[Python-checkins] r65335 - in python/trunk: Lib/test/seq_tests.py Lib/test/test_bigmem.py Lib/test/test_strop.py Lib/test/test_support.py Misc/NEWS Modules/gcmodule.c Modules/mmapmodule.c Modules/stropmodule.c Objects/bufferobject.c Objects/longobject.c Objects/stringobject.c Objects/tupleobject.c Objects/unicodeobject.c
neal.norwitz
python-checkins at python.org
Thu Jul 31 19:17:15 CEST 2008
Author: neal.norwitz
Date: Thu Jul 31 19:17:14 2008
New Revision: 65335
Log:
Security patches from Apple: prevent int overflow when allocating memory
Modified:
python/trunk/Lib/test/seq_tests.py
python/trunk/Lib/test/test_bigmem.py
python/trunk/Lib/test/test_strop.py
python/trunk/Lib/test/test_support.py
python/trunk/Misc/NEWS
python/trunk/Modules/gcmodule.c
python/trunk/Modules/mmapmodule.c
python/trunk/Modules/stropmodule.c
python/trunk/Objects/bufferobject.c
python/trunk/Objects/longobject.c
python/trunk/Objects/stringobject.c
python/trunk/Objects/tupleobject.c
python/trunk/Objects/unicodeobject.c
Modified: python/trunk/Lib/test/seq_tests.py
==============================================================================
--- python/trunk/Lib/test/seq_tests.py (original)
+++ python/trunk/Lib/test/seq_tests.py Thu Jul 31 19:17:14 2008
@@ -306,11 +306,13 @@
self.assertEqual(id(s), id(s*1))
def test_bigrepeat(self):
- x = self.type2test([0])
- x *= 2**16
- self.assertRaises(MemoryError, x.__mul__, 2**16)
- if hasattr(x, '__imul__'):
- self.assertRaises(MemoryError, x.__imul__, 2**16)
+ import sys
+ if sys.maxint <= 2147483647:
+ x = self.type2test([0])
+ x *= 2**16
+ self.assertRaises(MemoryError, x.__mul__, 2**16)
+ if hasattr(x, '__imul__'):
+ self.assertRaises(MemoryError, x.__imul__, 2**16)
def test_subscript(self):
a = self.type2test([10, 11])
Modified: python/trunk/Lib/test/test_bigmem.py
==============================================================================
--- python/trunk/Lib/test/test_bigmem.py (original)
+++ python/trunk/Lib/test/test_bigmem.py Thu Jul 31 19:17:14 2008
@@ -1,5 +1,5 @@
from test import test_support
-from test.test_support import bigmemtest, _1G, _2G
+from test.test_support import bigmemtest, _1G, _2G, _4G, precisionbigmemtest
import unittest
import operator
@@ -54,6 +54,22 @@
self.assertEquals(s[lpadsize:-rpadsize], SUBSTR)
self.assertEquals(s.strip(), SUBSTR.strip())
+ @precisionbigmemtest(size=_2G - 1, memuse=1)
+ def test_center_unicode(self, size):
+ SUBSTR = u' abc def ghi'
+ try:
+ s = SUBSTR.center(size)
+ except OverflowError:
+ pass # acceptable on 32-bit
+ else:
+ self.assertEquals(len(s), size)
+ lpadsize = rpadsize = (len(s) - len(SUBSTR)) // 2
+ if len(s) % 2:
+ lpadsize += 1
+ self.assertEquals(s[lpadsize:-rpadsize], SUBSTR)
+ self.assertEquals(s.strip(), SUBSTR.strip())
+ del s
+
@bigmemtest(minsize=_2G, memuse=2)
def test_count(self, size):
SUBSTR = ' abc def ghi'
@@ -70,10 +86,51 @@
s = '.' * size
self.assertEquals(len(s.decode('utf-8')), size)
+ def basic_encode_test(self, size, enc, c=u'.', expectedsize=None):
+ if expectedsize is None:
+ expectedsize = size
+
+ s = c * size
+ self.assertEquals(len(s.encode(enc)), expectedsize)
+
@bigmemtest(minsize=_2G + 2, memuse=3)
def test_encode(self, size):
- s = u'.' * size
- self.assertEquals(len(s.encode('utf-8')), size)
+ return self.basic_encode_test(size, 'utf-8')
+
+ @precisionbigmemtest(size=_4G / 6 + 2, memuse=2)
+ def test_encode_raw_unicode_escape(self, size):
+ try:
+ return self.basic_encode_test(size, 'raw_unicode_escape')
+ except MemoryError:
+ pass # acceptable on 32-bit
+
+ @precisionbigmemtest(size=_4G / 5 + 70, memuse=3)
+ def test_encode_utf7(self, size):
+ try:
+ return self.basic_encode_test(size, 'utf7')
+ except MemoryError:
+ pass # acceptable on 32-bit
+
+ @precisionbigmemtest(size=_4G / 4 + 5, memuse=6)
+ def test_encode_utf32(self, size):
+ try:
+ return self.basic_encode_test(size, 'utf32', expectedsize=4*size+4)
+ except MemoryError:
+ pass # acceptable on 32-bit
+
+ @precisionbigmemtest(size=_2G-1, memuse=2)
+ def test_decodeascii(self, size):
+ return self.basic_encode_test(size, 'ascii', c='A')
+
+ @precisionbigmemtest(size=_4G / 5, memuse=6+2)
+ def test_unicode_repr_oflw(self, size):
+ try:
+ s = u"\uAAAA"*size
+ r = repr(s)
+ except MemoryError:
+ pass # acceptable on 32-bit
+ else:
+ self.failUnless(s == eval(r))
@bigmemtest(minsize=_2G, memuse=2)
def test_endswith(self, size):
@@ -459,6 +516,11 @@
self.assertEquals(s.count('\\'), size)
self.assertEquals(s.count('0'), size * 2)
+ @bigmemtest(minsize=2**32 / 5, memuse=6+2)
+ def test_unicode_repr(self, size):
+ s = u"\uAAAA" * size
+ self.failUnless(len(repr(s)) > size)
+
# This test is meaningful even with size < 2G, as long as the
# doubled string is > 2G (but it tests more if both are > 2G :)
@bigmemtest(minsize=_1G + 2, memuse=3)
@@ -642,6 +704,35 @@
def test_repeat_large(self, size):
return self.basic_test_repeat(size)
+ @bigmemtest(minsize=_1G - 1, memuse=12)
+ def test_repeat_large_2(self, size):
+ return self.basic_test_repeat(size)
+
+ @precisionbigmemtest(size=_1G - 1, memuse=9)
+ def test_from_2G_generator(self, size):
+ try:
+ t = tuple(xrange(size))
+ except MemoryError:
+ pass # acceptable on 32-bit
+ else:
+ count = 0
+ for item in t:
+ self.assertEquals(item, count)
+ count += 1
+ self.assertEquals(count, size)
+
+ @precisionbigmemtest(size=_1G - 25, memuse=9)
+ def test_from_almost_2G_generator(self, size):
+ try:
+ t = tuple(xrange(size))
+ count = 0
+ for item in t:
+ self.assertEquals(item, count)
+ count += 1
+ self.assertEquals(count, size)
+ except MemoryError:
+ pass # acceptable, expected on 32-bit
+
# Like test_concat, split in two.
def basic_test_repr(self, size):
t = (0,) * size
@@ -957,8 +1048,23 @@
self.assertEquals(l[:10], [1] * 10)
self.assertEquals(l[-10:], [5] * 10)
+class BufferTest(unittest.TestCase):
+
+ @precisionbigmemtest(size=_1G, memuse=4)
+ def test_repeat(self, size):
+ try:
+ b = buffer("AAAA")*size
+ except MemoryError:
+ pass # acceptable on 32-bit
+ else:
+ count = 0
+ for c in b:
+ self.assertEquals(c, 'A')
+ count += 1
+ self.assertEquals(count, size*4)
+
def test_main():
- test_support.run_unittest(StrTest, TupleTest, ListTest)
+ test_support.run_unittest(StrTest, TupleTest, ListTest, BufferTest)
if __name__ == '__main__':
if len(sys.argv) > 1:
Modified: python/trunk/Lib/test/test_strop.py
==============================================================================
--- python/trunk/Lib/test/test_strop.py (original)
+++ python/trunk/Lib/test/test_strop.py Thu Jul 31 19:17:14 2008
@@ -115,6 +115,25 @@
strop.uppercase
strop.whitespace
+ @test_support.precisionbigmemtest(size=test_support._2G - 1, memuse=5)
+ def test_stropjoin_huge_list(self, size):
+ a = "A" * size
+ try:
+ r = strop.join([a, a], a)
+ except OverflowError:
+ pass
+ else:
+ self.assertEquals(len(r), len(a) * 3)
+
+ @test_support.precisionbigmemtest(size=test_support._2G - 1, memuse=1)
+ def test_stropjoin_huge_tup(self, size):
+ a = "A" * size
+ try:
+ r = strop.join((a, a), a)
+ except OverflowError:
+ pass # acceptable on 32-bit
+ else:
+ self.assertEquals(len(r), len(a) * 3)
transtable = '\000\001\002\003\004\005\006\007\010\011\012\013\014\015\016\017\020\021\022\023\024\025\026\027\030\031\032\033\034\035\036\037 !"#$%&\'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`xyzdefghijklmnopqrstuvwxyz{|}~\177\200\201\202\203\204\205\206\207\210\211\212\213\214\215\216\217\220\221\222\223\224\225\226\227\230\231\232\233\234\235\236\237\240\241\242\243\244\245\246\247\250\251\252\253\254\255\256\257\260\261\262\263\264\265\266\267\270\271\272\273\274\275\276\277\300\301\302\303\304\305\306\307\310\311\312\313\314\315\316\317\320\321\322\323\324\325\326\327\330\331\332\333\334\335\336\337\340\341\342\343\344\345\346\347\350\351\352\353\354\355\356\357\360\361\362\363\364\365\366\367\370\371\372\373\374\375\376\377'
Modified: python/trunk/Lib/test/test_support.py
==============================================================================
--- python/trunk/Lib/test/test_support.py (original)
+++ python/trunk/Lib/test/test_support.py Thu Jul 31 19:17:14 2008
@@ -67,6 +67,7 @@
use_resources = None # Flag set to [] by regrtest.py
max_memuse = 0 # Disable bigmem tests (they will still be run with
# small sizes, to make sure they work.)
+real_max_memuse = 0
# _original_stdout is meant to hold stdout at the time regrtest began.
# This may be "the real" stdout, or IDLE's emulation of stdout, or whatever.
@@ -611,12 +612,14 @@
_1M = 1024*1024
_1G = 1024 * _1M
_2G = 2 * _1G
+_4G = 4 * _1G
MAX_Py_ssize_t = sys.maxsize
def set_memlimit(limit):
import re
global max_memuse
+ global real_max_memuse
sizes = {
'k': 1024,
'm': _1M,
@@ -628,6 +631,7 @@
if m is None:
raise ValueError('Invalid memory limit %r' % (limit,))
memlimit = int(float(m.group(1)) * sizes[m.group(3).lower()])
+ real_max_memuse = memlimit
if memlimit > MAX_Py_ssize_t:
memlimit = MAX_Py_ssize_t
if memlimit < _2G - 1:
@@ -673,6 +677,27 @@
return wrapper
return decorator
+def precisionbigmemtest(size, memuse, overhead=5*_1M):
+ def decorator(f):
+ def wrapper(self):
+ if not real_max_memuse:
+ maxsize = 5147
+ else:
+ maxsize = size
+
+ if real_max_memuse and real_max_memuse < maxsize * memuse:
+ if verbose:
+ sys.stderr.write("Skipping %s because of memory "
+ "constraint\n" % (f.__name__,))
+ return
+
+ return f(self, maxsize)
+ wrapper.size = size
+ wrapper.memuse = memuse
+ wrapper.overhead = overhead
+ return wrapper
+ return decorator
+
def bigaddrspacetest(f):
"""Decorator for tests that fill the address space."""
def wrapper(self):
Modified: python/trunk/Misc/NEWS
==============================================================================
--- python/trunk/Misc/NEWS (original)
+++ python/trunk/Misc/NEWS Thu Jul 31 19:17:14 2008
@@ -12,6 +12,8 @@
Core and Builtins
-----------------
+- Apply security patches from Apple.
+
- Issue #2542: Now that issubclass() may call arbitrary code, ensure that
PyErr_ExceptionMatches returns 0 when an exception occurs there.
Modified: python/trunk/Modules/gcmodule.c
==============================================================================
--- python/trunk/Modules/gcmodule.c (original)
+++ python/trunk/Modules/gcmodule.c Thu Jul 31 19:17:14 2008
@@ -1342,7 +1342,10 @@
_PyObject_GC_Malloc(size_t basicsize)
{
PyObject *op;
- PyGC_Head *g = (PyGC_Head *)PyObject_MALLOC(
+ PyGC_Head *g;
+ if (basicsize > PY_SSIZE_T_MAX - sizeof(PyGC_Head))
+ return PyErr_NoMemory();
+ g = (PyGC_Head *)PyObject_MALLOC(
sizeof(PyGC_Head) + basicsize);
if (g == NULL)
return PyErr_NoMemory();
@@ -1385,6 +1388,8 @@
{
const size_t basicsize = _PyObject_VAR_SIZE(Py_TYPE(op), nitems);
PyGC_Head *g = AS_GC(op);
+ if (basicsize > PY_SSIZE_T_MAX - sizeof(PyGC_Head))
+ return (PyVarObject *)PyErr_NoMemory();
g = (PyGC_Head *)PyObject_REALLOC(g, sizeof(PyGC_Head) + basicsize);
if (g == NULL)
return (PyVarObject *)PyErr_NoMemory();
Modified: python/trunk/Modules/mmapmodule.c
==============================================================================
--- python/trunk/Modules/mmapmodule.c (original)
+++ python/trunk/Modules/mmapmodule.c Thu Jul 31 19:17:14 2008
@@ -239,7 +239,7 @@
return(NULL);
/* silently 'adjust' out-of-range requests */
- if ((self->pos + num_bytes) > self->size) {
+ if (num_bytes > self->size - self->pos) {
num_bytes -= (self->pos+num_bytes) - self->size;
}
result = Py_BuildValue("s#", self->data+self->pos, num_bytes);
Modified: python/trunk/Modules/stropmodule.c
==============================================================================
--- python/trunk/Modules/stropmodule.c (original)
+++ python/trunk/Modules/stropmodule.c Thu Jul 31 19:17:14 2008
@@ -216,6 +216,13 @@
return NULL;
}
slen = PyString_GET_SIZE(item);
+ if (slen > PY_SSIZE_T_MAX - reslen ||
+ seplen > PY_SSIZE_T_MAX - reslen - seplen) {
+ PyErr_SetString(PyExc_OverflowError,
+ "input too long");
+ Py_DECREF(res);
+ return NULL;
+ }
while (reslen + slen + seplen >= sz) {
if (_PyString_Resize(&res, sz * 2) < 0)
return NULL;
@@ -253,6 +260,14 @@
return NULL;
}
slen = PyString_GET_SIZE(item);
+ if (slen > PY_SSIZE_T_MAX - reslen ||
+ seplen > PY_SSIZE_T_MAX - reslen - seplen) {
+ PyErr_SetString(PyExc_OverflowError,
+ "input too long");
+ Py_DECREF(res);
+ Py_XDECREF(item);
+ return NULL;
+ }
while (reslen + slen + seplen >= sz) {
if (_PyString_Resize(&res, sz * 2) < 0) {
Py_DECREF(item);
Modified: python/trunk/Objects/bufferobject.c
==============================================================================
--- python/trunk/Objects/bufferobject.c (original)
+++ python/trunk/Objects/bufferobject.c Thu Jul 31 19:17:14 2008
@@ -431,6 +431,10 @@
count = 0;
if (!get_buf(self, &ptr, &size, ANY_BUFFER))
return NULL;
+ if (count > PY_SSIZE_T_MAX / size) {
+ PyErr_SetString(PyExc_MemoryError, "result too large");
+ return NULL;
+ }
ob = PyString_FromStringAndSize(NULL, size * count);
if ( ob == NULL )
return NULL;
Modified: python/trunk/Objects/longobject.c
==============================================================================
--- python/trunk/Objects/longobject.c (original)
+++ python/trunk/Objects/longobject.c Thu Jul 31 19:17:14 2008
@@ -70,6 +70,8 @@
return NULL;
}
/* coverity[ampersand_in_size] */
+ /* XXX(nnorwitz): This can overflow --
+ PyObject_NEW_VAR / _PyObject_VAR_SIZE need to detect overflow */
return PyObject_NEW_VAR(PyLongObject, &PyLong_Type, size);
}
Modified: python/trunk/Objects/stringobject.c
==============================================================================
--- python/trunk/Objects/stringobject.c (original)
+++ python/trunk/Objects/stringobject.c Thu Jul 31 19:17:14 2008
@@ -74,6 +74,11 @@
return (PyObject *)op;
}
+ if (size > PY_SSIZE_T_MAX - sizeof(PyStringObject)) {
+ PyErr_SetString(PyExc_OverflowError, "string is too large");
+ return NULL;
+ }
+
/* Inline PyObject_NewVar */
op = (PyStringObject *)PyObject_MALLOC(sizeof(PyStringObject) + size);
if (op == NULL)
@@ -109,7 +114,7 @@
assert(str != NULL);
size = strlen(str);
- if (size > PY_SSIZE_T_MAX) {
+ if (size > PY_SSIZE_T_MAX - sizeof(PyStringObject)) {
PyErr_SetString(PyExc_OverflowError,
"string is too long for a Python string");
return NULL;
@@ -977,13 +982,23 @@
return (PyObject *)a;
}
size = Py_SIZE(a) + Py_SIZE(b);
- if (size < 0) {
+ /* Check that string sizes are not negative, to prevent an
+ overflow in cases where we are passed incorrectly-created
+ strings with negative lengths (due to a bug in other code).
+ */
+ if (Py_SIZE(a) < 0 || Py_SIZE(b) < 0 ||
+ Py_SIZE(a) > PY_SSIZE_T_MAX - Py_SIZE(b)) {
PyErr_SetString(PyExc_OverflowError,
"strings are too large to concat");
return NULL;
}
/* Inline PyObject_NewVar */
+ if (size > PY_SSIZE_T_MAX - sizeof(PyStringObject)) {
+ PyErr_SetString(PyExc_OverflowError,
+ "strings are too large to concat");
+ return NULL;
+ }
op = (PyStringObject *)PyObject_MALLOC(sizeof(PyStringObject) + size);
if (op == NULL)
return PyErr_NoMemory();
Modified: python/trunk/Objects/tupleobject.c
==============================================================================
--- python/trunk/Objects/tupleobject.c (original)
+++ python/trunk/Objects/tupleobject.c Thu Jul 31 19:17:14 2008
@@ -60,11 +60,12 @@
Py_ssize_t nbytes = size * sizeof(PyObject *);
/* Check for overflow */
if (nbytes / sizeof(PyObject *) != (size_t)size ||
- (nbytes += sizeof(PyTupleObject) - sizeof(PyObject *))
- <= 0)
+ (nbytes > PY_SSIZE_T_MAX - sizeof(PyTupleObject) - sizeof(PyObject *)))
{
return PyErr_NoMemory();
}
+ nbytes += sizeof(PyTupleObject) - sizeof(PyObject *);
+
op = PyObject_GC_NewVar(PyTupleObject, &PyTuple_Type, size);
if (op == NULL)
return NULL;
Modified: python/trunk/Objects/unicodeobject.c
==============================================================================
--- python/trunk/Objects/unicodeobject.c (original)
+++ python/trunk/Objects/unicodeobject.c Thu Jul 31 19:17:14 2008
@@ -299,6 +299,11 @@
return unicode_empty;
}
+ /* Ensure we won't overflow the size. */
+ if (length > ((PY_SSIZE_T_MAX / sizeof(Py_UNICODE)) - 1)) {
+ return (PyUnicodeObject *)PyErr_NoMemory();
+ }
+
/* Unicode freelist & memory allocation */
if (free_list) {
unicode = free_list;
@@ -1651,6 +1656,9 @@
char * out;
char * start;
+ if (cbAllocated / 5 != size)
+ return PyErr_NoMemory();
+
if (size == 0)
return PyString_FromStringAndSize(NULL, 0);
@@ -2245,8 +2253,9 @@
{
PyObject *v;
unsigned char *p;
+ Py_ssize_t nsize, bytesize;
#ifndef Py_UNICODE_WIDE
- int i, pairs;
+ Py_ssize_t i, pairs;
#else
const int pairs = 0;
#endif
@@ -2274,8 +2283,11 @@
0xDC00 <= s[i+1] && s[i+1] <= 0xDFFF)
pairs++;
#endif
- v = PyString_FromStringAndSize(NULL,
- 4 * (size - pairs + (byteorder == 0)));
+ nsize = (size - pairs + (byteorder == 0));
+ bytesize = nsize * 4;
+ if (bytesize / 4 != nsize)
+ return PyErr_NoMemory();
+ v = PyString_FromStringAndSize(NULL, bytesize);
if (v == NULL)
return NULL;
@@ -2515,8 +2527,9 @@
{
PyObject *v;
unsigned char *p;
+ Py_ssize_t nsize, bytesize;
#ifdef Py_UNICODE_WIDE
- int i, pairs;
+ Py_ssize_t i, pairs;
#else
const int pairs = 0;
#endif
@@ -2539,8 +2552,15 @@
if (s[i] >= 0x10000)
pairs++;
#endif
- v = PyString_FromStringAndSize(NULL,
- 2 * (size + pairs + (byteorder == 0)));
+ /* 2 * (size + pairs + (byteorder == 0)) */
+ if (size > PY_SSIZE_T_MAX ||
+ size > PY_SSIZE_T_MAX - pairs - (byteorder == 0))
+ return PyErr_NoMemory();
+ nsize = size + pairs + (byteorder == 0);
+ bytesize = nsize * 2;
+ if (bytesize / 2 != nsize)
+ return PyErr_NoMemory();
+ v = PyString_FromStringAndSize(NULL, bytesize);
if (v == NULL)
return NULL;
@@ -2868,6 +2888,11 @@
char *p;
static const char *hexdigit = "0123456789abcdef";
+#ifdef Py_UNICODE_WIDE
+ const Py_ssize_t expandsize = 10;
+#else
+ const Py_ssize_t expandsize = 6;
+#endif
/* XXX(nnorwitz): rather than over-allocating, it would be
better to choose a different scheme. Perhaps scan the
@@ -2887,13 +2912,12 @@
escape.
*/
+ if (size > (PY_SSIZE_T_MAX - 2 - 1) / expandsize)
+ return PyErr_NoMemory();
+
repr = PyString_FromStringAndSize(NULL,
2
-#ifdef Py_UNICODE_WIDE
- + 10*size
-#else
- + 6*size
-#endif
+ + expandsize*size
+ 1);
if (repr == NULL)
return NULL;
@@ -3146,12 +3170,16 @@
char *q;
static const char *hexdigit = "0123456789abcdef";
-
#ifdef Py_UNICODE_WIDE
- repr = PyString_FromStringAndSize(NULL, 10 * size);
+ const Py_ssize_t expandsize = 10;
#else
- repr = PyString_FromStringAndSize(NULL, 6 * size);
+ const Py_ssize_t expandsize = 6;
#endif
+
+ if (size > PY_SSIZE_T_MAX / expandsize)
+ return PyErr_NoMemory();
+
+ repr = PyString_FromStringAndSize(NULL, expandsize * size);
if (repr == NULL)
return NULL;
if (size == 0)
@@ -5574,6 +5602,11 @@
return self;
}
+ if (left > PY_SSIZE_T_MAX - self->length ||
+ right > PY_SSIZE_T_MAX - (left + self->length)) {
+ PyErr_SetString(PyExc_OverflowError, "padded string is too long");
+ return NULL;
+ }
u = _PyUnicode_New(left + self->length + right);
if (u) {
if (left)
More information about the Python-checkins
mailing list