Add blobhash opcode

r0qs · cameel · r0qs · commit 9a7639ed48a1 · 2024-01-11T14:37:28.000-03:00
Co-authored-by: Kamil Śliwak &lt;kamil.sliwak@codepoets.it&gt;
diff --git a/Changelog.md b/Changelog.md
@@ -1,6 +1,7 @@
 ### 0.8.24 (unreleased)
 
 Language Features:
+ * Yul: Introduce builtin ``blobhash()`` for retrieving a versioned blob hash of a given blob index within the transaction.
 
 
 Compiler Features:
diff --git a/docs/grammar/SolidityLexer.g4 b/docs/grammar/SolidityLexer.g4
@@ -305,7 +305,7 @@ YulEVMBuiltin:
 	| 'delegatecall' | 'staticcall' | 'return' | 'revert' | 'selfdestruct' | 'invalid'
 	| 'log0' | 'log1' | 'log2' | 'log3' | 'log4' | 'chainid' | 'origin' | 'gasprice'
 	| 'blockhash' | 'coinbase' | 'timestamp' | 'number' | 'difficulty' | 'prevrandao'
-	| 'gaslimit' | 'basefee';
+	| 'gaslimit' | 'basefee' | 'blobhash';
 
 YulLBrace: '{' -> pushMode(YulMode);
 YulRBrace: '}' -> popMode;
diff --git a/docs/yul.rst b/docs/yul.rst
@@ -752,8 +752,8 @@ This document does not want to be a full description of the Ethereum virtual mac
 Please refer to a different document if you are interested in the precise semantics.
 
 Opcodes marked with ``-`` do not return a result and all others return exactly one value.
-Opcodes marked with ``F``, ``H``, ``B``, ``C``, ``I``, ``L`` and ``P`` are present since Frontier,
-Homestead, Byzantium, Constantinople, Istanbul, London or Paris respectively.
+Opcodes marked with ``F``, ``H``, ``B``, ``C``, ``I``, ``L``, ``P`` and ``N`` are present since Frontier,
+Homestead, Byzantium, Constantinople, Istanbul, London, Paris or Cancun respectively.
 
 In the following, ``mem[a...b)`` signifies the bytes of memory starting at position ``a`` up to
 but not including position ``b`` and ``storage[p]`` signifies the storage contents at slot ``p``.
@@ -925,6 +925,8 @@ the ``dup`` and ``swap`` instructions as well as ``jump`` instructions, labels a
 +-------------------------+-----+---+-----------------------------------------------------------------+
 | blockhash(b)            |     | F | hash of block nr b - only for last 256 blocks excluding current |
 +-------------------------+-----+---+-----------------------------------------------------------------+
+| blobhash(i)             |     | N | versioned hash of transaction's i-th blob                       |
++-------------------------+-----+---+-----------------------------------------------------------------+
 | coinbase()              |     | F | current mining beneficiary                                      |
 +-------------------------+-----+---+-----------------------------------------------------------------+
 | timestamp()             |     | F | timestamp of the current block in seconds since the epoch       |
diff --git a/libevmasm/Instruction.cpp b/libevmasm/Instruction.cpp
@@ -81,6 +81,7 @@ std::map<std::string, Instruction> const solidity::evmasm::c_instructions =
 	{ "CHAINID", Instruction::CHAINID },
 	{ "SELFBALANCE", Instruction::SELFBALANCE },
 	{ "BASEFEE", Instruction::BASEFEE },
+	{ "BLOBHASH", Instruction::BLOBHASH },
 	{ "POP", Instruction::POP },
 	{ "MLOAD", Instruction::MLOAD },
 	{ "MSTORE", Instruction::MSTORE },
@@ -230,6 +231,7 @@ static std::map<Instruction, InstructionInfo> const c_instructionInfo =
 	{ Instruction::CHAINID,		{ "CHAINID",		0, 0, 1, false, Tier::Base } },
 	{ Instruction::SELFBALANCE,	{ "SELFBALANCE",	0, 0, 1, false, Tier::Low } },
 	{ Instruction::BASEFEE,     { "BASEFEE",        0, 0, 1, false, Tier::Base } },
+	{ Instruction::BLOBHASH,    { "BLOBHASH",       0, 1, 1, false, Tier::VeryLow } },
 	{ Instruction::POP,			{ "POP",			0, 1, 0, false, Tier::Base } },
 	{ Instruction::MLOAD,		{ "MLOAD",			0, 1, 1, true, Tier::VeryLow } },
 	{ Instruction::MSTORE,		{ "MSTORE",			0, 2, 0, true, Tier::VeryLow } },
diff --git a/libevmasm/Instruction.h b/libevmasm/Instruction.h
@@ -89,6 +89,7 @@ enum class Instruction: uint8_t
 	CHAINID,			///< get the config's chainid param
 	SELFBALANCE,		///< get balance of the current account
 	BASEFEE,            ///< get the block's basefee
+	BLOBHASH = 0x49,    ///< get a versioned hash of one of the blobs associated with the transaction
 
 	POP = 0x50,			///< remove item from stack
 	MLOAD,				///< load word from memory
diff --git a/libevmasm/SemanticInformation.cpp b/libevmasm/SemanticInformation.cpp
@@ -473,6 +473,7 @@ bool SemanticInformation::invalidInPureFunctions(Instruction _instruction)
 	case Instruction::CALLVALUE:
 	case Instruction::CHAINID:
 	case Instruction::BASEFEE:
+	case Instruction::BLOBHASH:
 	case Instruction::GAS:
 	case Instruction::GASPRICE:
 	case Instruction::EXTCODESIZE:
diff --git a/libevmasm/SimplificationRule.h b/libevmasm/SimplificationRule.h
@@ -127,6 +127,7 @@ struct EVMBuiltins
 	static auto constexpr CHAINID = PatternGenerator<Instruction::CHAINID>{};
 	static auto constexpr SELFBALANCE = PatternGenerator<Instruction::SELFBALANCE>{};
 	static auto constexpr BASEFEE = PatternGenerator<Instruction::BASEFEE>{};
+	static auto constexpr BLOBHASH = PatternGenerator<Instruction::BLOBHASH>{};
 	static auto constexpr POP = PatternGenerator<Instruction::POP>{};
 	static auto constexpr MLOAD = PatternGenerator<Instruction::MLOAD>{};
 	static auto constexpr MSTORE = PatternGenerator<Instruction::MSTORE>{};
diff --git a/liblangutil/EVMVersion.cpp b/liblangutil/EVMVersion.cpp
@@ -49,6 +49,8 @@ bool EVMVersion::hasOpcode(Instruction _opcode) const
 		return hasSelfBalance();
 	case Instruction::BASEFEE:
 		return hasBaseFee();
+	case Instruction::BLOBHASH:
+		return hasBlobHash();
 	default:
 		return true;
 	}
diff --git a/liblangutil/EVMVersion.h b/liblangutil/EVMVersion.h
@@ -103,6 +103,7 @@ class EVMVersion:
 	bool hasBaseFee() const { return *this >= london(); }
 	bool hasPrevRandao() const { return *this >= paris(); }
 	bool hasPush0() const { return *this >= shanghai(); }
+	bool hasBlobHash() const { return *this >= cancun(); }
 
 	bool hasOpcode(evmasm::Instruction _opcode) const;
 
diff --git a/libyul/AsmAnalysis.cpp b/libyul/AsmAnalysis.cpp
@@ -727,6 +727,8 @@ bool AsmAnalyzer::validateInstructions(evmasm::Instruction _instr, SourceLocatio
 		errorForVM(7721_error, "only available for Istanbul-compatible");
 	else if (_instr == evmasm::Instruction::BASEFEE && !m_evmVersion.hasBaseFee())
 		errorForVM(5430_error, "only available for London-compatible");
+	else if (_instr == evmasm::Instruction::BLOBHASH && !m_evmVersion.hasBlobHash())
+		errorForVM(3502_error, "only available for Cancun-compatible");
 	else if (_instr == evmasm::Instruction::PC)
 		m_errorReporter.error(
 			2450_error,
diff --git a/libyul/backends/evm/EVMDialect.cpp b/libyul/backends/evm/EVMDialect.cpp
@@ -128,11 +128,22 @@ std::set<YulString> createReservedIdentifiers(langutil::EVMVersion _evmVersion)
 		return _instrName == "prevrandao" && _evmVersion < langutil::EVMVersion::paris();
 	};
 
+	// TODO remove this in 0.9.0. We allow creating functions or identifiers in Yul with the name
+	// blobhash for VMs before cancun.
+	auto blobHashException = [&](evmasm::Instruction _instr) -> bool
+	{
+		return _instr == evmasm::Instruction::BLOBHASH && _evmVersion < langutil::EVMVersion::cancun();
+	};
+
 	std::set<YulString> reserved;
 	for (auto const& instr: evmasm::c_instructions)
 	{
 		std::string name = toLower(instr.first);
-		if (!baseFeeException(instr.second) && !prevRandaoException(name))
+		if (
+			!baseFeeException(instr.second) &&
+			!prevRandaoException(name) &&
+			!blobHashException(instr.second)
+		)
 			reserved.emplace(name);
 	}
 	reserved += std::vector<YulString>{
diff --git a/scripts/error_codes.py b/scripts/error_codes.py
@@ -203,6 +203,7 @@ def examine_id_coverage(top_dir, source_id_to_file_names, new_ids_only=False):
                 # Due to 3805, the warning lists look different for different compiler builds.
         "1834", # Unimplemented feature error, as we do not test it anymore via cmdLineTests
         "5430", # basefee being used in inline assembly for EVMVersion < london
+        "3502", # blobhash being used in inline assembly for EVMVersion < cancun
         "1180", # SMTChecker, covered by CL tests
         "2339", # SMTChecker, covered by CL tests
         "2961", # SMTChecker, covered by CL tests
diff --git a/scripts/test_antlr_grammar.sh b/scripts/test_antlr_grammar.sh
@@ -124,6 +124,8 @@ done < <(
       # Skipping tests with "let prevrandao := ..."
       grep -v -E 'inlineAssembly/prevrandao_allowed_function_pre_paris.sol' |
       grep -v -E 'inlineAssembly/prevrandao_disallowed_function_post_paris.sol' |
+      # Skipping a test with "let blobhash := ..."
+      grep -v -E 'inlineAssembly/blobhash_pre_cancun.sol' |
       # Skipping license error, unrelated to the grammar
       grep -v -E 'license/license_double5.sol' |
       grep -v -E 'license/license_hidden_unicode.sol' |
diff --git a/test/EVMHost.cpp b/test/EVMHost.cpp
@@ -137,6 +137,13 @@ EVMHost::EVMHost(langutil::EVMVersion _evmVersion, evmc::VM& _vm):
 	// The minimum value of basefee
 	tx_context.block_base_fee = evmc::bytes32{7};
 
+	static evmc_bytes32 const blob_hashes_array[] = {
+		0x0100000000000000000000000000000000000000000000000000000000000001_bytes32,
+		0x0100000000000000000000000000000000000000000000000000000000000002_bytes32
+	};
+	tx_context.blob_hashes = blob_hashes_array;
+	tx_context.blob_hashes_count = sizeof(blob_hashes_array) / sizeof(blob_hashes_array[0]);
+
 	// Reserve space for recording calls.
 	if (!recorded_calls.capacity())
 		recorded_calls.reserve(max_recorded_calls);
diff --git a/test/libsolidity/semanticTests/inlineAssembly/blobhash.sol b/test/libsolidity/semanticTests/inlineAssembly/blobhash.sol
@@ -0,0 +1,11 @@
+contract C {
+    function f() public view returns (bytes32 ret) {
+        assembly {
+            ret := blobhash(0)
+        }
+    }
+}
+// ====
+// EVMVersion: >=cancun
+// ----
+// f() -> 0x0100000000000000000000000000000000000000000000000000000000000001
diff --git a/test/libsolidity/semanticTests/inlineAssembly/blobhash_index_exceeding_blob_count.sol b/test/libsolidity/semanticTests/inlineAssembly/blobhash_index_exceeding_blob_count.sol
@@ -0,0 +1,14 @@
+contract C {
+    function f() public view returns (bytes32 ret) {
+        assembly {
+            // EIP-4844 specifies that if `index < len(tx.blob_versioned_hashes)`, `blobhash(index)` should return 0.
+            // Thus, as we injected only two blob hashes in the transaction context in EVMHost,
+            // the return value of the function below MUST be zero.
+            ret := blobhash(2)
+        }
+    }
+}
+// ====
+// EVMVersion: >=cancun
+// ----
+// f() -> 0x00
diff --git a/test/libsolidity/semanticTests/inlineAssembly/blobhash_pre_cancun.sol b/test/libsolidity/semanticTests/inlineAssembly/blobhash_pre_cancun.sol
@@ -0,0 +1,21 @@
+contract C {
+    function f() public pure returns (uint ret) {
+        assembly {
+            let blobhash := 1
+            ret := blobhash
+        }
+    }
+    function g() public pure returns (uint ret) {
+        assembly {
+            function blobhash() -> r {
+                r := 1000
+            }
+            ret := blobhash()
+        }
+    }
+}
+// ====
+// EVMVersion: <=shanghai
+// ----
+// f() -> 1
+// g() -> 1000
diff --git a/test/libsolidity/syntaxTests/inlineAssembly/blobhash.sol b/test/libsolidity/syntaxTests/inlineAssembly/blobhash.sol
@@ -0,0 +1,10 @@
+contract C {
+    function f() public view returns (bytes32 ret) {
+        assembly {
+            ret := blobhash(1)
+        }
+    }
+}
+// ====
+// EVMVersion: >=cancun
+// ----
diff --git a/test/libsolidity/syntaxTests/inlineAssembly/blobhash_pre_cancun.sol b/test/libsolidity/syntaxTests/inlineAssembly/blobhash_pre_cancun.sol
@@ -0,0 +1,12 @@
+contract C {
+    function f() pure external returns (bytes32 ret) {
+        assembly {
+            ret := blobhash()
+        }
+    }
+}
+// ====
+// EVMVersion: <=shanghai
+// ----
+// DeclarationError 4619: (106-114): Function "blobhash" not found.
+// DeclarationError 8678: (99-116): Variable count for assignment to "ret" does not match number of values (1 vs. 0)
diff --git a/test/libsolidity/syntaxTests/inlineAssembly/blobhash_reserved_cancun.sol b/test/libsolidity/syntaxTests/inlineAssembly/blobhash_reserved_cancun.sol
@@ -0,0 +1,14 @@
+contract C {
+    function f() public pure returns (uint ret) {
+        assembly {
+            function blobhash() -> r {
+                r := 1000
+            }
+            ret := blobhash()
+        }
+    }
+}
+// ====
+// EVMVersion: >=cancun
+// ----
+// ParserError 5568: (103-111): Cannot use builtin function name "blobhash" as identifier name.
diff --git a/test/libsolidity/syntaxTests/viewPureChecker/blobhash_not_pure.sol b/test/libsolidity/syntaxTests/viewPureChecker/blobhash_not_pure.sol
@@ -0,0 +1,9 @@
+contract C {
+    function f() public pure {
+        assembly { pop(blobhash(0)) }
+    }
+}
+// ====
+// EVMVersion: >=cancun
+// ----
+// TypeError 2527: (67-78): Function declared as pure, but this expression (potentially) reads from the environment or state and thus requires "view".
diff --git a/test/libyul/yulInterpreterTests/blobhash.yul b/test/libyul/yulInterpreterTests/blobhash.yul
@@ -0,0 +1,13 @@
+{
+    sstore(0, blobhash(0))
+    sstore(1, blobhash(1))
+    sstore(2, blobhash(2)) // should store 0 since EVMHost has only two blob hashes injected in the block the transaction is being executed.
+}
+// ====
+// EVMVersion: >=cancun
+// ----
+// Trace:
+// Memory dump:
+// Storage dump:
+//   0000000000000000000000000000000000000000000000000000000000000000: 014916dd28fc4c10d78e287ca5d9cc51ee1ae73cbfde08c6b37324cbfaac8bc5
+//   0000000000000000000000000000000000000000000000000000000000000001: 0167d3dbed802941483f1afa2a6bc68de5f653128aca9bf1461c5d0a3ad36ed2
diff --git a/test/libyul/yulSyntaxTests/blobhash.yul b/test/libyul/yulSyntaxTests/blobhash.yul
@@ -0,0 +1,15 @@
+{
+    {
+        let blobhash := 1
+    }
+
+    {
+        function blobhash() {}
+        blobhash()
+    }
+}
+
+// ====
+// EVMVersion: >=cancun
+// ----
+// ParserError 5568: (20-28): Cannot use builtin function name "blobhash" as identifier name.
diff --git a/test/libyul/yulSyntaxTests/blobhash_pre_cancun.sol b/test/libyul/yulSyntaxTests/blobhash_pre_cancun.sol
@@ -0,0 +1,13 @@
+{
+    {
+        let blobhash := 1
+    }
+
+    {
+        function blobhash() {}
+        blobhash()
+    }
+}
+// ====
+// EVMVersion: <=shanghai
+// ----
diff --git a/test/tools/yulInterpreter/EVMInstructionInterpreter.cpp b/test/tools/yulInterpreter/EVMInstructionInterpreter.cpp
@@ -31,6 +31,7 @@
 
 #include <libsolutil/Keccak256.h>
 #include <libsolutil/Numeric.h>
+#include <libsolutil/picosha2.h>
 
 #include <limits>
 
@@ -233,6 +234,8 @@ u256 EVMInstructionInterpreter::eval(
 		return m_state.chainid;
 	case Instruction::BASEFEE:
 		return m_state.basefee;
+	case Instruction::BLOBHASH:
+		return blobHash(arg[0]);
 	case Instruction::EXTCODESIZE:
 		return u256(keccak256(h256(arg[0]))) & 0xffffff;
 	case Instruction::EXTCODEHASH:
@@ -647,3 +650,16 @@ std::pair<bool, size_t> EVMInstructionInterpreter::isInputMemoryPtrModified(
 	else
 		return {false, 0};
 }
+
+h256 EVMInstructionInterpreter::blobHash(u256 const& _index)
+{
+	yulAssert(m_evmVersion.hasBlobHash());
+	if (_index >= m_state.blobCommitments.size())
+		return util::FixedHash<32>{};
+
+	h256 hashedCommitment = h256(picosha2::hash256(toBigEndian(m_state.blobCommitments[static_cast<size_t>(_index)])));
+	yulAssert(m_state.blobHashVersion.size == 1);
+	hashedCommitment[0] = *m_state.blobHashVersion.data();
+	yulAssert(hashedCommitment.size == 32);
+	return hashedCommitment;
+}
diff --git a/test/tools/yulInterpreter/EVMInstructionInterpreter.h b/test/tools/yulInterpreter/EVMInstructionInterpreter.h
@@ -24,6 +24,7 @@
 #include <libyul/ASTForward.h>
 
 #include <libsolutil/CommonData.h>
+#include <libsolutil/FixedHash.h>
 #include <libsolutil/Numeric.h>
 
 #include <liblangutil/EVMVersion.h>
@@ -90,6 +91,9 @@ class EVMInstructionInterpreter
 		std::vector<u256> const& _evaluatedArguments
 	);
 
+	/// @returns the blob versioned hash
+	util::h256 blobHash(u256 const& _index);
+
 private:
 	/// Checks if the memory access is valid and adjusts msize accordingly.
 	/// @returns true if memory access is valid, false otherwise
diff --git a/test/tools/yulInterpreter/Interpreter.h b/test/tools/yulInterpreter/Interpreter.h
@@ -111,6 +111,11 @@ struct InterpreterState
 	/// Number of the current state instance, used for recursion protection
 	size_t numInstance = 0;
 
+	// Blob commitment hash version
+	util::FixedHash<1> const blobHashVersion = util::FixedHash<1>(1);
+	// Blob commitments
+	std::array<u256, 2> const blobCommitments = {0x01, 0x02};
+
 	/// Prints execution trace and non-zero storage to @param _out.
 	/// Flag @param _disableMemoryTrace, if set, does not produce a memory dump. This
 	/// avoids false positives reports by the fuzzer when certain optimizer steps are

Original file line number	Diff line number	Diff line change
`@@ -49,6 +49,8 @@ bool EVMVersion::hasOpcode(Instruction _opcode) const`
`49`	`49`	`return hasSelfBalance();`
`50`	`50`	`case Instruction::BASEFEE:`
`51`	`51`	`return hasBaseFee();`
	`52`	`+ case Instruction::BLOBHASH:`
	`53`	`+ return hasBlobHash();`
`52`	`54`	`default:`
`53`	`55`	`return true;`
`54`	`56`	`}`