Further progress in parsing format

src/bin/unsquashfs.zig

@@ -1,3 +1,132 @@
-const import = @import("std");
+const std = @import("std");
+const config = @import("config");
+const squashfs = @import("squashfs");
 
-pub fn main() void {}
+const help_msg =
+    \\Basic Usage: zig-unsquashfs [Options] SQUASHFS_FILE <EXTRACT_LOCATION>
+    \\
+    \\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.
+    \\
+    \\Extraction options:
+    \\  --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.
+    \\
+    \\Listing Options:
+    \\  -l                  List files instead of extracting. When used, you do not need to specify an extraction location.
+    \\  -ll                 Similiar to -l, but with file attributes.
+    \\  -lln                Similiar to -ll, but with numeric uids and gids.
+    \\
+    \\Other:
+    \\  --help              Prints this help message.
+    \\  -h                  Same as --help
+    \\  --version           Print version number.
+    \\
+;
+
+const stdout = std.io.getStdOut();
+
+var extr_files: std.ArrayList([]const u8) = undefined;
+var offset: u64 = 0;
+var verbose: bool = false;
+var unbreak: bool = false;
+var deref: bool = false;
+var processors: u16 = 0;
+var list: ListTypes = .None;
+
+var filename: []const u8 = "";
+var extr_location: []const u8 = "";
+
+const ListTypes = enum {
+    None,
+    List,
+    ListAttr,
+    ListNumeric,
+};
+
+pub fn main() !void {
+    var alloc: std.heap.GeneralPurposeAllocator(.{}) = .init;
+    extr_files = .init(alloc.allocator());
+    defer extr_files.deinit();
+    var args = std.process.argsWithAllocator(alloc.allocator()) catch {
+        _ = try stdout.writeAll("Unable to allocate memory");
+        return;
+    };
+    defer args.deinit();
+    _ = args.next();
+    while (args.next()) |arg| {
+        if (std.mem.eql(u8, arg, "--help") or std.mem.eql(u8, arg, "-h")) {
+            _ = try stdout.writeAll(help_msg);
+            return;
+        } else if (std.mem.eql(u8, arg, "--version")) {
+            try config.version.format("", .{}, stdout.writer());
+            _ = try stdout.write("\n");
+            return;
+        } else if (std.mem.eql(u8, arg, "-v")) {
+            verbose = true;
+        } else if (std.mem.eql(u8, arg, "--unbreak-symlinks") or std.mem.eql(u8, arg, "-us")) {
+            unbreak = true;
+        } else if (std.mem.eql(u8, arg, "--deref-symlinks") or std.mem.eql(u8, arg, "-ds")) {
+            deref = true;
+        } else if (std.mem.eql(u8, arg, "-l")) {
+            list = .List;
+        } else if (std.mem.eql(u8, arg, "-ll")) {
+            list = .ListAttr;
+        } else if (std.mem.eql(u8, arg, "-lln")) {
+            list = .ListNumeric;
+        } else if (std.mem.eql(u8, arg, "-e")) {
+            const next = args.next();
+            if (next == null) {
+                _ = try stdout.writeAll("path required after -e\n");
+                return;
+            }
+            try extr_files.append(next.?);
+        } else if (std.mem.eql(u8, arg, "-o")) {
+            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;
+        }
+    }
+    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, .{});
+    var rdr = squashfs.FileReader.init(
+        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
+}

src/directory.zig

@@ -0,0 +1,70 @@
+const std = @import("std");
+
+const InodeType = @import("inode.zig").Type;
+const Compression = @import("superblock.zig").Compression;
+
+const Header = extern struct { //use extern instead of packed, due to bit alignment
+    count: u32,
+    block: u32,
+    num: u32,
+};
+
+const RawEntry = struct {
+    offset: u16,
+    num_offset: i16,
+    type: InodeType,
+    size: u16,
+    name: []const u8,
+
+    pub fn init(alloc: std.mem.Allocator, rdr: anytype) !RawEntry {
+        const 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 = 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 {
+    const 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 = hdr.num + raw_ent.num_offset,
+                .type = raw_ent.type,
+                .name = raw_ent.name,
+            });
+        }
+    }
+    return entries.toOwnedSlice();
+}

src/fragment.zig

@@ -0,0 +1,7 @@
+const BlockSize = @import("inode/file.zig").BlockSize;
+
+pub const FragEntry = packed struct {
+    block: u64,
+    size: BlockSize,
+    _: u32,
+};

src/inode.zig

@@ -4,13 +4,17 @@ const dir = @import("inode/dir.zig");
 const file = @import("inode/file.zig");
 const misc = @import("inode/misc.zig");
 
+const ToRead = @import("reader/to_read.zig").ToRead;
+const Compression = @import("superblock.zig").Compression;
+const MetadataReader = @import("reader/metadata.zig").MetadataReader;
+
 pub const Ref = packed struct {
     offset: u16,
     block: u32,
     _: u16,
 };
 
-const Type = enum(u16) {
+pub const Type = enum(u16) {
     dir = 1,
     file,
     symlink,
@@ -27,7 +31,7 @@ const Type = enum(u16) {
     ext_socket,
 };
 
-const Header = packed struct {
+pub const Header = packed struct {
     type: Type,
     perm: u16,
     uid_idx: u16,
@@ -36,7 +40,7 @@ const Header = packed struct {
     num: u32,
 };
 
-const Data = union(enum) {
+pub const Data = union(enum) {
     dir: dir.Dir,
     file: file.File,
     symlink: misc.Symlink,
@@ -59,7 +63,6 @@ hdr: Header,
 data: Data,
 
 pub fn init(rdr: anytype, alloc: std.mem.Allocator, block_size: u32) !Self {
-    std.debug.assert(std.meta.hasFn(@TypeOf(rdr), "read"));
     var hdr: Header = undefined;
     _ = try rdr.read(std.mem.asBytes(&hdr));
     const data = switch (hdr.type) {
@@ -83,3 +86,10 @@ pub fn init(rdr: anytype, alloc: std.mem.Allocator, block_size: u32) !Self {
         .data = data,
     };
 }
+pub fn initFromRef(p_rdr: anytype, comp: Compression, ref: Ref, table_start: u64, alloc: std.mem.Allocator, block_size: u32) !Self {
+    const rdr: ToRead(@TypeOf(p_rdr)) = .init(p_rdr, ref.block + table_start);
+    const meta_rdr: MetadataReader(ToRead(@TypeOf(p_rdr))) = try .init(alloc, comp, rdr);
+    defer meta_rdr.deinit();
+    try meta_rdr.skip(ref.offset);
+    return init(meta_rdr, alloc, block_size);
+}

src/reader.zig

@@ -1,25 +1,53 @@
 const std = @import("std");
 
+const Inode = @import("inode.zig");
+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;
 
 pub fn Reader(comptime T: type) type {
-    std.debug.assert(std.meta.hasFn(T, "pread"));
+    comptime std.debug.assert(std.meta.hasFn(T, "pread"));
 
     return struct {
         const Self = @This();
 
         alloc: std.mem.Allocator,
-        rdr: T,
+        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,
+        root: ?Inode = null,
 
-        pub fn init(alloc: std.mem.Allocator, rdr: T) Self {
-            const out = Self{
+        pub fn init(alloc: std.mem.Allocator, rdr: T, offset: u64) !Self {
+            var out: Self = .{
                 .alloc = alloc,
-                .rdr = rdr,
+                .rdr = .init(rdr, offset),
             };
             _ = try rdr.pread(std.mem.asBytes(&out.super), 0);
+            out.frag_table = .init(alloc, rdr, out.super.frag_start, out.super.frag_count);
+            out.id_table = .init(alloc, rdr, out.super.id_start, out.super.id_count);
+            out.export_table = .init(alloc, rdr, 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();
+        }
+
+        /// Returns the inode with the given Inode Number.
+        /// Requires for the archive to have an export table.
+        pub fn InodeAt(self: Self, num: u32) !Inode {
+            const ref = try self.export_table.get(num - 1);
+            _ = ref;
+            return error{TODO}.TODO;
+        }
     };
 }

src/reader/metadata.zig

@@ -8,6 +8,7 @@ const MetaHeader = packed struct {
 };
 
 pub fn MetadataReader(comptime T: type) type {
+    comptime std.debug.assert(std.meta.hasFn(T, "read"));
     return struct {
         const Self = @This();
 
@@ -38,6 +39,31 @@ pub fn MetadataReader(comptime T: type) type {
                 std.io.limitedReader(self.rdr, hdr.size),
                 self.block,
             );
+            self.block_offset = 0;
+        }
+
+        pub fn skip(self: *Self, offset: u32) !void {
+            var skipped = 0;
+            var to_skip = 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 += to_read;
+            }
+            return cur_red;
         }
     };
 }

src/reader/p_read.zig

@@ -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);
+        }
+    };
+}

src/reader/to_read.zig

@@ -0,0 +1,24 @@
+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();
+
+        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;
+        }
+    };
+}

src/root.zig

@@ -0,0 +1,5 @@
+const std = @import("std");
+
+pub const Reader = @import("reader.zig").Reader;
+
+pub const FileReader = Reader(std.fs.File);

src/superblock.zig

@@ -1,5 +1,7 @@
 const std = @import("std");
 
+const InodeRef = @import("inode.zig").Ref;
+
 pub const Superblock = packed struct {
     magic: u32,
     inode_count: u32,
@@ -26,7 +28,7 @@ pub const Superblock = packed struct {
     id_count: u16,
     ver_maj: u16,
     ver_min: u16,
-    root_ref: u64,
+    root_ref: InodeRef,
     size: u64,
     id_start: u64,
     xattr_start: u64,
@@ -57,11 +59,13 @@ pub const Compression = enum(u16) {
             },
             .lzma => {
                 const decomp = try std.compress.lzma.decompress(alloc, source);
+                defer decomp.deinit();
                 return decomp.read(dest);
             },
             .lzo => return DecompressError.LzoUnavailable,
             .xz => {
                 const decomp = try std.compress.xz.decompress(alloc, source);
+                defer decomp.deinit();
                 return decomp.read(dest);
             },
             .lz4 => return DecompressError.Lz4Unavailable,

src/table.zig

@@ -0,0 +1,68 @@
+const std = @import("std");
+
+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: R,
+
+        offset: u64,
+        table_count: u32,
+        mut: std.Thread.RwLock = .{},
+
+        table: []T = &[0]T{},
+
+        pub fn init(alloc: std.mem.Allocator, rdr: R, offset: u64, table_count: u32) Self {
+            return .{
+                .alloc = alloc,
+                .rdr = rdr,
+                .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;
+                _ = try self.rdr.pread(std.mem.sliceAsBytes(self.table[self.table.len - to_read ..]), offset);
+            }
+            return self.table[idx];
+        }
+    };
+}