You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: compiler/src/jdk.graal.compiler/src/jdk/graal/compiler/replacements/nodes/ResolvedMethodHandleCallTargetNode.java
+8-9Lines changed: 8 additions & 9 deletions
Original file line number
Diff line number
Diff line change
@@ -41,15 +41,14 @@
41
41
42
42
/**
43
43
* A call target that replaces itself in the graph when being lowered by restoring the original
44
-
* {@link MethodHandle} invocation target. Prior to
45
-
* https://bugs.openjdk.java.net/browse/JDK-8072008, this is required for when a
46
-
* {@link MethodHandle} call is resolved to a constant target but the target was not inlined. In
47
-
* that case, the original invocation must be restored with all of its original arguments. Why?
48
-
* HotSpot linkage for {@link MethodHandle} intrinsics (see
49
-
* {@code MethodHandles::generate_method_handle_dispatch}) expects certain implicit arguments to be
50
-
* on the stack such as the MemberName suffix argument for a call to one of the MethodHandle.linkTo*
51
-
* methods. An {@linkplain MethodHandleNode#tryResolveTargetInvoke resolved} {@link MethodHandle}
52
-
* invocation drops these arguments which means the interpreter won't find them.
44
+
* {@link MethodHandle} invocation target. This is required for when a {@link MethodHandle} call is
45
+
* resolved to a constant target but the target was not inlined. In that case, the original
46
+
* invocation must be restored with all of its original arguments. Why? HotSpot linkage for
47
+
* {@link MethodHandle} intrinsics (see {@code MethodHandles::generate_method_handle_dispatch})
48
+
* expects certain implicit arguments to be on the stack such as the MemberName suffix argument for
49
+
* a call to one of the MethodHandle.linkTo* methods. An
0 commit comments