From nobody Thu Nov 28 13:43:49 2024 Received: from mail-pj1-f54.google.com (mail-pj1-f54.google.com [209.85.216.54]) (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 B0D3D1E1330 for ; Tue, 1 Oct 2024 23:53:55 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.216.54 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1727826837; cv=none; b=Lyv+TCImTwpokR5IPjRg0xQ1brWCDLhS3NwXHwCQN3WX/ekgHuRd+Nojs9GnfSCcERJoWHIXXZke0ei2QNFxX3v7ZiZsQCb6mDdEQKLzQSqyngf+bySP1sQWTzs2W5pTE48/tHw4e7K9sAWVR1nP3LUTjKmUJrFTKRYvPGP/kFA= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1727826837; c=relaxed/simple; bh=zto/izfmmguhKhVHdEZdiijUzaBXpzkbp1KqyItqIxo=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=QYOI6s5mFeDBUEKEN35Di48XcBVBk1dAq4YF4kWXEz7CzgWjE5Pg5kLf9iXChmZPPpihGDbqn5qEuKaZWvbQGopuwm4T1suW26QOgEx29vJD7fWb5V/liJvHAiK2HrRngjnPh0gvZNnUu15ASqwbWs67QUIkO7uwK9gbwIhor+s= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=fastly.com; spf=pass smtp.mailfrom=fastly.com; dkim=pass (1024-bit key) header.d=fastly.com header.i=@fastly.com header.b=iE9GiHBY; arc=none smtp.client-ip=209.85.216.54 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=reject dis=none) header.from=fastly.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=fastly.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=fastly.com header.i=@fastly.com header.b="iE9GiHBY" Received: by mail-pj1-f54.google.com with SMTP id 98e67ed59e1d1-2e0be1afa85so4391672a91.1 for ; Tue, 01 Oct 2024 16:53:55 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=fastly.com; s=google; t=1727826835; x=1728431635; darn=vger.kernel.org; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=wlIWZP5hxY7Z7ed2/y3QCNWpfTA3UWwpTpt+lr7sIBg=; b=iE9GiHBYF+WhLpT3RBv1V+mfW8Wgm3Yu9wy/VmhIK74mxvKpN1VWtseoOuV1eLE7/f P5femz9h5P+u4RYDvI8lrmuunCnYT6+XaZC6YfEBjejNc6GRbVuKdaMj4KXtymAmjhGp 2ndF8394YUpVGLHchbKmNl/S4Y5/Ld1PQ/uI8= X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1727826835; x=1728431635; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=wlIWZP5hxY7Z7ed2/y3QCNWpfTA3UWwpTpt+lr7sIBg=; b=Ptrv5ADqLJK7ZPrghfUDg8MVeH/nuaMG1VuI1GPI9LCE7WJLWbm6FLL+dCsCUT0c8Q VwKkjF0VnAb5o4SeTZS6hyaYr8a0gCCUauyA7sa1Sjdlx0vISsot5tVrhaIg1p2sIL4O cgy/ClpPt8EI6Yk9JzMGpqQ3E1BBUWEL4jZmspGZz0Hm+1CJRM7ze9H1PWTMfup1FWIg HtCcYb9d90SacgcCUVrTYVG1Tu1QOdxlw98v4bNBRVJdN2ztdwp7gUrOizZxRtHz1XzO xz6ht/vl3Iqmp1t0mDTIFK/iPzz4e9gsfX7V0UVCOrUX4KBqmYDDC84qzrXYXtxFGSIr Lktg== X-Forwarded-Encrypted: i=1; AJvYcCULovbXObLQYPLN0kEy/zBVZLCQOFOTFZYfDW7jFXnTiSOsQ8CNiJIK4WXEIIG165tBGNWwGC3QZGwQS/E=@vger.kernel.org X-Gm-Message-State: AOJu0Yxyj2Oc2uzFUBtA7Tk7fsBXqwvp55EwGe5wOwXG2AvlTFLftv/V yC0qo7SCsgRrohHf54pZ7naa5NMui5cvXD8Pi32h5M1RUhCqOZGGqUExujNN4iA= X-Google-Smtp-Source: AGHT+IFPu7lvhG4uUYOO2d1rHoWdVGH9+8y4NJPM7S8ebIb+AWvoza2AwcPTUtXv++qlhFapQqZP5w== X-Received: by 2002:a17:90a:ff0d:b0:2e0:875a:f72d with SMTP id 98e67ed59e1d1-2e182c80442mr1777548a91.0.1727826834967; Tue, 01 Oct 2024 16:53:54 -0700 (PDT) Received: from localhost.localdomain ([2620:11a:c019:0:65e:3115:2f58:c5fd]) by smtp.gmail.com with ESMTPSA id 98e67ed59e1d1-2e18f89e973sm213130a91.29.2024.10.01.16.53.53 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 01 Oct 2024 16:53:54 -0700 (PDT) From: Joe Damato To: netdev@vger.kernel.org Cc: mkarsten@uwaterloo.ca, skhawaja@google.com, sdf@fomichev.me, bjorn@rivosinc.com, amritha.nambiar@intel.com, sridhar.samudrala@intel.com, willemdebruijn.kernel@gmail.com, Joe Damato , "David S. Miller" , Eric Dumazet , Jakub Kicinski , Paolo Abeni , Jonathan Corbet , Jiri Pirko , Sebastian Andrzej Siewior , Lorenzo Bianconi , Johannes Berg , linux-doc@vger.kernel.org (open list:DOCUMENTATION), linux-kernel@vger.kernel.org (open list) Subject: [RFC net-next v4 5/9] net: napi: Add napi_config Date: Tue, 1 Oct 2024 23:52:36 +0000 Message-Id: <20241001235302.57609-6-jdamato@fastly.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20241001235302.57609-1-jdamato@fastly.com> References: <20241001235302.57609-1-jdamato@fastly.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Add a persistent NAPI config area for NAPI configuration to the core. Drivers opt-in to setting the persistent config for a NAPI by passing an index when calling netif_napi_add_config. napi_config is allocated in alloc_netdev_mqs, freed in free_netdev (after the NAPIs are deleted). Drivers which call netif_napi_add_config will have persistent per-NAPI settings: NAPI IDs, gro_flush_timeout, and defer_hard_irq settings. Per-NAPI settings are saved in napi_disable and restored in napi_enable. Co-developed-by: Martin Karsten Signed-off-by: Martin Karsten Signed-off-by: Joe Damato --- .../networking/net_cachelines/net_device.rst | 1 + include/linux/netdevice.h | 32 ++++++++ net/core/dev.c | 79 +++++++++++++++++-- net/core/dev.h | 14 ++++ 4 files changed, 119 insertions(+), 7 deletions(-) diff --git a/Documentation/networking/net_cachelines/net_device.rst b/Docum= entation/networking/net_cachelines/net_device.rst index 3d02ae79c850..11d659051f5e 100644 --- a/Documentation/networking/net_cachelines/net_device.rst +++ b/Documentation/networking/net_cachelines/net_device.rst @@ -183,3 +183,4 @@ struct hlist_head page_pools struct dim_irq_moder* irq_moder unsigned_long gro_flush_timeout u32 napi_defer_hard_irqs +struct napi_config* napi_config diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 33897edd16c8..51cff55e7ab8 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -342,6 +342,15 @@ struct gro_list { */ #define GRO_HASH_BUCKETS 8 =20 +/* + * Structure for per-NAPI config + */ +struct napi_config { + u64 gro_flush_timeout; + u32 defer_hard_irqs; + unsigned int napi_id; +}; + /* * Structure for NAPI scheduling similar to tasklet but with weighting */ @@ -379,6 +388,8 @@ struct napi_struct { int irq; unsigned long gro_flush_timeout; u32 defer_hard_irqs; + int index; + struct napi_config *config; }; =20 enum { @@ -2011,6 +2022,9 @@ enum netdev_reg_state { * @dpll_pin: Pointer to the SyncE source pin of a DPLL subsystem, * where the clock is recovered. * + * @napi_config: An array of napi_config structures containing per-NAPI + * settings. + * * FIXME: cleanup struct net_device such that network protocol info * moves out. */ @@ -2400,6 +2414,7 @@ struct net_device { struct dim_irq_moder *irq_moder; unsigned long gro_flush_timeout; u32 napi_defer_hard_irqs; + struct napi_config *napi_config; =20 u8 priv[] ____cacheline_aligned __counted_by(priv_len); @@ -2650,6 +2665,23 @@ netif_napi_add_tx_weight(struct net_device *dev, netif_napi_add_weight(dev, napi, poll, weight); } =20 +/** + * netif_napi_add_config - initialize a NAPI context with persistent config + * @dev: network device + * @napi: NAPI context + * @poll: polling function + * @weight: the poll weight of this NAPI + * @index: the NAPI index + */ +static inline void +netif_napi_add_config(struct net_device *dev, struct napi_struct *napi, + int (*poll)(struct napi_struct *, int), int index) +{ + napi->index =3D index; + napi->config =3D &dev->napi_config[index]; + netif_napi_add_weight(dev, napi, poll, NAPI_POLL_WEIGHT); +} + /** * netif_napi_add_tx() - initialize a NAPI context to be used for Tx only * @dev: network device diff --git a/net/core/dev.c b/net/core/dev.c index 056ed44f766f..9acb6db19200 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -6499,6 +6499,22 @@ EXPORT_SYMBOL(napi_busy_loop); =20 #endif /* CONFIG_NET_RX_BUSY_POLL */ =20 +static void __napi_hash_add_with_id(struct napi_struct *napi, + unsigned int napi_id) +{ + napi->napi_id =3D napi_id; + hlist_add_head_rcu(&napi->napi_hash_node, + &napi_hash[napi->napi_id % HASH_SIZE(napi_hash)]); +} + +static void napi_hash_add_with_id(struct napi_struct *napi, + unsigned int napi_id) +{ + spin_lock(&napi_hash_lock); + __napi_hash_add_with_id(napi, napi_id); + spin_unlock(&napi_hash_lock); +} + static void napi_hash_add(struct napi_struct *napi) { if (test_bit(NAPI_STATE_NO_BUSY_POLL, &napi->state)) @@ -6511,10 +6527,8 @@ static void napi_hash_add(struct napi_struct *napi) if (unlikely(++napi_gen_id < MIN_NAPI_ID)) napi_gen_id =3D MIN_NAPI_ID; } while (napi_by_id(napi_gen_id)); - napi->napi_id =3D napi_gen_id; =20 - hlist_add_head_rcu(&napi->napi_hash_node, - &napi_hash[napi->napi_id % HASH_SIZE(napi_hash)]); + __napi_hash_add_with_id(napi, napi_gen_id); =20 spin_unlock(&napi_hash_lock); } @@ -6637,6 +6651,28 @@ void netif_queue_set_napi(struct net_device *dev, un= signed int queue_index, } EXPORT_SYMBOL(netif_queue_set_napi); =20 +static void napi_restore_config(struct napi_struct *n) +{ + n->defer_hard_irqs =3D n->config->defer_hard_irqs; + n->gro_flush_timeout =3D n->config->gro_flush_timeout; + /* a NAPI ID might be stored in the config, if so use it. if not, use + * napi_hash_add to generate one for us. It will be saved to the config + * in napi_disable. + */ + if (n->config->napi_id) + napi_hash_add_with_id(n, n->config->napi_id); + else + napi_hash_add(n); +} + +static void napi_save_config(struct napi_struct *n) +{ + n->config->defer_hard_irqs =3D n->defer_hard_irqs; + n->config->gro_flush_timeout =3D n->gro_flush_timeout; + n->config->napi_id =3D n->napi_id; + napi_hash_del(n); +} + void netif_napi_add_weight(struct net_device *dev, struct napi_struct *nap= i, int (*poll)(struct napi_struct *, int), int weight) { @@ -6647,8 +6683,6 @@ void netif_napi_add_weight(struct net_device *dev, st= ruct napi_struct *napi, INIT_HLIST_NODE(&napi->napi_hash_node); hrtimer_init(&napi->timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL_PINNED); napi->timer.function =3D napi_watchdog; - napi_set_defer_hard_irqs(napi, READ_ONCE(dev->napi_defer_hard_irqs)); - napi_set_gro_flush_timeout(napi, READ_ONCE(dev->gro_flush_timeout)); init_gro_hash(napi); napi->skb =3D NULL; INIT_LIST_HEAD(&napi->rx_list); @@ -6666,7 +6700,13 @@ void netif_napi_add_weight(struct net_device *dev, s= truct napi_struct *napi, set_bit(NAPI_STATE_SCHED, &napi->state); set_bit(NAPI_STATE_NPSVC, &napi->state); list_add_rcu(&napi->dev_list, &dev->napi_list); - napi_hash_add(napi); + + /* default settings from sysfs are applied to all NAPIs. any per-NAPI + * configuration will be loaded in napi_enable + */ + napi_set_defer_hard_irqs(napi, READ_ONCE(dev->napi_defer_hard_irqs)); + napi_set_gro_flush_timeout(napi, READ_ONCE(dev->gro_flush_timeout)); + napi_get_frags_check(napi); /* Create kthread for this napi if dev->threaded is set. * Clear dev->threaded if kthread creation failed so that @@ -6698,6 +6738,11 @@ void napi_disable(struct napi_struct *n) =20 hrtimer_cancel(&n->timer); =20 + if (n->config) + napi_save_config(n); + else + napi_hash_del(n); + clear_bit(NAPI_STATE_DISABLE, &n->state); } EXPORT_SYMBOL(napi_disable); @@ -6713,6 +6758,11 @@ void napi_enable(struct napi_struct *n) { unsigned long new, val =3D READ_ONCE(n->state); =20 + if (n->config) + napi_restore_config(n); + else + napi_hash_add(n); + do { BUG_ON(!test_bit(NAPI_STATE_SCHED, &val)); =20 @@ -6742,7 +6792,11 @@ void __netif_napi_del(struct napi_struct *napi) if (!test_and_clear_bit(NAPI_STATE_LISTED, &napi->state)) return; =20 - napi_hash_del(napi); + if (napi->config) { + napi->index =3D -1; + napi->config =3D NULL; + } + list_del_rcu(&napi->dev_list); napi_free_frags(napi); =20 @@ -11079,6 +11133,8 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv= , const char *name, unsigned int txqs, unsigned int rxqs) { struct net_device *dev; + size_t napi_config_sz; + unsigned int maxqs; =20 BUG_ON(strlen(name) >=3D sizeof(dev->name)); =20 @@ -11092,6 +11148,8 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv= , const char *name, return NULL; } =20 + maxqs =3D max(txqs, rxqs); + dev =3D kvzalloc(struct_size(dev, priv, sizeof_priv), GFP_KERNEL_ACCOUNT | __GFP_RETRY_MAYFAIL); if (!dev) @@ -11166,6 +11224,11 @@ struct net_device *alloc_netdev_mqs(int sizeof_pri= v, const char *name, if (!dev->ethtool) goto free_all; =20 + napi_config_sz =3D array_size(maxqs, sizeof(*dev->napi_config)); + dev->napi_config =3D kvzalloc(napi_config_sz, GFP_KERNEL_ACCOUNT); + if (!dev->napi_config) + goto free_all; + strscpy(dev->name, name); dev->name_assign_type =3D name_assign_type; dev->group =3D INIT_NETDEV_GROUP; @@ -11227,6 +11290,8 @@ void free_netdev(struct net_device *dev) list_for_each_entry_safe(p, n, &dev->napi_list, dev_list) netif_napi_del(p); =20 + kvfree(dev->napi_config); + ref_tracker_dir_exit(&dev->refcnt_tracker); #ifdef CONFIG_PCPU_DEV_REFCNT free_percpu(dev->pcpu_refcnt); diff --git a/net/core/dev.h b/net/core/dev.h index 26e598aa56c3..7365fa0ffdc7 100644 --- a/net/core/dev.h +++ b/net/core/dev.h @@ -167,11 +167,18 @@ static inline void napi_set_defer_hard_irqs(struct na= pi_struct *n, u32 defer) static inline void netdev_set_defer_hard_irqs(struct net_device *netdev, u32 defer) { + unsigned int count =3D max(netdev->num_rx_queues, + netdev->num_tx_queues); struct napi_struct *napi; + int i; =20 WRITE_ONCE(netdev->napi_defer_hard_irqs, defer); list_for_each_entry(napi, &netdev->napi_list, dev_list) napi_set_defer_hard_irqs(napi, defer); + + if (netdev->napi_config) + for (i =3D 0; i < count; i++) + netdev->napi_config[i].defer_hard_irqs =3D defer; } =20 /** @@ -207,11 +214,18 @@ static inline void napi_set_gro_flush_timeout(struct = napi_struct *n, static inline void netdev_set_gro_flush_timeout(struct net_device *netdev, unsigned long timeout) { + unsigned int count =3D max(netdev->num_rx_queues, + netdev->num_tx_queues); struct napi_struct *napi; + int i; =20 WRITE_ONCE(netdev->gro_flush_timeout, timeout); list_for_each_entry(napi, &netdev->napi_list, dev_list) napi_set_gro_flush_timeout(napi, timeout); + + if (netdev->napi_config) + for (i =3D 0; i < count; i++) + netdev->napi_config[i].gro_flush_timeout =3D timeout; } =20 int rps_cpumask_housekeeping(struct cpumask *mask); --=20 2.25.1