lamfold-squash 0.1.0

no_std clean-room SquashFS reader — the lamfold compressed-read-only frontend (the live-ISO root)
Documentation
  • Coverage
  • 100%
    2 out of 2 items documented0 out of 0 items with examples
  • Size
  • Source code size: 33.41 kB This is the summed size of all the files inside the crates.io package for this release.
  • Documentation size: 396.36 kB This is the summed size of all files generated by rustdoc for all configured targets
  • Ø build duration
  • this release: 9s Average build duration of successful builds.
  • all releases: 11s Average build duration of successful builds in releases after 2024-10-23.
  • Links
  • Homepage
  • lamco-admin/lamfold
    1 0 0
  • crates.io
  • Dependencies
  • Versions
  • Owners
  • glamberson

lamfold-squash

The SquashFS frontend of the lamfold read-only media stack — clean-room SquashFS 4.0 (gzip / xz / zstd / lz4 / lzo).

Reads over a lamfold::BlockSource and implements lamfold::FoldFrontend, so it composes through LamBoot's dispatch_fs_over_source — including recursively over a file inside another volume. SquashFS is the root filesystem inside almost every Linux live ISO, so this frontend is the recursion payoff: mounted over the squashfs-root file inside an ISO9660 image (a FileBlockSource loopback), it reads the live root directly. Spec: the lamfold design spec §4.

Status — compressed read-only frontend

Layer State
Superblock — magic, block size, compression id → substrate codec, table offsets ✅ done + tested
Compressed metadata blocks — 2-byte header, substrate decompress, cross-block spans ✅ done + tested
Inodes — basic + extended directories, files, symlinks ✅ done + tested
Directory listings — header runs + entries, child inode-ref reconstruction ✅ done + tested
Data blocks — per-block compressed/uncompressed flag, multi-block reassembly, sparse zero-fill ✅ done + tested
Fragments — the shared tail-block table, fragment read + decompress ✅ done + tested
read_link (symlink targets) ✅ done
xattrs, export/lookup table, directory index (we linear-scan), streaming per-block read_at ⏳ future

Complete. Verified against a real mksquashfs -comp gzip image: probe, tree walk, a nested directory, two fragment-packed small files, and a 300 KB three-block file (full + partial reads, EOF). Every compressed metadata, data, and fragment block is decompressed through the shared substrate codec. Builds default + --no-default-features, no_std, #![forbid(unsafe_code)] — every on-disk field read through bounds-checked little-endian helpers, every allocation through the substrate read cap.

Compression features

A SquashFS image is built with exactly one compressor; each maps to a substrate codec behind an opt-in cargo feature (all on by default):

feature compression id substrate codec
gzip 1 Codec::Zlib (miniz_oxide)
lzo 3 Codec::Lzo (lzokay)
xz 4 Codec::Xz (lzma-rust2)
lz4 5 Codec::Lz4 (lz4_flex)
zstd 6 Codec::Zstd (ruzstd)

Legacy lzma1 (id 2) is rejected with Unsupported. Disable the codecs you don't need to shrink the binary; an image whose compressor is disabled fails cleanly with Unsupported rather than mis-reading.

Clean-room posture

Derived only from the public SquashFS 4.0 format docs and the permissive references in NOTICE. The GPL implementations (Linux fs/squashfs, squashfs-tools) are fenced off — never read or copied.

Build / test

cargo build
cargo test                                   # walks tests/fixtures/gzip.sqfs
cargo build --no-default-features            # no_std check

MIT OR Apache-2.0.