diff options
| author | Stanislav Fomichev <sdf@google.com> | 2021-01-15 08:34:59 -0800 |
|---|---|---|
| committer | Alexei Starovoitov <ast@kernel.org> | 2021-01-20 14:23:00 -0800 |
| commit | 9cacf81f8161111db25f98e78a7a0e32ae142b3f (patch) | |
| tree | f20ef5d2014e204f3084ebbea54515427fdef43e /net | |
| parent | 13ca51d5eb358edcb673afccb48c3440b9fda21b (diff) | |
| download | linux-9cacf81f8161111db25f98e78a7a0e32ae142b3f.tar.bz2 | |
bpf: Remove extra lock_sock for TCP_ZEROCOPY_RECEIVE
Add custom implementation of getsockopt hook for TCP_ZEROCOPY_RECEIVE.
We skip generic hooks for TCP_ZEROCOPY_RECEIVE and have a custom
call in do_tcp_getsockopt using the on-stack data. This removes
3% overhead for locking/unlocking the socket.
Without this patch:
3.38% 0.07% tcp_mmap [kernel.kallsyms] [k] __cgroup_bpf_run_filter_getsockopt
|
--3.30%--__cgroup_bpf_run_filter_getsockopt
|
--0.81%--__kmalloc
With the patch applied:
0.52% 0.12% tcp_mmap [kernel.kallsyms] [k] __cgroup_bpf_run_filter_getsockopt_kern
Note, exporting uapi/tcp.h requires removing netinet/tcp.h
from test_progs.h because those headers have confliciting
definitions.
Signed-off-by: Stanislav Fomichev <sdf@google.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/20210115163501.805133-2-sdf@google.com
Diffstat (limited to 'net')
| -rw-r--r-- | net/ipv4/tcp.c | 14 | ||||
| -rw-r--r-- | net/ipv4/tcp_ipv4.c | 1 | ||||
| -rw-r--r-- | net/ipv6/tcp_ipv6.c | 1 | ||||
| -rw-r--r-- | net/socket.c | 3 |
4 files changed, 19 insertions, 0 deletions
diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index 856ae516ac18..26aa923cf522 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -4099,6 +4099,8 @@ static int do_tcp_getsockopt(struct sock *sk, int level, return -EFAULT; lock_sock(sk); err = tcp_zerocopy_receive(sk, &zc); + err = BPF_CGROUP_RUN_PROG_GETSOCKOPT_KERN(sk, level, optname, + &zc, &len, err); release_sock(sk); if (len >= offsetofend(struct tcp_zerocopy_receive, err)) goto zerocopy_rcv_sk_err; @@ -4133,6 +4135,18 @@ zerocopy_rcv_out: return 0; } +bool tcp_bpf_bypass_getsockopt(int level, int optname) +{ + /* TCP do_tcp_getsockopt has optimized getsockopt implementation + * to avoid extra socket lock for TCP_ZEROCOPY_RECEIVE. + */ + if (level == SOL_TCP && optname == TCP_ZEROCOPY_RECEIVE) + return true; + + return false; +} +EXPORT_SYMBOL(tcp_bpf_bypass_getsockopt); + int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval, int __user *optlen) { diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index 777306b5bc22..62b6fd385a47 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -2793,6 +2793,7 @@ struct proto tcp_prot = { .shutdown = tcp_shutdown, .setsockopt = tcp_setsockopt, .getsockopt = tcp_getsockopt, + .bpf_bypass_getsockopt = tcp_bpf_bypass_getsockopt, .keepalive = tcp_set_keepalive, .recvmsg = tcp_recvmsg, .sendmsg = tcp_sendmsg, diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c index 0e1509b02cb3..8539715ff035 100644 --- a/net/ipv6/tcp_ipv6.c +++ b/net/ipv6/tcp_ipv6.c @@ -2121,6 +2121,7 @@ struct proto tcpv6_prot = { .shutdown = tcp_shutdown, .setsockopt = tcp_setsockopt, .getsockopt = tcp_getsockopt, + .bpf_bypass_getsockopt = tcp_bpf_bypass_getsockopt, .keepalive = tcp_set_keepalive, .recvmsg = tcp_recvmsg, .sendmsg = tcp_sendmsg, diff --git a/net/socket.c b/net/socket.c index 33e8b6c4e1d3..7f0617ab5437 100644 --- a/net/socket.c +++ b/net/socket.c @@ -2126,6 +2126,9 @@ SYSCALL_DEFINE5(setsockopt, int, fd, int, level, int, optname, return __sys_setsockopt(fd, level, optname, optval, optlen); } +INDIRECT_CALLABLE_DECLARE(bool tcp_bpf_bypass_getsockopt(int level, + int optname)); + /* * Get a socket option. Because we don't know the option lengths we have * to pass a user mode parameter for the protocols to sort out. |