Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next

Pull networking updates from David Miller:

 1) BPF debugger and asm tool by Daniel Borkmann.

 2) Speed up create/bind in AF_PACKET, also from Daniel Borkmann.

 3) Correct reciprocal_divide and update users, from Hannes Frederic
    Sowa and Daniel Borkmann.

 4) Currently we only have a "set" operation for the hw timestamp socket
    ioctl, add a "get" operation to match.  From Ben Hutchings.

 5) Add better trace events for debugging driver datapath problems, also
    from Ben Hutchings.

 6) Implement auto corking in TCP, from Eric Dumazet.  Basically, if we
    have a small send and a previous packet is already in the qdisc or
    device queue, defer until TX completion or we get more data.

 7) Allow userspace to manage ipv6 temporary addresses, from Jiri Pirko.

 8) Add a qdisc bypass option for AF_PACKET sockets, from Daniel
    Borkmann.

 9) Share IP header compression code between Bluetooth and IEEE802154
    layers, from Jukka Rissanen.

10) Fix ipv6 router reachability probing, from Jiri Benc.

11) Allow packets to be captured on macvtap devices, from Vlad Yasevich.

12) Support tunneling in GRO layer, from Jerry Chu.

13) Allow bonding to be configured fully using netlink, from Scott
    Feldman.

14) Allow AF_PACKET users to obtain the VLAN TPID, just like they can
    already get the TCI.  From Atzm Watanabe.

15) New "Heavy Hitter" qdisc, from Terry Lam.

16) Significantly improve the IPSEC support in pktgen, from Fan Du.

17) Allow ipv4 tunnels to cache routes, just like sockets.  From Tom
    Herbert.

18) Add Proportional Integral Enhanced packet scheduler, from Vijay
    Subramanian.

19) Allow openvswitch to mmap'd netlink, from Thomas Graf.

20) Key TCP metrics blobs also by source address, not just destination
    address.  From Christoph Paasch.

21) Support 10G in generic phylib.  From Andy Fleming.

22) Try to short-circuit GRO flow compares using device provided RX
    hash, if provided.  From Tom Herbert.

The wireless and netfilter folks have been busy little bees too.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (2064 commits)
  net/cxgb4: Fix referencing freed adapter
  ipv6: reallocate addrconf router for ipv6 address when lo device up
  fib_frontend: fix possible NULL pointer dereference
  rtnetlink: remove IFLA_BOND_SLAVE definition
  rtnetlink: remove check for fill_slave_info in rtnl_have_link_slave_info
  qlcnic: update version to 5.3.55
  qlcnic: Enhance logic to calculate msix vectors.
  qlcnic: Refactor interrupt coalescing code for all adapters.
  qlcnic: Update poll controller code path
  qlcnic: Interrupt code cleanup
  qlcnic: Enhance Tx timeout debugging.
  qlcnic: Use bool for rx_mac_learn.
  bonding: fix u64 division
  rtnetlink: add missing IFLA_BOND_AD_INFO_UNSPEC
  sfc: Use the correct maximum TX DMA ring size for SFC9100
  Add Shradha Shah as the sfc driver maintainer.
  net/vxlan: Share RX skb de-marking and checksum checks with ovs
  tulip: cleanup by using ARRAY_SIZE()
  ip_tunnel: clear IPCB in ip_tunnel_xmit() in case dst_link_failure() is called
  net/cxgb4: Don't retrieve stats during recovery
  ...

1 files changed, 410 insertions, 149 deletions

diff --git a/drivers/net/ethernet/brocade/bna/bnad.c b/drivers/net/ethernet/brocade/bna/bnad.c
index 248bc37cb41b..cf64f3d0b60d 100644
--- a/drivers/net/ethernet/brocade/bna/bnad.c
+++ b/drivers/net/ethernet/brocade/bna/bnad.c
@@ -52,7 +52,7 @@ MODULE_PARM_DESC(bna_debugfs_enable, "Enables debugfs feature, default=1,"
 /*
  * Global variables
  */
-u32 bnad_rxqs_per_cq = 2;
+static u32 bnad_rxqs_per_cq = 2;
 static u32 bna_id;
 static struct mutex bnad_list_mutex;
 static LIST_HEAD(bnad_list);
@@ -142,7 +142,8 @@ bnad_tx_buff_unmap(struct bnad *bnad,
 
 		dma_unmap_page(&bnad->pcidev->dev,
 			dma_unmap_addr(&unmap->vectors[vector], dma_addr),
-			skb_shinfo(skb)->frags[nvecs].size, DMA_TO_DEVICE);
+			dma_unmap_len(&unmap->vectors[vector], dma_len),
+			DMA_TO_DEVICE);
 		dma_unmap_addr_set(&unmap->vectors[vector], dma_addr, 0);
 		nvecs--;
 	}
@@ -282,27 +283,32 @@ static int
 bnad_rxq_alloc_init(struct bnad *bnad, struct bna_rcb *rcb)
 {
 	struct bnad_rx_unmap_q *unmap_q = rcb->unmap_q;
-	int mtu, order;
+	int order;
 
 	bnad_rxq_alloc_uninit(bnad, rcb);
 
-	mtu = bna_enet_mtu_get(&bnad->bna.enet);
-	order = get_order(mtu);
+	order = get_order(rcb->rxq->buffer_size);
+
+	unmap_q->type = BNAD_RXBUF_PAGE;
 
 	if (bna_is_small_rxq(rcb->id)) {
 		unmap_q->alloc_order = 0;
 		unmap_q->map_size = rcb->rxq->buffer_size;
 	} else {
-		unmap_q->alloc_order = order;
-		unmap_q->map_size =
-			(rcb->rxq->buffer_size > 2048) ?
-			PAGE_SIZE << order : 2048;
+		if (rcb->rxq->multi_buffer) {
+			unmap_q->alloc_order = 0;
+			unmap_q->map_size = rcb->rxq->buffer_size;
+			unmap_q->type = BNAD_RXBUF_MULTI_BUFF;
+		} else {
+			unmap_q->alloc_order = order;
+			unmap_q->map_size =
+				(rcb->rxq->buffer_size > 2048) ?
+				PAGE_SIZE << order : 2048;
+		}
 	}
 
 	BUG_ON(((PAGE_SIZE << order) % unmap_q->map_size));
 
-	unmap_q->type = BNAD_RXBUF_PAGE;
-
 	return 0;
 }
 
@@ -345,10 +351,10 @@ bnad_rxq_cleanup(struct bnad *bnad, struct bna_rcb *rcb)
 	for (i = 0; i < rcb->q_depth; i++) {
 		struct bnad_rx_unmap *unmap = &unmap_q->unmap[i];
 
-		if (BNAD_RXBUF_IS_PAGE(unmap_q->type))
-			bnad_rxq_cleanup_page(bnad, unmap);
-		else
+		if (BNAD_RXBUF_IS_SK_BUFF(unmap_q->type))
 			bnad_rxq_cleanup_skb(bnad, unmap);
+		else
+			bnad_rxq_cleanup_page(bnad, unmap);
 	}
 	bnad_rxq_alloc_uninit(bnad, rcb);
 }
@@ -480,10 +486,10 @@ bnad_rxq_post(struct bnad *bnad, struct bna_rcb *rcb)
 	if (!(to_alloc >> BNAD_RXQ_REFILL_THRESHOLD_SHIFT))
 		return;
 
-	if (BNAD_RXBUF_IS_PAGE(unmap_q->type))
-		bnad_rxq_refill_page(bnad, rcb, to_alloc);
-	else
+	if (BNAD_RXBUF_IS_SK_BUFF(unmap_q->type))
 		bnad_rxq_refill_skb(bnad, rcb, to_alloc);
+	else
+		bnad_rxq_refill_page(bnad, rcb, to_alloc);
 }
 
 #define flags_cksum_prot_mask (BNA_CQ_EF_IPV4 | BNA_CQ_EF_L3_CKSUM_OK | \
@@ -500,72 +506,114 @@ bnad_rxq_post(struct bnad *bnad, struct bna_rcb *rcb)
 #define flags_udp6 (BNA_CQ_EF_IPV6 | \
 				BNA_CQ_EF_UDP | BNA_CQ_EF_L4_CKSUM_OK)
 
-static inline struct sk_buff *
-bnad_cq_prepare_skb(struct bnad_rx_ctrl *rx_ctrl,
-		struct bnad_rx_unmap_q *unmap_q,
-		struct bnad_rx_unmap *unmap,
-		u32 length, u32 flags)
+static void
+bnad_cq_drop_packet(struct bnad *bnad, struct bna_rcb *rcb,
+		    u32 sop_ci, u32 nvecs)
 {
-	struct bnad *bnad = rx_ctrl->bnad;
-	struct sk_buff *skb;
+	struct bnad_rx_unmap_q *unmap_q;
+	struct bnad_rx_unmap *unmap;
+	u32 ci, vec;
 
-	if (BNAD_RXBUF_IS_PAGE(unmap_q->type)) {
-		skb = napi_get_frags(&rx_ctrl->napi);
-		if (unlikely(!skb))
-			return NULL;
+	unmap_q = rcb->unmap_q;
+	for (vec = 0, ci = sop_ci; vec < nvecs; vec++) {
+		unmap = &unmap_q->unmap[ci];
+		BNA_QE_INDX_INC(ci, rcb->q_depth);
+
+		if (BNAD_RXBUF_IS_SK_BUFF(unmap_q->type))
+			bnad_rxq_cleanup_skb(bnad, unmap);
+		else
+			bnad_rxq_cleanup_page(bnad, unmap);
+	}
+}
+
+static void
+bnad_cq_setup_skb_frags(struct bna_rcb *rcb, struct sk_buff *skb,
+			u32 sop_ci, u32 nvecs, u32 last_fraglen)
+{
+	struct bnad *bnad;
+	u32 ci, vec, len, totlen = 0;
+	struct bnad_rx_unmap_q *unmap_q;
+	struct bnad_rx_unmap *unmap;
+
+	unmap_q = rcb->unmap_q;
+	bnad = rcb->bnad;
+
+	/* prefetch header */
+	prefetch(page_address(unmap_q->unmap[sop_ci].page) +
+			unmap_q->unmap[sop_ci].page_offset);
+
+	for (vec = 1, ci = sop_ci; vec <= nvecs; vec++) {
+		unmap = &unmap_q->unmap[ci];
+		BNA_QE_INDX_INC(ci, rcb->q_depth);
 
 		dma_unmap_page(&bnad->pcidev->dev,
 				dma_unmap_addr(&unmap->vector, dma_addr),
 				unmap->vector.len, DMA_FROM_DEVICE);
+
+		len = (vec == nvecs) ?
+			last_fraglen : unmap->vector.len;
+		totlen += len;
+
 		skb_fill_page_desc(skb, skb_shinfo(skb)->nr_frags,
-				unmap->page, unmap->page_offset, length);
-		skb->len += length;
-		skb->data_len += length;
-		skb->truesize += length;
+				unmap->page, unmap->page_offset, len);
 
 		unmap->page = NULL;
 		unmap->vector.len = 0;
-
-		return skb;
 	}
 
-	skb = unmap->skb;
-	BUG_ON(!skb);
+	skb->len += totlen;
+	skb->data_len += totlen;
+	skb->truesize += totlen;
+}
+
+static inline void
+bnad_cq_setup_skb(struct bnad *bnad, struct sk_buff *skb,
+		  struct bnad_rx_unmap *unmap, u32 len)
+{
+	prefetch(skb->data);
 
 	dma_unmap_single(&bnad->pcidev->dev,
 			dma_unmap_addr(&unmap->vector, dma_addr),
 			unmap->vector.len, DMA_FROM_DEVICE);
 
-	skb_put(skb, length);
-
+	skb_put(skb, len);
 	skb->protocol = eth_type_trans(skb, bnad->netdev);
 
 	unmap->skb = NULL;
 	unmap->vector.len = 0;
-	return skb;
 }
 
 static u32
 bnad_cq_process(struct bnad *bnad, struct bna_ccb *ccb, int budget)
 {
-	struct bna_cq_entry *cq, *cmpl;
+	struct bna_cq_entry *cq, *cmpl, *next_cmpl;
 	struct bna_rcb *rcb = NULL;
 	struct bnad_rx_unmap_q *unmap_q;
-	struct bnad_rx_unmap *unmap;
-	struct sk_buff *skb;
+	struct bnad_rx_unmap *unmap = NULL;
+	struct sk_buff *skb = NULL;
 	struct bna_pkt_rate *pkt_rt = &ccb->pkt_rate;
 	struct bnad_rx_ctrl *rx_ctrl = ccb->ctrl;
-	u32 packets = 0, length = 0, flags, masked_flags;
+	u32 packets = 0, len = 0, totlen = 0;
+	u32 pi, vec, sop_ci = 0, nvecs = 0;
+	u32 flags, masked_flags;
 
 	prefetch(bnad->netdev);
 
 	cq = ccb->sw_q;
 	cmpl = &cq[ccb->producer_index];
 
-	while (cmpl->valid && (packets < budget)) {
-		packets++;
-		flags = ntohl(cmpl->flags);
-		length = ntohs(cmpl->length);
+	while (packets < budget) {
+		if (!cmpl->valid)
+			break;
+		/* The 'valid' field is set by the adapter, only after writing
+		 * the other fields of completion entry. Hence, do not load
+		 * other fields of completion entry *before* the 'valid' is
+		 * loaded. Adding the rmb() here prevents the compiler and/or
+		 * CPU from reordering the reads which would potentially result
+		 * in reading stale values in completion entry.
+		 */
+		rmb();
+
 		BNA_UPDATE_PKT_CNT(pkt_rt, ntohs(cmpl->length));
 
 		if (bna_is_small_rxq(cmpl->rxq_id))
@@ -574,25 +622,78 @@ bnad_cq_process(struct bnad *bnad, struct bna_ccb *ccb, int budget)
 			rcb = ccb->rcb[0];
 
 		unmap_q = rcb->unmap_q;
-		unmap = &unmap_q->unmap[rcb->consumer_index];
 
-		if (unlikely(flags & (BNA_CQ_EF_MAC_ERROR |
-					BNA_CQ_EF_FCS_ERROR |
-					BNA_CQ_EF_TOO_LONG))) {
-			if (BNAD_RXBUF_IS_PAGE(unmap_q->type))
-				bnad_rxq_cleanup_page(bnad, unmap);
-			else
-				bnad_rxq_cleanup_skb(bnad, unmap);
+		/* start of packet ci */
+		sop_ci = rcb->consumer_index;
 
+		if (BNAD_RXBUF_IS_SK_BUFF(unmap_q->type)) {
+			unmap = &unmap_q->unmap[sop_ci];
+			skb = unmap->skb;
+		} else {
+			skb = napi_get_frags(&rx_ctrl->napi);
+			if (unlikely(!skb))
+				break;
+		}
+		prefetch(skb);
+
+		flags = ntohl(cmpl->flags);
+		len = ntohs(cmpl->length);
+		totlen = len;
+		nvecs = 1;
+
+		/* Check all the completions for this frame.
+		 * busy-wait doesn't help much, break here.
+		 */
+		if (BNAD_RXBUF_IS_MULTI_BUFF(unmap_q->type) &&
+		    (flags & BNA_CQ_EF_EOP) == 0) {
+			pi = ccb->producer_index;
+			do {
+				BNA_QE_INDX_INC(pi, ccb->q_depth);
+				next_cmpl = &cq[pi];
+
+				if (!next_cmpl->valid)
+					break;
+				/* The 'valid' field is set by the adapter, only
+				 * after writing the other fields of completion
+				 * entry. Hence, do not load other fields of
+				 * completion entry *before* the 'valid' is
+				 * loaded. Adding the rmb() here prevents the
+				 * compiler and/or CPU from reordering the reads
+				 * which would potentially result in reading
+				 * stale values in completion entry.
+				 */
+				rmb();
+
+				len = ntohs(next_cmpl->length);
+				flags = ntohl(next_cmpl->flags);
+
+				nvecs++;
+				totlen += len;
+			} while ((flags & BNA_CQ_EF_EOP) == 0);
+
+			if (!next_cmpl->valid)
+				break;
+		}
+
+		/* TODO: BNA_CQ_EF_LOCAL ? */
+		if (unlikely(flags & (BNA_CQ_EF_MAC_ERROR |
+						BNA_CQ_EF_FCS_ERROR |
+						BNA_CQ_EF_TOO_LONG))) {
+			bnad_cq_drop_packet(bnad, rcb, sop_ci, nvecs);
 			rcb->rxq->rx_packets_with_error++;
+
 			goto next;
 		}
 
-		skb = bnad_cq_prepare_skb(ccb->ctrl, unmap_q, unmap,
-				length, flags);
+		if (BNAD_RXBUF_IS_SK_BUFF(unmap_q->type))
+			bnad_cq_setup_skb(bnad, skb, unmap, len);
+		else
+			bnad_cq_setup_skb_frags(rcb, skb, sop_ci, nvecs, len);
 
-		if (unlikely(!skb))
-			break;
+		packets++;
+		rcb->rxq->rx_packets++;
+		rcb->rxq->rx_bytes += totlen;
+		ccb->bytes_per_intr += totlen;
 
 		masked_flags = flags & flags_cksum_prot_mask;
 
@@ -606,21 +707,21 @@ bnad_cq_process(struct bnad *bnad, struct bna_ccb *ccb, int budget)
 		else
 			skb_checksum_none_assert(skb);
 
-		rcb->rxq->rx_packets++;
-		rcb->rxq->rx_bytes += length;
-
 		if (flags & BNA_CQ_EF_VLAN)
 			__vlan_hwaccel_put_tag(skb, htons(ETH_P_8021Q), ntohs(cmpl->vlan_tag));
 
-		if (BNAD_RXBUF_IS_PAGE(unmap_q->type))
-			napi_gro_frags(&rx_ctrl->napi);
-		else
+		if (BNAD_RXBUF_IS_SK_BUFF(unmap_q->type))
 			netif_receive_skb(skb);
+		else
+			napi_gro_frags(&rx_ctrl->napi);
 
 next:
-		cmpl->valid = 0;
-		BNA_QE_INDX_INC(rcb->consumer_index, rcb->q_depth);
-		BNA_QE_INDX_INC(ccb->producer_index, ccb->q_depth);
+		BNA_QE_INDX_ADD(rcb->consumer_index, nvecs, rcb->q_depth);
+		for (vec = 0; vec < nvecs; vec++) {
+			cmpl = &cq[ccb->producer_index];
+			cmpl->valid = 0;
+			BNA_QE_INDX_INC(ccb->producer_index, ccb->q_depth);
+		}
 		cmpl = &cq[ccb->producer_index];
 	}
 
@@ -1899,8 +2000,10 @@ bnad_setup_tx(struct bnad *bnad, u32 tx_id)
 	tx = bna_tx_create(&bnad->bna, bnad, tx_config, &tx_cbfn, res_info,
 			tx_info);
 	spin_unlock_irqrestore(&bnad->bna_lock, flags);
-	if (!tx)
+	if (!tx) {
+		err = -ENOMEM;
 		goto err_return;
+	}
 	tx_info->tx = tx;
 
 	INIT_DELAYED_WORK(&tx_info->tx_cleanup_work,
@@ -1911,7 +2014,7 @@ bnad_setup_tx(struct bnad *bnad, u32 tx_id)
 		err = bnad_tx_msix_register(bnad, tx_info,
 			tx_id, bnad->num_txq_per_tx);
 		if (err)
-			goto err_return;
+			goto cleanup_tx;
 	}
 
 	spin_lock_irqsave(&bnad->bna_lock, flags);
@@ -1920,6 +2023,12 @@ bnad_setup_tx(struct bnad *bnad, u32 tx_id)
 
 	return 0;
 
+cleanup_tx:
+	spin_lock_irqsave(&bnad->bna_lock, flags);
+	bna_tx_destroy(tx_info->tx);
+	spin_unlock_irqrestore(&bnad->bna_lock, flags);
+	tx_info->tx = NULL;
+	tx_info->tx_id = 0;
 err_return:
 	bnad_tx_res_free(bnad, res_info);
 	return err;
@@ -1930,6 +2039,7 @@ err_return:
 static void
 bnad_init_rx_config(struct bnad *bnad, struct bna_rx_config *rx_config)
 {
+	memset(rx_config, 0, sizeof(*rx_config));
 	rx_config->rx_type = BNA_RX_T_REGULAR;
 	rx_config->num_paths = bnad->num_rxp_per_rx;
 	rx_config->coalescing_timeo = bnad->rx_coalescing_timeo;
@@ -1950,10 +2060,39 @@ bnad_init_rx_config(struct bnad *bnad, struct bna_rx_config *rx_config)
 		memset(&rx_config->rss_config, 0,
 		       sizeof(rx_config->rss_config));
 	}
+
+	rx_config->frame_size = BNAD_FRAME_SIZE(bnad->netdev->mtu);
+	rx_config->q0_multi_buf = BNA_STATUS_T_DISABLED;
+
+	/* BNA_RXP_SINGLE - one data-buffer queue
+	 * BNA_RXP_SLR - one small-buffer and one large-buffer queues
+	 * BNA_RXP_HDS - one header-buffer and one data-buffer queues
+	 */
+	/* TODO: configurable param for queue type */
 	rx_config->rxp_type = BNA_RXP_SLR;
-	rx_config->q_depth = bnad->rxq_depth;
 
-	rx_config->small_buff_size = BFI_SMALL_RXBUF_SIZE;
+	if (BNAD_PCI_DEV_IS_CAT2(bnad) &&
+	    rx_config->frame_size > 4096) {
+		/* though size_routing_enable is set in SLR,
+		 * small packets may get routed to same rxq.
+		 * set buf_size to 2048 instead of PAGE_SIZE.
+		 */
+		rx_config->q0_buf_size = 2048;
+		/* this should be in multiples of 2 */
+		rx_config->q0_num_vecs = 4;
+		rx_config->q0_depth = bnad->rxq_depth * rx_config->q0_num_vecs;
+		rx_config->q0_multi_buf = BNA_STATUS_T_ENABLED;
+	} else {
+		rx_config->q0_buf_size = rx_config->frame_size;
+		rx_config->q0_num_vecs = 1;
+		rx_config->q0_depth = bnad->rxq_depth;
+	}
+
+	/* initialize for q1 for BNA_RXP_SLR/BNA_RXP_HDS */
+	if (rx_config->rxp_type == BNA_RXP_SLR) {
+		rx_config->q1_depth = bnad->rxq_depth;
+		rx_config->q1_buf_size = BFI_SMALL_RXBUF_SIZE;
+	}
 
 	rx_config->vlan_strip_status = BNA_STATUS_T_ENABLED;
 }
@@ -1969,6 +2108,49 @@ bnad_rx_ctrl_init(struct bnad *bnad, u32 rx_id)
 }
 
 /* Called with mutex_lock(&bnad->conf_mutex) held */
+static u32
+bnad_reinit_rx(struct bnad *bnad)
+{
+	struct net_device *netdev = bnad->netdev;
+	u32 err = 0, current_err = 0;
+	u32 rx_id = 0, count = 0;
+	unsigned long flags;
+
+	/* destroy and create new rx objects */
+	for (rx_id = 0; rx_id < bnad->num_rx; rx_id++) {
+		if (!bnad->rx_info[rx_id].rx)
+			continue;
+		bnad_destroy_rx(bnad, rx_id);
+	}
+
+	spin_lock_irqsave(&bnad->bna_lock, flags);
+	bna_enet_mtu_set(&bnad->bna.enet,
+			 BNAD_FRAME_SIZE(bnad->netdev->mtu), NULL);
+	spin_unlock_irqrestore(&bnad->bna_lock, flags);
+
+	for (rx_id = 0; rx_id < bnad->num_rx; rx_id++) {
+		count++;
+		current_err = bnad_setup_rx(bnad, rx_id);
+		if (current_err && !err) {
+			err = current_err;
+			pr_err("RXQ:%u setup failed\n", rx_id);
+		}
+	}
+
+	/* restore rx configuration */
+	if (bnad->rx_info[0].rx && !err) {
+		bnad_restore_vlans(bnad, 0);
+		bnad_enable_default_bcast(bnad);
+		spin_lock_irqsave(&bnad->bna_lock, flags);
+		bnad_mac_addr_set_locked(bnad, netdev->dev_addr);
+		spin_unlock_irqrestore(&bnad->bna_lock, flags);
+		bnad_set_rx_mode(netdev);
+	}
+
+	return count;
+}
+
+/* Called with bnad_conf_lock() held */
 void
 bnad_destroy_rx(struct bnad *bnad, u32 rx_id)
 {
@@ -2047,13 +2229,19 @@ bnad_setup_rx(struct bnad *bnad, u32 rx_id)
 	spin_unlock_irqrestore(&bnad->bna_lock, flags);
 
 	/* Fill Unmap Q memory requirements */
-	BNAD_FILL_UNMAPQ_MEM_REQ(&res_info[BNA_RX_RES_MEM_T_UNMAPQ],
-			rx_config->num_paths +
-			((rx_config->rxp_type == BNA_RXP_SINGLE) ?
-			 0 : rx_config->num_paths),
-			((bnad->rxq_depth * sizeof(struct bnad_rx_unmap)) +
-			 sizeof(struct bnad_rx_unmap_q)));
-
+	BNAD_FILL_UNMAPQ_MEM_REQ(&res_info[BNA_RX_RES_MEM_T_UNMAPDQ],
+				 rx_config->num_paths,
+			(rx_config->q0_depth *
+			 sizeof(struct bnad_rx_unmap)) +
+			 sizeof(struct bnad_rx_unmap_q));
+
+	if (rx_config->rxp_type != BNA_RXP_SINGLE) {
+		BNAD_FILL_UNMAPQ_MEM_REQ(&res_info[BNA_RX_RES_MEM_T_UNMAPHQ],
+					 rx_config->num_paths,
+				(rx_config->q1_depth *
+				 sizeof(struct bnad_rx_unmap) +
+				 sizeof(struct bnad_rx_unmap_q)));
+	}
 	/* Allocate resource */
 	err = bnad_rx_res_alloc(bnad, res_info, rx_id);
 	if (err)
@@ -2548,7 +2736,6 @@ bnad_open(struct net_device *netdev)
 	int err;
 	struct bnad *bnad = netdev_priv(netdev);
 	struct bna_pause_config pause_config;
-	int mtu;
 	unsigned long flags;
 
 	mutex_lock(&bnad->conf_mutex);
@@ -2567,10 +2754,9 @@ bnad_open(struct net_device *netdev)
 	pause_config.tx_pause = 0;
 	pause_config.rx_pause = 0;
 
-	mtu = ETH_HLEN + VLAN_HLEN + bnad->netdev->mtu + ETH_FCS_LEN;
-
 	spin_lock_irqsave(&bnad->bna_lock, flags);
-	bna_enet_mtu_set(&bnad->bna.enet, mtu, NULL);
+	bna_enet_mtu_set(&bnad->bna.enet,
+			 BNAD_FRAME_SIZE(bnad->netdev->mtu), NULL);
 	bna_enet_pause_config(&bnad->bna.enet, &pause_config, NULL);
 	bna_enet_enable(&bnad->bna.enet);
 	spin_unlock_irqrestore(&bnad->bna_lock, flags);
@@ -2624,9 +2810,6 @@ bnad_stop(struct net_device *netdev)
 	bnad_destroy_tx(bnad, 0);
 	bnad_destroy_rx(bnad, 0);
 
-	/* These config flags are cleared in the hardware */
-	bnad->cfg_flags &= ~(BNAD_CF_ALLMULTI | BNAD_CF_PROMISC);
-
 	/* Synchronize mailbox IRQ */
 	bnad_mbox_irq_sync(bnad);
 
@@ -2784,21 +2967,21 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)
 	}
 
 	tcb = bnad->tx_info[0].tcb[txq_id];
-	q_depth = tcb->q_depth;
-	prod = tcb->producer_index;
-
-	unmap_q = tcb->unmap_q;
 
 	/*
 	 * Takes care of the Tx that is scheduled between clearing the flag
 	 * and the netif_tx_stop_all_queues() call.
 	 */
-	if (unlikely(!test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags))) {
+	if (unlikely(!tcb || !test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags))) {
 		dev_kfree_skb(skb);
 		BNAD_UPDATE_CTR(bnad, tx_skb_stopping);
 		return NETDEV_TX_OK;
 	}
 
+	q_depth = tcb->q_depth;
+	prod = tcb->producer_index;
+	unmap_q = tcb->unmap_q;
+
 	vectors = 1 + skb_shinfo(skb)->nr_frags;
 	wis = BNA_TXQ_WI_NEEDED(vectors);	/* 4 vectors per work item */
 
@@ -2863,7 +3046,7 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)
 
 	for (i = 0, vect_id = 0; i < vectors - 1; i++) {
 		const struct skb_frag_struct *frag = &skb_shinfo(skb)->frags[i];
-		u16		size = skb_frag_size(frag);
+		u32		size = skb_frag_size(frag);
 
 		if (unlikely(size == 0)) {
 			/* Undo the changes starting at tcb->producer_index */
@@ -2888,10 +3071,11 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)
 
 		dma_addr = skb_frag_dma_map(&bnad->pcidev->dev, frag,
 					    0, size, DMA_TO_DEVICE);
+		dma_unmap_len_set(&unmap->vectors[vect_id], dma_len, size);
 		BNA_SET_DMA_ADDR(dma_addr, &txqent->vector[vect_id].host_addr);
 		txqent->vector[vect_id].length = htons(size);
 		dma_unmap_addr_set(&unmap->vectors[vect_id], dma_addr,
-						dma_addr);
+				   dma_addr);
 		head_unmap->nvecs++;
 	}
 
@@ -2911,6 +3095,8 @@ bnad_start_xmit(struct sk_buff *skb, struct net_device *netdev)
 	if (unlikely(!test_bit(BNAD_TXQ_TX_STARTED, &tcb->flags)))
 		return NETDEV_TX_OK;
 
+	skb_tx_timestamp(skb);
+
 	bna_txq_prod_indx_doorbell(tcb);
 	smp_mb();
 
@@ -2937,73 +3123,133 @@ bnad_get_stats64(struct net_device *netdev, struct rtnl_link_stats64 *stats)
 	return stats;
 }
 
+static void
+bnad_set_rx_ucast_fltr(struct bnad *bnad)
+{
+	struct net_device *netdev = bnad->netdev;
+	int uc_count = netdev_uc_count(netdev);
+	enum bna_cb_status ret;
+	u8 *mac_list;
+	struct netdev_hw_addr *ha;
+	int entry;
+
+	if (netdev_uc_empty(bnad->netdev)) {
+		bna_rx_ucast_listset(bnad->rx_info[0].rx, 0, NULL, NULL);
+		return;
+	}
+
+	if (uc_count > bna_attr(&bnad->bna)->num_ucmac)
+		goto mode_default;
+
+	mac_list = kzalloc(uc_count * ETH_ALEN, GFP_ATOMIC);
+	if (mac_list == NULL)
+		goto mode_default;
+
+	entry = 0;
+	netdev_for_each_uc_addr(ha, netdev) {
+		memcpy(&mac_list[entry * ETH_ALEN],
+		       &ha->addr[0], ETH_ALEN);
+		entry++;
+	}
+
+	ret = bna_rx_ucast_listset(bnad->rx_info[0].rx, entry,
+			mac_list, NULL);
+	kfree(mac_list);
+
+	if (ret != BNA_CB_SUCCESS)
+		goto mode_default;
+
+	return;
+
+	/* ucast packets not in UCAM are routed to default function */
+mode_default:
+	bnad->cfg_flags |= BNAD_CF_DEFAULT;
+	bna_rx_ucast_listset(bnad->rx_info[0].rx, 0, NULL, NULL);
+}
+
+static void
+bnad_set_rx_mcast_fltr(struct bnad *bnad)
+{
+	struct net_device *netdev = bnad->netdev;
+	int mc_count = netdev_mc_count(netdev);
+	enum bna_cb_status ret;
+	u8 *mac_list;
+
+	if (netdev->flags & IFF_ALLMULTI)
+		goto mode_allmulti;
+
+	if (netdev_mc_empty(netdev))
+		return;
+
+	if (mc_count > bna_attr(&bnad->bna)->num_mcmac)
+		goto mode_allmulti;
+
+	mac_list = kzalloc((mc_count + 1) * ETH_ALEN, GFP_ATOMIC);
+
+	if (mac_list == NULL)
+		goto mode_allmulti;
+
+	memcpy(&mac_list[0], &bnad_bcast_addr[0], ETH_ALEN);
+
+	/* copy rest of the MCAST addresses */
+	bnad_netdev_mc_list_get(netdev, mac_list);
+	ret = bna_rx_mcast_listset(bnad->rx_info[0].rx, mc_count + 1,
+			mac_list, NULL);
+	kfree(mac_list);
+
+	if (ret != BNA_CB_SUCCESS)
+		goto mode_allmulti;
+
+	return;
+
+mode_allmulti:
+	bnad->cfg_flags |= BNAD_CF_ALLMULTI;
+	bna_rx_mcast_delall(bnad->rx_info[0].rx, NULL);
+}
+
 void
 bnad_set_rx_mode(struct net_device *netdev)
 {
 	struct bnad *bnad = netdev_priv(netdev);
-	u32	new_mask, valid_mask;
+	enum bna_rxmode new_mode, mode_mask;
 	unsigned long flags;
 
 	spin_lock_irqsave(&bnad->bna_lock, flags);
 
-	new_mask = valid_mask = 0;
-
-	if (netdev->flags & IFF_PROMISC) {
-		if (!(bnad->cfg_flags & BNAD_CF_PROMISC)) {
-			new_mask = BNAD_RXMODE_PROMISC_DEFAULT;
-			valid_mask = BNAD_RXMODE_PROMISC_DEFAULT;
-			bnad->cfg_flags |= BNAD_CF_PROMISC;
-		}
-	} else {
-		if (bnad->cfg_flags & BNAD_CF_PROMISC) {
-			new_mask = ~BNAD_RXMODE_PROMISC_DEFAULT;
-			valid_mask = BNAD_RXMODE_PROMISC_DEFAULT;
-			bnad->cfg_flags &= ~BNAD_CF_PROMISC;
-		}
-	}
-
-	if (netdev->flags & IFF_ALLMULTI) {
-		if (!(bnad->cfg_flags & BNAD_CF_ALLMULTI)) {
-			new_mask |= BNA_RXMODE_ALLMULTI;
-			valid_mask |= BNA_RXMODE_ALLMULTI;
-			bnad->cfg_flags |= BNAD_CF_ALLMULTI;
-		}
-	} else {
-		if (bnad->cfg_flags & BNAD_CF_ALLMULTI) {
-			new_mask &= ~BNA_RXMODE_ALLMULTI;
-			valid_mask |= BNA_RXMODE_ALLMULTI;
-			bnad->cfg_flags &= ~BNAD_CF_ALLMULTI;
-		}
+	if (bnad->rx_info[0].rx == NULL) {
+		spin_unlock_irqrestore(&bnad->bna_lock, flags);
+		return;
 	}
 
-	if (bnad->rx_info[0].rx == NULL)
-		goto unlock;
+	/* clear bnad flags to update it with new settings */
+	bnad->cfg_flags &= ~(BNAD_CF_PROMISC | BNAD_CF_DEFAULT |
+			BNAD_CF_ALLMULTI);
 
-	bna_rx_mode_set(bnad->rx_info[0].rx, new_mask, valid_mask, NULL);
+	new_mode = 0;
+	if (netdev->flags & IFF_PROMISC) {
+		new_mode |= BNAD_RXMODE_PROMISC_DEFAULT;
+		bnad->cfg_flags |= BNAD_CF_PROMISC;
+	} else {
+		bnad_set_rx_mcast_fltr(bnad);
 
-	if (!netdev_mc_empty(netdev)) {
-		u8 *mcaddr_list;
-		int mc_count = netdev_mc_count(netdev);
+		if (bnad->cfg_flags & BNAD_CF_ALLMULTI)
+			new_mode |= BNA_RXMODE_ALLMULTI;
 
-		/* Index 0 holds the broadcast address */
-		mcaddr_list =
-			kzalloc((mc_count + 1) * ETH_ALEN,
-				GFP_ATOMIC);
-		if (!mcaddr_list)
-			goto unlock;
+		bnad_set_rx_ucast_fltr(bnad);
 
-		memcpy(&mcaddr_list[0], &bnad_bcast_addr[0], ETH_ALEN);
+		if (bnad->cfg_flags & BNAD_CF_DEFAULT)
+			new_mode |= BNA_RXMODE_DEFAULT;
+	}
 
-		/* Copy rest of the MC addresses */
-		bnad_netdev_mc_list_get(netdev, mcaddr_list);
+	mode_mask = BNA_RXMODE_PROMISC | BNA_RXMODE_DEFAULT |
+			BNA_RXMODE_ALLMULTI;
+	bna_rx_mode_set(bnad->rx_info[0].rx, new_mode, mode_mask, NULL);
 
-		bna_rx_mcast_listset(bnad->rx_info[0].rx, mc_count + 1,
-					mcaddr_list, NULL);
+	if (bnad->cfg_flags & BNAD_CF_PROMISC)
+		bna_rx_vlan_strip_disable(bnad->rx_info[0].rx);
+	else
+		bna_rx_vlan_strip_enable(bnad->rx_info[0].rx);
 
-		/* Should we enable BNAD_CF_ALLMULTI for err != 0 ? */
-		kfree(mcaddr_list);
-	}
-unlock:
 	spin_unlock_irqrestore(&bnad->bna_lock, flags);
 }
 
@@ -3033,14 +3279,14 @@ bnad_set_mac_address(struct net_device *netdev, void *mac_addr)
 }
 
 static int
-bnad_mtu_set(struct bnad *bnad, int mtu)
+bnad_mtu_set(struct bnad *bnad, int frame_size)
 {
 	unsigned long flags;
 
 	init_completion(&bnad->bnad_completions.mtu_comp);
 
 	spin_lock_irqsave(&bnad->bna_lock, flags);
-	bna_enet_mtu_set(&bnad->bna.enet, mtu, bnad_cb_enet_mtu_set);
+	bna_enet_mtu_set(&bnad->bna.enet, frame_size, bnad_cb_enet_mtu_set);
 	spin_unlock_irqrestore(&bnad->bna_lock, flags);
 
 	wait_for_completion(&bnad->bnad_completions.mtu_comp);
@@ -3051,18 +3297,34 @@ bnad_mtu_set(struct bnad *bnad, int mtu)
 static int
 bnad_change_mtu(struct net_device *netdev, int new_mtu)
 {
-	int err, mtu = netdev->mtu;
+	int err, mtu;
 	struct bnad *bnad = netdev_priv(netdev);
+	u32 rx_count = 0, frame, new_frame;
 
 	if (new_mtu + ETH_HLEN < ETH_ZLEN || new_mtu > BNAD_JUMBO_MTU)
 		return -EINVAL;
 
 	mutex_lock(&bnad->conf_mutex);
 
+	mtu = netdev->mtu;
 	netdev->mtu = new_mtu;
 
-	mtu = ETH_HLEN + VLAN_HLEN + new_mtu + ETH_FCS_LEN;
-	err = bnad_mtu_set(bnad, mtu);
+	frame = BNAD_FRAME_SIZE(mtu);
+	new_frame = BNAD_FRAME_SIZE(new_mtu);
+
+	/* check if multi-buffer needs to be enabled */
+	if (BNAD_PCI_DEV_IS_CAT2(bnad) &&
+	    netif_running(bnad->netdev)) {
+		/* only when transition is over 4K */
+		if ((frame <= 4096 && new_frame > 4096) ||
+		    (frame > 4096 && new_frame <= 4096))
+			rx_count = bnad_reinit_rx(bnad);
+	}
+
+	/* rx_count > 0 - new rx created
+	 *	- Linux set err = 0 and return
+	 */
+	err = bnad_mtu_set(bnad, new_frame);
 	if (err)
 		err = -EBUSY;
 
@@ -3262,7 +3524,6 @@ bnad_uninit(struct bnad *bnad)
 
 	if (bnad->bar0)
 		iounmap(bnad->bar0);
-	pci_set_drvdata(bnad->pcidev, NULL);
 }
 
 /*