"""physobj_owner_scan.py <dump.dmp>

Identify owners of leaked CPhysicsObj instances (EoR vtable 0x007c78e0).

Method:
  1. Enumerate all CPhysicsObj instances by scanning RW memory for DWORD = 0x007c78e0
     at offset 0 of any heap-aligned 4-byte slot. Each hit's address is a CPhysicsObj*.
  2. For each instance, scan ALL committed RW memory for DWORDs equal to that
     CPhysicsObj's base address (these are owners holding strong refs).
  3. For each owner-pointer hit, walk BACKWARDS up to 0x400 bytes within the same
     region looking for a DWORD that points into image memory (the owner's vtable).
  4. Histogram (vtable, field_offset) — these are the leak holders.

Output: ranked owner vtables and ranked (vtable, field_offset) pairs.
"""
import struct, sys
from collections import Counter, defaultdict
from minidump.minidumpfile import MinidumpFile

CPHYSOBJ_VT = 0x007c78e0


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()

    mods = []
    for m in md.modules.modules:
        mods.append((m.baseaddress, m.size, m.name))

    def mod_of(addr):
        for b, s, n in mods:
            if b <= addr < b + s:
                return n.split("\\")[-1]
        return None

    image_ranges = []
    for r in md.memory_info.infos:
        st, ty = _ei(r.State), _ei(r.Type)
        if st == 0x1000 and ty == 0x1000000:
            image_ranges.append((r.BaseAddress, r.BaseAddress + r.RegionSize))
    image_ranges.sort()

    def is_image(addr):
        for lo, hi in image_ranges:
            if lo <= addr < hi:
                return True
            if addr < lo:
                return False
        return False

    # Gather all RW private/mapped regions and cache buffers
    scan_regions = []
    for r in md.memory_info.infos:
        st, ty, pr = _ei(r.State), _ei(r.Type), _ei(r.Protect) & 0xff
        if st != 0x1000:
            continue
        if ty == 0x1000000:
            continue
        if pr not in (0x04, 0x40):
            continue
        scan_regions.append((r.BaseAddress, r.RegionSize))
    total_bytes = sum(s for _, s in scan_regions)
    print(f"scanning {len(scan_regions)} regions ({total_bytes / (1024 * 1024):.1f} MB)")

    region_bufs = []  # (base, buf)
    physobj_addrs = []
    for base, size in scan_regions:
        try:
            reader.move(base)
            buf = reader.read(size)
        except Exception:
            continue
        if not buf:
            continue
        region_bufs.append((base, buf))
        end = (len(buf) // 4) * 4
        for off in range(0, end, 4):
            v = struct.unpack_from("<I", buf, off)[0]
            if v == CPHYSOBJ_VT:
                physobj_addrs.append(base + off)

    print(f"found {len(physobj_addrs)} CPhysicsObj instances")
    if not physobj_addrs:
        return

    physobj_set = set(physobj_addrs)

    # Now scan ALL regions for DWORDs whose value is in physobj_set
    # For each hit, walk back 0x400 bytes to find a vtable.
    LOOKBACK = 0x400
    vtable_hits = Counter()
    vt_off_hits = Counter()
    examples = defaultdict(list)
    field_offsets_per_vtable = defaultdict(Counter)
    no_vtable = 0
    raw_hits = 0
    self_skipped = 0

    for base, buf in region_bufs:
        end = (len(buf) // 4) * 4
        for off in range(0, end, 4):
            v = struct.unpack_from("<I", buf, off)[0]
            if v not in physobj_set:
                continue
            hit_va = base + off
            if hit_va == v:
                # vtable slot of the physobj itself; ignore
                self_skipped += 1
                continue
            raw_hits += 1
            start = max(0, off - LOOKBACK)
            found = False
            for back in range(off - 4, start - 4, -4):
                if back < 0:
                    break
                vv = struct.unpack_from("<I", buf, back)[0]
                if vv < 0x00400000 or vv > 0x10000000:
                    continue
                if is_image(vv):
                    field_off = off - back
                    vtable_hits[vv] += 1
                    vt_off_hits[(vv, field_off)] += 1
                    field_offsets_per_vtable[vv][field_off] += 1
                    if len(examples[(vv, field_off)]) < 3:
                        examples[(vv, field_off)].append((hit_va, v))
                    found = True
                    break
            if not found:
                no_vtable += 1

    print(f"raw owner-ptr hits: {raw_hits}")
    print(f"self-vtable slot skipped: {self_skipped}")
    print(f"hits with no preceding vtable in lookback: {no_vtable}")
    print()
    print(f"=== Top owner vtables (regardless of offset) ===")
    for vt, cnt in vtable_hits.most_common(20):
        owner = mod_of(vt) or "?"
        top_offs = field_offsets_per_vtable[vt].most_common(4)
        offs_str = " ".join(f"+0x{o:x}={c}" for o, c in top_offs)
        print(f"  vt=0x{vt:08x}  hits={cnt:<6}  ({owner})  {offs_str}")

    print()
    print(f"=== Top (owner_vtable, field_offset) pairs ===")
    for (vt, off), cnt in vt_off_hits.most_common(30):
        owner = mod_of(vt) or "?"
        ex = examples[(vt, off)][0]
        print(f"  vt=0x{vt:08x}  +0x{off:03x}  hits={cnt:<6}  ({owner})  e.g. owner@0x{(ex[0] - off):08x} ptr@0x{ex[0]:08x} -> physobj@0x{ex[1]:08x}")


if __name__ == "__main__":
    main()