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 480341 - Memory leak at fakesink (regression in cvs)
Memory leak at fakesink (regression in cvs)
Status: RESOLVED FIXED
Product: GStreamer
Classification: Platform
Component: gstreamer (core)
git master
Other Linux
: Normal critical
: 0.10.15
Assigned To: Stefan Sauer (gstreamer, gtkdoc dev)
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2007-09-25 19:29 UTC by Josep Torra Valles
Modified: 2007-09-26 18:05 UTC
See Also:
GNOME target: ---
GNOME version: ---

Attachments
Valgrind report (70.54 KB, text/plain)
2007-09-25 19:30 UTC, Josep Torra Valles 		Details
Valgrind report using cvs from 3 weeks ago (28.43 KB, text/plain)
2007-09-25 19:35 UTC, Josep Torra Valles 		Details

Description Josep Torra Valles 2007-09-25 19:29:03 UTC 
A change introduced on cvs during last 3 weeks causes a big memory leak on fakesink. 

I suspect that the change/bug is at basesink.

Can be reproduced with the following command:

G_DEBUG=gc-friendly G_SLICE=always-malloc valgrind --leak-resolution=high --leak-check=full --num-callers=20 --alignment=8 gst-launch-0.10 videotestsrc num-buffers=200 ! fakesink 2>leak.txt

The ammount of memory leak is linear dependent on num-buffers.
Comment 1 Josep Torra Valles 2007-09-25 19:30:26 UTC 
Created attachment 96191 [details]
Valgrind report

This is the valgrind report obtained with:

G_DEBUG=gc-friendly G_SLICE=always-malloc valgrind --leak-resolution=high --leak-check=full --num-callers=20 --alignment=8 gst-launch-0.10 videotestsrc num-buffers=200 ! fakesink 2>leak.txt
Comment 2 Josep Torra Valles 2007-09-25 19:35:15 UTC 
Created attachment 96193 [details]
Valgrind report using cvs from 3 weeks ago

Same valgrind command of comment #2 but using gstreamer cvs three weeks older.
Comment 3 Sebastian Dröge (slomo) 2007-09-25 19:39:35 UTC 
I saw some leaks with fakesink today too, probably basesink related... I'll provide a valgrind log tomorrow.
Comment 4 Sebastian Dröge (slomo) 2007-09-25 19:40:12 UTC 
(or could be basesrc)
Comment 5 Stefan Sauer (gstreamer, gtkdoc dev) 2007-09-25 19:40:39 UTC 
I saw this too. I will investigate.
Comment 6 Tim-Philipp Müller 2007-09-26 15:55:50 UTC 
Does this also happen if you install a GLib with full debugging symbols, ie.  libglib2.0-0-dbg on ubuntu/debian?

I've seen valgrind report false positives when using a stripped GLib (don't really understand why it happens though).
Comment 7 Michael Smith 2007-09-26 17:40:15 UTC 
Fixed with:

        * gst/gstbuffer.c: (gst_buffer_finalize):
          Make it once again possible to free GstBuffers in the default
          build.
          The poisoning scribbles on parts of the miniobject we need in
          order to free it.