Let jfr2flame support AOT

[xiejf2020]

@xiejf2020 How do you think AOT support should look like and what problem/question this should solve?

I hope there's a [aot_comp] in addition to [Interpreter], [c1_comp] and [c2_comp].

Currently, all code (including Interpreter and C1) seems to be identified as C2 when AOT is on (maybe the whole logic used to identify the compilation is broken due to AOT). The following program was run on OpenJDK16.

java -XX:+UnlockExperimentalVMOptions -XX:+UseAOT -XX:AOTLibrary=$PWD/whatever.so -jar spring-petclinic-2.7.3.jar

Whatever whatever.so is here (maybe even if it's an empty lib with no compiled code) results in all code being recognized as C2:

And this is the flamegraph without AOT (C1 and C2 are correctly identified):

I know there are probably not many people who use AOT, so if you don't have time to fix it, maybe give me some tips about how to fix it :)

Originally posted by @xiejf2020 in #719 (comment)

[xiejf2020]

I think I may have found (part of) the problem.

    static NMethod* findNMethod(const void* pc) {
        if (contains(_code_heap[0], pc)) return findNMethod(_code_heap[0], pc);
        if (contains(_code_heap[1], pc)) return findNMethod(_code_heap[1], pc);
        if (contains(_code_heap[2], pc)) return findNMethod(_code_heap[2], pc);
        return NULL;
    }

It cannot find code in aot_code_heap because it does not traverse all codeheaps, instead it traverses only the first three.

So it failed at:

    if (event_type == 0) {
        // Async events
        int java_frames = getJavaTraceAsync(ucontext, frames + num_frames, _max_stack_depth, &java_ctx);
        if (java_frames > 0 && java_ctx.pc != NULL && VMStructs::hasMethodStructs()) {
            NMethod* nmethod = CodeHeap::findNMethod(java_ctx.pc);  // <- here
            if (nmethod != NULL) {
                fillFrameTypes(frames + num_frames, java_frames, nmethod);
            }
        }
        num_frames += java_frames;
    }

But that cannot explain why the interpreter and C1 are also recognized as C2.

[xiejf2020]

I see. The profiler won't init the _code_heap if code_heap_count > 3:

    if (_code_heap_addr != NULL && _code_heap_low_addr != NULL && _code_heap_high_addr != NULL) {
        char* code_heaps = *_code_heap_addr;
        unsigned int code_heap_count = *(unsigned int*)code_heaps;
        if (code_heap_count <= 3 && _array_data_offset >= 0) {
            char* code_heap_array = *(char**)(code_heaps + _array_data_offset);
            memcpy(_code_heap, code_heap_array, code_heap_count * sizeof(_code_heap[0]));
        }
        _code_heap_low = *_code_heap_low_addr;
        _code_heap_high = *_code_heap_high_addr;
    }