LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Slackware (https://www.linuxquestions.org/questions/slackware-14/)
-   -   google-chrome segfaults on Slackware64 -current (https://www.linuxquestions.org/questions/slackware-14/google-chrome-segfaults-on-slackware64-current-937022/)

TL_CLD 03-29-2012 02:53 AM

google-chrome segfaults on Slackware64 -current
 
I did a slackpkg update/upgrade-all on my boxes this morning (Thinkpad T420 and Sun Ultra 27) and now google-chrome 17.0.963.83 dies with a segmentation fault on both boxes.

I've created the package using the google-chrome slackbuild from 13.37.

Anybody else experiencing this? Is there perhaps a fix floating around out there? :)

willysr 03-29-2012 08:28 AM

try rebuild the Google Chrome and you probably want to upgrade to Google Chrome 18 :)

TL_CLD 03-29-2012 09:38 AM

Same deal. It still dies:

Code:

[2:2:90167802330:ERROR:nacl_fork_delegate_linux.cc(104)] Bad NaCl helper startup ack (0 bytes)
[16377:16390:90167819279:ERROR:object_proxy.cc(239)] Failed to call method: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.NetworkManager was not provided by any .service files
--2012-03-29 16:36:13--  https://clients2.google.com/cr/report
Resolving clients2.google.com (clients2.google.com)... 173.194.69.139, 173.194.69.138, 173.194.69.113, ...
Connecting to clients2.google.com (clients2.google.com)|173.194.69.139|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [text/html]
Saving to: `/dev/fd/3'

    [<=>                                                                                          ] 0          --.-K/s             
Crash dump id: 98285fec9d4c0db8
    [ <=>                                                                                        ] 16          --.-K/s  in 0s     

2012-03-29 16:36:15 (323 KB/s) - `/dev/fd/3' saved [16]

Segmentation fault

Pretty odd stuff.

willysr 03-29-2012 07:02 PM

Yeah indeed because my Google Chrome installation works fine here on -Current (32 bit systems)
have you make sure that you have updated your -Current to the latest batch of updates?

TL_CLD 03-30-2012 01:06 AM

Yup, I've got the latest batch of updates, with the glibc update included.

Maybe it's a 64 bit issue?

Oh well, I'll survive. I expect Pat and crew will have it sorted out sooner or later, and in the meantime I'll just use Opera instead.

willysr 03-30-2012 01:13 AM

did you use multilib?

TL_CLD 03-30-2012 07:34 AM

No multilib. Pure Slackware64 -current.

budman85 05-12-2012 07:34 AM

Quote:

Originally Posted by TL_CLD (Post 4640590)
No multilib. Pure Slackware64 -current.

Same problem after rebooting.

google-chrome
[2:2:504006861:ERROR:nacl_fork_delegate_linux.cc(104)] Bad NaCl helper startup ack (0 bytes)
[3459:3472:504019769:ERROR:object_proxy.cc(239)] Failed to call method: org.freedesktop.DBus.Error.ServiceUnknown: The name org.freedesktop.NetworkManager was not provided by any .service files
Segmentation fault


However, I can run google-chrome as root

root> google-chrome --user-data-dir /home/budman

Works fine.

Digging deeper to see why. :) will let you know if I find anything


EDIT: Some results

root> mount -o exec /dev/shm

google-chrome runs fine - I close it and restart, back to same issue

related bug
http://code.google.com/p/nativeclien...detail?id=2533

willysr 05-12-2012 09:09 AM

do you have shared memory partition on your /etc/fstab?
it's something like this on my system
Code:

tmpfs            /dev/shm        tmpfs      defaults        0  0

budman85 05-12-2012 03:36 PM

Quote:

Originally Posted by willysr (Post 4676637)
do you have shared memory partition on your /etc/fstab?
it's something like this on my system
Code:

tmpfs            /dev/shm        tmpfs      defaults        0  0

Yes, thats ok.

I renamed the google-chrome dir under .config and had a good version that wouldn't segfault.
Started adding back directories to see if anything triggers.
I am able to run google-chrome now, however, I see the message, but no longer segfaults.

Edit: found the problem file that was causing segfault
".config/google-chrome/Service State"
Once I renamed it, no more segfault.

I did try reinstalling glib2, and downloaded lastest version in installed that.
Same issue, but removing the file seems to have worked for now.

TL_CLD 05-14-2012 02:44 AM

Renaming the Service State file on my system changes nothing; Google Chrome still segfaults on a regular basis.

It's really dreadfully annoying, as all tabs are lost when it segfaults.

budman85 05-23-2012 06:17 PM

Quote:

Originally Posted by TL_CLD (Post 4677777)
Renaming the Service State file on my system changes nothing; Google Chrome still segfaults on a regular basis.

It's really dreadfully annoying, as all tabs are lost when it segfaults.


Can you check your sound device order.


I found mine added a new module for hdmi, I forced the option sequence order and now google-chrome runs fine.
No more segfaults.

/etc/modprobe.d/sound.conf
# force to higher card number
options snd_hda_intel index=1
options snd_hda_codec_hdmi index=2


> aplay -l| grep card
card 0: Live [SB Live! 5.1], device 0: emu10k1 [ADC Capture/Standard PCM Playback]
card 0: Live [SB Live! 5.1], device 2: emu10k1 efx [Multichannel Capture/PT Playback]
card 0: Live [SB Live! 5.1], device 3: emu10k1 [Multichannel Playback]

> cat /proc/asound/modules
0 snd_emu10k1

I noticed the card order changed when clementine and amarok crashed.
When I ran alsamixer, it was showing HDMI as the default, even when I ran alasconf.

See if this helps.

TL_CLD 05-24-2012 01:28 AM

I only have snd_hda_intel, which also is the only option in alsamixer.

So this seems to not be the solution for me. Sadly. I would've loved if it was something as simple as you described.

I guess I'm just going to have to wait for Pat to update the parts of the system that is causing this, be it the kernel or something else.

smoooth103 05-24-2012 02:46 AM

I'm running chrome with current-64 with no issues. Running it on a couple machines -- a fresh install and an old install, all appears ok here.

TL_CLD 05-24-2012 06:29 AM

When I start google-chrome with strace, it dies with this:

Code:

write(43, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(33, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(33, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
+++ killed by SIGSEGV +++
Segmentation fault

Without strace I simply get:

Code:

[2:2:952748919:ERROR:nacl_fork_delegate_linux.cc(106)] Bad NaCl helper startup ack (0 bytes)
Segmentation fault

The WM used does not appear to matter. I've tried KDE, XFCE and xmonad.

This happens consistently on two different systems: A Lenovo T420 laptop and a Sun Ultra27 workstation.

It is driving me nuts.

I've tried deleting .config/google-chrome, to no avail. I'm currently running the latest google-chrome 19.something.

budman85 05-24-2012 09:42 PM

Quote:

Originally Posted by TL_CLD (Post 4686292)
When I start google-chrome with strace, it dies with this:

Code:

write(43, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(33, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
write(33, "\0", 1)                      = 1
write(43, "\0", 1)                      = 1
+++ killed by SIGSEGV +++
Segmentation fault

Without strace I simply get:

Code:

[2:2:952748919:ERROR:nacl_fork_delegate_linux.cc(106)] Bad NaCl helper startup ack (0 bytes)
Segmentation fault

The WM used does not appear to matter. I've tried KDE, XFCE and xmonad.

This happens consistently on two different systems: A Lenovo T420 laptop and a Sun Ultra27 workstation.

It is driving me nuts.

I've tried deleting .config/google-chrome, to no avail. I'm currently running the latest google-chrome 19.something.




I saw a thread under ArchLinux. They found it was because the nscd daemon is no longer running. The quick work around is to run the daemon again, remember to create the dirs /var/run/nscd and /var/db/nscd before starting.

The real problem has been fixed in glibc 2.15.5 - current is using 2.15.2. I sent Patrick a note about this.

here is a link to the article https://bbs.archlinux.org/viewtopic.php?id=133021

ruario 05-25-2012 12:21 AM

@TL_CLD: You have installed the other dependencies from /extra right (google-chrome-pam-solibs, GConf and ORBit2)?

TL_CLD 05-25-2012 02:21 AM

Quote:

Originally Posted by ruario (Post 4687001)
@TL_CLD: You have installed the other dependencies from /extra right (google-chrome-pam-solibs, GConf and ORBit2)?

Yes. I've got all three. Note that Google Chrome is working, it just takes 10-15 crashes during startup before it finally stays alive. Once it is running, it is rock solid.

TL_CLD 05-25-2012 02:35 AM

Quote:

Originally Posted by budman85 (Post 4686931)
I saw a thread under ArchLinux. They found it was because the nscd daemon is no longer running. The quick work around is to run the daemon again, remember to create the dirs /var/run/nscd and /var/db/nscd before starting.

The real problem has been fixed in glibc 2.15.5 - current is using 2.15.2. I sent Patrick a note about this.

here is a link to the article https://bbs.archlinux.org/viewtopic.php?id=133021

So if/when Pat updates glibc to 2.15.5 the problem will hopefully go away. That is good to know.

I've no scripts for nscd in /etc/rc.d/ so I'll just leave that as is and hope for the glibc fix. I don't want to muck around too much with -current.

budman85 05-25-2012 07:26 AM

Quote:

Originally Posted by ruario (Post 4687001)
@TL_CLD: You have installed the other dependencies from /extra right (google-chrome-pam-solibs, GConf and ORBit2)?

I checked when the packages were installed, the glibc upgrade was in April.
I was using Chrome 16.0.912 without issue.
When I upgraded to Chrome 18.0.1025, the segfaults were continuous.
I saw some notes on Google Chrome dev forums about them making glib 2.18 the minimum version.
Maybe this has some residuals. glib2 2.28.6 was recently added.

When I check ldd /opt/google/chrome/chrome I only see one lib that is not from lib64 dir
linux-vdso.so.1 (0x00007fff287ff000)

I will try downgrading to version 16 to see if that segfaults on startup.

volkerdi 05-25-2012 10:25 AM

Hello, glibc-2.15 is the latest, and there is no such thing as glibc-2.15.5. The Arch bug report refers to their patched package, glibc-2.15-5.

This is the actual location to get the patch:

http://sourceware.org/bugzilla/show_bug.cgi?id=13594

budman85 05-26-2012 12:59 AM

Quote:

Originally Posted by volkerdi (Post 4687465)
Hello, glibc-2.15 is the latest, and there is no such thing as glibc-2.15.5. The Arch bug report refers to their patched package, glibc-2.15-5.

This is the actual location to get the patch:

http://sourceware.org/bugzilla/show_bug.cgi?id=13594

Thanks.

I rebooted the system today, same segfault.
This time strace showed it trying to read the cache dir.

~/.cache/google-chrome

I renamed the dir and it loaded with no segfault.

willysr 05-26-2012 02:14 AM

Pat has updated glibc with the patch mentioned above
can you confirmed that the patch work or does not work?

budman85 05-26-2012 06:07 PM

Quote:

Originally Posted by willysr (Post 4687961)
Pat has updated glibc with the patch mentioned above
can you confirmed that the patch work or does not work?

downloading now - will try in a few - thanks :)

Nice! No more segfault!

I rebooted, chrome loaded right up.

The reoccurring issue was:
Reboot
start - segfault
restart - runs fine, close it
restart - segfault
restart - runs fine, close it
restart - segfault
and so on..

After update, I tried the same steps:
Reboot
start - ran fine, close it
restart - ran fine, close it
and so on...

Looks like the patch worked. :)

Tell Pat thanks!

Also, tell him thanks for upgrading xscreensaver to 5.15, that fixes a hard crash bug for newer Nvidia cards and prop. divers.


Rich

TL_CLD 05-27-2012 02:24 AM

The latest glibc -current update fixes it completely. It works flawlessly on both my computers now. Thank you!

willysr 05-27-2012 04:22 AM

Quote:

Originally Posted by budman85 (Post 4688435)

Tell Pat thanks!

Also, tell him thanks for upgrading xscreensaver to 5.15, that fixes a hard crash bug for newer Nvidia cards and prop. divers.
Rich

You can thank him directly via his email :)

TL_CLD 05-27-2012 10:13 AM

He's also on Google+

https://plus.google.com/u/0/114480574919795916970/posts


All times are GMT -5. The time now is 08:53 AM.