Byte Code Verification cs205: engineering software university of virginia fall 2006 Byte Code Verification

Java Byte Code Instructions 0: nop 1-20: putting constants on the stack 96-119: arithmetic on ints, longs, floats, doubles What other kinds of instructions do we need?

Other Instructions Loading and Storing Variables (65 instructions) Control Flow (~20 instructions) if, goto, return Method Calls (4 instructions) Creating objects (1 instruction) Using object fields (4 instructions) Arrays (3 instructions) checkcast, instanceof

Referencing Memory iload <varnum> istore <varnum> Pushes the int in local variable <varnum> (1 byte) on the stack istore <varnum> Pops the int on the top of the stack and stores it in local variable <varnum> What if you have more than 256 local variables?

Referencing Example public class Locals1 { Method void main(java.lang.String[]) 0 iconst_2 1 istore_1 2 iconst_3 3 istore_2 4 iload_1 5 iload_2 6 iadd 7 istore_3 8 getstatic #2 <Field java.io.PrintStream err> 11 new #3 <Class java.lang.StringBuffer> 14 dup 15 invokespecial #4 <Method java.lang.StringBuffer()> 18 ldc #5 <String "c: "> 20 invokevirtual #6 <Method java.lang.StringBuffer append(java.lang.String)> 23 iload_3 24 invokevirtual #7 <Method java.lang.StringBuffer append(int)> 27 invokevirtual #8 <Method java.lang.String toString()> 30 invokevirtual #9 <Method void println(java.lang.String)> 33 return public class Locals1 { static public void main (String args[]) { int a = 2; int b = 3; int c = a + b; System.err.println ("c: " + c); } }

Control Flow ifeq <label> if_icmple <label> Pop an int off the stack. If it is zero, jump to the label. Otherwise, continue normally. if_icmple <label> Pop two ints off the stack. If the second one is <= the first one, jump to the label. Otherwise, continue normally.

invokevirtual <Method void println(java.lang.String)> Method Calls invokevirtual <method> Invokes the method <method> on the parameters and object on the top of the stack. Finds the appropriate method at run-time based on the actual type of the this object. invokevirtual <Method void println(java.lang.String)>

Method Calls invokestatic <method> Invokes a static (class) method <method> on the parameters on the top of the stack. Finds the appropriate method at run-time based on the actual type of the this object.

Example public class Sample1 { static public void main (String args[]) { System.err.println ("Hello!"); System.exit (1); }

> javap -c Sample1 Compiled from Sample1.java public class Sample1 { static public void main (String args[]) { System.err.println ("Hello!"); System.exit (1); } } > javap -c Sample1 Compiled from Sample1.java public class Sample1 extends java.lang.Object { public Sample1(); public static void main(java.lang.String[]); } Method Sample1() 0 aload_0 1 invokespecial #1 <Method java.lang.Object()> 4 return Method void main(java.lang.String[]) 0 getstatic #2 <Field java.io.PrintStream err> 3 ldc #3 <String "Hello!"> 5 invokevirtual #4 <Method void println(java.lang.String)> 8 iconst_1 9 invokestatic #5 <Method void exit(int)> 12 return

The Worst Instruction Jump subroutine Format jsr branchbyte1 Operation Jump subroutine Format The Worst Instruction jsr branchbyte1 branchbyte2 http://java.sun.com/docs/books/vmspec/2nd-edition/html/Instructions2.doc7.html jsr [branchbyte1] [branchbyte2] Operand Stack ...  ..., address Description The address of the opcode of the instruction immediately following this jsr instruction is pushed onto the operand stack as a value of type returnAddress. The unsigned branchbyte1 and branchbyte2 are used to construct a signed 16-bit offset, where the offset is (branchbyte1 << 8) | branchbyte2. Execution proceeds at that offset from the address of this jsr instruction. The target address must be that of an opcode of an instruction within the method that contains this jsr instruction. Notes The jsr instruction is used with the ret instruction in the implementation of the finally clauses of the Java programming language. Note that jsr pushes the address onto the operand stack and ret gets it out of a local variable. This asymmetry is intentional.

Try-Catch-Finally public class JSR { static public void main (String args[]) { try { System.out.println("hello"); } catch (Exception e) { System.out.println ("There was an exception!"); } finally { System.out.println ("I am finally here!"); }

0 8 14 <Class java.lang.Exception> 0 11 29 any 14 26 29 any Method void main(java.lang.String[]) 0 getstatic #2 <Field java.io.PrintStream out> 3 ldc #3 <String "hello"> 5 invokevirtual #4 <Method void println(java.lang.String)> 8 jsr 35 11 goto 46 14 astore_1 15 getstatic #2 <Field java.io.PrintStream out> 18 ldc #6 <String "There was an exception!"> 20 invokevirtual #4 <Method void println(java.lang.String)> 23 jsr 35 26 goto 46 29 astore_2 30 jsr 35 33 aload_2 34 athrow 35 astore_3 36 getstatic #2 <Field java.io.PrintStream out> 39 ldc #7 <String "I am finally here!"> 41 invokevirtual #4 <Method void println(java.lang.String)> 44 ret 3 46 return public class JSR { static public void main (String args[]) { try { System.out.println("hello"); } catch (Exception e) { System.out.println (“... exception!"); } finally { System.out.println ("I am finally"); } Exception table: from to target type 0 8 14 <Class java.lang.Exception> 0 11 29 any 14 26 29 any 29 33 29 any

Java Bytecode Verifier malcode.java javac Compiler malcode.class JVML Trusted Computing Base Java Bytecode Verifier Invalid Joe User “Okay” STOP JavaVM

Running Mistyped Code .method public static main([Ljava/lang/String;)V … iconst_2 istore_0 aload_0 iconst_3 iadd return .end method > java Simple Exception in thread "main" java.lang.VerifyError: (class: Simple, method: main signature: ([Ljava/lang/String;)V) Register 0 contains wrong type > java –noverify Simple result: 5

Running Mistyped Code ldc 205 .method public static main([Ljava/lang/String;)V … ldc 205 istore_0 aload_0 iconst_2 iconst_3 iadd .end method > java –noverify Simple Unexpected Signal : EXCEPTION_ACCESS_VIOLATION (0xc0000005) occurred at PC=0x809DCEB Function=JVM_FindSignal+0x1105F Library=C:\j2sdk1.4.2\jre\bin\client\jvm.dll Current Java thread: at Simple.main(Simple.java:7) … # # HotSpot Virtual Machine Error : EXCEPTION_ACCESS_VIOLATION # Error ID : 4F530E43505002EF # Please report this error at # http://java.sun.com/cgi-bin/bugreport.cgi # Java VM: Java HotSpot(TM) Client VM (1.4.2-b28 mixed mode)

Bytecode Verifier Checks class file is formatted correctly Magic number: class file starts with 0xCAFEBABE String table, code, methods, etc. Checks JVML code satisfies safety properties Simulates program execution to know types are correct, but doesn’t need to examine any instruction more than once

Verifying Safety Properties Type safe Stack and variable slots must store and load as same type Memory safe Must not attempt to pop more values from stack than are on it Doesn’t access private fields and methods outside class implementation Control flow safe Jumps must be to valid addresses within function, or call/return

Simulating All Paths The bytecode verifier verifies type safety for all possible executions of the program Since there are infinitely many paths through the program, how is this possible?

Verifier (should be) Conservative JVML programs Safe programs Verifiable programs (Slide from Nate Paul’s ACSAC talk)

Complexity Increases Risk JVML programs Safe programs Verifiable programs 7/8 titles confusing Bug (Slide from Nate Paul’s ACSAC talk)

Vulnerabilities in JavaVM 45 40 35 30 25 Vulnerabilities Reported 20 15 10 5 1 2 3 4 5 6 7 8 9 July 1996 Years Since First Release July 2005

Where are They? Verification 12 API bugs 10 Class loading 8 Other or unknown 2 Missing policy checks 3 Configuration 4 DoS attacks (crash, consumption) 5 several of these were because of jsr complexity

Summary: Low-level vs. Policy Security Low-level Code Safety: Type safety, memory safety, control flow safety Needed to prevent malcode from circumventing any policy mechanism Policy Security: Control access and use of resources (files, network, display, etc.) Enforced by Java class Hard part is deciding on a good policy

Charge Monday: Friday: project design documents due Quiz 4: will cover through Friday’s lecture on the Java bytecode verifier Using CVS (Dan) Friday: project design documents due