From 9fc3d1245bec49e29013b8120340e87adeaaf11a Mon Sep 17 00:00:00 2001
From: Maxime Coquelin <maxime.coquelin@redhat.com>
Date: Mon, 23 Apr 2018 11:33:42 +0200
Subject: [PATCH 05/11] vhost: add support for non-contiguous indirect descs
 tables

This patch adds support for non-contiguous indirect descriptor
tables in VA space.

When it happens, which is unlikely, a table is allocated and the
non-contiguous content is copied into it.

This issue has been assigned CVE-2018-1059.

Reported-by: Yongji Xie <xieyongji@baidu.com>
Signed-off-by: Maxime Coquelin <maxime.coquelin@redhat.com>
---
 lib/librte_vhost/virtio_net.c | 108 +++++++++++++++++++++++++++++++++++++++---
 1 file changed, 101 insertions(+), 7 deletions(-)

diff --git a/lib/librte_vhost/virtio_net.c b/lib/librte_vhost/virtio_net.c
index 79bac59..13252e6 100644
--- a/lib/librte_vhost/virtio_net.c
+++ b/lib/librte_vhost/virtio_net.c
@@ -46,4 +46,5 @@
 #include <rte_arp.h>
 #include <rte_spinlock.h>
+#include <rte_malloc.h>
 
 #include "iotlb.h"
@@ -60,4 +61,44 @@
 }
 
+static __rte_always_inline struct vring_desc *
+alloc_copy_ind_table(struct virtio_net *dev, struct vhost_virtqueue *vq,
+					 struct vring_desc *desc)
+{
+	struct vring_desc *idesc;
+	uint64_t src, dst;
+	uint64_t len, remain = desc->len;
+	uint64_t desc_addr = desc->addr;
+
+	idesc = rte_malloc(__func__, desc->len, 0);
+	if (unlikely(!idesc))
+		return 0;
+
+	dst = (uint64_t)(uintptr_t)idesc;
+
+	while (remain) {
+		len = remain;
+		src = vhost_iova_to_vva(dev, vq, desc_addr, &len,
+				VHOST_ACCESS_RO);
+		if (unlikely(!src || !len)) {
+			rte_free(idesc);
+			return 0;
+		}
+
+		rte_memcpy((void *)(uintptr_t)dst, (void *)(uintptr_t)src, len);
+
+		remain -= len;
+		dst += len;
+		desc_addr += len;
+	}
+
+	return idesc;
+}
+
+static __rte_always_inline void
+free_ind_table(struct vring_desc *idesc)
+{
+	rte_free(idesc);
+}
+
 static __rte_always_inline void
 do_flush_shadow_used_ring(struct virtio_net *dev, struct vhost_virtqueue *vq,
@@ -376,4 +417,5 @@
 	rte_prefetch0(&vq->desc[desc_indexes[0]]);
 	for (i = 0; i < count; i++) {
+		struct vring_desc *idesc = NULL;
 		uint16_t desc_idx = desc_indexes[i];
 		int err;
@@ -385,10 +427,22 @@
 						vq, vq->desc[desc_idx].addr,
 						&dlen, VHOST_ACCESS_RO);
-			if (unlikely(!descs ||
-					dlen != vq->desc[desc_idx].len)) {
+			if (unlikely(!descs)) {
 				count = i;
 				break;
 			}
 
+			if (unlikely(dlen < vq->desc[desc_idx].len)) {
+				/*
+				 * The indirect desc table is not contiguous
+				 * in process VA space, we have to copy it.
+				 */
+				idesc = alloc_copy_ind_table(dev, vq,
+							&vq->desc[desc_idx]);
+				if (unlikely(!idesc))
+					break;
+
+				descs = idesc;
+			}
+
 			desc_idx = 0;
 			sz = vq->desc[desc_idx].len / sizeof(*descs);
@@ -401,4 +455,5 @@
 		if (unlikely(err)) {
 			count = i;
+			free_ind_table(idesc);
 			break;
 		}
@@ -406,4 +461,7 @@
 		if (i + 1 < count)
 			rte_prefetch0(&vq->desc[desc_indexes[i+1]]);
+
+		if (unlikely(!!idesc))
+			free_ind_table(idesc);
 	}
 
@@ -446,4 +504,5 @@
 	uint64_t dlen;
 	struct vring_desc *descs = vq->desc;
+	struct vring_desc *idesc = NULL;
 
 	*desc_chain_head = idx;
@@ -455,13 +514,27 @@
 						&dlen,
 						VHOST_ACCESS_RO);
-		if (unlikely(!descs || dlen != vq->desc[idx].len))
+		if (unlikely(!descs))
 			return -1;
 
+		if (unlikely(dlen < vq->desc[idx].len)) {
+			/*
+			 * The indirect desc table is not contiguous
+			 * in process VA space, we have to copy it.
+			 */
+			idesc = alloc_copy_ind_table(dev, vq, &vq->desc[idx]);
+			if (unlikely(!idesc))
+				return -1;
+
+			descs = idesc;
+		}
+
 		idx = 0;
 	}
 
 	while (1) {
-		if (unlikely(vec_id >= BUF_VECTOR_MAX || idx >= vq->size))
+		if (unlikely(vec_id >= BUF_VECTOR_MAX || idx >= vq->size)) {
+			free_ind_table(idesc);
 			return -1;
+		}
 
 		len += descs[idx].len;
@@ -480,4 +553,7 @@
 	*vec_idx = vec_id;
 
+	if (unlikely(!!idesc))
+		free_ind_table(idesc);
+
 	return 0;
 }
@@ -1333,5 +1409,5 @@
 	rte_prefetch0(&vq->desc[desc_indexes[0]]);
 	for (i = 0; i < count; i++) {
-		struct vring_desc *desc;
+		struct vring_desc *desc, *idesc = NULL;
 		uint16_t sz, idx;
 		uint64_t dlen;
@@ -1348,8 +1424,20 @@
 						&dlen,
 						VHOST_ACCESS_RO);
-			if (unlikely(!desc ||
-					dlen != vq->desc[desc_indexes[i]].len))
+			if (unlikely(!desc))
 				break;
 
+			if (unlikely(dlen < vq->desc[desc_indexes[i]].len)) {
+				/*
+				 * The indirect desc table is not contiguous
+				 * in process VA space, we have to copy it.
+				 */
+				idesc = alloc_copy_ind_table(dev, vq,
+						&vq->desc[desc_indexes[i]]);
+				if (unlikely(!idesc))
+					break;
+
+				desc = idesc;
+			}
+
 			rte_prefetch0(desc);
 			sz = vq->desc[desc_indexes[i]].len / sizeof(*desc);
@@ -1365,4 +1453,5 @@
 			RTE_LOG(ERR, VHOST_DATA,
 				"Failed to allocate memory for mbuf.\n");
+			free_ind_table(idesc);
 			break;
 		}
@@ -1372,4 +1461,5 @@
 		if (unlikely(err)) {
 			rte_pktmbuf_free(pkts[i]);
+			free_ind_table(idesc);
 			break;
 		}
@@ -1381,4 +1471,5 @@
 			if (!zmbuf) {
 				rte_pktmbuf_free(pkts[i]);
+				free_ind_table(idesc);
 				break;
 			}
@@ -1397,4 +1488,7 @@
 			TAILQ_INSERT_TAIL(&vq->zmbuf_list, zmbuf, next);
 		}
+
+		if (unlikely(!!idesc))
+			free_ind_table(idesc);
 	}
 	vq->last_avail_idx += i;
-- 
1.8.3.1