"""
rendersurface_refcount_v2.py <dump.dmp>

Definitive empirical test of F2 hypothesis using the CORRECT EoR
RenderSurface addresses/offsets found in Phase 7:

  - RenderSurface vtable (full): 0x0079bfe8
  - RenderSurface vtable (sub):   0x0079bffc
  - sizeof(RenderSurface):        0x14c bytes
  - refcount offset:              +0x04
  - m_pSurfaceBits offset:        +0x128
  - sourceData.sourceBits offset: +0xa0

Method:
  1. Scan all committed writable memory for DWORDs == 0x0079bfe8 or
     0x0079bffc — those are vtable pointers at offset 0 of RenderSurface
     instances.
  2. For each match (the object's base address), read:
       - refcount at +0x04
       - m_pSurfaceBits at +0x128
       - sourceData.sourceBits at +0xa0
  3. Histogram refcounts.
  4. Cross-reference with the leaked 256-512 KB private regions: how
     many RenderSurface instances actually have a leaked-region
     pointer in their m_pSurfaceBits field?

Verdict:
  - If refcount mode is 1: surfaces are unreferenced (could still leak if
    nothing destroys them, but unlikely)
  - If refcount mode > 1: external holders exist (F2 confirmed)
  - If most surfaces have leaked m_pSurfaceBits AND refcount > 1: that's
    the smoking gun for F2 — the surface is alive (held) AND its buffer
    is in the leaked set
"""
import os, struct, sys
from collections import Counter, defaultdict
from minidump.minidumpfile import MinidumpFile


VTABLE_FULL = 0x0079bfe8        # base RenderSurface, sizeof 0x14c — not used at runtime
VTABLE_SUB  = 0x0079bffc        # base RS subobject vtable
VTABLE_D3D  = 0x007e68a4        # RenderSurfaceD3D — THE RUNTIME VTABLE, sizeof 0x164
OFF_REFCOUNT      = 0x04
OFF_SOURCEBITS    = 0xa0
OFF_SOURCE_FLAGS  = 0xa4
OFF_SURFACEBITS   = 0x128


def _ei(v):
    if v is None: return 0
    if hasattr(v, 'value'): return int(v.value)
    return int(v)


def main():
    md = MinidumpFile.parse(sys.argv[1])
    reader = md.get_reader().get_buffered_reader()

    # Find leaked 256-512 KB regions
    leaked_regions = set()
    for r in md.memory_info.infos:
        st = _ei(r.State); ty = _ei(r.Type); pr = _ei(r.Protect) & 0xff
        if st == 0x1000 and ty == 0x20000 and pr in (0x04, 0x40) \
                and 256*1024 <= r.RegionSize < 512*1024:
            leaked_regions.add(r.BaseAddress)
    print(f"leaked 256-512KB regions: {len(leaked_regions)}")

    # Scan all committed writable memory for vtable pointer values
    scan_regions = []
    for r in md.memory_info.infos:
        st = _ei(r.State); ty = _ei(r.Type); pr = _ei(r.Protect) & 0xff
        if st != 0x1000: continue
        if ty == 0x1000000: continue
        if pr in (0x04, 0x40):
            scan_regions.append((r.BaseAddress, r.RegionSize))
    print(f"scanning {len(scan_regions)} writable regions")

    target_vtables = {VTABLE_FULL, VTABLE_SUB, VTABLE_D3D}
    objects = []
    total = 0
    for base, size in scan_regions:
        try:
            reader.move(base)
            buf = reader.read(size)
        except Exception:
            continue
        if not buf: continue
        total += len(buf)
        end = (len(buf) // 4) * 4
        for off in range(0, end - 0x130, 4):  # leave room to read up to +0x130
            vtbl = struct.unpack_from("<I", buf, off)[0]
            if vtbl not in target_vtables:
                continue
            obj_addr = base + off
            try:
                refcount = struct.unpack_from("<I", buf, off + OFF_REFCOUNT)[0]
                surfacebits = struct.unpack_from("<I", buf, off + OFF_SURFACEBITS)[0]
                sourcebits  = struct.unpack_from("<I", buf, off + OFF_SOURCEBITS)[0]
                source_flag = struct.unpack_from("<I", buf, off + OFF_SOURCE_FLAGS)[0]
            except struct.error:
                continue
            # Also read D3D extra fields (0x14c-0x164) for the D3D variant
            try:
                d3d_fields = [struct.unpack_from("<I", buf, off + 0x14c + 4*i)[0]
                              for i in range(6)]
            except struct.error:
                d3d_fields = [0]*6
            objects.append({
                "addr": obj_addr,
                "vtbl": vtbl,
                "refcount": refcount,
                "surfacebits": surfacebits,
                "sourcebits": sourcebits,
                "source_flag": source_flag,
                "d3d_fields": d3d_fields,
            })
    print(f"scanned {total/(1024*1024):.1f} MB")
    print(f"RenderSurface-like objects found (vtbl match): {len(objects)}")

    # Histogram by vtable
    by_vt = defaultdict(list)
    for o in objects:
        by_vt[o["vtbl"]].append(o)

    for vt, objs in by_vt.items():
        print(f"\n=== vtable 0x{vt:08x} ({len(objs)} objects) ===")
        refcount_hist = Counter(o["refcount"] for o in objs)
        print("  refcount histogram (top 10):")
        for rc, cnt in refcount_hist.most_common(10):
            pct = 100.0 * cnt / len(objs)
            print(f"    refcount={rc:<6}  {cnt:>5} ({pct:.1f}%)")

        # How many have m_pSurfaceBits pointing into the leaked set?
        leaked_owners = [o for o in objs if o["surfacebits"] in
                         {region for region in leaked_regions} or
                         any(region <= o["surfacebits"] < region + 0x80000 for region in leaked_regions)]
        # The pointer may be region_base + 0x30 (heap header offset), so check
        # if surfacebits is just past any leaked region's base
        precise = [o for o in objs if o["surfacebits"] - 0x30 in leaked_regions
                                   or o["surfacebits"] in leaked_regions]
        print(f"  RenderSurfaces with m_pSurfaceBits in leaked range: {len(leaked_owners)}")
        print(f"  RenderSurfaces with m_pSurfaceBits == leaked_base or +0x30: {len(precise)}")

        if precise:
            rc_leaked = Counter(o["refcount"] for o in precise)
            print(f"  refcount distribution for LEAKED-buffer owners:")
            for rc, cnt in rc_leaked.most_common(8):
                pct = 100.0 * cnt / len(precise)
                print(f"    refcount={rc:<6}  {cnt:>5} ({pct:.1f}%)")

    # Check D3D extra fields for leaked-region pointers
    print(f"\n=== D3D extra fields (+0x14c..+0x160) leaked-region check ===")
    for i in range(6):
        offset = 0x14c + 4*i
        non_null = sum(1 for o in objects if o["d3d_fields"][i] != 0)
        in_leaked = sum(1 for o in objects
                        if o["d3d_fields"][i] - 0x30 in leaked_regions
                        or o["d3d_fields"][i] in leaked_regions)
        # Also check the field is just any address in heap of leaked region size
        near_leaked = sum(1 for o in objects
                          if any(rb <= o["d3d_fields"][i] < rb + 0x80000 for rb in leaked_regions))
        print(f"  +0x{offset:03x}: non-null={non_null}/{len(objects)}, in_leaked={in_leaked}, near_leaked_region={near_leaked}")

    # Total RenderSurfaces
    if objects:
        print(f"\n=== ALL RenderSurfaces combined ===")
        rc_all = Counter(o["refcount"] for o in objects)
        for rc, cnt in rc_all.most_common(10):
            pct = 100.0 * cnt / len(objects)
            print(f"  refcount={rc:<6}  {cnt:>5} ({pct:.1f}%)")
        # Pointer status
        with_bits = sum(1 for o in objects if o["surfacebits"] != 0)
        print(f"  RenderSurfaces with non-null m_pSurfaceBits: {with_bits} / {len(objects)}")
        with_source = sum(1 for o in objects if o["sourcebits"] != 0)
        print(f"  RenderSurfaces with non-null sourceData.sourceBits: {with_source} / {len(objects)}")


if __name__ == "__main__":
    main()