Merge pull request #7601 from ethereum/binaryTransform

EWasm Binary transform

libyul/AssemblyStack.cpp

@@ -200,7 +200,7 @@ MachineAssemblyObject AssemblyStack::assemble(Machine _machine) const
 		Dialect const& dialect = languageToDialect(m_language, EVMVersion{});
 
 		MachineAssemblyObject object;
-		object.assembly = EWasmObjectCompiler::compile(*m_parserResult, dialect);
+		object.assembly = EWasmObjectCompiler::compile(*m_parserResult, dialect).first;
 		return object;
 	}
 	}

libyul/CMakeLists.txt

@@ -50,6 +50,8 @@ add_library(yul
 	backends/wasm/EWasmObjectCompiler.h
 	backends/wasm/EWasmToText.cpp
 	backends/wasm/EWasmToText.h
+	backends/wasm/BinaryTransform.cpp
+	backends/wasm/BinaryTransform.h
 	backends/wasm/WasmDialect.cpp
 	backends/wasm/WasmDialect.h
 	backends/wasm/WordSizeTransform.cpp

libyul/backends/wasm/BinaryTransform.cpp

@@ -0,0 +1,582 @@
+/*
+	This file is part of solidity.
+
+	solidity is free software: you can redistribute it and/or modify
+	it under the terms of the GNU General Public License as published by
+	the Free Software Foundation, either version 3 of the License, or
+	(at your option) any later version.
+
+	solidity is distributed in the hope that it will be useful,
+	but WITHOUT ANY WARRANTY; without even the implied warranty of
+	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+	GNU General Public License for more details.
+
+	You should have received a copy of the GNU General Public License
+	along with solidity.  If not, see <http://www.gnu.org/licenses/>.
+*/
+/**
+ * EWasm to binary encoder.
+ */
+
+#include <libyul/backends/wasm/BinaryTransform.h>
+
+#include <libyul/Exceptions.h>
+#include <libdevcore/CommonData.h>
+
+#include <boost/range/adaptor/reversed.hpp>
+
+using namespace std;
+using namespace yul;
+using namespace dev;
+using namespace yul::wasm;
+
+namespace
+{
+
+bytes toBytes(uint8_t _b)
+{
+	return bytes(1, _b);
+}
+
+enum class Section: uint8_t
+{
+	CUSTOM = 0x00,
+	TYPE = 0x01,
+	IMPORT = 0x02,
+	FUNCTION = 0x03,
+	MEMORY = 0x05,
+	GLOBAL = 0x06,
+	EXPORT = 0x07,
+	CODE = 0x0a
+};
+
+bytes toBytes(Section _s)
+{
+	return toBytes(uint8_t(_s));
+}
+
+enum class ValueType: uint8_t
+{
+	Void = 0x40,
+	Function = 0x60,
+	I64 = 0x7e,
+	I32 = 0x7f
+};
+
+bytes toBytes(ValueType _vt)
+{
+	return toBytes(uint8_t(_vt));
+}
+
+enum class Export: uint8_t
+{
+	Function = 0x0,
+	Memory = 0x2
+};
+
+bytes toBytes(Export _export)
+{
+	return toBytes(uint8_t(_export));
+}
+
+enum class Opcode: uint8_t
+{
+	Unreachable = 0x00,
+	Nop = 0x01,
+	Block = 0x02,
+	Loop = 0x03,
+	If = 0x04,
+	Else = 0x05,
+	Try = 0x06,
+	Catch = 0x07,
+	Throw = 0x08,
+	Rethrow = 0x09,
+	BrOnExn = 0x0a,
+	End = 0x0b,
+	Br = 0x0c,
+	BrIf = 0x0d,
+	BrTable = 0x0e,
+	Return = 0x0f,
+	Call = 0x10,
+	CallIndirect = 0x11,
+	ReturnCall = 0x12,
+	ReturnCallIndirect = 0x13,
+	Drop = 0x1a,
+	Select = 0x1b,
+	LocalGet = 0x20,
+	LocalSet = 0x21,
+	LocalTee = 0x22,
+	GlobalGet = 0x23,
+	GlobalSet = 0x24,
+	I32Const = 0x41,
+	I64Const = 0x42,
+};
+
+bytes toBytes(Opcode _o)
+{
+	return toBytes(uint8_t(_o));
+}
+
+static std::map<string, uint8_t> const builtins = {
+	{"i32.load", 0x28},
+	{"i64.load", 0x29},
+	{"i32.load8_s", 0x2c},
+	{"i32.load8_u", 0x2d},
+	{"i32.load16_s", 0x2e},
+	{"i32.load16_u", 0x2f},
+	{"i64.load8_s", 0x30},
+	{"i64.load8_u", 0x31},
+	{"i64.load16_s", 0x32},
+	{"i64.load16_u", 0x33},
+	{"i64.load32_s", 0x34},
+	{"i64.load32_u", 0x35},
+	{"i32.store", 0x36},
+	{"i64.store", 0x37},
+	{"i32.store8", 0x3a},
+	{"i32.store16", 0x3b},
+	{"i64.store8", 0x3c},
+	{"i64.store16", 0x3d},
+	{"i64.store32", 0x3e},
+	{"memory.size", 0x3f},
+	{"memory.grow", 0x40},
+	{"i32.eqz", 0x45},
+	{"i32.eq", 0x46},
+	{"i32.ne", 0x47},
+	{"i32.lt_s", 0x48},
+	{"i32.lt_u", 0x49},
+	{"i32.gt_s", 0x4a},
+	{"i32.gt_u", 0x4b},
+	{"i32.le_s", 0x4c},
+	{"i32.le_u", 0x4d},
+	{"i32.ge_s", 0x4e},
+	{"i32.ge_u", 0x4f},
+	{"i64.eqz", 0x50},
+	{"i64.eq", 0x51},
+	{"i64.ne", 0x52},
+	{"i64.lt_s", 0x53},
+	{"i64.lt_u", 0x54},
+	{"i64.gt_s", 0x55},
+	{"i64.gt_u", 0x56},
+	{"i64.le_s", 0x57},
+	{"i64.le_u", 0x58},
+	{"i64.ge_s", 0x59},
+	{"i64.ge_u", 0x5a},
+	{"i32.clz", 0x67},
+	{"i32.ctz", 0x68},
+	{"i32.popcnt", 0x69},
+	{"i32.add", 0x6a},
+	{"i32.sub", 0x6b},
+	{"i32.mul", 0x6c},
+	{"i32.div_s", 0x6d},
+	{"i32.div_u", 0x6e},
+	{"i32.rem_s", 0x6f},
+	{"i32.rem_u", 0x70},
+	{"i32.and", 0x71},
+	{"i32.or", 0x72},
+	{"i32.xor", 0x73},
+	{"i32.shl", 0x74},
+	{"i32.shr_s", 0x75},
+	{"i32.shr_u", 0x76},
+	{"i32.rotl", 0x77},
+	{"i32.rotr", 0x78},
+	{"i64.clz", 0x79},
+	{"i64.ctz", 0x7a},
+	{"i64.popcnt", 0x7b},
+	{"i64.add", 0x7c},
+	{"i64.sub", 0x7d},
+	{"i64.mul", 0x7e},
+	{"i64.div_s", 0x7f},
+	{"i64.div_u", 0x80},
+	{"i64.rem_s", 0x81},
+	{"i64.rem_u", 0x82},
+	{"i64.and", 0x83},
+	{"i64.or", 0x84},
+	{"i64.xor", 0x85},
+	{"i64.shl", 0x86},
+	{"i64.shr_s", 0x87},
+	{"i64.shr_u", 0x88},
+	{"i64.rotl", 0x89},
+	{"i64.rotr", 0x8a},
+	{"i32.wrap_i64", 0xa7},
+	{"i64.extend_i32_s", 0xac},
+	{"i64.extend_i32_u", 0xad},
+};
+
+bytes lebEncode(uint64_t _n)
+{
+	bytes encoded;
+	while (_n > 0x7f)
+	{
+		encoded.emplace_back(uint8_t(0x80 | (_n & 0x7f)));
+		_n >>= 7;
+	}
+	encoded.emplace_back(_n);
+	return encoded;
+}
+
+bytes lebEncodeSigned(int64_t _n)
+{
+	if (_n >= 0 && _n < 0x40)
+		return toBytes(uint8_t(uint64_t(_n) & 0xff));
+	else if (-_n > 0 && -_n < 0x40)
+		return toBytes(uint8_t(uint64_t(_n + 0x80) & 0xff));
+	else
+		return toBytes(uint8_t(0x80 | uint8_t(_n & 0x7f))) + lebEncodeSigned(_n / 0x80);
+}
+
+bytes prefixSize(bytes _data)
+{
+	size_t size = _data.size();
+	return lebEncode(size) + std::move(_data);
+}
+
+bytes makeSection(Section _section, bytes _data)
+{
+	return toBytes(_section) + prefixSize(std::move(_data));
+}
+
+}
+
+bytes BinaryTransform::run(Module const& _module)
+{
+	BinaryTransform bt;
+
+	for (size_t i = 0; i < _module.globals.size(); ++i)
+		bt.m_globals[_module.globals[i].variableName] = i;
+
+	size_t funID = 0;
+	for (FunctionImport const& fun: _module.imports)
+		bt.m_functions[fun.internalName] = funID++;
+	for (FunctionDefinition const& fun: _module.functions)
+		bt.m_functions[fun.name] = funID++;
+
+	bytes ret{0, 'a', 's', 'm'};
+	// version
+	ret += bytes{1, 0, 0, 0};
+	ret += bt.typeSection(_module.imports, _module.functions);
+	ret += bt.importSection(_module.imports);
+	ret += bt.functionSection(_module.functions);
+	ret += bt.memorySection();
+	ret += bt.globalSection();
+	ret += bt.exportSection();
+	for (auto const& sub: _module.subModules)
+	{
+		// TODO should we prefix and / or shorten the name?
+		bytes data = BinaryTransform::run(sub.second);
+		size_t length = data.size();
+		ret += bt.customSection(sub.first, std::move(data));
+		bt.m_subModulePosAndSize[sub.first] = {ret.size() - length, length};
+	}
+	ret += bt.codeSection(_module.functions);
+	return ret;
+}
+
+bytes BinaryTransform::operator()(Literal const& _literal)
+{
+	return toBytes(Opcode::I64Const) + lebEncodeSigned(_literal.value);
+}
+
+bytes BinaryTransform::operator()(StringLiteral const&)
+{
+	// TODO is this used?
+	yulAssert(false, "String literals not yet implemented");
+}
+
+bytes BinaryTransform::operator()(LocalVariable const& _variable)
+{
+	return toBytes(Opcode::LocalGet) + lebEncode(m_locals.at(_variable.name));
+}
+
+bytes BinaryTransform::operator()(GlobalVariable const& _variable)
+{
+	return toBytes(Opcode::GlobalGet) + lebEncode(m_globals.at(_variable.name));
+}
+
+bytes BinaryTransform::operator()(BuiltinCall const& _call)
+{
+	// We need to avoid visiting the arguments of `dataoffset` and `datasize` because
+	// they are references to object names that should not end up in the code.
+	if (_call.functionName == "dataoffset")
+	{
+		string name = boost::get<StringLiteral>(_call.arguments.at(0)).value;
+		return toBytes(Opcode::I64Const) + lebEncodeSigned(m_subModulePosAndSize.at(name).first);
+	}
+	else if (_call.functionName == "datasize")
+	{
+		string name = boost::get<StringLiteral>(_call.arguments.at(0)).value;
+		return toBytes(Opcode::I64Const) + lebEncodeSigned(m_subModulePosAndSize.at(name).second);
+	}
+
+	bytes args = visit(_call.arguments);
+
+	if (_call.functionName == "unreachable")
+			return toBytes(Opcode::Unreachable);
+	else
+	{
+		yulAssert(builtins.count(_call.functionName), "Builtin " + _call.functionName + " not found");
+		bytes ret = std::move(args) + toBytes(builtins.at(_call.functionName));
+		if (
+			_call.functionName.find(".load") != string::npos ||
+			_call.functionName.find(".store") != string::npos
+		)
+			// alignment and offset
+			ret += bytes{{3, 0}};
+		return ret;
+	}
+}
+
+bytes BinaryTransform::operator()(FunctionCall const& _call)
+{
+	return visit(_call.arguments) + toBytes(Opcode::Call) + lebEncode(m_functions.at(_call.functionName));
+}
+
+bytes BinaryTransform::operator()(LocalAssignment const& _assignment)
+{
+	return
+		boost::apply_visitor(*this, *_assignment.value) +
+		toBytes(Opcode::LocalSet) +
+		lebEncode(m_locals.at(_assignment.variableName));
+}
+
+bytes BinaryTransform::operator()(GlobalAssignment const& _assignment)
+{
+	return
+		boost::apply_visitor(*this, *_assignment.value) +
+		toBytes(Opcode::GlobalSet) +
+		lebEncode(m_globals.at(_assignment.variableName));
+}
+
+bytes BinaryTransform::operator()(If const& _if)
+{
+	bytes result =
+		boost::apply_visitor(*this, *_if.condition) +
+		toBytes(Opcode::If) +
+		toBytes(ValueType::Void);
+
+	m_labels.push({});
+
+	result += visit(_if.statements);
+	if (_if.elseStatements)
+		result += toBytes(Opcode::Else) + visit(*_if.elseStatements);
+
+	m_labels.pop();
+
+	result += toBytes(Opcode::End);
+	return result;
+}
+
+bytes BinaryTransform::operator()(Loop const& _loop)
+{
+	bytes result = toBytes(Opcode::Loop) + toBytes(ValueType::Void);
+
+	m_labels.push(_loop.labelName);
+	result += visit(_loop.statements);
+	m_labels.pop();
+
+	result += toBytes(Opcode::End);
+	return result;
+}
+
+bytes BinaryTransform::operator()(Break const&)
+{
+	yulAssert(false, "br not yet implemented.");
+	// TODO the index is just the nesting depth.
+	return {};
+}
+
+bytes BinaryTransform::operator()(BreakIf const&)
+{
+	yulAssert(false, "br_if not yet implemented.");
+	// TODO the index is just the nesting depth.
+	return {};
+}
+
+bytes BinaryTransform::operator()(Block const& _block)
+{
+	return
+		toBytes(Opcode::Block) +
+		toBytes(ValueType::Void) +
+		visit(_block.statements) +
+		toBytes(Opcode::End);
+}
+
+bytes BinaryTransform::operator()(FunctionDefinition const& _function)
+{
+	bytes ret;
+
+	// This is a kind of run-length-encoding of local types. Has to be adapted once
+	// we have locals of different types.
+	ret += lebEncode(1); // number of locals groups
+	ret += lebEncode(_function.locals.size());
+	ret += toBytes(ValueType::I64);
+
+	m_locals.clear();
+	size_t varIdx = 0;
+	for (size_t i = 0; i < _function.parameterNames.size(); ++i)
+		m_locals[_function.parameterNames[i]] = varIdx++;
+	for (size_t i = 0; i < _function.locals.size(); ++i)
+		m_locals[_function.locals[i].variableName] = varIdx++;
+
+	ret += visit(_function.body);
+	ret += toBytes(Opcode::End);
+
+	return prefixSize(std::move(ret));
+}
+
+BinaryTransform::Type BinaryTransform::typeOf(FunctionImport const& _import)
+{
+	return {
+		encodeTypes(_import.paramTypes),
+		encodeTypes(_import.returnType ? vector<string>(1, *_import.returnType) : vector<string>())
+	};
+}
+
+BinaryTransform::Type BinaryTransform::typeOf(FunctionDefinition const& _funDef)
+{
+
+	return {
+		encodeTypes(vector<string>(_funDef.parameterNames.size(), "i64")),
+		encodeTypes(vector<string>(_funDef.returns ? 1 : 0, "i64"))
+	};
+}
+
+uint8_t BinaryTransform::encodeType(string const& _typeName)
+{
+	if (_typeName == "i32")
+		return uint8_t(ValueType::I32);
+	else if (_typeName == "i64")
+		return uint8_t(ValueType::I64);
+	else
+		yulAssert(false, "");
+	return 0;
+}
+
+vector<uint8_t> BinaryTransform::encodeTypes(vector<string> const& _typeNames)
+{
+	vector<uint8_t> result;
+	for (auto const& t: _typeNames)
+		result.emplace_back(encodeType(t));
+	return result;
+}
+
+bytes BinaryTransform::typeSection(
+	vector<FunctionImport> const& _imports,
+	vector<FunctionDefinition> const& _functions
+)
+{
+	map<Type, vector<string>> types;
+	for (auto const& import: _imports)
+		types[typeOf(import)].emplace_back(import.internalName);
+	for (auto const& fun: _functions)
+		types[typeOf(fun)].emplace_back(fun.name);
+
+	bytes result;
+	size_t index = 0;
+	for (auto const& [type, funNames]: types)
+	{
+		for (string const& name: funNames)
+			m_functionTypes[name] = index;
+		result += toBytes(ValueType::Function);
+		result += lebEncode(type.first.size()) + type.first;
+		result += lebEncode(type.second.size()) + type.second;
+
+		index++;
+	}
+
+	return makeSection(Section::TYPE, lebEncode(index) + std::move(result));
+}
+
+bytes BinaryTransform::importSection(
+	vector<FunctionImport> const& _imports
+)
+{
+	bytes result = lebEncode(_imports.size());
+	for (FunctionImport const& import: _imports)
+	{
+		uint8_t importKind = 0; // function
+		result +=
+			encodeName(import.module) +
+			encodeName(import.externalName) +
+			toBytes(importKind) +
+			lebEncode(m_functionTypes[import.internalName]);
+	}
+	return makeSection(Section::IMPORT, std::move(result));
+}
+
+bytes BinaryTransform::functionSection(vector<FunctionDefinition> const& _functions)
+{
+	bytes result = lebEncode(_functions.size());
+	for (auto const& fun: _functions)
+		result += lebEncode(m_functionTypes.at(fun.name));
+	return makeSection(Section::FUNCTION, std::move(result));
+}
+
+bytes BinaryTransform::memorySection()
+{
+	bytes result = lebEncode(1);
+	result.push_back(0); // flags
+	result.push_back(1); // initial
+	return makeSection(Section::MEMORY, std::move(result));
+}
+
+bytes BinaryTransform::globalSection()
+{
+	bytes result = lebEncode(m_globals.size());
+	for (size_t i = 0; i < m_globals.size(); ++i)
+		result +=
+			// mutable i64
+			bytes{uint8_t(ValueType::I64), 1} +
+			toBytes(Opcode::I64Const) +
+			lebEncodeSigned(0) +
+			toBytes(Opcode::End);
+
+	return makeSection(Section::GLOBAL, std::move(result));
+}
+
+bytes BinaryTransform::exportSection()
+{
+	bytes result = lebEncode(2);
+	result += encodeName("memory") + toBytes(Export::Memory) + lebEncode(0);
+	result += encodeName("main") + toBytes(Export::Function) + lebEncode(m_functions.at("main"));
+	return makeSection(Section::EXPORT, std::move(result));
+}
+
+bytes BinaryTransform::customSection(string const& _name, bytes _data)
+{
+	bytes result = encodeName(_name) + std::move(_data);
+	return makeSection(Section::CUSTOM, std::move(result));
+}
+
+bytes BinaryTransform::codeSection(vector<wasm::FunctionDefinition> const& _functions)
+{
+	bytes result = lebEncode(_functions.size());
+	for (FunctionDefinition const& fun: _functions)
+		result += (*this)(fun);
+	return makeSection(Section::CODE, std::move(result));
+}
+
+bytes BinaryTransform::visit(vector<Expression> const& _expressions)
+{
+	bytes result;
+	for (auto const& expr: _expressions)
+		result += boost::apply_visitor(*this, expr);
+	return result;
+}
+
+bytes BinaryTransform::visitReversed(vector<Expression> const& _expressions)
+{
+	bytes result;
+	for (auto const& expr: _expressions | boost::adaptors::reversed)
+		result += boost::apply_visitor(*this, expr);
+	return result;
+}
+
+bytes BinaryTransform::encodeName(std::string const& _name)
+{
+	// UTF-8 is allowed here by the Wasm spec, but since all names here should stem from
+	// Solidity or Yul identifiers or similar, non-ascii characters ending up here
+	// is a very bad sign.
+	for (char c: _name)
+		yulAssert(uint8_t(c) <= 0x7f, "Non-ascii character found.");
+	return lebEncode(_name.size()) + asBytes(_name);
+}

libyul/backends/wasm/BinaryTransform.h

@@ -0,0 +1,94 @@
+/*
+	This file is part of solidity.
+
+	solidity is free software: you can redistribute it and/or modify
+	it under the terms of the GNU General Public License as published by
+	the Free Software Foundation, either version 3 of the License, or
+	(at your option) any later version.
+
+	solidity is distributed in the hope that it will be useful,
+	but WITHOUT ANY WARRANTY; without even the implied warranty of
+	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+	GNU General Public License for more details.
+
+	You should have received a copy of the GNU General Public License
+	along with solidity.  If not, see <http://www.gnu.org/licenses/>.
+*/
+/**
+ * EWasm to binary encoder.
+ */
+
+#pragma once
+
+#include <libyul/backends/wasm/EWasmAST.h>
+
+#include <libdevcore/Common.h>
+
+#include <vector>
+#include <stack>
+
+namespace yul
+{
+namespace wasm
+{
+
+/**
+ * Web assembly to binary transform.
+ */
+class BinaryTransform: public boost::static_visitor<dev::bytes>
+{
+public:
+	static dev::bytes run(Module const& _module);
+
+	dev::bytes operator()(wasm::Literal const& _literal);
+	dev::bytes operator()(wasm::StringLiteral const& _literal);
+	dev::bytes operator()(wasm::LocalVariable const& _identifier);
+	dev::bytes operator()(wasm::GlobalVariable const& _identifier);
+	dev::bytes operator()(wasm::BuiltinCall const& _builinCall);
+	dev::bytes operator()(wasm::FunctionCall const& _functionCall);
+	dev::bytes operator()(wasm::LocalAssignment const& _assignment);
+	dev::bytes operator()(wasm::GlobalAssignment const& _assignment);
+	dev::bytes operator()(wasm::If const& _if);
+	dev::bytes operator()(wasm::Loop const& _loop);
+	dev::bytes operator()(wasm::Break const& _break);
+	dev::bytes operator()(wasm::BreakIf const& _break);
+	dev::bytes operator()(wasm::Block const& _block);
+	dev::bytes operator()(wasm::FunctionDefinition const& _function);
+
+private:
+	using Type = std::pair<std::vector<std::uint8_t>, std::vector<std::uint8_t>>;
+	static Type typeOf(wasm::FunctionImport const& _import);
+	static Type typeOf(wasm::FunctionDefinition const& _funDef);
+
+	static uint8_t encodeType(std::string const& _typeName);
+	static std::vector<uint8_t> encodeTypes(std::vector<std::string> const& _typeNames);
+	dev::bytes typeSection(
+		std::vector<wasm::FunctionImport> const& _imports,
+		std::vector<wasm::FunctionDefinition> const& _functions
+	);
+
+	dev::bytes importSection(std::vector<wasm::FunctionImport> const& _imports);
+	dev::bytes functionSection(std::vector<wasm::FunctionDefinition> const& _functions);
+	dev::bytes memorySection();
+	dev::bytes globalSection();
+	dev::bytes exportSection();
+	dev::bytes customSection(std::string const& _name, dev::bytes _data);
+	dev::bytes codeSection(std::vector<wasm::FunctionDefinition> const& _functions);
+
+	dev::bytes visit(std::vector<wasm::Expression> const& _expressions);
+	dev::bytes visitReversed(std::vector<wasm::Expression> const& _expressions);
+
+	static dev::bytes encodeName(std::string const& _name);
+
+	std::map<std::string, size_t> m_locals;
+	std::map<std::string, size_t> m_globals;
+	std::map<std::string, size_t> m_functions;
+	std::map<std::string, size_t> m_functionTypes;
+	std::stack<std::string> m_labels;
+	std::map<std::string, std::pair<size_t, size_t>> m_subModulePosAndSize;
+};
+
+
+}
+}
+

libyul/backends/wasm/EWasmAST.h

@@ -23,6 +23,8 @@
 #include <boost/variant.hpp>
 #include <string>
 #include <vector>
+#include <map>
+#include <memory>
 
 namespace yul
 {
@@ -86,6 +88,16 @@ struct FunctionDefinition
 	std::vector<Expression> body;
 };
 
+/**
+ * Abstract representation of a wasm module.
+ */
+struct Module
+{
+	std::vector<GlobalVariableDeclaration> globals;
+	std::vector<FunctionImport> imports;
+	std::vector<FunctionDefinition> functions;
+	std::map<std::string, Module> subModules;
+};
 
 }
 }

libyul/backends/wasm/EWasmCodeTransform.cpp

@@ -20,7 +20,6 @@
 
 #include <libyul/backends/wasm/EWasmCodeTransform.h>
 
-#include <libyul/backends/wasm/EWasmToText.h>
 #include <libyul/optimiser/NameCollector.h>
 
 #include <libyul/AsmData.h>
@@ -37,10 +36,11 @@ using namespace std;
 using namespace dev;
 using namespace yul;
 
-string EWasmCodeTransform::run(Dialect const& _dialect, yul::Block const& _ast)
+wasm::Module EWasmCodeTransform::run(Dialect const& _dialect, yul::Block const& _ast)
 {
+	wasm::Module module;
+
 	EWasmCodeTransform transform(_dialect, _ast);
-	vector<wasm::FunctionDefinition> functions;
 
 	for (auto const& statement: _ast.statements)
 	{
@@ -49,17 +49,14 @@ string EWasmCodeTransform::run(Dialect const& _dialect, yul::Block const& _ast)
 			"Expected only function definitions at the highest level."
 		);
 		if (statement.type() == typeid(yul::FunctionDefinition))
-			functions.emplace_back(transform.translateFunction(boost::get<yul::FunctionDefinition>(statement)));
+			module.functions.emplace_back(transform.translateFunction(boost::get<yul::FunctionDefinition>(statement)));
 	}
 
-	std::vector<wasm::FunctionImport> imports;
 	for (auto& imp: transform.m_functionsToImport)
-		imports.emplace_back(std::move(imp.second));
+		module.imports.emplace_back(std::move(imp.second));
-	return EWasmToText().run(
+	module.globals = transform.m_globalVariables;
-		transform.m_globalVariables,
+
-		imports,
+	return module;
-		functions
-	);
 }
 
 wasm::Expression EWasmCodeTransform::generateMultiAssignment(

libyul/backends/wasm/EWasmCodeTransform.h

@@ -35,7 +35,7 @@ struct AsmAnalysisInfo;
 class EWasmCodeTransform: public boost::static_visitor<wasm::Expression>
 {
 public:
-	static std::string run(Dialect const& _dialect, yul::Block const& _ast);
+	static wasm::Module run(Dialect const& _dialect, yul::Block const& _ast);
 
 public:
 	wasm::Expression operator()(yul::Instruction const& _instruction);

libyul/backends/wasm/EWasmObjectCompiler.cpp

@@ -21,32 +21,36 @@
 #include <libyul/backends/wasm/EWasmObjectCompiler.h>
 
 #include <libyul/backends/wasm/EWasmCodeTransform.h>
+#include <libyul/backends/wasm/BinaryTransform.h>
+#include <libyul/backends/wasm/EWasmToText.h>
 
 #include <libyul/Object.h>
 #include <libyul/Exceptions.h>
 
+#include <libdevcore/CommonData.h>
+
 using namespace yul;
 using namespace std;
 
-string EWasmObjectCompiler::compile(Object& _object, Dialect const& _dialect)
+pair<string, dev::bytes> EWasmObjectCompiler::compile(Object& _object, Dialect const& _dialect)
 {
 	EWasmObjectCompiler compiler(_dialect);
-	return compiler.run(_object);
+	wasm::Module module = compiler.run(_object);
+	return {EWasmToText().run(module), wasm::BinaryTransform::run(module)};
 }
 
-string EWasmObjectCompiler::run(Object& _object)
+wasm::Module EWasmObjectCompiler::run(Object& _object)
 {
-	string ret;
+	yulAssert(_object.analysisInfo, "No analysis info.");
+	yulAssert(_object.code, "No code.");
+
+	wasm::Module module = EWasmCodeTransform::run(m_dialect, *_object.code);
 
 	for (auto& subNode: _object.subObjects)
 		if (Object* subObject = dynamic_cast<Object*>(subNode.get()))
-			ret += compile(*subObject, m_dialect);
+			module.subModules[subObject->name.str()] = run(*subObject);
 		else
 			yulAssert(false, "Data is not yet supported for EWasm.");
 
-	yulAssert(_object.analysisInfo, "No analysis info.");
+	return module;
-	yulAssert(_object.code, "No code.");
-	ret += EWasmCodeTransform::run(m_dialect, *_object.code);
-
-	return ret;
 }

libyul/backends/wasm/EWasmObjectCompiler.h

@@ -21,22 +21,33 @@
 #pragma once
 
 #include <string>
+#include <vector>
+#include <tuple>
 
+namespace dev
+{
+using bytes = std::vector<uint8_t>;
+}
 namespace yul
 {
 struct Object;
 struct Dialect;
+namespace wasm
+{
+struct Module;
+}
 
 class EWasmObjectCompiler
 {
 public:
-	static std::string compile(Object& _object, Dialect const& _dialect);
+	/// Compiles the given object and returns the WAST and the binary representation.
+	static std::pair<std::string, dev::bytes> compile(Object& _object, Dialect const& _dialect);
 private:
 	EWasmObjectCompiler(Dialect const& _dialect):
 		m_dialect(_dialect)
 	{}
 
-	std::string run(Object& _object);
+	wasm::Module run(Object& _object);
 
 	Dialect const& m_dialect;
 };

libyul/backends/wasm/EWasmToText.cpp

@@ -30,14 +30,15 @@ using namespace std;
 using namespace yul;
 using namespace dev;
 
-string EWasmToText::run(
+string EWasmToText::run(wasm::Module const& _module)
-	vector<wasm::GlobalVariableDeclaration> const& _globals,
-	vector<wasm::FunctionImport> const& _imports,
-	vector<wasm::FunctionDefinition> const& _functions
-)
 {
 	string ret = "(module\n";
-	for (wasm::FunctionImport const& imp: _imports)
+	for (auto const& sub: _module.subModules)
+		ret +=
+			"    ;; sub-module \"" +
+			sub.first +
+			"\" will be encoded as custom section in binary here, but is skipped in text mode.\n";
+	for (wasm::FunctionImport const& imp: _module.imports)
 	{
 		ret += "    (import \"" + imp.module + "\" \"" + imp.externalName + "\" (func $" + imp.internalName;
 		if (!imp.paramTypes.empty())
@@ -52,10 +53,10 @@ string EWasmToText::run(
 	// export the main function
 	ret += "    (export \"main\" (func $main))\n";
 
-	for (auto const& g: _globals)
+	for (auto const& g: _module.globals)
 		ret += "    (global $" + g.variableName + " (mut i64) (i64.const 0))\n";
 	ret += "\n";
-	for (auto const& f: _functions)
+	for (auto const& f: _module.functions)
 		ret += transform(f) + "\n";
 	return move(ret) + ")\n";
 }

libyul/backends/wasm/EWasmToText.h

@@ -31,11 +31,7 @@ struct AsmAnalysisInfo;
 class EWasmToText: public boost::static_visitor<std::string>
 {
 public:
-	std::string run(
+	std::string run(wasm::Module const& _module);
-		std::vector<wasm::GlobalVariableDeclaration> const& _globals,
-		std::vector<wasm::FunctionImport> const& _imports,
-		std::vector<wasm::FunctionDefinition> const& _functions
-	);
 
 public:
 	std::string operator()(wasm::Literal const& _literal);

test/cmdlineTests/standard_eWasm_requested/output.json

@@ -1,63 +1,5 @@
 {"contracts":{"A":{"C":{"ewasm":{"wast":"(module
-    (import \"ethereum\" \"revert\" (func $eth.revert (param i32 i32)))
+    ;; sub-module \"C_2_deployed\" will be encoded as custom section in binary here, but is skipped in text mode.
-    (memory $memory (export \"memory\") 1)
-    (export \"main\" (func $main))
-
-(func $main
-    (local $_1 i64)
-    (local $_2 i64)
-    (local $_3 i64)
-    (local $hi i64)
-    (local $y i64)
-    (local $hi_1 i64)
-    (local.set $_1 (i64.const 0))
-    (local.set $_2 (i64.add (call $u256_to_i32 (local.get $_1) (local.get $_1) (local.get $_1) (i64.const 64)) (i64.const 64)))
-    (local.set $_3 (i64.const 65280))
-    (local.set $hi (i64.shl (i64.or (i64.shl (i64.or (i64.and (i64.shl (local.get $_1) (i64.const 8)) (local.get $_3)) (i64.and (i64.shr_u (local.get $_1) (i64.const 8)) (i64.const 255))) (i64.const 16)) (call $endian_swap_16 (i64.shr_u (local.get $_1) (i64.const 16)))) (i64.const 32)))
-    (local.set $y (i64.or (local.get $hi) (call $endian_swap_32 (i64.shr_u (local.get $_1) (i64.const 32)))))
-    (i64.store (i32.wrap_i64 (local.get $_2)) (local.get $y))
-    (i64.store (i32.wrap_i64 (i64.add (local.get $_2) (i64.const 8))) (local.get $y))
-    (i64.store (i32.wrap_i64 (i64.add (local.get $_2) (i64.const 16))) (local.get $y))
-    (local.set $hi_1 (i64.shl (i64.or (i64.shl (i64.or (i64.and (i64.shl (i64.const 128) (i64.const 8)) (local.get $_3)) (i64.and (i64.shr_u (i64.const 128) (i64.const 8)) (i64.const 255))) (i64.const 16)) (call $endian_swap_16 (i64.shr_u (i64.const 128) (i64.const 16)))) (i64.const 32)))
-    (i64.store (i32.wrap_i64 (i64.add (local.get $_2) (i64.const 24))) (i64.or (local.get $hi_1) (call $endian_swap_32 (i64.shr_u (i64.const 128) (i64.const 32)))))
-    (call $eth.revert (i32.wrap_i64 (i64.add (call $u256_to_i32 (local.get $_1) (local.get $_1) (local.get $_1) (local.get $_1)) (i64.const 64))) (i32.wrap_i64 (call $u256_to_i32 (local.get $_1) (local.get $_1) (local.get $_1) (local.get $_1))))
-)
-
-(func $u256_to_i32
-    (param $x1 i64)
-    (param $x2 i64)
-    (param $x3 i64)
-    (param $x4 i64)
-    (result i64)
-    (local $v i64)
-    (if (i64.ne (i64.extend_i32_u (i64.ne (local.get $v) (i64.or (i64.or (local.get $x1) (local.get $x2)) (local.get $x3)))) (i64.const 0)) (then
-        (unreachable)))
-    (if (i64.ne (i64.extend_i32_u (i64.ne (local.get $v) (i64.shr_u (local.get $x4) (i64.const 32)))) (i64.const 0)) (then
-        (unreachable)))
-    (local.set $v (local.get $x4))
-    (local.get $v)
-)
-
-(func $endian_swap_16
-    (param $x i64)
-    (result i64)
-    (local $y i64)
-    (local.set $y (i64.or (i64.and (i64.shl (local.get $x) (i64.const 8)) (i64.const 65280)) (i64.and (i64.shr_u (local.get $x) (i64.const 8)) (i64.const 255))))
-    (local.get $y)
-)
-
-(func $endian_swap_32
-    (param $x i64)
-    (result i64)
-    (local $y i64)
-    (local $hi i64)
-    (local.set $hi (i64.shl (call $endian_swap_16 (local.get $x)) (i64.const 16)))
-    (local.set $y (i64.or (local.get $hi) (call $endian_swap_16 (i64.shr_u (local.get $x) (i64.const 16)))))
-    (local.get $y)
-)
-
-)
-(module
     (import \"ethereum\" \"codeCopy\" (func $eth.codeCopy (param i32 i32 i32)))
     (import \"ethereum\" \"finish\" (func $eth.finish (param i32 i32)))
     (memory $memory (export \"memory\") 1)