vhost: make sure log_num < in_num
authoryongduan <yongduan@tencent.com>
Wed, 11 Sep 2019 09:44:24 +0000 (17:44 +0800)
committerMichael S. Tsirkin <mst@redhat.com>
Wed, 11 Sep 2019 19:15:26 +0000 (15:15 -0400)
The code assumes log_num < in_num everywhere, and that is true as long as
in_num is incremented by descriptor iov count, and log_num by 1. However
this breaks if there's a zero sized descriptor.

As a result, if a malicious guest creates a vring desc with desc.len = 0,
it may cause the host kernel to crash by overflowing the log array. This
bug can be triggered during the VM migration.

There's no need to log when desc.len = 0, so just don't increment log_num
in this case.

Fixes: 3a4d5c94e959 ("vhost_net: a kernel-level virtio server")
Cc: stable@vger.kernel.org
Reviewed-by: Lidong Chen <lidongchen@tencent.com>
Signed-off-by: ruippan <ruippan@tencent.com>
Signed-off-by: yongduan <yongduan@tencent.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Tyler Hicks <tyhicks@canonical.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
drivers/vhost/vhost.c

index 34ea219936e3f369ff77c52f073d9d6a1c051128..acabf20b069efabc216f33477ea8f77b588450e1 100644 (file)
@@ -2180,7 +2180,7 @@ static int get_indirect(struct vhost_virtqueue *vq,
                /* If this is an input descriptor, increment that count. */
                if (access == VHOST_ACCESS_WO) {
                        *in_num += ret;
-                       if (unlikely(log)) {
+                       if (unlikely(log && ret)) {
                                log[*log_num].addr = vhost64_to_cpu(vq, desc.addr);
                                log[*log_num].len = vhost32_to_cpu(vq, desc.len);
                                ++*log_num;
@@ -2321,7 +2321,7 @@ int vhost_get_vq_desc(struct vhost_virtqueue *vq,
                        /* If this is an input descriptor,
                         * increment that count. */
                        *in_num += ret;
-                       if (unlikely(log)) {
+                       if (unlikely(log && ret)) {
                                log[*log_num].addr = vhost64_to_cpu(vq, desc.addr);
                                log[*log_num].len = vhost32_to_cpu(vq, desc.len);
                                ++*log_num;