--- zzzz-none-000/linux-3.10.107/net/sunrpc/xprt.c 2017-06-27 09:49:32.000000000 +0000 +++ scorpion-7490-727/linux-3.10.107/net/sunrpc/xprt.c 2021-02-04 17:41:59.000000000 +0000 @@ -49,13 +49,15 @@ #include <linux/sunrpc/metrics.h> #include <linux/sunrpc/bc_xprt.h> +#include <trace/events/sunrpc.h> + #include "sunrpc.h" /* * Local variables */ -#ifdef RPC_DEBUG +#if IS_ENABLED(CONFIG_SUNRPC_DEBUG) # define RPCDBG_FACILITY RPCDBG_XPRT #endif @@ -66,29 +68,12 @@ static void xprt_request_init(struct rpc_task *, struct rpc_xprt *); static void xprt_connect_status(struct rpc_task *task); static int __xprt_get_cong(struct rpc_xprt *, struct rpc_task *); +static void __xprt_put_cong(struct rpc_xprt *, struct rpc_rqst *); static void xprt_destroy(struct rpc_xprt *xprt); static DEFINE_SPINLOCK(xprt_list_lock); static LIST_HEAD(xprt_list); -/* - * The transport code maintains an estimate on the maximum number of out- - * standing RPC requests, using a smoothed version of the congestion - * avoidance implemented in 44BSD. This is basically the Van Jacobson - * congestion algorithm: If a retransmit occurs, the congestion window is - * halved; otherwise, it is incremented by 1/cwnd when - * - * - a reply is received and - * - a full number of requests are outstanding and - * - the congestion window hasn't been updated recently. - */ -#define RPC_CWNDSHIFT (8U) -#define RPC_CWNDSCALE (1U << RPC_CWNDSHIFT) -#define RPC_INITCWND RPC_CWNDSCALE -#define RPC_MAXCWND(xprt) ((xprt)->max_reqs << RPC_CWNDSHIFT) - -#define RPCXPRT_CONGESTED(xprt) ((xprt)->cong >= (xprt)->cwnd) - /** * xprt_register_transport - register a transport implementation * @transport: transport to register @@ -205,10 +190,8 @@ goto out_sleep; } xprt->snd_task = task; - if (req != NULL) { - req->rq_bytes_sent = 0; + if (req != NULL) req->rq_ntrans++; - } return 1; @@ -232,9 +215,9 @@ { xprt->snd_task = NULL; if (!test_bit(XPRT_CLOSE_WAIT, &xprt->state)) { - smp_mb__before_clear_bit(); + smp_mb__before_atomic(); clear_bit(XPRT_LOCKED, &xprt->state); - smp_mb__after_clear_bit(); + smp_mb__after_atomic(); } else queue_work(rpciod_workqueue, &xprt->task_cleanup); } @@ -263,12 +246,13 @@ } if (__xprt_get_cong(xprt, task)) { xprt->snd_task = task; - req->rq_bytes_sent = 0; req->rq_ntrans++; return 1; } xprt_clear_locked(xprt); out_sleep: + if (req) + __xprt_put_cong(xprt, req); dprintk("RPC: %5u failed to lock transport %p\n", task->tk_pid, xprt); task->tk_timeout = 0; task->tk_status = -EAGAIN; @@ -300,10 +284,8 @@ req = task->tk_rqstp; xprt->snd_task = task; - if (req) { - req->rq_bytes_sent = 0; + if (req) req->rq_ntrans++; - } return true; } @@ -329,7 +311,6 @@ } if (__xprt_get_cong(xprt, task)) { xprt->snd_task = task; - req->rq_bytes_sent = 0; req->rq_ntrans++; return true; } @@ -348,6 +329,15 @@ xprt_clear_locked(xprt); } +static void xprt_task_clear_bytes_sent(struct rpc_task *task) +{ + if (task != NULL) { + struct rpc_rqst *req = task->tk_rqstp; + if (req != NULL) + req->rq_bytes_sent = 0; + } +} + /** * xprt_release_xprt - allow other requests to use a transport * @xprt: transport with other tasks potentially waiting @@ -358,6 +348,7 @@ void xprt_release_xprt(struct rpc_xprt *xprt, struct rpc_task *task) { if (xprt->snd_task == task) { + xprt_task_clear_bytes_sent(task); xprt_clear_locked(xprt); __xprt_lock_write_next(xprt); } @@ -375,6 +366,7 @@ void xprt_release_xprt_cong(struct rpc_xprt *xprt, struct rpc_task *task) { if (xprt->snd_task == task) { + xprt_task_clear_bytes_sent(task); xprt_clear_locked(xprt); __xprt_lock_write_next_cong(xprt); } @@ -442,7 +434,15 @@ * @task: recently completed RPC request used to adjust window * @result: result code of completed RPC request * - * We use a time-smoothed congestion estimator to avoid heavy oscillation. + * The transport code maintains an estimate on the maximum number of out- + * standing RPC requests, using a smoothed version of the congestion + * avoidance implemented in 44BSD. This is basically the Van Jacobson + * congestion algorithm: If a retransmit occurs, the congestion window is + * halved; otherwise, it is incremented by 1/cwnd when + * + * - a reply is received and + * - a full number of requests are outstanding and + * - the congestion window hasn't been updated recently. */ void xprt_adjust_cwnd(struct rpc_xprt *xprt, struct rpc_task *task, int result) { @@ -611,9 +611,10 @@ struct rpc_xprt *xprt = container_of(work, struct rpc_xprt, task_cleanup); - xprt->ops->close(xprt); clear_bit(XPRT_CLOSE_WAIT, &xprt->state); + xprt->ops->close(xprt); xprt_release_write(xprt, NULL); + wake_up_bit(&xprt->state, XPRT_LOCKED); } /** @@ -687,13 +688,45 @@ if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) goto out_abort; spin_unlock(&xprt->transport_lock); - set_bit(XPRT_CONNECTION_CLOSE, &xprt->state); queue_work(rpciod_workqueue, &xprt->task_cleanup); return; out_abort: spin_unlock(&xprt->transport_lock); } +bool xprt_lock_connect(struct rpc_xprt *xprt, + struct rpc_task *task, + void *cookie) +{ + bool ret = false; + + spin_lock_bh(&xprt->transport_lock); + if (!test_bit(XPRT_LOCKED, &xprt->state)) + goto out; + if (xprt->snd_task != task) + goto out; + xprt_task_clear_bytes_sent(task); + xprt->snd_task = cookie; + ret = true; +out: + spin_unlock_bh(&xprt->transport_lock); + return ret; +} + +void xprt_unlock_connect(struct rpc_xprt *xprt, void *cookie) +{ + spin_lock_bh(&xprt->transport_lock); + if (xprt->snd_task != cookie) + goto out; + if (!test_bit(XPRT_LOCKED, &xprt->state)) + goto out; + xprt->snd_task =NULL; + xprt->ops->release_xprt(xprt, NULL); +out: + spin_unlock_bh(&xprt->transport_lock); + wake_up_bit(&xprt->state, XPRT_LOCKED); +} + /** * xprt_connect - schedule a transport connect operation * @task: RPC task that is requesting the connect @@ -716,9 +749,7 @@ if (test_and_clear_bit(XPRT_CLOSE_WAIT, &xprt->state)) xprt->ops->close(xprt); - if (xprt_connected(xprt)) - xprt_release_write(xprt, task); - else { + if (!xprt_connected(xprt)) { task->tk_rqstp->rq_bytes_sent = 0; task->tk_timeout = task->tk_rqstp->rq_timeout; rpc_sleep_on(&xprt->pending, task, xprt_connect_status); @@ -730,6 +761,7 @@ xprt->stat.connect_start = jiffies; xprt->ops->connect(xprt, task); } + xprt_release_write(xprt, task); } static void xprt_connect_status(struct rpc_task *task) @@ -745,6 +777,12 @@ } switch (task->tk_status) { + case -ECONNREFUSED: + case -ECONNRESET: + case -ECONNABORTED: + case -ENETUNREACH: + case -EHOSTUNREACH: + case -EPIPE: case -EAGAIN: dprintk("RPC: %5u xprt_connect_status: retrying\n", task->tk_pid); break; @@ -756,7 +794,6 @@ dprintk("RPC: %5u xprt_connect_status: error %d connecting to " "server %s\n", task->tk_pid, -task->tk_status, xprt->servername); - xprt_release_write(xprt, task); task->tk_status = -EIO; } } @@ -772,11 +809,14 @@ struct rpc_rqst *entry; list_for_each_entry(entry, &xprt->recv, rq_list) - if (entry->rq_xid == xid) + if (entry->rq_xid == xid) { + trace_xprt_lookup_rqst(xprt, xid, 0); return entry; + } dprintk("RPC: xprt_lookup_rqst did not find xid %08x\n", ntohl(xid)); + trace_xprt_lookup_rqst(xprt, xid, -ENOENT); xprt->stat.bad_xids++; return NULL; } @@ -810,6 +850,7 @@ dprintk("RPC: %5u xid %08x complete (%d bytes received)\n", task->tk_pid, ntohl(req->rq_xid), copied); + trace_xprt_complete_rqst(xprt, req->rq_xid, copied); xprt->stat.recvs++; req->rq_rtt = ktime_sub(ktime_get(), req->rq_xtime); @@ -854,24 +895,36 @@ * @task: RPC task about to send a request * */ -int xprt_prepare_transmit(struct rpc_task *task) +bool xprt_prepare_transmit(struct rpc_task *task) { struct rpc_rqst *req = task->tk_rqstp; struct rpc_xprt *xprt = req->rq_xprt; - int err = 0; + bool ret = false; dprintk("RPC: %5u xprt_prepare_transmit\n", task->tk_pid); spin_lock_bh(&xprt->transport_lock); - if (req->rq_reply_bytes_recvd && !req->rq_bytes_sent) { - err = req->rq_reply_bytes_recvd; + if (!req->rq_bytes_sent) { + if (req->rq_reply_bytes_recvd) { + task->tk_status = req->rq_reply_bytes_recvd; + goto out_unlock; + } + if ((task->tk_flags & RPC_TASK_NO_RETRANS_TIMEOUT) + && xprt_connected(xprt) + && req->rq_connect_cookie == xprt->connect_cookie) { + xprt->ops->set_retrans_timeout(task); + rpc_sleep_on(&xprt->pending, task, xprt_timer); + goto out_unlock; + } + } + if (!xprt->ops->reserve_xprt(xprt, task)) { + task->tk_status = -EAGAIN; goto out_unlock; } - if (!xprt->ops->reserve_xprt(xprt, task)) - err = -EAGAIN; + ret = true; out_unlock: spin_unlock_bh(&xprt->transport_lock); - return err; + return ret; } void xprt_end_transmit(struct rpc_task *task) @@ -912,13 +965,14 @@ } else if (!req->rq_bytes_sent) return; - req->rq_connect_cookie = xprt->connect_cookie; req->rq_xtime = ktime_get(); status = xprt->ops->send_request(task); + trace_xprt_transmit(xprt, req->rq_xid, status); if (status != 0) { task->tk_status = status; return; } + xprt_inject_disconnect(xprt); dprintk("RPC: %5u xmit complete\n", task->tk_pid); task->tk_flags |= RPC_TASK_SENT; @@ -938,12 +992,14 @@ /* Don't race with disconnect */ if (!xprt_connected(xprt)) task->tk_status = -ENOTCONN; - else if (!req->rq_reply_bytes_recvd && rpc_reply_expected(task)) { + else { /* * Sleep on the pending queue since * we're expecting a reply. */ - rpc_sleep_on(&xprt->pending, task, xprt_timer); + if (!req->rq_reply_bytes_recvd && rpc_reply_expected(task)) + rpc_sleep_on(&xprt->pending, task, xprt_timer); + req->rq_connect_cookie = xprt->connect_cookie; } spin_unlock_bh(&xprt->transport_lock); } @@ -1087,11 +1143,9 @@ for (i = 0; i < num_prealloc; i++) { req = kzalloc(sizeof(struct rpc_rqst), GFP_KERNEL); if (!req) - break; + goto out_free; list_add(&req->rq_list, &xprt->free); } - if (i < num_prealloc) - goto out_free; if (max_alloc > num_prealloc) xprt->max_reqs = max_alloc; else @@ -1173,7 +1227,7 @@ static inline void xprt_init_xid(struct rpc_xprt *xprt) { - xprt->xid = net_random(); + xprt->xid = prandom_u32(); } static void xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt) @@ -1186,6 +1240,12 @@ req->rq_xprt = xprt; req->rq_buffer = NULL; req->rq_xid = xprt_alloc_xid(xprt); + req->rq_connect_cookie = xprt->connect_cookie - 1; + req->rq_bytes_sent = 0; + req->rq_snd_buf.len = 0; + req->rq_snd_buf.buflen = 0; + req->rq_rcv_buf.len = 0; + req->rq_rcv_buf.buflen = 0; req->rq_release_snd_buf = NULL; xprt_reset_majortimeo(req); dprintk("RPC: %5u reserved req %p xid %08x\n", task->tk_pid, @@ -1231,6 +1291,7 @@ spin_unlock_bh(&xprt->transport_lock); if (req->rq_buffer) xprt->ops->buf_free(req->rq_buffer); + xprt_inject_disconnect(xprt); if (req->rq_cred != NULL) put_rpccred(req->rq_cred); task->tk_rqstp = NULL; @@ -1319,6 +1380,8 @@ return ERR_PTR(-ENOMEM); } + rpc_xprt_debugfs_register(xprt); + dprintk("RPC: created transport %p with %u slots\n", xprt, xprt->max_reqs); out: @@ -1333,8 +1396,13 @@ static void xprt_destroy(struct rpc_xprt *xprt) { dprintk("RPC: destroying transport %p\n", xprt); + + /* Exclude transport connect/disconnect handlers */ + wait_on_bit_lock(&xprt->state, XPRT_LOCKED, TASK_UNINTERRUPTIBLE); + del_timer_sync(&xprt->timer); + rpc_xprt_debugfs_unregister(xprt); rpc_destroy_wait_queue(&xprt->binding); rpc_destroy_wait_queue(&xprt->pending); rpc_destroy_wait_queue(&xprt->sending); @@ -1357,15 +1425,3 @@ if (atomic_dec_and_test(&xprt->count)) xprt_destroy(xprt); } - -/** - * xprt_get - return a reference to an RPC transport. - * @xprt: pointer to the transport - * - */ -struct rpc_xprt *xprt_get(struct rpc_xprt *xprt) -{ - if (atomic_inc_not_zero(&xprt->count)) - return xprt; - return NULL; -}