Slackware64-15.0: Memory leak in kwin_x11?
by PJBrs from LinuxQuestions.org on (#6GS84)
Hi, I was wondering if anyone else suspects a memory leak in kwin_x11... I've tracked kwin_x11 memory use for some time using:
Code:ps -u -q `pidof kwin_x11`This results in the following log (polled every hour, except when suspended):
Code:USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 7.2 1.9 2064412 158576 ? Sl 17:52 0:11 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 1.8 2.0 2270272 165256 ? Sl 17:52 1:20 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 1.3 2.0 2262796 165872 ? Sl 17:52 1:43 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 1.0 2.0 2263372 166500 ? Sl 17:52 1:58 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.1 0.8 2341128 64352 ? Sl Nov22 3:33 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.3 1.0 2624312 82196 ? Sl Nov22 8:00 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 1.2 2638020 101372 ? Sl Nov22 13:08 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 1.4 2679912 119600 ? Sl Nov22 17:05 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 1.4 2680168 119892 ? Sl Nov22 17:20 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 1.4 2650388 119744 ? Sl Nov22 18:00 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.7 1.5 2659080 128292 ? Sl Nov22 21:22 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.5 2644772 121172 ? Sl Nov22 24:06 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 2.0 2873076 162784 ? Sl Nov22 28:38 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.9 2976480 157588 ? Sl Nov22 31:07 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3074340 147392 ? Sl Nov22 32:43 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3075364 148432 ? Sl Nov22 33:19 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3075608 148784 ? Sl Nov22 33:37 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3076140 149408 ? Sl Nov22 33:55 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3087844 149664 ? Sl Nov22 37:29 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3090080 152056 ? Sl Nov22 38:13 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.6 3100044 134456 ? Sl Nov22 39:57 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.9 3104272 158424 ? Rl Nov22 41:32 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 2.1 3163380 169104 ? Sl Nov22 44:09 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 2.2 3175708 180168 ? Sl Nov22 52:04 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 2.3 3193956 185756 ? Sl Nov22 58:02 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 2.5 3215012 205608 ? Sl Nov22 69:08 /usr/bin/kwin_x11The way I noticed this, was that after some time hibernation begins to fail. It happened this morning as well, and the last message in syslog reads:
Code:elogind-daemon[1929]: Failed to suspend system. System resumed again: No space left on deviceI'm using suspend-then-hibernate, and since Slackware64-15.0 I'm greeted with a new boot instead of a resume once every 5 or 10 days (and the risk that I've lost some work).
So, rather than just suspending and not hibernating, I thought it might be interesting to track memory use. I noticed that, after a couple of days, kwin_x11 comes up as the process that uses the most memory. So, does anyone else notice this? And does anyone perhaps know of a patch? A very cursory google search didn't yield anything conclusive...
Code:ps -u -q `pidof kwin_x11`This results in the following log (polled every hour, except when suspended):
Code:USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 7.2 1.9 2064412 158576 ? Sl 17:52 0:11 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 1.8 2.0 2270272 165256 ? Sl 17:52 1:20 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 1.3 2.0 2262796 165872 ? Sl 17:52 1:43 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 1.0 2.0 2263372 166500 ? Sl 17:52 1:58 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.1 0.8 2341128 64352 ? Sl Nov22 3:33 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.3 1.0 2624312 82196 ? Sl Nov22 8:00 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 1.2 2638020 101372 ? Sl Nov22 13:08 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 1.4 2679912 119600 ? Sl Nov22 17:05 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 1.4 2680168 119892 ? Sl Nov22 17:20 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 1.4 2650388 119744 ? Sl Nov22 18:00 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.7 1.5 2659080 128292 ? Sl Nov22 21:22 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.5 2644772 121172 ? Sl Nov22 24:06 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 2.0 2873076 162784 ? Sl Nov22 28:38 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.9 2976480 157588 ? Sl Nov22 31:07 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3074340 147392 ? Sl Nov22 32:43 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3075364 148432 ? Sl Nov22 33:19 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3075608 148784 ? Sl Nov22 33:37 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3076140 149408 ? Sl Nov22 33:55 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3087844 149664 ? Sl Nov22 37:29 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.8 3090080 152056 ? Sl Nov22 38:13 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.6 3100044 134456 ? Sl Nov22 39:57 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 1.9 3104272 158424 ? Rl Nov22 41:32 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.4 2.1 3163380 169104 ? Sl Nov22 44:09 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 2.2 3175708 180168 ? Sl Nov22 52:04 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.5 2.3 3193956 185756 ? Sl Nov22 58:02 /usr/bin/kwin_x11
USER PID %CPU %MEM VSZ RSS TTY STAT START TIME COMMAND
pj 2352 0.6 2.5 3215012 205608 ? Sl Nov22 69:08 /usr/bin/kwin_x11The way I noticed this, was that after some time hibernation begins to fail. It happened this morning as well, and the last message in syslog reads:
Code:elogind-daemon[1929]: Failed to suspend system. System resumed again: No space left on deviceI'm using suspend-then-hibernate, and since Slackware64-15.0 I'm greeted with a new boot instead of a resume once every 5 or 10 days (and the risk that I've lost some work).
So, rather than just suspending and not hibernating, I thought it might be interesting to track memory use. I noticed that, after a couple of days, kwin_x11 comes up as the process that uses the most memory. So, does anyone else notice this? And does anyone perhaps know of a patch? A very cursory google search didn't yield anything conclusive...