On Wed, 2008-12-24 at 12:11 +0100, Aurelien Jarno wrote:
> Hi,
> 
> qemu 0.9.1-9 fixes a remote DoS, please find the diff below. Could you
> please unblock it?

The same fix has also been applied to kvm in version kvm_72+dfsg-4.
Please unblock it, too.

Thanks,
Jan

diff -u kvm-72+dfsg/debian/changelog kvm-72+dfsg/debian/changelog
--- kvm-72+dfsg/debian/changelog
+++ kvm-72+dfsg/debian/changelog
@@ -1,3 +1,10 @@
+kvm (72+dfsg-4) unstable; urgency=high
+
+  * debian/patches/core-2008-1210.patch: fix remote DoS via VNC
+    (CORE-2008-1210/CVE-2008-2382).
+
+ -- Jan Lübbe <[email protected]>  Wed, 24 Dec 2008 12:23:06 +0100
+
 kvm (72+dfsg-3) unstable; urgency=medium
 
   * Apply patch from qemu (62_fix-ptyblocking.patch) which fixes a lockup
diff -u kvm-72+dfsg/debian/patches/series kvm-72+dfsg/debian/patches/series
--- kvm-72+dfsg/debian/patches/series
+++ kvm-72+dfsg/debian/patches/series
@@ -16,0 +17 @@
+core-2008-1210.patch
only in patch2:
unchanged:
--- kvm-72+dfsg.orig/debian/patches/core-2008-1210.patch
+++ kvm-72+dfsg/debian/patches/core-2008-1210.patch
@@ -0,0 +1,27 @@
+Fix CORE-2008-1210 VNC DoS
+
+If the client sends us a limit of zero, handle appropriately.
+
+Signed-off-by: Anthony Liguori <[email protected]>
+
+diff --git qemu/vnc.c qemu/vnc.c
+index 3a7d762..575fd68 100644
+--- a/qemu/vnc.c
++++ b/qemu/vnc.c
+@@ -1503,10 +1503,13 @@ static int protocol_client_msg(VncState *vs, uint8_t 
*data, size_t len)
+       if (len == 1)
+           return 4;
+ 
+-      if (len == 4)
+-          return 4 + (read_u16(data, 2) * 4);
++      if (len == 4) {
++            limit = read_u16(data, 2);
++            if (limit > 0)
++                return 4 + (limit * 4);
++        } else
++            limit = read_u16(data, 2);
+ 
+-      limit = read_u16(data, 2);
+       for (i = 0; i < limit; i++) {
+           int32_t val = read_s32(data, 4 + (i * 4));
+           memcpy(data + 4 + (i * 4), &val, sizeof(val));



-- 
To UNSUBSCRIBE, email to [email protected]
with a subject of "unsubscribe". Trouble? Contact [email protected]

Reply via email to