[libc-commits] [libc] [libc] Implement barriers for pthreads (PR #148948)

[Brooks Moses via libc-commits]

================
@@ -0,0 +1,83 @@
+//===-- Implementation of Barrier class ------------- ---------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+#include "src/__support/threads/barrier.h"
+#include "barrier.h"
+#include "hdr/errno_macros.h"
+#include "src/__support/threads/mutex.h"
+
+namespace LIBC_NAMESPACE_DECL {
+
+const int BARRIER_FIRST_EXITED = -1;
+
+int Barrier::init(Barrier *b,
+                  const pthread_barrierattr_t *attr __attribute__((unused)),
+                  unsigned count) {
+  LIBC_ASSERT(attr == nullptr); // TODO implement barrierattr
+  if (count == 0)
+    return EINVAL;
+
+  b->expected = count;
+  b->waiting = 0;
+  b->blocking = true;
+
+  int err;
+  err = CndVar::init(&b->entering);
+  if (err != 0)
+    return err;
+
+  err = CndVar::init(&b->exiting);
+  if (err != 0)
+    return err;
+
+  Mutex::init(&b->m, false, false, false, false);
----------------
brooksmoses wrote:

Why aren't we checking this for an error return?

(I see that currently `Mutex::init` always returns `MutexError::NONE`, but that could change in the future.  Alternately, if we don't expect it to change, then the `Mutex::init` function shouldn't have a return value.)

https://github.com/llvm/llvm-project/pull/148948