From cbda10fa97d72c7a1923be4426171aa90e8c6dab Mon Sep 17 00:00:00 2001 From: Vlad Dogaru Date: Thu, 13 Jan 2011 23:38:30 +0000 Subject: net_device: add support for network device groups Net devices can now be grouped, enabling simpler manipulation from userspace. This patch adds a group field to the net_device structure, as well as rtnetlink support to query and modify it. Signed-off-by: Vlad Dogaru Acked-by: Jamal Hadi Salim Signed-off-by: David S. Miller --- net/core/dev.c | 12 ++++++++++++ 1 file changed, 12 insertions(+) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 7741507429f..2b85d4ae981 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4571,6 +4571,17 @@ int dev_set_mtu(struct net_device *dev, int new_mtu) } EXPORT_SYMBOL(dev_set_mtu); +/** + * dev_set_group - Change group this device belongs to + * @dev: device + * @new_group: group this device should belong to + */ +void dev_set_group(struct net_device *dev, int new_group) +{ + dev->group = new_group; +} +EXPORT_SYMBOL(dev_set_group); + /** * dev_set_mac_address - Change Media Access Control Address * @dev: device @@ -5678,6 +5689,7 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name, dev->priv_flags = IFF_XMIT_DST_RELEASE; setup(dev); strcpy(dev->name, name); + dev->group = INIT_NETDEV_GROUP; return dev; free_pcpu: -- cgit v1.2.3-70-g09d2 From 4f57c087de9b46182545676d2c594120a20f2e58 Mon Sep 17 00:00:00 2001 From: John Fastabend Date: Mon, 17 Jan 2011 08:06:04 +0000 Subject: net: implement mechanism for HW based QOS This patch provides a mechanism for lower layer devices to steer traffic using skb->priority to tx queues. This allows for hardware based QOS schemes to use the default qdisc without incurring the penalties related to global state and the qdisc lock. While reliably receiving skbs on the correct tx ring to avoid head of line blocking resulting from shuffling in the LLD. Finally, all the goodness from txq caching and xps/rps can still be leveraged. Many drivers and hardware exist with the ability to implement QOS schemes in the hardware but currently these drivers tend to rely on firmware to reroute specific traffic, a driver specific select_queue or the queue_mapping action in the qdisc. By using select_queue for this drivers need to be updated for each and every traffic type and we lose the goodness of much of the upstream work. Firmware solutions are inherently inflexible. And finally if admins are expected to build a qdisc and filter rules to steer traffic this requires knowledge of how the hardware is currently configured. The number of tx queues and the queue offsets may change depending on resources. Also this approach incurs all the overhead of a qdisc with filters. With the mechanism in this patch users can set skb priority using expected methods ie setsockopt() or the stack can set the priority directly. Then the skb will be steered to the correct tx queues aligned with hardware QOS traffic classes. In the normal case with single traffic class and all queues in this class everything works as is until the LLD enables multiple tcs. To steer the skb we mask out the lower 4 bits of the priority and allow the hardware to configure upto 15 distinct classes of traffic. This is expected to be sufficient for most applications at any rate it is more then the 8021Q spec designates and is equal to the number of prio bands currently implemented in the default qdisc. This in conjunction with a userspace application such as lldpad can be used to implement 8021Q transmission selection algorithms one of these algorithms being the extended transmission selection algorithm currently being used for DCB. Signed-off-by: John Fastabend Signed-off-by: David S. Miller --- include/linux/netdevice.h | 68 +++++++++++++++++++++++++++++++++++++++++++++++ net/core/dev.c | 55 +++++++++++++++++++++++++++++++++++++- 2 files changed, 122 insertions(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 68a4627b74f..371fa8839d5 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -646,6 +646,14 @@ struct xps_dev_maps { (nr_cpu_ids * sizeof(struct xps_map *))) #endif /* CONFIG_XPS */ +#define TC_MAX_QUEUE 16 +#define TC_BITMASK 15 +/* HW offloaded queuing disciplines txq count and offset maps */ +struct netdev_tc_txq { + u16 count; + u16 offset; +}; + /* * This structure defines the management hooks for network devices. * The following hooks can be defined; unless noted otherwise, they are @@ -756,6 +764,11 @@ struct xps_dev_maps { * int (*ndo_set_vf_port)(struct net_device *dev, int vf, * struct nlattr *port[]); * int (*ndo_get_vf_port)(struct net_device *dev, int vf, struct sk_buff *skb); + * int (*ndo_setup_tc)(struct net_device *dev, u8 tc) + * Called to setup 'tc' number of traffic classes in the net device. This + * is always called from the stack with the rtnl lock held and netif tx + * queues stopped. This allows the netdevice to perform queue management + * safely. */ #define HAVE_NET_DEVICE_OPS struct net_device_ops { @@ -814,6 +827,7 @@ struct net_device_ops { struct nlattr *port[]); int (*ndo_get_vf_port)(struct net_device *dev, int vf, struct sk_buff *skb); + int (*ndo_setup_tc)(struct net_device *dev, u8 tc); #if defined(CONFIG_FCOE) || defined(CONFIG_FCOE_MODULE) int (*ndo_fcoe_enable)(struct net_device *dev); int (*ndo_fcoe_disable)(struct net_device *dev); @@ -1146,6 +1160,9 @@ struct net_device { /* Data Center Bridging netlink ops */ const struct dcbnl_rtnl_ops *dcbnl_ops; #endif + u8 num_tc; + struct netdev_tc_txq tc_to_txq[TC_MAX_QUEUE]; + u8 prio_tc_map[TC_BITMASK + 1]; #if defined(CONFIG_FCOE) || defined(CONFIG_FCOE_MODULE) /* max exchange id for FCoE LRO by ddp */ @@ -1164,6 +1181,57 @@ struct net_device { #define NETDEV_ALIGN 32 +static inline +int netdev_get_prio_tc_map(const struct net_device *dev, u32 prio) +{ + return dev->prio_tc_map[prio & TC_BITMASK]; +} + +static inline +int netdev_set_prio_tc_map(struct net_device *dev, u8 prio, u8 tc) +{ + if (tc >= dev->num_tc) + return -EINVAL; + + dev->prio_tc_map[prio & TC_BITMASK] = tc & TC_BITMASK; + return 0; +} + +static inline +void netdev_reset_tc(struct net_device *dev) +{ + dev->num_tc = 0; + memset(dev->tc_to_txq, 0, sizeof(dev->tc_to_txq)); + memset(dev->prio_tc_map, 0, sizeof(dev->prio_tc_map)); +} + +static inline +int netdev_set_tc_queue(struct net_device *dev, u8 tc, u16 count, u16 offset) +{ + if (tc >= dev->num_tc) + return -EINVAL; + + dev->tc_to_txq[tc].count = count; + dev->tc_to_txq[tc].offset = offset; + return 0; +} + +static inline +int netdev_set_num_tc(struct net_device *dev, u8 num_tc) +{ + if (num_tc > TC_MAX_QUEUE) + return -EINVAL; + + dev->num_tc = num_tc; + return 0; +} + +static inline +int netdev_get_num_tc(struct net_device *dev) +{ + return dev->num_tc; +} + static inline struct netdev_queue *netdev_get_tx_queue(const struct net_device *dev, unsigned int index) diff --git a/net/core/dev.c b/net/core/dev.c index 2b85d4ae981..8b1d886ed23 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1593,6 +1593,48 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) rcu_read_unlock(); } +/* netif_setup_tc - Handle tc mappings on real_num_tx_queues change + * @dev: Network device + * @txq: number of queues available + * + * If real_num_tx_queues is changed the tc mappings may no longer be + * valid. To resolve this verify the tc mapping remains valid and if + * not NULL the mapping. With no priorities mapping to this + * offset/count pair it will no longer be used. In the worst case TC0 + * is invalid nothing can be done so disable priority mappings. If is + * expected that drivers will fix this mapping if they can before + * calling netif_set_real_num_tx_queues. + */ +void netif_setup_tc(struct net_device *dev, unsigned int txq) +{ + int i; + struct netdev_tc_txq *tc = &dev->tc_to_txq[0]; + + /* If TC0 is invalidated disable TC mapping */ + if (tc->offset + tc->count > txq) { + pr_warning("Number of in use tx queues changed " + "invalidating tc mappings. Priority " + "traffic classification disabled!\n"); + dev->num_tc = 0; + return; + } + + /* Invalidated prio to tc mappings set to TC0 */ + for (i = 1; i < TC_BITMASK + 1; i++) { + int q = netdev_get_prio_tc_map(dev, i); + + tc = &dev->tc_to_txq[q]; + if (tc->offset + tc->count > txq) { + pr_warning("Number of in use tx queues " + "changed. Priority %i to tc " + "mapping %i is no longer valid " + "setting map to 0\n", + i, q); + netdev_set_prio_tc_map(dev, i, 0); + } + } +} + /* * Routine to help set real_num_tx_queues. To avoid skbs mapped to queues * greater then real_num_tx_queues stale skbs on the qdisc must be flushed. @@ -1612,6 +1654,9 @@ int netif_set_real_num_tx_queues(struct net_device *dev, unsigned int txq) if (rc) return rc; + if (dev->num_tc) + netif_setup_tc(dev, txq); + if (txq < dev->real_num_tx_queues) qdisc_reset_all_tx_gt(dev, txq); } @@ -2161,6 +2206,8 @@ u16 __skb_tx_hash(const struct net_device *dev, const struct sk_buff *skb, unsigned int num_tx_queues) { u32 hash; + u16 qoffset = 0; + u16 qcount = num_tx_queues; if (skb_rx_queue_recorded(skb)) { hash = skb_get_rx_queue(skb); @@ -2169,13 +2216,19 @@ u16 __skb_tx_hash(const struct net_device *dev, const struct sk_buff *skb, return hash; } + if (dev->num_tc) { + u8 tc = netdev_get_prio_tc_map(dev, skb->priority); + qoffset = dev->tc_to_txq[tc].offset; + qcount = dev->tc_to_txq[tc].count; + } + if (skb->sk && skb->sk->sk_hash) hash = skb->sk->sk_hash; else hash = (__force u16) skb->protocol ^ skb->rxhash; hash = jhash_1word(hash, hashrnd); - return (u16) (((u64) hash * num_tx_queues) >> 32); + return (u16) (((u64) hash * qcount) >> 32) + qoffset; } EXPORT_SYMBOL(__skb_tx_hash); -- cgit v1.2.3-70-g09d2 From 3fbd8758b027995b677046dae46f9b41ea88c88f Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 19 Jan 2011 21:23:22 +0000 Subject: net: dev_close_many() is static Signed-off-by: Eric Dumazet CC: Octavian Purdila Reviewed-by: Octavian Purdila Signed-off-by: David S. Miller --- net/core/dev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 8b1d886ed23..a4ccd47f319 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1285,7 +1285,7 @@ static int __dev_close(struct net_device *dev) return __dev_close_many(&single); } -int dev_close_many(struct list_head *head) +static int dev_close_many(struct list_head *head) { struct net_device *dev, *tmp; LIST_HEAD(tmp_list); -- cgit v1.2.3-70-g09d2 From a2da570d62fcb9e8816f6920e1ec02c706b289fa Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 20 Jan 2011 03:48:19 +0000 Subject: net_sched: RCU conversion of stab This patch converts stab qdisc management to RCU, so that we can perform the qdisc_calculate_pkt_len() call before getting qdisc lock. This shortens the lock's held time in __dev_xmit_skb(). This permits more qdiscs to get TCQ_F_CAN_BYPASS status, avoiding lot of cache misses and so reducing latencies. Signed-off-by: Eric Dumazet CC: Patrick McHardy CC: Jesper Dangaard Brouer CC: Jarek Poplawski CC: Jamal Hadi Salim CC: Stephen Hemminger Signed-off-by: David S. Miller --- include/net/sch_generic.h | 21 +++++++++++++++------ net/core/dev.c | 8 +++++--- net/sched/sch_api.c | 26 +++++++++++++++++--------- net/sched/sch_generic.c | 2 +- 4 files changed, 38 insertions(+), 19 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/net/sch_generic.h b/include/net/sch_generic.h index f6345f55041..d531baa2506 100644 --- a/include/net/sch_generic.h +++ b/include/net/sch_generic.h @@ -36,6 +36,7 @@ enum qdisc___state_t { }; struct qdisc_size_table { + struct rcu_head rcu; struct list_head list; struct tc_sizespec szopts; int refcnt; @@ -53,7 +54,7 @@ struct Qdisc { #define TCQ_F_WARN_NONWC (1 << 16) int padded; struct Qdisc_ops *ops; - struct qdisc_size_table *stab; + struct qdisc_size_table __rcu *stab; struct list_head list; u32 handle; u32 parent; @@ -349,8 +350,8 @@ extern struct Qdisc *qdisc_alloc(struct netdev_queue *dev_queue, struct Qdisc_ops *ops); extern struct Qdisc *qdisc_create_dflt(struct netdev_queue *dev_queue, struct Qdisc_ops *ops, u32 parentid); -extern void qdisc_calculate_pkt_len(struct sk_buff *skb, - struct qdisc_size_table *stab); +extern void __qdisc_calculate_pkt_len(struct sk_buff *skb, + const struct qdisc_size_table *stab); extern void tcf_destroy(struct tcf_proto *tp); extern void tcf_destroy_chain(struct tcf_proto **fl); @@ -429,12 +430,20 @@ enum net_xmit_qdisc_t { #define net_xmit_drop_count(e) (1) #endif -static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch) +static inline void qdisc_calculate_pkt_len(struct sk_buff *skb, + const struct Qdisc *sch) { #ifdef CONFIG_NET_SCHED - if (sch->stab) - qdisc_calculate_pkt_len(skb, sch->stab); + struct qdisc_size_table *stab = rcu_dereference_bh(sch->stab); + + if (stab) + __qdisc_calculate_pkt_len(skb, stab); #endif +} + +static inline int qdisc_enqueue(struct sk_buff *skb, struct Qdisc *sch) +{ + qdisc_calculate_pkt_len(skb, sch); return sch->enqueue(skb, sch); } diff --git a/net/core/dev.c b/net/core/dev.c index a4ccd47f319..2730352d2cc 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2325,15 +2325,18 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q, struct netdev_queue *txq) { spinlock_t *root_lock = qdisc_lock(q); - bool contended = qdisc_is_running(q); + bool contended; int rc; + qdisc_skb_cb(skb)->pkt_len = skb->len; + qdisc_calculate_pkt_len(skb, q); /* * Heuristic to force contended enqueues to serialize on a * separate lock before trying to get qdisc main lock. * This permits __QDISC_STATE_RUNNING owner to get the lock more often * and dequeue packets faster. */ + contended = qdisc_is_running(q); if (unlikely(contended)) spin_lock(&q->busylock); @@ -2351,7 +2354,6 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q, if (!(dev->priv_flags & IFF_XMIT_DST_RELEASE)) skb_dst_force(skb); - qdisc_skb_cb(skb)->pkt_len = skb->len; qdisc_bstats_update(q, skb); if (sch_direct_xmit(skb, q, dev, txq, root_lock)) { @@ -2366,7 +2368,7 @@ static inline int __dev_xmit_skb(struct sk_buff *skb, struct Qdisc *q, rc = NET_XMIT_SUCCESS; } else { skb_dst_force(skb); - rc = qdisc_enqueue_root(skb, q); + rc = q->enqueue(skb, q) & NET_XMIT_MASK; if (qdisc_run_begin(q)) { if (unlikely(contended)) { spin_unlock(&q->busylock); diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c index 374fcbef80e..15074157940 100644 --- a/net/sched/sch_api.c +++ b/net/sched/sch_api.c @@ -398,6 +398,11 @@ static struct qdisc_size_table *qdisc_get_stab(struct nlattr *opt) return stab; } +static void stab_kfree_rcu(struct rcu_head *head) +{ + kfree(container_of(head, struct qdisc_size_table, rcu)); +} + void qdisc_put_stab(struct qdisc_size_table *tab) { if (!tab) @@ -407,7 +412,7 @@ void qdisc_put_stab(struct qdisc_size_table *tab) if (--tab->refcnt == 0) { list_del(&tab->list); - kfree(tab); + call_rcu_bh(&tab->rcu, stab_kfree_rcu); } spin_unlock(&qdisc_stab_lock); @@ -430,7 +435,7 @@ nla_put_failure: return -1; } -void qdisc_calculate_pkt_len(struct sk_buff *skb, struct qdisc_size_table *stab) +void __qdisc_calculate_pkt_len(struct sk_buff *skb, const struct qdisc_size_table *stab) { int pkt_len, slot; @@ -456,7 +461,7 @@ out: pkt_len = 1; qdisc_skb_cb(skb)->pkt_len = pkt_len; } -EXPORT_SYMBOL(qdisc_calculate_pkt_len); +EXPORT_SYMBOL(__qdisc_calculate_pkt_len); void qdisc_warn_nonwc(char *txt, struct Qdisc *qdisc) { @@ -835,7 +840,7 @@ qdisc_create(struct net_device *dev, struct netdev_queue *dev_queue, err = PTR_ERR(stab); goto err_out4; } - sch->stab = stab; + rcu_assign_pointer(sch->stab, stab); } if (tca[TCA_RATE]) { spinlock_t *root_lock; @@ -875,7 +880,7 @@ err_out4: * Any broken qdiscs that would require a ops->reset() here? * The qdisc was never in action so it shouldn't be necessary. */ - qdisc_put_stab(sch->stab); + qdisc_put_stab(rtnl_dereference(sch->stab)); if (ops->destroy) ops->destroy(sch); goto err_out3; @@ -883,7 +888,7 @@ err_out4: static int qdisc_change(struct Qdisc *sch, struct nlattr **tca) { - struct qdisc_size_table *stab = NULL; + struct qdisc_size_table *ostab, *stab = NULL; int err = 0; if (tca[TCA_OPTIONS]) { @@ -900,8 +905,9 @@ static int qdisc_change(struct Qdisc *sch, struct nlattr **tca) return PTR_ERR(stab); } - qdisc_put_stab(sch->stab); - sch->stab = stab; + ostab = rtnl_dereference(sch->stab); + rcu_assign_pointer(sch->stab, stab); + qdisc_put_stab(ostab); if (tca[TCA_RATE]) { /* NB: ignores errors from replace_estimator @@ -1180,6 +1186,7 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid, struct nlmsghdr *nlh; unsigned char *b = skb_tail_pointer(skb); struct gnet_dump d; + struct qdisc_size_table *stab; nlh = NLMSG_NEW(skb, pid, seq, event, sizeof(*tcm), flags); tcm = NLMSG_DATA(nlh); @@ -1195,7 +1202,8 @@ static int tc_fill_qdisc(struct sk_buff *skb, struct Qdisc *q, u32 clid, goto nla_put_failure; q->qstats.qlen = q->q.qlen; - if (q->stab && qdisc_dump_stab(skb, q->stab) < 0) + stab = rtnl_dereference(q->stab); + if (stab && qdisc_dump_stab(skb, stab) < 0) goto nla_put_failure; if (gnet_stats_start_copy_compat(skb, TCA_STATS2, TCA_STATS, TCA_XSTATS, diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c index 2f1cb62130d..cc17e794c41 100644 --- a/net/sched/sch_generic.c +++ b/net/sched/sch_generic.c @@ -632,7 +632,7 @@ void qdisc_destroy(struct Qdisc *qdisc) #ifdef CONFIG_NET_SCHED qdisc_list_del(qdisc); - qdisc_put_stab(qdisc->stab); + qdisc_put_stab(rtnl_dereference(qdisc->stab)); #endif gen_kill_estimator(&qdisc->bstats, &qdisc->rate_est); if (ops->reset) -- cgit v1.2.3-70-g09d2 From bb134d2298b49f50cf6d9388410fba96272905dc Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Thu, 20 Jan 2011 19:18:08 +0000 Subject: net: netif_setup_tc() is static Signed-off-by: Eric Dumazet Acked-by: John Fastabend Signed-off-by: David S. Miller --- net/core/dev.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 2730352d2cc..47d3d78d541 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1605,7 +1605,7 @@ static void dev_queue_xmit_nit(struct sk_buff *skb, struct net_device *dev) * expected that drivers will fix this mapping if they can before * calling netif_set_real_num_tx_queues. */ -void netif_setup_tc(struct net_device *dev, unsigned int txq) +static void netif_setup_tc(struct net_device *dev, unsigned int txq) { int i; struct netdev_tc_txq *tc = &dev->tc_to_txq[0]; -- cgit v1.2.3-70-g09d2 From c445477d74ab3779d1386ab797fbb9b628eb9f64 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Wed, 19 Jan 2011 11:03:53 +0000 Subject: net: RPS: Enable hardware acceleration of RFS Allow drivers for multiqueue hardware with flow filter tables to accelerate RFS. The driver must: 1. Set net_device::rx_cpu_rmap to a cpu_rmap of the RX completion IRQs (in queue order). This will provide a mapping from CPUs to the queues for which completions are handled nearest to them. 2. Implement net_device_ops::ndo_rx_flow_steer. This operation adds or replaces a filter steering the given flow to the given RX queue, if possible. 3. Periodically remove filters for which rps_may_expire_flow() returns true. Signed-off-by: Ben Hutchings Signed-off-by: David S. Miller --- include/linux/netdevice.h | 33 ++++++++++++++-- net/Kconfig | 6 +++ net/core/dev.c | 97 ++++++++++++++++++++++++++++++++++++++++++++--- 3 files changed, 127 insertions(+), 9 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 371fa8839d5..a335f202269 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -554,14 +554,16 @@ struct rps_map { #define RPS_MAP_SIZE(_num) (sizeof(struct rps_map) + (_num * sizeof(u16))) /* - * The rps_dev_flow structure contains the mapping of a flow to a CPU and the - * tail pointer for that CPU's input queue at the time of last enqueue. + * The rps_dev_flow structure contains the mapping of a flow to a CPU, the + * tail pointer for that CPU's input queue at the time of last enqueue, and + * a hardware filter index. */ struct rps_dev_flow { u16 cpu; - u16 fill; + u16 filter; unsigned int last_qtail; }; +#define RPS_NO_FILTER 0xffff /* * The rps_dev_flow_table structure contains a table of flow mappings. @@ -611,6 +613,11 @@ static inline void rps_reset_sock_flow(struct rps_sock_flow_table *table, extern struct rps_sock_flow_table __rcu *rps_sock_flow_table; +#ifdef CONFIG_RFS_ACCEL +extern bool rps_may_expire_flow(struct net_device *dev, u16 rxq_index, + u32 flow_id, u16 filter_id); +#endif + /* This structure contains an instance of an RX queue. */ struct netdev_rx_queue { struct rps_map __rcu *rps_map; @@ -769,6 +776,13 @@ struct netdev_tc_txq { * is always called from the stack with the rtnl lock held and netif tx * queues stopped. This allows the netdevice to perform queue management * safely. + * + * RFS acceleration. + * int (*ndo_rx_flow_steer)(struct net_device *dev, const struct sk_buff *skb, + * u16 rxq_index, u32 flow_id); + * Set hardware filter for RFS. rxq_index is the target queue index; + * flow_id is a flow ID to be passed to rps_may_expire_flow() later. + * Return the filter ID on success, or a negative error code. */ #define HAVE_NET_DEVICE_OPS struct net_device_ops { @@ -842,6 +856,12 @@ struct net_device_ops { int (*ndo_fcoe_get_wwn)(struct net_device *dev, u64 *wwn, int type); #endif +#ifdef CONFIG_RFS_ACCEL + int (*ndo_rx_flow_steer)(struct net_device *dev, + const struct sk_buff *skb, + u16 rxq_index, + u32 flow_id); +#endif }; /* @@ -1056,6 +1076,13 @@ struct net_device { /* Number of RX queues currently active in device */ unsigned int real_num_rx_queues; + +#ifdef CONFIG_RFS_ACCEL + /* CPU reverse-mapping for RX completion interrupts, indexed + * by RX queue number. Assigned by driver. This must only be + * set if the ndo_rx_flow_steer operation is defined. */ + struct cpu_rmap *rx_cpu_rmap; +#endif #endif rx_handler_func_t __rcu *rx_handler; diff --git a/net/Kconfig b/net/Kconfig index 72840626284..79cabf1ee68 100644 --- a/net/Kconfig +++ b/net/Kconfig @@ -221,6 +221,12 @@ config RPS depends on SMP && SYSFS && USE_GENERIC_SMP_HELPERS default y +config RFS_ACCEL + boolean + depends on RPS && GENERIC_HARDIRQS + select CPU_RMAP + default y + config XPS boolean depends on SMP && SYSFS && USE_GENERIC_SMP_HELPERS diff --git a/net/core/dev.c b/net/core/dev.c index d162ba8d622..aa761472f9e 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -132,6 +132,7 @@ #include #include #include +#include #include "net-sysfs.h" @@ -2588,6 +2589,53 @@ EXPORT_SYMBOL(__skb_get_rxhash); struct rps_sock_flow_table __rcu *rps_sock_flow_table __read_mostly; EXPORT_SYMBOL(rps_sock_flow_table); +static struct rps_dev_flow * +set_rps_cpu(struct net_device *dev, struct sk_buff *skb, + struct rps_dev_flow *rflow, u16 next_cpu) +{ + u16 tcpu; + + tcpu = rflow->cpu = next_cpu; + if (tcpu != RPS_NO_CPU) { +#ifdef CONFIG_RFS_ACCEL + struct netdev_rx_queue *rxqueue; + struct rps_dev_flow_table *flow_table; + struct rps_dev_flow *old_rflow; + u32 flow_id; + u16 rxq_index; + int rc; + + /* Should we steer this flow to a different hardware queue? */ + if (!skb_rx_queue_recorded(skb) || !dev->rx_cpu_rmap) + goto out; + rxq_index = cpu_rmap_lookup_index(dev->rx_cpu_rmap, next_cpu); + if (rxq_index == skb_get_rx_queue(skb)) + goto out; + + rxqueue = dev->_rx + rxq_index; + flow_table = rcu_dereference(rxqueue->rps_flow_table); + if (!flow_table) + goto out; + flow_id = skb->rxhash & flow_table->mask; + rc = dev->netdev_ops->ndo_rx_flow_steer(dev, skb, + rxq_index, flow_id); + if (rc < 0) + goto out; + old_rflow = rflow; + rflow = &flow_table->flows[flow_id]; + rflow->cpu = next_cpu; + rflow->filter = rc; + if (old_rflow->filter == rflow->filter) + old_rflow->filter = RPS_NO_FILTER; + out: +#endif + rflow->last_qtail = + per_cpu(softnet_data, tcpu).input_queue_head; + } + + return rflow; +} + /* * get_rps_cpu is called from netif_receive_skb and returns the target * CPU from the RPS map of the receiving queue for a given skb. @@ -2658,12 +2706,9 @@ static int get_rps_cpu(struct net_device *dev, struct sk_buff *skb, if (unlikely(tcpu != next_cpu) && (tcpu == RPS_NO_CPU || !cpu_online(tcpu) || ((int)(per_cpu(softnet_data, tcpu).input_queue_head - - rflow->last_qtail)) >= 0)) { - tcpu = rflow->cpu = next_cpu; - if (tcpu != RPS_NO_CPU) - rflow->last_qtail = per_cpu(softnet_data, - tcpu).input_queue_head; - } + rflow->last_qtail)) >= 0)) + rflow = set_rps_cpu(dev, skb, rflow, next_cpu); + if (tcpu != RPS_NO_CPU && cpu_online(tcpu)) { *rflowp = rflow; cpu = tcpu; @@ -2684,6 +2729,46 @@ done: return cpu; } +#ifdef CONFIG_RFS_ACCEL + +/** + * rps_may_expire_flow - check whether an RFS hardware filter may be removed + * @dev: Device on which the filter was set + * @rxq_index: RX queue index + * @flow_id: Flow ID passed to ndo_rx_flow_steer() + * @filter_id: Filter ID returned by ndo_rx_flow_steer() + * + * Drivers that implement ndo_rx_flow_steer() should periodically call + * this function for each installed filter and remove the filters for + * which it returns %true. + */ +bool rps_may_expire_flow(struct net_device *dev, u16 rxq_index, + u32 flow_id, u16 filter_id) +{ + struct netdev_rx_queue *rxqueue = dev->_rx + rxq_index; + struct rps_dev_flow_table *flow_table; + struct rps_dev_flow *rflow; + bool expire = true; + int cpu; + + rcu_read_lock(); + flow_table = rcu_dereference(rxqueue->rps_flow_table); + if (flow_table && flow_id <= flow_table->mask) { + rflow = &flow_table->flows[flow_id]; + cpu = ACCESS_ONCE(rflow->cpu); + if (rflow->filter == filter_id && cpu != RPS_NO_CPU && + ((int)(per_cpu(softnet_data, cpu).input_queue_head - + rflow->last_qtail) < + (int)(10 * flow_table->mask))) + expire = false; + } + rcu_read_unlock(); + return expire; +} +EXPORT_SYMBOL(rps_may_expire_flow); + +#endif /* CONFIG_RFS_ACCEL */ + /* Called from hardirq (IPI) context */ static void rps_trigger_softirq(void *data) { -- cgit v1.2.3-70-g09d2 From 57422dc530115e427dff464cc0a32bcd0efb5008 Mon Sep 17 00:00:00 2001 From: Michał Mirosław Date: Sat, 22 Jan 2011 12:14:12 +0000 Subject: net: Move check of checksum features to netdev_fix_features() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Michał Mirosław Signed-off-by: David S. Miller --- net/core/dev.c | 32 +++++++++++++++++--------------- 1 file changed, 17 insertions(+), 15 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index aa761472f9e..ad374189858 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5215,6 +5215,23 @@ static void rollback_registered(struct net_device *dev) unsigned long netdev_fix_features(unsigned long features, const char *name) { + /* Fix illegal checksum combinations */ + if ((features & NETIF_F_HW_CSUM) && + (features & (NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) { + if (name) + printk(KERN_NOTICE "%s: mixed HW and IP checksum settings.\n", + name); + features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM); + } + + if ((features & NETIF_F_NO_CSUM) && + (features & (NETIF_F_HW_CSUM|NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) { + if (name) + printk(KERN_NOTICE "%s: mixed no checksumming and other settings.\n", + name); + features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM|NETIF_F_HW_CSUM); + } + /* Fix illegal SG+CSUM combinations. */ if ((features & NETIF_F_SG) && !(features & NETIF_F_ALL_CSUM)) { @@ -5390,21 +5407,6 @@ int register_netdevice(struct net_device *dev) if (dev->iflink == -1) dev->iflink = dev->ifindex; - /* Fix illegal checksum combinations */ - if ((dev->features & NETIF_F_HW_CSUM) && - (dev->features & (NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) { - printk(KERN_NOTICE "%s: mixed HW and IP checksum settings.\n", - dev->name); - dev->features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM); - } - - if ((dev->features & NETIF_F_NO_CSUM) && - (dev->features & (NETIF_F_HW_CSUM|NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) { - printk(KERN_NOTICE "%s: mixed no checksumming and other settings.\n", - dev->name); - dev->features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM|NETIF_F_HW_CSUM); - } - dev->features = netdev_fix_features(dev->features, dev->name); /* Enable software GSO if SG is supported. */ -- cgit v1.2.3-70-g09d2 From 04ed3e741d0f133e02bed7fa5c98edba128f90e7 Mon Sep 17 00:00:00 2001 From: Michał Mirosław Date: Mon, 24 Jan 2011 15:32:47 -0800 Subject: net: change netdev->features to u32 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Quoting Ben Hutchings: we presumably won't be defining features that can only be enabled on 64-bit architectures. Occurences found by `grep -r` on net/, drivers/net, include/ [ Move features and vlan_features next to each other in struct netdev, as per Eric Dumazet's suggestion -DaveM ] Signed-off-by: Michał Mirosław Signed-off-by: David S. Miller --- drivers/net/bnx2.c | 2 +- drivers/net/bonding/bond_main.c | 4 ++-- drivers/net/myri10ge/myri10ge.c | 4 ++-- drivers/net/sfc/ethtool.c | 4 ++-- drivers/net/sfc/net_driver.h | 2 +- drivers/net/tun.c | 2 +- include/linux/netdevice.h | 24 ++++++++++++------------ include/linux/skbuff.h | 2 +- include/net/protocol.h | 4 ++-- include/net/tcp.h | 2 +- include/net/udp.h | 2 +- net/8021q/vlan.c | 2 +- net/bridge/br_if.c | 2 +- net/bridge/br_private.h | 2 +- net/core/dev.c | 15 +++++++-------- net/core/ethtool.c | 2 +- net/core/net-sysfs.c | 2 +- net/core/skbuff.c | 4 ++-- net/ipv4/af_inet.c | 2 +- net/ipv4/tcp.c | 2 +- net/ipv4/udp.c | 2 +- net/ipv6/af_inet6.c | 2 +- net/ipv6/udp.c | 2 +- 23 files changed, 45 insertions(+), 46 deletions(-) (limited to 'net/core/dev.c') diff --git a/drivers/net/bnx2.c b/drivers/net/bnx2.c index df99edf3464..cab96fa4cd3 100644 --- a/drivers/net/bnx2.c +++ b/drivers/net/bnx2.c @@ -8312,7 +8312,7 @@ static const struct net_device_ops bnx2_netdev_ops = { #endif }; -static void inline vlan_features_add(struct net_device *dev, unsigned long flags) +static void inline vlan_features_add(struct net_device *dev, u32 flags) { dev->vlan_features |= flags; } diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 163e0b06eaa..7047b406b8b 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1372,8 +1372,8 @@ static int bond_compute_features(struct bonding *bond) { struct slave *slave; struct net_device *bond_dev = bond->dev; - unsigned long features = bond_dev->features; - unsigned long vlan_features = 0; + u32 features = bond_dev->features; + u32 vlan_features = 0; unsigned short max_hard_header_len = max((u16)ETH_HLEN, bond_dev->hard_header_len); int i; diff --git a/drivers/net/myri10ge/myri10ge.c b/drivers/net/myri10ge/myri10ge.c index ea5cfe2c3a0..a7f2eed9a08 100644 --- a/drivers/net/myri10ge/myri10ge.c +++ b/drivers/net/myri10ge/myri10ge.c @@ -253,7 +253,7 @@ struct myri10ge_priv { unsigned long serial_number; int vendor_specific_offset; int fw_multicast_support; - unsigned long features; + u32 features; u32 max_tso6; u32 read_dma; u32 write_dma; @@ -1776,7 +1776,7 @@ static int myri10ge_set_rx_csum(struct net_device *netdev, u32 csum_enabled) static int myri10ge_set_tso(struct net_device *netdev, u32 tso_enabled) { struct myri10ge_priv *mgp = netdev_priv(netdev); - unsigned long flags = mgp->features & (NETIF_F_TSO6 | NETIF_F_TSO); + u32 flags = mgp->features & (NETIF_F_TSO6 | NETIF_F_TSO); if (tso_enabled) netdev->features |= flags; diff --git a/drivers/net/sfc/ethtool.c b/drivers/net/sfc/ethtool.c index 0e8bb19ed60..713969accdb 100644 --- a/drivers/net/sfc/ethtool.c +++ b/drivers/net/sfc/ethtool.c @@ -502,7 +502,7 @@ static void efx_ethtool_get_stats(struct net_device *net_dev, static int efx_ethtool_set_tso(struct net_device *net_dev, u32 enable) { struct efx_nic *efx __attribute__ ((unused)) = netdev_priv(net_dev); - unsigned long features; + u32 features; features = NETIF_F_TSO; if (efx->type->offload_features & NETIF_F_V6_CSUM) @@ -519,7 +519,7 @@ static int efx_ethtool_set_tso(struct net_device *net_dev, u32 enable) static int efx_ethtool_set_tx_csum(struct net_device *net_dev, u32 enable) { struct efx_nic *efx = netdev_priv(net_dev); - unsigned long features = efx->type->offload_features & NETIF_F_ALL_CSUM; + u32 features = efx->type->offload_features & NETIF_F_ALL_CSUM; if (enable) net_dev->features |= features; diff --git a/drivers/net/sfc/net_driver.h b/drivers/net/sfc/net_driver.h index 28df8665256..c65270241d2 100644 --- a/drivers/net/sfc/net_driver.h +++ b/drivers/net/sfc/net_driver.h @@ -906,7 +906,7 @@ struct efx_nic_type { unsigned int phys_addr_channels; unsigned int tx_dc_base; unsigned int rx_dc_base; - unsigned long offload_features; + u32 offload_features; u32 reset_world_flags; }; diff --git a/drivers/net/tun.c b/drivers/net/tun.c index b100bd50a0d..55786a0efc4 100644 --- a/drivers/net/tun.c +++ b/drivers/net/tun.c @@ -1142,7 +1142,7 @@ static int tun_get_iff(struct net *net, struct tun_struct *tun, * privs required. */ static int set_offload(struct net_device *dev, unsigned long arg) { - unsigned int old_features, features; + u32 old_features, features; old_features = dev->features; /* Unset features, set them as we chew on the arg. */ diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index a335f202269..0de3c59720f 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -914,7 +914,11 @@ struct net_device { struct list_head unreg_list; /* Net device features */ - unsigned long features; + u32 features; + + /* VLAN feature mask */ + u32 vlan_features; + #define NETIF_F_SG 1 /* Scatter/gather IO. */ #define NETIF_F_IP_CSUM 2 /* Can checksum TCP/UDP over IPv4. */ #define NETIF_F_NO_CSUM 4 /* Does not require checksum. F.e. loopack. */ @@ -1176,9 +1180,6 @@ struct net_device { /* rtnetlink link ops */ const struct rtnl_link_ops *rtnl_link_ops; - /* VLAN feature mask */ - unsigned long vlan_features; - /* for setting kernel sock attribute on TCP connection setup */ #define GSO_MAX_SIZE 65536 unsigned int gso_max_size; @@ -1401,7 +1402,7 @@ struct packet_type { struct packet_type *, struct net_device *); struct sk_buff *(*gso_segment)(struct sk_buff *skb, - int features); + u32 features); int (*gso_send_check)(struct sk_buff *skb); struct sk_buff **(*gro_receive)(struct sk_buff **head, struct sk_buff *skb); @@ -2370,7 +2371,7 @@ extern int netdev_tstamp_prequeue; extern int weight_p; extern int netdev_set_master(struct net_device *dev, struct net_device *master); extern int skb_checksum_help(struct sk_buff *skb); -extern struct sk_buff *skb_gso_segment(struct sk_buff *skb, int features); +extern struct sk_buff *skb_gso_segment(struct sk_buff *skb, u32 features); #ifdef CONFIG_BUG extern void netdev_rx_csum_fault(struct net_device *dev); #else @@ -2397,22 +2398,21 @@ extern char *netdev_drivername(const struct net_device *dev, char *buffer, int l extern void linkwatch_run_queue(void); -unsigned long netdev_increment_features(unsigned long all, unsigned long one, - unsigned long mask); -unsigned long netdev_fix_features(unsigned long features, const char *name); +u32 netdev_increment_features(u32 all, u32 one, u32 mask); +u32 netdev_fix_features(u32 features, const char *name); void netif_stacked_transfer_operstate(const struct net_device *rootdev, struct net_device *dev); -int netif_skb_features(struct sk_buff *skb); +u32 netif_skb_features(struct sk_buff *skb); -static inline int net_gso_ok(int features, int gso_type) +static inline int net_gso_ok(u32 features, int gso_type) { int feature = gso_type << NETIF_F_GSO_SHIFT; return (features & feature) == feature; } -static inline int skb_gso_ok(struct sk_buff *skb, int features) +static inline int skb_gso_ok(struct sk_buff *skb, u32 features) { return net_gso_ok(features, skb_shinfo(skb)->gso_type) && (!skb_has_frag_list(skb) || (features & NETIF_F_FRAGLIST)); diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 6e946da9d1d..31f02d0b46a 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -1877,7 +1877,7 @@ extern void skb_split(struct sk_buff *skb, extern int skb_shift(struct sk_buff *tgt, struct sk_buff *skb, int shiftlen); -extern struct sk_buff *skb_segment(struct sk_buff *skb, int features); +extern struct sk_buff *skb_segment(struct sk_buff *skb, u32 features); static inline void *skb_header_pointer(const struct sk_buff *skb, int offset, int len, void *buffer) diff --git a/include/net/protocol.h b/include/net/protocol.h index dc07495bce4..6f7eb800974 100644 --- a/include/net/protocol.h +++ b/include/net/protocol.h @@ -38,7 +38,7 @@ struct net_protocol { void (*err_handler)(struct sk_buff *skb, u32 info); int (*gso_send_check)(struct sk_buff *skb); struct sk_buff *(*gso_segment)(struct sk_buff *skb, - int features); + u32 features); struct sk_buff **(*gro_receive)(struct sk_buff **head, struct sk_buff *skb); int (*gro_complete)(struct sk_buff *skb); @@ -57,7 +57,7 @@ struct inet6_protocol { int (*gso_send_check)(struct sk_buff *skb); struct sk_buff *(*gso_segment)(struct sk_buff *skb, - int features); + u32 features); struct sk_buff **(*gro_receive)(struct sk_buff **head, struct sk_buff *skb); int (*gro_complete)(struct sk_buff *skb); diff --git a/include/net/tcp.h b/include/net/tcp.h index 38509f04738..917911165e3 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -1404,7 +1404,7 @@ extern struct request_sock_ops tcp6_request_sock_ops; extern void tcp_v4_destroy_sock(struct sock *sk); extern int tcp_v4_gso_send_check(struct sk_buff *skb); -extern struct sk_buff *tcp_tso_segment(struct sk_buff *skb, int features); +extern struct sk_buff *tcp_tso_segment(struct sk_buff *skb, u32 features); extern struct sk_buff **tcp_gro_receive(struct sk_buff **head, struct sk_buff *skb); extern struct sk_buff **tcp4_gro_receive(struct sk_buff **head, diff --git a/include/net/udp.h b/include/net/udp.h index bb967dd59bf..e82f3a8c0f8 100644 --- a/include/net/udp.h +++ b/include/net/udp.h @@ -245,5 +245,5 @@ extern void udp4_proc_exit(void); extern void udp_init(void); extern int udp4_ufo_send_check(struct sk_buff *skb); -extern struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb, int features); +extern struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb, u32 features); #endif /* _UDP_H */ diff --git a/net/8021q/vlan.c b/net/8021q/vlan.c index 6e64f7c6a2e..7850412f52b 100644 --- a/net/8021q/vlan.c +++ b/net/8021q/vlan.c @@ -327,7 +327,7 @@ static void vlan_sync_address(struct net_device *dev, static void vlan_transfer_features(struct net_device *dev, struct net_device *vlandev) { - unsigned long old_features = vlandev->features; + u32 old_features = vlandev->features; vlandev->features &= ~dev->vlan_features; vlandev->features |= dev->features & dev->vlan_features; diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index d9d1e2bac1d..52ce4a30f8b 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -365,7 +365,7 @@ int br_min_mtu(const struct net_bridge *br) void br_features_recompute(struct net_bridge *br) { struct net_bridge_port *p; - unsigned long features, mask; + u32 features, mask; features = mask = br->feature_mask; if (list_empty(&br->port_list)) diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 84aac7734bf..9f22898c535 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -182,7 +182,7 @@ struct net_bridge struct br_cpu_netstats __percpu *stats; spinlock_t hash_lock; struct hlist_head hash[BR_HASH_SIZE]; - unsigned long feature_mask; + u32 feature_mask; #ifdef CONFIG_BRIDGE_NETFILTER struct rtable fake_rtable; bool nf_call_iptables; diff --git a/net/core/dev.c b/net/core/dev.c index ad374189858..7103f89fde0 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1858,7 +1858,7 @@ EXPORT_SYMBOL(skb_checksum_help); * It may return NULL if the skb requires no segmentation. This is * only possible when GSO is used for verifying header integrity. */ -struct sk_buff *skb_gso_segment(struct sk_buff *skb, int features) +struct sk_buff *skb_gso_segment(struct sk_buff *skb, u32 features) { struct sk_buff *segs = ERR_PTR(-EPROTONOSUPPORT); struct packet_type *ptype; @@ -2046,7 +2046,7 @@ static bool can_checksum_protocol(unsigned long features, __be16 protocol) protocol == htons(ETH_P_FCOE))); } -static int harmonize_features(struct sk_buff *skb, __be16 protocol, int features) +static u32 harmonize_features(struct sk_buff *skb, __be16 protocol, u32 features) { if (!can_checksum_protocol(features, protocol)) { features &= ~NETIF_F_ALL_CSUM; @@ -2058,10 +2058,10 @@ static int harmonize_features(struct sk_buff *skb, __be16 protocol, int features return features; } -int netif_skb_features(struct sk_buff *skb) +u32 netif_skb_features(struct sk_buff *skb) { __be16 protocol = skb->protocol; - int features = skb->dev->features; + u32 features = skb->dev->features; if (protocol == htons(ETH_P_8021Q)) { struct vlan_ethhdr *veh = (struct vlan_ethhdr *)skb->data; @@ -2106,7 +2106,7 @@ int dev_hard_start_xmit(struct sk_buff *skb, struct net_device *dev, int rc = NETDEV_TX_OK; if (likely(!skb->next)) { - int features; + u32 features; /* * If device doesnt need skb->dst, release it right now while @@ -5213,7 +5213,7 @@ static void rollback_registered(struct net_device *dev) rollback_registered_many(&single); } -unsigned long netdev_fix_features(unsigned long features, const char *name) +u32 netdev_fix_features(u32 features, const char *name) { /* Fix illegal checksum combinations */ if ((features & NETIF_F_HW_CSUM) && @@ -6143,8 +6143,7 @@ static int dev_cpu_callback(struct notifier_block *nfb, * @one to the master device with current feature set @all. Will not * enable anything that is off in @mask. Returns the new feature set. */ -unsigned long netdev_increment_features(unsigned long all, unsigned long one, - unsigned long mask) +u32 netdev_increment_features(u32 all, u32 one, u32 mask) { /* If device needs checksumming, downgrade to it. */ if (all & NETIF_F_NO_CSUM && !(one & NETIF_F_NO_CSUM)) diff --git a/net/core/ethtool.c b/net/core/ethtool.c index 17741782a34..bd1af99e112 100644 --- a/net/core/ethtool.c +++ b/net/core/ethtool.c @@ -1458,7 +1458,7 @@ int dev_ethtool(struct net *net, struct ifreq *ifr) void __user *useraddr = ifr->ifr_data; u32 ethcmd; int rc; - unsigned long old_features; + u32 old_features; if (!dev || !netif_device_present(dev)) return -ENODEV; diff --git a/net/core/net-sysfs.c b/net/core/net-sysfs.c index e23c01be5a5..81367ccf330 100644 --- a/net/core/net-sysfs.c +++ b/net/core/net-sysfs.c @@ -99,7 +99,7 @@ NETDEVICE_SHOW(addr_assign_type, fmt_dec); NETDEVICE_SHOW(addr_len, fmt_dec); NETDEVICE_SHOW(iflink, fmt_dec); NETDEVICE_SHOW(ifindex, fmt_dec); -NETDEVICE_SHOW(features, fmt_long_hex); +NETDEVICE_SHOW(features, fmt_hex); NETDEVICE_SHOW(type, fmt_dec); NETDEVICE_SHOW(link_mode, fmt_dec); diff --git a/net/core/skbuff.c b/net/core/skbuff.c index d31bb36ae0d..436c4c43924 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -2497,7 +2497,7 @@ EXPORT_SYMBOL_GPL(skb_pull_rcsum); * a pointer to the first in a list of new skbs for the segments. * In case of error it returns ERR_PTR(err). */ -struct sk_buff *skb_segment(struct sk_buff *skb, int features) +struct sk_buff *skb_segment(struct sk_buff *skb, u32 features) { struct sk_buff *segs = NULL; struct sk_buff *tail = NULL; @@ -2507,7 +2507,7 @@ struct sk_buff *skb_segment(struct sk_buff *skb, int features) unsigned int offset = doffset; unsigned int headroom; unsigned int len; - int sg = features & NETIF_F_SG; + int sg = !!(features & NETIF_F_SG); int nfrags = skb_shinfo(skb)->nr_frags; int err = -ENOMEM; int i = 0; diff --git a/net/ipv4/af_inet.c b/net/ipv4/af_inet.c index f2b61107df6..e5e2d9d64ab 100644 --- a/net/ipv4/af_inet.c +++ b/net/ipv4/af_inet.c @@ -1215,7 +1215,7 @@ out: return err; } -static struct sk_buff *inet_gso_segment(struct sk_buff *skb, int features) +static struct sk_buff *inet_gso_segment(struct sk_buff *skb, u32 features) { struct sk_buff *segs = ERR_PTR(-EINVAL); struct iphdr *iph; diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 6c11eece262..f9867d2dbef 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -2653,7 +2653,7 @@ int compat_tcp_getsockopt(struct sock *sk, int level, int optname, EXPORT_SYMBOL(compat_tcp_getsockopt); #endif -struct sk_buff *tcp_tso_segment(struct sk_buff *skb, int features) +struct sk_buff *tcp_tso_segment(struct sk_buff *skb, u32 features) { struct sk_buff *segs = ERR_PTR(-EINVAL); struct tcphdr *th; diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c index 8157b17959e..d37baaa1dbe 100644 --- a/net/ipv4/udp.c +++ b/net/ipv4/udp.c @@ -2199,7 +2199,7 @@ int udp4_ufo_send_check(struct sk_buff *skb) return 0; } -struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb, int features) +struct sk_buff *udp4_ufo_fragment(struct sk_buff *skb, u32 features) { struct sk_buff *segs = ERR_PTR(-EINVAL); unsigned int mss; diff --git a/net/ipv6/af_inet6.c b/net/ipv6/af_inet6.c index 978e80e2c4a..3194aa90987 100644 --- a/net/ipv6/af_inet6.c +++ b/net/ipv6/af_inet6.c @@ -772,7 +772,7 @@ out: return err; } -static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb, int features) +static struct sk_buff *ipv6_gso_segment(struct sk_buff *skb, u32 features) { struct sk_buff *segs = ERR_PTR(-EINVAL); struct ipv6hdr *ipv6h; diff --git a/net/ipv6/udp.c b/net/ipv6/udp.c index 9a009c66c8a..a419a787eb6 100644 --- a/net/ipv6/udp.c +++ b/net/ipv6/udp.c @@ -1299,7 +1299,7 @@ static int udp6_ufo_send_check(struct sk_buff *skb) return 0; } -static struct sk_buff *udp6_ufo_fragment(struct sk_buff *skb, int features) +static struct sk_buff *udp6_ufo_fragment(struct sk_buff *skb, u32 features) { struct sk_buff *segs = ERR_PTR(-EINVAL); unsigned int mss; -- cgit v1.2.3-70-g09d2 From acd1130e8793fb150fb522da8ec51675839eb4b1 Mon Sep 17 00:00:00 2001 From: Michał Mirosław Date: Mon, 24 Jan 2011 15:45:15 -0800 Subject: net: reduce and unify printk level in netdev_fix_features() MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Reduce printk() levels to KERN_INFO in netdev_fix_features() as this will be used by ethtool and might spam dmesg unnecessarily. This converts the function to use netdev_info() instead of plain printk(). As a side effect, bonding and bridge devices will now log dropped features on every slave device change. Signed-off-by: Michał Mirosław Signed-off-by: David S. Miller --- drivers/net/bonding/bond_main.c | 4 ++-- include/linux/netdevice.h | 2 +- net/bridge/br_if.c | 2 +- net/core/dev.c | 33 ++++++++++++--------------------- 4 files changed, 16 insertions(+), 25 deletions(-) (limited to 'net/core/dev.c') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 7047b406b8b..1df9f0ea918 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1400,8 +1400,8 @@ static int bond_compute_features(struct bonding *bond) done: features |= (bond_dev->features & BOND_VLAN_FEATURES); - bond_dev->features = netdev_fix_features(features, NULL); - bond_dev->vlan_features = netdev_fix_features(vlan_features, NULL); + bond_dev->features = netdev_fix_features(bond_dev, features); + bond_dev->vlan_features = netdev_fix_features(bond_dev, vlan_features); bond_dev->hard_header_len = max_hard_header_len; return 0; diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 0de3c59720f..8858422c5c5 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2399,7 +2399,7 @@ extern char *netdev_drivername(const struct net_device *dev, char *buffer, int l extern void linkwatch_run_queue(void); u32 netdev_increment_features(u32 all, u32 one, u32 mask); -u32 netdev_fix_features(u32 features, const char *name); +u32 netdev_fix_features(struct net_device *dev, u32 features); void netif_stacked_transfer_operstate(const struct net_device *rootdev, struct net_device *dev); diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c index 52ce4a30f8b..2a6801d8b72 100644 --- a/net/bridge/br_if.c +++ b/net/bridge/br_if.c @@ -379,7 +379,7 @@ void br_features_recompute(struct net_bridge *br) } done: - br->dev->features = netdev_fix_features(features, NULL); + br->dev->features = netdev_fix_features(br->dev, features); } /* called with RTNL */ diff --git a/net/core/dev.c b/net/core/dev.c index 7103f89fde0..1b4c07fe295 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5213,58 +5213,49 @@ static void rollback_registered(struct net_device *dev) rollback_registered_many(&single); } -u32 netdev_fix_features(u32 features, const char *name) +u32 netdev_fix_features(struct net_device *dev, u32 features) { /* Fix illegal checksum combinations */ if ((features & NETIF_F_HW_CSUM) && (features & (NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) { - if (name) - printk(KERN_NOTICE "%s: mixed HW and IP checksum settings.\n", - name); + netdev_info(dev, "mixed HW and IP checksum settings.\n"); features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM); } if ((features & NETIF_F_NO_CSUM) && (features & (NETIF_F_HW_CSUM|NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) { - if (name) - printk(KERN_NOTICE "%s: mixed no checksumming and other settings.\n", - name); + netdev_info(dev, "mixed no checksumming and other settings.\n"); features &= ~(NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM|NETIF_F_HW_CSUM); } /* Fix illegal SG+CSUM combinations. */ if ((features & NETIF_F_SG) && !(features & NETIF_F_ALL_CSUM)) { - if (name) - printk(KERN_NOTICE "%s: Dropping NETIF_F_SG since no " - "checksum feature.\n", name); + netdev_info(dev, + "Dropping NETIF_F_SG since no checksum feature.\n"); features &= ~NETIF_F_SG; } /* TSO requires that SG is present as well. */ if ((features & NETIF_F_TSO) && !(features & NETIF_F_SG)) { - if (name) - printk(KERN_NOTICE "%s: Dropping NETIF_F_TSO since no " - "SG feature.\n", name); + netdev_info(dev, "Dropping NETIF_F_TSO since no SG feature.\n"); features &= ~NETIF_F_TSO; } + /* UFO needs SG and checksumming */ if (features & NETIF_F_UFO) { /* maybe split UFO into V4 and V6? */ if (!((features & NETIF_F_GEN_CSUM) || (features & (NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM)) == (NETIF_F_IP_CSUM|NETIF_F_IPV6_CSUM))) { - if (name) - printk(KERN_ERR "%s: Dropping NETIF_F_UFO " - "since no checksum offload features.\n", - name); + netdev_info(dev, + "Dropping NETIF_F_UFO since no checksum offload features.\n"); features &= ~NETIF_F_UFO; } if (!(features & NETIF_F_SG)) { - if (name) - printk(KERN_ERR "%s: Dropping NETIF_F_UFO " - "since no NETIF_F_SG feature.\n", name); + netdev_info(dev, + "Dropping NETIF_F_UFO since no NETIF_F_SG feature.\n"); features &= ~NETIF_F_UFO; } } @@ -5407,7 +5398,7 @@ int register_netdevice(struct net_device *dev) if (dev->iflink == -1) dev->iflink = dev->ifindex; - dev->features = netdev_fix_features(dev->features, dev->name); + dev->features = netdev_fix_features(dev, dev->features); /* Enable software GSO if SG is supported. */ if (dev->features & NETIF_F_SG) -- cgit v1.2.3-70-g09d2 From ccf434380d1a67df2dcb9113206b77d0cb0a1cef Mon Sep 17 00:00:00 2001 From: Eric Dumazet Date: Wed, 26 Jan 2011 18:08:02 +0000 Subject: net: fix dev_seq_next() Commit c6d14c84566d (net: Introduce for_each_netdev_rcu() iterator) added a race in dev_seq_next(). The rcu_dereference() call should be done _before_ testing the end of list, or we might return a wrong net_device if a concurrent thread changes net_device list under us. Note : discovered thanks to a sparse warning : net/core/dev.c:3919:9: error: incompatible types in comparison expression (different address spaces) Signed-off-by: Eric Dumazet CC: Paul E. McKenney Signed-off-by: David S. Miller --- include/linux/netdevice.h | 9 ++++++++- net/core/dev.c | 11 +++++++---- 2 files changed, 15 insertions(+), 5 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 8858422c5c5..c7d70745222 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -1447,7 +1447,7 @@ static inline struct net_device *next_net_device_rcu(struct net_device *dev) struct net *net; net = dev_net(dev); - lh = rcu_dereference(dev->dev_list.next); + lh = rcu_dereference(list_next_rcu(&dev->dev_list)); return lh == &net->dev_base_head ? NULL : net_device_entry(lh); } @@ -1457,6 +1457,13 @@ static inline struct net_device *first_net_device(struct net *net) net_device_entry(net->dev_base_head.next); } +static inline struct net_device *first_net_device_rcu(struct net *net) +{ + struct list_head *lh = rcu_dereference(list_next_rcu(&net->dev_base_head)); + + return lh == &net->dev_base_head ? NULL : net_device_entry(lh); +} + extern int netdev_boot_setup_check(struct net_device *dev); extern unsigned long netdev_boot_base(const char *prefix, int unit); extern struct net_device *dev_getbyhwaddr_rcu(struct net *net, unsigned short type, diff --git a/net/core/dev.c b/net/core/dev.c index 1b4c07fe295..ddd5df2b61d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -4051,12 +4051,15 @@ void *dev_seq_start(struct seq_file *seq, loff_t *pos) void *dev_seq_next(struct seq_file *seq, void *v, loff_t *pos) { - struct net_device *dev = (v == SEQ_START_TOKEN) ? - first_net_device(seq_file_net(seq)) : - next_net_device((struct net_device *)v); + struct net_device *dev = v; + + if (v == SEQ_START_TOKEN) + dev = first_net_device_rcu(seq_file_net(seq)); + else + dev = next_net_device_rcu(dev); ++*pos; - return rcu_dereference(dev); + return dev; } void dev_seq_stop(struct seq_file *seq, void *v) -- cgit v1.2.3-70-g09d2 From d59cfde2fb960b5970ccb5a38cea25d38b37a8e8 Mon Sep 17 00:00:00 2001 From: Jiri Pirko Date: Sat, 12 Feb 2011 00:46:06 +0000 Subject: net: remove the unnecessary dance around skb_bond_should_drop MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit No need to check (master) twice and to drive in and out the header file. Signed-off-by: Jiri Pirko Reviewed-by: Nicolas de Pesloüan Signed-off-by: David S. Miller --- include/linux/netdevice.h | 11 ----------- net/core/dev.c | 6 +++--- 2 files changed, 3 insertions(+), 14 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index c7d70745222..5a5baeaaa50 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2437,17 +2437,6 @@ static inline void netif_set_gso_max_size(struct net_device *dev, dev->gso_max_size = size; } -extern int __skb_bond_should_drop(struct sk_buff *skb, - struct net_device *master); - -static inline int skb_bond_should_drop(struct sk_buff *skb, - struct net_device *master) -{ - if (master) - return __skb_bond_should_drop(skb, master); - return 0; -} - extern struct pernet_operations __net_initdata loopback_net_ops; static inline int dev_ethtool_get_settings(struct net_device *dev, diff --git a/net/core/dev.c b/net/core/dev.c index 6392ea0a591..d874fd1baf4 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3105,7 +3105,8 @@ static inline void skb_bond_set_mac_by_master(struct sk_buff *skb, * duplicates except for 802.3ad ETH_P_SLOW, alb non-mcast/bcast, and * ARP on active-backup slaves with arp_validate enabled. */ -int __skb_bond_should_drop(struct sk_buff *skb, struct net_device *master) +static int __skb_bond_should_drop(struct sk_buff *skb, + struct net_device *master) { struct net_device *dev = skb->dev; @@ -3139,7 +3140,6 @@ int __skb_bond_should_drop(struct sk_buff *skb, struct net_device *master) } return 0; } -EXPORT_SYMBOL(__skb_bond_should_drop); static int __netif_receive_skb(struct sk_buff *skb) { @@ -3177,7 +3177,7 @@ static int __netif_receive_skb(struct sk_buff *skb) if (skb->deliver_no_wcard) null_or_orig = orig_dev; else if (master) { - if (skb_bond_should_drop(skb, master)) { + if (__skb_bond_should_drop(skb, master)) { skb->deliver_no_wcard = 1; null_or_orig = orig_dev; /* deliver only exact match */ } else -- cgit v1.2.3-70-g09d2 From 1765a575334f1a232c1478accdee5c7d19f4b3e3 Mon Sep 17 00:00:00 2001 From: Jiri Pirko Date: Sat, 12 Feb 2011 06:48:36 +0000 Subject: net: make dev->master general dev->master is now tightly connected to bonding driver. This patch makes this pointer more general and ready to be used by others. - netdev_set_master() - bond specifics moved to new function netdev_set_bond_master() - introduced netif_is_bond_slave() to check if device is a bonding slave Signed-off-by: Jiri Pirko Signed-off-by: David S. Miller --- drivers/infiniband/hw/nes/nes.c | 3 ++- drivers/infiniband/hw/nes/nes_cm.c | 2 +- drivers/net/bonding/bond_main.c | 10 ++++---- drivers/net/cxgb3/cxgb3_offload.c | 3 ++- include/linux/netdevice.h | 7 ++++++ net/core/dev.c | 49 ++++++++++++++++++++++++++++---------- 6 files changed, 54 insertions(+), 20 deletions(-) (limited to 'net/core/dev.c') diff --git a/drivers/infiniband/hw/nes/nes.c b/drivers/infiniband/hw/nes/nes.c index 3b4ec3238ce..3d7f3664b67 100644 --- a/drivers/infiniband/hw/nes/nes.c +++ b/drivers/infiniband/hw/nes/nes.c @@ -153,7 +153,8 @@ static int nes_inetaddr_event(struct notifier_block *notifier, nesdev, nesdev->netdev[0]->name); netdev = nesdev->netdev[0]; nesvnic = netdev_priv(netdev); - is_bonded = (netdev->master == event_netdev); + is_bonded = netif_is_bond_slave(netdev) && + (netdev->master == event_netdev); if ((netdev == event_netdev) || is_bonded) { if (nesvnic->rdma_enabled == 0) { nes_debug(NES_DBG_NETDEV, "Returning without processing event for %s since" diff --git a/drivers/infiniband/hw/nes/nes_cm.c b/drivers/infiniband/hw/nes/nes_cm.c index 009ec814d51..ec3aa11c36c 100644 --- a/drivers/infiniband/hw/nes/nes_cm.c +++ b/drivers/infiniband/hw/nes/nes_cm.c @@ -1118,7 +1118,7 @@ static int nes_addr_resolve_neigh(struct nes_vnic *nesvnic, u32 dst_ip, int arpi return rc; } - if (nesvnic->netdev->master) + if (netif_is_bond_slave(netdev)) netdev = nesvnic->netdev->master; else netdev = nesvnic->netdev; diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 1df9f0ea918..9f877878d63 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1594,9 +1594,9 @@ int bond_enslave(struct net_device *bond_dev, struct net_device *slave_dev) } } - res = netdev_set_master(slave_dev, bond_dev); + res = netdev_set_bond_master(slave_dev, bond_dev); if (res) { - pr_debug("Error %d calling netdev_set_master\n", res); + pr_debug("Error %d calling netdev_set_bond_master\n", res); goto err_restore_mac; } /* open the slave since the application closed it */ @@ -1812,7 +1812,7 @@ err_close: dev_close(slave_dev); err_unset_master: - netdev_set_master(slave_dev, NULL); + netdev_set_bond_master(slave_dev, NULL); err_restore_mac: if (!bond->params.fail_over_mac) { @@ -1992,7 +1992,7 @@ int bond_release(struct net_device *bond_dev, struct net_device *slave_dev) netif_addr_unlock_bh(bond_dev); } - netdev_set_master(slave_dev, NULL); + netdev_set_bond_master(slave_dev, NULL); #ifdef CONFIG_NET_POLL_CONTROLLER read_lock_bh(&bond->lock); @@ -2114,7 +2114,7 @@ static int bond_release_all(struct net_device *bond_dev) netif_addr_unlock_bh(bond_dev); } - netdev_set_master(slave_dev, NULL); + netdev_set_bond_master(slave_dev, NULL); /* close slave before restoring its mac address */ dev_close(slave_dev); diff --git a/drivers/net/cxgb3/cxgb3_offload.c b/drivers/net/cxgb3/cxgb3_offload.c index 7ea94b5205f..862804f32b6 100644 --- a/drivers/net/cxgb3/cxgb3_offload.c +++ b/drivers/net/cxgb3/cxgb3_offload.c @@ -186,9 +186,10 @@ static struct net_device *get_iff_from_mac(struct adapter *adapter, dev = NULL; if (grp) dev = vlan_group_get_device(grp, vlan); - } else + } else if (netif_is_bond_slave(dev)) { while (dev->master) dev = dev->master; + } return dev; } } diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 5a5baeaaa50..5a42b100376 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2377,6 +2377,8 @@ extern int netdev_max_backlog; extern int netdev_tstamp_prequeue; extern int weight_p; extern int netdev_set_master(struct net_device *dev, struct net_device *master); +extern int netdev_set_bond_master(struct net_device *dev, + struct net_device *master); extern int skb_checksum_help(struct sk_buff *skb); extern struct sk_buff *skb_gso_segment(struct sk_buff *skb, u32 features); #ifdef CONFIG_BUG @@ -2437,6 +2439,11 @@ static inline void netif_set_gso_max_size(struct net_device *dev, dev->gso_max_size = size; } +static inline int netif_is_bond_slave(struct net_device *dev) +{ + return dev->flags & IFF_SLAVE && dev->priv_flags & IFF_BONDING; +} + extern struct pernet_operations __net_initdata loopback_net_ops; static inline int dev_ethtool_get_settings(struct net_device *dev, diff --git a/net/core/dev.c b/net/core/dev.c index d874fd1baf4..a4132766d36 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3146,7 +3146,6 @@ static int __netif_receive_skb(struct sk_buff *skb) struct packet_type *ptype, *pt_prev; rx_handler_func_t *rx_handler; struct net_device *orig_dev; - struct net_device *master; struct net_device *null_or_orig; struct net_device *orig_or_bond; int ret = NET_RX_DROP; @@ -3173,15 +3172,19 @@ static int __netif_receive_skb(struct sk_buff *skb) */ null_or_orig = NULL; orig_dev = skb->dev; - master = ACCESS_ONCE(orig_dev->master); if (skb->deliver_no_wcard) null_or_orig = orig_dev; - else if (master) { - if (__skb_bond_should_drop(skb, master)) { - skb->deliver_no_wcard = 1; - null_or_orig = orig_dev; /* deliver only exact match */ - } else - skb->dev = master; + else if (netif_is_bond_slave(orig_dev)) { + struct net_device *bond_master = ACCESS_ONCE(orig_dev->master); + + if (likely(bond_master)) { + if (__skb_bond_should_drop(skb, bond_master)) { + skb->deliver_no_wcard = 1; + /* deliver only exact match */ + null_or_orig = orig_dev; + } else + skb->dev = bond_master; + } } __this_cpu_inc(softnet_data.processed); @@ -4346,15 +4349,14 @@ static int __init dev_proc_init(void) /** - * netdev_set_master - set up master/slave pair + * netdev_set_master - set up master pointer * @slave: slave device * @master: new master device * * Changes the master device of the slave. Pass %NULL to break the * bonding. The caller must hold the RTNL semaphore. On a failure * a negative errno code is returned. On success the reference counts - * are adjusted, %RTM_NEWLINK is sent to the routing socket and the - * function returns zero. + * are adjusted and the function returns zero. */ int netdev_set_master(struct net_device *slave, struct net_device *master) { @@ -4374,6 +4376,29 @@ int netdev_set_master(struct net_device *slave, struct net_device *master) synchronize_net(); dev_put(old); } + return 0; +} +EXPORT_SYMBOL(netdev_set_master); + +/** + * netdev_set_bond_master - set up bonding master/slave pair + * @slave: slave device + * @master: new master device + * + * Changes the master device of the slave. Pass %NULL to break the + * bonding. The caller must hold the RTNL semaphore. On a failure + * a negative errno code is returned. On success %RTM_NEWLINK is sent + * to the routing socket and the function returns zero. + */ +int netdev_set_bond_master(struct net_device *slave, struct net_device *master) +{ + int err; + + ASSERT_RTNL(); + + err = netdev_set_master(slave, master); + if (err) + return err; if (master) slave->flags |= IFF_SLAVE; else @@ -4382,7 +4407,7 @@ int netdev_set_master(struct net_device *slave, struct net_device *master) rtmsg_ifinfo(RTM_NEWLINK, slave, IFF_SLAVE); return 0; } -EXPORT_SYMBOL(netdev_set_master); +EXPORT_SYMBOL(netdev_set_bond_master); static void dev_change_rx_flags(struct net_device *dev, int flags) { -- cgit v1.2.3-70-g09d2 From 5c56580b74e57e56f30e3c5bbc9d7ab487858497 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Tue, 15 Feb 2011 19:39:21 +0000 Subject: net: Adjust TX queue kobjects if number of queues changes during unregister If the root qdisc for a net device is mqprio, and the driver's ndo_setup_tc() operation dynamically adds and remvoes TX queues, netif_set_real_num_tx_queues() will be called during device unregistration to remove the extra TX queues when the qdisc is destroyed. Currently this causes the corresponding kobjects to be leaked, and the device's reference count never drops to 0. Signed-off-by: Ben Hutchings --- net/core/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 6392ea0a591..30c71f9b041 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1648,7 +1648,8 @@ int netif_set_real_num_tx_queues(struct net_device *dev, unsigned int txq) if (txq < 1 || txq > dev->num_tx_queues) return -EINVAL; - if (dev->reg_state == NETREG_REGISTERED) { + if (dev->reg_state == NETREG_REGISTERED || + dev->reg_state == NETREG_UNREGISTERING) { ASSERT_RTNL(); rc = netdev_queue_update_kobjects(dev, dev->real_num_tx_queues, -- cgit v1.2.3-70-g09d2 From 69a19ee60d5d5adc0addbdffd254f83b60660a07 Mon Sep 17 00:00:00 2001 From: Ben Hutchings Date: Tue, 15 Feb 2011 20:32:04 +0000 Subject: net: RPS: Make hardware-accelerated RFS conditional on NETIF_F_NTUPLE For testing and debugging purposes it is useful to be able to disable hardware acceleration of RFS without disabling RFS altogether. Since this is a similar feature to 'n-tuple' flow steering through the ethtool API, test the same feature flag that controls that. Signed-off-by: Ben Hutchings --- net/core/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 30c71f9b041..54aaca69a02 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2607,7 +2607,8 @@ set_rps_cpu(struct net_device *dev, struct sk_buff *skb, int rc; /* Should we steer this flow to a different hardware queue? */ - if (!skb_rx_queue_recorded(skb) || !dev->rx_cpu_rmap) + if (!skb_rx_queue_recorded(skb) || !dev->rx_cpu_rmap || + !(dev->features & NETIF_F_NTUPLE)) goto out; rxq_index = cpu_rmap_lookup_index(dev->rx_cpu_rmap, next_cpu); if (rxq_index == skb_get_rx_queue(skb)) -- cgit v1.2.3-70-g09d2 From 212b573f5552c60265da721ff9ce32e3462a2cdd Mon Sep 17 00:00:00 2001 From: Michał Mirosław Date: Tue, 15 Feb 2011 16:59:16 +0000 Subject: ethtool: enable GSO and GRO by default MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Signed-off-by: Michał Mirosław Signed-off-by: David S. Miller --- include/linux/netdevice.h | 3 +++ net/core/dev.c | 18 ++++++++++++++---- 2 files changed, 17 insertions(+), 4 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index d08ef653857..168e3ad14da 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -984,6 +984,9 @@ struct net_device { NETIF_F_SG | NETIF_F_HIGHDMA | \ NETIF_F_FRAGLIST) + /* changeable features with no special hardware requirements */ +#define NETIF_F_SOFT_FEATURES (NETIF_F_GSO | NETIF_F_GRO) + /* Interface index. Unique device identifier */ int ifindex; int iflink; diff --git a/net/core/dev.c b/net/core/dev.c index 4580460ebec..8686f6ffe7f 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5274,6 +5274,12 @@ u32 netdev_fix_features(struct net_device *dev, u32 features) features &= ~NETIF_F_TSO; } + /* Software GSO depends on SG. */ + if ((features & NETIF_F_GSO) && !(features & NETIF_F_SG)) { + netdev_info(dev, "Dropping NETIF_F_GSO since no SG feature.\n"); + features &= ~NETIF_F_GSO; + } + /* UFO needs SG and checksumming */ if (features & NETIF_F_UFO) { /* maybe split UFO into V4 and V6? */ @@ -5430,11 +5436,15 @@ int register_netdevice(struct net_device *dev) if (dev->iflink == -1) dev->iflink = dev->ifindex; - dev->features = netdev_fix_features(dev, dev->features); + /* Enable software offloads by default - will be stripped in + * netdev_fix_features() if not supported. */ + dev->features |= NETIF_F_SOFT_FEATURES; - /* Enable software GSO if SG is supported. */ - if (dev->features & NETIF_F_SG) - dev->features |= NETIF_F_GSO; + /* Avoid warning from netdev_fix_features() for GSO without SG */ + if (!(dev->features & NETIF_F_SG)) + dev->features &= ~NETIF_F_GSO; + + dev->features = netdev_fix_features(dev, dev->features); /* Enable GRO and NETIF_F_HIGHDMA for vlans by default, * vlan_dev_init() will do the dev->features check, so these features -- cgit v1.2.3-70-g09d2 From 5455c6998d34dc983a8693500e4dffefc3682dc5 Mon Sep 17 00:00:00 2001 From: Michał Mirosław Date: Tue, 15 Feb 2011 16:59:17 +0000 Subject: net: Introduce new feature setting ops MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This introduces a new framework to handle device features setting. It consists of: - new fields in struct net_device: + hw_features - features that hw/driver supports toggling + wanted_features - features that user wants enabled, when possible - new netdev_ops: + feat = ndo_fix_features(dev, feat) - API checking constraints for enabling features or their combinations + ndo_set_features(dev) - API updating hardware state to match changed dev->features - new ethtool commands: + ETHTOOL_GFEATURES/ETHTOOL_SFEATURES: get/set dev->wanted_features and trigger device reconfiguration if resulting dev->features changed + ETHTOOL_GSTRINGS(ETH_SS_FEATURES): get feature bits names (meaning) Signed-off-by: Michał Mirosław Signed-off-by: David S. Miller --- include/linux/ethtool.h | 85 +++++++++++++++++++++++++++++++ include/linux/netdevice.h | 37 +++++++++++++- net/core/dev.c | 46 ++++++++++++++--- net/core/ethtool.c | 125 +++++++++++++++++++++++++++++++++++++++++++++- 4 files changed, 283 insertions(+), 10 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/linux/ethtool.h b/include/linux/ethtool.h index 1908929204a..806e716bb4f 100644 --- a/include/linux/ethtool.h +++ b/include/linux/ethtool.h @@ -251,6 +251,7 @@ enum ethtool_stringset { ETH_SS_STATS, ETH_SS_PRIV_FLAGS, ETH_SS_NTUPLE_FILTERS, + ETH_SS_FEATURES, }; /* for passing string sets for data tagging */ @@ -523,6 +524,87 @@ struct ethtool_flash { char data[ETHTOOL_FLASH_MAX_FILENAME]; }; +/* for returning and changing feature sets */ + +/** + * struct ethtool_get_features_block - block with state of 32 features + * @available: mask of changeable features + * @requested: mask of features requested to be enabled if possible + * @active: mask of currently enabled features + * @never_changed: mask of features not changeable for any device + */ +struct ethtool_get_features_block { + __u32 available; + __u32 requested; + __u32 active; + __u32 never_changed; +}; + +/** + * struct ethtool_gfeatures - command to get state of device's features + * @cmd: command number = %ETHTOOL_GFEATURES + * @size: in: number of elements in the features[] array; + * out: number of elements in features[] needed to hold all features + * @features: state of features + */ +struct ethtool_gfeatures { + __u32 cmd; + __u32 size; + struct ethtool_get_features_block features[0]; +}; + +/** + * struct ethtool_set_features_block - block with request for 32 features + * @valid: mask of features to be changed + * @requested: values of features to be changed + */ +struct ethtool_set_features_block { + __u32 valid; + __u32 requested; +}; + +/** + * struct ethtool_sfeatures - command to request change in device's features + * @cmd: command number = %ETHTOOL_SFEATURES + * @size: array size of the features[] array + * @features: feature change masks + */ +struct ethtool_sfeatures { + __u32 cmd; + __u32 size; + struct ethtool_set_features_block features[0]; +}; + +/* + * %ETHTOOL_SFEATURES changes features present in features[].valid to the + * values of corresponding bits in features[].requested. Bits in .requested + * not set in .valid or not changeable are ignored. + * + * Returns %EINVAL when .valid contains undefined or never-changable bits + * or size is not equal to required number of features words (32-bit blocks). + * Returns >= 0 if request was completed; bits set in the value mean: + * %ETHTOOL_F_UNSUPPORTED - there were bits set in .valid that are not + * changeable (not present in %ETHTOOL_GFEATURES' features[].available) + * those bits were ignored. + * %ETHTOOL_F_WISH - some or all changes requested were recorded but the + * resulting state of bits masked by .valid is not equal to .requested. + * Probably there are other device-specific constraints on some features + * in the set. When %ETHTOOL_F_UNSUPPORTED is set, .valid is considered + * here as though ignored bits were cleared. + * + * Meaning of bits in the masks are obtained by %ETHTOOL_GSSET_INFO (number of + * bits in the arrays - always multiple of 32) and %ETHTOOL_GSTRINGS commands + * for ETH_SS_FEATURES string set. First entry in the table corresponds to least + * significant bit in features[0] fields. Empty strings mark undefined features. + */ +enum ethtool_sfeatures_retval_bits { + ETHTOOL_F_UNSUPPORTED__BIT, + ETHTOOL_F_WISH__BIT, +}; + +#define ETHTOOL_F_UNSUPPORTED (1 << ETHTOOL_F_UNSUPPORTED__BIT) +#define ETHTOOL_F_WISH (1 << ETHTOOL_F_WISH__BIT) + #ifdef __KERNEL__ #include @@ -744,6 +826,9 @@ struct ethtool_ops { #define ETHTOOL_GRXFHINDIR 0x00000038 /* Get RX flow hash indir'n table */ #define ETHTOOL_SRXFHINDIR 0x00000039 /* Set RX flow hash indir'n table */ +#define ETHTOOL_GFEATURES 0x0000003a /* Get device offload settings */ +#define ETHTOOL_SFEATURES 0x0000003b /* Change device offload settings */ + /* compatibility with older code */ #define SPARC_ETH_GSET ETHTOOL_GSET #define SPARC_ETH_SSET ETHTOOL_SSET diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index dede3fdbb4b..85f67e225f6 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -791,6 +791,18 @@ struct netdev_tc_txq { * * int (*ndo_del_slave)(struct net_device *dev, struct net_device *slave_dev); * Called to release previously enslaved netdev. + * + * Feature/offload setting functions. + * u32 (*ndo_fix_features)(struct net_device *dev, u32 features); + * Adjusts the requested feature flags according to device-specific + * constraints, and returns the resulting flags. Must not modify + * the device state. + * + * int (*ndo_set_features)(struct net_device *dev, u32 features); + * Called to update device configuration to new features. Passed + * feature set might be less than what was returned by ndo_fix_features()). + * Must return >0 or -errno if it changed dev->features itself. + * */ #define HAVE_NET_DEVICE_OPS struct net_device_ops { @@ -874,6 +886,10 @@ struct net_device_ops { struct net_device *slave_dev); int (*ndo_del_slave)(struct net_device *dev, struct net_device *slave_dev); + u32 (*ndo_fix_features)(struct net_device *dev, + u32 features); + int (*ndo_set_features)(struct net_device *dev, + u32 features); }; /* @@ -925,12 +941,18 @@ struct net_device { struct list_head napi_list; struct list_head unreg_list; - /* Net device features */ + /* currently active device features */ u32 features; - + /* user-changeable features */ + u32 hw_features; + /* user-requested features */ + u32 wanted_features; /* VLAN feature mask */ u32 vlan_features; + /* Net device feature bits; if you change something, + * also update netdev_features_strings[] in ethtool.c */ + #define NETIF_F_SG 1 /* Scatter/gather IO. */ #define NETIF_F_IP_CSUM 2 /* Can checksum TCP/UDP over IPv4. */ #define NETIF_F_NO_CSUM 4 /* Does not require checksum. F.e. loopack. */ @@ -966,6 +988,12 @@ struct net_device { #define NETIF_F_TSO6 (SKB_GSO_TCPV6 << NETIF_F_GSO_SHIFT) #define NETIF_F_FSO (SKB_GSO_FCOE << NETIF_F_GSO_SHIFT) + /* Features valid for ethtool to change */ + /* = all defined minus driver/device-class-related */ +#define NETIF_F_NEVER_CHANGE (NETIF_F_HIGHDMA | NETIF_F_VLAN_CHALLENGED | \ + NETIF_F_LLTX | NETIF_F_NETNS_LOCAL) +#define NETIF_F_ETHTOOL_BITS (0x1f3fffff & ~NETIF_F_NEVER_CHANGE) + /* List of features with software fallbacks. */ #define NETIF_F_GSO_SOFTWARE (NETIF_F_TSO | NETIF_F_TSO_ECN | \ NETIF_F_TSO6 | NETIF_F_UFO) @@ -2428,8 +2456,13 @@ extern char *netdev_drivername(const struct net_device *dev, char *buffer, int l extern void linkwatch_run_queue(void); +static inline u32 netdev_get_wanted_features(struct net_device *dev) +{ + return (dev->features & ~dev->hw_features) | dev->wanted_features; +} u32 netdev_increment_features(u32 all, u32 one, u32 mask); u32 netdev_fix_features(struct net_device *dev, u32 features); +void netdev_update_features(struct net_device *dev); void netif_stacked_transfer_operstate(const struct net_device *rootdev, struct net_device *dev); diff --git a/net/core/dev.c b/net/core/dev.c index 8686f6ffe7f..4f6943928fe 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5302,6 +5302,37 @@ u32 netdev_fix_features(struct net_device *dev, u32 features) } EXPORT_SYMBOL(netdev_fix_features); +void netdev_update_features(struct net_device *dev) +{ + u32 features; + int err = 0; + + features = netdev_get_wanted_features(dev); + + if (dev->netdev_ops->ndo_fix_features) + features = dev->netdev_ops->ndo_fix_features(dev, features); + + /* driver might be less strict about feature dependencies */ + features = netdev_fix_features(dev, features); + + if (dev->features == features) + return; + + netdev_info(dev, "Features changed: 0x%08x -> 0x%08x\n", + dev->features, features); + + if (dev->netdev_ops->ndo_set_features) + err = dev->netdev_ops->ndo_set_features(dev, features); + + if (!err) + dev->features = features; + else if (err < 0) + netdev_err(dev, + "set_features() failed (%d); wanted 0x%08x, left 0x%08x\n", + err, features, dev->features); +} +EXPORT_SYMBOL(netdev_update_features); + /** * netif_stacked_transfer_operstate - transfer operstate * @rootdev: the root or lower level device to transfer state from @@ -5436,15 +5467,18 @@ int register_netdevice(struct net_device *dev) if (dev->iflink == -1) dev->iflink = dev->ifindex; - /* Enable software offloads by default - will be stripped in - * netdev_fix_features() if not supported. */ - dev->features |= NETIF_F_SOFT_FEATURES; + /* Transfer changeable features to wanted_features and enable + * software offloads (GSO and GRO). + */ + dev->hw_features |= NETIF_F_SOFT_FEATURES; + dev->wanted_features = (dev->features & dev->hw_features) + | NETIF_F_SOFT_FEATURES; /* Avoid warning from netdev_fix_features() for GSO without SG */ - if (!(dev->features & NETIF_F_SG)) - dev->features &= ~NETIF_F_GSO; + if (!(dev->wanted_features & NETIF_F_SG)) + dev->wanted_features &= ~NETIF_F_GSO; - dev->features = netdev_fix_features(dev, dev->features); + netdev_update_features(dev); /* Enable GRO and NETIF_F_HIGHDMA for vlans by default, * vlan_dev_init() will do the dev->features check, so these features diff --git a/net/core/ethtool.c b/net/core/ethtool.c index c3fb8f90de6..95773960dc7 100644 --- a/net/core/ethtool.c +++ b/net/core/ethtool.c @@ -172,10 +172,120 @@ EXPORT_SYMBOL(ethtool_ntuple_flush); /* Handlers for each ethtool command */ +#define ETHTOOL_DEV_FEATURE_WORDS 1 + +static int ethtool_get_features(struct net_device *dev, void __user *useraddr) +{ + struct ethtool_gfeatures cmd = { + .cmd = ETHTOOL_GFEATURES, + .size = ETHTOOL_DEV_FEATURE_WORDS, + }; + struct ethtool_get_features_block features[ETHTOOL_DEV_FEATURE_WORDS] = { + { + .available = dev->hw_features, + .requested = dev->wanted_features, + .active = dev->features, + .never_changed = NETIF_F_NEVER_CHANGE, + }, + }; + u32 __user *sizeaddr; + u32 copy_size; + + sizeaddr = useraddr + offsetof(struct ethtool_gfeatures, size); + if (get_user(copy_size, sizeaddr)) + return -EFAULT; + + if (copy_size > ETHTOOL_DEV_FEATURE_WORDS) + copy_size = ETHTOOL_DEV_FEATURE_WORDS; + + if (copy_to_user(useraddr, &cmd, sizeof(cmd))) + return -EFAULT; + useraddr += sizeof(cmd); + if (copy_to_user(useraddr, features, copy_size * sizeof(*features))) + return -EFAULT; + + return 0; +} + +static int ethtool_set_features(struct net_device *dev, void __user *useraddr) +{ + struct ethtool_sfeatures cmd; + struct ethtool_set_features_block features[ETHTOOL_DEV_FEATURE_WORDS]; + int ret = 0; + + if (copy_from_user(&cmd, useraddr, sizeof(cmd))) + return -EFAULT; + useraddr += sizeof(cmd); + + if (cmd.size != ETHTOOL_DEV_FEATURE_WORDS) + return -EINVAL; + + if (copy_from_user(features, useraddr, sizeof(features))) + return -EFAULT; + + if (features[0].valid & ~NETIF_F_ETHTOOL_BITS) + return -EINVAL; + + if (features[0].valid & ~dev->hw_features) { + features[0].valid &= dev->hw_features; + ret |= ETHTOOL_F_UNSUPPORTED; + } + + dev->wanted_features &= ~features[0].valid; + dev->wanted_features |= features[0].valid & features[0].requested; + netdev_update_features(dev); + + if ((dev->wanted_features ^ dev->features) & features[0].valid) + ret |= ETHTOOL_F_WISH; + + return ret; +} + +static const char netdev_features_strings[ETHTOOL_DEV_FEATURE_WORDS * 32][ETH_GSTRING_LEN] = { + /* NETIF_F_SG */ "tx-scatter-gather", + /* NETIF_F_IP_CSUM */ "tx-checksum-ipv4", + /* NETIF_F_NO_CSUM */ "tx-checksum-unneeded", + /* NETIF_F_HW_CSUM */ "tx-checksum-ip-generic", + /* NETIF_F_IPV6_CSUM */ "tx_checksum-ipv6", + /* NETIF_F_HIGHDMA */ "highdma", + /* NETIF_F_FRAGLIST */ "tx-scatter-gather-fraglist", + /* NETIF_F_HW_VLAN_TX */ "tx-vlan-hw-insert", + + /* NETIF_F_HW_VLAN_RX */ "rx-vlan-hw-parse", + /* NETIF_F_HW_VLAN_FILTER */ "rx-vlan-filter", + /* NETIF_F_VLAN_CHALLENGED */ "vlan-challenged", + /* NETIF_F_GSO */ "tx-generic-segmentation", + /* NETIF_F_LLTX */ "tx-lockless", + /* NETIF_F_NETNS_LOCAL */ "netns-local", + /* NETIF_F_GRO */ "rx-gro", + /* NETIF_F_LRO */ "rx-lro", + + /* NETIF_F_TSO */ "tx-tcp-segmentation", + /* NETIF_F_UFO */ "tx-udp-fragmentation", + /* NETIF_F_GSO_ROBUST */ "tx-gso-robust", + /* NETIF_F_TSO_ECN */ "tx-tcp-ecn-segmentation", + /* NETIF_F_TSO6 */ "tx-tcp6-segmentation", + /* NETIF_F_FSO */ "tx-fcoe-segmentation", + "", + "", + + /* NETIF_F_FCOE_CRC */ "tx-checksum-fcoe-crc", + /* NETIF_F_SCTP_CSUM */ "tx-checksum-sctp", + /* NETIF_F_FCOE_MTU */ "fcoe-mtu", + /* NETIF_F_NTUPLE */ "rx-ntuple-filter", + /* NETIF_F_RXHASH */ "rx-hashing", + "", + "", + "", +}; + static int __ethtool_get_sset_count(struct net_device *dev, int sset) { const struct ethtool_ops *ops = dev->ethtool_ops; + if (sset == ETH_SS_FEATURES) + return ARRAY_SIZE(netdev_features_strings); + if (ops && ops->get_sset_count && ops->get_strings) return ops->get_sset_count(dev, sset); else @@ -187,8 +297,12 @@ static void __ethtool_get_strings(struct net_device *dev, { const struct ethtool_ops *ops = dev->ethtool_ops; - /* ops->get_strings is valid because checked earlier */ - ops->get_strings(dev, stringset, data); + if (stringset == ETH_SS_FEATURES) + memcpy(data, netdev_features_strings, + sizeof(netdev_features_strings)); + else + /* ops->get_strings is valid because checked earlier */ + ops->get_strings(dev, stringset, data); } static u32 ethtool_get_feature_mask(u32 eth_cmd) @@ -1533,6 +1647,7 @@ int dev_ethtool(struct net *net, struct ifreq *ifr) case ETHTOOL_GRXCLSRLCNT: case ETHTOOL_GRXCLSRULE: case ETHTOOL_GRXCLSRLALL: + case ETHTOOL_GFEATURES: break; default: if (!capable(CAP_NET_ADMIN)) @@ -1678,6 +1793,12 @@ int dev_ethtool(struct net *net, struct ifreq *ifr) case ETHTOOL_SRXFHINDIR: rc = ethtool_set_rxfh_indir(dev, useraddr); break; + case ETHTOOL_GFEATURES: + rc = ethtool_get_features(dev, useraddr); + break; + case ETHTOOL_SFEATURES: + rc = ethtool_set_features(dev, useraddr); + break; case ETHTOOL_GTXCSUM: case ETHTOOL_GSG: case ETHTOOL_GTSO: -- cgit v1.2.3-70-g09d2 From 8e9b59b219e520cfc2f80af471c6b0e67ad9dd75 Mon Sep 17 00:00:00 2001 From: Michał Mirosław Date: Tue, 22 Feb 2011 16:52:28 +0000 Subject: Fix "(unregistered net_device): Features changed" message MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Fix netdev_update_features() messages on register time by moving the call further in register_netdevice(). When netdev->reg_state != NETREG_REGISTERED, netdev_name() returns "(unregistered netdevice)" even if the dev's name is already filled. Signed-off-by: Michał Mirosław Signed-off-by: David S. Miller --- net/core/dev.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 578415c1ef7..77e5edb724f 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5483,8 +5483,6 @@ int register_netdevice(struct net_device *dev) if (!(dev->wanted_features & NETIF_F_SG)) dev->wanted_features &= ~NETIF_F_GSO; - netdev_update_features(dev); - /* Enable GRO and NETIF_F_HIGHDMA for vlans by default, * vlan_dev_init() will do the dev->features check, so these features * are enabled only if supported by underlying device. @@ -5501,6 +5499,8 @@ int register_netdevice(struct net_device *dev) goto err_uninit; dev->reg_state = NETREG_REGISTERED; + netdev_update_features(dev); + /* * Default initial state at registry is that the * device is present. -- cgit v1.2.3-70-g09d2 From 14d1232f490c1c696582909fb3b69e67a8d38a34 Mon Sep 17 00:00:00 2001 From: Michał Mirosław Date: Tue, 22 Feb 2011 16:52:28 +0000 Subject: net: avoid initial "Features changed" message MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Avoid "Features changed" message and ndo_set_features call on device registration caused by automatic enabling of GSO and GRO. Driver should have enabled hardware offloads it set in features, so the ndo_set_features() is not needed at registration time. Signed-off-by: Michał Mirosław Signed-off-by: David S. Miller --- net/core/dev.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 77e5edb724f..69a3c0817d6 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5476,12 +5476,14 @@ int register_netdevice(struct net_device *dev) * software offloads (GSO and GRO). */ dev->hw_features |= NETIF_F_SOFT_FEATURES; - dev->wanted_features = (dev->features & dev->hw_features) - | NETIF_F_SOFT_FEATURES; + dev->features |= NETIF_F_SOFT_FEATURES; + dev->wanted_features = dev->features & dev->hw_features; /* Avoid warning from netdev_fix_features() for GSO without SG */ - if (!(dev->wanted_features & NETIF_F_SG)) + if (!(dev->wanted_features & NETIF_F_SG)) { dev->wanted_features &= ~NETIF_F_GSO; + dev->features &= ~NETIF_F_GSO; + } /* Enable GRO and NETIF_F_HIGHDMA for vlans by default, * vlan_dev_init() will do the dev->features check, so these features -- cgit v1.2.3-70-g09d2 From 63d8ea7f93e1fb9d1aa9509ab3e1a71199245c80 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Mon, 28 Feb 2011 10:48:59 -0800 Subject: net: Forgot to commit net/core/dev.c part of Jiri's ->rx_handler patch. Signed-off-by: David S. Miller --- net/core/dev.c | 119 +++++++++++++++------------------------------------------ 1 file changed, 31 insertions(+), 88 deletions(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 69a3c0817d6..30440e7b296 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3096,63 +3096,31 @@ void netdev_rx_handler_unregister(struct net_device *dev) } EXPORT_SYMBOL_GPL(netdev_rx_handler_unregister); -static inline void skb_bond_set_mac_by_master(struct sk_buff *skb, - struct net_device *master) +static void vlan_on_bond_hook(struct sk_buff *skb) { - if (skb->pkt_type == PACKET_HOST) { - u16 *dest = (u16 *) eth_hdr(skb)->h_dest; + /* + * Make sure ARP frames received on VLAN interfaces stacked on + * bonding interfaces still make their way to any base bonding + * device that may have registered for a specific ptype. + */ + if (skb->dev->priv_flags & IFF_802_1Q_VLAN && + vlan_dev_real_dev(skb->dev)->priv_flags & IFF_BONDING && + skb->protocol == htons(ETH_P_ARP)) { + struct sk_buff *skb2 = skb_clone(skb, GFP_ATOMIC); - memcpy(dest, master->dev_addr, ETH_ALEN); + if (!skb2) + return; + skb2->dev = vlan_dev_real_dev(skb->dev); + netif_rx(skb2); } } -/* On bonding slaves other than the currently active slave, suppress - * duplicates except for 802.3ad ETH_P_SLOW, alb non-mcast/bcast, and - * ARP on active-backup slaves with arp_validate enabled. - */ -static int __skb_bond_should_drop(struct sk_buff *skb, - struct net_device *master) -{ - struct net_device *dev = skb->dev; - - if (master->priv_flags & IFF_MASTER_ARPMON) - dev->last_rx = jiffies; - - if ((master->priv_flags & IFF_MASTER_ALB) && - (master->priv_flags & IFF_BRIDGE_PORT)) { - /* Do address unmangle. The local destination address - * will be always the one master has. Provides the right - * functionality in a bridge. - */ - skb_bond_set_mac_by_master(skb, master); - } - - if (dev->priv_flags & IFF_SLAVE_INACTIVE) { - if ((dev->priv_flags & IFF_SLAVE_NEEDARP) && - skb->protocol == __cpu_to_be16(ETH_P_ARP)) - return 0; - - if (master->priv_flags & IFF_MASTER_ALB) { - if (skb->pkt_type != PACKET_BROADCAST && - skb->pkt_type != PACKET_MULTICAST) - return 0; - } - if (master->priv_flags & IFF_MASTER_8023AD && - skb->protocol == __cpu_to_be16(ETH_P_SLOW)) - return 0; - - return 1; - } - return 0; -} - static int __netif_receive_skb(struct sk_buff *skb) { struct packet_type *ptype, *pt_prev; rx_handler_func_t *rx_handler; struct net_device *orig_dev; - struct net_device *null_or_orig; - struct net_device *orig_or_bond; + struct net_device *null_or_dev; int ret = NET_RX_DROP; __be16 type; @@ -3167,32 +3135,8 @@ static int __netif_receive_skb(struct sk_buff *skb) if (!skb->skb_iif) skb->skb_iif = skb->dev->ifindex; - - /* - * bonding note: skbs received on inactive slaves should only - * be delivered to pkt handlers that are exact matches. Also - * the deliver_no_wcard flag will be set. If packet handlers - * are sensitive to duplicate packets these skbs will need to - * be dropped at the handler. - */ - null_or_orig = NULL; orig_dev = skb->dev; - if (skb->deliver_no_wcard) - null_or_orig = orig_dev; - else if (netif_is_bond_slave(orig_dev)) { - struct net_device *bond_master = ACCESS_ONCE(orig_dev->master); - - if (likely(bond_master)) { - if (__skb_bond_should_drop(skb, bond_master)) { - skb->deliver_no_wcard = 1; - /* deliver only exact match */ - null_or_orig = orig_dev; - } else - skb->dev = bond_master; - } - } - __this_cpu_inc(softnet_data.processed); skb_reset_network_header(skb); skb_reset_transport_header(skb); skb->mac_len = skb->network_header - skb->mac_header; @@ -3201,6 +3145,10 @@ static int __netif_receive_skb(struct sk_buff *skb) rcu_read_lock(); +another_round: + + __this_cpu_inc(softnet_data.processed); + #ifdef CONFIG_NET_CLS_ACT if (skb->tc_verd & TC_NCLS) { skb->tc_verd = CLR_TC_NCLS(skb->tc_verd); @@ -3209,8 +3157,7 @@ static int __netif_receive_skb(struct sk_buff *skb) #endif list_for_each_entry_rcu(ptype, &ptype_all, list) { - if (ptype->dev == null_or_orig || ptype->dev == skb->dev || - ptype->dev == orig_dev) { + if (!ptype->dev || ptype->dev == skb->dev) { if (pt_prev) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; @@ -3224,16 +3171,20 @@ static int __netif_receive_skb(struct sk_buff *skb) ncls: #endif - /* Handle special case of bridge or macvlan */ rx_handler = rcu_dereference(skb->dev->rx_handler); if (rx_handler) { + struct net_device *prev_dev; + if (pt_prev) { ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = NULL; } + prev_dev = skb->dev; skb = rx_handler(skb); if (!skb) goto out; + if (skb->dev != prev_dev) + goto another_round; } if (vlan_tx_tag_present(skb)) { @@ -3248,24 +3199,16 @@ ncls: goto out; } - /* - * Make sure frames received on VLAN interfaces stacked on - * bonding interfaces still make their way to any base bonding - * device that may have registered for a specific ptype. The - * handler may have to adjust skb->dev and orig_dev. - */ - orig_or_bond = orig_dev; - if ((skb->dev->priv_flags & IFF_802_1Q_VLAN) && - (vlan_dev_real_dev(skb->dev)->priv_flags & IFF_BONDING)) { - orig_or_bond = vlan_dev_real_dev(skb->dev); - } + vlan_on_bond_hook(skb); + + /* deliver only exact match when indicated */ + null_or_dev = skb->deliver_no_wcard ? skb->dev : NULL; type = skb->protocol; list_for_each_entry_rcu(ptype, &ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) { - if (ptype->type == type && (ptype->dev == null_or_orig || - ptype->dev == skb->dev || ptype->dev == orig_dev || - ptype->dev == orig_or_bond)) { + if (ptype->type == type && + (ptype->dev == null_or_dev || ptype->dev == skb->dev)) { if (pt_prev) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; -- cgit v1.2.3-70-g09d2 From e3f48d37cf87a4a94e9f05fddc39b0e5f2307c27 Mon Sep 17 00:00:00 2001 From: Jiri Pirko Date: Mon, 28 Feb 2011 20:26:31 +0000 Subject: net: allow handlers to be processed for orig_dev MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This was there before, I forgot about this. Allows deliveries to ptype_base handlers registered for orig_dev. I presume this is still desired. Signed-off-by: Jiri Pirko Reviewed-by: Nicolas de Pesloüan Signed-off-by: David S. Miller --- net/core/dev.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'net/core/dev.c') diff --git a/net/core/dev.c b/net/core/dev.c index 30440e7b296..9f66de9c057 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3208,7 +3208,8 @@ ncls: list_for_each_entry_rcu(ptype, &ptype_base[ntohs(type) & PTYPE_HASH_MASK], list) { if (ptype->type == type && - (ptype->dev == null_or_dev || ptype->dev == skb->dev)) { + (ptype->dev == null_or_dev || ptype->dev == skb->dev || + ptype->dev == orig_dev)) { if (pt_prev) ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = ptype; -- cgit v1.2.3-70-g09d2 From 8909c9ad8ff03611c9c96c9a92656213e4bb495b Mon Sep 17 00:00:00 2001 From: Vasiliy Kulikov Date: Wed, 2 Mar 2011 00:33:13 +0300 Subject: net: don't allow CAP_NET_ADMIN to load non-netdev kernel modules Since a8f80e8ff94ecba629542d9b4b5f5a8ee3eb565c any process with CAP_NET_ADMIN may load any module from /lib/modules/. This doesn't mean that CAP_NET_ADMIN is a superset of CAP_SYS_MODULE as modules are limited to /lib/modules/**. However, CAP_NET_ADMIN capability shouldn't allow anybody load any module not related to networking. This patch restricts an ability of autoloading modules to netdev modules with explicit aliases. This fixes CVE-2011-1019. Arnd Bergmann suggested to leave untouched the old pre-v2.6.32 behavior of loading netdev modules by name (without any prefix) for processes with CAP_SYS_MODULE to maintain the compatibility with network scripts that use autoloading netdev modules by aliases like "eth0", "wlan0". Currently there are only three users of the feature in the upstream kernel: ipip, ip_gre and sit. root@albatros:~# capsh --drop=$(seq -s, 0 11),$(seq -s, 13 34) -- root@albatros:~# grep Cap /proc/$$/status CapInh: 0000000000000000 CapPrm: fffffff800001000 CapEff: fffffff800001000 CapBnd: fffffff800001000 root@albatros:~# modprobe xfs FATAL: Error inserting xfs (/lib/modules/2.6.38-rc6-00001-g2bf4ca3/kernel/fs/xfs/xfs.ko): Operation not permitted root@albatros:~# lsmod | grep xfs root@albatros:~# ifconfig xfs xfs: error fetching interface information: Device not found root@albatros:~# lsmod | grep xfs root@albatros:~# lsmod | grep sit root@albatros:~# ifconfig sit sit: error fetching interface information: Device not found root@albatros:~# lsmod | grep sit root@albatros:~# ifconfig sit0 sit0 Link encap:IPv6-in-IPv4 NOARP MTU:1480 Metric:1 root@albatros:~# lsmod | grep sit sit 10457 0 tunnel4 2957 1 sit For CAP_SYS_MODULE module loading is still relaxed: root@albatros:~# grep Cap /proc/$$/status CapInh: 0000000000000000 CapPrm: ffffffffffffffff CapEff: ffffffffffffffff CapBnd: ffffffffffffffff root@albatros:~# ifconfig xfs xfs: error fetching interface information: Device not found root@albatros:~# lsmod | grep xfs xfs 745319 0 Reference: https://lkml.org/lkml/2011/2/24/203 Signed-off-by: Vasiliy Kulikov Signed-off-by: Michael Tokarev Acked-by: David S. Miller Acked-by: Kees Cook Signed-off-by: James Morris --- include/linux/netdevice.h | 3 +++ net/core/dev.c | 12 ++++++++++-- net/ipv4/ip_gre.c | 2 +- net/ipv4/ipip.c | 2 +- net/ipv6/sit.c | 2 +- 5 files changed, 16 insertions(+), 5 deletions(-) (limited to 'net/core/dev.c') diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index d971346b034..71caf7a5e6c 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -2392,6 +2392,9 @@ extern int netdev_notice(const struct net_device *dev, const char *format, ...) extern int netdev_info(const struct net_device *dev, const char *format, ...) __attribute__ ((format (printf, 2, 3))); +#define MODULE_ALIAS_NETDEV(device) \ + MODULE_ALIAS("netdev-" device) + #if defined(DEBUG) #define netdev_dbg(__dev, format, args...) \ netdev_printk(KERN_DEBUG, __dev, format, ##args) diff --git a/net/core/dev.c b/net/core/dev.c index 8ae6631abcc..6561021d22d 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -1114,13 +1114,21 @@ EXPORT_SYMBOL(netdev_bonding_change); void dev_load(struct net *net, const char *name) { struct net_device *dev; + int no_module; rcu_read_lock(); dev = dev_get_by_name_rcu(net, name); rcu_read_unlock(); - if (!dev && capable(CAP_NET_ADMIN)) - request_module("%s", name); + no_module = !dev; + if (no_module && capable(CAP_NET_ADMIN)) + no_module = request_module("netdev-%s", name); + if (no_module && capable(CAP_SYS_MODULE)) { + if (!request_module("%s", name)) + pr_err("Loading kernel module for a network device " +"with CAP_SYS_MODULE (deprecated). Use CAP_NET_ADMIN and alias netdev-%s " +"instead\n", name); + } } EXPORT_SYMBOL(dev_load); diff --git a/net/ipv4/ip_gre.c b/net/ipv4/ip_gre.c index 6613edfac28..d1d0e2c256f 100644 --- a/net/ipv4/ip_gre.c +++ b/net/ipv4/ip_gre.c @@ -1765,4 +1765,4 @@ module_exit(ipgre_fini); MODULE_LICENSE("GPL"); MODULE_ALIAS_RTNL_LINK("gre"); MODULE_ALIAS_RTNL_LINK("gretap"); -MODULE_ALIAS("gre0"); +MODULE_ALIAS_NETDEV("gre0"); diff --git a/net/ipv4/ipip.c b/net/ipv4/ipip.c index 988f52fba54..a5f58e7cbb2 100644 --- a/net/ipv4/ipip.c +++ b/net/ipv4/ipip.c @@ -913,4 +913,4 @@ static void __exit ipip_fini(void) module_init(ipip_init); module_exit(ipip_fini); MODULE_LICENSE("GPL"); -MODULE_ALIAS("tunl0"); +MODULE_ALIAS_NETDEV("tunl0"); diff --git a/net/ipv6/sit.c b/net/ipv6/sit.c index 8ce38f10a54..d2c16e10f65 100644 --- a/net/ipv6/sit.c +++ b/net/ipv6/sit.c @@ -1290,4 +1290,4 @@ static int __init sit_init(void) module_init(sit_init); module_exit(sit_cleanup); MODULE_LICENSE("GPL"); -MODULE_ALIAS("sit0"); +MODULE_ALIAS_NETDEV("sit0"); -- cgit v1.2.3-70-g09d2 From 8a4eb5734e8d1dc60a8c28576bbbdfdcc643626d Mon Sep 17 00:00:00 2001 From: Jiri Pirko Date: Sat, 12 Mar 2011 03:14:39 +0000 Subject: net: introduce rx_handler results and logic around that MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit This patch allows rx_handlers to better signalize what to do next to it's caller. That makes skb->deliver_no_wcard no longer needed. kernel-doc for rx_handler_result is taken from Nicolas' patch. Signed-off-by: Jiri Pirko Reviewed-by: Nicolas de Pesloüan Signed-off-by: David S. Miller --- drivers/net/bonding/bond_main.c | 22 +++++++++--------- drivers/net/macvlan.c | 11 ++++----- include/linux/netdevice.h | 50 ++++++++++++++++++++++++++++++++++++++++- include/linux/skbuff.h | 5 +---- net/bridge/br_input.c | 25 ++++++++++++--------- net/bridge/br_private.h | 2 +- net/core/dev.c | 21 +++++++++++------ net/core/skbuff.c | 1 - 8 files changed, 98 insertions(+), 39 deletions(-) (limited to 'net/core/dev.c') diff --git a/drivers/net/bonding/bond_main.c b/drivers/net/bonding/bond_main.c index 04119b1e7cd..27c413aa15d 100644 --- a/drivers/net/bonding/bond_main.c +++ b/drivers/net/bonding/bond_main.c @@ -1480,20 +1480,23 @@ static bool bond_should_deliver_exact_match(struct sk_buff *skb, return false; } -static struct sk_buff *bond_handle_frame(struct sk_buff *skb) +static rx_handler_result_t bond_handle_frame(struct sk_buff **pskb) { + struct sk_buff *skb = *pskb; struct slave *slave; struct net_device *bond_dev; struct bonding *bond; - skb = skb_share_check(skb, GFP_ATOMIC); - if (unlikely(!skb)) - return NULL; - slave = bond_slave_get_rcu(skb->dev); bond_dev = ACCESS_ONCE(slave->dev->master); if (unlikely(!bond_dev)) - return skb; + return RX_HANDLER_PASS; + + skb = skb_share_check(skb, GFP_ATOMIC); + if (unlikely(!skb)) + return RX_HANDLER_CONSUMED; + + *pskb = skb; bond = netdev_priv(bond_dev); @@ -1501,8 +1504,7 @@ static struct sk_buff *bond_handle_frame(struct sk_buff *skb) slave->dev->last_rx = jiffies; if (bond_should_deliver_exact_match(skb, slave, bond)) { - skb->deliver_no_wcard = 1; - return skb; + return RX_HANDLER_EXACT; } skb->dev = bond_dev; @@ -1514,12 +1516,12 @@ static struct sk_buff *bond_handle_frame(struct sk_buff *skb) if (unlikely(skb_cow_head(skb, skb->data - skb_mac_header(skb)))) { kfree_skb(skb); - return NULL; + return RX_HANDLER_CONSUMED; } memcpy(eth_hdr(skb)->h_dest, bond_dev->dev_addr, ETH_ALEN); } - return skb; + return RX_HANDLER_ANOTHER; } /* enslave device to bond device */ diff --git a/drivers/net/macvlan.c b/drivers/net/macvlan.c index 497991bd3b6..5b37d3c191e 100644 --- a/drivers/net/macvlan.c +++ b/drivers/net/macvlan.c @@ -152,9 +152,10 @@ static void macvlan_broadcast(struct sk_buff *skb, } /* called under rcu_read_lock() from netif_receive_skb */ -static struct sk_buff *macvlan_handle_frame(struct sk_buff *skb) +static rx_handler_result_t macvlan_handle_frame(struct sk_buff **pskb) { struct macvlan_port *port; + struct sk_buff *skb = *pskb; const struct ethhdr *eth = eth_hdr(skb); const struct macvlan_dev *vlan; const struct macvlan_dev *src; @@ -184,7 +185,7 @@ static struct sk_buff *macvlan_handle_frame(struct sk_buff *skb) */ macvlan_broadcast(skb, port, src->dev, MACVLAN_MODE_VEPA); - return skb; + return RX_HANDLER_PASS; } if (port->passthru) @@ -192,12 +193,12 @@ static struct sk_buff *macvlan_handle_frame(struct sk_buff *skb) else vlan = macvlan_hash_lookup(port, eth->h_dest); if (vlan == NULL) - return skb; + return RX_HANDLER_PASS; dev = vlan->dev; if (unlikely(!(dev->flags & IFF_UP))) { kfree_skb(skb); - return NULL; + return RX_HANDLER_CONSUMED; } len = skb->len + ETH_HLEN; skb = skb_share_check(skb, GFP_ATOMIC); @@ -211,7 +212,7 @@ static struct sk_buff *macvlan_handle_frame(struct sk_buff *skb) out: macvlan_count_rx(vlan, len, ret == NET_RX_SUCCESS, 0); - return NULL; + return RX_HANDLER_CONSUMED; } static int macvlan_queue_xmit(struct sk_buff *skb, struct net_device *dev) diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h index 604dbf5051d..5eeb2cd3631 100644 --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -390,7 +390,55 @@ enum gro_result { }; typedef enum gro_result gro_result_t; -typedef struct sk_buff *rx_handler_func_t(struct sk_buff *skb); +/* + * enum rx_handler_result - Possible return values for rx_handlers. + * @RX_HANDLER_CONSUMED: skb was consumed by rx_handler, do not process it + * further. + * @RX_HANDLER_ANOTHER: Do another round in receive path. This is indicated in + * case skb->dev was changed by rx_handler. + * @RX_HANDLER_EXACT: Force exact delivery, no wildcard. + * @RX_HANDLER_PASS: Do nothing, passe the skb as if no rx_handler was called. + * + * rx_handlers are functions called from inside __netif_receive_skb(), to do + * special processing of the skb, prior to delivery to protocol handlers. + * + * Currently, a net_device can only have a single rx_handler registered. Trying + * to register a second rx_handler will return -EBUSY. + * + * To register a rx_handler on a net_device, use netdev_rx_handler_register(). + * To unregister a rx_handler on a net_device, use + * netdev_rx_handler_unregister(). + * + * Upon return, rx_handler is expected to tell __netif_receive_skb() what to + * do with the skb. + * + * If the rx_handler consumed to skb in some way, it should return + * RX_HANDLER_CONSUMED. This is appropriate when the rx_handler arranged for + * the skb to be delivered in some other ways. + * + * If the rx_handler changed skb->dev, to divert the skb to another + * net_device, it should return RX_HANDLER_ANOTHER. The rx_handler for the + * new device will be called if it exists. + * + * If the rx_handler consider the skb should be ignored, it should return + * RX_HANDLER_EXACT. The skb will only be delivered to protocol handlers that + * are registred on exact device (ptype->dev == skb->dev). + * + * If the rx_handler didn't changed skb->dev, but want the skb to be normally + * delivered, it should return RX_HANDLER_PASS. + * + * A device without a registered rx_handler will behave as if rx_handler + * returned RX_HANDLER_PASS. + */ + +enum rx_handler_result { + RX_HANDLER_CONSUMED, + RX_HANDLER_ANOTHER, + RX_HANDLER_EXACT, + RX_HANDLER_PASS, +}; +typedef enum rx_handler_result rx_handler_result_t; +typedef rx_handler_result_t rx_handler_func_t(struct sk_buff **pskb); extern void __napi_schedule(struct napi_struct *n); diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h index 31f02d0b46a..24cfa626931 100644 --- a/include/linux/skbuff.h +++ b/include/linux/skbuff.h @@ -388,10 +388,7 @@ struct sk_buff { kmemcheck_bitfield_begin(flags2); __u16 queue_mapping:16; #ifdef CONFIG_IPV6_NDISC_NODETYPE - __u8 ndisc_nodetype:2, - deliver_no_wcard:1; -#else - __u8 deliver_no_wcard:1; + __u8 ndisc_nodetype:2; #endif __u8 ooo_okay:1; kmemcheck_bitfield_end(flags2); diff --git a/net/bridge/br_input.c b/net/bridge/br_input.c index 88e4aa9cb1f..e2160792e1b 100644 --- a/net/bridge/br_input.c +++ b/net/bridge/br_input.c @@ -139,21 +139,22 @@ static inline int is_link_local(const unsigned char *dest) * Return NULL if skb is handled * note: already called with rcu_read_lock */ -struct sk_buff *br_handle_frame(struct sk_buff *skb) +rx_handler_result_t br_handle_frame(struct sk_buff **pskb) { struct net_bridge_port *p; + struct sk_buff *skb = *pskb; const unsigned char *dest = eth_hdr(skb)->h_dest; br_should_route_hook_t *rhook; if (unlikely(skb->pkt_type == PACKET_LOOPBACK)) - return skb; + return RX_HANDLER_PASS; if (!is_valid_ether_addr(eth_hdr(skb)->h_source)) goto drop; skb = skb_share_check(skb, GFP_ATOMIC); if (!skb) - return NULL; + return RX_HANDLER_CONSUMED; p = br_port_get_rcu(skb->dev); @@ -167,10 +168,12 @@ struct sk_buff *br_handle_frame(struct sk_buff *skb) goto forward; if (NF_HOOK(NFPROTO_BRIDGE, NF_BR_LOCAL_IN, skb, skb->dev, - NULL, br_handle_local_finish)) - return NULL; /* frame consumed by filter */ - else - return skb; /* continue processing */ + NULL, br_handle_local_finish)) { + return RX_HANDLER_CONSUMED; /* consumed by filter */ + } else { + *pskb = skb; + return RX_HANDLER_PASS; /* continue processing */ + } } forward: @@ -178,8 +181,10 @@ forward: case BR_STATE_FORWARDING: rhook = rcu_dereference(br_should_route_hook); if (rhook) { - if ((*rhook)(skb)) - return skb; + if ((*rhook)(skb)) { + *pskb = skb; + return RX_HANDLER_PASS; + } dest = eth_hdr(skb)->h_dest; } /* fall through */ @@ -194,5 +199,5 @@ forward: drop: kfree_skb(skb); } - return NULL; + return RX_HANDLER_CONSUMED; } diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index f7afc364d77..19e2f46ed08 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -379,7 +379,7 @@ extern void br_features_recompute(struct net_bridge *br); /* br_input.c */ extern int br_handle_frame_finish(struct sk_buff *skb); -extern struct sk_buff *br_handle_frame(struct sk_buff *skb); +extern rx_handler_result_t br_handle_frame(struct sk_buff **pskb); /* br_ioctl.c */ extern int br_dev_ioctl(struct net_device *dev, struct ifreq *rq, int cmd); diff --git a/net/core/dev.c b/net/core/dev.c index 0d39032e962..0b88eba97da 100644 --- a/net/core/dev.c +++ b/net/core/dev.c @@ -3070,6 +3070,8 @@ out: * on a failure. * * The caller must hold the rtnl_mutex. + * + * For a general description of rx_handler, see enum rx_handler_result. */ int netdev_rx_handler_register(struct net_device *dev, rx_handler_func_t *rx_handler, @@ -3129,6 +3131,7 @@ static int __netif_receive_skb(struct sk_buff *skb) rx_handler_func_t *rx_handler; struct net_device *orig_dev; struct net_device *null_or_dev; + bool deliver_exact = false; int ret = NET_RX_DROP; __be16 type; @@ -3181,18 +3184,22 @@ ncls: rx_handler = rcu_dereference(skb->dev->rx_handler); if (rx_handler) { - struct net_device *prev_dev; - if (pt_prev) { ret = deliver_skb(skb, pt_prev, orig_dev); pt_prev = NULL; } - prev_dev = skb->dev; - skb = rx_handler(skb); - if (!skb) + switch (rx_handler(&skb)) { + case RX_HANDLER_CONSUMED: goto out; - if (skb->dev != prev_dev) + case RX_HANDLER_ANOTHER: goto another_round; + case RX_HANDLER_EXACT: + deliver_exact = true; + case RX_HANDLER_PASS: + break; + default: + BUG(); + } } if (vlan_tx_tag_present(skb)) { @@ -3210,7 +3217,7 @@ ncls: vlan_on_bond_hook(skb); /* deliver only exact match when indicated */ - null_or_dev = skb->deliver_no_wcard ? skb->dev : NULL; + null_or_dev = deliver_exact ? skb->dev : NULL; type = skb->protocol; list_for_each_entry_rcu(ptype, diff --git a/net/core/skbuff.c b/net/core/skbuff.c index 1eb526a848f..801dd08908f 100644 --- a/net/core/skbuff.c +++ b/net/core/skbuff.c @@ -523,7 +523,6 @@ static void __copy_skb_header(struct sk_buff *new, const struct sk_buff *old) new->ip_summed = old->ip_summed; skb_copy_queue_mapping(new, old); new->priority = old->priority; - new->deliver_no_wcard = old->deliver_no_wcard; #if defined(CONFIG_IP_VS) || defined(CONFIG_IP_VS_MODULE) new->ipvs_property = old->ipvs_property; #endif -- cgit v1.2.3-70-g09d2