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 781021 - gst-plugins-base.supp: add valgrind suppressions for one off gl init allocations
gst-plugins-base.supp: add valgrind suppressions for one off gl init allocations
Status: RESOLVED FIXED
Product: GStreamer
Classification: Platform
Component: gst-plugins-base
git master
Other Linux
: Normal normal
: 1.13.1
Assigned To: GStreamer Maintainers
GStreamer Maintainers
Depends on:
Blocks:
 
 
Reported: 2017-04-07 09:34 UTC by Vincent Penquerc'h
Modified: 2018-01-24 00:49 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
gst-plugins-bad.supp: add gl suppressions for one off init allocations (1.82 KB, patch)
2017-04-07 09:35 UTC, Vincent Penquerc'h
none Details | Review
gst-plugins-bad.supp: add gl suppressions for one off init allocations (1.69 KB, patch)
2017-04-07 10:10 UTC, Vincent Penquerc'h
committed Details | Review

Description Vincent Penquerc'h 2017-04-07 09:34:28 UTC
gst-plugins-bad.supp: add gl suppressions for one off init allocations
Comment 1 Vincent Penquerc'h 2017-04-07 09:35:06 UTC
Created attachment 349438 [details] [review]
gst-plugins-bad.supp: add gl suppressions for one off init allocations
Comment 2 Sebastian Dröge (slomo) 2017-04-07 09:52:14 UTC
Review of attachment 349438 [details] [review]:

::: tests/check/gst-plugins-bad.supp
@@ +42,3 @@
+   fun:calloc
+   ...
+   fun:__pthread_once_slow

What is this leak exactly?

@@ +83,3 @@
+
+{
+   <dodgy one, seems like a xcb bug>

Did you report?

@@ +98,3 @@
+   fun:g_malloc0
+   ...
+   fun:call_init.part.0

What are these exactly? What do we leak there?
Comment 3 Vincent Penquerc'h 2017-04-07 09:57:29 UTC
I did not report. The others are from once-only initializations. I will repro and post the stacks.
Comment 4 Vincent Penquerc'h 2017-04-07 10:06:19 UTC
XCB:

==7611== 36 bytes in 1 blocks are definitely lost in loss record 980 of 1,709
==7611==    at 0x4C2AFA7: malloc (vg_replace_malloc.c:296)
==7611==    by 0x587B75B: ??? (in /usr/lib64/libxcb.so.1.1.0)
==7611==    by 0x58794E0: ??? (in /usr/lib64/libxcb.so.1.1.0)
==7611==    by 0x587AC26: ??? (in /usr/lib64/libxcb.so.1.1.0)
==7611==    by 0x587AD30: xcb_wait_for_reply (in /usr/lib64/libxcb.so.1.1.0)
==7611==    by 0x50E32C4: ??? (in /usr/lib64/libGL.so.1.2.0)
==7611==    by 0x50DE0B7: ??? (in /usr/lib64/libGL.so.1.2.0)
==7611==    by 0x50B29B0: ??? (in /usr/lib64/libGL.so.1.2.0)
==7611==    by 0x50AE180: glXQueryVersion (in /usr/lib64/libGL.so.1.2.0)
==7611==    by 0x4E6EE2A: gst_gl_context_glx_choose_format (gstglcontext_glx.c:356)
==7611==    by 0x4E4A6D1: gst_gl_context_create_thread (gstglcontext.c:1193)
==7611==    by 0x81A18E4: ??? (in /usr/lib64/libglib-2.0.so.0.4600.2)
==7611==    by 0x6BF8619: start_thread (in /usr/lib64/libpthread-2.22.so)
==7611==    by 0x856C5FC: clone (in /usr/lib64/libc-2.22.so)


I'm not sure what the first one was about. In my previous comment I was thinking of the several *_once ones, but I'll remove that one. The last suppression is, I suspect, from the dynamic loader.
Comment 5 Vincent Penquerc'h 2017-04-07 10:10:39 UTC
Created attachment 349449 [details] [review]
gst-plugins-bad.supp: add gl suppressions for one off init allocations

Minus the first pthread_once, I can't recall which test threw it right now, it was a few days ago, I'll find out when I re-enable all through valgrind and post again if I still see it.
Comment 6 Tim-Philipp Müller 2018-01-24 00:49:20 UTC
Added these to gst-plugins-base.supp, but without the trailing line referring to ld-2.22.so

commit 63a83768253bc5dbad500b506b5aca269ec46667 (HEAD -> master)
Author: Vincent Penquerc'h <vincent.penquerch@collabora.co.uk>
Date:   Fri Apr 7 10:33:40 2017 +0100

    gst-plugins-base.supp: add gl suppressions for one off init allocations
    
    Add to valgrind suppressions file.
    
    https://bugzilla.gnome.org/show_bug.cgi?id=781021