[SOLVED] Frequent Browser/Electron crashes

NoisyFlake · 2023-11-29 07:22:44

Hi everyone,

ever since I switched to Arch a few weeks ago, my Firefox frequently crashes. Most of the time it's just a single tab, but less often the whole browser crashes. I don't want to switch to a Chromium-based browser, but the crashes every day are getting pretty annoying.
There doesn't seem to be any pattern for the crashes, sometimes it's even happening while idling (e.g. reading a news article).

Things I've tried without any luck so far:
- Running Firefox in safe-mode without any extensions
- Switching from regular to LTS kernel
- Disable Hardware Acceleration in Firefox
- Change RAM speed and timings
- Run Memtest successfully
- Replace entire RAM with a new certified kit
- Use only a single RAM slot
- Apply Ryzen fixes (iommu=soft, limit c-states)
- Use only a single CPU core (maxcpus=1)
- Downgrade Nvidia driver to 535xx
- Use Nouveau instead of the nvidia driver
- Use Openbox instead of KDE
- Disable zswap and THP

Here are the two latest Firefox tab crashes:

https://crash-stats.mozilla.org/report/ … c0e0231129
https://crash-stats.mozilla.org/report/ … e3d0231128

And the two latest whole browser crashes:

https://crash-stats.mozilla.org/report/ … 76c0231129
https://crash-stats.mozilla.org/report/ … 7940231129

And a full journalctl from a day until both Spotify and Firefox crashed a few seconds after each other:

https://pastebin.com/BH0LMnD9

Some more info about my system:

- Ryzen 5 3600X
- MSI B450M PRO-VDH Max
- 32GB RAM @ 3200MHz
- Geforce RTX 2070 SUPER (using nvidia-dkms)
- Plasma 5.27.10 on X11

Let me know if you need any more information that might help solve this issue.
Thanks!

Edit: I'm pretty sure that it's not hardware related, because I've booted up a Debian 12 live image where everything ran for several hours without a crash. But it seems to be Arch related, as I also booted up a fresh EndeavourOS live image (so basically Arch), where applications also randomly segfaulted. Any idea why everything works fine on Debian but not on Arch? Debian uses the 6.1 kernel, which I already tried, so that's not it.

Edit: It's not just Electron-based apps, but all apps. I've had plasmashell and old games like Half-Life also crash.

EDIT (SOLUTION): I had to disable Core Performance Boost in the UEFI/BIOS settings. After that, no more crashes.

Last edited by NoisyFlake (2024-02-07 16:50:33)

Durden · 2023-11-29 17:22:03

Delete the Firefox profile on in your user account? It's usually at ~/.mozilla (Back it up first of course)

You may also see a "Crash Reports" folder in there with some logs that may help you find the problem. In settings under general you may be able to toggle off things like DRM to see if thats the problem, smooth scrolling also. There is a section about letting Firefox use the recommended performance settings. You can toy with that too and see if its a hardware acceleration issue.

That's all I've got

Last edited by Durden (2023-11-29 17:22:25)

xerxes_ · 2023-11-29 20:54:06

Post output of commands:

coredumpctl
coredumpctl info
coredumpctl dump

You can start Firefox from terminal and see if there are some errors in terminal when it crash.
You can go to about:support in address bar in Firefox and see at the end of "Graphics" section if there are some errors (or post here that section in [ code ] tags [ / code ] (without spaces)).
You can try disable hardware acceleration by going to about:config and setting:

gfx.webrender.all - false
gfx.webrender.enabled - false
gfx.webrender.software - true
gfx.webrender.software.opengl - true
webgl.disabled - true
webgl.disabled-wgl - true
layers.acceleration.force-enabled - false

and restart Firefox. To see results go to about:support and see in section "Graphics" what is enabled and what not.

seth · 2023-11-29 21:42:29

The crashes on crash-stats.mozilla.org are all over the place. Probably RAM or CPU.

I don't want to switch to a Chromium-based browser

Ok, don't switch, but install chromium and see whether it shows similar instabilities (assuming it'll cause similar RAM/CPU pressure)
Also https://wiki.archlinux.org/title/Stress … MemTest86+
For 32GB that probably has to run for 24h+ to produce meaningful results.
You can go on a limb, configure the RAM for more conservative timings and frequencies and see whether that stabilizes FF.

NoisyFlake · 2023-11-30 09:04:55

Durden wrote:

Delete the Firefox profile on in your user account? It's usually at ~/.mozilla (Back it up first of course)

Shouldn't that be accomplished by running Firefox in safe-mode? Or is this different?

xerxes_ wrote:

Post output of commands:
coredumpctl
coredumpctl info
coredumpctl dump

coredumpctl:

Wed 2023-11-29 07:36:58 CET   9418 1000 1000 SIGSEGV present  /usr/lib/firefox/firefox                                                                              146.5M
Wed 2023-11-29 08:12:32 CET   9203 1000 1000 SIGSEGV present  /usr/lib/firefox/firefox                                                                              170.2M
Wed 2023-11-29 09:53:39 CET  16433 1000 1000 SIGSEGV present  /usr/lib/firefox/firefox                                                                              152.2M
Wed 2023-11-29 10:19:24 CET   2126 1000 1000 SIGABRT present  /home/noisyflake/.local/share/Steam/ubuntu12_32/steam                                                  45.7M
Wed 2023-11-29 10:19:32 CET   2144 1000 1000 SIGILL  present  /home/noisyflake/.local/share/Steam/ubuntu12_64/steamwebhelper                                          5.6M
Wed 2023-11-29 10:19:55 CET  36988 1000 1000 SIGSEGV present  /usr/lib/firefox/firefox                                                                              157.7M
Wed 2023-11-29 12:30:52 CET 123376 1000 1000 SIGSEGV present  /usr/lib/firefox/firefox                                                                              156.3M
Wed 2023-11-29 12:49:58 CET 123265 1000 1000 SIGSEGV present  /usr/lib/firefox/firefox                                                                              145.7M
Wed 2023-11-29 12:56:16 CET 180705 1000 1000 SIGSEGV present  /usr/lib/firefox/firefox                                                                              147.3M

coredumpctl info:

           PID: 180705 (Isolated Web Co)
           UID: 1000 (noisyflake)
           GID: 1000 (noisyflake)
        Signal: 11 (SEGV)
     Timestamp: Wed 2023-11-29 12:56:12 CET (20h ago)
  Command Line: /usr/lib/firefox/firefox -contentproc -childID 173 -isForBrowser -prefsLen 31503 -prefMapSize 236509 -jsInitLen 228948 -parentBuildID 20231120231815 -greomni /usr/lib/firefox/omni.ja -app>
    Executable: /usr/lib/firefox/firefox
 Control Group: /user.slice/user-1000.slice/user@1000.service/app.slice/app-pactl-95ea5a6be9364aac8fde0eb4462832d8.scope
          Unit: user@1000.service
     User Unit: app-pactl-95ea5a6be9364aac8fde0eb4462832d8.scope
         Slice: user-1000.slice
     Owner UID: 1000 (noisyflake)
       Boot ID: f4939c7aefd1431392e785d82db63c3b
    Machine ID: 7e142e2f5e47437c82f7d1297a941601
      Hostname: arch
       Storage: /var/lib/systemd/coredump/core.Isolated\x20Web\x20Co.1000.f4939c7aefd1431392e785d82db63c3b.180705.1701258972000000.zst (present)
  Size on Disk: 147.3M
       Message: Process 180705 (Isolated Web Co) of user 1000 dumped core.
                
                Stack trace of thread 182653:
                #0  0x00007f73fbca3f86 n/a (libxul.so + 0x34a3f86)
                #1  0x00007f73fbc979b4 n/a (libxul.so + 0x34979b4)
                #2  0x00007f73fc363ed9 n/a (libxul.so + 0x3b63ed9)
                #3  0x00007f73fbce5288 n/a (libxul.so + 0x34e5288)
                #4  0x00007f73fbcdfdad n/a (libxul.so + 0x34dfdad)
                #5  0x00007f73fbcdeba1 n/a (libxul.so + 0x34deba1)
                #6  0x00007f73fcbf1790 n/a (libxul.so + 0x43f1790)
                #7  0x00007f73fcbf15c0 n/a (libxul.so + 0x43f15c0)
                #8  0x00007f73fc776dba n/a (libxul.so + 0x3f76dba)
                #9  0x00007f73ff3d742f n/a (libxul.so + 0x6bd742f)
                #10 0x00007f73fc4ce349 n/a (libxul.so + 0x3cce349)
                #11 0x00007f73fc67007b n/a (libxul.so + 0x3e7007b)
                #12 0x00007f73fbf07bcf n/a (libxul.so + 0x3707bcf)
                #13 0x00007f73fbcb3e1c n/a (libxul.so + 0x34b3e1c)
                #14 0x00007f73fc60c839 n/a (libxul.so + 0x3e0c839)
                #15 0x00007f7406cd2c7b n/a (libnspr4.so + 0x2bc7b)
                #16 0x0000556928e620c3 n/a (firefox + 0x720c3)
                #17 0x00007f74068aa9eb n/a (libc.so.6 + 0x8c9eb)
                #18 0x00007f740692e654 __clone (libc.so.6 + 0x110654)
                
                Stack trace of thread 180714:
                #0  0x00007f74068a74ae n/a (libc.so.6 + 0x894ae)
                #1  0x00007f74068aa055 pthread_cond_timedwait (libc.so.6 + 0x8c055)
                #2  0x0000556928e966ff _ZN7mozilla6detail21ConditionVariableImpl8wait_forERNS0_9MutexImplERKNS_16BaseTimeDurationINS_27TimeDurationValueCalculatorEEE (firefox + 0xa66ff)
                #3  0x00007f73fccab62f n/a (libxul.so + 0x44ab62f)
                #4  0x00007f73fbf07bcf n/a (libxul.so + 0x3707bcf)
                #5  0x00007f73fbcb3e1c n/a (libxul.so + 0x34b3e1c)
                #6  0x00007f73fc60c839 n/a (libxul.so + 0x3e0c839)
                #7  0x00007f7406cd2c7b n/a (libnspr4.so + 0x2bc7b)
                #8  0x0000556928e620c3 n/a (firefox + 0x720c3)
                #9  0x00007f74068aa9eb n/a (libc.so.6 + 0x8c9eb)
                #10 0x00007f740692e7cc n/a (libc.so.6 + 0x1107cc)

                [...] more stack traces here, let me know if you need them

coredumpctl dump gives the same result as coredumpctl info.

xerxes_ wrote:

You can go to about:support in address bar in Firefox and see at the end of "Graphics" section if there are some errors (or post here that section in [ code ] tags [ / code ] (without spaces)).

There are no errors in that section.

xerxes_ wrote:

You can start Firefox from terminal and see if there are some errors in terminal when it crash.
You can try disable hardware acceleration by going to about:config and setting:
gfx.webrender.all - false
gfx.webrender.enabled - false
gfx.webrender.software - true
gfx.webrender.software.opengl - true
webgl.disabled - true
webgl.disabled-wgl - true
layers.acceleration.force-enabled - false
and restart Firefox. To see results go to about:support and see in section "Graphics" what is enabled and what not.

I'll try that and report back.

seth wrote:

The crashes on crash-stats.mozilla.org are all over the place. Probably RAM or CPU.

That's what I thought as well, but if that's really the case, shouldn't it affect other applications? I'm only having problems with Firefox, and before I switched to Arch, I ran Firefox on Windows 10 just fine.
Thought I gotta admit, my Windows 10 crashed about once a month with a bluescreen.

seth wrote:

Also https://wiki.archlinux.org/title/Stress … MemTest86+
For 32GB that probably has to run for 24h+ to produce meaningful results.

I ran MemTest about half a year ago because of the mentioned Windows crashes, but everything was fine. Though I think I only let it ran for about 10-12 hours.

seth wrote:

You can go on a limb, configure the RAM for more conservative timings and frequencies and see whether that stabilizes FF.

Alright, just did that. Ran my DDR4-3200 RAM with 3200MHz using the X-AMP profile so far, I just reduced the frequency to 2933. Do you think that should be enough, or should I go even further?

Edit: A tab just crashed again with 2933. Guess I'll have to go lower? But what good does 3200MHz RAM do if I can't use the full 3200 reliably?

Last edited by NoisyFlake (2023-11-30 09:54:25)

Lone_Wolf · 2023-11-30 10:49:45

Is the memory stick brand & model verified to work on your motherboard/processor combo at 3200 MHz by Motherboard or stick manufacturer ?

seth · 2023-11-30 13:59:58

sudo dmidecode

It does good if we can attribute the problem there.
The strategy would be to pick the slowest frequency and timings possible, see whether that stabilizes it and if so, work up from there.

The "special" thing about FF is probably that it uses more RAM than everything else, causing unstable DIMM combinations.
In theory.

NoisyFlake · 2023-11-30 14:16:45

Lone_Wolf wrote:

Is the memory stick brand & model verified to work on your motherboard/processor combo at 3200 MHz by Motherboard or stick manufacturer ?

I just looked it up, and apparently this specific stick type is not listed as officially supported, ugh. Though a ton of other sticks with the same size, frequency and brand are certified, but I guess that doesn't mean anything.

seth wrote:

sudo dmidecode
It does good if we can attribute the problem there.
The strategy would be to pick the slowest frequency and timings possible, see whether that stabilizes it and if so, work up from there.

Memory Device
        Array Handle: 0x0011
        Error Information Handle: 0x0018
        Total Width: 64 bits
        Data Width: 64 bits
        Size: 8 GB
        Form Factor: DIMM
        Set: None
        Locator: DIMM 0
        Bank Locator: P0 CHANNEL A
        Type: DDR4
        Type Detail: Synchronous Unbuffered (Unregistered)
        Speed: 2933 MT/s
        Manufacturer: Unknown
        Serial Number: 00000000
        Asset Tag: Not Specified
        Part Number: F4-3200C16-8GTZRX
        Rank: 1
        Configured Memory Speed: 2933 MT/s
        Minimum Voltage: 1.2 V
        Maximum Voltage: 1.2 V
        Configured Voltage: 1.2 V

The other 3 sticks are exactly the same. I'll try lowering it and see if it improves.

seth wrote:

The "special" thing about FF is probably that it uses more RAM than everything else, causing unstable DIMM combinations. In theory.

Hm, I think games like Apex Legends will definitely use more RAM than Firefox (which uses about 1-2GB on average for me).

seth · 2023-11-30 15:42:49

Hence also the suggestion to

seth wrote:

install chromium and see whether it shows similar instabilities (assuming it'll cause similar RAM/CPU pressure)

NoisyFlake · 2023-12-01 08:40:08

I went down to as low as 2133MHz, at which point the system became basically unusable because about 20% of the time, keyboard strokes were ignored. No idea how this can even happen, but I'm back up again to the original 3200 and will test Chromium now.

NoisyFlake · 2023-12-01 17:01:31

So I tested Chromium today, and it seems that it's really not a Firefox issue. A tab crashed once, and the whole browser once as well.

Does this mean this is definitely a RAM issue? So far, no other application did crash even once, and I'm gaming almost every day. Also Electron-based apps like Discord don't crash.

seth · 2023-12-01 17:18:55

If you want to know whether it's the RAM, you'll have to memtest86 it for a while.
Can you remove 2-3 DIMMs and see how the remaining one(s) perform?
Also, did yuo only alter the frequency or also the timings?

And oc. there's https://wiki.archlinux.org/title/Ryzen#Troubleshooting - though the effects of that don't tend to be process-isolated.

NoisyFlake · 2023-12-02 15:47:39

I ran Memtest86+ for about 15 hours (I needed the PC after that) and it had 12 passes with 0 errors.

I only altered the frequency and enabled/disabled XMP, as I have no knowledge about tweaking timings manually. I'm really running out of ideas here.

seth · 2023-12-02 15:49:26

You just use the slowest timings you can set, that should™ be some option in your BIOS (but you'll likely have to disable xmp to get access to it)

NoisyFlake · 2023-12-04 09:55:26

Some updates:

I figured out how to manually set the timings, but setting it to anything else (slower or faster) than the ones recommended by the RAM vendor (16-18-18-18-38) results in the system freezing here and there for a second, so it's basically unusable.
Using the recommended timings with any frequency (I went as low as 1600MHz) still results in my browser crashing often.

I've now removed 2 DIMMs and will see if anything changes. At this point, are you sure that it has to be a RAM issue? Or could it be something entirely different?

Last edited by NoisyFlake (2023-12-04 10:04:14)

NoisyFlake · 2023-12-04 10:45:31

Another update: A tab just crashed again with only 2 DIMMs (2x8GB) installed. Any other things I could try?

seth · 2023-12-04 11:29:26

You could throw https://archlinux.org/packages/extra/x86_64/stress/ at the system, and try "iommu=soft" and limit the c-states, you saw https://wiki.archlinux.org/title/Ryzen#Troubleshooting ?

the system freezing here and there for a second

1. that's rather odd
2. but does it also crash?

NoisyFlake · 2023-12-07 09:16:29

Yeah, I saw the Ryzen troubleshooting page, but fail to understand how any of it is relevant for my case? I'm not experiencing any of the problems listed in that article.

seth wrote:

the system freezing here and there for a second
1. that's rather odd
2. but does it also crash?

I'm giving it another try right now, I'll let you know.

I'll run the stress test soon though.

Last edited by NoisyFlake (2023-12-07 11:10:41)

seth · 2023-12-07 14:15:55

Ryzen is notoriously unstable and possible mitigations (depending on the cause) are to
* disable the IOMMU
* limit the c-states
* increase the CPU voltage

If either of that helps, it also hints at the cause of your specific situation.

NoisyFlake · 2023-12-08 08:45:53

For the last 24h, I didn't experience any crashes with the default timings (and the default 2133MHz). I also did a BIOS update, but I'm not sure if it has anything to do with it since my BIOS was already pretty up-to-date (from mid 2023).

What's weird is that the micro-freezes I talked about are apparently starting to disappear, meaning when I first started the PC with these settings, it would freeze every few seconds, and it got better and better over time. A few hours later and it's like they don't even exist at all, only sometimes a mouse click or scroll gets "ignored" because I executed it during the freeze.

I'll keep testing a bit longer and will apply the Ryzen fixes if a crash does happen again though.

Last edited by NoisyFlake (2023-12-08 09:01:20)

NoisyFlake · 2023-12-11 12:37:31

So I went back up to 3200MHz with the XMP timings for testing purposes, and it crashed, even with the Ryzen fixes applied. I'm now back to default timings at 2133MHz and hope that it's fixed for good.

But the freezes on these timings still are a bit annoying. As I said, they get better over time, but on the initial login screen they are the worst. Most of the time I have to put in my password very carefully, as sometimes a character input is not recognized and therefore skipped, and sometimes I type a single character and the system seems to miss the key release event, resulting in a continuous input of this character until I press it again. Any idea what could cause this?

Edit: Oh man, it just crashed again for the first time with 2133MHz and default timings. So I guess it isn't related to the RAM at all?

Last edited by NoisyFlake (2023-12-11 12:59:45)

seth · 2023-12-11 16:31:57

Using the recommended timings with any frequency (I went as low as 1600MHz) still results in my browser crashing often.

setting it to anything else (slower or faster) than the ones recommended by the RAM vendor (16-18-18-18-38) results in the system freezing here and there for a second

The RAM configuration clearly has some impact here.
Remove all but one DIMM and test them one at a time, keep the frequency at default/auto and use the most conservative possible timings.
If you still get occasional freezes, please post a system journal covering those.

NoisyFlake · 2024-01-02 14:53:47

Because I was fed up with this issue, I ordered new RAM where the exact model was marked as compatible with my CPU and Mainboard. Inserted it, activated X-AMP so that it runs at the certificied 3200MHz, and not an hour later, the first Firefox tab crashed again.

So I'm pretty certain that this cannot be a RAM issue, as it's brand new and certified to work with my hardware. Any other ideas what could cause this?

Edit: Forgot to mention, I also noticed now that it's not just Firefox/Chrome, but any Electron-based app. I noticed that Discord, Spotify and VSCode also crash randomly (but apparently not as often as the browser). This also happened before the new RAM, so it's nothing new, just wanted to mention it.

Last edited by NoisyFlake (2024-01-02 15:03:30)

xerxes_ · 2024-01-03 19:38:31

Post full journalctl log from boot when crash was, for example: 'sudo journalctl -b', or may it be 'sudo journalctl -b-1'. You can check which by command 'sudo journalctl --list-boots' which will list all available boot logs.

NoisyFlake · 2024-01-05 07:40:26

Here you go: https://pastebin.com/BH0LMnD9

A Spotify and a Firefox tab crash happened right at the end of it, only seconds away from each other.

Arch Linux

#1 2023-11-29 07:22:44

[SOLVED] Frequent Browser/Electron crashes

#2 2023-11-29 17:22:03

Re: [SOLVED] Frequent Browser/Electron crashes

#3 2023-11-29 20:54:06

Re: [SOLVED] Frequent Browser/Electron crashes

#4 2023-11-29 21:42:29

Re: [SOLVED] Frequent Browser/Electron crashes

#5 2023-11-30 09:04:55

Re: [SOLVED] Frequent Browser/Electron crashes

#6 2023-11-30 10:49:45

Re: [SOLVED] Frequent Browser/Electron crashes

#7 2023-11-30 13:59:58

Re: [SOLVED] Frequent Browser/Electron crashes

#8 2023-11-30 14:16:45

Re: [SOLVED] Frequent Browser/Electron crashes

#9 2023-11-30 15:42:49

Re: [SOLVED] Frequent Browser/Electron crashes

#10 2023-12-01 08:40:08

Re: [SOLVED] Frequent Browser/Electron crashes

#11 2023-12-01 17:01:31

Re: [SOLVED] Frequent Browser/Electron crashes

#12 2023-12-01 17:18:55

Re: [SOLVED] Frequent Browser/Electron crashes

#13 2023-12-02 15:47:39

Re: [SOLVED] Frequent Browser/Electron crashes

#14 2023-12-02 15:49:26

Re: [SOLVED] Frequent Browser/Electron crashes

#15 2023-12-04 09:55:26

Re: [SOLVED] Frequent Browser/Electron crashes

#16 2023-12-04 10:45:31

Re: [SOLVED] Frequent Browser/Electron crashes

#17 2023-12-04 11:29:26

Re: [SOLVED] Frequent Browser/Electron crashes

#18 2023-12-07 09:16:29

Re: [SOLVED] Frequent Browser/Electron crashes

#19 2023-12-07 14:15:55

Re: [SOLVED] Frequent Browser/Electron crashes

#20 2023-12-08 08:45:53

Re: [SOLVED] Frequent Browser/Electron crashes

#21 2023-12-11 12:37:31

Re: [SOLVED] Frequent Browser/Electron crashes

#22 2023-12-11 16:31:57

Re: [SOLVED] Frequent Browser/Electron crashes

#23 2024-01-02 14:53:47

Re: [SOLVED] Frequent Browser/Electron crashes

#24 2024-01-03 19:38:31

Re: [SOLVED] Frequent Browser/Electron crashes

#25 2024-01-05 07:40:26

Re: [SOLVED] Frequent Browser/Electron crashes

Board footer