Status: Untriaged
Owner: ----
Labels: Type-Bug Pri-2 OS-Linux Area-Misc Size-Medium

New issue 18278 by [email protected]: Chrome on Linux goes invisible and  
chews CPU and RAM after a day
http://code.google.com/p/chromium/issues/detail?id=18278

This has never happened to me before, but chrome
is sitting and spinning with no visible window.

$ cat /etc/issue
Ubuntu 9.04 \n \l

$ uname -a
Linux slave3 2.6.28-14-generic #47-Ubuntu SMP Sat Jul 25 01:19:55 UTC 2009
x86_64 GNU/Linux

$ dpkg-query -W google-chrome-unstable
google-chrome-unstable  3.0.195.1-r21222

$ ps | grep chrome

dank     23837 93.3  1.3 227144 55624 ?        Sl   Aug01 405:54
/opt/google/chrome/chrome
dank     23847  0.0  0.1  96868  7128 ?        S    Aug01   0:00
/opt/google/chrome/chrome
dank     23848  0.0  0.1  55184  7064 ?        S    Aug01   0:00
/opt/google/chrome/chrome --type=zygote

$ ps | grep chrome
dank     23837 93.4  1.3 227144 55624 ?        Sl   Aug01 408:40
/opt/google/chrome/chrome
dank     23847  0.0  0.1  96868  7128 ?        S    Aug01   0:00
/opt/google/chrome/chrome
dank     23848  0.0  0.1  55184  7064 ?        S    Aug01   0:00
/opt/google/chrome/chrome --type=zygote

i.e. process 23837 is chewing up lots of cpu.

I first noticed this when I got back to my computer
after being away for six hours; there was no browser
window, so I tried starting one, and it just printed
Created new window in existing browser session.

$ ps -eLF | grep chrome
dank     23837     1 23837  0    7 56786 55624   0 Aug01 ?        00:00:23
/opt/google/chrome/chrome
dank     23837     1 23841  0    7 56786 55624   0 Aug01 ?        00:00:00
/opt/google/chrome/chrome
dank     23837     1 23843  0    7 56786 55624   2 Aug01 ?        00:00:04
/opt/google/chrome/chrome
dank     23837     1 23845  0    7 56786 55624   0 Aug01 ?        00:00:03
/opt/google/chrome/chrome
dank     23837     1 23846  0    7 56786 55624   0 Aug01 ?        00:00:00
/opt/google/chrome/chrome
dank     23837     1 24846  0    7 56786 55624   3 Aug01 ?        00:00:31
/opt/google/chrome/chrome
dank     23837     1 24868 99    7 56786 55624   2 Aug01 ?        06:54:26
/opt/google/chrome/chrome
dank     23847 23837 23847  0    1 24217  7128   1 Aug01 ?        00:00:00
/opt/google/chrome/chrome
dank     23848 23837 23848  0    1 13796  7064   1 Aug01 ?        00:00:00
/opt/google/chrome/chrome --type=zygote

$ strace -p 24868
poll([{fd=105, events=POLLIN|POLLERR|POLLNVAL}], 1, 4294967295) = 1
([{fd=105, revents=POLLIN}])
poll([{fd=105, events=POLLIN|POLLERR|POLLNVAL}], 1, 4294967295) = 1
([{fd=105, revents=POLLIN}])
poll([{fd=105, events=POLLIN|POLLERR|POLLNVAL}], 1, 4294967295) = 1
([{fd=105, revents=POLLIN}])
poll([{fd=105, events=POLLIN|POLLERR|POLLNVAL}], 1, 4294967295) = 1
([{fd=105, revents=POLLIN}])
...
very quickly; seems to be stuck in a busy loop, getting notified
about fd 105, but not clearing the readiness notification?

$ ls -l /proc/23837/fd/105
lr-x------ 1 dank dank 64 2009-08-02 06:53 /proc/23837/fd/105 ->  
pipe:[1397024]

$ gdb /opt/google/chrome/chrome
(gdb) thread 2
[Switching to thread 2 (Thread 0xf3741b90 (LWP 24868))]#0  0xf7f95430 in
__kernel_vsyscall ()
(gdb) bt
#0  0xf7f95430 in __kernel_vsyscall ()
#1  0xf7301147 in *__GI___poll (fds=0xf3740970, nfds=1, timeout=-1) at
../sysdeps/unix/sysv/linux/poll.c:87
#2  0xf74d17bd in ?? () from /usr/lib32/libasound.so.2
#3  0xf74d1934 in snd_pcm_wait () from /usr/lib32/libasound.so.2
#4  0xf74d1b45 in ?? () from /usr/lib32/libasound.so.2
#5  0xf7516ea5 in ?? () from /usr/lib32/libasound.so.2
#6  0xf74cc3a4 in snd_pcm_writei () from /usr/lib32/libasound.so.2
#7  0x090563c0 in ?? ()
#8  0x084902b9 in ?? ()
#9  0x08490396 in ?? ()
#10 0x084905ba in ?? ()
#11 0x08491558 in ?? ()
#12 0x08490b12 in ?? ()
#13 0x08490c06 in ?? ()
#14 0x084b4ee0 in ?? ()
#15 0x0849c631 in ?? ()
#16 0xf76514ff in start_thread (arg=0xf3741b90) at pthread_create.c:297
#17 0xf730bb9e in clone () at ../sysdeps/unix/sysv/linux/i386/clone.S:130
(gdb) quit

So it kind of looks like the thread that's using all the cpu
is outputting (silent) audio nonstop.  The last thing I remember doing
is visiting youtube.com/html5, so perhaps that code is not yet
happy.

Here's what audio card I have, if it matters:

$ lspci | grep -i audio
00:09.0 Audio device: nVidia Corporation MCP73 High Definition Audio (rev  
a1)

And here's a little info about alsa's state (dunno if useful):

d...@slave3:/proc/asound/card0/pcm0p/sub0$ cat status
state: RUNNING
trigger_time: 307628.680062937
tstamp      : 333163.919990922
delay       : 3467
avail       : 117
avail_max   : 117
-----
hw_ptr      : 1126023942
appl_ptr    : 1126027409
d...@slave3:/proc/asound/card0/pcm0p/sub0$ cat status
state: RUNNING
trigger_time: 307628.680062937
tstamp      : 333174.429047120
delay       : 3282
avail       : 302
avail_max   : 302
-----
hw_ptr      : 1126487359
appl_ptr    : 1126490641

I'll leave this running for a while in case someone wants more info.

--
You received this message because you are listed in the owner
or CC fields of this issue, or because you starred this issue.
You may adjust your issue notification preferences at:
http://code.google.com/hosting/settings

--~--~---------~--~----~------------~-------~--~----~
Automated mail from issue updates at http://crbug.com/
Subscription options: http://groups.google.com/group/chromium-bugs
-~----------~----~----~----~------~----~------~--~---

Reply via email to