retitle 584968 libssl1.0.0: memory leak in load_builtin_compressions
forwarded 584968 
http://rt.openssl.org/Ticket/Display.html?id=2561&user=guest&pass=guest
reassign 584968 libssl1.0.0 1.0.0f-1
kthxbye

Hi,

On Mon, Jun 07, 2010 at 03:41:37PM -0700, Johnny Luong wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Package: libcurl3
> Version: 7.18.2-8lenny4
> 
> It's not clear to me but the library appears to leak memory; using the
> command line tool appears to leak as well.  This is on Debian Lenny
> using the AMD64 port.
> 
> [...]

This is what valgrind reports as of libcurl3 v7.23.1-3 and libssl1.0.0 
v1.0.0f-1:

==25191== HEAP SUMMARY:
==25191==     in use at exit: 88 bytes in 3 blocks
==25191==   total heap usage: 3,016 allocs, 3,013 frees, 148,896 bytes allocated
==25191== 
==25191== 24 bytes in 1 blocks are still reachable in loss record 1 of 3
==25191==    at 0x4C27673: malloc (in 
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==25191==    by 0x65F8B83: CRYPTO_malloc (mem.c:306)
==25191==    by 0x636993B: load_builtin_compressions (ssl_ciph.c:435)
==25191==    by 0x636B7F8: SSL_COMP_get_compression_methods (ssl_ciph.c:1689)
==25191==    by 0x63708A7: SSL_library_init (ssl_algs.c:134)
==25191==    by 0x4E567F2: Curl_ossl_init (ssluse.c:689)
==25191==    by 0x4E5E1FD: curl_global_init (easy.c:229)
==25191==    by 0x40833D: ??? (in /usr/bin/curl)
==25191==    by 0x40267E: ??? (in /usr/bin/curl)
==25191==    by 0x54D3EAC: (below main) (libc-start.c:228)
==25191== 
==25191== 32 bytes in 1 blocks are still reachable in loss record 2 of 3
==25191==    at 0x4C27673: malloc (in 
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==25191==    by 0x65F8B83: CRYPTO_malloc (mem.c:306)
==25191==    by 0x66621CE: sk_new (stack.c:125)
==25191==    by 0x6369919: load_builtin_compressions (ssl_ciph.c:432)
==25191==    by 0x636B7F8: SSL_COMP_get_compression_methods (ssl_ciph.c:1689)
==25191==    by 0x63708A7: SSL_library_init (ssl_algs.c:134)
==25191==    by 0x4E567F2: Curl_ossl_init (ssluse.c:689)
==25191==    by 0x4E5E1FD: curl_global_init (easy.c:229)
==25191==    by 0x40833D: ??? (in /usr/bin/curl)
==25191==    by 0x40267E: ??? (in /usr/bin/curl)
==25191==    by 0x54D3EAC: (below main) (libc-start.c:228)
==25191== 
==25191== 32 bytes in 1 blocks are still reachable in loss record 3 of 3
==25191==    at 0x4C27673: malloc (in 
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==25191==    by 0x65F8B83: CRYPTO_malloc (mem.c:306)
==25191==    by 0x66621EC: sk_new (stack.c:127)
==25191==    by 0x6369919: load_builtin_compressions (ssl_ciph.c:432)
==25191==    by 0x636B7F8: SSL_COMP_get_compression_methods (ssl_ciph.c:1689)
==25191==    by 0x63708A7: SSL_library_init (ssl_algs.c:134)
==25191==    by 0x4E567F2: Curl_ossl_init (ssluse.c:689)
==25191==    by 0x4E5E1FD: curl_global_init (easy.c:229)
==25191==    by 0x40833D: ??? (in /usr/bin/curl)
==25191==    by 0x40267E: ??? (in /usr/bin/curl)
==25191==    by 0x54D3EAC: (below main) (libc-start.c:228)
==25191== 
==25191== LEAK SUMMARY:
==25191==    definitely lost: 0 bytes in 0 blocks
==25191==    indirectly lost: 0 bytes in 0 blocks
==25191==      possibly lost: 0 bytes in 0 blocks
==25191==    still reachable: 88 bytes in 3 blocks
==25191==         suppressed: 0 bytes in 0 blocks
==25191== 
==25191== For counts of detected and suppressed errors, rerun with: -v
==25191== ERROR SUMMARY: 0 errors from 0 contexts (suppressed: 10 from 6)

The memory leak originally reported seems to have been fixed or it can't be 
reproduced (I suspect it has been fixed in the libssl 0.9.8 -> 1.0.0 update
but I didn't test this).

The leak above is caused by the fact the the compression algorithms 
allocated in the load_builtin_compressions function do not get freed, and
has been already reported upstream [0] [1] (hence reassigning to the libssl
package).

Cheers

[0] http://rt.openssl.org/Ticket/Display.html?id=2561&user=guest&pass=guest
[1] http://rt.openssl.org/Ticket/Display.html?id=2439&user=guest&pass=guest

-- 
perl -E'$_=q;$/= @{[@_]};and s;\S+;<inidehG ordnasselA>;eg;say~~reverse'



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to