dm-crypt.c

/*
 * Copyright (C) 2003 Christophe Saout <christophe@saout.de>
 * Copyright (C) 2004 Clemens Fruhwirth <clemens@endorphin.org>
 * Copyright (C) 2006-2009 Red Hat, Inc. All rights reserved.
 *
 * This file is released under the GPL.
 */

#include <linux/completion.h>
#include <linux/err.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include <linux/mempool.h>
#include <linux/slab.h>
#include <linux/crypto.h>
#include <linux/workqueue.h>
#include <linux/backing-dev.h>
#include <linux/percpu.h>
#include <asm/atomic.h>
#include <linux/scatterlist.h>
#include <asm/page.h>
#include <asm/unaligned.h>

#include <linux/device-mapper.h>

#define DM_MSG_PREFIX "crypt"
#define MESG_STR(x) x, sizeof(x)

/*
 * context holding the current state of a multi-part conversion
 */
struct convert_context {
	struct completion restart;
	struct bio *bio_in;
	struct bio *bio_out;
	unsigned int offset_in;
	unsigned int offset_out;
	unsigned int idx_in;
	unsigned int idx_out;
	sector_t sector;
	atomic_t pending;
};

/*
 * per bio private data
 */
struct dm_crypt_io {
	struct dm_target *target;
	struct bio *base_bio;
	struct work_struct work;

	struct convert_context ctx;

	atomic_t pending;
	int error;
	sector_t sector;
	struct dm_crypt_io *base_io;
};

struct dm_crypt_request {
	struct convert_context *ctx;
	struct scatterlist sg_in;
	struct scatterlist sg_out;
	sector_t iv_sector;
};

struct crypt_config;

struct crypt_iv_operations {
	int (*ctr)(struct crypt_config *cc, struct dm_target *ti,
		   const char *opts);
	void (*dtr)(struct crypt_config *cc);
	int (*init)(struct crypt_config *cc);
	int (*wipe)(struct crypt_config *cc);
	int (*generator)(struct crypt_config *cc, u8 *iv,
			 struct dm_crypt_request *dmreq);
	int (*post)(struct crypt_config *cc, u8 *iv,
		    struct dm_crypt_request *dmreq);
};

struct iv_essiv_private {
	struct crypto_hash *hash_tfm;
	u8 *salt;
};

struct iv_benbi_private {
	int shift;
};

/*
 * Crypt: maps a linear range of a block device
 * and encrypts / decrypts at the same time.
 */
enum flags { DM_CRYPT_SUSPENDED, DM_CRYPT_KEY_VALID };

/*
 * Duplicated per-CPU state for cipher.
 */
struct crypt_cpu {
	struct ablkcipher_request *req;
	struct crypto_ablkcipher *tfm;

	/* ESSIV: struct crypto_cipher *essiv_tfm */
	void *iv_private;
};

/*
 * The fields in here must be read only after initialization,
 * changing state should be in crypt_cpu.
 */
struct crypt_config {
	struct dm_dev *dev;
	sector_t start;

	/*
	 * pool for per bio private data, crypto requests and
	 * encryption requeusts/buffer pages
	 */
	mempool_t *io_pool;
	mempool_t *req_pool;
	mempool_t *page_pool;
	struct bio_set *bs;

	struct workqueue_struct *io_queue;
	struct workqueue_struct *crypt_queue;

	char *cipher;
	char *cipher_string;

	struct crypt_iv_operations *iv_gen_ops;
	union {
		struct iv_essiv_private essiv;
		struct iv_benbi_private benbi;
	} iv_gen_private;
	sector_t iv_offset;
	unsigned int iv_size;

	/*
	 * Duplicated per cpu state. Access through
	 * per_cpu_ptr() only.
	 */
	struct crypt_cpu __percpu *cpu;

	/*
	 * Layout of each crypto request:
	 *
	 *   struct ablkcipher_request
	 *      context
	 *      padding
	 *   struct dm_crypt_request
	 *      padding
	 *   IV
	 *
	 * The padding is added so that dm_crypt_request and the IV are
	 * correctly aligned.
	 */
	unsigned int dmreq_start;

	unsigned long flags;
	unsigned int key_size;
	u8 key[0];
};

#define MIN_IOS        16
#define MIN_POOL_PAGES 32
#define MIN_BIO_PAGES  8

static struct kmem_cache *_crypt_io_pool;

static void clone_init(struct dm_crypt_io *, struct bio *);
static void kcryptd_queue_crypt(struct dm_crypt_io *io);
static u8 *iv_of_dmreq(struct crypt_config *cc, struct dm_crypt_request *dmreq);

static struct crypt_cpu *this_crypt_config(struct crypt_config *cc)
{
	return this_cpu_ptr(cc->cpu);
}

/*
 * Use this to access cipher attributes that are the same for each CPU.
 */
static struct crypto_ablkcipher *any_tfm(struct crypt_config *cc)
{
	return __this_cpu_ptr(cc->cpu)->tfm;
}

/*
 * Different IV generation algorithms:
 *
 * plain: the initial vector is the 32-bit little-endian version of the sector
 *        number, padded with zeros if necessary.
 *
 * plain64: the initial vector is the 64-bit little-endian version of the sector
 *        number, padded with zeros if necessary.
 *
 * essiv: "encrypted sector|salt initial vector", the sector number is
 *        encrypted with the bulk cipher using a salt as key. The salt
 *        should be derived from the bulk cipher's key via hashing.
 *
 * benbi: the 64-bit "big-endian 'narrow block'-count", starting at 1
 *        (needed for LRW-32-AES and possible other narrow block modes)
 *
 * null: the initial vector is always zero.  Provides compatibility with
 *       obsolete loop_fish2 devices.  Do not use for new devices.
 *
 * plumb: unimplemented, see:
 * http://article.gmane.org/gmane.linux.kernel.device-mapper.dm-crypt/454
 */

static int crypt_iv_plain_gen(struct crypt_config *cc, u8 *iv,
			      struct dm_crypt_request *dmreq)
{
	memset(iv, 0, cc->iv_size);
	*(u32 *)iv = cpu_to_le32(dmreq->iv_sector & 0xffffffff);

	return 0;
}

static int crypt_iv_plain64_gen(struct crypt_config *cc, u8 *iv,
				struct dm_crypt_request *dmreq)
{
	memset(iv, 0, cc->iv_size);
	*(u64 *)iv = cpu_to_le64(dmreq->iv_sector);

	return 0;
}

/* Initialise ESSIV - compute salt but no local memory allocations */
static int crypt_iv_essiv_init(struct crypt_config *cc)
{
	struct iv_essiv_private *essiv = &cc->iv_gen_private.essiv;
	struct hash_desc desc;
	struct scatterlist sg;
	struct crypto_cipher *essiv_tfm;
	int err, cpu;

	sg_init_one(&sg, cc->key, cc->key_size);
	desc.tfm = essiv->hash_tfm;
	desc.flags = CRYPTO_TFM_REQ_MAY_SLEEP;

	err = crypto_hash_digest(&desc, &sg, cc->key_size, essiv->salt);
	if (err)
		return err;

	for_each_possible_cpu(cpu) {
		essiv_tfm = per_cpu_ptr(cc->cpu, cpu)->iv_private,

		err = crypto_cipher_setkey(essiv_tfm, essiv->salt,
				    crypto_hash_digestsize(essiv->hash_tfm));
		if (err)
			return err;
	}

	return 0;
}

/* Wipe salt and reset key derived from volume key */
static int crypt_iv_essiv_wipe(struct crypt_config *cc)
{
	struct iv_essiv_private *essiv = &cc->iv_gen_private.essiv;
	unsigned salt_size = crypto_hash_digestsize(essiv->hash_tfm);
	struct crypto_cipher *essiv_tfm;
	int cpu, r, err = 0;

	memset(essiv->salt, 0, salt_size);

	for_each_possible_cpu(cpu) {
		essiv_tfm = per_cpu_ptr(cc->cpu, cpu)->iv_private;
		r = crypto_cipher_setkey(essiv_tfm, essiv->salt, salt_size);
		if (r)
			err = r;
	}

	return err;
}

/* Set up per cpu cipher state */
static struct crypto_cipher *setup_essiv_cpu(struct crypt_config *cc,
					     struct dm_target *ti,
					     u8 *salt, unsigned saltsize)
{
	struct crypto_cipher *essiv_tfm;
	int err;

	/* Setup the essiv_tfm with the given salt */
	essiv_tfm = crypto_alloc_cipher(cc->cipher, 0, CRYPTO_ALG_ASYNC);
	if (IS_ERR(essiv_tfm)) {
		ti->error = "Error allocating crypto tfm for ESSIV";
		return essiv_tfm;
	}

	if (crypto_cipher_blocksize(essiv_tfm) !=
	    crypto_ablkcipher_ivsize(any_tfm(cc))) {
		ti->error = "Block size of ESSIV cipher does "
			    "not match IV size of block cipher";
		crypto_free_cipher(essiv_tfm);
		return ERR_PTR(-EINVAL);
	}

	err = crypto_cipher_setkey(essiv_tfm, salt, saltsize);
	if (err) {
		ti->error = "Failed to set key for ESSIV cipher";
		crypto_free_cipher(essiv_tfm);
		return ERR_PTR(err);
	}

	return essiv_tfm;
}

static void crypt_iv_essiv_dtr(struct crypt_config *cc)
{
	int cpu;
	struct crypt_cpu *cpu_cc;
	struct crypto_cipher *essiv_tfm;
	struct iv_essiv_private *essiv = &cc->iv_gen_private.essiv;

	crypto_free_hash(essiv->hash_tfm);
	essiv->hash_tfm = NULL;

	kzfree(essiv->salt);
	essiv->salt = NULL;

	for_each_possible_cpu(cpu) {
		cpu_cc = per_cpu_ptr(cc->cpu, cpu);
		essiv_tfm = cpu_cc->iv_private;

		if (essiv_tfm)
			crypto_free_cipher(essiv_tfm);

		cpu_cc->iv_private = NULL;
	}
}

static int crypt_iv_essiv_ctr(struct crypt_config *cc, struct dm_target *ti,
			      const char *opts)
{
	struct crypto_cipher *essiv_tfm = NULL;
	struct crypto_hash *hash_tfm = NULL;
	u8 *salt = NULL;
	int err, cpu;

	if (!opts) {
		ti->error = "Digest algorithm missing for ESSIV mode";
		return -EINVAL;
	}

	/* Allocate hash algorithm */
	hash_tfm = crypto_alloc_hash(opts, 0, CRYPTO_ALG_ASYNC);
	if (IS_ERR(hash_tfm)) {
		ti->error = "Error initializing ESSIV hash";
		err = PTR_ERR(hash_tfm);
		goto bad;
	}

	salt = kzalloc(crypto_hash_digestsize(hash_tfm), GFP_KERNEL);
	if (!salt) {
		ti->error = "Error kmallocing salt storage in ESSIV";
		err = -ENOMEM;
		goto bad;
	}

	cc->iv_gen_private.essiv.salt = salt;
	cc->iv_gen_private.essiv.hash_tfm = hash_tfm;

	for_each_possible_cpu(cpu) {
		essiv_tfm = setup_essiv_cpu(cc, ti, salt,
					crypto_hash_digestsize(hash_tfm));
		if (IS_ERR(essiv_tfm)) {
			crypt_iv_essiv_dtr(cc);
			return PTR_ERR(essiv_tfm);
		}
		per_cpu_ptr(cc->cpu, cpu)->iv_private = essiv_tfm;
	}

	return 0;

bad:
	if (hash_tfm && !IS_ERR(hash_tfm))
		crypto_free_hash(hash_tfm);
	kfree(salt);
	return err;
}

static int crypt_iv_essiv_gen(struct crypt_config *cc, u8 *iv,
			      struct dm_crypt_request *dmreq)
{
	struct crypto_cipher *essiv_tfm = this_crypt_config(cc)->iv_private;

	memset(iv, 0, cc->iv_size);
	*(u64 *)iv = cpu_to_le64(dmreq->iv_sector);
	crypto_cipher_encrypt_one(essiv_tfm, iv, iv);

	return 0;
}

static int crypt_iv_benbi_ctr(struct crypt_config *cc, struct dm_target *ti,
			      const char *opts)
{
	unsigned bs = crypto_ablkcipher_blocksize(any_tfm(cc));
	int log = ilog2(bs);

	/* we need to calculate how far we must shift the sector count
	 * to get the cipher block count, we use this shift in _gen */

	if (1 << log != bs) {
		ti->error = "cypher blocksize is not a power of 2";
		return -EINVAL;
	}

	if (log > 9) {
		ti->error = "cypher blocksize is > 512";
		return -EINVAL;
	}

	cc->iv_gen_private.benbi.shift = 9 - log;

	return 0;
}

static void crypt_iv_benbi_dtr(struct crypt_config *cc)
{
}

static int crypt_iv_benbi_gen(struct crypt_config *cc, u8 *iv,
			      struct dm_crypt_request *dmreq)
{
	__be64 val;

	memset(iv, 0, cc->iv_size - sizeof(u64)); /* rest is cleared below */

	val = cpu_to_be64(((u64)dmreq->iv_sector << cc->iv_gen_private.benbi.shift) + 1);
	put_unaligned(val, (__be64 *)(iv + cc->iv_size - sizeof(u64)));

	return 0;
}

static int crypt_iv_null_gen(struct crypt_config *cc, u8 *iv,
			     struct dm_crypt_request *dmreq)
{
	memset(iv, 0, cc->iv_size);

	return 0;
}

static struct crypt_iv_operations crypt_iv_plain_ops = {
	.generator = crypt_iv_plain_gen
};

static struct crypt_iv_operations crypt_iv_plain64_ops = {
	.generator = crypt_iv_plain64_gen
};

static struct crypt_iv_operations crypt_iv_essiv_ops = {
	.ctr       = crypt_iv_essiv_ctr,
	.dtr       = crypt_iv_essiv_dtr,
	.init      = crypt_iv_essiv_init,
	.wipe      = crypt_iv_essiv_wipe,
	.generator = crypt_iv_essiv_gen
};

static struct crypt_iv_operations crypt_iv_benbi_ops = {
	.ctr	   = crypt_iv_benbi_ctr,
	.dtr	   = crypt_iv_benbi_dtr,
	.generator = crypt_iv_benbi_gen
};

static struct crypt_iv_operations crypt_iv_null_ops = {
	.generator = crypt_iv_null_gen
};

static void crypt_convert_init(struct crypt_config *cc,
			       struct convert_context *ctx,
			       struct bio *bio_out, struct bio *bio_in,
			       sector_t sector)
{
	ctx->bio_in = bio_in;
	ctx->bio_out = bio_out;
	ctx->offset_in = 0;
	ctx->offset_out = 0;
	ctx->idx_in = bio_in ? bio_in->bi_idx : 0;
	ctx->idx_out = bio_out ? bio_out->bi_idx : 0;
	ctx->sector = sector + cc->iv_offset;
	init_completion(&ctx->restart);
}

static struct dm_crypt_request *dmreq_of_req(struct crypt_config *cc,
					     struct ablkcipher_request *req)
{
	return (struct dm_crypt_request *)((char *)req + cc->dmreq_start);
}

static struct ablkcipher_request *req_of_dmreq(struct crypt_config *cc,
					       struct dm_crypt_request *dmreq)
{
	return (struct ablkcipher_request *)((char *)dmreq - cc->dmreq_start);
}

static u8 *iv_of_dmreq(struct crypt_config *cc,
		       struct dm_crypt_request *dmreq)
{
	return (u8 *)ALIGN((unsigned long)(dmreq + 1),
		crypto_ablkcipher_alignmask(any_tfm(cc)) + 1);
}

static int crypt_convert_block(struct crypt_config *cc,
			       struct convert_context *ctx,
			       struct ablkcipher_request *req)
{
	struct bio_vec *bv_in = bio_iovec_idx(ctx->bio_in, ctx->idx_in);
	struct bio_vec *bv_out = bio_iovec_idx(ctx->bio_out, ctx->idx_out);
	struct dm_crypt_request *dmreq;
	u8 *iv;
	int r = 0;

	dmreq = dmreq_of_req(cc, req);
	iv = iv_of_dmreq(cc, dmreq);

	dmreq->iv_sector = ctx->sector;
	dmreq->ctx = ctx;
	sg_init_table(&dmreq->sg_in, 1);
	sg_set_page(&dmreq->sg_in, bv_in->bv_page, 1 << SECTOR_SHIFT,
		    bv_in->bv_offset + ctx->offset_in);

	sg_init_table(&dmreq->sg_out, 1);
	sg_set_page(&dmreq->sg_out, bv_out->bv_page, 1 << SECTOR_SHIFT,
		    bv_out->bv_offset + ctx->offset_out);

	ctx->offset_in += 1 << SECTOR_SHIFT;
	if (ctx->offset_in >= bv_in->bv_len) {
		ctx->offset_in = 0;
		ctx->idx_in++;
	}

	ctx->offset_out += 1 << SECTOR_SHIFT;
	if (ctx->offset_out >= bv_out->bv_len) {
		ctx->offset_out = 0;
		ctx->idx_out++;
	}

	if (cc->iv_gen_ops) {
		r = cc->iv_gen_ops->generator(cc, iv, dmreq);
		if (r < 0)
			return r;
	}

	ablkcipher_request_set_crypt(req, &dmreq->sg_in, &dmreq->sg_out,
				     1 << SECTOR_SHIFT, iv);

	if (bio_data_dir(ctx->bio_in) == WRITE)
		r = crypto_ablkcipher_encrypt(req);
	else
		r = crypto_ablkcipher_decrypt(req);

	if (!r && cc->iv_gen_ops && cc->iv_gen_ops->post)
		r = cc->iv_gen_ops->post(cc, iv, dmreq);

	return r;
}

static void kcryptd_async_done(struct crypto_async_request *async_req,
			       int error);

static void crypt_alloc_req(struct crypt_config *cc,
			    struct convert_context *ctx)
{
	struct crypt_cpu *this_cc = this_crypt_config(cc);

	if (!this_cc->req)
		this_cc->req = mempool_alloc(cc->req_pool, GFP_NOIO);

	ablkcipher_request_set_tfm(this_cc->req, this_cc->tfm);
	ablkcipher_request_set_callback(this_cc->req,
	    CRYPTO_TFM_REQ_MAY_BACKLOG | CRYPTO_TFM_REQ_MAY_SLEEP,
	    kcryptd_async_done, dmreq_of_req(cc, this_cc->req));
}

/*
 * Encrypt / decrypt data from one bio to another one (can be the same one)
 */
static int crypt_convert(struct crypt_config *cc,
			 struct convert_context *ctx)
{
	struct crypt_cpu *this_cc = this_crypt_config(cc);
	int r;

	atomic_set(&ctx->pending, 1);

	while(ctx->idx_in < ctx->bio_in->bi_vcnt &&
	      ctx->idx_out < ctx->bio_out->bi_vcnt) {

		crypt_alloc_req(cc, ctx);

		atomic_inc(&ctx->pending);

		r = crypt_convert_block(cc, ctx, this_cc->req);

		switch (r) {
		/* async */
		case -EBUSY:
			wait_for_completion(&ctx->restart);
			INIT_COMPLETION(ctx->restart);
			/* fall through*/
		case -EINPROGRESS:
			this_cc->req = NULL;
			ctx->sector++;
			continue;

		/* sync */
		case 0:
			atomic_dec(&ctx->pending);
			ctx->sector++;
			cond_resched();
			continue;

		/* error */
		default:
			atomic_dec(&ctx->pending);
			return r;
		}
	}

	return 0;
}

static void dm_crypt_bio_destructor(struct bio *bio)
{
	struct dm_crypt_io *io = bio->bi_private;
	struct crypt_config *cc = io->target->private;

	bio_free(bio, cc->bs);
}

/*
 * Generate a new unfragmented bio with the given size
 * This should never violate the device limitations
 * May return a smaller bio when running out of pages, indicated by
 * *out_of_pages set to 1.
 */
static struct bio *crypt_alloc_buffer(struct dm_crypt_io *io, unsigned size,
				      unsigned *out_of_pages)
{
	struct crypt_config *cc = io->target->private;
	struct bio *clone;
	unsigned int nr_iovecs = (size + PAGE_SIZE - 1) >> PAGE_SHIFT;
	gfp_t gfp_mask = GFP_NOIO | __GFP_HIGHMEM;
	unsigned i, len;
	struct page *page;

	clone = bio_alloc_bioset(GFP_NOIO, nr_iovecs, cc->bs);
	if (!clone)
		return NULL;

	clone_init(io, clone);
	*out_of_pages = 0;

	for (i = 0; i < nr_iovecs; i++) {
		page = mempool_alloc(cc->page_pool, gfp_mask);
		if (!page) {
			*out_of_pages = 1;
			break;
		}

		/*
		 * if additional pages cannot be allocated without waiting,
		 * return a partially allocated bio, the caller will then try
		 * to allocate additional bios while submitting this partial bio
		 */
		if (i == (MIN_BIO_PAGES - 1))
			gfp_mask = (gfp_mask | __GFP_NOWARN) & ~__GFP_WAIT;

		len = (size > PAGE_SIZE) ? PAGE_SIZE : size;

		if (!bio_add_page(clone, page, len, 0)) {
			mempool_free(page, cc->page_pool);
			break;
		}

		size -= len;
	}

	if (!clone->bi_size) {
		bio_put(clone);
		return NULL;
	}

	return clone;
}

static void crypt_free_buffer_pages(struct crypt_config *cc, struct bio *clone)
{
	unsigned int i;
	struct bio_vec *bv;

	for (i = 0; i < clone->bi_vcnt; i++) {
		bv = bio_iovec_idx(clone, i);
		BUG_ON(!bv->bv_page);
		mempool_free(bv->bv_page, cc->page_pool);
		bv->bv_page = NULL;
	}
}

static struct dm_crypt_io *crypt_io_alloc(struct dm_target *ti,
					  struct bio *bio, sector_t sector)
{
	struct crypt_config *cc = ti->private;
	struct dm_crypt_io *io;

	io = mempool_alloc(cc->io_pool, GFP_NOIO);
	io->target = ti;
	io->base_bio = bio;
	io->sector = sector;
	io->error = 0;
	io->base_io = NULL;
	atomic_set(&io->pending, 0);

	return io;
}

static void crypt_inc_pending(struct dm_crypt_io *io)
{
	atomic_inc(&io->pending);
}

/*
 * One of the bios was finished. Check for completion of
 * the whole request and correctly clean up the buffer.
 * If base_io is set, wait for the last fragment to complete.
 */
static void crypt_dec_pending(struct dm_crypt_io *io)
{
	struct crypt_config *cc = io->target->private;
	struct bio *base_bio = io->base_bio;
	struct dm_crypt_io *base_io = io->base_io;
	int error = io->error;

	if (!atomic_dec_and_test(&io->pending))
		return;

	mempool_free(io, cc->io_pool);

	if (likely(!base_io))
		bio_endio(base_bio, error);
	else {
		if (error && !base_io->error)
			base_io->error = error;
		crypt_dec_pending(base_io);
	}
}

/*
 * kcryptd/kcryptd_io:
 *
 * Needed because it would be very unwise to do decryption in an
 * interrupt context.
 *
 * kcryptd performs the actual encryption or decryption.
 *
 * kcryptd_io performs the IO submission.
 *
 * They must be separated as otherwise the final stages could be
 * starved by new requests which can block in the first stages due
 * to memory allocation.
 *
 * The work is done per CPU global for all dm-crypt instances.
 * They should not depend on each other and do not block.
 */
static void crypt_endio(struct bio *clone, int error)
{
	struct dm_crypt_io *io = clone->bi_private;
	struct crypt_config *cc = io->target->private;
	unsigned rw = bio_data_dir(clone);

	if (unlikely(!bio_flagged(clone, BIO_UPTODATE) && !error))
		error = -EIO;

	/*
	 * free the processed pages
	 */
	if (rw == WRITE)
		crypt_free_buffer_pages(cc, clone);

	bio_put(clone);

	if (rw == READ && !error) {
		kcryptd_queue_crypt(io);
		return;
	}

	if (unlikely(error))
		io->error = error;

	crypt_dec_pending(io);
}

static void clone_init(struct dm_crypt_io *io, struct bio *clone)
{
	struct crypt_config *cc = io->target->private;

	clone->bi_private = io;
	clone->bi_end_io  = crypt_endio;
	clone->bi_bdev    = cc->dev->bdev;
	clone->bi_rw      = io->base_bio->bi_rw;
	clone->bi_destructor = dm_crypt_bio_destructor;
}

static void kcryptd_unplug(struct crypt_config *cc)
{
	blk_unplug(bdev_get_queue(cc->dev->bdev));
}

static int kcryptd_io_read(struct dm_crypt_io *io, gfp_t gfp)
{
	struct crypt_config *cc = io->target->private;
	struct bio *base_bio = io->base_bio;
	struct bio *clone;

	/*
	 * The block layer might modify the bvec array, so always
	 * copy the required bvecs because we need the original
	 * one in order to decrypt the whole bio data *afterwards*.
	 */
	clone = bio_alloc_bioset(gfp, bio_segments(base_bio), cc->bs);
	if (!clone) {
		kcryptd_unplug(cc);
		return 1;
	}

	crypt_inc_pending(io);

	clone_init(io, clone);
	clone->bi_idx = 0;
	clone->bi_vcnt = bio_segments(base_bio);
	clone->bi_size = base_bio->bi_size;
	clone->bi_sector = cc->start + io->sector;
	memcpy(clone->bi_io_vec, bio_iovec(base_bio),
	       sizeof(struct bio_vec) * clone->bi_vcnt);

	generic_make_request(clone);
	return 0;
}

static void kcryptd_io_write(struct dm_crypt_io *io)
{
	struct bio *clone = io->ctx.bio_out;
	generic_make_request(clone);
}

static void kcryptd_io(struct work_struct *work)
{
	struct dm_crypt_io *io = container_of(work, struct dm_crypt_io, work);

	if (bio_data_dir(io->base_bio) == READ) {
		crypt_inc_pending(io);
		if (kcryptd_io_read(io, GFP_NOIO))
			io->error = -ENOMEM;
		crypt_dec_pending(io);
	} else
		kcryptd_io_write(io);
}

static void kcryptd_queue_io(struct dm_crypt_io *io)
{
	struct crypt_config *cc = io->target->private;

	INIT_WORK(&io->work, kcryptd_io);
	queue_work(cc->io_queue, &io->work);
}

static void kcryptd_crypt_write_io_submit(struct dm_crypt_io *io,
					  int error, int async)
{
	struct bio *clone = io->ctx.bio_out;
	struct crypt_config *cc = io->target->private;

	if (unlikely(error < 0)) {
		crypt_free_buffer_pages(cc, clone);
		bio_put(clone);
		io->error = -EIO;
		crypt_dec_pending(io);
		return;
	}

	/* crypt_convert should have filled the clone bio */
	BUG_ON(io->ctx.idx_out < clone->bi_vcnt);

	clone->bi_sector = cc->start + io->sector;

	if (async)
		kcryptd_queue_io(io);
	else
		generic_make_request(clone);
}

static void kcryptd_crypt_write_convert(struct dm_crypt_io *io)
{
	struct crypt_config *cc = io->target->private;
	struct bio *clone;
	struct dm_crypt_io *new_io;
	int crypt_finished;
	unsigned out_of_pages = 0;
	unsigned remaining = io->base_bio->bi_size;
	sector_t sector = io->sector;
	int r;

	/*
	 * Prevent io from disappearing until this function completes.
	 */
	crypt_inc_pending(io);
	crypt_convert_init(cc, &io->ctx, NULL, io->base_bio, sector);

	/*
	 * The allocated buffers can be smaller than the whole bio,
	 * so repeat the whole process until all the data can be handled.
	 */
	while (remaining) {
		clone = crypt_alloc_buffer(io, remaining, &out_of_pages);
		if (unlikely(!clone)) {
			io->error = -ENOMEM;
			break;
		}

		io->ctx.bio_out = clone;
		io->ctx.idx_out = 0;

		remaining -= clone->bi_size;
		sector += bio_sectors(clone);

		crypt_inc_pending(io);
		r = crypt_convert(cc, &io->ctx);
		crypt_finished = atomic_dec_and_test(&io->ctx.pending);

		/* Encryption was already finished, submit io now */
		if (crypt_finished) {
			kcryptd_crypt_write_io_submit(io, r, 0);

			/*
			 * If there was an error, do not try next fragments.
			 * For async, error is processed in async handler.
			 */
			if (unlikely(r < 0))
				break;

			io->sector = sector;
		}

		/*
		 * Out of memory -> run queues
		 * But don't wait if split was due to the io size restriction
		 */
		if (unlikely(out_of_pages))
			congestion_wait(BLK_RW_ASYNC, HZ/100);

		/*
		 * With async crypto it is unsafe to share the crypto context
		 * between fragments, so switch to a new dm_crypt_io structure.
		 */
		if (unlikely(!crypt_finished && remaining)) {
			new_io = crypt_io_alloc(io->target, io->base_bio,
						sector);
			crypt_inc_pending(new_io);
			crypt_convert_init(cc, &new_io->ctx, NULL,
					   io->base_bio, sector);
			new_io->ctx.idx_in = io->ctx.idx_in;
			new_io->ctx.offset_in = io->ctx.offset_in;

			/*
			 * Fragments after the first use the base_io
			 * pending count.
			 */
			if (!io->base_io)
				new_io->base_io = io;
			else {
				new_io->base_io = io->base_io;
				crypt_inc_pending(io->base_io);
				crypt_dec_pending(io);
			}

			io = new_io;
		}
	}

	crypt_dec_pending(io);
}

static void kcryptd_crypt_read_done(struct dm_crypt_io *io, int error)
{
	if (unlikely(error < 0))
		io->error = -EIO;

	crypt_dec_pending(io);
}

static void kcryptd_crypt_read_convert(struct dm_crypt_io *io)
{
	struct crypt_config *cc = io->target->private;
	int r = 0;