Description of problem: My VM is steadily 15% busy when idle at login screen, mainly due to sddm_greeter: PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 766 sddm 20 0 517860 100512 46072 S 9.0 5.0 38:48.08 sddm-greeter 712 root 20 0 193288 26208 16820 S 1.0 1.3 2:57.03 Xorg.bin strace loop: [pid 766] clock_gettime(CLOCK_MONOTONIC, {26826, 501372095}) = 0 [pid 766] clock_gettime(CLOCK_MONOTONIC, {26826, 501408388}) = 0 [pid 766] clock_gettime(CLOCK_MONOTONIC, {26826, 501444230}) = 0 [pid 766] clock_gettime(CLOCK_MONOTONIC, {26826, 501479954}) = 0 [pid 766] poll([{fd=5, events=POLLIN}, {fd=17, events=POLLIN}, {fd=19, events=POLLIN}, {fd=20, events=POLLIN}], 4, 5) = 0 (Timeout) [pid 766] clock_gettime(CLOCK_MONOTONIC, {26826, 506633469}) = 0 [pid 766] clock_gettime(CLOCK_MONOTONIC, {26826, 506658942}) = 0 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"\16\2\2\0\16\0 \0", 8}, {NULL, 0}, {"", 0}], 3) = 8 [pid 776] <... poll resumed> ) = 1 ([{fd=3, revents=POLLIN}]) [pid 776] futex(0x7fcfe2996008, FUTEX_WAIT_PRIVATE, 2, NULL <unfinished ...> [pid 766] futex(0x7fcfe2996008, FUTEX_WAKE_PRIVATE, 1 <unfinished ...> [pid 776] <... futex resumed> ) = 0 [pid 776] recvmsg(3, {msg_name(0)=NULL, msg_iov(1)=[{"\1\30<m\0\0\0\0{\1\0\0\0\0\0\0\0\4\0\3\0\0\0\0\0\0\0\0\0\0\0\0", 4096}], msg_controllen=0, msg_flags=0}, 0) = 32 [pid 776] futex(0x7fcfe2996008, FUTEX_WAKE_PRIVATE, 1) = 0 [pid 776] poll([{fd=3, events=POLLIN}], 1, 4294967295 <unfinished ...> [pid 766] <... futex resumed> ) = 1 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0\0\0\0\30 \0", 24}, {"\30\3\4\377\30\3\4\377\30\3\4\377\31\3\4\377\30\3\4\377\30\3\4\377\30\3\4\377\31\3\4\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0?\0\0\30 \0", 24}, {"3\23\n\3774\24\n\3774\24\n\3775\25\v\3775\25\v\3774\24\n\3774\24\n\3774\24\n\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0~\0\0\30 \0", 24}, {"6\26\v\3776\26\v\3775\25\v\3776\26\v\3775\25\v\3775\25\n\3775\25\n\3776\26\v\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0\275\0\0\30 \0", 24}, {"6\26\v\3775\26\v\3775\26\v\3776\26\f\3776\26\v\3775\25\v\3775\25\v\3776\26\f\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0\374\0\0\30 \0", 24}, {"5\25\v\3775\25\v\3775\25\v\3775\25\v\3775\25\n\3775\25\v\3775\25\n\3776\26\v\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0;\1\0\30 \0", 24}, {".\f\10\377.\r\t\377-\f\10\377.\f\10\377.\f\10\377.\f\10\377-\f\10\377.\f\10\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0z\1\0\30 \0", 24}, {"1\20\t\3771\21\n\3771\20\n\3771\21\t\3771\20\t\3772\21\n\3771\21\t\3772\21\t\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0\271\1\0\30 \0", 24}, {"6\27\f\3776\27\f\3777\27\f\3776\27\v\3777\27\f\3776\27\v\3776\27\v\3776\26\v\377"..., 258048}, {"", 0}], 3) = 258072 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0\370\1\0\30 \0", 24}, {"7\31\f\3777\30\f\3777\30\f\3777\30\f\3778\30\f\3777\30\f\3776\30\f\3777\27\v\377"..., 258048}, {"", 0}], 3) = 219264 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377"..., 38808}, {"", 0}], 2) = 38808 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0007\2\0\30 \0", 24}, {"9\32\r\3779\31\f\3778\32\f\3778\31\f\3779\32\f\3778\32\f\3778\31\f\3779\31\f\377"..., 258048}, {"", 0}], 3) = 219264 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"D'\20\377D'\20\377E'\20\377E'\20\377E(\21\377E(\20\377F(\21\377F(\21\377"..., 38808}, {"", 0}], 2) = 38808 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0v\2\0\30 \0", 24}, {"z`-\377z`-\377za-\377z`-\377z`,\377za-\377za-\377y`-\377"..., 258048}, {"", 0}], 3) = 219264 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"N1\23\377N0\23\377N0\23\377N0\23\377N1\23\377N1\23\377M0\23\377M1\23\377"..., 38808}, {"", 0}], 2) = 38808 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\6\374\16\0 \0\26\0 \0\0\4?\0\0\0\265\2\0\30 \0", 24}, {"\200f.\377\200g.\377\200g.\377\200f.\377\200f-\377\200g.\377\200g.\377\200f.\377"..., 258048}, {"", 0}], 3) = 219264 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"cG\31\377cF\31\377cF\31\377cF\31\377cF\32\377cF\31\377cF\32\377cF\31\377"..., 38808}, {"", 0}], 2) = 38808 [pid 766] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 766] writev(3, [{"H\2\0060\16\0 \0\26\0 \0\0\4\f\0\0\0\364\2\0\30 \0", 24}, {"tX\36\377tX\37\377tX\37\377tX\37\377tX\36\377tX\36\377sX\37\377sW\36\377"..., 49152}, {"", 0}], 3) = 49176 [pid 766] clock_gettime(CLOCK_MONOTONIC, {26826, 558686308}) = 0 [pid 766] poll([{fd=5, events=POLLIN}, {fd=17, events=POLLIN}, {fd=19, events=POLLIN}, {fd=20, events=POLLIN}], 4, 442^CProcess 766 detached <detached ...> Version-Release number of selected component (if applicable): sddm-0.10.0-2.fc21.x86_64
Hi Orion, can you please take a look what are those files under FD 3 and 5?
Rebooted and now the loop appears to use fds 3 and 6: sddm-gree 725 sddm 3u unix 0xffff880077724380 0t0 17507 socket sddm-gree 725 sddm 6u a_inode 0,9 0 6459 [eventfd] # lsof | grep 0xffff880077724380 sddm-gree 725 sddm 3u unix 0xffff880077724380 0t0 17507 socket QXcbEvent 725 749 sddm 3u unix 0xffff880077724380 0t0 17507 socket QQmlThrea 725 759 sddm 3u unix 0xffff880077724380 0t0 17507 socket 5/6 seems to be an inotify handler? [pid 725] poll([{fd=6, events=POLLIN}, {fd=17, events=POLLIN}, {fd=19, events=POLLIN}, {fd=20, events=POLLIN}], 4, 437) = 0 (Timeout) ... [pid 725] poll([{fd=6, events=POLLIN}, {fd=17, events=POLLIN}, {fd=19, events=POLLIN}, {fd=20, events=POLLIN}], 4, 5) = 0 (Timeout) can't we block instead of having short timeouts? (gdb) thr appl all bt Thread 3 (Thread 0x7fd2badf8700 (LWP 749)): #0 0x00007fd2c68b11dd in poll () from /lib64/libc.so.6 #1 0x00007fd2c92f3182 in _xcb_conn_wait () from /lib64/libxcb.so.1 #2 0x00007fd2c92f4cff in xcb_wait_for_event () from /lib64/libxcb.so.1 #3 0x00007fd2bd432089 in QXcbEventReader::run() () from /usr/lib64/qt5/plugins/platforms/libqxcb.so #4 0x00007fd2c744393e in QThreadPrivate::start(void*) () from /lib64/libQt5Core.so.5 #5 0x00007fd2c63a352a in start_thread () from /lib64/libpthread.so.0 #6 0x00007fd2c68bc77d in clone () from /lib64/libc.so.6 Thread 2 (Thread 0x7fd2b9008700 (LWP 759)): #0 0x00007fd2c68b11dd in poll () from /lib64/libc.so.6 #1 0x00007fd2c2680e24 in g_main_context_iterate.isra () from /lib64/libglib-2.0.so.0 #2 0x00007fd2c2680f3c in g_main_context_iteration () from /lib64/libglib-2.0.so.0 #3 0x00007fd2c7687897 in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () from /lib64/libQt5Core.so.5 #4 0x00007fd2c762da92 in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () from /lib64/libQt5Core.so.5 #5 0x00007fd2c743eae4 in QThread::exec() () from /lib64/libQt5Core.so.5 #6 0x00007fd2c744393e in QThreadPrivate::start(void*) () from /lib64/libQt5Core.so.5 #7 0x00007fd2c63a352a in start_thread () from /lib64/libpthread.so.0 #8 0x00007fd2c68bc77d in clone () from /lib64/libc.so.6 Thread 1 (Thread 0x7fd2c96e7880 (LWP 725)): #0 0x00007fd2c68b11dd in poll () from /lib64/libc.so.6 #1 0x00007fd2c2680e24 in g_main_context_iterate.isra () from /lib64/libglib-2.0.so.0 #2 0x00007fd2c2680f3c in g_main_context_iteration () from /lib64/libglib-2.0.so.0 #3 0x00007fd2c7687897 in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () from /lib64/libQt5Core.so.5 #4 0x00007fd2c762da92 in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () from /lib64/libQt5Core.so.5 #5 0x00007fd2c763550c in QCoreApplication::exec() () from /lib64/libQt5Core.so.5 #6 0x00007fd2c9741f47 in main (argc=5, argv=0x7fffd29191f8) at /usr/src/debug/sddm-0.10.0/src/greeter/GreeterApp.cpp:185
Seems better on my i3 laptop 0.3-0.7%. strace shows: [pid 756] <... poll resumed> ) = 1 ([{fd=3, revents=POLLIN}]) [pid 756] recvmsg(3, {msg_name(0)=NULL, msg_iov(1)=[{"#\223\363\16\0\0\0\0\2\0\0\0\26\0 \0\r\0 \0\320\v\0\0\27\0 \0\30\0 \0"..., 4096}], msg_controllen=0, msg_flags=0}, 0) = 72 [pid 756] futex(0x7f14292ebf3c, FUTEX_WAKE_OP_PRIVATE, 1, 1, 0x7f14292ebf38, {FUTEX_OP_SET, 0, FUTEX_OP_CMP_GT, 1}) = 1 [pid 755] <... futex resumed> ) = 0 [pid 756] poll([{fd=3, events=POLLIN}], 1, 4294967295 <unfinished ...> [pid 755] futex(0x7f142899db58, FUTEX_WAKE_PRIVATE, 1) = 0 [pid 755] ioctl(16, 0x400c645f, 0x7fff5cb0d7a0) = 0 [pid 755] ioctl(16, 0xc0086457, 0x7fff5cb0d570) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d560) = 0 [pid 755] ioctl(16, 0x4020645d, 0x7fff5cb0d5f0) = 0 [pid 755] ioctl(16, 0xc0086457, 0x7fff5cb0d550) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d540) = 0 [pid 755] ioctl(16, 0x4020645d, 0x7fff5cb0d5d0) = 0 [pid 755] ioctl(16, 0xc0086457, 0x7fff5cb0d850) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d840) = 0 [pid 755] ioctl(16, 0x4020645d, 0x7fff5cb0d8d0) = 0 [pid 755] ioctl(16, 0x40046460, 0x7fff5cb0da30) = 0 [pid 755] ioctl(16, 0x40406469, 0x7fff5cb0d9c0) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d980) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d980) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d980) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d980) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d9b0) = 0 [pid 755] ioctl(16, 0xc0086457, 0x7fff5cb0d9a0) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d990) = 0 [pid 755] ioctl(16, 0x400c645f, 0x7fff5cb0d9a0) = 0 [pid 755] poll([{fd=3, events=POLLIN|POLLOUT}], 1, 4294967295) = 1 ([{fd=3, revents=POLLOUT}]) [pid 755] writev(3, [{"\223\1\22\0\r\0 \0\27\0 \0\321\v\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 72}], 1) = 72 [pid 755] poll([{fd=5, events=POLLIN}, {fd=18, events=POLLIN}, {fd=20, events=POLLIN}, {fd=21, events=POLLIN}], 4, 0) = 0 (Timeout) [pid 755] poll([{fd=5, events=POLLIN}, {fd=18, events=POLLIN}, {fd=20, events=POLLIN}, {fd=21, events=POLLIN}], 4, 5) = 0 (Timeout) [pid 755] ioctl(16, 0xc0086457, 0x7fff5cb0d6d0) = 0 [pid 755] ioctl(16, VIDIOC_INT_RESET, 0x7fff5cb0d6c0) = 0 [pid 755] ioctl(16, 0x4020645d, 0x7fff5cb0d750) = 0 [pid 755] futex(0x7f14292ebf3c, FUTEX_WAIT_PRIVATE, 1233, NULL so perhaps video device related. Still getting tickled every second.
I don't think this is a bug. Since the driver is virtualized, it seems the CPU is being utilized at the graphical driver, somewhere in XCB. If there's an actual problem, it's in XCB, not SDDM, in my opinion.
SDDM is now using QML 2, right? That implies it's using OpenGL. I wonder how much of that CPU time is spent in llvmpipe.
Yeah, after poking at this some more, I'm thinking that this is the effect of using software rendering to draw the blinking cursor. And I'm not sure if there is any way to disable to blinking cursor. I see an option with QApplication, but not QGuiApplication.
FWIW - gdm on EL7 doesn't have a blinking cursor, although kdm does.
Qt Quick gets the cursor blink time from qApp->styleHints()->cursorFlashTime(). Unfortunately, this is officially read-only. There is, however, an undocumented internal "void setCursorFlashTime(int cursorFlashTime);" in the QStyleHints class that we may be able to use. (The whole class is exported, so the symbol should be visible.) The official way to set it is through the platform theme and/or platform integration mechanisms; I'd have to look through the KDE platform plugin to see whether it gets the value from KConfig or something. In any case, the fact that even something as simple as blinking the cursor sucks 15% CPU is really worrisome though, it looks like relying on OpenGL and thus llvmpipe was a very wrong thing for upstream to do. :-( But there isn't much we can do about that.
This situation can be observed in Fedora 22 as well. I tried with VirtualBox 4.3.28 as virtualization software. # rpm -qf /usr/bin/sddm-greeter sddm-0.10.0-4.fc22.x86_64 Either press Ctrl+Alt+F2 or ssh into the F22 machine. Then execute top. The CPU is 91% idle at login screen. %Cpu(s): 7.0 us, 1.7 sy, 0.0 ni, 91.0 id, 0.0 wa, 0.0 hi, 0.3 si, 0.0 st After a user logs in, sddm-greeter vanishes and CPU usage drops to nearly 0%. If you switch to GDM: $ sudo systemctl enable --force gdm.service ; reboot or LightDM: $ sudo systemctl enable --force lightdm.service ; reboot or KDM: $ sudo systemctl enable --force kdm.service ; reboot you can observe nearly 100% idle at login screen with top. %Cpu(s): 0.0 us, 0.3 sy, 0.0 ni, 99.7 id, 0.0 wa, 0.0 hi, 0.0 si, 0.0 st KDM and LightDM have a blinking cursor in text input fields. GDM has a always-visible cursor in text input fields. Please, note that CPU is used by sddm-greeter only, if cursor is blinking in text input box. Manually clicking outside a text input box is enough to let CPU usage of sddm-greeter drop to nearly 0%. Furthermore, screen locker of KDE/Plasma seems to be affected by the same problem by looking at the output of top. However the blinking cursor is always present. So unlocking the screen locker was the only way to test for me. To test execute (no need for KDE/Plasma session): $ /usr/libexec/kscreenlocker_greet --immediateLock --ksldfd 28 Shall I file a dedicated bug report for it? For what it is worth, in the title of this bug report 'sddm_greeter' should be changed to 'sddm-greeter'.
This message is a reminder that Fedora 21 is nearing its end of life. Approximately 4 (four) weeks from now Fedora will stop maintaining and issuing updates for Fedora 21. It is Fedora's policy to close all bug reports from releases that are no longer maintained. At that time this bug will be closed as EOL if it remains open with a Fedora 'version' of '21'. Package Maintainer: If you wish for this bug to remain open because you plan to fix it in a currently maintained version, simply change the 'version' to a later Fedora version. Thank you for reporting this issue and we are sorry that we were not able to fix it before Fedora 21 is end of life. If you would still like to see this bug fixed and are able to reproduce it against a later version of Fedora, you are encouraged change the 'version' to a later Fedora version prior this bug is closed as described in the policy above. Although we aim to fix as many bugs as possible during every release's lifetime, sometimes those efforts are overtaken by events. Often a more recent Fedora release includes newer upstream software that fixes bugs or makes them obsolete.
problem present in Fedora 23 x86_64 as well
problem might be resolved in Qt 5.8 according to upstream bug report: https://github.com/sddm/sddm/issues/323#issuecomment-228710103 F24 uses qt5-qtbase-5.6.0-21.fc24 at the moment
Fedora 22 changed to end-of-life (EOL) status on 2016-07-19. Fedora 22 is no longer maintained, which means that it will not receive any further security or bug fix updates. As a result we are closing this bug. If you can reproduce this bug against a currently maintained version of Fedora please feel free to reopen this bug against that version. If you are unable to reopen this bug, please file a new report against the current release. If you experience problems, please add a comment to this bug. Thank you for reporting this bug and we are sorry it could not be fixed.
I'm seeing this on F25 (KDE spin).
This solved it for me on KDE neon inside Hyper-V: Add to: /etc/security/pam_env.conf Avoid high cpu usage with sddm-greeter https://github.com/sddm/sddm/issues/323 ``` QT_QUICK_BACKEND DEFAULT=software ```