Files
zig/src/link/Elf/Atom.zig
Jakub Konka 1a6d12ea92 elf: clean up and unify symbol ref handling in relocs
Also, this lets us re-enable proper undefined symbols tracking.
2023-09-12 23:27:14 +02:00

604 lines
22 KiB
Zig

/// Address allocated for this Atom.
value: u64 = 0,
/// Name of this Atom.
name_offset: u32 = 0,
/// Index into linker's input file table.
file_index: File.Index = 0,
/// Size of this atom
size: u64 = 0,
/// Alignment of this atom as a power of two.
alignment: u8 = 0,
/// Index of the input section.
input_section_index: Index = 0,
/// Index of the output section.
output_section_index: Index = 0,
/// Index of the input section containing this atom's relocs.
relocs_section_index: Index = 0,
/// Index of this atom in the linker's atoms table.
atom_index: Index = 0,
/// Specifies whether this atom is alive or has been garbage collected.
alive: bool = false,
/// Specifies if the atom has been visited during garbage collection.
visited: bool = false,
/// Start index of FDEs referencing this atom.
fde_start: u32 = 0,
/// End index of FDEs referencing this atom.
fde_end: u32 = 0,
/// Points to the previous and next neighbors, based on the `text_offset`.
/// This can be used to find, for example, the capacity of this `TextBlock`.
prev_index: Index = 0,
next_index: Index = 0,
pub fn name(self: Atom, elf_file: *Elf) []const u8 {
return elf_file.strtab.getAssumeExists(self.name_offset);
}
pub fn inputShdr(self: Atom, elf_file: *Elf) elf.Elf64_Shdr {
const object = elf_file.file(self.file_index).?.object;
return object.shdrs.items[self.input_section_index];
}
pub fn codeInObject(self: Atom, elf_file: *Elf) []const u8 {
const object = elf_file.file(self.file_index).?.object;
return object.shdrContents(self.input_section_index);
}
/// Returns atom's code and optionally uncompresses data if required (for compressed sections).
/// Caller owns the memory.
pub fn codeInObjectUncompressAlloc(self: Atom, elf_file: *Elf) ![]u8 {
const gpa = elf_file.base.allocator;
const data = self.codeInObject(elf_file);
const shdr = self.inputShdr(elf_file);
if (shdr.sh_flags & elf.SHF_COMPRESSED != 0) {
const chdr = @as(*align(1) const elf.Elf64_Chdr, @ptrCast(data.ptr)).*;
switch (chdr.ch_type) {
.ZLIB => {
var stream = std.io.fixedBufferStream(data[@sizeOf(elf.Elf64_Chdr)..]);
var zlib_stream = std.compress.zlib.decompressStream(gpa, stream.reader()) catch
return error.InputOutput;
defer zlib_stream.deinit();
const decomp = try gpa.alloc(u8, chdr.ch_size);
const nread = zlib_stream.reader().readAll(decomp) catch return error.InputOutput;
if (nread != decomp.len) {
return error.InputOutput;
}
return decomp;
},
else => @panic("TODO unhandled compression scheme"),
}
} else return gpa.dupe(u8, data);
}
pub fn priority(self: Atom, elf_file: *Elf) u64 {
const index = elf_file.file(self.file_index).?.index();
return (@as(u64, @intCast(index)) << 32) | @as(u64, @intCast(self.input_section_index));
}
/// Returns how much room there is to grow in virtual address space.
/// File offset relocation happens transparently, so it is not included in
/// this calculation.
pub fn capacity(self: Atom, elf_file: *Elf) u64 {
const next_value = if (elf_file.atom(self.next_index)) |next| next.value else std.math.maxInt(u32);
return next_value - self.value;
}
pub fn freeListEligible(self: Atom, elf_file: *Elf) bool {
// No need to keep a free list node for the last block.
const next = elf_file.atom(self.next_index) orelse return false;
const cap = next.value - self.value;
const ideal_cap = Elf.padToIdeal(self.size);
if (cap <= ideal_cap) return false;
const surplus = cap - ideal_cap;
return surplus >= Elf.min_text_capacity;
}
pub fn allocate(self: *Atom, elf_file: *Elf) !void {
const shdr = &elf_file.shdrs.items[self.output_section_index];
const meta = elf_file.last_atom_and_free_list_table.getPtr(self.output_section_index).?;
const free_list = &meta.free_list;
const last_atom_index = &meta.last_atom_index;
const new_atom_ideal_capacity = Elf.padToIdeal(self.size);
const alignment = try std.math.powi(u64, 2, self.alignment);
// We use these to indicate our intention to update metadata, placing the new atom,
// and possibly removing a free list node.
// It would be simpler to do it inside the for loop below, but that would cause a
// problem if an error was returned later in the function. So this action
// is actually carried out at the end of the function, when errors are no longer possible.
var atom_placement: ?Atom.Index = null;
var free_list_removal: ?usize = null;
// First we look for an appropriately sized free list node.
// The list is unordered. We'll just take the first thing that works.
self.value = blk: {
var i: usize = if (elf_file.base.child_pid == null) 0 else free_list.items.len;
while (i < free_list.items.len) {
const big_atom_index = free_list.items[i];
const big_atom = elf_file.atom(big_atom_index).?;
// We now have a pointer to a live atom that has too much capacity.
// Is it enough that we could fit this new atom?
const cap = big_atom.capacity(elf_file);
const ideal_capacity = Elf.padToIdeal(cap);
const ideal_capacity_end_vaddr = std.math.add(u64, big_atom.value, ideal_capacity) catch ideal_capacity;
const capacity_end_vaddr = big_atom.value + cap;
const new_start_vaddr_unaligned = capacity_end_vaddr - new_atom_ideal_capacity;
const new_start_vaddr = std.mem.alignBackward(u64, new_start_vaddr_unaligned, alignment);
if (new_start_vaddr < ideal_capacity_end_vaddr) {
// Additional bookkeeping here to notice if this free list node
// should be deleted because the block that it points to has grown to take up
// more of the extra capacity.
if (!big_atom.freeListEligible(elf_file)) {
_ = free_list.swapRemove(i);
} else {
i += 1;
}
continue;
}
// At this point we know that we will place the new block here. But the
// remaining question is whether there is still yet enough capacity left
// over for there to still be a free list node.
const remaining_capacity = new_start_vaddr - ideal_capacity_end_vaddr;
const keep_free_list_node = remaining_capacity >= Elf.min_text_capacity;
// Set up the metadata to be updated, after errors are no longer possible.
atom_placement = big_atom_index;
if (!keep_free_list_node) {
free_list_removal = i;
}
break :blk new_start_vaddr;
} else if (elf_file.atom(last_atom_index.*)) |last| {
const ideal_capacity = Elf.padToIdeal(last.size);
const ideal_capacity_end_vaddr = last.value + ideal_capacity;
const new_start_vaddr = std.mem.alignForward(u64, ideal_capacity_end_vaddr, alignment);
// Set up the metadata to be updated, after errors are no longer possible.
atom_placement = last.atom_index;
break :blk new_start_vaddr;
} else {
break :blk shdr.sh_addr;
}
};
const expand_section = if (atom_placement) |placement_index|
elf_file.atom(placement_index).?.next_index == 0
else
true;
if (expand_section) {
const needed_size = (self.value + self.size) - shdr.sh_addr;
try elf_file.growAllocSection(self.output_section_index, needed_size);
last_atom_index.* = self.atom_index;
if (elf_file.dwarf) |_| {
// The .debug_info section has `low_pc` and `high_pc` values which is the virtual address
// range of the compilation unit. When we expand the text section, this range changes,
// so the DW_TAG.compile_unit tag of the .debug_info section becomes dirty.
elf_file.debug_info_header_dirty = true;
// This becomes dirty for the same reason. We could potentially make this more
// fine-grained with the addition of support for more compilation units. It is planned to
// model each package as a different compilation unit.
elf_file.debug_aranges_section_dirty = true;
}
}
shdr.sh_addralign = @max(shdr.sh_addralign, alignment);
// This function can also reallocate an atom.
// In this case we need to "unplug" it from its previous location before
// plugging it in to its new location.
if (elf_file.atom(self.prev_index)) |prev| {
prev.next_index = self.next_index;
}
if (elf_file.atom(self.next_index)) |next| {
next.prev_index = self.prev_index;
}
if (atom_placement) |big_atom_index| {
const big_atom = elf_file.atom(big_atom_index).?;
self.prev_index = big_atom_index;
self.next_index = big_atom.next_index;
big_atom.next_index = self.atom_index;
} else {
self.prev_index = 0;
self.next_index = 0;
}
if (free_list_removal) |i| {
_ = free_list.swapRemove(i);
}
}
pub fn shrink(self: *Atom, elf_file: *Elf) void {
_ = self;
_ = elf_file;
}
pub fn grow(self: *Atom, elf_file: *Elf) !void {
const alignment = try std.math.powi(u64, 2, self.alignment);
const align_ok = std.mem.alignBackward(u64, self.value, alignment) == self.value;
const need_realloc = !align_ok or self.size > self.capacity(elf_file);
if (need_realloc) try self.allocate(elf_file);
}
pub fn free(self: *Atom, elf_file: *Elf) void {
log.debug("freeAtom {d} ({s})", .{ self.atom_index, self.name(elf_file) });
const gpa = elf_file.base.allocator;
const zig_module = elf_file.file(self.file_index).?.zig_module;
const shndx = self.output_section_index;
const meta = elf_file.last_atom_and_free_list_table.getPtr(shndx).?;
const free_list = &meta.free_list;
const last_atom_index = &meta.last_atom_index;
var already_have_free_list_node = false;
{
var i: usize = 0;
// TODO turn free_list into a hash map
while (i < free_list.items.len) {
if (free_list.items[i] == self.atom_index) {
_ = free_list.swapRemove(i);
continue;
}
if (free_list.items[i] == self.prev_index) {
already_have_free_list_node = true;
}
i += 1;
}
}
if (elf_file.atom(last_atom_index.*)) |last_atom| {
if (last_atom.atom_index == self.atom_index) {
if (elf_file.atom(self.prev_index)) |_| {
// TODO shrink the section size here
last_atom_index.* = self.prev_index;
} else {
last_atom_index.* = 0;
}
}
}
if (elf_file.atom(self.prev_index)) |prev| {
prev.next_index = self.next_index;
if (!already_have_free_list_node and prev.*.freeListEligible(elf_file)) {
// The free list is heuristics, it doesn't have to be perfect, so we can
// ignore the OOM here.
free_list.append(gpa, prev.atom_index) catch {};
}
} else {
self.prev_index = 0;
}
if (elf_file.atom(self.next_index)) |next| {
next.prev_index = self.prev_index;
} else {
self.next_index = 0;
}
// TODO create relocs free list
self.freeRelocs(elf_file);
assert(zig_module.atoms.swapRemove(self.atom_index));
self.* = .{};
}
pub fn relocs(self: Atom, elf_file: *Elf) []align(1) const elf.Elf64_Rela {
return switch (elf_file.file(self.file_index).?) {
.zig_module => |x| x.relocs.items[self.relocs_section_index].items,
.object => |x| x.getRelocs(self.relocs_section_index),
else => unreachable,
};
}
pub fn addReloc(self: Atom, elf_file: *Elf, reloc: elf.Elf64_Rela) !void {
const gpa = elf_file.base.allocator;
const file_ptr = elf_file.file(self.file_index).?;
assert(file_ptr == .zig_module);
const zig_module = file_ptr.zig_module;
const rels = &zig_module.relocs.items[self.relocs_section_index];
try rels.append(gpa, reloc);
}
pub fn freeRelocs(self: Atom, elf_file: *Elf) void {
const file_ptr = elf_file.file(self.file_index).?;
assert(file_ptr == .zig_module);
const zig_module = file_ptr.zig_module;
zig_module.relocs.items[self.relocs_section_index].clearRetainingCapacity();
}
pub fn scanRelocs(self: Atom, elf_file: *Elf, undefs: anytype) !void {
const file_ptr = elf_file.file(self.file_index).?;
const rels = self.relocs(elf_file);
var i: usize = 0;
while (i < rels.len) : (i += 1) {
const rel = rels[i];
if (rel.r_type() == elf.R_X86_64_NONE) continue;
const symbol = switch (file_ptr) {
.zig_module => |x| elf_file.symbol(x.symbol(rel.r_sym())),
.object => |x| elf_file.symbol(x.symbols.items[rel.r_sym()]),
else => unreachable,
};
// Check for violation of One Definition Rule for COMDATs.
if (symbol.file(elf_file) == null) {
// TODO convert into an error
log.debug("{}: {s}: {s} refers to a discarded COMDAT section", .{
file_ptr.fmtPath(),
self.name(elf_file),
symbol.name(elf_file),
});
continue;
}
// Report an undefined symbol.
try self.reportUndefined(elf_file, symbol, rel, undefs);
// While traversing relocations, mark symbols that require special handling such as
// pointer indirection via GOT, or a stub trampoline via PLT.
switch (rel.r_type()) {
elf.R_X86_64_64 => {},
elf.R_X86_64_32,
elf.R_X86_64_32S,
=> {},
elf.R_X86_64_GOT32,
elf.R_X86_64_GOT64,
elf.R_X86_64_GOTPC32,
elf.R_X86_64_GOTPC64,
elf.R_X86_64_GOTPCREL,
elf.R_X86_64_GOTPCREL64,
elf.R_X86_64_GOTPCRELX,
elf.R_X86_64_REX_GOTPCRELX,
=> {
symbol.flags.needs_got = true;
},
elf.R_X86_64_PLT32,
elf.R_X86_64_PLTOFF64,
=> {
if (symbol.flags.import) {
symbol.flags.needs_plt = true;
}
},
elf.R_X86_64_PC32 => {},
else => @panic("TODO"),
}
}
}
// This function will report any undefined non-weak symbols that are not imports.
fn reportUndefined(self: Atom, elf_file: *Elf, sym: *const Symbol, rel: elf.Elf64_Rela, undefs: anytype) !void {
const rel_esym = switch (elf_file.file(self.file_index).?) {
.zig_module => |x| x.elfSym(rel.r_sym()).*,
.object => |x| x.symtab[rel.r_sym()],
else => unreachable,
};
const esym = sym.elfSym(elf_file);
if (rel_esym.st_shndx == elf.SHN_UNDEF and
rel_esym.st_bind() == elf.STB_GLOBAL and
sym.esym_index > 0 and
!sym.flags.import and
esym.st_shndx == elf.SHN_UNDEF)
{
const gop = try undefs.getOrPut(sym.index);
if (!gop.found_existing) {
gop.value_ptr.* = std.ArrayList(Atom.Index).init(elf_file.base.allocator);
}
try gop.value_ptr.append(self.atom_index);
}
}
/// TODO mark relocs dirty
pub fn resolveRelocs(self: Atom, elf_file: *Elf, code: []u8) !void {
relocs_log.debug("0x{x}: {s}", .{ self.value, self.name(elf_file) });
const file_ptr = elf_file.file(self.file_index).?;
var stream = std.io.fixedBufferStream(code);
const cwriter = stream.writer();
for (self.relocs(elf_file)) |rel| {
const r_type = rel.r_type();
if (r_type == elf.R_X86_64_NONE) continue;
const target = switch (file_ptr) {
.zig_module => |x| elf_file.symbol(x.symbol(rel.r_sym())),
.object => |x| elf_file.symbol(x.symbols.items[rel.r_sym()]),
else => unreachable,
};
// We will use equation format to resolve relocations:
// https://intezer.com/blog/malware-analysis/executable-and-linkable-format-101-part-3-relocations/
//
// Address of the source atom.
const P = @as(i64, @intCast(self.value + rel.r_offset));
// Addend from the relocation.
const A = rel.r_addend;
// Address of the target symbol - can be address of the symbol within an atom or address of PLT stub.
const S = @as(i64, @intCast(target.address(.{}, elf_file)));
// Address of the global offset table.
const GOT = blk: {
const shndx = if (elf_file.got_plt_section_index) |shndx|
shndx
else if (elf_file.got_section_index) |shndx|
shndx
else
null;
break :blk if (shndx) |index| @as(i64, @intCast(elf_file.shdrs.items[index].sh_addr)) else 0;
};
// Relative offset to the start of the global offset table.
const G = @as(i64, @intCast(target.gotAddress(elf_file))) - GOT;
// // Address of the thread pointer.
// const TP = @as(i64, @intCast(elf_file.getTpAddress()));
// // Address of the dynamic thread pointer.
// const DTP = @as(i64, @intCast(elf_file.getDtpAddress()));
relocs_log.debug(" {s}: {x}: [{x} => {x}] G({x}) ({s})", .{
fmtRelocType(r_type),
rel.r_offset,
P,
S + A,
G + GOT + A,
target.name(elf_file),
});
try stream.seekTo(rel.r_offset);
switch (rel.r_type()) {
elf.R_X86_64_NONE => unreachable,
elf.R_X86_64_64 => try cwriter.writeIntLittle(i64, S + A),
elf.R_X86_64_PLT32,
elf.R_X86_64_PC32,
=> try cwriter.writeIntLittle(i32, @as(i32, @intCast(S + A - P))),
else => {
log.err("TODO: unhandled relocation type {}", .{fmtRelocType(rel.r_type())});
@panic("TODO unhandled relocation type");
},
}
}
}
pub fn fmtRelocType(r_type: u32) std.fmt.Formatter(formatRelocType) {
return .{ .data = r_type };
}
fn formatRelocType(
r_type: u32,
comptime unused_fmt_string: []const u8,
options: std.fmt.FormatOptions,
writer: anytype,
) !void {
_ = unused_fmt_string;
_ = options;
const str = switch (r_type) {
elf.R_X86_64_NONE => "R_X86_64_NONE",
elf.R_X86_64_64 => "R_X86_64_64",
elf.R_X86_64_PC32 => "R_X86_64_PC32",
elf.R_X86_64_GOT32 => "R_X86_64_GOT32",
elf.R_X86_64_PLT32 => "R_X86_64_PLT32",
elf.R_X86_64_COPY => "R_X86_64_COPY",
elf.R_X86_64_GLOB_DAT => "R_X86_64_GLOB_DAT",
elf.R_X86_64_JUMP_SLOT => "R_X86_64_JUMP_SLOT",
elf.R_X86_64_RELATIVE => "R_X86_64_RELATIVE",
elf.R_X86_64_GOTPCREL => "R_X86_64_GOTPCREL",
elf.R_X86_64_32 => "R_X86_64_32",
elf.R_X86_64_32S => "R_X86_64_32S",
elf.R_X86_64_16 => "R_X86_64_16",
elf.R_X86_64_PC16 => "R_X86_64_PC16",
elf.R_X86_64_8 => "R_X86_64_8",
elf.R_X86_64_PC8 => "R_X86_64_PC8",
elf.R_X86_64_DTPMOD64 => "R_X86_64_DTPMOD64",
elf.R_X86_64_DTPOFF64 => "R_X86_64_DTPOFF64",
elf.R_X86_64_TPOFF64 => "R_X86_64_TPOFF64",
elf.R_X86_64_TLSGD => "R_X86_64_TLSGD",
elf.R_X86_64_TLSLD => "R_X86_64_TLSLD",
elf.R_X86_64_DTPOFF32 => "R_X86_64_DTPOFF32",
elf.R_X86_64_GOTTPOFF => "R_X86_64_GOTTPOFF",
elf.R_X86_64_TPOFF32 => "R_X86_64_TPOFF32",
elf.R_X86_64_PC64 => "R_X86_64_PC64",
elf.R_X86_64_GOTOFF64 => "R_X86_64_GOTOFF64",
elf.R_X86_64_GOTPC32 => "R_X86_64_GOTPC32",
elf.R_X86_64_GOT64 => "R_X86_64_GOT64",
elf.R_X86_64_GOTPCREL64 => "R_X86_64_GOTPCREL64",
elf.R_X86_64_GOTPC64 => "R_X86_64_GOTPC64",
elf.R_X86_64_GOTPLT64 => "R_X86_64_GOTPLT64",
elf.R_X86_64_PLTOFF64 => "R_X86_64_PLTOFF64",
elf.R_X86_64_SIZE32 => "R_X86_64_SIZE32",
elf.R_X86_64_SIZE64 => "R_X86_64_SIZE64",
elf.R_X86_64_GOTPC32_TLSDESC => "R_X86_64_GOTPC32_TLSDESC",
elf.R_X86_64_TLSDESC_CALL => "R_X86_64_TLSDESC_CALL",
elf.R_X86_64_TLSDESC => "R_X86_64_TLSDESC",
elf.R_X86_64_IRELATIVE => "R_X86_64_IRELATIVE",
elf.R_X86_64_RELATIVE64 => "R_X86_64_RELATIVE64",
elf.R_X86_64_GOTPCRELX => "R_X86_64_GOTPCRELX",
elf.R_X86_64_REX_GOTPCRELX => "R_X86_64_REX_GOTPCRELX",
elf.R_X86_64_NUM => "R_X86_64_NUM",
else => "R_X86_64_UNKNOWN",
};
try writer.print("{s}", .{str});
}
pub fn format(
atom: Atom,
comptime unused_fmt_string: []const u8,
options: std.fmt.FormatOptions,
writer: anytype,
) !void {
_ = atom;
_ = unused_fmt_string;
_ = options;
_ = writer;
@compileError("do not format symbols directly");
}
pub fn fmt(atom: Atom, elf_file: *Elf) std.fmt.Formatter(format2) {
return .{ .data = .{
.atom = atom,
.elf_file = elf_file,
} };
}
const FormatContext = struct {
atom: Atom,
elf_file: *Elf,
};
fn format2(
ctx: FormatContext,
comptime unused_fmt_string: []const u8,
options: std.fmt.FormatOptions,
writer: anytype,
) !void {
_ = options;
_ = unused_fmt_string;
const atom = ctx.atom;
const elf_file = ctx.elf_file;
try writer.print("atom({d}) : {s} : @{x} : sect({d}) : align({x}) : size({x})", .{
atom.atom_index, atom.name(elf_file), atom.value,
atom.output_section_index, atom.alignment, atom.size,
});
// if (atom.fde_start != atom.fde_end) {
// try writer.writeAll(" : fdes{ ");
// for (atom.getFdes(elf_file), atom.fde_start..) |fde, i| {
// try writer.print("{d}", .{i});
// if (!fde.alive) try writer.writeAll("([*])");
// if (i < atom.fde_end - 1) try writer.writeAll(", ");
// }
// try writer.writeAll(" }");
// }
const gc_sections = if (elf_file.base.options.gc_sections) |gc_sections| gc_sections else false;
if (gc_sections and !atom.alive) {
try writer.writeAll(" : [*]");
}
}
// TODO this has to be u32 but for now, to avoid redesigning elfSym machinery for
// ZigModule, keep it at u16 with the intention of bumping it to u32 in the near
// future.
pub const Index = u16;
const std = @import("std");
const assert = std.debug.assert;
const elf = std.elf;
const log = std.log.scoped(.link);
const relocs_log = std.log.scoped(.link_relocs);
const Allocator = std.mem.Allocator;
const Atom = @This();
const Elf = @import("../Elf.zig");
const File = @import("file.zig").File;
const Symbol = @import("Symbol.zig");