Netdev List
 help / color / mirror / Atom feed
From: Steffen Klassert <steffen.klassert@secunet.com>
To: David Miller <davem@davemloft.net>, Jakub Kicinski <kuba@kernel.org>
Cc: Herbert Xu <herbert@gondor.apana.org.au>,
	Steffen Klassert <steffen.klassert@secunet.com>,
	<netdev@vger.kernel.org>
Subject: [PATCH 08/10] xfrm: move policy_bydst RCU sync from per-netns .exit to .pre_exit
Date: Fri, 29 May 2026 11:26:10 +0200	[thread overview]
Message-ID: <20260529092648.3878973-9-steffen.klassert@secunet.com> (raw)
In-Reply-To: <20260529092648.3878973-1-steffen.klassert@secunet.com>

From: Usama Arif <usama.arif@linux.dev>

The struct pernet_operations docstring in include/net/net_namespace.h
explicitly warns against blocking RCU primitives in .exit handlers:

    Exit methods using blocking RCU primitives, such as
    synchronize_rcu(), should be implemented via exit_batch.
    [...]
    Please, avoid synchronize_rcu() at all, where it's possible.

    Note that a combination of pre_exit() and exit() can
    be used, since a synchronize_rcu() is guaranteed between
    the calls.

xfrm_policy_fini() violates this: it calls synchronize_rcu() before
freeing the policy_bydst hash tables (so no RCU reader is mid-
traversal at free time), but runs from xfrm_net_ops.exit -- once per
namespace -- so a cleanup_net() of N namespaces pays N full RCU
grace periods serially.

Use the documented pre_exit/exit split. Move the policy flush (and
the workqueue drains it depends on) into a new .pre_exit handler;
xfrm_policy_fini() then runs in .exit and frees the hash tables
after the synchronize_rcu_expedited() that cleanup_net() guarantees
between the two phases. Providing O(1) RCU grace periods per batch
instead of O(N).

Observed on Linux 6.18 with a workload doing unshare(CLONE_NEWNET)
at ~13/sec sustained: cleanup_net() and the netns_wq rescuer kthread
both stuck in xfrm_policy_fini()'s synchronize_rcu(), >300k struct
net accumulated in the cleanup queue, Percpu in /proc/meminfo climbed
to 130+ GB on 256-CPU hosts, and memcg OOMs followed. setup_net and
__put_net counts were balanced, ruling out a refcount leak.

Fixes: 069daad4f2ae ("xfrm: Wait for RCU readers during policy netns exit")
Signed-off-by: Usama Arif <usama.arif@linux.dev>
Signed-off-by: Steffen Klassert <steffen.klassert@secunet.com>
---
 net/xfrm/xfrm_policy.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/net/xfrm/xfrm_policy.c b/net/xfrm/xfrm_policy.c
index 59968dcbafe1..dd09d2063da2 100644
--- a/net/xfrm/xfrm_policy.c
+++ b/net/xfrm/xfrm_policy.c
@@ -4276,21 +4276,21 @@ static int __net_init xfrm_policy_init(struct net *net)
 	return -ENOMEM;
 }
 
-static void xfrm_policy_fini(struct net *net)
+static void __net_exit xfrm_net_pre_exit(struct net *net)
 {
-	struct xfrm_pol_inexact_bin *b, *t;
-	unsigned int sz;
-	int dir;
-
 	disable_work_sync(&net->xfrm.policy_hthresh.work);
-
 	flush_work(&net->xfrm.policy_hash_work);
 #ifdef CONFIG_XFRM_SUB_POLICY
 	xfrm_policy_flush(net, XFRM_POLICY_TYPE_SUB, false);
 #endif
 	xfrm_policy_flush(net, XFRM_POLICY_TYPE_MAIN, false);
+}
 
-	synchronize_rcu();
+static void xfrm_policy_fini(struct net *net)
+{
+	struct xfrm_pol_inexact_bin *b, *t;
+	unsigned int sz;
+	int dir;
 
 	WARN_ON(!list_empty(&net->xfrm.policy_all));
 
@@ -4368,6 +4368,7 @@ static void __net_exit xfrm_net_exit(struct net *net)
 
 static struct pernet_operations __net_initdata xfrm_net_ops = {
 	.init = xfrm_net_init,
+	.pre_exit = xfrm_net_pre_exit,
 	.exit = xfrm_net_exit,
 };
 
-- 
2.43.0


  parent reply	other threads:[~2026-05-29  9:27 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-05-29  9:26 [PATCH 0/10] pull request (net): ipsec 2026-05-29 Steffen Klassert
2026-05-29  9:26 ` [PATCH 01/10] xfrm: route MIGRATE notifications to caller's netns Steffen Klassert
2026-05-29 20:20   ` patchwork-bot+netdevbpf
2026-05-29  9:26 ` [PATCH 02/10] xfrm: ipcomp: Free destination pages on acomp errors Steffen Klassert
2026-05-29  9:26 ` [PATCH 03/10] xfrm: Check for underflow in xfrm_state_mtu Steffen Klassert
2026-05-29  9:26 ` [PATCH 04/10] xfrm: ah: use skb_to_full_sk in async output callbacks Steffen Klassert
2026-05-29  9:26 ` [PATCH 05/10] esp: fix page frag reference leak on skb_to_sgvec failure Steffen Klassert
2026-05-29  9:26 ` [PATCH 06/10] xfrm: esp: restore combined single-frag length gate Steffen Klassert
2026-05-29  9:26 ` [PATCH 07/10] xfrm: iptfs: reset runtime state when cloning SAs Steffen Klassert
2026-05-29  9:26 ` Steffen Klassert [this message]
2026-05-29  9:26 ` [PATCH 09/10] xfrm: input: hold netns during deferred transport reinjection Steffen Klassert
2026-05-29  9:26 ` [PATCH 10/10] Revert "esp: fix page frag reference leak on skb_to_sgvec failure" Steffen Klassert

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=20260529092648.3878973-9-steffen.klassert@secunet.com \
    --to=steffen.klassert@secunet.com \
    --cc=davem@davemloft.net \
    --cc=herbert@gondor.apana.org.au \
    --cc=kuba@kernel.org \
    --cc=netdev@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