 0 bytes in 0 blocks
There is still some loss because of absense of ebl_closebackend for
ebl_openbackend at
line 600: dwfl_module_getdwarf.c : mod->ebl = ebl_openbackend (mod->main.elf);
Is it worth to fix this one ?
Thanks,
--
Rakesh Pandit
https://fedoraproject.org/
freedom, friends, features, first