block: move integrity information into queue_limits

Move the integrity information into the queue limits so that it can be
set atomically with other queue limits, and that the sysfs changes to
the read_verify and write_generate flags are properly synchronized.
This also allows to provide a more useful helper to stack the integrity
fields, although it still is separate from the main stacking function
as not all stackable devices want to inherit the integrity settings.
Even with that it greatly simplifies the code in md and dm.

Note that the integrity field is moved as-is into the queue limits.
While there are good arguments for removing the separate blk_integrity
structure, this would cause a lot of churn and might better be done at a
later time if desired. However the integrity field in the queue_limits
structure is now unconditional so that various ifdefs can be avoided or
replaced with IS_ENABLED(). Given that tiny size of it that seems like
a worthwhile trade off.

Signed-off-by: Christoph Hellwig <[e-mail]>
Reviewed-by: Hannes Reinecke <[e-mail]>
Reviewed-by: Martin K. Petersen <[e-mail]>
Link: https://lore.kernel.org/r/[e-mail]
Signed-off-by: Jens Axboe <[e-mail]>

Documentation/block/data-integrity.rst

···
 4.2 Block Device
 ----------------
 
-Because the format of the protection data is tied to the physical
-disk, each block device has been extended with a block integrity
-profile (struct blk_integrity).  This optional profile is registered
-with the block layer using blk_integrity_register().
-
-The profile contains callback functions for generating and verifying
-the protection data, as well as getting and setting application tags.
-The profile also contains a few constants to aid in completing,
-merging and splitting the integrity metadata.
+Block devices can set up the integrity information in the integrity
+sub-struture of the queue_limits structure.
 
 Layered block devices will need to pick a profile that's appropriate
-for all subdevices.  blk_integrity_compare() can help with that.  DM
+for all subdevices.  queue_limits_stack_integrity() can help with that.  DM
 and MD linear, RAID0 and RAID1 are currently supported.  RAID4/5/6
 will require extra work due to the application tag.
 
···
       It is up to the receiver to process them and verify data
       integrity upon completion.
 
-
-5.4 Registering A Block Device As Capable Of Exchanging Integrity Metadata
---------------------------------------------------------------------------
-
-    To enable integrity exchange on a block device the gendisk must be
-    registered as capable:
-
-    `int blk_integrity_register(gendisk, blk_integrity);`
-
-      The blk_integrity struct is a template and should contain the
-      following::
-
-        static struct blk_integrity my_profile = {
-            .name                   = "STANDARDSBODY-TYPE-VARIANT-CSUM",
-            .generate_fn            = my_generate_fn,
-	    .verify_fn              = my_verify_fn,
-	    .tuple_size             = sizeof(struct my_tuple_size),
-	    .tag_size               = <tag bytes per hw sector>,
-        };
-
-      'name' is a text string which will be visible in sysfs.  This is
-      part of the userland API so chose it carefully and never change
-      it.  The format is standards body-type-variant.
-      E.g. T10-DIF-TYPE1-IP or T13-EPP-0-CRC.
-
-      'generate_fn' generates appropriate integrity metadata (for WRITE).
-
-      'verify_fn' verifies that the data buffer matches the integrity
-      metadata.
-
-      'tuple_size' must be set to match the size of the integrity
-      metadata per sector.  I.e. 8 for DIF and EPP.
-
-      'tag_size' must be set to identify how many bytes of tag space
-      are available per hardware sector.  For DIF this is either 2 or
-      0 depending on the value of the Control Mode Page ATO bit.
 
 ----------------------------------------------------------------------
 

block/blk-integrity.c

···
 }
 EXPORT_SYMBOL(blk_rq_map_integrity_sg);
 
-/**
- * blk_integrity_compare - Compare integrity profile of two disks
- * @gd1:	Disk to compare
- * @gd2:	Disk to compare
- *
- * Description: Meta-devices like DM and MD need to verify that all
- * sub-devices use the same integrity format before advertising to
- * upper layers that they can send/receive integrity metadata.  This
- * function can be used to check whether two gendisk devices have
- * compatible integrity formats.
- */
-int blk_integrity_compare(struct gendisk *gd1, struct gendisk *gd2)
-{
-	struct blk_integrity *b1 = &gd1->queue->integrity;
-	struct blk_integrity *b2 = &gd2->queue->integrity;
-
-	if (!b1->tuple_size && !b2->tuple_size)
-		return 0;
-
-	if (!b1->tuple_size || !b2->tuple_size)
-		return -1;
-
-	if (b1->interval_exp != b2->interval_exp) {
-		pr_err("%s: %s/%s protection interval %u != %u\n",
-		       __func__, gd1->disk_name, gd2->disk_name,
-		       1 << b1->interval_exp, 1 << b2->interval_exp);
-		return -1;
-	}
-
-	if (b1->tuple_size != b2->tuple_size) {
-		pr_err("%s: %s/%s tuple sz %u != %u\n", __func__,
-		       gd1->disk_name, gd2->disk_name,
-		       b1->tuple_size, b2->tuple_size);
-		return -1;
-	}
-
-	if (b1->tag_size && b2->tag_size && (b1->tag_size != b2->tag_size)) {
-		pr_err("%s: %s/%s tag sz %u != %u\n", __func__,
-		       gd1->disk_name, gd2->disk_name,
-		       b1->tag_size, b2->tag_size);
-		return -1;
-	}
-
-	if (b1->csum_type != b2->csum_type ||
-	    (b1->flags & BLK_INTEGRITY_REF_TAG) !=
-	    (b2->flags & BLK_INTEGRITY_REF_TAG)) {
-		pr_err("%s: %s/%s type %s != %s\n", __func__,
-		       gd1->disk_name, gd2->disk_name,
-		       blk_integrity_profile_name(b1),
-		       blk_integrity_profile_name(b2));
-		return -1;
-	}
-
-	return 0;
-}
-EXPORT_SYMBOL(blk_integrity_compare);
-
 bool blk_integrity_merge_rq(struct request_queue *q, struct request *req,
 			    struct request *next)
 {
···
 
 static inline struct blk_integrity *dev_to_bi(struct device *dev)
 {
-	return &dev_to_disk(dev)->queue->integrity;
+	return &dev_to_disk(dev)->queue->limits.integrity;
 }
 
 const char *blk_integrity_profile_name(struct blk_integrity *bi)
···
 static ssize_t flag_store(struct device *dev, struct device_attribute *attr,
 		const char *page, size_t count, unsigned char flag)
 {
-	struct blk_integrity *bi = dev_to_bi(dev);
+	struct request_queue *q = dev_to_disk(dev)->queue;
+	struct queue_limits lim;
 	unsigned long val;
 	int err;
 
···
 	if (err)
 		return err;
 
-	/* the flags are inverted vs the values in the sysfs files */
+	/* note that the flags are inverted vs the values in the sysfs files */
+	lim = queue_limits_start_update(q);
 	if (val)
-		bi->flags &= ~flag;
+		lim.integrity.flags &= ~flag;
 	else
-		bi->flags |= flag;
+		lim.integrity.flags |= flag;
+
+	blk_mq_freeze_queue(q);
+	err = queue_limits_commit_update(q, &lim);
+	blk_mq_unfreeze_queue(q);
+	if (err)
+		return err;
 	return count;
 }
 
···
 	.name = "integrity",
 	.attrs = integrity_attrs,
 };
-
-/**
- * blk_integrity_register - Register a gendisk as being integrity-capable
- * @disk:	struct gendisk pointer to make integrity-aware
- * @template:	block integrity profile to register
- *
- * Description: When a device needs to advertise itself as being able to
- * send/receive integrity metadata it must use this function to register
- * the capability with the block layer. The template is a blk_integrity
- * struct with values appropriate for the underlying hardware. See
- * Documentation/block/data-integrity.rst.
- */
-void blk_integrity_register(struct gendisk *disk, struct blk_integrity *template)
-{
-	struct blk_integrity *bi = &disk->queue->integrity;
-
-	bi->csum_type = template->csum_type;
-	bi->flags = template->flags;
-	bi->interval_exp = template->interval_exp ? :
-		ilog2(queue_logical_block_size(disk->queue));
-	bi->tuple_size = template->tuple_size;
-	bi->tag_size = template->tag_size;
-	bi->pi_offset = template->pi_offset;
-
-#ifdef CONFIG_BLK_INLINE_ENCRYPTION
-	if (disk->queue->crypto_profile) {
-		pr_warn("blk-integrity: Integrity and hardware inline encryption are not supported together. Disabling hardware inline encryption.\n");
-		disk->queue->crypto_profile = NULL;
-	}
-#endif
-}
-EXPORT_SYMBOL(blk_integrity_register);
-
-/**
- * blk_integrity_unregister - Unregister block integrity profile
- * @disk:	disk whose integrity profile to unregister
- *
- * Description: This function unregisters the integrity capability from
- * a block device.
- */
-void blk_integrity_unregister(struct gendisk *disk)
-{
-	struct blk_integrity *bi = &disk->queue->integrity;
-
-	if (!bi->tuple_size)
-		return;
-	memset(bi, 0, sizeof(*bi));
-}
-EXPORT_SYMBOL(blk_integrity_unregister);

block/blk-settings.c

···
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/bio.h>
-#include <linux/blkdev.h>
+#include <linux/blk-integrity.h>
 #include <linux/pagemap.h>
 #include <linux/backing-dev-defs.h>
 #include <linux/gcd.h>
···
 	return 0;
 }
 
+static int blk_validate_integrity_limits(struct queue_limits *lim)
+{
+	struct blk_integrity *bi = &lim->integrity;
+
+	if (!bi->tuple_size) {
+		if (bi->csum_type != BLK_INTEGRITY_CSUM_NONE ||
+		    bi->tag_size || ((bi->flags & BLK_INTEGRITY_REF_TAG))) {
+			pr_warn("invalid PI settings.\n");
+			return -EINVAL;
+		}
+		return 0;
+	}
+
+	if (!IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY)) {
+		pr_warn("integrity support disabled.\n");
+		return -EINVAL;
+	}
+
+	if (bi->csum_type == BLK_INTEGRITY_CSUM_NONE &&
+	    (bi->flags & BLK_INTEGRITY_REF_TAG)) {
+		pr_warn("ref tag not support without checksum.\n");
+		return -EINVAL;
+	}
+
+	if (!bi->interval_exp)
+		bi->interval_exp = ilog2(lim->logical_block_size);
+
+	return 0;
+}
+
 /*
  * Check that the limits in lim are valid, initialize defaults for unset
  * values, and cap values based on others where needed.
···
 {
 	unsigned int max_hw_sectors;
 	unsigned int logical_block_sectors;
+	int err;
 
 	/*
 	 * Unless otherwise specified, default to 512 byte logical blocks and a
···
 		lim->misaligned = 0;
 	}
 
+	err = blk_validate_integrity_limits(lim);
+	if (err)
+		return err;
 	return blk_validate_zoned_limits(lim);
 }
 
···
 		struct queue_limits *lim)
 	__releases(q->limits_lock)
 {
-	int error = blk_validate_limits(lim);
+	int error;
 
-	if (!error) {
-		q->limits = *lim;
-		if (q->disk)
-			blk_apply_bdi_limits(q->disk->bdi, lim);
+	error = blk_validate_limits(lim);
+	if (error)
+		goto out_unlock;
+
+#ifdef CONFIG_BLK_INLINE_ENCRYPTION
+	if (q->crypto_profile && lim->integrity.tag_size) {
+		pr_warn("blk-integrity: Integrity and hardware inline encryption are not supported together.\n");
+		error = -EINVAL;
+		goto out_unlock;
 	}
+#endif
+
+	q->limits = *lim;
+	if (q->disk)
+		blk_apply_bdi_limits(q->disk->bdi, lim);
+out_unlock:
 	mutex_unlock(&q->limits_lock);
 	return error;
 }
···
 			pfx, bdev);
 }
 EXPORT_SYMBOL_GPL(queue_limits_stack_bdev);
+
+/**
+ * queue_limits_stack_integrity - stack integrity profile
+ * @t: target queue limits
+ * @b: base queue limits
+ *
+ * Check if the integrity profile in the @b can be stacked into the
+ * target @t.  Stacking is possible if either:
+ *
+ *   a) does not have any integrity information stacked into it yet
+ *   b) the integrity profile in @b is identical to the one in @t
+ *
+ * If @b can be stacked into @t, return %true.  Else return %false and clear the
+ * integrity information in @t.
+ */
+bool queue_limits_stack_integrity(struct queue_limits *t,
+		struct queue_limits *b)
+{
+	struct blk_integrity *ti = &t->integrity;
+	struct blk_integrity *bi = &b->integrity;
+
+	if (!IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY))
+		return true;
+
+	if (!ti->tuple_size) {
+		/* inherit the settings from the first underlying device */
+		if (!(ti->flags & BLK_INTEGRITY_STACKED)) {
+			ti->flags = BLK_INTEGRITY_DEVICE_CAPABLE |
+				(bi->flags & BLK_INTEGRITY_REF_TAG);
+			ti->csum_type = bi->csum_type;
+			ti->tuple_size = bi->tuple_size;
+			ti->pi_offset = bi->pi_offset;
+			ti->interval_exp = bi->interval_exp;
+			ti->tag_size = bi->tag_size;
+			goto done;
+		}
+		if (!bi->tuple_size)
+			goto done;
+	}
+
+	if (ti->tuple_size != bi->tuple_size)
+		goto incompatible;
+	if (ti->interval_exp != bi->interval_exp)
+		goto incompatible;
+	if (ti->tag_size != bi->tag_size)
+		goto incompatible;
+	if (ti->csum_type != bi->csum_type)
+		goto incompatible;
+	if ((ti->flags & BLK_INTEGRITY_REF_TAG) !=
+	    (bi->flags & BLK_INTEGRITY_REF_TAG))
+		goto incompatible;
+
+done:
+	ti->flags |= BLK_INTEGRITY_STACKED;
+	return true;
+
+incompatible:
+	memset(ti, 0, sizeof(*ti));
+	return false;
+}
+EXPORT_SYMBOL_GPL(queue_limits_stack_integrity);
 
 /**
  * blk_queue_update_dma_pad - update pad mask

block/t10-pi.c

···
  */
 static void t10_pi_type1_prepare(struct request *rq)
 {
-	struct blk_integrity *bi = &rq->q->integrity;
+	struct blk_integrity *bi = &rq->q->limits.integrity;
 	const int tuple_sz = bi->tuple_size;
 	u32 ref_tag = t10_pi_ref_tag(rq);
 	u8 offset = bi->pi_offset;
···
  */
 static void t10_pi_type1_complete(struct request *rq, unsigned int nr_bytes)
 {
-	struct blk_integrity *bi = &rq->q->integrity;
+	struct blk_integrity *bi = &rq->q->limits.integrity;
 	unsigned intervals = nr_bytes >> bi->interval_exp;
 	const int tuple_sz = bi->tuple_size;
 	u32 ref_tag = t10_pi_ref_tag(rq);
···
 
 static void ext_pi_type1_prepare(struct request *rq)
 {
-	struct blk_integrity *bi = &rq->q->integrity;
+	struct blk_integrity *bi = &rq->q->limits.integrity;
 	const int tuple_sz = bi->tuple_size;
 	u64 ref_tag = ext_pi_ref_tag(rq);
 	u8 offset = bi->pi_offset;
···
 
 static void ext_pi_type1_complete(struct request *rq, unsigned int nr_bytes)
 {
-	struct blk_integrity *bi = &rq->q->integrity;
+	struct blk_integrity *bi = &rq->q->limits.integrity;
 	unsigned intervals = nr_bytes >> bi->interval_exp;
 	const int tuple_sz = bi->tuple_size;
 	u64 ref_tag = ext_pi_ref_tag(rq);
···
 
 void blk_integrity_prepare(struct request *rq)
 {
-	struct blk_integrity *bi = &rq->q->integrity;
+	struct blk_integrity *bi = &rq->q->limits.integrity;
 
 	if (!(bi->flags & BLK_INTEGRITY_REF_TAG))
 		return;
···
 
 void blk_integrity_complete(struct request *rq, unsigned int nr_bytes)
 {
-	struct blk_integrity *bi = &rq->q->integrity;
+	struct blk_integrity *bi = &rq->q->limits.integrity;
 
 	if (!(bi->flags & BLK_INTEGRITY_REF_TAG))
 		return;

drivers/md/dm-core.h

···
 
 	bool integrity_supported:1;
 	bool singleton:1;
-	unsigned integrity_added:1;
 
 	/*
 	 * Indicates the rw permissions for the new logical device.  This

drivers/md/dm-integrity.c

···
 		limits->dma_alignment = limits->logical_block_size - 1;
 		limits->discard_granularity = ic->sectors_per_block << SECTOR_SHIFT;
 	}
+
+	if (!ic->internal_hash) {
+		struct blk_integrity *bi = &limits->integrity;
+
+		memset(bi, 0, sizeof(*bi));
+		bi->tuple_size = ic->tag_size;
+		bi->tag_size = bi->tuple_size;
+		bi->interval_exp =
+			ic->sb->log2_sectors_per_block + SECTOR_SHIFT;
+	}
+
 	limits->max_integrity_segments = USHRT_MAX;
 }
 
···
 	sb_set_version(ic);
 
 	return 0;
-}
-
-static void dm_integrity_set(struct dm_target *ti, struct dm_integrity_c *ic)
-{
-	struct gendisk *disk = dm_disk(dm_table_get_md(ti->table));
-	struct blk_integrity bi;
-
-	memset(&bi, 0, sizeof(bi));
-	bi.tuple_size = ic->tag_size;
-	bi.tag_size = bi.tuple_size;
-	bi.interval_exp = ic->sb->log2_sectors_per_block + SECTOR_SHIFT;
-
-	blk_integrity_register(disk, &bi);
 }
 
 static void dm_integrity_free_page_list(struct page_list *pl)
···
 				goto bad;
 		}
 	}
-
-	if (!ic->internal_hash)
-		dm_integrity_set(ti, ic);
 
 	ti->num_flush_bios = 1;
 	ti->flush_supported = true;

drivers/md/dm-table.c

···
 		       q->limits.logical_block_size,
 		       q->limits.alignment_offset,
 		       (unsigned long long) start << SECTOR_SHIFT);
+
+	/*
+	 * Only stack the integrity profile if the target doesn't have native
+	 * integrity support.
+	 */
+	if (!dm_target_has_integrity(ti->type))
+		queue_limits_stack_integrity_bdev(limits, bdev);
 	return 0;
 }
 
···
 		}
 		t->immutable_target_type = ti->type;
 	}
-
-	if (dm_target_has_integrity(ti->type))
-		t->integrity_added = 1;
 
 	ti->table = t;
 	ti->begin = start;
···
 	return r;
 }
 
-static bool integrity_profile_exists(struct gendisk *disk)
-{
-	return !!blk_get_integrity(disk);
-}
-
-/*
- * Get a disk whose integrity profile reflects the table's profile.
- * Returns NULL if integrity support was inconsistent or unavailable.
- */
-static struct gendisk *dm_table_get_integrity_disk(struct dm_table *t)
-{
-	struct list_head *devices = dm_table_get_devices(t);
-	struct dm_dev_internal *dd = NULL;
-	struct gendisk *prev_disk = NULL, *template_disk = NULL;
-
-	for (unsigned int i = 0; i < t->num_targets; i++) {
-		struct dm_target *ti = dm_table_get_target(t, i);
-
-		if (!dm_target_passes_integrity(ti->type))
-			goto no_integrity;
-	}
-
-	list_for_each_entry(dd, devices, list) {
-		template_disk = dd->dm_dev->bdev->bd_disk;
-		if (!integrity_profile_exists(template_disk))
-			goto no_integrity;
-		else if (prev_disk &&
-			 blk_integrity_compare(prev_disk, template_disk) < 0)
-			goto no_integrity;
-		prev_disk = template_disk;
-	}
-
-	return template_disk;
-
-no_integrity:
-	if (prev_disk)
-		DMWARN("%s: integrity not set: %s and %s profile mismatch",
-		       dm_device_name(t->md),
-		       prev_disk->disk_name,
-		       template_disk->disk_name);
-	return NULL;
-}
-
-/*
- * Register the mapped device for blk_integrity support if the
- * underlying devices have an integrity profile.  But all devices may
- * not have matching profiles (checking all devices isn't reliable
- * during table load because this table may use other DM device(s) which
- * must be resumed before they will have an initialized integity
- * profile).  Consequently, stacked DM devices force a 2 stage integrity
- * profile validation: First pass during table load, final pass during
- * resume.
- */
-static int dm_table_register_integrity(struct dm_table *t)
-{
-	struct mapped_device *md = t->md;
-	struct gendisk *template_disk = NULL;
-
-	/* If target handles integrity itself do not register it here. */
-	if (t->integrity_added)
-		return 0;
-
-	template_disk = dm_table_get_integrity_disk(t);
-	if (!template_disk)
-		return 0;
-
-	if (!integrity_profile_exists(dm_disk(md))) {
-		t->integrity_supported = true;
-		/*
-		 * Register integrity profile during table load; we can do
-		 * this because the final profile must match during resume.
-		 */
-		blk_integrity_register(dm_disk(md),
-				       blk_get_integrity(template_disk));
-		return 0;
-	}
-
-	/*
-	 * If DM device already has an initialized integrity
-	 * profile the new profile should not conflict.
-	 */
-	if (blk_integrity_compare(dm_disk(md), template_disk) < 0) {
-		DMERR("%s: conflict with existing integrity profile: %s profile mismatch",
-		      dm_device_name(t->md),
-		      template_disk->disk_name);
-		return 1;
-	}
-
-	/* Preserve existing integrity profile */
-	t->integrity_supported = true;
-	return 0;
-}
-
 #ifdef CONFIG_BLK_INLINE_ENCRYPTION
 
 struct dm_crypto_profile {
···
 	r = dm_table_build_index(t);
 	if (r) {
 		DMERR("unable to build btrees");
-		return r;
-	}
-
-	r = dm_table_register_integrity(t);
-	if (r) {
-		DMERR("could not register integrity profile.");
 		return r;
 	}
 
···
 
 	blk_set_stacking_limits(limits);
 
+	t->integrity_supported = true;
+	for (unsigned int i = 0; i < t->num_targets; i++) {
+		struct dm_target *ti = dm_table_get_target(t, i);
+
+		if (!dm_target_passes_integrity(ti->type))
+			t->integrity_supported = false;
+	}
+
 	for (unsigned int i = 0; i < t->num_targets; i++) {
 		struct dm_target *ti = dm_table_get_target(t, i);
 
···
 			       dm_device_name(t->md),
 			       (unsigned long long) ti->begin,
 			       (unsigned long long) ti->len);
+
+		if (t->integrity_supported ||
+		    dm_target_has_integrity(ti->type)) {
+			if (!queue_limits_stack_integrity(limits, &ti_limits)) {
+				DMWARN("%s: adding target device (start sect %llu len %llu) "
+				       "disabled integrity support due to incompatibility",
+				       dm_device_name(t->md),
+				       (unsigned long long) ti->begin,
+				       (unsigned long long) ti->len);
+				t->integrity_supported = false;
+			}
+		}
 	}
 
 	/*
···
 		return -EINVAL;
 
 	return validate_hardware_logical_block_alignment(t, limits);
-}
-
-/*
- * Verify that all devices have an integrity profile that matches the
- * DM device's registered integrity profile.  If the profiles don't
- * match then unregister the DM device's integrity profile.
- */
-static void dm_table_verify_integrity(struct dm_table *t)
-{
-	struct gendisk *template_disk = NULL;
-
-	if (t->integrity_added)
-		return;
-
-	if (t->integrity_supported) {
-		/*
-		 * Verify that the original integrity profile
-		 * matches all the devices in this table.
-		 */
-		template_disk = dm_table_get_integrity_disk(t);
-		if (template_disk &&
-		    blk_integrity_compare(dm_disk(t->md), template_disk) >= 0)
-			return;
-	}
-
-	if (integrity_profile_exists(dm_disk(t->md))) {
-		DMWARN("%s: unable to establish an integrity profile",
-		       dm_device_name(t->md));
-		blk_integrity_unregister(dm_disk(t->md));
-	}
 }
 
 static int device_flush_capable(struct dm_target *ti, struct dm_dev *dev,
···
 		blk_queue_flag_clear(QUEUE_FLAG_NONROT, q);
 	else
 		blk_queue_flag_set(QUEUE_FLAG_NONROT, q);
-
-	dm_table_verify_integrity(t);
 
 	/*
 	 * Some devices don't use blk_integrity but still want stable pages

drivers/md/md.c

···
  */
 int md_integrity_register(struct mddev *mddev)
 {
-	struct md_rdev *rdev, *reference = NULL;
-
 	if (list_empty(&mddev->disks))
 		return 0; /* nothing to do */
-	if (mddev_is_dm(mddev) || blk_get_integrity(mddev->gendisk))
-		return 0; /* shouldn't register, or already is */
-	rdev_for_each(rdev, mddev) {
-		/* skip spares and non-functional disks */
-		if (test_bit(Faulty, &rdev->flags))
-			continue;
-		if (rdev->raid_disk < 0)
-			continue;
-		if (!reference) {
-			/* Use the first rdev as the reference */
-			reference = rdev;
-			continue;
-		}
-		/* does this rdev's profile match the reference profile? */
-		if (blk_integrity_compare(reference->bdev->bd_disk,
-				rdev->bdev->bd_disk) < 0)
-			return -EINVAL;
-	}
-	if (!reference || !bdev_get_integrity(reference->bdev))
-		return 0;
-	/*
-	 * All component devices are integrity capable and have matching
-	 * profiles, register the common profile for the md device.
-	 */
-	blk_integrity_register(mddev->gendisk,
-			       bdev_get_integrity(reference->bdev));
+	if (mddev_is_dm(mddev) || !blk_get_integrity(mddev->gendisk))
+		return 0; /* shouldn't register */
 
 	pr_debug("md: data integrity enabled on %s\n", mdname(mddev));
 	if (bioset_integrity_create(&mddev->bio_set, BIO_POOL_SIZE) ||
···
 	return 0;
 }
 EXPORT_SYMBOL(md_integrity_register);
-
-/*
- * Attempt to add an rdev, but only if it is consistent with the current
- * integrity profile
- */
-int md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev)
-{
-	struct blk_integrity *bi_mddev;
-
-	if (mddev_is_dm(mddev))
-		return 0;
-
-	bi_mddev = blk_get_integrity(mddev->gendisk);
-
-	if (!bi_mddev) /* nothing to do */
-		return 0;
-
-	if (blk_integrity_compare(mddev->gendisk, rdev->bdev->bd_disk) != 0) {
-		pr_err("%s: incompatible integrity profile for %pg\n",
-		       mdname(mddev), rdev->bdev);
-		return -ENXIO;
-	}
-
-	return 0;
-}
-EXPORT_SYMBOL(md_integrity_add_rdev);
 
 static bool rdev_read_only(struct md_rdev *rdev)
 {
···
 int mdp_major = 0;
 
 /* stack the limit for all rdevs into lim */
-void mddev_stack_rdev_limits(struct mddev *mddev, struct queue_limits *lim)
+int mddev_stack_rdev_limits(struct mddev *mddev, struct queue_limits *lim,
+		unsigned int flags)
 {
 	struct md_rdev *rdev;
 
 	rdev_for_each(rdev, mddev) {
 		queue_limits_stack_bdev(lim, rdev->bdev, rdev->data_offset,
 					mddev->gendisk->disk_name);
+		if ((flags & MDDEV_STACK_INTEGRITY) &&
+		    !queue_limits_stack_integrity_bdev(lim, rdev->bdev))
+			return -EINVAL;
 	}
+
+	return 0;
 }
 EXPORT_SYMBOL_GPL(mddev_stack_rdev_limits);
 
···
 	lim = queue_limits_start_update(mddev->gendisk->queue);
 	queue_limits_stack_bdev(&lim, rdev->bdev, rdev->data_offset,
 				mddev->gendisk->disk_name);
+
+	if (!queue_limits_stack_integrity_bdev(&lim, rdev->bdev)) {
+		pr_err("%s: incompatible integrity profile for %pg\n",
+		       mdname(mddev), rdev->bdev);
+		queue_limits_cancel_update(mddev->gendisk->queue);
+		return -ENXIO;
+	}
+
 	return queue_limits_commit_update(mddev->gendisk->queue, &lim);
 }
 EXPORT_SYMBOL_GPL(mddev_stack_new_rdev);

drivers/md/md.h

···
 extern void md_set_array_sectors(struct mddev *mddev, sector_t array_sectors);
 extern int md_check_no_bitmap(struct mddev *mddev);
 extern int md_integrity_register(struct mddev *mddev);
-extern int md_integrity_add_rdev(struct md_rdev *rdev, struct mddev *mddev);
 extern int strict_strtoul_scaled(const char *cp, unsigned long *res, int scale);
 
 extern int mddev_init(struct mddev *mddev);
···
 int md_set_array_info(struct mddev *mddev, struct mdu_array_info_s *info);
 int md_add_new_disk(struct mddev *mddev, struct mdu_disk_info_s *info);
 int do_md_run(struct mddev *mddev);
-void mddev_stack_rdev_limits(struct mddev *mddev, struct queue_limits *lim);
+#define MDDEV_STACK_INTEGRITY	(1u << 0)
+int mddev_stack_rdev_limits(struct mddev *mddev, struct queue_limits *lim,
+		unsigned int flags);
 int mddev_stack_new_rdev(struct mddev *mddev, struct md_rdev *rdev);
 void mddev_update_io_opt(struct mddev *mddev, unsigned int nr_stripes);
 

drivers/md/raid0.c

···
 static int raid0_set_limits(struct mddev *mddev)
 {
 	struct queue_limits lim;
+	int err;
 
 	blk_set_stacking_limits(&lim);
 	lim.max_hw_sectors = mddev->chunk_sectors;
 	lim.max_write_zeroes_sectors = mddev->chunk_sectors;
 	lim.io_min = mddev->chunk_sectors << 9;
 	lim.io_opt = lim.io_min * mddev->raid_disks;
-	mddev_stack_rdev_limits(mddev, &lim);
+	err = mddev_stack_rdev_limits(mddev, &lim, MDDEV_STACK_INTEGRITY);
+	if (err) {
+		queue_limits_cancel_update(mddev->gendisk->queue);
+		return err;
+	}
 	return queue_limits_set(mddev->gendisk->queue, &lim);
 }
 

drivers/md/raid1.c

···
 	if (mddev->recovery_disabled == conf->recovery_disabled)
 		return -EBUSY;
 
-	if (md_integrity_add_rdev(rdev, mddev))
-		return -ENXIO;
-
 	if (rdev->raid_disk >= 0)
 		first = last = rdev->raid_disk;
 
···
 static int raid1_set_limits(struct mddev *mddev)
 {
 	struct queue_limits lim;
+	int err;
 
 	blk_set_stacking_limits(&lim);
 	lim.max_write_zeroes_sectors = 0;
-	mddev_stack_rdev_limits(mddev, &lim);
+	err = mddev_stack_rdev_limits(mddev, &lim, MDDEV_STACK_INTEGRITY);
+	if (err) {
+		queue_limits_cancel_update(mddev->gendisk->queue);
+		return err;
+	}
 	return queue_limits_set(mddev->gendisk->queue, &lim);
 }
 

drivers/md/raid10.c

···
 	if (rdev->saved_raid_disk < 0 && !_enough(conf, 1, -1))
 		return -EINVAL;
 
-	if (md_integrity_add_rdev(rdev, mddev))
-		return -ENXIO;
-
 	if (rdev->raid_disk >= 0)
 		first = last = rdev->raid_disk;
 
···
 {
 	struct r10conf *conf = mddev->private;
 	struct queue_limits lim;
+	int err;
 
 	blk_set_stacking_limits(&lim);
 	lim.max_write_zeroes_sectors = 0;
 	lim.io_min = mddev->chunk_sectors << 9;
 	lim.io_opt = lim.io_min * raid10_nr_stripes(conf);
-	mddev_stack_rdev_limits(mddev, &lim);
+	err = mddev_stack_rdev_limits(mddev, &lim, MDDEV_STACK_INTEGRITY);
+	if (err) {
+		queue_limits_cancel_update(mddev->gendisk->queue);
+		return err;
+	}
 	return queue_limits_set(mddev->gendisk->queue, &lim);
 }
 

drivers/md/raid5.c

···
 	lim.raid_partial_stripes_expensive = 1;
 	lim.discard_granularity = stripe;
 	lim.max_write_zeroes_sectors = 0;
-	mddev_stack_rdev_limits(mddev, &lim);
+	mddev_stack_rdev_limits(mddev, &lim, 0);
 	rdev_for_each(rdev, mddev)
 		queue_limits_stack_bdev(&lim, rdev->bdev, rdev->new_data_offset,
 				mddev->gendisk->disk_name);

drivers/nvdimm/btt.c

···
 	};
 	int rc;
 
+	if (btt_meta_size(btt) && IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY)) {
+		lim.integrity.tuple_size = btt_meta_size(btt);
+		lim.integrity.tag_size = btt_meta_size(btt);
+	}
+
 	btt->btt_disk = blk_alloc_disk(&lim, NUMA_NO_NODE);
 	if (IS_ERR(btt->btt_disk))
 		return PTR_ERR(btt->btt_disk);
···
 
 	blk_queue_flag_set(QUEUE_FLAG_NONROT, btt->btt_disk->queue);
 	blk_queue_flag_set(QUEUE_FLAG_SYNCHRONOUS, btt->btt_disk->queue);
-
-	if (btt_meta_size(btt) && IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY)) {
-		struct blk_integrity bi = {
-			.tuple_size	= btt_meta_size(btt),
-			.tag_size	= btt_meta_size(btt),
-		};
-		blk_integrity_register(btt->btt_disk, &bi);
-	}
 
 	set_capacity(btt->btt_disk, btt->nlba * btt->sector_size >> 9);
 	rc = device_add_disk(&btt->nd_btt->dev, btt->btt_disk, NULL);

drivers/nvme/host/core.c

···
 	return 0;
 }
 
-static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
+static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head,
+		struct queue_limits *lim)
 {
-	struct blk_integrity integrity = { };
+	struct blk_integrity *bi = &lim->integrity;
 
-	blk_integrity_unregister(disk);
+	memset(bi, 0, sizeof(*bi));
 
 	if (!head->ms)
 		return true;
···
 	case NVME_NS_DPS_PI_TYPE3:
 		switch (head->guard_type) {
 		case NVME_NVM_NS_16B_GUARD:
-			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC;
-			integrity.tag_size = sizeof(u16) + sizeof(u32);
-			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
+			bi->csum_type = BLK_INTEGRITY_CSUM_CRC;
+			bi->tag_size = sizeof(u16) + sizeof(u32);
+			bi->flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
 			break;
 		case NVME_NVM_NS_64B_GUARD:
-			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC64;
-			integrity.tag_size = sizeof(u16) + 6;
-			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
+			bi->csum_type = BLK_INTEGRITY_CSUM_CRC64;
+			bi->tag_size = sizeof(u16) + 6;
+			bi->flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
 			break;
 		default:
 			break;
···
 	case NVME_NS_DPS_PI_TYPE2:
 		switch (head->guard_type) {
 		case NVME_NVM_NS_16B_GUARD:
-			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC;
-			integrity.tag_size = sizeof(u16);
-			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE |
-					   BLK_INTEGRITY_REF_TAG;
+			bi->csum_type = BLK_INTEGRITY_CSUM_CRC;
+			bi->tag_size = sizeof(u16);
+			bi->flags |= BLK_INTEGRITY_DEVICE_CAPABLE |
+				     BLK_INTEGRITY_REF_TAG;
 			break;
 		case NVME_NVM_NS_64B_GUARD:
-			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC64;
-			integrity.tag_size = sizeof(u16);
-			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE |
-					   BLK_INTEGRITY_REF_TAG;
+			bi->csum_type = BLK_INTEGRITY_CSUM_CRC64;
+			bi->tag_size = sizeof(u16);
+			bi->flags |= BLK_INTEGRITY_DEVICE_CAPABLE |
+				     BLK_INTEGRITY_REF_TAG;
 			break;
 		default:
 			break;
···
 		break;
 	}
 
-	integrity.tuple_size = head->ms;
-	integrity.pi_offset = head->pi_offset;
-	blk_integrity_register(disk, &integrity);
+	bi->tuple_size = head->ms;
+	bi->pi_offset = head->pi_offset;
 	return true;
 }
 
···
 	if (IS_ENABLED(CONFIG_BLK_DEV_ZONED) &&
 	    ns->head->ids.csi == NVME_CSI_ZNS)
 		nvme_update_zone_info(ns, &lim, &zi);
-	ret = queue_limits_commit_update(ns->disk->queue, &lim);
-	if (ret) {
-		blk_mq_unfreeze_queue(ns->disk->queue);
-		goto out;
-	}
 
 	/*
 	 * Register a metadata profile for PI, or the plain non-integrity NVMe
···
 	 * I/O to namespaces with metadata except when the namespace supports
 	 * PI, as it can strip/insert in that case.
 	 */
-	if (!nvme_init_integrity(ns->disk, ns->head))
+	if (!nvme_init_integrity(ns->disk, ns->head, &lim))
 		capacity = 0;
+
+	ret = queue_limits_commit_update(ns->disk->queue, &lim);
+	if (ret) {
+		blk_mq_unfreeze_queue(ns->disk->queue);
+		goto out;
+	}
 
 	set_capacity_and_notify(ns->disk, capacity);
 
···
 		struct queue_limits lim;
 
 		blk_mq_freeze_queue(ns->head->disk->queue);
-		if (unsupported)
-			ns->head->disk->flags |= GENHD_FL_HIDDEN;
-		else
-			nvme_init_integrity(ns->head->disk, ns->head);
-		set_capacity_and_notify(ns->head->disk, get_capacity(ns->disk));
-		set_disk_ro(ns->head->disk, nvme_ns_is_readonly(ns, info));
-		nvme_mpath_revalidate_paths(ns);
-
 		/*
 		 * queue_limits mixes values that are the hardware limitations
 		 * for bio splitting with what is the device configuration.
···
 		lim.io_opt = ns_lim->io_opt;
 		queue_limits_stack_bdev(&lim, ns->disk->part0, 0,
 					ns->head->disk->disk_name);
+		if (unsupported)
+			ns->head->disk->flags |= GENHD_FL_HIDDEN;
+		else
+			nvme_init_integrity(ns->head->disk, ns->head, &lim);
 		ret = queue_limits_commit_update(ns->head->disk->queue, &lim);
+
+		set_capacity_and_notify(ns->head->disk, get_capacity(ns->disk));
+		set_disk_ro(ns->head->disk, nvme_ns_is_readonly(ns, info));
+		nvme_mpath_revalidate_paths(ns);
+
 		blk_mq_unfreeze_queue(ns->head->disk->queue);
 	}
 

drivers/scsi/sd.c

···
 	return 0;
 }
 
-static void sd_config_protection(struct scsi_disk *sdkp)
+static void sd_config_protection(struct scsi_disk *sdkp,
+		struct queue_limits *lim)
 {
 	struct scsi_device *sdp = sdkp->device;
 
-	sd_dif_config_host(sdkp);
+	if (IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY))
+		sd_dif_config_host(sdkp, lim);
 
 	if (!sdkp->protection_type)
 		return;
···
 		sd_read_app_tag_own(sdkp, buffer);
 		sd_read_write_same(sdkp, buffer);
 		sd_read_security(sdkp, buffer);
-		sd_config_protection(sdkp);
+		sd_config_protection(sdkp, &lim);
 	}
 
 	/*

drivers/scsi/sd.h

···
 	return sector >> (ilog2(sdev->sector_size) - 9);
 }
 
-#ifdef CONFIG_BLK_DEV_INTEGRITY
-
-extern void sd_dif_config_host(struct scsi_disk *);
-
-#else /* CONFIG_BLK_DEV_INTEGRITY */
-
-static inline void sd_dif_config_host(struct scsi_disk *disk)
-{
-}
-
-#endif /* CONFIG_BLK_DEV_INTEGRITY */
+void sd_dif_config_host(struct scsi_disk *sdkp, struct queue_limits *lim);
 
 static inline int sd_is_zoned(struct scsi_disk *sdkp)
 {

drivers/scsi/sd_dif.c

···
 /*
  * Configure exchange of protection information between OS and HBA.
  */
-void sd_dif_config_host(struct scsi_disk *sdkp)
+void sd_dif_config_host(struct scsi_disk *sdkp, struct queue_limits *lim)
 {
 	struct scsi_device *sdp = sdkp->device;
-	struct gendisk *disk = sdkp->disk;
 	u8 type = sdkp->protection_type;
-	struct blk_integrity bi;
+	struct blk_integrity *bi = &lim->integrity;
 	int dif, dix;
+
+	memset(bi, 0, sizeof(*bi));
 
 	dif = scsi_host_dif_capable(sdp->host, type);
 	dix = scsi_host_dix_capable(sdp->host, type);
···
 		dif = 0; dix = 1;
 	}
 
-	if (!dix) {
-		blk_integrity_unregister(disk);
+	if (!dix)
 		return;
-	}
-
-	memset(&bi, 0, sizeof(bi));
 
 	/* Enable DMA of protection information */
 	if (scsi_host_get_guard(sdkp->device->host) & SHOST_DIX_GUARD_IP)
-		bi.csum_type = BLK_INTEGRITY_CSUM_IP;
+		bi->csum_type = BLK_INTEGRITY_CSUM_IP;
 	else
-		bi.csum_type = BLK_INTEGRITY_CSUM_CRC;
+		bi->csum_type = BLK_INTEGRITY_CSUM_CRC;
 
 	if (type != T10_PI_TYPE3_PROTECTION)
-		bi.flags |= BLK_INTEGRITY_REF_TAG;
+		bi->flags |= BLK_INTEGRITY_REF_TAG;
 
-	bi.tuple_size = sizeof(struct t10_pi_tuple);
+	bi->tuple_size = sizeof(struct t10_pi_tuple);
 
 	if (dif && type) {
-		bi.flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
+		bi->flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
 
 		if (!sdkp->ATO)
-			goto out;
+			return;
 
 		if (type == T10_PI_TYPE3_PROTECTION)
-			bi.tag_size = sizeof(u16) + sizeof(u32);
+			bi->tag_size = sizeof(u16) + sizeof(u32);
 		else
-			bi.tag_size = sizeof(u16);
+			bi->tag_size = sizeof(u16);
 	}
 
 	sd_first_printk(KERN_NOTICE, sdkp,
 			"Enabling DIX %s, application tag size %u bytes\n",
-			blk_integrity_profile_name(&bi), bi.tag_size);
-out:
-	blk_integrity_register(disk, &bi);
+			blk_integrity_profile_name(bi), bi->tag_size);
 }
-

include/linux/blk-integrity.h

···
 	BLK_INTEGRITY_NOGENERATE	= 1 << 1,
 	BLK_INTEGRITY_DEVICE_CAPABLE	= 1 << 2,
 	BLK_INTEGRITY_REF_TAG		= 1 << 3,
+	BLK_INTEGRITY_STACKED		= 1 << 4,
 };
 
 struct blk_integrity_iter {
···
 };
 
 const char *blk_integrity_profile_name(struct blk_integrity *bi);
+bool queue_limits_stack_integrity(struct queue_limits *t,
+		struct queue_limits *b);
+static inline bool queue_limits_stack_integrity_bdev(struct queue_limits *t,
+		struct block_device *bdev)
+{
+	return queue_limits_stack_integrity(t, &bdev->bd_disk->queue->limits);
+}
 
 #ifdef CONFIG_BLK_DEV_INTEGRITY
-void blk_integrity_register(struct gendisk *, struct blk_integrity *);
-void blk_integrity_unregister(struct gendisk *);
-int blk_integrity_compare(struct gendisk *, struct gendisk *);
 int blk_rq_map_integrity_sg(struct request_queue *, struct bio *,
 				   struct scatterlist *);
 int blk_rq_count_integrity_sg(struct request_queue *, struct bio *);
···
 static inline bool
 blk_integrity_queue_supports_integrity(struct request_queue *q)
 {
-	return q->integrity.tuple_size;
+	return q->limits.integrity.tuple_size;
 }
 
 static inline struct blk_integrity *blk_get_integrity(struct gendisk *disk)
 {
 	if (!blk_integrity_queue_supports_integrity(disk->queue))
 		return NULL;
-	return &disk->queue->integrity;
+	return &disk->queue->limits.integrity;
 }
 
 static inline struct blk_integrity *
···
 {
 	return false;
 }
-static inline int blk_integrity_compare(struct gendisk *a, struct gendisk *b)
-{
-	return 0;
-}
-static inline void blk_integrity_register(struct gendisk *d,
-					 struct blk_integrity *b)
-{
-}
-static inline void blk_integrity_unregister(struct gendisk *d)
-{
-}
 static inline unsigned short
 queue_max_integrity_segments(const struct request_queue *q)
 {
···
 	return NULL;
 }
 #endif /* CONFIG_BLK_DEV_INTEGRITY */
+
 #endif /* _LINUX_BLK_INTEGRITY_H */

include/linux/blkdev.h

···
 	 * due to possible offsets.
 	 */
 	unsigned int		dma_alignment;
+
+	struct blk_integrity	integrity;
 };
 
 typedef int (*report_zones_cb)(struct blk_zone *zone, unsigned int idx,
···
 	struct kobject *mq_kobj;
 
 	struct queue_limits	limits;
-
-#ifdef  CONFIG_BLK_DEV_INTEGRITY
-	struct blk_integrity integrity;
-#endif	/* CONFIG_BLK_DEV_INTEGRITY */
 
 #ifdef CONFIG_PM
 	struct device		*dev;
···
 {
 	struct request_queue *q = bdev_get_queue(bdev);
 
-#ifdef CONFIG_BLK_DEV_INTEGRITY
-	/* BLK_INTEGRITY_CSUM_NONE is not available in blkdev.h */
-	if (q->integrity.csum_type != 0)
+	if (IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY) &&
+	    q->limits.integrity.csum_type != BLK_INTEGRITY_CSUM_NONE)
 		return true;
-#endif
 	return test_bit(QUEUE_FLAG_STABLE_WRITES, &q->queue_flags);
 }
 

include/linux/t10-pi.h

···
 
 static inline u32 t10_pi_ref_tag(struct request *rq)
 {
-	unsigned int shift = ilog2(queue_logical_block_size(rq->q));
+	unsigned int shift = rq->q->limits.integrity.interval_exp;
 
-#ifdef CONFIG_BLK_DEV_INTEGRITY
-	if (rq->q->integrity.interval_exp)
-		shift = rq->q->integrity.interval_exp;
-#endif
 	return blk_rq_pos(rq) >> (shift - SECTOR_SHIFT) & 0xffffffff;
 }
 
···
 
 static inline u64 ext_pi_ref_tag(struct request *rq)
 {
-	unsigned int shift = ilog2(queue_logical_block_size(rq->q));
+	unsigned int shift = rq->q->limits.integrity.interval_exp;
 
-#ifdef CONFIG_BLK_DEV_INTEGRITY
-	if (rq->q->integrity.interval_exp)
-		shift = rq->q->integrity.interval_exp;
-#endif
 	return lower_48_bits(blk_rq_pos(rq) >> (shift - SECTOR_SHIFT));
 }