Check for yields while waiting for lock in a loop (#13498)

Summary:
Acquiring a lock here can take a long time and cause a user mode scheduler to hold up, as it relies on explicit yielding. Hence, forcing a check here but ignoring any abort requests. Would rely on upstream to take action on aborts.

Pull Request resolved: #13498

Reviewed By: pdillinger

Differential Revision: D71987173

Pulled By: jainpr

fbshipit-source-id: 4aec40bdf0bc657e29f72c306c576b3117f97a25

Author: jainpr

utilities/transactions/lock/point/point_lock_manager.cc

@@ -318,6 +318,9 @@ Status PointLockManager::AcquireWithTimeout(
       } else {
         uint64_t now = env->NowMicros();
         if (static_cast<uint64_t>(cv_end_time) > now) {
+          // This may be invoked multiple times since we divide
+          // the time into smaller intervals.
+          (void)ROCKSDB_THREAD_YIELD_CHECK_ABORT();
           result = stripe->stripe_cv->WaitFor(stripe->stripe_mutex,
                                               cv_end_time - now);
         }