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
|
/*
* Copyright (c) 2016 Oracle. All rights reserved.
*
* Use the core R/W API to move RPC-over-RDMA Read and Write chunks.
*/
#include <linux/sunrpc/rpc_rdma.h>
#include <linux/sunrpc/svc_rdma.h>
#include <linux/sunrpc/debug.h>
#include <rdma/rw.h>
#define RPCDBG_FACILITY RPCDBG_SVCXPRT
/* Each R/W context contains state for one chain of RDMA Read or
* Write Work Requests.
*
* Each WR chain handles a single contiguous server-side buffer,
* because scatterlist entries after the first have to start on
* page alignment. xdr_buf iovecs cannot guarantee alignment.
*
* Each WR chain handles only one R_key. Each RPC-over-RDMA segment
* from a client may contain a unique R_key, so each WR chain moves
* up to one segment at a time.
*
* The scatterlist makes this data structure over 4KB in size. To
* make it less likely to fail, and to handle the allocation for
* smaller I/O requests without disabling bottom-halves, these
* contexts are created on demand, but cached and reused until the
* controlling svcxprt_rdma is destroyed.
*/
struct svc_rdma_rw_ctxt {
struct list_head rw_list;
struct rdma_rw_ctx rw_ctx;
int rw_nents;
struct sg_table rw_sg_table;
struct scatterlist rw_first_sgl[0];
};
static inline struct svc_rdma_rw_ctxt *
svc_rdma_next_ctxt(struct list_head *list)
{
return list_first_entry_or_null(list, struct svc_rdma_rw_ctxt,
rw_list);
}
static struct svc_rdma_rw_ctxt *
svc_rdma_get_rw_ctxt(struct svcxprt_rdma *rdma, unsigned int sges)
{
struct svc_rdma_rw_ctxt *ctxt;
spin_lock(&rdma->sc_rw_ctxt_lock);
ctxt = svc_rdma_next_ctxt(&rdma->sc_rw_ctxts);
if (ctxt) {
list_del(&ctxt->rw_list);
spin_unlock(&rdma->sc_rw_ctxt_lock);
} else {
spin_unlock(&rdma->sc_rw_ctxt_lock);
ctxt = kmalloc(sizeof(*ctxt) +
SG_CHUNK_SIZE * sizeof(struct scatterlist),
GFP_KERNEL);
if (!ctxt)
goto out;
INIT_LIST_HEAD(&ctxt->rw_list);
}
ctxt->rw_sg_table.sgl = ctxt->rw_first_sgl;
if (sg_alloc_table_chained(&ctxt->rw_sg_table, sges,
ctxt->rw_sg_table.sgl)) {
kfree(ctxt);
ctxt = NULL;
}
out:
return ctxt;
}
static void svc_rdma_put_rw_ctxt(struct svcxprt_rdma *rdma,
struct svc_rdma_rw_ctxt *ctxt)
{
sg_free_table_chained(&ctxt->rw_sg_table, true);
spin_lock(&rdma->sc_rw_ctxt_lock);
list_add(&ctxt->rw_list, &rdma->sc_rw_ctxts);
spin_unlock(&rdma->sc_rw_ctxt_lock);
}
/**
* svc_rdma_destroy_rw_ctxts - Free accumulated R/W contexts
* @rdma: transport about to be destroyed
*
*/
void svc_rdma_destroy_rw_ctxts(struct svcxprt_rdma *rdma)
{
struct svc_rdma_rw_ctxt *ctxt;
while ((ctxt = svc_rdma_next_ctxt(&rdma->sc_rw_ctxts)) != NULL) {
list_del(&ctxt->rw_list);
kfree(ctxt);
}
}
/* A chunk context tracks all I/O for moving one Read or Write
* chunk. This is a a set of rdma_rw's that handle data movement
* for all segments of one chunk.
*
* These are small, acquired with a single allocator call, and
* no more than one is needed per chunk. They are allocated on
* demand, and not cached.
*/
struct svc_rdma_chunk_ctxt {
struct ib_cqe cc_cqe;
struct svcxprt_rdma *cc_rdma;
struct list_head cc_rwctxts;
int cc_sqecount;
enum dma_data_direction cc_dir;
};
static void svc_rdma_cc_init(struct svcxprt_rdma *rdma,
struct svc_rdma_chunk_ctxt *cc,
enum dma_data_direction dir)
{
cc->cc_rdma = rdma;
svc_xprt_get(&rdma->sc_xprt);
INIT_LIST_HEAD(&cc->cc_rwctxts);
cc->cc_sqecount = 0;
cc->cc_dir = dir;
}
static void svc_rdma_cc_release(struct svc_rdma_chunk_ctxt *cc)
{
struct svcxprt_rdma *rdma = cc->cc_rdma;
struct svc_rdma_rw_ctxt *ctxt;
while ((ctxt = svc_rdma_next_ctxt(&cc->cc_rwctxts)) != NULL) {
list_del(&ctxt->rw_list);
rdma_rw_ctx_destroy(&ctxt->rw_ctx, rdma->sc_qp,
rdma->sc_port_num, ctxt->rw_sg_table.sgl,
ctxt->rw_nents, cc->cc_dir);
svc_rdma_put_rw_ctxt(rdma, ctxt);
}
svc_xprt_put(&rdma->sc_xprt);
}
/* State for sending a Write or Reply chunk.
* - Tracks progress of writing one chunk over all its segments
* - Stores arguments for the SGL constructor functions
*/
struct svc_rdma_write_info {
/* write state of this chunk */
unsigned int wi_seg_off;
unsigned int wi_seg_no;
unsigned int wi_nsegs;
__be32 *wi_segs;
/* SGL constructor arguments */
struct xdr_buf *wi_xdr;
unsigned char *wi_base;
unsigned int wi_next_off;
struct svc_rdma_chunk_ctxt wi_cc;
};
static struct svc_rdma_write_info *
svc_rdma_write_info_alloc(struct svcxprt_rdma *rdma, __be32 *chunk)
{
struct svc_rdma_write_info *info;
info = kmalloc(sizeof(*info), GFP_KERNEL);
if (!info)
return info;
info->wi_seg_off = 0;
info->wi_seg_no = 0;
info->wi_nsegs = be32_to_cpup(++chunk);
info->wi_segs = ++chunk;
svc_rdma_cc_init(rdma, &info->wi_cc, DMA_TO_DEVICE);
return info;
}
static void svc_rdma_write_info_free(struct svc_rdma_write_info *info)
{
svc_rdma_cc_release(&info->wi_cc);
kfree(info);
}
/**
* svc_rdma_write_done - Write chunk completion
* @cq: controlling Completion Queue
* @wc: Work Completion
*
* Pages under I/O are freed by a subsequent Send completion.
*/
static void svc_rdma_write_done(struct ib_cq *cq, struct ib_wc *wc)
{
struct ib_cqe *cqe = wc->wr_cqe;
struct svc_rdma_chunk_ctxt *cc =
container_of(cqe, struct svc_rdma_chunk_ctxt, cc_cqe);
struct svcxprt_rdma *rdma = cc->cc_rdma;
struct svc_rdma_write_info *info =
container_of(cc, struct svc_rdma_write_info, wi_cc);
atomic_add(cc->cc_sqecount, &rdma->sc_sq_avail);
wake_up(&rdma->sc_send_wait);
if (unlikely(wc->status != IB_WC_SUCCESS)) {
set_bit(XPT_CLOSE, &rdma->sc_xprt.xpt_flags);
if (wc->status != IB_WC_WR_FLUSH_ERR)
pr_err("svcrdma: write ctx: %s (%u/0x%x)\n",
ib_wc_status_msg(wc->status),
wc->status, wc->vendor_err);
}
svc_rdma_write_info_free(info);
}
/* This function sleeps when the transport's Send Queue is congested.
*
* Assumptions:
* - If ib_post_send() succeeds, only one completion is expected,
* even if one or more WRs are flushed. This is true when posting
* an rdma_rw_ctx or when posting a single signaled WR.
*/
static int svc_rdma_post_chunk_ctxt(struct svc_rdma_chunk_ctxt *cc)
{
struct svcxprt_rdma *rdma = cc->cc_rdma;
struct svc_xprt *xprt = &rdma->sc_xprt;
struct ib_send_wr *first_wr, *bad_wr;
struct list_head *tmp;
struct ib_cqe *cqe;
int ret;
if (cc->cc_sqecount > rdma->sc_sq_depth)
return -EINVAL;
first_wr = NULL;
cqe = &cc->cc_cqe;
list_for_each(tmp, &cc->cc_rwctxts) {
struct svc_rdma_rw_ctxt *ctxt;
ctxt = list_entry(tmp, struct svc_rdma_rw_ctxt, rw_list);
first_wr = rdma_rw_ctx_wrs(&ctxt->rw_ctx, rdma->sc_qp,
rdma->sc_port_num, cqe, first_wr);
cqe = NULL;
}
do {
if (atomic_sub_return(cc->cc_sqecount,
&rdma->sc_sq_avail) > 0) {
ret = ib_post_send(rdma->sc_qp, first_wr, &bad_wr);
if (ret)
break;
return 0;
}
atomic_inc(&rdma_stat_sq_starve);
atomic_add(cc->cc_sqecount, &rdma->sc_sq_avail);
wait_event(rdma->sc_send_wait,
atomic_read(&rdma->sc_sq_avail)
|