em_txrx.c 22.4 KB
Newer Older
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
/*-
 * Copyright (c) 2016 Matt Macy <mmacy@nextbsd.org>
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 */

27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
/* $FreeBSD$ */
#include "if_em.h"

#ifdef	RSS
#include <net/rss_config.h>
#include <netinet/in_rss.h>
#endif

#ifdef VERBOSE_DEBUG
#define DPRINTF device_printf
#else
#define DPRINTF(...)
#endif

/*********************************************************************
 *  Local Function prototypes
 *********************************************************************/
static int em_tso_setup(struct adapter *adapter, if_pkt_info_t pi, u32 *txd_upper, u32 *txd_lower);
static int em_transmit_checksum_setup(struct adapter *adapter, if_pkt_info_t pi, u32 *txd_upper, u32 *txd_lower);
static int em_isc_txd_encap(void *arg, if_pkt_info_t pi);
static void em_isc_txd_flush(void *arg, uint16_t txqid, uint32_t pidx);
static int em_isc_txd_credits_update(void *arg, uint16_t txqid, uint32_t cidx_init, bool clear);
static void em_isc_rxd_refill(void *arg, uint16_t rxqid, uint8_t flid __unused,
			      uint32_t pidx, uint64_t *paddrs, caddr_t *vaddrs __unused, uint16_t count, uint16_t buflen __unused);
static void em_isc_rxd_flush(void *arg, uint16_t rxqid, uint8_t flid __unused, uint32_t pidx);
static int em_isc_rxd_available(void *arg, uint16_t rxqid, uint32_t idx,
				int budget);
static int em_isc_rxd_pkt_get(void *arg, if_rxd_info_t ri);

static void lem_isc_rxd_refill(void *arg, uint16_t rxqid, uint8_t flid __unused,
			      uint32_t pidx, uint64_t *paddrs, caddr_t *vaddrs __unused, uint16_t count, uint16_t buflen __unused);

static int lem_isc_rxd_available(void *arg, uint16_t rxqid, uint32_t idx,
				int budget);
static int lem_isc_rxd_pkt_get(void *arg, if_rxd_info_t ri);

static void lem_receive_checksum(int status, int errors, if_rxd_info_t ri);
static void em_receive_checksum(uint32_t status, if_rxd_info_t ri);
extern int em_intr(void *arg);

struct if_txrx em_txrx  = {
	em_isc_txd_encap,
	em_isc_txd_flush,
	em_isc_txd_credits_update,
	em_isc_rxd_available,
	em_isc_rxd_pkt_get,
	em_isc_rxd_refill,
	em_isc_rxd_flush,
	em_intr
};

struct if_txrx lem_txrx  = {
	em_isc_txd_encap,
	em_isc_txd_flush,
	em_isc_txd_credits_update,
	lem_isc_rxd_available,
	lem_isc_rxd_pkt_get,
	lem_isc_rxd_refill,
	em_isc_rxd_flush,
	em_intr
};

extern if_shared_ctx_t em_sctx; 

/**********************************************************************
 *
 *  Setup work for hardware segmentation offload (TSO) on
 *  adapters using advanced tx descriptors
 *
 **********************************************************************/
static int
em_tso_setup(struct adapter *adapter, if_pkt_info_t pi, u32 *txd_upper, u32 *txd_lower)
{
	if_softc_ctx_t scctx = adapter->shared;
        struct em_tx_queue *que = &adapter->tx_queues[pi->ipi_qsidx];
        struct tx_ring *txr = &que->txr;
	struct e1000_context_desc *TXD;
	struct em_txbuffer  *tx_buffer;
        int cur, hdr_len;

	hdr_len = pi->ipi_ehdrlen + pi->ipi_ip_hlen + pi->ipi_tcp_hlen;
  	*txd_lower = (E1000_TXD_CMD_DEXT |	/* Extended descr type */
		      E1000_TXD_DTYP_D |	/* Data descr type */
		      E1000_TXD_CMD_TSE);	/* Do TSE on this packet */

	/* IP and/or TCP header checksum calculation and insertion. */
	*txd_upper = (E1000_TXD_POPTS_IXSM | E1000_TXD_POPTS_TXSM) << 8;

	cur = pi->ipi_pidx;
        TXD = (struct e1000_context_desc *)&txr->tx_base[cur];
        tx_buffer = &txr->tx_buffers[cur];
	
	 /*
	 * Start offset for header checksum calculation.
	 * End offset for header checksum calculation.
	 * Offset of place put the checksum.
	 */
	TXD->lower_setup.ip_fields.ipcss = pi->ipi_ehdrlen;
	TXD->lower_setup.ip_fields.ipcse =
	    htole16(pi->ipi_ehdrlen + pi->ipi_ip_hlen - 1);
	TXD->lower_setup.ip_fields.ipcso = pi->ipi_ehdrlen + offsetof(struct ip, ip_sum);

         /*
	 * Start offset for payload checksum calculation.
	 * End offset for payload checksum calculation.
	 * Offset of place to put the checksum.
	 */
	TXD->upper_setup.tcp_fields.tucss = pi->ipi_ehdrlen + pi->ipi_ip_hlen;
	TXD->upper_setup.tcp_fields.tucse = 0;
	TXD->upper_setup.tcp_fields.tucso =
	    pi->ipi_ehdrlen + pi->ipi_ip_hlen + offsetof(struct tcphdr, th_sum);
	
         /*
	 * Payload size per packet w/o any headers.
	 * Length of all headers up to payload.
	 */
	TXD->tcp_seg_setup.fields.mss = htole16(pi->ipi_tso_segsz);
	TXD->tcp_seg_setup.fields.hdr_len = hdr_len;

	TXD->cmd_and_length = htole32(adapter->txd_cmd |
				E1000_TXD_CMD_DEXT |	/* Extended descr */
				E1000_TXD_CMD_TSE |	/* TSE context */
				E1000_TXD_CMD_IP |	/* Do IP csum */
				E1000_TXD_CMD_TCP |	/* Do TCP checksum */
				      (pi->ipi_len - hdr_len)); /* Total len */
	tx_buffer->eop = -1;
	txr->tx_tso = TRUE;

	if (++cur == scctx->isc_ntxd[0]) {
		cur = 0;
	}
	DPRINTF(iflib_get_dev(adapter->ctx), "%s: pidx: %d cur: %d\n", __FUNCTION__, pi->ipi_pidx, cur);
	return (cur);
}

#define TSO_WORKAROUND 4
#define DONT_FORCE_CTX 1


/*********************************************************************
 *  The offload context is protocol specific (TCP/UDP) and thus
 *  only needs to be set when the protocol changes. The occasion
 *  of a context change can be a performance detriment, and
 *  might be better just disabled. The reason arises in the way
 *  in which the controller supports pipelined requests from the
 *  Tx data DMA. Up to four requests can be pipelined, and they may
 *  belong to the same packet or to multiple packets. However all
 *  requests for one packet are issued before a request is issued
 *  for a subsequent packet and if a request for the next packet
 *  requires a context change, that request will be stalled
 *  until the previous request completes. This means setting up
 *  a new context effectively disables pipelined Tx data DMA which
 *  in turn greatly slow down performance to send small sized
 *  frames. 
 **********************************************************************/

static int
em_transmit_checksum_setup(struct adapter *adapter, if_pkt_info_t pi, u32 *txd_upper, u32 *txd_lower)
{
        struct e1000_context_desc   *TXD = NULL;
	if_softc_ctx_t              scctx = adapter->shared;
 	struct em_tx_queue          *que = &adapter->tx_queues[pi->ipi_qsidx];
	struct tx_ring              *txr = &que->txr;
	struct em_txbuffer          *tx_buffer;
	int                         csum_flags = pi->ipi_csum_flags;
	int                         cur, hdr_len;
	u32                         cmd;
	
	cur = pi->ipi_pidx;
	hdr_len = pi->ipi_ehdrlen + pi->ipi_ip_hlen;
	cmd = adapter->txd_cmd;

	/*
	 * The 82574L can only remember the *last* context used
	 * regardless of queue that it was use for.  We cannot reuse
	 * contexts on this hardware platform and must generate a new
	 * context every time.  82574L hardware spec, section 7.2.6,
	 * second note.
	 */
	if (DONT_FORCE_CTX &&
	    adapter->tx_num_queues == 1 &&
	    txr->csum_lhlen == pi->ipi_ehdrlen &&
	    txr->csum_iphlen == pi->ipi_ip_hlen &&
	    txr->csum_flags == csum_flags) {
		/*
		 * Same csum offload context as the previous packets;
		 * just return.
		 */
		*txd_upper = txr->csum_txd_upper;
		*txd_lower = txr->csum_txd_lower;
		return (cur);
	}

	TXD = (struct e1000_context_desc *)&txr->tx_base[cur];
	if (csum_flags & CSUM_IP) {
	  	*txd_upper |= E1000_TXD_POPTS_IXSM << 8;
		/*
		 * Start offset for header checksum calculation.
		 * End offset for header checksum calculation.
		 * Offset of place to put the checksum.
		 */
		TXD->lower_setup.ip_fields.ipcss = pi->ipi_ehdrlen;
		TXD->lower_setup.ip_fields.ipcse = htole16(hdr_len);
		TXD->lower_setup.ip_fields.ipcso = pi->ipi_ehdrlen + offsetof(struct ip, ip_sum);
		cmd |= E1000_TXD_CMD_IP;
	}

	if (csum_flags & (CSUM_TCP|CSUM_UDP)) {
		uint8_t tucso;

 		*txd_upper |= E1000_TXD_POPTS_TXSM << 8;
		*txd_lower = E1000_TXD_CMD_DEXT | E1000_TXD_DTYP_D;

		if (csum_flags & CSUM_TCP) {
			tucso = hdr_len + offsetof(struct tcphdr, th_sum);
			cmd |= E1000_TXD_CMD_TCP;
		} else
			tucso = hdr_len + offsetof(struct udphdr, uh_sum);
 		TXD->upper_setup.tcp_fields.tucss = hdr_len;
 		TXD->upper_setup.tcp_fields.tucse = htole16(0);
 		TXD->upper_setup.tcp_fields.tucso = tucso;
	}

	txr->csum_lhlen = pi->ipi_ehdrlen;
	txr->csum_iphlen = pi->ipi_ip_hlen;
	txr->csum_flags = csum_flags;
	txr->csum_txd_upper = *txd_upper;
	txr->csum_txd_lower = *txd_lower;

	TXD->tcp_seg_setup.data = htole32(0);
	TXD->cmd_and_length =
		htole32(E1000_TXD_CMD_IFCS | E1000_TXD_CMD_DEXT | cmd);

	tx_buffer = &txr->tx_buffers[cur];
	tx_buffer->eop = -1;

	if (++cur == scctx->isc_ntxd[0]) {
		cur = 0;
	}
	DPRINTF(iflib_get_dev(adapter->ctx), "checksum_setup csum_flags=%x txd_upper=%x txd_lower=%x hdr_len=%d cmd=%x\n",
		      csum_flags, *txd_upper, *txd_lower, hdr_len, cmd);
	return (cur);
}

static int
em_isc_txd_encap(void *arg, if_pkt_info_t pi)
{
        struct adapter *sc       = arg;
	if_softc_ctx_t scctx     = sc->shared;
	struct em_tx_queue *que  = &sc->tx_queues[pi->ipi_qsidx];
	struct tx_ring *txr      = &que->txr;
	bus_dma_segment_t *segs  = pi->ipi_segs;
	int nsegs                = pi->ipi_nsegs;
	int csum_flags           = pi->ipi_csum_flags;
        int i, j, first, pidx_last; 
	u32                     txd_upper = 0, txd_lower = 0; 
	
	struct em_txbuffer *tx_buffer;
	struct e1000_tx_desc *ctxd = NULL;
	bool do_tso, tso_desc; 
	
	i = first = pi->ipi_pidx;         
	do_tso = (csum_flags & CSUM_TSO);
	tso_desc = FALSE;
           /*
	 * TSO Hardware workaround, if this packet is not
	 * TSO, and is only a single descriptor long, and
	 * it follows a TSO burst, then we need to add a
	 * sentinel descriptor to prevent premature writeback.
	 */
	if ((!do_tso) && (txr->tx_tso == TRUE)) {
		if (nsegs == 1)
			tso_desc = TRUE;
		txr->tx_tso = FALSE;
	}

	/* Do hardware assists */
	if (do_tso) {
		i = em_tso_setup(sc, pi, &txd_upper, &txd_lower);
		tso_desc = TRUE;
307
	} else if (csum_flags & EM_CSUM_OFFLOAD) {
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
		i = em_transmit_checksum_setup(sc, pi, &txd_upper, &txd_lower);
	}

	if (pi->ipi_mflags & M_VLANTAG) {
	  /* Set the vlan id. */
		txd_upper |= htole16(pi->ipi_vtag) << 16;
                /* Tell hardware to add tag */
                txd_lower |= htole32(E1000_TXD_CMD_VLE);
	}

	DPRINTF(iflib_get_dev(sc->ctx), "encap: set up tx: nsegs=%d first=%d i=%d\n", nsegs, first, i);
	/* XXX adapter->pcix_82544 -- lem_fill_descriptors */

	/* Set up our transmit descriptors */
	for (j = 0; j < nsegs; j++) {
		bus_size_t seg_len;
		bus_addr_t seg_addr;
		uint32_t cmd;

		ctxd = &txr->tx_base[i];
		tx_buffer = &txr->tx_buffers[i];
		seg_addr = segs[j].ds_addr;
		seg_len = segs[j].ds_len;
		cmd = E1000_TXD_CMD_IFCS | sc->txd_cmd;

		/*
		** TSO Workaround:
		** If this is the last descriptor, we want to
		** split it so we have a small final sentinel
		*/
		if (tso_desc && (j == (nsegs - 1)) && (seg_len > 8)) {
			seg_len -= TSO_WORKAROUND;
			ctxd->buffer_addr = htole64(seg_addr);
			ctxd->lower.data = htole32(cmd | txd_lower | seg_len);
			ctxd->upper.data = htole32(txd_upper);

                        if (++i == scctx->isc_ntxd[0])
				i = 0;

			/* Now make the sentinel */
			ctxd = &txr->tx_base[i];
			tx_buffer = &txr->tx_buffers[i];
			ctxd->buffer_addr = htole64(seg_addr + seg_len);
			ctxd->lower.data = htole32(cmd | txd_lower | TSO_WORKAROUND);
			ctxd->upper.data = htole32(txd_upper);
			pidx_last = i;
			if (++i == scctx->isc_ntxd[0])
				i = 0;
			DPRINTF(iflib_get_dev(sc->ctx), "TSO path pidx_last=%d i=%d ntxd[0]=%d\n", pidx_last, i, scctx->isc_ntxd[0]);
		} else {
			ctxd->buffer_addr = htole64(seg_addr);
			ctxd->lower.data = htole32(cmd | txd_lower | seg_len);
			ctxd->upper.data = htole32(txd_upper);
			pidx_last = i;
			if (++i == scctx->isc_ntxd[0])
				i = 0;
			DPRINTF(iflib_get_dev(sc->ctx), "pidx_last=%d i=%d ntxd[0]=%d\n", pidx_last, i, scctx->isc_ntxd[0]);
		}
		tx_buffer->eop = -1;
	}

	/*
         * Last Descriptor of Packet
	 * needs End Of Packet (EOP)
	 * and Report Status (RS)
         */
        ctxd->lower.data |=
		htole32(E1000_TXD_CMD_EOP | E1000_TXD_CMD_RS);

	tx_buffer = &txr->tx_buffers[first];
	tx_buffer->eop = pidx_last;
	DPRINTF(iflib_get_dev(sc->ctx), "tx_buffers[%d]->eop = %d ipi_new_pidx=%d\n", first, pidx_last, i);
	pi->ipi_new_pidx = i;

	return (0); 
}

static void
em_isc_txd_flush(void *arg, uint16_t txqid, uint32_t pidx)
{
	struct adapter *adapter = arg;
	struct em_tx_queue *que = &adapter->tx_queues[txqid];
	struct tx_ring *txr = &que->txr;

	E1000_WRITE_REG(&adapter->hw, E1000_TDT(txr->me), pidx);
}

static int
em_isc_txd_credits_update(void *arg, uint16_t txqid, uint32_t cidx_init, bool clear)
{
	struct adapter *adapter = arg;
	if_softc_ctx_t scctx = adapter->shared;
	struct em_tx_queue *que = &adapter->tx_queues[txqid];
	struct tx_ring *txr = &que->txr;

	u32 cidx, processed = 0;
	int last, done;
	struct em_txbuffer *buf;
	struct e1000_tx_desc *tx_desc, *eop_desc;

	cidx = cidx_init;
	buf = &txr->tx_buffers[cidx];
	tx_desc = &txr->tx_base[cidx];
Sean Bruno's avatar
iflib:    
Sean Bruno committed
411
412
413
	last = buf->eop;
	if (last == -1)
		return (processed);
414
415
	eop_desc = &txr->tx_base[last];

Sean Bruno's avatar
iflib:    
Sean Bruno committed
416
417
	DPRINTF(iflib_get_dev(adapter->ctx),
		      "credits_update: cidx_init=%d clear=%d last=%d\n",
418
419
420
421
422
423
424
425
		      cidx_init, clear, last);
	/*
	 * What this does is get the index of the
	 * first descriptor AFTER the EOP of the
	 * first packet, that way we can do the
	 * simple comparison on the inner while loop.
	 */
	if (++last == scctx->isc_ntxd[0])
Sean Bruno's avatar
iflib:    
Sean Bruno committed
426
		last = 0;
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
	done = last;


	while (eop_desc->upper.fields.status & E1000_TXD_STAT_DD) {
		/* We clean the range of the packet */
		while (cidx != done) {
			if (clear) {
				tx_desc->upper.data = 0;
				tx_desc->lower.data = 0;
				tx_desc->buffer_addr = 0;
				buf->eop = -1;
			}
			tx_desc++;
			buf++;
			processed++;
Sean Bruno's avatar
iflib:    
Sean Bruno committed
442

443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
			/* wrap the ring ? */
			if (++cidx == scctx->isc_ntxd[0]) {
				cidx = 0;
			}
			buf = &txr->tx_buffers[cidx];
			tx_desc = &txr->tx_base[cidx];
		}
		/* See if we can continue to the next packet */
		last = buf->eop;
		if (last == -1)
			break;
		eop_desc = &txr->tx_base[last];
		/* Get new done point */
		if (++last == scctx->isc_ntxd[0])
			last = 0;
		done = last;
	}

	DPRINTF(iflib_get_dev(adapter->ctx), "Processed %d credits update\n", processed);
	return(processed);
}

static void
lem_isc_rxd_refill(void *arg, uint16_t rxqid, uint8_t flid __unused,
		  uint32_t pidx, uint64_t *paddrs, caddr_t *vaddrs __unused, uint16_t count, uint16_t buflen __unused)
{
	struct adapter *sc = arg;
	if_softc_ctx_t scctx = sc->shared;
	struct em_rx_queue *que = &sc->rx_queues[rxqid];
	struct rx_ring *rxr = &que->rxr;
	struct e1000_rx_desc *rxd;
	int i;
	uint32_t next_pidx;

	for (i = 0, next_pidx = pidx; i < count; i++) {
		rxd = (struct e1000_rx_desc *)&rxr->rx_base[next_pidx];
		rxd->buffer_addr = htole64(paddrs[i]);
		/* status bits must be cleared */
		rxd->status = 0;

		if (++next_pidx == scctx->isc_nrxd[0])
			next_pidx = 0;
	}
}

static void
em_isc_rxd_refill(void *arg, uint16_t rxqid, uint8_t flid __unused,
		  uint32_t pidx, uint64_t *paddrs, caddr_t *vaddrs __unused, uint16_t count, uint16_t buflen __unused)
{
	struct adapter *sc = arg;
	if_softc_ctx_t scctx = sc->shared;
	struct em_rx_queue *que = &sc->rx_queues[rxqid];
	struct rx_ring *rxr = &que->rxr;
	union e1000_rx_desc_extended *rxd;
	int i;
	uint32_t next_pidx;

	for (i = 0, next_pidx = pidx; i < count; i++) {
		rxd = &rxr->rx_base[next_pidx];
		rxd->read.buffer_addr = htole64(paddrs[i]);
		/* DD bits must be cleared */
		rxd->wb.upper.status_error = 0; 
	
		if (++next_pidx == scctx->isc_nrxd[0])
			next_pidx = 0;
	}
}

static void
em_isc_rxd_flush(void *arg, uint16_t rxqid, uint8_t flid __unused, uint32_t pidx)
{
    	struct adapter *sc       = arg;
	struct em_rx_queue *que     = &sc->rx_queues[rxqid];
	struct rx_ring *rxr      = &que->rxr;

        E1000_WRITE_REG(&sc->hw, E1000_RDT(rxr->me), pidx);
}

static int
lem_isc_rxd_available(void *arg, uint16_t rxqid, uint32_t idx, int budget)
{
	struct adapter *sc         = arg;
	if_softc_ctx_t scctx = sc->shared;
	struct em_rx_queue *que   = &sc->rx_queues[rxqid];
	struct rx_ring *rxr        = &que->rxr;
	struct e1000_rx_desc *rxd;
	u32                      staterr = 0;
	int                      cnt, i;

	for (cnt = 0, i = idx; cnt < scctx->isc_nrxd[0] && cnt <= budget;) {
		rxd = (struct e1000_rx_desc *)&rxr->rx_base[i];
		staterr = rxd->status;

		if ((staterr & E1000_RXD_STAT_DD) == 0)
			break;

		if (++i == scctx->isc_nrxd[0])
			i = 0;

		if (staterr & E1000_RXD_STAT_EOP)
			cnt++;
	}
	return (cnt);
}

static int
em_isc_rxd_available(void *arg, uint16_t rxqid, uint32_t idx, int budget)
{
       	struct adapter *sc         = arg;
	if_softc_ctx_t scctx = sc->shared;
	struct em_rx_queue *que   = &sc->rx_queues[rxqid];
	struct rx_ring *rxr        = &que->rxr;
	union e1000_rx_desc_extended *rxd;
	u32                      staterr = 0;
	int                      cnt, i;

	for (cnt = 0, i = idx; cnt < scctx->isc_nrxd[0] && cnt <= budget;) {
		rxd = &rxr->rx_base[i];
		staterr = le32toh(rxd->wb.upper.status_error);

		if ((staterr & E1000_RXD_STAT_DD) == 0)
			break;
		
		if (++i == scctx->isc_nrxd[0]) {
			i = 0;
		}

		if (staterr & E1000_RXD_STAT_EOP)
			cnt++;

	}
	return (cnt);
}

static int
lem_isc_rxd_pkt_get(void *arg, if_rxd_info_t ri)
{
	struct adapter           *adapter = arg;
	if_softc_ctx_t           scctx = adapter->shared;
	struct em_rx_queue       *que = &adapter->rx_queues[ri->iri_qsidx];
	struct rx_ring           *rxr = &que->rxr;
	struct e1000_rx_desc *rxd;
	u16                      len;
	u32                      status, errors;
	bool                     eop;
	int                      i, cidx;

	status = errors = i = 0;
	cidx = ri->iri_cidx;

	do {
		rxd = (struct e1000_rx_desc *)&rxr->rx_base[cidx];
		status = rxd->status;
		errors = rxd->errors;

		/* Error Checking then decrement count */
		MPASS ((status & E1000_RXD_STAT_DD) != 0);

		len = le16toh(rxd->length);
		ri->iri_len += len;

		eop = (status & E1000_RXD_STAT_EOP) != 0;

		/* Make sure bad packets are discarded */
		if (errors & E1000_RXD_ERR_FRAME_ERR_MASK) {
			adapter->dropped_pkts++;
			/* XXX fixup if common */
			return (EBADMSG);
		}

		ri->iri_frags[i].irf_flid = 0;
		ri->iri_frags[i].irf_idx = cidx;
		ri->iri_frags[i].irf_len = len;
		/* Zero out the receive descriptors status. */
		rxd->status = 0;

		if (++cidx == scctx->isc_nrxd[0])
			cidx = 0;
		i++;
	} while (!eop);

	/* XXX add a faster way to look this up */
	if (adapter->hw.mac.type >= e1000_82543 && !(status & E1000_RXD_STAT_IXSM))
		lem_receive_checksum(status, errors, ri);

	if (status & E1000_RXD_STAT_VP) {
		ri->iri_vtag = le16toh(rxd->special);
		ri->iri_flags |= M_VLANTAG;
	}

	ri->iri_nfrags = i;

	return (0);
}

static int
em_isc_rxd_pkt_get(void *arg, if_rxd_info_t ri)
{
      	struct adapter           *adapter = arg;
	if_softc_ctx_t           scctx = adapter->shared;
	struct em_rx_queue       *que = &adapter->rx_queues[ri->iri_qsidx];
	struct rx_ring           *rxr = &que->rxr;
	union e1000_rx_desc_extended *rxd;

	u16                      len; 
	u32                      staterr = 0;
	bool                     eop;
	int                      i, cidx, vtag;

	i = vtag = 0;
	cidx = ri->iri_cidx;

	do {
		rxd = &rxr->rx_base[cidx];
		staterr = le32toh(rxd->wb.upper.status_error);	
	
		/* Error Checking then decrement count */
		MPASS ((staterr & E1000_RXD_STAT_DD) != 0);

		len = le16toh(rxd->wb.upper.length);
		ri->iri_len += len; 

		eop = (staterr & E1000_RXD_STAT_EOP) != 0;

		/* Make sure bad packets are discarded */
		if (staterr & E1000_RXDEXT_ERR_FRAME_ERR_MASK) {
			adapter->dropped_pkts++;
			return EBADMSG;
		}

		ri->iri_frags[i].irf_flid = 0;
		ri->iri_frags[i].irf_idx = cidx;
		ri->iri_frags[i].irf_len = len;
		/* Zero out the receive descriptors status. */
		rxd->wb.upper.status_error &= htole32(~0xFF);

		if (++cidx == scctx->isc_nrxd[0])
			cidx = 0;
		i++;
	} while (!eop);

	/* XXX add a faster way to look this up */
	if (adapter->hw.mac.type >= e1000_82543)
		em_receive_checksum(staterr, ri);

	if (staterr & E1000_RXD_STAT_VP) {
		vtag = le16toh(rxd->wb.upper.vlan);
	} 
	
	ri->iri_vtag = vtag;
	ri->iri_nfrags = i;
	if (vtag)
		ri->iri_flags |= M_VLANTAG;
		
	return (0);
}

/*********************************************************************
 *
 *  Verify that the hardware indicated that the checksum is valid.
 *  Inform the stack about the status of checksum so that stack
 *  doesn't spend time verifying the checksum.
 *
 *********************************************************************/
static void
lem_receive_checksum(int status, int errors, if_rxd_info_t ri)
{
	/* Did it pass? */
	if (status & E1000_RXD_STAT_IPCS && !(errors & E1000_RXD_ERR_IPE))
		ri->iri_csum_flags = (CSUM_IP_CHECKED|CSUM_IP_VALID);

	if (status & E1000_RXD_STAT_TCPCS) {
		/* Did it pass? */
		if (!(errors & E1000_RXD_ERR_TCPE)) {
			ri->iri_csum_flags |=
			(CSUM_DATA_VALID | CSUM_PSEUDO_HDR);
			ri->iri_csum_data = htons(0xffff);
		}
	}
}

static void
em_receive_checksum(uint32_t status, if_rxd_info_t ri)
{
	ri->iri_csum_flags = 0;

	/* Ignore Checksum bit is set */
	if (status & E1000_RXD_STAT_IXSM)
		return;

	/* If the IP checksum exists and there is no IP Checksum error */
	if ((status & (E1000_RXD_STAT_IPCS | E1000_RXDEXT_STATERR_IPE)) ==
		E1000_RXD_STAT_IPCS) {
		ri->iri_csum_flags = (CSUM_IP_CHECKED | CSUM_IP_VALID);
	}

	/* TCP or UDP checksum */
	if ((status & (E1000_RXD_STAT_TCPCS | E1000_RXDEXT_STATERR_TCPE)) ==
	    E1000_RXD_STAT_TCPCS) {
		ri->iri_csum_flags |= (CSUM_DATA_VALID | CSUM_PSEUDO_HDR);
		ri->iri_csum_data = htons(0xffff);
	}
	if (status & E1000_RXD_STAT_UDPCS) {
		ri->iri_csum_flags |= (CSUM_DATA_VALID | CSUM_PSEUDO_HDR);
		ri->iri_csum_data = htons(0xffff);
	}
}