From patchwork Tue May 18 14:23:56 2021 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jussi Maki X-Patchwork-Id: 12264909 X-Patchwork-Delegate: bpf@iogearbox.net Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-15.8 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FORGED_FROMDOMAIN,FREEMAIL_FROM, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E3C2CC433ED for ; Tue, 18 May 2021 14:24:05 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id BF20761209 for ; Tue, 18 May 2021 14:24:05 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S230167AbhEROZW (ORCPT ); Tue, 18 May 2021 10:25:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:53984 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S230040AbhEROZV (ORCPT ); Tue, 18 May 2021 10:25:21 -0400 Received: from mail-wr1-x430.google.com (mail-wr1-x430.google.com [IPv6:2a00:1450:4864:20::430]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 70C40C061573 for ; Tue, 18 May 2021 07:24:03 -0700 (PDT) Received: by mail-wr1-x430.google.com with SMTP id n2so10531308wrm.0 for ; Tue, 18 May 2021 07:24:03 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=WUoX60AeNc4a7gZR22b8PjgFiFj8sdsjesxdRg9FsZk=; b=mGX6fhdEfBQsZpdFF+r8M73S5hTqW1Yov/69K/RjxYerNqEXuNL1TPfzVIPC3eRux2 2k12fEPGFQkgG9M4DUYcZHrgEOBl8jjjhQcUIHUzqmXDIYrh5Ep7JxAmVzUN1ljcoIzf hDJ5Ry8hQ8iIe5EnLnltzidTBFDuejbjuKocuyth2sIjfvoIl2lt0nPr9i8BYJ+Tr0k1 fGe4lMh1KZOxA470kcVpeea29It3sNreuzHYGByF2OHNmxotxaAK2ptHrcrqKfZRONmj UDtAUo16BkAjLdd8Equ5YvsnVXh5Bzdy4/SuhlLoiEde0W8FocrVc7Y1LTSCIR0nX/ki oSBw== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=WUoX60AeNc4a7gZR22b8PjgFiFj8sdsjesxdRg9FsZk=; b=pIRejxVab0nEh5oJ+/JUKywwUeVdbi2U1nKH7F3Lz5irBs7RgD2BK9UCUMRb6qXi7l Qwynr+FjUCGfq3fL8vxBgOwS+3nVuLKoxmvFhJXK7mqdFLv1TYGUoNmZrlFzVb6TNIGj Jv+Zu43hjSMGXNzFntVn3sjH2CMXNC2YfERvxUFoP40Lhjk114lskWJRXZ+qQK9c0w8R wWY1x+2rIC3if7kVbGB/nvrzL8SKa0+Id+uIy39mMWXOHzSvFb/FHrAo1L/YQPJCoxnA /wnhjkB4bYEiwpjtifhMZiV8b85cngQWO4b/OLs+4006yE3aP6C1lKmdI8pWVUGImYaF k46w== X-Gm-Message-State: AOAM531aCw5uMItP+Lp7i/XrmBSVKXb8H+oJ1xbNDNkf0+p9rD7fPe4x x8ois4563qXXwDsxQ5tsdyXyRDw1nnIq X-Google-Smtp-Source: ABdhPJwouosPyUFQnOOI4bfwZ/BG2zG2nIXPDvTOYPbNLhu6eOWiW9G79ywlFpLbRxNSXErJ5LYI6A== X-Received: by 2002:a5d:4e8c:: with SMTP id e12mr7465443wru.94.1621347841441; Tue, 18 May 2021 07:24:01 -0700 (PDT) Received: from balnab.. ([37.17.237.224]) by smtp.gmail.com with ESMTPSA id u126sm3443332wmb.9.2021.05.18.07.24.00 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Tue, 18 May 2021 07:24:00 -0700 (PDT) From: Jussi Maki To: bpf@vger.kernel.org Cc: daniel@iogearbox.net, andrii.nakryiko@gmail.com, Jussi Maki Subject: [PATCH bpf v3 2/2] selftests/bpf: Add test for l3 use of bpf_redirect_peer Date: Tue, 18 May 2021 14:23:56 +0000 Message-Id: <20210518142356.1852779-3-joamaki@gmail.com> X-Mailer: git-send-email 2.30.2 In-Reply-To: <20210518142356.1852779-1-joamaki@gmail.com> References: <20210427135550.807355-1-joamaki@gmail.com> <20210518142356.1852779-1-joamaki@gmail.com> MIME-Version: 1.0 Precedence: bulk List-ID: X-Mailing-List: bpf@vger.kernel.org X-Patchwork-Delegate: bpf@iogearbox.net Add a test case for using bpf_skb_change_head in combination with bpf_redirect_peer to redirect a packet from a L3 device to veth and back. The test uses a BPF program that adds L2 headers to the packet coming from a L3 device and then calls bpf_redirect_peer to redirect the packet to a veth device. The test fails as skb->mac_len is not set properly and thus the ethernet headers are not properly skb_pull'd in cls_bpf_classify, causing tcp_v4_rcv to point the TCP header into middle of the IP header. Signed-off-by: Jussi Maki --- .../selftests/bpf/prog_tests/tc_redirect.c | 538 ++++++++++++------ .../selftests/bpf/progs/test_tc_peer.c | 31 + 2 files changed, 391 insertions(+), 178 deletions(-) diff --git a/tools/testing/selftests/bpf/prog_tests/tc_redirect.c b/tools/testing/selftests/bpf/prog_tests/tc_redirect.c index 95ef9fcd31d8..c5e84b7a9c06 100644 --- a/tools/testing/selftests/bpf/prog_tests/tc_redirect.c +++ b/tools/testing/selftests/bpf/prog_tests/tc_redirect.c @@ -11,14 +11,17 @@ */ #define _GNU_SOURCE -#include + +#include #include #include +#include +#include #include #include #include #include -#include +#include #include "test_progs.h" #include "network_helpers.h" @@ -32,18 +35,25 @@ #define IP4_SRC "172.16.1.100" #define IP4_DST "172.16.2.100" +#define IP4_TUN_SRC "172.17.1.100" +#define IP4_TUN_FWD "172.17.1.200" #define IP4_PORT 9004 -#define IP6_SRC "::1:dead:beef:cafe" -#define IP6_DST "::2:dead:beef:cafe" +#define IP6_SRC "0::1:dead:beef:cafe" +#define IP6_DST "0::2:dead:beef:cafe" +#define IP6_TUN_SRC "1::1:dead:beef:cafe" +#define IP6_TUN_FWD "1::2:dead:beef:cafe" #define IP6_PORT 9006 #define IP4_SLL "169.254.0.1" #define IP4_DLL "169.254.0.2" #define IP4_NET "169.254.0.0" +#define MAC_DST_FWD "00:11:22:33:44:55" +#define MAC_DST "00:22:33:44:55:66" + #define IFADDR_STR_LEN 18 -#define PING_ARGS "-c 3 -w 10 -q" +#define PING_ARGS "-i 0.2 -c 3 -w 10 -q" #define SRC_PROG_PIN_FILE "/sys/fs/bpf/test_tc_src" #define DST_PROG_PIN_FILE "/sys/fs/bpf/test_tc_dst" @@ -51,120 +61,104 @@ #define TIMEOUT_MILLIS 10000 -#define MAX_PROC_MODS 128 -#define MAX_PROC_VALUE_LEN 16 - #define log_err(MSG, ...) \ fprintf(stderr, "(%s:%d: errno: %s) " MSG "\n", \ __FILE__, __LINE__, strerror(errno), ##__VA_ARGS__) -struct proc_mod { - char path[PATH_MAX]; - char oldval[MAX_PROC_VALUE_LEN]; - int oldlen; -}; - static const char * const namespaces[] = {NS_SRC, NS_FWD, NS_DST, NULL}; -static int root_netns_fd = -1; -static int num_proc_mods; -static struct proc_mod proc_mods[MAX_PROC_MODS]; -/** - * modify_proc() - Modify entry in /proc - * - * Modifies an entry in /proc and saves the original value for later - * restoration with restore_proc(). - */ -static int modify_proc(const char *path, const char *newval) +static int write_file(const char *path, const char *newval) { - struct proc_mod *mod; FILE *f; - if (num_proc_mods + 1 > MAX_PROC_MODS) - return -1; - f = fopen(path, "r+"); if (!f) return -1; - - mod = &proc_mods[num_proc_mods]; - num_proc_mods++; - - strncpy(mod->path, path, PATH_MAX); - - if (!fread(mod->oldval, 1, MAX_PROC_VALUE_LEN, f)) { - log_err("reading from %s failed", path); - goto fail; - } - rewind(f); if (fwrite(newval, strlen(newval), 1, f) != 1) { log_err("writing to %s failed", path); - goto fail; + fclose(f); + return -1; } - fclose(f); return 0; - -fail: - fclose(f); - num_proc_mods--; - return -1; } -/** - * restore_proc() - Restore all /proc modifications - */ -static void restore_proc(void) +struct nstoken { + int orig_netns_fd; +}; + +static int setns_by_fd(int nsfd) { - int i; + int err; - for (i = 0; i < num_proc_mods; i++) { - struct proc_mod *mod = &proc_mods[i]; - FILE *f; + err = setns(nsfd, CLONE_NEWNET); + close(nsfd); - f = fopen(mod->path, "w"); - if (!f) { - log_err("fopen of %s failed", mod->path); - continue; - } + if (!ASSERT_OK(err, "setns")) + return err; - if (fwrite(mod->oldval, mod->oldlen, 1, f) != 1) - log_err("fwrite to %s failed", mod->path); + /* Switch /sys to the new namespace so that e.g. /sys/class/net + * reflects the devices in the new namespace. + */ + err = unshare(CLONE_NEWNS); + if (!ASSERT_OK(err, "unshare")) + return err; - fclose(f); - } - num_proc_mods = 0; + err = umount2("/sys", MNT_DETACH); + if (!ASSERT_OK(err, "umount2 /sys")) + return err; + + err = mount("sysfs", "/sys", "sysfs", 0, NULL); + if (!ASSERT_OK(err, "mount /sys")) + return err; + + err = mount("bpffs", "/sys/fs/bpf", "bpf", 0, NULL); + if (!ASSERT_OK(err, "mount /sys/fs/bpf")) + return err; + + return 0; } /** - * setns_by_name() - Set networks namespace by name + * open_netns() - Switch to specified network namespace by name. + * + * Returns token with which to restore the original namespace + * using close_netns(). */ -static int setns_by_name(const char *name) +static struct nstoken *open_netns(const char *name) { int nsfd; char nspath[PATH_MAX]; int err; + struct nstoken *token; + + token = malloc(sizeof(struct nstoken)); + if (!ASSERT_OK_PTR(token, "malloc token")) + return NULL; + + token->orig_netns_fd = open("/proc/self/ns/net", O_RDONLY); + if (!ASSERT_GE(token->orig_netns_fd, 0, "open /proc/self/ns/net")) + goto fail; snprintf(nspath, sizeof(nspath), "%s/%s", "/var/run/netns", name); nsfd = open(nspath, O_RDONLY | O_CLOEXEC); - if (nsfd < 0) - return nsfd; + if (!ASSERT_GE(nsfd, 0, "open netns fd")) + goto fail; - err = setns(nsfd, CLONE_NEWNET); - close(nsfd); + err = setns_by_fd(nsfd); + if (!ASSERT_OK(err, "setns_by_fd")) + goto fail; - return err; + return token; +fail: + free(token); + return NULL; } -/** - * setns_root() - Set network namespace to original (root) namespace - * - * Not expected to ever fail, so error not returned, but failure logged - * and test marked as failed. - */ -static void setns_root(void) +static void close_netns(struct nstoken *token) { - ASSERT_OK(setns(root_netns_fd, CLONE_NEWNET), "setns root"); + ASSERT_OK(setns_by_fd(token->orig_netns_fd), "setns_by_fd"); + free(token); } static int netns_setup_namespaces(const char *verb) @@ -237,15 +231,17 @@ static int get_ifindex(const char *name) static int netns_setup_links_and_routes(struct netns_setup_result *result) { + struct nstoken *nstoken = NULL; char veth_src_fwd_addr[IFADDR_STR_LEN+1] = {}; - char veth_dst_fwd_addr[IFADDR_STR_LEN+1] = {}; SYS("ip link add veth_src type veth peer name veth_src_fwd"); SYS("ip link add veth_dst type veth peer name veth_dst_fwd"); + + SYS("ip link set veth_dst_fwd address " MAC_DST_FWD); + SYS("ip link set veth_dst address " MAC_DST); + if (get_ifaddr("veth_src_fwd", veth_src_fwd_addr)) goto fail; - if (get_ifaddr("veth_dst_fwd", veth_dst_fwd_addr)) - goto fail; result->ifindex_veth_src_fwd = get_ifindex("veth_src_fwd"); if (result->ifindex_veth_src_fwd < 0) @@ -260,7 +256,8 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result) SYS("ip link set veth_dst netns " NS_DST); /** setup in 'src' namespace */ - if (!ASSERT_OK(setns_by_name(NS_SRC), "setns src")) + nstoken = open_netns(NS_SRC); + if (!ASSERT_OK_PTR(nstoken, "setns src")) goto fail; SYS("ip addr add " IP4_SRC "/32 dev veth_src"); @@ -276,8 +273,11 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result) SYS("ip neigh add " IP6_DST " dev veth_src lladdr %s", veth_src_fwd_addr); + close_netns(nstoken); + /** setup in 'fwd' namespace */ - if (!ASSERT_OK(setns_by_name(NS_FWD), "setns fwd")) + nstoken = open_netns(NS_FWD); + if (!ASSERT_OK_PTR(nstoken, "setns fwd")) goto fail; /* The fwd netns automatically gets a v6 LL address / routes, but also @@ -294,8 +294,11 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result) SYS("ip route add " IP4_DST "/32 dev veth_dst_fwd scope global"); SYS("ip route add " IP6_DST "/128 dev veth_dst_fwd scope global"); + close_netns(nstoken); + /** setup in 'dst' namespace */ - if (!ASSERT_OK(setns_by_name(NS_DST), "setns dst")) + nstoken = open_netns(NS_DST); + if (!ASSERT_OK_PTR(nstoken, "setns dst")) goto fail; SYS("ip addr add " IP4_DST "/32 dev veth_dst"); @@ -306,23 +309,20 @@ static int netns_setup_links_and_routes(struct netns_setup_result *result) SYS("ip route add " IP4_NET "/16 dev veth_dst scope global"); SYS("ip route add " IP6_SRC "/128 dev veth_dst scope global"); - SYS("ip neigh add " IP4_SRC " dev veth_dst lladdr %s", - veth_dst_fwd_addr); - SYS("ip neigh add " IP6_SRC " dev veth_dst lladdr %s", - veth_dst_fwd_addr); + SYS("ip neigh add " IP4_SRC " dev veth_dst lladdr " MAC_DST_FWD); + SYS("ip neigh add " IP6_SRC " dev veth_dst lladdr " MAC_DST_FWD); + + close_netns(nstoken); - setns_root(); return 0; fail: - setns_root(); + if (nstoken) + close_netns(nstoken); return -1; } static int netns_load_bpf(void) { - if (!ASSERT_OK(setns_by_name(NS_FWD), "setns fwd")) - return -1; - SYS("tc qdisc add dev veth_src_fwd clsact"); SYS("tc filter add dev veth_src_fwd ingress bpf da object-pinned " SRC_PROG_PIN_FILE); @@ -335,42 +335,29 @@ static int netns_load_bpf(void) SYS("tc filter add dev veth_dst_fwd egress bpf da object-pinned " CHK_PROG_PIN_FILE); - setns_root(); - return -1; -fail: - setns_root(); - return -1; -} - -static int netns_unload_bpf(void) -{ - if (!ASSERT_OK(setns_by_name(NS_FWD), "setns fwd")) - goto fail; - SYS("tc qdisc delete dev veth_src_fwd clsact"); - SYS("tc qdisc delete dev veth_dst_fwd clsact"); - - setns_root(); return 0; fail: - setns_root(); return -1; } - static void test_tcp(int family, const char *addr, __u16 port) { int listen_fd = -1, accept_fd = -1, client_fd = -1; char buf[] = "testing testing"; int n; + struct nstoken *nstoken; - if (!ASSERT_OK(setns_by_name(NS_DST), "setns dst")) + nstoken = open_netns(NS_DST); + if (!ASSERT_OK_PTR(nstoken, "setns dst")) return; listen_fd = start_server(family, SOCK_STREAM, addr, port, 0); if (!ASSERT_GE(listen_fd, 0, "listen")) goto done; - if (!ASSERT_OK(setns_by_name(NS_SRC), "setns src")) + close_netns(nstoken); + nstoken = open_netns(NS_SRC); + if (!ASSERT_OK_PTR(nstoken, "setns src")) goto done; client_fd = connect_to_fd(listen_fd, TIMEOUT_MILLIS); @@ -392,7 +379,8 @@ static void test_tcp(int family, const char *addr, __u16 port) ASSERT_EQ(n, sizeof(buf), "recv from server"); done: - setns_root(); + if (nstoken) + close_netns(nstoken); if (listen_fd >= 0) close(listen_fd); if (accept_fd >= 0) @@ -405,7 +393,7 @@ static int test_ping(int family, const char *addr) { const char *ping = family == AF_INET6 ? "ping6" : "ping"; - SYS("ip netns exec " NS_SRC " %s " PING_ARGS " %s", ping, addr); + SYS("ip netns exec " NS_SRC " %s " PING_ARGS " %s > /dev/null", ping, addr); return 0; fail: return -1; @@ -419,19 +407,37 @@ static void test_connectivity(void) test_ping(AF_INET6, IP6_DST); } +static int set_forwarding(bool enable) +{ + int err; + + err = write_file("/proc/sys/net/ipv4/ip_forward", enable ? "1" : "0"); + if (!ASSERT_OK(err, "set ipv4.ip_forward=0")) + return err; + + err = write_file("/proc/sys/net/ipv6/conf/all/forwarding", enable ? "1" : "0"); + if (!ASSERT_OK(err, "set ipv6.forwarding=0")) + return err; + + return 0; +} + static void test_tc_redirect_neigh_fib(struct netns_setup_result *setup_result) { - struct test_tc_neigh_fib *skel; + struct nstoken *nstoken = NULL; + struct test_tc_neigh_fib *skel = NULL; int err; + nstoken = open_netns(NS_FWD); + if (!ASSERT_OK_PTR(nstoken, "setns fwd")) + return; + skel = test_tc_neigh_fib__open(); if (!ASSERT_OK_PTR(skel, "test_tc_neigh_fib__open")) - return; + goto done; - if (!ASSERT_OK(test_tc_neigh_fib__load(skel), "test_tc_neigh_fib__load")) { - test_tc_neigh_fib__destroy(skel); - return; - } + if (!ASSERT_OK(test_tc_neigh_fib__load(skel), "test_tc_neigh_fib__load")) + goto done; err = bpf_program__pin(skel->progs.tc_src, SRC_PROG_PIN_FILE); if (!ASSERT_OK(err, "pin " SRC_PROG_PIN_FILE)) @@ -449,46 +455,37 @@ static void test_tc_redirect_neigh_fib(struct netns_setup_result *setup_result) goto done; /* bpf_fib_lookup() checks if forwarding is enabled */ - if (!ASSERT_OK(setns_by_name(NS_FWD), "setns fwd")) - goto done; - - err = modify_proc("/proc/sys/net/ipv4/ip_forward", "1"); - if (!ASSERT_OK(err, "set ipv4.ip_forward")) + if (!ASSERT_OK(set_forwarding(true), "enable forwarding")) goto done; - err = modify_proc("/proc/sys/net/ipv6/conf/all/forwarding", "1"); - if (!ASSERT_OK(err, "set ipv6.forwarding")) - goto done; - setns_root(); - test_connectivity(); + done: - bpf_program__unpin(skel->progs.tc_src, SRC_PROG_PIN_FILE); - bpf_program__unpin(skel->progs.tc_chk, CHK_PROG_PIN_FILE); - bpf_program__unpin(skel->progs.tc_dst, DST_PROG_PIN_FILE); - test_tc_neigh_fib__destroy(skel); - netns_unload_bpf(); - setns_root(); - restore_proc(); + if (skel) + test_tc_neigh_fib__destroy(skel); + close_netns(nstoken); } static void test_tc_redirect_neigh(struct netns_setup_result *setup_result) { - struct test_tc_neigh *skel; + struct nstoken *nstoken = NULL; + struct test_tc_neigh *skel = NULL; int err; + nstoken = open_netns(NS_FWD); + if (!ASSERT_OK_PTR(nstoken, "setns fwd")) + return; + skel = test_tc_neigh__open(); if (!ASSERT_OK_PTR(skel, "test_tc_neigh__open")) - return; + goto done; skel->rodata->IFINDEX_SRC = setup_result->ifindex_veth_src_fwd; skel->rodata->IFINDEX_DST = setup_result->ifindex_veth_dst_fwd; err = test_tc_neigh__load(skel); - if (!ASSERT_OK(err, "test_tc_neigh__load")) { - test_tc_neigh__destroy(skel); - return; - } + if (!ASSERT_OK(err, "test_tc_neigh__load")) + goto done; err = bpf_program__pin(skel->progs.tc_src, SRC_PROG_PIN_FILE); if (!ASSERT_OK(err, "pin " SRC_PROG_PIN_FILE)) @@ -505,34 +502,37 @@ static void test_tc_redirect_neigh(struct netns_setup_result *setup_result) if (netns_load_bpf()) goto done; + if (!ASSERT_OK(set_forwarding(false), "disable forwarding")) + goto done; + test_connectivity(); done: - bpf_program__unpin(skel->progs.tc_src, SRC_PROG_PIN_FILE); - bpf_program__unpin(skel->progs.tc_chk, CHK_PROG_PIN_FILE); - bpf_program__unpin(skel->progs.tc_dst, DST_PROG_PIN_FILE); - test_tc_neigh__destroy(skel); - netns_unload_bpf(); - setns_root(); + if (skel) + test_tc_neigh__destroy(skel); + close_netns(nstoken); } static void test_tc_redirect_peer(struct netns_setup_result *setup_result) { + struct nstoken *nstoken; struct test_tc_peer *skel; int err; + nstoken = open_netns(NS_FWD); + if (!ASSERT_OK_PTR(nstoken, "setns fwd")) + return; + skel = test_tc_peer__open(); if (!ASSERT_OK_PTR(skel, "test_tc_peer__open")) - return; + goto done; skel->rodata->IFINDEX_SRC = setup_result->ifindex_veth_src_fwd; skel->rodata->IFINDEX_DST = setup_result->ifindex_veth_dst_fwd; err = test_tc_peer__load(skel); - if (!ASSERT_OK(err, "test_tc_peer__load")) { - test_tc_peer__destroy(skel); - return; - } + if (!ASSERT_OK(err, "test_tc_peer__load")) + goto done; err = bpf_program__pin(skel->progs.tc_src, SRC_PROG_PIN_FILE); if (!ASSERT_OK(err, "pin " SRC_PROG_PIN_FILE)) @@ -549,41 +549,223 @@ static void test_tc_redirect_peer(struct netns_setup_result *setup_result) if (netns_load_bpf()) goto done; + if (!ASSERT_OK(set_forwarding(false), "disable forwarding")) + goto done; + test_connectivity(); done: - bpf_program__unpin(skel->progs.tc_src, SRC_PROG_PIN_FILE); - bpf_program__unpin(skel->progs.tc_chk, CHK_PROG_PIN_FILE); - bpf_program__unpin(skel->progs.tc_dst, DST_PROG_PIN_FILE); - test_tc_peer__destroy(skel); - netns_unload_bpf(); - setns_root(); + if (skel) + test_tc_peer__destroy(skel); + close_netns(nstoken); } -void test_tc_redirect(void) +static int tun_open(char *name) { - struct netns_setup_result setup_result; + struct ifreq ifr; + int fd, err; + + fd = open("/dev/net/tun", O_RDWR); + if (!ASSERT_GE(fd, 0, "open /dev/net/tun")) + return -1; + + memset(&ifr, 0, sizeof(ifr)); + + ifr.ifr_flags = IFF_TUN | IFF_NO_PI; + if (*name) + strncpy(ifr.ifr_name, name, IFNAMSIZ); + + err = ioctl(fd, TUNSETIFF, &ifr); + if (!ASSERT_OK(err, "ioctl TUNSETIFF")) + goto fail; + + SYS("ip link set dev %s up", name); + + return fd; +fail: + close(fd); + return -1; +} + +#define MAX(a, b) ((a) > (b) ? (a) : (b)) +enum { + SRC_TO_TARGET = 0, + TARGET_TO_SRC = 1, +}; + +static int tun_relay_loop(int src_fd, int target_fd) +{ + fd_set rfds, wfds; + + FD_ZERO(&rfds); + FD_ZERO(&wfds); - root_netns_fd = open("/proc/self/ns/net", O_RDONLY); - if (!ASSERT_GE(root_netns_fd, 0, "open /proc/self/ns/net")) + for (;;) { + char buf[1500]; + int direction, nread, nwrite; + + FD_SET(src_fd, &rfds); + FD_SET(target_fd, &rfds); + + if (select(1 + MAX(src_fd, target_fd), &rfds, NULL, NULL, NULL) < 0) { + log_err("select failed"); + return 1; + } + + direction = FD_ISSET(src_fd, &rfds) ? SRC_TO_TARGET : TARGET_TO_SRC; + + nread = read(direction == SRC_TO_TARGET ? src_fd : target_fd, buf, sizeof(buf)); + if (nread < 0) { + log_err("read failed"); + return 1; + } + + nwrite = write(direction == SRC_TO_TARGET ? target_fd : src_fd, buf, nread); + if (nwrite != nread) { + log_err("write failed"); + return 1; + } + } +} + +static void test_tc_redirect_peer_l3(struct netns_setup_result *setup_result) +{ + struct test_tc_peer *skel = NULL; + struct nstoken *nstoken = NULL; + int err; + int tunnel_pid = -1; + int src_fd, target_fd; + int ifindex; + + /* Start a L3 TUN/TAP tunnel between the src and dst namespaces. + * This test is using TUN/TAP instead of e.g. IPIP or GRE tunnel as those + * expose the L2 headers encapsulating the IP packet to BPF and hence + * don't have skb in suitable state for this test. Alternative to TUN/TAP + * would be e.g. Wireguard which would appear as a pure L3 device to BPF, + * but that requires much more complicated setup. + */ + nstoken = open_netns(NS_SRC); + if (!ASSERT_OK_PTR(nstoken, "setns " NS_SRC)) return; - if (netns_setup_namespaces("add")) - goto done; + src_fd = tun_open("tun_src"); + if (!ASSERT_GE(src_fd, 0, "tun_open tun_src")) + goto fail; - if (netns_setup_links_and_routes(&setup_result)) - goto done; + close_netns(nstoken); - if (test__start_subtest("tc_redirect_peer")) - test_tc_redirect_peer(&setup_result); + nstoken = open_netns(NS_FWD); + if (!ASSERT_OK_PTR(nstoken, "setns " NS_FWD)) + goto fail; - if (test__start_subtest("tc_redirect_neigh")) - test_tc_redirect_neigh(&setup_result); + target_fd = tun_open("tun_fwd"); + if (!ASSERT_GE(target_fd, 0, "tun_open tun_fwd")) + goto fail; - if (test__start_subtest("tc_redirect_neigh_fib")) - test_tc_redirect_neigh_fib(&setup_result); + tunnel_pid = fork(); + if (!ASSERT_GE(tunnel_pid, 0, "fork tun_relay_loop")) + goto fail; -done: - close(root_netns_fd); - netns_setup_namespaces("delete"); + if (tunnel_pid == 0) + exit(tun_relay_loop(src_fd, target_fd)); + + skel = test_tc_peer__open(); + if (!ASSERT_OK_PTR(skel, "test_tc_peer__open")) + goto fail; + + ifindex = get_ifindex("tun_fwd"); + if (!ASSERT_GE(ifindex, 0, "get_ifindex tun_fwd")) + goto fail; + + skel->rodata->IFINDEX_SRC = ifindex; + skel->rodata->IFINDEX_DST = setup_result->ifindex_veth_dst_fwd; + + err = test_tc_peer__load(skel); + if (!ASSERT_OK(err, "test_tc_peer__load")) + goto fail; + + err = bpf_program__pin(skel->progs.tc_src_l3, SRC_PROG_PIN_FILE); + if (!ASSERT_OK(err, "pin " SRC_PROG_PIN_FILE)) + goto fail; + + err = bpf_program__pin(skel->progs.tc_dst_l3, DST_PROG_PIN_FILE); + if (!ASSERT_OK(err, "pin " DST_PROG_PIN_FILE)) + goto fail; + + err = bpf_program__pin(skel->progs.tc_chk, CHK_PROG_PIN_FILE); + if (!ASSERT_OK(err, "pin " CHK_PROG_PIN_FILE)) + goto fail; + + /* Load "tc_src_l3" to the tun_fwd interface to redirect packets + * towards dst, and "tc_dst" to redirect packets + * and "tc_chk" on veth_dst_fwd to drop non-redirected packets. + */ + SYS("tc qdisc add dev tun_fwd clsact"); + SYS("tc filter add dev tun_fwd ingress bpf da object-pinned " + SRC_PROG_PIN_FILE); + + SYS("tc qdisc add dev veth_dst_fwd clsact"); + SYS("tc filter add dev veth_dst_fwd ingress bpf da object-pinned " + DST_PROG_PIN_FILE); + SYS("tc filter add dev veth_dst_fwd egress bpf da object-pinned " + CHK_PROG_PIN_FILE); + + /* Setup route and neigh tables */ + SYS("ip -netns " NS_SRC " addr add dev tun_src " IP4_TUN_SRC "/24"); + SYS("ip -netns " NS_FWD " addr add dev tun_fwd " IP4_TUN_FWD "/24"); + + SYS("ip -netns " NS_SRC " addr add dev tun_src " IP6_TUN_SRC "/64 nodad"); + SYS("ip -netns " NS_FWD " addr add dev tun_fwd " IP6_TUN_FWD "/64 nodad"); + + SYS("ip -netns " NS_SRC " route del " IP4_DST "/32 dev veth_src scope global"); + SYS("ip -netns " NS_SRC " route add " IP4_DST "/32 via " IP4_TUN_FWD + " dev tun_src scope global"); + SYS("ip -netns " NS_DST " route add " IP4_TUN_SRC "/32 dev veth_dst scope global"); + SYS("ip -netns " NS_SRC " route del " IP6_DST "/128 dev veth_src scope global"); + SYS("ip -netns " NS_SRC " route add " IP6_DST "/128 via " IP6_TUN_FWD + " dev tun_src scope global"); + SYS("ip -netns " NS_DST " route add " IP6_TUN_SRC "/128 dev veth_dst scope global"); + + SYS("ip -netns " NS_DST " neigh add " IP4_TUN_SRC " dev veth_dst lladdr " MAC_DST_FWD); + SYS("ip -netns " NS_DST " neigh add " IP6_TUN_SRC " dev veth_dst lladdr " MAC_DST_FWD); + + if (!ASSERT_OK(set_forwarding(false), "disable forwarding")) + goto fail; + + test_connectivity(); + +fail: + if (tunnel_pid > 0) { + kill(tunnel_pid, SIGTERM); + waitpid(tunnel_pid, NULL, 0); + } + if (src_fd >= 0) + close(src_fd); + if (target_fd >= 0) + close(target_fd); + if (skel) + test_tc_peer__destroy(skel); + if (nstoken) + close_netns(nstoken); +} + +#define RUN_TEST(name) \ + ({ \ + if (test__start_subtest(#name)) \ + if (ASSERT_OK(netns_setup_namespaces("add"), "setup namespaces")) { \ + if (ASSERT_OK(netns_setup_links_and_routes(&setup_result), \ + "setup links and routes")) \ + test_ ## name(&setup_result); \ + netns_setup_namespaces("delete"); \ + } \ + }) + +void test_tc_redirect(void) +{ + struct netns_setup_result setup_result; + + RUN_TEST(tc_redirect_peer); + RUN_TEST(tc_redirect_peer_l3); + RUN_TEST(tc_redirect_neigh); + RUN_TEST(tc_redirect_neigh_fib); } diff --git a/tools/testing/selftests/bpf/progs/test_tc_peer.c b/tools/testing/selftests/bpf/progs/test_tc_peer.c index 72c72950c3bb..ad70f4ce9254 100644 --- a/tools/testing/selftests/bpf/progs/test_tc_peer.c +++ b/tools/testing/selftests/bpf/progs/test_tc_peer.c @@ -5,12 +5,17 @@ #include #include #include +#include +#include #include static volatile const __u32 IFINDEX_SRC; static volatile const __u32 IFINDEX_DST; +static const __u8 src_mac[] = {0x00, 0x11, 0x22, 0x33, 0x44, 0x55}; +static const __u8 dst_mac[] = {0x00, 0x22, 0x33, 0x44, 0x55, 0x66}; + SEC("classifier/chk_egress") int tc_chk(struct __sk_buff *skb) { @@ -29,4 +34,30 @@ int tc_src(struct __sk_buff *skb) return bpf_redirect_peer(IFINDEX_DST, 0); } +SEC("classifier/dst_ingress_l3") +int tc_dst_l3(struct __sk_buff *skb) +{ + return bpf_redirect(IFINDEX_SRC, 0); +} + +SEC("classifier/src_ingress_l3") +int tc_src_l3(struct __sk_buff *skb) +{ + __u16 proto = skb->protocol; + + if (bpf_skb_change_head(skb, ETH_HLEN, 0) != 0) + return TC_ACT_SHOT; + + if (bpf_skb_store_bytes(skb, 0, &src_mac, ETH_ALEN, 0) != 0) + return TC_ACT_SHOT; + + if (bpf_skb_store_bytes(skb, ETH_ALEN, &dst_mac, ETH_ALEN, 0) != 0) + return TC_ACT_SHOT; + + if (bpf_skb_store_bytes(skb, ETH_ALEN + ETH_ALEN, &proto, sizeof(__u16), 0) != 0) + return TC_ACT_SHOT; + + return bpf_redirect_peer(IFINDEX_DST, 0); +} + char __license[] SEC("license") = "GPL";