71 Commits

Author SHA1 Message Date
Caleb J. Gardner 451e53264d Basic extraction complete 2025-08-16 01:25:38 -05:00
Caleb J. Gardner 99e3b81100 Re-writing extraction (again) 2025-08-15 06:48:55 -05:00
Caleb J. Gardner 1eceb8d899 Futher tweaks & fixes 2025-08-08 06:34:31 -05:00
Caleb Gardner b50f28026f STUFF 2025-08-07 21:45:03 -05:00
Caleb Gardner 1269d3e30d More work on data reader 2025-08-07 20:47:04 -05:00
Caleb J. Gardner b3f4a02b72 More work on the data decompilation 2025-08-01 20:04:09 -05:00
Caleb Gardner ebaacf57a3 MORE for data reader (only a bit) 2025-07-31 12:58:35 -05:00
Caleb Gardner 13f92f2e83 Re-writing data reader. Again. 2025-07-30 20:54:11 -05:00
Caleb Gardner 3fb95dd3fa Started work on data reader 2025-07-29 20:23:27 -05:00
Caleb J. Gardner 9871b0b2c0 Started re-work of data reader 2025-07-29 06:43:28 -05:00
Caleb Gardner 61d194e80a Moving data reader to arrays instead of slices 2025-07-25 06:42:22 -05:00
Caleb Gardner 8c44c77456 Merge pull request #1 from CalebQ42/reset-4
Reset 4
2025-07-25 06:07:55 -05:00
Caleb Gardner a96ad46a6c Things are working (though slowly) 2025-07-22 05:13:34 -05:00
Caleb Gardner d5c50b19f2 Further progress on extraction 2025-07-21 06:40:18 -05:00
Caleb Gardner 8998d28253 Data reader fixes
Re-started extraction logic
2025-07-19 16:40:54 -05:00
Caleb Gardner 9c1d90f60b Reset on extraction logic
Fixed not using MetadataReader for Tables
Fixed MetadataReaders not respecting uncompressed blocks
2025-07-19 06:30:24 -05:00
Caleb Gardner eb214feefa Further progress on extraction 2025-07-18 22:30:30 -05:00
Caleb Gardner de988f083f Finished (?) file extraction 2025-07-18 05:41:27 -05:00
Caleb Gardner b4af1233e5 Started work on file extraction 2025-07-17 22:17:58 -05:00
Caleb Gardner 4d52627d5d Further tweaks & fixes 2025-07-17 09:16:32 -05:00
Caleb Gardner d6b136bc8f Added basic test
Fixed various bugs
2025-07-17 03:50:09 -05:00
Caleb Gardner 87563e43a5 Finished (?) data reader 2025-07-17 02:24:57 -05:00
Caleb Gardner b0dced90bc MORE PROGRESS 2025-07-14 01:40:04 -05:00
Caleb Gardner 69d90242ba Further progress
Mainly focusing on getting things more consistent
2025-07-12 02:13:41 -05:00
Caleb Gardner 61c86c9fea Laying down some framework 2025-07-11 06:45:51 -05:00
Caleb Gardner 23687eabb0 Further progress in parsing format 2025-07-11 04:32:02 -05:00
Caleb Gardner 5c14b7db48 Try 4, lol. 2025-07-09 06:42:02 -05:00
Caleb Gardner 60e183512b (some) listing for zig-unsquashfs 2025-05-31 22:26:20 -05:00
Caleb Gardner fd1f83d855 Added uid & gid 2025-05-31 21:57:18 -05:00
Caleb Gardner d48ed4259e Added parent_path to Files
Should allow for (future) better path parsing
2025-05-31 21:49:09 -05:00
Caleb Gardner 8cc576a7fd Remove random .continue folder 2025-05-29 00:10:23 -05:00
Caleb Gardner 10304139e4 Fix DataReader
Don't use DataExtractor when set to only use 1 thread
2025-05-28 21:17:04 -05:00
Caleb Gardner 4af3e0373e Init data reader for Files
Fixed File.Size()
Don't use extractor for small files
2025-05-28 20:47:28 -05:00
Caleb Gardner 5be59be220 Increase buffer size for decompressTo 2025-05-28 02:49:53 -05:00
Caleb Gardner f122d1b4be Fixed the final, lingering issues.
Updated README
2025-05-28 00:57:12 -05:00
Caleb Gardner e4a6c32528 So close yet so far away 2025-05-27 22:26:58 -05:00
Caleb Gardner c057099591 Fragment data is currently broken, starting to fix 2025-05-27 16:17:20 -05:00
Caleb Gardner 17dbda3326 Nearly there...
Various tweaks & fixes to get extraction working properly
2025-05-27 14:21:40 -05:00
Caleb Gardner 985e2bd7e5 Extraction?
There's still a problem with fragments...
2025-05-26 07:06:03 -05:00
Caleb Gardner 7a4105bebd Added Zig-unsquashfs arg parsing.
Tweaks to build, including the ability to specifiy version
2025-05-26 04:05:53 -05:00
Caleb Gardner b0ecbe16bd Small tweaks 2025-05-25 08:22:33 -05:00
Caleb Gardner c4e2dab3f7 Fixed block/char/fifo types 2025-05-25 06:33:03 -05:00
Caleb Gardner dd452060cb Several fixes for File.extract
Tweaks and fixes.
2025-05-25 05:34:32 -05:00
Caleb Gardner 82011a092c Small tweaks and fixes 2025-05-24 23:30:23 -05:00
Caleb Gardner 3bfd262824 Tweak build 2025-05-24 14:35:53 -05:00
Caleb Gardner 6f02f9f14d Start work on File data extraction 2025-05-23 03:58:26 -05:00
Caleb Gardner 66f6cfa069 Added File.reader() 2025-05-22 06:27:38 -05:00
Caleb Gardner af06021b1b Added some doc 2025-05-21 21:38:05 -05:00
Caleb Gardner 213dfa8b92 Started on zig-unquashfs 2025-05-21 20:50:09 -05:00
Caleb Gardner e91d75458e Fixed table reading 2025-05-21 20:35:45 -05:00
Caleb Gardner 41a6b0d6f3 Mostly finished data reader 2025-05-20 09:15:00 -05:00
Caleb Gardner 128ed9f001 Started work on data readers 2025-05-19 09:49:23 -05:00
Caleb Gardner 1150b0d427 Fixed early file.deinit causing memory leaks 2025-05-19 06:08:43 -05:00
Caleb Gardner 1b0a0221c4 Now just for the memory leaks 2025-05-17 16:16:54 -05:00
Caleb Gardner 9f345e5fdb Almost there? 2025-05-17 14:33:16 -05:00
Caleb Gardner f77c2ecf48 Making things work properly 2025-05-17 09:05:24 -05:00
Caleb Gardner bbf3539dcf Messing around with thing.
Things not working.
2025-05-16 15:16:46 -05:00
Caleb Gardner 43295fb823 Finished? inodes 2025-05-16 11:02:59 -05:00
Caleb Gardner ff2ef6feaa INODES! 2025-05-16 06:59:21 -05:00
Caleb Gardner 986f308c60 Start over to clear my head 2025-05-16 05:55:40 -05:00
Caleb Gardner fc068fdbd9 Tweaking to get stuff workign 2025-05-15 05:59:38 -05:00
Caleb Gardner e010763fc6 Directory table parsing 2025-05-14 22:23:17 -05:00
Caleb Gardner 5daffdafc7 Proper zig project organization 2025-05-14 11:46:09 -05:00
Caleb Gardner 6dd3054006 Started work on File 2025-05-14 06:59:56 -05:00
Caleb Gardner b0c71c59f8 Inode parsing works! 2025-05-14 05:16:12 -05:00
Caleb Gardner 3684a958a0 Almost kind of working 2025-05-14 01:59:19 -05:00
Caleb Gardner a866804853 A bit more work on metadata reader 2025-05-13 22:29:45 -05:00
Caleb Gardner 246d63d48a A bit more messing around. 2025-05-13 10:24:17 -05:00
Caleb Gardner b4848de95d Further work on inode decoding 2025-05-13 00:23:53 -05:00
Caleb Gardner b3a5ff8f94 Starting work on inode decoding 2025-05-12 13:46:46 -05:00
Caleb Gardner 58e89c0981 Initial commit 2025-05-11 09:39:24 -05:00
45 changed files with 1617 additions and 2415 deletions
-31
View File
@@ -1,31 +0,0 @@
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
-1
View File
@@ -2,4 +2,3 @@ testing/
.zig-cache/ .zig-cache/
zig-out/ zig-out/
zig-pkg/
-25
View File
@@ -1,25 +0,0 @@
// 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) 2026 Caleb Gardner Copyright (c) 2025 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
+3 -71
View File
@@ -1,75 +1,7 @@
# zig-squashfs # zig-squashfs
This is my experiments to learn Zig. Might amount to something. Might not. Messing around with zig via making a squashfs library. May amount to something. Or not.
A library and application to decompress or view squashfs archives. ## Current state
## Current State Performance is reatively bad (when compared to the official [squashfs-tools](https://github.com/plougher/squashfs-tools), but the basics should fully work.
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;
```
+35 -58
View File
@@ -1,76 +1,53 @@
const std = @import("std"); const std = @import("std");
/// version if version isn't provided during build
const def_version = "0.0.0+testing";
pub fn build(b: *std.Build) !void { 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 opt = b.addOptions();
// const allow_lzo = b.option(bool, "allow_lzo", "Compile with lzo support") orelse false; const ver = b.option([]const u8, "version", "sematic version") orelse def_version;
const debug = b.option(bool, "debug", "Enable options to make debugging easier."); const sem_ver = try std.SemanticVersion.parse(ver);
const version_string_option = b.option([]const u8, "version", "Version of the library/binary"); opt.addOption(std.SemanticVersion, "version", sem_ver);
// 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 target = b.standardTargetOptions(.{}); const target = b.standardTargetOptions(.{});
const optimize = b.standardOptimizeOption(.{}); const optimize = b.standardOptimizeOption(.{});
const lib_mod = b.createModule(.{
.root_source_file = b.path("src/root.zig"),
.target = target,
.optimize = optimize,
});
const lib = b.addLibrary(.{ const lib = b.addLibrary(.{
.name = "squashfs", .linkage = .static,
.root_module = b.createModule(.{ .name = "zig_squashfs",
.optimize = if (debug == true) .Debug else optimize, .root_module = lib_mod,
.target = target, .version = sem_ver,
.valgrind = debug,
.root_source_file = b.path("src/root.zig"),
}),
.use_llvm = debug,
}); });
var version = version_string_option orelse "0.0.0-testing"; const exe_mod = b.createModule(.{
if (version[0] == 'v') version = version[1..]; .root_source_file = b.path("src/bin/unsquashfs.zig"),
const unsquashfs_options = b.addOptions(); .target = target,
unsquashfs_options.addOption( .optimize = optimize,
std.SemanticVersion, });
"version", exe_mod.addImport("squashfs", lib_mod);
try std.SemanticVersion.parse(version), exe_mod.addOptions("config", opt);
);
const exe = b.addExecutable(.{ const exe = b.addExecutable(.{
.linkage = .static,
.name = "unsquashfs", .name = "unsquashfs",
.root_module = b.createModule(.{ .root_module = exe_mod,
.optimize = if (debug == true) .Debug else optimize, .version = sem_ver,
.target = target,
.valgrind = debug,
.root_source_file = b.path("src/bin/unsquashfs.zig"),
.imports = &.{
.{ .name = "zig_squashfs", .module = lib.root_module },
},
}),
.use_llvm = debug,
}); });
exe.root_module.addOptions("config", unsquashfs_options);
b.installArtifact(lib); b.installArtifact(lib);
b.installArtifact(exe); b.installArtifact(exe);
const mod_tests = b.addTest(.{ const lib_unit_tests = b.addTest(.{
.root_module = b.createModule(.{ .root_module = lib_mod,
.optimize = optimize,
.target = target,
.root_source_file = b.path("src/test.zig"),
}),
}); });
const run_mod_tests = b.addRunArtifact(mod_tests); const exe_unit_test = b.addTest(.{
const test_step = b.step("test", "Run tests"); .root_module = exe_mod,
test_step.dependOn(&run_mod_tests.step);
// zls build check steps
const lib_check = b.addLibrary(.{
.name = "squashfs",
.root_module = exe.root_module,
}); });
const exe_check = b.addExecutable(.{ const run_lib_unit_tests = b.addRunArtifact(lib_unit_tests);
.name = "unsquashfs", const run_exe_unit_tests = b.addRunArtifact(exe_unit_test);
.root_module = lib.root_module, const test_step = b.step("test", "Run unit tests");
}); test_step.dependOn(&run_lib_unit_tests.step);
const check = b.step("check", "Check if unsquashfs compiles"); test_step.dependOn(&run_exe_unit_tests.step);
check.dependOn(&lib_check.step);
check.dependOn(&exe_check.step);
} }
+49 -20
View File
@@ -1,26 +1,55 @@
.{ .{
.name = .squashfs, .name = .zig_squashfs,
.version = "0.0.6", .version = "0.0.1",
.fingerprint = 0x37ba29474b87f145, // Changing this has security and trust implications. .fingerprint = 0x527960c72c03ffe3, // Changing this has security and trust implications.
.minimum_zig_version = "0.15.2",
.minimum_zig_version = "0.14.0",
// This field is optional.
// Each dependency must either provide a `url` and `hash`, or a `path`.
// `zig build --fetch` can be used to fetch all dependencies of a package, recursively.
// Once all dependencies are fetched, `zig build` no longer requires
// internet connectivity.
.dependencies = .{ .dependencies = .{
.zlib_ng = .{ // See `zig fetch --save <url>` for a command-line interface for adding dependencies.
.url = "git+https://github.com/CalebQ42/zig-zlib-ng#5f2f02dfb28acca2517dacbbd09e9b987f57b133", //.example = .{
.hash = "zlib_ng-2.3.3-pre1-2HYS4ClFAABW8KlHMyBHtlNKE3V7kCS8wqfxawG7xeaa", // // When updating this field to a new URL, be sure to delete the corresponding
}, // // `hash`, otherwise you are communicating that you expect to find the old hash at
.zstd = .{ // // the new URL. If the contents of a URL change this will result in a hash mismatch
.url = "git+https://github.com/allyourcodebase/zstd.git?ref=1.5.7-1#e1a501be57f42c541e8a5597e4b59a074dfd09a3", // // which will prevent zig from using it.
.hash = "zstd-1.5.7-1-KEItkAMwAAD6OKY3m0OOmXG7aL-aLUfrDqbP5J5oYapU", // .url = "https://example.com/foo.tar.gz",
}, //
.lz4 = .{ // // This is computed from the file contents of the directory of files that is
.url = "git+https://github.com/allyourcodebase/lz4.git?ref=1.10.0-6#41f52ab227caf9d48cf88c89a4d2946caa12b102", // // obtained after fetching `url` and applying the inclusion rules given by
.hash = "lz4-1.10.0-6-ewyzw-4NAAAWDpY4xpiqr4LQhZQAC0x_rGnW2iPh6jk2", // // `paths`.
}, // //
.minilzo = .{ // // This field is the source of truth; packages do not come from a `url`; they
.url = "git+https://github.com/CalebQ42/zig-minilzo.git#7cbae997b91a44d74b7cd6c073584dc9562a6c90", // // come from a `hash`. `url` is just one of many possible mirrors for how to
.hash = "minilzo-2.10.0-Ij7BO8wLAADeWI4Pe4jp8XTDsDaquZR14oZ7_9yKKDWP", // // obtain a package matching this `hash`.
}, // //
// // Uses the [multihash](https://multiformats.io/multihash/) format.
// .hash = "...",
//
// // When this is provided, the package is found in a directory relative to the
// // build root. In this case the package's hash is irrelevant and therefore not
// // computed. This field and `url` are mutually exclusive.
// .path = "foo",
//
// // When this is set to `true`, a package is declared to be lazily
// // fetched. This makes the dependency only get fetched if it is
// // actually used.
// .lazy = false,
//},
}, },
// Specifies the set of files and directories that are included in this package.
// Only files and directories listed here are included in the `hash` that
// is computed for this package. Only files listed here will remain on disk
// when using the zig package manager. As a rule of thumb, one should list
// files required for compilation plus any license(s).
// Paths are relative to the build root. Use the empty string (`""`) to refer to
// the build root itself.
// A directory listed here means that all files within, recursively, are included.
.paths = .{ .paths = .{
"build.zig", "build.zig",
"build.zig.zon", "build.zig.zon",
-10
View File
@@ -1,10 +0,0 @@
#!/bin/sh
zig test \
-lc \
-lz \
-llzma \
-lminilzo \
-llz4 \
-lzstd \
src/test.zig
-181
View File
@@ -1,181 +0,0 @@
const std = @import("std");
const Io = std.Io;
const ExtractionOptions = @import("options.zig");
const File = @import("file.zig");
const Inode = @import("inode.zig");
const LookupTable = @import("lookup_table.zig");
const Decompressor = @import("util/decompressor.zig");
const MetadataReader = @import("util/metadata.zig");
const Utils = @import("util/misc.zig");
const OffsetFile = @import("util/offset_file.zig");
const Archive = @This();
file: OffsetFile,
super: Superblock,
stateless_decomp: Decompressor,
pub fn init(io: Io, file: std.Io.File, offset: u64) !Archive {
var rdr = file.reader(io, &[0]u8{});
try rdr.seekTo(offset);
var super: Superblock = undefined;
try rdr.interface.readSliceEndian(Superblock, @ptrCast(&super), .little);
return .{
.file = .init(file, offset),
.super = super,
.stateless_decomp = switch (super.compression) {
.gzip => @import("decomp/zlib.zig").stateless_decompressor,
.lzma => @import("decomp/lzma.zig").stateless_decompressor,
.lzo => return error.LzoUnsupported,
.xz => @import("decomp/xz.zig").stateless_decompressor,
.lz4 => return error.Lz4Unsupported,
.zstd => @import("decomp/zstd.zig").stateless_decompressor,
},
};
}
/// The root folder of the Archive. Used to open other Files.
pub fn root(self: Archive, alloc: std.mem.Allocator, io: Io) !File {
const root_inode = try Utils.inodeFromRef(
alloc,
io,
self.file,
&self.stateless_decomp,
self.super.inode_start,
self.super.block_size,
self.super.root_ref,
);
return .init(alloc, self, root_inode, "");
}
/// Opens a File within the archive.
pub fn open(self: Archive, alloc: std.mem.Allocator, io: Io, filepath: []const u8) !File {
const root_file = try self.root(alloc, io);
const path = std.mem.trim(u8, filepath, "/");
if (Utils.pathIsSelf(path))
return root_file;
defer root_file.deinit();
return root_file.open(alloc, io, filepath);
}
/// Extract the entire archive contents to the given directory.
pub fn extract(self: Archive, alloc: std.mem.Allocator, io: Io, extract_dir: []const u8, options: ExtractionOptions) !void {
const root_inode = try Utils.inodeFromRef(
alloc,
io,
self.file,
&self.stateless_decomp,
self.super.inode_start,
self.super.block_size,
self.super.root_ref,
);
_ = root_inode;
_ = extract_dir;
_ = options;
return error.TODO;
}
/// Returns the inode with the given inode number.
/// Requires that the archive is exportable (has an export lookup table).
pub fn inode(self: Archive, alloc: std.mem.Allocator, io: Io, num: u32) !Inode {
if (!self.super.flags.exportable)
return error.NotExportable;
const ref = try LookupTable.lookupValue(
Inode.Ref,
alloc,
io,
&self.stateless_decomp,
self.file,
self.super.export_start,
num + 1,
);
return Utils.inodeFromRef(
alloc,
io,
self.file,
&self.stateless_decomp,
self.super.inode_start,
self.super.block_size,
ref,
);
}
/// Returns a value at the given index from the Archive's id (uid/gid) table.
pub fn idTable(self: Archive, alloc: std.mem.Allocator, io: Io, idx: u32) !u16 {
return LookupTable.lookupValue(
u16,
alloc,
io,
&self.stateless_decomp,
self.file,
self.super.id_start,
idx,
);
}
// Superblock
const SQUASHFS_MAGIC: u32 = std.mem.readInt(u32, "hsqs", .little);
const SuperblockError = error{
InvalidMagic,
InvalidBlockLog,
InvalidVersion,
InvalidCheck,
};
/// A squashfs Superblock
pub const Superblock = extern struct {
magic: u32,
inode_count: u32,
mod_time: u32,
block_size: u32,
frag_count: u32,
compression: enum(u16) {
gzip = 1,
lzma,
lzo,
xz,
lz4,
zstd,
},
block_log: u16,
flags: packed struct(u16) {
inode_uncompressed: bool,
data_uncompressed: bool,
check: bool,
frag_uncompressed: bool,
fragment_never: bool,
fragment_always: bool,
duplicates: bool,
exportable: bool,
xattr_uncompressed: bool,
xattr_never: bool,
compression_options: bool,
ids_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,
/// Validate the Superblock. If an error is returned, it's likely the archive is corrupted or not a squashfs archive.
pub fn validate(self: Superblock) !void {
if (self.magic != SQUASHFS_MAGIC)
return SuperblockError.InvalidMagic;
if (self.flags.check)
return SuperblockError.InvalidCheck;
if (self.ver_maj != 4 or self.ver_min != 0)
return SuperblockError.InvalidVersion;
if (std.math.log2(self.block_size) != self.block_log)
return SuperblockError.InvalidBlockLog;
}
};
+124 -112
View File
@@ -1,134 +1,146 @@
const std = @import("std"); const std = @import("std");
const Io = std.Io;
const Writer = Io.Writer;
const builtin = @import("builtin");
const config = @import("config"); const config = @import("config");
const squashfs = @import("zig_squashfs"); const squashfs = @import("squashfs");
//TODO: Add more options const help_msg =
const help_mgs = \\Basic Usage: zig-unsquashfs [Options] SQUASHFS_FILE <EXTRACT_LOCATION>
\\ \\
\\Usage: unsquashfs [options] <archive> \\General options:
\\ -e <path> Path to a file or directory inside the archive to extract instead of the whole archive.
\\ Can be given multiple times.
\\ -o <bytes> Skip <bytes> before reading from the archive.
\\ -v Verbose output.
\\ \\
\\Options: \\Extraction options:
\\ -d <location> Extract to the given location instead of "squashfs-root" \\ --unbreak-symlinks Attempt extract symlink targets along with symlinks. Will not place files outside of the extraction location.
\\ -us Same as --unbreak-symlinks
\\ --deref-symlinks Replace symlink files with their target.
\\ -ds Same as --deref-symlinks
\\ -p <#> Use at most # of processors. Defaults to logical core count.
\\ \\
\\ -o <offset> Start reading the archive at the given offset. \\Listing Options:
\\ -dx Don't set xattr values \\ -l List files instead of extracting. When used, you do not need to specify an extraction location.
\\ -dp Don't set permissions (includes setting uid & gid owner) \\ -ll Similiar to -l, but with file attributes.
\\ -lln Similiar to -ll, but with numeric uids and gids.
\\ \\
\\ -p <threads> Specify how many threads to use. If no present or zero, the system's logical cores count is used. \\Other:
\\ -v Verbose \\ --help Prints this help message.
\\ \\ -h Same as --help
\\ --force Force extraction. If the destination already exists, it will be deleted. \\ --version Print version number.
\\
\\ --help Display this messages
\\ --version Display the version
\\ \\
; ;
const errors = error{InvalidArguments}; const stdout = std.io.getStdOut();
var archive: []const u8 = ""; var extr_files: std.ArrayList([]const u8) = undefined;
var extLoc: []const u8 = "squashfs-root";
var offset: u64 = 0; var offset: u64 = 0;
var threads: u32 = 0;
var verbose: bool = false; var verbose: bool = false;
var ignore_xattrs: bool = false; var unbreak: bool = false;
var ignore_permissions: bool = false; var deref: bool = false;
var force: bool = false; var processors: u16 = 0;
var list: ListTypes = .None;
pub fn main(init: std.process.Init) !void { var filename: []const u8 = "";
const alloc = init.gpa; var extr_location: []const u8 = "";
const io = init.io;
var stdout = std.Io.File.stdout(); const ListTypes = enum {
var out = stdout.writer(io, &[0]u8{}); None,
defer out.interface.flush() catch {}; List,
try handleArgs(init.minimal.args, &out.interface); ListAttr,
if (archive.len == 0) { ListNumeric,
try out.interface.print("You must provide a squashfs archive\n", .{}); };
try out.interface.print(help_mgs, .{});
pub fn main() !void {
const alloc = std.heap.smp_allocator;
extr_files = .init(alloc);
defer extr_files.deinit();
var args = std.process.argsWithAllocator(alloc) catch {
_ = try stdout.writeAll("Unable to allocate memory");
return; return;
}
var fil = try Io.Dir.cwd().openFile(io, archive, .{}); //TODO: Handle error gracefully.
defer fil.close(io);
var arc: squashfs.Archive = try .init(io, fil, offset); //TODO: Update when memory size matters. //TODO: Handle error gracefully.
const options: squashfs.ExtractionOptions = .{
.threads = if (threads == 0) try std.Thread.getCpuCount() else threads,
.verbose = verbose,
.verbose_writer = if (verbose) &out.interface else null,
.ignore_xattr = ignore_xattrs,
.ignore_permissions = ignore_permissions,
}; };
if (force) defer args.deinit();
try Io.Dir.cwd().deleteTree(io, extLoc); _ = args.next();
try arc.extract(alloc, io, extLoc, options); //TODO: Handle error gracefully. while (args.next()) |arg| {
} if (std.mem.eql(u8, arg, "--help") or std.mem.eql(u8, arg, "-h")) {
_ = try stdout.writeAll(help_msg);
fn handleArgs(args: std.process.Args, out: *Writer) !void { return;
var arg_iter = args.iterate(); } else if (std.mem.eql(u8, arg, "--version")) {
defer arg_iter.deinit(); try config.version.format("", .{}, stdout.writer());
_ = arg_iter.next(); // args[0] is the application launch command. _ = try stdout.write("\n");
while (arg_iter.next()) |arg| { return;
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")) { } else if (std.mem.eql(u8, arg, "-v")) {
verbose = true; verbose = true;
continue; } else if (std.mem.eql(u8, arg, "--unbreak-symlinks") or std.mem.eql(u8, arg, "-us")) {
} else if (std.mem.eql(u8, arg, "-dx")) { unbreak = true;
ignore_xattrs = true; } else if (std.mem.eql(u8, arg, "--deref-symlinks") or std.mem.eql(u8, arg, "-ds")) {
continue; deref = true;
} else if (std.mem.eql(u8, arg, "-dp")) { } else if (std.mem.eql(u8, arg, "-l")) {
ignore_permissions = true; list = .List;
continue; } else if (std.mem.eql(u8, arg, "-ll")) {
} else if (std.mem.eql(u8, arg, "--force")) { list = .ListAttr;
force = true; } else if (std.mem.eql(u8, arg, "-lln")) {
continue; list = .ListNumeric;
} else if (std.mem.eql(u8, arg, "--version")) { } else if (std.mem.eql(u8, arg, "-e")) {
try out.print("zig-unsquashfs v", .{}); const next = args.next();
try config.version.format(out); if (next == null) {
try out.print("\nBuilt using Zig {s} in {} mode\n", .{ builtin.zig_version_string, builtin.mode }); _ = try stdout.writeAll("path required after -e\n");
std.process.exit(0); return;
return; }
} else if (std.mem.eql(u8, arg, "--help")) { try extr_files.append(next.?);
try out.print(help_mgs, .{}); } else if (std.mem.eql(u8, arg, "-o")) {
std.process.exit(0); const next = args.next();
if (next == null) {
_ = try stdout.writeAll("offset required after -o\n");
return;
}
offset = try std.fmt.parseInt(u64, next.?, 10);
} else if (std.mem.eql(u8, arg, "-p")) {
const next = args.next();
if (next == null) {
_ = try stdout.writeAll("number required after -p\n");
return;
}
processors = try std.fmt.parseInt(u16, next.?, 10);
} else if (filename.len == 0) {
filename = arg;
} else if (extr_location.len == 0) {
extr_location = arg;
} else {
_ = try stdout.writeAll("invalid or too many arguments\n");
return; 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;
} }
if (filename.len == 0) {
_ = try stdout.writeAll("no archive given\n");
return;
}
if (list == .None and extr_location.len == 0) {
_ = try stdout.writeAll("no extract location given\n");
return;
}
const fil = try std.fs.cwd().openFile(filename, .{});
defer fil.close();
var th_alloc: std.heap.ThreadSafeAllocator = .{ .child_allocator = std.heap.smp_allocator };
var rdr = squashfs.SfsFile.init(
th_alloc.allocator(),
fil,
offset,
) catch |err| {
try std.fmt.format(stdout.writer(), "Error opening {s} as squashfs: {any}\n", .{ filename, err });
return;
};
defer rdr.deinit();
//TODO: list and extr_files;
var op: squashfs.ExtractionOptions = squashfs.ExtractionOptions.init() catch |err| {
try std.fmt.format(stdout.writer(), "Error setting extraction options: {any}\n", .{err});
return;
};
op.verbose = verbose;
op.dereference_symlinks = deref;
op.unbreak_symlinks = unbreak;
if (processors != 0) op.thread_count = processors;
rdr.extract(op, extr_location) catch |err| {
try std.fmt.format(stdout.writer(), "Error extracting archive: {any}\n", .{err});
return;
};
} }
-77
View File
@@ -1,77 +0,0 @@
const std = @import("std");
const Reader = std.Io.Reader;
const lzma = std.compress.lzma;
const Node = std.SinglyLinkedList.Node;
const Decompressor = @import("../util/decompressor.zig");
const Error = Decompressor.Error;
const Self = @This();
const Buffer = struct {
node: Node,
buf: []u8,
};
interface: Decompressor = .{ .decomp_fn = decomp },
alloc: std.mem.Allocator,
block_size: u32,
buffers: std.ArrayList(Buffer),
buffer_queue: std.SinglyLinkedList,
pub fn init(alloc: std.mem.Allocator, block_size: u32) !Self {
return .{
.alloc = alloc,
.block_size = block_size,
.buffers = try .initCapacity(alloc, 5),
};
}
pub fn deinit(self: Self) void {
for (self.buffers) |buf|
self.alloc.free(buf);
self.buffers.deinit(self.alloc);
}
fn decomp(d: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
if (d == null) {
const buf = try alloc.alloc(u8, in.len * 2);
defer alloc.free(buf);
return lzmaDecomp(buf, in, out);
}
var self: Self = @fieldParentPtr("interface", d.?);
const buf_node = self.buffer_queue.popFirst();
var buf: *Buffer = undefined;
if (buf_node == null) {
const new_buf = try self.buffers.addOne(self.alloc);
new_buf.* = .{ .{}, try self.alloc.alloc(u8, self.block_size + lzma.block_size_max) };
buf = new_buf;
} else {
buf = @fieldParentPtr("node", buf_node);
}
defer self.buffer_queue.prepend(&buf.node);
return lzmaDecomp(self.alloc, &buf.buf, in, out);
}
inline fn lzmaDecomp(alloc: std.mem.Allocator, buffer: *[]u8, in: []u8, out: []u8) !usize {
var rdr: Reader = .fixed(in);
var d = try lzma.Decompress.initOptions(&rdr, alloc, buffer.*, .{ .allow_incomplete = true }, 3 * 1024 * 1024);
defer {
buffer.* = d.takeBuffer();
d.deinit();
}
return d.reader.readSliceShort(out);
}
// Stateless
pub const stateless_decompressor: Decompressor = .{ .decomp_fn = statelessDecomp };
fn statelessDecomp(_: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var buf = try alloc.alloc(u8, in.len);
defer alloc.free(buf);
return lzmaDecomp(alloc, &buf, in, out) catch return Error.ReadFailed;
}
-77
View File
@@ -1,77 +0,0 @@
const std = @import("std");
const Reader = std.Io.Reader;
const xz = std.compress.xz;
const Node = std.SinglyLinkedList.Node;
const Decompressor = @import("../util/decompressor.zig");
const Error = Decompressor.Error;
const Self = @This();
const Buffer = struct {
node: Node,
buf: []u8,
};
interface: Decompressor = .{ .decomp_fn = decomp },
alloc: std.mem.Allocator,
block_size: u32,
buffers: std.ArrayList(Buffer),
buffer_queue: std.SinglyLinkedList,
pub fn init(alloc: std.mem.Allocator, block_size: u32) !Self {
return .{
.alloc = alloc,
.block_size = block_size,
.buffers = try .initCapacity(alloc, 5),
};
}
pub fn deinit(self: Self) void {
for (self.buffers) |buf|
self.alloc.free(buf);
self.buffers.deinit(self.alloc);
}
fn decomp(d: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
if (d == null) {
const buf = try alloc.alloc(u8, in.len * 2);
defer alloc.free(buf);
return lzmaDecomp(buf, in, out);
}
var self: Self = @fieldParentPtr("interface", d.?);
const buf_node = self.buffer_queue.popFirst();
var buf: *Buffer = undefined;
if (buf_node == null) {
const new_buf = try self.buffers.addOne(self.alloc);
new_buf.* = .{ .{}, try self.alloc.alloc(u8, self.block_size + xz.block_size_max) };
buf = new_buf;
} else {
buf = @fieldParentPtr("node", buf_node);
}
defer self.buffer_queue.prepend(&buf.node);
return lzmaDecomp(self.alloc, &buf.buf, in, out);
}
inline fn lzmaDecomp(alloc: std.mem.Allocator, buffer: *[]u8, in: []u8, out: []u8) !usize {
var rdr: Reader = .fixed(in);
var d = try xz.Decompress.init(&rdr, alloc, buffer.*);
defer {
buffer.* = d.takeBuffer();
d.deinit();
}
return d.reader.readSliceShort(out);
}
// Stateless
pub const stateless_decompressor: Decompressor = .{ .decomp_fn = statelessDecomp };
fn statelessDecomp(_: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
var buf = try alloc.alloc(u8, in.len);
defer alloc.free(buf);
return lzmaDecomp(alloc, &buf, in, out) catch return Error.ReadFailed;
}
-73
View File
@@ -1,73 +0,0 @@
const std = @import("std");
const Reader = std.Io.Reader;
const flate = std.compress.flate;
const Node = std.SinglyLinkedList.Node;
const Decompressor = @import("../util/decompressor.zig");
const Error = Decompressor.Error;
const Self = @This();
const Buffer = struct {
node: Node,
buf: []u8,
};
interface: Decompressor = .{ .decomp_fn = decomp },
alloc: std.mem.Allocator,
block_size: u32,
buffers: std.ArrayList(Buffer),
buffer_queue: std.SinglyLinkedList,
pub fn init(alloc: std.mem.Allocator, block_size: u32) !Self {
return .{
.alloc = alloc,
.block_size = block_size,
.buffers = try .initCapacity(alloc, 5),
};
}
pub fn deinit(self: Self) void {
for (self.buffers) |buf|
self.alloc.free(buf);
self.buffers.deinit(self.alloc);
}
fn decomp(d: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
if (d == null) {
const buf = try alloc.alloc(u8, in.len * 2);
defer alloc.free(buf);
return zlibDecomp(buf, in, out);
}
var self: Self = @fieldParentPtr("interface", d.?);
const buf_node = self.buffer_queue.popFirst();
var buf: *Buffer = undefined;
if (buf_node == null) {
const new_buf = try self.buffers.addOne(self.alloc);
new_buf.* = .{ .{}, try self.alloc.alloc(u8, self.block_size) };
buf = new_buf;
} else {
buf = @fieldParentPtr("node", buf_node);
}
defer self.buffer_queue.prepend(&buf.node);
return zlibDecomp(buf.buf, in, out);
}
inline fn zlibDecomp(buffer: []u8, in: []u8, out: []u8) !usize {
var rdr: Reader = .fixed(in);
var d = flate.Decompress.init(&rdr, .zlib, buffer);
return d.reader.readSliceShort(out);
}
// Stateless
pub const stateless_decompressor: Decompressor = .{ .decomp_fn = statelessDecomp };
fn statelessDecomp(_: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
const buf = try alloc.alloc(u8, in.len * 2);
defer alloc.free(buf);
return zlibDecomp(buf, in, out);
}
-73
View File
@@ -1,73 +0,0 @@
const std = @import("std");
const Reader = std.Io.Reader;
const zstd = std.compress.zstd;
const Node = std.SinglyLinkedList.Node;
const Decompressor = @import("../util/decompressor.zig");
const Error = Decompressor.Error;
const Self = @This();
const Buffer = struct {
node: Node,
buf: []u8,
};
interface: Decompressor = .{ .decomp_fn = decomp },
alloc: std.mem.Allocator,
block_size: u32,
buffers: std.ArrayList(Buffer),
buffer_queue: std.SinglyLinkedList,
pub fn init(alloc: std.mem.Allocator, block_size: u32) !Self {
return .{
.alloc = alloc,
.block_size = block_size,
.buffers = try .initCapacity(alloc, 5),
};
}
pub fn deinit(self: Self) void {
for (self.buffers) |buf|
self.alloc.free(buf);
self.buffers.deinit(self.alloc);
}
fn decomp(d: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
if (d == null) {
const buf = try alloc.alloc(u8, in.len * 2);
defer alloc.free(buf);
return zstdDecomp(buf, in, out);
}
var self: Self = @fieldParentPtr("interface", d.?);
const buf_node = self.buffer_queue.popFirst();
var buf: *Buffer = undefined;
if (buf_node == null) {
const new_buf = try self.buffers.addOne(self.alloc);
new_buf.* = .{ .{}, try self.alloc.alloc(u8, self.block_size + zstd.block_size_max) };
buf = new_buf;
} else {
buf = @fieldParentPtr("node", buf_node);
}
defer self.buffer_queue.prepend(&buf.node);
return zstdDecomp(buf.buf, in, out);
}
inline fn zstdDecomp(buffer: []u8, in: []u8, out: []u8) !usize {
var rdr: Reader = .fixed(in);
var d = zstd.Decompress.init(&rdr, buffer, .{ .window_len = @truncate(out.len) });
return d.reader.readSliceShort(out);
}
// Stateless
pub const stateless_decompressor: Decompressor = .{ .decomp_fn = statelessDecomp };
fn statelessDecomp(_: ?*const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
const buf = try alloc.alloc(u8, out.len + zstd.block_size_max);
defer alloc.free(buf);
return zstdDecomp(buf, in, out);
}
+61 -58
View File
@@ -1,67 +1,70 @@
const std = @import("std"); const std = @import("std");
const Reader = std.Io.Reader;
const Inode = @import("inode.zig"); const InodeType = @import("inode.zig").Type;
const Compression = @import("superblock.zig").Compression;
const DirEntry = @This(); const Header = extern struct { //use extern instead of packed, due to bit alignment
block_start: u32,
block_offset: u16,
type: Inode.Type,
name: []const u8,
pub fn deinit(self: DirEntry, alloc: std.mem.Allocator) void {
alloc.free(self.name);
}
pub fn readDirectory(alloc: std.mem.Allocator, rdr: *Reader, size: u32) ![]DirEntry {
var hdr: Header = undefined;
var raw: RawEntry = undefined;
var out: std.ArrayList(DirEntry) = try .initCapacity(alloc, 30);
errdefer {
for (out.items) |ent|
alloc.free(ent.name);
out.deinit(alloc);
}
var tot_red: u32 = 3;
while (tot_red < size) {
try rdr.readSliceEndian(Header, @ptrCast(&hdr), .little);
try out.ensureUnusedCapacity(alloc, hdr.count + 1);
tot_red += @sizeOf(Header);
for (hdr.count + 1) |_| {
try rdr.readSliceEndian(RawEntry, @ptrCast(&raw), .little);
const new_name = try alloc.alloc(u8, raw.name_size + 1);
try rdr.readSliceEndian(u8, new_name, .little);
const new = out.addOneAssumeCapacity();
new.* = .{
.block_start = hdr.block_start,
.block_offset = raw.block_offset,
.type = raw.type,
.name = new_name,
};
tot_red += @sizeOf(RawEntry) + raw.name_size + 1;
}
}
return out.toOwnedSlice(alloc);
}
// Types
const Header = extern struct {
count: u32, count: u32,
block_start: u32, block: u32,
num: u32, num: u32,
}; };
const RawEntry = extern struct { const RawEntry = struct {
block_offset: u16, offset: u16,
num_offset: i16, num_offset: i16,
type: Inode.Type, type: InodeType,
name_size: u16, size: u16,
name: []const u8,
pub fn init(alloc: std.mem.Allocator, rdr: anytype) !RawEntry {
var fixed: [8]u8 = undefined;
_ = try rdr.read(&fixed);
const size = std.mem.readInt(u16, fixed[6..8], .little);
const name = try alloc.alloc(u8, size + 1);
_ = try rdr.read(name);
return .{
.offset = std.mem.readInt(u16, fixed[0..2], .little),
.num_offset = std.mem.readInt(i16, fixed[2..4], .little),
.type = @enumFromInt(std.mem.readInt(u16, fixed[4..6], .little)),
.size = size,
.name = name,
};
}
}; };
pub const Entry = struct {
block: u32,
offset: u16,
num: u32,
type: InodeType,
name: []const u8,
pub fn deinit(self: Entry, alloc: std.mem.Allocator) void {
alloc.free(self.name);
}
};
pub fn readDirectory(alloc: std.mem.Allocator, rdr: anytype, size: u32) ![]Entry {
var entries: std.ArrayList(Entry) = .init(alloc);
errdefer entries.deinit();
var cur_red: u32 = 3; // dir size includes "." & "..", so its actual size is off by 3.
var hdr: Header = undefined;
while (cur_red < size) {
_ = try rdr.read(std.mem.asBytes(&hdr));
cur_red += 12;
try entries.ensureUnusedCapacity(hdr.count + 1);
for (0..hdr.count + 1) |_| {
const raw_ent: RawEntry = try .init(alloc, rdr);
cur_red += 9 + raw_ent.size;
errdefer alloc.free(raw_ent.name);
entries.appendAssumeCapacity(.{
.block = hdr.block,
.offset = raw_ent.offset,
.num = @truncate(@abs(@as(i64, hdr.num) + raw_ent.num_offset)),
.type = raw_ent.type,
.name = raw_ent.name,
});
}
}
return entries.toOwnedSlice();
}
+23
View File
@@ -0,0 +1,23 @@
const std = @import("std");
const Self = @This();
/// Replace symlinks with their targets
dereference_symlinks: bool = false,
/// Always extract a symlink's target if it's part of the archive.
/// May result in the symlink's target being changed.
unbreak_symlinks: bool = false,
/// Do not set file's permissions & owner when extracted.
ignore_permissions: bool = false,
/// Verbose logging
verbose: bool = false,
/// Verbose logging writer. If not set, stdout is used.
verbose_logger: std.io.AnyWriter = std.io.getStdOut().writer().any(),
/// Number of threads used during extraction. Defualts to std.Thread.getCpuCount().
thread_count: usize,
pub fn init() !Self {
return .{
.thread_count = try std.Thread.getCpuCount(),
};
}
+341 -110
View File
@@ -1,124 +1,355 @@
//! An easier to use wrapper around an inode.
const std = @import("std"); const std = @import("std");
const Io = std.Io; const builtin = @import("builtin");
const Archive = @import("archive.zig"); const dir = @import("directory.zig");
const DirEntry = @import("directory.zig");
const ExtractionOptions = @import("options.zig"); const DirEntry = dir.Entry;
const Inode = @import("inode.zig"); const Inode = @import("inode.zig");
const DataExtractor = @import("util/data_extractor.zig"); const SfsReader = @import("reader.zig").SfsReader;
const Decompressor = @import("util/decompressor.zig"); const ToReader = @import("reader/to_read.zig").ToRead;
const MetadataReader = @import("util/metadata.zig"); const ExtractionOptions = @import("extract_options.zig");
const SharedCache = @import("util/shared_cache.zig"); const DataReader = @import("reader/data.zig").DataReader;
const Compression = @import("superblock.zig").Compression;
const MetadataReader = @import("reader/metadata.zig").MetadataReader;
const File = @This(); pub fn File(comptime T: type) type {
return struct {
pub const FileError = error{
NotRegular,
NotDirectory,
NotFound,
};
alloc: std.mem.Allocator, const Self = @This();
archive: Archive, rdr: *SfsReader(T),
// parent: *File(T),
inode: Inode, inode: Inode,
name: []const u8, name: []const u8,
/// Creates a new File from an inode. Takes ownership of the Inode and creates a copy of the given name. /// Directory entries. Only populated on directories.
/// Requires the given allocator was used to create the Inode. entries: ?[]DirEntry = null,
pub fn init(alloc: std.mem.Allocator, archive: Archive, in: Inode, name: []const u8) !File { /// File reader. Only populated on regular files.
const new_name = try alloc.alloc(u8, name.len); data_reader: ?DataReader(T) = null,
@memcpy(new_name, name);
return .{
.alloc = alloc,
.archive = archive, pub fn init(rdr: *SfsReader(T), inode: Inode, name: []const u8) !Self {
const name_cpy: []u8 = try rdr.alloc.alloc(u8, name.len);
.inode = in, @memcpy(name_cpy, name);
.name = new_name, var out = Self{
}; .rdr = rdr,
} .inode = inode,
pub fn fromDirEntry(alloc: std.mem.Allocator, io: Io, archive: Archive, ent: DirEntry) !File { .name = name_cpy,
var rdr = try archive.file.readerAt(io, archive.super.inode_start + ent.block_start, &[0]u8{}); };
var meta: MetadataReader = .init(alloc, &rdr.interface, &archive.stateless_decomp); switch (inode.data) {
try meta.interface.discardAll(ent.block_offset); .dir => |d| {
var meta = MetadataReader(T).init(
var in: Inode = try .read(alloc, &meta.interface, archive.super.block_size); rdr.alloc,
errdefer in.deinit(alloc); rdr.super.comp,
return .init(alloc, archive, in, ent.name); rdr.rdr,
} d.block + rdr.super.dir_start,
pub fn deinit(self: File) void { );
self.alloc.free(self.name); try meta.skip(d.offset);
self.inode.deinit(self.alloc); out.entries = try dir.readDirectory(rdr.alloc, &meta, d.size);
} },
.ext_dir => |d| {
pub fn open(self: File, alloc: std.mem.Allocator, io: Io, filepath: []const u8) !File { var meta = MetadataReader(T).init(
const entries = try self.inode.readDirectory( rdr.alloc,
alloc, rdr.super.comp,
io, rdr.rdr,
self.archive.file, d.block + rdr.super.dir_start,
&self.archive.stateless_decomp, );
self.archive.super.dir_start, try meta.skip(d.offset);
); out.entries = try dir.readDirectory(rdr.alloc, &meta, d.size);
defer { },
for (entries) |ent| .file => |f| {
alloc.free(ent.name); out.data_reader = try .init(rdr, inode);
alloc.free(entries); _ = f;
} //TODO: fragments
const path = std.mem.trim(u8, filepath, "/"); // if (f.hasFragment()) {
const first_element: []u8 = std.mem.sliceTo(path, "/"); // try out.data_reader.?.addFragment(
// try rdr.frag_table.get(f.frag_idx),
var search_slice = entries; // f.frag_offset,
var idx: usize = undefined; // );
while (search_slice.len > 0) { // }
idx = search_slice / 2; },
const middle = search_slice[idx]; .ext_file => |f| {
switch (std.mem.order(u8, first_element, middle.name)) { out.data_reader = try .init(rdr, inode);
.eq => break, _ = f;
.lt => search_slice = search_slice[0..idx], //TODO: Fragments
.gt => search_slice = search_slice[idx + 1 ..], // if (f.hasFragment()) {
// try out.data_reader.?.addFragment(
// try rdr.frag_table.get(f.frag_idx),
// f.frag_offset,
// );
// }
},
else => {},
}
return out;
}
pub fn initFromRef(rdr: *SfsReader(T), ref: Inode.Ref, name: []const u8) !Self {
var meta: MetadataReader(T) = .init(rdr.alloc, rdr.super.comp, rdr.rdr, ref.block + rdr.super.inode_start);
try meta.skip(ref.offset);
const inode: Inode = try .init(&meta, rdr.alloc, rdr.super.block_size);
return .init(rdr, inode, name);
}
pub fn initFromEntry(rdr: *SfsReader(T), ent: DirEntry) !Self {
var meta: MetadataReader(T) = .init(rdr.alloc, rdr.super.comp, rdr.rdr, ent.block + rdr.super.inode_start);
try meta.skip(ent.offset);
const inode: Inode = try .init(&meta, rdr.alloc, rdr.super.block_size);
return .init(rdr, inode, ent.name);
}
pub fn deinit(self: *Self) void {
self.rdr.alloc.free(self.name);
self.inode.deinit(self.rdr.alloc);
if (self.entries != null) {
for (self.entries.?) |e| {
e.deinit(self.rdr.alloc);
}
self.rdr.alloc.free(self.entries.?);
}
if (self.data_reader != null) {
self.data_reader.?.deinit();
}
} }
} else return Error.FileNotFound;
const first_elem_file = try fromDirEntry(alloc, io, self.archive, search_slice[idx]); pub fn uid(self: Self) !u32 {
if (first_element.len == path.len) return self.rdr.id_table.get(self.inode.hdr.uid_idx);
return first_elem_file; }
defer first_elem_file.deinit(); pub fn gid(self: Self) !u32 {
return first_elem_file.open(alloc, io, path[first_element.len + 1 ..]); return self.rdr.id_table.get(self.inode.hdr.uid_idx);
} }
pub fn extract(self: File, alloc: std.mem.Allocator, io: Io, filepath: []const u8, options: ExtractionOptions) !void { const Reader = std.io.GenericReader(*DataReader(T), anyerror, DataReader(T).read);
var cache: SharedCache = try .init(alloc, 10); // TODO: calculate a good initial cache size.
defer cache.deinit(); pub fn read(self: *Self, buf: []u8) !usize {
var decomp = switch (self.archive.super.compression) { if (self.data_reader == null) return FileError.NotRegular;
.gzip => {}, return self.data_reader.?.read(buf);
.lzma => {}, }
.xz => {}, pub fn reader(self: *Self) !Reader {
.zstd => {}, if (self.data_reader == null) return FileError.NotRegular;
else => unreachable, return self.data_reader.?.reader();
}
pub fn open(self: Self, path: []const u8) !Self {
if (self.entries == null) return FileError.NotDirectory;
if (path.len == 0) return self;
const idx = std.mem.indexOf(u8, path, "/") orelse path.len;
if (idx == 0) return self.open(path[1..]);
const name = path[0..idx];
for (self.entries.?) |e| {
if (std.mem.eql(u8, e.name, name)) {
var fil: Self = try .initFromEntry(self.rdr, e);
if (idx >= path.len - 1) return fil;
defer fil.deinit();
return fil.open(path[idx + 1 ..]);
}
}
return FileError.NotFound;
}
pub fn iterate(self: Self) Iterator {
return .{
.rdr = self.rdr,
.entries = self.entries.?,
};
}
const Iterator = struct {
rdr: *SfsReader(T),
entries: []DirEntry,
idx: u32 = 0,
pub fn next(self: *Iterator) !?File(T) {
if (self.idx >= self.entries.len) return null;
const out = try Self.initFromEntry(self.rdr, self.entries[self.idx]);
self.idx += 1;
return out;
}
pub fn reset(self: *Iterator) void {
self.idx = 0;
}
};
const WaitGroup = std.Thread.WaitGroup;
const Pool = std.Thread.Pool;
const Mutex = std.Thread.Mutex;
pub const ExtractError = error{FileExists};
pub fn extract(self: *Self, op: ExtractionOptions, path: []const u8) !void {
var wg: WaitGroup = .{};
var pol: Pool = undefined;
try pol.init(.{
.n_jobs = op.thread_count,
.allocator = self.rdr.alloc,
});
defer pol.deinit();
var errs: std.ArrayList(anyerror) = .init(self.rdr.alloc);
defer errs.deinit();
try self.extractInode(op, &wg, &errs, &pol, self.inode, path);
wg.wait();
if (errs.items.len > 0) return errs.items[0];
}
fn extractInode(
self: *Self,
op: ExtractionOptions,
wg: *WaitGroup,
errs: *std.ArrayList(anyerror),
pol: *Pool,
inode: Inode,
path: []const u8,
) !void {
wg.start();
defer wg.finish(); //TODO: When everthing is threaded, this will need to be handled by the threads, not here.
switch (inode.hdr.type) {
.file, .ext_file => {
var fil = try std.fs.cwd().createFile(path, .{});
defer fil.close();
var data: DataReader(T) = try .init(self.rdr, inode);
defer data.deinit();
try data.writeTo(fil); // TODO: Thread
const fil_uid = self.rdr.id_table.get(inode.hdr.uid_idx) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error getting uid {} from table: {}\n", .{ inode.hdr.uid_idx, err }) catch {};
}
return;
};
const fil_gid = self.rdr.id_table.get(inode.hdr.gid_idx) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error getting gid {} from table: {}\n", .{ inode.hdr.gid_idx, err }) catch {};
}
return;
};
fil.chmod(inode.hdr.perm) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error chmod {s}: {}\n", .{ path, err }) catch {};
}
return;
};
fil.chown(fil_uid, fil_gid) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error chmod {s}: {}\n", .{ path, err }) catch {};
}
return;
};
//TODO: update mtime.
},
.dir, .ext_dir => {
std.fs.cwd().makeDir(path) catch |err| {
if (err != std.fs.Dir.MakeError.PathAlreadyExists) {
return err;
}
};
var dir_block: u32 = 0;
var dir_offset: u16 = 0;
var dir_size: u32 = 0;
switch (inode.data) {
.dir => |d| {
dir_block = d.block;
dir_offset = d.offset;
dir_size = d.size;
},
.ext_dir => |d| {
dir_block = d.block;
dir_offset = d.offset;
dir_size = d.size;
},
else => unreachable,
}
var meta: MetadataReader(T) = .init(self.rdr.alloc, self.rdr.super.comp, self.rdr.rdr, dir_block + self.rdr.super.dir_start);
try meta.skip(dir_offset);
const entries = try dir.readDirectory(self.rdr.alloc, &meta, dir_size);
defer self.rdr.alloc.free(entries);
for (entries) |ent| {
defer ent.deinit(self.rdr.alloc);
var new_path: []u8 = undefined;
if (path[path.len - 1] == '/') {
new_path = self.rdr.alloc.alloc(u8, path.len + ent.name.len) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error allocating memory: {}\n", .{err}) catch {};
}
errs.append(err) catch {};
continue;
};
@memcpy(new_path[0..path.len], path);
@memcpy(new_path[path.len..], ent.name);
} else {
new_path = self.rdr.alloc.alloc(u8, path.len + ent.name.len + 1) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error allocating memory: {}\n", .{err}) catch {};
}
errs.append(err) catch {};
continue;
};
@memcpy(new_path[0..path.len], path);
new_path[path.len] = '/';
@memcpy(new_path[path.len + 1 ..], ent.name);
}
defer self.rdr.alloc.free(new_path);
meta = .init(self.rdr.alloc, self.rdr.super.comp, self.rdr.rdr, ent.block + self.rdr.super.inode_start);
meta.skip(ent.offset) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error reading inode: {}\n", .{err}) catch {};
}
errs.append(err) catch {};
continue;
};
const new_inode = Inode.init(&meta, self.rdr.alloc, self.rdr.super.block_size) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error reading inode: {}\n", .{err}) catch {};
}
errs.append(err) catch {};
continue;
};
defer new_inode.deinit(self.rdr.alloc);
self.extractInode(op, wg, errs, pol, new_inode, new_path) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error extracting {s}: {}\n", .{ new_path, err }) catch {};
}
errs.append(err) catch {};
continue;
};
}
var fil = std.fs.cwd().openDir(path, .{ .iterate = true }) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error openning {s} to set permissions: {}\n", .{ path, err }) catch {};
}
return;
};
const fil_uid = self.rdr.id_table.get(inode.hdr.uid_idx) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error getting uid {} from table: {}\n", .{ inode.hdr.uid_idx, err }) catch {};
}
return;
};
const fil_gid = self.rdr.id_table.get(inode.hdr.gid_idx) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error getting gid {} from table: {}\n", .{ inode.hdr.gid_idx, err }) catch {};
}
return;
};
fil.chmod(inode.hdr.perm) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error chmod {s}: {}\n", .{ path, err }) catch {};
}
return;
};
fil.chown(fil_uid, fil_gid) catch |err| {
if (op.verbose) {
std.fmt.format(op.verbose_logger, "error chmod {s}: {}\n", .{ path, err }) catch {};
}
return;
};
},
// .symlink, .ext_symlink => {},
else => {
std.debug.print("TODO: {}\n", .{inode.hdr.type});
},
}
}
}; };
return self.extractReal(alloc, io, &cache, &decomp.interface, filepath, options);
} }
fn extractReal(self: File, alloc: std.mem.Allocator, io: Io, cache: *SharedCache, decomp: *const Decompressor, filepath: []const u8, options: ExtractionOptions) !void {
_ = options;
switch (self.inode.hdr.inode_type) {
.file, .ext_file => {
var ext = try self.inode.dataExtractor(
self.archive.file,
cache,
decomp,
self.archive.super.block_size,
);
var atomic_file = try Io.Dir.cwd().createFileAtomic(io, filepath, .{});
defer atomic_file.deinit(io);
try ext.extract(alloc, io, atomic_file.file);
try atomic_file.link(io);
},
else => return error.TODO,
}
}
// Types
pub const Error = error{
FileNotFound,
} || Inode.Error;
-7
View File
@@ -1,7 +0,0 @@
const BlockSize = @import("inode_data/file.zig").BlockSize;
pub const FragEntry = extern struct {
start: u64,
size: BlockSize,
_: u32,
};
+7
View File
@@ -0,0 +1,7 @@
const BlockSize = @import("inode/file.zig").BlockSize;
pub const FragEntry = packed struct {
block: u64,
size: BlockSize,
_: u32,
};
+52 -165
View File
@@ -1,155 +1,12 @@
//! A file-system object. Represents a File or directory.
const std = @import("std"); const std = @import("std");
const Reader = std.Io.Reader;
const Io = std.Io;
const Archive = @import("archive.zig"); const dir = @import("inode/dir.zig");
const DirEntry = @import("directory.zig"); const file = @import("inode/file.zig");
const ExtractionOptions = @import("options.zig"); const misc = @import("inode/misc.zig");
const dir = @import("inode_data/dir.zig");
const file = @import("inode_data/file.zig");
const misc = @import("inode_data/misc.zig");
const LookupTable = @import("lookup_table.zig");
const DataExtractor = @import("util/data_extractor.zig");
const DataReader = @import("util/data_reader.zig");
const Decompressor = @import("util/decompressor.zig");
const MetadataReader = @import("util/metadata.zig");
const OffsetFile = @import("util/offset_file.zig");
const SharedCache = @import("util/shared_cache.zig");
const XattrTable = @import("xattr_table.zig");
const Inode = @This(); pub const Ref = packed struct {
offset: u16,
hdr: Header, block: u32,
data: Data,
pub fn read(alloc: std.mem.Allocator, rdr: *Reader, block_size: u32) !Inode {
var hdr: Header = undefined;
try rdr.readSliceEndian(Header, @ptrCast(&hdr), .little);
return .{
.hdr = hdr,
.data = switch (hdr.inode_type) {
.dir => .{ .dir = try .read(rdr) },
.file => .{ .file = try .read(alloc, rdr, block_size) },
.symlink => .{ .symlink = try .read(alloc, rdr) },
.block_dev => .{ .block_dev = try .read(rdr) },
.char_dev => .{ .char_dev = try .read(rdr) },
.fifo => .{ .fifo = try .read(rdr) },
.socket => .{ .socket = try .read(rdr) },
.ext_dir => .{ .ext_dir = try .read(rdr) },
.ext_file => .{ .ext_file = try .read(alloc, rdr, block_size) },
.ext_symlink => .{ .ext_symlink = try .read(alloc, rdr) },
.ext_block_dev => .{ .ext_block_dev = try .read(rdr) },
.ext_char_dev => .{ .ext_char_dev = try .read(rdr) },
.ext_fifo => .{ .ext_fifo = try .read(rdr) },
.ext_socket => .{ .ext_socket = try .read(rdr) },
},
};
}
pub fn deinit(self: Inode, alloc: std.mem.Allocator) void {
switch (self.data) {
.file => |d| d.deinit(alloc),
.symlink => |d| d.deinit(alloc),
.ext_file => |d| d.deinit(alloc),
.ext_symlink => |d| d.deinit(alloc),
else => {},
}
}
// Utility Functions
/// Read the directory entries
pub fn readDirectory(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, decomp: *const Decompressor, dir_offset: u64) ![]DirEntry {
return switch (self.data) {
.dir => |d| readDirFromData(alloc, io, fil, decomp, dir_offset, d),
.ext_dir => |d| readDirFromData(alloc, io, fil, decomp, dir_offset, d),
else => Error.NotDirectory,
};
}
fn readDirFromData(alloc: std.mem.Allocator, io: Io, fil: OffsetFile, decomp: *const Decompressor, dir_offset: u64, d: anytype) ![]DirEntry {
var rdr = try fil.readerAt(io, dir_offset + d.block_start, &[0]u8{});
var meta: MetadataReader = .init(alloc, &rdr.interface, decomp);
try meta.interface.discardAll(d.block_offset);
return DirEntry.readDirectory(alloc, &meta.interface, d.size);
}
/// Get a reader for a regular file's data.
pub fn dataReader(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, cache: *SharedCache, decomp: *const Decompressor, block_size: u32) !DataReader {
return switch (self.data) {
.file => |f| getReaderFromData(alloc, io, fil, cache, decomp, block_size, f),
.ext_file => |f| getReaderFromData(alloc, io, fil, cache, decomp, block_size, f),
else => Error.NotRegularFile,
};
}
fn getReaderFromData(alloc: std.mem.Allocator, io: Io, fil: OffsetFile, cache: *SharedCache, decomp: *const Decompressor, block_size: u32, d: anytype) !DataReader {
const ext: DataReader = .init(alloc, io, fil, cache, decomp, block_size, d.size, d.block_start, d.blocks);
if (d.frag_block_offset == 0xFFFFFFFF) {
// TODO:
return error.TODO;
}
return ext;
}
/// Get an extractor for a regular file's data.
pub fn dataExtractor(self: Inode, fil: OffsetFile, cache: *SharedCache, decomp: *const Decompressor, block_size: u32) !DataExtractor {
return switch (self.data) {
.file => |f| getExtractorFromData(fil, cache, decomp, block_size, f),
.ext_file => |f| getExtractorFromData(fil, cache, decomp, block_size, f),
else => Error.NotRegularFile,
};
}
fn getExtractorFromData(fil: OffsetFile, cache: *SharedCache, decomp: *const Decompressor, block_size: u32, d: anytype) !DataExtractor {
const ext: DataExtractor = .init(fil, cache, decomp, block_size, d.size, d.block_start, d.blocks);
if (d.frag_block_offset == 0xFFFFFFFF) {
// TODO:
return error.TODO;
}
return ext;
}
// Get a symlink's target path
pub fn symlinkTarget(self: Inode) ![]const u8 {
return switch (self.data) {
.symlink => |s| s.target,
.ext_symlink => |s| s.target,
else => Error.NotSymlink,
};
}
// Get inode's gid
pub fn gid(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, decomp: *const Decompressor, id_table_start: u64) !u16 {
return LookupTable.lookupValue(u16, alloc, io, decomp, fil, id_table_start, self.hdr.gid_idx);
}
// Get inode's uid
pub fn uid(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, decomp: *const Decompressor, id_table_start: u64) !u16 {
return LookupTable.lookupValue(u16, alloc, io, decomp, fil, id_table_start, self.hdr.uid_idx);
}
// Get an inode's xattr values. If the inode does not have xattr values (including if the inode is not an extended type), an empty slice is returned.
pub fn xattrValues(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, decomp: *const Decompressor, xattr_table_start: u64) ![]XattrTable.XattrOwned {
const idx = switch (self.data) {
.ext_dir => |e| e.xattr_idx,
.ext_file => |e| e.xattr_idx,
.ext_symlink => |e| e.xattr_idx,
.ext_block_dev => |e| e.xattr_idx,
.ext_char_dev => |e| e.xattr_idx,
.ext_fifo => |e| e.xattr_idx,
.ext_socket => |e| e.xattr_idx,
else => return &[0]XattrTable.XattrOwned{},
};
if (idx == 0xFFFFFFFF) return &[0]XattrTable.XattrOwned{};
return XattrTable.statelessLookup(alloc, io, decomp, fil, xattr_table_start, idx);
}
// Types
pub const Error = error{
NotDirectory,
NotRegularFile,
NotSymlink,
NotExtended,
};
pub const Ref = packed struct(u64) {
block_offset: u16,
block_start: u32,
_: u16, _: u16,
}; };
@@ -170,7 +27,16 @@ pub const Type = enum(u16) {
ext_socket, ext_socket,
}; };
pub const Data = union(Type) { pub const Header = packed struct {
type: Type,
perm: u16,
uid_idx: u16,
gid_idx: u16,
mod_time: u32,
num: u32,
};
pub const Data = union(enum) {
dir: dir.Dir, dir: dir.Dir,
file: file.File, file: file.File,
symlink: misc.Symlink, symlink: misc.Symlink,
@@ -187,20 +53,41 @@ pub const Data = union(Type) {
ext_socket: misc.ExtIPC, ext_socket: misc.ExtIPC,
}; };
pub const Header = extern struct { const Self = @This();
inode_type: Type,
permissions: u16,
uid_idx: u16,
gid_idx: u16,
mod_time: u32,
num: u32,
};
// Extract hdr: Header,
data: Data,
pub fn extract(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, super: Archive.Superblock, path: []const u8, options: ExtractionOptions) !void {} pub fn init(rdr: anytype, alloc: std.mem.Allocator, block_size: u32) !Self {
pub fn extractDir(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, super: Archive.Superblock, path: []const u8, options: ExtractionOptions) !void {} var hdr: Header = undefined;
pub fn extractRegFile(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, super: Archive.Superblock, path: []const u8, options: ExtractionOptions) !void {} _ = try rdr.read(std.mem.asBytes(&hdr));
pub fn extractSymlink(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, super: Archive.Superblock, path: []const u8, options: ExtractionOptions) !void {} const data: Data = switch (hdr.type) {
pub fn extractDevice(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, super: Archive.Superblock, path: []const u8, options: ExtractionOptions) !void {} .dir => .{ .dir = try .init(rdr) },
pub fn extractIPC(self: Inode, alloc: std.mem.Allocator, io: Io, fil: OffsetFile, super: Archive.Superblock, path: []const u8, options: ExtractionOptions) !void {} .file => .{ .file = try .init(rdr, alloc, block_size) },
.symlink => .{ .symlink = try .init(rdr, alloc) },
.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(rdr, alloc, block_size) },
.ext_symlink => .{ .ext_symlink = try .init(rdr, alloc) },
.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 deinit(self: Self, alloc: std.mem.Allocator) void {
switch (self.data) {
.file => |f| alloc.free(f.block_sizes),
.ext_file => |f| alloc.free(f.block_sizes),
.symlink => |s| alloc.free(s.target),
.ext_symlink => |s| alloc.free(s.target),
else => {},
}
}
+31
View File
@@ -0,0 +1,31 @@
const std = @import("std");
pub const Dir = packed struct {
block: u32,
hard_link: u32,
size: u16,
offset: u16,
parent_num: u32,
pub fn init(rdr: anytype) !Dir {
var out: Dir = undefined;
_ = try rdr.read(std.mem.asBytes(&out));
return out;
}
};
pub const ExtDir = packed struct {
hard_link: u32,
size: u32,
block: u32,
parent_num: u32,
idx_count: u16,
offset: u16,
xattr_idx: u32,
pub fn init(rdr: anytype) !ExtDir {
var out: ExtDir = undefined;
_ = try rdr.read(std.mem.asBytes(&out));
return out;
}
};
+77
View File
@@ -0,0 +1,77 @@
const std = @import("std");
pub const BlockSize = packed struct {
size: u24,
uncompressed: bool,
_: u7,
};
pub const File = struct {
block: u32,
frag_idx: u32,
frag_offset: u32,
size: u32,
block_sizes: []BlockSize,
pub fn init(rdr: anytype, alloc: std.mem.Allocator, block_size: u32) !File {
var fixed: [16]u8 = undefined;
_ = try rdr.read(&fixed);
const frag_idx = std.mem.readInt(u32, fixed[4..8], .little);
const size = std.mem.readInt(u32, fixed[12..16], .little);
var blocks: u32 = size / block_size;
if (size % block_size > 0 and frag_idx == 0xffffffff) {
blocks += 1;
}
const block_sizes = try alloc.alloc(BlockSize, blocks);
errdefer alloc.free(block_sizes);
_ = try rdr.read(std.mem.sliceAsBytes(block_sizes));
return .{
.block = std.mem.readInt(u32, fixed[0..4], .little),
.frag_idx = frag_idx,
.frag_offset = std.mem.readInt(u32, fixed[8..12], .little),
.size = size,
.block_sizes = block_sizes,
};
}
pub fn hasFragment(self: File) bool {
return self.frag_idx != 0xffffffff;
}
};
pub const ExtFile = struct {
block: u64,
size: u64,
sparse: u64,
hard_link: u32,
frag_idx: u32,
frag_offset: u32,
xattr_idx: u32,
block_sizes: []BlockSize,
pub fn init(rdr: anytype, alloc: std.mem.Allocator, block_size: u32) !ExtFile {
var fixed: [40]u8 = undefined;
_ = try rdr.read(&fixed);
const size = std.mem.readInt(u64, fixed[8..16], .little);
const frag_idx = std.mem.readInt(u32, fixed[28..32], .little);
var blocks: u32 = @truncate(size / block_size);
if (size % block_size > 0 and frag_idx == 0xffffffff) {
blocks += 1;
}
const block_sizes = try alloc.alloc(BlockSize, blocks);
errdefer alloc.free(block_sizes);
_ = try rdr.read(std.mem.sliceAsBytes(block_sizes));
return .{
.block = std.mem.readInt(u64, fixed[0..8], .little),
.size = size,
.sparse = std.mem.readInt(u64, fixed[16..24], .little),
.hard_link = std.mem.readInt(u32, fixed[24..28], .little),
.frag_idx = frag_idx,
.frag_offset = std.mem.readInt(u32, fixed[32..36], .little),
.xattr_idx = std.mem.readInt(u32, fixed[36..40], .little),
.block_sizes = block_sizes,
};
}
pub fn hasFragment(self: ExtFile) bool {
return self.frag_idx != 0xffffffff;
}
};
+87
View File
@@ -0,0 +1,87 @@
const std = @import("std");
pub const Symlink = struct {
hard_link: u32,
// size: u32,
target: []const u8,
pub fn init(rdr: anytype, alloc: std.mem.Allocator) !Symlink {
var fixed: [8]u8 = undefined;
_ = try rdr.read(&fixed);
const size = std.mem.readInt(u32, fixed[4..8], .little);
const target = try alloc.alloc(u8, size);
errdefer alloc.free(target);
_ = try rdr.read(target);
return .{
.hard_link = std.mem.readInt(u32, fixed[0..4], .little),
.target = target,
};
}
};
pub const ExtSymlink = struct {
hard_link: u32,
// size: u32,
target: []const u8,
xattr_idx: u32,
pub fn init(rdr: anytype, alloc: std.mem.Allocator) !ExtSymlink {
var fixed: [8]u8 = undefined;
_ = try rdr.read(&fixed);
const size = std.mem.readInt(u32, fixed[4..8], .little);
const target = try alloc.alloc(u8, size);
errdefer alloc.free(target);
_ = try rdr.read(target);
var xattr_idx: u32 = 0;
_ = try rdr.read(std.mem.asBytes(&xattr_idx));
return .{
.hard_link = std.mem.readInt(u32, fixed[0..4], .little),
.target = target,
.xattr_idx = xattr_idx,
};
}
};
pub const Dev = packed struct {
hard_link: u32,
device: u32,
pub fn init(rdr: anytype) !Dev {
var out: Dev = undefined;
_ = try rdr.read(std.mem.asBytes(&out));
return out;
}
};
pub const ExtDev = packed struct {
hard_link: u32,
device: u32,
xattr_idx: u32,
pub fn init(rdr: anytype) !ExtDev {
var out: ExtDev = undefined;
_ = try rdr.read(std.mem.asBytes(&out));
return out;
}
};
pub const IPC = packed struct {
hard_link: u32,
pub fn init(rdr: anytype) !IPC {
var out: IPC = undefined;
_ = try rdr.read(std.mem.asBytes(&out));
return out;
}
};
pub const ExtIPC = packed struct {
hard_link: u32,
xattr_idx: u32,
pub fn init(rdr: anytype) !ExtIPC {
var out: ExtIPC = undefined;
_ = try rdr.read(std.mem.asBytes(&out));
return out;
}
};
-32
View File
@@ -1,32 +0,0 @@
const Reader = @import("std").Io.Reader;
pub const Dir = extern struct {
block_start: u32,
hard_links: u32,
size: u16,
block_offset: u16,
parent_num: u32,
pub fn read(rdr: *Reader) !Dir {
var d: Dir = undefined;
try rdr.readSliceEndian(Dir, @ptrCast(&d), .little);
return d;
}
};
pub const ExtDir = extern struct {
hard_links: u32,
size: u32,
block_start: u32,
parent_num: u32,
idx_count: u16,
block_offset: u16,
xattr_id: u32,
// index: []DirIndex
pub fn read(rdr: *Reader) !ExtDir {
var d: ExtDir = undefined;
try rdr.readSliceEndian(ExtDir, @ptrCast(&d), .little);
return d;
}
};
-97
View File
@@ -1,97 +0,0 @@
const std = @import("std");
const Reader = std.Io.Reader;
pub const BlockSize = packed struct(u32) {
size: u24,
uncompressed: bool,
_: u7,
};
const FileRawRead = extern struct {
block_start: u32,
frag_idx: u32,
frag_block_offset: u32,
size: u32,
};
pub const File = struct {
block_start: u32,
frag_idx: u32,
frag_block_offset: u32,
size: u32,
block_sizes: []BlockSize,
pub fn read(alloc: std.mem.Allocator, rdr: *Reader, block_size: u32) !File {
var raw: FileRawRead = undefined;
try rdr.readSliceEndian(FileRawRead, @ptrCast(&raw), .little);
var num_blocks: u32 = raw.size / block_size;
if (raw.size % block_size != 0 and raw.frag_idx == 0xFFFFFFFF)
num_blocks += 1;
const sizes = try alloc.alloc(BlockSize, num_blocks);
errdefer alloc.free(sizes);
try rdr.readSliceEndian(BlockSize, sizes, .little);
return .{
.block_start = raw.block_start,
.frag_idx = raw.frag_idx,
.frag_block_offset = raw.frag_block_offset,
.size = raw.size,
.block_sizes = sizes,
};
}
pub fn deinit(self: File, alloc: std.mem.Allocator) void {
alloc.free(self.block_sizes);
}
};
const ExtFileRawRead = extern struct {
block_start: u64,
size: u64,
sparse: u64,
hard_links: u32,
frag_idx: u32,
frag_block_offset: u32,
xattr_idx: u32,
};
pub const ExtFile = struct {
block_start: u64,
size: u64,
sparse: u64,
hard_links: u32,
frag_idx: u32,
frag_block_offset: u32,
xattr_idx: u32,
block_sizes: []BlockSize,
pub fn read(alloc: std.mem.Allocator, rdr: *Reader, block_size: u32) !ExtFile {
var raw: ExtFileRawRead = undefined;
try rdr.readSliceEndian(ExtFileRawRead, @ptrCast(&raw), .little);
var num_blocks: u32 = @truncate(raw.size / block_size);
if (raw.size % block_size != 0 and raw.frag_idx == 0xFFFFFFFF)
num_blocks += 1;
const sizes = try alloc.alloc(BlockSize, num_blocks);
errdefer alloc.free(sizes);
try rdr.readSliceEndian(BlockSize, sizes, .little);
return .{
.block_start = raw.block_start,
.size = raw.size,
.sparse = raw.sparse,
.hard_links = raw.hard_links,
.frag_idx = raw.frag_idx,
.frag_block_offset = raw.frag_block_offset,
.xattr_idx = raw.xattr_idx,
.block_sizes = sizes,
};
}
pub fn deinit(self: ExtFile, alloc: std.mem.Allocator) void {
alloc.free(self.block_sizes);
}
};
-98
View File
@@ -1,98 +0,0 @@
const std = @import("std");
const Reader = std.Io.Reader;
pub const Symlink = struct {
hard_links: u32,
target: []const u8,
pub fn read(alloc: std.mem.Allocator, rdr: *Reader) !Symlink {
var start: [8]u8 = undefined;
try rdr.readSliceAll(&start);
const target_size = std.mem.readInt(u32, start[4..8], .little);
const target = try alloc.alloc(u8, target_size + 1);
errdefer alloc.free(target);
try rdr.readSliceEndian(u8, target, .little);
return .{
.hard_links = std.mem.readInt(u32, start[0..4], .little),
.target = target,
};
}
pub fn deinit(self: Symlink, alloc: std.mem.Allocator) void {
alloc.free(self.target);
}
};
pub const ExtSymlink = struct {
hard_links: u32,
target: []const u8,
xattr_idx: u32,
pub fn read(alloc: std.mem.Allocator, rdr: *Reader) !ExtSymlink {
var start: [8]u8 = undefined;
try rdr.readSliceAll(&start);
const target_size = std.mem.readInt(u32, start[4..8], .little);
const target = try alloc.alloc(u8, target_size + 1);
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 = std.mem.readInt(u32, start[0..4], .little),
.target = target,
.xattr_idx = xattr_idx,
};
}
pub fn deinit(self: ExtSymlink, alloc: std.mem.Allocator) void {
alloc.free(self.target);
}
};
/// A block or character device.
pub const Dev = extern struct {
hard_links: u32,
dev: u32,
pub fn read(rdr: *Reader) !Dev {
var d: Dev = undefined;
try rdr.readSliceEndian(Dev, @ptrCast(&d), .little);
return d;
}
};
/// An extended block or character device.
pub const ExtDev = extern struct {
hard_links: u32,
dev: u32,
xattr_idx: u32,
pub fn read(rdr: *Reader) !ExtDev {
var d: ExtDev = undefined;
try rdr.readSliceEndian(ExtDev, @ptrCast(&d), .little);
return d;
}
};
/// A socket or FIFO file.
pub const IPC = extern struct {
hard_links: u32,
pub fn read(rdr: *Reader) !IPC {
var d: IPC = undefined;
try rdr.readSliceEndian(IPC, @ptrCast(&d), .little);
return d;
}
};
/// An extended socket or FIFO file.
pub const ExtIPC = extern struct {
hard_links: u32,
xattr_idx: u32,
pub fn read(rdr: *Reader) !ExtIPC {
var d: ExtIPC = undefined;
try rdr.readSliceEndian(ExtIPC, @ptrCast(&d), .little);
return d;
}
};
-84
View File
@@ -1,84 +0,0 @@
const std = @import("std");
const Io = std.Io;
const Decompressor = @import("util/decompressor.zig");
const MetadataReader = @import("util/metadata.zig");
const OffsetFile = @import("util/offset_file.zig");
pub fn lookupValue(comptime T: anytype, alloc: std.mem.Allocator, io: Io, decomp: *Decompressor, file: OffsetFile, table_start: u64, idx: u16) !T {
const T_PER_BLOCK: u16 = 8192 / @sizeOf(T);
const block = idx / T_PER_BLOCK;
const block_offset = idx % T_PER_BLOCK;
var rdr = try file.readerAt(io, table_start + (8 * block), &[0]u8{});
var offset: u64 = undefined;
try rdr.interface.readSliceEndian(u64, @ptrCast(&offset), .little);
rdr = try file.readerAt(io, offset, &[0]u8{});
var meta: MetadataReader = .init(alloc, &rdr, decomp);
try meta.interface.discardAll(@sizeOf(T) * block_offset);
var out: T = undefined;
try meta.interface.readSliceEndian(T, @ptrCast(&out), .little);
return out;
}
pub fn CachedTable(comptime T: anytype) type {
return struct {
const T_PER_BLOCK: u16 = 8192 / @sizeOf(T);
const Table = @This();
alloc: std.mem.Allocator,
fil: OffsetFile,
table_start: u64,
total_num: u32,
table: std.AutoHashMap(u32, []T),
mut: Io.Mutex = .init,
pub fn init(alloc: std.mem.Allocator, fil: OffsetFile, offset: u64, total_num: u32) Table {
return .{
.alloc = alloc,
.fil = fil,
.table_start = offset,
.total_num = total_num,
.table = .init(alloc),
};
}
pub fn deinit(self: *Table, io: Io) void {
self.mut.lockUncancelable(io);
self.table.deinit();
}
pub fn get(self: *Table, io: Io, idx: u32) !T {
const block = idx / T_PER_BLOCK;
const block_offset = idx % T_PER_BLOCK;
if (self.table.contains(block))
return self.table.get(block).?[block_offset];
try self.mut.lock(io);
defer self.mut.unlock(io);
if (self.table.contains(block))
return self.table.get(block).?[block_offset];
var rdr = try self.fil.readerAt(io, self.table_start + (8 * block), &[0]u8{});
var offset: u64 = undefined;
try rdr.interface.readSliceEndian(u64, @ptrCast(&offset), .little);
const len: u16 = if (self.total_num % T_PER_BLOCK != 0 and block == (self.total_num - 1) / T_PER_BLOCK)
self.total_num % T_PER_BLOCK
else
T_PER_BLOCK;
rdr = try self.fil.readerAt(io, offset, &[0]u8{});
var meta: MetadataReader = .init(self.alloc, &rdr, self.decomp);
try self.table.put(block, try meta.interface.readSliceEndianAlloc(self.alloc, T, len, .little));
}
};
}
-33
View File
@@ -1,33 +0,0 @@
//! Options for file/directory extraction.
const std = @import("std");
const Writer = std.Io.Writer;
const ExtractionOptions = @This();
/// The number of threads used for extraction. 0 implies single threaded.
threads: usize = 1,
/// 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 SingleThreadedDefault: ExtractionOptions = .{};
pub fn Default() !ExtractionOptions {
return .{
.threads = try std.Thread.getCpuCount(),
};
}
pub fn VerboseDefault(wrt: *Writer) !ExtractionOptions {
return .{
.verbose = true,
.verbose_writer = wrt,
.threads = try std.Thread.getCpuCount(),
};
}
+85
View File
@@ -0,0 +1,85 @@
const std = @import("std");
const Inode = @import("inode.zig");
const File = @import("file.zig").File;
const Table = @import("table.zig").Table;
const PRead = @import("reader/p_read.zig").PRead;
const FragEntry = @import("fragment.zig").FragEntry;
const Superblock = @import("superblock.zig").Superblock;
const ExtractionOptions = @import("extract_options.zig");
const MetadataReader = @import("reader/metadata.zig").MetadataReader;
pub const SfsError = error{
NotExportable,
};
pub fn SfsReader(comptime T: type) type {
comptime std.debug.assert(std.meta.hasFn(T, "pread"));
return struct {
const Self = @This();
alloc: std.mem.Allocator,
rdr: PRead(T),
super: Superblock = undefined,
/// ID table. Can be accessed directly
id_table: Table(u32, T) = undefined,
/// Fragment table. Can be accessed directly
frag_table: Table(FragEntry, T) = undefined,
/// Export table. Each element is an inode referce.
/// If accessing directly, keep in mind, the table starts at inode 1, as such it's recommended to use the InodeAt function instead.
export_table: Table(Inode.Ref, T) = undefined,
pub fn init(alloc: std.mem.Allocator, rdr: T, offset: u64) !Self {
var out: Self = .{
.alloc = alloc,
.rdr = .init(rdr, offset),
};
_ = try rdr.pread(std.mem.asBytes(&out.super), 0);
out.frag_table = .init(alloc, out.rdr, out.super.comp, out.super.frag_start, out.super.frag_count);
out.id_table = .init(alloc, out.rdr, out.super.comp, out.super.id_start, out.super.id_count);
out.export_table = .init(alloc, out.rdr, out.super.comp, out.super.export_start, out.super.inode_count);
return out;
}
pub fn deinit(self: *Self) void {
self.id_table.deinit();
self.frag_table.deinit();
self.export_table.deinit();
}
/// A representation of the archives root folder.
pub fn root(self: *Self) !File(T) {
return .initFromRef(self, self.super.root_ref, "");
}
/// Get the file at path. Equivelent to calling open on the root File.
pub fn open(self: *Self, path: []const u8) !File(T) {
var rt = try self.root();
if (path.len == 0 or (path.len == 1 and path[0] == '/') or path.len == 1 and path[0] == '.') return rt;
defer rt.deinit();
return rt.open(path);
}
/// Extract the entire archive to the given path & with the given options.
/// Equivelent to calling extract on the root File.
pub fn extract(self: *Self, op: ExtractionOptions, path: []const u8) !void {
var rt = try self.root();
defer rt.deinit();
return rt.extract(op, path);
}
/// Returns the Inode with the given Inode Number.
/// Requires the archive to have an export table.
pub fn inodeAt(self: Self, num: u32) !Inode {
if (!self.super.flags.has_export) return SfsError.NotExportable;
const ref = try self.export_table.get(num - 1);
var meta = MetadataReader(T).init(
self.alloc,
self.super.comp,
self.rdr,
self.super.inode_start + ref.block,
);
try meta.skip(ref.offset);
return .init(meta, self.alloc, self.super.block_size);
}
};
}
+271
View File
@@ -0,0 +1,271 @@
const std = @import("std");
const Inode = @import("../inode.zig");
const PRead = @import("p_read.zig").PRead;
const SfsReader = @import("../reader.zig").SfsReader;
const FragEntry = @import("../fragment.zig").FragEntry;
const BlockSize = @import("../inode/file.zig").BlockSize;
const Compression = @import("../superblock.zig").Compression;
const DataReaderError = error{
EOF,
InvalidIndex,
ExtractionActive,
};
const DecompCompletion = struct {
errs: std.ArrayList(anyerror),
map: std.AutoArrayHashMap(usize, []u8),
mut: std.Thread.Mutex = .{},
cond: std.Thread.Condition = .{},
fn init(alloc: std.mem.Allocator) DecompCompletion {
return .{
.errs = .init(alloc),
.map = .init(alloc),
};
}
fn deinit(self: *DecompCompletion) void {
self.errs.deinit();
self.map.deinit();
}
fn clear(self: *DecompCompletion) void {
self.mut.lock();
defer self.mut.unlock();
self.errs.clearAndFree();
self.map.clearAndFree();
}
fn add(self: *DecompCompletion, idx: usize, data: []u8) !void {
self.mut.lock();
defer self.mut.unlock();
defer self.cond.signal();
try self.map.put(idx, data);
}
fn addErr(self: *DecompCompletion, err: anyerror) void {
self.mut.lock();
defer self.mut.unlock();
defer self.cond.signal();
self.errs.append(err) catch {};
}
fn getBlock(self: *DecompCompletion, idx: usize) ?[]u8 {
const res = self.map.fetchSwapRemove(idx);
if (res == null) return null;
return res.?.value;
}
fn hasErrs(self: DecompCompletion) bool {
return self.errs.items.len > 0;
}
fn condWait(self: *DecompCompletion) void {
self.cond.wait(&self.mut);
}
};
pub fn DataReader(comptime T: type) type {
return struct {
const Self = @This();
alloc: std.mem.Allocator,
rdr: PRead(T),
comp: Compression,
block_size: u32,
sizes: []BlockSize,
offsets: []u64,
file_size: u64,
frag: ?[]u8 = null,
completion: DecompCompletion,
pub fn init(rdr: *SfsReader(T), inode: Inode) !Self {
var sizes: []BlockSize = undefined;
var file_size: u64 = 0;
var offsets: []u64 = undefined;
switch (inode.data) {
.file => |f| {
sizes = f.block_sizes;
file_size = f.size;
offsets = try rdr.alloc.alloc(u64, sizes.len);
if (sizes.len > 0) offsets[0] = f.block;
},
.ext_file => |f| {
sizes = f.block_sizes;
file_size = f.size;
offsets = try rdr.alloc.alloc(u64, sizes.len);
if (sizes.len > 0) offsets[0] = f.block;
},
else => unreachable,
}
if (offsets.len > 1) {
for (1..offsets.len) |i| {
offsets[i] = offsets[i - 1] + sizes[i - 1].size;
}
}
return .{
.alloc = rdr.alloc,
.rdr = rdr.rdr,
.comp = rdr.super.comp,
.block_size = rdr.super.block_size,
.sizes = sizes,
.offsets = offsets,
.file_size = file_size,
.completion = .init(rdr.alloc),
};
}
pub fn deinit(self: *Self) void {
self.alloc.free(self.offsets);
self.completion.deinit();
}
pub fn addFragment(self: *Self, data: []u8) void {
self.frag = data;
}
pub fn writeTo(self: *Self, wrt: anytype) !void {
comptime std.debug.assert(std.meta.hasFn(@TypeOf(wrt), "write") or std.meta.hasFn(@TypeOf(wrt), "pwrite"));
var write_thr = try std.Thread.spawn(
.{ .allocator = self.alloc },
writeThread,
.{ self, wrt, null, null },
);
defer self.completion.clear();
for (0..self.numBlocks()) |i| {
var thr = std.Thread.spawn(
.{ .allocator = self.alloc },
decompThread,
.{ self, i },
) catch |err| {
self.completion.addErr(err);
continue;
};
thr.detach();
}
write_thr.join();
if (self.completion.hasErrs()) return self.completion.errs.items[0];
}
pub fn writeToNoBlock(self: *Self, wrt: anytype, comptime finish: anytype, finish_args: anytype) !void {
comptime std.debug.assert(std.meta.hasFn(@TypeOf(wrt), "write") or std.meta.hasFn(@TypeOf(wrt), "pwrite"));
errdefer self.completion.clear();
var write_thr = try std.Thread.spawn(
.{ .allocator = self.alloc },
writeThread,
.{ self, wrt, finish, finish_args },
);
write_thr.detach();
for (0..self.numBlocks()) |i| {
var thr = std.Thread.spawn(
.{ .allocator = self.alloc },
decompThread,
.{ self, i },
) catch |err| {
self.completion.addErr(err);
continue;
};
thr.detach();
}
}
fn numBlocks(self: Self) usize {
var out = self.sizes.len;
if (self.frag != null) out += 1;
return out;
}
/// Returns the decompressed data block at the given idx.
/// If the block is sparse (filled with 0s), a zero length slice is returned.
fn blockAt(self: Self, idx: usize) ![]u8 {
if (idx >= self.numBlocks()) return DataReaderError.InvalidIndex;
const size = self.sizes[idx];
if (size.size == 0) return &[0]u8{};
const block = try self.alloc.alloc(u8, blk: {
if (idx == self.numBlocks() - 1) break :blk self.file_size % self.block_size;
break :blk self.block_size;
});
if (idx == self.sizes.len and self.frag != null) {
@memcpy(block, self.frag.?);
return block;
}
if (size.uncompressed) {
_ = try self.rdr.pread(block, self.offsets[idx]);
return block;
}
_ = try self.comp.decompress(
1024 * 1024,
self.alloc,
self.rdr.readerAt(self.offsets[idx]).reader(),
block,
);
return block;
}
fn writeThread(
self: *Self,
wrt: anytype,
comptime finish: anytype,
finish_args: anytype,
) void {
var cur_idx: usize = 0;
self.completion.mut.lock();
defer self.completion.mut.unlock();
while (cur_idx < self.numBlocks() and !self.completion.hasErrs()) {
self.completion.condWait();
if (self.completion.hasErrs()) break;
if (comptime std.meta.hasFn(@TypeOf(wrt), "pwrite")) {
for (self.completion.map.keys()) |_| {
const k = self.completion.map.keys()[0];
const blk = self.completion.getBlock(k).?;
defer self.alloc.free(blk);
if (blk.len > 0) {
_ = wrt.pwrite(blk, self.block_size * k) catch |err| {
self.completion.addErr(err);
break;
};
} else {
_ = wrt.pwrite(&[1]u8{0}, (self.block_size * (k + 1)) - 1) catch |err| {
self.completion.addErr(err);
break;
};
}
cur_idx += 1;
}
continue;
} else {
while (self.completion.getBlock(cur_idx)) |blk| {
defer self.alloc.free(blk);
if (blk.len > 0) {
_ = wrt.write(blk) catch |err| {
self.completion.addErr(err);
break;
};
} else {
const blank: [1024 * 1024]u8 = [1]u8{0} ** (1024 * 1024);
_ = wrt.write(blank[0..self.block_size]) catch |err| {
self.completion.addErr(err);
break;
};
}
cur_idx += 1;
}
}
}
if (comptime @TypeOf(finish) != @TypeOf(null) and @TypeOf(finish_args) != @TypeOf(null)) @call(.auto, finish, finish_args);
}
fn decompThread(
self: *Self,
idx: usize,
) void {
if (self.completion.hasErrs()) return;
defer self.completion.cond.signal();
const block = self.blockAt(idx) catch |err| {
self.completion.addErr(err);
return;
};
self.completion.add(idx, block) catch |err| {
self.completion.addErr(err);
};
}
};
}
+82
View File
@@ -0,0 +1,82 @@
const std = @import("std");
const PRead = @import("p_read.zig").PRead;
const Compression = @import("../superblock.zig").Compression;
const MetaHeader = packed struct {
size: u15,
uncompressed: bool,
};
pub fn MetadataReader(comptime T: type) type {
comptime std.debug.assert(std.meta.hasFn(T, "read"));
return struct {
const Self = @This();
alloc: std.mem.Allocator,
comp: Compression,
rdr: PRead(T),
offset: u64,
block: [8192]u8 = undefined,
block_size: usize = 0,
block_offset: u32 = 0,
pub fn init(alloc: std.mem.Allocator, comp: Compression, rdr: PRead(T), offset: u64) Self {
return .{
.alloc = alloc,
.comp = comp,
.rdr = rdr,
.offset = offset,
};
}
fn readNextBlock(self: *Self) !void {
var hdr: MetaHeader = undefined;
_ = try self.rdr.pread(std.mem.asBytes(&hdr), self.offset);
self.offset += 2;
if (hdr.uncompressed) {
self.block_size = try self.rdr.pread(self.block[0..hdr.size], self.offset);
} else {
self.block_size = try self.comp.decompress(
8192,
self.alloc,
self.rdr.readerAt(self.offset).reader(),
&self.block,
);
}
self.offset += hdr.size;
self.block_offset = 0;
}
pub fn skip(self: *Self, offset: u32) !void {
var skipped: u32 = 0;
var hdr: MetaHeader = undefined;
while (offset - skipped >= 8192) {
_ = try self.rdr.pread(std.mem.asBytes(&hdr), self.offset);
self.offset += 2 + hdr.size;
skipped += 8192;
}
var to_skip: u32 = 0;
while (skipped < offset) {
if (self.block_offset >= self.block_size) try self.readNextBlock();
to_skip = @min(self.block_size - self.block_offset, offset - skipped);
self.block_offset += to_skip;
skipped += to_skip;
}
}
pub fn read(self: *Self, buf: []u8) !usize {
var cur_red: usize = 0;
var to_read: usize = 0;
while (cur_red < buf.len) {
if (self.block_offset >= self.block_size) try self.readNextBlock();
to_read = @min(buf.len - cur_red, self.block_size - self.block_offset);
@memcpy(buf[cur_red .. cur_red + to_read], self.block[self.block_offset .. self.block_offset + to_read]);
cur_red += to_read;
self.block_offset += @truncate(to_read);
}
return cur_red;
}
};
}
+29
View File
@@ -0,0 +1,29 @@
const std = @import("std");
const ToRead = @import("to_read.zig").ToRead;
/// A simple wrapper around a type with the pread([]u8, u64) function.
/// Provides a couple useful utility functions.
pub fn PRead(comptime T: type) type {
comptime std.debug.assert(std.meta.hasFn(T, "pread"));
return struct {
const Self = @This();
rdr: T,
offset: u64,
pub fn init(rdr: T, offset: u64) Self {
return .{
.rdr = rdr,
.offset = offset,
};
}
pub fn pread(self: Self, buf: []u8, offset: u64) !usize {
return self.rdr.pread(buf, self.offset + offset);
}
pub fn readerAt(self: Self, offset: u64) ToRead(T) {
return .init(self.rdr, self.offset + offset);
}
};
}
+43
View File
@@ -0,0 +1,43 @@
const std = @import("std");
pub fn ToRead(comptime T: type) type {
comptime std.debug.assert(std.meta.hasFn(T, "pread"));
return struct {
const Self = @This();
pub const Error = anyerror;
rdr: T,
offset: u64,
pub fn init(rdr: T, init_offset: u64) Self {
return .{
.rdr = rdr,
.offset = init_offset,
};
}
pub fn read(self: *Self, buf: []u8) !usize {
const red = try self.rdr.pread(buf, self.offset);
self.offset += red;
return red;
}
pub fn readAll(self: *Self, buf: []u8) !usize {
var cur_red = try self.read(buf);
if (cur_red == 0) return cur_red;
var res: usize = 0;
while (cur_red < buf.len) {
res = try self.read(buf[cur_red..]);
if (res == 0) break;
cur_red += res;
}
return cur_red;
}
const Reader = std.io.GenericReader(*Self, anyerror, read);
pub fn reader(self: anytype) Reader {
return .{
.context = @constCast(self),
};
}
};
}
+62 -2
View File
@@ -1,2 +1,62 @@
pub const Archive = @import("archive.zig"); const std = @import("std");
pub const ExtractionOptions = @import("options.zig");
pub const SfsReader = @import("reader.zig").SfsReader;
pub const ExtractionOptions = @import("extract_options.zig");
pub const SfsFile = SfsReader(std.fs.File);
const test_archive = "testing/LinuxPATest.sfs";
test "OpenFile" {
const sfs_fil = try std.fs.cwd().openFile(test_archive, .{});
defer sfs_fil.close();
var rdr: SfsFile = try .init(std.testing.allocator, sfs_fil, 0);
defer rdr.deinit();
_ = try rdr.frag_table.get(rdr.super.frag_count - 1);
_ = try rdr.id_table.get(rdr.super.id_count - 1);
_ = try rdr.export_table.get(rdr.super.inode_count - 1);
std.debug.print("{}\n", .{rdr.super});
var root = try rdr.root();
defer root.deinit();
var iter = root.iterate();
while (true) {
var f = try iter.next();
if (f == null) break;
defer f.?.deinit();
std.debug.print("{s}\n", .{f.?.name});
}
std.debug.print("Finished OpenFile test\n", .{});
}
test "ExtractSingleFile" {
const single_file = "PortableApps/Notepad++Portable/App/Notepad++/doLocalConf.xml";
const single_file_extr_loc = "testing/doLocalConf.xml";
std.fs.cwd().deleteFile(single_file_extr_loc) catch {};
const sfs_fil = try std.fs.cwd().openFile(test_archive, .{});
defer sfs_fil.close();
var rdr: SfsFile = try .init(std.testing.allocator, sfs_fil, 0);
defer rdr.deinit();
var fil = try rdr.open(single_file);
defer fil.deinit();
var op: ExtractionOptions = try .init();
op.verbose = true;
try fil.extract(op, single_file_extr_loc);
std.debug.print("Finished ExtractSingleFile test\n", .{});
}
test "ExtractAll" {
const extr_dir = "testing/testExtract";
std.fs.cwd().deleteTree(extr_dir) catch {};
const sfs_fil = try std.fs.cwd().openFile(test_archive, .{});
defer sfs_fil.close();
var rdr: SfsFile = try .init(std.testing.allocator, sfs_fil, 0);
defer rdr.deinit();
const op: ExtractionOptions = try .init();
// op.verbose = true;
try rdr.extract(op, extr_dir);
std.debug.print("Finished ExtractAll test\n", .{});
}
+79
View File
@@ -0,0 +1,79 @@
const std = @import("std");
const InodeRef = @import("inode.zig").Ref;
pub const Superblock = packed struct {
magic: u32,
inode_count: u32,
mod_time: u32,
block_size: u32,
frag_count: u32,
comp: Compression,
block_log: u16,
flags: packed struct {
_: u4,
id_uncomp: bool,
comp_options: bool,
no_xattr: bool,
xattr_uncomp: bool,
has_export: bool,
de_dupe: bool,
frag_always: bool,
no_frag: bool,
frag_uncomp: bool,
check: bool,
data_uncomp: bool,
inode_uncomp: bool,
},
id_count: u16,
ver_maj: u16,
ver_min: u16,
root_ref: InodeRef,
size: u64,
id_start: u64,
xattr_start: u64,
inode_start: u64,
dir_start: u64,
frag_start: u64,
export_start: u64,
};
pub const DecompressError = error{
LzoUnavailable,
Lz4Unavailable,
};
pub const Compression = enum(u16) {
gzip = 1,
lzma,
lzo,
xz,
lz4,
zstd,
pub fn decompress(self: Compression, comptime max_size: u32, alloc: std.mem.Allocator, source: anytype, dest: []u8) !usize {
switch (self) {
.gzip => {
var decomp = std.compress.zlib.decompressor(source);
return decomp.read(dest);
},
.lzma => {
var decomp = try std.compress.lzma.decompress(alloc, source);
defer decomp.deinit();
return decomp.read(dest);
},
.lzo => return DecompressError.LzoUnavailable,
.xz => {
var decomp = try std.compress.xz.decompress(alloc, source);
defer decomp.deinit();
return decomp.read(dest);
},
.lz4 => return DecompressError.Lz4Unavailable,
.zstd => {
var window: [max_size]u8 = undefined;
var decomp = std.compress.zstd.decompressor(source, .{ .window_buffer = &window });
return decomp.read(dest);
},
}
}
};
+75
View File
@@ -0,0 +1,75 @@
const std = @import("std");
const PRead = @import("reader/p_read.zig").PRead;
const Compression = @import("superblock.zig").Compression;
const MetadataReader = @import("reader/metadata.zig").MetadataReader;
pub const TableError = error{
InvalidIndex,
};
pub fn Table(comptime T: type, comptime R: type) type {
comptime std.debug.assert(std.meta.hasFn(R, "pread"));
return struct {
const Self = @This();
alloc: std.mem.Allocator,
rdr: PRead(R),
comp: Compression,
offset: u64,
table_count: u32,
mut: std.Thread.RwLock = .{},
table: []T = &[0]T{},
pub fn init(alloc: std.mem.Allocator, rdr: PRead(R), comp: Compression, offset: u64, table_count: u32) Self {
return .{
.alloc = alloc,
.rdr = rdr,
.comp = comp,
.offset = offset,
.table_count = table_count,
};
}
pub fn deinit(self: Self) void {
self.alloc.free(self.table);
}
fn resize(self: *Self, to_add: usize) !void {
if (!self.alloc.resize(self.table, self.table.len + to_add)) {
const new_table = try self.alloc.alloc(T, self.table.len + to_add);
@memcpy(new_table[0..self.table.len], self.table);
self.alloc.free(self.table);
self.table = new_table;
}
}
pub fn get(self: *Self, idx: u32) !T {
if (idx >= self.table_count) return TableError.InvalidIndex;
self.mut.lockShared();
defer self.mut.unlockShared();
if (idx >= self.table.len) {
return self.getAndFill(idx);
}
return self.table[idx];
}
fn getAndFill(self: *Self, idx: u32) !T {
self.mut.unlockShared();
defer self.mut.lockShared();
self.mut.lock();
defer self.mut.unlock();
var to_read: usize = 0;
var offset: u64 = 0;
while (idx >= self.table.len) {
to_read = @min(self.table_count - self.table.len, comptime 8192 / @sizeOf(T));
try self.resize(to_read);
_ = try self.rdr.pread(std.mem.asBytes(&offset), self.offset);
self.offset += 8;
var meta: MetadataReader(R) = .init(self.alloc, self.comp, self.rdr, offset);
_ = try meta.read(std.mem.sliceAsBytes(self.table[self.table.len - to_read ..]));
}
return self.table[idx];
}
};
}
-83
View File
@@ -1,83 +0,0 @@
const std = @import("std");
const Io = std.Io;
const io = std.testing.io;
const alloc = std.testing.allocator;
const stuff = @import("builtin");
const Archive = @import("archive.zig");
const Superblock = Archive.Superblock;
const TestArchive = "testing/LinuxPATest.sfs";
test "Basics" {
var fil = try Io.Dir.cwd().openFile(io, TestArchive, .{});
defer fil.close(io);
var sfs: Archive = try .init(io, fil, 0);
try std.testing.expectEqualDeep(sfs.super, LinuxPATestCorrectSuperblock);
const root_file = try sfs.root(alloc, io);
defer root_file.deinit();
}
const TestFile = "Start.exe";
const TestFileExtractLocation = "testing/Start.exe";
test "ExtractSingleFile" {
Io.Dir.cwd().deleteFile(io, TestFileExtractLocation) catch {};
var fil = try Io.Dir.cwd().openFile(io, TestArchive, .{});
defer fil.close(io);
var sfs: Archive = try .init(io, fil, 0);
var test_fil = try sfs.open(alloc, io, TestFile);
defer test_fil.deinit();
try test_fil.extract(alloc, io, TestFileExtractLocation, try .Default());
//TODO: validate extracted file.
}
const TestFullExtractLocation = "testing/TestExtract";
test "ExtractCompleteArchive" {
Io.Dir.cwd().deleteTree(io, TestFullExtractLocation) catch {};
var fil = try Io.Dir.cwd().openFile(io, TestArchive, .{});
defer fil.close(io);
var sfs: Archive = try .init(io, fil, 0);
try sfs.extract(alloc, io, TestFullExtractLocation, try .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,
};
-119
View File
@@ -1,119 +0,0 @@
//! The DataExtractor is meant to extract a regular file's data to a given file asyncronously.
const std = @import("std");
const Io = std.Io;
const FragEntry = @import("../frag.zig").FragEntry;
const BlockSize = @import("../inode_data/file.zig").BlockSize;
const Decompressor = @import("decompressor.zig");
const OffsetFile = @import("offset_file.zig");
const SharedCache = @import("shared_cache.zig");
const DataExtractor = @This();
fil: OffsetFile,
cache: *SharedCache,
decomp: *const Decompressor,
block_size: u32,
file_size: u64,
start: u64,
blocks: []BlockSize,
frag_offset: u32 = 0,
frag_entry: ?FragEntry = null,
pub fn init(fil: OffsetFile, cache: *SharedCache, decomp: *const Decompressor, block_size: u32, file_size: u64, data_start: u64, blocks: []BlockSize) DataExtractor {
return .{
.fil = fil,
.cache = cache,
.decomp = decomp,
.block_size = block_size,
.file_size = file_size,
.start = data_start,
.blocks = blocks,
};
}
pub fn addFrag(self: *DataExtractor, frag_offset: u32, entry: FragEntry) void {
self.frag_offset = frag_offset;
self.frag_entry = entry;
}
fn numBlocks(self: DataExtractor) usize {
var num = self.blocks.len;
if (self.frag_entry != null) num += 1;
return num;
}
/// Starts extracting the data using the given group to spawn async tasks.
pub fn extractAsync(self: DataExtractor, alloc: std.mem.Allocator, io: Io, group: *Io.Group, fil: Io.File) void {
var read_offset: u64 = self.start;
for (0..self.blocks.len) |idx| {
group.async(io, blockThread, .{ self, alloc, io, fil, read_offset, idx });
read_offset += self.blocks[idx].size;
}
if (self.frag_entry != null)
group.async(io, fragThread, .{ self, alloc, io, fil });
}
fn blockThread(self: DataExtractor, alloc: std.mem.Allocator, io: Io, fil: Io.File, read_offset: u64, idx: u32) !void {
const block = self.blocks[idx];
const cur_block_size = if (idx == self.numBlocks() - 1)
self.file_size % self.block_size
else
self.block_size;
var wrt = fil.writer(io, &[0]u8{});
try wrt.seekTo(self.block_size * idx);
defer wrt.flush() catch {};
if (block.size == 0) {
try wrt.interface.splatByteAll(0, cur_block_size);
return;
}
var rdr = try self.fil.readerAt(io, read_offset, &[0]u8{});
if (block.uncompressed) {
try rdr.interface.streamExact(&wrt, cur_block_size);
return;
} else {
@branchHint(.likely);
var cache = try self.cache.getCache(io);
defer self.cache.returnCache(cache);
var tmp = try self.cache.getCache(io);
defer self.cache.returnCache(tmp);
try rdr.interface.readSliceAll(cache.cache[0..block.size]);
_ = try self.decomp.Decompress(alloc, cache.cache[0..block.size], tmp.cache[0..cur_block_size]);
try wrt.interface.writeAll(tmp.cache[0..cur_block_size]);
}
}
fn fragThread(self: DataExtractor, alloc: std.mem.Allocator, io: Io, fil: Io.File) !void {
const frag = self.frag_entry.?;
const cur_block_size = self.file_size % self.block_size;
var wrt = fil.writer(io, &[0]u8{});
try wrt.seekTo(self.blocks.len * self.block_size);
defer wrt.flush() catch {};
var rdr = try self.fil.readerAt(io, frag.start, &[0]u8{});
if (frag.size.uncompressed) {
try rdr.interface.discardAll(self.frag_offset);
try rdr.interface.streamExact(&wrt, cur_block_size);
return;
} else {
@branchHint(.likely);
var cache = try self.cache.getCache(io);
defer self.cache.returnCache(cache);
var tmp = try self.cache.getCache(io);
defer self.cache.returnCache(tmp);
try rdr.interface.readSliceAll(cache.cache[0..frag.size.size]);
_ = try self.decomp.Decompress(alloc, cache.cache[0..frag.size.size], tmp.cache[0..self.block_size]);
try wrt.interface.writeAll(tmp.cache[0..cur_block_size]);
}
}
-209
View File
@@ -1,209 +0,0 @@
//! DataReader reads a regular file's data linearly from start to finish using Io.Reader interface.
const std = @import("std");
const Io = std.Io;
const Reader = Io.Reader;
const Writer = Io.Writer;
const Limit = Io.Limit;
const FragEntry = @import("../frag.zig").FragEntry;
const BlockSize = @import("../inode_data/file.zig").BlockSize;
const Decompressor = @import("decompressor.zig");
const OffsetFile = @import("offset_file.zig");
const SharedCache = @import("shared_cache.zig");
const DataReader = @This();
alloc: std.mem.Allocator,
fil: OffsetFile,
io: Io,
cache: *SharedCache,
decomp: *const Decompressor,
block_size: u32,
file_size: u64,
cur_offset: u64,
blocks: []BlockSize,
frag_offset: u32 = 0,
frag_entry: ?FragEntry = null,
block_idx: usize = 0,
sparse_block: bool = false,
interface: Io.Reader,
pub fn init(alloc: std.mem.Allocator, io: Io, fil: OffsetFile, cache: *SharedCache, decomp: *const Decompressor, block_size: u32, file_size: u64, data_start: u64, blocks: []BlockSize) !DataReader {
return .{
.alloc = alloc,
.fil = fil,
.io = io,
.decomp = decomp,
.block_size = block_size,
.file_size = file_size,
.cur_offset = data_start,
.blocks = blocks,
.interface = .{
.buffer = try cache.getCache(io),
.seek = 0,
.end = 0,
.vtable = &.{
.stream = stream,
.discard = discard,
.readVec = readVec,
},
},
};
}
pub fn deinit(self: *DataReader) void {
if (self.interface.buffer.len > 0) {
const buf_nod: *SharedCache.BufferNode = @fieldParentPtr("cache", self.interface.buffer);
self.cache.returnCache(buf_nod);
}
}
pub fn addFrag(self: *DataReader, frag_offset: u32, entry: FragEntry) void {
self.frag_offset = frag_offset;
self.frag_entry = entry;
}
fn numBlocks(self: DataReader) usize {
var num = self.blocks.len;
if (self.frag_entry != null) num += 1;
return num;
}
fn advanceBuffer(self: *DataReader) !void {
if (self.block_idx >= self.numBlocks()) {
return Reader.Error.EndOfStream;
}
defer self.block_idx += 1;
self.interface.end = if (self.block_idx == self.numBlocks() - 1)
self.size % self.block_size
else
self.block_size;
// Fragment
if (self.block_idx == self.blocks.len) {
const entry = self.frag_entry.?;
if (entry.size.uncompressed) {
var rdr = try self.fil.readerAt(self.io, entry.start + self.frag_offset, &[0]u8{});
try rdr.interface.readSliceAll(self.interface.buffer[0..self.interface.end]);
} else {
@branchHint(.likely);
const tmp = try self.cache.getCache(self.io);
defer self.cache.returnCache(tmp);
var rdr = try self.fil.readerAt(self.io, entry.start, &[0]u8{});
try rdr.interface.readSliceAll(tmp.cache[0..entry.size.size]);
_ = try self.decomp.Decompress(self.alloc, tmp.cache[0..entry.size.size], self.interface.buffer[0..self.block_size]);
@memmove(self.interface.buffer[0..self.interface.end], self.interface.buffer[self.frag_offset .. self.frag_offset + self.interface.end]);
}
self.interface.seek = 0;
return;
}
// Normal Block
const block = self.blocks[self.block_idx];
if (block.size == 0) {
self.interface.seek = 0;
self.sparse_block = true;
return;
} else {
self.sparse_block = false;
}
if (block.uncompressed) {
try self.fil.readAt(self.io, self.cur_offset, self.interface.buffer[0..self.interface.end]);
self.cur_offset += self.interface.end;
} else {
@branchHint(.likely);
const tmp = try self.cache.getCache(self.io);
defer self.cache.returnCache(tmp);
var rdr = try self.fil.readerAt(self.io, self.cur_offset, &[0]u8{});
try rdr.interface.readSliceAll(tmp.cache[0..block.size]);
self.cur_offset += block.size;
_ = try self.decomp.Decompress(self.alloc, tmp.cache[0..block.size], self.interface.buffer[0..self.interface.end]);
}
self.interface.seek = 0;
}
fn stream(rdr: *Reader, wrt: *Writer, limit: Limit) Reader.StreamError!usize {
var data: *DataReader = @fieldParentPtr("interface", rdr);
if (rdr.seek == rdr.end)
data.advanceBuffer() catch |err| return switch (err) {
error.ReadFailed => error.ReadFailed,
error.EndOfStream => error.EndOfStream,
else => error.ReadFailed,
};
switch (limit) {
.nothing => return 0,
.unlimited => {
const wrote = if (data.sparse_block)
try wrt.splatByte(0, rdr.end - rdr.seek)
else
try wrt.write(rdr.buffer[rdr.seek..rdr.end]);
rdr.seek += wrote;
return wrote;
},
else => {
const to_read = @min(rdr.end - rdr.seek, @intFromEnum(limit));
const wrote = if (data.sparse_block)
try wrt.splatByte(0, to_read)
else
try wrt.write(rdr.buffer[rdr.seek .. rdr.seek + to_read]);
rdr.seek += wrote;
return wrote;
},
}
}
fn discard(rdr: *Reader, limit: Limit) Reader.Error!usize {
var data: *DataReader = @fieldParentPtr("interface", rdr);
if (rdr.seek == rdr.end)
data.advanceBuffer() catch |err| return switch (err) {
error.ReadFailed => error.ReadFailed,
error.EndOfStream => error.EndOfStream,
else => error.ReadFailed,
};
switch (limit) {
.nothing => return 0,
.unlimited => {
const adv = rdr.end - rdr.seek;
rdr.seek = rdr.end;
return adv;
},
else => {
const adv = @min(rdr.end - rdr.seek, @intFromEnum(limit));
rdr.seek += adv;
return adv;
},
}
}
fn readVec(rdr: *Reader, vec: [][]u8) Reader.Error!usize {
var data: *DataReader = @fieldParentPtr("interface", rdr);
if (rdr.seek == rdr.end)
data.advanceBuffer() catch |err| return switch (err) {
error.ReadFailed => error.ReadFailed,
error.EndOfStream => error.EndOfStream,
else => error.ReadFailed,
};
var wrote: usize = 0;
for (vec) |buf| {
if (rdr.seek == rdr.end) break;
const to_copy = @min(rdr.end - rdr.seek, buf.len);
if (data.sparse_block)
@memset(buf[0..to_copy], 0)
else
@memcpy(buf[0..to_copy], rdr.buffer[rdr.seek .. rdr.seek + to_copy]);
rdr.seek += to_copy;
wrote += to_copy;
}
return wrote;
}
-15
View File
@@ -1,15 +0,0 @@
//! A decompression interface
const std = @import("std");
const Decompressor = @This();
pub const Error = std.Io.Reader.StreamError || std.mem.Allocator.Error;
/// The actual decompression function.
/// If the given decompressor is null, then the decompression should be done "stateless" without lasting allocations.
decomp_fn: *const fn (?*const Decompressor, std.mem.Allocator, in: []u8, out: []u8) Error!usize,
pub fn Decompress(self: *const Decompressor, alloc: std.mem.Allocator, in: []u8, out: []u8) Error!usize {
return self.decomp_fn(self, alloc, in, out);
}
-103
View File
@@ -1,103 +0,0 @@
const std = @import("std");
const Reader = std.Io.Reader;
const Writer = std.Io.Writer;
const Limit = std.Io.Limit;
const StreamError = std.Io.Reader.StreamError;
const Decompressor = @import("decompressor.zig");
const BlockHeader = packed struct(u16) {
size: u15,
uncompressed: bool,
};
const This = @This();
alloc: std.mem.Allocator,
rdr: *Reader,
decomp: *const Decompressor,
cur_block_start: u32 = 0,
next_start_start: u32 = 0,
buf: [8192]u8 = undefined,
err: ?anyerror = null,
interface: Reader = .{
.buffer = &[0]u8{},
.end = 0,
.seek = 0,
.vtable = &.{
.stream = stream,
.discard = discard,
.readVec = readVec,
// TODO: Potentially add rebase so that we can guarentee that self.block_start & interface.seek is correct.
},
},
pub fn init(alloc: std.mem.Allocator, rdr: *Reader, decomp: *const Decompressor) This {
return .{
.alloc = alloc,
.rdr = rdr,
.decomp = decomp,
};
}
fn advance(self: *This) !void {
self.interface.seek = 0;
var hdr: BlockHeader = undefined;
try self.rdr.readSliceEndian(BlockHeader, @ptrCast(&hdr), .little);
self.cur_block_start = self.next_start_start;
self.next_start_start += hdr.size;
if (hdr.uncompressed) {
try self.rdr.readSliceEndian(u8, self.buf[0..hdr.size], .little);
self.interface.end = hdr.size;
self.interface.buffer = self.buf[0..hdr.size];
return;
} else {
@branchHint(.likely);
var tmp_buf: [8192]u8 = undefined;
try self.rdr.readSliceAll(tmp_buf[0..hdr.size]);
self.interface.end = try self.decomp.Decompress(self.alloc, tmp_buf[0..hdr.size], &self.buf);
self.interface.buffer = self.buf[0..self.interface.end];
}
}
fn stream(rdr: *Reader, wrt: *Writer, limit: Limit) StreamError!usize {
const self: *This = @fieldParentPtr("interface", rdr);
if (rdr.end == rdr.seek) self.advance() catch |err| {
self.err = err;
return StreamError.ReadFailed;
};
if (@intFromEnum(limit) == 0) return 0;
const to_write = @min(rdr.end - rdr.seek, @intFromEnum(limit));
const wrote = try wrt.write(self.buf[rdr.seek .. rdr.seek + to_write]);
self.interface.seek += wrote;
return wrote;
}
fn discard(rdr: *Reader, limit: Limit) Reader.Error!usize {
const self: *This = @fieldParentPtr("interface", rdr);
if (rdr.end == rdr.seek) self.advance() catch |err| {
self.err = err;
return error.ReadFailed;
};
if (@intFromEnum(limit) == 0) return 0;
const to_skip = @min(rdr.end - rdr.seek, @intFromEnum(limit));
rdr.seek += to_skip;
return to_skip;
}
fn readVec(rdr: *Reader, vec: [][]u8) Reader.Error!usize {
const self: *This = @fieldParentPtr("interface", rdr);
if (rdr.end == rdr.seek) self.advance() catch |err| {
self.err = err;
return error.ReadFailed;
};
var cur_red: usize = 0;
for (vec) |s| {
const to_copy: usize = @min(rdr.end - rdr.seek, s.len);
@memcpy(s[0..to_copy], self.buf[rdr.seek .. rdr.seek + to_copy]);
rdr.seek += to_copy;
cur_red += to_copy;
if (rdr.end == rdr.seek) break;
}
return cur_red;
}
-25
View File
@@ -1,25 +0,0 @@
//! Miscellaneous utility functions.
const std = @import("std");
const Io = std.Io;
const Inode = @import("../inode.zig");
const Decompressor = @import("decompressor.zig");
const MetadataReader = @import("metadata.zig");
const OffsetFile = @import("offset_file.zig");
/// check is the path is referencing itself ("" or ".").
/// separators must be trimmed before calling this function for it to work properly.
pub fn pathIsSelf(path: []const u8) bool {
if (path.len == 0) return true;
if (path.len > 1) return false;
return path[0] == '.';
}
/// Creates an Inode from an Inode.Ref.
pub fn inodeFromRef(alloc: std.mem.Allocator, io: Io, file: OffsetFile, decomp: *const Decompressor, inode_start: u64, block_size: u32, ref: Inode.Ref) !Inode {
var rdr = try file.readerAt(io, inode_start + ref.block_start, &[0]u8{});
var meta: MetadataReader = .init(alloc, &rdr.interface, decomp);
try meta.interface.discardAll(ref.block_offset);
return .read(alloc, &meta.interface, block_size);
}
-29
View File
@@ -1,29 +0,0 @@
//! A File where it's meaningful (to us) content starts at a given offset.
const std = @import("std");
const Io = std.Io;
const File = Io.File;
const Reader = File.Reader;
const OffsetFile = @This();
fil: File,
offset: u64,
pub fn init(fil: File, init_offset: u64) OffsetFile {
return .{ .fil = fil, .offset = init_offset };
}
pub fn readerAt(self: OffsetFile, io: Io, offset: u64, buffer: []u8) !Reader {
var rdr = self.fil.reader(io, buffer);
try rdr.seekTo(self.offset + offset);
return rdr;
}
pub fn readAt(self: OffsetFile, io: Io, offset: u64, buf: []u8) !void {
_ = try self.fil.readPositionalAll(io, buf, self.offset + offset);
}
pub fn readValueAt(self: OffsetFile, comptime T: anytype, io: Io, offset: u64) !void {
//TODO: check for endianess and decode accordingly.
var new: T = undefined;
_ = try self.fil.readPositionalAll(io, @ptrCast(&new), self.offset + offset);
}
-52
View File
@@ -1,52 +0,0 @@
const std = @import("std");
const Io = std.Io;
const Node = std.SinglyLinkedList.Node;
const SharedCache = @This();
pub const CACHE_SIZE = 1024 * 1024;
pub const BufferNode = struct {
node: Node,
cache: [CACHE_SIZE]u8,
};
alloc: std.mem.Allocator,
caches: std.ArrayList(BufferNode),
cache_queue: std.SinglyLinkedList,
queue_mut: Io.Mutex,
pub fn init(alloc: std.mem.Allocator, init_cache_size: u32) !SharedCache {
const caches: std.ArrayList(BufferNode) = try .initCapacity(alloc, init_cache_size);
var queue: std.SinglyLinkedList = .{};
for (caches.items) |item|
queue.prepend(&item.node);
return .{
.alloc = alloc,
.caches = caches,
.cache_queue = queue,
};
}
pub fn deinit(self: *SharedCache) void {
self.caches.deinit(self.alloc);
}
pub fn getCache(self: *SharedCache, io: Io) !*BufferNode {
self.queue_mut.lock(io);
const nxt = self.cache_queue.popFirst();
self.queue_mut.unlock(io);
if (nxt == null) {
const new = try self.caches.addOne(self.alloc);
new.* = .{
.node = .{},
.cache = undefined,
};
return new;
}
return @fieldParentPtr("node", nxt.?);
}
pub fn returnCache(self: *SharedCache, buf: *BufferNode) void {
self.cache_queue.prepend(buf);
}
-284
View File
@@ -1,284 +0,0 @@
const std = @import("std");
const Io = std.Io;
const InodeRef = @import("inode.zig").Ref;
const LookupTable = @import("lookup_table.zig");
const Decompressor = @import("util/decompressor.zig");
const MetadataReader = @import("util/metadata.zig");
const OffsetFile = @import("util/offset_file.zig");
const XattrCachedTable = @This();
alloc: std.mem.Allocator,
fil: OffsetFile,
decomp: *const Decompressor,
kv_start: u64,
table: LookupTable.CachedTable(TableValue),
value_cache: std.AutoHashMap(InodeRef, []const u8),
value_mut: Io.Mutex,
pub fn init(alloc: std.mem.Allocator, io: Io, fil: OffsetFile, decomp: *const Decompressor, xattr_start: u64) !XattrCachedTable {
var rdr = try fil.readerAt(io, xattr_start, &[0]u8{});
var start: u64 = undefined;
try rdr.interface.readSliceEndian(u64, @ptrCast(&start), .little);
var num: u32 = undefined;
try rdr.interface.readSliceEndian(u32, @ptrCast(&num), .little);
return .{
.alloc = alloc,
.fil = fil,
.decomp = decomp,
.kv_start = start,
.table = .init(alloc, fil, xattr_start + 16, num),
.value_cache = .init(alloc),
};
}
pub fn deinit(self: *XattrCachedTable) void {
self.table.deinit();
self.value_cache.deinit();
}
pub fn get(self: *XattrCachedTable, alloc: std.mem.Allocator, io: Io, idx: u32) ![]XattrSemiOwned {
const lookup = try self.table.get(io, idx);
var rdr = try self.fil.readerAt(io, self.kv_start + lookup.ref.block_start, &[0]u8{});
var meta: MetadataReader = .init(alloc, &rdr.interface, self.decomp);
try meta.interface.discardAll(lookup.ref.block_offset);
const out = try alloc.alloc(XattrSemiOwned, lookup.count);
errdefer alloc.free(out);
for (0..lookup.count) |i| {
const key_entry: KeyEntry = undefined;
try meta.interface.readSliceEndian(KeyEntry, @ptrCast(&key_entry), .little);
const key = switch (key_entry.type.namespace) {
.user => blk: {
const tmp = try alloc.alloc(u8, key_entry.name_size + 1 + 5);
errdefer alloc.free(tmp);
try meta.interface.readSliceEndian(u8, tmp[5 .. tmp.len - 1], .little);
@memset(tmp[0..5], "user.");
break :blk tmp;
},
.trusted => blk: {
const tmp = try alloc.alloc(u8, key_entry.name_size + 1 + 8);
errdefer alloc.free(tmp);
try meta.interface.readSliceEndian(u8, tmp[8 .. tmp.len - 1], .little);
@memset(tmp[0..8], "trusted.");
break :blk tmp;
},
.security => blk: {
const tmp = try alloc.alloc(u8, key_entry.name_size + 1 + 9);
errdefer alloc.free(tmp);
try meta.interface.readSliceEndian(u8, tmp[9 .. tmp.len - 1], .little);
@memset(tmp[0..9], "security.");
break :blk tmp;
},
};
key[key.len - 1] = 0;
errdefer alloc.free(key);
if (key_entry.type.out_of_line) {
const value: ValueOutOfLineEntry = undefined;
try meta.interface.readSliceEndian(ValueOutOfLineEntry, @ptrCast(&value), .little);
out[i] = .{
.key = key,
.value = try self.valueAt(io, value.ref),
};
continue;
}
const val_ref: InodeRef = .{ .block_start = meta.cur_block_start, .block_offset = meta.interface.seek };
try self.value_mut.lock(io);
defer self.value_mut.unlock(io);
if (self.value_cache.contains(val_ref)) {
out[i] = .{
.key = key,
.value = try self.valueAt(io, val_ref),
};
continue;
}
var val_size: u32 = undefined;
try meta.interface.readSliceEndian(val_size, @ptrCast(&val_size), .little);
const val = try self.alloc.alloc(u8, val_size);
errdefer alloc.free(val);
try meta.interface.readSliceEndian(u8, val, .little);
try self.value_cache.put(val_ref, val);
out[i] = .{
.key = key,
.value = val,
};
}
return out;
}
fn valueAt(self: *XattrCachedTable, io: Io, ref: InodeRef) ![]const u8 {
try self.value_mut.lock(io);
defer self.value_mut.unlock(io);
if (self.value_cache.contains(ref)) return self.value_cache.get(ref).?;
var rdr = try self.fil.readerAt(io, self.kv_start + ref.block_start, &[0]u8{});
var meta: MetadataReader = .init(self.alloc, &rdr.interface, self.decomp);
try meta.interface.discardAll(ref.block_offset);
var val_size: u32 = undefined;
try meta.interface.readSliceEndian(val_size, @ptrCast(&val_size), .little);
const val = try self.alloc.alloc(u8, val_size);
errdefer self.alloc.free(val);
try meta.interface.readSliceEndian(u8, val, .little);
try self.value_cache.put(ref, val);
return val;
}
// Types
/// An Xattr return value where the reciever only owns the key value.
pub const XattrSemiOwned = struct {
key: [:0]const u8,
value: []const u8,
pub fn deinit(self: XattrSemiOwned, alloc: std.mem.Allocator) void {
alloc.free(self.key);
}
};
/// An Xattr return value where the reciever owns both the key & value.
pub const XattrOwned = struct {
key: [:0]const u8,
value: []const u8,
pub fn deinit(self: XattrSemiOwned, alloc: std.mem.Allocator) void {
alloc.free(self.key);
alloc.free(self.value);
}
};
const TableValue = extern struct {
ref: InodeRef,
count: u32,
size: u32,
};
const KeyEntry = extern struct {
type: XattrPrefix,
name_size: u16,
};
const ValueOutOfLineEntry = extern struct {
_: u32,
ref: InodeRef,
};
const XattrPrefix = packed struct(u16) {
namespace: enum(u8) {
user,
trusted,
security,
fn prefixSize(self: @This()) u16 {
return switch (self) {
.user => 5,
.trusted => 8,
.security => 9,
};
}
},
out_of_line: bool,
_: u7,
};
// Stateless
pub fn statelessLookup(alloc: std.mem.Allocator, io: Io, decomp: *const Decompressor, fil: OffsetFile, table_start: u64, idx: u16) ![]XattrOwned {
var rdr = try fil.readerAt(io, table_start, &[0]u8{});
var kv_start: u64 = undefined;
try rdr.interface.readSliceEndian(u64, @ptrCast(&kv_start), .little);
const lookup = try LookupTable.lookupValue(TableValue, alloc, io, decomp, fil, table_start + 16, idx);
rdr = try fil.readerAt(io, kv_start + lookup.ref.block_start, &[0]u8{});
var meta: MetadataReader = .init(alloc, &rdr.interface, decomp);
try meta.interface.discardAll(lookup.ref.block_offset);
const out = try alloc.alloc(XattrOwned, lookup.count);
errdefer alloc.free(out);
for (0..lookup.count) |i| {
const key_entry: KeyEntry = undefined;
try meta.interface.readSliceEndian(KeyEntry, @ptrCast(&key_entry), .little);
const key = switch (key_entry.type.namespace) {
.user => blk: {
const tmp = try alloc.alloc(u8, key_entry.name_size + 1 + 5);
errdefer alloc.free(tmp);
try meta.interface.readSliceEndian(u8, tmp[5 .. tmp.len - 1], .little);
@memset(tmp[0..5], "user.");
break :blk tmp;
},
.trusted => blk: {
const tmp = try alloc.alloc(u8, key_entry.name_size + 1 + 8);
errdefer alloc.free(tmp);
try meta.interface.readSliceEndian(u8, tmp[8 .. tmp.len - 1], .little);
@memset(tmp[0..8], "trusted.");
break :blk tmp;
},
.security => blk: {
const tmp = try alloc.alloc(u8, key_entry.name_size + 1 + 9);
errdefer alloc.free(tmp);
try meta.interface.readSliceEndian(u8, tmp[9 .. tmp.len - 1], .little);
@memset(tmp[0..9], "security.");
break :blk tmp;
},
};
key[key.len - 1] = 0;
errdefer alloc.free(key);
if (key_entry.type.out_of_line) {
const value: ValueOutOfLineEntry = undefined;
try meta.interface.readSliceEndian(ValueOutOfLineEntry, @ptrCast(&value), .little);
var ool_rdr = try fil.readerAt(io, kv_start + value.ref.block_start, &[0]u8{});
var ool_meta: MetadataReader = .init(alloc, &ool_rdr.interface, decomp);
try ool_meta.interface.discardAll(value.ref.block_offset);
var val_size: u32 = undefined;
try ool_meta.interface.readSliceEndian(val_size, @ptrCast(&val_size), .little);
const val = try alloc.alloc(u8, val_size);
errdefer alloc.free(val);
try ool_meta.interface.readSliceEndian(u8, val, .little);
out[i] = .{
.key = key,
.value = val,
};
continue;
}
var val_size: u32 = undefined;
try meta.interface.readSliceEndian(val_size, @ptrCast(&val_size), .little);
const val = try alloc.alloc(u8, val_size);
errdefer alloc.free(val);
try meta.interface.readSliceEndian(u8, val, .little);
out[i] = .{
.key = key,
.value = val,
};
}
return out;
}