arm_bypass_bd/arm_bypass_uhd now also Set(Hrl, off) (revocation skip) so a
revoked/absent cert never trips the HRL check on a keyless bypass rip; tests
updated for the new SET/verify order.
FreemkvUnlocker::set sent the vendor SET with DataDirection::None + an
empty buffer, but the SET CDB advertises a MIN_ALLOC_LEN (64B) data-in
allocation and the drive ABORTS a no-data transfer (CHECK CONDITION /
Aborted Command — HW-confirmed on BU40N fw 0.8.1). Every Set(feature)
failed, so the load-bearing Set(Ake,off) failed and the whole freemkv
unlock aborted, silently falling back to the AACS/online key route
(autorip showed 'Unlockers — freemkv: no'). Mirror the working
FirmwareControl::set (exec_in): read the state table back over a
FromDevice phase and ignore it. Adds a regression test modelling the
drive's abort-without-data-phase contract.
Remove the HRL_WIPE_ONCE const, the wipe_hrl_once() setter, and their tests/doc references from the host library so wipe-once is no longer part of the host-facing firmware ABI. HRL exposes only skip (0x01).
Lockstep with freemkv-fw BD-refuse stealth fix: the BD force-disable wire
value moves from STATE_OFF (0x00, the SRAM boot value) to a distinct
sentinel STATE_BD_DISABLE (0x02), so an unarmed firmware image stays
OEM-identical. disable_bd() now sends 0x02; add the const and align the
Feature::Bd / Feature::Uhd ABI-mirror docs (enable dir / STATE_OFF are
OEM no-ops).
Fix two re-audit findings:
1. LOW (correctness): run_cert_handshake_with_anchors computed the
up-front cert-skip has_v2 gate WITHOUT allow_p256, unlike
attempt_one_cert. In production (allow_p256=false) a dead-v1 cert
carrying v2 creds was not skipped up front, reached the drive, failed
the step-7 self-verify, and consumed a MAX_CERT_ATTEMPTS slot -- so
[dead+v2 x3, valid] exhausted the cap and never tried the valid cert.
Mirror attempt_one_cert's gate exactly. Add a regression test.
2. LOW (key hygiene): add the zeroize crate and wipe sensitive key
material on drop -- HostCert private keys, AacsAuth/CertHandshake
derived bus_key/volume_id/read_data_key (ZeroizeOnDrop, non-secret
fields skipped), the ephemeral AKE scalars and raw private-scalar byte
buffers (Zeroizing). BigUint scalars/shared-x are a documented
best-effort residual (num-bigint exposes no zeroization). Debug
redaction and the EC math are untouched.
MEDIUM (coupled):
- aacs/handshake: on the AACS 1.0 path a type-0x11 drive cert is now
REJECTED (verify_cert + step-6 drive-key signature always run) instead
of accepted unverified — the old 0x11-skip did neither yet still ran
ECDH, an attacker-chosen-bus-key hole. run_cert_handshake routes a
genuine 0x11 drive to the native P-256 AKE via the v2-cred fallback.
- aacs/handshake: a post-auth VID/MAC failure (AacsVidMac) with v2 creds
present now falls through to aacs2_authenticate_p256 instead of a
terminal VidUnavailable, so a backward-compat 2.0 drive that fails the
1.0 VID MAC still reaches the P-256 path.
LOW:
- Gate the native AACS 2.0 (P-256) path behind AACS2_P256_EXPERIMENTAL
(default off) at the run_cert_handshake boundary: its cert offsets are
provisional, so production can't silently mis-verify a real disc.
- Release the AGID on the fully-successful handshake too (was leaked,
slowly draining the drive's 4-AGID pool).
- Rejection-sample the ephemeral ECDH private key (drop the modulo bias
of raw % n) in both generate_host_key_pair{,_p256}.
- Guard shared.infinity in compute_bus_key{,_p256} (return None).
- Remove the dead, truncating AacsAuth.drive_cert field.
- Wire UnlockError::NoUsableHostCert for the all-certs-skipped / empty
path (more accurate than HandshakeRejected).
- css: promote crypt_key's debug_assert! bounds checks to real assert!.
Handshake tests reworked onto an anchor-parametrised seam
(run_cert_handshake_with_anchors / *_with_anchor) so the DriveEmu drives
a genuine LA-signed 1.0 AKE under a self-generated test anchor; adds the
0x11-rejection, VID-MAC→P-256 fallback (HybridDrive), AGID-release,
infinity-guard, NoUsableHostCert and crypt_key-assert tests.
Add src/firmware: a hand-kept mirror of freemkv-fw/src/abi.rs (Verb/Feature
enums with pinned repr(u8) values, state constants, CDB builders) plus an
ergonomic FirmwareControl<&mut dyn ScsiTransport> with identity/get/set/reset/
dump, typed setters, states(), and the named recipes (arm_oem_bd, arm_oem_uhd,
arm_bypass_bd, arm_bypass_uhd, arm_stealth_oem). A drift test pins every wire
value against abi.rs.
Migrate FreemkvUnlocker off the old SubFn frame onto the new verb/feature
grammar via the shared builders (region-free=Set(Region,on),
riplock=Set(Speed,max), transport unlock=Set(Ake,null), bus-off=Set(Bus,off)).
Wire opt-in arming into AacsUnlocker::arm_before_unlock (OFF by default): when
set, unlock() detects freemkv via IDENTITY and applies a recipe before the AKE;
a non-freemkv drive or a refused recipe is left untouched.
Add `aacs1_keypair_matches(private_key, cert)` — true iff a stored AACS 1.0
host private key matches its certificate's public key (`priv·G ==
cert_pub_key`), reusing the live handshake's own curve math. A genuine,
LA-signed cert can still be paired in a keydb with the wrong private key,
which never establishes a bus key ("KEY NOT ESTABLISHED"); this primitive
distinguishes that DEAD pairing from a live one. Exposed for the key
service's host-cert health gate. Add `AacsHostSignVerify` (code 7016) for a
host-side self-verify guard on our own produced signature.
Also rework `run_cert_handshake` cert selection: on a rejected/revoked cert
the handshake now backs off and tries the next candidate instead of failing,
with a `cert_host_id_hex` helper and DriveEmu extensions + tests covering
single-valid, back-off-to-second, and roll-past-revoked paths.
find_by_drive_id gated the product-id-blind four-field fallback behind an empty
product_id in 1.7.0, so a drive reporting a specific product_id (e.g. an LG BU40N
reporting "BD-RE BU40N") no longer matched its catalog entry (generic "BD-RE"),
silently skipping the LibreDrive firmware unlocker — the rip fell through to the
AACS host-cert path (LD: no, AACS: yes).
Run the four-field fallback (vendor/revision/vendor_specific/firmware_date)
unconditionally but ambiguity-guarded: bind a UNIQUE catalog match regardless of
the reported product_id (the real-drive case), and require the exact product_id
only when catalogued siblings genuinely share the four-field identity — preserving
the anti-mis-bind protection the 1.7.0 gate intended. Regression test added.
Bump to 1.7.1.
identify() requested a 32-byte data-in allocation while the firmware commits
KNOCK_RESP_LEN (64) bytes on every subfn. The short allocation desyncs the first
command's data-in phase and wedges the drive, so the later bare 0xAD VID read
aborts (CHECK CONDITION) — the "Raw Read GOOD but no VID" failure that left the
keyserver with vid=none. Request KNOCK_RESP_LEN like every other knock; update the
CDB-shape tests. Also sync Cargo.lock to the 1.7.0 manifest.
Every unlocker now implements one method — unlock(&self, scsi, ctx) ->
Result<Option<Unlocked>> — with identical workflow and error handling,
differing only in CDBs/checks. Drops the unlock_features/unlock_bus split,
per-unlocker result state, and the drive_unlocked flag ("unlocked" is just
Some). Only a dead bus is Err; everything else falls through as Ok(None).
- VID read is best-effort everywhere via the shared crate::vid reader: a
VID miss no longer discards an unlock that already removed the bus — the
bug that dropped a freemkv-unlocked drive to the cert path and sent the
keyserver vid=none.
- AACS: host certs injected at construction (AacsUnlocker::new(certs)), not
passed per-call; a successful cert AKE is a real unlock.
- Renesas/Pioneer: the vendor open enables raw reads (LibreDrive parity),
then reads the VID via the same 0xAD reader — no host cert, matching the
firmware routes.
- Rename the MT1959 unlocker to LD (matches the ld/ module); the MediaTek
MT1959 platform driver keeps its name.
- Consumer assembles its own dispatch list from the exported types; no
central factory. Adds a uniform dead-bus contract test over all unlockers.
Fixed
- ld/profile: find_by_drive_id no longer mis-binds a drive whose reported
product_id is real but uncatalogued to a sibling model's profile (and its wrong
LibreDrive microcode); the product-id-blind loose match is gated on an empty id.
Docs / CI
- Third-party NOTICE for the LibreDrive microcode bundled in profiles.json;
cargo-deny gate (#15) and audience-based comment-guard (#17).
Matches the ld/ platform-folder convention (MediaTek lives in src/ld/); the
module/struct spelling is normalized renesis->renesas to match the drives it
detects. Pure move + identifier rename, no behavior change.
Declared rust-version and the ci.yml/qa.yml toolchain pins both sit at 1.88 (the
edition-2024 floor once zip 8 and let-chains are accounted for), so CI builds on
exactly the MSRV: an accidental bump to a newer toolchain fails CI by design and
moving the floor takes a deliberate pin edit. The uninlined_format_args allow
reconciles 1.88's clippy with the allow-by-default it became in 1.97+.
Matthew Jackson