From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail-qt1-f202.google.com (mail-qt1-f202.google.com [209.85.160.202]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 5BECA1F7916 for ; Fri, 6 Mar 2026 11:34:05 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.160.202 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1772796847; cv=none; b=tayEqaioC5Z+6HhQfc9qcMb1KNhxL3oDpgrXTRvEfNpew52/vJysSoJ3RpPi3wh+Vz5vDPKCp01StJA1SqWFUUE3QR5opgxoB7kx5IQzGo+B6VYEdaJtw5FFTtdFuwuIRzlS81UNujN3MM3htDgnP7/OoSWQGjkWK3WqXaktYCo= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1772796847; c=relaxed/simple; bh=0capAzWAeqw/LRABhb2AMbjYe+KV8ujvpye/S3kjFVE=; h=Date:Mime-Version:Message-ID:Subject:From:To:Cc:Content-Type; b=f0vCRYu7ZN0uYseLC2CidIFVlLqHdLigI9IVcXcePr3lYbFjo+4MZNNf8DQLB7K5qpcbO472HABvkFhZwzkBlVQqUWA6ByjFtHBDijD6QZ/FIbj+bzTRBuduoGnwe9LZgb9kaQTiy2RUl7f/e3VkmZMyx7yci8iSUYFoK0BWbVE= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com; spf=pass smtp.mailfrom=flex--edumazet.bounces.google.com; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b=XvMdicNC; arc=none smtp.client-ip=209.85.160.202 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=google.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=flex--edumazet.bounces.google.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="XvMdicNC" Received: by mail-qt1-f202.google.com with SMTP id d75a77b69052e-5069ba5af5bso514881821cf.2 for ; Fri, 06 Mar 2026 03:34:05 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20230601; t=1772796845; x=1773401645; darn=vger.kernel.org; h=cc:to:from:subject:message-id:mime-version:date:from:to:cc:subject :date:message-id:reply-to; bh=FHVLYlMUm/Yi7WUUBmUmUNil+uywKm5OLA05N2fjOsQ=; b=XvMdicNC0EvTtR1gIWT+yyxdkEzEl468Dy8P2MgHN1p0jU/02s3+QnwCFW6DXTg0+Z RVvhABLXP8IsvZaU4RMdyC4YxQj4qzCD8DTWYHKmPJE0ixiukIHwQ3zVgmUjo6Hm1KAV EiH+nr6uiG6isAcd7eBj27mlyvatrX2MnrSUkPliF4QmfZAW/2t3t3yHlLjzf0loZpFG YfuUQsb4hcmuiyuhJdYsI4I2zsX0WQGNDCfQIUrTpGi65/29uDDQ9lF+dsiszVPf3woU KkazI/jrgHBIL2LKMgx+KQrsZaIqX3BiXK+8KEd7Q+Kt9iNDwcSDY+V8+F3pgvRwx1Va L9ZA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1772796845; x=1773401645; h=cc:to:from:subject:message-id:mime-version:date:x-gm-message-state :from:to:cc:subject:date:message-id:reply-to; bh=FHVLYlMUm/Yi7WUUBmUmUNil+uywKm5OLA05N2fjOsQ=; b=U8wT519bbFLmH6yqyYB9R6yAIPqAnyN/1mnTUtBpJGd8Nra++69v90Gl4wTBGrcEOH iOAoA3o7oNulnLKFGiWSh2mb+h33Zr//5DMQ3nayNlLGJzv63Kz0FPVlyFn0txgZAGvL DmCoSKGXXdV2veMSnZov6w+OJrf4+MjYNsatPf/QaayZHReyxqkJyai94KGErcVxdHXl R9sZABOryPIhrg0Nr30aZrQHoJqeEx9kH0LJLEoB1cVaVAEli1QnyYP7g+PRzP1dCp+x VFzJOMrRZekZCkhNQjdbCXcASZ88gPMq/nhT91lUAGYadV7Ehs+IMf8DMvX0wzac9nVt IC2g== X-Forwarded-Encrypted: i=1; AJvYcCU9D3geQj0F+umi6RgHQgG63ArFPw77wt7wF20pKyG5rzbBI53n3KinfQ02PQ68wEuDVc/Z4rI=@vger.kernel.org X-Gm-Message-State: AOJu0YyjkLqxYewJHreQ/oYs/AvLtNvzcLw4GUxCxTf/8K/HhygY59Aj euBc0n1FaPaqdZKQ8/iGuY85bMdBknnOtEPksbkLdM9Op/mFr1TwP0Iguo6FdYuda3M/lTgTgVE a5tZEftUTZwyBbg== X-Received: from qtpp20-n1.prod.google.com ([2002:a05:622a:9d4:10b0:508:f3c1:694]) (user=edumazet job=prod-delivery.src-stubby-dispatcher) by 2002:a05:622a:14cd:b0:4ff:9688:7dca with SMTP id d75a77b69052e-508f470587fmr19484951cf.8.1772796844838; Fri, 06 Mar 2026 03:34:04 -0800 (PST) Date: Fri, 6 Mar 2026 11:34:02 +0000 Precedence: bulk X-Mailing-List: netdev@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Mime-Version: 1.0 X-Mailer: git-send-email 2.53.0.473.g4a7958ca14-goog Message-ID: <20260306113402.435055-1-edumazet@google.com> Subject: [PATCH net-next] net/sched: do not reset queues in graft operations From: Eric Dumazet To: "David S . Miller" , Jakub Kicinski , Paolo Abeni Cc: Simon Horman , Jamal Hadi Salim , Jiri Pirko , netdev@vger.kernel.org, eric.dumazet@gmail.com, Eric Dumazet Content-Type: text/plain; charset="UTF-8" Following typical script is extremely disruptive, because each graft operation calls dev_deactivate() which resets all the queues of the device. QPARAM="limit 100000 flow_limit 1000 buckets 4096" TXQS=64 for ETH in eth1 do tc qd del dev $ETH root 2>/dev/null tc qd add dev $ETH root handle 1: mq for i in `seq 1 $TXQS` do slot=$( printf %x $(( i )) ) tc qd add dev $ETH parent 1:$slot fq $QPARAM done done One can add "ip link set dev $ETH down/up" to reduce the disruption time: QPARAM="limit 100000 flow_limit 1000 buckets 4096" TXQS=64 for ETH in eth1 do ip link set dev $ETH down tc qd del dev $ETH root 2>/dev/null tc qd add dev $ETH root handle 1: mq for i in `seq 1 $TXQS` do slot=$( printf %x $(( i )) ) tc qd add dev $ETH parent 1:$slot fq $QPARAM done ip link set dev $ETH up done Or we can add a @reset_needed flag to dev_deactivate() and dev_deactivate_many(). This flag is set to true at device dismantle or linkwatch_do_dev(), and to false for graft operations. In the future, we might only stop one queue instead of the whole device, ie call dev_deactivate_queue() instead of dev_deactivate(). Signed-off-by: Eric Dumazet --- include/net/sch_generic.h | 4 ++-- net/core/dev.c | 2 +- net/core/link_watch.c | 2 +- net/sched/sch_api.c | 2 +- net/sched/sch_generic.c | 19 +++++++++++-------- net/sched/sch_htb.c | 4 ++-- net/sched/sch_mq.c | 2 +- net/sched/sch_mqprio.c | 2 +- net/sched/sch_taprio.c | 2 +- 9 files changed, 21 insertions(+), 18 deletions(-) diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index c355300893a3921740cf6b919cc58de953cd610c..16beba40914ee5a9454f9eec2ad6e600543c89f2 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -710,8 +710,8 @@ void dev_qdisc_change_real_num_tx(struct net_device *dev, void dev_init_scheduler(struct net_device *dev); void dev_shutdown(struct net_device *dev); void dev_activate(struct net_device *dev); -void dev_deactivate(struct net_device *dev); -void dev_deactivate_many(struct list_head *head); +void dev_deactivate(struct net_device *dev, bool reset_needed); +void dev_deactivate_many(struct list_head *head, bool reset_needed); struct Qdisc *dev_graft_qdisc(struct netdev_queue *dev_queue, struct Qdisc *qdisc); void qdisc_reset(struct Qdisc *qdisc); diff --git a/net/core/dev.c b/net/core/dev.c index 203dc36aaed55e706f5d978ec02c6bf48201e3d4..6fc9350f0be8756ae849d5b0ca9b222055d96ced 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1756,7 +1756,7 @@ static void __dev_close_many(struct list_head *head) smp_mb__after_atomic(); /* Commit netif_running(). */ } - dev_deactivate_many(head); + dev_deactivate_many(head, true); list_for_each_entry(dev, head, close_list) { const struct net_device_ops *ops = dev->netdev_ops; diff --git a/net/core/link_watch.c b/net/core/link_watch.c index 25c455c10a01cf535d6a7d2952d51434fe462690..ff2c1d4538efbc89cd49d8d8477542a6e7bacbad 100644 --- a/net/core/link_watch.c +++ b/net/core/link_watch.c @@ -181,7 +181,7 @@ static void linkwatch_do_dev(struct net_device *dev) if (netif_carrier_ok(dev)) dev_activate(dev); else - dev_deactivate(dev); + dev_deactivate(dev, true); netif_state_change(dev); } diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index cc43e3f7574fae203989f5c28b4934f0720e64c2..c0bab092ea809955b3a5aaa08c169bca55cfaf7f 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -1120,7 +1120,7 @@ static int qdisc_graft(struct net_device *dev, struct Qdisc *parent, } if (dev->flags & IFF_UP) - dev_deactivate(dev); + dev_deactivate(dev, false); qdisc_offload_graft_root(dev, new, old, extack); diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 556e0d8003161d3dfe68e484f6349714cf989c0c..55e17f9e5636f2d9e798889ff59b2f69c9c1709c 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -1374,7 +1374,7 @@ static bool some_qdisc_is_busy(struct net_device *dev) * This function returns only when all outstanding transmissions * have completed, unless all devices are in dismantle phase. */ -void dev_deactivate_many(struct list_head *head) +void dev_deactivate_many(struct list_head *head, bool reset_needed) { bool sync_needed = false; struct net_device *dev; @@ -1393,11 +1393,14 @@ void dev_deactivate_many(struct list_head *head) if (sync_needed) synchronize_net(); - list_for_each_entry(dev, head, close_list) { - netdev_for_each_tx_queue(dev, dev_reset_queue, NULL); + if (reset_needed) { + list_for_each_entry(dev, head, close_list) { + netdev_for_each_tx_queue(dev, dev_reset_queue, NULL); - if (dev_ingress_queue(dev)) - dev_reset_queue(dev, dev_ingress_queue(dev), NULL); + if (dev_ingress_queue(dev)) + dev_reset_queue(dev, dev_ingress_queue(dev), + NULL); + } } /* Wait for outstanding qdisc_run calls. */ @@ -1412,12 +1415,12 @@ void dev_deactivate_many(struct list_head *head) } } -void dev_deactivate(struct net_device *dev) +void dev_deactivate(struct net_device *dev, bool reset_needed) { LIST_HEAD(single); list_add(&dev->close_list, &single); - dev_deactivate_many(&single); + dev_deactivate_many(&single, reset_needed); list_del(&single); } EXPORT_SYMBOL(dev_deactivate); @@ -1473,7 +1476,7 @@ int dev_qdisc_change_tx_queue_len(struct net_device *dev) int ret = 0; if (up) - dev_deactivate(dev); + dev_deactivate(dev, false); for (i = 0; i < dev->num_tx_queues; i++) { ret = qdisc_change_tx_queue_len(dev, &dev->_tx[i]); diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index cf6cd4ccfa2029d9a45e35d6780520290690732d..eb12381795ce1bb0f3b8c5f502e16ad64c4408c8 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1387,7 +1387,7 @@ htb_graft_helper(struct netdev_queue *dev_queue, struct Qdisc *new_q) struct Qdisc *old_q; if (dev->flags & IFF_UP) - dev_deactivate(dev); + dev_deactivate(dev, false); old_q = dev_graft_qdisc(dev_queue, new_q); if (new_q) new_q->flags |= TCQ_F_ONETXQUEUE | TCQ_F_NOPARENT; @@ -1421,7 +1421,7 @@ static void htb_offload_move_qdisc(struct Qdisc *sch, struct htb_class *cl_old, struct Qdisc *qdisc; if (dev->flags & IFF_UP) - dev_deactivate(dev); + dev_deactivate(dev, false); qdisc = dev_graft_qdisc(queue_old, NULL); WARN_ON(qdisc != cl_old->leaf.q); } diff --git a/net/sched/sch_mq.c b/net/sched/sch_mq.c index 0ed199fa18f04a06197116b6ab64ff3647dbace9..a0133a7b9d3b09a0d2a6064234c8fdef60dbf955 100644 --- a/net/sched/sch_mq.c +++ b/net/sched/sch_mq.c @@ -201,7 +201,7 @@ static int mq_graft(struct Qdisc *sch, unsigned long cl, struct Qdisc *new, struct net_device *dev = qdisc_dev(sch); if (dev->flags & IFF_UP) - dev_deactivate(dev); + dev_deactivate(dev, false); *old = dev_graft_qdisc(dev_queue, new); if (new) diff --git a/net/sched/sch_mqprio.c b/net/sched/sch_mqprio.c index b83276409416fabeccc441e5127211632ddcfedb..002add5ce9e0ab04a6260495d1bec02983c2a204 100644 --- a/net/sched/sch_mqprio.c +++ b/net/sched/sch_mqprio.c @@ -469,7 +469,7 @@ static int mqprio_graft(struct Qdisc *sch, unsigned long cl, struct Qdisc *new, return -EINVAL; if (dev->flags & IFF_UP) - dev_deactivate(dev); + dev_deactivate(dev, false); *old = dev_graft_qdisc(dev_queue, new); diff --git a/net/sched/sch_taprio.c b/net/sched/sch_taprio.c index f721c03514f6008ecc59fe4c4ca4a082099dc125..8e375281195061da848fb2bfaf79cf125afccac0 100644 --- a/net/sched/sch_taprio.c +++ b/net/sched/sch_taprio.c @@ -2184,7 +2184,7 @@ static int taprio_graft(struct Qdisc *sch, unsigned long cl, return -EINVAL; if (dev->flags & IFF_UP) - dev_deactivate(dev); + dev_deactivate(dev, false); /* In offload mode, the child Qdisc is directly attached to the netdev * TX queue, and thus, we need to keep its refcount elevated in order -- 2.53.0.473.g4a7958ca14-goog