zig/src/link/C.zig
Andrew Kelley db33ee45b7 rework generic function calls
Abridged summary:

 * Move `Module.Fn` into `InternPool`.
 * Delete a lot of confusing and problematic `Sema` logic related to
   generic function calls.

This commit removes `Module.Fn` and replaces it with two new
`InternPool.Tag` values:

 * `func_decl` - corresponding to a function declared in the source
   code. This one contains line/column numbers, zir_body_inst, etc.

 * `func_instance` - one for each monomorphization of a generic
   function. Contains a reference to the `func_decl` from whence the
   instantiation came, along with the `comptime` parameter values (or
   types in the case of `anytype`)

Since `InternPool` provides deduplication on these values, these fields
are now deleted from `Module`:

 * `monomorphed_func_keys`
 * `monomorphed_funcs`
 * `align_stack_fns`

Instead of these, Sema logic for generic function instantiation now
unconditionally evaluates the function prototype expression for every
generic callsite. This is technically required in order for type
coercions to work. The previous code had some dubious, probably wrong
hacks to make things work, such as `hashUncoerced`. I'm not 100% sure
how we were able to eliminate that function and still pass all the
behavior tests, but I'm pretty sure things were still broken without
doing type coercion for every generic function call argument.

After the function prototype is evaluated, it produces a deduplicated
`func_instance` `InternPool.Index` which can then be used for the
generic function call.

Some other nice things made by this simplification are the removal of
`comptime_args_fn_inst` and `preallocated_new_func` from `Sema`, and the
messy logic associated with them.

I have not yet been able to measure the perf of this against master
branch. On one hand, it reduces memory usage and pointer chasing of the
most heavily used `InternPool` Tag - function bodies - but on the other
hand, it does evaluate function prototype expressions more than before.
We will soon find out.
2023-07-18 19:02:05 -07:00

633 lines
21 KiB
Zig

const std = @import("std");
const mem = std.mem;
const assert = std.debug.assert;
const Allocator = std.mem.Allocator;
const fs = std.fs;
const C = @This();
const Module = @import("../Module.zig");
const InternPool = @import("../InternPool.zig");
const Compilation = @import("../Compilation.zig");
const codegen = @import("../codegen/c.zig");
const link = @import("../link.zig");
const trace = @import("../tracy.zig").trace;
const Type = @import("../type.zig").Type;
const Air = @import("../Air.zig");
const Liveness = @import("../Liveness.zig");
pub const base_tag: link.File.Tag = .c;
pub const zig_h = "#include \"zig.h\"\n";
base: link.File,
/// This linker backend does not try to incrementally link output C source code.
/// Instead, it tracks all declarations in this table, and iterates over it
/// in the flush function, stitching pre-rendered pieces of C code together.
decl_table: std.AutoArrayHashMapUnmanaged(Module.Decl.Index, DeclBlock) = .{},
/// Per-declaration data.
const DeclBlock = struct {
code: std.ArrayListUnmanaged(u8) = .{},
fwd_decl: std.ArrayListUnmanaged(u8) = .{},
/// Each `Decl` stores a set of used `CType`s. In `flush()`, we iterate
/// over each `Decl` and generate the definition for each used `CType` once.
ctypes: codegen.CType.Store = .{},
/// Key and Value storage use the ctype arena.
lazy_fns: codegen.LazyFnMap = .{},
fn deinit(db: *DeclBlock, gpa: Allocator) void {
db.lazy_fns.deinit(gpa);
db.ctypes.deinit(gpa);
db.fwd_decl.deinit(gpa);
db.code.deinit(gpa);
db.* = undefined;
}
};
pub fn openPath(gpa: Allocator, sub_path: []const u8, options: link.Options) !*C {
assert(options.target.ofmt == .c);
if (options.use_llvm) return error.LLVMHasNoCBackend;
if (options.use_lld) return error.LLDHasNoCBackend;
const file = try options.emit.?.directory.handle.createFile(sub_path, .{
// Truncation is done on `flush`.
.truncate = false,
.mode = link.determineMode(options),
});
errdefer file.close();
var c_file = try gpa.create(C);
errdefer gpa.destroy(c_file);
c_file.* = C{
.base = .{
.tag = .c,
.options = options,
.file = file,
.allocator = gpa,
},
};
return c_file;
}
pub fn deinit(self: *C) void {
const gpa = self.base.allocator;
for (self.decl_table.values()) |*db| {
db.deinit(gpa);
}
self.decl_table.deinit(gpa);
}
pub fn freeDecl(self: *C, decl_index: Module.Decl.Index) void {
const gpa = self.base.allocator;
if (self.decl_table.fetchSwapRemove(decl_index)) |kv| {
var decl_block = kv.value;
decl_block.deinit(gpa);
}
}
pub fn updateFunc(self: *C, module: *Module, func_index: InternPool.Index, air: Air, liveness: Liveness) !void {
const tracy = trace(@src());
defer tracy.end();
const gpa = self.base.allocator;
const func = module.funcInfo(func_index);
const decl_index = func.owner_decl;
const gop = try self.decl_table.getOrPut(gpa, decl_index);
if (!gop.found_existing) {
gop.value_ptr.* = .{};
}
const ctypes = &gop.value_ptr.ctypes;
const lazy_fns = &gop.value_ptr.lazy_fns;
const fwd_decl = &gop.value_ptr.fwd_decl;
const code = &gop.value_ptr.code;
ctypes.clearRetainingCapacity(gpa);
lazy_fns.clearRetainingCapacity();
fwd_decl.shrinkRetainingCapacity(0);
code.shrinkRetainingCapacity(0);
var function: codegen.Function = .{
.value_map = codegen.CValueMap.init(gpa),
.air = air,
.liveness = liveness,
.func_index = func_index,
.object = .{
.dg = .{
.gpa = gpa,
.module = module,
.error_msg = null,
.decl_index = decl_index.toOptional(),
.decl = module.declPtr(decl_index),
.fwd_decl = fwd_decl.toManaged(gpa),
.ctypes = ctypes.*,
},
.code = code.toManaged(gpa),
.indent_writer = undefined, // set later so we can get a pointer to object.code
},
.lazy_fns = lazy_fns.*,
};
function.object.indent_writer = .{ .underlying_writer = function.object.code.writer() };
defer function.deinit();
codegen.genFunc(&function) catch |err| switch (err) {
error.AnalysisFail => {
try module.failed_decls.put(gpa, decl_index, function.object.dg.error_msg.?);
return;
},
else => |e| return e,
};
ctypes.* = function.object.dg.ctypes.move();
lazy_fns.* = function.lazy_fns.move();
fwd_decl.* = function.object.dg.fwd_decl.moveToUnmanaged();
code.* = function.object.code.moveToUnmanaged();
// Free excess allocated memory for this Decl.
ctypes.shrinkAndFree(gpa, ctypes.count());
lazy_fns.shrinkAndFree(gpa, lazy_fns.count());
fwd_decl.shrinkAndFree(gpa, fwd_decl.items.len);
code.shrinkAndFree(gpa, code.items.len);
}
pub fn updateDecl(self: *C, module: *Module, decl_index: Module.Decl.Index) !void {
const tracy = trace(@src());
defer tracy.end();
const gpa = self.base.allocator;
const gop = try self.decl_table.getOrPut(gpa, decl_index);
if (!gop.found_existing) {
gop.value_ptr.* = .{};
}
const ctypes = &gop.value_ptr.ctypes;
const fwd_decl = &gop.value_ptr.fwd_decl;
const code = &gop.value_ptr.code;
ctypes.clearRetainingCapacity(gpa);
fwd_decl.shrinkRetainingCapacity(0);
code.shrinkRetainingCapacity(0);
const decl = module.declPtr(decl_index);
var object: codegen.Object = .{
.dg = .{
.gpa = gpa,
.module = module,
.error_msg = null,
.decl_index = decl_index.toOptional(),
.decl = decl,
.fwd_decl = fwd_decl.toManaged(gpa),
.ctypes = ctypes.*,
},
.code = code.toManaged(gpa),
.indent_writer = undefined, // set later so we can get a pointer to object.code
};
object.indent_writer = .{ .underlying_writer = object.code.writer() };
defer {
object.code.deinit();
object.dg.ctypes.deinit(object.dg.gpa);
object.dg.fwd_decl.deinit();
}
codegen.genDecl(&object) catch |err| switch (err) {
error.AnalysisFail => {
try module.failed_decls.put(gpa, decl_index, object.dg.error_msg.?);
return;
},
else => |e| return e,
};
ctypes.* = object.dg.ctypes.move();
fwd_decl.* = object.dg.fwd_decl.moveToUnmanaged();
code.* = object.code.moveToUnmanaged();
// Free excess allocated memory for this Decl.
ctypes.shrinkAndFree(gpa, ctypes.count());
fwd_decl.shrinkAndFree(gpa, fwd_decl.items.len);
code.shrinkAndFree(gpa, code.items.len);
}
pub fn updateDeclLineNumber(self: *C, module: *Module, decl_index: Module.Decl.Index) !void {
// The C backend does not have the ability to fix line numbers without re-generating
// the entire Decl.
_ = self;
_ = module;
_ = decl_index;
}
pub fn flush(self: *C, comp: *Compilation, prog_node: *std.Progress.Node) !void {
return self.flushModule(comp, prog_node);
}
fn abiDefines(self: *C, target: std.Target) !std.ArrayList(u8) {
var defines = std.ArrayList(u8).init(self.base.allocator);
errdefer defines.deinit();
const writer = defines.writer();
switch (target.abi) {
.msvc => try writer.writeAll("#define ZIG_TARGET_ABI_MSVC\n"),
else => {},
}
try writer.print("#define ZIG_TARGET_MAX_INT_ALIGNMENT {d}\n", .{target.maxIntAlignment()});
return defines;
}
pub fn flushModule(self: *C, _: *Compilation, prog_node: *std.Progress.Node) !void {
const tracy = trace(@src());
defer tracy.end();
var sub_prog_node = prog_node.start("Flush Module", 0);
sub_prog_node.activate();
defer sub_prog_node.end();
const gpa = self.base.allocator;
const module = self.base.options.module.?;
// This code path happens exclusively with -ofmt=c. The flush logic for
// emit-h is in `flushEmitH` below.
var f: Flush = .{};
defer f.deinit(gpa);
const abi_defines = try self.abiDefines(module.getTarget());
defer abi_defines.deinit();
// Covers defines, zig.h, ctypes, asm, lazy fwd.
try f.all_buffers.ensureUnusedCapacity(gpa, 5);
f.appendBufAssumeCapacity(abi_defines.items);
f.appendBufAssumeCapacity(zig_h);
const ctypes_index = f.all_buffers.items.len;
f.all_buffers.items.len += 1;
{
var asm_buf = f.asm_buf.toManaged(gpa);
defer f.asm_buf = asm_buf.moveToUnmanaged();
try codegen.genGlobalAsm(module, asm_buf.writer());
f.appendBufAssumeCapacity(asm_buf.items);
}
const lazy_index = f.all_buffers.items.len;
f.all_buffers.items.len += 1;
try self.flushErrDecls(&f.lazy_db);
// `CType`s, forward decls, and non-functions first.
// Unlike other backends, the .c code we are emitting is order-dependent. Therefore
// we must traverse the set of Decls that we are emitting according to their dependencies.
// Our strategy is to populate a set of remaining decls, pop Decls one by one,
// recursively chasing their dependencies.
try f.remaining_decls.ensureUnusedCapacity(gpa, self.decl_table.count());
const decl_keys = self.decl_table.keys();
const decl_values = self.decl_table.values();
for (decl_keys) |decl_index| {
assert(module.declPtr(decl_index).has_tv);
f.remaining_decls.putAssumeCapacityNoClobber(decl_index, {});
}
{
var export_names: std.AutoHashMapUnmanaged(InternPool.NullTerminatedString, void) = .{};
defer export_names.deinit(gpa);
try export_names.ensureTotalCapacity(gpa, @as(u32, @intCast(module.decl_exports.entries.len)));
for (module.decl_exports.values()) |exports| for (exports.items) |@"export"|
try export_names.put(gpa, @"export".opts.name, {});
while (f.remaining_decls.popOrNull()) |kv| {
const decl_index = kv.key;
try self.flushDecl(&f, decl_index, export_names);
}
}
{
// We need to flush lazy ctypes after flushing all decls but before flushing any decl ctypes.
// This ensures that every lazy CType.Index exactly matches the global CType.Index.
assert(f.ctypes.count() == 0);
try self.flushCTypes(&f, .none, f.lazy_db.ctypes);
var it = self.decl_table.iterator();
while (it.next()) |entry|
try self.flushCTypes(&f, entry.key_ptr.toOptional(), entry.value_ptr.ctypes);
}
f.all_buffers.items[ctypes_index] = .{
.iov_base = if (f.ctypes_buf.items.len > 0) f.ctypes_buf.items.ptr else "",
.iov_len = f.ctypes_buf.items.len,
};
f.file_size += f.ctypes_buf.items.len;
f.all_buffers.items[lazy_index] = .{
.iov_base = if (f.lazy_db.fwd_decl.items.len > 0) f.lazy_db.fwd_decl.items.ptr else "",
.iov_len = f.lazy_db.fwd_decl.items.len,
};
f.file_size += f.lazy_db.fwd_decl.items.len;
// Now the code.
try f.all_buffers.ensureUnusedCapacity(gpa, 1 + decl_values.len);
f.appendBufAssumeCapacity(f.lazy_db.code.items);
for (decl_values) |decl| f.appendBufAssumeCapacity(decl.code.items);
const file = self.base.file.?;
try file.setEndPos(f.file_size);
try file.pwritevAll(f.all_buffers.items, 0);
}
const Flush = struct {
remaining_decls: std.AutoArrayHashMapUnmanaged(Module.Decl.Index, void) = .{},
ctypes: codegen.CType.Store = .{},
ctypes_map: std.ArrayListUnmanaged(codegen.CType.Index) = .{},
ctypes_buf: std.ArrayListUnmanaged(u8) = .{},
lazy_db: DeclBlock = .{},
lazy_fns: LazyFns = .{},
asm_buf: std.ArrayListUnmanaged(u8) = .{},
/// We collect a list of buffers to write, and write them all at once with pwritev 😎
all_buffers: std.ArrayListUnmanaged(std.os.iovec_const) = .{},
/// Keeps track of the total bytes of `all_buffers`.
file_size: u64 = 0,
const LazyFns = std.AutoHashMapUnmanaged(codegen.LazyFnKey, void);
fn appendBufAssumeCapacity(f: *Flush, buf: []const u8) void {
if (buf.len == 0) return;
f.all_buffers.appendAssumeCapacity(.{ .iov_base = buf.ptr, .iov_len = buf.len });
f.file_size += buf.len;
}
fn deinit(f: *Flush, gpa: Allocator) void {
f.all_buffers.deinit(gpa);
f.asm_buf.deinit(gpa);
f.lazy_fns.deinit(gpa);
f.lazy_db.deinit(gpa);
f.ctypes_buf.deinit(gpa);
f.ctypes_map.deinit(gpa);
f.ctypes.deinit(gpa);
f.remaining_decls.deinit(gpa);
}
};
const FlushDeclError = error{
OutOfMemory,
};
fn flushCTypes(
self: *C,
f: *Flush,
decl_index: Module.Decl.OptionalIndex,
decl_ctypes: codegen.CType.Store,
) FlushDeclError!void {
const gpa = self.base.allocator;
const mod = self.base.options.module.?;
const decl_ctypes_len = decl_ctypes.count();
f.ctypes_map.clearRetainingCapacity();
try f.ctypes_map.ensureTotalCapacity(gpa, decl_ctypes_len);
var global_ctypes = f.ctypes.promote(gpa);
defer f.ctypes.demote(global_ctypes);
var ctypes_buf = f.ctypes_buf.toManaged(gpa);
defer f.ctypes_buf = ctypes_buf.moveToUnmanaged();
const writer = ctypes_buf.writer();
const slice = decl_ctypes.set.map.entries.slice();
for (slice.items(.key), 0..) |decl_cty, decl_i| {
const Context = struct {
arena: Allocator,
ctypes_map: []codegen.CType.Index,
cached_hash: codegen.CType.Store.Set.Map.Hash,
idx: codegen.CType.Index,
pub fn hash(ctx: @This(), _: codegen.CType) codegen.CType.Store.Set.Map.Hash {
return ctx.cached_hash;
}
pub fn eql(ctx: @This(), lhs: codegen.CType, rhs: codegen.CType, _: usize) bool {
return lhs.eqlContext(rhs, ctx);
}
pub fn eqlIndex(
ctx: @This(),
lhs_idx: codegen.CType.Index,
rhs_idx: codegen.CType.Index,
) bool {
if (lhs_idx < codegen.CType.Tag.no_payload_count or
rhs_idx < codegen.CType.Tag.no_payload_count) return lhs_idx == rhs_idx;
const lhs_i = lhs_idx - codegen.CType.Tag.no_payload_count;
if (lhs_i >= ctx.ctypes_map.len) return false;
return ctx.ctypes_map[lhs_i] == rhs_idx;
}
pub fn copyIndex(ctx: @This(), idx: codegen.CType.Index) codegen.CType.Index {
if (idx < codegen.CType.Tag.no_payload_count) return idx;
return ctx.ctypes_map[idx - codegen.CType.Tag.no_payload_count];
}
};
const decl_idx = @as(codegen.CType.Index, @intCast(codegen.CType.Tag.no_payload_count + decl_i));
const ctx = Context{
.arena = global_ctypes.arena.allocator(),
.ctypes_map = f.ctypes_map.items,
.cached_hash = decl_ctypes.indexToHash(decl_idx),
.idx = decl_idx,
};
const gop = try global_ctypes.set.map.getOrPutContextAdapted(gpa, decl_cty, ctx, .{
.store = &global_ctypes.set,
});
const global_idx =
@as(codegen.CType.Index, @intCast(codegen.CType.Tag.no_payload_count + gop.index));
f.ctypes_map.appendAssumeCapacity(global_idx);
if (!gop.found_existing) {
errdefer _ = global_ctypes.set.map.pop();
gop.key_ptr.* = try decl_cty.copyContext(ctx);
}
if (std.debug.runtime_safety) {
const global_cty = &global_ctypes.set.map.entries.items(.key)[gop.index];
assert(global_cty == gop.key_ptr);
assert(decl_cty.eqlContext(global_cty.*, ctx));
assert(decl_cty.hash(decl_ctypes.set) == global_cty.hash(global_ctypes.set));
}
try codegen.genTypeDecl(
mod,
writer,
global_ctypes.set,
global_idx,
decl_index,
decl_ctypes.set,
decl_idx,
gop.found_existing,
);
}
}
fn flushErrDecls(self: *C, db: *DeclBlock) FlushDeclError!void {
const gpa = self.base.allocator;
const fwd_decl = &db.fwd_decl;
const ctypes = &db.ctypes;
const code = &db.code;
var object = codegen.Object{
.dg = .{
.gpa = gpa,
.module = self.base.options.module.?,
.error_msg = null,
.decl_index = .none,
.decl = null,
.fwd_decl = fwd_decl.toManaged(gpa),
.ctypes = ctypes.*,
},
.code = code.toManaged(gpa),
.indent_writer = undefined, // set later so we can get a pointer to object.code
};
object.indent_writer = .{ .underlying_writer = object.code.writer() };
defer {
object.code.deinit();
object.dg.ctypes.deinit(gpa);
object.dg.fwd_decl.deinit();
}
codegen.genErrDecls(&object) catch |err| switch (err) {
error.AnalysisFail => unreachable,
else => |e| return e,
};
fwd_decl.* = object.dg.fwd_decl.moveToUnmanaged();
ctypes.* = object.dg.ctypes.move();
code.* = object.code.moveToUnmanaged();
}
fn flushLazyFn(self: *C, db: *DeclBlock, lazy_fn: codegen.LazyFnMap.Entry) FlushDeclError!void {
const gpa = self.base.allocator;
const fwd_decl = &db.fwd_decl;
const ctypes = &db.ctypes;
const code = &db.code;
var object = codegen.Object{
.dg = .{
.gpa = gpa,
.module = self.base.options.module.?,
.error_msg = null,
.decl_index = .none,
.decl = null,
.fwd_decl = fwd_decl.toManaged(gpa),
.ctypes = ctypes.*,
},
.code = code.toManaged(gpa),
.indent_writer = undefined, // set later so we can get a pointer to object.code
};
object.indent_writer = .{ .underlying_writer = object.code.writer() };
defer {
object.code.deinit();
object.dg.ctypes.deinit(gpa);
object.dg.fwd_decl.deinit();
}
codegen.genLazyFn(&object, lazy_fn) catch |err| switch (err) {
error.AnalysisFail => unreachable,
else => |e| return e,
};
fwd_decl.* = object.dg.fwd_decl.moveToUnmanaged();
ctypes.* = object.dg.ctypes.move();
code.* = object.code.moveToUnmanaged();
}
fn flushLazyFns(self: *C, f: *Flush, lazy_fns: codegen.LazyFnMap) FlushDeclError!void {
const gpa = self.base.allocator;
try f.lazy_fns.ensureUnusedCapacity(gpa, @as(Flush.LazyFns.Size, @intCast(lazy_fns.count())));
var it = lazy_fns.iterator();
while (it.next()) |entry| {
const gop = f.lazy_fns.getOrPutAssumeCapacity(entry.key_ptr.*);
if (gop.found_existing) continue;
gop.value_ptr.* = {};
try self.flushLazyFn(&f.lazy_db, entry);
}
}
/// Assumes `decl` was in the `remaining_decls` set, and has already been removed.
fn flushDecl(
self: *C,
f: *Flush,
decl_index: Module.Decl.Index,
export_names: std.AutoHashMapUnmanaged(InternPool.NullTerminatedString, void),
) FlushDeclError!void {
const gpa = self.base.allocator;
const mod = self.base.options.module.?;
const decl = mod.declPtr(decl_index);
// Before flushing any particular Decl we must ensure its
// dependencies are already flushed, so that the order in the .c
// file comes out correctly.
for (decl.dependencies.keys()) |dep| {
if (f.remaining_decls.swapRemove(dep)) {
try flushDecl(self, f, dep, export_names);
}
}
const decl_block = self.decl_table.getPtr(decl_index).?;
try self.flushLazyFns(f, decl_block.lazy_fns);
try f.all_buffers.ensureUnusedCapacity(gpa, 1);
if (!(decl.isExtern(mod) and export_names.contains(decl.name)))
f.appendBufAssumeCapacity(decl_block.fwd_decl.items);
}
pub fn flushEmitH(module: *Module) !void {
const tracy = trace(@src());
defer tracy.end();
const emit_h = module.emit_h orelse return;
// We collect a list of buffers to write, and write them all at once with pwritev 😎
const num_buffers = emit_h.decl_table.count() + 1;
var all_buffers = try std.ArrayList(std.os.iovec_const).initCapacity(module.gpa, num_buffers);
defer all_buffers.deinit();
var file_size: u64 = zig_h.len;
if (zig_h.len != 0) {
all_buffers.appendAssumeCapacity(.{
.iov_base = zig_h,
.iov_len = zig_h.len,
});
}
for (emit_h.decl_table.keys()) |decl_index| {
const decl_emit_h = emit_h.declPtr(decl_index);
const buf = decl_emit_h.fwd_decl.items;
if (buf.len != 0) {
all_buffers.appendAssumeCapacity(.{
.iov_base = buf.ptr,
.iov_len = buf.len,
});
file_size += buf.len;
}
}
const directory = emit_h.loc.directory orelse module.comp.local_cache_directory;
const file = try directory.handle.createFile(emit_h.loc.basename, .{
// We set the end position explicitly below; by not truncating the file, we possibly
// make it easier on the file system by doing 1 reallocation instead of two.
.truncate = false,
});
defer file.close();
try file.setEndPos(file_size);
try file.pwritevAll(all_buffers.items, 0);
}
pub fn updateDeclExports(
self: *C,
module: *Module,
decl_index: Module.Decl.Index,
exports: []const *Module.Export,
) !void {
_ = exports;
_ = decl_index;
_ = module;
_ = self;
}