| View previous topic :: View next topic |
| Author |
Message |
tbart
Apprentice
Joined: 31 Oct 2004
Posts: 151
|
 Posted: Mon Dec 31, 2018 3:13 pm Post subject: [SOLVED] Kernel 4.14+: segfaults, hangup in chromium/firefox |
 |
|
Hi there!
I have a rock solid system here when run on 4.14 kernels. Thinkpad T470s (Intel(R) Core(TM) i7-7600U CPU @ 2.80GHz), integrated intel GPU, crashes happen both when using the modesetting and the intel driver.
Running enlightenment-0.22.4 on X.
After only a few minutes into a browser session (firefox-bin, firefox .tar.gz from mozilla.org, chromium-71.0.3578.80), either the browser crashes or the whole system hangs.
The only thing I get in the logs are segfaults when firefox crashes.
| Code: | [ 518.818032] Chrome_~dThread[4366]: segfault at 0 ip 00007fc566be8861 sp 00007fc574f19a90 error 6 in libxul.so[7fc56384e000+48d8000]
[ 518.818039] Code: 8b 0d 0b 75 42 03 48 89 01 c7 04 25 00 00 00 00 73 01 00 00 e8 70 14 c7 fc 48 8d 05 1d 9c 68 01 48 8b 0d ea 74 42 03 48 89 01 <c7> 04 25 00 00 00 00 6d 0a 00 00 e8 4f 14 c7 fc 66 2e 0f 1f 84 00
[ 518.818260] Chrome_~dThread[4474]: segfault at 0 ip 00007f95b5f49861 sp 00007f95c4311a90 error 6 in libxul.so[7f95b2baf000+48d8000]
[ 518.818267] Code: 8b 0d 0b 75 42 03 48 89 01 c7 04 25 00 00 00 00 73 01 00 00 e8 70 14 c7 fc 48 8d 05 1d 9c 68 01 48 8b 0d ea 74 42 03 48 89 01 <c7> 04 25 00 00 00 00 6d 0a 00 00 e8 4f 14 c7 fc 66 2e 0f 1f 84 00
[ 518.819280] Chrome_~dThread[4488]: segfault at 0 ip 00007ff2f7749861 sp 00007ff305ac0a90 error 6 in libxul.so[7ff2f43af000+48d8000]
[ 518.819286] Chrome_~dThread[4498]: segfault at 0 ip 00007f3de6c49861 sp 00007f3df4f78a90 error 6
[ 518.819288] Code: 8b 0d 0b 75 42 03 48 89 01 c7 04 25 00 00 00 00 73 01 00 00 e8 70 14 c7 fc 48 8d 05 1d 9c 68 01 48 8b 0d ea 74 42 03 48 89 01 <c7> 04 25 00 00 00 00 6d 0a 00 00 e8 4f 14 c7 fc 66 2e 0f 1f 84 00
[ 518.819291] in libxul.so[7f3de38af000+48d8000]
[ 518.819298] Code: 8b 0d 0b 75 42 03 48 89 01 c7 04 25 00 00 00 00 73 01 00 00 e8 70 14 c7 fc 48 8d 05 1d 9c 68 01 48 8b 0d ea 74 42 03 48 89 01 <c7> 04 25 00 00 00 00 6d 0a 00 00 e8 4f 14 c7 fc 66 2e 0f 1f 84 00
[ 518.819503] Chrome_~dThread[4391]: segfault at 0 ip 00007fe076349861 sp 00007fe08470ca90 error 6 in libxul.so[7fe072faf000+48d8000]
[ 518.819508] Code: 8b 0d 0b 75 42 03 48 89 01 c7 04 25 00 00 00 00 73 01 00 00 e8 70 14 c7 fc 48 8d 05 1d 9c 68 01 48 8b 0d ea 74 42 03 48 89 01 <c7> 04 25 00 00 00 00 6d 0a 00 00 e8 4f 14 c7 fc 66 2e 0f 1f 84 00
[ 710.614810] Web Content[5210]: segfault at 7fa83c6168c0 ip 00007fa7db496dd6 sp 00007ffc02afa850 error 4 in libxul.so[7fa7d864e000+48d8000]
[ 710.614814] Code: f0 31 c8 44 69 e8 b9 79 37 9e 49 69 cd 23 f7 f2 0a 48 c1 e9 20 44 89 ee 29 ce d1 ee 01 ce c1 ee 08 69 ce eb 01 00 00 41 29 cd <4e> 8b b4 eb a8 00 00 00 4d 85 f6 74 17 4d 39 26 75 12 4c 89 f0 48 |
No kernel-related errors whatsoever.
Mostly the system will fully hang, not even pingable. I also do not see anything useful when using dmesg -w from another machine.
Other than using browsers, I can emerge world (including the hefty chromium build) or do other random stuff without any error.
Why do I use 4.19? 4.14 has some errors related to my GPU (flickers, short image distortions, error messages regarding cpu hangs), which are gone with 4.19.
I manually configure and compile my kernel. Every kernel in the 4.19.* series shows the same behavior.
I'd very much appreciate how to track this down, so I can report it correctly.
Thanks in advance!
Kernel config: https://pastebin.com/PmMNiNcV
Last edited by tbart on Tue Jan 01, 2019 1:39 pm; edited 3 times in total |
|
| Back to top |
|
 |
Jaglover
Watchman
Joined: 29 May 2005
Posts: 8291
Location: Saint Amant, Acadiana
|
| Posted: Mon Dec 31, 2018 3:20 pm Post subject: |
|
|
Please remove this config from your post, people hate all that scrolling down for nothing and kernel config won't fit the forum post anyway. Use wgetpaste if you need to post it.
However, I'd suspect a hardware issue here, possibly RAM. Your new kernel probably is just exposing the problem you have.
_________________
My Gentoo installation notes.
Please learn how to denote units correctly! |
|
| Back to top |
|
|
tbart
Apprentice
Joined: 31 Oct 2004
Posts: 151
|
| Posted: Mon Dec 31, 2018 4:30 pm Post subject: |
|
|
Sorry, corrected that.
This does not seem too likely, as I've really maxed out my 16G of RAM with compilation stuff recently, either with 4.19 (tried 4.20 just now as well) without a browser, or on 4.14 even with a browser.
I'll do a memtest nevertheless. |
|
| Back to top |
|
|
tbart
Apprentice
Joined: 31 Oct 2004
Posts: 151
|
| Posted: Tue Jan 01, 2019 12:48 pm Post subject: |
|
|
Alright, full memtest86 over night (4 passes, all 13 test patterns)
Lenovo supplied hardware test: CPU, Mainboard, Storage, PCIe, RAM
All passed without errors.
I somehow need to rule out stuff, but I am unsure where to start!
Segfaults normally only occur due to programming errors. Considering how stable firefox and chromium are on other boxes, I doubt the error is on these products (as they work flawlessly on a 4.14.x kernel).
Both browser have a default config and no extensions for testing, errors are still there.
I think I'll try 4.15, 4.16 and 4.17 next to further narrow down the release. Or are there any tools that might help in getting some kind of debug output? |
|
| Back to top |
|
|
tbart
Apprentice
Joined: 31 Oct 2004
Posts: 151
|
| Posted: Tue Jan 01, 2019 1:38 pm Post subject: |
|
|
Sorry for the noise, problem seems to be solved.
This CPU and GPU are undervolted, and although the mentioned combinations did not crash a single time over the past months, it seems I have to increase either voltage for kernels after 4.14. Which is sad in a way as I nearly got a full hour of longer battery life by undervolting 
That's not exactly an energy efficient development as it stands. I am pretty sure this won't be regarded as a valid bug report so I'll most likely not report it. Which is a pity. |
|
| Back to top |
|
|
Hu
Administrator
Joined: 06 Mar 2007
Posts: 23062
|
| Posted: Wed Jan 02, 2019 1:47 am Post subject: |
|
|
| What settings were you using that were unstable before? What settings are you using now to make the system stable on new kernels? I am curious by how much you undervolted it relative to the documented capabilities of the hardware. |
|
| Back to top |
|
|
tbart
Apprentice
Joined: 31 Oct 2004
Posts: 151
|
| Posted: Thu Jan 03, 2019 6:25 pm Post subject: |
|
|
-100mV for both CPU and GPU. Which is a lot, considering this CPU has a mere 600mV Vcore when idling!
With every kernel 4.19+ I cannot even get -50mV to run reliably, which is a shame.
Undervolting on this machine has another massive pro. The cooling system is not really capable of getting rid of all of the dissipated power of the CPU when running on full turbo on both cores.
When undervolting I can get to 3.5GHz (of 3.9GHz) instead of just 3.2GHz because of thermal capping (70°C, untouched stock firmware config). Upping the temperature to 90°C and undervolting gets me to 3.8GHz, but I don't really feel it's worth the increased aging and energy consumption for those last 300MHz.
Appling better paste might be an option as well but I didn't feel like opening a brand new machine till now  |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Kernel & Hardware
