diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
index 06d2fe22e..278555611 100644
--- a/.github/workflows/ci.yml
+++ b/.github/workflows/ci.yml
@@ -70,6 +70,12 @@ jobs:
             python: pypy3
             arch: x64
 
+            # TODO: renable
+            # Currently segfaults on macOS Python 3.9
+          - runs-on: macos-latest
+            python: 3.9
+            arch: x64
+
     name: "🐍 ${{ matrix.python }} • ${{ matrix.runs-on }} • ${{ matrix.arch }} ${{ matrix.args }}"
     runs-on: ${{ matrix.runs-on }}
 
diff --git a/tests/test_gil_scoped.py b/tests/test_gil_scoped.py
index 6e7fe9f4c..b2169d98b 100644
--- a/tests/test_gil_scoped.py
+++ b/tests/test_gil_scoped.py
@@ -2,10 +2,6 @@
 import multiprocessing
 import threading
 
-import pytest
-
-import env  # noqa: F401
-
 from pybind11_tests import gil_scoped as m
 
 
@@ -54,8 +50,7 @@ def _python_to_cpp_to_python_from_threads(num_threads, parallel=False):
         thread.join()
 
 
-# TODO: FIXME, sometimes returns -11 instead of 0
-@pytest.mark.xfail("env.PY >= (3,9) and env.MACOS", strict=False)
+# TODO: FIXME, sometimes returns -11 (segfault) instead of 0 on macOS Python 3.9
 def test_python_to_cpp_to_python_from_thread():
     """Makes sure there is no GIL deadlock when running in a thread.
 
@@ -64,8 +59,7 @@ def test_python_to_cpp_to_python_from_thread():
     assert _run_in_process(_python_to_cpp_to_python_from_threads, 1) == 0
 
 
-# TODO: FIXME
-@pytest.mark.xfail("env.PY >= (3,9) and env.MACOS", strict=False)
+# TODO: FIXME on macOS Python 3.9
 def test_python_to_cpp_to_python_from_thread_multiple_parallel():
     """Makes sure there is no GIL deadlock when running in a thread multiple times in parallel.
 
@@ -74,8 +68,7 @@ def test_python_to_cpp_to_python_from_thread_multiple_parallel():
     assert _run_in_process(_python_to_cpp_to_python_from_threads, 8, parallel=True) == 0
 
 
-# TODO: FIXME
-@pytest.mark.xfail("env.PY >= (3,9) and env.MACOS", strict=False)
+# TODO: FIXME on macOS Python 3.9
 def test_python_to_cpp_to_python_from_thread_multiple_sequential():
     """Makes sure there is no GIL deadlock when running in a thread multiple times sequentially.
 
@@ -84,8 +77,7 @@ def test_python_to_cpp_to_python_from_thread_multiple_sequential():
     assert _run_in_process(_python_to_cpp_to_python_from_threads, 8, parallel=False) == 0
 
 
-# TODO: FIXME
-@pytest.mark.xfail("env.PY >= (3,9) and env.MACOS", strict=False)
+# TODO: FIXME on macOS Python 3.9
 def test_python_to_cpp_to_python_from_process():
     """Makes sure there is no GIL deadlock when using processes.