[libcxx-commits] [PATCH] D142422: [libc++] [FreeBSD] only use _umtx_op(2) on 64bit arches

[Konstantin Belousov via Phabricator via libcxx-commits]

kib updated this revision to Diff 495703.
kib added a comment.

Add comment explaining why umtx is only used on 64bit arches.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D142422/new/

https://reviews.llvm.org/D142422

Files:
  libcxx/src/atomic.cpp


Index: libcxx/src/atomic.cpp
===================================================================
--- libcxx/src/atomic.cpp
+++ libcxx/src/atomic.cpp
@@ -77,20 +77,25 @@
                  const_cast<__cxx_atomic_contention_t*>(__ptr), 0);
 }
 
-#elif defined(__FreeBSD__)
+#elif defined(__FreeBSD__) && __SIZEOF_LONG__ == 8
+/*
+ * Since __cxx_contention_t is int64_t even on 32bit FreeBSD
+ * platforms, we have to use umtx ops that work on the long type, and
+ * limit its use to architectures where long and int64_t are synonyms.
+ */
 
 static void __libcpp_platform_wait_on_address(__cxx_atomic_contention_t const volatile* __ptr,
                                               __cxx_contention_t __val)
 {
     _umtx_op(const_cast<__cxx_atomic_contention_t*>(__ptr),
-             UMTX_OP_WAIT_UINT_PRIVATE, __val, NULL, NULL);
+	     UMTX_OP_WAIT, __val, NULL, NULL);
 }
 
 static void __libcpp_platform_wake_by_address(__cxx_atomic_contention_t const volatile* __ptr,
                                               bool __notify_one)
 {
     _umtx_op(const_cast<__cxx_atomic_contention_t*>(__ptr),
-             UMTX_OP_WAKE_PRIVATE, __notify_one ? 1 : INT_MAX, NULL, NULL);
+	     UMTX_OP_WAKE, __notify_one ? 1 : INT_MAX, NULL, NULL);
 }
 
 #else // <- Add other operating systems here


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D142422.495703.patch
Type: text/x-patch
Size: 1295 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/libcxx-commits/attachments/20230208/03b98ac8/attachment.bin>