Use SharedMutex instead of MicroLock in AtomicObserver

Summary:
The initialization of `folly::MicroLock` *looks* correct. MSAN instrumentation probably needs to be disabled in its implementation.

Regardless, the main point of using `folly::MicroLock` was its low memory usage (1 byte). I'm now changing it to `folly::SharedMutex` (4 bytes) which will be the same size due to alignment. Although `folly::SharedMutex` is not ideal for write-only locks, we don't care too much about write performance here.

Differential Revision: D25487925

fbshipit-source-id: 4c30650cf22dafbdaa9fea0c2252d29caf477d54

folly/experimental/observer/Observer-inl.h

@@ -110,9 +110,7 @@ Observer<T> makeStaticObserver(std::shared_ptr<T> value) {
 
 template <typename T>
 AtomicObserver<T>::AtomicObserver(Observer<T> observer)
-    : observer_(std::move(observer)) {
-  refreshLock_.init();
-}
+    : observer_(std::move(observer)) {}
 
 template <typename T>
 AtomicObserver<T>::AtomicObserver(const AtomicObserver<T>& other)
@@ -145,7 +143,7 @@ template <typename T>
 T AtomicObserver<T>::get() const {
   auto version = cachedVersion_.load(std::memory_order_acquire);
   if (UNLIKELY(observer_.needRefresh(version))) {
-    std::lock_guard<folly::MicroLock> guard{refreshLock_};
+    folly::SharedMutex::WriteHolder guard{refreshLock_};
     version = cachedVersion_.load(std::memory_order_acquire);
     if (LIKELY(observer_.needRefresh(version))) {
       auto snapshot = *observer_;

folly/experimental/observer/Observer.h

@@ -16,7 +16,7 @@
 
 #pragma once
 
-#include <folly/MicroLock.h>
+#include <folly/SharedMutex.h>
 #include <folly/ThreadLocal.h>
 #include <folly/experimental/ReadMostlySharedPtr.h>
 #include <folly/experimental/observer/Observer-pre.h>
@@ -271,7 +271,7 @@ class AtomicObserver {
  private:
   mutable std::atomic<T> cachedValue_{};
   mutable std::atomic<size_t> cachedVersion_{};
-  mutable folly::MicroLock refreshLock_;
+  mutable folly::SharedMutex refreshLock_;
   Observer<T> observer_;
 };