^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) // SPDX-License-Identifier: GPL-2.0-only
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) /* Copyright (c) 2017 Covalent IO, Inc. http://covalent.io
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) /* Devmaps primary use is as a backend map for XDP BPF helper call
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) * bpf_redirect_map(). Because XDP is mostly concerned with performance we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * spent some effort to ensure the datapath with redirect maps does not use
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) * any locking. This is a quick note on the details.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) * We have three possible paths to get into the devmap control plane bpf
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) * syscalls, bpf programs, and driver side xmit/flush operations. A bpf syscall
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) * will invoke an update, delete, or lookup operation. To ensure updates and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) * deletes appear atomic from the datapath side xchg() is used to modify the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) * netdev_map array. Then because the datapath does a lookup into the netdev_map
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) * array (read-only) from an RCU critical section we use call_rcu() to wait for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) * an rcu grace period before free'ing the old data structures. This ensures the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) * datapath always has a valid copy. However, the datapath does a "flush"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) * operation that pushes any pending packets in the driver outside the RCU
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) * critical section. Each bpf_dtab_netdev tracks these pending operations using
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) * a per-cpu flush list. The bpf_dtab_netdev object will not be destroyed until
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) * this list is empty, indicating outstanding flush operations have completed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) * BPF syscalls may race with BPF program calls on any of the update, delete
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) * or lookup operations. As noted above the xchg() operation also keep the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) * netdev_map consistent in this case. From the devmap side BPF programs
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) * calling into these operations are the same as multiple user space threads
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) * making system calls.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) * Finally, any of the above may race with a netdev_unregister notifier. The
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) * unregister notifier must search for net devices in the map structure that
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) * contain a reference to the net device and remove them. This is a two step
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) * process (a) dereference the bpf_dtab_netdev object in netdev_map and (b)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) * check to see if the ifindex is the same as the net_device being removed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) * When removing the dev a cmpxchg() is used to ensure the correct dev is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) * removed, in the case of a concurrent update or delete operation it is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) * possible that the initially referenced dev is no longer in the map. As the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) * notifier hook walks the map we know that new dev references can not be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) * added by the user because core infrastructure ensures dev_get_by_index()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) * calls will fail at this point.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) * The devmap_hash type is a map type which interprets keys as ifindexes and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) * indexes these using a hashmap. This allows maps that use ifindex as key to be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) * densely packed instead of having holes in the lookup array for unused
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) * ifindexes. The setup and packet enqueue/send code is shared between the two
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) * types of devmap; only the lookup and insertion is different.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47) #include <linux/bpf.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) #include <net/xdp.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) #include <linux/filter.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) #include <trace/events/xdp.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) #define DEV_CREATE_FLAG_MASK \
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) (BPF_F_NUMA_NODE | BPF_F_RDONLY | BPF_F_WRONLY)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) struct xdp_dev_bulk_queue {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) struct xdp_frame *q[DEV_MAP_BULK_SIZE];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) struct list_head flush_node;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) struct net_device *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) struct net_device *dev_rx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) unsigned int count;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) struct bpf_dtab_netdev {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) struct net_device *dev; /* must be first member, due to tracepoint */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) struct hlist_node index_hlist;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) struct bpf_dtab *dtab;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) struct bpf_prog *xdp_prog;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) struct rcu_head rcu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) unsigned int idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) struct bpf_devmap_val val;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) struct bpf_dtab {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) struct bpf_map map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) struct bpf_dtab_netdev **netdev_map; /* DEVMAP type only */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) struct list_head list;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) /* these are only used for DEVMAP_HASH type maps */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) struct hlist_head *dev_index_head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) spinlock_t index_lock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) unsigned int items;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) u32 n_buckets;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) static DEFINE_PER_CPU(struct list_head, dev_flush_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) static DEFINE_SPINLOCK(dev_map_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) static LIST_HEAD(dev_map_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) static struct hlist_head *dev_map_create_hash(unsigned int entries,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) int numa_node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) struct hlist_head *hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) hash = bpf_map_area_alloc((u64) entries * sizeof(*hash), numa_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) if (hash != NULL)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) for (i = 0; i < entries; i++)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) INIT_HLIST_HEAD(&hash[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) return hash;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) static inline struct hlist_head *dev_map_index_hash(struct bpf_dtab *dtab,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) int idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) return &dtab->dev_index_head[idx & (dtab->n_buckets - 1)];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) static int dev_map_init_map(struct bpf_dtab *dtab, union bpf_attr *attr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) u32 valsize = attr->value_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) u64 cost = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) int err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) /* check sanity of attributes. 2 value sizes supported:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) * 4 bytes: ifindex
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) * 8 bytes: ifindex + prog fd
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) if (attr->max_entries == 0 || attr->key_size != 4 ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) (valsize != offsetofend(struct bpf_devmap_val, ifindex) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) valsize != offsetofend(struct bpf_devmap_val, bpf_prog.fd)) ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) attr->map_flags & ~DEV_CREATE_FLAG_MASK)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) /* Lookup returns a pointer straight to dev->ifindex, so make sure the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) * verifier prevents writes from the BPF side
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) attr->map_flags |= BPF_F_RDONLY_PROG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) bpf_map_init_from_attr(&dtab->map, attr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) if (attr->map_type == BPF_MAP_TYPE_DEVMAP_HASH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) dtab->n_buckets = roundup_pow_of_two(dtab->map.max_entries);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) if (!dtab->n_buckets) /* Overflow check */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) cost += (u64) sizeof(struct hlist_head) * dtab->n_buckets;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) cost += (u64) dtab->map.max_entries * sizeof(struct bpf_dtab_netdev *);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) /* if map size is larger than memlock limit, reject it */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) err = bpf_map_charge_init(&dtab->map.memory, cost);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) if (attr->map_type == BPF_MAP_TYPE_DEVMAP_HASH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) dtab->dev_index_head = dev_map_create_hash(dtab->n_buckets,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) dtab->map.numa_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) if (!dtab->dev_index_head)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) goto free_charge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) spin_lock_init(&dtab->index_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) dtab->netdev_map = bpf_map_area_alloc((u64) dtab->map.max_entries *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) sizeof(struct bpf_dtab_netdev *),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) dtab->map.numa_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) if (!dtab->netdev_map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) goto free_charge;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) free_charge:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) bpf_map_charge_finish(&dtab->map.memory);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) return -ENOMEM;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) static struct bpf_map *dev_map_alloc(union bpf_attr *attr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) struct bpf_dtab *dtab;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) int err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) if (!capable(CAP_NET_ADMIN))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) return ERR_PTR(-EPERM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) dtab = kzalloc(sizeof(*dtab), GFP_USER);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) if (!dtab)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) return ERR_PTR(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) err = dev_map_init_map(dtab, attr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) if (err) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) kfree(dtab);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) return ERR_PTR(err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) spin_lock(&dev_map_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) list_add_tail_rcu(&dtab->list, &dev_map_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) spin_unlock(&dev_map_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) return &dtab->map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) static void dev_map_free(struct bpf_map *map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) /* At this point bpf_prog->aux->refcnt == 0 and this map->refcnt == 0,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) * so the programs (can be more than one that used this map) were
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) * disconnected from events. The following synchronize_rcu() guarantees
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) * both rcu read critical sections complete and waits for
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) * preempt-disable regions (NAPI being the relevant context here) so we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) * are certain there will be no further reads against the netdev_map and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) * all flush operations are complete. Flush operations can only be done
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) * from NAPI context for this reason.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) spin_lock(&dev_map_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) list_del_rcu(&dtab->list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) spin_unlock(&dev_map_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) bpf_clear_redirect_map(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) synchronize_rcu();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) /* Make sure prior __dev_map_entry_free() have completed. */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) rcu_barrier();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) if (dtab->map.map_type == BPF_MAP_TYPE_DEVMAP_HASH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) for (i = 0; i < dtab->n_buckets; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) struct bpf_dtab_netdev *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) struct hlist_head *head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) struct hlist_node *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) head = dev_map_index_hash(dtab, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) hlist_for_each_entry_safe(dev, next, head, index_hlist) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) hlist_del_rcu(&dev->index_hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) if (dev->xdp_prog)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) bpf_prog_put(dev->xdp_prog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) dev_put(dev->dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) kfree(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) bpf_map_area_free(dtab->dev_index_head);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) for (i = 0; i < dtab->map.max_entries; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) struct bpf_dtab_netdev *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) dev = dtab->netdev_map[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) if (!dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246) if (dev->xdp_prog)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247) bpf_prog_put(dev->xdp_prog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248) dev_put(dev->dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) kfree(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) bpf_map_area_free(dtab->netdev_map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) kfree(dtab);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) static int dev_map_get_next_key(struct bpf_map *map, void *key, void *next_key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) u32 index = key ? *(u32 *)key : U32_MAX;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) u32 *next = next_key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) if (index >= dtab->map.max_entries) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) *next = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) if (index == dtab->map.max_entries - 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) return -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) *next = index + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) struct bpf_dtab_netdev *__dev_map_hash_lookup_elem(struct bpf_map *map, u32 key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) struct hlist_head *head = dev_map_index_hash(dtab, key);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) struct bpf_dtab_netdev *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) hlist_for_each_entry_rcu(dev, head, index_hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) lockdep_is_held(&dtab->index_lock))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) if (dev->idx == key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) return dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) static int dev_map_hash_get_next_key(struct bpf_map *map, void *key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) void *next_key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) u32 idx, *next = next_key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) struct bpf_dtab_netdev *dev, *next_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) struct hlist_head *head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) int i = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) if (!key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) goto find_first;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) idx = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) dev = __dev_map_hash_lookup_elem(map, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) if (!dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) goto find_first;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) next_dev = hlist_entry_safe(rcu_dereference_raw(hlist_next_rcu(&dev->index_hlist)),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) struct bpf_dtab_netdev, index_hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) if (next_dev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) *next = next_dev->idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) i = idx & (dtab->n_buckets - 1);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) i++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) find_first:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) for (; i < dtab->n_buckets; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) head = dev_map_index_hash(dtab, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) next_dev = hlist_entry_safe(rcu_dereference_raw(hlist_first_rcu(head)),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) struct bpf_dtab_netdev,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) index_hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) if (next_dev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) *next = next_dev->idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) return -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) bool dev_map_can_have_prog(struct bpf_map *map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) if ((map->map_type == BPF_MAP_TYPE_DEVMAP ||
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) map->map_type == BPF_MAP_TYPE_DEVMAP_HASH) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) map->value_size != offsetofend(struct bpf_devmap_val, ifindex))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) return true;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) return false;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) static void bq_xmit_all(struct xdp_dev_bulk_queue *bq, u32 flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) struct net_device *dev = bq->dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) int sent = 0, drops = 0, err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) int i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) if (unlikely(!bq->count))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) for (i = 0; i < bq->count; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) struct xdp_frame *xdpf = bq->q[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) prefetch(xdpf);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) sent = dev->netdev_ops->ndo_xdp_xmit(dev, bq->count, bq->q, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) if (sent < 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) err = sent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) sent = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) goto error;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 364) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 365) drops = bq->count - sent;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 366) out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 367) bq->count = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 368)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 369) trace_xdp_devmap_xmit(bq->dev_rx, dev, sent, drops, err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 370) bq->dev_rx = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 371) __list_del_clearprev(&bq->flush_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 372) return;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 373) error:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 374) /* If ndo_xdp_xmit fails with an errno, no frames have been
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 375) * xmit'ed and it's our responsibility to them free all.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 376) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 377) for (i = 0; i < bq->count; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 378) struct xdp_frame *xdpf = bq->q[i];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 379)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 380) xdp_return_frame_rx_napi(xdpf);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 381) drops++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 382) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 383) goto out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 384) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 385)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 386) /* __dev_flush is called from xdp_do_flush() which _must_ be signaled
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 387) * from the driver before returning from its napi->poll() routine. The poll()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 388) * routine is called either from busy_poll context or net_rx_action signaled
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 389) * from NET_RX_SOFTIRQ. Either way the poll routine must complete before the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 390) * net device can be torn down. On devmap tear down we ensure the flush list
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 391) * is empty before completing to ensure all flush operations have completed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 392) * When drivers update the bpf program they may need to ensure any flush ops
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 393) * are also complete. Using synchronize_rcu or call_rcu will suffice for this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 394) * because both wait for napi context to exit.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 395) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 396) void __dev_flush(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 397) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 398) struct list_head *flush_list = this_cpu_ptr(&dev_flush_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 399) struct xdp_dev_bulk_queue *bq, *tmp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 400)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 401) list_for_each_entry_safe(bq, tmp, flush_list, flush_node)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 402) bq_xmit_all(bq, XDP_XMIT_FLUSH);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 403) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 404)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 405) /* rcu_read_lock (from syscall and BPF contexts) ensures that if a delete and/or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 406) * update happens in parallel here a dev_put wont happen until after reading the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 407) * ifindex.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 408) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 409) struct bpf_dtab_netdev *__dev_map_lookup_elem(struct bpf_map *map, u32 key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 410) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 411) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 412) struct bpf_dtab_netdev *obj;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 413)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 414) if (key >= map->max_entries)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 415) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 416)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 417) obj = READ_ONCE(dtab->netdev_map[key]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 418) return obj;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 419) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 420)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 421) /* Runs under RCU-read-side, plus in softirq under NAPI protection.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 422) * Thus, safe percpu variable access.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 423) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 424) static void bq_enqueue(struct net_device *dev, struct xdp_frame *xdpf,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 425) struct net_device *dev_rx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 426) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 427) struct list_head *flush_list = this_cpu_ptr(&dev_flush_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 428) struct xdp_dev_bulk_queue *bq = this_cpu_ptr(dev->xdp_bulkq);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 429)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 430) if (unlikely(bq->count == DEV_MAP_BULK_SIZE))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 431) bq_xmit_all(bq, 0);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 432)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 433) /* Ingress dev_rx will be the same for all xdp_frame's in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 434) * bulk_queue, because bq stored per-CPU and must be flushed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 435) * from net_device drivers NAPI func end.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 436) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 437) if (!bq->dev_rx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 438) bq->dev_rx = dev_rx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 439)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 440) bq->q[bq->count++] = xdpf;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 441)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 442) if (!bq->flush_node.prev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 443) list_add(&bq->flush_node, flush_list);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 444) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 445)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 446) static inline int __xdp_enqueue(struct net_device *dev, struct xdp_buff *xdp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 447) struct net_device *dev_rx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 448) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 449) struct xdp_frame *xdpf;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 450) int err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 451)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 452) if (!dev->netdev_ops->ndo_xdp_xmit)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 453) return -EOPNOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 454)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 455) err = xdp_ok_fwd_dev(dev, xdp->data_end - xdp->data);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 456) if (unlikely(err))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 457) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 458)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 459) xdpf = xdp_convert_buff_to_frame(xdp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 460) if (unlikely(!xdpf))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 461) return -EOVERFLOW;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 462)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 463) bq_enqueue(dev, xdpf, dev_rx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 464) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 465) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 466)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 467) static struct xdp_buff *dev_map_run_prog(struct net_device *dev,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 468) struct xdp_buff *xdp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 469) struct bpf_prog *xdp_prog)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 470) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 471) struct xdp_txq_info txq = { .dev = dev };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 472) u32 act;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 473)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 474) xdp_set_data_meta_invalid(xdp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 475) xdp->txq = &txq;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 476)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 477) act = bpf_prog_run_xdp(xdp_prog, xdp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 478) switch (act) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 479) case XDP_PASS:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 480) return xdp;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 481) case XDP_DROP:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 482) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 483) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 484) bpf_warn_invalid_xdp_action(act);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 485) fallthrough;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 486) case XDP_ABORTED:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 487) trace_xdp_exception(dev, xdp_prog, act);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 488) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 489) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 490)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 491) xdp_return_buff(xdp);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 492) return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 493) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 494)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 495) int dev_xdp_enqueue(struct net_device *dev, struct xdp_buff *xdp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 496) struct net_device *dev_rx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 497) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 498) return __xdp_enqueue(dev, xdp, dev_rx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 499) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 500)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 501) int dev_map_enqueue(struct bpf_dtab_netdev *dst, struct xdp_buff *xdp,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 502) struct net_device *dev_rx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 503) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 504) struct net_device *dev = dst->dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 505)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 506) if (dst->xdp_prog) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 507) xdp = dev_map_run_prog(dev, xdp, dst->xdp_prog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 508) if (!xdp)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 509) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 510) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 511) return __xdp_enqueue(dev, xdp, dev_rx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 512) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 513)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 514) int dev_map_generic_redirect(struct bpf_dtab_netdev *dst, struct sk_buff *skb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 515) struct bpf_prog *xdp_prog)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 516) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 517) int err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 518)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 519) err = xdp_ok_fwd_dev(dst->dev, skb->len);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 520) if (unlikely(err))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 521) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 522) skb->dev = dst->dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 523) generic_xdp_tx(skb, xdp_prog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 524)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 525) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 526) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 527)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 528) static void *dev_map_lookup_elem(struct bpf_map *map, void *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 529) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 530) struct bpf_dtab_netdev *obj = __dev_map_lookup_elem(map, *(u32 *)key);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 531)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 532) return obj ? &obj->val : NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 533) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 534)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 535) static void *dev_map_hash_lookup_elem(struct bpf_map *map, void *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 536) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 537) struct bpf_dtab_netdev *obj = __dev_map_hash_lookup_elem(map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 538) *(u32 *)key);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 539) return obj ? &obj->val : NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 540) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 541)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 542) static void __dev_map_entry_free(struct rcu_head *rcu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 543) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 544) struct bpf_dtab_netdev *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 545)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 546) dev = container_of(rcu, struct bpf_dtab_netdev, rcu);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 547) if (dev->xdp_prog)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 548) bpf_prog_put(dev->xdp_prog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 549) dev_put(dev->dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 550) kfree(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 551) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 552)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 553) static int dev_map_delete_elem(struct bpf_map *map, void *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 554) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 555) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 556) struct bpf_dtab_netdev *old_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 557) int k = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 558)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 559) if (k >= map->max_entries)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 560) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 561)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 562) /* Use call_rcu() here to ensure any rcu critical sections have
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 563) * completed as well as any flush operations because call_rcu
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 564) * will wait for preempt-disable region to complete, NAPI in this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 565) * context. And additionally, the driver tear down ensures all
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 566) * soft irqs are complete before removing the net device in the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 567) * case of dev_put equals zero.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 568) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 569) old_dev = xchg(&dtab->netdev_map[k], NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 570) if (old_dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 571) call_rcu(&old_dev->rcu, __dev_map_entry_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 572) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 573) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 574)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 575) static int dev_map_hash_delete_elem(struct bpf_map *map, void *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 576) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 577) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 578) struct bpf_dtab_netdev *old_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 579) int k = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 580) unsigned long flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 581) int ret = -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 582)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 583) spin_lock_irqsave(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 584)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 585) old_dev = __dev_map_hash_lookup_elem(map, k);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 586) if (old_dev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 587) dtab->items--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 588) hlist_del_init_rcu(&old_dev->index_hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 589) call_rcu(&old_dev->rcu, __dev_map_entry_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 590) ret = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 591) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 592) spin_unlock_irqrestore(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 593)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 594) return ret;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 595) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 596)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 597) static struct bpf_dtab_netdev *__dev_map_alloc_node(struct net *net,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 598) struct bpf_dtab *dtab,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 599) struct bpf_devmap_val *val,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 600) unsigned int idx)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 601) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 602) struct bpf_prog *prog = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 603) struct bpf_dtab_netdev *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 604)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 605) dev = kmalloc_node(sizeof(*dev), GFP_ATOMIC | __GFP_NOWARN,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 606) dtab->map.numa_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 607) if (!dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 608) return ERR_PTR(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 609)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 610) dev->dev = dev_get_by_index(net, val->ifindex);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 611) if (!dev->dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 612) goto err_out;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 613)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 614) if (val->bpf_prog.fd > 0) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 615) prog = bpf_prog_get_type_dev(val->bpf_prog.fd,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 616) BPF_PROG_TYPE_XDP, false);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 617) if (IS_ERR(prog))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 618) goto err_put_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 619) if (prog->expected_attach_type != BPF_XDP_DEVMAP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 620) goto err_put_prog;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 621) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 622)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 623) dev->idx = idx;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 624) dev->dtab = dtab;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 625) if (prog) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 626) dev->xdp_prog = prog;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 627) dev->val.bpf_prog.id = prog->aux->id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 628) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 629) dev->xdp_prog = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 630) dev->val.bpf_prog.id = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 631) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 632) dev->val.ifindex = val->ifindex;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 633)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 634) return dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 635) err_put_prog:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 636) bpf_prog_put(prog);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 637) err_put_dev:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 638) dev_put(dev->dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 639) err_out:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 640) kfree(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 641) return ERR_PTR(-EINVAL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 642) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 643)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 644) static int __dev_map_update_elem(struct net *net, struct bpf_map *map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 645) void *key, void *value, u64 map_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 646) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 647) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 648) struct bpf_dtab_netdev *dev, *old_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 649) struct bpf_devmap_val val = {};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 650) u32 i = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 651)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 652) if (unlikely(map_flags > BPF_EXIST))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 653) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 654) if (unlikely(i >= dtab->map.max_entries))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 655) return -E2BIG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 656) if (unlikely(map_flags == BPF_NOEXIST))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 657) return -EEXIST;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 658)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 659) /* already verified value_size <= sizeof val */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 660) memcpy(&val, value, map->value_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 661)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 662) if (!val.ifindex) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 663) dev = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 664) /* can not specify fd if ifindex is 0 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 665) if (val.bpf_prog.fd > 0)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 666) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 667) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 668) dev = __dev_map_alloc_node(net, dtab, &val, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 669) if (IS_ERR(dev))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 670) return PTR_ERR(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 671) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 672)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 673) /* Use call_rcu() here to ensure rcu critical sections have completed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 674) * Remembering the driver side flush operation will happen before the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 675) * net device is removed.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 676) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 677) old_dev = xchg(&dtab->netdev_map[i], dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 678) if (old_dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 679) call_rcu(&old_dev->rcu, __dev_map_entry_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 680)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 681) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 682) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 683)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 684) static int dev_map_update_elem(struct bpf_map *map, void *key, void *value,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 685) u64 map_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 686) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 687) return __dev_map_update_elem(current->nsproxy->net_ns,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 688) map, key, value, map_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 689) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 690)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 691) static int __dev_map_hash_update_elem(struct net *net, struct bpf_map *map,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 692) void *key, void *value, u64 map_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 693) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 694) struct bpf_dtab *dtab = container_of(map, struct bpf_dtab, map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 695) struct bpf_dtab_netdev *dev, *old_dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 696) struct bpf_devmap_val val = {};
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 697) u32 idx = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 698) unsigned long flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 699) int err = -EEXIST;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 700)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 701) /* already verified value_size <= sizeof val */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 702) memcpy(&val, value, map->value_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 703)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 704) if (unlikely(map_flags > BPF_EXIST || !val.ifindex))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 705) return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 706)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 707) spin_lock_irqsave(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 708)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 709) old_dev = __dev_map_hash_lookup_elem(map, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 710) if (old_dev && (map_flags & BPF_NOEXIST))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 711) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 712)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 713) dev = __dev_map_alloc_node(net, dtab, &val, idx);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 714) if (IS_ERR(dev)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 715) err = PTR_ERR(dev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 716) goto out_err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 717) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 718)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 719) if (old_dev) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 720) hlist_del_rcu(&old_dev->index_hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 721) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 722) if (dtab->items >= dtab->map.max_entries) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 723) spin_unlock_irqrestore(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 724) call_rcu(&dev->rcu, __dev_map_entry_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 725) return -E2BIG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 726) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 727) dtab->items++;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 728) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 729)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 730) hlist_add_head_rcu(&dev->index_hlist,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 731) dev_map_index_hash(dtab, idx));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 732) spin_unlock_irqrestore(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 733)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 734) if (old_dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 735) call_rcu(&old_dev->rcu, __dev_map_entry_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 736)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 737) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 738)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 739) out_err:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 740) spin_unlock_irqrestore(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 741) return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 742) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 743)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 744) static int dev_map_hash_update_elem(struct bpf_map *map, void *key, void *value,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 745) u64 map_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 746) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 747) return __dev_map_hash_update_elem(current->nsproxy->net_ns,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 748) map, key, value, map_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 749) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 750)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 751) static int dev_map_btf_id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 752) const struct bpf_map_ops dev_map_ops = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 753) .map_meta_equal = bpf_map_meta_equal,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 754) .map_alloc = dev_map_alloc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 755) .map_free = dev_map_free,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 756) .map_get_next_key = dev_map_get_next_key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 757) .map_lookup_elem = dev_map_lookup_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 758) .map_update_elem = dev_map_update_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 759) .map_delete_elem = dev_map_delete_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 760) .map_check_btf = map_check_no_btf,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 761) .map_btf_name = "bpf_dtab",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 762) .map_btf_id = &dev_map_btf_id,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 763) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 764)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 765) static int dev_map_hash_map_btf_id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 766) const struct bpf_map_ops dev_map_hash_ops = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 767) .map_meta_equal = bpf_map_meta_equal,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 768) .map_alloc = dev_map_alloc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 769) .map_free = dev_map_free,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 770) .map_get_next_key = dev_map_hash_get_next_key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 771) .map_lookup_elem = dev_map_hash_lookup_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 772) .map_update_elem = dev_map_hash_update_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 773) .map_delete_elem = dev_map_hash_delete_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 774) .map_check_btf = map_check_no_btf,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 775) .map_btf_name = "bpf_dtab",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 776) .map_btf_id = &dev_map_hash_map_btf_id,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 777) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 778)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 779) static void dev_map_hash_remove_netdev(struct bpf_dtab *dtab,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 780) struct net_device *netdev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 781) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 782) unsigned long flags;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 783) u32 i;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 784)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 785) spin_lock_irqsave(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 786) for (i = 0; i < dtab->n_buckets; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 787) struct bpf_dtab_netdev *dev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 788) struct hlist_head *head;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 789) struct hlist_node *next;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 790)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 791) head = dev_map_index_hash(dtab, i);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 792)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 793) hlist_for_each_entry_safe(dev, next, head, index_hlist) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 794) if (netdev != dev->dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 795) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 796)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 797) dtab->items--;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 798) hlist_del_rcu(&dev->index_hlist);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 799) call_rcu(&dev->rcu, __dev_map_entry_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 800) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 801) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 802) spin_unlock_irqrestore(&dtab->index_lock, flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 803) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 804)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 805) static int dev_map_notification(struct notifier_block *notifier,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 806) ulong event, void *ptr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 807) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 808) struct net_device *netdev = netdev_notifier_info_to_dev(ptr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 809) struct bpf_dtab *dtab;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 810) int i, cpu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 811)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 812) switch (event) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 813) case NETDEV_REGISTER:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 814) if (!netdev->netdev_ops->ndo_xdp_xmit || netdev->xdp_bulkq)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 815) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 816)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 817) /* will be freed in free_netdev() */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 818) netdev->xdp_bulkq = alloc_percpu(struct xdp_dev_bulk_queue);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 819) if (!netdev->xdp_bulkq)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 820) return NOTIFY_BAD;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 821)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 822) for_each_possible_cpu(cpu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 823) per_cpu_ptr(netdev->xdp_bulkq, cpu)->dev = netdev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 824) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 825) case NETDEV_UNREGISTER:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 826) /* This rcu_read_lock/unlock pair is needed because
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 827) * dev_map_list is an RCU list AND to ensure a delete
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 828) * operation does not free a netdev_map entry while we
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 829) * are comparing it against the netdev being unregistered.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 830) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 831) rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 832) list_for_each_entry_rcu(dtab, &dev_map_list, list) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 833) if (dtab->map.map_type == BPF_MAP_TYPE_DEVMAP_HASH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 834) dev_map_hash_remove_netdev(dtab, netdev);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 835) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 836) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 837)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 838) for (i = 0; i < dtab->map.max_entries; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 839) struct bpf_dtab_netdev *dev, *odev;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 840)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 841) dev = READ_ONCE(dtab->netdev_map[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 842) if (!dev || netdev != dev->dev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 843) continue;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 844) odev = cmpxchg(&dtab->netdev_map[i], dev, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 845) if (dev == odev)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 846) call_rcu(&dev->rcu,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 847) __dev_map_entry_free);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 848) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 849) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 850) rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 851) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 852) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 853) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 854) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 855) return NOTIFY_OK;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 856) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 857)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 858) static struct notifier_block dev_map_notifier = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 859) .notifier_call = dev_map_notification,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 860) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 861)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 862) static int __init dev_map_init(void)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 863) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 864) int cpu;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 865)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 866) /* Assure tracepoint shadow struct _bpf_dtab_netdev is in sync */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 867) BUILD_BUG_ON(offsetof(struct bpf_dtab_netdev, dev) !=
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 868) offsetof(struct _bpf_dtab_netdev, dev));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 869) register_netdevice_notifier(&dev_map_notifier);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 870)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 871) for_each_possible_cpu(cpu)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 872) INIT_LIST_HEAD(&per_cpu(dev_flush_list, cpu));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 873) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 874) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 875)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 876) subsys_initcall(dev_map_init);
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags