Skip to content

gh-159320: Enable stack size detection on Android #159321

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
Yhg1s merged 3 commits into from
Oct 14, 2026
Merged

gh-159320: Enable stack size detection on Android #159321

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
0c51afe
Extend the use of pthread_getattr_np in stack size detection from any…
Yhg1s Oct 13, 2026
f114aa6
Update tests for android's new recursion limit.
Yhg1s Oct 13, 2026
d3516fe
Apply batched suggestions from code review
Yhg1s Oct 14, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 2 additions & 4 deletions Lib/test/test_ast/test_ast.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1107,13 +1107,11 @@ def next(self):
enum._test_simple_enum(_Precedence, _ast_unparse._Precedence)

@support.cpython_only
@support.run_with_limited_c_stack(
100_000 if sys.platform == "android" else 500_000)
@support.run_with_limited_c_stack(500_000)
@skip_wasi_stack_overflow()
@skip_emscripten_stack_overflow()
def test_ast_recursion_limit(self):
# Android test devices have less memory.
crash_depth = 100_000 if sys.platform == "android" else 500_000
crash_depth = 500_000
success_depth = 200
if _testinternalcapi is not None:
remaining = _testinternalcapi.get_c_recursion_remaining()
Expand Down
9 changes: 3 additions & 6 deletions Lib/test/test_compile.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -773,14 +773,12 @@ def test_yet_more_evil_still_undecodable(self):

@support.cpython_only
@unittest.skipIf(support.is_wasi, "exhausts limited stack on WASI")
@support.run_with_limited_c_stack(
100_000 if sys.platform == "android" else 500_000)
@support.run_with_limited_c_stack(500_000)
@support.skip_emscripten_stack_overflow()
def test_compiler_recursion_limit(self):
# Compiler frames are small
limit = 100
# Android test devices have less memory.
crash_depth = limit * (1000 if sys.platform == "android" else 5000)
crash_depth = limit * 5000
success_depth = limit

def check_limit(prefix, repeated, mode="single"):
Expand Down Expand Up @@ -1092,8 +1090,7 @@ def test_path_like_objects(self):
# complex statements.
@support.requires_resource('cpu')
def test_stack_overflow(self):
# Android test devices have less memory.
size = 100_000 if sys.platform == "android" else 200_000
size = 200_000
compile("if a: b\n" * size, "<dummy>", "exec")

# Multiple users rely on the fact that CPython does not generate
Expand Down
2 changes: 2 additions & 0 deletions Misc/NEWS.d/next/Core_and_Builtins/2026-10-13-16-38-00.gh-issue-159320._Yhg1s.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,2 @@
Add dynamic detection of stack sizes on Android, allowing safer deep
recrusion.
Comment thread
Yhg1s marked this conversation as resolved.
Resolved
11 changes: 7 additions & 4 deletions Python/ceval.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -142,11 +142,14 @@ hardware_stack_limits(uintptr_t *base, uintptr_t *top, uintptr_t sp)
*top = (uintptr_t)stack_addr;
*base = ((uintptr_t)stack_addr) - stack_size;
#else
/// XXX musl supports HAVE_PTHRED_GETATTR_NP, but the resulting stack size
/// (on alpine at least) is much smaller than expected and imposes undue limits
/// compared to the old stack size estimation. (We assume musl is not glibc.)
// musl supports HAVE_PTHRED_GETATTR_NP, but the resulting stack size
Comment thread
Yhg1s marked this conversation as resolved.
Resolved
// (on alpine at least) is much smaller than expected and imposes undue
// limits compared to the old stack size estimation. musl doesn't give
// us a way to detect it, so we have to assume any non-glibc,
// non-Android Linux is musl and mustn't rely on pthread_getattr_np.
# if defined(HAVE_PTHREAD_GETATTR_NP) && !defined(_AIX) && \
!defined(__NetBSD__) && (defined(__GLIBC__) || !defined(__linux__))
!defined(__NetBSD__) && \
(defined(__GLIBC__) || defined(__ANDROID__) || !defined(__linux__))
size_t stack_size, guard_size;
void *stack_addr;
pthread_attr_t attr;
Expand Down
Loading