Re: very slow jni code callbacks

[Andrew Haley <aph@redhat.com>]

Bryce McKinlay wrote:
> On Mon, Sep 28, 2009 at 1:03 PM, Andrew Haley <aph@redhat.com> wrote:
> 
>>> When I use oprofile I see at lot of calls to execute_cfa_program
>>> and _Unwind_IteratePhdrCallback.
>> I think this is a bug.  I've had a look at the code, and it's in
>> jni.cc: _Jv_JNI_CallAnyVoidMethodV
>>
>> This calls _Jv_GetTypesFromSignature, which in turn iterates though
>> all the class loaders looking up the types of all the arguments.
>> But as far as I can see this is a pointless waste of time: the only
>> use of the type of each argument is to determine whether it's a
>> primitive type or not, and we could certainly have done that when
>> we created the jmethodID.  There is a field in the method struct called
>> ffi_arg_types, but I think it isn't used when calling C to Java via JNI.
> 
> I think the underlying issue is that jMethodID is typedef'ed to
> libgcj's internal _Jv_Method. It really needs to use something more
> like the interpreter's _Jv_ResolvedMethod or _Jv_JNIMethod so that the
> arg types and return type (and, in fact, the ffi_cif itself) can be
> resolved once and reused.

Well, sure, but it seems to me like this is obvious and almost trivial,
so I don't understand what the problem is.  I'm guessing that it's simply
that no-one ever got around to doing it.

Andrew.