Merge git://git.infradead.org/battery-2.6
[sfrench/cifs-2.6.git] / net / sunrpc / xprtrdma / svc_rdma_recvfrom.c
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  */
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 struct chunk_sge {
116         int start;              /* sge no for this chunk */
117         int count;              /* sge count for this chunk */
118 };
119
120 /* Encode a read-chunk-list as an array of IB SGE
121  *
122  * Assumptions:
123  * - chunk[0]->position points to pages[0] at an offset of 0
124  * - pages[] is not physically or virtually contigous and consists of
125  *   PAGE_SIZE elements.
126  *
127  * Output:
128  * - sge array pointing into pages[] array.
129  * - chunk_sge array specifying sge index and count for each
130  *   chunk in the read list
131  *
132  */
133 static int rdma_rcl_to_sge(struct svcxprt_rdma *xprt,
134                            struct svc_rqst *rqstp,
135                            struct svc_rdma_op_ctxt *head,
136                            struct rpcrdma_msg *rmsgp,
137                            struct ib_sge *sge,
138                            struct chunk_sge *ch_sge_ary,
139                            int ch_count,
140                            int byte_count)
141 {
142         int sge_no;
143         int sge_bytes;
144         int page_off;
145         int page_no;
146         int ch_bytes;
147         int ch_no;
148         struct rpcrdma_read_chunk *ch;
149
150         sge_no = 0;
151         page_no = 0;
152         page_off = 0;
153         ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0];
154         ch_no = 0;
155         ch_bytes = ch->rc_target.rs_length;
156         head->arg.head[0] = rqstp->rq_arg.head[0];
157         head->arg.tail[0] = rqstp->rq_arg.tail[0];
158         head->arg.pages = &head->pages[head->count];
159         head->sge[0].length = head->count; /* save count of hdr pages */
160         head->arg.page_base = 0;
161         head->arg.page_len = ch_bytes;
162         head->arg.len = rqstp->rq_arg.len + ch_bytes;
163         head->arg.buflen = rqstp->rq_arg.buflen + ch_bytes;
164         head->count++;
165         ch_sge_ary[0].start = 0;
166         while (byte_count) {
167                 sge_bytes = min_t(int, PAGE_SIZE-page_off, ch_bytes);
168                 sge[sge_no].addr =
169                         ib_dma_map_page(xprt->sc_cm_id->device,
170                                         rqstp->rq_arg.pages[page_no],
171                                         page_off, sge_bytes,
172                                         DMA_FROM_DEVICE);
173                 sge[sge_no].length = sge_bytes;
174                 sge[sge_no].lkey = xprt->sc_phys_mr->lkey;
175                 /*
176                  * Don't bump head->count here because the same page
177                  * may be used by multiple SGE.
178                  */
179                 head->arg.pages[page_no] = rqstp->rq_arg.pages[page_no];
180                 rqstp->rq_respages = &rqstp->rq_arg.pages[page_no+1];
181
182                 byte_count -= sge_bytes;
183                 ch_bytes -= sge_bytes;
184                 sge_no++;
185                 /*
186                  * If all bytes for this chunk have been mapped to an
187                  * SGE, move to the next SGE
188                  */
189                 if (ch_bytes == 0) {
190                         ch_sge_ary[ch_no].count =
191                                 sge_no - ch_sge_ary[ch_no].start;
192                         ch_no++;
193                         ch++;
194                         ch_sge_ary[ch_no].start = sge_no;
195                         ch_bytes = ch->rc_target.rs_length;
196                         /* If bytes remaining account for next chunk */
197                         if (byte_count) {
198                                 head->arg.page_len += ch_bytes;
199                                 head->arg.len += ch_bytes;
200                                 head->arg.buflen += ch_bytes;
201                         }
202                 }
203                 /*
204                  * If this SGE consumed all of the page, move to the
205                  * next page
206                  */
207                 if ((sge_bytes + page_off) == PAGE_SIZE) {
208                         page_no++;
209                         page_off = 0;
210                         /*
211                          * If there are still bytes left to map, bump
212                          * the page count
213                          */
214                         if (byte_count)
215                                 head->count++;
216                 } else
217                         page_off += sge_bytes;
218         }
219         BUG_ON(byte_count != 0);
220         return sge_no;
221 }
222
223 static void rdma_set_ctxt_sge(struct svc_rdma_op_ctxt *ctxt,
224                               struct ib_sge *sge,
225                               u64 *sgl_offset,
226                               int count)
227 {
228         int i;
229
230         ctxt->count = count;
231         for (i = 0; i < count; i++) {
232                 ctxt->sge[i].addr = sge[i].addr;
233                 ctxt->sge[i].length = sge[i].length;
234                 *sgl_offset = *sgl_offset + sge[i].length;
235         }
236 }
237
238 static int rdma_read_max_sge(struct svcxprt_rdma *xprt, int sge_count)
239 {
240 #ifdef RDMA_TRANSPORT_IWARP
241         if ((RDMA_TRANSPORT_IWARP ==
242              rdma_node_get_transport(xprt->sc_cm_id->
243                                      device->node_type))
244             && sge_count > 1)
245                 return 1;
246         else
247 #endif
248                 return min_t(int, sge_count, xprt->sc_max_sge);
249 }
250
251 /*
252  * Use RDMA_READ to read data from the advertised client buffer into the
253  * XDR stream starting at rq_arg.head[0].iov_base.
254  * Each chunk in the array
255  * contains the following fields:
256  * discrim      - '1', This isn't used for data placement
257  * position     - The xdr stream offset (the same for every chunk)
258  * handle       - RMR for client memory region
259  * length       - data transfer length
260  * offset       - 64 bit tagged offset in remote memory region
261  *
262  * On our side, we need to read into a pagelist. The first page immediately
263  * follows the RPC header.
264  *
265  * This function returns 1 to indicate success. The data is not yet in
266  * the pagelist and therefore the RPC request must be deferred. The
267  * I/O completion will enqueue the transport again and
268  * svc_rdma_recvfrom will complete the request.
269  *
270  * NOTE: The ctxt must not be touched after the last WR has been posted
271  * because the I/O completion processing may occur on another
272  * processor and free / modify the context. Ne touche pas!
273  */
274 static int rdma_read_xdr(struct svcxprt_rdma *xprt,
275                          struct rpcrdma_msg *rmsgp,
276                          struct svc_rqst *rqstp,
277                          struct svc_rdma_op_ctxt *hdr_ctxt)
278 {
279         struct ib_send_wr read_wr;
280         int err = 0;
281         int ch_no;
282         struct ib_sge *sge;
283         int ch_count;
284         int byte_count;
285         int sge_count;
286         u64 sgl_offset;
287         struct rpcrdma_read_chunk *ch;
288         struct svc_rdma_op_ctxt *ctxt = NULL;
289         struct svc_rdma_op_ctxt *head;
290         struct svc_rdma_op_ctxt *tmp_sge_ctxt;
291         struct svc_rdma_op_ctxt *tmp_ch_ctxt;
292         struct chunk_sge *ch_sge_ary;
293
294         /* If no read list is present, return 0 */
295         ch = svc_rdma_get_read_chunk(rmsgp);
296         if (!ch)
297                 return 0;
298
299         /* Allocate temporary contexts to keep SGE */
300         BUG_ON(sizeof(struct ib_sge) < sizeof(struct chunk_sge));
301         tmp_sge_ctxt = svc_rdma_get_context(xprt);
302         sge = tmp_sge_ctxt->sge;
303         tmp_ch_ctxt = svc_rdma_get_context(xprt);
304         ch_sge_ary = (struct chunk_sge *)tmp_ch_ctxt->sge;
305
306         svc_rdma_rcl_chunk_counts(ch, &ch_count, &byte_count);
307         sge_count = rdma_rcl_to_sge(xprt, rqstp, hdr_ctxt, rmsgp,
308                                     sge, ch_sge_ary,
309                                     ch_count, byte_count);
310         head = svc_rdma_get_context(xprt);
311         sgl_offset = 0;
312         ch_no = 0;
313
314         for (ch = (struct rpcrdma_read_chunk *)&rmsgp->rm_body.rm_chunks[0];
315              ch->rc_discrim != 0; ch++, ch_no++) {
316 next_sge:
317                 if (!ctxt)
318                         ctxt = head;
319                 else {
320                         ctxt->next = svc_rdma_get_context(xprt);
321                         ctxt = ctxt->next;
322                 }
323                 ctxt->next = NULL;
324                 ctxt->direction = DMA_FROM_DEVICE;
325                 clear_bit(RDMACTXT_F_READ_DONE, &ctxt->flags);
326                 clear_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags);
327                 if ((ch+1)->rc_discrim == 0) {
328                         /*
329                          * Checked in sq_cq_reap to see if we need to
330                          * be enqueued
331                          */
332                         set_bit(RDMACTXT_F_LAST_CTXT, &ctxt->flags);
333                         ctxt->next = hdr_ctxt;
334                         hdr_ctxt->next = head;
335                 }
336
337                 /* Prepare READ WR */
338                 memset(&read_wr, 0, sizeof read_wr);
339                 ctxt->wr_op = IB_WR_RDMA_READ;
340                 read_wr.wr_id = (unsigned long)ctxt;
341                 read_wr.opcode = IB_WR_RDMA_READ;
342                 read_wr.send_flags = IB_SEND_SIGNALED;
343                 read_wr.wr.rdma.rkey = ch->rc_target.rs_handle;
344                 read_wr.wr.rdma.remote_addr =
345                         get_unaligned(&(ch->rc_target.rs_offset)) +
346                         sgl_offset;
347                 read_wr.sg_list = &sge[ch_sge_ary[ch_no].start];
348                 read_wr.num_sge =
349                         rdma_read_max_sge(xprt, ch_sge_ary[ch_no].count);
350                 rdma_set_ctxt_sge(ctxt, &sge[ch_sge_ary[ch_no].start],
351                                   &sgl_offset,
352                                   read_wr.num_sge);
353
354                 /* Post the read */
355                 err = svc_rdma_send(xprt, &read_wr);
356                 if (err) {
357                         printk(KERN_ERR "svcrdma: Error posting send = %d\n",
358                                err);
359                         /*
360                          * Break the circular list so free knows when
361                          * to stop if the error happened to occur on
362                          * the last read
363                          */
364                         ctxt->next = NULL;
365                         goto out;
366                 }
367                 atomic_inc(&rdma_stat_read);
368
369                 if (read_wr.num_sge < ch_sge_ary[ch_no].count) {
370                         ch_sge_ary[ch_no].count -= read_wr.num_sge;
371                         ch_sge_ary[ch_no].start += read_wr.num_sge;
372                         goto next_sge;
373                 }
374                 sgl_offset = 0;
375                 err = 0;
376         }
377
378  out:
379         svc_rdma_put_context(tmp_sge_ctxt, 0);
380         svc_rdma_put_context(tmp_ch_ctxt, 0);
381
382         /* Detach arg pages. svc_recv will replenish them */
383         for (ch_no = 0; &rqstp->rq_pages[ch_no] < rqstp->rq_respages; ch_no++)
384                 rqstp->rq_pages[ch_no] = NULL;
385
386         /*
387          * Detach res pages. svc_release must see a resused count of
388          * zero or it will attempt to put them.
389          */
390         while (rqstp->rq_resused)
391                 rqstp->rq_respages[--rqstp->rq_resused] = NULL;
392
393         if (err) {
394                 printk(KERN_ERR "svcrdma : RDMA_READ error = %d\n", err);
395                 set_bit(XPT_CLOSE, &xprt->sc_xprt.xpt_flags);
396                 /* Free the linked list of read contexts */
397                 while (head != NULL) {
398                         ctxt = head->next;
399                         svc_rdma_put_context(head, 1);
400                         head = ctxt;
401                 }
402                 return 0;
403         }
404
405         return 1;
406 }
407
408 static int rdma_read_complete(struct svc_rqst *rqstp,
409                               struct svc_rdma_op_ctxt *data)
410 {
411         struct svc_rdma_op_ctxt *head = data->next;
412         int page_no;
413         int ret;
414
415         BUG_ON(!head);
416
417         /* Copy RPC pages */
418         for (page_no = 0; page_no < head->count; page_no++) {
419                 put_page(rqstp->rq_pages[page_no]);
420                 rqstp->rq_pages[page_no] = head->pages[page_no];
421         }
422         /* Point rq_arg.pages past header */
423         rqstp->rq_arg.pages = &rqstp->rq_pages[head->sge[0].length];
424         rqstp->rq_arg.page_len = head->arg.page_len;
425         rqstp->rq_arg.page_base = head->arg.page_base;
426
427         /* rq_respages starts after the last arg page */
428         rqstp->rq_respages = &rqstp->rq_arg.pages[page_no];
429         rqstp->rq_resused = 0;
430
431         /* Rebuild rq_arg head and tail. */
432         rqstp->rq_arg.head[0] = head->arg.head[0];
433         rqstp->rq_arg.tail[0] = head->arg.tail[0];
434         rqstp->rq_arg.len = head->arg.len;
435         rqstp->rq_arg.buflen = head->arg.buflen;
436
437         /* XXX: What should this be? */
438         rqstp->rq_prot = IPPROTO_MAX;
439
440         /*
441          * Free the contexts we used to build the RDMA_READ. We have
442          * to be careful here because the context list uses the same
443          * next pointer used to chain the contexts associated with the
444          * RDMA_READ
445          */
446         data->next = NULL;      /* terminate circular list */
447         do {
448                 data = head->next;
449                 svc_rdma_put_context(head, 0);
450                 head = data;
451         } while (head != NULL);
452
453         ret = rqstp->rq_arg.head[0].iov_len
454                 + rqstp->rq_arg.page_len
455                 + rqstp->rq_arg.tail[0].iov_len;
456         dprintk("svcrdma: deferred read ret=%d, rq_arg.len =%d, "
457                 "rq_arg.head[0].iov_base=%p, rq_arg.head[0].iov_len = %zd\n",
458                 ret, rqstp->rq_arg.len, rqstp->rq_arg.head[0].iov_base,
459                 rqstp->rq_arg.head[0].iov_len);
460
461         /* Indicate that we've consumed an RQ credit */
462         rqstp->rq_xprt_ctxt = rqstp->rq_xprt;
463         svc_xprt_received(rqstp->rq_xprt);
464         return ret;
465 }
466
467 /*
468  * Set up the rqstp thread context to point to the RQ buffer. If
469  * necessary, pull additional data from the client with an RDMA_READ
470  * request.
471  */
472 int svc_rdma_recvfrom(struct svc_rqst *rqstp)
473 {
474         struct svc_xprt *xprt = rqstp->rq_xprt;
475         struct svcxprt_rdma *rdma_xprt =
476                 container_of(xprt, struct svcxprt_rdma, sc_xprt);
477         struct svc_rdma_op_ctxt *ctxt = NULL;
478         struct rpcrdma_msg *rmsgp;
479         int ret = 0;
480         int len;
481
482         dprintk("svcrdma: rqstp=%p\n", rqstp);
483
484         /*
485          * The rq_xprt_ctxt indicates if we've consumed an RQ credit
486          * or not. It is used in the rdma xpo_release_rqst function to
487          * determine whether or not to return an RQ WQE to the RQ.
488          */
489         rqstp->rq_xprt_ctxt = NULL;
490
491         spin_lock_bh(&rdma_xprt->sc_read_complete_lock);
492         if (!list_empty(&rdma_xprt->sc_read_complete_q)) {
493                 ctxt = list_entry(rdma_xprt->sc_read_complete_q.next,
494                                   struct svc_rdma_op_ctxt,
495                                   dto_q);
496                 list_del_init(&ctxt->dto_q);
497         }
498         spin_unlock_bh(&rdma_xprt->sc_read_complete_lock);
499         if (ctxt)
500                 return rdma_read_complete(rqstp, ctxt);
501
502         spin_lock_bh(&rdma_xprt->sc_rq_dto_lock);
503         if (!list_empty(&rdma_xprt->sc_rq_dto_q)) {
504                 ctxt = list_entry(rdma_xprt->sc_rq_dto_q.next,
505                                   struct svc_rdma_op_ctxt,
506                                   dto_q);
507                 list_del_init(&ctxt->dto_q);
508         } else {
509                 atomic_inc(&rdma_stat_rq_starve);
510                 clear_bit(XPT_DATA, &xprt->xpt_flags);
511                 ctxt = NULL;
512         }
513         spin_unlock_bh(&rdma_xprt->sc_rq_dto_lock);
514         if (!ctxt) {
515                 /* This is the EAGAIN path. The svc_recv routine will
516                  * return -EAGAIN, the nfsd thread will go to call into
517                  * svc_recv again and we shouldn't be on the active
518                  * transport list
519                  */
520                 if (test_bit(XPT_CLOSE, &xprt->xpt_flags))
521                         goto close_out;
522
523                 BUG_ON(ret);
524                 goto out;
525         }
526         dprintk("svcrdma: processing ctxt=%p on xprt=%p, rqstp=%p, status=%d\n",
527                 ctxt, rdma_xprt, rqstp, ctxt->wc_status);
528         BUG_ON(ctxt->wc_status != IB_WC_SUCCESS);
529         atomic_inc(&rdma_stat_recv);
530
531         /* Build up the XDR from the receive buffers. */
532         rdma_build_arg_xdr(rqstp, ctxt, ctxt->byte_len);
533
534         /* Decode the RDMA header. */
535         len = svc_rdma_xdr_decode_req(&rmsgp, rqstp);
536         rqstp->rq_xprt_hlen = len;
537
538         /* If the request is invalid, reply with an error */
539         if (len < 0) {
540                 if (len == -ENOSYS)
541                         (void)svc_rdma_send_error(rdma_xprt, rmsgp, ERR_VERS);
542                 goto close_out;
543         }
544
545         /* Read read-list data. If we would need to wait, defer
546          * it. Not that in this case, we don't return the RQ credit
547          * until after the read completes.
548          */
549         if (rdma_read_xdr(rdma_xprt, rmsgp, rqstp, ctxt)) {
550                 svc_xprt_received(xprt);
551                 return 0;
552         }
553
554         /* Indicate we've consumed an RQ credit */
555         rqstp->rq_xprt_ctxt = rqstp->rq_xprt;
556
557         ret = rqstp->rq_arg.head[0].iov_len
558                 + rqstp->rq_arg.page_len
559                 + rqstp->rq_arg.tail[0].iov_len;
560         svc_rdma_put_context(ctxt, 0);
561  out:
562         dprintk("svcrdma: ret = %d, rq_arg.len =%d, "
563                 "rq_arg.head[0].iov_base=%p, rq_arg.head[0].iov_len = %zd\n",
564                 ret, rqstp->rq_arg.len,
565                 rqstp->rq_arg.head[0].iov_base,
566                 rqstp->rq_arg.head[0].iov_len);
567         rqstp->rq_prot = IPPROTO_MAX;
568         svc_xprt_copy_addrs(rqstp, xprt);
569         svc_xprt_received(xprt);
570         return ret;
571
572  close_out:
573         if (ctxt) {
574                 svc_rdma_put_context(ctxt, 1);
575                 /* Indicate we've consumed an RQ credit */
576                 rqstp->rq_xprt_ctxt = rqstp->rq_xprt;
577         }
578         dprintk("svcrdma: transport %p is closing\n", xprt);
579         /*
580          * Set the close bit and enqueue it. svc_recv will see the
581          * close bit and call svc_xprt_delete
582          */
583         set_bit(XPT_CLOSE, &xprt->xpt_flags);
584         svc_xprt_received(xprt);
585         return 0;
586 }