nicvf_queues.c 38.5 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 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
/*
 * Copyright (C) 2015 Cavium, Inc.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of version 2 of the GNU General Public License
 * as published by the Free Software Foundation.
 */

#include <linux/pci.h>
#include <linux/netdevice.h>
#include <linux/ip.h>
#include <linux/etherdevice.h>
#include <net/ip.h>
#include <net/tso.h>

#include "nic_reg.h"
#include "nic.h"
#include "q_struct.h"
#include "nicvf_queues.h"

struct rbuf_info {
	struct page *page;
	void	*data;
	u64	offset;
};

#define GET_RBUF_INFO(x) ((struct rbuf_info *)(x - NICVF_RCV_BUF_ALIGN_BYTES))

/* Poll a register for a specific value */
static int nicvf_poll_reg(struct nicvf *nic, int qidx,
			  u64 reg, int bit_pos, int bits, int val)
{
	u64 bit_mask;
	u64 reg_val;
	int timeout = 10;

	bit_mask = (1ULL << bits) - 1;
	bit_mask = (bit_mask << bit_pos);

	while (timeout) {
		reg_val = nicvf_queue_reg_read(nic, reg, qidx);
		if (((reg_val & bit_mask) >> bit_pos) == val)
			return 0;
		usleep_range(1000, 2000);
		timeout--;
	}
	netdev_err(nic->netdev, "Poll on reg 0x%llx failed\n", reg);
	return 1;
}

/* Allocate memory for a queue's descriptors */
static int nicvf_alloc_q_desc_mem(struct nicvf *nic, struct q_desc_mem *dmem,
				  int q_len, int desc_size, int align_bytes)
{
	dmem->q_len = q_len;
	dmem->size = (desc_size * q_len) + align_bytes;
	/* Save address, need it while freeing */
	dmem->unalign_base = dma_zalloc_coherent(&nic->pdev->dev, dmem->size,
						&dmem->dma, GFP_KERNEL);
	if (!dmem->unalign_base)
		return -ENOMEM;

	/* Align memory address for 'align_bytes' */
	dmem->phys_base = NICVF_ALIGNED_ADDR((u64)dmem->dma, align_bytes);
65
	dmem->base = dmem->unalign_base + (dmem->phys_base - dmem->dma);
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
	return 0;
}

/* Free queue's descriptor memory */
static void nicvf_free_q_desc_mem(struct nicvf *nic, struct q_desc_mem *dmem)
{
	if (!dmem)
		return;

	dma_free_coherent(&nic->pdev->dev, dmem->size,
			  dmem->unalign_base, dmem->dma);
	dmem->unalign_base = NULL;
	dmem->base = NULL;
}

/* Allocate buffer for packet reception
 * HW returns memory address where packet is DMA'ed but not a pointer
 * into RBDR ring, so save buffer address at the start of fragment and
 * align the start address to a cache aligned address
 */
static inline int nicvf_alloc_rcv_buffer(struct nicvf *nic, gfp_t gfp,
					 u32 buf_len, u64 **rbuf)
{
	u64 data;
	struct rbuf_info *rinfo;
	int order = get_order(buf_len);

	/* Check if request can be accomodated in previous allocated page */
	if (nic->rb_page) {
		if ((nic->rb_page_offset + buf_len + buf_len) >
		    (PAGE_SIZE << order)) {
			nic->rb_page = NULL;
		} else {
			nic->rb_page_offset += buf_len;
			get_page(nic->rb_page);
		}
	}

	/* Allocate a new page */
	if (!nic->rb_page) {
106 107
		nic->rb_page = alloc_pages(gfp | __GFP_COMP | __GFP_NOWARN,
					   order);
108
		if (!nic->rb_page) {
109 110
			netdev_err(nic->netdev,
				   "Failed to allocate new rcv buffer\n");
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
			return -ENOMEM;
		}
		nic->rb_page_offset = 0;
	}

	data = (u64)page_address(nic->rb_page) + nic->rb_page_offset;

	/* Align buffer addr to cache line i.e 128 bytes */
	rinfo = (struct rbuf_info *)(data + NICVF_RCV_BUF_ALIGN_LEN(data));
	/* Save page address for reference updation */
	rinfo->page = nic->rb_page;
	/* Store start address for later retrieval */
	rinfo->data = (void *)data;
	/* Store alignment offset */
	rinfo->offset = NICVF_RCV_BUF_ALIGN_LEN(data);

	data += rinfo->offset;

	/* Give next aligned address to hw for DMA */
	*rbuf = (u64 *)(data + NICVF_RCV_BUF_ALIGN_BYTES);
	return 0;
}

/* Retrieve actual buffer start address and build skb for received packet */
static struct sk_buff *nicvf_rb_ptr_to_skb(struct nicvf *nic,
					   u64 rb_ptr, int len)
{
	struct sk_buff *skb;
	struct rbuf_info *rinfo;

	rb_ptr = (u64)phys_to_virt(rb_ptr);
	/* Get buffer start address and alignment offset */
	rinfo = GET_RBUF_INFO(rb_ptr);

	/* Now build an skb to give to stack */
	skb = build_skb(rinfo->data, RCV_FRAG_LEN);
	if (!skb) {
		put_page(rinfo->page);
		return NULL;
	}

	/* Set correct skb->data */
	skb_reserve(skb, rinfo->offset + NICVF_RCV_BUF_ALIGN_BYTES);

	prefetch((void *)rb_ptr);
	return skb;
}

/* Allocate RBDR ring and populate receive buffers */
static int  nicvf_init_rbdr(struct nicvf *nic, struct rbdr *rbdr,
			    int ring_len, int buf_size)
{
	int idx;
	u64 *rbuf;
	struct rbdr_entry_t *desc;
	int err;

	err = nicvf_alloc_q_desc_mem(nic, &rbdr->dmem, ring_len,
				     sizeof(struct rbdr_entry_t),
				     NICVF_RCV_BUF_ALIGN_BYTES);
	if (err)
		return err;

	rbdr->desc = rbdr->dmem.base;
	/* Buffer size has to be in multiples of 128 bytes */
	rbdr->dma_size = buf_size;
	rbdr->enable = true;
	rbdr->thresh = RBDR_THRESH;

	nic->rb_page = NULL;
	for (idx = 0; idx < ring_len; idx++) {
		err = nicvf_alloc_rcv_buffer(nic, GFP_KERNEL, RCV_FRAG_LEN,
					     &rbuf);
		if (err)
			return err;

		desc = GET_RBDR_DESC(rbdr, idx);
		desc->buf_addr = virt_to_phys(rbuf) >> NICVF_RCV_BUF_ALIGN;
	}
	return 0;
}

/* Free RBDR ring and its receive buffers */
static void nicvf_free_rbdr(struct nicvf *nic, struct rbdr *rbdr)
{
	int head, tail;
	u64 buf_addr;
	struct rbdr_entry_t *desc;
	struct rbuf_info *rinfo;

	if (!rbdr)
		return;

	rbdr->enable = false;
	if (!rbdr->dmem.base)
		return;

	head = rbdr->head;
	tail = rbdr->tail;

	/* Free SKBs */
	while (head != tail) {
		desc = GET_RBDR_DESC(rbdr, head);
		buf_addr = desc->buf_addr << NICVF_RCV_BUF_ALIGN;
		rinfo = GET_RBUF_INFO((u64)phys_to_virt(buf_addr));
		put_page(rinfo->page);
		head++;
		head &= (rbdr->dmem.q_len - 1);
	}
	/* Free SKB of tail desc */
	desc = GET_RBDR_DESC(rbdr, tail);
	buf_addr = desc->buf_addr << NICVF_RCV_BUF_ALIGN;
	rinfo = GET_RBUF_INFO((u64)phys_to_virt(buf_addr));
	put_page(rinfo->page);

	/* Free RBDR ring */
	nicvf_free_q_desc_mem(nic, &rbdr->dmem);
}

/* Refill receive buffer descriptors with new buffers.
 */
A
Aleksey Makarov 已提交
232
static void nicvf_refill_rbdr(struct nicvf *nic, gfp_t gfp)
233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
{
	struct queue_set *qs = nic->qs;
	int rbdr_idx = qs->rbdr_cnt;
	int tail, qcount;
	int refill_rb_cnt;
	struct rbdr *rbdr;
	struct rbdr_entry_t *desc;
	u64 *rbuf;
	int new_rb = 0;

refill:
	if (!rbdr_idx)
		return;
	rbdr_idx--;
	rbdr = &qs->rbdr[rbdr_idx];
	/* Check if it's enabled */
	if (!rbdr->enable)
		goto next_rbdr;

	/* Get no of desc's to be refilled */
	qcount = nicvf_queue_reg_read(nic, NIC_QSET_RBDR_0_1_STATUS0, rbdr_idx);
	qcount &= 0x7FFFF;
	/* Doorbell can be ringed with a max of ring size minus 1 */
	if (qcount >= (qs->rbdr_len - 1))
		goto next_rbdr;
	else
		refill_rb_cnt = qs->rbdr_len - qcount - 1;

	/* Start filling descs from tail */
	tail = nicvf_queue_reg_read(nic, NIC_QSET_RBDR_0_1_TAIL, rbdr_idx) >> 3;
	while (refill_rb_cnt) {
		tail++;
		tail &= (rbdr->dmem.q_len - 1);

		if (nicvf_alloc_rcv_buffer(nic, gfp, RCV_FRAG_LEN, &rbuf))
			break;

		desc = GET_RBDR_DESC(rbdr, tail);
		desc->buf_addr = virt_to_phys(rbuf) >> NICVF_RCV_BUF_ALIGN;
		refill_rb_cnt--;
		new_rb++;
	}

	/* make sure all memory stores are done before ringing doorbell */
	smp_wmb();

	/* Check if buffer allocation failed */
	if (refill_rb_cnt)
		nic->rb_alloc_fail = true;
	else
		nic->rb_alloc_fail = false;

	/* Notify HW */
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_DOOR,
			      rbdr_idx, new_rb);
next_rbdr:
	/* Re-enable RBDR interrupts only if buffer allocation is success */
	if (!nic->rb_alloc_fail && rbdr->enable)
		nicvf_enable_intr(nic, NICVF_INTR_RBDR, rbdr_idx);

	if (rbdr_idx)
		goto refill;
}

/* Alloc rcv buffers in non-atomic mode for better success */
void nicvf_rbdr_work(struct work_struct *work)
{
	struct nicvf *nic = container_of(work, struct nicvf, rbdr_work.work);

	nicvf_refill_rbdr(nic, GFP_KERNEL);
	if (nic->rb_alloc_fail)
		schedule_delayed_work(&nic->rbdr_work, msecs_to_jiffies(10));
	else
		nic->rb_work_scheduled = false;
}

/* In Softirq context, alloc rcv buffers in atomic mode */
void nicvf_rbdr_task(unsigned long data)
{
	struct nicvf *nic = (struct nicvf *)data;

	nicvf_refill_rbdr(nic, GFP_ATOMIC);
	if (nic->rb_alloc_fail) {
		nic->rb_work_scheduled = true;
		schedule_delayed_work(&nic->rbdr_work, msecs_to_jiffies(10));
	}
}

/* Initialize completion queue */
static int nicvf_init_cmp_queue(struct nicvf *nic,
				struct cmp_queue *cq, int q_len)
{
	int err;

	err = nicvf_alloc_q_desc_mem(nic, &cq->dmem, q_len, CMP_QUEUE_DESC_SIZE,
				     NICVF_CQ_BASE_ALIGN_BYTES);
	if (err)
		return err;

	cq->desc = cq->dmem.base;
	cq->thresh = CMP_QUEUE_CQE_THRESH;
	nic->cq_coalesce_usecs = (CMP_QUEUE_TIMER_THRESH * 0.05) - 1;

	return 0;
}

static void nicvf_free_cmp_queue(struct nicvf *nic, struct cmp_queue *cq)
{
	if (!cq)
		return;
	if (!cq->dmem.base)
		return;

	nicvf_free_q_desc_mem(nic, &cq->dmem);
}

/* Initialize transmit queue */
static int nicvf_init_snd_queue(struct nicvf *nic,
				struct snd_queue *sq, int q_len)
{
	int err;

	err = nicvf_alloc_q_desc_mem(nic, &sq->dmem, q_len, SND_QUEUE_DESC_SIZE,
				     NICVF_SQ_BASE_ALIGN_BYTES);
	if (err)
		return err;

	sq->desc = sq->dmem.base;
361
	sq->skbuff = kcalloc(q_len, sizeof(u64), GFP_KERNEL);
362 363
	if (!sq->skbuff)
		return -ENOMEM;
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386
	sq->head = 0;
	sq->tail = 0;
	atomic_set(&sq->free_cnt, q_len - 1);
	sq->thresh = SND_QUEUE_THRESH;

	/* Preallocate memory for TSO segment's header */
	sq->tso_hdrs = dma_alloc_coherent(&nic->pdev->dev,
					  q_len * TSO_HEADER_SIZE,
					  &sq->tso_hdrs_phys, GFP_KERNEL);
	if (!sq->tso_hdrs)
		return -ENOMEM;

	return 0;
}

static void nicvf_free_snd_queue(struct nicvf *nic, struct snd_queue *sq)
{
	if (!sq)
		return;
	if (!sq->dmem.base)
		return;

	if (sq->tso_hdrs)
387 388
		dma_free_coherent(&nic->pdev->dev,
				  sq->dmem.q_len * TSO_HEADER_SIZE,
389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868
				  sq->tso_hdrs, sq->tso_hdrs_phys);

	kfree(sq->skbuff);
	nicvf_free_q_desc_mem(nic, &sq->dmem);
}

static void nicvf_reclaim_snd_queue(struct nicvf *nic,
				    struct queue_set *qs, int qidx)
{
	/* Disable send queue */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_CFG, qidx, 0);
	/* Check if SQ is stopped */
	if (nicvf_poll_reg(nic, qidx, NIC_QSET_SQ_0_7_STATUS, 21, 1, 0x01))
		return;
	/* Reset send queue */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_CFG, qidx, NICVF_SQ_RESET);
}

static void nicvf_reclaim_rcv_queue(struct nicvf *nic,
				    struct queue_set *qs, int qidx)
{
	union nic_mbx mbx = {};

	/* Make sure all packets in the pipeline are written back into mem */
	mbx.msg.msg = NIC_MBOX_MSG_RQ_SW_SYNC;
	nicvf_send_msg_to_pf(nic, &mbx);
}

static void nicvf_reclaim_cmp_queue(struct nicvf *nic,
				    struct queue_set *qs, int qidx)
{
	/* Disable timer threshold (doesn't get reset upon CQ reset */
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_CFG2, qidx, 0);
	/* Disable completion queue */
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_CFG, qidx, 0);
	/* Reset completion queue */
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_CFG, qidx, NICVF_CQ_RESET);
}

static void nicvf_reclaim_rbdr(struct nicvf *nic,
			       struct rbdr *rbdr, int qidx)
{
	u64 tmp, fifo_state;
	int timeout = 10;

	/* Save head and tail pointers for feeing up buffers */
	rbdr->head = nicvf_queue_reg_read(nic,
					  NIC_QSET_RBDR_0_1_HEAD,
					  qidx) >> 3;
	rbdr->tail = nicvf_queue_reg_read(nic,
					  NIC_QSET_RBDR_0_1_TAIL,
					  qidx) >> 3;

	/* If RBDR FIFO is in 'FAIL' state then do a reset first
	 * before relaiming.
	 */
	fifo_state = nicvf_queue_reg_read(nic, NIC_QSET_RBDR_0_1_STATUS0, qidx);
	if (((fifo_state >> 62) & 0x03) == 0x3)
		nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_CFG,
				      qidx, NICVF_RBDR_RESET);

	/* Disable RBDR */
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_CFG, qidx, 0);
	if (nicvf_poll_reg(nic, qidx, NIC_QSET_RBDR_0_1_STATUS0, 62, 2, 0x00))
		return;
	while (1) {
		tmp = nicvf_queue_reg_read(nic,
					   NIC_QSET_RBDR_0_1_PREFETCH_STATUS,
					   qidx);
		if ((tmp & 0xFFFFFFFF) == ((tmp >> 32) & 0xFFFFFFFF))
			break;
		usleep_range(1000, 2000);
		timeout--;
		if (!timeout) {
			netdev_err(nic->netdev,
				   "Failed polling on prefetch status\n");
			return;
		}
	}
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_CFG,
			      qidx, NICVF_RBDR_RESET);

	if (nicvf_poll_reg(nic, qidx, NIC_QSET_RBDR_0_1_STATUS0, 62, 2, 0x02))
		return;
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_CFG, qidx, 0x00);
	if (nicvf_poll_reg(nic, qidx, NIC_QSET_RBDR_0_1_STATUS0, 62, 2, 0x00))
		return;
}

/* Configures receive queue */
static void nicvf_rcv_queue_config(struct nicvf *nic, struct queue_set *qs,
				   int qidx, bool enable)
{
	union nic_mbx mbx = {};
	struct rcv_queue *rq;
	struct rq_cfg rq_cfg;

	rq = &qs->rq[qidx];
	rq->enable = enable;

	/* Disable receive queue */
	nicvf_queue_reg_write(nic, NIC_QSET_RQ_0_7_CFG, qidx, 0);

	if (!rq->enable) {
		nicvf_reclaim_rcv_queue(nic, qs, qidx);
		return;
	}

	rq->cq_qs = qs->vnic_id;
	rq->cq_idx = qidx;
	rq->start_rbdr_qs = qs->vnic_id;
	rq->start_qs_rbdr_idx = qs->rbdr_cnt - 1;
	rq->cont_rbdr_qs = qs->vnic_id;
	rq->cont_qs_rbdr_idx = qs->rbdr_cnt - 1;
	/* all writes of RBDR data to be loaded into L2 Cache as well*/
	rq->caching = 1;

	/* Send a mailbox msg to PF to config RQ */
	mbx.rq.msg = NIC_MBOX_MSG_RQ_CFG;
	mbx.rq.qs_num = qs->vnic_id;
	mbx.rq.rq_num = qidx;
	mbx.rq.cfg = (rq->caching << 26) | (rq->cq_qs << 19) |
			  (rq->cq_idx << 16) | (rq->cont_rbdr_qs << 9) |
			  (rq->cont_qs_rbdr_idx << 8) |
			  (rq->start_rbdr_qs << 1) | (rq->start_qs_rbdr_idx);
	nicvf_send_msg_to_pf(nic, &mbx);

	mbx.rq.msg = NIC_MBOX_MSG_RQ_BP_CFG;
	mbx.rq.cfg = (1ULL << 63) | (1ULL << 62) | (qs->vnic_id << 0);
	nicvf_send_msg_to_pf(nic, &mbx);

	/* RQ drop config
	 * Enable CQ drop to reserve sufficient CQEs for all tx packets
	 */
	mbx.rq.msg = NIC_MBOX_MSG_RQ_DROP_CFG;
	mbx.rq.cfg = (1ULL << 62) | (RQ_CQ_DROP << 8);
	nicvf_send_msg_to_pf(nic, &mbx);

	nicvf_queue_reg_write(nic, NIC_QSET_RQ_GEN_CFG, qidx, 0x00);

	/* Enable Receive queue */
	rq_cfg.ena = 1;
	rq_cfg.tcp_ena = 0;
	nicvf_queue_reg_write(nic, NIC_QSET_RQ_0_7_CFG, qidx, *(u64 *)&rq_cfg);
}

/* Configures completion queue */
void nicvf_cmp_queue_config(struct nicvf *nic, struct queue_set *qs,
			    int qidx, bool enable)
{
	struct cmp_queue *cq;
	struct cq_cfg cq_cfg;

	cq = &qs->cq[qidx];
	cq->enable = enable;

	if (!cq->enable) {
		nicvf_reclaim_cmp_queue(nic, qs, qidx);
		return;
	}

	/* Reset completion queue */
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_CFG, qidx, NICVF_CQ_RESET);

	if (!cq->enable)
		return;

	spin_lock_init(&cq->lock);
	/* Set completion queue base address */
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_BASE,
			      qidx, (u64)(cq->dmem.phys_base));

	/* Enable Completion queue */
	cq_cfg.ena = 1;
	cq_cfg.reset = 0;
	cq_cfg.caching = 0;
	cq_cfg.qsize = CMP_QSIZE;
	cq_cfg.avg_con = 0;
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_CFG, qidx, *(u64 *)&cq_cfg);

	/* Set threshold value for interrupt generation */
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_THRESH, qidx, cq->thresh);
	nicvf_queue_reg_write(nic, NIC_QSET_CQ_0_7_CFG2,
			      qidx, nic->cq_coalesce_usecs);
}

/* Configures transmit queue */
static void nicvf_snd_queue_config(struct nicvf *nic, struct queue_set *qs,
				   int qidx, bool enable)
{
	union nic_mbx mbx = {};
	struct snd_queue *sq;
	struct sq_cfg sq_cfg;

	sq = &qs->sq[qidx];
	sq->enable = enable;

	if (!sq->enable) {
		nicvf_reclaim_snd_queue(nic, qs, qidx);
		return;
	}

	/* Reset send queue */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_CFG, qidx, NICVF_SQ_RESET);

	sq->cq_qs = qs->vnic_id;
	sq->cq_idx = qidx;

	/* Send a mailbox msg to PF to config SQ */
	mbx.sq.msg = NIC_MBOX_MSG_SQ_CFG;
	mbx.sq.qs_num = qs->vnic_id;
	mbx.sq.sq_num = qidx;
	mbx.sq.cfg = (sq->cq_qs << 3) | sq->cq_idx;
	nicvf_send_msg_to_pf(nic, &mbx);

	/* Set queue base address */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_BASE,
			      qidx, (u64)(sq->dmem.phys_base));

	/* Enable send queue  & set queue size */
	sq_cfg.ena = 1;
	sq_cfg.reset = 0;
	sq_cfg.ldwb = 0;
	sq_cfg.qsize = SND_QSIZE;
	sq_cfg.tstmp_bgx_intf = 0;
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_CFG, qidx, *(u64 *)&sq_cfg);

	/* Set threshold value for interrupt generation */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_THRESH, qidx, sq->thresh);

	/* Set queue:cpu affinity for better load distribution */
	if (cpu_online(qidx)) {
		cpumask_set_cpu(qidx, &sq->affinity_mask);
		netif_set_xps_queue(nic->netdev,
				    &sq->affinity_mask, qidx);
	}
}

/* Configures receive buffer descriptor ring */
static void nicvf_rbdr_config(struct nicvf *nic, struct queue_set *qs,
			      int qidx, bool enable)
{
	struct rbdr *rbdr;
	struct rbdr_cfg rbdr_cfg;

	rbdr = &qs->rbdr[qidx];
	nicvf_reclaim_rbdr(nic, rbdr, qidx);
	if (!enable)
		return;

	/* Set descriptor base address */
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_BASE,
			      qidx, (u64)(rbdr->dmem.phys_base));

	/* Enable RBDR  & set queue size */
	/* Buffer size should be in multiples of 128 bytes */
	rbdr_cfg.ena = 1;
	rbdr_cfg.reset = 0;
	rbdr_cfg.ldwb = 0;
	rbdr_cfg.qsize = RBDR_SIZE;
	rbdr_cfg.avg_con = 0;
	rbdr_cfg.lines = rbdr->dma_size / 128;
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_CFG,
			      qidx, *(u64 *)&rbdr_cfg);

	/* Notify HW */
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_DOOR,
			      qidx, qs->rbdr_len - 1);

	/* Set threshold value for interrupt generation */
	nicvf_queue_reg_write(nic, NIC_QSET_RBDR_0_1_THRESH,
			      qidx, rbdr->thresh - 1);
}

/* Requests PF to assign and enable Qset */
void nicvf_qset_config(struct nicvf *nic, bool enable)
{
	union nic_mbx mbx = {};
	struct queue_set *qs = nic->qs;
	struct qs_cfg *qs_cfg;

	if (!qs) {
		netdev_warn(nic->netdev,
			    "Qset is still not allocated, don't init queues\n");
		return;
	}

	qs->enable = enable;
	qs->vnic_id = nic->vf_id;

	/* Send a mailbox msg to PF to config Qset */
	mbx.qs.msg = NIC_MBOX_MSG_QS_CFG;
	mbx.qs.num = qs->vnic_id;

	mbx.qs.cfg = 0;
	qs_cfg = (struct qs_cfg *)&mbx.qs.cfg;
	if (qs->enable) {
		qs_cfg->ena = 1;
#ifdef __BIG_ENDIAN
		qs_cfg->be = 1;
#endif
		qs_cfg->vnic = qs->vnic_id;
	}
	nicvf_send_msg_to_pf(nic, &mbx);
}

static void nicvf_free_resources(struct nicvf *nic)
{
	int qidx;
	struct queue_set *qs = nic->qs;

	/* Free receive buffer descriptor ring */
	for (qidx = 0; qidx < qs->rbdr_cnt; qidx++)
		nicvf_free_rbdr(nic, &qs->rbdr[qidx]);

	/* Free completion queue */
	for (qidx = 0; qidx < qs->cq_cnt; qidx++)
		nicvf_free_cmp_queue(nic, &qs->cq[qidx]);

	/* Free send queue */
	for (qidx = 0; qidx < qs->sq_cnt; qidx++)
		nicvf_free_snd_queue(nic, &qs->sq[qidx]);
}

static int nicvf_alloc_resources(struct nicvf *nic)
{
	int qidx;
	struct queue_set *qs = nic->qs;

	/* Alloc receive buffer descriptor ring */
	for (qidx = 0; qidx < qs->rbdr_cnt; qidx++) {
		if (nicvf_init_rbdr(nic, &qs->rbdr[qidx], qs->rbdr_len,
				    DMA_BUFFER_LEN))
			goto alloc_fail;
	}

	/* Alloc send queue */
	for (qidx = 0; qidx < qs->sq_cnt; qidx++) {
		if (nicvf_init_snd_queue(nic, &qs->sq[qidx], qs->sq_len))
			goto alloc_fail;
	}

	/* Alloc completion queue */
	for (qidx = 0; qidx < qs->cq_cnt; qidx++) {
		if (nicvf_init_cmp_queue(nic, &qs->cq[qidx], qs->cq_len))
			goto alloc_fail;
	}

	return 0;
alloc_fail:
	nicvf_free_resources(nic);
	return -ENOMEM;
}

int nicvf_set_qset_resources(struct nicvf *nic)
{
	struct queue_set *qs;

	qs = devm_kzalloc(&nic->pdev->dev, sizeof(*qs), GFP_KERNEL);
	if (!qs)
		return -ENOMEM;
	nic->qs = qs;

	/* Set count of each queue */
	qs->rbdr_cnt = RBDR_CNT;
	qs->rq_cnt = RCV_QUEUE_CNT;
	qs->sq_cnt = SND_QUEUE_CNT;
	qs->cq_cnt = CMP_QUEUE_CNT;

	/* Set queue lengths */
	qs->rbdr_len = RCV_BUF_COUNT;
	qs->sq_len = SND_QUEUE_LEN;
	qs->cq_len = CMP_QUEUE_LEN;
	return 0;
}

int nicvf_config_data_transfer(struct nicvf *nic, bool enable)
{
	bool disable = false;
	struct queue_set *qs = nic->qs;
	int qidx;

	if (!qs)
		return 0;

	if (enable) {
		if (nicvf_alloc_resources(nic))
			return -ENOMEM;

		for (qidx = 0; qidx < qs->sq_cnt; qidx++)
			nicvf_snd_queue_config(nic, qs, qidx, enable);
		for (qidx = 0; qidx < qs->cq_cnt; qidx++)
			nicvf_cmp_queue_config(nic, qs, qidx, enable);
		for (qidx = 0; qidx < qs->rbdr_cnt; qidx++)
			nicvf_rbdr_config(nic, qs, qidx, enable);
		for (qidx = 0; qidx < qs->rq_cnt; qidx++)
			nicvf_rcv_queue_config(nic, qs, qidx, enable);
	} else {
		for (qidx = 0; qidx < qs->rq_cnt; qidx++)
			nicvf_rcv_queue_config(nic, qs, qidx, disable);
		for (qidx = 0; qidx < qs->rbdr_cnt; qidx++)
			nicvf_rbdr_config(nic, qs, qidx, disable);
		for (qidx = 0; qidx < qs->sq_cnt; qidx++)
			nicvf_snd_queue_config(nic, qs, qidx, disable);
		for (qidx = 0; qidx < qs->cq_cnt; qidx++)
			nicvf_cmp_queue_config(nic, qs, qidx, disable);

		nicvf_free_resources(nic);
	}

	return 0;
}

/* Get a free desc from SQ
 * returns descriptor ponter & descriptor number
 */
static inline int nicvf_get_sq_desc(struct snd_queue *sq, int desc_cnt)
{
	int qentry;

	qentry = sq->tail;
	atomic_sub(desc_cnt, &sq->free_cnt);
	sq->tail += desc_cnt;
	sq->tail &= (sq->dmem.q_len - 1);

	return qentry;
}

/* Free descriptor back to SQ for future use */
void nicvf_put_sq_desc(struct snd_queue *sq, int desc_cnt)
{
	atomic_add(desc_cnt, &sq->free_cnt);
	sq->head += desc_cnt;
	sq->head &= (sq->dmem.q_len - 1);
}

static inline int nicvf_get_nxt_sqentry(struct snd_queue *sq, int qentry)
{
	qentry++;
	qentry &= (sq->dmem.q_len - 1);
	return qentry;
}

void nicvf_sq_enable(struct nicvf *nic, struct snd_queue *sq, int qidx)
{
	u64 sq_cfg;

	sq_cfg = nicvf_queue_reg_read(nic, NIC_QSET_SQ_0_7_CFG, qidx);
	sq_cfg |= NICVF_SQ_EN;
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_CFG, qidx, sq_cfg);
	/* Ring doorbell so that H/W restarts processing SQEs */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_DOOR, qidx, 0);
}

void nicvf_sq_disable(struct nicvf *nic, int qidx)
{
	u64 sq_cfg;

	sq_cfg = nicvf_queue_reg_read(nic, NIC_QSET_SQ_0_7_CFG, qidx);
	sq_cfg &= ~NICVF_SQ_EN;
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_CFG, qidx, sq_cfg);
}

void nicvf_sq_free_used_descs(struct net_device *netdev, struct snd_queue *sq,
			      int qidx)
{
	u64 head, tail;
	struct sk_buff *skb;
	struct nicvf *nic = netdev_priv(netdev);
	struct sq_hdr_subdesc *hdr;

	head = nicvf_queue_reg_read(nic, NIC_QSET_SQ_0_7_HEAD, qidx) >> 4;
	tail = nicvf_queue_reg_read(nic, NIC_QSET_SQ_0_7_TAIL, qidx) >> 4;
	while (sq->head != head) {
		hdr = (struct sq_hdr_subdesc *)GET_SQ_DESC(sq, sq->head);
		if (hdr->subdesc_type != SQ_DESC_TYPE_HEADER) {
			nicvf_put_sq_desc(sq, 1);
			continue;
		}
		skb = (struct sk_buff *)sq->skbuff[sq->head];
869 870
		if (skb)
			dev_kfree_skb_any(skb);
871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998
		atomic64_add(1, (atomic64_t *)&netdev->stats.tx_packets);
		atomic64_add(hdr->tot_len,
			     (atomic64_t *)&netdev->stats.tx_bytes);
		nicvf_put_sq_desc(sq, hdr->subdesc_cnt + 1);
	}
}

/* Calculate no of SQ subdescriptors needed to transmit all
 * segments of this TSO packet.
 * Taken from 'Tilera network driver' with a minor modification.
 */
static int nicvf_tso_count_subdescs(struct sk_buff *skb)
{
	struct skb_shared_info *sh = skb_shinfo(skb);
	unsigned int sh_len = skb_transport_offset(skb) + tcp_hdrlen(skb);
	unsigned int data_len = skb->len - sh_len;
	unsigned int p_len = sh->gso_size;
	long f_id = -1;    /* id of the current fragment */
	long f_size = skb_headlen(skb) - sh_len;  /* current fragment size */
	long f_used = 0;  /* bytes used from the current fragment */
	long n;            /* size of the current piece of payload */
	int num_edescs = 0;
	int segment;

	for (segment = 0; segment < sh->gso_segs; segment++) {
		unsigned int p_used = 0;

		/* One edesc for header and for each piece of the payload. */
		for (num_edescs++; p_used < p_len; num_edescs++) {
			/* Advance as needed. */
			while (f_used >= f_size) {
				f_id++;
				f_size = skb_frag_size(&sh->frags[f_id]);
				f_used = 0;
			}

			/* Use bytes from the current fragment. */
			n = p_len - p_used;
			if (n > f_size - f_used)
				n = f_size - f_used;
			f_used += n;
			p_used += n;
		}

		/* The last segment may be less than gso_size. */
		data_len -= p_len;
		if (data_len < p_len)
			p_len = data_len;
	}

	/* '+ gso_segs' for SQ_HDR_SUDESCs for each segment */
	return num_edescs + sh->gso_segs;
}

/* Get the number of SQ descriptors needed to xmit this skb */
static int nicvf_sq_subdesc_required(struct nicvf *nic, struct sk_buff *skb)
{
	int subdesc_cnt = MIN_SQ_DESC_PER_PKT_XMIT;

	if (skb_shinfo(skb)->gso_size) {
		subdesc_cnt = nicvf_tso_count_subdescs(skb);
		return subdesc_cnt;
	}

	if (skb_shinfo(skb)->nr_frags)
		subdesc_cnt += skb_shinfo(skb)->nr_frags;

	return subdesc_cnt;
}

/* Add SQ HEADER subdescriptor.
 * First subdescriptor for every send descriptor.
 */
static inline void
nicvf_sq_add_hdr_subdesc(struct snd_queue *sq, int qentry,
			 int subdesc_cnt, struct sk_buff *skb, int len)
{
	int proto;
	struct sq_hdr_subdesc *hdr;

	hdr = (struct sq_hdr_subdesc *)GET_SQ_DESC(sq, qentry);
	sq->skbuff[qentry] = (u64)skb;

	memset(hdr, 0, SND_QUEUE_DESC_SIZE);
	hdr->subdesc_type = SQ_DESC_TYPE_HEADER;
	/* Enable notification via CQE after processing SQE */
	hdr->post_cqe = 1;
	/* No of subdescriptors following this */
	hdr->subdesc_cnt = subdesc_cnt;
	hdr->tot_len = len;

	/* Offload checksum calculation to HW */
	if (skb->ip_summed == CHECKSUM_PARTIAL) {
		if (skb->protocol != htons(ETH_P_IP))
			return;

		hdr->csum_l3 = 1; /* Enable IP csum calculation */
		hdr->l3_offset = skb_network_offset(skb);
		hdr->l4_offset = skb_transport_offset(skb);

		proto = ip_hdr(skb)->protocol;
		switch (proto) {
		case IPPROTO_TCP:
			hdr->csum_l4 = SEND_L4_CSUM_TCP;
			break;
		case IPPROTO_UDP:
			hdr->csum_l4 = SEND_L4_CSUM_UDP;
			break;
		case IPPROTO_SCTP:
			hdr->csum_l4 = SEND_L4_CSUM_SCTP;
			break;
		}
	}
}

/* SQ GATHER subdescriptor
 * Must follow HDR descriptor
 */
static inline void nicvf_sq_add_gather_subdesc(struct snd_queue *sq, int qentry,
					       int size, u64 data)
{
	struct sq_gather_subdesc *gather;

	qentry &= (sq->dmem.q_len - 1);
	gather = (struct sq_gather_subdesc *)GET_SQ_DESC(sq, qentry);

	memset(gather, 0, SND_QUEUE_DESC_SIZE);
	gather->subdesc_type = SQ_DESC_TYPE_GATHER;
999
	gather->ld_type = NIC_SEND_LD_TYPE_E_LDD;
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
	gather->size = size;
	gather->addr = data;
}

/* Segment a TSO packet into 'gso_size' segments and append
 * them to SQ for transfer
 */
static int nicvf_sq_append_tso(struct nicvf *nic, struct snd_queue *sq,
			       int qentry, struct sk_buff *skb)
{
	struct tso_t tso;
	int seg_subdescs = 0, desc_cnt = 0;
	int seg_len, total_len, data_left;
	int hdr_qentry = qentry;
	int hdr_len = skb_transport_offset(skb) + tcp_hdrlen(skb);

	tso_start(skb, &tso);
	total_len = skb->len - hdr_len;
	while (total_len > 0) {
		char *hdr;

		/* Save Qentry for adding HDR_SUBDESC at the end */
		hdr_qentry = qentry;

		data_left = min_t(int, skb_shinfo(skb)->gso_size, total_len);
		total_len -= data_left;

		/* Add segment's header */
		qentry = nicvf_get_nxt_sqentry(sq, qentry);
		hdr = sq->tso_hdrs + qentry * TSO_HEADER_SIZE;
		tso_build_hdr(skb, hdr, &tso, data_left, total_len == 0);
		nicvf_sq_add_gather_subdesc(sq, qentry, hdr_len,
					    sq->tso_hdrs_phys +
					    qentry * TSO_HEADER_SIZE);
		/* HDR_SUDESC + GATHER */
		seg_subdescs = 2;
		seg_len = hdr_len;

		/* Add segment's payload fragments */
		while (data_left > 0) {
			int size;

			size = min_t(int, tso.size, data_left);

			qentry = nicvf_get_nxt_sqentry(sq, qentry);
			nicvf_sq_add_gather_subdesc(sq, qentry, size,
						    virt_to_phys(tso.data));
			seg_subdescs++;
			seg_len += size;

			data_left -= size;
			tso_build_data(skb, &tso, size);
		}
		nicvf_sq_add_hdr_subdesc(sq, hdr_qentry,
					 seg_subdescs - 1, skb, seg_len);
1055
		sq->skbuff[hdr_qentry] = (u64)NULL;
1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068
		qentry = nicvf_get_nxt_sqentry(sq, qentry);

		desc_cnt += seg_subdescs;
	}
	/* Save SKB in the last segment for freeing */
	sq->skbuff[hdr_qentry] = (u64)skb;

	/* make sure all memory stores are done before ringing doorbell */
	smp_wmb();

	/* Inform HW to xmit all TSO segments */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_DOOR,
			      skb_get_queue_mapping(skb), desc_cnt);
1069
	nic->drv_stats.tx_tso++;
1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550
	return 1;
}

/* Append an skb to a SQ for packet transfer. */
int nicvf_sq_append_skb(struct nicvf *nic, struct sk_buff *skb)
{
	int i, size;
	int subdesc_cnt;
	int sq_num, qentry;
	struct queue_set *qs = nic->qs;
	struct snd_queue *sq;

	sq_num = skb_get_queue_mapping(skb);
	sq = &qs->sq[sq_num];

	subdesc_cnt = nicvf_sq_subdesc_required(nic, skb);
	if (subdesc_cnt > atomic_read(&sq->free_cnt))
		goto append_fail;

	qentry = nicvf_get_sq_desc(sq, subdesc_cnt);

	/* Check if its a TSO packet */
	if (skb_shinfo(skb)->gso_size)
		return nicvf_sq_append_tso(nic, sq, qentry, skb);

	/* Add SQ header subdesc */
	nicvf_sq_add_hdr_subdesc(sq, qentry, subdesc_cnt - 1, skb, skb->len);

	/* Add SQ gather subdescs */
	qentry = nicvf_get_nxt_sqentry(sq, qentry);
	size = skb_is_nonlinear(skb) ? skb_headlen(skb) : skb->len;
	nicvf_sq_add_gather_subdesc(sq, qentry, size, virt_to_phys(skb->data));

	/* Check for scattered buffer */
	if (!skb_is_nonlinear(skb))
		goto doorbell;

	for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
		const struct skb_frag_struct *frag;

		frag = &skb_shinfo(skb)->frags[i];

		qentry = nicvf_get_nxt_sqentry(sq, qentry);
		size = skb_frag_size(frag);
		nicvf_sq_add_gather_subdesc(sq, qentry, size,
					    virt_to_phys(
					    skb_frag_address(frag)));
	}

doorbell:
	/* make sure all memory stores are done before ringing doorbell */
	smp_wmb();

	/* Inform HW to xmit new packet */
	nicvf_queue_reg_write(nic, NIC_QSET_SQ_0_7_DOOR,
			      sq_num, subdesc_cnt);
	return 1;

append_fail:
	netdev_dbg(nic->netdev, "Not enough SQ descriptors to xmit pkt\n");
	return 0;
}

static inline unsigned frag_num(unsigned i)
{
#ifdef __BIG_ENDIAN
	return (i & ~3) + 3 - (i & 3);
#else
	return i;
#endif
}

/* Returns SKB for a received packet */
struct sk_buff *nicvf_get_rcv_skb(struct nicvf *nic, struct cqe_rx_t *cqe_rx)
{
	int frag;
	int payload_len = 0;
	struct sk_buff *skb = NULL;
	struct sk_buff *skb_frag = NULL;
	struct sk_buff *prev_frag = NULL;
	u16 *rb_lens = NULL;
	u64 *rb_ptrs = NULL;

	rb_lens = (void *)cqe_rx + (3 * sizeof(u64));
	rb_ptrs = (void *)cqe_rx + (6 * sizeof(u64));

	netdev_dbg(nic->netdev, "%s rb_cnt %d rb0_ptr %llx rb0_sz %d\n",
		   __func__, cqe_rx->rb_cnt, cqe_rx->rb0_ptr, cqe_rx->rb0_sz);

	for (frag = 0; frag < cqe_rx->rb_cnt; frag++) {
		payload_len = rb_lens[frag_num(frag)];
		if (!frag) {
			/* First fragment */
			skb = nicvf_rb_ptr_to_skb(nic,
						  *rb_ptrs - cqe_rx->align_pad,
						  payload_len);
			if (!skb)
				return NULL;
			skb_reserve(skb, cqe_rx->align_pad);
			skb_put(skb, payload_len);
		} else {
			/* Add fragments */
			skb_frag = nicvf_rb_ptr_to_skb(nic, *rb_ptrs,
						       payload_len);
			if (!skb_frag) {
				dev_kfree_skb(skb);
				return NULL;
			}

			if (!skb_shinfo(skb)->frag_list)
				skb_shinfo(skb)->frag_list = skb_frag;
			else
				prev_frag->next = skb_frag;

			prev_frag = skb_frag;
			skb->len += payload_len;
			skb->data_len += payload_len;
			skb_frag->len = payload_len;
		}
		/* Next buffer pointer */
		rb_ptrs++;
	}
	return skb;
}

/* Enable interrupt */
void nicvf_enable_intr(struct nicvf *nic, int int_type, int q_idx)
{
	u64 reg_val;

	reg_val = nicvf_reg_read(nic, NIC_VF_ENA_W1S);

	switch (int_type) {
	case NICVF_INTR_CQ:
		reg_val |= ((1ULL << q_idx) << NICVF_INTR_CQ_SHIFT);
		break;
	case NICVF_INTR_SQ:
		reg_val |= ((1ULL << q_idx) << NICVF_INTR_SQ_SHIFT);
		break;
	case NICVF_INTR_RBDR:
		reg_val |= ((1ULL << q_idx) << NICVF_INTR_RBDR_SHIFT);
		break;
	case NICVF_INTR_PKT_DROP:
		reg_val |= (1ULL << NICVF_INTR_PKT_DROP_SHIFT);
		break;
	case NICVF_INTR_TCP_TIMER:
		reg_val |= (1ULL << NICVF_INTR_TCP_TIMER_SHIFT);
		break;
	case NICVF_INTR_MBOX:
		reg_val |= (1ULL << NICVF_INTR_MBOX_SHIFT);
		break;
	case NICVF_INTR_QS_ERR:
		reg_val |= (1ULL << NICVF_INTR_QS_ERR_SHIFT);
		break;
	default:
		netdev_err(nic->netdev,
			   "Failed to enable interrupt: unknown type\n");
		break;
	}

	nicvf_reg_write(nic, NIC_VF_ENA_W1S, reg_val);
}

/* Disable interrupt */
void nicvf_disable_intr(struct nicvf *nic, int int_type, int q_idx)
{
	u64 reg_val = 0;

	switch (int_type) {
	case NICVF_INTR_CQ:
		reg_val |= ((1ULL << q_idx) << NICVF_INTR_CQ_SHIFT);
		break;
	case NICVF_INTR_SQ:
		reg_val |= ((1ULL << q_idx) << NICVF_INTR_SQ_SHIFT);
		break;
	case NICVF_INTR_RBDR:
		reg_val |= ((1ULL << q_idx) << NICVF_INTR_RBDR_SHIFT);
		break;
	case NICVF_INTR_PKT_DROP:
		reg_val |= (1ULL << NICVF_INTR_PKT_DROP_SHIFT);
		break;
	case NICVF_INTR_TCP_TIMER:
		reg_val |= (1ULL << NICVF_INTR_TCP_TIMER_SHIFT);
		break;
	case NICVF_INTR_MBOX:
		reg_val |= (1ULL << NICVF_INTR_MBOX_SHIFT);
		break;
	case NICVF_INTR_QS_ERR:
		reg_val |= (1ULL << NICVF_INTR_QS_ERR_SHIFT);
		break;
	default:
		netdev_err(nic->netdev,
			   "Failed to disable interrupt: unknown type\n");
		break;
	}

	nicvf_reg_write(nic, NIC_VF_ENA_W1C, reg_val);
}

/* Clear interrupt */
void nicvf_clear_intr(struct nicvf *nic, int int_type, int q_idx)
{
	u64 reg_val = 0;

	switch (int_type) {
	case NICVF_INTR_CQ:
		reg_val = ((1ULL << q_idx) << NICVF_INTR_CQ_SHIFT);
		break;
	case NICVF_INTR_SQ:
		reg_val = ((1ULL << q_idx) << NICVF_INTR_SQ_SHIFT);
		break;
	case NICVF_INTR_RBDR:
		reg_val = ((1ULL << q_idx) << NICVF_INTR_RBDR_SHIFT);
		break;
	case NICVF_INTR_PKT_DROP:
		reg_val = (1ULL << NICVF_INTR_PKT_DROP_SHIFT);
		break;
	case NICVF_INTR_TCP_TIMER:
		reg_val = (1ULL << NICVF_INTR_TCP_TIMER_SHIFT);
		break;
	case NICVF_INTR_MBOX:
		reg_val = (1ULL << NICVF_INTR_MBOX_SHIFT);
		break;
	case NICVF_INTR_QS_ERR:
		reg_val |= (1ULL << NICVF_INTR_QS_ERR_SHIFT);
		break;
	default:
		netdev_err(nic->netdev,
			   "Failed to clear interrupt: unknown type\n");
		break;
	}

	nicvf_reg_write(nic, NIC_VF_INT, reg_val);
}

/* Check if interrupt is enabled */
int nicvf_is_intr_enabled(struct nicvf *nic, int int_type, int q_idx)
{
	u64 reg_val;
	u64 mask = 0xff;

	reg_val = nicvf_reg_read(nic, NIC_VF_ENA_W1S);

	switch (int_type) {
	case NICVF_INTR_CQ:
		mask = ((1ULL << q_idx) << NICVF_INTR_CQ_SHIFT);
		break;
	case NICVF_INTR_SQ:
		mask = ((1ULL << q_idx) << NICVF_INTR_SQ_SHIFT);
		break;
	case NICVF_INTR_RBDR:
		mask = ((1ULL << q_idx) << NICVF_INTR_RBDR_SHIFT);
		break;
	case NICVF_INTR_PKT_DROP:
		mask = NICVF_INTR_PKT_DROP_MASK;
		break;
	case NICVF_INTR_TCP_TIMER:
		mask = NICVF_INTR_TCP_TIMER_MASK;
		break;
	case NICVF_INTR_MBOX:
		mask = NICVF_INTR_MBOX_MASK;
		break;
	case NICVF_INTR_QS_ERR:
		mask = NICVF_INTR_QS_ERR_MASK;
		break;
	default:
		netdev_err(nic->netdev,
			   "Failed to check interrupt enable: unknown type\n");
		break;
	}

	return (reg_val & mask);
}

void nicvf_update_rq_stats(struct nicvf *nic, int rq_idx)
{
	struct rcv_queue *rq;

#define GET_RQ_STATS(reg) \
	nicvf_reg_read(nic, NIC_QSET_RQ_0_7_STAT_0_1 |\
			    (rq_idx << NIC_Q_NUM_SHIFT) | (reg << 3))

	rq = &nic->qs->rq[rq_idx];
	rq->stats.bytes = GET_RQ_STATS(RQ_SQ_STATS_OCTS);
	rq->stats.pkts = GET_RQ_STATS(RQ_SQ_STATS_PKTS);
}

void nicvf_update_sq_stats(struct nicvf *nic, int sq_idx)
{
	struct snd_queue *sq;

#define GET_SQ_STATS(reg) \
	nicvf_reg_read(nic, NIC_QSET_SQ_0_7_STAT_0_1 |\
			    (sq_idx << NIC_Q_NUM_SHIFT) | (reg << 3))

	sq = &nic->qs->sq[sq_idx];
	sq->stats.bytes = GET_SQ_STATS(RQ_SQ_STATS_OCTS);
	sq->stats.pkts = GET_SQ_STATS(RQ_SQ_STATS_PKTS);
}

/* Check for errors in the receive cmp.queue entry */
int nicvf_check_cqe_rx_errs(struct nicvf *nic,
			    struct cmp_queue *cq, struct cqe_rx_t *cqe_rx)
{
	struct cmp_queue_stats *stats = &cq->stats;

	if (!cqe_rx->err_level && !cqe_rx->err_opcode) {
		stats->rx.errop.good++;
		return 0;
	}

	if (netif_msg_rx_err(nic))
		netdev_err(nic->netdev,
			   "%s: RX error CQE err_level 0x%x err_opcode 0x%x\n",
			   nic->netdev->name,
			   cqe_rx->err_level, cqe_rx->err_opcode);

	switch (cqe_rx->err_level) {
	case CQ_ERRLVL_MAC:
		stats->rx.errlvl.mac_errs++;
		break;
	case CQ_ERRLVL_L2:
		stats->rx.errlvl.l2_errs++;
		break;
	case CQ_ERRLVL_L3:
		stats->rx.errlvl.l3_errs++;
		break;
	case CQ_ERRLVL_L4:
		stats->rx.errlvl.l4_errs++;
		break;
	}

	switch (cqe_rx->err_opcode) {
	case CQ_RX_ERROP_RE_PARTIAL:
		stats->rx.errop.partial_pkts++;
		break;
	case CQ_RX_ERROP_RE_JABBER:
		stats->rx.errop.jabber_errs++;
		break;
	case CQ_RX_ERROP_RE_FCS:
		stats->rx.errop.fcs_errs++;
		break;
	case CQ_RX_ERROP_RE_TERMINATE:
		stats->rx.errop.terminate_errs++;
		break;
	case CQ_RX_ERROP_RE_RX_CTL:
		stats->rx.errop.bgx_rx_errs++;
		break;
	case CQ_RX_ERROP_PREL2_ERR:
		stats->rx.errop.prel2_errs++;
		break;
	case CQ_RX_ERROP_L2_FRAGMENT:
		stats->rx.errop.l2_frags++;
		break;
	case CQ_RX_ERROP_L2_OVERRUN:
		stats->rx.errop.l2_overruns++;
		break;
	case CQ_RX_ERROP_L2_PFCS:
		stats->rx.errop.l2_pfcs++;
		break;
	case CQ_RX_ERROP_L2_PUNY:
		stats->rx.errop.l2_puny++;
		break;
	case CQ_RX_ERROP_L2_MAL:
		stats->rx.errop.l2_hdr_malformed++;
		break;
	case CQ_RX_ERROP_L2_OVERSIZE:
		stats->rx.errop.l2_oversize++;
		break;
	case CQ_RX_ERROP_L2_UNDERSIZE:
		stats->rx.errop.l2_undersize++;
		break;
	case CQ_RX_ERROP_L2_LENMISM:
		stats->rx.errop.l2_len_mismatch++;
		break;
	case CQ_RX_ERROP_L2_PCLP:
		stats->rx.errop.l2_pclp++;
		break;
	case CQ_RX_ERROP_IP_NOT:
		stats->rx.errop.non_ip++;
		break;
	case CQ_RX_ERROP_IP_CSUM_ERR:
		stats->rx.errop.ip_csum_err++;
		break;
	case CQ_RX_ERROP_IP_MAL:
		stats->rx.errop.ip_hdr_malformed++;
		break;
	case CQ_RX_ERROP_IP_MALD:
		stats->rx.errop.ip_payload_malformed++;
		break;
	case CQ_RX_ERROP_IP_HOP:
		stats->rx.errop.ip_hop_errs++;
		break;
	case CQ_RX_ERROP_L3_ICRC:
		stats->rx.errop.l3_icrc_errs++;
		break;
	case CQ_RX_ERROP_L3_PCLP:
		stats->rx.errop.l3_pclp++;
		break;
	case CQ_RX_ERROP_L4_MAL:
		stats->rx.errop.l4_malformed++;
		break;
	case CQ_RX_ERROP_L4_CHK:
		stats->rx.errop.l4_csum_errs++;
		break;
	case CQ_RX_ERROP_UDP_LEN:
		stats->rx.errop.udp_len_err++;
		break;
	case CQ_RX_ERROP_L4_PORT:
		stats->rx.errop.bad_l4_port++;
		break;
	case CQ_RX_ERROP_TCP_FLAG:
		stats->rx.errop.bad_tcp_flag++;
		break;
	case CQ_RX_ERROP_TCP_OFFSET:
		stats->rx.errop.tcp_offset_errs++;
		break;
	case CQ_RX_ERROP_L4_PCLP:
		stats->rx.errop.l4_pclp++;
		break;
	case CQ_RX_ERROP_RBDR_TRUNC:
		stats->rx.errop.pkt_truncated++;
		break;
	}

	return 1;
}

/* Check for errors in the send cmp.queue entry */
int nicvf_check_cqe_tx_errs(struct nicvf *nic,
			    struct cmp_queue *cq, struct cqe_send_t *cqe_tx)
{
	struct cmp_queue_stats *stats = &cq->stats;

	switch (cqe_tx->send_status) {
	case CQ_TX_ERROP_GOOD:
		stats->tx.good++;
		return 0;
	case CQ_TX_ERROP_DESC_FAULT:
		stats->tx.desc_fault++;
		break;
	case CQ_TX_ERROP_HDR_CONS_ERR:
		stats->tx.hdr_cons_err++;
		break;
	case CQ_TX_ERROP_SUBDC_ERR:
		stats->tx.subdesc_err++;
		break;
	case CQ_TX_ERROP_IMM_SIZE_OFLOW:
		stats->tx.imm_size_oflow++;
		break;
	case CQ_TX_ERROP_DATA_SEQUENCE_ERR:
		stats->tx.data_seq_err++;
		break;
	case CQ_TX_ERROP_MEM_SEQUENCE_ERR:
		stats->tx.mem_seq_err++;
		break;
	case CQ_TX_ERROP_LOCK_VIOL:
		stats->tx.lock_viol++;
		break;
	case CQ_TX_ERROP_DATA_FAULT:
		stats->tx.data_fault++;
		break;
	case CQ_TX_ERROP_TSTMP_CONFLICT:
		stats->tx.tstmp_conflict++;
		break;
	case CQ_TX_ERROP_TSTMP_TIMEOUT:
		stats->tx.tstmp_timeout++;
		break;
	case CQ_TX_ERROP_MEM_FAULT:
		stats->tx.mem_fault++;
		break;
	case CQ_TX_ERROP_CK_OVERLAP:
		stats->tx.csum_overlap++;
		break;
	case CQ_TX_ERROP_CK_OFLOW:
		stats->tx.csum_overflow++;
		break;
	}

	return 1;
}