Re: [RFC PATCH v3 8/9] vmci/vsock: check SO_RCVLOWAT before wake up reader

[Stefano Garzarella]

On Wed, Aug 03, 2022 at 02:05:52PM +0000, Arseniy Krasnov wrote:
> This adds extra condition to wake up data reader: do it only when number
> of readable bytes >= SO_RCVLOWAT. Otherwise, there is no sense to kick
> user,because it will wait until SO_RCVLOWAT bytes will be dequeued.

Ditto as previous patch.

@Bryan, @Vishnu, plaese, can you review/ack also this patch?

Thanks,
Stefano

> Signed-off-by: Arseniy Krasnov <AVKrasnov@xxxxxxxxxxxxxx>
> ---
> net/vmw_vsock/vmci_transport_notify.c        | 2 +-
> net/vmw_vsock/vmci_transport_notify_qstate.c | 4 ++--
> 2 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/net/vmw_vsock/vmci_transport_notify.c b/net/vmw_vsock/vmci_transport_notify.c
> index 852097e2b9e6..7c3a7db134b2 100644
> --- a/net/vmw_vsock/vmci_transport_notify.c
> +++ b/net/vmw_vsock/vmci_transport_notify.c
> @@ -307,7 +307,7 @@ vmci_transport_handle_wrote(struct sock *sk,
> 	struct vsock_sock *vsk = vsock_sk(sk);
> 	PKT_FIELD(vsk, sent_waiting_read) = false;
> #endif
> -	sk->sk_data_ready(sk);
> +	vsock_data_ready(sk);
> }
>
> static void vmci_transport_notify_pkt_socket_init(struct sock *sk)
> diff --git a/net/vmw_vsock/vmci_transport_notify_qstate.c b/net/vmw_vsock/vmci_transport_notify_qstate.c
> index 12f0cb8fe998..e96a88d850a8 100644
> --- a/net/vmw_vsock/vmci_transport_notify_qstate.c
> +++ b/net/vmw_vsock/vmci_transport_notify_qstate.c
> @@ -84,7 +84,7 @@ vmci_transport_handle_wrote(struct sock *sk,
> 			    bool bottom_half,
> 			    struct sockaddr_vm *dst, struct sockaddr_vm *src)
> {
> -	sk->sk_data_ready(sk);
> +	vsock_data_ready(sk);
> }
>
> static void vsock_block_update_write_window(struct sock *sk)
> @@ -282,7 +282,7 @@ vmci_transport_notify_pkt_recv_post_dequeue(
> 		/* See the comment in
> 		 * vmci_transport_notify_pkt_send_post_enqueue().
> 		 */
> -		sk->sk_data_ready(sk);
> +		vsock_data_ready(sk);
> 	}
>
> 	return err;
> --
> 2.25.1