1 /*
2  * Copyright (c) 2005-2006 Network Appliance, Inc. All rights reserved.
3  *
4  * This software is available to you under a choice of one of two
5  * licenses.  You may choose to be licensed under the terms of the GNU
6  * General Public License (GPL) Version 2, available from the file
7  * COPYING in the main directory of this source tree, or the BSD-type
8  * license below:
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  *
14  *      Redistributions of source code must retain the above copyright
15  *      notice, this list of conditions and the following disclaimer.
16  *
17  *      Redistributions in binary form must reproduce the above
18  *      copyright notice, this list of conditions and the following
19  *      disclaimer in the documentation and/or other materials provided
20  *      with the distribution.
21  *
22  *      Neither the name of the Network Appliance, Inc. nor the names of
23  *      its contributors may be used to endorse or promote products
24  *      derived from this software without specific prior written
25  *      permission.
26  *
27  * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
28  * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
29  * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
30  * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
31  * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
32  * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
33  * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
34  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
35  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
36  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
37  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
38  *
39  * Author: Tom Tucker <tom@opengridcomputing.com>
40  */
41 
42 #include <linux/sunrpc/debug.h>
43 #include <linux/sunrpc/rpc_rdma.h>
44 #include <linux/spinlock.h>
45 #include <asm/unaligned.h>
46 #include <rdma/ib_verbs.h>
47 #include <rdma/rdma_cm.h>
48 #include <linux/sunrpc/svc_rdma.h>
49 
50 #define RPCDBG_FACILITY	RPCDBG_SVCXPRT
51 
52 /*
53  * Replace the pages in the rq_argpages array with the pages from the SGE in
54  * the RDMA_RECV completion. The SGL should contain full pages up until the
55  * last one.
56  */
rdma_build_arg_xdr(struct svc_rqst * rqstp,struct svc_rdma_op_ctxt * ctxt,u32 byte_count)57 static void rdma_build_arg_xdr(struct svc_rqst *rqstp,
58 			       struct svc_rdma_op_ctxt *ctxt,
59 			       u32 byte_count)
60 {
61 	struct page *page;
62 	u32 bc;
63 	int sge_no;
64 
65 	/* Swap the page in the SGE with the page in argpages */
66 	page = ctxt->pages[0];
67 	put_page(rqstp->rq_pages[0]);
68 	rqstp->rq_pages[0] = page;
69 
70 	/* Set up the XDR head */
71 	rqstp->rq_arg.head[0].iov_base = page_address(page);
72 	rqstp->rq_arg.head[0].iov_len = min(byte_count, ctxt->sge[0].length);
73 	rqstp->rq_arg.len = byte_count;
74 	rqstp->rq_arg.buflen = byte_count;
75 
76 	/* Compute bytes past head in the SGL */
77 	bc = byte_count - rqstp->rq_arg.head[0].iov_len;
78 
79 	/* If data remains, store it in the pagelist */
80 	rqstp->rq_arg.page_len = bc;
81 	rqstp->rq_arg.page_base = 0;
82 	rqstp->rq_arg.pages = &rqstp->rq_pages[1];
83 	sge_no = 1;
84 	while (bc && sge_no < ctxt->count) {
85 		page = ctxt->pages[sge_no];
86 		put_page(rqstp->rq_pages[sge_no]);
87 		rqstp->rq_pages[sge_no] = page;
88 		bc -= min(bc, ctxt->sge[sge_no].length);
89 		rqstp->rq_arg.buflen += ctxt->sge[sge_no].length;
90 		sge_no++;
91 	}
92 	rqstp->rq_respages = &rqstp->rq_pages[sge_no];
93 
94 	/* We should never run out of SGE because the limit is defined to
95 	 * support the max allowed RPC data length
96 	 */
97 	BUG_ON(bc && (sge_no == ctxt->count));
98 	BUG_ON((rqstp->rq_arg.head[0].iov_len + rqstp->rq_arg.page_len)
99 	       != byte_count);
100 	BUG_ON(rqstp->rq_arg.len != byte_count);
101 
102 	/* If not all pages were used from the SGL, free the remaining ones */
103 	bc = sge_no;
104 	while (sge_no < ctxt->count) {
105 		page = ctxt->pages[sge_no++];
106 		put_page(page);
107 	}
108 	ctxt->count = bc;
109 
110 	/* Set up tail */
111 	rqstp->rq_arg.tail[0].iov_base = NULL;
112 	rqstp->rq_arg.tail[0].iov_len = 0;
113 }
114 
115 /* Encode a read-chunk-list as an array of IB SGE
116  *
117  * Assumptions:
118  * - chunk[0]->position points to pages[0] at an offset of 0
119  * - pages[] is not physically or virtually contiguous and consists of
120  *   PAGE_SIZE elements.
121  *
122  * Output:
123  * - sge array pointing into pages[] array.
124  * - chunk_sge array specifying sge index and count for each
125  *   chunk in the read list
126  *
127  */
map_read_chunks(struct svcxprt_rdma * xprt,struct svc_rqst * rqstp,struct svc_rdma_op_ctxt * head,struct rpcrdma_msg * rmsgp,struct svc_rdma_req_map * rpl_map,struct svc_rdma_req_map * chl_map,int ch_count,int byte_count)128 static int map_read_chunks(struct svcxprt_rdma *xprt,
129 			   struct svc_rqst *rqstp,
130 			   struct svc_rdma_op_ctxt *head,
131 			   struct rpcrdma_msg *rmsgp,
132 			   struct svc_rdma_req_map *rpl_map,
133 			   struct svc_rdma_req_map *chl_map,
134 			   int ch_count,
135 			   int byte_count)
136 {
137 	int sge_no;
138 	int sge_bytes;
139 	int page_off;
140 	int page_no;
141 	int ch_bytes;
142 	int ch_no;
143 	struct rpcrdma_read_chunk *ch;
144 
145 	sge_no = 0;
146 	page_no = 0;
147 	page_off = 0;
148 	ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0];
149 	ch_no = 0;
150 	ch_bytes = ntohl(ch->rc_target.rs_length);
151 	head->arg.head[0] = rqstp->rq_arg.head[0];
152 	head->arg.tail[0] = rqstp->rq_arg.tail[0];
153 	head->arg.pages = &head->pages[head->count];
154 	head->hdr_count = head->count; /* save count of hdr pages */
155 	head->arg.page_base = 0;
156 	head->arg.page_len = ch_bytes;
157 	head->arg.len = rqstp->rq_arg.len + ch_bytes;
158 	head->arg.buflen = rqstp->rq_arg.buflen + ch_bytes;
159 	head->count++;
160 	chl_map->ch[0].start = 0;
161 	while (byte_count) {
162 		rpl_map->sge[sge_no].iov_base =
163 			page_address(rqstp->rq_arg.pages[page_no]) + page_off;
164 		sge_bytes = min_t(int, PAGE_SIZE-page_off, ch_bytes);
165 		rpl_map->sge[sge_no].iov_len = sge_bytes;
166 		/*
167 		 * Don't bump head->count here because the same page
168 		 * may be used by multiple SGE.
169 		 */
170 		head->arg.pages[page_no] = rqstp->rq_arg.pages[page_no];
171 		rqstp->rq_respages = &rqstp->rq_arg.pages[page_no+1];
172 
173 		byte_count -= sge_bytes;
174 		ch_bytes -= sge_bytes;
175 		sge_no++;
176 		/*
177 		 * If all bytes for this chunk have been mapped to an
178 		 * SGE, move to the next SGE
179 		 */
180 		if (ch_bytes == 0) {
181 			chl_map->ch[ch_no].count =
182 				sge_no - chl_map->ch[ch_no].start;
183 			ch_no++;
184 			ch++;
185 			chl_map->ch[ch_no].start = sge_no;
186 			ch_bytes = ntohl(ch->rc_target.rs_length);
187 			/* If bytes remaining account for next chunk */
188 			if (byte_count) {
189 				head->arg.page_len += ch_bytes;
190 				head->arg.len += ch_bytes;
191 				head->arg.buflen += ch_bytes;
192 			}
193 		}
194 		/*
195 		 * If this SGE consumed all of the page, move to the
196 		 * next page
197 		 */
198 		if ((sge_bytes + page_off) == PAGE_SIZE) {
199 			page_no++;
200 			page_off = 0;
201 			/*
202 			 * If there are still bytes left to map, bump
203 			 * the page count
204 			 */
205 			if (byte_count)
206 				head->count++;
207 		} else
208 			page_off += sge_bytes;
209 	}
210 	BUG_ON(byte_count != 0);
211 	return sge_no;
212 }
213 
214 /* Map a read-chunk-list to an XDR and fast register the page-list.
215  *
216  * Assumptions:
217  * - chunk[0]	position points to pages[0] at an offset of 0
218  * - pages[]	will be made physically contiguous by creating a one-off memory
219  *		region using the fastreg verb.
220  * - byte_count is # of bytes in read-chunk-list
221  * - ch_count	is # of chunks in read-chunk-list
222  *
223  * Output:
224  * - sge array pointing into pages[] array.
225  * - chunk_sge array specifying sge index and count for each
226  *   chunk in the read list
227  */
fast_reg_read_chunks(struct svcxprt_rdma * xprt,struct svc_rqst * rqstp,struct svc_rdma_op_ctxt * head,struct rpcrdma_msg * rmsgp,struct svc_rdma_req_map * rpl_map,struct svc_rdma_req_map * chl_map,int ch_count,int byte_count)228 static int fast_reg_read_chunks(struct svcxprt_rdma *xprt,
229 				struct svc_rqst *rqstp,
230 				struct svc_rdma_op_ctxt *head,
231 				struct rpcrdma_msg *rmsgp,
232 				struct svc_rdma_req_map *rpl_map,
233 				struct svc_rdma_req_map *chl_map,
234 				int ch_count,
235 				int byte_count)
236 {
237 	int page_no;
238 	int ch_no;
239 	u32 offset;
240 	struct rpcrdma_read_chunk *ch;
241 	struct svc_rdma_fastreg_mr *frmr;
242 	int ret = 0;
243 
244 	frmr = svc_rdma_get_frmr(xprt);
245 	if (IS_ERR(frmr))
246 		return -ENOMEM;
247 
248 	head->frmr = frmr;
249 	head->arg.head[0] = rqstp->rq_arg.head[0];
250 	head->arg.tail[0] = rqstp->rq_arg.tail[0];
251 	head->arg.pages = &head->pages[head->count];
252 	head->hdr_count = head->count; /* save count of hdr pages */
253 	head->arg.page_base = 0;
254 	head->arg.page_len = byte_count;
255 	head->arg.len = rqstp->rq_arg.len + byte_count;
256 	head->arg.buflen = rqstp->rq_arg.buflen + byte_count;
257 
258 	/* Fast register the page list */
259 	frmr->kva = page_address(rqstp->rq_arg.pages[0]);
260 	frmr->direction = DMA_FROM_DEVICE;
261 	frmr->access_flags = (IB_ACCESS_LOCAL_WRITE|IB_ACCESS_REMOTE_WRITE);
262 	frmr->map_len = byte_count;
263 	frmr->page_list_len = PAGE_ALIGN(byte_count) >> PAGE_SHIFT;
264 	for (page_no = 0; page_no < frmr->page_list_len; page_no++) {
265 		frmr->page_list->page_list[page_no] =
266 			ib_dma_map_page(xprt->sc_cm_id->device,
267 					rqstp->rq_arg.pages[page_no], 0,
268 					PAGE_SIZE, DMA_FROM_DEVICE);
269 		if (ib_dma_mapping_error(xprt->sc_cm_id->device,
270 					 frmr->page_list->page_list[page_no]))
271 			goto fatal_err;
272 		atomic_inc(&xprt->sc_dma_used);
273 		head->arg.pages[page_no] = rqstp->rq_arg.pages[page_no];
274 	}
275 	head->count += page_no;
276 
277 	/* rq_respages points one past arg pages */
278 	rqstp->rq_respages = &rqstp->rq_arg.pages[page_no];
279 
280 	/* Create the reply and chunk maps */
281 	offset = 0;
282 	ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0];
283 	for (ch_no = 0; ch_no < ch_count; ch_no++) {
284 		int len = ntohl(ch->rc_target.rs_length);
285 		rpl_map->sge[ch_no].iov_base = frmr->kva + offset;
286 		rpl_map->sge[ch_no].iov_len = len;
287 		chl_map->ch[ch_no].count = 1;
288 		chl_map->ch[ch_no].start = ch_no;
289 		offset += len;
290 		ch++;
291 	}
292 
293 	ret = svc_rdma_fastreg(xprt, frmr);
294 	if (ret)
295 		goto fatal_err;
296 
297 	return ch_no;
298 
299  fatal_err:
300 	printk("svcrdma: error fast registering xdr for xprt %p", xprt);
301 	svc_rdma_put_frmr(xprt, frmr);
302 	return -EIO;
303 }
304 
rdma_set_ctxt_sge(struct svcxprt_rdma * xprt,struct svc_rdma_op_ctxt * ctxt,struct svc_rdma_fastreg_mr * frmr,struct kvec * vec,u64 * sgl_offset,int count)305 static int rdma_set_ctxt_sge(struct svcxprt_rdma *xprt,
306 			     struct svc_rdma_op_ctxt *ctxt,
307 			     struct svc_rdma_fastreg_mr *frmr,
308 			     struct kvec *vec,
309 			     u64 *sgl_offset,
310 			     int count)
311 {
312 	int i;
313 	unsigned long off;
314 
315 	ctxt->count = count;
316 	ctxt->direction = DMA_FROM_DEVICE;
317 	for (i = 0; i < count; i++) {
318 		ctxt->sge[i].length = 0; /* in case map fails */
319 		if (!frmr) {
320 			BUG_ON(!virt_to_page(vec[i].iov_base));
321 			off = (unsigned long)vec[i].iov_base & ~PAGE_MASK;
322 			ctxt->sge[i].addr =
323 				ib_dma_map_page(xprt->sc_cm_id->device,
324 						virt_to_page(vec[i].iov_base),
325 						off,
326 						vec[i].iov_len,
327 						DMA_FROM_DEVICE);
328 			if (ib_dma_mapping_error(xprt->sc_cm_id->device,
329 						 ctxt->sge[i].addr))
330 				return -EINVAL;
331 			ctxt->sge[i].lkey = xprt->sc_dma_lkey;
332 			atomic_inc(&xprt->sc_dma_used);
333 		} else {
334 			ctxt->sge[i].addr = (unsigned long)vec[i].iov_base;
335 			ctxt->sge[i].lkey = frmr->mr->lkey;
336 		}
337 		ctxt->sge[i].length = vec[i].iov_len;
338 		*sgl_offset = *sgl_offset + vec[i].iov_len;
339 	}
340 	return 0;
341 }
342 
rdma_read_max_sge(struct svcxprt_rdma * xprt,int sge_count)343 static int rdma_read_max_sge(struct svcxprt_rdma *xprt, int sge_count)
344 {
345 	if ((rdma_node_get_transport(xprt->sc_cm_id->device->node_type) ==
346 	     RDMA_TRANSPORT_IWARP) &&
347 	    sge_count > 1)
348 		return 1;
349 	else
350 		return min_t(int, sge_count, xprt->sc_max_sge);
351 }
352 
353 /*
354  * Use RDMA_READ to read data from the advertised client buffer into the
355  * XDR stream starting at rq_arg.head[0].iov_base.
356  * Each chunk in the array
357  * contains the following fields:
358  * discrim      - '1', This isn't used for data placement
359  * position     - The xdr stream offset (the same for every chunk)
360  * handle       - RMR for client memory region
361  * length       - data transfer length
362  * offset       - 64 bit tagged offset in remote memory region
363  *
364  * On our side, we need to read into a pagelist. The first page immediately
365  * follows the RPC header.
366  *
367  * This function returns:
368  * 0 - No error and no read-list found.
369  *
370  * 1 - Successful read-list processing. The data is not yet in
371  * the pagelist and therefore the RPC request must be deferred. The
372  * I/O completion will enqueue the transport again and
373  * svc_rdma_recvfrom will complete the request.
374  *
375  * <0 - Error processing/posting read-list.
376  *
377  * NOTE: The ctxt must not be touched after the last WR has been posted
378  * because the I/O completion processing may occur on another
379  * processor and free / modify the context. Ne touche pas!
380  */
rdma_read_xdr(struct svcxprt_rdma * xprt,struct rpcrdma_msg * rmsgp,struct svc_rqst * rqstp,struct svc_rdma_op_ctxt * hdr_ctxt)381 static int rdma_read_xdr(struct svcxprt_rdma *xprt,
382 			 struct rpcrdma_msg *rmsgp,
383 			 struct svc_rqst *rqstp,
384 			 struct svc_rdma_op_ctxt *hdr_ctxt)
385 {
386 	struct ib_send_wr read_wr;
387 	struct ib_send_wr inv_wr;
388 	int err = 0;
389 	int ch_no;
390 	int ch_count;
391 	int byte_count;
392 	int sge_count;
393 	u64 sgl_offset;
394 	struct rpcrdma_read_chunk *ch;
395 	struct svc_rdma_op_ctxt *ctxt = NULL;
396 	struct svc_rdma_req_map *rpl_map;
397 	struct svc_rdma_req_map *chl_map;
398 
399 	/* If no read list is present, return 0 */
400 	ch = svc_rdma_get_read_chunk(rmsgp);
401 	if (!ch)
402 		return 0;
403 
404 	svc_rdma_rcl_chunk_counts(ch, &ch_count, &byte_count);
405 	if (ch_count > RPCSVC_MAXPAGES)
406 		return -EINVAL;
407 
408 	/* Allocate temporary reply and chunk maps */
409 	rpl_map = svc_rdma_get_req_map();
410 	chl_map = svc_rdma_get_req_map();
411 
412 	if (!xprt->sc_frmr_pg_list_len)
413 		sge_count = map_read_chunks(xprt, rqstp, hdr_ctxt, rmsgp,
414 					    rpl_map, chl_map, ch_count,
415 					    byte_count);
416 	else
417 		sge_count = fast_reg_read_chunks(xprt, rqstp, hdr_ctxt, rmsgp,
418 						 rpl_map, chl_map, ch_count,
419 						 byte_count);
420 	if (sge_count < 0) {
421 		err = -EIO;
422 		goto out;
423 	}
424 
425 	sgl_offset = 0;
426 	ch_no = 0;
427 
428 	for (ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0];
429 	     ch->rc_discrim != 0; ch++, ch_no++) {
430 		u64 rs_offset;
431 next_sge:
432 		ctxt = svc_rdma_get_context(xprt);
433 		ctxt->direction = DMA_FROM_DEVICE;
434 		ctxt->frmr = hdr_ctxt->frmr;
435 		ctxt->read_hdr = NULL;
436 		clear_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags);
437 		clear_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags);
438 
439 		/* Prepare READ WR */
440 		memset(&read_wr, 0, sizeof read_wr);
441 		read_wr.wr_id = (unsigned long)ctxt;
442 		read_wr.opcode = IB_WR_RDMA_READ;
443 		ctxt->wr_op = read_wr.opcode;
444 		read_wr.send_flags = IB_SEND_SIGNALED;
445 		read_wr.wr.rdma.rkey = ntohl(ch->rc_target.rs_handle);
446 		xdr_decode_hyper((__be32 *)&ch->rc_target.rs_offset,
447 				 &rs_offset);
448 		read_wr.wr.rdma.remote_addr = rs_offset + sgl_offset;
449 		read_wr.sg_list = ctxt->sge;
450 		read_wr.num_sge =
451 			rdma_read_max_sge(xprt, chl_map->ch[ch_no].count);
452 		err = rdma_set_ctxt_sge(xprt, ctxt, hdr_ctxt->frmr,
453 					&rpl_map->sge[chl_map->ch[ch_no].start],
454 					&sgl_offset,
455 					read_wr.num_sge);
456 		if (err) {
457 			svc_rdma_unmap_dma(ctxt);
458 			svc_rdma_put_context(ctxt, 0);
459 			goto out;
460 		}
461 		if (((ch+1)->rc_discrim == 0) &&
462 		    (read_wr.num_sge == chl_map->ch[ch_no].count)) {
463 			/*
464 			 * Mark the last RDMA_READ with a bit to
465 			 * indicate all RPC data has been fetched from
466 			 * the client and the RPC needs to be enqueued.
467 			 */
468 			set_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags);
469 			if (hdr_ctxt->frmr) {
470 				set_bit(RDMACTXT_F_FAST_UNREG, &ctxt->flags);
471 				/*
472 				 * Invalidate the local MR used to map the data
473 				 * sink.
474 				 */
475 				if (xprt->sc_dev_caps &
476 				    SVCRDMA_DEVCAP_READ_W_INV) {
477 					read_wr.opcode =
478 						IB_WR_RDMA_READ_WITH_INV;
479 					ctxt->wr_op = read_wr.opcode;
480 					read_wr.ex.invalidate_rkey =
481 						ctxt->frmr->mr->lkey;
482 				} else {
483 					/* Prepare INVALIDATE WR */
484 					memset(&inv_wr, 0, sizeof inv_wr);
485 					inv_wr.opcode = IB_WR_LOCAL_INV;
486 					inv_wr.send_flags = IB_SEND_SIGNALED;
487 					inv_wr.ex.invalidate_rkey =
488 						hdr_ctxt->frmr->mr->lkey;
489 					read_wr.next = &inv_wr;
490 				}
491 			}
492 			ctxt->read_hdr = hdr_ctxt;
493 		}
494 		/* Post the read */
495 		err = svc_rdma_send(xprt, &read_wr);
496 		if (err) {
497 			printk(KERN_ERR "svcrdma: Error %d posting RDMA_READ\n",
498 			       err);
499 			set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags);
500 			svc_rdma_unmap_dma(ctxt);
501 			svc_rdma_put_context(ctxt, 0);
502 			goto out;
503 		}
504 		atomic_inc(&rdma_stat_read);
505 
506 		if (read_wr.num_sge < chl_map->ch[ch_no].count) {
507 			chl_map->ch[ch_no].count -= read_wr.num_sge;
508 			chl_map->ch[ch_no].start += read_wr.num_sge;
509 			goto next_sge;
510 		}
511 		sgl_offset = 0;
512 		err = 1;
513 	}
514 
515  out:
516 	svc_rdma_put_req_map(rpl_map);
517 	svc_rdma_put_req_map(chl_map);
518 
519 	/* Detach arg pages. svc_recv will replenish them */
520 	for (ch_no = 0; &rqstp->rq_pages[ch_no] < rqstp->rq_respages; ch_no++)
521 		rqstp->rq_pages[ch_no] = NULL;
522 
523 	/*
524 	 * Detach res pages. svc_release must see a resused count of
525 	 * zero or it will attempt to put them.
526 	 */
527 	while (rqstp->rq_resused)
528 		rqstp->rq_respages[--rqstp->rq_resused] = NULL;
529 
530 	return err;
531 }
532 
rdma_read_complete(struct svc_rqst * rqstp,struct svc_rdma_op_ctxt * head)533 static int rdma_read_complete(struct svc_rqst *rqstp,
534 			      struct svc_rdma_op_ctxt *head)
535 {
536 	int page_no;
537 	int ret;
538 
539 	BUG_ON(!head);
540 
541 	/* Copy RPC pages */
542 	for (page_no = 0; page_no < head->count; page_no++) {
543 		put_page(rqstp->rq_pages[page_no]);
544 		rqstp->rq_pages[page_no] = head->pages[page_no];
545 	}
546 	/* Point rq_arg.pages past header */
547 	rqstp->rq_arg.pages = &rqstp->rq_pages[head->hdr_count];
548 	rqstp->rq_arg.page_len = head->arg.page_len;
549 	rqstp->rq_arg.page_base = head->arg.page_base;
550 
551 	/* rq_respages starts after the last arg page */
552 	rqstp->rq_respages = &rqstp->rq_arg.pages[page_no];
553 	rqstp->rq_resused = 0;
554 
555 	/* Rebuild rq_arg head and tail. */
556 	rqstp->rq_arg.head[0] = head->arg.head[0];
557 	rqstp->rq_arg.tail[0] = head->arg.tail[0];
558 	rqstp->rq_arg.len = head->arg.len;
559 	rqstp->rq_arg.buflen = head->arg.buflen;
560 
561 	/* Free the context */
562 	svc_rdma_put_context(head, 0);
563 
564 	/* XXX: What should this be? */
565 	rqstp->rq_prot = IPPROTO_MAX;
566 	svc_xprt_copy_addrs(rqstp, rqstp->rq_xprt);
567 
568 	ret = rqstp->rq_arg.head[0].iov_len
569 		+ rqstp->rq_arg.page_len
570 		+ rqstp->rq_arg.tail[0].iov_len;
571 	dprintk("svcrdma: deferred read ret=%d, rq_arg.len =%d, "
572 		"rq_arg.head[0].iov_base=%p, rq_arg.head[0].iov_len = %zd\n",
573 		ret, rqstp->rq_arg.len,	rqstp->rq_arg.head[0].iov_base,
574 		rqstp->rq_arg.head[0].iov_len);
575 
576 	return ret;
577 }
578 
579 /*
580  * Set up the rqstp thread context to point to the RQ buffer. If
581  * necessary, pull additional data from the client with an RDMA_READ
582  * request.
583  */
svc_rdma_recvfrom(struct svc_rqst * rqstp)584 int svc_rdma_recvfrom(struct svc_rqst *rqstp)
585 {
586 	struct svc_xprt *xprt = rqstp->rq_xprt;
587 	struct svcxprt_rdma *rdma_xprt =
588 		container_of(xprt, struct svcxprt_rdma, sc_xprt);
589 	struct svc_rdma_op_ctxt *ctxt = NULL;
590 	struct rpcrdma_msg *rmsgp;
591 	int ret = 0;
592 	int len;
593 
594 	dprintk("svcrdma: rqstp=%p\n", rqstp);
595 
596 	spin_lock_bh(&rdma_xprt->sc_rq_dto_lock);
597 	if (!list_empty(&rdma_xprt->sc_read_complete_q)) {
598 		ctxt = list_entry(rdma_xprt->sc_read_complete_q.next,
599 				  struct svc_rdma_op_ctxt,
600 				  dto_q);
601 		list_del_init(&ctxt->dto_q);
602 	}
603 	if (ctxt) {
604 		spin_unlock_bh(&rdma_xprt->sc_rq_dto_lock);
605 		return rdma_read_complete(rqstp, ctxt);
606 	}
607 
608 	if (!list_empty(&rdma_xprt->sc_rq_dto_q)) {
609 		ctxt = list_entry(rdma_xprt->sc_rq_dto_q.next,
610 				  struct svc_rdma_op_ctxt,
611 				  dto_q);
612 		list_del_init(&ctxt->dto_q);
613 	} else {
614 		atomic_inc(&rdma_stat_rq_starve);
615 		clear_bit(XPT_DATA, &xprt->xpt_flags);
616 		ctxt = NULL;
617 	}
618 	spin_unlock_bh(&rdma_xprt->sc_rq_dto_lock);
619 	if (!ctxt) {
620 		/* This is the EAGAIN path. The svc_recv routine will
621 		 * return -EAGAIN, the nfsd thread will go to call into
622 		 * svc_recv again and we shouldn't be on the active
623 		 * transport list
624 		 */
625 		if (test_bit(XPT_CLOSE, &xprt->xpt_flags))
626 			goto close_out;
627 
628 		BUG_ON(ret);
629 		goto out;
630 	}
631 	dprintk("svcrdma: processing ctxt=%p on xprt=%p, rqstp=%p, status=%d\n",
632 		ctxt, rdma_xprt, rqstp, ctxt->wc_status);
633 	BUG_ON(ctxt->wc_status != IB_WC_SUCCESS);
634 	atomic_inc(&rdma_stat_recv);
635 
636 	/* Build up the XDR from the receive buffers. */
637 	rdma_build_arg_xdr(rqstp, ctxt, ctxt->byte_len);
638 
639 	/* Decode the RDMA header. */
640 	len = svc_rdma_xdr_decode_req(&rmsgp, rqstp);
641 	rqstp->rq_xprt_hlen = len;
642 
643 	/* If the request is invalid, reply with an error */
644 	if (len < 0) {
645 		if (len == -ENOSYS)
646 			svc_rdma_send_error(rdma_xprt, rmsgp, ERR_VERS);
647 		goto close_out;
648 	}
649 
650 	/* Read read-list data. */
651 	ret = rdma_read_xdr(rdma_xprt, rmsgp, rqstp, ctxt);
652 	if (ret > 0) {
653 		/* read-list posted, defer until data received from client. */
654 		goto defer;
655 	}
656 	if (ret < 0) {
657 		/* Post of read-list failed, free context. */
658 		svc_rdma_put_context(ctxt, 1);
659 		return 0;
660 	}
661 
662 	ret = rqstp->rq_arg.head[0].iov_len
663 		+ rqstp->rq_arg.page_len
664 		+ rqstp->rq_arg.tail[0].iov_len;
665 	svc_rdma_put_context(ctxt, 0);
666  out:
667 	dprintk("svcrdma: ret = %d, rq_arg.len =%d, "
668 		"rq_arg.head[0].iov_base=%p, rq_arg.head[0].iov_len = %zd\n",
669 		ret, rqstp->rq_arg.len,
670 		rqstp->rq_arg.head[0].iov_base,
671 		rqstp->rq_arg.head[0].iov_len);
672 	rqstp->rq_prot = IPPROTO_MAX;
673 	svc_xprt_copy_addrs(rqstp, xprt);
674 	return ret;
675 
676  close_out:
677 	if (ctxt)
678 		svc_rdma_put_context(ctxt, 1);
679 	dprintk("svcrdma: transport %p is closing\n", xprt);
680 	/*
681 	 * Set the close bit and enqueue it. svc_recv will see the
682 	 * close bit and call svc_xprt_delete
683 	 */
684 	set_bit(XPT_CLOSE, &xprt->xpt_flags);
685 defer:
686 	return 0;
687 }
688