The Linux Kernel/Networking

networkingsockets accessaddress families: inet, unix, ...network storageprotocolsnetwork interfacesnetwork drivers Linux kernel network functionality spans from sockets interface through protocols to network cards. ⚲ Shell interfaces: 8 netstat prints network connections, routing tables, interface statistics and other details 8 ip shows and configures routing, network devices, interfaces and tunnels 8 ss - socket statistics utility Sockets ⚲ API: sys/socket.h – main user mode sockets header Basic common and client side interface: 2 socket ↪ __sys_socket creates an endpoint for communication struct sockaddr - abstract socket address 2 connect ↪ __sys_connect; 2 shutdown shuts down part of a full-duplex connection 2 send ↪ __sys_sendto sends a message on a socket 2 recv ↪ __sys_recvfrom, __sys_recvmsg receives a message from a socket 2 setsockopt, 2 getsockopt ↪ __sys_setsockopt – set/get socket options 2 socketpair ↪ __sys_socketpair – create a pair of connected sockets Additional server side interface: 2 bind ↪ __sys_bind - binds a sockaddr to a socket 2 listen ↪ __sys_listen - listens for connections on a socket 2 accept ↪ __sys_accept4 - accepts a connection on a socket ⚙️ Internals struct socket @ linux/net.h contains struct proto_ops - abstract protocols interface struct sock - network layer representation of sockets net/sock.h __sys_socket ↯ call hierarchy: sock_create __sock_create security_socket_create sock_alloc net_proto_family->create. for example inet_create. See Address families for other options. __sys_connect ↯ call hierarchy: move_addr_to_kernel audit_sockaddr __sys_connect_file sock_from_file security_socket_connect proto_ops->connect. for example inet_stream_connect. See Protocols for other options. net/socket.c – socket system calls implementation net/core/sock.c – sock core: allocation, options, buffer management 📚 References 7 socket linux/socket.h Berkeley sockets 🔨 stress-ng --sock, --sockpair Network storage 🚀 advanced topic 🔧 TODO ⚲ API: 2 sendfile ↪ do_sendfile. See also Zero-copy between file descriptors Application layer Application layer: Network File System NFS filesystems/nfs, linux/nfs_fs.h, fs/nfs init_nfs_fs, nfs4_fs_type, nfs_fs_type, init_nfsd, nfsd_fs_type CIFS admin-guide/cifs init_cifs cifs_fs_type, smb3_fs_type cifs_smb3_do_mount target and iSCSI Interfaces Guide driver-api/target.html 📚 Further reading https://deepwiki.com/torvalds/linux/4.3-network-filesystems-(cifssmb) Transport and Network Names ⚲ API: 2 uname, 2 sethostname, 2 gethostname, 2 setdomainname 2 getdomainname ↪ utsname ⚙️ Details utsname returns writable pointer to new_utsname from uts_namespace from nsproxy from current task_struct. CLONE_NEWUTS, setns kernel/utsname.c – UTS namespace management 📚 References: 7 namespaces 7 network_namespaces 7 uts_namespaces Address families ⚲ API 2 getsockname 2 getpeername Address Family (AF) domain defines address format and address length socklen_t. 3 inet_ntop, 3 inet_pton (derive socklen_t from AF) Common AF: AF_UNIX, AF_INET, AF_NETLINK. PF - Protocol Family index (PF_MAX) actually is the same as Address Family index (AF). ⚙️ Internals of some AF 7 unix ↪ unix_family_ops - sockets for local IPC, net/unix unix_create 7 ip ↪ inet_family_ops - IPv4 inet_create 7 netlink ↪ netlink_family_ops - communication between kernel and user space, net/netlink linux/netlink.h, linux/rtnetlink.h – netlink and routing netlink messages netlink_create 🔨 stress-ng --netlink-proc 7 vsock ↪ vsock_family_ops - communication between VM and hypervisor vsock_create 7 packet ↪ packet_family_ops - device level interface packet_create bt_sock_family_ops - Bluetooth bt_sock_create There are more than 40 AFs in total (see AF_MAX) ⚙️ Internals sock_register - registers net_proto_family. See references to this identifiers to find more than 30 protocol families. __sock_create 📚 Further reading 8 ip-address – protocol address management Internet layer 7 address_families Protocols Each Protocol Family (PF, same index as Address Family AF) consists of several protocol implementations. Directory /proc/net contains various files and subdirectories containing information about the networking layer. File /proc/net/protocols lists available and used protocols. In each PF protocols are classified to different types sock_type, for example stream, datagram and raw socket. TCP is type of stream, UDP is type of datagram, raw and ping are type of raw. proto_register - registers struct proto - protocol implementations: In inet_init initcall, inetsw_array, proto_ops and proto : inet_stream_ops & tcp_prot tcp_sendmsg ... inet_dgram_ops & udp_prot udp_sendmsg ... inet_sockraw_ops raw_prot raw_sendmsg ... ping_prot ping_v4_sendmsg ... In af_unix_init initcall: unix_family_ops unix_create unix_stream_ops unix_stream_sendmsg ... unix_dgram_ops unix_dgram_sendmsg ... unix_seqpacket_ops unix_seqpacket_sendmsg ... ⚙️ Internals: linux/in.h, linux/ip.h – IPv4 address, protocol and header definitions net/ipv4 – IPv4 protocol stack net/ipv4/af_inet.c – PF_INET socket family (inet_family_ops, inet_init) net/ipv4/protocol.c – IPv4 protocol handler registration (inet_add_protocol) net/ipv4/ip_input.c – IPv4 receive path (ip_rcv, ip_local_deliver) net/ipv4/ip_output.c – IPv4 transmit path (ip_output, ip_queue_xmit) net/ipv4/ip_forward.c – IPv4 packet forwarding (ip_forward) net/ipv4/ip_fragment.c – fragmentation and reassembly (ip_do_fragment) net/ipv4/ip_options.c – IP options parsing and building net/ipv4/ip_sockglue.c – IP socket options (ip_setsockopt) Addressing: net/ipv4/devinet.c – per-device IPv4 address management (inetdev_init) net/ipv4/arp.c – Address Resolution Protocol (arp_rcv, arp_send) Routing: net/ipv4/route.c – IPv4 route lookup and cache (ip_route_output_key, rtable) net/ipv4/fib_frontend.c – FIB user interface: netlink and ioctl (fib_table_lookup) net/ipv4/fib_trie.c – LC-trie FIB implementation (fib_trie_table) net/ipv4/fib_semantics.c – FIB entry semantics: nexthop, scope, metrics net/ipv4/fib_rules.c – IPv4 routing policy rules (ip rule) net/ipv4/nexthop.c – generic nexthop objects (since 5.3, nexthop_find_by_id) net/ipv4/inetpeer.c – peer storage for per-destination state (inet_getpeer) ICMP / IGMP: net/ipv4/icmp.c – ICMPv4 (icmp_rcv, icmp_send) net/ipv4/igmp.c – Internet Group Management Protocol, MLDv1 (ip_mc_join_group) net/ipv4/ipmr.c – IPv4 multicast routing (ipmr_mfc_add) Transport protocols: net/tcp.h, net/ipv4/tcp.c – TCP core (connect, send, recv, close) tcp_sock – per-connection TCP state, tcp_sk – cast from sock tcphdr – TCP header struct tcp_congestion_ops – pluggable congestion control (TCP congestion control cubic, bbr, reno) net/ipv4/tcp_ipv4.c – TCP over IPv4 (tcp_v4_rcv, tcp_prot) net/ipv4/tcp_input.c – TCP receive processing (tcp_rcv_established, ACK, SACK) net/ipv4/tcp_output.c – TCP transmit: segment building, retransmit (tcp_write_xmit) net/ipv4/tcp_timer.c – TCP timers: retransmit, keepalive, TIME_WAIT (tcp_retransmit_timer) net/ipv4/tcp_minisocks.c – TIME_WAIT and SYN_RECV mini sockets (tcp_time_wait) net/ipv4/tcp_fastopen.c – TCP Fast Open (tcp_fastopen_cookie, since 3.6) net/ipv4/tcp_ao.c – TCP Authentication Option (RFC 5925, since 6.7) net/ipv4/tcp_metrics.c – per-destination TCP metrics cache net/ipv4/tcp_cong.c – congestion control framework (tcp_register_congestion_control) net/ipv4/tcp_cubic.c – CUBIC congestion control (default) net/ipv4/tcp_bbr.c – BBR congestion control (since 4.9) net/ipv4/syncookies.c – TCP SYN cookies (cookie_v4_check) net/ipv4/udp.c – UDP (udp_rcv, udp_prot) net/ipv4/raw.c – raw IP sockets (raw_prot) net/ipv4/ping.c – ping socket / ICMP echo (ping_prot) net/ipv4/datagram.c – common UDP/RAW datagram helpers (ip4_datagram_connect) Socket infrastructure: net/ipv4/inet_connection_sock.c – connection-oriented socket base (inet_csk_accept) net/ipv4/inet_hashtables.c – TCP/UDP socket hash table lookup (inet_lookup_established) net/ipv4/inet_timewait_sock.c – TIME_WAIT socket recycling net/ipv4/inet_fragment.c – generic IP fragment queue management net/ipv4/inet_diag.c – socket diagnostics via netlink (ss, inet_diag_handler) Tunnels: net/ipv4/ip_tunnel.c – generic IP tunnel framework (ip_tunnel_xmit) net/ipv4/ip_gre.c – GRE over IPv4 (gre, gretap) net/ipv4/ipip.c – IP-in-IP tunnel (ipip) net/ipv4/ip_vti.c – virtual tunnel interface (vti, IPsec-aware) net/ipv4/fou_core.c – Foo-over-UDP encapsulation (since 3.18) net/ipv4/gre_demux.c – GRE protocol demultiplexer net/ipv4/tunnel4.c – IPv4 tunnel protocol registration IPsec / XFRM: net/ipv4/xfrm4_policy.c – IPv4 IPsec policy (xfrm4_dst_ops) net/ipv4/xfrm4_state.c – IPv4 IPsec state management net/ipv4/xfrm4_input.c – IPv4 IPsec inbound processing net/ipv4/xfrm4_output.c – IPv4 IPsec outbound processing net/ipv4/ah4.c – Authentication Header (AH) for IPv4 net/ipv4/esp4.c – Encapsulating Security Payload (ESP) for IPv4 net/ipv4/ipcomp.c – IP Payload Compression (IPComp, RFC 3173) Security: net/ipv4/cipso_ipv4.c – CIPSO Commercial IP Security Option (cipso_v4_validate) sysctl / proc: net/ipv4/sysctl_net_ipv4.c – /proc/sys/net/ipv4/ tunables net/ipv4/proc.c – /proc/net/snmp and /proc/net/netstat statistics net/ipv4/netfilter – IPv4 netfilter hooks (see Netfilter) linux/ipv6.h – IPv6 definitions net/ipv6 – IPv6 protocol stack net/ipv6/af_inet6.c – PF_INET6 socket family (inet6_family_ops, inet6_init) net/ipv6/protocol.c – IPv6 protocol handler registration (inet6_add_protocol) net/ipv6/ip6_input.c – IPv6 receive path (ipv6_rcv, ip6_forward) net/ipv6/ip6_output.c – IPv6 transmit path (ip6_output, ip6_xmit) net/ipv6/ip6_checksum.c – UDP/TCP checksum over IPv6 net/ipv6/ila – IPv6 address Identifier Locator Addressing (ILA): address translation for scalable IPv6 overlays net/ipv6/netfilter – IPv6 netfilter hooks (see Netfilter) Addressing: net/ipv6/addrconf.c – address autoconfiguration: SLAAC, DAD, RA handling (addrconf_notify) net/ipv6/addrconf_core.c – address scope helpers for static builds (no full IPv6 module) net/ipv6/ndisc.c – Neighbor Discovery Protocol (Neighbor Discovery Protocol NDP, RFC 4861) net/ipv6/anycast.c – anycast address support net/ipv6/addrlabel.c – address label subsystem for source address selection (RFC 6724) Routing: net/ipv6/route.c – IPv6 routing table lookup and route management (ip6_route_output, rt6_info) net/ipv6/ip6_fib.c – IPv6 FIB trie (fib6_table, fib6_info) net/ipv6/fib6_rules.c – IPv6 routing policy rules (ip rule) net/ipv6/ip6_flowlabel.c – flow label manager (ip6_flowlabel) Extension headers: net/ipv6/exthdrs.c – extension header processing (hop-by-hop, routing, destination) Transport protocols: net/ipv6/tcp_ipv6.c – TCP over IPv6 (tcp_v6_rcv, tcpv6_prot, tcp_v6_syn_recv_sock) net/ipv6/udp.c – UDP over IPv6 (udpv6_prot, udpv6_rcv) net/ipv6/raw.c – raw sockets over IPv6 net/ipv6/ping.c – ping socket (ICMPv6 echo) over IPv6 net/ipv6/datagram.c – common UDP/RAW datagram helpers (ip6_datagram_connect) net/ipv6/syncookies.c – TCP SYN cookie support for IPv6 net/ipv6/reassembly.c – fragment reassembly (ipv6_frag_rcv) ICMPv6: net/ipv6/icmp.c – ICMPv6 (RFC 4443): error messages, echo, icmpv6_send Multicast: net/ipv6/mcast.c – multicast group management, MLDv1/v2 (ipv6_dev_mc_inc) net/ipv6/ip6mr.c – IPv6 multicast routing (ip6mr_table) net/ipv6/mcast_snoop.c – MLD snooping for bridge integration Tunnels: net/ipv6/ip6_tunnel.c – IPv6-in-IPv6 tunnel (ip6tnl) net/ipv6/ip6_gre.c – GRE over IPv6 (ip6gre, ip6gretap) net/ipv6/sit.c – IPv4-in-IPv6 Simple Internet Transition tunnel (sit) net/ipv6/ip6_vti.c – virtual tunnel interface over IPv6 (vti6, IPsec-aware) net/ipv6/ip6_udp_tunnel.c – UDP encapsulation helpers for IPv6 tunnels net/ipv6/fou6.c – Foo-over-UDP encapsulation for IPv6 IPsec / XFRM: net/ipv6/xfrm6_policy.c – IPv6 IPsec policy lookup (xfrm6_dst_ops) net/ipv6/xfrm6_state.c – IPv6 IPsec state management net/ipv6/xfrm6_input.c – IPv6 IPsec inbound processing net/ipv6/xfrm6_output.c – IPv6 IPsec outbound processing net/ipv6/ah6.c – Authentication Header (AH) transform for IPv6 net/ipv6/esp6.c – Encapsulating Security Payload (ESP) transform for IPv6 net/ipv6/ipcomp6.c – IP Payload Compression (IPComp) for IPv6 (RFC 3173) net/ipv6/mip6.c – Mobile IPv6 extension header handling Segment Routing: net/ipv6/seg6.c – Segment Routing IPv6 (SRv6) core: genl API, encap helpers net/ipv6/seg6_local.c – SRv6 local endpoint behaviors (End, End.DX4, End.DT6, …) net/ipv6/seg6_iptunnel.c – SRv6 tunnel encapsulation via lwtunnel net/ipv6/seg6_hmac.c – SRv6 HMAC integrity verification (RFC 8754) net/ipv6/rpl.c – RPL source routing header IOAM: net/ipv6/ioam6.c – In-situ OAM for IPv6 (RFC 9197): namespace and schema management net/ipv6/ioam6_iptunnel.c – IOAM insertion via lwtunnel encapsulation Security: net/ipv6/calipso.c – CALIPSO Mandatory Access Control option (RFC 5570) sysctl / proc: net/ipv6/sysctl_net_ipv6.c – /proc/sys/net/ipv6/ tunables net/ipv6/proc.c – /proc/net/snmp6 and /proc/net/dev_snmp6 statistics net/tls – kernel TLS (since 4.13) net/xdp – Express Data Path fast packet processing (since 4.8) net/sched – traffic control and QoS A Qdisc (queuing discipline) sits on each netdev_queue and controls packet scheduling on the TX path. struct Qdisc_ops – enqueue / dequeue / peek callbacks, registered via register_qdisc net/sched/sch_generic.c – default pfifo_fast qdisc, qdisc_run, dev_watchdog net/sched/sch_api.c – tc netlink API: qdisc/class/filter CRUD Common qdiscs: net/sched/sch_fifo.c – pfifo / bfifo - simple FIFO, default fallback net/sched/sch_fq.c – FQ - per-flow fair queue with pacing (fq_sched_data) net/sched/sch_fq_codel.c – FQ-CoDel - fair queue + CoDel AQM (default on many distros) net/sched/sch_htb.c – HTB - Hierarchical Token Bucket, bandwidth allocation net/sched/sch_cake.c – CAKE - combined AQM + shaper + flow isolation (since 4.19) net/sched/sch_codel.c – CoDel - Controlled Delay AQM net/sched/sch_tbf.c – TBF - Token Bucket Filter, rate limiter net/sch_generic.h – Qdisc, Qdisc_ops, tcf_proto definitions Socket performance options: TCP_NODELAY – disable Nagle's algorithm for low-latency sends TCP_CORK – accumulate small writes into full-size segments Network determinism tips, RHEL RT Reducing TCP performance spikes – TCP timestamp tuning 📚 References: 7 tcp 7 udp 7 raw 8 tc – traffic control Transport layer and Transmission_Control_Protocol TCP 🔨 stress-ng --udp, --udp-flood, --sctp RDMA 🚀 advanced topic 🗝️ Acronyms: IB — InfiniBand, an interconnect standard, competes with Ethernet, Fibre Channel IPoIB — IP network emulation layer over InfiniBand networks SRP — SCSI RDMA Protocol ULP — Upper-layer protocols iSER — iSCSI Extensions for RDMA ⚲ Interfaces: https://github.com/linux-rdma/rdma-core 8 rdma 7 rdma_cm — RDMA communication manager include/uapi/rdma – user-space RDMA API definitions include/rdma – kernel RDMA API definitions ⚙️ Internals: drivers/infiniband – InfiniBand core and drivers drivers/infiniband/ulp — Upper-layer protocols drivers/infiniband/sw — software drivers drivers/infiniband/hw — hardware device drivers 📚 References: InfiniBand infiniband InfiniBand and RDMA Interfaces driver-api/infiniband.html Netfilter Netfilter is the kernel’s packet filtering and manipulation framework. It provides hook points in the network stack where modules can inspect, modify, drop, or redirect packets. nftables is the current user-facing interface, replacing the legacy iptables/ip6tables/ebtables/arptables tools. Netfilter#Connection_tracking connection tracking (conntrack) tracks network connections for stateful filtering and NAT. ⚲ API 8 nft – nftables rule management 8 iptables, 8 ip6tables – legacy packet filter 8 ebtables, 8 arptables – bridge and ARP filtering 8 ebtables-nft, 8 arptables-nft, 8 xtables-nft – nftables-backed compatibility ipset – IP set management 8 conntrack – conntrack table management /proc/net/nf_conntrack – active connections /proc/sys/net/netfilter/ – conntrack tunables linux/netfilter.h – hook API nf_hook_state, nf_register_net_hook uapi/linux/netfilter, linux/netfilter net/netfilter, net/netns/netfilter.h linux/netfilter/x_tables.h – xtables match/target registration ⚙️ Internals net/netfilter – packet filtering framework net/netfilter/core.c – hook registration and dispatch net/netfilter/nf_tables_api.c – nftables kernel API net/netfilter/nf_tables_core.c – nftables rule evaluation net/netfilter/nf_conntrack_core.c – connection tracking core net/netfilter/nf_nat_core.c – Network address translation NAT core net/netfilter/nf_flow_table_core.c – hardware offload flowtable 📖 References Netfilter Sysfs variables networking/netfilter-sysctl.html Netfilter Conntrack Sysfs variables networking/nf_conntrack-sysctl.html Netfilter’s flowtable infrastructure networking/nf_flowtable.html nftables, https://wiki.nftables.org/ https://lwn.net/Kernel/Index/#Networking-Packet_filtering Network device Network interface interfaces ⚲ Interfaces ip -brief link show ls -l /sys/class/net devm_register_netdev registers net_device, net_device_ops dev_queue_xmit queues socket buffers into transmit queue linux/netdevice.h linux/skbuff.h sk_buff socket buffer (skb) /sys/class/net/*/statistics/ – per-interface packet and byte counters 8 devlink – manage switchdev, eswitch, shared buffers and firmware of network devices 8 bridge – show/manipulate bridge forwarding, MAC and VLAN tables sos networking plugin – diagnostics collection for network subsystem 👁 Example: drivers/net/loopback.c - the most famous and simple interface lo ⚙️ Internals net/core – core networking infrastructure Device: net/core/dev.c – network device operations core: xmit, receive, register/unregister (__dev_queue_xmit, netif_receive_skb) ⚙️ Internals: softnet_data – per-CPU: poll_list, input_pkt_queue, completion queue net_dev_init – registers NET_RX_SOFTIRQ / NET_TX_SOFTIRQ handlers net_tx_action – TX softirq handler: frees completed skbs from softnet_data→completion_queue drains softnet_data→output_queue of pending qdiscs (qdisc_run → driver ndo_start_xmit) net_rx_action – RX softirq handler; drains up to netdev_budget (300) packets napi_poll → driver napi->poll() netif_receive_skb → __netif_receive_skb_core → ip_rcv / ipv6_rcv … Legacy (pre-NAPI): netif_rx → input_pkt_queue → backlog napi (same drain) net/neighbour.h, net/core/neighbour.c – generic neighbour (ARP/NDP) cache (neigh_lookup, neighbour) net/core/rtnetlink.c – rtnetlink: netlink interface for routing and device config (__rtnl_register_many) net/core/net-sysfs.c – /sys/class/net/ device attributes net/core/sysctl_net_core.c – /proc/sys/net/core/ tunables Sockets: net/sock.h, net/core/sock.c – sock core: allocation, options, buffer management linux/skbuff.h, net/core/skbuff.c – sk_buff allocation, cloning, manipulation (alloc_skb, skb_clone) net/core/datagram.c – generic datagram recv helpers (skb_recv_datagram) net/core/stream.c – generic stream socket helpers (sk_stream_wait_connect) net/core/scm.c – socket control message (ancillary data) handling net/sock_reuseport.h, net/core/sock_reuseport.c – SO_REUSEPORT socket selection (reuseport_add_sock) linux/filter.h, net/core/filter.c – socket filtering: BPF/cBPF/eBPF attach to sockets (sk_filter_trim_cap) Routing / forwarding: net/dst.h, net/core/dst.c – destination cache (dst_entry, dst_alloc) net/flow_dissector.h, net/core/flow_dissector.c – packet field extraction for hashing and steering (skb_flow_dissector_init, skb_flow_get_ports) net/lwtunnel.h, net/core/lwtunnel.c – lightweight tunnel infrastructure (lwtunnel_build_state) net/net_namespace.h, net/core/net_namespace.c – network namespace lifecycle (copy_net_ns, net_ns_ops) Offload: net/core/gro.c – Generic Receive Offload (gro_receive_skb, dev_gro_receive) net/core/gso.c – Generic Segmentation Offload (__skb_gso_segment, skb_mac_gso_segment) net/xdp.h, net/core/xdp.c – XDP buffer and memory model (xdp_rxq_info, xdp_buff) net/page_pool/types.h, net/core/page_pool.c – page pool allocator for zero-copy RX (page_pool_create) Diagnostics / tools: net/core/secure_seq.c – ISN and port randomization (secure_tcp_seq_and_ts_off, secure_ipv4_port_ephemeral) net/core/netpoll.c – network console / crash dump transport (netpoll_send_skb) net/core/drop_monitor.c – kfree_skb drop event monitor (send_dm_alert, net_dm_alert_summary_ops) net/core/pktgen.c – in-kernel packet generator (8 pktgen) net/core/timestamping.c – SO_TIMESTAMPING and PHY timestamping helpers function loopback_xmit receives skb and passes it back with netif_rx 🔨 stress-ng --rawsock, --icmp-flood 📚 Further reading 8 ip-link – network device configuration 8 ip-stats – manage and show interface statistics 7 netdevice – low-level access to Linux network devices 7 packet – packet interface on device level Queueing in the Linux Network Stack 💾 Historical LDP TLK Chapter 10 Networks Network drivers linux/if_ether.h – IEEE 802.3 Ethernet constants and header linux/etherdevice.h New_API NAPI NAPI Driver design ⚲ API: netif_napi_add – registers napi_struct, stores driver poll_fn as napi->poll napi_schedule – raises NET_RX_SOFTIRQ from IRQ handler netif_receive_skb – delivers skb to protocol handlers napi_complete_done – signals poll() exhausted its quota ⚙️ Internals: net_dev_init – registers softirq handlers, sets up per-CPU softnet_data napi_poll – calls driver napi->poll(), enforces budget 👁 example e1000_intr → __napi_schedule e1000e_poll (napi->poll()) → napi_complete_done ether_setup setups Ethernet network device 👁 An example of Ethernet driver: e1000_probe ⚙️ Internals: drivers/net – network device drivers drivers/net/wireless – wireless drivers drivers/net/ethernet – Ethernet drivers net/wireless – cfg80211, wireless configuration API and userspace interface via 8 iw net/mac80211 – mac80211, software 802.11 MAC layer used by most WiFi drivers net/bluetooth – Bluetooth protocol stack net/bridge – Ethernet bridging networking/bridge.html 📚 References: 8 ethtool – query or control network driver and hardware settings Data link layer: Ethernet GRO - Generic Receive Offload Segmentation Offloads networking/segmentation-offloads.html https://wireless.wiki.kernel.org 📚 Further reading LKMPG: Network Drivers 💾 Historical: LDD2:Network Drivers LDD3:Network Drivers Kernel Analysis: Networking, 2003 network_overview 📖 Further reading about networking Networking interfaces subsystem-apis.html#networking-interfaces Networking networking https://lwn.net/Kernel/Index/#Networking https://lartc.org/ – Linux Advanced Routing & Traffic Control 8 ip – show / manipulate routing, network devices, interfaces and tunnels 8 tc – show / manipulate traffic control settings bcc/ebpf networking tools eBPF-based Networking, Security, and Observability Retis – tracing packets in the Linux networking stack & friends