82 Commits

Author SHA1 Message Date
Caleb Gardner c52fb15609 Finished (theoretically) file extraction. 2026-05-30 17:29:50 -05:00
Caleb Gardner 578911ba67 Work on extraction
Created DataExtractor & DataReader
Created Lookup tables
2026-05-30 06:22:26 -05:00
Caleb Gardner 56ad79ba94 Updated unsquashfs for zig 0.16.0
Fixed a couple bugs
Added scaffold for extraction
2026-05-29 18:50:45 -05:00
Caleb Gardner 2cb0863cc1 A BUNCH OF STUFF
Metadata reading
Directory reading
Decompile cache
Actual SfsFile implementation
Finished Inode implementation
Actually build unsquashfs
2026-05-29 06:20:06 -05:00
Caleb Gardner 2c47c7492e Reset for 0.16 (redux)
Starting from scratch
2026-05-28 06:13:00 -05:00
Caleb J. Gardner 4b2b7021c7 Moved & organized decompression
Fully implemented Decompressor vtable
2026-04-02 06:27:34 -05:00
Caleb J. Gardner a1b9828578 Finished (?) decompression restructuring 2026-03-26 06:40:17 -05:00
Caleb J. Gardner 8e4661c4c6 Moving decompression to a vtable interface 2026-03-22 06:35:25 -05:00
Caleb J. Gardner 54aaf30ea5 Working on re-doing decompression 2026-03-21 02:13:36 -05:00
Caleb J. Gardner df22cf6529 Started work on stateful decompression 2026-03-20 01:55:00 -05:00
Caleb J. Gardner 6b5c830234 Actually fix build 2026-03-18 06:17:07 -05:00
Caleb J. Gardner 116234cf9c Fixed build 2026-03-18 06:15:52 -05:00
Caleb J. Gardner 4601e8f323 Updated build 2026-03-18 06:13:19 -05:00
Caleb J. Gardner 50cae8b63d Use zig packed versions of zlib-ng, lz4, and zstd.
Changed use_c_libs to use_zig_decomp so c libraries are now default
2026-03-18 05:24:58 -05:00
Caleb J. Gardner 8b8c9a772f Added -Ddebug build option
Re-added errors from settings xattr values
2026-03-09 00:16:19 -05:00
Caleb Gardner f563648b20 Merge pull request #4 from CalebQ42/inode_finish
Re-do much of extraction
2026-03-08 22:36:37 -05:00
Caleb Gardner 8d4f3d72f8 Merge branch 'main' into inode_finish 2026-03-08 22:34:50 -05:00
Caleb J. Gardner 0b6129f1ae Switch to zlib-ng 2026-03-08 22:15:28 -05:00
Caleb J. Gardner 7308faad36 Moved kv free's in setMetadata 2026-03-05 12:54:25 -06:00
Caleb J. Gardner c9499251f8 Moved lookup tables into separate struct to fix some race conditions
Fixed lingering issues due to zero work size InodeFinish
Fixed xattrs not applying due to the keys sometimes not being
null-terminated.
Updated performance numbers
2026-03-05 12:20:30 -06:00
Caleb J. Gardner d470ca98e3 Added --force to unsquashfs
Fixing race condition bugs (yay)
2026-03-05 07:04:24 -06:00
Caleb J. Gardner a606f5e11a Move extract logic to util/extract.zig to make it easier to read. 2026-03-05 03:03:37 -06:00
Caleb J. Gardner a4e23a840d Updated performance values in README.
Added ability to ignore xattrs & permissions.
Ignore setting xattr errors due to an unknown issues.
2026-03-05 03:03:34 -06:00
Caleb J. Gardner 3a10572953 Updated performance values in README.
Added ability to ignore xattrs & permissions.
Ignore setting xattr errors due to an unknown issues.
2026-03-04 13:28:29 -06:00
Caleb J. Gardner edfe919c1b Expirementation with a new way to finish threads. Currently not working. 2026-03-04 06:39:44 -06:00
Caleb J. Gardner 4515610082 Added xattr support (currently untested)
Use stack fallback allocator on extraction for better performance.
2026-03-04 03:24:30 -06:00
Caleb J. Gardner beca6a2ae6 Small tweaks 2026-03-01 02:27:54 -06:00
Caleb J. Gardner 760e11df0b Fixed github workflow 2026-02-28 22:38:47 -06:00
Caleb J. Gardner 5f629df47c alloc-ified many functions.
Updated README
2026-02-28 22:33:57 -06:00
Caleb J. Gardner b0160e005b Some fixes 2026-02-17 05:54:32 -06:00
Caleb J. Gardner b7b99325da Slight improvement to how permissions are applied to folders
unsquashfs Verbose flag
2026-02-13 01:35:20 -06:00
Caleb J. Gardner b22e4d003d Update times 2026-02-12 08:57:36 -06:00
Caleb J. Gardner a41c37fef4 Updated unsquashfs version print out 2026-02-12 06:54:20 -06:00
Caleb J. Gardner 2d079d77f7 Exclusive file creation 2026-02-12 05:19:25 -06:00
Caleb J. Gardner 48f4235875 Fixed threads == 0 causing single threaded extraction.
Set exclusive file creation
2026-02-12 05:08:16 -06:00
Caleb J. Gardner 567dea8a0b Re-fix Action name 2026-02-12 04:28:32 -06:00
Caleb J. Gardner f78e5c7386 Rever to previous Action 2026-02-12 04:27:06 -06:00
Caleb J. Gardner 81e975c0d9 Try doing upload in two separate steps 2026-02-12 04:22:20 -06:00
Caleb J. Gardner fd274a8072 Different try to upload release files 2026-02-12 04:18:19 -06:00
Caleb J. Gardner 50ae79637e Removed broken ZSTD error value 2026-02-12 04:06:18 -06:00
Caleb J. Gardner ee41dc7278 Added import of zstd_errors.h 2026-02-12 03:58:08 -06:00
Caleb J. Gardner c34acebf51 Only link lzo if allow_lzo 2026-02-12 03:46:36 -06:00
Caleb J. Gardner bdbda29d39 Fixed wrong zig action line 2026-02-12 03:42:48 -06:00
Caleb J. Gardner ca4e867ddc Use setup zig Action 2026-02-12 03:39:05 -06:00
Caleb J. Gardner b066835066 Update apt 2026-02-12 02:51:01 -06:00
Caleb J. Gardner 2363bd7d10 sudo apt Actions 2026-02-12 02:43:55 -06:00
Caleb J. Gardner e619005b77 Fix action 2026-02-12 02:40:37 -06:00
Caleb J. Gardner 28b44891b3 Forgot Zig itself to the build 2026-02-12 02:32:34 -06:00
Caleb J. Gardner 4829c802a3 Clean version string 2026-02-12 02:32:06 -06:00
Caleb Gardner 570db9632a GH Action to create a release when release tagged 2026-02-12 02:30:21 -06:00
Caleb J. Gardner 0076294675 Small fix to testing to check if libc is linked 2026-02-12 02:19:38 -06:00
Caleb J. Gardner fd9e3d595b A bit of README cleanup
Added allow_lzo build option due to lzo build issues
2026-02-11 06:57:52 -06:00
Caleb J. Gardner b8189490eb Small change/fix when applying permissions 2026-02-09 13:50:56 -06:00
Caleb J. Gardner 6adc1d5c0c Fixes for threaded extraction.
Archive.extract now uses threaded extraction
2026-02-09 13:45:38 -06:00
Caleb J. Gardner 5ec12b5786 Finished adding multi-threaded extraction.
Added option in unsquashfs to specify the number of threads used.
Changed some functions to accept an allocator instead of just using
Archive's
Fixed run_tests.sh due to new c libraries
2026-02-08 15:14:35 -06:00
Caleb J. Gardner b892adacd7 Some work for threaded extraction 2026-02-08 10:27:35 -06:00
Caleb J. Gardner 2760ad6ccb Finished adding (untested) C decompression libraries. 2026-02-08 06:52:14 -06:00
Caleb J. Gardner 61311433b9 unsquashfs --help 2026-02-07 17:20:03 -06:00
Caleb J. Gardner 053d64a954 Some cleanup to build zon.
Minor start on threaded extraction
2026-02-07 17:20:03 -06:00
Caleb Gardner 0e0222cd02 Fix formatting for build flags in README 2026-02-07 11:34:34 -06:00
Caleb Gardner 9c0dfbadc2 Merge pull request #2 from CalebQ42/extract_expirement
Extract expirement
2026-02-07 11:31:51 -06:00
Caleb J. Gardner db2fb4b9f2 Fixed compilation when using zig test.
Remove option for static since -static should suffice in most situations
2026-02-07 11:24:58 -06:00
Caleb J. Gardner 067eaa87c2 You can now set when building to use c or zig libraries. 2026-02-07 10:58:32 -06:00
Caleb J. Gardner b64a3ec44a Archive.extract now directly uses Inode instead of File. 2026-02-07 06:57:55 -06:00
Caleb J. Gardner 704215e1a9 Remove updateTimes because I was running into integer overlow issues. 2026-02-07 06:42:34 -06:00
Caleb J. Gardner bcfd983f8d Fixed a handful of errors when extracting.
Fixed issues with unsquashfs
2026-02-07 06:28:27 -06:00
Caleb J. Gardner 75502da1d0 Remove DecompMgr in favor of a much simpler fn ptr.
Moved more functionality to Inode instead of File.
Started doing some optimization around allocation.
Slight rework of ExtractionOptions.
2026-02-07 05:09:17 -06:00
Caleb J. Gardner a316ba569f Renamed File.dev to File.devNum 2026-02-06 23:22:21 -06:00
Caleb J. Gardner a0f3f45885 Unsquashfs 2026-02-06 09:51:27 -06:00
Caleb J. Gardner f771ef7623 Work on extraction 2026-02-06 06:56:28 -06:00
Caleb J. Gardner 0d2576f5ee Fixed tables getting re-created 2026-01-31 06:39:41 -06:00
Caleb J. Gardner a76803aad1 Comments! 2026-01-31 05:14:00 -06:00
Caleb J. Gardner 1ff1e91d5e Added ExtractCompleteArchive test 2026-01-29 22:23:46 -06:00
Caleb J. Gardner 2bcbc16613 Extraction Finished-ish 2026-01-29 06:46:52 -06:00
Caleb J. Gardner 3c98cf2cdb EXTRACTION 2026-01-28 11:19:38 -06:00
Caleb J. Gardner 2c392cf250 Fixed a few bugs preventing basic functions
Worked on extraction, including creating DataReader
Added proper access to id, fragment, and export tables
2026-01-28 06:55:01 -06:00
Caleb J. Gardner 5d4e7b1435 Fixed things 2026-01-25 07:06:30 -06:00
Caleb J. Gardner 7aed59b5b1 Some fixes 2026-01-17 05:30:19 -06:00
Caleb J. Gardner f3fb8a128f Inodes! ExtractionOptions! Files! Directories! 2026-01-16 06:53:10 -06:00
Caleb J. Gardner 23bb19644b Finished (?) decompression, metadata reader, and Table. 2026-01-15 11:22:02 -06:00
Caleb J. Gardner ed14f13d9a Some work on tests & metadata 2026-01-15 06:56:43 -06:00
Caleb J. Gardner 428f938c3a Restart (once again) 2026-01-15 06:40:59 -06:00
26 changed files with 2351 additions and 1 deletions
+31
View File
@@ -0,0 +1,31 @@
name: Release Build
on:
push:
tags:
- "v*.*.*"
jobs:
build:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v6
- uses: mlugg/setup-zig@v2
- name: Install deps
run: sudo apt update && sudo apt install -y liblzma-dev liblzo2-dev
- name: Build normal version
run: zig build --release=fast -Duse_zig_decomp=true -Dversion=${{ github.ref_name }}
- name: Move zig build out
run: mv zig-out/bin/unsquashfs ./unsquashfs-x86_64-zig-libs
- name: Rebuild with C libraries
run: zig build --release=fast -Dversion="${{ github.ref_name }}"
- name: Move C build out
run: mv zig-out/bin/unsquashfs ./unsquashfs-x86_64-c-libs
- name: Release
uses: softprops/action-gh-release@v2
with:
prerelease: true
files: |
unsquashfs-x86_64-zig-libs
unsquashfs-x86_64-c-libs
+3
View File
@@ -1,2 +1,5 @@
testing/
.zig-cache/ .zig-cache/
zig-out/ zig-out/
zig-pkg/
+25
View File
@@ -0,0 +1,25 @@
// Project-local debug tasks
//
// For more documentation on how to configure debug tasks,
// see: https://zed.dev/docs/debugger
[
{
"label": "Build & Run",
"adapter": "CodeLLDB",
"request": "launch",
"build": {
"command": "zig",
"args": ["build", "-Ddebug=true"],
},
"program": "zig-out/bin/unsquashfs",
"args": [
"--force",
"-d",
"testing/TestExtractUnsquashfs",
"testing/LinuxPATest.sfs",
],
},
]
+1 -1
View File
@@ -1,6 +1,6 @@
MIT License MIT License
Copyright (c) 2025 Caleb Gardner Copyright (c) 2026 Caleb Gardner
Permission is hereby granted, free of charge, to any person obtaining a copy Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal of this software and associated documentation files (the "Software"), to deal
+75
View File
@@ -0,0 +1,75 @@
# zig-squashfs
This is my experiments to learn Zig. Might amount to something. Might not.
A library and application to decompress or view squashfs archives.
## Current State
Overall works, but currently is missing some features ([see below](#capabilities)) and has significantly slow performance compared to `unsquashfs` ([see below](#performance)).
## Build options
> `-Duse_c_libs=true`
Instead of using Zig's standard library for decompression, use the system's C libraries. Has the benefit of being much faster and enabling LZO and LZ4 decompression.
> `-Dallow_lzo=true`
Enable compiling with LZO decompression support. The LZO library currently has some issues with Zig when imported so it's easier to just disable it by default. Only has an effect when using `-Duse_c_libs=true`.
> `-Ddebug=true`
Sets various build options that make debugging easier. Specifically, debug optimization is forced, valgrind support is enabled, error tracing is enabled, stipping is disabled, and copmilation uses LLVM (this is due to some linking issues when on Debug optimization and is required for debugging tools such as `lldb`. In the future this may be removed from the debug flag).
> `-Dversion=0.0.0`
Sets the version of `unsquashfs` shown when `--version` is passed.
## Capabilities
Most features are present except for the following:
* When using Zig decompression libraries then lzo and lz4 compression types are unavailable. I don't _currently_ plan on spending the time to find and validate a library since neither is popular.
* When using C decompression libraries, lzo is not supported by default due to [some issues](#build-considerations). If it's needed it's trivial to fix, but it's easiest to just leave it disabled.
## Performance
This is some basic observation's I've made about this library's performance when compared to `unsquashfs`. Unless otherwise stated, most observations were made when extracting my test archive (which is fairly small and uses zstd compression) and with `--release=fast`.
Currently, my only performance checks are checking execution time, nothing deeper.
* Under ideal circumstances, my library is ~70% slower (.12s vs .20s).
* Using Zig decompression libraries *significantly* increases decompression time by ~600%. Under ideal circumstances.
* Performance improvements/regressions will be common. I'm still learning Zig.
Example Times:
* *unsquashfs, multi-threaded*: .12s
* *unsquashfs, single-threaded*: .13s
* *C-libs, single-threaded*: .45s
* *C-libs, multi-threaded*: .20s
* *Zig-libs, single-threaded*: 5.78s
* *Zig-libs, multi-threaded*: 1.08s
## Build considerations
Compilation without `use_c_libs` works completely fine, but Zig has issues with some symbols from the lzo library that needs to be manually fixed. In particular you need to fix the definitions for `lzo_bytep` and `lzo_voidp` to be `*u8` and `?*anyopaque` respectively. Due to this, you have to manually enable LZO decompression using `-Dallow_lzo=true` when building.
```zig
pub const lzo_bytep = @compileError("unable to translate C expr: unexpected token ''");
// /usr/include/lzo/lzoconf.h:148:9
pub const lzo_charp = @compileError("unable to translate C expr: unexpected token ''");
// /usr/include/lzo/lzoconf.h:149:9
pub const lzo_voidp = @compileError("unable to translate C expr: unexpected token ''");
```
to
```zig
pub const lzo_bytep = *u8;
// /usr/include/lzo/lzoconf.h:148:9
pub const lzo_charp = @compileError("unable to translate C expr: unexpected token ''");
// /usr/include/lzo/lzoconf.h:149:9
pub const lzo_voidp = ?*anyopaque;
```
+125
View File
@@ -0,0 +1,125 @@
const std = @import("std");
const Build = std.Build;
pub fn build(b: *std.Build) !void {
const use_zig_decomp = b.option(bool, "use_zig_decomp", "Use zig standard library for decompression.") orelse false;
const allow_lzo = b.option(bool, "allow_lzo", "Compile with lzo decompression support.") orelse false;
const dynamic = b.option(bool, "dynamic", "Dynamic link C decompression libraries.") orelse false;
const debug = b.option(bool, "debug", "Enable options to make debugging easier.") orelse false;
const version_string = b.option([]const u8, "version", "Version of the library/binary") orelse "0.0.0-testing";
const version: std.SemanticVersion = try .parse(version_string);
const zig_squashfs_options = b.addOptions();
zig_squashfs_options.addOption(bool, "use_zig_decomp", use_zig_decomp);
zig_squashfs_options.addOption(bool, "allow_lzo", allow_lzo);
const optimize = b.standardOptimizeOption(.{});
const target = b.standardTargetOptions(.{});
const c = b.addTranslateC(.{
.optimize = optimize,
.target = target,
.root_source_file = b.path("src/c.h"),
});
if (allow_lzo)
c.defineCMacro("ALLOW_LZO", null);
if (dynamic) {
c.linkSystemLibrary("z", .{});
c.linkSystemLibrary("lzma", .{});
c.linkSystemLibrary("lz4", .{});
c.linkSystemLibrary("zstd", .{});
if (allow_lzo)
c.linkSystemLibrary("minilzo", .{});
}
const deps = try getDependencies(b, optimize, target, allow_lzo);
const lib = b.addLibrary(.{
.name = "squashfs",
.use_llvm = debug,
.version = version,
.root_module = b.createModule(.{
.imports = &.{
.{ .name = "options", .module = zig_squashfs_options.createModule() },
.{ .name = "c", .module = c.createModule() },
},
.optimize = optimize,
.target = target,
.root_source_file = b.path("src/root.zig"),
.valgrind = debug,
}),
});
for (deps) |d|
lib.root_module.linkLibrary(d);
b.installArtifact(lib);
const exe_config = b.addOptions();
exe_config.addOption(std.SemanticVersion,"version", version);
const exe = b.addExecutable(.{
.name = "unsquashfs",
.use_llvm = debug,
.version = version,
.root_module = b.createModule(.{
.optimize = optimize,
.target = target,
.root_source_file = b.path("src/bin/unsquashfs.zig"),
.valgrind = debug,
.imports = &.{
.{ .name = "config", .module = exe_config.createModule() },
.{ .name = "squashfs", .module = lib.root_module }
},
}),
});
b.installArtifact(exe);
const lib_test = b.addTest(.{
.name = "squashfs-test",
.root_module = lib.root_module,
});
const test_step = b.step("test", "Run tests");
test_step.dependOn(&lib_test.step);
// zls check step
const lib_check = b.addLibrary(.{
.name = "squashfs-check",
.root_module = lib.root_module,
});
const exe_check = b.addLibrary(.{
.name = "unsquashfs-check",
.root_module = exe.root_module,
});
const check = b.step("check", "Check if squashfs compiles");
check.dependOn(&lib_check.step);
check.dependOn(&exe_check.step);
}
fn getDependencies(b: *Build, optimize: std.builtin.OptimizeMode, target: Build.ResolvedTarget, allow_lzo: bool) ![]*Build.Step.Compile {
const alloc = b.allocator;
var list: std.ArrayList(*Build.Step.Compile) = .empty;
const zlib_ng = b.dependency("zlib_ng", .{ .optimize = optimize, .target = target });
try list.append(alloc, zlib_ng.artifact("zng"));
const xz = b.dependency("xz", .{ .optimize = optimize, .target = target });
try list.append(alloc, xz.artifact("lzma"));
const lz4 = b.dependency("lz4", .{ .optimize = optimize, .target = target });
try list.append(alloc, lz4.artifact("lz4"));
const zstd = b.dependency("zstd", .{ .optimize = optimize, .target = target });
try list.append(alloc, zstd.artifact("zstd"));
if (allow_lzo) {
const minilzo = b.dependency("minilzo", .{ .optimize = optimize, .target = target });
try list.append(alloc, minilzo.artifact("minilzo"));
}
return list.toOwnedSlice(b.allocator);
}
+36
View File
@@ -0,0 +1,36 @@
.{
.name = .squashfs,
.version = "0.0.6",
.fingerprint = 0x37ba29474b87f145, // Changing this has security and trust implications.
.minimum_zig_version = "0.16.0",
.dependencies = .{
.zlib_ng = .{
.url = "git+https://github.com/CalebQ42/zig-zlib-ng#5f2f02dfb28acca2517dacbbd09e9b987f57b133",
.hash = "zlib_ng-2.3.3-pre1-2HYS4ClFAABW8KlHMyBHtlNKE3V7kCS8wqfxawG7xeaa",
},
.zstd = .{
.url = "git+https://github.com/allyourcodebase/zstd.git?ref=1.5.7-1#e1a501be57f42c541e8a5597e4b59a074dfd09a3",
.hash = "zstd-1.5.7-1-KEItkAMwAAD6OKY3m0OOmXG7aL-aLUfrDqbP5J5oYapU",
},
.lz4 = .{
.url = "git+https://github.com/allyourcodebase/lz4.git?ref=1.10.0-6#41f52ab227caf9d48cf88c89a4d2946caa12b102",
.hash = "lz4-1.10.0-6-ewyzw-4NAAAWDpY4xpiqr4LQhZQAC0x_rGnW2iPh6jk2",
},
.minilzo = .{
.url = "git+https://github.com/CalebQ42/zig-minilzo.git#7cbae997b91a44d74b7cd6c073584dc9562a6c90",
.hash = "minilzo-2.10.0-Ij7BO8wLAADeWI4Pe4jp8XTDsDaquZR14oZ7_9yKKDWP",
},
.xz = .{
.url = "git+https://github.com/akunaakwei/zig-xz.git#e2d389262c8291907e3e4c6fb119819141c16c0f",
.hash = "xz-5.8.2-6v47_JYeAABSL-jonprpL5-E_YaaGc4B5xrbe93WsJ3G",
},
},
.paths = .{
"build.zig",
"build.zig.zon",
"src",
"LICENSE",
"README.md",
},
}
Executable
+10
View File
@@ -0,0 +1,10 @@
#!/bin/sh
zig test \
-lc \
-lz \
-llzma \
-lminilzo \
-llz4 \
-lzstd \
src/test.zig
+185
View File
@@ -0,0 +1,185 @@
const std = @import("std");
const Io = std.Io;
const File = Io.File;
const MemoryMap = File.MemoryMap;
const Decomp = @import("decomp.zig");
const DecompCache = @import("decomp_cache.zig");
const Extract = @import("extract.zig");
const ExtractionOptions = @import("options.zig");
const Inode = @import("inode.zig");
const SfsFile = @import("file.zig");
const Archive = @This();
const CACHE_MEM_MAX = 1024 * 1024 * 1024;
super: Superblock,
cache: DecompCache,
pub fn init(alloc: std.mem.Allocator, io: Io, fil: File) !Archive {
return initAdvanced(alloc, io, fil, 0, 0);
}
pub fn initAdvanced(alloc: std.mem.Allocator, io: Io, fil: File, offset: u64, cache_memory_max: u64) !Archive {
var rdr = fil.reader(io, &[0]u8{});
try rdr.seekTo(offset);
var super: Superblock = undefined;
try rdr.interface.readSliceEndian(Superblock, @ptrCast(&super), .little);
try super.validate();
const map = try fil.createMemoryMap(io, .{
.offset = offset,
.len = super.size,
.protection = .{ .read = true },
});
return .{
.super = super,
.cache = try .init(
alloc,
map,
super.compression,
if (cache_memory_max != 0)
cache_memory_max
else
@min(CACHE_MEM_MAX, (try std.process.totalSystemMemory()) / 2),
),
};
}
pub fn deinit(self: *Archive, io: Io) void {
self.cache.deinit(io);
}
pub fn root(self: *Archive, alloc: std.mem.Allocator, io: Io) !SfsFile {
const inode: Inode = try .initRef(
alloc,
io,
&self.cache,
self.super.inode_start,
self.super.block_size,
self.super.root_ref,
);
return .init(alloc, self, inode, "");
}
pub fn open(self: *Archive, alloc: std.mem.Allocator, io: Io, filepath: []const u8) !SfsFile {
const path = std.mem.trim(u8, filepath, "/");
var root_file = try self.root(alloc, io);
if (path.len == 0 or path[0] == '.') return root_file;
defer root_file.deinit();
return root_file.open(alloc, io, filepath);
}
pub fn extract(self: *Archive, alloc: std.mem.Allocator, io: Io, ext_loc: []const u8, options: ExtractionOptions) !void {
const root_inode: Inode = try .initRef(
alloc,
io,
&self.cache,
self.super.inode_start,
self.super.block_size,
self.super.root_ref,
);
return Extract.extract(alloc, io, root_inode, &self.cache, self.super, ext_loc, options);
}
// Superblock
pub const Superblock = extern struct {
magic: u32,
inode_count: u32,
mod_time: u32,
block_size: u32,
frag_count: u32,
compression: Decomp.Enum,
block_log: u16,
flags: packed struct(u16) {
inode_uncompressed: bool,
data_uncompressed: bool,
check: bool,
frag_uncompressed: bool,
frag_never: bool,
frag_always: bool,
de_dupe: bool,
exportable: bool,
xattr_uncompressed: bool,
xattr_never: bool,
compression_options: bool,
id_uncompressed: bool,
_: u4,
},
id_count: u16,
ver_maj: u16,
ver_min: u16,
root_ref: Inode.Ref,
size: u64,
id_start: u64,
xattr_start: u64,
inode_start: u64,
dir_start: u64,
frag_start: u64,
export_start: u64,
pub fn validate(self: Superblock) !void {
if (self.magic != std.mem.readInt(u32, "hsqs", .little))
return error.BadMagic;
if (self.ver_maj != 4 or self.ver_min != 0)
return error.InvalidVersion;
if (self.block_log != std.math.log2(self.block_size))
return error.BadBlockLog;
if (self.flags.check)
return error.BadCheckFlag;
}
};
// Test
const TestArchive = "testing/LinuxPATest.sfs";
test "Basics" {
const alloc = std.testing.allocator;
const io = std.testing.io;
var archive_file = try Io.Dir.cwd().openFile(io, TestArchive, .{});
defer archive_file.close(io);
var arc: Archive = try .init(alloc, io, archive_file);
defer arc.deinit(io);
var root_file = try arc.root(alloc, io);
defer root_file.deinit();
}
const TestFile = "Start.exe";
const TestFileExtractLocation = "testing/Start.exe";
test "SingleFileExtraction" {
const alloc = std.testing.allocator;
const io = std.testing.io;
var archive_file = try Io.Dir.cwd().openFile(io, TestArchive, .{});
defer archive_file.close(io);
var arc: Archive = try .init(alloc, io, archive_file);
defer arc.deinit(io);
var ext_file = try arc.open(alloc, io, TestFile);
defer ext_file.deinit();
try ext_file.extract(alloc, io, TestFileExtractLocation, .default);
}
const TestFullExtractLocation = "testing/TestExtract";
test "FullExtraction" {
const alloc = std.testing.allocator;
const io = std.testing.io;
var archive_file = try Io.Dir.cwd().openFile(io, TestArchive, .{});
defer archive_file.close(io);
var arc: Archive = try .init(alloc, io, archive_file);
defer arc.deinit(io);
try arc.extract(alloc, io, TestFullExtractLocation, .default);
}
+136
View File
@@ -0,0 +1,136 @@
const std = @import("std");
const Io = std.Io;
const Writer = Io.Writer;
const File = Io.File;
const builtin = @import("builtin");
const config = @import("config");
const squashfs = @import("squashfs");
//TODO: Add more options
const help_mgs =
\\
\\Usage: unsquashfs [options] <archive>
\\
\\Options:
\\ -d <location> Extract to the given location instead of "squashfs-root"
\\
\\ -o <offset> Start reading the archive at the given offset.
\\ -dx Don't set xattr values
\\ -dp Don't set permissions (includes setting uid & gid owner)
\\
\\ -p <threads> Specify how many threads to use. If no present or zero, the system's logical cores count is used.
\\ -v Verbose
\\
\\ --force Force extraction. If the destination already exists, it will be deleted.
\\
\\ --help Display this messages
\\ --version Display the version
\\
;
const errors = error{InvalidArguments};
var archive: []const u8 = "";
var extLoc: []const u8 = "squashfs-root";
var offset: u64 = 0;
var threads: u32 = 0;
var verbose: bool = false;
var ignore_xattrs: bool = false;
var ignore_permissions: bool = false;
var force: bool = false;
pub fn main(init: std.process.Init) !void {
const alloc = init.gpa;
const io = init.io;
var stdout = File.stdout();
var out = stdout.writer(io, &[0]u8{});
defer out.interface.flush() catch {};
try handleArgs(init.minimal.args, &out.interface);
if (archive.len == 0) {
try out.interface.print("You must provide a squashfs archive\n", .{});
try out.interface.print(help_mgs, .{});
return;
}
var fil: File = try Io.Dir.cwd().openFile(io, archive, .{}); //TODO: Handle error gracefully.
defer fil.close(io);
var arc: squashfs.Archive = try .initAdvanced(alloc, io, fil, offset, 0); //TODO: Update when memory size matters. //TODO: Handle error gracefully.
defer arc.deinit(io);
const options: squashfs.ExtractionOptions = .{
.verbose = verbose,
.verbose_writer = if (verbose) &out.interface else null,
.ignore_xattr = ignore_xattrs,
.ignore_permissions = ignore_permissions,
};
if (force)
try Io.Dir.cwd().deleteTree(io, extLoc);
try arc.extract(alloc, io, extLoc, options); //TODO: Handle error gracefully.
}
fn handleArgs(args: std.process.Args, out: *Writer) !void {
var arg_iter = args.iterate();
defer arg_iter.deinit();
_ = arg_iter.skip(); // args[0] is the application launch command.
while (arg_iter.next()) |arg| {
if (std.mem.eql(u8, arg, "-o")) {
const nxt = arg_iter.next();
if (nxt == null or nxt.?.len == 0) {
try out.print("-o must be followed by a number\n", .{});
return errors.InvalidArguments;
}
offset = std.fmt.parseInt(u64, nxt.?, 10) catch {
try out.print("-o must be followed by a number\n", .{});
return errors.InvalidArguments;
};
continue;
} else if (std.mem.eql(u8, arg, "-d")) {
const nxt = arg_iter.next();
if (nxt == null or nxt.?.len == 0) {
try out.print("-d must be followed by a location\n", .{});
return errors.InvalidArguments;
}
extLoc = nxt.?;
continue;
} else if (std.mem.eql(u8, arg, "-p")) {
const nxt = arg_iter.next();
if (nxt == null or nxt.?.len == 0) {
try out.print("-p must be followed by a number\n", .{});
return errors.InvalidArguments;
}
threads = std.fmt.parseInt(u32, nxt.?, 10) catch {
try out.print("-p must be followed by a number\n", .{});
return errors.InvalidArguments;
};
continue;
} else if (std.mem.eql(u8, arg, "-v")) {
verbose = true;
continue;
} else if (std.mem.eql(u8, arg, "-dx")) {
ignore_xattrs = true;
continue;
} else if (std.mem.eql(u8, arg, "-dp")) {
ignore_permissions = true;
continue;
} else if (std.mem.eql(u8, arg, "--force")) {
force = true;
continue;
} else if (std.mem.eql(u8, arg, "--version")) {
try out.print("zig-unsquashfs v", .{});
try config.version.format(out);
try out.print("\nBuilt using Zig {s} in {} mode\n", .{ builtin.zig_version_string, builtin.mode });
std.process.exit(0);
return;
} else if (std.mem.eql(u8, arg, "--help")) {
try out.print(help_mgs, .{});
std.process.exit(0);
return;
}
if (archive.len > 0) {
try out.print("you can only provide one file at a time\n", .{});
try out.print(help_mgs, .{});
return errors.InvalidArguments;
}
archive = arg;
}
}
+7
View File
@@ -0,0 +1,7 @@
#include <zlib.h>
#include <lzma.h>
#include <lz4.h>
#include <zstd.h>
#ifdef ALLOW_LZO
#include <lzo/minilzo.h>
#endif
+61
View File
@@ -0,0 +1,61 @@
const std = @import("std");
const c = @import("c");
const Error = @import("decomp.zig").Error;
pub fn zlibDecompress(_: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var strem: c.z_stream = .{
.next_in = in.ptr,
.avail_in = @truncate(in.len),
.next_out = out.ptr,
.avail_out = @truncate(out.len),
};
var res = c.inflateInit(&strem);
if (res != c.Z_OK) return Error.ReadFailed;
defer _ = c.inflateEnd(&strem);
res = c.inflate(&strem, c.Z_FULL_FLUSH);
if (res != c.Z_OK) return Error.ReadFailed;
return strem.total_out;
}
pub fn lzmaDecompress(_: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var strem: c.lzma_stream = .{
.next_in = in.ptr,
.avail_in = in.len,
.next_out = out.ptr,
.avail_out = out.len,
};
var res = c.lzma_auto_decoder(&strem, out.len * 2, 0);
if (res != c.LZMA_OK) return Error.ReadFailed;
defer c.lzma_end(&strem);
while (res == c.LZMA_OK)
res = c.lzma_code(&strem, c.LZMA_RUN);
if (res != c.LZMA_FINISH) return Error.ReadFailed;
return strem.total_out;
}
pub fn lzoDecompress(_: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var out_len = out.len;
const res = c.lzo1x_decompress(in.ptr, in.len, out.ptr, &out_len, null);
if (res != c.LZO_E_OK) return Error.ReadFailed;
return out_len;
}
pub fn lz4Decompress(_: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
const res = c.LZ4_decompress_safe(
in.ptr,
out.ptr,
@bitCast(@as(u32, @truncate(in.len))),
@bitCast(@as(u32, @truncate(out.len))),
);
if (res < 0) return Error.ReadFailed;
return @abs(res);
}
pub fn zstdDecompress(_: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
const res = c.ZSTD_decompress(out.ptr, out.len, in.ptr, in.len);
if (c.ZSTD_isError(res) != 0)
return Error.ReadFailed;
return res;
}
+102
View File
@@ -0,0 +1,102 @@
const std = @import("std");
const Io = std.Io;
const DecompCache = @import("../decomp_cache.zig");
const DataBlock = @import("../inode.zig").DataBlock;
const Extractor = @This();
cache: *DecompCache,
block_size: u32,
start: u64,
size: u64,
blocks: []DataBlock,
frag_data: ?[]u8 = null,
frag_offset: u32 = 0,
pub fn init(cache: *DecompCache, block_size: u32, size: u64, start: u64, blocks: []DataBlock) Extractor {
return .{
.cache = cache,
.block_size = block_size,
.start = start,
.size = size,
.blocks = blocks,
};
}
pub fn addFragment(self: *Extractor, data: []u8, offset: u32) void {
self.frag_data = data;
self.frag_offset = offset;
}
pub fn asyncExtract(self: Extractor, io: Io, fil: Io.File) Error!void {
try fil.writePositionalAll(io, &.{&.{0}}, self.size - 1);
var map = try fil.createMemoryMap(io, .{ .len = self.size, .protection = .{ .write = true } });
defer map.destroy(io);
var group: Io.Group = .init;
defer group.cancel(io);
var ret_err: ?Error = null;
var offset = self.start;
for (0..self.blocks.len) |i| {
group.async(io, blockThread, .{ self, io, map, offset, i, &ret_err });
offset += self.blocks[i].size;
}
if (self.frag_data != null)
group.async(io, fragThread, .{ self, map });
group.await(io) catch |err| return ret_err orelse err;
try map.write(io);
}
fn blockThread(self: Extractor, io: Io, map: Io.File.MemoryMap, read_offset: u64, idx: usize, ret_err: *?Error) error{Canceled}!void {
const write_pos = idx * self.block_size;
const size = if (self.frag_data == null and idx == self.block_size.len - 1)
self.size % self.block_size
else
self.block_size;
const block = self.blocks[idx];
if (block.size == 0) {
@memset(map.memory[write_pos..][0..size], 0);
return;
}
if (block.uncompressed) {
@memcpy(map[write_pos..][0..size], self.cache.map.memory[read_offset..][0..size]);
return;
}
const data = self.cache.get(io, read_offset, block.size, size) catch |err| switch (err) {
error.Canceled => {
io.recancel();
return error.Canceled;
},
else => |e| {
ret_err.* = e;
return error.Canceled;
},
};
defer self.cache.finished(io, read_offset);
if (data.len != size) {
std.debug.print("Size of decompression at {} is {} and should be {}\n", .{ read_offset, data.len, size });
return Error.BadDecompressionSize;
}
@memcpy(map[write_pos..][0..size], data);
}
fn fragThread(self: Extractor, map: Io.File.MemoryMap) error{Canceled}!void {
const write_pos = self.blocks.len * self.block_size;
const size = self.size % self.block_size;
@memcpy(map.memory[write_pos..][0..size], self.frag_data.?[self.frag_offset..][0..size]);
}
// Types
pub const Error = error{BadDecompressionSize} || Io.File.WritePositionalError || Io.File.MemoryMap.CreateError;
+151
View File
@@ -0,0 +1,151 @@
const std = @import("std");
const Io = std.Io;
const DecompCache = @import("../decomp_cache.zig");
const DataBlock = @import("../inode.zig").DataBlock;
const Reader = @This();
io: Io,
cache: *DecompCache,
block_size: u32,
size: u64,
blocks: []DataBlock,
frag_data: ?[]u8 = null,
frag_offset: u32 = 0,
cur_offset: u64 = 0,
next_offset: u64,
idx: u32 = 0,
cur_block_sparse: bool = false,
interface: Io.Reader = .{
.buffer = &[0]u8{},
.end = 0,
.seek = 0,
.vtable = &.{
.stream = stream,
.discard = discard,
.readVec = readVec,
},
},
pub fn init(io: Io, cache: *DecompCache, block_size: u32, size: u64, start: u64, blocks: []DataBlock) Reader {
return .{
.io = io,
.cache = cache,
.block_size = block_size,
.size = size,
.blocks = blocks,
.next_offset = start,
};
}
pub fn deinit(self: Reader) void {
self.cache.finished(self.io);
}
pub fn addFragment(self: *Reader, data: []u8, offset: u32) void {
self.frag_data = data;
self.frag_offset = offset;
}
fn advance(self: *Reader) Io.Reader.Error!void {
errdefer self.interface.end = 0;
self.interface.seek = 0;
if (self.idx > self.blocks.len) return error.EndOfStream;
defer self.idx += 1;
self.cache.finished(self.io, self.cur_offset);
if (self.idx == self.blocks.len) {
if (self.frag_data == null) return error.EndOfStream;
self.cur_offset = 0;
const size = self.size % self.block_size;
self.interface.buffer = self.frag_data.?[self.frag_offset..][0..size];
self.interface.end = size;
return;
}
const block = self.blocks[self.idx];
const size = if (self.idx == self.blocks.len - 1 and self.frag_data == null)
self.size % self.block_size
else
self.block_size;
if (block.size == 0) {
self.interface.buffer = &[0]u8{};
self.cur_block_sparse = true;
self.interface.end = size;
return;
} else {
self.cur_block_sparse = false;
}
self.cur_offset = self.next_offset;
self.next_offset = self.cur_offset + block.size;
if (block.uncompressed) {
self.interface.buffer = self.cache.map.memory[self.cur_offset..][0..size];
self.interface.end = size;
return;
}
const data = self.cache.get(self.io, self.cur_offset, block.size, size);
if (data.len != size) {
std.debug.print("Size of decompression at {} is {} and should be {}\n", .{ self.cur_offset, data.len, size });
return Io.Reader.Error.ReadFailed;
}
self.interface.buffer = data;
self.interface.end = size;
}
fn stream(r: *Io.Reader, w: *Io.Writer, limit: Io.Limit) Io.Reader.StreamError!usize {
const self: *Reader = @fieldParentPtr("interface", r);
if (r.seek >= r.end) {
try self.advance();
}
const to_write = @min(@intFromEnum(limit), r.end - r.seek);
const wrote = try if (self.cur_block_sparse)
w.splatByte(0, to_write)
else
w.write(r.buffer[r.seek..][0..to_write]);
r.seek += wrote;
return wrote;
}
fn discard(r: *Io.Reader, limit: Io.Limit) Io.Reader.Error!usize {
if (r.seek >= r.end) {
const self: *Reader = @fieldParentPtr("interface", r);
try self.advance();
}
const to_discard = @min(@intFromEnum(limit), r.end - r.seek);
r.seek += to_discard;
return to_discard;
}
fn readVec(r: *Io.Reader, vec: [][]u8) Io.Reader.Error!usize {
const self: *Reader = @fieldParentPtr("interface", r);
if (r.seek >= r.end) {
try self.advance();
}
var total: usize = 0;
for (vec) |v| {
const to_copy = @min(v.len, r.end - r.seek);
if (self.cur_block_sparse) {
@memset(v[0..to_copy], 0);
} else {
@memcpy(v[0..to_copy], r.buffer[r.seek..][0..to_copy]);
}
total += to_copy;
r.seek += to_copy;
if (r.seek >= r.end) break;
}
return total;
}
+41
View File
@@ -0,0 +1,41 @@
const std = @import("std");
const options = @import("options");
const c_decomp = @import("c_decomp.zig");
const zig_decomp = @import("zig_decomp.zig");
pub const Error = error{} || std.Io.Reader.UnlimitedAllocError;
pub const Enum = enum(u16) {
zlib = 1,
lzma,
lzo,
xz,
lz4,
zstd,
};
pub const Fn = *const fn (std.mem.Allocator, in: []u8, out: []u8) Error!usize;
pub fn DecompFn(comp: Enum) !Fn {
return if (options.use_zig_decomp)
switch (comp) {
.zlib => zig_decomp.zlibDecompress,
.lzma => zig_decomp.lzmaDecompress,
.xz => zig_decomp.xzDecompress,
.zstd => zig_decomp.zstdDecompress,
.lz4 => error.Lz4Unsupported,
.lzo => error.LzoUnsupported,
}
else switch (comp) {
.zlib => c_decomp.zlibDecompress,
.lzma, .xz => c_decomp.lzmaDecompress,
.zstd => c_decomp.zstdDecompress,
.lz4 => c_decomp.lz4Decompress,
.lzo => if (options.allow_lzo)
c_decomp.zstdDecompress
else
error.LzoUnsupported,
};
}
+116
View File
@@ -0,0 +1,116 @@
const std = @import("std");
const Io = std.Io;
const File = Io.File;
const MemoryMap = File.MemoryMap;
const Atomic = std.atomic.Value;
const Decomp = @import("decomp.zig");
const DecompCache = @This();
alloc: std.mem.Allocator,
map: MemoryMap,
decomp_fn: Decomp.Fn,
cache: std.AutoHashMap(u64, Cache),
mut: std.Io.RwLock = .init,
cond: std.Io.Condition = .init,
max_mem: u64,
cur_mem: u64 = 0,
pub fn init(alloc: std.mem.Allocator, map: MemoryMap, compression: Decomp.Enum, max_mem: u64) !DecompCache {
return .{
.alloc = alloc,
.map = map,
.decomp_fn = try Decomp.DecompFn(compression),
.cache = .init(alloc),
.max_mem = max_mem,
};
}
pub fn deinit(self: *DecompCache, io: Io) void {
self.mut.lockUncancelable(io);
var iter = self.cache.valueIterator();
while (iter.next()) |v|
self.alloc.free(v.data);
self.cache.deinit();
}
pub fn get(self: *DecompCache, io: Io, offset: u64, compressed_size: u32, max_size: u32) ![]u8 {
{
try self.mut.lockShared(io);
defer self.mut.unlockShared(io);
const cache = self.cache.getPtr(offset);
if (cache != null) {
_ = cache.?.usage.fetchAdd(1, .acquire);
return cache.?.data;
}
}
try self.mut.lock(io);
defer self.mut.unlock(io);
const cache = try self.cache.getOrPut(offset);
if (cache.found_existing) {
_ = cache.value_ptr.usage.fetchAdd(1, .acquire);
return cache.value_ptr.data;
}
errdefer self.cache.removeByPtr(cache.key_ptr);
try self.ensureSpace(io, max_size);
var out = try self.alloc.alloc(u8, max_size);
errdefer self.alloc.free(out);
const decomp_size = try self.decomp_fn(self.alloc, self.map.memory[offset..][0..compressed_size], out);
if (decomp_size != max_size) {
if (!self.alloc.resize(out, decomp_size)) {
const new_out = try self.alloc.alloc(u8, decomp_size);
@memcpy(new_out, out[0..decomp_size]);
out = new_out;
} else {
out.len = decomp_size;
}
}
self.cur_mem += decomp_size;
cache.value_ptr.data = out;
_ = cache.value_ptr.usage.fetchAdd(1, .acquire);
return out;
}
pub fn finished(self: *DecompCache, io: Io, offset: u64) void {
const cache = self.cache.getPtr(offset);
if (cache == null) {
std.debug.print("Finished using cache, but cache does not exist: {}\n", .{offset});
return;
}
const use = cache.?.usage.fetchSub(1, .acquire);
if (use == 0)
self.cond.broadcast(io);
}
fn ensureSpace(self: *DecompCache, io: Io, size: u64) !void {
while (self.cur_mem + size > self.max_mem) {
var iter = self.cache.valueIterator();
while (iter.next()) |cache| {
if (cache.usage.load(.unordered) == 0) {
self.alloc.free(cache.data);
self.cur_mem -= cache.data.len;
if (self.cur_mem + size <= self.max_mem) return;
}
}
if (self.cur_mem + size <= self.max_mem) return;
try self.cond.wait(io, &self.mut.mutex);
}
}
// Types
const Cache = struct {
data: []u8,
usage: Atomic(u32),
};
+83
View File
@@ -0,0 +1,83 @@
const std = @import("std");
const Io = std.Io;
const Reader = Io.Reader;
const Inode = @import("inode.zig");
const Directory = @This();
entries: []Entry,
pub fn init(alloc: std.mem.Allocator, rdr: *Reader, size: u32) Error!Directory {
if (size <= 3) return .{ .entries = &[0]Entry{} };
var entries: std.ArrayList(Entry) = try .initCapacity(alloc, 50);
errdefer {
for (entries.items) |ent|
ent.deinit(alloc);
entries.deinit(alloc);
}
var read: u32 = 3;
while (read < size) {
var hdr: Header = undefined;
try rdr.readSliceEndian(Header, @ptrCast(&hdr), .little);
read += @sizeOf(Header);
try entries.ensureUnusedCapacity(alloc, hdr.count + 1);
for (0..hdr.count + 1) |_| {
var raw: RawEntry = undefined;
try rdr.readSliceEndian(RawEntry, @ptrCast(&raw), .little);
const name = try alloc.alloc(u8, raw.name_size + 1);
errdefer alloc.free(name);
try rdr.readSliceEndian(u8, name, .little);
entries.appendAssumeCapacity(.{
.inode_num = if (raw.inode_num_offset > 0)
hdr.inode_num + @abs(raw.inode_num_offset)
else
hdr.inode_num - @abs(raw.inode_num_offset),
.block_start = hdr.block_start,
.block_offset = raw.block_offset,
.type = raw.type,
.name = name,
});
}
}
return .{ .entries = try entries.toOwnedSlice(alloc) };
}
pub fn deinit(self: Directory, alloc: std.mem.Allocator) void {
for (self.entries) |entry|
entry.deinit(alloc);
alloc.free(self.entries);
}
// Types
pub const Error = Reader.Error || std.mem.Allocator.Error;
pub const Entry = struct {
inode_num: u32,
block_start: u32,
block_offset: u16,
type: Inode.Enum,
name: []const u8,
pub fn deinit(self: Entry, alloc: std.mem.Allocator) void {
alloc.free(self.name);
}
};
const Header = extern struct {
count: u32,
block_start: u32,
inode_num: u32,
};
const RawEntry = extern struct {
block_offset: u16,
inode_num_offset: i16,
type: Inode.Enum,
name_size: u16,
};
+283
View File
@@ -0,0 +1,283 @@
const std = @import("std");
const Io = std.Io;
const Atomic = std.atomic.Value;
const DecompCache = @import("decomp_cache.zig");
const ExtractionOptions = @import("options.zig");
const Inode = @import("inode.zig");
const Superblock = @import("archive.zig").Superblock;
const Directory = @import("directory.zig");
const DataExtractor = @import("data/extractor.zig");
const DataReader = @import("data/reader.zig");
const Lookup = @import("lookup.zig");
pub fn extract(alloc: std.mem.Allocator, io: Io, inode: Inode, cache: *DecompCache, super: Superblock, ext_loc: []const u8, options: ExtractionOptions) !void {
const path = std.mem.trim(u8, ext_loc, "/");
var buf: [50]ReturnUnion = undefined;
var sel: Io.Select(ReturnUnion) = .init(io, &buf);
defer {
while (sel.cancel()) |ret| {
switch (ret) {
.dir_ret => |d| {
const res = d catch continue;
alloc.free(res.path);
},
.file_ret => |f| {
const res = f catch continue;
alloc.free(res.path);
},
else => {},
}
}
}
var frag_table: Lookup.Table(Lookup.FragmentEntry) = .init(alloc, cache, super.frag_start, super.frag_count);
defer frag_table.deinit();
var ret_loop = io.async(returnLoop, .{ alloc, &sel, options });
try extractReal(alloc, io, cache, super, &sel, &frag_table, path, inode, null, false);
ret_loop.await(io) catch |err| {
// TODO: Drain sel
return err;
};
}
fn extractReal(
alloc: std.mem.Allocator,
io: Io,
cache: *DecompCache,
super: Superblock,
sel: *Io.Select(ReturnUnion),
frag_table: *Lookup.Table(Lookup.FragmentEntry),
path: []const u8,
inode: Inode,
parent: ?*Atomic(usize),
origin: bool,
) Error!void {
try io.checkCancel();
switch (inode.data) {
.dir, .ext_dir => sel.async(
.dir_ret,
extractDir,
.{ alloc, io, cache, super, sel, frag_table, path, inode, parent, origin },
),
.file, .ext_file => sel.async(
.file_ret,
extractFile,
.{ alloc, io, cache, super.block_size, frag_table, path, inode, parent, origin },
),
else => return error.Canceled,
}
}
fn extractDir(
alloc: std.mem.Allocator,
io: Io,
cache: *DecompCache,
super: Superblock,
sel: *Io.Select(ReturnUnion),
frag_table: *Lookup.Table(Lookup.FragmentEntry),
path: []const u8,
inode: Inode,
parent: ?*Atomic(usize),
origin: bool,
) Error!DirReturn {
defer {
if (parent != null)
_ = parent.?.fetchSub(1, .acquire);
if (!origin) inode.deinit(alloc);
}
errdefer if (!origin) alloc.free(path);
const dir = inode.directory(alloc, io, cache, super.dir_start) catch |err| switch (err) {
error.NotDirectory => unreachable,
else => |e| return e,
};
defer dir.deinit(alloc);
const sub_files = try alloc.create(Atomic(usize));
sub_files.* = .init(dir.entries.len);
const ret: DirReturn = .{
.path = path,
.sub_files = sub_files,
.origin = origin,
.uid_idx = inode.hdr.uid_idx,
.gid_idx = inode.hdr.gid_idx,
.mod_time = inode.hdr.mod_time,
.permissions = inode.hdr.permission,
.xattr_idx = switch (inode.data) {
.ext_dir => |d| if (d.xattr_idx != 0xFFFFFFFF) d.xattr_idx else null,
else => null,
},
};
for (dir.entries) |entry| {
const new_inode: Inode = try .initDirEntry(alloc, io, cache, super.inode_start, super.block_size, entry);
errdefer new_inode.deinit(alloc);
const new_path = try std.mem.concat(alloc, u8, &.{ path, "/", entry.name });
try extractReal(
alloc,
io,
cache,
super,
sel,
frag_table,
new_path,
new_inode,
sub_files,
false,
);
}
return ret;
}
fn extractFile(
alloc: std.mem.Allocator,
io: Io,
cache: *DecompCache,
block_size: u32,
frag_table: *Lookup.Table(Lookup.FragmentEntry),
path: []const u8,
inode: Inode,
parent: ?*Atomic(usize),
origin: bool,
) Error!FileReturn {
defer {
if (parent != null)
_ = parent.?.fetchSub(1, .acquire);
if (!origin) inode.deinit(alloc);
}
errdefer if (!origin) alloc.free(path);
var atomic = try Io.Dir.cwd().createFileAtomic(io, path, .{});
defer atomic.deinit(io);
var ret: FileReturn = .{
.path = path,
.origin = origin,
.uid_idx = inode.hdr.uid_idx,
.gid_idx = inode.hdr.gid_idx,
.permissions = inode.hdr.permission,
.mod_time = inode.hdr.mod_time,
};
const data: DataExtractor = switch (inode.data) {
.file => |f| blk: {
var data: DataExtractor = .init(cache, block_size, f.size, f.data_start, f.blocks);
if (f.frag_idx != 0xFFFFFFFF) {
const entry: Lookup.FragmentEntry = try frag_table.get(io, f.frag_idx);
if (entry.size.uncompressed) {
data.addFragment(cache.map.memory[entry.start..][0..entry.size.size], f.frag_offset);
} else {
const block = try cache.get(io, entry.start, entry.size.size, block_size);
data.addFragment(block, f.frag_offset);
}
}
break :blk data;
},
.ext_file => |f| blk: {
var data: DataExtractor = .init(cache, block_size, f.size, f.data_start, f.blocks);
if (f.frag_idx != 0xFFFFFFFF) {
const entry: Lookup.FragmentEntry = try frag_table.get(io, f.frag_idx);
if (entry.size.uncompressed) {
data.addFragment(cache.map.memory[entry.start..][0..entry.size.size], f.frag_offset);
} else {
const block = try cache.get(io, entry.start, entry.size.size, block_size);
data.addFragment(block, f.frag_offset);
}
}
if (f.xattr_idx != 0xFFFFFFFF)
ret.xattr_idx = f.xattr_idx;
break :blk data;
},
else => unreachable,
};
try data.asyncExtract(io, atomic.file);
try atomic.link(io);
return ret;
}
// Loop
fn returnLoop(alloc: std.mem.Allocator, sel: *Io.Select(ReturnUnion), options: ExtractionOptions) !void {
while (true) {
const finished = try sel.await();
switch (finished) {
.dir_ret => |d| {
const ret = try d;
if (ret.sub_files.load(.unordered) != 0) {
sel.queue.putOne(sel.io, .{ .dir_ret = ret }) catch |err| {
if (!ret.origin) alloc.free(ret.path);
return err;
};
continue;
}
if (!ret.origin) alloc.free(ret.path);
alloc.destroy(ret.sub_files);
if (!options.ignore_permissions and !options.ignore_xattr) {
// TODO: set permissions & xattr.
}
},
.file_ret => |f| {
const ret = try f;
if (!ret.origin) alloc.free(ret.path);
if (!options.ignore_permissions and !options.ignore_xattr) {
// TODO: set permissions & xattr.
}
},
.void_ret => |v| try v,
}
if (sel.group.token.load(.unordered) == null) break;
}
}
// Utility types
const ReturnUnion = union(enum) {
file_ret: Error!FileReturn,
dir_ret: Error!DirReturn,
void_ret: Error!void,
};
const Error = error{Canceled} || Directory.Error || Io.Dir.CreateFileAtomicError || Io.File.Atomic.LinkError || DataExtractor.Error;
const FileReturn = struct {
path: []const u8,
origin: bool,
uid_idx: u32,
gid_idx: u32,
mod_time: u32,
permissions: u16,
xattr_idx: ?u32 = null,
};
const DirReturn = struct {
path: []const u8,
sub_files: *Atomic(usize),
origin: bool,
uid_idx: u32,
gid_idx: u32,
mod_time: u32,
permissions: u16,
xattr_idx: ?u32 = null,
};
+102
View File
@@ -0,0 +1,102 @@
const std = @import("std");
const Io = std.Io;
const Archive = @import("archive.zig");
const Directory = @import("directory.zig");
const ExtractionOptions = @import("options.zig");
const Inode = @import("inode.zig");
const SfsFile = @This();
alloc: std.mem.Allocator,
archive: *Archive,
inode: Inode,
name: []const u8,
/// The given allocator must have been used to create the Inode and name.
pub fn init(alloc: std.mem.Allocator, archive: *Archive, inode: Inode, name: []const u8) SfsFile {
return .{
.alloc = alloc,
.archive = archive,
.inode = inode,
.name = name,
};
}
pub fn initDirEntry(alloc: std.mem.Allocator, io: Io, archive: *Archive, entry: Directory.Entry) !SfsFile {
const new_name = try alloc.alloc(u8, entry.name.len);
defer alloc.free(new_name);
@memcpy(new_name, entry.name);
return .{
.alloc = alloc,
.archive = archive,
.inode = try .initDirEntry(
alloc,
io,
&archive.cache,
archive.super.inode_start,
archive.super.block_size,
entry,
),
.name = new_name,
};
}
/// Creates a new copy of the given SfsFile using the given allocator
pub fn copy(self: SfsFile, alloc: std.mem.Allocator) !SfsFile {
const new_name = try alloc(u8, self.name.len);
errdefer alloc.free(new_name);
return .{
.alloc = alloc,
.archive = self.archive,
.inode = try self.inode.copy(alloc),
.name = new_name,
};
}
pub fn deinit(self: SfsFile) void {
self.inode.deinit(self.alloc);
self.alloc.free(self.name);
}
/// Attempts to open the filepath if the SfsFile is a directory.
/// If the given path refers to itself (such as "" or "."), a copied SfsFile is returned.
pub fn open(self: SfsFile, alloc: std.mem.Allocator, io: Io, filepath: []const u8) !SfsFile {
const path = std.mem.trim(u8, filepath, "/");
const first_element: []const u8 = std.mem.sliceTo(path, '/');
const dir: Directory = try self.inode.directory(alloc, io, &self.archive.cache, self.archive.super.dir_start);
defer dir.deinit(alloc);
var cur_slice = dir.entries;
var idx: usize = undefined;
while (cur_slice.len > 0) {
idx = cur_slice.len / 2;
switch (std.mem.order(u8, first_element, cur_slice[idx].name)) {
.eq => break,
.lt => cur_slice = cur_slice[0..idx],
.gt => cur_slice = cur_slice[idx..],
}
} else {
return error.NotFound;
}
if (first_element.len == path.len) return .initDirEntry(alloc, io, self.archive, cur_slice[idx]);
if (cur_slice[idx].type != .dir) return error.NotFound;
const tmp_file: SfsFile = try .initDirEntry(alloc, io, self.archive, cur_slice[idx]);
defer tmp_file.deinit();
return tmp_file.open(alloc, io, path[first_element.len..]);
}
pub fn extract(self: SfsFile, alloc: std.mem.Allocator, io: Io, ext_dir: []const u8, options: ExtractionOptions) !void {
_ = self;
_ = alloc;
_ = io;
_ = ext_dir;
_ = options;
return error.TODO;
}
+398
View File
@@ -0,0 +1,398 @@
const std = @import("std");
const Io = std.Io;
const Reader = Io.Reader;
const DecompCache = @import("decomp_cache.zig");
const Directory = @import("directory.zig");
const MetadataReader = @import("meta_rdr.zig");
const Inode = @This();
hdr: Header,
data: Data,
/// Read an inode given an inode Ref.
pub fn initRef(alloc: std.mem.Allocator, io: Io, cache: *DecompCache, inode_start: u64, block_size: u32, ref: Ref) !Inode {
var meta: MetadataReader = .init(io, cache, inode_start + ref.block_start);
defer meta.deinit(io);
try meta.interface.discardAll(ref.block_offset);
return .init(alloc, &meta.interface, block_size);
}
pub fn initDirEntry(alloc: std.mem.Allocator, io: Io, cache: *DecompCache, inode_start: u64, block_size: u32, entry: Directory.Entry) !Inode {
var meta: MetadataReader = .init(io, cache, inode_start + entry.block_start);
defer meta.deinit(io);
try meta.interface.discardAll(entry.block_offset);
return .init(alloc, &meta.interface, block_size);
}
/// Read the inode from the given Reader.
pub fn init(alloc: std.mem.Allocator, rdr: *Reader, block_size: u32) !Inode {
var hdr: Header = undefined;
try rdr.readSliceEndian(Header, @ptrCast(&hdr), .little);
const data: Data = switch (hdr.type) {
.dir => .{ .dir = try .init(rdr) },
.file => .{ .file = try .init(alloc, rdr, block_size) },
.symlink => .{ .symlink = try .init(alloc, rdr) },
.block_dev => .{ .block_dev = try .init(rdr) },
.char_dev => .{ .char_dev = try .init(rdr) },
.fifo => .{ .fifo = try .init(rdr) },
.socket => .{ .socket = try .init(rdr) },
.ext_dir => .{ .ext_dir = try .init(rdr) },
.ext_file => .{ .ext_file = try .init(alloc, rdr, block_size) },
.ext_symlink => .{ .ext_symlink = try .init(alloc, rdr) },
.ext_block_dev => .{ .ext_block_dev = try .init(rdr) },
.ext_char_dev => .{ .ext_char_dev = try .init(rdr) },
.ext_fifo => .{ .ext_fifo = try .init(rdr) },
.ext_socket => .{ .ext_socket = try .init(rdr) },
};
return .{
.hdr = hdr,
.data = data,
};
}
pub fn copy(self: Inode, alloc: std.mem.Allocator) !Inode {
var new_inode = self;
switch (new_inode.data) {
.file => |*f| {
if (f.blocks.len > 0) {
f.blocks = try alloc.alloc(DataBlock, f.blocks.len);
@memcpy(f.blocks, self.data.file.blocks);
}
},
.ext_file => |*f| {
if (f.blocks.len > 0) {
f.blocks = try alloc.alloc(DataBlock, f.blocks.len);
@memcpy(f.blocks, self.data.ext_file.blocks);
}
},
.symlink => |*s| {
s.target = try alloc.alloc(u8, s.target.len);
@memcpy(s.target, self.data.symlink.target);
},
.ext_symlink => |*s| {
s.target = try alloc.alloc(u8, s.target.len);
@memcpy(s.target, self.data.ext_symlink.target);
},
}
return new_inode;
}
pub fn deinit(self: Inode, alloc: std.mem.Allocator) void {
switch (self.data) {
.file => |f| f.deinit(alloc),
.ext_file => |f| f.deinit(alloc),
.symlink => |s| s.deinit(alloc),
.ext_symlink => |s| s.deinit(alloc),
else => {},
}
}
// Utility functions
pub fn directory(self: Inode, alloc: std.mem.Allocator, io: Io, cache: *DecompCache, dir_start: u64) !Directory {
return switch (self.data) {
.dir => |d| readDirectory(alloc, io, cache, dir_start, d),
.ext_dir => |d| readDirectory(alloc, io, cache, dir_start, d),
else => error.NotDirectory,
};
}
fn readDirectory(alloc: std.mem.Allocator, io: Io, cache: *DecompCache, dir_start: u64, d: anytype) !Directory {
var meta: MetadataReader = .init(io, cache, dir_start + d.block_start);
defer meta.deinit(io);
try meta.interface.discardAll(d.block_offset);
return .init(alloc, &meta.interface, d.size);
}
// Types
pub const Ref = packed struct(u64) {
block_offset: u16,
block_start: u32,
_: u16,
};
pub const Enum = enum(u16) {
dir = 1,
file,
symlink,
block_dev,
char_dev,
fifo,
socket,
ext_dir,
ext_file,
ext_symlink,
ext_block_dev,
ext_char_dev,
ext_fifo,
ext_socket,
};
pub const Header = extern struct {
type: Enum,
permission: u16,
uid_idx: u16,
gid_idx: u16,
mod_time: u32,
num: u32,
};
pub const Data = union(Enum) {
dir: Dir,
file: File,
symlink: Symlink,
block_dev: Device,
char_dev: Device,
fifo: IPC,
socket: IPC,
ext_dir: ExtDir,
ext_file: ExtFile,
ext_symlink: ExtSymlink,
ext_block_dev: ExtDevice,
ext_char_dev: ExtDevice,
ext_fifo: ExtIPC,
ext_socket: ExtIPC,
};
pub const DataBlock = packed struct(u32) {
size: u24,
uncompressed: bool,
_: u7,
};
const Dir = extern struct {
block_start: u32,
hard_links: u32,
size: u16,
block_offset: u16,
parent: u32,
const Self = @This();
fn init(rdr: *Reader) !Self {
var dir: Self = undefined;
try rdr.readSliceEndian(Self, @ptrCast(&dir), .little);
return dir;
}
};
const ExtDir = extern struct {
hard_links: u32,
size: u32,
block_start: u32,
parent: u32,
idx_count: u16,
block_offset: u16,
xattr_idx: u32,
// []DirIndex
const Self = @This();
fn init(rdr: *Reader) !Self {
var dir: Self = undefined;
try rdr.readSliceEndian(Self, @ptrCast(&dir), .little);
return dir;
}
};
const File = struct {
data_start: u32,
frag_idx: u32,
frag_offset: u32,
size: u32,
blocks: []DataBlock,
const Raw = extern struct {
data_start: u32,
frag_idx: u32,
frag_offset: u32,
size: u32,
};
const Self = @This();
pub fn init(alloc: std.mem.Allocator, rdr: *Reader, block_size: u32) !Self {
var raw: Raw = undefined;
try rdr.readSliceEndian(Raw, @ptrCast(&raw), .little);
var blocks_num = raw.size / block_size;
if (raw.frag_idx == 0xFFFFFFFF and raw.size % block_size > 0)
blocks_num += 1;
const blocks: []DataBlock = try alloc.alloc(DataBlock, blocks_num);
errdefer alloc.free(blocks);
try rdr.readSliceEndian(DataBlock, blocks, .little);
return .{
.data_start = raw.data_start,
.frag_idx = raw.frag_idx,
.frag_offset = raw.frag_offset,
.size = raw.size,
.blocks = blocks,
};
}
pub fn deinit(self: File, alloc: std.mem.Allocator) void {
alloc.free(self.blocks);
}
};
const ExtFile = struct {
data_start: u64,
size: u64,
sparse: u64,
hard_links: u32,
frag_idx: u32,
frag_offset: u32,
xattr_idx: u32,
blocks: []DataBlock,
const Raw = extern struct {
data_start: u64,
size: u64,
sparse: u64,
hard_links: u32,
frag_idx: u32,
frag_offset: u32,
xattr_idx: u32,
};
const Self = @This();
pub fn init(alloc: std.mem.Allocator, rdr: *Reader, block_size: u32) !Self {
var raw: Raw = undefined;
try rdr.readSliceEndian(Raw, @ptrCast(&raw), .little);
var blocks_num = raw.size / block_size;
if (raw.frag_idx == 0xFFFFFFFF and raw.size % block_size > 0)
blocks_num += 1;
const blocks: []DataBlock = try alloc.alloc(DataBlock, blocks_num);
errdefer alloc.free(blocks);
try rdr.readSliceEndian(DataBlock, blocks, .little);
return .{
.data_start = raw.data_start,
.size = raw.size,
.sparse = raw.sparse,
.hard_links = raw.hard_links,
.frag_idx = raw.frag_idx,
.frag_offset = raw.frag_offset,
.xattr_idx = raw.xattr_idx,
.blocks = blocks,
};
}
pub fn deinit(self: ExtFile, alloc: std.mem.Allocator) void {
alloc.free(self.blocks);
}
};
const Symlink = struct {
hard_links: u32,
target: []const u8,
const Raw = extern struct {
hard_links: u32,
target_size: u32,
};
const Self = @This();
pub fn init(alloc: std.mem.Allocator, rdr: *Reader) !Self {
var raw: Raw = undefined;
try rdr.readSliceEndian(Raw, @ptrCast(&raw), .little);
const target = try alloc.alloc(u8, raw.target_size);
try rdr.readSliceEndian(u8, target, .little);
return .{
.hard_links = raw.hard_links,
.target = target,
};
}
pub fn deinit(self: Symlink, alloc: std.mem.Allocator) void {
alloc.free(self.target);
}
};
const ExtSymlink = struct {
hard_links: u32,
xattr_idx: u32,
target: []const u8,
const Raw = extern struct {
hard_links: u32,
target_size: u32,
};
const Self = @This();
pub fn init(alloc: std.mem.Allocator, rdr: *Reader) !Self {
var raw: Raw = undefined;
try rdr.readSliceEndian(Raw, @ptrCast(&raw), .little);
const target = try alloc.alloc(u8, raw.target_size);
errdefer alloc.free(target);
try rdr.readSliceEndian(u8, target, .little);
var xattr_idx: u32 = undefined;
try rdr.readSliceEndian(u32, @ptrCast(&xattr_idx), .little);
return .{
.hard_links = raw.hard_links,
.target = target,
.xattr_idx = xattr_idx,
};
}
pub fn deinit(self: ExtSymlink, alloc: std.mem.Allocator) void {
alloc.free(self.target);
}
};
const Device = extern struct {
hard_links: u32,
device: u32,
const Self = @This();
fn init(rdr: *Reader) !Self {
var dir: Self = undefined;
try rdr.readSliceEndian(Self, @ptrCast(&dir), .little);
return dir;
}
};
const ExtDevice = extern struct {
hard_links: u32,
device: u32,
xattr_idx: u32,
const Self = @This();
fn init(rdr: *Reader) !Self {
var dir: Self = undefined;
try rdr.readSliceEndian(Self, @ptrCast(&dir), .little);
return dir;
}
};
const IPC = extern struct {
hard_links: u32,
const Self = @This();
fn init(rdr: *Reader) !Self {
var dir: Self = undefined;
try rdr.readSliceEndian(Self, @ptrCast(&dir), .little);
return dir;
}
};
const ExtIPC = extern struct {
hard_links: u32,
xattr_idx: u32,
const Self = @This();
fn init(rdr: *Reader) !Self {
var dir: Self = undefined;
try rdr.readSliceEndian(Self, @ptrCast(&dir), .little);
return dir;
}
};
+114
View File
@@ -0,0 +1,114 @@
const std = @import("std");
const Io = std.Io;
const DataBlock = @import("inode.zig").DataBlock;
const InodeRef = @import("inode.zig").Ref;
const DecompCache = @import("decomp_cache.zig");
const MetadataReader = @import("meta_rdr.zig");
pub fn stateless(comptime T: anytype, io: Io, cache: *DecompCache, table_start: u64, idx: u32) !T {
const PER_BLOCK = 8192 / @sizeOf(T);
const block = idx / PER_BLOCK;
const block_idx = idx % PER_BLOCK;
const offset_offset = table_start + (block * 8);
const offset: u64 = std.mem.readInt(u64, cache.map.memory[offset_offset..][0..2], .little);
var meta: MetadataReader = .init(io, cache, offset);
defer meta.deinit(io);
try meta.discardAll(block_idx * @sizeOf(T));
var new: T = undefined;
try meta.interface.readSliceEndian(T, @ptrCast(&new), .little);
return new;
}
pub fn Table(comptime T: anytype) type {
return struct {
const PER_BLOCK = 8192 / @sizeOf(T);
const LookupTable = @This();
alloc: std.mem.Allocator,
cache: *DecompCache,
table_start: u64,
num: u32,
values: std.AutoHashMap(u32, []T),
mut: Io.RwLock = .init,
pub fn init(alloc: std.mem.Allocator, cache: *DecompCache, table_start: u64, num_values: u32) LookupTable {
return .{
.alloc = alloc,
.cache = cache,
.table_start = table_start,
.num = num_values,
.values = .init(alloc),
};
}
pub fn deinit(self: *LookupTable) void {
var iter = self.values.valueIterator();
while (iter.next()) |v|
self.alloc.free(v);
self.values.deinit();
}
pub fn get(self: *LookupTable, io: Io, idx: u32) Error!T {
const block = idx / PER_BLOCK;
const block_idx = idx % PER_BLOCK;
{
try self.mut.lockShared(io);
defer self.mut.unlockShared(io);
const val = self.values.get(block);
if (val != null) return val.*[block_idx];
}
try self.mut.lock(io);
defer self.mut.unlock(io);
const val = try self.values.getOrPut(block);
if (val.found_existing)
return val.value_ptr.*[block_idx];
errdefer self.values.removeByPtr(val.key_ptr);
const offset_offset = self.table_start + (block * 8);
const offset: u64 = std.mem.readInt(u64, self.cache.map.memory[offset_offset..][0..2], .little);
var meta: MetadataReader = .init(io, self.cache, offset);
defer meta.deinit(io);
const size = if (block == ((self.num - 1) / PER_BLOCK))
self.num % PER_BLOCK
else
PER_BLOCK;
const new_block = try self.alloc.alloc(T, size);
errdefer self.alloc.free(new_block);
try meta.interface.readSliceEndian(T, new_block, .little);
val.value_ptr.* = new_block;
return new_block[block_idx];
}
};
}
// Types
pub const Error = error{} || std.mem.Allocator.Error;
pub const FragmentEntry = extern struct {
start: u64,
size: DataBlock,
_: u32,
};
pub const XattrEntry = extern struct {
ref: InodeRef,
count: u32,
size: u32,
};
+109
View File
@@ -0,0 +1,109 @@
const std = @import("std");
const Io = std.Io;
const Reader = Io.Reader;
const Writer = Io.Writer;
const Limit = Io.Limit;
const DecompCache = @import("decomp_cache.zig");
const MetadataReader = @This();
io: Io,
cache: *DecompCache,
cur_offset: u64 = 0,
next_offset: u64,
interface: Reader = .{
.buffer = &[0]u8{},
.end = 0,
.seek = 0,
.vtable = &.{
.stream = stream,
.discard = discard,
.readVec = readVec,
},
},
pub fn init(io: Io, cache: *DecompCache, start: u64) MetadataReader {
return .{
.io = io,
.cache = cache,
.next_offset = start,
};
}
pub fn deinit(self: *MetadataReader, io: Io) void {
self.cache.finished(io, self.cur_offset);
}
fn advance(self: *MetadataReader) !void {
self.cache.finished(self.io, self.cur_offset);
self.interface.seek = 0;
errdefer self.interface.end = 0;
const hdr: Header = @bitCast(std.mem.readInt(u16, self.cache.map.memory[self.next_offset..][0..2], .little));
self.cur_offset = self.next_offset + 2;
self.next_offset = self.cur_offset + hdr.size;
if (hdr.uncompressed) {
self.interface.buffer = self.cache.map.memory[self.cur_offset..][0..hdr.size];
self.interface.end = hdr.size;
return;
}
self.interface.buffer = try self.cache.get(self.io, self.cur_offset, hdr.size, 8192);
self.interface.end = self.interface.buffer.len;
}
fn stream(r: *Reader, w: *Writer, limit: Limit) Reader.StreamError!usize {
if (r.seek >= r.end) {
const self: *MetadataReader = @fieldParentPtr("interface", r);
self.advance() catch |err| {
std.debug.print("error advancing metadata reader: {}\n", .{err});
return Reader.Error.ReadFailed;
};
}
const to_write = @min(r.end - r.seek, @intFromEnum(limit));
const wrote = try w.write(r.buffer[r.seek..][0..to_write]);
r.seek += wrote;
return wrote;
}
fn discard(r: *Reader, limit: Limit) Reader.Error!usize {
if (r.seek >= r.end) {
const self: *MetadataReader = @fieldParentPtr("interface", r);
self.advance() catch |err| {
std.debug.print("error advancing metadata reader: {}\n", .{err});
return Reader.Error.ReadFailed;
};
}
const to_discard = @min(r.end - r.seek, @intFromEnum(limit));
r.seek += to_discard;
return to_discard;
}
fn readVec(r: *Reader, vec: [][]u8) Reader.Error!usize {
if (r.seek >= r.end) {
const self: *MetadataReader = @fieldParentPtr("interface", r);
self.advance() catch |err| {
std.debug.print("error advancing metadata reader: {}\n", .{err});
return Reader.Error.ReadFailed;
};
}
var total: usize = 0;
for (vec) |v| {
const to_copy = @min(r.end - r.seek, v.len);
@memcpy(v[0..to_copy], r.buffer[r.seek..][0..to_copy]);
r.seek += to_copy;
total += to_copy;
if (r.seek >= r.end)
break;
}
return total;
}
// Types
const Header = packed struct(u16) {
size: u15,
uncompressed: bool,
};
+25
View File
@@ -0,0 +1,25 @@
//! Options for file/directory extraction.
const std = @import("std");
const Writer = std.Io.Writer;
const ExtractionOptions = @This();
/// Don't set the file's owner & permissions after extraction
ignore_permissions: bool = false,
/// Don't set xattr values. Currently xattrs are never set anyway.
ignore_xattr: bool = false,
/// Replace symlinks with their target.
dereference_symlinks: bool = false,
/// Verbose logging. If true, verbose_writer must be set
verbose: bool = false,
/// Where to print verbose log.
verbose_writer: ?*Writer = null,
pub const default: ExtractionOptions = .{};
pub fn VerboseDefault(wrt: *Writer) !ExtractionOptions {
return .{
.verbose = true,
.verbose_writer = wrt,
};
}
+8
View File
@@ -0,0 +1,8 @@
pub const Archive = @import("archive.zig");
pub const ExtractionOptions = @import("options.zig");
test {
const std = @import("std");
std.testing.refAllDecls(Archive);
}
+84
View File
@@ -0,0 +1,84 @@
const std = @import("std");
const stuff = @import("builtin");
const Archive = @import("archive.zig");
const Superblock = @import("super.zig").Superblock;
const TestArchive = "testing/LinuxPATest.sfs";
test "Basics" {
var fil = try std.fs.cwd().openFile(TestArchive, .{});
defer fil.close();
var sfs: Archive = try .init(std.testing.allocator, fil);
defer sfs.deinit();
if (sfs.super != LinuxPATestCorrectSuperblock) {
std.debug.print("Superblock wrong\nShould be: {}\n\nis: {}\n", .{ LinuxPATestCorrectSuperblock, sfs.super });
return error.BadSuperblock;
}
}
const TestFile = "Start.exe";
const TestFileExtractLocation = "testing/Start.exe";
test "ExtractSingleFile" {
std.fs.cwd().deleteFile(TestFileExtractLocation) catch {};
var fil = try std.fs.cwd().openFile(TestArchive, .{});
defer fil.close();
var sfs: Archive = try .init(std.testing.allocator, fil);
defer sfs.deinit();
var test_fil = try sfs.open(TestFile);
defer test_fil.deinit();
try test_fil.extract(TestFileExtractLocation, .Default);
//TODO: validate extracted file.
}
const TestFullExtractLocation = "testing/TestExtract";
test "ExtractCompleteArchive" {
std.fs.cwd().deleteTree(TestFullExtractLocation) catch {};
var fil = try std.fs.cwd().openFile(TestArchive, .{});
defer fil.close();
var sfs: Archive = try .init(std.testing.allocator, fil);
defer sfs.deinit();
try sfs.extract(TestFullExtractLocation, .Default);
}
const LinuxPATestCorrectSuperblock: Superblock = .{
.magic = std.mem.readInt(u32, "hsqs", .little),
.inode_count = 2974,
.mod_time = 1632696724,
.block_size = 131072,
.frag_count = 264,
.compression = .zstd,
.block_log = 17,
.flags = .{
.inode_uncompressed = false,
.data_uncompressed = false,
.check = false,
.frag_uncompressed = false,
.fragment_never = false,
.fragment_always = false,
.duplicates = true,
.exportable = true,
.xattr_uncompressed = false,
.xattr_never = false,
.compression_options = false,
.ids_uncompressed = false,
._ = 0,
},
.id_count = 1,
.ver_maj = 4,
.ver_min = 0,
.root_ref = .{
.block_offset = 1363,
.block_start = 29237,
._ = 0,
},
.size = 106841744,
.id_start = 106841632,
.xattr_start = 106841720,
.inode_start = 106778274,
.dir_start = 106807998,
.frag_start = 106837747,
.export_start = 106841602,
};
+40
View File
@@ -0,0 +1,40 @@
const std = @import("std");
const Reader = std.Io.Reader;
const flate = std.compress.flate;
const zstd = std.compress.zstd;
const xz = std.compress.xz;
const lzma = std.compress.lzma;
const Error = @import("decomp.zig").Error;
pub fn zlibDecompress(_: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var buf: [flate.max_window_len]u8 = undefined;
var rdr: Reader = .fixed(in);
var decomp: flate.Decompress = .init(&rdr, .zlib, &buf);
return decomp.reader.readSliceShort(out);
}
pub fn zstdDecompress(alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
const buf = try alloc.alloc(u8, in.len + zstd.block_size_max);
defer alloc.free(buf);
var rdr: Reader = .fixed(in);
var decomp: zstd.Decompress = .init(&rdr, buf, .{ .window_len = in.len });
return decomp.reader.readSliceShort(out);
}
pub fn lzmaDecompress(alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var rdr: Reader = .fixed(in);
var decomp: lzma.Decompress = .initOptions(&rdr, alloc, &[0]u8{}, .{}, 2 * out.len);
defer decomp.deinit();
return decomp.reader.readSliceShort(out);
}
pub fn xzDecompress(alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var rdr: Reader = .fixed(in);
var decomp: xz.Decompress = .init(&rdr, alloc, &[0]u8{});
defer decomp.deinit();
return decomp.reader.readSliceShort(out);
}