diff mbox series

[net-next,v6,1/2] net: sched: use queue_mapping to pick tx queue

Message ID 20211222120809.2222-2-xiangxia.m.yue@gmail.com (mailing list archive)
State Superseded
Delegated to: Netdev Maintainers
Headers show
Series net: sched: allow user to select txqueue | expand

Checks

Context Check Description
netdev/tree_selection success Clearly marked for net-next
netdev/fixes_present success Fixes tag not required for -next series
netdev/subject_prefix success Link
netdev/cover_letter success Series has a cover letter
netdev/patch_count success Link
netdev/header_inline success No static functions without inline keyword in header files
netdev/build_32bit success Errors and warnings before: 4765 this patch: 4765
netdev/cc_maintainers success CCed 8 of 8 maintainers
netdev/build_clang success Errors and warnings before: 841 this patch: 841
netdev/module_param success Was 0 now: 0
netdev/verify_signedoff success Signed-off-by tag matches author and committer
netdev/verify_fixes success No Fixes tag
netdev/build_allmodconfig_warn success Errors and warnings before: 4918 this patch: 4918
netdev/checkpatch warning WARNING: please, no spaces at the start of a line
netdev/kdoc success Errors and warnings before: 0 this patch: 0
netdev/source_inline fail Was 0 now: 2

Commit Message

Tonghao Zhang Dec. 22, 2021, 12:08 p.m. UTC 
From: Tonghao Zhang <xiangxia.m.yue@gmail.com>

This patch fixes issue:
* If we install tc filters with act_skbedit in clsact hook.
  It doesn't work, because netdev_core_pick_tx() overwrites
  queue_mapping.

  $ tc filter ... action skbedit queue_mapping 1

And this patch is useful:
* We can use FQ + EDT to implement efficient policies. Tx queues
  are picked by xps, ndo_select_queue of netdev driver, or skb hash
  in netdev_core_pick_tx(). In fact, the netdev driver, and skb
  hash are _not_ under control. xps uses the CPUs map to select Tx
  queues, but we can't figure out which task_struct of pod/containter
  running on this cpu in most case. We can use clsact filters to classify
  one pod/container traffic to one Tx queue. Why ?

  In containter networking environment, there are two kinds of pod/
  containter/net-namespace. One kind (e.g. P1, P2), the high throughput
  is key in these applications. But avoid running out of network resource,
  the outbound traffic of these pods is limited, using or sharing one
  dedicated Tx queues assigned HTB/TBF/FQ Qdisc. Other kind of pods
  (e.g. Pn), the low latency of data access is key. And the traffic is not
  limited. Pods use or share other dedicated Tx queues assigned FIFO Qdisc.
  This choice provides two benefits. First, contention on the HTB/FQ Qdisc
  lock is significantly reduced since fewer CPUs contend for the same queue.
  More importantly, Qdisc contention can be eliminated completely if each
  CPU has its own FIFO Qdisc for the second kind of pods.

  There must be a mechanism in place to support classifying traffic based on
  pods/container to different Tx queues. Note that clsact is outside of Qdisc
  while Qdisc can run a classifier to select a sub-queue under the lock.

  In general recording the decision in the skb seems a little heavy handed.
  This patch introduces a per-CPU variable, suggested by Eric.

  The xmit.skip_txqueue flag is firstly cleared in __dev_queue_xmit().
  - Tx Qdisc may install that skbedit actions, then xmit.skip_txqueue flag
    is set in qdisc->enqueue() though tx queue has been selected in
    netdev_tx_queue_mapping() or netdev_core_pick_tx(). That flag is cleared
    firstly in __dev_queue_xmit(), is useful:
  - Avoid picking Tx queue with netdev_tx_queue_mapping() in next netdev
    in such case: eth0 macvlan - eth0.3 vlan - eth0 ixgbe-phy:
    For example, eth0, macvlan in pod, which root Qdisc install skbedit
    queue_mapping, send packets to eth0.3, vlan in host. In __dev_queue_xmit() of
    eth0.3, clear the flag, does not select tx queue according to skb->queue_mapping
    because there is no filters in clsact or tx Qdisc of this netdev.
    Same action taked in eth0, ixgbe in Host.
  - Avoid picking Tx queue for next packet. If we set xmit.skip_txqueue
    in tx Qdisc (qdisc->enqueue()), the proper way to clear it is clearing it
    in __dev_queue_xmit when processing next packets.

  For performance reasons, use the static key. If user does not config the NET_EGRESS,
  the patch will not be compiled.

  +----+      +----+      +----+
  | P1 |      | P2 |      | Pn |
  +----+      +----+      +----+
    |           |           |
    +-----------+-----------+
                |
                | clsact/skbedit
                |      MQ
                v
    +-----------+-----------+
    | q0        | q1        | qn
    v           v           v
  HTB/FQ      HTB/FQ  ...  FIFO

Cc: Jamal Hadi Salim <jhs@mojatatu.com>
Cc: Cong Wang <xiyou.wangcong@gmail.com>
Cc: Jiri Pirko <jiri@resnulli.us>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Jonathan Lemon <jonathan.lemon@gmail.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Alexander Lobakin <alobakin@pm.me>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Talal Ahmad <talalahmad@google.com>
Cc: Kevin Hao <haokexin@gmail.com>
Cc: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Cc: Antoine Tenart <atenart@kernel.org>
Cc: Wei Wang <weiwan@google.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Tonghao Zhang <xiangxia.m.yue@gmail.com>
---
 include/linux/netdevice.h |  3 +++
 include/linux/rtnetlink.h |  3 +++
 net/core/dev.c            | 44 ++++++++++++++++++++++++++++++++++++++-
 net/sched/act_skbedit.c   | 18 ++++++++++++++--
 4 files changed, 65 insertions(+), 3 deletions(-)

Comments

kernel test robot Dec. 22, 2021, 10:44 p.m. UTC | #1 
Hi,

Thank you for the patch! Perhaps something to improve:

[auto build test WARNING on net-next/master]

url:    https://github.com/0day-ci/linux/commits/xiangxia-m-yue-gmail-com/net-sched-allow-user-to-select-txqueue/20211222-201128
base:   https://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next.git f4f2970dfd87e5132c436e6125148914596a9863
config: i386-randconfig-m031-20211222 (https://download.01.org/0day-ci/archive/20211223/202112230652.grNJivfH-lkp@intel.com/config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0

If you fix the issue, kindly add following tag as appropriate
Reported-by: kernel test robot <lkp@intel.com>

smatch warnings:
net/core/dev.c:4114 __dev_queue_xmit() warn: inconsistent indenting

vim +4114 net/core/dev.c

638b2a699fd3ec9 Jiri Pirko             2015-05-12  4036  
d29f749e252bcdb Dave Jones             2008-07-22  4037  /**
9d08dd3d320fab4 Jason Wang             2014-01-20  4038   *	__dev_queue_xmit - transmit a buffer
d29f749e252bcdb Dave Jones             2008-07-22  4039   *	@skb: buffer to transmit
eadec877ce9ca46 Alexander Duyck        2018-07-09  4040   *	@sb_dev: suboordinate device used for L2 forwarding offload
d29f749e252bcdb Dave Jones             2008-07-22  4041   *
d29f749e252bcdb Dave Jones             2008-07-22  4042   *	Queue a buffer for transmission to a network device. The caller must
d29f749e252bcdb Dave Jones             2008-07-22  4043   *	have set the device and priority and built the buffer before calling
d29f749e252bcdb Dave Jones             2008-07-22  4044   *	this function. The function can be called from an interrupt.
d29f749e252bcdb Dave Jones             2008-07-22  4045   *
d29f749e252bcdb Dave Jones             2008-07-22  4046   *	A negative errno code is returned on a failure. A success does not
d29f749e252bcdb Dave Jones             2008-07-22  4047   *	guarantee the frame will be transmitted as it may be dropped due
d29f749e252bcdb Dave Jones             2008-07-22  4048   *	to congestion or traffic shaping.
d29f749e252bcdb Dave Jones             2008-07-22  4049   *
d29f749e252bcdb Dave Jones             2008-07-22  4050   * -----------------------------------------------------------------------------------
d29f749e252bcdb Dave Jones             2008-07-22  4051   *      I notice this method can also return errors from the queue disciplines,
d29f749e252bcdb Dave Jones             2008-07-22  4052   *      including NET_XMIT_DROP, which is a positive value.  So, errors can also
d29f749e252bcdb Dave Jones             2008-07-22  4053   *      be positive.
d29f749e252bcdb Dave Jones             2008-07-22  4054   *
d29f749e252bcdb Dave Jones             2008-07-22  4055   *      Regardless of the return value, the skb is consumed, so it is currently
d29f749e252bcdb Dave Jones             2008-07-22  4056   *      difficult to retry a send to this method.  (You can bump the ref count
d29f749e252bcdb Dave Jones             2008-07-22  4057   *      before sending to hold a reference for retry if you are careful.)
d29f749e252bcdb Dave Jones             2008-07-22  4058   *
d29f749e252bcdb Dave Jones             2008-07-22  4059   *      When calling this method, interrupts MUST be enabled.  This is because
d29f749e252bcdb Dave Jones             2008-07-22  4060   *      the BH enable code must have IRQs enabled so that it will not deadlock.
d29f749e252bcdb Dave Jones             2008-07-22  4061   *          --BLG
d29f749e252bcdb Dave Jones             2008-07-22  4062   */
eadec877ce9ca46 Alexander Duyck        2018-07-09  4063  static int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev)
^1da177e4c3f415 Linus Torvalds         2005-04-16  4064  {
^1da177e4c3f415 Linus Torvalds         2005-04-16  4065  	struct net_device *dev = skb->dev;
dc2b48475a0a36f David S. Miller        2008-07-08  4066  	struct netdev_queue *txq;
^1da177e4c3f415 Linus Torvalds         2005-04-16  4067  	struct Qdisc *q;
^1da177e4c3f415 Linus Torvalds         2005-04-16  4068  	int rc = -ENOMEM;
f53c723902d1ac5 Steffen Klassert       2017-12-20  4069  	bool again = false;
^1da177e4c3f415 Linus Torvalds         2005-04-16  4070  
6d1ccff62780682 Eric Dumazet           2013-02-05  4071  	skb_reset_mac_header(skb);
6d1ccff62780682 Eric Dumazet           2013-02-05  4072  
e7fd2885385157d Willem de Bruijn       2014-08-04  4073  	if (unlikely(skb_shinfo(skb)->tx_flags & SKBTX_SCHED_TSTAMP))
e7ed11ee945438b Yousuk Seung           2021-01-20  4074  		__skb_tstamp_tx(skb, NULL, NULL, skb->sk, SCM_TSTAMP_SCHED);
e7fd2885385157d Willem de Bruijn       2014-08-04  4075  
^1da177e4c3f415 Linus Torvalds         2005-04-16  4076  	/* Disable soft irqs for various locks below. Also
^1da177e4c3f415 Linus Torvalds         2005-04-16  4077  	 * stops preemption for RCU.
^1da177e4c3f415 Linus Torvalds         2005-04-16  4078  	 */
d4828d85d188dc7 Herbert Xu             2006-06-22  4079  	rcu_read_lock_bh();
^1da177e4c3f415 Linus Torvalds         2005-04-16  4080  
5bc1421e34ecfe0 Neil Horman            2011-11-22  4081  	skb_update_prio(skb);
5bc1421e34ecfe0 Neil Horman            2011-11-22  4082  
1f211a1b929c804 Daniel Borkmann        2016-01-07  4083  	qdisc_pkt_len_init(skb);
1f211a1b929c804 Daniel Borkmann        2016-01-07  4084  #ifdef CONFIG_NET_CLS_ACT
8dc07fdbf2054f1 Willem de Bruijn       2017-01-07  4085  	skb->tc_at_ingress = 0;
42df6e1d221dddc Lukas Wunner           2021-10-08  4086  #endif
1f211a1b929c804 Daniel Borkmann        2016-01-07  4087  #ifdef CONFIG_NET_EGRESS
3435309167e51b1 Tonghao Zhang          2021-12-22  4088  	if (static_branch_unlikely(&txqueue_needed_key))
3435309167e51b1 Tonghao Zhang          2021-12-22  4089  		netdev_xmit_skip_txqueue(false);
3435309167e51b1 Tonghao Zhang          2021-12-22  4090  
aabf6772cc745f9 Davidlohr Bueso        2018-05-08  4091  	if (static_branch_unlikely(&egress_needed_key)) {
42df6e1d221dddc Lukas Wunner           2021-10-08  4092  		if (nf_hook_egress_active()) {
42df6e1d221dddc Lukas Wunner           2021-10-08  4093  			skb = nf_hook_egress(skb, &rc, dev);
42df6e1d221dddc Lukas Wunner           2021-10-08  4094  			if (!skb)
42df6e1d221dddc Lukas Wunner           2021-10-08  4095  				goto out;
42df6e1d221dddc Lukas Wunner           2021-10-08  4096  		}
42df6e1d221dddc Lukas Wunner           2021-10-08  4097  		nf_skip_egress(skb, true);
1f211a1b929c804 Daniel Borkmann        2016-01-07  4098  		skb = sch_handle_egress(skb, &rc, dev);
1f211a1b929c804 Daniel Borkmann        2016-01-07  4099  		if (!skb)
1f211a1b929c804 Daniel Borkmann        2016-01-07  4100  			goto out;
42df6e1d221dddc Lukas Wunner           2021-10-08  4101  		nf_skip_egress(skb, false);
1f211a1b929c804 Daniel Borkmann        2016-01-07  4102  	}
3435309167e51b1 Tonghao Zhang          2021-12-22  4103  
3435309167e51b1 Tonghao Zhang          2021-12-22  4104  	if (static_branch_unlikely(&txqueue_needed_key) &&
3435309167e51b1 Tonghao Zhang          2021-12-22  4105  	    netdev_xmit_txqueue_skipped())
3435309167e51b1 Tonghao Zhang          2021-12-22  4106  		txq = netdev_tx_queue_mapping(dev, skb);
3435309167e51b1 Tonghao Zhang          2021-12-22  4107  	else
1f211a1b929c804 Daniel Borkmann        2016-01-07  4108  #endif
3435309167e51b1 Tonghao Zhang          2021-12-22  4109  		txq = netdev_core_pick_tx(dev, skb, sb_dev);
3435309167e51b1 Tonghao Zhang          2021-12-22  4110  
0287587884b1504 Eric Dumazet           2014-10-05  4111  	/* If device/qdisc don't need skb->dst, release it right now while
0287587884b1504 Eric Dumazet           2014-10-05  4112  	 * its hot in this cpu cache.
0287587884b1504 Eric Dumazet           2014-10-05  4113  	 */
0287587884b1504 Eric Dumazet           2014-10-05 @4114  	if (dev->priv_flags & IFF_XMIT_DST_RELEASE)
0287587884b1504 Eric Dumazet           2014-10-05  4115  		skb_dst_drop(skb);
0287587884b1504 Eric Dumazet           2014-10-05  4116  	else
0287587884b1504 Eric Dumazet           2014-10-05  4117  		skb_dst_force(skb);
0287587884b1504 Eric Dumazet           2014-10-05  4118  
a898def29e4119b Paul E. McKenney       2010-02-22  4119  	q = rcu_dereference_bh(txq->qdisc);
37437bb2e1ae8af David S. Miller        2008-07-16  4120  
cf66ba58b5cb8b1 Koki Sanagi            2010-08-23  4121  	trace_net_dev_queue(skb);
^1da177e4c3f415 Linus Torvalds         2005-04-16  4122  	if (q->enqueue) {
bbd8a0d3a3b65d3 Krishna Kumar          2009-08-06  4123  		rc = __dev_xmit_skb(skb, q, dev, txq);
^1da177e4c3f415 Linus Torvalds         2005-04-16  4124  		goto out;
^1da177e4c3f415 Linus Torvalds         2005-04-16  4125  	}
^1da177e4c3f415 Linus Torvalds         2005-04-16  4126  
^1da177e4c3f415 Linus Torvalds         2005-04-16  4127  	/* The device has no queue. Common case for software devices:
eb13da1a103a808 tcharding              2017-02-09  4128  	 * loopback, all the sorts of tunnels...
^1da177e4c3f415 Linus Torvalds         2005-04-16  4129  
eb13da1a103a808 tcharding              2017-02-09  4130  	 * Really, it is unlikely that netif_tx_lock protection is necessary
eb13da1a103a808 tcharding              2017-02-09  4131  	 * here.  (f.e. loopback and IP tunnels are clean ignoring statistics
eb13da1a103a808 tcharding              2017-02-09  4132  	 * counters.)
eb13da1a103a808 tcharding              2017-02-09  4133  	 * However, it is possible, that they rely on protection
eb13da1a103a808 tcharding              2017-02-09  4134  	 * made by us here.
^1da177e4c3f415 Linus Torvalds         2005-04-16  4135  
eb13da1a103a808 tcharding              2017-02-09  4136  	 * Check this and shot the lock. It is not prone from deadlocks.
eb13da1a103a808 tcharding              2017-02-09  4137  	 *Either shot noqueue qdisc, it is even simpler 8)
^1da177e4c3f415 Linus Torvalds         2005-04-16  4138  	 */
^1da177e4c3f415 Linus Torvalds         2005-04-16  4139  	if (dev->flags & IFF_UP) {
^1da177e4c3f415 Linus Torvalds         2005-04-16  4140  		int cpu = smp_processor_id(); /* ok because BHs are off */
^1da177e4c3f415 Linus Torvalds         2005-04-16  4141  
7a10d8c810cfad3 Eric Dumazet           2021-11-30  4142  		/* Other cpus might concurrently change txq->xmit_lock_owner
7a10d8c810cfad3 Eric Dumazet           2021-11-30  4143  		 * to -1 or to their cpu id, but not to our id.
7a10d8c810cfad3 Eric Dumazet           2021-11-30  4144  		 */
7a10d8c810cfad3 Eric Dumazet           2021-11-30  4145  		if (READ_ONCE(txq->xmit_lock_owner) != cpu) {
97cdcf37b57e3f2 Florian Westphal       2019-04-01  4146  			if (dev_xmit_recursion())
745e20f1b626b1b Eric Dumazet           2010-09-29  4147  				goto recursion_alert;
745e20f1b626b1b Eric Dumazet           2010-09-29  4148  
f53c723902d1ac5 Steffen Klassert       2017-12-20  4149  			skb = validate_xmit_skb(skb, dev, &again);
1f59533f9ca5634 Jesper Dangaard Brouer 2014-09-03  4150  			if (!skb)
d21fd63ea385620 Eric Dumazet           2016-04-12  4151  				goto out;
1f59533f9ca5634 Jesper Dangaard Brouer 2014-09-03  4152  
3744741adab6d91 Willy Tarreau          2020-08-10  4153  			PRANDOM_ADD_NOISE(skb, dev, txq, jiffies);
c773e847ea8f681 David S. Miller        2008-07-08  4154  			HARD_TX_LOCK(dev, txq, cpu);
^1da177e4c3f415 Linus Torvalds         2005-04-16  4155  
7346649826382b7 Tom Herbert            2011-11-28  4156  			if (!netif_xmit_stopped(txq)) {
97cdcf37b57e3f2 Florian Westphal       2019-04-01  4157  				dev_xmit_recursion_inc();
ce93718fb7cdbc0 David S. Miller        2014-08-30  4158  				skb = dev_hard_start_xmit(skb, dev, txq, &rc);
97cdcf37b57e3f2 Florian Westphal       2019-04-01  4159  				dev_xmit_recursion_dec();
572a9d7b6fc7f20 Patrick McHardy        2009-11-10  4160  				if (dev_xmit_complete(rc)) {
c773e847ea8f681 David S. Miller        2008-07-08  4161  					HARD_TX_UNLOCK(dev, txq);
^1da177e4c3f415 Linus Torvalds         2005-04-16  4162  					goto out;
^1da177e4c3f415 Linus Torvalds         2005-04-16  4163  				}
^1da177e4c3f415 Linus Torvalds         2005-04-16  4164  			}
c773e847ea8f681 David S. Miller        2008-07-08  4165  			HARD_TX_UNLOCK(dev, txq);
e87cc4728f0e2fb Joe Perches            2012-05-13  4166  			net_crit_ratelimited("Virtual device %s asks to queue packet!\n",
7b6cd1ce72176e2 Joe Perches            2012-02-01  4167  					     dev->name);
^1da177e4c3f415 Linus Torvalds         2005-04-16  4168  		} else {
^1da177e4c3f415 Linus Torvalds         2005-04-16  4169  			/* Recursion is detected! It is possible,
745e20f1b626b1b Eric Dumazet           2010-09-29  4170  			 * unfortunately
745e20f1b626b1b Eric Dumazet           2010-09-29  4171  			 */
745e20f1b626b1b Eric Dumazet           2010-09-29  4172  recursion_alert:
e87cc4728f0e2fb Joe Perches            2012-05-13  4173  			net_crit_ratelimited("Dead loop on virtual device %s, fix it urgently!\n",
7b6cd1ce72176e2 Joe Perches            2012-02-01  4174  					     dev->name);
^1da177e4c3f415 Linus Torvalds         2005-04-16  4175  		}
^1da177e4c3f415 Linus Torvalds         2005-04-16  4176  	}
^1da177e4c3f415 Linus Torvalds         2005-04-16  4177  
^1da177e4c3f415 Linus Torvalds         2005-04-16  4178  	rc = -ENETDOWN;
d4828d85d188dc7 Herbert Xu             2006-06-22  4179  	rcu_read_unlock_bh();
^1da177e4c3f415 Linus Torvalds         2005-04-16  4180  
015f0688f57ca4d Eric Dumazet           2014-03-27  4181  	atomic_long_inc(&dev->tx_dropped);
1f59533f9ca5634 Jesper Dangaard Brouer 2014-09-03  4182  	kfree_skb_list(skb);
^1da177e4c3f415 Linus Torvalds         2005-04-16  4183  	return rc;
^1da177e4c3f415 Linus Torvalds         2005-04-16  4184  out:
d4828d85d188dc7 Herbert Xu             2006-06-22  4185  	rcu_read_unlock_bh();
^1da177e4c3f415 Linus Torvalds         2005-04-16  4186  	return rc;
^1da177e4c3f415 Linus Torvalds         2005-04-16  4187  }
f663dd9aaf9ed12 Jason Wang             2014-01-10  4188  

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org
diff mbox series

Patch

diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index 8b0bdeb4734e..708e9f4cca01 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -3010,6 +3010,9 @@  struct softnet_data {
 	struct {
 		u16 recursion;
 		u8  more;
+#ifdef CONFIG_NET_EGRESS
+		u8  skip_txqueue;
+#endif
 	} xmit;
 #ifdef CONFIG_RPS
 	/* input_queue_head should be written by cpu owning this struct,
diff --git a/include/linux/rtnetlink.h b/include/linux/rtnetlink.h
index bb9cb84114c1..256bf78daea6 100644
--- a/include/linux/rtnetlink.h
+++ b/include/linux/rtnetlink.h
@@ -100,6 +100,9 @@  void net_dec_ingress_queue(void);
 #ifdef CONFIG_NET_EGRESS
 void net_inc_egress_queue(void);
 void net_dec_egress_queue(void);
+void net_inc_queue_mapping(void);
+void net_dec_queue_mapping(void);
+void netdev_xmit_skip_txqueue(bool skip);
 #endif
 
 void rtnetlink_init(void);
diff --git a/net/core/dev.c b/net/core/dev.c
index a855e41bbe39..b197dabcd721 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -1998,6 +1998,20 @@  void net_dec_egress_queue(void)
 	static_branch_dec(&egress_needed_key);
 }
 EXPORT_SYMBOL_GPL(net_dec_egress_queue);
+
+static DEFINE_STATIC_KEY_FALSE(txqueue_needed_key);
+
+void net_inc_queue_mapping(void)
+{
+	static_branch_inc(&txqueue_needed_key);
+}
+EXPORT_SYMBOL_GPL(net_inc_queue_mapping);
+
+void net_dec_queue_mapping(void)
+{
+	static_branch_dec(&txqueue_needed_key);
+}
+EXPORT_SYMBOL_GPL(net_dec_queue_mapping);
 #endif
 
 static DEFINE_STATIC_KEY_FALSE(netstamp_needed_key);
@@ -3860,6 +3874,25 @@  sch_handle_egress(struct sk_buff *skb, int *ret, struct net_device *dev)
 
 	return skb;
 }
+
+static inline struct netdev_queue *
+netdev_tx_queue_mapping(struct net_device *dev, struct sk_buff *skb)
+{
+       int qm = skb_get_queue_mapping(skb);
+
+       return netdev_get_tx_queue(dev, netdev_cap_txqueue(dev, qm));
+}
+
+static inline bool netdev_xmit_txqueue_skipped(void)
+{
+	return __this_cpu_read(softnet_data.xmit.skip_txqueue);
+}
+
+void netdev_xmit_skip_txqueue(bool skip)
+{
+	__this_cpu_write(softnet_data.xmit.skip_txqueue, skip);
+}
+EXPORT_SYMBOL_GPL(netdev_xmit_skip_txqueue);
 #endif /* CONFIG_NET_EGRESS */
 
 #ifdef CONFIG_XPS
@@ -4052,6 +4085,9 @@  static int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev)
 	skb->tc_at_ingress = 0;
 #endif
 #ifdef CONFIG_NET_EGRESS
+	if (static_branch_unlikely(&txqueue_needed_key))
+		netdev_xmit_skip_txqueue(false);
+
 	if (static_branch_unlikely(&egress_needed_key)) {
 		if (nf_hook_egress_active()) {
 			skb = nf_hook_egress(skb, &rc, dev);
@@ -4064,7 +4100,14 @@  static int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev)
 			goto out;
 		nf_skip_egress(skb, false);
 	}
+
+	if (static_branch_unlikely(&txqueue_needed_key) &&
+	    netdev_xmit_txqueue_skipped())
+		txq = netdev_tx_queue_mapping(dev, skb);
+	else
 #endif
+		txq = netdev_core_pick_tx(dev, skb, sb_dev);
+
 	/* If device/qdisc don't need skb->dst, release it right now while
 	 * its hot in this cpu cache.
 	 */
@@ -4073,7 +4116,6 @@  static int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev)
 	else
 		skb_dst_force(skb);
 
-	txq = netdev_core_pick_tx(dev, skb, sb_dev);
 	q = rcu_dereference_bh(txq->qdisc);
 
 	trace_net_dev_queue(skb);
diff --git a/net/sched/act_skbedit.c b/net/sched/act_skbedit.c
index ceba11b198bb..325991080a8a 100644
--- a/net/sched/act_skbedit.c
+++ b/net/sched/act_skbedit.c
@@ -58,8 +58,12 @@  static int tcf_skbedit_act(struct sk_buff *skb, const struct tc_action *a,
 		}
 	}
 	if (params->flags & SKBEDIT_F_QUEUE_MAPPING &&
-	    skb->dev->real_num_tx_queues > params->queue_mapping)
+	    skb->dev->real_num_tx_queues > params->queue_mapping) {
+#ifdef CONFIG_NET_EGRESS
+		netdev_xmit_skip_txqueue(true);
+#endif
 		skb_set_queue_mapping(skb, params->queue_mapping);
+	}
 	if (params->flags & SKBEDIT_F_MARK) {
 		skb->mark &= ~params->mask;
 		skb->mark |= params->mark & params->mask;
@@ -225,6 +229,11 @@  static int tcf_skbedit_init(struct net *net, struct nlattr *nla,
 	if (goto_ch)
 		tcf_chain_put_by_act(goto_ch);
 
+#ifdef CONFIG_NET_EGRESS
+	if (flags & SKBEDIT_F_QUEUE_MAPPING)
+		net_inc_queue_mapping();
+#endif
+
 	return ret;
 put_chain:
 	if (goto_ch)
@@ -295,8 +304,13 @@  static void tcf_skbedit_cleanup(struct tc_action *a)
 	struct tcf_skbedit_params *params;
 
 	params = rcu_dereference_protected(d->params, 1);
-	if (params)
+	if (params) {
+#ifdef CONFIG_NET_EGRESS
+		if (params->flags & SKBEDIT_F_QUEUE_MAPPING)
+			net_dec_queue_mapping();
+#endif
 		kfree_rcu(params, rcu);
+	}
 }
 
 static int tcf_skbedit_walker(struct net *net, struct sk_buff *skb,