]> git.proxmox.com Git - mirror_ubuntu-artful-kernel.git/blob - drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c
projects / mirror_ubuntu-artful-kernel.git / blob
? search:
summary | shortlog | log | commit | commitdiff | tree
blame | history | raw | HEAD
Merge tag 'powerpc-4.13-8' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
[mirror_ubuntu-artful-kernel.git] / drivers / staging / lustre / lnet / klnds / o2iblnd / o2iblnd_cb.c
1 /*
2 * GPL HEADER START
3 *
4 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
5 *
6 * This program is free software; you can redistribute it and/or modify
7 * it under the terms of the GNU General Public License version 2 only,
8 * as published by the Free Software Foundation.
9 *
10 * This program is distributed in the hope that it will be useful, but
11 * WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
13 * General Public License version 2 for more details (a copy is included
14 * in the LICENSE file that accompanied this code).
15 *
16 * You should have received a copy of the GNU General Public License
17 * version 2 along with this program; If not, see
18 * http://www.gnu.org/licenses/gpl-2.0.html
19 *
20 * GPL HEADER END
21 */
22 /*
23 * Copyright (c) 2007, 2010, Oracle and/or its affiliates. All rights reserved.
24 * Use is subject to license terms.
25 *
26 * Copyright (c) 2012, 2015, Intel Corporation.
27 */
28 /*
29 * This file is part of Lustre, http://www.lustre.org/
30 * Lustre is a trademark of Sun Microsystems, Inc.
31 *
32 * lnet/klnds/o2iblnd/o2iblnd_cb.c
33 *
34 * Author: Eric Barton <eric@bartonsoftware.com>
35 */
36
37 #include "o2iblnd.h"
38
39 #define MAX_CONN_RACES_BEFORE_ABORT 20
40
41 static void kiblnd_peer_alive(struct kib_peer *peer);
42 static void kiblnd_peer_connect_failed(struct kib_peer *peer, int active, int error);
43 static void kiblnd_init_tx_msg(struct lnet_ni *ni, struct kib_tx *tx,
44 int type, int body_nob);
45 static int kiblnd_init_rdma(struct kib_conn *conn, struct kib_tx *tx, int type,
46 int resid, struct kib_rdma_desc *dstrd,
47 __u64 dstcookie);
48 static void kiblnd_queue_tx_locked(struct kib_tx *tx, struct kib_conn *conn);
49 static void kiblnd_queue_tx(struct kib_tx *tx, struct kib_conn *conn);
50 static void kiblnd_unmap_tx(struct lnet_ni *ni, struct kib_tx *tx);
51 static void kiblnd_check_sends_locked(struct kib_conn *conn);
52
53 static void
54 kiblnd_tx_done(struct lnet_ni *ni, struct kib_tx *tx)
55 {
56 struct lnet_msg *lntmsg[2];
57 struct kib_net *net = ni->ni_data;
58 int rc;
59 int i;
60
61 LASSERT(net);
62 LASSERT(!in_interrupt());
63 LASSERT(!tx->tx_queued); /* mustn't be queued for sending */
64 LASSERT(!tx->tx_sending); /* mustn't be awaiting sent callback */
65 LASSERT(!tx->tx_waiting); /* mustn't be awaiting peer response */
66 LASSERT(tx->tx_pool);
67
68 kiblnd_unmap_tx(ni, tx);
69
70 /* tx may have up to 2 lnet msgs to finalise */
71 lntmsg[0] = tx->tx_lntmsg[0]; tx->tx_lntmsg[0] = NULL;
72 lntmsg[1] = tx->tx_lntmsg[1]; tx->tx_lntmsg[1] = NULL;
73 rc = tx->tx_status;
74
75 if (tx->tx_conn) {
76 LASSERT(ni == tx->tx_conn->ibc_peer->ibp_ni);
77
78 kiblnd_conn_decref(tx->tx_conn);
79 tx->tx_conn = NULL;
80 }
81
82 tx->tx_nwrq = 0;
83 tx->tx_status = 0;
84
85 kiblnd_pool_free_node(&tx->tx_pool->tpo_pool, &tx->tx_list);
86
87 /* delay finalize until my descs have been freed */
88 for (i = 0; i < 2; i++) {
89 if (!lntmsg[i])
90 continue;
91
92 lnet_finalize(ni, lntmsg[i], rc);
93 }
94 }
95
96 void
97 kiblnd_txlist_done(struct lnet_ni *ni, struct list_head *txlist, int status)
98 {
99 struct kib_tx *tx;
100
101 while (!list_empty(txlist)) {
102 tx = list_entry(txlist->next, struct kib_tx, tx_list);
103
104 list_del(&tx->tx_list);
105 /* complete now */
106 tx->tx_waiting = 0;
107 tx->tx_status = status;
108 kiblnd_tx_done(ni, tx);
109 }
110 }
111
112 static struct kib_tx *
113 kiblnd_get_idle_tx(struct lnet_ni *ni, lnet_nid_t target)
114 {
115 struct kib_net *net = (struct kib_net *)ni->ni_data;
116 struct list_head *node;
117 struct kib_tx *tx;
118 struct kib_tx_poolset *tps;
119
120 tps = net->ibn_tx_ps[lnet_cpt_of_nid(target)];
121 node = kiblnd_pool_alloc_node(&tps->tps_poolset);
122 if (!node)
123 return NULL;
124 tx = list_entry(node, struct kib_tx, tx_list);
125
126 LASSERT(!tx->tx_nwrq);
127 LASSERT(!tx->tx_queued);
128 LASSERT(!tx->tx_sending);
129 LASSERT(!tx->tx_waiting);
130 LASSERT(!tx->tx_status);
131 LASSERT(!tx->tx_conn);
132 LASSERT(!tx->tx_lntmsg[0]);
133 LASSERT(!tx->tx_lntmsg[1]);
134 LASSERT(!tx->tx_nfrags);
135
136 return tx;
137 }
138
139 static void
140 kiblnd_drop_rx(struct kib_rx *rx)
141 {
142 struct kib_conn *conn = rx->rx_conn;
143 struct kib_sched_info *sched = conn->ibc_sched;
144 unsigned long flags;
145
146 spin_lock_irqsave(&sched->ibs_lock, flags);
147 LASSERT(conn->ibc_nrx > 0);
148 conn->ibc_nrx--;
149 spin_unlock_irqrestore(&sched->ibs_lock, flags);
150
151 kiblnd_conn_decref(conn);
152 }
153
154 int
155 kiblnd_post_rx(struct kib_rx *rx, int credit)
156 {
157 struct kib_conn *conn = rx->rx_conn;
158 struct kib_net *net = conn->ibc_peer->ibp_ni->ni_data;
159 struct ib_recv_wr *bad_wrq = NULL;
160 int rc;
161
162 LASSERT(net);
163 LASSERT(!in_interrupt());
164 LASSERT(credit == IBLND_POSTRX_NO_CREDIT ||
165 credit == IBLND_POSTRX_PEER_CREDIT ||
166 credit == IBLND_POSTRX_RSRVD_CREDIT);
167
168 rx->rx_sge.lkey = conn->ibc_hdev->ibh_pd->local_dma_lkey;
169 rx->rx_sge.addr = rx->rx_msgaddr;
170 rx->rx_sge.length = IBLND_MSG_SIZE;
171
172 rx->rx_wrq.next = NULL;
173 rx->rx_wrq.sg_list = &rx->rx_sge;
174 rx->rx_wrq.num_sge = 1;
175 rx->rx_wrq.wr_id = kiblnd_ptr2wreqid(rx, IBLND_WID_RX);
176
177 LASSERT(conn->ibc_state >= IBLND_CONN_INIT);
178 LASSERT(rx->rx_nob >= 0); /* not posted */
179
180 if (conn->ibc_state > IBLND_CONN_ESTABLISHED) {
181 kiblnd_drop_rx(rx); /* No more posts for this rx */
182 return 0;
183 }
184
185 rx->rx_nob = -1; /* flag posted */
186
187 /* NB: need an extra reference after ib_post_recv because we don't
188 * own this rx (and rx::rx_conn) anymore, LU-5678.
189 */
190 kiblnd_conn_addref(conn);
191 rc = ib_post_recv(conn->ibc_cmid->qp, &rx->rx_wrq, &bad_wrq);
192 if (unlikely(rc)) {
193 CERROR("Can't post rx for %s: %d, bad_wrq: %p\n",
194 libcfs_nid2str(conn->ibc_peer->ibp_nid), rc, bad_wrq);
195 rx->rx_nob = 0;
196 }
197
198 if (conn->ibc_state < IBLND_CONN_ESTABLISHED) /* Initial post */
199 goto out;
200
201 if (unlikely(rc)) {
202 kiblnd_close_conn(conn, rc);
203 kiblnd_drop_rx(rx); /* No more posts for this rx */
204 goto out;
205 }
206
207 if (credit == IBLND_POSTRX_NO_CREDIT)
208 goto out;
209
210 spin_lock(&conn->ibc_lock);
211 if (credit == IBLND_POSTRX_PEER_CREDIT)
212 conn->ibc_outstanding_credits++;
213 else
214 conn->ibc_reserved_credits++;
215 kiblnd_check_sends_locked(conn);
216 spin_unlock(&conn->ibc_lock);
217
218 out:
219 kiblnd_conn_decref(conn);
220 return rc;
221 }
222
223 static struct kib_tx *
224 kiblnd_find_waiting_tx_locked(struct kib_conn *conn, int txtype, __u64 cookie)
225 {
226 struct list_head *tmp;
227
228 list_for_each(tmp, &conn->ibc_active_txs) {
229 struct kib_tx *tx = list_entry(tmp, struct kib_tx, tx_list);
230
231 LASSERT(!tx->tx_queued);
232 LASSERT(tx->tx_sending || tx->tx_waiting);
233
234 if (tx->tx_cookie != cookie)
235 continue;
236
237 if (tx->tx_waiting &&
238 tx->tx_msg->ibm_type == txtype)
239 return tx;
240
241 CWARN("Bad completion: %swaiting, type %x (wanted %x)\n",
242 tx->tx_waiting ? "" : "NOT ",
243 tx->tx_msg->ibm_type, txtype);
244 }
245 return NULL;
246 }
247
248 static void
249 kiblnd_handle_completion(struct kib_conn *conn, int txtype, int status, __u64 cookie)
250 {
251 struct kib_tx *tx;
252 struct lnet_ni *ni = conn->ibc_peer->ibp_ni;
253 int idle;
254
255 spin_lock(&conn->ibc_lock);
256
257 tx = kiblnd_find_waiting_tx_locked(conn, txtype, cookie);
258 if (!tx) {
259 spin_unlock(&conn->ibc_lock);
260
261 CWARN("Unmatched completion type %x cookie %#llx from %s\n",
262 txtype, cookie, libcfs_nid2str(conn->ibc_peer->ibp_nid));
263 kiblnd_close_conn(conn, -EPROTO);
264 return;
265 }
266
267 if (!tx->tx_status) { /* success so far */
268 if (status < 0) /* failed? */
269 tx->tx_status = status;
270 else if (txtype == IBLND_MSG_GET_REQ)
271 lnet_set_reply_msg_len(ni, tx->tx_lntmsg[1], status);
272 }
273
274 tx->tx_waiting = 0;
275
276 idle = !tx->tx_queued && !tx->tx_sending;
277 if (idle)
278 list_del(&tx->tx_list);
279
280 spin_unlock(&conn->ibc_lock);
281
282 if (idle)
283 kiblnd_tx_done(ni, tx);
284 }
285
286 static void
287 kiblnd_send_completion(struct kib_conn *conn, int type, int status, __u64 cookie)
288 {
289 struct lnet_ni *ni = conn->ibc_peer->ibp_ni;
290 struct kib_tx *tx = kiblnd_get_idle_tx(ni, conn->ibc_peer->ibp_nid);
291
292 if (!tx) {
293 CERROR("Can't get tx for completion %x for %s\n",
294 type, libcfs_nid2str(conn->ibc_peer->ibp_nid));
295 return;
296 }
297
298 tx->tx_msg->ibm_u.completion.ibcm_status = status;
299 tx->tx_msg->ibm_u.completion.ibcm_cookie = cookie;
300 kiblnd_init_tx_msg(ni, tx, type, sizeof(struct kib_completion_msg));
301
302 kiblnd_queue_tx(tx, conn);
303 }
304
305 static void
306 kiblnd_handle_rx(struct kib_rx *rx)
307 {
308 struct kib_msg *msg = rx->rx_msg;
309 struct kib_conn *conn = rx->rx_conn;
310 struct lnet_ni *ni = conn->ibc_peer->ibp_ni;
311 int credits = msg->ibm_credits;
312 struct kib_tx *tx;
313 int rc = 0;
314 int rc2;
315 int post_credit;
316
317 LASSERT(conn->ibc_state >= IBLND_CONN_ESTABLISHED);
318
319 CDEBUG(D_NET, "Received %x[%d] from %s\n",
320 msg->ibm_type, credits,
321 libcfs_nid2str(conn->ibc_peer->ibp_nid));
322
323 if (credits) {
324 /* Have I received credits that will let me send? */
325 spin_lock(&conn->ibc_lock);
326
327 if (conn->ibc_credits + credits >
328 conn->ibc_queue_depth) {
329 rc2 = conn->ibc_credits;
330 spin_unlock(&conn->ibc_lock);
331
332 CERROR("Bad credits from %s: %d + %d > %d\n",
333 libcfs_nid2str(conn->ibc_peer->ibp_nid),
334 rc2, credits, conn->ibc_queue_depth);
335
336 kiblnd_close_conn(conn, -EPROTO);
337 kiblnd_post_rx(rx, IBLND_POSTRX_NO_CREDIT);
338 return;
339 }
340
341 conn->ibc_credits += credits;
342
343 /* This ensures the credit taken by NOOP can be returned */
344 if (msg->ibm_type == IBLND_MSG_NOOP &&
345 !IBLND_OOB_CAPABLE(conn->ibc_version)) /* v1 only */
346 conn->ibc_outstanding_credits++;
347
348 kiblnd_check_sends_locked(conn);
349 spin_unlock(&conn->ibc_lock);
350 }
351
352 switch (msg->ibm_type) {
353 default:
354 CERROR("Bad IBLND message type %x from %s\n",
355 msg->ibm_type, libcfs_nid2str(conn->ibc_peer->ibp_nid));
356 post_credit = IBLND_POSTRX_NO_CREDIT;
357 rc = -EPROTO;
358 break;
359
360 case IBLND_MSG_NOOP:
361 if (IBLND_OOB_CAPABLE(conn->ibc_version)) {
362 post_credit = IBLND_POSTRX_NO_CREDIT;
363 break;
364 }
365
366 if (credits) /* credit already posted */
367 post_credit = IBLND_POSTRX_NO_CREDIT;
368 else /* a keepalive NOOP */
369 post_credit = IBLND_POSTRX_PEER_CREDIT;
370 break;
371
372 case IBLND_MSG_IMMEDIATE:
373 post_credit = IBLND_POSTRX_DONT_POST;
374 rc = lnet_parse(ni, &msg->ibm_u.immediate.ibim_hdr,
375 msg->ibm_srcnid, rx, 0);
376 if (rc < 0) /* repost on error */
377 post_credit = IBLND_POSTRX_PEER_CREDIT;
378 break;
379
380 case IBLND_MSG_PUT_REQ:
381 post_credit = IBLND_POSTRX_DONT_POST;
382 rc = lnet_parse(ni, &msg->ibm_u.putreq.ibprm_hdr,
383 msg->ibm_srcnid, rx, 1);
384 if (rc < 0) /* repost on error */
385 post_credit = IBLND_POSTRX_PEER_CREDIT;
386 break;
387
388 case IBLND_MSG_PUT_NAK:
389 CWARN("PUT_NACK from %s\n",
390 libcfs_nid2str(conn->ibc_peer->ibp_nid));
391 post_credit = IBLND_POSTRX_RSRVD_CREDIT;
392 kiblnd_handle_completion(conn, IBLND_MSG_PUT_REQ,
393 msg->ibm_u.completion.ibcm_status,
394 msg->ibm_u.completion.ibcm_cookie);
395 break;
396
397 case IBLND_MSG_PUT_ACK:
398 post_credit = IBLND_POSTRX_RSRVD_CREDIT;
399
400 spin_lock(&conn->ibc_lock);
401 tx = kiblnd_find_waiting_tx_locked(conn, IBLND_MSG_PUT_REQ,
402 msg->ibm_u.putack.ibpam_src_cookie);
403 if (tx)
404 list_del(&tx->tx_list);
405 spin_unlock(&conn->ibc_lock);
406
407 if (!tx) {
408 CERROR("Unmatched PUT_ACK from %s\n",
409 libcfs_nid2str(conn->ibc_peer->ibp_nid));
410 rc = -EPROTO;
411 break;
412 }
413
414 LASSERT(tx->tx_waiting);
415 /*
416 * CAVEAT EMPTOR: I could be racing with tx_complete, but...
417 * (a) I can overwrite tx_msg since my peer has received it!
418 * (b) tx_waiting set tells tx_complete() it's not done.
419 */
420 tx->tx_nwrq = 0; /* overwrite PUT_REQ */
421
422 rc2 = kiblnd_init_rdma(conn, tx, IBLND_MSG_PUT_DONE,
423 kiblnd_rd_size(&msg->ibm_u.putack.ibpam_rd),
424 &msg->ibm_u.putack.ibpam_rd,
425 msg->ibm_u.putack.ibpam_dst_cookie);
426 if (rc2 < 0)
427 CERROR("Can't setup rdma for PUT to %s: %d\n",
428 libcfs_nid2str(conn->ibc_peer->ibp_nid), rc2);
429
430 spin_lock(&conn->ibc_lock);
431 tx->tx_waiting = 0; /* clear waiting and queue atomically */
432 kiblnd_queue_tx_locked(tx, conn);
433 spin_unlock(&conn->ibc_lock);
434 break;
435
436 case IBLND_MSG_PUT_DONE:
437 post_credit = IBLND_POSTRX_PEER_CREDIT;
438 kiblnd_handle_completion(conn, IBLND_MSG_PUT_ACK,
439 msg->ibm_u.completion.ibcm_status,
440 msg->ibm_u.completion.ibcm_cookie);
441 break;
442
443 case IBLND_MSG_GET_REQ:
444 post_credit = IBLND_POSTRX_DONT_POST;
445 rc = lnet_parse(ni, &msg->ibm_u.get.ibgm_hdr,
446 msg->ibm_srcnid, rx, 1);
447 if (rc < 0) /* repost on error */
448 post_credit = IBLND_POSTRX_PEER_CREDIT;
449 break;
450
451 case IBLND_MSG_GET_DONE:
452 post_credit = IBLND_POSTRX_RSRVD_CREDIT;
453 kiblnd_handle_completion(conn, IBLND_MSG_GET_REQ,
454 msg->ibm_u.completion.ibcm_status,
455 msg->ibm_u.completion.ibcm_cookie);
456 break;
457 }
458
459 if (rc < 0) /* protocol error */
460 kiblnd_close_conn(conn, rc);
461
462 if (post_credit != IBLND_POSTRX_DONT_POST)
463 kiblnd_post_rx(rx, post_credit);
464 }
465
466 static void
467 kiblnd_rx_complete(struct kib_rx *rx, int status, int nob)
468 {
469 struct kib_msg *msg = rx->rx_msg;
470 struct kib_conn *conn = rx->rx_conn;
471 struct lnet_ni *ni = conn->ibc_peer->ibp_ni;
472 struct kib_net *net = ni->ni_data;
473 int rc;
474 int err = -EIO;
475
476 LASSERT(net);
477 LASSERT(rx->rx_nob < 0); /* was posted */
478 rx->rx_nob = 0; /* isn't now */
479
480 if (conn->ibc_state > IBLND_CONN_ESTABLISHED)
481 goto ignore;
482
483 if (status != IB_WC_SUCCESS) {
484 CNETERR("Rx from %s failed: %d\n",
485 libcfs_nid2str(conn->ibc_peer->ibp_nid), status);
486 goto failed;
487 }
488
489 LASSERT(nob >= 0);
490 rx->rx_nob = nob;
491
492 rc = kiblnd_unpack_msg(msg, rx->rx_nob);
493 if (rc) {
494 CERROR("Error %d unpacking rx from %s\n",
495 rc, libcfs_nid2str(conn->ibc_peer->ibp_nid));
496 goto failed;
497 }
498
499 if (msg->ibm_srcnid != conn->ibc_peer->ibp_nid ||
500 msg->ibm_dstnid != ni->ni_nid ||
501 msg->ibm_srcstamp != conn->ibc_incarnation ||
502 msg->ibm_dststamp != net->ibn_incarnation) {
503 CERROR("Stale rx from %s\n",
504 libcfs_nid2str(conn->ibc_peer->ibp_nid));
505 err = -ESTALE;
506 goto failed;
507 }
508
509 /* set time last known alive */
510 kiblnd_peer_alive(conn->ibc_peer);
511
512 /* racing with connection establishment/teardown! */
513
514 if (conn->ibc_state < IBLND_CONN_ESTABLISHED) {
515 rwlock_t *g_lock = &kiblnd_data.kib_global_lock;
516 unsigned long flags;
517
518 write_lock_irqsave(g_lock, flags);
519 /* must check holding global lock to eliminate race */
520 if (conn->ibc_state < IBLND_CONN_ESTABLISHED) {
521 list_add_tail(&rx->rx_list, &conn->ibc_early_rxs);
522 write_unlock_irqrestore(g_lock, flags);
523 return;
524 }
525 write_unlock_irqrestore(g_lock, flags);
526 }
527 kiblnd_handle_rx(rx);
528 return;
529
530 failed:
531 CDEBUG(D_NET, "rx %p conn %p\n", rx, conn);
532 kiblnd_close_conn(conn, err);
533 ignore:
534 kiblnd_drop_rx(rx); /* Don't re-post rx. */
535 }
536
537 static struct page *
538 kiblnd_kvaddr_to_page(unsigned long vaddr)
539 {
540 struct page *page;
541
542 if (is_vmalloc_addr((void *)vaddr)) {
543 page = vmalloc_to_page((void *)vaddr);
544 LASSERT(page);
545 return page;
546 }
547 #ifdef CONFIG_HIGHMEM
548 if (vaddr >= PKMAP_BASE &&
549 vaddr < (PKMAP_BASE + LAST_PKMAP * PAGE_SIZE)) {
550 /* No highmem pages only used for bulk (kiov) I/O */
551 CERROR("find page for address in highmem\n");
552 LBUG();
553 }
554 #endif
555 page = virt_to_page(vaddr);
556 LASSERT(page);
557 return page;
558 }
559
560 static int
561 kiblnd_fmr_map_tx(struct kib_net *net, struct kib_tx *tx, struct kib_rdma_desc *rd, __u32 nob)
562 {
563 struct kib_hca_dev *hdev;
564 struct kib_fmr_poolset *fps;
565 int cpt;
566 int rc;
567
568 LASSERT(tx->tx_pool);
569 LASSERT(tx->tx_pool->tpo_pool.po_owner);
570
571 hdev = tx->tx_pool->tpo_hdev;
572 cpt = tx->tx_pool->tpo_pool.po_owner->ps_cpt;
573
574 fps = net->ibn_fmr_ps[cpt];
575 rc = kiblnd_fmr_pool_map(fps, tx, rd, nob, 0, &tx->fmr);
576 if (rc) {
577 CERROR("Can't map %u bytes: %d\n", nob, rc);
578 return rc;
579 }
580
581 /*
582 * If rd is not tx_rd, it's going to get sent to a peer, who will need
583 * the rkey
584 */
585 rd->rd_key = tx->fmr.fmr_key;
586 rd->rd_frags[0].rf_addr &= ~hdev->ibh_page_mask;
587 rd->rd_frags[0].rf_nob = nob;
588 rd->rd_nfrags = 1;
589
590 return 0;
591 }
592
593 static void kiblnd_unmap_tx(struct lnet_ni *ni, struct kib_tx *tx)
594 {
595 struct kib_net *net = ni->ni_data;
596
597 LASSERT(net);
598
599 if (net->ibn_fmr_ps)
600 kiblnd_fmr_pool_unmap(&tx->fmr, tx->tx_status);
601
602 if (tx->tx_nfrags) {
603 kiblnd_dma_unmap_sg(tx->tx_pool->tpo_hdev->ibh_ibdev,
604 tx->tx_frags, tx->tx_nfrags, tx->tx_dmadir);
605 tx->tx_nfrags = 0;
606 }
607 }
608
609 static int kiblnd_map_tx(struct lnet_ni *ni, struct kib_tx *tx,
610 struct kib_rdma_desc *rd, int nfrags)
611 {
612 struct kib_net *net = ni->ni_data;
613 struct kib_hca_dev *hdev = net->ibn_dev->ibd_hdev;
614 __u32 nob;
615 int i;
616
617 /*
618 * If rd is not tx_rd, it's going to get sent to a peer and I'm the
619 * RDMA sink
620 */
621 tx->tx_dmadir = (rd != tx->tx_rd) ? DMA_FROM_DEVICE : DMA_TO_DEVICE;
622 tx->tx_nfrags = nfrags;
623
624 rd->rd_nfrags = kiblnd_dma_map_sg(hdev->ibh_ibdev, tx->tx_frags,
625 tx->tx_nfrags, tx->tx_dmadir);
626
627 for (i = 0, nob = 0; i < rd->rd_nfrags; i++) {
628 rd->rd_frags[i].rf_nob = kiblnd_sg_dma_len(
629 hdev->ibh_ibdev, &tx->tx_frags[i]);
630 rd->rd_frags[i].rf_addr = kiblnd_sg_dma_address(
631 hdev->ibh_ibdev, &tx->tx_frags[i]);
632 nob += rd->rd_frags[i].rf_nob;
633 }
634
635 if (net->ibn_fmr_ps)
636 return kiblnd_fmr_map_tx(net, tx, rd, nob);
637
638 return -EINVAL;
639 }
640
641 static int
642 kiblnd_setup_rd_iov(struct lnet_ni *ni, struct kib_tx *tx,
643 struct kib_rdma_desc *rd, unsigned int niov,
644 const struct kvec *iov, int offset, int nob)
645 {
646 struct kib_net *net = ni->ni_data;
647 struct page *page;
648 struct scatterlist *sg;
649 unsigned long vaddr;
650 int fragnob;
651 int page_offset;
652
653 LASSERT(nob > 0);
654 LASSERT(niov > 0);
655 LASSERT(net);
656
657 while (offset >= iov->iov_len) {
658 offset -= iov->iov_len;
659 niov--;
660 iov++;
661 LASSERT(niov > 0);
662 }
663
664 sg = tx->tx_frags;
665 do {
666 LASSERT(niov > 0);
667
668 vaddr = ((unsigned long)iov->iov_base) + offset;
669 page_offset = vaddr & (PAGE_SIZE - 1);
670 page = kiblnd_kvaddr_to_page(vaddr);
671 if (!page) {
672 CERROR("Can't find page\n");
673 return -EFAULT;
674 }
675
676 fragnob = min((int)(iov->iov_len - offset), nob);
677 fragnob = min(fragnob, (int)PAGE_SIZE - page_offset);
678
679 sg_set_page(sg, page, fragnob, page_offset);
680 sg = sg_next(sg);
681 if (!sg) {
682 CERROR("lacking enough sg entries to map tx\n");
683 return -EFAULT;
684 }
685
686 if (offset + fragnob < iov->iov_len) {
687 offset += fragnob;
688 } else {
689 offset = 0;
690 iov++;
691 niov--;
692 }
693 nob -= fragnob;
694 } while (nob > 0);
695
696 return kiblnd_map_tx(ni, tx, rd, sg - tx->tx_frags);
697 }
698
699 static int
700 kiblnd_setup_rd_kiov(struct lnet_ni *ni, struct kib_tx *tx,
701 struct kib_rdma_desc *rd, int nkiov,
702 const struct bio_vec *kiov, int offset, int nob)
703 {
704 struct kib_net *net = ni->ni_data;
705 struct scatterlist *sg;
706 int fragnob;
707
708 CDEBUG(D_NET, "niov %d offset %d nob %d\n", nkiov, offset, nob);
709
710 LASSERT(nob > 0);
711 LASSERT(nkiov > 0);
712 LASSERT(net);
713
714 while (offset >= kiov->bv_len) {
715 offset -= kiov->bv_len;
716 nkiov--;
717 kiov++;
718 LASSERT(nkiov > 0);
719 }
720
721 sg = tx->tx_frags;
722 do {
723 LASSERT(nkiov > 0);
724
725 fragnob = min((int)(kiov->bv_len - offset), nob);
726
727 sg_set_page(sg, kiov->bv_page, fragnob,
728 kiov->bv_offset + offset);
729 sg = sg_next(sg);
730 if (!sg) {
731 CERROR("lacking enough sg entries to map tx\n");
732 return -EFAULT;
733 }
734
735 offset = 0;
736 kiov++;
737 nkiov--;
738 nob -= fragnob;
739 } while (nob > 0);
740
741 return kiblnd_map_tx(ni, tx, rd, sg - tx->tx_frags);
742 }
743
744 static int
745 kiblnd_post_tx_locked(struct kib_conn *conn, struct kib_tx *tx, int credit)
746 __must_hold(&conn->ibc_lock)
747 {
748 struct kib_msg *msg = tx->tx_msg;
749 struct kib_peer *peer = conn->ibc_peer;
750 struct lnet_ni *ni = peer->ibp_ni;
751 int ver = conn->ibc_version;
752 int rc;
753 int done;
754
755 LASSERT(tx->tx_queued);
756 /* We rely on this for QP sizing */
757 LASSERT(tx->tx_nwrq > 0);
758
759 LASSERT(!credit || credit == 1);
760 LASSERT(conn->ibc_outstanding_credits >= 0);
761 LASSERT(conn->ibc_outstanding_credits <= conn->ibc_queue_depth);
762 LASSERT(conn->ibc_credits >= 0);
763 LASSERT(conn->ibc_credits <= conn->ibc_queue_depth);
764
765 if (conn->ibc_nsends_posted == kiblnd_concurrent_sends(ver, ni)) {
766 /* tx completions outstanding... */
767 CDEBUG(D_NET, "%s: posted enough\n",
768 libcfs_nid2str(peer->ibp_nid));
769 return -EAGAIN;
770 }
771
772 if (credit && !conn->ibc_credits) { /* no credits */
773 CDEBUG(D_NET, "%s: no credits\n",
774 libcfs_nid2str(peer->ibp_nid));
775 return -EAGAIN;
776 }
777
778 if (credit && !IBLND_OOB_CAPABLE(ver) &&
779 conn->ibc_credits == 1 && /* last credit reserved */
780 msg->ibm_type != IBLND_MSG_NOOP) { /* for NOOP */
781 CDEBUG(D_NET, "%s: not using last credit\n",
782 libcfs_nid2str(peer->ibp_nid));
783 return -EAGAIN;
784 }
785
786 /* NB don't drop ibc_lock before bumping tx_sending */
787 list_del(&tx->tx_list);
788 tx->tx_queued = 0;
789
790 if (msg->ibm_type == IBLND_MSG_NOOP &&
791 (!kiblnd_need_noop(conn) || /* redundant NOOP */
792 (IBLND_OOB_CAPABLE(ver) && /* posted enough NOOP */
793 conn->ibc_noops_posted == IBLND_OOB_MSGS(ver)))) {
794 /*
795 * OK to drop when posted enough NOOPs, since
796 * kiblnd_check_sends_locked will queue NOOP again when
797 * posted NOOPs complete
798 */
799 spin_unlock(&conn->ibc_lock);
800 kiblnd_tx_done(peer->ibp_ni, tx);
801 spin_lock(&conn->ibc_lock);
802 CDEBUG(D_NET, "%s(%d): redundant or enough NOOP\n",
803 libcfs_nid2str(peer->ibp_nid),
804 conn->ibc_noops_posted);
805 return 0;
806 }
807
808 kiblnd_pack_msg(peer->ibp_ni, msg, ver, conn->ibc_outstanding_credits,
809 peer->ibp_nid, conn->ibc_incarnation);
810
811 conn->ibc_credits -= credit;
812 conn->ibc_outstanding_credits = 0;
813 conn->ibc_nsends_posted++;
814 if (msg->ibm_type == IBLND_MSG_NOOP)
815 conn->ibc_noops_posted++;
816
817 /*
818 * CAVEAT EMPTOR! This tx could be the PUT_DONE of an RDMA
819 * PUT. If so, it was first queued here as a PUT_REQ, sent and
820 * stashed on ibc_active_txs, matched by an incoming PUT_ACK,
821 * and then re-queued here. It's (just) possible that
822 * tx_sending is non-zero if we've not done the tx_complete()
823 * from the first send; hence the ++ rather than = below.
824 */
825 tx->tx_sending++;
826 list_add(&tx->tx_list, &conn->ibc_active_txs);
827
828 /* I'm still holding ibc_lock! */
829 if (conn->ibc_state != IBLND_CONN_ESTABLISHED) {
830 rc = -ECONNABORTED;
831 } else if (tx->tx_pool->tpo_pool.po_failed ||
832 conn->ibc_hdev != tx->tx_pool->tpo_hdev) {
833 /* close_conn will launch failover */
834 rc = -ENETDOWN;
835 } else {
836 struct kib_fast_reg_descriptor *frd = tx->fmr.fmr_frd;
837 struct ib_send_wr *bad = &tx->tx_wrq[tx->tx_nwrq - 1].wr;
838 struct ib_send_wr *wrq = &tx->tx_wrq[0].wr;
839
840 if (frd) {
841 if (!frd->frd_valid) {
842 wrq = &frd->frd_inv_wr;
843 wrq->next = &frd->frd_fastreg_wr.wr;
844 } else {
845 wrq = &frd->frd_fastreg_wr.wr;
846 }
847 frd->frd_fastreg_wr.wr.next = &tx->tx_wrq[0].wr;
848 }
849
850 LASSERTF(bad->wr_id == kiblnd_ptr2wreqid(tx, IBLND_WID_TX),
851 "bad wr_id %llx, opc %d, flags %d, peer: %s\n",
852 bad->wr_id, bad->opcode, bad->send_flags,
853 libcfs_nid2str(conn->ibc_peer->ibp_nid));
854 bad = NULL;
855 rc = ib_post_send(conn->ibc_cmid->qp, wrq, &bad);
856 }
857
858 conn->ibc_last_send = jiffies;
859
860 if (!rc)
861 return 0;
862
863 /*
864 * NB credits are transferred in the actual
865 * message, which can only be the last work item
866 */
867 conn->ibc_credits += credit;
868 conn->ibc_outstanding_credits += msg->ibm_credits;
869 conn->ibc_nsends_posted--;
870 if (msg->ibm_type == IBLND_MSG_NOOP)
871 conn->ibc_noops_posted--;
872
873 tx->tx_status = rc;
874 tx->tx_waiting = 0;
875 tx->tx_sending--;
876
877 done = !tx->tx_sending;
878 if (done)
879 list_del(&tx->tx_list);
880
881 spin_unlock(&conn->ibc_lock);
882
883 if (conn->ibc_state == IBLND_CONN_ESTABLISHED)
884 CERROR("Error %d posting transmit to %s\n",
885 rc, libcfs_nid2str(peer->ibp_nid));
886 else
887 CDEBUG(D_NET, "Error %d posting transmit to %s\n",
888 rc, libcfs_nid2str(peer->ibp_nid));
889
890 kiblnd_close_conn(conn, rc);
891
892 if (done)
893 kiblnd_tx_done(peer->ibp_ni, tx);
894
895 spin_lock(&conn->ibc_lock);
896
897 return -EIO;
898 }
899
900 static void
901 kiblnd_check_sends_locked(struct kib_conn *conn)
902 {
903 int ver = conn->ibc_version;
904 struct lnet_ni *ni = conn->ibc_peer->ibp_ni;
905 struct kib_tx *tx;
906
907 /* Don't send anything until after the connection is established */
908 if (conn->ibc_state < IBLND_CONN_ESTABLISHED) {
909 CDEBUG(D_NET, "%s too soon\n",
910 libcfs_nid2str(conn->ibc_peer->ibp_nid));
911 return;
912 }
913
914 LASSERT(conn->ibc_nsends_posted <= kiblnd_concurrent_sends(ver, ni));
915 LASSERT(!IBLND_OOB_CAPABLE(ver) ||
916 conn->ibc_noops_posted <= IBLND_OOB_MSGS(ver));
917 LASSERT(conn->ibc_reserved_credits >= 0);
918
919 while (conn->ibc_reserved_credits > 0 &&
920 !list_empty(&conn->ibc_tx_queue_rsrvd)) {
921 tx = list_entry(conn->ibc_tx_queue_rsrvd.next,
922 struct kib_tx, tx_list);
923 list_del(&tx->tx_list);
924 list_add_tail(&tx->tx_list, &conn->ibc_tx_queue);
925 conn->ibc_reserved_credits--;
926 }
927
928 if (kiblnd_need_noop(conn)) {
929 spin_unlock(&conn->ibc_lock);
930
931 tx = kiblnd_get_idle_tx(ni, conn->ibc_peer->ibp_nid);
932 if (tx)
933 kiblnd_init_tx_msg(ni, tx, IBLND_MSG_NOOP, 0);
934
935 spin_lock(&conn->ibc_lock);
936 if (tx)
937 kiblnd_queue_tx_locked(tx, conn);
938 }
939
940 for (;;) {
941 int credit;
942
943 if (!list_empty(&conn->ibc_tx_queue_nocred)) {
944 credit = 0;
945 tx = list_entry(conn->ibc_tx_queue_nocred.next,
946 struct kib_tx, tx_list);
947 } else if (!list_empty(&conn->ibc_tx_noops)) {
948 LASSERT(!IBLND_OOB_CAPABLE(ver));
949 credit = 1;
950 tx = list_entry(conn->ibc_tx_noops.next,
951 struct kib_tx, tx_list);
952 } else if (!list_empty(&conn->ibc_tx_queue)) {
953 credit = 1;
954 tx = list_entry(conn->ibc_tx_queue.next,
955 struct kib_tx, tx_list);
956 } else {
957 break;
958 }
959
960 if (kiblnd_post_tx_locked(conn, tx, credit))
961 break;
962 }
963 }
964
965 static void
966 kiblnd_tx_complete(struct kib_tx *tx, int status)
967 {
968 int failed = (status != IB_WC_SUCCESS);
969 struct kib_conn *conn = tx->tx_conn;
970 int idle;
971
972 LASSERT(tx->tx_sending > 0);
973
974 if (failed) {
975 if (conn->ibc_state == IBLND_CONN_ESTABLISHED)
976 CNETERR("Tx -> %s cookie %#llx sending %d waiting %d: failed %d\n",
977 libcfs_nid2str(conn->ibc_peer->ibp_nid),
978 tx->tx_cookie, tx->tx_sending, tx->tx_waiting,
979 status);
980
981 kiblnd_close_conn(conn, -EIO);
982 } else {
983 kiblnd_peer_alive(conn->ibc_peer);
984 }
985
986 spin_lock(&conn->ibc_lock);
987
988 /*
989 * I could be racing with rdma completion. Whoever makes 'tx' idle
990 * gets to free it, which also drops its ref on 'conn'.
991 */
992 tx->tx_sending--;
993 conn->ibc_nsends_posted--;
994 if (tx->tx_msg->ibm_type == IBLND_MSG_NOOP)
995 conn->ibc_noops_posted--;
996
997 if (failed) {
998 tx->tx_waiting = 0; /* don't wait for peer */
999 tx->tx_status = -EIO;
1000 }
1001
1002 idle = !tx->tx_sending && /* This is the final callback */
1003 !tx->tx_waiting && /* Not waiting for peer */
1004 !tx->tx_queued; /* Not re-queued (PUT_DONE) */
1005 if (idle)
1006 list_del(&tx->tx_list);
1007
1008 kiblnd_check_sends_locked(conn);
1009 spin_unlock(&conn->ibc_lock);
1010
1011 if (idle)
1012 kiblnd_tx_done(conn->ibc_peer->ibp_ni, tx);
1013 }
1014
1015 static void
1016 kiblnd_init_tx_msg(struct lnet_ni *ni, struct kib_tx *tx, int type,
1017 int body_nob)
1018 {
1019 struct kib_hca_dev *hdev = tx->tx_pool->tpo_hdev;
1020 struct ib_sge *sge = &tx->tx_sge[tx->tx_nwrq];
1021 struct ib_rdma_wr *wrq = &tx->tx_wrq[tx->tx_nwrq];
1022 int nob = offsetof(struct kib_msg, ibm_u) + body_nob;
1023
1024 LASSERT(tx->tx_nwrq >= 0);
1025 LASSERT(tx->tx_nwrq < IBLND_MAX_RDMA_FRAGS + 1);
1026 LASSERT(nob <= IBLND_MSG_SIZE);
1027
1028 kiblnd_init_msg(tx->tx_msg, type, body_nob);
1029
1030 sge->lkey = hdev->ibh_pd->local_dma_lkey;
1031 sge->addr = tx->tx_msgaddr;
1032 sge->length = nob;
1033
1034 memset(wrq, 0, sizeof(*wrq));
1035
1036 wrq->wr.next = NULL;
1037 wrq->wr.wr_id = kiblnd_ptr2wreqid(tx, IBLND_WID_TX);
1038 wrq->wr.sg_list = sge;
1039 wrq->wr.num_sge = 1;
1040 wrq->wr.opcode = IB_WR_SEND;
1041 wrq->wr.send_flags = IB_SEND_SIGNALED;
1042
1043 tx->tx_nwrq++;
1044 }
1045
1046 static int
1047 kiblnd_init_rdma(struct kib_conn *conn, struct kib_tx *tx, int type,
1048 int resid, struct kib_rdma_desc *dstrd, __u64 dstcookie)
1049 {
1050 struct kib_msg *ibmsg = tx->tx_msg;
1051 struct kib_rdma_desc *srcrd = tx->tx_rd;
1052 struct ib_sge *sge = &tx->tx_sge[0];
1053 struct ib_rdma_wr *wrq, *next;
1054 int rc = resid;
1055 int srcidx = 0;
1056 int dstidx = 0;
1057 int wrknob;
1058
1059 LASSERT(!in_interrupt());
1060 LASSERT(!tx->tx_nwrq);
1061 LASSERT(type == IBLND_MSG_GET_DONE ||
1062 type == IBLND_MSG_PUT_DONE);
1063
1064 if (kiblnd_rd_size(srcrd) > conn->ibc_max_frags << PAGE_SHIFT) {
1065 CERROR("RDMA is too large for peer %s (%d), src size: %d dst size: %d\n",
1066 libcfs_nid2str(conn->ibc_peer->ibp_nid),
1067 conn->ibc_max_frags << PAGE_SHIFT,
1068 kiblnd_rd_size(srcrd), kiblnd_rd_size(dstrd));
1069 rc = -EMSGSIZE;
1070 goto too_big;
1071 }
1072
1073 while (resid > 0) {
1074 if (srcidx >= srcrd->rd_nfrags) {
1075 CERROR("Src buffer exhausted: %d frags\n", srcidx);
1076 rc = -EPROTO;
1077 break;
1078 }
1079
1080 if (dstidx == dstrd->rd_nfrags) {
1081 CERROR("Dst buffer exhausted: %d frags\n", dstidx);
1082 rc = -EPROTO;
1083 break;
1084 }
1085
1086 if (tx->tx_nwrq >= IBLND_MAX_RDMA_FRAGS) {
1087 CERROR("RDMA has too many fragments for peer %s (%d), src idx/frags: %d/%d dst idx/frags: %d/%d\n",
1088 libcfs_nid2str(conn->ibc_peer->ibp_nid),
1089 IBLND_MAX_RDMA_FRAGS,
1090 srcidx, srcrd->rd_nfrags,
1091 dstidx, dstrd->rd_nfrags);
1092 rc = -EMSGSIZE;
1093 break;
1094 }
1095
1096 wrknob = min3(kiblnd_rd_frag_size(srcrd, srcidx),
1097 kiblnd_rd_frag_size(dstrd, dstidx),
1098 (__u32)resid);
1099
1100 sge = &tx->tx_sge[tx->tx_nwrq];
1101 sge->addr = kiblnd_rd_frag_addr(srcrd, srcidx);
1102 sge->lkey = kiblnd_rd_frag_key(srcrd, srcidx);
1103 sge->length = wrknob;
1104
1105 wrq = &tx->tx_wrq[tx->tx_nwrq];
1106 next = wrq + 1;
1107
1108 wrq->wr.next = &next->wr;
1109 wrq->wr.wr_id = kiblnd_ptr2wreqid(tx, IBLND_WID_RDMA);
1110 wrq->wr.sg_list = sge;
1111 wrq->wr.num_sge = 1;
1112 wrq->wr.opcode = IB_WR_RDMA_WRITE;
1113 wrq->wr.send_flags = 0;
1114
1115 wrq->remote_addr = kiblnd_rd_frag_addr(dstrd, dstidx);
1116 wrq->rkey = kiblnd_rd_frag_key(dstrd, dstidx);
1117
1118 srcidx = kiblnd_rd_consume_frag(srcrd, srcidx, wrknob);
1119 dstidx = kiblnd_rd_consume_frag(dstrd, dstidx, wrknob);
1120
1121 resid -= wrknob;
1122
1123 tx->tx_nwrq++;
1124 wrq++;
1125 sge++;
1126 }
1127 too_big:
1128 if (rc < 0) /* no RDMA if completing with failure */
1129 tx->tx_nwrq = 0;
1130
1131 ibmsg->ibm_u.completion.ibcm_status = rc;
1132 ibmsg->ibm_u.completion.ibcm_cookie = dstcookie;
1133 kiblnd_init_tx_msg(conn->ibc_peer->ibp_ni, tx,
1134 type, sizeof(struct kib_completion_msg));
1135
1136 return rc;
1137 }
1138
1139 static void
1140 kiblnd_queue_tx_locked(struct kib_tx *tx, struct kib_conn *conn)
1141 {
1142 struct list_head *q;
1143
1144 LASSERT(tx->tx_nwrq > 0); /* work items set up */
1145 LASSERT(!tx->tx_queued); /* not queued for sending already */
1146 LASSERT(conn->ibc_state >= IBLND_CONN_ESTABLISHED);
1147
1148 tx->tx_queued = 1;
1149 tx->tx_deadline = jiffies +
1150 msecs_to_jiffies(*kiblnd_tunables.kib_timeout *
1151 MSEC_PER_SEC);
1152
1153 if (!tx->tx_conn) {
1154 kiblnd_conn_addref(conn);
1155 tx->tx_conn = conn;
1156 LASSERT(tx->tx_msg->ibm_type != IBLND_MSG_PUT_DONE);
1157 } else {
1158 /* PUT_DONE first attached to conn as a PUT_REQ */
1159 LASSERT(tx->tx_conn == conn);
1160 LASSERT(tx->tx_msg->ibm_type == IBLND_MSG_PUT_DONE);
1161 }
1162
1163 switch (tx->tx_msg->ibm_type) {
1164 default:
1165 LBUG();
1166
1167 case IBLND_MSG_PUT_REQ:
1168 case IBLND_MSG_GET_REQ:
1169 q = &conn->ibc_tx_queue_rsrvd;
1170 break;
1171
1172 case IBLND_MSG_PUT_NAK:
1173 case IBLND_MSG_PUT_ACK:
1174 case IBLND_MSG_PUT_DONE:
1175 case IBLND_MSG_GET_DONE:
1176 q = &conn->ibc_tx_queue_nocred;
1177 break;
1178
1179 case IBLND_MSG_NOOP:
1180 if (IBLND_OOB_CAPABLE(conn->ibc_version))
1181 q = &conn->ibc_tx_queue_nocred;
1182 else
1183 q = &conn->ibc_tx_noops;
1184 break;
1185
1186 case IBLND_MSG_IMMEDIATE:
1187 q = &conn->ibc_tx_queue;
1188 break;
1189 }
1190
1191 list_add_tail(&tx->tx_list, q);
1192 }
1193
1194 static void
1195 kiblnd_queue_tx(struct kib_tx *tx, struct kib_conn *conn)
1196 {
1197 spin_lock(&conn->ibc_lock);
1198 kiblnd_queue_tx_locked(tx, conn);
1199 kiblnd_check_sends_locked(conn);
1200 spin_unlock(&conn->ibc_lock);
1201 }
1202
1203 static int kiblnd_resolve_addr(struct rdma_cm_id *cmid,
1204 struct sockaddr_in *srcaddr,
1205 struct sockaddr_in *dstaddr,
1206 int timeout_ms)
1207 {
1208 unsigned short port;
1209 int rc;
1210
1211 /* allow the port to be reused */
1212 rc = rdma_set_reuseaddr(cmid, 1);
1213 if (rc) {
1214 CERROR("Unable to set reuse on cmid: %d\n", rc);
1215 return rc;
1216 }
1217
1218 /* look for a free privileged port */
1219 for (port = PROT_SOCK - 1; port > 0; port--) {
1220 srcaddr->sin_port = htons(port);
1221 rc = rdma_resolve_addr(cmid,
1222 (struct sockaddr *)srcaddr,
1223 (struct sockaddr *)dstaddr,
1224 timeout_ms);
1225 if (!rc) {
1226 CDEBUG(D_NET, "bound to port %hu\n", port);
1227 return 0;
1228 } else if (rc == -EADDRINUSE || rc == -EADDRNOTAVAIL) {
1229 CDEBUG(D_NET, "bind to port %hu failed: %d\n",
1230 port, rc);
1231 } else {
1232 return rc;
1233 }
1234 }
1235
1236 CERROR("Failed to bind to a free privileged port\n");
1237 return rc;
1238 }
1239
1240 static void
1241 kiblnd_connect_peer(struct kib_peer *peer)
1242 {
1243 struct rdma_cm_id *cmid;
1244 struct kib_dev *dev;
1245 struct kib_net *net = peer->ibp_ni->ni_data;
1246 struct sockaddr_in srcaddr;
1247 struct sockaddr_in dstaddr;
1248 int rc;
1249
1250 LASSERT(net);
1251 LASSERT(peer->ibp_connecting > 0);
1252 LASSERT(!peer->ibp_reconnecting);
1253
1254 cmid = kiblnd_rdma_create_id(kiblnd_cm_callback, peer, RDMA_PS_TCP,
1255 IB_QPT_RC);
1256
1257 if (IS_ERR(cmid)) {
1258 CERROR("Can't create CMID for %s: %ld\n",
1259 libcfs_nid2str(peer->ibp_nid), PTR_ERR(cmid));
1260 rc = PTR_ERR(cmid);
1261 goto failed;
1262 }
1263
1264 dev = net->ibn_dev;
1265 memset(&srcaddr, 0, sizeof(srcaddr));
1266 srcaddr.sin_family = AF_INET;
1267 srcaddr.sin_addr.s_addr = htonl(dev->ibd_ifip);
1268
1269 memset(&dstaddr, 0, sizeof(dstaddr));
1270 dstaddr.sin_family = AF_INET;
1271 dstaddr.sin_port = htons(*kiblnd_tunables.kib_service);
1272 dstaddr.sin_addr.s_addr = htonl(LNET_NIDADDR(peer->ibp_nid));
1273
1274 kiblnd_peer_addref(peer); /* cmid's ref */
1275
1276 if (*kiblnd_tunables.kib_use_priv_port) {
1277 rc = kiblnd_resolve_addr(cmid, &srcaddr, &dstaddr,
1278 *kiblnd_tunables.kib_timeout * 1000);
1279 } else {
1280 rc = rdma_resolve_addr(cmid,
1281 (struct sockaddr *)&srcaddr,
1282 (struct sockaddr *)&dstaddr,
1283 *kiblnd_tunables.kib_timeout * 1000);
1284 }
1285 if (rc) {
1286 /* Can't initiate address resolution: */
1287 CERROR("Can't resolve addr for %s: %d\n",
1288 libcfs_nid2str(peer->ibp_nid), rc);
1289 goto failed2;
1290 }
1291
1292 LASSERT(cmid->device);
1293 CDEBUG(D_NET, "%s: connection bound to %s:%pI4h:%s\n",
1294 libcfs_nid2str(peer->ibp_nid), dev->ibd_ifname,
1295 &dev->ibd_ifip, cmid->device->name);
1296
1297 return;
1298
1299 failed2:
1300 kiblnd_peer_connect_failed(peer, 1, rc);
1301 kiblnd_peer_decref(peer); /* cmid's ref */
1302 rdma_destroy_id(cmid);
1303 return;
1304 failed:
1305 kiblnd_peer_connect_failed(peer, 1, rc);
1306 }
1307
1308 bool
1309 kiblnd_reconnect_peer(struct kib_peer *peer)
1310 {
1311 rwlock_t *glock = &kiblnd_data.kib_global_lock;
1312 char *reason = NULL;
1313 struct list_head txs;
1314 unsigned long flags;
1315
1316 INIT_LIST_HEAD(&txs);
1317
1318 write_lock_irqsave(glock, flags);
1319 if (!peer->ibp_reconnecting) {
1320 if (peer->ibp_accepting)
1321 reason = "accepting";
1322 else if (peer->ibp_connecting)
1323 reason = "connecting";
1324 else if (!list_empty(&peer->ibp_conns))
1325 reason = "connected";
1326 else /* connected then closed */
1327 reason = "closed";
1328
1329 goto no_reconnect;
1330 }
1331
1332 LASSERT(!peer->ibp_accepting && !peer->ibp_connecting &&
1333 list_empty(&peer->ibp_conns));
1334 peer->ibp_reconnecting = 0;
1335
1336 if (!kiblnd_peer_active(peer)) {
1337 list_splice_init(&peer->ibp_tx_queue, &txs);
1338 reason = "unlinked";
1339 goto no_reconnect;
1340 }
1341
1342 peer->ibp_connecting++;
1343 peer->ibp_reconnected++;
1344 write_unlock_irqrestore(glock, flags);
1345
1346 kiblnd_connect_peer(peer);
1347 return true;
1348
1349 no_reconnect:
1350 write_unlock_irqrestore(glock, flags);
1351
1352 CWARN("Abort reconnection of %s: %s\n",
1353 libcfs_nid2str(peer->ibp_nid), reason);
1354 kiblnd_txlist_done(peer->ibp_ni, &txs, -ECONNABORTED);
1355 return false;
1356 }
1357
1358 void
1359 kiblnd_launch_tx(struct lnet_ni *ni, struct kib_tx *tx, lnet_nid_t nid)
1360 {
1361 struct kib_peer *peer;
1362 struct kib_peer *peer2;
1363 struct kib_conn *conn;
1364 rwlock_t *g_lock = &kiblnd_data.kib_global_lock;
1365 unsigned long flags;
1366 int rc;
1367
1368 /*
1369 * If I get here, I've committed to send, so I complete the tx with
1370 * failure on any problems
1371 */
1372 LASSERT(!tx || !tx->tx_conn); /* only set when assigned a conn */
1373 LASSERT(!tx || tx->tx_nwrq > 0); /* work items have been set up */
1374
1375 /*
1376 * First time, just use a read lock since I expect to find my peer
1377 * connected
1378 */
1379 read_lock_irqsave(g_lock, flags);
1380
1381 peer = kiblnd_find_peer_locked(nid);
1382 if (peer && !list_empty(&peer->ibp_conns)) {
1383 /* Found a peer with an established connection */
1384 conn = kiblnd_get_conn_locked(peer);
1385 kiblnd_conn_addref(conn); /* 1 ref for me... */
1386
1387 read_unlock_irqrestore(g_lock, flags);
1388
1389 if (tx)
1390 kiblnd_queue_tx(tx, conn);
1391 kiblnd_conn_decref(conn); /* ...to here */
1392 return;
1393 }
1394
1395 read_unlock(g_lock);
1396 /* Re-try with a write lock */
1397 write_lock(g_lock);
1398
1399 peer = kiblnd_find_peer_locked(nid);
1400 if (peer) {
1401 if (list_empty(&peer->ibp_conns)) {
1402 /* found a peer, but it's still connecting... */
1403 LASSERT(kiblnd_peer_connecting(peer));
1404 if (tx)
1405 list_add_tail(&tx->tx_list,
1406 &peer->ibp_tx_queue);
1407 write_unlock_irqrestore(g_lock, flags);
1408 } else {
1409 conn = kiblnd_get_conn_locked(peer);
1410 kiblnd_conn_addref(conn); /* 1 ref for me... */
1411
1412 write_unlock_irqrestore(g_lock, flags);
1413
1414 if (tx)
1415 kiblnd_queue_tx(tx, conn);
1416 kiblnd_conn_decref(conn); /* ...to here */
1417 }
1418 return;
1419 }
1420
1421 write_unlock_irqrestore(g_lock, flags);
1422
1423 /* Allocate a peer ready to add to the peer table and retry */
1424 rc = kiblnd_create_peer(ni, &peer, nid);
1425 if (rc) {
1426 CERROR("Can't create peer %s\n", libcfs_nid2str(nid));
1427 if (tx) {
1428 tx->tx_status = -EHOSTUNREACH;
1429 tx->tx_waiting = 0;
1430 kiblnd_tx_done(ni, tx);
1431 }
1432 return;
1433 }
1434
1435 write_lock_irqsave(g_lock, flags);
1436
1437 peer2 = kiblnd_find_peer_locked(nid);
1438 if (peer2) {
1439 if (list_empty(&peer2->ibp_conns)) {
1440 /* found a peer, but it's still connecting... */
1441 LASSERT(kiblnd_peer_connecting(peer2));
1442 if (tx)
1443 list_add_tail(&tx->tx_list,
1444 &peer2->ibp_tx_queue);
1445 write_unlock_irqrestore(g_lock, flags);
1446 } else {
1447 conn = kiblnd_get_conn_locked(peer2);
1448 kiblnd_conn_addref(conn); /* 1 ref for me... */
1449
1450 write_unlock_irqrestore(g_lock, flags);
1451
1452 if (tx)
1453 kiblnd_queue_tx(tx, conn);
1454 kiblnd_conn_decref(conn); /* ...to here */
1455 }
1456
1457 kiblnd_peer_decref(peer);
1458 return;
1459 }
1460
1461 /* Brand new peer */
1462 LASSERT(!peer->ibp_connecting);
1463 peer->ibp_connecting = 1;
1464
1465 /* always called with a ref on ni, which prevents ni being shutdown */
1466 LASSERT(!((struct kib_net *)ni->ni_data)->ibn_shutdown);
1467
1468 if (tx)
1469 list_add_tail(&tx->tx_list, &peer->ibp_tx_queue);
1470
1471 kiblnd_peer_addref(peer);
1472 list_add_tail(&peer->ibp_list, kiblnd_nid2peerlist(nid));
1473
1474 write_unlock_irqrestore(g_lock, flags);
1475
1476 kiblnd_connect_peer(peer);
1477 kiblnd_peer_decref(peer);
1478 }
1479
1480 int
1481 kiblnd_send(struct lnet_ni *ni, void *private, struct lnet_msg *lntmsg)
1482 {
1483 struct lnet_hdr *hdr = &lntmsg->msg_hdr;
1484 int type = lntmsg->msg_type;
1485 struct lnet_process_id target = lntmsg->msg_target;
1486 int target_is_router = lntmsg->msg_target_is_router;
1487 int routing = lntmsg->msg_routing;
1488 unsigned int payload_niov = lntmsg->msg_niov;
1489 struct kvec *payload_iov = lntmsg->msg_iov;
1490 struct bio_vec *payload_kiov = lntmsg->msg_kiov;
1491 unsigned int payload_offset = lntmsg->msg_offset;
1492 unsigned int payload_nob = lntmsg->msg_len;
1493 struct iov_iter from;
1494 struct kib_msg *ibmsg;
1495 struct kib_rdma_desc *rd;
1496 struct kib_tx *tx;
1497 int nob;
1498 int rc;
1499
1500 /* NB 'private' is different depending on what we're sending.... */
1501
1502 CDEBUG(D_NET, "sending %d bytes in %d frags to %s\n",
1503 payload_nob, payload_niov, libcfs_id2str(target));
1504
1505 LASSERT(!payload_nob || payload_niov > 0);
1506 LASSERT(payload_niov <= LNET_MAX_IOV);
1507
1508 /* Thread context */
1509 LASSERT(!in_interrupt());
1510 /* payload is either all vaddrs or all pages */
1511 LASSERT(!(payload_kiov && payload_iov));
1512
1513 if (payload_kiov)
1514 iov_iter_bvec(&from, ITER_BVEC | WRITE,
1515 payload_kiov, payload_niov,
1516 payload_nob + payload_offset);
1517 else
1518 iov_iter_kvec(&from, ITER_KVEC | WRITE,
1519 payload_iov, payload_niov,
1520 payload_nob + payload_offset);
1521
1522 iov_iter_advance(&from, payload_offset);
1523
1524 switch (type) {
1525 default:
1526 LBUG();
1527 return -EIO;
1528
1529 case LNET_MSG_ACK:
1530 LASSERT(!payload_nob);
1531 break;
1532
1533 case LNET_MSG_GET:
1534 if (routing || target_is_router)
1535 break; /* send IMMEDIATE */
1536
1537 /* is the REPLY message too small for RDMA? */
1538 nob = offsetof(struct kib_msg, ibm_u.immediate.ibim_payload[lntmsg->msg_md->md_length]);
1539 if (nob <= IBLND_MSG_SIZE)
1540 break; /* send IMMEDIATE */
1541
1542 tx = kiblnd_get_idle_tx(ni, target.nid);
1543 if (!tx) {
1544 CERROR("Can't allocate txd for GET to %s\n",
1545 libcfs_nid2str(target.nid));
1546 return -ENOMEM;
1547 }
1548
1549 ibmsg = tx->tx_msg;
1550 rd = &ibmsg->ibm_u.get.ibgm_rd;
1551 if (!(lntmsg->msg_md->md_options & LNET_MD_KIOV))
1552 rc = kiblnd_setup_rd_iov(ni, tx, rd,
1553 lntmsg->msg_md->md_niov,
1554 lntmsg->msg_md->md_iov.iov,
1555 0, lntmsg->msg_md->md_length);
1556 else
1557 rc = kiblnd_setup_rd_kiov(ni, tx, rd,
1558 lntmsg->msg_md->md_niov,
1559 lntmsg->msg_md->md_iov.kiov,
1560 0, lntmsg->msg_md->md_length);
1561 if (rc) {
1562 CERROR("Can't setup GET sink for %s: %d\n",
1563 libcfs_nid2str(target.nid), rc);
1564 kiblnd_tx_done(ni, tx);
1565 return -EIO;
1566 }
1567
1568 nob = offsetof(struct kib_get_msg, ibgm_rd.rd_frags[rd->rd_nfrags]);
1569 ibmsg->ibm_u.get.ibgm_cookie = tx->tx_cookie;
1570 ibmsg->ibm_u.get.ibgm_hdr = *hdr;
1571
1572 kiblnd_init_tx_msg(ni, tx, IBLND_MSG_GET_REQ, nob);
1573
1574 tx->tx_lntmsg[1] = lnet_create_reply_msg(ni, lntmsg);
1575 if (!tx->tx_lntmsg[1]) {
1576 CERROR("Can't create reply for GET -> %s\n",
1577 libcfs_nid2str(target.nid));
1578 kiblnd_tx_done(ni, tx);
1579 return -EIO;
1580 }
1581
1582 tx->tx_lntmsg[0] = lntmsg; /* finalise lntmsg[0,1] on completion */
1583 tx->tx_waiting = 1; /* waiting for GET_DONE */
1584 kiblnd_launch_tx(ni, tx, target.nid);
1585 return 0;
1586
1587 case LNET_MSG_REPLY:
1588 case LNET_MSG_PUT:
1589 /* Is the payload small enough not to need RDMA? */
1590 nob = offsetof(struct kib_msg, ibm_u.immediate.ibim_payload[payload_nob]);
1591 if (nob <= IBLND_MSG_SIZE)
1592 break; /* send IMMEDIATE */
1593
1594 tx = kiblnd_get_idle_tx(ni, target.nid);
1595 if (!tx) {
1596 CERROR("Can't allocate %s txd for %s\n",
1597 type == LNET_MSG_PUT ? "PUT" : "REPLY",
1598 libcfs_nid2str(target.nid));
1599 return -ENOMEM;
1600 }
1601
1602 if (!payload_kiov)
1603 rc = kiblnd_setup_rd_iov(ni, tx, tx->tx_rd,
1604 payload_niov, payload_iov,
1605 payload_offset, payload_nob);
1606 else
1607 rc = kiblnd_setup_rd_kiov(ni, tx, tx->tx_rd,
1608 payload_niov, payload_kiov,
1609 payload_offset, payload_nob);
1610 if (rc) {
1611 CERROR("Can't setup PUT src for %s: %d\n",
1612 libcfs_nid2str(target.nid), rc);
1613 kiblnd_tx_done(ni, tx);
1614 return -EIO;
1615 }
1616
1617 ibmsg = tx->tx_msg;
1618 ibmsg->ibm_u.putreq.ibprm_hdr = *hdr;
1619 ibmsg->ibm_u.putreq.ibprm_cookie = tx->tx_cookie;
1620 kiblnd_init_tx_msg(ni, tx, IBLND_MSG_PUT_REQ, sizeof(struct kib_putreq_msg));
1621
1622 tx->tx_lntmsg[0] = lntmsg; /* finalise lntmsg on completion */
1623 tx->tx_waiting = 1; /* waiting for PUT_{ACK,NAK} */
1624 kiblnd_launch_tx(ni, tx, target.nid);
1625 return 0;
1626 }
1627
1628 /* send IMMEDIATE */
1629
1630 LASSERT(offsetof(struct kib_msg, ibm_u.immediate.ibim_payload[payload_nob])
1631 <= IBLND_MSG_SIZE);
1632
1633 tx = kiblnd_get_idle_tx(ni, target.nid);
1634 if (!tx) {
1635 CERROR("Can't send %d to %s: tx descs exhausted\n",
1636 type, libcfs_nid2str(target.nid));
1637 return -ENOMEM;
1638 }
1639
1640 ibmsg = tx->tx_msg;
1641 ibmsg->ibm_u.immediate.ibim_hdr = *hdr;
1642
1643 rc = copy_from_iter(&ibmsg->ibm_u.immediate.ibim_payload, payload_nob,
1644 &from);
1645 if (rc != payload_nob) {
1646 kiblnd_pool_free_node(&tx->tx_pool->tpo_pool, &tx->tx_list);
1647 return -EFAULT;
1648 }
1649
1650 nob = offsetof(struct kib_immediate_msg, ibim_payload[payload_nob]);
1651 kiblnd_init_tx_msg(ni, tx, IBLND_MSG_IMMEDIATE, nob);
1652
1653 tx->tx_lntmsg[0] = lntmsg; /* finalise lntmsg on completion */
1654 kiblnd_launch_tx(ni, tx, target.nid);
1655 return 0;
1656 }
1657
1658 static void
1659 kiblnd_reply(struct lnet_ni *ni, struct kib_rx *rx, struct lnet_msg *lntmsg)
1660 {
1661 struct lnet_process_id target = lntmsg->msg_target;
1662 unsigned int niov = lntmsg->msg_niov;
1663 struct kvec *iov = lntmsg->msg_iov;
1664 struct bio_vec *kiov = lntmsg->msg_kiov;
1665 unsigned int offset = lntmsg->msg_offset;
1666 unsigned int nob = lntmsg->msg_len;
1667 struct kib_tx *tx;
1668 int rc;
1669
1670 tx = kiblnd_get_idle_tx(ni, rx->rx_conn->ibc_peer->ibp_nid);
1671 if (!tx) {
1672 CERROR("Can't get tx for REPLY to %s\n",
1673 libcfs_nid2str(target.nid));
1674 goto failed_0;
1675 }
1676
1677 if (!nob)
1678 rc = 0;
1679 else if (!kiov)
1680 rc = kiblnd_setup_rd_iov(ni, tx, tx->tx_rd,
1681 niov, iov, offset, nob);
1682 else
1683 rc = kiblnd_setup_rd_kiov(ni, tx, tx->tx_rd,
1684 niov, kiov, offset, nob);
1685
1686 if (rc) {
1687 CERROR("Can't setup GET src for %s: %d\n",
1688 libcfs_nid2str(target.nid), rc);
1689 goto failed_1;
1690 }
1691
1692 rc = kiblnd_init_rdma(rx->rx_conn, tx,
1693 IBLND_MSG_GET_DONE, nob,
1694 &rx->rx_msg->ibm_u.get.ibgm_rd,
1695 rx->rx_msg->ibm_u.get.ibgm_cookie);
1696 if (rc < 0) {
1697 CERROR("Can't setup rdma for GET from %s: %d\n",
1698 libcfs_nid2str(target.nid), rc);
1699 goto failed_1;
1700 }
1701
1702 if (!nob) {
1703 /* No RDMA: local completion may happen now! */
1704 lnet_finalize(ni, lntmsg, 0);
1705 } else {
1706 /* RDMA: lnet_finalize(lntmsg) when it completes */
1707 tx->tx_lntmsg[0] = lntmsg;
1708 }
1709
1710 kiblnd_queue_tx(tx, rx->rx_conn);
1711 return;
1712
1713 failed_1:
1714 kiblnd_tx_done(ni, tx);
1715 failed_0:
1716 lnet_finalize(ni, lntmsg, -EIO);
1717 }
1718
1719 int
1720 kiblnd_recv(struct lnet_ni *ni, void *private, struct lnet_msg *lntmsg,
1721 int delayed, struct iov_iter *to, unsigned int rlen)
1722 {
1723 struct kib_rx *rx = private;
1724 struct kib_msg *rxmsg = rx->rx_msg;
1725 struct kib_conn *conn = rx->rx_conn;
1726 struct kib_tx *tx;
1727 int nob;
1728 int post_credit = IBLND_POSTRX_PEER_CREDIT;
1729 int rc = 0;
1730
1731 LASSERT(iov_iter_count(to) <= rlen);
1732 LASSERT(!in_interrupt());
1733 /* Either all pages or all vaddrs */
1734
1735 switch (rxmsg->ibm_type) {
1736 default:
1737 LBUG();
1738
1739 case IBLND_MSG_IMMEDIATE:
1740 nob = offsetof(struct kib_msg, ibm_u.immediate.ibim_payload[rlen]);
1741 if (nob > rx->rx_nob) {
1742 CERROR("Immediate message from %s too big: %d(%d)\n",
1743 libcfs_nid2str(rxmsg->ibm_u.immediate.ibim_hdr.src_nid),
1744 nob, rx->rx_nob);
1745 rc = -EPROTO;
1746 break;
1747 }
1748
1749 rc = copy_to_iter(&rxmsg->ibm_u.immediate.ibim_payload, rlen,
1750 to);
1751 if (rc != rlen) {
1752 rc = -EFAULT;
1753 break;
1754 }
1755
1756 rc = 0;
1757 lnet_finalize(ni, lntmsg, 0);
1758 break;
1759
1760 case IBLND_MSG_PUT_REQ: {
1761 struct kib_msg *txmsg;
1762 struct kib_rdma_desc *rd;
1763
1764 if (!iov_iter_count(to)) {
1765 lnet_finalize(ni, lntmsg, 0);
1766 kiblnd_send_completion(rx->rx_conn, IBLND_MSG_PUT_NAK, 0,
1767 rxmsg->ibm_u.putreq.ibprm_cookie);
1768 break;
1769 }
1770
1771 tx = kiblnd_get_idle_tx(ni, conn->ibc_peer->ibp_nid);
1772 if (!tx) {
1773 CERROR("Can't allocate tx for %s\n",
1774 libcfs_nid2str(conn->ibc_peer->ibp_nid));
1775 /* Not replying will break the connection */
1776 rc = -ENOMEM;
1777 break;
1778 }
1779
1780 txmsg = tx->tx_msg;
1781 rd = &txmsg->ibm_u.putack.ibpam_rd;
1782 if (!(to->type & ITER_BVEC))
1783 rc = kiblnd_setup_rd_iov(ni, tx, rd,
1784 to->nr_segs, to->kvec,
1785 to->iov_offset,
1786 iov_iter_count(to));
1787 else
1788 rc = kiblnd_setup_rd_kiov(ni, tx, rd,
1789 to->nr_segs, to->bvec,
1790 to->iov_offset,
1791 iov_iter_count(to));
1792 if (rc) {
1793 CERROR("Can't setup PUT sink for %s: %d\n",
1794 libcfs_nid2str(conn->ibc_peer->ibp_nid), rc);
1795 kiblnd_tx_done(ni, tx);
1796 /* tell peer it's over */
1797 kiblnd_send_completion(rx->rx_conn, IBLND_MSG_PUT_NAK, rc,
1798 rxmsg->ibm_u.putreq.ibprm_cookie);
1799 break;
1800 }
1801
1802 nob = offsetof(struct kib_putack_msg, ibpam_rd.rd_frags[rd->rd_nfrags]);
1803 txmsg->ibm_u.putack.ibpam_src_cookie = rxmsg->ibm_u.putreq.ibprm_cookie;
1804 txmsg->ibm_u.putack.ibpam_dst_cookie = tx->tx_cookie;
1805
1806 kiblnd_init_tx_msg(ni, tx, IBLND_MSG_PUT_ACK, nob);
1807
1808 tx->tx_lntmsg[0] = lntmsg; /* finalise lntmsg on completion */
1809 tx->tx_waiting = 1; /* waiting for PUT_DONE */
1810 kiblnd_queue_tx(tx, conn);
1811
1812 /* reposted buffer reserved for PUT_DONE */
1813 post_credit = IBLND_POSTRX_NO_CREDIT;
1814 break;
1815 }
1816
1817 case IBLND_MSG_GET_REQ:
1818 if (lntmsg) {
1819 /* Optimized GET; RDMA lntmsg's payload */
1820 kiblnd_reply(ni, rx, lntmsg);
1821 } else {
1822 /* GET didn't match anything */
1823 kiblnd_send_completion(rx->rx_conn, IBLND_MSG_GET_DONE,
1824 -ENODATA,
1825 rxmsg->ibm_u.get.ibgm_cookie);
1826 }
1827 break;
1828 }
1829
1830 kiblnd_post_rx(rx, post_credit);
1831 return rc;
1832 }
1833
1834 int
1835 kiblnd_thread_start(int (*fn)(void *arg), void *arg, char *name)
1836 {
1837 struct task_struct *task = kthread_run(fn, arg, "%s", name);
1838
1839 if (IS_ERR(task))
1840 return PTR_ERR(task);
1841
1842 atomic_inc(&kiblnd_data.kib_nthreads);
1843 return 0;
1844 }
1845
1846 static void
1847 kiblnd_thread_fini(void)
1848 {
1849 atomic_dec(&kiblnd_data.kib_nthreads);
1850 }
1851
1852 static void
1853 kiblnd_peer_alive(struct kib_peer *peer)
1854 {
1855 /* This is racy, but everyone's only writing cfs_time_current() */
1856 peer->ibp_last_alive = cfs_time_current();
1857 mb();
1858 }
1859
1860 static void
1861 kiblnd_peer_notify(struct kib_peer *peer)
1862 {
1863 int error = 0;
1864 unsigned long last_alive = 0;
1865 unsigned long flags;
1866
1867 read_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
1868
1869 if (kiblnd_peer_idle(peer) && peer->ibp_error) {
1870 error = peer->ibp_error;
1871 peer->ibp_error = 0;
1872
1873 last_alive = peer->ibp_last_alive;
1874 }
1875
1876 read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
1877
1878 if (error)
1879 lnet_notify(peer->ibp_ni,
1880 peer->ibp_nid, 0, last_alive);
1881 }
1882
1883 void
1884 kiblnd_close_conn_locked(struct kib_conn *conn, int error)
1885 {
1886 /*
1887 * This just does the immediate housekeeping. 'error' is zero for a
1888 * normal shutdown which can happen only after the connection has been
1889 * established. If the connection is established, schedule the
1890 * connection to be finished off by the connd. Otherwise the connd is
1891 * already dealing with it (either to set it up or tear it down).
1892 * Caller holds kib_global_lock exclusively in irq context
1893 */
1894 struct kib_peer *peer = conn->ibc_peer;
1895 struct kib_dev *dev;
1896 unsigned long flags;
1897
1898 LASSERT(error || conn->ibc_state >= IBLND_CONN_ESTABLISHED);
1899
1900 if (error && !conn->ibc_comms_error)
1901 conn->ibc_comms_error = error;
1902
1903 if (conn->ibc_state != IBLND_CONN_ESTABLISHED)
1904 return; /* already being handled */
1905
1906 if (!error &&
1907 list_empty(&conn->ibc_tx_noops) &&
1908 list_empty(&conn->ibc_tx_queue) &&
1909 list_empty(&conn->ibc_tx_queue_rsrvd) &&
1910 list_empty(&conn->ibc_tx_queue_nocred) &&
1911 list_empty(&conn->ibc_active_txs)) {
1912 CDEBUG(D_NET, "closing conn to %s\n",
1913 libcfs_nid2str(peer->ibp_nid));
1914 } else {
1915 CNETERR("Closing conn to %s: error %d%s%s%s%s%s\n",
1916 libcfs_nid2str(peer->ibp_nid), error,
1917 list_empty(&conn->ibc_tx_queue) ? "" : "(sending)",
1918 list_empty(&conn->ibc_tx_noops) ? "" : "(sending_noops)",
1919 list_empty(&conn->ibc_tx_queue_rsrvd) ? "" : "(sending_rsrvd)",
1920 list_empty(&conn->ibc_tx_queue_nocred) ? "" : "(sending_nocred)",
1921 list_empty(&conn->ibc_active_txs) ? "" : "(waiting)");
1922 }
1923
1924 dev = ((struct kib_net *)peer->ibp_ni->ni_data)->ibn_dev;
1925 list_del(&conn->ibc_list);
1926 /* connd (see below) takes over ibc_list's ref */
1927
1928 if (list_empty(&peer->ibp_conns) && /* no more conns */
1929 kiblnd_peer_active(peer)) { /* still in peer table */
1930 kiblnd_unlink_peer_locked(peer);
1931
1932 /* set/clear error on last conn */
1933 peer->ibp_error = conn->ibc_comms_error;
1934 }
1935
1936 kiblnd_set_conn_state(conn, IBLND_CONN_CLOSING);
1937
1938 if (error &&
1939 kiblnd_dev_can_failover(dev)) {
1940 list_add_tail(&dev->ibd_fail_list,
1941 &kiblnd_data.kib_failed_devs);
1942 wake_up(&kiblnd_data.kib_failover_waitq);
1943 }
1944
1945 spin_lock_irqsave(&kiblnd_data.kib_connd_lock, flags);
1946
1947 list_add_tail(&conn->ibc_list, &kiblnd_data.kib_connd_conns);
1948 wake_up(&kiblnd_data.kib_connd_waitq);
1949
1950 spin_unlock_irqrestore(&kiblnd_data.kib_connd_lock, flags);
1951 }
1952
1953 void
1954 kiblnd_close_conn(struct kib_conn *conn, int error)
1955 {
1956 unsigned long flags;
1957
1958 write_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
1959
1960 kiblnd_close_conn_locked(conn, error);
1961
1962 write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
1963 }
1964
1965 static void
1966 kiblnd_handle_early_rxs(struct kib_conn *conn)
1967 {
1968 unsigned long flags;
1969 struct kib_rx *rx;
1970 struct kib_rx *tmp;
1971
1972 LASSERT(!in_interrupt());
1973 LASSERT(conn->ibc_state >= IBLND_CONN_ESTABLISHED);
1974
1975 write_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
1976 list_for_each_entry_safe(rx, tmp, &conn->ibc_early_rxs, rx_list) {
1977 list_del(&rx->rx_list);
1978 write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
1979
1980 kiblnd_handle_rx(rx);
1981
1982 write_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
1983 }
1984 write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
1985 }
1986
1987 static void
1988 kiblnd_abort_txs(struct kib_conn *conn, struct list_head *txs)
1989 {
1990 LIST_HEAD(zombies);
1991 struct list_head *tmp;
1992 struct list_head *nxt;
1993 struct kib_tx *tx;
1994
1995 spin_lock(&conn->ibc_lock);
1996
1997 list_for_each_safe(tmp, nxt, txs) {
1998 tx = list_entry(tmp, struct kib_tx, tx_list);
1999
2000 if (txs == &conn->ibc_active_txs) {
2001 LASSERT(!tx->tx_queued);
2002 LASSERT(tx->tx_waiting || tx->tx_sending);
2003 } else {
2004 LASSERT(tx->tx_queued);
2005 }
2006
2007 tx->tx_status = -ECONNABORTED;
2008 tx->tx_waiting = 0;
2009
2010 if (!tx->tx_sending) {
2011 tx->tx_queued = 0;
2012 list_del(&tx->tx_list);
2013 list_add(&tx->tx_list, &zombies);
2014 }
2015 }
2016
2017 spin_unlock(&conn->ibc_lock);
2018
2019 kiblnd_txlist_done(conn->ibc_peer->ibp_ni, &zombies, -ECONNABORTED);
2020 }
2021
2022 static void
2023 kiblnd_finalise_conn(struct kib_conn *conn)
2024 {
2025 LASSERT(!in_interrupt());
2026 LASSERT(conn->ibc_state > IBLND_CONN_INIT);
2027
2028 kiblnd_set_conn_state(conn, IBLND_CONN_DISCONNECTED);
2029
2030 /*
2031 * abort_receives moves QP state to IB_QPS_ERR. This is only required
2032 * for connections that didn't get as far as being connected, because
2033 * rdma_disconnect() does this for free.
2034 */
2035 kiblnd_abort_receives(conn);
2036
2037 /*
2038 * Complete all tx descs not waiting for sends to complete.
2039 * NB we should be safe from RDMA now that the QP has changed state
2040 */
2041 kiblnd_abort_txs(conn, &conn->ibc_tx_noops);
2042 kiblnd_abort_txs(conn, &conn->ibc_tx_queue);
2043 kiblnd_abort_txs(conn, &conn->ibc_tx_queue_rsrvd);
2044 kiblnd_abort_txs(conn, &conn->ibc_tx_queue_nocred);
2045 kiblnd_abort_txs(conn, &conn->ibc_active_txs);
2046
2047 kiblnd_handle_early_rxs(conn);
2048 }
2049
2050 static void
2051 kiblnd_peer_connect_failed(struct kib_peer *peer, int active, int error)
2052 {
2053 LIST_HEAD(zombies);
2054 unsigned long flags;
2055
2056 LASSERT(error);
2057 LASSERT(!in_interrupt());
2058
2059 write_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
2060
2061 if (active) {
2062 LASSERT(peer->ibp_connecting > 0);
2063 peer->ibp_connecting--;
2064 } else {
2065 LASSERT(peer->ibp_accepting > 0);
2066 peer->ibp_accepting--;
2067 }
2068
2069 if (kiblnd_peer_connecting(peer)) {
2070 /* another connection attempt under way... */
2071 write_unlock_irqrestore(&kiblnd_data.kib_global_lock,
2072 flags);
2073 return;
2074 }
2075
2076 peer->ibp_reconnected = 0;
2077 if (list_empty(&peer->ibp_conns)) {
2078 /* Take peer's blocked transmits to complete with error */
2079 list_add(&zombies, &peer->ibp_tx_queue);
2080 list_del_init(&peer->ibp_tx_queue);
2081
2082 if (kiblnd_peer_active(peer))
2083 kiblnd_unlink_peer_locked(peer);
2084
2085 peer->ibp_error = error;
2086 } else {
2087 /* Can't have blocked transmits if there are connections */
2088 LASSERT(list_empty(&peer->ibp_tx_queue));
2089 }
2090
2091 write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
2092
2093 kiblnd_peer_notify(peer);
2094
2095 if (list_empty(&zombies))
2096 return;
2097
2098 CNETERR("Deleting messages for %s: connection failed\n",
2099 libcfs_nid2str(peer->ibp_nid));
2100
2101 kiblnd_txlist_done(peer->ibp_ni, &zombies, -EHOSTUNREACH);
2102 }
2103
2104 static void
2105 kiblnd_connreq_done(struct kib_conn *conn, int status)
2106 {
2107 struct kib_peer *peer = conn->ibc_peer;
2108 struct kib_tx *tx;
2109 struct kib_tx *tmp;
2110 struct list_head txs;
2111 unsigned long flags;
2112 int active;
2113
2114 active = (conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT);
2115
2116 CDEBUG(D_NET, "%s: active(%d), version(%x), status(%d)\n",
2117 libcfs_nid2str(peer->ibp_nid), active,
2118 conn->ibc_version, status);
2119
2120 LASSERT(!in_interrupt());
2121 LASSERT((conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT &&
2122 peer->ibp_connecting > 0) ||
2123 (conn->ibc_state == IBLND_CONN_PASSIVE_WAIT &&
2124 peer->ibp_accepting > 0));
2125
2126 LIBCFS_FREE(conn->ibc_connvars, sizeof(*conn->ibc_connvars));
2127 conn->ibc_connvars = NULL;
2128
2129 if (status) {
2130 /* failed to establish connection */
2131 kiblnd_peer_connect_failed(peer, active, status);
2132 kiblnd_finalise_conn(conn);
2133 return;
2134 }
2135
2136 /* connection established */
2137 write_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
2138
2139 conn->ibc_last_send = jiffies;
2140 kiblnd_set_conn_state(conn, IBLND_CONN_ESTABLISHED);
2141 kiblnd_peer_alive(peer);
2142
2143 /*
2144 * Add conn to peer's list and nuke any dangling conns from a different
2145 * peer instance...
2146 */
2147 kiblnd_conn_addref(conn); /* +1 ref for ibc_list */
2148 list_add(&conn->ibc_list, &peer->ibp_conns);
2149 peer->ibp_reconnected = 0;
2150 if (active)
2151 peer->ibp_connecting--;
2152 else
2153 peer->ibp_accepting--;
2154
2155 if (!peer->ibp_version) {
2156 peer->ibp_version = conn->ibc_version;
2157 peer->ibp_incarnation = conn->ibc_incarnation;
2158 }
2159
2160 if (peer->ibp_version != conn->ibc_version ||
2161 peer->ibp_incarnation != conn->ibc_incarnation) {
2162 kiblnd_close_stale_conns_locked(peer, conn->ibc_version,
2163 conn->ibc_incarnation);
2164 peer->ibp_version = conn->ibc_version;
2165 peer->ibp_incarnation = conn->ibc_incarnation;
2166 }
2167
2168 /* grab pending txs while I have the lock */
2169 list_add(&txs, &peer->ibp_tx_queue);
2170 list_del_init(&peer->ibp_tx_queue);
2171
2172 if (!kiblnd_peer_active(peer) || /* peer has been deleted */
2173 conn->ibc_comms_error) { /* error has happened already */
2174 struct lnet_ni *ni = peer->ibp_ni;
2175
2176 /* start to shut down connection */
2177 kiblnd_close_conn_locked(conn, -ECONNABORTED);
2178 write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
2179
2180 kiblnd_txlist_done(ni, &txs, -ECONNABORTED);
2181
2182 return;
2183 }
2184
2185 /*
2186 * +1 ref for myself, this connection is visible to other threads
2187 * now, refcount of peer:ibp_conns can be released by connection
2188 * close from either a different thread, or the calling of
2189 * kiblnd_check_sends_locked() below. See bz21911 for details.
2190 */
2191 kiblnd_conn_addref(conn);
2192 write_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
2193
2194 /* Schedule blocked txs */
2195 spin_lock(&conn->ibc_lock);
2196 list_for_each_entry_safe(tx, tmp, &txs, tx_list) {
2197 list_del(&tx->tx_list);
2198
2199 kiblnd_queue_tx_locked(tx, conn);
2200 }
2201 kiblnd_check_sends_locked(conn);
2202 spin_unlock(&conn->ibc_lock);
2203
2204 /* schedule blocked rxs */
2205 kiblnd_handle_early_rxs(conn);
2206
2207 kiblnd_conn_decref(conn);
2208 }
2209
2210 static void
2211 kiblnd_reject(struct rdma_cm_id *cmid, struct kib_rej *rej)
2212 {
2213 int rc;
2214
2215 rc = rdma_reject(cmid, rej, sizeof(*rej));
2216
2217 if (rc)
2218 CWARN("Error %d sending reject\n", rc);
2219 }
2220
2221 static int
2222 kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob)
2223 {
2224 rwlock_t *g_lock = &kiblnd_data.kib_global_lock;
2225 struct kib_msg *reqmsg = priv;
2226 struct kib_msg *ackmsg;
2227 struct kib_dev *ibdev;
2228 struct kib_peer *peer;
2229 struct kib_peer *peer2;
2230 struct kib_conn *conn;
2231 struct lnet_ni *ni = NULL;
2232 struct kib_net *net = NULL;
2233 lnet_nid_t nid;
2234 struct rdma_conn_param cp;
2235 struct kib_rej rej;
2236 int version = IBLND_MSG_VERSION;
2237 unsigned long flags;
2238 int max_frags;
2239 int rc;
2240 struct sockaddr_in *peer_addr;
2241
2242 LASSERT(!in_interrupt());
2243
2244 /* cmid inherits 'context' from the corresponding listener id */
2245 ibdev = (struct kib_dev *)cmid->context;
2246 LASSERT(ibdev);
2247
2248 memset(&rej, 0, sizeof(rej));
2249 rej.ibr_magic = IBLND_MSG_MAGIC;
2250 rej.ibr_why = IBLND_REJECT_FATAL;
2251 rej.ibr_cp.ibcp_max_msg_size = IBLND_MSG_SIZE;
2252
2253 peer_addr = (struct sockaddr_in *)&cmid->route.addr.dst_addr;
2254 if (*kiblnd_tunables.kib_require_priv_port &&
2255 ntohs(peer_addr->sin_port) >= PROT_SOCK) {
2256 __u32 ip = ntohl(peer_addr->sin_addr.s_addr);
2257
2258 CERROR("Peer's port (%pI4h:%hu) is not privileged\n",
2259 &ip, ntohs(peer_addr->sin_port));
2260 goto failed;
2261 }
2262
2263 if (priv_nob < offsetof(struct kib_msg, ibm_type)) {
2264 CERROR("Short connection request\n");
2265 goto failed;
2266 }
2267
2268 /*
2269 * Future protocol version compatibility support! If the
2270 * o2iblnd-specific protocol changes, or when LNET unifies
2271 * protocols over all LNDs, the initial connection will
2272 * negotiate a protocol version. I trap this here to avoid
2273 * console errors; the reject tells the peer which protocol I
2274 * speak.
2275 */
2276 if (reqmsg->ibm_magic == LNET_PROTO_MAGIC ||
2277 reqmsg->ibm_magic == __swab32(LNET_PROTO_MAGIC))
2278 goto failed;
2279 if (reqmsg->ibm_magic == IBLND_MSG_MAGIC &&
2280 reqmsg->ibm_version != IBLND_MSG_VERSION &&
2281 reqmsg->ibm_version != IBLND_MSG_VERSION_1)
2282 goto failed;
2283 if (reqmsg->ibm_magic == __swab32(IBLND_MSG_MAGIC) &&
2284 reqmsg->ibm_version != __swab16(IBLND_MSG_VERSION) &&
2285 reqmsg->ibm_version != __swab16(IBLND_MSG_VERSION_1))
2286 goto failed;
2287
2288 rc = kiblnd_unpack_msg(reqmsg, priv_nob);
2289 if (rc) {
2290 CERROR("Can't parse connection request: %d\n", rc);
2291 goto failed;
2292 }
2293
2294 nid = reqmsg->ibm_srcnid;
2295 ni = lnet_net2ni(LNET_NIDNET(reqmsg->ibm_dstnid));
2296
2297 if (ni) {
2298 net = (struct kib_net *)ni->ni_data;
2299 rej.ibr_incarnation = net->ibn_incarnation;
2300 }
2301
2302 if (!ni || /* no matching net */
2303 ni->ni_nid != reqmsg->ibm_dstnid || /* right NET, wrong NID! */
2304 net->ibn_dev != ibdev) { /* wrong device */
2305 CERROR("Can't accept conn from %s on %s (%s:%d:%pI4h): bad dst nid %s\n",
2306 libcfs_nid2str(nid),
2307 !ni ? "NA" : libcfs_nid2str(ni->ni_nid),
2308 ibdev->ibd_ifname, ibdev->ibd_nnets,
2309 &ibdev->ibd_ifip,
2310 libcfs_nid2str(reqmsg->ibm_dstnid));
2311
2312 goto failed;
2313 }
2314
2315 /* check time stamp as soon as possible */
2316 if (reqmsg->ibm_dststamp &&
2317 reqmsg->ibm_dststamp != net->ibn_incarnation) {
2318 CWARN("Stale connection request\n");
2319 rej.ibr_why = IBLND_REJECT_CONN_STALE;
2320 goto failed;
2321 }
2322
2323 /* I can accept peer's version */
2324 version = reqmsg->ibm_version;
2325
2326 if (reqmsg->ibm_type != IBLND_MSG_CONNREQ) {
2327 CERROR("Unexpected connreq msg type: %x from %s\n",
2328 reqmsg->ibm_type, libcfs_nid2str(nid));
2329 goto failed;
2330 }
2331
2332 if (reqmsg->ibm_u.connparams.ibcp_queue_depth >
2333 kiblnd_msg_queue_size(version, ni)) {
2334 CERROR("Can't accept conn from %s, queue depth too large: %d (<=%d wanted)\n",
2335 libcfs_nid2str(nid),
2336 reqmsg->ibm_u.connparams.ibcp_queue_depth,
2337 kiblnd_msg_queue_size(version, ni));
2338
2339 if (version == IBLND_MSG_VERSION)
2340 rej.ibr_why = IBLND_REJECT_MSG_QUEUE_SIZE;
2341
2342 goto failed;
2343 }
2344
2345 max_frags = reqmsg->ibm_u.connparams.ibcp_max_frags >> IBLND_FRAG_SHIFT;
2346 if (max_frags > kiblnd_rdma_frags(version, ni)) {
2347 CWARN("Can't accept conn from %s (version %x): max message size %d is too large (%d wanted)\n",
2348 libcfs_nid2str(nid), version, max_frags,
2349 kiblnd_rdma_frags(version, ni));
2350
2351 if (version >= IBLND_MSG_VERSION)
2352 rej.ibr_why = IBLND_REJECT_RDMA_FRAGS;
2353
2354 goto failed;
2355 } else if (max_frags < kiblnd_rdma_frags(version, ni) &&
2356 !net->ibn_fmr_ps) {
2357 CWARN("Can't accept conn from %s (version %x): max message size %d incompatible without FMR pool (%d wanted)\n",
2358 libcfs_nid2str(nid), version, max_frags,
2359 kiblnd_rdma_frags(version, ni));
2360
2361 if (version == IBLND_MSG_VERSION)
2362 rej.ibr_why = IBLND_REJECT_RDMA_FRAGS;
2363
2364 goto failed;
2365 }
2366
2367 if (reqmsg->ibm_u.connparams.ibcp_max_msg_size > IBLND_MSG_SIZE) {
2368 CERROR("Can't accept %s: message size %d too big (%d max)\n",
2369 libcfs_nid2str(nid),
2370 reqmsg->ibm_u.connparams.ibcp_max_msg_size,
2371 IBLND_MSG_SIZE);
2372 goto failed;
2373 }
2374
2375 /* assume 'nid' is a new peer; create */
2376 rc = kiblnd_create_peer(ni, &peer, nid);
2377 if (rc) {
2378 CERROR("Can't create peer for %s\n", libcfs_nid2str(nid));
2379 rej.ibr_why = IBLND_REJECT_NO_RESOURCES;
2380 goto failed;
2381 }
2382
2383 /* We have validated the peer's parameters so use those */
2384 peer->ibp_max_frags = max_frags;
2385 peer->ibp_queue_depth = reqmsg->ibm_u.connparams.ibcp_queue_depth;
2386
2387 write_lock_irqsave(g_lock, flags);
2388
2389 peer2 = kiblnd_find_peer_locked(nid);
2390 if (peer2) {
2391 if (!peer2->ibp_version) {
2392 peer2->ibp_version = version;
2393 peer2->ibp_incarnation = reqmsg->ibm_srcstamp;
2394 }
2395
2396 /* not the guy I've talked with */
2397 if (peer2->ibp_incarnation != reqmsg->ibm_srcstamp ||
2398 peer2->ibp_version != version) {
2399 kiblnd_close_peer_conns_locked(peer2, -ESTALE);
2400
2401 if (kiblnd_peer_active(peer2)) {
2402 peer2->ibp_incarnation = reqmsg->ibm_srcstamp;
2403 peer2->ibp_version = version;
2404 }
2405 write_unlock_irqrestore(g_lock, flags);
2406
2407 CWARN("Conn stale %s version %x/%x incarnation %llu/%llu\n",
2408 libcfs_nid2str(nid), peer2->ibp_version, version,
2409 peer2->ibp_incarnation, reqmsg->ibm_srcstamp);
2410
2411 kiblnd_peer_decref(peer);
2412 rej.ibr_why = IBLND_REJECT_CONN_STALE;
2413 goto failed;
2414 }
2415
2416 /*
2417 * Tie-break connection race in favour of the higher NID.
2418 * If we keep running into a race condition multiple times,
2419 * we have to assume that the connection attempt with the
2420 * higher NID is stuck in a connecting state and will never
2421 * recover. As such, we pass through this if-block and let
2422 * the lower NID connection win so we can move forward.
2423 */
2424 if (peer2->ibp_connecting &&
2425 nid < ni->ni_nid && peer2->ibp_races <
2426 MAX_CONN_RACES_BEFORE_ABORT) {
2427 peer2->ibp_races++;
2428 write_unlock_irqrestore(g_lock, flags);
2429
2430 CDEBUG(D_NET, "Conn race %s\n",
2431 libcfs_nid2str(peer2->ibp_nid));
2432
2433 kiblnd_peer_decref(peer);
2434 rej.ibr_why = IBLND_REJECT_CONN_RACE;
2435 goto failed;
2436 }
2437 if (peer2->ibp_races >= MAX_CONN_RACES_BEFORE_ABORT)
2438 CNETERR("Conn race %s: unresolved after %d attempts, letting lower NID win\n",
2439 libcfs_nid2str(peer2->ibp_nid),
2440 MAX_CONN_RACES_BEFORE_ABORT);
2441 /**
2442 * passive connection is allowed even this peer is waiting for
2443 * reconnection.
2444 */
2445 peer2->ibp_reconnecting = 0;
2446 peer2->ibp_races = 0;
2447 peer2->ibp_accepting++;
2448 kiblnd_peer_addref(peer2);
2449
2450 /**
2451 * Race with kiblnd_launch_tx (active connect) to create peer
2452 * so copy validated parameters since we now know what the
2453 * peer's limits are
2454 */
2455 peer2->ibp_max_frags = peer->ibp_max_frags;
2456 peer2->ibp_queue_depth = peer->ibp_queue_depth;
2457
2458 write_unlock_irqrestore(g_lock, flags);
2459 kiblnd_peer_decref(peer);
2460 peer = peer2;
2461 } else {
2462 /* Brand new peer */
2463 LASSERT(!peer->ibp_accepting);
2464 LASSERT(!peer->ibp_version &&
2465 !peer->ibp_incarnation);
2466
2467 peer->ibp_accepting = 1;
2468 peer->ibp_version = version;
2469 peer->ibp_incarnation = reqmsg->ibm_srcstamp;
2470
2471 /* I have a ref on ni that prevents it being shutdown */
2472 LASSERT(!net->ibn_shutdown);
2473
2474 kiblnd_peer_addref(peer);
2475 list_add_tail(&peer->ibp_list, kiblnd_nid2peerlist(nid));
2476
2477 write_unlock_irqrestore(g_lock, flags);
2478 }
2479
2480 conn = kiblnd_create_conn(peer, cmid, IBLND_CONN_PASSIVE_WAIT,
2481 version);
2482 if (!conn) {
2483 kiblnd_peer_connect_failed(peer, 0, -ENOMEM);
2484 kiblnd_peer_decref(peer);
2485 rej.ibr_why = IBLND_REJECT_NO_RESOURCES;
2486 goto failed;
2487 }
2488
2489 /*
2490 * conn now "owns" cmid, so I return success from here on to ensure the
2491 * CM callback doesn't destroy cmid.
2492 */
2493 conn->ibc_incarnation = reqmsg->ibm_srcstamp;
2494 conn->ibc_credits = conn->ibc_queue_depth;
2495 conn->ibc_reserved_credits = conn->ibc_queue_depth;
2496 LASSERT(conn->ibc_credits + conn->ibc_reserved_credits +
2497 IBLND_OOB_MSGS(version) <= IBLND_RX_MSGS(conn));
2498
2499 ackmsg = &conn->ibc_connvars->cv_msg;
2500 memset(ackmsg, 0, sizeof(*ackmsg));
2501
2502 kiblnd_init_msg(ackmsg, IBLND_MSG_CONNACK,
2503 sizeof(ackmsg->ibm_u.connparams));
2504 ackmsg->ibm_u.connparams.ibcp_queue_depth = conn->ibc_queue_depth;
2505 ackmsg->ibm_u.connparams.ibcp_max_frags = conn->ibc_max_frags << IBLND_FRAG_SHIFT;
2506 ackmsg->ibm_u.connparams.ibcp_max_msg_size = IBLND_MSG_SIZE;
2507
2508 kiblnd_pack_msg(ni, ackmsg, version, 0, nid, reqmsg->ibm_srcstamp);
2509
2510 memset(&cp, 0, sizeof(cp));
2511 cp.private_data = ackmsg;
2512 cp.private_data_len = ackmsg->ibm_nob;
2513 cp.responder_resources = 0; /* No atomic ops or RDMA reads */
2514 cp.initiator_depth = 0;
2515 cp.flow_control = 1;
2516 cp.retry_count = *kiblnd_tunables.kib_retry_count;
2517 cp.rnr_retry_count = *kiblnd_tunables.kib_rnr_retry_count;
2518
2519 CDEBUG(D_NET, "Accept %s\n", libcfs_nid2str(nid));
2520
2521 rc = rdma_accept(cmid, &cp);
2522 if (rc) {
2523 CERROR("Can't accept %s: %d\n", libcfs_nid2str(nid), rc);
2524 rej.ibr_version = version;
2525 rej.ibr_why = IBLND_REJECT_FATAL;
2526
2527 kiblnd_reject(cmid, &rej);
2528 kiblnd_connreq_done(conn, rc);
2529 kiblnd_conn_decref(conn);
2530 }
2531
2532 lnet_ni_decref(ni);
2533 return 0;
2534
2535 failed:
2536 if (ni) {
2537 rej.ibr_cp.ibcp_queue_depth = kiblnd_msg_queue_size(version, ni);
2538 rej.ibr_cp.ibcp_max_frags = kiblnd_rdma_frags(version, ni);
2539 lnet_ni_decref(ni);
2540 }
2541
2542 rej.ibr_version = version;
2543 kiblnd_reject(cmid, &rej);
2544
2545 return -ECONNREFUSED;
2546 }
2547
2548 static void
2549 kiblnd_check_reconnect(struct kib_conn *conn, int version,
2550 __u64 incarnation, int why, struct kib_connparams *cp)
2551 {
2552 rwlock_t *glock = &kiblnd_data.kib_global_lock;
2553 struct kib_peer *peer = conn->ibc_peer;
2554 char *reason;
2555 int msg_size = IBLND_MSG_SIZE;
2556 int frag_num = -1;
2557 int queue_dep = -1;
2558 bool reconnect;
2559 unsigned long flags;
2560
2561 LASSERT(conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT);
2562 LASSERT(peer->ibp_connecting > 0); /* 'conn' at least */
2563 LASSERT(!peer->ibp_reconnecting);
2564
2565 if (cp) {
2566 msg_size = cp->ibcp_max_msg_size;
2567 frag_num = cp->ibcp_max_frags << IBLND_FRAG_SHIFT;
2568 queue_dep = cp->ibcp_queue_depth;
2569 }
2570
2571 write_lock_irqsave(glock, flags);
2572 /**
2573 * retry connection if it's still needed and no other connection
2574 * attempts (active or passive) are in progress
2575 * NB: reconnect is still needed even when ibp_tx_queue is
2576 * empty if ibp_version != version because reconnect may be
2577 * initiated by kiblnd_query()
2578 */
2579 reconnect = (!list_empty(&peer->ibp_tx_queue) ||
2580 peer->ibp_version != version) &&
2581 peer->ibp_connecting == 1 &&
2582 !peer->ibp_accepting;
2583 if (!reconnect) {
2584 reason = "no need";
2585 goto out;
2586 }
2587
2588 switch (why) {
2589 default:
2590 reason = "Unknown";
2591 break;
2592
2593 case IBLND_REJECT_RDMA_FRAGS: {
2594 struct lnet_ioctl_config_lnd_tunables *tunables;
2595
2596 if (!cp) {
2597 reason = "can't negotiate max frags";
2598 goto out;
2599 }
2600 tunables = peer->ibp_ni->ni_lnd_tunables;
2601 if (!tunables->lt_tun_u.lt_o2ib.lnd_map_on_demand) {
2602 reason = "map_on_demand must be enabled";
2603 goto out;
2604 }
2605 if (conn->ibc_max_frags <= frag_num) {
2606 reason = "unsupported max frags";
2607 goto out;
2608 }
2609
2610 peer->ibp_max_frags = frag_num;
2611 reason = "rdma fragments";
2612 break;
2613 }
2614 case IBLND_REJECT_MSG_QUEUE_SIZE:
2615 if (!cp) {
2616 reason = "can't negotiate queue depth";
2617 goto out;
2618 }
2619 if (conn->ibc_queue_depth <= queue_dep) {
2620 reason = "unsupported queue depth";
2621 goto out;
2622 }
2623
2624 peer->ibp_queue_depth = queue_dep;
2625 reason = "queue depth";
2626 break;
2627
2628 case IBLND_REJECT_CONN_STALE:
2629 reason = "stale";
2630 break;
2631
2632 case IBLND_REJECT_CONN_RACE:
2633 reason = "conn race";
2634 break;
2635
2636 case IBLND_REJECT_CONN_UNCOMPAT:
2637 reason = "version negotiation";
2638 break;
2639 }
2640
2641 conn->ibc_reconnect = 1;
2642 peer->ibp_reconnecting = 1;
2643 peer->ibp_version = version;
2644 if (incarnation)
2645 peer->ibp_incarnation = incarnation;
2646 out:
2647 write_unlock_irqrestore(glock, flags);
2648
2649 CNETERR("%s: %s (%s), %x, %x, msg_size: %d, queue_depth: %d/%d, max_frags: %d/%d\n",
2650 libcfs_nid2str(peer->ibp_nid),
2651 reconnect ? "reconnect" : "don't reconnect",
2652 reason, IBLND_MSG_VERSION, version, msg_size,
2653 conn->ibc_queue_depth, queue_dep,
2654 conn->ibc_max_frags, frag_num);
2655 /**
2656 * if conn::ibc_reconnect is TRUE, connd will reconnect to the peer
2657 * while destroying the zombie
2658 */
2659 }
2660
2661 static void
2662 kiblnd_rejected(struct kib_conn *conn, int reason, void *priv, int priv_nob)
2663 {
2664 struct kib_peer *peer = conn->ibc_peer;
2665
2666 LASSERT(!in_interrupt());
2667 LASSERT(conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT);
2668
2669 switch (reason) {
2670 case IB_CM_REJ_STALE_CONN:
2671 kiblnd_check_reconnect(conn, IBLND_MSG_VERSION, 0,
2672 IBLND_REJECT_CONN_STALE, NULL);
2673 break;
2674
2675 case IB_CM_REJ_INVALID_SERVICE_ID:
2676 CNETERR("%s rejected: no listener at %d\n",
2677 libcfs_nid2str(peer->ibp_nid),
2678 *kiblnd_tunables.kib_service);
2679 break;
2680
2681 case IB_CM_REJ_CONSUMER_DEFINED:
2682 if (priv_nob >= offsetof(struct kib_rej, ibr_padding)) {
2683 struct kib_rej *rej = priv;
2684 struct kib_connparams *cp = NULL;
2685 int flip = 0;
2686 __u64 incarnation = -1;
2687
2688 /* NB. default incarnation is -1 because:
2689 * a) V1 will ignore dst incarnation in connreq.
2690 * b) V2 will provide incarnation while rejecting me,
2691 * -1 will be overwrote.
2692 *
2693 * if I try to connect to a V1 peer with V2 protocol,
2694 * it rejected me then upgrade to V2, I have no idea
2695 * about the upgrading and try to reconnect with V1,
2696 * in this case upgraded V2 can find out I'm trying to
2697 * talk to the old guy and reject me(incarnation is -1).
2698 */
2699
2700 if (rej->ibr_magic == __swab32(IBLND_MSG_MAGIC) ||
2701 rej->ibr_magic == __swab32(LNET_PROTO_MAGIC)) {
2702 __swab32s(&rej->ibr_magic);
2703 __swab16s(&rej->ibr_version);
2704 flip = 1;
2705 }
2706
2707 if (priv_nob >= sizeof(struct kib_rej) &&
2708 rej->ibr_version > IBLND_MSG_VERSION_1) {
2709 /*
2710 * priv_nob is always 148 in current version
2711 * of OFED, so we still need to check version.
2712 * (define of IB_CM_REJ_PRIVATE_DATA_SIZE)
2713 */
2714 cp = &rej->ibr_cp;
2715
2716 if (flip) {
2717 __swab64s(&rej->ibr_incarnation);
2718 __swab16s(&cp->ibcp_queue_depth);
2719 __swab16s(&cp->ibcp_max_frags);
2720 __swab32s(&cp->ibcp_max_msg_size);
2721 }
2722
2723 incarnation = rej->ibr_incarnation;
2724 }
2725
2726 if (rej->ibr_magic != IBLND_MSG_MAGIC &&
2727 rej->ibr_magic != LNET_PROTO_MAGIC) {
2728 CERROR("%s rejected: consumer defined fatal error\n",
2729 libcfs_nid2str(peer->ibp_nid));
2730 break;
2731 }
2732
2733 if (rej->ibr_version != IBLND_MSG_VERSION &&
2734 rej->ibr_version != IBLND_MSG_VERSION_1) {
2735 CERROR("%s rejected: o2iblnd version %x error\n",
2736 libcfs_nid2str(peer->ibp_nid),
2737 rej->ibr_version);
2738 break;
2739 }
2740
2741 if (rej->ibr_why == IBLND_REJECT_FATAL &&
2742 rej->ibr_version == IBLND_MSG_VERSION_1) {
2743 CDEBUG(D_NET, "rejected by old version peer %s: %x\n",
2744 libcfs_nid2str(peer->ibp_nid), rej->ibr_version);
2745
2746 if (conn->ibc_version != IBLND_MSG_VERSION_1)
2747 rej->ibr_why = IBLND_REJECT_CONN_UNCOMPAT;
2748 }
2749
2750 switch (rej->ibr_why) {
2751 case IBLND_REJECT_CONN_RACE:
2752 case IBLND_REJECT_CONN_STALE:
2753 case IBLND_REJECT_CONN_UNCOMPAT:
2754 case IBLND_REJECT_MSG_QUEUE_SIZE:
2755 case IBLND_REJECT_RDMA_FRAGS:
2756 kiblnd_check_reconnect(conn, rej->ibr_version,
2757 incarnation,
2758 rej->ibr_why, cp);
2759 break;
2760
2761 case IBLND_REJECT_NO_RESOURCES:
2762 CERROR("%s rejected: o2iblnd no resources\n",
2763 libcfs_nid2str(peer->ibp_nid));
2764 break;
2765
2766 case IBLND_REJECT_FATAL:
2767 CERROR("%s rejected: o2iblnd fatal error\n",
2768 libcfs_nid2str(peer->ibp_nid));
2769 break;
2770
2771 default:
2772 CERROR("%s rejected: o2iblnd reason %d\n",
2773 libcfs_nid2str(peer->ibp_nid),
2774 rej->ibr_why);
2775 break;
2776 }
2777 break;
2778 }
2779 /* fall through */
2780 default:
2781 CNETERR("%s rejected: reason %d, size %d\n",
2782 libcfs_nid2str(peer->ibp_nid), reason, priv_nob);
2783 break;
2784 }
2785
2786 kiblnd_connreq_done(conn, -ECONNREFUSED);
2787 }
2788
2789 static void
2790 kiblnd_check_connreply(struct kib_conn *conn, void *priv, int priv_nob)
2791 {
2792 struct kib_peer *peer = conn->ibc_peer;
2793 struct lnet_ni *ni = peer->ibp_ni;
2794 struct kib_net *net = ni->ni_data;
2795 struct kib_msg *msg = priv;
2796 int ver = conn->ibc_version;
2797 int rc = kiblnd_unpack_msg(msg, priv_nob);
2798 unsigned long flags;
2799
2800 LASSERT(net);
2801
2802 if (rc) {
2803 CERROR("Can't unpack connack from %s: %d\n",
2804 libcfs_nid2str(peer->ibp_nid), rc);
2805 goto failed;
2806 }
2807
2808 if (msg->ibm_type != IBLND_MSG_CONNACK) {
2809 CERROR("Unexpected message %d from %s\n",
2810 msg->ibm_type, libcfs_nid2str(peer->ibp_nid));
2811 rc = -EPROTO;
2812 goto failed;
2813 }
2814
2815 if (ver != msg->ibm_version) {
2816 CERROR("%s replied version %x is different with requested version %x\n",
2817 libcfs_nid2str(peer->ibp_nid), msg->ibm_version, ver);
2818 rc = -EPROTO;
2819 goto failed;
2820 }
2821
2822 if (msg->ibm_u.connparams.ibcp_queue_depth >
2823 conn->ibc_queue_depth) {
2824 CERROR("%s has incompatible queue depth %d (<=%d wanted)\n",
2825 libcfs_nid2str(peer->ibp_nid),
2826 msg->ibm_u.connparams.ibcp_queue_depth,
2827 conn->ibc_queue_depth);
2828 rc = -EPROTO;
2829 goto failed;
2830 }
2831
2832 if ((msg->ibm_u.connparams.ibcp_max_frags >> IBLND_FRAG_SHIFT) >
2833 conn->ibc_max_frags) {
2834 CERROR("%s has incompatible max_frags %d (<=%d wanted)\n",
2835 libcfs_nid2str(peer->ibp_nid),
2836 msg->ibm_u.connparams.ibcp_max_frags >> IBLND_FRAG_SHIFT,
2837 conn->ibc_max_frags);
2838 rc = -EPROTO;
2839 goto failed;
2840 }
2841
2842 if (msg->ibm_u.connparams.ibcp_max_msg_size > IBLND_MSG_SIZE) {
2843 CERROR("%s max message size %d too big (%d max)\n",
2844 libcfs_nid2str(peer->ibp_nid),
2845 msg->ibm_u.connparams.ibcp_max_msg_size,
2846 IBLND_MSG_SIZE);
2847 rc = -EPROTO;
2848 goto failed;
2849 }
2850
2851 read_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
2852 if (msg->ibm_dstnid == ni->ni_nid &&
2853 msg->ibm_dststamp == net->ibn_incarnation)
2854 rc = 0;
2855 else
2856 rc = -ESTALE;
2857 read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
2858
2859 if (rc) {
2860 CERROR("Bad connection reply from %s, rc = %d, version: %x max_frags: %d\n",
2861 libcfs_nid2str(peer->ibp_nid), rc,
2862 msg->ibm_version, msg->ibm_u.connparams.ibcp_max_frags);
2863 goto failed;
2864 }
2865
2866 conn->ibc_incarnation = msg->ibm_srcstamp;
2867 conn->ibc_credits = msg->ibm_u.connparams.ibcp_queue_depth;
2868 conn->ibc_reserved_credits = msg->ibm_u.connparams.ibcp_queue_depth;
2869 conn->ibc_queue_depth = msg->ibm_u.connparams.ibcp_queue_depth;
2870 conn->ibc_max_frags = msg->ibm_u.connparams.ibcp_max_frags >> IBLND_FRAG_SHIFT;
2871 LASSERT(conn->ibc_credits + conn->ibc_reserved_credits +
2872 IBLND_OOB_MSGS(ver) <= IBLND_RX_MSGS(conn));
2873
2874 kiblnd_connreq_done(conn, 0);
2875 return;
2876
2877 failed:
2878 /*
2879 * NB My QP has already established itself, so I handle anything going
2880 * wrong here by setting ibc_comms_error.
2881 * kiblnd_connreq_done(0) moves the conn state to ESTABLISHED, but then
2882 * immediately tears it down.
2883 */
2884 LASSERT(rc);
2885 conn->ibc_comms_error = rc;
2886 kiblnd_connreq_done(conn, 0);
2887 }
2888
2889 static int
2890 kiblnd_active_connect(struct rdma_cm_id *cmid)
2891 {
2892 struct kib_peer *peer = (struct kib_peer *)cmid->context;
2893 struct kib_conn *conn;
2894 struct kib_msg *msg;
2895 struct rdma_conn_param cp;
2896 int version;
2897 __u64 incarnation;
2898 unsigned long flags;
2899 int rc;
2900
2901 read_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
2902
2903 incarnation = peer->ibp_incarnation;
2904 version = !peer->ibp_version ? IBLND_MSG_VERSION :
2905 peer->ibp_version;
2906
2907 read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
2908
2909 conn = kiblnd_create_conn(peer, cmid, IBLND_CONN_ACTIVE_CONNECT,
2910 version);
2911 if (!conn) {
2912 kiblnd_peer_connect_failed(peer, 1, -ENOMEM);
2913 kiblnd_peer_decref(peer); /* lose cmid's ref */
2914 return -ENOMEM;
2915 }
2916
2917 /*
2918 * conn "owns" cmid now, so I return success from here on to ensure the
2919 * CM callback doesn't destroy cmid. conn also takes over cmid's ref
2920 * on peer
2921 */
2922 msg = &conn->ibc_connvars->cv_msg;
2923
2924 memset(msg, 0, sizeof(*msg));
2925 kiblnd_init_msg(msg, IBLND_MSG_CONNREQ, sizeof(msg->ibm_u.connparams));
2926 msg->ibm_u.connparams.ibcp_queue_depth = conn->ibc_queue_depth;
2927 msg->ibm_u.connparams.ibcp_max_frags = conn->ibc_max_frags << IBLND_FRAG_SHIFT;
2928 msg->ibm_u.connparams.ibcp_max_msg_size = IBLND_MSG_SIZE;
2929
2930 kiblnd_pack_msg(peer->ibp_ni, msg, version,
2931 0, peer->ibp_nid, incarnation);
2932
2933 memset(&cp, 0, sizeof(cp));
2934 cp.private_data = msg;
2935 cp.private_data_len = msg->ibm_nob;
2936 cp.responder_resources = 0; /* No atomic ops or RDMA reads */
2937 cp.initiator_depth = 0;
2938 cp.flow_control = 1;
2939 cp.retry_count = *kiblnd_tunables.kib_retry_count;
2940 cp.rnr_retry_count = *kiblnd_tunables.kib_rnr_retry_count;
2941
2942 LASSERT(cmid->context == (void *)conn);
2943 LASSERT(conn->ibc_cmid == cmid);
2944
2945 rc = rdma_connect(cmid, &cp);
2946 if (rc) {
2947 CERROR("Can't connect to %s: %d\n",
2948 libcfs_nid2str(peer->ibp_nid), rc);
2949 kiblnd_connreq_done(conn, rc);
2950 kiblnd_conn_decref(conn);
2951 }
2952
2953 return 0;
2954 }
2955
2956 int
2957 kiblnd_cm_callback(struct rdma_cm_id *cmid, struct rdma_cm_event *event)
2958 {
2959 struct kib_peer *peer;
2960 struct kib_conn *conn;
2961 int rc;
2962
2963 switch (event->event) {
2964 default:
2965 CERROR("Unexpected event: %d, status: %d\n",
2966 event->event, event->status);
2967 LBUG();
2968
2969 case RDMA_CM_EVENT_CONNECT_REQUEST:
2970 /* destroy cmid on failure */
2971 rc = kiblnd_passive_connect(cmid,
2972 (void *)KIBLND_CONN_PARAM(event),
2973 KIBLND_CONN_PARAM_LEN(event));
2974 CDEBUG(D_NET, "connreq: %d\n", rc);
2975 return rc;
2976
2977 case RDMA_CM_EVENT_ADDR_ERROR:
2978 peer = (struct kib_peer *)cmid->context;
2979 CNETERR("%s: ADDR ERROR %d\n",
2980 libcfs_nid2str(peer->ibp_nid), event->status);
2981 kiblnd_peer_connect_failed(peer, 1, -EHOSTUNREACH);
2982 kiblnd_peer_decref(peer);
2983 return -EHOSTUNREACH; /* rc destroys cmid */
2984
2985 case RDMA_CM_EVENT_ADDR_RESOLVED:
2986 peer = (struct kib_peer *)cmid->context;
2987
2988 CDEBUG(D_NET, "%s Addr resolved: %d\n",
2989 libcfs_nid2str(peer->ibp_nid), event->status);
2990
2991 if (event->status) {
2992 CNETERR("Can't resolve address for %s: %d\n",
2993 libcfs_nid2str(peer->ibp_nid), event->status);
2994 rc = event->status;
2995 } else {
2996 rc = rdma_resolve_route(
2997 cmid, *kiblnd_tunables.kib_timeout * 1000);
2998 if (!rc)
2999 return 0;
3000 /* Can't initiate route resolution */
3001 CERROR("Can't resolve route for %s: %d\n",
3002 libcfs_nid2str(peer->ibp_nid), rc);
3003 }
3004 kiblnd_peer_connect_failed(peer, 1, rc);
3005 kiblnd_peer_decref(peer);
3006 return rc; /* rc destroys cmid */
3007
3008 case RDMA_CM_EVENT_ROUTE_ERROR:
3009 peer = (struct kib_peer *)cmid->context;
3010 CNETERR("%s: ROUTE ERROR %d\n",
3011 libcfs_nid2str(peer->ibp_nid), event->status);
3012 kiblnd_peer_connect_failed(peer, 1, -EHOSTUNREACH);
3013 kiblnd_peer_decref(peer);
3014 return -EHOSTUNREACH; /* rc destroys cmid */
3015
3016 case RDMA_CM_EVENT_ROUTE_RESOLVED:
3017 peer = (struct kib_peer *)cmid->context;
3018 CDEBUG(D_NET, "%s Route resolved: %d\n",
3019 libcfs_nid2str(peer->ibp_nid), event->status);
3020
3021 if (!event->status)
3022 return kiblnd_active_connect(cmid);
3023
3024 CNETERR("Can't resolve route for %s: %d\n",
3025 libcfs_nid2str(peer->ibp_nid), event->status);
3026 kiblnd_peer_connect_failed(peer, 1, event->status);
3027 kiblnd_peer_decref(peer);
3028 return event->status; /* rc destroys cmid */
3029
3030 case RDMA_CM_EVENT_UNREACHABLE:
3031 conn = (struct kib_conn *)cmid->context;
3032 LASSERT(conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT ||
3033 conn->ibc_state == IBLND_CONN_PASSIVE_WAIT);
3034 CNETERR("%s: UNREACHABLE %d\n",
3035 libcfs_nid2str(conn->ibc_peer->ibp_nid), event->status);
3036 kiblnd_connreq_done(conn, -ENETDOWN);
3037 kiblnd_conn_decref(conn);
3038 return 0;
3039
3040 case RDMA_CM_EVENT_CONNECT_ERROR:
3041 conn = (struct kib_conn *)cmid->context;
3042 LASSERT(conn->ibc_state == IBLND_CONN_ACTIVE_CONNECT ||
3043 conn->ibc_state == IBLND_CONN_PASSIVE_WAIT);
3044 CNETERR("%s: CONNECT ERROR %d\n",
3045 libcfs_nid2str(conn->ibc_peer->ibp_nid), event->status);
3046 kiblnd_connreq_done(conn, -ENOTCONN);
3047 kiblnd_conn_decref(conn);
3048 return 0;
3049
3050 case RDMA_CM_EVENT_REJECTED:
3051 conn = (struct kib_conn *)cmid->context;
3052 switch (conn->ibc_state) {
3053 default:
3054 LBUG();
3055
3056 case IBLND_CONN_PASSIVE_WAIT:
3057 CERROR("%s: REJECTED %d\n",
3058 libcfs_nid2str(conn->ibc_peer->ibp_nid),
3059 event->status);
3060 kiblnd_connreq_done(conn, -ECONNRESET);
3061 break;
3062
3063 case IBLND_CONN_ACTIVE_CONNECT:
3064 kiblnd_rejected(conn, event->status,
3065 (void *)KIBLND_CONN_PARAM(event),
3066 KIBLND_CONN_PARAM_LEN(event));
3067 break;
3068 }
3069 kiblnd_conn_decref(conn);
3070 return 0;
3071
3072 case RDMA_CM_EVENT_ESTABLISHED:
3073 conn = (struct kib_conn *)cmid->context;
3074 switch (conn->ibc_state) {
3075 default:
3076 LBUG();
3077
3078 case IBLND_CONN_PASSIVE_WAIT:
3079 CDEBUG(D_NET, "ESTABLISHED (passive): %s\n",
3080 libcfs_nid2str(conn->ibc_peer->ibp_nid));
3081 kiblnd_connreq_done(conn, 0);
3082 break;
3083
3084 case IBLND_CONN_ACTIVE_CONNECT:
3085 CDEBUG(D_NET, "ESTABLISHED(active): %s\n",
3086 libcfs_nid2str(conn->ibc_peer->ibp_nid));
3087 kiblnd_check_connreply(conn,
3088 (void *)KIBLND_CONN_PARAM(event),
3089 KIBLND_CONN_PARAM_LEN(event));
3090 break;
3091 }
3092 /* net keeps its ref on conn! */
3093 return 0;
3094
3095 case RDMA_CM_EVENT_TIMEWAIT_EXIT:
3096 CDEBUG(D_NET, "Ignore TIMEWAIT_EXIT event\n");
3097 return 0;
3098 case RDMA_CM_EVENT_DISCONNECTED:
3099 conn = (struct kib_conn *)cmid->context;
3100 if (conn->ibc_state < IBLND_CONN_ESTABLISHED) {
3101 CERROR("%s DISCONNECTED\n",
3102 libcfs_nid2str(conn->ibc_peer->ibp_nid));
3103 kiblnd_connreq_done(conn, -ECONNRESET);
3104 } else {
3105 kiblnd_close_conn(conn, 0);
3106 }
3107 kiblnd_conn_decref(conn);
3108 cmid->context = NULL;
3109 return 0;
3110
3111 case RDMA_CM_EVENT_DEVICE_REMOVAL:
3112 LCONSOLE_ERROR_MSG(0x131,
3113 "Received notification of device removal\n"
3114 "Please shutdown LNET to allow this to proceed\n");
3115 /*
3116 * Can't remove network from underneath LNET for now, so I have
3117 * to ignore this
3118 */
3119 return 0;
3120
3121 case RDMA_CM_EVENT_ADDR_CHANGE:
3122 LCONSOLE_INFO("Physical link changed (eg hca/port)\n");
3123 return 0;
3124 }
3125 }
3126
3127 static int
3128 kiblnd_check_txs_locked(struct kib_conn *conn, struct list_head *txs)
3129 {
3130 struct kib_tx *tx;
3131 struct list_head *ttmp;
3132
3133 list_for_each(ttmp, txs) {
3134 tx = list_entry(ttmp, struct kib_tx, tx_list);
3135
3136 if (txs != &conn->ibc_active_txs) {
3137 LASSERT(tx->tx_queued);
3138 } else {
3139 LASSERT(!tx->tx_queued);
3140 LASSERT(tx->tx_waiting || tx->tx_sending);
3141 }
3142
3143 if (cfs_time_aftereq(jiffies, tx->tx_deadline)) {
3144 CERROR("Timed out tx: %s, %lu seconds\n",
3145 kiblnd_queue2str(conn, txs),
3146 cfs_duration_sec(jiffies - tx->tx_deadline));
3147 return 1;
3148 }
3149 }
3150
3151 return 0;
3152 }
3153
3154 static int
3155 kiblnd_conn_timed_out_locked(struct kib_conn *conn)
3156 {
3157 return kiblnd_check_txs_locked(conn, &conn->ibc_tx_queue) ||
3158 kiblnd_check_txs_locked(conn, &conn->ibc_tx_noops) ||
3159 kiblnd_check_txs_locked(conn, &conn->ibc_tx_queue_rsrvd) ||
3160 kiblnd_check_txs_locked(conn, &conn->ibc_tx_queue_nocred) ||
3161 kiblnd_check_txs_locked(conn, &conn->ibc_active_txs);
3162 }
3163
3164 static void
3165 kiblnd_check_conns(int idx)
3166 {
3167 LIST_HEAD(closes);
3168 LIST_HEAD(checksends);
3169 struct list_head *peers = &kiblnd_data.kib_peers[idx];
3170 struct list_head *ptmp;
3171 struct kib_peer *peer;
3172 struct kib_conn *conn;
3173 struct kib_conn *temp;
3174 struct kib_conn *tmp;
3175 struct list_head *ctmp;
3176 unsigned long flags;
3177
3178 /*
3179 * NB. We expect to have a look at all the peers and not find any
3180 * RDMAs to time out, so we just use a shared lock while we
3181 * take a look...
3182 */
3183 read_lock_irqsave(&kiblnd_data.kib_global_lock, flags);
3184
3185 list_for_each(ptmp, peers) {
3186 peer = list_entry(ptmp, struct kib_peer, ibp_list);
3187
3188 list_for_each(ctmp, &peer->ibp_conns) {
3189 int timedout;
3190 int sendnoop;
3191
3192 conn = list_entry(ctmp, struct kib_conn, ibc_list);
3193
3194 LASSERT(conn->ibc_state == IBLND_CONN_ESTABLISHED);
3195
3196 spin_lock(&conn->ibc_lock);
3197
3198 sendnoop = kiblnd_need_noop(conn);
3199 timedout = kiblnd_conn_timed_out_locked(conn);
3200 if (!sendnoop && !timedout) {
3201 spin_unlock(&conn->ibc_lock);
3202 continue;
3203 }
3204
3205 if (timedout) {
3206 CERROR("Timed out RDMA with %s (%lu): c: %u, oc: %u, rc: %u\n",
3207 libcfs_nid2str(peer->ibp_nid),
3208 cfs_duration_sec(cfs_time_current() -
3209 peer->ibp_last_alive),
3210 conn->ibc_credits,
3211 conn->ibc_outstanding_credits,
3212 conn->ibc_reserved_credits);
3213 list_add(&conn->ibc_connd_list, &closes);
3214 } else {
3215 list_add(&conn->ibc_connd_list, &checksends);
3216 }
3217 /* +ref for 'closes' or 'checksends' */
3218 kiblnd_conn_addref(conn);
3219
3220 spin_unlock(&conn->ibc_lock);
3221 }
3222 }
3223
3224 read_unlock_irqrestore(&kiblnd_data.kib_global_lock, flags);
3225
3226 /*
3227 * Handle timeout by closing the whole
3228 * connection. We can only be sure RDMA activity
3229 * has ceased once the QP has been modified.
3230 */
3231 list_for_each_entry_safe(conn, tmp, &closes, ibc_connd_list) {
3232 list_del(&conn->ibc_connd_list);
3233 kiblnd_close_conn(conn, -ETIMEDOUT);
3234 kiblnd_conn_decref(conn);
3235 }
3236
3237 /*
3238 * In case we have enough credits to return via a
3239 * NOOP, but there were no non-blocking tx descs
3240 * free to do it last time...
3241 */
3242 list_for_each_entry_safe(conn, temp, &checksends, ibc_connd_list) {
3243 list_del(&conn->ibc_connd_list);
3244
3245 spin_lock(&conn->ibc_lock);
3246 kiblnd_check_sends_locked(conn);
3247 spin_unlock(&conn->ibc_lock);
3248
3249 kiblnd_conn_decref(conn);
3250 }
3251 }
3252
3253 static void
3254 kiblnd_disconnect_conn(struct kib_conn *conn)
3255 {
3256 LASSERT(!in_interrupt());
3257 LASSERT(current == kiblnd_data.kib_connd);
3258 LASSERT(conn->ibc_state == IBLND_CONN_CLOSING);
3259
3260 rdma_disconnect(conn->ibc_cmid);
3261 kiblnd_finalise_conn(conn);
3262
3263 kiblnd_peer_notify(conn->ibc_peer);
3264 }
3265
3266 /**
3267 * High-water for reconnection to the same peer, reconnection attempt should
3268 * be delayed after trying more than KIB_RECONN_HIGH_RACE.
3269 */
3270 #define KIB_RECONN_HIGH_RACE 10
3271 /**
3272 * Allow connd to take a break and handle other things after consecutive
3273 * reconnection attempts.
3274 */
3275 #define KIB_RECONN_BREAK 100
3276
3277 int
3278 kiblnd_connd(void *arg)
3279 {
3280 spinlock_t *lock = &kiblnd_data.kib_connd_lock;
3281 wait_queue_entry_t wait;
3282 unsigned long flags;
3283 struct kib_conn *conn;
3284 int timeout;
3285 int i;
3286 int dropped_lock;
3287 int peer_index = 0;
3288 unsigned long deadline = jiffies;
3289
3290 cfs_block_allsigs();
3291
3292 init_waitqueue_entry(&wait, current);
3293 kiblnd_data.kib_connd = current;
3294
3295 spin_lock_irqsave(lock, flags);
3296
3297 while (!kiblnd_data.kib_shutdown) {
3298 int reconn = 0;
3299
3300 dropped_lock = 0;
3301
3302 if (!list_empty(&kiblnd_data.kib_connd_zombies)) {
3303 struct kib_peer *peer = NULL;
3304
3305 conn = list_entry(kiblnd_data.kib_connd_zombies.next,
3306 struct kib_conn, ibc_list);
3307 list_del(&conn->ibc_list);
3308 if (conn->ibc_reconnect) {
3309 peer = conn->ibc_peer;
3310 kiblnd_peer_addref(peer);
3311 }
3312
3313 spin_unlock_irqrestore(lock, flags);
3314 dropped_lock = 1;
3315
3316 kiblnd_destroy_conn(conn, !peer);
3317
3318 spin_lock_irqsave(lock, flags);
3319 if (!peer)
3320 continue;
3321
3322 conn->ibc_peer = peer;
3323 if (peer->ibp_reconnected < KIB_RECONN_HIGH_RACE)
3324 list_add_tail(&conn->ibc_list,
3325 &kiblnd_data.kib_reconn_list);
3326 else
3327 list_add_tail(&conn->ibc_list,
3328 &kiblnd_data.kib_reconn_wait);
3329 }
3330
3331 if (!list_empty(&kiblnd_data.kib_connd_conns)) {
3332 conn = list_entry(kiblnd_data.kib_connd_conns.next,
3333 struct kib_conn, ibc_list);
3334 list_del(&conn->ibc_list);
3335
3336 spin_unlock_irqrestore(lock, flags);
3337 dropped_lock = 1;
3338
3339 kiblnd_disconnect_conn(conn);
3340 kiblnd_conn_decref(conn);
3341
3342 spin_lock_irqsave(lock, flags);
3343 }
3344
3345 while (reconn < KIB_RECONN_BREAK) {
3346 if (kiblnd_data.kib_reconn_sec !=
3347 ktime_get_real_seconds()) {
3348 kiblnd_data.kib_reconn_sec = ktime_get_real_seconds();
3349 list_splice_init(&kiblnd_data.kib_reconn_wait,
3350 &kiblnd_data.kib_reconn_list);
3351 }
3352
3353 if (list_empty(&kiblnd_data.kib_reconn_list))
3354 break;
3355
3356 conn = list_entry(kiblnd_data.kib_reconn_list.next,
3357 struct kib_conn, ibc_list);
3358 list_del(&conn->ibc_list);
3359
3360 spin_unlock_irqrestore(lock, flags);
3361 dropped_lock = 1;
3362
3363 reconn += kiblnd_reconnect_peer(conn->ibc_peer);
3364 kiblnd_peer_decref(conn->ibc_peer);
3365 LIBCFS_FREE(conn, sizeof(*conn));
3366
3367 spin_lock_irqsave(lock, flags);
3368 }
3369
3370 /* careful with the jiffy wrap... */
3371 timeout = (int)(deadline - jiffies);
3372 if (timeout <= 0) {
3373 const int n = 4;
3374 const int p = 1;
3375 int chunk = kiblnd_data.kib_peer_hash_size;
3376
3377 spin_unlock_irqrestore(lock, flags);
3378 dropped_lock = 1;
3379
3380 /*
3381 * Time to check for RDMA timeouts on a few more
3382 * peers: I do checks every 'p' seconds on a
3383 * proportion of the peer table and I need to check
3384 * every connection 'n' times within a timeout
3385 * interval, to ensure I detect a timeout on any
3386 * connection within (n+1)/n times the timeout
3387 * interval.
3388 */
3389 if (*kiblnd_tunables.kib_timeout > n * p)
3390 chunk = (chunk * n * p) /
3391 *kiblnd_tunables.kib_timeout;
3392 if (!chunk)
3393 chunk = 1;
3394
3395 for (i = 0; i < chunk; i++) {
3396 kiblnd_check_conns(peer_index);
3397 peer_index = (peer_index + 1) %
3398 kiblnd_data.kib_peer_hash_size;
3399 }
3400
3401 deadline += msecs_to_jiffies(p * MSEC_PER_SEC);
3402 spin_lock_irqsave(lock, flags);
3403 }
3404
3405 if (dropped_lock)
3406 continue;
3407
3408 /* Nothing to do for 'timeout' */
3409 set_current_state(TASK_INTERRUPTIBLE);
3410 add_wait_queue(&kiblnd_data.kib_connd_waitq, &wait);
3411 spin_unlock_irqrestore(lock, flags);
3412
3413 schedule_timeout(timeout);
3414
3415 remove_wait_queue(&kiblnd_data.kib_connd_waitq, &wait);
3416 spin_lock_irqsave(lock, flags);
3417 }
3418
3419 spin_unlock_irqrestore(lock, flags);
3420
3421 kiblnd_thread_fini();
3422 return 0;
3423 }
3424
3425 void
3426 kiblnd_qp_event(struct ib_event *event, void *arg)
3427 {
3428 struct kib_conn *conn = arg;
3429
3430 switch (event->event) {
3431 case IB_EVENT_COMM_EST:
3432 CDEBUG(D_NET, "%s established\n",
3433 libcfs_nid2str(conn->ibc_peer->ibp_nid));
3434 /*
3435 * We received a packet but connection isn't established
3436 * probably handshake packet was lost, so free to
3437 * force make connection established
3438 */
3439 rdma_notify(conn->ibc_cmid, IB_EVENT_COMM_EST);
3440 return;
3441
3442 default:
3443 CERROR("%s: Async QP event type %d\n",
3444 libcfs_nid2str(conn->ibc_peer->ibp_nid), event->event);
3445 return;
3446 }
3447 }
3448
3449 static void
3450 kiblnd_complete(struct ib_wc *wc)
3451 {
3452 switch (kiblnd_wreqid2type(wc->wr_id)) {
3453 default:
3454 LBUG();
3455
3456 case IBLND_WID_MR:
3457 if (wc->status != IB_WC_SUCCESS &&
3458 wc->status != IB_WC_WR_FLUSH_ERR)
3459 CNETERR("FastReg failed: %d\n", wc->status);
3460 break;
3461
3462 case IBLND_WID_RDMA:
3463 /*
3464 * We only get RDMA completion notification if it fails. All
3465 * subsequent work items, including the final SEND will fail
3466 * too. However we can't print out any more info about the
3467 * failing RDMA because 'tx' might be back on the idle list or
3468 * even reused already if we didn't manage to post all our work
3469 * items
3470 */
3471 CNETERR("RDMA (tx: %p) failed: %d\n",
3472 kiblnd_wreqid2ptr(wc->wr_id), wc->status);
3473 return;
3474
3475 case IBLND_WID_TX:
3476 kiblnd_tx_complete(kiblnd_wreqid2ptr(wc->wr_id), wc->status);
3477 return;
3478
3479 case IBLND_WID_RX:
3480 kiblnd_rx_complete(kiblnd_wreqid2ptr(wc->wr_id), wc->status,
3481 wc->byte_len);
3482 return;
3483 }
3484 }
3485
3486 void
3487 kiblnd_cq_completion(struct ib_cq *cq, void *arg)
3488 {
3489 /*
3490 * NB I'm not allowed to schedule this conn once its refcount has
3491 * reached 0. Since fundamentally I'm racing with scheduler threads
3492 * consuming my CQ I could be called after all completions have
3493 * occurred. But in this case, !ibc_nrx && !ibc_nsends_posted
3494 * and this CQ is about to be destroyed so I NOOP.
3495 */
3496 struct kib_conn *conn = arg;
3497 struct kib_sched_info *sched = conn->ibc_sched;
3498 unsigned long flags;
3499
3500 LASSERT(cq == conn->ibc_cq);
3501
3502 spin_lock_irqsave(&sched->ibs_lock, flags);
3503
3504 conn->ibc_ready = 1;
3505
3506 if (!conn->ibc_scheduled &&
3507 (conn->ibc_nrx > 0 ||
3508 conn->ibc_nsends_posted > 0)) {
3509 kiblnd_conn_addref(conn); /* +1 ref for sched_conns */
3510 conn->ibc_scheduled = 1;
3511 list_add_tail(&conn->ibc_sched_list, &sched->ibs_conns);
3512
3513 if (waitqueue_active(&sched->ibs_waitq))
3514 wake_up(&sched->ibs_waitq);
3515 }
3516
3517 spin_unlock_irqrestore(&sched->ibs_lock, flags);
3518 }
3519
3520 void
3521 kiblnd_cq_event(struct ib_event *event, void *arg)
3522 {
3523 struct kib_conn *conn = arg;
3524
3525 CERROR("%s: async CQ event type %d\n",
3526 libcfs_nid2str(conn->ibc_peer->ibp_nid), event->event);
3527 }
3528
3529 int
3530 kiblnd_scheduler(void *arg)
3531 {
3532 long id = (long)arg;
3533 struct kib_sched_info *sched;
3534 struct kib_conn *conn;
3535 wait_queue_entry_t wait;
3536 unsigned long flags;
3537 struct ib_wc wc;
3538 int did_something;
3539 int busy_loops = 0;
3540 int rc;
3541
3542 cfs_block_allsigs();
3543
3544 init_waitqueue_entry(&wait, current);
3545
3546 sched = kiblnd_data.kib_scheds[KIB_THREAD_CPT(id)];
3547
3548 rc = cfs_cpt_bind(lnet_cpt_table(), sched->ibs_cpt);
3549 if (rc) {
3550 CWARN("Unable to bind on CPU partition %d, please verify whether all CPUs are healthy and reload modules if necessary, otherwise your system might under risk of low performance\n",
3551 sched->ibs_cpt);
3552 }
3553
3554 spin_lock_irqsave(&sched->ibs_lock, flags);
3555
3556 while (!kiblnd_data.kib_shutdown) {
3557 if (busy_loops++ >= IBLND_RESCHED) {
3558 spin_unlock_irqrestore(&sched->ibs_lock, flags);
3559
3560 cond_resched();
3561 busy_loops = 0;
3562
3563 spin_lock_irqsave(&sched->ibs_lock, flags);
3564 }
3565
3566 did_something = 0;
3567
3568 if (!list_empty(&sched->ibs_conns)) {
3569 conn = list_entry(sched->ibs_conns.next, struct kib_conn,
3570 ibc_sched_list);
3571 /* take over kib_sched_conns' ref on conn... */
3572 LASSERT(conn->ibc_scheduled);
3573 list_del(&conn->ibc_sched_list);
3574 conn->ibc_ready = 0;
3575
3576 spin_unlock_irqrestore(&sched->ibs_lock, flags);
3577
3578 wc.wr_id = IBLND_WID_INVAL;
3579
3580 rc = ib_poll_cq(conn->ibc_cq, 1, &wc);
3581 if (!rc) {
3582 rc = ib_req_notify_cq(conn->ibc_cq,
3583 IB_CQ_NEXT_COMP);
3584 if (rc < 0) {
3585 CWARN("%s: ib_req_notify_cq failed: %d, closing connection\n",
3586 libcfs_nid2str(conn->ibc_peer->ibp_nid), rc);
3587 kiblnd_close_conn(conn, -EIO);
3588 kiblnd_conn_decref(conn);
3589 spin_lock_irqsave(&sched->ibs_lock,
3590 flags);
3591 continue;
3592 }
3593
3594 rc = ib_poll_cq(conn->ibc_cq, 1, &wc);
3595 }
3596
3597 if (unlikely(rc > 0 && wc.wr_id == IBLND_WID_INVAL)) {
3598 LCONSOLE_ERROR("ib_poll_cq (rc: %d) returned invalid wr_id, opcode %d, status: %d, vendor_err: %d, conn: %s status: %d\nplease upgrade firmware and OFED or contact vendor.\n",
3599 rc, wc.opcode, wc.status,
3600 wc.vendor_err,
3601 libcfs_nid2str(conn->ibc_peer->ibp_nid),
3602 conn->ibc_state);
3603 rc = -EINVAL;
3604 }
3605
3606 if (rc < 0) {
3607 CWARN("%s: ib_poll_cq failed: %d, closing connection\n",
3608 libcfs_nid2str(conn->ibc_peer->ibp_nid),
3609 rc);
3610 kiblnd_close_conn(conn, -EIO);
3611 kiblnd_conn_decref(conn);
3612 spin_lock_irqsave(&sched->ibs_lock, flags);
3613 continue;
3614 }
3615
3616 spin_lock_irqsave(&sched->ibs_lock, flags);
3617
3618 if (rc || conn->ibc_ready) {
3619 /*
3620 * There may be another completion waiting; get
3621 * another scheduler to check while I handle
3622 * this one...
3623 */
3624 /* +1 ref for sched_conns */
3625 kiblnd_conn_addref(conn);
3626 list_add_tail(&conn->ibc_sched_list,
3627 &sched->ibs_conns);
3628 if (waitqueue_active(&sched->ibs_waitq))
3629 wake_up(&sched->ibs_waitq);
3630 } else {
3631 conn->ibc_scheduled = 0;
3632 }
3633
3634 if (rc) {
3635 spin_unlock_irqrestore(&sched->ibs_lock, flags);
3636 kiblnd_complete(&wc);
3637
3638 spin_lock_irqsave(&sched->ibs_lock, flags);
3639 }
3640
3641 kiblnd_conn_decref(conn); /* ...drop my ref from above */
3642 did_something = 1;
3643 }
3644
3645 if (did_something)
3646 continue;
3647
3648 set_current_state(TASK_INTERRUPTIBLE);
3649 add_wait_queue_exclusive(&sched->ibs_waitq, &wait);
3650 spin_unlock_irqrestore(&sched->ibs_lock, flags);
3651
3652 schedule();
3653 busy_loops = 0;
3654
3655 remove_wait_queue(&sched->ibs_waitq, &wait);
3656 spin_lock_irqsave(&sched->ibs_lock, flags);
3657 }
3658
3659 spin_unlock_irqrestore(&sched->ibs_lock, flags);
3660
3661 kiblnd_thread_fini();
3662 return 0;
3663 }
3664
3665 int
3666 kiblnd_failover_thread(void *arg)
3667 {
3668 rwlock_t *glock = &kiblnd_data.kib_global_lock;
3669 struct kib_dev *dev;
3670 wait_queue_entry_t wait;
3671 unsigned long flags;
3672 int rc;
3673
3674 LASSERT(*kiblnd_tunables.kib_dev_failover);
3675
3676 cfs_block_allsigs();
3677
3678 init_waitqueue_entry(&wait, current);
3679 write_lock_irqsave(glock, flags);
3680
3681 while (!kiblnd_data.kib_shutdown) {
3682 int do_failover = 0;
3683 int long_sleep;
3684
3685 list_for_each_entry(dev, &kiblnd_data.kib_failed_devs,
3686 ibd_fail_list) {
3687 if (time_before(cfs_time_current(),
3688 dev->ibd_next_failover))
3689 continue;
3690 do_failover = 1;
3691 break;
3692 }
3693
3694 if (do_failover) {
3695 list_del_init(&dev->ibd_fail_list);
3696 dev->ibd_failover = 1;
3697 write_unlock_irqrestore(glock, flags);
3698
3699 rc = kiblnd_dev_failover(dev);
3700
3701 write_lock_irqsave(glock, flags);
3702
3703 LASSERT(dev->ibd_failover);
3704 dev->ibd_failover = 0;
3705 if (rc >= 0) { /* Device is OK or failover succeed */
3706 dev->ibd_next_failover = cfs_time_shift(3);
3707 continue;
3708 }
3709
3710 /* failed to failover, retry later */
3711 dev->ibd_next_failover =
3712 cfs_time_shift(min(dev->ibd_failed_failover, 10));
3713 if (kiblnd_dev_can_failover(dev)) {
3714 list_add_tail(&dev->ibd_fail_list,
3715 &kiblnd_data.kib_failed_devs);
3716 }
3717
3718 continue;
3719 }
3720
3721 /* long sleep if no more pending failover */
3722 long_sleep = list_empty(&kiblnd_data.kib_failed_devs);
3723
3724 set_current_state(TASK_INTERRUPTIBLE);
3725 add_wait_queue(&kiblnd_data.kib_failover_waitq, &wait);
3726 write_unlock_irqrestore(glock, flags);
3727
3728 rc = schedule_timeout(long_sleep ? cfs_time_seconds(10) :
3729 cfs_time_seconds(1));
3730 remove_wait_queue(&kiblnd_data.kib_failover_waitq, &wait);
3731 write_lock_irqsave(glock, flags);
3732
3733 if (!long_sleep || rc)
3734 continue;
3735
3736 /*
3737 * have a long sleep, routine check all active devices,
3738 * we need checking like this because if there is not active
3739 * connection on the dev and no SEND from local, we may listen
3740 * on wrong HCA for ever while there is a bonding failover
3741 */
3742 list_for_each_entry(dev, &kiblnd_data.kib_devs, ibd_list) {
3743 if (kiblnd_dev_can_failover(dev)) {
3744 list_add_tail(&dev->ibd_fail_list,
3745 &kiblnd_data.kib_failed_devs);
3746 }
3747 }
3748 }
3749
3750 write_unlock_irqrestore(glock, flags);
3751
3752 kiblnd_thread_fini();
3753 return 0;
3754 }
Ubuntu Artful kernel mirror