From: Pablo Neira Ayuso <pablo@netfilter.org>
To: netfilter-devel@vger.kernel.org
Cc: davem@davemloft.net, netdev@vger.kernel.org
Subject: [PATCH 26/29] netfilter: conntrack: resched gc again if eviction rate is high
Date: Mon, 5 Sep 2016 12:58:41 +0200 [thread overview]
Message-ID: <1473073124-5015-27-git-send-email-pablo@netfilter.org> (raw)
In-Reply-To: <1473073124-5015-1-git-send-email-pablo@netfilter.org>
From: Florian Westphal <fw@strlen.de>
If we evicted a large fraction of the scanned conntrack entries re-schedule
the next gc cycle for immediate execution.
This triggers during tests where load is high, then drops to zero and
many connections will be in TW/CLOSE state with < 30 second timeouts.
Without this change it will take several minutes until conntrack count
comes back to normal.
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
---
net/netfilter/nf_conntrack_core.c | 6 ++++++
1 file changed, 6 insertions(+)
diff --git a/net/netfilter/nf_conntrack_core.c b/net/netfilter/nf_conntrack_core.c
index f95a9e9..7c66ce40 100644
--- a/net/netfilter/nf_conntrack_core.c
+++ b/net/netfilter/nf_conntrack_core.c
@@ -945,6 +945,7 @@ static void gc_worker(struct work_struct *work)
{
unsigned int i, goal, buckets = 0, expired_count = 0;
unsigned long next_run = GC_INTERVAL;
+ unsigned int ratio, scanned = 0;
struct conntrack_gc_work *gc_work;
gc_work = container_of(work, struct conntrack_gc_work, dwork.work);
@@ -969,6 +970,7 @@ static void gc_worker(struct work_struct *work)
hlist_nulls_for_each_entry_rcu(h, n, &ct_hash[i], hnnode) {
tmp = nf_ct_tuplehash_to_ctrack(h);
+ scanned++;
if (nf_ct_is_expired(tmp)) {
nf_ct_gc_expired(tmp);
expired_count++;
@@ -988,6 +990,10 @@ static void gc_worker(struct work_struct *work)
if (gc_work->exiting)
return;
+ ratio = scanned ? expired_count * 100 / scanned : 0;
+ if (ratio >= 90)
+ next_run = 0;
+
gc_work->last_bucket = i;
schedule_delayed_work(&gc_work->dwork, next_run);
}
--
2.1.4
next prev parent reply other threads:[~2016-09-05 10:59 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-05 10:58 [PATCH 00/29] Netfilter updates for net-next Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 01/29] netfilter: conntrack: Only need first 4 bytes to get l4proto ports Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 02/29] netfilter: physdev: add missed blank Pablo Neira Ayuso
2016-09-05 17:43 ` Joe Perches
2016-09-05 10:58 ` [PATCH 03/29] netfilter: nf_dup4: remove redundant checksum recalculation Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 04/29] netfilter: use_nf_conn_expires helper in more places Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 05/29] ipvs: use nf_ct_kill helper Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 06/29] netfilter: nf_tables: rename set implementations Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 07/29] netfilter: nf_tables: add hash expression Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 08/29] netfilter: remove ip_conntrack* sysctl compat code Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 09/29] netfilter: conntrack: simplify the code by using nf_conntrack_get_ht Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 10/29] netfilter: nf_conntrack: restore nf_conntrack_htable_size as exported symbol Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 11/29] netfilter: nf_tables: add quota expression Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 12/29] netfilter: nf_tables: add number generator expression Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 13/29] netfilter: fix spelling mistake: "delimitter" -> "delimiter" Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 14/29] netfilter: nft_hash: fix non static symbol warning Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 15/29] netfilter: nf_tables: typo in trace attribute definition Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 16/29] netfilter: nf_tables: introduce nft_chain_parse_hook() Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 17/29] netfilter: nf_tables: reject hook configuration updates on existing chains Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 18/29] rhashtable: add rhashtable_lookup_get_insert_key() Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 19/29] netfilter: nf_tables: honor NLM_F_EXCL flag in set element insertion Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 20/29] netfilter: nf_tables: Use nla_put_be32() to dump immediate parameters Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 21/29] netfilter: restart search if moved to other chain Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 22/29] netfilter: don't rely on DYING bit to detect when destroy event was sent Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 23/29] netfilter: conntrack: get rid of conntrack timer Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 24/29] netfilter: evict stale entries on netlink dumps Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 25/29] netfilter: conntrack: add gc worker to remove timed-out entries Pablo Neira Ayuso
2016-09-05 10:58 ` Pablo Neira Ayuso [this message]
2016-09-05 10:58 ` [PATCH 27/29] netfilter: remove __nf_ct_kill_acct helper Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 28/29] netfilter: log_arp: Use ARPHRD_ETHER instead of literal '1' Pablo Neira Ayuso
2016-09-05 10:58 ` [PATCH 29/29] netfilter: log: Check param to avoid overflow in nf_log_set Pablo Neira Ayuso
2016-09-06 19:47 ` [PATCH 00/29] Netfilter updates for net-next David Miller
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=1473073124-5015-27-git-send-email-pablo@netfilter.org \
--to=pablo@netfilter.org \
--cc=davem@davemloft.net \
--cc=netdev@vger.kernel.org \
--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).