From d62607c3fe45911b2331fac073355a8c914bbde2 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Tue, 7 Jun 2022 21:39:55 -0700 Subject: net: rename reference+tracking helpers Netdev reference helpers have a dev_ prefix for historic reasons. Renaming the old helpers would be too much churn but we can rename the tracking ones which are relatively recent and should be the default for new code. Rename: dev_hold_track() -> netdev_hold() dev_put_track() -> netdev_put() dev_replace_track() -> netdev_ref_replace() Link: https://lore.kernel.org/r/20220608043955.919359-1-kuba@kernel.org Signed-off-by: Jakub Kicinski --- include/linux/netdevice.h | 24 ++++++++++++------------ 1 file changed, 12 insertions(+), 12 deletions(-) (limited to 'include/linux/netdevice.h') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index f615a66c89e9..e2e5088888b1 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -3981,8 +3981,8 @@ static inline void netdev_tracker_free(struct net_device *dev, #endif } -static inline void dev_hold_track(struct net_device *dev, - netdevice_tracker *tracker, gfp_t gfp) +static inline void netdev_hold(struct net_device *dev, + netdevice_tracker *tracker, gfp_t gfp) { if (dev) { __dev_hold(dev); @@ -3990,8 +3990,8 @@ static inline void dev_hold_track(struct net_device *dev, } } -static inline void dev_put_track(struct net_device *dev, - netdevice_tracker *tracker) +static inline void netdev_put(struct net_device *dev, + netdevice_tracker *tracker) { if (dev) { netdev_tracker_free(dev, tracker); @@ -4004,11 +4004,11 @@ static inline void dev_put_track(struct net_device *dev, * @dev: network device * * Hold reference to device to keep it from being freed. - * Try using dev_hold_track() instead. + * Try using netdev_hold() instead. */ static inline void dev_hold(struct net_device *dev) { - dev_hold_track(dev, NULL, GFP_ATOMIC); + netdev_hold(dev, NULL, GFP_ATOMIC); } /** @@ -4016,17 +4016,17 @@ static inline void dev_hold(struct net_device *dev) * @dev: network device * * Release reference to device to allow it to be freed. - * Try using dev_put_track() instead. + * Try using netdev_put() instead. */ static inline void dev_put(struct net_device *dev) { - dev_put_track(dev, NULL); + netdev_put(dev, NULL); } -static inline void dev_replace_track(struct net_device *odev, - struct net_device *ndev, - netdevice_tracker *tracker, - gfp_t gfp) +static inline void netdev_ref_replace(struct net_device *odev, + struct net_device *ndev, + netdevice_tracker *tracker, + gfp_t gfp) { if (odev) netdev_tracker_free(odev, tracker); -- cgit v1.2.3-59-g8ed1b From 9962acefbcb92736c268aafe5f52200948f60f3e Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 8 Jun 2022 08:46:37 -0700 Subject: net: adopt u64_stats_t in struct pcpu_sw_netstats As explained in commit 316580b69d0a ("u64_stats: provide u64_stats_t type") we should use u64_stats_t and related accessors to avoid load/store tearing. Signed-off-by: Eric Dumazet Signed-off-by: Jakub Kicinski --- drivers/net/macsec.c | 8 ++++---- drivers/net/usb/usbnet.c | 8 ++++---- drivers/net/vxlan/vxlan_core.c | 8 ++++---- include/linux/netdevice.h | 16 ++++++++-------- include/net/ip_tunnels.h | 4 ++-- net/bridge/br_netlink.c | 8 ++++---- net/bridge/br_vlan.c | 36 ++++++++++++++++++++---------------- net/core/dev.c | 18 +++++++++--------- net/dsa/slave.c | 8 ++++---- 9 files changed, 59 insertions(+), 55 deletions(-) (limited to 'include/linux/netdevice.h') diff --git a/drivers/net/macsec.c b/drivers/net/macsec.c index 815738c0e067..c881e1bf6f6e 100644 --- a/drivers/net/macsec.c +++ b/drivers/net/macsec.c @@ -523,8 +523,8 @@ static void count_tx(struct net_device *dev, int ret, int len) struct pcpu_sw_netstats *stats = this_cpu_ptr(dev->tstats); u64_stats_update_begin(&stats->syncp); - stats->tx_packets++; - stats->tx_bytes += len; + u64_stats_inc(&stats->tx_packets); + u64_stats_add(&stats->tx_bytes, len); u64_stats_update_end(&stats->syncp); } } @@ -825,8 +825,8 @@ static void count_rx(struct net_device *dev, int len) struct pcpu_sw_netstats *stats = this_cpu_ptr(dev->tstats); u64_stats_update_begin(&stats->syncp); - stats->rx_packets++; - stats->rx_bytes += len; + u64_stats_inc(&stats->rx_packets); + u64_stats_add(&stats->rx_bytes, len); u64_stats_update_end(&stats->syncp); } diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c index 1cb6dab3e2d0..dc79811535c2 100644 --- a/drivers/net/usb/usbnet.c +++ b/drivers/net/usb/usbnet.c @@ -337,8 +337,8 @@ void usbnet_skb_return (struct usbnet *dev, struct sk_buff *skb) skb->protocol = eth_type_trans (skb, dev->net); flags = u64_stats_update_begin_irqsave(&stats64->syncp); - stats64->rx_packets++; - stats64->rx_bytes += skb->len; + u64_stats_inc(&stats64->rx_packets); + u64_stats_add(&stats64->rx_bytes, skb->len); u64_stats_update_end_irqrestore(&stats64->syncp, flags); netif_dbg(dev, rx_status, dev->net, "< rx, len %zu, type 0x%x\n", @@ -1258,8 +1258,8 @@ static void tx_complete (struct urb *urb) unsigned long flags; flags = u64_stats_update_begin_irqsave(&stats64->syncp); - stats64->tx_packets += entry->packets; - stats64->tx_bytes += entry->length; + u64_stats_add(&stats64->tx_packets, entry->packets); + u64_stats_add(&stats64->tx_bytes, entry->length); u64_stats_update_end_irqrestore(&stats64->syncp, flags); } else { dev->net->stats.tx_errors++; diff --git a/drivers/net/vxlan/vxlan_core.c b/drivers/net/vxlan/vxlan_core.c index 265d4a0245e7..8b0710b576c2 100644 --- a/drivers/net/vxlan/vxlan_core.c +++ b/drivers/net/vxlan/vxlan_core.c @@ -2385,15 +2385,15 @@ static void vxlan_encap_bypass(struct sk_buff *skb, struct vxlan_dev *src_vxlan, vxlan_snoop(dev, &loopback, eth_hdr(skb)->h_source, 0, vni); u64_stats_update_begin(&tx_stats->syncp); - tx_stats->tx_packets++; - tx_stats->tx_bytes += len; + u64_stats_inc(&tx_stats->tx_packets); + u64_stats_add(&tx_stats->tx_bytes, len); u64_stats_update_end(&tx_stats->syncp); vxlan_vnifilter_count(src_vxlan, vni, NULL, VXLAN_VNI_STATS_TX, len); if (__netif_rx(skb) == NET_RX_SUCCESS) { u64_stats_update_begin(&rx_stats->syncp); - rx_stats->rx_packets++; - rx_stats->rx_bytes += len; + u64_stats_inc(&rx_stats->rx_packets); + u64_stats_add(&rx_stats->rx_bytes, len); u64_stats_update_end(&rx_stats->syncp); vxlan_vnifilter_count(dst_vxlan, vni, NULL, VXLAN_VNI_STATS_RX, len); diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index e2e5088888b1..89afa4f7747d 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2636,10 +2636,10 @@ struct packet_offload { /* often modified stats are per-CPU, other are shared (netdev->stats) */ struct pcpu_sw_netstats { - u64 rx_packets; - u64 rx_bytes; - u64 tx_packets; - u64 tx_bytes; + u64_stats_t rx_packets; + u64_stats_t rx_bytes; + u64_stats_t tx_packets; + u64_stats_t tx_bytes; struct u64_stats_sync syncp; } __aligned(4 * sizeof(u64)); @@ -2656,8 +2656,8 @@ static inline void dev_sw_netstats_rx_add(struct net_device *dev, unsigned int l struct pcpu_sw_netstats *tstats = this_cpu_ptr(dev->tstats); u64_stats_update_begin(&tstats->syncp); - tstats->rx_bytes += len; - tstats->rx_packets++; + u64_stats_add(&tstats->rx_bytes, len); + u64_stats_inc(&tstats->rx_packets); u64_stats_update_end(&tstats->syncp); } @@ -2668,8 +2668,8 @@ static inline void dev_sw_netstats_tx_add(struct net_device *dev, struct pcpu_sw_netstats *tstats = this_cpu_ptr(dev->tstats); u64_stats_update_begin(&tstats->syncp); - tstats->tx_bytes += len; - tstats->tx_packets += packets; + u64_stats_add(&tstats->tx_bytes, len); + u64_stats_add(&tstats->tx_packets, packets); u64_stats_update_end(&tstats->syncp); } diff --git a/include/net/ip_tunnels.h b/include/net/ip_tunnels.h index c24fa934221d..70cbc4a72669 100644 --- a/include/net/ip_tunnels.h +++ b/include/net/ip_tunnels.h @@ -456,8 +456,8 @@ static inline void iptunnel_xmit_stats(struct net_device *dev, int pkt_len) struct pcpu_sw_netstats *tstats = get_cpu_ptr(dev->tstats); u64_stats_update_begin(&tstats->syncp); - tstats->tx_bytes += pkt_len; - tstats->tx_packets++; + u64_stats_add(&tstats->tx_bytes, pkt_len); + u64_stats_inc(&tstats->tx_packets); u64_stats_update_end(&tstats->syncp); put_cpu_ptr(tstats); } else { diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c index bb01776d2d88..1ef14a099c6b 100644 --- a/net/bridge/br_netlink.c +++ b/net/bridge/br_netlink.c @@ -1770,10 +1770,10 @@ static int br_fill_linkxstats(struct sk_buff *skb, if (v->vid == pvid) vxi.flags |= BRIDGE_VLAN_INFO_PVID; br_vlan_get_stats(v, &stats); - vxi.rx_bytes = stats.rx_bytes; - vxi.rx_packets = stats.rx_packets; - vxi.tx_bytes = stats.tx_bytes; - vxi.tx_packets = stats.tx_packets; + vxi.rx_bytes = u64_stats_read(&stats.rx_bytes); + vxi.rx_packets = u64_stats_read(&stats.rx_packets); + vxi.tx_bytes = u64_stats_read(&stats.tx_bytes); + vxi.tx_packets = u64_stats_read(&stats.tx_packets); if (nla_put(skb, BRIDGE_XSTATS_VLAN, sizeof(vxi), &vxi)) goto nla_put_failure; diff --git a/net/bridge/br_vlan.c b/net/bridge/br_vlan.c index 0f5e75ccac79..6e53dc991409 100644 --- a/net/bridge/br_vlan.c +++ b/net/bridge/br_vlan.c @@ -505,8 +505,8 @@ struct sk_buff *br_handle_vlan(struct net_bridge *br, if (br_opt_get(br, BROPT_VLAN_STATS_ENABLED)) { stats = this_cpu_ptr(v->stats); u64_stats_update_begin(&stats->syncp); - stats->tx_bytes += skb->len; - stats->tx_packets++; + u64_stats_add(&stats->tx_bytes, skb->len); + u64_stats_inc(&stats->tx_packets); u64_stats_update_end(&stats->syncp); } @@ -624,8 +624,8 @@ static bool __allowed_ingress(const struct net_bridge *br, if (br_opt_get(br, BROPT_VLAN_STATS_ENABLED)) { stats = this_cpu_ptr(v->stats); u64_stats_update_begin(&stats->syncp); - stats->rx_bytes += skb->len; - stats->rx_packets++; + u64_stats_add(&stats->rx_bytes, skb->len); + u64_stats_inc(&stats->rx_packets); u64_stats_update_end(&stats->syncp); } @@ -1379,16 +1379,16 @@ void br_vlan_get_stats(const struct net_bridge_vlan *v, cpu_stats = per_cpu_ptr(v->stats, i); do { start = u64_stats_fetch_begin_irq(&cpu_stats->syncp); - rxpackets = cpu_stats->rx_packets; - rxbytes = cpu_stats->rx_bytes; - txbytes = cpu_stats->tx_bytes; - txpackets = cpu_stats->tx_packets; + rxpackets = u64_stats_read(&cpu_stats->rx_packets); + rxbytes = u64_stats_read(&cpu_stats->rx_bytes); + txbytes = u64_stats_read(&cpu_stats->tx_bytes); + txpackets = u64_stats_read(&cpu_stats->tx_packets); } while (u64_stats_fetch_retry_irq(&cpu_stats->syncp, start)); - stats->rx_packets += rxpackets; - stats->rx_bytes += rxbytes; - stats->tx_bytes += txbytes; - stats->tx_packets += txpackets; + u64_stats_add(&stats->rx_packets, rxpackets); + u64_stats_add(&stats->rx_bytes, rxbytes); + u64_stats_add(&stats->tx_bytes, txbytes); + u64_stats_add(&stats->tx_packets, txpackets); } } @@ -1779,14 +1779,18 @@ static bool br_vlan_stats_fill(struct sk_buff *skb, return false; br_vlan_get_stats(v, &stats); - if (nla_put_u64_64bit(skb, BRIDGE_VLANDB_STATS_RX_BYTES, stats.rx_bytes, + if (nla_put_u64_64bit(skb, BRIDGE_VLANDB_STATS_RX_BYTES, + u64_stats_read(&stats.rx_bytes), BRIDGE_VLANDB_STATS_PAD) || nla_put_u64_64bit(skb, BRIDGE_VLANDB_STATS_RX_PACKETS, - stats.rx_packets, BRIDGE_VLANDB_STATS_PAD) || - nla_put_u64_64bit(skb, BRIDGE_VLANDB_STATS_TX_BYTES, stats.tx_bytes, + u64_stats_read(&stats.rx_packets), + BRIDGE_VLANDB_STATS_PAD) || + nla_put_u64_64bit(skb, BRIDGE_VLANDB_STATS_TX_BYTES, + u64_stats_read(&stats.tx_bytes), BRIDGE_VLANDB_STATS_PAD) || nla_put_u64_64bit(skb, BRIDGE_VLANDB_STATS_TX_PACKETS, - stats.tx_packets, BRIDGE_VLANDB_STATS_PAD)) + u64_stats_read(&stats.tx_packets), + BRIDGE_VLANDB_STATS_PAD)) goto out_err; nla_nest_end(skb, nest); diff --git a/net/core/dev.c b/net/core/dev.c index c9d96b85a825..511cf5d3aade 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -10459,23 +10459,23 @@ void dev_fetch_sw_netstats(struct rtnl_link_stats64 *s, int cpu; for_each_possible_cpu(cpu) { + u64 rx_packets, rx_bytes, tx_packets, tx_bytes; const struct pcpu_sw_netstats *stats; - struct pcpu_sw_netstats tmp; unsigned int start; stats = per_cpu_ptr(netstats, cpu); do { start = u64_stats_fetch_begin_irq(&stats->syncp); - tmp.rx_packets = stats->rx_packets; - tmp.rx_bytes = stats->rx_bytes; - tmp.tx_packets = stats->tx_packets; - tmp.tx_bytes = stats->tx_bytes; + rx_packets = u64_stats_read(&stats->rx_packets); + rx_bytes = u64_stats_read(&stats->rx_bytes); + tx_packets = u64_stats_read(&stats->tx_packets); + tx_bytes = u64_stats_read(&stats->tx_bytes); } while (u64_stats_fetch_retry_irq(&stats->syncp, start)); - s->rx_packets += tmp.rx_packets; - s->rx_bytes += tmp.rx_bytes; - s->tx_packets += tmp.tx_packets; - s->tx_bytes += tmp.tx_bytes; + s->rx_packets += rx_packets; + s->rx_bytes += rx_bytes; + s->tx_packets += tx_packets; + s->tx_bytes += tx_bytes; } } EXPORT_SYMBOL_GPL(dev_fetch_sw_netstats); diff --git a/net/dsa/slave.c b/net/dsa/slave.c index 801a5d445833..2e1ac638d135 100644 --- a/net/dsa/slave.c +++ b/net/dsa/slave.c @@ -935,10 +935,10 @@ static void dsa_slave_get_ethtool_stats(struct net_device *dev, s = per_cpu_ptr(dev->tstats, i); do { start = u64_stats_fetch_begin_irq(&s->syncp); - tx_packets = s->tx_packets; - tx_bytes = s->tx_bytes; - rx_packets = s->rx_packets; - rx_bytes = s->rx_bytes; + tx_packets = u64_stats_read(&s->tx_packets); + tx_bytes = u64_stats_read(&s->tx_bytes); + rx_packets = u64_stats_read(&s->rx_packets); + rx_bytes = u64_stats_read(&s->rx_bytes); } while (u64_stats_fetch_retry_irq(&s->syncp, start)); data[0] += tx_packets; data[1] += tx_bytes; -- cgit v1.2.3-59-g8ed1b From af67508ea6cbf0e4ea27f8120056fa2efce127dd Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Tue, 23 Aug 2022 10:46:56 -0700 Subject: net: Fix data-races around sysctl_fb_tunnels_only_for_init_net. While reading sysctl_fb_tunnels_only_for_init_net, it can be changed concurrently. Thus, we need to add READ_ONCE() to its readers. Fixes: 79134e6ce2c9 ("net: do not create fallback tunnels for non-default namespaces") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller --- include/linux/netdevice.h | 11 ++++++++--- 1 file changed, 8 insertions(+), 3 deletions(-) (limited to 'include/linux/netdevice.h') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 1a3cb93c3dcc..6d3a33fd0cdb 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -640,9 +640,14 @@ extern int sysctl_devconf_inherit_init_net; */ static inline bool net_has_fallback_tunnels(const struct net *net) { - return !IS_ENABLED(CONFIG_SYSCTL) || - !sysctl_fb_tunnels_only_for_init_net || - (net == &init_net && sysctl_fb_tunnels_only_for_init_net == 1); +#if IS_ENABLED(CONFIG_SYSCTL) + int fb_tunnels_only_for_init_net = READ_ONCE(sysctl_fb_tunnels_only_for_init_net); + + return !fb_tunnels_only_for_init_net || + (net_eq(net, &init_net) && fb_tunnels_only_for_init_net == 1); +#else + return true; +#endif } static inline int netdev_queue_numa_node_read(const struct netdev_queue *q) -- cgit v1.2.3-59-g8ed1b From a5612ca10d1aa05624ebe72633e0c8c792970833 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Tue, 23 Aug 2022 10:46:57 -0700 Subject: net: Fix data-races around sysctl_devconf_inherit_init_net. While reading sysctl_devconf_inherit_init_net, it can be changed concurrently. Thus, we need to add READ_ONCE() to its readers. Fixes: 856c395cfa63 ("net: introduce a knob to control whether to inherit devconf config") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller --- include/linux/netdevice.h | 9 +++++++++ net/ipv4/devinet.c | 16 ++++++++++------ net/ipv6/addrconf.c | 5 ++--- 3 files changed, 21 insertions(+), 9 deletions(-) (limited to 'include/linux/netdevice.h') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 6d3a33fd0cdb..05d6f3facd5a 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -650,6 +650,15 @@ static inline bool net_has_fallback_tunnels(const struct net *net) #endif } +static inline int net_inherit_devconf(void) +{ +#if IS_ENABLED(CONFIG_SYSCTL) + return READ_ONCE(sysctl_devconf_inherit_init_net); +#else + return 0; +#endif +} + static inline int netdev_queue_numa_node_read(const struct netdev_queue *q) { #if defined(CONFIG_XPS) && defined(CONFIG_NUMA) diff --git a/net/ipv4/devinet.c b/net/ipv4/devinet.c index 92b778e423df..e8b9a9202fec 100644 --- a/net/ipv4/devinet.c +++ b/net/ipv4/devinet.c @@ -2682,23 +2682,27 @@ static __net_init int devinet_init_net(struct net *net) #endif if (!net_eq(net, &init_net)) { - if (IS_ENABLED(CONFIG_SYSCTL) && - sysctl_devconf_inherit_init_net == 3) { + switch (net_inherit_devconf()) { + case 3: /* copy from the current netns */ memcpy(all, current->nsproxy->net_ns->ipv4.devconf_all, sizeof(ipv4_devconf)); memcpy(dflt, current->nsproxy->net_ns->ipv4.devconf_dflt, sizeof(ipv4_devconf_dflt)); - } else if (!IS_ENABLED(CONFIG_SYSCTL) || - sysctl_devconf_inherit_init_net != 2) { - /* inherit == 0 or 1: copy from init_net */ + break; + case 0: + case 1: + /* copy from init_net */ memcpy(all, init_net.ipv4.devconf_all, sizeof(ipv4_devconf)); memcpy(dflt, init_net.ipv4.devconf_dflt, sizeof(ipv4_devconf_dflt)); + break; + case 2: + /* use compiled values */ + break; } - /* else inherit == 2: use compiled values */ } #ifdef CONFIG_SYSCTL diff --git a/net/ipv6/addrconf.c b/net/ipv6/addrconf.c index b624e3d8c5f0..e15f64f22fa8 100644 --- a/net/ipv6/addrconf.c +++ b/net/ipv6/addrconf.c @@ -7162,9 +7162,8 @@ static int __net_init addrconf_init_net(struct net *net) if (!dflt) goto err_alloc_dflt; - if (IS_ENABLED(CONFIG_SYSCTL) && - !net_eq(net, &init_net)) { - switch (sysctl_devconf_inherit_init_net) { + if (!net_eq(net, &init_net)) { + switch (net_inherit_devconf()) { case 1: /* copy from init_net */ memcpy(all, init_net.ipv6.devconf_all, sizeof(ipv6_devconf)); -- cgit v1.2.3-59-g8ed1b