android_kernel_samsung_sm8650-modules/dp/wifi3.0/dp_rx.c

/*
 * Copyright (c) 2016-2017 The Linux Foundation. All rights reserved.
 *
 * Permission to use, copy, modify, and/or distribute this software for
 * any purpose with or without fee is hereby granted, provided that the
 * above copyright notice and this permission notice appear in all
 * copies.
 *
 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL
 * WARRANTIES WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED
 * WARRANTIES OF MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE
 * AUTHOR BE LIABLE FOR ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL
 * DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR
 * PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
 * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR
 * PERFORMANCE OF THIS SOFTWARE.
 */

#include "dp_types.h"
#include "dp_rx.h"
#include "dp_peer.h"
#include "hal_rx.h"
#include "hal_api.h"
#include "qdf_nbuf.h"
#include <ieee80211.h>
#ifdef MESH_MODE_SUPPORT
#include "if_meta_hdr.h"
#endif
#include "dp_internal.h"

/*
 * dp_rx_buffers_replenish() - replenish rxdma ring with rx nbufs
 *			       called during dp rx initialization
 *			       and at the end of dp_rx_process.
 *
 * @soc: core txrx main context
 * @mac_id: mac_id which is one of 3 mac_ids
 * @dp_rxdma_srng: dp rxdma circular ring
 * @rx_desc_pool: Poiter to free Rx descriptor pool
 * @num_req_buffers: number of buffer to be replenished
 * @desc_list: list of descs if called from dp_rx_process
 *	       or NULL during dp rx initialization or out of buffer
 *	       interrupt.
 * @tail: tail of descs list
 * @owner: who owns the nbuf (host, NSS etc...)
 * Return: return success or failure
 */
QDF_STATUS dp_rx_buffers_replenish(struct dp_soc *dp_soc, uint32_t mac_id,
				struct dp_srng *dp_rxdma_srng,
				struct rx_desc_pool *rx_desc_pool,
				uint32_t num_req_buffers,
				union dp_rx_desc_list_elem_t **desc_list,
				union dp_rx_desc_list_elem_t **tail,
				uint8_t owner)
{
	uint32_t num_alloc_desc;
	uint16_t num_desc_to_free = 0;
	struct dp_pdev *dp_pdev = dp_soc->pdev_list[mac_id];
	uint32_t num_entries_avail;
	uint32_t count;
	int sync_hw_ptr = 1;
	qdf_dma_addr_t paddr;
	qdf_nbuf_t rx_netbuf;
	void *rxdma_ring_entry;
	union dp_rx_desc_list_elem_t *next;
	QDF_STATUS ret;

	void *rxdma_srng;

	rxdma_srng = dp_rxdma_srng->hal_srng;

	if (!rxdma_srng) {
		QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
			"rxdma srng not initialized");
		DP_STATS_INC(dp_pdev, err.rxdma_unitialized, 1);
		return QDF_STATUS_E_FAILURE;
	}

	QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_DEBUG,
		"requested %d buffers for replenish", num_req_buffers);

	/*
	 * if desc_list is NULL, allocate the descs from freelist
	 */
	if (!(*desc_list)) {

		num_alloc_desc = dp_rx_get_free_desc_list(dp_soc, mac_id,
							  rx_desc_pool,
							  num_req_buffers,
							  desc_list,
							  tail);

		if (!num_alloc_desc) {
			QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
				"no free rx_descs in freelist");
			DP_STATS_INC(dp_pdev, err.desc_alloc_fail,
					num_alloc_desc);
			return QDF_STATUS_E_NOMEM;
		}

		QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_DEBUG,
			"%d rx desc allocated", num_alloc_desc);
		num_req_buffers = num_alloc_desc;
	}

	hal_srng_access_start(dp_soc->hal_soc, rxdma_srng);
	num_entries_avail = hal_srng_src_num_avail(dp_soc->hal_soc,
						   rxdma_srng,
						   sync_hw_ptr);

	QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_DEBUG,
			"no of availble entries in rxdma ring: %d",
			num_entries_avail);

	if (num_entries_avail < num_req_buffers) {
		num_desc_to_free = num_req_buffers - num_entries_avail;
		num_req_buffers = num_entries_avail;
	}

	count = 0;

	while (count < num_req_buffers) {
		rx_netbuf = qdf_nbuf_alloc(dp_pdev->osif_pdev,
					RX_BUFFER_SIZE,
					RX_BUFFER_RESERVATION,
					RX_BUFFER_ALIGNMENT,
					FALSE);

		if (rx_netbuf == NULL)
			continue;

		qdf_nbuf_map_single(dp_soc->osdev, rx_netbuf,
				    QDF_DMA_BIDIRECTIONAL);

		paddr = qdf_nbuf_get_frag_paddr(rx_netbuf, 0);

		/*
		 * check if the physical address of nbuf->data is
		 * less then 0x50000000 then free the nbuf and try
		 * allocating new nbuf. We can try for 100 times.
		 * this is a temp WAR till we fix it properly.
		 */
		ret = check_x86_paddr(dp_soc, &rx_netbuf, &paddr, dp_pdev);
		if (ret == QDF_STATUS_E_FAILURE)
			break;

		count++;

		rxdma_ring_entry = hal_srng_src_get_next(dp_soc->hal_soc,
								rxdma_srng);

		next = (*desc_list)->next;

		(*desc_list)->rx_desc.nbuf = rx_netbuf;
		DP_STATS_INC_PKT(dp_pdev, replenished, 1,
				qdf_nbuf_len(rx_netbuf));

		QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_DEBUG,
				"rx_netbuf=%p, buf=%p, paddr=0x%llx, cookie=%d\n",
			rx_netbuf, qdf_nbuf_data(rx_netbuf),
			(unsigned long long)paddr, (*desc_list)->rx_desc.cookie);

		hal_rxdma_buff_addr_info_set(rxdma_ring_entry, paddr,
						(*desc_list)->rx_desc.cookie,
						owner);

		*desc_list = next;
	}

	hal_srng_access_end(dp_soc->hal_soc, rxdma_srng);

	QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_DEBUG,
		"successfully replenished %d buffers", num_req_buffers);
	QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_DEBUG,
		"%d rx desc added back to free list", num_desc_to_free);
	DP_STATS_INC(dp_pdev, buf_freelist, num_desc_to_free);

	/*
	 * add any available free desc back to the free list
	 */

	if (*desc_list)
		dp_rx_add_desc_list_to_free_list(dp_soc, desc_list, tail,
			mac_id, rx_desc_pool);

	return QDF_STATUS_SUCCESS;
}

/*
 * dp_rx_deliver_raw() - process RAW mode pkts and hand over the
 *				pkts to RAW mode simulation to
 *				decapsulate the pkt.
 *
 * @vdev: vdev on which RAW mode is enabled
 * @nbuf_list: list of RAW pkts to process
 *
 * Return: void
 */
static void
dp_rx_deliver_raw(struct dp_vdev *vdev, qdf_nbuf_t nbuf_list)
{
	qdf_nbuf_t deliver_list_head = NULL;
	qdf_nbuf_t deliver_list_tail = NULL;
	qdf_nbuf_t nbuf;

	nbuf = nbuf_list;
	while (nbuf) {
		qdf_nbuf_t next = qdf_nbuf_next(nbuf);

		DP_RX_LIST_APPEND(deliver_list_head, deliver_list_tail, nbuf);

		/*
		 * reset the chfrag_start and chfrag_end bits in nbuf cb
		 * as this is a non-amsdu pkt and RAW mode simulation expects
		 * these bit s to be 0 for non-amsdu pkt.
		 */
		if (qdf_nbuf_is_chfrag_start(nbuf) &&
			 qdf_nbuf_is_chfrag_end(nbuf)) {
			qdf_nbuf_set_chfrag_start(nbuf, 0);
			qdf_nbuf_set_chfrag_end(nbuf, 0);
		}

		nbuf = next;
	}

	vdev->osif_rsim_rx_decap(vdev->osif_vdev, &deliver_list_head,
				 &deliver_list_tail);

	vdev->osif_rx(vdev->osif_vdev, deliver_list_head);
}


#ifdef DP_LFR
/*
 * In case of LFR, data of a new peer might be sent up
 * even before peer is added.
 */
static inline struct dp_vdev *
dp_get_vdev_from_peer(struct dp_soc *soc,
			uint16_t peer_id,
			struct dp_peer *peer,
			struct hal_rx_mpdu_desc_info mpdu_desc_info)
{
	struct dp_vdev *vdev;
	uint8_t vdev_id;

	if (unlikely(!peer)) {
		if (peer_id != HTT_INVALID_PEER) {
			vdev_id = DP_PEER_METADATA_ID_GET(
					mpdu_desc_info.peer_meta_data);
			QDF_TRACE(QDF_MODULE_ID_DP,
				QDF_TRACE_LEVEL_ERROR,
				FL("PeerID %d not found use vdevID %d"),
				peer_id, vdev_id);
			vdev = dp_get_vdev_from_soc_vdev_id_wifi3(soc,
							vdev_id);
		} else {
			QDF_TRACE(QDF_MODULE_ID_DP,
				QDF_TRACE_LEVEL_ERROR,
				FL("Invalid PeerID %d"),
				peer_id);
			return NULL;
		}
	} else {
		vdev = peer->vdev;
	}
	return vdev;
}
#else
static inline struct dp_vdev *
dp_get_vdev_from_peer(struct dp_soc *soc,
			uint16_t peer_id,
			struct dp_peer *peer,
			struct hal_rx_mpdu_desc_info mpdu_desc_info)
{
	if (unlikely(!peer)) {
		QDF_TRACE(QDF_MODULE_ID_DP,
			QDF_TRACE_LEVEL_ERROR,
			FL("Peer not found for peerID %d"),
			peer_id);
		return NULL;
	} else {
		return peer->vdev;
	}
}
#endif
/**
 * dp_rx_intrabss_fwd() - Implements the Intra-BSS forwarding logic
 *
 * @soc: core txrx main context
 * @sa_peer	: source peer entry
 * @rx_tlv_hdr	: start address of rx tlvs
 * @nbuf	: nbuf that has to be intrabss forwarded
 *
 * Return: bool: true if it is forwarded else false
 */
static bool
dp_rx_intrabss_fwd(struct dp_soc *soc,
			struct dp_peer *sa_peer,
			uint8_t *rx_tlv_hdr,
			qdf_nbuf_t nbuf)
{
	uint16_t da_idx;
	uint16_t len;
	struct dp_peer *da_peer;
	struct dp_ast_entry *ast_entry;
	qdf_nbuf_t nbuf_copy;

	/* check if the destination peer is available in peer table
	 * and also check if the source peer and destination peer
	 * belong to the same vap and destination peer is not bss peer.
	 */
	if ((hal_rx_msdu_end_da_is_valid_get(rx_tlv_hdr) &&
	   !hal_rx_msdu_end_da_is_mcbc_get(rx_tlv_hdr))) {
		da_idx = hal_rx_msdu_end_da_idx_get(rx_tlv_hdr);

		ast_entry = soc->ast_table[da_idx];
		if (!ast_entry)
			return false;

		da_peer = ast_entry->peer;

		if (!da_peer)
			return false;

		if (da_peer->vdev == sa_peer->vdev && !da_peer->bss_peer) {
			memset(nbuf->cb, 0x0, sizeof(nbuf->cb));
			len = qdf_nbuf_len(nbuf);
			if (!dp_tx_send(sa_peer->vdev, nbuf)) {
				DP_STATS_INC_PKT(sa_peer, rx.intra_bss, 1, len);
				return true;
			} else
				return false;
		}
	}
	/* if it is a broadcast pkt (eg: ARP) and it is not its own
	 * source, then clone the pkt and send the cloned pkt for
	 * intra BSS forwarding and original pkt up the network stack
	 * Note: how do we handle multicast pkts. do we forward
	 * all multicast pkts as is or let a higher layer module
	 * like igmpsnoop decide whether to forward or not with
	 * Mcast enhancement.
	 */
	else if (qdf_unlikely((hal_rx_msdu_end_da_is_mcbc_get(rx_tlv_hdr) &&
		!sa_peer->bss_peer))) {
		nbuf_copy = qdf_nbuf_copy(nbuf);
		if (!nbuf_copy)
			return false;
		memset(nbuf_copy->cb, 0x0, sizeof(nbuf_copy->cb));
		len = qdf_nbuf_len(nbuf_copy);
		if (dp_tx_send(sa_peer->vdev, nbuf_copy))
			qdf_nbuf_free(nbuf_copy);
		else
			DP_STATS_INC_PKT(sa_peer, rx.intra_bss, 1, len);
	}
	/* return false as we have to still send the original pkt
	 * up the stack
	 */
	return false;
}

#ifdef MESH_MODE_SUPPORT

/**
 * dp_rx_fill_mesh_stats() - Fills the mesh per packet receive stats
 *
 * @vdev: DP Virtual device handle
 * @nbuf: Buffer pointer
 *
 * This function allocated memory for mesh receive stats and fill the
 * required stats. Stores the memory address in skb cb.
 *
 * Return: void
 */
static
void dp_rx_fill_mesh_stats(struct dp_vdev *vdev, qdf_nbuf_t nbuf)
{
	struct mesh_recv_hdr_s *rx_info = NULL;
	uint32_t pkt_type;
	uint32_t nss;
	uint32_t rate_mcs;
	uint8_t *rx_tlv_hdr = qdf_nbuf_data(nbuf);

	/* fill recv mesh stats */
	rx_info = qdf_mem_malloc(sizeof(struct mesh_recv_hdr_s));

	/* upper layers are resposible to free this memory */

	if (rx_info == NULL) {
		QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
			"Memory allocation failed for mesh rx stats");
		return;
	}

	if (qdf_nbuf_is_chfrag_start(nbuf))
		rx_info->rs_flags |= MESH_RX_FIRST_MSDU;

	if (qdf_nbuf_is_chfrag_end(nbuf))
		rx_info->rs_flags |= MESH_RX_LAST_MSDU;

	if (hal_rx_attn_msdu_get_is_decrypted(rx_tlv_hdr)) {
		rx_info->rs_flags |= MESH_RX_DECRYPTED;
		rx_info->rs_keyix = hal_rx_msdu_get_keyid(rx_tlv_hdr);
		rx_info->rs_flags |= MESH_KEY_NOTFILLED;
	}

	rx_info->rs_rssi = hal_rx_msdu_start_get_rssi(rx_tlv_hdr);
	rx_info->rs_channel = hal_rx_msdu_start_get_freq(rx_tlv_hdr);
	pkt_type = hal_rx_msdu_start_get_pkt_type(rx_tlv_hdr);
	rate_mcs = hal_rx_msdu_start_rate_mcs_get(rx_tlv_hdr);
	nss = hal_rx_msdu_start_nss_get(rx_tlv_hdr);
	rx_info->rs_ratephy1 = rate_mcs | (nss << 0x4) | (pkt_type << 6);

	qdf_nbuf_set_fctx_type(nbuf, (void *)rx_info, CB_FTYPE_MESH_RX_INFO);

	QDF_TRACE(QDF_MODULE_ID_TXRX, QDF_TRACE_LEVEL_DEBUG,
		FL("Mesh rx stats: flags %x, rssi %x, chn %x, rate %x, kix %x"),
						rx_info->rs_flags,
						rx_info->rs_rssi,
						rx_info->rs_channel,
						rx_info->rs_ratephy1,
						rx_info->rs_keyix);

}

/**
 * dp_rx_fill_mesh_stats() - Filters mesh unwanted packets
 *
 * @vdev: DP Virtual device handle
 * @nbuf: Buffer pointer
 *
 * This checks if the received packet is matching any filter out
 * catogery and and drop the packet if it matches.
 *
 * Return: status(0 indicates drop, 1 indicate to no drop)
 */

static inline
QDF_STATUS dp_rx_filter_mesh_packets(struct dp_vdev *vdev, qdf_nbuf_t nbuf)
{
	uint8_t *rx_tlv_hdr = qdf_nbuf_data(nbuf);
	union dp_align_mac_addr mac_addr;

	if (qdf_unlikely(vdev->mesh_rx_filter)) {
		if (vdev->mesh_rx_filter & MESH_FILTER_OUT_FROMDS)
			if (hal_rx_mpdu_get_fr_ds(rx_tlv_hdr))
				return  QDF_STATUS_SUCCESS;

		if (vdev->mesh_rx_filter & MESH_FILTER_OUT_TODS)
			if (hal_rx_mpdu_get_to_ds(rx_tlv_hdr))
				return  QDF_STATUS_SUCCESS;

		if (vdev->mesh_rx_filter & MESH_FILTER_OUT_NODS)
			if (!hal_rx_mpdu_get_fr_ds(rx_tlv_hdr)
				&& !hal_rx_mpdu_get_to_ds(rx_tlv_hdr))
				return  QDF_STATUS_SUCCESS;

		if (vdev->mesh_rx_filter & MESH_FILTER_OUT_RA) {
			if (hal_rx_mpdu_get_addr1(rx_tlv_hdr,
					&mac_addr.raw[0]))
				return QDF_STATUS_E_FAILURE;

			if (!qdf_mem_cmp(&mac_addr.raw[0],
					&vdev->mac_addr.raw[0],
					DP_MAC_ADDR_LEN))
				return  QDF_STATUS_SUCCESS;
		}

		if (vdev->mesh_rx_filter & MESH_FILTER_OUT_TA) {
			if (hal_rx_mpdu_get_addr2(rx_tlv_hdr,
					&mac_addr.raw[0]))
				return QDF_STATUS_E_FAILURE;

			if (!qdf_mem_cmp(&mac_addr.raw[0],
					&vdev->mac_addr.raw[0],
					DP_MAC_ADDR_LEN))
				return  QDF_STATUS_SUCCESS;
		}
	}

	return QDF_STATUS_E_FAILURE;
}

#else
static
void dp_rx_fill_mesh_stats(struct dp_vdev *vdev, qdf_nbuf_t nbuf)
{
}

static inline
QDF_STATUS dp_rx_filter_mesh_packets(struct dp_vdev *vdev, qdf_nbuf_t nbuf)
{
	return QDF_STATUS_E_FAILURE;
}

#endif

#ifdef CONFIG_WIN
/**
 * dp_rx_process_invalid_peer(): Function to pass invalid peer list to umac
 * @soc: DP SOC handle
 * @nbuf: nbuf for which peer is invalid
 *
 * return: integer type
 */
uint8_t dp_rx_process_invalid_peer(struct dp_soc *soc, qdf_nbuf_t nbuf)
{
	struct dp_invalid_peer_msg msg;
	struct dp_vdev *vdev = NULL;
	struct dp_pdev *pdev = NULL;
	struct ieee80211_frame *wh;
	uint8_t i;
	uint8_t *rx_pkt_hdr;

	rx_pkt_hdr = qdf_nbuf_data(nbuf);
	wh = (struct ieee80211_frame *)rx_pkt_hdr;

	if (!DP_FRAME_IS_DATA(wh)) {
		QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_DEBUG,
				"NAWDS valid only for data frames");
		return 1;
	}

	if (qdf_nbuf_len(nbuf) < sizeof(struct ieee80211_frame)) {
		QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
				"Invalid nbuf length");
		return 1;
	}


	for (i = 0; i < MAX_PDEV_CNT; i++) {
		pdev = soc->pdev_list[i];
		if (!pdev) {
			QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
					"PDEV not found");
			continue;
		}

		TAILQ_FOREACH(vdev, &pdev->vdev_list, vdev_list_elem) {
			if (qdf_mem_cmp(wh->i_addr1, vdev->mac_addr.raw,
						DP_MAC_ADDR_LEN) == 0) {
				goto out;
			}
		}
	}

	if (!vdev) {
		QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
				"VDEV not found");
		return 1;
	}

out:
	msg.wh = wh;
	msg.nbuf = nbuf;
	msg.vdev_id = vdev->vdev_id;
	if (pdev->soc->cdp_soc.ol_ops->rx_invalid_peer)
		return pdev->soc->cdp_soc.ol_ops->rx_invalid_peer(
				pdev->osif_pdev, &msg);

	return 0;
}
#else
uint8_t dp_rx_process_invalid_peer(struct dp_soc *soc, qdf_nbuf_t nbuf)
{
	return 0;
}
#endif

/**
 * dp_rx_process() - Brain of the Rx processing functionality
 *		     Called from the bottom half (tasklet/NET_RX_SOFTIRQ)
 * @soc: core txrx main context
 * @hal_ring: opaque pointer to the HAL Rx Ring, which will be serviced
 * @quota: No. of units (packets) that can be serviced in one shot.
 *
 * This function implements the core of Rx functionality. This is
 * expected to handle only non-error frames.
 *
 * Return: uint32_t: No. of elements processed
 */
uint32_t
dp_rx_process(struct dp_soc *soc, void *hal_ring, uint32_t quota)
{
	void *hal_soc;
	void *ring_desc;
	struct dp_rx_desc *rx_desc;
	qdf_nbuf_t nbuf;
	union dp_rx_desc_list_elem_t *head[MAX_PDEV_CNT] = { NULL };
	union dp_rx_desc_list_elem_t *tail[MAX_PDEV_CNT] = { NULL };
	uint32_t rx_bufs_used = 0, rx_buf_cookie, l2_hdr_offset;
	uint16_t msdu_len;
	uint16_t peer_id;
	struct dp_peer *peer = NULL;
	struct dp_vdev *vdev = NULL;
	struct dp_vdev *vdev_list[WLAN_UMAC_PSOC_MAX_VDEVS] = { NULL };
	uint32_t pkt_len;
	struct hal_rx_mpdu_desc_info mpdu_desc_info;
	struct hal_rx_msdu_desc_info msdu_desc_info;
	enum hal_reo_error_status error;
	static uint32_t peer_mdata;
	uint8_t *rx_tlv_hdr;
	uint32_t rx_bufs_reaped[MAX_PDEV_CNT] = { 0 };
	uint32_t sgi, rate_mcs, tid, nss, bw, reception_type;
	uint64_t vdev_map = 0;
	uint8_t mac_id;
	uint16_t i, vdev_cnt = 0;
	uint32_t ampdu_flag, amsdu_flag;
	struct ether_header *eh;
	struct dp_pdev *pdev;
	struct dp_srng *dp_rxdma_srng;
	struct rx_desc_pool *rx_desc_pool;

	DP_HIST_INIT();
	/* Debug -- Remove later */
	qdf_assert(soc && hal_ring);

	hal_soc = soc->hal_soc;

	/* Debug -- Remove later */
	qdf_assert(hal_soc);

	if (qdf_unlikely(hal_srng_access_start(hal_soc, hal_ring))) {

		/*
		 * Need API to convert from hal_ring pointer to
		 * Ring Type / Ring Id combo
		 */
		DP_STATS_INC(soc, rx.err.hal_ring_access_fail, 1);
		QDF_TRACE(QDF_MODULE_ID_TXRX, QDF_TRACE_LEVEL_ERROR,
			FL("HAL RING Access Failed -- %p"), hal_ring);
		hal_srng_access_end(hal_soc, hal_ring);
		goto done;
	}

	/*
	 * start reaping the buffers from reo ring and queue
	 * them in per vdev queue.
	 * Process the received pkts in a different per vdev loop.
	 */
	while (qdf_likely((ring_desc =
				hal_srng_dst_get_next(hal_soc, hal_ring))
				&& quota--)) {

		error = HAL_RX_ERROR_STATUS_GET(ring_desc);

		if (qdf_unlikely(error == HAL_REO_ERROR_DETECTED)) {
			QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
			FL("HAL RING 0x%p:error %d"), hal_ring, error);
			/* Don't know how to deal with this -- assert */
			qdf_assert(0);
		}

		rx_buf_cookie = HAL_RX_REO_BUF_COOKIE_GET(ring_desc);

		rx_desc = dp_rx_cookie_2_va_rxdma_buf(soc, rx_buf_cookie);

		qdf_assert(rx_desc);
		rx_bufs_reaped[rx_desc->pool_id]++;

		/* TODO */
		/*
		 * Need a separate API for unmapping based on
		 * phyiscal address
		 */
		qdf_nbuf_unmap_single(soc->osdev, rx_desc->nbuf,
					QDF_DMA_BIDIRECTIONAL);

		/* Get MPDU DESC info */
		hal_rx_mpdu_desc_info_get(ring_desc, &mpdu_desc_info);
		peer_id = DP_PEER_METADATA_PEER_ID_GET(
				mpdu_desc_info.peer_meta_data);

		peer = dp_peer_find_by_id(soc, peer_id);

		vdev = dp_get_vdev_from_peer(soc, peer_id, peer,
						mpdu_desc_info);

		if (!vdev) {
			QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
				FL("vdev is NULL"));
			qdf_nbuf_free(rx_desc->nbuf);
			goto fail;

		}

		if (!((vdev_map >> vdev->vdev_id) & 1)) {
			vdev_map |= 1 << vdev->vdev_id;
			vdev_list[vdev_cnt] = vdev;
			vdev_cnt++;
		}

		/* Get MSDU DESC info */
		hal_rx_msdu_desc_info_get(ring_desc, &msdu_desc_info);

		/*
		 * save msdu flags first, last and continuation msdu in
		 * nbuf->cb
		 */
		if (msdu_desc_info.msdu_flags & HAL_MSDU_F_FIRST_MSDU_IN_MPDU)
			qdf_nbuf_set_chfrag_start(rx_desc->nbuf, 1);

		if (msdu_desc_info.msdu_flags & HAL_MSDU_F_MSDU_CONTINUATION)
			qdf_nbuf_set_chfrag_cont(rx_desc->nbuf, 1);

		if (msdu_desc_info.msdu_flags & HAL_MSDU_F_LAST_MSDU_IN_MPDU)
			qdf_nbuf_set_chfrag_end(rx_desc->nbuf, 1);

		DP_STATS_INC_PKT(peer, rx.rcvd_reo, 1,
				qdf_nbuf_len(rx_desc->nbuf));

		ampdu_flag = (mpdu_desc_info.mpdu_flags &
				HAL_MPDU_F_AMPDU_FLAG);

		DP_STATS_INCC(peer, rx.ampdu_cnt, 1, ampdu_flag);
		DP_STATS_INCC(peer, rx.non_ampdu_cnt, 1, !(ampdu_flag));

		hal_rx_msdu_desc_info_get(ring_desc, &msdu_desc_info);
		amsdu_flag = ((msdu_desc_info.msdu_flags &
				HAL_MSDU_F_FIRST_MSDU_IN_MPDU) &&
				(msdu_desc_info.msdu_flags &
					HAL_MSDU_F_LAST_MSDU_IN_MPDU));

		DP_STATS_INCC(peer, rx.non_amsdu_cnt, 1,
				amsdu_flag);
		DP_STATS_INCC(peer, rx.amsdu_cnt, 1,
				!(amsdu_flag));

		DP_HIST_PACKET_COUNT_INC(vdev->pdev->pdev_id);
		qdf_nbuf_queue_add(&vdev->rxq, rx_desc->nbuf);
fail:
		dp_rx_add_to_free_desc_list(&head[rx_desc->pool_id],
						&tail[rx_desc->pool_id],
						rx_desc);
	}
done:
	hal_srng_access_end(hal_soc, hal_ring);

	/* Update histogram statistics by looping through pdev's */
	DP_RX_HIST_STATS_PER_PDEV();

	for (mac_id = 0; mac_id < MAX_PDEV_CNT; mac_id++) {
		/*
		 * continue with next mac_id if no pkts were reaped
		 * from that pool
		 */
		if (!rx_bufs_reaped[mac_id])
			continue;

		pdev = soc->pdev_list[mac_id];
		dp_rxdma_srng = &pdev->rx_refill_buf_ring;
		rx_desc_pool = &soc->rx_desc_buf[mac_id];

		dp_rx_buffers_replenish(soc, mac_id, dp_rxdma_srng,
					rx_desc_pool, rx_bufs_reaped[mac_id],
					&head[mac_id], &tail[mac_id],
					HAL_RX_BUF_RBM_SW3_BM);
	}

	for (i = 0; i < vdev_cnt; i++) {
		qdf_nbuf_t deliver_list_head = NULL;
		qdf_nbuf_t deliver_list_tail = NULL;

		vdev = vdev_list[i];
		while ((nbuf = qdf_nbuf_queue_remove(&vdev->rxq))) {
			rx_tlv_hdr = qdf_nbuf_data(nbuf);
			eh = (struct ether_header *)qdf_nbuf_data(nbuf);

			/*
			 * Check if DMA completed -- msdu_done is the last bit
			 * to be written
			 */
			if (!hal_rx_attn_msdu_done_get(rx_tlv_hdr)) {

				QDF_TRACE(QDF_MODULE_ID_DP,
						QDF_TRACE_LEVEL_ERROR,
						FL("MSDU DONE failure"));

				hal_rx_dump_pkt_tlvs(rx_tlv_hdr,
							QDF_TRACE_LEVEL_INFO);
				qdf_assert(0);
			}

			if (qdf_nbuf_is_chfrag_start(nbuf))
				peer_mdata = hal_rx_mpdu_peer_meta_data_get(rx_tlv_hdr);

			peer_id = DP_PEER_METADATA_PEER_ID_GET(peer_mdata);
			peer = dp_peer_find_by_id(soc, peer_id);

			/* TODO */
			/*
			 * In case of roaming peer object may not be
			 * immediately available -- need to handle this
			 * Cannot drop these packets right away.
			 */
			/* Peer lookup failed */
			if (!peer && !vdev) {

				dp_rx_process_invalid_peer(soc, nbuf);
				DP_STATS_INC_PKT(soc, rx.err.rx_invalid_peer, 1,
						qdf_nbuf_len(nbuf));
				/* Drop & free packet */
				qdf_nbuf_free(nbuf);

				/* Statistics */
				continue;
			}

			if (peer && qdf_unlikely(peer->bss_peer)) {
				QDF_TRACE(QDF_MODULE_ID_DP,
					QDF_TRACE_LEVEL_INFO,
					FL("received pkt with same src MAC"));

				/* Drop & free packet */
				qdf_nbuf_free(nbuf);
				/* Statistics */
				continue;
			}

			sgi = hal_rx_msdu_start_sgi_get(rx_tlv_hdr);
			rate_mcs = hal_rx_msdu_start_rate_mcs_get(rx_tlv_hdr);
			tid = hal_rx_mpdu_start_tid_get(rx_tlv_hdr);

			QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_INFO,
				"%s: %d, SGI: %d, rate_mcs: %d, tid: %d",
				__func__, __LINE__, sgi, rate_mcs, tid);

			bw = hal_rx_msdu_start_bw_get(rx_tlv_hdr);
			reception_type = hal_rx_msdu_start_reception_type_get(
					rx_tlv_hdr);
			nss = hal_rx_msdu_start_nss_get(rx_tlv_hdr);

			DP_STATS_INC(vdev->pdev, rx.bw[bw], 1);
			DP_STATS_INC(vdev->pdev,
					rx.reception_type[reception_type], 1);
			DP_STATS_INCC(vdev->pdev, rx.nss[nss], 1,
					((reception_type == REPT_MU_MIMO) ||
					 (reception_type == REPT_MU_OFDMA_MIMO))
					);
			DP_STATS_INC(peer, rx.sgi_count[sgi], 1);
			DP_STATS_INC(peer, rx.mcs_count[rate_mcs], 1);
			DP_STATS_INCC(peer, rx.err.mic_err, 1,
					hal_rx_mpdu_end_mic_err_get(
						rx_tlv_hdr));
			DP_STATS_INCC(peer, rx.err.decrypt_err, 1,
					hal_rx_mpdu_end_decrypt_err_get(
						rx_tlv_hdr));

			DP_STATS_INC(peer, rx.wme_ac_type[TID_TO_WME_AC(tid)],
					1);
			DP_STATS_INC(peer, rx.bw[bw], 1);
			DP_STATS_INC(peer, rx.reception_type[reception_type],
					1);

			/*
			 * HW structures call this L3 header padding --
			 * even though this is actually the offset from
			 * the buffer beginning where the L2 header
			 * begins.
			 */
			QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_INFO,
				FL("rxhash: flow id toeplitz: 0x%x\n"),
				hal_rx_msdu_start_toeplitz_get(rx_tlv_hdr));

			l2_hdr_offset =
				hal_rx_msdu_end_l3_hdr_padding_get(rx_tlv_hdr);

			msdu_len = hal_rx_msdu_start_msdu_len_get(rx_tlv_hdr);
			pkt_len = msdu_len + l2_hdr_offset + RX_PKT_TLVS_LEN;

			/* Set length in nbuf */
			qdf_nbuf_set_pktlen(nbuf, pkt_len);

			if (qdf_unlikely(vdev->mesh_vdev)) {
				if (dp_rx_filter_mesh_packets(vdev, nbuf)
						== QDF_STATUS_SUCCESS) {
					QDF_TRACE(QDF_MODULE_ID_DP,
						QDF_TRACE_LEVEL_INFO_MED,
						FL("mesh pkt filtered"));

					qdf_nbuf_free(nbuf);
					continue;
				}
				dp_rx_fill_mesh_stats(vdev, nbuf);
			}

			/*
			 * Advance the packet start pointer by total size of
			 * pre-header TLV's
			 */
			qdf_nbuf_pull_head(nbuf,
					   RX_PKT_TLVS_LEN + l2_hdr_offset);

#ifdef QCA_WIFI_NAPIER_EMULATION /* Debug code, remove later */
			QDF_TRACE(QDF_MODULE_ID_DP, QDF_TRACE_LEVEL_ERROR,
				"p_id %d msdu_len %d hdr_off %d",
				peer_id, msdu_len, l2_hdr_offset);

			print_hex_dump(KERN_ERR,
				       "\t Pkt Data:", DUMP_PREFIX_NONE, 32, 4,
					qdf_nbuf_data(nbuf), 128, false);
#endif /* NAPIER_EMULATION */

			/* WDS Source Port Learning */
			dp_rx_wds_srcport_learn(soc, rx_tlv_hdr, peer, nbuf);

			/* Intrabss-fwd */
			if (vdev->opmode != wlan_op_mode_sta)
				if (dp_rx_intrabss_fwd(soc, peer, rx_tlv_hdr,
									nbuf))
					continue; /* Get next descriptor */

			rx_bufs_used++;
			DP_RX_LIST_APPEND(deliver_list_head,
						deliver_list_tail,
						nbuf);

			DP_STATS_INCC_PKT(peer, rx.multicast, 1, pkt_len,
					DP_FRAME_IS_MULTICAST((eh)->ether_dhost
						));
			DP_STATS_INCC_PKT(peer, rx.unicast, 1, pkt_len,
					!(DP_FRAME_IS_MULTICAST(
							(eh)->ether_dhost)));
			DP_STATS_INC_PKT(peer, rx.to_stack, 1,
					pkt_len);

			if (hal_rx_attn_first_mpdu_get(rx_tlv_hdr)) {
				if (soc->cdp_soc.ol_ops->update_dp_stats)
					soc->cdp_soc.ol_ops->update_dp_stats(
							vdev->pdev->osif_pdev,
							&peer->stats,
							peer_id,
							UPDATE_PEER_STATS);

				dp_aggregate_vdev_stats(peer->vdev);

				if (soc->cdp_soc.ol_ops->update_dp_stats)
					soc->cdp_soc.ol_ops->update_dp_stats(
							vdev->pdev->osif_pdev,
							&peer->vdev->stats,
							peer->vdev->vdev_id,
							UPDATE_VDEV_STATS);
			}
		}

		if (qdf_unlikely(vdev->rx_decap_type == htt_cmn_pkt_type_raw))
			dp_rx_deliver_raw(vdev, deliver_list_head);
		else if (qdf_likely(vdev->osif_rx) && deliver_list_head)
			vdev->osif_rx(vdev->osif_vdev, deliver_list_head);

	}
	return rx_bufs_used; /* Assume no scale factor for now */
}

/**
 * dp_rx_detach() - detach dp rx
 * @pdev: core txrx pdev context
 *
 * This function will detach DP RX into main device context
 * will free DP Rx resources.
 *
 * Return: void
 */
void
dp_rx_pdev_detach(struct dp_pdev *pdev)
{
	uint8_t pdev_id = pdev->pdev_id;
	struct dp_soc *soc = pdev->soc;
	struct rx_desc_pool *rx_desc_pool;

	rx_desc_pool = &soc->rx_desc_buf[pdev_id];

	dp_rx_desc_pool_free(soc, pdev_id, rx_desc_pool);
	qdf_spinlock_destroy(&soc->rx_desc_mutex[pdev_id]);

	return;
}

/**
 * dp_rx_attach() - attach DP RX
 * @pdev: core txrx pdev context
 *
 * This function will attach a DP RX instance into the main
 * device (SOC) context. Will allocate dp rx resource and
 * initialize resources.
 *
 * Return: QDF_STATUS_SUCCESS: success
 *         QDF_STATUS_E_RESOURCES: Error return
 */
QDF_STATUS
dp_rx_pdev_attach(struct dp_pdev *pdev)
{
	uint8_t pdev_id = pdev->pdev_id;
	struct dp_soc *soc = pdev->soc;
	struct dp_srng rxdma_srng;
	uint32_t rxdma_entries;
	union dp_rx_desc_list_elem_t *desc_list = NULL;
	union dp_rx_desc_list_elem_t *tail = NULL;
	struct dp_srng *dp_rxdma_srng;
	struct rx_desc_pool *rx_desc_pool;

	qdf_spinlock_create(&soc->rx_desc_mutex[pdev_id]);
	pdev = soc->pdev_list[pdev_id];
	rxdma_srng = pdev->rx_refill_buf_ring;

	rxdma_entries = rxdma_srng.alloc_size/hal_srng_get_entrysize(
						     soc->hal_soc, RXDMA_BUF);

	rx_desc_pool = &soc->rx_desc_buf[pdev_id];

	dp_rx_desc_pool_alloc(soc, pdev_id, rxdma_entries*3, rx_desc_pool);
	/* For Rx buffers, WBM release ring is SW RING 3,for all pdev's */
	dp_rxdma_srng = &pdev->rx_refill_buf_ring;
	dp_rx_buffers_replenish(soc, pdev_id, dp_rxdma_srng, rx_desc_pool,
		rxdma_entries, &desc_list, &tail, HAL_RX_BUF_RBM_SW3_BM);

	return QDF_STATUS_SUCCESS;
}