Repository navigation
Conversation
reporting the stack limit of the main thread incorrectly. On platforms where pthread_getattr_np underreports the main thread's stack size, use RLIMIT_STACK for the main thread instead.
Yhg1s
requested review from
AA-Turner,
Eclips4,
corona10,
emmatyping,
erlend-aasland,
itamaro,
markshannon and
tomasr8
as code owners
October 13, 2026 21:31
Member
Author
|
This is a more general fix than #159321, and incidentally also lets us have proper stack overflow protections on musl. |
Member
Author
|
For the record, I think this should not be backported. |
encukou
reviewed
Oct 16, 2026
…doesn't depend on explicitly setting the stack size. It's correct for new threads regardless of that call.
Member
|
Can you please make the new "HAVE" macro private? Replace the macro in configure.ac: - [AC_DEFINE([HAVE_UNRELIABLE_PTHREAD_GETATTR_NP], [1],
+ [AC_DEFINE([_Py_HAVE_UNRELIABLE_PTHREAD_GETATTR_NP], [1],Something like that: Detailsdiff --git a/Python/ceval.c b/Python/ceval.c
index 3ed070a5846..e0623293670 100644
--- a/Python/ceval.c
+++ b/Python/ceval.c
@@ -65,7 +65,7 @@ _Py_ReachedRecursionLimitWithMargin(PyThreadState *tstate, int margin_count)
# define Py_C_STACK_SIZE 4000000
#endif
-#ifdef HAVE_UNRELIABLE_PTHREAD_GETATTR_NP
+#ifdef _Py_HAVE_UNRELIABLE_PTHREAD_GETATTR_NP
#ifdef HAVE_SYS_RESOURCE_H
# include <sys/resource.h>
#endif
@@ -165,7 +165,7 @@ hardware_stack_limits(uintptr_t *base, uintptr_t *top, uintptr_t sp)
err |= pthread_attr_destroy(&attr);
}
if (err == 0) {
-# ifdef HAVE_UNRELIABLE_PTHREAD_GETATTR_NP
+# ifdef _Py_HAVE_UNRELIABLE_PTHREAD_GETATTR_NP
/* pthread_getattr_np on the main thread of a musl process reports a
* stack size based on the current stack rather than RLIMIT_STACK.
* New threads are reported correctly even on musl, so this fallback
diff --git a/Tools/build/check_capi_macros_ignored.txt b/Tools/build/check_capi_macros_ignored.txt
index cfa37c2b137..238c29b7039 100644
--- a/Tools/build/check_capi_macros_ignored.txt
+++ b/Tools/build/check_capi_macros_ignored.txt
@@ -627,7 +627,6 @@ HAVE_UNAME
HAVE_UNISTD_H
HAVE_UNLINKAT
HAVE_UNLOCKPT
-HAVE_UNRELIABLE_PTHREAD_GETATTR_NP
HAVE_UNSHARE
HAVE_USABLE_WCHAR_T
HAVE_UTIL_H
diff --git a/configure b/configure
index 221907add48..744ee2beef4 100755
--- a/configure
+++ b/configure
@@ -21875,7 +21875,7 @@ printf "%s\n" "$ac_cv_unreliable_pthread_getattr_np" >&6; }
if test "x$ac_cv_unreliable_pthread_getattr_np" = xyes
then :
-printf "%s\n" "#define HAVE_UNRELIABLE_PTHREAD_GETATTR_NP 1" >>confdefs.h
+printf "%s\n" "#define _Py_HAVE_UNRELIABLE_PTHREAD_GETATTR_NP 1" >>confdefs.h
fi
fi
diff --git a/configure.ac b/configure.ac
index e523314b492..42844ab64ca 100644
--- a/configure.ac
+++ b/configure.ac
@@ -5620,7 +5620,7 @@ int main(void) {
[ac_cv_unreliable_pthread_getattr_np=yes],
[ac_cv_unreliable_pthread_getattr_np=no])])
AS_VAR_IF([ac_cv_unreliable_pthread_getattr_np], [yes],
- [AC_DEFINE([HAVE_UNRELIABLE_PTHREAD_GETATTR_NP], [1],
+ [AC_DEFINE([_Py_HAVE_UNRELIABLE_PTHREAD_GETATTR_NP], [1],
[Define if pthread_getattr_np under-reports the main thread's stack size (musl behaviour).])])])
# os.statx uses Linux's statx function. AIX also has a function named statx,
diff --git a/pyconfig.h.in b/pyconfig.h.in
index 4193d536f3b..1fdeb7d9f2f 100644
--- a/pyconfig.h.in
+++ b/pyconfig.h.in
@@ -1706,10 +1706,6 @@
/* Define to 1 if you have the 'unlockpt' function. */
#undef HAVE_UNLOCKPT
-/* Define if pthread_getattr_np under-reports the main thread's stack size
- (musl behaviour). */
-#undef HAVE_UNRELIABLE_PTHREAD_GETATTR_NP
-
/* Define to 1 if you have the 'unshare' function. */
#undef HAVE_UNSHARE
@@ -2227,6 +2223,10 @@
/* Define if you have the 'tanpi' function. */
#undef _Py_HAVE_TANPI
+/* Define if pthread_getattr_np under-reports the main thread's stack size
+ (musl behaviour). */
+#undef _Py_HAVE_UNRELIABLE_PTHREAD_GETATTR_NP
+
/* Thread stack size set by the linker (in bytes). */
#undef _Py_LINKER_THREAD_STACK_SIZE
|
Member
|
See also #159604 for FreeBSD. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adjust the use of pthread_getattr_np for stack size limit detection to avoid the problem on musl of it reporting the stack limit of the main thread incorrectly. On platforms where pthread_getattr_np underreports the main thread's stack size, use RLIMIT_STACK for the main thread instead. (Cross-compiles targeting musl will have to pass
ac_cv_unreliable_pthread_getattr_np=yesto configure.)