blob: fe785ee819ddb195db524124c854decf6937d23b [file] [log] [blame]
Andy Grover5c115592009-02-24 15:30:27 +00001/*
Ka-Cheong Poona43cced2018-04-11 00:57:25 -07002 * Copyright (c) 2006, 2018 Oracle and/or its affiliates. All rights reserved.
Andy Grover5c115592009-02-24 15:30:27 +00003 *
4 * This software is available to you under a choice of one of two
5 * licenses. You may choose to be licensed under the terms of the GNU
6 * General Public License (GPL) Version 2, available from the file
7 * COPYING in the main directory of this source tree, or the
8 * OpenIB.org BSD license below:
9 *
10 * Redistribution and use in source and binary forms, with or
11 * without modification, are permitted provided that the following
12 * conditions are met:
13 *
14 * - Redistributions of source code must retain the above
15 * copyright notice, this list of conditions and the following
16 * disclaimer.
17 *
18 * - Redistributions in binary form must reproduce the above
19 * copyright notice, this list of conditions and the following
20 * disclaimer in the documentation and/or other materials
21 * provided with the distribution.
22 *
23 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
24 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
25 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
26 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
27 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
28 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
29 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
30 * SOFTWARE.
31 *
32 */
33#include <linux/kernel.h>
Paul Gortmakerd9b93842011-09-18 13:21:27 -040034#include <linux/moduleparam.h>
Tejun Heo5a0e3ad2010-03-24 17:04:11 +090035#include <linux/gfp.h>
Andy Grover5c115592009-02-24 15:30:27 +000036#include <net/sock.h>
37#include <linux/in.h>
38#include <linux/list.h>
Manuel Zerpiescb0a6052011-06-16 02:09:57 +000039#include <linux/ratelimit.h>
Paul Gortmakerbc3b2d72011-07-15 11:47:34 -040040#include <linux/export.h>
Santosh Shilimkar4bebdd72015-09-10 11:57:14 -070041#include <linux/sizes.h>
Andy Grover5c115592009-02-24 15:30:27 +000042
43#include "rds.h"
Andy Grover5c115592009-02-24 15:30:27 +000044
45/* When transmitting messages in rds_send_xmit, we need to emerge from
46 * time to time and briefly release the CPU. Otherwise the softlock watchdog
47 * will kick our shin.
48 * Also, it seems fairer to not let one busy connection stall all the
49 * others.
50 *
51 * send_batch_count is the number of times we'll loop in send_xmit. Setting
52 * it to 0 will restore the old behavior (where we looped until we had
53 * drained the queue).
54 */
Santosh Shilimkar4bebdd72015-09-10 11:57:14 -070055static int send_batch_count = SZ_1K;
Andy Grover5c115592009-02-24 15:30:27 +000056module_param(send_batch_count, int, 0444);
57MODULE_PARM_DESC(send_batch_count, " batch factor when working the send queue");
58
stephen hemmingerff51bf82010-10-19 08:08:33 +000059static void rds_send_remove_from_sock(struct list_head *messages, int status);
60
Andy Grover5c115592009-02-24 15:30:27 +000061/*
Zach Brown0f4b1c72010-06-04 14:41:41 -070062 * Reset the send state. Callers must ensure that this doesn't race with
63 * rds_send_xmit().
Andy Grover5c115592009-02-24 15:30:27 +000064 */
Sowmini Varadhand769ef82016-06-13 09:44:41 -070065void rds_send_path_reset(struct rds_conn_path *cp)
Andy Grover5c115592009-02-24 15:30:27 +000066{
67 struct rds_message *rm, *tmp;
68 unsigned long flags;
69
Sowmini Varadhan4e9b5512016-06-13 09:44:30 -070070 if (cp->cp_xmit_rm) {
71 rm = cp->cp_xmit_rm;
72 cp->cp_xmit_rm = NULL;
Andy Grover5c115592009-02-24 15:30:27 +000073 /* Tell the user the RDMA op is no longer mapped by the
74 * transport. This isn't entirely true (it's flushed out
75 * independently) but as the connection is down, there's
76 * no ongoing RDMA to/from that memory */
Chris Mason7e3f2952010-05-11 15:11:11 -070077 rds_message_unmapped(rm);
Chris Mason7e3f2952010-05-11 15:11:11 -070078 rds_message_put(rm);
Andy Grover5c115592009-02-24 15:30:27 +000079 }
Chris Mason7e3f2952010-05-11 15:11:11 -070080
Sowmini Varadhan4e9b5512016-06-13 09:44:30 -070081 cp->cp_xmit_sg = 0;
82 cp->cp_xmit_hdr_off = 0;
83 cp->cp_xmit_data_off = 0;
84 cp->cp_xmit_atomic_sent = 0;
85 cp->cp_xmit_rdma_sent = 0;
86 cp->cp_xmit_data_sent = 0;
Andy Grover5c115592009-02-24 15:30:27 +000087
Sowmini Varadhan4e9b5512016-06-13 09:44:30 -070088 cp->cp_conn->c_map_queued = 0;
Andy Grover5c115592009-02-24 15:30:27 +000089
Sowmini Varadhan4e9b5512016-06-13 09:44:30 -070090 cp->cp_unacked_packets = rds_sysctl_max_unacked_packets;
91 cp->cp_unacked_bytes = rds_sysctl_max_unacked_bytes;
Andy Grover5c115592009-02-24 15:30:27 +000092
93 /* Mark messages as retransmissions, and move them to the send q */
Sowmini Varadhan4e9b5512016-06-13 09:44:30 -070094 spin_lock_irqsave(&cp->cp_lock, flags);
95 list_for_each_entry_safe(rm, tmp, &cp->cp_retrans, m_conn_item) {
Andy Grover5c115592009-02-24 15:30:27 +000096 set_bit(RDS_MSG_ACK_REQUIRED, &rm->m_flags);
97 set_bit(RDS_MSG_RETRANSMITTED, &rm->m_flags);
98 }
Sowmini Varadhan4e9b5512016-06-13 09:44:30 -070099 list_splice_init(&cp->cp_retrans, &cp->cp_send_queue);
100 spin_unlock_irqrestore(&cp->cp_lock, flags);
101}
Sowmini Varadhand769ef82016-06-13 09:44:41 -0700102EXPORT_SYMBOL_GPL(rds_send_path_reset);
Andy Grover5c115592009-02-24 15:30:27 +0000103
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700104static int acquire_in_xmit(struct rds_conn_path *cp)
Zach Brown0f4b1c72010-06-04 14:41:41 -0700105{
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700106 return test_and_set_bit(RDS_IN_XMIT, &cp->cp_flags) == 0;
Zach Brown0f4b1c72010-06-04 14:41:41 -0700107}
108
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700109static void release_in_xmit(struct rds_conn_path *cp)
Zach Brown0f4b1c72010-06-04 14:41:41 -0700110{
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700111 clear_bit(RDS_IN_XMIT, &cp->cp_flags);
Peter Zijlstra4e857c52014-03-17 18:06:10 +0100112 smp_mb__after_atomic();
Zach Brown0f4b1c72010-06-04 14:41:41 -0700113 /*
114 * We don't use wait_on_bit()/wake_up_bit() because our waking is in a
115 * hot path and finding waiters is very rare. We don't want to walk
116 * the system-wide hashed waitqueue buckets in the fast path only to
117 * almost never find waiters.
118 */
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700119 if (waitqueue_active(&cp->cp_waitq))
120 wake_up_all(&cp->cp_waitq);
Zach Brown0f4b1c72010-06-04 14:41:41 -0700121}
122
Andy Grover5c115592009-02-24 15:30:27 +0000123/*
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300124 * We're making the conscious trade-off here to only send one message
Andy Grover5c115592009-02-24 15:30:27 +0000125 * down the connection at a time.
126 * Pro:
127 * - tx queueing is a simple fifo list
128 * - reassembly is optional and easily done by transports per conn
129 * - no per flow rx lookup at all, straight to the socket
130 * - less per-frag memory and wire overhead
131 * Con:
132 * - queued acks can be delayed behind large messages
133 * Depends:
134 * - small message latency is higher behind queued large messages
135 * - large message latency isn't starved by intervening small sends
136 */
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700137int rds_send_xmit(struct rds_conn_path *cp)
Andy Grover5c115592009-02-24 15:30:27 +0000138{
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700139 struct rds_connection *conn = cp->cp_conn;
Andy Grover5c115592009-02-24 15:30:27 +0000140 struct rds_message *rm;
141 unsigned long flags;
142 unsigned int tmp;
Andy Grover5c115592009-02-24 15:30:27 +0000143 struct scatterlist *sg;
144 int ret = 0;
Andy Grover5c115592009-02-24 15:30:27 +0000145 LIST_HEAD(to_be_dropped);
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400146 int batch_count;
147 unsigned long send_gen = 0;
Andy Grover5c115592009-02-24 15:30:27 +0000148
Andy Groverfcc54502010-03-29 17:08:49 -0700149restart:
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400150 batch_count = 0;
Andy Grover049ee3f2010-03-23 17:39:07 -0700151
Andy Grover5c115592009-02-24 15:30:27 +0000152 /*
153 * sendmsg calls here after having queued its message on the send
154 * queue. We only have one task feeding the connection at a time. If
155 * another thread is already feeding the queue then we back off. This
156 * avoids blocking the caller and trading per-connection data between
157 * caches per message.
Andy Grover5c115592009-02-24 15:30:27 +0000158 */
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700159 if (!acquire_in_xmit(cp)) {
Andy Grover049ee3f2010-03-23 17:39:07 -0700160 rds_stats_inc(s_send_lock_contention);
Andy Grover5c115592009-02-24 15:30:27 +0000161 ret = -ENOMEM;
162 goto out;
163 }
Zach Brown0f4b1c72010-06-04 14:41:41 -0700164
Sowmini Varadhanebeeb1a2018-02-03 04:26:51 -0800165 if (rds_destroy_pending(cp->cp_conn)) {
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -0800166 release_in_xmit(cp);
167 ret = -ENETUNREACH; /* dont requeue send work */
168 goto out;
169 }
170
Zach Brown0f4b1c72010-06-04 14:41:41 -0700171 /*
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400172 * we record the send generation after doing the xmit acquire.
173 * if someone else manages to jump in and do some work, we'll use
174 * this to avoid a goto restart farther down.
175 *
176 * The acquire_in_xmit() check above ensures that only one
177 * caller can increment c_send_gen at any time.
178 */
HÃ¥kon Buggee623a482017-07-20 12:28:55 +0200179 send_gen = READ_ONCE(cp->cp_send_gen) + 1;
180 WRITE_ONCE(cp->cp_send_gen, send_gen);
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400181
182 /*
Zach Brown0f4b1c72010-06-04 14:41:41 -0700183 * rds_conn_shutdown() sets the conn state and then tests RDS_IN_XMIT,
184 * we do the opposite to avoid races.
185 */
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700186 if (!rds_conn_path_up(cp)) {
187 release_in_xmit(cp);
Zach Brown0f4b1c72010-06-04 14:41:41 -0700188 ret = 0;
189 goto out;
190 }
Andy Grover5c115592009-02-24 15:30:27 +0000191
Sowmini Varadhan226f7a72016-06-30 16:11:10 -0700192 if (conn->c_trans->xmit_path_prepare)
193 conn->c_trans->xmit_path_prepare(cp);
Andy Grover5c115592009-02-24 15:30:27 +0000194
195 /*
196 * spin trying to push headers and data down the connection until
Andy Grover5b2366b2010-02-03 19:36:44 -0800197 * the connection doesn't make forward progress.
Andy Grover5c115592009-02-24 15:30:27 +0000198 */
Andy Groverfcc54502010-03-29 17:08:49 -0700199 while (1) {
Andy Grover5c115592009-02-24 15:30:27 +0000200
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700201 rm = cp->cp_xmit_rm;
Andy Grover5c115592009-02-24 15:30:27 +0000202
Andy Grover5b2366b2010-02-03 19:36:44 -0800203 /*
204 * If between sending messages, we can send a pending congestion
205 * map update.
Andy Grover5c115592009-02-24 15:30:27 +0000206 */
Andy Grover8690bfa2010-01-12 11:56:44 -0800207 if (!rm && test_and_clear_bit(0, &conn->c_map_queued)) {
Andy Grover77dd5502010-03-22 15:22:04 -0700208 rm = rds_cong_update_alloc(conn);
209 if (IS_ERR(rm)) {
210 ret = PTR_ERR(rm);
211 break;
Andy Grover5b2366b2010-02-03 19:36:44 -0800212 }
Andy Grover77dd5502010-03-22 15:22:04 -0700213 rm->data.op_active = 1;
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700214 rm->m_inc.i_conn_path = cp;
215 rm->m_inc.i_conn = cp->cp_conn;
Andy Grover77dd5502010-03-22 15:22:04 -0700216
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700217 cp->cp_xmit_rm = rm;
Andy Grover5c115592009-02-24 15:30:27 +0000218 }
219
220 /*
Andy Grover5b2366b2010-02-03 19:36:44 -0800221 * If not already working on one, grab the next message.
Andy Grover5c115592009-02-24 15:30:27 +0000222 *
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700223 * cp_xmit_rm holds a ref while we're sending this message down
Andy Grover5c115592009-02-24 15:30:27 +0000224 * the connction. We can use this ref while holding the
225 * send_sem.. rds_send_reset() is serialized with it.
226 */
Andy Grover8690bfa2010-01-12 11:56:44 -0800227 if (!rm) {
Andy Grover5c115592009-02-24 15:30:27 +0000228 unsigned int len;
229
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400230 batch_count++;
231
232 /* we want to process as big a batch as we can, but
233 * we also want to avoid softlockups. If we've been
234 * through a lot of messages, lets back off and see
235 * if anyone else jumps in
236 */
Santosh Shilimkar4bebdd72015-09-10 11:57:14 -0700237 if (batch_count >= send_batch_count)
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400238 goto over_batch;
239
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700240 spin_lock_irqsave(&cp->cp_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000241
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700242 if (!list_empty(&cp->cp_send_queue)) {
243 rm = list_entry(cp->cp_send_queue.next,
Andy Grover5c115592009-02-24 15:30:27 +0000244 struct rds_message,
245 m_conn_item);
246 rds_message_addref(rm);
247
248 /*
249 * Move the message from the send queue to the retransmit
250 * list right away.
251 */
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700252 list_move_tail(&rm->m_conn_item,
253 &cp->cp_retrans);
Andy Grover5c115592009-02-24 15:30:27 +0000254 }
255
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700256 spin_unlock_irqrestore(&cp->cp_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000257
Andy Groverfcc54502010-03-29 17:08:49 -0700258 if (!rm)
Andy Grover5c115592009-02-24 15:30:27 +0000259 break;
Andy Grover5c115592009-02-24 15:30:27 +0000260
261 /* Unfortunately, the way Infiniband deals with
262 * RDMA to a bad MR key is by moving the entire
263 * queue pair to error state. We cold possibly
264 * recover from that, but right now we drop the
265 * connection.
266 * Therefore, we never retransmit messages with RDMA ops.
267 */
Sowmini Varadhan905dd412016-11-16 13:29:49 -0800268 if (test_bit(RDS_MSG_FLUSH, &rm->m_flags) ||
269 (rm->rdma.op_active &&
270 test_bit(RDS_MSG_RETRANSMITTED, &rm->m_flags))) {
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700271 spin_lock_irqsave(&cp->cp_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000272 if (test_and_clear_bit(RDS_MSG_ON_CONN, &rm->m_flags))
273 list_move(&rm->m_conn_item, &to_be_dropped);
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700274 spin_unlock_irqrestore(&cp->cp_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000275 continue;
276 }
277
278 /* Require an ACK every once in a while */
279 len = ntohl(rm->m_inc.i_hdr.h_len);
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700280 if (cp->cp_unacked_packets == 0 ||
281 cp->cp_unacked_bytes < len) {
HÃ¥kon Buggef530f392017-09-05 17:42:01 +0200282 set_bit(RDS_MSG_ACK_REQUIRED, &rm->m_flags);
Andy Grover5c115592009-02-24 15:30:27 +0000283
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700284 cp->cp_unacked_packets =
285 rds_sysctl_max_unacked_packets;
286 cp->cp_unacked_bytes =
287 rds_sysctl_max_unacked_bytes;
Andy Grover5c115592009-02-24 15:30:27 +0000288 rds_stats_inc(s_send_ack_required);
289 } else {
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700290 cp->cp_unacked_bytes -= len;
291 cp->cp_unacked_packets--;
Andy Grover5c115592009-02-24 15:30:27 +0000292 }
293
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700294 cp->cp_xmit_rm = rm;
Andy Grover5c115592009-02-24 15:30:27 +0000295 }
296
Andy Grover2c3a5f92010-03-01 16:10:40 -0800297 /* The transport either sends the whole rdma or none of it */
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700298 if (rm->rdma.op_active && !cp->cp_xmit_rdma_sent) {
Andy Groverff3d7d32010-03-01 14:03:09 -0800299 rm->m_final_op = &rm->rdma;
santosh.shilimkar@oracle.com4f731132015-08-22 15:45:29 -0700300 /* The transport owns the mapped memory for now.
301 * You can't unmap it while it's on the send queue
302 */
303 set_bit(RDS_MSG_MAPPED, &rm->m_flags);
Andy Grover2c3a5f92010-03-01 16:10:40 -0800304 ret = conn->c_trans->xmit_rdma(conn, &rm->rdma);
santosh.shilimkar@oracle.com4f731132015-08-22 15:45:29 -0700305 if (ret) {
306 clear_bit(RDS_MSG_MAPPED, &rm->m_flags);
307 wake_up_interruptible(&rm->m_flush_wait);
Andy Grover2c3a5f92010-03-01 16:10:40 -0800308 break;
santosh.shilimkar@oracle.com4f731132015-08-22 15:45:29 -0700309 }
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700310 cp->cp_xmit_rdma_sent = 1;
Andy Grover2c3a5f92010-03-01 16:10:40 -0800311
Andy Grover2c3a5f92010-03-01 16:10:40 -0800312 }
313
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700314 if (rm->atomic.op_active && !cp->cp_xmit_atomic_sent) {
Andy Groverff3d7d32010-03-01 14:03:09 -0800315 rm->m_final_op = &rm->atomic;
santosh.shilimkar@oracle.com4f731132015-08-22 15:45:29 -0700316 /* The transport owns the mapped memory for now.
317 * You can't unmap it while it's on the send queue
318 */
319 set_bit(RDS_MSG_MAPPED, &rm->m_flags);
Andy Groverff3d7d32010-03-01 14:03:09 -0800320 ret = conn->c_trans->xmit_atomic(conn, &rm->atomic);
santosh.shilimkar@oracle.com4f731132015-08-22 15:45:29 -0700321 if (ret) {
322 clear_bit(RDS_MSG_MAPPED, &rm->m_flags);
323 wake_up_interruptible(&rm->m_flush_wait);
Andy Grover15133f62010-01-12 14:33:38 -0800324 break;
santosh.shilimkar@oracle.com4f731132015-08-22 15:45:29 -0700325 }
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700326 cp->cp_xmit_atomic_sent = 1;
Andy Groverff3d7d32010-03-01 14:03:09 -0800327
Andy Grover15133f62010-01-12 14:33:38 -0800328 }
329
Andy Grover2c3a5f92010-03-01 16:10:40 -0800330 /*
331 * A number of cases require an RDS header to be sent
332 * even if there is no data.
333 * We permit 0-byte sends; rds-ping depends on this.
334 * However, if there are exclusively attached silent ops,
335 * we skip the hdr/data send, to enable silent operation.
336 */
337 if (rm->data.op_nents == 0) {
338 int ops_present;
339 int all_ops_are_silent = 1;
Andy Grover241eef32010-01-19 21:25:26 -0800340
Andy Grover2c3a5f92010-03-01 16:10:40 -0800341 ops_present = (rm->atomic.op_active || rm->rdma.op_active);
342 if (rm->atomic.op_active && !rm->atomic.op_silent)
343 all_ops_are_silent = 0;
344 if (rm->rdma.op_active && !rm->rdma.op_silent)
345 all_ops_are_silent = 0;
Andy Grover241eef32010-01-19 21:25:26 -0800346
Andy Grover2c3a5f92010-03-01 16:10:40 -0800347 if (ops_present && all_ops_are_silent
348 && !rm->m_rdma_cookie)
349 rm->data.op_active = 0;
Andy Grover5c115592009-02-24 15:30:27 +0000350 }
351
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700352 if (rm->data.op_active && !cp->cp_xmit_data_sent) {
Andy Groverff3d7d32010-03-01 14:03:09 -0800353 rm->m_final_op = &rm->data;
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700354
Andy Grover5c115592009-02-24 15:30:27 +0000355 ret = conn->c_trans->xmit(conn, rm,
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700356 cp->cp_xmit_hdr_off,
357 cp->cp_xmit_sg,
358 cp->cp_xmit_data_off);
Andy Grover5c115592009-02-24 15:30:27 +0000359 if (ret <= 0)
360 break;
361
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700362 if (cp->cp_xmit_hdr_off < sizeof(struct rds_header)) {
Andy Grover5c115592009-02-24 15:30:27 +0000363 tmp = min_t(int, ret,
364 sizeof(struct rds_header) -
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700365 cp->cp_xmit_hdr_off);
366 cp->cp_xmit_hdr_off += tmp;
Andy Grover5c115592009-02-24 15:30:27 +0000367 ret -= tmp;
368 }
369
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700370 sg = &rm->data.op_sg[cp->cp_xmit_sg];
Andy Grover5c115592009-02-24 15:30:27 +0000371 while (ret) {
372 tmp = min_t(int, ret, sg->length -
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700373 cp->cp_xmit_data_off);
374 cp->cp_xmit_data_off += tmp;
Andy Grover5c115592009-02-24 15:30:27 +0000375 ret -= tmp;
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700376 if (cp->cp_xmit_data_off == sg->length) {
377 cp->cp_xmit_data_off = 0;
Andy Grover5c115592009-02-24 15:30:27 +0000378 sg++;
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700379 cp->cp_xmit_sg++;
380 BUG_ON(ret != 0 && cp->cp_xmit_sg ==
381 rm->data.op_nents);
Andy Grover5c115592009-02-24 15:30:27 +0000382 }
383 }
Andy Grover5b2366b2010-02-03 19:36:44 -0800384
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700385 if (cp->cp_xmit_hdr_off == sizeof(struct rds_header) &&
386 (cp->cp_xmit_sg == rm->data.op_nents))
387 cp->cp_xmit_data_sent = 1;
Andy Grover5b2366b2010-02-03 19:36:44 -0800388 }
389
390 /*
391 * A rm will only take multiple times through this loop
392 * if there is a data op. Thus, if the data is sent (or there was
393 * none), then we're done with the rm.
394 */
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700395 if (!rm->data.op_active || cp->cp_xmit_data_sent) {
396 cp->cp_xmit_rm = NULL;
397 cp->cp_xmit_sg = 0;
398 cp->cp_xmit_hdr_off = 0;
399 cp->cp_xmit_data_off = 0;
400 cp->cp_xmit_rdma_sent = 0;
401 cp->cp_xmit_atomic_sent = 0;
402 cp->cp_xmit_data_sent = 0;
Andy Grover5b2366b2010-02-03 19:36:44 -0800403
404 rds_message_put(rm);
Andy Grover5c115592009-02-24 15:30:27 +0000405 }
406 }
407
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400408over_batch:
Sowmini Varadhan226f7a72016-06-30 16:11:10 -0700409 if (conn->c_trans->xmit_path_complete)
410 conn->c_trans->xmit_path_complete(cp);
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -0700411 release_in_xmit(cp);
Andy Grover5c115592009-02-24 15:30:27 +0000412
Andy Grover2ad80992010-03-23 17:48:04 -0700413 /* Nuke any messages we decided not to retransmit. */
414 if (!list_empty(&to_be_dropped)) {
415 /* irqs on here, so we can put(), unlike above */
416 list_for_each_entry(rm, &to_be_dropped, m_conn_item)
417 rds_message_put(rm);
418 rds_send_remove_from_sock(&to_be_dropped, RDS_RDMA_DROPPED);
419 }
420
Andy Groverfcc54502010-03-29 17:08:49 -0700421 /*
Zach Brown0f4b1c72010-06-04 14:41:41 -0700422 * Other senders can queue a message after we last test the send queue
423 * but before we clear RDS_IN_XMIT. In that case they'd back off and
424 * not try and send their newly queued message. We need to check the
425 * send queue after having cleared RDS_IN_XMIT so that their message
426 * doesn't get stuck on the send queue.
Andy Groverfcc54502010-03-29 17:08:49 -0700427 *
428 * If the transport cannot continue (i.e ret != 0), then it must
429 * call us when more room is available, such as from the tx
430 * completion handler.
Sowmini Varadhan443be0e2015-04-08 12:33:47 -0400431 *
432 * We have an extra generation check here so that if someone manages
433 * to jump in after our release_in_xmit, we'll see that they have done
434 * some work and we will skip our goto
Andy Groverfcc54502010-03-29 17:08:49 -0700435 */
436 if (ret == 0) {
HÃ¥kon Bugge126f7602017-09-06 18:35:51 +0200437 bool raced;
438
Chris Mason9e29db02010-04-15 16:38:14 -0400439 smp_mb();
HÃ¥kon Bugge126f7602017-09-06 18:35:51 +0200440 raced = send_gen != READ_ONCE(cp->cp_send_gen);
441
santosh.shilimkar@oracle.com0c484242015-08-22 15:45:27 -0700442 if ((test_bit(0, &conn->c_map_queued) ||
HÃ¥kon Bugge126f7602017-09-06 18:35:51 +0200443 !list_empty(&cp->cp_send_queue)) && !raced) {
Santosh Shilimkar4bebdd72015-09-10 11:57:14 -0700444 if (batch_count < send_batch_count)
445 goto restart;
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -0800446 rcu_read_lock();
Sowmini Varadhanebeeb1a2018-02-03 04:26:51 -0800447 if (rds_destroy_pending(cp->cp_conn))
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -0800448 ret = -ENETUNREACH;
449 else
450 queue_delayed_work(rds_wq, &cp->cp_send_w, 1);
451 rcu_read_unlock();
HÃ¥kon Bugge126f7602017-09-06 18:35:51 +0200452 } else if (raced) {
453 rds_stats_inc(s_send_lock_queue_raced);
Andy Grover5c115592009-02-24 15:30:27 +0000454 }
Andy Grover5c115592009-02-24 15:30:27 +0000455 }
456out:
457 return ret;
458}
Santosh Shilimkar0c28c042015-09-06 02:18:51 -0400459EXPORT_SYMBOL_GPL(rds_send_xmit);
Andy Grover5c115592009-02-24 15:30:27 +0000460
461static void rds_send_sndbuf_remove(struct rds_sock *rs, struct rds_message *rm)
462{
463 u32 len = be32_to_cpu(rm->m_inc.i_hdr.h_len);
464
465 assert_spin_locked(&rs->rs_lock);
466
467 BUG_ON(rs->rs_snd_bytes < len);
468 rs->rs_snd_bytes -= len;
469
470 if (rs->rs_snd_bytes == 0)
471 rds_stats_inc(s_send_queue_empty);
472}
473
474static inline int rds_send_is_acked(struct rds_message *rm, u64 ack,
475 is_acked_func is_acked)
476{
477 if (is_acked)
478 return is_acked(rm, ack);
479 return be64_to_cpu(rm->m_inc.i_hdr.h_sequence) <= ack;
480}
481
482/*
Andy Grover5c115592009-02-24 15:30:27 +0000483 * This is pretty similar to what happens below in the ACK
484 * handling code - except that we call here as soon as we get
485 * the IB send completion on the RDMA op and the accompanying
486 * message.
487 */
488void rds_rdma_send_complete(struct rds_message *rm, int status)
489{
490 struct rds_sock *rs = NULL;
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800491 struct rm_rdma_op *ro;
Andy Grover5c115592009-02-24 15:30:27 +0000492 struct rds_notifier *notifier;
Andy Grover9de08642010-03-29 16:50:54 -0700493 unsigned long flags;
Santosh Shilimkar941f8d52016-02-18 20:06:47 -0800494 unsigned int notify = 0;
Andy Grover5c115592009-02-24 15:30:27 +0000495
Andy Grover9de08642010-03-29 16:50:54 -0700496 spin_lock_irqsave(&rm->m_rs_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000497
Santosh Shilimkar941f8d52016-02-18 20:06:47 -0800498 notify = rm->rdma.op_notify | rm->data.op_notify;
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800499 ro = &rm->rdma;
Joe Perchesf64f9e72009-11-29 16:55:45 -0800500 if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags) &&
Santosh Shilimkar941f8d52016-02-18 20:06:47 -0800501 ro->op_active && notify && ro->op_notifier) {
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800502 notifier = ro->op_notifier;
Andy Grover5c115592009-02-24 15:30:27 +0000503 rs = rm->m_rs;
504 sock_hold(rds_rs_to_sk(rs));
505
506 notifier->n_status = status;
507 spin_lock(&rs->rs_lock);
508 list_add_tail(&notifier->n_list, &rs->rs_notify_queue);
509 spin_unlock(&rs->rs_lock);
510
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800511 ro->op_notifier = NULL;
Andy Grover5c115592009-02-24 15:30:27 +0000512 }
513
Andy Grover9de08642010-03-29 16:50:54 -0700514 spin_unlock_irqrestore(&rm->m_rs_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000515
516 if (rs) {
517 rds_wake_sk_sleep(rs);
518 sock_put(rds_rs_to_sk(rs));
519 }
520}
Andy Grover616b7572009-08-21 12:28:32 +0000521EXPORT_SYMBOL_GPL(rds_rdma_send_complete);
Andy Grover5c115592009-02-24 15:30:27 +0000522
523/*
Andy Grover15133f62010-01-12 14:33:38 -0800524 * Just like above, except looks at atomic op
525 */
526void rds_atomic_send_complete(struct rds_message *rm, int status)
527{
528 struct rds_sock *rs = NULL;
529 struct rm_atomic_op *ao;
530 struct rds_notifier *notifier;
Andy Grovercf4b7382010-03-29 16:50:54 -0700531 unsigned long flags;
Andy Grover15133f62010-01-12 14:33:38 -0800532
Andy Grovercf4b7382010-03-29 16:50:54 -0700533 spin_lock_irqsave(&rm->m_rs_lock, flags);
Andy Grover15133f62010-01-12 14:33:38 -0800534
535 ao = &rm->atomic;
536 if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags)
537 && ao->op_active && ao->op_notify && ao->op_notifier) {
538 notifier = ao->op_notifier;
539 rs = rm->m_rs;
540 sock_hold(rds_rs_to_sk(rs));
541
542 notifier->n_status = status;
543 spin_lock(&rs->rs_lock);
544 list_add_tail(&notifier->n_list, &rs->rs_notify_queue);
545 spin_unlock(&rs->rs_lock);
546
547 ao->op_notifier = NULL;
548 }
549
Andy Grovercf4b7382010-03-29 16:50:54 -0700550 spin_unlock_irqrestore(&rm->m_rs_lock, flags);
Andy Grover15133f62010-01-12 14:33:38 -0800551
552 if (rs) {
553 rds_wake_sk_sleep(rs);
554 sock_put(rds_rs_to_sk(rs));
555 }
556}
557EXPORT_SYMBOL_GPL(rds_atomic_send_complete);
558
559/*
Andy Grover5c115592009-02-24 15:30:27 +0000560 * This is the same as rds_rdma_send_complete except we
561 * don't do any locking - we have all the ingredients (message,
562 * socket, socket lock) and can just move the notifier.
563 */
564static inline void
Andy Grover940786e2010-02-19 18:04:58 -0800565__rds_send_complete(struct rds_sock *rs, struct rds_message *rm, int status)
Andy Grover5c115592009-02-24 15:30:27 +0000566{
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800567 struct rm_rdma_op *ro;
Andy Grover940786e2010-02-19 18:04:58 -0800568 struct rm_atomic_op *ao;
Andy Grover5c115592009-02-24 15:30:27 +0000569
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800570 ro = &rm->rdma;
571 if (ro->op_active && ro->op_notify && ro->op_notifier) {
572 ro->op_notifier->n_status = status;
573 list_add_tail(&ro->op_notifier->n_list, &rs->rs_notify_queue);
574 ro->op_notifier = NULL;
Andy Grover5c115592009-02-24 15:30:27 +0000575 }
576
Andy Grover940786e2010-02-19 18:04:58 -0800577 ao = &rm->atomic;
578 if (ao->op_active && ao->op_notify && ao->op_notifier) {
579 ao->op_notifier->n_status = status;
580 list_add_tail(&ao->op_notifier->n_list, &rs->rs_notify_queue);
581 ao->op_notifier = NULL;
582 }
583
Andy Grover5c115592009-02-24 15:30:27 +0000584 /* No need to wake the app - caller does this */
585}
586
587/*
Andy Grover5c115592009-02-24 15:30:27 +0000588 * This removes messages from the socket's list if they're on it. The list
589 * argument must be private to the caller, we must be able to modify it
590 * without locks. The messages must have a reference held for their
591 * position on the list. This function will drop that reference after
592 * removing the messages from the 'messages' list regardless of if it found
593 * the messages on the socket list or not.
594 */
stephen hemmingerff51bf82010-10-19 08:08:33 +0000595static void rds_send_remove_from_sock(struct list_head *messages, int status)
Andy Grover5c115592009-02-24 15:30:27 +0000596{
Andy Grover561c7df2010-03-11 13:50:06 +0000597 unsigned long flags;
Andy Grover5c115592009-02-24 15:30:27 +0000598 struct rds_sock *rs = NULL;
599 struct rds_message *rm;
600
Andy Grover5c115592009-02-24 15:30:27 +0000601 while (!list_empty(messages)) {
Andy Grover561c7df2010-03-11 13:50:06 +0000602 int was_on_sock = 0;
603
Andy Grover5c115592009-02-24 15:30:27 +0000604 rm = list_entry(messages->next, struct rds_message,
605 m_conn_item);
606 list_del_init(&rm->m_conn_item);
607
608 /*
609 * If we see this flag cleared then we're *sure* that someone
610 * else beat us to removing it from the sock. If we race
611 * with their flag update we'll get the lock and then really
612 * see that the flag has been cleared.
613 *
614 * The message spinlock makes sure nobody clears rm->m_rs
615 * while we're messing with it. It does not prevent the
616 * message from being removed from the socket, though.
617 */
Andy Grover561c7df2010-03-11 13:50:06 +0000618 spin_lock_irqsave(&rm->m_rs_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000619 if (!test_bit(RDS_MSG_ON_SOCK, &rm->m_flags))
620 goto unlock_and_drop;
621
622 if (rs != rm->m_rs) {
623 if (rs) {
Andy Grover5c115592009-02-24 15:30:27 +0000624 rds_wake_sk_sleep(rs);
625 sock_put(rds_rs_to_sk(rs));
626 }
627 rs = rm->m_rs;
Herton R. Krzesinski593cbb32014-10-01 18:49:54 -0300628 if (rs)
629 sock_hold(rds_rs_to_sk(rs));
Andy Grover5c115592009-02-24 15:30:27 +0000630 }
Herton R. Krzesinski593cbb32014-10-01 18:49:54 -0300631 if (!rs)
632 goto unlock_and_drop;
Tina Yang048c15e2010-03-11 13:50:00 +0000633 spin_lock(&rs->rs_lock);
Andy Grover5c115592009-02-24 15:30:27 +0000634
635 if (test_and_clear_bit(RDS_MSG_ON_SOCK, &rm->m_flags)) {
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800636 struct rm_rdma_op *ro = &rm->rdma;
Andy Grover5c115592009-02-24 15:30:27 +0000637 struct rds_notifier *notifier;
638
639 list_del_init(&rm->m_sock_item);
640 rds_send_sndbuf_remove(rs, rm);
641
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800642 if (ro->op_active && ro->op_notifier &&
643 (ro->op_notify || (ro->op_recverr && status))) {
644 notifier = ro->op_notifier;
Andy Grover5c115592009-02-24 15:30:27 +0000645 list_add_tail(&notifier->n_list,
646 &rs->rs_notify_queue);
647 if (!notifier->n_status)
648 notifier->n_status = status;
Andy Groverf8b3aaf2010-03-01 14:11:53 -0800649 rm->rdma.op_notifier = NULL;
Andy Grover5c115592009-02-24 15:30:27 +0000650 }
Andy Grover561c7df2010-03-11 13:50:06 +0000651 was_on_sock = 1;
Andy Grover5c115592009-02-24 15:30:27 +0000652 }
Tina Yang048c15e2010-03-11 13:50:00 +0000653 spin_unlock(&rs->rs_lock);
Andy Grover5c115592009-02-24 15:30:27 +0000654
655unlock_and_drop:
Andy Grover561c7df2010-03-11 13:50:06 +0000656 spin_unlock_irqrestore(&rm->m_rs_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000657 rds_message_put(rm);
Andy Grover561c7df2010-03-11 13:50:06 +0000658 if (was_on_sock)
659 rds_message_put(rm);
Andy Grover5c115592009-02-24 15:30:27 +0000660 }
661
662 if (rs) {
Andy Grover5c115592009-02-24 15:30:27 +0000663 rds_wake_sk_sleep(rs);
664 sock_put(rds_rs_to_sk(rs));
665 }
Andy Grover5c115592009-02-24 15:30:27 +0000666}
667
668/*
669 * Transports call here when they've determined that the receiver queued
670 * messages up to, and including, the given sequence number. Messages are
671 * moved to the retrans queue when rds_send_xmit picks them off the send
672 * queue. This means that in the TCP case, the message may not have been
673 * assigned the m_ack_seq yet - but that's fine as long as tcp_is_acked
674 * checks the RDS_MSG_HAS_ACK_SEQ bit.
Andy Grover5c115592009-02-24 15:30:27 +0000675 */
Sowmini Varadhan5c3d2742016-06-13 09:44:31 -0700676void rds_send_path_drop_acked(struct rds_conn_path *cp, u64 ack,
677 is_acked_func is_acked)
Andy Grover5c115592009-02-24 15:30:27 +0000678{
679 struct rds_message *rm, *tmp;
680 unsigned long flags;
681 LIST_HEAD(list);
682
Sowmini Varadhan5c3d2742016-06-13 09:44:31 -0700683 spin_lock_irqsave(&cp->cp_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000684
Sowmini Varadhan5c3d2742016-06-13 09:44:31 -0700685 list_for_each_entry_safe(rm, tmp, &cp->cp_retrans, m_conn_item) {
Andy Grover5c115592009-02-24 15:30:27 +0000686 if (!rds_send_is_acked(rm, ack, is_acked))
687 break;
688
689 list_move(&rm->m_conn_item, &list);
690 clear_bit(RDS_MSG_ON_CONN, &rm->m_flags);
691 }
692
693 /* order flag updates with spin locks */
694 if (!list_empty(&list))
Peter Zijlstra4e857c52014-03-17 18:06:10 +0100695 smp_mb__after_atomic();
Andy Grover5c115592009-02-24 15:30:27 +0000696
Sowmini Varadhan5c3d2742016-06-13 09:44:31 -0700697 spin_unlock_irqrestore(&cp->cp_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000698
699 /* now remove the messages from the sock list as needed */
700 rds_send_remove_from_sock(&list, RDS_RDMA_SUCCESS);
701}
Sowmini Varadhan5c3d2742016-06-13 09:44:31 -0700702EXPORT_SYMBOL_GPL(rds_send_path_drop_acked);
703
704void rds_send_drop_acked(struct rds_connection *conn, u64 ack,
705 is_acked_func is_acked)
706{
707 WARN_ON(conn->c_trans->t_mp_capable);
708 rds_send_path_drop_acked(&conn->c_path[0], ack, is_acked);
709}
Andy Grover616b7572009-08-21 12:28:32 +0000710EXPORT_SYMBOL_GPL(rds_send_drop_acked);
Andy Grover5c115592009-02-24 15:30:27 +0000711
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -0700712void rds_send_drop_to(struct rds_sock *rs, struct sockaddr_in6 *dest)
Andy Grover5c115592009-02-24 15:30:27 +0000713{
714 struct rds_message *rm, *tmp;
715 struct rds_connection *conn;
Sowmini Varadhan01ff34e2016-06-13 09:44:35 -0700716 struct rds_conn_path *cp;
Andy Grover7c82eaf2010-02-19 18:01:41 -0800717 unsigned long flags;
Andy Grover5c115592009-02-24 15:30:27 +0000718 LIST_HEAD(list);
Andy Grover5c115592009-02-24 15:30:27 +0000719
720 /* get all the messages we're dropping under the rs lock */
721 spin_lock_irqsave(&rs->rs_lock, flags);
722
723 list_for_each_entry_safe(rm, tmp, &rs->rs_send_queue, m_sock_item) {
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -0700724 if (dest &&
725 (!ipv6_addr_equal(&dest->sin6_addr, &rm->m_daddr) ||
726 dest->sin6_port != rm->m_inc.i_hdr.h_dport))
Andy Grover5c115592009-02-24 15:30:27 +0000727 continue;
728
Andy Grover5c115592009-02-24 15:30:27 +0000729 list_move(&rm->m_sock_item, &list);
730 rds_send_sndbuf_remove(rs, rm);
731 clear_bit(RDS_MSG_ON_SOCK, &rm->m_flags);
Andy Grover5c115592009-02-24 15:30:27 +0000732 }
733
734 /* order flag updates with the rs lock */
Peter Zijlstra4e857c52014-03-17 18:06:10 +0100735 smp_mb__after_atomic();
Andy Grover5c115592009-02-24 15:30:27 +0000736
737 spin_unlock_irqrestore(&rs->rs_lock, flags);
738
Andy Grover7c82eaf2010-02-19 18:01:41 -0800739 if (list_empty(&list))
740 return;
Andy Grover5c115592009-02-24 15:30:27 +0000741
Andy Grover7c82eaf2010-02-19 18:01:41 -0800742 /* Remove the messages from the conn */
Andy Grover5c115592009-02-24 15:30:27 +0000743 list_for_each_entry(rm, &list, m_sock_item) {
Andy Grover7c82eaf2010-02-19 18:01:41 -0800744
745 conn = rm->m_inc.i_conn;
Sowmini Varadhan01ff34e2016-06-13 09:44:35 -0700746 if (conn->c_trans->t_mp_capable)
747 cp = rm->m_inc.i_conn_path;
748 else
749 cp = &conn->c_path[0];
Andy Grover7c82eaf2010-02-19 18:01:41 -0800750
Sowmini Varadhan01ff34e2016-06-13 09:44:35 -0700751 spin_lock_irqsave(&cp->cp_lock, flags);
Andy Grover7c82eaf2010-02-19 18:01:41 -0800752 /*
753 * Maybe someone else beat us to removing rm from the conn.
754 * If we race with their flag update we'll get the lock and
755 * then really see that the flag has been cleared.
756 */
757 if (!test_and_clear_bit(RDS_MSG_ON_CONN, &rm->m_flags)) {
Sowmini Varadhan01ff34e2016-06-13 09:44:35 -0700758 spin_unlock_irqrestore(&cp->cp_lock, flags);
Andy Grover7c82eaf2010-02-19 18:01:41 -0800759 continue;
760 }
Andy Grover9de08642010-03-29 16:50:54 -0700761 list_del_init(&rm->m_conn_item);
Sowmini Varadhan01ff34e2016-06-13 09:44:35 -0700762 spin_unlock_irqrestore(&cp->cp_lock, flags);
Andy Grover7c82eaf2010-02-19 18:01:41 -0800763
764 /*
765 * Couldn't grab m_rs_lock in top loop (lock ordering),
766 * but we can now.
767 */
Andy Grover9de08642010-03-29 16:50:54 -0700768 spin_lock_irqsave(&rm->m_rs_lock, flags);
Andy Grover7c82eaf2010-02-19 18:01:41 -0800769
Tina Yang550a8002010-03-11 13:50:03 +0000770 spin_lock(&rs->rs_lock);
Andy Grover940786e2010-02-19 18:04:58 -0800771 __rds_send_complete(rs, rm, RDS_RDMA_CANCELED);
Tina Yang550a8002010-03-11 13:50:03 +0000772 spin_unlock(&rs->rs_lock);
Andy Grover7c82eaf2010-02-19 18:01:41 -0800773
Andy Grover9de08642010-03-29 16:50:54 -0700774 spin_unlock_irqrestore(&rm->m_rs_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +0000775
Andy Grover7c82eaf2010-02-19 18:01:41 -0800776 rds_message_put(rm);
Andy Grover5c115592009-02-24 15:30:27 +0000777 }
778
Andy Grover7c82eaf2010-02-19 18:01:41 -0800779 rds_wake_sk_sleep(rs);
Tina Yang550a8002010-03-11 13:50:03 +0000780
Andy Grover5c115592009-02-24 15:30:27 +0000781 while (!list_empty(&list)) {
782 rm = list_entry(list.next, struct rds_message, m_sock_item);
783 list_del_init(&rm->m_sock_item);
Andy Grover5c115592009-02-24 15:30:27 +0000784 rds_message_wait(rm);
santosh.shilimkar@oracle.comdfcec252015-08-22 15:45:33 -0700785
786 /* just in case the code above skipped this message
787 * because RDS_MSG_ON_CONN wasn't set, run it again here
788 * taking m_rs_lock is the only thing that keeps us
789 * from racing with ack processing.
790 */
791 spin_lock_irqsave(&rm->m_rs_lock, flags);
792
793 spin_lock(&rs->rs_lock);
794 __rds_send_complete(rs, rm, RDS_RDMA_CANCELED);
795 spin_unlock(&rs->rs_lock);
796
santosh.shilimkar@oracle.comdfcec252015-08-22 15:45:33 -0700797 spin_unlock_irqrestore(&rm->m_rs_lock, flags);
798
Andy Grover5c115592009-02-24 15:30:27 +0000799 rds_message_put(rm);
800 }
801}
802
803/*
804 * we only want this to fire once so we use the callers 'queued'. It's
805 * possible that another thread can race with us and remove the
806 * message from the flow with RDS_CANCEL_SENT_TO.
807 */
808static int rds_send_queue_rm(struct rds_sock *rs, struct rds_connection *conn,
Sowmini Varadhan780a6d92016-06-13 09:44:33 -0700809 struct rds_conn_path *cp,
Andy Grover5c115592009-02-24 15:30:27 +0000810 struct rds_message *rm, __be16 sport,
811 __be16 dport, int *queued)
812{
813 unsigned long flags;
814 u32 len;
815
816 if (*queued)
817 goto out;
818
819 len = be32_to_cpu(rm->m_inc.i_hdr.h_len);
820
821 /* this is the only place which holds both the socket's rs_lock
822 * and the connection's c_lock */
823 spin_lock_irqsave(&rs->rs_lock, flags);
824
825 /*
826 * If there is a little space in sndbuf, we don't queue anything,
827 * and userspace gets -EAGAIN. But poll() indicates there's send
828 * room. This can lead to bad behavior (spinning) if snd_bytes isn't
829 * freed up by incoming acks. So we check the *old* value of
830 * rs_snd_bytes here to allow the last msg to exceed the buffer,
831 * and poll() now knows no more data can be sent.
832 */
833 if (rs->rs_snd_bytes < rds_sk_sndbuf(rs)) {
834 rs->rs_snd_bytes += len;
835
836 /* let recv side know we are close to send space exhaustion.
837 * This is probably not the optimal way to do it, as this
838 * means we set the flag on *all* messages as soon as our
839 * throughput hits a certain threshold.
840 */
841 if (rs->rs_snd_bytes >= rds_sk_sndbuf(rs) / 2)
HÃ¥kon Buggef530f392017-09-05 17:42:01 +0200842 set_bit(RDS_MSG_ACK_REQUIRED, &rm->m_flags);
Andy Grover5c115592009-02-24 15:30:27 +0000843
844 list_add_tail(&rm->m_sock_item, &rs->rs_send_queue);
845 set_bit(RDS_MSG_ON_SOCK, &rm->m_flags);
846 rds_message_addref(rm);
Sowmini Varadhanea8994c2018-02-15 10:49:33 -0800847 sock_hold(rds_rs_to_sk(rs));
Andy Grover5c115592009-02-24 15:30:27 +0000848 rm->m_rs = rs;
849
850 /* The code ordering is a little weird, but we're
851 trying to minimize the time we hold c_lock */
852 rds_message_populate_header(&rm->m_inc.i_hdr, sport, dport, 0);
853 rm->m_inc.i_conn = conn;
Sowmini Varadhan780a6d92016-06-13 09:44:33 -0700854 rm->m_inc.i_conn_path = cp;
Andy Grover5c115592009-02-24 15:30:27 +0000855 rds_message_addref(rm);
856
Sowmini Varadhan780a6d92016-06-13 09:44:33 -0700857 spin_lock(&cp->cp_lock);
858 rm->m_inc.i_hdr.h_sequence = cpu_to_be64(cp->cp_next_tx_seq++);
859 list_add_tail(&rm->m_conn_item, &cp->cp_send_queue);
Andy Grover5c115592009-02-24 15:30:27 +0000860 set_bit(RDS_MSG_ON_CONN, &rm->m_flags);
Sowmini Varadhan780a6d92016-06-13 09:44:33 -0700861 spin_unlock(&cp->cp_lock);
Andy Grover5c115592009-02-24 15:30:27 +0000862
863 rdsdebug("queued msg %p len %d, rs %p bytes %d seq %llu\n",
864 rm, len, rs, rs->rs_snd_bytes,
865 (unsigned long long)be64_to_cpu(rm->m_inc.i_hdr.h_sequence));
866
867 *queued = 1;
868 }
869
870 spin_unlock_irqrestore(&rs->rs_lock, flags);
871out:
872 return *queued;
873}
874
Andy Groverfc445082010-01-12 12:56:06 -0800875/*
876 * rds_message is getting to be quite complicated, and we'd like to allocate
877 * it all in one go. This figures out how big it needs to be up front.
878 */
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -0800879static int rds_rm_size(struct msghdr *msg, int num_sgs)
Andy Groverfc445082010-01-12 12:56:06 -0800880{
Andy Groverff87e972010-01-12 14:13:15 -0800881 struct cmsghdr *cmsg;
Andy Groverfc445082010-01-12 12:56:06 -0800882 int size = 0;
Andy Groveraa0a4ef2010-04-13 12:00:35 -0700883 int cmsg_groups = 0;
Andy Groverff87e972010-01-12 14:13:15 -0800884 int retval;
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -0800885 bool zcopy_cookie = false;
Andy Groverfc445082010-01-12 12:56:06 -0800886
Gu Zhengf95b4142014-12-11 11:22:04 +0800887 for_each_cmsghdr(cmsg, msg) {
Andy Groverff87e972010-01-12 14:13:15 -0800888 if (!CMSG_OK(msg, cmsg))
889 return -EINVAL;
890
891 if (cmsg->cmsg_level != SOL_RDS)
892 continue;
893
894 switch (cmsg->cmsg_type) {
895 case RDS_CMSG_RDMA_ARGS:
Andy Groveraa0a4ef2010-04-13 12:00:35 -0700896 cmsg_groups |= 1;
Andy Groverff87e972010-01-12 14:13:15 -0800897 retval = rds_rdma_extra_size(CMSG_DATA(cmsg));
898 if (retval < 0)
899 return retval;
900 size += retval;
Andy Groveraa0a4ef2010-04-13 12:00:35 -0700901
Andy Groverff87e972010-01-12 14:13:15 -0800902 break;
903
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -0800904 case RDS_CMSG_ZCOPY_COOKIE:
905 zcopy_cookie = true;
Gustavo A. R. Silvaf9053112018-02-19 12:10:20 -0600906 /* fall through */
907
Andy Groverff87e972010-01-12 14:13:15 -0800908 case RDS_CMSG_RDMA_DEST:
909 case RDS_CMSG_RDMA_MAP:
Andy Groveraa0a4ef2010-04-13 12:00:35 -0700910 cmsg_groups |= 2;
Andy Groverff87e972010-01-12 14:13:15 -0800911 /* these are valid but do no add any size */
912 break;
913
Andy Grover15133f62010-01-12 14:33:38 -0800914 case RDS_CMSG_ATOMIC_CSWP:
915 case RDS_CMSG_ATOMIC_FADD:
Andy Grover20c72bd2010-08-25 05:51:28 -0700916 case RDS_CMSG_MASKED_ATOMIC_CSWP:
917 case RDS_CMSG_MASKED_ATOMIC_FADD:
Andy Groveraa0a4ef2010-04-13 12:00:35 -0700918 cmsg_groups |= 1;
Andy Grover15133f62010-01-12 14:33:38 -0800919 size += sizeof(struct scatterlist);
920 break;
921
Andy Groverff87e972010-01-12 14:13:15 -0800922 default:
923 return -EINVAL;
924 }
925
926 }
927
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -0800928 if ((msg->msg_flags & MSG_ZEROCOPY) && !zcopy_cookie)
929 return -EINVAL;
930
931 size += num_sgs * sizeof(struct scatterlist);
Andy Groverfc445082010-01-12 12:56:06 -0800932
Andy Groveraa0a4ef2010-04-13 12:00:35 -0700933 /* Ensure (DEST, MAP) are never used with (ARGS, ATOMIC) */
934 if (cmsg_groups == 3)
935 return -EINVAL;
936
Andy Groverfc445082010-01-12 12:56:06 -0800937 return size;
938}
939
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -0800940static int rds_cmsg_zcopy(struct rds_sock *rs, struct rds_message *rm,
941 struct cmsghdr *cmsg)
942{
943 u32 *cookie;
944
Sowmini Varadhan79a5b972018-02-22 13:40:27 -0800945 if (cmsg->cmsg_len < CMSG_LEN(sizeof(*cookie)) ||
946 !rm->data.op_mmp_znotifier)
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -0800947 return -EINVAL;
948 cookie = CMSG_DATA(cmsg);
949 rm->data.op_mmp_znotifier->z_cookie = *cookie;
950 return 0;
951}
952
Andy Grover5c115592009-02-24 15:30:27 +0000953static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm,
954 struct msghdr *msg, int *allocated_mr)
955{
956 struct cmsghdr *cmsg;
957 int ret = 0;
958
Gu Zhengf95b4142014-12-11 11:22:04 +0800959 for_each_cmsghdr(cmsg, msg) {
Andy Grover5c115592009-02-24 15:30:27 +0000960 if (!CMSG_OK(msg, cmsg))
961 return -EINVAL;
962
963 if (cmsg->cmsg_level != SOL_RDS)
964 continue;
965
966 /* As a side effect, RDMA_DEST and RDMA_MAP will set
Andy Grover15133f62010-01-12 14:33:38 -0800967 * rm->rdma.m_rdma_cookie and rm->rdma.m_rdma_mr.
Andy Grover5c115592009-02-24 15:30:27 +0000968 */
969 switch (cmsg->cmsg_type) {
970 case RDS_CMSG_RDMA_ARGS:
971 ret = rds_cmsg_rdma_args(rs, rm, cmsg);
972 break;
973
974 case RDS_CMSG_RDMA_DEST:
975 ret = rds_cmsg_rdma_dest(rs, rm, cmsg);
976 break;
977
978 case RDS_CMSG_RDMA_MAP:
979 ret = rds_cmsg_rdma_map(rs, rm, cmsg);
980 if (!ret)
981 *allocated_mr = 1;
Santosh Shilimkar584a8272016-07-04 17:04:37 -0700982 else if (ret == -ENODEV)
983 /* Accommodate the get_mr() case which can fail
984 * if connection isn't established yet.
985 */
986 ret = -EAGAIN;
Andy Grover5c115592009-02-24 15:30:27 +0000987 break;
Andy Grover15133f62010-01-12 14:33:38 -0800988 case RDS_CMSG_ATOMIC_CSWP:
989 case RDS_CMSG_ATOMIC_FADD:
Andy Grover20c72bd2010-08-25 05:51:28 -0700990 case RDS_CMSG_MASKED_ATOMIC_CSWP:
991 case RDS_CMSG_MASKED_ATOMIC_FADD:
Andy Grover15133f62010-01-12 14:33:38 -0800992 ret = rds_cmsg_atomic(rs, rm, cmsg);
993 break;
Andy Grover5c115592009-02-24 15:30:27 +0000994
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -0800995 case RDS_CMSG_ZCOPY_COOKIE:
996 ret = rds_cmsg_zcopy(rs, rm, cmsg);
997 break;
998
Andy Grover5c115592009-02-24 15:30:27 +0000999 default:
1000 return -EINVAL;
1001 }
1002
1003 if (ret)
1004 break;
1005 }
1006
1007 return ret;
1008}
1009
Ka-Cheong Poon9a4890b2018-10-08 09:17:11 -07001010static int rds_send_mprds_hash(struct rds_sock *rs,
1011 struct rds_connection *conn, int nonblock)
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001012{
1013 int hash;
1014
1015 if (conn->c_npaths == 0)
1016 hash = RDS_MPATH_HASH(rs, RDS_MPATH_WORKERS);
1017 else
1018 hash = RDS_MPATH_HASH(rs, conn->c_npaths);
1019 if (conn->c_npaths == 0 && hash != 0) {
Sowmini Varadhan69b92b52017-06-21 13:40:12 -07001020 rds_send_ping(conn, 0);
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001021
Ka-Cheong Poona43cced2018-04-11 00:57:25 -07001022 /* The underlying connection is not up yet. Need to wait
1023 * until it is up to be sure that the non-zero c_path can be
1024 * used. But if we are interrupted, we have to use the zero
1025 * c_path in case the connection ends up being non-MP capable.
1026 */
Ka-Cheong Poon9a4890b2018-10-08 09:17:11 -07001027 if (conn->c_npaths == 0) {
1028 /* Cannot wait for the connection be made, so just use
1029 * the base c_path.
1030 */
1031 if (nonblock)
1032 return 0;
Ka-Cheong Poona43cced2018-04-11 00:57:25 -07001033 if (wait_event_interruptible(conn->c_hs_waitq,
1034 conn->c_npaths != 0))
1035 hash = 0;
Ka-Cheong Poon9a4890b2018-10-08 09:17:11 -07001036 }
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001037 if (conn->c_npaths == 1)
1038 hash = 0;
1039 }
1040 return hash;
1041}
1042
Avinash Repakaf9fb69a2016-02-29 15:30:57 -08001043static int rds_rdma_bytes(struct msghdr *msg, size_t *rdma_bytes)
1044{
1045 struct rds_rdma_args *args;
1046 struct cmsghdr *cmsg;
1047
1048 for_each_cmsghdr(cmsg, msg) {
1049 if (!CMSG_OK(msg, cmsg))
1050 return -EINVAL;
1051
1052 if (cmsg->cmsg_level != SOL_RDS)
1053 continue;
1054
1055 if (cmsg->cmsg_type == RDS_CMSG_RDMA_ARGS) {
Avinash Repaka14e138a2017-12-21 20:17:04 -08001056 if (cmsg->cmsg_len <
1057 CMSG_LEN(sizeof(struct rds_rdma_args)))
1058 return -EINVAL;
Avinash Repakaf9fb69a2016-02-29 15:30:57 -08001059 args = CMSG_DATA(cmsg);
1060 *rdma_bytes += args->remote_vec.bytes;
1061 }
1062 }
1063 return 0;
1064}
1065
Ying Xue1b784142015-03-02 15:37:48 +08001066int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len)
Andy Grover5c115592009-02-24 15:30:27 +00001067{
1068 struct sock *sk = sock->sk;
1069 struct rds_sock *rs = rds_sk_to_rs(sk);
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001070 DECLARE_SOCKADDR(struct sockaddr_in6 *, sin6, msg->msg_name);
Steffen Hurrle342dfc32014-01-17 22:53:15 +01001071 DECLARE_SOCKADDR(struct sockaddr_in *, usin, msg->msg_name);
Andy Grover5c115592009-02-24 15:30:27 +00001072 __be16 dport;
1073 struct rds_message *rm = NULL;
1074 struct rds_connection *conn;
1075 int ret = 0;
1076 int queued = 0, allocated_mr = 0;
1077 int nonblock = msg->msg_flags & MSG_DONTWAIT;
Andy Grover1123fd72010-03-11 13:49:56 +00001078 long timeo = sock_sndtimeo(sk, nonblock);
Sowmini Varadhan780a6d92016-06-13 09:44:33 -07001079 struct rds_conn_path *cpath;
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001080 struct in6_addr daddr;
1081 __u32 scope_id = 0;
Avinash Repakaf9fb69a2016-02-29 15:30:57 -08001082 size_t total_payload_len = payload_len, rdma_payload_len = 0;
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -08001083 bool zcopy = ((msg->msg_flags & MSG_ZEROCOPY) &&
1084 sock_flag(rds_rs_to_sk(rs), SOCK_ZEROCOPY));
1085 int num_sgs = ceil(payload_len, PAGE_SIZE);
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001086 int namelen;
Andy Grover5c115592009-02-24 15:30:27 +00001087
1088 /* Mirror Linux UDP mirror of BSD error message compatibility */
1089 /* XXX: Perhaps MSG_MORE someday */
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -08001090 if (msg->msg_flags & ~(MSG_DONTWAIT | MSG_CMSG_COMPAT | MSG_ZEROCOPY)) {
Andy Grover5c115592009-02-24 15:30:27 +00001091 ret = -EOPNOTSUPP;
1092 goto out;
1093 }
1094
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001095 namelen = msg->msg_namelen;
1096 if (namelen != 0) {
1097 if (namelen < sizeof(*usin)) {
Andy Grover5c115592009-02-24 15:30:27 +00001098 ret = -EINVAL;
1099 goto out;
1100 }
Ka-Cheong Poon1e2b44e2018-07-23 20:51:22 -07001101 switch (usin->sin_family) {
1102 case AF_INET:
1103 if (usin->sin_addr.s_addr == htonl(INADDR_ANY) ||
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001104 usin->sin_addr.s_addr == htonl(INADDR_BROADCAST) ||
1105 IN_MULTICAST(ntohl(usin->sin_addr.s_addr))) {
1106 ret = -EINVAL;
1107 goto out;
1108 }
1109 ipv6_addr_set_v4mapped(usin->sin_addr.s_addr, &daddr);
1110 dport = usin->sin_port;
1111 break;
1112
Ka-Cheong Poone65d4d92018-07-30 22:48:42 -07001113#if IS_ENABLED(CONFIG_IPV6)
Ka-Cheong Poon1e2b44e2018-07-23 20:51:22 -07001114 case AF_INET6: {
1115 int addr_type;
1116
1117 if (namelen < sizeof(*sin6)) {
1118 ret = -EINVAL;
1119 goto out;
1120 }
1121 addr_type = ipv6_addr_type(&sin6->sin6_addr);
1122 if (!(addr_type & IPV6_ADDR_UNICAST)) {
1123 __be32 addr4;
1124
1125 if (!(addr_type & IPV6_ADDR_MAPPED)) {
1126 ret = -EINVAL;
1127 goto out;
1128 }
1129
1130 /* It is a mapped address. Need to do some
1131 * sanity checks.
1132 */
1133 addr4 = sin6->sin6_addr.s6_addr32[3];
1134 if (addr4 == htonl(INADDR_ANY) ||
1135 addr4 == htonl(INADDR_BROADCAST) ||
1136 IN_MULTICAST(ntohl(addr4))) {
Gustavo A. R. Silvadc66fe42018-07-25 10:22:27 -05001137 ret = -EINVAL;
Ka-Cheong Poon1e2b44e2018-07-23 20:51:22 -07001138 goto out;
1139 }
1140 }
1141 if (addr_type & IPV6_ADDR_LINKLOCAL) {
1142 if (sin6->sin6_scope_id == 0) {
1143 ret = -EINVAL;
1144 goto out;
1145 }
1146 scope_id = sin6->sin6_scope_id;
1147 }
1148
1149 daddr = sin6->sin6_addr;
1150 dport = sin6->sin6_port;
1151 break;
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001152 }
Ka-Cheong Poone65d4d92018-07-30 22:48:42 -07001153#endif
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001154
1155 default:
1156 ret = -EINVAL;
1157 goto out;
1158 }
Andy Grover5c115592009-02-24 15:30:27 +00001159 } else {
1160 /* We only care about consistency with ->connect() */
1161 lock_sock(sk);
1162 daddr = rs->rs_conn_addr;
1163 dport = rs->rs_conn_port;
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001164 scope_id = rs->rs_bound_scope_id;
Andy Grover5c115592009-02-24 15:30:27 +00001165 release_sock(sk);
1166 }
1167
Quentin Casasnovas8c7188b2015-11-24 17:13:21 -05001168 lock_sock(sk);
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001169 if (ipv6_addr_any(&rs->rs_bound_addr) || ipv6_addr_any(&daddr)) {
Quentin Casasnovas8c7188b2015-11-24 17:13:21 -05001170 release_sock(sk);
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001171 ret = -ENOTCONN;
Andy Grover5c115592009-02-24 15:30:27 +00001172 goto out;
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001173 } else if (namelen != 0) {
1174 /* Cannot send to an IPv4 address using an IPv6 source
1175 * address and cannot send to an IPv6 address using an
1176 * IPv4 source address.
1177 */
1178 if (ipv6_addr_v4mapped(&daddr) ^
1179 ipv6_addr_v4mapped(&rs->rs_bound_addr)) {
1180 release_sock(sk);
1181 ret = -EOPNOTSUPP;
1182 goto out;
1183 }
Ka-Cheong Poon1e2b44e2018-07-23 20:51:22 -07001184 /* If the socket is already bound to a link local address,
1185 * it can only send to peers on the same link. But allow
1186 * communicating beween link local and non-link local address.
1187 */
1188 if (scope_id != rs->rs_bound_scope_id) {
1189 if (!scope_id) {
1190 scope_id = rs->rs_bound_scope_id;
1191 } else if (rs->rs_bound_scope_id) {
1192 release_sock(sk);
1193 ret = -EINVAL;
1194 goto out;
1195 }
1196 }
Andy Grover5c115592009-02-24 15:30:27 +00001197 }
Quentin Casasnovas8c7188b2015-11-24 17:13:21 -05001198 release_sock(sk);
Andy Grover5c115592009-02-24 15:30:27 +00001199
Avinash Repakaf9fb69a2016-02-29 15:30:57 -08001200 ret = rds_rdma_bytes(msg, &rdma_payload_len);
1201 if (ret)
1202 goto out;
1203
1204 total_payload_len += rdma_payload_len;
1205 if (max_t(size_t, payload_len, rdma_payload_len) > RDS_MAX_MSG_SIZE) {
1206 ret = -EMSGSIZE;
1207 goto out;
1208 }
1209
Mukesh Kacker06e89412015-08-22 15:45:34 -07001210 if (payload_len > rds_sk_sndbuf(rs)) {
1211 ret = -EMSGSIZE;
1212 goto out;
1213 }
1214
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -08001215 if (zcopy) {
1216 if (rs->rs_transport->t_type != RDS_TRANS_TCP) {
1217 ret = -EOPNOTSUPP;
1218 goto out;
1219 }
1220 num_sgs = iov_iter_npages(&msg->msg_iter, INT_MAX);
1221 }
Andy Groverfc445082010-01-12 12:56:06 -08001222 /* size of rm including all sgs */
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -08001223 ret = rds_rm_size(msg, num_sgs);
Andy Groverfc445082010-01-12 12:56:06 -08001224 if (ret < 0)
1225 goto out;
1226
1227 rm = rds_message_alloc(ret, GFP_KERNEL);
1228 if (!rm) {
1229 ret = -ENOMEM;
Andy Grover5c115592009-02-24 15:30:27 +00001230 goto out;
1231 }
1232
Andy Grover372cd7d2010-02-03 19:40:32 -08001233 /* Attach data to the rm */
1234 if (payload_len) {
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -08001235 rm->data.op_sg = rds_message_alloc_sgs(rm, num_sgs);
Andy Groverd139ff02010-10-28 15:40:59 +00001236 if (!rm->data.op_sg) {
1237 ret = -ENOMEM;
1238 goto out;
1239 }
Sowmini Varadhan0cebacc2018-02-15 10:49:36 -08001240 ret = rds_message_copy_from_user(rm, &msg->msg_iter, zcopy);
Andy Grover372cd7d2010-02-03 19:40:32 -08001241 if (ret)
1242 goto out;
1243 }
1244 rm->data.op_active = 1;
Andy Groverfc445082010-01-12 12:56:06 -08001245
Andy Grover5c115592009-02-24 15:30:27 +00001246 rm->m_daddr = daddr;
1247
Andy Grover5c115592009-02-24 15:30:27 +00001248 /* rds_conn_create has a spinlock that runs with IRQ off.
1249 * Caching the conn in the socket helps a lot. */
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001250 if (rs->rs_conn && ipv6_addr_equal(&rs->rs_conn->c_faddr, &daddr))
Andy Grover5c115592009-02-24 15:30:27 +00001251 conn = rs->rs_conn;
1252 else {
Sowmini Varadhand5a8ac22015-08-05 01:43:25 -04001253 conn = rds_conn_create_outgoing(sock_net(sock->sk),
Ka-Cheong Pooneee2fa62018-07-23 20:51:21 -07001254 &rs->rs_bound_addr, &daddr,
1255 rs->rs_transport,
1256 sock->sk->sk_allocation,
1257 scope_id);
Andy Grover5c115592009-02-24 15:30:27 +00001258 if (IS_ERR(conn)) {
1259 ret = PTR_ERR(conn);
1260 goto out;
1261 }
1262 rs->rs_conn = conn;
1263 }
1264
Avinash Repaka9e630bc2018-07-24 20:31:58 -07001265 if (conn->c_trans->t_mp_capable)
Ka-Cheong Poon9a4890b2018-10-08 09:17:11 -07001266 cpath = &conn->c_path[rds_send_mprds_hash(rs, conn, nonblock)];
Avinash Repaka9e630bc2018-07-24 20:31:58 -07001267 else
1268 cpath = &conn->c_path[0];
1269
1270 rm->m_conn_path = cpath;
1271
Andy Grover49f69692009-04-09 14:09:41 +00001272 /* Parse any control messages the user may have included. */
1273 ret = rds_cmsg_send(rs, rm, msg, &allocated_mr);
Santosh Shilimkar584a8272016-07-04 17:04:37 -07001274 if (ret) {
1275 /* Trigger connection so that its ready for the next retry */
1276 if (ret == -EAGAIN)
1277 rds_conn_connect_if_down(conn);
Andy Grover49f69692009-04-09 14:09:41 +00001278 goto out;
Santosh Shilimkar584a8272016-07-04 17:04:37 -07001279 }
Andy Grover49f69692009-04-09 14:09:41 +00001280
Andy Grover2c3a5f92010-03-01 16:10:40 -08001281 if (rm->rdma.op_active && !conn->c_trans->xmit_rdma) {
Manuel Zerpiescb0a6052011-06-16 02:09:57 +00001282 printk_ratelimited(KERN_NOTICE "rdma_op %p conn xmit_rdma %p\n",
Andy Groverf8b3aaf2010-03-01 14:11:53 -08001283 &rm->rdma, conn->c_trans->xmit_rdma);
Andy Grover15133f62010-01-12 14:33:38 -08001284 ret = -EOPNOTSUPP;
1285 goto out;
1286 }
1287
1288 if (rm->atomic.op_active && !conn->c_trans->xmit_atomic) {
Manuel Zerpiescb0a6052011-06-16 02:09:57 +00001289 printk_ratelimited(KERN_NOTICE "atomic_op %p conn xmit_atomic %p\n",
Andy Grover15133f62010-01-12 14:33:38 -08001290 &rm->atomic, conn->c_trans->xmit_atomic);
Andy Grover5c115592009-02-24 15:30:27 +00001291 ret = -EOPNOTSUPP;
1292 goto out;
1293 }
1294
Sowmini Varadhanebeeb1a2018-02-03 04:26:51 -08001295 if (rds_destroy_pending(conn)) {
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -08001296 ret = -EAGAIN;
1297 goto out;
1298 }
1299
Sowmini Varadhan3c0a5902016-06-13 09:44:37 -07001300 rds_conn_path_connect_if_down(cpath);
Andy Grover5c115592009-02-24 15:30:27 +00001301
1302 ret = rds_cong_wait(conn->c_fcong, dport, nonblock, rs);
Andy Groverb98ba522010-03-11 13:50:04 +00001303 if (ret) {
1304 rs->rs_seen_congestion = 1;
Andy Grover5c115592009-02-24 15:30:27 +00001305 goto out;
Andy Groverb98ba522010-03-11 13:50:04 +00001306 }
Sowmini Varadhan780a6d92016-06-13 09:44:33 -07001307 while (!rds_send_queue_rm(rs, conn, cpath, rm, rs->rs_bound_port,
Andy Grover5c115592009-02-24 15:30:27 +00001308 dport, &queued)) {
1309 rds_stats_inc(s_send_queue_full);
Mukesh Kacker06e89412015-08-22 15:45:34 -07001310
Andy Grover5c115592009-02-24 15:30:27 +00001311 if (nonblock) {
1312 ret = -EAGAIN;
1313 goto out;
1314 }
1315
Eric Dumazetaa395142010-04-20 13:03:51 +00001316 timeo = wait_event_interruptible_timeout(*sk_sleep(sk),
Sowmini Varadhan780a6d92016-06-13 09:44:33 -07001317 rds_send_queue_rm(rs, conn, cpath, rm,
Andy Grover5c115592009-02-24 15:30:27 +00001318 rs->rs_bound_port,
1319 dport,
1320 &queued),
1321 timeo);
1322 rdsdebug("sendmsg woke queued %d timeo %ld\n", queued, timeo);
1323 if (timeo > 0 || timeo == MAX_SCHEDULE_TIMEOUT)
1324 continue;
1325
1326 ret = timeo;
1327 if (ret == 0)
1328 ret = -ETIMEDOUT;
1329 goto out;
1330 }
1331
1332 /*
1333 * By now we've committed to the send. We reuse rds_send_worker()
1334 * to retry sends in the rds thread if the transport asks us to.
1335 */
1336 rds_stats_inc(s_send_queued);
1337
Sowmini Varadhan1f9ecd72016-06-13 09:44:34 -07001338 ret = rds_send_xmit(cpath);
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -08001339 if (ret == -ENOMEM || ret == -EAGAIN) {
1340 ret = 0;
1341 rcu_read_lock();
Sowmini Varadhanebeeb1a2018-02-03 04:26:51 -08001342 if (rds_destroy_pending(cpath->cp_conn))
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -08001343 ret = -ENETUNREACH;
1344 else
1345 queue_delayed_work(rds_wq, &cpath->cp_send_w, 1);
1346 rcu_read_unlock();
1347 }
1348 if (ret)
1349 goto out;
Andy Grover5c115592009-02-24 15:30:27 +00001350 rds_message_put(rm);
1351 return payload_len;
1352
1353out:
1354 /* If the user included a RDMA_MAP cmsg, we allocated a MR on the fly.
1355 * If the sendmsg goes through, we keep the MR. If it fails with EAGAIN
1356 * or in any other way, we need to destroy the MR again */
1357 if (allocated_mr)
1358 rds_rdma_unuse(rs, rds_rdma_cookie_key(rm->m_rdma_cookie), 1);
1359
1360 if (rm)
1361 rds_message_put(rm);
1362 return ret;
1363}
1364
1365/*
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001366 * send out a probe. Can be shared by rds_send_ping,
1367 * rds_send_pong, rds_send_hb.
1368 * rds_send_hb should use h_flags
1369 * RDS_FLAG_HB_PING|RDS_FLAG_ACK_REQUIRED
1370 * or
1371 * RDS_FLAG_HB_PONG|RDS_FLAG_ACK_REQUIRED
Andy Grover5c115592009-02-24 15:30:27 +00001372 */
Santosh Shilimkarbb789762016-12-04 16:41:29 -08001373static int
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001374rds_send_probe(struct rds_conn_path *cp, __be16 sport,
1375 __be16 dport, u8 h_flags)
Andy Grover5c115592009-02-24 15:30:27 +00001376{
1377 struct rds_message *rm;
1378 unsigned long flags;
1379 int ret = 0;
1380
1381 rm = rds_message_alloc(0, GFP_ATOMIC);
Andy Grover8690bfa2010-01-12 11:56:44 -08001382 if (!rm) {
Andy Grover5c115592009-02-24 15:30:27 +00001383 ret = -ENOMEM;
1384 goto out;
1385 }
1386
Sowmini Varadhan45997e92016-06-13 09:44:36 -07001387 rm->m_daddr = cp->cp_conn->c_faddr;
Andy Groveracfcd4d2010-03-31 18:56:25 -07001388 rm->data.op_active = 1;
Andy Grover5c115592009-02-24 15:30:27 +00001389
Sowmini Varadhan3c0a5902016-06-13 09:44:37 -07001390 rds_conn_path_connect_if_down(cp);
Andy Grover5c115592009-02-24 15:30:27 +00001391
Sowmini Varadhan45997e92016-06-13 09:44:36 -07001392 ret = rds_cong_wait(cp->cp_conn->c_fcong, dport, 1, NULL);
Andy Grover5c115592009-02-24 15:30:27 +00001393 if (ret)
1394 goto out;
1395
Sowmini Varadhan45997e92016-06-13 09:44:36 -07001396 spin_lock_irqsave(&cp->cp_lock, flags);
1397 list_add_tail(&rm->m_conn_item, &cp->cp_send_queue);
Andy Grover5c115592009-02-24 15:30:27 +00001398 set_bit(RDS_MSG_ON_CONN, &rm->m_flags);
1399 rds_message_addref(rm);
Sowmini Varadhan45997e92016-06-13 09:44:36 -07001400 rm->m_inc.i_conn = cp->cp_conn;
1401 rm->m_inc.i_conn_path = cp;
Andy Grover5c115592009-02-24 15:30:27 +00001402
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001403 rds_message_populate_header(&rm->m_inc.i_hdr, sport, dport,
Sowmini Varadhan45997e92016-06-13 09:44:36 -07001404 cp->cp_next_tx_seq);
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001405 rm->m_inc.i_hdr.h_flags |= h_flags;
Sowmini Varadhan45997e92016-06-13 09:44:36 -07001406 cp->cp_next_tx_seq++;
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001407
Sowmini Varadhan00354de2017-06-15 11:28:54 -07001408 if (RDS_HS_PROBE(be16_to_cpu(sport), be16_to_cpu(dport)) &&
1409 cp->cp_conn->c_trans->t_mp_capable) {
1410 u16 npaths = cpu_to_be16(RDS_MPATH_WORKERS);
1411 u32 my_gen_num = cpu_to_be32(cp->cp_conn->c_my_gen_num);
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001412
1413 rds_message_add_extension(&rm->m_inc.i_hdr,
1414 RDS_EXTHDR_NPATHS, &npaths,
1415 sizeof(npaths));
Sowmini Varadhan905dd412016-11-16 13:29:49 -08001416 rds_message_add_extension(&rm->m_inc.i_hdr,
1417 RDS_EXTHDR_GEN_NUM,
Sowmini Varadhan00354de2017-06-15 11:28:54 -07001418 &my_gen_num,
Sowmini Varadhan905dd412016-11-16 13:29:49 -08001419 sizeof(u32));
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001420 }
Sowmini Varadhan45997e92016-06-13 09:44:36 -07001421 spin_unlock_irqrestore(&cp->cp_lock, flags);
Andy Grover5c115592009-02-24 15:30:27 +00001422
1423 rds_stats_inc(s_send_queued);
1424 rds_stats_inc(s_send_pong);
1425
santosh.shilimkar@oracle.com7b4b0002015-10-16 22:13:21 -04001426 /* schedule the send work on rds_wq */
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -08001427 rcu_read_lock();
Sowmini Varadhanebeeb1a2018-02-03 04:26:51 -08001428 if (!rds_destroy_pending(cp->cp_conn))
Sowmini Varadhan3db6e0d2018-01-04 06:53:00 -08001429 queue_delayed_work(rds_wq, &cp->cp_send_w, 1);
1430 rcu_read_unlock();
Andy Groveracfcd4d2010-03-31 18:56:25 -07001431
Andy Grover5c115592009-02-24 15:30:27 +00001432 rds_message_put(rm);
1433 return 0;
1434
1435out:
1436 if (rm)
1437 rds_message_put(rm);
1438 return ret;
1439}
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001440
1441int
1442rds_send_pong(struct rds_conn_path *cp, __be16 dport)
1443{
1444 return rds_send_probe(cp, 0, dport, 0);
1445}
1446
Sowmini Varadhan69b92b52017-06-21 13:40:12 -07001447void
1448rds_send_ping(struct rds_connection *conn, int cp_index)
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001449{
1450 unsigned long flags;
Sowmini Varadhan69b92b52017-06-21 13:40:12 -07001451 struct rds_conn_path *cp = &conn->c_path[cp_index];
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001452
1453 spin_lock_irqsave(&cp->cp_lock, flags);
1454 if (conn->c_ping_triggered) {
1455 spin_unlock_irqrestore(&cp->cp_lock, flags);
1456 return;
1457 }
1458 conn->c_ping_triggered = 1;
1459 spin_unlock_irqrestore(&cp->cp_lock, flags);
Sowmini Varadhan69b92b52017-06-21 13:40:12 -07001460 rds_send_probe(cp, cpu_to_be16(RDS_FLAG_PROBE_PORT), 0, 0);
Sowmini Varadhan5916e2c2016-07-14 03:51:03 -07001461}
Sowmini Varadhan69b92b52017-06-21 13:40:12 -07001462EXPORT_SYMBOL_GPL(rds_send_ping);