Orange Pi5 kernel

^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   1) // SPDX-License-Identifier: GPL-2.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   2) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   3)  * Copyright (c) 2018 Facebook
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   4)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   5) #include <linux/bpf.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   6) #include <linux/err.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   7) #include <linux/sock_diag.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   8) #include <net/sock_reuseport.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300   9) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  10) struct reuseport_array {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  11) 	struct bpf_map map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  12) 	struct sock __rcu *ptrs[];
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  13) };
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  14) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  15) static struct reuseport_array *reuseport_array(struct bpf_map *map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  16) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  17) 	return (struct reuseport_array *)map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  18) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  19) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  20) /* The caller must hold the reuseport_lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  21) void bpf_sk_reuseport_detach(struct sock *sk)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  22) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  23) 	uintptr_t sk_user_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  24) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  25) 	write_lock_bh(&sk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  26) 	sk_user_data = (uintptr_t)sk->sk_user_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  27) 	if (sk_user_data & SK_USER_DATA_BPF) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  28) 		struct sock __rcu **socks;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  29) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  30) 		socks = (void *)(sk_user_data & SK_USER_DATA_PTRMASK);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  31) 		WRITE_ONCE(sk->sk_user_data, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  32) 		/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  33) 		 * Do not move this NULL assignment outside of
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  34) 		 * sk->sk_callback_lock because there is
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  35) 		 * a race with reuseport_array_free()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  36) 		 * which does not hold the reuseport_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  37) 		 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  38) 		RCU_INIT_POINTER(*socks, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  39) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  40) 	write_unlock_bh(&sk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  41) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  42) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  43) static int reuseport_array_alloc_check(union bpf_attr *attr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  44) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  45) 	if (attr->value_size != sizeof(u32) &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  46) 	    attr->value_size != sizeof(u64))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  47) 		return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  48) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  49) 	return array_map_alloc_check(attr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  50) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  51) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  52) static void *reuseport_array_lookup_elem(struct bpf_map *map, void *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  53) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  54) 	struct reuseport_array *array = reuseport_array(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  55) 	u32 index = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  56) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  57) 	if (unlikely(index >= array->map.max_entries))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  58) 		return NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  59) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  60) 	return rcu_dereference(array->ptrs[index]);
^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) /* Called from syscall only */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  64) static int reuseport_array_delete_elem(struct bpf_map *map, void *key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  65) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  66) 	struct reuseport_array *array = reuseport_array(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  67) 	u32 index = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  68) 	struct sock *sk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  69) 	int err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  70) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  71) 	if (index >= map->max_entries)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  72) 		return -E2BIG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  73) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  74) 	if (!rcu_access_pointer(array->ptrs[index]))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  75) 		return -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  76) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  77) 	spin_lock_bh(&reuseport_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  78) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  79) 	sk = rcu_dereference_protected(array->ptrs[index],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  80) 				       lockdep_is_held(&reuseport_lock));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  81) 	if (sk) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  82) 		write_lock_bh(&sk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  83) 		WRITE_ONCE(sk->sk_user_data, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  84) 		RCU_INIT_POINTER(array->ptrs[index], NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  85) 		write_unlock_bh(&sk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  86) 		err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  87) 	} else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  88) 		err = -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  89) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  90) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  91) 	spin_unlock_bh(&reuseport_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  92) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  93) 	return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  94) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  95) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  96) static void reuseport_array_free(struct bpf_map *map)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  97) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  98) 	struct reuseport_array *array = reuseport_array(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300  99) 	struct sock *sk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) 	u32 i;
^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) 	 * ops->map_*_elem() will not be able to access this
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) 	 * array now. Hence, this function only races with
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 105) 	 * bpf_sk_reuseport_detach() which was triggerred by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 106) 	 * close() or disconnect().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 107) 	 *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 108) 	 * This function and bpf_sk_reuseport_detach() are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 109) 	 * both removing sk from "array".  Who removes it
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 110) 	 * first does not matter.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 111) 	 *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 112) 	 * The only concern here is bpf_sk_reuseport_detach()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 113) 	 * may access "array" which is being freed here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 114) 	 * bpf_sk_reuseport_detach() access this "array"
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 115) 	 * through sk->sk_user_data _and_ with sk->sk_callback_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 116) 	 * held which is enough because this "array" is not freed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 117) 	 * until all sk->sk_user_data has stopped referencing this "array".
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 118) 	 *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 119) 	 * Hence, due to the above, taking "reuseport_lock" is not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 120) 	 * needed here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 121) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 122) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 123) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 124) 	 * Since reuseport_lock is not taken, sk is accessed under
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 125) 	 * rcu_read_lock()
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 126) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 127) 	rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 128) 	for (i = 0; i < map->max_entries; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 129) 		sk = rcu_dereference(array->ptrs[i]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 130) 		if (sk) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 131) 			write_lock_bh(&sk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 132) 			/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 133) 			 * No need for WRITE_ONCE(). At this point,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 134) 			 * no one is reading it without taking the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 135) 			 * sk->sk_callback_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 136) 			 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 137) 			sk->sk_user_data = NULL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 138) 			write_unlock_bh(&sk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 139) 			RCU_INIT_POINTER(array->ptrs[i], NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 140) 		}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 141) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 142) 	rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 143) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 144) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 145) 	 * Once reaching here, all sk->sk_user_data is not
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 146) 	 * referenceing this "array".  "array" can be freed now.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 147) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 148) 	bpf_map_area_free(array);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 149) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 150) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 151) static struct bpf_map *reuseport_array_alloc(union bpf_attr *attr)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 152) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 153) 	int err, numa_node = bpf_map_attr_numa_node(attr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 154) 	struct reuseport_array *array;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 155) 	struct bpf_map_memory mem;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 156) 	u64 array_size;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 157) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 158) 	if (!bpf_capable())
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 159) 		return ERR_PTR(-EPERM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 160) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 161) 	array_size = sizeof(*array);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 162) 	array_size += (u64)attr->max_entries * sizeof(struct sock *);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 163) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 164) 	err = bpf_map_charge_init(&mem, array_size);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 165) 	if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 166) 		return ERR_PTR(err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 167) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 168) 	/* allocate all map elements and zero-initialize them */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 169) 	array = bpf_map_area_alloc(array_size, numa_node);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 170) 	if (!array) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 171) 		bpf_map_charge_finish(&mem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 172) 		return ERR_PTR(-ENOMEM);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 173) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 174) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 175) 	/* copy mandatory map attributes */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 176) 	bpf_map_init_from_attr(&array->map, attr);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 177) 	bpf_map_charge_move(&array->map.memory, &mem);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 178) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 179) 	return &array->map;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 180) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 181) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 182) int bpf_fd_reuseport_array_lookup_elem(struct bpf_map *map, void *key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 183) 				       void *value)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 184) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 185) 	struct sock *sk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 186) 	int err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 187) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 188) 	if (map->value_size != sizeof(u64))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 189) 		return -ENOSPC;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 190) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 191) 	rcu_read_lock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 192) 	sk = reuseport_array_lookup_elem(map, key);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 193) 	if (sk) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 194) 		*(u64 *)value = __sock_gen_cookie(sk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 195) 		err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 196) 	} else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 197) 		err = -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 198) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 199) 	rcu_read_unlock();
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 200) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 201) 	return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 202) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 203) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 204) static int
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 205) reuseport_array_update_check(const struct reuseport_array *array,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 206) 			     const struct sock *nsk,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 207) 			     const struct sock *osk,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 208) 			     const struct sock_reuseport *nsk_reuse,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 209) 			     u32 map_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 210) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 211) 	if (osk && map_flags == BPF_NOEXIST)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 212) 		return -EEXIST;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 213) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 214) 	if (!osk && map_flags == BPF_EXIST)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 215) 		return -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 216) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 217) 	if (nsk->sk_protocol != IPPROTO_UDP && nsk->sk_protocol != IPPROTO_TCP)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 218) 		return -ENOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 219) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 220) 	if (nsk->sk_family != AF_INET && nsk->sk_family != AF_INET6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 221) 		return -ENOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 222) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 223) 	if (nsk->sk_type != SOCK_STREAM && nsk->sk_type != SOCK_DGRAM)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 224) 		return -ENOTSUPP;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 225) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 226) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 227) 	 * sk must be hashed (i.e. listening in the TCP case or binded
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 228) 	 * in the UDP case) and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 229) 	 * it must also be a SO_REUSEPORT sk (i.e. reuse cannot be NULL).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 230) 	 *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 231) 	 * Also, sk will be used in bpf helper that is protected by
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 232) 	 * rcu_read_lock().
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 233) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 234) 	if (!sock_flag(nsk, SOCK_RCU_FREE) || !sk_hashed(nsk) || !nsk_reuse)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 235) 		return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 236) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 237) 	/* READ_ONCE because the sk->sk_callback_lock may not be held here */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 238) 	if (READ_ONCE(nsk->sk_user_data))
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 239) 		return -EBUSY;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 240) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 241) 	return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 242) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 243) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 244) /*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 245)  * Called from syscall only.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 246)  * The "nsk" in the fd refcnt.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 247)  * The "osk" and "reuse" are protected by reuseport_lock.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 248)  */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 249) int bpf_fd_reuseport_array_update_elem(struct bpf_map *map, void *key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 250) 				       void *value, u64 map_flags)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 251) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 252) 	struct reuseport_array *array = reuseport_array(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 253) 	struct sock *free_osk = NULL, *osk, *nsk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 254) 	struct sock_reuseport *reuse;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 255) 	u32 index = *(u32 *)key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 256) 	uintptr_t sk_user_data;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 257) 	struct socket *socket;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 258) 	int err, fd;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 259) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 260) 	if (map_flags > BPF_EXIST)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 261) 		return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 262) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 263) 	if (index >= map->max_entries)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 264) 		return -E2BIG;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 265) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 266) 	if (map->value_size == sizeof(u64)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 267) 		u64 fd64 = *(u64 *)value;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 268) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 269) 		if (fd64 > S32_MAX)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 270) 			return -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 271) 		fd = fd64;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 272) 	} else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 273) 		fd = *(int *)value;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 274) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 275) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 276) 	socket = sockfd_lookup(fd, &err);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 277) 	if (!socket)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 278) 		return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 279) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 280) 	nsk = socket->sk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 281) 	if (!nsk) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 282) 		err = -EINVAL;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 283) 		goto put_file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 284) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 285) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 286) 	/* Quick checks before taking reuseport_lock */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 287) 	err = reuseport_array_update_check(array, nsk,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 288) 					   rcu_access_pointer(array->ptrs[index]),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 289) 					   rcu_access_pointer(nsk->sk_reuseport_cb),
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 290) 					   map_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 291) 	if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 292) 		goto put_file;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 293) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 294) 	spin_lock_bh(&reuseport_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 295) 	/*
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 296) 	 * Some of the checks only need reuseport_lock
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 297) 	 * but it is done under sk_callback_lock also
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 298) 	 * for simplicity reason.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 299) 	 */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 300) 	write_lock_bh(&nsk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 301) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 302) 	osk = rcu_dereference_protected(array->ptrs[index],
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 303) 					lockdep_is_held(&reuseport_lock));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 304) 	reuse = rcu_dereference_protected(nsk->sk_reuseport_cb,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 305) 					  lockdep_is_held(&reuseport_lock));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 306) 	err = reuseport_array_update_check(array, nsk, osk, reuse, map_flags);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 307) 	if (err)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 308) 		goto put_file_unlock;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 309) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 310) 	sk_user_data = (uintptr_t)&array->ptrs[index] | SK_USER_DATA_NOCOPY |
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 311) 		SK_USER_DATA_BPF;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 312) 	WRITE_ONCE(nsk->sk_user_data, (void *)sk_user_data);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 313) 	rcu_assign_pointer(array->ptrs[index], nsk);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 314) 	free_osk = osk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 315) 	err = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 316) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 317) put_file_unlock:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 318) 	write_unlock_bh(&nsk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 319) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 320) 	if (free_osk) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 321) 		write_lock_bh(&free_osk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 322) 		WRITE_ONCE(free_osk->sk_user_data, NULL);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 323) 		write_unlock_bh(&free_osk->sk_callback_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 324) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 325) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 326) 	spin_unlock_bh(&reuseport_lock);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 327) put_file:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 328) 	fput(socket->file);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 329) 	return err;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 330) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 331) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 332) /* Called from syscall */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 333) static int reuseport_array_get_next_key(struct bpf_map *map, void *key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 334) 					void *next_key)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 335) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 336) 	struct reuseport_array *array = reuseport_array(map);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 337) 	u32 index = key ? *(u32 *)key : U32_MAX;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 338) 	u32 *next = (u32 *)next_key;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 339) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 340) 	if (index >= array->map.max_entries) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 341) 		*next = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 342) 		return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 343) 	}
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 344) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 345) 	if (index == array->map.max_entries - 1)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 346) 		return -ENOENT;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 347) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 348) 	*next = index + 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 349) 	return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 350) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 351) 
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 352) static int reuseport_array_map_btf_id;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 353) const struct bpf_map_ops reuseport_array_ops = {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 354) 	.map_meta_equal = bpf_map_meta_equal,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 355) 	.map_alloc_check = reuseport_array_alloc_check,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 356) 	.map_alloc = reuseport_array_alloc,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 357) 	.map_free = reuseport_array_free,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 358) 	.map_lookup_elem = reuseport_array_lookup_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 359) 	.map_get_next_key = reuseport_array_get_next_key,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 360) 	.map_delete_elem = reuseport_array_delete_elem,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 361) 	.map_btf_name = "reuseport_array",
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 362) 	.map_btf_id = &reuseport_array_map_btf_id,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 363) };