core/vm: Limit JUMPDEST analysis only to code section of EOF

Author: gumb0

core/vm/analysis.go

@@ -61,20 +61,21 @@ func (bits *bitvec) codeSegment(pos uint64) bool {
 }
 
 // codeBitmap collects data locations in code.
-func codeBitmap(code []byte) bitvec {
+func codeBitmap(c *Contract) bitvec {
 	// The bitmap is 4 bytes longer than necessary, in case the code
 	// ends with a PUSH32, the algorithm will push zeroes onto the
 	// bitvector outside the bounds of the actual code.
-	bits := make(bitvec, len(code)/8+1+4)
-	return codeBitmapInternal(code, bits)
+	bits := make(bitvec, c.CodeSize()/8+1+4)
+	return codeBitmapInternal(c, bits)
 }
 
 // codeBitmapInternal is the internal implementation of codeBitmap.
 // It exists for the purpose of being able to run benchmark tests
 // without dynamic allocations affecting the results.
-func codeBitmapInternal(code, bits bitvec) bitvec {
-	for pc := uint64(0); pc < uint64(len(code)); {
-		op := OpCode(code[pc])
+func codeBitmapInternal(c *Contract, bits bitvec) bitvec {
+	codeBegin := c.CodeBeginOffset()
+	for pc := uint64(0); pc < c.CodeSize(); {
+		op := OpCode(c.Code[codeBegin+pc])
 		pc++
 		if int8(op) < int8(PUSH1) { // If not PUSH (the int8(op) > int(PUSH32) is always false).
 			continue

core/vm/analysis_test.go

@@ -20,6 +20,7 @@ import (
 	"math/bits"
 	"testing"
 
+	"github.com/ethereum/go-ethereum/common"
 	"github.com/ethereum/go-ethereum/crypto"
 )
 
@@ -52,7 +53,15 @@ func TestJumpDestAnalysis(t *testing.T) {
 		{[]byte{byte(PUSH32)}, 0b0000_0001, 4},
 	}
 	for i, test := range tests {
-		ret := codeBitmap(test.code)
+		var (
+			contract Contract
+			addr     common.Address
+			hash     common.Hash
+			header   EOF1Header
+		)
+		contract.SetCallCode(&addr, hash, test.code, &header)
+
+		ret := codeBitmap(&contract)
 		if ret[test.which] != test.exp {
 			t.Fatalf("test %d: expected %x, got %02x", i, test.exp, ret[test.which])
 		}
@@ -64,10 +73,17 @@ const analysisCodeSize = 1200 * 1024
 func BenchmarkJumpdestAnalysis_1200k(bench *testing.B) {
 	// 1.4 ms
 	code := make([]byte, analysisCodeSize)
-	bench.SetBytes(analysisCodeSize)
+	var (
+		contract Contract
+		addr     common.Address
+		hash     common.Hash
+		header   EOF1Header
+	)
+	contract.SetCallCode(&addr, hash, code, &header)
 	bench.ResetTimer()
+	bench.SetBytes(analysisCodeSize)
 	for i := 0; i < bench.N; i++ {
-		codeBitmap(code)
+		codeBitmap(&contract)
 	}
 	bench.StopTimer()
 }
@@ -83,20 +99,27 @@ func BenchmarkJumpdestHashing_1200k(bench *testing.B) {
 }
 
 func BenchmarkJumpdestOpAnalysis(bench *testing.B) {
-	var op OpCode
+	var (
+		contract Contract
+		addr     common.Address
+		hash     common.Hash
+		header   EOF1Header
+		op       OpCode
+	)
 	bencher := func(b *testing.B) {
 		code := make([]byte, analysisCodeSize)
 		b.SetBytes(analysisCodeSize)
 		for i := range code {
 			code[i] = byte(op)
 		}
 		bits := make(bitvec, len(code)/8+1+4)
+		contract.SetCallCode(&addr, hash, code, &header)
 		b.ResetTimer()
 		for i := 0; i < b.N; i++ {
 			for j := range bits {
 				bits[j] = 0
 			}
-			codeBitmapInternal(code, bits)
+			codeBitmapInternal(&contract, bits)
 		}
 	}
 	for op = PUSH1; op <= PUSH32; op++ {

core/vm/contract.go

@@ -88,11 +88,11 @@ func (c *Contract) validJumpdest(dest *uint256.Int) bool {
 	udest, overflow := dest.Uint64WithOverflow()
 	// PC cannot go beyond len(code) and certainly can't be bigger than 63bits.
 	// Don't bother checking for JUMPDEST in that case.
-	if overflow || udest >= uint64(len(c.Code)) {
+	if overflow || udest >= c.CodeSize() {
 		return false
 	}
 	// Only JUMPDESTs allowed for destinations
-	if OpCode(c.Code[udest]) != JUMPDEST {
+	if OpCode(c.Code[c.CodeBeginOffset()+udest]) != JUMPDEST {
 		return false
 	}
 	return c.isCode(udest)
@@ -114,7 +114,7 @@ func (c *Contract) isCode(udest uint64) bool {
 		if !exist {
 			// Do the analysis and save in parent context
 			// We do not need to store it in c.analysis
-			analysis = codeBitmap(c.Code)
+			analysis = codeBitmap(c)
 			c.jumpdests[c.CodeHash] = analysis
 		}
 		// Also stash it in current contract for faster access
@@ -126,7 +126,7 @@ func (c *Contract) isCode(udest uint64) bool {
 	// we don't have to recalculate it for every JUMP instruction in the execution
 	// However, we don't save it within the parent context
 	if c.analysis == nil {
-		c.analysis = codeBitmap(c.Code)
+		c.analysis = codeBitmap(c)
 	}
 	return c.analysis.codeSegment(udest)
 }
@@ -201,6 +201,13 @@ func (c *Contract) CodeEndOffset() uint64 {
 	return c.header.CodeEndOffset()
 }
 
+func (c *Contract) CodeSize() uint64 {
+	if c.IsLegacy() {
+		return uint64(len(c.Code))
+	}
+	return uint64(c.header.codeSize)
+}
+
 // SetCallCode sets the code of the contract and address of the backing data
 // object
 func (c *Contract) SetCallCode(addr *common.Address, hash common.Hash, code []byte, header *EOF1Header) {