diff options
Diffstat (limited to 'net/packet/af_packet.c')
-rw-r--r-- | net/packet/af_packet.c | 157 |
1 files changed, 103 insertions, 54 deletions
diff --git a/net/packet/af_packet.c b/net/packet/af_packet.c index bad144dfabc5..9bf81352c66c 100644 --- a/net/packet/af_packet.c +++ b/net/packet/af_packet.c @@ -57,6 +57,7 @@ #include <linux/if_packet.h> #include <linux/wireless.h> #include <linux/kernel.h> +#include <linux/delay.h> #include <linux/kmod.h> #include <linux/slab.h> #include <linux/vmalloc.h> @@ -384,7 +385,7 @@ static void __packet_set_status(struct packet_sock *po, void *frame, int status) smp_wmb(); } -static int __packet_get_status(struct packet_sock *po, void *frame) +static int __packet_get_status(const struct packet_sock *po, void *frame) { union tpacket_uhdr h; @@ -460,10 +461,10 @@ static __u32 __packet_set_timestamp(struct packet_sock *po, void *frame, return ts_status; } -static void *packet_lookup_frame(struct packet_sock *po, - struct packet_ring_buffer *rb, - unsigned int position, - int status) +static void *packet_lookup_frame(const struct packet_sock *po, + const struct packet_ring_buffer *rb, + unsigned int position, + int status) { unsigned int pg_vec_pos, frame_offset; union tpacket_uhdr h; @@ -544,7 +545,8 @@ static int prb_calc_retire_blk_tmo(struct packet_sock *po, msec = 1; div = ecmd.base.speed / 1000; } - } + } else + return DEFAULT_PRB_RETIRE_TOV; mbits = (blk_size_in_bytes * 8) / (1024 * 1024); @@ -659,7 +661,7 @@ static void prb_retire_rx_blk_timer_expired(struct timer_list *t) if (BLOCK_NUM_PKTS(pbd)) { while (atomic_read(&pkc->blk_fill_in_prog)) { /* Waiting for skb_copy_bits to finish... */ - cpu_relax(); + cpu_chill(); } } @@ -758,7 +760,7 @@ static void prb_close_block(struct tpacket_kbdq_core *pkc1, struct tpacket_hdr_v1 *h1 = &pbd1->hdr.bh1; struct sock *sk = &po->sk; - if (po->stats.stats3.tp_drops) + if (atomic_read(&po->tp_drops)) status |= TP_STATUS_LOSING; last_pkt = (struct tpacket3_hdr *)pkc1->prev; @@ -921,7 +923,7 @@ static void prb_retire_current_block(struct tpacket_kbdq_core *pkc, if (!(status & TP_STATUS_BLK_TMO)) { while (atomic_read(&pkc->blk_fill_in_prog)) { /* Waiting for skb_copy_bits to finish... */ - cpu_relax(); + cpu_chill(); } } prb_close_block(pkc, pbd, po, status); @@ -1083,10 +1085,10 @@ static void *packet_current_rx_frame(struct packet_sock *po, } } -static void *prb_lookup_block(struct packet_sock *po, - struct packet_ring_buffer *rb, - unsigned int idx, - int status) +static void *prb_lookup_block(const struct packet_sock *po, + const struct packet_ring_buffer *rb, + unsigned int idx, + int status) { struct tpacket_kbdq_core *pkc = GET_PBDQC_FROM_RB(rb); struct tpacket_block_desc *pbd = GET_PBLOCK_DESC(pkc, idx); @@ -1199,12 +1201,12 @@ static void packet_free_pending(struct packet_sock *po) #define ROOM_LOW 0x1 #define ROOM_NORMAL 0x2 -static bool __tpacket_has_room(struct packet_sock *po, int pow_off) +static bool __tpacket_has_room(const struct packet_sock *po, int pow_off) { int idx, len; - len = po->rx_ring.frame_max + 1; - idx = po->rx_ring.head; + len = READ_ONCE(po->rx_ring.frame_max) + 1; + idx = READ_ONCE(po->rx_ring.head); if (pow_off) idx += len >> pow_off; if (idx >= len) @@ -1212,12 +1214,12 @@ static bool __tpacket_has_room(struct packet_sock *po, int pow_off) return packet_lookup_frame(po, &po->rx_ring, idx, TP_STATUS_KERNEL); } -static bool __tpacket_v3_has_room(struct packet_sock *po, int pow_off) +static bool __tpacket_v3_has_room(const struct packet_sock *po, int pow_off) { int idx, len; - len = po->rx_ring.prb_bdqc.knum_blocks; - idx = po->rx_ring.prb_bdqc.kactive_blk_num; + len = READ_ONCE(po->rx_ring.prb_bdqc.knum_blocks); + idx = READ_ONCE(po->rx_ring.prb_bdqc.kactive_blk_num); if (pow_off) idx += len >> pow_off; if (idx >= len) @@ -1225,15 +1227,18 @@ static bool __tpacket_v3_has_room(struct packet_sock *po, int pow_off) return prb_lookup_block(po, &po->rx_ring, idx, TP_STATUS_KERNEL); } -static int __packet_rcv_has_room(struct packet_sock *po, struct sk_buff *skb) +static int __packet_rcv_has_room(const struct packet_sock *po, + const struct sk_buff *skb) { - struct sock *sk = &po->sk; + const struct sock *sk = &po->sk; int ret = ROOM_NONE; if (po->prot_hook.func != tpacket_rcv) { - int avail = sk->sk_rcvbuf - atomic_read(&sk->sk_rmem_alloc) - - (skb ? skb->truesize : 0); - if (avail > (sk->sk_rcvbuf >> ROOM_POW_OFF)) + int rcvbuf = READ_ONCE(sk->sk_rcvbuf); + int avail = rcvbuf - atomic_read(&sk->sk_rmem_alloc) + - (skb ? skb->truesize : 0); + + if (avail > (rcvbuf >> ROOM_POW_OFF)) return ROOM_NORMAL; else if (avail > 0) return ROOM_LOW; @@ -1258,19 +1263,24 @@ static int __packet_rcv_has_room(struct packet_sock *po, struct sk_buff *skb) static int packet_rcv_has_room(struct packet_sock *po, struct sk_buff *skb) { - int ret; - bool has_room; + int pressure, ret; - spin_lock_bh(&po->sk.sk_receive_queue.lock); ret = __packet_rcv_has_room(po, skb); - has_room = ret == ROOM_NORMAL; - if (po->pressure == has_room) - po->pressure = !has_room; - spin_unlock_bh(&po->sk.sk_receive_queue.lock); + pressure = ret != ROOM_NORMAL; + + if (READ_ONCE(po->pressure) != pressure) + WRITE_ONCE(po->pressure, pressure); return ret; } +static void packet_rcv_try_clear_pressure(struct packet_sock *po) +{ + if (READ_ONCE(po->pressure) && + __packet_rcv_has_room(po, NULL) == ROOM_NORMAL) + WRITE_ONCE(po->pressure, 0); +} + static void packet_sock_destruct(struct sock *sk) { skb_queue_purge(&sk->sk_error_queue); @@ -1288,15 +1298,21 @@ static void packet_sock_destruct(struct sock *sk) static bool fanout_flow_is_huge(struct packet_sock *po, struct sk_buff *skb) { - u32 rxhash; + u32 *history = po->rollover->history; + u32 victim, rxhash; int i, count = 0; rxhash = skb_get_hash(skb); for (i = 0; i < ROLLOVER_HLEN; i++) - if (po->rollover->history[i] == rxhash) + if (READ_ONCE(history[i]) == rxhash) count++; - po->rollover->history[prandom_u32() % ROLLOVER_HLEN] = rxhash; + victim = prandom_u32() % ROLLOVER_HLEN; + + /* Avoid dirtying the cache line if possible */ + if (READ_ONCE(history[victim]) != rxhash) + WRITE_ONCE(history[victim], rxhash); + return count > (ROLLOVER_HLEN >> 1); } @@ -1351,7 +1367,7 @@ static unsigned int fanout_demux_rollover(struct packet_fanout *f, i = j = min_t(int, po->rollover->sock, num - 1); do { po_next = pkt_sk(f->arr[i]); - if (po_next != po_skip && !po_next->pressure && + if (po_next != po_skip && !READ_ONCE(po_next->pressure) && packet_rcv_has_room(po_next, skb) == ROOM_NORMAL) { if (i != j) po->rollover->sock = i; @@ -2126,10 +2142,8 @@ static int packet_rcv(struct sk_buff *skb, struct net_device *dev, drop_n_acct: is_drop_n_account = true; - spin_lock(&sk->sk_receive_queue.lock); - po->stats.stats1.tp_drops++; + atomic_inc(&po->tp_drops); atomic_inc(&sk->sk_drops); - spin_unlock(&sk->sk_receive_queue.lock); drop_n_restore: if (skb_head != skb->data && skb_shared(skb)) { @@ -2155,11 +2169,13 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, int skb_len = skb->len; unsigned int snaplen, res; unsigned long status = TP_STATUS_USER; - unsigned short macoff, netoff, hdrlen; + unsigned short macoff, hdrlen; + unsigned int netoff; struct sk_buff *copy_skb = NULL; struct timespec ts; __u32 ts_status; bool is_drop_n_account = false; + unsigned int slot_id = 0; bool do_vnet = false; /* struct tpacket{2,3}_hdr is aligned to a multiple of TPACKET_ALIGNMENT. @@ -2193,6 +2209,12 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, if (!res) goto drop_n_restore; + /* If we are flooded, just give up */ + if (__packet_rcv_has_room(po, skb) == ROOM_NONE) { + atomic_inc(&po->tp_drops); + goto drop_n_restore; + } + if (skb->ip_summed == CHECKSUM_PARTIAL) status |= TP_STATUS_CSUMNOTREADY; else if (skb->pkt_type != PACKET_OUTGOING && @@ -2217,6 +2239,10 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, } macoff = netoff - maclen; } + if (netoff > USHRT_MAX) { + atomic_inc(&po->tp_drops); + goto drop_n_restore; + } if (po->tp_version <= TPACKET_V2) { if (macoff + snaplen > po->rx_ring.frame_size) { if (po->copy_thresh && @@ -2255,6 +2281,20 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, TP_STATUS_KERNEL, (macoff+snaplen)); if (!h.raw) goto drop_n_account; + + if (po->tp_version <= TPACKET_V2) { + slot_id = po->rx_ring.head; + if (test_bit(slot_id, po->rx_ring.rx_owner_map)) + goto drop_n_account; + __set_bit(slot_id, po->rx_ring.rx_owner_map); + } + + if (do_vnet && + virtio_net_hdr_from_skb(skb, h.raw + macoff - + sizeof(struct virtio_net_hdr), + vio_le(), true, 0)) + goto drop_n_account; + if (po->tp_version <= TPACKET_V2) { packet_increment_rx_head(po, &po->rx_ring); /* @@ -2263,16 +2303,10 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, * Anyways, moving it for V1/V2 only as V3 doesn't need this * at packet level. */ - if (po->stats.stats1.tp_drops) + if (atomic_read(&po->tp_drops)) status |= TP_STATUS_LOSING; } - if (do_vnet && - virtio_net_hdr_from_skb(skb, h.raw + macoff - - sizeof(struct virtio_net_hdr), - vio_le(), true, 0)) - goto drop_n_account; - po->stats.stats1.tp_packets++; if (copy_skb) { status |= TP_STATUS_COPY; @@ -2360,7 +2394,10 @@ static int tpacket_rcv(struct sk_buff *skb, struct net_device *dev, #endif if (po->tp_version <= TPACKET_V2) { + spin_lock(&sk->sk_receive_queue.lock); __packet_set_status(po, h.raw, status); + __clear_bit(slot_id, po->rx_ring.rx_owner_map); + spin_unlock(&sk->sk_receive_queue.lock); sk->sk_data_ready(sk); } else { prb_clear_blk_fill_status(&po->rx_ring); @@ -2379,9 +2416,9 @@ drop: return 0; drop_n_account: - is_drop_n_account = true; - po->stats.stats1.tp_drops++; spin_unlock(&sk->sk_receive_queue.lock); + atomic_inc(&po->tp_drops); + is_drop_n_account = true; sk->sk_data_ready(sk); kfree_skb(copy_skb); @@ -3325,8 +3362,7 @@ static int packet_recvmsg(struct socket *sock, struct msghdr *msg, size_t len, if (skb == NULL) goto out; - if (pkt_sk(sk)->pressure) - packet_rcv_has_room(pkt_sk(sk), NULL); + packet_rcv_try_clear_pressure(pkt_sk(sk)); if (pkt_sk(sk)->has_vnet_hdr) { err = packet_rcv_vnet(msg, skb, &len); @@ -3898,6 +3934,7 @@ static int packet_getsockopt(struct socket *sock, int level, int optname, void *data = &val; union tpacket_stats_u st; struct tpacket_rollover_stats rstats; + int drops; if (level != SOL_PACKET) return -ENOPROTOOPT; @@ -3914,14 +3951,17 @@ static int packet_getsockopt(struct socket *sock, int level, int optname, memcpy(&st, &po->stats, sizeof(st)); memset(&po->stats, 0, sizeof(po->stats)); spin_unlock_bh(&sk->sk_receive_queue.lock); + drops = atomic_xchg(&po->tp_drops, 0); if (po->tp_version == TPACKET_V3) { lv = sizeof(struct tpacket_stats_v3); - st.stats3.tp_packets += st.stats3.tp_drops; + st.stats3.tp_drops = drops; + st.stats3.tp_packets += drops; data = &st.stats3; } else { lv = sizeof(struct tpacket_stats); - st.stats1.tp_packets += st.stats1.tp_drops; + st.stats1.tp_drops = drops; + st.stats1.tp_packets += drops; data = &st.stats1; } @@ -4140,8 +4180,7 @@ static __poll_t packet_poll(struct file *file, struct socket *sock, TP_STATUS_KERNEL)) mask |= EPOLLIN | EPOLLRDNORM; } - if (po->pressure && __packet_rcv_has_room(po, NULL) == ROOM_NORMAL) - po->pressure = 0; + packet_rcv_try_clear_pressure(po); spin_unlock_bh(&sk->sk_receive_queue.lock); spin_lock_bh(&sk->sk_write_queue.lock); if (po->tx_ring.pg_vec) { @@ -4255,6 +4294,7 @@ static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u, { struct pgv *pg_vec = NULL; struct packet_sock *po = pkt_sk(sk); + unsigned long *rx_owner_map = NULL; int was_running, order = 0; struct packet_ring_buffer *rb; struct sk_buff_head *rb_queue; @@ -4340,6 +4380,12 @@ static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u, } break; default: + if (!tx_ring) { + rx_owner_map = bitmap_alloc(req->tp_frame_nr, + GFP_KERNEL | __GFP_NOWARN | __GFP_ZERO); + if (!rx_owner_map) + goto out_free_pg_vec; + } break; } } @@ -4369,6 +4415,8 @@ static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u, err = 0; spin_lock_bh(&rb_queue->lock); swap(rb->pg_vec, pg_vec); + if (po->tp_version <= TPACKET_V2) + swap(rb->rx_owner_map, rx_owner_map); rb->frame_max = (req->tp_frame_nr - 1); rb->head = 0; rb->frame_size = req->tp_frame_size; @@ -4400,6 +4448,7 @@ static int packet_set_ring(struct sock *sk, union tpacket_req_u *req_u, } out_free_pg_vec: + bitmap_free(rx_owner_map); if (pg_vec) free_pg_vec(pg_vec, order, req->tp_block_nr); out: |