Driver for IBM System i/p VNIC protocol

This is a new device driver for a high performance SR-IOV assisted virtual
network for IBM System p and IBM System i systems.  The SR-IOV VF will be
attached to the VIOS partition and mapped to the Linux client via the
hypervisor's VNIC protocol that this driver implements.

This driver is able to perform basic tx and rx, new features
and improvements will be added as they are being developed and tested.

Signed-off-by: Thomas Falcon <tlfalcon@linux.vnet.ibm.com>
Signed-off-by: John Allen <jallen@linux.vnet.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

4 files changed, 4642 insertions, 0 deletions

diff --git a/drivers/net/ethernet/ibm/Kconfig b/drivers/net/ethernet/ibm/Kconfig
index 99c1cebd002d..37dceabf8861 100644
--- a/drivers/net/ethernet/ibm/Kconfig
+++ b/drivers/net/ethernet/ibm/Kconfig
@@ -37,4 +37,14 @@ config EHEA
 	  To compile the driver as a module, choose M here. The module
 	  will be called ehea.
 
+config IBMVNIC
+	tristate "IBM Virtual NIC support"
+	depends on PPC_PSERIES
+	---help---
+	  This driver supports Virtual NIC adapters on IBM i and IBM System p
+	  systems.
+
+	  To compile this driver as a module, choose M here. The module will
+	  be called ibmvnic.
+
 endif # NET_VENDOR_IBM
diff --git a/drivers/net/ethernet/ibm/Makefile b/drivers/net/ethernet/ibm/Makefile
index 2f04e71a5926..447865c8b632 100644
--- a/drivers/net/ethernet/ibm/Makefile
+++ b/drivers/net/ethernet/ibm/Makefile
@@ -3,5 +3,6 @@
 #
 
 obj-$(CONFIG_IBMVETH) += ibmveth.o
+obj-$(CONFIG_IBMVNIC) += ibmvnic.o
 obj-$(CONFIG_IBM_EMAC) += emac/
 obj-$(CONFIG_EHEA) += ehea/
diff --git a/drivers/net/ethernet/ibm/ibmvnic.c b/drivers/net/ethernet/ibm/ibmvnic.c
new file mode 100644
index 000000000000..7d6570843723
--- /dev/null
+++ b/drivers/net/ethernet/ibm/ibmvnic.c
@@ -0,0 +1,3585 @@
+/**************************************************************************/
+/*                                                                        */
+/*  IBM System i and System p Virtual NIC Device Driver                   */
+/*  Copyright (C) 2014 IBM Corp.                                          */
+/*  Santiago Leon (santi_leon@yahoo.com)                                  */
+/*  Thomas Falcon (tlfalcon@linux.vnet.ibm.com)                           */
+/*  John Allen (jallen@linux.vnet.ibm.com)                                */
+/*                                                                        */
+/*  This program is free software; you can redistribute it and/or modify  */
+/*  it under the terms of the GNU General Public License as published by  */
+/*  the Free Software Foundation; either version 2 of the License, or     */
+/*  (at your option) any later version.                                   */
+/*                                                                        */
+/*  This program is distributed in the hope that it will be useful,       */
+/*  but WITHOUT ANY WARRANTY; without even the implied warranty of        */
+/*  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the         */
+/*  GNU General Public License for more details.                          */
+/*                                                                        */
+/*  You should have received a copy of the GNU General Public License     */
+/*  along with this program.                                              */
+/*                                                                        */
+/* This module contains the implementation of a virtual ethernet device   */
+/* for use with IBM i/p Series LPAR Linux. It utilizes the logical LAN    */
+/* option of the RS/6000 Platform Architecture to interface with virtual  */
+/* ethernet NICs that are presented to the partition by the hypervisor.   */
+/*									   */
+/* Messages are passed between the VNIC driver and the VNIC server using  */
+/* Command/Response Queues (CRQs) and sub CRQs (sCRQs). CRQs are used to  */
+/* issue and receive commands that initiate communication with the server */
+/* on driver initialization. Sub CRQs (sCRQs) are similar to CRQs, but    */
+/* are used by the driver to notify the server that a packet is           */
+/* ready for transmission or that a buffer has been added to receive a    */
+/* packet. Subsequently, sCRQs are used by the server to notify the       */
+/* driver that a packet transmission has been completed or that a packet  */
+/* has been received and placed in a waiting buffer.                      */
+/*                                                                        */
+/* In lieu of a more conventional "on-the-fly" DMA mapping strategy in    */
+/* which skbs are DMA mapped and immediately unmapped when the transmit   */
+/* or receive has been completed, the VNIC driver is required to use      */
+/* "long term mapping". This entails that large, continuous DMA mapped    */
+/* buffers are allocated on driver initialization and these buffers are   */
+/* then continuously reused to pass skbs to and from the VNIC server.     */
+/*                                                                        */
+/**************************************************************************/
+
+#include <linux/module.h>
+#include <linux/moduleparam.h>
+#include <linux/types.h>
+#include <linux/errno.h>
+#include <linux/completion.h>
+#include <linux/ioport.h>
+#include <linux/dma-mapping.h>
+#include <linux/kernel.h>
+#include <linux/netdevice.h>
+#include <linux/etherdevice.h>
+#include <linux/skbuff.h>
+#include <linux/init.h>
+#include <linux/delay.h>
+#include <linux/mm.h>
+#include <linux/ethtool.h>
+#include <linux/proc_fs.h>
+#include <linux/in.h>
+#include <linux/ip.h>
+#include <linux/irq.h>
+#include <linux/kthread.h>
+#include <linux/seq_file.h>
+#include <linux/debugfs.h>
+#include <linux/interrupt.h>
+#include <net/net_namespace.h>
+#include <asm/hvcall.h>
+#include <linux/atomic.h>
+#include <asm/vio.h>
+#include <asm/iommu.h>
+#include <linux/uaccess.h>
+#include <asm/firmware.h>
+#include <linux/seq_file.h>
+
+#include "ibmvnic.h"
+
+static const char ibmvnic_driver_name[] = "ibmvnic";
+static const char ibmvnic_driver_string[] = "IBM System i/p Virtual NIC Driver";
+
+MODULE_AUTHOR("Santiago Leon <santi_leon@yahoo.com>");
+MODULE_DESCRIPTION("IBM System i/p Virtual NIC Driver");
+MODULE_LICENSE("GPL");
+MODULE_VERSION(IBMVNIC_DRIVER_VERSION);
+
+static int ibmvnic_version = IBMVNIC_INITIAL_VERSION;
+static int ibmvnic_remove(struct vio_dev *);
+static void release_sub_crqs(struct ibmvnic_adapter *);
+static int ibmvnic_reset_crq(struct ibmvnic_adapter *);
+static int ibmvnic_send_crq_init(struct ibmvnic_adapter *);
+static int ibmvnic_reenable_crq_queue(struct ibmvnic_adapter *);
+static int ibmvnic_send_crq(struct ibmvnic_adapter *, union ibmvnic_crq *);
+static int send_subcrq(struct ibmvnic_adapter *adapter, u64 remote_handle,
+		       union sub_crq *sub_crq);
+static irqreturn_t ibmvnic_interrupt_rx(int irq, void *instance);
+static int enable_scrq_irq(struct ibmvnic_adapter *,
+			   struct ibmvnic_sub_crq_queue *);
+static int disable_scrq_irq(struct ibmvnic_adapter *,
+			    struct ibmvnic_sub_crq_queue *);
+static int pending_scrq(struct ibmvnic_adapter *,
+			struct ibmvnic_sub_crq_queue *);
+static union sub_crq *ibmvnic_next_scrq(struct ibmvnic_adapter *,
+					struct ibmvnic_sub_crq_queue *);
+static int ibmvnic_poll(struct napi_struct *napi, int data);
+static void send_map_query(struct ibmvnic_adapter *adapter);
+static void send_request_map(struct ibmvnic_adapter *, dma_addr_t, __be32, u8);
+static void send_request_unmap(struct ibmvnic_adapter *, u8);
+
+struct ibmvnic_stat {
+	char name[ETH_GSTRING_LEN];
+	int offset;
+};
+
+#define IBMVNIC_STAT_OFF(stat) (offsetof(struct ibmvnic_adapter, stats) + \
+			     offsetof(struct ibmvnic_statistics, stat))
+#define IBMVNIC_GET_STAT(a, off) (*((u64 *)(((unsigned long)(a)) + off)))
+
+static const struct ibmvnic_stat ibmvnic_stats[] = {
+	{"rx_packets", IBMVNIC_STAT_OFF(rx_packets)},
+	{"rx_bytes", IBMVNIC_STAT_OFF(rx_bytes)},
+	{"tx_packets", IBMVNIC_STAT_OFF(tx_packets)},
+	{"tx_bytes", IBMVNIC_STAT_OFF(tx_bytes)},
+	{"ucast_tx_packets", IBMVNIC_STAT_OFF(ucast_tx_packets)},
+	{"ucast_rx_packets", IBMVNIC_STAT_OFF(ucast_rx_packets)},
+	{"mcast_tx_packets", IBMVNIC_STAT_OFF(mcast_tx_packets)},
+	{"mcast_rx_packets", IBMVNIC_STAT_OFF(mcast_rx_packets)},
+	{"bcast_tx_packets", IBMVNIC_STAT_OFF(bcast_tx_packets)},
+	{"bcast_rx_packets", IBMVNIC_STAT_OFF(bcast_rx_packets)},
+	{"align_errors", IBMVNIC_STAT_OFF(align_errors)},
+	{"fcs_errors", IBMVNIC_STAT_OFF(fcs_errors)},
+	{"single_collision_frames", IBMVNIC_STAT_OFF(single_collision_frames)},
+	{"multi_collision_frames", IBMVNIC_STAT_OFF(multi_collision_frames)},
+	{"sqe_test_errors", IBMVNIC_STAT_OFF(sqe_test_errors)},
+	{"deferred_tx", IBMVNIC_STAT_OFF(deferred_tx)},
+	{"late_collisions", IBMVNIC_STAT_OFF(late_collisions)},
+	{"excess_collisions", IBMVNIC_STAT_OFF(excess_collisions)},
+	{"internal_mac_tx_errors", IBMVNIC_STAT_OFF(internal_mac_tx_errors)},
+	{"carrier_sense", IBMVNIC_STAT_OFF(carrier_sense)},
+	{"too_long_frames", IBMVNIC_STAT_OFF(too_long_frames)},
+	{"internal_mac_rx_errors", IBMVNIC_STAT_OFF(internal_mac_rx_errors)},
+};
+
+static long h_reg_sub_crq(unsigned long unit_address, unsigned long token,
+			  unsigned long length, unsigned long *number,
+			  unsigned long *irq)
+{
+	unsigned long retbuf[PLPAR_HCALL_BUFSIZE];
+	long rc;
+
+	rc = plpar_hcall(H_REG_SUB_CRQ, retbuf, unit_address, token, length);
+	*number = retbuf[0];
+	*irq = retbuf[1];
+
+	return rc;
+}
+
+/* net_device_ops functions */
+
+static void init_rx_pool(struct ibmvnic_adapter *adapter,
+			 struct ibmvnic_rx_pool *rx_pool, int num, int index,
+			 int buff_size, int active)
+{
+	netdev_dbg(adapter->netdev,
+		   "Initializing rx_pool %d, %d buffs, %d bytes each\n",
+		   index, num, buff_size);
+	rx_pool->size = num;
+	rx_pool->index = index;
+	rx_pool->buff_size = buff_size;
+	rx_pool->active = active;
+}
+
+static int alloc_long_term_buff(struct ibmvnic_adapter *adapter,
+				struct ibmvnic_long_term_buff *ltb, int size)
+{
+	struct device *dev = &adapter->vdev->dev;
+
+	ltb->size = size;
+	ltb->buff = dma_alloc_coherent(dev, ltb->size, &ltb->addr,
+				       GFP_KERNEL);
+
+	if (!ltb->buff) {
+		dev_err(dev, "Couldn't alloc long term buffer\n");
+		return -ENOMEM;
+	}
+	ltb->map_id = adapter->map_id;
+	adapter->map_id++;
+	send_request_map(adapter, ltb->addr,
+			 ltb->size, ltb->map_id);
+	init_completion(&adapter->fw_done);
+	wait_for_completion(&adapter->fw_done);
+	return 0;
+}
+
+static void free_long_term_buff(struct ibmvnic_adapter *adapter,
+				struct ibmvnic_long_term_buff *ltb)
+{
+	struct device *dev = &adapter->vdev->dev;
+
+	dma_free_coherent(dev, ltb->size, ltb->buff, ltb->addr);
+	send_request_unmap(adapter, ltb->map_id);
+}
+
+static int alloc_rx_pool(struct ibmvnic_adapter *adapter,
+			 struct ibmvnic_rx_pool *pool)
+{
+	struct device *dev = &adapter->vdev->dev;
+	int i;
+
+	pool->free_map = kcalloc(pool->size, sizeof(int), GFP_KERNEL);
+	if (!pool->free_map)
+		return -ENOMEM;
+
+	pool->rx_buff = kcalloc(pool->size, sizeof(struct ibmvnic_rx_buff),
+				GFP_KERNEL);
+
+	if (!pool->rx_buff) {
+		dev_err(dev, "Couldn't alloc rx buffers\n");
+		kfree(pool->free_map);
+		return -ENOMEM;
+	}
+
+	if (alloc_long_term_buff(adapter, &pool->long_term_buff,
+				 pool->size * pool->buff_size)) {
+		kfree(pool->free_map);
+		kfree(pool->rx_buff);
+		return -ENOMEM;
+	}
+
+	for (i = 0; i < pool->size; ++i)
+		pool->free_map[i] = i;
+
+	atomic_set(&pool->available, 0);
+	pool->next_alloc = 0;
+	pool->next_free = 0;
+
+	return 0;
+}
+
+static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
+			      struct ibmvnic_rx_pool *pool)
+{
+	int count = pool->size - atomic_read(&pool->available);
+	struct device *dev = &adapter->vdev->dev;
+	int buffers_added = 0;
+	unsigned long lpar_rc;
+	union sub_crq sub_crq;
+	struct sk_buff *skb;
+	unsigned int offset;
+	dma_addr_t dma_addr;
+	unsigned char *dst;
+	u64 *handle_array;
+	int shift = 0;
+	int index;
+	int i;
+
+	handle_array = (u64 *)((u8 *)(adapter->login_rsp_buf) +
+				      be32_to_cpu(adapter->login_rsp_buf->
+				      off_rxadd_subcrqs));
+
+	for (i = 0; i < count; ++i) {
+		skb = alloc_skb(pool->buff_size, GFP_ATOMIC);
+		if (!skb) {
+			dev_err(dev, "Couldn't replenish rx buff\n");
+			adapter->replenish_no_mem++;
+			break;
+		}
+
+		index = pool->free_map[pool->next_free];
+
+		if (pool->rx_buff[index].skb)
+			dev_err(dev, "Inconsistent free_map!\n");
+
+		/* Copy the skb to the long term mapped DMA buffer */
+		offset = index * pool->buff_size;
+		dst = pool->long_term_buff.buff + offset;
+		memset(dst, 0, pool->buff_size);
+		dma_addr = pool->long_term_buff.addr + offset;
+		pool->rx_buff[index].data = dst;
+
+		pool->free_map[pool->next_free] = IBMVNIC_INVALID_MAP;
+		pool->rx_buff[index].dma = dma_addr;
+		pool->rx_buff[index].skb = skb;
+		pool->rx_buff[index].pool_index = pool->index;
+		pool->rx_buff[index].size = pool->buff_size;
+
+		memset(&sub_crq, 0, sizeof(sub_crq));
+		sub_crq.rx_add.first = IBMVNIC_CRQ_CMD;
+		sub_crq.rx_add.correlator =
+		    cpu_to_be64((u64)&pool->rx_buff[index]);
+		sub_crq.rx_add.ioba = cpu_to_be32(dma_addr);
+		sub_crq.rx_add.map_id = pool->long_term_buff.map_id;
+
+		/* The length field of the sCRQ is defined to be 24 bits so the
+		 * buffer size needs to be left shifted by a byte before it is
+		 * converted to big endian to prevent the last byte from being
+		 * truncated.
+		 */
+#ifdef __LITTLE_ENDIAN__
+		shift = 8;
+#endif
+		sub_crq.rx_add.len = cpu_to_be32(pool->buff_size << shift);
+
+		lpar_rc = send_subcrq(adapter, handle_array[pool->index],
+				      &sub_crq);
+		if (lpar_rc != H_SUCCESS)
+			goto failure;
+
+		buffers_added++;
+		adapter->replenish_add_buff_success++;
+		pool->next_free = (pool->next_free + 1) % pool->size;
+	}
+	atomic_add(buffers_added, &pool->available);
+	return;
+
+failure:
+	dev_info(dev, "replenish pools failure\n");
+	pool->free_map[pool->next_free] = index;
+	pool->rx_buff[index].skb = NULL;
+	if (!dma_mapping_error(dev, dma_addr))
+		dma_unmap_single(dev, dma_addr, pool->buff_size,
+				 DMA_FROM_DEVICE);
+
+	dev_kfree_skb_any(skb);
+	adapter->replenish_add_buff_failure++;
+	atomic_add(buffers_added, &pool->available);
+}
+
+static void replenish_pools(struct ibmvnic_adapter *adapter)
+{
+	int i;
+
+	if (adapter->migrated)
+		return;
+
+	adapter->replenish_task_cycles++;
+	for (i = 0; i < be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
+	     i++) {
+		if (adapter->rx_pool[i].active)
+			replenish_rx_pool(adapter, &adapter->rx_pool[i]);
+	}
+}
+
+static void free_rx_pool(struct ibmvnic_adapter *adapter,
+			 struct ibmvnic_rx_pool *pool)
+{
+	int i;
+
+	kfree(pool->free_map);
+	pool->free_map = NULL;
+
+	if (!pool->rx_buff)
+		return;
+
+	for (i = 0; i < pool->size; i++) {
+		if (pool->rx_buff[i].skb) {
+			dev_kfree_skb_any(pool->rx_buff[i].skb);
+			pool->rx_buff[i].skb = NULL;
+		}
+	}
+	kfree(pool->rx_buff);
+	pool->rx_buff = NULL;
+}
+
+static int ibmvnic_open(struct net_device *netdev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	struct device *dev = &adapter->vdev->dev;
+	struct ibmvnic_tx_pool *tx_pool;
+	union ibmvnic_crq crq;
+	int rxadd_subcrqs;
+	u64 *size_array;
+	int tx_subcrqs;
+	int i, j;
+
+	rxadd_subcrqs =
+	    be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
+	tx_subcrqs =
+	    be32_to_cpu(adapter->login_rsp_buf->num_txsubm_subcrqs);
+	size_array = (u64 *)((u8 *)(adapter->login_rsp_buf) +
+				  be32_to_cpu(adapter->login_rsp_buf->
+					      off_rxadd_buff_size));
+	adapter->map_id = 1;
+	adapter->napi = kcalloc(adapter->req_rx_queues,
+				sizeof(struct napi_struct), GFP_KERNEL);
+	if (!adapter->napi)
+		goto alloc_napi_failed;
+	for (i = 0; i < adapter->req_rx_queues; i++) {
+		netif_napi_add(netdev, &adapter->napi[i], ibmvnic_poll,
+			       NAPI_POLL_WEIGHT);
+		napi_enable(&adapter->napi[i]);
+	}
+	adapter->rx_pool =
+	    kcalloc(rxadd_subcrqs, sizeof(struct ibmvnic_rx_pool), GFP_KERNEL);
+
+	if (!adapter->rx_pool)
+		goto rx_pool_arr_alloc_failed;
+	send_map_query(adapter);
+	for (i = 0; i < rxadd_subcrqs; i++) {
+		init_rx_pool(adapter, &adapter->rx_pool[i],
+			     IBMVNIC_BUFFS_PER_POOL, i,
+			     be64_to_cpu(size_array[i]), 1);
+		if (alloc_rx_pool(adapter, &adapter->rx_pool[i])) {
+			dev_err(dev, "Couldn't alloc rx pool\n");
+			goto rx_pool_alloc_failed;
+		}
+	}
+	adapter->tx_pool =
+	    kcalloc(tx_subcrqs, sizeof(struct ibmvnic_tx_pool), GFP_KERNEL);
+
+	if (!adapter->tx_pool)
+		goto tx_pool_arr_alloc_failed;
+	for (i = 0; i < tx_subcrqs; i++) {
+		tx_pool = &adapter->tx_pool[i];
+		tx_pool->tx_buff =
+		    kcalloc(adapter->max_tx_entries_per_subcrq,
+			    sizeof(struct ibmvnic_tx_buff), GFP_KERNEL);
+		if (!tx_pool->tx_buff)
+			goto tx_pool_alloc_failed;
+
+		if (alloc_long_term_buff(adapter, &tx_pool->long_term_buff,
+					 adapter->max_tx_entries_per_subcrq *
+					 adapter->req_mtu))
+			goto tx_ltb_alloc_failed;
+
+		tx_pool->free_map =
+		    kcalloc(adapter->max_tx_entries_per_subcrq,
+			    sizeof(int), GFP_KERNEL);
+		if (!tx_pool->free_map)
+			goto tx_fm_alloc_failed;
+
+		for (j = 0; j < adapter->max_tx_entries_per_subcrq; j++)
+			tx_pool->free_map[j] = j;
+
+		tx_pool->consumer_index = 0;
+		tx_pool->producer_index = 0;
+	}
+	adapter->bounce_buffer_size =
+	    (netdev->mtu + ETH_HLEN - 1) / PAGE_SIZE + 1;
+	adapter->bounce_buffer = kmalloc(adapter->bounce_buffer_size,
+					 GFP_KERNEL);
+	if (!adapter->bounce_buffer)
+		goto bounce_alloc_failed;
+
+	adapter->bounce_buffer_dma = dma_map_single(dev, adapter->bounce_buffer,
+						    adapter->bounce_buffer_size,
+						    DMA_TO_DEVICE);
+	if (dma_mapping_error(dev, adapter->bounce_buffer_dma)) {
+		dev_err(dev, "Couldn't map tx bounce buffer\n");
+		goto bounce_map_failed;
+	}
+	replenish_pools(adapter);
+
+	/* We're ready to receive frames, enable the sub-crq interrupts and
+	 * set the logical link state to up
+	 */
+	for (i = 0; i < adapter->req_rx_queues; i++)
+		enable_scrq_irq(adapter, adapter->rx_scrq[i]);
+
+	for (i = 0; i < adapter->req_tx_queues; i++)
+		enable_scrq_irq(adapter, adapter->tx_scrq[i]);
+
+	memset(&crq, 0, sizeof(crq));
+	crq.logical_link_state.first = IBMVNIC_CRQ_CMD;
+	crq.logical_link_state.cmd = LOGICAL_LINK_STATE;
+	crq.logical_link_state.link_state = IBMVNIC_LOGICAL_LNK_UP;
+	ibmvnic_send_crq(adapter, &crq);
+
+	netif_start_queue(netdev);
+	return 0;
+
+bounce_map_failed:
+	kfree(adapter->bounce_buffer);
+bounce_alloc_failed:
+	i = tx_subcrqs - 1;
+	kfree(adapter->tx_pool[i].free_map);
+tx_fm_alloc_failed:
+	free_long_term_buff(adapter, &adapter->tx_pool[i].long_term_buff);
+tx_ltb_alloc_failed:
+	kfree(adapter->tx_pool[i].tx_buff);
+tx_pool_alloc_failed:
+	for (j = 0; j < i; j++) {
+		kfree(adapter->tx_pool[j].tx_buff);
+		free_long_term_buff(adapter,
+				    &adapter->tx_pool[j].long_term_buff);
+		kfree(adapter->tx_pool[j].free_map);
+	}
+	kfree(adapter->tx_pool);
+	adapter->tx_pool = NULL;
+tx_pool_arr_alloc_failed:
+	i = rxadd_subcrqs;
+rx_pool_alloc_failed:
+	for (j = 0; j < i; j++) {
+		free_rx_pool(adapter, &adapter->rx_pool[j]);
+		free_long_term_buff(adapter,
+				    &adapter->rx_pool[j].long_term_buff);
+	}
+	kfree(adapter->rx_pool);
+	adapter->rx_pool = NULL;
+rx_pool_arr_alloc_failed:
+	for (i = 0; i < adapter->req_rx_queues; i++)
+		napi_enable(&adapter->napi[i]);
+alloc_napi_failed:
+	return -ENOMEM;
+}
+
+static int ibmvnic_close(struct net_device *netdev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	struct device *dev = &adapter->vdev->dev;
+	union ibmvnic_crq crq;
+	int i;
+
+	adapter->closing = true;
+
+	for (i = 0; i < adapter->req_rx_queues; i++)
+		napi_disable(&adapter->napi[i]);
+
+	netif_stop_queue(netdev);
+
+	if (adapter->bounce_buffer) {
+		if (!dma_mapping_error(dev, adapter->bounce_buffer_dma)) {
+			dma_unmap_single(&adapter->vdev->dev,
+					 adapter->bounce_buffer_dma,
+					 adapter->bounce_buffer_size,
+					 DMA_BIDIRECTIONAL);
+			adapter->bounce_buffer_dma = DMA_ERROR_CODE;
+		}
+		kfree(adapter->bounce_buffer);
+		adapter->bounce_buffer = NULL;
+	}
+
+	memset(&crq, 0, sizeof(crq));
+	crq.logical_link_state.first = IBMVNIC_CRQ_CMD;
+	crq.logical_link_state.cmd = LOGICAL_LINK_STATE;
+	crq.logical_link_state.link_state = IBMVNIC_LOGICAL_LNK_DN;
+	ibmvnic_send_crq(adapter, &crq);
+
+	for (i = 0; i < be32_to_cpu(adapter->login_rsp_buf->num_txsubm_subcrqs);
+	     i++) {
+		kfree(adapter->tx_pool[i].tx_buff);
+		free_long_term_buff(adapter,
+				    &adapter->tx_pool[i].long_term_buff);
+		kfree(adapter->tx_pool[i].free_map);
+	}
+	kfree(adapter->tx_pool);
+	adapter->tx_pool = NULL;
+
+	for (i = 0; i < be32_to_cpu(adapter->login_rsp_buf->num_rxadd_subcrqs);
+	     i++) {
+		free_rx_pool(adapter, &adapter->rx_pool[i]);
+		free_long_term_buff(adapter,
+				    &adapter->rx_pool[i].long_term_buff);
+	}
+	kfree(adapter->rx_pool);
+	adapter->rx_pool = NULL;
+
+	adapter->closing = false;
+
+	return 0;
+}
+
+static int ibmvnic_xmit(struct sk_buff *skb, struct net_device *netdev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	int queue_num = skb_get_queue_mapping(skb);
+	struct device *dev = &adapter->vdev->dev;
+	struct ibmvnic_tx_buff *tx_buff = NULL;
+	struct ibmvnic_tx_pool *tx_pool;
+	unsigned int tx_send_failed = 0;
+	unsigned int tx_map_failed = 0;
+	unsigned int tx_dropped = 0;
+	unsigned int tx_packets = 0;
+	unsigned int tx_bytes = 0;
+	dma_addr_t data_dma_addr;
+	struct netdev_queue *txq;
+	bool used_bounce = false;
+	unsigned long lpar_rc;
+	union sub_crq tx_crq;
+	unsigned int offset;
+	unsigned char *dst;
+	u64 *handle_array;
+	int index = 0;
+	int ret = 0;
+
+	tx_pool = &adapter->tx_pool[queue_num];
+	txq = netdev_get_tx_queue(netdev, skb_get_queue_mapping(skb));
+	handle_array = (u64 *)((u8 *)(adapter->login_rsp_buf) +
+				   be32_to_cpu(adapter->login_rsp_buf->
+					       off_txsubm_subcrqs));
+	if (adapter->migrated) {
+		tx_send_failed++;
+		tx_dropped++;
+		ret = NETDEV_TX_BUSY;
+		goto out;
+	}
+
+	index = tx_pool->free_map[tx_pool->consumer_index];
+	offset = index * adapter->req_mtu;
+	dst = tx_pool->long_term_buff.buff + offset;
+	memset(dst, 0, adapter->req_mtu);
+	skb_copy_from_linear_data(skb, dst, skb->len);
+	data_dma_addr = tx_pool->long_term_buff.addr + offset;
+
+	tx_pool->consumer_index =
+	    (tx_pool->consumer_index + 1) %
+		adapter->max_tx_entries_per_subcrq;
+
+	tx_buff = &tx_pool->tx_buff[index];
+	tx_buff->skb = skb;
+	tx_buff->data_dma[0] = data_dma_addr;
+	tx_buff->data_len[0] = skb->len;
+	tx_buff->index = index;
+	tx_buff->pool_index = queue_num;
+	tx_buff->last_frag = true;
+	tx_buff->used_bounce = used_bounce;
+
+	memset(&tx_crq, 0, sizeof(tx_crq));
+	tx_crq.v1.first = IBMVNIC_CRQ_CMD;
+	tx_crq.v1.type = IBMVNIC_TX_DESC;
+	tx_crq.v1.n_crq_elem = 1;
+	tx_crq.v1.n_sge = 1;
+	tx_crq.v1.flags1 = IBMVNIC_TX_COMP_NEEDED;
+	tx_crq.v1.correlator = cpu_to_be32(index);
+	tx_crq.v1.dma_reg = cpu_to_be16(tx_pool->long_term_buff.map_id);
+	tx_crq.v1.sge_len = cpu_to_be32(skb->len);
+	tx_crq.v1.ioba = cpu_to_be64(data_dma_addr);
+
+	if (adapter->vlan_header_insertion) {
+		tx_crq.v1.flags2 |= IBMVNIC_TX_VLAN_INSERT;
+		tx_crq.v1.vlan_id = cpu_to_be16(skb->vlan_tci);
+	}
+
+	if (skb->protocol == htons(ETH_P_IP)) {
+		if (ip_hdr(skb)->version == 4)
+			tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_IPV4;
+		else if (ip_hdr(skb)->version == 6)
+			tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_IPV6;
+
+		if (ip_hdr(skb)->protocol == IPPROTO_TCP)
+			tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_TCP;
+		else if (ip_hdr(skb)->protocol != IPPROTO_TCP)
+			tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_UDP;
+	}
+
+	if (skb->ip_summed == CHECKSUM_PARTIAL)
+		tx_crq.v1.flags1 |= IBMVNIC_TX_CHKSUM_OFFLOAD;
+
+	lpar_rc = send_subcrq(adapter, handle_array[0], &tx_crq);
+
+	if (lpar_rc != H_SUCCESS) {
+		dev_err(dev, "tx failed with code %ld\n", lpar_rc);
+
+		if (tx_pool->consumer_index == 0)
+			tx_pool->consumer_index =
+				adapter->max_tx_entries_per_subcrq - 1;
+		else
+			tx_pool->consumer_index--;
+
+		tx_send_failed++;
+		tx_dropped++;
+		ret = NETDEV_TX_BUSY;
+		goto out;
+	}
+	tx_packets++;
+	tx_bytes += skb->len;
+	txq->trans_start = jiffies;
+	ret = NETDEV_TX_OK;
+
+out:
+	netdev->stats.tx_dropped += tx_dropped;
+	netdev->stats.tx_bytes += tx_bytes;
+	netdev->stats.tx_packets += tx_packets;
+	adapter->tx_send_failed += tx_send_failed;
+	adapter->tx_map_failed += tx_map_failed;
+
+	return ret;
+}
+
+static void ibmvnic_set_multi(struct net_device *netdev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	struct netdev_hw_addr *ha;
+	union ibmvnic_crq crq;
+
+	memset(&crq, 0, sizeof(crq));
+	crq.request_capability.first = IBMVNIC_CRQ_CMD;
+	crq.request_capability.cmd = REQUEST_CAPABILITY;
+
+	if (netdev->flags & IFF_PROMISC) {
+		if (!adapter->promisc_supported)
+			return;
+	} else {
+		if (netdev->flags & IFF_ALLMULTI) {
+			/* Accept all multicast */
+			memset(&crq, 0, sizeof(crq));
+			crq.multicast_ctrl.first = IBMVNIC_CRQ_CMD;
+			crq.multicast_ctrl.cmd = MULTICAST_CTRL;
+			crq.multicast_ctrl.flags = IBMVNIC_ENABLE_ALL;
+			ibmvnic_send_crq(adapter, &crq);
+		} else if (netdev_mc_empty(netdev)) {
+			/* Reject all multicast */
+			memset(&crq, 0, sizeof(crq));
+			crq.multicast_ctrl.first = IBMVNIC_CRQ_CMD;
+			crq.multicast_ctrl.cmd = MULTICAST_CTRL;
+			crq.multicast_ctrl.flags = IBMVNIC_DISABLE_ALL;
+			ibmvnic_send_crq(adapter, &crq);
+		} else {
+			/* Accept one or more multicast(s) */
+			netdev_for_each_mc_addr(ha, netdev) {
+				memset(&crq, 0, sizeof(crq));
+				crq.multicast_ctrl.first = IBMVNIC_CRQ_CMD;
+				crq.multicast_ctrl.cmd = MULTICAST_CTRL;
+				crq.multicast_ctrl.flags = IBMVNIC_ENABLE_MC;
+				ether_addr_copy(&crq.multicast_ctrl.mac_addr[0],
+						ha->addr);
+				ibmvnic_send_crq(adapter, &crq);
+			}
+		}
+	}
+}
+
+static int ibmvnic_set_mac(struct net_device *netdev, void *p)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	struct sockaddr *addr = p;
+	union ibmvnic_crq crq;
+
+	if (!is_valid_ether_addr(addr->sa_data))
+		return -EADDRNOTAVAIL;
+
+	memset(&crq, 0, sizeof(crq));
+	crq.change_mac_addr.first = IBMVNIC_CRQ_CMD;
+	crq.change_mac_addr.cmd = CHANGE_MAC_ADDR;
+	ether_addr_copy(&crq.change_mac_addr.mac_addr[0], addr->sa_data);
+	ibmvnic_send_crq(adapter, &crq);
+	/* netdev->dev_addr is changed in handle_change_mac_rsp function */
+	return 0;
+}
+
+static int ibmvnic_change_mtu(struct net_device *netdev, int new_mtu)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+
+	if (new_mtu > adapter->req_mtu || new_mtu < adapter->min_mtu)
+		return -EINVAL;
+
+	netdev->mtu = new_mtu;
+	return 0;
+}
+
+static void ibmvnic_tx_timeout(struct net_device *dev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(dev);
+	int rc;
+
+	/* Adapter timed out, resetting it */
+	release_sub_crqs(adapter);
+	rc = ibmvnic_reset_crq(adapter);
+	if (rc)
+		dev_err(&adapter->vdev->dev, "Adapter timeout, reset failed\n");
+	else
+		ibmvnic_send_crq_init(adapter);
+}
+
+static void remove_buff_from_pool(struct ibmvnic_adapter *adapter,
+				  struct ibmvnic_rx_buff *rx_buff)
+{
+	struct ibmvnic_rx_pool *pool = &adapter->rx_pool[rx_buff->pool_index];
+
+	rx_buff->skb = NULL;
+
+	pool->free_map[pool->next_alloc] = (int)(rx_buff - pool->rx_buff);
+	pool->next_alloc = (pool->next_alloc + 1) % pool->size;
+
+	atomic_dec(&pool->available);
+}
+
+static int ibmvnic_poll(struct napi_struct *napi, int budget)
+{
+	struct net_device *netdev = napi->dev;
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+	int scrq_num = (int)(napi - adapter->napi);
+	int frames_processed = 0;
+restart_poll:
+	while (frames_processed < budget) {
+		struct sk_buff *skb;
+		struct ibmvnic_rx_buff *rx_buff;
+		union sub_crq *next;
+		u32 length;
+		u16 offset;
+		u8 flags = 0;
+
+		if (!pending_scrq(adapter, adapter->rx_scrq[scrq_num]))
+			break;
+		next = ibmvnic_next_scrq(adapter, adapter->rx_scrq[scrq_num]);
+		rx_buff =
+		    (struct ibmvnic_rx_buff *)be64_to_cpu(next->
+							  rx_comp.correlator);
+		/* do error checking */
+		if (next->rx_comp.rc) {
+			netdev_err(netdev, "rx error %x\n", next->rx_comp.rc);
+			/* free the entry */
+			next->rx_comp.first = 0;
+			remove_buff_from_pool(adapter, rx_buff);
+			break;
+		}
+
+		length = be32_to_cpu(next->rx_comp.len);
+		offset = be16_to_cpu(next->rx_comp.off_frame_data);
+		flags = next->rx_comp.flags;
+		skb = rx_buff->skb;
+		skb_copy_to_linear_data(skb, rx_buff->data + offset,
+					length);
+		skb->vlan_tci = be16_to_cpu(next->rx_comp.vlan_tci);
+		/* free the entry */
+		next->rx_comp.first = 0;
+		remove_buff_from_pool(adapter, rx_buff);
+
+		skb_put(skb, length);
+		skb->protocol = eth_type_trans(skb, netdev);
+
+		if (flags & IBMVNIC_IP_CHKSUM_GOOD &&
+		    flags & IBMVNIC_TCP_UDP_CHKSUM_GOOD) {
+			skb->ip_summed = CHECKSUM_UNNECESSARY;
+		}
+
+		length = skb->len;
+		napi_gro_receive(napi, skb); /* send it up */
+		netdev->stats.rx_packets++;
+		netdev->stats.rx_bytes += length;
+		frames_processed++;
+	}
+	replenish_pools(adapter);
+
+	if (frames_processed < budget) {
+		enable_scrq_irq(adapter, adapter->rx_scrq[scrq_num]);
+		napi_complete(napi);
+		if (pending_scrq(adapter, adapter->rx_scrq[scrq_num]) &&
+		    napi_reschedule(napi)) {
+			disable_scrq_irq(adapter, adapter->rx_scrq[scrq_num]);
+			goto restart_poll;
+		}
+	}
+	return frames_processed;
+}
+
+#ifdef CONFIG_NET_POLL_CONTROLLER
+static void ibmvnic_netpoll_controller(struct net_device *dev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(dev);
+	int i;
+
+	replenish_pools(netdev_priv(dev));
+	for (i = 0; i < adapter->req_rx_queues; i++)
+		ibmvnic_interrupt_rx(adapter->rx_scrq[i]->irq,
+				     adapter->rx_scrq[i]);
+}
+#endif
+
+static const struct net_device_ops ibmvnic_netdev_ops = {
+	.ndo_open		= ibmvnic_open,
+	.ndo_stop		= ibmvnic_close,
+	.ndo_start_xmit		= ibmvnic_xmit,
+	.ndo_set_rx_mode	= ibmvnic_set_multi,
+	.ndo_set_mac_address	= ibmvnic_set_mac,
+	.ndo_validate_addr	= eth_validate_addr,
+	.ndo_change_mtu		= ibmvnic_change_mtu,
+	.ndo_tx_timeout		= ibmvnic_tx_timeout,
+#ifdef CONFIG_NET_POLL_CONTROLLER
+	.ndo_poll_controller	= ibmvnic_netpoll_controller,
+#endif
+};
+
+/* ethtool functions */
+
+static int ibmvnic_get_settings(struct net_device *netdev,
+				struct ethtool_cmd *cmd)
+{
+	cmd->supported = (SUPPORTED_1000baseT_Full | SUPPORTED_Autoneg |
+			  SUPPORTED_FIBRE);
+	cmd->advertising = (ADVERTISED_1000baseT_Full | ADVERTISED_Autoneg |
+			    ADVERTISED_FIBRE);
+	ethtool_cmd_speed_set(cmd, SPEED_1000);
+	cmd->duplex = DUPLEX_FULL;
+	cmd->port = PORT_FIBRE;
+	cmd->phy_address = 0;
+	cmd->transceiver = XCVR_INTERNAL;
+	cmd->autoneg = AUTONEG_ENABLE;
+	cmd->maxtxpkt = 0;
+	cmd->maxrxpkt = 1;
+	return 0;
+}
+
+static void ibmvnic_get_drvinfo(struct net_device *dev,
+				struct ethtool_drvinfo *info)
+{
+	strlcpy(info->driver, ibmvnic_driver_name, sizeof(info->driver));
+	strlcpy(info->version, IBMVNIC_DRIVER_VERSION, sizeof(info->version));
+}
+
+static u32 ibmvnic_get_msglevel(struct net_device *netdev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+
+	return adapter->msg_enable;
+}
+
+static void ibmvnic_set_msglevel(struct net_device *netdev, u32 data)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+
+	adapter->msg_enable = data;
+}
+
+static u32 ibmvnic_get_link(struct net_device *netdev)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
+
+	/* Don't need to send a query because we request a logical link up at
+	 * init and then we wait for link state indications
+	 */
+	return adapter->logical_link_state;
+}
+
+static void ibmvnic_get_ringparam(struct net_device *netdev,
+				  struct ethtool_ringparam *ring)
+{
+	ring->rx_max_pending = 0;
+	ring->tx_max_pending = 0;
+	ring->rx_mini_max_pending = 0;
+	ring->rx_jumbo_max_pending = 0;
+	ring->rx_pending = 0;
+	ring->tx_pending = 0;
+	ring->rx_mini_pending = 0;
+	ring->rx_jumbo_pending = 0;
+}
+
+static void ibmvnic_get_strings(struct net_device *dev, u32 stringset, u8 *data)
+{
+	int i;
+
+	if (stringset != ETH_SS_STATS)
+		return;
+
+	for (i = 0; i < ARRAY_SIZE(ibmvnic_stats); i++, data += ETH_GSTRING_LEN)
+		memcpy(data, ibmvnic_stats[i].name, ETH_GSTRING_LEN);
+}
+
+static int ibmvnic_get_sset_count(struct net_device *dev, int sset)
+{
+	switch (sset) {
+	case ETH_SS_STATS:
+		return ARRAY_SIZE(ibmvnic_stats);
+	default:
+		return -EOPNOTSUPP;
+	}
+}
+
+static void ibmvnic_get_ethtool_stats(struct net_device *dev,
+				      struct ethtool_stats *stats, u64 *data)
+{
+	struct ibmvnic_adapter *adapter = netdev_priv(dev);
+	union ibmvnic_crq crq;
+	int i;
+
+	memset(&crq, 0, sizeof(crq));
+	crq.request_statistics.first = IBMVNIC_CRQ_CMD;
+	crq.request_statistics.cmd = REQUEST_STATISTICS;
+	crq.request_statistics.ioba = cpu_to_be32(adapter->stats_token);
+	crq.request_statistics.len =
+	    cpu_to_be32(sizeof(struct ibmvnic_statistics));
+	ibmvnic_send_crq(adapter, &crq);
+
+	/* Wait for data to be written */
+	init_completion(&adapter->stats_done);
+	wait_for_completion(&adapter->stats_done);
+
+	for (i = 0; i < ARRAY_SIZE(ibmvnic_stats); i++)
+		data[i] = IBMVNIC_GET_STAT(adapter, ibmvnic_stats[i].offset);
+}
+
+static const struct ethtool_ops ibmvnic_ethtool_ops = {
+	.get_settings		= ibmvnic_get_settings,
+	.get_drvinfo		= ibmvnic_get_drvinfo,
+	.get_msglevel		= ibmvnic_get_msglevel,
+	.set_msglevel		= ibmvnic_set_msglevel,
+	.get_link		= ibmvnic_get_link,
+	.get_ringparam		= ibmvnic_get_ringparam,
+	.get_strings            = ibmvnic_get_strings,
+	.get_sset_count         = ibmvnic_get_sset_count,