Through the entire 0.13.x line, every CHECK CONDITION reply from the
drive (the standard way SCSI tells you why a sector failed) was being
collapsed into a synthetic status=0xFF, sense_key=0 transport-wedge
sentinel and the actual sense data was thrown away. Confirmed live on
the BU40N reading Dune 2 on 2026-04-27: drive returned host_status=0,
driver_status=8, status=2, exec_elapsed_ms=1416 on every bad sector
— a clean CHECK CONDITION carrying full sense data — and Disc::copy
was bailing on it as if the bridge had wedged.
Root cause: scsi/linux.rs's wedge check was
`host_status != 0 || driver_status != 0`
SG's DRIVER_SENSE bit (0x08) is set on every CHECK CONDITION reply
just to flag "sense buffer is populated" — it's not a transport
failure on its own. Pre-fix we conflated the two and silently lost
every drive-reported error reason. macOS and Windows backends had
the same shape: they extracted sense_key only, dropping ASC/ASCQ.
API restructure (clean separation):
Error::ScsiError {
opcode: u8,
status: u8, // 0xFF = synthetic transport-failure
sense: Option<ScsiSense>, // None ⇔ no sense delivered
}
pub struct ScsiSense { sense_key: u8, asc: u8, ascq: u8 }
impl ScsiSense {
pub fn is_marginal(&self) -> bool // keys 0/1/3/B
pub fn is_medium_error(&self) -> bool
pub fn is_hardware_error(&self) -> bool
pub fn is_unit_attention(&self) -> bool
pub fn is_data_protect(&self) -> bool
pub fn is_not_ready(&self) -> bool
pub fn is_illegal_request(&self) -> bool
pub fn is_aborted_command(&self) -> bool
}
impl Error {
pub fn scsi_sense(&self) -> Option<&ScsiSense>
pub fn is_scsi_transport_failure(&self) -> bool
pub fn is_marginal_read(&self) -> bool
}
SCSI protocol constants (SCSI_STATUS_*, SENSE_KEY_*) moved from
error.rs to scsi/mod.rs where they belong alongside SCSI_INQUIRY,
SCSI_READ_10, etc. parse_sense replaces parse_sense_key (returns the
full triple, not just the key); inline tests now exercise ASC/ASCQ
extraction at the right offsets for both descriptor (0x72/0x73) and
fixed (0x70/0x71) sense formats.
Disc::copy + Disc::patch sense-aware dispatch:
- marginal sense (MEDIUM ERROR / ABORTED COMMAND / RECOVERED ERROR
/ NO SENSE) → engage hysteresis (Block→Single, bpt=1)
- non-marginal sense (HARDWARE / DATA PROTECT / UNIT ATTENTION /
NOT READY / ILLEGAL REQUEST / transport failure / kernel
IoError) → bail with full sense info preserved; caller (autorip)
surfaces "physical replug" / "drive failing" / "media changed"
Pre-fix: every CHECK CONDITION → 0xFF synthetic → Disc::copy bailed
→ bytes_good froze at the bad zone. The hysteresis from v0.13.22
was correct but never got to run. This release unblocks it.
Disc::patch's wedged_threshold (50 consecutive failures) stays as
defense-in-depth for chains of marginal failures; a single
non-marginal sense now short-circuits it.
New phase=bail trace event records the bail reason with the sense
triple. phase=transport_err remains for genuine bridge wedges /
kernel timeouts; phase=scsi_err carries the parsed sense_key, asc,
ascq for drive-reported errors.
All 350 tests pass. Clippy clean across all targets.
Stream trait: read() returns PesFrame, write() accepts PesFrame.
A stream is a stream — you read from it or write to it.
No separate Input/Output traits.
API: libfreemkv::input(url) and libfreemkv::output(url, title, codecs)
Returns Box<dyn Stream>.
- Rename DriveSession → Drive across entire codebase
- find_drives() returns Vec<Drive>, find_drive() returns Option<Drive>
- resolve_device() now pub(crate) — internal only
- StreamUrl is now a typed enum (Disc, Mkv, M2ts, Iso, Network, Stdio, Null)
with scheme() and path_str() accessors, replacing struct of Strings
- Add lock_tray() / unlock_tray() for safe disc access during rips
- Improve reset() with eject cycle that clears LibreDrive stuck state
- Add Send bounds to ScsiTransport and PlatformDriver traits
- DiscOptions uses PathBuf instead of String for device/keydb paths
- Update doc example to use new Drive API
- Split AacsError { detail } into 13 specific error variants (AacsCertShort,
AacsAgidAlloc, AacsCertRejected, etc.) with unique error codes E7001-E7012
- Split DiscError { detail } into 7 specific variants (DiscRead, MplsParse,
ClpiParse, UdfNotFound, DiscNoTitles, DiscTitleRange, DiscNoExtents)
- Add WriteError (E5001), KeydbLoad (E8005), MuxLookahead (E9000), MuxWrite (E9001)
- Add OpenDisc API for single-call open+scan+rip workflow
- Remove all English text from error Display impl (code-only output)
- Normalize doc comments to use -- instead of em dash for ASCII consistency
Platform trait is no longer publicly exported. External code uses
DriveSession only — cannot call unlock, load_firmware, calibrate directly.
- Platform trait: pub(crate) with 3 methods only
- All handlers are private methods on Mt1959
- DriveStatus moved to mt1959 internal struct
- init() has guard: no re-init if already ready
- set_read_speed() has guard: no-op if not calibrated
- Removed open_unlocked() — open() is the only entry
- Removed Platform and DriveStatus from public exports
Prevents: out-of-sequence SCSI commands, double-init, wrong firmware writes.
- Removed open_unlocked() — open() is the only entry
- Removed redundant init() call from open_title()
- init() called once in open(), handles everything
- Each function does one thing: open→init→scan→read
status() (sub_cmd 0x13) returns ILLEGAL REQUEST on some drives.
Was retrying 6× with 30s timeouts = 180s hang during init().
Steps 1-10 all pass on hardware:
unlock: OK, load_firmware: OK, calibrate: OK,
register_a: OK, register_b: OK
Only status fails — not needed for reads.
A (): single WRITE_BUFFER → verify 0x45 → unlock×2
B (): WRITE handshake → READ 0x3000 → WRITE 16B → verify → unlock×5
9/10 handlers are identical A/B. Only load_firmware has different logic.
Both paths end with do_unlock() — firmware upload is a prerequisite for
unlock, not a substitute. init() tries unlock first, falls back to
load_firmware only on failure (cold boot).
Every handler traced instruction-by-instruction from operation: do_unlock with configurable response size
operation: WRITE_BUFFER + verify buf=0x45 + unlock×2
operation: do_unlock → validate → send pre-built CDB → [4:20]
operation: same with CDB B
operation: init → scan 0x0000-0x5800 → build table → triple speed
operation: ↔x86 VM only (host_write 16B), no SCSI
operation: do_unlock → validate → probe 0x13 → check sig → features
operation: 3 paths by param count (1/5/9), dynamic READ_BUFFER
operation: search 64-entry table → position probe →
set_cd_speed_max → custom SET_CD_SPEED with matched value
operation: ↔x86 VM only (host_read 8B), no SCSI
init() matches x86 dispatch exactly:
Phase 1: unlock → [load_fw] × 6
Phase 2: calibrate × 6
Phase 3: probe (drive info)
Phase 4: register A + B × 5
Phase 5: status × 6
Handlers 5/9 are VM communication (no SCSI equivalent in Rust).
All other handlers send real SCSI commands.
Back to basics: open, unlock, SET CD SPEED max, read.
Remove all calibration probes, register reads, maintain_speed calls.
This is closest to the build that hit 17 MB/s earlier.
Also: drive discovery moved to libfreemkv (find_drive, resolve_device),
AACS via UDF only, clean pipeline, sg device support.
Features:
- Open drive identification via SPC-4 INQUIRY + MMC-6 GET CONFIGURATION
- 141 supported drives with bundled profiles
- MT1959 platform: unlock, calibrate, raw sector reads
- DriveSpeed enum: BD1x-BD12x, DVD1x-DVD16x
- Field names follow SPC-4 §6.4.2 and MMC-6 §5.3.10 standards
- No proprietary fingerprints — open matching by SCSI fields
- Zero config: profiles compiled into binary
Tested on real hardware: HL-DT-ST BD-RE BU40N 1.03
MattJackson