After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 560410 - pulse plugin leaks memory
pulse plugin leaks memory
Status: RESOLVED DUPLICATE of bug 567746
Product: GStreamer
Classification: Platform
Component: gst-plugins-good
0.10.11
Other Linux
: Normal normal
: git master
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2008-11-11 22:03 UTC by Götz Waschk
Modified: 2009-01-19 13:18 UTC
See Also:
GNOME target: ---
GNOME version: 2.25/2.26



Description Götz Waschk 2008-11-11 22:03:58 UTC
This is on Mandriva Cooker with gst-plugins-good 0.10.11 and pulseaudio 0.9.14-0.20081105.1mdv2009.1. I've set the audio output plugin in gstreamer-properties to pulse and pressed the test button a few times. Every button press increases the memory usage as shown in top:

25680 goetz     20   0  106m  19m  15m S  0.7  1.9   0:00.53 gstreamer-prope
25680 goetz     20   0  131m  20m  15m S  4.7  2.0   0:00.71 gstreamer-prope
25680 goetz     20   0  251m  20m  15m S  5.6  2.1   0:01.54 gstreamer-prope
25680 goetz     20   0  307m  20m  15m S  2.3  2.1   0:01.96 gstreamer-prope
Comment 1 Wim Taymans 2008-11-12 18:01:52 UTC
top is a very bad way to check for memory leaks.

That said, valgrind detects two suspicious leaks:

==24075== 51,292 (4,337 direct, 46,955 indirect) bytes in 38 blocks are definitely lost in loss record 227 of 257
==24075==    at 0x4C265AE: malloc (vg_replace_malloc.c:207)
==24075==    by 0x6473452: g_malloc (gmem.c:131)
==24075==    by 0x6488EA7: g_slice_alloc (gslice.c:824)
==24075==    by 0x60168F9: g_value_array_new (gvaluearray.c:120)
==24075==    by 0x144F6F4D: gst_pulseprobe_get_values (pulseprobe.c:351)
==24075==    by 0x40357F: (within /usr/bin/gstreamer-properties)
==24075==    by 0x4040DE: (within /usr/bin/gstreamer-properties)
==24075==    by 0x404461: (within /usr/bin/gstreamer-properties)
==24075==    by 0x6B3D465: (below main) (in /lib/libc-2.8.90.so)
==24075== 

==24075== 64 bytes in 1 blocks are definitely lost in loss record 75 of 257
==24075==    at 0x4C24384: calloc (vg_replace_malloc.c:397)
==24075==    by 0x64733F9: g_malloc0 (gmem.c:151)
==24075==    by 0x5D5E0D3: gst_element_init (gstelement.c:274)
==24075==    by 0x6013DF0: g_type_create_instance (gtype.c:1666)
==24075==    by 0x5FF85CA: g_object_constructor (gobject.c:1334)
==24075==    by 0x5FF8C12: g_object_newv (gobject.c:1211)
==24075==    by 0x5FF9766: g_object_new_valist (gobject.c:1274)
==24075==    by 0x5FF98AB: g_object_new (gobject.c:1056)
==24075==    by 0x5D601C3: gst_element_factory_create (gstelementfactory.c:405)
==24075==    by 0x5D60A6E: gst_element_factory_make (gstelementfactory.c:474)
==24075==    by 0x5DAF1BE: _gst_parse_yyparse (grammar.y:568)
==24075==    by 0x5DB06E7: _gst_parse_launch (grammar.y:837)
==24075== 

presumable caused by gnome-properties not unreffing things.
Comment 2 Götz Waschk 2008-11-13 08:05:09 UTC
I had initially reported this for pidgin as it was eating *all* my machine's memory after some chatting:
http://developer.pidgin.im/ticket/7519

So it is not just gstreamer-properties.
Comment 3 Warren Togami 2008-12-03 22:56:32 UTC
https://bugzilla.redhat.com/show_bug.cgi?id=470000#c12
Someone wrote a minimal C program that outputs a .wav file via gstreamer every 5 seconds.

pulseaudio-0.9.13-6.fc10
gstreamer-0.10.21-2.fc10
gstreamer-plugins-good-0.10.11-1.fc10
On Fedora 10, you can see it leaks a thread every time it outputs sound via gstreamer.  F-10 with gstreamer outputting to ALSA reportedly also does not leak.

pulseaudio-0.9.10-2.fc9
gstreamer-0.10.20-1.fc9
gstreamer-plugins-pulse-0.9.5-0.5.svn20070924.fc9
The same test binary built on Fedora 9 does not leak any thread.
Comment 4 Christophe Fergeau 2009-01-19 12:25:17 UTC
This looks like bug #567746
Comment 5 Frederic Crozat 2009-01-19 13:18:33 UTC
confirmed as dup of #567746

*** This bug has been marked as a duplicate of 567746 ***