1/*
2 * blk-integrity.c - Block layer data integrity extensions
3 *
4 * Copyright (C) 2007, 2008 Oracle Corporation
5 * Written by: Martin K. Petersen <martin.petersen@oracle.com>
6 *
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.
10 *
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.
15 *
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,
19 * USA.
20 *
21 */
22
23#include <linux/blkdev.h>
24#include <linux/mempool.h>
25#include <linux/bio.h>
26#include <linux/scatterlist.h>
27#include <linux/export.h>
28#include <linux/slab.h>
29
30#include "blk.h"
31
32static struct kmem_cache *integrity_cachep;
33
34static const char *bi_unsupported_name = "unsupported";
35
36/**
37 * blk_rq_count_integrity_sg - Count number of integrity scatterlist elements
38 * @q:		request queue
39 * @bio:	bio with integrity metadata attached
40 *
41 * Description: Returns the number of elements required in a
42 * scatterlist corresponding to the integrity metadata in a bio.
43 */
44int blk_rq_count_integrity_sg(struct request_queue *q, struct bio *bio)
45{
46	struct bio_vec iv, ivprv = { NULL };
47	unsigned int segments = 0;
48	unsigned int seg_size = 0;
49	struct bvec_iter iter;
50	int prev = 0;
51
52	bio_for_each_integrity_vec(iv, bio, iter) {
53
54		if (prev) {
55			if (!BIOVEC_PHYS_MERGEABLE(&ivprv, &iv))
56				goto new_segment;
57
58			if (!BIOVEC_SEG_BOUNDARY(q, &ivprv, &iv))
59				goto new_segment;
60
61			if (seg_size + iv.bv_len > queue_max_segment_size(q))
62				goto new_segment;
63
64			seg_size += iv.bv_len;
65		} else {
66new_segment:
67			segments++;
68			seg_size = iv.bv_len;
69		}
70
71		prev = 1;
72		ivprv = iv;
73	}
74
75	return segments;
76}
77EXPORT_SYMBOL(blk_rq_count_integrity_sg);
78
79/**
80 * blk_rq_map_integrity_sg - Map integrity metadata into a scatterlist
81 * @q:		request queue
82 * @bio:	bio with integrity metadata attached
83 * @sglist:	target scatterlist
84 *
85 * Description: Map the integrity vectors in request into a
86 * scatterlist.  The scatterlist must be big enough to hold all
87 * elements.  I.e. sized using blk_rq_count_integrity_sg().
88 */
89int blk_rq_map_integrity_sg(struct request_queue *q, struct bio *bio,
90			    struct scatterlist *sglist)
91{
92	struct bio_vec iv, ivprv = { NULL };
93	struct scatterlist *sg = NULL;
94	unsigned int segments = 0;
95	struct bvec_iter iter;
96	int prev = 0;
97
98	bio_for_each_integrity_vec(iv, bio, iter) {
99
100		if (prev) {
101			if (!BIOVEC_PHYS_MERGEABLE(&ivprv, &iv))
102				goto new_segment;
103
104			if (!BIOVEC_SEG_BOUNDARY(q, &ivprv, &iv))
105				goto new_segment;
106
107			if (sg->length + iv.bv_len > queue_max_segment_size(q))
108				goto new_segment;
109
110			sg->length += iv.bv_len;
111		} else {
112new_segment:
113			if (!sg)
114				sg = sglist;
115			else {
116				sg_unmark_end(sg);
117				sg = sg_next(sg);
118			}
119
120			sg_set_page(sg, iv.bv_page, iv.bv_len, iv.bv_offset);
121			segments++;
122		}
123
124		prev = 1;
125		ivprv = iv;
126	}
127
128	if (sg)
129		sg_mark_end(sg);
130
131	return segments;
132}
133EXPORT_SYMBOL(blk_rq_map_integrity_sg);
134
135/**
136 * blk_integrity_compare - Compare integrity profile of two disks
137 * @gd1:	Disk to compare
138 * @gd2:	Disk to compare
139 *
140 * Description: Meta-devices like DM and MD need to verify that all
141 * sub-devices use the same integrity format before advertising to
142 * upper layers that they can send/receive integrity metadata.  This
143 * function can be used to check whether two gendisk devices have
144 * compatible integrity formats.
145 */
146int blk_integrity_compare(struct gendisk *gd1, struct gendisk *gd2)
147{
148	struct blk_integrity *b1 = gd1->integrity;
149	struct blk_integrity *b2 = gd2->integrity;
150
151	if (!b1 && !b2)
152		return 0;
153
154	if (!b1 || !b2)
155		return -1;
156
157	if (b1->interval != b2->interval) {
158		pr_err("%s: %s/%s protection interval %u != %u\n",
159		       __func__, gd1->disk_name, gd2->disk_name,
160		       b1->interval, b2->interval);
161		return -1;
162	}
163
164	if (b1->tuple_size != b2->tuple_size) {
165		printk(KERN_ERR "%s: %s/%s tuple sz %u != %u\n", __func__,
166		       gd1->disk_name, gd2->disk_name,
167		       b1->tuple_size, b2->tuple_size);
168		return -1;
169	}
170
171	if (b1->tag_size && b2->tag_size && (b1->tag_size != b2->tag_size)) {
172		printk(KERN_ERR "%s: %s/%s tag sz %u != %u\n", __func__,
173		       gd1->disk_name, gd2->disk_name,
174		       b1->tag_size, b2->tag_size);
175		return -1;
176	}
177
178	if (strcmp(b1->name, b2->name)) {
179		printk(KERN_ERR "%s: %s/%s type %s != %s\n", __func__,
180		       gd1->disk_name, gd2->disk_name,
181		       b1->name, b2->name);
182		return -1;
183	}
184
185	return 0;
186}
187EXPORT_SYMBOL(blk_integrity_compare);
188
189bool blk_integrity_merge_rq(struct request_queue *q, struct request *req,
190			    struct request *next)
191{
192	if (blk_integrity_rq(req) == 0 && blk_integrity_rq(next) == 0)
193		return true;
194
195	if (blk_integrity_rq(req) == 0 || blk_integrity_rq(next) == 0)
196		return false;
197
198	if (bio_integrity(req->bio)->bip_flags !=
199	    bio_integrity(next->bio)->bip_flags)
200		return false;
201
202	if (req->nr_integrity_segments + next->nr_integrity_segments >
203	    q->limits.max_integrity_segments)
204		return false;
205
206	return true;
207}
208EXPORT_SYMBOL(blk_integrity_merge_rq);
209
210bool blk_integrity_merge_bio(struct request_queue *q, struct request *req,
211			     struct bio *bio)
212{
213	int nr_integrity_segs;
214	struct bio *next = bio->bi_next;
215
216	if (blk_integrity_rq(req) == 0 && bio_integrity(bio) == NULL)
217		return true;
218
219	if (blk_integrity_rq(req) == 0 || bio_integrity(bio) == NULL)
220		return false;
221
222	if (bio_integrity(req->bio)->bip_flags != bio_integrity(bio)->bip_flags)
223		return false;
224
225	bio->bi_next = NULL;
226	nr_integrity_segs = blk_rq_count_integrity_sg(q, bio);
227	bio->bi_next = next;
228
229	if (req->nr_integrity_segments + nr_integrity_segs >
230	    q->limits.max_integrity_segments)
231		return false;
232
233	req->nr_integrity_segments += nr_integrity_segs;
234
235	return true;
236}
237EXPORT_SYMBOL(blk_integrity_merge_bio);
238
239struct integrity_sysfs_entry {
240	struct attribute attr;
241	ssize_t (*show)(struct blk_integrity *, char *);
242	ssize_t (*store)(struct blk_integrity *, const char *, size_t);
243};
244
245static ssize_t integrity_attr_show(struct kobject *kobj, struct attribute *attr,
246				   char *page)
247{
248	struct blk_integrity *bi =
249		container_of(kobj, struct blk_integrity, kobj);
250	struct integrity_sysfs_entry *entry =
251		container_of(attr, struct integrity_sysfs_entry, attr);
252
253	return entry->show(bi, page);
254}
255
256static ssize_t integrity_attr_store(struct kobject *kobj,
257				    struct attribute *attr, const char *page,
258				    size_t count)
259{
260	struct blk_integrity *bi =
261		container_of(kobj, struct blk_integrity, kobj);
262	struct integrity_sysfs_entry *entry =
263		container_of(attr, struct integrity_sysfs_entry, attr);
264	ssize_t ret = 0;
265
266	if (entry->store)
267		ret = entry->store(bi, page, count);
268
269	return ret;
270}
271
272static ssize_t integrity_format_show(struct blk_integrity *bi, char *page)
273{
274	if (bi != NULL && bi->name != NULL)
275		return sprintf(page, "%s\n", bi->name);
276	else
277		return sprintf(page, "none\n");
278}
279
280static ssize_t integrity_tag_size_show(struct blk_integrity *bi, char *page)
281{
282	if (bi != NULL)
283		return sprintf(page, "%u\n", bi->tag_size);
284	else
285		return sprintf(page, "0\n");
286}
287
288static ssize_t integrity_verify_store(struct blk_integrity *bi,
289				      const char *page, size_t count)
290{
291	char *p = (char *) page;
292	unsigned long val = simple_strtoul(p, &p, 10);
293
294	if (val)
295		bi->flags |= BLK_INTEGRITY_VERIFY;
296	else
297		bi->flags &= ~BLK_INTEGRITY_VERIFY;
298
299	return count;
300}
301
302static ssize_t integrity_verify_show(struct blk_integrity *bi, char *page)
303{
304	return sprintf(page, "%d\n", (bi->flags & BLK_INTEGRITY_VERIFY) != 0);
305}
306
307static ssize_t integrity_generate_store(struct blk_integrity *bi,
308					const char *page, size_t count)
309{
310	char *p = (char *) page;
311	unsigned long val = simple_strtoul(p, &p, 10);
312
313	if (val)
314		bi->flags |= BLK_INTEGRITY_GENERATE;
315	else
316		bi->flags &= ~BLK_INTEGRITY_GENERATE;
317
318	return count;
319}
320
321static ssize_t integrity_generate_show(struct blk_integrity *bi, char *page)
322{
323	return sprintf(page, "%d\n", (bi->flags & BLK_INTEGRITY_GENERATE) != 0);
324}
325
326static ssize_t integrity_device_show(struct blk_integrity *bi, char *page)
327{
328	return sprintf(page, "%u\n",
329		       (bi->flags & BLK_INTEGRITY_DEVICE_CAPABLE) != 0);
330}
331
332static struct integrity_sysfs_entry integrity_format_entry = {
333	.attr = { .name = "format", .mode = S_IRUGO },
334	.show = integrity_format_show,
335};
336
337static struct integrity_sysfs_entry integrity_tag_size_entry = {
338	.attr = { .name = "tag_size", .mode = S_IRUGO },
339	.show = integrity_tag_size_show,
340};
341
342static struct integrity_sysfs_entry integrity_verify_entry = {
343	.attr = { .name = "read_verify", .mode = S_IRUGO | S_IWUSR },
344	.show = integrity_verify_show,
345	.store = integrity_verify_store,
346};
347
348static struct integrity_sysfs_entry integrity_generate_entry = {
349	.attr = { .name = "write_generate", .mode = S_IRUGO | S_IWUSR },
350	.show = integrity_generate_show,
351	.store = integrity_generate_store,
352};
353
354static struct integrity_sysfs_entry integrity_device_entry = {
355	.attr = { .name = "device_is_integrity_capable", .mode = S_IRUGO },
356	.show = integrity_device_show,
357};
358
359static struct attribute *integrity_attrs[] = {
360	&integrity_format_entry.attr,
361	&integrity_tag_size_entry.attr,
362	&integrity_verify_entry.attr,
363	&integrity_generate_entry.attr,
364	&integrity_device_entry.attr,
365	NULL,
366};
367
368static const struct sysfs_ops integrity_ops = {
369	.show	= &integrity_attr_show,
370	.store	= &integrity_attr_store,
371};
372
373static int __init blk_dev_integrity_init(void)
374{
375	integrity_cachep = kmem_cache_create("blkdev_integrity",
376					     sizeof(struct blk_integrity),
377					     0, SLAB_PANIC, NULL);
378	return 0;
379}
380subsys_initcall(blk_dev_integrity_init);
381
382static void blk_integrity_release(struct kobject *kobj)
383{
384	struct blk_integrity *bi =
385		container_of(kobj, struct blk_integrity, kobj);
386
387	kmem_cache_free(integrity_cachep, bi);
388}
389
390static struct kobj_type integrity_ktype = {
391	.default_attrs	= integrity_attrs,
392	.sysfs_ops	= &integrity_ops,
393	.release	= blk_integrity_release,
394};
395
396bool blk_integrity_is_initialized(struct gendisk *disk)
397{
398	struct blk_integrity *bi = blk_get_integrity(disk);
399
400	return (bi && bi->name && strcmp(bi->name, bi_unsupported_name) != 0);
401}
402EXPORT_SYMBOL(blk_integrity_is_initialized);
403
404/**
405 * blk_integrity_register - Register a gendisk as being integrity-capable
406 * @disk:	struct gendisk pointer to make integrity-aware
407 * @template:	optional integrity profile to register
408 *
409 * Description: When a device needs to advertise itself as being able
410 * to send/receive integrity metadata it must use this function to
411 * register the capability with the block layer.  The template is a
412 * blk_integrity struct with values appropriate for the underlying
413 * hardware.  If template is NULL the new profile is allocated but
414 * not filled out. See Documentation/block/data-integrity.txt.
415 */
416int blk_integrity_register(struct gendisk *disk, struct blk_integrity *template)
417{
418	struct blk_integrity *bi;
419
420	BUG_ON(disk == NULL);
421
422	if (disk->integrity == NULL) {
423		bi = kmem_cache_alloc(integrity_cachep,
424				      GFP_KERNEL | __GFP_ZERO);
425		if (!bi)
426			return -1;
427
428		if (kobject_init_and_add(&bi->kobj, &integrity_ktype,
429					 &disk_to_dev(disk)->kobj,
430					 "%s", "integrity")) {
431			kmem_cache_free(integrity_cachep, bi);
432			return -1;
433		}
434
435		kobject_uevent(&bi->kobj, KOBJ_ADD);
436
437		bi->flags |= BLK_INTEGRITY_VERIFY | BLK_INTEGRITY_GENERATE;
438		bi->interval = queue_logical_block_size(disk->queue);
439		disk->integrity = bi;
440	} else
441		bi = disk->integrity;
442
443	/* Use the provided profile as template */
444	if (template != NULL) {
445		bi->name = template->name;
446		bi->generate_fn = template->generate_fn;
447		bi->verify_fn = template->verify_fn;
448		bi->tuple_size = template->tuple_size;
449		bi->tag_size = template->tag_size;
450		bi->flags |= template->flags;
451	} else
452		bi->name = bi_unsupported_name;
453
454	disk->queue->backing_dev_info.capabilities |= BDI_CAP_STABLE_WRITES;
455
456	return 0;
457}
458EXPORT_SYMBOL(blk_integrity_register);
459
460/**
461 * blk_integrity_unregister - Remove block integrity profile
462 * @disk:	disk whose integrity profile to deallocate
463 *
464 * Description: This function frees all memory used by the block
465 * integrity profile.  To be called at device teardown.
466 */
467void blk_integrity_unregister(struct gendisk *disk)
468{
469	struct blk_integrity *bi;
470
471	if (!disk || !disk->integrity)
472		return;
473
474	disk->queue->backing_dev_info.capabilities &= ~BDI_CAP_STABLE_WRITES;
475
476	bi = disk->integrity;
477
478	kobject_uevent(&bi->kobj, KOBJ_REMOVE);
479	kobject_del(&bi->kobj);
480	kobject_put(&bi->kobj);
481	disk->integrity = NULL;
482}
483EXPORT_SYMBOL(blk_integrity_unregister);
484