]>
Commit | Line | Data |
---|---|---|
5c115590 AG |
1 | /* |
2 | * Copyright (c) 2006 Oracle. All rights reserved. | |
3 | * | |
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> | |
d9b93842 | 34 | #include <linux/moduleparam.h> |
5a0e3ad6 | 35 | #include <linux/gfp.h> |
5c115590 AG |
36 | #include <net/sock.h> |
37 | #include <linux/in.h> | |
38 | #include <linux/list.h> | |
cb0a6056 | 39 | #include <linux/ratelimit.h> |
bc3b2d7f | 40 | #include <linux/export.h> |
4bebdd7a | 41 | #include <linux/sizes.h> |
5c115590 AG |
42 | |
43 | #include "rds.h" | |
5c115590 AG |
44 | |
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 | */ | |
4bebdd7a | 55 | static int send_batch_count = SZ_1K; |
5c115590 AG |
56 | module_param(send_batch_count, int, 0444); |
57 | MODULE_PARM_DESC(send_batch_count, " batch factor when working the send queue"); | |
58 | ||
ff51bf84 | 59 | static void rds_send_remove_from_sock(struct list_head *messages, int status); |
60 | ||
5c115590 | 61 | /* |
0f4b1c7e ZB |
62 | * Reset the send state. Callers must ensure that this doesn't race with |
63 | * rds_send_xmit(). | |
5c115590 AG |
64 | */ |
65 | void rds_send_reset(struct rds_connection *conn) | |
66 | { | |
67 | struct rds_message *rm, *tmp; | |
68 | unsigned long flags; | |
69 | ||
70 | if (conn->c_xmit_rm) { | |
7e3f2952 CM |
71 | rm = conn->c_xmit_rm; |
72 | conn->c_xmit_rm = NULL; | |
5c115590 AG |
73 | /* 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 */ | |
7e3f2952 | 77 | rds_message_unmapped(rm); |
7e3f2952 | 78 | rds_message_put(rm); |
5c115590 | 79 | } |
7e3f2952 | 80 | |
5c115590 AG |
81 | conn->c_xmit_sg = 0; |
82 | conn->c_xmit_hdr_off = 0; | |
83 | conn->c_xmit_data_off = 0; | |
15133f6e | 84 | conn->c_xmit_atomic_sent = 0; |
5b2366bd AG |
85 | conn->c_xmit_rdma_sent = 0; |
86 | conn->c_xmit_data_sent = 0; | |
5c115590 AG |
87 | |
88 | conn->c_map_queued = 0; | |
89 | ||
90 | conn->c_unacked_packets = rds_sysctl_max_unacked_packets; | |
91 | conn->c_unacked_bytes = rds_sysctl_max_unacked_bytes; | |
92 | ||
93 | /* Mark messages as retransmissions, and move them to the send q */ | |
94 | spin_lock_irqsave(&conn->c_lock, flags); | |
95 | list_for_each_entry_safe(rm, tmp, &conn->c_retrans, m_conn_item) { | |
96 | set_bit(RDS_MSG_ACK_REQUIRED, &rm->m_flags); | |
97 | set_bit(RDS_MSG_RETRANSMITTED, &rm->m_flags); | |
98 | } | |
99 | list_splice_init(&conn->c_retrans, &conn->c_send_queue); | |
100 | spin_unlock_irqrestore(&conn->c_lock, flags); | |
101 | } | |
0b6f760c | 102 | EXPORT_SYMBOL_GPL(rds_send_reset); |
5c115590 | 103 | |
0f4b1c7e ZB |
104 | static int acquire_in_xmit(struct rds_connection *conn) |
105 | { | |
106 | return test_and_set_bit(RDS_IN_XMIT, &conn->c_flags) == 0; | |
107 | } | |
108 | ||
109 | static void release_in_xmit(struct rds_connection *conn) | |
110 | { | |
111 | clear_bit(RDS_IN_XMIT, &conn->c_flags); | |
4e857c58 | 112 | smp_mb__after_atomic(); |
0f4b1c7e ZB |
113 | /* |
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 | */ | |
119 | if (waitqueue_active(&conn->c_waitq)) | |
120 | wake_up_all(&conn->c_waitq); | |
121 | } | |
122 | ||
5c115590 | 123 | /* |
25985edc | 124 | * We're making the conscious trade-off here to only send one message |
5c115590 AG |
125 | * 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 | */ | |
137 | int rds_send_xmit(struct rds_connection *conn) | |
138 | { | |
139 | struct rds_message *rm; | |
140 | unsigned long flags; | |
141 | unsigned int tmp; | |
5c115590 AG |
142 | struct scatterlist *sg; |
143 | int ret = 0; | |
5c115590 | 144 | LIST_HEAD(to_be_dropped); |
443be0e5 SV |
145 | int batch_count; |
146 | unsigned long send_gen = 0; | |
5c115590 | 147 | |
fcc5450c | 148 | restart: |
443be0e5 | 149 | batch_count = 0; |
049ee3f5 | 150 | |
5c115590 AG |
151 | /* |
152 | * sendmsg calls here after having queued its message on the send | |
153 | * queue. We only have one task feeding the connection at a time. If | |
154 | * another thread is already feeding the queue then we back off. This | |
155 | * avoids blocking the caller and trading per-connection data between | |
156 | * caches per message. | |
5c115590 | 157 | */ |
0f4b1c7e | 158 | if (!acquire_in_xmit(conn)) { |
049ee3f5 | 159 | rds_stats_inc(s_send_lock_contention); |
5c115590 AG |
160 | ret = -ENOMEM; |
161 | goto out; | |
162 | } | |
0f4b1c7e | 163 | |
443be0e5 SV |
164 | /* |
165 | * we record the send generation after doing the xmit acquire. | |
166 | * if someone else manages to jump in and do some work, we'll use | |
167 | * this to avoid a goto restart farther down. | |
168 | * | |
169 | * The acquire_in_xmit() check above ensures that only one | |
170 | * caller can increment c_send_gen at any time. | |
171 | */ | |
172 | conn->c_send_gen++; | |
173 | send_gen = conn->c_send_gen; | |
174 | ||
0f4b1c7e ZB |
175 | /* |
176 | * rds_conn_shutdown() sets the conn state and then tests RDS_IN_XMIT, | |
177 | * we do the opposite to avoid races. | |
178 | */ | |
179 | if (!rds_conn_up(conn)) { | |
180 | release_in_xmit(conn); | |
181 | ret = 0; | |
182 | goto out; | |
183 | } | |
5c115590 AG |
184 | |
185 | if (conn->c_trans->xmit_prepare) | |
186 | conn->c_trans->xmit_prepare(conn); | |
187 | ||
188 | /* | |
189 | * spin trying to push headers and data down the connection until | |
5b2366bd | 190 | * the connection doesn't make forward progress. |
5c115590 | 191 | */ |
fcc5450c | 192 | while (1) { |
5c115590 | 193 | |
5c115590 | 194 | rm = conn->c_xmit_rm; |
5c115590 | 195 | |
5b2366bd AG |
196 | /* |
197 | * If between sending messages, we can send a pending congestion | |
198 | * map update. | |
5c115590 | 199 | */ |
8690bfa1 | 200 | if (!rm && test_and_clear_bit(0, &conn->c_map_queued)) { |
77dd550e AG |
201 | rm = rds_cong_update_alloc(conn); |
202 | if (IS_ERR(rm)) { | |
203 | ret = PTR_ERR(rm); | |
204 | break; | |
5b2366bd | 205 | } |
77dd550e AG |
206 | rm->data.op_active = 1; |
207 | ||
208 | conn->c_xmit_rm = rm; | |
5c115590 AG |
209 | } |
210 | ||
211 | /* | |
5b2366bd | 212 | * If not already working on one, grab the next message. |
5c115590 AG |
213 | * |
214 | * c_xmit_rm holds a ref while we're sending this message down | |
215 | * the connction. We can use this ref while holding the | |
216 | * send_sem.. rds_send_reset() is serialized with it. | |
217 | */ | |
8690bfa1 | 218 | if (!rm) { |
5c115590 AG |
219 | unsigned int len; |
220 | ||
443be0e5 SV |
221 | batch_count++; |
222 | ||
223 | /* we want to process as big a batch as we can, but | |
224 | * we also want to avoid softlockups. If we've been | |
225 | * through a lot of messages, lets back off and see | |
226 | * if anyone else jumps in | |
227 | */ | |
4bebdd7a | 228 | if (batch_count >= send_batch_count) |
443be0e5 SV |
229 | goto over_batch; |
230 | ||
0f4b1c7e | 231 | spin_lock_irqsave(&conn->c_lock, flags); |
5c115590 AG |
232 | |
233 | if (!list_empty(&conn->c_send_queue)) { | |
234 | rm = list_entry(conn->c_send_queue.next, | |
235 | struct rds_message, | |
236 | m_conn_item); | |
237 | rds_message_addref(rm); | |
238 | ||
239 | /* | |
240 | * Move the message from the send queue to the retransmit | |
241 | * list right away. | |
242 | */ | |
243 | list_move_tail(&rm->m_conn_item, &conn->c_retrans); | |
244 | } | |
245 | ||
0f4b1c7e | 246 | spin_unlock_irqrestore(&conn->c_lock, flags); |
5c115590 | 247 | |
fcc5450c | 248 | if (!rm) |
5c115590 | 249 | break; |
5c115590 AG |
250 | |
251 | /* Unfortunately, the way Infiniband deals with | |
252 | * RDMA to a bad MR key is by moving the entire | |
253 | * queue pair to error state. We cold possibly | |
254 | * recover from that, but right now we drop the | |
255 | * connection. | |
256 | * Therefore, we never retransmit messages with RDMA ops. | |
257 | */ | |
f8b3aaf2 | 258 | if (rm->rdma.op_active && |
f64f9e71 | 259 | test_bit(RDS_MSG_RETRANSMITTED, &rm->m_flags)) { |
0f4b1c7e | 260 | spin_lock_irqsave(&conn->c_lock, flags); |
5c115590 AG |
261 | if (test_and_clear_bit(RDS_MSG_ON_CONN, &rm->m_flags)) |
262 | list_move(&rm->m_conn_item, &to_be_dropped); | |
0f4b1c7e | 263 | spin_unlock_irqrestore(&conn->c_lock, flags); |
5c115590 AG |
264 | continue; |
265 | } | |
266 | ||
267 | /* Require an ACK every once in a while */ | |
268 | len = ntohl(rm->m_inc.i_hdr.h_len); | |
f64f9e71 JP |
269 | if (conn->c_unacked_packets == 0 || |
270 | conn->c_unacked_bytes < len) { | |
5c115590 AG |
271 | __set_bit(RDS_MSG_ACK_REQUIRED, &rm->m_flags); |
272 | ||
273 | conn->c_unacked_packets = rds_sysctl_max_unacked_packets; | |
274 | conn->c_unacked_bytes = rds_sysctl_max_unacked_bytes; | |
275 | rds_stats_inc(s_send_ack_required); | |
276 | } else { | |
277 | conn->c_unacked_bytes -= len; | |
278 | conn->c_unacked_packets--; | |
279 | } | |
280 | ||
281 | conn->c_xmit_rm = rm; | |
282 | } | |
283 | ||
2c3a5f9a AG |
284 | /* The transport either sends the whole rdma or none of it */ |
285 | if (rm->rdma.op_active && !conn->c_xmit_rdma_sent) { | |
ff3d7d36 | 286 | rm->m_final_op = &rm->rdma; |
4f73113c | 287 | /* The transport owns the mapped memory for now. |
288 | * You can't unmap it while it's on the send queue | |
289 | */ | |
290 | set_bit(RDS_MSG_MAPPED, &rm->m_flags); | |
2c3a5f9a | 291 | ret = conn->c_trans->xmit_rdma(conn, &rm->rdma); |
4f73113c | 292 | if (ret) { |
293 | clear_bit(RDS_MSG_MAPPED, &rm->m_flags); | |
294 | wake_up_interruptible(&rm->m_flush_wait); | |
15133f6e | 295 | break; |
4f73113c | 296 | } |
2c3a5f9a AG |
297 | conn->c_xmit_rdma_sent = 1; |
298 | ||
15133f6e AG |
299 | } |
300 | ||
2c3a5f9a | 301 | if (rm->atomic.op_active && !conn->c_xmit_atomic_sent) { |
ff3d7d36 | 302 | rm->m_final_op = &rm->atomic; |
4f73113c | 303 | /* The transport owns the mapped memory for now. |
304 | * You can't unmap it while it's on the send queue | |
305 | */ | |
306 | set_bit(RDS_MSG_MAPPED, &rm->m_flags); | |
ff3d7d36 | 307 | ret = conn->c_trans->xmit_atomic(conn, &rm->atomic); |
4f73113c | 308 | if (ret) { |
309 | clear_bit(RDS_MSG_MAPPED, &rm->m_flags); | |
310 | wake_up_interruptible(&rm->m_flush_wait); | |
5c115590 | 311 | break; |
4f73113c | 312 | } |
2c3a5f9a | 313 | conn->c_xmit_atomic_sent = 1; |
ff3d7d36 | 314 | |
5c115590 AG |
315 | } |
316 | ||
2c3a5f9a AG |
317 | /* |
318 | * A number of cases require an RDS header to be sent | |
319 | * even if there is no data. | |
320 | * We permit 0-byte sends; rds-ping depends on this. | |
321 | * However, if there are exclusively attached silent ops, | |
322 | * we skip the hdr/data send, to enable silent operation. | |
323 | */ | |
324 | if (rm->data.op_nents == 0) { | |
325 | int ops_present; | |
326 | int all_ops_are_silent = 1; | |
327 | ||
328 | ops_present = (rm->atomic.op_active || rm->rdma.op_active); | |
329 | if (rm->atomic.op_active && !rm->atomic.op_silent) | |
330 | all_ops_are_silent = 0; | |
331 | if (rm->rdma.op_active && !rm->rdma.op_silent) | |
332 | all_ops_are_silent = 0; | |
333 | ||
334 | if (ops_present && all_ops_are_silent | |
335 | && !rm->m_rdma_cookie) | |
336 | rm->data.op_active = 0; | |
337 | } | |
338 | ||
5b2366bd | 339 | if (rm->data.op_active && !conn->c_xmit_data_sent) { |
ff3d7d36 | 340 | rm->m_final_op = &rm->data; |
5c115590 AG |
341 | ret = conn->c_trans->xmit(conn, rm, |
342 | conn->c_xmit_hdr_off, | |
343 | conn->c_xmit_sg, | |
344 | conn->c_xmit_data_off); | |
345 | if (ret <= 0) | |
346 | break; | |
347 | ||
348 | if (conn->c_xmit_hdr_off < sizeof(struct rds_header)) { | |
349 | tmp = min_t(int, ret, | |
350 | sizeof(struct rds_header) - | |
351 | conn->c_xmit_hdr_off); | |
352 | conn->c_xmit_hdr_off += tmp; | |
353 | ret -= tmp; | |
354 | } | |
355 | ||
6c7cc6e4 | 356 | sg = &rm->data.op_sg[conn->c_xmit_sg]; |
5c115590 AG |
357 | while (ret) { |
358 | tmp = min_t(int, ret, sg->length - | |
359 | conn->c_xmit_data_off); | |
360 | conn->c_xmit_data_off += tmp; | |
361 | ret -= tmp; | |
362 | if (conn->c_xmit_data_off == sg->length) { | |
363 | conn->c_xmit_data_off = 0; | |
364 | sg++; | |
365 | conn->c_xmit_sg++; | |
366 | BUG_ON(ret != 0 && | |
6c7cc6e4 | 367 | conn->c_xmit_sg == rm->data.op_nents); |
5c115590 AG |
368 | } |
369 | } | |
5b2366bd AG |
370 | |
371 | if (conn->c_xmit_hdr_off == sizeof(struct rds_header) && | |
372 | (conn->c_xmit_sg == rm->data.op_nents)) | |
373 | conn->c_xmit_data_sent = 1; | |
374 | } | |
375 | ||
376 | /* | |
377 | * A rm will only take multiple times through this loop | |
378 | * if there is a data op. Thus, if the data is sent (or there was | |
379 | * none), then we're done with the rm. | |
380 | */ | |
381 | if (!rm->data.op_active || conn->c_xmit_data_sent) { | |
382 | conn->c_xmit_rm = NULL; | |
383 | conn->c_xmit_sg = 0; | |
384 | conn->c_xmit_hdr_off = 0; | |
385 | conn->c_xmit_data_off = 0; | |
386 | conn->c_xmit_rdma_sent = 0; | |
387 | conn->c_xmit_atomic_sent = 0; | |
388 | conn->c_xmit_data_sent = 0; | |
389 | ||
390 | rds_message_put(rm); | |
5c115590 AG |
391 | } |
392 | } | |
393 | ||
443be0e5 | 394 | over_batch: |
5c115590 AG |
395 | if (conn->c_trans->xmit_complete) |
396 | conn->c_trans->xmit_complete(conn); | |
0f4b1c7e | 397 | release_in_xmit(conn); |
5c115590 | 398 | |
2ad8099b AG |
399 | /* Nuke any messages we decided not to retransmit. */ |
400 | if (!list_empty(&to_be_dropped)) { | |
401 | /* irqs on here, so we can put(), unlike above */ | |
402 | list_for_each_entry(rm, &to_be_dropped, m_conn_item) | |
403 | rds_message_put(rm); | |
404 | rds_send_remove_from_sock(&to_be_dropped, RDS_RDMA_DROPPED); | |
405 | } | |
406 | ||
fcc5450c | 407 | /* |
0f4b1c7e ZB |
408 | * Other senders can queue a message after we last test the send queue |
409 | * but before we clear RDS_IN_XMIT. In that case they'd back off and | |
410 | * not try and send their newly queued message. We need to check the | |
411 | * send queue after having cleared RDS_IN_XMIT so that their message | |
412 | * doesn't get stuck on the send queue. | |
fcc5450c AG |
413 | * |
414 | * If the transport cannot continue (i.e ret != 0), then it must | |
415 | * call us when more room is available, such as from the tx | |
416 | * completion handler. | |
443be0e5 SV |
417 | * |
418 | * We have an extra generation check here so that if someone manages | |
419 | * to jump in after our release_in_xmit, we'll see that they have done | |
420 | * some work and we will skip our goto | |
fcc5450c AG |
421 | */ |
422 | if (ret == 0) { | |
9e29db0e | 423 | smp_mb(); |
0c484240 | 424 | if ((test_bit(0, &conn->c_map_queued) || |
425 | !list_empty(&conn->c_send_queue)) && | |
443be0e5 | 426 | send_gen == conn->c_send_gen) { |
049ee3f5 | 427 | rds_stats_inc(s_send_lock_queue_raced); |
4bebdd7a SS |
428 | if (batch_count < send_batch_count) |
429 | goto restart; | |
430 | queue_delayed_work(rds_wq, &conn->c_send_w, 1); | |
5c115590 | 431 | } |
5c115590 AG |
432 | } |
433 | out: | |
434 | return ret; | |
435 | } | |
0c28c045 | 436 | EXPORT_SYMBOL_GPL(rds_send_xmit); |
5c115590 AG |
437 | |
438 | static void rds_send_sndbuf_remove(struct rds_sock *rs, struct rds_message *rm) | |
439 | { | |
440 | u32 len = be32_to_cpu(rm->m_inc.i_hdr.h_len); | |
441 | ||
442 | assert_spin_locked(&rs->rs_lock); | |
443 | ||
444 | BUG_ON(rs->rs_snd_bytes < len); | |
445 | rs->rs_snd_bytes -= len; | |
446 | ||
447 | if (rs->rs_snd_bytes == 0) | |
448 | rds_stats_inc(s_send_queue_empty); | |
449 | } | |
450 | ||
451 | static inline int rds_send_is_acked(struct rds_message *rm, u64 ack, | |
452 | is_acked_func is_acked) | |
453 | { | |
454 | if (is_acked) | |
455 | return is_acked(rm, ack); | |
456 | return be64_to_cpu(rm->m_inc.i_hdr.h_sequence) <= ack; | |
457 | } | |
458 | ||
5c115590 AG |
459 | /* |
460 | * This is pretty similar to what happens below in the ACK | |
461 | * handling code - except that we call here as soon as we get | |
462 | * the IB send completion on the RDMA op and the accompanying | |
463 | * message. | |
464 | */ | |
465 | void rds_rdma_send_complete(struct rds_message *rm, int status) | |
466 | { | |
467 | struct rds_sock *rs = NULL; | |
f8b3aaf2 | 468 | struct rm_rdma_op *ro; |
5c115590 | 469 | struct rds_notifier *notifier; |
9de0864c | 470 | unsigned long flags; |
5c115590 | 471 | |
9de0864c | 472 | spin_lock_irqsave(&rm->m_rs_lock, flags); |
5c115590 | 473 | |
f8b3aaf2 | 474 | ro = &rm->rdma; |
f64f9e71 | 475 | if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags) && |
f8b3aaf2 AG |
476 | ro->op_active && ro->op_notify && ro->op_notifier) { |
477 | notifier = ro->op_notifier; | |
5c115590 AG |
478 | rs = rm->m_rs; |
479 | sock_hold(rds_rs_to_sk(rs)); | |
480 | ||
481 | notifier->n_status = status; | |
482 | spin_lock(&rs->rs_lock); | |
483 | list_add_tail(¬ifier->n_list, &rs->rs_notify_queue); | |
484 | spin_unlock(&rs->rs_lock); | |
485 | ||
f8b3aaf2 | 486 | ro->op_notifier = NULL; |
5c115590 AG |
487 | } |
488 | ||
9de0864c | 489 | spin_unlock_irqrestore(&rm->m_rs_lock, flags); |
5c115590 AG |
490 | |
491 | if (rs) { | |
492 | rds_wake_sk_sleep(rs); | |
493 | sock_put(rds_rs_to_sk(rs)); | |
494 | } | |
495 | } | |
616b757a | 496 | EXPORT_SYMBOL_GPL(rds_rdma_send_complete); |
5c115590 | 497 | |
15133f6e AG |
498 | /* |
499 | * Just like above, except looks at atomic op | |
500 | */ | |
501 | void rds_atomic_send_complete(struct rds_message *rm, int status) | |
502 | { | |
503 | struct rds_sock *rs = NULL; | |
504 | struct rm_atomic_op *ao; | |
505 | struct rds_notifier *notifier; | |
cf4b7389 | 506 | unsigned long flags; |
15133f6e | 507 | |
cf4b7389 | 508 | spin_lock_irqsave(&rm->m_rs_lock, flags); |
15133f6e AG |
509 | |
510 | ao = &rm->atomic; | |
511 | if (test_bit(RDS_MSG_ON_SOCK, &rm->m_flags) | |
512 | && ao->op_active && ao->op_notify && ao->op_notifier) { | |
513 | notifier = ao->op_notifier; | |
514 | rs = rm->m_rs; | |
515 | sock_hold(rds_rs_to_sk(rs)); | |
516 | ||
517 | notifier->n_status = status; | |
518 | spin_lock(&rs->rs_lock); | |
519 | list_add_tail(¬ifier->n_list, &rs->rs_notify_queue); | |
520 | spin_unlock(&rs->rs_lock); | |
521 | ||
522 | ao->op_notifier = NULL; | |
523 | } | |
524 | ||
cf4b7389 | 525 | spin_unlock_irqrestore(&rm->m_rs_lock, flags); |
15133f6e AG |
526 | |
527 | if (rs) { | |
528 | rds_wake_sk_sleep(rs); | |
529 | sock_put(rds_rs_to_sk(rs)); | |
530 | } | |
531 | } | |
532 | EXPORT_SYMBOL_GPL(rds_atomic_send_complete); | |
533 | ||
5c115590 AG |
534 | /* |
535 | * This is the same as rds_rdma_send_complete except we | |
536 | * don't do any locking - we have all the ingredients (message, | |
537 | * socket, socket lock) and can just move the notifier. | |
538 | */ | |
539 | static inline void | |
940786eb | 540 | __rds_send_complete(struct rds_sock *rs, struct rds_message *rm, int status) |
5c115590 | 541 | { |
f8b3aaf2 | 542 | struct rm_rdma_op *ro; |
940786eb | 543 | struct rm_atomic_op *ao; |
5c115590 | 544 | |
f8b3aaf2 AG |
545 | ro = &rm->rdma; |
546 | if (ro->op_active && ro->op_notify && ro->op_notifier) { | |
547 | ro->op_notifier->n_status = status; | |
548 | list_add_tail(&ro->op_notifier->n_list, &rs->rs_notify_queue); | |
549 | ro->op_notifier = NULL; | |
5c115590 AG |
550 | } |
551 | ||
940786eb AG |
552 | ao = &rm->atomic; |
553 | if (ao->op_active && ao->op_notify && ao->op_notifier) { | |
554 | ao->op_notifier->n_status = status; | |
555 | list_add_tail(&ao->op_notifier->n_list, &rs->rs_notify_queue); | |
556 | ao->op_notifier = NULL; | |
557 | } | |
558 | ||
5c115590 AG |
559 | /* No need to wake the app - caller does this */ |
560 | } | |
561 | ||
562 | /* | |
563 | * This is called from the IB send completion when we detect | |
564 | * a RDMA operation that failed with remote access error. | |
565 | * So speed is not an issue here. | |
566 | */ | |
567 | struct rds_message *rds_send_get_message(struct rds_connection *conn, | |
f8b3aaf2 | 568 | struct rm_rdma_op *op) |
5c115590 AG |
569 | { |
570 | struct rds_message *rm, *tmp, *found = NULL; | |
571 | unsigned long flags; | |
572 | ||
573 | spin_lock_irqsave(&conn->c_lock, flags); | |
574 | ||
575 | list_for_each_entry_safe(rm, tmp, &conn->c_retrans, m_conn_item) { | |
f8b3aaf2 | 576 | if (&rm->rdma == op) { |
5c115590 AG |
577 | atomic_inc(&rm->m_refcount); |
578 | found = rm; | |
579 | goto out; | |
580 | } | |
581 | } | |
582 | ||
583 | list_for_each_entry_safe(rm, tmp, &conn->c_send_queue, m_conn_item) { | |
f8b3aaf2 | 584 | if (&rm->rdma == op) { |
5c115590 AG |
585 | atomic_inc(&rm->m_refcount); |
586 | found = rm; | |
587 | break; | |
588 | } | |
589 | } | |
590 | ||
591 | out: | |
592 | spin_unlock_irqrestore(&conn->c_lock, flags); | |
593 | ||
594 | return found; | |
595 | } | |
616b757a | 596 | EXPORT_SYMBOL_GPL(rds_send_get_message); |
5c115590 AG |
597 | |
598 | /* | |
599 | * This removes messages from the socket's list if they're on it. The list | |
600 | * argument must be private to the caller, we must be able to modify it | |
601 | * without locks. The messages must have a reference held for their | |
602 | * position on the list. This function will drop that reference after | |
603 | * removing the messages from the 'messages' list regardless of if it found | |
604 | * the messages on the socket list or not. | |
605 | */ | |
ff51bf84 | 606 | static void rds_send_remove_from_sock(struct list_head *messages, int status) |
5c115590 | 607 | { |
561c7df6 | 608 | unsigned long flags; |
5c115590 AG |
609 | struct rds_sock *rs = NULL; |
610 | struct rds_message *rm; | |
611 | ||
5c115590 | 612 | while (!list_empty(messages)) { |
561c7df6 AG |
613 | int was_on_sock = 0; |
614 | ||
5c115590 AG |
615 | rm = list_entry(messages->next, struct rds_message, |
616 | m_conn_item); | |
617 | list_del_init(&rm->m_conn_item); | |
618 | ||
619 | /* | |
620 | * If we see this flag cleared then we're *sure* that someone | |
621 | * else beat us to removing it from the sock. If we race | |
622 | * with their flag update we'll get the lock and then really | |
623 | * see that the flag has been cleared. | |
624 | * | |
625 | * The message spinlock makes sure nobody clears rm->m_rs | |
626 | * while we're messing with it. It does not prevent the | |
627 | * message from being removed from the socket, though. | |
628 | */ | |
561c7df6 | 629 | spin_lock_irqsave(&rm->m_rs_lock, flags); |
5c115590 AG |
630 | if (!test_bit(RDS_MSG_ON_SOCK, &rm->m_flags)) |
631 | goto unlock_and_drop; | |
632 | ||
633 | if (rs != rm->m_rs) { | |
634 | if (rs) { | |
5c115590 AG |
635 | rds_wake_sk_sleep(rs); |
636 | sock_put(rds_rs_to_sk(rs)); | |
637 | } | |
638 | rs = rm->m_rs; | |
593cbb3e HK |
639 | if (rs) |
640 | sock_hold(rds_rs_to_sk(rs)); | |
5c115590 | 641 | } |
593cbb3e HK |
642 | if (!rs) |
643 | goto unlock_and_drop; | |
048c15e6 | 644 | spin_lock(&rs->rs_lock); |
5c115590 AG |
645 | |
646 | if (test_and_clear_bit(RDS_MSG_ON_SOCK, &rm->m_flags)) { | |
f8b3aaf2 | 647 | struct rm_rdma_op *ro = &rm->rdma; |
5c115590 AG |
648 | struct rds_notifier *notifier; |
649 | ||
650 | list_del_init(&rm->m_sock_item); | |
651 | rds_send_sndbuf_remove(rs, rm); | |
652 | ||
f8b3aaf2 AG |
653 | if (ro->op_active && ro->op_notifier && |
654 | (ro->op_notify || (ro->op_recverr && status))) { | |
655 | notifier = ro->op_notifier; | |
5c115590 AG |
656 | list_add_tail(¬ifier->n_list, |
657 | &rs->rs_notify_queue); | |
658 | if (!notifier->n_status) | |
659 | notifier->n_status = status; | |
f8b3aaf2 | 660 | rm->rdma.op_notifier = NULL; |
5c115590 | 661 | } |
561c7df6 | 662 | was_on_sock = 1; |
5c115590 AG |
663 | rm->m_rs = NULL; |
664 | } | |
048c15e6 | 665 | spin_unlock(&rs->rs_lock); |
5c115590 AG |
666 | |
667 | unlock_and_drop: | |
561c7df6 | 668 | spin_unlock_irqrestore(&rm->m_rs_lock, flags); |
5c115590 | 669 | rds_message_put(rm); |
561c7df6 AG |
670 | if (was_on_sock) |
671 | rds_message_put(rm); | |
5c115590 AG |
672 | } |
673 | ||
674 | if (rs) { | |
5c115590 AG |
675 | rds_wake_sk_sleep(rs); |
676 | sock_put(rds_rs_to_sk(rs)); | |
677 | } | |
5c115590 AG |
678 | } |
679 | ||
680 | /* | |
681 | * Transports call here when they've determined that the receiver queued | |
682 | * messages up to, and including, the given sequence number. Messages are | |
683 | * moved to the retrans queue when rds_send_xmit picks them off the send | |
684 | * queue. This means that in the TCP case, the message may not have been | |
685 | * assigned the m_ack_seq yet - but that's fine as long as tcp_is_acked | |
686 | * checks the RDS_MSG_HAS_ACK_SEQ bit. | |
5c115590 AG |
687 | */ |
688 | void rds_send_drop_acked(struct rds_connection *conn, u64 ack, | |
689 | is_acked_func is_acked) | |
690 | { | |
691 | struct rds_message *rm, *tmp; | |
692 | unsigned long flags; | |
693 | LIST_HEAD(list); | |
694 | ||
695 | spin_lock_irqsave(&conn->c_lock, flags); | |
696 | ||
697 | list_for_each_entry_safe(rm, tmp, &conn->c_retrans, m_conn_item) { | |
698 | if (!rds_send_is_acked(rm, ack, is_acked)) | |
699 | break; | |
700 | ||
701 | list_move(&rm->m_conn_item, &list); | |
702 | clear_bit(RDS_MSG_ON_CONN, &rm->m_flags); | |
703 | } | |
704 | ||
705 | /* order flag updates with spin locks */ | |
706 | if (!list_empty(&list)) | |
4e857c58 | 707 | smp_mb__after_atomic(); |
5c115590 AG |
708 | |
709 | spin_unlock_irqrestore(&conn->c_lock, flags); | |
710 | ||
711 | /* now remove the messages from the sock list as needed */ | |
712 | rds_send_remove_from_sock(&list, RDS_RDMA_SUCCESS); | |
713 | } | |
616b757a | 714 | EXPORT_SYMBOL_GPL(rds_send_drop_acked); |
5c115590 AG |
715 | |
716 | void rds_send_drop_to(struct rds_sock *rs, struct sockaddr_in *dest) | |
717 | { | |
718 | struct rds_message *rm, *tmp; | |
719 | struct rds_connection *conn; | |
7c82eaf0 | 720 | unsigned long flags; |
5c115590 | 721 | LIST_HEAD(list); |
5c115590 AG |
722 | |
723 | /* get all the messages we're dropping under the rs lock */ | |
724 | spin_lock_irqsave(&rs->rs_lock, flags); | |
725 | ||
726 | list_for_each_entry_safe(rm, tmp, &rs->rs_send_queue, m_sock_item) { | |
727 | if (dest && (dest->sin_addr.s_addr != rm->m_daddr || | |
728 | dest->sin_port != rm->m_inc.i_hdr.h_dport)) | |
729 | continue; | |
730 | ||
5c115590 AG |
731 | list_move(&rm->m_sock_item, &list); |
732 | rds_send_sndbuf_remove(rs, rm); | |
733 | clear_bit(RDS_MSG_ON_SOCK, &rm->m_flags); | |
5c115590 AG |
734 | } |
735 | ||
736 | /* order flag updates with the rs lock */ | |
4e857c58 | 737 | smp_mb__after_atomic(); |
5c115590 AG |
738 | |
739 | spin_unlock_irqrestore(&rs->rs_lock, flags); | |
740 | ||
7c82eaf0 AG |
741 | if (list_empty(&list)) |
742 | return; | |
5c115590 | 743 | |
7c82eaf0 | 744 | /* Remove the messages from the conn */ |
5c115590 | 745 | list_for_each_entry(rm, &list, m_sock_item) { |
7c82eaf0 AG |
746 | |
747 | conn = rm->m_inc.i_conn; | |
5c115590 | 748 | |
9de0864c | 749 | spin_lock_irqsave(&conn->c_lock, flags); |
5c115590 | 750 | /* |
7c82eaf0 AG |
751 | * Maybe someone else beat us to removing rm from the conn. |
752 | * If we race with their flag update we'll get the lock and | |
753 | * then really see that the flag has been cleared. | |
5c115590 | 754 | */ |
7c82eaf0 AG |
755 | if (!test_and_clear_bit(RDS_MSG_ON_CONN, &rm->m_flags)) { |
756 | spin_unlock_irqrestore(&conn->c_lock, flags); | |
593cbb3e HK |
757 | spin_lock_irqsave(&rm->m_rs_lock, flags); |
758 | rm->m_rs = NULL; | |
759 | spin_unlock_irqrestore(&rm->m_rs_lock, flags); | |
5c115590 | 760 | continue; |
5c115590 | 761 | } |
9de0864c AG |
762 | list_del_init(&rm->m_conn_item); |
763 | spin_unlock_irqrestore(&conn->c_lock, flags); | |
5c115590 | 764 | |
7c82eaf0 AG |
765 | /* |
766 | * Couldn't grab m_rs_lock in top loop (lock ordering), | |
767 | * but we can now. | |
768 | */ | |
9de0864c | 769 | spin_lock_irqsave(&rm->m_rs_lock, flags); |
5c115590 | 770 | |
7c82eaf0 | 771 | spin_lock(&rs->rs_lock); |
940786eb | 772 | __rds_send_complete(rs, rm, RDS_RDMA_CANCELED); |
7c82eaf0 AG |
773 | spin_unlock(&rs->rs_lock); |
774 | ||
775 | rm->m_rs = NULL; | |
9de0864c | 776 | spin_unlock_irqrestore(&rm->m_rs_lock, flags); |
7c82eaf0 | 777 | |
7c82eaf0 | 778 | rds_message_put(rm); |
7c82eaf0 | 779 | } |
5c115590 | 780 | |
7c82eaf0 | 781 | rds_wake_sk_sleep(rs); |
550a8002 | 782 | |
5c115590 AG |
783 | while (!list_empty(&list)) { |
784 | rm = list_entry(list.next, struct rds_message, m_sock_item); | |
785 | list_del_init(&rm->m_sock_item); | |
5c115590 | 786 | rds_message_wait(rm); |
dfcec251 | 787 | |
788 | /* just in case the code above skipped this message | |
789 | * because RDS_MSG_ON_CONN wasn't set, run it again here | |
790 | * taking m_rs_lock is the only thing that keeps us | |
791 | * from racing with ack processing. | |
792 | */ | |
793 | spin_lock_irqsave(&rm->m_rs_lock, flags); | |
794 | ||
795 | spin_lock(&rs->rs_lock); | |
796 | __rds_send_complete(rs, rm, RDS_RDMA_CANCELED); | |
797 | spin_unlock(&rs->rs_lock); | |
798 | ||
799 | rm->m_rs = NULL; | |
800 | spin_unlock_irqrestore(&rm->m_rs_lock, flags); | |
801 | ||
5c115590 AG |
802 | rds_message_put(rm); |
803 | } | |
804 | } | |
805 | ||
806 | /* | |
807 | * we only want this to fire once so we use the callers 'queued'. It's | |
808 | * possible that another thread can race with us and remove the | |
809 | * message from the flow with RDS_CANCEL_SENT_TO. | |
810 | */ | |
811 | static int rds_send_queue_rm(struct rds_sock *rs, struct rds_connection *conn, | |
812 | struct rds_message *rm, __be16 sport, | |
813 | __be16 dport, int *queued) | |
814 | { | |
815 | unsigned long flags; | |
816 | u32 len; | |
817 | ||
818 | if (*queued) | |
819 | goto out; | |
820 | ||
821 | len = be32_to_cpu(rm->m_inc.i_hdr.h_len); | |
822 | ||
823 | /* this is the only place which holds both the socket's rs_lock | |
824 | * and the connection's c_lock */ | |
825 | spin_lock_irqsave(&rs->rs_lock, flags); | |
826 | ||
827 | /* | |
828 | * If there is a little space in sndbuf, we don't queue anything, | |
829 | * and userspace gets -EAGAIN. But poll() indicates there's send | |
830 | * room. This can lead to bad behavior (spinning) if snd_bytes isn't | |
831 | * freed up by incoming acks. So we check the *old* value of | |
832 | * rs_snd_bytes here to allow the last msg to exceed the buffer, | |
833 | * and poll() now knows no more data can be sent. | |
834 | */ | |
835 | if (rs->rs_snd_bytes < rds_sk_sndbuf(rs)) { | |
836 | rs->rs_snd_bytes += len; | |
837 | ||
838 | /* let recv side know we are close to send space exhaustion. | |
839 | * This is probably not the optimal way to do it, as this | |
840 | * means we set the flag on *all* messages as soon as our | |
841 | * throughput hits a certain threshold. | |
842 | */ | |
843 | if (rs->rs_snd_bytes >= rds_sk_sndbuf(rs) / 2) | |
844 | __set_bit(RDS_MSG_ACK_REQUIRED, &rm->m_flags); | |
845 | ||
846 | list_add_tail(&rm->m_sock_item, &rs->rs_send_queue); | |
847 | set_bit(RDS_MSG_ON_SOCK, &rm->m_flags); | |
848 | rds_message_addref(rm); | |
849 | rm->m_rs = rs; | |
850 | ||
851 | /* The code ordering is a little weird, but we're | |
852 | trying to minimize the time we hold c_lock */ | |
853 | rds_message_populate_header(&rm->m_inc.i_hdr, sport, dport, 0); | |
854 | rm->m_inc.i_conn = conn; | |
855 | rds_message_addref(rm); | |
856 | ||
857 | spin_lock(&conn->c_lock); | |
858 | rm->m_inc.i_hdr.h_sequence = cpu_to_be64(conn->c_next_tx_seq++); | |
859 | list_add_tail(&rm->m_conn_item, &conn->c_send_queue); | |
860 | set_bit(RDS_MSG_ON_CONN, &rm->m_flags); | |
861 | spin_unlock(&conn->c_lock); | |
862 | ||
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); | |
871 | out: | |
872 | return *queued; | |
873 | } | |
874 | ||
fc445084 AG |
875 | /* |
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 | */ | |
879 | static int rds_rm_size(struct msghdr *msg, int data_len) | |
880 | { | |
ff87e97a | 881 | struct cmsghdr *cmsg; |
fc445084 | 882 | int size = 0; |
aa0a4ef4 | 883 | int cmsg_groups = 0; |
ff87e97a AG |
884 | int retval; |
885 | ||
f95b414e | 886 | for_each_cmsghdr(cmsg, msg) { |
ff87e97a AG |
887 | if (!CMSG_OK(msg, cmsg)) |
888 | return -EINVAL; | |
889 | ||
890 | if (cmsg->cmsg_level != SOL_RDS) | |
891 | continue; | |
892 | ||
893 | switch (cmsg->cmsg_type) { | |
894 | case RDS_CMSG_RDMA_ARGS: | |
aa0a4ef4 | 895 | cmsg_groups |= 1; |
ff87e97a AG |
896 | retval = rds_rdma_extra_size(CMSG_DATA(cmsg)); |
897 | if (retval < 0) | |
898 | return retval; | |
899 | size += retval; | |
aa0a4ef4 | 900 | |
ff87e97a AG |
901 | break; |
902 | ||
903 | case RDS_CMSG_RDMA_DEST: | |
904 | case RDS_CMSG_RDMA_MAP: | |
aa0a4ef4 | 905 | cmsg_groups |= 2; |
ff87e97a AG |
906 | /* these are valid but do no add any size */ |
907 | break; | |
908 | ||
15133f6e AG |
909 | case RDS_CMSG_ATOMIC_CSWP: |
910 | case RDS_CMSG_ATOMIC_FADD: | |
20c72bd5 AG |
911 | case RDS_CMSG_MASKED_ATOMIC_CSWP: |
912 | case RDS_CMSG_MASKED_ATOMIC_FADD: | |
aa0a4ef4 | 913 | cmsg_groups |= 1; |
15133f6e AG |
914 | size += sizeof(struct scatterlist); |
915 | break; | |
916 | ||
ff87e97a AG |
917 | default: |
918 | return -EINVAL; | |
919 | } | |
920 | ||
921 | } | |
fc445084 | 922 | |
ff87e97a | 923 | size += ceil(data_len, PAGE_SIZE) * sizeof(struct scatterlist); |
fc445084 | 924 | |
aa0a4ef4 AG |
925 | /* Ensure (DEST, MAP) are never used with (ARGS, ATOMIC) */ |
926 | if (cmsg_groups == 3) | |
927 | return -EINVAL; | |
928 | ||
fc445084 AG |
929 | return size; |
930 | } | |
931 | ||
5c115590 AG |
932 | static int rds_cmsg_send(struct rds_sock *rs, struct rds_message *rm, |
933 | struct msghdr *msg, int *allocated_mr) | |
934 | { | |
935 | struct cmsghdr *cmsg; | |
936 | int ret = 0; | |
937 | ||
f95b414e | 938 | for_each_cmsghdr(cmsg, msg) { |
5c115590 AG |
939 | if (!CMSG_OK(msg, cmsg)) |
940 | return -EINVAL; | |
941 | ||
942 | if (cmsg->cmsg_level != SOL_RDS) | |
943 | continue; | |
944 | ||
945 | /* As a side effect, RDMA_DEST and RDMA_MAP will set | |
15133f6e | 946 | * rm->rdma.m_rdma_cookie and rm->rdma.m_rdma_mr. |
5c115590 AG |
947 | */ |
948 | switch (cmsg->cmsg_type) { | |
949 | case RDS_CMSG_RDMA_ARGS: | |
950 | ret = rds_cmsg_rdma_args(rs, rm, cmsg); | |
951 | break; | |
952 | ||
953 | case RDS_CMSG_RDMA_DEST: | |
954 | ret = rds_cmsg_rdma_dest(rs, rm, cmsg); | |
955 | break; | |
956 | ||
957 | case RDS_CMSG_RDMA_MAP: | |
958 | ret = rds_cmsg_rdma_map(rs, rm, cmsg); | |
959 | if (!ret) | |
960 | *allocated_mr = 1; | |
961 | break; | |
15133f6e AG |
962 | case RDS_CMSG_ATOMIC_CSWP: |
963 | case RDS_CMSG_ATOMIC_FADD: | |
20c72bd5 AG |
964 | case RDS_CMSG_MASKED_ATOMIC_CSWP: |
965 | case RDS_CMSG_MASKED_ATOMIC_FADD: | |
15133f6e AG |
966 | ret = rds_cmsg_atomic(rs, rm, cmsg); |
967 | break; | |
5c115590 AG |
968 | |
969 | default: | |
970 | return -EINVAL; | |
971 | } | |
972 | ||
973 | if (ret) | |
974 | break; | |
975 | } | |
976 | ||
977 | return ret; | |
978 | } | |
979 | ||
1b784140 | 980 | int rds_sendmsg(struct socket *sock, struct msghdr *msg, size_t payload_len) |
5c115590 AG |
981 | { |
982 | struct sock *sk = sock->sk; | |
983 | struct rds_sock *rs = rds_sk_to_rs(sk); | |
342dfc30 | 984 | DECLARE_SOCKADDR(struct sockaddr_in *, usin, msg->msg_name); |
5c115590 AG |
985 | __be32 daddr; |
986 | __be16 dport; | |
987 | struct rds_message *rm = NULL; | |
988 | struct rds_connection *conn; | |
989 | int ret = 0; | |
990 | int queued = 0, allocated_mr = 0; | |
991 | int nonblock = msg->msg_flags & MSG_DONTWAIT; | |
1123fd73 | 992 | long timeo = sock_sndtimeo(sk, nonblock); |
5c115590 AG |
993 | |
994 | /* Mirror Linux UDP mirror of BSD error message compatibility */ | |
995 | /* XXX: Perhaps MSG_MORE someday */ | |
996 | if (msg->msg_flags & ~(MSG_DONTWAIT | MSG_CMSG_COMPAT)) { | |
5c115590 AG |
997 | ret = -EOPNOTSUPP; |
998 | goto out; | |
999 | } | |
1000 | ||
1001 | if (msg->msg_namelen) { | |
1002 | /* XXX fail non-unicast destination IPs? */ | |
1003 | if (msg->msg_namelen < sizeof(*usin) || usin->sin_family != AF_INET) { | |
1004 | ret = -EINVAL; | |
1005 | goto out; | |
1006 | } | |
1007 | daddr = usin->sin_addr.s_addr; | |
1008 | dport = usin->sin_port; | |
1009 | } else { | |
1010 | /* We only care about consistency with ->connect() */ | |
1011 | lock_sock(sk); | |
1012 | daddr = rs->rs_conn_addr; | |
1013 | dport = rs->rs_conn_port; | |
1014 | release_sock(sk); | |
1015 | } | |
1016 | ||
8c7188b2 | 1017 | lock_sock(sk); |
5c115590 | 1018 | if (daddr == 0 || rs->rs_bound_addr == 0) { |
8c7188b2 | 1019 | release_sock(sk); |
5c115590 AG |
1020 | ret = -ENOTCONN; /* XXX not a great errno */ |
1021 | goto out; | |
1022 | } | |
8c7188b2 | 1023 | release_sock(sk); |
5c115590 | 1024 | |
06e8941e MK |
1025 | if (payload_len > rds_sk_sndbuf(rs)) { |
1026 | ret = -EMSGSIZE; | |
1027 | goto out; | |
1028 | } | |
1029 | ||
fc445084 AG |
1030 | /* size of rm including all sgs */ |
1031 | ret = rds_rm_size(msg, payload_len); | |
1032 | if (ret < 0) | |
1033 | goto out; | |
1034 | ||
1035 | rm = rds_message_alloc(ret, GFP_KERNEL); | |
1036 | if (!rm) { | |
1037 | ret = -ENOMEM; | |
5c115590 AG |
1038 | goto out; |
1039 | } | |
1040 | ||
372cd7de AG |
1041 | /* Attach data to the rm */ |
1042 | if (payload_len) { | |
1043 | rm->data.op_sg = rds_message_alloc_sgs(rm, ceil(payload_len, PAGE_SIZE)); | |
d139ff09 AG |
1044 | if (!rm->data.op_sg) { |
1045 | ret = -ENOMEM; | |
1046 | goto out; | |
1047 | } | |
c0371da6 | 1048 | ret = rds_message_copy_from_user(rm, &msg->msg_iter); |
372cd7de AG |
1049 | if (ret) |
1050 | goto out; | |
1051 | } | |
1052 | rm->data.op_active = 1; | |
fc445084 | 1053 | |
5c115590 AG |
1054 | rm->m_daddr = daddr; |
1055 | ||
5c115590 AG |
1056 | /* rds_conn_create has a spinlock that runs with IRQ off. |
1057 | * Caching the conn in the socket helps a lot. */ | |
1058 | if (rs->rs_conn && rs->rs_conn->c_faddr == daddr) | |
1059 | conn = rs->rs_conn; | |
1060 | else { | |
d5a8ac28 SV |
1061 | conn = rds_conn_create_outgoing(sock_net(sock->sk), |
1062 | rs->rs_bound_addr, daddr, | |
5c115590 AG |
1063 | rs->rs_transport, |
1064 | sock->sk->sk_allocation); | |
1065 | if (IS_ERR(conn)) { | |
1066 | ret = PTR_ERR(conn); | |
1067 | goto out; | |
1068 | } | |
1069 | rs->rs_conn = conn; | |
1070 | } | |
1071 | ||
49f69691 AG |
1072 | /* Parse any control messages the user may have included. */ |
1073 | ret = rds_cmsg_send(rs, rm, msg, &allocated_mr); | |
1074 | if (ret) | |
1075 | goto out; | |
1076 | ||
2c3a5f9a | 1077 | if (rm->rdma.op_active && !conn->c_trans->xmit_rdma) { |
cb0a6056 | 1078 | printk_ratelimited(KERN_NOTICE "rdma_op %p conn xmit_rdma %p\n", |
f8b3aaf2 | 1079 | &rm->rdma, conn->c_trans->xmit_rdma); |
15133f6e AG |
1080 | ret = -EOPNOTSUPP; |
1081 | goto out; | |
1082 | } | |
1083 | ||
1084 | if (rm->atomic.op_active && !conn->c_trans->xmit_atomic) { | |
cb0a6056 | 1085 | printk_ratelimited(KERN_NOTICE "atomic_op %p conn xmit_atomic %p\n", |
15133f6e | 1086 | &rm->atomic, conn->c_trans->xmit_atomic); |
5c115590 AG |
1087 | ret = -EOPNOTSUPP; |
1088 | goto out; | |
1089 | } | |
1090 | ||
f3c6808d | 1091 | rds_conn_connect_if_down(conn); |
5c115590 AG |
1092 | |
1093 | ret = rds_cong_wait(conn->c_fcong, dport, nonblock, rs); | |
b98ba52f AG |
1094 | if (ret) { |
1095 | rs->rs_seen_congestion = 1; | |
5c115590 | 1096 | goto out; |
b98ba52f | 1097 | } |
5c115590 AG |
1098 | |
1099 | while (!rds_send_queue_rm(rs, conn, rm, rs->rs_bound_port, | |
1100 | dport, &queued)) { | |
1101 | rds_stats_inc(s_send_queue_full); | |
06e8941e | 1102 | |
5c115590 AG |
1103 | if (nonblock) { |
1104 | ret = -EAGAIN; | |
1105 | goto out; | |
1106 | } | |
1107 | ||
aa395145 | 1108 | timeo = wait_event_interruptible_timeout(*sk_sleep(sk), |
5c115590 AG |
1109 | rds_send_queue_rm(rs, conn, rm, |
1110 | rs->rs_bound_port, | |
1111 | dport, | |
1112 | &queued), | |
1113 | timeo); | |
1114 | rdsdebug("sendmsg woke queued %d timeo %ld\n", queued, timeo); | |
1115 | if (timeo > 0 || timeo == MAX_SCHEDULE_TIMEOUT) | |
1116 | continue; | |
1117 | ||
1118 | ret = timeo; | |
1119 | if (ret == 0) | |
1120 | ret = -ETIMEDOUT; | |
1121 | goto out; | |
1122 | } | |
1123 | ||
1124 | /* | |
1125 | * By now we've committed to the send. We reuse rds_send_worker() | |
1126 | * to retry sends in the rds thread if the transport asks us to. | |
1127 | */ | |
1128 | rds_stats_inc(s_send_queued); | |
1129 | ||
db6526dc SS |
1130 | ret = rds_send_xmit(conn); |
1131 | if (ret == -ENOMEM || ret == -EAGAIN) | |
1132 | queue_delayed_work(rds_wq, &conn->c_send_w, 1); | |
5c115590 AG |
1133 | |
1134 | rds_message_put(rm); | |
1135 | return payload_len; | |
1136 | ||
1137 | out: | |
1138 | /* If the user included a RDMA_MAP cmsg, we allocated a MR on the fly. | |
1139 | * If the sendmsg goes through, we keep the MR. If it fails with EAGAIN | |
1140 | * or in any other way, we need to destroy the MR again */ | |
1141 | if (allocated_mr) | |
1142 | rds_rdma_unuse(rs, rds_rdma_cookie_key(rm->m_rdma_cookie), 1); | |
1143 | ||
1144 | if (rm) | |
1145 | rds_message_put(rm); | |
1146 | return ret; | |
1147 | } | |
1148 | ||
1149 | /* | |
1150 | * Reply to a ping packet. | |
1151 | */ | |
1152 | int | |
1153 | rds_send_pong(struct rds_connection *conn, __be16 dport) | |
1154 | { | |
1155 | struct rds_message *rm; | |
1156 | unsigned long flags; | |
1157 | int ret = 0; | |
1158 | ||
1159 | rm = rds_message_alloc(0, GFP_ATOMIC); | |
8690bfa1 | 1160 | if (!rm) { |
5c115590 AG |
1161 | ret = -ENOMEM; |
1162 | goto out; | |
1163 | } | |
1164 | ||
1165 | rm->m_daddr = conn->c_faddr; | |
acfcd4d4 | 1166 | rm->data.op_active = 1; |
5c115590 | 1167 | |
f3c6808d | 1168 | rds_conn_connect_if_down(conn); |
5c115590 AG |
1169 | |
1170 | ret = rds_cong_wait(conn->c_fcong, dport, 1, NULL); | |
1171 | if (ret) | |
1172 | goto out; | |
1173 | ||
1174 | spin_lock_irqsave(&conn->c_lock, flags); | |
1175 | list_add_tail(&rm->m_conn_item, &conn->c_send_queue); | |
1176 | set_bit(RDS_MSG_ON_CONN, &rm->m_flags); | |
1177 | rds_message_addref(rm); | |
1178 | rm->m_inc.i_conn = conn; | |
1179 | ||
1180 | rds_message_populate_header(&rm->m_inc.i_hdr, 0, dport, | |
1181 | conn->c_next_tx_seq); | |
1182 | conn->c_next_tx_seq++; | |
1183 | spin_unlock_irqrestore(&conn->c_lock, flags); | |
1184 | ||
1185 | rds_stats_inc(s_send_queued); | |
1186 | rds_stats_inc(s_send_pong); | |
1187 | ||
7b4b0009 | 1188 | /* schedule the send work on rds_wq */ |
1189 | queue_delayed_work(rds_wq, &conn->c_send_w, 1); | |
acfcd4d4 | 1190 | |
5c115590 AG |
1191 | rds_message_put(rm); |
1192 | return 0; | |
1193 | ||
1194 | out: | |
1195 | if (rm) | |
1196 | rds_message_put(rm); | |
1197 | return ret; | |
1198 | } |