blob: c63eb2529d4418b489074e5fb2c6969c294752d8 [file] [log] [blame]
/** @file
* @brief Network packet buffer descriptor API
*
* Network data is passed between different parts of the stack via
* net_buf struct.
*/
/*
* Copyright (c) 2016 Intel Corporation
*
* SPDX-License-Identifier: Apache-2.0
*/
/* Data buffer API - used for all data to/from net */
#ifndef __NET_PKT_H__
#define __NET_PKT_H__
#include <zephyr/types.h>
#include <stdbool.h>
#include <net/buf.h>
#include <net/net_core.h>
#include <net/net_linkaddr.h>
#include <net/net_ip.h>
#include <net/net_if.h>
#include <net/net_context.h>
#ifdef __cplusplus
extern "C" {
#endif
/**
* @brief Network packet management library
* @defgroup net_pkt Network Packet Library
* @ingroup networking
* @{
*/
struct net_context;
/* Note that if you add new fields into net_pkt, remember to update
* net_pkt_clone() function.
*/
struct net_pkt {
/** FIFO uses first 4 bytes itself, reserve space */
int _reserved;
/** Slab pointer from where it belongs to */
struct k_mem_slab *slab;
/** List of buffer fragments holding the packet */
struct net_buf *frags;
/** Network connection context */
struct net_context *context;
/** Network context token that user can set. This is passed
* to user callback when data has been sent.
*/
void *token;
/** Network interface */
struct net_if *iface;
/** @cond ignore */
u8_t *appdata; /* application data starts here */
u8_t *next_hdr; /* where is the next header */
/* Filled by layer 2 when network packet is received. */
struct net_linkaddr lladdr_src;
struct net_linkaddr lladdr_dst;
u16_t data_len; /* amount of payload data that can be added */
u16_t appdatalen;
u8_t ll_reserve; /* link layer header length */
u8_t ip_hdr_len; /* pre-filled in order to avoid func call */
#if defined(CONFIG_NET_TCP)
sys_snode_t sent_list;
#endif
u8_t sent_or_eof: 1; /* For outgoing packet: is this sent or not
* For incoming packet of a socket: last
* packet before EOF
* Used only if defined(CONFIG_NET_TCP)
*/
u8_t pkt_queued: 1; /* For outgoing packet: is this packet queued
* to be sent but has not reached the driver
* yet. Used only if defined(CONFIG_NET_TCP)
*/
u8_t forwarding : 1; /* Are we forwarding this pkt
* Used only if defined(CONFIG_NET_ROUTE)
*/
u8_t family : 4; /* IPv4 vs IPv6 */
u8_t _unused : 3;
union {
/* IPv6 hop limit or IPv4 ttl for this network packet.
* The value is shared between IPv6 and IPv4.
*/
u8_t ipv6_hop_limit;
u8_t ipv4_ttl;
};
#if defined(CONFIG_NET_IPV6)
u8_t ipv6_ext_len; /* length of extension headers */
u8_t ipv6_ext_opt_len; /* IPv6 ND option length */
/* Where is the start of the last header before payload data
* in IPv6 packet. This is offset value from start of the IPv6
* packet. Note that this value should be updated by who ever
* adds IPv6 extension headers to the network packet.
*/
u16_t ipv6_prev_hdr_start;
#if defined(CONFIG_NET_IPV6_FRAGMENT)
u16_t ipv6_fragment_offset; /* Fragment offset of this packet */
u32_t ipv6_fragment_id; /* Fragment id */
u8_t *ipv6_frag_hdr_start; /* Where starts the fragment header */
#endif /* CONFIG_NET_IPV6_FRAGMENT */
#endif /* CONFIG_NET_IPV6 */
#if defined(CONFIG_IEEE802154)
u8_t ieee802154_rssi; /* Received Signal Strength Indication */
u8_t ieee802154_lqi; /* Link Quality Indicator */
#endif
/* @endcond */
/** Reference counter */
u8_t ref;
};
/** @cond ignore */
/* The interface real ll address */
static inline struct net_linkaddr *net_pkt_ll_if(struct net_pkt *pkt)
{
return net_if_get_link_addr(pkt->iface);
}
static inline struct net_context *net_pkt_context(struct net_pkt *pkt)
{
return pkt->context;
}
static inline void net_pkt_set_context(struct net_pkt *pkt,
struct net_context *ctx)
{
pkt->context = ctx;
}
static inline void *net_pkt_token(struct net_pkt *pkt)
{
return pkt->token;
}
static inline void net_pkt_set_token(struct net_pkt *pkt, void *token)
{
pkt->token = token;
}
static inline struct net_if *net_pkt_iface(struct net_pkt *pkt)
{
return pkt->iface;
}
static inline void net_pkt_set_iface(struct net_pkt *pkt, struct net_if *iface)
{
pkt->iface = iface;
/* If the network interface is set in pkt, then also set the type of
* the network address that is stored in pkt. This is done here so
* that the address type is properly set and is not forgotten.
*/
pkt->lladdr_src.type = iface->link_addr.type;
pkt->lladdr_dst.type = iface->link_addr.type;
}
static inline u8_t net_pkt_family(struct net_pkt *pkt)
{
return pkt->family;
}
static inline void net_pkt_set_family(struct net_pkt *pkt, u8_t family)
{
pkt->family = family;
}
static inline u8_t net_pkt_ip_hdr_len(struct net_pkt *pkt)
{
return pkt->ip_hdr_len;
}
static inline void net_pkt_set_ip_hdr_len(struct net_pkt *pkt, u8_t len)
{
pkt->ip_hdr_len = len;
}
static inline u8_t *net_pkt_next_hdr(struct net_pkt *pkt)
{
return pkt->next_hdr;
}
static inline void net_pkt_set_next_hdr(struct net_pkt *pkt, u8_t *hdr)
{
pkt->next_hdr = hdr;
}
static inline u8_t net_pkt_sent(struct net_pkt *pkt)
{
return pkt->sent_or_eof;
}
static inline void net_pkt_set_sent(struct net_pkt *pkt, bool sent)
{
pkt->sent_or_eof = sent;
}
static inline u8_t net_pkt_queued(struct net_pkt *pkt)
{
return pkt->pkt_queued;
}
static inline void net_pkt_set_queued(struct net_pkt *pkt, bool send)
{
pkt->pkt_queued = send;
}
#if defined(CONFIG_NET_SOCKETS)
static inline u8_t net_pkt_eof(struct net_pkt *pkt)
{
return pkt->sent_or_eof;
}
static inline void net_pkt_set_eof(struct net_pkt *pkt, bool eof)
{
pkt->sent_or_eof = eof;
}
#endif
#if defined(CONFIG_NET_ROUTE)
static inline bool net_pkt_forwarding(struct net_pkt *pkt)
{
return pkt->forwarding;
}
static inline void net_pkt_set_forwarding(struct net_pkt *pkt, bool forward)
{
pkt->forwarding = forward;
}
#else
static inline bool net_pkt_forwarding(struct net_pkt *pkt)
{
return false;
}
#endif
#if defined(CONFIG_NET_IPV4)
static inline u8_t net_pkt_ipv4_ttl(struct net_pkt *pkt)
{
return pkt->ipv4_ttl;
}
static inline void net_pkt_set_ipv4_ttl(struct net_pkt *pkt,
u8_t ttl)
{
pkt->ipv4_ttl = ttl;
}
#endif
#if defined(CONFIG_NET_IPV6)
static inline u8_t net_pkt_ipv6_ext_opt_len(struct net_pkt *pkt)
{
return pkt->ipv6_ext_opt_len;
}
static inline void net_pkt_set_ipv6_ext_opt_len(struct net_pkt *pkt,
u8_t len)
{
pkt->ipv6_ext_opt_len = len;
}
static inline u8_t net_pkt_ipv6_ext_len(struct net_pkt *pkt)
{
return pkt->ipv6_ext_len;
}
static inline void net_pkt_set_ipv6_ext_len(struct net_pkt *pkt, u8_t len)
{
pkt->ipv6_ext_len = len;
}
static inline u16_t net_pkt_ipv6_hdr_prev(struct net_pkt *pkt)
{
return pkt->ipv6_prev_hdr_start;
}
static inline void net_pkt_set_ipv6_hdr_prev(struct net_pkt *pkt,
u16_t offset)
{
pkt->ipv6_prev_hdr_start = offset;
}
static inline u8_t net_pkt_ipv6_hop_limit(struct net_pkt *pkt)
{
return pkt->ipv6_hop_limit;
}
static inline void net_pkt_set_ipv6_hop_limit(struct net_pkt *pkt,
u8_t hop_limit)
{
pkt->ipv6_hop_limit = hop_limit;
}
#if defined(CONFIG_NET_IPV6_FRAGMENT)
static inline u8_t *net_pkt_ipv6_fragment_start(struct net_pkt *pkt)
{
return pkt->ipv6_frag_hdr_start;
}
static inline void net_pkt_set_ipv6_fragment_start(struct net_pkt *pkt,
u8_t *start)
{
pkt->ipv6_frag_hdr_start = start;
}
static inline u16_t net_pkt_ipv6_fragment_offset(struct net_pkt *pkt)
{
return pkt->ipv6_fragment_offset;
}
static inline void net_pkt_set_ipv6_fragment_offset(struct net_pkt *pkt,
u16_t offset)
{
pkt->ipv6_fragment_offset = offset;
}
static inline u32_t net_pkt_ipv6_fragment_id(struct net_pkt *pkt)
{
return pkt->ipv6_fragment_id;
}
static inline void net_pkt_set_ipv6_fragment_id(struct net_pkt *pkt,
u32_t id)
{
pkt->ipv6_fragment_id = id;
}
#endif /* CONFIG_NET_IPV6_FRAGMENT */
#else /* CONFIG_NET_IPV6 */
#define net_pkt_ipv6_ext_len(...) 0
#define net_pkt_set_ipv6_ext_len(...)
#endif /* CONFIG_NET_IPV6 */
static inline size_t net_pkt_get_len(struct net_pkt *pkt)
{
return net_buf_frags_len(pkt->frags);
}
static inline u8_t *net_pkt_ip_data(struct net_pkt *pkt)
{
return pkt->frags->data;
}
static inline u8_t *net_pkt_appdata(struct net_pkt *pkt)
{
return pkt->appdata;
}
static inline void net_pkt_set_appdata(struct net_pkt *pkt, u8_t *data)
{
pkt->appdata = data;
}
static inline u16_t net_pkt_appdatalen(struct net_pkt *pkt)
{
return pkt->appdatalen;
}
static inline void net_pkt_set_appdatalen(struct net_pkt *pkt, u16_t len)
{
pkt->appdatalen = len;
}
static inline u8_t net_pkt_ll_reserve(struct net_pkt *pkt)
{
return pkt->ll_reserve;
}
static inline void net_pkt_set_ll_reserve(struct net_pkt *pkt, u8_t len)
{
pkt->ll_reserve = len;
}
static inline u8_t *net_pkt_ll(struct net_pkt *pkt)
{
return net_pkt_ip_data(pkt) - net_pkt_ll_reserve(pkt);
}
static inline struct net_linkaddr *net_pkt_ll_src(struct net_pkt *pkt)
{
return &pkt->lladdr_src;
}
static inline struct net_linkaddr *net_pkt_ll_dst(struct net_pkt *pkt)
{
return &pkt->lladdr_dst;
}
static inline void net_pkt_ll_clear(struct net_pkt *pkt)
{
memset(net_pkt_ll(pkt), 0, net_pkt_ll_reserve(pkt));
net_pkt_ll_src(pkt)->addr = NULL;
net_pkt_ll_src(pkt)->len = 0;
}
static inline void net_pkt_ll_swap(struct net_pkt *pkt)
{
u8_t *addr = net_pkt_ll_src(pkt)->addr;
net_pkt_ll_src(pkt)->addr = net_pkt_ll_dst(pkt)->addr;
net_pkt_ll_dst(pkt)->addr = addr;
}
#if defined(CONFIG_IEEE802154) || defined(CONFIG_IEEE802154_RAW_MODE)
static inline u8_t net_pkt_ieee802154_rssi(struct net_pkt *pkt)
{
return pkt->ieee802154_rssi;
}
static inline void net_pkt_set_ieee802154_rssi(struct net_pkt *pkt,
u8_t rssi)
{
pkt->ieee802154_rssi = rssi;
}
static inline u8_t net_pkt_ieee802154_lqi(struct net_pkt *pkt)
{
return pkt->ieee802154_lqi;
}
static inline void net_pkt_set_ieee802154_lqi(struct net_pkt *pkt,
u8_t lqi)
{
pkt->ieee802154_lqi = lqi;
}
#endif
#define NET_IPV6_HDR(pkt) ((struct net_ipv6_hdr *)net_pkt_ip_data(pkt))
#define NET_IPV4_HDR(pkt) ((struct net_ipv4_hdr *)net_pkt_ip_data(pkt))
static inline void net_pkt_set_src_ipv6_addr(struct net_pkt *pkt)
{
net_if_ipv6_select_src_addr(net_context_get_iface(
net_pkt_context(pkt)),
&NET_IPV6_HDR(pkt)->src);
}
/* @endcond */
/**
* @brief Create a net_pkt slab
*
* A net_pkt slab is used to store meta-information about
* network packets. It must be coupled with a data fragment pool
* (:c:macro:`NET_PKT_DATA_POOL_DEFINE`) used to store the actual
* packet data. The macro can be used by an application to define
* additional custom per-context TX packet slabs (see
* :c:func:`net_context_setup_pools`).
*
* @param name Name of the slab.
* @param count Number of net_pkt in this slab.
*/
#define NET_PKT_SLAB_DEFINE(name, count) \
K_MEM_SLAB_DEFINE(name, sizeof(struct net_pkt), count, 4)
/* Backward compatibility macro */
#define NET_PKT_TX_SLAB_DEFINE(name, count) NET_PKT_SLAB_DEFINE(name, count)
/**
* @brief Create a data fragment net_buf pool
*
* A net_buf pool is used to store actual data for
* network packets. It must be coupled with a net_pkt slab
* (:c:macro:`NET_PKT_SLAB_DEFINE`) used to store the packet
* meta-information. The macro can be used by an application to
* define additional custom per-context TX packet pools (see
* :c:func:`net_context_setup_pools`).
*
* @param name Name of the pool.
* @param count Number of net_buf in this pool.
*/
#define NET_PKT_DATA_POOL_DEFINE(name, count) \
NET_BUF_POOL_DEFINE(name, count, CONFIG_NET_BUF_DATA_SIZE, \
CONFIG_NET_BUF_USER_DATA_SIZE, NULL)
#if defined(CONFIG_NET_DEBUG_NET_PKT)
/* Debug versions of the net_pkt functions that are used when tracking
* buffer usage.
*/
struct net_pkt *net_pkt_get_reserve_debug(struct k_mem_slab *slab,
u16_t reserve_head,
s32_t timeout,
const char *caller,
int line);
#define net_pkt_get_reserve(slab, reserve_head, timeout) \
net_pkt_get_reserve_debug(slab, reserve_head, timeout, \
__func__, __LINE__)
struct net_buf *net_pkt_get_reserve_data_debug(struct net_buf_pool *pool,
u16_t reserve_head,
s32_t timeout,
const char *caller,
int line);
#define net_pkt_get_reserve_data(pool, reserve_head, timeout) \
net_pkt_get_reserve_debug(pool, reserve_head, timeout, \
__func__, __LINE__)
struct net_pkt *net_pkt_get_rx_debug(struct net_context *context,
s32_t timeout,
const char *caller, int line);
#define net_pkt_get_rx(context, timeout) \
net_pkt_get_rx_debug(context, timeout, __func__, __LINE__)
struct net_pkt *net_pkt_get_tx_debug(struct net_context *context,
s32_t timeout,
const char *caller, int line);
#define net_pkt_get_tx(context, timeout) \
net_pkt_get_tx_debug(context, timeout, __func__, __LINE__)
struct net_buf *net_pkt_get_data_debug(struct net_context *context,
s32_t timeout,
const char *caller, int line);
#define net_pkt_get_data(context, timeout) \
net_pkt_get_data_debug(context, timeout, __func__, __LINE__)
struct net_pkt *net_pkt_get_reserve_rx_debug(u16_t reserve_head,
s32_t timeout,
const char *caller, int line);
#define net_pkt_get_reserve_rx(res, timeout) \
net_pkt_get_reserve_rx_debug(res, timeout, __func__, __LINE__)
struct net_pkt *net_pkt_get_reserve_tx_debug(u16_t reserve_head,
s32_t timeout,
const char *caller, int line);
#define net_pkt_get_reserve_tx(res, timeout) \
net_pkt_get_reserve_tx_debug(res, timeout, __func__, __LINE__)
struct net_buf *net_pkt_get_reserve_rx_data_debug(u16_t reserve_head,
s32_t timeout,
const char *caller,
int line);
#define net_pkt_get_reserve_rx_data(res, timeout) \
net_pkt_get_reserve_rx_data_debug(res, timeout, __func__, __LINE__)
struct net_buf *net_pkt_get_reserve_tx_data_debug(u16_t reserve_head,
s32_t timeout,
const char *caller,
int line);
#define net_pkt_get_reserve_tx_data(res, timeout) \
net_pkt_get_reserve_tx_data_debug(res, timeout, __func__, __LINE__)
struct net_buf *net_pkt_get_frag_debug(struct net_pkt *pkt,
s32_t timeout,
const char *caller, int line);
#define net_pkt_get_frag(pkt, timeout) \
net_pkt_get_frag_debug(pkt, timeout, __func__, __LINE__)
void net_pkt_unref_debug(struct net_pkt *pkt, const char *caller, int line);
#define net_pkt_unref(pkt) net_pkt_unref_debug(pkt, __func__, __LINE__)
struct net_pkt *net_pkt_ref_debug(struct net_pkt *pkt, const char *caller,
int line);
#define net_pkt_ref(pkt) net_pkt_ref_debug(pkt, __func__, __LINE__)
struct net_buf *net_pkt_frag_ref_debug(struct net_buf *frag,
const char *caller, int line);
#define net_pkt_frag_ref(frag) net_pkt_frag_ref_debug(frag, __func__, __LINE__)
void net_pkt_frag_unref_debug(struct net_buf *frag,
const char *caller, int line);
#define net_pkt_frag_unref(frag) \
net_pkt_frag_unref_debug(frag, __func__, __LINE__)
struct net_buf *net_pkt_frag_del_debug(struct net_pkt *pkt,
struct net_buf *parent,
struct net_buf *frag,
const char *caller, int line);
#define net_pkt_frag_del(pkt, parent, frag) \
net_pkt_frag_del_debug(pkt, parent, frag, __func__, __LINE__)
void net_pkt_frag_add_debug(struct net_pkt *pkt, struct net_buf *frag,
const char *caller, int line);
#define net_pkt_frag_add(pkt, frag) \
net_pkt_frag_add_debug(pkt, frag, __func__, __LINE__)
void net_pkt_frag_insert_debug(struct net_pkt *pkt, struct net_buf *frag,
const char *caller, int line);
#define net_pkt_frag_insert(pkt, frag) \
net_pkt_frag_insert_debug(pkt, frag, __func__, __LINE__)
/**
* @brief Print fragment list and the fragment sizes
*
* @details Only available if debugging is activated.
*
* @param pkt Network pkt.
*/
void net_pkt_print_frags(struct net_pkt *pkt);
#else /* CONFIG_NET_DEBUG_NET_PKT */
#define net_pkt_print_frags(...)
/**
* @brief Get packet from the given packet slab.
*
* @details Get network packet from the specific packet slab.
*
* @param slab Network packet slab.
* @param reserve_head How many bytes to reserve for headroom.
* @param timeout Affects the action taken should the net pkt slab be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network packet if successful, NULL otherwise.
*/
struct net_pkt *net_pkt_get_reserve(struct k_mem_slab *slab,
u16_t reserve_head,
s32_t timeout);
/**
* @brief Get packet from the RX packet slab.
*
* @details Get network packet from RX packet slab. You must have
* network context before able to use this function.
*
* @param context Network context that will be related to this packet.
* @param timeout Affects the action taken should the net pkt slab be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network packet if successful, NULL otherwise.
*/
struct net_pkt *net_pkt_get_rx(struct net_context *context,
s32_t timeout);
/**
* @brief Get packet from the TX packets slab.
*
* @details Get network packet from TX packet slab. You must have
* network context before able to use this function.
*
* @param context Network context that will be related to
* this packet.
* @param timeout Affects the action taken should the net pkt slab be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network packet if successful, NULL otherwise.
*/
struct net_pkt *net_pkt_get_tx(struct net_context *context,
s32_t timeout);
/**
* @brief Get buffer from the DATA buffers pool.
*
* @details Get network buffer from DATA buffer pool. You must have
* network context before able to use this function.
*
* @param context Network context that will be related to
* this buffer.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network buffer if successful, NULL otherwise.
*/
struct net_buf *net_pkt_get_data(struct net_context *context,
s32_t timeout);
/**
* @brief Get RX packet from slab but also reserve headroom for
* potential headers.
*
* @details Normally this version is not useful for applications
* but is mainly used by network fragmentation code.
*
* @param reserve_head How many bytes to reserve for headroom.
* @param timeout Affects the action taken should the net pkt slab be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network packet if successful, NULL otherwise.
*/
struct net_pkt *net_pkt_get_reserve_rx(u16_t reserve_head,
s32_t timeout);
/**
* @brief Get TX packet from slab but also reserve headroom for
* potential headers.
*
* @details Normally this version is not useful for applications
* but is mainly used by network fragmentation code.
*
* @param reserve_head How many bytes to reserve for headroom.
* @param timeout Affects the action taken should the net pkt slab be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network packet if successful, NULL otherwise.
*/
struct net_pkt *net_pkt_get_reserve_tx(u16_t reserve_head,
s32_t timeout);
/**
* @brief Get RX DATA buffer from pool but also reserve headroom for
* potential headers. Normally you should use net_pkt_get_frag() instead.
*
* @details Normally this version is not useful for applications
* but is mainly used by network fragmentation code.
*
* @param reserve_head How many bytes to reserve for headroom.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network buffer if successful, NULL otherwise.
*/
struct net_buf *net_pkt_get_reserve_rx_data(u16_t reserve_head,
s32_t timeout);
/**
* @brief Get TX DATA buffer from pool but also reserve headroom for
* potential headers. Normally you should use net_pkt_get_frag() instead.
*
* @details Normally this version is not useful for applications
* but is mainly used by network fragmentation code.
*
* @param reserve_head How many bytes to reserve for headroom.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network buffer if successful, NULL otherwise.
*/
struct net_buf *net_pkt_get_reserve_tx_data(u16_t reserve_head,
s32_t timeout);
/**
* @brief Get a data fragment that might be from user specific
* buffer pool or from global DATA pool.
*
* @param pkt Network packet.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Network buffer if successful, NULL otherwise.
*/
struct net_buf *net_pkt_get_frag(struct net_pkt *pkt, s32_t timeout);
/**
* @brief Place packet back into the available packets slab
*
* @details Releases the packet to other use. This needs to be
* called by application after it has finished with the packet.
*
* @param pkt Network packet to release.
*
*/
void net_pkt_unref(struct net_pkt *pkt);
/**
* @brief Increase the packet ref count
*
* @details Mark the packet to be used still.
*
* @param pkt Network packet to ref.
*
* @return Network packet if successful, NULL otherwise.
*/
struct net_pkt *net_pkt_ref(struct net_pkt *pkt);
/**
* @brief Increase the packet fragment ref count
*
* @details Mark the fragment to be used still.
*
* @param frag Network fragment to ref.
*
* @return a pointer on the referenced Network fragment.
*/
struct net_buf *net_pkt_frag_ref(struct net_buf *frag);
/**
* @brief Decrease the packet fragment ref count
*
* @param frag Network fragment to unref.
*/
void net_pkt_frag_unref(struct net_buf *frag);
/**
* @brief Delete existing fragment from a packet
*
* @param pkt Network packet from which frag belongs to.
* @param parent parent fragment of frag, or NULL if none.
* @param frag Fragment to delete.
*
* @return Pointer to the following fragment, or NULL if it had no
* further fragments.
*/
struct net_buf *net_pkt_frag_del(struct net_pkt *pkt,
struct net_buf *parent,
struct net_buf *frag);
/**
* @brief Add a fragment to a packet at the end of its fragment list
*
* @param pkt pkt Network packet where to add the fragment
* @param frag Fragment to add
*/
void net_pkt_frag_add(struct net_pkt *pkt, struct net_buf *frag);
/**
* @brief Insert a fragment to a packet at the beginning of its fragment list
*
* @param pkt pkt Network packet where to insert the fragment
* @param frag Fragment to insert
*/
void net_pkt_frag_insert(struct net_pkt *pkt, struct net_buf *frag);
#endif /* CONFIG_NET_DEBUG_NET_PKT */
/**
* @brief Copy a packet fragment list while reserving some extra space
* in destination buffer before a copy.
*
* @param pkt Network packet.
* @param amount Max amount of data to be copied.
* @param reserve Amount of extra data (this is not link layer header) in the
* first data fragment that is returned. The function will copy the original
* buffer right after the reserved bytes in the first destination fragment.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return New fragment list if successful, NULL otherwise.
*/
struct net_buf *net_pkt_copy(struct net_pkt *pkt, size_t amount,
size_t reserve, s32_t timeout);
/**
* @brief Copy a packet fragment list while reserving some extra space
* in destination buffer before a copy.
*
* @param pkt Network packet.
* @param reserve Amount of extra data (this is not link layer header) in the
* first data fragment that is returned. The function will copy the original
* buffer right after the reserved bytes in the first destination fragment.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return New fragment list if successful, NULL otherwise.
*/
static inline struct net_buf *net_pkt_copy_all(struct net_pkt *pkt,
size_t reserve,
s32_t timeout)
{
return net_pkt_copy(pkt, net_buf_frags_len(pkt->frags),
reserve, timeout);
}
/**
* @brief Copy len bytes from src starting from offset to dst
*
* This routine assumes that dst is formed of one fragment with enough space
* to store @a len bytes starting from offset at src.
*
* @param dst Destination buffer
* @param src Source buffer that may be fragmented
* @param offset Starting point to copy from
* @param len Number of bytes to copy
* @return 0 on success
* @return -ENOMEM on error
*/
int net_frag_linear_copy(struct net_buf *dst, struct net_buf *src,
u16_t offset, u16_t len);
/**
* @brief Copy len bytes from src starting from offset to dst buffer
*
* This routine assumes that dst is large enough to store @a len bytes
* starting from offset at src.
*
* @param dst Destination buffer
* @param dst_len Destination buffer max length
* @param src Source buffer that may be fragmented
* @param offset Starting point to copy from
* @param len Number of bytes to copy
* @return number of bytes copied if everything is ok
* @return -ENOMEM on error
*/
int net_frag_linearize(u8_t *dst, size_t dst_len,
struct net_pkt *src, u16_t offset, u16_t len);
/**
* @brief Compact the fragment list of a packet.
*
* @details After this there is no more any free space in individual fragments.
* @param pkt Network packet.
*
* @return True if compact success, False otherwise.
*/
bool net_pkt_compact(struct net_pkt *pkt);
/**
* @brief Append data to fragment list of a packet
*
* @details Append data to last fragment. If there is not enough space in
* last fragment then more data fragments will be added, unless there are
* no free fragments and timeout occurs.
*
* @param pkt Network packet.
* @param len Total length of input data
* @param data Data to be added
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Length of data actually added. This may be less than input
* length if other timeout than K_FOREVER was used, and there
* were no free fragments in a pool to accommodate all data.
*/
u16_t net_pkt_append(struct net_pkt *pkt, u16_t len, const u8_t *data,
s32_t timeout);
/**
* @brief Append all data to fragment list of a packet (or fail)
*
* @details Append data to last fragment. If there is not enough space in
* last fragment then more data fragments will be added. Return unsuccessful
* status if there are no free fragments to accommodate all data and timeout
* occurs.
*
* @param pkt Network packet.
* @param len Total length of input data
* @param data Data to be added
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return True if all the data is placed at end of fragment list,
* false otherwise (in which case packet may contain incomplete
* input data).
*/
static inline bool net_pkt_append_all(struct net_pkt *pkt, u16_t len,
const u8_t *data, s32_t timeout)
{
return net_pkt_append(pkt, len, data, timeout) == len;
}
/**
* @brief Append u8_t data to last fragment in fragment list of a packet
*
* @details Append data to last fragment. If there is not enough space in last
* fragment then new data fragment will be created and will be added to
* fragment list. Caller has to take care of endianness if needed.
*
* @param pkt Network packet.
* @param data Data to be added
*
* @return True if all the data is placed at end of fragment list,
* False otherwise (In-case of false pkt might contain input
* data in the process of placing into fragments).
*/
static inline bool net_pkt_append_u8(struct net_pkt *pkt, u8_t data)
{
return net_pkt_append_all(pkt, 1, &data, K_FOREVER);
}
/**
* @brief Append u16_t data to last fragment in fragment list of a packet
*
* @details Append data to last fragment. If there is not enough space in last
* fragment then new data fragment will be created and will be added to
* fragment list. Caller has to take care of endianness if needed.
*
* @param pkt Network packet.
* @param data Data to be added
*
* @return True if all the data is placed at end of fragment list,
* False otherwise (In-case of false pkt might contain input data
* in the process of placing into fragments).
*/
static inline bool net_pkt_append_be16(struct net_pkt *pkt, u16_t data)
{
u16_t value = sys_cpu_to_be16(data);
return net_pkt_append_all(pkt, sizeof(u16_t), (u8_t *)&value,
K_FOREVER);
}
/**
* @brief Append u32_t data to last fragment in fragment list of a packet
*
* @details Append data to last fragment. If there is not enough space in last
* fragment then new data fragment will be created and will be added to
* fragment list. Caller has to take care of endianness if needed.
*
* @param pkt Network packet.
* @param data Data to be added
*
* @return True if all the data is placed at end of fragment list,
* False otherwise (In-case of false pkt might contain input data
* in the process of placing into fragments).
*/
static inline bool net_pkt_append_be32(struct net_pkt *pkt, u32_t data)
{
u32_t value = sys_cpu_to_be32(data);
return net_pkt_append_all(pkt, sizeof(u32_t), (u8_t *)&value,
K_FOREVER);
}
/**
* @brief Append u32_t data to last fragment in fragment list
*
* @details Append data to last fragment. If there is not enough space in last
* fragment then new data fragment will be created and will be added to
* fragment list. Convert data to LE.
*
* @param pkt Network packet fragment list.
* @param data Data to be added
*
* @return True if all the data is placed at end of fragment list,
* False otherwise (In-case of false pkt might contain input data
* in the process of placing into fragments).
*/
static inline bool net_pkt_append_le32(struct net_pkt *pkt, u32_t data)
{
u32_t value = sys_cpu_to_le32(data);
return net_pkt_append_all(pkt, sizeof(u32_t), (u8_t *)&value,
K_FOREVER);
}
/**
* @brief Get data from buffer
*
* @details Get N number of bytes starting from fragment's offset. If the total
* length of data is placed in multiple fragments, this function will read from
* all fragments until it reaches N number of bytes. Caller has to take care of
* endianness if needed.
*
* @param frag Network buffer fragment.
* @param offset Offset of input buffer.
* @param pos Pointer to position of offset after reading n number of bytes,
* this is with respect to return buffer(fragment).
* @param len Total length of data to be read.
* @param data Data will be copied here.
*
* @return Pointer to the fragment or
* NULL and pos is 0 after successful read,
* NULL and pos is 0xffff otherwise.
*/
struct net_buf *net_frag_read(struct net_buf *frag, u16_t offset,
u16_t *pos, u16_t len, u8_t *data);
/**
* @brief Skip N number of bytes while reading buffer
*
* @details Skip N number of bytes starting from fragment's offset. If the total
* length of data is placed in multiple fragments, this function will skip from
* all fragments until it reaches N number of bytes. This function is useful
* when unwanted data (e.g. reserved or not supported data in message) is part
* of fragment and we want to skip it.
*
* @param frag Network buffer fragment.
* @param offset Offset of input buffer.
* @param pos Pointer to position of offset after reading n number of bytes,
* this is with respect to return buffer(fragment).
* @param len Total length of data to be read.
*
* @return Pointer to the fragment or
* NULL and pos is 0 after successful skip,
* NULL and pos is 0xffff otherwise.
*/
static inline struct net_buf *net_frag_skip(struct net_buf *frag,
u16_t offset,
u16_t *pos, u16_t len)
{
return net_frag_read(frag, offset, pos, len, NULL);
}
/**
* @brief Get a byte value from fragmented buffer
*
* @param frag Network buffer fragment.
* @param offset Offset of input buffer.
* @param pos Pointer to position of offset after reading 2 bytes,
* this is with respect to return buffer(fragment).
* @param value Value is returned
*
* @return Pointer to fragment after successful read,
* NULL otherwise (if pos is 0, NULL is not a failure case).
*/
static inline struct net_buf *net_frag_read_u8(struct net_buf *frag,
u16_t offset,
u16_t *pos,
u8_t *value)
{
return net_frag_read(frag, offset, pos, 1, value);
}
/**
* @brief Get 16 bit big endian value from fragmented buffer
*
* @param frag Network buffer fragment.
* @param offset Offset of input buffer.
* @param pos Pointer to position of offset after reading 2 bytes,
* this is with respect to return buffer(fragment).
* @param value Value is returned
*
* @return Pointer to fragment after successful read,
* NULL otherwise (if pos is 0, NULL is not a failure case).
*/
struct net_buf *net_frag_read_be16(struct net_buf *frag, u16_t offset,
u16_t *pos, u16_t *value);
/**
* @brief Get 32 bit big endian value from fragmented buffer
*
* @param frag Network buffer fragment.
* @param offset Offset of input buffer.
* @param pos Pointer to position of offset after reading 4 bytes,
* this is with respect to return buffer(fragment).
* @param value Value is returned
*
* @return Pointer to fragment after successful read,
* NULL otherwise (if pos is 0, NULL is not a failure case).
*/
struct net_buf *net_frag_read_be32(struct net_buf *frag, u16_t offset,
u16_t *pos, u32_t *value);
/**
* @brief Write data to an arbitrary offset in fragments list of a packet.
*
* @details Write data to an arbitrary offset in a series of fragments.
* Offset is based on fragment 'size' and calculates from input fragment
* starting position.
*
* Size in this context refers the fragment full size without link layer header
* part. The fragment might have user written data in it, the amount of such
* data is stored in frag->len variable (the frag->len is always <= frag->size).
* If using this API, the tailroom in the fragments will be taken into use.
*
* If offset is more than already allocated length in fragment, then empty space
* or extra empty fragments is created to reach proper offset.
* If there is any data present on input fragment offset, then it will be
* 'overwritten'. Use net_pkt_insert() api if you don't want to overwrite.
*
* Offset is calculated from starting point of data area in input fragment.
* e.g. Pkt(Tx/Rx) - Frag1 - Frag2 - Frag3 - Frag4
* (Assume FRAG DATA SIZE is 100 bytes after link layer header)
*
* 1) net_pkt_write(pkt, frag2, 20, &pos, 20, data, K_FOREVER)
* In this case write starts from "frag2->data + 20",
* returns frag2, pos = 40
*
* 2) net_pkt_write(pkt, frag1, 150, &pos, 60, data, K_FOREVER)
* In this case write starts from "frag2->data + 50"
* returns frag3, pos = 10
*
* 3) net_pkt_write(pkt, frag1, 350, &pos, 30, data, K_FOREVER)
* In this case write starts from "frag4->data + 50"
* returns frag4, pos = 80
*
* 4) net_pkt_write(pkt, frag2, 110, &pos, 90, data, K_FOREVER)
* In this case write starts from "frag3->data + 10"
* returns frag4, pos = 0
*
* 5) net_pkt_write(pkt, frag4, 110, &pos, 20, data, K_FOREVER)
* In this case write creates new data fragment and starts from
* "frag5->data + 10"
* returns frag5, pos = 30
*
* If input argument frag is NULL, it will create new data fragment
* and append at the end of fragment list.
*
* @param pkt Network packet.
* @param frag Network buffer fragment.
* @param offset Offset
* @param pos Position of offset after write completed (this will be
* relative to return fragment)
* @param len Length of the data to be written.
* @param data Data to be written
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return Pointer to the fragment and position (*pos) where write ended,
* NULL and pos is 0xffff otherwise.
*/
struct net_buf *net_pkt_write(struct net_pkt *pkt, struct net_buf *frag,
u16_t offset, u16_t *pos, u16_t len,
u8_t *data, s32_t timeout);
/* Write u8_t data to an arbitrary offset in fragment. */
static inline struct net_buf *net_pkt_write_u8(struct net_pkt *pkt,
struct net_buf *frag,
u16_t offset,
u16_t *pos,
u8_t data)
{
return net_pkt_write(pkt, frag, offset, pos, sizeof(u8_t),
&data, K_FOREVER);
}
/* Write u16_t big endian value to an arbitrary offset in fragment. */
static inline struct net_buf *net_pkt_write_be16(struct net_pkt *pkt,
struct net_buf *frag,
u16_t offset,
u16_t *pos,
u16_t data)
{
u16_t value = htons(data);
return net_pkt_write(pkt, frag, offset, pos, sizeof(u16_t),
(u8_t *)&value, K_FOREVER);
}
/* Write u32_t big endian value to an arbitrary offset in fragment. */
static inline struct net_buf *net_pkt_write_be32(struct net_pkt *pkt,
struct net_buf *frag,
u16_t offset,
u16_t *pos,
u32_t data)
{
u32_t value = htonl(data);
return net_pkt_write(pkt, frag, offset, pos, sizeof(u32_t),
(u8_t *)&value, K_FOREVER);
}
/**
* @brief Insert data at an arbitrary offset in a series of fragments.
*
* @details Insert data at an arbitrary offset in a series of fragments. Offset
* is based on fragment length (only user written data length, any tailroom
* in fragments does not come to consideration unlike net_pkt_write()) and
* calculates from input fragment starting position.
* If the data pointer is NULL, insert a sequence of zeros with the given
* length.
*
* Offset examples can be considered from net_pkt_write() api.
* If the offset is more than already allocated fragments length then it is an
* error case.
*
* @param pkt Network packet.
* @param frag Network buffer fragment.
* @param offset Offset of fragment where insertion will start.
* @param len Length of the data to be inserted.
* @param data Data to be inserted, can be NULL.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then
* wait as long as necessary. Otherwise, wait up to the specified
* number of milliseconds before timing out.
*
* @return True on success, False otherwise.
*/
bool net_pkt_insert(struct net_pkt *pkt, struct net_buf *frag,
u16_t offset, u16_t len, u8_t *data,
s32_t timeout);
/* Insert u8_t data at an arbitrary offset in a series of fragments. */
static inline bool net_pkt_insert_u8(struct net_pkt *pkt,
struct net_buf *frag,
u16_t offset,
u8_t data)
{
return net_pkt_insert(pkt, frag, offset, sizeof(u8_t), &data,
K_FOREVER);
}
/* Insert u16_t big endian value at an arbitrary offset in a series of
* fragments.
*/
static inline bool net_pkt_insert_be16(struct net_pkt *pkt,
struct net_buf *frag,
u16_t offset,
u16_t data)
{
u16_t value = htons(data);
return net_pkt_insert(pkt, frag, offset, sizeof(u16_t),
(u8_t *)&value, K_FOREVER);
}
/* Insert u32_t big endian value at an arbitrary offset in a series of
* fragments.
*/
static inline bool net_pkt_insert_be32(struct net_pkt *pkt,
struct net_buf *frag,
u16_t offset,
u32_t data)
{
u32_t value = htonl(data);
return net_pkt_insert(pkt, frag, offset, sizeof(u32_t),
(u8_t *)&value, K_FOREVER);
}
/**
* @brief Split a fragment to two parts at arbitrary offset.
*
* @details This will generate two new fragments (fragA and fragB) from
* one (orig_frag). The original fragment is not modified but two new
* fragments are allocated and returned to the caller. The original fragment
* must be part of the packet pointed by the pkt parameter. If the len parameter
* is larger than the amount of data in the orig fragment, then the fragA will
* contain all the data and fragB will be empty.
*
* @param pkt Network packet
* @param orig_frag Original network buffer fragment which is to be split.
* @param len Amount of data in the first returned fragment.
* @param fragA A fragment is returned. This will contain len bytes that
* are copied from start of orig_frag.
* @param fragB Another fragment is returned. This will contain remaining
* bytes (orig_frag->len - len) from the orig_frag or NULL if all the data
* was copied into fragA.
* @param timeout Affects the action taken should the net buf pool be empty.
* If K_NO_WAIT, then return immediately. If K_FOREVER, then wait as long as
* necessary. Otherwise, wait up to the specified number of milliseconds before
* timing out.
*
* @return 0 on success, <0 otherwise.
*/
int net_pkt_split(struct net_pkt *pkt, struct net_buf *orig_frag,
u16_t len, struct net_buf **fragA,
struct net_buf **fragB, s32_t timeout);
/**
* @brief Return the fragment and offset within it according to network
* packet offset.
*
* @details This is typically used to get the protocol header pointer when
* we know the offset. According to this information, the corresponding fragment
* and position within that fragment is returned.
*
* @param pkt Network packet
* @param offset Offset of desired location in network packet. For example, if
* we want to know where UDP header is located after the IPv6 header,
* the offset could have a value of sizeof(struct net_ipv6_hdr). Note that this
* is a simplified example that does not take into account the possible IPv6
* extension headers.
* @param pos Pointer to position within result fragment corresponding to
* offset param. For example, if the IPv6 header is split between two fragments,
* then if we want to know the start of UDP header, the returned pos variable
* would indicate how many bytes from second fragment the UDP header starts.
*
* @return Pointer to the fragment where the the offset is located or
* NULL if there is not enough bytes in the packet
*/
struct net_buf *net_frag_get_pos(struct net_pkt *pkt,
u16_t offset,
u16_t *pos);
/**
* @brief Clone pkt and its fragment chain.
*
* @param pkt Original pkt to be cloned
* @param timeout Timeout to wait for free net_buf
*
* @return NULL if error, clone fragment chain otherwise.
*/
struct net_pkt *net_pkt_clone(struct net_pkt *pkt, s32_t timeout);
/**
* @brief Get information about predefined RX, TX and DATA pools.
*
* @param rx Pointer to RX pool is returned.
* @param tx Pointer to TX pool is returned.
* @param rx_data Pointer to RX DATA pool is returned.
* @param tx_data Pointer to TX DATA pool is returned.
*/
void net_pkt_get_info(struct k_mem_slab **rx,
struct k_mem_slab **tx,
struct net_buf_pool **rx_data,
struct net_buf_pool **tx_data);
/**
* @brief Get source socket address.
*
* @param pkt Network packet
* @param addr Source socket address
* @param addrlen The length of source socket address
* @return 0 on success, <0 otherwise.
*/
int net_pkt_get_src_addr(struct net_pkt *pkt,
struct sockaddr *addr,
socklen_t addrlen);
/**
* @brief Get destination socket address.
*
* @param pkt Network packet
* @param addr Destination socket address
* @param addrlen The length of destination socket address
* @return 0 on success, <0 otherwise.
*/
int net_pkt_get_dst_addr(struct net_pkt *pkt,
struct sockaddr *addr,
socklen_t addrlen);
#if defined(CONFIG_NET_DEBUG_NET_PKT)
/**
* @brief Debug helper to print out the buffer allocations
*/
void net_pkt_print(void);
typedef void (*net_pkt_allocs_cb_t)(struct net_pkt *pkt,
struct net_buf *buf,
const char *func_alloc,
int line_alloc,
const char *func_free,
int line_free,
bool in_use,
void *user_data);
void net_pkt_allocs_foreach(net_pkt_allocs_cb_t cb, void *user_data);
const char *net_pkt_slab2str(struct k_mem_slab *slab);
const char *net_pkt_pool2str(struct net_buf_pool *pool);
#else
#define net_pkt_print(...)
#endif /* CONFIG_NET_DEBUG_NET_PKT */
/**
* @}
*/
#ifdef __cplusplus
}
#endif
#endif /* __NET_PKT_H__ */