1// SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB
2/*
3 * Copyright (c) 2016 Mellanox Technologies Ltd. All rights reserved.
4 * Copyright (c) 2015 System Fabric Works, Inc. All rights reserved.
5 */
6
7#include <linux/skbuff.h>
8
9#include "rxe.h"
10#include "rxe_loc.h"
11#include "rxe_queue.h"
12#include "rxe_task.h"
13
14enum comp_state {
15	COMPST_GET_ACK,
16	COMPST_GET_WQE,
17	COMPST_COMP_WQE,
18	COMPST_COMP_ACK,
19	COMPST_CHECK_PSN,
20	COMPST_CHECK_ACK,
21	COMPST_READ,
22	COMPST_ATOMIC,
23	COMPST_WRITE_SEND,
24	COMPST_UPDATE_COMP,
25	COMPST_ERROR_RETRY,
26	COMPST_RNR_RETRY,
27	COMPST_ERROR,
28	COMPST_EXIT, /* We have an issue, and we want to rerun the completer */
29	COMPST_DONE, /* The completer finished successflly */
30};
31
32static char *comp_state_name[] =  {
33	[COMPST_GET_ACK]		= "GET ACK",
34	[COMPST_GET_WQE]		= "GET WQE",
35	[COMPST_COMP_WQE]		= "COMP WQE",
36	[COMPST_COMP_ACK]		= "COMP ACK",
37	[COMPST_CHECK_PSN]		= "CHECK PSN",
38	[COMPST_CHECK_ACK]		= "CHECK ACK",
39	[COMPST_READ]			= "READ",
40	[COMPST_ATOMIC]			= "ATOMIC",
41	[COMPST_WRITE_SEND]		= "WRITE/SEND",
42	[COMPST_UPDATE_COMP]		= "UPDATE COMP",
43	[COMPST_ERROR_RETRY]		= "ERROR RETRY",
44	[COMPST_RNR_RETRY]		= "RNR RETRY",
45	[COMPST_ERROR]			= "ERROR",
46	[COMPST_EXIT]			= "EXIT",
47	[COMPST_DONE]			= "DONE",
48};
49
50static unsigned long rnrnak_usec[32] = {
51	[IB_RNR_TIMER_655_36] = 655360,
52	[IB_RNR_TIMER_000_01] = 10,
53	[IB_RNR_TIMER_000_02] = 20,
54	[IB_RNR_TIMER_000_03] = 30,
55	[IB_RNR_TIMER_000_04] = 40,
56	[IB_RNR_TIMER_000_06] = 60,
57	[IB_RNR_TIMER_000_08] = 80,
58	[IB_RNR_TIMER_000_12] = 120,
59	[IB_RNR_TIMER_000_16] = 160,
60	[IB_RNR_TIMER_000_24] = 240,
61	[IB_RNR_TIMER_000_32] = 320,
62	[IB_RNR_TIMER_000_48] = 480,
63	[IB_RNR_TIMER_000_64] = 640,
64	[IB_RNR_TIMER_000_96] = 960,
65	[IB_RNR_TIMER_001_28] = 1280,
66	[IB_RNR_TIMER_001_92] = 1920,
67	[IB_RNR_TIMER_002_56] = 2560,
68	[IB_RNR_TIMER_003_84] = 3840,
69	[IB_RNR_TIMER_005_12] = 5120,
70	[IB_RNR_TIMER_007_68] = 7680,
71	[IB_RNR_TIMER_010_24] = 10240,
72	[IB_RNR_TIMER_015_36] = 15360,
73	[IB_RNR_TIMER_020_48] = 20480,
74	[IB_RNR_TIMER_030_72] = 30720,
75	[IB_RNR_TIMER_040_96] = 40960,
76	[IB_RNR_TIMER_061_44] = 61410,
77	[IB_RNR_TIMER_081_92] = 81920,
78	[IB_RNR_TIMER_122_88] = 122880,
79	[IB_RNR_TIMER_163_84] = 163840,
80	[IB_RNR_TIMER_245_76] = 245760,
81	[IB_RNR_TIMER_327_68] = 327680,
82	[IB_RNR_TIMER_491_52] = 491520,
83};
84
85static inline unsigned long rnrnak_jiffies(u8 timeout)
86{
87	return max_t(unsigned long,
88		usecs_to_jiffies(rnrnak_usec[timeout]), 1);
89}
90
91static enum ib_wc_opcode wr_to_wc_opcode(enum ib_wr_opcode opcode)
92{
93	switch (opcode) {
94	case IB_WR_RDMA_WRITE:			return IB_WC_RDMA_WRITE;
95	case IB_WR_RDMA_WRITE_WITH_IMM:		return IB_WC_RDMA_WRITE;
96	case IB_WR_SEND:			return IB_WC_SEND;
97	case IB_WR_SEND_WITH_IMM:		return IB_WC_SEND;
98	case IB_WR_RDMA_READ:			return IB_WC_RDMA_READ;
99	case IB_WR_ATOMIC_CMP_AND_SWP:		return IB_WC_COMP_SWAP;
100	case IB_WR_ATOMIC_FETCH_AND_ADD:	return IB_WC_FETCH_ADD;
101	case IB_WR_LSO:				return IB_WC_LSO;
102	case IB_WR_SEND_WITH_INV:		return IB_WC_SEND;
103	case IB_WR_RDMA_READ_WITH_INV:		return IB_WC_RDMA_READ;
104	case IB_WR_LOCAL_INV:			return IB_WC_LOCAL_INV;
105	case IB_WR_REG_MR:			return IB_WC_REG_MR;
106
107	default:
108		return 0xff;
109	}
110}
111
112void retransmit_timer(struct timer_list *t)
113{
114	struct rxe_qp *qp = from_timer(qp, t, retrans_timer);
115
116	if (qp->valid) {
117		qp->comp.timeout = 1;
118		rxe_run_task(&qp->comp.task, 1);
119	}
120}
121
122void rxe_comp_queue_pkt(struct rxe_qp *qp, struct sk_buff *skb)
123{
124	int must_sched;
125
126	skb_queue_tail(&qp->resp_pkts, skb);
127
128	must_sched = skb_queue_len(&qp->resp_pkts) > 1;
129	if (must_sched != 0)
130		rxe_counter_inc(SKB_TO_PKT(skb)->rxe, RXE_CNT_COMPLETER_SCHED);
131
132	rxe_run_task(&qp->comp.task, must_sched);
133}
134
135static inline enum comp_state get_wqe(struct rxe_qp *qp,
136				      struct rxe_pkt_info *pkt,
137				      struct rxe_send_wqe **wqe_p)
138{
139	struct rxe_send_wqe *wqe;
140
141	/* we come here whether or not we found a response packet to see if
142	 * there are any posted WQEs
143	 */
144	wqe = queue_head(qp->sq.queue);
145	*wqe_p = wqe;
146
147	/* no WQE or requester has not started it yet */
148	if (!wqe || wqe->state == wqe_state_posted)
149		return pkt ? COMPST_DONE : COMPST_EXIT;
150
151	/* WQE does not require an ack */
152	if (wqe->state == wqe_state_done)
153		return COMPST_COMP_WQE;
154
155	/* WQE caused an error */
156	if (wqe->state == wqe_state_error)
157		return COMPST_ERROR;
158
159	/* we have a WQE, if we also have an ack check its PSN */
160	return pkt ? COMPST_CHECK_PSN : COMPST_EXIT;
161}
162
163static inline void reset_retry_counters(struct rxe_qp *qp)
164{
165	qp->comp.retry_cnt = qp->attr.retry_cnt;
166	qp->comp.rnr_retry = qp->attr.rnr_retry;
167	qp->comp.started_retry = 0;
168}
169
170static inline enum comp_state check_psn(struct rxe_qp *qp,
171					struct rxe_pkt_info *pkt,
172					struct rxe_send_wqe *wqe)
173{
174	s32 diff;
175
176	/* check to see if response is past the oldest WQE. if it is, complete
177	 * send/write or error read/atomic
178	 */
179	diff = psn_compare(pkt->psn, wqe->last_psn);
180	if (diff > 0) {
181		if (wqe->state == wqe_state_pending) {
182			if (wqe->mask & WR_ATOMIC_OR_READ_MASK)
183				return COMPST_ERROR_RETRY;
184
185			reset_retry_counters(qp);
186			return COMPST_COMP_WQE;
187		} else {
188			return COMPST_DONE;
189		}
190	}
191
192	/* compare response packet to expected response */
193	diff = psn_compare(pkt->psn, qp->comp.psn);
194	if (diff < 0) {
195		/* response is most likely a retried packet if it matches an
196		 * uncompleted WQE go complete it else ignore it
197		 */
198		if (pkt->psn == wqe->last_psn)
199			return COMPST_COMP_ACK;
200		else
201			return COMPST_DONE;
202	} else if ((diff > 0) && (wqe->mask & WR_ATOMIC_OR_READ_MASK)) {
203		return COMPST_DONE;
204	} else {
205		return COMPST_CHECK_ACK;
206	}
207}
208
209static inline enum comp_state check_ack(struct rxe_qp *qp,
210					struct rxe_pkt_info *pkt,
211					struct rxe_send_wqe *wqe)
212{
213	unsigned int mask = pkt->mask;
214	u8 syn;
215	struct rxe_dev *rxe = to_rdev(qp->ibqp.device);
216
217	/* Check the sequence only */
218	switch (qp->comp.opcode) {
219	case -1:
220		/* Will catch all *_ONLY cases. */
221		if (!(mask & RXE_START_MASK))
222			return COMPST_ERROR;
223
224		break;
225
226	case IB_OPCODE_RC_RDMA_READ_RESPONSE_FIRST:
227	case IB_OPCODE_RC_RDMA_READ_RESPONSE_MIDDLE:
228		if (pkt->opcode != IB_OPCODE_RC_RDMA_READ_RESPONSE_MIDDLE &&
229		    pkt->opcode != IB_OPCODE_RC_RDMA_READ_RESPONSE_LAST) {
230			/* read retries of partial data may restart from
231			 * read response first or response only.
232			 */
233			if ((pkt->psn == wqe->first_psn &&
234			     pkt->opcode ==
235			     IB_OPCODE_RC_RDMA_READ_RESPONSE_FIRST) ||
236			    (wqe->first_psn == wqe->last_psn &&
237			     pkt->opcode ==
238			     IB_OPCODE_RC_RDMA_READ_RESPONSE_ONLY))
239				break;
240
241			return COMPST_ERROR;
242		}
243		break;
244	default:
245		WARN_ON_ONCE(1);
246	}
247
248	/* Check operation validity. */
249	switch (pkt->opcode) {
250	case IB_OPCODE_RC_RDMA_READ_RESPONSE_FIRST:
251	case IB_OPCODE_RC_RDMA_READ_RESPONSE_LAST:
252	case IB_OPCODE_RC_RDMA_READ_RESPONSE_ONLY:
253		syn = aeth_syn(pkt);
254
255		if ((syn & AETH_TYPE_MASK) != AETH_ACK)
256			return COMPST_ERROR;
257
258		fallthrough;
259		/* (IB_OPCODE_RC_RDMA_READ_RESPONSE_MIDDLE doesn't have an AETH)
260		 */
261	case IB_OPCODE_RC_RDMA_READ_RESPONSE_MIDDLE:
262		if (wqe->wr.opcode != IB_WR_RDMA_READ &&
263		    wqe->wr.opcode != IB_WR_RDMA_READ_WITH_INV) {
264			wqe->status = IB_WC_FATAL_ERR;
265			return COMPST_ERROR;
266		}
267		reset_retry_counters(qp);
268		return COMPST_READ;
269
270	case IB_OPCODE_RC_ATOMIC_ACKNOWLEDGE:
271		syn = aeth_syn(pkt);
272
273		if ((syn & AETH_TYPE_MASK) != AETH_ACK)
274			return COMPST_ERROR;
275
276		if (wqe->wr.opcode != IB_WR_ATOMIC_CMP_AND_SWP &&
277		    wqe->wr.opcode != IB_WR_ATOMIC_FETCH_AND_ADD)
278			return COMPST_ERROR;
279		reset_retry_counters(qp);
280		return COMPST_ATOMIC;
281
282	case IB_OPCODE_RC_ACKNOWLEDGE:
283		syn = aeth_syn(pkt);
284		switch (syn & AETH_TYPE_MASK) {
285		case AETH_ACK:
286			reset_retry_counters(qp);
287			return COMPST_WRITE_SEND;
288
289		case AETH_RNR_NAK:
290			rxe_counter_inc(rxe, RXE_CNT_RCV_RNR);
291			return COMPST_RNR_RETRY;
292
293		case AETH_NAK:
294			switch (syn) {
295			case AETH_NAK_PSN_SEQ_ERROR:
296				/* a nak implicitly acks all packets with psns
297				 * before
298				 */
299				if (psn_compare(pkt->psn, qp->comp.psn) > 0) {
300					rxe_counter_inc(rxe,
301							RXE_CNT_RCV_SEQ_ERR);
302					qp->comp.psn = pkt->psn;
303					if (qp->req.wait_psn) {
304						qp->req.wait_psn = 0;
305						rxe_run_task(&qp->req.task, 0);
306					}
307				}
308				return COMPST_ERROR_RETRY;
309
310			case AETH_NAK_INVALID_REQ:
311				wqe->status = IB_WC_REM_INV_REQ_ERR;
312				return COMPST_ERROR;
313
314			case AETH_NAK_REM_ACC_ERR:
315				wqe->status = IB_WC_REM_ACCESS_ERR;
316				return COMPST_ERROR;
317
318			case AETH_NAK_REM_OP_ERR:
319				wqe->status = IB_WC_REM_OP_ERR;
320				return COMPST_ERROR;
321
322			default:
323				pr_warn("unexpected nak %x\n", syn);
324				wqe->status = IB_WC_REM_OP_ERR;
325				return COMPST_ERROR;
326			}
327
328		default:
329			return COMPST_ERROR;
330		}
331		break;
332
333	default:
334		pr_warn("unexpected opcode\n");
335	}
336
337	return COMPST_ERROR;
338}
339
340static inline enum comp_state do_read(struct rxe_qp *qp,
341				      struct rxe_pkt_info *pkt,
342				      struct rxe_send_wqe *wqe)
343{
344	int ret;
345
346	ret = copy_data(qp->pd, IB_ACCESS_LOCAL_WRITE,
347			&wqe->dma, payload_addr(pkt),
348			payload_size(pkt), to_mem_obj, NULL);
349	if (ret)
350		return COMPST_ERROR;
351
352	if (wqe->dma.resid == 0 && (pkt->mask & RXE_END_MASK))
353		return COMPST_COMP_ACK;
354	else
355		return COMPST_UPDATE_COMP;
356}
357
358static inline enum comp_state do_atomic(struct rxe_qp *qp,
359					struct rxe_pkt_info *pkt,
360					struct rxe_send_wqe *wqe)
361{
362	int ret;
363
364	u64 atomic_orig = atmack_orig(pkt);
365
366	ret = copy_data(qp->pd, IB_ACCESS_LOCAL_WRITE,
367			&wqe->dma, &atomic_orig,
368			sizeof(u64), to_mem_obj, NULL);
369	if (ret)
370		return COMPST_ERROR;
371	else
372		return COMPST_COMP_ACK;
373}
374
375static void make_send_cqe(struct rxe_qp *qp, struct rxe_send_wqe *wqe,
376			  struct rxe_cqe *cqe)
377{
378	memset(cqe, 0, sizeof(*cqe));
379
380	if (!qp->is_user) {
381		struct ib_wc		*wc	= &cqe->ibwc;
382
383		wc->wr_id		= wqe->wr.wr_id;
384		wc->status		= wqe->status;
385		wc->opcode		= wr_to_wc_opcode(wqe->wr.opcode);
386		if (wqe->wr.opcode == IB_WR_RDMA_WRITE_WITH_IMM ||
387		    wqe->wr.opcode == IB_WR_SEND_WITH_IMM)
388			wc->wc_flags = IB_WC_WITH_IMM;
389		wc->byte_len		= wqe->dma.length;
390		wc->qp			= &qp->ibqp;
391	} else {
392		struct ib_uverbs_wc	*uwc	= &cqe->uibwc;
393
394		uwc->wr_id		= wqe->wr.wr_id;
395		uwc->status		= wqe->status;
396		uwc->opcode		= wr_to_wc_opcode(wqe->wr.opcode);
397		if (wqe->wr.opcode == IB_WR_RDMA_WRITE_WITH_IMM ||
398		    wqe->wr.opcode == IB_WR_SEND_WITH_IMM)
399			uwc->wc_flags = IB_WC_WITH_IMM;
400		uwc->byte_len		= wqe->dma.length;
401		uwc->qp_num		= qp->ibqp.qp_num;
402	}
403}
404
405/*
406 * IBA Spec. Section 10.7.3.1 SIGNALED COMPLETIONS
407 * ---------8<---------8<-------------
408 * ...Note that if a completion error occurs, a Work Completion
409 * will always be generated, even if the signaling
410 * indicator requests an Unsignaled Completion.
411 * ---------8<---------8<-------------
412 */
413static void do_complete(struct rxe_qp *qp, struct rxe_send_wqe *wqe)
414{
415	struct rxe_dev *rxe = to_rdev(qp->ibqp.device);
416	struct rxe_cqe cqe;
417
418	if ((qp->sq_sig_type == IB_SIGNAL_ALL_WR) ||
419	    (wqe->wr.send_flags & IB_SEND_SIGNALED) ||
420	    wqe->status != IB_WC_SUCCESS) {
421		make_send_cqe(qp, wqe, &cqe);
422		advance_consumer(qp->sq.queue);
423		rxe_cq_post(qp->scq, &cqe, 0);
424	} else {
425		advance_consumer(qp->sq.queue);
426	}
427
428	if (wqe->wr.opcode == IB_WR_SEND ||
429	    wqe->wr.opcode == IB_WR_SEND_WITH_IMM ||
430	    wqe->wr.opcode == IB_WR_SEND_WITH_INV)
431		rxe_counter_inc(rxe, RXE_CNT_RDMA_SEND);
432
433	/*
434	 * we completed something so let req run again
435	 * if it is trying to fence
436	 */
437	if (qp->req.wait_fence) {
438		qp->req.wait_fence = 0;
439		rxe_run_task(&qp->req.task, 0);
440	}
441}
442
443static inline enum comp_state complete_ack(struct rxe_qp *qp,
444					   struct rxe_pkt_info *pkt,
445					   struct rxe_send_wqe *wqe)
446{
447	unsigned long flags;
448
449	if (wqe->has_rd_atomic) {
450		wqe->has_rd_atomic = 0;
451		atomic_inc(&qp->req.rd_atomic);
452		if (qp->req.need_rd_atomic) {
453			qp->comp.timeout_retry = 0;
454			qp->req.need_rd_atomic = 0;
455			rxe_run_task(&qp->req.task, 0);
456		}
457	}
458
459	if (unlikely(qp->req.state == QP_STATE_DRAIN)) {
460		/* state_lock used by requester & completer */
461		spin_lock_irqsave(&qp->state_lock, flags);
462		if ((qp->req.state == QP_STATE_DRAIN) &&
463		    (qp->comp.psn == qp->req.psn)) {
464			qp->req.state = QP_STATE_DRAINED;
465			spin_unlock_irqrestore(&qp->state_lock, flags);
466
467			if (qp->ibqp.event_handler) {
468				struct ib_event ev;
469
470				ev.device = qp->ibqp.device;
471				ev.element.qp = &qp->ibqp;
472				ev.event = IB_EVENT_SQ_DRAINED;
473				qp->ibqp.event_handler(&ev,
474					qp->ibqp.qp_context);
475			}
476		} else {
477			spin_unlock_irqrestore(&qp->state_lock, flags);
478		}
479	}
480
481	do_complete(qp, wqe);
482
483	if (psn_compare(pkt->psn, qp->comp.psn) >= 0)
484		return COMPST_UPDATE_COMP;
485	else
486		return COMPST_DONE;
487}
488
489static inline enum comp_state complete_wqe(struct rxe_qp *qp,
490					   struct rxe_pkt_info *pkt,
491					   struct rxe_send_wqe *wqe)
492{
493	if (pkt && wqe->state == wqe_state_pending) {
494		if (psn_compare(wqe->last_psn, qp->comp.psn) >= 0) {
495			qp->comp.psn = (wqe->last_psn + 1) & BTH_PSN_MASK;
496			qp->comp.opcode = -1;
497		}
498
499		if (qp->req.wait_psn) {
500			qp->req.wait_psn = 0;
501			rxe_run_task(&qp->req.task, 1);
502		}
503	}
504
505	do_complete(qp, wqe);
506
507	return COMPST_GET_WQE;
508}
509
510static void rxe_drain_resp_pkts(struct rxe_qp *qp, bool notify)
511{
512	struct sk_buff *skb;
513	struct rxe_send_wqe *wqe;
514
515	while ((skb = skb_dequeue(&qp->resp_pkts))) {
516		rxe_drop_ref(qp);
517		kfree_skb(skb);
518	}
519
520	while ((wqe = queue_head(qp->sq.queue))) {
521		if (notify) {
522			wqe->status = IB_WC_WR_FLUSH_ERR;
523			do_complete(qp, wqe);
524		} else {
525			advance_consumer(qp->sq.queue);
526		}
527	}
528}
529
530int rxe_completer(void *arg)
531{
532	struct rxe_qp *qp = (struct rxe_qp *)arg;
533	struct rxe_dev *rxe = to_rdev(qp->ibqp.device);
534	struct rxe_send_wqe *wqe = NULL;
535	struct sk_buff *skb = NULL;
536	struct rxe_pkt_info *pkt = NULL;
537	enum comp_state state;
538
539	rxe_add_ref(qp);
540
541	if (!qp->valid || qp->req.state == QP_STATE_ERROR ||
542	    qp->req.state == QP_STATE_RESET) {
543		rxe_drain_resp_pkts(qp, qp->valid &&
544				    qp->req.state == QP_STATE_ERROR);
545		goto exit;
546	}
547
548	if (qp->comp.timeout) {
549		qp->comp.timeout_retry = 1;
550		qp->comp.timeout = 0;
551	} else {
552		qp->comp.timeout_retry = 0;
553	}
554
555	if (qp->req.need_retry)
556		goto exit;
557
558	state = COMPST_GET_ACK;
559
560	while (1) {
561		pr_debug("qp#%d state = %s\n", qp_num(qp),
562			 comp_state_name[state]);
563		switch (state) {
564		case COMPST_GET_ACK:
565			skb = skb_dequeue(&qp->resp_pkts);
566			if (skb) {
567				pkt = SKB_TO_PKT(skb);
568				qp->comp.timeout_retry = 0;
569			}
570			state = COMPST_GET_WQE;
571			break;
572
573		case COMPST_GET_WQE:
574			state = get_wqe(qp, pkt, &wqe);
575			break;
576
577		case COMPST_CHECK_PSN:
578			state = check_psn(qp, pkt, wqe);
579			break;
580
581		case COMPST_CHECK_ACK:
582			state = check_ack(qp, pkt, wqe);
583			break;
584
585		case COMPST_READ:
586			state = do_read(qp, pkt, wqe);
587			break;
588
589		case COMPST_ATOMIC:
590			state = do_atomic(qp, pkt, wqe);
591			break;
592
593		case COMPST_WRITE_SEND:
594			if (wqe->state == wqe_state_pending &&
595			    wqe->last_psn == pkt->psn)
596				state = COMPST_COMP_ACK;
597			else
598				state = COMPST_UPDATE_COMP;
599			break;
600
601		case COMPST_COMP_ACK:
602			state = complete_ack(qp, pkt, wqe);
603			break;
604
605		case COMPST_COMP_WQE:
606			state = complete_wqe(qp, pkt, wqe);
607			break;
608
609		case COMPST_UPDATE_COMP:
610			if (pkt->mask & RXE_END_MASK)
611				qp->comp.opcode = -1;
612			else
613				qp->comp.opcode = pkt->opcode;
614
615			if (psn_compare(pkt->psn, qp->comp.psn) >= 0)
616				qp->comp.psn = (pkt->psn + 1) & BTH_PSN_MASK;
617
618			if (qp->req.wait_psn) {
619				qp->req.wait_psn = 0;
620				rxe_run_task(&qp->req.task, 1);
621			}
622
623			state = COMPST_DONE;
624			break;
625
626		case COMPST_DONE:
627			if (pkt) {
628				rxe_drop_ref(pkt->qp);
629				kfree_skb(skb);
630				skb = NULL;
631			}
632			goto done;
633
634		case COMPST_EXIT:
635			if (qp->comp.timeout_retry && wqe) {
636				state = COMPST_ERROR_RETRY;
637				break;
638			}
639
640			/* re reset the timeout counter if
641			 * (1) QP is type RC
642			 * (2) the QP is alive
643			 * (3) there is a packet sent by the requester that
644			 *     might be acked (we still might get spurious
645			 *     timeouts but try to keep them as few as possible)
646			 * (4) the timeout parameter is set
647			 */
648			if ((qp_type(qp) == IB_QPT_RC) &&
649			    (qp->req.state == QP_STATE_READY) &&
650			    (psn_compare(qp->req.psn, qp->comp.psn) > 0) &&
651			    qp->qp_timeout_jiffies)
652				mod_timer(&qp->retrans_timer,
653					  jiffies + qp->qp_timeout_jiffies);
654			goto exit;
655
656		case COMPST_ERROR_RETRY:
657			/* we come here if the retry timer fired and we did
658			 * not receive a response packet. try to retry the send
659			 * queue if that makes sense and the limits have not
660			 * been exceeded. remember that some timeouts are
661			 * spurious since we do not reset the timer but kick
662			 * it down the road or let it expire
663			 */
664
665			/* there is nothing to retry in this case */
666			if (!wqe || (wqe->state == wqe_state_posted))
667				goto exit;
668
669			/* if we've started a retry, don't start another
670			 * retry sequence, unless this is a timeout.
671			 */
672			if (qp->comp.started_retry &&
673			    !qp->comp.timeout_retry) {
674				if (pkt) {
675					rxe_drop_ref(pkt->qp);
676					kfree_skb(skb);
677					skb = NULL;
678				}
679
680				goto done;
681			}
682
683			if (qp->comp.retry_cnt > 0) {
684				if (qp->comp.retry_cnt != 7)
685					qp->comp.retry_cnt--;
686
687				/* no point in retrying if we have already
688				 * seen the last ack that the requester could
689				 * have caused
690				 */
691				if (psn_compare(qp->req.psn,
692						qp->comp.psn) > 0) {
693					/* tell the requester to retry the
694					 * send queue next time around
695					 */
696					rxe_counter_inc(rxe,
697							RXE_CNT_COMP_RETRY);
698					qp->req.need_retry = 1;
699					qp->comp.started_retry = 1;
700					rxe_run_task(&qp->req.task, 0);
701				}
702
703				if (pkt) {
704					rxe_drop_ref(pkt->qp);
705					kfree_skb(skb);
706					skb = NULL;
707				}
708
709				goto done;
710
711			} else {
712				rxe_counter_inc(rxe, RXE_CNT_RETRY_EXCEEDED);
713				wqe->status = IB_WC_RETRY_EXC_ERR;
714				state = COMPST_ERROR;
715			}
716			break;
717
718		case COMPST_RNR_RETRY:
719			if (qp->comp.rnr_retry > 0) {
720				if (qp->comp.rnr_retry != 7)
721					qp->comp.rnr_retry--;
722
723				qp->req.need_retry = 1;
724				pr_debug("qp#%d set rnr nak timer\n",
725					 qp_num(qp));
726				mod_timer(&qp->rnr_nak_timer,
727					  jiffies + rnrnak_jiffies(aeth_syn(pkt)
728						& ~AETH_TYPE_MASK));
729				rxe_drop_ref(pkt->qp);
730				kfree_skb(skb);
731				skb = NULL;
732				goto exit;
733			} else {
734				rxe_counter_inc(rxe,
735						RXE_CNT_RNR_RETRY_EXCEEDED);
736				wqe->status = IB_WC_RNR_RETRY_EXC_ERR;
737				state = COMPST_ERROR;
738			}
739			break;
740
741		case COMPST_ERROR:
742			WARN_ON_ONCE(wqe->status == IB_WC_SUCCESS);
743			do_complete(qp, wqe);
744			rxe_qp_error(qp);
745
746			if (pkt) {
747				rxe_drop_ref(pkt->qp);
748				kfree_skb(skb);
749				skb = NULL;
750			}
751
752			goto exit;
753		}
754	}
755
756exit:
757	/* we come here if we are done with processing and want the task to
758	 * exit from the loop calling us
759	 */
760	WARN_ON_ONCE(skb);
761	rxe_drop_ref(qp);
762	return -EAGAIN;
763
764done:
765	/* we come here if we have processed a packet we want the task to call
766	 * us again to see if there is anything else to do
767	 */
768	WARN_ON_ONCE(skb);
769	rxe_drop_ref(qp);
770	return 0;
771}
772