bio-integrity.c 14.2 KB
Newer Older
1 2 3
/*
 * bio-integrity.c - bio data integrity extensions
 *
4
 * Copyright (C) 2007, 2008, 2009 Oracle Corporation
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
 * Written by: Martin K. Petersen <martin.petersen@oracle.com>
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License version
 * 2 as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it will be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; see the file COPYING.  If not, write to
 * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139,
 * USA.
 *
 */

#include <linux/blkdev.h>
#include <linux/mempool.h>
25
#include <linux/export.h>
26 27
#include <linux/bio.h>
#include <linux/workqueue.h>
28
#include <linux/slab.h>
29

30
#define BIP_INLINE_VECS	4
31

32
static struct kmem_cache *bip_slab;
33 34 35
static struct workqueue_struct *kintegrityd_wq;

/**
36
 * bio_integrity_alloc - Allocate integrity payload and attach it to bio
37 38 39 40 41 42 43 44
 * @bio:	bio to attach integrity metadata to
 * @gfp_mask:	Memory allocation mask
 * @nr_vecs:	Number of integrity metadata scatter-gather elements
 *
 * Description: This function prepares a bio for attaching integrity
 * metadata.  nr_vecs specifies the maximum number of pages containing
 * integrity metadata that can be attached.
 */
45 46 47
struct bio_integrity_payload *bio_integrity_alloc(struct bio *bio,
						  gfp_t gfp_mask,
						  unsigned int nr_vecs)
48 49
{
	struct bio_integrity_payload *bip;
50
	struct bio_set *bs = bio->bi_pool;
51 52 53
	unsigned long idx = BIO_POOL_NONE;
	unsigned inline_vecs;

54
	if (!bs || !bs->bio_integrity_pool) {
55 56 57 58
		bip = kmalloc(sizeof(struct bio_integrity_payload) +
			      sizeof(struct bio_vec) * nr_vecs, gfp_mask);
		inline_vecs = nr_vecs;
	} else {
59
		bip = mempool_alloc(bs->bio_integrity_pool, gfp_mask);
60
		inline_vecs = BIP_INLINE_VECS;
61 62
	}

63 64 65
	if (unlikely(!bip))
		return NULL;

66 67
	memset(bip, 0, sizeof(*bip));

68 69 70 71 72
	if (nr_vecs > inline_vecs) {
		bip->bip_vec = bvec_alloc(gfp_mask, nr_vecs, &idx,
					  bs->bvec_integrity_pool);
		if (!bip->bip_vec)
			goto err;
73
		bip->bip_max_vcnt = bvec_nr_vecs(idx);
74 75
	} else {
		bip->bip_vec = bip->bip_inline_vecs;
76
		bip->bip_max_vcnt = inline_vecs;
77 78
	}

79
	bip->bip_slab = idx;
80 81
	bip->bip_bio = bio;
	bio->bi_integrity = bip;
82
	bio->bi_rw |= REQ_INTEGRITY;
83 84

	return bip;
85 86 87
err:
	mempool_free(bip, bs->bio_integrity_pool);
	return NULL;
88 89 90 91 92 93 94 95 96 97
}
EXPORT_SYMBOL(bio_integrity_alloc);

/**
 * bio_integrity_free - Free bio integrity payload
 * @bio:	bio containing bip to be freed
 *
 * Description: Used to free the integrity portion of a bio. Usually
 * called from bio_free().
 */
98
void bio_integrity_free(struct bio *bio)
99
{
100
	struct bio_integrity_payload *bip = bio_integrity(bio);
101 102
	struct bio_set *bs = bio->bi_pool;

103
	if (bip->bip_flags & BIP_BLOCK_INTEGRITY)
M
Martin K. Petersen 已提交
104 105
		kfree(page_address(bip->bip_vec->bv_page) +
		      bip->bip_vec->bv_offset);
106

107
	if (bs && bs->bio_integrity_pool) {
108 109 110 111
		if (bip->bip_slab != BIO_POOL_NONE)
			bvec_free(bs->bvec_integrity_pool, bip->bip_vec,
				  bip->bip_slab);

112
		mempool_free(bip, bs->bio_integrity_pool);
113 114 115
	} else {
		kfree(bip);
	}
116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132

	bio->bi_integrity = NULL;
}
EXPORT_SYMBOL(bio_integrity_free);

/**
 * bio_integrity_add_page - Attach integrity metadata
 * @bio:	bio to update
 * @page:	page containing integrity metadata
 * @len:	number of bytes of integrity metadata in page
 * @offset:	start offset within page
 *
 * Description: Attach a page containing integrity metadata to bio.
 */
int bio_integrity_add_page(struct bio *bio, struct page *page,
			   unsigned int len, unsigned int offset)
{
133
	struct bio_integrity_payload *bip = bio_integrity(bio);
134 135
	struct bio_vec *iv;

136
	if (bip->bip_vcnt >= bip->bip_max_vcnt) {
137 138 139 140
		printk(KERN_ERR "%s: bip_vec full\n", __func__);
		return 0;
	}

141
	iv = bip->bip_vec + bip->bip_vcnt;
142

143 144 145 146 147
	if (bip->bip_vcnt &&
	    bvec_gap_to_prev(bdev_get_queue(bio->bi_bdev),
			     &bip->bip_vec[bip->bip_vcnt - 1], offset))
		return 0;

148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165
	iv->bv_page = page;
	iv->bv_len = len;
	iv->bv_offset = offset;
	bip->bip_vcnt++;

	return len;
}
EXPORT_SYMBOL(bio_integrity_add_page);

/**
 * bio_integrity_enabled - Check whether integrity can be passed
 * @bio:	bio to check
 *
 * Description: Determines whether bio_integrity_prep() can be called
 * on this bio or not.	bio data direction and target device must be
 * set prior to calling.  The functions honors the write_generate and
 * read_verify flags in sysfs.
 */
166
bool bio_integrity_enabled(struct bio *bio)
167
{
168 169
	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);

170
	if (!bio_is_rw(bio))
171
		return false;
172

173 174
	/* Already protected? */
	if (bio_integrity(bio))
175 176 177 178 179
		return false;

	if (bi == NULL)
		return false;

180
	if (bio_data_dir(bio) == READ && bi->profile->verify_fn != NULL &&
181
	    (bi->flags & BLK_INTEGRITY_VERIFY))
182 183
		return true;

184
	if (bio_data_dir(bio) == WRITE && bi->profile->generate_fn != NULL &&
185
	    (bi->flags & BLK_INTEGRITY_GENERATE))
186
		return true;
187

188
	return false;
189 190 191 192
}
EXPORT_SYMBOL(bio_integrity_enabled);

/**
193
 * bio_integrity_intervals - Return number of integrity intervals for a bio
194
 * @bi:		blk_integrity profile for device
195
 * @sectors:	Size of the bio in 512-byte sectors
196 197
 *
 * Description: The block layer calculates everything in 512 byte
198 199 200
 * sectors but integrity metadata is done in terms of the data integrity
 * interval size of the storage device.  Convert the block layer sectors
 * to the appropriate number of integrity intervals.
201
 */
202 203
static inline unsigned int bio_integrity_intervals(struct blk_integrity *bi,
						   unsigned int sectors)
204
{
205
	return sectors >> (ilog2(bi->interval) - 9);
206 207
}

208 209 210
static inline unsigned int bio_integrity_bytes(struct blk_integrity *bi,
					       unsigned int sectors)
{
211
	return bio_integrity_intervals(bi, sectors) * bi->tuple_size;
212 213
}

214
/**
215
 * bio_integrity_process - Process integrity metadata for a bio
216
 * @bio:	bio to generate/verify integrity metadata for
217
 * @proc_fn:	Pointer to the relevant processing function
218
 */
219 220
static int bio_integrity_process(struct bio *bio,
				 integrity_processing_fn *proc_fn)
221 222
{
	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
223
	struct blk_integrity_iter iter;
224 225
	struct bvec_iter bviter;
	struct bio_vec bv;
M
Martin K. Petersen 已提交
226
	struct bio_integrity_payload *bip = bio_integrity(bio);
227
	unsigned int ret = 0;
M
Martin K. Petersen 已提交
228 229
	void *prot_buf = page_address(bip->bip_vec->bv_page) +
		bip->bip_vec->bv_offset;
230

231 232 233 234
	iter.disk_name = bio->bi_bdev->bd_disk->disk_name;
	iter.interval = bi->interval;
	iter.seed = bip_get_seed(bip);
	iter.prot_buf = prot_buf;
235

236 237
	bio_for_each_segment(bv, bio, bviter) {
		void *kaddr = kmap_atomic(bv.bv_page);
238

239 240
		iter.data_buf = kaddr + bv.bv_offset;
		iter.data_size = bv.bv_len;
241 242 243 244 245 246

		ret = proc_fn(&iter);
		if (ret) {
			kunmap_atomic(kaddr);
			return ret;
		}
247

248
		kunmap_atomic(kaddr);
249
	}
250 251 252
	return ret;
}

253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272
/**
 * bio_integrity_prep - Prepare bio for integrity I/O
 * @bio:	bio to prepare
 *
 * Description: Allocates a buffer for integrity metadata, maps the
 * pages and attaches them to a bio.  The bio must have data
 * direction, target device and start sector set priot to calling.  In
 * the WRITE case, integrity metadata will be generated using the
 * block device's integrity function.  In the READ case, the buffer
 * will be prepared for DMA and a suitable end_io handler set up.
 */
int bio_integrity_prep(struct bio *bio)
{
	struct bio_integrity_payload *bip;
	struct blk_integrity *bi;
	struct request_queue *q;
	void *buf;
	unsigned long start, end;
	unsigned int len, nr_pages;
	unsigned int bytes, offset, i;
273
	unsigned int intervals;
274 275 276 277 278 279

	bi = bdev_get_integrity(bio->bi_bdev);
	q = bdev_get_queue(bio->bi_bdev);
	BUG_ON(bi == NULL);
	BUG_ON(bio_integrity(bio));

280
	intervals = bio_integrity_intervals(bi, bio_sectors(bio));
281 282

	/* Allocate kernel buffer for protection data */
283
	len = intervals * bi->tuple_size;
284
	buf = kmalloc(len, GFP_NOIO | q->bounce_gfp);
285 286
	if (unlikely(buf == NULL)) {
		printk(KERN_ERR "could not allocate integrity buffer\n");
287
		return -ENOMEM;
288 289 290 291 292 293 294 295 296 297 298 299 300 301
	}

	end = (((unsigned long) buf) + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
	start = ((unsigned long) buf) >> PAGE_SHIFT;
	nr_pages = end - start;

	/* Allocate bio integrity payload and integrity vectors */
	bip = bio_integrity_alloc(bio, GFP_NOIO, nr_pages);
	if (unlikely(bip == NULL)) {
		printk(KERN_ERR "could not allocate data integrity bioset\n");
		kfree(buf);
		return -EIO;
	}

302
	bip->bip_flags |= BIP_BLOCK_INTEGRITY;
303
	bip->bip_iter.bi_size = len;
304
	bip_set_seed(bip, bio->bi_iter.bi_sector);
305

306 307 308
	if (bi->flags & BLK_INTEGRITY_IP_CHECKSUM)
		bip->bip_flags |= BIP_IP_CHECKSUM;

309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342
	/* Map it */
	offset = offset_in_page(buf);
	for (i = 0 ; i < nr_pages ; i++) {
		int ret;
		bytes = PAGE_SIZE - offset;

		if (len <= 0)
			break;

		if (bytes > len)
			bytes = len;

		ret = bio_integrity_add_page(bio, virt_to_page(buf),
					     bytes, offset);

		if (ret == 0)
			return 0;

		if (ret < bytes)
			break;

		buf += bytes;
		len -= bytes;
		offset = 0;
	}

	/* Install custom I/O completion handler if read verify is enabled */
	if (bio_data_dir(bio) == READ) {
		bip->bip_end_io = bio->bi_end_io;
		bio->bi_end_io = bio_integrity_endio;
	}

	/* Auto-generate integrity metadata if this is a write */
	if (bio_data_dir(bio) == WRITE)
343
		bio_integrity_process(bio, bi->profile->generate_fn);
344 345 346 347 348 349 350 351 352 353 354 355 356 357 358

	return 0;
}
EXPORT_SYMBOL(bio_integrity_prep);

/**
 * bio_integrity_verify_fn - Integrity I/O completion worker
 * @work:	Work struct stored in bio to be verified
 *
 * Description: This workqueue function is called to complete a READ
 * request.  The function verifies the transferred integrity metadata
 * and then calls the original bio end_io function.
 */
static void bio_integrity_verify_fn(struct work_struct *work)
{
359
	struct bio_integrity_payload *bip =
360 361
		container_of(work, struct bio_integrity_payload, bip_work);
	struct bio *bio = bip->bip_bio;
362
	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
363

364
	bio->bi_error = bio_integrity_process(bio, bi->profile->verify_fn);
365 366 367

	/* Restore original bio completion handler */
	bio->bi_end_io = bip->bip_end_io;
368
	bio_endio(bio);
369 370 371 372 373 374 375 376 377 378 379 380 381 382
}

/**
 * bio_integrity_endio - Integrity I/O completion function
 * @bio:	Protected bio
 * @error:	Pointer to errno
 *
 * Description: Completion for integrity I/O
 *
 * Normally I/O completion is done in interrupt context.  However,
 * verifying I/O integrity is a time-consuming task which must be run
 * in process context.	This function postpones completion
 * accordingly.
 */
383
void bio_integrity_endio(struct bio *bio)
384
{
385
	struct bio_integrity_payload *bip = bio_integrity(bio);
386 387 388

	BUG_ON(bip->bip_bio != bio);

389 390 391 392
	/* In case of an I/O error there is no point in verifying the
	 * integrity metadata.  Restore original bio end_io handler
	 * and run it.
	 */
393
	if (bio->bi_error) {
394
		bio->bi_end_io = bip->bip_end_io;
395
		bio_endio(bio);
396 397 398 399

		return;
	}

400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415
	INIT_WORK(&bip->bip_work, bio_integrity_verify_fn);
	queue_work(kintegrityd_wq, &bip->bip_work);
}
EXPORT_SYMBOL(bio_integrity_endio);

/**
 * bio_integrity_advance - Advance integrity vector
 * @bio:	bio whose integrity vector to update
 * @bytes_done:	number of data bytes that have been completed
 *
 * Description: This function calculates how many integrity bytes the
 * number of completed data bytes correspond to and advances the
 * integrity vector accordingly.
 */
void bio_integrity_advance(struct bio *bio, unsigned int bytes_done)
{
416
	struct bio_integrity_payload *bip = bio_integrity(bio);
417
	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
418
	unsigned bytes = bio_integrity_bytes(bi, bytes_done >> 9);
419

420
	bvec_iter_advance(bip->bip_vec, &bip->bip_iter, bytes);
421 422 423 424 425 426 427 428 429 430 431 432 433 434
}
EXPORT_SYMBOL(bio_integrity_advance);

/**
 * bio_integrity_trim - Trim integrity vector
 * @bio:	bio whose integrity vector to update
 * @offset:	offset to first data sector
 * @sectors:	number of data sectors
 *
 * Description: Used to trim the integrity vector in a cloned bio.
 * The ivec will be advanced corresponding to 'offset' data sectors
 * and the length will be truncated corresponding to 'len' data
 * sectors.
 */
435 436
void bio_integrity_trim(struct bio *bio, unsigned int offset,
			unsigned int sectors)
437
{
438
	struct bio_integrity_payload *bip = bio_integrity(bio);
439 440
	struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);

441 442
	bio_integrity_advance(bio, offset << 9);
	bip->bip_iter.bi_size = bio_integrity_bytes(bi, sectors);
443 444 445 446 447 448 449
}
EXPORT_SYMBOL(bio_integrity_trim);

/**
 * bio_integrity_clone - Callback for cloning bios with integrity metadata
 * @bio:	New bio
 * @bio_src:	Original bio
450
 * @gfp_mask:	Memory allocation mask
451 452 453
 *
 * Description:	Called to allocate a bip when cloning a bio
 */
454
int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
455
			gfp_t gfp_mask)
456
{
457
	struct bio_integrity_payload *bip_src = bio_integrity(bio_src);
458 459 460 461
	struct bio_integrity_payload *bip;

	BUG_ON(bip_src == NULL);

462
	bip = bio_integrity_alloc(bio, gfp_mask, bip_src->bip_vcnt);
463 464 465 466 467 468 469 470

	if (bip == NULL)
		return -EIO;

	memcpy(bip->bip_vec, bip_src->bip_vec,
	       bip_src->bip_vcnt * sizeof(struct bio_vec));

	bip->bip_vcnt = bip_src->bip_vcnt;
471
	bip->bip_iter = bip_src->bip_iter;
472 473 474 475 476

	return 0;
}
EXPORT_SYMBOL(bio_integrity_clone);

477
int bioset_integrity_create(struct bio_set *bs, int pool_size)
478
{
479 480 481
	if (bs->bio_integrity_pool)
		return 0;

482
	bs->bio_integrity_pool = mempool_create_slab_pool(pool_size, bip_slab);
483
	if (!bs->bio_integrity_pool)
484
		return -1;
485

486
	bs->bvec_integrity_pool = biovec_create_pool(pool_size);
487 488
	if (!bs->bvec_integrity_pool) {
		mempool_destroy(bs->bio_integrity_pool);
489
		return -1;
490
	}
491 492 493 494 495 496 497 498 499

	return 0;
}
EXPORT_SYMBOL(bioset_integrity_create);

void bioset_integrity_free(struct bio_set *bs)
{
	if (bs->bio_integrity_pool)
		mempool_destroy(bs->bio_integrity_pool);
500 501

	if (bs->bvec_integrity_pool)
502
		mempool_destroy(bs->bvec_integrity_pool);
503 504 505 506 507
}
EXPORT_SYMBOL(bioset_integrity_free);

void __init bio_integrity_init(void)
{
508 509 510 511 512 513
	/*
	 * kintegrityd won't block much but may burn a lot of CPU cycles.
	 * Make it highpri CPU intensive wq with max concurrency of 1.
	 */
	kintegrityd_wq = alloc_workqueue("kintegrityd", WQ_MEM_RECLAIM |
					 WQ_HIGHPRI | WQ_CPU_INTENSIVE, 1);
514 515
	if (!kintegrityd_wq)
		panic("Failed to create kintegrityd\n");
516

517 518 519 520
	bip_slab = kmem_cache_create("bio_integrity_payload",
				     sizeof(struct bio_integrity_payload) +
				     sizeof(struct bio_vec) * BIP_INLINE_VECS,
				     0, SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
521
}