summaryrefslogtreecommitdiff
path: root/glib
diff options
context:
space:
mode:
authorMatthias Clasen <mclasen@redhat.com>2009-03-12 13:30:55 +0000
committerMatthias Clasen <matthiasc@src.gnome.org>2009-03-12 13:30:55 +0000
commit2ad1af5b1ee5934a5d66be506b85157facc73a46 (patch)
tree7d904db459e61d2105310c3a012648cffba040df /glib
parentbe600f638f9807497e2932f7b5661bb12a20bbe2 (diff)
downloadglib-2ad1af5b1ee5934a5d66be506b85157facc73a46.tar.gz
Avoid integer overflows in the base64 functions. Fixes CVE-2008-4316
2009-03-12 Matthias Clasen <mclasen@redhat.com> * glib/gbase64.c: Avoid integer overflows in the base64 functions. Fixes CVE-2008-4316 svn path=/trunk/; revision=7973
Diffstat (limited to 'glib')
-rw-r--r--glib/gbase64.c25
1 files changed, 18 insertions, 7 deletions
diff --git a/glib/gbase64.c b/glib/gbase64.c
index 99adb85e7..bfdc76aa5 100644
--- a/glib/gbase64.c
+++ b/glib/gbase64.c
@@ -54,8 +54,9 @@ static const char base64_alphabet[] =
*
* The output buffer must be large enough to fit all the data that will
* be written to it. Due to the way base64 encodes you will need
- * at least: @len * 4 / 3 + 6 bytes. If you enable line-breaking you will
- * need at least: @len * 4 / 3 + @len * 4 / (3 * 72) + 7 bytes.
+ * at least: (@len / 3 + 1) * 4 + 4 bytes (+ 4 may be needed in case of
+ * non-zero state). If you enable line-breaking you will need at least:
+ * ((@len / 3 + 1) * 4 + 4) / 72 + 1 bytes of extra space.
*
* @break_lines is typically used when putting base64-encoded data in emails.
* It breaks the lines at 72 columns instead of putting all of the text on
@@ -233,8 +234,14 @@ g_base64_encode (const guchar *data,
g_return_val_if_fail (data != NULL, NULL);
g_return_val_if_fail (len > 0, NULL);
- /* We can use a smaller limit here, since we know the saved state is 0 */
- out = g_malloc (len * 4 / 3 + 4);
+ /* We can use a smaller limit here, since we know the saved state is 0,
+ +1 is needed for trailing \0, also check for unlikely integer overflow */
+ if (len >= ((G_MAXSIZE - 1) / 4 - 1) * 3)
+ g_error("%s: input too large for Base64 encoding (%"G_GSIZE_FORMAT" chars)",
+ G_STRLOC, len);
+
+ out = g_malloc ((len / 3 + 1) * 4 + 1);
+
outlen = g_base64_encode_step (data, len, FALSE, out, &state, &save);
outlen += g_base64_encode_close (FALSE, out + outlen, &state, &save);
out[outlen] = '\0';
@@ -275,7 +282,8 @@ static const unsigned char mime_base64_rank[256] = {
*
* The output buffer must be large enough to fit all the data that will
* be written to it. Since base64 encodes 3 bytes in 4 chars you need
- * at least: @len * 3 / 4 bytes.
+ * at least: (@len / 4) * 3 + 3 bytes (+ 3 may be needed in case of non-zero
+ * state).
*
* Return value: The number of bytes of output that was written
*
@@ -358,7 +366,8 @@ g_base64_decode (const gchar *text,
gsize *out_len)
{
guchar *ret;
- gint input_length, state = 0;
+ gsize input_length;
+ gint state = 0;
guint save = 0;
g_return_val_if_fail (text != NULL, NULL);
@@ -368,7 +377,9 @@ g_base64_decode (const gchar *text,
g_return_val_if_fail (input_length > 1, NULL);
- ret = g_malloc0 (input_length * 3 / 4);
+ /* We can use a smaller limit here, since we know the saved state is 0,
+ +1 used to avoid calling g_malloc0(0), and hence retruning NULL */
+ ret = g_malloc0 ((input_length / 4) * 3 + 1);
*out_len = g_base64_decode_step (text, input_length, ret, &state, &save);