Loading...
1/*
2 * Copyright (c) 2004 Topspin Communications. All rights reserved.
3 * Copyright (c) 2005 Intel Corporation. All rights reserved.
4 * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
5 * Copyright (c) 2005 Voltaire, Inc. All rights reserved.
6 *
7 * This software is available to you under a choice of one of two
8 * licenses. You may choose to be licensed under the terms of the GNU
9 * General Public License (GPL) Version 2, available from the file
10 * COPYING in the main directory of this source tree, or the
11 * OpenIB.org BSD license below:
12 *
13 * Redistribution and use in source and binary forms, with or
14 * without modification, are permitted provided that the following
15 * conditions are met:
16 *
17 * - Redistributions of source code must retain the above
18 * copyright notice, this list of conditions and the following
19 * disclaimer.
20 *
21 * - Redistributions in binary form must reproduce the above
22 * copyright notice, this list of conditions and the following
23 * disclaimer in the documentation and/or other materials
24 * provided with the distribution.
25 *
26 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
27 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
28 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
29 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
30 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
31 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
32 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
33 * SOFTWARE.
34 */
35
36#include <linux/if_vlan.h>
37#include <linux/errno.h>
38#include <linux/slab.h>
39#include <linux/workqueue.h>
40#include <linux/netdevice.h>
41#include <net/addrconf.h>
42
43#include <rdma/ib_cache.h>
44
45#include "core_priv.h"
46
47struct ib_pkey_cache {
48 int table_len;
49 u16 table[] __counted_by(table_len);
50};
51
52struct ib_update_work {
53 struct work_struct work;
54 struct ib_event event;
55 bool enforce_security;
56};
57
58union ib_gid zgid;
59EXPORT_SYMBOL(zgid);
60
61enum gid_attr_find_mask {
62 GID_ATTR_FIND_MASK_GID = 1UL << 0,
63 GID_ATTR_FIND_MASK_NETDEV = 1UL << 1,
64 GID_ATTR_FIND_MASK_DEFAULT = 1UL << 2,
65 GID_ATTR_FIND_MASK_GID_TYPE = 1UL << 3,
66};
67
68enum gid_table_entry_state {
69 GID_TABLE_ENTRY_INVALID = 1,
70 GID_TABLE_ENTRY_VALID = 2,
71 /*
72 * Indicates that entry is pending to be removed, there may
73 * be active users of this GID entry.
74 * When last user of the GID entry releases reference to it,
75 * GID entry is detached from the table.
76 */
77 GID_TABLE_ENTRY_PENDING_DEL = 3,
78};
79
80struct roce_gid_ndev_storage {
81 struct rcu_head rcu_head;
82 struct net_device *ndev;
83};
84
85struct ib_gid_table_entry {
86 struct kref kref;
87 struct work_struct del_work;
88 struct ib_gid_attr attr;
89 void *context;
90 /* Store the ndev pointer to release reference later on in
91 * call_rcu context because by that time gid_table_entry
92 * and attr might be already freed. So keep a copy of it.
93 * ndev_storage is freed by rcu callback.
94 */
95 struct roce_gid_ndev_storage *ndev_storage;
96 enum gid_table_entry_state state;
97};
98
99struct ib_gid_table {
100 int sz;
101 /* In RoCE, adding a GID to the table requires:
102 * (a) Find if this GID is already exists.
103 * (b) Find a free space.
104 * (c) Write the new GID
105 *
106 * Delete requires different set of operations:
107 * (a) Find the GID
108 * (b) Delete it.
109 *
110 **/
111 /* Any writer to data_vec must hold this lock and the write side of
112 * rwlock. Readers must hold only rwlock. All writers must be in a
113 * sleepable context.
114 */
115 struct mutex lock;
116 /* rwlock protects data_vec[ix]->state and entry pointer.
117 */
118 rwlock_t rwlock;
119 struct ib_gid_table_entry **data_vec;
120 /* bit field, each bit indicates the index of default GID */
121 u32 default_gid_indices;
122};
123
124static void dispatch_gid_change_event(struct ib_device *ib_dev, u32 port)
125{
126 struct ib_event event;
127
128 event.device = ib_dev;
129 event.element.port_num = port;
130 event.event = IB_EVENT_GID_CHANGE;
131
132 ib_dispatch_event_clients(&event);
133}
134
135static const char * const gid_type_str[] = {
136 /* IB/RoCE v1 value is set for IB_GID_TYPE_IB and IB_GID_TYPE_ROCE for
137 * user space compatibility reasons.
138 */
139 [IB_GID_TYPE_IB] = "IB/RoCE v1",
140 [IB_GID_TYPE_ROCE] = "IB/RoCE v1",
141 [IB_GID_TYPE_ROCE_UDP_ENCAP] = "RoCE v2",
142};
143
144const char *ib_cache_gid_type_str(enum ib_gid_type gid_type)
145{
146 if (gid_type < ARRAY_SIZE(gid_type_str) && gid_type_str[gid_type])
147 return gid_type_str[gid_type];
148
149 return "Invalid GID type";
150}
151EXPORT_SYMBOL(ib_cache_gid_type_str);
152
153/** rdma_is_zero_gid - Check if given GID is zero or not.
154 * @gid: GID to check
155 * Returns true if given GID is zero, returns false otherwise.
156 */
157bool rdma_is_zero_gid(const union ib_gid *gid)
158{
159 return !memcmp(gid, &zgid, sizeof(*gid));
160}
161EXPORT_SYMBOL(rdma_is_zero_gid);
162
163/** is_gid_index_default - Check if a given index belongs to
164 * reserved default GIDs or not.
165 * @table: GID table pointer
166 * @index: Index to check in GID table
167 * Returns true if index is one of the reserved default GID index otherwise
168 * returns false.
169 */
170static bool is_gid_index_default(const struct ib_gid_table *table,
171 unsigned int index)
172{
173 return index < 32 && (BIT(index) & table->default_gid_indices);
174}
175
176int ib_cache_gid_parse_type_str(const char *buf)
177{
178 unsigned int i;
179 size_t len;
180 int err = -EINVAL;
181
182 len = strlen(buf);
183 if (len == 0)
184 return -EINVAL;
185
186 if (buf[len - 1] == '\n')
187 len--;
188
189 for (i = 0; i < ARRAY_SIZE(gid_type_str); ++i)
190 if (gid_type_str[i] && !strncmp(buf, gid_type_str[i], len) &&
191 len == strlen(gid_type_str[i])) {
192 err = i;
193 break;
194 }
195
196 return err;
197}
198EXPORT_SYMBOL(ib_cache_gid_parse_type_str);
199
200static struct ib_gid_table *rdma_gid_table(struct ib_device *device, u32 port)
201{
202 return device->port_data[port].cache.gid;
203}
204
205static bool is_gid_entry_free(const struct ib_gid_table_entry *entry)
206{
207 return !entry;
208}
209
210static bool is_gid_entry_valid(const struct ib_gid_table_entry *entry)
211{
212 return entry && entry->state == GID_TABLE_ENTRY_VALID;
213}
214
215static void schedule_free_gid(struct kref *kref)
216{
217 struct ib_gid_table_entry *entry =
218 container_of(kref, struct ib_gid_table_entry, kref);
219
220 queue_work(ib_wq, &entry->del_work);
221}
222
223static void put_gid_ndev(struct rcu_head *head)
224{
225 struct roce_gid_ndev_storage *storage =
226 container_of(head, struct roce_gid_ndev_storage, rcu_head);
227
228 WARN_ON(!storage->ndev);
229 /* At this point its safe to release netdev reference,
230 * as all callers working on gid_attr->ndev are done
231 * using this netdev.
232 */
233 dev_put(storage->ndev);
234 kfree(storage);
235}
236
237static void free_gid_entry_locked(struct ib_gid_table_entry *entry)
238{
239 struct ib_device *device = entry->attr.device;
240 u32 port_num = entry->attr.port_num;
241 struct ib_gid_table *table = rdma_gid_table(device, port_num);
242
243 dev_dbg(&device->dev, "%s port=%u index=%u gid %pI6\n", __func__,
244 port_num, entry->attr.index, entry->attr.gid.raw);
245
246 write_lock_irq(&table->rwlock);
247
248 /*
249 * The only way to avoid overwriting NULL in table is
250 * by comparing if it is same entry in table or not!
251 * If new entry in table is added by the time we free here,
252 * don't overwrite the table entry.
253 */
254 if (entry == table->data_vec[entry->attr.index])
255 table->data_vec[entry->attr.index] = NULL;
256 /* Now this index is ready to be allocated */
257 write_unlock_irq(&table->rwlock);
258
259 if (entry->ndev_storage)
260 call_rcu(&entry->ndev_storage->rcu_head, put_gid_ndev);
261 kfree(entry);
262}
263
264static void free_gid_entry(struct kref *kref)
265{
266 struct ib_gid_table_entry *entry =
267 container_of(kref, struct ib_gid_table_entry, kref);
268
269 free_gid_entry_locked(entry);
270}
271
272/**
273 * free_gid_work - Release reference to the GID entry
274 * @work: Work structure to refer to GID entry which needs to be
275 * deleted.
276 *
277 * free_gid_work() frees the entry from the HCA's hardware table
278 * if provider supports it. It releases reference to netdevice.
279 */
280static void free_gid_work(struct work_struct *work)
281{
282 struct ib_gid_table_entry *entry =
283 container_of(work, struct ib_gid_table_entry, del_work);
284 struct ib_device *device = entry->attr.device;
285 u32 port_num = entry->attr.port_num;
286 struct ib_gid_table *table = rdma_gid_table(device, port_num);
287
288 mutex_lock(&table->lock);
289 free_gid_entry_locked(entry);
290 mutex_unlock(&table->lock);
291}
292
293static struct ib_gid_table_entry *
294alloc_gid_entry(const struct ib_gid_attr *attr)
295{
296 struct ib_gid_table_entry *entry;
297 struct net_device *ndev;
298
299 entry = kzalloc(sizeof(*entry), GFP_KERNEL);
300 if (!entry)
301 return NULL;
302
303 ndev = rcu_dereference_protected(attr->ndev, 1);
304 if (ndev) {
305 entry->ndev_storage = kzalloc(sizeof(*entry->ndev_storage),
306 GFP_KERNEL);
307 if (!entry->ndev_storage) {
308 kfree(entry);
309 return NULL;
310 }
311 dev_hold(ndev);
312 entry->ndev_storage->ndev = ndev;
313 }
314 kref_init(&entry->kref);
315 memcpy(&entry->attr, attr, sizeof(*attr));
316 INIT_WORK(&entry->del_work, free_gid_work);
317 entry->state = GID_TABLE_ENTRY_INVALID;
318 return entry;
319}
320
321static void store_gid_entry(struct ib_gid_table *table,
322 struct ib_gid_table_entry *entry)
323{
324 entry->state = GID_TABLE_ENTRY_VALID;
325
326 dev_dbg(&entry->attr.device->dev, "%s port=%u index=%u gid %pI6\n",
327 __func__, entry->attr.port_num, entry->attr.index,
328 entry->attr.gid.raw);
329
330 lockdep_assert_held(&table->lock);
331 write_lock_irq(&table->rwlock);
332 table->data_vec[entry->attr.index] = entry;
333 write_unlock_irq(&table->rwlock);
334}
335
336static void get_gid_entry(struct ib_gid_table_entry *entry)
337{
338 kref_get(&entry->kref);
339}
340
341static void put_gid_entry(struct ib_gid_table_entry *entry)
342{
343 kref_put(&entry->kref, schedule_free_gid);
344}
345
346static void put_gid_entry_locked(struct ib_gid_table_entry *entry)
347{
348 kref_put(&entry->kref, free_gid_entry);
349}
350
351static int add_roce_gid(struct ib_gid_table_entry *entry)
352{
353 const struct ib_gid_attr *attr = &entry->attr;
354 int ret;
355
356 if (!attr->ndev) {
357 dev_err(&attr->device->dev, "%s NULL netdev port=%u index=%u\n",
358 __func__, attr->port_num, attr->index);
359 return -EINVAL;
360 }
361 if (rdma_cap_roce_gid_table(attr->device, attr->port_num)) {
362 ret = attr->device->ops.add_gid(attr, &entry->context);
363 if (ret) {
364 dev_err(&attr->device->dev,
365 "%s GID add failed port=%u index=%u\n",
366 __func__, attr->port_num, attr->index);
367 return ret;
368 }
369 }
370 return 0;
371}
372
373/**
374 * del_gid - Delete GID table entry
375 *
376 * @ib_dev: IB device whose GID entry to be deleted
377 * @port: Port number of the IB device
378 * @table: GID table of the IB device for a port
379 * @ix: GID entry index to delete
380 *
381 */
382static void del_gid(struct ib_device *ib_dev, u32 port,
383 struct ib_gid_table *table, int ix)
384{
385 struct roce_gid_ndev_storage *ndev_storage;
386 struct ib_gid_table_entry *entry;
387
388 lockdep_assert_held(&table->lock);
389
390 dev_dbg(&ib_dev->dev, "%s port=%u index=%d gid %pI6\n", __func__, port,
391 ix, table->data_vec[ix]->attr.gid.raw);
392
393 write_lock_irq(&table->rwlock);
394 entry = table->data_vec[ix];
395 entry->state = GID_TABLE_ENTRY_PENDING_DEL;
396 /*
397 * For non RoCE protocol, GID entry slot is ready to use.
398 */
399 if (!rdma_protocol_roce(ib_dev, port))
400 table->data_vec[ix] = NULL;
401 write_unlock_irq(&table->rwlock);
402
403 if (rdma_cap_roce_gid_table(ib_dev, port))
404 ib_dev->ops.del_gid(&entry->attr, &entry->context);
405
406 ndev_storage = entry->ndev_storage;
407 if (ndev_storage) {
408 entry->ndev_storage = NULL;
409 rcu_assign_pointer(entry->attr.ndev, NULL);
410 call_rcu(&ndev_storage->rcu_head, put_gid_ndev);
411 }
412
413 put_gid_entry_locked(entry);
414}
415
416/**
417 * add_modify_gid - Add or modify GID table entry
418 *
419 * @table: GID table in which GID to be added or modified
420 * @attr: Attributes of the GID
421 *
422 * Returns 0 on success or appropriate error code. It accepts zero
423 * GID addition for non RoCE ports for HCA's who report them as valid
424 * GID. However such zero GIDs are not added to the cache.
425 */
426static int add_modify_gid(struct ib_gid_table *table,
427 const struct ib_gid_attr *attr)
428{
429 struct ib_gid_table_entry *entry;
430 int ret = 0;
431
432 /*
433 * Invalidate any old entry in the table to make it safe to write to
434 * this index.
435 */
436 if (is_gid_entry_valid(table->data_vec[attr->index]))
437 del_gid(attr->device, attr->port_num, table, attr->index);
438
439 /*
440 * Some HCA's report multiple GID entries with only one valid GID, and
441 * leave other unused entries as the zero GID. Convert zero GIDs to
442 * empty table entries instead of storing them.
443 */
444 if (rdma_is_zero_gid(&attr->gid))
445 return 0;
446
447 entry = alloc_gid_entry(attr);
448 if (!entry)
449 return -ENOMEM;
450
451 if (rdma_protocol_roce(attr->device, attr->port_num)) {
452 ret = add_roce_gid(entry);
453 if (ret)
454 goto done;
455 }
456
457 store_gid_entry(table, entry);
458 return 0;
459
460done:
461 put_gid_entry(entry);
462 return ret;
463}
464
465/* rwlock should be read locked, or lock should be held */
466static int find_gid(struct ib_gid_table *table, const union ib_gid *gid,
467 const struct ib_gid_attr *val, bool default_gid,
468 unsigned long mask, int *pempty)
469{
470 int i = 0;
471 int found = -1;
472 int empty = pempty ? -1 : 0;
473
474 while (i < table->sz && (found < 0 || empty < 0)) {
475 struct ib_gid_table_entry *data = table->data_vec[i];
476 struct ib_gid_attr *attr;
477 int curr_index = i;
478
479 i++;
480
481 /* find_gid() is used during GID addition where it is expected
482 * to return a free entry slot which is not duplicate.
483 * Free entry slot is requested and returned if pempty is set,
484 * so lookup free slot only if requested.
485 */
486 if (pempty && empty < 0) {
487 if (is_gid_entry_free(data) &&
488 default_gid ==
489 is_gid_index_default(table, curr_index)) {
490 /*
491 * Found an invalid (free) entry; allocate it.
492 * If default GID is requested, then our
493 * found slot must be one of the DEFAULT
494 * reserved slots or we fail.
495 * This ensures that only DEFAULT reserved
496 * slots are used for default property GIDs.
497 */
498 empty = curr_index;
499 }
500 }
501
502 /*
503 * Additionally find_gid() is used to find valid entry during
504 * lookup operation; so ignore the entries which are marked as
505 * pending for removal and the entries which are marked as
506 * invalid.
507 */
508 if (!is_gid_entry_valid(data))
509 continue;
510
511 if (found >= 0)
512 continue;
513
514 attr = &data->attr;
515 if (mask & GID_ATTR_FIND_MASK_GID_TYPE &&
516 attr->gid_type != val->gid_type)
517 continue;
518
519 if (mask & GID_ATTR_FIND_MASK_GID &&
520 memcmp(gid, &data->attr.gid, sizeof(*gid)))
521 continue;
522
523 if (mask & GID_ATTR_FIND_MASK_NETDEV &&
524 attr->ndev != val->ndev)
525 continue;
526
527 if (mask & GID_ATTR_FIND_MASK_DEFAULT &&
528 is_gid_index_default(table, curr_index) != default_gid)
529 continue;
530
531 found = curr_index;
532 }
533
534 if (pempty)
535 *pempty = empty;
536
537 return found;
538}
539
540static void make_default_gid(struct net_device *dev, union ib_gid *gid)
541{
542 gid->global.subnet_prefix = cpu_to_be64(0xfe80000000000000LL);
543 addrconf_ifid_eui48(&gid->raw[8], dev);
544}
545
546static int __ib_cache_gid_add(struct ib_device *ib_dev, u32 port,
547 union ib_gid *gid, struct ib_gid_attr *attr,
548 unsigned long mask, bool default_gid)
549{
550 struct ib_gid_table *table;
551 int ret = 0;
552 int empty;
553 int ix;
554
555 /* Do not allow adding zero GID in support of
556 * IB spec version 1.3 section 4.1.1 point (6) and
557 * section 12.7.10 and section 12.7.20
558 */
559 if (rdma_is_zero_gid(gid))
560 return -EINVAL;
561
562 table = rdma_gid_table(ib_dev, port);
563
564 mutex_lock(&table->lock);
565
566 ix = find_gid(table, gid, attr, default_gid, mask, &empty);
567 if (ix >= 0)
568 goto out_unlock;
569
570 if (empty < 0) {
571 ret = -ENOSPC;
572 goto out_unlock;
573 }
574 attr->device = ib_dev;
575 attr->index = empty;
576 attr->port_num = port;
577 attr->gid = *gid;
578 ret = add_modify_gid(table, attr);
579 if (!ret)
580 dispatch_gid_change_event(ib_dev, port);
581
582out_unlock:
583 mutex_unlock(&table->lock);
584 if (ret)
585 pr_warn("%s: unable to add gid %pI6 error=%d\n",
586 __func__, gid->raw, ret);
587 return ret;
588}
589
590int ib_cache_gid_add(struct ib_device *ib_dev, u32 port,
591 union ib_gid *gid, struct ib_gid_attr *attr)
592{
593 unsigned long mask = GID_ATTR_FIND_MASK_GID |
594 GID_ATTR_FIND_MASK_GID_TYPE |
595 GID_ATTR_FIND_MASK_NETDEV;
596
597 return __ib_cache_gid_add(ib_dev, port, gid, attr, mask, false);
598}
599
600static int
601_ib_cache_gid_del(struct ib_device *ib_dev, u32 port,
602 union ib_gid *gid, struct ib_gid_attr *attr,
603 unsigned long mask, bool default_gid)
604{
605 struct ib_gid_table *table;
606 int ret = 0;
607 int ix;
608
609 table = rdma_gid_table(ib_dev, port);
610
611 mutex_lock(&table->lock);
612
613 ix = find_gid(table, gid, attr, default_gid, mask, NULL);
614 if (ix < 0) {
615 ret = -EINVAL;
616 goto out_unlock;
617 }
618
619 del_gid(ib_dev, port, table, ix);
620 dispatch_gid_change_event(ib_dev, port);
621
622out_unlock:
623 mutex_unlock(&table->lock);
624 if (ret)
625 pr_debug("%s: can't delete gid %pI6 error=%d\n",
626 __func__, gid->raw, ret);
627 return ret;
628}
629
630int ib_cache_gid_del(struct ib_device *ib_dev, u32 port,
631 union ib_gid *gid, struct ib_gid_attr *attr)
632{
633 unsigned long mask = GID_ATTR_FIND_MASK_GID |
634 GID_ATTR_FIND_MASK_GID_TYPE |
635 GID_ATTR_FIND_MASK_DEFAULT |
636 GID_ATTR_FIND_MASK_NETDEV;
637
638 return _ib_cache_gid_del(ib_dev, port, gid, attr, mask, false);
639}
640
641int ib_cache_gid_del_all_netdev_gids(struct ib_device *ib_dev, u32 port,
642 struct net_device *ndev)
643{
644 struct ib_gid_table *table;
645 int ix;
646 bool deleted = false;
647
648 table = rdma_gid_table(ib_dev, port);
649
650 mutex_lock(&table->lock);
651
652 for (ix = 0; ix < table->sz; ix++) {
653 if (is_gid_entry_valid(table->data_vec[ix]) &&
654 table->data_vec[ix]->attr.ndev == ndev) {
655 del_gid(ib_dev, port, table, ix);
656 deleted = true;
657 }
658 }
659
660 mutex_unlock(&table->lock);
661
662 if (deleted)
663 dispatch_gid_change_event(ib_dev, port);
664
665 return 0;
666}
667
668/**
669 * rdma_find_gid_by_port - Returns the GID entry attributes when it finds
670 * a valid GID entry for given search parameters. It searches for the specified
671 * GID value in the local software cache.
672 * @ib_dev: The device to query.
673 * @gid: The GID value to search for.
674 * @gid_type: The GID type to search for.
675 * @port: The port number of the device where the GID value should be searched.
676 * @ndev: In RoCE, the net device of the device. NULL means ignore.
677 *
678 * Returns sgid attributes if the GID is found with valid reference or
679 * returns ERR_PTR for the error.
680 * The caller must invoke rdma_put_gid_attr() to release the reference.
681 */
682const struct ib_gid_attr *
683rdma_find_gid_by_port(struct ib_device *ib_dev,
684 const union ib_gid *gid,
685 enum ib_gid_type gid_type,
686 u32 port, struct net_device *ndev)
687{
688 int local_index;
689 struct ib_gid_table *table;
690 unsigned long mask = GID_ATTR_FIND_MASK_GID |
691 GID_ATTR_FIND_MASK_GID_TYPE;
692 struct ib_gid_attr val = {.ndev = ndev, .gid_type = gid_type};
693 const struct ib_gid_attr *attr;
694 unsigned long flags;
695
696 if (!rdma_is_port_valid(ib_dev, port))
697 return ERR_PTR(-ENOENT);
698
699 table = rdma_gid_table(ib_dev, port);
700
701 if (ndev)
702 mask |= GID_ATTR_FIND_MASK_NETDEV;
703
704 read_lock_irqsave(&table->rwlock, flags);
705 local_index = find_gid(table, gid, &val, false, mask, NULL);
706 if (local_index >= 0) {
707 get_gid_entry(table->data_vec[local_index]);
708 attr = &table->data_vec[local_index]->attr;
709 read_unlock_irqrestore(&table->rwlock, flags);
710 return attr;
711 }
712
713 read_unlock_irqrestore(&table->rwlock, flags);
714 return ERR_PTR(-ENOENT);
715}
716EXPORT_SYMBOL(rdma_find_gid_by_port);
717
718/**
719 * rdma_find_gid_by_filter - Returns the GID table attribute where a
720 * specified GID value occurs
721 * @ib_dev: The device to query.
722 * @gid: The GID value to search for.
723 * @port: The port number of the device where the GID value could be
724 * searched.
725 * @filter: The filter function is executed on any matching GID in the table.
726 * If the filter function returns true, the corresponding index is returned,
727 * otherwise, we continue searching the GID table. It's guaranteed that
728 * while filter is executed, ndev field is valid and the structure won't
729 * change. filter is executed in an atomic context. filter must not be NULL.
730 * @context: Private data to pass into the call-back.
731 *
732 * rdma_find_gid_by_filter() searches for the specified GID value
733 * of which the filter function returns true in the port's GID table.
734 *
735 */
736const struct ib_gid_attr *rdma_find_gid_by_filter(
737 struct ib_device *ib_dev, const union ib_gid *gid, u32 port,
738 bool (*filter)(const union ib_gid *gid, const struct ib_gid_attr *,
739 void *),
740 void *context)
741{
742 const struct ib_gid_attr *res = ERR_PTR(-ENOENT);
743 struct ib_gid_table *table;
744 unsigned long flags;
745 unsigned int i;
746
747 if (!rdma_is_port_valid(ib_dev, port))
748 return ERR_PTR(-EINVAL);
749
750 table = rdma_gid_table(ib_dev, port);
751
752 read_lock_irqsave(&table->rwlock, flags);
753 for (i = 0; i < table->sz; i++) {
754 struct ib_gid_table_entry *entry = table->data_vec[i];
755
756 if (!is_gid_entry_valid(entry))
757 continue;
758
759 if (memcmp(gid, &entry->attr.gid, sizeof(*gid)))
760 continue;
761
762 if (filter(gid, &entry->attr, context)) {
763 get_gid_entry(entry);
764 res = &entry->attr;
765 break;
766 }
767 }
768 read_unlock_irqrestore(&table->rwlock, flags);
769 return res;
770}
771
772static struct ib_gid_table *alloc_gid_table(int sz)
773{
774 struct ib_gid_table *table = kzalloc(sizeof(*table), GFP_KERNEL);
775
776 if (!table)
777 return NULL;
778
779 table->data_vec = kcalloc(sz, sizeof(*table->data_vec), GFP_KERNEL);
780 if (!table->data_vec)
781 goto err_free_table;
782
783 mutex_init(&table->lock);
784
785 table->sz = sz;
786 rwlock_init(&table->rwlock);
787 return table;
788
789err_free_table:
790 kfree(table);
791 return NULL;
792}
793
794static void release_gid_table(struct ib_device *device,
795 struct ib_gid_table *table)
796{
797 int i;
798
799 if (!table)
800 return;
801
802 for (i = 0; i < table->sz; i++) {
803 if (is_gid_entry_free(table->data_vec[i]))
804 continue;
805
806 WARN_ONCE(true,
807 "GID entry ref leak for dev %s index %d ref=%u\n",
808 dev_name(&device->dev), i,
809 kref_read(&table->data_vec[i]->kref));
810 }
811
812 mutex_destroy(&table->lock);
813 kfree(table->data_vec);
814 kfree(table);
815}
816
817static void cleanup_gid_table_port(struct ib_device *ib_dev, u32 port,
818 struct ib_gid_table *table)
819{
820 int i;
821
822 if (!table)
823 return;
824
825 mutex_lock(&table->lock);
826 for (i = 0; i < table->sz; ++i) {
827 if (is_gid_entry_valid(table->data_vec[i]))
828 del_gid(ib_dev, port, table, i);
829 }
830 mutex_unlock(&table->lock);
831}
832
833void ib_cache_gid_set_default_gid(struct ib_device *ib_dev, u32 port,
834 struct net_device *ndev,
835 unsigned long gid_type_mask,
836 enum ib_cache_gid_default_mode mode)
837{
838 union ib_gid gid = { };
839 struct ib_gid_attr gid_attr;
840 unsigned int gid_type;
841 unsigned long mask;
842
843 mask = GID_ATTR_FIND_MASK_GID_TYPE |
844 GID_ATTR_FIND_MASK_DEFAULT |
845 GID_ATTR_FIND_MASK_NETDEV;
846 memset(&gid_attr, 0, sizeof(gid_attr));
847 gid_attr.ndev = ndev;
848
849 for (gid_type = 0; gid_type < IB_GID_TYPE_SIZE; ++gid_type) {
850 if (1UL << gid_type & ~gid_type_mask)
851 continue;
852
853 gid_attr.gid_type = gid_type;
854
855 if (mode == IB_CACHE_GID_DEFAULT_MODE_SET) {
856 make_default_gid(ndev, &gid);
857 __ib_cache_gid_add(ib_dev, port, &gid,
858 &gid_attr, mask, true);
859 } else if (mode == IB_CACHE_GID_DEFAULT_MODE_DELETE) {
860 _ib_cache_gid_del(ib_dev, port, &gid,
861 &gid_attr, mask, true);
862 }
863 }
864}
865
866static void gid_table_reserve_default(struct ib_device *ib_dev, u32 port,
867 struct ib_gid_table *table)
868{
869 unsigned int i;
870 unsigned long roce_gid_type_mask;
871 unsigned int num_default_gids;
872
873 roce_gid_type_mask = roce_gid_type_mask_support(ib_dev, port);
874 num_default_gids = hweight_long(roce_gid_type_mask);
875 /* Reserve starting indices for default GIDs */
876 for (i = 0; i < num_default_gids && i < table->sz; i++)
877 table->default_gid_indices |= BIT(i);
878}
879
880
881static void gid_table_release_one(struct ib_device *ib_dev)
882{
883 u32 p;
884
885 rdma_for_each_port (ib_dev, p) {
886 release_gid_table(ib_dev, ib_dev->port_data[p].cache.gid);
887 ib_dev->port_data[p].cache.gid = NULL;
888 }
889}
890
891static int _gid_table_setup_one(struct ib_device *ib_dev)
892{
893 struct ib_gid_table *table;
894 u32 rdma_port;
895
896 rdma_for_each_port (ib_dev, rdma_port) {
897 table = alloc_gid_table(
898 ib_dev->port_data[rdma_port].immutable.gid_tbl_len);
899 if (!table)
900 goto rollback_table_setup;
901
902 gid_table_reserve_default(ib_dev, rdma_port, table);
903 ib_dev->port_data[rdma_port].cache.gid = table;
904 }
905 return 0;
906
907rollback_table_setup:
908 gid_table_release_one(ib_dev);
909 return -ENOMEM;
910}
911
912static void gid_table_cleanup_one(struct ib_device *ib_dev)
913{
914 u32 p;
915
916 rdma_for_each_port (ib_dev, p)
917 cleanup_gid_table_port(ib_dev, p,
918 ib_dev->port_data[p].cache.gid);
919}
920
921static int gid_table_setup_one(struct ib_device *ib_dev)
922{
923 int err;
924
925 err = _gid_table_setup_one(ib_dev);
926
927 if (err)
928 return err;
929
930 rdma_roce_rescan_device(ib_dev);
931
932 return err;
933}
934
935/**
936 * rdma_query_gid - Read the GID content from the GID software cache
937 * @device: Device to query the GID
938 * @port_num: Port number of the device
939 * @index: Index of the GID table entry to read
940 * @gid: Pointer to GID where to store the entry's GID
941 *
942 * rdma_query_gid() only reads the GID entry content for requested device,
943 * port and index. It reads for IB, RoCE and iWarp link layers. It doesn't
944 * hold any reference to the GID table entry in the HCA or software cache.
945 *
946 * Returns 0 on success or appropriate error code.
947 *
948 */
949int rdma_query_gid(struct ib_device *device, u32 port_num,
950 int index, union ib_gid *gid)
951{
952 struct ib_gid_table *table;
953 unsigned long flags;
954 int res;
955
956 if (!rdma_is_port_valid(device, port_num))
957 return -EINVAL;
958
959 table = rdma_gid_table(device, port_num);
960 read_lock_irqsave(&table->rwlock, flags);
961
962 if (index < 0 || index >= table->sz) {
963 res = -EINVAL;
964 goto done;
965 }
966
967 if (!is_gid_entry_valid(table->data_vec[index])) {
968 res = -ENOENT;
969 goto done;
970 }
971
972 memcpy(gid, &table->data_vec[index]->attr.gid, sizeof(*gid));
973 res = 0;
974
975done:
976 read_unlock_irqrestore(&table->rwlock, flags);
977 return res;
978}
979EXPORT_SYMBOL(rdma_query_gid);
980
981/**
982 * rdma_read_gid_hw_context - Read the HW GID context from GID attribute
983 * @attr: Potinter to the GID attribute
984 *
985 * rdma_read_gid_hw_context() reads the drivers GID HW context corresponding
986 * to the SGID attr. Callers are required to already be holding the reference
987 * to an existing GID entry.
988 *
989 * Returns the HW GID context
990 *
991 */
992void *rdma_read_gid_hw_context(const struct ib_gid_attr *attr)
993{
994 return container_of(attr, struct ib_gid_table_entry, attr)->context;
995}
996EXPORT_SYMBOL(rdma_read_gid_hw_context);
997
998/**
999 * rdma_find_gid - Returns SGID attributes if the matching GID is found.
1000 * @device: The device to query.
1001 * @gid: The GID value to search for.
1002 * @gid_type: The GID type to search for.
1003 * @ndev: In RoCE, the net device of the device. NULL means ignore.
1004 *
1005 * rdma_find_gid() searches for the specified GID value in the software cache.
1006 *
1007 * Returns GID attributes if a valid GID is found or returns ERR_PTR for the
1008 * error. The caller must invoke rdma_put_gid_attr() to release the reference.
1009 *
1010 */
1011const struct ib_gid_attr *rdma_find_gid(struct ib_device *device,
1012 const union ib_gid *gid,
1013 enum ib_gid_type gid_type,
1014 struct net_device *ndev)
1015{
1016 unsigned long mask = GID_ATTR_FIND_MASK_GID |
1017 GID_ATTR_FIND_MASK_GID_TYPE;
1018 struct ib_gid_attr gid_attr_val = {.ndev = ndev, .gid_type = gid_type};
1019 u32 p;
1020
1021 if (ndev)
1022 mask |= GID_ATTR_FIND_MASK_NETDEV;
1023
1024 rdma_for_each_port(device, p) {
1025 struct ib_gid_table *table;
1026 unsigned long flags;
1027 int index;
1028
1029 table = device->port_data[p].cache.gid;
1030 read_lock_irqsave(&table->rwlock, flags);
1031 index = find_gid(table, gid, &gid_attr_val, false, mask, NULL);
1032 if (index >= 0) {
1033 const struct ib_gid_attr *attr;
1034
1035 get_gid_entry(table->data_vec[index]);
1036 attr = &table->data_vec[index]->attr;
1037 read_unlock_irqrestore(&table->rwlock, flags);
1038 return attr;
1039 }
1040 read_unlock_irqrestore(&table->rwlock, flags);
1041 }
1042
1043 return ERR_PTR(-ENOENT);
1044}
1045EXPORT_SYMBOL(rdma_find_gid);
1046
1047int ib_get_cached_pkey(struct ib_device *device,
1048 u32 port_num,
1049 int index,
1050 u16 *pkey)
1051{
1052 struct ib_pkey_cache *cache;
1053 unsigned long flags;
1054 int ret = 0;
1055
1056 if (!rdma_is_port_valid(device, port_num))
1057 return -EINVAL;
1058
1059 read_lock_irqsave(&device->cache_lock, flags);
1060
1061 cache = device->port_data[port_num].cache.pkey;
1062
1063 if (!cache || index < 0 || index >= cache->table_len)
1064 ret = -EINVAL;
1065 else
1066 *pkey = cache->table[index];
1067
1068 read_unlock_irqrestore(&device->cache_lock, flags);
1069
1070 return ret;
1071}
1072EXPORT_SYMBOL(ib_get_cached_pkey);
1073
1074void ib_get_cached_subnet_prefix(struct ib_device *device, u32 port_num,
1075 u64 *sn_pfx)
1076{
1077 unsigned long flags;
1078
1079 read_lock_irqsave(&device->cache_lock, flags);
1080 *sn_pfx = device->port_data[port_num].cache.subnet_prefix;
1081 read_unlock_irqrestore(&device->cache_lock, flags);
1082}
1083EXPORT_SYMBOL(ib_get_cached_subnet_prefix);
1084
1085int ib_find_cached_pkey(struct ib_device *device, u32 port_num,
1086 u16 pkey, u16 *index)
1087{
1088 struct ib_pkey_cache *cache;
1089 unsigned long flags;
1090 int i;
1091 int ret = -ENOENT;
1092 int partial_ix = -1;
1093
1094 if (!rdma_is_port_valid(device, port_num))
1095 return -EINVAL;
1096
1097 read_lock_irqsave(&device->cache_lock, flags);
1098
1099 cache = device->port_data[port_num].cache.pkey;
1100 if (!cache) {
1101 ret = -EINVAL;
1102 goto err;
1103 }
1104
1105 *index = -1;
1106
1107 for (i = 0; i < cache->table_len; ++i)
1108 if ((cache->table[i] & 0x7fff) == (pkey & 0x7fff)) {
1109 if (cache->table[i] & 0x8000) {
1110 *index = i;
1111 ret = 0;
1112 break;
1113 } else {
1114 partial_ix = i;
1115 }
1116 }
1117
1118 if (ret && partial_ix >= 0) {
1119 *index = partial_ix;
1120 ret = 0;
1121 }
1122
1123err:
1124 read_unlock_irqrestore(&device->cache_lock, flags);
1125
1126 return ret;
1127}
1128EXPORT_SYMBOL(ib_find_cached_pkey);
1129
1130int ib_find_exact_cached_pkey(struct ib_device *device, u32 port_num,
1131 u16 pkey, u16 *index)
1132{
1133 struct ib_pkey_cache *cache;
1134 unsigned long flags;
1135 int i;
1136 int ret = -ENOENT;
1137
1138 if (!rdma_is_port_valid(device, port_num))
1139 return -EINVAL;
1140
1141 read_lock_irqsave(&device->cache_lock, flags);
1142
1143 cache = device->port_data[port_num].cache.pkey;
1144 if (!cache) {
1145 ret = -EINVAL;
1146 goto err;
1147 }
1148
1149 *index = -1;
1150
1151 for (i = 0; i < cache->table_len; ++i)
1152 if (cache->table[i] == pkey) {
1153 *index = i;
1154 ret = 0;
1155 break;
1156 }
1157
1158err:
1159 read_unlock_irqrestore(&device->cache_lock, flags);
1160
1161 return ret;
1162}
1163EXPORT_SYMBOL(ib_find_exact_cached_pkey);
1164
1165int ib_get_cached_lmc(struct ib_device *device, u32 port_num, u8 *lmc)
1166{
1167 unsigned long flags;
1168 int ret = 0;
1169
1170 if (!rdma_is_port_valid(device, port_num))
1171 return -EINVAL;
1172
1173 read_lock_irqsave(&device->cache_lock, flags);
1174 *lmc = device->port_data[port_num].cache.lmc;
1175 read_unlock_irqrestore(&device->cache_lock, flags);
1176
1177 return ret;
1178}
1179EXPORT_SYMBOL(ib_get_cached_lmc);
1180
1181int ib_get_cached_port_state(struct ib_device *device, u32 port_num,
1182 enum ib_port_state *port_state)
1183{
1184 unsigned long flags;
1185 int ret = 0;
1186
1187 if (!rdma_is_port_valid(device, port_num))
1188 return -EINVAL;
1189
1190 read_lock_irqsave(&device->cache_lock, flags);
1191 *port_state = device->port_data[port_num].cache.port_state;
1192 read_unlock_irqrestore(&device->cache_lock, flags);
1193
1194 return ret;
1195}
1196EXPORT_SYMBOL(ib_get_cached_port_state);
1197
1198/**
1199 * rdma_get_gid_attr - Returns GID attributes for a port of a device
1200 * at a requested gid_index, if a valid GID entry exists.
1201 * @device: The device to query.
1202 * @port_num: The port number on the device where the GID value
1203 * is to be queried.
1204 * @index: Index of the GID table entry whose attributes are to
1205 * be queried.
1206 *
1207 * rdma_get_gid_attr() acquires reference count of gid attributes from the
1208 * cached GID table. Caller must invoke rdma_put_gid_attr() to release
1209 * reference to gid attribute regardless of link layer.
1210 *
1211 * Returns pointer to valid gid attribute or ERR_PTR for the appropriate error
1212 * code.
1213 */
1214const struct ib_gid_attr *
1215rdma_get_gid_attr(struct ib_device *device, u32 port_num, int index)
1216{
1217 const struct ib_gid_attr *attr = ERR_PTR(-ENODATA);
1218 struct ib_gid_table *table;
1219 unsigned long flags;
1220
1221 if (!rdma_is_port_valid(device, port_num))
1222 return ERR_PTR(-EINVAL);
1223
1224 table = rdma_gid_table(device, port_num);
1225 if (index < 0 || index >= table->sz)
1226 return ERR_PTR(-EINVAL);
1227
1228 read_lock_irqsave(&table->rwlock, flags);
1229 if (!is_gid_entry_valid(table->data_vec[index]))
1230 goto done;
1231
1232 get_gid_entry(table->data_vec[index]);
1233 attr = &table->data_vec[index]->attr;
1234done:
1235 read_unlock_irqrestore(&table->rwlock, flags);
1236 return attr;
1237}
1238EXPORT_SYMBOL(rdma_get_gid_attr);
1239
1240/**
1241 * rdma_query_gid_table - Reads GID table entries of all the ports of a device up to max_entries.
1242 * @device: The device to query.
1243 * @entries: Entries where GID entries are returned.
1244 * @max_entries: Maximum number of entries that can be returned.
1245 * Entries array must be allocated to hold max_entries number of entries.
1246 *
1247 * Returns number of entries on success or appropriate error code.
1248 */
1249ssize_t rdma_query_gid_table(struct ib_device *device,
1250 struct ib_uverbs_gid_entry *entries,
1251 size_t max_entries)
1252{
1253 const struct ib_gid_attr *gid_attr;
1254 ssize_t num_entries = 0, ret;
1255 struct ib_gid_table *table;
1256 u32 port_num, i;
1257 struct net_device *ndev;
1258 unsigned long flags;
1259
1260 rdma_for_each_port(device, port_num) {
1261 table = rdma_gid_table(device, port_num);
1262 read_lock_irqsave(&table->rwlock, flags);
1263 for (i = 0; i < table->sz; i++) {
1264 if (!is_gid_entry_valid(table->data_vec[i]))
1265 continue;
1266 if (num_entries >= max_entries) {
1267 ret = -EINVAL;
1268 goto err;
1269 }
1270
1271 gid_attr = &table->data_vec[i]->attr;
1272
1273 memcpy(&entries->gid, &gid_attr->gid,
1274 sizeof(gid_attr->gid));
1275 entries->gid_index = gid_attr->index;
1276 entries->port_num = gid_attr->port_num;
1277 entries->gid_type = gid_attr->gid_type;
1278 ndev = rcu_dereference_protected(
1279 gid_attr->ndev,
1280 lockdep_is_held(&table->rwlock));
1281 if (ndev)
1282 entries->netdev_ifindex = ndev->ifindex;
1283
1284 num_entries++;
1285 entries++;
1286 }
1287 read_unlock_irqrestore(&table->rwlock, flags);
1288 }
1289
1290 return num_entries;
1291err:
1292 read_unlock_irqrestore(&table->rwlock, flags);
1293 return ret;
1294}
1295EXPORT_SYMBOL(rdma_query_gid_table);
1296
1297/**
1298 * rdma_put_gid_attr - Release reference to the GID attribute
1299 * @attr: Pointer to the GID attribute whose reference
1300 * needs to be released.
1301 *
1302 * rdma_put_gid_attr() must be used to release reference whose
1303 * reference is acquired using rdma_get_gid_attr() or any APIs
1304 * which returns a pointer to the ib_gid_attr regardless of link layer
1305 * of IB or RoCE.
1306 *
1307 */
1308void rdma_put_gid_attr(const struct ib_gid_attr *attr)
1309{
1310 struct ib_gid_table_entry *entry =
1311 container_of(attr, struct ib_gid_table_entry, attr);
1312
1313 put_gid_entry(entry);
1314}
1315EXPORT_SYMBOL(rdma_put_gid_attr);
1316
1317/**
1318 * rdma_hold_gid_attr - Get reference to existing GID attribute
1319 *
1320 * @attr: Pointer to the GID attribute whose reference
1321 * needs to be taken.
1322 *
1323 * Increase the reference count to a GID attribute to keep it from being
1324 * freed. Callers are required to already be holding a reference to attribute.
1325 *
1326 */
1327void rdma_hold_gid_attr(const struct ib_gid_attr *attr)
1328{
1329 struct ib_gid_table_entry *entry =
1330 container_of(attr, struct ib_gid_table_entry, attr);
1331
1332 get_gid_entry(entry);
1333}
1334EXPORT_SYMBOL(rdma_hold_gid_attr);
1335
1336/**
1337 * rdma_read_gid_attr_ndev_rcu - Read GID attribute netdevice
1338 * which must be in UP state.
1339 *
1340 * @attr:Pointer to the GID attribute
1341 *
1342 * Returns pointer to netdevice if the netdevice was attached to GID and
1343 * netdevice is in UP state. Caller must hold RCU lock as this API
1344 * reads the netdev flags which can change while netdevice migrates to
1345 * different net namespace. Returns ERR_PTR with error code otherwise.
1346 *
1347 */
1348struct net_device *rdma_read_gid_attr_ndev_rcu(const struct ib_gid_attr *attr)
1349{
1350 struct ib_gid_table_entry *entry =
1351 container_of(attr, struct ib_gid_table_entry, attr);
1352 struct ib_device *device = entry->attr.device;
1353 struct net_device *ndev = ERR_PTR(-EINVAL);
1354 u32 port_num = entry->attr.port_num;
1355 struct ib_gid_table *table;
1356 unsigned long flags;
1357 bool valid;
1358
1359 table = rdma_gid_table(device, port_num);
1360
1361 read_lock_irqsave(&table->rwlock, flags);
1362 valid = is_gid_entry_valid(table->data_vec[attr->index]);
1363 if (valid) {
1364 ndev = rcu_dereference(attr->ndev);
1365 if (!ndev)
1366 ndev = ERR_PTR(-ENODEV);
1367 }
1368 read_unlock_irqrestore(&table->rwlock, flags);
1369 return ndev;
1370}
1371EXPORT_SYMBOL(rdma_read_gid_attr_ndev_rcu);
1372
1373static int get_lower_dev_vlan(struct net_device *lower_dev,
1374 struct netdev_nested_priv *priv)
1375{
1376 u16 *vlan_id = (u16 *)priv->data;
1377
1378 if (is_vlan_dev(lower_dev))
1379 *vlan_id = vlan_dev_vlan_id(lower_dev);
1380
1381 /* We are interested only in first level vlan device, so
1382 * always return 1 to stop iterating over next level devices.
1383 */
1384 return 1;
1385}
1386
1387/**
1388 * rdma_read_gid_l2_fields - Read the vlan ID and source MAC address
1389 * of a GID entry.
1390 *
1391 * @attr: GID attribute pointer whose L2 fields to be read
1392 * @vlan_id: Pointer to vlan id to fill up if the GID entry has
1393 * vlan id. It is optional.
1394 * @smac: Pointer to smac to fill up for a GID entry. It is optional.
1395 *
1396 * rdma_read_gid_l2_fields() returns 0 on success and returns vlan id
1397 * (if gid entry has vlan) and source MAC, or returns error.
1398 */
1399int rdma_read_gid_l2_fields(const struct ib_gid_attr *attr,
1400 u16 *vlan_id, u8 *smac)
1401{
1402 struct netdev_nested_priv priv = {
1403 .data = (void *)vlan_id,
1404 };
1405 struct net_device *ndev;
1406
1407 rcu_read_lock();
1408 ndev = rcu_dereference(attr->ndev);
1409 if (!ndev) {
1410 rcu_read_unlock();
1411 return -ENODEV;
1412 }
1413 if (smac)
1414 ether_addr_copy(smac, ndev->dev_addr);
1415 if (vlan_id) {
1416 *vlan_id = 0xffff;
1417 if (is_vlan_dev(ndev)) {
1418 *vlan_id = vlan_dev_vlan_id(ndev);
1419 } else {
1420 /* If the netdev is upper device and if it's lower
1421 * device is vlan device, consider vlan id of
1422 * the lower vlan device for this gid entry.
1423 */
1424 netdev_walk_all_lower_dev_rcu(attr->ndev,
1425 get_lower_dev_vlan, &priv);
1426 }
1427 }
1428 rcu_read_unlock();
1429 return 0;
1430}
1431EXPORT_SYMBOL(rdma_read_gid_l2_fields);
1432
1433static int config_non_roce_gid_cache(struct ib_device *device,
1434 u32 port, struct ib_port_attr *tprops)
1435{
1436 struct ib_gid_attr gid_attr = {};
1437 struct ib_gid_table *table;
1438 int ret = 0;
1439 int i;
1440
1441 gid_attr.device = device;
1442 gid_attr.port_num = port;
1443 table = rdma_gid_table(device, port);
1444
1445 mutex_lock(&table->lock);
1446 for (i = 0; i < tprops->gid_tbl_len; ++i) {
1447 if (!device->ops.query_gid)
1448 continue;
1449 ret = device->ops.query_gid(device, port, i, &gid_attr.gid);
1450 if (ret) {
1451 dev_warn(&device->dev,
1452 "query_gid failed (%d) for index %d\n", ret,
1453 i);
1454 goto err;
1455 }
1456
1457 if (rdma_protocol_iwarp(device, port)) {
1458 struct net_device *ndev;
1459
1460 ndev = ib_device_get_netdev(device, port);
1461 if (!ndev)
1462 continue;
1463 RCU_INIT_POINTER(gid_attr.ndev, ndev);
1464 dev_put(ndev);
1465 }
1466
1467 gid_attr.index = i;
1468 tprops->subnet_prefix =
1469 be64_to_cpu(gid_attr.gid.global.subnet_prefix);
1470 add_modify_gid(table, &gid_attr);
1471 }
1472err:
1473 mutex_unlock(&table->lock);
1474 return ret;
1475}
1476
1477static int
1478ib_cache_update(struct ib_device *device, u32 port, bool update_gids,
1479 bool update_pkeys, bool enforce_security)
1480{
1481 struct ib_port_attr *tprops = NULL;
1482 struct ib_pkey_cache *pkey_cache = NULL;
1483 struct ib_pkey_cache *old_pkey_cache = NULL;
1484 int i;
1485 int ret;
1486
1487 if (!rdma_is_port_valid(device, port))
1488 return -EINVAL;
1489
1490 tprops = kmalloc(sizeof *tprops, GFP_KERNEL);
1491 if (!tprops)
1492 return -ENOMEM;
1493
1494 ret = ib_query_port(device, port, tprops);
1495 if (ret) {
1496 dev_warn(&device->dev, "ib_query_port failed (%d)\n", ret);
1497 goto err;
1498 }
1499
1500 if (!rdma_protocol_roce(device, port) && update_gids) {
1501 ret = config_non_roce_gid_cache(device, port,
1502 tprops);
1503 if (ret)
1504 goto err;
1505 }
1506
1507 update_pkeys &= !!tprops->pkey_tbl_len;
1508
1509 if (update_pkeys) {
1510 pkey_cache = kmalloc(struct_size(pkey_cache, table,
1511 tprops->pkey_tbl_len),
1512 GFP_KERNEL);
1513 if (!pkey_cache) {
1514 ret = -ENOMEM;
1515 goto err;
1516 }
1517
1518 pkey_cache->table_len = tprops->pkey_tbl_len;
1519
1520 for (i = 0; i < pkey_cache->table_len; ++i) {
1521 ret = ib_query_pkey(device, port, i,
1522 pkey_cache->table + i);
1523 if (ret) {
1524 dev_warn(&device->dev,
1525 "ib_query_pkey failed (%d) for index %d\n",
1526 ret, i);
1527 goto err;
1528 }
1529 }
1530 }
1531
1532 write_lock_irq(&device->cache_lock);
1533
1534 if (update_pkeys) {
1535 old_pkey_cache = device->port_data[port].cache.pkey;
1536 device->port_data[port].cache.pkey = pkey_cache;
1537 }
1538 device->port_data[port].cache.lmc = tprops->lmc;
1539 device->port_data[port].cache.port_state = tprops->state;
1540
1541 device->port_data[port].cache.subnet_prefix = tprops->subnet_prefix;
1542 write_unlock_irq(&device->cache_lock);
1543
1544 if (enforce_security)
1545 ib_security_cache_change(device,
1546 port,
1547 tprops->subnet_prefix);
1548
1549 kfree(old_pkey_cache);
1550 kfree(tprops);
1551 return 0;
1552
1553err:
1554 kfree(pkey_cache);
1555 kfree(tprops);
1556 return ret;
1557}
1558
1559static void ib_cache_event_task(struct work_struct *_work)
1560{
1561 struct ib_update_work *work =
1562 container_of(_work, struct ib_update_work, work);
1563 int ret;
1564
1565 /* Before distributing the cache update event, first sync
1566 * the cache.
1567 */
1568 ret = ib_cache_update(work->event.device, work->event.element.port_num,
1569 work->event.event == IB_EVENT_GID_CHANGE,
1570 work->event.event == IB_EVENT_PKEY_CHANGE,
1571 work->enforce_security);
1572
1573 /* GID event is notified already for individual GID entries by
1574 * dispatch_gid_change_event(). Hence, notifiy for rest of the
1575 * events.
1576 */
1577 if (!ret && work->event.event != IB_EVENT_GID_CHANGE)
1578 ib_dispatch_event_clients(&work->event);
1579
1580 kfree(work);
1581}
1582
1583static void ib_generic_event_task(struct work_struct *_work)
1584{
1585 struct ib_update_work *work =
1586 container_of(_work, struct ib_update_work, work);
1587
1588 ib_dispatch_event_clients(&work->event);
1589 kfree(work);
1590}
1591
1592static bool is_cache_update_event(const struct ib_event *event)
1593{
1594 return (event->event == IB_EVENT_PORT_ERR ||
1595 event->event == IB_EVENT_PORT_ACTIVE ||
1596 event->event == IB_EVENT_LID_CHANGE ||
1597 event->event == IB_EVENT_PKEY_CHANGE ||
1598 event->event == IB_EVENT_CLIENT_REREGISTER ||
1599 event->event == IB_EVENT_GID_CHANGE);
1600}
1601
1602/**
1603 * ib_dispatch_event - Dispatch an asynchronous event
1604 * @event:Event to dispatch
1605 *
1606 * Low-level drivers must call ib_dispatch_event() to dispatch the
1607 * event to all registered event handlers when an asynchronous event
1608 * occurs.
1609 */
1610void ib_dispatch_event(const struct ib_event *event)
1611{
1612 struct ib_update_work *work;
1613
1614 work = kzalloc(sizeof(*work), GFP_ATOMIC);
1615 if (!work)
1616 return;
1617
1618 if (is_cache_update_event(event))
1619 INIT_WORK(&work->work, ib_cache_event_task);
1620 else
1621 INIT_WORK(&work->work, ib_generic_event_task);
1622
1623 work->event = *event;
1624 if (event->event == IB_EVENT_PKEY_CHANGE ||
1625 event->event == IB_EVENT_GID_CHANGE)
1626 work->enforce_security = true;
1627
1628 queue_work(ib_wq, &work->work);
1629}
1630EXPORT_SYMBOL(ib_dispatch_event);
1631
1632int ib_cache_setup_one(struct ib_device *device)
1633{
1634 u32 p;
1635 int err;
1636
1637 err = gid_table_setup_one(device);
1638 if (err)
1639 return err;
1640
1641 rdma_for_each_port (device, p) {
1642 err = ib_cache_update(device, p, true, true, true);
1643 if (err) {
1644 gid_table_cleanup_one(device);
1645 return err;
1646 }
1647 }
1648
1649 return 0;
1650}
1651
1652void ib_cache_release_one(struct ib_device *device)
1653{
1654 u32 p;
1655
1656 /*
1657 * The release function frees all the cache elements.
1658 * This function should be called as part of freeing
1659 * all the device's resources when the cache could no
1660 * longer be accessed.
1661 */
1662 rdma_for_each_port (device, p)
1663 kfree(device->port_data[p].cache.pkey);
1664
1665 gid_table_release_one(device);
1666}
1667
1668void ib_cache_cleanup_one(struct ib_device *device)
1669{
1670 /* The cleanup function waits for all in-progress workqueue
1671 * elements and cleans up the GID cache. This function should be
1672 * called after the device was removed from the devices list and
1673 * all clients were removed, so the cache exists but is
1674 * non-functional and shouldn't be updated anymore.
1675 */
1676 flush_workqueue(ib_wq);
1677 gid_table_cleanup_one(device);
1678
1679 /*
1680 * Flush the wq second time for any pending GID delete work.
1681 */
1682 flush_workqueue(ib_wq);
1683}
1/*
2 * Copyright (c) 2004 Topspin Communications. All rights reserved.
3 * Copyright (c) 2005 Intel Corporation. All rights reserved.
4 * Copyright (c) 2005 Sun Microsystems, Inc. All rights reserved.
5 * Copyright (c) 2005 Voltaire, Inc. All rights reserved.
6 *
7 * This software is available to you under a choice of one of two
8 * licenses. You may choose to be licensed under the terms of the GNU
9 * General Public License (GPL) Version 2, available from the file
10 * COPYING in the main directory of this source tree, or the
11 * OpenIB.org BSD license below:
12 *
13 * Redistribution and use in source and binary forms, with or
14 * without modification, are permitted provided that the following
15 * conditions are met:
16 *
17 * - Redistributions of source code must retain the above
18 * copyright notice, this list of conditions and the following
19 * disclaimer.
20 *
21 * - Redistributions in binary form must reproduce the above
22 * copyright notice, this list of conditions and the following
23 * disclaimer in the documentation and/or other materials
24 * provided with the distribution.
25 *
26 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
27 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
28 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
29 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
30 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
31 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
32 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
33 * SOFTWARE.
34 */
35
36#include <linux/module.h>
37#include <linux/errno.h>
38#include <linux/slab.h>
39#include <linux/workqueue.h>
40#include <linux/netdevice.h>
41#include <net/addrconf.h>
42
43#include <rdma/ib_cache.h>
44
45#include "core_priv.h"
46
47struct ib_pkey_cache {
48 int table_len;
49 u16 table[0];
50};
51
52struct ib_update_work {
53 struct work_struct work;
54 struct ib_device *device;
55 u8 port_num;
56 bool enforce_security;
57};
58
59union ib_gid zgid;
60EXPORT_SYMBOL(zgid);
61
62enum gid_attr_find_mask {
63 GID_ATTR_FIND_MASK_GID = 1UL << 0,
64 GID_ATTR_FIND_MASK_NETDEV = 1UL << 1,
65 GID_ATTR_FIND_MASK_DEFAULT = 1UL << 2,
66 GID_ATTR_FIND_MASK_GID_TYPE = 1UL << 3,
67};
68
69enum gid_table_entry_state {
70 GID_TABLE_ENTRY_INVALID = 1,
71 GID_TABLE_ENTRY_VALID = 2,
72 /*
73 * Indicates that entry is pending to be removed, there may
74 * be active users of this GID entry.
75 * When last user of the GID entry releases reference to it,
76 * GID entry is detached from the table.
77 */
78 GID_TABLE_ENTRY_PENDING_DEL = 3,
79};
80
81struct roce_gid_ndev_storage {
82 struct rcu_head rcu_head;
83 struct net_device *ndev;
84};
85
86struct ib_gid_table_entry {
87 struct kref kref;
88 struct work_struct del_work;
89 struct ib_gid_attr attr;
90 void *context;
91 /* Store the ndev pointer to release reference later on in
92 * call_rcu context because by that time gid_table_entry
93 * and attr might be already freed. So keep a copy of it.
94 * ndev_storage is freed by rcu callback.
95 */
96 struct roce_gid_ndev_storage *ndev_storage;
97 enum gid_table_entry_state state;
98};
99
100struct ib_gid_table {
101 int sz;
102 /* In RoCE, adding a GID to the table requires:
103 * (a) Find if this GID is already exists.
104 * (b) Find a free space.
105 * (c) Write the new GID
106 *
107 * Delete requires different set of operations:
108 * (a) Find the GID
109 * (b) Delete it.
110 *
111 **/
112 /* Any writer to data_vec must hold this lock and the write side of
113 * rwlock. Readers must hold only rwlock. All writers must be in a
114 * sleepable context.
115 */
116 struct mutex lock;
117 /* rwlock protects data_vec[ix]->state and entry pointer.
118 */
119 rwlock_t rwlock;
120 struct ib_gid_table_entry **data_vec;
121 /* bit field, each bit indicates the index of default GID */
122 u32 default_gid_indices;
123};
124
125static void dispatch_gid_change_event(struct ib_device *ib_dev, u8 port)
126{
127 struct ib_event event;
128
129 event.device = ib_dev;
130 event.element.port_num = port;
131 event.event = IB_EVENT_GID_CHANGE;
132
133 ib_dispatch_event(&event);
134}
135
136static const char * const gid_type_str[] = {
137 [IB_GID_TYPE_IB] = "IB/RoCE v1",
138 [IB_GID_TYPE_ROCE_UDP_ENCAP] = "RoCE v2",
139};
140
141const char *ib_cache_gid_type_str(enum ib_gid_type gid_type)
142{
143 if (gid_type < ARRAY_SIZE(gid_type_str) && gid_type_str[gid_type])
144 return gid_type_str[gid_type];
145
146 return "Invalid GID type";
147}
148EXPORT_SYMBOL(ib_cache_gid_type_str);
149
150/** rdma_is_zero_gid - Check if given GID is zero or not.
151 * @gid: GID to check
152 * Returns true if given GID is zero, returns false otherwise.
153 */
154bool rdma_is_zero_gid(const union ib_gid *gid)
155{
156 return !memcmp(gid, &zgid, sizeof(*gid));
157}
158EXPORT_SYMBOL(rdma_is_zero_gid);
159
160/** is_gid_index_default - Check if a given index belongs to
161 * reserved default GIDs or not.
162 * @table: GID table pointer
163 * @index: Index to check in GID table
164 * Returns true if index is one of the reserved default GID index otherwise
165 * returns false.
166 */
167static bool is_gid_index_default(const struct ib_gid_table *table,
168 unsigned int index)
169{
170 return index < 32 && (BIT(index) & table->default_gid_indices);
171}
172
173int ib_cache_gid_parse_type_str(const char *buf)
174{
175 unsigned int i;
176 size_t len;
177 int err = -EINVAL;
178
179 len = strlen(buf);
180 if (len == 0)
181 return -EINVAL;
182
183 if (buf[len - 1] == '\n')
184 len--;
185
186 for (i = 0; i < ARRAY_SIZE(gid_type_str); ++i)
187 if (gid_type_str[i] && !strncmp(buf, gid_type_str[i], len) &&
188 len == strlen(gid_type_str[i])) {
189 err = i;
190 break;
191 }
192
193 return err;
194}
195EXPORT_SYMBOL(ib_cache_gid_parse_type_str);
196
197static struct ib_gid_table *rdma_gid_table(struct ib_device *device, u8 port)
198{
199 return device->port_data[port].cache.gid;
200}
201
202static bool is_gid_entry_free(const struct ib_gid_table_entry *entry)
203{
204 return !entry;
205}
206
207static bool is_gid_entry_valid(const struct ib_gid_table_entry *entry)
208{
209 return entry && entry->state == GID_TABLE_ENTRY_VALID;
210}
211
212static void schedule_free_gid(struct kref *kref)
213{
214 struct ib_gid_table_entry *entry =
215 container_of(kref, struct ib_gid_table_entry, kref);
216
217 queue_work(ib_wq, &entry->del_work);
218}
219
220static void put_gid_ndev(struct rcu_head *head)
221{
222 struct roce_gid_ndev_storage *storage =
223 container_of(head, struct roce_gid_ndev_storage, rcu_head);
224
225 WARN_ON(!storage->ndev);
226 /* At this point its safe to release netdev reference,
227 * as all callers working on gid_attr->ndev are done
228 * using this netdev.
229 */
230 dev_put(storage->ndev);
231 kfree(storage);
232}
233
234static void free_gid_entry_locked(struct ib_gid_table_entry *entry)
235{
236 struct ib_device *device = entry->attr.device;
237 u8 port_num = entry->attr.port_num;
238 struct ib_gid_table *table = rdma_gid_table(device, port_num);
239
240 dev_dbg(&device->dev, "%s port=%d index=%d gid %pI6\n", __func__,
241 port_num, entry->attr.index, entry->attr.gid.raw);
242
243 write_lock_irq(&table->rwlock);
244
245 /*
246 * The only way to avoid overwriting NULL in table is
247 * by comparing if it is same entry in table or not!
248 * If new entry in table is added by the time we free here,
249 * don't overwrite the table entry.
250 */
251 if (entry == table->data_vec[entry->attr.index])
252 table->data_vec[entry->attr.index] = NULL;
253 /* Now this index is ready to be allocated */
254 write_unlock_irq(&table->rwlock);
255
256 if (entry->ndev_storage)
257 call_rcu(&entry->ndev_storage->rcu_head, put_gid_ndev);
258 kfree(entry);
259}
260
261static void free_gid_entry(struct kref *kref)
262{
263 struct ib_gid_table_entry *entry =
264 container_of(kref, struct ib_gid_table_entry, kref);
265
266 free_gid_entry_locked(entry);
267}
268
269/**
270 * free_gid_work - Release reference to the GID entry
271 * @work: Work structure to refer to GID entry which needs to be
272 * deleted.
273 *
274 * free_gid_work() frees the entry from the HCA's hardware table
275 * if provider supports it. It releases reference to netdevice.
276 */
277static void free_gid_work(struct work_struct *work)
278{
279 struct ib_gid_table_entry *entry =
280 container_of(work, struct ib_gid_table_entry, del_work);
281 struct ib_device *device = entry->attr.device;
282 u8 port_num = entry->attr.port_num;
283 struct ib_gid_table *table = rdma_gid_table(device, port_num);
284
285 mutex_lock(&table->lock);
286 free_gid_entry_locked(entry);
287 mutex_unlock(&table->lock);
288}
289
290static struct ib_gid_table_entry *
291alloc_gid_entry(const struct ib_gid_attr *attr)
292{
293 struct ib_gid_table_entry *entry;
294 struct net_device *ndev;
295
296 entry = kzalloc(sizeof(*entry), GFP_KERNEL);
297 if (!entry)
298 return NULL;
299
300 ndev = rcu_dereference_protected(attr->ndev, 1);
301 if (ndev) {
302 entry->ndev_storage = kzalloc(sizeof(*entry->ndev_storage),
303 GFP_KERNEL);
304 if (!entry->ndev_storage) {
305 kfree(entry);
306 return NULL;
307 }
308 dev_hold(ndev);
309 entry->ndev_storage->ndev = ndev;
310 }
311 kref_init(&entry->kref);
312 memcpy(&entry->attr, attr, sizeof(*attr));
313 INIT_WORK(&entry->del_work, free_gid_work);
314 entry->state = GID_TABLE_ENTRY_INVALID;
315 return entry;
316}
317
318static void store_gid_entry(struct ib_gid_table *table,
319 struct ib_gid_table_entry *entry)
320{
321 entry->state = GID_TABLE_ENTRY_VALID;
322
323 dev_dbg(&entry->attr.device->dev, "%s port=%d index=%d gid %pI6\n",
324 __func__, entry->attr.port_num, entry->attr.index,
325 entry->attr.gid.raw);
326
327 lockdep_assert_held(&table->lock);
328 write_lock_irq(&table->rwlock);
329 table->data_vec[entry->attr.index] = entry;
330 write_unlock_irq(&table->rwlock);
331}
332
333static void get_gid_entry(struct ib_gid_table_entry *entry)
334{
335 kref_get(&entry->kref);
336}
337
338static void put_gid_entry(struct ib_gid_table_entry *entry)
339{
340 kref_put(&entry->kref, schedule_free_gid);
341}
342
343static void put_gid_entry_locked(struct ib_gid_table_entry *entry)
344{
345 kref_put(&entry->kref, free_gid_entry);
346}
347
348static int add_roce_gid(struct ib_gid_table_entry *entry)
349{
350 const struct ib_gid_attr *attr = &entry->attr;
351 int ret;
352
353 if (!attr->ndev) {
354 dev_err(&attr->device->dev, "%s NULL netdev port=%d index=%d\n",
355 __func__, attr->port_num, attr->index);
356 return -EINVAL;
357 }
358 if (rdma_cap_roce_gid_table(attr->device, attr->port_num)) {
359 ret = attr->device->ops.add_gid(attr, &entry->context);
360 if (ret) {
361 dev_err(&attr->device->dev,
362 "%s GID add failed port=%d index=%d\n",
363 __func__, attr->port_num, attr->index);
364 return ret;
365 }
366 }
367 return 0;
368}
369
370/**
371 * del_gid - Delete GID table entry
372 *
373 * @ib_dev: IB device whose GID entry to be deleted
374 * @port: Port number of the IB device
375 * @table: GID table of the IB device for a port
376 * @ix: GID entry index to delete
377 *
378 */
379static void del_gid(struct ib_device *ib_dev, u8 port,
380 struct ib_gid_table *table, int ix)
381{
382 struct roce_gid_ndev_storage *ndev_storage;
383 struct ib_gid_table_entry *entry;
384
385 lockdep_assert_held(&table->lock);
386
387 dev_dbg(&ib_dev->dev, "%s port=%d index=%d gid %pI6\n", __func__, port,
388 ix, table->data_vec[ix]->attr.gid.raw);
389
390 write_lock_irq(&table->rwlock);
391 entry = table->data_vec[ix];
392 entry->state = GID_TABLE_ENTRY_PENDING_DEL;
393 /*
394 * For non RoCE protocol, GID entry slot is ready to use.
395 */
396 if (!rdma_protocol_roce(ib_dev, port))
397 table->data_vec[ix] = NULL;
398 write_unlock_irq(&table->rwlock);
399
400 ndev_storage = entry->ndev_storage;
401 if (ndev_storage) {
402 entry->ndev_storage = NULL;
403 rcu_assign_pointer(entry->attr.ndev, NULL);
404 call_rcu(&ndev_storage->rcu_head, put_gid_ndev);
405 }
406
407 if (rdma_cap_roce_gid_table(ib_dev, port))
408 ib_dev->ops.del_gid(&entry->attr, &entry->context);
409
410 put_gid_entry_locked(entry);
411}
412
413/**
414 * add_modify_gid - Add or modify GID table entry
415 *
416 * @table: GID table in which GID to be added or modified
417 * @attr: Attributes of the GID
418 *
419 * Returns 0 on success or appropriate error code. It accepts zero
420 * GID addition for non RoCE ports for HCA's who report them as valid
421 * GID. However such zero GIDs are not added to the cache.
422 */
423static int add_modify_gid(struct ib_gid_table *table,
424 const struct ib_gid_attr *attr)
425{
426 struct ib_gid_table_entry *entry;
427 int ret = 0;
428
429 /*
430 * Invalidate any old entry in the table to make it safe to write to
431 * this index.
432 */
433 if (is_gid_entry_valid(table->data_vec[attr->index]))
434 del_gid(attr->device, attr->port_num, table, attr->index);
435
436 /*
437 * Some HCA's report multiple GID entries with only one valid GID, and
438 * leave other unused entries as the zero GID. Convert zero GIDs to
439 * empty table entries instead of storing them.
440 */
441 if (rdma_is_zero_gid(&attr->gid))
442 return 0;
443
444 entry = alloc_gid_entry(attr);
445 if (!entry)
446 return -ENOMEM;
447
448 if (rdma_protocol_roce(attr->device, attr->port_num)) {
449 ret = add_roce_gid(entry);
450 if (ret)
451 goto done;
452 }
453
454 store_gid_entry(table, entry);
455 return 0;
456
457done:
458 put_gid_entry(entry);
459 return ret;
460}
461
462/* rwlock should be read locked, or lock should be held */
463static int find_gid(struct ib_gid_table *table, const union ib_gid *gid,
464 const struct ib_gid_attr *val, bool default_gid,
465 unsigned long mask, int *pempty)
466{
467 int i = 0;
468 int found = -1;
469 int empty = pempty ? -1 : 0;
470
471 while (i < table->sz && (found < 0 || empty < 0)) {
472 struct ib_gid_table_entry *data = table->data_vec[i];
473 struct ib_gid_attr *attr;
474 int curr_index = i;
475
476 i++;
477
478 /* find_gid() is used during GID addition where it is expected
479 * to return a free entry slot which is not duplicate.
480 * Free entry slot is requested and returned if pempty is set,
481 * so lookup free slot only if requested.
482 */
483 if (pempty && empty < 0) {
484 if (is_gid_entry_free(data) &&
485 default_gid ==
486 is_gid_index_default(table, curr_index)) {
487 /*
488 * Found an invalid (free) entry; allocate it.
489 * If default GID is requested, then our
490 * found slot must be one of the DEFAULT
491 * reserved slots or we fail.
492 * This ensures that only DEFAULT reserved
493 * slots are used for default property GIDs.
494 */
495 empty = curr_index;
496 }
497 }
498
499 /*
500 * Additionally find_gid() is used to find valid entry during
501 * lookup operation; so ignore the entries which are marked as
502 * pending for removal and the entries which are marked as
503 * invalid.
504 */
505 if (!is_gid_entry_valid(data))
506 continue;
507
508 if (found >= 0)
509 continue;
510
511 attr = &data->attr;
512 if (mask & GID_ATTR_FIND_MASK_GID_TYPE &&
513 attr->gid_type != val->gid_type)
514 continue;
515
516 if (mask & GID_ATTR_FIND_MASK_GID &&
517 memcmp(gid, &data->attr.gid, sizeof(*gid)))
518 continue;
519
520 if (mask & GID_ATTR_FIND_MASK_NETDEV &&
521 attr->ndev != val->ndev)
522 continue;
523
524 if (mask & GID_ATTR_FIND_MASK_DEFAULT &&
525 is_gid_index_default(table, curr_index) != default_gid)
526 continue;
527
528 found = curr_index;
529 }
530
531 if (pempty)
532 *pempty = empty;
533
534 return found;
535}
536
537static void make_default_gid(struct net_device *dev, union ib_gid *gid)
538{
539 gid->global.subnet_prefix = cpu_to_be64(0xfe80000000000000LL);
540 addrconf_ifid_eui48(&gid->raw[8], dev);
541}
542
543static int __ib_cache_gid_add(struct ib_device *ib_dev, u8 port,
544 union ib_gid *gid, struct ib_gid_attr *attr,
545 unsigned long mask, bool default_gid)
546{
547 struct ib_gid_table *table;
548 int ret = 0;
549 int empty;
550 int ix;
551
552 /* Do not allow adding zero GID in support of
553 * IB spec version 1.3 section 4.1.1 point (6) and
554 * section 12.7.10 and section 12.7.20
555 */
556 if (rdma_is_zero_gid(gid))
557 return -EINVAL;
558
559 table = rdma_gid_table(ib_dev, port);
560
561 mutex_lock(&table->lock);
562
563 ix = find_gid(table, gid, attr, default_gid, mask, &empty);
564 if (ix >= 0)
565 goto out_unlock;
566
567 if (empty < 0) {
568 ret = -ENOSPC;
569 goto out_unlock;
570 }
571 attr->device = ib_dev;
572 attr->index = empty;
573 attr->port_num = port;
574 attr->gid = *gid;
575 ret = add_modify_gid(table, attr);
576 if (!ret)
577 dispatch_gid_change_event(ib_dev, port);
578
579out_unlock:
580 mutex_unlock(&table->lock);
581 if (ret)
582 pr_warn("%s: unable to add gid %pI6 error=%d\n",
583 __func__, gid->raw, ret);
584 return ret;
585}
586
587int ib_cache_gid_add(struct ib_device *ib_dev, u8 port,
588 union ib_gid *gid, struct ib_gid_attr *attr)
589{
590 unsigned long mask = GID_ATTR_FIND_MASK_GID |
591 GID_ATTR_FIND_MASK_GID_TYPE |
592 GID_ATTR_FIND_MASK_NETDEV;
593
594 return __ib_cache_gid_add(ib_dev, port, gid, attr, mask, false);
595}
596
597static int
598_ib_cache_gid_del(struct ib_device *ib_dev, u8 port,
599 union ib_gid *gid, struct ib_gid_attr *attr,
600 unsigned long mask, bool default_gid)
601{
602 struct ib_gid_table *table;
603 int ret = 0;
604 int ix;
605
606 table = rdma_gid_table(ib_dev, port);
607
608 mutex_lock(&table->lock);
609
610 ix = find_gid(table, gid, attr, default_gid, mask, NULL);
611 if (ix < 0) {
612 ret = -EINVAL;
613 goto out_unlock;
614 }
615
616 del_gid(ib_dev, port, table, ix);
617 dispatch_gid_change_event(ib_dev, port);
618
619out_unlock:
620 mutex_unlock(&table->lock);
621 if (ret)
622 pr_debug("%s: can't delete gid %pI6 error=%d\n",
623 __func__, gid->raw, ret);
624 return ret;
625}
626
627int ib_cache_gid_del(struct ib_device *ib_dev, u8 port,
628 union ib_gid *gid, struct ib_gid_attr *attr)
629{
630 unsigned long mask = GID_ATTR_FIND_MASK_GID |
631 GID_ATTR_FIND_MASK_GID_TYPE |
632 GID_ATTR_FIND_MASK_DEFAULT |
633 GID_ATTR_FIND_MASK_NETDEV;
634
635 return _ib_cache_gid_del(ib_dev, port, gid, attr, mask, false);
636}
637
638int ib_cache_gid_del_all_netdev_gids(struct ib_device *ib_dev, u8 port,
639 struct net_device *ndev)
640{
641 struct ib_gid_table *table;
642 int ix;
643 bool deleted = false;
644
645 table = rdma_gid_table(ib_dev, port);
646
647 mutex_lock(&table->lock);
648
649 for (ix = 0; ix < table->sz; ix++) {
650 if (is_gid_entry_valid(table->data_vec[ix]) &&
651 table->data_vec[ix]->attr.ndev == ndev) {
652 del_gid(ib_dev, port, table, ix);
653 deleted = true;
654 }
655 }
656
657 mutex_unlock(&table->lock);
658
659 if (deleted)
660 dispatch_gid_change_event(ib_dev, port);
661
662 return 0;
663}
664
665/**
666 * rdma_find_gid_by_port - Returns the GID entry attributes when it finds
667 * a valid GID entry for given search parameters. It searches for the specified
668 * GID value in the local software cache.
669 * @device: The device to query.
670 * @gid: The GID value to search for.
671 * @gid_type: The GID type to search for.
672 * @port_num: The port number of the device where the GID value should be
673 * searched.
674 * @ndev: In RoCE, the net device of the device. NULL means ignore.
675 *
676 * Returns sgid attributes if the GID is found with valid reference or
677 * returns ERR_PTR for the error.
678 * The caller must invoke rdma_put_gid_attr() to release the reference.
679 */
680const struct ib_gid_attr *
681rdma_find_gid_by_port(struct ib_device *ib_dev,
682 const union ib_gid *gid,
683 enum ib_gid_type gid_type,
684 u8 port, struct net_device *ndev)
685{
686 int local_index;
687 struct ib_gid_table *table;
688 unsigned long mask = GID_ATTR_FIND_MASK_GID |
689 GID_ATTR_FIND_MASK_GID_TYPE;
690 struct ib_gid_attr val = {.ndev = ndev, .gid_type = gid_type};
691 const struct ib_gid_attr *attr;
692 unsigned long flags;
693
694 if (!rdma_is_port_valid(ib_dev, port))
695 return ERR_PTR(-ENOENT);
696
697 table = rdma_gid_table(ib_dev, port);
698
699 if (ndev)
700 mask |= GID_ATTR_FIND_MASK_NETDEV;
701
702 read_lock_irqsave(&table->rwlock, flags);
703 local_index = find_gid(table, gid, &val, false, mask, NULL);
704 if (local_index >= 0) {
705 get_gid_entry(table->data_vec[local_index]);
706 attr = &table->data_vec[local_index]->attr;
707 read_unlock_irqrestore(&table->rwlock, flags);
708 return attr;
709 }
710
711 read_unlock_irqrestore(&table->rwlock, flags);
712 return ERR_PTR(-ENOENT);
713}
714EXPORT_SYMBOL(rdma_find_gid_by_port);
715
716/**
717 * rdma_find_gid_by_filter - Returns the GID table attribute where a
718 * specified GID value occurs
719 * @device: The device to query.
720 * @gid: The GID value to search for.
721 * @port: The port number of the device where the GID value could be
722 * searched.
723 * @filter: The filter function is executed on any matching GID in the table.
724 * If the filter function returns true, the corresponding index is returned,
725 * otherwise, we continue searching the GID table. It's guaranteed that
726 * while filter is executed, ndev field is valid and the structure won't
727 * change. filter is executed in an atomic context. filter must not be NULL.
728 *
729 * rdma_find_gid_by_filter() searches for the specified GID value
730 * of which the filter function returns true in the port's GID table.
731 *
732 */
733const struct ib_gid_attr *rdma_find_gid_by_filter(
734 struct ib_device *ib_dev, const union ib_gid *gid, u8 port,
735 bool (*filter)(const union ib_gid *gid, const struct ib_gid_attr *,
736 void *),
737 void *context)
738{
739 const struct ib_gid_attr *res = ERR_PTR(-ENOENT);
740 struct ib_gid_table *table;
741 unsigned long flags;
742 unsigned int i;
743
744 if (!rdma_is_port_valid(ib_dev, port))
745 return ERR_PTR(-EINVAL);
746
747 table = rdma_gid_table(ib_dev, port);
748
749 read_lock_irqsave(&table->rwlock, flags);
750 for (i = 0; i < table->sz; i++) {
751 struct ib_gid_table_entry *entry = table->data_vec[i];
752
753 if (!is_gid_entry_valid(entry))
754 continue;
755
756 if (memcmp(gid, &entry->attr.gid, sizeof(*gid)))
757 continue;
758
759 if (filter(gid, &entry->attr, context)) {
760 get_gid_entry(entry);
761 res = &entry->attr;
762 break;
763 }
764 }
765 read_unlock_irqrestore(&table->rwlock, flags);
766 return res;
767}
768
769static struct ib_gid_table *alloc_gid_table(int sz)
770{
771 struct ib_gid_table *table = kzalloc(sizeof(*table), GFP_KERNEL);
772
773 if (!table)
774 return NULL;
775
776 table->data_vec = kcalloc(sz, sizeof(*table->data_vec), GFP_KERNEL);
777 if (!table->data_vec)
778 goto err_free_table;
779
780 mutex_init(&table->lock);
781
782 table->sz = sz;
783 rwlock_init(&table->rwlock);
784 return table;
785
786err_free_table:
787 kfree(table);
788 return NULL;
789}
790
791static void release_gid_table(struct ib_device *device,
792 struct ib_gid_table *table)
793{
794 bool leak = false;
795 int i;
796
797 if (!table)
798 return;
799
800 for (i = 0; i < table->sz; i++) {
801 if (is_gid_entry_free(table->data_vec[i]))
802 continue;
803 if (kref_read(&table->data_vec[i]->kref) > 1) {
804 dev_err(&device->dev,
805 "GID entry ref leak for index %d ref=%d\n", i,
806 kref_read(&table->data_vec[i]->kref));
807 leak = true;
808 }
809 }
810 if (leak)
811 return;
812
813 mutex_destroy(&table->lock);
814 kfree(table->data_vec);
815 kfree(table);
816}
817
818static void cleanup_gid_table_port(struct ib_device *ib_dev, u8 port,
819 struct ib_gid_table *table)
820{
821 int i;
822 bool deleted = false;
823
824 if (!table)
825 return;
826
827 mutex_lock(&table->lock);
828 for (i = 0; i < table->sz; ++i) {
829 if (is_gid_entry_valid(table->data_vec[i])) {
830 del_gid(ib_dev, port, table, i);
831 deleted = true;
832 }
833 }
834 mutex_unlock(&table->lock);
835
836 if (deleted)
837 dispatch_gid_change_event(ib_dev, port);
838}
839
840void ib_cache_gid_set_default_gid(struct ib_device *ib_dev, u8 port,
841 struct net_device *ndev,
842 unsigned long gid_type_mask,
843 enum ib_cache_gid_default_mode mode)
844{
845 union ib_gid gid = { };
846 struct ib_gid_attr gid_attr;
847 unsigned int gid_type;
848 unsigned long mask;
849
850 mask = GID_ATTR_FIND_MASK_GID_TYPE |
851 GID_ATTR_FIND_MASK_DEFAULT |
852 GID_ATTR_FIND_MASK_NETDEV;
853 memset(&gid_attr, 0, sizeof(gid_attr));
854 gid_attr.ndev = ndev;
855
856 for (gid_type = 0; gid_type < IB_GID_TYPE_SIZE; ++gid_type) {
857 if (1UL << gid_type & ~gid_type_mask)
858 continue;
859
860 gid_attr.gid_type = gid_type;
861
862 if (mode == IB_CACHE_GID_DEFAULT_MODE_SET) {
863 make_default_gid(ndev, &gid);
864 __ib_cache_gid_add(ib_dev, port, &gid,
865 &gid_attr, mask, true);
866 } else if (mode == IB_CACHE_GID_DEFAULT_MODE_DELETE) {
867 _ib_cache_gid_del(ib_dev, port, &gid,
868 &gid_attr, mask, true);
869 }
870 }
871}
872
873static void gid_table_reserve_default(struct ib_device *ib_dev, u8 port,
874 struct ib_gid_table *table)
875{
876 unsigned int i;
877 unsigned long roce_gid_type_mask;
878 unsigned int num_default_gids;
879
880 roce_gid_type_mask = roce_gid_type_mask_support(ib_dev, port);
881 num_default_gids = hweight_long(roce_gid_type_mask);
882 /* Reserve starting indices for default GIDs */
883 for (i = 0; i < num_default_gids && i < table->sz; i++)
884 table->default_gid_indices |= BIT(i);
885}
886
887
888static void gid_table_release_one(struct ib_device *ib_dev)
889{
890 unsigned int p;
891
892 rdma_for_each_port (ib_dev, p) {
893 release_gid_table(ib_dev, ib_dev->port_data[p].cache.gid);
894 ib_dev->port_data[p].cache.gid = NULL;
895 }
896}
897
898static int _gid_table_setup_one(struct ib_device *ib_dev)
899{
900 struct ib_gid_table *table;
901 unsigned int rdma_port;
902
903 rdma_for_each_port (ib_dev, rdma_port) {
904 table = alloc_gid_table(
905 ib_dev->port_data[rdma_port].immutable.gid_tbl_len);
906 if (!table)
907 goto rollback_table_setup;
908
909 gid_table_reserve_default(ib_dev, rdma_port, table);
910 ib_dev->port_data[rdma_port].cache.gid = table;
911 }
912 return 0;
913
914rollback_table_setup:
915 gid_table_release_one(ib_dev);
916 return -ENOMEM;
917}
918
919static void gid_table_cleanup_one(struct ib_device *ib_dev)
920{
921 unsigned int p;
922
923 rdma_for_each_port (ib_dev, p)
924 cleanup_gid_table_port(ib_dev, p,
925 ib_dev->port_data[p].cache.gid);
926}
927
928static int gid_table_setup_one(struct ib_device *ib_dev)
929{
930 int err;
931
932 err = _gid_table_setup_one(ib_dev);
933
934 if (err)
935 return err;
936
937 rdma_roce_rescan_device(ib_dev);
938
939 return err;
940}
941
942/**
943 * rdma_query_gid - Read the GID content from the GID software cache
944 * @device: Device to query the GID
945 * @port_num: Port number of the device
946 * @index: Index of the GID table entry to read
947 * @gid: Pointer to GID where to store the entry's GID
948 *
949 * rdma_query_gid() only reads the GID entry content for requested device,
950 * port and index. It reads for IB, RoCE and iWarp link layers. It doesn't
951 * hold any reference to the GID table entry in the HCA or software cache.
952 *
953 * Returns 0 on success or appropriate error code.
954 *
955 */
956int rdma_query_gid(struct ib_device *device, u8 port_num,
957 int index, union ib_gid *gid)
958{
959 struct ib_gid_table *table;
960 unsigned long flags;
961 int res = -EINVAL;
962
963 if (!rdma_is_port_valid(device, port_num))
964 return -EINVAL;
965
966 table = rdma_gid_table(device, port_num);
967 read_lock_irqsave(&table->rwlock, flags);
968
969 if (index < 0 || index >= table->sz ||
970 !is_gid_entry_valid(table->data_vec[index]))
971 goto done;
972
973 memcpy(gid, &table->data_vec[index]->attr.gid, sizeof(*gid));
974 res = 0;
975
976done:
977 read_unlock_irqrestore(&table->rwlock, flags);
978 return res;
979}
980EXPORT_SYMBOL(rdma_query_gid);
981
982/**
983 * rdma_find_gid - Returns SGID attributes if the matching GID is found.
984 * @device: The device to query.
985 * @gid: The GID value to search for.
986 * @gid_type: The GID type to search for.
987 * @ndev: In RoCE, the net device of the device. NULL means ignore.
988 *
989 * rdma_find_gid() searches for the specified GID value in the software cache.
990 *
991 * Returns GID attributes if a valid GID is found or returns ERR_PTR for the
992 * error. The caller must invoke rdma_put_gid_attr() to release the reference.
993 *
994 */
995const struct ib_gid_attr *rdma_find_gid(struct ib_device *device,
996 const union ib_gid *gid,
997 enum ib_gid_type gid_type,
998 struct net_device *ndev)
999{
1000 unsigned long mask = GID_ATTR_FIND_MASK_GID |
1001 GID_ATTR_FIND_MASK_GID_TYPE;
1002 struct ib_gid_attr gid_attr_val = {.ndev = ndev, .gid_type = gid_type};
1003 unsigned int p;
1004
1005 if (ndev)
1006 mask |= GID_ATTR_FIND_MASK_NETDEV;
1007
1008 rdma_for_each_port(device, p) {
1009 struct ib_gid_table *table;
1010 unsigned long flags;
1011 int index;
1012
1013 table = device->port_data[p].cache.gid;
1014 read_lock_irqsave(&table->rwlock, flags);
1015 index = find_gid(table, gid, &gid_attr_val, false, mask, NULL);
1016 if (index >= 0) {
1017 const struct ib_gid_attr *attr;
1018
1019 get_gid_entry(table->data_vec[index]);
1020 attr = &table->data_vec[index]->attr;
1021 read_unlock_irqrestore(&table->rwlock, flags);
1022 return attr;
1023 }
1024 read_unlock_irqrestore(&table->rwlock, flags);
1025 }
1026
1027 return ERR_PTR(-ENOENT);
1028}
1029EXPORT_SYMBOL(rdma_find_gid);
1030
1031int ib_get_cached_pkey(struct ib_device *device,
1032 u8 port_num,
1033 int index,
1034 u16 *pkey)
1035{
1036 struct ib_pkey_cache *cache;
1037 unsigned long flags;
1038 int ret = 0;
1039
1040 if (!rdma_is_port_valid(device, port_num))
1041 return -EINVAL;
1042
1043 read_lock_irqsave(&device->cache.lock, flags);
1044
1045 cache = device->port_data[port_num].cache.pkey;
1046
1047 if (index < 0 || index >= cache->table_len)
1048 ret = -EINVAL;
1049 else
1050 *pkey = cache->table[index];
1051
1052 read_unlock_irqrestore(&device->cache.lock, flags);
1053
1054 return ret;
1055}
1056EXPORT_SYMBOL(ib_get_cached_pkey);
1057
1058int ib_get_cached_subnet_prefix(struct ib_device *device,
1059 u8 port_num,
1060 u64 *sn_pfx)
1061{
1062 unsigned long flags;
1063
1064 if (!rdma_is_port_valid(device, port_num))
1065 return -EINVAL;
1066
1067 read_lock_irqsave(&device->cache.lock, flags);
1068 *sn_pfx = device->port_data[port_num].cache.subnet_prefix;
1069 read_unlock_irqrestore(&device->cache.lock, flags);
1070
1071 return 0;
1072}
1073EXPORT_SYMBOL(ib_get_cached_subnet_prefix);
1074
1075int ib_find_cached_pkey(struct ib_device *device,
1076 u8 port_num,
1077 u16 pkey,
1078 u16 *index)
1079{
1080 struct ib_pkey_cache *cache;
1081 unsigned long flags;
1082 int i;
1083 int ret = -ENOENT;
1084 int partial_ix = -1;
1085
1086 if (!rdma_is_port_valid(device, port_num))
1087 return -EINVAL;
1088
1089 read_lock_irqsave(&device->cache.lock, flags);
1090
1091 cache = device->port_data[port_num].cache.pkey;
1092
1093 *index = -1;
1094
1095 for (i = 0; i < cache->table_len; ++i)
1096 if ((cache->table[i] & 0x7fff) == (pkey & 0x7fff)) {
1097 if (cache->table[i] & 0x8000) {
1098 *index = i;
1099 ret = 0;
1100 break;
1101 } else
1102 partial_ix = i;
1103 }
1104
1105 if (ret && partial_ix >= 0) {
1106 *index = partial_ix;
1107 ret = 0;
1108 }
1109
1110 read_unlock_irqrestore(&device->cache.lock, flags);
1111
1112 return ret;
1113}
1114EXPORT_SYMBOL(ib_find_cached_pkey);
1115
1116int ib_find_exact_cached_pkey(struct ib_device *device,
1117 u8 port_num,
1118 u16 pkey,
1119 u16 *index)
1120{
1121 struct ib_pkey_cache *cache;
1122 unsigned long flags;
1123 int i;
1124 int ret = -ENOENT;
1125
1126 if (!rdma_is_port_valid(device, port_num))
1127 return -EINVAL;
1128
1129 read_lock_irqsave(&device->cache.lock, flags);
1130
1131 cache = device->port_data[port_num].cache.pkey;
1132
1133 *index = -1;
1134
1135 for (i = 0; i < cache->table_len; ++i)
1136 if (cache->table[i] == pkey) {
1137 *index = i;
1138 ret = 0;
1139 break;
1140 }
1141
1142 read_unlock_irqrestore(&device->cache.lock, flags);
1143
1144 return ret;
1145}
1146EXPORT_SYMBOL(ib_find_exact_cached_pkey);
1147
1148int ib_get_cached_lmc(struct ib_device *device,
1149 u8 port_num,
1150 u8 *lmc)
1151{
1152 unsigned long flags;
1153 int ret = 0;
1154
1155 if (!rdma_is_port_valid(device, port_num))
1156 return -EINVAL;
1157
1158 read_lock_irqsave(&device->cache.lock, flags);
1159 *lmc = device->port_data[port_num].cache.lmc;
1160 read_unlock_irqrestore(&device->cache.lock, flags);
1161
1162 return ret;
1163}
1164EXPORT_SYMBOL(ib_get_cached_lmc);
1165
1166int ib_get_cached_port_state(struct ib_device *device,
1167 u8 port_num,
1168 enum ib_port_state *port_state)
1169{
1170 unsigned long flags;
1171 int ret = 0;
1172
1173 if (!rdma_is_port_valid(device, port_num))
1174 return -EINVAL;
1175
1176 read_lock_irqsave(&device->cache.lock, flags);
1177 *port_state = device->port_data[port_num].cache.port_state;
1178 read_unlock_irqrestore(&device->cache.lock, flags);
1179
1180 return ret;
1181}
1182EXPORT_SYMBOL(ib_get_cached_port_state);
1183
1184/**
1185 * rdma_get_gid_attr - Returns GID attributes for a port of a device
1186 * at a requested gid_index, if a valid GID entry exists.
1187 * @device: The device to query.
1188 * @port_num: The port number on the device where the GID value
1189 * is to be queried.
1190 * @index: Index of the GID table entry whose attributes are to
1191 * be queried.
1192 *
1193 * rdma_get_gid_attr() acquires reference count of gid attributes from the
1194 * cached GID table. Caller must invoke rdma_put_gid_attr() to release
1195 * reference to gid attribute regardless of link layer.
1196 *
1197 * Returns pointer to valid gid attribute or ERR_PTR for the appropriate error
1198 * code.
1199 */
1200const struct ib_gid_attr *
1201rdma_get_gid_attr(struct ib_device *device, u8 port_num, int index)
1202{
1203 const struct ib_gid_attr *attr = ERR_PTR(-EINVAL);
1204 struct ib_gid_table *table;
1205 unsigned long flags;
1206
1207 if (!rdma_is_port_valid(device, port_num))
1208 return ERR_PTR(-EINVAL);
1209
1210 table = rdma_gid_table(device, port_num);
1211 if (index < 0 || index >= table->sz)
1212 return ERR_PTR(-EINVAL);
1213
1214 read_lock_irqsave(&table->rwlock, flags);
1215 if (!is_gid_entry_valid(table->data_vec[index]))
1216 goto done;
1217
1218 get_gid_entry(table->data_vec[index]);
1219 attr = &table->data_vec[index]->attr;
1220done:
1221 read_unlock_irqrestore(&table->rwlock, flags);
1222 return attr;
1223}
1224EXPORT_SYMBOL(rdma_get_gid_attr);
1225
1226/**
1227 * rdma_put_gid_attr - Release reference to the GID attribute
1228 * @attr: Pointer to the GID attribute whose reference
1229 * needs to be released.
1230 *
1231 * rdma_put_gid_attr() must be used to release reference whose
1232 * reference is acquired using rdma_get_gid_attr() or any APIs
1233 * which returns a pointer to the ib_gid_attr regardless of link layer
1234 * of IB or RoCE.
1235 *
1236 */
1237void rdma_put_gid_attr(const struct ib_gid_attr *attr)
1238{
1239 struct ib_gid_table_entry *entry =
1240 container_of(attr, struct ib_gid_table_entry, attr);
1241
1242 put_gid_entry(entry);
1243}
1244EXPORT_SYMBOL(rdma_put_gid_attr);
1245
1246/**
1247 * rdma_hold_gid_attr - Get reference to existing GID attribute
1248 *
1249 * @attr: Pointer to the GID attribute whose reference
1250 * needs to be taken.
1251 *
1252 * Increase the reference count to a GID attribute to keep it from being
1253 * freed. Callers are required to already be holding a reference to attribute.
1254 *
1255 */
1256void rdma_hold_gid_attr(const struct ib_gid_attr *attr)
1257{
1258 struct ib_gid_table_entry *entry =
1259 container_of(attr, struct ib_gid_table_entry, attr);
1260
1261 get_gid_entry(entry);
1262}
1263EXPORT_SYMBOL(rdma_hold_gid_attr);
1264
1265/**
1266 * rdma_read_gid_attr_ndev_rcu - Read GID attribute netdevice
1267 * which must be in UP state.
1268 *
1269 * @attr:Pointer to the GID attribute
1270 *
1271 * Returns pointer to netdevice if the netdevice was attached to GID and
1272 * netdevice is in UP state. Caller must hold RCU lock as this API
1273 * reads the netdev flags which can change while netdevice migrates to
1274 * different net namespace. Returns ERR_PTR with error code otherwise.
1275 *
1276 */
1277struct net_device *rdma_read_gid_attr_ndev_rcu(const struct ib_gid_attr *attr)
1278{
1279 struct ib_gid_table_entry *entry =
1280 container_of(attr, struct ib_gid_table_entry, attr);
1281 struct ib_device *device = entry->attr.device;
1282 struct net_device *ndev = ERR_PTR(-ENODEV);
1283 u8 port_num = entry->attr.port_num;
1284 struct ib_gid_table *table;
1285 unsigned long flags;
1286 bool valid;
1287
1288 table = rdma_gid_table(device, port_num);
1289
1290 read_lock_irqsave(&table->rwlock, flags);
1291 valid = is_gid_entry_valid(table->data_vec[attr->index]);
1292 if (valid) {
1293 ndev = rcu_dereference(attr->ndev);
1294 if (!ndev ||
1295 (ndev && ((READ_ONCE(ndev->flags) & IFF_UP) == 0)))
1296 ndev = ERR_PTR(-ENODEV);
1297 }
1298 read_unlock_irqrestore(&table->rwlock, flags);
1299 return ndev;
1300}
1301EXPORT_SYMBOL(rdma_read_gid_attr_ndev_rcu);
1302
1303static int get_lower_dev_vlan(struct net_device *lower_dev, void *data)
1304{
1305 u16 *vlan_id = data;
1306
1307 if (is_vlan_dev(lower_dev))
1308 *vlan_id = vlan_dev_vlan_id(lower_dev);
1309
1310 /* We are interested only in first level vlan device, so
1311 * always return 1 to stop iterating over next level devices.
1312 */
1313 return 1;
1314}
1315
1316/**
1317 * rdma_read_gid_l2_fields - Read the vlan ID and source MAC address
1318 * of a GID entry.
1319 *
1320 * @attr: GID attribute pointer whose L2 fields to be read
1321 * @vlan_id: Pointer to vlan id to fill up if the GID entry has
1322 * vlan id. It is optional.
1323 * @smac: Pointer to smac to fill up for a GID entry. It is optional.
1324 *
1325 * rdma_read_gid_l2_fields() returns 0 on success and returns vlan id
1326 * (if gid entry has vlan) and source MAC, or returns error.
1327 */
1328int rdma_read_gid_l2_fields(const struct ib_gid_attr *attr,
1329 u16 *vlan_id, u8 *smac)
1330{
1331 struct net_device *ndev;
1332
1333 rcu_read_lock();
1334 ndev = rcu_dereference(attr->ndev);
1335 if (!ndev) {
1336 rcu_read_unlock();
1337 return -ENODEV;
1338 }
1339 if (smac)
1340 ether_addr_copy(smac, ndev->dev_addr);
1341 if (vlan_id) {
1342 *vlan_id = 0xffff;
1343 if (is_vlan_dev(ndev)) {
1344 *vlan_id = vlan_dev_vlan_id(ndev);
1345 } else {
1346 /* If the netdev is upper device and if it's lower
1347 * device is vlan device, consider vlan id of the
1348 * the lower vlan device for this gid entry.
1349 */
1350 netdev_walk_all_lower_dev_rcu(attr->ndev,
1351 get_lower_dev_vlan, vlan_id);
1352 }
1353 }
1354 rcu_read_unlock();
1355 return 0;
1356}
1357EXPORT_SYMBOL(rdma_read_gid_l2_fields);
1358
1359static int config_non_roce_gid_cache(struct ib_device *device,
1360 u8 port, int gid_tbl_len)
1361{
1362 struct ib_gid_attr gid_attr = {};
1363 struct ib_gid_table *table;
1364 int ret = 0;
1365 int i;
1366
1367 gid_attr.device = device;
1368 gid_attr.port_num = port;
1369 table = rdma_gid_table(device, port);
1370
1371 mutex_lock(&table->lock);
1372 for (i = 0; i < gid_tbl_len; ++i) {
1373 if (!device->ops.query_gid)
1374 continue;
1375 ret = device->ops.query_gid(device, port, i, &gid_attr.gid);
1376 if (ret) {
1377 dev_warn(&device->dev,
1378 "query_gid failed (%d) for index %d\n", ret,
1379 i);
1380 goto err;
1381 }
1382 gid_attr.index = i;
1383 add_modify_gid(table, &gid_attr);
1384 }
1385err:
1386 mutex_unlock(&table->lock);
1387 return ret;
1388}
1389
1390static void ib_cache_update(struct ib_device *device,
1391 u8 port,
1392 bool enforce_security)
1393{
1394 struct ib_port_attr *tprops = NULL;
1395 struct ib_pkey_cache *pkey_cache = NULL, *old_pkey_cache;
1396 int i;
1397 int ret;
1398
1399 if (!rdma_is_port_valid(device, port))
1400 return;
1401
1402 tprops = kmalloc(sizeof *tprops, GFP_KERNEL);
1403 if (!tprops)
1404 return;
1405
1406 ret = ib_query_port(device, port, tprops);
1407 if (ret) {
1408 dev_warn(&device->dev, "ib_query_port failed (%d)\n", ret);
1409 goto err;
1410 }
1411
1412 if (!rdma_protocol_roce(device, port)) {
1413 ret = config_non_roce_gid_cache(device, port,
1414 tprops->gid_tbl_len);
1415 if (ret)
1416 goto err;
1417 }
1418
1419 pkey_cache = kmalloc(struct_size(pkey_cache, table,
1420 tprops->pkey_tbl_len),
1421 GFP_KERNEL);
1422 if (!pkey_cache)
1423 goto err;
1424
1425 pkey_cache->table_len = tprops->pkey_tbl_len;
1426
1427 for (i = 0; i < pkey_cache->table_len; ++i) {
1428 ret = ib_query_pkey(device, port, i, pkey_cache->table + i);
1429 if (ret) {
1430 dev_warn(&device->dev,
1431 "ib_query_pkey failed (%d) for index %d\n",
1432 ret, i);
1433 goto err;
1434 }
1435 }
1436
1437 write_lock_irq(&device->cache.lock);
1438
1439 old_pkey_cache = device->port_data[port].cache.pkey;
1440
1441 device->port_data[port].cache.pkey = pkey_cache;
1442 device->port_data[port].cache.lmc = tprops->lmc;
1443 device->port_data[port].cache.port_state = tprops->state;
1444
1445 device->port_data[port].cache.subnet_prefix = tprops->subnet_prefix;
1446 write_unlock_irq(&device->cache.lock);
1447
1448 if (enforce_security)
1449 ib_security_cache_change(device,
1450 port,
1451 tprops->subnet_prefix);
1452
1453 kfree(old_pkey_cache);
1454 kfree(tprops);
1455 return;
1456
1457err:
1458 kfree(pkey_cache);
1459 kfree(tprops);
1460}
1461
1462static void ib_cache_task(struct work_struct *_work)
1463{
1464 struct ib_update_work *work =
1465 container_of(_work, struct ib_update_work, work);
1466
1467 ib_cache_update(work->device,
1468 work->port_num,
1469 work->enforce_security);
1470 kfree(work);
1471}
1472
1473static void ib_cache_event(struct ib_event_handler *handler,
1474 struct ib_event *event)
1475{
1476 struct ib_update_work *work;
1477
1478 if (event->event == IB_EVENT_PORT_ERR ||
1479 event->event == IB_EVENT_PORT_ACTIVE ||
1480 event->event == IB_EVENT_LID_CHANGE ||
1481 event->event == IB_EVENT_PKEY_CHANGE ||
1482 event->event == IB_EVENT_CLIENT_REREGISTER ||
1483 event->event == IB_EVENT_GID_CHANGE) {
1484 work = kmalloc(sizeof *work, GFP_ATOMIC);
1485 if (work) {
1486 INIT_WORK(&work->work, ib_cache_task);
1487 work->device = event->device;
1488 work->port_num = event->element.port_num;
1489 if (event->event == IB_EVENT_PKEY_CHANGE ||
1490 event->event == IB_EVENT_GID_CHANGE)
1491 work->enforce_security = true;
1492 else
1493 work->enforce_security = false;
1494
1495 queue_work(ib_wq, &work->work);
1496 }
1497 }
1498}
1499
1500int ib_cache_setup_one(struct ib_device *device)
1501{
1502 unsigned int p;
1503 int err;
1504
1505 rwlock_init(&device->cache.lock);
1506
1507 err = gid_table_setup_one(device);
1508 if (err)
1509 return err;
1510
1511 rdma_for_each_port (device, p)
1512 ib_cache_update(device, p, true);
1513
1514 INIT_IB_EVENT_HANDLER(&device->cache.event_handler,
1515 device, ib_cache_event);
1516 ib_register_event_handler(&device->cache.event_handler);
1517 return 0;
1518}
1519
1520void ib_cache_release_one(struct ib_device *device)
1521{
1522 unsigned int p;
1523
1524 /*
1525 * The release function frees all the cache elements.
1526 * This function should be called as part of freeing
1527 * all the device's resources when the cache could no
1528 * longer be accessed.
1529 */
1530 rdma_for_each_port (device, p)
1531 kfree(device->port_data[p].cache.pkey);
1532
1533 gid_table_release_one(device);
1534}
1535
1536void ib_cache_cleanup_one(struct ib_device *device)
1537{
1538 /* The cleanup function unregisters the event handler,
1539 * waits for all in-progress workqueue elements and cleans
1540 * up the GID cache. This function should be called after
1541 * the device was removed from the devices list and all
1542 * clients were removed, so the cache exists but is
1543 * non-functional and shouldn't be updated anymore.
1544 */
1545 ib_unregister_event_handler(&device->cache.event_handler);
1546 flush_workqueue(ib_wq);
1547 gid_table_cleanup_one(device);
1548
1549 /*
1550 * Flush the wq second time for any pending GID delete work.
1551 */
1552 flush_workqueue(ib_wq);
1553}