More specifically the entry frame performs the transition from C++
code to generated code, so every sequence of interpreted or compiled
frames always begins with an entry frame. It's built by the call_stub
and sets up all the invariants required by generated code.
tom
On Mar 12, 2009, at 4:04 AM, Gary Benson wrote:
> Colin(Du Li) wrote:
>> In hotspot, I usually see the terms "interpreted frame" and "entry
>> frame". What's the difference between them?
>
> Each interpreted method has a frame on the stack; these frames are
> interpreted frames. On entry, interpreted methods expect to find
> their arguments at the end of their caller's frame. For the very
> first call there won't be an interpreted frame preceding, so for
> the first call a thing called the call stub creates a special frame
> with the arguments at the end; this is the entry frame.
>
> Cheers,
> Gary
>
> --
>
http://gbenson.net/
