Thread (20 messages) 20 messages, 3 authors, 2021-03-27

Re: [Patch bpf-next v6 02/12] skmsg: introduce a spinlock to protect ingress_msg

From: Yunsheng Lin <hidden>
Date: 2021-03-23 01:25:58
Also in: bpf

On 2021/3/23 8:37, Cong Wang wrote:
quoted hunk ↗ jump to hunk
From: Cong Wang <redacted>

Currently we rely on lock_sock to protect ingress_msg,
it is too big for this, we can actually just use a spinlock
to protect this list like protecting other skb queues.

__tcp_bpf_recvmsg() is still special because of peeking,
it still has to use lock_sock.

Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Lorenz Bauer <redacted>
Acked-by: Jakub Sitnicki <jakub@cloudflare.com>
Signed-off-by: Cong Wang <redacted>
---
 include/linux/skmsg.h | 46 +++++++++++++++++++++++++++++++++++++++++++
 net/core/skmsg.c      |  3 +++
 net/ipv4/tcp_bpf.c    | 18 ++++++-----------
 3 files changed, 55 insertions(+), 12 deletions(-)
diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h
index 6c09d94be2e9..f2d45a73b2b2 100644
--- a/include/linux/skmsg.h
+++ b/include/linux/skmsg.h
@@ -89,6 +89,7 @@ struct sk_psock {
 #endif
 	struct sk_buff_head		ingress_skb;
 	struct list_head		ingress_msg;
+	spinlock_t			ingress_lock;
 	unsigned long			state;
 	struct list_head		link;
 	spinlock_t			link_lock;
@@ -284,7 +285,45 @@ static inline struct sk_psock *sk_psock(const struct sock *sk)
 static inline void sk_psock_queue_msg(struct sk_psock *psock,
 				      struct sk_msg *msg)
 {
+	spin_lock_bh(&psock->ingress_lock);
 	list_add_tail(&msg->list, &psock->ingress_msg);
+	spin_unlock_bh(&psock->ingress_lock);
+}
+
+static inline struct sk_msg *sk_psock_dequeue_msg(struct sk_psock *psock)
+{
+	struct sk_msg *msg;
+
+	spin_lock_bh(&psock->ingress_lock);
+	msg = list_first_entry_or_null(&psock->ingress_msg, struct sk_msg, list);
+	if (msg)
+		list_del(&msg->list);
+	spin_unlock_bh(&psock->ingress_lock);
+	return msg;
+}
+
+static inline struct sk_msg *sk_psock_peek_msg(struct sk_psock *psock)
+{
+	struct sk_msg *msg;
+
+	spin_lock_bh(&psock->ingress_lock);
+	msg = list_first_entry_or_null(&psock->ingress_msg, struct sk_msg, list);
+	spin_unlock_bh(&psock->ingress_lock);
+	return msg;
+}
+
+static inline struct sk_msg *sk_psock_next_msg(struct sk_psock *psock,
+					       struct sk_msg *msg)
+{
+	struct sk_msg *ret;
Nit:
Use msg instead of ret to be consistently with sk_psock_dequeue_msg()
and sk_psock_next_msg().
+
+	spin_lock_bh(&psock->ingress_lock);
+	if (list_is_last(&msg->list, &psock->ingress_msg))
+		ret = NULL;
+	else
+		ret = list_next_entry(msg, list);
+	spin_unlock_bh(&psock->ingress_lock);
+	return ret;
 }
  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help