- Issue #13703: oCERT-2011-003: add -R command-line option and PYTHONHASHSEED

  environment variable, to provide an opt-in way to protect against denial of
  service attacks due to hash collisions within the dict and set types.  Patch
  by David Malcolm, based on work by Victor Stinner.

4 files changed, 35 insertions, 3 deletions

diff --git a/Objects/bufferobject.c b/Objects/bufferobject.c
index 0f7e763..c52f0bc 100644
--- a/Objects/bufferobject.c
+++ b/Objects/bufferobject.c
@@ -334,10 +334,20 @@ buffer_hash(PyBufferObject *self)
         return -1;
     p = (unsigned char *) ptr;
     len = size;
-    x = *p << 7;
+    /*
+      We make the hash of the empty buffer be 0, rather than using
+      (prefix ^ suffix), since this slightly obfuscates the hash secret
+    */
+    if (len == 0) {
+        self->b_hash = 0;
+        return 0;
+    }
+    x = _Py_HashSecret.prefix;
+    x ^= *p << 7;
     while (--len >= 0)
         x = (1000003*x) ^ *p++;
     x ^= size;
+    x ^= _Py_HashSecret.suffix;
     if (x == -1)
         x = -2;
     self->b_hash = x;
diff --git a/Objects/object.c b/Objects/object.c
index 424f353..9816541 100644
--- a/Objects/object.c
+++ b/Objects/object.c
@@ -1101,6 +1101,8 @@ PyObject_HashNotImplemented(PyObject *self)
     return -1;
 }
 
+_Py_HashSecret_t _Py_HashSecret;
+
 long
 PyObject_Hash(PyObject *v)
 {
diff --git a/Objects/stringobject.c b/Objects/stringobject.c
index e37b579..9e2673d 100644
--- a/Objects/stringobject.c
+++ b/Objects/stringobject.c
@@ -1212,11 +1212,21 @@ string_hash(PyStringObject *a)
     if (a->ob_shash != -1)
         return a->ob_shash;
     len = Py_SIZE(a);
+    /*
+      We make the hash of the empty string be 0, rather than using
+      (prefix ^ suffix), since this slightly obfuscates the hash secret
+    */
+    if (len == 0) {
+        a->ob_shash = 0;
+        return 0;
+    }
     p = (unsigned char *) a->ob_sval;
-    x = *p << 7;
+    x = _Py_HashSecret.prefix;
+    x ^= *p << 7;
     while (--len >= 0)
         x = (1000003*x) ^ *p++;
     x ^= Py_SIZE(a);
+    x ^= _Py_HashSecret.suffix;
     if (x == -1)
         x = -2;
     a->ob_shash = x;
diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c
index d8dab67..2f80e59 100644
--- a/Objects/unicodeobject.c
+++ b/Objects/unicodeobject.c
@@ -6695,11 +6695,21 @@ unicode_hash(PyUnicodeObject *self)
     if (self->hash != -1)
         return self->hash;
     len = PyUnicode_GET_SIZE(self);
+    /*
+      We make the hash of the empty string be 0, rather than using
+      (prefix ^ suffix), since this slightly obfuscates the hash secret
+    */
+    if (len == 0) {
+        self->hash = 0;
+        return 0;
+    }
     p = PyUnicode_AS_UNICODE(self);
-    x = *p << 7;
+    x = _Py_HashSecret.prefix;
+    x ^= *p << 7;
     while (--len >= 0)
         x = (1000003*x) ^ *p++;
     x ^= PyUnicode_GET_SIZE(self);
+    x ^= _Py_HashSecret.suffix;
     if (x == -1)
         x = -2;
     self->hash = x;