Root/net/rds/iw_recv.c

1/*
2 * Copyright (c) 2006 Oracle. 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
8 * OpenIB.org BSD license below:
9 *
10 * Redistribution and use in source and binary forms, with or
11 * without modification, are permitted provided that the following
12 * conditions are met:
13 *
14 * - Redistributions of source code must retain the above
15 * copyright notice, this list of conditions and the following
16 * disclaimer.
17 *
18 * - Redistributions in binary form must reproduce the above
19 * copyright notice, this list of conditions and the following
20 * disclaimer in the documentation and/or other materials
21 * provided with the distribution.
22 *
23 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
24 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
25 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
26 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
27 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
28 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
29 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
30 * SOFTWARE.
31 *
32 */
33#include <linux/kernel.h>
34#include <linux/slab.h>
35#include <linux/pci.h>
36#include <linux/dma-mapping.h>
37#include <rdma/rdma_cm.h>
38
39#include "rds.h"
40#include "iw.h"
41
42static struct kmem_cache *rds_iw_incoming_slab;
43static struct kmem_cache *rds_iw_frag_slab;
44static atomic_t rds_iw_allocation = ATOMIC_INIT(0);
45
46static void rds_iw_frag_drop_page(struct rds_page_frag *frag)
47{
48    rdsdebug("frag %p page %p\n", frag, frag->f_page);
49    __free_page(frag->f_page);
50    frag->f_page = NULL;
51}
52
53static void rds_iw_frag_free(struct rds_page_frag *frag)
54{
55    rdsdebug("frag %p page %p\n", frag, frag->f_page);
56    BUG_ON(frag->f_page != NULL);
57    kmem_cache_free(rds_iw_frag_slab, frag);
58}
59
60/*
61 * We map a page at a time. Its fragments are posted in order. This
62 * is called in fragment order as the fragments get send completion events.
63 * Only the last frag in the page performs the unmapping.
64 *
65 * It's OK for ring cleanup to call this in whatever order it likes because
66 * DMA is not in flight and so we can unmap while other ring entries still
67 * hold page references in their frags.
68 */
69static void rds_iw_recv_unmap_page(struct rds_iw_connection *ic,
70                   struct rds_iw_recv_work *recv)
71{
72    struct rds_page_frag *frag = recv->r_frag;
73
74    rdsdebug("recv %p frag %p page %p\n", recv, frag, frag->f_page);
75    if (frag->f_mapped)
76        ib_dma_unmap_page(ic->i_cm_id->device,
77                   frag->f_mapped,
78                   RDS_FRAG_SIZE, DMA_FROM_DEVICE);
79    frag->f_mapped = 0;
80}
81
82void rds_iw_recv_init_ring(struct rds_iw_connection *ic)
83{
84    struct rds_iw_recv_work *recv;
85    u32 i;
86
87    for (i = 0, recv = ic->i_recvs; i < ic->i_recv_ring.w_nr; i++, recv++) {
88        struct ib_sge *sge;
89
90        recv->r_iwinc = NULL;
91        recv->r_frag = NULL;
92
93        recv->r_wr.next = NULL;
94        recv->r_wr.wr_id = i;
95        recv->r_wr.sg_list = recv->r_sge;
96        recv->r_wr.num_sge = RDS_IW_RECV_SGE;
97
98        sge = rds_iw_data_sge(ic, recv->r_sge);
99        sge->addr = 0;
100        sge->length = RDS_FRAG_SIZE;
101        sge->lkey = 0;
102
103        sge = rds_iw_header_sge(ic, recv->r_sge);
104        sge->addr = ic->i_recv_hdrs_dma + (i * sizeof(struct rds_header));
105        sge->length = sizeof(struct rds_header);
106        sge->lkey = 0;
107    }
108}
109
110static void rds_iw_recv_clear_one(struct rds_iw_connection *ic,
111                  struct rds_iw_recv_work *recv)
112{
113    if (recv->r_iwinc) {
114        rds_inc_put(&recv->r_iwinc->ii_inc);
115        recv->r_iwinc = NULL;
116    }
117    if (recv->r_frag) {
118        rds_iw_recv_unmap_page(ic, recv);
119        if (recv->r_frag->f_page)
120            rds_iw_frag_drop_page(recv->r_frag);
121        rds_iw_frag_free(recv->r_frag);
122        recv->r_frag = NULL;
123    }
124}
125
126void rds_iw_recv_clear_ring(struct rds_iw_connection *ic)
127{
128    u32 i;
129
130    for (i = 0; i < ic->i_recv_ring.w_nr; i++)
131        rds_iw_recv_clear_one(ic, &ic->i_recvs[i]);
132
133    if (ic->i_frag.f_page)
134        rds_iw_frag_drop_page(&ic->i_frag);
135}
136
137static int rds_iw_recv_refill_one(struct rds_connection *conn,
138                  struct rds_iw_recv_work *recv,
139                  gfp_t kptr_gfp, gfp_t page_gfp)
140{
141    struct rds_iw_connection *ic = conn->c_transport_data;
142    dma_addr_t dma_addr;
143    struct ib_sge *sge;
144    int ret = -ENOMEM;
145
146    if (recv->r_iwinc == NULL) {
147        if (!atomic_add_unless(&rds_iw_allocation, 1, rds_iw_sysctl_max_recv_allocation)) {
148            rds_iw_stats_inc(s_iw_rx_alloc_limit);
149            goto out;
150        }
151        recv->r_iwinc = kmem_cache_alloc(rds_iw_incoming_slab,
152                         kptr_gfp);
153        if (recv->r_iwinc == NULL) {
154            atomic_dec(&rds_iw_allocation);
155            goto out;
156        }
157        INIT_LIST_HEAD(&recv->r_iwinc->ii_frags);
158        rds_inc_init(&recv->r_iwinc->ii_inc, conn, conn->c_faddr);
159    }
160
161    if (recv->r_frag == NULL) {
162        recv->r_frag = kmem_cache_alloc(rds_iw_frag_slab, kptr_gfp);
163        if (recv->r_frag == NULL)
164            goto out;
165        INIT_LIST_HEAD(&recv->r_frag->f_item);
166        recv->r_frag->f_page = NULL;
167    }
168
169    if (ic->i_frag.f_page == NULL) {
170        ic->i_frag.f_page = alloc_page(page_gfp);
171        if (ic->i_frag.f_page == NULL)
172            goto out;
173        ic->i_frag.f_offset = 0;
174    }
175
176    dma_addr = ib_dma_map_page(ic->i_cm_id->device,
177                  ic->i_frag.f_page,
178                  ic->i_frag.f_offset,
179                  RDS_FRAG_SIZE,
180                  DMA_FROM_DEVICE);
181    if (ib_dma_mapping_error(ic->i_cm_id->device, dma_addr))
182        goto out;
183
184    /*
185     * Once we get the RDS_PAGE_LAST_OFF frag then rds_iw_frag_unmap()
186     * must be called on this recv. This happens as completions hit
187     * in order or on connection shutdown.
188     */
189    recv->r_frag->f_page = ic->i_frag.f_page;
190    recv->r_frag->f_offset = ic->i_frag.f_offset;
191    recv->r_frag->f_mapped = dma_addr;
192
193    sge = rds_iw_data_sge(ic, recv->r_sge);
194    sge->addr = dma_addr;
195    sge->length = RDS_FRAG_SIZE;
196
197    sge = rds_iw_header_sge(ic, recv->r_sge);
198    sge->addr = ic->i_recv_hdrs_dma + (recv - ic->i_recvs) * sizeof(struct rds_header);
199    sge->length = sizeof(struct rds_header);
200
201    get_page(recv->r_frag->f_page);
202
203    if (ic->i_frag.f_offset < RDS_PAGE_LAST_OFF) {
204        ic->i_frag.f_offset += RDS_FRAG_SIZE;
205    } else {
206        put_page(ic->i_frag.f_page);
207        ic->i_frag.f_page = NULL;
208        ic->i_frag.f_offset = 0;
209    }
210
211    ret = 0;
212out:
213    return ret;
214}
215
216/*
217 * This tries to allocate and post unused work requests after making sure that
218 * they have all the allocations they need to queue received fragments into
219 * sockets. The i_recv_mutex is held here so that ring_alloc and _unalloc
220 * pairs don't go unmatched.
221 *
222 * -1 is returned if posting fails due to temporary resource exhaustion.
223 */
224int rds_iw_recv_refill(struct rds_connection *conn, gfp_t kptr_gfp,
225               gfp_t page_gfp, int prefill)
226{
227    struct rds_iw_connection *ic = conn->c_transport_data;
228    struct rds_iw_recv_work *recv;
229    struct ib_recv_wr *failed_wr;
230    unsigned int posted = 0;
231    int ret = 0;
232    u32 pos;
233
234    while ((prefill || rds_conn_up(conn)) &&
235           rds_iw_ring_alloc(&ic->i_recv_ring, 1, &pos)) {
236        if (pos >= ic->i_recv_ring.w_nr) {
237            printk(KERN_NOTICE "Argh - ring alloc returned pos=%u\n",
238                    pos);
239            ret = -EINVAL;
240            break;
241        }
242
243        recv = &ic->i_recvs[pos];
244        ret = rds_iw_recv_refill_one(conn, recv, kptr_gfp, page_gfp);
245        if (ret) {
246            ret = -1;
247            break;
248        }
249
250        /* XXX when can this fail? */
251        ret = ib_post_recv(ic->i_cm_id->qp, &recv->r_wr, &failed_wr);
252        rdsdebug("recv %p iwinc %p page %p addr %lu ret %d\n", recv,
253             recv->r_iwinc, recv->r_frag->f_page,
254             (long) recv->r_frag->f_mapped, ret);
255        if (ret) {
256            rds_iw_conn_error(conn, "recv post on "
257                   "%pI4 returned %d, disconnecting and "
258                   "reconnecting\n", &conn->c_faddr,
259                   ret);
260            ret = -1;
261            break;
262        }
263
264        posted++;
265    }
266
267    /* We're doing flow control - update the window. */
268    if (ic->i_flowctl && posted)
269        rds_iw_advertise_credits(conn, posted);
270
271    if (ret)
272        rds_iw_ring_unalloc(&ic->i_recv_ring, 1);
273    return ret;
274}
275
276void rds_iw_inc_purge(struct rds_incoming *inc)
277{
278    struct rds_iw_incoming *iwinc;
279    struct rds_page_frag *frag;
280    struct rds_page_frag *pos;
281
282    iwinc = container_of(inc, struct rds_iw_incoming, ii_inc);
283    rdsdebug("purging iwinc %p inc %p\n", iwinc, inc);
284
285    list_for_each_entry_safe(frag, pos, &iwinc->ii_frags, f_item) {
286        list_del_init(&frag->f_item);
287        rds_iw_frag_drop_page(frag);
288        rds_iw_frag_free(frag);
289    }
290}
291
292void rds_iw_inc_free(struct rds_incoming *inc)
293{
294    struct rds_iw_incoming *iwinc;
295
296    iwinc = container_of(inc, struct rds_iw_incoming, ii_inc);
297
298    rds_iw_inc_purge(inc);
299    rdsdebug("freeing iwinc %p inc %p\n", iwinc, inc);
300    BUG_ON(!list_empty(&iwinc->ii_frags));
301    kmem_cache_free(rds_iw_incoming_slab, iwinc);
302    atomic_dec(&rds_iw_allocation);
303    BUG_ON(atomic_read(&rds_iw_allocation) < 0);
304}
305
306int rds_iw_inc_copy_to_user(struct rds_incoming *inc, struct iovec *first_iov,
307                size_t size)
308{
309    struct rds_iw_incoming *iwinc;
310    struct rds_page_frag *frag;
311    struct iovec *iov = first_iov;
312    unsigned long to_copy;
313    unsigned long frag_off = 0;
314    unsigned long iov_off = 0;
315    int copied = 0;
316    int ret;
317    u32 len;
318
319    iwinc = container_of(inc, struct rds_iw_incoming, ii_inc);
320    frag = list_entry(iwinc->ii_frags.next, struct rds_page_frag, f_item);
321    len = be32_to_cpu(inc->i_hdr.h_len);
322
323    while (copied < size && copied < len) {
324        if (frag_off == RDS_FRAG_SIZE) {
325            frag = list_entry(frag->f_item.next,
326                      struct rds_page_frag, f_item);
327            frag_off = 0;
328        }
329        while (iov_off == iov->iov_len) {
330            iov_off = 0;
331            iov++;
332        }
333
334        to_copy = min(iov->iov_len - iov_off, RDS_FRAG_SIZE - frag_off);
335        to_copy = min_t(size_t, to_copy, size - copied);
336        to_copy = min_t(unsigned long, to_copy, len - copied);
337
338        rdsdebug("%lu bytes to user [%p, %zu] + %lu from frag "
339             "[%p, %lu] + %lu\n",
340             to_copy, iov->iov_base, iov->iov_len, iov_off,
341             frag->f_page, frag->f_offset, frag_off);
342
343        /* XXX needs + offset for multiple recvs per page */
344        ret = rds_page_copy_to_user(frag->f_page,
345                        frag->f_offset + frag_off,
346                        iov->iov_base + iov_off,
347                        to_copy);
348        if (ret) {
349            copied = ret;
350            break;
351        }
352
353        iov_off += to_copy;
354        frag_off += to_copy;
355        copied += to_copy;
356    }
357
358    return copied;
359}
360
361/* ic starts out kzalloc()ed */
362void rds_iw_recv_init_ack(struct rds_iw_connection *ic)
363{
364    struct ib_send_wr *wr = &ic->i_ack_wr;
365    struct ib_sge *sge = &ic->i_ack_sge;
366
367    sge->addr = ic->i_ack_dma;
368    sge->length = sizeof(struct rds_header);
369    sge->lkey = rds_iw_local_dma_lkey(ic);
370
371    wr->sg_list = sge;
372    wr->num_sge = 1;
373    wr->opcode = IB_WR_SEND;
374    wr->wr_id = RDS_IW_ACK_WR_ID;
375    wr->send_flags = IB_SEND_SIGNALED | IB_SEND_SOLICITED;
376}
377
378/*
379 * You'd think that with reliable IB connections you wouldn't need to ack
380 * messages that have been received. The problem is that IB hardware generates
381 * an ack message before it has DMAed the message into memory. This creates a
382 * potential message loss if the HCA is disabled for any reason between when it
383 * sends the ack and before the message is DMAed and processed. This is only a
384 * potential issue if another HCA is available for fail-over.
385 *
386 * When the remote host receives our ack they'll free the sent message from
387 * their send queue. To decrease the latency of this we always send an ack
388 * immediately after we've received messages.
389 *
390 * For simplicity, we only have one ack in flight at a time. This puts
391 * pressure on senders to have deep enough send queues to absorb the latency of
392 * a single ack frame being in flight. This might not be good enough.
393 *
394 * This is implemented by have a long-lived send_wr and sge which point to a
395 * statically allocated ack frame. This ack wr does not fall under the ring
396 * accounting that the tx and rx wrs do. The QP attribute specifically makes
397 * room for it beyond the ring size. Send completion notices its special
398 * wr_id and avoids working with the ring in that case.
399 */
400#ifndef KERNEL_HAS_ATOMIC64
401static void rds_iw_set_ack(struct rds_iw_connection *ic, u64 seq,
402                int ack_required)
403{
404    unsigned long flags;
405
406    spin_lock_irqsave(&ic->i_ack_lock, flags);
407    ic->i_ack_next = seq;
408    if (ack_required)
409        set_bit(IB_ACK_REQUESTED, &ic->i_ack_flags);
410    spin_unlock_irqrestore(&ic->i_ack_lock, flags);
411}
412
413static u64 rds_iw_get_ack(struct rds_iw_connection *ic)
414{
415    unsigned long flags;
416    u64 seq;
417
418    clear_bit(IB_ACK_REQUESTED, &ic->i_ack_flags);
419
420    spin_lock_irqsave(&ic->i_ack_lock, flags);
421    seq = ic->i_ack_next;
422    spin_unlock_irqrestore(&ic->i_ack_lock, flags);
423
424    return seq;
425}
426#else
427static void rds_iw_set_ack(struct rds_iw_connection *ic, u64 seq,
428                int ack_required)
429{
430    atomic64_set(&ic->i_ack_next, seq);
431    if (ack_required) {
432        smp_mb__before_clear_bit();
433        set_bit(IB_ACK_REQUESTED, &ic->i_ack_flags);
434    }
435}
436
437static u64 rds_iw_get_ack(struct rds_iw_connection *ic)
438{
439    clear_bit(IB_ACK_REQUESTED, &ic->i_ack_flags);
440    smp_mb__after_clear_bit();
441
442    return atomic64_read(&ic->i_ack_next);
443}
444#endif
445
446
447static void rds_iw_send_ack(struct rds_iw_connection *ic, unsigned int adv_credits)
448{
449    struct rds_header *hdr = ic->i_ack;
450    struct ib_send_wr *failed_wr;
451    u64 seq;
452    int ret;
453
454    seq = rds_iw_get_ack(ic);
455
456    rdsdebug("send_ack: ic %p ack %llu\n", ic, (unsigned long long) seq);
457    rds_message_populate_header(hdr, 0, 0, 0);
458    hdr->h_ack = cpu_to_be64(seq);
459    hdr->h_credit = adv_credits;
460    rds_message_make_checksum(hdr);
461    ic->i_ack_queued = jiffies;
462
463    ret = ib_post_send(ic->i_cm_id->qp, &ic->i_ack_wr, &failed_wr);
464    if (unlikely(ret)) {
465        /* Failed to send. Release the WR, and
466         * force another ACK.
467         */
468        clear_bit(IB_ACK_IN_FLIGHT, &ic->i_ack_flags);
469        set_bit(IB_ACK_REQUESTED, &ic->i_ack_flags);
470
471        rds_iw_stats_inc(s_iw_ack_send_failure);
472        /* Need to finesse this later. */
473        BUG();
474    } else
475        rds_iw_stats_inc(s_iw_ack_sent);
476}
477
478/*
479 * There are 3 ways of getting acknowledgements to the peer:
480 * 1. We call rds_iw_attempt_ack from the recv completion handler
481 * to send an ACK-only frame.
482 * However, there can be only one such frame in the send queue
483 * at any time, so we may have to postpone it.
484 * 2. When another (data) packet is transmitted while there's
485 * an ACK in the queue, we piggyback the ACK sequence number
486 * on the data packet.
487 * 3. If the ACK WR is done sending, we get called from the
488 * send queue completion handler, and check whether there's
489 * another ACK pending (postponed because the WR was on the
490 * queue). If so, we transmit it.
491 *
492 * We maintain 2 variables:
493 * - i_ack_flags, which keeps track of whether the ACK WR
494 * is currently in the send queue or not (IB_ACK_IN_FLIGHT)
495 * - i_ack_next, which is the last sequence number we received
496 *
497 * Potentially, send queue and receive queue handlers can run concurrently.
498 * It would be nice to not have to use a spinlock to synchronize things,
499 * but the one problem that rules this out is that 64bit updates are
500 * not atomic on all platforms. Things would be a lot simpler if
501 * we had atomic64 or maybe cmpxchg64 everywhere.
502 *
503 * Reconnecting complicates this picture just slightly. When we
504 * reconnect, we may be seeing duplicate packets. The peer
505 * is retransmitting them, because it hasn't seen an ACK for
506 * them. It is important that we ACK these.
507 *
508 * ACK mitigation adds a header flag "ACK_REQUIRED"; any packet with
509 * this flag set *MUST* be acknowledged immediately.
510 */
511
512/*
513 * When we get here, we're called from the recv queue handler.
514 * Check whether we ought to transmit an ACK.
515 */
516void rds_iw_attempt_ack(struct rds_iw_connection *ic)
517{
518    unsigned int adv_credits;
519
520    if (!test_bit(IB_ACK_REQUESTED, &ic->i_ack_flags))
521        return;
522
523    if (test_and_set_bit(IB_ACK_IN_FLIGHT, &ic->i_ack_flags)) {
524        rds_iw_stats_inc(s_iw_ack_send_delayed);
525        return;
526    }
527
528    /* Can we get a send credit? */
529    if (!rds_iw_send_grab_credits(ic, 1, &adv_credits, 0, RDS_MAX_ADV_CREDIT)) {
530        rds_iw_stats_inc(s_iw_tx_throttle);
531        clear_bit(IB_ACK_IN_FLIGHT, &ic->i_ack_flags);
532        return;
533    }
534
535    clear_bit(IB_ACK_REQUESTED, &ic->i_ack_flags);
536    rds_iw_send_ack(ic, adv_credits);
537}
538
539/*
540 * We get here from the send completion handler, when the
541 * adapter tells us the ACK frame was sent.
542 */
543void rds_iw_ack_send_complete(struct rds_iw_connection *ic)
544{
545    clear_bit(IB_ACK_IN_FLIGHT, &ic->i_ack_flags);
546    rds_iw_attempt_ack(ic);
547}
548
549/*
550 * This is called by the regular xmit code when it wants to piggyback
551 * an ACK on an outgoing frame.
552 */
553u64 rds_iw_piggyb_ack(struct rds_iw_connection *ic)
554{
555    if (test_and_clear_bit(IB_ACK_REQUESTED, &ic->i_ack_flags))
556        rds_iw_stats_inc(s_iw_ack_send_piggybacked);
557    return rds_iw_get_ack(ic);
558}
559
560/*
561 * It's kind of lame that we're copying from the posted receive pages into
562 * long-lived bitmaps. We could have posted the bitmaps and rdma written into
563 * them. But receiving new congestion bitmaps should be a *rare* event, so
564 * hopefully we won't need to invest that complexity in making it more
565 * efficient. By copying we can share a simpler core with TCP which has to
566 * copy.
567 */
568static void rds_iw_cong_recv(struct rds_connection *conn,
569                  struct rds_iw_incoming *iwinc)
570{
571    struct rds_cong_map *map;
572    unsigned int map_off;
573    unsigned int map_page;
574    struct rds_page_frag *frag;
575    unsigned long frag_off;
576    unsigned long to_copy;
577    unsigned long copied;
578    uint64_t uncongested = 0;
579    void *addr;
580
581    /* catch completely corrupt packets */
582    if (be32_to_cpu(iwinc->ii_inc.i_hdr.h_len) != RDS_CONG_MAP_BYTES)
583        return;
584
585    map = conn->c_fcong;
586    map_page = 0;
587    map_off = 0;
588
589    frag = list_entry(iwinc->ii_frags.next, struct rds_page_frag, f_item);
590    frag_off = 0;
591
592    copied = 0;
593
594    while (copied < RDS_CONG_MAP_BYTES) {
595        uint64_t *src, *dst;
596        unsigned int k;
597
598        to_copy = min(RDS_FRAG_SIZE - frag_off, PAGE_SIZE - map_off);
599        BUG_ON(to_copy & 7); /* Must be 64bit aligned. */
600
601        addr = kmap_atomic(frag->f_page, KM_SOFTIRQ0);
602
603        src = addr + frag_off;
604        dst = (void *)map->m_page_addrs[map_page] + map_off;
605        for (k = 0; k < to_copy; k += 8) {
606            /* Record ports that became uncongested, ie
607             * bits that changed from 0 to 1. */
608            uncongested |= ~(*src) & *dst;
609            *dst++ = *src++;
610        }
611        kunmap_atomic(addr, KM_SOFTIRQ0);
612
613        copied += to_copy;
614
615        map_off += to_copy;
616        if (map_off == PAGE_SIZE) {
617            map_off = 0;
618            map_page++;
619        }
620
621        frag_off += to_copy;
622        if (frag_off == RDS_FRAG_SIZE) {
623            frag = list_entry(frag->f_item.next,
624                      struct rds_page_frag, f_item);
625            frag_off = 0;
626        }
627    }
628
629    /* the congestion map is in little endian order */
630    uncongested = le64_to_cpu(uncongested);
631
632    rds_cong_map_updated(map, uncongested);
633}
634
635/*
636 * Rings are posted with all the allocations they'll need to queue the
637 * incoming message to the receiving socket so this can't fail.
638 * All fragments start with a header, so we can make sure we're not receiving
639 * garbage, and we can tell a small 8 byte fragment from an ACK frame.
640 */
641struct rds_iw_ack_state {
642    u64 ack_next;
643    u64 ack_recv;
644    unsigned int ack_required:1;
645    unsigned int ack_next_valid:1;
646    unsigned int ack_recv_valid:1;
647};
648
649static void rds_iw_process_recv(struct rds_connection *conn,
650                struct rds_iw_recv_work *recv, u32 byte_len,
651                struct rds_iw_ack_state *state)
652{
653    struct rds_iw_connection *ic = conn->c_transport_data;
654    struct rds_iw_incoming *iwinc = ic->i_iwinc;
655    struct rds_header *ihdr, *hdr;
656
657    /* XXX shut down the connection if port 0,0 are seen? */
658
659    rdsdebug("ic %p iwinc %p recv %p byte len %u\n", ic, iwinc, recv,
660         byte_len);
661
662    if (byte_len < sizeof(struct rds_header)) {
663        rds_iw_conn_error(conn, "incoming message "
664               "from %pI4 didn't inclue a "
665               "header, disconnecting and "
666               "reconnecting\n",
667               &conn->c_faddr);
668        return;
669    }
670    byte_len -= sizeof(struct rds_header);
671
672    ihdr = &ic->i_recv_hdrs[recv - ic->i_recvs];
673
674    /* Validate the checksum. */
675    if (!rds_message_verify_checksum(ihdr)) {
676        rds_iw_conn_error(conn, "incoming message "
677               "from %pI4 has corrupted header - "
678               "forcing a reconnect\n",
679               &conn->c_faddr);
680        rds_stats_inc(s_recv_drop_bad_checksum);
681        return;
682    }
683
684    /* Process the ACK sequence which comes with every packet */
685    state->ack_recv = be64_to_cpu(ihdr->h_ack);
686    state->ack_recv_valid = 1;
687
688    /* Process the credits update if there was one */
689    if (ihdr->h_credit)
690        rds_iw_send_add_credits(conn, ihdr->h_credit);
691
692    if (ihdr->h_sport == 0 && ihdr->h_dport == 0 && byte_len == 0) {
693        /* This is an ACK-only packet. The fact that it gets
694         * special treatment here is that historically, ACKs
695         * were rather special beasts.
696         */
697        rds_iw_stats_inc(s_iw_ack_received);
698
699        /*
700         * Usually the frags make their way on to incs and are then freed as
701         * the inc is freed. We don't go that route, so we have to drop the
702         * page ref ourselves. We can't just leave the page on the recv
703         * because that confuses the dma mapping of pages and each recv's use
704         * of a partial page. We can leave the frag, though, it will be
705         * reused.
706         *
707         * FIXME: Fold this into the code path below.
708         */
709        rds_iw_frag_drop_page(recv->r_frag);
710        return;
711    }
712
713    /*
714     * If we don't already have an inc on the connection then this
715     * fragment has a header and starts a message.. copy its header
716     * into the inc and save the inc so we can hang upcoming fragments
717     * off its list.
718     */
719    if (iwinc == NULL) {
720        iwinc = recv->r_iwinc;
721        recv->r_iwinc = NULL;
722        ic->i_iwinc = iwinc;
723
724        hdr = &iwinc->ii_inc.i_hdr;
725        memcpy(hdr, ihdr, sizeof(*hdr));
726        ic->i_recv_data_rem = be32_to_cpu(hdr->h_len);
727
728        rdsdebug("ic %p iwinc %p rem %u flag 0x%x\n", ic, iwinc,
729             ic->i_recv_data_rem, hdr->h_flags);
730    } else {
731        hdr = &iwinc->ii_inc.i_hdr;
732        /* We can't just use memcmp here; fragments of a
733         * single message may carry different ACKs */
734        if (hdr->h_sequence != ihdr->h_sequence ||
735            hdr->h_len != ihdr->h_len ||
736            hdr->h_sport != ihdr->h_sport ||
737            hdr->h_dport != ihdr->h_dport) {
738            rds_iw_conn_error(conn,
739                "fragment header mismatch; forcing reconnect\n");
740            return;
741        }
742    }
743
744    list_add_tail(&recv->r_frag->f_item, &iwinc->ii_frags);
745    recv->r_frag = NULL;
746
747    if (ic->i_recv_data_rem > RDS_FRAG_SIZE)
748        ic->i_recv_data_rem -= RDS_FRAG_SIZE;
749    else {
750        ic->i_recv_data_rem = 0;
751        ic->i_iwinc = NULL;
752
753        if (iwinc->ii_inc.i_hdr.h_flags == RDS_FLAG_CONG_BITMAP)
754            rds_iw_cong_recv(conn, iwinc);
755        else {
756            rds_recv_incoming(conn, conn->c_faddr, conn->c_laddr,
757                      &iwinc->ii_inc, GFP_ATOMIC,
758                      KM_SOFTIRQ0);
759            state->ack_next = be64_to_cpu(hdr->h_sequence);
760            state->ack_next_valid = 1;
761        }
762
763        /* Evaluate the ACK_REQUIRED flag *after* we received
764         * the complete frame, and after bumping the next_rx
765         * sequence. */
766        if (hdr->h_flags & RDS_FLAG_ACK_REQUIRED) {
767            rds_stats_inc(s_recv_ack_required);
768            state->ack_required = 1;
769        }
770
771        rds_inc_put(&iwinc->ii_inc);
772    }
773}
774
775/*
776 * Plucking the oldest entry from the ring can be done concurrently with
777 * the thread refilling the ring. Each ring operation is protected by
778 * spinlocks and the transient state of refilling doesn't change the
779 * recording of which entry is oldest.
780 *
781 * This relies on IB only calling one cq comp_handler for each cq so that
782 * there will only be one caller of rds_recv_incoming() per RDS connection.
783 */
784void rds_iw_recv_cq_comp_handler(struct ib_cq *cq, void *context)
785{
786    struct rds_connection *conn = context;
787    struct rds_iw_connection *ic = conn->c_transport_data;
788
789    rdsdebug("conn %p cq %p\n", conn, cq);
790
791    rds_iw_stats_inc(s_iw_rx_cq_call);
792
793    tasklet_schedule(&ic->i_recv_tasklet);
794}
795
796static inline void rds_poll_cq(struct rds_iw_connection *ic,
797                   struct rds_iw_ack_state *state)
798{
799    struct rds_connection *conn = ic->conn;
800    struct ib_wc wc;
801    struct rds_iw_recv_work *recv;
802
803    while (ib_poll_cq(ic->i_recv_cq, 1, &wc) > 0) {
804        rdsdebug("wc wr_id 0x%llx status %u byte_len %u imm_data %u\n",
805             (unsigned long long)wc.wr_id, wc.status, wc.byte_len,
806             be32_to_cpu(wc.ex.imm_data));
807        rds_iw_stats_inc(s_iw_rx_cq_event);
808
809        recv = &ic->i_recvs[rds_iw_ring_oldest(&ic->i_recv_ring)];
810
811        rds_iw_recv_unmap_page(ic, recv);
812
813        /*
814         * Also process recvs in connecting state because it is possible
815         * to get a recv completion _before_ the rdmacm ESTABLISHED
816         * event is processed.
817         */
818        if (rds_conn_up(conn) || rds_conn_connecting(conn)) {
819            /* We expect errors as the qp is drained during shutdown */
820            if (wc.status == IB_WC_SUCCESS) {
821                rds_iw_process_recv(conn, recv, wc.byte_len, state);
822            } else {
823                rds_iw_conn_error(conn, "recv completion on "
824                       "%pI4 had status %u, disconnecting and "
825                       "reconnecting\n", &conn->c_faddr,
826                       wc.status);
827            }
828        }
829
830        rds_iw_ring_free(&ic->i_recv_ring, 1);
831    }
832}
833
834void rds_iw_recv_tasklet_fn(unsigned long data)
835{
836    struct rds_iw_connection *ic = (struct rds_iw_connection *) data;
837    struct rds_connection *conn = ic->conn;
838    struct rds_iw_ack_state state = { 0, };
839
840    rds_poll_cq(ic, &state);
841    ib_req_notify_cq(ic->i_recv_cq, IB_CQ_SOLICITED);
842    rds_poll_cq(ic, &state);
843
844    if (state.ack_next_valid)
845        rds_iw_set_ack(ic, state.ack_next, state.ack_required);
846    if (state.ack_recv_valid && state.ack_recv > ic->i_ack_recv) {
847        rds_send_drop_acked(conn, state.ack_recv, NULL);
848        ic->i_ack_recv = state.ack_recv;
849    }
850    if (rds_conn_up(conn))
851        rds_iw_attempt_ack(ic);
852
853    /* If we ever end up with a really empty receive ring, we're
854     * in deep trouble, as the sender will definitely see RNR
855     * timeouts. */
856    if (rds_iw_ring_empty(&ic->i_recv_ring))
857        rds_iw_stats_inc(s_iw_rx_ring_empty);
858
859    /*
860     * If the ring is running low, then schedule the thread to refill.
861     */
862    if (rds_iw_ring_low(&ic->i_recv_ring))
863        queue_delayed_work(rds_wq, &conn->c_recv_w, 0);
864}
865
866int rds_iw_recv(struct rds_connection *conn)
867{
868    struct rds_iw_connection *ic = conn->c_transport_data;
869    int ret = 0;
870
871    rdsdebug("conn %p\n", conn);
872
873    /*
874     * If we get a temporary posting failure in this context then
875     * we're really low and we want the caller to back off for a bit.
876     */
877    mutex_lock(&ic->i_recv_mutex);
878    if (rds_iw_recv_refill(conn, GFP_KERNEL, GFP_HIGHUSER, 0))
879        ret = -ENOMEM;
880    else
881        rds_iw_stats_inc(s_iw_rx_refill_from_thread);
882    mutex_unlock(&ic->i_recv_mutex);
883
884    if (rds_conn_up(conn))
885        rds_iw_attempt_ack(ic);
886
887    return ret;
888}
889
890int __init rds_iw_recv_init(void)
891{
892    struct sysinfo si;
893    int ret = -ENOMEM;
894
895    /* Default to 30% of all available RAM for recv memory */
896    si_meminfo(&si);
897    rds_iw_sysctl_max_recv_allocation = si.totalram / 3 * PAGE_SIZE / RDS_FRAG_SIZE;
898
899    rds_iw_incoming_slab = kmem_cache_create("rds_iw_incoming",
900                    sizeof(struct rds_iw_incoming),
901                    0, 0, NULL);
902    if (rds_iw_incoming_slab == NULL)
903        goto out;
904
905    rds_iw_frag_slab = kmem_cache_create("rds_iw_frag",
906                    sizeof(struct rds_page_frag),
907                    0, 0, NULL);
908    if (rds_iw_frag_slab == NULL)
909        kmem_cache_destroy(rds_iw_incoming_slab);
910    else
911        ret = 0;
912out:
913    return ret;
914}
915
916void rds_iw_recv_exit(void)
917{
918    kmem_cache_destroy(rds_iw_incoming_slab);
919    kmem_cache_destroy(rds_iw_frag_slab);
920}
921

Archive Download this file



interactive