[llvm-branch-commits] [libcxx] 2cff4ff - Make ios_base::xalloc non-atomic with LIBCXX_ENABLE_THREADS=OFF. (#208356)

[Douglas Yung via llvm-branch-commits]

Author: Eli Friedman
Date: 2026-07-22T13:43:38Z
New Revision: 2cff4ff7a21c91b3493eb0fd8a7f072aee77fe7d

URL: https://github.com/llvm/llvm-project/commit/2cff4ff7a21c91b3493eb0fd8a7f072aee77fe7d
DIFF: https://github.com/llvm/llvm-project/commit/2cff4ff7a21c91b3493eb0fd8a7f072aee77fe7d.diff

LOG: Make ios_base::xalloc non-atomic with LIBCXX_ENABLE_THREADS=OFF. (#208356)

762b77a moved the definition of "xindex" out of the header, and in the
process dropped the _LIBCPP_HAS_THREADS check. Re-add the check to
maintain the status quo.

The discussion on https://github.com/llvm/llvm-project/pull/198994
indicates it's not clear whether LIBCXX_ENABLE_THREADS=OFF is actually
supposed to mean single-threaded. But it clearly does in practice:
atomic_support.h uses non-atomic ops when threads are disabled, and a
few other APIs have explicit non-atomic fallback paths.

My team ran into this trying to run libc++ tests for a RISC-V core
without the "a" extension.

(cherry picked from commit 3eb929be5e17d66900020bd1caa7d2510d4f9601)

Added: 
    

Modified: 
    libcxx/src/ios.cpp

Removed: 
    


################################################################################
diff  --git a/libcxx/src/ios.cpp b/libcxx/src/ios.cpp
index 2e049098740dc..db6ca50f7f6ca 100644
--- a/libcxx/src/ios.cpp
+++ b/libcxx/src/ios.cpp
@@ -122,7 +122,14 @@ static size_t __ios_new_cap(size_t __req_size, size_t __current_cap) { // Precon
 }
 
 int ios_base::xalloc() {
+#if _LIBCPP_HAS_THREADS
   constinit static atomic<int> xindex = 0;
+#else
+  // If we don't have atomics, fall back to single-threaded implementation.
+  // FIXME: Should "single-threaded" be a separate option from
+  // _LIBCPP_HAS_THREADS?
+  static int xindex = 0;
+#endif // _LIBCPP_HAS_THREADS
   return xindex++;
 }