Files
zig/src/link/MachO/DebugSymbols.zig

468 lines
16 KiB
Zig

allocator: Allocator,
file: fs.File,
symtab_cmd: macho.symtab_command = .{},
uuid_cmd: macho.uuid_command = .{ .uuid = [_]u8{0} ** 16 },
segments: std.ArrayListUnmanaged(macho.segment_command_64) = .empty,
sections: std.ArrayListUnmanaged(macho.section_64) = .empty,
dwarf_segment_cmd_index: ?u8 = null,
linkedit_segment_cmd_index: ?u8 = null,
debug_info_section_index: ?u8 = null,
debug_abbrev_section_index: ?u8 = null,
debug_str_section_index: ?u8 = null,
debug_aranges_section_index: ?u8 = null,
debug_line_section_index: ?u8 = null,
debug_line_str_section_index: ?u8 = null,
debug_loclists_section_index: ?u8 = null,
debug_rnglists_section_index: ?u8 = null,
relocs: std.ArrayListUnmanaged(Reloc) = .empty,
/// Output synthetic sections
symtab: std.ArrayListUnmanaged(macho.nlist_64) = .empty,
strtab: std.ArrayListUnmanaged(u8) = .empty,
pub const Reloc = struct {
type: enum {
direct_load,
got_load,
},
target: u32,
offset: u64,
addend: u32,
};
/// You must call this function *after* `ZigObject.initMetadata()`
/// has been called to get a viable debug symbols output.
pub fn initMetadata(self: *DebugSymbols, macho_file: *MachO) !void {
try self.strtab.append(self.allocator, 0);
{
self.dwarf_segment_cmd_index = @as(u8, @intCast(self.segments.items.len));
const page_size = macho_file.getPageSize();
const off = @as(u64, @intCast(page_size));
const ideal_size: u16 = 200 + 128 + 160 + 250;
const needed_size = mem.alignForward(u64, padToIdeal(ideal_size), page_size);
log.debug("found __DWARF segment free space 0x{x} to 0x{x}", .{ off, off + needed_size });
try self.segments.append(self.allocator, .{
.segname = makeStaticString("__DWARF"),
.vmsize = needed_size,
.fileoff = off,
.filesize = needed_size,
.cmdsize = @sizeOf(macho.segment_command_64),
});
}
self.debug_str_section_index = try self.createSection("__debug_str", 0);
self.debug_info_section_index = try self.createSection("__debug_info", 0);
self.debug_abbrev_section_index = try self.createSection("__debug_abbrev", 0);
self.debug_aranges_section_index = try self.createSection("__debug_aranges", 4);
self.debug_line_section_index = try self.createSection("__debug_line", 0);
self.debug_line_str_section_index = try self.createSection("__debug_line_str", 0);
self.debug_loclists_section_index = try self.createSection("__debug_loclists", 0);
self.debug_rnglists_section_index = try self.createSection("__debug_rnglists", 0);
self.linkedit_segment_cmd_index = @intCast(self.segments.items.len);
try self.segments.append(self.allocator, .{
.segname = makeStaticString("__LINKEDIT"),
.maxprot = macho.PROT.READ,
.initprot = macho.PROT.READ,
.cmdsize = @sizeOf(macho.segment_command_64),
});
}
fn createSection(self: *DebugSymbols, sectname: []const u8, alignment: u16) !u8 {
const segment = self.getDwarfSegmentPtr();
var sect = macho.section_64{
.sectname = makeStaticString(sectname),
.segname = segment.segname,
.@"align" = alignment,
};
log.debug("create {s},{s} section", .{ sect.segName(), sect.sectName() });
const index: u8 = @intCast(self.sections.items.len);
try self.sections.append(self.allocator, sect);
segment.cmdsize += @sizeOf(macho.section_64);
segment.nsects += 1;
return index;
}
pub fn growSection(
self: *DebugSymbols,
sect_index: u8,
needed_size: u64,
requires_file_copy: bool,
macho_file: *MachO,
) !void {
const sect = self.getSectionPtr(sect_index);
const allocated_size = self.allocatedSize(sect.offset);
if (needed_size > allocated_size) {
const existing_size = sect.size;
sect.size = 0; // free the space
const new_offset = try self.findFreeSpace(needed_size, 1);
log.debug("moving {s} section: {} bytes from 0x{x} to 0x{x}", .{
sect.sectName(),
existing_size,
sect.offset,
new_offset,
});
if (requires_file_copy) {
const amt = try self.file.copyRangeAll(
sect.offset,
self.file,
new_offset,
existing_size,
);
if (amt != existing_size) return error.InputOutput;
}
sect.offset = @intCast(new_offset);
} else if (sect.offset + allocated_size == std.math.maxInt(u64)) {
try self.file.setEndPos(sect.offset + needed_size);
}
sect.size = needed_size;
self.markDirty(sect_index, macho_file);
}
pub fn markDirty(self: *DebugSymbols, sect_index: u8, macho_file: *MachO) void {
if (macho_file.getZigObject()) |zo| {
if (self.debug_info_section_index.? == sect_index) {
zo.debug_info_header_dirty = true;
} else if (self.debug_line_section_index.? == sect_index) {
zo.debug_line_header_dirty = true;
} else if (self.debug_abbrev_section_index.? == sect_index) {
zo.debug_abbrev_dirty = true;
} else if (self.debug_str_section_index.? == sect_index) {
zo.debug_strtab_dirty = true;
} else if (self.debug_aranges_section_index.? == sect_index) {
zo.debug_aranges_dirty = true;
}
}
}
fn detectAllocCollision(self: *DebugSymbols, start: u64, size: u64) !?u64 {
var at_end = true;
const end = start + padToIdeal(size);
for (self.sections.items) |section| {
const increased_size = padToIdeal(section.size);
const test_end = section.offset + increased_size;
if (start < test_end) {
if (end > section.offset) return test_end;
if (test_end < std.math.maxInt(u64)) at_end = false;
}
}
if (at_end) try self.file.setEndPos(end);
return null;
}
fn findFreeSpace(self: *DebugSymbols, object_size: u64, min_alignment: u64) !u64 {
const segment = self.getDwarfSegmentPtr();
var offset: u64 = segment.fileoff;
while (try self.detectAllocCollision(offset, object_size)) |item_end| {
offset = mem.alignForward(u64, item_end, min_alignment);
}
return offset;
}
pub fn flushModule(self: *DebugSymbols, macho_file: *MachO) !void {
const zo = macho_file.getZigObject().?;
for (self.relocs.items) |*reloc| {
const sym = zo.symbols.items[reloc.target];
const sym_name = sym.getName(macho_file);
const addr = switch (reloc.type) {
.direct_load => sym.getAddress(.{}, macho_file),
.got_load => sym.getGotAddress(macho_file),
};
const sect = &self.sections.items[self.debug_info_section_index.?];
const file_offset = sect.offset + reloc.offset;
log.debug("resolving relocation: {d}@{x} ('{s}') at offset {x}", .{
reloc.target,
addr,
sym_name,
file_offset,
});
try self.file.pwriteAll(mem.asBytes(&addr), file_offset);
}
self.finalizeDwarfSegment(macho_file);
try self.writeLinkeditSegmentData(macho_file);
// Write load commands
const ncmds, const sizeofcmds = try self.writeLoadCommands(macho_file);
try self.writeHeader(macho_file, ncmds, sizeofcmds);
}
pub fn deinit(self: *DebugSymbols) void {
const gpa = self.allocator;
self.file.close();
self.segments.deinit(gpa);
self.sections.deinit(gpa);
self.relocs.deinit(gpa);
self.symtab.deinit(gpa);
self.strtab.deinit(gpa);
}
pub fn swapRemoveRelocs(self: *DebugSymbols, target: u32) void {
// TODO re-implement using a hashmap with free lists
var last_index: usize = 0;
while (last_index < self.relocs.items.len) {
const reloc = self.relocs.items[last_index];
if (reloc.target == target) {
_ = self.relocs.swapRemove(last_index);
} else {
last_index += 1;
}
}
}
fn finalizeDwarfSegment(self: *DebugSymbols, macho_file: *MachO) void {
const base_vmaddr = blk: {
// Note that we purposely take the last VM address of the MachO binary including
// the binary's LINKEDIT segment. This is in contrast to how dsymutil does it
// which overwrites the the address space taken by the original MachO binary,
// however at the cost of having LINKEDIT preceed DWARF in dSYM binary which we
// do not want as we want to be able to incrementally move DWARF sections in the
// file as we please.
const last_seg = macho_file.getLinkeditSegment();
break :blk last_seg.vmaddr + last_seg.vmsize;
};
const dwarf_segment = self.getDwarfSegmentPtr();
var file_size: u64 = 0;
for (self.sections.items) |header| {
file_size = @max(file_size, header.offset + header.size);
}
const page_size = macho_file.getPageSize();
const aligned_size = mem.alignForward(u64, file_size, page_size);
dwarf_segment.vmaddr = base_vmaddr;
dwarf_segment.filesize = aligned_size;
dwarf_segment.vmsize = aligned_size;
const linkedit = self.getLinkeditSegmentPtr();
linkedit.vmaddr = mem.alignForward(
u64,
dwarf_segment.vmaddr + aligned_size,
page_size,
);
linkedit.fileoff = mem.alignForward(
u64,
dwarf_segment.fileoff + aligned_size,
page_size,
);
log.debug("found __LINKEDIT segment free space at 0x{x}", .{linkedit.fileoff});
}
fn writeLoadCommands(self: *DebugSymbols, macho_file: *MachO) !struct { usize, usize } {
const gpa = self.allocator;
const needed_size = load_commands.calcLoadCommandsSizeDsym(macho_file, self);
const buffer = try gpa.alloc(u8, needed_size);
defer gpa.free(buffer);
var stream = std.io.fixedBufferStream(buffer);
const writer = stream.writer();
var ncmds: usize = 0;
// UUID comes first presumably to speed up lookup by the consumer like lldb.
@memcpy(&self.uuid_cmd.uuid, &macho_file.uuid_cmd.uuid);
try writer.writeStruct(self.uuid_cmd);
ncmds += 1;
// Segment and section load commands
{
// Write segment/section headers from the binary file first.
const slice = macho_file.sections.slice();
var sect_id: usize = 0;
for (macho_file.segments.items, 0..) |seg, seg_id| {
if (seg_id == macho_file.linkedit_seg_index.?) break;
var out_seg = seg;
out_seg.fileoff = 0;
out_seg.filesize = 0;
try writer.writeStruct(out_seg);
for (slice.items(.header)[sect_id..][0..seg.nsects]) |header| {
var out_header = header;
out_header.offset = 0;
try writer.writeStruct(out_header);
}
sect_id += seg.nsects;
}
ncmds += macho_file.segments.items.len - 1;
// Next, commit DSYM's __LINKEDIT and __DWARF segments headers.
sect_id = 0;
for (self.segments.items) |seg| {
try writer.writeStruct(seg);
for (self.sections.items[sect_id..][0..seg.nsects]) |header| {
try writer.writeStruct(header);
}
sect_id += seg.nsects;
}
ncmds += self.segments.items.len;
}
try writer.writeStruct(self.symtab_cmd);
ncmds += 1;
assert(stream.pos == needed_size);
try self.file.pwriteAll(buffer, @sizeOf(macho.mach_header_64));
return .{ ncmds, buffer.len };
}
fn writeHeader(self: *DebugSymbols, macho_file: *MachO, ncmds: usize, sizeofcmds: usize) !void {
var header: macho.mach_header_64 = .{};
header.filetype = macho.MH_DSYM;
switch (macho_file.getTarget().cpu.arch) {
.aarch64 => {
header.cputype = macho.CPU_TYPE_ARM64;
header.cpusubtype = macho.CPU_SUBTYPE_ARM_ALL;
},
.x86_64 => {
header.cputype = macho.CPU_TYPE_X86_64;
header.cpusubtype = macho.CPU_SUBTYPE_X86_64_ALL;
},
else => return error.UnsupportedCpuArchitecture,
}
header.ncmds = @intCast(ncmds);
header.sizeofcmds = @intCast(sizeofcmds);
log.debug("writing Mach-O header {}", .{header});
try self.file.pwriteAll(mem.asBytes(&header), 0);
}
fn allocatedSize(self: *DebugSymbols, start: u64) u64 {
if (start == 0) return 0;
const seg = self.getDwarfSegmentPtr();
assert(start >= seg.fileoff);
var min_pos: u64 = std.math.maxInt(u64);
for (self.sections.items) |section| {
if (section.offset <= start) continue;
if (section.offset < min_pos) min_pos = section.offset;
}
return min_pos - start;
}
fn writeLinkeditSegmentData(self: *DebugSymbols, macho_file: *MachO) !void {
const tracy = trace(@src());
defer tracy.end();
const page_size = macho_file.getPageSize();
const seg = &self.segments.items[self.linkedit_segment_cmd_index.?];
var off = math.cast(u32, seg.fileoff) orelse return error.Overflow;
off = try self.writeSymtab(off, macho_file);
off = mem.alignForward(u32, off, @alignOf(u64));
off = try self.writeStrtab(off);
seg.filesize = off - seg.fileoff;
const aligned_size = mem.alignForward(u64, seg.filesize, page_size);
seg.vmsize = aligned_size;
}
pub fn writeSymtab(self: *DebugSymbols, off: u32, macho_file: *MachO) !u32 {
const tracy = trace(@src());
defer tracy.end();
const gpa = self.allocator;
const cmd = &self.symtab_cmd;
cmd.nsyms = macho_file.symtab_cmd.nsyms;
cmd.strsize = macho_file.symtab_cmd.strsize;
cmd.symoff = off;
try self.symtab.resize(gpa, cmd.nsyms);
try self.strtab.resize(gpa, cmd.strsize);
self.strtab.items[0] = 0;
if (macho_file.getZigObject()) |zo| {
zo.writeSymtab(macho_file, self);
}
for (macho_file.objects.items) |index| {
macho_file.getFile(index).?.writeSymtab(macho_file, self);
}
for (macho_file.dylibs.items) |index| {
macho_file.getFile(index).?.writeSymtab(macho_file, self);
}
if (macho_file.getInternalObject()) |internal| {
internal.writeSymtab(macho_file, self);
}
try self.file.pwriteAll(mem.sliceAsBytes(self.symtab.items), cmd.symoff);
return off + cmd.nsyms * @sizeOf(macho.nlist_64);
}
pub fn writeStrtab(self: *DebugSymbols, off: u32) !u32 {
const cmd = &self.symtab_cmd;
cmd.stroff = off;
try self.file.pwriteAll(self.strtab.items, cmd.stroff);
return off + cmd.strsize;
}
pub fn getSectionIndexes(self: *DebugSymbols, segment_index: u8) struct { start: u8, end: u8 } {
var start: u8 = 0;
const nsects: u8 = for (self.segments.items, 0..) |seg, i| {
if (i == segment_index) break @intCast(seg.nsects);
start += @intCast(seg.nsects);
} else 0;
return .{ .start = start, .end = start + nsects };
}
fn getDwarfSegmentPtr(self: *DebugSymbols) *macho.segment_command_64 {
const index = self.dwarf_segment_cmd_index.?;
return &self.segments.items[index];
}
fn getLinkeditSegmentPtr(self: *DebugSymbols) *macho.segment_command_64 {
const index = self.linkedit_segment_cmd_index.?;
return &self.segments.items[index];
}
pub fn getSectionPtr(self: *DebugSymbols, sect: u8) *macho.section_64 {
assert(sect < self.sections.items.len);
return &self.sections.items[sect];
}
pub fn getSection(self: DebugSymbols, sect: u8) macho.section_64 {
assert(sect < self.sections.items.len);
return self.sections.items[sect];
}
const DebugSymbols = @This();
const std = @import("std");
const build_options = @import("build_options");
const assert = std.debug.assert;
const fs = std.fs;
const link = @import("../../link.zig");
const load_commands = @import("load_commands.zig");
const log = std.log.scoped(.link_dsym);
const macho = std.macho;
const makeStaticString = MachO.makeStaticString;
const math = std.math;
const mem = std.mem;
const padToIdeal = MachO.padToIdeal;
const trace = @import("../../tracy.zig").trace;
const Allocator = mem.Allocator;
const MachO = @import("../MachO.zig");
const StringTable = @import("../StringTable.zig");
const Type = @import("../../Type.zig");