From mboxrd@z Thu Jan 1 00:00:00 1970 From: Christoph Hellwig Date: Wed, 13 May 2020 08:26:23 +0200 Subject: [Ocfs2-devel] [PATCH 08/33] net: add sock_set_rcvbuf In-Reply-To: <20200513062649.2100053-1-hch@lst.de> References: <20200513062649.2100053-1-hch@lst.de> Message-ID: <20200513062649.2100053-9-hch@lst.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: "David S. Miller" , Jakub Kicinski Cc: Eric Dumazet , Alexey Kuznetsov , Hideaki YOSHIFUJI , Vlad Yasevich , Neil Horman , Marcelo Ricardo Leitner , Jon Maloy , Ying Xue , drbd-dev@lists.linbit.com, linux-block@vger.kernel.org, linux-kernel@vger.kernel.org, linux-rdma@vger.kernel.org, linux-nvme@lists.infradead.org, target-devel@vger.kernel.org, linux-afs@lists.infradead.org, linux-cifs@vger.kernel.org, cluster-devel@redhat.com, ocfs2-devel@oss.oracle.com, netdev@vger.kernel.org, linux-sctp@vger.kernel.org, ceph-devel@vger.kernel.org, rds-devel@oss.oracle.com, linux-nfs@vger.kernel.org Add a helper to directly set the SO_RCVBUFFORCE sockopt from kernel space without going through a fake uaccess. Signed-off-by: Christoph Hellwig --- fs/dlm/lowcomms.c | 7 +----- include/net/sock.h | 1 + net/core/sock.c | 59 +++++++++++++++++++++++++--------------------- 3 files changed, 34 insertions(+), 33 deletions(-) diff --git a/fs/dlm/lowcomms.c b/fs/dlm/lowcomms.c index 16d616c180613..223c185ecd0c7 100644 --- a/fs/dlm/lowcomms.c +++ b/fs/dlm/lowcomms.c @@ -1297,7 +1297,6 @@ static int sctp_listen_for_all(void) struct socket *sock = NULL; int result = -EINVAL; struct connection *con = nodeid2con(0, GFP_NOFS); - int bufsize = NEEDED_RMEM; int one = 1; if (!con) @@ -1312,11 +1311,7 @@ static int sctp_listen_for_all(void) goto out; } - result = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUFFORCE, - (char *)&bufsize, sizeof(bufsize)); - if (result) - log_print("Error increasing buffer space on socket %d", result); - + sock_set_rcvbuf(sock->sk, NEEDED_RMEM); result = kernel_setsockopt(sock, SOL_SCTP, SCTP_NODELAY, (char *)&one, sizeof(one)); if (result < 0) diff --git a/include/net/sock.h b/include/net/sock.h index 4cedde585424f..e1ed40ff01312 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -2692,6 +2692,7 @@ void sock_set_linger(struct sock *sk, bool onoff, unsigned int linger); void sock_set_priority(struct sock *sk, u32 priority); void sock_set_sndtimeo(struct sock *sk, unsigned int secs); void sock_set_keepalive(struct sock *sk, bool keepalive); +void sock_set_rcvbuf(struct sock *sk, int val); int sock_bindtoindex(struct sock *sk, int ifindex); void sock_set_timestamps(struct sock *sk, bool val, bool new, bool ns); diff --git a/net/core/sock.c b/net/core/sock.c index dfd2b839f88bb..6af01b757cf24 100644 --- a/net/core/sock.c +++ b/net/core/sock.c @@ -804,6 +804,35 @@ void sock_set_keepalive(struct sock *sk, bool keepalive) } EXPORT_SYMBOL(sock_set_keepalive); +void __sock_set_rcvbuf(struct sock *sk, int val) +{ + /* Ensure val * 2 fits into an int, to prevent max_t() from treating it + * as a negative value. + */ + val = min_t(int, val, INT_MAX / 2); + sk->sk_userlocks |= SOCK_RCVBUF_LOCK; + + /* We double it on the way in to account for "struct sk_buff" etc. + * overhead. Applications assume that the SO_RCVBUF setting they make + * will allow that much actual data to be received on that socket. + * + * Applications are unaware that "struct sk_buff" and other overheads + * allocate from the receive buffer during socket buffer allocation. + * + * And after considering the possible alternatives, returning the value + * we actually used in getsockopt is the most desirable behavior. + */ + WRITE_ONCE(sk->sk_rcvbuf, max_t(int, val * 2, SOCK_MIN_RCVBUF)); +} + +void sock_set_rcvbuf(struct sock *sk, int val) +{ + lock_sock(sk); + __sock_set_rcvbuf(sk, val); + release_sock(sk); +} +EXPORT_SYMBOL(sock_set_rcvbuf); + /* * This is meant for all protocols to use and covers goings on * at the socket level. Everything here is generic. @@ -900,30 +929,7 @@ int sock_setsockopt(struct socket *sock, int level, int optname, * play 'guess the biggest size' games. RCVBUF/SNDBUF * are treated in BSD as hints */ - val = min_t(u32, val, sysctl_rmem_max); -set_rcvbuf: - /* Ensure val * 2 fits into an int, to prevent max_t() - * from treating it as a negative value. - */ - val = min_t(int, val, INT_MAX / 2); - sk->sk_userlocks |= SOCK_RCVBUF_LOCK; - /* - * We double it on the way in to account for - * "struct sk_buff" etc. overhead. Applications - * assume that the SO_RCVBUF setting they make will - * allow that much actual data to be received on that - * socket. - * - * Applications are unaware that "struct sk_buff" and - * other overheads allocate from the receive buffer - * during socket buffer allocation. - * - * And after considering the possible alternatives, - * returning the value we actually used in getsockopt - * is the most desirable behavior. - */ - WRITE_ONCE(sk->sk_rcvbuf, - max_t(int, val * 2, SOCK_MIN_RCVBUF)); + __sock_set_rcvbuf(sk, min_t(u32, val, sysctl_rmem_max)); break; case SO_RCVBUFFORCE: @@ -935,9 +941,8 @@ int sock_setsockopt(struct socket *sock, int level, int optname, /* No negative values (to prevent underflow, as val will be * multiplied by 2). */ - if (val < 0) - val = 0; - goto set_rcvbuf; + __sock_set_rcvbuf(sk, max(val, 0)); + break; case SO_KEEPALIVE: if (sk->sk_prot->keepalive) -- 2.26.2