Trainwreck crashlog. Usually happens on cell change or save.

[DocClox]

Here's the log. I can generally puzzle something out of these things but I'm not getting much from this one. Although there are a lot of size_t values on the stack with -1 values.

 

Here's my current load order.

 

I should also add that this is running on Linux using Proton GE 8-30. It was all working just fine. But then I tried to set up a second profile, which involved updating some shared mods and that broke the load order on my first profile, and now while it works it crashes a little too frequently.

 

Anyway, any and all help appreciated.

[traison]

The reason why it crashed was because it tried to dereference RDI which at the time of the crash was 0x128 - way outside the memory address space of the process, hence the access denied error. The slightly unfortunate thing about this is that 0x128 is not 0x0 (a traditional null-pointer). I'd say simply because of that this issue looks a bit different than the other crash logs I've seen on LL - its a unique characteristic, but you probably won't find anything on Google with it.

 

SoakingWet has a CTD causing issue in it, disable that, as you won't need it right now anyways. Edit: It's rare, and its fingerprint is different, but it is in the stack so: make it gone.

 

One typical situation where these BSMultiBoundRoom, ShadowSceneNode and BSParabolicCullingProcess keywords show up is when there's an incompatible dll installed. Another thing that may cause it is a lack of BEES (not applicable to your game version) and a misconfigured Engine Fixes SSE. Perhaps start with this and ifi t doesn't work I'll consider having a proper look at the crash log.

Edited March 30, 2024 by traison

[DocClox]

Engine Fixes could be a culprit. I had the problem where it crashed the game and cleared plugins.txt every time. I turned off the settings that were supposed to cause that particular problem, but the game still wouldn't start. So I renamed the d3dx9_42.dll loader and unloaded the SKSE mod. Maybe the tbbmalloc.dll is acting up?

 

I'll try disabling soaking wet. It's funny, I loaded that as a replacement for Wet And Cold as that was supposedly causing crashes. Never mind, as you say, I don't need it.

[traison]

4 minutes ago, DocClox said:

Maybe the tbbmalloc.dll is acting up?

 

I'd say with the Proton layer in between, and just generally running software on a kernel where it wasn't supposed to run is going to introduce all kinds of unusual and nonsensical issues. So lets put this way: I wouldn't blame tbbmalloc just because it shows up in practically every crash log - its not an unstable file - but it is doing something that Proton or the Linux kernel doesn't like? Seems unlikely, but not impossible. Unlikely because based on the name, this is some kind of memory allocation provider. Memory is allocated probably a million times a second - if it was going to break, I doubt it would break in... well whatever situation you were in when it broke. A bit like whats the odds of you getting hit by a meteorite while sitting in your sofa - not gonna happen right?

 

7 minutes ago, DocClox said:

I'll try disabling soaking wet. It's funny, I loaded that as a replacement for Wet And Cold as that was supposedly causing crashes.

 

SoakingWet is not an unstable mod. I've only ever had one situation where it crashed - it was a repeatable crash. Usually I just like to get the noise makers out of crash logs, especially if they're easy to remove - such as ENB, CBPC and SMP. Usually it makes no difference, but it also cleans up the next crash log - makes it easier to read because the information in it is more relevant.

[DocClox]

2 hours ago, traison said:

Unlikely because based on the name, this is some kind of memory allocation provider

 

Must admit, it wouldn't have been my first suspect either. I only mentioned it since it's the only part of engine fixes left - well that and tbb.dll, anywy.

 

3 hours ago, traison said:

Usually I just like to get the noise makers out of crash logs

 

Sensible enough.

 

Oddly enough, disabling Soaking Wet seems to have left things a bit more stable. I've had a couple of crashes; one of them after leaving the game paused on the console for an our after being called away unexpectedly (and then having to alt-tab to fix something on the other monitor), the other on exit from the game. The crash-on-exit I can live with, and the other one would have been a good way to get a crash, even if things were more or less perfect.

 

I#ll play on for a bit, I think, and see if I get a better candidate.

[DocClox]

Here's a repeatable crash trying to  travel to Falkreath. Tast travel says fast travel is unavailable, teleport with the Bat Travel mod crashes the game, as does taking a carriage. I suspect this particular save has issues. Anyway: Crashlog.

[traison]

1 hour ago, DocClox said:

Here's a repeatable crash trying to  travel to Falkreath.

 

Did you modify that log?

 

Edit: "What kind of a stupid question is that?" I hear you say. I mean the callstack and the stack are from 2 different crashes. They do not line up. Only the first (callstack entry #1) is correct:

 

SkyrimSE.exe+0624FE0 (faulting offset, or RIP, should not be part of the callstack if you ask me)

SkyrimSE.exe+0172E61

SkyrimSE.exe+0224E5E

SkyrimSE.exe+023943F

SkyrimSE.exe+02DA9E0

SkyrimSE.exe+0686F04

...

 

Then if I do the unwinding by hand:

 

✅SkyrimSE.exe+0172E61
❌SkyrimSE.exe+1AFA67E

❌SkyrimSE.exe+0CD0393

❌SkyrimSE.exe+01E022A
❌SkyrimSE.exe+0151853
❌SkyrimSE.exe+0CC41B2

...

 

3 options then:

1. There's something here that I don't understand.
2. Something went horribly wrong.
3. 3rd party interference.

 

Edit again: Give this thing a try instead: https://www.nexusmods.com/skyrimspecialedition/mods/59818

Edited April 1, 2024 by traison

[DocClox]

Well, I didn't edit the crash log, so no idea what happened. I'll try and get a log with that crashlogger instead.

 

Thanks for looking at this.