#!/usr/bin/env python3
"""Find guest addresses holding one value in snapshot A and another in snapshot B.

The classic differential search, which is how you locate a field whose ADDRESS is
unknown but whose VALUE is known at two moments. Used here for the GamePart slot:
`GP_EXTRAS` = 5 on the extras screen, `GP_MISSION_SELECT` = 7 on the next one, so
the slot is a word that reads 5 then 7.

Both snapshots are sparse copies of `/dev/shm/xenia_memory_*`, so the ~4.6 GB of
holes cost nothing: walk A's allocated extents with SEEK_DATA and only look at B
where A already matched.

    diff_words.py <snapA> <valA> <snapB> <valB> [--near <va> <span>]
"""
import os
import struct
import sys

sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from gmem import off_to_vas  # noqa: E402


def main():
    if len(sys.argv) < 5:
        print(__doc__)
        return 1
    pa, va_, pb, vb_ = sys.argv[1], int(sys.argv[2], 0), sys.argv[3], int(sys.argv[4], 0)
    pat_a = struct.pack(">I", va_)
    fa, fb = open(pa, "rb"), open(pb, "rb")
    size = min(os.path.getsize(pa), os.path.getsize(pb))

    hits, scanned = [], 0
    off = 0
    while off < size:
        try:
            off = os.lseek(fa.fileno(), off, os.SEEK_DATA)
        except OSError:
            break
        if off >= size:
            break
        try:
            end = min(os.lseek(fa.fileno(), off, os.SEEK_HOLE), size)
        except OSError:
            end = size
        while off < end:
            n = min(1 << 22, end - off)
            fa.seek(off)
            buf = fa.read(n)
            scanned += len(buf)
            i = 0
            while True:
                i = buf.find(pat_a, i)
                if i < 0:
                    break
                p = off + i
                if p % 4 == 0:
                    fb.seek(p)
                    if fb.read(4) == struct.pack(">I", vb_):
                        hits.append(p)
                i += 1
            off += n

    print(f"scanned {scanned/1e6:.0f} MB of allocated guest memory")
    print(f"addresses reading {va_} in A and {vb_} in B: {len(hits)}")
    for h in hits[:40]:
        vas = off_to_vas(h)
        print(f"   file off {h:#012x}   guest VA {[hex(v) for v in vas][:2]}")
    if len(hits) > 40:
        print(f"   … and {len(hits)-40} more")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())