Skip to contentSkip to conversation

virtio-net: correctly copy vnet header when flushing TX

PATCH In review Jason Wang 4 participants 7 messages Jan 2, 2024 – Jan 22, 2024

lore
Share & tools
lore

Jason Wang

author
#0 lore raw

PATCH 1/1 virtio-net: correctly copy vnet header when flushing TX

When HASH_REPORT is negotiated, the guest_hdr_len might be larger than the size of the mergeable rx buffer header. Using virtio_net_hdr_mrg_rxbuf during the header swap might lead a stack overflow in this case. Fixing this by using virtio_net_hdr_v1_hash instead.

Cc Yuri Benditovich <[email protected]>
Cc Mauro Matteo Cascella <[email protected]>
Fixes CVE-2023-6693
Fixes e22f0603fb2f ("virtio-net: reference implementation of hash report")
---
 hw/net/virtio-net.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
hw/net/virtio-net.c +9 -4
index 80c56f0cfc..73024babd4 100644
@@ -674,6 +674,11 @@ static void virtio_net_set_mrg_rx_bufs(VirtIONet *n, int mergeable_rx_bufs,
 
     n->mergeable_rx_bufs = mergeable_rx_bufs;
 
+    /*
+     * Note: when extending the vnet header, please make sure to
+     * change the vnet header copying logic in virtio_net_flush_tx()
+     * as well.
+     */
     if (version_1) {
         n->guest_hdr_len = hash_report ?
             sizeof(struct virtio_net_hdr_v1_hash) :
@@ -2693,7 +2698,7 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
         ssize_t ret;
         unsigned int out_num;
         struct iovec sg[VIRTQUEUE_MAX_SIZE], sg2[VIRTQUEUE_MAX_SIZE + 1], *out_sg;
-        struct virtio_net_hdr_mrg_rxbuf mhdr;
+        struct virtio_net_hdr_v1_hash vhdr;
 
         elem = virtqueue_pop(q->tx_vq, sizeof(VirtQueueElement));
         if (!elem) {
@@ -2710,7 +2715,7 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
         }
 
         if (n->has_vnet_hdr) {
-            if (iov_to_buf(out_sg, out_num, 0, &mhdr, n->guest_hdr_len) <
+            if (iov_to_buf(out_sg, out_num, 0, &vhdr, n->guest_hdr_len) <
                 n->guest_hdr_len) {
                 virtio_error(vdev, "virtio-net header incorrect");
                 virtqueue_detach_element(q->tx_vq, elem, 0);
@@ -2718,8 +2723,8 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
                 return -EINVAL;
             }
             if (n->needs_vnet_hdr_swap) {
-                virtio_net_hdr_swap(vdev, (void *) &mhdr);
-                sg2[0].iov_base = &mhdr;
+                virtio_net_hdr_swap(vdev, (void *) &vhdr);
+                sg2[0].iov_base = &vhdr;
                 sg2[0].iov_len = n->guest_hdr_len;
                 out_num = iov_copy(&sg2[1], ARRAY_SIZE(sg2) - 1,
                                    out_sg, out_num,
Signature · 1 line
2.42.0
#1 lore raw

I agree, thank you.

Where is this CVE-2023-6693 available?

Thanks,
Yuri

Quoted 68 lines · Jason Wang

When HASH_REPORT is negotiated, the guest_hdr_len might be larger than the size of the mergeable rx buffer header. Using virtio_net_hdr_mrg_rxbuf during the header swap might lead a stack overflow in this case. Fixing this by using virtio_net_hdr_v1_hash instead.

Cc Yuri Benditovich <[email protected]>
Cc Mauro Matteo Cascella <[email protected]>
Fixes CVE-2023-6693
Fixes e22f0603fb2f ("virtio-net: reference implementation of hash report")
---
 hw/net/virtio-net.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
hw/net/virtio-net.c +5 -0
index 80c56f0cfc..73024babd4 100644
@@ -674,6 +674,11 @@ static void virtio_net_set_mrg_rx_bufs(VirtIONet *n,
 int mergeable_rx_bufs,
 
     n->mergeable_rx_bufs = mergeable_rx_bufs;
 
+    /*
+     * Note: when extending the vnet header, please make sure to
+     * change the vnet header copying logic in virtio_net_flush_tx()
+     * as well.
+     */
     if (version_1) {
         n->guest_hdr_len = hash_report ?

sizeof(struct virtio_net_hdr_v1_hash) :

hw/net/virtio-net.c +1 -1
@@ -2693,7 +2698,7 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
         ssize_t ret;
         unsigned int out_num;
         struct iovec sg[VIRTQUEUE_MAX_SIZE], sg2[VIRTQUEUE_MAX_SIZE + 1],
 *out_sg;
-        struct virtio_net_hdr_mrg_rxbuf mhdr;
+        struct virtio_net_hdr_v1_hash vhdr;
 
         elem = virtqueue_pop(q->tx_vq, sizeof(VirtQueueElement));
         if (!elem) {
hw/net/virtio-net.c +3 -4
@@ -2710,7 +2715,7 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
         }
 
         if (n->has_vnet_hdr) {
-            if (iov_to_buf(out_sg, out_num, 0, &mhdr, n->guest_hdr_len) <
+            if (iov_to_buf(out_sg, out_num, 0, &vhdr, n->guest_hdr_len) <
                 n->guest_hdr_len) {
                 virtio_error(vdev, "virtio-net header incorrect");
                 virtqueue_detach_element(q->tx_vq, elem, 0);
@@ -2718,8 +2723,8 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
                 return -EINVAL;
             }
             if (n->needs_vnet_hdr_swap) {
-                virtio_net_hdr_swap(vdev, (void *) &mhdr);
-                sg2[0].iov_base = &mhdr;
+                virtio_net_hdr_swap(vdev, (void *) &vhdr);
+                sg2[0].iov_base = &vhdr;
                 sg2[0].iov_len = n->guest_hdr_len;
                 out_num = iov_copy(&sg2[1], ARRAY_SIZE(sg2) - 1,
                                    out_sg, out_num,
--

2.42.0

#2 lore raw
Quoted 4 lines · Yuri Benditovich

I agree, thank you.

Where is this CVE-2023-6693 available?

Here: https://bugzilla.redhat.com/show_bug.cgi?id=2254580.

Quoted 70 lines

Thanks,
Yuri

Quoted 66 lines · Jason Wang

When HASH_REPORT is negotiated, the guest_hdr_len might be larger than the size of the mergeable rx buffer header. Using virtio_net_hdr_mrg_rxbuf during the header swap might lead a stack overflow in this case. Fixing this by using virtio_net_hdr_v1_hash instead.

Cc Yuri Benditovich <[email protected]>
Cc Mauro Matteo Cascella <[email protected]>
Fixes CVE-2023-6693
Fixes e22f0603fb2f ("virtio-net: reference implementation of hash report")
---
 hw/net/virtio-net.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
hw/net/virtio-net.c +9 -5
index 80c56f0cfc..73024babd4 100644
@@ -674,6 +674,11 @@ static void virtio_net_set_mrg_rx_bufs(VirtIONet *n, int mergeable_rx_bufs,
 
     n->mergeable_rx_bufs = mergeable_rx_bufs;
 
+    /*
+     * Note: when extending the vnet header, please make sure to
+     * change the vnet header copying logic in virtio_net_flush_tx()
+     * as well.
+     */
     if (version_1) {
         n->guest_hdr_len = hash_report ?
             sizeof(struct virtio_net_hdr_v1_hash) :
@@ -2693,7 +2698,7 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
         ssize_t ret;
         unsigned int out_num;
         struct iovec sg[VIRTQUEUE_MAX_SIZE], sg2[VIRTQUEUE_MAX_SIZE + 1], *out_sg;
-        struct virtio_net_hdr_mrg_rxbuf mhdr;
+        struct virtio_net_hdr_v1_hash vhdr;
 
         elem = virtqueue_pop(q->tx_vq, sizeof(VirtQueueElement));
         if (!elem) {
@@ -2710,7 +2715,7 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
         }
 
         if (n->has_vnet_hdr) {
-            if (iov_to_buf(out_sg, out_num, 0, &mhdr, n->guest_hdr_len) <
+            if (iov_to_buf(out_sg, out_num, 0, &vhdr, n->guest_hdr_len) <
                 n->guest_hdr_len) {
                 virtio_error(vdev, "virtio-net header incorrect");
                 virtqueue_detach_element(q->tx_vq, elem, 0);
@@ -2718,8 +2723,8 @@ static int32_t virtio_net_flush_tx(VirtIONetQueue *q)
                 return -EINVAL;
             }
             if (n->needs_vnet_hdr_swap) {
-                virtio_net_hdr_swap(vdev, (void *) &mhdr);
-                sg2[0].iov_base = &mhdr;
+                virtio_net_hdr_swap(vdev, (void *) &vhdr);
+                sg2[0].iov_base = &vhdr;
                 sg2[0].iov_len = n->guest_hdr_len;
                 out_num = iov_copy(&sg2[1], ARRAY_SIZE(sg2) - 1,
                                    out_sg, out_num,
--

2.42.0

Signature · 3 lines
Mauro Matteo Cascella
Red Hat Product Security
PGP-Key ID: BB3410B0
author
#3 lore raw
In reply to Jason Wang
Quoted 7 lines · Jason Wang

When HASH_REPORT is negotiated, the guest_hdr_len might be larger than the size of the mergeable rx buffer header. Using virtio_net_hdr_mrg_rxbuf during the header swap might lead a stack overflow in this case. Fixing this by using virtio_net_hdr_v1_hash instead.

Quoted 7 lines
Cc Yuri Benditovich <[email protected]>
Cc Mauro Matteo Cascella <[email protected]>
Fixes CVE-2023-6693
Fixes e22f0603fb2f ("virtio-net: reference implementation of hash report")

Queued.

Thanks

#4 lore raw
In reply to Jason Wang

02.01.2024 06:29, Jason Wang :

Quoted 6 lines

When HASH_REPORT is negotiated, the guest_hdr_len might be larger than the size of the mergeable rx buffer header. Using virtio_net_hdr_mrg_rxbuf during the header swap might lead a stack overflow in this case. Fixing this by using virtio_net_hdr_v1_hash instead.

Quoted 7 lines
Cc Yuri Benditovich <[email protected]>
Cc Mauro Matteo Cascella <[email protected]>
Fixes CVE-2023-6693
Fixes e22f0603fb2f ("virtio-net: reference implementation of hash report")

Hi! Can we get this to master before Jan-27 please, so it's included in 8.2.1?

Thanks!

/mjt

author
#5 lore raw
Quoted 16 lines · Michael Tokarev

02.01.2024 06:29, Jason Wang :

Quoted 6 lines

When HASH_REPORT is negotiated, the guest_hdr_len might be larger than the size of the mergeable rx buffer header. Using virtio_net_hdr_mrg_rxbuf during the header swap might lead a stack overflow in this case. Fixing this by using virtio_net_hdr_v1_hash instead.

Quoted 7 lines
Cc Yuri Benditovich <[email protected]>
Cc Mauro Matteo Cascella <[email protected]>
Fixes CVE-2023-6693
Fixes e22f0603fb2f ("virtio-net: reference implementation of hash report")
Quoted 1 line

Hi! Can we get this to master before Jan-27 please, so it's included in 8.2.1?

I think it will be.

Thanks

Quoted 5 lines

Thanks!

/mjt

Reviewed-by on v1 1/1
#6 lore raw
In reply to Jason Wang

02.01.2024 06:29, Jason Wang :

Quoted 6 lines

When HASH_REPORT is negotiated, the guest_hdr_len might be larger than the size of the mergeable rx buffer header. Using virtio_net_hdr_mrg_rxbuf during the header swap might lead a stack overflow in this case. Fixing this by using virtio_net_hdr_v1_hash instead.

Quoted 7 lines
Cc Yuri Benditovich <[email protected]>
Cc Mauro Matteo Cascella <[email protected]>
Fixes CVE-2023-6693
Fixes e22f0603fb2f ("virtio-net: reference implementation of hash report")

/mjt