move integrity settings to queue_limits v2

move integrity settings to queue_limits v2

[Christoph Hellwig]

Hi Jens, hi Martin,

this series converts the blk-integrity settings to sit in the queue
limits and be updated through the atomic queue limits API.

I've mostly tested this with nvme, scsi is only covered by simple
scsi_debug based tests.

For MD I found an pre-existing error handling bug when combining PI
capable devices with not PI capable devices.  The fix was posted here
(and is included in the git branch below):

   https://lore.kernel.org/linux-raid/20240604172607.3185916-1-hch@lst.de/

For dm-integrity my testing showed that even the baseline fails to create
the luks-based dm-crypto with dm-integrity backing for the authentication
data.  As the failure is non-fatal I've not addressed it here.

Note that the support for native metadata in dm-crypt by Mikulas will
need a rebase on top of this, but as it already requires another
block layer patch and the changes in this series will simplify it a bit
I hope that is ok.

The series is based on top of my previously sent "convert the SCSI ULDs
to the atomic queue limits API v2" API.

A git tree is available here:

   git://git.infradead.org/users/hch/block.git block-integrity-limits

Gitweb:

   http://git.infradead.org/?p=users/hch/block.git;a=shortlog;h=refs/heads/block-integrity-limits

Changes since v1:
 - keep generating (empty) non-PI metadata
 - use a packed enum for the csum type
 - remove an unused flag left in the code

Diffstat:
 Documentation/block/data-integrity.rst |   49 ------
 block/Kconfig                          |    8 -
 block/Makefile                         |    3 
 block/bio-integrity.c                  |   33 +---
 block/blk-integrity.c                  |  229 ++++++++----------------------
 block/blk-mq.c                         |   13 -
 block/blk-settings.c                   |  118 ++++++++++++++-
 block/blk.h                            |    8 +
 block/t10-pi.c                         |  249 +++++++++++----------------------
 drivers/md/dm-core.h                   |    1 
 drivers/md/dm-crypt.c                  |    4 
 drivers/md/dm-integrity.c              |   47 +-----
 drivers/md/dm-table.c                  |  161 +++------------------
 drivers/md/md.c                        |   72 ++-------
 drivers/md/md.h                        |    5 
 drivers/md/raid0.c                     |    7 
 drivers/md/raid1.c                     |   10 -
 drivers/md/raid10.c                    |   10 -
 drivers/md/raid5.c                     |    2 
 drivers/nvdimm/btt.c                   |   13 -
 drivers/nvme/host/Kconfig              |    1 
 drivers/nvme/host/core.c               |   71 ++++-----
 drivers/nvme/host/multipath.c          |    3 
 drivers/nvme/target/Kconfig            |    1 
 drivers/nvme/target/io-cmd-bdev.c      |   16 +-
 drivers/scsi/Kconfig                   |    1 
 drivers/scsi/sd.c                      |   28 +--
 drivers/scsi/sd.h                      |   12 -
 drivers/scsi/sd_dif.c                  |   45 ++---
 drivers/target/target_core_iblock.c    |   49 +++---
 include/linux/bio.h                    |    7 
 include/linux/blk-integrity.h          |   62 ++------
 include/linux/blkdev.h                 |   26 ++-
 include/linux/t10-pi.h                 |   20 --
 34 files changed, 523 insertions(+), 861 deletions(-)

[PATCH 01/11] dm-integrity: use the nop integrity profile

[Christoph Hellwig]

Use the block layer built-in nop profile instead of duplicating it.

Tested by:

$ dd if=/dev/urandom of=key.bin bs=512 count=1

$ cryptsetup luksFormat -q --type luks2 --integrity hmac-sha256 \
 	--integrity-no-wipe /dev/nvme0n1 key.bin
$ cryptsetup luksOpen /dev/nvme0n1 luks-integrity --key-file key.bin

and then doing mkfs.xfs and simple I/O on the mount file system.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Milan Broz <gmazyland@gmail.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
---
 drivers/md/dm-crypt.c     |  4 ++--
 drivers/md/dm-integrity.c | 20 --------------------
 2 files changed, 2 insertions(+), 22 deletions(-)

diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index 1b7a97cc377943..1dfc462f29cd6f 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -1176,8 +1176,8 @@ static int crypt_integrity_ctr(struct crypt_config *cc, struct dm_target *ti)
 	struct blk_integrity *bi = blk_get_integrity(cc->dev->bdev->bd_disk);
 	struct mapped_device *md = dm_table_get_md(ti->table);
 
-	/* From now we require underlying device with our integrity profile */
-	if (!bi || strcasecmp(bi->profile->name, "DM-DIF-EXT-TAG")) {
+	/* We require an underlying device with non-PI metadata */
+	if (!bi || strcmp(bi->profile->name, "nop")) {
 		ti->error = "Integrity profile not supported.";
 		return -EINVAL;
 	}
diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c
index 417fddebe367a2..c1cc27541673c7 100644
--- a/drivers/md/dm-integrity.c
+++ b/drivers/md/dm-integrity.c
@@ -350,25 +350,6 @@ static struct kmem_cache *journal_io_cache;
 #define DEBUG_bytes(bytes, len, msg, ...)	do { } while (0)
 #endif
 
-static void dm_integrity_prepare(struct request *rq)
-{
-}
-
-static void dm_integrity_complete(struct request *rq, unsigned int nr_bytes)
-{
-}
-
-/*
- * DM Integrity profile, protection is performed layer above (dm-crypt)
- */
-static const struct blk_integrity_profile dm_integrity_profile = {
-	.name			= "DM-DIF-EXT-TAG",
-	.generate_fn		= NULL,
-	.verify_fn		= NULL,
-	.prepare_fn		= dm_integrity_prepare,
-	.complete_fn		= dm_integrity_complete,
-};
-
 static void dm_integrity_map_continue(struct dm_integrity_io *dio, bool from_map);
 static void integrity_bio_wait(struct work_struct *w);
 static void dm_integrity_dtr(struct dm_target *ti);
@@ -3656,7 +3637,6 @@ static void dm_integrity_set(struct dm_target *ti, struct dm_integrity_c *ic)
 	struct blk_integrity bi;
 
 	memset(&bi, 0, sizeof(bi));
-	bi.profile = &dm_integrity_profile;
 	bi.tuple_size = ic->tag_size;
 	bi.tag_size = bi.tuple_size;
 	bi.interval_exp = ic->sb->log2_sectors_per_block + SECTOR_SHIFT;
-- 
2.43.0

Re: [PATCH 01/11] dm-integrity: use the nop integrity profile

[Hannes Reinecke]

On 6/7/24 07:58, Christoph Hellwig wrote:
> Use the block layer built-in nop profile instead of duplicating it.
> 
> Tested by:
> 
> $ dd if=/dev/urandom of=key.bin bs=512 count=1
> 
> $ cryptsetup luksFormat -q --type luks2 --integrity hmac-sha256 \
>   	--integrity-no-wipe /dev/nvme0n1 key.bin
> $ cryptsetup luksOpen /dev/nvme0n1 luks-integrity --key-file key.bin
> 
> and then doing mkfs.xfs and simple I/O on the mount file system.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Milan Broz <gmazyland@gmail.com>
> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> ---
>   drivers/md/dm-crypt.c     |  4 ++--
>   drivers/md/dm-integrity.c | 20 --------------------
>   2 files changed, 2 insertions(+), 22 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

[PATCH 02/11] block: remove the unused BIP_{CTRL,DISK}_NOCHECK flags

[Christoph Hellwig]

Both flags are only checked, but never set.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Kanchan Joshi <joshi.k@samsung.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
---
 drivers/scsi/sd.c   | 14 +++-----------
 include/linux/bio.h |  2 --
 2 files changed, 3 insertions(+), 13 deletions(-)

diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index d957e29b17a98a..b477383ccc3b2a 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -806,25 +806,17 @@ static unsigned char sd_setup_protect_cmnd(struct scsi_cmnd *scmd,
 	if (dix) {				/* DIX Type 0, 1, 2, 3 */
 		if (bio_integrity_flagged(bio, BIP_IP_CHECKSUM))
 			scmd->prot_flags |= SCSI_PROT_IP_CHECKSUM;
-
-		if (bio_integrity_flagged(bio, BIP_CTRL_NOCHECK) == false)
-			scmd->prot_flags |= SCSI_PROT_GUARD_CHECK;
+		scmd->prot_flags |= SCSI_PROT_GUARD_CHECK;
 	}
 
 	if (dif != T10_PI_TYPE3_PROTECTION) {	/* DIX/DIF Type 0, 1, 2 */
 		scmd->prot_flags |= SCSI_PROT_REF_INCREMENT;
-
-		if (bio_integrity_flagged(bio, BIP_CTRL_NOCHECK) == false)
-			scmd->prot_flags |= SCSI_PROT_REF_CHECK;
+		scmd->prot_flags |= SCSI_PROT_REF_CHECK;
 	}
 
 	if (dif) {				/* DIX/DIF Type 1, 2, 3 */
 		scmd->prot_flags |= SCSI_PROT_TRANSFER_PI;
-
-		if (bio_integrity_flagged(bio, BIP_DISK_NOCHECK))
-			protect = 3 << 5;	/* Disable target PI checking */
-		else
-			protect = 1 << 5;	/* Enable target PI checking */
+		protect = 1 << 5;	/* Enable target PI checking */
 	}
 
 	scsi_set_prot_op(scmd, prot_op);
diff --git a/include/linux/bio.h b/include/linux/bio.h
index d5379548d684e1..ec5dcf8635ac66 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -324,8 +324,6 @@ static inline void bio_next_folio(struct folio_iter *fi, struct bio *bio)
 enum bip_flags {
 	BIP_BLOCK_INTEGRITY	= 1 << 0, /* block layer owns integrity data */
 	BIP_MAPPED_INTEGRITY	= 1 << 1, /* ref tag has been remapped */
-	BIP_CTRL_NOCHECK	= 1 << 2, /* disable HBA integrity checking */
-	BIP_DISK_NOCHECK	= 1 << 3, /* disable disk integrity checking */
 	BIP_IP_CHECKSUM		= 1 << 4, /* IP checksum */
 	BIP_INTEGRITY_USER	= 1 << 5, /* Integrity payload is user address */
 	BIP_COPY_USER		= 1 << 6, /* Kernel bounce buffer in use */
-- 
2.43.0

Re: [PATCH 02/11] block: remove the unused BIP_{CTRL,DISK}_NOCHECK flags

[Hannes Reinecke]

On 6/7/24 07:58, Christoph Hellwig wrote:
> Both flags are only checked, but never set.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Bart Van Assche <bvanassche@acm.org>
> Reviewed-by: Kanchan Joshi <joshi.k@samsung.com>
> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> ---
>   drivers/scsi/sd.c   | 14 +++-----------
>   include/linux/bio.h |  2 --
>   2 files changed, 3 insertions(+), 13 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 02/11] block: remove the unused BIP_{CTRL,DISK}_NOCHECK flags

[Martin K. Petersen]

Christoph,

> Both flags are only checked, but never set.

Sad to see all the DIX1.1 enablement removed when we're this close to
finally having a userland interface plumbed through. I've been working
on porting our test tooling to work on top of Kanchan's and Anuj's
series. The intent is to add the tests to blktests and fio.

Fundamentally, the biggest problem we had with the original
implementation was that the "integrity profile" was static on a per
controller+device basis. The purpose of 1.1 was to make sure that how to
handle integrity metadata was a per-I/O decision with what to check and
how to do it driven by whichever entity attached the PI. As opposed to
being inferred by controllers and targets (through INQUIRY snooping,
etc.).

We can add the flags back as part of the io_uring series but it does
seem like unnecessary churn to remove things in one release only to add
them back in the next (I'm assuming passthrough will be in 6.12).

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 02/11] block: remove the unused BIP_{CTRL,DISK}_NOCHECK flags

[Christoph Hellwig]

On Mon, Jun 10, 2024 at 07:48:52AM -0400, Martin K. Petersen wrote:
> Fundamentally, the biggest problem we had with the original
> implementation was that the "integrity profile" was static on a per
> controller+device basis. The purpose of 1.1 was to make sure that how to
> handle integrity metadata was a per-I/O decision with what to check and
> how to do it driven by whichever entity attached the PI. As opposed to
> being inferred by controllers and targets (through INQUIRY snooping,
> etc.).
> 
> We can add the flags back as part of the io_uring series but it does
> seem like unnecessary churn to remove things in one release only to add
> them back in the next (I'm assuming passthrough will be in 6.12).

I can just keep the flags in, they aren't really in the way of anything
else here.  That being said, if you want opt-in aren't they the wrong
polarity anyway?

Re: [PATCH 02/11] block: remove the unused BIP_{CTRL,DISK}_NOCHECK flags

[Martin K. Petersen]

Christoph,

> I can just keep the flags in, they aren't really in the way of anything
> else here.  That being said, if you want opt-in aren't they the wrong
> polarity anyway?

I don't particularly like the polarity. It is an artifact of the fact
that unless otherwise noted, checking will be enabled both at HBA and
storage device. So if we reverse the polarity, it would mean that sd.c,
somewhat counter-intuitively, would enable checking on a bio that has no
bip attached. Since checking is enabled by default, regardless of
whether a bip is provided, it seemed more appropriate to opt in to
disabling the checks.

I believe one of my review comments wrt. to the io_uring passthrough
series was that I'd prefer to see the userland flag have the right
polarity, though. Because at that level, explicitly enabling checking
makes more sense.

I don't really mind reversing the BIP flag polarity either. It's mostly
a historical artifact since non-DIX HBAs would snoop INQUIRY and READ
CAPACITY and transparently enable T10 PI on the wire. DIX moved that
decision to sd.c instead of being done by HBA firmware. But we'd still
want checking to be enabled by default even if no integrity was passed
down from the HBA.

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 02/11] block: remove the unused BIP_{CTRL,DISK}_NOCHECK flags

[Christoph Hellwig]

I can just leave them in for now.  Or I can remove them and we add them
back with the right polarity and support in nvme when we add an
actual user.  Either way is pretty simple unlike the weird ip checksum
thing.

[PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Christoph Hellwig]

Remove the BIP_IP_CHECKSUM as sd can just look at the per-disk
checksum type instead.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Kanchan Joshi <joshi.k@samsung.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
---
 block/bio-integrity.c | 3 ---
 drivers/scsi/sd.c     | 6 +++---
 include/linux/bio.h   | 5 ++---
 3 files changed, 5 insertions(+), 9 deletions(-)

diff --git a/block/bio-integrity.c b/block/bio-integrity.c
index af7f71d16114de..c69da65759af89 100644
--- a/block/bio-integrity.c
+++ b/block/bio-integrity.c
@@ -488,9 +488,6 @@ bool bio_integrity_prep(struct bio *bio)
 	bip->bip_flags |= BIP_BLOCK_INTEGRITY;
 	bip_set_seed(bip, bio->bi_iter.bi_sector);
 
-	if (bi->flags & BLK_INTEGRITY_IP_CHECKSUM)
-		bip->bip_flags |= BIP_IP_CHECKSUM;
-
 	/* Map it */
 	offset = offset_in_page(buf);
 	for (i = 0; i < nr_pages && len > 0; i++) {
diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index b477383ccc3b2a..e21b7df5c31b0d 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -43,7 +43,7 @@
 #include <linux/idr.h>
 #include <linux/interrupt.h>
 #include <linux/init.h>
-#include <linux/blkdev.h>
+#include <linux/blk-integrity.h>
 #include <linux/blkpg.h>
 #include <linux/blk-pm.h>
 #include <linux/delay.h>
@@ -799,12 +799,12 @@ static unsigned char sd_setup_protect_cmnd(struct scsi_cmnd *scmd,
 					   unsigned int dix, unsigned int dif)
 {
 	struct request *rq = scsi_cmd_to_rq(scmd);
-	struct bio *bio = rq->bio;
+	struct blk_integrity *bi = &rq->q->integrity;
 	unsigned int prot_op = sd_prot_op(rq_data_dir(rq), dix, dif);
 	unsigned int protect = 0;
 
 	if (dix) {				/* DIX Type 0, 1, 2, 3 */
-		if (bio_integrity_flagged(bio, BIP_IP_CHECKSUM))
+		if (bi->flags & BLK_INTEGRITY_IP_CHECKSUM)
 			scmd->prot_flags |= SCSI_PROT_IP_CHECKSUM;
 		scmd->prot_flags |= SCSI_PROT_GUARD_CHECK;
 	}
diff --git a/include/linux/bio.h b/include/linux/bio.h
index ec5dcf8635ac66..3295dd6021659b 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -324,9 +324,8 @@ static inline void bio_next_folio(struct folio_iter *fi, struct bio *bio)
 enum bip_flags {
 	BIP_BLOCK_INTEGRITY	= 1 << 0, /* block layer owns integrity data */
 	BIP_MAPPED_INTEGRITY	= 1 << 1, /* ref tag has been remapped */
-	BIP_IP_CHECKSUM		= 1 << 4, /* IP checksum */
-	BIP_INTEGRITY_USER	= 1 << 5, /* Integrity payload is user address */
-	BIP_COPY_USER		= 1 << 6, /* Kernel bounce buffer in use */
+	BIP_INTEGRITY_USER	= 1 << 2, /* Integrity payload is user address */
+	BIP_COPY_USER		= 1 << 3, /* Kernel bounce buffer in use */
 };
 
 /*
-- 
2.43.0

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Hannes Reinecke]

On 6/7/24 07:58, Christoph Hellwig wrote:
> Remove the BIP_IP_CHECKSUM as sd can just look at the per-disk
> checksum type instead.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Bart Van Assche <bvanassche@acm.org>
> Reviewed-by: Kanchan Joshi <joshi.k@samsung.com>
> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> ---
>   block/bio-integrity.c | 3 ---
>   drivers/scsi/sd.c     | 6 +++---
>   include/linux/bio.h   | 5 ++---
>   3 files changed, 5 insertions(+), 9 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Martin K. Petersen]

Christoph,

> Remove the BIP_IP_CHECKSUM as sd can just look at the per-disk
> checksum type instead.

This removes the ability to submit an individual I/O using a CRC instead
of the IP checksum. There are cases which can't be expressed when
the controller is operating in IP checksum mode.

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Christoph Hellwig]

On Mon, Jun 10, 2024 at 07:56:00AM -0400, Martin K. Petersen wrote:
> > Remove the BIP_IP_CHECKSUM as sd can just look at the per-disk
> > checksum type instead.
> 
> This removes the ability to submit an individual I/O using a CRC instead
> of the IP checksum. There are cases which can't be expressed when
> the controller is operating in IP checksum mode.

Huh, how?

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Martin K. Petersen]

Christoph,

>> This removes the ability to submit an individual I/O using a CRC
>> instead of the IP checksum. There are cases which can't be expressed
>> when the controller is operating in IP checksum mode.
>
> Huh, how?

On the wire between controller and target there's only CRC. If I want to
write a "bad" CRC to disk, I have switch the controller to CRC mode. The
controller can't convert a "bad" IP checksum to a "bad" CRC. The PI test
tooling relies heavily on being able to write "bad" things to disk and
read them back to validate that we detect the error.

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Christoph Hellwig]

On Mon, Jun 10, 2024 at 08:19:33AM -0400, Martin K. Petersen wrote:
> On the wire between controller and target there's only CRC. If I want to
> write a "bad" CRC to disk, I have switch the controller to CRC mode. The
> controller can't convert a "bad" IP checksum to a "bad" CRC. The PI test
> tooling relies heavily on being able to write "bad" things to disk and
> read them back to validate that we detect the error.

But how do you even toggle the flag?  There is no no code to do that.
And if you already have a special kernel module for that it really
should just use a passthrough request to take care of that.

Note that unlike the NOCHECK flag which I just cleaned up because they
were unused, this one actually does get in the way of the architecture
of the whole series :(  We could add a per-bip csum_type but it would
feel really weird.

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Martin K. Petersen]

Christoph,

Sorry about the delay. Travel got in the way.

>> On the wire between controller and target there's only CRC. If I want to
>> write a "bad" CRC to disk, I have switch the controller to CRC mode. The
>> controller can't convert a "bad" IP checksum to a "bad" CRC. The PI test
>> tooling relies heavily on being able to write "bad" things to disk and
>> read them back to validate that we detect the error.
>
> But how do you even toggle the flag?  There is no no code to do that.
> And if you already have a special kernel module for that it really
> should just use a passthrough request to take care of that.

A passthrough command to the controller?

> Note that unlike the NOCHECK flag which I just cleaned up because they
> were unused, this one actually does get in the way of the architecture
> of the whole series :( We could add a per-bip csum_type but it would
> feel really weird.

Why would it feel weird? That's how it currently works.

The qualification tool issues a flurry of commands injecting errors at
various places in the stack to identify that the right entity (block
layer, controller, storage device) catch a bad checksum, reference tag,
etc. Being able to enable/disable checking at each place in the stack is
important. I also have code for target that does the same thing in the
reverse direction.

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Christoph Hellwig]

On Tue, Jun 11, 2024 at 03:51:27PM -0400, Martin K. Petersen wrote:
> > But how do you even toggle the flag?  There is no no code to do that.
> > And if you already have a special kernel module for that it really
> > should just use a passthrough request to take care of that.
> 
> A passthrough command to the controller?

A passthrough command to the LU that has a mismatching checksum.

> > Note that unlike the NOCHECK flag which I just cleaned up because they
> > were unused, this one actually does get in the way of the architecture
> > of the whole series :( We could add a per-bip csum_type but it would
> > feel really weird.
> 
> Why would it feel weird? That's how it currently works.

Because there's no way to have it set to anything but the per-queue
one.  

> The qualification tool issues a flurry of commands injecting errors at
> various places in the stack to identify that the right entity (block
> layer, controller, storage device) catch a bad checksum, reference tag,
> etc.

How does it do that?  There's no actualy way to make it mismatch.

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Martin K. Petersen]

Christoph,

>> > Note that unlike the NOCHECK flag which I just cleaned up because they
>> > were unused, this one actually does get in the way of the architecture
>> > of the whole series :( We could add a per-bip csum_type but it would
>> > feel really weird.
>> 
>> Why would it feel weird? That's how it currently works.
>
> Because there's no way to have it set to anything but the per-queue
> one.

That's what the io_uring passthrough changes enable.

Note that the IP checksum is an optional performance feature. A SCSI
controller supporting IP-to-CRC conversion does not imply that all
submitted metadata must use IP checksum format.

The T10 CRC used to be painfully slow to calculate prior to processors
growing support for pclmulqdq or similar. Hence the optional IP
checksum. But on a modern CPU, the T10 CRC can often be calculated fast
enough that it is less of a performance impediment.

The interface was explicitly designed so that the entity which creates
the metadata decides which checksum it wants to use. And then it uses
the bip flag to communicate that to the HBA. The patch which allowed the
user to set the desired guard tag format for block layer-owned PI fell
by the wayside, apparently. Possibly lost track because the T10 CRC
hardware offload changes took a while to land.

Note that I would personally love to get rid of the IP checksum
altogether but I think it's too soon to make it obsolete. Still a lot of
SCSI stuff out there which runs in IP checksum mode. And it is still a
bit faster than CRC for many workloads. And as long as it is in use, we
need the ability to support it and qualify it.

All I'm asking is that we retain the ability to disable checking at the
controller level and at the target level. And that the optional IP
checksum can be selected on a per-I/O basis. IOW, please just retain the
three existing bip flags. Happy to look into what polarity-reversal
would look like but I don't think that should hold up your series.

>> The qualification tool issues a flurry of commands injecting errors at
>> various places in the stack to identify that the right entity (block
>> layer, controller, storage device) catch a bad checksum, reference tag,
>> etc.
>
> How does it do that?  There's no actualy way to make it mismatch.

Through a custom passthrough driver that we want to get rid of and
replace with the io_uring interface series.

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Christoph Hellwig]

On Wed, Jun 12, 2024 at 01:27:47PM -0400, Martin K. Petersen wrote:
> >> > Note that unlike the NOCHECK flag which I just cleaned up because they
> >> > were unused, this one actually does get in the way of the architecture
> >> > of the whole series :( We could add a per-bip csum_type but it would
> >> > feel really weird.
> >> 
> >> Why would it feel weird? That's how it currently works.
> >
> > Because there's no way to have it set to anything but the per-queue
> > one.
> 
> That's what the io_uring passthrough changes enable.

The checksum type?  How is that compatible with nvme?

Anyway, I'll just leave this flag in for the resend, but if we can't
come up with a coherent user for it in a merge cycle or two (which
I very much doubt) I'll send another patch to remove it.

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Martin K. Petersen]

Christoph,

> The checksum type?  How is that compatible with nvme?

NVMe does not support IP checksum.

-- 
Martin K. Petersen	Oracle Linux Engineering

Re: [PATCH 03/11] block: remove the BIP_IP_CHECKSUM flag

[Christoph Hellwig]

On Thu, Jun 13, 2024 at 08:57:26PM -0400, Martin K. Petersen wrote:
> 
> Christoph,
> 
> > The checksum type?  How is that compatible with nvme?
> 
> NVMe does not support IP checksum.

Yes, I know.  So how do you make an interface work where you can
arbitrarily force it?

[PATCH 04/11] block: remove the blk_integrity_profile structure

[Christoph Hellwig]

Block layer integrity configuration is a bit complex right now, as it
indirects through operation vectors for a simple two-dimensional
configuration:

 a) the checksum type of none, ip checksum, crc, crc64
 b) the presence or absence of a reference tag

Remove the integrity profile, and instead add a separate csum_type flag
which replaces the existing ip-checksum field and a new flag that
indicates the presence of the reference tag.

This removes up to two layers of indirect calls, remove the need to
offload the no-op verification of non-PI metadata to a workqueue and
generally simplifies the code. The downside is that block/t10-pi.c now
has to be built into the kernel when CONFIG_BLK_DEV_INTEGRITY is
supported.  Given that both nvme and SCSI require t10-pi.ko, it is loaded
for all usual configurations that enabled CONFIG_BLK_DEV_INTEGRITY
already, though.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/Kconfig                       |   8 +-
 block/Makefile                      |   3 +-
 block/bio-integrity.c               |  30 ++--
 block/blk-integrity.c               |  66 ++++----
 block/blk-mq.c                      |  13 +-
 block/blk.h                         |   8 +
 block/t10-pi.c                      | 241 ++++++++++------------------
 drivers/md/dm-crypt.c               |   2 +-
 drivers/nvme/host/Kconfig           |   1 -
 drivers/nvme/host/core.c            |  17 +-
 drivers/nvme/target/Kconfig         |   1 -
 drivers/nvme/target/io-cmd-bdev.c   |  16 +-
 drivers/scsi/Kconfig                |   1 -
 drivers/scsi/sd.c                   |   2 +-
 drivers/scsi/sd_dif.c               |  19 +--
 drivers/target/target_core_iblock.c |  49 +++---
 include/linux/blk-integrity.h       |  35 ++--
 include/linux/blkdev.h              |   9 +-
 include/linux/t10-pi.h              |   8 -
 19 files changed, 215 insertions(+), 314 deletions(-)

diff --git a/block/Kconfig b/block/Kconfig
index dc12af58dbaeca..5b623b876d3b4a 100644
--- a/block/Kconfig
+++ b/block/Kconfig
@@ -62,6 +62,8 @@ config BLK_DEV_BSGLIB
 
 config BLK_DEV_INTEGRITY
 	bool "Block layer data integrity support"
+	select CRC_T10DIF
+	select CRC64_ROCKSOFT
 	help
 	Some storage devices allow extra information to be
 	stored/retrieved to help protect the data.  The block layer
@@ -72,12 +74,6 @@ config BLK_DEV_INTEGRITY
 	T10/SCSI Data Integrity Field or the T13/ATA External Path
 	Protection.  If in doubt, say N.
 
-config BLK_DEV_INTEGRITY_T10
-	tristate
-	depends on BLK_DEV_INTEGRITY
-	select CRC_T10DIF
-	select CRC64_ROCKSOFT
-
 config BLK_DEV_WRITE_MOUNTED
 	bool "Allow writing to mounted block devices"
 	default y
diff --git a/block/Makefile b/block/Makefile
index 168150b9c51025..ddfd21c1a9ffc9 100644
--- a/block/Makefile
+++ b/block/Makefile
@@ -26,8 +26,7 @@ obj-$(CONFIG_MQ_IOSCHED_KYBER)	+= kyber-iosched.o
 bfq-y				:= bfq-iosched.o bfq-wf2q.o bfq-cgroup.o
 obj-$(CONFIG_IOSCHED_BFQ)	+= bfq.o
 
-obj-$(CONFIG_BLK_DEV_INTEGRITY) += bio-integrity.o blk-integrity.o
-obj-$(CONFIG_BLK_DEV_INTEGRITY_T10)	+= t10-pi.o
+obj-$(CONFIG_BLK_DEV_INTEGRITY) += bio-integrity.o blk-integrity.o t10-pi.o
 obj-$(CONFIG_BLK_MQ_PCI)	+= blk-mq-pci.o
 obj-$(CONFIG_BLK_MQ_VIRTIO)	+= blk-mq-virtio.o
 obj-$(CONFIG_BLK_DEV_ZONED)	+= blk-zoned.o
diff --git a/block/bio-integrity.c b/block/bio-integrity.c
index c69da65759af89..e23acf59c12de3 100644
--- a/block/bio-integrity.c
+++ b/block/bio-integrity.c
@@ -378,10 +378,9 @@ EXPORT_SYMBOL_GPL(bio_integrity_map_user);
  * bio_integrity_process - Process integrity metadata for a bio
  * @bio:	bio to generate/verify integrity metadata for
  * @proc_iter:  iterator to process
- * @proc_fn:	Pointer to the relevant processing function
  */
 static blk_status_t bio_integrity_process(struct bio *bio,
-		struct bvec_iter *proc_iter, integrity_processing_fn *proc_fn)
+		struct bvec_iter *proc_iter)
 {
 	struct blk_integrity *bi = blk_get_integrity(bio->bi_bdev->bd_disk);
 	struct blk_integrity_iter iter;
@@ -392,17 +391,18 @@ static blk_status_t bio_integrity_process(struct bio *bio,
 
 	iter.disk_name = bio->bi_bdev->bd_disk->disk_name;
 	iter.interval = 1 << bi->interval_exp;
-	iter.tuple_size = bi->tuple_size;
 	iter.seed = proc_iter->bi_sector;
 	iter.prot_buf = bvec_virt(bip->bip_vec);
-	iter.pi_offset = bi->pi_offset;
 
 	__bio_for_each_segment(bv, bio, bviter, *proc_iter) {
 		void *kaddr = bvec_kmap_local(&bv);
 
 		iter.data_buf = kaddr;
 		iter.data_size = bv.bv_len;
-		ret = proc_fn(&iter);
+		if (bio_data_dir(bio) == WRITE)
+			blk_integrity_generate(&iter, bi);
+		else
+			ret = blk_integrity_verify(&iter, bi);
 		kunmap_local(kaddr);
 
 		if (ret)
@@ -448,12 +448,10 @@ bool bio_integrity_prep(struct bio *bio)
 		return true;
 
 	if (bio_data_dir(bio) == READ) {
-		if (!bi->profile->verify_fn ||
-		    !(bi->flags & BLK_INTEGRITY_VERIFY))
+		if (!(bi->flags & BLK_INTEGRITY_VERIFY))
 			return true;
 	} else {
-		if (!bi->profile->generate_fn ||
-		    !(bi->flags & BLK_INTEGRITY_GENERATE))
+		if (!(bi->flags & BLK_INTEGRITY_GENERATE))
 			return true;
 
 		/*
@@ -508,12 +506,10 @@ bool bio_integrity_prep(struct bio *bio)
 	}
 
 	/* Auto-generate integrity metadata if this is a write */
-	if (bio_data_dir(bio) == WRITE) {
-		bio_integrity_process(bio, &bio->bi_iter,
-				      bi->profile->generate_fn);
-	} else {
+	if (bio_data_dir(bio) == WRITE)
+		bio_integrity_process(bio, &bio->bi_iter);
+	else
 		bip->bio_iter = bio->bi_iter;
-	}
 	return true;
 
 err_end_io:
@@ -536,15 +532,13 @@ static void bio_integrity_verify_fn(struct work_struct *work)
 	struct bio_integrity_payload *bip =
 		container_of(work, struct bio_integrity_payload, bip_work);
 	struct bio *bio = bip->bip_bio;
-	struct blk_integrity *bi = blk_get_integrity(bio->bi_bdev->bd_disk);
 
 	/*
 	 * 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.
 	 */
-	bio->bi_status = bio_integrity_process(bio, &bip->bio_iter,
-						bi->profile->verify_fn);
+	bio->bi_status = bio_integrity_process(bio, &bip->bio_iter);
 	bio_integrity_free(bio);
 	bio_endio(bio);
 }
@@ -566,7 +560,7 @@ bool __bio_integrity_endio(struct bio *bio)
 	struct bio_integrity_payload *bip = bio_integrity(bio);
 
 	if (bio_op(bio) == REQ_OP_READ && !bio->bi_status &&
-	    (bip->bip_flags & BIP_BLOCK_INTEGRITY) && bi->profile->verify_fn) {
+	    (bip->bip_flags & BIP_BLOCK_INTEGRITY) && bi->csum_type) {
 		INIT_WORK(&bip->bip_work, bio_integrity_verify_fn);
 		queue_work(kintegrityd_wq, &bip->bip_work);
 		return false;
diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index ccbeb6dfa87a4d..17d37badfbb8bc 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -123,10 +123,10 @@ 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->profile && !b2->profile)
+	if (!b1->tuple_size && !b2->tuple_size)
 		return 0;
 
-	if (!b1->profile || !b2->profile)
+	if (!b1->tuple_size || !b2->tuple_size)
 		return -1;
 
 	if (b1->interval_exp != b2->interval_exp) {
@@ -150,10 +150,13 @@ int blk_integrity_compare(struct gendisk *gd1, struct gendisk *gd2)
 		return -1;
 	}
 
-	if (b1->profile != b2->profile) {
+	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,
-		       b1->profile->name, b2->profile->name);
+		       blk_integrity_profile_name(b1),
+		       blk_integrity_profile_name(b2));
 		return -1;
 	}
 
@@ -217,14 +220,37 @@ static inline struct blk_integrity *dev_to_bi(struct device *dev)
 	return &dev_to_disk(dev)->queue->integrity;
 }
 
+const char *blk_integrity_profile_name(struct blk_integrity *bi)
+{
+	switch (bi->csum_type) {
+	case BLK_INTEGRITY_CSUM_IP:
+		if (bi->flags & BLK_INTEGRITY_REF_TAG)
+			return "T10-DIF-TYPE1-IP";
+		return "T10-DIF-TYPE3-IP";
+	case BLK_INTEGRITY_CSUM_CRC:
+		if (bi->flags & BLK_INTEGRITY_REF_TAG)
+			return "T10-DIF-TYPE1-CRC";
+		return "T10-DIF-TYPE3-CRC";
+	case BLK_INTEGRITY_CSUM_CRC64:
+		if (bi->flags & BLK_INTEGRITY_REF_TAG)
+			return "EXT-DIF-TYPE1-CRC64";
+		return "EXT-DIF-TYPE3-CRC64";
+	case BLK_INTEGRITY_CSUM_NONE:
+		break;
+	}
+
+	return "nop";
+}
+EXPORT_SYMBOL_GPL(blk_integrity_profile_name);
+
 static ssize_t format_show(struct device *dev, struct device_attribute *attr,
 			   char *page)
 {
 	struct blk_integrity *bi = dev_to_bi(dev);
 
-	if (bi->profile && bi->profile->name)
-		return sysfs_emit(page, "%s\n", bi->profile->name);
-	return sysfs_emit(page, "none\n");
+	if (!bi->tuple_size)
+		return sysfs_emit(page, "none\n");
+	return sysfs_emit(page, "%s\n", blk_integrity_profile_name(bi));
 }
 
 static ssize_t tag_size_show(struct device *dev, struct device_attribute *attr,
@@ -326,28 +352,6 @@ const struct attribute_group blk_integrity_attr_group = {
 	.attrs = integrity_attrs,
 };
 
-static blk_status_t blk_integrity_nop_fn(struct blk_integrity_iter *iter)
-{
-	return BLK_STS_OK;
-}
-
-static void blk_integrity_nop_prepare(struct request *rq)
-{
-}
-
-static void blk_integrity_nop_complete(struct request *rq,
-		unsigned int nr_bytes)
-{
-}
-
-static const struct blk_integrity_profile nop_profile = {
-	.name = "nop",
-	.generate_fn = blk_integrity_nop_fn,
-	.verify_fn = blk_integrity_nop_fn,
-	.prepare_fn = blk_integrity_nop_prepare,
-	.complete_fn = blk_integrity_nop_complete,
-};
-
 /**
  * blk_integrity_register - Register a gendisk as being integrity-capable
  * @disk:	struct gendisk pointer to make integrity-aware
@@ -363,11 +367,11 @@ 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 = BLK_INTEGRITY_VERIFY | BLK_INTEGRITY_GENERATE |
 		template->flags;
 	bi->interval_exp = template->interval_exp ? :
 		ilog2(queue_logical_block_size(disk->queue));
-	bi->profile = template->profile ? template->profile : &nop_profile;
 	bi->tuple_size = template->tuple_size;
 	bi->tag_size = template->tag_size;
 	bi->pi_offset = template->pi_offset;
@@ -394,7 +398,7 @@ void blk_integrity_unregister(struct gendisk *disk)
 {
 	struct blk_integrity *bi = &disk->queue->integrity;
 
-	if (!bi->profile)
+	if (!bi->tuple_size)
 		return;
 
 	/* ensure all bios are off the integrity workqueue */
diff --git a/block/blk-mq.c b/block/blk-mq.c
index 3b4df8e5ac9e5f..0d4cd39c3d25da 100644
--- a/block/blk-mq.c
+++ b/block/blk-mq.c
@@ -804,10 +804,8 @@ static void blk_complete_request(struct request *req)
 	if (!bio)
 		return;
 
-#ifdef CONFIG_BLK_DEV_INTEGRITY
 	if (blk_integrity_rq(req) && req_op(req) == REQ_OP_READ)
-		req->q->integrity.profile->complete_fn(req, total_bytes);
-#endif
+		blk_integrity_complete(req, total_bytes);
 
 	/*
 	 * Upper layers may call blk_crypto_evict_key() anytime after the last
@@ -875,11 +873,9 @@ bool blk_update_request(struct request *req, blk_status_t error,
 	if (!req->bio)
 		return false;
 
-#ifdef CONFIG_BLK_DEV_INTEGRITY
 	if (blk_integrity_rq(req) && req_op(req) == REQ_OP_READ &&
 	    error == BLK_STS_OK)
-		req->q->integrity.profile->complete_fn(req, nr_bytes);
-#endif
+		blk_integrity_complete(req, nr_bytes);
 
 	/*
 	 * Upper layers may call blk_crypto_evict_key() anytime after the last
@@ -1264,10 +1260,9 @@ void blk_mq_start_request(struct request *rq)
 	WRITE_ONCE(rq->state, MQ_RQ_IN_FLIGHT);
 	rq->mq_hctx->tags->rqs[rq->tag] = rq;
 
-#ifdef CONFIG_BLK_DEV_INTEGRITY
 	if (blk_integrity_rq(rq) && req_op(rq) == REQ_OP_WRITE)
-		q->integrity.profile->prepare_fn(rq);
-#endif
+		blk_integrity_prepare(rq);
+
 	if (rq->bio && rq->bio->bi_opf & REQ_POLLED)
 	        WRITE_ONCE(rq->bio->bi_cookie, rq->mq_hctx->queue_num);
 }
diff --git a/block/blk.h b/block/blk.h
index 189bc25beb502a..79e8d5d4fe0caf 100644
--- a/block/blk.h
+++ b/block/blk.h
@@ -9,6 +9,7 @@
 #include <xen/xen.h>
 #include "blk-crypto-internal.h"
 
+struct blk_integrity_iter;
 struct elevator_type;
 
 /* Max future timer expiry for timeouts */
@@ -673,4 +674,11 @@ int bdev_open(struct block_device *bdev, blk_mode_t mode, void *holder,
 	      const struct blk_holder_ops *hops, struct file *bdev_file);
 int bdev_permission(dev_t dev, blk_mode_t mode, void *holder);
 
+void blk_integrity_generate(struct blk_integrity_iter *iter,
+		struct blk_integrity *bi);
+blk_status_t blk_integrity_verify(struct blk_integrity_iter *iter,
+		struct blk_integrity *bi);
+void blk_integrity_prepare(struct request *rq);
+void blk_integrity_complete(struct request *rq, unsigned int nr_bytes);
+
 #endif /* BLK_INTERNAL_H */
diff --git a/block/t10-pi.c b/block/t10-pi.c
index f4cc91156da1f2..dadecf621497bb 100644
--- a/block/t10-pi.c
+++ b/block/t10-pi.c
@@ -11,17 +11,14 @@
 #include <linux/module.h>
 #include <net/checksum.h>
 #include <asm/unaligned.h>
+#include "blk.h"
 
-typedef __be16 (csum_fn) (__be16, void *, unsigned int);
-
-static __be16 t10_pi_crc_fn(__be16 crc, void *data, unsigned int len)
-{
-	return cpu_to_be16(crc_t10dif_update(be16_to_cpu(crc), data, len));
-}
-
-static __be16 t10_pi_ip_fn(__be16 csum, void *data, unsigned int len)
+static __be16 t10_pi_csum(__be16 csum, void *data, unsigned int len,
+		unsigned char csum_type)
 {
-	return (__force __be16)ip_compute_csum(data, len);
+	if (csum_type == BLK_INTEGRITY_CSUM_IP)
+		return (__force __be16)ip_compute_csum(data, len);
+	return cpu_to_be16(crc_t10dif_update(be16_to_cpu(csum), data, len));
 }
 
 /*
@@ -29,48 +26,44 @@ static __be16 t10_pi_ip_fn(__be16 csum, void *data, unsigned int len)
  * 16 bit app tag, 32 bit reference tag. Type 3 does not define the ref
  * tag.
  */
-static blk_status_t t10_pi_generate(struct blk_integrity_iter *iter,
-		csum_fn *fn, enum t10_dif_type type)
+static void t10_pi_generate(struct blk_integrity_iter *iter,
+		struct blk_integrity *bi)
 {
-	u8 offset = iter->pi_offset;
+	u8 offset = bi->pi_offset;
 	unsigned int i;
 
 	for (i = 0 ; i < iter->data_size ; i += iter->interval) {
 		struct t10_pi_tuple *pi = iter->prot_buf + offset;
 
-		pi->guard_tag = fn(0, iter->data_buf, iter->interval);
+		pi->guard_tag = t10_pi_csum(0, iter->data_buf, iter->interval,
+				bi->csum_type);
 		if (offset)
-			pi->guard_tag = fn(pi->guard_tag, iter->prot_buf,
-					   offset);
+			pi->guard_tag = t10_pi_csum(pi->guard_tag,
+					iter->prot_buf, offset, bi->csum_type);
 		pi->app_tag = 0;
 
-		if (type == T10_PI_TYPE1_PROTECTION)
+		if (bi->flags & BLK_INTEGRITY_REF_TAG)
 			pi->ref_tag = cpu_to_be32(lower_32_bits(iter->seed));
 		else
 			pi->ref_tag = 0;
 
 		iter->data_buf += iter->interval;
-		iter->prot_buf += iter->tuple_size;
+		iter->prot_buf += bi->tuple_size;
 		iter->seed++;
 	}
-
-	return BLK_STS_OK;
 }
 
 static blk_status_t t10_pi_verify(struct blk_integrity_iter *iter,
-		csum_fn *fn, enum t10_dif_type type)
+		struct blk_integrity *bi)
 {
-	u8 offset = iter->pi_offset;
+	u8 offset = bi->pi_offset;
 	unsigned int i;
 
-	BUG_ON(type == T10_PI_TYPE0_PROTECTION);
-
 	for (i = 0 ; i < iter->data_size ; i += iter->interval) {
 		struct t10_pi_tuple *pi = iter->prot_buf + offset;
 		__be16 csum;
 
-		if (type == T10_PI_TYPE1_PROTECTION ||
-		    type == T10_PI_TYPE2_PROTECTION) {
+		if (bi->flags & BLK_INTEGRITY_REF_TAG) {
 			if (pi->app_tag == T10_PI_APP_ESCAPE)
 				goto next;
 
@@ -82,15 +75,17 @@ static blk_status_t t10_pi_verify(struct blk_integrity_iter *iter,
 				       iter->seed, be32_to_cpu(pi->ref_tag));
 				return BLK_STS_PROTECTION;
 			}
-		} else if (type == T10_PI_TYPE3_PROTECTION) {
+		} else {
 			if (pi->app_tag == T10_PI_APP_ESCAPE &&
 			    pi->ref_tag == T10_PI_REF_ESCAPE)
 				goto next;
 		}
 
-		csum = fn(0, iter->data_buf, iter->interval);
+		csum = t10_pi_csum(0, iter->data_buf, iter->interval,
+				bi->csum_type);
 		if (offset)
-			csum = fn(csum, iter->prot_buf, offset);
+			csum = t10_pi_csum(csum, iter->prot_buf, offset,
+					bi->csum_type);
 
 		if (pi->guard_tag != csum) {
 			pr_err("%s: guard tag error at sector %llu " \
@@ -102,33 +97,13 @@ static blk_status_t t10_pi_verify(struct blk_integrity_iter *iter,
 
 next:
 		iter->data_buf += iter->interval;
-		iter->prot_buf += iter->tuple_size;
+		iter->prot_buf += bi->tuple_size;
 		iter->seed++;
 	}
 
 	return BLK_STS_OK;
 }
 
-static blk_status_t t10_pi_type1_generate_crc(struct blk_integrity_iter *iter)
-{
-	return t10_pi_generate(iter, t10_pi_crc_fn, T10_PI_TYPE1_PROTECTION);
-}
-
-static blk_status_t t10_pi_type1_generate_ip(struct blk_integrity_iter *iter)
-{
-	return t10_pi_generate(iter, t10_pi_ip_fn, T10_PI_TYPE1_PROTECTION);
-}
-
-static blk_status_t t10_pi_type1_verify_crc(struct blk_integrity_iter *iter)
-{
-	return t10_pi_verify(iter, t10_pi_crc_fn, T10_PI_TYPE1_PROTECTION);
-}
-
-static blk_status_t t10_pi_type1_verify_ip(struct blk_integrity_iter *iter)
-{
-	return t10_pi_verify(iter, t10_pi_ip_fn, T10_PI_TYPE1_PROTECTION);
-}
-
 /**
  * t10_pi_type1_prepare - prepare PI prior submitting request to device
  * @rq:              request with PI that should be prepared
@@ -225,81 +200,15 @@ static void t10_pi_type1_complete(struct request *rq, unsigned int nr_bytes)
 	}
 }
 
-static blk_status_t t10_pi_type3_generate_crc(struct blk_integrity_iter *iter)
-{
-	return t10_pi_generate(iter, t10_pi_crc_fn, T10_PI_TYPE3_PROTECTION);
-}
-
-static blk_status_t t10_pi_type3_generate_ip(struct blk_integrity_iter *iter)
-{
-	return t10_pi_generate(iter, t10_pi_ip_fn, T10_PI_TYPE3_PROTECTION);
-}
-
-static blk_status_t t10_pi_type3_verify_crc(struct blk_integrity_iter *iter)
-{
-	return t10_pi_verify(iter, t10_pi_crc_fn, T10_PI_TYPE3_PROTECTION);
-}
-
-static blk_status_t t10_pi_type3_verify_ip(struct blk_integrity_iter *iter)
-{
-	return t10_pi_verify(iter, t10_pi_ip_fn, T10_PI_TYPE3_PROTECTION);
-}
-
-/* Type 3 does not have a reference tag so no remapping is required. */
-static void t10_pi_type3_prepare(struct request *rq)
-{
-}
-
-/* Type 3 does not have a reference tag so no remapping is required. */
-static void t10_pi_type3_complete(struct request *rq, unsigned int nr_bytes)
-{
-}
-
-const struct blk_integrity_profile t10_pi_type1_crc = {
-	.name			= "T10-DIF-TYPE1-CRC",
-	.generate_fn		= t10_pi_type1_generate_crc,
-	.verify_fn		= t10_pi_type1_verify_crc,
-	.prepare_fn		= t10_pi_type1_prepare,
-	.complete_fn		= t10_pi_type1_complete,
-};
-EXPORT_SYMBOL(t10_pi_type1_crc);
-
-const struct blk_integrity_profile t10_pi_type1_ip = {
-	.name			= "T10-DIF-TYPE1-IP",
-	.generate_fn		= t10_pi_type1_generate_ip,
-	.verify_fn		= t10_pi_type1_verify_ip,
-	.prepare_fn		= t10_pi_type1_prepare,
-	.complete_fn		= t10_pi_type1_complete,
-};
-EXPORT_SYMBOL(t10_pi_type1_ip);
-
-const struct blk_integrity_profile t10_pi_type3_crc = {
-	.name			= "T10-DIF-TYPE3-CRC",
-	.generate_fn		= t10_pi_type3_generate_crc,
-	.verify_fn		= t10_pi_type3_verify_crc,
-	.prepare_fn		= t10_pi_type3_prepare,
-	.complete_fn		= t10_pi_type3_complete,
-};
-EXPORT_SYMBOL(t10_pi_type3_crc);
-
-const struct blk_integrity_profile t10_pi_type3_ip = {
-	.name			= "T10-DIF-TYPE3-IP",
-	.generate_fn		= t10_pi_type3_generate_ip,
-	.verify_fn		= t10_pi_type3_verify_ip,
-	.prepare_fn		= t10_pi_type3_prepare,
-	.complete_fn		= t10_pi_type3_complete,
-};
-EXPORT_SYMBOL(t10_pi_type3_ip);
-
 static __be64 ext_pi_crc64(u64 crc, void *data, unsigned int len)
 {
 	return cpu_to_be64(crc64_rocksoft_update(crc, data, len));
 }
 
-static blk_status_t ext_pi_crc64_generate(struct blk_integrity_iter *iter,
-					enum t10_dif_type type)
+static void ext_pi_crc64_generate(struct blk_integrity_iter *iter,
+		struct blk_integrity *bi)
 {
-	u8 offset = iter->pi_offset;
+	u8 offset = bi->pi_offset;
 	unsigned int i;
 
 	for (i = 0 ; i < iter->data_size ; i += iter->interval) {
@@ -311,17 +220,15 @@ static blk_status_t ext_pi_crc64_generate(struct blk_integrity_iter *iter,
 					iter->prot_buf, offset);
 		pi->app_tag = 0;
 
-		if (type == T10_PI_TYPE1_PROTECTION)
+		if (bi->flags & BLK_INTEGRITY_REF_TAG)
 			put_unaligned_be48(iter->seed, pi->ref_tag);
 		else
 			put_unaligned_be48(0ULL, pi->ref_tag);
 
 		iter->data_buf += iter->interval;
-		iter->prot_buf += iter->tuple_size;
+		iter->prot_buf += bi->tuple_size;
 		iter->seed++;
 	}
-
-	return BLK_STS_OK;
 }
 
 static bool ext_pi_ref_escape(u8 *ref_tag)
@@ -332,9 +239,9 @@ static bool ext_pi_ref_escape(u8 *ref_tag)
 }
 
 static blk_status_t ext_pi_crc64_verify(struct blk_integrity_iter *iter,
-				      enum t10_dif_type type)
+		struct blk_integrity *bi)
 {
-	u8 offset = iter->pi_offset;
+	u8 offset = bi->pi_offset;
 	unsigned int i;
 
 	for (i = 0; i < iter->data_size; i += iter->interval) {
@@ -342,7 +249,7 @@ static blk_status_t ext_pi_crc64_verify(struct blk_integrity_iter *iter,
 		u64 ref, seed;
 		__be64 csum;
 
-		if (type == T10_PI_TYPE1_PROTECTION) {
+		if (bi->flags & BLK_INTEGRITY_REF_TAG) {
 			if (pi->app_tag == T10_PI_APP_ESCAPE)
 				goto next;
 
@@ -353,7 +260,7 @@ static blk_status_t ext_pi_crc64_verify(struct blk_integrity_iter *iter,
 					iter->disk_name, seed, ref);
 				return BLK_STS_PROTECTION;
 			}
-		} else if (type == T10_PI_TYPE3_PROTECTION) {
+		} else {
 			if (pi->app_tag == T10_PI_APP_ESCAPE &&
 			    ext_pi_ref_escape(pi->ref_tag))
 				goto next;
@@ -374,23 +281,13 @@ static blk_status_t ext_pi_crc64_verify(struct blk_integrity_iter *iter,
 
 next:
 		iter->data_buf += iter->interval;
-		iter->prot_buf += iter->tuple_size;
+		iter->prot_buf += bi->tuple_size;
 		iter->seed++;
 	}
 
 	return BLK_STS_OK;
 }
 
-static blk_status_t ext_pi_type1_verify_crc64(struct blk_integrity_iter *iter)
-{
-	return ext_pi_crc64_verify(iter, T10_PI_TYPE1_PROTECTION);
-}
-
-static blk_status_t ext_pi_type1_generate_crc64(struct blk_integrity_iter *iter)
-{
-	return ext_pi_crc64_generate(iter, T10_PI_TYPE1_PROTECTION);
-}
-
 static void ext_pi_type1_prepare(struct request *rq)
 {
 	struct blk_integrity *bi = &rq->q->integrity;
@@ -467,33 +364,61 @@ static void ext_pi_type1_complete(struct request *rq, unsigned int nr_bytes)
 	}
 }
 
-static blk_status_t ext_pi_type3_verify_crc64(struct blk_integrity_iter *iter)
+void blk_integrity_generate(struct blk_integrity_iter *iter,
+		struct blk_integrity *bi)
 {
-	return ext_pi_crc64_verify(iter, T10_PI_TYPE3_PROTECTION);
+	switch (bi->csum_type) {
+	case BLK_INTEGRITY_CSUM_CRC64:
+		ext_pi_crc64_generate(iter, bi);
+		break;
+	case BLK_INTEGRITY_CSUM_CRC:
+	case BLK_INTEGRITY_CSUM_IP:
+		t10_pi_generate(iter, bi);
+		break;
+	default:
+		break;
+	}
 }
 
-static blk_status_t ext_pi_type3_generate_crc64(struct blk_integrity_iter *iter)
+blk_status_t blk_integrity_verify(struct blk_integrity_iter *iter,
+		struct blk_integrity *bi)
 {
-	return ext_pi_crc64_generate(iter, T10_PI_TYPE3_PROTECTION);
+	switch (bi->csum_type) {
+	case BLK_INTEGRITY_CSUM_CRC64:
+		return ext_pi_crc64_verify(iter, bi);
+	case BLK_INTEGRITY_CSUM_CRC:
+	case BLK_INTEGRITY_CSUM_IP:
+		return t10_pi_verify(iter, bi);
+	default:
+		return BLK_STS_OK;
+	}
 }
 
-const struct blk_integrity_profile ext_pi_type1_crc64 = {
-	.name			= "EXT-DIF-TYPE1-CRC64",
-	.generate_fn		= ext_pi_type1_generate_crc64,
-	.verify_fn		= ext_pi_type1_verify_crc64,
-	.prepare_fn		= ext_pi_type1_prepare,
-	.complete_fn		= ext_pi_type1_complete,
-};
-EXPORT_SYMBOL_GPL(ext_pi_type1_crc64);
-
-const struct blk_integrity_profile ext_pi_type3_crc64 = {
-	.name			= "EXT-DIF-TYPE3-CRC64",
-	.generate_fn		= ext_pi_type3_generate_crc64,
-	.verify_fn		= ext_pi_type3_verify_crc64,
-	.prepare_fn		= t10_pi_type3_prepare,
-	.complete_fn		= t10_pi_type3_complete,
-};
-EXPORT_SYMBOL_GPL(ext_pi_type3_crc64);
+void blk_integrity_prepare(struct request *rq)
+{
+	struct blk_integrity *bi = &rq->q->integrity;
+
+	if (!(bi->flags & BLK_INTEGRITY_REF_TAG))
+		return;
+
+	if (bi->csum_type == BLK_INTEGRITY_CSUM_CRC64)
+		ext_pi_type1_prepare(rq);
+	else
+		t10_pi_type1_prepare(rq);
+}
+
+void blk_integrity_complete(struct request *rq, unsigned int nr_bytes)
+{
+	struct blk_integrity *bi = &rq->q->integrity;
+
+	if (!(bi->flags & BLK_INTEGRITY_REF_TAG))
+		return;
+
+	if (bi->csum_type == BLK_INTEGRITY_CSUM_CRC64)
+		ext_pi_type1_complete(rq, nr_bytes);
+	else
+		t10_pi_type1_complete(rq, nr_bytes);
+}
 
 MODULE_DESCRIPTION("T10 Protection Information module");
 MODULE_LICENSE("GPL");
diff --git a/drivers/md/dm-crypt.c b/drivers/md/dm-crypt.c
index 1dfc462f29cd6f..6c013ceb0e5f1d 100644
--- a/drivers/md/dm-crypt.c
+++ b/drivers/md/dm-crypt.c
@@ -1177,7 +1177,7 @@ static int crypt_integrity_ctr(struct crypt_config *cc, struct dm_target *ti)
 	struct mapped_device *md = dm_table_get_md(ti->table);
 
 	/* We require an underlying device with non-PI metadata */
-	if (!bi || strcmp(bi->profile->name, "nop")) {
+	if (!bi || bi->csum_type != BLK_INTEGRITY_CSUM_NONE) {
 		ti->error = "Integrity profile not supported.";
 		return -EINVAL;
 	}
diff --git a/drivers/nvme/host/Kconfig b/drivers/nvme/host/Kconfig
index b309c8be720f47..a3caef75aa0a83 100644
--- a/drivers/nvme/host/Kconfig
+++ b/drivers/nvme/host/Kconfig
@@ -1,7 +1,6 @@
 # SPDX-License-Identifier: GPL-2.0-only
 config NVME_CORE
 	tristate
-	select BLK_DEV_INTEGRITY_T10 if BLK_DEV_INTEGRITY
 
 config BLK_DEV_NVME
 	tristate "NVM Express block device"
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index f5d150c62955d8..14bac248cde4ca 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1744,17 +1744,16 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
 	case NVME_NS_DPS_PI_TYPE3:
 		switch (head->guard_type) {
 		case NVME_NVM_NS_16B_GUARD:
-			integrity.profile = &t10_pi_type3_crc;
+			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC;
 			integrity.tag_size = sizeof(u16) + sizeof(u32);
 			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
 			break;
 		case NVME_NVM_NS_64B_GUARD:
-			integrity.profile = &ext_pi_type3_crc64;
+			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC64;
 			integrity.tag_size = sizeof(u16) + 6;
 			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
 			break;
 		default:
-			integrity.profile = NULL;
 			break;
 		}
 		break;
@@ -1762,22 +1761,22 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
 	case NVME_NS_DPS_PI_TYPE2:
 		switch (head->guard_type) {
 		case NVME_NVM_NS_16B_GUARD:
-			integrity.profile = &t10_pi_type1_crc;
+			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC;
 			integrity.tag_size = sizeof(u16);
-			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
+			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE |
+					   BLK_INTEGRITY_REF_TAG;
 			break;
 		case NVME_NVM_NS_64B_GUARD:
-			integrity.profile = &ext_pi_type1_crc64;
+			integrity.csum_type = BLK_INTEGRITY_CSUM_CRC64;
 			integrity.tag_size = sizeof(u16);
-			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE;
+			integrity.flags |= BLK_INTEGRITY_DEVICE_CAPABLE |
+					   BLK_INTEGRITY_REF_TAG;
 			break;
 		default:
-			integrity.profile = NULL;
 			break;
 		}
 		break;
 	default:
-		integrity.profile = NULL;
 		break;
 	}
 
diff --git a/drivers/nvme/target/Kconfig b/drivers/nvme/target/Kconfig
index 872dd1a0acd804..c42aec41cc7b1f 100644
--- a/drivers/nvme/target/Kconfig
+++ b/drivers/nvme/target/Kconfig
@@ -6,7 +6,6 @@ config NVME_TARGET
 	depends on CONFIGFS_FS
 	select NVME_KEYRING if NVME_TARGET_TCP_TLS
 	select KEYS if NVME_TARGET_TCP_TLS
-	select BLK_DEV_INTEGRITY_T10 if BLK_DEV_INTEGRITY
 	select SGL_ALLOC
 	help
 	  This enabled target side support for the NVMe protocol, that is
diff --git a/drivers/nvme/target/io-cmd-bdev.c b/drivers/nvme/target/io-cmd-bdev.c
index 6426aac2634aeb..b628bc5ee99847 100644
--- a/drivers/nvme/target/io-cmd-bdev.c
+++ b/drivers/nvme/target/io-cmd-bdev.c
@@ -61,15 +61,17 @@ static void nvmet_bdev_ns_enable_integrity(struct nvmet_ns *ns)
 {
 	struct blk_integrity *bi = bdev_get_integrity(ns->bdev);
 
-	if (bi) {
+	if (!bi)
+		return;
+
+	if (bi->csum_type == BLK_INTEGRITY_CSUM_CRC) {
 		ns->metadata_size = bi->tuple_size;
-		if (bi->profile == &t10_pi_type1_crc)
+		if (bi->flags & BLK_INTEGRITY_REF_TAG)
 			ns->pi_type = NVME_NS_DPS_PI_TYPE1;
-		else if (bi->profile == &t10_pi_type3_crc)
-			ns->pi_type = NVME_NS_DPS_PI_TYPE3;
 		else
-			/* Unsupported metadata type */
-			ns->metadata_size = 0;
+			ns->pi_type = NVME_NS_DPS_PI_TYPE3;
+	} else {
+		ns->metadata_size = 0;
 	}
 }
 
@@ -102,7 +104,7 @@ int nvmet_bdev_ns_enable(struct nvmet_ns *ns)
 
 	ns->pi_type = 0;
 	ns->metadata_size = 0;
-	if (IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY_T10))
+	if (IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY))
 		nvmet_bdev_ns_enable_integrity(ns);
 
 	if (bdev_is_zoned(ns->bdev)) {
diff --git a/drivers/scsi/Kconfig b/drivers/scsi/Kconfig
index 065db86d602164..37c24ffea65cc0 100644
--- a/drivers/scsi/Kconfig
+++ b/drivers/scsi/Kconfig
@@ -82,7 +82,6 @@ comment "SCSI support type (disk, tape, CD-ROM)"
 config BLK_DEV_SD
 	tristate "SCSI disk support"
 	depends on SCSI
-	select BLK_DEV_INTEGRITY_T10 if BLK_DEV_INTEGRITY
 	help
 	  If you want to use SCSI hard disks, Fibre Channel disks,
 	  Serial ATA (SATA) or Parallel ATA (PATA) hard disks,
diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index e21b7df5c31b0d..67f24607b862a3 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -804,7 +804,7 @@ static unsigned char sd_setup_protect_cmnd(struct scsi_cmnd *scmd,
 	unsigned int protect = 0;
 
 	if (dix) {				/* DIX Type 0, 1, 2, 3 */
-		if (bi->flags & BLK_INTEGRITY_IP_CHECKSUM)
+		if (bi->csum_type == BLK_INTEGRITY_CSUM_IP)
 			scmd->prot_flags |= SCSI_PROT_IP_CHECKSUM;
 		scmd->prot_flags |= SCSI_PROT_GUARD_CHECK;
 	}
diff --git a/drivers/scsi/sd_dif.c b/drivers/scsi/sd_dif.c
index 1df847b5f74764..6f0921c7db787b 100644
--- a/drivers/scsi/sd_dif.c
+++ b/drivers/scsi/sd_dif.c
@@ -47,18 +47,13 @@ void sd_dif_config_host(struct scsi_disk *sdkp)
 	memset(&bi, 0, sizeof(bi));
 
 	/* Enable DMA of protection information */
-	if (scsi_host_get_guard(sdkp->device->host) & SHOST_DIX_GUARD_IP) {
-		if (type == T10_PI_TYPE3_PROTECTION)
-			bi.profile = &t10_pi_type3_ip;
-		else
-			bi.profile = &t10_pi_type1_ip;
+	if (scsi_host_get_guard(sdkp->device->host) & SHOST_DIX_GUARD_IP)
+		bi.csum_type = BLK_INTEGRITY_CSUM_IP;
+	else
+		bi.csum_type = BLK_INTEGRITY_CSUM_CRC;
 
-		bi.flags |= BLK_INTEGRITY_IP_CHECKSUM;
-	} else
-		if (type == T10_PI_TYPE3_PROTECTION)
-			bi.profile = &t10_pi_type3_crc;
-		else
-			bi.profile = &t10_pi_type1_crc;
+	if (type != T10_PI_TYPE3_PROTECTION)
+		bi.flags |= BLK_INTEGRITY_REF_TAG;
 
 	bi.tuple_size = sizeof(struct t10_pi_tuple);
 
@@ -76,7 +71,7 @@ void sd_dif_config_host(struct scsi_disk *sdkp)
 
 	sd_first_printk(KERN_NOTICE, sdkp,
 			"Enabling DIX %s, application tag size %u bytes\n",
-			bi.profile->name, bi.tag_size);
+			blk_integrity_profile_name(&bi), bi.tag_size);
 out:
 	blk_integrity_register(disk, &bi);
 }
diff --git a/drivers/target/target_core_iblock.c b/drivers/target/target_core_iblock.c
index 7f6ca81778453b..a3e09adc4e767c 100644
--- a/drivers/target/target_core_iblock.c
+++ b/drivers/target/target_core_iblock.c
@@ -148,35 +148,38 @@ static int iblock_configure_device(struct se_device *dev)
 		dev->dev_attrib.is_nonrot = 1;
 
 	bi = bdev_get_integrity(bd);
-	if (bi) {
-		struct bio_set *bs = &ib_dev->ibd_bio_set;
-
-		if (!strcmp(bi->profile->name, "T10-DIF-TYPE3-IP") ||
-		    !strcmp(bi->profile->name, "T10-DIF-TYPE1-IP")) {
-			pr_err("IBLOCK export of blk_integrity: %s not"
-			       " supported\n", bi->profile->name);
-			ret = -ENOSYS;
-			goto out_blkdev_put;
-		}
+	if (!bi)
+		return 0;
 
-		if (!strcmp(bi->profile->name, "T10-DIF-TYPE3-CRC")) {
-			dev->dev_attrib.pi_prot_type = TARGET_DIF_TYPE3_PROT;
-		} else if (!strcmp(bi->profile->name, "T10-DIF-TYPE1-CRC")) {
+	switch (bi->csum_type) {
+	case BLK_INTEGRITY_CSUM_IP:
+		pr_err("IBLOCK export of blk_integrity: %s not supported\n",
+			blk_integrity_profile_name(bi));
+		ret = -ENOSYS;
+		goto out_blkdev_put;
+	case BLK_INTEGRITY_CSUM_CRC:
+		if (bi->flags & BLK_INTEGRITY_REF_TAG)
 			dev->dev_attrib.pi_prot_type = TARGET_DIF_TYPE1_PROT;
-		}
+		else
+			dev->dev_attrib.pi_prot_type = TARGET_DIF_TYPE3_PROT;
+		break;
+	default:
+		break;
+	}
 
-		if (dev->dev_attrib.pi_prot_type) {
-			if (bioset_integrity_create(bs, IBLOCK_BIO_POOL_SIZE) < 0) {
-				pr_err("Unable to allocate bioset for PI\n");
-				ret = -ENOMEM;
-				goto out_blkdev_put;
-			}
-			pr_debug("IBLOCK setup BIP bs->bio_integrity_pool: %p\n",
-				 &bs->bio_integrity_pool);
+	if (dev->dev_attrib.pi_prot_type) {
+		struct bio_set *bs = &ib_dev->ibd_bio_set;
+
+		if (bioset_integrity_create(bs, IBLOCK_BIO_POOL_SIZE) < 0) {
+			pr_err("Unable to allocate bioset for PI\n");
+			ret = -ENOMEM;
+			goto out_blkdev_put;
 		}
-		dev->dev_attrib.hw_pi_prot_type = dev->dev_attrib.pi_prot_type;
+		pr_debug("IBLOCK setup BIP bs->bio_integrity_pool: %p\n",
+			 &bs->bio_integrity_pool);
 	}
 
+	dev->dev_attrib.hw_pi_prot_type = dev->dev_attrib.pi_prot_type;
 	return 0;
 
 out_blkdev_put:
diff --git a/include/linux/blk-integrity.h b/include/linux/blk-integrity.h
index 7428cb43952da0..56ce1ae355805d 100644
--- a/include/linux/blk-integrity.h
+++ b/include/linux/blk-integrity.h
@@ -10,7 +10,7 @@ enum blk_integrity_flags {
 	BLK_INTEGRITY_VERIFY		= 1 << 0,
 	BLK_INTEGRITY_GENERATE		= 1 << 1,
 	BLK_INTEGRITY_DEVICE_CAPABLE	= 1 << 2,
-	BLK_INTEGRITY_IP_CHECKSUM	= 1 << 3,
+	BLK_INTEGRITY_REF_TAG		= 1 << 3,
 };
 
 struct blk_integrity_iter {
@@ -19,22 +19,10 @@ struct blk_integrity_iter {
 	sector_t		seed;
 	unsigned int		data_size;
 	unsigned short		interval;
-	unsigned char		tuple_size;
-	unsigned char		pi_offset;
 	const char		*disk_name;
 };
 
-typedef blk_status_t (integrity_processing_fn) (struct blk_integrity_iter *);
-typedef void (integrity_prepare_fn) (struct request *);
-typedef void (integrity_complete_fn) (struct request *, unsigned int);
-
-struct blk_integrity_profile {
-	integrity_processing_fn		*generate_fn;
-	integrity_processing_fn		*verify_fn;
-	integrity_prepare_fn		*prepare_fn;
-	integrity_complete_fn		*complete_fn;
-	const char			*name;
-};
+const char *blk_integrity_profile_name(struct blk_integrity *bi);
 
 #ifdef CONFIG_BLK_DEV_INTEGRITY
 void blk_integrity_register(struct gendisk *, struct blk_integrity *);
@@ -44,14 +32,17 @@ 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 struct blk_integrity *blk_get_integrity(struct gendisk *disk)
+static inline bool
+blk_integrity_queue_supports_integrity(struct request_queue *q)
 {
-	struct blk_integrity *bi = &disk->queue->integrity;
+	return q->integrity.tuple_size;
+}
 
-	if (!bi->profile)
+static inline struct blk_integrity *blk_get_integrity(struct gendisk *disk)
+{
+	if (!blk_integrity_queue_supports_integrity(disk->queue))
 		return NULL;
-
-	return bi;
+	return &disk->queue->integrity;
 }
 
 static inline struct blk_integrity *
@@ -60,12 +51,6 @@ bdev_get_integrity(struct block_device *bdev)
 	return blk_get_integrity(bdev->bd_disk);
 }
 
-static inline bool
-blk_integrity_queue_supports_integrity(struct request_queue *q)
-{
-	return q->integrity.profile;
-}
-
 static inline unsigned short
 queue_max_integrity_segments(const struct request_queue *q)
 {
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index ac8e0cb2353a0e..bdd33388e1ced8 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -105,9 +105,16 @@ enum {
 struct disk_events;
 struct badblocks;
 
+enum blk_integrity_checksum {
+	BLK_INTEGRITY_CSUM_NONE		= 0,
+	BLK_INTEGRITY_CSUM_IP		= 1,
+	BLK_INTEGRITY_CSUM_CRC		= 2,
+	BLK_INTEGRITY_CSUM_CRC64	= 3,
+} __packed ;
+
 struct blk_integrity {
-	const struct blk_integrity_profile	*profile;
 	unsigned char				flags;
+	enum blk_integrity_checksum		csum_type;
 	unsigned char				tuple_size;
 	unsigned char				pi_offset;
 	unsigned char				interval_exp;
diff --git a/include/linux/t10-pi.h b/include/linux/t10-pi.h
index 248f4ac9564258..d2bafb76badfb9 100644
--- a/include/linux/t10-pi.h
+++ b/include/linux/t10-pi.h
@@ -48,11 +48,6 @@ static inline u32 t10_pi_ref_tag(struct request *rq)
 	return blk_rq_pos(rq) >> (shift - SECTOR_SHIFT) & 0xffffffff;
 }
 
-extern const struct blk_integrity_profile t10_pi_type1_crc;
-extern const struct blk_integrity_profile t10_pi_type1_ip;
-extern const struct blk_integrity_profile t10_pi_type3_crc;
-extern const struct blk_integrity_profile t10_pi_type3_ip;
-
 struct crc64_pi_tuple {
 	__be64 guard_tag;
 	__be16 app_tag;
@@ -79,7 +74,4 @@ static inline u64 ext_pi_ref_tag(struct request *rq)
 	return lower_48_bits(blk_rq_pos(rq) >> (shift - SECTOR_SHIFT));
 }
 
-extern const struct blk_integrity_profile ext_pi_type1_crc64;
-extern const struct blk_integrity_profile ext_pi_type3_crc64;
-
 #endif
-- 
2.43.0

Re: [PATCH 04/11] block: remove the blk_integrity_profile structure

[Hannes Reinecke]

On 6/7/24 07:58, Christoph Hellwig wrote:
> Block layer integrity configuration is a bit complex right now, as it
> indirects through operation vectors for a simple two-dimensional
> configuration:
> 
>   a) the checksum type of none, ip checksum, crc, crc64
>   b) the presence or absence of a reference tag
> 
> Remove the integrity profile, and instead add a separate csum_type flag
> which replaces the existing ip-checksum field and a new flag that
> indicates the presence of the reference tag.
> 
> This removes up to two layers of indirect calls, remove the need to
> offload the no-op verification of non-PI metadata to a workqueue and
> generally simplifies the code. The downside is that block/t10-pi.c now
> has to be built into the kernel when CONFIG_BLK_DEV_INTEGRITY is
> supported.  Given that both nvme and SCSI require t10-pi.ko, it is loaded
> for all usual configurations that enabled CONFIG_BLK_DEV_INTEGRITY
> already, though.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   block/Kconfig                       |   8 +-
>   block/Makefile                      |   3 +-
>   block/bio-integrity.c               |  30 ++--
>   block/blk-integrity.c               |  66 ++++----
>   block/blk-mq.c                      |  13 +-
>   block/blk.h                         |   8 +
>   block/t10-pi.c                      | 241 ++++++++++------------------
>   drivers/md/dm-crypt.c               |   2 +-
>   drivers/nvme/host/Kconfig           |   1 -
>   drivers/nvme/host/core.c            |  17 +-
>   drivers/nvme/target/Kconfig         |   1 -
>   drivers/nvme/target/io-cmd-bdev.c   |  16 +-
>   drivers/scsi/Kconfig                |   1 -
>   drivers/scsi/sd.c                   |   2 +-
>   drivers/scsi/sd_dif.c               |  19 +--
>   drivers/target/target_core_iblock.c |  49 +++---
>   include/linux/blk-integrity.h       |  35 ++--
>   include/linux/blkdev.h              |   9 +-
>   include/linux/t10-pi.h              |   8 -
>   19 files changed, 215 insertions(+), 314 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 04/11] block: remove the blk_integrity_profile structure

[Kanchan Joshi]

On 6/7/2024 11:28 AM, Christoph Hellwig wrote:
> --- a/drivers/md/dm-crypt.c
> +++ b/drivers/md/dm-crypt.c
> @@ -1177,7 +1177,7 @@ static int crypt_integrity_ctr(struct crypt_config *cc, struct dm_target *ti)
>   	struct mapped_device *md = dm_table_get_md(ti->table);
>   
>   	/* We require an underlying device with non-PI metadata */
> -	if (!bi || strcmp(bi->profile->name, "nop")) {
> +	if (!bi || bi->csum_type != BLK_INTEGRITY_CSUM_NONE) {
>   		ti->error = "Integrity profile not supported.";
>   		return -EINVAL;

I'd rename BLK_INTEGRITY_CSUM_NONE to BLK_INTEGRITY_CSUM_NOP. Overall.

Current choice is a bit confusing as it indicates that code is trying to 
handle "none" case while it is actually trying to handle/support "nop" 
profile.

With extended format off:
# nvme format /dev/nvme0n1 -l 5 -m 0 -i 0 -f
Success formatting namespace:1
# cat /sys/block/nvme0n1/integrity/format
nop

With extended format on:
# nvme format /dev/nvme0n1 -l 5 -m 1 -i 0 -f
Success formatting namespace:1
# cat /sys/block/nvme0n1/integrity/format
none

nop is the case when bi->tuple_size is perfectly valid (i.e. not zero), 
and the code needs to have support for it.
none is the case when bi->tuple_size is zero, and the code only needs to 
ensure that it does nothing.

That said, the change can be deferred to a future patch as well.

So, looks good!

Reviewed-by: Kanchan Joshi <joshi.k@samsung.com>

Re: [PATCH 04/11] block: remove the blk_integrity_profile structure

[Christoph Hellwig]

On Sat, Jun 08, 2024 at 12:01:43AM +0530, Kanchan Joshi wrote:
> > +	if (!bi || bi->csum_type != BLK_INTEGRITY_CSUM_NONE) {
> >   		ti->error = "Integrity profile not supported.";
> >   		return -EINVAL;
> 
> I'd rename BLK_INTEGRITY_CSUM_NONE to BLK_INTEGRITY_CSUM_NOP. Overall.

Well, we don't do any checksumming here, and not a no-op checksum.
So in terms of the checksum field I think the name is correct.  But
it is indeed confusing vs the format string, but that is an ABI
we can't change..

Re: [PATCH 04/11] block: remove the blk_integrity_profile structure

[Martin K. Petersen]

Christoph,

> Block layer integrity configuration is a bit complex right now, as it
> indirects through operation vectors for a simple two-dimensional
> configuration:
>
>  a) the checksum type of none, ip checksum, crc, crc64
>  b) the presence or absence of a reference tag
>
> Remove the integrity profile, and instead add a separate csum_type flag
> which replaces the existing ip-checksum field and a new flag that
> indicates the presence of the reference tag.
>
> This removes up to two layers of indirect calls, remove the need to
> offload the no-op verification of non-PI metadata to a workqueue and
> generally simplifies the code. The downside is that block/t10-pi.c now
> has to be built into the kernel when CONFIG_BLK_DEV_INTEGRITY is
> supported.  Given that both nvme and SCSI require t10-pi.ko, it is loaded
> for all usual configurations that enabled CONFIG_BLK_DEV_INTEGRITY
> already, though.

Glad to see the profiles go! Vestiges of a time where it looked like ATA
might have an entirely different format and NVMe did not exist yet.

I'll test when I get back home later today.

-- 
Martin K. Petersen	Oracle Linux Engineering

[PATCH 05/11] block: remove the blk_flush_integrity call in blk_integrity_unregister

[Christoph Hellwig]

Now that there are no indirect calls for PI processing there is no
way to dereference a NULL pointer here.  Additionally drivers now always
freeze the queue (or in case of stacking drivers use their internal
equivalent) around changing the integrity profile.

This is effectively a revert of commit 3df49967f6f1 ("block: flush the
integrity workqueue in blk_integrity_unregister").

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-integrity.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index 17d37badfbb8bc..24f04575096d39 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -401,8 +401,6 @@ void blk_integrity_unregister(struct gendisk *disk)
 	if (!bi->tuple_size)
 		return;
 
-	/* ensure all bios are off the integrity workqueue */
-	blk_flush_integrity();
 	blk_queue_flag_clear(QUEUE_FLAG_STABLE_WRITES, disk->queue);
 	memset(bi, 0, sizeof(*bi));
 }
-- 
2.43.0

Re: [PATCH 05/11] block: remove the blk_flush_integrity call in blk_integrity_unregister

[Hannes Reinecke]

On 6/7/24 07:58, Christoph Hellwig wrote:
> Now that there are no indirect calls for PI processing there is no
> way to dereference a NULL pointer here.  Additionally drivers now always
> freeze the queue (or in case of stacking drivers use their internal
> equivalent) around changing the integrity profile.
> 
> This is effectively a revert of commit 3df49967f6f1 ("block: flush the
> integrity workqueue in blk_integrity_unregister").
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   block/blk-integrity.c | 2 --
>   1 file changed, 2 deletions(-)
> 
> diff --git a/block/blk-integrity.c b/block/blk-integrity.c
> index 17d37badfbb8bc..24f04575096d39 100644
> --- a/block/blk-integrity.c
> +++ b/block/blk-integrity.c
> @@ -401,8 +401,6 @@ void blk_integrity_unregister(struct gendisk *disk)
>   	if (!bi->tuple_size)
>   		return;
>   
> -	/* ensure all bios are off the integrity workqueue */
> -	blk_flush_integrity();
>   	blk_queue_flag_clear(QUEUE_FLAG_STABLE_WRITES, disk->queue);
>   	memset(bi, 0, sizeof(*bi));
>   }

Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 05/11] block: remove the blk_flush_integrity call in blk_integrity_unregister

[Martin K. Petersen]

Christoph,

> Now that there are no indirect calls for PI processing there is no
> way to dereference a NULL pointer here.  Additionally drivers now always
> freeze the queue (or in case of stacking drivers use their internal
> equivalent) around changing the integrity profile.
>
> This is effectively a revert of commit 3df49967f6f1 ("block: flush the
> integrity workqueue in blk_integrity_unregister").

Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>

-- 
Martin K. Petersen	Oracle Linux Engineering

[PATCH 06/11] block: factor out flag_{store,show} helper for integrity

[Christoph Hellwig]

Factor the duplicate code for the generate and verify attributes into
common helpers.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
---
 block/blk-integrity.c | 53 +++++++++++++++++++++----------------------
 1 file changed, 26 insertions(+), 27 deletions(-)

diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index 24f04575096d39..24671d9f90a124 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -243,6 +243,28 @@ const char *blk_integrity_profile_name(struct blk_integrity *bi)
 }
 EXPORT_SYMBOL_GPL(blk_integrity_profile_name);
 
+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);
+	char *p = (char *) page;
+	unsigned long val = simple_strtoul(p, &p, 10);
+
+	if (val)
+		bi->flags |= flag;
+	else
+		bi->flags &= ~flag;
+	return count;
+}
+
+static ssize_t flag_show(struct device *dev, struct device_attribute *attr,
+		char *page, unsigned char flag)
+{
+	struct blk_integrity *bi = dev_to_bi(dev);
+
+	return sysfs_emit(page, "%d\n", !!(bi->flags & flag));
+}
+
 static ssize_t format_show(struct device *dev, struct device_attribute *attr,
 			   char *page)
 {
@@ -275,49 +297,26 @@ static ssize_t read_verify_store(struct device *dev,
 				 struct device_attribute *attr,
 				 const char *page, size_t count)
 {
-	struct blk_integrity *bi = dev_to_bi(dev);
-	char *p = (char *) page;
-	unsigned long val = simple_strtoul(p, &p, 10);
-
-	if (val)
-		bi->flags |= BLK_INTEGRITY_VERIFY;
-	else
-		bi->flags &= ~BLK_INTEGRITY_VERIFY;
-
-	return count;
+	return flag_store(dev, attr, page, count, BLK_INTEGRITY_VERIFY);
 }
 
 static ssize_t read_verify_show(struct device *dev,
 				struct device_attribute *attr, char *page)
 {
-	struct blk_integrity *bi = dev_to_bi(dev);
-
-	return sysfs_emit(page, "%d\n", !!(bi->flags & BLK_INTEGRITY_VERIFY));
+	return flag_show(dev, attr, page, BLK_INTEGRITY_VERIFY);
 }
 
 static ssize_t write_generate_store(struct device *dev,
 				    struct device_attribute *attr,
 				    const char *page, size_t count)
 {
-	struct blk_integrity *bi = dev_to_bi(dev);
-
-	char *p = (char *) page;
-	unsigned long val = simple_strtoul(p, &p, 10);
-
-	if (val)
-		bi->flags |= BLK_INTEGRITY_GENERATE;
-	else
-		bi->flags &= ~BLK_INTEGRITY_GENERATE;
-
-	return count;
+	return flag_store(dev, attr, page, count, BLK_INTEGRITY_GENERATE);
 }
 
 static ssize_t write_generate_show(struct device *dev,
 				   struct device_attribute *attr, char *page)
 {
-	struct blk_integrity *bi = dev_to_bi(dev);
-
-	return sysfs_emit(page, "%d\n", !!(bi->flags & BLK_INTEGRITY_GENERATE));
+	return flag_show(dev, attr, page, BLK_INTEGRITY_GENERATE);
 }
 
 static ssize_t device_is_integrity_capable_show(struct device *dev,
-- 
2.43.0

Re: [PATCH 06/11] block: factor out flag_{store,show} helper for integrity

[Hannes Reinecke]

On 6/7/24 07:59, Christoph Hellwig wrote:
> Factor the duplicate code for the generate and verify attributes into
> common helpers.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Keith Busch <kbusch@kernel.org>
> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> ---
>   block/blk-integrity.c | 53 +++++++++++++++++++++----------------------
>   1 file changed, 26 insertions(+), 27 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 06/11] block: factor out flag_{store,show} helper for integrity

[Martin K. Petersen]

Christoph,

> Factor the duplicate code for the generate and verify attributes into
> common helpers.

Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>

-- 
Martin K. Petersen	Oracle Linux Engineering

[PATCH 07/11] block: use kstrtoul in flag_store

[Christoph Hellwig]

Use the text to integer helper that has error handling and doesn't modify
the input pointer.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
---
 block/blk-integrity.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index 24671d9f90a124..58760a6d6b2209 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -247,8 +247,12 @@ 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);
-	char *p = (char *) page;
-	unsigned long val = simple_strtoul(p, &p, 10);
+	unsigned long val;
+	int err;
+
+	err = kstrtoul(page, 10, &val);
+	if (err)
+		return err;
 
 	if (val)
 		bi->flags |= flag;
-- 
2.43.0

Re: [PATCH 07/11] block: use kstrtoul in flag_store

[Hannes Reinecke]

On 6/7/24 07:59, Christoph Hellwig wrote:
> Use the text to integer helper that has error handling and doesn't modify
> the input pointer.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Keith Busch <kbusch@kernel.org>
> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> ---
>   block/blk-integrity.c | 8 ++++++--
>   1 file changed, 6 insertions(+), 2 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 07/11] block: use kstrtoul in flag_store

[Martin K. Petersen]

Christoph,

> Use the text to integer helper that has error handling and doesn't modify
> the input pointer.

Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>

-- 
Martin K. Petersen	Oracle Linux Engineering

[PATCH 08/11] block: don't require stable pages for non-PI metadata

[Christoph Hellwig]

Non-PI metadata doesn't contain checksums and thus doesn't require
stable pages.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-integrity.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index 58760a6d6b2209..1d2d371cd632d3 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -379,7 +379,8 @@ void blk_integrity_register(struct gendisk *disk, struct blk_integrity *template
 	bi->tag_size = template->tag_size;
 	bi->pi_offset = template->pi_offset;
 
-	blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES, disk->queue);
+	if (bi->csum_type != BLK_INTEGRITY_CSUM_NONE)
+		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES, disk->queue);
 
 #ifdef CONFIG_BLK_INLINE_ENCRYPTION
 	if (disk->queue->crypto_profile) {
@@ -404,7 +405,8 @@ void blk_integrity_unregister(struct gendisk *disk)
 	if (!bi->tuple_size)
 		return;
 
-	blk_queue_flag_clear(QUEUE_FLAG_STABLE_WRITES, disk->queue);
+	if (bi->csum_type != BLK_INTEGRITY_CSUM_NONE)
+		blk_queue_flag_clear(QUEUE_FLAG_STABLE_WRITES, disk->queue);
 	memset(bi, 0, sizeof(*bi));
 }
 EXPORT_SYMBOL(blk_integrity_unregister);
-- 
2.43.0

Re: [PATCH 08/11] block: don't require stable pages for non-PI metadata

[Hannes Reinecke]

On 6/7/24 07:59, Christoph Hellwig wrote:
> Non-PI metadata doesn't contain checksums and thus doesn't require
> stable pages.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   block/blk-integrity.c | 6 ++++--
>   1 file changed, 4 insertions(+), 2 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 08/11] block: don't require stable pages for non-PI metadata

[Martin K. Petersen]

Christoph,

> Non-PI metadata doesn't contain checksums and thus doesn't require
> stable pages.

Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>

-- 
Martin K. Petersen	Oracle Linux Engineering

[PATCH 09/11] block: bypass the STABLE_WRITES flag for protection information

[Christoph Hellwig]

Currently registering a checksum-enabled (aka PI) integrity profile sets
the QUEUE_FLAG_STABLE_WRITE flag, and unregistering it clears the flag.
This can incorrectly clear the flag when the driver requires stable
writes even without PI, e.g. in case of iSCSI or NVMe/TCP with data
digest enabled.

Fix this by looking at the csum_type directly in bdev_stable_writes and
not setting the queue flag.  Also remove the blk_queue_stable_writes
helper as the only user in nvme wants to only look at the actual
QUEUE_FLAG_STABLE_WRITE flag as it inherits the integrity configuration
by other means.

Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 block/blk-integrity.c         |  6 ------
 drivers/nvme/host/multipath.c |  3 ++-
 include/linux/blkdev.h        | 12 ++++++++----
 3 files changed, 10 insertions(+), 11 deletions(-)

diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index 1d2d371cd632d3..bec0d1df387ce9 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -379,9 +379,6 @@ void blk_integrity_register(struct gendisk *disk, struct blk_integrity *template
 	bi->tag_size = template->tag_size;
 	bi->pi_offset = template->pi_offset;
 
-	if (bi->csum_type != BLK_INTEGRITY_CSUM_NONE)
-		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES, disk->queue);
-
 #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");
@@ -404,9 +401,6 @@ void blk_integrity_unregister(struct gendisk *disk)
 
 	if (!bi->tuple_size)
 		return;
-
-	if (bi->csum_type != BLK_INTEGRITY_CSUM_NONE)
-		blk_queue_flag_clear(QUEUE_FLAG_STABLE_WRITES, disk->queue);
 	memset(bi, 0, sizeof(*bi));
 }
 EXPORT_SYMBOL(blk_integrity_unregister);
diff --git a/drivers/nvme/host/multipath.c b/drivers/nvme/host/multipath.c
index d8b6b4648eaff9..12c59db02539e5 100644
--- a/drivers/nvme/host/multipath.c
+++ b/drivers/nvme/host/multipath.c
@@ -875,7 +875,8 @@ void nvme_mpath_add_disk(struct nvme_ns *ns, __le32 anagrpid)
 		nvme_mpath_set_live(ns);
 	}
 
-	if (blk_queue_stable_writes(ns->queue) && ns->head->disk)
+	if (test_bit(QUEUE_FLAG_STABLE_WRITES, &ns->queue->queue_flags) &&
+	    ns->head->disk)
 		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES,
 				   ns->head->disk->queue);
 #ifdef CONFIG_BLK_DEV_ZONED
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index bdd33388e1ced8..f9089750919c6b 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -571,8 +571,6 @@ bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q);
 #define blk_queue_noxmerges(q)	\
 	test_bit(QUEUE_FLAG_NOXMERGES, &(q)->queue_flags)
 #define blk_queue_nonrot(q)	test_bit(QUEUE_FLAG_NONROT, &(q)->queue_flags)
-#define blk_queue_stable_writes(q) \
-	test_bit(QUEUE_FLAG_STABLE_WRITES, &(q)->queue_flags)
 #define blk_queue_io_stat(q)	test_bit(QUEUE_FLAG_IO_STAT, &(q)->queue_flags)
 #define blk_queue_add_random(q)	test_bit(QUEUE_FLAG_ADD_RANDOM, &(q)->queue_flags)
 #define blk_queue_zone_resetall(q)	\
@@ -1300,8 +1298,14 @@ static inline bool bdev_synchronous(struct block_device *bdev)
 
 static inline bool bdev_stable_writes(struct block_device *bdev)
 {
-	return test_bit(QUEUE_FLAG_STABLE_WRITES,
-			&bdev_get_queue(bdev)->queue_flags);
+	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)
+		return true;
+#endif
+	return test_bit(QUEUE_FLAG_STABLE_WRITES, &q->queue_flags);
 }
 
 static inline bool bdev_write_cache(struct block_device *bdev)
-- 
2.43.0

Re: [PATCH 09/11] block: bypass the STABLE_WRITES flag for protection information

[Hannes Reinecke]

On 6/7/24 07:59, Christoph Hellwig wrote:
> Currently registering a checksum-enabled (aka PI) integrity profile sets
> the QUEUE_FLAG_STABLE_WRITE flag, and unregistering it clears the flag.
> This can incorrectly clear the flag when the driver requires stable
> writes even without PI, e.g. in case of iSCSI or NVMe/TCP with data
> digest enabled.
> 
> Fix this by looking at the csum_type directly in bdev_stable_writes and
> not setting the queue flag.  Also remove the blk_queue_stable_writes
> helper as the only user in nvme wants to only look at the actual
> QUEUE_FLAG_STABLE_WRITE flag as it inherits the integrity configuration
> by other means.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> ---
>   block/blk-integrity.c         |  6 ------
>   drivers/nvme/host/multipath.c |  3 ++-
>   include/linux/blkdev.h        | 12 ++++++++----
>   3 files changed, 10 insertions(+), 11 deletions(-)
> 
> diff --git a/block/blk-integrity.c b/block/blk-integrity.c
> index 1d2d371cd632d3..bec0d1df387ce9 100644
> --- a/block/blk-integrity.c
> +++ b/block/blk-integrity.c
> @@ -379,9 +379,6 @@ void blk_integrity_register(struct gendisk *disk, struct blk_integrity *template
>   	bi->tag_size = template->tag_size;
>   	bi->pi_offset = template->pi_offset;
>   
> -	if (bi->csum_type != BLK_INTEGRITY_CSUM_NONE)
> -		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES, disk->queue);
> -
>   #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");
> @@ -404,9 +401,6 @@ void blk_integrity_unregister(struct gendisk *disk)
>   
>   	if (!bi->tuple_size)
>   		return;
> -
> -	if (bi->csum_type != BLK_INTEGRITY_CSUM_NONE)
> -		blk_queue_flag_clear(QUEUE_FLAG_STABLE_WRITES, disk->queue);
>   	memset(bi, 0, sizeof(*bi));
>   }
>   EXPORT_SYMBOL(blk_integrity_unregister);
> diff --git a/drivers/nvme/host/multipath.c b/drivers/nvme/host/multipath.c
> index d8b6b4648eaff9..12c59db02539e5 100644
> --- a/drivers/nvme/host/multipath.c
> +++ b/drivers/nvme/host/multipath.c
> @@ -875,7 +875,8 @@ void nvme_mpath_add_disk(struct nvme_ns *ns, __le32 anagrpid)
>   		nvme_mpath_set_live(ns);
>   	}
>   
> -	if (blk_queue_stable_writes(ns->queue) && ns->head->disk)
> +	if (test_bit(QUEUE_FLAG_STABLE_WRITES, &ns->queue->queue_flags) &&
> +	    ns->head->disk)
>   		blk_queue_flag_set(QUEUE_FLAG_STABLE_WRITES,
>   				   ns->head->disk->queue);
>   #ifdef CONFIG_BLK_DEV_ZONED
> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
> index bdd33388e1ced8..f9089750919c6b 100644
> --- a/include/linux/blkdev.h
> +++ b/include/linux/blkdev.h
> @@ -571,8 +571,6 @@ bool blk_queue_flag_test_and_set(unsigned int flag, struct request_queue *q);
>   #define blk_queue_noxmerges(q)	\
>   	test_bit(QUEUE_FLAG_NOXMERGES, &(q)->queue_flags)
>   #define blk_queue_nonrot(q)	test_bit(QUEUE_FLAG_NONROT, &(q)->queue_flags)
> -#define blk_queue_stable_writes(q) \
> -	test_bit(QUEUE_FLAG_STABLE_WRITES, &(q)->queue_flags)
>   #define blk_queue_io_stat(q)	test_bit(QUEUE_FLAG_IO_STAT, &(q)->queue_flags)
>   #define blk_queue_add_random(q)	test_bit(QUEUE_FLAG_ADD_RANDOM, &(q)->queue_flags)
>   #define blk_queue_zone_resetall(q)	\
> @@ -1300,8 +1298,14 @@ static inline bool bdev_synchronous(struct block_device *bdev)
>   
>   static inline bool bdev_stable_writes(struct block_device *bdev)
>   {
> -	return test_bit(QUEUE_FLAG_STABLE_WRITES,
> -			&bdev_get_queue(bdev)->queue_flags);
> +	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)
> +		return true;
> +#endif
> +	return test_bit(QUEUE_FLAG_STABLE_WRITES, &q->queue_flags);
>   }
>   
>   static inline bool bdev_write_cache(struct block_device *bdev)

Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 09/11] block: bypass the STABLE_WRITES flag for protection information

[Martin K. Petersen]

Christoph,

> Currently registering a checksum-enabled (aka PI) integrity profile sets
> the QUEUE_FLAG_STABLE_WRITE flag, and unregistering it clears the flag.
> This can incorrectly clear the flag when the driver requires stable
> writes even without PI, e.g. in case of iSCSI or NVMe/TCP with data
> digest enabled.
>
> Fix this by looking at the csum_type directly in bdev_stable_writes and
> not setting the queue flag.  Also remove the blk_queue_stable_writes
> helper as the only user in nvme wants to only look at the actual
> QUEUE_FLAG_STABLE_WRITE flag as it inherits the integrity configuration
> by other means.

Looks OK.

Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>

-- 
Martin K. Petersen	Oracle Linux Engineering

[PATCH 10/11] block: invert the BLK_INTEGRITY_{GENERATE,VERIFY} flags

[Christoph Hellwig]

Invert the flags so that user set values will be able to persist
revalidating the integrity information once we switch the integrity
information to queue_limits.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
---
 block/bio-integrity.c         |  4 ++--
 block/blk-integrity.c         | 18 +++++++++---------
 include/linux/blk-integrity.h |  4 ++--
 3 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/block/bio-integrity.c b/block/bio-integrity.c
index e23acf59c12de3..48ddbc703cd8ff 100644
--- a/block/bio-integrity.c
+++ b/block/bio-integrity.c
@@ -448,10 +448,10 @@ bool bio_integrity_prep(struct bio *bio)
 		return true;
 
 	if (bio_data_dir(bio) == READ) {
-		if (!(bi->flags & BLK_INTEGRITY_VERIFY))
+		if (bi->flags & BLK_INTEGRITY_NOVERIFY)
 			return true;
 	} else {
-		if (!(bi->flags & BLK_INTEGRITY_GENERATE))
+		if (bi->flags & BLK_INTEGRITY_NOGENERATE)
 			return true;
 
 		/*
diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index bec0d1df387ce9..b37b8855eed147 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -254,10 +254,11 @@ static ssize_t flag_store(struct device *dev, struct device_attribute *attr,
 	if (err)
 		return err;
 
+	/* the flags are inverted vs the values in the sysfs files */
 	if (val)
-		bi->flags |= flag;
-	else
 		bi->flags &= ~flag;
+	else
+		bi->flags |= flag;
 	return count;
 }
 
@@ -266,7 +267,7 @@ static ssize_t flag_show(struct device *dev, struct device_attribute *attr,
 {
 	struct blk_integrity *bi = dev_to_bi(dev);
 
-	return sysfs_emit(page, "%d\n", !!(bi->flags & flag));
+	return sysfs_emit(page, "%d\n", !(bi->flags & flag));
 }
 
 static ssize_t format_show(struct device *dev, struct device_attribute *attr,
@@ -301,26 +302,26 @@ static ssize_t read_verify_store(struct device *dev,
 				 struct device_attribute *attr,
 				 const char *page, size_t count)
 {
-	return flag_store(dev, attr, page, count, BLK_INTEGRITY_VERIFY);
+	return flag_store(dev, attr, page, count, BLK_INTEGRITY_NOVERIFY);
 }
 
 static ssize_t read_verify_show(struct device *dev,
 				struct device_attribute *attr, char *page)
 {
-	return flag_show(dev, attr, page, BLK_INTEGRITY_VERIFY);
+	return flag_show(dev, attr, page, BLK_INTEGRITY_NOVERIFY);
 }
 
 static ssize_t write_generate_store(struct device *dev,
 				    struct device_attribute *attr,
 				    const char *page, size_t count)
 {
-	return flag_store(dev, attr, page, count, BLK_INTEGRITY_GENERATE);
+	return flag_store(dev, attr, page, count, BLK_INTEGRITY_NOGENERATE);
 }
 
 static ssize_t write_generate_show(struct device *dev,
 				   struct device_attribute *attr, char *page)
 {
-	return flag_show(dev, attr, page, BLK_INTEGRITY_GENERATE);
+	return flag_show(dev, attr, page, BLK_INTEGRITY_NOGENERATE);
 }
 
 static ssize_t device_is_integrity_capable_show(struct device *dev,
@@ -371,8 +372,7 @@ 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 = BLK_INTEGRITY_VERIFY | BLK_INTEGRITY_GENERATE |
-		template->flags;
+	bi->flags = template->flags;
 	bi->interval_exp = template->interval_exp ? :
 		ilog2(queue_logical_block_size(disk->queue));
 	bi->tuple_size = template->tuple_size;
diff --git a/include/linux/blk-integrity.h b/include/linux/blk-integrity.h
index 56ce1ae355805d..bafa01d4e7f95b 100644
--- a/include/linux/blk-integrity.h
+++ b/include/linux/blk-integrity.h
@@ -7,8 +7,8 @@
 struct request;
 
 enum blk_integrity_flags {
-	BLK_INTEGRITY_VERIFY		= 1 << 0,
-	BLK_INTEGRITY_GENERATE		= 1 << 1,
+	BLK_INTEGRITY_NOVERIFY		= 1 << 0,
+	BLK_INTEGRITY_NOGENERATE	= 1 << 1,
 	BLK_INTEGRITY_DEVICE_CAPABLE	= 1 << 2,
 	BLK_INTEGRITY_REF_TAG		= 1 << 3,
 };
-- 
2.43.0

Re: [PATCH 10/11] block: invert the BLK_INTEGRITY_{GENERATE,VERIFY} flags

[Hannes Reinecke]

On 6/7/24 07:59, Christoph Hellwig wrote:
> Invert the flags so that user set values will be able to persist
> revalidating the integrity information once we switch the integrity
> information to queue_limits.
> 
> Signed-off-by: Christoph Hellwig <hch@lst.de>
> Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
> ---
>   block/bio-integrity.c         |  4 ++--
>   block/blk-integrity.c         | 18 +++++++++---------
>   include/linux/blk-integrity.h |  4 ++--
>   3 files changed, 13 insertions(+), 13 deletions(-)
> 
Reviewed-by: Hannes Reinecke <hare@suse.de>

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

[PATCH 11/11] block: move integrity information into queue_limits

[Christoph Hellwig]

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 <hch@lst.de>
---
 Documentation/block/data-integrity.rst |  49 +-------
 block/blk-integrity.c                  | 124 ++-----------------
 block/blk-settings.c                   | 118 +++++++++++++++++-
 block/t10-pi.c                         |  12 +-
 drivers/md/dm-core.h                   |   1 -
 drivers/md/dm-integrity.c              |  27 ++---
 drivers/md/dm-table.c                  | 161 +++++--------------------
 drivers/md/md.c                        |  72 +++--------
 drivers/md/md.h                        |   5 +-
 drivers/md/raid0.c                     |   7 +-
 drivers/md/raid1.c                     |  10 +-
 drivers/md/raid10.c                    |  10 +-
 drivers/md/raid5.c                     |   2 +-
 drivers/nvdimm/btt.c                   |  13 +-
 drivers/nvme/host/core.c               |  70 +++++------
 drivers/scsi/sd.c                      |  10 +-
 drivers/scsi/sd.h                      |  12 +-
 drivers/scsi/sd_dif.c                  |  34 +++---
 include/linux/blk-integrity.h          |  27 ++---
 include/linux/blkdev.h                 |  11 +-
 include/linux/t10-pi.h                 |  12 +-
 21 files changed, 290 insertions(+), 497 deletions(-)

diff --git a/Documentation/block/data-integrity.rst b/Documentation/block/data-integrity.rst
index 6a760c0eb1924e..99905e880a0e56 100644
--- a/Documentation/block/data-integrity.rst
+++ b/Documentation/block/data-integrity.rst
@@ -153,18 +153,11 @@ bio_free() will automatically free the bip.
 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.
 
@@ -250,42 +243,6 @@ will require extra work due to the application tag.
       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.
-
 ----------------------------------------------------------------------
 
 2007-12-24 Martin K. Petersen <martin.petersen@oracle.com>
diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index b37b8855eed147..05a48689a424b2 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -107,63 +107,6 @@ int blk_rq_map_integrity_sg(struct request_queue *q, struct bio *bio,
 }
 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)
 {
@@ -217,7 +160,7 @@ bool blk_integrity_merge_bio(struct request_queue *q, struct request *req,
 
 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)
@@ -246,7 +189,8 @@ EXPORT_SYMBOL_GPL(blk_integrity_profile_name);
 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;
 
@@ -254,11 +198,18 @@ static ssize_t flag_store(struct device *dev, struct device_attribute *attr,
 	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;
 }
 
@@ -355,52 +306,3 @@ const struct attribute_group blk_integrity_attr_group = {
 	.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);
diff --git a/block/blk-settings.c b/block/blk-settings.c
index 996f247fc98e80..f11c8676eb4c67 100644
--- a/block/blk-settings.c
+++ b/block/blk-settings.c
@@ -6,7 +6,7 @@
 #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>
@@ -97,6 +97,36 @@ static int blk_validate_zoned_limits(struct queue_limits *lim)
 	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.
@@ -105,6 +135,7 @@ static int blk_validate_limits(struct queue_limits *lim)
 {
 	unsigned int max_hw_sectors;
 	unsigned int logical_block_sectors;
+	int err;
 
 	/*
 	 * Unless otherwise specified, default to 512 byte logical blocks and a
@@ -230,6 +261,9 @@ static int blk_validate_limits(struct queue_limits *lim)
 		lim->misaligned = 0;
 	}
 
+	err = blk_validate_integrity_limits(lim);
+	if (err)
+		return err;
 	return blk_validate_zoned_limits(lim);
 }
 
@@ -263,13 +297,24 @@ int queue_limits_commit_update(struct request_queue *q,
 		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;
 }
@@ -575,6 +620,67 @@ void queue_limits_stack_bdev(struct queue_limits *t, struct block_device *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
  * @q:     the request queue for the device
diff --git a/block/t10-pi.c b/block/t10-pi.c
index dadecf621497bb..cd7fa60d63ff21 100644
--- a/block/t10-pi.c
+++ b/block/t10-pi.c
@@ -116,7 +116,7 @@ static blk_status_t t10_pi_verify(struct blk_integrity_iter *iter,
  */
 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;
@@ -167,7 +167,7 @@ static void t10_pi_type1_prepare(struct request *rq)
  */
 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);
@@ -290,7 +290,7 @@ static blk_status_t ext_pi_crc64_verify(struct blk_integrity_iter *iter,
 
 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;
@@ -330,7 +330,7 @@ static void ext_pi_type1_prepare(struct request *rq)
 
 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);
@@ -396,7 +396,7 @@ blk_status_t blk_integrity_verify(struct blk_integrity_iter *iter,
 
 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;
@@ -409,7 +409,7 @@ void blk_integrity_prepare(struct request *rq)
 
 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;
diff --git a/drivers/md/dm-core.h b/drivers/md/dm-core.h
index 08700bfc3e2343..14a44c0f82868b 100644
--- a/drivers/md/dm-core.h
+++ b/drivers/md/dm-core.h
@@ -206,7 +206,6 @@ struct dm_table {
 
 	bool integrity_supported:1;
 	bool singleton:1;
-	unsigned integrity_added:1;
 
 	/*
 	 * Indicates the rw permissions for the new logical device.  This
diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c
index c1cc27541673c7..2a89f8eb4713c9 100644
--- a/drivers/md/dm-integrity.c
+++ b/drivers/md/dm-integrity.c
@@ -3475,6 +3475,17 @@ static void dm_integrity_io_hints(struct dm_target *ti, struct queue_limits *lim
 		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;
 }
 
@@ -3631,19 +3642,6 @@ static int initialize_superblock(struct dm_integrity_c *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)
 {
 	unsigned int i;
@@ -4629,9 +4627,6 @@ static int dm_integrity_ctr(struct dm_target *ti, unsigned int argc, char **argv
 		}
 	}
 
-	if (!ic->internal_hash)
-		dm_integrity_set(ti, ic);
-
 	ti->num_flush_bios = 1;
 	ti->flush_supported = true;
 	if (ic->discard)
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index b2d5246cff2102..fd789eeb62d943 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -425,6 +425,13 @@ static int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev,
 		       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;
 }
 
@@ -702,9 +709,6 @@ int dm_table_add_target(struct dm_table *t, const char *type,
 		t->immutable_target_type = ti->type;
 	}
 
-	if (dm_target_has_integrity(ti->type))
-		t->integrity_added = 1;
-
 	ti->table = t;
 	ti->begin = start;
 	ti->len = len;
@@ -1119,99 +1123,6 @@ static int dm_table_build_index(struct dm_table *t)
 	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 {
@@ -1423,12 +1334,6 @@ int dm_table_complete(struct dm_table *t)
 		return r;
 	}
 
-	r = dm_table_register_integrity(t);
-	if (r) {
-		DMERR("could not register integrity profile.");
-		return r;
-	}
-
 	r = dm_table_construct_crypto_profile(t);
 	if (r) {
 		DMERR("could not construct crypto profile.");
@@ -1688,6 +1593,14 @@ int dm_calculate_queue_limits(struct dm_table *t,
 
 	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);
 
@@ -1738,6 +1651,18 @@ int dm_calculate_queue_limits(struct dm_table *t,
 			       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;
+			}
+		}
 	}
 
 	/*
@@ -1761,36 +1686,6 @@ int dm_calculate_queue_limits(struct dm_table *t,
 	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,
 				sector_t start, sector_t len, void *data)
 {
@@ -2004,8 +1899,6 @@ int dm_table_set_restrictions(struct dm_table *t, struct request_queue *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
 	 * because they do their own checksumming.
diff --git a/drivers/md/md.c b/drivers/md/md.c
index aff9118ff69750..67ece2cd725f50 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -2410,36 +2410,10 @@ static LIST_HEAD(pending_raid_disks);
  */
 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) ||
@@ -2459,32 +2433,6 @@ int md_integrity_register(struct mddev *mddev)
 }
 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)
 {
 	return bdev_read_only(rdev->bdev) ||
@@ -5755,14 +5703,20 @@ static const struct kobj_type md_ktype = {
 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);
 
@@ -5777,6 +5731,14 @@ int mddev_stack_new_rdev(struct mddev *mddev, struct md_rdev *rdev)
 	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);
diff --git a/drivers/md/md.h b/drivers/md/md.h
index ca085ecad50449..6733b0b0abf999 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -809,7 +809,6 @@ extern void md_wait_for_blocked_rdev(struct md_rdev *rdev, struct mddev *mddev);
 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);
@@ -908,7 +907,9 @@ void md_autostart_arrays(int part);
 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);
 
diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
index 81c01347cd24e6..62634e2a33bd0f 100644
--- a/drivers/md/raid0.c
+++ b/drivers/md/raid0.c
@@ -377,13 +377,18 @@ static void raid0_free(struct mddev *mddev, void *priv)
 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);
 }
 
diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
index 1f321826ef02ba..779cad62f6f8c0 100644
--- a/drivers/md/raid1.c
+++ b/drivers/md/raid1.c
@@ -1907,9 +1907,6 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 	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;
 
@@ -3197,10 +3194,15 @@ static struct r1conf *setup_conf(struct mddev *mddev)
 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);
 }
 
diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
index a4556d2e46bf95..5f6885b53b691a 100644
--- a/drivers/md/raid10.c
+++ b/drivers/md/raid10.c
@@ -2083,9 +2083,6 @@ static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev)
 	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;
 
@@ -3980,12 +3977,17 @@ static int raid10_set_queue_limits(struct mddev *mddev)
 {
 	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);
 }
 
diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
index 2bd1ce9b39226a..675c68fa6c6403 100644
--- a/drivers/md/raid5.c
+++ b/drivers/md/raid5.c
@@ -7708,7 +7708,7 @@ static int raid5_set_limits(struct mddev *mddev)
 	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);
diff --git a/drivers/nvdimm/btt.c b/drivers/nvdimm/btt.c
index 1e5aedaf8c7bd9..c5f8451b494d6c 100644
--- a/drivers/nvdimm/btt.c
+++ b/drivers/nvdimm/btt.c
@@ -1504,6 +1504,11 @@ static int btt_blk_init(struct btt *btt)
 	};
 	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);
@@ -1516,14 +1521,6 @@ static int btt_blk_init(struct btt *btt)
 	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);
 	if (rc)
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 14bac248cde4ca..5a673fa5cb2612 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1723,11 +1723,12 @@ int nvme_getgeo(struct block_device *bdev, struct hd_geometry *geo)
 	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;
@@ -1744,14 +1745,14 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
 	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;
@@ -1761,16 +1762,16 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
 	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;
@@ -1780,9 +1781,8 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
 		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;
 }
 
@@ -2105,11 +2105,6 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,
 	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
@@ -2117,9 +2112,15 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,
 	 * 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);
 
 	/*
@@ -2191,14 +2192,6 @@ static int nvme_update_ns_info(struct nvme_ns *ns, struct nvme_ns_info *info)
 		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.
@@ -2221,7 +2214,16 @@ static int nvme_update_ns_info(struct nvme_ns *ns, struct nvme_ns_info *info)
 		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);
 	}
 
diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index 67f24607b862a3..85b45345a27739 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -799,7 +799,7 @@ static unsigned char sd_setup_protect_cmnd(struct scsi_cmnd *scmd,
 					   unsigned int dix, unsigned int dif)
 {
 	struct request *rq = scsi_cmd_to_rq(scmd);
-	struct blk_integrity *bi = &rq->q->integrity;
+	struct blk_integrity *bi = &rq->q->limits.integrity;
 	unsigned int prot_op = sd_prot_op(rq_data_dir(rq), dix, dif);
 	unsigned int protect = 0;
 
@@ -2474,11 +2474,13 @@ static int sd_read_protection_type(struct scsi_disk *sdkp, unsigned char *buffer
 	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;
@@ -3669,7 +3671,7 @@ static int sd_revalidate_disk(struct gendisk *disk)
 		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);
 	}
 
 	/*
diff --git a/drivers/scsi/sd.h b/drivers/scsi/sd.h
index b4170b17bad47a..726f1613f6cb56 100644
--- a/drivers/scsi/sd.h
+++ b/drivers/scsi/sd.h
@@ -220,17 +220,7 @@ static inline sector_t sectors_to_logical(struct scsi_device *sdev, sector_t sec
 	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)
 {
diff --git a/drivers/scsi/sd_dif.c b/drivers/scsi/sd_dif.c
index 6f0921c7db787b..ae6ce6f5d622d9 100644
--- a/drivers/scsi/sd_dif.c
+++ b/drivers/scsi/sd_dif.c
@@ -24,14 +24,15 @@
 /*
  * 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);
 
@@ -39,40 +40,33 @@ void sd_dif_config_host(struct scsi_disk *sdkp)
 		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);
 }
-
diff --git a/include/linux/blk-integrity.h b/include/linux/blk-integrity.h
index bafa01d4e7f95b..d201140d77a336 100644
--- a/include/linux/blk-integrity.h
+++ b/include/linux/blk-integrity.h
@@ -11,6 +11,7 @@ enum blk_integrity_flags {
 	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 {
@@ -23,11 +24,15 @@ 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 *);
@@ -35,14 +40,14 @@ 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 *
@@ -119,17 +124,6 @@ blk_integrity_queue_supports_integrity(struct request_queue *q)
 {
 	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)
 {
@@ -157,4 +151,5 @@ static inline struct bio_vec *rq_integrity_vec(struct request *rq)
 	return NULL;
 }
 #endif /* CONFIG_BLK_DEV_INTEGRITY */
+
 #endif /* _LINUX_BLK_INTEGRITY_H */
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index f9089750919c6b..c792d4d81e5fcc 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -334,6 +334,8 @@ struct queue_limits {
 	 * due to possible offsets.
 	 */
 	unsigned int		dma_alignment;
+
+	struct blk_integrity	integrity;
 };
 
 typedef int (*report_zones_cb)(struct blk_zone *zone, unsigned int idx,
@@ -419,10 +421,6 @@ struct request_queue {
 
 	struct queue_limits	limits;
 
-#ifdef  CONFIG_BLK_DEV_INTEGRITY
-	struct blk_integrity integrity;
-#endif	/* CONFIG_BLK_DEV_INTEGRITY */
-
 #ifdef CONFIG_PM
 	struct device		*dev;
 	enum rpm_status		rpm_status;
@@ -1300,11 +1298,10 @@ static inline bool bdev_stable_writes(struct block_device *bdev)
 {
 	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 != 0)
 		return true;
-#endif
 	return test_bit(QUEUE_FLAG_STABLE_WRITES, &q->queue_flags);
 }
 
diff --git a/include/linux/t10-pi.h b/include/linux/t10-pi.h
index d2bafb76badfb9..1773610010ebaf 100644
--- a/include/linux/t10-pi.h
+++ b/include/linux/t10-pi.h
@@ -39,12 +39,8 @@ struct t10_pi_tuple {
 
 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;
 }
 
@@ -65,12 +61,8 @@ static inline u64 lower_48_bits(u64 n)
 
 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));
 }
 
-- 
2.43.0

Re: [PATCH 11/11] block: move integrity information into queue_limits

[Hannes Reinecke]

On 6/7/24 07:59, Christoph Hellwig wrote:
> 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 <hch@lst.de>
> ---
>   Documentation/block/data-integrity.rst |  49 +-------
>   block/blk-integrity.c                  | 124 ++-----------------
>   block/blk-settings.c                   | 118 +++++++++++++++++-
>   block/t10-pi.c                         |  12 +-
>   drivers/md/dm-core.h                   |   1 -
>   drivers/md/dm-integrity.c              |  27 ++---
>   drivers/md/dm-table.c                  | 161 +++++--------------------
>   drivers/md/md.c                        |  72 +++--------
>   drivers/md/md.h                        |   5 +-
>   drivers/md/raid0.c                     |   7 +-
>   drivers/md/raid1.c                     |  10 +-
>   drivers/md/raid10.c                    |  10 +-
>   drivers/md/raid5.c                     |   2 +-
>   drivers/nvdimm/btt.c                   |  13 +-
>   drivers/nvme/host/core.c               |  70 +++++------
>   drivers/scsi/sd.c                      |  10 +-
>   drivers/scsi/sd.h                      |  12 +-
>   drivers/scsi/sd_dif.c                  |  34 +++---
>   include/linux/blk-integrity.h          |  27 ++---
>   include/linux/blkdev.h                 |  11 +-
>   include/linux/t10-pi.h                 |  12 +-
>   21 files changed, 290 insertions(+), 497 deletions(-)
> 
> diff --git a/Documentation/block/data-integrity.rst b/Documentation/block/data-integrity.rst
> index 6a760c0eb1924e..99905e880a0e56 100644
> --- a/Documentation/block/data-integrity.rst
> +++ b/Documentation/block/data-integrity.rst
> @@ -153,18 +153,11 @@ bio_free() will automatically free the bip.
>   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.
>   
> @@ -250,42 +243,6 @@ will require extra work due to the application tag.
>         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.
> -
>   ----------------------------------------------------------------------
>   
>   2007-12-24 Martin K. Petersen <martin.petersen@oracle.com>
> diff --git a/block/blk-integrity.c b/block/blk-integrity.c
> index b37b8855eed147..05a48689a424b2 100644
> --- a/block/blk-integrity.c
> +++ b/block/blk-integrity.c
> @@ -107,63 +107,6 @@ int blk_rq_map_integrity_sg(struct request_queue *q, struct bio *bio,
>   }
>   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)
>   {
> @@ -217,7 +160,7 @@ bool blk_integrity_merge_bio(struct request_queue *q, struct request *req,
>   
>   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)
> @@ -246,7 +189,8 @@ EXPORT_SYMBOL_GPL(blk_integrity_profile_name);
>   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;
>   
> @@ -254,11 +198,18 @@ static ssize_t flag_store(struct device *dev, struct device_attribute *attr,
>   	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;
>   }
>   
> @@ -355,52 +306,3 @@ const struct attribute_group blk_integrity_attr_group = {
>   	.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);
> diff --git a/block/blk-settings.c b/block/blk-settings.c
> index 996f247fc98e80..f11c8676eb4c67 100644
> --- a/block/blk-settings.c
> +++ b/block/blk-settings.c
> @@ -6,7 +6,7 @@
>   #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>
> @@ -97,6 +97,36 @@ static int blk_validate_zoned_limits(struct queue_limits *lim)
>   	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;
> +	}
> +
Why is that an error?
Surely 'validate' should not return an error if BLK_DEV_INTEGRITY is 
disabled and no limits are set?

> +	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.
> @@ -105,6 +135,7 @@ static int blk_validate_limits(struct queue_limits *lim)
>   {
>   	unsigned int max_hw_sectors;
>   	unsigned int logical_block_sectors;
> +	int err;
>   
>   	/*
>   	 * Unless otherwise specified, default to 512 byte logical blocks and a
> @@ -230,6 +261,9 @@ static int blk_validate_limits(struct queue_limits *lim)
>   		lim->misaligned = 0;
>   	}
>   
> +	err = blk_validate_integrity_limits(lim);
> +	if (err)
> +		return err;
Wouldn't we always fail to validate the limits if BLK_DEV_INTEGRITY is 
disabled, given the check above?

>   	return blk_validate_zoned_limits(lim);
>   }
>   
> @@ -263,13 +297,24 @@ int queue_limits_commit_update(struct request_queue *q,
>   		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;
>   }
> @@ -575,6 +620,67 @@ void queue_limits_stack_bdev(struct queue_limits *t, struct block_device *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
>    * @q:     the request queue for the device
> diff --git a/block/t10-pi.c b/block/t10-pi.c
> index dadecf621497bb..cd7fa60d63ff21 100644
> --- a/block/t10-pi.c
> +++ b/block/t10-pi.c
> @@ -116,7 +116,7 @@ static blk_status_t t10_pi_verify(struct blk_integrity_iter *iter,
>    */
>   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;
> @@ -167,7 +167,7 @@ static void t10_pi_type1_prepare(struct request *rq)
>    */
>   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);
> @@ -290,7 +290,7 @@ static blk_status_t ext_pi_crc64_verify(struct blk_integrity_iter *iter,
>   
>   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;
> @@ -330,7 +330,7 @@ static void ext_pi_type1_prepare(struct request *rq)
>   
>   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);
> @@ -396,7 +396,7 @@ blk_status_t blk_integrity_verify(struct blk_integrity_iter *iter,
>   
>   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;
> @@ -409,7 +409,7 @@ void blk_integrity_prepare(struct request *rq)
>   
>   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;
> diff --git a/drivers/md/dm-core.h b/drivers/md/dm-core.h
> index 08700bfc3e2343..14a44c0f82868b 100644
> --- a/drivers/md/dm-core.h
> +++ b/drivers/md/dm-core.h
> @@ -206,7 +206,6 @@ struct dm_table {
>   
>   	bool integrity_supported:1;
>   	bool singleton:1;
> -	unsigned integrity_added:1;
>   
>   	/*
>   	 * Indicates the rw permissions for the new logical device.  This
> diff --git a/drivers/md/dm-integrity.c b/drivers/md/dm-integrity.c
> index c1cc27541673c7..2a89f8eb4713c9 100644
> --- a/drivers/md/dm-integrity.c
> +++ b/drivers/md/dm-integrity.c
> @@ -3475,6 +3475,17 @@ static void dm_integrity_io_hints(struct dm_target *ti, struct queue_limits *lim
>   		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;
>   }
>   
> @@ -3631,19 +3642,6 @@ static int initialize_superblock(struct dm_integrity_c *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)
>   {
>   	unsigned int i;
> @@ -4629,9 +4627,6 @@ static int dm_integrity_ctr(struct dm_target *ti, unsigned int argc, char **argv
>   		}
>   	}
>   
> -	if (!ic->internal_hash)
> -		dm_integrity_set(ti, ic);
> -
>   	ti->num_flush_bios = 1;
>   	ti->flush_supported = true;
>   	if (ic->discard)
> diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
> index b2d5246cff2102..fd789eeb62d943 100644
> --- a/drivers/md/dm-table.c
> +++ b/drivers/md/dm-table.c
> @@ -425,6 +425,13 @@ static int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev,
>   		       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;
>   }
>   
> @@ -702,9 +709,6 @@ int dm_table_add_target(struct dm_table *t, const char *type,
>   		t->immutable_target_type = ti->type;
>   	}
>   
> -	if (dm_target_has_integrity(ti->type))
> -		t->integrity_added = 1;
> -
>   	ti->table = t;
>   	ti->begin = start;
>   	ti->len = len;
> @@ -1119,99 +1123,6 @@ static int dm_table_build_index(struct dm_table *t)
>   	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 {
> @@ -1423,12 +1334,6 @@ int dm_table_complete(struct dm_table *t)
>   		return r;
>   	}
>   
> -	r = dm_table_register_integrity(t);
> -	if (r) {
> -		DMERR("could not register integrity profile.");
> -		return r;
> -	}
> -
>   	r = dm_table_construct_crypto_profile(t);
>   	if (r) {
>   		DMERR("could not construct crypto profile.");
> @@ -1688,6 +1593,14 @@ int dm_calculate_queue_limits(struct dm_table *t,
>   
>   	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);
>   
> @@ -1738,6 +1651,18 @@ int dm_calculate_queue_limits(struct dm_table *t,
>   			       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;
> +			}
> +		}
>   	}
>   
>   	/*
> @@ -1761,36 +1686,6 @@ int dm_calculate_queue_limits(struct dm_table *t,
>   	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,
>   				sector_t start, sector_t len, void *data)
>   {
> @@ -2004,8 +1899,6 @@ int dm_table_set_restrictions(struct dm_table *t, struct request_queue *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
>   	 * because they do their own checksumming.
> diff --git a/drivers/md/md.c b/drivers/md/md.c
> index aff9118ff69750..67ece2cd725f50 100644
> --- a/drivers/md/md.c
> +++ b/drivers/md/md.c
> @@ -2410,36 +2410,10 @@ static LIST_HEAD(pending_raid_disks);
>    */
>   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) ||
> @@ -2459,32 +2433,6 @@ int md_integrity_register(struct mddev *mddev)
>   }
>   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)
>   {
>   	return bdev_read_only(rdev->bdev) ||
> @@ -5755,14 +5703,20 @@ static const struct kobj_type md_ktype = {
>   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);
>   
> @@ -5777,6 +5731,14 @@ int mddev_stack_new_rdev(struct mddev *mddev, struct md_rdev *rdev)
>   	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);
> diff --git a/drivers/md/md.h b/drivers/md/md.h
> index ca085ecad50449..6733b0b0abf999 100644
> --- a/drivers/md/md.h
> +++ b/drivers/md/md.h
> @@ -809,7 +809,6 @@ extern void md_wait_for_blocked_rdev(struct md_rdev *rdev, struct mddev *mddev);
>   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);
> @@ -908,7 +907,9 @@ void md_autostart_arrays(int part);
>   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);
>   
> diff --git a/drivers/md/raid0.c b/drivers/md/raid0.c
> index 81c01347cd24e6..62634e2a33bd0f 100644
> --- a/drivers/md/raid0.c
> +++ b/drivers/md/raid0.c
> @@ -377,13 +377,18 @@ static void raid0_free(struct mddev *mddev, void *priv)
>   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);
>   }
>   
> diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c
> index 1f321826ef02ba..779cad62f6f8c0 100644
> --- a/drivers/md/raid1.c
> +++ b/drivers/md/raid1.c
> @@ -1907,9 +1907,6 @@ static int raid1_add_disk(struct mddev *mddev, struct md_rdev *rdev)
>   	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;
>   
> @@ -3197,10 +3194,15 @@ static struct r1conf *setup_conf(struct mddev *mddev)
>   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);
>   }
>   
> diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c
> index a4556d2e46bf95..5f6885b53b691a 100644
> --- a/drivers/md/raid10.c
> +++ b/drivers/md/raid10.c
> @@ -2083,9 +2083,6 @@ static int raid10_add_disk(struct mddev *mddev, struct md_rdev *rdev)
>   	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;
>   
> @@ -3980,12 +3977,17 @@ static int raid10_set_queue_limits(struct mddev *mddev)
>   {
>   	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);
>   }
>   
> diff --git a/drivers/md/raid5.c b/drivers/md/raid5.c
> index 2bd1ce9b39226a..675c68fa6c6403 100644
> --- a/drivers/md/raid5.c
> +++ b/drivers/md/raid5.c
> @@ -7708,7 +7708,7 @@ static int raid5_set_limits(struct mddev *mddev)
>   	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);
> diff --git a/drivers/nvdimm/btt.c b/drivers/nvdimm/btt.c
> index 1e5aedaf8c7bd9..c5f8451b494d6c 100644
> --- a/drivers/nvdimm/btt.c
> +++ b/drivers/nvdimm/btt.c
> @@ -1504,6 +1504,11 @@ static int btt_blk_init(struct btt *btt)
>   	};
>   	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);
> @@ -1516,14 +1521,6 @@ static int btt_blk_init(struct btt *btt)
>   	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);
>   	if (rc)
> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
> index 14bac248cde4ca..5a673fa5cb2612 100644
> --- a/drivers/nvme/host/core.c
> +++ b/drivers/nvme/host/core.c
> @@ -1723,11 +1723,12 @@ int nvme_getgeo(struct block_device *bdev, struct hd_geometry *geo)
>   	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;
> @@ -1744,14 +1745,14 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
>   	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;
> @@ -1761,16 +1762,16 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
>   	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;
> @@ -1780,9 +1781,8 @@ static bool nvme_init_integrity(struct gendisk *disk, struct nvme_ns_head *head)
>   		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;
>   }
>   
> @@ -2105,11 +2105,6 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,
>   	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
> @@ -2117,9 +2112,15 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,
>   	 * 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);
>   
>   	/*
> @@ -2191,14 +2192,6 @@ static int nvme_update_ns_info(struct nvme_ns *ns, struct nvme_ns_info *info)
>   		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.
> @@ -2221,7 +2214,16 @@ static int nvme_update_ns_info(struct nvme_ns *ns, struct nvme_ns_info *info)
>   		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);
>   	}
>   
> diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
> index 67f24607b862a3..85b45345a27739 100644
> --- a/drivers/scsi/sd.c
> +++ b/drivers/scsi/sd.c
> @@ -799,7 +799,7 @@ static unsigned char sd_setup_protect_cmnd(struct scsi_cmnd *scmd,
>   					   unsigned int dix, unsigned int dif)
>   {
>   	struct request *rq = scsi_cmd_to_rq(scmd);
> -	struct blk_integrity *bi = &rq->q->integrity;
> +	struct blk_integrity *bi = &rq->q->limits.integrity;
>   	unsigned int prot_op = sd_prot_op(rq_data_dir(rq), dix, dif);
>   	unsigned int protect = 0;
>   
> @@ -2474,11 +2474,13 @@ static int sd_read_protection_type(struct scsi_disk *sdkp, unsigned char *buffer
>   	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;
> @@ -3669,7 +3671,7 @@ static int sd_revalidate_disk(struct gendisk *disk)
>   		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);
>   	}
>   
>   	/*
> diff --git a/drivers/scsi/sd.h b/drivers/scsi/sd.h
> index b4170b17bad47a..726f1613f6cb56 100644
> --- a/drivers/scsi/sd.h
> +++ b/drivers/scsi/sd.h
> @@ -220,17 +220,7 @@ static inline sector_t sectors_to_logical(struct scsi_device *sdev, sector_t sec
>   	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)
>   {
> diff --git a/drivers/scsi/sd_dif.c b/drivers/scsi/sd_dif.c
> index 6f0921c7db787b..ae6ce6f5d622d9 100644
> --- a/drivers/scsi/sd_dif.c
> +++ b/drivers/scsi/sd_dif.c
> @@ -24,14 +24,15 @@
>   /*
>    * 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);
>   
> @@ -39,40 +40,33 @@ void sd_dif_config_host(struct scsi_disk *sdkp)
>   		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);
>   }
> -
> diff --git a/include/linux/blk-integrity.h b/include/linux/blk-integrity.h
> index bafa01d4e7f95b..d201140d77a336 100644
> --- a/include/linux/blk-integrity.h
> +++ b/include/linux/blk-integrity.h
> @@ -11,6 +11,7 @@ enum blk_integrity_flags {
>   	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 {
> @@ -23,11 +24,15 @@ 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 *);
> @@ -35,14 +40,14 @@ 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 *
> @@ -119,17 +124,6 @@ blk_integrity_queue_supports_integrity(struct request_queue *q)
>   {
>   	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)
>   {
> @@ -157,4 +151,5 @@ static inline struct bio_vec *rq_integrity_vec(struct request *rq)
>   	return NULL;
>   }
>   #endif /* CONFIG_BLK_DEV_INTEGRITY */
> +
>   #endif /* _LINUX_BLK_INTEGRITY_H */
> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
> index f9089750919c6b..c792d4d81e5fcc 100644
> --- a/include/linux/blkdev.h
> +++ b/include/linux/blkdev.h
> @@ -334,6 +334,8 @@ struct queue_limits {
>   	 * due to possible offsets.
>   	 */
>   	unsigned int		dma_alignment;
> +
> +	struct blk_integrity	integrity;
>   };
>   
>   typedef int (*report_zones_cb)(struct blk_zone *zone, unsigned int idx,
> @@ -419,10 +421,6 @@ struct request_queue {
>   
>   	struct queue_limits	limits;
>   
> -#ifdef  CONFIG_BLK_DEV_INTEGRITY
> -	struct blk_integrity integrity;
> -#endif	/* CONFIG_BLK_DEV_INTEGRITY */
> -
>   #ifdef CONFIG_PM
>   	struct device		*dev;
>   	enum rpm_status		rpm_status;
> @@ -1300,11 +1298,10 @@ static inline bool bdev_stable_writes(struct block_device *bdev)
>   {
>   	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 != 0)
>   		return true;
> -#endif
>   	return test_bit(QUEUE_FLAG_STABLE_WRITES, &q->queue_flags);
>   }
>   
> diff --git a/include/linux/t10-pi.h b/include/linux/t10-pi.h
> index d2bafb76badfb9..1773610010ebaf 100644
> --- a/include/linux/t10-pi.h
> +++ b/include/linux/t10-pi.h
> @@ -39,12 +39,8 @@ struct t10_pi_tuple {
>   
>   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;
>   }
>   
> @@ -65,12 +61,8 @@ static inline u64 lower_48_bits(u64 n)
>   
>   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));
>   }
>   
Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 11/11] block: move integrity information into queue_limits

[Christoph Hellwig]

I just found a fullquota.  If you have anthing to say please trim your
reply so that it can be found.

Re: [PATCH 11/11] block: move integrity information into queue_limits

[Hannes Reinecke]

On 6/7/24 08:32, Christoph Hellwig wrote:
> I just found a fullquota.  If you have anthing to say please trim your
> reply so that it can be found.
> 
Sorry.

Trimming reply next time.

Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 11/11] block: move integrity information into queue_limits

[Hannes Reinecke]

(Trimming reply for better readability)

On 6/7/24 08:29, Hannes Reinecke wrote:
> On 6/7/24 07:59, Christoph Hellwig wrote:
[ .. ]
>> diff --git a/block/blk-settings.c b/block/blk-settings.c
>> index 996f247fc98e80..f11c8676eb4c67 100644
>> --- a/block/blk-settings.c
>> +++ b/block/blk-settings.c
>> @@ -6,7 +6,7 @@
>>   #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>
>> @@ -97,6 +97,36 @@ static int blk_validate_zoned_limits(struct 
>> queue_limits *lim)
>>       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;
>> +    }
>> +
> Why is that an error?
> Surely 'validate' should not return an error if BLK_DEV_INTEGRITY is 
> disabled and no limits are set?
> 
>> +    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.
>> @@ -105,6 +135,7 @@ static int blk_validate_limits(struct queue_limits 
>> *lim)
>>   {
>>       unsigned int max_hw_sectors;
>>       unsigned int logical_block_sectors;
>> +    int err;
>>       /*
>>        * Unless otherwise specified, default to 512 byte logical 
>> blocks and a
>> @@ -230,6 +261,9 @@ static int blk_validate_limits(struct queue_limits 
>> *lim)
>>           lim->misaligned = 0;
>>       }
>> +    err = blk_validate_integrity_limits(lim);
>> +    if (err)
>> +        return err;
> Wouldn't we always fail to validate the limits if BLK_DEV_INTEGRITY is 
> disabled, given the check above?
> 
Cheers,

Hannes
-- 
Dr. Hannes Reinecke                  Kernel Storage Architect
hare@suse.de                                +49 911 74053 688
SUSE Software Solutions GmbH, Frankenstr. 146, 90461 Nürnberg
HRB 36809 (AG Nürnberg), GF: I. Totev, A. McDonald, W. Knoblich

Re: [PATCH 11/11] block: move integrity information into queue_limits

[Christoph Hellwig]

On Fri, Jun 07, 2024 at 09:35:02AM +0200, Hannes Reinecke wrote:
>>> +?????? if (!IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY)) {
>>> +?????????????? pr_warn("integrity support disabled.\n");
>>> +?????????????? return -EINVAL;
>>> +?????? }
>>> +
>> Why is that an error?
>> Surely 'validate' should not return an error if BLK_DEV_INTEGRITY is 
>> disabled and no limits are set?

Look a few lines above, we'll never get here if no tuple_size is set,
which is now the indicator for metadata support.

Re: move integrity settings to queue_limits v2

[Mike Snitzer]

On Fri, Jun 07, 2024 at 07:58:54AM +0200, Christoph Hellwig wrote:
> Hi Jens, hi Martin,
> 
> this series converts the blk-integrity settings to sit in the queue
> limits and be updated through the atomic queue limits API.
> 
> I've mostly tested this with nvme, scsi is only covered by simple
> scsi_debug based tests.
> 
> For MD I found an pre-existing error handling bug when combining PI
> capable devices with not PI capable devices.  The fix was posted here
> (and is included in the git branch below):
> 
>    https://lore.kernel.org/linux-raid/20240604172607.3185916-1-hch@lst.de/
> 
> For dm-integrity my testing showed that even the baseline fails to create
> the luks-based dm-crypto with dm-integrity backing for the authentication
> data.  As the failure is non-fatal I've not addressed it here.

Setup is complicated. Did you test in terms of cryptsetup's testsuite?
Or something else?

Would really like to see these changes verified to work, with no
cryptsetup regressions, before they go in.
 
> Note that the support for native metadata in dm-crypt by Mikulas will
> need a rebase on top of this, but as it already requires another
> block layer patch and the changes in this series will simplify it a bit
> I hope that is ok.

Should be fine, Mikulas can you verify this series to pass
cryptsetup's testsuite before you rebase?

Thanks,
Mike

Re: move integrity settings to queue_limits v2

[Mikulas Patocka]

On Fri, 7 Jun 2024, Mike Snitzer wrote:

> On Fri, Jun 07, 2024 at 07:58:54AM +0200, Christoph Hellwig wrote:
> > Hi Jens, hi Martin,
> > 
> > this series converts the blk-integrity settings to sit in the queue
> > limits and be updated through the atomic queue limits API.
> > 
> > I've mostly tested this with nvme, scsi is only covered by simple
> > scsi_debug based tests.
> > 
> > For MD I found an pre-existing error handling bug when combining PI
> > capable devices with not PI capable devices.  The fix was posted here
> > (and is included in the git branch below):
> > 
> >    https://lore.kernel.org/linux-raid/20240604172607.3185916-1-hch@lst.de/
> > 
> > For dm-integrity my testing showed that even the baseline fails to create
> > the luks-based dm-crypto with dm-integrity backing for the authentication
> > data.  As the failure is non-fatal I've not addressed it here.
> 
> Setup is complicated. Did you test in terms of cryptsetup's testsuite?
> Or something else?
> 
> Would really like to see these changes verified to work, with no
> cryptsetup regressions, before they go in.
>  
> > Note that the support for native metadata in dm-crypt by Mikulas will
> > need a rebase on top of this, but as it already requires another
> > block layer patch and the changes in this series will simplify it a bit
> > I hope that is ok.
> 
> Should be fine, Mikulas can you verify this series to pass
> cryptsetup's testsuite before you rebase?

Yes - it passes the cryptsetup testsuite.

Mikulas

> Thanks,
> Mike
> 

Re: move integrity settings to queue_limits v2

[Christoph Hellwig]

On Fri, Jun 07, 2024 at 11:42:54AM -0400, Mike Snitzer wrote:
> > For dm-integrity my testing showed that even the baseline fails to create
> > the luks-based dm-crypto with dm-integrity backing for the authentication
> > data.  As the failure is non-fatal I've not addressed it here.
> 
> Setup is complicated. Did you test in terms of cryptsetup's testsuite?

I didn't realize cryptsetup has yet another testsuite..

> Or something else?

I just followed a random guid on the internet on how to set up the
authenticated encryption, wrote a script to do it and used that for
testing.