1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
|
// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (c) 2015 HGST, a Western Digital Company.
*/
#include <linux/module.h>
#include <linux/err.h>
#include <linux/slab.h>
#include <rdma/ib_verbs.h>
#include "core_priv.h"
#include <trace/events/rdma_core.h>
/* Max size for shared CQ, may require tuning */
#define IB_MAX_SHARED_CQ_SZ 4096U
/* # of WCs to poll for with a single call to ib_poll_cq */
#define IB_POLL_BATCH 16
#define IB_POLL_BATCH_DIRECT 8
/* # of WCs to iterate over before yielding */
#define IB_POLL_BUDGET_IRQ 256
#define IB_POLL_BUDGET_WORKQUEUE 65536
#define IB_POLL_FLAGS \
(IB_CQ_NEXT_COMP | IB_CQ_REPORT_MISSED_EVENTS)
static const struct dim_cq_moder
rdma_dim_prof[RDMA_DIM_PARAMS_NUM_PROFILES] = {
{1, 0, 1, 0},
{1, 0, 4, 0},
{2, 0, 4, 0},
{2, 0, 8, 0},
{4, 0, 8, 0},
{16, 0, 8, 0},
{16, 0, 16, 0},
{32, 0, 16, 0},
{32, 0, 32, 0},
};
static void ib_cq_rdma_dim_work(struct work_struct *w)
{
struct dim *dim = container_of(w, struct dim, work);
struct ib_cq *cq = dim->priv;
u16 usec = rdma_dim_prof[dim->profile_ix].usec;
u16 comps = rdma_dim_prof[dim->profile_ix].comps;
dim->state = DIM_START_MEASURE;
trace_cq_modify(cq, comps, usec);
cq->device->ops.modify_cq(cq, comps, usec);
}
static void rdma_dim_init(struct ib_cq *cq)
{
struct dim *dim;
if (!cq->device->ops.modify_cq || !cq->device->use_cq_dim ||
cq->poll_ctx == IB_POLL_DIRECT)
return;
dim = kzalloc(sizeof(struct dim), GFP_KERNEL);
if (!dim)
return;
dim->state = DIM_START_MEASURE;
dim->tune_state = DIM_GOING_RIGHT;
dim->profile_ix = RDMA_DIM_START_PROFILE;
dim->priv = cq;
cq->dim = dim;
INIT_WORK(&dim->work, ib_cq_rdma_dim_work);
}
static void rdma_dim_destroy(struct ib_cq *cq)
{
if (!cq->dim)
return;
cancel_work_sync(&cq->dim->work);
kfree(cq->dim);
}
static int __poll_cq(struct ib_cq *cq, int num_entries, struct ib_wc *wc)
{
int rc;
rc = ib_poll_cq(cq, num_entries, wc);
trace_cq_poll(cq, num_entries, rc);
return rc;
}
static int __ib_process_cq(struct ib_cq *cq, int budget, struct ib_wc *wcs,
int batch)
{
int i, n, completed = 0;
trace_cq_process(cq);
/*
* budget might be (-1) if the caller does not
* want to bound this call, thus we need unsigned
* minimum here.
*/
while ((n = __poll_cq(cq, min_t(u32, batch,
budget - completed), wcs)) > 0) {
for (i = 0; i < n; i++) {
struct ib_wc *wc = &wcs[i];
if (wc->wr_cqe)
wc->wr_cqe->done(cq, wc);
else
WARN_ON_ONCE(wc->status == IB_WC_SUCCESS);
}
completed += n;
if (n != batch || (budget != -1 && completed >= budget))
break;
}
return completed;
}
/**
* ib_process_direct_cq - process a CQ in caller context
* @cq: CQ to process
* @budget: number of CQEs to poll for
*
* This function is used to process all outstanding CQ entries.
* It does not offload CQ processing to a different context and does
* not ask for completion interrupts from the HCA.
* Using direct processing on CQ with non IB_POLL_DIRECT type may trigger
* concurrent processing.
*
* Note: do not pass -1 as %budget unless it is guaranteed that the number
* of completions that will be processed is small.
*/
int ib_process_cq_direct(struct ib_cq *cq, int budget)
{
struct ib_wc wcs[IB_POLL_BATCH_DIRECT];
return __ib_process_cq(cq, budget, wcs, IB_POLL_BATCH_DIRECT);
}
EXPORT_SYMBOL(ib_process_cq_direct);
static void ib_cq_completion_direct(struct ib_cq *cq, void *private)
{
WARN_ONCE(1, "got unsolicited completion for CQ 0x%p\n", cq);
}
static int ib_poll_handler(struct irq_poll *iop, int budget)
{
struct ib_cq *cq = container_of(iop, struct ib_cq, iop);
struct dim *dim = cq->dim;
int completed;
completed = __ib_process_cq(cq, budget, cq->wc, IB_POLL_BATCH);
if (completed < budget) {
irq_poll_complete(&cq->iop);
if (ib_req_notify_cq(cq, IB_POLL_FLAGS) > 0) {
trace_cq_reschedule(cq);
irq_poll_sched(&cq->iop);
}
}
if (dim)
rdma_dim(dim, completed);
return completed;
}
static void ib_cq_completion_softirq(struct ib_cq *cq, void *private)
{
trace_cq_schedule(cq);
irq_poll_sched(&cq->iop);
}
static void ib_cq_poll_work(struct work_struct *work)
{
struct ib_cq *cq = container_of(work, struct ib_cq, work);
int completed;
completed = __ib_process_cq(cq, IB_POLL_BUDGET_WORKQUEUE, cq->wc,
IB_POLL_BATCH);
if (completed >= IB_POLL_BUDGET_WORKQUEUE ||
ib_req_notify_cq(cq, IB_POLL_FLAGS) > 0)
queue_work(cq->comp_wq, &cq->work);
else if (cq->dim)
rdma_dim(cq->dim, completed);
}
static void ib_cq_completion_workqueue(struct ib_cq *cq, void *private)
{
trace_cq_schedule(cq);
queue_work(cq->comp_wq, &cq->work);
}
/**
* __ib_alloc_cq_user - allocate a completion queue
* @dev: device to allocate the CQ for
* @private: driver private data, accessible from cq->cq_context
* @nr_cqe: number of CQEs to allocate
* @comp_vector: HCA completion vectors for this CQ
* @poll_ctx: context to poll the CQ from.
* @caller: module owner name.
* @udata: Valid user data or NULL for kernel object
*
* This is the proper interface to allocate a CQ for in-kernel users. A
* CQ allocated with this interface will automatically be polled from the
* specified context. The ULP must use wr->wr_cqe instead of wr->wr_id
* to use this CQ abstraction.
*/
struct ib_cq *__ib_alloc_cq_user(struct ib_device *dev, void *private,
int nr_cqe, int comp_vector,
enum ib_poll_context poll_ctx,
const char *caller, struct ib_udata *udata)
{
struct ib_cq_init_attr cq_attr = {
.cqe = nr_cqe,
.comp_vector = comp_vector,
};
struct ib_cq *cq;
int ret = -ENOMEM;
cq = rdma_zalloc_drv_obj(dev, ib_cq);
if (!cq)
return ERR_PTR(ret);
cq->device = dev;
cq->cq_context = private;
cq->poll_ctx = poll_ctx;
atomic_set(&cq->usecnt, 0);
cq->comp_vector = comp_vector;
cq->wc = kmalloc_array(IB_POLL_BATCH, sizeof(*cq->wc), GFP_KERNEL);
if (!cq->wc)
goto out_free_cq;
cq->res.type = RDMA_RESTRACK_CQ;
rdma_restrack_set_task(&cq->res,
|