[v3,bpf-next,0/2] Add support to set window_clamp from bpf setsockops

This patch contains support to set tcp window_field field from bpf setsockops.

v2: Used TCP_WINDOW_CLAMP setsockopt logic for bpf_setsockopt (review comment addressed)

v3: Created a common function for duplicated code (review comment addressed)


Prankur gupta (2):
  bpf: Adds support for setting window clamp
  selftests/bpf: Add Userspace tests for TCP_WINDOW_CLAMP

 include/net/tcp.h                             |  1 +
 net/core/filter.c                             |  3 ++
 net/ipv4/tcp.c                                | 23 ++++++++-----
 tools/testing/selftests/bpf/bpf_tcp_helpers.h |  1 +
 .../selftests/bpf/prog_tests/tcpbpf_user.c    |  4 +++
 .../selftests/bpf/progs/test_tcpbpf_kern.c    | 33 +++++++++++++++++++
 tools/testing/selftests/bpf/test_tcpbpf.h     |  2 ++
 7 files changed, 58 insertions(+), 9 deletions(-)

On 12/2/20 9:29 PM, Prankur gupta wrote:
> Adds a new bpf_setsockopt for TCP sockets, TCP_BPF_WINDOW_CLAMP,
> which sets the maximum receiver window size. It will be useful for
> limiting receiver window based on RTT.
> 
> Signed-off-by: Prankur gupta <prankgup@fb.com>
> ---
>  include/net/tcp.h |  1 +
>  net/core/filter.c |  3 +++
>  net/ipv4/tcp.c    | 23 ++++++++++++++---------
>  3 files changed, 18 insertions(+), 9 deletions(-)
> 
> diff --git a/include/net/tcp.h b/include/net/tcp.h
> index 4aba0f069b05..39ced5882fe3 100644
> --- a/include/net/tcp.h
> +++ b/include/net/tcp.h
> @@ -406,6 +406,7 @@ void tcp_syn_ack_timeout(const struct request_sock *req);
>  int tcp_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, int nonblock,
>  		int flags, int *addr_len);
>  int tcp_set_rcvlowat(struct sock *sk, int val);
> +int tcp_set_window_clamp(struct sock *sk, struct tcp_sock *tp, int val);
>  void tcp_data_ready(struct sock *sk);
>  #ifdef CONFIG_MMU
>  int tcp_mmap(struct file *file, struct socket *sock,
> diff --git a/net/core/filter.c b/net/core/filter.c
> index 2ca5eecebacf..d6225842cfb1 100644
> --- a/net/core/filter.c
> +++ b/net/core/filter.c
> @@ -4910,6 +4910,9 @@ static int _bpf_setsockopt(struct sock *sk, int level, int optname,
>  				tp->notsent_lowat = val;
>  				sk->sk_write_space(sk);
>  				break;
> +			case TCP_WINDOW_CLAMP:
> +				ret = tcp_set_window_clamp(sk, tp, val);
> +				break;
>  			default:
>  				ret = -EINVAL;
>  			}
> diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c
> index b2bc3d7fe9e8..312feb8fcae5 100644
> --- a/net/ipv4/tcp.c
> +++ b/net/ipv4/tcp.c
> @@ -3022,6 +3022,19 @@ int tcp_sock_set_keepcnt(struct sock *sk, int val)
>  }
>  EXPORT_SYMBOL(tcp_sock_set_keepcnt);
>  
> +int tcp_set_window_clamp(struct sock *sk, struct tcp_sock *tp, int val)

No TCP function pass both sk and tp (which are identical values)

Prefer :

int tcp_set_window_clamp(struct sock *sk, int val)
{
        struct tcp_sock *tp = tcp_sk(sk);
        ...

This will allow optimal code generation.

> +{
> +	if (!val) {
> +		if (sk->sk_state != TCP_CLOSE)
> +			return -EINVAL;
> +		tp->window_clamp = 0;
> +	} else {
> +		tp->window_clamp = val < SOCK_MIN_RCVBUF / 2 ?
> +			SOCK_MIN_RCVBUF / 2 : val;
> +	}
> +	return 0;
> +}
> +
>  /*
>   *	Socket option code for TCP.
>   */
> @@ -3235,15 +3248,7 @@ static int do_tcp_setsockopt(struct sock *sk, int level, int optname,
>  		break;
>  
>  	case TCP_WINDOW_CLAMP:
> -		if (!val) {
> -			if (sk->sk_state != TCP_CLOSE) {
> -				err = -EINVAL;
> -				break;
> -			}
> -			tp->window_clamp = 0;
> -		} else
> -			tp->window_clamp = val < SOCK_MIN_RCVBUF / 2 ?
> -						SOCK_MIN_RCVBUF / 2 : val;
> +		err = tcp_set_window_clamp(sk, tp, val);
>  		break;
>  
>  	case TCP_QUICKACK:
>