[BUG]: Control Flow Guard violation at komorebi.exe + 0x9B34C4 (FASTFAIL_GUARD_ICALL_CHECK_FAILURE)
Crash Report: Control Flow Guard violation at komorebi.exe + 0x9B34C4
Environment
- komorebi version: 0.1.41 (installed via scoop extras bucket)
- OS: Windows 10 22H2 (build 19045)
- CPU: 12 threads/processors
- Architecture: x64
Crash dump analysis
Two crash dumps were captured, both showing the same crash type at the same code offset:
| # | Dump file | Date | Offset in komorebi.exe |
|---|---|---|---|
| 1 | komorebi.exe.12408.dmp |
2025-07-24 | 0x9B34C4 |
| 2 | komorebi.exe.24496.dmp |
2025-07-27 (latest) | 0x9B34C4 |
Exception details
Exception code: 0xC0000409 (STATUS_STACK_BUFFER_OVERRUN)
FastFail subcode: 7 (FASTFAIL_GUARD_ICALL_CHECK_FAILURE)
Exception flags: 0x1 (non-continuable)
Crash RIP offset: 0x9B34C4 (from komorebi.exe base)
RDI (vtable/target) 0xBCE9C0 (from komorebi.exe base)
Number of threads: 40 (first crash was 42)What this means
This is a Control Flow Guard (CFG) protection fault. Windows detected that komorebi attempted to make an indirect call (through a function pointer, vtable dispatch, or trait object dispatch) to an address that was not registered as a valid call target.
In Rust context, this typically indicates one of:
- Trait object (
dyn) vtable corruption — the vtable pointer was overwritten to point to an invalid memory location - Use-after-free — a trait object was dropped/freed but a reference to its vtable was still used
- Unsafe code — a raw pointer or transmuted value was used as an indirect call target
- Stack/heap buffer overflow — adjacent memory corruption overwrote a function pointer
Consistency
Both crash dumps show the identical crash signature:
- Same exception code (
0xC0000409) - Same FastFail subcode (
7) - Same offset within komorebi.exe (
0x9B34C4)
This means the crash is deterministically reproducible, not random memory corruption. This is a definite bug, not a hardware issue.
Stack trace hints
The stack (raw scan of RSP) shows a call chain primarily within komorebi.exe with involvement of combase.dll and ntdll.dll, suggesting this may be triggered during Windows COM / window message processing.
How to reproduce
In my usage, komorebi crashed twice in the span of a few days (July 24 and July 27) during normal desktop usage — window management, switching workspaces, etc. I was unable to identify a specific trigger action; the crash seems to happen spontaneously under normal workload.
Crash dump files
Both are full crash dumps (~25 MB each).
⚠️ Note: These are full process memory dumps and may contain sensitive information (window titles, file paths, etc.). I will not upload them publicly to this issue. If you (the maintainer) need them for debugging, please let me know a private channel (email / Discord DM / etc.) to send them to.
Generated from crash dump analysis via minidump Python library on 2025-07-27.
Source: LGUG2Z/komorebi