M-Labs/nac3
M-Labs
/
nac3
8
0
Fork 10
Code Issues 97 Pull Requests 19 Releases Wiki Activity

nac3core/codegen: refactor according to #23

This commit is contained in:
pca006132 2021-10-16 22:17:36 +08:00
parent c4259d14d1
commit 1f5bea2448
7 changed files with 954 additions and 732 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
nac3artiq/src/lib.rs
View File

@ -19,7 +19,7 @@ use rustpython_parser::{
use parking_lot::{Mutex, RwLock}; use parking_lot::{Mutex, RwLock};
use nac3core::{ use nac3core::{
codegen::{CodeGenTask, WithCall, WorkerRegistry}, codegen::{CodeGenTask, DefaultCodeGenerator, WithCall, WorkerRegistry},
symbol_resolver::SymbolResolver, symbol_resolver::SymbolResolver,
toplevel::{composer::TopLevelComposer, DefinitionId, GenCall, TopLevelContext, TopLevelDef}, toplevel::{composer::TopLevelComposer, DefinitionId, GenCall, TopLevelContext, TopLevelDef},
typecheck::typedef::{FunSignature, FuncArg}, typecheck::typedef::{FunSignature, FuncArg},
@ -423,11 +423,13 @@ impl Nac3 {
.expect("couldn't write module to file"); .expect("couldn't write module to file");
}))); })));
let thread_names: Vec<String> = (0..4).map(|i| format!("module{}", i)).collect(); let thread_names: Vec<String> = (0..4).map(|i| format!("module{}", i)).collect();
let threads: Vec<_> = thread_names.iter().map(|s| s.as_str()).collect(); let threads: Vec<_> = thread_names
.iter()
.map(|s| Box::new(DefaultCodeGenerator::new(s.to_string())))
.collect();
py.allow_threads(|| { py.allow_threads(|| {
let (registry, handles) = let (registry, handles) = WorkerRegistry::create_workers(threads, top_level.clone(), f);
WorkerRegistry::create_workers(&threads, top_level.clone(), f);
registry.add_task(task); registry.add_task(task);
registry.wait_tasks_complete(handles); registry.wait_tasks_complete(handles);
}); });

617
nac3core/src/codegen/expr.rs
View File

@ -14,6 +14,8 @@ use inkwell::{
use itertools::{chain, izip, zip, Itertools}; use itertools::{chain, izip, zip, Itertools};
use rustpython_parser::ast::{self, Boolop, Constant, Expr, ExprKind, Operator, StrRef}; use rustpython_parser::ast::{self, Boolop, Constant, Expr, ExprKind, Operator, StrRef};
use super::CodeGenerator;
pub fn assert_int_val(val: BasicValueEnum<'_>) -> IntValue<'_> { pub fn assert_int_val(val: BasicValueEnum<'_>) -> IntValue<'_> {
if let BasicValueEnum::IntValue(v) = val { if let BasicValueEnum::IntValue(v) = val {
v v
@ -102,172 +104,6 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
get_llvm_type(self.ctx, &mut self.unifier, self.top_level, &mut self.type_cache, ty) get_llvm_type(self.ctx, &mut self.unifier, self.top_level, &mut self.type_cache, ty)
} }
fn gen_call(
&mut self,
obj: Option<(Type, BasicValueEnum<'ctx>)>,
fun: (&FunSignature, DefinitionId),
params: Vec<(Option<StrRef>, BasicValueEnum<'ctx>)>,
) -> Option<BasicValueEnum<'ctx>> {
let definition = self.top_level.definitions.read().get(fun.1 .0).cloned().unwrap();
let mut task = None;
let key = self.get_subst_key(obj.map(|a| a.0), fun.0, None);
let symbol = {
// make sure this lock guard is dropped at the end of this scope...
let def = definition.read();
match &*def {
TopLevelDef::Function { instance_to_symbol, codegen_callback, .. } => {
if let Some(callback) = codegen_callback {
return callback.run(self, obj, fun, params);
}
instance_to_symbol.get(&key).cloned()
}
TopLevelDef::Class { methods, .. } => {
// TODO: what about other fields that require alloca?
let mut fun_id = None;
for (name, _, id) in methods.iter() {
if name == &"__init__".into() {
fun_id = Some(*id);
}
}
let ty = self.get_llvm_type(fun.0.ret).into_pointer_type();
let zelf_ty: BasicTypeEnum = ty.get_element_type().try_into().unwrap();
let zelf = self.builder.build_alloca(zelf_ty, "alloca").into();
// call `__init__` if there is one
if let Some(fun_id) = fun_id {
let mut sign = fun.0.clone();
sign.ret = self.primitives.none;
self.gen_call(Some((fun.0.ret, zelf)), (&sign, fun_id), params);
}
return Some(zelf);
}
}
}
.unwrap_or_else(|| {
if let TopLevelDef::Function {
name,
instance_to_symbol,
instance_to_stmt,
var_id,
resolver,
..
} = &mut *definition.write()
{
instance_to_symbol.get(&key).cloned().unwrap_or_else(|| {
let symbol = format!("{}.{}", name, instance_to_symbol.len());
instance_to_symbol.insert(key, symbol.clone());
let key = self.get_subst_key(obj.map(|a| a.0), fun.0, Some(var_id));
let instance = instance_to_stmt.get(&key).unwrap();
let unifiers = self.top_level.unifiers.read();
let (unifier, primitives) = &unifiers[instance.unifier_id];
let mut unifier = Unifier::from_shared_unifier(unifier);
let mut type_cache = [
(self.primitives.int32, primitives.int32),
(self.primitives.int64, primitives.int64),
(self.primitives.float, primitives.float),
(self.primitives.bool, primitives.bool),
(self.primitives.none, primitives.none),
]
.iter()
.map(|(a, b)| {
(self.unifier.get_representative(*a), unifier.get_representative(*b))
})
.collect();
let subst = fun
.0
.vars
.iter()
.map(|(id, ty)| {
(
*instance.subst.get(id).unwrap(),
unifier.copy_from(&mut self.unifier, *ty, &mut type_cache),
)
})
.collect();
let mut signature = FunSignature {
args: fun
.0
.args
.iter()
.map(|arg| FuncArg {
name: arg.name,
ty: unifier.copy_from(&mut self.unifier, arg.ty, &mut type_cache),
default_value: arg.default_value.clone(),
})
.collect(),
ret: unifier.copy_from(&mut self.unifier, fun.0.ret, &mut type_cache),
vars: fun
.0
.vars
.iter()
.map(|(id, ty)| {
(*id, unifier.copy_from(&mut self.unifier, *ty, &mut type_cache))
})
.collect(),
};
if let Some(obj) = &obj {
signature.args.insert(
0,
FuncArg { name: "self".into(), ty: obj.0, default_value: None },
);
}
let unifier = (unifier.get_shared_unifier(), *primitives);
task = Some(CodeGenTask {
symbol_name: symbol.clone(),
body: instance.body.clone(),
resolver: resolver.as_ref().unwrap().clone(),
calls: instance.calls.clone(),
subst,
signature,
unifier,
});
symbol
})
} else {
unreachable!()
}
});
if let Some(task) = task {
self.registry.add_task(task);
}
let fun_val = self.module.get_function(&symbol).unwrap_or_else(|| {
let mut args = fun.0.args.clone();
if let Some(obj) = &obj {
args.insert(0, FuncArg { name: "self".into(), ty: obj.0, default_value: None });
}
let params = args.iter().map(|arg| self.get_llvm_type(arg.ty)).collect_vec();
let fun_ty = if self.unifier.unioned(fun.0.ret, self.primitives.none) {
self.ctx.void_type().fn_type(&params, false)
} else {
self.get_llvm_type(fun.0.ret).fn_type(&params, false)
};
self.module.add_function(&symbol, fun_ty, None)
});
let mut keys = fun.0.args.clone();
let mut mapping = HashMap::new();
for (key, value) in params.into_iter() {
mapping.insert(key.unwrap_or_else(|| keys.remove(0).name), value);
}
// default value handling
for k in keys.into_iter() {
mapping.insert(k.name, self.gen_symbol_val(&k.default_value.unwrap()));
}
// reorder the parameters
let mut params =
fun.0.args.iter().map(|arg| mapping.remove(&arg.name).unwrap()).collect_vec();
if let Some(obj) = obj {
params.insert(0, obj.1);
}
self.builder.build_call(fun_val, &params, "call").try_as_basic_value().left()
}
fn gen_const(&mut self, value: &Constant, ty: Type) -> BasicValueEnum<'ctx> { fn gen_const(&mut self, value: &Constant, ty: Type) -> BasicValueEnum<'ctx> {
match value { match value {
Constant::Bool(v) => { Constant::Bool(v) => {
@ -378,165 +214,347 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
_ => unimplemented!(), _ => unimplemented!(),
} }
} }
}
pub fn gen_expr(&mut self, expr: &Expr<Option<Type>>) -> Option<BasicValueEnum<'ctx>> { pub fn gen_constructor<'ctx, 'a, G: CodeGenerator + ?Sized>(
let zero = self.ctx.i32_type().const_int(0, false); generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
signature: &FunSignature,
def: &TopLevelDef,
params: Vec<(Option<StrRef>, BasicValueEnum<'ctx>)>,
) -> BasicValueEnum<'ctx> {
match def {
TopLevelDef::Class { methods, .. } => {
// TODO: what about other fields that require alloca?
let mut fun_id = None;
for (name, _, id) in methods.iter() {
if name == &"__init__".into() {
fun_id = Some(*id);
}
}
let ty = ctx.get_llvm_type(signature.ret).into_pointer_type();
let zelf_ty: BasicTypeEnum = ty.get_element_type().try_into().unwrap();
let zelf = ctx.builder.build_alloca(zelf_ty, "alloca").into();
// call `__init__` if there is one
if let Some(fun_id) = fun_id {
let mut sign = signature.clone();
sign.ret = ctx.primitives.none;
generator.gen_call(ctx, Some((signature.ret, zelf)), (&sign, fun_id), params);
}
zelf
}
_ => unreachable!(),
}
}
pub fn gen_func_instance<'ctx, 'a>(
ctx: &mut CodeGenContext<'ctx, 'a>,
obj: Option<(Type, BasicValueEnum<'ctx>)>,
fun: (&FunSignature, &mut TopLevelDef, String),
) -> String {
if let (
sign,
TopLevelDef::Function {
name, instance_to_symbol, instance_to_stmt, var_id, resolver, ..
},
key,
) = fun
{
instance_to_symbol.get(&key).cloned().unwrap_or_else(|| {
let symbol = format!("{}.{}", name, instance_to_symbol.len());
instance_to_symbol.insert(key, symbol.clone());
let key = ctx.get_subst_key(obj.map(|a| a.0), sign, Some(var_id));
let instance = instance_to_stmt.get(&key).unwrap();
let unifiers = ctx.top_level.unifiers.read();
let (unifier, primitives) = &unifiers[instance.unifier_id];
let mut unifier = Unifier::from_shared_unifier(unifier);
let mut type_cache = [
(ctx.primitives.int32, primitives.int32),
(ctx.primitives.int64, primitives.int64),
(ctx.primitives.float, primitives.float),
(ctx.primitives.bool, primitives.bool),
(ctx.primitives.none, primitives.none),
]
.iter()
.map(|(a, b)| (ctx.unifier.get_representative(*a), unifier.get_representative(*b)))
.collect();
let subst = sign
.vars
.iter()
.map(|(id, ty)| {
(
*instance.subst.get(id).unwrap(),
unifier.copy_from(&mut ctx.unifier, *ty, &mut type_cache),
)
})
.collect();
let mut signature = FunSignature {
args: sign
.args
.iter()
.map(|arg| FuncArg {
name: arg.name,
ty: unifier.copy_from(&mut ctx.unifier, arg.ty, &mut type_cache),
default_value: arg.default_value.clone(),
})
.collect(),
ret: unifier.copy_from(&mut ctx.unifier, sign.ret, &mut type_cache),
vars: sign
.vars
.iter()
.map(|(id, ty)| {
(*id, unifier.copy_from(&mut ctx.unifier, *ty, &mut type_cache))
})
.collect(),
};
if let Some(obj) = &obj {
signature
.args
.insert(0, FuncArg { name: "self".into(), ty: obj.0, default_value: None });
}
let unifier = (unifier.get_shared_unifier(), *primitives);
ctx.registry.add_task(CodeGenTask {
symbol_name: symbol.clone(),
body: instance.body.clone(),
resolver: resolver.as_ref().unwrap().clone(),
calls: instance.calls.clone(),
subst,
signature,
unifier,
});
symbol
})
} else {
unreachable!()
}
}
pub fn gen_call<'ctx, 'a, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
obj: Option<(Type, BasicValueEnum<'ctx>)>,
fun: (&FunSignature, DefinitionId),
params: Vec<(Option<StrRef>, BasicValueEnum<'ctx>)>,
) -> Option<BasicValueEnum<'ctx>> {
let definition = ctx.top_level.definitions.read().get(fun.1 .0).cloned().unwrap();
let key = ctx.get_subst_key(obj.map(|a| a.0), fun.0, None);
let symbol = {
// make sure this lock guard is dropped at the end of this scope...
let def = definition.read();
match &*def {
TopLevelDef::Function { instance_to_symbol, codegen_callback, .. } => {
if let Some(callback) = codegen_callback {
return callback.run(ctx, obj, fun, params);
}
instance_to_symbol.get(&key).cloned()
}
TopLevelDef::Class { .. } => {
return Some(generator.gen_constructor(ctx, fun.0, &*def, params))
}
}
}
.unwrap_or_else(|| {
generator.gen_func_instance(ctx, obj, (fun.0, &mut *definition.write(), key))
});
let fun_val = ctx.module.get_function(&symbol).unwrap_or_else(|| {
let mut args = fun.0.args.clone();
if let Some(obj) = &obj {
args.insert(0, FuncArg { name: "self".into(), ty: obj.0, default_value: None });
}
let params = args.iter().map(|arg| ctx.get_llvm_type(arg.ty)).collect_vec();
let fun_ty = if ctx.unifier.unioned(fun.0.ret, ctx.primitives.none) {
ctx.ctx.void_type().fn_type(&params, false)
} else {
ctx.get_llvm_type(fun.0.ret).fn_type(&params, false)
};
ctx.module.add_function(&symbol, fun_ty, None)
});
let mut keys = fun.0.args.clone();
let mut mapping = HashMap::new();
for (key, value) in params.into_iter() {
mapping.insert(key.unwrap_or_else(|| keys.remove(0).name), value);
}
// default value handling
for k in keys.into_iter() {
mapping.insert(k.name, ctx.gen_symbol_val(&k.default_value.unwrap()));
}
// reorder the parameters
let mut params = fun.0.args.iter().map(|arg| mapping.remove(&arg.name).unwrap()).collect_vec();
if let Some(obj) = obj {
params.insert(0, obj.1);
}
ctx.builder.build_call(fun_val, &params, "call").try_as_basic_value().left()
}
pub fn gen_expr<'ctx, 'a, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
expr: &Expr<Option<Type>>,
) -> Option<BasicValueEnum<'ctx>> {
let zero = ctx.ctx.i32_type().const_int(0, false);
Some(match &expr.node { Some(match &expr.node {
ExprKind::Constant { value, .. } => { ExprKind::Constant { value, .. } => {
let ty = expr.custom.unwrap(); let ty = expr.custom.unwrap();
self.gen_const(value, ty) ctx.gen_const(value, ty)
} }
ExprKind::Name { id, .. } => { ExprKind::Name { id, .. } => {
let ptr = self.var_assignment.get(id); let ptr = ctx.var_assignment.get(id);
if let Some(ptr) = ptr { if let Some(ptr) = ptr {
self.builder.build_load(*ptr, "load") ctx.builder.build_load(*ptr, "load")
} else { } else {
let resolver = self.resolver.clone(); let resolver = ctx.resolver.clone();
resolver.get_symbol_value(*id, self).unwrap() resolver.get_symbol_value(*id, ctx).unwrap()
} }
} }
ExprKind::List { elts, .. } => { ExprKind::List { elts, .. } => {
// this shall be optimized later for constant primitive lists... // this shall be optimized later for constant primitive lists...
// we should use memcpy for that instead of generating thousands of stores // we should use memcpy for that instead of generating thousands of stores
let elements = elts.iter().map(|x| self.gen_expr(x).unwrap()).collect_vec(); let elements = elts.iter().map(|x| generator.gen_expr(ctx, x).unwrap()).collect_vec();
let ty = if elements.is_empty() { let ty = if elements.is_empty() {
self.ctx.i32_type().into() ctx.ctx.i32_type().into()
} else { } else {
elements[0].get_type() elements[0].get_type()
}; };
let arr_ptr = self.builder.build_array_alloca( let arr_ptr = ctx.builder.build_array_alloca(
ty, ty,
self.ctx.i32_type().const_int(elements.len() as u64, false), ctx.ctx.i32_type().const_int(elements.len() as u64, false),
"tmparr", "tmparr",
); );
let arr_ty = self.ctx.struct_type( let arr_ty = ctx.ctx.struct_type(
&[self.ctx.i32_type().into(), ty.ptr_type(AddressSpace::Generic).into()], &[ctx.ctx.i32_type().into(), ty.ptr_type(AddressSpace::Generic).into()],
false, false,
); );
let arr_str_ptr = self.builder.build_alloca(arr_ty, "tmparrstr"); let arr_str_ptr = ctx.builder.build_alloca(arr_ty, "tmparrstr");
unsafe { unsafe {
let len_ptr = let len_ptr =
self.builder.build_in_bounds_gep(arr_str_ptr, &[zero, zero], "len_ptr"); ctx.builder.build_in_bounds_gep(arr_str_ptr, &[zero, zero], "len_ptr");
self.builder.build_store( ctx.builder.build_store(
len_ptr, len_ptr,
self.ctx.i32_type().const_int(elements.len() as u64, false), ctx.ctx.i32_type().const_int(elements.len() as u64, false),
); );
let ptr_to_arr = self.builder.build_in_bounds_gep( let ptr_to_arr = ctx.builder.build_in_bounds_gep(
arr_str_ptr, arr_str_ptr,
&[zero, self.ctx.i32_type().const_int(1, false)], &[zero, ctx.ctx.i32_type().const_int(1, false)],
"ptr_to_arr", "ptr_to_arr",
); );
self.builder.build_store(ptr_to_arr, arr_ptr); ctx.builder.build_store(ptr_to_arr, arr_ptr);
let i32_type = self.ctx.i32_type(); let i32_type = ctx.ctx.i32_type();
for (i, v) in elements.iter().enumerate() { for (i, v) in elements.iter().enumerate() {
let elem_ptr = self.builder.build_in_bounds_gep( let elem_ptr = ctx.builder.build_in_bounds_gep(
arr_ptr, arr_ptr,
&[i32_type.const_int(i as u64, false)], &[i32_type.const_int(i as u64, false)],
"elem_ptr", "elem_ptr",
); );
self.builder.build_store(elem_ptr, *v); ctx.builder.build_store(elem_ptr, *v);
} }
} }
arr_str_ptr.into() arr_str_ptr.into()
} }
ExprKind::Tuple { elts, .. } => { ExprKind::Tuple { elts, .. } => {
let element_val = elts.iter().map(|x| self.gen_expr(x).unwrap()).collect_vec(); let element_val =
elts.iter().map(|x| generator.gen_expr(ctx, x).unwrap()).collect_vec();
let element_ty = element_val.iter().map(BasicValueEnum::get_type).collect_vec(); let element_ty = element_val.iter().map(BasicValueEnum::get_type).collect_vec();
let tuple_ty = self.ctx.struct_type(&element_ty, false); let tuple_ty = ctx.ctx.struct_type(&element_ty, false);
let tuple_ptr = self.builder.build_alloca(tuple_ty, "tuple"); let tuple_ptr = ctx.builder.build_alloca(tuple_ty, "tuple");
for (i, v) in element_val.into_iter().enumerate() { for (i, v) in element_val.into_iter().enumerate() {
unsafe { unsafe {
let ptr = self.builder.build_in_bounds_gep( let ptr = ctx.builder.build_in_bounds_gep(
tuple_ptr, tuple_ptr,
&[zero, self.ctx.i32_type().const_int(i as u64, false)], &[zero, ctx.ctx.i32_type().const_int(i as u64, false)],
"ptr", "ptr",
); );
self.builder.build_store(ptr, v); ctx.builder.build_store(ptr, v);
} }
} }
tuple_ptr.into() tuple_ptr.into()
} }
ExprKind::Attribute { value, attr, .. } => { ExprKind::Attribute { value, attr, .. } => {
// note that we would handle class methods directly in calls // note that we would handle class methods directly in calls
let index = self.get_attr_index(value.custom.unwrap(), *attr); let index = ctx.get_attr_index(value.custom.unwrap(), *attr);
let val = self.gen_expr(value).unwrap(); let val = generator.gen_expr(ctx, value).unwrap();
let ptr = assert_pointer_val(val); let ptr = assert_pointer_val(val);
unsafe { unsafe {
let ptr = self.builder.build_in_bounds_gep( let ptr = ctx.builder.build_in_bounds_gep(
ptr, ptr,
&[zero, self.ctx.i32_type().const_int(index as u64, false)], &[zero, ctx.ctx.i32_type().const_int(index as u64, false)],
"attr", "attr",
); );
self.builder.build_load(ptr, "field") ctx.builder.build_load(ptr, "field")
} }
} }
ExprKind::BoolOp { op, values } => { ExprKind::BoolOp { op, values } => {
// requires conditional branches for short-circuiting... // requires conditional branches for short-circuiting...
let left = assert_int_val(self.gen_expr(&values[0]).unwrap()); let left = assert_int_val(generator.gen_expr(ctx, &values[0]).unwrap());
let current = self.builder.get_insert_block().unwrap().get_parent().unwrap(); let current = ctx.builder.get_insert_block().unwrap().get_parent().unwrap();
let a_bb = self.ctx.append_basic_block(current, "a"); let a_bb = ctx.ctx.append_basic_block(current, "a");
let b_bb = self.ctx.append_basic_block(current, "b"); let b_bb = ctx.ctx.append_basic_block(current, "b");
let cont_bb = self.ctx.append_basic_block(current, "cont"); let cont_bb = ctx.ctx.append_basic_block(current, "cont");
self.builder.build_conditional_branch(left, a_bb, b_bb); ctx.builder.build_conditional_branch(left, a_bb, b_bb);
let (a, b) = match op { let (a, b) = match op {
Boolop::Or => { Boolop::Or => {
self.builder.position_at_end(a_bb); ctx.builder.position_at_end(a_bb);
let a = self.ctx.bool_type().const_int(1, false); let a = ctx.ctx.bool_type().const_int(1, false);
self.builder.build_unconditional_branch(cont_bb); ctx.builder.build_unconditional_branch(cont_bb);
self.builder.position_at_end(b_bb); ctx.builder.position_at_end(b_bb);
let b = assert_int_val(self.gen_expr(&values[1]).unwrap()); let b = assert_int_val(generator.gen_expr(ctx, &values[1]).unwrap());
self.builder.build_unconditional_branch(cont_bb); ctx.builder.build_unconditional_branch(cont_bb);
(a, b) (a, b)
} }
Boolop::And => { Boolop::And => {
self.builder.position_at_end(a_bb); ctx.builder.position_at_end(a_bb);
let a = assert_int_val(self.gen_expr(&values[1]).unwrap()); let a = assert_int_val(generator.gen_expr(ctx, &values[1]).unwrap());
self.builder.build_unconditional_branch(cont_bb); ctx.builder.build_unconditional_branch(cont_bb);
self.builder.position_at_end(b_bb); ctx.builder.position_at_end(b_bb);
let b = self.ctx.bool_type().const_int(0, false); let b = ctx.ctx.bool_type().const_int(0, false);
self.builder.build_unconditional_branch(cont_bb); ctx.builder.build_unconditional_branch(cont_bb);
(a, b) (a, b)
} }
}; };
self.builder.position_at_end(cont_bb); ctx.builder.position_at_end(cont_bb);
let phi = self.builder.build_phi(self.ctx.bool_type(), "phi"); let phi = ctx.builder.build_phi(ctx.ctx.bool_type(), "phi");
phi.add_incoming(&[(&a, a_bb), (&b, b_bb)]); phi.add_incoming(&[(&a, a_bb), (&b, b_bb)]);
phi.as_basic_value() phi.as_basic_value()
} }
ExprKind::BinOp { op, left, right } => { ExprKind::BinOp { op, left, right } => {
let ty1 = self.unifier.get_representative(left.custom.unwrap()); let ty1 = ctx.unifier.get_representative(left.custom.unwrap());
let ty2 = self.unifier.get_representative(right.custom.unwrap()); let ty2 = ctx.unifier.get_representative(right.custom.unwrap());
let left = self.gen_expr(left).unwrap(); let left = generator.gen_expr(ctx, left).unwrap();
let right = self.gen_expr(right).unwrap(); let right = generator.gen_expr(ctx, right).unwrap();
// we can directly compare the types, because we've got their representatives // we can directly compare the types, because we've got their representatives
// which would be unchanged until further unification, which we would never do // which would be unchanged until further unification, which we would never do
// when doing code generation for function instances // when doing code generation for function instances
if ty1 == ty2 && [self.primitives.int32, self.primitives.int64].contains(&ty1) { if ty1 == ty2 && [ctx.primitives.int32, ctx.primitives.int64].contains(&ty1) {
self.gen_int_ops(op, left, right) ctx.gen_int_ops(op, left, right)
} else if ty1 == ty2 && self.primitives.float == ty1 { } else if ty1 == ty2 && ctx.primitives.float == ty1 {
self.gen_float_ops(op, left, right) ctx.gen_float_ops(op, left, right)
} else { } else {
unimplemented!() unimplemented!()
} }
} }
ExprKind::UnaryOp { op, operand } => { ExprKind::UnaryOp { op, operand } => {
let ty = self.unifier.get_representative(operand.custom.unwrap()); let ty = ctx.unifier.get_representative(operand.custom.unwrap());
let val = self.gen_expr(operand).unwrap(); let val = generator.gen_expr(ctx, operand).unwrap();
if ty == self.primitives.bool { if ty == ctx.primitives.bool {
let val = assert_int_val(val); let val = assert_int_val(val);
match op { match op {
ast::Unaryop::Invert | ast::Unaryop::Not => { ast::Unaryop::Invert | ast::Unaryop::Not => {
self.builder.build_not(val, "not").into() ctx.builder.build_not(val, "not").into()
} }
_ => val.into(), _ => val.into(),
} }
} else if [self.primitives.int32, self.primitives.int64].contains(&ty) { } else if [ctx.primitives.int32, ctx.primitives.int64].contains(&ty) {
let val = assert_int_val(val); let val = assert_int_val(val);
match op { match op {
ast::Unaryop::USub => self.builder.build_int_neg(val, "neg").into(), ast::Unaryop::USub => ctx.builder.build_int_neg(val, "neg").into(),
ast::Unaryop::Invert => self.builder.build_not(val, "not").into(), ast::Unaryop::Invert => ctx.builder.build_not(val, "not").into(),
ast::Unaryop::Not => self ast::Unaryop::Not => ctx
.builder .builder
.build_int_compare( .build_int_compare(
inkwell::IntPredicate::EQ, inkwell::IntPredicate::EQ,
@ -547,15 +565,12 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
.into(), .into(),
_ => val.into(), _ => val.into(),
} }
} else if ty == self.primitives.float { } else if ty == ctx.primitives.float {
let val = if let BasicValueEnum::FloatValue(val) = val { let val =
val if let BasicValueEnum::FloatValue(val) = val { val } else { unreachable!() };
} else {
unreachable!()
};
match op { match op {
ast::Unaryop::USub => self.builder.build_float_neg(val, "neg").into(), ast::Unaryop::USub => ctx.builder.build_float_neg(val, "neg").into(),
ast::Unaryop::Not => self ast::Unaryop::Not => ctx
.builder .builder
.build_float_compare( .build_float_compare(
inkwell::FloatPredicate::OEQ, inkwell::FloatPredicate::OEQ,
@ -571,23 +586,20 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
} }
} }
ExprKind::Compare { left, ops, comparators } => { ExprKind::Compare { left, ops, comparators } => {
izip!( izip!(chain(once(left.as_ref()), comparators.iter()), comparators.iter(), ops.iter(),)
chain(once(left.as_ref()), comparators.iter()),
comparators.iter(),
ops.iter(),
)
.fold(None, |prev, (lhs, rhs, op)| { .fold(None, |prev, (lhs, rhs, op)| {
let ty = self.unifier.get_representative(lhs.custom.unwrap()); let ty = ctx.unifier.get_representative(lhs.custom.unwrap());
let current = let current =
if [self.primitives.int32, self.primitives.int64, self.primitives.bool] if [ctx.primitives.int32, ctx.primitives.int64, ctx.primitives.bool]
.contains(&ty) .contains(&ty)
{ {
let (lhs, rhs) = if let ( let (lhs, rhs) = if let (
BasicValueEnum::IntValue(lhs), BasicValueEnum::IntValue(lhs),
BasicValueEnum::IntValue(rhs), BasicValueEnum::IntValue(rhs),
) = ) = (
(self.gen_expr(lhs).unwrap(), self.gen_expr(rhs).unwrap()) generator.gen_expr(ctx, lhs).unwrap(),
{ generator.gen_expr(ctx, rhs).unwrap(),
) {
(lhs, rhs) (lhs, rhs)
} else { } else {
unreachable!() unreachable!()
@ -601,14 +613,15 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
ast::Cmpop::GtE => inkwell::IntPredicate::SGE, ast::Cmpop::GtE => inkwell::IntPredicate::SGE,
_ => unreachable!(), _ => unreachable!(),
}; };
self.builder.build_int_compare(op, lhs, rhs, "cmp") ctx.builder.build_int_compare(op, lhs, rhs, "cmp")
} else if ty == self.primitives.float { } else if ty == ctx.primitives.float {
let (lhs, rhs) = if let ( let (lhs, rhs) = if let (
BasicValueEnum::FloatValue(lhs), BasicValueEnum::FloatValue(lhs),
BasicValueEnum::FloatValue(rhs), BasicValueEnum::FloatValue(rhs),
) = ) = (
(self.gen_expr(lhs).unwrap(), self.gen_expr(rhs).unwrap()) generator.gen_expr(ctx, lhs).unwrap(),
{ generator.gen_expr(ctx, rhs).unwrap(),
) {
(lhs, rhs) (lhs, rhs)
} else { } else {
unreachable!() unreachable!()
@ -622,46 +635,49 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
ast::Cmpop::GtE => inkwell::FloatPredicate::OGE, ast::Cmpop::GtE => inkwell::FloatPredicate::OGE,
_ => unreachable!(), _ => unreachable!(),
}; };
self.builder.build_float_compare(op, lhs, rhs, "cmp") ctx.builder.build_float_compare(op, lhs, rhs, "cmp")
} else { } else {
unimplemented!() unimplemented!()
}; };
prev.map(|v| self.builder.build_and(v, current, "cmp")).or(Some(current)) prev.map(|v| ctx.builder.build_and(v, current, "cmp")).or(Some(current))
}) })
.unwrap() .unwrap()
.into() // as there should be at least 1 element, it should never be none .into() // as there should be at least 1 element, it should never be none
} }
ExprKind::IfExp { test, body, orelse } => { ExprKind::IfExp { test, body, orelse } => {
let test = assert_int_val(self.gen_expr(test).unwrap()); let test = assert_int_val(generator.gen_expr(ctx, test).unwrap());
let current = self.builder.get_insert_block().unwrap().get_parent().unwrap(); let current = ctx.builder.get_insert_block().unwrap().get_parent().unwrap();
let then_bb = self.ctx.append_basic_block(current, "then"); let then_bb = ctx.ctx.append_basic_block(current, "then");
let else_bb = self.ctx.append_basic_block(current, "else"); let else_bb = ctx.ctx.append_basic_block(current, "else");
let cont_bb = self.ctx.append_basic_block(current, "cont"); let cont_bb = ctx.ctx.append_basic_block(current, "cont");
self.builder.build_conditional_branch(test, then_bb, else_bb); ctx.builder.build_conditional_branch(test, then_bb, else_bb);
self.builder.position_at_end(then_bb); ctx.builder.position_at_end(then_bb);
let a = self.gen_expr(body).unwrap(); let a = generator.gen_expr(ctx, body).unwrap();
self.builder.build_unconditional_branch(cont_bb); ctx.builder.build_unconditional_branch(cont_bb);
self.builder.position_at_end(else_bb); ctx.builder.position_at_end(else_bb);
let b = self.gen_expr(orelse).unwrap(); let b = generator.gen_expr(ctx, orelse).unwrap();
self.builder.build_unconditional_branch(cont_bb); ctx.builder.build_unconditional_branch(cont_bb);
self.builder.position_at_end(cont_bb); ctx.builder.position_at_end(cont_bb);
let phi = self.builder.build_phi(a.get_type(), "ifexpr"); let phi = ctx.builder.build_phi(a.get_type(), "ifexpr");
phi.add_incoming(&[(&a, then_bb), (&b, else_bb)]); phi.add_incoming(&[(&a, then_bb), (&b, else_bb)]);
phi.as_basic_value() phi.as_basic_value()
} }
ExprKind::Call { func, args, keywords } => { ExprKind::Call { func, args, keywords } => {
let mut params = let mut params =
args.iter().map(|arg| (None, self.gen_expr(arg).unwrap())).collect_vec(); args.iter().map(|arg| (None, generator.gen_expr(ctx, arg).unwrap())).collect_vec();
let kw_iter = keywords.iter().map(|kw| { let kw_iter = keywords.iter().map(|kw| {
(Some(*kw.node.arg.as_ref().unwrap()), self.gen_expr(&kw.node.value).unwrap()) (
Some(*kw.node.arg.as_ref().unwrap()),
generator.gen_expr(ctx, &kw.node.value).unwrap(),
)
}); });
params.extend(kw_iter); params.extend(kw_iter);
let call = self.calls.get(&expr.location.into()); let call = ctx.calls.get(&expr.location.into());
let signature = match call { let signature = match call {
Some(call) => self.unifier.get_call_signature(*call).unwrap(), Some(call) => ctx.unifier.get_call_signature(*call).unwrap(),
None => { None => {
let ty = func.custom.unwrap(); let ty = func.custom.unwrap();
if let TypeEnum::TFunc(sign) = &*self.unifier.get_ty(ty) { if let TypeEnum::TFunc(sign) = &*ctx.unifier.get_ty(ty) {
sign.borrow().clone() sign.borrow().clone()
} else { } else {
unreachable!() unreachable!()
@ -671,21 +687,20 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
match &func.as_ref().node { match &func.as_ref().node {
ExprKind::Name { id, .. } => { ExprKind::Name { id, .. } => {
// TODO: handle primitive casts and function pointers // TODO: handle primitive casts and function pointers
let fun = let fun = ctx.resolver.get_identifier_def(*id).expect("Unknown identifier");
self.resolver.get_identifier_def(*id).expect("Unknown identifier"); return generator.gen_call(ctx, None, (&signature, fun), params);
return self.gen_call(None, (&signature, fun), params);
} }
ExprKind::Attribute { value, attr, .. } => { ExprKind::Attribute { value, attr, .. } => {
let val = self.gen_expr(value).unwrap(); let val = generator.gen_expr(ctx, value).unwrap();
let id = if let TypeEnum::TObj { obj_id, .. } = let id = if let TypeEnum::TObj { obj_id, .. } =
&*self.unifier.get_ty(value.custom.unwrap()) &*ctx.unifier.get_ty(value.custom.unwrap())
{ {
*obj_id *obj_id
} else { } else {
unreachable!() unreachable!()
}; };
let fun_id = { let fun_id = {
let defs = self.top_level.definitions.read(); let defs = ctx.top_level.definitions.read();
let obj_def = defs.get(id.0).unwrap().read(); let obj_def = defs.get(id.0).unwrap().read();
if let TopLevelDef::Class { methods, .. } = &*obj_def { if let TopLevelDef::Class { methods, .. } = &*obj_def {
let mut fun_id = None; let mut fun_id = None;
@ -699,7 +714,8 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
unreachable!() unreachable!()
} }
}; };
return self.gen_call( return generator.gen_call(
ctx,
Some((value.custom.unwrap(), val)), Some((value.custom.unwrap(), val)),
(&signature, fun_id), (&signature, fun_id),
params, params,
@ -709,41 +725,40 @@ impl<'ctx, 'a> CodeGenContext<'ctx, 'a> {
} }
} }
ExprKind::Subscript { value, slice, .. } => { ExprKind::Subscript { value, slice, .. } => {
if let TypeEnum::TList { .. } = &*self.unifier.get_ty(value.custom.unwrap()) { if let TypeEnum::TList { .. } = &*ctx.unifier.get_ty(value.custom.unwrap()) {
if let ExprKind::Slice { .. } = slice.node { if let ExprKind::Slice { .. } = slice.node {
unimplemented!() unimplemented!()
} else { } else {
// TODO: bound check // TODO: bound check
let i32_type = self.ctx.i32_type(); let i32_type = ctx.ctx.i32_type();
let v = assert_pointer_val(self.gen_expr(value).unwrap()); let v = assert_pointer_val(generator.gen_expr(ctx, value).unwrap());
let index = assert_int_val(self.gen_expr(slice).unwrap()); let index = assert_int_val(generator.gen_expr(ctx, slice).unwrap());
unsafe { unsafe {
let ptr_to_arr = self.builder.build_in_bounds_gep( let ptr_to_arr = ctx.builder.build_in_bounds_gep(
v, v,
&[i32_type.const_zero(), i32_type.const_int(1, false)], &[i32_type.const_zero(), i32_type.const_int(1, false)],
"ptr_to_arr", "ptr_to_arr",
); );
let arr_ptr = let arr_ptr =
assert_pointer_val(self.builder.build_load(ptr_to_arr, "loadptr")); assert_pointer_val(ctx.builder.build_load(ptr_to_arr, "loadptr"));
let ptr = self.builder.build_gep(arr_ptr, &[index], "loadarrgep"); let ptr = ctx.builder.build_gep(arr_ptr, &[index], "loadarrgep");
self.builder.build_load(ptr, "loadarr") ctx.builder.build_load(ptr, "loadarr")
} }
} }
} else { } else {
let i32_type = self.ctx.i32_type(); let i32_type = ctx.ctx.i32_type();
let v = assert_pointer_val(self.gen_expr(value).unwrap()); let v = assert_pointer_val(generator.gen_expr(ctx, value).unwrap());
let index = assert_int_val(self.gen_expr(slice).unwrap()); let index = assert_int_val(generator.gen_expr(ctx, slice).unwrap());
unsafe { unsafe {
let ptr_to_elem = self.builder.build_in_bounds_gep( let ptr_to_elem = ctx.builder.build_in_bounds_gep(
v, v,
&[i32_type.const_zero(), index], &[i32_type.const_zero(), index],
"ptr_to_elem", "ptr_to_elem",
); );
self.builder.build_load(ptr_to_elem, "loadelem") ctx.builder.build_load(ptr_to_elem, "loadelem")
} }
} }
} }
_ => unimplemented!(), _ => unimplemented!(),
}) })
}
} }

142
nac3core/src/codegen/generator.rs Normal file
View File

@ -0,0 +1,142 @@
use crate::{
codegen::{expr::*, stmt::*, CodeGenContext},
toplevel::{DefinitionId, TopLevelDef},
typecheck::typedef::{FunSignature, Type},
};
use inkwell::values::{BasicValueEnum, PointerValue};
use rustpython_parser::ast::{Expr, Stmt, StrRef};
pub trait CodeGenerator {
/// Return the module name for the code generator.
fn get_name(&self) -> &str;
/// Generate function call and returns the function return value.
/// - obj: Optional object for method call.
/// - fun: Function signature and definition ID.
/// - params: Function parameters. Note that this does not include the object even if the
/// function is a class method.
fn gen_call<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
obj: Option<(Type, BasicValueEnum<'ctx>)>,
fun: (&FunSignature, DefinitionId),
params: Vec<(Option<StrRef>, BasicValueEnum<'ctx>)>,
) -> Option<BasicValueEnum<'ctx>> {
gen_call(self, ctx, obj, fun, params)
}
/// Generate object constructor and returns the constructed object.
/// - signature: Function signature of the contructor.
/// - def: Class definition for the constructor class.
/// - params: Function parameters.
fn gen_constructor<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
signature: &FunSignature,
def: &TopLevelDef,
params: Vec<(Option<StrRef>, BasicValueEnum<'ctx>)>,
) -> BasicValueEnum<'ctx> {
gen_constructor(self, ctx, signature, def, params)
}
/// Generate a function instance.
/// - obj: Optional object for method call.
/// - fun: Function signature, definition ID and the substitution key.
/// - params: Function parameters. Note that this does not include the object even if the
/// function is a class method.
/// Note that this function should check if the function is generated in another thread (due to
/// possible race condition), see the default implementation for an example.
fn gen_func_instance<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
obj: Option<(Type, BasicValueEnum<'ctx>)>,
fun: (&FunSignature, &mut TopLevelDef, String),
) -> String {
gen_func_instance(ctx, obj, fun)
}
/// Generate the code for an expression.
fn gen_expr<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
expr: &Expr<Option<Type>>,
) -> Option<BasicValueEnum<'ctx>> {
gen_expr(self, ctx, expr)
}
/// Allocate memory for a variable and return a pointer pointing to it.
/// The default implementation places the allocations at the start of the function.
fn gen_var_alloc<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
ty: Type,
) -> PointerValue<'ctx> {
gen_var(ctx, ty)
}
/// Return a pointer pointing to the target of the expression.
fn gen_store_target<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
pattern: &Expr<Option<Type>>,
) -> PointerValue<'ctx> {
gen_store_target(self, ctx, pattern)
}
/// Generate code for an assignment expression.
fn gen_assign<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
target: &Expr<Option<Type>>,
value: BasicValueEnum<'ctx>,
) {
gen_assign(self, ctx, target, value)
}
/// Generate code for a while expression.
/// Return true if the while loop must early return
fn gen_while<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
stmt: &Stmt<Option<Type>>,
) -> bool {
gen_while(self, ctx, stmt);
false
}
/// Generate code for an if expression.
/// Return true if the statement must early return
fn gen_if<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
stmt: &Stmt<Option<Type>>,
) -> bool {
gen_if(self, ctx, stmt)
}
/// Generate code for a statement
/// Return true if the statement must early return
fn gen_stmt<'ctx, 'a>(
&mut self,
ctx: &mut CodeGenContext<'ctx, 'a>,
stmt: &Stmt<Option<Type>>,
) -> bool {
gen_stmt(self, ctx, stmt)
}
}
pub struct DefaultCodeGenerator {
name: String,
}
impl DefaultCodeGenerator {
pub fn new(name: String) -> DefaultCodeGenerator {
DefaultCodeGenerator { name }
}
}
impl CodeGenerator for DefaultCodeGenerator {
fn get_name(&self) -> &str {
&self.name
}
}

28
nac3core/src/codegen/mod.rs
View File

@ -27,11 +27,14 @@ use std::sync::{
use std::thread; use std::thread;
mod expr; mod expr;
mod generator;
mod stmt; mod stmt;
#[cfg(test)] #[cfg(test)]
mod test; mod test;
pub use generator::{CodeGenerator, DefaultCodeGenerator};
pub struct CodeGenContext<'ctx, 'a> { pub struct CodeGenContext<'ctx, 'a> {
pub ctx: &'ctx Context, pub ctx: &'ctx Context,
pub builder: Builder<'ctx>, pub builder: Builder<'ctx>,
@ -77,8 +80,8 @@ pub struct WorkerRegistry {
} }
impl WorkerRegistry { impl WorkerRegistry {
pub fn create_workers( pub fn create_workers<G: CodeGenerator + Send + 'static>(
names: &[&str], generators: Vec<Box<G>>,
top_level_ctx: Arc<TopLevelContext>, top_level_ctx: Arc<TopLevelContext>,
f: Arc<WithCall>, f: Arc<WithCall>,
) -> (Arc<WorkerRegistry>, Vec<thread::JoinHandle<()>>) { ) -> (Arc<WorkerRegistry>, Vec<thread::JoinHandle<()>>) {
@ -89,21 +92,20 @@ impl WorkerRegistry {
let registry = Arc::new(WorkerRegistry { let registry = Arc::new(WorkerRegistry {
sender: Arc::new(sender), sender: Arc::new(sender),
receiver: Arc::new(receiver), receiver: Arc::new(receiver),
thread_count: names.len(), thread_count: generators.len(),
panicked: AtomicBool::new(false), panicked: AtomicBool::new(false),
task_count, task_count,
wait_condvar, wait_condvar,
}); });
let mut handles = Vec::new(); let mut handles = Vec::new();
for name in names.iter() { for mut generator in generators.into_iter() {
let top_level_ctx = top_level_ctx.clone(); let top_level_ctx = top_level_ctx.clone();
let registry = registry.clone(); let registry = registry.clone();
let registry2 = registry.clone(); let registry2 = registry.clone();
let name = name.to_string();
let f = f.clone(); let f = f.clone();
let handle = thread::spawn(move || { let handle = thread::spawn(move || {
registry.worker_thread(name, top_level_ctx, f); registry.worker_thread(generator.as_mut(), top_level_ctx, f);
}); });
let handle = thread::spawn(move || { let handle = thread::spawn(move || {
if let Err(e) = handle.join() { if let Err(e) = handle.join() {
@ -156,18 +158,19 @@ impl WorkerRegistry {
self.sender.send(Some(task)).unwrap(); self.sender.send(Some(task)).unwrap();
} }
fn worker_thread( fn worker_thread<G: CodeGenerator>(
&self, &self,
module_name: String, generator: &mut G,
top_level_ctx: Arc<TopLevelContext>, top_level_ctx: Arc<TopLevelContext>,
f: Arc<WithCall>, f: Arc<WithCall>,
) { ) {
let context = Context::create(); let context = Context::create();
let mut builder = context.create_builder(); let mut builder = context.create_builder();
let mut module = context.create_module(&module_name); let mut module = context.create_module(generator.get_name());
while let Some(task) = self.receiver.recv().unwrap() { while let Some(task) = self.receiver.recv().unwrap() {
let result = gen_func(&context, self, builder, module, task, top_level_ctx.clone()); let result =
gen_func(&context, generator, self, builder, module, task, top_level_ctx.clone());
builder = result.0; builder = result.0;
module = result.1; module = result.1;
*self.task_count.lock() -= 1; *self.task_count.lock() -= 1;
@ -243,8 +246,9 @@ fn get_llvm_type<'ctx>(
}) })
} }
pub fn gen_func<'ctx>( pub fn gen_func<'ctx, G: CodeGenerator + ?Sized>(
context: &'ctx Context, context: &'ctx Context,
generator: &mut G,
registry: &WorkerRegistry, registry: &WorkerRegistry,
builder: Builder<'ctx>, builder: Builder<'ctx>,
module: Module<'ctx>, module: Module<'ctx>,
@ -351,7 +355,7 @@ pub fn gen_func<'ctx>(
let mut returned = false; let mut returned = false;
for stmt in task.body.iter() { for stmt in task.body.iter() {
returned = code_gen_context.gen_stmt(stmt); returned = generator.gen_stmt(&mut code_gen_context, stmt);
if returned { if returned {
break; break;
} }

311
nac3core/src/codegen/stmt.rs
View File

@ -1,221 +1,250 @@
use super::{ use super::{
expr::{assert_int_val, assert_pointer_val}, expr::{assert_int_val, assert_pointer_val},
CodeGenContext, CodeGenContext, CodeGenerator,
}; };
use crate::typecheck::typedef::Type; use crate::typecheck::typedef::Type;
use inkwell::values::{BasicValue, BasicValueEnum, PointerValue}; use inkwell::values::{BasicValue, BasicValueEnum, PointerValue};
use rustpython_parser::ast::{Expr, ExprKind, Stmt, StmtKind}; use rustpython_parser::ast::{Expr, ExprKind, Stmt, StmtKind};
impl<'ctx, 'a> CodeGenContext<'ctx, 'a> { pub fn gen_var<'ctx, 'a>(ctx: &mut CodeGenContext<'ctx, 'a>, ty: Type) -> PointerValue<'ctx> {
fn gen_var(&mut self, ty: Type) -> PointerValue<'ctx> {
// put the alloca in init block // put the alloca in init block
let current = self.builder.get_insert_block().unwrap(); let current = ctx.builder.get_insert_block().unwrap();
// position before the last branching instruction... // position before the last branching instruction...
self.builder.position_before(&self.init_bb.get_last_instruction().unwrap()); ctx.builder.position_before(&ctx.init_bb.get_last_instruction().unwrap());
let ty = self.get_llvm_type(ty); let ty = ctx.get_llvm_type(ty);
let ptr = self.builder.build_alloca(ty, "tmp"); let ptr = ctx.builder.build_alloca(ty, "tmp");
self.builder.position_at_end(current); ctx.builder.position_at_end(current);
ptr ptr
} }
fn parse_pattern(&mut self, pattern: &Expr<Option<Type>>) -> PointerValue<'ctx> { pub fn gen_store_target<'ctx, 'a, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
pattern: &Expr<Option<Type>>,
) -> PointerValue<'ctx> {
// very similar to gen_expr, but we don't do an extra load at the end // very similar to gen_expr, but we don't do an extra load at the end
// and we flatten nested tuples // and we flatten nested tuples
match &pattern.node { match &pattern.node {
ExprKind::Name { id, .. } => { ExprKind::Name { id, .. } => ctx.var_assignment.get(id).cloned().unwrap_or_else(|| {
self.var_assignment.get(id).cloned().unwrap_or_else(|| { let ptr = generator.gen_var_alloc(ctx, pattern.custom.unwrap());
let ptr = self.gen_var(pattern.custom.unwrap()); ctx.var_assignment.insert(*id, ptr);
self.var_assignment.insert(*id, ptr);
ptr ptr
}) }),
}
ExprKind::Attribute { value, attr, .. } => { ExprKind::Attribute { value, attr, .. } => {
let index = self.get_attr_index(value.custom.unwrap(), *attr); let index = ctx.get_attr_index(value.custom.unwrap(), *attr);
let val = self.gen_expr(value).unwrap(); let val = generator.gen_expr(ctx, value).unwrap();
let ptr = if let BasicValueEnum::PointerValue(v) = val { let ptr = if let BasicValueEnum::PointerValue(v) = val {
v v
} else { } else {
unreachable!(); unreachable!();
}; };
unsafe { unsafe {
self.builder.build_in_bounds_gep( ctx.builder.build_in_bounds_gep(
ptr, ptr,
&[ &[
self.ctx.i32_type().const_zero(), ctx.ctx.i32_type().const_zero(),
self.ctx.i32_type().const_int(index as u64, false), ctx.ctx.i32_type().const_int(index as u64, false),
], ],
"attr", "attr",
) )
} }
} }
ExprKind::Subscript { value, slice, .. } => { ExprKind::Subscript { value, slice, .. } => {
let i32_type = self.ctx.i32_type(); let i32_type = ctx.ctx.i32_type();
let v = assert_pointer_val(self.gen_expr(value).unwrap()); let v = assert_pointer_val(generator.gen_expr(ctx, value).unwrap());
let index = assert_int_val(self.gen_expr(slice).unwrap()); let index = assert_int_val(generator.gen_expr(ctx, slice).unwrap());
unsafe { unsafe {
let ptr_to_arr = self.builder.build_in_bounds_gep( let ptr_to_arr = ctx.builder.build_in_bounds_gep(
v, v,
&[i32_type.const_zero(), i32_type.const_int(1, false)], &[i32_type.const_zero(), i32_type.const_int(1, false)],
"ptr_to_arr", "ptr_to_arr",
); );
let arr_ptr = let arr_ptr = assert_pointer_val(ctx.builder.build_load(ptr_to_arr, "loadptr"));
assert_pointer_val(self.builder.build_load(ptr_to_arr, "loadptr")); ctx.builder.build_gep(arr_ptr, &[index], "loadarrgep")
self.builder.build_gep(arr_ptr, &[index], "loadarrgep")
} }
} }
_ => unreachable!(), _ => unreachable!(),
} }
} }
fn gen_assignment(&mut self, target: &Expr<Option<Type>>, value: BasicValueEnum<'ctx>) { pub fn gen_assign<'ctx, 'a, G: CodeGenerator + ?Sized>(
let i32_type = self.ctx.i32_type(); generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
target: &Expr<Option<Type>>,
value: BasicValueEnum<'ctx>,
) {
let i32_type = ctx.ctx.i32_type();
if let ExprKind::Tuple { elts, .. } = &target.node { if let ExprKind::Tuple { elts, .. } = &target.node {
if let BasicValueEnum::PointerValue(ptr) = value { if let BasicValueEnum::PointerValue(ptr) = value {
for (i, elt) in elts.iter().enumerate() { for (i, elt) in elts.iter().enumerate() {
unsafe { unsafe {
let t = self.builder.build_in_bounds_gep( let t = ctx.builder.build_in_bounds_gep(
ptr, ptr,
&[i32_type.const_zero(), i32_type.const_int(i as u64, false)], &[i32_type.const_zero(), i32_type.const_int(i as u64, false)],
"elem", "elem",
); );
let v = self.builder.build_load(t, "tmpload"); let v = ctx.builder.build_load(t, "tmpload");
self.gen_assignment(elt, v); generator.gen_assign(ctx, elt, v);
} }
} }
} else { } else {
unreachable!() unreachable!()
} }
} else { } else {
let ptr = self.parse_pattern(target); let ptr = generator.gen_store_target(ctx, target);
self.builder.build_store(ptr, value); ctx.builder.build_store(ptr, value);
}
} }
}
// return true if it contains terminator pub fn gen_while<'ctx, 'a, G: CodeGenerator + ?Sized>(
pub fn gen_stmt(&mut self, stmt: &Stmt<Option<Type>>) -> bool { generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
stmt: &Stmt<Option<Type>>,
) {
if let StmtKind::While { test, body, orelse } = &stmt.node {
let current = ctx.builder.get_insert_block().unwrap().get_parent().unwrap();
let test_bb = ctx.ctx.append_basic_block(current, "test");
let body_bb = ctx.ctx.append_basic_block(current, "body");
let cont_bb = ctx.ctx.append_basic_block(current, "cont");
// if there is no orelse, we just go to cont_bb
let orelse_bb =
if orelse.is_empty() { cont_bb } else { ctx.ctx.append_basic_block(current, "orelse") };
// store loop bb information and restore it later
let loop_bb = ctx.loop_bb.replace((test_bb, cont_bb));
ctx.builder.build_unconditional_branch(test_bb);
ctx.builder.position_at_end(test_bb);
let test = generator.gen_expr(ctx, test).unwrap();
if let BasicValueEnum::IntValue(test) = test {
ctx.builder.build_conditional_branch(test, body_bb, orelse_bb);
} else {
unreachable!()
};
ctx.builder.position_at_end(body_bb);
for stmt in body.iter() {
generator.gen_stmt(ctx, stmt);
}
ctx.builder.build_unconditional_branch(test_bb);
if !orelse.is_empty() {
ctx.builder.position_at_end(orelse_bb);
for stmt in orelse.iter() {
generator.gen_stmt(ctx, stmt);
}
ctx.builder.build_unconditional_branch(cont_bb);
}
ctx.builder.position_at_end(cont_bb);
ctx.loop_bb = loop_bb;
} else {
unreachable!()
}
}
pub fn gen_if<'ctx, 'a, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
stmt: &Stmt<Option<Type>>,
) -> bool {
if let StmtKind::If { test, body, orelse } = &stmt.node {
let current = ctx.builder.get_insert_block().unwrap().get_parent().unwrap();
let test_bb = ctx.ctx.append_basic_block(current, "test");
let body_bb = ctx.ctx.append_basic_block(current, "body");
let mut cont_bb = None;
// if there is no orelse, we just go to cont_bb
let orelse_bb = if orelse.is_empty() {
cont_bb = Some(ctx.ctx.append_basic_block(current, "cont"));
cont_bb.unwrap()
} else {
ctx.ctx.append_basic_block(current, "orelse")
};
ctx.builder.build_unconditional_branch(test_bb);
ctx.builder.position_at_end(test_bb);
let test = generator.gen_expr(ctx, test).unwrap();
if let BasicValueEnum::IntValue(test) = test {
ctx.builder.build_conditional_branch(test, body_bb, orelse_bb);
} else {
unreachable!()
};
ctx.builder.position_at_end(body_bb);
let mut exited = false;
for stmt in body.iter() {
exited = generator.gen_stmt(ctx, stmt);
if exited {
break;
}
}
if !exited {
if cont_bb.is_none() {
cont_bb = Some(ctx.ctx.append_basic_block(current, "cont"));
}
ctx.builder.build_unconditional_branch(cont_bb.unwrap());
}
let then_exited = exited;
let else_exited = if !orelse.is_empty() {
exited = false;
ctx.builder.position_at_end(orelse_bb);
for stmt in orelse.iter() {
exited = generator.gen_stmt(ctx, stmt);
if exited {
break;
}
}
if !exited {
if cont_bb.is_none() {
cont_bb = Some(ctx.ctx.append_basic_block(current, "cont"));
}
ctx.builder.build_unconditional_branch(cont_bb.unwrap());
}
exited
} else {
false
};
if let Some(cont_bb) = cont_bb {
ctx.builder.position_at_end(cont_bb);
}
then_exited && else_exited
} else {
unreachable!()
}
}
pub fn gen_stmt<'ctx, 'a, G: CodeGenerator + ?Sized>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
stmt: &Stmt<Option<Type>>,
) -> bool {
match &stmt.node { match &stmt.node {
StmtKind::Pass => {}, StmtKind::Pass => {}
StmtKind::Expr { value } => { StmtKind::Expr { value } => {
self.gen_expr(value); generator.gen_expr(ctx, value);
} }
StmtKind::Return { value } => { StmtKind::Return { value } => {
let value = value.as_ref().map(|v| self.gen_expr(v).unwrap()); let value = value.as_ref().map(|v| generator.gen_expr(ctx, v).unwrap());
let value = value.as_ref().map(|v| v as &dyn BasicValue); let value = value.as_ref().map(|v| v as &dyn BasicValue);
self.builder.build_return(value); ctx.builder.build_return(value);
return true; return true;
} }
StmtKind::AnnAssign { target, value, .. } => { StmtKind::AnnAssign { target, value, .. } => {
if let Some(value) = value { if let Some(value) = value {
let value = self.gen_expr(value).unwrap(); let value = generator.gen_expr(ctx, value).unwrap();
self.gen_assignment(target, value); generator.gen_assign(ctx, target, value);
} }
} }
StmtKind::Assign { targets, value, .. } => { StmtKind::Assign { targets, value, .. } => {
let value = self.gen_expr(value).unwrap(); let value = generator.gen_expr(ctx, value).unwrap();
for target in targets.iter() { for target in targets.iter() {
self.gen_assignment(target, value); generator.gen_assign(ctx, target, value);
} }
} }
StmtKind::Continue => { StmtKind::Continue => {
self.builder.build_unconditional_branch(self.loop_bb.unwrap().0); ctx.builder.build_unconditional_branch(ctx.loop_bb.unwrap().0);
return true; return true;
} }
StmtKind::Break => { StmtKind::Break => {
self.builder.build_unconditional_branch(self.loop_bb.unwrap().1); ctx.builder.build_unconditional_branch(ctx.loop_bb.unwrap().1);
return true; return true;
} }
StmtKind::If { test, body, orelse } => { StmtKind::If { .. } => return generator.gen_if(ctx, stmt),
let current = self.builder.get_insert_block().unwrap().get_parent().unwrap(); StmtKind::While { .. } => return generator.gen_while(ctx, stmt),
let test_bb = self.ctx.append_basic_block(current, "test"); _ => unimplemented!()
let body_bb = self.ctx.append_basic_block(current, "body");
let mut cont_bb = None; // self.ctx.append_basic_block(current, "cont");
// if there is no orelse, we just go to cont_bb
let orelse_bb = if orelse.is_empty() {
cont_bb = Some(self.ctx.append_basic_block(current, "cont"));
cont_bb.unwrap()
} else {
self.ctx.append_basic_block(current, "orelse")
};
self.builder.build_unconditional_branch(test_bb);
self.builder.position_at_end(test_bb);
let test = self.gen_expr(test).unwrap();
if let BasicValueEnum::IntValue(test) = test {
self.builder.build_conditional_branch(test, body_bb, orelse_bb);
} else {
unreachable!()
};
self.builder.position_at_end(body_bb);
let mut exited = false;
for stmt in body.iter() {
exited = self.gen_stmt(stmt);
if exited {
break;
}
}
if !exited {
if cont_bb.is_none() {
cont_bb = Some(self.ctx.append_basic_block(current, "cont"));
}
self.builder.build_unconditional_branch(cont_bb.unwrap());
}
if !orelse.is_empty() {
exited = false;
self.builder.position_at_end(orelse_bb);
for stmt in orelse.iter() {
exited = self.gen_stmt(stmt);
if exited {
break;
}
}
if !exited {
if cont_bb.is_none() {
cont_bb = Some(self.ctx.append_basic_block(current, "cont"));
}
self.builder.build_unconditional_branch(cont_bb.unwrap());
}
}
if let Some(cont_bb) = cont_bb {
self.builder.position_at_end(cont_bb);
}
}
StmtKind::While { test, body, orelse } => {
let current = self.builder.get_insert_block().unwrap().get_parent().unwrap();
let test_bb = self.ctx.append_basic_block(current, "test");
let body_bb = self.ctx.append_basic_block(current, "body");
let cont_bb = self.ctx.append_basic_block(current, "cont");
// if there is no orelse, we just go to cont_bb
let orelse_bb = if orelse.is_empty() {
cont_bb
} else {
self.ctx.append_basic_block(current, "orelse")
};
// store loop bb information and restore it later
let loop_bb = self.loop_bb.replace((test_bb, cont_bb));
self.builder.build_unconditional_branch(test_bb);
self.builder.position_at_end(test_bb);
let test = self.gen_expr(test).unwrap();
if let BasicValueEnum::IntValue(test) = test {
self.builder.build_conditional_branch(test, body_bb, orelse_bb);
} else {
unreachable!()
};
self.builder.position_at_end(body_bb);
for stmt in body.iter() {
self.gen_stmt(stmt);
}
self.builder.build_unconditional_branch(test_bb);
if !orelse.is_empty() {
self.builder.position_at_end(orelse_bb);
for stmt in orelse.iter() {
self.gen_stmt(stmt);
}
self.builder.build_unconditional_branch(cont_bb);
}
self.builder.position_at_end(cont_bb);
self.loop_bb = loop_bb;
}
_ => unimplemented!("{:?}", stmt),
}; };
false false
}
} }

10
nac3core/src/codegen/test.rs
View File

@ -1,5 +1,5 @@
use crate::{ use crate::{
codegen::{CodeGenTask, WithCall, WorkerRegistry, CodeGenContext}, codegen::{CodeGenTask, WithCall, WorkerRegistry, CodeGenContext, DefaultCodeGenerator},
location::Location, location::Location,
symbol_resolver::SymbolResolver, symbol_resolver::SymbolResolver,
toplevel::{ toplevel::{
@ -72,7 +72,7 @@ fn test_primitives() {
class_names: Default::default(), class_names: Default::default(),
}) as Arc<dyn SymbolResolver + Send + Sync>; }) as Arc<dyn SymbolResolver + Send + Sync>;
let threads = ["test"]; let threads = vec![DefaultCodeGenerator::new("test".into()).into()];
let signature = FunSignature { let signature = FunSignature {
args: vec![ args: vec![
FuncArg { name: "a".into(), ty: primitives.int32, default_value: None }, FuncArg { name: "a".into(), ty: primitives.int32, default_value: None },
@ -186,7 +186,7 @@ fn test_primitives() {
.trim(); .trim();
assert_eq!(expected, module.print_to_string().to_str().unwrap().trim()); assert_eq!(expected, module.print_to_string().to_str().unwrap().trim());
}))); })));
let (registry, handles) = WorkerRegistry::create_workers(&threads, top_level, f); let (registry, handles) = WorkerRegistry::create_workers(threads, top_level, f);
registry.add_task(task); registry.add_task(task);
registry.wait_tasks_complete(handles); registry.wait_tasks_complete(handles);
} }
@ -245,7 +245,7 @@ fn test_simple_call() {
unreachable!() unreachable!()
} }
let threads = ["test"]; let threads = vec![DefaultCodeGenerator::new("test".into()).into()];
let mut function_data = FunctionData { let mut function_data = FunctionData {
resolver: resolver.clone(), resolver: resolver.clone(),
bound_variables: Vec::new(), bound_variables: Vec::new(),
@ -351,7 +351,7 @@ fn test_simple_call() {
.trim(); .trim();
assert_eq!(expected, module.print_to_string().to_str().unwrap().trim()); assert_eq!(expected, module.print_to_string().to_str().unwrap().trim());
}))); })));
let (registry, handles) = WorkerRegistry::create_workers(&threads, top_level, f); let (registry, handles) = WorkerRegistry::create_workers(threads, top_level, f);
registry.add_task(task); registry.add_task(task);
registry.wait_tasks_complete(handles); registry.wait_tasks_complete(handles);
} }

78
nac3standalone/src/main.rs
View File

@ -1,12 +1,16 @@
use std::env; use inkwell::{
use std::fs; passes::{PassManager, PassManagerBuilder},
use inkwell::{OptimizationLevel, passes::{PassManager, PassManagerBuilder}, targets::*}; targets::*,
OptimizationLevel,
};
use nac3core::typecheck::type_inferencer::PrimitiveStore; use nac3core::typecheck::type_inferencer::PrimitiveStore;
use rustpython_parser::parser; use rustpython_parser::parser;
use std::env;
use std::fs;
use std::{collections::HashMap, path::Path, sync::Arc, time::SystemTime}; use std::{collections::HashMap, path::Path, sync::Arc, time::SystemTime};
use nac3core::{ use nac3core::{
codegen::{CodeGenTask, WithCall, WorkerRegistry}, codegen::{CodeGenTask, DefaultCodeGenerator, WithCall, WorkerRegistry},
symbol_resolver::SymbolResolver, symbol_resolver::SymbolResolver,
toplevel::{composer::TopLevelComposer, TopLevelDef}, toplevel::{composer::TopLevelComposer, TopLevelDef},
typecheck::typedef::FunSignature, typecheck::typedef::FunSignature,
@ -17,7 +21,10 @@ use basic_symbol_resolver::*;
fn main() { fn main() {
let demo_name = env::args().nth(1).unwrap(); let demo_name = env::args().nth(1).unwrap();
let threads: u32 = env::args().nth(2).map(|s| str::parse(&s).unwrap()).unwrap_or(1); let threads: u32 = env::args()
.nth(2)
.map(|s| str::parse(&s).unwrap())
.unwrap_or(1);
let start = SystemTime::now(); let start = SystemTime::now();
@ -38,23 +45,27 @@ fn main() {
id_to_type: builtins_ty.into(), id_to_type: builtins_ty.into(),
id_to_def: builtins_def.into(), id_to_def: builtins_def.into(),
class_names: Default::default(), class_names: Default::default(),
}.into(); }
let resolver = Arc::new( .into();
Resolver(internal_resolver.clone()) let resolver =
) as Arc<dyn SymbolResolver + Send + Sync>; Arc::new(Resolver(internal_resolver.clone())) as Arc<dyn SymbolResolver + Send + Sync>;
let setup_time = SystemTime::now(); let setup_time = SystemTime::now();
println!("setup time: {}ms", setup_time.duration_since(start).unwrap().as_millis()); println!(
"setup time: {}ms",
setup_time.duration_since(start).unwrap().as_millis()
);
let parser_result = parser::parse_program(&program).unwrap(); let parser_result = parser::parse_program(&program).unwrap();
let parse_time = SystemTime::now(); let parse_time = SystemTime::now();
println!("parse time: {}ms", parse_time.duration_since(setup_time).unwrap().as_millis()); println!(
"parse time: {}ms",
parse_time.duration_since(setup_time).unwrap().as_millis()
);
for stmt in parser_result.into_iter() { for stmt in parser_result.into_iter() {
let (name, def_id, ty) = composer.register_top_level( let (name, def_id, ty) = composer
stmt, .register_top_level(stmt, Some(resolver.clone()), "__main__".into())
Some(resolver.clone()), .unwrap();
"__main__".into(),
).unwrap();
internal_resolver.add_id_def(name, def_id); internal_resolver.add_id_def(name, def_id);
if let Some(ty) = ty { if let Some(ty) = ty {
@ -64,7 +75,13 @@ fn main() {
composer.start_analysis(true).unwrap(); composer.start_analysis(true).unwrap();
let analysis_time = SystemTime::now(); let analysis_time = SystemTime::now();
println!("analysis time: {}ms", analysis_time.duration_since(parse_time).unwrap().as_millis()); println!(
"analysis time: {}ms",
analysis_time
.duration_since(parse_time)
.unwrap()
.as_millis()
);
let top_level = Arc::new(composer.make_top_level_context()); let top_level = Arc::new(composer.make_top_level_context());
@ -119,19 +136,32 @@ fn main() {
) )
.expect("couldn't create target machine"); .expect("couldn't create target machine");
target_machine target_machine
.write_to_file(module, FileType::Object, Path::new(&format!("{}.o", module.get_name().to_str().unwrap()))) .write_to_file(
module,
FileType::Object,
Path::new(&format!("{}.o", module.get_name().to_str().unwrap())),
)
.expect("couldn't write module to file"); .expect("couldn't write module to file");
// println!("IR:\n{}", module.print_to_string().to_str().unwrap()); // println!("IR:\n{}", module.print_to_string().to_str().unwrap());
}))); })));
let threads: Vec<String> = (0..threads).map(|i| format!("module{}", i)).collect(); let threads = (0..threads)
let threads: Vec<_> = threads.iter().map(|s| s.as_str()).collect(); .map(|i| Box::new(DefaultCodeGenerator::new(format!("module{}", i))))
let (registry, handles) = WorkerRegistry::create_workers(&threads, top_level, f); .collect();
let (registry, handles) = WorkerRegistry::create_workers(threads, top_level, f);
registry.add_task(task); registry.add_task(task);
registry.wait_tasks_complete(handles); registry.wait_tasks_complete(handles);
let final_time = SystemTime::now(); let final_time = SystemTime::now();
println!("codegen time (including LLVM): {}ms", final_time.duration_since(analysis_time).unwrap().as_millis()); println!(
println!("total time: {}ms", final_time.duration_since(start).unwrap().as_millis()); "codegen time (including LLVM): {}ms",
final_time
.duration_since(analysis_time)
.unwrap()
.as_millis()
);
println!(
"total time: {}ms",
final_time.duration_since(start).unwrap().as_millis()
);
} }