From: Maxime Coquelin <maxime.coquelin@redhat.com>
To: dev@dpdk.org
Cc: Maxime Coquelin <maxime.coquelin@redhat.com>
Subject: [dpdk-dev] [PATCH 05/12] vhost: add support for non-contiguous indirect descs tables
Date: Mon, 23 Apr 2018 17:58:11 +0200 [thread overview]
Message-ID: <20180423155818.21285-6-maxime.coquelin@redhat.com> (raw)
In-Reply-To: <20180423155818.21285-1-maxime.coquelin@redhat.com>
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 2be3e7a7e..e43df8cb6 100644
--- a/lib/librte_vhost/virtio_net.c
+++ b/lib/librte_vhost/virtio_net.c
@@ -16,6 +16,7 @@
#include <rte_sctp.h>
#include <rte_arp.h>
#include <rte_spinlock.h>
+#include <rte_malloc.h>
#include "iotlb.h"
#include "vhost.h"
@@ -30,6 +31,46 @@ is_valid_virt_queue_idx(uint32_t idx, int is_tx, uint32_t nr_vring)
return (is_tx ^ (idx & 1)) == 0 && idx < nr_vring;
}
+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,
uint16_t to, uint16_t from, uint16_t size)
@@ -351,6 +392,7 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id,
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;
@@ -360,12 +402,24 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id,
vhost_iova_to_vva(dev,
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);
} else {
@@ -376,11 +430,15 @@ virtio_dev_rx(struct virtio_net *dev, uint16_t queue_id,
err = copy_mbuf_to_desc(dev, vq, descs, pkts[i], desc_idx, sz);
if (unlikely(err)) {
count = i;
+ free_ind_table(idesc);
break;
}
if (i + 1 < count)
rte_prefetch0(&vq->desc[desc_indexes[i+1]]);
+
+ if (unlikely(!!idesc))
+ free_ind_table(idesc);
}
do_data_copy_enqueue(dev, vq);
@@ -415,6 +473,7 @@ fill_vec_buf(struct virtio_net *dev, struct vhost_virtqueue *vq,
uint32_t len = 0;
uint64_t dlen;
struct vring_desc *descs = vq->desc;
+ struct vring_desc *idesc = NULL;
*desc_chain_head = idx;
@@ -424,15 +483,29 @@ fill_vec_buf(struct virtio_net *dev, struct vhost_virtqueue *vq,
vhost_iova_to_vva(dev, vq, vq->desc[idx].addr,
&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;
buf_vec[vec_id].buf_addr = descs[idx].addr;
@@ -449,6 +522,9 @@ fill_vec_buf(struct virtio_net *dev, struct vhost_virtqueue *vq,
*desc_chain_len = len;
*vec_idx = vec_id;
+ if (unlikely(!!idesc))
+ free_ind_table(idesc);
+
return 0;
}
@@ -1265,7 +1341,7 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
/* Prefetch descriptor index. */
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;
int err;
@@ -1280,10 +1356,22 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
vq->desc[desc_indexes[i]].addr,
&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);
idx = 0;
@@ -1297,6 +1385,7 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
if (unlikely(pkts[i] == NULL)) {
RTE_LOG(ERR, VHOST_DATA,
"Failed to allocate memory for mbuf.\n");
+ free_ind_table(idesc);
break;
}
@@ -1304,6 +1393,7 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
mbuf_pool);
if (unlikely(err)) {
rte_pktmbuf_free(pkts[i]);
+ free_ind_table(idesc);
break;
}
@@ -1313,6 +1403,7 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
zmbuf = get_zmbuf(vq);
if (!zmbuf) {
rte_pktmbuf_free(pkts[i]);
+ free_ind_table(idesc);
break;
}
zmbuf->mbuf = pkts[i];
@@ -1329,6 +1420,9 @@ rte_vhost_dequeue_burst(int vid, uint16_t queue_id,
vq->nr_zmbuf += 1;
TAILQ_INSERT_TAIL(&vq->zmbuf_list, zmbuf, next);
}
+
+ if (unlikely(!!idesc))
+ free_ind_table(idesc);
}
vq->last_avail_idx += i;
--
2.14.3
next prev parent reply other threads:[~2018-04-23 15:58 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-04-23 15:58 [dpdk-dev] [PATCH 00/12] Vhost: CVE-2018-1059 fixes Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 01/12] vhost: fix indirect descriptors table translation size Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 02/12] vhost: check all range is mapped when translating GPAs Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 03/12] vhost: introduce safe API for GPA translation Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 04/12] vhost: ensure all range is mapped when translating QVAs Maxime Coquelin
2018-04-23 15:58 ` Maxime Coquelin [this message]
2018-04-23 15:58 ` [dpdk-dev] [PATCH 06/12] vhost: handle virtually non-contiguous buffers in Tx Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 07/12] vhost: handle virtually non-contiguous buffers in Rx Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 08/12] vhost: handle virtually non-contiguous buffers in Rx-mrg Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 09/12] examples/vhost: move to safe GPA translation API Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 10/12] examples/vhost_scsi: " Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 11/12] vhost/crypto: " Maxime Coquelin
2018-04-23 15:58 ` [dpdk-dev] [PATCH 12/12] vhost: deprecate unsafe " Maxime Coquelin
2018-05-02 5:08 ` [dpdk-dev] [PATCH 00/12] Vhost: CVE-2018-1059 fixes Yao, Lei A
2018-05-02 9:20 ` Maxime Coquelin
2018-05-02 12:10 ` Yao, Lei A
2018-05-18 2:02 ` Yao, Lei A
2018-05-18 7:15 ` Maxime Coquelin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20180423155818.21285-6-maxime.coquelin@redhat.com \
--to=maxime.coquelin@redhat.com \
--cc=dev@dpdk.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).