How to Remove Memory Leak

[Stefan Sauer]

On 12/15/2011 12:22 PM, Umakant Goyal wrote:
> I have run the valgrind with option
> --suppressions=gstreamer/common/gst.supp. But still valgrind output
> showing lot of memory leaks like:

Please google a bit. This is an FAQ. Also set these env-vars:
G_SLICE=always-malloc G_DEBUG=gc-friendly
Then understand what a leak is and what a one-time allocation is.

Stefan
> ==31550== 100,184 bytes in 21 blocks are possibly lost in loss record
> 3,352 of 3,355
> ==31550==    at 0x4005046: memalign (vg_replace_malloc.c:581)
> ==31550==    by 0x40050A0: posix_memalign (vg_replace_malloc.c:709)
> ==31550==    by 0x42BAD68: slab_allocator_alloc_chunk (gslice.c:1158)
> ==31550==    by 0x42BBF2B: g_slice_alloc (gslice.c:684)
> ==31550==    by 0x42BC0A4: g_slice_alloc0 (gslice.c:854)
> ==31550==    by 0x410EE46: g_type_create_instance (gtype.c:1865)
> ==31550==    by 0x40F3A11: g_object_constructor (gobject.c:1615)
> ==31550==    by 0x40F4281: g_object_newv (gobject.c:1479)
> ==31550==    by 0x40F4DC4: g_object_new_valist (gobject.c:1596)
> ==31550==    by 0x40F4EC6: g_object_new (gobject.c:1311)
> ==31550==    by 0x4055131: gst_element_factory_create (gstelementfactory.c:380)
> ==31550==    by 0x40564F2: gst_element_factory_make (gstelementfactory.c:451)
> ==31550==
> ==31550== 106,080 bytes in 52 blocks are still reachable in loss
> record 3,353 of 3,355
> ==31550==    at 0x4005046: memalign (vg_replace_malloc.c:581)
> ==31550==    by 0x40050A0: posix_memalign (vg_replace_malloc.c:709)
> ==31550==    by 0x42BAD68: slab_allocator_alloc_chunk (gslice.c:1158)
> ==31550==    by 0x42BBF2B: g_slice_alloc (gslice.c:684)
> ==31550==    by 0x42BC0A4: g_slice_alloc0 (gslice.c:854)
> ==31550==    by 0x410EE46: g_type_create_instance (gtype.c:1865)
> ==31550==    by 0x40F3A11: g_object_constructor (gobject.c:1615)
> ==31550==    by 0x40F4281: g_object_newv (gobject.c:1479)
> ==31550==    by 0x4088B60: _priv_gst_registry_chunks_load_plugin
> (gstregistrychunks.c:559)
> ==31550==    by 0x40AFE7F: gst_registry_binary_read_cache
> (gstregistrybinary.c:600)
> ==31550==    by 0x4086E33: gst_update_registry (gstregistry.c:1600)
>
> Please Suggest me how to remove this.
>
> Thanks in Advance
>
> On 12/13/11, Nicolas Dufresne <nicolas.dufresne at collabora.co.uk> wrote:
>> Le mardi 13 décembre 2011 à 16:43 +0530, Umakant Goyal a écrit :
>>
>>> Hi All,
>>>
>>> I am initializing the GStreaming using gst_init (NULL, NULL) API and
>>> then deinit the same using gst_deinit () API.
>>> I run my binary using valgrind and found that Leak Summary of Valgrind
>>> showing lot of possibly lost and still reachable in Gstreamer.  Below
>>> Sharing the some logs of valgrind:
>> When using valgrind on GStreamer you should add the option
>> --suppressions=gstreamer/common/gst.supp . This file is in the common
>> directory of each Gstreamer repositories (a git submodule). This file
>> contains description of false positives shown by valgrind.
>>
>> Nicolas
>>
> _______________________________________________
> gstreamer-devel mailing list
> gstreamer-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel