/* SPDX-License-Identifier: GPL-2.0 */ /* * LiMon Monitor (LiMon) - Network. * * Copyright 1994 - 2000 Neil Russell. * (See License) * * History * 9/16/00 bor adapted to TQM823L/STK8xxL board, RARP/TFTP boot added */ #ifndef __NET_H__ #define __NET_H__ #include #include #include /* for nton* / ntoh* stuff */ #include #include #include #include #include struct bd_info; struct cmd_tbl; struct udevice; #define DEBUG_LL_STATE 0 /* Link local state machine changes */ #define DEBUG_DEV_PKT 0 /* Packets or info directed to the device */ #define DEBUG_NET_PKT 0 /* Packets on info on the network at large */ #define DEBUG_INT_STATE 0 /* Internal network state changes */ /* * The number of receive packet buffers, and the required packet buffer * alignment in memory. * */ #define PKTBUFSRX CONFIG_SYS_RX_ETH_BUFFER #define PKTALIGN ARCH_DMA_MINALIGN /* Number of packets processed together */ #define ETH_PACKETS_BATCH_RECV 32 /* ARP hardware address length */ #define ARP_HLEN 6 /* * The size of a MAC address in string form, each digit requires two chars * and five separator characters to form '00:00:00:00:00:00'. */ #define ARP_HLEN_ASCII (ARP_HLEN * 2) + (ARP_HLEN - 1) /* IPv4 addresses are always 32 bits in size */ struct in_addr { __be32 s_addr; }; /** * do_tftpb - Run the tftpboot command * * @cmdtp: Command information for tftpboot * @flag: Command flags (CMD_FLAG_...) * @argc: Number of arguments * @argv: List of arguments * Return: result (see enum command_ret_t) */ int do_tftpb(struct cmd_tbl *cmdtp, int flag, int argc, char *const argv[]); /** * dhcp_run() - Run DHCP on the current ethernet device * * This sets the autoload variable, then puts it back to similar to its original * state (y, n or unset). * * @addr: Address to load the file into (0 if @autoload is false) * @fname: Filename of file to load (NULL if @autoload is false or to use the * default filename) * @autoload: true to load the file, false to just get the network IP * @return 0 if OK, -EINVAL if the environment failed, -ENOENT if ant file was * not found */ int dhcp_run(ulong addr, const char *fname, bool autoload); /** * An incoming packet handler. * @param pkt pointer to the application packet * @param dport destination UDP port * @param sip source IP address * @param sport source UDP port * @param len packet length */ typedef void rxhand_f(uchar *pkt, unsigned dport, struct in_addr sip, unsigned sport, unsigned len); /** * An incoming ICMP packet handler. * @param type ICMP type * @param code ICMP code * @param dport destination UDP port * @param sip source IP address * @param sport source UDP port * @param pkt pointer to the ICMP packet data * @param len packet length */ typedef void rxhand_icmp_f(unsigned type, unsigned code, unsigned dport, struct in_addr sip, unsigned sport, uchar *pkt, unsigned len); /* * A timeout handler. Called after time interval has expired. */ typedef void thand_f(void); enum eth_state_t { ETH_STATE_INIT, ETH_STATE_PASSIVE, ETH_STATE_ACTIVE }; /** * struct eth_pdata - Platform data for Ethernet MAC controllers * * @iobase: The base address of the hardware registers * @enetaddr: The Ethernet MAC address that is loaded from EEPROM or env * @phy_interface: PHY interface to use - see PHY_INTERFACE_MODE_... * @max_speed: Maximum speed of Ethernet connection supported by MAC * @priv_pdata: device specific plat */ struct eth_pdata { phys_addr_t iobase; unsigned char enetaddr[ARP_HLEN]; int phy_interface; int max_speed; void *priv_pdata; }; enum eth_recv_flags { /* * Check hardware device for new packets (otherwise only return those * which are already in the memory buffer ready to process) */ ETH_RECV_CHECK_DEVICE = 1 << 0, }; /** * struct eth_ops - functions of Ethernet MAC controllers * * start: Prepare the hardware to send and receive packets * send: Send the bytes passed in "packet" as a packet on the wire * recv: Check if the hardware received a packet. If so, set the pointer to the * packet buffer in the packetp parameter. If not, return an error or 0 to * indicate that the hardware receive FIFO is empty. If 0 is returned, the * network stack will not process the empty packet, but free_pkt() will be * called if supplied * free_pkt: Give the driver an opportunity to manage its packet buffer memory * when the network stack is finished processing it. This will only be * called when no error was returned from recv - optional * stop: Stop the hardware from looking for packets - may be called even if * state == PASSIVE * mcast: Join or leave a multicast group (for TFTP) - optional * write_hwaddr: Write a MAC address to the hardware (used to pass it to Linux * on some platforms like ARM). This function expects the * eth_pdata::enetaddr field to be populated. The method can * return -ENOSYS to indicate that this is not implemented for this hardware - optional. * read_rom_hwaddr: Some devices have a backup of the MAC address stored in a * ROM on the board. This is how the driver should expose it * to the network stack. This function should fill in the * eth_pdata::enetaddr field - optional * set_promisc: Enable or Disable promiscuous mode */ struct eth_ops { int (*start)(struct udevice *dev); int (*send)(struct udevice *dev, void *packet, int length); int (*recv)(struct udevice *dev, int flags, uchar **packetp); int (*free_pkt)(struct udevice *dev, uchar *packet, int length); void (*stop)(struct udevice *dev); int (*mcast)(struct udevice *dev, const u8 *enetaddr, int join); int (*write_hwaddr)(struct udevice *dev); int (*read_rom_hwaddr)(struct udevice *dev); int (*set_promisc)(struct udevice *dev, bool enable); }; #define eth_get_ops(dev) ((struct eth_ops *)(dev)->driver->ops) struct udevice *eth_get_dev(void); /* get the current device */ /* * The devname can be either an exact name given by the driver or device tree * or it can be an alias of the form "eth%d" */ struct udevice *eth_get_dev_by_name(const char *devname); unsigned char *eth_get_ethaddr(void); /* get the current device MAC */ /* Used only when NetConsole is enabled */ int eth_is_active(struct udevice *dev); /* Test device for active state */ int eth_init_state_only(void); /* Set active state */ void eth_halt_state_only(void); /* Set passive state */ int eth_initialize(void); /* Initialize network subsystem */ void eth_try_another(int first_restart); /* Change the device */ void eth_set_current(void); /* set nterface to ethcur var */ int eth_get_dev_index(void); /* get the device index */ /** * eth_env_set_enetaddr_by_index() - set the MAC address environment variable * * This sets up an environment variable with the given MAC address (@enetaddr). * The environment variable to be set is defined by <@base_name><@index>addr. * If @index is 0 it is omitted. For common Ethernet this means ethaddr, * eth1addr, etc. * * @base_name: Base name for variable, typically "eth" * @index: Index of interface being updated (>=0) * @enetaddr: Pointer to MAC address to put into the variable * Return: 0 if OK, other value on error */ int eth_env_set_enetaddr_by_index(const char *base_name, int index, uchar *enetaddr); /* * Initialize USB ethernet device with CONFIG_DM_ETH * Returns: * 0 is success, non-zero is error status. */ int usb_ether_init(void); /* * Get the hardware address for an ethernet interface . * Args: * base_name - base name for device (normally "eth") * index - device index number (0 for first) * enetaddr - returns 6 byte hardware address * Returns: * Return true if the address is valid. */ int eth_env_get_enetaddr_by_index(const char *base_name, int index, uchar *enetaddr); int eth_init(void); /* Initialize the device */ int eth_send(void *packet, int length); /* Send a packet */ #if defined(CONFIG_API) || defined(CONFIG_EFI_LOADER) int eth_receive(void *packet, int length); /* Receive a packet*/ extern void (*push_packet)(void *packet, int length); #endif int eth_rx(void); /* Check for received packets */ void eth_halt(void); /* stop SCC */ const char *eth_get_name(void); /* get name of current device */ int eth_mcast_join(struct in_addr mcast_addr, int join); /**********************************************************************/ /* * Protocol headers. */ /* * Ethernet header */ struct ethernet_hdr { u8 et_dest[ARP_HLEN]; /* Destination node */ u8 et_src[ARP_HLEN]; /* Source node */ u16 et_protlen; /* Protocol or length */ } __attribute__((packed)); /* Ethernet header size */ #define ETHER_HDR_SIZE (sizeof(struct ethernet_hdr)) #define ETH_FCS_LEN 4 /* Octets in the FCS */ struct e802_hdr { u8 et_dest[ARP_HLEN]; /* Destination node */ u8 et_src[ARP_HLEN]; /* Source node */ u16 et_protlen; /* Protocol or length */ u8 et_dsap; /* 802 DSAP */ u8 et_ssap; /* 802 SSAP */ u8 et_ctl; /* 802 control */ u8 et_snap1; /* SNAP */ u8 et_snap2; u8 et_snap3; u16 et_prot; /* 802 protocol */ } __attribute__((packed)); /* 802 + SNAP + ethernet header size */ #define E802_HDR_SIZE (sizeof(struct e802_hdr)) /* * Virtual LAN Ethernet header */ struct vlan_ethernet_hdr { u8 vet_dest[ARP_HLEN]; /* Destination node */ u8 vet_src[ARP_HLEN]; /* Source node */ u16 vet_vlan_type; /* PROT_VLAN */ u16 vet_tag; /* TAG of VLAN */ u16 vet_type; /* protocol type */ } __attribute__((packed)); /* VLAN Ethernet header size */ #define VLAN_ETHER_HDR_SIZE (sizeof(struct vlan_ethernet_hdr)) #define PROT_IP 0x0800 /* IP protocol */ #define PROT_ARP 0x0806 /* IP ARP protocol */ #define PROT_WOL 0x0842 /* ether-wake WoL protocol */ #define PROT_RARP 0x8035 /* IP ARP protocol */ #define PROT_VLAN 0x8100 /* IEEE 802.1q protocol */ #define PROT_IPV6 0x86dd /* IPv6 over bluebook */ #define PROT_PPP_SES 0x8864 /* PPPoE session messages */ #define PROT_NCSI 0x88f8 /* NC-SI control packets */ #define IPPROTO_ICMP 1 /* Internet Control Message Protocol */ #define IPPROTO_TCP 6 /* Transmission Control Protocol */ #define IPPROTO_UDP 17 /* User Datagram Protocol */ /* * Internet Protocol (IP) header. */ struct ip_hdr { u8 ip_hl_v; /* header length and version */ u8 ip_tos; /* type of service */ u16 ip_len; /* total length */ u16 ip_id; /* identification */ u16 ip_off; /* fragment offset field */ u8 ip_ttl; /* time to live */ u8 ip_p; /* protocol */ u16 ip_sum; /* checksum */ struct in_addr ip_src; /* Source IP address */ struct in_addr ip_dst; /* Destination IP address */ } __attribute__((packed)); #define IP_OFFS 0x1fff /* ip offset *= 8 */ #define IP_FLAGS 0xe000 /* first 3 bits */ #define IP_FLAGS_RES 0x8000 /* reserved */ #define IP_FLAGS_DFRAG 0x4000 /* don't fragments */ #define IP_FLAGS_MFRAG 0x2000 /* more fragments */ #define IP_HDR_SIZE (sizeof(struct ip_hdr)) #define IP_MIN_FRAG_DATAGRAM_SIZE (IP_HDR_SIZE + 8) /* * Internet Protocol (IP) + UDP header. */ struct ip_udp_hdr { u8 ip_hl_v; /* header length and version */ u8 ip_tos; /* type of service */ u16 ip_len; /* total length */ u16 ip_id; /* identification */ u16 ip_off; /* fragment offset field */ u8 ip_ttl; /* time to live */ u8 ip_p; /* protocol */ u16 ip_sum; /* checksum */ struct in_addr ip_src; /* Source IP address */ struct in_addr ip_dst; /* Destination IP address */ u16 udp_src; /* UDP source port */ u16 udp_dst; /* UDP destination port */ u16 udp_len; /* Length of UDP packet */ u16 udp_xsum; /* Checksum */ } __attribute__((packed)); #define IP_UDP_HDR_SIZE (sizeof(struct ip_udp_hdr)) #define UDP_HDR_SIZE (IP_UDP_HDR_SIZE - IP_HDR_SIZE) /* * Address Resolution Protocol (ARP) header. */ struct arp_hdr { u16 ar_hrd; /* Format of hardware address */ # define ARP_ETHER 1 /* Ethernet hardware address */ u16 ar_pro; /* Format of protocol address */ u8 ar_hln; /* Length of hardware address */ u8 ar_pln; /* Length of protocol address */ # define ARP_PLEN 4 u16 ar_op; /* Operation */ # define ARPOP_REQUEST 1 /* Request to resolve address */ # define ARPOP_REPLY 2 /* Response to previous request */ # define RARPOP_REQUEST 3 /* Request to resolve address */ # define RARPOP_REPLY 4 /* Response to previous request */ /* * The remaining fields are variable in size, according to * the sizes above, and are defined as appropriate for * specific hardware/protocol combinations. */ u8 ar_data[0]; #define ar_sha ar_data[0] #define ar_spa ar_data[ARP_HLEN] #define ar_tha ar_data[ARP_HLEN + ARP_PLEN] #define ar_tpa ar_data[ARP_HLEN + ARP_PLEN + ARP_HLEN] #if 0 u8 ar_sha[]; /* Sender hardware address */ u8 ar_spa[]; /* Sender protocol address */ u8 ar_tha[]; /* Target hardware address */ u8 ar_tpa[]; /* Target protocol address */ #endif /* 0 */ } __attribute__((packed)); #define ARP_HDR_SIZE (8+20) /* Size assuming ethernet */ /* * ICMP stuff (just enough to handle (host) redirect messages) */ #define ICMP_ECHO_REPLY 0 /* Echo reply */ #define ICMP_NOT_REACH 3 /* Detination unreachable */ #define ICMP_REDIRECT 5 /* Redirect (change route) */ #define ICMP_ECHO_REQUEST 8 /* Echo request */ /* Codes for REDIRECT. */ #define ICMP_REDIR_NET 0 /* Redirect Net */ #define ICMP_REDIR_HOST 1 /* Redirect Host */ /* Codes for NOT_REACH */ #define ICMP_NOT_REACH_PORT 3 /* Port unreachable */ struct icmp_hdr { u8 type; u8 code; u16 checksum; union { struct { u16 id; u16 sequence; } echo; u32 gateway; struct { u16 unused; u16 mtu; } frag; u8 data[0]; } un; } __attribute__((packed)); #define ICMP_HDR_SIZE (sizeof(struct icmp_hdr)) #define IP_ICMP_HDR_SIZE (IP_HDR_SIZE + ICMP_HDR_SIZE) /* * Maximum packet size; used to allocate packet storage. Use * the maxium Ethernet frame size as specified by the Ethernet * standard including the 802.1Q tag (VLAN tagging). * maximum packet size = 1522 * maximum packet size and multiple of 32 bytes = 1536 */ #define PKTSIZE 1522 #ifndef CONFIG_DM_DSA #define PKTSIZE_ALIGN 1536 #else /* Maximum DSA tagging overhead (headroom and/or tailroom) */ #define DSA_MAX_OVR 256 #define PKTSIZE_ALIGN (1536 + DSA_MAX_OVR) #endif /* * Maximum receive ring size; that is, the number of packets * we can buffer before overflow happens. Basically, this just * needs to be enough to prevent a packet being discarded while * we are processing the previous one. */ #define RINGSZ 4 #define RINGSZ_LOG2 2 /**********************************************************************/ /* * Globals. * * Note: * * All variables of type struct in_addr are stored in NETWORK byte order * (big endian). */ /* net.c */ /** BOOTP EXTENTIONS **/ extern struct in_addr net_gateway; /* Our gateway IP address */ extern struct in_addr net_netmask; /* Our subnet mask (0 = unknown) */ /* Our Domain Name Server (0 = unknown) */ extern struct in_addr net_dns_server; #if defined(CONFIG_BOOTP_DNS2) /* Our 2nd Domain Name Server (0 = unknown) */ extern struct in_addr net_dns_server2; #endif extern char net_nis_domain[32]; /* Our IS domain */ extern char net_hostname[32]; /* Our hostname */ #ifdef CONFIG_NET extern char net_root_path[CONFIG_BOOTP_MAX_ROOT_PATH_LEN]; /* Our root path */ #endif /* Indicates whether the pxe path prefix / config file was specified in dhcp option */ extern char *pxelinux_configfile; /** END OF BOOTP EXTENTIONS **/ extern u8 net_ethaddr[ARP_HLEN]; /* Our ethernet address */ extern u8 net_server_ethaddr[ARP_HLEN]; /* Boot server enet address */ extern struct in_addr net_ip; /* Our IP addr (0 = unknown) */ extern struct in_addr net_server_ip; /* Server IP addr (0 = unknown) */ extern uchar *net_tx_packet; /* THE transmit packet */ extern uchar *net_rx_packets[PKTBUFSRX]; /* Receive packets */ extern uchar *net_rx_packet; /* Current receive packet */ extern int net_rx_packet_len; /* Current rx packet length */ extern const u8 net_bcast_ethaddr[ARP_HLEN]; /* Ethernet broadcast address */ extern const u8 net_null_ethaddr[ARP_HLEN]; #define VLAN_NONE 4095 /* untagged */ #define VLAN_IDMASK 0x0fff /* mask of valid vlan id */ extern ushort net_our_vlan; /* Our VLAN */ extern ushort net_native_vlan; /* Our Native VLAN */ extern int net_restart_wrap; /* Tried all network devices */ enum proto_t { BOOTP, RARP, ARP, TFTPGET, DHCP, DHCP6, PING, PING6, DNS, NFS, CDP, NETCONS, SNTP, TFTPSRV, TFTPPUT, LINKLOCAL, FASTBOOT, WOL, UDP, NCSI, WGET }; extern char net_boot_file_name[1024];/* Boot File name */ /* Indicates whether the file name was specified on the command line */ extern bool net_boot_file_name_explicit; /* The actual transferred size of the bootfile (in bytes) */ extern u32 net_boot_file_size; /* Boot file size in blocks as reported by the DHCP server */ extern u32 net_boot_file_expected_size_in_blocks; #if defined(CONFIG_CMD_DNS) extern char *net_dns_resolve; /* The host to resolve */ extern char *net_dns_env_var; /* the env var to put the ip into */ #endif #if defined(CONFIG_CMD_PING) extern struct in_addr net_ping_ip; /* the ip address to ping */ #endif #if defined(CONFIG_CMD_CDP) /* when CDP completes these hold the return values */ extern ushort cdp_native_vlan; /* CDP returned native VLAN */ extern ushort cdp_appliance_vlan; /* CDP returned appliance VLAN */ /* * Check for a CDP packet by examining the received MAC address field */ static inline int is_cdp_packet(const uchar *ethaddr) { extern const u8 net_cdp_ethaddr[ARP_HLEN]; return memcmp(ethaddr, net_cdp_ethaddr, ARP_HLEN) == 0; } #endif #if defined(CONFIG_CMD_SNTP) extern struct in_addr net_ntp_server; /* the ip address to NTP */ extern int net_ntp_time_offset; /* offset time from UTC */ #endif /* Initialize the network adapter */ int net_init(void); int net_loop(enum proto_t); /* Load failed. Start again. */ int net_start_again(void); /* Get size of the ethernet header when we send */ int net_eth_hdr_size(void); /* Set ethernet header; returns the size of the header */ int net_set_ether(uchar *xet, const uchar *dest_ethaddr, uint prot); int net_update_ether(struct ethernet_hdr *et, uchar *addr, uint prot); /* Set IP header */ void net_set_ip_header(uchar *pkt, struct in_addr dest, struct in_addr source, u16 pkt_len, u8 proto); void net_set_udp_header(uchar *pkt, struct in_addr dest, int dport, int sport, int len); /** * compute_ip_checksum() - Compute IP checksum * * @addr: Address to check (must be 16-bit aligned) * @nbytes: Number of bytes to check (normally a multiple of 2) * Return: 16-bit IP checksum */ unsigned compute_ip_checksum(const void *addr, unsigned nbytes); /** * add_ip_checksums() - add two IP checksums * * @offset: Offset of first sum (if odd we do a byte-swap) * @sum: First checksum * @new_sum: New checksum to add * Return: updated 16-bit IP checksum */ unsigned add_ip_checksums(unsigned offset, unsigned sum, unsigned new_sum); /** * ip_checksum_ok() - check if a checksum is correct * * This works by making sure the checksum sums to 0 * * @addr: Address to check (must be 16-bit aligned) * @nbytes: Number of bytes to check (normally a multiple of 2) * Return: true if the checksum matches, false if not */ int ip_checksum_ok(const void *addr, unsigned nbytes); /* Callbacks */ rxhand_f *net_get_udp_handler(void); /* Get UDP RX packet handler */ void net_set_udp_handler(rxhand_f *); /* Set UDP RX packet handler */ rxhand_f *net_get_arp_handler(void); /* Get ARP RX packet handler */ void net_set_arp_handler(rxhand_f *); /* Set ARP RX packet handler */ bool arp_is_waiting(void); /* Waiting for ARP reply? */ void net_set_icmp_handler(rxhand_icmp_f *f); /* Set ICMP RX handler */ void net_set_timeout_handler(ulong, thand_f *);/* Set timeout handler */ /* Network loop state */ enum net_loop_state { NETLOOP_CONTINUE, NETLOOP_RESTART, NETLOOP_SUCCESS, NETLOOP_FAIL }; extern enum net_loop_state net_state; static inline void net_set_state(enum net_loop_state state) { debug_cond(DEBUG_INT_STATE, "--- NetState set to %d\n", state); net_state = state; } /* * net_get_async_tx_pkt_buf - Get a packet buffer that is not in use for * sending an asynchronous reply * * returns - ptr to packet buffer */ uchar * net_get_async_tx_pkt_buf(void); /* Transmit a packet */ static inline void net_send_packet(uchar *pkt, int len) { /* Currently no way to return errors from eth_send() */ (void) eth_send(pkt, len); } /** * net_send_ip_packet() - Transmit "net_tx_packet" as UDP or TCP packet, * send ARP request if needed (ether will be populated) * @ether: Raw packet buffer * @dest: IP address to send the datagram to * @dport: Destination UDP port * @sport: Source UDP port * @payload_len: Length of data after the UDP header * @action: TCP action to be performed * @tcp_seq_num: TCP sequence number of this transmission * @tcp_ack_num: TCP stream acknolegement number * * Return: 0 on success, other value on failure */ int net_send_ip_packet(uchar *ether, struct in_addr dest, int dport, int sport, int payload_len, int proto, u8 action, u32 tcp_seq_num, u32 tcp_ack_num); /** * net_send_tcp_packet() - Transmit TCP packet. * @payload_len: length of payload * @dport: Destination TCP port * @sport: Source TCP port * @action: TCP action to be performed * @tcp_seq_num: TCP sequence number of this transmission * @tcp_ack_num: TCP stream acknolegement number * * Return: 0 on success, other value on failure */ int net_send_tcp_packet(int payload_len, int dport, int sport, u8 action, u32 tcp_seq_num, u32 tcp_ack_num); int net_send_udp_packet(uchar *ether, struct in_addr dest, int dport, int sport, int payload_len); /* Processes a received packet */ void net_process_received_packet(uchar *in_packet, int len); #if defined(CONFIG_NETCONSOLE) && !defined(CONFIG_SPL_BUILD) void nc_start(void); int nc_input_packet(uchar *pkt, struct in_addr src_ip, unsigned dest_port, unsigned src_port, unsigned len); #endif static __always_inline int eth_is_on_demand_init(void) { #if defined(CONFIG_NETCONSOLE) && !defined(CONFIG_SPL_BUILD) extern enum proto_t net_loop_last_protocol; return net_loop_last_protocol != NETCONS; #else return 1; #endif } static inline void eth_set_last_protocol(int protocol) { #if defined(CONFIG_NETCONSOLE) && !defined(CONFIG_SPL_BUILD) extern enum proto_t net_loop_last_protocol; net_loop_last_protocol = protocol; #endif } /* * Check if autoload is enabled. If so, use either NFS or TFTP to download * the boot file. */ void net_auto_load(void); /* * The following functions are a bit ugly, but necessary to deal with * alignment restrictions on ARM. * * We're using inline functions, which had the smallest memory * footprint in our tests. */ /* return IP *in network byteorder* */ static inline struct in_addr net_read_ip(void *from) { struct in_addr ip; memcpy((void *)&ip, (void *)from, sizeof(ip)); return ip; } /* return ulong *in network byteorder* */ static inline u32 net_read_u32(void *from) { u32 l; memcpy((void *)&l, (void *)from, sizeof(l)); return l; } /* write IP *in network byteorder* */ static inline void net_write_ip(void *to, struct in_addr ip) { memcpy(to, (void *)&ip, sizeof(ip)); } /* copy IP */ static inline void net_copy_ip(void *to, void *from) { memcpy((void *)to, from, sizeof(struct in_addr)); } /* copy ulong */ static inline void net_copy_u32(void *to, void *from) { memcpy((void *)to, (void *)from, sizeof(u32)); } /** * is_zero_ethaddr - Determine if give Ethernet address is all zeros. * @addr: Pointer to a six-byte array containing the Ethernet address * * Return true if the address is all zeroes. */ static inline int is_zero_ethaddr(const u8 *addr) { return !(addr[0] | addr[1] | addr[2] | addr[3] | addr[4] | addr[5]); } /** * is_multicast_ethaddr - Determine if the Ethernet address is a multicast. * @addr: Pointer to a six-byte array containing the Ethernet address * * Return true if the address is a multicast address. * By definition the broadcast address is also a multicast address. */ static inline int is_multicast_ethaddr(const u8 *addr) { return 0x01 & addr[0]; } /* * is_broadcast_ethaddr - Determine if the Ethernet address is broadcast * @addr: Pointer to a six-byte array containing the Ethernet address * * Return true if the address is the broadcast address. */ static inline int is_broadcast_ethaddr(const u8 *addr) { return (addr[0] & addr[1] & addr[2] & addr[3] & addr[4] & addr[5]) == 0xff; } /* * is_valid_ethaddr - Determine if the given Ethernet address is valid * @addr: Pointer to a six-byte array containing the Ethernet address * * Check that the Ethernet address (MAC) is not 00:00:00:00:00:00, is not * a multicast address, and is not FF:FF:FF:FF:FF:FF. * * Return true if the address is valid. */ static inline int is_valid_ethaddr(const u8 *addr) { /* FF:FF:FF:FF:FF:FF is a multicast address so we don't need to * explicitly check for it here. */ return !is_multicast_ethaddr(addr) && !is_zero_ethaddr(addr); } /** * net_random_ethaddr - Generate software assigned random Ethernet address * @addr: Pointer to a six-byte array containing the Ethernet address * * Generate a random Ethernet address (MAC) that is not multicast * and has the local assigned bit set. */ static inline void net_random_ethaddr(uchar *addr) { int i; unsigned int seed = get_ticks(); for (i = 0; i < 6; i++) addr[i] = rand_r(&seed); addr[0] &= 0xfe; /* clear multicast bit */ addr[0] |= 0x02; /* set local assignment bit (IEEE802) */ } /** * string_to_enetaddr() - Parse a MAC address * * Convert a string MAC address * * Implemented in lib/net_utils.c (built unconditionally) * * @addr: MAC address in aa:bb:cc:dd:ee:ff format, where each part is a 2-digit * hex value * @enetaddr: Place to put MAC address (6 bytes) */ void string_to_enetaddr(const char *addr, uint8_t *enetaddr); /* Convert an IP address to a string */ void ip_to_string(struct in_addr x, char *s); /** * string_to_ip() - Convert a string to ip address * * Implemented in lib/net_utils.c (built unconditionally) * * @s: Input string to parse * @return: in_addr struct containing the parsed IP address */ struct in_addr string_to_ip(const char *s); /* Convert a VLAN id to a string */ void vlan_to_string(ushort x, char *s); /* Convert a string to a vlan id */ ushort string_to_vlan(const char *s); /* read a VLAN id from an environment variable */ ushort env_get_vlan(char *); /* copy a filename (allow for "..." notation, limit length) */ void copy_filename(char *dst, const char *src, int size); /* check if serverip is specified in filename from the command line */ int is_serverip_in_cmd(void); /** * net_parse_bootfile - Parse the bootfile env var / cmd line param * * @param ipaddr - a pointer to the ipaddr to populate if included in bootfile * @param filename - a pointer to the string to save the filename part * @param max_len - The longest - 1 that the filename part can be * * return 1 if parsed, 0 if bootfile is empty */ int net_parse_bootfile(struct in_addr *ipaddr, char *filename, int max_len); /** * update_tftp - Update firmware over TFTP (via DFU) * * This function updates board's firmware via TFTP * * @param addr - memory address where data is stored * @param interface - the DFU medium name - e.g. "mmc" * @param devstring - the DFU medium number - e.g. "1" * * Return: - 0 on success, other value on failure */ int update_tftp(ulong addr, char *interface, char *devstring); /** * env_get_ip() - Convert an environment value to to an ip address * * @var: Environment variable to convert. The value of this variable must be * in the format format a.b.c.d, where each value is a decimal number from * 0 to 255 * Return: IP address, or 0 if invalid */ static inline struct in_addr env_get_ip(char *var) { return string_to_ip(env_get(var)); } /** * reset_phy() - Reset the Ethernet PHY * * This should be implemented by boards if CONFIG_RESET_PHY_R is enabled */ void reset_phy(void); #if CONFIG_IS_ENABLED(NET) /** * eth_set_enable_bootdevs() - Enable or disable binding of Ethernet bootdevs * * These get in the way of bootstd testing, so are normally disabled by tests. * This provide control of this setting. It only affects binding of Ethernet * devices, so if that has already happened, this flag does nothing. * * @enable: true to enable binding of bootdevs when binding new Ethernet * devices, false to disable it */ void eth_set_enable_bootdevs(bool enable); #else static inline void eth_set_enable_bootdevs(bool enable) {} #endif #endif /* __NET_H__ */