Various debug ART GC performance improvements

Removed unnecessary read barriers in SizeOf and a few other places.
Disabled disallow read barrier check.

Before:
GC time: 15.817s
Real 0m26.113s
user  1m16.780s
sys 0m3.152s

After:
GC time: 9.212s
real  0m19.875s
user  1m9.916s
sys 0m1.916s

Bug: 35644369
Test: test-art-host

Change-Id: I79a65259deff2a478a96e02ae69b14730b6dcbe6
diff --git a/runtime/gc/heap.cc b/runtime/gc/heap.cc
index 357541f..a853b98 100644
--- a/runtime/gc/heap.cc
+++ b/runtime/gc/heap.cc
@@ -2337,9 +2337,7 @@
     size_t bin_size = object_addr - context->prev_;
     // Add the bin consisting of the end of the previous object to the start of the current object.
     collector->AddBin(bin_size, context->prev_);
-    // Turn off read barrier. ZygoteCompactingCollector doesn't use it (even in the CC build.)
-    context->prev_ = object_addr + RoundUp(obj->SizeOf<kDefaultVerifyFlags, kWithoutReadBarrier>(),
-                                           kObjectAlignment);
+    context->prev_ = object_addr + RoundUp(obj->SizeOf<kDefaultVerifyFlags>(), kObjectAlignment);
   }
 
   void AddBin(size_t size, uintptr_t position) {
@@ -2359,8 +2357,7 @@
 
   virtual mirror::Object* MarkNonForwardedObject(mirror::Object* obj)
       REQUIRES(Locks::heap_bitmap_lock_, Locks::mutator_lock_) {
-    // Turn off read barrier. ZygoteCompactingCollector doesn't use it (even in the CC build.)
-    size_t obj_size = obj->SizeOf<kDefaultVerifyFlags, kWithoutReadBarrier>();
+    size_t obj_size = obj->SizeOf<kDefaultVerifyFlags>();
     size_t alloc_size = RoundUp(obj_size, kObjectAlignment);
     mirror::Object* forward_address;
     // Find the smallest bin which we can move obj in.