From: Florian Westphal <fw@strlen.de>
To: <netfilter-devel@vger.kernel.org>
Cc: Florian Westphal <fw@strlen.de>
Subject: [PATCH nf-next 3/4] netfilter: conntrack: add clash resolution stat counter
Date: Wed, 26 Aug 2020 00:52:44 +0200 [thread overview]
Message-ID: <20200825225245.8072-4-fw@strlen.de> (raw)
In-Reply-To: <20200825225245.8072-1-fw@strlen.de>
There is a misconception about what "insert_failed" means.
We increment this even when a clash got resolved, so it might not indicate
a problem.
Add a dedicated counter for clash resolution and only increment
insert_failed if a clash cannot be resolved.
For the old /proc interface, export this in place of an older stat
that got removed a while back.
For ctnetlink, export this with a new attribute.
Also correct an outdated comment that implies we add a duplicate tuple --
we only add the (unique) reply direction.
Signed-off-by: Florian Westphal <fw@strlen.de>
---
include/linux/netfilter/nf_conntrack_common.h | 1 +
include/uapi/linux/netfilter/nfnetlink_conntrack.h | 1 +
net/netfilter/nf_conntrack_core.c | 9 +++++----
net/netfilter/nf_conntrack_netlink.c | 4 +++-
net/netfilter/nf_conntrack_standalone.c | 2 +-
5 files changed, 11 insertions(+), 6 deletions(-)
diff --git a/include/linux/netfilter/nf_conntrack_common.h b/include/linux/netfilter/nf_conntrack_common.h
index 96b90d7e361f..0c7d8d1e945d 100644
--- a/include/linux/netfilter/nf_conntrack_common.h
+++ b/include/linux/netfilter/nf_conntrack_common.h
@@ -10,6 +10,7 @@ struct ip_conntrack_stat {
unsigned int invalid;
unsigned int insert;
unsigned int insert_failed;
+ unsigned int clash_resolve;
unsigned int drop;
unsigned int early_drop;
unsigned int error;
diff --git a/include/uapi/linux/netfilter/nfnetlink_conntrack.h b/include/uapi/linux/netfilter/nfnetlink_conntrack.h
index 3e471558da82..d8484be72fdc 100644
--- a/include/uapi/linux/netfilter/nfnetlink_conntrack.h
+++ b/include/uapi/linux/netfilter/nfnetlink_conntrack.h
@@ -256,6 +256,7 @@ enum ctattr_stats_cpu {
CTA_STATS_EARLY_DROP,
CTA_STATS_ERROR,
CTA_STATS_SEARCH_RESTART,
+ CTA_STATS_CLASH_RESOLVE,
__CTA_STATS_MAX,
};
#define CTA_STATS_MAX (__CTA_STATS_MAX - 1)
diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
index a111bcf1b93c..93e77ca0efad 100644
--- a/net/netfilter/nf_conntrack_core.c
+++ b/net/netfilter/nf_conntrack_core.c
@@ -859,7 +859,6 @@ nf_conntrack_hash_check_insert(struct nf_conn *ct)
out:
nf_conntrack_double_unlock(hash, reply_hash);
- NF_CT_STAT_INC(net, insert_failed);
local_bh_enable();
return -EEXIST;
}
@@ -934,7 +933,7 @@ static int __nf_ct_resolve_clash(struct sk_buff *skb,
nf_conntrack_put(&loser_ct->ct_general);
nf_ct_set(skb, ct, ctinfo);
- NF_CT_STAT_INC(net, insert_failed);
+ NF_CT_STAT_INC(net, clash_resolve);
return NF_ACCEPT;
}
@@ -998,6 +997,8 @@ static int nf_ct_resolve_clash_harder(struct sk_buff *skb, u32 repl_idx)
hlist_nulls_add_head_rcu(&loser_ct->tuplehash[IP_CT_DIR_REPLY].hnnode,
&nf_conntrack_hash[repl_idx]);
+
+ NF_CT_STAT_INC(net, clash_resolve);
return NF_ACCEPT;
}
@@ -1027,10 +1028,10 @@ static int nf_ct_resolve_clash_harder(struct sk_buff *skb, u32 repl_idx)
*
* Failing that, the new, unconfirmed conntrack is still added to the table
* provided that the collision only occurs in the ORIGINAL direction.
- * The new entry will be added after the existing one in the hash list,
+ * The new entry will be added only in the non-clashing REPLY direction,
* so packets in the ORIGINAL direction will continue to match the existing
* entry. The new entry will also have a fixed timeout so it expires --
- * due to the collision, it will not see bidirectional traffic.
+ * due to the collision, it will only see reply traffic.
*
* Returns NF_DROP if the clash could not be resolved.
*/
diff --git a/net/netfilter/nf_conntrack_netlink.c b/net/netfilter/nf_conntrack_netlink.c
index c64f23a8f373..89d99f6dfd0a 100644
--- a/net/netfilter/nf_conntrack_netlink.c
+++ b/net/netfilter/nf_conntrack_netlink.c
@@ -2516,7 +2516,9 @@ ctnetlink_ct_stat_cpu_fill_info(struct sk_buff *skb, u32 portid, u32 seq,
nla_put_be32(skb, CTA_STATS_EARLY_DROP, htonl(st->early_drop)) ||
nla_put_be32(skb, CTA_STATS_ERROR, htonl(st->error)) ||
nla_put_be32(skb, CTA_STATS_SEARCH_RESTART,
- htonl(st->search_restart)))
+ htonl(st->search_restart)) ||
+ nla_put_be32(skb, CTA_STATS_CLASH_RESOLVE,
+ htonl(st->clash_resolve)))
goto nla_put_failure;
nlmsg_end(skb, nlh);
diff --git a/net/netfilter/nf_conntrack_standalone.c b/net/netfilter/nf_conntrack_standalone.c
index b673a03624d2..0ff39740797d 100644
--- a/net/netfilter/nf_conntrack_standalone.c
+++ b/net/netfilter/nf_conntrack_standalone.c
@@ -435,7 +435,7 @@ static int ct_cpu_seq_show(struct seq_file *seq, void *v)
seq_printf(seq, "%08x %08x %08x %08x %08x %08x %08x %08x "
"%08x %08x %08x %08x %08x %08x %08x %08x %08x\n",
nr_conntracks,
- 0,
+ st->clash_resolve, /* was: searched */
st->found,
0,
st->invalid,
--
2.26.2
next prev parent reply other threads:[~2020-08-25 22:53 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-08-25 22:52 [PATCH nf-next 0/4] netfilter: revisit conntrack statistics Florian Westphal
2020-08-25 22:52 ` [PATCH nf-next 1/4] netfilter: conntrack: do not increment two error counters at same time Florian Westphal
2020-08-25 22:52 ` [PATCH nf-next 2/4] netfilter: conntrack: remove ignore stats Florian Westphal
2020-08-25 22:52 ` Florian Westphal [this message]
2020-08-25 22:52 ` [PATCH nf-next 4/4] netfilter: conntrack: remove unneeded nf_ct_put Florian Westphal
2020-08-28 17:52 ` [PATCH nf-next 0/4] netfilter: revisit conntrack statistics Pablo Neira Ayuso
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200825225245.8072-4-fw@strlen.de \
--to=fw@strlen.de \
--cc=netfilter-devel@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).