[ocaml5-issue] Segfault in Lin Effect test in framepointer mode

[jmid]

In the first CI run for #415 we triggered a segfault in the 5.1 framepointer mode:
https://github.com/ocaml-multicore/multicoretests/actions/runs/6956703096/job/18928093937#logs

random seed: 285592073
generated error  fail  pass / total     time test name

[ ]     0     0     0     0 / 20000     0.0s Lin ref int test with Effect
[ ]     0     0     0     0 / 20000     0.0s Lin ref int test with Effect (generating)
[✓] 20000     0     0 20000 / 20000     0.3s Lin ref int test with Effect

[ ]     0     0     0     0 / 20000     0.0s Lin ref int64 test with Effect
[✓] 20000     0     0 20000 / 20000     0.3s Lin ref int64 test with Effect

[ ]     0     0     0     0 / 20000     0.0s Lin CList int test with Effect
[✓] 20000     0     0 20000 / 20000     0.4s Lin CList int test with Effect

[ ]     0     0     0     0 / 20000     0.0s Lin CList int64 test with Effect
[✓] 20000     0     0 20000 / 20000     0.5s Lin CList int64 test with Effect

[ ]     0     0     0     0 / 20000     0.0s negative Lin ref int test with Effect
[✓]     1     0     1     0 / 20000     0.0s negative Lin ref int test with Effect

[ ]     0     0     0     0 / 20000     0.0s negative Lin ref int64 test with Effect
[✓]     1     0     1     0 / 20000     0.0s negative Lin ref int64 test with Effect

[ ]     0     0     0     0 / 20000     0.0s negative Lin CList int test with Effect
[✓]     1     0     1     0 / 20000     0.0s negative Lin CList int test with Effect

File "src/neg_tests/dune", line 92, characters 7-23:
92 |  (name lin_tests_effect)
            ^^^^^^^^^^^^^^^^
(cd _build/default/src/neg_tests && ./lin_tests_effect.exe --verbose)
Command got signal SEGV.
[ ]     0     0     0     0 / 20000     0.0s negative Lin CList int64 test with Effect

Note: I pushed an update to add a CHANGES entry to the PR underway, which triggers a cancellation of ongoing and queued jobs. I'm unsure whether that caused it (is it implemented as a signal?) or whether this is reproducible locally... 🤔

[jmid]

I just recalled that ocaml/ocaml#12535 was not included in the September 5.1.0 release.
I also confirmed locally: it is indeed reproducible on 5.1.0 but not on trunk.

Duplicate of #394