My L compiler's toolchain is now complete, in that every necessary transformation pass is here. The various passes are parsing, macro-expansion, type checking and inference, CPS transformation, closure conversion, and compilation to LLVM instructions.
Most of the passes are still simple, and a lot of work remains to obtain something usable. For instance I do not propagate informations about locations, so typing error does not explain where the error is. All values, including integers, are boxed, allocated with malloc and never freed; and L code cannot call external C functions. The CPS transformations are not very efficient, and do not carry type informations. These are the points I am going to improve next.
However having a complete toolchain is nice: it gives a complete overview so now I know how changing a pass can benefit to both the above and below layers.
The nice thing about the passes being simple is that they are easy to understand, so this is a good opportunity to publish the code. To further improve the comprehension, I have decided for the last pass I wrote, which is the transformation from CPS to LLVM, to give a try at literate programming. It basically consists in writing your code in the manner of a text book.
There is a nice tool to do literate programming in ocaml, named
ocamlweb. It allows to write the literate parts in standard comments,
so that the Ocaml files can either be compiled or transformed into a
document. The default HTML output of ocamlweb (based on HEVEA is not
very nice however, but some configuration allows to improve it. Here
is is mine, that I put in a file heveaprefix.tex. This file changes
the HTML output of the code parts of OcamlWeb to look like the HTML
output of source code in Emacs Org-mode (to maintain consistency with
this blog).
%% Note: The colors code are those of Emacs org-mode output (which I %% think just put those of Emacs). %% This makes \url links as clickables urls. \input{urlhref.hva} %% Big code blocks. \renewcommand{\ocwbegincode}{% \begin{rawhtml} <div class="ocamlweb-src"><code> \end{rawhtml}} \renewcommand{\ocwendcode}{\begin{rawhtml}</code></div>\end{rawhtml}} %% Inline code blocks inside comments (given with []) \renewcommand{\ocwbegindcode}{\begin{rawhtml}<code>\end{rawhtml}} \renewcommand{\ocwenddcode}{\begin{rawhtml}</code>\end{rawhtml}} %% Keywords. We distinguish some keywords (those that ``create'' %% something, and begin, in blue). We rely on HEVEA native support for %% the ifthen package. \newcommand{\spanpurple}[1]{% \begin{rawhtml}<span style="color: #a020f0; ">\end{rawhtml}#1% \begin{rawhtml}</span>\end{rawhtml}} \newcommand{\spanred}[1]{% \begin{rawhtml}<span style="color: #a52a2a; ">\end{rawhtml}#1% \begin{rawhtml}</span>\end{rawhtml}} \newcommand{\spanboldblue}[1]{% \begin{rawhtml}<span style="color: #0000ff; font-weight: bold;">\end{rawhtml}#1% \begin{rawhtml}</span>\end{rawhtml}} \renewcommand{\ocwkw}[1]{% \ifthenelse{\equal{#1}{let}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{and}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{rec}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{in}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{type}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{of}}{\spanred{#1}}{% \ifthenelse{\equal{#1}{open}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{struct}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{sig}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{functor}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{module}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{val}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{begin}}{\spanboldblue{#1}}{% \ifthenelse{\equal{#1}{end}}{\spanboldblue{#1}}{% \spanpurple{#1}}}}}}}}}}}}}}}} %% Ids that begin in lower case. The textrm command (note: Hevea does %% not know about mathrm) allows non-italic typesetting. We also %% consider failwith as a keyword (even if it is a function that calls %% raise). \renewcommand{\ocwlowerid}[1]{% \ifthenelse{\equal{#1}{failwith}}{\spanpurple{\textrm{#1}}}{% \textrm{#1}}} %% Ids that begin in upper case. \newcommand{\spangreen}[1]{% \begin{rawhtml}<span style="color: #228b22; ">\end{rawhtml}#1% \begin{rawhtml}</span>\end{rawhtml}} \renewcommand{\ocwupperid}[1]{\spangreen{\textrm{#1}}} %% Comments are type set in red, with the leading (* and closing *). \renewcommand{\ocwbc}{\begin{rawhtml}<span style="color: #b22222">(∗ \end{rawhtml}} \renewcommand{\ocwec}{\begin{rawhtml} ∗)</span>\end{rawhtml}} %% Strings are typeset in brown. \newcommand{\spanbrown}[1]{% \begin{rawhtml}<span style="color: #8b2252; ">\end{rawhtml}#1% \begin{rawhtml}</span>\end{rawhtml}} \renewcommand{\ocwstring}[1]{\spanbrown{\textrm{#1}}} %% Base types and type variables are in green. \renewcommand{\ocwbt}[1]{\spangreen{\textrm{#1}}} \renewcommand{\ocwtv}[1]{\spangreen{#1e}}
The compilation command I use to perform the Ocaml to HTML transformation is:
ocamlweb -p "\usepackage{hevea}\usepackage{url}" --no-index \ heveaprefix.tex cps/cpsbase.ml llvm/cpsllvm.mli llvm/cpsllvm.ml > web/cpsllvm.tex \ && cd web && hevea -I /usr/share/texmf/tex/latex/misc ocamlweb.sty cpsllvm.tex
Below is the result (no need to explain it since this is literate programming! :))
Update: Apparently editing the post with blogger's editor mixes up the HTML, but this should be fixed now.
1. These definitions originates from the "compiling with continuations, continued" paper, by Andrew Kennedy (we currently use the simplified, non-graph version).
CPS (for continuation passing style) puts constraints on functional programs so that a function f never returns; instead it is passed a continuation k, which is a function that represents what is executed on f has finished its execution. So instead of returning a value x, f "returns" by calling k(x). CPS style makes returning from functions, and more generally control flow, explicit, at the expense or more verbosity.
This file presents a particular representation of CPS terms that
separates continuations, calling a continuations, variables holding
continations from respectively normal functions, normal function
calls, and normal variables. This distinction allows to compile the
CPS program using a stack (see the Cpsllvm module for an
implementation of that).
The representation also forces all values (including constants such as integers) to be held in variables, which simplify later transformation algorithms.
2.
We define variables and continuation variables a unique, to avoid
any need for alpha conversion.
module UniqueCPSVarId = Unique.Make(struct end)
module UniqueCPSContVarId = Unique.Make(struct end)
type var = Var of UniqueCPSVarId.t
type contvar = ContVar of UniqueCPSContVarId.t
module VarMap = Map.Make(struct
type t = var
let compare = compare
end)
module VarSet = Set.Make(struct
type t = var
let compare = compare
end)
module ContVarMap = Map.Make(struct
type t = contvar
let compare = compare
end)
module ContVarSet = Set.Make(struct
type t = contvar
let compare = compare
end)
type value =
∣ Void
∣ Constant of Constant.t
∣ Tuple of var list
∣ Lambda of contvar × var × term
x, which is the result of
the computation, to the caller.
and term =
∣ Let_value of var × value × term
∣ Let_primop of var × primitive_operation × term
∣ Let_cont of contvar × var × term × term
∣ Apply_cont of contvar × var
∣ Apply of var × contvar × var
∣ Halt of var
The various operations are:
Note that there are no primitive that would allow to write let x = y, where y is a variable; thus there cannot be two variables that directly share the same value.
and primitive_operation =
∣ Projection of var × int
∣ Integer_binary_op of Constant.integer_binary_op × var × var
6. This module translates CPS representation to the LLVM IR. CPS terms must observe that
7.
All translations are done using Llvm.global_context(), and in a
single Llvm module named the_module.
val the_module : Llvm.llmodule
build_nodef name expr builds an expr, an expression in CPS form
that is not part of a function, (for instance if it was typed in the
interactive prompt). It is translated to a Llvm function that take
no argument, named name.
val build_nodef : string → Cpsbase.term → Llvm.llvalue
9. This module translates a term written in CPS representation to LLVM instructions in SSA form.
The CPS representations stems from the paper "Compiling with continuations, continued" by Andrew Kennedy. In particular this representation separates continuations from standard lambda functions, which allows calling and returning from functions using the normal stack.
This module assumes that functions have no free variables (or continuation variables). Closure conversion removes free variables from functions. Free continuation variables should never happen when translating normal terms to CPS.
The module also assumes that the CPS values do not refer to primitive operations, such as +,-,*,/. Previous passes must transform calls to primitive operations to let x = primitive(args); and η-expand primitive operations passed as functions (e.g. let x = f() must have been transformed).
To keep things simple in this first version, no external functions is called (only lambdas defined in the body of the expression, and primitive operations, can be called).
In addition, all data is boxed, allocated using malloc (and never
freed; this could be improved by using libgc). Unboxed data would
requires to carry typing information in the CPS terms.
10.
To get an overview of the translation algorithm, the best is to
understand how the CPS concepts are mapped to the SSA concepts. In
the following, we denote by [x] the translation of x.
Llvm.llvalue). CPS variables are introduced as arguments to lambda
and continuations, and in the let x = ... form.Llvm.basicblock). The x formal argument of k corresponds to a
phi node at the start of [t1]. A call k( y to this continuation
with a value y is translated to a "jmp" instruction to the basic
block [t1], that binds [y] to the phi node at the start of
[t1].Note that the SSA representation are well-formed only if "the
definition of a variable %x does not dominate all of its uses"
(http://llvm.org/docs/LangRef.html#introduction). The translation
from a CPS term (without free variables) ensures that.
11.
Here is a simplified example of how the translation from CPS to
SSA works.
The CPS code:
let v = 3; let k(x) = k(2+x); k(11)
Is translated to SSA (ignoring boxing):
entry:
v = 3
n_ = 11
jmp k
k:
x = phi (entry n_) (k o_)
m_ = 2
o_ = m_ + x
jmp k This shows how k is translated to a separate basic block, and the argument x to a phi node connected to all the uses of k.
12.
If one encounters segmentation faults when changing the LLVM
related code, this may be caused by:
Llvm.build_call on a value which does not have the
function lltype, or Llvm.build_gep with operations that do not
correspond to the lltype of the value.build_phi with an empty list of "incoming".ExecutionEngine.create the_module before calling
Llvm_executionengine.initialize_native_target() can also segfault.Using valgrind or gdb allows to quickly locate the problematic Ocaml Llvm binding.
let context = Llvm.global_context()
let the_module = Llvm.create_module context "my jitted module"
let void_type = Llvm.void_type context
let i32_type = Llvm.i32_type context
let i32star_type = Llvm.pointer_type i32_type
13.
These helper functions create or read-from memory object.
Currently LLVM compiles using a very simple strategy: every value is
boxed (including integers and floats). This simplifies compilation a
lot: every value we create has type void *, and we cast the type
from void * according to how we use it.
LLVM does not (yet?) know how to replace heap allocations with stack
allocations, so we should do that (using an escape analysis). But
LLVM has passes that allow promotion of stack allocations to
register ("mem2reg" and "scalarrepl"), so once this is done (plus
passing and returning arguments in registers), many values should be
unboxed by the compiler (and this would not be that inefficient).
Additional performances could then be obtained by monomorphizing the
code.
14.
Store llvalue in heap-allocated memory.
let build_box llvalue name builder =
let lltype = Llvm.type_of llvalue in
let pointer = Llvm.build_malloc lltype name builder in
ignore(Llvm.build_store llvalue pointer builder);
Llvm.build_bitcast pointer anystar_type (name ^ "box") builder
llvalue of type lltype.
let build_unbox llvalue lltype name builder =
let typeptr = Llvm.pointer_type lltype in
let castedptr = Llvm.build_bitcast llvalue typeptr (name ^ "castedptr") builder in
Llvm.build_load castedptr (name ^ "unbox") builder
anystar_type. Each
element of the array contains the llvalue in l.
let build_tuple l builder =
let length = List.length l in
let array_type = Llvm.array_type anystar_type length in
let pointer = Llvm.build_malloc array_type "tuple" builder in
let f () (int,elem) =
(∗ Note: the first 0 is because pointer is not the start of
the array, but a pointer to the start of the array, that
must thus be dereferenced. ∗)
let path = [| (Llvm.const_int i32_type 0); (Llvm.const_int i32_type int) |] in
let gep_ptr = Llvm.build_gep pointer path "gep" builder in
ignore(Llvm.build_store elem gep_ptr builder) in
Utils.Int.fold_with_list f () (0,l);
Llvm.build_bitcast pointer anystar_type ("tuplecast") builder
let build_letproj pointer i builder =
let stringi = (string_of_int i) in
(∗ First we compute an acceptable LLvm type, and cast the pointer to
that type (failure to do that makes Llvm.build_gep segfault).
As we try to access the ith element, we assume we are accessing
an array of size i+1. ∗)
let array_type = Llvm.array_type anystar_type (i+1) in
let arraystar_type = Llvm.pointer_type array_type in
let cast_pointer = Llvm.build_bitcast pointer arraystar_type ("castptr") builder in
let gep_ptr = Llvm.build_gep cast_pointer [| (Llvm.const_int i32_type 0);
(Llvm.const_int i32_type i) |]
("gep" ^ stringi) builder in
let result = Llvm.build_load gep_ptr ("builder" ^ stringi) builder in
result
let build_integer_binary_op op a b builder =
let build_fn = match op with
∣ Constant.IAdd → Llvm.build_add
∣ Constant.ISub → Llvm.build_sub
∣ Constant.IMul → Llvm.build_mul
∣ Constant.IDiv → Llvm.build_udiv in
let a_unbox = (build_unbox a i32_type "a" builder) in
let b_unbox = (build_unbox b i32_type "b" builder) in
let res = build_fn a_unbox b_unbox "bop" builder in
build_box res "res" builder
caller to a function pointer.
let build_call caller callee builder =
let function_type = Llvm.pointer_type (Llvm.function_type anystar_type [| anystar_type |]) in
let casted_caller = Llvm.build_bitcast caller function_type "function" builder in
let retval = Llvm.build_call casted_caller [| callee |] "retval" builder in
retval
20.
This special value is used to ensure, via the type checker, that
compilation to LLVM never leaves a basic-block halfly built. LLVM
basic blocks should all end with a terminator instruction; whenever
one is inserted, the function should return End_of_block. When
building non-terminator instructions, the code must continue
building the basic block.
type termination = End_of_block
Note that LLVM basic blocks are associated to a parent function,
that we need to retrieve to create a new basic block.
let new_block builder =
let current_bb = Llvm.insertion_block builder in
let the_function = Llvm.block_parent current_bb in
let new_bb = Llvm.append_block context "k" the_function in
new_bb
Some(phi) if the block already begins with a phi instruction,
or None otherwise.
let begin_with_phi_node basic_block =
let pos = Llvm.instr_begin basic_block in
match pos with
∣ Llvm.At_end(_) → None
∣ Llvm.Before(inst) →
(match Llvm.instr_opcode inst with
∣ Llvm.Opcode.PHI → Some(inst)
∣ _ → None)
destination_block, also passing
the v value. This is achieved by setting v as an incoming value
for the phi instruction that begins destination_block. If
destination_block does not start with a phi node, then it is the
first time that destination_block is called, and we create this
phi node.
let build_jmp_to_and_add_incoming destination_block v builder =
let add_incoming_to_block basic_block (value,curblock) =
match begin_with_phi_node basic_block with
∣ Some(phi) → Llvm.add_incoming (value,curblock) phi
∣ None →
(∗ Temporarily create a builder to build the phi instruction. ∗)
let builder = Llvm.builder_at context (Llvm.instr_begin basic_block) in
ignore(Llvm.build_phi [value,curblock] "phi" builder) in
let current_basic_block = Llvm.insertion_block builder in
add_incoming_to_block destination_block (v, current_basic_block);
ignore(Llvm.build_br destination_block builder);
End_of_block
x of the call k( x)
is returned;x of
the call k( x) is passed to the phi node starting this basic
block.
The CPS→LLVM translation maps continuation variables to dest_types.
type dest_type =
∣ Ret
∣ Jmp_to of Llvm.llbasicblockk x.
let build_applycont k x builder =
match k with
∣ Ret → ignore(Llvm.build_ret x builder); End_of_block
∣ Jmp_to(destination) → build_jmp_to_and_add_incoming destination x builder
It is important for LLVM that function names are unique.
module UniqueFunctionId = Unique.Make(struct end)
cps, in the current block
pointed to by builder. varmap maps CPS variables to LLVM
llvalues. contvarmap maps CPS continuation variables to values of
type contvar_type.All the free variables or continuation variables in cps must be
in contvarmap or in varmap. cps can contain lambda, but they
must not contain any free variables or free continuation variables
(even the one in varmap and contvarmap). Closure conversion
deals with this. Note: previously-defined global variables are not
considered free.
let rec build_term cps (contvarmap, varmap) builder =
let lookup_var x =
try VarMap.find x varmap
with _ → failwith "in lookup" in
let lookup_contvar k =
try ContVarMap.find k contvarmap
with _ → failwith "in contvar lookup" in
let add_to_varmap var value = VarMap.add var value varmap in
let add_to_contvarmap contvar block = ContVarMap.add contvar (Jmp_to block) contvarmap in
To keep the implementation simple, all values are boxed (i.e. put
in the heap and accessed through a pointer), and of llvm type "i8
*". Pointer conversions are done according to the use of the
value.
match cps with
∣ Let_value(x, value, body) →
let newllvalue =
(match value with
∣ Constant(Constant.Int i) →
let llvalue = Llvm.const_int i32_type i in
build_box llvalue ("int" ^ string_of_int i) builder
∣ Tuple(l) →
let llvalues = List.map lookup_var l in
build_tuple llvalues builder
Note that build_function will use a new builder, so the
lambda can be built in parallel with the current
function. Also it will use new variables and continuation
variable maps (with only the x parameter), so the lambda
expression must not contain any free variables.
∣ Lambda(k,x,body) →
let f = build_function "lambda" k x body in
Llvm.set_linkage Llvm.Linkage.Private f;
Llvm.build_bitcast f anystar_type "lambdacast" builderlet x = primitive should have
been translated into something like let x = (a,b) ->
primitiveop( a,b) in previous compilation stage, so
should fail here.
∣ Constant(c) →
assert( Constant.is_function c);
failwith "ICE: primitive operations as value in LLVM translation."
)
in build_term body (contvarmap, (add_to_varmap x newllvalue)) builder
∣ Let_primop(x,prim,body) →
let result = (match prim with
∣ Integer_binary_op(op,xa,xb) →
build_integer_binary_op op (lookup_var xa) (lookup_var xb) builder
∣ Projection(x,i) → build_letproj (lookup_var x) i builder
) in
build_term body (contvarmap, (add_to_varmap x result)) builder
The tricky part is that the llvm bindings do not allow to create an "empty" phi node (even if it would, in future implementations which would not box everything we would still have to know the llvm type of the phi node, and that llvm type is not known until we have processed the jumps to that node). So it is the calls to k that create or change the phi node; no phi node means k is never called.
Doing the operations in this order ensures that calls to k are
processed before k is built.
∣ Let_cont(k,x,term,body) →
let new_bb = new_block builder in
let newcvm = add_to_contvarmap k new_bb in
let End_of_block = build_term body (newcvm, varmap) builder in
Llvm.position_at_end new_bb builder;
(match begin_with_phi_node new_bb with
∣ None → End_of_block
∣ Some(phi) → build_term term (newcvm, (add_to_varmap x phi)) builder)
∣ Apply_cont(k,x) →
build_applycont (lookup_contvar k) (lookup_var x) builder
∣ Apply(caller,k,callee) →
let retval = build_call (lookup_var caller) (lookup_var callee) builder in
build_applycont (lookup_contvar k) retval builder
∣ Halt(x) → ignore(Llvm.build_ret (lookup_var x) builder); End_of_block
void -> void" function.
and build_nodef name cpsbody =
prepare_build name cpsbody None
and build_function name contparam param cpsbody =
prepare_build name cpsbody (Some (contparam,param))
and prepare_build name cpsbody param =
let params_type = match param with None → [| |] ∣ _ → [| anystar_type |] in
let function_type = Llvm.function_type anystar_type params_type in
(∗ Note: it is important for LLVM that function names are unique. ∗)
let funname = name ^ "#" ^ (UniqueFunctionId.to_string (UniqueFunctionId.fresh())) in
let the_function = Llvm.declare_function funname function_type the_module in
let bb = Llvm.append_block context "entry" the_function in
(∗ Note that we use a new builder. We could even build the functions in parallel. ∗)
let builder = Llvm.builder context in
Llvm.position_at_end bb builder;
try
let initial_varmaps =
match param with
∣ None → (ContVarMap.empty, VarMap.empty)
∣ Some(k,x) → (ContVarMap.singleton k Ret,
VarMap.singleton x (Llvm.param the_function 0)) in
ignore(build_term cpsbody initial_varmaps builder);
(∗ Prints the textual representation of the function to stderr. ∗)
Llvm.dump_value the_function;
(∗ Validate the code we just generated. ∗)
Llvm_analysis.assert_valid_function the_function;
the_function
(∗ Normally, no exception should be thrown, be we never know. ∗)
with e → Llvm.delete_function the_function; raise e