PDA

View Full Version : [SOLVED] segfault exiting chrome



VernDog
29th June 2017, 12:14 AM
Clearing 'dmesg' and then opening chrome, I see nothing reported from 'dmesg'. Right after exiting chrome, 'dmesg' reports chrome segfault. Here's the results launching chrome from terminal, then dumping 'dmesg'.


$ google-chrome
Gkr-Message: secret service operation failed: Failed to execute program org.freedesktop.secrets: Permission denied
Gkr-Message: secret service operation failed: Failed to execute program org.freedesktop.secrets: Permission denied
[1472:1507:0628/155110.465805:ERROR:connection_factory_impl.cc(386 )] Failed to connect to MCS endpoint with error -105
[1472:1507:0628/155110.565353:ERROR:checkin_request.cc(211)] Failed to get checkin response. Fetcher failed. Retrying.
[1:11:0628/155113.451191:ERROR:adm_helpers.cc(62)] Failed to query stereo recording.
[1472:1507:0628/155133.655115:ERROR:checkin_request.cc(211)] Failed to get checkin response. Fetcher failed. Retrying.
[1472:1507:0628/155136.197255:ERROR:connection_factory_impl.cc(386 )] Failed to connect to MCS endpoint with error -105
[1472:1472:0628/155143.359447:ERROR:account_info_fetcher.cc(58)] OnGetTokenFailure: Request canceled.
$ sudo dmesg
[ 115.871098] audit: type=1106 audit(1498690240.761:202): pid=1465 uid=0 auid=1000 ses=2 subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 msg='op=PAM:session_close grantors=pam_keyinit,pam_limits,pam_keyinit,pam_li mits,pam_systemd,pam_unix acct="root" exe="/usr/bin/sudo" hostname=? addr=? terminal=/dev/pts/0 res=success'
[ 115.871101] audit: type=1104 audit(1498690240.761:203): pid=1465 uid=0 auid=1000 ses=2 subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 msg='op=PAM:setcred grantors=pam_unix acct="root" exe="/usr/bin/sudo" hostname=? addr=? terminal=/dev/pts/0 res=success'
[ 177.568237] chrome[1548]: segfault at 0 ip 00007efcb2436795 sp 00007fffe17552e0 error 4 in nouveau_dri.so[7efcb1f00000+aab000]
[ 177.568315] audit: type=1701 audit(1498690302.456:204): auid=1000 uid=1000 gid=1000 ses=2 subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 pid=1548 comm="chrome" exe="/opt/google/chrome/chrome" sig=11 res=1
[ 191.197228] audit: type=1123 audit(1498690316.084:205): pid=1694 uid=1000 auid=1000 ses=2 subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 msg='cwd="/home/vmc" cmd="dmesg" terminal=pts/0 res=success'
[ 191.197232] audit: type=1110 audit(1498690316.084:206): pid=1694 uid=0 auid=1000 ses=2 subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 msg='op=PAM:setcred grantors=pam_env,pam_unix acct="root" exe="/usr/bin/sudo" hostname=? addr=? terminal=/dev/pts/0 res=success'
[ 191.205032] audit: type=1105 audit(1498690316.090:207): pid=1694 uid=0 auid=1000 ses=2 subj=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 msg='op=PAM:session_open grantors=pam_keyinit,pam_limits,pam_keyinit,pam_li mits,pam_systemd,pam_unix acct="root" exe="/usr/bin/sudo" hostname=? addr=? terminal=/dev/pts/0 res=success'

The segfault only happens on fedora. ubuntu, debian, and suse are ok.

hardware:
amd athlon 11 x2 250
4gb ram
nvidia geforce 6159 se

edit:

$ journalctl -p err -b|grep chrome
Jun 28 15:51:42 localhost.localdomain abrt-hook-ccpp[1682]: Process 1548 (chrome) of user 1000 killed by SIGSEGV - dumping core
Jun 28 15:56:04 cpe-23-241-175-59.socal.res.rr.com abrt-hook-ccpp[2115]: Process 1976 (chrome) of user 1000 killed by SIGSEGV - dumping core
Jun 28 16:14:46 cpe-23-241-175-59.socal.res.rr.com abrt-hook-ccpp[2527]: Process 2117 (chrome) of user 1000 killed by SIGSEGV - dumping core

Not sure what nouveau_dri.so has to do with it. googling with little relevant info.

nobody
29th June 2017, 05:20 AM
The segfault only happens on fedora. ubuntu, debian, and suse are ok.

hardware:
amd athlon 11 x2 250
4gb ram
nvidia geforce 6159 se

Not sure what nouveau_dri.so has to do with it. googling with little relevant info.

It is very hard to imagine what kind of error/bug produces this.

Hello Dog,

Let me try to imagine. It is very tough thread, but, nothing to be scared of.

Let me go through possibilities, one by one.

[1] nvidia geforce 6159 se is old hardware, as I see. I guess, from time to time they (nvidia) give updates to the 6159 driver: http://nvidia.windowsdriversdownloadutility.com/nvidia_geforce_6159_le_driver.html

This one is for WINdows OSes, but it has, as we see, some limitations. so, do you have the latest nvidia driver for geforce 6159 se?

[2] Fedora differs from other Linux distros you mentioned, since it follows kernel.org by much faster pace then others. As kernel maintainers change interfaces to DRIlib and (beneath) to DRMlib, it is questionable do nvidia maintainers follow these kernel changes.

You should use older kernels (from GRUB2 menu, you by default have them 3), and try to see if older ones work. Bottom line, you also have fallback kernel, and you should try with that one as well, to see if this error repeats (I bet it does NOT). I do NOT see kernel numbers, but I see that this is x86_64 kernel (registers are 64 bit long, and external address bus is 48 bit wide). :D

Here is the task for you: what is the latest kernel your browser is working seamlessly with???

[3] Once you determine the first kernel it fails, the following picture will show you the approximate relationship with your X application, X11 server and GFX components:

https://upload.wikimedia.org/wikipedia/commons/thumb/5/5f/Vaapi_illustration.svg/550px-Vaapi_illustration.svg.png

On it, you see where the potential problem lies with DRI layer... Not sure if you use the video while using browser, problem upon exit very well might be with releasing sources by some System Call (I bet the shared memory causes this problem). Wild guess: https://bugs.chromium.org/p/chromium/issues/detail?id=415681

One more thing you can do, if BIOS allows: to add some 2GB or 4GB of extra DDR2/3 memory, and see if problem persists.

It is so far enough... I have also other potential problems (which might cause this) in my head. Few at the time. :p

_nobody_

nobody
29th June 2017, 05:26 AM
As separate post (on purpose) to this problem:


$ google-chrome
Gkr-Message: secret service operation failed: Failed to execute program org.freedesktop.secrets: Permission denied
Gkr-Message: secret service operation failed: Failed to execute program org.freedesktop.secrets: Permission denied

You might want also to try to switch SElinux off (kernel command line addition: selinux=0) and retry? :D

_nobody_

VernDog
29th June 2017, 06:55 PM
Thanks for your replies!

How to launch chrome with LIBGL_DRI3_DISABLE=1 ?

VernDog
29th June 2017, 07:27 PM
As separate post (on purpose) to this problem:



You might want also to try to switch SElinux off (kernel command line addition: selinux=0) and retry? :D

_nobody_

I switched of SElinux. That didn't work. Switching it back on took a long time in initializes, but I'm back to square-one.

nobody
29th June 2017, 07:57 PM
What about this test? :confused:


[2] Fedora differs from other Linux distros you mentioned, since it follows kernel.org by much faster pace then others. As kernel maintainers change interfaces to DRIlib and (beneath) to DRMlib, it is questionable do nvidia maintainers follow these kernel changes.

You should use older kernels (from GRUB2 menu, you by default have them 3), and try to see if older ones work. Bottom line, you also have fallback kernel, and you should try with that one as well, to see if this error repeats (I bet it does NOT). I do NOT see kernel numbers, but I see that this is x86_64 kernel (registers are 64 bit long, and external address bus is 48 bit wide).

Here is the task for you: what is the latest kernel your browser is working seamlessly with???


BTW, what are the kernels, which are failing with chrome? :D

_nobody_

nobody
29th June 2017, 08:38 PM
[root@localhost ~]# uname -r
4.11.7-300.fc26.x86_64
[root@localhost ~]# google-chrome-stable
[0629/212904.720612:ERROR:nss_util.cc(94)] Failed to create /home/user/.pki/nssdb directory.
--2017-06-29 21:29:04-- https://clients2.google.com/cr/report
Resolving clients2.google.com (clients2.google.com)... 216.58.207.142, 2a00:1450:4016:806::200e
Connecting to clients2.google.com (clients2.google.com)|216.58.207.142|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [text/html]
Saving to: /dev/fd/4

0K
Crash dump id: bcdb019e08000000
530K=0s

2017-06-29 21:29:05 (530 KB/s) - /dev/fd/4 saved [16]

--2017-06-29 21:29:05-- https://clients2.google.com/cr/report
Resolving clients2.google.com (clients2.google.com)... 216.58.207.142, 2a00:1450:4016:806::200e
Connecting to clients2.google.com (clients2.google.com)|216.58.207.142|:443... connected.
HTTP request sent, awaiting response... 200 OK
Length: unspecified [text/html]
Saving to: /dev/fd/4

0K
Crash dump id: efffc7bc68000000
131K=0s

2017-06-29 21:29:06 (131 KB/s) - /dev/fd/4 saved [16]

Illegal instruction (core dumped)
[root@localhost ~]# exit
exit
[user@localhost ~]$ google-chrome-stable
[10356:10389:0629/212911.801974:ERROR:nss_util.cc(706)] Error initializing NSS with a persistent database (sql:/home/user/.pki/nssdb): NSS error code: -8174
libGL error: failed to open drm device: Permission denied
libGL error: failed to load driver: vmwgfx
[10448:10448:0629/212912.191727:ERROR:sandbox_linux.cc(343)] InitializeSandbox() called with multiple threads in process gpu-process.
[user@localhost ~]$


[34511.057678] traps: chrome[9651] trap int3 ip:558d8985d386 sp:7ffcd45687b0 error:0 in chrome[558d88cf8000+6adb000]
[34513.380297] traps: chrome[9642] trap invalid opcode ip:55fc1ced641a sp:7ffc6b374520 error:0 in chrome[55fc1bc0b000+6adb000]
[34561.859225] traps: chrome[9819] trap int3 ip:55d7d76a3386 sp:7ffd0730a560 error:0 in chrome[55d7d6b3e000+6adb000]
[34563.942815] traps: chrome[9810] trap invalid opcode ip:55623f4c341a sp:7fff3b4ed860 error:0 in chrome[55623e1f8000+6adb000]
[34615.054603] traps: chrome[9995] trap int3 ip:565520448386 sp:7ffe836c2b20 error:0 in chrome[56551f8e3000+6adb000]
[34617.190150] traps: chrome[9986] trap invalid opcode ip:555e354b641a sp:7fffa2dce180 error:0 in chrome[555e341eb000+6adb000]
[34640.234067] traps: chrome[10216] trap int3 ip:55b3bfffb386 sp:7ffd9d19e1a0 error:0 in chrome[55b3bf496000+6adb000]
[34642.262423] traps: chrome[10207] trap invalid opcode ip:55a7ae73341a sp:7fff16f803f0 error:0 in chrome[55a7ad468000+6adb000]
[34909.354518] traps: chrome[10628] trap int3 ip:55682c6c3386 sp:7fff51169a30 error:0 in chrome[55682bb5e000+6adb000]
[34911.484324] traps: chrome[10619] trap invalid opcode ip:55c6b750841a sp:7fff1cd456c0 error:0 in chrome[55c6b623d000+6adb000]
[root@localhost ~]#

Running chrome as root, I always also experience a crash, while as user just drm and loadable driver module problems, which do not produce crashes?! :D :D :D

_nobody_

nobody
29th June 2017, 09:03 PM
Thanks for your replies!

How to launch chrome with LIBGL_DRI3_DISABLE=1 ?

This is the actual problem, for you (not sure if it is for me)??? :doh: :doh: :doh:
https://keithp.com/blogs/chromium-dri3/

F@ck G@@gle! :C

_nobody_

VernDog
1st November 2017, 12:25 AM
I just want to report that I found the fix for me. Somewhere there was reference to keyring. So I checked my '/usr/bin/gnome-keyring-daemon' and found it RW only. After adding execute, I no longer have chrome segfault.

edit:also if I executed chrome from a terminal before I fix it, then this error message come up:

Gkr-Message: secret service operation failed: Failed to execute program org.freedesktop.secrets: Permission denied