zig/lib/compiler/objcopy.zig
2024-10-12 10:44:17 -07:00

1672 lines
70 KiB
Zig

const builtin = @import("builtin");
const std = @import("std");
const mem = std.mem;
const fs = std.fs;
const elf = std.elf;
const Allocator = std.mem.Allocator;
const File = std.fs.File;
const assert = std.debug.assert;
const fatal = std.zig.fatal;
const Server = std.zig.Server;
pub fn main() !void {
var arena_instance = std.heap.ArenaAllocator.init(std.heap.page_allocator);
defer arena_instance.deinit();
const arena = arena_instance.allocator();
var general_purpose_allocator: std.heap.GeneralPurposeAllocator(.{}) = .init;
const gpa = general_purpose_allocator.allocator();
const args = try std.process.argsAlloc(arena);
return cmdObjCopy(gpa, arena, args[1..]);
}
fn cmdObjCopy(
gpa: Allocator,
arena: Allocator,
args: []const []const u8,
) !void {
var i: usize = 0;
var opt_out_fmt: ?std.Target.ObjectFormat = null;
var opt_input: ?[]const u8 = null;
var opt_output: ?[]const u8 = null;
var opt_extract: ?[]const u8 = null;
var opt_add_debuglink: ?[]const u8 = null;
var only_section: ?[]const u8 = null;
var pad_to: ?u64 = null;
var strip_all: bool = false;
var strip_debug: bool = false;
var only_keep_debug: bool = false;
var compress_debug_sections: bool = false;
var listen = false;
var add_section: ?AddSection = null;
var set_section_alignment: ?SetSectionAlignment = null;
var set_section_flags: ?SetSectionFlags = null;
while (i < args.len) : (i += 1) {
const arg = args[i];
if (!mem.startsWith(u8, arg, "-")) {
if (opt_input == null) {
opt_input = arg;
} else if (opt_output == null) {
opt_output = arg;
} else {
fatal("unexpected positional argument: '{s}'", .{arg});
}
} else if (mem.eql(u8, arg, "-h") or mem.eql(u8, arg, "--help")) {
return std.io.getStdOut().writeAll(usage);
} else if (mem.eql(u8, arg, "-O") or mem.eql(u8, arg, "--output-target")) {
i += 1;
if (i >= args.len) fatal("expected another argument after '{s}'", .{arg});
const next_arg = args[i];
if (mem.eql(u8, next_arg, "binary")) {
opt_out_fmt = .raw;
} else {
opt_out_fmt = std.meta.stringToEnum(std.Target.ObjectFormat, next_arg) orelse
fatal("invalid output format: '{s}'", .{next_arg});
}
} else if (mem.startsWith(u8, arg, "--output-target=")) {
const next_arg = arg["--output-target=".len..];
if (mem.eql(u8, next_arg, "binary")) {
opt_out_fmt = .raw;
} else {
opt_out_fmt = std.meta.stringToEnum(std.Target.ObjectFormat, next_arg) orelse
fatal("invalid output format: '{s}'", .{next_arg});
}
} else if (mem.eql(u8, arg, "-j") or mem.eql(u8, arg, "--only-section")) {
i += 1;
if (i >= args.len) fatal("expected another argument after '{s}'", .{arg});
only_section = args[i];
} else if (mem.eql(u8, arg, "--listen=-")) {
listen = true;
} else if (mem.startsWith(u8, arg, "--only-section=")) {
only_section = arg["--only-section=".len..];
} else if (mem.eql(u8, arg, "--pad-to")) {
i += 1;
if (i >= args.len) fatal("expected another argument after '{s}'", .{arg});
pad_to = std.fmt.parseInt(u64, args[i], 0) catch |err| {
fatal("unable to parse: '{s}': {s}", .{ args[i], @errorName(err) });
};
} else if (mem.eql(u8, arg, "-g") or mem.eql(u8, arg, "--strip-debug")) {
strip_debug = true;
} else if (mem.eql(u8, arg, "-S") or mem.eql(u8, arg, "--strip-all")) {
strip_all = true;
} else if (mem.eql(u8, arg, "--only-keep-debug")) {
only_keep_debug = true;
} else if (mem.eql(u8, arg, "--compress-debug-sections")) {
compress_debug_sections = true;
} else if (mem.startsWith(u8, arg, "--add-gnu-debuglink=")) {
opt_add_debuglink = arg["--add-gnu-debuglink=".len..];
} else if (mem.eql(u8, arg, "--add-gnu-debuglink")) {
i += 1;
if (i >= args.len) fatal("expected another argument after '{s}'", .{arg});
opt_add_debuglink = args[i];
} else if (mem.startsWith(u8, arg, "--extract-to=")) {
opt_extract = arg["--extract-to=".len..];
} else if (mem.eql(u8, arg, "--extract-to")) {
i += 1;
if (i >= args.len) fatal("expected another argument after '{s}'", .{arg});
opt_extract = args[i];
} else if (mem.eql(u8, arg, "--set-section-alignment")) {
i += 1;
if (i >= args.len) fatal("expected section name and alignment arguments after '{s}'", .{arg});
if (splitOption(args[i])) |split| {
const alignment = std.fmt.parseInt(u32, split.second, 10) catch |err| {
fatal("unable to parse alignment number: '{s}': {s}", .{ split.second, @errorName(err) });
};
if (!std.math.isPowerOfTwo(alignment)) fatal("alignment must be a power of two", .{});
set_section_alignment = .{ .section_name = split.first, .alignment = alignment };
} else {
fatal("unrecognized argument: '{s}', expecting <name>=<alignment>", .{args[i]});
}
} else if (mem.eql(u8, arg, "--set-section-flags")) {
i += 1;
if (i >= args.len) fatal("expected section name and filename arguments after '{s}'", .{arg});
if (splitOption(args[i])) |split| {
set_section_flags = .{ .section_name = split.first, .flags = parseSectionFlags(split.second) };
} else {
fatal("unrecognized argument: '{s}', expecting <name>=<flags>", .{args[i]});
}
} else if (mem.eql(u8, arg, "--add-section")) {
i += 1;
if (i >= args.len) fatal("expected section name and filename arguments after '{s}'", .{arg});
if (splitOption(args[i])) |split| {
add_section = .{ .section_name = split.first, .file_path = split.second };
} else {
fatal("unrecognized argument: '{s}', expecting <name>=<file>", .{args[i]});
}
} else {
fatal("unrecognized argument: '{s}'", .{arg});
}
}
const input = opt_input orelse fatal("expected input parameter", .{});
const output = opt_output orelse fatal("expected output parameter", .{});
var in_file = fs.cwd().openFile(input, .{}) catch |err|
fatal("unable to open '{s}': {s}", .{ input, @errorName(err) });
defer in_file.close();
const elf_hdr = std.elf.Header.read(in_file) catch |err| switch (err) {
error.InvalidElfMagic => fatal("not an ELF file: '{s}'", .{input}),
else => fatal("unable to read '{s}': {s}", .{ input, @errorName(err) }),
};
const in_ofmt = .elf;
const out_fmt: std.Target.ObjectFormat = opt_out_fmt orelse ofmt: {
if (mem.endsWith(u8, output, ".hex") or std.mem.endsWith(u8, output, ".ihex")) {
break :ofmt .hex;
} else if (mem.endsWith(u8, output, ".bin")) {
break :ofmt .raw;
} else if (mem.endsWith(u8, output, ".elf")) {
break :ofmt .elf;
} else {
break :ofmt in_ofmt;
}
};
const mode = mode: {
if (out_fmt != .elf or only_keep_debug)
break :mode fs.File.default_mode;
if (in_file.stat()) |stat|
break :mode stat.mode
else |_|
break :mode fs.File.default_mode;
};
var out_file = try fs.cwd().createFile(output, .{ .mode = mode });
defer out_file.close();
switch (out_fmt) {
.hex, .raw => {
if (strip_debug or strip_all or only_keep_debug)
fatal("zig objcopy: ELF to RAW or HEX copying does not support --strip", .{});
if (opt_extract != null)
fatal("zig objcopy: ELF to RAW or HEX copying does not support --extract-to", .{});
if (add_section != null)
fatal("zig objcopy: ELF to RAW or HEX copying does not support --add-section", .{});
if (set_section_alignment != null)
fatal("zig objcopy: ELF to RAW or HEX copying does not support --set_section_alignment", .{});
if (set_section_flags != null)
fatal("zig objcopy: ELF to RAW or HEX copying does not support --set_section_flags", .{});
try emitElf(arena, in_file, out_file, elf_hdr, .{
.ofmt = out_fmt,
.only_section = only_section,
.pad_to = pad_to,
});
},
.elf => {
if (elf_hdr.endian != builtin.target.cpu.arch.endian())
fatal("zig objcopy: ELF to ELF copying only supports native endian", .{});
if (elf_hdr.phoff == 0) // no program header
fatal("zig objcopy: ELF to ELF copying only supports programs", .{});
if (only_section) |_|
fatal("zig objcopy: ELF to ELF copying does not support --only-section", .{});
if (pad_to) |_|
fatal("zig objcopy: ELF to ELF copying does not support --pad-to", .{});
try stripElf(arena, in_file, out_file, elf_hdr, .{
.strip_debug = strip_debug,
.strip_all = strip_all,
.only_keep_debug = only_keep_debug,
.add_debuglink = opt_add_debuglink,
.extract_to = opt_extract,
.compress_debug = compress_debug_sections,
.add_section = add_section,
.set_section_alignment = set_section_alignment,
.set_section_flags = set_section_flags,
});
return std.process.cleanExit();
},
else => fatal("unsupported output object format: {s}", .{@tagName(out_fmt)}),
}
if (listen) {
var server = try Server.init(.{
.gpa = gpa,
.in = std.io.getStdIn(),
.out = std.io.getStdOut(),
.zig_version = builtin.zig_version_string,
});
defer server.deinit();
var seen_update = false;
while (true) {
const hdr = try server.receiveMessage();
switch (hdr.tag) {
.exit => {
return std.process.cleanExit();
},
.update => {
if (seen_update) fatal("zig objcopy only supports 1 update for now", .{});
seen_update = true;
// The build system already knows what the output is at this point, we
// only need to communicate that the process has finished.
// Use the empty error bundle to indicate that the update is done.
try server.serveErrorBundle(std.zig.ErrorBundle.empty);
},
else => fatal("unsupported message: {s}", .{@tagName(hdr.tag)}),
}
}
}
return std.process.cleanExit();
}
const usage =
\\Usage: zig objcopy [options] input output
\\
\\Options:
\\ -h, --help Print this help and exit
\\ --output-target=<value> Format of the output file
\\ -O <value> Alias for --output-target
\\ --only-section=<section> Remove all but <section>
\\ -j <value> Alias for --only-section
\\ --pad-to <addr> Pad the last section up to address <addr>
\\ --strip-debug, -g Remove all debug sections from the output.
\\ --strip-all, -S Remove all debug sections and symbol table from the output.
\\ --only-keep-debug Strip a file, removing contents of any sections that would not be stripped by --strip-debug and leaving the debugging sections intact.
\\ --add-gnu-debuglink=<file> Creates a .gnu_debuglink section which contains a reference to <file> and adds it to the output file.
\\ --extract-to <file> Extract the removed sections into <file>, and add a .gnu-debuglink section.
\\ --compress-debug-sections Compress DWARF debug sections with zlib
\\ --set-section-alignment <name>=<align> Set alignment of section <name> to <align> bytes. Must be a power of two.
\\ --set-section-flags <name>=<file> Set flags of section <name> to <flags> represented as a comma separated set of flags.
\\ --add-section <name>=<file> Add file content from <file> with the a new section named <name>.
\\
;
pub const EmitRawElfOptions = struct {
ofmt: std.Target.ObjectFormat,
only_section: ?[]const u8 = null,
pad_to: ?u64 = null,
add_section: ?AddSection = null,
set_section_alignment: ?SetSectionAlignment = null,
set_section_flags: ?SetSectionFlags = null,
};
const AddSection = struct {
section_name: []const u8,
file_path: []const u8,
};
const SetSectionAlignment = struct {
section_name: []const u8,
alignment: u32,
};
const SetSectionFlags = struct {
section_name: []const u8,
flags: SectionFlags,
};
fn emitElf(
arena: Allocator,
in_file: File,
out_file: File,
elf_hdr: elf.Header,
options: EmitRawElfOptions,
) !void {
var binary_elf_output = try BinaryElfOutput.parse(arena, in_file, elf_hdr);
defer binary_elf_output.deinit();
if (options.ofmt == .elf) {
fatal("zig objcopy: ELF to ELF copying is not implemented yet", .{});
}
if (options.only_section) |target_name| {
switch (options.ofmt) {
.hex => fatal("zig objcopy: hex format with sections is not implemented yet", .{}),
.raw => {
for (binary_elf_output.sections.items) |section| {
if (section.name) |curr_name| {
if (!std.mem.eql(u8, curr_name, target_name))
continue;
} else {
continue;
}
try writeBinaryElfSection(in_file, out_file, section);
try padFile(out_file, options.pad_to);
return;
}
},
else => unreachable,
}
return error.SectionNotFound;
}
switch (options.ofmt) {
.raw => {
for (binary_elf_output.sections.items) |section| {
try out_file.seekTo(section.binaryOffset);
try writeBinaryElfSection(in_file, out_file, section);
}
try padFile(out_file, options.pad_to);
},
.hex => {
if (binary_elf_output.segments.items.len == 0) return;
if (!containsValidAddressRange(binary_elf_output.segments.items)) {
return error.InvalidHexfileAddressRange;
}
var hex_writer = HexWriter{ .out_file = out_file };
for (binary_elf_output.segments.items) |segment| {
try hex_writer.writeSegment(segment, in_file);
}
if (options.pad_to) |_| {
// Padding to a size in hex files isn't applicable
return error.InvalidArgument;
}
try hex_writer.writeEOF();
},
else => unreachable,
}
}
const BinaryElfSection = struct {
elfOffset: u64,
binaryOffset: u64,
fileSize: usize,
name: ?[]const u8,
segment: ?*BinaryElfSegment,
};
const BinaryElfSegment = struct {
physicalAddress: u64,
virtualAddress: u64,
elfOffset: u64,
binaryOffset: u64,
fileSize: u64,
firstSection: ?*BinaryElfSection,
};
const BinaryElfOutput = struct {
segments: std.ArrayListUnmanaged(*BinaryElfSegment),
sections: std.ArrayListUnmanaged(*BinaryElfSection),
allocator: Allocator,
shstrtab: ?[]const u8,
const Self = @This();
pub fn deinit(self: *Self) void {
if (self.shstrtab) |shstrtab|
self.allocator.free(shstrtab);
self.sections.deinit(self.allocator);
self.segments.deinit(self.allocator);
}
pub fn parse(allocator: Allocator, elf_file: File, elf_hdr: elf.Header) !Self {
var self: Self = .{
.segments = .{},
.sections = .{},
.allocator = allocator,
.shstrtab = null,
};
errdefer self.sections.deinit(allocator);
errdefer self.segments.deinit(allocator);
self.shstrtab = blk: {
if (elf_hdr.shstrndx >= elf_hdr.shnum) break :blk null;
var section_headers = elf_hdr.section_header_iterator(&elf_file);
var section_counter: usize = 0;
while (section_counter < elf_hdr.shstrndx) : (section_counter += 1) {
_ = (try section_headers.next()).?;
}
const shstrtab_shdr = (try section_headers.next()).?;
const buffer = try allocator.alloc(u8, @intCast(shstrtab_shdr.sh_size));
errdefer allocator.free(buffer);
const num_read = try elf_file.preadAll(buffer, shstrtab_shdr.sh_offset);
if (num_read != buffer.len) return error.EndOfStream;
break :blk buffer;
};
errdefer if (self.shstrtab) |shstrtab| allocator.free(shstrtab);
var section_headers = elf_hdr.section_header_iterator(&elf_file);
while (try section_headers.next()) |section| {
if (sectionValidForOutput(section)) {
const newSection = try allocator.create(BinaryElfSection);
newSection.binaryOffset = 0;
newSection.elfOffset = section.sh_offset;
newSection.fileSize = @intCast(section.sh_size);
newSection.segment = null;
newSection.name = if (self.shstrtab) |shstrtab|
std.mem.span(@as([*:0]const u8, @ptrCast(&shstrtab[section.sh_name])))
else
null;
try self.sections.append(allocator, newSection);
}
}
var program_headers = elf_hdr.program_header_iterator(&elf_file);
while (try program_headers.next()) |phdr| {
if (phdr.p_type == elf.PT_LOAD) {
const newSegment = try allocator.create(BinaryElfSegment);
newSegment.physicalAddress = phdr.p_paddr;
newSegment.virtualAddress = phdr.p_vaddr;
newSegment.fileSize = @intCast(phdr.p_filesz);
newSegment.elfOffset = phdr.p_offset;
newSegment.binaryOffset = 0;
newSegment.firstSection = null;
for (self.sections.items) |section| {
if (sectionWithinSegment(section, phdr)) {
if (section.segment) |sectionSegment| {
if (sectionSegment.elfOffset > newSegment.elfOffset) {
section.segment = newSegment;
}
} else {
section.segment = newSegment;
}
if (newSegment.firstSection == null) {
newSegment.firstSection = section;
}
}
}
try self.segments.append(allocator, newSegment);
}
}
mem.sort(*BinaryElfSegment, self.segments.items, {}, segmentSortCompare);
for (self.segments.items, 0..) |firstSegment, i| {
if (firstSegment.firstSection) |firstSection| {
const diff = firstSection.elfOffset - firstSegment.elfOffset;
firstSegment.elfOffset += diff;
firstSegment.fileSize += diff;
firstSegment.physicalAddress += diff;
const basePhysicalAddress = firstSegment.physicalAddress;
for (self.segments.items[i + 1 ..]) |segment| {
segment.binaryOffset = segment.physicalAddress - basePhysicalAddress;
}
break;
}
}
for (self.sections.items) |section| {
if (section.segment) |segment| {
section.binaryOffset = segment.binaryOffset + (section.elfOffset - segment.elfOffset);
}
}
mem.sort(*BinaryElfSection, self.sections.items, {}, sectionSortCompare);
return self;
}
fn sectionWithinSegment(section: *BinaryElfSection, segment: elf.Elf64_Phdr) bool {
return segment.p_offset <= section.elfOffset and (segment.p_offset + segment.p_filesz) >= (section.elfOffset + section.fileSize);
}
fn sectionValidForOutput(shdr: anytype) bool {
return shdr.sh_type != elf.SHT_NOBITS and
((shdr.sh_flags & elf.SHF_ALLOC) == elf.SHF_ALLOC);
}
fn segmentSortCompare(context: void, left: *BinaryElfSegment, right: *BinaryElfSegment) bool {
_ = context;
if (left.physicalAddress < right.physicalAddress) {
return true;
}
if (left.physicalAddress > right.physicalAddress) {
return false;
}
return false;
}
fn sectionSortCompare(context: void, left: *BinaryElfSection, right: *BinaryElfSection) bool {
_ = context;
return left.binaryOffset < right.binaryOffset;
}
};
fn writeBinaryElfSection(elf_file: File, out_file: File, section: *BinaryElfSection) !void {
try out_file.writeFileAll(elf_file, .{
.in_offset = section.elfOffset,
.in_len = section.fileSize,
});
}
const HexWriter = struct {
prev_addr: ?u32 = null,
out_file: File,
/// Max data bytes per line of output
const MAX_PAYLOAD_LEN: u8 = 16;
fn addressParts(address: u16) [2]u8 {
const msb: u8 = @truncate(address >> 8);
const lsb: u8 = @truncate(address);
return [2]u8{ msb, lsb };
}
const Record = struct {
const Type = enum(u8) {
Data = 0,
EOF = 1,
ExtendedSegmentAddress = 2,
ExtendedLinearAddress = 4,
};
address: u16,
payload: union(Type) {
Data: []const u8,
EOF: void,
ExtendedSegmentAddress: [2]u8,
ExtendedLinearAddress: [2]u8,
},
fn EOF() Record {
return Record{
.address = 0,
.payload = .EOF,
};
}
fn Data(address: u32, data: []const u8) Record {
return Record{
.address = @intCast(address % 0x10000),
.payload = .{ .Data = data },
};
}
fn Address(address: u32) Record {
assert(address > 0xFFFF);
const segment: u16 = @intCast(address / 0x10000);
if (address > 0xFFFFF) {
return Record{
.address = 0,
.payload = .{ .ExtendedLinearAddress = addressParts(segment) },
};
} else {
return Record{
.address = 0,
.payload = .{ .ExtendedSegmentAddress = addressParts(segment << 12) },
};
}
}
fn getPayloadBytes(self: *const Record) []const u8 {
return switch (self.payload) {
.Data => |d| d,
.EOF => @as([]const u8, &.{}),
.ExtendedSegmentAddress, .ExtendedLinearAddress => |*seg| seg,
};
}
fn checksum(self: Record) u8 {
const payload_bytes = self.getPayloadBytes();
var sum: u8 = @intCast(payload_bytes.len);
const parts = addressParts(self.address);
sum +%= parts[0];
sum +%= parts[1];
sum +%= @intFromEnum(self.payload);
for (payload_bytes) |byte| {
sum +%= byte;
}
return (sum ^ 0xFF) +% 1;
}
fn write(self: Record, file: File) File.WriteError!void {
const linesep = "\r\n";
// colon, (length, address, type, payload, checksum) as hex, CRLF
const BUFSIZE = 1 + (1 + 2 + 1 + MAX_PAYLOAD_LEN + 1) * 2 + linesep.len;
var outbuf: [BUFSIZE]u8 = undefined;
const payload_bytes = self.getPayloadBytes();
assert(payload_bytes.len <= MAX_PAYLOAD_LEN);
const line = try std.fmt.bufPrint(&outbuf, ":{0X:0>2}{1X:0>4}{2X:0>2}{3s}{4X:0>2}" ++ linesep, .{
@as(u8, @intCast(payload_bytes.len)),
self.address,
@intFromEnum(self.payload),
std.fmt.fmtSliceHexUpper(payload_bytes),
self.checksum(),
});
try file.writeAll(line);
}
};
pub fn writeSegment(self: *HexWriter, segment: *const BinaryElfSegment, elf_file: File) !void {
var buf: [MAX_PAYLOAD_LEN]u8 = undefined;
var bytes_read: usize = 0;
while (bytes_read < segment.fileSize) {
const row_address: u32 = @intCast(segment.physicalAddress + bytes_read);
const remaining = segment.fileSize - bytes_read;
const to_read: usize = @intCast(@min(remaining, MAX_PAYLOAD_LEN));
const did_read = try elf_file.preadAll(buf[0..to_read], segment.elfOffset + bytes_read);
if (did_read < to_read) return error.UnexpectedEOF;
try self.writeDataRow(row_address, buf[0..did_read]);
bytes_read += did_read;
}
}
fn writeDataRow(self: *HexWriter, address: u32, data: []const u8) File.WriteError!void {
const record = Record.Data(address, data);
if (address > 0xFFFF and (self.prev_addr == null or record.address != self.prev_addr.?)) {
try Record.Address(address).write(self.out_file);
}
try record.write(self.out_file);
self.prev_addr = @intCast(record.address + data.len);
}
fn writeEOF(self: HexWriter) File.WriteError!void {
try Record.EOF().write(self.out_file);
}
};
fn containsValidAddressRange(segments: []*BinaryElfSegment) bool {
const max_address = std.math.maxInt(u32);
for (segments) |segment| {
if (segment.fileSize > max_address or
segment.physicalAddress > max_address - segment.fileSize) return false;
}
return true;
}
fn padFile(f: File, opt_size: ?u64) !void {
const size = opt_size orelse return;
try f.setEndPos(size);
}
test "HexWriter.Record.Address has correct payload and checksum" {
const record = HexWriter.Record.Address(0x0800_0000);
const payload = record.getPayloadBytes();
const sum = record.checksum();
try std.testing.expect(sum == 0xF2);
try std.testing.expect(payload.len == 2);
try std.testing.expect(payload[0] == 8);
try std.testing.expect(payload[1] == 0);
}
test "containsValidAddressRange" {
var segment = BinaryElfSegment{
.physicalAddress = 0,
.virtualAddress = 0,
.elfOffset = 0,
.binaryOffset = 0,
.fileSize = 0,
.firstSection = null,
};
var buf: [1]*BinaryElfSegment = .{&segment};
// segment too big
segment.fileSize = std.math.maxInt(u32) + 1;
try std.testing.expect(!containsValidAddressRange(&buf));
// start address too big
segment.physicalAddress = std.math.maxInt(u32) + 1;
segment.fileSize = 2;
try std.testing.expect(!containsValidAddressRange(&buf));
// max address too big
segment.physicalAddress = std.math.maxInt(u32) - 1;
segment.fileSize = 2;
try std.testing.expect(!containsValidAddressRange(&buf));
// is ok
segment.physicalAddress = std.math.maxInt(u32) - 1;
segment.fileSize = 1;
try std.testing.expect(containsValidAddressRange(&buf));
}
// -------------
// ELF to ELF stripping
const StripElfOptions = struct {
extract_to: ?[]const u8 = null,
add_debuglink: ?[]const u8 = null,
strip_all: bool = false,
strip_debug: bool = false,
only_keep_debug: bool = false,
compress_debug: bool = false,
add_section: ?AddSection,
set_section_alignment: ?SetSectionAlignment,
set_section_flags: ?SetSectionFlags,
};
fn stripElf(
allocator: Allocator,
in_file: File,
out_file: File,
elf_hdr: elf.Header,
options: StripElfOptions,
) !void {
const Filter = ElfFileHelper.Filter;
const DebugLink = ElfFileHelper.DebugLink;
const filter: Filter = filter: {
if (options.only_keep_debug) break :filter .debug;
if (options.strip_all) break :filter .program;
if (options.strip_debug) break :filter .program_and_symbols;
break :filter .all;
};
const filter_complement: ?Filter = blk: {
if (options.extract_to) |_| {
break :blk switch (filter) {
.program => .debug_and_symbols,
.debug => .program_and_symbols,
.program_and_symbols => .debug,
.debug_and_symbols => .program,
.all => fatal("zig objcopy: nothing to extract", .{}),
};
} else {
break :blk null;
}
};
const debuglink_path = path: {
if (options.add_debuglink) |path| break :path path;
if (options.extract_to) |path| break :path path;
break :path null;
};
switch (elf_hdr.is_64) {
inline else => |is_64| {
var elf_file = try ElfFile(is_64).parse(allocator, in_file, elf_hdr);
defer elf_file.deinit();
if (options.add_section) |user_section| {
for (elf_file.sections) |section| {
if (std.mem.eql(u8, section.name, user_section.section_name)) {
fatal("zig objcopy: unable to add section '{s}'. Section already exists in input", .{user_section.section_name});
}
}
}
if (filter_complement) |flt| {
// write the .dbg file and close it, so it can be read back to compute the debuglink checksum.
const path = options.extract_to.?;
const dbg_file = std.fs.cwd().createFile(path, .{}) catch |err| {
fatal("zig objcopy: unable to create '{s}': {s}", .{ path, @errorName(err) });
};
defer dbg_file.close();
try elf_file.emit(allocator, dbg_file, in_file, .{ .section_filter = flt, .compress_debug = options.compress_debug });
}
const debuglink: ?DebugLink = if (debuglink_path) |path| ElfFileHelper.createDebugLink(path) else null;
try elf_file.emit(allocator, out_file, in_file, .{
.section_filter = filter,
.debuglink = debuglink,
.compress_debug = options.compress_debug,
.add_section = options.add_section,
.set_section_alignment = options.set_section_alignment,
.set_section_flags = options.set_section_flags,
});
},
}
}
// note: this is "a minimal effort implementation"
// It doesn't support all possibile elf files: some sections type may need fixups, the program header may need fix up, ...
// It was written for a specific use case (strip debug info to a sperate file, for linux 64-bits executables built with `zig` or `zig c++` )
// It moves and reoders the sections as little as possible to avoid having to do fixups.
// TODO: support non-native endianess
fn ElfFile(comptime is_64: bool) type {
const Elf_Ehdr = if (is_64) elf.Elf64_Ehdr else elf.Elf32_Ehdr;
const Elf_Phdr = if (is_64) elf.Elf64_Phdr else elf.Elf32_Phdr;
const Elf_Shdr = if (is_64) elf.Elf64_Shdr else elf.Elf32_Shdr;
const Elf_Chdr = if (is_64) elf.Elf64_Chdr else elf.Elf32_Chdr;
const Elf_Sym = if (is_64) elf.Elf64_Sym else elf.Elf32_Sym;
const Elf_OffSize = if (is_64) elf.Elf64_Off else elf.Elf32_Off;
return struct {
raw_elf_header: Elf_Ehdr,
program_segments: []const Elf_Phdr,
sections: []const Section,
arena: std.heap.ArenaAllocator,
const SectionCategory = ElfFileHelper.SectionCategory;
const section_memory_align = @alignOf(Elf_Sym); // most restrictive of what we may load in memory
const Section = struct {
section: Elf_Shdr,
name: []const u8 = "",
segment: ?*const Elf_Phdr = null, // if the section is used by a program segment (there can be more than one)
payload: ?[]align(section_memory_align) const u8 = null, // if we need the data in memory
category: SectionCategory = .none, // should the section be kept in the exe or stripped to the debug database, or both.
};
const Self = @This();
pub fn parse(gpa: Allocator, in_file: File, header: elf.Header) !Self {
var arena = std.heap.ArenaAllocator.init(gpa);
errdefer arena.deinit();
const allocator = arena.allocator();
var raw_header: Elf_Ehdr = undefined;
{
const bytes_read = try in_file.preadAll(std.mem.asBytes(&raw_header), 0);
if (bytes_read < @sizeOf(Elf_Ehdr))
return error.TRUNCATED_ELF;
}
// program header: list of segments
const program_segments = blk: {
if (@sizeOf(Elf_Phdr) != header.phentsize)
fatal("zig objcopy: unsupported ELF file, unexpected phentsize ({d})", .{header.phentsize});
const program_header = try allocator.alloc(Elf_Phdr, header.phnum);
const bytes_read = try in_file.preadAll(std.mem.sliceAsBytes(program_header), header.phoff);
if (bytes_read < @sizeOf(Elf_Phdr) * header.phnum)
return error.TRUNCATED_ELF;
break :blk program_header;
};
// section header
const sections = blk: {
if (@sizeOf(Elf_Shdr) != header.shentsize)
fatal("zig objcopy: unsupported ELF file, unexpected shentsize ({d})", .{header.shentsize});
const section_header = try allocator.alloc(Section, header.shnum);
const raw_section_header = try allocator.alloc(Elf_Shdr, header.shnum);
defer allocator.free(raw_section_header);
const bytes_read = try in_file.preadAll(std.mem.sliceAsBytes(raw_section_header), header.shoff);
if (bytes_read < @sizeOf(Elf_Phdr) * header.shnum)
return error.TRUNCATED_ELF;
for (section_header, raw_section_header) |*section, hdr| {
section.* = .{ .section = hdr };
}
break :blk section_header;
};
// load data to memory for some sections:
// string tables for access
// sections than need modifications when other sections move.
for (sections, 0..) |*section, idx| {
const need_data = switch (section.section.sh_type) {
elf.DT_VERSYM => true,
elf.SHT_SYMTAB, elf.SHT_DYNSYM => true,
else => false,
};
const need_strings = (idx == header.shstrndx);
if (need_data or need_strings) {
const buffer = try allocator.alignedAlloc(u8, section_memory_align, @intCast(section.section.sh_size));
const bytes_read = try in_file.preadAll(buffer, section.section.sh_offset);
if (bytes_read != section.section.sh_size) return error.TRUNCATED_ELF;
section.payload = buffer;
}
}
// fill-in sections info:
// resolve the name
// find if a program segment uses the section
// categorize sections usage (used by program segments, debug datadase, common metadata, symbol table)
for (sections) |*section| {
section.segment = for (program_segments) |*seg| {
if (sectionWithinSegment(section.section, seg.*)) break seg;
} else null;
if (section.section.sh_name != 0 and header.shstrndx != elf.SHN_UNDEF)
section.name = std.mem.span(@as([*:0]const u8, @ptrCast(&sections[header.shstrndx].payload.?[section.section.sh_name])));
const category_from_program: SectionCategory = if (section.segment != null) .exe else .debug;
section.category = switch (section.section.sh_type) {
elf.SHT_NOTE => .common,
elf.SHT_SYMTAB => .symbols, // "strip all" vs "strip only debug"
elf.SHT_DYNSYM => .exe,
elf.SHT_PROGBITS => cat: {
if (std.mem.eql(u8, section.name, ".comment")) break :cat .exe;
if (std.mem.eql(u8, section.name, ".gnu_debuglink")) break :cat .none;
break :cat category_from_program;
},
elf.SHT_LOPROC...elf.SHT_HIPROC => .common, // don't strip unknown sections
elf.SHT_LOUSER...elf.SHT_HIUSER => .common, // don't strip unknown sections
else => category_from_program,
};
}
sections[0].category = .common; // mandatory null section
if (header.shstrndx != elf.SHN_UNDEF)
sections[header.shstrndx].category = .common; // string table for the headers
// recursively propagate section categories to their linked sections, so that they are kept together
var dirty: u1 = 1;
while (dirty != 0) {
dirty = 0;
for (sections) |*section| {
if (section.section.sh_link != elf.SHN_UNDEF)
dirty |= ElfFileHelper.propagateCategory(&sections[section.section.sh_link].category, section.category);
if ((section.section.sh_flags & elf.SHF_INFO_LINK) != 0 and section.section.sh_info != elf.SHN_UNDEF)
dirty |= ElfFileHelper.propagateCategory(&sections[section.section.sh_info].category, section.category);
}
}
return Self{
.arena = arena,
.raw_elf_header = raw_header,
.program_segments = program_segments,
.sections = sections,
};
}
pub fn deinit(self: *Self) void {
self.arena.deinit();
}
const Filter = ElfFileHelper.Filter;
const DebugLink = ElfFileHelper.DebugLink;
const EmitElfOptions = struct {
section_filter: Filter = .all,
debuglink: ?DebugLink = null,
compress_debug: bool = false,
add_section: ?AddSection = null,
set_section_alignment: ?SetSectionAlignment = null,
set_section_flags: ?SetSectionFlags = null,
};
fn emit(self: *const Self, gpa: Allocator, out_file: File, in_file: File, options: EmitElfOptions) !void {
var arena = std.heap.ArenaAllocator.init(gpa);
defer arena.deinit();
const allocator = arena.allocator();
// when emitting the stripped exe:
// - unused sections are removed
// when emitting the debug file:
// - all sections are kept, but some are emptied and their types is changed to SHT_NOBITS
// the program header is kept unchanged. (`strip` does update it, but `eu-strip` does not, and it still works)
const Update = struct {
action: ElfFileHelper.Action,
// remap the indexs after omitting the filtered sections
remap_idx: u16,
// optionally overrides the payload from the source file
payload: ?[]align(section_memory_align) const u8 = null,
section: ?Elf_Shdr = null,
};
const sections_update = try allocator.alloc(Update, self.sections.len);
const new_shnum = blk: {
var next_idx: u16 = 0;
for (self.sections, sections_update) |section, *update| {
const action = ElfFileHelper.selectAction(section.category, options.section_filter);
const remap_idx = idx: {
if (action == .strip) break :idx elf.SHN_UNDEF;
next_idx += 1;
break :idx next_idx - 1;
};
update.* = Update{ .action = action, .remap_idx = remap_idx };
}
if (options.debuglink != null)
next_idx += 1;
if (options.add_section != null) {
next_idx += 1;
}
break :blk next_idx;
};
// add a ".gnu_debuglink" to the string table if needed
const debuglink_name: u32 = blk: {
if (options.debuglink == null) break :blk elf.SHN_UNDEF;
if (self.raw_elf_header.e_shstrndx == elf.SHN_UNDEF)
fatal("zig objcopy: no strtab, cannot add the debuglink section", .{}); // TODO add the section if needed?
const strtab = &self.sections[self.raw_elf_header.e_shstrndx];
const update = &sections_update[self.raw_elf_header.e_shstrndx];
const name: []const u8 = ".gnu_debuglink";
const new_offset: u32 = @intCast(strtab.payload.?.len);
const buf = try allocator.alignedAlloc(u8, section_memory_align, new_offset + name.len + 1);
@memcpy(buf[0..new_offset], strtab.payload.?);
@memcpy(buf[new_offset..][0..name.len], name);
buf[new_offset + name.len] = 0;
assert(update.action == .keep);
update.payload = buf;
break :blk new_offset;
};
// add user section to the string table if needed
const user_section_name: u32 = blk: {
if (options.add_section == null) break :blk elf.SHN_UNDEF;
if (self.raw_elf_header.e_shstrndx == elf.SHN_UNDEF)
fatal("zig objcopy: no strtab, cannot add the user section", .{}); // TODO add the section if needed?
const strtab = &self.sections[self.raw_elf_header.e_shstrndx];
const update = &sections_update[self.raw_elf_header.e_shstrndx];
const name = options.add_section.?.section_name;
const new_offset: u32 = @intCast(strtab.payload.?.len);
const buf = try allocator.alignedAlloc(u8, section_memory_align, new_offset + name.len + 1);
@memcpy(buf[0..new_offset], strtab.payload.?);
@memcpy(buf[new_offset..][0..name.len], name);
buf[new_offset + name.len] = 0;
assert(update.action == .keep);
update.payload = buf;
break :blk new_offset;
};
// maybe compress .debug sections
if (options.compress_debug) {
for (self.sections[1..], sections_update[1..]) |section, *update| {
if (update.action != .keep) continue;
if (!std.mem.startsWith(u8, section.name, ".debug_")) continue;
if ((section.section.sh_flags & elf.SHF_COMPRESSED) != 0) continue; // already compressed
const chdr = Elf_Chdr{
.ch_type = elf.COMPRESS.ZLIB,
.ch_size = section.section.sh_size,
.ch_addralign = section.section.sh_addralign,
};
const compressed_payload = try ElfFileHelper.tryCompressSection(allocator, in_file, section.section.sh_offset, section.section.sh_size, std.mem.asBytes(&chdr));
if (compressed_payload) |payload| {
update.payload = payload;
update.section = section.section;
update.section.?.sh_addralign = @alignOf(Elf_Chdr);
update.section.?.sh_size = @intCast(payload.len);
update.section.?.sh_flags |= elf.SHF_COMPRESSED;
}
}
}
var cmdbuf = std.ArrayList(ElfFileHelper.WriteCmd).init(allocator);
defer cmdbuf.deinit();
try cmdbuf.ensureUnusedCapacity(3 + new_shnum);
var eof_offset: Elf_OffSize = 0; // track the end of the data written so far.
// build the updated headers
// nb: updated_elf_header will be updated before the actual write
var updated_elf_header = self.raw_elf_header;
if (updated_elf_header.e_shstrndx != elf.SHN_UNDEF)
updated_elf_header.e_shstrndx = sections_update[updated_elf_header.e_shstrndx].remap_idx;
cmdbuf.appendAssumeCapacity(.{ .write_data = .{ .data = std.mem.asBytes(&updated_elf_header), .out_offset = 0 } });
eof_offset = @sizeOf(Elf_Ehdr);
// program header as-is.
// nb: for only-debug files, removing it appears to work, but is invalid by ELF specifcation.
{
assert(updated_elf_header.e_phoff == @sizeOf(Elf_Ehdr));
const data = std.mem.sliceAsBytes(self.program_segments);
assert(data.len == @as(usize, updated_elf_header.e_phentsize) * updated_elf_header.e_phnum);
cmdbuf.appendAssumeCapacity(.{ .write_data = .{ .data = data, .out_offset = updated_elf_header.e_phoff } });
eof_offset = updated_elf_header.e_phoff + @as(Elf_OffSize, @intCast(data.len));
}
// update sections and queue payload writes
const updated_section_header = blk: {
const dest_sections = try allocator.alloc(Elf_Shdr, new_shnum);
{
// the ELF format doesn't specify the order for all sections.
// this code only supports when they are in increasing file order.
var offset: u64 = eof_offset;
for (self.sections[1..]) |section| {
if (section.section.sh_type == elf.SHT_NOBITS)
continue;
if (section.section.sh_offset < offset) {
fatal("zig objcopy: unsupported ELF file", .{});
}
offset = section.section.sh_offset;
}
}
dest_sections[0] = self.sections[0].section;
var dest_section_idx: u32 = 1;
for (self.sections[1..], sections_update[1..]) |section, update| {
if (update.action == .strip) continue;
assert(update.remap_idx == dest_section_idx);
const src = if (update.section) |*s| s else &section.section;
const dest = &dest_sections[dest_section_idx];
const payload = if (update.payload) |data| data else section.payload;
dest_section_idx += 1;
dest.* = src.*;
if (src.sh_link != elf.SHN_UNDEF)
dest.sh_link = sections_update[src.sh_link].remap_idx;
if ((src.sh_flags & elf.SHF_INFO_LINK) != 0 and src.sh_info != elf.SHN_UNDEF)
dest.sh_info = sections_update[src.sh_info].remap_idx;
if (payload) |data|
dest.sh_size = @intCast(data.len);
const addralign = if (src.sh_addralign == 0 or dest.sh_type == elf.SHT_NOBITS) 1 else src.sh_addralign;
dest.sh_offset = std.mem.alignForward(Elf_OffSize, eof_offset, addralign);
if (src.sh_offset != dest.sh_offset and section.segment != null and update.action != .empty and dest.sh_type != elf.SHT_NOTE and dest.sh_type != elf.SHT_NOBITS) {
if (src.sh_offset > dest.sh_offset) {
dest.sh_offset = src.sh_offset; // add padding to avoid modifing the program segments
} else {
fatal("zig objcopy: cannot adjust program segments", .{});
}
}
assert(dest.sh_addr % addralign == dest.sh_offset % addralign);
if (update.action == .empty)
dest.sh_type = elf.SHT_NOBITS;
if (dest.sh_type != elf.SHT_NOBITS) {
if (payload) |src_data| {
// update sections payload and write
const dest_data = switch (src.sh_type) {
elf.DT_VERSYM => dst_data: {
const data = try allocator.alignedAlloc(u8, section_memory_align, src_data.len);
@memcpy(data, src_data);
const defs = @as([*]elf.Verdef, @ptrCast(data))[0 .. @as(usize, @intCast(src.sh_size)) / @sizeOf(elf.Verdef)];
for (defs) |*def| switch (def.ndx) {
.LOCAL, .GLOBAL => {},
else => def.ndx = @enumFromInt(sections_update[src.sh_info].remap_idx),
};
break :dst_data data;
},
elf.SHT_SYMTAB, elf.SHT_DYNSYM => dst_data: {
const data = try allocator.alignedAlloc(u8, section_memory_align, src_data.len);
@memcpy(data, src_data);
const syms = @as([*]Elf_Sym, @ptrCast(data))[0 .. @as(usize, @intCast(src.sh_size)) / @sizeOf(Elf_Sym)];
for (syms) |*sym| {
if (sym.st_shndx != elf.SHN_UNDEF and sym.st_shndx < elf.SHN_LORESERVE)
sym.st_shndx = sections_update[sym.st_shndx].remap_idx;
}
break :dst_data data;
},
else => src_data,
};
assert(dest_data.len == dest.sh_size);
cmdbuf.appendAssumeCapacity(.{ .write_data = .{ .data = dest_data, .out_offset = dest.sh_offset } });
eof_offset = dest.sh_offset + dest.sh_size;
} else {
// direct contents copy
cmdbuf.appendAssumeCapacity(.{ .copy_range = .{ .in_offset = src.sh_offset, .len = dest.sh_size, .out_offset = dest.sh_offset } });
eof_offset = dest.sh_offset + dest.sh_size;
}
} else {
// account for alignment padding even in empty sections to keep logical section order
eof_offset = dest.sh_offset;
}
}
// add a ".gnu_debuglink" section
if (options.debuglink) |link| {
const payload = payload: {
const crc_offset = std.mem.alignForward(usize, link.name.len + 1, 4);
const buf = try allocator.alignedAlloc(u8, 4, crc_offset + 4);
@memcpy(buf[0..link.name.len], link.name);
@memset(buf[link.name.len..crc_offset], 0);
@memcpy(buf[crc_offset..], std.mem.asBytes(&link.crc32));
break :payload buf;
};
dest_sections[dest_section_idx] = Elf_Shdr{
.sh_name = debuglink_name,
.sh_type = elf.SHT_PROGBITS,
.sh_flags = 0,
.sh_addr = 0,
.sh_offset = eof_offset,
.sh_size = @intCast(payload.len),
.sh_link = elf.SHN_UNDEF,
.sh_info = elf.SHN_UNDEF,
.sh_addralign = 4,
.sh_entsize = 0,
};
dest_section_idx += 1;
cmdbuf.appendAssumeCapacity(.{ .write_data = .{ .data = payload, .out_offset = eof_offset } });
eof_offset += @as(Elf_OffSize, @intCast(payload.len));
}
// --add-section
if (options.add_section) |add_section| {
var section_file = fs.cwd().openFile(add_section.file_path, .{}) catch |err|
fatal("unable to open '{s}': {s}", .{ add_section.file_path, @errorName(err) });
defer section_file.close();
const payload = try section_file.readToEndAlloc(arena.allocator(), std.math.maxInt(usize));
dest_sections[dest_section_idx] = Elf_Shdr{
.sh_name = user_section_name,
.sh_type = elf.SHT_PROGBITS,
.sh_flags = 0,
.sh_addr = 0,
.sh_offset = eof_offset,
.sh_size = @intCast(payload.len),
.sh_link = elf.SHN_UNDEF,
.sh_info = elf.SHN_UNDEF,
.sh_addralign = 4,
.sh_entsize = 0,
};
dest_section_idx += 1;
cmdbuf.appendAssumeCapacity(.{ .write_data = .{ .data = payload, .out_offset = eof_offset } });
eof_offset += @as(Elf_OffSize, @intCast(payload.len));
}
assert(dest_section_idx == new_shnum);
break :blk dest_sections;
};
// --set-section-alignment: overwrite alignment
if (options.set_section_alignment) |set_align| {
if (self.raw_elf_header.e_shstrndx == elf.SHN_UNDEF)
fatal("zig objcopy: no strtab, cannot add the user section", .{}); // TODO add the section if needed?
const strtab = &sections_update[self.raw_elf_header.e_shstrndx];
for (updated_section_header) |*section| {
const section_name = std.mem.span(@as([*:0]const u8, @ptrCast(&strtab.payload.?[section.sh_name])));
if (std.mem.eql(u8, section_name, set_align.section_name)) {
section.sh_addralign = set_align.alignment;
break;
}
} else std.log.warn("Skipping --set-section-alignment. Section '{s}' not found", .{set_align.section_name});
}
// --set-section-flags: overwrite flags
if (options.set_section_flags) |set_flags| {
if (self.raw_elf_header.e_shstrndx == elf.SHN_UNDEF)
fatal("zig objcopy: no strtab, cannot add the user section", .{}); // TODO add the section if needed?
const strtab = &sections_update[self.raw_elf_header.e_shstrndx];
for (updated_section_header) |*section| {
const section_name = std.mem.span(@as([*:0]const u8, @ptrCast(&strtab.payload.?[section.sh_name])));
if (std.mem.eql(u8, section_name, set_flags.section_name)) {
section.sh_flags = std.elf.SHF_WRITE; // default is writable cleared by "readonly"
const f = set_flags.flags;
// Supporting a subset of GNU and LLVM objcopy for ELF only
// GNU:
// alloc: add SHF_ALLOC
// contents: if section is SHT_NOBITS, set SHT_PROGBITS, otherwise do nothing
// load: if section is SHT_NOBITS, set SHT_PROGBITS, otherwise do nothing (same as contents)
// noload: not ELF relevant
// readonly: clear default SHF_WRITE flag
// code: add SHF_EXECINSTR
// data: not ELF relevant
// rom: ignored
// exclude: add SHF_EXCLUDE
// share: not ELF relevant
// debug: not ELF relevant
// large: add SHF_X86_64_LARGE. Fatal error if target is not x86_64
if (f.alloc) section.sh_flags |= std.elf.SHF_ALLOC;
if (f.contents or f.load) {
if (section.sh_type == std.elf.SHT_NOBITS) section.sh_type = std.elf.SHT_PROGBITS;
}
if (f.readonly) section.sh_flags &= ~@as(@TypeOf(section.sh_type), std.elf.SHF_WRITE);
if (f.code) section.sh_flags |= std.elf.SHF_EXECINSTR;
if (f.exclude) section.sh_flags |= std.elf.SHF_EXCLUDE;
if (f.large) {
if (updated_elf_header.e_machine != std.elf.EM.X86_64)
fatal("zig objcopy: 'large' section flag is only supported on x86_64 targets", .{});
section.sh_flags |= std.elf.SHF_X86_64_LARGE;
}
// LLVM:
// merge: add SHF_MERGE
// strings: add SHF_STRINGS
if (f.merge) section.sh_flags |= std.elf.SHF_MERGE;
if (f.strings) section.sh_flags |= std.elf.SHF_STRINGS;
break;
}
} else std.log.warn("Skipping --set-section-flags. Section '{s}' not found", .{set_flags.section_name});
}
// write the section header at the tail
{
const offset = std.mem.alignForward(Elf_OffSize, eof_offset, @alignOf(Elf_Shdr));
const data = std.mem.sliceAsBytes(updated_section_header);
assert(data.len == @as(usize, updated_elf_header.e_shentsize) * new_shnum);
updated_elf_header.e_shoff = offset;
updated_elf_header.e_shnum = new_shnum;
cmdbuf.appendAssumeCapacity(.{ .write_data = .{ .data = data, .out_offset = updated_elf_header.e_shoff } });
}
try ElfFileHelper.write(allocator, out_file, in_file, cmdbuf.items);
}
fn sectionWithinSegment(section: Elf_Shdr, segment: Elf_Phdr) bool {
const file_size = if (section.sh_type == elf.SHT_NOBITS) 0 else section.sh_size;
return segment.p_offset <= section.sh_offset and (segment.p_offset + segment.p_filesz) >= (section.sh_offset + file_size);
}
};
}
const ElfFileHelper = struct {
const DebugLink = struct { name: []const u8, crc32: u32 };
const Filter = enum { all, program, debug, program_and_symbols, debug_and_symbols };
const SectionCategory = enum { common, exe, debug, symbols, none };
fn propagateCategory(cur: *SectionCategory, new: SectionCategory) u1 {
const cat: SectionCategory = switch (cur.*) {
.none => new,
.common => .common,
.debug => switch (new) {
.none, .debug => .debug,
else => new,
},
.exe => switch (new) {
.common => .common,
.none, .debug, .exe => .exe,
.symbols => .exe,
},
.symbols => switch (new) {
.none, .common, .debug, .exe => unreachable,
.symbols => .symbols,
},
};
if (cur.* != cat) {
cur.* = cat;
return 1;
} else {
return 0;
}
}
const Action = enum { keep, strip, empty };
fn selectAction(category: SectionCategory, filter: Filter) Action {
if (category == .none) return .strip;
return switch (filter) {
.all => switch (category) {
.none => .strip,
else => .keep,
},
.program => switch (category) {
.common, .exe => .keep,
else => .strip,
},
.program_and_symbols => switch (category) {
.common, .exe, .symbols => .keep,
else => .strip,
},
.debug => switch (category) {
.exe, .symbols => .empty,
.none => .strip,
else => .keep,
},
.debug_and_symbols => switch (category) {
.exe => .empty,
.none => .strip,
else => .keep,
},
};
}
const WriteCmd = union(enum) {
copy_range: struct { in_offset: u64, len: u64, out_offset: u64 },
write_data: struct { data: []const u8, out_offset: u64 },
};
fn write(allocator: Allocator, out_file: File, in_file: File, cmds: []const WriteCmd) !void {
// consolidate holes between writes:
// by coping original padding data from in_file (by fusing contiguous ranges)
// by writing zeroes otherwise
const zeroes = [1]u8{0} ** 4096;
var consolidated = std.ArrayList(WriteCmd).init(allocator);
defer consolidated.deinit();
try consolidated.ensureUnusedCapacity(cmds.len * 2);
var offset: u64 = 0;
var fused_cmd: ?WriteCmd = null;
for (cmds) |cmd| {
switch (cmd) {
.write_data => |data| {
assert(data.out_offset >= offset);
if (fused_cmd) |prev| {
consolidated.appendAssumeCapacity(prev);
fused_cmd = null;
}
if (data.out_offset > offset) {
consolidated.appendAssumeCapacity(.{ .write_data = .{ .data = zeroes[0..@intCast(data.out_offset - offset)], .out_offset = offset } });
}
consolidated.appendAssumeCapacity(cmd);
offset = data.out_offset + data.data.len;
},
.copy_range => |range| {
assert(range.out_offset >= offset);
if (fused_cmd) |prev| {
if (range.in_offset >= prev.copy_range.in_offset + prev.copy_range.len and (range.out_offset - prev.copy_range.out_offset == range.in_offset - prev.copy_range.in_offset)) {
fused_cmd = .{ .copy_range = .{
.in_offset = prev.copy_range.in_offset,
.out_offset = prev.copy_range.out_offset,
.len = (range.out_offset + range.len) - prev.copy_range.out_offset,
} };
} else {
consolidated.appendAssumeCapacity(prev);
if (range.out_offset > offset) {
consolidated.appendAssumeCapacity(.{ .write_data = .{ .data = zeroes[0..@intCast(range.out_offset - offset)], .out_offset = offset } });
}
fused_cmd = cmd;
}
} else {
fused_cmd = cmd;
}
offset = range.out_offset + range.len;
},
}
}
if (fused_cmd) |cmd| {
consolidated.appendAssumeCapacity(cmd);
}
// write the output file
for (consolidated.items) |cmd| {
switch (cmd) {
.write_data => |data| {
var iovec = [_]std.posix.iovec_const{.{ .base = data.data.ptr, .len = data.data.len }};
try out_file.pwritevAll(&iovec, data.out_offset);
},
.copy_range => |range| {
const copied_bytes = try in_file.copyRangeAll(range.in_offset, out_file, range.out_offset, range.len);
if (copied_bytes < range.len) return error.TRUNCATED_ELF;
},
}
}
}
fn tryCompressSection(allocator: Allocator, in_file: File, offset: u64, size: u64, prefix: []const u8) !?[]align(8) const u8 {
if (size < prefix.len) return null;
try in_file.seekTo(offset);
var section_reader = std.io.limitedReader(in_file.reader(), size);
// allocate as large as decompressed data. if the compression doesn't fit, keep the data uncompressed.
const compressed_data = try allocator.alignedAlloc(u8, 8, @intCast(size));
var compressed_stream = std.io.fixedBufferStream(compressed_data);
try compressed_stream.writer().writeAll(prefix);
{
var compressor = try std.compress.zlib.compressor(compressed_stream.writer(), .{});
var buf: [8000]u8 = undefined;
while (true) {
const bytes_read = try section_reader.read(&buf);
if (bytes_read == 0) break;
const bytes_written = compressor.write(buf[0..bytes_read]) catch |err| switch (err) {
error.NoSpaceLeft => {
allocator.free(compressed_data);
return null;
},
else => return err,
};
std.debug.assert(bytes_written == bytes_read);
}
compressor.finish() catch |err| switch (err) {
error.NoSpaceLeft => {
allocator.free(compressed_data);
return null;
},
else => return err,
};
}
const compressed_len: usize = @intCast(compressed_stream.getPos() catch unreachable);
const data = allocator.realloc(compressed_data, compressed_len) catch compressed_data;
return data[0..compressed_len];
}
fn createDebugLink(path: []const u8) DebugLink {
const file = std.fs.cwd().openFile(path, .{}) catch |err| {
fatal("zig objcopy: could not open `{s}`: {s}\n", .{ path, @errorName(err) });
};
defer file.close();
const crc = ElfFileHelper.computeFileCrc(file) catch |err| {
fatal("zig objcopy: could not read `{s}`: {s}\n", .{ path, @errorName(err) });
};
return .{
.name = std.fs.path.basename(path),
.crc32 = crc,
};
}
fn computeFileCrc(file: File) !u32 {
var buf: [8000]u8 = undefined;
try file.seekTo(0);
var hasher = std.hash.Crc32.init();
while (true) {
const bytes_read = try file.read(&buf);
if (bytes_read == 0) break;
hasher.update(buf[0..bytes_read]);
}
return hasher.final();
}
};
const SectionFlags = packed struct {
alloc: bool = false,
contents: bool = false,
load: bool = false,
noload: bool = false,
readonly: bool = false,
code: bool = false,
data: bool = false,
rom: bool = false,
exclude: bool = false,
shared: bool = false,
debug: bool = false,
large: bool = false,
merge: bool = false,
strings: bool = false,
};
fn parseSectionFlags(comma_separated_flags: []const u8) SectionFlags {
const P = struct {
fn parse(flags: *SectionFlags, string: []const u8) void {
if (string.len == 0) return;
if (std.mem.eql(u8, string, "alloc")) {
flags.alloc = true;
} else if (std.mem.eql(u8, string, "contents")) {
flags.contents = true;
} else if (std.mem.eql(u8, string, "load")) {
flags.load = true;
} else if (std.mem.eql(u8, string, "noload")) {
flags.noload = true;
} else if (std.mem.eql(u8, string, "readonly")) {
flags.readonly = true;
} else if (std.mem.eql(u8, string, "code")) {
flags.code = true;
} else if (std.mem.eql(u8, string, "data")) {
flags.data = true;
} else if (std.mem.eql(u8, string, "rom")) {
flags.rom = true;
} else if (std.mem.eql(u8, string, "exclude")) {
flags.exclude = true;
} else if (std.mem.eql(u8, string, "shared")) {
flags.shared = true;
} else if (std.mem.eql(u8, string, "debug")) {
flags.debug = true;
} else if (std.mem.eql(u8, string, "large")) {
flags.large = true;
} else if (std.mem.eql(u8, string, "merge")) {
flags.merge = true;
} else if (std.mem.eql(u8, string, "strings")) {
flags.strings = true;
} else {
std.log.warn("Skipping unrecognized section flag '{s}'", .{string});
}
}
};
var flags = SectionFlags{};
var offset: usize = 0;
for (comma_separated_flags, 0..) |c, i| {
if (c == ',') {
defer offset = i + 1;
const string = comma_separated_flags[offset..i];
P.parse(&flags, string);
}
}
P.parse(&flags, comma_separated_flags[offset..]);
return flags;
}
test "Parse section flags" {
const F = SectionFlags;
try std.testing.expectEqual(F{}, parseSectionFlags(""));
try std.testing.expectEqual(F{}, parseSectionFlags(","));
try std.testing.expectEqual(F{}, parseSectionFlags("abc"));
try std.testing.expectEqual(F{ .alloc = true }, parseSectionFlags("alloc"));
try std.testing.expectEqual(F{ .data = true }, parseSectionFlags("data,"));
try std.testing.expectEqual(F{ .alloc = true, .code = true }, parseSectionFlags("alloc,code"));
try std.testing.expectEqual(F{ .alloc = true, .code = true }, parseSectionFlags("alloc,code,not_supported"));
}
const SplitResult = struct { first: []const u8, second: []const u8 };
fn splitOption(option: []const u8) ?SplitResult {
const separator = '=';
if (option.len < 3) return null; // minimum "a=b"
for (1..option.len - 1) |i| {
if (option[i] == separator) return .{
.first = option[0..i],
.second = option[i + 1 ..],
};
}
return null;
}
test "Split option" {
{
const split = splitOption(".abc=123");
try std.testing.expect(split != null);
try std.testing.expectEqualStrings(".abc", split.?.first);
try std.testing.expectEqualStrings("123", split.?.second);
}
try std.testing.expectEqual(null, splitOption(""));
try std.testing.expectEqual(null, splitOption("=abc"));
try std.testing.expectEqual(null, splitOption("abc="));
try std.testing.expectEqual(null, splitOption("abc"));
}