2 * bio-integrity.c - bio data integrity extensions
4 * Copyright (C) 2007, 2008, 2009 Oracle Corporation
5 * Written by: Martin K. Petersen <martin.petersen@oracle.com>
7 * This program is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU General Public License version
9 * 2 as published by the Free Software Foundation.
11 * This program is distributed in the hope that it will be useful, but
12 * WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * General Public License for more details.
16 * You should have received a copy of the GNU General Public License
17 * along with this program; see the file COPYING. If not, write to
18 * the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139,
23 #include <linux/blkdev.h>
24 #include <linux/mempool.h>
25 #include <linux/export.h>
26 #include <linux/bio.h>
27 #include <linux/workqueue.h>
28 #include <linux/slab.h>
30 #define BIP_INLINE_VECS 4
32 static struct kmem_cache *bip_slab;
33 static struct workqueue_struct *kintegrityd_wq;
36 * bio_integrity_alloc - Allocate integrity payload and attach it to bio
37 * @bio: bio to attach integrity metadata to
38 * @gfp_mask: Memory allocation mask
39 * @nr_vecs: Number of integrity metadata scatter-gather elements
41 * Description: This function prepares a bio for attaching integrity
42 * metadata. nr_vecs specifies the maximum number of pages containing
43 * integrity metadata that can be attached.
45 struct bio_integrity_payload *bio_integrity_alloc(struct bio *bio,
49 struct bio_integrity_payload *bip;
50 struct bio_set *bs = bio->bi_pool;
51 unsigned long idx = BIO_POOL_NONE;
55 bip = kmalloc(sizeof(struct bio_integrity_payload) +
56 sizeof(struct bio_vec) * nr_vecs, gfp_mask);
57 inline_vecs = nr_vecs;
59 bip = mempool_alloc(bs->bio_integrity_pool, gfp_mask);
60 inline_vecs = BIP_INLINE_VECS;
66 memset(bip, 0, sizeof(*bip));
68 if (nr_vecs > inline_vecs) {
69 bip->bip_vec = bvec_alloc(gfp_mask, nr_vecs, &idx,
70 bs->bvec_integrity_pool);
73 bip->bip_max_vcnt = bvec_nr_vecs(idx);
75 bip->bip_vec = bip->bip_inline_vecs;
76 bip->bip_max_vcnt = inline_vecs;
81 bio->bi_integrity = bip;
82 bio->bi_rw |= REQ_INTEGRITY;
86 mempool_free(bip, bs->bio_integrity_pool);
89 EXPORT_SYMBOL(bio_integrity_alloc);
92 * bio_integrity_free - Free bio integrity payload
93 * @bio: bio containing bip to be freed
95 * Description: Used to free the integrity portion of a bio. Usually
96 * called from bio_free().
98 void bio_integrity_free(struct bio *bio)
100 struct bio_integrity_payload *bip = bio_integrity(bio);
101 struct bio_set *bs = bio->bi_pool;
103 if (bip->bip_owns_buf)
104 kfree(page_address(bip->bip_vec->bv_page) +
105 bip->bip_vec->bv_offset);
108 if (bip->bip_slab != BIO_POOL_NONE)
109 bvec_free(bs->bvec_integrity_pool, bip->bip_vec,
112 mempool_free(bip, bs->bio_integrity_pool);
117 bio->bi_integrity = NULL;
119 EXPORT_SYMBOL(bio_integrity_free);
122 * bio_integrity_add_page - Attach integrity metadata
123 * @bio: bio to update
124 * @page: page containing integrity metadata
125 * @len: number of bytes of integrity metadata in page
126 * @offset: start offset within page
128 * Description: Attach a page containing integrity metadata to bio.
130 int bio_integrity_add_page(struct bio *bio, struct page *page,
131 unsigned int len, unsigned int offset)
133 struct bio_integrity_payload *bip = bio_integrity(bio);
136 if (bip->bip_vcnt >= bip->bip_max_vcnt) {
137 printk(KERN_ERR "%s: bip_vec full\n", __func__);
141 iv = bip->bip_vec + bip->bip_vcnt;
145 iv->bv_offset = offset;
150 EXPORT_SYMBOL(bio_integrity_add_page);
153 * bio_integrity_enabled - Check whether integrity can be passed
156 * Description: Determines whether bio_integrity_prep() can be called
157 * on this bio or not. bio data direction and target device must be
158 * set prior to calling. The functions honors the write_generate and
159 * read_verify flags in sysfs.
161 bool bio_integrity_enabled(struct bio *bio)
163 struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
168 /* Already protected? */
169 if (bio_integrity(bio))
175 if (bio_data_dir(bio) == READ && bi->verify_fn != NULL &&
176 (bi->flags & INTEGRITY_FLAG_READ))
179 if (bio_data_dir(bio) == WRITE && bi->generate_fn != NULL &&
180 (bi->flags & INTEGRITY_FLAG_WRITE))
185 EXPORT_SYMBOL(bio_integrity_enabled);
188 * bio_integrity_intervals - Return number of integrity intervals for a bio
189 * @bi: blk_integrity profile for device
190 * @sectors: Size of the bio in 512-byte sectors
192 * Description: The block layer calculates everything in 512 byte
193 * sectors but integrity metadata is done in terms of the data integrity
194 * interval size of the storage device. Convert the block layer sectors
195 * to the appropriate number of integrity intervals.
197 static inline unsigned int bio_integrity_intervals(struct blk_integrity *bi,
198 unsigned int sectors)
200 return sectors >> (ilog2(bi->interval) - 9);
203 static inline unsigned int bio_integrity_bytes(struct blk_integrity *bi,
204 unsigned int sectors)
206 return bio_integrity_intervals(bi, sectors) * bi->tuple_size;
210 * bio_integrity_process - Process integrity metadata for a bio
211 * @bio: bio to generate/verify integrity metadata for
212 * @proc_fn: Pointer to the relevant processing function
214 static int bio_integrity_process(struct bio *bio,
215 integrity_processing_fn *proc_fn)
217 struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
218 struct blk_integrity_iter iter;
220 struct bio_integrity_payload *bip = bio_integrity(bio);
221 unsigned int i, ret = 0;
222 void *prot_buf = page_address(bip->bip_vec->bv_page) +
223 bip->bip_vec->bv_offset;
225 iter.disk_name = bio->bi_bdev->bd_disk->disk_name;
226 iter.interval = bi->interval;
227 iter.seed = bip_get_seed(bip);
228 iter.prot_buf = prot_buf;
230 bio_for_each_segment_all(bv, bio, i) {
231 void *kaddr = kmap_atomic(bv->bv_page);
233 iter.data_buf = kaddr + bv->bv_offset;
234 iter.data_size = bv->bv_len;
236 ret = proc_fn(&iter);
238 kunmap_atomic(kaddr);
242 kunmap_atomic(kaddr);
248 * bio_integrity_prep - Prepare bio for integrity I/O
249 * @bio: bio to prepare
251 * Description: Allocates a buffer for integrity metadata, maps the
252 * pages and attaches them to a bio. The bio must have data
253 * direction, target device and start sector set priot to calling. In
254 * the WRITE case, integrity metadata will be generated using the
255 * block device's integrity function. In the READ case, the buffer
256 * will be prepared for DMA and a suitable end_io handler set up.
258 int bio_integrity_prep(struct bio *bio)
260 struct bio_integrity_payload *bip;
261 struct blk_integrity *bi;
262 struct request_queue *q;
264 unsigned long start, end;
265 unsigned int len, nr_pages;
266 unsigned int bytes, offset, i;
267 unsigned int intervals;
269 bi = bdev_get_integrity(bio->bi_bdev);
270 q = bdev_get_queue(bio->bi_bdev);
272 BUG_ON(bio_integrity(bio));
274 intervals = bio_integrity_intervals(bi, bio_sectors(bio));
276 /* Allocate kernel buffer for protection data */
277 len = intervals * bi->tuple_size;
278 buf = kmalloc(len, GFP_NOIO | q->bounce_gfp);
279 if (unlikely(buf == NULL)) {
280 printk(KERN_ERR "could not allocate integrity buffer\n");
284 end = (((unsigned long) buf) + len + PAGE_SIZE - 1) >> PAGE_SHIFT;
285 start = ((unsigned long) buf) >> PAGE_SHIFT;
286 nr_pages = end - start;
288 /* Allocate bio integrity payload and integrity vectors */
289 bip = bio_integrity_alloc(bio, GFP_NOIO, nr_pages);
290 if (unlikely(bip == NULL)) {
291 printk(KERN_ERR "could not allocate data integrity bioset\n");
296 bip->bip_owns_buf = 1;
297 bip->bip_iter.bi_size = len;
298 bip_set_seed(bip, bio->bi_iter.bi_sector);
301 offset = offset_in_page(buf);
302 for (i = 0 ; i < nr_pages ; i++) {
304 bytes = PAGE_SIZE - offset;
312 ret = bio_integrity_add_page(bio, virt_to_page(buf),
326 /* Install custom I/O completion handler if read verify is enabled */
327 if (bio_data_dir(bio) == READ) {
328 bip->bip_end_io = bio->bi_end_io;
329 bio->bi_end_io = bio_integrity_endio;
332 /* Auto-generate integrity metadata if this is a write */
333 if (bio_data_dir(bio) == WRITE)
334 bio_integrity_process(bio, bi->generate_fn);
338 EXPORT_SYMBOL(bio_integrity_prep);
341 * bio_integrity_verify_fn - Integrity I/O completion worker
342 * @work: Work struct stored in bio to be verified
344 * Description: This workqueue function is called to complete a READ
345 * request. The function verifies the transferred integrity metadata
346 * and then calls the original bio end_io function.
348 static void bio_integrity_verify_fn(struct work_struct *work)
350 struct bio_integrity_payload *bip =
351 container_of(work, struct bio_integrity_payload, bip_work);
352 struct bio *bio = bip->bip_bio;
353 struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
356 error = bio_integrity_process(bio, bi->verify_fn);
358 /* Restore original bio completion handler */
359 bio->bi_end_io = bip->bip_end_io;
360 bio_endio_nodec(bio, error);
364 * bio_integrity_endio - Integrity I/O completion function
365 * @bio: Protected bio
366 * @error: Pointer to errno
368 * Description: Completion for integrity I/O
370 * Normally I/O completion is done in interrupt context. However,
371 * verifying I/O integrity is a time-consuming task which must be run
372 * in process context. This function postpones completion
375 void bio_integrity_endio(struct bio *bio, int error)
377 struct bio_integrity_payload *bip = bio_integrity(bio);
379 BUG_ON(bip->bip_bio != bio);
381 /* In case of an I/O error there is no point in verifying the
382 * integrity metadata. Restore original bio end_io handler
386 bio->bi_end_io = bip->bip_end_io;
387 bio_endio_nodec(bio, error);
392 INIT_WORK(&bip->bip_work, bio_integrity_verify_fn);
393 queue_work(kintegrityd_wq, &bip->bip_work);
395 EXPORT_SYMBOL(bio_integrity_endio);
398 * bio_integrity_advance - Advance integrity vector
399 * @bio: bio whose integrity vector to update
400 * @bytes_done: number of data bytes that have been completed
402 * Description: This function calculates how many integrity bytes the
403 * number of completed data bytes correspond to and advances the
404 * integrity vector accordingly.
406 void bio_integrity_advance(struct bio *bio, unsigned int bytes_done)
408 struct bio_integrity_payload *bip = bio_integrity(bio);
409 struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
410 unsigned bytes = bio_integrity_bytes(bi, bytes_done >> 9);
412 bvec_iter_advance(bip->bip_vec, &bip->bip_iter, bytes);
414 EXPORT_SYMBOL(bio_integrity_advance);
417 * bio_integrity_trim - Trim integrity vector
418 * @bio: bio whose integrity vector to update
419 * @offset: offset to first data sector
420 * @sectors: number of data sectors
422 * Description: Used to trim the integrity vector in a cloned bio.
423 * The ivec will be advanced corresponding to 'offset' data sectors
424 * and the length will be truncated corresponding to 'len' data
427 void bio_integrity_trim(struct bio *bio, unsigned int offset,
428 unsigned int sectors)
430 struct bio_integrity_payload *bip = bio_integrity(bio);
431 struct blk_integrity *bi = bdev_get_integrity(bio->bi_bdev);
433 bio_integrity_advance(bio, offset << 9);
434 bip->bip_iter.bi_size = bio_integrity_bytes(bi, sectors);
436 EXPORT_SYMBOL(bio_integrity_trim);
439 * bio_integrity_clone - Callback for cloning bios with integrity metadata
441 * @bio_src: Original bio
442 * @gfp_mask: Memory allocation mask
444 * Description: Called to allocate a bip when cloning a bio
446 int bio_integrity_clone(struct bio *bio, struct bio *bio_src,
449 struct bio_integrity_payload *bip_src = bio_integrity(bio_src);
450 struct bio_integrity_payload *bip;
452 BUG_ON(bip_src == NULL);
454 bip = bio_integrity_alloc(bio, gfp_mask, bip_src->bip_vcnt);
459 memcpy(bip->bip_vec, bip_src->bip_vec,
460 bip_src->bip_vcnt * sizeof(struct bio_vec));
462 bip->bip_vcnt = bip_src->bip_vcnt;
463 bip->bip_iter = bip_src->bip_iter;
467 EXPORT_SYMBOL(bio_integrity_clone);
469 int bioset_integrity_create(struct bio_set *bs, int pool_size)
471 if (bs->bio_integrity_pool)
474 bs->bio_integrity_pool = mempool_create_slab_pool(pool_size, bip_slab);
475 if (!bs->bio_integrity_pool)
478 bs->bvec_integrity_pool = biovec_create_pool(pool_size);
479 if (!bs->bvec_integrity_pool) {
480 mempool_destroy(bs->bio_integrity_pool);
486 EXPORT_SYMBOL(bioset_integrity_create);
488 void bioset_integrity_free(struct bio_set *bs)
490 if (bs->bio_integrity_pool)
491 mempool_destroy(bs->bio_integrity_pool);
493 if (bs->bvec_integrity_pool)
494 mempool_destroy(bs->bvec_integrity_pool);
496 EXPORT_SYMBOL(bioset_integrity_free);
498 void __init bio_integrity_init(void)
501 * kintegrityd won't block much but may burn a lot of CPU cycles.
502 * Make it highpri CPU intensive wq with max concurrency of 1.
504 kintegrityd_wq = alloc_workqueue("kintegrityd", WQ_MEM_RECLAIM |
505 WQ_HIGHPRI | WQ_CPU_INTENSIVE, 1);
507 panic("Failed to create kintegrityd\n");
509 bip_slab = kmem_cache_create("bio_integrity_payload",
510 sizeof(struct bio_integrity_payload) +
511 sizeof(struct bio_vec) * BIP_INLINE_VECS,
512 0, SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
514 panic("Failed to create slab\n");