[mirror_zfs-debian.git] / module / zfs / zio_checksum.c

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */
/*
 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
 * Copyright (c) 2013, 2016 by Delphix. All rights reserved.
 * Copyright 2013 Saso Kiselkov. All rights reserved.
 */

#include <sys/zfs_context.h>
#include <sys/spa.h>
#include <sys/spa_impl.h>
#include <sys/zio.h>
#include <sys/zio_checksum.h>
#include <sys/zil.h>
#include <sys/abd.h>
#include <zfs_fletcher.h>

/*
 * Checksum vectors.
 *
 * In the SPA, everything is checksummed.  We support checksum vectors
 * for three distinct reasons:
 *
 *   1. Different kinds of data need different levels of protection.
 *	For SPA metadata, we always want a very strong checksum.
 *	For user data, we let users make the trade-off between speed
 *	and checksum strength.
 *
 *   2. Cryptographic hash and MAC algorithms are an area of active research.
 *	It is likely that in future hash functions will be at least as strong
 *	as current best-of-breed, and may be substantially faster as well.
 *	We want the ability to take advantage of these new hashes as soon as
 *	they become available.
 *
 *   3. If someone develops hardware that can compute a strong hash quickly,
 *	we want the ability to take advantage of that hardware.
 *
 * Of course, we don't want a checksum upgrade to invalidate existing
 * data, so we store the checksum *function* in eight bits of the bp.
 * This gives us room for up to 256 different checksum functions.
 *
 * When writing a block, we always checksum it with the latest-and-greatest
 * checksum function of the appropriate strength.  When reading a block,
 * we compare the expected checksum against the actual checksum, which we
 * compute via the checksum function specified by BP_GET_CHECKSUM(bp).
 *
 * SALTED CHECKSUMS
 *
 * To enable the use of less secure hash algorithms with dedup, we
 * introduce the notion of salted checksums (MACs, really).  A salted
 * checksum is fed both a random 256-bit value (the salt) and the data
 * to be checksummed.  This salt is kept secret (stored on the pool, but
 * never shown to the user).  Thus even if an attacker knew of collision
 * weaknesses in the hash algorithm, they won't be able to mount a known
 * plaintext attack on the DDT, since the actual hash value cannot be
 * known ahead of time.  How the salt is used is algorithm-specific
 * (some might simply prefix it to the data block, others might need to
 * utilize a full-blown HMAC).  On disk the salt is stored in a ZAP
 * object in the MOS (DMU_POOL_CHECKSUM_SALT).
 *
 * CONTEXT TEMPLATES
 *
 * Some hashing algorithms need to perform a substantial amount of
 * initialization work (e.g. salted checksums above may need to pre-hash
 * the salt) before being able to process data.  Performing this
 * redundant work for each block would be wasteful, so we instead allow
 * a checksum algorithm to do the work once (the first time it's used)
 * and then keep this pre-initialized context as a template inside the
 * spa_t (spa_cksum_tmpls).  If the zio_checksum_info_t contains
 * non-NULL ci_tmpl_init and ci_tmpl_free callbacks, they are used to
 * construct and destruct the pre-initialized checksum context.  The
 * pre-initialized context is then reused during each checksum
 * invocation and passed to the checksum function.
 */

/*ARGSUSED*/
static void
abd_checksum_off(abd_t *abd, uint64_t size,
    const void *ctx_template, zio_cksum_t *zcp)
{
	ZIO_SET_CHECKSUM(zcp, 0, 0, 0, 0);
}

/*ARGSUSED*/
void
abd_fletcher_2_native(abd_t *abd, uint64_t size,
    const void *ctx_template, zio_cksum_t *zcp)
{
	fletcher_init(zcp);
	(void) abd_iterate_func(abd, 0, size,
	    fletcher_2_incremental_native, zcp);
}

/*ARGSUSED*/
void
abd_fletcher_2_byteswap(abd_t *abd, uint64_t size,
    const void *ctx_template, zio_cksum_t *zcp)
{
	fletcher_init(zcp);
	(void) abd_iterate_func(abd, 0, size,
	    fletcher_2_incremental_byteswap, zcp);
}

static inline void
abd_fletcher_4_impl(abd_t *abd, uint64_t size, zio_abd_checksum_data_t *acdp)
{
	fletcher_4_abd_ops.acf_init(acdp);
	abd_iterate_func(abd, 0, size, fletcher_4_abd_ops.acf_iter, acdp);
	fletcher_4_abd_ops.acf_fini(acdp);
}

/*ARGSUSED*/
void
abd_fletcher_4_native(abd_t *abd, uint64_t size,
    const void *ctx_template, zio_cksum_t *zcp)
{
	fletcher_4_ctx_t ctx;

	zio_abd_checksum_data_t acd = {
		.acd_byteorder	= ZIO_CHECKSUM_NATIVE,
		.acd_zcp 	= zcp,
		.acd_ctx	= &ctx
	};

	abd_fletcher_4_impl(abd, size, &acd);

}

/*ARGSUSED*/
void
abd_fletcher_4_byteswap(abd_t *abd, uint64_t size,
    const void *ctx_template, zio_cksum_t *zcp)
{
	fletcher_4_ctx_t ctx;

	zio_abd_checksum_data_t acd = {
		.acd_byteorder	= ZIO_CHECKSUM_BYTESWAP,
		.acd_zcp 	= zcp,
		.acd_ctx	= &ctx
	};

	abd_fletcher_4_impl(abd, size, &acd);
}

zio_checksum_info_t zio_checksum_table[ZIO_CHECKSUM_FUNCTIONS] = {
	{{NULL, NULL}, NULL, NULL, 0, "inherit"},
	{{NULL, NULL}, NULL, NULL, 0, "on"},
	{{abd_checksum_off,		abd_checksum_off},
	    NULL, NULL, 0, "off"},
	{{abd_checksum_SHA256,		abd_checksum_SHA256},
	    NULL, NULL, ZCHECKSUM_FLAG_METADATA | ZCHECKSUM_FLAG_EMBEDDED,
	    "label"},
	{{abd_checksum_SHA256,		abd_checksum_SHA256},
	    NULL, NULL, ZCHECKSUM_FLAG_METADATA | ZCHECKSUM_FLAG_EMBEDDED,
	    "gang_header"},
	{{abd_fletcher_2_native,	abd_fletcher_2_byteswap},
	    NULL, NULL, ZCHECKSUM_FLAG_EMBEDDED, "zilog"},
	{{abd_fletcher_2_native,	abd_fletcher_2_byteswap},
	    NULL, NULL, 0, "fletcher2"},
	{{abd_fletcher_4_native,	abd_fletcher_4_byteswap},
	    NULL, NULL, ZCHECKSUM_FLAG_METADATA, "fletcher4"},
	{{abd_checksum_SHA256,		abd_checksum_SHA256},
	    NULL, NULL, ZCHECKSUM_FLAG_METADATA | ZCHECKSUM_FLAG_DEDUP |
	    ZCHECKSUM_FLAG_NOPWRITE, "sha256"},
	{{abd_fletcher_4_native,	abd_fletcher_4_byteswap},
	    NULL, NULL, ZCHECKSUM_FLAG_EMBEDDED, "zilog2"},
	{{abd_checksum_off,		abd_checksum_off},
	    NULL, NULL, 0, "noparity"},
	{{abd_checksum_SHA512_native,	abd_checksum_SHA512_byteswap},
	    NULL, NULL, ZCHECKSUM_FLAG_METADATA | ZCHECKSUM_FLAG_DEDUP |
	    ZCHECKSUM_FLAG_NOPWRITE, "sha512"},
	{{abd_checksum_skein_native,	abd_checksum_skein_byteswap},
	    abd_checksum_skein_tmpl_init, abd_checksum_skein_tmpl_free,
	    ZCHECKSUM_FLAG_METADATA | ZCHECKSUM_FLAG_DEDUP |
	    ZCHECKSUM_FLAG_SALTED | ZCHECKSUM_FLAG_NOPWRITE, "skein"},
	{{abd_checksum_edonr_native,	abd_checksum_edonr_byteswap},
	    abd_checksum_edonr_tmpl_init, abd_checksum_edonr_tmpl_free,
	    ZCHECKSUM_FLAG_METADATA | ZCHECKSUM_FLAG_SALTED |
	    ZCHECKSUM_FLAG_NOPWRITE, "edonr"},
};

/*
 * The flag corresponding to the "verify" in dedup=[checksum,]verify
 * must be cleared first, so callers should use ZIO_CHECKSUM_MASK.
 */
spa_feature_t
zio_checksum_to_feature(enum zio_checksum cksum)
{
	VERIFY((cksum & ~ZIO_CHECKSUM_MASK) == 0);

	switch (cksum) {
	case ZIO_CHECKSUM_SHA512:
		return (SPA_FEATURE_SHA512);
	case ZIO_CHECKSUM_SKEIN:
		return (SPA_FEATURE_SKEIN);
	case ZIO_CHECKSUM_EDONR:
		return (SPA_FEATURE_EDONR);
	default:
		return (SPA_FEATURE_NONE);
	}
}

enum zio_checksum
zio_checksum_select(enum zio_checksum child, enum zio_checksum parent)
{
	ASSERT(child < ZIO_CHECKSUM_FUNCTIONS);
	ASSERT(parent < ZIO_CHECKSUM_FUNCTIONS);
	ASSERT(parent != ZIO_CHECKSUM_INHERIT && parent != ZIO_CHECKSUM_ON);

	if (child == ZIO_CHECKSUM_INHERIT)
		return (parent);

	if (child == ZIO_CHECKSUM_ON)
		return (ZIO_CHECKSUM_ON_VALUE);

	return (child);
}

enum zio_checksum
zio_checksum_dedup_select(spa_t *spa, enum zio_checksum child,
    enum zio_checksum parent)
{
	ASSERT((child & ZIO_CHECKSUM_MASK) < ZIO_CHECKSUM_FUNCTIONS);
	ASSERT((parent & ZIO_CHECKSUM_MASK) < ZIO_CHECKSUM_FUNCTIONS);
	ASSERT(parent != ZIO_CHECKSUM_INHERIT && parent != ZIO_CHECKSUM_ON);

	if (child == ZIO_CHECKSUM_INHERIT)
		return (parent);

	if (child == ZIO_CHECKSUM_ON)
		return (spa_dedup_checksum(spa));

	if (child == (ZIO_CHECKSUM_ON | ZIO_CHECKSUM_VERIFY))
		return (spa_dedup_checksum(spa) | ZIO_CHECKSUM_VERIFY);

	ASSERT((zio_checksum_table[child & ZIO_CHECKSUM_MASK].ci_flags &
	    ZCHECKSUM_FLAG_DEDUP) ||
	    (child & ZIO_CHECKSUM_VERIFY) || child == ZIO_CHECKSUM_OFF);

	return (child);
}

/*
 * Set the external verifier for a gang block based on <vdev, offset, txg>,
 * a tuple which is guaranteed to be unique for the life of the pool.
 */
static void
zio_checksum_gang_verifier(zio_cksum_t *zcp, const blkptr_t *bp)
{
	const dva_t *dva = BP_IDENTITY(bp);
	uint64_t txg = BP_PHYSICAL_BIRTH(bp);

	ASSERT(BP_IS_GANG(bp));

	ZIO_SET_CHECKSUM(zcp, DVA_GET_VDEV(dva), DVA_GET_OFFSET(dva), txg, 0);
}

/*
 * Set the external verifier for a label block based on its offset.
 * The vdev is implicit, and the txg is unknowable at pool open time --
 * hence the logic in vdev_uberblock_load() to find the most recent copy.
 */
static void
zio_checksum_label_verifier(zio_cksum_t *zcp, uint64_t offset)
{
	ZIO_SET_CHECKSUM(zcp, offset, 0, 0, 0);
}

/*
 * Calls the template init function of a checksum which supports context
 * templates and installs the template into the spa_t.
 */
static void
zio_checksum_template_init(enum zio_checksum checksum, spa_t *spa)
{
	zio_checksum_info_t *ci = &zio_checksum_table[checksum];

	if (ci->ci_tmpl_init == NULL)
		return;
	if (spa->spa_cksum_tmpls[checksum] != NULL)
		return;

	VERIFY(ci->ci_tmpl_free != NULL);
	mutex_enter(&spa->spa_cksum_tmpls_lock);
	if (spa->spa_cksum_tmpls[checksum] == NULL) {
		spa->spa_cksum_tmpls[checksum] =
		    ci->ci_tmpl_init(&spa->spa_cksum_salt);
		VERIFY(spa->spa_cksum_tmpls[checksum] != NULL);
	}
	mutex_exit(&spa->spa_cksum_tmpls_lock);
}

/*
 * Generate the checksum.
 */
void
zio_checksum_compute(zio_t *zio, enum zio_checksum checksum,
    abd_t *abd, uint64_t size)
{
	static const uint64_t zec_magic = ZEC_MAGIC;
	blkptr_t *bp = zio->io_bp;
	uint64_t offset = zio->io_offset;
	zio_checksum_info_t *ci = &zio_checksum_table[checksum];
	zio_cksum_t cksum;
	spa_t *spa = zio->io_spa;

	ASSERT((uint_t)checksum < ZIO_CHECKSUM_FUNCTIONS);
	ASSERT(ci->ci_func[0] != NULL);

	zio_checksum_template_init(checksum, spa);

	if (ci->ci_flags & ZCHECKSUM_FLAG_EMBEDDED) {
		zio_eck_t eck;
		size_t eck_offset;

		if (checksum == ZIO_CHECKSUM_ZILOG2) {
			zil_chain_t zilc;
			abd_copy_to_buf(&zilc, abd, sizeof (zil_chain_t));

			size = P2ROUNDUP_TYPED(zilc.zc_nused, ZIL_MIN_BLKSZ,
			    uint64_t);
			eck = zilc.zc_eck;
			eck_offset = offsetof(zil_chain_t, zc_eck);
		} else {
			eck_offset = size - sizeof (zio_eck_t);
			abd_copy_to_buf_off(&eck, abd, eck_offset,
			    sizeof (zio_eck_t));
		}

		if (checksum == ZIO_CHECKSUM_GANG_HEADER) {
			zio_checksum_gang_verifier(&eck.zec_cksum, bp);
			abd_copy_from_buf_off(abd, &eck.zec_cksum,
			    eck_offset + offsetof(zio_eck_t, zec_cksum),
			    sizeof (zio_cksum_t));
		} else if (checksum == ZIO_CHECKSUM_LABEL) {
			zio_checksum_label_verifier(&eck.zec_cksum, offset);
			abd_copy_from_buf_off(abd, &eck.zec_cksum,
			    eck_offset + offsetof(zio_eck_t, zec_cksum),
			    sizeof (zio_cksum_t));
		} else {
			bp->blk_cksum = eck.zec_cksum;
		}

		abd_copy_from_buf_off(abd, &zec_magic,
		    eck_offset + offsetof(zio_eck_t, zec_magic),
		    sizeof (zec_magic));

		ci->ci_func[0](abd, size, spa->spa_cksum_tmpls[checksum],
		    &cksum);

		abd_copy_from_buf_off(abd, &cksum,
		    eck_offset + offsetof(zio_eck_t, zec_cksum),
		    sizeof (zio_cksum_t));
	} else {
		ci->ci_func[0](abd, size, spa->spa_cksum_tmpls[checksum],
		    &bp->blk_cksum);
	}
}

int
zio_checksum_error_impl(spa_t *spa, const blkptr_t *bp,
    enum zio_checksum checksum, abd_t *abd, uint64_t size, uint64_t offset,
    zio_bad_cksum_t *info)
{
	zio_checksum_info_t *ci = &zio_checksum_table[checksum];
	zio_cksum_t actual_cksum, expected_cksum;
	zio_eck_t eck;
	int byteswap;

	if (checksum >= ZIO_CHECKSUM_FUNCTIONS || ci->ci_func[0] == NULL)
		return (SET_ERROR(EINVAL));

	zio_checksum_template_init(checksum, spa);

	if (ci->ci_flags & ZCHECKSUM_FLAG_EMBEDDED) {
		zio_cksum_t verifier;
		size_t eck_offset;

		if (checksum == ZIO_CHECKSUM_ZILOG2) {
			zil_chain_t zilc;
			uint64_t nused;

			abd_copy_to_buf(&zilc, abd, sizeof (zil_chain_t));

			eck = zilc.zc_eck;
			eck_offset = offsetof(zil_chain_t, zc_eck) +
			    offsetof(zio_eck_t, zec_cksum);

			if (eck.zec_magic == ZEC_MAGIC) {
				nused = zilc.zc_nused;
			} else if (eck.zec_magic == BSWAP_64(ZEC_MAGIC)) {
				nused = BSWAP_64(zilc.zc_nused);
			} else {
				return (SET_ERROR(ECKSUM));
			}

			if (nused > size) {
				return (SET_ERROR(ECKSUM));
			}

			size = P2ROUNDUP_TYPED(nused, ZIL_MIN_BLKSZ, uint64_t);
		} else {
			eck_offset = size - sizeof (zio_eck_t);
			abd_copy_to_buf_off(&eck, abd, eck_offset,
			    sizeof (zio_eck_t));
			eck_offset += offsetof(zio_eck_t, zec_cksum);
		}

		if (checksum == ZIO_CHECKSUM_GANG_HEADER)
			zio_checksum_gang_verifier(&verifier, bp);
		else if (checksum == ZIO_CHECKSUM_LABEL)
			zio_checksum_label_verifier(&verifier, offset);
		else
			verifier = bp->blk_cksum;

		byteswap = (eck.zec_magic == BSWAP_64(ZEC_MAGIC));

		if (byteswap)
			byteswap_uint64_array(&verifier, sizeof (zio_cksum_t));

		expected_cksum = eck.zec_cksum;

		abd_copy_from_buf_off(abd, &verifier, eck_offset,
		    sizeof (zio_cksum_t));

		ci->ci_func[byteswap](abd, size,
		    spa->spa_cksum_tmpls[checksum], &actual_cksum);

		abd_copy_from_buf_off(abd, &expected_cksum, eck_offset,
		    sizeof (zio_cksum_t));

		if (byteswap) {
			byteswap_uint64_array(&expected_cksum,
			    sizeof (zio_cksum_t));
		}
	} else {
		byteswap = BP_SHOULD_BYTESWAP(bp);
		expected_cksum = bp->blk_cksum;
		ci->ci_func[byteswap](abd, size,
		    spa->spa_cksum_tmpls[checksum], &actual_cksum);
	}

	if (info != NULL) {
		info->zbc_expected = expected_cksum;
		info->zbc_actual = actual_cksum;
		info->zbc_checksum_name = ci->ci_name;
		info->zbc_byteswapped = byteswap;
		info->zbc_injected = 0;
		info->zbc_has_cksum = 1;
	}

	if (!ZIO_CHECKSUM_EQUAL(actual_cksum, expected_cksum))
		return (SET_ERROR(ECKSUM));

	return (0);
}

int
zio_checksum_error(zio_t *zio, zio_bad_cksum_t *info)
{
	blkptr_t *bp = zio->io_bp;
	uint_t checksum = (bp == NULL ? zio->io_prop.zp_checksum :
	    (BP_IS_GANG(bp) ? ZIO_CHECKSUM_GANG_HEADER : BP_GET_CHECKSUM(bp)));
	int error;
	uint64_t size = (bp == NULL ? zio->io_size :
	    (BP_IS_GANG(bp) ? SPA_GANGBLOCKSIZE : BP_GET_PSIZE(bp)));
	uint64_t offset = zio->io_offset;
	abd_t *data = zio->io_abd;
	spa_t *spa = zio->io_spa;

	error = zio_checksum_error_impl(spa, bp, checksum, data, size,
	    offset, info);

	if (zio_injection_enabled && error == 0 && zio->io_error == 0) {
		error = zio_handle_fault_injection(zio, ECKSUM);
		if (error != 0)
			info->zbc_injected = 1;
	}

	return (error);
}

/*
 * Called by a spa_t that's about to be deallocated. This steps through
 * all of the checksum context templates and deallocates any that were
 * initialized using the algorithm-specific template init function.
 */
void
zio_checksum_templates_free(spa_t *spa)
{
	enum zio_checksum checksum;
	for (checksum = 0; checksum < ZIO_CHECKSUM_FUNCTIONS;
	    checksum++) {
		if (spa->spa_cksum_tmpls[checksum] != NULL) {
			zio_checksum_info_t *ci = &zio_checksum_table[checksum];

			VERIFY(ci->ci_tmpl_free != NULL);
			ci->ci_tmpl_free(spa->spa_cksum_tmpls[checksum]);
			spa->spa_cksum_tmpls[checksum] = NULL;
		}
	}
}
Commit	Line	Data
34dc7c2f BB	1	/*
	2	* CDDL HEADER START
	3	*
	4	* The contents of this file are subject to the terms of the
	5	* Common Development and Distribution License (the "License").
	6	* You may not use this file except in compliance with the License.
	7	*
	8	* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
	9	* or http://www.opensolaris.org/os/licensing.
	10	* See the License for the specific language governing permissions
	11	* and limitations under the License.
	12	*
	13	* When distributing Covered Code, include this CDDL HEADER in each
	14	* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
	15	* If applicable, add the following below this CDDL HEADER, with the
	16	* fields enclosed by brackets "[]" replaced with your own identifying
	17	* information: Portions Copyright [yyyy] [name of copyright owner]
	18	*
	19	* CDDL HEADER END
	20	*/
	21	/*
428870ff	22	* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
cae5b340 AX	23	* Copyright (c) 2013, 2016 by Delphix. All rights reserved.
cae5b340 AX	24	* Copyright 2013 Saso Kiselkov. All rights reserved.
34dc7c2f BB	25	*/
34dc7c2f BB	26
34dc7c2f BB	27	#include <sys/zfs_context.h>
34dc7c2f BB	28	#include <sys/spa.h>
cae5b340	29	#include <sys/spa_impl.h>
34dc7c2f BB	30	#include <sys/zio.h>
34dc7c2f BB	31	#include <sys/zio_checksum.h>
428870ff	32	#include <sys/zil.h>
cae5b340	33	#include <sys/abd.h>
428870ff	34	#include <zfs_fletcher.h>
34dc7c2f BB	35
	36	/*
	37	* Checksum vectors.
	38	*
	39	* In the SPA, everything is checksummed. We support checksum vectors
	40	* for three distinct reasons:
	41	*
	42	* 1. Different kinds of data need different levels of protection.
	43	* For SPA metadata, we always want a very strong checksum.
	44	* For user data, we let users make the trade-off between speed
	45	* and checksum strength.
	46	*
	47	* 2. Cryptographic hash and MAC algorithms are an area of active research.
	48	* It is likely that in future hash functions will be at least as strong
	49	* as current best-of-breed, and may be substantially faster as well.
	50	* We want the ability to take advantage of these new hashes as soon as
	51	* they become available.
	52	*
	53	* 3. If someone develops hardware that can compute a strong hash quickly,
	54	* we want the ability to take advantage of that hardware.
	55	*
	56	* Of course, we don't want a checksum upgrade to invalidate existing
428870ff BB	57	* data, so we store the checksum function in eight bits of the bp.
428870ff BB	58	* This gives us room for up to 256 different checksum functions.
34dc7c2f BB	59	*
	60	* When writing a block, we always checksum it with the latest-and-greatest
	61	* checksum function of the appropriate strength. When reading a block,
	62	* we compare the expected checksum against the actual checksum, which we
428870ff	63	* compute via the checksum function specified by BP_GET_CHECKSUM(bp).
cae5b340 AX	64	*
	65	* SALTED CHECKSUMS
	66	*
	67	* To enable the use of less secure hash algorithms with dedup, we
	68	* introduce the notion of salted checksums (MACs, really). A salted
	69	* checksum is fed both a random 256-bit value (the salt) and the data
	70	* to be checksummed. This salt is kept secret (stored on the pool, but
	71	* never shown to the user). Thus even if an attacker knew of collision
	72	* weaknesses in the hash algorithm, they won't be able to mount a known
	73	* plaintext attack on the DDT, since the actual hash value cannot be
	74	* known ahead of time. How the salt is used is algorithm-specific
	75	* (some might simply prefix it to the data block, others might need to
	76	* utilize a full-blown HMAC). On disk the salt is stored in a ZAP
	77	* object in the MOS (DMU_POOL_CHECKSUM_SALT).
	78	*
	79	* CONTEXT TEMPLATES
	80	*
	81	* Some hashing algorithms need to perform a substantial amount of
	82	* initialization work (e.g. salted checksums above may need to pre-hash
	83	* the salt) before being able to process data. Performing this
	84	* redundant work for each block would be wasteful, so we instead allow
	85	* a checksum algorithm to do the work once (the first time it's used)
	86	* and then keep this pre-initialized context as a template inside the
	87	* spa_t (spa_cksum_tmpls). If the zio_checksum_info_t contains
	88	* non-NULL ci_tmpl_init and ci_tmpl_free callbacks, they are used to
	89	* construct and destruct the pre-initialized checksum context. The
	90	* pre-initialized context is then reused during each checksum
	91	* invocation and passed to the checksum function.
34dc7c2f BB	92	*/
	93
	94	/ARGSUSED/
	95	static void
cae5b340 AX	96	abd_checksum_off(abd_t *abd, uint64_t size,
cae5b340 AX	97	const void ctx_template, zio_cksum_t zcp)
34dc7c2f BB	98	{
	99	ZIO_SET_CHECKSUM(zcp, 0, 0, 0, 0);
	100	}
	101
cae5b340 AX	102	/ARGSUSED/
	103	void
	104	abd_fletcher_2_native(abd_t *abd, uint64_t size,
	105	const void ctx_template, zio_cksum_t zcp)
	106	{
	107	fletcher_init(zcp);
	108	(void) abd_iterate_func(abd, 0, size,
	109	fletcher_2_incremental_native, zcp);
	110	}
	111
	112	/ARGSUSED/
	113	void
	114	abd_fletcher_2_byteswap(abd_t *abd, uint64_t size,
	115	const void ctx_template, zio_cksum_t zcp)
	116	{
	117	fletcher_init(zcp);
	118	(void) abd_iterate_func(abd, 0, size,
	119	fletcher_2_incremental_byteswap, zcp);
	120	}
	121
	122	static inline void
	123	abd_fletcher_4_impl(abd_t abd, uint64_t size, zio_abd_checksum_data_t acdp)
	124	{
	125	fletcher_4_abd_ops.acf_init(acdp);
	126	abd_iterate_func(abd, 0, size, fletcher_4_abd_ops.acf_iter, acdp);
	127	fletcher_4_abd_ops.acf_fini(acdp);
	128	}
	129
	130	/ARGSUSED/
	131	void
	132	abd_fletcher_4_native(abd_t *abd, uint64_t size,
	133	const void ctx_template, zio_cksum_t zcp)
	134	{
	135	fletcher_4_ctx_t ctx;
	136
	137	zio_abd_checksum_data_t acd = {
	138	.acd_byteorder = ZIO_CHECKSUM_NATIVE,
	139	.acd_zcp = zcp,
	140	.acd_ctx = &ctx
	141	};
	142
	143	abd_fletcher_4_impl(abd, size, &acd);
	144
	145	}
	146
	147	/ARGSUSED/
	148	void
	149	abd_fletcher_4_byteswap(abd_t *abd, uint64_t size,
	150	const void ctx_template, zio_cksum_t zcp)
	151	{
	152	fletcher_4_ctx_t ctx;
	153
	154	zio_abd_checksum_data_t acd = {
	155	.acd_byteorder = ZIO_CHECKSUM_BYTESWAP,
	156	.acd_zcp = zcp,
	157	.acd_ctx = &ctx
	158	};
	159
	160	abd_fletcher_4_impl(abd, size, &acd);
	161	}
	162
34dc7c2f	163	zio_checksum_info_t zio_checksum_table[ZIO_CHECKSUM_FUNCTIONS] = {
cae5b340 AX	164	{{NULL, NULL}, NULL, NULL, 0, "inherit"},
	165	{{NULL, NULL}, NULL, NULL, 0, "on"},
	166	{{abd_checksum_off, abd_checksum_off},
	167	NULL, NULL, 0, "off"},
	168	{{abd_checksum_SHA256, abd_checksum_SHA256},
	169	NULL, NULL, ZCHECKSUM_FLAG_METADATA \| ZCHECKSUM_FLAG_EMBEDDED,
	170	"label"},
	171	{{abd_checksum_SHA256, abd_checksum_SHA256},
	172	NULL, NULL, ZCHECKSUM_FLAG_METADATA \| ZCHECKSUM_FLAG_EMBEDDED,
	173	"gang_header"},
	174	{{abd_fletcher_2_native, abd_fletcher_2_byteswap},
	175	NULL, NULL, ZCHECKSUM_FLAG_EMBEDDED, "zilog"},
	176	{{abd_fletcher_2_native, abd_fletcher_2_byteswap},
	177	NULL, NULL, 0, "fletcher2"},
	178	{{abd_fletcher_4_native, abd_fletcher_4_byteswap},
	179	NULL, NULL, ZCHECKSUM_FLAG_METADATA, "fletcher4"},
	180	{{abd_checksum_SHA256, abd_checksum_SHA256},
	181	NULL, NULL, ZCHECKSUM_FLAG_METADATA \| ZCHECKSUM_FLAG_DEDUP \|
	182	ZCHECKSUM_FLAG_NOPWRITE, "sha256"},
	183	{{abd_fletcher_4_native, abd_fletcher_4_byteswap},
	184	NULL, NULL, ZCHECKSUM_FLAG_EMBEDDED, "zilog2"},
	185	{{abd_checksum_off, abd_checksum_off},
	186	NULL, NULL, 0, "noparity"},
	187	{{abd_checksum_SHA512_native, abd_checksum_SHA512_byteswap},
	188	NULL, NULL, ZCHECKSUM_FLAG_METADATA \| ZCHECKSUM_FLAG_DEDUP \|
	189	ZCHECKSUM_FLAG_NOPWRITE, "sha512"},
	190	{{abd_checksum_skein_native, abd_checksum_skein_byteswap},
	191	abd_checksum_skein_tmpl_init, abd_checksum_skein_tmpl_free,
	192	ZCHECKSUM_FLAG_METADATA \| ZCHECKSUM_FLAG_DEDUP \|
	193	ZCHECKSUM_FLAG_SALTED \| ZCHECKSUM_FLAG_NOPWRITE, "skein"},
	194	{{abd_checksum_edonr_native, abd_checksum_edonr_byteswap},
	195	abd_checksum_edonr_tmpl_init, abd_checksum_edonr_tmpl_free,
	196	ZCHECKSUM_FLAG_METADATA \| ZCHECKSUM_FLAG_SALTED \|
	197	ZCHECKSUM_FLAG_NOPWRITE, "edonr"},
34dc7c2f BB	198	};
34dc7c2f BB	199
cae5b340 AX	200	/*
	201	* The flag corresponding to the "verify" in dedup=[checksum,]verify
	202	* must be cleared first, so callers should use ZIO_CHECKSUM_MASK.
	203	*/
	204	spa_feature_t
	205	zio_checksum_to_feature(enum zio_checksum cksum)
	206	{
	207	VERIFY((cksum & ~ZIO_CHECKSUM_MASK) == 0);
	208
	209	switch (cksum) {
	210	case ZIO_CHECKSUM_SHA512:
	211	return (SPA_FEATURE_SHA512);
	212	case ZIO_CHECKSUM_SKEIN:
	213	return (SPA_FEATURE_SKEIN);
	214	case ZIO_CHECKSUM_EDONR:
	215	return (SPA_FEATURE_EDONR);
	216	default:
	217	return (SPA_FEATURE_NONE);
	218	}
	219	}
	220
428870ff BB	221	enum zio_checksum
428870ff BB	222	zio_checksum_select(enum zio_checksum child, enum zio_checksum parent)
34dc7c2f BB	223	{
	224	ASSERT(child < ZIO_CHECKSUM_FUNCTIONS);
	225	ASSERT(parent < ZIO_CHECKSUM_FUNCTIONS);
	226	ASSERT(parent != ZIO_CHECKSUM_INHERIT && parent != ZIO_CHECKSUM_ON);
	227
	228	if (child == ZIO_CHECKSUM_INHERIT)
	229	return (parent);
	230
	231	if (child == ZIO_CHECKSUM_ON)
	232	return (ZIO_CHECKSUM_ON_VALUE);
	233
	234	return (child);
	235	}
	236
428870ff BB	237	enum zio_checksum
	238	zio_checksum_dedup_select(spa_t *spa, enum zio_checksum child,
	239	enum zio_checksum parent)
	240	{
	241	ASSERT((child & ZIO_CHECKSUM_MASK) < ZIO_CHECKSUM_FUNCTIONS);
	242	ASSERT((parent & ZIO_CHECKSUM_MASK) < ZIO_CHECKSUM_FUNCTIONS);
	243	ASSERT(parent != ZIO_CHECKSUM_INHERIT && parent != ZIO_CHECKSUM_ON);
	244
	245	if (child == ZIO_CHECKSUM_INHERIT)
	246	return (parent);
	247
	248	if (child == ZIO_CHECKSUM_ON)
	249	return (spa_dedup_checksum(spa));
	250
	251	if (child == (ZIO_CHECKSUM_ON \| ZIO_CHECKSUM_VERIFY))
	252	return (spa_dedup_checksum(spa) \| ZIO_CHECKSUM_VERIFY);
	253
cae5b340 AX	254	ASSERT((zio_checksum_table[child & ZIO_CHECKSUM_MASK].ci_flags &
cae5b340 AX	255	ZCHECKSUM_FLAG_DEDUP) \|\|
428870ff BB	256	(child & ZIO_CHECKSUM_VERIFY) \|\| child == ZIO_CHECKSUM_OFF);
	257
	258	return (child);
	259	}
	260
b128c09f BB	261	/*
	262	* Set the external verifier for a gang block based on <vdev, offset, txg>,
	263	* a tuple which is guaranteed to be unique for the life of the pool.
	264	*/
	265	static void
cae5b340	266	zio_checksum_gang_verifier(zio_cksum_t zcp, const blkptr_t bp)
b128c09f	267	{
ea04106b	268	const dva_t *dva = BP_IDENTITY(bp);
428870ff	269	uint64_t txg = BP_PHYSICAL_BIRTH(bp);
b128c09f BB	270
	271	ASSERT(BP_IS_GANG(bp));
	272
	273	ZIO_SET_CHECKSUM(zcp, DVA_GET_VDEV(dva), DVA_GET_OFFSET(dva), txg, 0);
	274	}
	275
	276	/*
	277	* Set the external verifier for a label block based on its offset.
	278	* The vdev is implicit, and the txg is unknowable at pool open time --
	279	* hence the logic in vdev_uberblock_load() to find the most recent copy.
	280	*/
	281	static void
	282	zio_checksum_label_verifier(zio_cksum_t *zcp, uint64_t offset)
	283	{
	284	ZIO_SET_CHECKSUM(zcp, offset, 0, 0, 0);
	285	}
	286
cae5b340 AX	287	/*
	288	* Calls the template init function of a checksum which supports context
	289	* templates and installs the template into the spa_t.
	290	*/
	291	static void
	292	zio_checksum_template_init(enum zio_checksum checksum, spa_t *spa)
	293	{
	294	zio_checksum_info_t *ci = &zio_checksum_table[checksum];
	295
	296	if (ci->ci_tmpl_init == NULL)
	297	return;
	298	if (spa->spa_cksum_tmpls[checksum] != NULL)
	299	return;
	300
	301	VERIFY(ci->ci_tmpl_free != NULL);
	302	mutex_enter(&spa->spa_cksum_tmpls_lock);
	303	if (spa->spa_cksum_tmpls[checksum] == NULL) {
	304	spa->spa_cksum_tmpls[checksum] =
	305	ci->ci_tmpl_init(&spa->spa_cksum_salt);
	306	VERIFY(spa->spa_cksum_tmpls[checksum] != NULL);
	307	}
	308	mutex_exit(&spa->spa_cksum_tmpls_lock);
	309	}
	310
34dc7c2f BB	311	/*
	312	* Generate the checksum.
	313	*/
	314	void
b128c09f	315	zio_checksum_compute(zio_t *zio, enum zio_checksum checksum,
cae5b340	316	abd_t *abd, uint64_t size)
34dc7c2f	317	{
cae5b340	318	static const uint64_t zec_magic = ZEC_MAGIC;
b128c09f BB	319	blkptr_t *bp = zio->io_bp;
b128c09f BB	320	uint64_t offset = zio->io_offset;
34dc7c2f	321	zio_checksum_info_t *ci = &zio_checksum_table[checksum];
428870ff	322	zio_cksum_t cksum;
cae5b340	323	spa_t *spa = zio->io_spa;
34dc7c2f	324
b128c09f	325	ASSERT((uint_t)checksum < ZIO_CHECKSUM_FUNCTIONS);
34dc7c2f BB	326	ASSERT(ci->ci_func[0] != NULL);
34dc7c2f BB	327
cae5b340 AX	328	zio_checksum_template_init(checksum, spa);
	329
	330	if (ci->ci_flags & ZCHECKSUM_FLAG_EMBEDDED) {
	331	zio_eck_t eck;
	332	size_t eck_offset;
428870ff BB	333
428870ff BB	334	if (checksum == ZIO_CHECKSUM_ZILOG2) {
cae5b340 AX	335	zil_chain_t zilc;
cae5b340 AX	336	abd_copy_to_buf(&zilc, abd, sizeof (zil_chain_t));
428870ff	337
cae5b340	338	size = P2ROUNDUP_TYPED(zilc.zc_nused, ZIL_MIN_BLKSZ,
428870ff	339	uint64_t);
cae5b340 AX	340	eck = zilc.zc_eck;
cae5b340 AX	341	eck_offset = offsetof(zil_chain_t, zc_eck);
428870ff	342	} else {
cae5b340 AX	343	eck_offset = size - sizeof (zio_eck_t);
	344	abd_copy_to_buf_off(&eck, abd, eck_offset,
	345	sizeof (zio_eck_t));
428870ff	346	}
cae5b340 AX	347
	348	if (checksum == ZIO_CHECKSUM_GANG_HEADER) {
	349	zio_checksum_gang_verifier(&eck.zec_cksum, bp);
	350	abd_copy_from_buf_off(abd, &eck.zec_cksum,
	351	eck_offset + offsetof(zio_eck_t, zec_cksum),
	352	sizeof (zio_cksum_t));
	353	} else if (checksum == ZIO_CHECKSUM_LABEL) {
	354	zio_checksum_label_verifier(&eck.zec_cksum, offset);
	355	abd_copy_from_buf_off(abd, &eck.zec_cksum,
	356	eck_offset + offsetof(zio_eck_t, zec_cksum),
	357	sizeof (zio_cksum_t));
	358	} else {
	359	bp->blk_cksum = eck.zec_cksum;
	360	}
	361
	362	abd_copy_from_buf_off(abd, &zec_magic,
	363	eck_offset + offsetof(zio_eck_t, zec_magic),
	364	sizeof (zec_magic));
	365
	366	ci->ci_func[0](abd, size, spa->spa_cksum_tmpls[checksum],
	367	&cksum);
	368
	369	abd_copy_from_buf_off(abd, &cksum,
	370	eck_offset + offsetof(zio_eck_t, zec_cksum),
	371	sizeof (zio_cksum_t));
34dc7c2f	372	} else {
cae5b340 AX	373	ci->ci_func[0](abd, size, spa->spa_cksum_tmpls[checksum],
cae5b340 AX	374	&bp->blk_cksum);
34dc7c2f BB	375	}
	376	}
	377
	378	int
cae5b340 AX	379	zio_checksum_error_impl(spa_t spa, const blkptr_t bp,
	380	enum zio_checksum checksum, abd_t *abd, uint64_t size, uint64_t offset,
	381	zio_bad_cksum_t *info)
34dc7c2f	382	{
34dc7c2f	383	zio_checksum_info_t *ci = &zio_checksum_table[checksum];
cae5b340 AX	384	zio_cksum_t actual_cksum, expected_cksum;
	385	zio_eck_t eck;
	386	int byteswap;
34dc7c2f BB	387
34dc7c2f BB	388	if (checksum >= ZIO_CHECKSUM_FUNCTIONS \|\| ci->ci_func[0] == NULL)
a08ee875	389	return (SET_ERROR(EINVAL));
34dc7c2f	390
cae5b340 AX	391	zio_checksum_template_init(checksum, spa);
	392
	393	if (ci->ci_flags & ZCHECKSUM_FLAG_EMBEDDED) {
	394	zio_cksum_t verifier;
	395	size_t eck_offset;
428870ff BB	396
428870ff BB	397	if (checksum == ZIO_CHECKSUM_ZILOG2) {
cae5b340	398	zil_chain_t zilc;
428870ff BB	399	uint64_t nused;
428870ff BB	400
cae5b340 AX	401	abd_copy_to_buf(&zilc, abd, sizeof (zil_chain_t));
	402
	403	eck = zilc.zc_eck;
	404	eck_offset = offsetof(zil_chain_t, zc_eck) +
	405	offsetof(zio_eck_t, zec_cksum);
	406
	407	if (eck.zec_magic == ZEC_MAGIC) {
	408	nused = zilc.zc_nused;
	409	} else if (eck.zec_magic == BSWAP_64(ZEC_MAGIC)) {
	410	nused = BSWAP_64(zilc.zc_nused);
	411	} else {
a08ee875	412	return (SET_ERROR(ECKSUM));
cae5b340	413	}
428870ff	414
cae5b340	415	if (nused > size) {
a08ee875	416	return (SET_ERROR(ECKSUM));
cae5b340	417	}
428870ff BB	418
	419	size = P2ROUNDUP_TYPED(nused, ZIL_MIN_BLKSZ, uint64_t);
	420	} else {
cae5b340 AX	421	eck_offset = size - sizeof (zio_eck_t);
	422	abd_copy_to_buf_off(&eck, abd, eck_offset,
	423	sizeof (zio_eck_t));
	424	eck_offset += offsetof(zio_eck_t, zec_cksum);
428870ff BB	425	}
428870ff BB	426
34dc7c2f	427	if (checksum == ZIO_CHECKSUM_GANG_HEADER)
b128c09f BB	428	zio_checksum_gang_verifier(&verifier, bp);
	429	else if (checksum == ZIO_CHECKSUM_LABEL)
	430	zio_checksum_label_verifier(&verifier, offset);
	431	else
	432	verifier = bp->blk_cksum;
	433
cae5b340	434	byteswap = (eck.zec_magic == BSWAP_64(ZEC_MAGIC));
34dc7c2f	435
b128c09f BB	436	if (byteswap)
	437	byteswap_uint64_array(&verifier, sizeof (zio_cksum_t));
	438
cae5b340	439	expected_cksum = eck.zec_cksum;
b128c09f	440
cae5b340 AX	441	abd_copy_from_buf_off(abd, &verifier, eck_offset,
	442	sizeof (zio_cksum_t));
	443
	444	ci->ci_func[byteswap](abd, size,
	445	spa->spa_cksum_tmpls[checksum], &actual_cksum);
	446
	447	abd_copy_from_buf_off(abd, &expected_cksum, eck_offset,
	448	sizeof (zio_cksum_t));
	449
	450	if (byteswap) {
34dc7c2f BB	451	byteswap_uint64_array(&expected_cksum,
34dc7c2f BB	452	sizeof (zio_cksum_t));
cae5b340	453	}
34dc7c2f	454	} else {
b128c09f BB	455	byteswap = BP_SHOULD_BYTESWAP(bp);
b128c09f BB	456	expected_cksum = bp->blk_cksum;
cae5b340 AX	457	ci->ci_func[byteswap](abd, size,
cae5b340 AX	458	spa->spa_cksum_tmpls[checksum], &actual_cksum);
34dc7c2f BB	459	}
34dc7c2f BB	460
cae5b340 AX	461	if (info != NULL) {
	462	info->zbc_expected = expected_cksum;
	463	info->zbc_actual = actual_cksum;
	464	info->zbc_checksum_name = ci->ci_name;
	465	info->zbc_byteswapped = byteswap;
	466	info->zbc_injected = 0;
	467	info->zbc_has_cksum = 1;
	468	}
428870ff	469
b128c09f	470	if (!ZIO_CHECKSUM_EQUAL(actual_cksum, expected_cksum))
a08ee875	471	return (SET_ERROR(ECKSUM));
34dc7c2f	472
cae5b340 AX	473	return (0);
cae5b340 AX	474	}
428870ff	475
cae5b340 AX	476	int
	477	zio_checksum_error(zio_t zio, zio_bad_cksum_t info)
	478	{
	479	blkptr_t *bp = zio->io_bp;
	480	uint_t checksum = (bp == NULL ? zio->io_prop.zp_checksum :
	481	(BP_IS_GANG(bp) ? ZIO_CHECKSUM_GANG_HEADER : BP_GET_CHECKSUM(bp)));
	482	int error;
	483	uint64_t size = (bp == NULL ? zio->io_size :
	484	(BP_IS_GANG(bp) ? SPA_GANGBLOCKSIZE : BP_GET_PSIZE(bp)));
	485	uint64_t offset = zio->io_offset;
	486	abd_t *data = zio->io_abd;
	487	spa_t *spa = zio->io_spa;
	488
	489	error = zio_checksum_error_impl(spa, bp, checksum, data, size,
	490	offset, info);
	491
	492	if (zio_injection_enabled && error == 0 && zio->io_error == 0) {
	493	error = zio_handle_fault_injection(zio, ECKSUM);
	494	if (error != 0)
	495	info->zbc_injected = 1;
428870ff	496	}
34dc7c2f	497
cae5b340 AX	498	return (error);
	499	}
	500
	501	/*
	502	* Called by a spa_t that's about to be deallocated. This steps through
	503	* all of the checksum context templates and deallocates any that were
	504	* initialized using the algorithm-specific template init function.
	505	*/
	506	void
	507	zio_checksum_templates_free(spa_t *spa)
	508	{
	509	enum zio_checksum checksum;
	510	for (checksum = 0; checksum < ZIO_CHECKSUM_FUNCTIONS;
	511	checksum++) {
	512	if (spa->spa_cksum_tmpls[checksum] != NULL) {
	513	zio_checksum_info_t *ci = &zio_checksum_table[checksum];
	514
	515	VERIFY(ci->ci_tmpl_free != NULL);
	516	ci->ci_tmpl_free(spa->spa_cksum_tmpls[checksum]);
	517	spa->spa_cksum_tmpls[checksum] = NULL;
	518	}
	519	}
34dc7c2f	520	}