From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from unimail.uni-dortmund.de (mx1.hrz.uni-dortmund.de [129.217.128.51]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 528E03E556C; Wed, 25 Mar 2026 14:47:46 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=129.217.128.51 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1774450070; cv=none; b=WdjeM/j/5fXfdRv1xRoKtKXFEweOg+4XbDonpyVVa20b1Wsswo3L8B7TOymy8kwVwOnbKS0PALYbaboD5/WXlqmswvNaIjYBLCh0piY4RD2M0aix71II4w8RMTgHECiyB0OI97/6L8wgNUp9g07z5OME/dOz01XNSm1X5TfZDQc= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1774450070; c=relaxed/simple; bh=3vMzfTSY0WKYes7D51AP4NEkC2SMGLTu3ppzV8FUDEs=; h=Message-ID:Date:MIME-Version:Subject:From:To:Cc:References: In-Reply-To:Content-Type; b=LFZ6KNvJVY0yoN7j0yqwEI1LXWT53v4H9Gxi/YnPyRUf4KW3YRWazanskCPc/t2aC9TDOOqLekMrb7gHTvgivcuvE9T0JKUNQ8ENaHf4dWZnyqbNrKblS5QX4KftepMi9gkPkTNuuFf4Nylt5t1bxOmkwr608JDhn+U5KNZjA7Q= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=tu-dortmund.de; spf=pass smtp.mailfrom=tu-dortmund.de; dkim=pass (1024-bit key) header.d=tu-dortmund.de header.i=@tu-dortmund.de header.b=DWi1ClUA; arc=none smtp.client-ip=129.217.128.51 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=tu-dortmund.de Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=tu-dortmund.de Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=tu-dortmund.de header.i=@tu-dortmund.de header.b="DWi1ClUA" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=tu-dortmund.de; s=unimail; t=1774450056; bh=3vMzfTSY0WKYes7D51AP4NEkC2SMGLTu3ppzV8FUDEs=; h=Date:Subject:From:To:Cc:References:In-Reply-To; b=DWi1ClUACRz9BAa7yPQZpoUB7jTfbj+svASA7scdHS/lY7zkJ1+zaw0my/Lt/xuWh NAOCrMFuXgmhlHOAjG3geuhelBVlQ+RPDAF84srKnc39oo8XRkNSwlnMcUJez7299h hx54hUBesw9N481Us3gprmiM5bFEdDE8Cg1eKnt4= Received: from [192.168.96.127] (dhcp-ip-251.ipt.rwth-aachen.de [137.226.82.251]) (authenticated bits=0) by unimail.uni-dortmund.de (8.18.2/8.18.2) with ESMTPSA id 62PElYOM028631 (version=TLSv1.3 cipher=TLS_AES_256_GCM_SHA384 bits=256 verify=NOT); Wed, 25 Mar 2026 15:47:34 +0100 (CET) Message-ID: Date: Wed, 25 Mar 2026 15:47:34 +0100 Precedence: bulk X-Mailing-List: virtualization@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 User-Agent: Mozilla Thunderbird Subject: Re: [PATCH net-next v8 4/4] tun/tap & vhost-net: avoid ptr_ring tail-drop when a qdisc is present From: Simon Schippers To: Jason Wang Cc: willemdebruijn.kernel@gmail.com, andrew+netdev@lunn.ch, davem@davemloft.net, edumazet@google.com, kuba@kernel.org, pabeni@redhat.com, mst@redhat.com, eperezma@redhat.com, leiyang@redhat.com, stephen@networkplumber.org, jon@nutanix.com, tim.gebauer@tu-dortmund.de, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, kvm@vger.kernel.org, virtualization@lists.linux.dev References: <20260312130639.138988-1-simon.schippers@tu-dortmund.de> <20260312130639.138988-5-simon.schippers@tu-dortmund.de> <0908392d-6314-4141-b908-6c9a880ba0a4@tu-dortmund.de> Content-Language: en-US In-Reply-To: <0908392d-6314-4141-b908-6c9a880ba0a4@tu-dortmund.de> Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit On 3/24/26 11:14, Simon Schippers wrote: > On 3/24/26 02:47, Jason Wang wrote: >> On Thu, Mar 12, 2026 at 9:07 PM Simon Schippers >> wrote: >>> >>> This commit prevents tail-drop when a qdisc is present and the ptr_ring >>> becomes full. Once an entry is successfully produced and the ptr_ring >>> reaches capacity, the netdev queue is stopped instead of dropping >>> subsequent packets. >>> >>> If producing an entry fails anyways due to a race, tun_net_xmit returns >>> NETDEV_TX_BUSY, again avoiding a drop. Such races are expected because >>> LLTX is enabled and the transmit path operates without the usual locking. >>> >>> The existing __tun_wake_queue() function wakes the netdev queue. Races >>> between this wakeup and the queue-stop logic could leave the queue >>> stopped indefinitely. To prevent this, a memory barrier is enforced >>> (as discussed in a similar implementation in [1]), followed by a recheck >>> that wakes the queue if space is already available. >>> >>> If no qdisc is present, the previous tail-drop behavior is preserved. >> >> I wonder if we need a dedicated TUN flag to enable this. With this new >> flag, we can even prevent TUN from using noqueue (not sure if it's >> possible or not). >> > > Except of the slight regressions because of this patchset I do not see > a reason for such a flag. > > I have never seen that the driver prevents noqueue. For example you can > set noqueue to your ethernet interface and under load you soon get > > net_crit_ratelimited("Virtual device %s asks to queue packet!\n", > dev->name); > > followed by a -ENETDOWN. And this is not prevented even though it is > clearly not something a user wants. > >>> >>> Benchmarks: >>> The benchmarks show a slight regression in raw transmission performance, >>> though no packets are lost anymore. >>> >>> The previously introduced threshold to only wake after the queue stopped >>> and half of the ring was consumed showed to be a descent choice: >>> Waking the queue whenever a consume made space in the ring strongly >>> degrades performance for tap, while waking only when the ring is empty >>> is too late and also hurts throughput for tap & tap+vhost-net. >>> Other ratios (3/4, 7/8) showed similar results (not shown here), so >>> 1/2 was chosen for the sake of simplicity for both tun/tap and >>> tun/tap+vhost-net. >>> >>> Test setup: >>> AMD Ryzen 5 5600X at 4.3 GHz, 3200 MHz RAM, isolated QEMU threads; >>> Average over 20 runs @ 100,000,000 packets. SRSO and spectre v2 >>> mitigations disabled. >>> >>> Note for tap+vhost-net: >>> XDP drop program active in VM -> ~2.5x faster, slower for tap due to >>> more syscalls (high utilization of entry_SYSRETQ_unsafe_stack in perf) >>> >>> +--------------------------+--------------+----------------+----------+ >>> | 1 thread | Stock | Patched with | diff | >>> | sending | | fq_codel qdisc | | >>> +------------+-------------+--------------+----------------+----------+ >>> | TAP | Transmitted | 1.151 Mpps | 1.139 Mpps | -1.1% | >>> | +-------------+--------------+----------------+----------+ >>> | | Lost/s | 3.606 Mpps | 0 pps | | >>> +------------+-------------+--------------+----------------+----------+ >>> | TAP | Transmitted | 3.948 Mpps | 3.738 Mpps | -5.3% | >>> | +-------------+--------------+----------------+----------+ >>> | +vhost-net | Lost/s | 496.5 Kpps | 0 pps | | >>> +------------+-------------+--------------+----------------+----------+ >>> >>> +--------------------------+--------------+----------------+----------+ >>> | 2 threads | Stock | Patched with | diff | >>> | sending | | fq_codel qdisc | | >>> +------------+-------------+--------------+----------------+----------+ >>> | TAP | Transmitted | 1.133 Mpps | 1.109 Mpps | -2.1% | >>> | +-------------+--------------+----------------+----------+ >>> | | Lost/s | 8.269 Mpps | 0 pps | | >>> +------------+-------------+--------------+----------------+----------+ >>> | TAP | Transmitted | 3.820 Mpps | 3.513 Mpps | -8.0% | >>> | +-------------+--------------+----------------+----------+ >>> | +vhost-net | Lost/s | 4.961 Mpps | 0 pps | | >>> +------------+-------------+--------------+----------------+----------+ >>> >>> [1] Link: https://lore.kernel.org/all/20250424085358.75d817ae@kernel.org/ >>> >>> Co-developed-by: Tim Gebauer >>> Signed-off-by: Tim Gebauer >>> Signed-off-by: Simon Schippers >>> --- >>> drivers/net/tun.c | 30 ++++++++++++++++++++++++++++-- >>> 1 file changed, 28 insertions(+), 2 deletions(-) >>> >>> diff --git a/drivers/net/tun.c b/drivers/net/tun.c >>> index b86582cc6cb6..9b7daec69acd 100644 >>> --- a/drivers/net/tun.c >>> +++ b/drivers/net/tun.c >>> @@ -1011,6 +1011,8 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev) >>> struct netdev_queue *queue; >>> struct tun_file *tfile; >>> int len = skb->len; >>> + bool qdisc_present; >>> + int ret; >>> >>> rcu_read_lock(); >>> tfile = rcu_dereference(tun->tfiles[txq]); >>> @@ -1063,13 +1065,37 @@ static netdev_tx_t tun_net_xmit(struct sk_buff *skb, struct net_device *dev) >>> >>> nf_reset_ct(skb); >>> >>> - if (ptr_ring_produce(&tfile->tx_ring, skb)) { >>> + queue = netdev_get_tx_queue(dev, txq); >>> + qdisc_present = !qdisc_txq_has_no_queue(queue); >>> + >>> + spin_lock(&tfile->tx_ring.producer_lock); >>> + ret = __ptr_ring_produce(&tfile->tx_ring, skb); >>> + if (__ptr_ring_produce_peek(&tfile->tx_ring) && qdisc_present) { >> >> So, it's possible that the administrator is switching between noqueue >> and another qdisc. So ptr_ring_produce() can fail here, do we need to >> check that or not? >> > > Do you mean that? My thoughts: > > Switching from noqueue to some qdisc can cause a > > net_crit_ratelimited("Virtual device %s asks to queue packet!\n", > dev->name); > > followed by a return of -ENETDOWN in __dev_queue_xmit(). > This is because tun_net_xmit detects some qdisc with > > qdisc_present = !qdisc_txq_has_no_queue(queue); > > and returns NETDEV_TX_BUSY even though __dev_queue_xmit() did still > detect noqueue. > > I am not sure how to solve this/if this has to be solved. > I do not see a proper way to avoid parallel execution of ndo_start_xmit > and a qdisc change (dev_graft_qdisc only takes qdisc_skb_head lock). > > And from my understanding the veth implementation faces the same issue. How about rechecking if a qdisc is connected? This would avoid -ENETDOWN. diff --git a/net/core/dev.c b/net/core/dev.c index f48dc299e4b2..2731a1a70732 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4845,10 +4845,17 @@ int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev) if (is_list) rc = NETDEV_TX_OK; } + bool qdisc_present = !qdisc_txq_has_no_queue(txq); HARD_TX_UNLOCK(dev, txq); if (!skb) /* xmit completed */ goto out; + /* Maybe a qdisc was connected in the meantime */ + if (qdisc_present) { + kfree_skb(skb); + goto out; + } + net_crit_ratelimited("Virtual device %s asks to queue packet!\n", dev->name); /* NETDEV_TX_BUSY or queue was stopped */ > > > Switching from some qdisc to noqueue is no problem I think. > >>> + netif_tx_stop_queue(queue); >>> + /* Avoid races with queue wake-ups in __tun_wake_queue by >>> + * waking if space is available in a re-check. >>> + * The barrier makes sure that the stop is visible before >>> + * we re-check. >>> + */ >>> + smp_mb__after_atomic(); >> >> Let's document which barrier is paired with this. >> > > I am basically copying the (old) logic of veth [1] proposed by > Jakub Kicinski. I must admit I am not 100% sure what it pairs with. > > [1] Link: https://lore.kernel.org/all/20250424085358.75d817ae@kernel.org/ > >>> + if (!__ptr_ring_produce_peek(&tfile->tx_ring)) >>> + netif_tx_wake_queue(queue); >>> + } >>> + spin_unlock(&tfile->tx_ring.producer_lock); >>> + >>> + if (ret) { >>> + /* If a qdisc is attached to our virtual device, >>> + * returning NETDEV_TX_BUSY is allowed. >>> + */ >>> + if (qdisc_present) { >>> + rcu_read_unlock(); >>> + return NETDEV_TX_BUSY; >>> + } >>> drop_reason = SKB_DROP_REASON_FULL_RING; >>> goto drop; >>> } >>> >>> /* dev->lltx requires to do our own update of trans_start */ >>> - queue = netdev_get_tx_queue(dev, txq); >>> txq_trans_cond_update(queue); >>> >>> /* Notify and wake up reader process */ >>> -- >>> 2.43.0 >>> >> >> Thanks >>