AstGen: back to index-based for loops

This commit is contained in:
Andrew Kelley 2023-02-17 16:39:45 -07:00
parent 841add6890
commit 293d6bdc73
4 changed files with 106 additions and 143 deletions

View File

@ -88,7 +88,6 @@ fn setExtra(astgen: *AstGen, index: usize, extra: anytype) void {
Zir.Inst.BuiltinCall.Flags => @bitCast(u32, @field(extra, field.name)),
Zir.Inst.SwitchBlock.Bits => @bitCast(u32, @field(extra, field.name)),
Zir.Inst.FuncFancy.Bits => @bitCast(u32, @field(extra, field.name)),
Zir.Inst.ElemPtrImm.Bits => @bitCast(u32, @field(extra, field.name)),
else => @compileError("bad field type"),
};
i += 1;
@ -1566,9 +1565,7 @@ fn arrayInitExprRlPtrInner(
for (elements) |elem_init, i| {
const elem_ptr = try gz.addPlNode(.elem_ptr_imm, elem_init, Zir.Inst.ElemPtrImm{
.ptr = result_ptr,
.bits = .{
.index = @intCast(u31, i),
},
.index = @intCast(u32, i),
});
astgen.extra.items[extra_index] = refToIndex(elem_ptr).?;
extra_index += 1;
@ -2601,6 +2598,7 @@ fn addEnsureResult(gz: *GenZir, maybe_unused_result: Zir.Inst.Ref, statement: As
.field_base_ptr,
.ret_ptr,
.ret_type,
.for_len,
.@"try",
.try_ptr,
//.try_inline,
@ -2669,7 +2667,6 @@ fn addEnsureResult(gz: *GenZir, maybe_unused_result: Zir.Inst.Ref, statement: As
.validate_deref,
.save_err_ret_index,
.restore_err_ret_index,
.for_check_lens,
=> break :b true,
.@"defer" => unreachable,
@ -6305,23 +6302,26 @@ fn forExpr(
const node_data = tree.nodes.items(.data);
const gpa = astgen.gpa;
const allocs = try gpa.alloc(Zir.Inst.Ref, for_full.ast.inputs.len);
defer gpa.free(allocs);
// For counters, this is the start value; for indexables, this is the base
// pointer that can be used with elem_ptr and similar instructions.
// Special value `none` means that this is a counter and its start value is
// zero, indicating that the main index counter can be used directly.
const indexables = try gpa.alloc(Zir.Inst.Ref, for_full.ast.inputs.len);
defer gpa.free(indexables);
// elements of this array can be `none`, indicating no length check.
const lens = try gpa.alloc(Zir.Inst.Ref, for_full.ast.inputs.len);
defer gpa.free(lens);
const alloc_tag: Zir.Inst.Tag = if (is_inline) .alloc_comptime_mut else .alloc_mut;
// We will use a single zero-based counter no matter how many indexables there are.
const index_ptr = blk: {
const alloc_tag: Zir.Inst.Tag = if (is_inline) .alloc_comptime_mut else .alloc;
const index_ptr = try parent_gz.addUnNode(alloc_tag, .usize_type, node);
// initialize to zero
_ = try parent_gz.addBin(.store, index_ptr, .zero_usize);
break :blk index_ptr;
};
// Tracks the index of allocs/lens that has a length to be checked and is
// used for the end value.
// If this is null, there are no len checks.
var end_input_index: ?u32 = null;
// This is a value to use to find out if the for loop has reached the end
// yet. It prefers to use a counter since the end value is provided directly,
// and otherwise falls back to adding ptr+len of a slice to compute end.
// Corresponds to end_input_index and will be .none in case that value is null.
var cond_end_val: Zir.Inst.Ref = .none;
var any_len_checks = false;
{
var capture_token = for_full.payload_token;
@ -6341,10 +6341,8 @@ fn forExpr(
if (capture_is_ref) {
return astgen.failTok(ident_tok, "cannot capture reference to range", .{});
}
const counter_ptr = try parent_gz.addUnNode(alloc_tag, .usize_type, node);
const start_node = node_data[input].lhs;
const start_val = try expr(parent_gz, scope, .{ .rl = .none }, start_node);
_ = try parent_gz.addBin(.store, counter_ptr, start_val);
const end_node = node_data[input].rhs;
const end_val = if (end_node != 0)
@ -6352,7 +6350,8 @@ fn forExpr(
else
.none;
const range_len = if (end_val == .none or nodeIsTriviallyZero(tree, start_node))
const start_is_zero = nodeIsTriviallyZero(tree, start_node);
const range_len = if (end_val == .none or start_is_zero)
end_val
else
try parent_gz.addPlNode(.sub, input, Zir.Inst.Bin{
@ -6360,61 +6359,33 @@ fn forExpr(
.rhs = start_val,
});
if (range_len != .none and cond_end_val == .none) {
end_input_index = i;
cond_end_val = end_val;
}
allocs[i] = counter_ptr;
any_len_checks = any_len_checks or range_len != .none;
indexables[i] = if (start_is_zero) .none else start_val;
lens[i] = range_len;
} else {
const indexable = try expr(parent_gz, scope, .{ .rl = .none }, input);
// This instruction has nice compile errors so we put it before the other ones
// even though it is not needed until later in the block.
const ptr_len = try parent_gz.addUnNode(.indexable_ptr_len, indexable, input);
const base_ptr = try parent_gz.addPlNode(.elem_ptr_imm, input, Zir.Inst.ElemPtrImm{
.ptr = indexable,
.bits = .{
.index = 0,
.manyptr = true,
},
});
const alloc_ty_inst = try parent_gz.addUnNode(.typeof, base_ptr, node);
const alloc = try parent_gz.addUnNode(alloc_tag, alloc_ty_inst, node);
_ = try parent_gz.addBin(.store, alloc, base_ptr);
const indexable_len = try parent_gz.addUnNode(.indexable_ptr_len, indexable, input);
if (end_input_index == null) {
end_input_index = i;
assert(cond_end_val == .none);
}
allocs[i] = alloc;
lens[i] = ptr_len;
any_len_checks = true;
indexables[i] = indexable;
lens[i] = indexable_len;
}
}
}
// In case there are no counters which already have an end computed, we
// compute an end from base pointer plus length.
if (end_input_index) |i| {
if (cond_end_val == .none) {
cond_end_val = try parent_gz.addPlNode(.add, for_full.ast.inputs[i], Zir.Inst.Bin{
.lhs = allocs[i],
.rhs = lens[i],
});
}
}
// We use a dedicated ZIR instruction to assert the lengths to assist with
// nicer error reporting as well as fewer ZIR bytes emitted.
if (end_input_index != null) {
const len: Zir.Inst.Ref = len: {
if (!any_len_checks) break :len .none;
const lens_len = @intCast(u32, lens.len);
try astgen.extra.ensureUnusedCapacity(gpa, @typeInfo(Zir.Inst.MultiOp).Struct.fields.len + lens_len);
_ = try parent_gz.addPlNode(.for_check_lens, node, Zir.Inst.MultiOp{
const len = try parent_gz.addPlNode(.for_len, node, Zir.Inst.MultiOp{
.operands_len = lens_len,
});
appendRefsAssumeCapacity(astgen, lens);
}
break :len len;
};
const loop_tag: Zir.Inst.Tag = if (is_inline) .block_inline else .loop;
const loop_block = try parent_gz.makeBlockInst(loop_tag, node);
@ -6429,22 +6400,14 @@ fn forExpr(
var cond_scope = parent_gz.makeSubBlock(&loop_scope.base);
defer cond_scope.unstack();
// Load all the iterables.
const loaded_ptrs = try gpa.alloc(Zir.Inst.Ref, allocs.len);
defer gpa.free(loaded_ptrs);
for (allocs) |alloc, i| {
loaded_ptrs[i] = try cond_scope.addUnNode(.load, alloc, for_full.ast.inputs[i]);
}
// Check the condition.
const input_index = end_input_index orelse {
if (!any_len_checks) {
return astgen.failNode(node, "TODO: handle infinite for loop", .{});
};
assert(cond_end_val != .none);
const cond = try cond_scope.addPlNode(.cmp_neq, for_full.ast.inputs[input_index], Zir.Inst.Bin{
.lhs = loaded_ptrs[input_index],
.rhs = cond_end_val,
}
const index = try cond_scope.addUnNode(.load, index_ptr, node);
const cond = try cond_scope.addPlNode(.cmp_lt, node, Zir.Inst.Bin{
.lhs = index,
.rhs = len,
});
const condbr_tag: Zir.Inst.Tag = if (is_inline) .condbr_inline else .condbr;
@ -6455,14 +6418,12 @@ fn forExpr(
// cond_block unstacked now, can add new instructions to loop_scope
try loop_scope.instructions.append(gpa, cond_block);
// Increment the loop variables.
for (allocs) |alloc, i| {
const incremented = try loop_scope.addPlNode(.add, node, Zir.Inst.Bin{
.lhs = loaded_ptrs[i],
.rhs = .one_usize,
});
_ = try loop_scope.addBin(.store, alloc, incremented);
}
// Increment the index variable.
const index_plus_one = try loop_scope.addPlNode(.add, node, Zir.Inst.Bin{
.lhs = index,
.rhs = .one_usize,
});
_ = try loop_scope.addBin(.store, index_ptr, index_plus_one);
const repeat_tag: Zir.Inst.Tag = if (is_inline) .repeat_inline else .repeat;
_ = try loop_scope.addNode(repeat_tag, node);
@ -6500,21 +6461,43 @@ fn forExpr(
const name_str_index = try astgen.identAsString(ident_tok);
try astgen.detectLocalShadowing(capture_sub_scope, name_str_index, ident_tok, capture_name, .capture);
const loaded = if (capture_is_ref)
loaded_ptrs[i]
else
try then_scope.addUnNode(.load, loaded_ptrs[i], input);
const capture_inst = inst: {
const is_counter = node_tags[input] == .for_range;
if (indexables[i] == .none) {
// Special case: the main index can be used directly.
assert(is_counter);
assert(!capture_is_ref);
break :inst index;
}
// For counters, we add the index variable to the start value; for
// indexables, we use it as an element index. This is so similar
// that they can share the same code paths, branching only on the
// ZIR tag.
const switch_cond = (@as(u2, @boolToInt(capture_is_ref)) << 1) | @boolToInt(is_counter);
const tag: Zir.Inst.Tag = switch (switch_cond) {
0b00 => .elem_val,
0b01 => .add,
0b10 => .elem_ptr,
0b11 => unreachable, // compile error emitted already
};
break :inst try then_scope.addPlNode(tag, input, Zir.Inst.Bin{
.lhs = indexables[i],
.rhs = index,
});
};
capture_scopes[i] = .{
.parent = capture_sub_scope,
.gen_zir = &then_scope,
.name = name_str_index,
.inst = loaded,
.inst = capture_inst,
.token_src = ident_tok,
.id_cat = .capture,
};
try then_scope.addDbgVar(.dbg_var_val, name_str_index, loaded);
try then_scope.addDbgVar(.dbg_var_val, name_str_index, capture_inst);
capture_sub_scope = &capture_scopes[i].base;
}

View File

@ -1035,6 +1035,7 @@ fn analyzeBodyInner(
.@"await" => try sema.zirAwait(block, inst),
.array_base_ptr => try sema.zirArrayBasePtr(block, inst),
.field_base_ptr => try sema.zirFieldBasePtr(block, inst),
.for_len => try sema.zirForLen(block, inst),
.clz => try sema.zirBitCount(block, inst, .clz, Value.clz),
.ctz => try sema.zirBitCount(block, inst, .ctz, Value.ctz),
@ -1386,11 +1387,6 @@ fn analyzeBodyInner(
i += 1;
continue;
},
.for_check_lens => {
try sema.zirForCheckLens(block, inst);
i += 1;
continue;
},
// Special case instructions to handle comptime control flow.
.@"break" => {
@ -3924,6 +3920,16 @@ fn zirFieldBasePtr(
return sema.failWithStructInitNotSupported(block, src, sema.typeOf(start_ptr).childType());
}
fn zirForLen(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError!Air.Inst.Ref {
const inst_data = sema.code.instructions.items(.data)[inst].pl_node;
const extra = sema.code.extraData(Zir.Inst.MultiOp, inst_data.payload_index);
const args = sema.code.refSlice(extra.end, extra.data.operands_len);
const src = inst_data.src();
_ = args;
return sema.fail(block, src, "TODO implement zirForCheckLens", .{});
}
fn validateArrayInitTy(
sema: *Sema,
block: *Block,
@ -9649,7 +9655,7 @@ fn zirElemPtr(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError!Air
const extra = sema.code.extraData(Zir.Inst.Bin, inst_data.payload_index).data;
const array_ptr = try sema.resolveInst(extra.lhs);
const elem_index = try sema.resolveInst(extra.rhs);
return sema.elemPtr(block, src, array_ptr, elem_index, src, false, .One);
return sema.elemPtr(block, src, array_ptr, elem_index, src, false);
}
fn zirElemPtrNode(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError!Air.Inst.Ref {
@ -9662,7 +9668,7 @@ fn zirElemPtrNode(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError
const extra = sema.code.extraData(Zir.Inst.Bin, inst_data.payload_index).data;
const array_ptr = try sema.resolveInst(extra.lhs);
const elem_index = try sema.resolveInst(extra.rhs);
return sema.elemPtr(block, src, array_ptr, elem_index, elem_index_src, false, .One);
return sema.elemPtr(block, src, array_ptr, elem_index, elem_index_src, false);
}
fn zirElemPtrImm(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError!Air.Inst.Ref {
@ -9673,9 +9679,8 @@ fn zirElemPtrImm(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError!
const src = inst_data.src();
const extra = sema.code.extraData(Zir.Inst.ElemPtrImm, inst_data.payload_index).data;
const array_ptr = try sema.resolveInst(extra.ptr);
const elem_index = try sema.addIntUnsigned(Type.usize, extra.bits.index);
const size: std.builtin.Type.Pointer.Size = if (extra.bits.manyptr) .Many else .One;
return sema.elemPtr(block, src, array_ptr, elem_index, src, true, size);
const elem_index = try sema.addIntUnsigned(Type.usize, extra.index);
return sema.elemPtr(block, src, array_ptr, elem_index, src, true);
}
fn zirSliceStart(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError!Air.Inst.Ref {
@ -17102,16 +17107,6 @@ fn zirRestoreErrRetIndex(sema: *Sema, start_block: *Block, inst: Zir.Inst.Index)
return sema.popErrorReturnTrace(start_block, src, operand, saved_index);
}
fn zirForCheckLens(sema: *Sema, block: *Block, inst: Zir.Inst.Index) CompileError!void {
const inst_data = sema.code.instructions.items(.data)[inst].pl_node;
const extra = sema.code.extraData(Zir.Inst.MultiOp, inst_data.payload_index);
const args = sema.code.refSlice(extra.end, extra.data.operands_len);
const src = inst_data.src();
_ = args;
return sema.fail(block, src, "TODO implement zirForCheckLens", .{});
}
fn addToInferredErrorSet(sema: *Sema, uncasted_operand: Air.Inst.Ref) !void {
assert(sema.fn_ret_ty.zigTypeTag() == .ErrorUnion);
@ -22906,7 +22901,7 @@ fn panicSentinelMismatch(
const actual_sentinel = if (ptr_ty.isSlice())
try parent_block.addBinOp(.slice_elem_val, ptr, sentinel_index)
else blk: {
const elem_ptr_ty = try sema.elemPtrType(ptr_ty, null, .One);
const elem_ptr_ty = try sema.elemPtrType(ptr_ty, null);
const sentinel_ptr = try parent_block.addPtrElemPtr(ptr, sentinel_index, elem_ptr_ty);
break :blk try parent_block.addTyOp(.load, sentinel_ty, sentinel_ptr);
};
@ -24073,7 +24068,6 @@ fn elemPtr(
elem_index: Air.Inst.Ref,
elem_index_src: LazySrcLoc,
init: bool,
size: std.builtin.Type.Pointer.Size,
) CompileError!Air.Inst.Ref {
const indexable_ptr_src = src; // TODO better source location
const indexable_ptr_ty = sema.typeOf(indexable_ptr);
@ -24100,12 +24094,13 @@ fn elemPtr(
const index_val = maybe_index_val orelse break :rs elem_index_src;
const index = @intCast(usize, index_val.toUnsignedInt(target));
const elem_ptr = try ptr_val.elemPtr(indexable_ty, sema.arena, index, sema.mod);
const elem_ptr_ty = try sema.elemPtrType(indexable_ty, index, size);
return sema.addConstant(elem_ptr_ty, elem_ptr);
const result_ty = try sema.elemPtrType(indexable_ty, index);
return sema.addConstant(result_ty, elem_ptr);
};
const elem_ptr_ty = try sema.elemPtrType(indexable_ty, null, size);
const result_ty = try sema.elemPtrType(indexable_ty, null);
try sema.requireRuntimeBlock(block, src, runtime_src);
return block.addPtrElemPtr(indexable, elem_index, elem_ptr_ty);
return block.addPtrElemPtr(indexable, elem_index, result_ty);
},
.One => {
assert(indexable_ty.childType().zigTypeTag() == .Array); // Guaranteed by isIndexable
@ -24167,7 +24162,7 @@ fn elemVal(
},
.One => {
assert(indexable_ty.childType().zigTypeTag() == .Array); // Guaranteed by isIndexable
const elem_ptr = try sema.elemPtr(block, indexable_src, indexable, elem_index, elem_index_src, false, .One);
const elem_ptr = try sema.elemPtr(block, indexable_src, indexable, elem_index, elem_index_src, false);
return sema.analyzeLoad(block, indexable_src, elem_ptr, elem_index_src);
},
},
@ -24405,7 +24400,7 @@ fn elemPtrArray(
break :o index;
} else null;
const elem_ptr_ty = try sema.elemPtrType(array_ptr_ty, offset, .One);
const elem_ptr_ty = try sema.elemPtrType(array_ptr_ty, offset);
if (maybe_undef_array_ptr_val) |array_ptr_val| {
if (array_ptr_val.isUndef()) {
@ -24510,7 +24505,7 @@ fn elemPtrSlice(
break :o index;
} else null;
const elem_ptr_ty = try sema.elemPtrType(slice_ty, offset, .One);
const elem_ptr_ty = try sema.elemPtrType(slice_ty, offset);
if (maybe_undef_slice_val) |slice_val| {
if (slice_val.isUndef()) {
@ -26240,7 +26235,7 @@ fn storePtr2(
const elem_src = operand_src; // TODO better source location
const elem = try sema.tupleField(block, operand_src, uncasted_operand, elem_src, i);
const elem_index = try sema.addIntUnsigned(Type.usize, i);
const elem_ptr = try sema.elemPtr(block, ptr_src, ptr, elem_index, elem_src, false, .One);
const elem_ptr = try sema.elemPtr(block, ptr_src, ptr, elem_index, elem_src, false);
try sema.storePtr2(block, src, elem_ptr, elem_src, elem, elem_src, .store);
}
return;
@ -33277,12 +33272,7 @@ fn compareVector(
/// For []T, returns *T
/// Handles const-ness and address spaces in particular.
/// This code is duplicated in `analyzePtrArithmetic`.
fn elemPtrType(
sema: *Sema,
ptr_ty: Type,
offset: ?usize,
size: std.builtin.Type.Pointer.Size,
) !Type {
fn elemPtrType(sema: *Sema, ptr_ty: Type, offset: ?usize) !Type {
const ptr_info = ptr_ty.ptrInfo().data;
const elem_ty = ptr_ty.elemType2();
const allow_zero = ptr_info.@"allowzero" and (offset orelse 0) == 0;
@ -33327,7 +33317,6 @@ fn elemPtrType(
break :a new_align;
};
return try Type.ptr(sema.arena, sema.mod, .{
.size = size,
.pointee_type = elem_ty,
.mutable = ptr_info.mutable,
.@"addrspace" = ptr_info.@"addrspace",

View File

@ -79,7 +79,6 @@ pub fn extraData(code: Zir, comptime T: type, index: usize) struct { data: T, en
Inst.BuiltinCall.Flags => @bitCast(Inst.BuiltinCall.Flags, code.extra[i]),
Inst.SwitchBlock.Bits => @bitCast(Inst.SwitchBlock.Bits, code.extra[i]),
Inst.FuncFancy.Bits => @bitCast(Inst.FuncFancy.Bits, code.extra[i]),
Inst.ElemPtrImm.Bits => @bitCast(Inst.ElemPtrImm.Bits, code.extra[i]),
else => @compileError("bad field type"),
};
i += 1;
@ -501,14 +500,14 @@ pub const Inst = struct {
/// Uses the `node` field.
repeat_inline,
/// Asserts that all the lengths provided match. Used to build a for loop.
/// Return value is always void.
/// Return value is the length as a usize.
/// Uses the `pl_node` field with payload `MultiOp`.
/// There is exactly one item corresponding to each AST node inside the for
/// loop condition. Each item may be `none`, indicating an unbounded range.
/// loop condition. Any item may be `none`, indicating an unbounded range.
/// Illegal behaviors:
/// * If all lengths are unbounded ranges (always a compile error).
/// * If any two lengths do not match each other.
for_check_lens,
for_len,
/// Merge two error sets into one, `E1 || E2`.
/// Uses the `pl_node` field with payload `Bin`.
merge_error_sets,
@ -1254,7 +1253,7 @@ pub const Inst = struct {
.defer_err_code,
.save_err_ret_index,
.restore_err_ret_index,
.for_check_lens,
.for_len,
=> false,
.@"break",
@ -1322,7 +1321,6 @@ pub const Inst = struct {
.memcpy,
.memset,
.check_comptime_control_flow,
.for_check_lens,
.@"defer",
.defer_err_code,
.restore_err_ret_index,
@ -1547,6 +1545,7 @@ pub const Inst = struct {
.repeat_inline,
.panic,
.panic_comptime,
.for_len,
.@"try",
.try_ptr,
//.try_inline,
@ -1602,7 +1601,7 @@ pub const Inst = struct {
.@"break" = .@"break",
.break_inline = .@"break",
.check_comptime_control_flow = .un_node,
.for_check_lens = .pl_node,
.for_len = .pl_node,
.call = .pl_node,
.cmp_lt = .pl_node,
.cmp_lte = .pl_node,
@ -2975,13 +2974,7 @@ pub const Inst = struct {
pub const ElemPtrImm = struct {
ptr: Ref,
bits: Bits,
pub const Bits = packed struct(u32) {
index: u31,
/// Controls whether the type returned is `*T` or `[*]T`.
manyptr: bool = false,
};
index: u32,
};
/// 0. multi_cases_len: u32 // If has_multi_cases is set.

View File

@ -355,7 +355,7 @@ const Writer = struct {
.array_type,
=> try self.writePlNodeBin(stream, inst),
.for_check_lens => try self.writePlNodeMultiOp(stream, inst),
.for_len => try self.writePlNodeMultiOp(stream, inst),
.elem_ptr_imm => try self.writeElemPtrImm(stream, inst),
@ -888,9 +888,7 @@ const Writer = struct {
const extra = self.code.extraData(Zir.Inst.ElemPtrImm, inst_data.payload_index).data;
try self.writeInstRef(stream, extra.ptr);
try stream.print(", {d}", .{extra.bits.index});
try self.writeFlag(stream, ", manyptr", extra.bits.manyptr);
try stream.writeAll(") ");
try stream.print(", {d}) ", .{extra.index});
try self.writeSrc(stream, inst_data.src());
}