^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) /* Copyright (c) 2017 Facebook
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) * This program is free software; you can redistribute it and/or
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) * modify it under the terms of version 2 of the GNU General Public
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) * License as published by the Free Software Foundation.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) * Sample BPF program to set send and receive buffers to 150KB, sndcwnd clamp
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) * to 100 packets and SYN and SYN_ACK RTOs to 10ms when both hosts are within
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9) * the same datacenter. For his example, we assume they are within the same
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) * datacenter when the first 5.5 bytes of their IPv6 addresses are the same.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) *
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12) * Use "bpftool cgroup attach $cg sock_ops $prog" to load this BPF program.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15) #include <uapi/linux/bpf.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) #include <uapi/linux/if_ether.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17) #include <uapi/linux/if_packet.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) #include <uapi/linux/ip.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) #include <linux/socket.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) #include <bpf/bpf_helpers.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) #include <bpf/bpf_endian.h>
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23) #define DEBUG 1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) SEC("sockops")
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) int bpf_clamp(struct bpf_sock_ops *skops)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28) int bufsize = 150000;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) int to_init = 10;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) int clamp = 100;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) int rv = 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) int op;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) /* For testing purposes, only execute rest of BPF program
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) * if neither port numberis 55601
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) if (bpf_ntohl(skops->remote_port) != 55601 && skops->local_port != 55601) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) skops->reply = -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39) return 0;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) op = (int) skops->op;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) #ifdef DEBUG
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) bpf_printk("BPF command: %d\n", op);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) /* Check that both hosts are within same datacenter. For this example
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) * it is the case when the first 5.5 bytes of their IPv6 addresses are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50) * the same.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) if (skops->family == AF_INET6 &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) skops->local_ip6[0] == skops->remote_ip6[0] &&
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54) (bpf_ntohl(skops->local_ip6[1]) & 0xfff00000) ==
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) (bpf_ntohl(skops->remote_ip6[1]) & 0xfff00000)) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56) switch (op) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) case BPF_SOCK_OPS_TIMEOUT_INIT:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58) rv = to_init;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) case BPF_SOCK_OPS_TCP_CONNECT_CB:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) /* Set sndbuf and rcvbuf of active connections */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) rv = bpf_setsockopt(skops, SOL_SOCKET, SO_SNDBUF,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) &bufsize, sizeof(bufsize));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) rv += bpf_setsockopt(skops, SOL_SOCKET,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) SO_RCVBUF, &bufsize,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66) sizeof(bufsize));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68) case BPF_SOCK_OPS_ACTIVE_ESTABLISHED_CB:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) rv = bpf_setsockopt(skops, SOL_TCP,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70) TCP_BPF_SNDCWND_CLAMP,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) &clamp, sizeof(clamp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) case BPF_SOCK_OPS_PASSIVE_ESTABLISHED_CB:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) /* Set sndbuf and rcvbuf of passive connections */
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) rv = bpf_setsockopt(skops, SOL_TCP,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) TCP_BPF_SNDCWND_CLAMP,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) &clamp, sizeof(clamp));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) rv += bpf_setsockopt(skops, SOL_SOCKET,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79) SO_SNDBUF, &bufsize,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) sizeof(bufsize));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) rv += bpf_setsockopt(skops, SOL_SOCKET,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) SO_RCVBUF, &bufsize,
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) sizeof(bufsize));
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84) break;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) default:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) rv = -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) } else {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) rv = -1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) #ifdef DEBUG
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) bpf_printk("Returning %d\n", rv);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) #endif
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) skops->reply = rv;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) return 1;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) char _license[] SEC("license") = "GPL";