Initialize _Py_HashSecret

Author: msimacek

graalpython/com.oracle.graal.python.cext/src/pyhash.c

@@ -47,11 +47,13 @@ Py_hash_t _Py_HashDouble(double value) {
 long _PyHASH_INF;
 long _PyHASH_NAN;
 long _PyHASH_IMAG;
+_Py_HashSecret_t _Py_HashSecret = {{0}};
 
 void initialize_hashes() {
     _PyHASH_INF = UPCALL_L(PY_BUILTIN, polyglot_from_string("hash", SRC_CS), INFINITY);
     _PyHASH_NAN = UPCALL_L(PY_BUILTIN, polyglot_from_string("hash", SRC_CS), NAN);
     _PyHASH_IMAG = UPCALL_L(PY_TRUFFLE_CEXT, polyglot_from_string("PyHash_Imag", SRC_CS));
+    ((void (*)(int8_t *))polyglot_get_member(PY_TRUFFLE_CEXT, polyglot_from_string("PyTruffleHash_InitSecret", SRC_CS)))(polyglot_from_i8_array((int8_t *)&_Py_HashSecret, sizeof(_Py_HashSecret)));
 }
 
 Py_hash_t _Py_HashBytes(const void *src, Py_ssize_t len) {

graalpython/com.oracle.graal.python.shell/src/com/oracle/graal/python/shell/GraalPythonMain.java

@@ -532,6 +532,11 @@ protected void launch(Builder contextBuilder) {
             unbufferedIO = unbufferedIO || System.getenv("PYTHONUNBUFFERED") != null;
             dontWriteBytecode = dontWriteBytecode || System.getenv("PYTHONDONTWRITEBYTECODE") != null;
 
+            String hashSeed = System.getenv("PYTHONHASHSEED");
+            if (hashSeed != null) {
+                contextBuilder.option("python.HashSeed", hashSeed);
+            }
+
             String envWarnOptions = System.getenv("PYTHONWARNINGS");
             if (envWarnOptions != null && !envWarnOptions.isEmpty()) {
                 if (warnOptions == null) {

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/builtins/modules/PythonCextBuiltins.java

@@ -1453,6 +1453,26 @@ long getHash() {
         }
     }
 
+    @Builtin(name = "PyTruffleHash_InitSecret", minNumOfPositionalArgs = 1)
+    @GenerateNodeFactory
+    abstract static class PyTruffleHashGetSecret extends PythonUnaryBuiltinNode {
+        @Specialization
+        @TruffleBoundary
+        Object get(Object secretPtr) {
+            try {
+                InteropLibrary lib = InteropLibrary.getUncached(secretPtr);
+                byte[] secret = getContext().getHashSecret();
+                int len = (int) lib.getArraySize(secretPtr);
+                for (int i = 0; i < len; i++) {
+                    lib.writeArrayElement(secretPtr, i, secret[i]);
+                }
+                return 0;
+            } catch (UnsupportedMessageException | UnsupportedTypeException | InvalidArrayIndexException e) {
+                throw CompilerDirectives.shouldNotReachHere(e);
+            }
+        }
+    }
+
     @Builtin(name = "PyTruffleFrame_New", minNumOfPositionalArgs = 4)
     @GenerateNodeFactory
     abstract static class PyTruffleFrameNewNode extends PythonBuiltinNode {

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/runtime/PythonContext.java

@@ -37,6 +37,8 @@
 import java.io.PrintWriter;
 import java.lang.ref.WeakReference;
 import java.nio.file.LinkOption;
+import java.security.NoSuchAlgorithmException;
+import java.security.SecureRandom;
 import java.text.MessageFormat;
 import java.util.ArrayDeque;
 import java.util.ArrayList;
@@ -419,6 +421,9 @@ PythonThreadState getThreadState() {
     private final ThreadGroup threadGroup = new ThreadGroup(GRAALPYTHON_THREADS);
     private final IDUtils idUtils = new IDUtils();
 
+    // Equivalent of _Py_HashSecret
+    @CompilationFinal(dimensions = 1) private byte[] hashSecret = new byte[24];
+
     // ctypes' used native libraries/functions.
     private final ConcurrentHashMap<Long, Object> ptrAdrMap = new ConcurrentHashMap<>();
 
@@ -1074,6 +1079,11 @@ public long getPerfCounterStart() {
         return perfCounterStart;
     }
 
+    public byte[] getHashSecret() {
+        assert !ImageInfo.inImageBuildtimeCode();
+        return hashSecret;
+    }
+
     public boolean isInitialized() {
         if (PythonUtils.ASSERTIONS_ENABLED && isInitializedNonCompilationFinal != isInitialized) {
             // We cannot use normal assertion, because those are removed in compilation
@@ -1229,6 +1239,9 @@ private void patchPackagePaths(String from, String to) {
     }
 
     private void setupRuntimeInformation(boolean isPatching) {
+        if (!ImageInfo.inImageBuildtimeCode()) {
+            initializeHashSecret();
+        }
         nativeZlib = NFIZlibSupport.createNative(this, "");
         nativeBz2lib = NFIBz2Support.createNative(this, "");
         nativeLZMA = NFILZMASupport.createNative(this, "");
@@ -1255,6 +1268,38 @@ private void setupRuntimeInformation(boolean isPatching) {
         isInitializedNonCompilationFinal = true;
     }
 
+    private void initializeHashSecret() {
+        String hashSeed = getOption(PythonOptions.HashSeed);
+        if (hashSeed.equals("random")) {
+            try {
+                SecureRandom.getInstance("NativePRNGNonBlocking").nextBytes(hashSecret);
+            } catch (NoSuchAlgorithmException e) {
+                throw new RuntimeException("Unable to obtain entropy source for hash randomization (NativePRNGNonBlocking)");
+            }
+        } else {
+            try {
+                long hashSeedValue = Long.parseLong(hashSeed);
+                if (hashSeedValue < 0 || hashSeedValue > 4294967295L) {
+                    throw new NumberFormatException();
+                }
+                // 0 disables the option, leaving the secret at 0
+                if (hashSeedValue != 0) {
+                    // Generate the whole secret from the seed number the same way as CPython
+                    // Taken from bootstrap_hash.c:lcg_urandom
+                    int x = (int) hashSeedValue;
+                    for (int i = 0; i < hashSecret.length; i++) {
+                        x *= 214013;
+                        x += 2531011;
+                        /* modulo 2 ^ (8 * sizeof(int)) */
+                        hashSecret[i] = (byte) ((x >>> 16) & 0xff);
+                    }
+                }
+            } catch (NumberFormatException e) {
+                throw new RuntimeException("PYTHONHASHSEED must be \"random\" or an integer in range [0; 4294967295]");
+            }
+        }
+    }
+
     private void initializePosixSupport() {
         String option = getLanguage().getEngineOption(PythonOptions.PosixModuleBackend);
         PosixSupport result;

graalpython/com.oracle.graal.python/src/com/oracle/graal/python/runtime/PythonOptions.java

@@ -135,6 +135,9 @@ private PythonOptions() {
     @Option(category = OptionCategory.USER, help = "Equivalent to setting the PYTHONWARNINGS environment variable for the standard launcher.", stability = OptionStability.STABLE) //
     public static final OptionKey<String> WarnOptions = new OptionKey<>("");
 
+    @Option(category = OptionCategory.USER, help = "Equivalent to setting PYTHONHASHSEED environment variable", stability = OptionStability.STABLE) //
+    public static final OptionKey<String> HashSeed = new OptionKey<>("random");
+
     @EngineOption @Option(category = OptionCategory.USER, help = "Choose the backend for the POSIX module. Valid values are 'java', 'native', 'llvm'.") //
     public static final OptionKey<String> PosixModuleBackend = new OptionKey<>("java");