diff --git a/libsolidity/codegen/YulUtilFunctions.cpp b/libsolidity/codegen/YulUtilFunctions.cpp
index 4c8b20636..a791cae35 100644
--- a/libsolidity/codegen/YulUtilFunctions.cpp
+++ b/libsolidity/codegen/YulUtilFunctions.cpp
@@ -379,27 +379,32 @@ string YulUtilFunctions::roundUpFunction()
 	});
 }
 
-string YulUtilFunctions::overflowCheckedUIntAddFunction(size_t _bits)
+string YulUtilFunctions::overflowCheckedIntAddFunction(IntegerType const& _type)
 {
-	solAssert(0 < _bits && _bits <= 256 && _bits % 8 == 0, "");
-	string functionName = "checked_add_uint_" + to_string(_bits);
+	string functionName = "checked_add_" + _type.identifier();
+	// TODO: Consider to add a special case for unsigned 256-bit integers
+	//       and use the following instead:
+	//       sum := add(x, y) if lt(sum, x) { revert(0, 0) }
 	return m_functionCollector->createFunction(functionName, [&]() {
 		return
 			Whiskers(R"(
 			function <functionName>(x, y) -> sum {
-				<?shortType>
-					let mask := <mask>
-					sum := add(and(x, mask), and(y, mask))
-					if and(sum, not(mask)) { revert(0, 0) }
-				<!shortType>
-					sum := add(x, y)
-					if lt(sum, x) { revert(0, 0) }
-				</shortType>
+				<?signed>
+					// overflow, if x >= 0 and y > (maxValue - x)
+					if and(iszero(slt(x, 0)), sgt(y, sub(<maxValue>, x))) { revert(0, 0) }
+					// underflow, if x < 0 and y < (minValue - x)
+					if and(slt(x, 0), slt(y, sub(<minValue>, x))) { revert(0, 0) }
+				<!signed>
+					// overflow, if x > (maxValue - y)
+					if gt(x, sub(<maxValue>, y)) { revert(0, 0) }
+				</signed>
+				sum := add(x, y)
 			}
 			)")
-			("shortType", _bits < 256)
 			("functionName", functionName)
-			("mask", toCompactHexWithPrefix((u256(1) << _bits) - 1))
+			("signed", _type.isSigned())
+			("maxValue", toCompactHexWithPrefix(u256(_type.maxValue())))
+			("minValue", toCompactHexWithPrefix(u256(_type.minValue())))
 			.render();
 	});
 }
diff --git a/libsolidity/codegen/YulUtilFunctions.h b/libsolidity/codegen/YulUtilFunctions.h
index ec828ce4f..fcc168626 100644
--- a/libsolidity/codegen/YulUtilFunctions.h
+++ b/libsolidity/codegen/YulUtilFunctions.h
@@ -93,7 +93,7 @@ public:
 	std::string roundUpFunction();
 
 	/// signature: (x, y) -> sum
-	std::string overflowCheckedUIntAddFunction(size_t _bits);
+	std::string overflowCheckedIntAddFunction(IntegerType const& _type);
 
 	/// signature: (x, y) -> product
 	std::string overflowCheckedUIntMulFunction(size_t _bits);
diff --git a/libsolidity/codegen/ir/IRGeneratorForStatements.cpp b/libsolidity/codegen/ir/IRGeneratorForStatements.cpp
index 80a09ccbd..1c64c2fdc 100644
--- a/libsolidity/codegen/ir/IRGeneratorForStatements.cpp
+++ b/libsolidity/codegen/ir/IRGeneratorForStatements.cpp
@@ -1126,18 +1126,27 @@ string IRGeneratorForStatements::binaryOperation(
 	if (IntegerType const* type = dynamic_cast<IntegerType const*>(&_type))
 	{
 		string fun;
-		// TODO: Only division is implemented for signed integers for now.
-		if (!type->isSigned())
+		// TODO: Implement all operations for signed and unsigned types.
+		switch (_operator)
 		{
-			if (_operator == Token::Add)
-				fun = m_utils.overflowCheckedUIntAddFunction(type->numBits());
-			else if (_operator == Token::Sub)
-				fun = m_utils.overflowCheckedUIntSubFunction();
-			else if (_operator == Token::Mul)
-				fun = m_utils.overflowCheckedUIntMulFunction(type->numBits());
+			case Token::Add:
+				fun = m_utils.overflowCheckedIntAddFunction(*type);
+				break;
+			case Token::Sub:
+				if (!type->isSigned())
+					fun = m_utils.overflowCheckedUIntSubFunction();
+				break;
+			case Token::Mul:
+				if (!type->isSigned())
+					fun = m_utils.overflowCheckedUIntMulFunction(type->numBits());
+				break;
+			case Token::Div:
+				fun = m_utils.overflowCheckedIntDivFunction(*type);
+				break;
+			default:
+				break;
 		}
-		if (_operator == Token::Div)
-			fun = m_utils.overflowCheckedIntDivFunction(*type);
+
 		solUnimplementedAssert(!fun.empty(), "");
 		return fun + "(" + _left + ", " + _right + ")\n";
 	}
diff --git a/test/libsolidity/semanticTests/viaYul/detect_add_overflow.yul b/test/libsolidity/semanticTests/viaYul/detect_add_overflow.sol
similarity index 100%
rename from test/libsolidity/semanticTests/viaYul/detect_add_overflow.yul
rename to test/libsolidity/semanticTests/viaYul/detect_add_overflow.sol
diff --git a/test/libsolidity/semanticTests/viaYul/detect_add_overflow_signed.sol b/test/libsolidity/semanticTests/viaYul/detect_add_overflow_signed.sol
new file mode 100644
index 000000000..6e44c44ba
--- /dev/null
+++ b/test/libsolidity/semanticTests/viaYul/detect_add_overflow_signed.sol
@@ -0,0 +1,37 @@
+contract C {
+    function f(int a, int b) public pure returns (int x) {
+        x = a + b;
+    }
+    function g(int8 a, int8 b) public pure returns (int8 x) {
+        x = a + b;
+    }
+}
+// ====
+// compileViaYul: true
+// ----
+// f(int256,int256): 5, 6 -> 11
+// f(int256,int256): -2, 1 -> -1
+// f(int256,int256): -2, 2 -> 0
+// f(int256,int256): 2, -2 -> 0
+// f(int256,int256): -5, -6 -> -11
+// f(int256,int256): 0x7FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF0, 0x0F -> 0x7FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
+// f(int256,int256): 0x0F, 0x7FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF0 -> 0x7FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF
+// f(int256,int256): 0x7FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF, 1 -> FAILURE
+// f(int256,int256): 1, 0x7FFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFFF -> FAILURE
+// f(int256,int256): 0x8000000000000000000000000000000000000000000000000000000000000001, -1 -> 0x8000000000000000000000000000000000000000000000000000000000000000
+// f(int256,int256): -1, 0x8000000000000000000000000000000000000000000000000000000000000001 -> 0x8000000000000000000000000000000000000000000000000000000000000000
+// f(int256,int256): 0x8000000000000000000000000000000000000000000000000000000000000000, -1 -> FAILURE
+// f(int256,int256): -1, 0x8000000000000000000000000000000000000000000000000000000000000000 -> FAILURE
+// g(int8,int8): 5, 6 -> 11
+// g(int8,int8): -2, 1 -> -1
+// g(int8,int8): -2, 2 -> 0
+// g(int8,int8): 2, -2 -> 0
+// g(int8,int8): -5, -6 -> -11
+// g(int8,int8): 126, 1 -> 127
+// g(int8,int8): 1, 126 -> 127
+// g(int8,int8): 127, 1 -> FAILURE
+// g(int8,int8): 1, 127 -> FAILURE
+// g(int8,int8): -127, -1 -> -128
+// g(int8,int8): -1, -127 -> -128
+// g(int8,int8): -127, -2 -> FAILURE
+// g(int8,int8): -2, -127 -> FAILURE