Fix a memory ordering bug when updating the inline cache for predicted chaining.

Also collect the statistics about fast/queued/dropped IC update. Example
output:

D/dalvikvm(   86): JIT: ICPatch: 603 fast, 6 queued; 0 dropped

Bug: 2642019

Change-Id: I8cffdc9e422189bfd94bf845bcb2cff516f37077
diff --git a/vm/compiler/Compiler.c b/vm/compiler/Compiler.c
index 6cb1c77..6a58d8b 100644
--- a/vm/compiler/Compiler.c
+++ b/vm/compiler/Compiler.c
@@ -340,6 +340,7 @@
 #if defined(WITH_JIT_TUNING)
     /* Track method-level compilation statistics */
     gDvmJit.methodStatsTable =  dvmHashTableCreate(32, NULL);
+    gDvm.verboseShutdown = true;
 #endif
 
     dvmUnlockMutex(&gDvmJit.compilerLock);