GHCi segfault with GHC 9.4.5

Summary

I've been looking at a segfault that seems to be triggered by some interpreted code calling into object code. This started appearing after upgrading from GHC 9.2.5 to 9.4.5. I've managed to extract an independent reproducer that I've linked below.

Steps to reproduce

Checkout the reproducer from: https://gitlab.haskell.org/teo/ghci-segfault-9.4 Then

nix develop
cabal repl ghci-segfault-test
:main

Expected behavior

Not seg fault

Environment

GHC version used: 9.4.5

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information