diff options
author | 2022-03-25 12:14:28 +0000 | |
---|---|---|
committer | 2022-03-25 18:09:48 +0000 | |
commit | 02af7366063fe6ae9ac20e0521de2ca7b33e50ef (patch) | |
tree | 7d9ad4833a6aa8cc5f5318564f0229691a4ebb6c /compiler/driver/compiler_options.h | |
parent | 1dd1f479a0f8539e0f6cf1e30122dc32c61a5933 (diff) |
Fix a couple of places to check for instrumented entry points
There are a couple of places where we missed the check for instrumented
entry points:
- In resoultion stubs, when we get the code to execute we need to check
if we need an instrumentation entry point. Resolution stubs could
suspend so it is possible that we instrumented after resolution stub
started executing.
- When undeoptimizing a method, though the method itself is
undeoptimized we may still need instrumentation entry points.
Instrumentation entry points are also required to deoptimize the
caller.
Bug: 222479430
Test: art/test.py
Change-Id: Id520b5f6ff47fddee67e2bf8d3396d6d9db7fff6
Diffstat (limited to 'compiler/driver/compiler_options.h')
0 files changed, 0 insertions, 0 deletions