Files
apk-reverse/tests/unit/test_trace_addr2sym.py
T
root ab22b225ff Hardware pass: hardware watchpoints, ETE differential tracing, kernel-side event injection
Add the layer under the software tooling, in the repository's shape: what runs
on a host is measured, and every device-bound or kernel-bound piece is a gated,
labelled template.

- scripts/hw_watch.py: arm64 hardware watchpoints/execution breakpoints. WCR/
  BCR encoders with the 8-byte-window and alignment rules (unit-tested to the
  canonical 0x1FFD/0x3F05), a self-contained on-device ptrace helper that
  compiles clean as a static aarch64 ELF (observed: cross gcc 16.2), adb
  orchestration, and hit decoding to module+offset. The stealth ledger is
  stated three times on purpose: the ptrace route leaves TracerPid visible.
- scripts/ete_capture.py: CoreSight ETE capture with the five gates (kernel
  era, coresight sysfs, cs_etm PMU, perf-with-CORESIGHT, permission), the
  perf/OpenCSD build recipe, and the per-thread record wrapper whose
  --control one-instruction run must pass before any real capture.
- scripts/trace_diff.py: differential reduction of two execution traces into
  first-divergence / divergent regions / blocks-unique-to-each-side.
  Producer-agnostic: Stalker BLK logs (raw or run_probe-wrapped) and decoded
  cs_etm address streams alike, host-tested on synthetic flattened-dispatcher
  fixtures.
- scripts/kernel_instrument_scaffold.py: KPM/LKM watchpoint + register-rewrite
  templates plus an honest userspace carrier; every kernel file carries a
  NOT COMPILED AND NOT LOADED banner that verify enforces, and the
  register-rewrite action is fenced as experimental in the template itself.
- references/hardware-watchpoints.md, references/ete-diff-tracing.md,
  references/kernel-event-injection.md: the mechanism, the gates, the traps,
  and the composition recipes (watch the key buffer; diff two runs; flip the
  decision at the trap), with hand-offs instead of duplication.
- Registration: 3 symptom rows (SKILL.md + routing mirror), 3 reference rows,
  4 script rows, coverage/route-inventory/evidence rows, capability-matrix
  entries, known-limitations entry, eval case 4, README trees and coverage
  paragraphs in both languages, and three EXTENSION evidence records stating
  exactly what was measured on this host (encoders, compile, fixtures,
  generate/verify) and what was not (no watchpoint armed, no cs_etm stream,
  nothing kernel-side compiled or loaded anywhere).

Hardware pass 2: SPE/BRBE crash forensics, address symbolization, VM de-virtualization

Calibrated against the two public obfuscator implementations the user supplied
(dpkg123/ollvm-aosp-toolchain: Hikari-family and Pluto-family flattening, the
xvmp IR virtualizer at 63fec41f/9b7cf75e; fetched read-only into
tools/_work/refs, git-ignored). Source facts observed, runtime inferred.

- scripts/trace_addr2sym.py: runtime addresses -> symbol + static offset,
  zero-dependency ELF64 (PT_LOAD vaddr<->file-off conversion, .dynsym/.symtab,
  stripped section+off fallback). Host-verified against a real cross-compiled
  aarch64 .so with readelf as the independent cross-check, plus a hand-built
  minimal ELF covering vaddr!=offset; accepts hw_watch hits, trace mod+off
  keys, or raw address lists. Fixes on the way: --base int coercion, and
  nearest-below symbol semantics for PCs in inter-symbol gaps.
- scripts/vm_handler_slice.py: slice an interpreter trace into handler
  segments (dispatcher detection = histogram top + predecessor diversity; the
  first 4x-margin rule refused the calibrated fixture itself and was replaced),
  signature clustering, open-tail handling, and the guest-opcode trajectory via
  an optional handler->opcode map. Calibrated against xvmp's documented shape:
  vm_exec(code,size,vm_args,num) stubs, boxed arguments observable before
  interpretation, 11 opcodes + 19 arithmetic sub-ops, const-materialization
  dispatch as a distinct segment.
- references/hardware-sampling-and-crash.md: SPE (data flow without a known
  address; PC + data address per sample), BRBE (branch records when a crash
  stack cannot be trusted), the crash loop (tombstone -> branch records ->
  symbolize), and what each adds to the OLLVM/VMP routes. All
  documented-and-gated: no SPE sample or BRBE record was captured here.
- Flattening-recipe corrections from source: Hikari keeps a frame-resident
  switchVar written through an indirect pointer with scramble32-permuted cases;
  Pluto stores switchVar/switchXor rolling-XOR so memory only ever holds noise
  -- hardware-watchpoints.md's state-variable recipe now distinguishes the
  families, and ete-diff-tracing.md carries the family table plus the xvmp
  shape. AntiDebugging injects raw svc inline asm with randomized fields, so
  byte signatures cannot survive it; semantic site-location can.
- Registration: 3 symptom rows (untrustworthy crash stack; runtime addresses
  needing symbols; interpreter loop reading as noise), reference/script rows,
  coverage/evidence/capability-matrix entries, README trees in both languages,
  and EXTENSION-sampling-crash.md recording exactly what was measured (real .so
  symbolization, slicer fixtures, the slicer's self-found dispatcher-rule bug)
  and what was not (no SPE/BRBE capture, no hardened binary ever traced).

Fix all PR #1 review findings by severity; un-entrench the tests that pinned them

Major (device-breaking / wrong results):
- hw_watch: match the arm64 user_hwdebug_state slot ABI (address register
  first, control second -- the swap armed nothing correctly)
- hw_watch: every post-seize error path now clears slots and detaches;
  PTRACE_O_EXITKILL turned an analysis failure into a SIGKILL of the target
- hw_watch: forward genuine target SIGTRAPs (raise/tgkill/int3); only ptrace
  event stops are consumed -- the reference's signal-preservation promise
- hw_watch: run checks push/build/chmod and the helper's exit status; a dead
  capture is RESULT=blocked, never ok-no-hits
- hw_watch: decode offsets from the module base, not each mapping's start
- ete_capture: --control now requires the control record to DECODE
  (perf script event lines), the documented gate, not just a zero exit
- ete_capture: -o before the -- workload separator in --cmd mode; all
  values shell-quoted into su -c; stale remote perf.data removed first;
  perf exit status checked (timeout's 124 is the normal window end)
- kernel scaffold: --set-reg restricted to x0..x30; the rewrite is gated by
  its own breakpoint event so a watchpoint hit cannot run it; LKM pins the
  task (get/put_task_struct) across registration and calls the four-argument
  register_user_hw_breakpoint; --break rows initialize .addr/.len/.bp_type
  (there is no .pc member); --id validated against the module-id grammar
- trace_addr2sym: exact module matching (path or basename, never substring);
  load bias aligned against the PT_LOAD (bias = start - (foff - p_offset)
  - p_vaddr); PCs outside the ELF span are refused instead of landing on the
  last symbol; --json stdout is a pure document (RESULT= in text mode only)
- trace_diff: block keys need 0x or >=8 hex digits (PIDs/counts no longer
  parse); vm_handler_slice: dispatcher needs predecessor diversity, no
  frequency-margin fallback; both JSON outputs pure (check_commands' rule)

Minor/nitpick:
- capability-matrix + evidence-summary: trace reduction status ok->partial
  (no real-target producer ran it; the file's own definition of ok)
- hardware-sampling-and-crash.md and the script examples carry --elf
- kernel-event-injection.md no longer claims the callback's first statement
  is a filter (it prints unfiltered; the filter is a required change)
- Elf64 caches symbol addresses for per-row bisect
- docstring coverage of the touched surface: 16.67% -> 100%

Tests: every json.loads workaround stripping RESULT= is gone; new regression
tests cover the slot ABI, release-on-error, SIGTRAP forwarding, non-first
mapping offsets, -o placement and su -c quoting, the control decode gate,
x31/--id refusals, the guarded rewrite, outside-ELF PCs and bias math,
short-token refusal and the dispatcher margin rule. The minimal-ELF fixture
now packs a full 8-field Elf64_Phdr (its 7-field pack silently shifted
p_memsz onto p_align). 739 passed; flake8/mypy and all five gates clean.

Address the follow-up review on 0bd8dfe (6 findings)

- ete_capture: two-layer quoting for the control-record, decode and cleanup
  commands -- shlex.quote spliced between literal single quotes broke out of
  su -c for any spaced output name; --seconds must be positive (timeout 0
  disables the timer), refused before any adb check
- kernel scaffold: --set-reg now refuses sp/pc as well (the template has no
  rewrite path for them; a generate that silently drops the action is the
  misleading kind of success); both templates arm with a sampling period
  (LKM via hw_breakpoint_init, KPM spells out sample_period=1) -- a zero
  period is a non-sampling event whose callback never fires: armed, silent
- trace_addr2sym: contains_vaddr tests each PT_LOAD's loaded span instead
  of the whole-file extent; a PC in the unmapped gap between segments is
  refused rather than named after the nearest symbol below
- trace_diff: a bare all-digit token is indistinguishable from a date or
  count, so block keys need 0x or a non-decimal hex digit; a date-only log
  stays refused as a capture failure

Tests: two-layer quoting with a spaced control name, --seconds refusal,
sp/pc in the rejection list, sampling-period assertions for both templates,
a two-segment gap fixture, and a date-only parser case. 743 passed; flake8,
mypy and all five gates clean.
2026-10-15 15:33:28 +08:00

261 lines
12 KiB
Python

"""Unit tests for trace_addr2sym.py: the zero-dependency ELF64 walker, both offset
spaces, and the input adapters.
Two evidence layers, matching docs/tool-verification/EXTENSION-sampling-crash.md:
a hand-built minimal ELF (always runs, covers vaddr!=offset and the stripped
fallback) and a real cross-compiled aarch64 .so (runs where the toolchain exists,
cross-checked against readelf).
"""
import json
import shutil
import struct
import subprocess
import pytest
from conftest import load_script, run_script
tas = load_script("trace_addr2sym.py")
# ---------------------------------------------------------------------------
# a hand-built minimal ELF64: one PT_LOAD with vaddr != offset, one dynsym
# ---------------------------------------------------------------------------
def build_minimal_elf(path, vaddr_delta=0x1000, with_gap=False):
"""Two-symbol dynsym; one PT_LOAD, or two with an unmapped gap.
Symbols: sym_a @ 0x2000 (size 0x100), sym_b @ 0x2400 (size 0x10).
The load segment maps file [0x0,0x4000) at vaddr 0x1000..0x5000, so
file offset F <-> vaddr F + vaddr_delta for the segment, while the symbols
live at vaddr 0x2000/0x2400 regardless (they sit inside the segment).
with_gap=True shrinks segment 1 to vaddr 0x1000..0x3000 and adds a second
PT_LOAD at vaddr 0x4000..0x5000: the span [0x3000,0x4000) is loaded by
NEITHER segment, the shape real shared objects have between PT_LOADs.
"""
assert vaddr_delta == 0x1000
data = bytearray(b"\x00" * 0x4000)
# ELF header
data[0:4] = b"\x7fELF"
struct.pack_into("<B", data, 4, 2) # ELFCLASS64
struct.pack_into("<B", data, 5, 1) # little-endian
struct.pack_into("<H", data, 16, 3) # e_type = ET_DYN
struct.pack_into("<H", data, 18, 0xB7) # e_machine = AArch64
e_phoff, e_shoff = 0x40, 0x3000
struct.pack_into("<Q", data, 0x20, e_phoff)
struct.pack_into("<Q", data, 0x28, e_shoff)
struct.pack_into("<H", data, 0x36, 56) # e_phentsize
struct.pack_into("<H", data, 0x38, 1) # e_phnum
struct.pack_into("<H", data, 0x3A, 64) # e_shentsize
struct.pack_into("<H", data, 0x3C, 3) # e_shnum: null, dynsym, dynstr
struct.pack_into("<H", data, 0x3E, 2) # e_shstrndx (unused: names empty)
if with_gap:
struct.pack_into("<H", data, 0x38, 2) # e_phnum: two PT_LOADs
struct.pack_into("<IIQQQQQQ", data, e_phoff,
1, 5, 0, 0x1000, 0, 0x2000, 0x2000, 0x1000)
struct.pack_into("<IIQQQQQQ", data, e_phoff + 56,
1, 5, 0x2500, 0x4000, 0, 0x1000, 0x1000, 0x1000)
else:
# PT_LOAD file[0,0x4000) @ vaddr 0x1000 (8 fields, Elf64_Phdr:
# p_paddr included -- a 7-field pack silently shifts p_memsz onto p_align)
struct.pack_into("<IIQQQQQQ", data, e_phoff,
1, 5, 0, 0x1000, 0, 0x4000, 0x4000, 0x1000)
# dynstr @ 0x200, dynsym @ 0x400
strtab = b"\x00sym_a\x00sym_b\x00"
data[0x200:0x200 + len(strtab)] = strtab
# Elf64_Sym is 24B: st_name, st_info, st_other/shndx, st_value, st_size
def sym(name_off, info, value, size):
"""Sym."""
return struct.pack("<IBBHQQ", name_off, info, 0, 1, value, size)
syms = sym(1, (1 << 4) | 2, 0x2000, 0x100) + sym(7, (1 << 4) | 2, 0x2400, 0x10)
data[0x400:0x400 + len(syms)] = syms
# section headers @ 0x3000: null, dynsym(link=dynstr idx 2), dynstr
def shdr(sh_type, sh_offset, sh_size, sh_link=0, sh_entsize=24):
"""Shdr."""
return struct.pack("<IIQQQQIIQQ", 0, sh_type, 0, 0, sh_offset, sh_size,
sh_link, 0, 0, sh_entsize)
sh = (shdr(0, 0, 0)
+ shdr(11, 0x400, len(syms), sh_link=2)
+ shdr(3, 0x200, len(strtab), 0, 0))
data[e_shoff:e_shoff + len(sh)] = sh
with open(path, "wb") as fh:
fh.write(bytes(data))
def test_minimal_elf_roundtrip(tmp_path):
"""Minimal elf roundtrip."""
p = tmp_path / "min.so"
build_minimal_elf(p)
elf = tas.Elf64(str(p))
assert [s["name"] for s in elf.symbols] == ["sym_a", "sym_b"]
# vaddr -> file: the segment adds 0x1000 going file->vaddr
assert elf.vaddr_to_file(0x2000) == 0x1000 # 0x2000 - 0x1000
assert elf.file_to_vaddr(0x1000) == 0x2000
sym, off = elf.resolve(0x2042)
assert (sym, off) == ("sym_a", 0x42)
sym, off = elf.resolve(0x2405)
assert (sym, off) == ("sym_b", 5)
# past every symbol but still inside the load segment: nearest-below
# fallback (addr2line-style, gaps included) -- a PC OUTSIDE the segment is
# refused by describe(), not folded onto the last symbol (see below)
sym, off = elf.resolve(0x3f00)
assert (sym, off) == ("sym_b", 0x1b00)
# the span gate: a PC from another module must not be attributed here
assert elf.contains_vaddr(0x3f00) and not elf.contains_vaddr(0xdead0000)
assert "outside-elf" in tas.describe(elf, 0xdead0000)
def test_offs_are_file_flag_and_vaddr_space(tmp_path):
"""file offset input and vaddr input must not be conflated."""
p = tmp_path / "min.so"
build_minimal_elf(p)
r1 = run_script("trace_addr2sym.py", ["--elf", str(p), "--offs", "0x2042"])
assert r1.returncode == 0 and "sym_a+0x42" in r1.output
r2 = run_script("trace_addr2sym.py", ["--elf", str(p), "--offs", "0x2042",
"--offs-are-file"])
assert r2.returncode == 0
# file 0x2042 -> vaddr 0x3042: past sym_a's 0x100-size, nearest-below is sym_b
assert "sym_b+0xc42" in r2.output
def test_absolute_addrs_need_a_base(tmp_path):
"""Absolute addrs need a base."""
p = tmp_path / "min.so"
build_minimal_elf(p)
r = run_script("trace_addr2sym.py", ["--elf", str(p), "--addrs", "0x7b2000002042"])
assert r.returncode == 2
assert "load base" in r.output # ASLR: refuse rather than guess
r = run_script("trace_addr2sym.py", ["--elf", str(p), "--addrs", "0x7b2000002042",
"--base", "0x7b2000000000"])
assert r.returncode == 0 and "sym_a+0x42" in r.output
# ---------------------------------------------------------------------------
# real cross-compiled library (skipped where no aarch64 gcc exists)
# ---------------------------------------------------------------------------
@pytest.fixture(scope="module")
def real_so(tmp_path_factory):
"""Real so."""
cc = shutil.which("gcc")
if cc is None:
pytest.skip("no C compiler on this host")
probe = subprocess.run([cc, "-dumpmachine"], capture_output=True, text=True)
if "aarch64" not in (probe.stdout or ""):
pytest.skip("host gcc does not target aarch64 (%r)" % probe.stdout.strip())
d = tmp_path_factory.mktemp("so")
src = d / "tlib.c"
src.write_text("int encrypt_block(unsigned char *b,int n){for(int i=0;i<n;i++)"
"b[i]^=0x5A;return n;}\n"
"int decrypt_block(unsigned char *b,int n){return encrypt_block(b,n);}\n",
encoding="utf-8")
so = d / "libtlib.so"
built = subprocess.run([cc, "-shared", "-fPIC", "-O2", "-o", str(so), str(src)],
capture_output=True, text=True)
assert built.returncode == 0, built.stderr
return so
def test_real_so_symbols_agree_with_readelf(real_so):
"""Real so symbols agree with readelf."""
syms = subprocess.run(["readelf", "--syms", "-W", str(real_so)],
capture_output=True, text=True).stdout
addr = None
for line in syms.splitlines():
parts = line.split()
if len(parts) > 7 and parts[7] == "encrypt_block" and parts[3] == "FUNC":
addr = int(parts[1], 16)
break
assert addr is not None
r = run_script("trace_addr2sym.py", ["--elf", str(real_so), "--offs",
"0x%x" % (addr + 0x30), "--json"])
assert r.returncode == 0
doc = json.loads(r.stdout) # --json stdout is a pure JSON document
assert doc["resolved"][0]["where"].startswith("encrypt_block+0x30")
def test_real_so_stripped_fallback(real_so, tmp_path):
"""Real so stripped fallback."""
stripped = tmp_path / "stripped.so"
shutil.copy(real_so, stripped)
if shutil.which("strip"):
subprocess.run(["strip", str(stripped)], capture_output=True)
r = run_script("trace_addr2sym.py", ["--elf", str(stripped), "--offs", "0x610"])
assert r.returncode == 0
assert ".text+" in r.output # section+offset fallback, no crash
def test_hits_and_trace_adapters(tmp_path):
"""Hits and trace adapters."""
p = tmp_path / "min.so"
build_minimal_elf(p)
hits = tmp_path / "hits.log"
hits.write_text("HIT tid=1 kind=watch pc=0x7b2000002042 addr=0x7b200000feedd0 "
"sp=0x0 lr=0x0 x0=0x0 x1=0x0 x2=0x0 x3=0x0 x4=0x0 x5=0x0 x6=0x0 x7=0x0\n",
encoding="utf-8")
maps = tmp_path / "maps.txt"
# physically correct for this ELF: the segment (p_offset 0, p_vaddr 0x1000)
# is mapped at 0x7b2000001000, so the load bias is 0x7b2000001000 - 0x1000
maps.write_text("7b2000001000-7b2000005000 r--p 00000000 fd:01 1 "
"/data/app/min.so\n", encoding="utf-8")
r = run_script("trace_addr2sym.py", ["--hits", str(hits), "--maps", str(maps),
"--elf", str(p)])
assert r.returncode == 0, r.output
assert "sym_a+0x42" in r.output
# a neighbour whose name merely CONTAINS the hint must not be matched
maps.write_text("7b3000001000-7b3000005000 r--p 00000000 fd:01 1 "
"/data/app/min.so.v2\n", encoding="utf-8")
r = run_script("trace_addr2sym.py", ["--hits", str(hits), "--maps", str(maps),
"--elf", str(p)])
assert r.returncode == 2 and "no load bias" in r.output
trace = tmp_path / "t.log"
trace.write_text("BLK 0 min.so+0x2042\nBLK 1 xmin.so+0x2405\n", encoding="utf-8")
r = run_script("trace_addr2sym.py", ["--trace", str(trace), "--mod", "min.so",
"--elf", str(p)])
assert r.returncode == 0
assert "sym_a+0x42" in r.output and "sym_b+0x5" not in r.output
def test_load_bias_subtracts_segment_vaddr(tmp_path):
"""p_vaddr != 0: the bias is map_start - (foff - p_offset) - p_vaddr,
not the mapping start."""
p = tmp_path / "min.so"
build_minimal_elf(p)
elf = tas.Elf64(str(p))
maps = tmp_path / "maps.txt"
maps.write_text("7b2000001000-7b2000005000 r--p 00000000 fd:01 1 /data/app/min.so\n",
encoding="utf-8")
assert tas.load_bias_from_maps(str(maps), "min.so", elf) == 0x7b2000000000
# full-path hint matches too; a mere-substring neighbour does not
maps.write_text("7b2000001000-7b2000005000 r--p 00000000 fd:01 1 /data/app/min.so\n"
"7b3000000000-7b3000004000 r--p 00000000 fd:01 1 /data/app/min.so.v2\n",
encoding="utf-8")
assert tas.load_bias_from_maps(str(maps), "min.so", elf) == 0x7b2000000000
assert tas.load_bias_from_maps(str(maps), "/data/app/min.so", elf) == 0x7b2000000000
def test_pcs_in_the_gap_between_segments_are_refused(tmp_path):
"""Real shared objects leave unmapped gaps between PT_LOADs: a PC in a
gap is outside every segment and must be refused, not named after the
nearest symbol below it."""
p = tmp_path / "gap.so"
build_minimal_elf(p, with_gap=True)
elf = tas.Elf64(str(p))
assert elf.contains_vaddr(0x2000) # segment 1
assert elf.contains_vaddr(0x4500) # segment 2
assert not elf.contains_vaddr(0x3800) # the gap
assert "outside-elf" in tas.describe(elf, 0x3800)
def test_json_stdout_is_pure(tmp_path):
"""--json mode must be parseable as one document (no trailing RESULT=)."""
p = tmp_path / "min.so"
build_minimal_elf(p)
r = run_script("trace_addr2sym.py", ["--elf", str(p), "--offs", "0x2042", "--json"])
assert r.returncode == 0
doc = json.loads(r.stdout)
assert doc["status"] == "ok" and "RESULT" not in r.stdout