GNOME Bugzilla – Bug 708147
memory leak
Last modified: 2014-03-02 20:38:17 UTC
Created attachment 255010 [details] valgrind details I'm logged with gnome-shell, and often get gnome-mahjongg using high cpu resources on that system having 4 cores. Sometimes i need to kill the gnome-mahjongg process to get back a normal cpu/cores usage (even when gnome-mahjongg is loaded but not played). reported on launchpad: https://bugs.launchpad.net/ubuntu/+source/gnome-mahjongg/+bug/1225919
Thanks for taking the time to report this bug! That log only shows 280 bytes definitely lost; hardly a big deal. They're both GTK+ leaks: ==9092== 160 bytes in 8 blocks are definitely lost in loss record 4,554 of 4,936 ==9092== at 0x402B965: calloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so) ==9092== by 0x4A2E50A: g_malloc0 (in /lib/i386-linux-gnu/libglib-2.0.so.0.3707.0) ==9092== by 0x4641B61: ??? (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x4623268: ??? (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x45FE065: gdk_display_manager_open_display (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x45FD363: gdk_display_open (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x4B1B904: (below main) (libc-start.c:260) and ==9092== 165 (120 direct, 45 indirect) bytes in 6 blocks are definitely lost in loss record 4,556 of 4,936 ==9092== at 0x402B965: calloc (in /usr/lib/valgrind/vgpreload_memcheck-x86-linux.so) ==9092== by 0x4A2E50A: g_malloc0 (in /lib/i386-linux-gnu/libglib-2.0.so.0.3707.0) ==9092== by 0x4641AE3: ??? (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x4623268: ??? (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x45FE065: gdk_display_manager_open_display (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x45FD363: gdk_display_open (in /usr/lib/i386-linux-gnu/libgdk-3.so.0.914.0) ==9092== by 0x4B1B904: (below main) (libc-start.c:260) I am worried about the excessive CPU usage, but I don't see any evidence that it's caused by a memory leak. If you're still getting excessive CPU usage on 3.10, please open another bug for that. Thanks!