Project

General

Profile

Wiki » History » Version 4

Andrei Tatarnikov, 09/28/2012 06:52 PM

1 1 Andrei Tatarnikov
h1. Constraint Solver API
2
3 2 Andrei Tatarnikov
h2. Basic Concepts
4 1 Andrei Tatarnikov
5 2 Andrei Tatarnikov
Constraint Solver Java API provides a Java API for generating pseudorandom values that satisfy certain constraints. This feature is important for test generators that aim at creating directed tests. At logical level, a constraint is represented by a set of expressions that specify limitations for input values (assertions that must be hold for those values). If there are values satisfying all of the specified assertions they will be used a solution for the constraint. If there is a multitude of values satisfying the constraint, specific values will be selected from the range of possible solutions on random basis.
6
7 4 Andrei Tatarnikov
From an implementational point of view, the API represents a wrapper around some kind of an openly distributed SMT solver engine (in the current version, we use the Z3 solver by Microsoft Research). It can be extended to support other solver engines and it provides a possibility to interact with different solver engines in a uniform way. Also, it facilitates creating task-specific custom solvers and extending functionality of existing solver engines by adding custom operations (macros based on built-in operations).
8 2 Andrei Tatarnikov
9 3 Andrei Tatarnikov
h2. SMT-LIB
10 2 Andrei Tatarnikov
11
In SMT solvers, a special functional language is used to specify constraints. The constraint solver subsystem generates constructions in the SMT language and runs the engine to process them and produce the results (find values of unknown input variables).
12
13
// TODO: more info
14 1 Andrei Tatarnikov
15
h2. Constraints and SMT
16
17
Constrains specified as an SMT model are represented by a set of assertions (formulas) that must be satisfied. An SMT solver checks the satisfiability of the model and suggests a solution (variable values) that would satisfy the model. In the example below, we specify a model that should help us create a test that will cause a MIPS processor to generate an exception. We want to find values of the rs and rt general purpose registers that will cause the ADD instruction to raise an integer overflow exception. It should be correct 32-bit signed integers that are not equal to each other. Here is an SMT script:
18
19
<pre>
20
(define-sort        Int_t () (_ BitVec 64))
21
22
(define-fun      INT_ZERO () Int_t (_ bv0 64))
23
(define-fun INT_BASE_SIZE () Int_t (_ bv32 64))
24
(define-fun INT_SIGN_MASK () Int_t (bvshl (bvnot INT_ZERO) INT_BASE_SIZE))
25
26
(define-fun IsValidPos ((x!1 Int_t)) Bool (ite (= (bvand x!1 INT_SIGN_MASK) INT_ZERO) true false))
27
(define-fun IsValidNeg ((x!1 Int_t)) Bool (ite (= (bvand x!1 INT_SIGN_MASK) INT_SIGN_MASK) true false))
28
(define-fun IsValidSignedInt ((x!1 Int_t)) Bool (ite (or (IsValidPos x!1) (IsValidNeg x!1)) true false))
29
30
(declare-const rs Int_t)
31
(declare-const rt Int_t)
32
33
; rt and rs must contain valid sign-extended 32-bit values (bits 63..31 equal)
34
(assert (IsValidSignedInt rs))
35
(assert (IsValidSignedInt rt))
36
37
; the condition for an overflow: the summation result is not a valid sign-extended 32-bit value
38
(assert (not (IsValidSignedInt (bvadd rs rt))))
39
40
; just in case: rs and rt are not equal (to make the results more interesting)
41
(assert (not (= rs rt)))
42
43
(check-sat)
44
45
(echo "Values that lead to an overflow:")
46
(get-value (rs rt))
47
</pre>
48
49
In an ideal case, each run of an SMT solver should return random values from the set of possible solutions. This should improve test coverage. Unfortunately, the current implementation is limited to a single solution that is constant for all run. This should be improved in the final version.   
50
51
h2. Tree Representation
52
53
In our system, we use context-independent syntax trees to represent constraints. These trees are then used to generate a representation that can be understood by a particular SMT solver. Generally, it is an SMT model that uses some limited set of solver features applicable to microprocessor verification. The syntax tree contains nodes of the following types:
54
# Constraint. This is the root node of the tree. It holds the list of unknown variables and the list of assertions (formulas) for these variables.
55
# Formula. Represents an assertion expression. Can be combined with other formulas to build a more complex expression (by applying logic OR, AND or NOT to it). The underlying expression must be a logic expression that can be solved to true or false.
56
# Operation. Represents an unary or binary operation with some unknown variable, some value or some expression as parameters.
57
# Variable.Represents an input variable. It can have an assigned value and, in such a case, will be treated as a value. Otherwise, it is an unknown variable. A variable includes a type as an attribute.
58
# Value. Specifies some known value of the specified type which can be accessed as an attribute.
59
60
Note: Operation, Variables and Value are designed to be treated polymorphically. This allows combining them to build complex expressions.
61
62
h2. Constraint Solver Java Library
63
64
The Constraint Solver subsystem is implemented in Java. The source code files are located in the "microtesk++/constraint-solver" folder. The Java classes are organized in the following packages:
65
# ru.ispras.microtesk.constraints - contains SMT model generation logic and solver implementations.
66
# ru.ispras.microtesk.constraints.syntax - contains classes implementing syntax tree nodes.
67
# ru.ispras.microtesk.constraints.syntax.types - contains code that specifies particular data types and operation types.
68
# ru.ispras.microtesk.constraints.tests - contains JUnit test cases.
69
70
h3. Core classes/interfaces
71
72
*Syntax Tree Implementation*
73
74
The syntax tree nodes are implemented in the following classes:
75
* Constraint. Parameterized by a collection of Variable objects and a collection of Formula objects.
76
* Formula. Parameterized by an Operation object.
77
* Operation. Implements SyntaxElement. Parameterized by operand objects implementing SyntaxElement and an operation type object implementing OperationType.
78
* Variable. Implements SyntaxElement. Parameterized by the variable name string, a data type object implemeting DataType and a BigInteger value object.   
79
* Value. Implements SyntaxElement. Parameterized a data type object implemeting DataType and a BigInteger value object.
80
81
The SyntaxElement interface provides the ability to combine different kinds of elements into expressions.
82
83
The current implementation supports operations with the following data types: (1) Bit vectors, (2) Booleans. They are implemented in the BitVector and LogicBoolean classes. The BitVectorOperation and LogicBooleanOperation classes specify supported operation with these types. For example, the LogicBooleanOperation class looks like this:
84
85
<pre>
86
public final class LogicBooleanOperation extends OperationType
87
{
88
	private LogicBooleanOperation() {}
89
	
90
	/** Operation: Logic - Equality */
91
	public static final OperationType EQ = new LogicBooleanOperation();
92
	/** Operation: Logic - AND */
93
	public static final OperationType AND = new LogicBooleanOperation();
94
	/** Operation: Logic - OR */
95
	public static final OperationType OR  = new LogicBooleanOperation();
96
	/** Operation: Logic - NOT */
97
	public static final OperationType NOT = new LogicBooleanOperation();
98
	/** Operation: Logic - XOR */
99
	public static final OperationType XOR = new LogicBooleanOperation();
100
	/** Operation: Logic - Implication */
101
	public static final OperationType IMPL= new LogicBooleanOperation();
102
} 
103
</pre>
104
105
The code below demonstrates how we can build a syntax tree representation for the integer overflow constraint:
106
107
<pre>
108
class BitVectorIntegerOverflowTestCase implements SolverTestCase
109
{
110
	private static final int      BIT_VECTOR_LENGTH = 64;
111
	private static final DataType   BIT_VECTOR_TYPE = DataType.getBitVector(BIT_VECTOR_LENGTH);
112
	private static final Value             INT_ZERO = new Value(new BigInteger("0"), BIT_VECTOR_TYPE);
113
	private static final Value        INT_BASE_SIZE = new Value(new BigInteger("32"), BIT_VECTOR_TYPE);
114
115
	private static final Operation    INT_SIGN_MASK =
116
		new Operation(BitVectorOperation.BVSHL, new Operation(BitVectorOperation.BVNOT, INT_ZERO, null), INT_BASE_SIZE);
117
	
118
	private Operation IsValidPos(SyntaxElement arg)
119
	{
120
		return new Operation(LogicBooleanOperation.EQ, new Operation(BitVectorOperation.BVAND, arg, INT_SIGN_MASK), INT_ZERO);
121
	}
122
	
123
	private Operation IsValidNeg(SyntaxElement arg)
124
	{
125
		return new Operation(LogicBooleanOperation.EQ, new Operation(BitVectorOperation.BVAND, arg, INT_SIGN_MASK), INT_SIGN_MASK);
126
	}
127
	
128
	private Operation IsValidSignedInt(SyntaxElement arg)
129
	{
130
		return new Operation(LogicBooleanOperation.OR, IsValidPos(arg), IsValidNeg(arg));
131
	}
132
	
133
	public Constraint getConstraint()
134
	{
135
		Constraint constraint = new Constraint();
136
		
137
		Variable rs = new Variable("rs", BIT_VECTOR_TYPE, null);
138
		constraint.addVariable(rs);
139
		
140
		Variable rt = new Variable("rt", BIT_VECTOR_TYPE, null);
141
		constraint.addVariable(rt);
142
		
143
		
144
		constraint.addFormula(
145
			new Formula(
146
				IsValidSignedInt(rs)
147
			)
148
		);
149
		
150
		constraint.addFormula(
151
			new Formula(
152
				IsValidSignedInt(rt)
153
			)
154
		);
155
156
		constraint.addFormula(
157
			new Formula(
158
				new Operation(
159
					LogicBooleanOperation.NOT,
160
					IsValidSignedInt(new Operation(BitVectorOperation.BVADD, rs, rt)),
161
					null
162
				) 
163
			)
164
		);
165
166
		constraint.addFormula(
167
			new Formula(
168
				new Operation(LogicBooleanOperation.NOT, new Operation(LogicBooleanOperation.EQ, rs, rt), null)
169
			)
170
		);
171
172
		return constraint;
173
	}
174
	
175
	public Vector<Variable> getExpectedVariables()	
176
	{
177
		Vector<Variable> result = new Vector<Variable>();
178
		
179
		result.add(new Variable("rs", BIT_VECTOR_TYPE, new BigInteger("000000009b91b193", 16)));
180
		result.add(new Variable("rt", BIT_VECTOR_TYPE, new BigInteger("000000009b91b1b3", 16)));
181
		
182
		return result;	
183
	}
184
}
185
</pre>
186
187
*Representation Translation*
188
189
The logic that translates a tree representation into an SMT representation is implemented in the following way: Methods of the Translator class traverse the constraint syntax tree and use methods of the RepresentationBuilder interface to translate information about its nodes into a representation that can be understood by a particular solver. The RepresentationBuilder interface looks like follows:
190
191
<pre>
192
public interface RepresentationBuilder
193
{	
194
	public void addVariableDeclaration(Variable variable);
195
196
	public void beginConstraint();
197
	public void endConstraint();
198
199
	public void beginFormula();
200
	public void endFormula();
201
202
	public void beginExpression();
203
	public void endExpression();
204
205
	public void appendValue(Value value);
206
	public void appendVariable(Variable variable);
207
	public void appendOperation(OperationType type);
208
}
209
</pre>
210
211
*Solver Implementation*
212
213
Solvers use the Translator class and a specific implementation of the RepresentationBuilder interface to generate an SMT representation of a constraint. Then they run a solver engine to solve the constraint and produce the results. Solver implement a common interface called Solver that looks like this:
214
215
<pre>
216
public interface Solver
217
{
218
	public boolean solveConstraint(Constraint constraint);
219
	
220
	public boolean isSolved();
221
	public boolean isSatisfiable();
222
	
223
	public int getErrorCount();
224
	public String getErrorText(int index);
225
	
226
	public int getVariableCount();
227
	public Variable getVariable(int index);
228
}
229
</pre>