From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: stable@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
patches@lists.linux.dev, Eric Dumazet <edumazet@google.com>,
Antoine Tenart <atenart@kernel.org>,
Jakub Kicinski <kuba@kernel.org>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 5.4 56/91] net: add exit_batch_rtnl() method
Date: Thu, 30 Jan 2025 15:01:15 +0100 [thread overview]
Message-ID: <20250130140135.912017826@linuxfoundation.org> (raw)
In-Reply-To: <20250130140133.662535583@linuxfoundation.org>
5.4-stable review patch. If anyone has any objections, please let me know.
------------------
From: Eric Dumazet <edumazet@google.com>
[ Upstream commit fd4f101edbd9f99567ab2adb1f2169579ede7c13 ]
Many (struct pernet_operations)->exit_batch() methods have
to acquire rtnl.
In presence of rtnl mutex pressure, this makes cleanup_net()
very slow.
This patch adds a new exit_batch_rtnl() method to reduce
number of rtnl acquisitions from cleanup_net().
exit_batch_rtnl() handlers are called while rtnl is locked,
and devices to be killed can be queued in a list provided
as their second argument.
A single unregister_netdevice_many() is called right
before rtnl is released.
exit_batch_rtnl() handlers are called before ->exit() and
->exit_batch() handlers.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Antoine Tenart <atenart@kernel.org>
Link: https://lore.kernel.org/r/20240206144313.2050392-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 46841c7053e6 ("gtp: Use for_each_netdev_rcu() in gtp_genl_dump_pdp().")
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
include/net/net_namespace.h | 3 +++
net/core/net_namespace.c | 31 ++++++++++++++++++++++++++++++-
2 files changed, 33 insertions(+), 1 deletion(-)
diff --git a/include/net/net_namespace.h b/include/net/net_namespace.h
index 167e390ac9d4e..0d61b452b9082 100644
--- a/include/net/net_namespace.h
+++ b/include/net/net_namespace.h
@@ -382,6 +382,9 @@ struct pernet_operations {
void (*pre_exit)(struct net *net);
void (*exit)(struct net *net);
void (*exit_batch)(struct list_head *net_exit_list);
+ /* Following method is called with RTNL held. */
+ void (*exit_batch_rtnl)(struct list_head *net_exit_list,
+ struct list_head *dev_kill_list);
unsigned int *id;
size_t size;
};
diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
index c4bcedc06822b..e6585a758edd4 100644
--- a/net/core/net_namespace.c
+++ b/net/core/net_namespace.c
@@ -325,8 +325,9 @@ static __net_init int setup_net(struct net *net, struct user_namespace *user_ns)
{
/* Must be called with pernet_ops_rwsem held */
const struct pernet_operations *ops, *saved_ops;
- int error = 0;
LIST_HEAD(net_exit_list);
+ LIST_HEAD(dev_kill_list);
+ int error = 0;
refcount_set(&net->count, 1);
refcount_set(&net->passive, 1);
@@ -359,6 +360,15 @@ static __net_init int setup_net(struct net *net, struct user_namespace *user_ns)
synchronize_rcu();
+ ops = saved_ops;
+ rtnl_lock();
+ list_for_each_entry_continue_reverse(ops, &pernet_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();
+
ops = saved_ops;
list_for_each_entry_continue_reverse(ops, &pernet_list, list)
ops_exit_list(ops, &net_exit_list);
@@ -563,6 +573,7 @@ static void cleanup_net(struct work_struct *work)
struct net *net, *tmp, *last;
struct llist_node *net_kill_list;
LIST_HEAD(net_exit_list);
+ LIST_HEAD(dev_kill_list);
/* Atomically snapshot the list of namespaces to cleanup */
net_kill_list = llist_del_all(&cleanup_list);
@@ -603,6 +614,14 @@ static void cleanup_net(struct work_struct *work)
*/
synchronize_rcu();
+ rtnl_lock();
+ list_for_each_entry_reverse(ops, &pernet_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();
+
/* Run all of the network namespace exit methods */
list_for_each_entry_reverse(ops, &pernet_list, list)
ops_exit_list(ops, &net_exit_list);
@@ -1150,7 +1169,17 @@ static void free_exit_list(struct pernet_operations *ops, struct list_head *net_
{
ops_pre_exit_list(ops, net_exit_list);
synchronize_rcu();
+
+ if (ops->exit_batch_rtnl) {
+ LIST_HEAD(dev_kill_list);
+
+ rtnl_lock();
+ ops->exit_batch_rtnl(net_exit_list, &dev_kill_list);
+ unregister_netdevice_many(&dev_kill_list);
+ rtnl_unlock();
+ }
ops_exit_list(ops, net_exit_list);
+
ops_free_list(ops, net_exit_list);
}
--
2.39.5
next prev parent reply other threads:[~2025-01-30 14:19 UTC|newest]
Thread overview: 97+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-01-30 14:00 [PATCH 5.4 00/91] 5.4.290-rc1 review Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 01/91] jbd2: flush filesystem device before updating tail sequence Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 02/91] dm array: fix releasing a faulty array block twice in dm_array_cursor_end Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 03/91] dm array: fix unreleased btree blocks on closing a faulty array cursor Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 04/91] dm array: fix cursor index when skipping across block boundaries Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 05/91] ieee802154: ca8210: Add missing check for kfifo_alloc() in ca8210_probe() Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 06/91] net: 802: LLC+SNAP OID:PID lookup on start of skb data Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 07/91] tcp/dccp: complete lockless accesses to sk->sk_max_ack_backlog Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 08/91] tcp/dccp: allow a connection when sk_max_ack_backlog is zero Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 09/91] net_sched: cls_flow: validate TCA_FLOW_RSHIFT attribute Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 10/91] tls: Fix tls_sw_sendmsg error handling Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 11/91] dm thin: make get_first_thin use rcu-safe list first function Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 12/91] sctp: sysctl: cookie_hmac_alg: avoid using current->nsproxy Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 13/91] sctp: sysctl: auth_enable: " Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 14/91] drm/amd/display: Add check for granularity in dml ceil/floor helpers Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 15/91] ACPI: resource: Add TongFang GM5HG0A to irq1_edge_low_force_override[] Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 16/91] ACPI: resource: Add Asus Vivobook X1504VAP to irq1_level_low_skip_override[] Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 17/91] drm/amd/display: increase MAX_SURFACES to the value supported by hw Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 18/91] USB: serial: option: add MeiG Smart SRM815 Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 19/91] USB: serial: option: add Neoway N723-EA support Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 20/91] staging: iio: ad9834: Correct phase range check Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 21/91] staging: iio: ad9832: " Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 22/91] usb-storage: Add max sectors quirk for Nokia 208 Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 23/91] USB: serial: cp210x: add Phoenix Contact UPS Device Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 24/91] usb: gadget: u_serial: Disable ep before setting port to null to fix the crash caused by port being null Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 25/91] USB: usblp: return error when setting unsupported protocol Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 26/91] USB: core: Disable LPM only for non-suspended ports Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 27/91] usb: fix reference leak in usb_new_device() Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 28/91] usb: gadget: f_fs: Remove WARN_ON in functionfs_bind Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 29/91] iio: pressure: zpa2326: fix information leak in triggered buffer Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 30/91] iio: dummy: iio_simply_dummy_buffer: " Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 31/91] iio: light: vcnl4035: " Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 32/91] iio: imu: kmx61: " Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 33/91] iio: adc: ti-ads8688: " Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 34/91] iio: gyro: fxas21002c: Fix missing data update in trigger handler Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 35/91] iio: adc: ti-ads124s08: Use gpiod_set_value_cansleep() Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 36/91] iio: adc: at91: call input_free_device() on allocated iio_dev Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 37/91] iio: inkern: call iio_device_put() only on mapped devices Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 38/91] arm64: dts: rockchip: fix defines in pd_vio node for rk3399 Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 39/91] arm64: dts: rockchip: fix pd_tcpc0 and pd_tcpc1 node position on rk3399 Greg Kroah-Hartman
2025-01-30 14:00 ` [PATCH 5.4 40/91] arm64: dts: rockchip: add #power-domain-cells to power domain nodes Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 41/91] arm64: dts: rockchip: add hevc power domain clock to rk3328 Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 42/91] phy: core: fix code style in devm_of_phy_provider_unregister Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 43/91] phy: core: Fix that API devm_of_phy_provider_unregister() fails to unregister the phy provider Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 44/91] ocfs2: correct return value of ocfs2_local_free_info() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 45/91] ocfs2: fix slab-use-after-free due to dangling pointer dqi_priv Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 46/91] sctp: sysctl: rto_min/max: avoid using current->nsproxy Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 47/91] riscv: abstract out CSR names for supervisor vs machine mode Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 48/91] riscv: remove unused handle_exception symbol Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 49/91] RISC-V: Avoid dereferening NULL regs in die() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 50/91] riscv: Avoid enabling interrupts " Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 51/91] riscv: Fix sleeping in invalid context " Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 52/91] riscv: prefix IRQ_ macro names with an RV_ namespace Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 53/91] RISC-V: Dont enable all interrupts in trap_init() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 54/91] net: ethernet: ti: cpsw_ale: Fix cpsw_ale_get_field() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 55/91] net: net_namespace: Optimize the code Greg Kroah-Hartman
2025-01-30 14:01 ` Greg Kroah-Hartman [this message]
2025-01-30 14:01 ` [PATCH 5.4 57/91] gtp: use exit_batch_rtnl() method Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 58/91] gtp: Use for_each_netdev_rcu() in gtp_genl_dump_pdp() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 59/91] gtp: Destroy device along with udp sockets netns dismantle Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 60/91] nfp: bpf: prevent integer overflow in nfp_bpf_event_output() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 61/91] drm/v3d: Ensure job pointer is set to NULL after job completion Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 62/91] i2c: mux: demux-pinctrl: check initial mux selection, too Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 63/91] mac802154: check local interfaces before deleting sdata list Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 64/91] hfs: Sanity check the root record Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 65/91] kheaders: Ignore silly-rename files Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 66/91] poll_wait: add mb() to fix theoretical race between waitqueue_active() and .poll() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 67/91] nvmet: propagate npwg topology Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 68/91] net: ethernet: xgbe: re-add aneg to supported features in PHY quirks Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 69/91] fs/proc: fix softlockup in __read_vmcore (part 2) Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 70/91] irqchip/gic-v3: Handle CPU_PM_ENTER_FAILED correctly Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 71/91] hrtimers: Handle CPU state correctly on hotplug Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 72/91] ipv6: avoid possible NULL deref in rt6_uncached_list_flush_dev() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 73/91] scsi: sg: Fix slab-use-after-free read in sg_release() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 74/91] net: fix data-races around sk->sk_forward_alloc Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 75/91] ASoC: wm8994: Add depends on MFD core Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 76/91] scsi: iscsi: Fix redundant response for ISCSI_UEVENT_GET_HOST_STATS request Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 77/91] irqchip/sunxi-nmi: Add missing SKIP_WAKE flag Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 78/91] gfs2: Truncate address space when flipping GFS2_DIF_JDATA flag Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 79/91] m68k: Update ->thread.esp0 before calling syscall_trace() in ret_from_signal Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 80/91] m68k: Add missing mmap_read_lock() to sys_cacheflush() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 81/91] signal/m68k: Use force_sigsegv(SIGSEGV) in fpsp040_die Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 82/91] net: xen-netback: hash.c: Use built-in RCU list checking Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 83/91] net/xen-netback: prevent UAF in xenvif_flush_hash() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 84/91] vfio/platform: check the bounds of read/write syscalls Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 85/91] ext4: avoid ext4_error()s caused by ENOMEM in the truncate path Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 86/91] ext4: fix slab-use-after-free in ext4_split_extent_at() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 87/91] USB: serial: quatech2: fix null-ptr-deref in qt2_process_read_urb() Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 88/91] Revert "usb: gadget: u_serial: Disable ep before setting port to null to fix the crash caused by port being null" Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 89/91] Input: atkbd - map F23 key to support default copilot shortcut Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 90/91] Input: xpad - add unofficial Xbox 360 wireless receiver clone Greg Kroah-Hartman
2025-01-30 14:01 ` [PATCH 5.4 91/91] Input: xpad - add support for wooting two he (arm) Greg Kroah-Hartman
2025-01-30 19:35 ` [PATCH 5.4 00/91] 5.4.290-rc1 review Naresh Kamboju
2025-01-31 8:57 ` Greg Kroah-Hartman
2025-01-30 20:24 ` Florian Fainelli
2025-01-30 22:21 ` Pavel Machek
2025-01-31 5:38 ` Jon Hunter
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=20250130140135.912017826@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=atenart@kernel.org \
--cc=edumazet@google.com \
--cc=kuba@kernel.org \
--cc=patches@lists.linux.dev \
--cc=sashal@kernel.org \
--cc=stable@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