Skip to contentSkip to conversation

vsock/virtio: bind uarg before filling zerocopy skb

PATCH Applied Lin Ma 2 participants 2 messages May 27 – May 29

Subsystems (6)
lore
Share & tools
lore

Lin Ma

author
#0 lore raw

PATCH v2 1/1 vsock/virtio: bind uarg before filling zerocopy skb

From: Jingguo Tan <[email protected]>

virtio_transport_send_pkt_info() allocates or reuses the zerocopy uarg before entering the send loop, but virtio_transport_alloc_skb() still fills the skb before it inherits that uarg. When fixed-buffer vectored zerocopy hits MAX_SKB_FRAGS, io_sg_from_iter() may partially attach managed frags and return -EMSGSIZE. The rollback path call kfree_skb() to free an skb that carries SKBFL_MANAGED_FRAG_REFS but no uarg, so skb_release_data() falls through to ordinary frag unref.

Pass the uarg into virtio_transport_alloc_skb() and bind it immediately before virtio_transport_fill_skb(). This keeps control or no-payload skbs untouched while ensuring success and rollback share one lifetime rule.

---
v1->v2 changes:
* add acked and review tags.
* repost to avoid duplicated subject.

 net/vmw_vsock/virtio_transport_common.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
net/vmw_vsock/virtio_transport_common.c +8 -3
index df3b418e0392..73f58925ff72 100644
@@ -205,6 +205,7 @@ static u16 virtio_transport_get_type(struct sock *sk)
 static struct sk_buff *virtio_transport_alloc_skb(struct virtio_vsock_pkt_info *info,
 						  size_t payload_len,
 						  bool zcopy,
+						  struct ubuf_info *uarg,
 						  u32 src_cid,
 						  u32 src_port,
 						  u32 dst_cid,
@@ -245,6 +246,11 @@ static struct sk_buff *virtio_transport_alloc_skb(struct virtio_vsock_pkt_info *
 	if (info->msg && payload_len > 0) {
 		int err;
 
+		/* Bind the zerocopy lifetime before filling frags so error rollback
+		 * frees managed fixed-buffer pages through the uarg-aware path.
+		 */
+		skb_zcopy_set(skb, uarg, NULL);
+
 		err = virtio_transport_fill_skb(skb, info, payload_len, zcopy);
 		if (err)
 			goto out;
@@ -364,6 +370,7 @@ static int virtio_transport_send_pkt_info(struct vsock_sock *vsk,
 		skb_len = min(max_skb_len, rest_len);
 
 		skb = virtio_transport_alloc_skb(info, skb_len, can_zcopy,
+						 uarg,
 						 src_cid, src_port,
 						 dst_cid, dst_port);
 		if (!skb) {
@@ -371,8 +378,6 @@ static int virtio_transport_send_pkt_info(struct vsock_sock *vsk,
 			break;
 		}
 
-		skb_zcopy_set(skb, uarg, NULL);
-
 		virtio_transport_inc_tx_pkt(vvs, skb);
 
 		ret = t_ops->send_pkt(skb, info->net);
@@ -1183,7 +1188,7 @@ static int virtio_transport_reset_no_sock(const struct virtio_transport *t,
 	if (!t)
 		return -ENOTCONN;
 
-	reply = virtio_transport_alloc_skb(&info, 0, false,
+	reply = virtio_transport_alloc_skb(&info, 0, false, NULL,
 					   le64_to_cpu(hdr->dst_cid),
 					   le32_to_cpu(hdr->dst_port),
 					   le64_to_cpu(hdr->src_cid),
Signature · 1 line
2.53.0
#1 lore raw
This patch was applied to netdev/net.git (main) by Jakub Kicinski <[email protected]>:

Hello:

This patch was applied to netdev/net.git (main)
by Jakub Kicinski <[email protected]>:

Quoted 11 lines · On Wed, 27 May 2026 10:33:01 +0800 you wrote:

From: Jingguo Tan <[email protected]>

virtio_transport_send_pkt_info() allocates or reuses the zerocopy uarg before entering the send loop, but virtio_transport_alloc_skb() still fills the skb before it inherits that uarg. When fixed-buffer vectored zerocopy hits MAX_SKB_FRAGS, io_sg_from_iter() may partially attach managed frags and return -EMSGSIZE. The rollback path call kfree_skb() to free an skb that carries SKBFL_MANAGED_FRAG_REFS but no uarg, so skb_release_data() falls through to ordinary frag unref.

[...]

Here is the summary with links:
- [net,v2] vsock/virtio: bind uarg before filling zerocopy skb

    https://git.kernel.org/netdev/net/c/1e584c304cfb

You are awesome, thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html