stage2: prepare for mainining Decl references to ZIR indexes

This commit is contained in:
Andrew Kelley 2021-04-28 16:55:22 -07:00
parent f86469bc5e
commit 3462193d30
4 changed files with 181 additions and 120 deletions

View File

@ -426,14 +426,6 @@ pub fn analyzeNamespace(
/// * C.FnBlock
/// * Wasm.FnData
/// * SpirV.FnData
/// This name is relative to the containing namespace of the decl.
/// The memory is owned by the containing File ZIR.
pub fn getName(decl: Decl) ?[:0]const u8 {
const zir = decl.namespace.file_scope.zir;
const name_index = zir.extra[decl.zir_decl_index + 4];
if (name_index <= 1) return null;
return zir.nullTerminatedString(name_index);
}
extra_index += @boolToInt(has_align);

View File

@ -154,13 +154,9 @@ pub const DeclPlusEmitH = struct {
};
pub const Decl = struct {
/// This name is relative to the containing namespace of the decl.
/// All Decls have names, even values that are not bound to a zig namespace.
/// This is necessary for mapping them to an address in the output file.
/// Memory is owned by this decl, using Module's allocator.
/// Note that this cannot be changed to reference ZIR memory because when
/// ZIR updates, it would change the Decl name, but we still need the previous
/// name to delete the Decl from the hash maps it has been inserted into.
/// For declarations that have corresponding source code, this is identical to
/// `getName().?`. For anonymous declarations this is allocated with Module's
/// allocator.
name: [*:0]const u8,
/// The most recent Type of the Decl after a successful semantic analysis.
/// Populated when `has_tv`.
@ -186,10 +182,10 @@ pub const Decl = struct {
/// The AST node index of this declaration.
/// Must be recomputed when the corresponding source file is modified.
src_node: ast.Node.Index,
/// Index to ZIR `extra` array to the block of ZIR code that encodes the Decl expression.
zir_block_index: Zir.Inst.Index,
zir_align_ref: Zir.Inst.Ref = .none,
zir_linksection_ref: Zir.Inst.Ref = .none,
/// Index to ZIR `extra` array to the entry in the parent's decl structure
/// (the part that says "for every decls_len"). The first item at this index is
/// the contents hash, followed by the name.
zir_decl_index: Zir.Inst.Index,
/// Represents the "shallow" analysis status. For example, for decls that are functions,
/// the function type is analyzed with this set to `in_progress`, however, the semantic
@ -234,6 +230,10 @@ pub const Decl = struct {
is_pub: bool,
/// Whether the corresponding AST decl has a `export` keyword.
is_exported: bool,
/// Whether the ZIR code provides an align instruction.
has_align: bool,
/// Whether the ZIR code provides a linksection instruction.
has_linksection: bool,
/// Represents the position of the code in the output file.
/// This is populated regardless of semantic analysis and code generation.
@ -246,11 +246,6 @@ pub const Decl = struct {
/// to save on memory usage.
fn_link: link.File.LinkFn,
/// This is stored separately in addition to being available via `zir_decl_index`
/// because when the underlying ZIR code is updated, this field is used to find
/// out if anything changed.
contents_hash: std.zig.SrcHash,
/// The shallow set of other decls whose typed_value could possibly change if this Decl's
/// typed_value is modified.
dependants: DepsTable = .{},
@ -260,7 +255,13 @@ pub const Decl = struct {
/// The reason this is not `std.AutoArrayHashMapUnmanaged` is a workaround for
/// stage1 compiler giving me: `error: struct 'Module.Decl' depends on itself`
pub const DepsTable = std.ArrayHashMapUnmanaged(*Decl, void, std.array_hash_map.getAutoHashFn(*Decl), std.array_hash_map.getAutoEqlFn(*Decl), false);
pub const DepsTable = std.ArrayHashMapUnmanaged(
*Decl,
void,
std.array_hash_map.getAutoHashFn(*Decl),
std.array_hash_map.getAutoEqlFn(*Decl),
false,
);
pub fn destroy(decl: *Decl, module: *Module) void {
const gpa = module.gpa;
@ -283,6 +284,48 @@ pub const Decl = struct {
}
}
/// This name is relative to the containing namespace of the decl.
/// The memory is owned by the containing File ZIR.
pub fn getName(decl: Decl) ?[:0]const u8 {
const zir = decl.namespace.file_scope.zir;
return decl.getNameZir(zir);
}
pub fn getNameZir(decl: Decl, zir: Zir) ?[:0]const u8 {
const name_index = zir.extra[decl.zir_decl_index + 4];
if (name_index <= 1) return null;
return zir.nullTerminatedString(name_index);
}
pub fn contentsHash(decl: Decl) std.zig.SrcHash {
const zir = decl.namespace.file_scope.zir;
return decl.contentsHashZir(zir);
}
pub fn contentsHashZir(decl: Decl, zir: Zir) std.zig.SrcHash {
const hash_u32s = zir.extra[decl.zir_decl_index..][0..4];
const contents_hash = @bitCast(std.zig.SrcHash, hash_u32s.*);
return contents_hash;
}
pub fn zirBlockIndex(decl: Decl) Zir.Inst.Index {
const zir = decl.namespace.file_scope.zir;
return zir.extra[decl.zir_decl_index + 5];
}
pub fn zirAlignRef(decl: Decl) Zir.Inst.Ref {
if (!decl.has_align) return .none;
const zir = decl.namespace.file_scope.zir;
return @intToEnum(Zir.Inst.Ref, zir.extra[decl.zir_decl_index + 6]);
}
pub fn zirLinkSectionRef(decl: Decl) Zir.Inst.Ref {
if (!decl.has_linksection) return .none;
const zir = decl.namespace.file_scope.zir;
const extra_index = decl.zir_decl_index + 6 + @boolToInt(decl.has_align);
return @intToEnum(Zir.Inst.Ref, zir.extra[extra_index]);
}
pub fn relativeToNodeIndex(decl: Decl, offset: i32) ast.Node.Index {
return @bitCast(ast.Node.Index, offset + @bitCast(i32, decl.src_node));
}
@ -653,9 +696,6 @@ pub const Scope = struct {
/// TODO save memory with https://github.com/ziglang/zig/issues/8619.
/// Does not contain anonymous decls.
decls: std.StringArrayHashMapUnmanaged(*Decl) = .{},
/// Names imported into the namespace via `usingnamespace`.
/// The key memory is owned by the ZIR of the `File` containing the `Namespace`.
usingnamespace_decls: std.StringArrayHashMapUnmanaged(*Namespace) = .{},
pub fn deinit(ns: *Namespace, mod: *Module) void {
const gpa = mod.gpa;
@ -715,6 +755,11 @@ pub const Scope = struct {
/// The namespace of the struct that represents this file.
/// Populated only when status is success.
namespace: *Namespace,
/// All namespaces that this file contains. This is here so that
/// when a file is updated, and new ZIR code is generated, the
/// old and new ZIR code can be compared side by side and references
/// to old ZIR updated to new ZIR, and a changelist generated.
namespace_set: std.AutoArrayHashMapUnmanaged(*Namespace, void) = .{},
pub fn unload(file: *File, gpa: *Allocator) void {
file.unloadTree(gpa);
@ -2919,17 +2964,15 @@ pub fn astGenFile(mod: *Module, file: *Scope.File, prog_node: *std.Progress.Node
},
};
// Clear compile error for this file.
switch (file.status) {
.success, .retryable_failure => {},
.never_loaded, .parse_failure, .astgen_failure => {
const lock = comp.mutex.acquire();
defer lock.release();
if (mod.failed_files.swapRemove(file)) |entry| {
if (entry.value) |msg| msg.destroy(gpa); // Delete previous error message.
}
},
}
mod.lockAndClearFileCompileError(file);
// Move previous ZIR to a local variable so we can compare it with the new one.
var prev_zir = file.zir;
const prev_zir_loaded = file.zir_loaded;
file.zir_loaded = false;
file.zir = undefined;
defer if (prev_zir_loaded) prev_zir.deinit(gpa);
file.unload(gpa);
if (stat.size > std.math.maxInt(u32))
@ -3041,6 +3084,18 @@ pub fn astGenFile(mod: *Module, file: *Scope.File, prog_node: *std.Progress.Node
});
};
if (prev_zir_loaded) {
// Iterate over all Namespace objects contained within this File, looking at the
// previous and new ZIR together and update the references to point
// to the new one. For example, Decl name, Decl zir_decl_index, and Namespace
// decl_table keys need to get updated to point to the new memory, even if the
// underlying source code is unchanged.
// We do not need to hold any locks at this time because all the Decl and Namespace
// objects being touched are specific to this File, and the only other concurrent
// tasks are touching other File objects.
@panic("TODO implement update references from old ZIR to new ZIR");
}
// TODO don't report compile errors until Sema @importFile
if (file.zir.hasCompileErrors()) {
{
@ -3168,10 +3223,11 @@ pub fn semaFile(mod: *Module, file: *Scope.File) InnerError!void {
.deletion_flag = false,
.is_pub = true,
.is_exported = false,
.has_linksection = false,
.has_align = false,
.link = undefined, // don't try to codegen this
.fn_link = undefined, // not a function
.contents_hash = undefined, // top-level struct has no contents hash
.zir_block_index = undefined,
.zir_decl_index = undefined,
.has_tv = false,
.ty = undefined,
@ -3263,15 +3319,16 @@ fn semaDecl(mod: *Module, decl: *Decl) !bool {
};
defer block_scope.instructions.deinit(gpa);
const inst_data = zir.instructions.items(.data)[decl.zir_block_index].pl_node;
const zir_block_index = decl.zirBlockIndex();
const inst_data = zir.instructions.items(.data)[zir_block_index].pl_node;
const extra = zir.extraData(Zir.Inst.Block, inst_data.payload_index);
const body = zir.extra[extra.end..][0..extra.data.body_len];
const break_index = try sema.analyzeBody(&block_scope, body);
if (decl.zir_align_ref != .none) {
if (decl.zirAlignRef() != .none) {
@panic("TODO implement decl align");
}
if (decl.zir_linksection_ref != .none) {
if (decl.zirLinkSectionRef() != .none) {
@panic("TODO implement decl linksection");
}
@ -3457,51 +3514,25 @@ pub fn scanNamespace(
var bit_bag_index: usize = extra_start;
var cur_bit_bag: u32 = undefined;
var decl_i: u32 = 0;
var scan_decl_iter: ScanDeclIter = .{
.module = mod,
.namespace = namespace,
.deleted_decls = &deleted_decls,
.outdated_decls = &outdated_decls,
.parent_decl = parent_decl,
};
while (decl_i < decls_len) : (decl_i += 1) {
if (decl_i % 8 == 0) {
cur_bit_bag = zir.extra[bit_bag_index];
bit_bag_index += 1;
}
const is_pub = @truncate(u1, cur_bit_bag) != 0;
cur_bit_bag >>= 1;
const is_exported = @truncate(u1, cur_bit_bag) != 0;
cur_bit_bag >>= 1;
const has_align = @truncate(u1, cur_bit_bag) != 0;
cur_bit_bag >>= 1;
const has_section = @truncate(u1, cur_bit_bag) != 0;
cur_bit_bag >>= 1;
const flags = @truncate(u4, cur_bit_bag);
const decl_sub_index = extra_index;
extra_index += 6;
extra_index += @truncate(u1, flags >> 2);
extra_index += @truncate(u1, flags >> 3);
const hash_u32s = zir.extra[extra_index..][0..4];
extra_index += 4;
const decl_name_index = zir.extra[extra_index];
extra_index += 1;
const decl_index = zir.extra[extra_index];
extra_index += 1;
const align_inst: Zir.Inst.Ref = if (!has_align) .none else inst: {
const inst = @intToEnum(Zir.Inst.Ref, zir.extra[extra_index]);
extra_index += 1;
break :inst inst;
};
const section_inst: Zir.Inst.Ref = if (!has_section) .none else inst: {
const inst = @intToEnum(Zir.Inst.Ref, zir.extra[extra_index]);
extra_index += 1;
break :inst inst;
};
const contents_hash = @bitCast(std.zig.SrcHash, hash_u32s.*);
try mod.scanDecl(
namespace,
&deleted_decls,
&outdated_decls,
contents_hash,
decl_name_index,
decl_index,
is_pub,
is_exported,
align_inst,
section_inst,
parent_decl,
);
try scanDecl(&scan_decl_iter, decl_sub_index, flags);
}
// Handle explicitly deleted decls from the source code. This is one of two
// places that Decl deletions happen. The other is in `Compilation`, after
@ -3528,62 +3559,80 @@ pub fn scanNamespace(
return extra_index;
}
fn scanDecl(
mod: *Module,
const ScanDeclIter = struct {
module: *Module,
namespace: *Scope.Namespace,
deleted_decls: *std.AutoArrayHashMap(*Decl, void),
outdated_decls: *std.AutoArrayHashMap(*Decl, void),
contents_hash: std.zig.SrcHash,
decl_name_index: u32,
decl_index: Zir.Inst.Index,
is_pub: bool,
is_exported: bool,
align_inst: Zir.Inst.Ref,
section_inst: Zir.Inst.Ref,
parent_decl: *Decl,
) InnerError!void {
usingnamespace_index: usize = 0,
comptime_index: usize = 0,
unnamed_test_index: usize = 0,
};
fn scanDecl(iter: *ScanDeclIter, decl_sub_index: usize, flags: u4) InnerError!void {
const tracy = trace(@src());
defer tracy.end();
const mod = iter.module;
const namespace = iter.namespace;
const gpa = mod.gpa;
const zir = namespace.file_scope.zir;
const decl_block_inst_data = zir.instructions.items(.data)[decl_index].pl_node;
const decl_node = parent_decl.relativeToNodeIndex(decl_block_inst_data.src_node);
// zig fmt: off
const is_pub = (flags & 0b0001) != 0;
const is_exported = (flags & 0b0010) != 0;
const has_align = (flags & 0b0100) != 0;
const has_linksection = (flags & 0b1000) != 0;
// zig fmt: on
const decl_name: ?[]const u8 = if (decl_name_index > 1)
zir.nullTerminatedString(decl_name_index)
else
null;
const decl_name_index = zir.extra[decl_sub_index + 4];
const decl_index = zir.extra[decl_sub_index + 5];
const decl_block_inst_data = zir.instructions.items(.data)[decl_index].pl_node;
const decl_node = iter.parent_decl.relativeToNodeIndex(decl_block_inst_data.src_node);
// Every Decl needs a name.
const decl_name: [:0]const u8 = switch (decl_name_index) {
0 => name: {
if (is_exported) {
const i = iter.usingnamespace_index;
iter.usingnamespace_index += 1;
break :name try std.fmt.allocPrintZ(gpa, "usingnamespace${d}", .{i});
} else {
const i = iter.comptime_index;
iter.comptime_index += 1;
break :name try std.fmt.allocPrintZ(gpa, "comptime${d}", .{i});
}
},
1 => name: {
const i = iter.unnamed_test_index;
iter.unnamed_test_index += 1;
break :name try std.fmt.allocPrintZ(gpa, "test${d}", .{i});
},
else => zir.nullTerminatedString(decl_name_index),
};
// We create a Decl for it regardless of analysis status.
// Decls that have names are keyed in the namespace by the name. Decls without
// names are keyed by their contents hash. This way we can detect if, for example,
// a comptime decl gets moved around in the file.
const decl_key = decl_name orelse &contents_hash;
const gop = try namespace.decls.getOrPut(gpa, decl_key);
const gop = try namespace.decls.getOrPut(gpa, decl_name);
if (!gop.found_existing) {
const new_decl = try mod.allocateNewDecl(namespace, decl_node);
new_decl.contents_hash = contents_hash;
new_decl.name = try gpa.dupeZ(u8, decl_key);
// Update the key reference to the longer-lived memory.
gop.entry.key = &new_decl.contents_hash;
new_decl.name = decl_name;
gop.entry.value = new_decl;
// Exported decls, comptime decls, usingnamespace decls, and
// test decls if in test mode, get analyzed.
const want_analysis = is_exported or switch (decl_name_index) {
0 => true, // comptime decl
1 => mod.comp.bin_file.options.is_test, // test decl
else => false,
else => false, // TODO set to true for named tests when testing
};
if (want_analysis) {
mod.comp.work_queue.writeItemAssumeCapacity(.{ .analyze_decl = new_decl });
}
new_decl.is_exported = is_exported;
new_decl.is_pub = is_pub;
new_decl.zir_block_index = decl_index;
new_decl.zir_align_ref = align_inst;
new_decl.zir_linksection_ref = section_inst;
new_decl.is_exported = is_exported;
new_decl.has_align = has_align;
new_decl.has_linksection = has_linksection;
new_decl.zir_decl_index = @intCast(u32, decl_sub_index);
return;
}
const decl = gop.entry.value;
@ -3591,12 +3640,13 @@ fn scanDecl(
// have been re-ordered.
const prev_src_node = decl.src_node;
decl.src_node = decl_node;
decl.is_pub = is_pub;
decl.is_exported = is_exported;
decl.zir_block_index = decl_index;
decl.zir_align_ref = align_inst;
decl.zir_linksection_ref = section_inst;
if (deleted_decls.swapRemove(decl) == null) {
decl.has_align = has_align;
decl.has_linksection = has_linksection;
decl.zir_decl_index = @intCast(u32, decl_sub_index);
if (iter.deleted_decls.swapRemove(decl) == null) {
if (true) {
@panic("TODO I think this code path is unreachable; should be caught by AstGen.");
}
@ -3615,8 +3665,11 @@ fn scanDecl(
try mod.errNoteNonLazy(other_src_loc, msg, "previously declared here", .{});
try mod.failed_decls.putNoClobber(gpa, decl, msg);
} else {
if (true) {
@panic("TODO reimplement scanDecl with regards to incremental compilation.");
}
if (!std.zig.srcHashEql(decl.contents_hash, contents_hash)) {
try outdated_decls.put(decl, {});
try iter.outdated_decls.put(decl, {});
decl.contents_hash = contents_hash;
} else if (try decl.isFunction()) switch (mod.comp.bin_file.tag) {
.coff => {
@ -3848,8 +3901,7 @@ fn allocateNewDecl(mod: *Module, namespace: *Scope.Namespace, src_node: ast.Node
.linksection_val = undefined,
.analysis = .unreferenced,
.deletion_flag = false,
.contents_hash = undefined,
.zir_block_index = undefined,
.zir_decl_index = undefined,
.link = switch (mod.comp.bin_file.tag) {
.coff => .{ .coff = link.File.Coff.TextBlock.empty },
.elf => .{ .elf = link.File.Elf.TextBlock.empty },
@ -3869,6 +3921,8 @@ fn allocateNewDecl(mod: *Module, namespace: *Scope.Namespace, src_node: ast.Node
.generation = 0,
.is_pub = false,
.is_exported = false,
.has_linksection = false,
.has_align = false,
};
return new_decl;
}
@ -4602,3 +4656,16 @@ pub fn getTarget(mod: Module) Target {
pub fn optimizeMode(mod: Module) std.builtin.Mode {
return mod.comp.bin_file.options.optimize_mode;
}
fn lockAndClearFileCompileError(mod: *Module, file: *Scope.File) void {
switch (file.status) {
.success, .retryable_failure => {},
.never_loaded, .parse_failure, .astgen_failure => {
const lock = mod.comp.mutex.acquire();
defer lock.release();
if (mod.failed_files.swapRemove(file)) |entry| {
if (entry.value) |msg| msg.destroy(mod.gpa); // Delete previous error message.
}
},
}
}

View File

@ -79,6 +79,7 @@ pub fn deinit(self: *C) void {
pub fn allocateDeclIndexes(self: *C, decl: *Module.Decl) !void {}
pub fn freeDecl(self: *C, decl: *Module.Decl) void {
self.decl_table.removeAssertDiscard(decl);
decl.link.c.code.deinit(self.base.allocator);
decl.fn_link.c.fwd_decl.deinit(self.base.allocator);
var it = decl.fn_link.c.typedefs.iterator();

View File

@ -127,6 +127,7 @@ pub fn updateDeclExports(
) !void {}
pub fn freeDecl(self: *SpirV, decl: *Module.Decl) void {
self.decl_table.removeAssertDiscard(decl);
var fn_data = decl.fn_link.spirv;
fn_data.code.deinit(self.base.allocator);
if (fn_data.id) |id| self.spirv_module.freeId(id);