Re: [RFA] Suspend counting for JDWP

[Bryce McKinlay <mckinlay at redhat dot com>]

Keith Seitz wrote:
> > Are we likely to need to associate other JVMTI/JDWP debugging data 
> > with a thread? If so then it might be wise to put these in a 
> > _Jv_ThreadDebugInfo (or some such) struct which would be initialized 
> > only when JDWP is in use.
>
> JVMTI defines the following thread-related functions:
> All of these things basically mirror JDWP information, and all of it 
> is contained within the thread. JVMTI does not give access to suspend 
> counts the way JDWP does.

OK, it looks like a separate ThreadDebugInfo struct is not required.

> Of course, nothing stops us from implementing JVMTI right along with 
> JDWP, since the two do obviously overlap in many ways. But that was 
> well outside the scope of the project.

You could think of JVMTI as the low-level interface through which the 
JDWP implementation should communicate with the VM. If we were to 
implement JDWP without JVMTI, then we'd just end up duplicating much of 
JVMTI anyway.

Clearly, "jvmtiError SuspendThread(jvmtiEnv* env, jthread thread)" is 
more-or-less equivalent to "_Jv_DebugSuspendThread(_Jv_Thread_t)".

I'd argue that anything additional to what is provided by JVMTI, such as 
suspend counts, should be implemented in the JDWP code itself and 
doesn't need to be dealt with in the core runtime.

> So, the question is: do maintainers want to start implementing JVMTI 
> and switch over to use it now or wait and implement the whole thing? 
> I'm sure that these threading issues are not the only place this route 
> could have been chosen.

Its your call. Certainly if you think JVMTI is going to hold up JDWP 
work, I wouldn't want to see too much time spent on it. But, I think 
that implementing the parts needed by JDWP shouldn't be a huge 
undertaking, It should actually be helpful, by clearly defining the 
interfaces through which JDWP should communicate with the VM, rather 
than ending up with bits and pieces of JDWP support code scattered around.

I'm not suggesting should we be trying to implement the full JVMTI spec 
at this stage - just the bits that are directly used by the JDWP 
features we are going to support.

Bryce