mlx5/mlx5_ib/mlx5_ib_mr.c

/*-
 * Copyright (c) 2013-2015, Mellanox Technologies, Ltd.  All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 * 1. Redistributions of source code must retain the above copyright
 *    notice, this list of conditions and the following disclaimer.
 * 2. Redistributions in binary form must reproduce the above copyright
 *    notice, this list of conditions and the following disclaimer in the
 *    documentation and/or other materials provided with the distribution.
 *
 * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS `AS IS' AND
 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
 * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
 * SUCH DAMAGE.
 *
 * $FreeBSD$
 */

#include <linux/kref.h>
#include <linux/random.h>
#include <linux/delay.h>
#include <linux/sched.h>
#include <rdma/ib_umem.h>
#include <rdma/ib_umem_odp.h>
#include <rdma/ib_verbs.h>
#include "mlx5_ib.h"

enum {
	MAX_PENDING_REG_MR = 8,
};

#define MLX5_UMR_ALIGN 2048
#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
static __be64 mlx5_ib_update_mtt_emergency_buffer[
		MLX5_UMR_MTT_MIN_CHUNK_SIZE/sizeof(__be64)]
	__aligned(MLX5_UMR_ALIGN);
static DEFINE_MUTEX(mlx5_ib_update_mtt_emergency_buffer_mutex);
#endif

static int clean_mr(struct mlx5_ib_mr *mr);

static int destroy_mkey(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr)
{
	int err = mlx5_core_destroy_mkey(dev->mdev, &mr->mmkey);

#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
	/* Wait until all page fault handlers using the mr complete. */
	synchronize_srcu(&dev->mr_srcu);
#endif

	return err;
}

static int order2idx(struct mlx5_ib_dev *dev, int order)
{
	struct mlx5_mr_cache *cache = &dev->cache;

	if (order < cache->ent[0].order)
		return 0;
	else
		return order - cache->ent[0].order;
}

static bool use_umr_mtt_update(struct mlx5_ib_mr *mr, u64 start, u64 length)
{
	return ((u64)1 << mr->order) * MLX5_ADAPTER_PAGE_SIZE >=
		length + (start & (MLX5_ADAPTER_PAGE_SIZE - 1));
}

#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
static void update_odp_mr(struct mlx5_ib_mr *mr)
{
	if (mr->umem->odp_data) {
		/*
		 * This barrier prevents the compiler from moving the
		 * setting of umem->odp_data->private to point to our
		 * MR, before reg_umr finished, to ensure that the MR
		 * initialization have finished before starting to
		 * handle invalidations.
		 */
		smp_wmb();
		mr->umem->odp_data->private = mr;
		/*
		 * Make sure we will see the new
		 * umem->odp_data->private value in the invalidation
		 * routines, before we can get page faults on the
		 * MR. Page faults can happen once we put the MR in
		 * the tree, below this line. Without the barrier,
		 * there can be a fault handling and an invalidation
		 * before umem->odp_data->private == mr is visible to
		 * the invalidation handler.
		 */
		smp_wmb();
	}
}
#endif

static void reg_mr_callback(int status, struct mlx5_async_work *context)
{
	struct mlx5_ib_mr *mr =
		container_of(context, struct mlx5_ib_mr, cb_work);
	struct mlx5_ib_dev *dev = mr->dev;
	struct mlx5_mr_cache *cache = &dev->cache;
	int c = order2idx(dev, mr->order);
	struct mlx5_cache_ent *ent = &cache->ent[c];
	u8 key;
	unsigned long flags;
	struct mlx5_mr_table *table = &dev->mdev->priv.mr_table;
	int err;

	spin_lock_irqsave(&ent->lock, flags);
	ent->pending--;
	spin_unlock_irqrestore(&ent->lock, flags);
	if (status) {
		mlx5_ib_warn(dev, "async reg mr failed. status %d\n", status);
		kfree(mr);
		dev->fill_delay = 1;
		mod_timer(&dev->delay_timer, jiffies + HZ);
		return;
	}

	spin_lock_irqsave(&dev->mdev->priv.mkey_lock, flags);
	key = dev->mdev->priv.mkey_key++;
	spin_unlock_irqrestore(&dev->mdev->priv.mkey_lock, flags);
	mr->mmkey.key = mlx5_idx_to_mkey(MLX5_GET(create_mkey_out, mr->out, mkey_index)) | key;

	cache->last_add = jiffies;

	spin_lock_irqsave(&ent->lock, flags);
	list_add_tail(&mr->list, &ent->head);
	ent->cur++;
	ent->size++;
	spin_unlock_irqrestore(&ent->lock, flags);

	spin_lock_irqsave(&table->lock, flags);
	err = radix_tree_insert(&table->tree, mlx5_mkey_to_idx(mr->mmkey.key),
				&mr->mmkey);
	if (err)
		pr_err("Error inserting to mkey tree. 0x%x\n", -err);
	spin_unlock_irqrestore(&table->lock, flags);
}

static int add_keys(struct mlx5_ib_dev *dev, int c, int num)
{
	struct mlx5_mr_cache *cache = &dev->cache;
	struct mlx5_cache_ent *ent = &cache->ent[c];
	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
	struct mlx5_ib_mr *mr;
	int npages = 1 << ent->order;
	void *mkc;
	u32 *in;
	int err = 0;
	int i;

	in = kzalloc(inlen, GFP_KERNEL);
	if (!in)
		return -ENOMEM;

	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
	for (i = 0; i < num; i++) {
		if (ent->pending >= MAX_PENDING_REG_MR) {
			err = -EAGAIN;
			break;
		}

		mr = kzalloc(sizeof(*mr), GFP_KERNEL);
		if (!mr) {
			err = -ENOMEM;
			break;
		}
		mr->order = ent->order;
		mr->umred = 1;
		mr->dev = dev;

		MLX5_SET(mkc, mkc, free, 1);
		MLX5_SET(mkc, mkc, umr_en, 1);
		MLX5_SET(mkc, mkc, access_mode, MLX5_ACCESS_MODE_MTT);

		MLX5_SET(mkc, mkc, qpn, 0xffffff);
		MLX5_SET(mkc, mkc, translations_octword_size, (npages + 1) / 2);
		MLX5_SET(mkc, mkc, log_page_size, 12);

		spin_lock_irq(&ent->lock);
		ent->pending++;
		spin_unlock_irq(&ent->lock);
		err = mlx5_core_create_mkey_cb(dev->mdev, &mr->mmkey,
					       &dev->async_ctx, in, inlen,
					       mr->out, sizeof(mr->out),
					       reg_mr_callback, &mr->cb_work);
		if (err) {
			spin_lock_irq(&ent->lock);
			ent->pending--;
			spin_unlock_irq(&ent->lock);
			mlx5_ib_warn(dev, "create mkey failed %d\n", err);
			kfree(mr);
			break;
		}
	}

	kfree(in);
	return err;
}

static void remove_keys(struct mlx5_ib_dev *dev, int c, int num)
{
	struct mlx5_mr_cache *cache = &dev->cache;
	struct mlx5_cache_ent *ent = &cache->ent[c];
	struct mlx5_ib_mr *mr;
	int err;
	int i;

	for (i = 0; i < num; i++) {
		spin_lock_irq(&ent->lock);
		if (list_empty(&ent->head)) {
			spin_unlock_irq(&ent->lock);
			return;
		}
		mr = list_first_entry(&ent->head, struct mlx5_ib_mr, list);
		list_del(&mr->list);
		ent->cur--;
		ent->size--;
		spin_unlock_irq(&ent->lock);
		err = destroy_mkey(dev, mr);
		if (err)
			mlx5_ib_warn(dev, "failed destroy mkey\n");
		else
			kfree(mr);
	}
}

static int someone_adding(struct mlx5_mr_cache *cache)
{
	int i;

	for (i = 0; i < MAX_MR_CACHE_ENTRIES; i++) {
		if (cache->ent[i].cur < cache->ent[i].limit)
			return 1;
	}

	return 0;
}

static void __cache_work_func(struct mlx5_cache_ent *ent)
{
	struct mlx5_ib_dev *dev = ent->dev;
	struct mlx5_mr_cache *cache = &dev->cache;
	int i = order2idx(dev, ent->order);
	int err;

	if (cache->stopped)
		return;

	ent = &dev->cache.ent[i];
	if (ent->cur < 2 * ent->limit && !dev->fill_delay) {
		err = add_keys(dev, i, 1);
		if (ent->cur < 2 * ent->limit) {
			if (err == -EAGAIN) {
				mlx5_ib_dbg(dev, "returned eagain, order %d\n",
					    i + 2);
				queue_delayed_work(cache->wq, &ent->dwork,
						   msecs_to_jiffies(3));
			} else if (err) {
				mlx5_ib_warn(dev, "command failed order %d, err %d\n",
					     i + 2, err);
				queue_delayed_work(cache->wq, &ent->dwork,
						   msecs_to_jiffies(1000));
			} else {
				queue_work(cache->wq, &ent->work);
			}
		}
	} else if (ent->cur > 2 * ent->limit) {
		/*
		 * The remove_keys() logic is performed as garbage collection
		 * task. Such task is intended to be run when no other active
		 * processes are running.
		 *
		 * The need_resched() will return TRUE if there are user tasks
		 * to be activated in near future.
		 *
		 * In such case, we don't execute remove_keys() and postpone
		 * the garbage collection work to try to run in next cycle,
		 * in order to free CPU resources to other tasks.
		 */
		if (!need_resched() && !someone_adding(cache) &&
		    time_after(jiffies, cache->last_add + 300 * HZ)) {
			remove_keys(dev, i, 1);
			if (ent->cur > ent->limit)
				queue_work(cache->wq, &ent->work);
		} else {
			queue_delayed_work(cache->wq, &ent->dwork, 300 * HZ);
		}
	}
}

static void delayed_cache_work_func(struct work_struct *work)
{
	struct mlx5_cache_ent *ent;

	ent = container_of(work, struct mlx5_cache_ent, dwork.work);
	__cache_work_func(ent);
}

static void cache_work_func(struct work_struct *work)
{
	struct mlx5_cache_ent *ent;

	ent = container_of(work, struct mlx5_cache_ent, work);
	__cache_work_func(ent);
}

static struct mlx5_ib_mr *alloc_cached_mr(struct mlx5_ib_dev *dev, int order)
{
	struct mlx5_mr_cache *cache = &dev->cache;
	struct mlx5_ib_mr *mr = NULL;
	struct mlx5_cache_ent *ent;
	int c;
	int i;

	c = order2idx(dev, order);
	if (c < 0 || c >= MAX_MR_CACHE_ENTRIES) {
		mlx5_ib_warn(dev, "order %d, cache index %d\n", order, c);
		return NULL;
	}

	for (i = c; i < MAX_MR_CACHE_ENTRIES; i++) {
		ent = &cache->ent[i];

		mlx5_ib_dbg(dev, "order %d, cache index %d\n", ent->order, i);

		spin_lock_irq(&ent->lock);
		if (!list_empty(&ent->head)) {
			mr = list_first_entry(&ent->head, struct mlx5_ib_mr,
					      list);
			list_del(&mr->list);
			ent->cur--;
			spin_unlock_irq(&ent->lock);
			if (ent->cur < ent->limit)
				queue_work(cache->wq, &ent->work);
			break;
		}
		spin_unlock_irq(&ent->lock);

		queue_work(cache->wq, &ent->work);
	}

	if (!mr)
		cache->ent[c].miss++;

	return mr;
}

static void free_cached_mr(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr)
{
	struct mlx5_mr_cache *cache = &dev->cache;
	struct mlx5_cache_ent *ent;
	int shrink = 0;
	int c;

	c = order2idx(dev, mr->order);
	if (c < 0 || c >= MAX_MR_CACHE_ENTRIES) {
		mlx5_ib_warn(dev, "order %d, cache index %d\n", mr->order, c);
		return;
	}
	ent = &cache->ent[c];
	spin_lock_irq(&ent->lock);
	list_add_tail(&mr->list, &ent->head);
	ent->cur++;
	if (ent->cur > 2 * ent->limit)
		shrink = 1;
	spin_unlock_irq(&ent->lock);

	if (shrink)
		queue_work(cache->wq, &ent->work);
}

static void clean_keys(struct mlx5_ib_dev *dev, int c)
{
	struct mlx5_mr_cache *cache = &dev->cache;
	struct mlx5_cache_ent *ent = &cache->ent[c];
	struct mlx5_ib_mr *mr;
	int err;

	cancel_delayed_work(&ent->dwork);
	while (1) {
		spin_lock_irq(&ent->lock);
		if (list_empty(&ent->head)) {
			spin_unlock_irq(&ent->lock);
			return;
		}
		mr = list_first_entry(&ent->head, struct mlx5_ib_mr, list);
		list_del(&mr->list);
		ent->cur--;
		ent->size--;
		spin_unlock_irq(&ent->lock);
		err = destroy_mkey(dev, mr);
		if (err)
			mlx5_ib_warn(dev, "failed destroy mkey\n");
		else
			kfree(mr);
	}
}

static void delay_time_func(unsigned long ctx)
{
	struct mlx5_ib_dev *dev = (struct mlx5_ib_dev *)ctx;

	dev->fill_delay = 0;
}

int mlx5_mr_cache_init(struct mlx5_ib_dev *dev)
{
	struct mlx5_mr_cache *cache = &dev->cache;
	struct mlx5_cache_ent *ent;
	int limit;
	int i;

	mutex_init(&dev->slow_path_mutex);
	cache->wq = alloc_ordered_workqueue("mkey_cache", WQ_MEM_RECLAIM);
	if (!cache->wq) {
		mlx5_ib_warn(dev, "failed to create work queue\n");
		return -ENOMEM;
	}

	mlx5_cmd_init_async_ctx(dev->mdev, &dev->async_ctx);
	setup_timer(&dev->delay_timer, delay_time_func, (unsigned long)dev);
	for (i = 0; i < MAX_MR_CACHE_ENTRIES; i++) {
		INIT_LIST_HEAD(&cache->ent[i].head);
		spin_lock_init(&cache->ent[i].lock);

		ent = &cache->ent[i];
		INIT_LIST_HEAD(&ent->head);
		spin_lock_init(&ent->lock);
		ent->order = i + 2;
		ent->dev = dev;

		if (dev->mdev->profile->mask & MLX5_PROF_MASK_MR_CACHE)
			limit = dev->mdev->profile->mr_cache[i].limit;
		else
			limit = 0;

		INIT_WORK(&ent->work, cache_work_func);
		INIT_DELAYED_WORK(&ent->dwork, delayed_cache_work_func);
		ent->limit = limit;
		queue_work(cache->wq, &ent->work);
	}

	return 0;
}

int mlx5_mr_cache_cleanup(struct mlx5_ib_dev *dev)
{
	int i;

	dev->cache.stopped = 1;
	flush_workqueue(dev->cache.wq);
	mlx5_cmd_cleanup_async_ctx(&dev->async_ctx);

	for (i = 0; i < MAX_MR_CACHE_ENTRIES; i++)
		clean_keys(dev, i);

	destroy_workqueue(dev->cache.wq);
	del_timer_sync(&dev->delay_timer);

	return 0;
}

struct ib_mr *mlx5_ib_get_dma_mr(struct ib_pd *pd, int acc)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
	struct mlx5_core_dev *mdev = dev->mdev;
	struct mlx5_ib_mr *mr;
	void *mkc;
	u32 *in;
	int err;

	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
	if (!mr)
		return ERR_PTR(-ENOMEM);

	in = kzalloc(inlen, GFP_KERNEL);
	if (!in) {
		err = -ENOMEM;
		goto err_free;
	}

	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);

	MLX5_SET(mkc, mkc, access_mode, MLX5_ACCESS_MODE_PA);
	MLX5_SET(mkc, mkc, a, !!(acc & IB_ACCESS_REMOTE_ATOMIC));
	MLX5_SET(mkc, mkc, rw, !!(acc & IB_ACCESS_REMOTE_WRITE));
	MLX5_SET(mkc, mkc, rr, !!(acc & IB_ACCESS_REMOTE_READ));
	MLX5_SET(mkc, mkc, lw, !!(acc & IB_ACCESS_LOCAL_WRITE));
	MLX5_SET(mkc, mkc, lr, 1);

	MLX5_SET(mkc, mkc, length64, 1);
	MLX5_SET(mkc, mkc, pd, to_mpd(pd)->pdn);
	MLX5_SET(mkc, mkc, qpn, 0xffffff);
	MLX5_SET64(mkc, mkc, start_addr, 0);

	err = mlx5_core_create_mkey(mdev, &mr->mmkey, in, inlen);
	if (err)
		goto err_in;

	kfree(in);
	mr->ibmr.lkey = mr->mmkey.key;
	mr->ibmr.rkey = mr->mmkey.key;
	mr->umem = NULL;

	return &mr->ibmr;

err_in:
	kfree(in);

err_free:
	kfree(mr);

	return ERR_PTR(err);
}

static int get_octo_len(u64 addr, u64 len, int page_size)
{
	u64 offset;
	int npages;

	offset = addr & (page_size - 1);
	npages = ALIGN(len + offset, page_size) >> ilog2(page_size);
	return (npages + 1) / 2;
}

static int use_umr(int order)
{
	return order <= MLX5_MAX_UMR_SHIFT;
}

static int dma_map_mr_pas(struct mlx5_ib_dev *dev, struct ib_umem *umem,
			  int npages, int page_shift, int *size,
			  __be64 **mr_pas, dma_addr_t *dma)
{
	__be64 *pas;
	struct device *ddev = dev->ib_dev.dma_device;

	/*
	 * UMR copies MTTs in units of MLX5_UMR_MTT_ALIGNMENT bytes.
	 * To avoid copying garbage after the pas array, we allocate
	 * a little more.
	 */
	*size = ALIGN(sizeof(u64) * npages, MLX5_UMR_MTT_ALIGNMENT);
	*mr_pas = kmalloc(*size + MLX5_UMR_ALIGN - 1, GFP_KERNEL);
	if (!(*mr_pas))
		return -ENOMEM;

	pas = PTR_ALIGN(*mr_pas, MLX5_UMR_ALIGN);
	mlx5_ib_populate_pas(dev, umem, page_shift, pas, MLX5_IB_MTT_PRESENT);
	/* Clear padding after the actual pages. */
	memset(pas + npages, 0, *size - npages * sizeof(u64));

	*dma = dma_map_single(ddev, pas, *size, DMA_TO_DEVICE);
	if (dma_mapping_error(ddev, *dma)) {
		kfree(*mr_pas);
		return -ENOMEM;
	}

	return 0;
}

static void prep_umr_wqe_common(struct ib_pd *pd, struct ib_send_wr *wr,
				struct ib_sge *sg, u64 dma, int n, u32 key,
				int page_shift)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	struct mlx5_umr_wr *umrwr = umr_wr(wr);

	sg->addr = dma;
	sg->length = ALIGN(sizeof(u64) * n, 64);
	sg->lkey = dev->umrc.pd->local_dma_lkey;

	wr->next = NULL;
	wr->sg_list = sg;
	if (n)
		wr->num_sge = 1;
	else
		wr->num_sge = 0;

	wr->opcode = MLX5_IB_WR_UMR;

	umrwr->npages = n;
	umrwr->page_shift = page_shift;
	umrwr->mkey = key;
}

static void prep_umr_reg_wqe(struct ib_pd *pd, struct ib_send_wr *wr,
			     struct ib_sge *sg, u64 dma, int n, u32 key,
			     int page_shift, u64 virt_addr, u64 len,
			     int access_flags)
{
	struct mlx5_umr_wr *umrwr = umr_wr(wr);

	prep_umr_wqe_common(pd, wr, sg, dma, n, key, page_shift);

	wr->send_flags = 0;

	umrwr->target.virt_addr = virt_addr;
	umrwr->length = len;
	umrwr->access_flags = access_flags;
	umrwr->pd = pd;
}

static void prep_umr_unreg_wqe(struct mlx5_ib_dev *dev,
			       struct ib_send_wr *wr, u32 key)
{
	struct mlx5_umr_wr *umrwr = umr_wr(wr);

	wr->send_flags = MLX5_IB_SEND_UMR_UNREG | MLX5_IB_SEND_UMR_FAIL_IF_FREE;
	wr->opcode = MLX5_IB_WR_UMR;
	umrwr->mkey = key;
}

static struct ib_umem *mr_umem_get(struct ib_pd *pd, u64 start, u64 length,
				   int access_flags, int *npages,
				   int *page_shift, int *ncont, int *order)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	struct ib_umem *umem = ib_umem_get(pd->uobject->context, start, length,
					   access_flags, 0);
	if (IS_ERR(umem)) {
		mlx5_ib_err(dev, "umem get failed (%ld)\n", PTR_ERR(umem));
		return (void *)umem;
	}

	mlx5_ib_cont_pages(umem, start, npages, page_shift, ncont, order);
	if (!*npages) {
		mlx5_ib_warn(dev, "avoid zero region\n");
		ib_umem_release(umem);
		return ERR_PTR(-EINVAL);
	}

	mlx5_ib_dbg(dev, "npages %d, ncont %d, order %d, page_shift %d\n",
		    *npages, *ncont, *order, *page_shift);

	return umem;
}

static void mlx5_ib_umr_done(struct ib_cq *cq, struct ib_wc *wc)
{
	struct mlx5_ib_umr_context *context =
		container_of(wc->wr_cqe, struct mlx5_ib_umr_context, cqe);

	context->status = wc->status;
	complete(&context->done);
}

static inline void mlx5_ib_init_umr_context(struct mlx5_ib_umr_context *context)
{
	context->cqe.done = mlx5_ib_umr_done;
	context->status = -1;
	init_completion(&context->done);
}

static struct mlx5_ib_mr *reg_umr(struct ib_pd *pd, struct ib_umem *umem,
				  u64 virt_addr, u64 len, int npages,
				  int page_shift, int order, int access_flags)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	struct device *ddev = dev->ib_dev.dma_device;
	struct umr_common *umrc = &dev->umrc;
	struct mlx5_ib_umr_context umr_context;
	struct mlx5_umr_wr umrwr = {};
	struct ib_send_wr *bad;
	struct mlx5_ib_mr *mr;
	struct ib_sge sg;
	int size;
	__be64 *mr_pas;
	dma_addr_t dma;
	int err = 0;
	int i;

	for (i = 0; i < 1; i++) {
		mr = alloc_cached_mr(dev, order);
		if (mr)
			break;

		err = add_keys(dev, order2idx(dev, order), 1);
		if (err && err != -EAGAIN) {
			mlx5_ib_warn(dev, "add_keys failed, err %d\n", err);
			break;
		}
	}

	if (!mr)
		return ERR_PTR(-EAGAIN);

	err = dma_map_mr_pas(dev, umem, npages, page_shift, &size, &mr_pas,
			     &dma);
	if (err)
		goto free_mr;

	mlx5_ib_init_umr_context(&umr_context);

	umrwr.wr.wr_cqe = &umr_context.cqe;
	prep_umr_reg_wqe(pd, &umrwr.wr, &sg, dma, npages, mr->mmkey.key,
			 page_shift, virt_addr, len, access_flags);

	down(&umrc->sem);
	err = ib_post_send(umrc->qp, &umrwr.wr, &bad);
	if (err) {
		mlx5_ib_warn(dev, "post send failed, err %d\n", err);
		goto unmap_dma;
	} else {
		wait_for_completion(&umr_context.done);
		if (umr_context.status != IB_WC_SUCCESS) {
			mlx5_ib_warn(dev, "reg umr failed\n");
			err = -EFAULT;
		}
	}

	mr->mmkey.iova = virt_addr;
	mr->mmkey.size = len;
	mr->mmkey.pd = to_mpd(pd)->pdn;

	mr->live = 1;

unmap_dma:
	up(&umrc->sem);
	dma_unmap_single(ddev, dma, size, DMA_TO_DEVICE);

	kfree(mr_pas);

free_mr:
	if (err) {
		free_cached_mr(dev, mr);
		return ERR_PTR(err);
	}

	return mr;
}

#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
int mlx5_ib_update_mtt(struct mlx5_ib_mr *mr, u64 start_page_index, int npages,
		       int zap)
{
	struct mlx5_ib_dev *dev = mr->dev;
	struct device *ddev = dev->ib_dev.dma_device;
	struct umr_common *umrc = &dev->umrc;
	struct mlx5_ib_umr_context umr_context;
	struct ib_umem *umem = mr->umem;
	int size;
	__be64 *pas;
	dma_addr_t dma;
	struct ib_send_wr *bad;
	struct mlx5_umr_wr wr;
	struct ib_sge sg;
	int err = 0;
	const int page_index_alignment = MLX5_UMR_MTT_ALIGNMENT / sizeof(u64);
	const int page_index_mask = page_index_alignment - 1;
	size_t pages_mapped = 0;
	size_t pages_to_map = 0;
	size_t pages_iter = 0;
	int use_emergency_buf = 0;

	/* UMR copies MTTs in units of MLX5_UMR_MTT_ALIGNMENT bytes,
	 * so we need to align the offset and length accordingly */
	if (start_page_index & page_index_mask) {
		npages += start_page_index & page_index_mask;
		start_page_index &= ~page_index_mask;
	}

	pages_to_map = ALIGN(npages, page_index_alignment);

	if (start_page_index + pages_to_map > MLX5_MAX_UMR_PAGES)
		return -EINVAL;

	size = sizeof(u64) * pages_to_map;
	size = min_t(int, PAGE_SIZE, size);
	/* We allocate with GFP_ATOMIC to avoid recursion into page-reclaim
	 * code, when we are called from an invalidation. The pas buffer must
	 * be 2k-aligned for Connect-IB. */
	pas = (__be64 *)get_zeroed_page(GFP_ATOMIC);
	if (!pas) {
		mlx5_ib_warn(dev, "unable to allocate memory during MTT update, falling back to slower chunked mechanism.\n");
		pas = mlx5_ib_update_mtt_emergency_buffer;
		size = MLX5_UMR_MTT_MIN_CHUNK_SIZE;
		use_emergency_buf = 1;
		mutex_lock(&mlx5_ib_update_mtt_emergency_buffer_mutex);
		memset(pas, 0, size);
	}
	pages_iter = size / sizeof(u64);
	dma = dma_map_single(ddev, pas, size, DMA_TO_DEVICE);
	if (dma_mapping_error(ddev, dma)) {
		mlx5_ib_err(dev, "unable to map DMA during MTT update.\n");
		err = -ENOMEM;
		goto free_pas;
	}

	for (pages_mapped = 0;
	     pages_mapped < pages_to_map && !err;
	     pages_mapped += pages_iter, start_page_index += pages_iter) {
		dma_sync_single_for_cpu(ddev, dma, size, DMA_TO_DEVICE);

		npages = min_t(size_t,
			       pages_iter,
			       ib_umem_num_pages(umem) - start_page_index);

		if (!zap) {
			__mlx5_ib_populate_pas(dev, umem, PAGE_SHIFT,
					       start_page_index, npages, pas,
					       MLX5_IB_MTT_PRESENT);
			/* Clear padding after the pages brought from the
			 * umem. */
			memset(pas + npages, 0, size - npages * sizeof(u64));
		}

		dma_sync_single_for_device(ddev, dma, size, DMA_TO_DEVICE);

		mlx5_ib_init_umr_context(&umr_context);

		memset(&wr, 0, sizeof(wr));
		wr.wr.wr_cqe = &umr_context.cqe;

		sg.addr = dma;
		sg.length = ALIGN(npages * sizeof(u64),
				MLX5_UMR_MTT_ALIGNMENT);
		sg.lkey = dev->umrc.pd->local_dma_lkey;

		wr.wr.send_flags = MLX5_IB_SEND_UMR_FAIL_IF_FREE |
				MLX5_IB_SEND_UMR_UPDATE_MTT;
		wr.wr.sg_list = &sg;
		wr.wr.num_sge = 1;
		wr.wr.opcode = MLX5_IB_WR_UMR;
		wr.npages = sg.length / sizeof(u64);
		wr.page_shift = PAGE_SHIFT;
		wr.mkey = mr->mmkey.key;
		wr.target.offset = start_page_index;

		down(&umrc->sem);
		err = ib_post_send(umrc->qp, &wr.wr, &bad);
		if (err) {
			mlx5_ib_err(dev, "UMR post send failed, err %d\n", err);
		} else {
			wait_for_completion(&umr_context.done);
			if (umr_context.status != IB_WC_SUCCESS) {
				mlx5_ib_err(dev, "UMR completion failed, code %d\n",
					    umr_context.status);
				err = -EFAULT;
			}
		}
		up(&umrc->sem);
	}
	dma_unmap_single(ddev, dma, size, DMA_TO_DEVICE);

free_pas:
	if (!use_emergency_buf)
		free_page((unsigned long)pas);
	else
		mutex_unlock(&mlx5_ib_update_mtt_emergency_buffer_mutex);

	return err;
}
#endif

/*
 * If ibmr is NULL it will be allocated by reg_create.
 * Else, the given ibmr will be used.
 */
static struct mlx5_ib_mr *reg_create(struct ib_mr *ibmr, struct ib_pd *pd,
				     u64 virt_addr, u64 length,
				     struct ib_umem *umem, int npages,
				     int page_shift, int access_flags)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	struct mlx5_ib_mr *mr;
	__be64 *pas;
	void *mkc;
	int inlen;
	u32 *in;
	int err;
	bool pg_cap = !!(MLX5_CAP_GEN(dev->mdev, pg));

	mr = ibmr ? to_mmr(ibmr) : kzalloc(sizeof(*mr), GFP_KERNEL);
	if (!mr)
		return ERR_PTR(-ENOMEM);

	inlen = MLX5_ST_SZ_BYTES(create_mkey_in) +
		sizeof(*pas) * ((npages + 1) / 2) * 2;
	in = mlx5_vzalloc(inlen);
	if (!in) {
		err = -ENOMEM;
		goto err_1;
	}
	pas = (__be64 *)MLX5_ADDR_OF(create_mkey_in, in, klm_pas_mtt);
	mlx5_ib_populate_pas(dev, umem, page_shift, pas,
			     pg_cap ? MLX5_IB_MTT_PRESENT : 0);

	/* The pg_access bit allows setting the access flags
	 * in the page list submitted with the command. */
	MLX5_SET(create_mkey_in, in, pg_access, !!(pg_cap));

	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
	MLX5_SET(mkc, mkc, access_mode, MLX5_ACCESS_MODE_MTT);
	MLX5_SET(mkc, mkc, a, !!(access_flags & IB_ACCESS_REMOTE_ATOMIC));
	MLX5_SET(mkc, mkc, rw, !!(access_flags & IB_ACCESS_REMOTE_WRITE));
	MLX5_SET(mkc, mkc, rr, !!(access_flags & IB_ACCESS_REMOTE_READ));
	MLX5_SET(mkc, mkc, lw, !!(access_flags & IB_ACCESS_LOCAL_WRITE));
	MLX5_SET(mkc, mkc, lr, 1);

	MLX5_SET64(mkc, mkc, start_addr, virt_addr);
	MLX5_SET64(mkc, mkc, len, length);
	MLX5_SET(mkc, mkc, pd, to_mpd(pd)->pdn);
	MLX5_SET(mkc, mkc, bsf_octword_size, 0);
	MLX5_SET(mkc, mkc, translations_octword_size,
		 get_octo_len(virt_addr, length, 1 << page_shift));
	MLX5_SET(mkc, mkc, log_page_size, page_shift);
	MLX5_SET(mkc, mkc, qpn, 0xffffff);
	MLX5_SET(create_mkey_in, in, translations_octword_actual_size,
		 get_octo_len(virt_addr, length, 1 << page_shift));

	err = mlx5_core_create_mkey(dev->mdev, &mr->mmkey, in, inlen);
	if (err) {
		mlx5_ib_warn(dev, "create mkey failed\n");
		goto err_2;
	}
	mr->umem = umem;
	mr->dev = dev;
	mr->live = 1;
	kvfree(in);

	mlx5_ib_dbg(dev, "mkey = 0x%x\n", mr->mmkey.key);

	return mr;

err_2:
	kvfree(in);

err_1:
	if (!ibmr)
		kfree(mr);

	return ERR_PTR(err);
}

static void set_mr_fileds(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr,
			  int npages, u64 length, int access_flags)
{
	mr->npages = npages;
	atomic_add(npages, &dev->mdev->priv.reg_pages);
	mr->ibmr.lkey = mr->mmkey.key;
	mr->ibmr.rkey = mr->mmkey.key;
	mr->ibmr.length = length;
	mr->access_flags = access_flags;
}

struct ib_mr *mlx5_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
				  u64 virt_addr, int access_flags,
				  struct ib_udata *udata)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	struct mlx5_ib_mr *mr = NULL;
	struct ib_umem *umem;
	int page_shift;
	int npages;
	int ncont;
	int order;
	int err;

	mlx5_ib_dbg(dev, "start 0x%llx, virt_addr 0x%llx, length 0x%llx, access_flags 0x%x\n",
		    (long long)start, (long long)virt_addr, (long long)length, access_flags);
	umem = mr_umem_get(pd, start, length, access_flags, &npages,
			   &page_shift, &ncont, &order);

	if (IS_ERR(umem))
		return (void *)umem;

	if (use_umr(order)) {
		mr = reg_umr(pd, umem, virt_addr, length, ncont, page_shift,
			     order, access_flags);
		if (PTR_ERR(mr) == -EAGAIN) {
			mlx5_ib_dbg(dev, "cache empty for order %d", order);
			mr = NULL;
		}
	} else if (access_flags & IB_ACCESS_ON_DEMAND) {
		err = -EINVAL;
		pr_err("Got MR registration for ODP MR > 512MB, not supported for Connect-IB");
		goto error;
	}

	if (!mr) {
		mutex_lock(&dev->slow_path_mutex);
		mr = reg_create(NULL, pd, virt_addr, length, umem, ncont,
				page_shift, access_flags);
		mutex_unlock(&dev->slow_path_mutex);
	}

	if (IS_ERR(mr)) {
		err = PTR_ERR(mr);
		goto error;
	}

	mlx5_ib_dbg(dev, "mkey 0x%x\n", mr->mmkey.key);

	mr->umem = umem;
	set_mr_fileds(dev, mr, npages, length, access_flags);

#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
	update_odp_mr(mr);
#endif

	return &mr->ibmr;

error:
	ib_umem_release(umem);
	return ERR_PTR(err);
}

static int unreg_umr(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr)
{
	struct mlx5_core_dev *mdev = dev->mdev;
	struct umr_common *umrc = &dev->umrc;
	struct mlx5_ib_umr_context umr_context;
	struct mlx5_umr_wr umrwr = {};
	struct ib_send_wr *bad;
	int err;

	if (mdev->state == MLX5_DEVICE_STATE_INTERNAL_ERROR)
		return 0;

	mlx5_ib_init_umr_context(&umr_context);

	umrwr.wr.wr_cqe = &umr_context.cqe;
	prep_umr_unreg_wqe(dev, &umrwr.wr, mr->mmkey.key);

	down(&umrc->sem);
	err = ib_post_send(umrc->qp, &umrwr.wr, &bad);
	if (err) {
		up(&umrc->sem);
		mlx5_ib_dbg(dev, "err %d\n", err);
		goto error;
	} else {
		wait_for_completion(&umr_context.done);
		up(&umrc->sem);
	}
	if (umr_context.status != IB_WC_SUCCESS) {
		mlx5_ib_warn(dev, "unreg umr failed\n");
		err = -EFAULT;
		goto error;
	}
	return 0;

error:
	return err;
}

static int rereg_umr(struct ib_pd *pd, struct mlx5_ib_mr *mr, u64 virt_addr,
		     u64 length, int npages, int page_shift, int order,
		     int access_flags, int flags)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	struct device *ddev = dev->ib_dev.dma_device;
	struct mlx5_ib_umr_context umr_context;
	struct ib_send_wr *bad;
	struct mlx5_umr_wr umrwr = {};
	struct ib_sge sg;
	struct umr_common *umrc = &dev->umrc;
	dma_addr_t dma = 0;
	__be64 *mr_pas = NULL;
	int size;
	int err;

	mlx5_ib_init_umr_context(&umr_context);

	umrwr.wr.wr_cqe = &umr_context.cqe;
	umrwr.wr.send_flags = MLX5_IB_SEND_UMR_FAIL_IF_FREE;

	if (flags & IB_MR_REREG_TRANS) {
		err = dma_map_mr_pas(dev, mr->umem, npages, page_shift, &size,
				     &mr_pas, &dma);
		if (err)
			return err;

		umrwr.target.virt_addr = virt_addr;
		umrwr.length = length;
		umrwr.wr.send_flags |= MLX5_IB_SEND_UMR_UPDATE_TRANSLATION;
	}

	prep_umr_wqe_common(pd, &umrwr.wr, &sg, dma, npages, mr->mmkey.key,
			    page_shift);

	if (flags & IB_MR_REREG_PD) {
		umrwr.pd = pd;
		umrwr.wr.send_flags |= MLX5_IB_SEND_UMR_UPDATE_PD;
	}

	if (flags & IB_MR_REREG_ACCESS) {
		umrwr.access_flags = access_flags;
		umrwr.wr.send_flags |= MLX5_IB_SEND_UMR_UPDATE_ACCESS;
	}

	/* post send request to UMR QP */
	down(&umrc->sem);
	err = ib_post_send(umrc->qp, &umrwr.wr, &bad);

	if (err) {
		mlx5_ib_warn(dev, "post send failed, err %d\n", err);
	} else {
		wait_for_completion(&umr_context.done);
		if (umr_context.status != IB_WC_SUCCESS) {
			mlx5_ib_warn(dev, "reg umr failed (%u)\n",
				     umr_context.status);
			err = -EFAULT;
		}
	}

	up(&umrc->sem);
	if (flags & IB_MR_REREG_TRANS) {
		dma_unmap_single(ddev, dma, size, DMA_TO_DEVICE);
		kfree(mr_pas);
	}
	return err;
}

int mlx5_ib_rereg_user_mr(struct ib_mr *ib_mr, int flags, u64 start,
			  u64 length, u64 virt_addr, int new_access_flags,
			  struct ib_pd *new_pd, struct ib_udata *udata)
{
	struct mlx5_ib_dev *dev = to_mdev(ib_mr->device);
	struct mlx5_ib_mr *mr = to_mmr(ib_mr);
	struct ib_pd *pd = (flags & IB_MR_REREG_PD) ? new_pd : ib_mr->pd;
	int access_flags = flags & IB_MR_REREG_ACCESS ?
			    new_access_flags :
			    mr->access_flags;
	u64 addr = (flags & IB_MR_REREG_TRANS) ? virt_addr : mr->umem->address;
	u64 len = (flags & IB_MR_REREG_TRANS) ? length : mr->umem->length;
	int page_shift = 0;
	int npages = 0;
	int ncont = 0;
	int order = 0;
	int err;

	mlx5_ib_dbg(dev, "start 0x%llx, virt_addr 0x%llx, length 0x%llx, access_flags 0x%x\n",
		    (long long)start, (long long)virt_addr, (long long)length, access_flags);

	if (flags != IB_MR_REREG_PD) {
		/*
		 * Replace umem. This needs to be done whether or not UMR is
		 * used.
		 */
		flags |= IB_MR_REREG_TRANS;
		ib_umem_release(mr->umem);
		mr->umem = mr_umem_get(pd, addr, len, access_flags, &npages,
				       &page_shift, &ncont, &order);
		if (IS_ERR(mr->umem)) {
			err = PTR_ERR(mr->umem);
			mr->umem = NULL;
			return err;
		}
	}

	if (flags & IB_MR_REREG_TRANS && !use_umr_mtt_update(mr, addr, len)) {
		/*
		 * UMR can't be used - MKey needs to be replaced.
		 */
		if (mr->umred) {
			err = unreg_umr(dev, mr);
			if (err)
				mlx5_ib_warn(dev, "Failed to unregister MR\n");
		} else {
			err = destroy_mkey(dev, mr);
			if (err)
				mlx5_ib_warn(dev, "Failed to destroy MKey\n");
		}
		if (err)
			return err;

		mr = reg_create(ib_mr, pd, addr, len, mr->umem, ncont,
				page_shift, access_flags);

		if (IS_ERR(mr))
			return PTR_ERR(mr);

		mr->umred = 0;
	} else {
		/*
		 * Send a UMR WQE
		 */
		err = rereg_umr(pd, mr, addr, len, npages, page_shift,
				order, access_flags, flags);
		if (err) {
			mlx5_ib_warn(dev, "Failed to rereg UMR\n");
			return err;
		}
	}

	if (flags & IB_MR_REREG_PD) {
		ib_mr->pd = pd;
		mr->mmkey.pd = to_mpd(pd)->pdn;
	}

	if (flags & IB_MR_REREG_ACCESS)
		mr->access_flags = access_flags;

	if (flags & IB_MR_REREG_TRANS) {
		atomic_sub(mr->npages, &dev->mdev->priv.reg_pages);
		set_mr_fileds(dev, mr, npages, len, access_flags);
		mr->mmkey.iova = addr;
		mr->mmkey.size = len;
	}
#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
	update_odp_mr(mr);
#endif

	return 0;
}

static int
mlx5_alloc_priv_descs(struct ib_device *device,
		      struct mlx5_ib_mr *mr,
		      int ndescs,
		      int desc_size)
{
	int size = ndescs * desc_size;
	int add_size;
	int ret;

	add_size = max_t(int, MLX5_UMR_ALIGN - 1, 0);

	mr->descs_alloc = kzalloc(size + add_size, GFP_KERNEL);
	if (!mr->descs_alloc)
		return -ENOMEM;

	mr->descs = PTR_ALIGN(mr->descs_alloc, MLX5_UMR_ALIGN);

	mr->desc_map = dma_map_single(device->dma_device, mr->descs,
				      size, DMA_TO_DEVICE);
	if (dma_mapping_error(device->dma_device, mr->desc_map)) {
		ret = -ENOMEM;
		goto err;
	}

	return 0;
err:
	kfree(mr->descs_alloc);

	return ret;
}

static void
mlx5_free_priv_descs(struct mlx5_ib_mr *mr)
{
	if (mr->descs) {
		struct ib_device *device = mr->ibmr.device;
		int size = mr->max_descs * mr->desc_size;

		dma_unmap_single(device->dma_device, mr->desc_map,
				 size, DMA_TO_DEVICE);
		kfree(mr->descs_alloc);
		mr->descs = NULL;
	}
}

static int clean_mr(struct mlx5_ib_mr *mr)
{
	struct mlx5_ib_dev *dev = to_mdev(mr->ibmr.device);
	int umred = mr->umred;
	int err;

	if (mr->sig) {
		if (mlx5_core_destroy_psv(dev->mdev,
					  mr->sig->psv_memory.psv_idx))
			mlx5_ib_warn(dev, "failed to destroy mem psv %d\n",
				     mr->sig->psv_memory.psv_idx);
		if (mlx5_core_destroy_psv(dev->mdev,
					  mr->sig->psv_wire.psv_idx))
			mlx5_ib_warn(dev, "failed to destroy wire psv %d\n",
				     mr->sig->psv_wire.psv_idx);
		kfree(mr->sig);
		mr->sig = NULL;
	}

	mlx5_free_priv_descs(mr);

	if (!umred) {
		err = destroy_mkey(dev, mr);
		if (err) {
			mlx5_ib_warn(dev, "failed to destroy mkey 0x%x (%d)\n",
				     mr->mmkey.key, err);
			return err;
		}
	} else {
		err = unreg_umr(dev, mr);
		if (err) {
			mlx5_ib_warn(dev, "failed unregister\n");
			return err;
		}
		free_cached_mr(dev, mr);
	}

	if (!umred)
		kfree(mr);

	return 0;
}

int mlx5_ib_dereg_mr(struct ib_mr *ibmr)
{
	struct mlx5_ib_dev *dev = to_mdev(ibmr->device);
	struct mlx5_ib_mr *mr = to_mmr(ibmr);
	int npages = mr->npages;
	struct ib_umem *umem = mr->umem;

#ifdef CONFIG_INFINIBAND_ON_DEMAND_PAGING
	if (umem && umem->odp_data) {
		/* Prevent new page faults from succeeding */
		mr->live = 0;
		/* Wait for all running page-fault handlers to finish. */
		synchronize_srcu(&dev->mr_srcu);
		/* Destroy all page mappings */
		mlx5_ib_invalidate_range(umem, ib_umem_start(umem),
					 ib_umem_end(umem));
		/*
		 * We kill the umem before the MR for ODP,
		 * so that there will not be any invalidations in
		 * flight, looking at the *mr struct.
		 */
		ib_umem_release(umem);
		atomic_sub(npages, &dev->mdev->priv.reg_pages);

		/* Avoid double-freeing the umem. */
		umem = NULL;
	}
#endif

	clean_mr(mr);

	if (umem) {
		ib_umem_release(umem);
		atomic_sub(npages, &dev->mdev->priv.reg_pages);
	}

	return 0;
}

struct ib_mr *mlx5_ib_alloc_mr(struct ib_pd *pd,
			       enum ib_mr_type mr_type,
			       u32 max_num_sg)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
	int ndescs = ALIGN(max_num_sg, 4);
	struct mlx5_ib_mr *mr;
	void *mkc;
	u32 *in;
	int err;

	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
	if (!mr)
		return ERR_PTR(-ENOMEM);

	in = kzalloc(inlen, GFP_KERNEL);
	if (!in) {
		err = -ENOMEM;
		goto err_free;
	}

	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);
	MLX5_SET(mkc, mkc, free, 1);
	MLX5_SET(mkc, mkc, translations_octword_size, ndescs);
	MLX5_SET(mkc, mkc, qpn, 0xffffff);
	MLX5_SET(mkc, mkc, pd, to_mpd(pd)->pdn);

	if (mr_type == IB_MR_TYPE_MEM_REG) {
		mr->access_mode = MLX5_ACCESS_MODE_MTT;
		MLX5_SET(mkc, mkc, log_page_size, PAGE_SHIFT);
		err = mlx5_alloc_priv_descs(pd->device, mr,
					    ndescs, sizeof(u64));
		if (err)
			goto err_free_in;

		mr->desc_size = sizeof(u64);
		mr->max_descs = ndescs;
	} else if (mr_type == IB_MR_TYPE_SG_GAPS) {
		mr->access_mode = MLX5_ACCESS_MODE_KLM;

		err = mlx5_alloc_priv_descs(pd->device, mr,
					    ndescs, sizeof(struct mlx5_klm));
		if (err)
			goto err_free_in;
		mr->desc_size = sizeof(struct mlx5_klm);
		mr->max_descs = ndescs;
	} else if (mr_type == IB_MR_TYPE_SIGNATURE) {
		u32 psv_index[2];

		MLX5_SET(mkc, mkc, bsf_en, 1);
		MLX5_SET(mkc, mkc, bsf_octword_size, MLX5_MKEY_BSF_OCTO_SIZE);
		mr->sig = kzalloc(sizeof(*mr->sig), GFP_KERNEL);
		if (!mr->sig) {
			err = -ENOMEM;
			goto err_free_in;
		}

		/* create mem & wire PSVs */
		err = mlx5_core_create_psv(dev->mdev, to_mpd(pd)->pdn,
					   2, psv_index);
		if (err)
			goto err_free_sig;

		mr->access_mode = MLX5_ACCESS_MODE_KLM;
		mr->sig->psv_memory.psv_idx = psv_index[0];
		mr->sig->psv_wire.psv_idx = psv_index[1];

		mr->sig->sig_status_checked = true;
		mr->sig->sig_err_exists = false;
		/* Next UMR, Arm SIGERR */
		++mr->sig->sigerr_count;
	} else {
		mlx5_ib_warn(dev, "Invalid mr type %d\n", mr_type);
		err = -EINVAL;
		goto err_free_in;
	}

	MLX5_SET(mkc, mkc, access_mode, mr->access_mode);
	MLX5_SET(mkc, mkc, umr_en, 1);

	err = mlx5_core_create_mkey(dev->mdev, &mr->mmkey, in, inlen);
	if (err)
		goto err_destroy_psv;

	mr->ibmr.lkey = mr->mmkey.key;
	mr->ibmr.rkey = mr->mmkey.key;
	mr->umem = NULL;
	kfree(in);

	return &mr->ibmr;

err_destroy_psv:
	if (mr->sig) {
		if (mlx5_core_destroy_psv(dev->mdev,
					  mr->sig->psv_memory.psv_idx))
			mlx5_ib_warn(dev, "failed to destroy mem psv %d\n",
				     mr->sig->psv_memory.psv_idx);
		if (mlx5_core_destroy_psv(dev->mdev,
					  mr->sig->psv_wire.psv_idx))
			mlx5_ib_warn(dev, "failed to destroy wire psv %d\n",
				     mr->sig->psv_wire.psv_idx);
	}
	mlx5_free_priv_descs(mr);
err_free_sig:
	kfree(mr->sig);
err_free_in:
	kfree(in);
err_free:
	kfree(mr);
	return ERR_PTR(err);
}

struct ib_mw *mlx5_ib_alloc_mw(struct ib_pd *pd, enum ib_mw_type type,
			       struct ib_udata *udata)
{
	struct mlx5_ib_dev *dev = to_mdev(pd->device);
	int inlen = MLX5_ST_SZ_BYTES(create_mkey_in);
	struct mlx5_ib_mw *mw = NULL;
	u32 *in = NULL;
	void *mkc;
	int ndescs;
	int err;
	struct mlx5_ib_alloc_mw req = {};
	struct {
		__u32	comp_mask;
		__u32	response_length;
	} resp = {};

	err = ib_copy_from_udata(&req, udata, min(udata->inlen, sizeof(req)));
	if (err)
		return ERR_PTR(err);

	if (req.comp_mask || req.reserved1 || req.reserved2)
		return ERR_PTR(-EOPNOTSUPP);

	if (udata->inlen > sizeof(req) &&
	    !ib_is_udata_cleared(udata, sizeof(req),
				 udata->inlen - sizeof(req)))
		return ERR_PTR(-EOPNOTSUPP);

	ndescs = req.num_klms ? roundup(req.num_klms, 4) : roundup(1, 4);

	mw = kzalloc(sizeof(*mw), GFP_KERNEL);
	in = kzalloc(inlen, GFP_KERNEL);
	if (!mw || !in) {
		err = -ENOMEM;
		goto free;
	}

	mkc = MLX5_ADDR_OF(create_mkey_in, in, memory_key_mkey_entry);

	MLX5_SET(mkc, mkc, free, 1);
	MLX5_SET(mkc, mkc, translations_octword_size, ndescs);
	MLX5_SET(mkc, mkc, pd, to_mpd(pd)->pdn);
	MLX5_SET(mkc, mkc, umr_en, 1);
	MLX5_SET(mkc, mkc, lr, 1);
	MLX5_SET(mkc, mkc, access_mode, MLX5_ACCESS_MODE_KLM);
	MLX5_SET(mkc, mkc, en_rinval, !!((type == IB_MW_TYPE_2)));
	MLX5_SET(mkc, mkc, qpn, 0xffffff);

	err = mlx5_core_create_mkey(dev->mdev, &mw->mmkey, in, inlen);
	if (err)
		goto free;

	mw->ibmw.rkey = mw->mmkey.key;

	resp.response_length = min(offsetof(typeof(resp), response_length) +
				   sizeof(resp.response_length), udata->outlen);
	if (resp.response_length) {
		err = ib_copy_to_udata(udata, &resp, resp.response_length);
		if (err) {
			mlx5_core_destroy_mkey(dev->mdev, &mw->mmkey);
			goto free;
		}
	}

	kfree(in);
	return &mw->ibmw;

free:
	kfree(mw);
	kfree(in);
	return ERR_PTR(err);
}

int mlx5_ib_dealloc_mw(struct ib_mw *mw)
{
	struct mlx5_ib_mw *mmw = to_mmw(mw);
	int err;

	err =  mlx5_core_destroy_mkey((to_mdev(mw->device))->mdev,
				      &mmw->mmkey);
	if (!err)
		kfree(mmw);
	return err;
}

int mlx5_ib_check_mr_status(struct ib_mr *ibmr, u32 check_mask,
			    struct ib_mr_status *mr_status)
{
	struct mlx5_ib_mr *mmr = to_mmr(ibmr);
	int ret = 0;

	if (check_mask & ~IB_MR_CHECK_SIG_STATUS) {
		pr_err("Invalid status check mask\n");
		ret = -EINVAL;
		goto done;
	}

	mr_status->fail_status = 0;
	if (check_mask & IB_MR_CHECK_SIG_STATUS) {
		if (!mmr->sig) {
			ret = -EINVAL;
			pr_err("signature status check requested on a non-signature enabled MR\n");
			goto done;
		}

		mmr->sig->sig_status_checked = true;
		if (!mmr->sig->sig_err_exists)
			goto done;

		if (ibmr->lkey == mmr->sig->err_item.key)
			memcpy(&mr_status->sig_err, &mmr->sig->err_item,
			       sizeof(mr_status->sig_err));
		else {
			mr_status->sig_err.err_type = IB_SIG_BAD_GUARD;
			mr_status->sig_err.sig_err_offset = 0;
			mr_status->sig_err.key = mmr->sig->err_item.key;
		}

		mmr->sig->sig_err_exists = false;
		mr_status->fail_status |= IB_MR_CHECK_SIG_STATUS;
	}

done:
	return ret;
}

static int
mlx5_ib_sg_to_klms(struct mlx5_ib_mr *mr,
		   struct scatterlist *sgl,
		   unsigned short sg_nents,
		   unsigned int *sg_offset_p)
{
	struct scatterlist *sg = sgl;
	struct mlx5_klm *klms = mr->descs;
	unsigned int sg_offset = sg_offset_p ? *sg_offset_p : 0;
	u32 lkey = mr->ibmr.pd->local_dma_lkey;
	int i;

	mr->ibmr.iova = sg_dma_address(sg) + sg_offset;
	mr->ibmr.length = 0;
	mr->ndescs = sg_nents;

	for_each_sg(sgl, sg, sg_nents, i) {
		if (unlikely(i > mr->max_descs))
			break;
		klms[i].va = cpu_to_be64(sg_dma_address(sg) + sg_offset);
		klms[i].bcount = cpu_to_be32(sg_dma_len(sg) - sg_offset);
		klms[i].key = cpu_to_be32(lkey);
		mr->ibmr.length += sg_dma_len(sg);

		sg_offset = 0;
	}

	if (sg_offset_p)
		*sg_offset_p = sg_offset;

	return i;
}

static int mlx5_set_page(struct ib_mr *ibmr, u64 addr)
{
	struct mlx5_ib_mr *mr = to_mmr(ibmr);
	__be64 *descs;

	if (unlikely(mr->ndescs == mr->max_descs))
		return -ENOMEM;

	descs = mr->descs;
	descs[mr->ndescs++] = cpu_to_be64(addr | MLX5_EN_RD | MLX5_EN_WR);

	return 0;
}

int mlx5_ib_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents,
		      unsigned int *sg_offset)
{
	struct mlx5_ib_mr *mr = to_mmr(ibmr);
	int n;

	mr->ndescs = 0;

	ib_dma_sync_single_for_cpu(ibmr->device, mr->desc_map,
				   mr->desc_size * mr->max_descs,
				   DMA_TO_DEVICE);

	if (mr->access_mode == MLX5_ACCESS_MODE_KLM)
		n = mlx5_ib_sg_to_klms(mr, sg, sg_nents, sg_offset);
	else
		n = ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset,
				mlx5_set_page);

	ib_dma_sync_single_for_device(ibmr->device, mr->desc_map,
				      mr->desc_size * mr->max_descs,
				      DMA_TO_DEVICE);

	return n;
}