From bd736ed3e2d1088d9b4050f727342e1e619c3841 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 15 Aug 2019 17:26:17 -0400 Subject: SUNRPC: Don't handle errors if the bind/connect succeeded Don't handle errors in call_bind_status()/call_connect_status() if it turns out that a previous call caused it to succeed. Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org # v5.1+ --- net/sunrpc/clnt.c | 35 ++++++++++++++++++++++++----------- 1 file changed, 24 insertions(+), 11 deletions(-) (limited to 'net/sunrpc') diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index d8679b6027e9..3b731411d8e8 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -1970,6 +1970,7 @@ call_bind(struct rpc_task *task) static void call_bind_status(struct rpc_task *task) { + struct rpc_xprt *xprt = task->tk_rqstp->rq_xprt; int status = -EIO; if (rpc_task_transmitted(task)) { @@ -1977,14 +1978,15 @@ call_bind_status(struct rpc_task *task) return; } - if (task->tk_status >= 0) { - dprint_status(task); + dprint_status(task); + trace_rpc_bind_status(task); + if (task->tk_status >= 0) + goto out_next; + if (xprt_bound(xprt)) { task->tk_status = 0; - task->tk_action = call_connect; - return; + goto out_next; } - trace_rpc_bind_status(task); switch (task->tk_status) { case -ENOMEM: dprintk("RPC: %5u rpcbind out of memory\n", task->tk_pid); @@ -2043,7 +2045,9 @@ call_bind_status(struct rpc_task *task) rpc_call_rpcerror(task, status); return; - +out_next: + task->tk_action = call_connect; + return; retry_timeout: task->tk_status = 0; task->tk_action = call_bind; @@ -2090,6 +2094,7 @@ call_connect(struct rpc_task *task) static void call_connect_status(struct rpc_task *task) { + struct rpc_xprt *xprt = task->tk_rqstp->rq_xprt; struct rpc_clnt *clnt = task->tk_client; int status = task->tk_status; @@ -2099,8 +2104,17 @@ call_connect_status(struct rpc_task *task) } dprint_status(task); - trace_rpc_connect_status(task); + + if (task->tk_status == 0) { + clnt->cl_stats->netreconn++; + goto out_next; + } + if (xprt_connected(xprt)) { + task->tk_status = 0; + goto out_next; + } + task->tk_status = 0; switch (status) { case -ECONNREFUSED: @@ -2131,13 +2145,12 @@ call_connect_status(struct rpc_task *task) case -EAGAIN: case -ETIMEDOUT: goto out_retry; - case 0: - clnt->cl_stats->netreconn++; - task->tk_action = call_transmit; - return; } rpc_call_rpcerror(task, status); return; +out_next: + task->tk_action = call_transmit; + return; out_retry: /* Check for timeouts before looping back to call_bind */ task->tk_action = call_bind; -- cgit v1.2.3 From 80f455da6cd0998a5be30a8af24ea2a22815c212 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Thu, 15 Aug 2019 22:55:19 -0400 Subject: SUNRPC: Handle EADDRINUSE and ENOBUFS correctly If a connect or bind attempt returns EADDRINUSE, that means we want to retry with a different port. It is not a fatal connection error. Similarly, ENOBUFS is not fatal, but just indicates a memory allocation issue. Retry after a short delay. Signed-off-by: Trond Myklebust --- net/sunrpc/clnt.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'net/sunrpc') diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 3b731411d8e8..8a25440b771c 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -2005,6 +2005,9 @@ call_bind_status(struct rpc_task *task) task->tk_rebind_retry--; rpc_delay(task, 3*HZ); goto retry_timeout; + case -ENOBUFS: + rpc_delay(task, HZ >> 2); + goto retry_timeout; case -EAGAIN: goto retry_timeout; case -ETIMEDOUT: @@ -2028,7 +2031,6 @@ call_bind_status(struct rpc_task *task) case -ENETDOWN: case -EHOSTUNREACH: case -ENETUNREACH: - case -ENOBUFS: case -EPIPE: dprintk("RPC: %5u remote rpcbind unreachable: %d\n", task->tk_pid, task->tk_status); @@ -2131,8 +2133,6 @@ call_connect_status(struct rpc_task *task) case -ENETDOWN: case -ENETUNREACH: case -EHOSTUNREACH: - case -EADDRINUSE: - case -ENOBUFS: case -EPIPE: xprt_conditional_disconnect(task->tk_rqstp->rq_xprt, task->tk_rqstp->rq_connect_cookie); @@ -2141,10 +2141,14 @@ call_connect_status(struct rpc_task *task) /* retry with existing socket, after a delay */ rpc_delay(task, 3*HZ); /* fall through */ + case -EADDRINUSE: case -ENOTCONN: case -EAGAIN: case -ETIMEDOUT: goto out_retry; + case -ENOBUFS: + rpc_delay(task, HZ >> 2); + goto out_retry; } rpc_call_rpcerror(task, status); return; -- cgit v1.2.3 From d5711920ec6e578f51db95caa6f185f5090b865e Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 16 Aug 2019 08:37:26 -0400 Subject: Revert "NFSv4/flexfiles: Abort I/O early if the layout segment was invalidated" This reverts commit a79f194aa4879e9baad118c3f8bb2ca24dbef765. The mechanism for aborting I/O is racy, since we are not guaranteed that the request is asleep while we're changing both task->tk_status and task->tk_action. Signed-off-by: Trond Myklebust Cc: stable@vger.kernel.org # v5.1 --- fs/nfs/flexfilelayout/flexfilelayout.c | 17 ----------------- include/linux/sunrpc/sched.h | 1 - net/sunrpc/xprt.c | 7 ------- 3 files changed, 25 deletions(-) (limited to 'net/sunrpc') diff --git a/fs/nfs/flexfilelayout/flexfilelayout.c b/fs/nfs/flexfilelayout/flexfilelayout.c index b04e20d28162..2c7e1eca1ed7 100644 --- a/fs/nfs/flexfilelayout/flexfilelayout.c +++ b/fs/nfs/flexfilelayout/flexfilelayout.c @@ -1148,8 +1148,6 @@ static int ff_layout_async_handle_error_v4(struct rpc_task *task, break; case -NFS4ERR_RETRY_UNCACHED_REP: break; - case -EAGAIN: - return -NFS4ERR_RESET_TO_PNFS; /* Invalidate Layout errors */ case -NFS4ERR_PNFS_NO_LAYOUT: case -ESTALE: /* mapped NFS4ERR_STALE */ @@ -1210,7 +1208,6 @@ static int ff_layout_async_handle_error_v3(struct rpc_task *task, case -EBADHANDLE: case -ELOOP: case -ENOSPC: - case -EAGAIN: break; case -EJUKEBOX: nfs_inc_stats(lseg->pls_layout->plh_inode, NFSIOS_DELAY); @@ -1445,16 +1442,6 @@ static void ff_layout_read_prepare_v4(struct rpc_task *task, void *data) ff_layout_read_prepare_common(task, hdr); } -static void -ff_layout_io_prepare_transmit(struct rpc_task *task, - void *data) -{ - struct nfs_pgio_header *hdr = data; - - if (!pnfs_is_valid_lseg(hdr->lseg)) - rpc_exit(task, -EAGAIN); -} - static void ff_layout_read_call_done(struct rpc_task *task, void *data) { struct nfs_pgio_header *hdr = data; @@ -1740,7 +1727,6 @@ static void ff_layout_commit_release(void *data) static const struct rpc_call_ops ff_layout_read_call_ops_v3 = { .rpc_call_prepare = ff_layout_read_prepare_v3, - .rpc_call_prepare_transmit = ff_layout_io_prepare_transmit, .rpc_call_done = ff_layout_read_call_done, .rpc_count_stats = ff_layout_read_count_stats, .rpc_release = ff_layout_read_release, @@ -1748,7 +1734,6 @@ static const struct rpc_call_ops ff_layout_read_call_ops_v3 = { static const struct rpc_call_ops ff_layout_read_call_ops_v4 = { .rpc_call_prepare = ff_layout_read_prepare_v4, - .rpc_call_prepare_transmit = ff_layout_io_prepare_transmit, .rpc_call_done = ff_layout_read_call_done, .rpc_count_stats = ff_layout_read_count_stats, .rpc_release = ff_layout_read_release, @@ -1756,7 +1741,6 @@ static const struct rpc_call_ops ff_layout_read_call_ops_v4 = { static const struct rpc_call_ops ff_layout_write_call_ops_v3 = { .rpc_call_prepare = ff_layout_write_prepare_v3, - .rpc_call_prepare_transmit = ff_layout_io_prepare_transmit, .rpc_call_done = ff_layout_write_call_done, .rpc_count_stats = ff_layout_write_count_stats, .rpc_release = ff_layout_write_release, @@ -1764,7 +1748,6 @@ static const struct rpc_call_ops ff_layout_write_call_ops_v3 = { static const struct rpc_call_ops ff_layout_write_call_ops_v4 = { .rpc_call_prepare = ff_layout_write_prepare_v4, - .rpc_call_prepare_transmit = ff_layout_io_prepare_transmit, .rpc_call_done = ff_layout_write_call_done, .rpc_count_stats = ff_layout_write_count_stats, .rpc_release = ff_layout_write_release, diff --git a/include/linux/sunrpc/sched.h b/include/linux/sunrpc/sched.h index baa3ecdb882f..27536b961552 100644 --- a/include/linux/sunrpc/sched.h +++ b/include/linux/sunrpc/sched.h @@ -98,7 +98,6 @@ typedef void (*rpc_action)(struct rpc_task *); struct rpc_call_ops { void (*rpc_call_prepare)(struct rpc_task *, void *); - void (*rpc_call_prepare_transmit)(struct rpc_task *, void *); void (*rpc_call_done)(struct rpc_task *, void *); void (*rpc_count_stats)(struct rpc_task *, void *); void (*rpc_release)(void *); diff --git a/net/sunrpc/xprt.c b/net/sunrpc/xprt.c index 783748dc5e6f..2e71f5455c6c 100644 --- a/net/sunrpc/xprt.c +++ b/net/sunrpc/xprt.c @@ -1408,13 +1408,6 @@ xprt_request_transmit(struct rpc_rqst *req, struct rpc_task *snd_task) status = -EBADMSG; goto out_dequeue; } - if (task->tk_ops->rpc_call_prepare_transmit) { - task->tk_ops->rpc_call_prepare_transmit(task, - task->tk_calldata); - status = task->tk_status; - if (status < 0) - goto out_dequeue; - } if (RPC_SIGNALLED(task)) { status = -ERESTARTSYS; goto out_dequeue; -- cgit v1.2.3 From c82e5472c9980e0e483f4b689044150eefaca408 Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Fri, 16 Aug 2019 08:58:48 -0400 Subject: SUNRPC: Handle connection breakages correctly in call_status() If the connection breaks while we're waiting for a reply from the server, then we want to immediately try to reconnect. Fixes: ec6017d90359 ("SUNRPC fix regression in umount of a secure mount") Signed-off-by: Trond Myklebust --- net/sunrpc/clnt.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/sunrpc') diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 8a25440b771c..a07b516e503a 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -2382,7 +2382,7 @@ call_status(struct rpc_task *task) case -ECONNABORTED: case -ENOTCONN: rpc_force_rebind(clnt); - /* fall through */ + break; case -EADDRINUSE: rpc_delay(task, 3*HZ); /* fall through */ -- cgit v1.2.3