Introduce a flag to check if JITed code has instrumentation support

Introduce a new flag to identify if JITed code was compiled with
instrumentation support. We used to check if the runtime is java
debuggable to check for instrumentation support of JITed code. We only
set the java debuggable at runtime init and never changed it after. So
this check was sufficient since we always JIT code with instrumentation
support in debuggable runtimes.

We want to be able to change the runtime to debuggable after the runtime
has started. As a first step, introduce a new flag to explicitly check
if JITed code was compiled with instrumentation support. Use this flag
to check if code needs entry / exit stubs and to check if code is async
deoptimizeable.

Bug: 222479430
Test: art/test.py
Change-Id: Ibcaeab869aa8ce153920a801dcc60988411c775b

1 files changed, 6 insertions, 1 deletions

diff --git a/compiler/exception_test.cc b/compiler/exception_test.cc
index 4471b93f17..a49c6c630e 100644
--- a/compiler/exception_test.cc
+++ b/compiler/exception_test.cc
@@ -78,7 +78,12 @@ class ExceptionTest : public CommonRuntimeTest {
     ArenaStack arena_stack(&pool);
     ScopedArenaAllocator allocator(&arena_stack);
     StackMapStream stack_maps(&allocator, kRuntimeISA);
-    stack_maps.BeginMethod(4 * sizeof(void*), 0u, 0u, 0u);
+    stack_maps.BeginMethod(/* frame_size_in_bytes= */ 4 * sizeof(void*),
+                           /* core_spill_mask= */ 0u,
+                           /* fp_spill_mask= */ 0u,
+                           /* num_dex_registers= */ 0u,
+                           /* baseline= */ false,
+                           /* debuggable= */ false);
     stack_maps.BeginStackMapEntry(kDexPc, native_pc_offset);
     stack_maps.EndStackMapEntry();
     stack_maps.EndMethod(code_size);