healthd: notify listeners using local copy of list, drop lock

Binder currently may service an incoming oneway transaction whenever
an outbound oneway call is made (if there is already a pending
incoming oneway call waiting).  The unexpected nested method call
blocks forever on a recursive mutex acquire because healthd is single-
threaded.  The binder behavior is considered a bug and may change in
the future.  For now, work around this in healthd.

Make a local copy of the listeners list, then drop the lock and
perform the outbound calls on the local copy of the list.

Bug: 38201220
Test: Marlin with modified client calling scheduleUpdate() repeatedly
Change-Id: If35c2847556245921e2aff808ff747bb60356811

1 files changed, 13 insertions, 3 deletions

diff --git a/healthd/BatteryPropertiesRegistrar.cpp b/healthd/BatteryPropertiesRegistrar.cpp
index 523e1f136..e51a06d5e 100644
--- a/healthd/BatteryPropertiesRegistrar.cpp
+++ b/healthd/BatteryPropertiesRegistrar.cpp
@@ -36,9 +36,19 @@ void BatteryPropertiesRegistrar::publish(
 }
 
 void BatteryPropertiesRegistrar::notifyListeners(const struct BatteryProperties& props) {
-    Mutex::Autolock _l(mRegistrationLock);
-    for (size_t i = 0; i < mListeners.size(); i++) {
-        mListeners[i]->batteryPropertiesChanged(props);
+    Vector<sp<IBatteryPropertiesListener> > listenersCopy;
+
+    // Binder currently may service an incoming oneway transaction whenever an
+    // outbound oneway call is made (if there is already a pending incoming
+    // oneway call waiting).  This is considered a bug and may change in the
+    // future.  For now, avoid recursive mutex lock while making outbound
+    // calls by making a local copy of the current list of listeners.
+    {
+        Mutex::Autolock _l(mRegistrationLock);
+        listenersCopy = mListeners;
+    }
+    for (size_t i = 0; i < listenersCopy.size(); i++) {
+        listenersCopy[i]->batteryPropertiesChanged(props);
     }
 }