Crash with unmanged DLL did not create a StackTrace in Eventlog

[jogibear9988]

Description

We have an error, were we access Pointers from an unmanged DLL wich we freeed before.
If we do this in a Net4.8 Application in Debug Mode, we get a Stacktrace in the Eventlog.

If we du this in a Net6 Application in Debug Mode we get only this:

    Application: MCC.Common.Services.NTServiceHost.exe
    CoreCLR Version: 6.0.822.36306
    .NET Version: 6.0.8
    Description: The process was terminated due to an internal error in the .NET Runtime at IP 00007FFE9F791FDD (00007FFE9F5C0000) with exit code c0000005.

If we create a MiniDump we could see the Stacktrace in the Dump

Reproduction Steps

Crash a Net6 Application in unmanged COde

Expected behavior

We should see a Stacktrace in Eventlog

Actual behavior

No Stacktrace

Regression?

No response

Known Workarounds

No response

Configuration

No response

Other information

see #67897

Tagging subscribers to this area: @tommcdon
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Description

We have an error, were we access Pointers from an unmanged DLL wich we freeed before.
If we do this in a Net4.8 Application in Debug Mode, we get a Stacktrace in the Eventlog.

If we du this in a Net6 Application in Debug Mode we get only this:

    Application: MCC.Common.Services.NTServiceHost.exe
    CoreCLR Version: 6.0.822.36306
    .NET Version: 6.0.8
    Description: The process was terminated due to an internal error in the .NET Runtime at IP 00007FFE9F791FDD (00007FFE9F5C0000) with exit code c0000005.

If we create a MiniDump we could see the Stacktrace in the Dump

Reproduction Steps

Crash a Net6 Application in unmanged COde

Expected behavior

We should see a Stacktrace in Eventlog

Actual behavior

No Stacktrace

Regression?

No response

Known Workarounds

No response

Configuration

No response

Other information

see #67897

Author:	jogibear9988
Assignees:	-
Labels:	area-Diagnostics-coreclr
Milestone:	-

[hoyosjs]

@noahfalk @janvorli @jkotas Do we always do EventLog reporting in core and do we do so for unmanaged resources? I don't know if our SEH logic tries to do anything for such cases or if that's a Windows capability

[jogibear9988]

I get an Eventlog Entry, but without any Stacktrace

[hoyosjs]

Is your app deployed as a self contained publish?

[jogibear9988]

no. I'll try if I could recreate a sample app

[hoyosjs]

How are you running the app?

[noahfalk]

This is probably relevant code for whether we get a callstack in the error report: https://github.com/dotnet/runtime/blob/main/src/coreclr/vm/eepolicy.cpp#L458-L464

I don't recall that we made an intentional change to change the error report text, but it is certainly possible that it is a byproduct of other changes .NET Core made to have a simpler and more consistent cross-platform exception handling story. .NET Framework has a complex mechanism that can marshal exceptions across native<->managed interop boundaries whereas .NET Core treats exceptions crossing the boundary as fatal. If the stacktrace on .NET Framework was created by allowing a managed exception to unwind through native code .NET Core probably isn't going to replicate that behavior.

@jogibear9988 - are you able to show us the error log message produced by .NET Framework and what stacktrace you see in the minidump on .NET Core?

[jogibear9988]

in minidump I get the stacktrace.
I will post the messages from eventlog when I've created the sample.

The Stacktrace is in manged code. And this one I also get in fullframework in the eventlog

[jogibear9988]

If I debug the minidump mixed, I get this:

But I only need the managed stacktrace, wich was in the eventlog in fullFramework

[tommcdon]

@jogibear9988 just a friendly reminder to share the event log output once you have it available.