From nobody Tue Oct 7 02:15:31 2025 Received: from mxhk.zte.com.cn (mxhk.zte.com.cn [160.30.148.35]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 804F4257452; Tue, 15 Jul 2025 15:52:59 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=160.30.148.35 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1752594783; cv=none; b=pq9JiQBo/osGk73pXLw7mAPzueVgk89cf04Fp8m9rbB0Mrw4IjPtRkrvr9v6b42jupZuXiy9i31e1HWLRNhnG/GlOMMs4teGS8pBeE41Kvji0v55dUkBMzdrFSV2oRCSQUo6Z7khuqWPc72F9/Ew61IW/q7M02mXO6+U390ZQJw= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1752594783; c=relaxed/simple; bh=hpah6/rE5DqZh8G1bxYLjWz7PdoI4oZdrcR4eeIx4Mc=; h=Date:Message-ID:Mime-Version:From:To:Cc:Subject:Content-Type; b=X+8Nk/0P/UsXXG3rFxpUgu8vyzgYTRWTZN5l6WY5iOd38jgd+5hGB/VOqEx6u+oWJ7SKr47j0MCiSw6/xQthilg/v+jt+FmPktzAAP28SHHGimqI9D1n+DPjBEEH2glSRExryG5DjSEzaX+HjNDs+/IPJe1XBlWzMjcUajD/YwY= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=zte.com.cn; spf=pass smtp.mailfrom=zte.com.cn; arc=none smtp.client-ip=160.30.148.35 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=zte.com.cn Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=zte.com.cn Received: from mse-fl1.zte.com.cn (unknown [10.5.228.132]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mxhk.zte.com.cn (FangMail) with ESMTPS id 4bhNxc62Kbz8R048; Tue, 15 Jul 2025 23:52:52 +0800 (CST) Received: from xaxapp04.zte.com.cn ([10.99.98.157]) by mse-fl1.zte.com.cn with SMTP id 56FFqkdV061208; Tue, 15 Jul 2025 23:52:46 +0800 (+08) (envelope-from fan.yu9@zte.com.cn) Received: from mapi (xaxapp02[null]) by mapi (Zmail) with MAPI id mid32; Tue, 15 Jul 2025 23:52:49 +0800 (CST) Date: Tue, 15 Jul 2025 23:52:49 +0800 (CST) X-Zmail-TransId: 2afa68767951ffffffff8b8-f3315 X-Mailer: Zmail v1.0 Message-ID: <20250715235249811YJbb1y6_zOrLbsh6PTlXv@zte.com.cn> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 From: To: , , , , , , , Cc: , , , , , , , , , Subject: =?UTF-8?B?W1BBVENIIG5ldC1uZXh0IHY1XSB0Y3A6IHRyYWNlIHJldHJhbnNtaXQgZmFpbHVyZXMgaW4gdGNwX3JldHJhbnNtaXRfc2ti?= X-MAIL: mse-fl1.zte.com.cn 56FFqkdV061208 X-Fangmail-Anti-Spam-Filtered: true X-Fangmail-MID-QID: 68767954.001/4bhNxc62Kbz8R048 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" From: Fan Yu Background =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D When TCP retransmits a packet due to missing ACKs, the retransmission may fail for various reasons (e.g., packets stuck in driver queues, receiver zero windows, or routing issues). The original tcp_retransmit_skb tracepoint: 'commit e086101b150a ("tcp: add a tracepoint for tcp retransmission")&= apos; lacks visibility into these failure causes, making production diagnostics difficult. Solution =3D=3D=3D=3D=3D=3D=3D=3D Adds the retval("err") to the tcp_retransmit_skb tracepoint. Enables users to know why some tcp retransmission failed and users can filter retransmission failures by retval. Compatibility description =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D This patch extends the tcp_retransmit_skb tracepoint by adding a new "err" field at the end of its existing structure (within TP_STRUCT__entry). The compatibility implications are detailed as follows: 1) Structural compatibility for legacy user-space tools Legacy tools/BPF programs accessing existing fields (by offset or name) can still work without modification or recompilation.The new field is appended to the end, preserving original memory layout. 2) Note: semantic changes The original tracepoint primarily only focused on successfully retransmitted packets. With this patch, the tracepoint now can figure out packets that may terminate early due to specific reasons. For accurate statistics, users should filter using "err" to distinguish outcomes. Before patched: # cat /sys/kernel/debug/tracing/events/tcp/tcp_retransmit_skb/format field:const void * skbaddr; offset:8; size:8; signed:0; field:const void * skaddr; offset:16; size:8; signed:0; field:int state; offset:24; size:4; signed:1; field:__u16 sport; offset:28; size:2; signed:0; field:__u16 dport; offset:30; size:2; signed:0; field:__u16 family; offset:32; size:2; signed:0; field:__u8 saddr[4]; offset:34; size:4; signed:0; field:__u8 daddr[4]; offset:38; size:4; signed:0; field:__u8 saddr_v6[16]; offset:42; size:16; signed:0; field:__u8 daddr_v6[16]; offset:58; size:16; signed:0; print fmt: "skbaddr=3D%p skaddr=3D%p family=3D%s sport=3D%hu dport=3D%hu sa= ddr=3D%pI4 daddr=3D%pI4 saddrv6=3D%pI6c daddrv6=3D%pI6c state=3D%s" After patched: # cat /sys/kernel/debug/tracing/events/tcp/tcp_retransmit_skb/format field:const void * skbaddr; offset:8; size:8; signed:0; field:const void * skaddr; offset:16; size:8; signed:0; field:int state; offset:24; size:4; signed:1; field:__u16 sport; offset:28; size:2; signed:0; field:__u16 dport; offset:30; size:2; signed:0; field:__u16 family; offset:32; size:2; signed:0; field:__u8 saddr[4]; offset:34; size:4; signed:0; field:__u8 daddr[4]; offset:38; size:4; signed:0; field:__u8 saddr_v6[16]; offset:42; size:16; signed:0; field:__u8 daddr_v6[16]; offset:58; size:16; signed:0; field:int err; offset:76; size:4; signed:1; print fmt: "skbaddr=3D%p skaddr=3D%p family=3D%s sport=3D%hu dport=3D%hu sa= ddr=3D%pI4 daddr=3D%pI4 saddrv6=3D%pI6c daddrv6=3D%pI6c state=3D%s err=3D%d" Change Log =3D=3D=3D=3D=3D=3D=3D=3D=3D v4->v5: Some fixes according to https://lore.kernel.org/all/20250715072058.12f343bb@kernel.org/ 1. Instead of introducing new TCP_RETRANS_* enums, directly passing the retval to the tracepoint. v3->v4: Some fixes according to https://lore.kernel.org/all/CANn89i+JGSt=3D_CtWfhDXypWW-34a6SoP3RAzWQ9B9VL4= +PHjDw@mail.gmail.com/ 1. Consolidate ENOMEMs into a unified TCP_RETRANS_NOMEM. v2->v3: Some fixes according to https://lore.kernel.org/all/CANn89iJvyYjiweCESQL8E-Si7M=3DgosYvh1BAVWwAWycX= W8GSdg@mail.gmail.com/ 1. Rename "quit_reason" to "result". Also, keep "key=3Dval" format concise(= no space in vals). v1->v2: Some fixes according to https://lore.kernel.org/all/CANn89iK-6kT-ZUpNRMjPY9_TkQj-dLuKrDQtvO1140q4EU= sjFg@mail.gmail.com/ 1.Rename TCP_RETRANS_QUIT_UNDEFINED to TCP_RETRANS_ERR_DEFAULT. 2.Added detailed compatibility consequences section. Suggested-by: Jakub Kicinski Suggested-by: Eric Dumazet Co-developed-by: xu xin Signed-off-by: xu xin Signed-off-by: Fan Yu --- include/trace/events/tcp.h | 27 ++++++++-------------- net/ipv4/tcp_output.c | 46 ++++++++++++++++++++++++-------------- 2 files changed, 38 insertions(+), 35 deletions(-) diff --git a/include/trace/events/tcp.h b/include/trace/events/tcp.h index 54e60c6009e3..9d2c36c6a0ed 100644 --- a/include/trace/events/tcp.h +++ b/include/trace/events/tcp.h @@ -13,17 +13,11 @@ #include #include -/* - * tcp event with arguments sk and skb - * - * Note: this class requires a valid sk pointer; while skb pointer could - * be NULL. - */ -DECLARE_EVENT_CLASS(tcp_event_sk_skb, +TRACE_EVENT(tcp_retransmit_skb, - TP_PROTO(const struct sock *sk, const struct sk_buff *skb), + TP_PROTO(const struct sock *sk, const struct sk_buff *skb, int err), - TP_ARGS(sk, skb), + TP_ARGS(sk, skb, err), TP_STRUCT__entry( __field(const void *, skbaddr) @@ -36,6 +30,7 @@ DECLARE_EVENT_CLASS(tcp_event_sk_skb, __array(__u8, daddr, 4) __array(__u8, saddr_v6, 16) __array(__u8, daddr_v6, 16) + __field(int, err) ), TP_fast_assign( @@ -58,21 +53,17 @@ DECLARE_EVENT_CLASS(tcp_event_sk_skb, TP_STORE_ADDRS(__entry, inet->inet_saddr, inet->inet_daddr, sk->sk_v6_rcv_saddr, sk->sk_v6_daddr); + + __entry->err =3D err; ), - TP_printk("skbaddr=3D%p skaddr=3D%p family=3D%s sport=3D%hu dport=3D%hu s= addr=3D%pI4 daddr=3D%pI4 saddrv6=3D%pI6c daddrv6=3D%pI6c state=3D%s", + TP_printk("skbaddr=3D%p skaddr=3D%p family=3D%s sport=3D%hu dport=3D%hu s= addr=3D%pI4 daddr=3D%pI4 saddrv6=3D%pI6c daddrv6=3D%pI6c state=3D%s err=3D%= d", __entry->skbaddr, __entry->skaddr, show_family_name(__entry->family), __entry->sport, __entry->dport, __entry->saddr, __entry->daddr, __entry->saddr_v6, __entry->daddr_v6, - show_tcp_state_name(__entry->state)) -); - -DEFINE_EVENT(tcp_event_sk_skb, tcp_retransmit_skb, - - TP_PROTO(const struct sock *sk, const struct sk_buff *skb), - - TP_ARGS(sk, skb) + show_tcp_state_name(__entry->state), + __entry->err) ); #undef FN diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index b616776e3354..c31e164693d5 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -3330,8 +3330,10 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_= buff *skb, int segs) if (icsk->icsk_mtup.probe_size) icsk->icsk_mtup.probe_size =3D 0; - if (skb_still_in_host_queue(sk, skb)) - return -EBUSY; + if (skb_still_in_host_queue(sk, skb)) { + err =3D -EBUSY; + goto out; + } start: if (before(TCP_SKB_CB(skb)->seq, tp->snd_una)) { @@ -3342,14 +3344,19 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk= _buff *skb, int segs) } if (unlikely(before(TCP_SKB_CB(skb)->end_seq, tp->snd_una))) { WARN_ON_ONCE(1); - return -EINVAL; + err =3D -EINVAL; + goto out; + } + if (tcp_trim_head(sk, skb, tp->snd_una - TCP_SKB_CB(skb)->seq)) { + err =3D -ENOMEM; + goto out; } - if (tcp_trim_head(sk, skb, tp->snd_una - TCP_SKB_CB(skb)->seq)) - return -ENOMEM; } - if (inet_csk(sk)->icsk_af_ops->rebuild_header(sk)) - return -EHOSTUNREACH; /* Routing failure or similar. */ + if (inet_csk(sk)->icsk_af_ops->rebuild_header(sk)) { + err =3D -EHOSTUNREACH; /* Routing failure or similar. */ + goto out; + } cur_mss =3D tcp_current_mss(sk); avail_wnd =3D tcp_wnd_end(tp) - TCP_SKB_CB(skb)->seq; @@ -3360,8 +3367,10 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk_= buff *skb, int segs) * our retransmit of one segment serves as a zero window probe. */ if (avail_wnd <=3D 0) { - if (TCP_SKB_CB(skb)->seq !=3D tp->snd_una) - return -EAGAIN; + if (TCP_SKB_CB(skb)->seq !=3D tp->snd_una) { + err =3D -EAGAIN; + goto out; + } avail_wnd =3D cur_mss; } @@ -3373,11 +3382,15 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk= _buff *skb, int segs) } if (skb->len > len) { if (tcp_fragment(sk, TCP_FRAG_IN_RTX_QUEUE, skb, len, - cur_mss, GFP_ATOMIC)) - return -ENOMEM; /* We'll try again later. */ + cur_mss, GFP_ATOMIC)) { + err =3D -ENOMEM; /* We'll try again later. */ + goto out; + } } else { - if (skb_unclone_keeptruesize(skb, GFP_ATOMIC)) - return -ENOMEM; + if (skb_unclone_keeptruesize(skb, GFP_ATOMIC)) { + err =3D -ENOMEM; + goto out; + } diff =3D tcp_skb_pcount(skb); tcp_set_skb_tso_segs(skb, cur_mss); @@ -3431,17 +3444,16 @@ int __tcp_retransmit_skb(struct sock *sk, struct sk= _buff *skb, int segs) tcp_call_bpf_3arg(sk, BPF_SOCK_OPS_RETRANS_CB, TCP_SKB_CB(skb)->seq, segs, err); - if (likely(!err)) { - trace_tcp_retransmit_skb(sk, skb); - } else if (err !=3D -EBUSY) { + if (err !=3D -EBUSY) NET_ADD_STATS(sock_net(sk), LINUX_MIB_TCPRETRANSFAIL, segs); - } /* To avoid taking spuriously low RTT samples based on a timestamp * for a transmit that never happened, always mark EVER_RETRANS */ TCP_SKB_CB(skb)->sacked |=3D TCPCB_EVER_RETRANS; +out: + trace_tcp_retransmit_skb(sk, skb, err); return err; } -- 2.25.1