CTD when loading Saves - need Help with Crash Log

[astrum_xeno]

Hello,

 

I'm having a Problem with CTDs when loading Save Files, the crashes aren't happening every Time (basically a 50/50 if it loads or crashes) but the Crash Log Output is always the same so the cause is consistent at least. Here's the Link to the Log: https://pastebin.com/3z3KDVJs

I'm very bad at reading these, and Google wasn't very helpful so far so i hope someone here can help me solve this Issue.

Things i tried so far (based on the little info i could actually "understand"):

-reinstall FSMP

-repair/reinstall Visual c++

 

any Help is greatly appreciated!

[Tlam99]

I never download crashlogs.

Put in a spoiler togetzer with LO.

Mqybe others are willing to download, but not me.

[traison]

Stack is missing from the log.

[astrum_xeno]

I dunno what to say, but this is everything that CrashLogger has spit out sadly. And no matter how many times this crash happens it refuses to give me more info than what's in the pastebin

[traison]

Must be some internal error then. The callstack does not exist without the stack. There's newer versions of Crash Logger available.

[astrum_xeno]

Alright forget what i said i just got another (hopefully more helpful) log... just kept throwing my head at the wall until something happened lol.

 

here's the link: https://pastebin.com/Ab51XcV9

[traison]

This new log is a different crash than the first one. The stack in this new one is all over the place, so not sure what the game is actually doing here.

1. Figure out what form or reference id 0xFF001335 is.
2. Figure out what form id 0x000001C0 is.

These look like they could be vanilla game objects, but they're not present in my 1.5.97 Skyrim.esm. Either they were added in 640, or there's something else going on here.

 

Your issue may be in form id 0xFF000EF9, reference id 0xFF000EF9 "Cragkin Frostbite Goblin Melee".

[astrum_xeno]

huh, that's very weird. i changed literally nothing about my game since my first post.

the "cragkin frostbite goblin melee" is from the demonic creatures mod, I'll have a look about that. still, it leaves open the question what the other crash is about then and why it only produces such short incomplete logs... you said there's a newer version of crash logger available, haven't looked in a while, i'll try updating that and see if anything changes.

 

in any case thanks for trying to help!

[astrum_xeno]

So i tried a few more things, updated crash logger, there was an update for demonic creatures available, installed that and so far i could not get another crash related to it, but the crash related to the first log file still persists, always the exact same log so no extra info from that.

 

I took a look at the windows event log to see if there's any info on what's happening, found an error relating to ucrtbase.dll and tried running the system file check in case the dll was corrupted but the sfc did not find anything so i have no idea what's causing it to give out an error or if that is even close to being the culprit... i'm out of ideas.

[traison]

Generally when you see crashes occuring in a system file like ucrtbase, kernel32 or vcruntime what you're actually seeing is something else passing an exported function malformed input. So when people reinstall the vc++ runtime as a result of a crash with a faulting offset in vcruntime its generally the wrong thing to do. While not impossible, it's quite unlikely that a 5 year old game would be finding actual bugs in large frameworks like this. The far more likely answer to these issues is to look down the callstack to what came before.

 

In your case the thing that came before is SkyrimSE.exe. Without the stack we have no peek into the parameters getting passed around, so it would be essential to get a debugger or at least a disassembler in there. From my point of view there's a couple of issues with this:

1. I'd have to obtain the 640 exe from somwhere. Unaltered.
2. I can't use a debugger on it without a large amount of work, since my game version is different. If this crash is related to mods, which it likely is, I couldn't trigger the crash on my end anyways.
3. Throwing the exe as-is into a disassembler could take 6-12 hours before any actual work can be started. I suspect this is because the exe is packed, confusing the automated analyzer. Also since it's packed, there's a chance the offsets mentioned in the callstack may not even be visible.

Options?

1. Try Trainwreck.
2. I remote control your computer and do the debugging locally. I'd have to bring my tools obviously. You'd have to consider the security implications. Generally this is not the popular option. 100% of my clients have survived so far, but all you really have is my word and my reputation on LL.

 

[astrum_xeno]

Makes sense, i wasn't really expecting it to actually be a corrupted system file but it was worth a shot to mention it just in case.

 

So for the Options you mentioned, i installed Trainwreck, here's what it spit out: https://pastebin.com/0WUFg0ni

 

I believe your good intentions, but letting someone i don't personally know remote control my computer is a big no-no because said security risks, so if this trainwreck log doesn't have anything useful i guess we're out of options and i have to start disassembling my load order one by one to find the culprit... what a pain.

 

in any case thank you again for your assistance and patience!

[traison]

Disable po3_SpellPerkItemDistributor.dll and test again.

[astrum_xeno]

Tried it, still crashes. https://pastebin.com/AWK7Mwwf

[traison]

As expected but it was worth a shot.

 

Quite annoying that the answer is most likely in RSP+F0 but we can't see it. A major shortcoming of Trainwreck, but it is what it is.

[astrum_xeno]

I assume that RSP+F0 is related to the missing stacks in the first crash logger log, right? I still don't understand why this particular crash only creates these incomplete logs...

 

anyway, i guess i'll start disassembling my modlist then. If anything new comes up or i find a solution, i'll post it here.

[astrum_xeno]

Well... i did... something i guess and produced another crashlogger log, no idea if this is related to the first one or if it's another issue entirely which wouldn't surprise me at all at this point lol. anyway another crash another log: https://pastebin.com/1mQY2W67

 

i'm tired. i'll start my disassembly project another day

[traison]

That's a new issue.

 

Reference id 0xFF001335 has a form id of 0x00000000. Looks like an actor with no base form which as far as I know is not valid. You most likely got a mod with a dll doing something it really shouldn't be doing.