[PATCH net-next 0/9] net: add skb drop reasons to TCP packet receive

[PATCH net-next 0/9] net: add skb drop reasons to TCP packet receive

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

In this series patches, reasons for skb drops are added to TCP layer, and
both TCPv4 and TCPv6 are considered.

in this series patches, the process of packet ingress in TCP layer is
considered, as skb drops hardly happens in the egress path.

However, it's a little complex for TCP state processing, as I find that
it's hard to report skb drop reasons to where it is freed. For example,
when skb is dropped in tcp_rcv_state_process(), the reason can be caused
by the call of tcp_v4_conn_request(), and it's hard to return a drop
reason from tcp_v4_conn_request(). So I just skip such case for this
moment.


Menglong Dong (9):
  net: tcp: introduce tcp_drop_reason()
  net: tcp: add skb drop reasons to tcp_v4_rcv()
  net: tcp: use kfree_skb_reason() for tcp_v6_rcv()
  net: tcp: add skb drop reasons to tcp_v{4,6}_inbound_md5_hash()
  net: tcp: add skb drop reasons to tcp_add_backlog()
  net: tcp: use kfree_skb_reason() for tcp_v{4,6}_do_rcv()
  net: tcp: use tcp_drop_reason() for tcp_rcv_established()
  net: tcp: use tcp_drop_reason() for tcp_data_queue()
  net: tcp: use tcp_drop_reason() for tcp_data_queue_ofo()

 include/linux/skbuff.h     | 28 +++++++++++++++++++++++++
 include/net/tcp.h          |  3 ++-
 include/trace/events/skb.h | 10 +++++++++
 net/ipv4/tcp_input.c       | 42 +++++++++++++++++++++++++++++---------
 net/ipv4/tcp_ipv4.c        | 36 ++++++++++++++++++++++++--------
 net/ipv6/tcp_ipv6.c        | 42 +++++++++++++++++++++++++++++---------
 6 files changed, 131 insertions(+), 30 deletions(-)

-- 
2.34.1

[PATCH net-next 1/9] net: tcp: introduce tcp_drop_reason()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

For TCP protocol, tcp_drop() is used to free the skb when it needs
to be dropped. To make use of kfree_skb_reason() and pass the drop
reason to it, introduce the function tcp_drop_reason(). Meanwhile,
make tcp_drop() an inline call to tcp_drop_reason().

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 net/ipv4/tcp_input.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index af94a6d22a9d..0a2740add404 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -4684,10 +4684,16 @@ static bool tcp_ooo_try_coalesce(struct sock *sk,
 	return res;
 }
 
-static void tcp_drop(struct sock *sk, struct sk_buff *skb)
+static void tcp_drop_reason(struct sock *sk, struct sk_buff *skb,
+			    enum skb_drop_reason reason)
 {
 	sk_drops_add(sk, skb);
-	__kfree_skb(skb);
+	kfree_skb_reason(skb, reason);
+}
+
+static inline void tcp_drop(struct sock *sk, struct sk_buff *skb)
+{
+	tcp_drop_reason(sk, skb, SKB_DROP_REASON_NOT_SPECIFIED);
 }
 
 /* This one checks to see if we can put data from the
-- 
2.34.1

Re: [PATCH net-next 1/9] net: tcp: introduce tcp_drop_reason()

[Jakub Kicinski]

On Wed, 16 Feb 2022 11:54:18 +0800 menglong8.dong@gmail.com wrote:
> +static inline void tcp_drop(struct sock *sk, struct sk_buff *skb)
> +{
> +	tcp_drop_reason(sk, skb, SKB_DROP_REASON_NOT_SPECIFIED);
>  }

Please make this non-inline. The compiler will inline it anyway, and 
if it's a static inline compiler will not warn us that it should be
removed once all callers are gone.

Re: [PATCH net-next 1/9] net: tcp: introduce tcp_drop_reason()

[Steven Rostedt]

On Wed, 16 Feb 2022 07:58:21 -0800
Jakub Kicinski <kuba@kernel.org> wrote:

> On Wed, 16 Feb 2022 11:54:18 +0800 menglong8.dong@gmail.com wrote:
> > +static inline void tcp_drop(struct sock *sk, struct sk_buff *skb)
> > +{
> > +	tcp_drop_reason(sk, skb, SKB_DROP_REASON_NOT_SPECIFIED);
> >  }  
> 
> Please make this non-inline. The compiler will inline it anyway, and 
> if it's a static inline compiler will not warn us that it should be
> removed once all callers are gone.

I guess that's no longer true for C files.

  https://lore.kernel.org/all/202202132037.4aN017dU-lkp@intel.com/

-- Steve

Re: [PATCH net-next 1/9] net: tcp: introduce tcp_drop_reason()

[Menglong Dong]

On Thu, Feb 17, 2022 at 12:50 AM Steven Rostedt <rostedt@goodmis.org> wrote:
>
> On Wed, 16 Feb 2022 07:58:21 -0800
> Jakub Kicinski <kuba@kernel.org> wrote:
>
> > On Wed, 16 Feb 2022 11:54:18 +0800 menglong8.dong@gmail.com wrote:
> > > +static inline void tcp_drop(struct sock *sk, struct sk_buff *skb)
> > > +{
> > > +   tcp_drop_reason(sk, skb, SKB_DROP_REASON_NOT_SPECIFIED);
> > >  }
> >
> > Please make this non-inline. The compiler will inline it anyway, and
> > if it's a static inline compiler will not warn us that it should be
> > removed once all callers are gone.
>
> I guess that's no longer true for C files.
>
>   https://lore.kernel.org/all/202202132037.4aN017dU-lkp@intel.com/
>

Ok, seems we can keep this inline still.

> -- Steve

[PATCH net-next 2/9] net: tcp: add skb drop reasons to tcp_v4_rcv()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Use kfree_skb_reason() for some path in tcp_v4_rcv() that missed before,
including:

SKB_DROP_REASON_SOCKET_FILTER
SKB_DROP_REASON_XFRM_POLICY

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 net/ipv4/tcp_ipv4.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 6873f46fc8ba..a93921fb498f 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -2057,6 +2057,8 @@ int tcp_v4_rcv(struct sk_buff *skb)
 			iph = ip_hdr(skb);
 			tcp_v4_fill_cb(skb, iph, th);
 			nsk = tcp_check_req(sk, skb, req, false, &req_stolen);
+		} else {
+			drop_reason = SKB_DROP_REASON_SOCKET_FILTER;
 		}
 		if (!nsk) {
 			reqsk_put(req);
@@ -2092,8 +2094,10 @@ int tcp_v4_rcv(struct sk_buff *skb)
 		}
 	}
 
-	if (!xfrm4_policy_check(sk, XFRM_POLICY_IN, skb))
+	if (!xfrm4_policy_check(sk, XFRM_POLICY_IN, skb)) {
+		drop_reason = SKB_DROP_REASON_XFRM_POLICY;
 		goto discard_and_relse;
+	}
 
 	if (tcp_v4_inbound_md5_hash(sk, skb, dif, sdif))
 		goto discard_and_relse;
@@ -2137,8 +2141,10 @@ int tcp_v4_rcv(struct sk_buff *skb)
 
 no_tcp_socket:
 	drop_reason = SKB_DROP_REASON_NO_SOCKET;
-	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb))
+	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
+		drop_reason = SKB_DROP_REASON_XFRM_POLICY;
 		goto discard_it;
+	}
 
 	tcp_v4_fill_cb(skb, iph, th);
 
@@ -2166,6 +2172,7 @@ int tcp_v4_rcv(struct sk_buff *skb)
 
 do_time_wait:
 	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
+		drop_reason = SKB_DROP_REASON_XFRM_POLICY;
 		inet_twsk_put(inet_twsk(sk));
 		goto discard_it;
 	}
-- 
2.34.1

Re: [PATCH net-next 2/9] net: tcp: add skb drop reasons to tcp_v4_rcv()

[David Ahern]

On 2/15/22 8:54 PM, menglong8.dong@gmail.com wrote:
> @@ -2137,8 +2141,10 @@ int tcp_v4_rcv(struct sk_buff *skb)
>  
>  no_tcp_socket:
>  	drop_reason = SKB_DROP_REASON_NO_SOCKET;
> -	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb))
> +	if (!xfrm4_policy_check(NULL, XFRM_POLICY_IN, skb)) {
> +		drop_reason = SKB_DROP_REASON_XFRM_POLICY;

no socket reason trumps the xfrm failure; it was first.

>  		goto discard_it;
> +	}
>  
>  	tcp_v4_fill_cb(skb, iph, th);
>  

[PATCH net-next 3/9] net: tcp: use kfree_skb_reason() for tcp_v6_rcv()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Replace kfree_skb() used in tcp_v6_rcv() with kfree_skb_reason().

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 net/ipv6/tcp_ipv6.c | 25 ++++++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)

diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 0c648bf07f39..402ffbacc371 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1627,6 +1627,7 @@ static void tcp_v6_fill_cb(struct sk_buff *skb, const struct ipv6hdr *hdr,
 
 INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 {
+	enum skb_drop_reason drop_reason;
 	int sdif = inet6_sdif(skb);
 	int dif = inet6_iif(skb);
 	const struct tcphdr *th;
@@ -1636,6 +1637,7 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 	int ret;
 	struct net *net = dev_net(skb->dev);
 
+	drop_reason = SKB_DROP_REASON_NOT_SPECIFIED;
 	if (skb->pkt_type != PACKET_HOST)
 		goto discard_it;
 
@@ -1649,8 +1651,10 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 
 	th = (const struct tcphdr *)skb->data;
 
-	if (unlikely(th->doff < sizeof(struct tcphdr)/4))
+	if (unlikely(th->doff < sizeof(struct tcphdr) / 4)) {
+		drop_reason = SKB_DROP_REASON_PKT_TOO_SMALL;
 		goto bad_packet;
+	}
 	if (!pskb_may_pull(skb, th->doff*4))
 		goto discard_it;
 
@@ -1706,6 +1710,8 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 			hdr = ipv6_hdr(skb);
 			tcp_v6_fill_cb(skb, hdr, th);
 			nsk = tcp_check_req(sk, skb, req, false, &req_stolen);
+		} else {
+			drop_reason = SKB_DROP_REASON_SOCKET_FILTER;
 		}
 		if (!nsk) {
 			reqsk_put(req);
@@ -1741,14 +1747,18 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 		}
 	}
 
-	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb))
+	if (!xfrm6_policy_check(sk, XFRM_POLICY_IN, skb)) {
+		drop_reason = SKB_DROP_REASON_XFRM_POLICY;
 		goto discard_and_relse;
+	}
 
 	if (tcp_v6_inbound_md5_hash(sk, skb, dif, sdif))
 		goto discard_and_relse;
 
-	if (tcp_filter(sk, skb))
+	if (tcp_filter(sk, skb)) {
+		drop_reason = SKB_DROP_REASON_SOCKET_FILTER;
 		goto discard_and_relse;
+	}
 	th = (const struct tcphdr *)skb->data;
 	hdr = ipv6_hdr(skb);
 	tcp_v6_fill_cb(skb, hdr, th);
@@ -1779,13 +1789,17 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 	return ret ? -1 : 0;
 
 no_tcp_socket:
-	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
+	drop_reason = SKB_DROP_REASON_NO_SOCKET;
+	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
+		drop_reason = SKB_DROP_REASON_XFRM_POLICY;
 		goto discard_it;
+	}
 
 	tcp_v6_fill_cb(skb, hdr, th);
 
 	if (tcp_checksum_complete(skb)) {
 csum_error:
+		drop_reason = SKB_DROP_REASON_TCP_CSUM;
 		trace_tcp_bad_csum(skb);
 		__TCP_INC_STATS(net, TCP_MIB_CSUMERRORS);
 bad_packet:
@@ -1795,7 +1809,7 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 	}
 
 discard_it:
-	kfree_skb(skb);
+	kfree_skb_reason(skb, drop_reason);
 	return 0;
 
 discard_and_relse:
@@ -1806,6 +1820,7 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 
 do_time_wait:
 	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
+		drop_reason = SKB_DROP_REASON_XFRM_POLICY;
 		inet_twsk_put(inet_twsk(sk));
 		goto discard_it;
 	}
-- 
2.34.1

Re: [PATCH net-next 3/9] net: tcp: use kfree_skb_reason() for tcp_v6_rcv()

[David Ahern]

On 2/15/22 8:54 PM, menglong8.dong@gmail.com wrote:
> @@ -1779,13 +1789,17 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
>  	return ret ? -1 : 0;
>  
>  no_tcp_socket:
> -	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb))
> +	drop_reason = SKB_DROP_REASON_NO_SOCKET;
> +	if (!xfrm6_policy_check(NULL, XFRM_POLICY_IN, skb)) {
> +		drop_reason = SKB_DROP_REASON_XFRM_POLICY;

same here. First failure takes precedence.

[PATCH net-next 4/9] net: tcp: add skb drop reasons to tcp_v{4,6}_inbound_md5_hash()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Pass the address of drop reason to tcp_v4_inbound_md5_hash() and
tcp_v6_inbound_md5_hash() to store the reasons for skb drops when this
function fails. Therefore, the drop reason can be passed to
kfree_skb_reason() when the skb needs to be freed.

Following drop reasons are added:

SKB_DROP_REASON_TCP_MD5NOTFOUND
SKB_DROP_REASON_TCP_MD5UNEXPECTED
SKB_DROP_REASON_TCP_MD5FAILURE

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 include/linux/skbuff.h     |  7 +++++++
 include/trace/events/skb.h |  4 ++++
 net/ipv4/tcp_ipv4.c        | 13 +++++++++----
 net/ipv6/tcp_ipv6.c        | 11 ++++++++---
 4 files changed, 28 insertions(+), 7 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index a5adbf6b51e8..aea46b38cffa 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -346,6 +346,13 @@ enum skb_drop_reason {
 					 * udp packet drop out of
 					 * udp_memory_allocated.
 					 */
+	SKB_DROP_REASON_TCP_MD5NOTFOUND,	/* No MD5 hash and one
+						 * expected
+						 */
+	SKB_DROP_REASON_TCP_MD5UNEXPECTED,	/* MD5 hash and we're not
+						 * expecting one
+						 */
+	SKB_DROP_REASON_TCP_MD5FAILURE,	/* MD5 hash and its wrong */
 	SKB_DROP_REASON_MAX,
 };
 
diff --git a/include/trace/events/skb.h b/include/trace/events/skb.h
index cfcfd26399f7..46c06b0be850 100644
--- a/include/trace/events/skb.h
+++ b/include/trace/events/skb.h
@@ -27,6 +27,10 @@
 	EM(SKB_DROP_REASON_IP_NOPROTO, IP_NOPROTO)		\
 	EM(SKB_DROP_REASON_SOCKET_RCVBUFF, SOCKET_RCVBUFF)	\
 	EM(SKB_DROP_REASON_PROTO_MEM, PROTO_MEM)		\
+	EM(SKB_DROP_REASON_TCP_MD5NOTFOUND, TCP_MD5NOTFOUND)	\
+	EM(SKB_DROP_REASON_TCP_MD5UNEXPECTED,			\
+	   TCP_MD5UNEXPECTED)					\
+	EM(SKB_DROP_REASON_TCP_MD5FAILURE, TCP_MD5FAILURE)	\
 	EMe(SKB_DROP_REASON_MAX, MAX)
 
 #undef EM
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index a93921fb498f..3e7ab605dddc 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1412,7 +1412,8 @@ EXPORT_SYMBOL(tcp_v4_md5_hash_skb);
 /* Called with rcu_read_lock() */
 static bool tcp_v4_inbound_md5_hash(const struct sock *sk,
 				    const struct sk_buff *skb,
-				    int dif, int sdif)
+				    int dif, int sdif,
+				    enum skb_drop_reason *reason)
 {
 #ifdef CONFIG_TCP_MD5SIG
 	/*
@@ -1445,11 +1446,13 @@ static bool tcp_v4_inbound_md5_hash(const struct sock *sk,
 		return false;
 
 	if (hash_expected && !hash_location) {
+		*reason = SKB_DROP_REASON_TCP_MD5NOTFOUND;
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5NOTFOUND);
 		return true;
 	}
 
 	if (!hash_expected && hash_location) {
+		*reason = SKB_DROP_REASON_TCP_MD5UNEXPECTED;
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
 		return true;
 	}
@@ -1462,6 +1465,7 @@ static bool tcp_v4_inbound_md5_hash(const struct sock *sk,
 				      NULL, skb);
 
 	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
+		*reason = SKB_DROP_REASON_TCP_MD5FAILURE;
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5FAILURE);
 		net_info_ratelimited("MD5 Hash failed for (%pI4, %d)->(%pI4, %d)%s L3 index %d\n",
 				     &iph->saddr, ntohs(th->source),
@@ -1971,13 +1975,13 @@ static void tcp_v4_fill_cb(struct sk_buff *skb, const struct iphdr *iph,
 int tcp_v4_rcv(struct sk_buff *skb)
 {
 	struct net *net = dev_net(skb->dev);
+	enum skb_drop_reason drop_reason;
 	int sdif = inet_sdif(skb);
 	int dif = inet_iif(skb);
 	const struct iphdr *iph;
 	const struct tcphdr *th;
 	bool refcounted;
 	struct sock *sk;
-	int drop_reason;
 	int ret;
 
 	drop_reason = SKB_DROP_REASON_NOT_SPECIFIED;
@@ -2025,7 +2029,8 @@ int tcp_v4_rcv(struct sk_buff *skb)
 		struct sock *nsk;
 
 		sk = req->rsk_listener;
-		if (unlikely(tcp_v4_inbound_md5_hash(sk, skb, dif, sdif))) {
+		if (unlikely(tcp_v4_inbound_md5_hash(sk, skb, dif, sdif,
+						     &drop_reason))) {
 			sk_drops_add(sk, skb);
 			reqsk_put(req);
 			goto discard_it;
@@ -2099,7 +2104,7 @@ int tcp_v4_rcv(struct sk_buff *skb)
 		goto discard_and_relse;
 	}
 
-	if (tcp_v4_inbound_md5_hash(sk, skb, dif, sdif))
+	if (tcp_v4_inbound_md5_hash(sk, skb, dif, sdif, &drop_reason))
 		goto discard_and_relse;
 
 	nf_reset_ct(skb);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 402ffbacc371..dfefbc1eac5d 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -775,7 +775,8 @@ static int tcp_v6_md5_hash_skb(char *md5_hash,
 
 static bool tcp_v6_inbound_md5_hash(const struct sock *sk,
 				    const struct sk_buff *skb,
-				    int dif, int sdif)
+				    int dif, int sdif,
+				    enum skb_drop_reason *reason)
 {
 #ifdef CONFIG_TCP_MD5SIG
 	const __u8 *hash_location = NULL;
@@ -798,11 +799,13 @@ static bool tcp_v6_inbound_md5_hash(const struct sock *sk,
 		return false;
 
 	if (hash_expected && !hash_location) {
+		*reason = SKB_DROP_REASON_TCP_MD5NOTFOUND;
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5NOTFOUND);
 		return true;
 	}
 
 	if (!hash_expected && hash_location) {
+		*reason = SKB_DROP_REASON_TCP_MD5UNEXPECTED;
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5UNEXPECTED);
 		return true;
 	}
@@ -813,6 +816,7 @@ static bool tcp_v6_inbound_md5_hash(const struct sock *sk,
 				      NULL, skb);
 
 	if (genhash || memcmp(hash_location, newhash, 16) != 0) {
+		*reason = SKB_DROP_REASON_TCP_MD5FAILURE;
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPMD5FAILURE);
 		net_info_ratelimited("MD5 Hash %s for [%pI6c]:%u->[%pI6c]:%u L3 index %d\n",
 				     genhash ? "failed" : "mismatch",
@@ -1681,7 +1685,8 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 		struct sock *nsk;
 
 		sk = req->rsk_listener;
-		if (tcp_v6_inbound_md5_hash(sk, skb, dif, sdif)) {
+		if (tcp_v6_inbound_md5_hash(sk, skb, dif, sdif,
+					    &drop_reason)) {
 			sk_drops_add(sk, skb);
 			reqsk_put(req);
 			goto discard_it;
@@ -1752,7 +1757,7 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 		goto discard_and_relse;
 	}
 
-	if (tcp_v6_inbound_md5_hash(sk, skb, dif, sdif))
+	if (tcp_v6_inbound_md5_hash(sk, skb, dif, sdif, &drop_reason))
 		goto discard_and_relse;
 
 	if (tcp_filter(sk, skb)) {
-- 
2.34.1

[PATCH net-next 5/9] net: tcp: add skb drop reasons to tcp_add_backlog()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Pass the address of drop_reason to tcp_add_backlog() to store the
reasons for skb drops when fails. Following drop reasons are
introduced:

SKB_DROP_REASON_SOCKET_BACKLOG

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 include/linux/skbuff.h     | 4 ++++
 include/net/tcp.h          | 3 ++-
 include/trace/events/skb.h | 1 +
 net/ipv4/tcp_ipv4.c        | 7 +++++--
 net/ipv6/tcp_ipv6.c        | 2 +-
 5 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index aea46b38cffa..9a4424ceb7cb 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -353,6 +353,10 @@ enum skb_drop_reason {
 						 * expecting one
 						 */
 	SKB_DROP_REASON_TCP_MD5FAILURE,	/* MD5 hash and its wrong */
+	SKB_DROP_REASON_SOCKET_BACKLOG,	/* failed to add skb to socket
+					 * backlog (see
+					 * LINUX_MIB_TCPBACKLOGDROP)
+					 */
 	SKB_DROP_REASON_MAX,
 };
 
diff --git a/include/net/tcp.h b/include/net/tcp.h
index eff2487d972d..04f4650e0ff0 100644
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -1367,7 +1367,8 @@ static inline bool tcp_checksum_complete(struct sk_buff *skb)
 		__skb_checksum_complete(skb);
 }
 
-bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb);
+bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb,
+		     enum skb_drop_reason *reason);
 
 #ifdef CONFIG_INET
 void __sk_defer_free_flush(struct sock *sk);
diff --git a/include/trace/events/skb.h b/include/trace/events/skb.h
index 46c06b0be850..bfccd77e9071 100644
--- a/include/trace/events/skb.h
+++ b/include/trace/events/skb.h
@@ -31,6 +31,7 @@
 	EM(SKB_DROP_REASON_TCP_MD5UNEXPECTED,			\
 	   TCP_MD5UNEXPECTED)					\
 	EM(SKB_DROP_REASON_TCP_MD5FAILURE, TCP_MD5FAILURE)	\
+	EM(SKB_DROP_REASON_SOCKET_BACKLOG, SOCKET_BACKLOG)	\
 	EMe(SKB_DROP_REASON_MAX, MAX)
 
 #undef EM
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 3e7ab605dddc..15ef1bcff84f 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1811,7 +1811,8 @@ int tcp_v4_early_demux(struct sk_buff *skb)
 	return 0;
 }
 
-bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
+bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb,
+		     enum skb_drop_reason *reason)
 {
 	u32 limit, tail_gso_size, tail_gso_segs;
 	struct skb_shared_info *shinfo;
@@ -1837,6 +1838,7 @@ bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
 	if (unlikely(tcp_checksum_complete(skb))) {
 		bh_unlock_sock(sk);
 		trace_tcp_bad_csum(skb);
+		*reason = SKB_DROP_REASON_TCP_CSUM;
 		__TCP_INC_STATS(sock_net(sk), TCP_MIB_CSUMERRORS);
 		__TCP_INC_STATS(sock_net(sk), TCP_MIB_INERRS);
 		return true;
@@ -1925,6 +1927,7 @@ bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb)
 
 	if (unlikely(sk_add_backlog(sk, skb, limit))) {
 		bh_unlock_sock(sk);
+		*reason = SKB_DROP_REASON_SOCKET_BACKLOG;
 		__NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPBACKLOGDROP);
 		return true;
 	}
@@ -2133,7 +2136,7 @@ int tcp_v4_rcv(struct sk_buff *skb)
 	if (!sock_owned_by_user(sk)) {
 		ret = tcp_v4_do_rcv(sk, skb);
 	} else {
-		if (tcp_add_backlog(sk, skb))
+		if (tcp_add_backlog(sk, skb, &drop_reason))
 			goto discard_and_relse;
 	}
 	bh_unlock_sock(sk);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index dfefbc1eac5d..15090fe5af90 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1784,7 +1784,7 @@ INDIRECT_CALLABLE_SCOPE int tcp_v6_rcv(struct sk_buff *skb)
 	if (!sock_owned_by_user(sk)) {
 		ret = tcp_v6_do_rcv(sk, skb);
 	} else {
-		if (tcp_add_backlog(sk, skb))
+		if (tcp_add_backlog(sk, skb, &drop_reason))
 			goto discard_and_relse;
 	}
 	bh_unlock_sock(sk);
-- 
2.34.1

[PATCH net-next 6/9] net: tcp: use kfree_skb_reason() for tcp_v{4,6}_do_rcv()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Replace kfree_skb() used in tcp_v4_do_rcv() and tcp_v6_do_rcv() with
kfree_skb_reason().

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 net/ipv4/tcp_ipv4.c | 5 ++++-
 net/ipv6/tcp_ipv6.c | 4 +++-
 2 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index 15ef1bcff84f..036b855c1b14 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -1708,6 +1708,7 @@ INDIRECT_CALLABLE_DECLARE(struct dst_entry *ipv4_dst_check(struct dst_entry *,
  */
 int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb)
 {
+	enum skb_drop_reason reason;
 	struct sock *rsk;
 
 	if (sk->sk_state == TCP_ESTABLISHED) { /* Fast path */
@@ -1730,6 +1731,7 @@ int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb)
 		return 0;
 	}
 
+	reason = SKB_DROP_REASON_NOT_SPECIFIED;
 	if (tcp_checksum_complete(skb))
 		goto csum_err;
 
@@ -1757,7 +1759,7 @@ int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb)
 reset:
 	tcp_v4_send_reset(rsk, skb);
 discard:
-	kfree_skb(skb);
+	kfree_skb_reason(skb, reason);
 	/* Be careful here. If this function gets more complicated and
 	 * gcc suffers from register pressure on the x86, sk (in %ebx)
 	 * might be destroyed here. This current version compiles correctly,
@@ -1766,6 +1768,7 @@ int tcp_v4_do_rcv(struct sock *sk, struct sk_buff *skb)
 	return 0;
 
 csum_err:
+	reason = SKB_DROP_REASON_TCP_CSUM;
 	trace_tcp_bad_csum(skb);
 	TCP_INC_STATS(sock_net(sk), TCP_MIB_CSUMERRORS);
 	TCP_INC_STATS(sock_net(sk), TCP_MIB_INERRS);
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index 15090fe5af90..4d4af9b15c83 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1476,6 +1476,7 @@ int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb)
 {
 	struct ipv6_pinfo *np = tcp_inet6_sk(sk);
 	struct sk_buff *opt_skb = NULL;
+	enum skb_drop_reason reason;
 	struct tcp_sock *tp;
 
 	/* Imagine: socket is IPv6. IPv4 packet arrives,
@@ -1563,9 +1564,10 @@ int tcp_v6_do_rcv(struct sock *sk, struct sk_buff *skb)
 discard:
 	if (opt_skb)
 		__kfree_skb(opt_skb);
-	kfree_skb(skb);
+	kfree_skb_reason(skb, reason);
 	return 0;
 csum_err:
+	reason = SKB_DROP_REASON_TCP_CSUM;
 	trace_tcp_bad_csum(skb);
 	TCP_INC_STATS(sock_net(sk), TCP_MIB_CSUMERRORS);
 	TCP_INC_STATS(sock_net(sk), TCP_MIB_INERRS);
-- 
2.34.1

Re: [PATCH net-next 6/9] net: tcp: use kfree_skb_reason() for tcp_v{4,6}_do_rcv()

[Jakub Kicinski]

On Wed, 16 Feb 2022 11:54:23 +0800 menglong8.dong@gmail.com wrote:
> From: Menglong Dong <imagedong@tencent.com>
> 
> Replace kfree_skb() used in tcp_v4_do_rcv() and tcp_v6_do_rcv() with
> kfree_skb_reason().
> 
> Signed-off-by: Menglong Dong <imagedong@tencent.com>

Clang says:

net/ipv6/tcp_ipv6.c:1556:6: warning: variable 'reason' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
        if (tcp_rcv_state_process(sk, skb))
            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
net/ipv6/tcp_ipv6.c:1567:24: note: uninitialized use occurs here
        kfree_skb_reason(skb, reason);
                              ^~~~~~
net/ipv6/tcp_ipv6.c:1556:2: note: remove the 'if' if its condition is always false
        if (tcp_rcv_state_process(sk, skb))
        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
net/ipv6/tcp_ipv6.c:1547:8: warning: variable 'reason' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
                        if (tcp_child_process(sk, nsk, skb))
                            ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
net/ipv6/tcp_ipv6.c:1567:24: note: uninitialized use occurs here
        kfree_skb_reason(skb, reason);
                              ^~~~~~
net/ipv6/tcp_ipv6.c:1547:4: note: remove the 'if' if its condition is always false
                        if (tcp_child_process(sk, nsk, skb))
                        ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
net/ipv6/tcp_ipv6.c:1543:7: warning: variable 'reason' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
                if (!nsk)
                    ^~~~
net/ipv6/tcp_ipv6.c:1567:24: note: uninitialized use occurs here
        kfree_skb_reason(skb, reason);
                              ^~~~~~
net/ipv6/tcp_ipv6.c:1543:3: note: remove the 'if' if its condition is always false
                if (!nsk)
                ^~~~~~~~~
net/ipv6/tcp_ipv6.c:1479:2: note: variable 'reason' is declared here
        enum skb_drop_reason reason;
        ^

[PATCH net-next 7/9] net: tcp: use tcp_drop_reason() for tcp_rcv_established()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Replace tcp_drop() used in tcp_rcv_established() with tcp_drop_reason().
Following drop reasons are added:

SKB_DROP_REASON_TCP_FLAGS

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 include/linux/skbuff.h     | 1 +
 include/trace/events/skb.h | 1 +
 net/ipv4/tcp_input.c       | 9 +++++++--
 3 files changed, 9 insertions(+), 2 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 9a4424ceb7cb..dcf9d8bd0079 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -357,6 +357,7 @@ enum skb_drop_reason {
 					 * backlog (see
 					 * LINUX_MIB_TCPBACKLOGDROP)
 					 */
+	SKB_DROP_REASON_TCP_FLAGS,	/* TCP flags invalid */
 	SKB_DROP_REASON_MAX,
 };
 
diff --git a/include/trace/events/skb.h b/include/trace/events/skb.h
index bfccd77e9071..d332e7313a61 100644
--- a/include/trace/events/skb.h
+++ b/include/trace/events/skb.h
@@ -32,6 +32,7 @@
 	   TCP_MD5UNEXPECTED)					\
 	EM(SKB_DROP_REASON_TCP_MD5FAILURE, TCP_MD5FAILURE)	\
 	EM(SKB_DROP_REASON_SOCKET_BACKLOG, SOCKET_BACKLOG)	\
+	EM(SKB_DROP_REASON_TCP_FLAGS, TCP_FLAGS)		\
 	EMe(SKB_DROP_REASON_MAX, MAX)
 
 #undef EM
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 0a2740add404..16ee1127e25d 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -5787,6 +5787,7 @@ static bool tcp_validate_incoming(struct sock *sk, struct sk_buff *skb,
  */
 void tcp_rcv_established(struct sock *sk, struct sk_buff *skb)
 {
+	enum skb_drop_reason reason = SKB_DROP_REASON_NOT_SPECIFIED;
 	const struct tcphdr *th = (const struct tcphdr *)skb->data;
 	struct tcp_sock *tp = tcp_sk(sk);
 	unsigned int len = skb->len;
@@ -5875,6 +5876,7 @@ void tcp_rcv_established(struct sock *sk, struct sk_buff *skb)
 				tp->rcv_rtt_last_tsecr = tp->rx_opt.rcv_tsecr;
 				return;
 			} else { /* Header too small */
+				reason = SKB_DROP_REASON_PKT_TOO_SMALL;
 				TCP_INC_STATS(sock_net(sk), TCP_MIB_INERRS);
 				goto discard;
 			}
@@ -5930,8 +5932,10 @@ void tcp_rcv_established(struct sock *sk, struct sk_buff *skb)
 	if (len < (th->doff << 2) || tcp_checksum_complete(skb))
 		goto csum_error;
 
-	if (!th->ack && !th->rst && !th->syn)
+	if (!th->ack && !th->rst && !th->syn) {
+		reason = SKB_DROP_REASON_TCP_FLAGS;
 		goto discard;
+	}
 
 	/*
 	 *	Standard slow path.
@@ -5957,12 +5961,13 @@ void tcp_rcv_established(struct sock *sk, struct sk_buff *skb)
 	return;
 
 csum_error:
+	reason = SKB_DROP_REASON_TCP_CSUM;
 	trace_tcp_bad_csum(skb);
 	TCP_INC_STATS(sock_net(sk), TCP_MIB_CSUMERRORS);
 	TCP_INC_STATS(sock_net(sk), TCP_MIB_INERRS);
 
 discard:
-	tcp_drop(sk, skb);
+	tcp_drop_reason(sk, skb, reason);
 }
 EXPORT_SYMBOL(tcp_rcv_established);
 
-- 
2.34.1

[PATCH net-next 8/9] net: tcp: use tcp_drop_reason() for tcp_data_queue()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Replace tcp_drop() used in tcp_data_queue() with tcp_drop_reason().
Following drop reasons are introduced:

SKB_DROP_REASON_TCP_ZEROWINDOW
SKB_DROP_REASON_TCP_OLD_DATA
SKB_DROP_REASON_TCP_OVERWINDOW

SKB_DROP_REASON_TCP_OLD_DATA is used for the case that end_seq of skb
less than the left edges of receive window. (Maybe there is a better
name?)

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 include/linux/skbuff.h     | 13 +++++++++++++
 include/trace/events/skb.h |  3 +++
 net/ipv4/tcp_input.c       | 13 +++++++++++--
 3 files changed, 27 insertions(+), 2 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index dcf9d8bd0079..62a0d7d78f6f 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -358,6 +358,19 @@ enum skb_drop_reason {
 					 * LINUX_MIB_TCPBACKLOGDROP)
 					 */
 	SKB_DROP_REASON_TCP_FLAGS,	/* TCP flags invalid */
+	SKB_DROP_REASON_TCP_ZEROWINDOW,	/* TCP receive window size is zero,
+					 * see LINUX_MIB_TCPZEROWINDOWDROP
+					 */
+	SKB_DROP_REASON_TCP_OLD_DATA,	/* the TCP data reveived is already
+					 * received before (spurious retrans
+					 * may happened), see
+					 * LINUX_MIB_DELAYEDACKLOST
+					 */
+	SKB_DROP_REASON_TCP_OVERWINDOW,	/* the TCP data is out of window,
+					 * the seq of the first byte exceed
+					 * the right edges of receive
+					 * window
+					 */
 	SKB_DROP_REASON_MAX,
 };
 
diff --git a/include/trace/events/skb.h b/include/trace/events/skb.h
index d332e7313a61..cc1c8f7eaf72 100644
--- a/include/trace/events/skb.h
+++ b/include/trace/events/skb.h
@@ -33,6 +33,9 @@
 	EM(SKB_DROP_REASON_TCP_MD5FAILURE, TCP_MD5FAILURE)	\
 	EM(SKB_DROP_REASON_SOCKET_BACKLOG, SOCKET_BACKLOG)	\
 	EM(SKB_DROP_REASON_TCP_FLAGS, TCP_FLAGS)		\
+	EM(SKB_DROP_REASON_TCP_ZEROWINDOW, TCP_ZEROWINDOW)	\
+	EM(SKB_DROP_REASON_TCP_OLD_DATA, TCP_OLD_DATA)		\
+	EM(SKB_DROP_REASON_TCP_OVERWINDOW, TCP_OVERWINDOW)	\
 	EMe(SKB_DROP_REASON_MAX, MAX)
 
 #undef EM
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 16ee1127e25d..0a4ca818d580 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -4988,6 +4988,7 @@ void tcp_data_ready(struct sock *sk)
 static void tcp_data_queue(struct sock *sk, struct sk_buff *skb)
 {
 	struct tcp_sock *tp = tcp_sk(sk);
+	enum skb_drop_reason reason;
 	bool fragstolen;
 	int eaten;
 
@@ -5006,6 +5007,7 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb)
 	skb_dst_drop(skb);
 	__skb_pull(skb, tcp_hdr(skb)->doff * 4);
 
+	reason = SKB_DROP_REASON_NOT_SPECIFIED;
 	tp->rx_opt.dsack = 0;
 
 	/*  Queue data for delivery to the user.
@@ -5014,6 +5016,7 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb)
 	 */
 	if (TCP_SKB_CB(skb)->seq == tp->rcv_nxt) {
 		if (tcp_receive_window(tp) == 0) {
+			reason = SKB_DROP_REASON_TCP_ZEROWINDOW;
 			NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPZEROWINDOWDROP);
 			goto out_of_window;
 		}
@@ -5023,6 +5026,7 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb)
 		if (skb_queue_len(&sk->sk_receive_queue) == 0)
 			sk_forced_mem_schedule(sk, skb->truesize);
 		else if (tcp_try_rmem_schedule(sk, skb, skb->truesize)) {
+			reason = SKB_DROP_REASON_PROTO_MEM;
 			NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPRCVQDROP);
 			sk->sk_data_ready(sk);
 			goto drop;
@@ -5059,6 +5063,7 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb)
 	if (!after(TCP_SKB_CB(skb)->end_seq, tp->rcv_nxt)) {
 		tcp_rcv_spurious_retrans(sk, skb);
 		/* A retransmit, 2nd most common case.  Force an immediate ack. */
+		reason = SKB_DROP_REASON_TCP_OLD_DATA;
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_DELAYEDACKLOST);
 		tcp_dsack_set(sk, TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq);
 
@@ -5066,13 +5071,16 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb)
 		tcp_enter_quickack_mode(sk, TCP_MAX_QUICKACKS);
 		inet_csk_schedule_ack(sk);
 drop:
-		tcp_drop(sk, skb);
+		tcp_drop_reason(sk, skb, reason);
 		return;
 	}
 
 	/* Out of window. F.e. zero window probe. */
-	if (!before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt + tcp_receive_window(tp)))
+	if (!before(TCP_SKB_CB(skb)->seq,
+		    tp->rcv_nxt + tcp_receive_window(tp))) {
+		reason = SKB_DROP_REASON_TCP_OVERWINDOW;
 		goto out_of_window;
+	}
 
 	if (before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) {
 		/* Partial packet, seq < rcv_next < end_seq */
@@ -5082,6 +5090,7 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb)
 		 * remembering D-SACK for its head made in previous line.
 		 */
 		if (!tcp_receive_window(tp)) {
+			reason = SKB_DROP_REASON_TCP_ZEROWINDOW;
 			NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPZEROWINDOWDROP);
 			goto out_of_window;
 		}
-- 
2.34.1

[PATCH net-next 9/9] net: tcp: use tcp_drop_reason() for tcp_data_queue_ofo()

[menglong8.dong]

From: Menglong Dong <imagedong@tencent.com>

Replace tcp_drop() used in tcp_data_queue_ofo with tcp_drop_reason().
Following drop reasons are introduced:

SKB_DROP_REASON_TCP_OFOMERGE

Signed-off-by: Menglong Dong <imagedong@tencent.com>
---
 include/linux/skbuff.h     |  3 +++
 include/trace/events/skb.h |  1 +
 net/ipv4/tcp_input.c       | 10 ++++++----
 3 files changed, 10 insertions(+), 4 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 62a0d7d78f6f..73ed01d87e43 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -371,6 +371,9 @@ enum skb_drop_reason {
 					 * the right edges of receive
 					 * window
 					 */
+	SKB_DROP_REASON_TCP_OFOMERGE,	/* the data of skb is already in
+					 * the ofo queue.
+					 */
 	SKB_DROP_REASON_MAX,
 };
 
diff --git a/include/trace/events/skb.h b/include/trace/events/skb.h
index cc1c8f7eaf72..2ab7193313aa 100644
--- a/include/trace/events/skb.h
+++ b/include/trace/events/skb.h
@@ -36,6 +36,7 @@
 	EM(SKB_DROP_REASON_TCP_ZEROWINDOW, TCP_ZEROWINDOW)	\
 	EM(SKB_DROP_REASON_TCP_OLD_DATA, TCP_OLD_DATA)		\
 	EM(SKB_DROP_REASON_TCP_OVERWINDOW, TCP_OVERWINDOW)	\
+	EM(SKB_DROP_REASON_TCP_OFOMERGE, TCP_OFOMERGE)		\
 	EMe(SKB_DROP_REASON_MAX, MAX)
 
 #undef EM
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 0a4ca818d580..1fc422af11f8 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -4779,7 +4779,7 @@ static void tcp_data_queue_ofo(struct sock *sk, struct sk_buff *skb)
 	if (unlikely(tcp_try_rmem_schedule(sk, skb, skb->truesize))) {
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPOFODROP);
 		sk->sk_data_ready(sk);
-		tcp_drop(sk, skb);
+		tcp_drop_reason(sk, skb, SKB_DROP_REASON_PROTO_MEM);
 		return;
 	}
 
@@ -4842,7 +4842,8 @@ static void tcp_data_queue_ofo(struct sock *sk, struct sk_buff *skb)
 				/* All the bits are present. Drop. */
 				NET_INC_STATS(sock_net(sk),
 					      LINUX_MIB_TCPOFOMERGE);
-				tcp_drop(sk, skb);
+				tcp_drop_reason(sk, skb,
+						SKB_DROP_REASON_TCP_OFOMERGE);
 				skb = NULL;
 				tcp_dsack_set(sk, seq, end_seq);
 				goto add_sack;
@@ -4861,7 +4862,8 @@ static void tcp_data_queue_ofo(struct sock *sk, struct sk_buff *skb)
 						 TCP_SKB_CB(skb1)->end_seq);
 				NET_INC_STATS(sock_net(sk),
 					      LINUX_MIB_TCPOFOMERGE);
-				tcp_drop(sk, skb1);
+				tcp_drop_reason(sk, skb1,
+						SKB_DROP_REASON_TCP_OFOMERGE);
 				goto merge_right;
 			}
 		} else if (tcp_ooo_try_coalesce(sk, skb1,
@@ -4889,7 +4891,7 @@ static void tcp_data_queue_ofo(struct sock *sk, struct sk_buff *skb)
 		tcp_dsack_extend(sk, TCP_SKB_CB(skb1)->seq,
 				 TCP_SKB_CB(skb1)->end_seq);
 		NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPOFOMERGE);
-		tcp_drop(sk, skb1);
+		tcp_drop_reason(sk, skb1, SKB_DROP_REASON_TCP_OFOMERGE);
 	}
 	/* If there is no skb after us, we are the last_skb ! */
 	if (!skb1)
-- 
2.34.1

Re: [PATCH net-next 0/9] net: add skb drop reasons to TCP packet receive

[Eric Dumazet]

On Tue, Feb 15, 2022 at 7:54 PM <menglong8.dong@gmail.com> wrote:
>
> From: Menglong Dong <imagedong@tencent.com>
>
> In this series patches, reasons for skb drops are added to TCP layer, and
> both TCPv4 and TCPv6 are considered.
>
> in this series patches, the process of packet ingress in TCP layer is
> considered, as skb drops hardly happens in the egress path.
>
> However, it's a little complex for TCP state processing, as I find that
> it's hard to report skb drop reasons to where it is freed. For example,
> when skb is dropped in tcp_rcv_state_process(), the reason can be caused
> by the call of tcp_v4_conn_request(), and it's hard to return a drop
> reason from tcp_v4_conn_request(). So I just skip such case for this
> moment.
>

I think you should add at least in this cover letter, or better in a
document that can be amended,
how this can be used on a typical TCP session.
For someone who is having issues with TCP flows, what would they need to do.
Think of something that we (kernel dev) could copy paste to future
email replies.
It might be mostly clear for some of us reviewing patches at this
moment, but in one year we will all forget about the details.


>
> Menglong Dong (9):
>   net: tcp: introduce tcp_drop_reason()
>   net: tcp: add skb drop reasons to tcp_v4_rcv()
>   net: tcp: use kfree_skb_reason() for tcp_v6_rcv()
>   net: tcp: add skb drop reasons to tcp_v{4,6}_inbound_md5_hash()
>   net: tcp: add skb drop reasons to tcp_add_backlog()
>   net: tcp: use kfree_skb_reason() for tcp_v{4,6}_do_rcv()
>   net: tcp: use tcp_drop_reason() for tcp_rcv_established()
>   net: tcp: use tcp_drop_reason() for tcp_data_queue()
>   net: tcp: use tcp_drop_reason() for tcp_data_queue_ofo()
>
>  include/linux/skbuff.h     | 28 +++++++++++++++++++++++++
>  include/net/tcp.h          |  3 ++-
>  include/trace/events/skb.h | 10 +++++++++
>  net/ipv4/tcp_input.c       | 42 +++++++++++++++++++++++++++++---------
>  net/ipv4/tcp_ipv4.c        | 36 ++++++++++++++++++++++++--------
>  net/ipv6/tcp_ipv6.c        | 42 +++++++++++++++++++++++++++++---------
>  6 files changed, 131 insertions(+), 30 deletions(-)
>
> --
> 2.34.1
>

Re: [PATCH net-next 0/9] net: add skb drop reasons to TCP packet receive

[Menglong Dong]

On Thu, Feb 17, 2022 at 1:05 AM Eric Dumazet <edumazet@google.com> wrote:
>
> On Tue, Feb 15, 2022 at 7:54 PM <menglong8.dong@gmail.com> wrote:
> >
> > From: Menglong Dong <imagedong@tencent.com>
> >
> > In this series patches, reasons for skb drops are added to TCP layer, and
> > both TCPv4 and TCPv6 are considered.
> >
> > in this series patches, the process of packet ingress in TCP layer is
> > considered, as skb drops hardly happens in the egress path.
> >
> > However, it's a little complex for TCP state processing, as I find that
> > it's hard to report skb drop reasons to where it is freed. For example,
> > when skb is dropped in tcp_rcv_state_process(), the reason can be caused
> > by the call of tcp_v4_conn_request(), and it's hard to return a drop
> > reason from tcp_v4_conn_request(). So I just skip such case for this
> > moment.
> >
>
> I think you should add at least in this cover letter, or better in a
> document that can be amended,
> how this can be used on a typical TCP session.
> For someone who is having issues with TCP flows, what would they need to do.
> Think of something that we (kernel dev) could copy paste to future
> email replies.
> It might be mostly clear for some of us reviewing patches at this
> moment, but in one year we will all forget about the details.
>

Yeah, this cover letter seems too simple to explain what we
are doing.

I'll describe in detail what this series patches do and how
they can be used in the cover letter, and give some examples.


Thanks!
Menglong Dong
>
> >
> > Menglong Dong (9):
> >   net: tcp: introduce tcp_drop_reason()
> >   net: tcp: add skb drop reasons to tcp_v4_rcv()
> >   net: tcp: use kfree_skb_reason() for tcp_v6_rcv()
> >   net: tcp: add skb drop reasons to tcp_v{4,6}_inbound_md5_hash()
> >   net: tcp: add skb drop reasons to tcp_add_backlog()
> >   net: tcp: use kfree_skb_reason() for tcp_v{4,6}_do_rcv()
> >   net: tcp: use tcp_drop_reason() for tcp_rcv_established()
> >   net: tcp: use tcp_drop_reason() for tcp_data_queue()
> >   net: tcp: use tcp_drop_reason() for tcp_data_queue_ofo()
> >
> >  include/linux/skbuff.h     | 28 +++++++++++++++++++++++++
> >  include/net/tcp.h          |  3 ++-
> >  include/trace/events/skb.h | 10 +++++++++
> >  net/ipv4/tcp_input.c       | 42 +++++++++++++++++++++++++++++---------
> >  net/ipv4/tcp_ipv4.c        | 36 ++++++++++++++++++++++++--------
> >  net/ipv6/tcp_ipv6.c        | 42 +++++++++++++++++++++++++++++---------
> >  6 files changed, 131 insertions(+), 30 deletions(-)
> >
> > --
> > 2.34.1
> >