Re: [PATCH 31/33] sctp: add sctp_sock_set_nodelay
From: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Date: 2020-05-20 23:10:08
Also in:
ceph-devel, linux-cifs, linux-nfs, linux-nvme, linux-rdma, linux-sctp, lkml, ocfs2-devel, target-devel
On Wed, May 20, 2020 at 09:55:07PM +0200, Christoph Hellwig wrote:
quoted hunk ↗ jump to hunk
Add a helper to directly set the SCTP_NODELAY sockopt from kernel space without going through a fake uaccess. Signed-off-by: Christoph Hellwig <hch@lst.de> --- fs/dlm/lowcomms.c | 10 ++-------- include/net/sctp/sctp.h | 7 +++++++ 2 files changed, 9 insertions(+), 8 deletions(-)diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c index 69333728d871b..9f1c3cdc9d653 100644 --- a/fs/dlm/lowcomms.c +++ b/fs/dlm/lowcomms.c@@ -914,7 +914,6 @@ static int sctp_bind_addrs(struct connection *con, uint16_t port) static void sctp_connect_to_sock(struct connection *con) { struct sockaddr_storage daddr; - int one = 1; int result; int addr_len; struct socket *sock;@@ -961,8 +960,7 @@ static void sctp_connect_to_sock(struct connection *con) log_print("connecting to %d", con->nodeid); /* Turn off Nagle's algorithm */ - kernel_setsockopt(sock, SOL_SCTP, SCTP_NODELAY, (char *)&one, - sizeof(one)); + sctp_sock_set_nodelay(sock->sk); /* * Make sock->ops->connect() function return in specified time,@@ -1176,7 +1174,6 @@ static int sctp_listen_for_all(void) struct socket *sock = NULL; int result = -EINVAL; struct connection *con = nodeid2con(0, GFP_NOFS); - int one = 1; if (!con) return -ENOMEM;@@ -1191,10 +1188,7 @@ static int sctp_listen_for_all(void) } sock_set_rcvbuf(sock->sk, NEEDED_RMEM); - result = kernel_setsockopt(sock, SOL_SCTP, SCTP_NODELAY, (char *)&one, - sizeof(one)); - if (result < 0) - log_print("Could not set SCTP NODELAY error %d\n", result); + sctp_sock_set_nodelay(sock->sk); write_lock_bh(&sock->sk->sk_callback_lock); /* Init con struct */diff --git a/include/net/sctp/sctp.h b/include/net/sctp/sctp.h index 3ab5c6bbb90bd..f8bcb75bb0448 100644 --- a/include/net/sctp/sctp.h +++ b/include/net/sctp/sctp.h@@ -615,4 +615,11 @@ static inline bool sctp_newsk_ready(const struct sock *sk) return sock_flag(sk, SOCK_DEAD) || sk->sk_socket; } +static inline void sctp_sock_set_nodelay(struct sock *sk) +{ + lock_sock(sk); + sctp_sk(sk)->nodelay = true; + release_sock(sk); +} +
The duplication with sctp_setsockopt_nodelay() is quite silly/bad. Also, why have the 'true' hardcoded? It's what dlm uses, yes, but the API could be a bit more complete than that. Like for the other patch, this one could build on David's patch, do the ternary check before calling sctp_setsockopt_nodelay and then move sctp_setsockopt_nodelay to the header, or something like that.