[PATCH v2 net-next 0/9] ipv6: snmp: avoid performance issue with RATELIMITHOST

[PATCH v2 net-next 0/9] ipv6: snmp: avoid performance issue with RATELIMITHOST

[Eric Dumazet]

Addition of ICMP6_MIB_RATELIMITHOST in commit d0941130c9351
("icmp: Add counters for rate limits") introduced a performance
drop in case of DOS (like receiving UDP packets
to closed ports).

Per netns ICMP6_MIB_RATELIMITHOST tracking uses per-cpu
storage and is enough, we do not need per-device and slow tracking
for this metric.

In v2 of this series, I completed the removal of SNMP_MIB_SENTINEL
in all the kernel for consistency.

v1: had an issue : https://lore.kernel.org/netdev/20250904092432.113c4940@kernel.org/

Eric Dumazet (9):
  ipv6: snmp: remove icmp6type2name[]
  ipv6: snmp: do not use SNMP_MIB_SENTINEL anymore
  ipv6: snmp: do not track per idev ICMP6_MIB_RATELIMITHOST
  ipv4: snmp: do not use SNMP_MIB_SENTINEL anymore
  mptcp: snmp: do not use SNMP_MIB_SENTINEL anymore
  sctp: snmp: do not use SNMP_MIB_SENTINEL anymore
  tls: snmp: do not use SNMP_MIB_SENTINEL anymore
  xfrm: snmp: do not use SNMP_MIB_SENTINEL anymore
  net: snmp: remove SNMP_MIB_SENTINEL

 include/net/ip.h	  |  9 +++--
 include/net/snmp.h   |  5 ---
 net/ipv4/proc.c	  | 65 +++++++++++++++----------------
 net/ipv6/icmp.c	  |  3 +-
 net/ipv6/proc.c	  | 91 ++++++++++++++++++++++++--------------------
 net/mptcp/mib.c	  | 12 +++---
 net/sctp/proc.c	  | 12 +++---
 net/tls/tls_proc.c   | 10 +++--
 net/xfrm/xfrm_proc.c | 12 +++---
 9 files changed, 113 insertions(+), 106 deletions(-)

--
2.51.0.355.g5224444f11-goog

[PATCH v2 net-next 1/9] ipv6: snmp: remove icmp6type2name[]

[Eric Dumazet]

This 2KB array can be replaced by a switch() to save space.

Before:
$ size net/ipv6/proc.o
   text	   data	    bss	    dec	    hex	filename
   6410	    624	      0	   7034	   1b7a	net/ipv6/proc.o

After:
$ size net/ipv6/proc.o
   text	   data	    bss	    dec	    hex	filename
   5516	    592	      0	   6108	   17dc	net/ipv6/proc.o

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 net/ipv6/proc.c | 44 ++++++++++++++++++++++----------------------
 1 file changed, 22 insertions(+), 22 deletions(-)

diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c
index 752327b10dde..e96f14a36834 100644
--- a/net/ipv6/proc.c
+++ b/net/ipv6/proc.c
@@ -99,26 +99,6 @@ static const struct snmp_mib snmp6_icmp6_list[] = {
 	SNMP_MIB_SENTINEL
 };
 
-/* RFC 4293 v6 ICMPMsgStatsTable; named items for RFC 2466 compatibility */
-static const char *const icmp6type2name[256] = {
-	[ICMPV6_DEST_UNREACH] = "DestUnreachs",
-	[ICMPV6_PKT_TOOBIG] = "PktTooBigs",
-	[ICMPV6_TIME_EXCEED] = "TimeExcds",
-	[ICMPV6_PARAMPROB] = "ParmProblems",
-	[ICMPV6_ECHO_REQUEST] = "Echos",
-	[ICMPV6_ECHO_REPLY] = "EchoReplies",
-	[ICMPV6_MGM_QUERY] = "GroupMembQueries",
-	[ICMPV6_MGM_REPORT] = "GroupMembResponses",
-	[ICMPV6_MGM_REDUCTION] = "GroupMembReductions",
-	[ICMPV6_MLD2_REPORT] = "MLDv2Reports",
-	[NDISC_ROUTER_ADVERTISEMENT] = "RouterAdvertisements",
-	[NDISC_ROUTER_SOLICITATION] = "RouterSolicits",
-	[NDISC_NEIGHBOUR_ADVERTISEMENT] = "NeighborAdvertisements",
-	[NDISC_NEIGHBOUR_SOLICITATION] = "NeighborSolicits",
-	[NDISC_REDIRECT] = "Redirects",
-};
-
-
 static const struct snmp_mib snmp6_udp6_list[] = {
 	SNMP_MIB_ITEM("Udp6InDatagrams", UDP_MIB_INDATAGRAMS),
 	SNMP_MIB_ITEM("Udp6NoPorts", UDP_MIB_NOPORTS),
@@ -151,11 +131,31 @@ static void snmp6_seq_show_icmpv6msg(struct seq_file *seq, atomic_long_t *smib)
 
 	/* print by name -- deprecated items */
 	for (i = 0; i < ICMP6MSG_MIB_MAX; i++) {
+		const char *p = NULL;
 		int icmptype;
-		const char *p;
+
+#define CASE(TYP, STR) case TYP: p = STR; break;
 
 		icmptype = i & 0xff;
-		p = icmp6type2name[icmptype];
+		switch (icmptype) {
+/* RFC 4293 v6 ICMPMsgStatsTable; named items for RFC 2466 compatibility */
+		CASE(ICMPV6_DEST_UNREACH,	"DestUnreachs")
+		CASE(ICMPV6_PKT_TOOBIG,		"PktTooBigs")
+		CASE(ICMPV6_TIME_EXCEED,	"TimeExcds")
+		CASE(ICMPV6_PARAMPROB,		"ParmProblems")
+		CASE(ICMPV6_ECHO_REQUEST,	"Echos")
+		CASE(ICMPV6_ECHO_REPLY,		"EchoReplies")
+		CASE(ICMPV6_MGM_QUERY,		"GroupMembQueries")
+		CASE(ICMPV6_MGM_REPORT,		"GroupMembResponses")
+		CASE(ICMPV6_MGM_REDUCTION,	"GroupMembReductions")
+		CASE(ICMPV6_MLD2_REPORT,	"MLDv2Reports")
+		CASE(NDISC_ROUTER_ADVERTISEMENT, "RouterAdvertisements")
+		CASE(NDISC_ROUTER_SOLICITATION, "RouterSolicits")
+		CASE(NDISC_NEIGHBOUR_ADVERTISEMENT, "NeighborAdvertisements")
+		CASE(NDISC_NEIGHBOUR_SOLICITATION, "NeighborSolicits")
+		CASE(NDISC_REDIRECT,		"Redirects")
+		}
+#undef CASE
 		if (!p)	/* don't print un-named types here */
 			continue;
 		snprintf(name, sizeof(name), "Icmp6%s%s",
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 1/9] ipv6: snmp: remove icmp6type2name[]

[Sabrina Dubroca]

2025-09-05, 16:58:05 +0000, Eric Dumazet wrote:
> This 2KB array can be replaced by a switch() to save space.
> 
> Before:
> $ size net/ipv6/proc.o
>    text	   data	    bss	    dec	    hex	filename
>    6410	    624	      0	   7034	   1b7a	net/ipv6/proc.o
> 
> After:
> $ size net/ipv6/proc.o
>    text	   data	    bss	    dec	    hex	filename
>    5516	    592	      0	   6108	   17dc	net/ipv6/proc.o
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> ---
>  net/ipv6/proc.c | 44 ++++++++++++++++++++++----------------------
>  1 file changed, 22 insertions(+), 22 deletions(-)

Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>

-- 
Sabrina

[PATCH v2 net-next 2/9] ipv6: snmp: do not use SNMP_MIB_SENTINEL anymore

[Eric Dumazet]

Use ARRAY_SIZE(), so that we know the limit at compile time.

Following patch needs this preliminary change.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/ip.h | 24 ++++++++++++++++++++++++
 net/ipv6/proc.c  | 43 ++++++++++++++++++++++++-------------------
 2 files changed, 48 insertions(+), 19 deletions(-)

diff --git a/include/net/ip.h b/include/net/ip.h
index 6dbd2bf8fa9c..a1624e8db1ab 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -338,6 +338,19 @@ static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_o
 	} \
 }
 
+#define snmp_get_cpu_field64_batch_cnt(buff64, stats_list, cnt,	\
+				       mib_statistic, offset)	\
+{ \
+	int i, c; \
+	for_each_possible_cpu(c) { \
+		for (i = 0; i < cnt; i++) \
+			buff64[i] += snmp_get_cpu_field64( \
+					mib_statistic, \
+					c, stats_list[i].entry, \
+					offset); \
+	} \
+}
+
 #define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \
 { \
 	int i, c; \
@@ -349,6 +362,17 @@ static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_o
 	} \
 }
 
+#define snmp_get_cpu_field_batch_cnt(buff, stats_list, cnt, mib_statistic) \
+{ \
+	int i, c; \
+	for_each_possible_cpu(c) { \
+		for (i = 0; i < cnt; i++) \
+			buff[i] += snmp_get_cpu_field( \
+						mib_statistic, \
+						c, stats_list[i].entry); \
+	} \
+}
+
 static inline void inet_get_local_port_range(const struct net *net, int *low, int *high)
 {
 	u32 range = READ_ONCE(net->ipv4.ip_local_ports.range);
diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c
index e96f14a36834..92ed04729c2f 100644
--- a/net/ipv6/proc.c
+++ b/net/ipv6/proc.c
@@ -85,7 +85,6 @@ static const struct snmp_mib snmp6_ipstats_list[] = {
 	SNMP_MIB_ITEM("Ip6InECT0Pkts", IPSTATS_MIB_ECT0PKTS),
 	SNMP_MIB_ITEM("Ip6InCEPkts", IPSTATS_MIB_CEPKTS),
 	SNMP_MIB_ITEM("Ip6OutTransmits", IPSTATS_MIB_OUTPKTS),
-	SNMP_MIB_SENTINEL
 };
 
 static const struct snmp_mib snmp6_icmp6_list[] = {
@@ -96,7 +95,6 @@ static const struct snmp_mib snmp6_icmp6_list[] = {
 	SNMP_MIB_ITEM("Icmp6OutErrors", ICMP6_MIB_OUTERRORS),
 	SNMP_MIB_ITEM("Icmp6InCsumErrors", ICMP6_MIB_CSUMERRORS),
 	SNMP_MIB_ITEM("Icmp6OutRateLimitHost", ICMP6_MIB_RATELIMITHOST),
-	SNMP_MIB_SENTINEL
 };
 
 static const struct snmp_mib snmp6_udp6_list[] = {
@@ -109,7 +107,6 @@ static const struct snmp_mib snmp6_udp6_list[] = {
 	SNMP_MIB_ITEM("Udp6InCsumErrors", UDP_MIB_CSUMERRORS),
 	SNMP_MIB_ITEM("Udp6IgnoredMulti", UDP_MIB_IGNOREDMULTI),
 	SNMP_MIB_ITEM("Udp6MemErrors", UDP_MIB_MEMERRORS),
-	SNMP_MIB_SENTINEL
 };
 
 static const struct snmp_mib snmp6_udplite6_list[] = {
@@ -121,7 +118,6 @@ static const struct snmp_mib snmp6_udplite6_list[] = {
 	SNMP_MIB_ITEM("UdpLite6SndbufErrors", UDP_MIB_SNDBUFERRORS),
 	SNMP_MIB_ITEM("UdpLite6InCsumErrors", UDP_MIB_CSUMERRORS),
 	SNMP_MIB_ITEM("UdpLite6MemErrors", UDP_MIB_MEMERRORS),
-	SNMP_MIB_SENTINEL
 };
 
 static void snmp6_seq_show_icmpv6msg(struct seq_file *seq, atomic_long_t *smib)
@@ -182,35 +178,37 @@ static void snmp6_seq_show_icmpv6msg(struct seq_file *seq, atomic_long_t *smib)
  */
 static void snmp6_seq_show_item(struct seq_file *seq, void __percpu *pcpumib,
 				atomic_long_t *smib,
-				const struct snmp_mib *itemlist)
+				const struct snmp_mib *itemlist,
+				int cnt)
 {
 	unsigned long buff[SNMP_MIB_MAX];
 	int i;
 
 	if (pcpumib) {
-		memset(buff, 0, sizeof(unsigned long) * SNMP_MIB_MAX);
+		memset(buff, 0, sizeof(unsigned long) * cnt);
 
-		snmp_get_cpu_field_batch(buff, itemlist, pcpumib);
-		for (i = 0; itemlist[i].name; i++)
+		snmp_get_cpu_field_batch_cnt(buff, itemlist, cnt, pcpumib);
+		for (i = 0; i < cnt; i++)
 			seq_printf(seq, "%-32s\t%lu\n",
 				   itemlist[i].name, buff[i]);
 	} else {
-		for (i = 0; itemlist[i].name; i++)
+		for (i = 0; i < cnt; i++)
 			seq_printf(seq, "%-32s\t%lu\n", itemlist[i].name,
 				   atomic_long_read(smib + itemlist[i].entry));
 	}
 }
 
 static void snmp6_seq_show_item64(struct seq_file *seq, void __percpu *mib,
-				  const struct snmp_mib *itemlist, size_t syncpoff)
+				  const struct snmp_mib *itemlist,
+				  int cnt, size_t syncpoff)
 {
 	u64 buff64[SNMP_MIB_MAX];
 	int i;
 
-	memset(buff64, 0, sizeof(u64) * SNMP_MIB_MAX);
+	memset(buff64, 0, sizeof(u64) * cnt);
 
-	snmp_get_cpu_field64_batch(buff64, itemlist, mib, syncpoff);
-	for (i = 0; itemlist[i].name; i++)
+	snmp_get_cpu_field64_batch_cnt(buff64, itemlist, cnt, mib, syncpoff);
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, "%-32s\t%llu\n", itemlist[i].name, buff64[i]);
 }
 
@@ -219,14 +217,19 @@ static int snmp6_seq_show(struct seq_file *seq, void *v)
 	struct net *net = (struct net *)seq->private;
 
 	snmp6_seq_show_item64(seq, net->mib.ipv6_statistics,
-			    snmp6_ipstats_list, offsetof(struct ipstats_mib, syncp));
+			      snmp6_ipstats_list,
+			      ARRAY_SIZE(snmp6_ipstats_list),
+			      offsetof(struct ipstats_mib, syncp));
 	snmp6_seq_show_item(seq, net->mib.icmpv6_statistics,
-			    NULL, snmp6_icmp6_list);
+			    NULL, snmp6_icmp6_list,
+			    ARRAY_SIZE(snmp6_icmp6_list));
 	snmp6_seq_show_icmpv6msg(seq, net->mib.icmpv6msg_statistics->mibs);
 	snmp6_seq_show_item(seq, net->mib.udp_stats_in6,
-			    NULL, snmp6_udp6_list);
+			    NULL, snmp6_udp6_list,
+			    ARRAY_SIZE(snmp6_udp6_list));
 	snmp6_seq_show_item(seq, net->mib.udplite_stats_in6,
-			    NULL, snmp6_udplite6_list);
+			    NULL, snmp6_udplite6_list,
+			    ARRAY_SIZE(snmp6_udplite6_list));
 	return 0;
 }
 
@@ -236,9 +239,11 @@ static int snmp6_dev_seq_show(struct seq_file *seq, void *v)
 
 	seq_printf(seq, "%-32s\t%u\n", "ifIndex", idev->dev->ifindex);
 	snmp6_seq_show_item64(seq, idev->stats.ipv6,
-			    snmp6_ipstats_list, offsetof(struct ipstats_mib, syncp));
+			      snmp6_ipstats_list,
+			      ARRAY_SIZE(snmp6_ipstats_list),
+			      offsetof(struct ipstats_mib, syncp));
 	snmp6_seq_show_item(seq, NULL, idev->stats.icmpv6dev->mibs,
-			    snmp6_icmp6_list);
+			    snmp6_icmp6_list, ARRAY_SIZE(snmp6_icmp6_list));
 	snmp6_seq_show_icmpv6msg(seq, idev->stats.icmpv6msgdev->mibs);
 	return 0;
 }
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 2/9] ipv6: snmp: do not use SNMP_MIB_SENTINEL anymore

[Sabrina Dubroca]

2025-09-05, 16:58:06 +0000, Eric Dumazet wrote:
> Use ARRAY_SIZE(), so that we know the limit at compile time.
> 
> Following patch needs this preliminary change.
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> ---
>  include/net/ip.h | 24 ++++++++++++++++++++++++
>  net/ipv6/proc.c  | 43 ++++++++++++++++++++++++-------------------
>  2 files changed, 48 insertions(+), 19 deletions(-)

Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>

-- 
Sabrina

[PATCH v2 net-next 3/9] ipv6: snmp: do not track per idev ICMP6_MIB_RATELIMITHOST

[Eric Dumazet]

Blamed commit added a critical false sharing on a single
atomic_long_t under DOS, like receiving UDP packets
to closed ports.

Per netns ICMP6_MIB_RATELIMITHOST tracking uses per-cpu
storage and is enough, we do not need per-device and slow tracking.

Fixes: d0941130c9351 ("icmp: Add counters for rate limits")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jamie Bainbridge <jamie.bainbridge@gmail.com>
Cc: Abhishek Rawal <rawal.abhishek92@gmail.com>
---
 net/ipv6/icmp.c | 3 +--
 net/ipv6/proc.c | 6 +++++-
 2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/net/ipv6/icmp.c b/net/ipv6/icmp.c
index 95cdd4cacb00..56c974cf75d1 100644
--- a/net/ipv6/icmp.c
+++ b/net/ipv6/icmp.c
@@ -230,8 +230,7 @@ static bool icmpv6_xrlim_allow(struct sock *sk, u8 type,
 	}
 	rcu_read_unlock();
 	if (!res)
-		__ICMP6_INC_STATS(net, ip6_dst_idev(dst),
-				  ICMP6_MIB_RATELIMITHOST);
+		__ICMP6_INC_STATS(net, NULL, ICMP6_MIB_RATELIMITHOST);
 	else
 		icmp_global_consume(net);
 	dst_release(dst);
diff --git a/net/ipv6/proc.c b/net/ipv6/proc.c
index 92ed04729c2f..73296f38c252 100644
--- a/net/ipv6/proc.c
+++ b/net/ipv6/proc.c
@@ -94,6 +94,7 @@ static const struct snmp_mib snmp6_icmp6_list[] = {
 	SNMP_MIB_ITEM("Icmp6OutMsgs", ICMP6_MIB_OUTMSGS),
 	SNMP_MIB_ITEM("Icmp6OutErrors", ICMP6_MIB_OUTERRORS),
 	SNMP_MIB_ITEM("Icmp6InCsumErrors", ICMP6_MIB_CSUMERRORS),
+/* ICMP6_MIB_RATELIMITHOST needs to be last, see snmp6_dev_seq_show(). */
 	SNMP_MIB_ITEM("Icmp6OutRateLimitHost", ICMP6_MIB_RATELIMITHOST),
 };
 
@@ -242,8 +243,11 @@ static int snmp6_dev_seq_show(struct seq_file *seq, void *v)
 			      snmp6_ipstats_list,
 			      ARRAY_SIZE(snmp6_ipstats_list),
 			      offsetof(struct ipstats_mib, syncp));
+
+	/* Per idev icmp stats do not have ICMP6_MIB_RATELIMITHOST */
 	snmp6_seq_show_item(seq, NULL, idev->stats.icmpv6dev->mibs,
-			    snmp6_icmp6_list, ARRAY_SIZE(snmp6_icmp6_list));
+			    snmp6_icmp6_list, ARRAY_SIZE(snmp6_icmp6_list) - 1);
+
 	snmp6_seq_show_icmpv6msg(seq, idev->stats.icmpv6msgdev->mibs);
 	return 0;
 }
-- 
2.51.0.355.g5224444f11-goog

[PATCH v2 net-next 4/9] ipv4: snmp: do not use SNMP_MIB_SENTINEL anymore

[Eric Dumazet]

Use ARRAY_SIZE(), so that we know the limit at compile time.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 net/ipv4/proc.c | 65 +++++++++++++++++++++++++------------------------
 1 file changed, 33 insertions(+), 32 deletions(-)

diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c
index 65b0d0ab0084..974afc4ecbe2 100644
--- a/net/ipv4/proc.c
+++ b/net/ipv4/proc.c
@@ -95,7 +95,6 @@ static const struct snmp_mib snmp4_ipstats_list[] = {
 	SNMP_MIB_ITEM("FragFails", IPSTATS_MIB_FRAGFAILS),
 	SNMP_MIB_ITEM("FragCreates", IPSTATS_MIB_FRAGCREATES),
 	SNMP_MIB_ITEM("OutTransmits", IPSTATS_MIB_OUTPKTS),
-	SNMP_MIB_SENTINEL
 };
 
 /* Following items are displayed in /proc/net/netstat */
@@ -119,7 +118,6 @@ static const struct snmp_mib snmp4_ipextstats_list[] = {
 	SNMP_MIB_ITEM("InECT0Pkts", IPSTATS_MIB_ECT0PKTS),
 	SNMP_MIB_ITEM("InCEPkts", IPSTATS_MIB_CEPKTS),
 	SNMP_MIB_ITEM("ReasmOverlaps", IPSTATS_MIB_REASM_OVERLAPS),
-	SNMP_MIB_SENTINEL
 };
 
 static const struct {
@@ -157,7 +155,6 @@ static const struct snmp_mib snmp4_tcp_list[] = {
 	SNMP_MIB_ITEM("InErrs", TCP_MIB_INERRS),
 	SNMP_MIB_ITEM("OutRsts", TCP_MIB_OUTRSTS),
 	SNMP_MIB_ITEM("InCsumErrors", TCP_MIB_CSUMERRORS),
-	SNMP_MIB_SENTINEL
 };
 
 static const struct snmp_mib snmp4_udp_list[] = {
@@ -170,7 +167,6 @@ static const struct snmp_mib snmp4_udp_list[] = {
 	SNMP_MIB_ITEM("InCsumErrors", UDP_MIB_CSUMERRORS),
 	SNMP_MIB_ITEM("IgnoredMulti", UDP_MIB_IGNOREDMULTI),
 	SNMP_MIB_ITEM("MemErrors", UDP_MIB_MEMERRORS),
-	SNMP_MIB_SENTINEL
 };
 
 static const struct snmp_mib snmp4_net_list[] = {
@@ -309,7 +305,6 @@ static const struct snmp_mib snmp4_net_list[] = {
 	SNMP_MIB_ITEM("TCPAOKeyNotFound", LINUX_MIB_TCPAOKEYNOTFOUND),
 	SNMP_MIB_ITEM("TCPAOGood", LINUX_MIB_TCPAOGOOD),
 	SNMP_MIB_ITEM("TCPAODroppedIcmps", LINUX_MIB_TCPAODROPPEDICMPS),
-	SNMP_MIB_SENTINEL
 };
 
 static void icmpmsg_put_line(struct seq_file *seq, unsigned long *vals,
@@ -389,14 +384,15 @@ static void icmp_put(struct seq_file *seq)
  */
 static int snmp_seq_show_ipstats(struct seq_file *seq, void *v)
 {
+	const int cnt = ARRAY_SIZE(snmp4_ipstats_list);
+	u64 buff64[ARRAY_SIZE(snmp4_ipstats_list)];
 	struct net *net = seq->private;
-	u64 buff64[IPSTATS_MIB_MAX];
 	int i;
 
-	memset(buff64, 0, IPSTATS_MIB_MAX * sizeof(u64));
+	memset(buff64, 0, sizeof(buff64));
 
 	seq_puts(seq, "Ip: Forwarding DefaultTTL");
-	for (i = 0; snmp4_ipstats_list[i].name; i++)
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, " %s", snmp4_ipstats_list[i].name);
 
 	seq_printf(seq, "\nIp: %d %d",
@@ -404,10 +400,10 @@ static int snmp_seq_show_ipstats(struct seq_file *seq, void *v)
 		   READ_ONCE(net->ipv4.sysctl_ip_default_ttl));
 
 	BUILD_BUG_ON(offsetof(struct ipstats_mib, mibs) != 0);
-	snmp_get_cpu_field64_batch(buff64, snmp4_ipstats_list,
-				   net->mib.ip_statistics,
-				   offsetof(struct ipstats_mib, syncp));
-	for (i = 0; snmp4_ipstats_list[i].name; i++)
+	snmp_get_cpu_field64_batch_cnt(buff64, snmp4_ipstats_list, cnt,
+				       net->mib.ip_statistics,
+				       offsetof(struct ipstats_mib, syncp));
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, " %llu", buff64[i]);
 
 	return 0;
@@ -415,20 +411,23 @@ static int snmp_seq_show_ipstats(struct seq_file *seq, void *v)
 
 static int snmp_seq_show_tcp_udp(struct seq_file *seq, void *v)
 {
+	const int udp_cnt = ARRAY_SIZE(snmp4_udp_list);
+	const int tcp_cnt = ARRAY_SIZE(snmp4_tcp_list);
 	unsigned long buff[TCPUDP_MIB_MAX];
 	struct net *net = seq->private;
 	int i;
 
-	memset(buff, 0, TCPUDP_MIB_MAX * sizeof(unsigned long));
+	memset(buff, 0, tcp_cnt * sizeof(unsigned long));
 
 	seq_puts(seq, "\nTcp:");
-	for (i = 0; snmp4_tcp_list[i].name; i++)
+	for (i = 0; i < tcp_cnt; i++)
 		seq_printf(seq, " %s", snmp4_tcp_list[i].name);
 
 	seq_puts(seq, "\nTcp:");
-	snmp_get_cpu_field_batch(buff, snmp4_tcp_list,
-				 net->mib.tcp_statistics);
-	for (i = 0; snmp4_tcp_list[i].name; i++) {
+	snmp_get_cpu_field_batch_cnt(buff, snmp4_tcp_list,
+				     tcp_cnt,
+				     net->mib.tcp_statistics);
+	for (i = 0; i < tcp_cnt; i++) {
 		/* MaxConn field is signed, RFC 2012 */
 		if (snmp4_tcp_list[i].entry == TCP_MIB_MAXCONN)
 			seq_printf(seq, " %ld", buff[i]);
@@ -436,27 +435,29 @@ static int snmp_seq_show_tcp_udp(struct seq_file *seq, void *v)
 			seq_printf(seq, " %lu", buff[i]);
 	}
 
-	memset(buff, 0, TCPUDP_MIB_MAX * sizeof(unsigned long));
+	memset(buff, 0, udp_cnt * sizeof(unsigned long));
 
-	snmp_get_cpu_field_batch(buff, snmp4_udp_list,
-				 net->mib.udp_statistics);
+	snmp_get_cpu_field_batch_cnt(buff, snmp4_udp_list,
+				     udp_cnt,
+				     net->mib.udp_statistics);
 	seq_puts(seq, "\nUdp:");
-	for (i = 0; snmp4_udp_list[i].name; i++)
+	for (i = 0; i < udp_cnt; i++)
 		seq_printf(seq, " %s", snmp4_udp_list[i].name);
 	seq_puts(seq, "\nUdp:");
-	for (i = 0; snmp4_udp_list[i].name; i++)
+	for (i = 0; i < udp_cnt; i++)
 		seq_printf(seq, " %lu", buff[i]);
 
-	memset(buff, 0, TCPUDP_MIB_MAX * sizeof(unsigned long));
+	memset(buff, 0, udp_cnt * sizeof(unsigned long));
 
 	/* the UDP and UDP-Lite MIBs are the same */
 	seq_puts(seq, "\nUdpLite:");
-	snmp_get_cpu_field_batch(buff, snmp4_udp_list,
-				 net->mib.udplite_statistics);
-	for (i = 0; snmp4_udp_list[i].name; i++)
+	snmp_get_cpu_field_batch_cnt(buff, snmp4_udp_list,
+				     udp_cnt,
+				     net->mib.udplite_statistics);
+	for (i = 0; i < udp_cnt; i++)
 		seq_printf(seq, " %s", snmp4_udp_list[i].name);
 	seq_puts(seq, "\nUdpLite:");
-	for (i = 0; snmp4_udp_list[i].name; i++)
+	for (i = 0; i < udp_cnt; i++)
 		seq_printf(seq, " %lu", buff[i]);
 
 	seq_putc(seq, '\n');
@@ -480,8 +481,8 @@ static int snmp_seq_show(struct seq_file *seq, void *v)
  */
 static int netstat_seq_show(struct seq_file *seq, void *v)
 {
-	const int ip_cnt = ARRAY_SIZE(snmp4_ipextstats_list) - 1;
-	const int tcp_cnt = ARRAY_SIZE(snmp4_net_list) - 1;
+	const int ip_cnt = ARRAY_SIZE(snmp4_ipextstats_list);
+	const int tcp_cnt = ARRAY_SIZE(snmp4_net_list);
 	struct net *net = seq->private;
 	unsigned long *buff;
 	int i;
@@ -494,8 +495,8 @@ static int netstat_seq_show(struct seq_file *seq, void *v)
 	buff = kzalloc(max(tcp_cnt * sizeof(long), ip_cnt * sizeof(u64)),
 		       GFP_KERNEL);
 	if (buff) {
-		snmp_get_cpu_field_batch(buff, snmp4_net_list,
-					 net->mib.net_statistics);
+		snmp_get_cpu_field_batch_cnt(buff, snmp4_net_list, tcp_cnt,
+					     net->mib.net_statistics);
 		for (i = 0; i < tcp_cnt; i++)
 			seq_printf(seq, " %lu", buff[i]);
 	} else {
@@ -513,7 +514,7 @@ static int netstat_seq_show(struct seq_file *seq, void *v)
 		u64 *buff64 = (u64 *)buff;
 
 		memset(buff64, 0, ip_cnt * sizeof(u64));
-		snmp_get_cpu_field64_batch(buff64, snmp4_ipextstats_list,
+		snmp_get_cpu_field64_batch_cnt(buff64, snmp4_ipextstats_list, ip_cnt,
 					   net->mib.ip_statistics,
 					   offsetof(struct ipstats_mib, syncp));
 		for (i = 0; i < ip_cnt; i++)
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 4/9] ipv4: snmp: do not use SNMP_MIB_SENTINEL anymore

[Sabrina Dubroca]

2025-09-05, 16:58:08 +0000, Eric Dumazet wrote:
> Use ARRAY_SIZE(), so that we know the limit at compile time.
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> ---
>  net/ipv4/proc.c | 65 +++++++++++++++++++++++++------------------------
>  1 file changed, 33 insertions(+), 32 deletions(-)

Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>

-- 
Sabrina

[PATCH v2 net-next 5/9] mptcp: snmp: do not use SNMP_MIB_SENTINEL anymore

[Eric Dumazet]

Use ARRAY_SIZE(), so that we know the limit at compile time.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Matthieu Baerts <matttbe@kernel.org>
Cc: Mat Martineau <martineau@kernel.org>
Cc: Geliang Tang <geliang@kernel.org>
---
 net/mptcp/mib.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/net/mptcp/mib.c b/net/mptcp/mib.c
index cf879c188ca2..6003e47c770a 100644
--- a/net/mptcp/mib.c
+++ b/net/mptcp/mib.c
@@ -85,7 +85,6 @@ static const struct snmp_mib mptcp_snmp_list[] = {
 	SNMP_MIB_ITEM("DssFallback", MPTCP_MIB_DSSFALLBACK),
 	SNMP_MIB_ITEM("SimultConnectFallback", MPTCP_MIB_SIMULTCONNFALLBACK),
 	SNMP_MIB_ITEM("FallbackFailed", MPTCP_MIB_FALLBACKFAILED),
-	SNMP_MIB_SENTINEL
 };
 
 /* mptcp_mib_alloc - allocate percpu mib counters
@@ -108,22 +107,23 @@ bool mptcp_mib_alloc(struct net *net)
 
 void mptcp_seq_show(struct seq_file *seq)
 {
-	unsigned long sum[ARRAY_SIZE(mptcp_snmp_list) - 1];
+	unsigned long sum[ARRAY_SIZE(mptcp_snmp_list)];
+	const int cnt = ARRAY_SIZE(mptcp_snmp_list);
 	struct net *net = seq->private;
 	int i;
 
 	seq_puts(seq, "MPTcpExt:");
-	for (i = 0; mptcp_snmp_list[i].name; i++)
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, " %s", mptcp_snmp_list[i].name);
 
 	seq_puts(seq, "\nMPTcpExt:");
 
 	memset(sum, 0, sizeof(sum));
 	if (net->mib.mptcp_statistics)
-		snmp_get_cpu_field_batch(sum, mptcp_snmp_list,
-					 net->mib.mptcp_statistics);
+		snmp_get_cpu_field_batch_cnt(sum, mptcp_snmp_list, cnt,
+					     net->mib.mptcp_statistics);
 
-	for (i = 0; mptcp_snmp_list[i].name; i++)
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, " %lu", sum[i]);
 
 	seq_putc(seq, '\n');
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 5/9] mptcp: snmp: do not use SNMP_MIB_SENTINEL anymore

[Matthieu Baerts]

Hi Eric,

On 05/09/2025 18:58, Eric Dumazet wrote:
> Use ARRAY_SIZE(), so that we know the limit at compile time.

Thank you for having done this modification in MPTCP as well!

Reviewed-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>

Cheers,
Matt
-- 
Sponsored by the NGI0 Core fund.

[PATCH v2 net-next 6/9] sctp: snmp: do not use SNMP_MIB_SENTINEL anymore

[Eric Dumazet]

Use ARRAY_SIZE(), so that we know the limit at compile time.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Cc: Xin Long <lucien.xin@gmail.com>
---
 net/sctp/proc.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/net/sctp/proc.c b/net/sctp/proc.c
index 74bff317e205..1ed281f3c355 100644
--- a/net/sctp/proc.c
+++ b/net/sctp/proc.c
@@ -52,21 +52,21 @@ static const struct snmp_mib sctp_snmp_list[] = {
 	SNMP_MIB_ITEM("SctpInPktBacklog", SCTP_MIB_IN_PKT_BACKLOG),
 	SNMP_MIB_ITEM("SctpInPktDiscards", SCTP_MIB_IN_PKT_DISCARDS),
 	SNMP_MIB_ITEM("SctpInDataChunkDiscards", SCTP_MIB_IN_DATA_CHUNK_DISCARDS),
-	SNMP_MIB_SENTINEL
 };
 
 /* Display sctp snmp mib statistics(/proc/net/sctp/snmp). */
 static int sctp_snmp_seq_show(struct seq_file *seq, void *v)
 {
-	unsigned long buff[SCTP_MIB_MAX];
+	unsigned long buff[ARRAY_SIZE(sctp_snmp_list)];
+	const int cnt = ARRAY_SIZE(sctp_snmp_list);
 	struct net *net = seq->private;
 	int i;
 
-	memset(buff, 0, sizeof(unsigned long) * SCTP_MIB_MAX);
+	memset(buff, 0, sizeof(buff));
 
-	snmp_get_cpu_field_batch(buff, sctp_snmp_list,
-				 net->sctp.sctp_statistics);
-	for (i = 0; sctp_snmp_list[i].name; i++)
+	snmp_get_cpu_field_batch_cnt(buff, sctp_snmp_list, cnt,
+				     net->sctp.sctp_statistics);
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, "%-32s\t%ld\n", sctp_snmp_list[i].name,
 						buff[i]);
 
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 6/9] sctp: snmp: do not use SNMP_MIB_SENTINEL anymore

[Sabrina Dubroca]

2025-09-05, 16:58:10 +0000, Eric Dumazet wrote:
> Use ARRAY_SIZE(), so that we know the limit at compile time.
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> Cc: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> Cc: Xin Long <lucien.xin@gmail.com>
> ---
>  net/sctp/proc.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)

Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>

-- 
Sabrina

Re: [PATCH v2 net-next 6/9] sctp: snmp: do not use SNMP_MIB_SENTINEL anymore

[Xin Long]

On Fri, Sep 5, 2025 at 12:58 PM Eric Dumazet <edumazet@google.com> wrote:
>
> Use ARRAY_SIZE(), so that we know the limit at compile time.
>
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> Cc: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> Cc: Xin Long <lucien.xin@gmail.com>
Acked-by: Xin Long <lucien.xin@gmail.com>

[PATCH v2 net-next 7/9] tls: snmp: do not use SNMP_MIB_SENTINEL anymore

[Eric Dumazet]

Use ARRAY_SIZE(), so that we know the limit at compile time.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Sabrina Dubroca <sd@queasysnail.net>
---
 net/tls/tls_proc.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/net/tls/tls_proc.c b/net/tls/tls_proc.c
index 367666aa07b8..4012c4372d4c 100644
--- a/net/tls/tls_proc.c
+++ b/net/tls/tls_proc.c
@@ -27,17 +27,19 @@ static const struct snmp_mib tls_mib_list[] = {
 	SNMP_MIB_ITEM("TlsTxRekeyOk", LINUX_MIB_TLSTXREKEYOK),
 	SNMP_MIB_ITEM("TlsTxRekeyError", LINUX_MIB_TLSTXREKEYERROR),
 	SNMP_MIB_ITEM("TlsRxRekeyReceived", LINUX_MIB_TLSRXREKEYRECEIVED),
-	SNMP_MIB_SENTINEL
 };
 
 static int tls_statistics_seq_show(struct seq_file *seq, void *v)
 {
-	unsigned long buf[LINUX_MIB_TLSMAX] = {};
+	unsigned long buf[ARRAY_SIZE(tls_mib_list)];
+	const int cnt = ARRAY_SIZE(tls_mib_list);
 	struct net *net = seq->private;
 	int i;
 
-	snmp_get_cpu_field_batch(buf, tls_mib_list, net->mib.tls_statistics);
-	for (i = 0; tls_mib_list[i].name; i++)
+	memset(buf, 0, sizeof(buf));
+	snmp_get_cpu_field_batch_cnt(buf, tls_mib_list, cnt,
+				     net->mib.tls_statistics);
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, "%-32s\t%lu\n", tls_mib_list[i].name, buf[i]);
 
 	return 0;
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 7/9] tls: snmp: do not use SNMP_MIB_SENTINEL anymore

[Sabrina Dubroca]

2025-09-05, 16:58:11 +0000, Eric Dumazet wrote:
> Use ARRAY_SIZE(), so that we know the limit at compile time.
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> Cc: John Fastabend <john.fastabend@gmail.com>
> Cc: Sabrina Dubroca <sd@queasysnail.net>
> ---
>  net/tls/tls_proc.c | 10 ++++++----
>  1 file changed, 6 insertions(+), 4 deletions(-)

Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>

-- 
Sabrina

[PATCH v2 net-next 8/9] xfrm: snmp: do not use SNMP_MIB_SENTINEL anymore

[Eric Dumazet]

Use ARRAY_SIZE(), so that we know the limit at compile time.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Steffen Klassert <steffen.klassert@secunet.com>
---
 net/xfrm/xfrm_proc.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/net/xfrm/xfrm_proc.c b/net/xfrm/xfrm_proc.c
index 8e07dd614b0b..5e1fd6b1d503 100644
--- a/net/xfrm/xfrm_proc.c
+++ b/net/xfrm/xfrm_proc.c
@@ -45,21 +45,21 @@ static const struct snmp_mib xfrm_mib_list[] = {
 	SNMP_MIB_ITEM("XfrmInStateDirError", LINUX_MIB_XFRMINSTATEDIRERROR),
 	SNMP_MIB_ITEM("XfrmInIptfsError", LINUX_MIB_XFRMINIPTFSERROR),
 	SNMP_MIB_ITEM("XfrmOutNoQueueSpace", LINUX_MIB_XFRMOUTNOQSPACE),
-	SNMP_MIB_SENTINEL
 };
 
 static int xfrm_statistics_seq_show(struct seq_file *seq, void *v)
 {
-	unsigned long buff[LINUX_MIB_XFRMMAX];
+	unsigned long buff[ARRAY_SIZE(xfrm_mib_list)];
+	const int cnt = ARRAY_SIZE(xfrm_mib_list);
 	struct net *net = seq->private;
 	int i;
 
-	memset(buff, 0, sizeof(unsigned long) * LINUX_MIB_XFRMMAX);
+	memset(buff, 0, sizeof(buff));
 
 	xfrm_state_update_stats(net);
-	snmp_get_cpu_field_batch(buff, xfrm_mib_list,
-				 net->mib.xfrm_statistics);
-	for (i = 0; xfrm_mib_list[i].name; i++)
+	snmp_get_cpu_field_batch_cnt(buff, xfrm_mib_list, cnt,
+				     net->mib.xfrm_statistics);
+	for (i = 0; i < cnt; i++)
 		seq_printf(seq, "%-24s\t%lu\n", xfrm_mib_list[i].name,
 						buff[i]);
 
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 8/9] xfrm: snmp: do not use SNMP_MIB_SENTINEL anymore

[Sabrina Dubroca]

2025-09-05, 16:58:12 +0000, Eric Dumazet wrote:
> Use ARRAY_SIZE(), so that we know the limit at compile time.
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> Cc: Steffen Klassert <steffen.klassert@secunet.com>
> ---
>  net/xfrm/xfrm_proc.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)

Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>

-- 
Sabrina

[PATCH v2 net-next 9/9] net: snmp: remove SNMP_MIB_SENTINEL

[Eric Dumazet]

No more user of SNMP_MIB_SENTINEL, we can remove it.

Also remove snmp_get_cpu_field[64]_batch() helpers.

Signed-off-by: Eric Dumazet <edumazet@google.com>
---
 include/net/ip.h   | 23 -----------------------
 include/net/snmp.h |  5 -----
 2 files changed, 28 deletions(-)

diff --git a/include/net/ip.h b/include/net/ip.h
index a1624e8db1ab..380afb691c41 100644
--- a/include/net/ip.h
+++ b/include/net/ip.h
@@ -326,18 +326,6 @@ static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_o
 }
 #endif
 
-#define snmp_get_cpu_field64_batch(buff64, stats_list, mib_statistic, offset) \
-{ \
-	int i, c; \
-	for_each_possible_cpu(c) { \
-		for (i = 0; stats_list[i].name; i++) \
-			buff64[i] += snmp_get_cpu_field64( \
-					mib_statistic, \
-					c, stats_list[i].entry, \
-					offset); \
-	} \
-}
-
 #define snmp_get_cpu_field64_batch_cnt(buff64, stats_list, cnt,	\
 				       mib_statistic, offset)	\
 { \
@@ -351,17 +339,6 @@ static inline u64 snmp_fold_field64(void __percpu *mib, int offt, size_t syncp_o
 	} \
 }
 
-#define snmp_get_cpu_field_batch(buff, stats_list, mib_statistic) \
-{ \
-	int i, c; \
-	for_each_possible_cpu(c) { \
-		for (i = 0; stats_list[i].name; i++) \
-			buff[i] += snmp_get_cpu_field( \
-						mib_statistic, \
-						c, stats_list[i].entry); \
-	} \
-}
-
 #define snmp_get_cpu_field_batch_cnt(buff, stats_list, cnt, mib_statistic) \
 { \
 	int i, c; \
diff --git a/include/net/snmp.h b/include/net/snmp.h
index 4cb4326dfebe..584e70742e9b 100644
--- a/include/net/snmp.h
+++ b/include/net/snmp.h
@@ -36,11 +36,6 @@ struct snmp_mib {
 	.entry = _entry,			\
 }
 
-#define SNMP_MIB_SENTINEL {	\
-	.name = NULL,		\
-	.entry = 0,		\
-}
-
 /*
  * We use unsigned longs for most mibs but u64 for ipstats.
  */
-- 
2.51.0.355.g5224444f11-goog

Re: [PATCH v2 net-next 9/9] net: snmp: remove SNMP_MIB_SENTINEL

[Sabrina Dubroca]

2025-09-05, 16:58:13 +0000, Eric Dumazet wrote:
> No more user of SNMP_MIB_SENTINEL, we can remove it.
> 
> Also remove snmp_get_cpu_field[64]_batch() helpers.
> 
> Signed-off-by: Eric Dumazet <edumazet@google.com>
> ---
>  include/net/ip.h   | 23 -----------------------
>  include/net/snmp.h |  5 -----
>  2 files changed, 28 deletions(-)

Reviewed-by: Sabrina Dubroca <sd@queasysnail.net>

-- 
Sabrina

Re: [PATCH v2 net-next 0/9] ipv6: snmp: avoid performance issue with RATELIMITHOST

[patchwork-bot+netdevbpf]

Hello:

This series was applied to netdev/net-next.git (main)
by Jakub Kicinski <kuba@kernel.org>:

On Fri,  5 Sep 2025 16:58:04 +0000 you wrote:
> Addition of ICMP6_MIB_RATELIMITHOST in commit d0941130c9351
> ("icmp: Add counters for rate limits") introduced a performance
> drop in case of DOS (like receiving UDP packets
> to closed ports).
> 
> Per netns ICMP6_MIB_RATELIMITHOST tracking uses per-cpu
> storage and is enough, we do not need per-device and slow tracking
> for this metric.
> 
> [...]

Here is the summary with links:
  - [v2,net-next,1/9] ipv6: snmp: remove icmp6type2name[]
    https://git.kernel.org/netdev/net-next/c/b7fe8c1be776
  - [v2,net-next,2/9] ipv6: snmp: do not use SNMP_MIB_SENTINEL anymore
    https://git.kernel.org/netdev/net-next/c/ceac1fb2290d
  - [v2,net-next,3/9] ipv6: snmp: do not track per idev ICMP6_MIB_RATELIMITHOST
    https://git.kernel.org/netdev/net-next/c/2fab94bcf313
  - [v2,net-next,4/9] ipv4: snmp: do not use SNMP_MIB_SENTINEL anymore
    https://git.kernel.org/netdev/net-next/c/b7b74953f834
  - [v2,net-next,5/9] mptcp: snmp: do not use SNMP_MIB_SENTINEL anymore
    https://git.kernel.org/netdev/net-next/c/35cb2da0abaf
  - [v2,net-next,6/9] sctp: snmp: do not use SNMP_MIB_SENTINEL anymore
    https://git.kernel.org/netdev/net-next/c/52a33cae6a6f
  - [v2,net-next,7/9] tls: snmp: do not use SNMP_MIB_SENTINEL anymore
    https://git.kernel.org/netdev/net-next/c/3a951f95202c
  - [v2,net-next,8/9] xfrm: snmp: do not use SNMP_MIB_SENTINEL anymore
    https://git.kernel.org/netdev/net-next/c/c73d583e7008
  - [v2,net-next,9/9] net: snmp: remove SNMP_MIB_SENTINEL
    https://git.kernel.org/netdev/net-next/c/20d3d2681544

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html