Improve the allocation speed.
- Improves the Ritz MemAllocTest benchmark result by ~500 ms (or ~5%) on Nexus 4.
- Move the memset() call that zeroes the allocated memory out of the lock region.
- De-virtualize/Inline the allocation call chains into Heap::AllocObject().
- Turn Heap::measure_allocation_time_ into a static const variable.
- Surround the VerifyObject() call with kIsDebugBuild.
Bug: 9986565
Change-Id: Ib70b6d051a80ec329788b30256565561f031da2a
diff --git a/runtime/gc/space/large_object_space.h b/runtime/gc/space/large_object_space.h
index 8cd5088..31659db 100644
--- a/runtime/gc/space/large_object_space.h
+++ b/runtime/gc/space/large_object_space.h
@@ -83,7 +83,7 @@
// Return the storage space required by obj.
size_t AllocationSize(const mirror::Object* obj);
- mirror::Object* Alloc(Thread* self, size_t num_bytes);
+ mirror::Object* Alloc(Thread* self, size_t num_bytes, size_t* bytes_allocated);
size_t Free(Thread* self, mirror::Object* ptr);
void Walk(DlMallocSpace::WalkCallback, void* arg);
// TODO: disabling thread safety analysis as this may be called when we already hold lock_.
@@ -110,7 +110,7 @@
static FreeListSpace* Create(const std::string& name, byte* requested_begin, size_t capacity);
size_t AllocationSize(const mirror::Object* obj) EXCLUSIVE_LOCKS_REQUIRED(lock_);
- mirror::Object* Alloc(Thread* self, size_t num_bytes);
+ mirror::Object* Alloc(Thread* self, size_t num_bytes, size_t* bytes_allocated);
size_t Free(Thread* self, mirror::Object* obj);
bool Contains(const mirror::Object* obj) const;
void Walk(DlMallocSpace::WalkCallback callback, void* arg);