[2.27 COMMITTED] pthread_cond_broadcast: Fix waiters-after-spinning case [BZ #23538]

[Florian Weimer]

From: Martin Kuchta <martin.kuchta@netapp.com>

(cherry picked from commit 99ea93ca31795469d2a1f1570f17a5c39c2eb7e2)

2018-08-27 Martin Kuchta  <martin.kuchta@netapp.com>
	   Torvald Riegel  <triegel@redhat.com>

	[BZ #23538]
	* nptl/pthread_cond_common.c (__condvar_quiesce_and_switch_g1):
	Update r to include the set wake-request flag if waiters are
	remaining after spinning.

diff --git a/NEWS b/NEWS
index bc150b63f9..a2e31b45cc 100644
--- a/NEWS
+++ b/NEWS
@@ -89,6 +89,7 @@ The following bugs are resolved with this release:
   [23363] stdio-common/tst-printf.c has non-free license
   [23456] Wrong index_cpu_LZCNT
   [23459] COMMON_CPUID_INDEX_80000001 isn't populated for Intel processors
+  [23538] pthread_cond_broadcast: Fix waiters-after-spinning case
 
 
 Version 2.27
diff --git a/nptl/pthread_cond_common.c b/nptl/pthread_cond_common.c
index 8e425eb01e..479e54febb 100644
--- a/nptl/pthread_cond_common.c
+++ b/nptl/pthread_cond_common.c
@@ -405,8 +405,12 @@ __condvar_quiesce_and_switch_g1 (pthread_cond_t *cond, uint64_t wseq,
 	{
 	  /* There is still a waiter after spinning.  Set the wake-request
 	     flag and block.  Relaxed MO is fine because this is just about
-	     this futex word.  */
-	  r = atomic_fetch_or_relaxed (cond->__data.__g_refs + g1, 1);
+	     this futex word.
+
+	     Update r to include the set wake-request flag so that the upcoming
+	     futex_wait only blocks if the flag is still set (otherwise, we'd
+	     violate the basic client-side futex protocol).  */
+	  r = atomic_fetch_or_relaxed (cond->__data.__g_refs + g1, 1) | 1;
 
 	  if ((r >> 1) > 0)
 	    futex_wait_simple (cond->__data.__g_refs + g1, r, private);