1192477 - Firefox nightly on Linux uses 100% of CPU while idle

Status: RESOLVED WORKSFORME

(Core :: General, defect)

Description

[Stephen Moehle]

After running only a few minutes, Firefox nightly on Fedora 22 will start using 100% of the CPU while idle. The only window open is about:blank and Firefox was started in safe mode. It seems to start using 100% of the CPU only after one or more tabs have been closed. Other than turning my laptop into a space heater, Firefox itself seems to be running fine. There is no lag or sluggishness suggesting that it is bogging down processing something.

The stack trace I get from GDB is always the same when I break:

#0  0x00007fad633362fd in poll () at /lib64/libc.so.6
#1  0x00007fad523210d6 in PollWrapper(_GPollFD*, unsigned int, int) ()
    at /home/stephe/mozilla/nightly/libxul.so
#2  0x00007fad5e3ecdbc in g_main_context_iterate.isra ()
    at /lib64/libglib-2.0.so.0
#3  0x00007fad5e3ececc in g_main_context_iteration ()
    at /lib64/libglib-2.0.so.0
#4  0x00007fad5232106f in nsAppShell::ProcessNextNativeEvent(bool) ()
    at /home/stephe/mozilla/nightly/libxul.so
#5  0x00007fad537abffd in nsBaseAppShell::DoProcessNextNativeEvent(bool, unsigned int) () at /home/stephe/mozilla/nightly/libxul.so
#6  0x00007fad52320811 in nsBaseAppShell::OnProcessNextEvent(nsIThreadInternal*, bool, unsigned int) () at /home/stephe/mozilla/nightly/libxul.so
#7  0x00007fad52264c5d in nsThread::ProcessNextEvent(bool, bool*) ()
    at /home/stephe/mozilla/nightly/libxul.so
#8  0x00007fad5226da6e in NS_ProcessNextEvent(nsIThread*, bool) ()
    at /home/stephe/mozilla/nightly/libxul.so
#9  0x00007fad522853ff in mozilla::ipc::MessagePump::Run(base::MessagePump::Delegate*) () at /home/stephe/mozilla/nightly/libxul.so
#10 0x00007fad527d88a3 in MessageLoop::Run() ()
    at /home/stephe/mozilla/nightly/libxul.so
#11 0x00007fad537a6d58 in nsBaseAppShell::Run() ()
    at /home/stephe/mozilla/nightly/libxul.so
#12 0x00007fad53d518fc in nsAppStartup::Run() ()
    at /home/stephe/mozilla/nightly/libxul.so
#13 0x00007fad53d8e5dd in XREMain::XRE_mainRun() ()
    at /home/stephe/mozilla/nightly/libxul.so
#14 0x00007fad53d90c29 in XREMain::XRE_main(int, char**, nsXREAppData const*) () at /home/stephe/mozilla/nightly/libxul.so
#15 0x00007fad53d90ed8 in XRE_main ()
    at /home/stephe/mozilla/nightly/libxul.so
#16 0x0000000000408eb6 in do_main(int, char**, nsIFile*) ()
#17 0x000000000040548e in main ()

Strace shows the following repeated indefinitely with some differences on some iterations:

recvmsg(4, 0x7fffa8f753f0, 0)           = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=23, events=POLLIN}, {fd=24, events=POLLIN}, {fd=127, events=POLLIN}], 6, 0) = 0 (Timeout)
futex(0x7fad568617cc, FUTEX_CMP_REQUEUE_PRIVATE, 1, 2147483647, 0x7fad568fd950, 1040268) = 1
recvmsg(4, 0x7fffa8f753c0, 0)           = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=23, events=POLLIN}, {fd=24, events=POLLIN}, {fd=127, events=POLLIN}], 6, 0) = 0 (Timeout)
futex(0x7fad568617cc, FUTEX_CMP_REQUEUE_PRIVATE, 1, 2147483647, 0x7fad568fd950, 1040270) = 1
recvmsg(4, 0x7fffa8f753f0, 0)           = -1 EAGAIN (Resource temporarily unavailable)
futex(0x7fad568617cc, FUTEX_CMP_REQUEUE_PRIVATE, 1, 2147483647, 0x7fad568fd950, 1040272) = 1
poll([{fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=23, events=POLLIN}, {fd=24, events=POLLIN}, {fd=127, events=POLLIN}], 6, 0) = 0 (Timeout)
futex(0x7fad568617cc, FUTEX_CMP_REQUEUE_PRIVATE, 1, 2147483647, 0x7fad568fd950, 1040274) = 1
recvmsg(4, 0x7fffa8f753c0, 0)           = -1 EAGAIN (Resource temporarily unavailable)
poll([{fd=3, events=POLLIN}, {fd=4, events=POLLIN}, {fd=5, events=POLLIN}, {fd=23, events=POLLIN}, {fd=24, events=POLLIN}, {fd=127, events=POLLIN}], 6, 4294967295) = 1 ([{fd=24, revents=POLLIN}])
futex(0x7fad568617cc, FUTEX_CMP_REQUEUE_PRIVATE, 1, 2147483647, 0x7fad568fd950, 1040276) = 1
read(24, "\372", 1)                     = 1
write(11, "\0", 1)                      = 1

Comment 1

[Paul Oiegas [:poiegas], PM for (Ecosystem QA, Services QA, WebCompat QA, Ads QA, and Solo QA teams)]

Hi Stephen,

I have tried to reproduce your issue with the latest Nightly build (45.0a1) on Fedora 23 and could not reproduce it. Left the browser opened with just the default tab or with multiple tabs opened while observing the activity monitor. In none of the cases the processor was used at 100%.

Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Firefox/45.0 - 20151203053521

Can you please test this with the latest Nightly build or latest Firefox release 42.0 and let me know if it still reproduces on your end ?
Also, if try to do this, please use a new profile or maybe even safe mode (https://support.mozilla.org/en-US/kb/troubleshoot-and-diagnose-firefox-problems). Sometimes custom settings or addons could cause this kind of issues.


Thanks,
Paul.

Comment 2

[Stephen Moehle]

I use the latest Nightly builds, and I have not seen this problem for quite a while now. It must have been fixed somewhere along the line.

Comment 3

[Paul Oiegas [:poiegas], PM for (Ecosystem QA, Services QA, WebCompat QA, Ads QA, and Solo QA teams)]

Hi Stephen,

Great, thanks for the reply. Considering this I will mark this issue as resolved-worksforme. Feel free to reopen this bug if you encounter this again.


Regards,
Paul.

Comment 4

[Firefox Bug Husbandry Bot]

Moving from Core::Untriaged to Core::General https://bugzilla.mozilla.org/show_bug.cgi?id=1407598