bio-integrity.c 12.5 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
2 3 4
/*
 * bio-integrity.c - bio data integrity extensions
 *
5
 * Copyright (C) 2007, 2008, 2009 Oracle Corporation
6 7 8 9 10
 * Written by: Martin K. Petersen <martin.petersen@oracle.com>
 */

#include <linux/blkdev.h>
#include <linux/mempool.h>
11
#include <linux/export.h>
12 13
#include <linux/bio.h>
#include <linux/workqueue.h>
14
#include <linux/slab.h>
15
#include "blk.h"
16

17
#define BIP_INLINE_VECS	4
18

19
static struct kmem_cache *bip_slab;
20 21
static struct workqueue_struct *kintegrityd_wq;

22 23 24 25 26
void blk_flush_integrity(void)
{
	flush_workqueue(kintegrityd_wq);
}

27 28 29 30 31 32 33 34 35 36 37 38
void __bio_integrity_free(struct bio_set *bs, struct bio_integrity_payload *bip)
{
	if (bs && mempool_initialized(&bs->bio_integrity_pool)) {
		if (bip->bip_vec)
			bvec_free(&bs->bvec_integrity_pool, bip->bip_vec,
				  bip->bip_slab);
		mempool_free(bip, &bs->bio_integrity_pool);
	} else {
		kfree(bip);
	}
}

39
/**
40
 * bio_integrity_alloc - Allocate integrity payload and attach it to bio
41 42 43 44 45 46 47 48
 * @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.
 */
49 50 51
struct bio_integrity_payload *bio_integrity_alloc(struct bio *bio,
						  gfp_t gfp_mask,
						  unsigned int nr_vecs)
52 53
{
	struct bio_integrity_payload *bip;
54
	struct bio_set *bs = bio->bi_pool;
55 56
	unsigned inline_vecs;

57 58 59
	if (WARN_ON_ONCE(bio_has_crypt_ctx(bio)))
		return ERR_PTR(-EOPNOTSUPP);

60
	if (!bs || !mempool_initialized(&bs->bio_integrity_pool)) {
61
		bip = kmalloc(struct_size(bip, bip_inline_vecs, nr_vecs), gfp_mask);
62 63
		inline_vecs = nr_vecs;
	} else {
64
		bip = mempool_alloc(&bs->bio_integrity_pool, gfp_mask);
65
		inline_vecs = BIP_INLINE_VECS;
66 67
	}

68
	if (unlikely(!bip))
69
		return ERR_PTR(-ENOMEM);
70

71 72
	memset(bip, 0, sizeof(*bip));

73
	if (nr_vecs > inline_vecs) {
74 75
		unsigned long idx = 0;

76
		bip->bip_vec = bvec_alloc(gfp_mask, nr_vecs, &idx,
77
					  &bs->bvec_integrity_pool);
78 79
		if (!bip->bip_vec)
			goto err;
80
		bip->bip_max_vcnt = bvec_nr_vecs(idx);
81
		bip->bip_slab = idx;
82 83
	} else {
		bip->bip_vec = bip->bip_inline_vecs;
84
		bip->bip_max_vcnt = inline_vecs;
85 86
	}

87 88
	bip->bip_bio = bio;
	bio->bi_integrity = bip;
J
Jens Axboe 已提交
89
	bio->bi_opf |= REQ_INTEGRITY;
90 91

	return bip;
92
err:
93
	__bio_integrity_free(bs, bip);
94
	return ERR_PTR(-ENOMEM);
95 96 97 98 99 100 101 102 103 104
}
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().
 */
105
void bio_integrity_free(struct bio *bio)
106
{
107
	struct bio_integrity_payload *bip = bio_integrity(bio);
108 109
	struct bio_set *bs = bio->bi_pool;

110
	if (bip->bip_flags & BIP_BLOCK_INTEGRITY)
M
Martin K. Petersen 已提交
111 112
		kfree(page_address(bip->bip_vec->bv_page) +
		      bip->bip_vec->bv_offset);
113

114
	__bio_integrity_free(bs, bip);
115
	bio->bi_integrity = NULL;
116
	bio->bi_opf &= ~REQ_INTEGRITY;
117 118 119 120 121 122 123 124 125 126 127 128 129 130
}

/**
 * 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)
{
131
	struct bio_integrity_payload *bip = bio_integrity(bio);
132 133
	struct bio_vec *iv;

134
	if (bip->bip_vcnt >= bip->bip_max_vcnt) {
135 136 137 138
		printk(KERN_ERR "%s: bip_vec full\n", __func__);
		return 0;
	}

139
	iv = bip->bip_vec + bip->bip_vcnt;
140

141
	if (bip->bip_vcnt &&
142
	    bvec_gap_to_prev(bio->bi_disk->queue,
143 144 145
			     &bip->bip_vec[bip->bip_vcnt - 1], offset))
		return 0;

146 147 148 149 150 151 152 153 154 155
	iv->bv_page = page;
	iv->bv_len = len;
	iv->bv_offset = offset;
	bip->bip_vcnt++;

	return len;
}
EXPORT_SYMBOL(bio_integrity_add_page);

/**
156
 * bio_integrity_process - Process integrity metadata for a bio
157
 * @bio:	bio to generate/verify integrity metadata for
158
 * @proc_iter:  iterator to process
159
 * @proc_fn:	Pointer to the relevant processing function
160
 */
161
static blk_status_t bio_integrity_process(struct bio *bio,
162
		struct bvec_iter *proc_iter, integrity_processing_fn *proc_fn)
163
{
164
	struct blk_integrity *bi = blk_get_integrity(bio->bi_disk);
165
	struct blk_integrity_iter iter;
166 167
	struct bvec_iter bviter;
	struct bio_vec bv;
M
Martin K. Petersen 已提交
168
	struct bio_integrity_payload *bip = bio_integrity(bio);
169
	blk_status_t ret = BLK_STS_OK;
M
Martin K. Petersen 已提交
170 171
	void *prot_buf = page_address(bip->bip_vec->bv_page) +
		bip->bip_vec->bv_offset;
172

173
	iter.disk_name = bio->bi_disk->disk_name;
174
	iter.interval = 1 << bi->interval_exp;
175
	iter.seed = proc_iter->bi_sector;
176
	iter.prot_buf = prot_buf;
177

178
	__bio_for_each_segment(bv, bio, bviter, *proc_iter) {
179
		void *kaddr = kmap_atomic(bv.bv_page);
180

181 182
		iter.data_buf = kaddr + bv.bv_offset;
		iter.data_size = bv.bv_len;
183 184 185 186 187 188

		ret = proc_fn(&iter);
		if (ret) {
			kunmap_atomic(kaddr);
			return ret;
		}
189

190
		kunmap_atomic(kaddr);
191
	}
192 193 194
	return ret;
}

195 196 197 198
/**
 * bio_integrity_prep - Prepare bio for integrity I/O
 * @bio:	bio to prepare
 *
199 200 201 202 203 204
 * Description:  Checks if the bio already has an integrity payload attached.
 * If it does, the payload has been generated by another kernel subsystem,
 * and we just pass it through. Otherwise allocates integrity payload.
 * 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
205 206
 * will be prepared for DMA and a suitable end_io handler set up.
 */
207
bool bio_integrity_prep(struct bio *bio)
208 209
{
	struct bio_integrity_payload *bip;
210 211
	struct blk_integrity *bi = blk_get_integrity(bio->bi_disk);
	struct request_queue *q = bio->bi_disk->queue;
212 213 214 215
	void *buf;
	unsigned long start, end;
	unsigned int len, nr_pages;
	unsigned int bytes, offset, i;
216
	unsigned int intervals;
217
	blk_status_t status;
218

219 220 221
	if (!bi)
		return true;

222 223 224 225 226
	if (bio_op(bio) != REQ_OP_READ && bio_op(bio) != REQ_OP_WRITE)
		return true;

	if (!bio_sectors(bio))
		return true;
227

228 229 230 231 232 233 234 235 236 237 238 239 240
	/* Already protected? */
	if (bio_integrity(bio))
		return true;

	if (bio_data_dir(bio) == READ) {
		if (!bi->profile->verify_fn ||
		    !(bi->flags & BLK_INTEGRITY_VERIFY))
			return true;
	} else {
		if (!bi->profile->generate_fn ||
		    !(bi->flags & BLK_INTEGRITY_GENERATE))
			return true;
	}
241
	intervals = bio_integrity_intervals(bi, bio_sectors(bio));
242 243

	/* Allocate kernel buffer for protection data */
244
	len = intervals * bi->tuple_size;
245
	buf = kmalloc(len, GFP_NOIO | q->bounce_gfp);
246
	status = BLK_STS_RESOURCE;
247 248
	if (unlikely(buf == NULL)) {
		printk(KERN_ERR "could not allocate integrity buffer\n");
249
		goto err_end_io;
250 251 252 253 254 255 256 257
	}

	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);
258
	if (IS_ERR(bip)) {
259 260
		printk(KERN_ERR "could not allocate data integrity bioset\n");
		kfree(buf);
261 262
		status = BLK_STS_RESOURCE;
		goto err_end_io;
263 264
	}

265
	bip->bip_flags |= BIP_BLOCK_INTEGRITY;
266
	bip->bip_iter.bi_size = len;
267
	bip_set_seed(bip, bio->bi_iter.bi_sector);
268

269 270 271
	if (bi->flags & BLK_INTEGRITY_IP_CHECKSUM)
		bip->bip_flags |= BIP_IP_CHECKSUM;

272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
	/* 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);

287 288 289 290 291
		if (ret == 0) {
			printk(KERN_ERR "could not attach integrity payload\n");
			status = BLK_STS_RESOURCE;
			goto err_end_io;
		}
292 293 294 295 296 297 298 299 300 301

		if (ret < bytes)
			break;

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

	/* Auto-generate integrity metadata if this is a write */
302 303 304
	if (bio_data_dir(bio) == WRITE) {
		bio_integrity_process(bio, &bio->bi_iter,
				      bi->profile->generate_fn);
M
Ming Lei 已提交
305 306
	} else {
		bip->bio_iter = bio->bi_iter;
307
	}
308 309 310 311 312 313
	return true;

err_end_io:
	bio->bi_status = status;
	bio_endio(bio);
	return false;
314 315 316 317 318 319 320 321 322 323 324 325 326 327

}
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)
{
328
	struct bio_integrity_payload *bip =
329 330
		container_of(work, struct bio_integrity_payload, bip_work);
	struct bio *bio = bip->bip_bio;
331
	struct blk_integrity *bi = blk_get_integrity(bio->bi_disk);
332 333 334 335 336 337

	/*
	 * At the moment verify is called bio's iterator was advanced
	 * during split and completion, we need to rewind iterator to
	 * it's original position.
	 */
M
Ming Lei 已提交
338 339
	bio->bi_status = bio_integrity_process(bio, &bip->bio_iter,
						bi->profile->verify_fn);
340
	bio_integrity_free(bio);
341
	bio_endio(bio);
342 343 344
}

/**
345
 * __bio_integrity_endio - Integrity I/O completion function
346 347 348 349 350 351 352 353 354
 * @bio:	Protected bio
 *
 * 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.
 */
355
bool __bio_integrity_endio(struct bio *bio)
356
{
357
	struct blk_integrity *bi = blk_get_integrity(bio->bi_disk);
358
	struct bio_integrity_payload *bip = bio_integrity(bio);
359 360

	if (bio_op(bio) == REQ_OP_READ && !bio->bi_status &&
361
	    (bip->bip_flags & BIP_BLOCK_INTEGRITY) && bi->profile->verify_fn) {
362 363 364
		INIT_WORK(&bip->bip_work, bio_integrity_verify_fn);
		queue_work(kintegrityd_wq, &bip->bip_work);
		return false;
365 366
	}

367 368
	bio_integrity_free(bio);
	return true;
369 370 371 372 373 374 375 376 377 378 379 380 381
}

/**
 * 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)
{
382
	struct bio_integrity_payload *bip = bio_integrity(bio);
383
	struct blk_integrity *bi = blk_get_integrity(bio->bi_disk);
384
	unsigned bytes = bio_integrity_bytes(bi, bytes_done >> 9);
385

386
	bip->bip_iter.bi_sector += bytes_done >> 9;
387
	bvec_iter_advance(bip->bip_vec, &bip->bip_iter, bytes);
388 389 390 391 392 393 394 395
}

/**
 * bio_integrity_trim - Trim integrity vector
 * @bio:	bio whose integrity vector to update
 *
 * Description: Used to trim the integrity vector in a cloned bio.
 */
396
void bio_integrity_trim(struct bio *bio)
397
{
398
	struct bio_integrity_payload *bip = bio_integrity(bio);
399
	struct blk_integrity *bi = blk_get_integrity(bio->bi_disk);
400

401
	bip->bip_iter.bi_size = bio_integrity_bytes(bi, bio_sectors(bio));
402 403 404 405 406 407 408
}
EXPORT_SYMBOL(bio_integrity_trim);

/**
 * bio_integrity_clone - Callback for cloning bios with integrity metadata
 * @bio:	New bio
 * @bio_src:	Original bio
409
 * @gfp_mask:	Memory allocation mask
410 411 412
 *
 * Description:	Called to allocate a bip when cloning a bio
 */
413
int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
414
			gfp_t gfp_mask)
415
{
416
	struct bio_integrity_payload *bip_src = bio_integrity(bio_src);
417 418 419 420
	struct bio_integrity_payload *bip;

	BUG_ON(bip_src == NULL);

421
	bip = bio_integrity_alloc(bio, gfp_mask, bip_src->bip_vcnt);
422 423
	if (IS_ERR(bip))
		return PTR_ERR(bip);
424 425 426 427 428

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

	bip->bip_vcnt = bip_src->bip_vcnt;
429
	bip->bip_iter = bip_src->bip_iter;
430 431 432 433 434

	return 0;
}
EXPORT_SYMBOL(bio_integrity_clone);

435
int bioset_integrity_create(struct bio_set *bs, int pool_size)
436
{
437
	if (mempool_initialized(&bs->bio_integrity_pool))
438 439
		return 0;

440 441
	if (mempool_init_slab_pool(&bs->bio_integrity_pool,
				   pool_size, bip_slab))
442
		return -1;
443

444 445
	if (biovec_init_pool(&bs->bvec_integrity_pool, pool_size)) {
		mempool_exit(&bs->bio_integrity_pool);
446
		return -1;
447
	}
448 449 450 451 452 453 454

	return 0;
}
EXPORT_SYMBOL(bioset_integrity_create);

void bioset_integrity_free(struct bio_set *bs)
{
455 456
	mempool_exit(&bs->bio_integrity_pool);
	mempool_exit(&bs->bvec_integrity_pool);
457 458 459 460
}

void __init bio_integrity_init(void)
{
461 462 463 464 465 466
	/*
	 * 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);
467 468
	if (!kintegrityd_wq)
		panic("Failed to create kintegrityd\n");
469

470 471 472 473
	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);
474
}