aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorLinus Torvalds2018-02-09 14:55:30 -0800
committerLinus Torvalds2018-02-09 14:55:30 -0800
commit82f0a41e1980318ea4cdae20cdce7b33cb9c8946 (patch)
tree633222659a5b180f546d63af36f2db533b86455d
parent858f45bff3b8be61d91e87ef90dddd68433cbffa (diff)
parent0afa6b4412988019db14c6bfb8c6cbdf120ca9ad (diff)
Merge tag 'nfs-for-4.16-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Pull more NFS client updates from Trond Myklebust: "A few bugfixes and some small sunrpc latency/performance improvements before the merge window closes: Stable fixes: - fix an incorrect calculation of the RDMA send scatter gather element limit - fix an Oops when attempting to free resources after RDMA device removal Bugfixes: - SUNRPC: Ensure we always release the TCP socket in a timely fashion when the connection is shut down. - SUNRPC: Don't call __UDPX_INC_STATS() from a preemptible context Latency/Performance: - SUNRPC: Queue latency sensitive socket tasks to the less contended xprtiod queue - SUNRPC: Make the xprtiod workqueue unbounded. - SUNRPC: Make the rpciod workqueue unbounded" * tag 'nfs-for-4.16-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: SUNRPC: Don't call __UDPX_INC_STATS() from a preemptible context fix parallelism for rpc tasks Make the xprtiod workqueue unbounded. SUNRPC: Queue latency-sensitive socket tasks to xprtiod SUNRPC: Ensure we always close the socket after a connection shuts down xprtrdma: Fix BUG after a device removal xprtrdma: Fix calculation of ri_max_send_sges
-rw-r--r--include/linux/sunrpc/sched.h3
-rw-r--r--net/sunrpc/sched.c16
-rw-r--r--net/sunrpc/xprt.c3
-rw-r--r--net/sunrpc/xprtrdma/rpc_rdma.c2
-rw-r--r--net/sunrpc/xprtrdma/verbs.c8
-rw-r--r--net/sunrpc/xprtsock.c27
6 files changed, 36 insertions, 23 deletions
diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h
index d96e74e114c0..592653becd91 100644
--- a/include/linux/sunrpc/sched.h
+++ b/include/linux/sunrpc/sched.h
@@ -229,6 +229,9 @@ void rpc_sleep_on_priority(struct rpc_wait_queue *,
struct rpc_task *,
rpc_action action,
int priority);
+void rpc_wake_up_queued_task_on_wq(struct workqueue_struct *wq,
+ struct rpc_wait_queue *queue,
+ struct rpc_task *task);
void rpc_wake_up_queued_task(struct rpc_wait_queue *,
struct rpc_task *);
void rpc_wake_up(struct rpc_wait_queue *);
diff --git a/net/sunrpc/sched.c b/net/sunrpc/sched.c
index 896691afbb1a..d9db2eab3a8d 100644
--- a/net/sunrpc/sched.c
+++ b/net/sunrpc/sched.c
@@ -461,6 +461,18 @@ static void rpc_wake_up_task_queue_locked(struct rpc_wait_queue *queue, struct r
/*
* Wake up a task on a specific queue
*/
+void rpc_wake_up_queued_task_on_wq(struct workqueue_struct *wq,
+ struct rpc_wait_queue *queue,
+ struct rpc_task *task)
+{
+ spin_lock_bh(&queue->lock);
+ rpc_wake_up_task_on_wq_queue_locked(wq, queue, task);
+ spin_unlock_bh(&queue->lock);
+}
+
+/*
+ * Wake up a task on a specific queue
+ */
void rpc_wake_up_queued_task(struct rpc_wait_queue *queue, struct rpc_task *task)
{
spin_lock_bh(&queue->lock);
@@ -1092,12 +1104,12 @@ static int rpciod_start(void)
* Create the rpciod thread and wait for it to start.
*/
dprintk("RPC: creating workqueue rpciod\n");
- wq = alloc_workqueue("rpciod", WQ_MEM_RECLAIM, 0);
+ wq = alloc_workqueue("rpciod", WQ_MEM_RECLAIM | WQ_UNBOUND, 0);
if (!wq)
goto out_failed;
rpciod_workqueue = wq;
/* Note: highpri because network receive is latency sensitive */
- wq = alloc_workqueue("xprtiod", WQ_MEM_RECLAIM | WQ_HIGHPRI, 0);
+ wq = alloc_workqueue("xprtiod", WQ_UNBOUND|WQ_MEM_RECLAIM|WQ_HIGHPRI, 0);
if (!wq)
goto free_rpciod;
xprtiod_workqueue = wq;
diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c
index 2436fd1125fc..8f0ad4f268da 100644
--- a/net/sunrpc/xprt.c
+++ b/net/sunrpc/xprt.c
@@ -517,7 +517,8 @@ void xprt_write_space(struct rpc_xprt *xprt)
if (xprt->snd_task) {
dprintk("RPC: write space: waking waiting task on "
"xprt %p\n", xprt);
- rpc_wake_up_queued_task(&xprt->pending, xprt->snd_task);
+ rpc_wake_up_queued_task_on_wq(xprtiod_workqueue,
+ &xprt->pending, xprt->snd_task);
}
spin_unlock_bh(&xprt->transport_lock);
}
diff --git a/net/sunrpc/xprtrdma/rpc_rdma.c b/net/sunrpc/xprtrdma/rpc_rdma.c
index 162e5dd82466..f0855a959a27 100644
--- a/net/sunrpc/xprtrdma/rpc_rdma.c
+++ b/net/sunrpc/xprtrdma/rpc_rdma.c
@@ -143,7 +143,7 @@ static bool rpcrdma_args_inline(struct rpcrdma_xprt *r_xprt,
if (xdr->page_len) {
remaining = xdr->page_len;
offset = offset_in_page(xdr->page_base);
- count = 0;
+ count = RPCRDMA_MIN_SEND_SGES;
while (remaining) {
remaining -= min_t(unsigned int,
PAGE_SIZE - offset, remaining);
diff --git a/net/sunrpc/xprtrdma/verbs.c b/net/sunrpc/xprtrdma/verbs.c
index f4eb63e8e689..e6f84a6434a0 100644
--- a/net/sunrpc/xprtrdma/verbs.c
+++ b/net/sunrpc/xprtrdma/verbs.c
@@ -505,7 +505,7 @@ rpcrdma_ep_create(struct rpcrdma_ep *ep, struct rpcrdma_ia *ia,
pr_warn("rpcrdma: HCA provides only %d send SGEs\n", max_sge);
return -ENOMEM;
}
- ia->ri_max_send_sges = max_sge - RPCRDMA_MIN_SEND_SGES;
+ ia->ri_max_send_sges = max_sge;
if (ia->ri_device->attrs.max_qp_wr <= RPCRDMA_BACKWARD_WRS) {
dprintk("RPC: %s: insufficient wqe's available\n",
@@ -1502,6 +1502,9 @@ __rpcrdma_dma_map_regbuf(struct rpcrdma_ia *ia, struct rpcrdma_regbuf *rb)
static void
rpcrdma_dma_unmap_regbuf(struct rpcrdma_regbuf *rb)
{
+ if (!rb)
+ return;
+
if (!rpcrdma_regbuf_is_mapped(rb))
return;
@@ -1517,9 +1520,6 @@ rpcrdma_dma_unmap_regbuf(struct rpcrdma_regbuf *rb)
void
rpcrdma_free_regbuf(struct rpcrdma_regbuf *rb)
{
- if (!rb)
- return;
-
rpcrdma_dma_unmap_regbuf(rb);
kfree(rb);
}
diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c
index 18803021f242..a6b8c1f8f92a 100644
--- a/net/sunrpc/xprtsock.c
+++ b/net/sunrpc/xprtsock.c
@@ -807,13 +807,6 @@ static void xs_sock_reset_connection_flags(struct rpc_xprt *xprt)
smp_mb__after_atomic();
}
-static void xs_sock_mark_closed(struct rpc_xprt *xprt)
-{
- xs_sock_reset_connection_flags(xprt);
- /* Mark transport as closed and wake up all pending tasks */
- xprt_disconnect_done(xprt);
-}
-
/**
* xs_error_report - callback to handle TCP socket state errors
* @sk: socket
@@ -833,9 +826,6 @@ static void xs_error_report(struct sock *sk)
err = -sk->sk_err;
if (err == 0)
goto out;
- /* Is this a reset event? */
- if (sk->sk_state == TCP_CLOSE)
- xs_sock_mark_closed(xprt);
dprintk("RPC: xs_error_report client %p, error=%d...\n",
xprt, -err);
trace_rpc_socket_error(xprt, sk->sk_socket, err);
@@ -1078,18 +1068,18 @@ static void xs_udp_data_read_skb(struct rpc_xprt *xprt,
/* Suck it into the iovec, verify checksum if not done by hw. */
if (csum_partial_copy_to_xdr(&rovr->rq_private_buf, skb)) {
- __UDPX_INC_STATS(sk, UDP_MIB_INERRORS);
spin_lock(&xprt->recv_lock);
+ __UDPX_INC_STATS(sk, UDP_MIB_INERRORS);
goto out_unpin;
}
- __UDPX_INC_STATS(sk, UDP_MIB_INDATAGRAMS);
spin_lock_bh(&xprt->transport_lock);
xprt_adjust_cwnd(xprt, task, copied);
spin_unlock_bh(&xprt->transport_lock);
spin_lock(&xprt->recv_lock);
xprt_complete_rqst(task, copied);
+ __UDPX_INC_STATS(sk, UDP_MIB_INDATAGRAMS);
out_unpin:
xprt_unpin_rqst(rovr);
out_unlock:
@@ -1655,9 +1645,11 @@ static void xs_tcp_state_change(struct sock *sk)
if (test_and_clear_bit(XPRT_SOCK_CONNECTING,
&transport->sock_state))
xprt_clear_connecting(xprt);
+ clear_bit(XPRT_CLOSING, &xprt->state);
if (sk->sk_err)
xprt_wake_pending_tasks(xprt, -sk->sk_err);
- xs_sock_mark_closed(xprt);
+ /* Trigger the socket release */
+ xs_tcp_force_close(xprt);
}
out:
read_unlock_bh(&sk->sk_callback_lock);
@@ -2265,14 +2257,19 @@ static void xs_tcp_shutdown(struct rpc_xprt *xprt)
{
struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
struct socket *sock = transport->sock;
+ int skst = transport->inet ? transport->inet->sk_state : TCP_CLOSE;
if (sock == NULL)
return;
- if (xprt_connected(xprt)) {
+ switch (skst) {
+ default:
kernel_sock_shutdown(sock, SHUT_RDWR);
trace_rpc_socket_shutdown(xprt, sock);
- } else
+ break;
+ case TCP_CLOSE:
+ case TCP_TIME_WAIT:
xs_reset_transport(transport);
+ }
}
static void xs_tcp_set_socket_timeouts(struct rpc_xprt *xprt,