Large GET DATA returns 61 xx. Concatenate clear 00 C0 bodies before
unwrap so 255-byte AES and 2K3DES objects print instead of exit 246.
Co-Authored-By: Cursor Grok 4.6 <noreply@cursor.com>
hf seos pacs on 2K3DES cards exited 236 with no message after mutual
auth. Print ExchangeAPDU14a failures, check mbedtls DES-CMAC without
mutating the caller key, decode SELECT 85 at its real length, and emit
short-form 85 when the cryptogram is under 128 bytes.
Co-Authored-By: Cursor Grok 4.6 <noreply@cursor.com>
Bootloader memory-dump cleanup sends CMD_PING to wake the communication
thread after clearing its run flag. The thread can be waiting for another
transmit after the dump ACK, so it sends this command before exiting.
The non-BWM bootrom does not handle CMD_PING: its default dispatch enters
Fatal(), leaving the device unresponsive even though the dump succeeded.
Use CMD_DEVICE_INFO for the wakeup instead. It is already supported by the
bootloader and used to identify it before the dump. Keep the wakeup so
closing the client does not leave the communication thread waiting.
Reproduced on PM5: the dump completed, but the following reset and device
probe failed. With only the rebuilt client changed, the same test script
passed five OS-to-bootloader, memory-dump and reset-to-OS cycles, with a
successful ping after each. No firmware was flashed for this comparison.
Built the client with AppleClang through Make, CMake and experimental_lib;
client/check passed. GCC 16 compiled the changed translation unit with
-Wall -Werror. Existing macOS linker warnings remain. The full GCC host
build was not repeated because of the previously observed Apple SDK
alignment error. Linux and Windows were not tested; this change replaces
an existing protocol command without changing platform-specific code.
Problem
The native hf mfu countertear used a device-side
CMD_HF_MFU_COUNTER_TEAROFF path with a static increment payload,
decided once per run. The pyscript mfuev1_counter_reset.py implements
the same attack with auto-calibration, a 3-phase attack loop,
per-attempt BD delay tuning (hunting) and stuck-value handling,
which the native command lacked.
Change
- rewrite the client command around the generic tear-off hook
(hw tearoff) and raw INCREASE_CNT frames, mirroring the pyscript
step by step: preflight checks, CHECK_TEARING_EVENT support gate,
targeted repair/normalize, BD sweep calibration (working delay
accepted only on 0xBD), and the DECREMENT -> PREPARE (0x00FFFF)
-> RESET phases with the same increment payloads (000100 for
decrement steps, 010000 for reset/repair)
- same delay tuning on counter movement: +1 when unchanged, -1 when
a jump exceeds 10, delay locked after a successful decrement
- remove CMD_HF_MFU_COUNTER_TEAROFF usage from the client; device
side stays untouched
- add Lua port client/luascripts/mfuev1_counter_reset.lua with the
same behaviour, using core.console capture and Enter-to-stop
- regenerate doc/commands.md and doc/commands.json
Behaviour change
Same CLI shape as the pyscript (-c/-f/--DelayBD/--Delay00, legacy
-i/-l/-s/-x options removed). Console output now matches the pyscript.
Testing
gcc warningless for client, armsrc and bootrom (RDV4). make check
green. doc regenerated with GD enabled. Verified 1:1 against the
pyscript and the new Lua port on hardware (EV1 card). Not tested
here: clang, Windows/ProxSpace and macOS. No build files are touched,
so the usual client-side review applies.
Co-Authored-By: opencode agent (GLM-5.3-Flash-Free) <noreply@opencode.ai>
Validate counter choices 0-2, treat only 0x00/0xBD as valid
CHECK_TEARING_EVENT answers, and detect NACK replies from the raw
payload, so retry logic does not mistake a NACK for a real answer.
Building the client with -march=x86-64-v3 (the CentOS/RHEL 10
baseline) fails with gcc 14:
src/emv/emvcore.c:211:24: error: writing 32 bytes into a region of
size 22 [-Werror=stringop-overflow=]
gcc vectorizes the dash fill loop in aidlist_rule() into 32-byte
stores and loses the pos + 2 < dstlen bound. The bound was correct.
Compute the dash count with the same limit and fill it with memset,
so there is no vectorized loop store to mis-size. Output is unchanged.
Reported in #3671.
Co-Authored-By: Grok Bot <noreply@users.noreply.github.com>
The cache is keyed on a hash of the graph buffer contents: the buffer has
many writers and no change counter.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Add an opt-in simulator probe that checks a reader's AES authentication continuation against a supplied candidate and stops before the final card response. Expose the result through the firmware protocol and etest bench command, and document the evidence and handshake scope.
Co-Authored-By: OpenAI Codex <noreply@openai.com>
New CLI command `lf cotag sim` with COTAG tag simulation implementation.
Detects COTAG reader activation sequence and emulates a COTAG tag.
Signed-off-by: andrej@towelbyte.net
save_graphbuffer() and save_buffer8() calloc a copy, and the restore_*()
helpers only copy out of it, so every owner leaked its save state.
check_chiptype() lost 4 * g_GraphTraceLen bytes plus the packed demod
buffer on each online `lf search` (~2 MB after the COTAG probe), and the
Visa2000 demod leaked on every `lf search`, online or offline.
Add free_buffer_savestate() and call it after the last restore in
check_chiptype(), demodTI(), demodVisa2k(), demodzx(),
try_detect_modulation() and t55xx_psk_untrim_head(), including the
early-return and success paths that never restored.
Reported with the check_chiptype() diagnosis and fix by @BeheSQ.
Fixes#3675
Co-Authored-By: Grok Bot <noreply@users.noreply.github.com>
At fc/2 the subcarrier is a one-sample alternation and pskRawDemod_ext finds
wave tops one sample at a time, unfiltered. On a weak tag, sample noise erases
tops; a wave then reads as four samples and counts as a phase shift, and the
bit stream fills with errors although every bit is still in the trace. A T5577
Indala at 22 mm gave the wrong card on 14 of 50 lf search reads.
PSKDemod keeps the tracker's result whenever it is clean, so nothing changes
where master already works. Otherwise, for fc/2, it tries the existing coherent
receiver pm3_psk_demod (now with an optional per-symbol magnitude output) and
keeps that only if it scores well above what level-step leakage can produce,
its bits aren't doubled or constant, and no symbol is weak. The receiver's
output is differential, so one bad decision would flip every later bit: a weak
symbol refuses the trace rather than risk a wrong card (absolute decisions with
weak symbols left to maxErr were tried and gave wrong IDs on dropouts). With no
usable clock from the caller, candidate clocks are tried largest first. Polarity is voted from the
raw level over each phase reversal, the cue pskFindFirstPhaseShift reads at the
first one. Client only; firmware is unchanged.
Base vs this change, same captures:
- T5577 Indala, live lf search: 22 mm 3 right / 14 wrong -> 50 / 0; 24 mm
0 / 2 -> 50 / 0; 26 mm and 0-20 mm unchanged
- lf t55xx dump, live: 22 and 24 mm blocks 0-2 and page 1 correct 5/5 where
master reads nothing; unchanged where master reads
- emulated Indala (Chameleon Ultra, clock drifting against the reader):
lf search right / wrong 56 / 130 -> 91 / 95 of 200; every one of the 95 is
master's own wrong ID on the same capture
- 27 PSK traces in traces/, weakened with Gaussian noise: correct 341 -> 636 of
810, wrong IDs 83 -> 21; with recorded Proxmark noise 367 -> 553 of 648,
33 -> 27. No lost reads and no new wrong IDs in either
- traces/lf_*.pm3 (99): lf search -1 and -1 -c output identical; 1,581
non-PSK probes: lf search and lf t55xx detect identical
- 200 empty-field captures: no decode, either build
tools/pm3_tests.sh: a weak 24 mm Indala trace must read its ID (fails on
master), and a synthetic NRZ with an fc/2 ripple must not decode as a tag
(the level-step leakage the score check rejects).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>