summaryrefslogtreecommitdiffstats
path: root/net/sunrpc/xprtrdma/rpc_rdma.c
blob: c10d9699441c3390d384e90f7dfc15ef9cd0ee27 (plain)
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
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
/*
 * Copyright (c) 2003-2007 Network Appliance, Inc. All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the BSD-type
 * license below:
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 *      Redistributions of source code must retain the above copyright
 *      notice, this list of conditions and the following disclaimer.
 *
 *      Redistributions in binary form must reproduce the above
 *      copyright notice, this list of conditions and the following
 *      disclaimer in the documentation and/or other materials provided
 *      with the distribution.
 *
 *      Neither the name of the Network Appliance, Inc. nor the names of
 *      its contributors may be used to endorse or promote products
 *      derived from this software without specific prior written
 *      permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
 * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/*
 * rpc_rdma.c
 *
 * This file contains the guts of the RPC RDMA protocol, and
 * does marshaling/unmarshaling, etc. It is also where interfacing
 * to the Linux RPC framework lives.
 */

#include "xprt_rdma.h"

#include <linux/highmem.h>

#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
# define RPCDBG_FACILITY	RPCDBG_TRANS
#endif

enum rpcrdma_chunktype {
	rpcrdma_noch = 0,
	rpcrdma_readch,
	rpcrdma_areadch,
	rpcrdma_writech,
	rpcrdma_replych
};

#if IS_ENABLED(CONFIG_SUNRPC_DEBUG)
static const char transfertypes[][12] = {
	"pure inline",	/* no chunks */
	" read chunk",	/* some argument via rdma read */
	"*read chunk",	/* entire request via rdma read */
	"write chunk",	/* some result via rdma write */
	"reply chunk"	/* entire reply via rdma write */
};
#endif

/* The client can send a request inline as long as the RPCRDMA header
 * plus the RPC call fit under the transport's inline limit. If the
 * combined call message size exceeds that limit, the client must use
 * the read chunk list for this operation.
 */
static bool rpcrdma_args_inline(struct rpc_rqst *rqst)
{
	unsigned int callsize = RPCRDMA_HDRLEN_MIN + rqst->rq_snd_buf.len;

	return callsize <= RPCRDMA_INLINE_WRITE_THRESHOLD(rqst);
}

/* The client can't know how large the actual reply will be. Thus it
 * plans for the largest possible reply for that particular ULP
 * operation. If the maximum combined reply message size exceeds that
 * limit, the client must provide a write list or a reply chunk for
 * this request.
 */
static bool rpcrdma_results_inline(struct rpc_rqst *rqst)
{
	unsigned int repsize = RPCRDMA_HDRLEN_MIN + rqst->rq_rcv_buf.buflen;

	return repsize <= RPCRDMA_INLINE_READ_THRESHOLD(rqst);
}

static int
rpcrdma_tail_pullup(struct xdr_buf *buf)
{
	size_t tlen = buf->tail[0].iov_len;
	size_t skip = tlen & 3;

	/* Do not include the tail if it is only an XDR pad */
	if (tlen < 4)
		return 0;

	/* xdr_write_pages() adds a pad at the beginning of the tail
	 * if the content in "buf->pages" is unaligned. Force the
	 * tail's actual content to land at the next XDR position
	 * after the head instead.
	 */
	if (skip) {
		unsigned char *src, *dst;
		unsigned int count;

		src = buf->tail[0].iov_base;
		dst = buf->head[0].iov_base;
		dst += buf->head[0].iov_len;

		src += skip;
		tlen -= skip;

		dprintk("RPC:       %s: skip=%zu, memmove(%p, %p, %zu)\n",
			__func__, skip, dst, src, tlen);

		for (count = tlen; count; count--)
			*dst++ = *src++;
	}

	return tlen;
}

/*
 * Chunk assembly from upper layer xdr_buf.
 *
 * Prepare the passed-in xdr_buf into representation as RPC/RDMA chunk
 * elements. Segments are then coalesced when registered, if possible
 * within the selected memreg mode.
 *
 * Returns positive number of segments converted, or a negative errno.
 */

static int
rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos,
	enum rpcrdma_chunktype type, struct rpcrdma_mr_seg *seg, int nsegs)
{
	int len, n = 0, p;
	int page_base;
	struct page **ppages;

	if (pos == 0 && xdrbuf->head[0].iov_len) {
		seg[n].mr_page = NULL;
		seg[n].mr_offset = xdrbuf->head[0].iov_base;
		seg[n].mr_len = xdrbuf->head[0].iov_len;
		++n;
	}

	len = xdrbuf->page_len;
	ppages = xdrbuf->pages + (xdrbuf->page_base >> PAGE_SHIFT);
	page_base = xdrbuf->page_base & ~PAGE_MASK;
	p = 0;
	while (len && n < nsegs) {
		if (!ppages[p]) {
			/* alloc the pagelist for receiving buffer */
			ppages[p] = alloc_page(GFP_ATOMIC);
			if (!ppages[p])
				return -ENOMEM;
		}
		seg[n].mr_page = ppages[p];
		seg[n].mr_offset = (void *)(unsigned long) page_base;
		seg[n].mr_len = min_t(u32, PAGE_SIZE - page_base, len);
		if (seg[n].mr_len > PAGE_SIZE)
			return -EIO;
		len -= seg[n].mr_len;
		++n;
		++p;
		page_base = 0;	/* page offset only applies to first page */
	}

	/* Message overflows the seg array */
	if (len && n == nsegs)
		return -EIO;

	/* When encoding the read list, the tail is always sent inline */
	if (type == rpcrdma_readch)
		return n;

	if (xdrbuf->tail[0].iov_len) {
		/* the rpcrdma protocol allows us to omit any trailing
		 * xdr pad bytes, saving the server an RDMA operation. */
		if (xdrbuf->tail[0].iov_len < 4 && xprt_rdma_pad_optimize)
			return n;
		if (n == nsegs)
			/* Tail remains, but we're out of segments */
			return -EIO;
		seg[n].mr_page = NULL;
		seg[n].mr_offset = xdrbuf->tail[0].iov_base;
		seg[n].mr_len = xdrbuf->tail[0].iov_len;
		++n;
	}

	return n;
}

/*
 * Create read/write chunk lists, and reply chunks, for RDMA
 *
 *   Assume check against THRESHOLD has been done, and chunks are required.
 *   Assume only encoding one list entry for read|write chunks. The NFSv3
 *     protocol is simple enough to allow this as it only has a single "bulk
 *     result" in each procedure - complicated NFSv4 COMPOUNDs are not. (The
 *     RDMA/Sessions NFSv4 proposal addresses this for future v4 revs.)
 *
 * When used for a single reply chunk (which is a special write
 * chunk used for the entire reply, rather than just the data), it
 * is used primarily for READDIR and READLINK which would otherwise
 * be severely size-limited by a small rdma inline read max. The server
 * response will come back as an RDMA Write, followed by a message
 * of type RDMA_NOMSG carrying the xid and length. As a result, reply
 * chunks do not provide data alignment, however they do not require
 * "fixup" (moving the response to the upper layer buffer) either.
 *
 * Encoding key for single-list chunks (HLOO = Handle32 Length32 Offset64):
 *
 *  Read chunklist (a linked list):
 *   N elements, position P (same P for all chunks of same arg!):
 *    1 - PHLOO - 1 - PHLOO - ... - 1 - PHLOO - 0
 *
 *  Write chunklist (a list of (one) counted array):
 *   N elements:
 *    1 - N - HLOO - HLOO - ... - HLOO - 0
 *
 *  Reply chunk (a counted array):
 *   N elements:
 *    1 - N - HLOO - HLOO - ... - HLOO
 *
 * Returns positive RPC/RDMA header size, or negative errno.
 */

static ssize_t
rpcrdma_create_chunks(struct rpc_rqst *rqst, struct xdr_buf *target,
		struct rpcrdma_msg *headerp, enum rpcrdma_chunktype type)
{
	struct rpcrdma_req *req = rpcr_to_rdmar(rqst);
	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(rqst->rq_xprt);
	int n, nsegs, nchunks = 0;
	unsigned int pos;
	struct rpcrdma_mr_seg *seg = req->rl_segments;
	struct rpcrdma_read_chunk *cur_rchunk = NULL;
	struct rpcrdma_write_array *warray = NULL;
	struct rpcrdma_write_chunk *cur_wchunk = NULL;
	__be32 *iptr = headerp->rm_body.rm_chunks;
	int (*map)(struct rpcrdma_xprt *, struct rpcrdma_mr_seg *, int, bool);

	if (type == rpcrdma_readch || type == rpcrdma_areadch) {
		/* a read chunk - server will RDMA Read our memory */
		cur_rchunk = (struct rpcrdma_read_chunk *) iptr;
	} else {
		/* a write or reply chunk - server will RDMA Write our memory */
		*iptr++ = xdr_zero;	/* encode a NULL read chunk list */
		if (type == rpcrdma_replych)
			*iptr++ = xdr_zero;	/* a NULL write chunk list */
		warray = (struct rpcrdma_write_array *) iptr;
		cur_wchunk = (struct rpcrdma_write_chunk *) (warray + 1);
	}

	if (type == rpcrdma_replych || type == rpcrdma_areadch)
		pos = 0;
	else
		pos = target->head[0].iov_len;

	nsegs = rpcrdma_convert_iovs(target, pos, type, seg, RPCRDMA_MAX_SEGS);
	if (nsegs < 0)
		return nsegs;

	map = r_xprt->rx_ia.ri_ops->ro_map;
	do {
		n = map(r_xprt, seg, nsegs, cur_wchunk != NULL);
		if (n <= 0)
			goto out;
		if (cur_rchunk) {	/* read */
			cur_rchunk->rc_discrim = xdr_one;
			/* all read chunks have the same "position" */
			cur_rchunk->rc_position = cpu_to_be32(pos);
			cur_rchunk->rc_target.rs_handle =
						cpu_to_be32(seg->mr_rkey);
			cur_rchunk->rc_target.rs_length =
						cpu_to_be32(seg->mr_len);
			xdr_encode_hyper(
					(__be32 *)&cur_rchunk->rc_target.rs_offset,
					seg->mr_base);
			dprintk("RPC:       %s: read chunk "
				"elem %d@0x%llx:0x%x pos %u (%s)\n", __func__,
				seg->mr_len, (unsigned long long)seg->mr_base,
				seg->mr_rkey, pos, n < nsegs ? "more" : "last");
			cur_rchunk++;
			r_xprt->rx_stats.read_chunk_count++;
		} else {		/* write/reply */
			cur_wchunk->wc_target.rs_handle =
						cpu_to_be32(seg->mr_rkey);
			cur_wchunk->wc_target.rs_length =
						cpu_to_be32(seg->mr_len);
			xdr_encode_hyper(
					(__be32 *)&cur_wchunk->wc_target.rs_offset,
					seg->mr_base);
			dprintk("RPC:       %s: %s chunk "
				"elem %d@0x%llx:0x%x (%s)\n", __func__,
				(type == rpcrdma_replych) ? "reply" : "write",
				seg->mr_len, (unsigned long long)seg->mr_base,
				seg->mr_rkey, n < nsegs ? "more" : "last");
			cur_wchunk++;
			if (type == rpcrdma_replych)
				r_xprt->rx_stats.reply_chunk_count++;
			else
				r_xprt->rx_stats.write_chunk_count++;
			r_xprt->rx_stats.total_rdma_request += seg->mr_len;
		}
		nchunks++;
		seg   += n;
		nsegs -= n;
	} while (nsegs);

	/* success. all failures return above */
	req->rl_nchunks = nchunks;

	/*
	 * finish off header. If write, marshal discrim and nchunks.
	 */
	if (cur_rchunk) {
		iptr = (__be32 *) cur_rchunk;
		*iptr++ = xdr_zero;	/* finish the read chunk list */
		*iptr++ = xdr_zero;	/* encode a NULL write chunk list */
		*iptr++ = xdr_zero;	/* encode a NULL reply chunk */
	} else {
		warray->wc_discrim = xdr_one;
		warray->wc_nchunks = cpu_to_be32(nchunks);
		iptr = (__be32 *) cur_wchunk;
		if (type == rpcrdma_writech) {
			*iptr++ = xdr_zero; /* finish the write chunk list */
			*iptr++ = xdr_zero; /* encode a NULL reply chunk */
		}
	}

	/*
	 * Return header size.
	 */
	return (unsigned char *)iptr - (unsigned char *)headerp;

out:
	for (pos = 0; nchunks--;)
		pos += r_xprt->rx_ia.ri_ops->ro_unmap(r_xprt,
						      &req->rl_segments[pos]);
	return n;
}

/*
 * Copy write data inline.
 * This function is used for "small" requests. Data which is passed
 * to RPC via iovecs (or page list) is copied directly into the
 * pre-registered memory buffer for this request. For small amounts
 * of data, this is efficient. The cutoff value is tunable.
 */
static void rpcrdma_inline_pullup(struct rpc_rqst *rqst)
{
	int i, npages, curlen;
	int copy_len;
	unsigned char *srcp, *destp;
	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(rqst->rq_xprt);
	int page_base;
	struct page **ppages;

	destp = rqst->rq_svec[0].iov_base;
	curlen = rqst->rq_svec[0].iov_len;
	destp += curlen;

	dprintk("RPC:       %s: destp 0x%p len %d hdrlen %d\n",
		__func__, destp, rqst->rq_slen, curlen);

	copy_len = rqst->rq_snd_buf.page_len;

	if (rqst->rq_snd_buf.tail[0].iov_len) {
		curlen = rqst->rq_snd_buf.tail[0].iov_len;
		if (destp + copy_len != rqst->rq_snd_buf.tail[0].iov_base) {
			memmove(destp + copy_len,
				rqst->rq_snd_buf.tail[0].iov_base, curlen);
			r_xprt->rx_stats.pullup_copy_count += curlen;
		}
		dprintk("RPC:       %s: tail destp 0x%p len %d\n",
			__func__, destp + copy_len, curlen);
		rqst->rq_svec[0].iov_len += curlen;
	}
	r_xprt->rx_stats.pullup_copy_count += copy_len;

	page_base = rqst->rq_snd_buf.page_base;
	ppages = rqst->rq_snd_buf.pages + (page_base >> PAGE_SHIFT);
	page_base &= ~PAGE_MASK;
	npages = PAGE_ALIGN(page_base+copy_len) >> PAGE_SHIFT;
	for (i = 0; copy_len && i < npages; i++) {
		curlen = PAGE_SIZE - page_base;
		if (curlen > copy_len)
			curlen = copy_len;
		dprintk("RPC:       %s: page %d destp 0x%p len %d curlen %d\n",
			__func__, i, destp, copy_len, curlen);
		srcp = kmap_atomic(ppages[i]);
		memcpy(destp, srcp+page_base, curlen);
		kunmap_atomic(srcp);
		rqst->rq_svec[0].iov_len += curlen;
		destp += curlen;
		copy_len -= curlen;
		page_base = 0;
	}
	/* header now contains entire send message */
}

/*
 * Marshal a request: the primary job of this routine is to choose
 * the transfer modes. See comments below.
 *
 * Uses multiple RDMA IOVs for a request:
 *  [0] -- RPC RDMA header, which uses memory from the *start* of the
 *         preregistered buffer that already holds the RPC data in
 *         its middle.
 *  [1] -- the RPC header/data, marshaled by RPC and the NFS protocol.
 *  [2] -- optional padding.
 *  [3] -- if padded, header only in [1] and data here.
 *
 * Returns zero on success, otherwise a negative errno.
 */

int
rpcrdma_marshal_req(struct rpc_rqst *rqst)
{
	struct rpc_xprt *xprt = rqst->rq_xprt;
	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(xprt);
	struct rpcrdma_req *req = rpcr_to_rdmar(rqst);
	char *base;
	size_t rpclen;
	ssize_t hdrlen;
	enum rpcrdma_chunktype rtype, wtype;
	struct rpcrdma_msg *headerp;

#if defined(CONFIG_SUNRPC_BACKCHANNEL)
	if (test_bit(RPC_BC_PA_IN_USE, &rqst->rq_bc_pa_state))
		return rpcrdma_bc_marshal_reply(rqst);
#endif

	/*
	 * rpclen gets amount of data in first buffer, which is the
	 * pre-registered buffer.
	 */
	base = rqst->rq_svec[0].iov_base;
	rpclen = rqst->rq_svec[0].iov_len;

	headerp = rdmab_to_msg(req->rl_rdmabuf);
	/* don't byte-swap XID, it's already done in request */
	headerp->rm_xid = rqst->rq_xid;
	headerp->rm_vers = rpcrdma_version;
	headerp->rm_credit = cpu_to_be32(r_xprt->rx_buf.rb_max_requests);
	headerp->rm_type = rdma_msg;

	/*
	 * Chunks needed for results?
	 *
	 * o Read ops return data as write chunk(s), header as inline.
	 * o If the expected result is under the inline threshold, all ops
	 *   return as inline.
	 * o Large non-read ops return as a single reply chunk.
	 */
	if (rqst->rq_rcv_buf.flags & XDRBUF_READ)
		wtype = rpcrdma_writech;
	else if (rpcrdma_results_inline(rqst))
		wtype = rpcrdma_noch;
	else
		wtype = rpcrdma_replych;

	/*
	 * Chunks needed for arguments?
	 *
	 * o If the total request is under the inline threshold, all ops
	 *   are sent as inline.
	 * o Large write ops transmit data as read chunk(s), header as
	 *   inline.
	 * o Large non-write ops are sent with the entire message as a
	 *   single read chunk (protocol 0-position special case).
	 *
	 * This assumes that the upper layer does not present a request
	 * that both has a data payload, and whose non-data arguments
	 * by themselves are larger than the inline threshold.
	 */
	if (rpcrdma_args_inline(rqst)) {
		rtype = rpcrdma_noch;
	} else if (rqst->rq_snd_buf.flags & XDRBUF_WRITE) {
		rtype = rpcrdma_readch;
	} else {
		r_xprt->rx_stats.nomsg_call_count++;
		headerp->rm_type = htonl(RDMA_NOMSG);
		rtype = rpcrdma_areadch;
		rpclen = 0;
	}

	/* The following simplification is not true forever */
	if (rtype != rpcrdma_noch && wtype == rpcrdma_replych)
		wtype = rpcrdma_noch;
	if (rtype != rpcrdma_noch && wtype != rpcrdma_noch) {
		dprintk("RPC:       %s: cannot marshal multiple chunk lists\n",
			__func__);
		return -EIO;
	}

	hdrlen = RPCRDMA_HDRLEN_MIN;

	/*
	 * Pull up any extra send data into the preregistered buffer.
	 * When padding is in use and applies to the transfer, insert
	 * it and change the message type.
	 */
	if (rtype == rpcrdma_noch) {

		rpcrdma_inline_pullup(rqst);

		headerp->rm_body.rm_nochunks.rm_empty[0] = xdr_zero;
		headerp->rm_body.rm_nochunks.rm_empty[1] = xdr_zero;
		headerp->rm_body.rm_nochunks.rm_empty[2] = xdr_zero;
		/* new length after pullup */
		rpclen = rqst->rq_svec[0].iov_len;
	} else if (rtype == rpcrdma_readch)
		rpclen += rpcrdma_tail_pullup(&rqst->rq_snd_buf);
	if (rtype != rpcrdma_noch) {
		hdrlen = rpcrdma_create_chunks(rqst, &rqst->rq_snd_buf,
					       headerp, rtype);
		wtype = rtype;	/* simplify dprintk */

	} else if (wtype != rpcrdma_noch) {
		hdrlen = rpcrdma_create_chunks(rqst, &rqst->rq_rcv_buf,
					       headerp, wtype);
	}
	if (hdrlen < 0)
		return hdrlen;

	dprintk("RPC:       %s: %s: hdrlen %zd rpclen %zd"
		" headerp 0x%p base 0x%p lkey 0x%x\n",
		__func__, transfertypes[wtype], hdrlen, rpclen,
		headerp, base, rdmab_lkey(req->rl_rdmabuf));

	/*
	 * initialize send_iov's - normally only two: rdma chunk header and
	 * single preregistered RPC header buffer, but if padding is present,
	 * then use a preregistered (and zeroed) pad buffer between the RPC
	 * header and any write data. In all non-rdma cases, any following
	 * data has been copied into the RPC header buffer.
	 */
	req->rl_send_iov[0].addr = rdmab_addr(req->rl_rdmabuf);
	req->rl_send_iov[0].length = hdrlen;
	req->rl_send_iov[0].lkey = rdmab_lkey(req->rl_rdmabuf);

	req->rl_niovs = 1;
	if (rtype == rpcrdma_areadch)
		return 0;

	req->rl_send_iov[1].addr = rdmab_addr(req->rl_sendbuf);
	req->rl_send_iov[1].length = rpclen;
	req->rl_send_iov[1].lkey = rdmab_lkey(req->rl_sendbuf);

	req->rl_niovs = 2;
	return 0;
}

/*
 * Chase down a received write or reply chunklist to get length
 * RDMA'd by server. See map at rpcrdma_create_chunks()! :-)
 */
static int
rpcrdma_count_chunks(struct rpcrdma_rep *rep, unsigned int max, int wrchunk, __be32 **iptrp)
{
	unsigned int i, total_len;
	struct rpcrdma_write_chunk *cur_wchunk;
	char *base = (char *)rdmab_to_msg(rep->rr_rdmabuf);

	i = be32_to_cpu(**iptrp);
	if (i > max)
		return -1;
	cur_wchunk = (struct rpcrdma_write_chunk *) (*iptrp + 1);
	total_len = 0;
	while (i--) {
		struct rpcrdma_segment *seg = &cur_wchunk->wc_target;
		ifdebug(FACILITY) {
			u64 off;
			xdr_decode_hyper((__be32 *)&seg->rs_offset, &off);
			dprintk("RPC:       %s: chunk %d@0x%llx:0x%x\n",
				__func__,
				be32_to_cpu(seg->rs_length),
				(unsigned long long)off,
				be32_to_cpu(seg->rs_handle));
		}
		total_len += be32_to_cpu(seg->rs_length);
		++cur_wchunk;
	}
	/* check and adjust for properly terminated write chunk */
	if (wrchunk) {
		__be32 *w = (__be32 *) cur_wchunk;
		if (*w++ != xdr_zero)
			return -1;
		cur_wchunk = (struct rpcrdma_write_chunk *) w;
	}
	if ((char *)cur_wchunk > base + rep->rr_len)
		return -1;

	*iptrp = (__be32 *) cur_wchunk;
	return total_len;
}

/*
 * Scatter inline received data back into provided iov's.
 */
static void
rpcrdma_inline_fixup(struct rpc_rqst *rqst, char *srcp, int copy_len, int pad)
{
	int i, npages, curlen, olen;
	char *destp;
	struct page **ppages;
	int page_base;

	curlen = rqst->rq_rcv_buf.head[0].iov_len;
	if (curlen > copy_len) {	/* write chunk header fixup */
		curlen = copy_len;
		rqst->rq_rcv_buf.head[0].iov_len = curlen;
	}

	dprintk("RPC:       %s: srcp 0x%p len %d hdrlen %d\n",
		__func__, srcp, copy_len, curlen);

	/* Shift pointer for first receive segment only */
	rqst->rq_rcv_buf.head[0].iov_base = srcp;
	srcp += curlen;
	copy_len -= curlen;

	olen = copy_len;
	i = 0;
	rpcx_to_rdmax(rqst->rq_xprt)->rx_stats.fixup_copy_count += olen;
	page_base = rqst->rq_rcv_buf.page_base;
	ppages = rqst->rq_rcv_buf.pages + (page_base >> PAGE_SHIFT);
	page_base &= ~PAGE_MASK;

	if (copy_len && rqst->rq_rcv_buf.page_len) {
		npages = PAGE_ALIGN(page_base +
			rqst->rq_rcv_buf.page_len) >> PAGE_SHIFT;
		for (; i < npages; i++) {
			curlen = PAGE_SIZE - page_base;
			if (curlen > copy_len)
				curlen = copy_len;
			dprintk("RPC:       %s: page %d"
				" srcp 0x%p len %d curlen %d\n",
				__func__, i, srcp, copy_len, curlen);
			destp = kmap_atomic(ppages[i]);
			memcpy(destp + page_base, srcp, curlen);
			flush_dcache_page(ppages[i]);
			kunmap_atomic(destp);
			srcp += curlen;
			copy_len -= curlen;
			if (copy_len == 0)
				break;
			page_base = 0;
		}
	}

	if (copy_len && rqst->rq_rcv_buf.tail[0].iov_len) {
		curlen = copy_len;
		if (curlen > rqst->rq_rcv_buf.tail[0].iov_len)
			curlen = rqst->rq_rcv_buf.tail[0].iov_len;
		if (rqst->rq_rcv_buf.tail[0].iov_base != srcp)
			memmove(rqst->rq_rcv_buf.tail[0].iov_base, srcp, curlen);
		dprintk("RPC:       %s: tail srcp 0x%p len %d curlen %d\n",
			__func__, srcp, copy_len, curlen);
		rqst->rq_rcv_buf.tail[0].iov_len = curlen;
		copy_len -= curlen; ++i;
	} else
		rqst->rq_rcv_buf.tail[0].iov_len = 0;

	if (pad) {
		/* implicit padding on terminal chunk */
		unsigned char *p = rqst->rq_rcv_buf.tail[0].iov_base;
		while (pad--)
			p[rqst->rq_rcv_buf.tail[0].iov_len++] = 0;
	}

	if (copy_len)
		dprintk("RPC:       %s: %d bytes in"
			" %d extra segments (%d lost)\n",
			__func__, olen, i, copy_len);

	/* TBD avoid a warning from call_decode() */
	rqst->rq_private_buf = rqst->rq_rcv_buf;
}

void
rpcrdma_connect_worker(struct work_struct *work)
{
	struct rpcrdma_ep *ep =
		container_of(work, struct rpcrdma_ep, rep_connect_worker.work);
	struct rpcrdma_xprt *r_xprt =
		container_of(ep, struct rpcrdma_xprt, rx_ep);
	struct rpc_xprt *xprt = &r_xprt->rx_xprt;

	spin_lock_bh(&xprt->transport_lock);
	if (++xprt->connect_cookie == 0)	/* maintain a reserved value */
		++xprt->connect_cookie;
	if (ep->rep_connected > 0) {
		if (!xprt_test_and_set_connected(xprt))
			xprt_wake_pending_tasks(xprt, 0);
	} else {
		if (xprt_test_and_clear_connected(xprt))
			xprt_wake_pending_tasks(xprt, -ENOTCONN);
	}
	spin_unlock_bh(&xprt->transport_lock);
}

#if defined(CONFIG_SUNRPC_BACKCHANNEL)
/* By convention, backchannel calls arrive via rdma_msg type
 * messages, and never populate the chunk lists. This makes
 * the RPC/RDMA header small and fixed in size, so it is
 * straightforward to check the RPC header's direction field.
 */
static bool
rpcrdma_is_bcall(struct rpcrdma_msg *headerp)
{
	__be32 *p = (__be32 *)headerp;

	if (headerp->rm_type != rdma_msg)
		return false;
	if (headerp->rm_body.rm_chunks[0] != xdr_zero)
		return false;
	if (headerp->rm_body.rm_chunks[1] != xdr_zero)
		return false;
	if (headerp->rm_body.rm_chunks[2] != xdr_zero)
		return false;

	/* sanity */
	if (p[7] != headerp->rm_xid)
		return false;
	/* call direction */
	if (p[8] != cpu_to_be32(RPC_CALL))
		return false;

	return true;
}
#endif	/* CONFIG_SUNRPC_BACKCHANNEL */

/*
 * This function is called when an async event is posted to
 * the connection which changes the connection state. All it
 * does at this point is mark the connection up/down, the rpc
 * timers do the rest.
 */
void
rpcrdma_conn_func(struct rpcrdma_ep *ep)
{
	schedule_delayed_work(&ep->rep_connect_worker, 0);
}

/* Process received RPC/RDMA messages.
 *
 * Errors must result in the RPC task either being awakened, or
 * allowed to timeout, to discover the errors at that time.
 */
void
rpcrdma_reply_handler(struct rpcrdma_rep *rep)
{
	struct rpcrdma_msg *headerp;
	struct rpcrdma_req *req;
	struct rpc_rqst *rqst;
	struct rpcrdma_xprt *r_xprt = rep->rr_rxprt;
	struct rpc_xprt *xprt = &r_xprt->rx_xprt;
	__be32 *iptr;
	int rdmalen, status;
	unsigned long cwnd;
	u32 credits;

	dprintk("RPC:       %s: incoming rep %p\n", __func__, rep);

	if (rep->rr_len == RPCRDMA_BAD_LEN)
		goto out_badstatus;
	if (rep->rr_len < RPCRDMA_HDRLEN_MIN)
		goto out_shortreply;

	headerp = rdmab_to_msg(rep->rr_rdmabuf);
	if (headerp->rm_vers != rpcrdma_version)
		goto out_badversion;
#if defined(CONFIG_SUNRPC_BACKCHANNEL)
	if (rpcrdma_is_bcall(headerp))
		goto out_bcall;
#endif

	/* Match incoming rpcrdma_rep to an rpcrdma_req to
	 * get context for handling any incoming chunks.
	 */
	spin_lock_bh(&xprt->transport_lock);
	rqst = xprt_lookup_rqst(xprt, headerp->rm_xid);
	if (!rqst)
		goto out_nomatch;

	req = rpcr_to_rdmar(rqst);
	if (req->rl_reply)
		goto out_duplicate;

	dprintk("RPC:       %s: reply 0x%p completes request 0x%p\n"
		"                   RPC request 0x%p xid 0x%08x\n",
			__func__, rep, req, rqst,
			be32_to_cpu(headerp->rm_xid));

	/* from here on, the reply is no longer an orphan */
	req->rl_reply = rep;
	xprt->reestablish_timeout = 0;

	/* check for expected message types */
	/* The order of some of these tests is important. */
	switch (headerp->rm_type) {
	case rdma_msg:
		/* never expect read chunks */
		/* never expect reply chunks (two ways to check) */
		/* never expect write chunks without having offered RDMA */
		if (headerp->rm_body.rm_chunks[0] != xdr_zero ||
		    (headerp->rm_body.rm_chunks[1] == xdr_zero &&
		     headerp->rm_body.rm_chunks[2] != xdr_zero) ||
		    (headerp->rm_body.rm_chunks[1] != xdr_zero &&
		     req->rl_nchunks == 0))
			goto badheader;
		if (headerp->rm_body.rm_chunks[1] != xdr_zero) {
			/* count any expected write chunks in read reply */
			/* start at write chunk array count */
			iptr = &headerp->rm_body.rm_chunks[2];
			rdmalen = rpcrdma_count_chunks(rep,
						req->rl_nchunks, 1, &iptr);
			/* check for validity, and no reply chunk after */
			if (rdmalen < 0 || *iptr++ != xdr_zero)
				goto badheader;
			rep->rr_len -=
			    ((unsigned char *)iptr - (unsigned char *)headerp);
			status = rep->rr_len + rdmalen;
			r_xprt->rx_stats.total_rdma_reply += rdmalen;
			/* special case - last chunk may omit padding */
			if (rdmalen &= 3) {
				rdmalen = 4 - rdmalen;
				status += rdmalen;
			}
		} else {
			/* else ordinary inline */
			rdmalen = 0;
			iptr = (__be32 *)((unsigned char *)headerp +
							RPCRDMA_HDRLEN_MIN);
			rep->rr_len -= RPCRDMA_HDRLEN_MIN;
			status = rep->rr_len;
		}
		/* Fix up the rpc results for upper layer */
		rpcrdma_inline_fixup(rqst, (char *)iptr, rep->rr_len, rdmalen);
		break;

	case rdma_nomsg:
		/* never expect read or write chunks, always reply chunks */
		if (headerp->rm_body.rm_chunks[0] != xdr_zero ||
		    headerp->rm_body.rm_chunks[1] != xdr_zero ||
		    headerp->rm_body.rm_chunks[2] != xdr_one ||
		    req->rl_nchunks == 0)
			goto badheader;
		iptr = (__be32 *)((unsigned char *)headerp +
							RPCRDMA_HDRLEN_MIN);
		rdmalen = rpcrdma_count_chunks(rep, req->rl_nchunks, 0, &iptr);
		if (rdmalen < 0)
			goto badheader;
		r_xprt->rx_stats.total_rdma_reply += rdmalen;
		/* Reply chunk buffer already is the reply vector - no fixup. */
		status = rdmalen;
		break;

badheader:
	default:
		dprintk("%s: invalid rpcrdma reply header (type %d):"
				" chunks[012] == %d %d %d"
				" expected chunks <= %d\n",
				__func__, be32_to_cpu(headerp->rm_type),
				headerp->rm_body.rm_chunks[0],
				headerp->rm_body.rm_chunks[1],
				headerp->rm_body.rm_chunks[2],
				req->rl_nchunks);
		status = -EIO;
		r_xprt->rx_stats.bad_reply_count++;
		break;
	}

	credits = be32_to_cpu(headerp->rm_credit);
	if (credits == 0)
		credits = 1;	/* don't deadlock */
	else if (credits > r_xprt->rx_buf.rb_max_requests)
		credits = r_xprt->rx_buf.rb_max_requests;

	cwnd = xprt->cwnd;
	xprt->cwnd = credits << RPC_CWNDSHIFT;
	if (xprt->cwnd > cwnd)
		xprt_release_rqst_cong(rqst->rq_task);

	xprt_complete_rqst(rqst->rq_task, status);
	spin_unlock_bh(&xprt->transport_lock);
	dprintk("RPC:       %s: xprt_complete_rqst(0x%p, 0x%p, %d)\n",
			__func__, xprt, rqst, status);
	return;

out_badstatus:
	rpcrdma_recv_buffer_put(rep);
	if (r_xprt->rx_ep.rep_connected == 1) {
		r_xprt->rx_ep.rep_connected = -EIO;
		rpcrdma_conn_func(&r_xprt->rx_ep);
	}
	return;

#if defined(CONFIG_SUNRPC_BACKCHANNEL)
out_bcall:
	rpcrdma_bc_receive_call(r_xprt, rep);
	return;
#endif

out_shortreply:
	dprintk("RPC:       %s: short/invalid reply\n", __func__);
	goto repost;

out_badversion:
	dprintk("RPC:       %s: invalid version %d\n",
		__func__, be32_to_cpu(headerp->rm_vers));
	goto repost;

out_nomatch:
	spin_unlock_bh(&xprt->transport_lock);
	dprintk("RPC:       %s: no match for incoming xid 0x%08x len %d\n",
		__func__, be32_to_cpu(headerp->rm_xid),
		rep->rr_len);
	goto repost;

out_duplicate:
	spin_unlock_bh(&xprt->transport_lock);
	dprintk("RPC:       %s: "
		"duplicate reply %p to RPC request %p: xid 0x%08x\n",
		__func__, rep, req, be32_to_cpu(headerp->rm_xid));

repost:
	r_xprt->rx_stats.bad_reply_count++;
	if (rpcrdma_ep_post_recv(&r_xprt->rx_ia, &r_xprt->rx_ep, rep))
		rpcrdma_recv_buffer_put(rep);
}
OpenPOWER on IntegriCloud