// SPDX-License-Identifier: GPL-2.0-only /* Copyright(c) 2017 Jesper Dangaard Brouer, Red Hat, Inc. */ static const char *__doc__ = "XDP CPU redirect tool, using BPF_MAP_TYPE_CPUMAP\n" "Usage: xdp_redirect_cpu -d -c 0 ... -c N\n" "Valid specification for CPUMAP BPF program:\n" " --mprog-name/-e pass (use built-in XDP_PASS program)\n" " --mprog-name/-e drop (use built-in XDP_DROP program)\n" " --redirect-device/-r (use built-in DEVMAP redirect program)\n" " Custom CPUMAP BPF program:\n" " --mprog-filename/-f --mprog-name/-e \n" " Optionally, also pass --redirect-map/-m and --redirect-device/-r together\n" " to configure DEVMAP in BPF object \n"; #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "bpf_util.h" #include "xdp_sample_user.h" #include "xdp_redirect_cpu.skel.h" static int map_fd; static int avail_fd; static int count_fd; static int mask = SAMPLE_RX_CNT | SAMPLE_REDIRECT_ERR_MAP_CNT | SAMPLE_CPUMAP_ENQUEUE_CNT | SAMPLE_CPUMAP_KTHREAD_CNT | SAMPLE_EXCEPTION_CNT; DEFINE_SAMPLE_INIT(xdp_redirect_cpu); static const struct option long_options[] = { { "help", no_argument, NULL, 'h' }, { "dev", required_argument, NULL, 'd' }, { "skb-mode", no_argument, NULL, 'S' }, { "progname", required_argument, NULL, 'p' }, { "qsize", required_argument, NULL, 'q' }, { "cpu", required_argument, NULL, 'c' }, { "stress-mode", no_argument, NULL, 'x' }, { "force", no_argument, NULL, 'F' }, { "interval", required_argument, NULL, 'i' }, { "verbose", no_argument, NULL, 'v' }, { "stats", no_argument, NULL, 's' }, { "mprog-name", required_argument, NULL, 'e' }, { "mprog-filename", required_argument, NULL, 'f' }, { "redirect-device", required_argument, NULL, 'r' }, { "redirect-map", required_argument, NULL, 'm' }, {} }; static void print_avail_progs(struct bpf_object *obj) { struct bpf_program *pos; printf(" Programs to be used for -p/--progname:\n"); bpf_object__for_each_program(pos, obj) { if (bpf_program__is_xdp(pos)) { if (!strncmp(bpf_program__name(pos), "xdp_prognum", sizeof("xdp_prognum") - 1)) printf(" %s\n", bpf_program__name(pos)); } } } static void usage(char *argv[], const struct option *long_options, const char *doc, int mask, bool error, struct bpf_object *obj) { sample_usage(argv, long_options, doc, mask, error); print_avail_progs(obj); } static int create_cpu_entry(__u32 cpu, struct bpf_cpumap_val *value, __u32 avail_idx, bool new) { __u32 curr_cpus_count = 0; __u32 key = 0; int ret; /* Add a CPU entry to cpumap, as this allocate a cpu entry in * the kernel for the cpu. */ ret = bpf_map_update_elem(map_fd, &cpu, value, 0); if (ret < 0) { fprintf(stderr, "Create CPU entry failed: %s\n", strerror(errno)); return ret; } /* Inform bpf_prog's that a new CPU is available to select * from via some control maps. */ ret = bpf_map_update_elem(avail_fd, &avail_idx, &cpu, 0); if (ret < 0) { fprintf(stderr, "Add to avail CPUs failed: %s\n", strerror(errno)); return ret; } /* When not replacing/updating existing entry, bump the count */ ret = bpf_map_lookup_elem(count_fd, &key, &curr_cpus_count); if (ret < 0) { fprintf(stderr, "Failed reading curr cpus_count: %s\n", strerror(errno)); return ret; } if (new) { curr_cpus_count++; ret = bpf_map_update_elem(count_fd, &key, &curr_cpus_count, 0); if (ret < 0) { fprintf(stderr, "Failed write curr cpus_count: %s\n", strerror(errno)); return ret; } } printf("%s CPU: %u as idx: %u qsize: %d cpumap_prog_fd: %d (cpus_count: %u)\n", new ? "Add new" : "Replace", cpu, avail_idx, value->qsize, value->bpf_prog.fd, curr_cpus_count); return 0; } /* CPUs are zero-indexed. Thus, add a special sentinel default value * in map cpus_available to mark CPU index'es not configured */ static int mark_cpus_unavailable(void) { int ret, i, n_cpus = libbpf_num_possible_cpus(); __u32 invalid_cpu = n_cpus; for (i = 0; i < n_cpus; i++) { ret = bpf_map_update_elem(avail_fd, &i, &invalid_cpu, 0); if (ret < 0) { fprintf(stderr, "Failed marking CPU unavailable: %s\n", strerror(errno)); return ret; } } return 0; } /* Stress cpumap management code by concurrently changing underlying cpumap */ static void stress_cpumap(void *ctx) { struct bpf_cpumap_val *value = ctx; /* Changing qsize will cause kernel to free and alloc a new * bpf_cpu_map_entry, with an associated/complicated tear-down * procedure. */ value->qsize = 1024; create_cpu_entry(1, value, 0, false); value->qsize = 8; create_cpu_entry(1, value, 0, false); value->qsize = 16000; create_cpu_entry(1, value, 0, false); } static int set_cpumap_prog(struct xdp_redirect_cpu *skel, const char *redir_interface, const char *redir_map, const char *mprog_filename, const char *mprog_name) { if (mprog_filename) { struct bpf_program *prog; struct bpf_object *obj; int ret; if (!mprog_name) { fprintf(stderr, "BPF program not specified for file %s\n", mprog_filename); goto end; } if ((redir_interface && !redir_map) || (!redir_interface && redir_map)) { fprintf(stderr, "--redirect-%s specified but --redirect-%s not specified\n", redir_interface ? "device" : "map", redir_interface ? "map" : "device"); goto end; } /* Custom BPF program */ obj = bpf_object__open_file(mprog_filename, NULL); if (!obj) { ret = -errno; fprintf(stderr, "Failed to bpf_prog_load_xattr: %s\n", strerror(errno)); return ret; } ret = bpf_object__load(obj); if (ret < 0) { ret = -errno; fprintf(stderr, "Failed to bpf_object__load: %s\n", strerror(errno)); return ret; } if (redir_map) { int err, redir_map_fd, ifindex_out, key = 0; redir_map_fd = bpf_object__find_map_fd_by_name(obj, redir_map); if (redir_map_fd < 0) { fprintf(stderr, "Failed to bpf_object__find_map_fd_by_name: %s\n", strerror(errno)); return redir_map_fd; } ifindex_out = if_nametoindex(redir_interface); if (!ifindex_out) ifindex_out = strtoul(redir_interface, NULL, 0); if (!ifindex_out) { fprintf(stderr, "Bad interface name or index\n"); return -EINVAL; } err = bpf_map_update_elem(redir_map_fd, &key, &ifindex_out, 0); if (err < 0) return err; } prog = bpf_object__find_program_by_name(obj, mprog_name); if (!prog) { ret = -errno; fprintf(stderr, "Failed to bpf_object__find_program_by_name: %s\n", strerror(errno)); return ret; } return bpf_program__fd(prog); } else { if (mprog_name) { if (redir_interface || redir_map) { fprintf(stderr, "Need to specify --mprog-filename/-f\n"); goto end; } if (!strcmp(mprog_name, "pass") || !strcmp(mprog_name, "drop")) { /* Use built-in pass/drop programs */ return *mprog_name == 'p' ? bpf_program__fd(skel->progs.xdp_redirect_cpu_pass) : bpf_program__fd(skel->progs.xdp_redirect_cpu_drop); } else { fprintf(stderr, "Unknown name \"%s\" for built-in BPF program\n", mprog_name); goto end; } } else { if (redir_map) { fprintf(stderr, "Need to specify --mprog-filename, --mprog-name and" " --redirect-device with --redirect-map\n"); goto end; } if (redir_interface) { /* Use built-in devmap redirect */ struct bpf_devmap_val val = {}; int ifindex_out, err; __u32 key = 0; if (!redir_interface) return 0; ifindex_out = if_nametoindex(redir_interface); if (!ifindex_out) ifindex_out = strtoul(redir_interface, NULL, 0); if (!ifindex_out) { fprintf(stderr, "Bad interface name or index\n"); return -EINVAL; } if (get_mac_addr(ifindex_out, skel->bss->tx_mac_addr) < 0) { printf("Get interface %d mac failed\n", ifindex_out); return -EINVAL; } val.ifindex = ifindex_out; val.bpf_prog.fd = bpf_program__fd(skel->progs.xdp_redirect_egress_prog); err = bpf_map_update_elem(bpf_map__fd(skel->maps.tx_port), &key, &val, 0); if (err < 0) return -errno; return bpf_program__fd(skel->progs.xdp_redirect_cpu_devmap); } } } /* Disabled */ return 0; end: fprintf(stderr, "Invalid options for CPUMAP BPF program\n"); return -EINVAL; } int main(int argc, char **argv) { const char *redir_interface = NULL, *redir_map = NULL; const char *mprog_filename = NULL, *mprog_name = NULL; struct xdp_redirect_cpu *skel; struct bpf_map_info info = {}; char ifname_buf[IF_NAMESIZE]; struct bpf_cpumap_val value; __u32 infosz = sizeof(info); int ret = EXIT_FAIL_OPTION; unsigned long interval = 2; bool stress_mode = false; struct bpf_program *prog; const char *prog_name; bool generic = false; bool force = false; int added_cpus = 0; bool error = true; int longindex = 0; int add_cpu = -1; int ifindex = -1; int *cpu, i, opt; __u32 qsize; int n_cpus; n_cpus = libbpf_num_possible_cpus(); /* Notice: Choosing the queue size is very important when CPU is * configured with power-saving states. * * If deepest state take 133 usec to wakeup from (133/10^6). When link * speed is 10Gbit/s ((10*10^9/8) in bytes/sec). How many bytes can * arrive with in 133 usec at this speed: (10*10^9/8)*(133/10^6) = * 166250 bytes. With MTU size packets this is 110 packets, and with * minimum Ethernet (MAC-preamble + intergap) 84 bytes is 1979 packets. * * Setting default cpumap queue to 2048 as worst-case (small packet) * should be +64 packet due kthread wakeup call (due to xdp_do_flush) * worst-case is 2043 packets. * * Sysadm can configured system to avoid deep-sleep via: * tuned-adm profile network-latency */ qsize = 2048; skel = xdp_redirect_cpu__open(); if (!skel) { fprintf(stderr, "Failed to xdp_redirect_cpu__open: %s\n", strerror(errno)); ret = EXIT_FAIL_BPF; goto end; } ret = sample_init_pre_load(skel); if (ret < 0) { fprintf(stderr, "Failed to sample_init_pre_load: %s\n", strerror(-ret)); ret = EXIT_FAIL_BPF; goto end_destroy; } if (bpf_map__set_max_entries(skel->maps.cpu_map, n_cpus) < 0) { fprintf(stderr, "Failed to set max entries for cpu_map map: %s", strerror(errno)); ret = EXIT_FAIL_BPF; goto end_destroy; } if (bpf_map__set_max_entries(skel->maps.cpus_available, n_cpus) < 0) { fprintf(stderr, "Failed to set max entries for cpus_available map: %s", strerror(errno)); ret = EXIT_FAIL_BPF; goto end_destroy; } cpu = calloc(n_cpus, sizeof(int)); if (!cpu) { fprintf(stderr, "Failed to allocate cpu array\n"); goto end_destroy; } prog = skel->progs.xdp_prognum5_lb_hash_ip_pairs; while ((opt = getopt_long(argc, argv, "d:si:Sxp:f:e:r:m:c:q:Fvh", long_options, &longindex)) != -1) { switch (opt) { case 'd': if (strlen(optarg) >= IF_NAMESIZE) { fprintf(stderr, "-d/--dev name too long\n"); goto end_cpu; } safe_strncpy(ifname_buf, optarg, strlen(ifname_buf)); ifindex = if_nametoindex(ifname_buf); if (!ifindex) ifindex = strtoul(optarg, NULL, 0); if (!ifindex) { fprintf(stderr, "Bad interface index or name (%d): %s\n", errno, strerror(errno)); usage(argv, long_options, __doc__, mask, true, skel->obj); goto end_cpu; } break; case 's': mask |= SAMPLE_REDIRECT_MAP_CNT; break; case 'i': interval = strtoul(optarg, NULL, 0); break; case 'S': generic = true; break; case 'x': stress_mode = true; break; case 'p': /* Selecting eBPF prog to load */ prog_name = optarg; prog = bpf_object__find_program_by_name(skel->obj, prog_name); if (!prog) { fprintf(stderr, "Failed to find program %s specified by" " option -p/--progname\n", prog_name); print_avail_progs(skel->obj); goto end_cpu; } break; case 'f': mprog_filename = optarg; break; case 'e': mprog_name = optarg; break; case 'r': redir_interface = optarg; mask |= SAMPLE_DEVMAP_XMIT_CNT_MULTI; break; case 'm': redir_map = optarg; break; case 'c': /* Add multiple CPUs */ add_cpu = strtoul(optarg, NULL, 0); if (add_cpu >= n_cpus) { fprintf(stderr, "--cpu nr too large for cpumap err (%d):%s\n", errno, strerror(errno)); usage(argv, long_options, __doc__, mask, true, skel->obj); goto end_cpu; } cpu[added_cpus++] = add_cpu; break; case 'q': qsize = strtoul(optarg, NULL, 0); break; case 'F': force = true; break; case 'v': sample_switch_mode(); break; case 'h': error = false; default: usage(argv, long_options, __doc__, mask, error, skel->obj); goto end_cpu; } } ret = EXIT_FAIL_OPTION; if (ifindex == -1) { fprintf(stderr, "Required option --dev missing\n"); usage(argv, long_options, __doc__, mask, true, skel->obj); goto end_cpu; } if (add_cpu == -1) { fprintf(stderr, "Required option --cpu missing\n" "Specify multiple --cpu option to add more\n"); usage(argv, long_options, __doc__, mask, true, skel->obj); goto end_cpu; } skel->rodata->from_match[0] = ifindex; if (redir_interface) skel->rodata->to_match[0] = if_nametoindex(redir_interface); ret = xdp_redirect_cpu__load(skel); if (ret < 0) { fprintf(stderr, "Failed to xdp_redirect_cpu__load: %s\n", strerror(errno)); goto end_cpu; } ret = bpf_obj_get_info_by_fd(bpf_map__fd(skel->maps.cpu_map), &info, &infosz); if (ret < 0) { fprintf(stderr, "Failed bpf_obj_get_info_by_fd for cpumap: %s\n", strerror(errno)); goto end_cpu; } skel->bss->cpumap_map_id = info.id; map_fd = bpf_map__fd(skel->maps.cpu_map); avail_fd = bpf_map__fd(skel->maps.cpus_available); count_fd = bpf_map__fd(skel->maps.cpus_count); ret = mark_cpus_unavailable(); if (ret < 0) { fprintf(stderr, "Unable to mark CPUs as unavailable\n"); goto end_cpu; } ret = sample_init(skel, mask); if (ret < 0) { fprintf(stderr, "Failed to initialize sample: %s\n", strerror(-ret)); ret = EXIT_FAIL; goto end_cpu; } value.bpf_prog.fd = set_cpumap_prog(skel, redir_interface, redir_map, mprog_filename, mprog_name); if (value.bpf_prog.fd < 0) { fprintf(stderr, "Failed to set CPUMAP BPF program: %s\n", strerror(-value.bpf_prog.fd)); usage(argv, long_options, __doc__, mask, true, skel->obj); ret = EXIT_FAIL_BPF; goto end_cpu; } value.qsize = qsize; for (i = 0; i < added_cpus; i++) { if (create_cpu_entry(cpu[i], &value, i, true) < 0) { fprintf(stderr, "Cannot proceed, exiting\n"); usage(argv, long_options, __doc__, mask, true, skel->obj); goto end_cpu; } } ret = EXIT_FAIL_XDP; if (sample_install_xdp(prog, ifindex, generic, force) < 0) goto end_cpu; ret = sample_run(interval, stress_mode ? stress_cpumap : NULL, &value); if (ret < 0) { fprintf(stderr, "Failed during sample run: %s\n", strerror(-ret)); ret = EXIT_FAIL; goto end_cpu; } ret = EXIT_OK; end_cpu: free(cpu); end_destroy: xdp_redirect_cpu__destroy(skel); end: sample_exit(ret); }