drivers/net/ethernet/ibm/ibmvnic.c

// SPDX-License-Identifier: GPL-2.0-or-later
/**************************************************************************/
/*                                                                        */
/*  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 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/if_arp.h>
#include <linux/in.h>
#include <linux/ip.h>
#include <linux/ipv6.h>
#include <linux/irq.h>
#include <linux/kthread.h>
#include <linux/seq_file.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/workqueue.h>
#include <linux/if_vlan.h>
#include <linux/utsname.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");
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 void release_sub_crqs(struct ibmvnic_adapter *, bool);
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_indirect(struct ibmvnic_adapter *, u64, u64, u64);
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_query_map(struct ibmvnic_adapter *adapter);
static int send_request_map(struct ibmvnic_adapter *, dma_addr_t, u32, u8);
static int send_request_unmap(struct ibmvnic_adapter *, u8);
static int send_login(struct ibmvnic_adapter *adapter);
static void send_query_cap(struct ibmvnic_adapter *adapter);
static int init_sub_crqs(struct ibmvnic_adapter *);
static int init_sub_crq_irqs(struct ibmvnic_adapter *adapter);
static int ibmvnic_reset_init(struct ibmvnic_adapter *, bool reset);
static void release_crq_queue(struct ibmvnic_adapter *);
static int __ibmvnic_set_mac(struct net_device *, u8 *);
static int init_crq_queue(struct ibmvnic_adapter *adapter);
static int send_query_phys_parms(struct ibmvnic_adapter *adapter);
static void ibmvnic_tx_scrq_clean_buffer(struct ibmvnic_adapter *adapter,
					 struct ibmvnic_sub_crq_queue *tx_scrq);
static void free_long_term_buff(struct ibmvnic_adapter *adapter,
				struct ibmvnic_long_term_buff *ltb);

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 int send_crq_init_complete(struct ibmvnic_adapter *adapter)
{
	union ibmvnic_crq crq;

	memset(&crq, 0, sizeof(crq));
	crq.generic.first = IBMVNIC_CRQ_INIT_CMD;
	crq.generic.cmd = IBMVNIC_CRQ_INIT_COMPLETE;

	return ibmvnic_send_crq(adapter, &crq);
}

static int send_version_xchg(struct ibmvnic_adapter *adapter)
{
	union ibmvnic_crq crq;

	memset(&crq, 0, sizeof(crq));
	crq.version_exchange.first = IBMVNIC_CRQ_CMD;
	crq.version_exchange.cmd = VERSION_EXCHANGE;
	crq.version_exchange.version = cpu_to_be16(ibmvnic_version);

	return ibmvnic_send_crq(adapter, &crq);
}

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;
}

/**
 * ibmvnic_wait_for_completion - Check device state and wait for completion
 * @adapter: private device data
 * @comp_done: completion structure to wait for
 * @timeout: time to wait in milliseconds
 *
 * Wait for a completion signal or until the timeout limit is reached
 * while checking that the device is still active.
 */
static int ibmvnic_wait_for_completion(struct ibmvnic_adapter *adapter,
				       struct completion *comp_done,
				       unsigned long timeout)
{
	struct net_device *netdev;
	unsigned long div_timeout;
	u8 retry;

	netdev = adapter->netdev;
	retry = 5;
	div_timeout = msecs_to_jiffies(timeout / retry);
	while (true) {
		if (!adapter->crq.active) {
			netdev_err(netdev, "Device down!\n");
			return -ENODEV;
		}
		if (!retry--)
			break;
		if (wait_for_completion_timeout(comp_done, div_timeout))
			return 0;
	}
	netdev_err(netdev, "Operation timed out.\n");
	return -ETIMEDOUT;
}

/**
 * reuse_ltb() - Check if a long term buffer can be reused
 * @ltb:  The long term buffer to be checked
 * @size: The size of the long term buffer.
 *
 * An LTB can be reused unless its size has changed.
 *
 * Return: Return true if the LTB can be reused, false otherwise.
 */
static bool reuse_ltb(struct ibmvnic_long_term_buff *ltb, int size)
{
	return (ltb->buff && ltb->size == size);
}

/**
 * alloc_long_term_buff() - Allocate a long term buffer (LTB)
 *
 * @adapter: ibmvnic adapter associated to the LTB
 * @ltb:     container object for the LTB
 * @size:    size of the LTB
 *
 * Allocate an LTB of the specified size and notify VIOS.
 *
 * If the given @ltb already has the correct size, reuse it. Otherwise if
 * its non-NULL, free it. Then allocate a new one of the correct size.
 * Notify the VIOS either way since we may now be working with a new VIOS.
 *
 * Allocating larger chunks of memory during resets, specially LPM or under
 * low memory situations can cause resets to fail/timeout and for LPAR to
 * lose connectivity. So hold onto the LTB even if we fail to communicate
 * with the VIOS and reuse it on next open. Free LTB when adapter is closed.
 *
 * Return: 0 if we were able to allocate the LTB and notify the VIOS and
 *	   a negative value otherwise.
 */
static int alloc_long_term_buff(struct ibmvnic_adapter *adapter,
				struct ibmvnic_long_term_buff *ltb, int size)
{
	struct device *dev = &adapter->vdev->dev;
	int rc;

	if (!reuse_ltb(ltb, size)) {
		dev_dbg(dev,
			"LTB size changed from 0x%llx to 0x%x, reallocating\n",
			 ltb->size, size);
		free_long_term_buff(adapter, ltb);
	}

	if (ltb->buff) {
		dev_dbg(dev, "Reusing LTB [map %d, size 0x%llx]\n",
			ltb->map_id, ltb->size);
	} else {
		ltb->buff = dma_alloc_coherent(dev, size, &ltb->addr,
					       GFP_KERNEL);
		if (!ltb->buff) {
			dev_err(dev, "Couldn't alloc long term buffer\n");
			return -ENOMEM;
		}
		ltb->size = size;

		ltb->map_id = find_first_zero_bit(adapter->map_ids,
						  MAX_MAP_ID);
		bitmap_set(adapter->map_ids, ltb->map_id, 1);

		dev_dbg(dev,
			"Allocated new LTB [map %d, size 0x%llx]\n",
			 ltb->map_id, ltb->size);
	}

	/* Ensure ltb is zeroed - specially when reusing it. */
	memset(ltb->buff, 0, ltb->size);

	mutex_lock(&adapter->fw_lock);
	adapter->fw_done_rc = 0;
	reinit_completion(&adapter->fw_done);

	rc = send_request_map(adapter, ltb->addr, ltb->size, ltb->map_id);
	if (rc) {
		dev_err(dev, "send_request_map failed, rc = %d\n", rc);
		goto out;
	}

	rc = ibmvnic_wait_for_completion(adapter, &adapter->fw_done, 10000);
	if (rc) {
		dev_err(dev, "LTB map request aborted or timed out, rc = %d\n",
			rc);
		goto out;
	}

	if (adapter->fw_done_rc) {
		dev_err(dev, "Couldn't map LTB, rc = %d\n",
			adapter->fw_done_rc);
		rc = -1;
		goto out;
	}
	rc = 0;
out:
	/* don't free LTB on communication error - see function header */
	mutex_unlock(&adapter->fw_lock);
	return rc;
}

static void free_long_term_buff(struct ibmvnic_adapter *adapter,
				struct ibmvnic_long_term_buff *ltb)
{
	struct device *dev = &adapter->vdev->dev;

	if (!ltb->buff)
		return;

	/* VIOS automatically unmaps the long term buffer at remote
	 * end for the following resets:
	 * FAILOVER, MOBILITY, TIMEOUT.
	 */
	if (adapter->reset_reason != VNIC_RESET_FAILOVER &&
	    adapter->reset_reason != VNIC_RESET_MOBILITY &&
	    adapter->reset_reason != VNIC_RESET_TIMEOUT)
		send_request_unmap(adapter, ltb->map_id);

	dma_free_coherent(dev, ltb->size, ltb->buff, ltb->addr);

	ltb->buff = NULL;
	/* mark this map_id free */
	bitmap_clear(adapter->map_ids, ltb->map_id, 1);
	ltb->map_id = 0;
}

static void deactivate_rx_pools(struct ibmvnic_adapter *adapter)
{
	int i;

	for (i = 0; i < adapter->num_active_rx_pools; i++)
		adapter->rx_pool[i].active = 0;
}

static void replenish_rx_pool(struct ibmvnic_adapter *adapter,
			      struct ibmvnic_rx_pool *pool)
{
	int count = pool->size - atomic_read(&pool->available);
	u64 handle = adapter->rx_scrq[pool->index]->handle;
	struct device *dev = &adapter->vdev->dev;
	struct ibmvnic_ind_xmit_queue *ind_bufp;
	struct ibmvnic_sub_crq_queue *rx_scrq;
	union sub_crq *sub_crq;
	int buffers_added = 0;
	unsigned long lpar_rc;
	struct sk_buff *skb;
	unsigned int offset;
	dma_addr_t dma_addr;
	unsigned char *dst;
	int shift = 0;
	int index;
	int i;

	if (!pool->active)
		return;

	rx_scrq = adapter->rx_scrq[pool->index];
	ind_bufp = &rx_scrq->ind_buf;

	/* netdev_skb_alloc() could have failed after we saved a few skbs
	 * in the indir_buf and we would not have sent them to VIOS yet.
	 * To account for them, start the loop at ind_bufp->index rather
	 * than 0. If we pushed all the skbs to VIOS, ind_bufp->index will
	 * be 0.
	 */
	for (i = ind_bufp->index; i < count; ++i) {
		index = pool->free_map[pool->next_free];

		/* We maybe reusing the skb from earlier resets. Allocate
		 * only if necessary. But since the LTB may have changed
		 * during reset (see init_rx_pools()), update LTB below
		 * even if reusing skb.
		 */
		skb = pool->rx_buff[index].skb;
		if (!skb) {
			skb = netdev_alloc_skb(adapter->netdev,
					       pool->buff_size);
			if (!skb) {
				dev_err(dev, "Couldn't replenish rx buff\n");
				adapter->replenish_no_mem++;
				break;
			}
		}

		pool->free_map[pool->next_free] = IBMVNIC_INVALID_MAP;
		pool->next_free = (pool->next_free + 1) % pool->size;

		/* 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;

		/* add the skb to an rx_buff in the pool */
		pool->rx_buff[index].data = dst;
		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;

		/* queue the rx_buff for the next send_subcrq_indirect */
		sub_crq = &ind_bufp->indir_arr[ind_bufp->index++];
		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);

		/* if send_subcrq_indirect queue is full, flush to VIOS */
		if (ind_bufp->index == IBMVNIC_MAX_IND_DESCS ||
		    i == count - 1) {
			lpar_rc =
				send_subcrq_indirect(adapter, handle,
						     (u64)ind_bufp->indir_dma,
						     (u64)ind_bufp->index);
			if (lpar_rc != H_SUCCESS)
				goto failure;
			buffers_added += ind_bufp->index;
			adapter->replenish_add_buff_success += ind_bufp->index;
			ind_bufp->index = 0;
		}
	}
	atomic_add(buffers_added, &pool->available);
	return;

failure:
	if (lpar_rc != H_PARAMETER && lpar_rc != H_CLOSED)
		dev_err_ratelimited(dev, "rx: replenish packet buffer failed\n");
	for (i = ind_bufp->index - 1; i >= 0; --i) {
		struct ibmvnic_rx_buff *rx_buff;

		pool->next_free = pool->next_free == 0 ?
				  pool->size - 1 : pool->next_free - 1;
		sub_crq = &ind_bufp->indir_arr[i];
		rx_buff = (struct ibmvnic_rx_buff *)
				be64_to_cpu(sub_crq->rx_add.correlator);
		index = (int)(rx_buff - pool->rx_buff);
		pool->free_map[pool->next_free] = index;
		dev_kfree_skb_any(pool->rx_buff[index].skb);
		pool->rx_buff[index].skb = NULL;
	}
	adapter->replenish_add_buff_failure += ind_bufp->index;
	atomic_add(buffers_added, &pool->available);
	ind_bufp->index = 0;
	if (lpar_rc == H_CLOSED || adapter->failover_pending) {
		/* Disable buffer pool replenishment and report carrier off if
		 * queue is closed or pending failover.
		 * Firmware guarantees that a signal will be sent to the
		 * driver, triggering a reset.
		 */
		deactivate_rx_pools(adapter);
		netif_carrier_off(adapter->netdev);
	}
}

static void replenish_pools(struct ibmvnic_adapter *adapter)
{
	int i;

	adapter->replenish_task_cycles++;
	for (i = 0; i < adapter->num_active_rx_pools; i++) {
		if (adapter->rx_pool[i].active)
			replenish_rx_pool(adapter, &adapter->rx_pool[i]);
	}

	netdev_dbg(adapter->netdev, "Replenished %d pools\n", i);
}

static void release_stats_buffers(struct ibmvnic_adapter *adapter)
{
	kfree(adapter->tx_stats_buffers);
	kfree(adapter->rx_stats_buffers);
	adapter->tx_stats_buffers = NULL;
	adapter->rx_stats_buffers = NULL;
}

static int init_stats_buffers(struct ibmvnic_adapter *adapter)
{
	adapter->tx_stats_buffers =
				kcalloc(IBMVNIC_MAX_QUEUES,
					sizeof(struct ibmvnic_tx_queue_stats),
					GFP_KERNEL);
	if (!adapter->tx_stats_buffers)
		return -ENOMEM;

	adapter->rx_stats_buffers =
				kcalloc(IBMVNIC_MAX_QUEUES,
					sizeof(struct ibmvnic_rx_queue_stats),
					GFP_KERNEL);
	if (!adapter->rx_stats_buffers)
		return -ENOMEM;

	return 0;
}

static void release_stats_token(struct ibmvnic_adapter *adapter)
{
	struct device *dev = &adapter->vdev->dev;

	if (!adapter->stats_token)
		return;

	dma_unmap_single(dev, adapter->stats_token,
			 sizeof(struct ibmvnic_statistics),
			 DMA_FROM_DEVICE);
	adapter->stats_token = 0;
}

static int init_stats_token(struct ibmvnic_adapter *adapter)
{
	struct device *dev = &adapter->vdev->dev;
	dma_addr_t stok;

	stok = dma_map_single(dev, &adapter->stats,
			      sizeof(struct ibmvnic_statistics),
			      DMA_FROM_DEVICE);
	if (dma_mapping_error(dev, stok)) {
		dev_err(dev, "Couldn't map stats buffer\n");
		return -1;
	}

	adapter->stats_token = stok;
	netdev_dbg(adapter->netdev, "Stats token initialized (%llx)\n", stok);
	return 0;
}

/**
 * release_rx_pools() - Release any rx pools attached to @adapter.
 * @adapter: ibmvnic adapter
 *
 * Safe to call this multiple times - even if no pools are attached.
 */
static void release_rx_pools(struct ibmvnic_adapter *adapter)
{
	struct ibmvnic_rx_pool *rx_pool;
	int i, j;

	if (!adapter->rx_pool)
		return;

	for (i = 0; i < adapter->num_active_rx_pools; i++) {
		rx_pool = &adapter->rx_pool[i];

		netdev_dbg(adapter->netdev, "Releasing rx_pool[%d]\n", i);

		kfree(rx_pool->free_map);

		free_long_term_buff(adapter, &rx_pool->long_term_buff);

		if (!rx_pool->rx_buff)
			continue;

		for (j = 0; j < rx_pool->size; j++) {
			if (rx_pool->rx_buff[j].skb) {
				dev_kfree_skb_any(rx_pool->rx_buff[j].skb);
				rx_pool->rx_buff[j].skb = NULL;
			}
		}

		kfree(rx_pool->rx_buff);
	}

	kfree(adapter->rx_pool);
	adapter->rx_pool = NULL;
	adapter->num_active_rx_pools = 0;
	adapter->prev_rx_pool_size = 0;
}

/**
 * reuse_rx_pools() - Check if the existing rx pools can be reused.
 * @adapter: ibmvnic adapter
 *
 * Check if the existing rx pools in the adapter can be reused. The
 * pools can be reused if the pool parameters (number of pools,
 * number of buffers in the pool and size of each buffer) have not
 * changed.
 *
 * NOTE: This assumes that all pools have the same number of buffers
 *       which is the case currently. If that changes, we must fix this.
 *
 * Return: true if the rx pools can be reused, false otherwise.
 */
static bool reuse_rx_pools(struct ibmvnic_adapter *adapter)
{
	u64 old_num_pools, new_num_pools;
	u64 old_pool_size, new_pool_size;
	u64 old_buff_size, new_buff_size;

	if (!adapter->rx_pool)
		return false;

	old_num_pools = adapter->num_active_rx_pools;
	new_num_pools = adapter->req_rx_queues;

	old_pool_size = adapter->prev_rx_pool_size;
	new_pool_size = adapter->req_rx_add_entries_per_subcrq;

	old_buff_size = adapter->prev_rx_buf_sz;
	new_buff_size = adapter->cur_rx_buf_sz;

	/* Require buff size to be exactly same for now */
	if (old_buff_size != new_buff_size)
		return false;

	if (old_num_pools == new_num_pools && old_pool_size == new_pool_size)
		return true;

	if (old_num_pools < adapter->min_rx_queues ||
	    old_num_pools > adapter->max_rx_queues ||
	    old_pool_size < adapter->min_rx_add_entries_per_subcrq ||
	    old_pool_size > adapter->max_rx_add_entries_per_subcrq)
		return false;

	return true;
}

/**
 * init_rx_pools(): Initialize the set of receiver pools in the adapter.
 * @netdev: net device associated with the vnic interface
 *
 * Initialize the set of receiver pools in the ibmvnic adapter associated
 * with the net_device @netdev. If possible, reuse the existing rx pools.
 * Otherwise free any existing pools and  allocate a new set of pools
 * before initializing them.
 *
 * Return: 0 on success and negative value on error.
 */
static int init_rx_pools(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	struct device *dev = &adapter->vdev->dev;
	struct ibmvnic_rx_pool *rx_pool;
	u64 num_pools;
	u64 pool_size;		/* # of buffers in one pool */
	u64 buff_size;
	int i, j;

	pool_size = adapter->req_rx_add_entries_per_subcrq;
	num_pools = adapter->req_rx_queues;
	buff_size = adapter->cur_rx_buf_sz;

	if (reuse_rx_pools(adapter)) {
		dev_dbg(dev, "Reusing rx pools\n");
		goto update_ltb;
	}

	/* Allocate/populate the pools. */
	release_rx_pools(adapter);

	adapter->rx_pool = kcalloc(num_pools,
				   sizeof(struct ibmvnic_rx_pool),
				   GFP_KERNEL);
	if (!adapter->rx_pool) {
		dev_err(dev, "Failed to allocate rx pools\n");
		return -1;
	}

	/* Set num_active_rx_pools early. If we fail below after partial
	 * allocation, release_rx_pools() will know how many to look for.
	 */
	adapter->num_active_rx_pools = num_pools;

	for (i = 0; i < num_pools; i++) {
		rx_pool = &adapter->rx_pool[i];

		netdev_dbg(adapter->netdev,
			   "Initializing rx_pool[%d], %lld buffs, %lld bytes each\n",
			   i, pool_size, buff_size);

		rx_pool->size = pool_size;
		rx_pool->index = i;
		rx_pool->buff_size = ALIGN(buff_size, L1_CACHE_BYTES);

		rx_pool->free_map = kcalloc(rx_pool->size, sizeof(int),
					    GFP_KERNEL);
		if (!rx_pool->free_map) {
			dev_err(dev, "Couldn't alloc free_map %d\n", i);
			goto out_release;
		}

		rx_pool->rx_buff = kcalloc(rx_pool->size,
					   sizeof(struct ibmvnic_rx_buff),
					   GFP_KERNEL);
		if (!rx_pool->rx_buff) {
			dev_err(dev, "Couldn't alloc rx buffers\n");
			goto out_release;
		}
	}

	adapter->prev_rx_pool_size = pool_size;
	adapter->prev_rx_buf_sz = adapter->cur_rx_buf_sz;

update_ltb:
	for (i = 0; i < num_pools; i++) {
		rx_pool = &adapter->rx_pool[i];
		dev_dbg(dev, "Updating LTB for rx pool %d [%d, %d]\n",
			i, rx_pool->size, rx_pool->buff_size);

		if (alloc_long_term_buff(adapter, &rx_pool->long_term_buff,
					 rx_pool->size * rx_pool->buff_size))
			goto out;

		for (j = 0; j < rx_pool->size; ++j) {
			struct ibmvnic_rx_buff *rx_buff;

			rx_pool->free_map[j] = j;

			/* NOTE: Don't clear rx_buff->skb here - will leak
			 * memory! replenish_rx_pool() will reuse skbs or
			 * allocate as necessary.
			 */
			rx_buff = &rx_pool->rx_buff[j];
			rx_buff->dma = 0;
			rx_buff->data = 0;
			rx_buff->size = 0;
			rx_buff->pool_index = 0;
		}

		/* Mark pool "empty" so replenish_rx_pools() will
		 * update the LTB info for each buffer
		 */
		atomic_set(&rx_pool->available, 0);
		rx_pool->next_alloc = 0;
		rx_pool->next_free = 0;
		/* replenish_rx_pool() may have called deactivate_rx_pools()
		 * on failover. Ensure pool is active now.
		 */
		rx_pool->active = 1;
	}
	return 0;
out_release:
	release_rx_pools(adapter);
out:
	/* We failed to allocate one or more LTBs or map them on the VIOS.
	 * Hold onto the pools and any LTBs that we did allocate/map.
	 */
	return -1;
}

static void release_vpd_data(struct ibmvnic_adapter *adapter)
{
	if (!adapter->vpd)
		return;

	kfree(adapter->vpd->buff);
	kfree(adapter->vpd);

	adapter->vpd = NULL;
}

static void release_one_tx_pool(struct ibmvnic_adapter *adapter,
				struct ibmvnic_tx_pool *tx_pool)
{
	kfree(tx_pool->tx_buff);
	kfree(tx_pool->free_map);
	free_long_term_buff(adapter, &tx_pool->long_term_buff);
}

/**
 * release_tx_pools() - Release any tx pools attached to @adapter.
 * @adapter: ibmvnic adapter
 *
 * Safe to call this multiple times - even if no pools are attached.
 */
static void release_tx_pools(struct ibmvnic_adapter *adapter)
{
	int i;

	/* init_tx_pools() ensures that ->tx_pool and ->tso_pool are
	 * both NULL or both non-NULL. So we only need to check one.
	 */
	if (!adapter->tx_pool)
		return;

	for (i = 0; i < adapter->num_active_tx_pools; i++) {
		release_one_tx_pool(adapter, &adapter->tx_pool[i]);
		release_one_tx_pool(adapter, &adapter->tso_pool[i]);
	}

	kfree(adapter->tx_pool);
	adapter->tx_pool = NULL;
	kfree(adapter->tso_pool);
	adapter->tso_pool = NULL;
	adapter->num_active_tx_pools = 0;
	adapter->prev_tx_pool_size = 0;
}

static int init_one_tx_pool(struct net_device *netdev,
			    struct ibmvnic_tx_pool *tx_pool,
			    int pool_size, int buf_size)
{
	int i;

	tx_pool->tx_buff = kcalloc(pool_size,
				   sizeof(struct ibmvnic_tx_buff),
				   GFP_KERNEL);
	if (!tx_pool->tx_buff)
		return -1;

	tx_pool->free_map = kcalloc(pool_size, sizeof(int), GFP_KERNEL);
	if (!tx_pool->free_map) {
		kfree(tx_pool->tx_buff);
		tx_pool->tx_buff = NULL;
		return -1;
	}

	for (i = 0; i < pool_size; i++)
		tx_pool->free_map[i] = i;

	tx_pool->consumer_index = 0;
	tx_pool->producer_index = 0;
	tx_pool->num_buffers = pool_size;
	tx_pool->buf_size = buf_size;

	return 0;
}

/**
 * reuse_tx_pools() - Check if the existing tx pools can be reused.
 * @adapter: ibmvnic adapter
 *
 * Check if the existing tx pools in the adapter can be reused. The
 * pools can be reused if the pool parameters (number of pools,
 * number of buffers in the pool and mtu) have not changed.
 *
 * NOTE: This assumes that all pools have the same number of buffers
 *       which is the case currently. If that changes, we must fix this.
 *
 * Return: true if the tx pools can be reused, false otherwise.
 */
static bool reuse_tx_pools(struct ibmvnic_adapter *adapter)
{
	u64 old_num_pools, new_num_pools;
	u64 old_pool_size, new_pool_size;
	u64 old_mtu, new_mtu;

	if (!adapter->tx_pool)
		return false;

	old_num_pools = adapter->num_active_tx_pools;
	new_num_pools = adapter->num_active_tx_scrqs;
	old_pool_size = adapter->prev_tx_pool_size;
	new_pool_size = adapter->req_tx_entries_per_subcrq;
	old_mtu = adapter->prev_mtu;
	new_mtu = adapter->req_mtu;

	/* Require MTU to be exactly same to reuse pools for now */
	if (old_mtu != new_mtu)
		return false;

	if (old_num_pools == new_num_pools && old_pool_size == new_pool_size)
		return true;

	if (old_num_pools < adapter->min_tx_queues ||
	    old_num_pools > adapter->max_tx_queues ||
	    old_pool_size < adapter->min_tx_entries_per_subcrq ||
	    old_pool_size > adapter->max_tx_entries_per_subcrq)
		return false;

	return true;
}

/**
 * init_tx_pools(): Initialize the set of transmit pools in the adapter.
 * @netdev: net device associated with the vnic interface
 *
 * Initialize the set of transmit pools in the ibmvnic adapter associated
 * with the net_device @netdev. If possible, reuse the existing tx pools.
 * Otherwise free any existing pools and  allocate a new set of pools
 * before initializing them.
 *
 * Return: 0 on success and negative value on error.
 */
static int init_tx_pools(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	struct device *dev = &adapter->vdev->dev;
	int num_pools;
	u64 pool_size;		/* # of buffers in pool */
	u64 buff_size;
	int i, j, rc;

	num_pools = adapter->req_tx_queues;

	/* We must notify the VIOS about the LTB on all resets - but we only
	 * need to alloc/populate pools if either the number of buffers or
	 * size of each buffer in the pool has changed.
	 */
	if (reuse_tx_pools(adapter)) {
		netdev_dbg(netdev, "Reusing tx pools\n");
		goto update_ltb;
	}

	/* Allocate/populate the pools. */
	release_tx_pools(adapter);

	pool_size = adapter->req_tx_entries_per_subcrq;
	num_pools = adapter->num_active_tx_scrqs;

	adapter->tx_pool = kcalloc(num_pools,
				   sizeof(struct ibmvnic_tx_pool), GFP_KERNEL);
	if (!adapter->tx_pool)
		return -1;

	adapter->tso_pool = kcalloc(num_pools,
				    sizeof(struct ibmvnic_tx_pool), GFP_KERNEL);
	/* To simplify release_tx_pools() ensure that ->tx_pool and
	 * ->tso_pool are either both NULL or both non-NULL.
	 */
	if (!adapter->tso_pool) {
		kfree(adapter->tx_pool);
		adapter->tx_pool = NULL;
		return -1;
	}

	/* Set num_active_tx_pools early. If we fail below after partial
	 * allocation, release_tx_pools() will know how many to look for.
	 */
	adapter->num_active_tx_pools = num_pools;

	buff_size = adapter->req_mtu + VLAN_HLEN;
	buff_size = ALIGN(buff_size, L1_CACHE_BYTES);

	for (i = 0; i < num_pools; i++) {
		dev_dbg(dev, "Init tx pool %d [%llu, %llu]\n",
			i, adapter->req_tx_entries_per_subcrq, buff_size);

		rc = init_one_tx_pool(netdev, &adapter->tx_pool[i],
				      pool_size, buff_size);
		if (rc)
			goto out_release;

		rc = init_one_tx_pool(netdev, &adapter->tso_pool[i],
				      IBMVNIC_TSO_BUFS,
				      IBMVNIC_TSO_BUF_SZ);
		if (rc)
			goto out_release;
	}

	adapter->prev_tx_pool_size = pool_size;
	adapter->prev_mtu = adapter->req_mtu;

update_ltb:
	/* NOTE: All tx_pools have the same number of buffers (which is
	 *       same as pool_size). All tso_pools have IBMVNIC_TSO_BUFS
	 *       buffers (see calls init_one_tx_pool() for these).
	 *       For consistency, we use tx_pool->num_buffers and
	 *       tso_pool->num_buffers below.
	 */
	rc = -1;
	for (i = 0; i < num_pools; i++) {
		struct ibmvnic_tx_pool *tso_pool;
		struct ibmvnic_tx_pool *tx_pool;
		u32 ltb_size;

		tx_pool = &adapter->tx_pool[i];
		ltb_size = tx_pool->num_buffers * tx_pool->buf_size;
		if (alloc_long_term_buff(adapter, &tx_pool->long_term_buff,
					 ltb_size))
			goto out;

		dev_dbg(dev, "Updated LTB for tx pool %d [%p, %d, %d]\n",
			i, tx_pool->long_term_buff.buff,
			tx_pool->num_buffers, tx_pool->buf_size);

		tx_pool->consumer_index = 0;
		tx_pool->producer_index = 0;

		for (j = 0; j < tx_pool->num_buffers; j++)
			tx_pool->free_map[j] = j;

		tso_pool = &adapter->tso_pool[i];
		ltb_size = tso_pool->num_buffers * tso_pool->buf_size;
		if (alloc_long_term_buff(adapter, &tso_pool->long_term_buff,
					 ltb_size))
			goto out;

		dev_dbg(dev, "Updated LTB for tso pool %d [%p, %d, %d]\n",
			i, tso_pool->long_term_buff.buff,
			tso_pool->num_buffers, tso_pool->buf_size);

		tso_pool->consumer_index = 0;
		tso_pool->producer_index = 0;

		for (j = 0; j < tso_pool->num_buffers; j++)
			tso_pool->free_map[j] = j;
	}

	return 0;
out_release:
	release_tx_pools(adapter);
out:
	/* We failed to allocate one or more LTBs or map them on the VIOS.
	 * Hold onto the pools and any LTBs that we did allocate/map.
	 */
	return rc;
}

static void ibmvnic_napi_enable(struct ibmvnic_adapter *adapter)
{
	int i;

	if (adapter->napi_enabled)
		return;

	for (i = 0; i < adapter->req_rx_queues; i++)
		napi_enable(&adapter->napi[i]);

	adapter->napi_enabled = true;
}

static void ibmvnic_napi_disable(struct ibmvnic_adapter *adapter)
{
	int i;

	if (!adapter->napi_enabled)
		return;

	for (i = 0; i < adapter->req_rx_queues; i++) {
		netdev_dbg(adapter->netdev, "Disabling napi[%d]\n", i);
		napi_disable(&adapter->napi[i]);
	}

	adapter->napi_enabled = false;
}

static int init_napi(struct ibmvnic_adapter *adapter)
{
	int i;

	adapter->napi = kcalloc(adapter->req_rx_queues,
				sizeof(struct napi_struct), GFP_KERNEL);
	if (!adapter->napi)
		return -ENOMEM;

	for (i = 0; i < adapter->req_rx_queues; i++) {
		netdev_dbg(adapter->netdev, "Adding napi[%d]\n", i);
		netif_napi_add(adapter->netdev, &adapter->napi[i],
			       ibmvnic_poll, NAPI_POLL_WEIGHT);
	}

	adapter->num_active_rx_napi = adapter->req_rx_queues;
	return 0;
}

static void release_napi(struct ibmvnic_adapter *adapter)
{
	int i;

	if (!adapter->napi)
		return;

	for (i = 0; i < adapter->num_active_rx_napi; i++) {
		netdev_dbg(adapter->netdev, "Releasing napi[%d]\n", i);
		netif_napi_del(&adapter->napi[i]);
	}

	kfree(adapter->napi);
	adapter->napi = NULL;
	adapter->num_active_rx_napi = 0;
	adapter->napi_enabled = false;
}

static const char *adapter_state_to_string(enum vnic_state state)
{
	switch (state) {
	case VNIC_PROBING:
		return "PROBING";
	case VNIC_PROBED:
		return "PROBED";
	case VNIC_OPENING:
		return "OPENING";
	case VNIC_OPEN:
		return "OPEN";
	case VNIC_CLOSING:
		return "CLOSING";
	case VNIC_CLOSED:
		return "CLOSED";
	case VNIC_REMOVING:
		return "REMOVING";
	case VNIC_REMOVED:
		return "REMOVED";
	case VNIC_DOWN:
		return "DOWN";
	}
	return "UNKNOWN";
}

static int ibmvnic_login(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	unsigned long timeout = msecs_to_jiffies(20000);
	int retry_count = 0;
	int retries = 10;
	bool retry;
	int rc;

	do {
		retry = false;
		if (retry_count > retries) {
			netdev_warn(netdev, "Login attempts exceeded\n");
			return -1;
		}

		adapter->init_done_rc = 0;
		reinit_completion(&adapter->init_done);
		rc = send_login(adapter);
		if (rc)
			return rc;

		if (!wait_for_completion_timeout(&adapter->init_done,
						 timeout)) {
			netdev_warn(netdev, "Login timed out, retrying...\n");
			retry = true;
			adapter->init_done_rc = 0;
			retry_count++;
			continue;
		}

		if (adapter->init_done_rc == ABORTED) {
			netdev_warn(netdev, "Login aborted, retrying...\n");
			retry = true;
			adapter->init_done_rc = 0;
			retry_count++;
			/* FW or device may be busy, so
			 * wait a bit before retrying login
			 */
			msleep(500);
		} else if (adapter->init_done_rc == PARTIALSUCCESS) {
			retry_count++;
			release_sub_crqs(adapter, 1);

			retry = true;
			netdev_dbg(netdev,
				   "Received partial success, retrying...\n");
			adapter->init_done_rc = 0;
			reinit_completion(&adapter->init_done);
			send_query_cap(adapter);
			if (!wait_for_completion_timeout(&adapter->init_done,
							 timeout)) {
				netdev_warn(netdev,
					    "Capabilities query timed out\n");
				return -1;
			}

			rc = init_sub_crqs(adapter);
			if (rc) {
				netdev_warn(netdev,
					    "SCRQ initialization failed\n");
				return -1;
			}

			rc = init_sub_crq_irqs(adapter);
			if (rc) {
				netdev_warn(netdev,
					    "SCRQ irq initialization failed\n");
				return -1;
			}
		} else if (adapter->init_done_rc) {
			netdev_warn(netdev, "Adapter login failed\n");
			return -1;
		}
	} while (retry);

	__ibmvnic_set_mac(netdev, adapter->mac_addr);

	netdev_dbg(netdev, "[S:%s] Login succeeded\n", adapter_state_to_string(adapter->state));
	return 0;
}

static void release_login_buffer(struct ibmvnic_adapter *adapter)
{
	kfree(adapter->login_buf);
	adapter->login_buf = NULL;
}

static void release_login_rsp_buffer(struct ibmvnic_adapter *adapter)
{
	kfree(adapter->login_rsp_buf);
	adapter->login_rsp_buf = NULL;
}

static void release_resources(struct ibmvnic_adapter *adapter)
{
	release_vpd_data(adapter);

	release_napi(adapter);
	release_login_buffer(adapter);
	release_login_rsp_buffer(adapter);
}

static int set_link_state(struct ibmvnic_adapter *adapter, u8 link_state)
{
	struct net_device *netdev = adapter->netdev;
	unsigned long timeout = msecs_to_jiffies(20000);
	union ibmvnic_crq crq;
	bool resend;
	int rc;

	netdev_dbg(netdev, "setting link state %d\n", link_state);

	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 = link_state;

	do {
		resend = false;

		reinit_completion(&adapter->init_done);
		rc = ibmvnic_send_crq(adapter, &crq);
		if (rc) {
			netdev_err(netdev, "Failed to set link state\n");
			return rc;
		}

		if (!wait_for_completion_timeout(&adapter->init_done,
						 timeout)) {
			netdev_err(netdev, "timeout setting link state\n");
			return -1;
		}

		if (adapter->init_done_rc == PARTIALSUCCESS) {
			/* Partuial success, delay and re-send */
			mdelay(1000);
			resend = true;
		} else if (adapter->init_done_rc) {
			netdev_warn(netdev, "Unable to set link state, rc=%d\n",
				    adapter->init_done_rc);
			return adapter->init_done_rc;
		}
	} while (resend);

	return 0;
}

static int set_real_num_queues(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	int rc;

	netdev_dbg(netdev, "Setting real tx/rx queues (%llx/%llx)\n",
		   adapter->req_tx_queues, adapter->req_rx_queues);

	rc = netif_set_real_num_tx_queues(netdev, adapter->req_tx_queues);
	if (rc) {
		netdev_err(netdev, "failed to set the number of tx queues\n");
		return rc;
	}

	rc = netif_set_real_num_rx_queues(netdev, adapter->req_rx_queues);
	if (rc)
		netdev_err(netdev, "failed to set the number of rx queues\n");

	return rc;
}

static int ibmvnic_get_vpd(struct ibmvnic_adapter *adapter)
{
	struct device *dev = &adapter->vdev->dev;
	union ibmvnic_crq crq;
	int len = 0;
	int rc;

	if (adapter->vpd->buff)
		len = adapter->vpd->len;

	mutex_lock(&adapter->fw_lock);
	adapter->fw_done_rc = 0;
	reinit_completion(&adapter->fw_done);

	crq.get_vpd_size.first = IBMVNIC_CRQ_CMD;
	crq.get_vpd_size.cmd = GET_VPD_SIZE;
	rc = ibmvnic_send_crq(adapter, &crq);
	if (rc) {
		mutex_unlock(&adapter->fw_lock);
		return rc;
	}

	rc = ibmvnic_wait_for_completion(adapter, &adapter->fw_done, 10000);
	if (rc) {
		dev_err(dev, "Could not retrieve VPD size, rc = %d\n", rc);
		mutex_unlock(&adapter->fw_lock);
		return rc;
	}
	mutex_unlock(&adapter->fw_lock);

	if (!adapter->vpd->len)
		return -ENODATA;

	if (!adapter->vpd->buff)
		adapter->vpd->buff = kzalloc(adapter->vpd->len, GFP_KERNEL);
	else if (adapter->vpd->len != len)
		adapter->vpd->buff =
			krealloc(adapter->vpd->buff,
				 adapter->vpd->len, GFP_KERNEL);

	if (!adapter->vpd->buff) {
		dev_err(dev, "Could allocate VPD buffer\n");
		return -ENOMEM;
	}

	adapter->vpd->dma_addr =
		dma_map_single(dev, adapter->vpd->buff, adapter->vpd->len,
			       DMA_FROM_DEVICE);
	if (dma_mapping_error(dev, adapter->vpd->dma_addr)) {
		dev_err(dev, "Could not map VPD buffer\n");
		kfree(adapter->vpd->buff);
		adapter->vpd->buff = NULL;
		return -ENOMEM;
	}

	mutex_lock(&adapter->fw_lock);
	adapter->fw_done_rc = 0;
	reinit_completion(&adapter->fw_done);

	crq.get_vpd.first = IBMVNIC_CRQ_CMD;
	crq.get_vpd.cmd = GET_VPD;
	crq.get_vpd.ioba = cpu_to_be32(adapter->vpd->dma_addr);
	crq.get_vpd.len = cpu_to_be32((u32)adapter->vpd->len);
	rc = ibmvnic_send_crq(adapter, &crq);
	if (rc) {
		kfree(adapter->vpd->buff);
		adapter->vpd->buff = NULL;
		mutex_unlock(&adapter->fw_lock);
		return rc;
	}

	rc = ibmvnic_wait_for_completion(adapter, &adapter->fw_done, 10000);
	if (rc) {
		dev_err(dev, "Unable to retrieve VPD, rc = %d\n", rc);
		kfree(adapter->vpd->buff);
		adapter->vpd->buff = NULL;
		mutex_unlock(&adapter->fw_lock);
		return rc;
	}

	mutex_unlock(&adapter->fw_lock);
	return 0;
}

static int init_resources(struct ibmvnic_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	int rc;

	rc = set_real_num_queues(netdev);
	if (rc)
		return rc;

	adapter->vpd = kzalloc(sizeof(*adapter->vpd), GFP_KERNEL);
	if (!adapter->vpd)
		return -ENOMEM;

	/* Vital Product Data (VPD) */
	rc = ibmvnic_get_vpd(adapter);
	if (rc) {
		netdev_err(netdev, "failed to initialize Vital Product Data (VPD)\n");
		return rc;
	}

	rc = init_napi(adapter);
	if (rc)
		return rc;

	send_query_map(adapter);

	rc = init_rx_pools(netdev);
	if (rc)
		return rc;

	rc = init_tx_pools(netdev);
	return rc;
}

static int __ibmvnic_open(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	enum vnic_state prev_state = adapter->state;
	int i, rc;

	adapter->state = VNIC_OPENING;
	replenish_pools(adapter);
	ibmvnic_napi_enable(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++) {
		netdev_dbg(netdev, "Enabling rx_scrq[%d] irq\n", i);
		if (prev_state == VNIC_CLOSED)
			enable_irq(adapter->rx_scrq[i]->irq);
		enable_scrq_irq(adapter, adapter->rx_scrq[i]);
	}

	for (i = 0; i < adapter->req_tx_queues; i++) {
		netdev_dbg(netdev, "Enabling tx_scrq[%d] irq\n", i);
		if (prev_state == VNIC_CLOSED)
			enable_irq(adapter->tx_scrq[i]->irq);
		enable_scrq_irq(adapter, adapter->tx_scrq[i]);
		netdev_tx_reset_queue(netdev_get_tx_queue(netdev, i));
	}

	rc = set_link_state(adapter, IBMVNIC_LOGICAL_LNK_UP);
	if (rc) {
		ibmvnic_napi_disable(adapter);
		release_resources(adapter);
		return rc;
	}

	netif_tx_start_all_queues(netdev);

	if (prev_state == VNIC_CLOSED) {
		for (i = 0; i < adapter->req_rx_queues; i++)
			napi_schedule(&adapter->napi[i]);
	}

	adapter->state = VNIC_OPEN;
	return rc;
}

static int ibmvnic_open(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	int rc;

	ASSERT_RTNL();

	/* If device failover is pending or we are about to reset, just set
	 * device state and return. Device operation will be handled by reset
	 * routine.
	 *
	 * It should be safe to overwrite the adapter->state here. Since
	 * we hold the rtnl, either the reset has not actually started or
	 * the rtnl got dropped during the set_link_state() in do_reset().
	 * In the former case, no one else is changing the state (again we
	 * have the rtnl) and in the latter case, do_reset() will detect and
	 * honor our setting below.
	 */
	if (adapter->failover_pending || (test_bit(0, &adapter->resetting))) {
		netdev_dbg(netdev, "[S:%s FOP:%d] Resetting, deferring open\n",
			   adapter_state_to_string(adapter->state),
			   adapter->failover_pending);
		adapter->state = VNIC_OPEN;
		rc = 0;
		goto out;
	}

	if (adapter->state != VNIC_CLOSED) {
		rc = ibmvnic_login(netdev);
		if (rc)
			goto out;

		rc = init_resources(adapter);
		if (rc) {
			netdev_err(netdev, "failed to initialize resources\n");
			release_resources(adapter);
			release_rx_pools(adapter);
			release_tx_pools(adapter);
			goto out;
		}
	}

	rc = __ibmvnic_open(netdev);

out:
	/* If open failed and there is a pending failover or in-progress reset,
	 * set device state and return. Device operation will be handled by
	 * reset routine. See also comments above regarding rtnl.
	 */
	if (rc &&
	    (adapter->failover_pending || (test_bit(0, &adapter->resetting)))) {
		adapter->state = VNIC_OPEN;
		rc = 0;
	}
	return rc;
}

static void clean_rx_pools(struct ibmvnic_adapter *adapter)
{
	struct ibmvnic_rx_pool *rx_pool;
	struct ibmvnic_rx_buff *rx_buff;
	u64 rx_entries;
	int rx_scrqs;
	int i, j;

	if (!adapter->rx_pool)
		return;

	rx_scrqs = adapter->num_active_rx_pools;
	rx_entries = adapter->req_rx_add_entries_per_subcrq;

	/* Free any remaining skbs in the rx buffer pools */
	for (i = 0; i < rx_scrqs; i++) {
		rx_pool = &adapter->rx_pool[i];
		if (!rx_pool || !rx_pool->rx_buff)
			continue;

		netdev_dbg(adapter->netdev, "Cleaning rx_pool[%d]\n", i);
		for (j = 0; j < rx_entries; j++) {
			rx_buff = &rx_pool->rx_buff[j];
			if (rx_buff && rx_buff->skb) {
				dev_kfree_skb_any(rx_buff->skb);
				rx_buff->skb = NULL;
			}
		}
	}
}

static void clean_one_tx_pool(struct ibmvnic_adapter *adapter,
			      struct ibmvnic_tx_pool *tx_pool)
{
	struct ibmvnic_tx_buff *tx_buff;
	u64 tx_entries;
	int i;

	if (!tx_pool || !tx_pool->tx_buff)
		return;

	tx_entries = tx_pool->num_buffers;

	for (i = 0; i < tx_entries; i++) {
		tx_buff = &tx_pool->tx_buff[i];
		if (tx_buff && tx_buff->skb) {
			dev_kfree_skb_any(tx_buff->skb);
			tx_buff->skb = NULL;
		}
	}
}

static void clean_tx_pools(struct ibmvnic_adapter *adapter)
{
	int tx_scrqs;
	int i;

	if (!adapter->tx_pool || !adapter->tso_pool)
		return;

	tx_scrqs = adapter->num_active_tx_pools;

	/* Free any remaining skbs in the tx buffer pools */
	for (i = 0; i < tx_scrqs; i++) {
		netdev_dbg(adapter->netdev, "Cleaning tx_pool[%d]\n", i);
		clean_one_tx_pool(adapter, &adapter->tx_pool[i]);
		clean_one_tx_pool(adapter, &adapter->tso_pool[i]);
	}
}

static void ibmvnic_disable_irqs(struct ibmvnic_adapter *adapter)
{
	struct net_device *netdev = adapter->netdev;
	int i;

	if (adapter->tx_scrq) {
		for (i = 0; i < adapter->req_tx_queues; i++)
			if (adapter->tx_scrq[i]->irq) {
				netdev_dbg(netdev,
					   "Disabling tx_scrq[%d] irq\n", i);
				disable_scrq_irq(adapter, adapter->tx_scrq[i]);
				disable_irq(adapter->tx_scrq[i]->irq);
			}
	}

	if (adapter->rx_scrq) {
		for (i = 0; i < adapter->req_rx_queues; i++) {
			if (adapter->rx_scrq[i]->irq) {
				netdev_dbg(netdev,
					   "Disabling rx_scrq[%d] irq\n", i);
				disable_scrq_irq(adapter, adapter->rx_scrq[i]);
				disable_irq(adapter->rx_scrq[i]->irq);
			}
		}
	}
}

static void ibmvnic_cleanup(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);

	/* ensure that transmissions are stopped if called by do_reset */
	if (test_bit(0, &adapter->resetting))
		netif_tx_disable(netdev);
	else
		netif_tx_stop_all_queues(netdev);

	ibmvnic_napi_disable(adapter);
	ibmvnic_disable_irqs(adapter);
}

static int __ibmvnic_close(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	int rc = 0;

	adapter->state = VNIC_CLOSING;
	rc = set_link_state(adapter, IBMVNIC_LOGICAL_LNK_DN);
	adapter->state = VNIC_CLOSED;
	return rc;
}

static int ibmvnic_close(struct net_device *netdev)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	int rc;

	netdev_dbg(netdev, "[S:%s FOP:%d FRR:%d] Closing\n",
		   adapter_state_to_string(adapter->state),
		   adapter->failover_pending,
		   adapter->force_reset_recovery);

	/* If device failover is pending, just set device state and return.
	 * Device operation will be handled by reset routine.
	 */
	if (adapter->failover_pending) {
		adapter->state = VNIC_CLOSED;
		return 0;
	}

	rc = __ibmvnic_close(netdev);
	ibmvnic_cleanup(netdev);
	clean_rx_pools(adapter);
	clean_tx_pools(adapter);

	return rc;
}

/**
 * build_hdr_data - creates L2/L3/L4 header data buffer
 * @hdr_field: bitfield determining needed headers
 * @skb: socket buffer
 * @hdr_len: array of header lengths
 * @hdr_data: buffer to write the header to
 *
 * Reads hdr_field to determine which headers are needed by firmware.
 * Builds a buffer containing these headers.  Saves individual header
 * lengths and total buffer length to be used to build descriptors.
 */
static int build_hdr_data(u8 hdr_field, struct sk_buff *skb,
			  int *hdr_len, u8 *hdr_data)
{
	int len = 0;
	u8 *hdr;

	if (skb_vlan_tagged(skb) && !skb_vlan_tag_present(skb))
		hdr_len[0] = sizeof(struct vlan_ethhdr);
	else
		hdr_len[0] = sizeof(struct ethhdr);

	if (skb->protocol == htons(ETH_P_IP)) {
		hdr_len[1] = ip_hdr(skb)->ihl * 4;
		if (ip_hdr(skb)->protocol == IPPROTO_TCP)
			hdr_len[2] = tcp_hdrlen(skb);
		else if (ip_hdr(skb)->protocol == IPPROTO_UDP)
			hdr_len[2] = sizeof(struct udphdr);
	} else if (skb->protocol == htons(ETH_P_IPV6)) {
		hdr_len[1] = sizeof(struct ipv6hdr);
		if (ipv6_hdr(skb)->nexthdr == IPPROTO_TCP)
			hdr_len[2] = tcp_hdrlen(skb);
		else if (ipv6_hdr(skb)->nexthdr == IPPROTO_UDP)
			hdr_len[2] = sizeof(struct udphdr);
	} else if (skb->protocol == htons(ETH_P_ARP)) {
		hdr_len[1] = arp_hdr_len(skb->dev);
		hdr_len[2] = 0;
	}

	memset(hdr_data, 0, 120);
	if ((hdr_field >> 6) & 1) {
		hdr = skb_mac_header(skb);
		memcpy(hdr_data, hdr, hdr_len[0]);
		len += hdr_len[0];
	}

	if ((hdr_field >> 5) & 1) {
		hdr = skb_network_header(skb);
		memcpy(hdr_data + len, hdr, hdr_len[1]);
		len += hdr_len[1];
	}

	if ((hdr_field >> 4) & 1) {
		hdr = skb_transport_header(skb);
		memcpy(hdr_data + len, hdr, hdr_len[2]);
		len += hdr_len[2];
	}
	return len;
}

/**
 * create_hdr_descs - create header and header extension descriptors
 * @hdr_field: bitfield determining needed headers
 * @hdr_data: buffer containing header data
 * @len: length of data buffer
 * @hdr_len: array of individual header lengths
 * @scrq_arr: descriptor array
 *
 * Creates header and, if needed, header extension descriptors and
 * places them in a descriptor array, scrq_arr
 */

static int create_hdr_descs(u8 hdr_field, u8 *hdr_data, int len, int *hdr_len,
			    union sub_crq *scrq_arr)
{
	union sub_crq hdr_desc;
	int tmp_len = len;
	int num_descs = 0;
	u8 *data, *cur;
	int tmp;

	while (tmp_len > 0) {
		cur = hdr_data + len - tmp_len;

		memset(&hdr_desc, 0, sizeof(hdr_desc));
		if (cur != hdr_data) {
			data = hdr_desc.hdr_ext.data;
			tmp = tmp_len > 29 ? 29 : tmp_len;
			hdr_desc.hdr_ext.first = IBMVNIC_CRQ_CMD;
			hdr_desc.hdr_ext.type = IBMVNIC_HDR_EXT_DESC;
			hdr_desc.hdr_ext.len = tmp;
		} else {
			data = hdr_desc.hdr.data;
			tmp = tmp_len > 24 ? 24 : tmp_len;
			hdr_desc.hdr.first = IBMVNIC_CRQ_CMD;
			hdr_desc.hdr.type = IBMVNIC_HDR_DESC;
			hdr_desc.hdr.len = tmp;
			hdr_desc.hdr.l2_len = (u8)hdr_len[0];
			hdr_desc.hdr.l3_len = cpu_to_be16((u16)hdr_len[1]);
			hdr_desc.hdr.l4_len = (u8)hdr_len[2];
			hdr_desc.hdr.flag = hdr_field << 1;
		}
		memcpy(data, cur, tmp);
		tmp_len -= tmp;
		*scrq_arr = hdr_desc;
		scrq_arr++;
		num_descs++;
	}

	return num_descs;
}

/**
 * build_hdr_descs_arr - build a header descriptor array
 * @skb: tx socket buffer
 * @indir_arr: indirect array
 * @num_entries: number of descriptors to be sent
 * @hdr_field: bit field determining which headers will be sent
 *
 * This function will build a TX descriptor array with applicable
 * L2/L3/L4 packet header descriptors to be sent by send_subcrq_indirect.
 */

static void build_hdr_descs_arr(struct sk_buff *skb,
				union sub_crq *indir_arr,
				int *num_entries, u8 hdr_field)
{
	int hdr_len[3] = {0, 0, 0};
	u8 hdr_data[140] = {0};
	int tot_len;

	tot_len = build_hdr_data(hdr_field, skb, hdr_len,
				 hdr_data);
	*num_entries += create_hdr_descs(hdr_field, hdr_data, tot_len, hdr_len,
					 indir_arr + 1);
}

static int ibmvnic_xmit_workarounds(struct sk_buff *skb,
				    struct net_device *netdev)
{
	/* For some backing devices, mishandling of small packets
	 * can result in a loss of connection or TX stall. Device
	 * architects recommend that no packet should be smaller
	 * than the minimum MTU value provided to the driver, so
	 * pad any packets to that length
	 */
	if (skb->len < netdev->min_mtu)
		return skb_put_padto(skb, netdev->min_mtu);

	return 0;
}

static void ibmvnic_tx_scrq_clean_buffer(struct ibmvnic_adapter *adapter,
					 struct ibmvnic_sub_crq_queue *tx_scrq)
{
	struct ibmvnic_ind_xmit_queue *ind_bufp;
	struct ibmvnic_tx_buff *tx_buff;
	struct ibmvnic_tx_pool *tx_pool;
	union sub_crq tx_scrq_entry;
	int queue_num;
	int entries;
	int index;
	int i;

	ind_bufp = &tx_scrq->ind_buf;
	entries = (u64)ind_bufp->index;
	queue_num = tx_scrq->pool_index;

	for (i = entries - 1; i >= 0; --i) {
		tx_scrq_entry = ind_bufp->indir_arr[i];
		if (tx_scrq_entry.v1.type != IBMVNIC_TX_DESC)
			continue;
		index = be32_to_cpu(tx_scrq_entry.v1.correlator);
		if (index & IBMVNIC_TSO_POOL_MASK) {
			tx_pool = &adapter->tso_pool[queue_num];
			index &= ~IBMVNIC_TSO_POOL_MASK;
		} else {
			tx_pool = &adapter->tx_pool[queue_num];
		}
		tx_pool->free_map[tx_pool->consumer_index] = index;
		tx_pool->consumer_index = tx_pool->consumer_index == 0 ?
					  tx_pool->num_buffers - 1 :
					  tx_pool->consumer_index - 1;
		tx_buff = &tx_pool->tx_buff[index];
		adapter->netdev->stats.tx_packets--;
		adapter->netdev->stats.tx_bytes -= tx_buff->skb->len;
		adapter->tx_stats_buffers[queue_num].packets--;
		adapter->tx_stats_buffers[queue_num].bytes -=
						tx_buff->skb->len;
		dev_kfree_skb_any(tx_buff->skb);
		tx_buff->skb = NULL;
		adapter->netdev->stats.tx_dropped++;
	}
	ind_bufp->index = 0;
	if (atomic_sub_return(entries, &tx_scrq->used) <=
	    (adapter->req_tx_entries_per_subcrq / 2) &&
	    __netif_subqueue_stopped(adapter->netdev, queue_num) &&
	    !test_bit(0, &adapter->resetting)) {
		netif_wake_subqueue(adapter->netdev, queue_num);
		netdev_dbg(adapter->netdev, "Started queue %d\n",
			   queue_num);
	}
}

static int ibmvnic_tx_scrq_flush(struct ibmvnic_adapter *adapter,
				 struct ibmvnic_sub_crq_queue *tx_scrq)
{
	struct ibmvnic_ind_xmit_queue *ind_bufp;
	u64 dma_addr;
	u64 entries;
	u64 handle;
	int rc;

	ind_bufp = &tx_scrq->ind_buf;
	dma_addr = (u64)ind_bufp->indir_dma;
	entries = (u64)ind_bufp->index;
	handle = tx_scrq->handle;

	if (!entries)
		return 0;
	rc = send_subcrq_indirect(adapter, handle, dma_addr, entries);
	if (rc)
		ibmvnic_tx_scrq_clean_buffer(adapter, tx_scrq);
	else
		ind_bufp->index = 0;
	return 0;
}

static netdev_tx_t 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);
	u8 *hdrs = (u8 *)&adapter->tx_rx_desc_req;
	struct device *dev = &adapter->vdev->dev;
	struct ibmvnic_ind_xmit_queue *ind_bufp;
	struct ibmvnic_tx_buff *tx_buff = NULL;
	struct ibmvnic_sub_crq_queue *tx_scrq;
	struct ibmvnic_tx_pool *tx_pool;
	unsigned int tx_send_failed = 0;
	netdev_tx_t ret = NETDEV_TX_OK;
	unsigned int tx_map_failed = 0;
	union sub_crq indir_arr[16];
	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;
	unsigned long lpar_rc;
	union sub_crq tx_crq;
	unsigned int offset;
	int num_entries = 1;
	unsigned char *dst;
	int index = 0;
	u8 proto = 0;

	tx_scrq = adapter->tx_scrq[queue_num];
	txq = netdev_get_tx_queue(netdev, queue_num);
	ind_bufp = &tx_scrq->ind_buf;

	if (test_bit(0, &adapter->resetting)) {
		dev_kfree_skb_any(skb);

		tx_send_failed++;
		tx_dropped++;
		ret = NETDEV_TX_OK;
		goto out;
	}

	if (ibmvnic_xmit_workarounds(skb, netdev)) {
		tx_dropped++;
		tx_send_failed++;
		ret = NETDEV_TX_OK;
		ibmvnic_tx_scrq_flush(adapter, tx_scrq);
		goto out;
	}
	if (skb_is_gso(skb))
		tx_pool = &adapter->tso_pool[queue_num];
	else
		tx_pool = &adapter->tx_pool[queue_num];

	index = tx_pool->free_map[tx_pool->consumer_index];

	if (index == IBMVNIC_INVALID_MAP) {
		dev_kfree_skb_any(skb);
		tx_send_failed++;
		tx_dropped++;
		ibmvnic_tx_scrq_flush(adapter, tx_scrq);
		ret = NETDEV_TX_OK;
		goto out;
	}

	tx_pool->free_map[tx_pool->consumer_index] = IBMVNIC_INVALID_MAP;

	offset = index * tx_pool->buf_size;
	dst = tx_pool->long_term_buff.buff + offset;
	memset(dst, 0, tx_pool->buf_size);
	data_dma_addr = tx_pool->long_term_buff.addr + offset;

	if (skb_shinfo(skb)->nr_frags) {
		int cur, i;

		/* Copy the head */
		skb_copy_from_linear_data(skb, dst, skb_headlen(skb));
		cur = skb_headlen(skb);

		/* Copy the frags */
		for (i = 0; i < skb_shinfo(skb)->nr_frags; i++) {
			const skb_frag_t *frag = &skb_shinfo(skb)->frags[i];

			memcpy(dst + cur, skb_frag_address(frag),
			       skb_frag_size(frag));
			cur += skb_frag_size(frag);
		}
	} else {
		skb_copy_from_linear_data(skb, dst, skb->len);
	}

	/* post changes to long_term_buff *dst before VIOS accessing it */
	dma_wmb();

	tx_pool->consumer_index =
	    (tx_pool->consumer_index + 1) % tx_pool->num_buffers;

	tx_buff = &tx_pool->tx_buff[index];
	tx_buff->skb = skb;
	tx_buff->index = index;
	tx_buff->pool_index = queue_num;

	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;

	if (skb_is_gso(skb))
		tx_crq.v1.correlator =
			cpu_to_be32(index | IBMVNIC_TSO_POOL_MASK);
	else
		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 && skb_vlan_tag_present(skb)) {
		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)) {
		tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_IPV4;
		proto = ip_hdr(skb)->protocol;
	} else if (skb->protocol == htons(ETH_P_IPV6)) {
		tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_IPV6;
		proto = ipv6_hdr(skb)->nexthdr;
	}

	if (proto == IPPROTO_TCP)
		tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_TCP;
	else if (proto == IPPROTO_UDP)
		tx_crq.v1.flags1 |= IBMVNIC_TX_PROT_UDP;

	if (skb->ip_summed == CHECKSUM_PARTIAL) {
		tx_crq.v1.flags1 |= IBMVNIC_TX_CHKSUM_OFFLOAD;
		hdrs += 2;
	}
	if (skb_is_gso(skb)) {
		tx_crq.v1.flags1 |= IBMVNIC_TX_LSO;
		tx_crq.v1.mss = cpu_to_be16(skb_shinfo(skb)->gso_size);
		hdrs += 2;
	}

	if ((*hdrs >> 7) & 1)
		build_hdr_descs_arr(skb, indir_arr, &num_entries, *hdrs);

	tx_crq.v1.n_crq_elem = num_entries;
	tx_buff->num_entries = num_entries;
	/* flush buffer if current entry can not fit */
	if (num_entries + ind_bufp->index > IBMVNIC_MAX_IND_DESCS) {
		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq);
		if (lpar_rc != H_SUCCESS)
			goto tx_flush_err;
	}

	indir_arr[0] = tx_crq;
	memcpy(&ind_bufp->indir_arr[ind_bufp->index], &indir_arr[0],
	       num_entries * sizeof(struct ibmvnic_generic_scrq));
	ind_bufp->index += num_entries;
	if (__netdev_tx_sent_queue(txq, skb->len,
				   netdev_xmit_more() &&
				   ind_bufp->index < IBMVNIC_MAX_IND_DESCS)) {
		lpar_rc = ibmvnic_tx_scrq_flush(adapter, tx_scrq);
		if (lpar_rc != H_SUCCESS)
			goto tx_err;
	}

	if (atomic_add_return(num_entries, &tx_scrq->used)
					>= adapter->req_tx_entries_per_subcrq) {
		netdev_dbg(netdev, "Stopping queue %d\n", queue_num);
		netif_stop_subqueue(netdev, queue_num);
	}

	tx_packets++;
	tx_bytes += skb->len;
	txq->trans_start = jiffies;
	ret = NETDEV_TX_OK;
	goto out;

tx_flush_err:
	dev_kfree_skb_any(skb);
	tx_buff->skb = NULL;
	tx_pool->consumer_index = tx_pool->consumer_index == 0 ?
				  tx_pool->num_buffers - 1 :
				  tx_pool->consumer_index - 1;
	tx_dropped++;
tx_err:
	if (lpar_rc != H_CLOSED && lpar_rc != H_PARAMETER)
		dev_err_ratelimited(dev, "tx: send failed\n");

	if (lpar_rc == H_CLOSED || adapter->failover_pending) {
		/* Disable TX and report carrier off if queue is closed
		 * or pending failover.
		 * Firmware guarantees that a signal will be sent to the
		 * driver, triggering a reset or some other action.
		 */
		netif_tx_stop_all_queues(netdev);
		netif_carrier_off(netdev);
	}
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;
	adapter->tx_stats_buffers[queue_num].packets += tx_packets;
	adapter->tx_stats_buffers[queue_num].bytes += tx_bytes;
	adapter->tx_stats_buffers[queue_num].dropped_packets += tx_dropped;

	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, u8 *dev_addr)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	union ibmvnic_crq crq;
	int rc;

	if (!is_valid_ether_addr(dev_addr)) {
		rc = -EADDRNOTAVAIL;
		goto err;
	}

	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], dev_addr);

	mutex_lock(&adapter->fw_lock);
	adapter->fw_done_rc = 0;
	reinit_completion(&adapter->fw_done);

	rc = ibmvnic_send_crq(adapter, &crq);
	if (rc) {
		rc = -EIO;
		mutex_unlock(&adapter->fw_lock);
		goto err;
	}

	rc = ibmvnic_wait_for_completion(adapter, &adapter->fw_done, 10000);
	/* netdev->dev_addr is changed in handle_change_mac_rsp function */
	if (rc || adapter->fw_done_rc) {
		rc = -EIO;
		mutex_unlock(&adapter->fw_lock);
		goto err;
	}
	mutex_unlock(&adapter->fw_lock);
	return 0;
err:
	ether_addr_copy(adapter->mac_addr, netdev->dev_addr);
	return rc;
}

static int ibmvnic_set_mac(struct net_device *netdev, void *p)
{
	struct ibmvnic_adapter *adapter = netdev_priv(netdev);
	struct sockaddr *addr = p;
	int rc;

	rc = 0;
	if (!is_valid_ether_addr(addr->sa_data))
		return -EADDRNOTAVAIL;

	ether_addr_copy(adapter->mac_addr, addr->sa_data);
	if (adapter->state != VNIC_PROBED)
		rc = __ibmvnic_set_mac(netdev, addr->sa_data);

	return rc;
}

static const char *reset_reason_to_string(enum ibmvnic_reset_reason reason)
{
	switch (reason) {
	case VNIC_RESET_FAILOVER:
		return "FAILOVER";
	case VNIC_RESET_MOBILITY:
		return "MOBILITY";
	case VNIC_RESET_FATAL:
		return "FATAL";
	case VNIC_RESET_NON_FATAL:
		return "NON_FATAL";
	case VNIC_RESET_TIMEOUT:
		return "TIMEOUT";
	case VNIC_RESET_CHANGE_PARAM:
		return "CHANGE_PARAM";
	case VNIC_RESET_PASSIVE_INIT:
		return "PASSIVE_INIT";
	}
	return "UNKNOWN";
}

/*
 * do_reset returns zero if we are able to keep processing reset events, or
 * non-zero if we hit a fatal error and must halt.
 */
static int do_reset(struct ibmvnic_adapter *adapter,
		    struct ibmvnic_rwi *rwi, u32 reset_state)
{
	struct net_device *netdev = adapter->netdev;
	u64 old_num_rx_queues, old_num_tx_queues;
	u64 old_num_rx_slots, old_num_tx_slots;
	int rc;

	netdev_dbg(adapter->netdev,
		   "[S:%s FOP:%d] Reset reason: %s, reset_state: %s\n",
		   adapter_state_to_string(adapter->state),
		   adapter->failover_pending,
		   reset_reason_to_string(rwi->reset_reason),
		   adapter_state_to_string(reset_state));

	adapter->reset_reason = rwi->reset_reason;
	/* requestor of VNIC_RESET_CHANGE_PARAM already has the rtnl lock */
	if (!(adapter->reset_reason == VNIC_RESET_CHANGE_PARAM))
		rtnl_lock();

	/* Now that we have the rtnl lock, clear any pending failover.
	 * This will ensure ibmvnic_open() has either completed or will
	 * block until failover is complete.
	 */
	if (rwi->reset_reason == VNIC_RESET_FAILOVER)
		adapter->failover_pending = false;

	/* read the state and check (again) after getting rtnl */
	reset_state = adapter->state;

	if (reset_state == VNIC_REMOVING || reset_state == VNIC_REMOVED) {
		rc = -EBUSY;
		goto out;
	}

	netif_carrier_off(netdev);

	old_num_rx_queues = adapter->req_rx_queues;
	old_num_tx_queues = adapter->req_tx_queues;
	old_num_rx_slots = adapter->req_rx_add_entries_per_subcrq;
	old_num_tx_slots = adapter->req_tx_entries_per_subcrq;

	ibmvnic_cleanup(netdev);

	if (reset_state == VNIC_OPEN &&
	    adapter->reset_reason != VNIC_RESET_MOBILITY &&
	    adapter->reset_reason != VNIC_RESET_FAILOVER) {
		if (adapter->reset_reason == VNIC_RESET_CHANGE_PARAM) {
			rc = __ibmvnic_close(netdev);
			if (rc)
				goto out;
		} else {
			adapter->state = VNIC_CLOSING;

			/* Release the RTNL lock before link state change and
			 * re-acquire after the link state change to allow
			 * linkwatch_event to grab the RTNL lock and run during
			 * a reset.
			 */
			rtnl_unlock();
			rc = set_link_state(adapter, IBMVNIC_LOGICAL_LNK_DN);
			rtnl_lock();
			if (rc)
				goto out;

			if (adapter->state == VNIC_OPEN) {
				/* When we dropped rtnl, ibmvnic_open() got
				 * it and noticed that we are resetting and
				 * set the adapter state to OPEN. Update our
				 * new "target" state, and resume the reset
				 * from VNIC_CLOSING state.
				 */
				netdev_dbg(netdev,
					   "Open changed state from %s, updating.\n",
					   adapter_state_to_string(reset_state));
				reset_state = VNIC_OPEN;
				adapter->state = VNIC_CLOSING;
			}

			if (adapter->state != VNIC_CLOSING) {
				/* If someone else changed the adapter state
				 * when we dropped the rtnl, fail the reset
				 */
				rc = -1;
				goto out;
			}
			adapter->state = VNIC_CLOSED;
		}
	}

	if (adapter->reset_reason == VNIC_RESET_CHANGE_PARAM) {
		release_resources(adapter);
		release_sub_crqs(adapter, 1);
		release_crq_queue(adapter);
	}

	if (adapter->reset_reason != VNIC_RESET_NON_FATAL) {
		/* remove the closed state so when we call open it appears
		 * we are coming from the probed state.
		 */
		adapter->state = VNIC_PROBED;

		if (adapter->reset_reason == VNIC_RESET_CHANGE_PARAM) {
			rc = init_crq_queue(adapter);
		} else if (adapter->reset_reason == VNIC_RESET_MOBILITY) {
			rc = ibmvnic_reenable_crq_queue(adapter);
			release_sub_crqs(adapter, 1);
		} else {
			rc = ibmvnic_reset_crq(adapter);
			if (rc == H_CLOSED || rc == H_SUCCESS) {
				rc = vio_enable_interrupts(adapter->vdev);
				if (rc)
					netdev_err(adapter->netdev,
						   "Reset failed to enable interrupts. rc=%d\n",
						   rc);
			}
		}

		if (rc) {
			netdev_err(adapter->netdev,
				   "Reset couldn't initialize crq. rc=%d\n", rc);
			goto out;
		}

		rc = ibmvnic_reset_init(adapter, true);
		if (rc) {
			rc = IBMVNIC_INIT_FAILED;
			goto out;
		}

		/* If the adapter was in PROBE or DOWN state prior to the reset,
		 * exit here.
		 */
		if (reset_state == VNIC_PROBED || reset_state == VNIC_DOWN) {
			rc = 0;
			goto out;
		}

		rc = ibmvnic_login(netdev);
		if (rc)
			goto out;

		if (adapter->reset_reason == VNIC_RESET_CHANGE_PARAM) {
			rc = init_resources(adapter);
			if (rc)
				goto out;
		} else if (adapter->req_rx_queues != old_num_rx_queues ||
		    adapter->req_tx_queues != old_num_tx_queues ||
		    adapter->req_rx_add_entries_per_subcrq !=
		    old_num_rx_slots ||
		    adapter->req_tx_entries_per_subcrq !=
		    old_num_tx_slots ||
		    !adapter->rx_pool ||
		    !adapter->tso_pool ||
		    !adapter->tx_pool) {
			release_napi(adapter);
			release_vpd_data(adapter);

			rc = init_resources(adapter);
			if (rc)
				goto out;

		} else {
			rc = init_tx_pools(netdev);
			if (rc) {
				netdev_dbg(netdev,
					   "init tx pools failed (%d)\n",
					   rc);
				goto out;
			}

			rc = init_rx_pools(netdev);
			if (rc) {
				netdev_dbg(netdev,
					   "init rx pools failed (%d)\n",
					   rc);
				goto out;
			}
		}
		ibmvnic_disable_irqs(adapter);
	}
	adapter->state = VNIC_CLOSED;

	if (reset_state == VNIC_CLOSED) {
		rc = 0;
		goto out;
	}

	rc = __ibmvnic_open(netdev);
	if (rc) {
		rc = IBMVNIC_OPEN_FAILED;
		goto out;
	}

	/* refresh device's multicast list */
	ibmvnic_set_multi(netdev);

	if (adapter->reset_reason == VNIC_RESET_FAILOVER ||
	    adapter->reset_reason == VNIC_RESET_MOBILITY)
		__netdev_notify_peers(netdev);

	rc = 0;

out:
	/* restore the adapter state if reset failed */
	if (rc)
		adapter->state = reset_state;
	/* requestor of VNIC_RESET_CHANGE_PARAM should still hold the rtnl lock */
	if (!(adapter->reset_reason == VNIC_RESET_CHANGE_PARAM))
		rtnl_unlock();

	netdev_dbg(adapter->netdev, "[S:%s FOP:%d] Reset done, rc %d\n",
		   adapter_state_to_string(adapter->state),
		   adapter->failover_pending, rc);
	return rc;
}

static int do_hard_reset(struct ibmvnic_adapter *adapter,
			 struct ibmvnic_rwi *rwi, u32 reset_state)
{
	struct net_device *netdev = adapter->netdev;
	int rc;

	netdev_dbg(adapter->netdev, "Hard resetting driver (%s)\n",
		   reset_reason_to_string(rwi->reset_reason));

	/* read the state and check (again) after getting rtnl */
	reset_state = adapter->state;

	if (reset_state == VNIC_REMOVING || reset_state == VNIC_REMOVED) {
		rc = -EBUSY;
		goto out;
	}

	netif_carrier_off(netdev);
	adapter->reset_reason = rwi->reset_reason;

	ibmvnic_cleanup(netdev);
	release_resources(adapter);
	release_sub_crqs(adapter, 0);
	release_crq_queue(adapter);

	/* remove the closed state so when we call open it appears
	 * we are coming from the probed state.
	 */
	adapter->state = VNIC_PROBED;

	reinit_completion(&adapter->init_done);
	rc = init_crq_queue(adapter);
	if (rc) {
		netdev_err(adapter->netdev,
			   "Couldn't initialize crq. rc=%d\n", rc);
		goto out;
	}

	rc = ibmvnic_reset_init(adapter, false);
	if (rc)
		goto out;

	/* If the adapter was in PROBE or DOWN state prior to the reset,
	 * exit here.
	 */
	if (reset_state == VNIC_PROBED || reset_state == VNIC_DOWN)
		goto out;

	rc = ibmvnic_login(netdev);
	if (rc)
		goto out;

	rc = init_resources(adapter);
	if (rc)
		goto out;

	ibmvnic_disable_irqs(adapter);
	adapter->state = VNIC_CLOSED;

	if (reset_state == VNIC_CLOSED)
		goto out;

	rc = __ibmvnic_open(netdev);
	if (rc) {
		rc = IBMVNIC_OPEN_FAILED;
		goto out;
	}

	__netdev_notify_peers(netdev);
out:
	/* restore adapter state if reset failed */
	if (rc)
		adapter->state = reset_state;
	netdev_dbg(adapter->netdev, "[S:%s FOP:%d] Hard reset done, rc %d\n",
		   adapter_state_to_string(adapter->state),
		   adapter->failover_pending, rc);
	return rc;
}

static struct ibmvnic_rwi *get_next_rwi(struct ibmvnic_adapter *adapter)
{
	struct ibmvnic_rwi *rwi;
	unsigned long flags;

	spin_lock_irqsave(&adapter->rwi_lock, flags);

	if (!list_empty(&adapter->rwi_list)) {
		rwi = list_first_entry(&adapter->rwi_list, struct ibmvnic_rwi,
				       list);
		list_del(&rwi->list);
	} else {
		rwi = NULL;
	}

	spin_unlock_irqrestore(&adapter->rwi_lock, flags);
	return rwi;
}

/**
 * do_passive_init - complete probing when partner device is detected.
 * @adapter: ibmvnic_adapter struct
 *
 * If the ibmvnic device does not have a partner device to communicate with at boot
 * and that partner device comes online at a later time, this function is called
 * to complete the initialization process of ibmvnic device.
 * Caller is expected to hold rtnl_lock().
 *
 * Returns non-zero if sub-CRQs are not initialized properly leaving the device
 * in the down state.
 * Returns 0 upon success and the device is in PROBED state.
 */

static int do_passive_init(struct ibmvnic_adapter *adapter)
{
	unsigned long timeout = msecs_to_jiffies(30000);
	struct net_device *netdev = adapter->netdev;
	struct device *dev = &adapter->vdev->dev;
	int rc;

	netdev_dbg(netdev, "Partner device found, probing.\n");

	adapter->state = VNIC_PROBING;
	reinit_completion(&adapter->init_done);
	adapter->init_done_rc = 0;
	adapter->crq.active = true;

	rc = send_crq_init_complete(adapter);
	if (rc)
		goto out;

	rc = send_version_xchg(adapter);
	if (rc)
		netdev_dbg(adapter->netdev, "send_version_xchg failed, rc=%d\n", rc);

	if (!wait_for_completion_timeout(&adapter->init_done, timeout)) {
		dev_err(dev, "Initialization sequence timed out\n");
		rc = -ETIMEDOUT;
		goto out;
	}

	rc = init_sub_crqs(adapter);
	if (rc) {
		dev_err(dev, "Initialization of sub crqs failed, rc=%d\n", rc);
		goto out;
	}

	rc = init_sub_crq_irqs(adapter);
	if (rc) {
		dev_err(dev, "Failed to initialize sub crq irqs\n, rc=%d", rc);
		goto init_failed;
	}

	netdev->mtu = adapter->req_mtu - ETH_HLEN;
	netdev->min_mtu = adapter->min_mtu - ETH_HLEN;
	netdev->max_mtu = adapter->max_mtu - ETH_HLEN;

	adapter->state = VNIC_PROBED;
	netdev_dbg(netdev, "Probed successfully. Waiting for signal from partner device.\n");

	return 0;

init_failed:
	release_sub_crqs(adapter, 1);
out:
	adapter->state = VNIC_DOWN;
	return rc;
}

static void __ibmvnic_reset(struct work_struct *work)
{
	struct ibmvnic_adapter *adapter;
	bool saved_state = false;
	struct ibmvnic_rwi *tmprwi;
	struct ibmvnic_rwi *rwi;
	unsigned long flags;
	u32 reset_state;
	int rc = 0;

	adapter = container_of(work, struct ibmvnic_adapter, ibmvnic_reset);

	if (test_and_set_bit_lock(0, &adapter->resetting)) {
		queue_delayed_work(system_long_wq,
				   &adapter->ibmvnic_delayed_reset,
				   IBMVNIC_RESET_DELAY);
		return;
	}

	rwi = get_next_rwi(adapter);
	while (rwi) {
		spin_lock_irqsave(&adapter->state_lock, flags);

		if (adapter->state == VNIC_REMOVING ||
		    adapter->state == VNIC_REMOVED) {
			spin_unlock_irqrestore(&adapter->state_lock, flags);
			kfree(rwi);
			rc = EBUSY;
			break;
		}

		if (!saved_state) {
			reset_state = adapter->state;
			saved_state = true;
		}
		spin_unlock_irqrestore(&adapter->state_lock, flags);

		if (rwi->reset_reason == VNIC_RESET_PASSIVE_INIT) {
			rtnl_lock();
			rc = do_passive_init(adapter);
			rtnl_unlock();
			if (!rc)
				netif_carrier_on(adapter->netdev);
		} else if (adapter->force_reset_recovery) {
			/* Since we are doing a hard reset now, clear the
			 * failover_pending flag so we don't ignore any
			 * future MOBILITY or other resets.
			 */
			adapter->failover_pending = false;

			/* Transport event occurred during previous reset */
			if (adapter->wait_for_reset) {
				/* Previous was CHANGE_PARAM; caller locked */
				adapter->force_reset_recovery = false;
				rc = do_hard_reset(adapter, rwi, reset_state);
			} else {
				rtnl_lock();
				adapter->force_reset_recovery = false;
				rc = do_hard_reset(adapter, rwi, reset_state);
				rtnl_unlock();
			}
			if (rc) {
				/* give backing device time to settle down */
				netdev_dbg(adapter->netdev,
					   "[S:%s] Hard reset failed, waiting 60 secs\n",
					   adapter_state_to_string(adapter->state));
				set_current_state(TASK_UNINTERRUPTIBLE);
				schedule_timeout(60 * HZ);
			}
		} else {
			rc = do_reset(adapter, rwi, reset_state);
		}
		tmprwi = rwi;
		adapter->last_reset_time = jiffies;

		if (rc)
			netdev_dbg(adapter->netdev, "Reset failed, rc=%d\n", rc);

		rwi = get_next_rwi(adapter);

		/*
		 * If there is another reset queued, free the previous rwi
		 * and process the new reset even if previous reset failed
		 * (the previous reset could have failed because of a fail
		 * over for instance, so process the fail over).
		 *
		 * If there are no resets queued and the previous reset failed,
		 * the adapter would be in an undefined state. So retry the
		 * previous reset as a hard reset.
		 */
		if (rwi)
			kfree(tmprwi);
		else if (rc)
			rwi = tmprwi;

		if (rwi && (rwi->reset_reason == VNIC_RESET_FAILOVER ||
			    rwi->reset_reason == VNIC_RESET_MOBILITY || rc))
			adapter->force_reset_recovery = true;
	}

	if (adapter->wait_for_reset) {
		adapter->reset_done_rc = rc;
		complete(&adapter->reset_done);
	}

	clear_bit_unlock(0, &adapter->resetting);

	netdev_dbg(adapter->netdev,
		   "[S:%s FRR:%d WFR:%d] Done processing resets\n",
		   adapter_state_to_string(adapter->state),
		   adapter->force_reset_recovery,
		   adapter->wait_for_reset);
}

static void __ibmvnic_delayed_reset(struct work_struct *work)
{
	struct ibmvnic_adapter *adapter;

	adapter = container_of(work, struct ibmvnic_adapter,
			       ibmvnic_delayed_reset.work);
	__ibmvnic_reset(&adapter->ibmvnic_reset);
}

static int ibmvnic_reset(struct ibmvnic_adapter *adapter,
			 enum ibmvnic_reset_reason reason)
{
	struct list_head *entry, *tmp_entry;
	struct ibmvnic_rwi *rwi, *tmp;
	struct net_device *netdev = adapter->netdev;
	unsigned long flags;
	int ret;

	spin_lock_irqsave(&adapter->rwi_lock, flags);

	/* If failover is pending don't schedule any other reset.
	 * Instead let the failover complete. If there is already a
	 * a failover reset scheduled, we will detect and drop the
	 * duplicate reset when walking the ->rwi_list below.
	 */
	if (adapter->state == VNIC_REMOVING ||
	    adapter->state == VNIC_REMOVED ||
	    (adapter->failover_pending && reason != VNIC_RESET_FAILOVER)) {
		ret = EBUSY;
		netdev_dbg(netdev, "Adapter removing or pending failover, skipping reset\n");
		goto err;
	}

	if (adapter->state == VNIC_PROBING) {
		netdev_warn(netdev, "Adapter reset during probe\n");
		adapter->init_done_rc = -EAGAIN;
		ret = EAGAIN;