From beecf11bc2188067824591612151c4dc6ec383c7 Mon Sep 17 00:00:00 2001 From: Stanislav Fomichev Date: Thu, 30 Apr 2020 16:31:52 -0700 Subject: bpf: Bpf_{g,s}etsockopt for struct bpf_sock_addr Currently, bpf_getsockopt and bpf_setsockopt helpers operate on the 'struct bpf_sock_ops' context in BPF_PROG_TYPE_SOCK_OPS program. Let's generalize them and make them available for 'struct bpf_sock_addr'. That way, in the future, we can allow those helpers in more places. As an example, let's expose those 'struct bpf_sock_addr' based helpers to BPF_CGROUP_INET{4,6}_CONNECT hooks. That way we can override CC before the connection is made. v3: * Expose custom helpers for bpf_sock_addr context instead of doing generic bpf_sock argument (as suggested by Daniel). Even with try_socket_lock that doesn't sleep we have a problem where context sk is already locked and socket lock is non-nestable. v2: * s/BPF_PROG_TYPE_CGROUP_SOCKOPT/BPF_PROG_TYPE_SOCK_OPS/ Signed-off-by: Stanislav Fomichev Signed-off-by: Alexei Starovoitov Acked-by: Martin KaFai Lau Acked-by: John Fastabend Link: https://lore.kernel.org/bpf/20200430233152.199403-1-sdf@google.com --- tools/testing/selftests/bpf/progs/connect4_prog.c | 46 +++++++++++++++++++++++ 1 file changed, 46 insertions(+) (limited to 'tools/testing/selftests/bpf/progs/connect4_prog.c') diff --git a/tools/testing/selftests/bpf/progs/connect4_prog.c b/tools/testing/selftests/bpf/progs/connect4_prog.c index ad3c498a8150..972918cd2d7f 100644 --- a/tools/testing/selftests/bpf/progs/connect4_prog.c +++ b/tools/testing/selftests/bpf/progs/connect4_prog.c @@ -8,6 +8,7 @@ #include #include #include +#include #include #include @@ -16,6 +17,10 @@ #define DST_REWRITE_IP4 0x7f000001U #define DST_REWRITE_PORT4 4444 +#ifndef TCP_CA_NAME_MAX +#define TCP_CA_NAME_MAX 16 +#endif + int _version SEC("version") = 1; __attribute__ ((noinline)) @@ -33,6 +38,43 @@ int do_bind(struct bpf_sock_addr *ctx) return 1; } +static __inline int verify_cc(struct bpf_sock_addr *ctx, + char expected[TCP_CA_NAME_MAX]) +{ + char buf[TCP_CA_NAME_MAX]; + int i; + + if (bpf_getsockopt(ctx, SOL_TCP, TCP_CONGESTION, &buf, sizeof(buf))) + return 1; + + for (i = 0; i < TCP_CA_NAME_MAX; i++) { + if (buf[i] != expected[i]) + return 1; + if (buf[i] == 0) + break; + } + + return 0; +} + +static __inline int set_cc(struct bpf_sock_addr *ctx) +{ + char dctcp[TCP_CA_NAME_MAX] = "dctcp"; + char cubic[TCP_CA_NAME_MAX] = "cubic"; + + if (bpf_setsockopt(ctx, SOL_TCP, TCP_CONGESTION, &dctcp, sizeof(dctcp))) + return 1; + if (verify_cc(ctx, dctcp)) + return 1; + + if (bpf_setsockopt(ctx, SOL_TCP, TCP_CONGESTION, &cubic, sizeof(cubic))) + return 1; + if (verify_cc(ctx, cubic)) + return 1; + + return 0; +} + SEC("cgroup/connect4") int connect_v4_prog(struct bpf_sock_addr *ctx) { @@ -66,6 +108,10 @@ int connect_v4_prog(struct bpf_sock_addr *ctx) bpf_sk_release(sk); + /* Rewrite congestion control. */ + if (ctx->type == SOCK_STREAM && set_cc(ctx)) + return 0; + /* Rewrite destination. */ ctx->user_ip4 = bpf_htonl(DST_REWRITE_IP4); ctx->user_port = bpf_htons(DST_REWRITE_PORT4); -- cgit v1.2.3