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

Pull networking changes from David Miller:

1) Allow to dump, monitor, and change the bridge multicast database
   using netlink.  From Cong Wang.

2) RFC 5961 TCP blind data injection attack mitigation, from Eric
   Dumazet.

3) Networking user namespace support from Eric W. Biederman.

4) tuntap/virtio-net multiqueue support by Jason Wang.

5) Support for checksum offload of encapsulated packets (basically,
   tunneled traffic can still be checksummed by HW).  From Joseph
   Gasparakis.

6) Allow BPF filter access to VLAN tags, from Eric Dumazet and
   Daniel Borkmann.

7) Bridge port parameters over netlink and BPDU blocking support
   from Stephen Hemminger.

8) Improve data access patterns during inet socket demux by rearranging
   socket layout, from Eric Dumazet.

9) TIPC protocol updates and cleanups from Ying Xue, Paul Gortmaker, and
   Jon Maloy.

10) Update TCP socket hash sizing to be more in line with current day
    realities.  The existing heurstics were choosen a decade ago.
    From Eric Dumazet.

11) Fix races, queue bloat, and excessive wakeups in ATM and
    associated drivers, from Krzysztof Mazur and David Woodhouse.

12) Support DOVE (Distributed Overlay Virtual Ethernet) extensions
    in VXLAN driver, from David Stevens.

13) Add "oops_only" mode to netconsole, from Amerigo Wang.

14) Support set and query of VEB/VEPA bridge mode via PF_BRIDGE, also
    allow DCB netlink to work on namespaces other than the initial
    namespace.  From John Fastabend.

15) Support PTP in the Tigon3 driver, from Matt Carlson.

16) tun/vhost zero copy fixes and improvements, plus turn it on
    by default, from Michael S. Tsirkin.

17) Support per-association statistics in SCTP, from Michele
    Baldessari.

And many, many, driver updates, cleanups, and improvements.  Too
numerous to mention individually.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1722 commits)
  net/mlx4_en: Add support for destination MAC in steering rules
  net/mlx4_en: Use generic etherdevice.h functions.
  net: ethtool: Add destination MAC address to flow steering API
  bridge: add support of adding and deleting mdb entries
  bridge: notify mdb changes via netlink
  ndisc: Unexport ndisc_{build,send}_skb().
  uapi: add missing netconf.h to export list
  pkt_sched: avoid requeues if possible
  solos-pci: fix double-free of TX skb in DMA mode
  bnx2: Fix accidental reversions.
  bna: Driver Version Updated to 3.1.2.1
  bna: Firmware update
  bna: Add RX State
  bna: Rx Page Based Allocation
  bna: TX Intr Coalescing Fix
  bna: Tx and Rx Optimizations
  bna: Code Cleanup and Enhancements
  ath9k: check pdata variable before dereferencing it
  ath5k: RX timestamp is reported at end of frame
  ath9k_htc: RX timestamp is reported at end of frame
  ...

1 files changed, 307 insertions, 106 deletions

diff --git a/drivers/net/wireless/iwlwifi/pcie/rx.c b/drivers/net/wireless/iwlwifi/pcie/rx.c
index bb69f8f90b3b..dad4c4aad91f 100644
--- a/drivers/net/wireless/iwlwifi/pcie/rx.c
+++ b/drivers/net/wireless/iwlwifi/pcie/rx.c
@@ -76,7 +76,7 @@
  * + A list of pre-allocated SKBs is stored in iwl->rxq->rx_free.  When
  *   iwl->rxq->free_count drops to or below RX_LOW_WATERMARK, work is scheduled
  *   to replenish the iwl->rxq->rx_free.
- * + In iwl_rx_replenish (scheduled) if 'processed' != 'read' then the
+ * + In iwl_pcie_rx_replenish (scheduled) if 'processed' != 'read' then the
  *   iwl->rxq is replenished and the READ INDEX is updated (updating the
  *   'processed' and 'read' driver indexes as well)
  * + A received packet is processed and handed to the kernel network stack,
@@ -89,28 +89,28 @@
  *
  * Driver sequence:
  *
- * iwl_rx_queue_alloc()   Allocates rx_free
- * iwl_rx_replenish()     Replenishes rx_free list from rx_used, and calls
- *                            iwl_rx_queue_restock
- * iwl_rx_queue_restock() Moves available buffers from rx_free into Rx
+ * iwl_rxq_alloc()            Allocates rx_free
+ * iwl_pcie_rx_replenish()    Replenishes rx_free list from rx_used, and calls
+ *                            iwl_pcie_rxq_restock
+ * iwl_pcie_rxq_restock()     Moves available buffers from rx_free into Rx
  *                            queue, updates firmware pointers, and updates
  *                            the WRITE index.  If insufficient rx_free buffers
- *                            are available, schedules iwl_rx_replenish
+ *                            are available, schedules iwl_pcie_rx_replenish
  *
  * -- enable interrupts --
- * ISR - iwl_rx()         Detach iwl_rx_mem_buffers from pool up to the
+ * ISR - iwl_rx()             Detach iwl_rx_mem_buffers from pool up to the
  *                            READ INDEX, detaching the SKB from the pool.
  *                            Moves the packet buffer from queue to rx_used.
- *                            Calls iwl_rx_queue_restock to refill any empty
+ *                            Calls iwl_pcie_rxq_restock to refill any empty
  *                            slots.
  * ...
  *
  */
 
-/**
- * iwl_rx_queue_space - Return number of free slots available in queue.
+/*
+ * iwl_rxq_space - Return number of free slots available in queue.
  */
-static int iwl_rx_queue_space(const struct iwl_rx_queue *q)
+static int iwl_rxq_space(const struct iwl_rxq *q)
 {
 	int s = q->read - q->write;
 	if (s <= 0)
@@ -122,11 +122,28 @@ static int iwl_rx_queue_space(const struct iwl_rx_queue *q)
 	return s;
 }
 
-/**
- * iwl_rx_queue_update_write_ptr - Update the write pointer for the RX queue
+/*
+ * iwl_dma_addr2rbd_ptr - convert a DMA address to a uCode read buffer ptr
+ */
+static inline __le32 iwl_pcie_dma_addr2rbd_ptr(dma_addr_t dma_addr)
+{
+	return cpu_to_le32((u32)(dma_addr >> 8));
+}
+
+/*
+ * iwl_pcie_rx_stop - stops the Rx DMA
+ */
+int iwl_pcie_rx_stop(struct iwl_trans *trans)
+{
+	iwl_write_direct32(trans, FH_MEM_RCSR_CHNL0_CONFIG_REG, 0);
+	return iwl_poll_direct_bit(trans, FH_MEM_RSSR_RX_STATUS_REG,
+				   FH_RSSR_CHNL0_RX_STATUS_CHNL_IDLE, 1000);
+}
+
+/*
+ * iwl_pcie_rxq_inc_wr_ptr - Update the write pointer for the RX queue
  */
-void iwl_rx_queue_update_write_ptr(struct iwl_trans *trans,
-				   struct iwl_rx_queue *q)
+static void iwl_pcie_rxq_inc_wr_ptr(struct iwl_trans *trans, struct iwl_rxq *q)
 {
 	unsigned long flags;
 	u32 reg;
@@ -176,16 +193,8 @@ void iwl_rx_queue_update_write_ptr(struct iwl_trans *trans,
 	spin_unlock_irqrestore(&q->lock, flags);
 }
 
-/**
- * iwl_dma_addr2rbd_ptr - convert a DMA address to a uCode read buffer ptr
- */
-static inline __le32 iwl_dma_addr2rbd_ptr(dma_addr_t dma_addr)
-{
-	return cpu_to_le32((u32)(dma_addr >> 8));
-}
-
-/**
- * iwl_rx_queue_restock - refill RX queue from pre-allocated pool
+/*
+ * iwl_pcie_rxq_restock - refill RX queue from pre-allocated pool
  *
  * If there are slots in the RX queue that need to be restocked,
  * and we have free pre-allocated buffers, fill the ranks as much
@@ -195,11 +204,10 @@ static inline __le32 iwl_dma_addr2rbd_ptr(dma_addr_t dma_addr)
  * also updates the memory address in the firmware to reference the new
  * target buffer.
  */
-static void iwl_rx_queue_restock(struct iwl_trans *trans)
+static void iwl_pcie_rxq_restock(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
-	struct iwl_rx_queue *rxq = &trans_pcie->rxq;
-	struct list_head *element;
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
 	struct iwl_rx_mem_buffer *rxb;
 	unsigned long flags;
 
@@ -215,18 +223,18 @@ static void iwl_rx_queue_restock(struct iwl_trans *trans)
 		return;
 
 	spin_lock_irqsave(&rxq->lock, flags);
-	while ((iwl_rx_queue_space(rxq) > 0) && (rxq->free_count)) {
+	while ((iwl_rxq_space(rxq) > 0) && (rxq->free_count)) {
 		/* The overwritten rxb must be a used one */
 		rxb = rxq->queue[rxq->write];
 		BUG_ON(rxb && rxb->page);
 
 		/* Get next free Rx buffer, remove from free list */
-		element = rxq->rx_free.next;
-		rxb = list_entry(element, struct iwl_rx_mem_buffer, list);
-		list_del(element);
+		rxb = list_first_entry(&rxq->rx_free, struct iwl_rx_mem_buffer,
+				       list);
+		list_del(&rxb->list);
 
 		/* Point to Rx buffer via next RBD in circular buffer */
-		rxq->bd[rxq->write] = iwl_dma_addr2rbd_ptr(rxb->page_dma);
+		rxq->bd[rxq->write] = iwl_pcie_dma_addr2rbd_ptr(rxb->page_dma);
 		rxq->queue[rxq->write] = rxb;
 		rxq->write = (rxq->write + 1) & RX_QUEUE_MASK;
 		rxq->free_count--;
@@ -243,24 +251,23 @@ static void iwl_rx_queue_restock(struct iwl_trans *trans)
 		spin_lock_irqsave(&rxq->lock, flags);
 		rxq->need_update = 1;
 		spin_unlock_irqrestore(&rxq->lock, flags);
-		iwl_rx_queue_update_write_ptr(trans, rxq);
+		iwl_pcie_rxq_inc_wr_ptr(trans, rxq);
 	}
 }
 
 /*
- * iwl_rx_allocate - allocate a page for each used RBD
+ * iwl_pcie_rxq_alloc_rbs - allocate a page for each used RBD
  *
  * A used RBD is an Rx buffer that has been given to the stack. To use it again
  * a page must be allocated and the RBD must point to the page. This function
  * doesn't change the HW pointer but handles the list of pages that is used by
- * iwl_rx_queue_restock. The latter function will update the HW to use the newly
+ * iwl_pcie_rxq_restock. The latter function will update the HW to use the newly
  * allocated buffers.
  */
-static void iwl_rx_allocate(struct iwl_trans *trans, gfp_t priority)
+static void iwl_pcie_rxq_alloc_rbs(struct iwl_trans *trans, gfp_t priority)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
-	struct iwl_rx_queue *rxq = &trans_pcie->rxq;
-	struct list_head *element;
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
 	struct iwl_rx_mem_buffer *rxb;
 	struct page *page;
 	unsigned long flags;
@@ -308,10 +315,9 @@ static void iwl_rx_allocate(struct iwl_trans *trans, gfp_t priority)
 			__free_pages(page, trans_pcie->rx_page_order);
 			return;
 		}
-		element = rxq->rx_used.next;
-		rxb = list_entry(element, struct iwl_rx_mem_buffer, list);
-		list_del(element);
-
+		rxb = list_first_entry(&rxq->rx_used, struct iwl_rx_mem_buffer,
+				       list);
+		list_del(&rxb->list);
 		spin_unlock_irqrestore(&rxq->lock, flags);
 
 		BUG_ON(rxb->page);
@@ -343,47 +349,227 @@ static void iwl_rx_allocate(struct iwl_trans *trans, gfp_t priority)
 	}
 }
 
+static void iwl_pcie_rxq_free_rbs(struct iwl_trans *trans)
+{
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
+	int i;
+
+	/* Fill the rx_used queue with _all_ of the Rx buffers */
+	for (i = 0; i < RX_FREE_BUFFERS + RX_QUEUE_SIZE; i++) {
+		/* In the reset function, these buffers may have been allocated
+		 * to an SKB, so we need to unmap and free potential storage */
+		if (rxq->pool[i].page != NULL) {
+			dma_unmap_page(trans->dev, rxq->pool[i].page_dma,
+				       PAGE_SIZE << trans_pcie->rx_page_order,
+				       DMA_FROM_DEVICE);
+			__free_pages(rxq->pool[i].page,
+				     trans_pcie->rx_page_order);
+			rxq->pool[i].page = NULL;
+		}
+		list_add_tail(&rxq->pool[i].list, &rxq->rx_used);
+	}
+}
+
 /*
- * iwl_rx_replenish - Move all used buffers from rx_used to rx_free
+ * iwl_pcie_rx_replenish - Move all used buffers from rx_used to rx_free
  *
  * When moving to rx_free an page is allocated for the slot.
  *
- * Also restock the Rx queue via iwl_rx_queue_restock.
+ * Also restock the Rx queue via iwl_pcie_rxq_restock.
  * This is called as a scheduled work item (except for during initialization)
  */
-void iwl_rx_replenish(struct iwl_trans *trans)
+static void iwl_pcie_rx_replenish(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 	unsigned long flags;
 
-	iwl_rx_allocate(trans, GFP_KERNEL);
+	iwl_pcie_rxq_alloc_rbs(trans, GFP_KERNEL);
 
 	spin_lock_irqsave(&trans_pcie->irq_lock, flags);
-	iwl_rx_queue_restock(trans);
+	iwl_pcie_rxq_restock(trans);
 	spin_unlock_irqrestore(&trans_pcie->irq_lock, flags);
 }
 
-static void iwl_rx_replenish_now(struct iwl_trans *trans)
+static void iwl_pcie_rx_replenish_now(struct iwl_trans *trans)
 {
-	iwl_rx_allocate(trans, GFP_ATOMIC);
+	iwl_pcie_rxq_alloc_rbs(trans, GFP_ATOMIC);
 
-	iwl_rx_queue_restock(trans);
+	iwl_pcie_rxq_restock(trans);
 }
 
-void iwl_bg_rx_replenish(struct work_struct *data)
+static void iwl_pcie_rx_replenish_work(struct work_struct *data)
 {
 	struct iwl_trans_pcie *trans_pcie =
 	    container_of(data, struct iwl_trans_pcie, rx_replenish);
 
-	iwl_rx_replenish(trans_pcie->trans);
+	iwl_pcie_rx_replenish(trans_pcie->trans);
+}
+
+static int iwl_pcie_rx_alloc(struct iwl_trans *trans)
+{
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
+	struct device *dev = trans->dev;
+
+	memset(&trans_pcie->rxq, 0, sizeof(trans_pcie->rxq));
+
+	spin_lock_init(&rxq->lock);
+
+	if (WARN_ON(rxq->bd || rxq->rb_stts))
+		return -EINVAL;
+
+	/* Allocate the circular buffer of Read Buffer Descriptors (RBDs) */
+	rxq->bd = dma_zalloc_coherent(dev, sizeof(__le32) * RX_QUEUE_SIZE,
+				      &rxq->bd_dma, GFP_KERNEL);
+	if (!rxq->bd)
+		goto err_bd;
+
+	/*Allocate the driver's pointer to receive buffer status */
+	rxq->rb_stts = dma_zalloc_coherent(dev, sizeof(*rxq->rb_stts),
+					   &rxq->rb_stts_dma, GFP_KERNEL);
+	if (!rxq->rb_stts)
+		goto err_rb_stts;
+
+	return 0;
+
+err_rb_stts:
+	dma_free_coherent(dev, sizeof(__le32) * RX_QUEUE_SIZE,
+			  rxq->bd, rxq->bd_dma);
+	memset(&rxq->bd_dma, 0, sizeof(rxq->bd_dma));
+	rxq->bd = NULL;
+err_bd:
+	return -ENOMEM;
+}
+
+static void iwl_pcie_rx_hw_init(struct iwl_trans *trans, struct iwl_rxq *rxq)
+{
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
+	u32 rb_size;
+	const u32 rfdnlog = RX_QUEUE_SIZE_LOG; /* 256 RBDs */
+
+	if (trans_pcie->rx_buf_size_8k)
+		rb_size = FH_RCSR_RX_CONFIG_REG_VAL_RB_SIZE_8K;
+	else
+		rb_size = FH_RCSR_RX_CONFIG_REG_VAL_RB_SIZE_4K;
+
+	/* Stop Rx DMA */
+	iwl_write_direct32(trans, FH_MEM_RCSR_CHNL0_CONFIG_REG, 0);
+
+	/* Reset driver's Rx queue write index */
+	iwl_write_direct32(trans, FH_RSCSR_CHNL0_RBDCB_WPTR_REG, 0);
+
+	/* Tell device where to find RBD circular buffer in DRAM */
+	iwl_write_direct32(trans, FH_RSCSR_CHNL0_RBDCB_BASE_REG,
+			   (u32)(rxq->bd_dma >> 8));
+
+	/* Tell device where in DRAM to update its Rx status */
+	iwl_write_direct32(trans, FH_RSCSR_CHNL0_STTS_WPTR_REG,
+			   rxq->rb_stts_dma >> 4);
+
+	/* Enable Rx DMA
+	 * FH_RCSR_CHNL0_RX_IGNORE_RXF_EMPTY is set because of HW bug in
+	 *      the credit mechanism in 5000 HW RX FIFO
+	 * Direct rx interrupts to hosts
+	 * Rx buffer size 4 or 8k
+	 * RB timeout 0x10
+	 * 256 RBDs
+	 */
+	iwl_write_direct32(trans, FH_MEM_RCSR_CHNL0_CONFIG_REG,
+			   FH_RCSR_RX_CONFIG_CHNL_EN_ENABLE_VAL |
+			   FH_RCSR_CHNL0_RX_IGNORE_RXF_EMPTY |
+			   FH_RCSR_CHNL0_RX_CONFIG_IRQ_DEST_INT_HOST_VAL |
+			   rb_size|
+			   (RX_RB_TIMEOUT << FH_RCSR_RX_CONFIG_REG_IRQ_RBTH_POS)|
+			   (rfdnlog << FH_RCSR_RX_CONFIG_RBDCB_SIZE_POS));
+
+	/* Set interrupt coalescing timer to default (2048 usecs) */
+	iwl_write8(trans, CSR_INT_COALESCING, IWL_HOST_INT_TIMEOUT_DEF);
+}
+
+int iwl_pcie_rx_init(struct iwl_trans *trans)
+{
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
+
+	int i, err;
+	unsigned long flags;
+
+	if (!rxq->bd) {
+		err = iwl_pcie_rx_alloc(trans);
+		if (err)
+			return err;
+	}
+
+	spin_lock_irqsave(&rxq->lock, flags);
+	INIT_LIST_HEAD(&rxq->rx_free);
+	INIT_LIST_HEAD(&rxq->rx_used);
+
+	INIT_WORK(&trans_pcie->rx_replenish,
+		  iwl_pcie_rx_replenish_work);
+
+	iwl_pcie_rxq_free_rbs(trans);
+
+	for (i = 0; i < RX_QUEUE_SIZE; i++)
+		rxq->queue[i] = NULL;
+
+	/* Set us so that we have processed and used all buffers, but have
+	 * not restocked the Rx queue with fresh buffers */
+	rxq->read = rxq->write = 0;
+	rxq->write_actual = 0;
+	rxq->free_count = 0;
+	spin_unlock_irqrestore(&rxq->lock, flags);
+
+	iwl_pcie_rx_replenish(trans);
+
+	iwl_pcie_rx_hw_init(trans, rxq);
+
+	spin_lock_irqsave(&trans_pcie->irq_lock, flags);
+	rxq->need_update = 1;
+	iwl_pcie_rxq_inc_wr_ptr(trans, rxq);
+	spin_unlock_irqrestore(&trans_pcie->irq_lock, flags);
+
+	return 0;
 }
 
-static void iwl_rx_handle_rxbuf(struct iwl_trans *trans,
+void iwl_pcie_rx_free(struct iwl_trans *trans)
+{
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
+	unsigned long flags;
+
+	/*if rxq->bd is NULL, it means that nothing has been allocated,
+	 * exit now */
+	if (!rxq->bd) {
+		IWL_DEBUG_INFO(trans, "Free NULL rx context\n");
+		return;
+	}
+
+	spin_lock_irqsave(&rxq->lock, flags);
+	iwl_pcie_rxq_free_rbs(trans);
+	spin_unlock_irqrestore(&rxq->lock, flags);
+
+	dma_free_coherent(trans->dev, sizeof(__le32) * RX_QUEUE_SIZE,
+			  rxq->bd, rxq->bd_dma);
+	memset(&rxq->bd_dma, 0, sizeof(rxq->bd_dma));
+	rxq->bd = NULL;
+
+	if (rxq->rb_stts)
+		dma_free_coherent(trans->dev,
+				  sizeof(struct iwl_rb_status),
+				  rxq->rb_stts, rxq->rb_stts_dma);
+	else
+		IWL_DEBUG_INFO(trans, "Free rxq->rb_stts which is NULL\n");
+	memset(&rxq->rb_stts_dma, 0, sizeof(rxq->rb_stts_dma));
+	rxq->rb_stts = NULL;
+}
+
+static void iwl_pcie_rx_handle_rb(struct iwl_trans *trans,
 				struct iwl_rx_mem_buffer *rxb)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
-	struct iwl_rx_queue *rxq = &trans_pcie->rxq;
-	struct iwl_tx_queue *txq = &trans_pcie->txq[trans_pcie->cmd_queue];
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
+	struct iwl_txq *txq = &trans_pcie->txq[trans_pcie->cmd_queue];
 	unsigned long flags;
 	bool page_stolen = false;
 	int max_len = PAGE_SIZE << trans_pcie->rx_page_order;
@@ -413,13 +599,13 @@ static void iwl_rx_handle_rxbuf(struct iwl_trans *trans,
 			break;
 
 		IWL_DEBUG_RX(trans, "cmd at offset %d: %s (0x%.2x)\n",
-			rxcb._offset,
-			trans_pcie_get_cmd_string(trans_pcie, pkt->hdr.cmd),
+			rxcb._offset, get_cmd_string(trans_pcie, pkt->hdr.cmd),
 			pkt->hdr.cmd);
 
 		len = le32_to_cpu(pkt->len_n_flags) & FH_RSCSR_FRAME_SIZE_MSK;
 		len += sizeof(u32); /* account for status word */
-		trace_iwlwifi_dev_rx(trans->dev, pkt, len);
+		trace_iwlwifi_dev_rx(trans->dev, trans, pkt, len);
+		trace_iwlwifi_dev_rx_data(trans->dev, trans, pkt, len);
 
 		/* Reclaim a command buffer only if this packet is a response
 		 *   to a (driver-originated) command.
@@ -445,7 +631,7 @@ static void iwl_rx_handle_rxbuf(struct iwl_trans *trans,
 		cmd_index = get_cmd_index(&txq->q, index);
 
 		if (reclaim) {
-			struct iwl_pcie_tx_queue_entry *ent;
+			struct iwl_pcie_txq_entry *ent;
 			ent = &txq->entries[cmd_index];
 			cmd = ent->copy_cmd;
 			WARN_ON_ONCE(!cmd && ent->meta.flags & CMD_WANT_HCMD);
@@ -459,6 +645,9 @@ static void iwl_rx_handle_rxbuf(struct iwl_trans *trans,
 			/* The original command isn't needed any more */
 			kfree(txq->entries[cmd_index].copy_cmd);
 			txq->entries[cmd_index].copy_cmd = NULL;
+			/* nor is the duplicated part of the command */
+			kfree(txq->entries[cmd_index].free_buf);
+			txq->entries[cmd_index].free_buf = NULL;
 		}
 
 		/*
@@ -472,7 +661,7 @@ static void iwl_rx_handle_rxbuf(struct iwl_trans *trans,
 			 * iwl_trans_send_cmd()
 			 * as we reclaim the driver command queue */
 			if (!rxcb._page_stolen)
-				iwl_tx_cmd_complete(trans, &rxcb, err);
+				iwl_pcie_hcmd_complete(trans, &rxcb, err);
 			else
 				IWL_WARN(trans, "Claim null rxb?\n");
 		}
@@ -514,17 +703,13 @@ static void iwl_rx_handle_rxbuf(struct iwl_trans *trans,
 	spin_unlock_irqrestore(&rxq->lock, flags);
 }
 
-/**
- * iwl_rx_handle - Main entry function for receiving responses from uCode
- *
- * Uses the priv->rx_handlers callback function array to invoke
- * the appropriate handlers, including command responses,
- * frame-received notifications, and other notifications.
+/*
+ * iwl_pcie_rx_handle - Main entry function for receiving responses from fw
  */
-static void iwl_rx_handle(struct iwl_trans *trans)
+static void iwl_pcie_rx_handle(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
-	struct iwl_rx_queue *rxq = &trans_pcie->rxq;
+	struct iwl_rxq *rxq = &trans_pcie->rxq;
 	u32 r, i;
 	u8 fill_rx = 0;
 	u32 count = 8;
@@ -532,7 +717,7 @@ static void iwl_rx_handle(struct iwl_trans *trans)
 
 	/* uCode's read index (stored in shared DRAM) indicates the last Rx
 	 * buffer that the driver may process (last buffer filled by ucode). */
-	r = le16_to_cpu(rxq->rb_stts->closed_rb_num) &  0x0FFF;
+	r = le16_to_cpu(ACCESS_ONCE(rxq->rb_stts->closed_rb_num)) & 0x0FFF;
 	i = rxq->read;
 
 	/* Rx interrupt, but nothing sent from uCode */
@@ -555,7 +740,7 @@ static void iwl_rx_handle(struct iwl_trans *trans)
 
 		IWL_DEBUG_RX(trans, "rxbuf: HW = %d, SW = %d (%p)\n",
 			     r, i, rxb);
-		iwl_rx_handle_rxbuf(trans, rxb);
+		iwl_pcie_rx_handle_rb(trans, rxb);
 
 		i = (i + 1) & RX_QUEUE_MASK;
 		/* If there are a lot of unused frames,
@@ -564,7 +749,7 @@ static void iwl_rx_handle(struct iwl_trans *trans)
 			count++;
 			if (count >= 8) {
 				rxq->read = i;
-				iwl_rx_replenish_now(trans);
+				iwl_pcie_rx_replenish_now(trans);
 				count = 0;
 			}
 		}
@@ -573,39 +758,41 @@ static void iwl_rx_handle(struct iwl_trans *trans)
 	/* Backtrack one entry */
 	rxq->read = i;
 	if (fill_rx)
-		iwl_rx_replenish_now(trans);
+		iwl_pcie_rx_replenish_now(trans);
 	else
-		iwl_rx_queue_restock(trans);
+		iwl_pcie_rxq_restock(trans);
 }
 
-/**
- * iwl_irq_handle_error - called for HW or SW error interrupt from card
+/*
+ * iwl_pcie_irq_handle_error - called for HW or SW error interrupt from card
  */
-static void iwl_irq_handle_error(struct iwl_trans *trans)
+static void iwl_pcie_irq_handle_error(struct iwl_trans *trans)
 {
+	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
+
 	/* W/A for WiFi/WiMAX coex and WiMAX own the RF */
 	if (trans->cfg->internal_wimax_coex &&
 	    (!(iwl_read_prph(trans, APMG_CLK_CTRL_REG) &
 			     APMS_CLK_VAL_MRB_FUNC_MODE) ||
 	     (iwl_read_prph(trans, APMG_PS_CTRL_REG) &
 			    APMG_PS_CTRL_VAL_RESET_REQ))) {
-		struct iwl_trans_pcie *trans_pcie =
-			IWL_TRANS_GET_PCIE_TRANS(trans);
-
 		clear_bit(STATUS_HCMD_ACTIVE, &trans_pcie->status);
 		iwl_op_mode_wimax_active(trans->op_mode);
-		wake_up(&trans->wait_command_queue);
+		wake_up(&trans_pcie->wait_command_queue);
 		return;
 	}
 
-	iwl_dump_csr(trans);
-	iwl_dump_fh(trans, NULL);
+	iwl_pcie_dump_csr(trans);
+	iwl_pcie_dump_fh(trans, NULL);
+
+	set_bit(STATUS_FW_ERROR, &trans_pcie->status);
+	clear_bit(STATUS_HCMD_ACTIVE, &trans_pcie->status);
+	wake_up(&trans_pcie->wait_command_queue);
 
 	iwl_op_mode_nic_error(trans->op_mode);
 }
 
-/* tasklet for iwlagn interrupt */
-void iwl_irq_tasklet(struct iwl_trans *trans)
+void iwl_pcie_tasklet(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 	struct isr_statistics *isr_stats = &trans_pcie->isr_stats;
@@ -657,7 +844,7 @@ void iwl_irq_tasklet(struct iwl_trans *trans)
 		iwl_disable_interrupts(trans);
 
 		isr_stats->hw++;
-		iwl_irq_handle_error(trans);
+		iwl_pcie_irq_handle_error(trans);
 
 		handled |= CSR_INT_BIT_HW_ERR;
 
@@ -694,6 +881,16 @@ void iwl_irq_tasklet(struct iwl_trans *trans)
 		isr_stats->rfkill++;
 
 		iwl_op_mode_hw_rf_kill(trans->op_mode, hw_rfkill);
+		if (hw_rfkill) {
+			set_bit(STATUS_RFKILL, &trans_pcie->status);
+			if (test_and_clear_bit(STATUS_HCMD_ACTIVE,
+					       &trans_pcie->status))
+				IWL_DEBUG_RF_KILL(trans,
+						  "Rfkill while SYNC HCMD in flight\n");
+			wake_up(&trans_pcie->wait_command_queue);
+		} else {
+			clear_bit(STATUS_RFKILL, &trans_pcie->status);
+		}
 
 		handled |= CSR_INT_BIT_RF_KILL;
 	}
@@ -710,17 +907,16 @@ void iwl_irq_tasklet(struct iwl_trans *trans)
 		IWL_ERR(trans, "Microcode SW error detected. "
 			" Restarting 0x%X.\n", inta);
 		isr_stats->sw++;
-		iwl_irq_handle_error(trans);
+		iwl_pcie_irq_handle_error(trans);
 		handled |= CSR_INT_BIT_SW_ERR;
 	}
 
 	/* uCode wakes up after power-down sleep */
 	if (inta & CSR_INT_BIT_WAKEUP) {
 		IWL_DEBUG_ISR(trans, "Wakeup interrupt\n");
-		iwl_rx_queue_update_write_ptr(trans, &trans_pcie->rxq);
+		iwl_pcie_rxq_inc_wr_ptr(trans, &trans_pcie->rxq);
 		for (i = 0; i < trans->cfg->base_params->num_of_queues; i++)
-			iwl_txq_update_write_ptr(trans,
-						 &trans_pcie->txq[i]);
+			iwl_pcie_txq_inc_wr_ptr(trans, &trans_pcie->txq[i]);
 
 		isr_stats->wakeup++;
 
@@ -758,7 +954,7 @@ void iwl_irq_tasklet(struct iwl_trans *trans)
 		iwl_write8(trans, CSR_INT_PERIODIC_REG,
 			    CSR_INT_PERIODIC_DIS);
 
-		iwl_rx_handle(trans);
+		iwl_pcie_rx_handle(trans);
 
 		/*
 		 * Enable periodic interrupt in 8 msec only if we received
@@ -816,7 +1012,7 @@ void iwl_irq_tasklet(struct iwl_trans *trans)
 #define ICT_COUNT	(ICT_SIZE / sizeof(u32))
 
 /* Free dram table */
-void iwl_free_isr_ict(struct iwl_trans *trans)
+void iwl_pcie_free_ict(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 
@@ -829,13 +1025,12 @@ void iwl_free_isr_ict(struct iwl_trans *trans)
 	}
 }
 
-
 /*
  * allocate dram shared table, it is an aligned memory
  * block of ICT_SIZE.
  * also reset all data related to ICT table interrupt.
  */
-int iwl_alloc_isr_ict(struct iwl_trans *trans)
+int iwl_pcie_alloc_ict(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 
@@ -848,7 +1043,7 @@ int iwl_alloc_isr_ict(struct iwl_trans *trans)
 
 	/* just an API sanity check ... it is guaranteed to be aligned */
 	if (WARN_ON(trans_pcie->ict_tbl_dma & (ICT_SIZE - 1))) {
-		iwl_free_isr_ict(trans);
+		iwl_pcie_free_ict(trans);
 		return -EINVAL;
 	}
 
@@ -869,7 +1064,7 @@ int iwl_alloc_isr_ict(struct iwl_trans *trans)
 /* Device is going up inform it about using ICT interrupt table,
  * also we need to tell the driver to start using ICT interrupt.
  */
-void iwl_reset_ict(struct iwl_trans *trans)
+void iwl_pcie_reset_ict(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 	u32 val;
@@ -899,7 +1094,7 @@ void iwl_reset_ict(struct iwl_trans *trans)
 }
 
 /* Device is going down disable ict interrupt usage */
-void iwl_disable_ict(struct iwl_trans *trans)
+void iwl_pcie_disable_ict(struct iwl_trans *trans)
 {
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
 	unsigned long flags;
@@ -910,7 +1105,7 @@ void iwl_disable_ict(struct iwl_trans *trans)
 }
 
 /* legacy (non-ICT) ISR. Assumes that trans_pcie->irq_lock is held */
-static irqreturn_t iwl_isr(int irq, void *data)
+static irqreturn_t iwl_pcie_isr(int irq, void *data)
 {
 	struct iwl_trans *trans = data;
 	struct iwl_trans_pcie *trans_pcie = IWL_TRANS_GET_PCIE_TRANS(trans);
@@ -927,12 +1122,20 @@ static irqreturn_t iwl_isr(int irq, void *data)
 	 *    back-to-back ISRs and sporadic interrupts from our NIC.
 	 * If we have something to service, the tasklet will re-enable ints.
 	 * If we *don't* have something, we'll re-enable before leaving here. */
-	inta_mask = iwl_read32(trans, CSR_INT_MASK);  /* just for debug */
+	inta_mask = iwl_read32(trans, CSR_INT_MASK);
 	iwl_write32(trans, CSR_INT_MASK, 0x00000000);
 
 	/* Discover which interrupts are active/pending */
 	inta = iwl_read32(trans, CSR_INT);
 
+	if (inta & (~inta_mask)) {
+		IWL_DEBUG_ISR(trans,
+			      "We got a masked interrupt (0x%08x)...Ack and ignore\n",
+			      inta & (~inta_mask));
+		iwl_write32(trans, CSR_INT, inta & (~inta_mask));
+		inta &= inta_mask;
+	}
+
 	/* Ignore interrupt if there's nothing in NIC to service.
 	 * This may be due to IRQ shared with another device,
 	 * or due to sporadic interrupts thrown from our NIC. */
@@ -957,7 +1160,7 @@ static irqreturn_t iwl_isr(int irq, void *data)
 #endif
 
 	trans_pcie->inta |= inta;
-	/* iwl_irq_tasklet() will service interrupts and re-enable them */
+	/* iwl_pcie_tasklet() will service interrupts and re-enable them */
 	if (likely(inta))
 		tasklet_schedule(&trans_pcie->irq_tasklet);
 	else if (test_bit(STATUS_INT_ENABLED, &trans_pcie->status) &&
@@ -982,7 +1185,7 @@ none:
  * the interrupt we need to service, driver will set the entries back to 0 and
  * set index.
  */
-irqreturn_t iwl_isr_ict(int irq, void *data)
+irqreturn_t iwl_pcie_isr_ict(int irq, void *data)
 {
 	struct iwl_trans *trans = data;
 	struct iwl_trans_pcie *trans_pcie;
@@ -1002,23 +1205,21 @@ irqreturn_t iwl_isr_ict(int irq, void *data)
 	 * use legacy interrupt.
 	 */
 	if (unlikely(!trans_pcie->use_ict)) {
-		irqreturn_t ret = iwl_isr(irq, data);
+		irqreturn_t ret = iwl_pcie_isr(irq, data);
 		spin_unlock_irqrestore(&trans_pcie->irq_lock, flags);
 		return ret;
 	}
 
 	trace_iwlwifi_dev_irq(trans->dev);
 
-
 	/* Disable (but don't clear!) interrupts here to avoid
 	 * back-to-back ISRs and sporadic interrupts from our NIC.
 	 * If we have something to service, the tasklet will re-enable ints.
 	 * If we *don't* have something, we'll re-enable before leaving here.
 	 */
-	inta_mask = iwl_read32(trans, CSR_INT_MASK);  /* just for debug */
+	inta_mask = iwl_read32(trans, CSR_INT_MASK);
 	iwl_write32(trans, CSR_INT_MASK, 0x00000000);
 
-
 	/* Ignore interrupt if there's nothing in NIC to service.
 	 * This may be due to IRQ shared with another device,
 	 * or due to sporadic interrupts thrown from our NIC. */
@@ -1067,7 +1268,7 @@ irqreturn_t iwl_isr_ict(int irq, void *data)
 	inta &= trans_pcie->inta_mask;
 	trans_pcie->inta |= inta;
 
-	/* iwl_irq_tasklet() will service interrupts and re-enable them */
+	/* iwl_pcie_tasklet() will service interrupts and re-enable them */
 	if (likely(inta))
 		tasklet_schedule(&trans_pcie->irq_tasklet);
 	else if (test_bit(STATUS_INT_ENABLED, &trans_pcie->status) &&