Re: [PATCH net v3] tcp: check skb is non-NULL in tcp_rto_delta_us()

From: Josh Hunt
Date: Tue Sep 24 2024 - 14:15:49 EST


On 9/19/24 2:05 AM, Paolo Abeni wrote:
!-------------------------------------------------------------------|
 This Message Is From an External Sender
 This message came from outside your organization.
|-------------------------------------------------------------------!

On 9/10/24 21:08, Josh Hunt wrote:
diff --git a/include/net/tcp.h b/include/net/tcp.h
index 2aac11e7e1cc..196c148fce8a 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -2434,9 +2434,26 @@ static inline s64 tcp_rto_delta_us(const struct sock *sk)
  {
      const struct sk_buff *skb = tcp_rtx_queue_head(sk);
      u32 rto = inet_csk(sk)->icsk_rto;
-    u64 rto_time_stamp_us = tcp_skb_timestamp_us(skb) + jiffies_to_usecs(rto);
-    return rto_time_stamp_us - tcp_sk(sk)->tcp_mstamp;
+    if (likely(skb)) {
+        u64 rto_time_stamp_us = tcp_skb_timestamp_us(skb) + jiffies_to_usecs(rto);
+
+        return rto_time_stamp_us - tcp_sk(sk)->tcp_mstamp;
+    } else {
+        WARN_ONCE(1,
+            "rtx queue emtpy: "
+            "out:%u sacked:%u lost:%u retrans:%u "
+            "tlp_high_seq:%u sk_state:%u ca_state:%u "
+            "advmss:%u mss_cache:%u pmtu:%u\n",
+            tcp_sk(sk)->packets_out, tcp_sk(sk)->sacked_out,
+            tcp_sk(sk)->lost_out, tcp_sk(sk)->retrans_out,
+            tcp_sk(sk)->tlp_high_seq, sk->sk_state,
+            inet_csk(sk)->icsk_ca_state,
+            tcp_sk(sk)->advmss, tcp_sk(sk)->mss_cache,
+            inet_csk(sk)->icsk_pmtu_cookie);

As the underlying issue here share the same root cause as the one covered by the WARN_ONCE() in tcp_send_loss_probe(), I'm wondering if it would make sense do move the info dumping in a common helper, so that we get the verbose warning on either cases.

Thanks,

Paolo

Thanks for the review Paolo. Sorry for the delay in replying I was OOO. I can send a follow-up commit to create a common helper.

Josh