Re: [PATCH mptcp-next v4 05/11] mptcp: simplify push_pending

[Mat Martineau <mathew.j.martineau@linux.intel.com>]

On Sun, 2 Oct 2022, Geliang Tang wrote:

> This patch moves the duplicate code from __mptcp_push_pending() and
> __mptcp_subflow_push_pending() into a new helper function, named
> __subflow_push_pending(). And simplify __mptcp_push_pending() by
> invoking this helper.
>
> Signed-off-by: Geliang Tang <geliang.tang@suse.com>
> ---
> net/mptcp/protocol.c | 90 +++++++++++++++++++++++---------------------
> 1 file changed, 48 insertions(+), 42 deletions(-)
>
> diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c
> index 296b7135e9cf..66c97ac06a4e 100644
> --- a/net/mptcp/protocol.c
> +++ b/net/mptcp/protocol.c
> @@ -1480,12 +1480,6 @@ struct sock *mptcp_subflow_get_send(struct mptcp_sock *msk)
> 	return ssk;
> }
>
> -static void mptcp_push_release(struct sock *ssk, struct mptcp_sendmsg_info *info)
> -{
> -	tcp_push(ssk, 0, info->mss_now, tcp_sk(ssk)->nonagle, info->size_goal);
> -	release_sock(ssk);
> -}
> -
> static void mptcp_update_post_push(struct mptcp_sock *msk,
> 				   struct mptcp_data_frag *dfrag,
> 				   u32 sent)
> @@ -1517,6 +1511,52 @@ void mptcp_check_and_set_pending(struct sock *sk)
> 		mptcp_sk(sk)->push_pending |= BIT(MPTCP_PUSH_PENDING);
> }
>
> +static int __subflow_push_pending(struct sock *sk, struct sock *ssk,
> +				  struct mptcp_sendmsg_info *info)
> +{
> +	struct mptcp_sock *msk = mptcp_sk(sk);
> +	struct mptcp_data_frag *dfrag;
> +	int len, copied = 0;
> +
> +	while ((dfrag = mptcp_send_head(sk))) {
> +		info->sent = dfrag->already_sent;
> +		info->limit = dfrag->data_len;
> +		len = dfrag->data_len - dfrag->already_sent;
> +		while (len > 0) {
> +			int ret = 0;
> +
> +			ret = mptcp_sendmsg_frag(sk, ssk, dfrag, info);
> +			if (ret <= 0) {
> +				if (ret == -EAGAIN)
> +					continue;

In the older code, the scheduler got called again in the -EAGAIN case. 
With this refactored code, this could get in a very tight loop if 
mptcp_sendmsg_frag() returns. I think it's necessary to cleanly finish 
this loop iteration and let the scheduler run again in this -EAGAIN case.

> +				goto out;
> +			}
> +
> +			info->sent += ret;
> +			copied += ret;
> +			len -= ret;
> +
> +			mptcp_update_post_push(msk, dfrag, ret);
> +		}
> +		WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
> +
> +		if (msk->snd_burst <= 0 ||
> +		    !sk_stream_memory_free(ssk) ||
> +		    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
> +			goto out;
> +		}
> +		mptcp_set_timeout(sk);
> +	}
> +
> +out:
> +	if (copied) {
> +		tcp_push(ssk, 0, info->mss_now, tcp_sk(ssk)->nonagle,
> +			 info->size_goal);
> +	}
> +
> +	return copied;
> +}
> +
> void __mptcp_push_pending(struct sock *sk, unsigned int flags)
> {
> 	struct mptcp_sock *msk = mptcp_sk(sk);
> @@ -1524,48 +1564,14 @@ void __mptcp_push_pending(struct sock *sk, unsigned int flags)
> 				.flags = flags,
> 	};
> 	bool do_check_data_fin = false;
> -	struct mptcp_data_frag *dfrag;
> 	struct sock *ssk;
> -	int len;
>
> 	while (mptcp_send_head(sk) && (ssk = mptcp_subflow_get_send(msk))) {
> 		lock_sock(ssk);
> -
> -		while ((dfrag = mptcp_send_head(sk))) {
> -			info.sent = dfrag->already_sent;
> -			info.limit = dfrag->data_len;
> -			len = dfrag->data_len - dfrag->already_sent;
> -			while (len > 0) {
> -				int ret = 0;
> -
> -				ret = mptcp_sendmsg_frag(sk, ssk, dfrag, &info);
> -				if (ret <= 0) {
> -					if (ret == -EAGAIN)
> -						continue;
> -					mptcp_push_release(ssk, &info);
> -					goto out;
> -				}
> -
> -				do_check_data_fin = true;
> -				info.sent += ret;
> -				len -= ret;
> -
> -				mptcp_update_post_push(msk, dfrag, ret);
> -			}
> -			WRITE_ONCE(msk->first_pending, mptcp_send_next(sk));
> -
> -			if (msk->snd_burst <= 0 ||
> -			    !sk_stream_memory_free(ssk) ||
> -			    !mptcp_subflow_active(mptcp_subflow_ctx(ssk))) {
> -				goto out;
> -			}
> -			mptcp_set_timeout(sk);
> -		}
> -
> -		mptcp_push_release(ssk, &info);
> +		do_check_data_fin = __subflow_push_pending(sk, ssk, &info);

do_check_data_fin is a bool, but __subflow_push_pending() returns an 
int (the number of bytes copied). Looking at the other patches, the int 
value is needed. I suggest assigning the int return value to an int 
variable:

 		copied = __subflow_push_pending(sk, ssk, &info);

and updating the 'if' statement below...

> +		release_sock(ssk);
> 	}
>
> -out:
> 	/* ensure the rtx timer is running */
> 	if (!mptcp_timer_pending(sk))
> 		mptcp_reset_timer(sk);
>	if (do_check_data_fin)
> 		__mptcp_check_send_data_fin(sk);

...like this:

 	if (copied > 0)
  		__mptcp_check_send_data_fin(sk);


Thanks,

--
Mat Martineau
Intel