From: Kuniyuki Iwashima <kuniyu@amazon.com>
To: "David S. Miller" <davem@davemloft.net>,
Eric Dumazet <edumazet@google.com>,
Jakub Kicinski <kuba@kernel.org>, Paolo Abeni <pabeni@redhat.com>
Cc: Simon Horman <horms@kernel.org>,
Kuniyuki Iwashima <kuniyu@amazon.com>,
Kuniyuki Iwashima <kuni1840@gmail.com>, <netdev@vger.kernel.org>
Subject: [PATCH v1 net-next 03/14] net: Add ->exit_rtnl() hook to struct pernet_operations.
Date: Wed, 9 Apr 2025 19:19:24 -0700 [thread overview]
Message-ID: <20250410022004.8668-4-kuniyu@amazon.com> (raw)
In-Reply-To: <20250410022004.8668-1-kuniyu@amazon.com>
struct pernet_operations provides two batching hooks; ->exit_batch()
and ->exit_batch_rtnl().
The batching variant is beneficial if ->exit() meets any of the
following conditions:
1) ->exit() repeatedly acquires a global lock for each netns
2) ->exit() has a time-consuming operation that can be factored
out (e.g. synchronize_rcu(), smp_mb(), etc)
3) ->exit() does not need to repeat the same iterations for each
netns (e.g. inet_twsk_purge())
Currently, none of the ->exit_batch_rtnl() functions satisfy any of
the above conditions because RTNL is factored out and held by the
caller and all of these functions iterate over the dying netns list.
Also, we want to hold per-netns RTNL there but avoid spreading
__rtnl_net_lock() across multiple locations.
Let's add ->exit_rtnl() hook and run it under __rtnl_net_lock().
The following patches will convert all ->exit_batch_rtnl() users
to ->exit_rtnl().
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
---
include/net/net_namespace.h | 2 ++
net/core/net_namespace.c | 53 +++++++++++++++++++++++++++----------
2 files changed, 41 insertions(+), 14 deletions(-)
diff --git a/include/net/net_namespace.h b/include/net/net_namespace.h
index bd57d8fb54f1..b071e6eed9d5 100644
--- a/include/net/net_namespace.h
+++ b/include/net/net_namespace.h
@@ -475,6 +475,8 @@ struct pernet_operations {
void (*exit)(struct net *net);
void (*exit_batch)(struct list_head *net_exit_list);
/* Following method is called with RTNL held. */
+ void (*exit_rtnl)(struct net *net,
+ struct list_head *dev_kill_list);
void (*exit_batch_rtnl)(struct list_head *net_exit_list,
struct list_head *dev_kill_list);
unsigned int * const id;
diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
index 5e0ee117cf25..7ea818342721 100644
--- a/net/core/net_namespace.c
+++ b/net/core/net_namespace.c
@@ -163,16 +163,51 @@ static void ops_pre_exit_list(const struct pernet_operations *ops,
}
}
+static void ops_exit_rtnl_list(const struct list_head *ops_list,
+ const struct pernet_operations *ops,
+ struct list_head *net_exit_list)
+{
+ const struct pernet_operations *saved_ops = ops;
+ LIST_HEAD(dev_kill_list);
+ struct net *net;
+
+ rtnl_lock();
+
+ list_for_each_entry(net, net_exit_list, exit_list) {
+ __rtnl_net_lock(net);
+
+ ops = saved_ops;
+ list_for_each_entry_continue_reverse(ops, ops_list, list) {
+ if (ops->exit_rtnl)
+ ops->exit_rtnl(net, &dev_kill_list);
+ }
+
+ __rtnl_net_unlock(net);
+ }
+
+ ops = saved_ops;
+ list_for_each_entry_continue_reverse(ops, ops_list, list) {
+ if (ops->exit_batch_rtnl)
+ ops->exit_batch_rtnl(net_exit_list, &dev_kill_list);
+ }
+
+ unregister_netdevice_many(&dev_kill_list);
+
+ rtnl_unlock();
+}
+
static void ops_exit_list(const struct pernet_operations *ops,
struct list_head *net_exit_list)
{
- struct net *net;
if (ops->exit) {
+ struct net *net;
+
list_for_each_entry(net, net_exit_list, exit_list) {
ops->exit(net);
cond_resched();
}
}
+
if (ops->exit_batch)
ops->exit_batch(net_exit_list);
}
@@ -213,18 +248,8 @@ static void ops_undo_list(const struct list_head *ops_list,
else
synchronize_rcu();
- if (hold_rtnl) {
- LIST_HEAD(dev_kill_list);
-
- ops = saved_ops;
- rtnl_lock();
- list_for_each_entry_continue_reverse(ops, ops_list, list) {
- if (ops->exit_batch_rtnl)
- ops->exit_batch_rtnl(net_exit_list, &dev_kill_list);
- }
- unregister_netdevice_many(&dev_kill_list);
- rtnl_unlock();
- }
+ if (hold_rtnl)
+ ops_exit_rtnl_list(ops_list, saved_ops, net_exit_list);
ops = saved_ops;
list_for_each_entry_continue_reverse(ops, ops_list, list)
@@ -238,7 +263,7 @@ static void ops_undo_list(const struct list_head *ops_list,
static void ops_undo_single(struct pernet_operations *ops,
struct list_head *net_exit_list)
{
- bool hold_rtnl = !!ops->exit_batch_rtnl;
+ bool hold_rtnl = ops->exit_rtnl || ops->exit_batch_rtnl;
LIST_HEAD(ops_list);
list_add(&ops->list, &ops_list);
--
2.49.0
next prev parent reply other threads:[~2025-04-10 2:21 UTC|newest]
Thread overview: 25+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-04-10 2:19 [PATCH v1 net-next 00/14] net: Convert ->exit_batch_rtnl() to ->exit_rtnl() Kuniyuki Iwashima
2025-04-10 2:19 ` [PATCH v1 net-next 01/14] net: Factorise setup_net() and cleanup_net() Kuniyuki Iwashima
2025-04-10 2:19 ` [PATCH v1 net-next 02/14] net: Add ops_undo_single for module load/unload Kuniyuki Iwashima
2025-04-11 3:16 ` kernel test robot
2025-04-11 3:32 ` OA[PATCH " Kuniyuki Iwashima
2025-04-11 4:37 ` [PATCH " kernel test robot
2025-04-10 2:19 ` Kuniyuki Iwashima [this message]
2025-04-10 2:19 ` [PATCH v1 net-next 04/14] nexthop: Convert nexthop_net_exit_batch_rtnl() to ->exit_rtnl() Kuniyuki Iwashima
2025-04-10 14:34 ` David Ahern
2025-04-10 2:19 ` [PATCH v1 net-next 05/14] vxlan: Convert vxlan_exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 9:07 ` Nikolay Aleksandrov
2025-04-10 2:19 ` [PATCH v1 net-next 06/14] ipv4: ip_tunnel: Convert ip_tunnel_delete_nets() callers " Kuniyuki Iwashima
2025-04-10 14:46 ` David Ahern
2025-04-10 2:19 ` [PATCH v1 net-next 07/14] ipv6: Convert tunnel devices' ->exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 14:50 ` David Ahern
2025-04-10 2:19 ` [PATCH v1 net-next 08/14] xfrm: Convert xfrmi_exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 2:19 ` [PATCH v1 net-next 09/14] bridge: Convert br_net_exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 7:05 ` Ido Schimmel
2025-04-10 9:04 ` Nikolay Aleksandrov
2025-04-10 2:19 ` [PATCH v1 net-next 10/14] bonding: Convert bond_net_exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 9:44 ` Nikolay Aleksandrov
2025-04-10 2:19 ` [PATCH v1 net-next 11/14] gtp: Convert gtp_net_exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 2:19 ` [PATCH v1 net-next 12/14] bareudp: Convert bareudp_exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 2:19 ` [PATCH v1 net-next 13/14] geneve: Convert geneve_exit_batch_rtnl() " Kuniyuki Iwashima
2025-04-10 2:19 ` [PATCH v1 net-next 14/14] net: Remove ->exit_batch_rtnl() Kuniyuki Iwashima
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=20250410022004.8668-4-kuniyu@amazon.com \
--to=kuniyu@amazon.com \
--cc=davem@davemloft.net \
--cc=edumazet@google.com \
--cc=horms@kernel.org \
--cc=kuba@kernel.org \
--cc=kuni1840@gmail.com \
--cc=netdev@vger.kernel.org \
--cc=pabeni@redhat.com \
/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).