All of lore.kernel.org
 help / color / mirror / Atom feed
From: Guenter Roeck <linux@roeck-us.net>
To: Mike Snitzer <snitzer@redhat.com>
Cc: Mike Snitzer <snitzer@kernel.org>,
	keescook@chromium.org, sarthakkukreti@google.com,
	Greg KH <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org, Oleksandr Tymoshenko <ovt@google.com>,
	dm-devel@redhat.com, regressions@lists.linux.dev
Subject: Re: [PATCH 5.4 26/34] dm verity: set DM_TARGET_IMMUTABLE feature flag
Date: Thu, 16 Jun 2022 16:22:48 -0700	[thread overview]
Message-ID: <4dea87ca-b478-829e-2337-3d82725b6aec@roeck-us.net> (raw)
In-Reply-To: <Yqo63CvFpTDFnH3x@redhat.com>

On 6/15/22 13:02, Mike Snitzer wrote:
> On Wed, Jun 15 2022 at  1:50P -0400,
> Guenter Roeck <linux@roeck-us.net> wrote:
> 
>> On 6/15/22 08:29, Mike Snitzer wrote:
>>> On Wed, Jun 15 2022 at 10:36P -0400,
>>> Guenter Roeck <linux@roeck-us.net> wrote:
>>>
>>>> On Mon, Jun 13, 2022 at 11:13:21AM +0200, Greg KH wrote:
>>>>> On Fri, Jun 10, 2022 at 11:11:00AM -0400, Mike Snitzer wrote:
>>>>>> On Fri, Jun 10 2022 at  1:15P -0400,
>>>>>> Greg KH <gregkh@linuxfoundation.org> wrote:
>>>>>>
>>>>>>> On Fri, Jun 10, 2022 at 04:22:00AM +0000, Oleksandr Tymoshenko wrote:
>>>>>>>> I believe this commit introduced a regression in dm verity on systems
>>>>>>>> where data device is an NVME one. Loading table fails with the
>>>>>>>> following diagnostics:
>>>>>>>>
>>>>>>>> device-mapper: table: table load rejected: including non-request-stackable devices
>>>>>>>>
>>>>>>>> The same kernel works with the same data drive on the SCSI interface.
>>>>>>>> NVME-backed dm verity works with just this commit reverted.
>>>>>>>>
>>>>>>>> I believe the presence of the immutable partition is used as an indicator
>>>>>>>> of special case NVME configuration and if the data device's name starts
>>>>>>>> with "nvme" the code tries to switch the target type to
>>>>>>>> DM_TYPE_NVME_BIO_BASED (drivers/md/dm-table.c lines 1003-1010).
>>>>>>>>
>>>>>>>> The special NVME optimization case was removed in
>>>>>>>> 5.10 by commit 9c37de297f6590937f95a28bec1b7ac68a38618f, so only 5.4 is
>>>>>>>> affected.
>>>>>>>>
>>>>>>>
>>>>>>> Why wouldn't 4.9, 4.14, and 4.19 also be affected here?  Should I also
>>>>>>> just queue up 9c37de297f65 ("dm: remove special-casing of bio-based
>>>>>>> immutable singleton target on NVMe") to those older kernels?  If so,
>>>>>>> have you tested this and verified that it worked?
>>>>>>
>>>>>> Sorry for the unforeseen stable@ troubles here!
>>>>>>
>>>>>> In general we'd be fine to apply commit 9c37de297f65 but to do it
>>>>>> properly would require also making sure commits that remove
>>>>>> "DM_TYPE_NVME_BIO_BASED", like 8d47e65948dd ("dm mpath: remove
>>>>>> unnecessary NVMe branching in favor of scsi_dh checks") are applied --
>>>>>> basically any lingering references to DM_TYPE_NVME_BIO_BASED need to
>>>>>> be removed.
>>>>>>
>>>>>> The commit header for 8d47e65948dd documents what
>>>>>> DM_TYPE_NVME_BIO_BASED was used for.. it was dm-mpath specific and
>>>>>> "nvme" mode really never got used by any userspace that I'm aware of.
>>>>>>
>>>>>> Sadly I currently don't have the time to do this backport for all N
>>>>>> stable kernels... :(
>>>>>>
>>>>>> But if that backport gets out of control: A simpler, albeit stable@
>>>>>> unicorn, way to resolve this is to simply revert 9c37de297f65 and make
>>>>
>>>> 9c37de297f65 can not be reverted in 5.4 and older because it isn't there,
>>>> and trying to apply it results in conflicts which at least I can not
>>>> resolve.
>>>>
>>>>>> it so that DM-mpath and DM core just used bio-based if "nvme" is
>>>>>> requested by dm-mpath, so also in drivers/md/dm-mpath.c e.g.:
>>>>>>
>>>>>> @@ -1091,8 +1088,6 @@ static int parse_features(struct dm_arg_set *as, struct multipath *m)
>>>>>>
>>>>>>                           if (!strcasecmp(queue_mode_name, "bio"))
>>>>>>                                   m->queue_mode = DM_TYPE_BIO_BASED;
>>>>>> 			else if (!strcasecmp(queue_mode_name, "nvme"))
>>>>>> -                               m->queue_mode = DM_TYPE_NVME_BIO_BASED;
>>>>>> +                               m->queue_mode = DM_TYPE_BIO_BASED;
>>>>>>                           else if (!strcasecmp(queue_mode_name, "rq"))
>>>>>>                                   m->queue_mode = DM_TYPE_REQUEST_BASED;
>>>>>>                           else if (!strcasecmp(queue_mode_name, "mq"))
>>>>>>
>>>>>> Mike
>>>>>>
>>>>>
>>>>> Ok, please submit a working patch for the kernels that need it so that
>>>>> we can review and apply it to solve this regression.
>>>>>
>>>>
>>>> So, effectively, v5.4.y and older are broken right now for use cases
>>>> with dm on NVME drives.
>>>>
>>>> Given that the regression does affect older branches, and given that we
>>>> have to revert this patch to avoid regressions in ChromeOS, would it be
>>>> possible to revert it from v5.4.y and older until a fix is found ?
>>>
>>> I obviously would prefer to not have this false-start.
>>>
>> The false start has already happened since we had to revert the patch
>> from chromeos-5.4 and older branches.
> 
> OK, well this is pretty easy to fix in general.  If there are slight
> differences across older trees they are easily resolved.  Fact that
> stable@ couldn't cope with backporting 9c37de297f65 is.. what it is.
> 
> But this will fix the issue on 5.4.y:
> 
> From: Mike Snitzer <snitzer@kernel.org>
> Date: Wed, 15 Jun 2022 14:07:09 -0400
> Subject: [5.4.y PATCH] dm: remove special-casing of bio-based immutable singleton target on NVMe
> 
> Commit 9c37de297f6590937f95a28bec1b7ac68a38618f upstream.
> 
> There is no benefit to DM special-casing NVMe. Remove all code used to
> establish DM_TYPE_NVME_BIO_BASED.
> 
> Signed-off-by: Mike Snitzer <snitzer@kernel.org>

This patch passes our tests after I removed the unused variable.

Tested-by: Guenter Roeck <linux@roeck-us.net>

Thanks a lot for the backport!

Guenter

> ---
>   drivers/md/dm-table.c         | 32 ++----------------
>   drivers/md/dm.c               | 64 +++--------------------------------
>   include/linux/device-mapper.h |  1 -
>   3 files changed, 7 insertions(+), 90 deletions(-)
> 
> diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
> index 06b382304d92..81bc36a43b32 100644
> --- a/drivers/md/dm-table.c
> +++ b/drivers/md/dm-table.c
> @@ -872,8 +872,7 @@ EXPORT_SYMBOL(dm_consume_args);
>   static bool __table_type_bio_based(enum dm_queue_mode table_type)
>   {
>   	return (table_type == DM_TYPE_BIO_BASED ||
> -		table_type == DM_TYPE_DAX_BIO_BASED ||
> -		table_type == DM_TYPE_NVME_BIO_BASED);
> +		table_type == DM_TYPE_DAX_BIO_BASED);
>   }
>   
>   static bool __table_type_request_based(enum dm_queue_mode table_type)
> @@ -929,8 +928,6 @@ bool dm_table_supports_dax(struct dm_table *t,
>   	return true;
>   }
>   
> -static bool dm_table_does_not_support_partial_completion(struct dm_table *t);
> -
>   static int device_is_rq_stackable(struct dm_target *ti, struct dm_dev *dev,
>   				  sector_t start, sector_t len, void *data)
>   {
> @@ -960,7 +957,6 @@ static int dm_table_determine_type(struct dm_table *t)
>   			goto verify_bio_based;
>   		}
>   		BUG_ON(t->type == DM_TYPE_DAX_BIO_BASED);
> -		BUG_ON(t->type == DM_TYPE_NVME_BIO_BASED);
>   		goto verify_rq_based;
>   	}
>   
> @@ -999,15 +995,6 @@ static int dm_table_determine_type(struct dm_table *t)
>   		if (dm_table_supports_dax(t, device_not_dax_capable, &page_size) ||
>   		    (list_empty(devices) && live_md_type == DM_TYPE_DAX_BIO_BASED)) {
>   			t->type = DM_TYPE_DAX_BIO_BASED;
> -		} else {
> -			/* Check if upgrading to NVMe bio-based is valid or required */
> -			tgt = dm_table_get_immutable_target(t);
> -			if (tgt && !tgt->max_io_len && dm_table_does_not_support_partial_completion(t)) {
> -				t->type = DM_TYPE_NVME_BIO_BASED;
> -				goto verify_rq_based; /* must be stacked directly on NVMe (blk-mq) */
> -			} else if (list_empty(devices) && live_md_type == DM_TYPE_NVME_BIO_BASED) {
> -				t->type = DM_TYPE_NVME_BIO_BASED;
> -			}
>   		}
>   		return 0;
>   	}
> @@ -1024,8 +1011,7 @@ static int dm_table_determine_type(struct dm_table *t)
>   	 * (e.g. request completion process for partial completion.)
>   	 */
>   	if (t->num_targets > 1) {
> -		DMERR("%s DM doesn't support multiple targets",
> -		      t->type == DM_TYPE_NVME_BIO_BASED ? "nvme bio-based" : "request-based");
> +		DMERR("request-based DM doesn't support multiple targets");
>   		return -EINVAL;
>   	}
>   
> @@ -1714,20 +1700,6 @@ static int device_is_not_random(struct dm_target *ti, struct dm_dev *dev,
>   	return q && !blk_queue_add_random(q);
>   }
>   
> -static int device_is_partial_completion(struct dm_target *ti, struct dm_dev *dev,
> -					sector_t start, sector_t len, void *data)
> -{
> -	char b[BDEVNAME_SIZE];
> -
> -	/* For now, NVMe devices are the only devices of this class */
> -	return (strncmp(bdevname(dev->bdev, b), "nvme", 4) != 0);
> -}
> -
> -static bool dm_table_does_not_support_partial_completion(struct dm_table *t)
> -{
> -	return !dm_table_any_dev_attr(t, device_is_partial_completion, NULL);
> -}
> -
>   static int device_not_write_same_capable(struct dm_target *ti, struct dm_dev *dev,
>   					 sector_t start, sector_t len, void *data)
>   {
> diff --git a/drivers/md/dm.c b/drivers/md/dm.c
> index 37b8bb4d80f0..3c45c389ded9 100644
> --- a/drivers/md/dm.c
> +++ b/drivers/md/dm.c
> @@ -1000,7 +1000,7 @@ static void clone_endio(struct bio *bio)
>   	struct mapped_device *md = tio->io->md;
>   	dm_endio_fn endio = tio->ti->type->end_io;
>   
> -	if (unlikely(error == BLK_STS_TARGET) && md->type != DM_TYPE_NVME_BIO_BASED) {
> +	if (unlikely(error == BLK_STS_TARGET)) {
>   		if (bio_op(bio) == REQ_OP_DISCARD &&
>   		    !bio->bi_disk->queue->limits.max_discard_sectors)
>   			disable_discard(md);
> @@ -1340,10 +1340,7 @@ static blk_qc_t __map_bio(struct dm_target_io *tio)
>   		/* the bio has been remapped so dispatch it */
>   		trace_block_bio_remap(clone->bi_disk->queue, clone,
>   				      bio_dev(io->orig_bio), sector);
> -		if (md->type == DM_TYPE_NVME_BIO_BASED)
> -			ret = direct_make_request(clone);
> -		else
> -			ret = generic_make_request(clone);
> +		ret = generic_make_request(clone);
>   		break;
>   	case DM_MAPIO_KILL:
>   		if (unlikely(swap_bios_limit(ti, clone))) {
> @@ -1732,51 +1729,6 @@ static blk_qc_t __split_and_process_bio(struct mapped_device *md,
>   	return ret;
>   }
>   
> -/*
> - * Optimized variant of __split_and_process_bio that leverages the
> - * fact that targets that use it do _not_ have a need to split bios.
> - */
> -static blk_qc_t __process_bio(struct mapped_device *md, struct dm_table *map,
> -			      struct bio *bio, struct dm_target *ti)
> -{
> -	struct clone_info ci;
> -	blk_qc_t ret = BLK_QC_T_NONE;
> -	int error = 0;
> -
> -	init_clone_info(&ci, md, map, bio);
> -
> -	if (bio->bi_opf & REQ_PREFLUSH) {
> -		struct bio flush_bio;
> -
> -		/*
> -		 * Use an on-stack bio for this, it's safe since we don't
> -		 * need to reference it after submit. It's just used as
> -		 * the basis for the clone(s).
> -		 */
> -		bio_init(&flush_bio, NULL, 0);
> -		flush_bio.bi_opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC;
> -		ci.bio = &flush_bio;
> -		ci.sector_count = 0;
> -		error = __send_empty_flush(&ci);
> -		bio_uninit(ci.bio);
> -		/* dec_pending submits any data associated with flush */
> -	} else {
> -		struct dm_target_io *tio;
> -
> -		ci.bio = bio;
> -		ci.sector_count = bio_sectors(bio);
> -		if (__process_abnormal_io(&ci, ti, &error))
> -			goto out;
> -
> -		tio = alloc_tio(&ci, ti, 0, GFP_NOIO);
> -		ret = __clone_and_map_simple_bio(&ci, tio, NULL);
> -	}
> -out:
> -	/* drop the extra reference count */
> -	dec_pending(ci.io, errno_to_blk_status(error));
> -	return ret;
> -}
> -
>   static blk_qc_t dm_process_bio(struct mapped_device *md,
>   			       struct dm_table *map, struct bio *bio)
>   {
> @@ -1807,8 +1759,6 @@ static blk_qc_t dm_process_bio(struct mapped_device *md,
>   		/* regular IO is split by __split_and_process_bio */
>   	}
>   
> -	if (dm_get_md_type(md) == DM_TYPE_NVME_BIO_BASED)
> -		return __process_bio(md, map, bio, ti);
>   	return __split_and_process_bio(md, map, bio);
>   }
>   
> @@ -2200,12 +2150,10 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t,
>   	if (request_based)
>   		dm_stop_queue(q);
>   
> -	if (request_based || md->type == DM_TYPE_NVME_BIO_BASED) {
> +	if (request_based) {
>   		/*
> -		 * Leverage the fact that request-based DM targets and
> -		 * NVMe bio based targets are immutable singletons
> -		 * - used to optimize both dm_request_fn and dm_mq_queue_rq;
> -		 *   and __process_bio.
> +		 * Leverage the fact that request-based DM targets are
> +		 * immutable singletons - used to optimize dm_mq_queue_rq.
>   		 */
>   		md->immutable_target = dm_table_get_immutable_target(t);
>   	}
> @@ -2334,7 +2282,6 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t)
>   		break;
>   	case DM_TYPE_BIO_BASED:
>   	case DM_TYPE_DAX_BIO_BASED:
> -	case DM_TYPE_NVME_BIO_BASED:
>   		dm_init_congested_fn(md);
>   		break;
>   	case DM_TYPE_NONE:
> @@ -3070,7 +3017,6 @@ struct dm_md_mempools *dm_alloc_md_mempools(struct mapped_device *md, enum dm_qu
>   	switch (type) {
>   	case DM_TYPE_BIO_BASED:
>   	case DM_TYPE_DAX_BIO_BASED:
> -	case DM_TYPE_NVME_BIO_BASED:
>   		pool_size = max(dm_get_reserved_bio_based_ios(), min_pool_size);
>   		front_pad = roundup(per_io_data_size, __alignof__(struct dm_target_io)) + offsetof(struct dm_target_io, clone);
>   		io_front_pad = roundup(front_pad,  __alignof__(struct dm_io)) + offsetof(struct dm_io, tio);
> diff --git a/include/linux/device-mapper.h b/include/linux/device-mapper.h
> index a53d7d2c2d95..60631f3abddb 100644
> --- a/include/linux/device-mapper.h
> +++ b/include/linux/device-mapper.h
> @@ -28,7 +28,6 @@ enum dm_queue_mode {
>   	DM_TYPE_BIO_BASED	 = 1,
>   	DM_TYPE_REQUEST_BASED	 = 2,
>   	DM_TYPE_DAX_BIO_BASED	 = 3,
> -	DM_TYPE_NVME_BIO_BASED	 = 4,
>   };
>   
>   typedef enum { STATUSTYPE_INFO, STATUSTYPE_TABLE } status_type_t;


WARNING: multiple messages have this Message-ID (diff)
From: Guenter Roeck <linux@roeck-us.net>
To: Mike Snitzer <snitzer@redhat.com>
Cc: keescook@chromium.org, sarthakkukreti@google.com,
	Greg KH <gregkh@linuxfoundation.org>,
	Mike Snitzer <snitzer@kernel.org>,
	stable@vger.kernel.org, Oleksandr Tymoshenko <ovt@google.com>,
	dm-devel@redhat.com, regressions@lists.linux.dev
Subject: Re: [dm-devel] [PATCH 5.4 26/34] dm verity: set DM_TARGET_IMMUTABLE feature flag
Date: Thu, 16 Jun 2022 16:22:48 -0700	[thread overview]
Message-ID: <4dea87ca-b478-829e-2337-3d82725b6aec@roeck-us.net> (raw)
In-Reply-To: <Yqo63CvFpTDFnH3x@redhat.com>

On 6/15/22 13:02, Mike Snitzer wrote:
> On Wed, Jun 15 2022 at  1:50P -0400,
> Guenter Roeck <linux@roeck-us.net> wrote:
> 
>> On 6/15/22 08:29, Mike Snitzer wrote:
>>> On Wed, Jun 15 2022 at 10:36P -0400,
>>> Guenter Roeck <linux@roeck-us.net> wrote:
>>>
>>>> On Mon, Jun 13, 2022 at 11:13:21AM +0200, Greg KH wrote:
>>>>> On Fri, Jun 10, 2022 at 11:11:00AM -0400, Mike Snitzer wrote:
>>>>>> On Fri, Jun 10 2022 at  1:15P -0400,
>>>>>> Greg KH <gregkh@linuxfoundation.org> wrote:
>>>>>>
>>>>>>> On Fri, Jun 10, 2022 at 04:22:00AM +0000, Oleksandr Tymoshenko wrote:
>>>>>>>> I believe this commit introduced a regression in dm verity on systems
>>>>>>>> where data device is an NVME one. Loading table fails with the
>>>>>>>> following diagnostics:
>>>>>>>>
>>>>>>>> device-mapper: table: table load rejected: including non-request-stackable devices
>>>>>>>>
>>>>>>>> The same kernel works with the same data drive on the SCSI interface.
>>>>>>>> NVME-backed dm verity works with just this commit reverted.
>>>>>>>>
>>>>>>>> I believe the presence of the immutable partition is used as an indicator
>>>>>>>> of special case NVME configuration and if the data device's name starts
>>>>>>>> with "nvme" the code tries to switch the target type to
>>>>>>>> DM_TYPE_NVME_BIO_BASED (drivers/md/dm-table.c lines 1003-1010).
>>>>>>>>
>>>>>>>> The special NVME optimization case was removed in
>>>>>>>> 5.10 by commit 9c37de297f6590937f95a28bec1b7ac68a38618f, so only 5.4 is
>>>>>>>> affected.
>>>>>>>>
>>>>>>>
>>>>>>> Why wouldn't 4.9, 4.14, and 4.19 also be affected here?  Should I also
>>>>>>> just queue up 9c37de297f65 ("dm: remove special-casing of bio-based
>>>>>>> immutable singleton target on NVMe") to those older kernels?  If so,
>>>>>>> have you tested this and verified that it worked?
>>>>>>
>>>>>> Sorry for the unforeseen stable@ troubles here!
>>>>>>
>>>>>> In general we'd be fine to apply commit 9c37de297f65 but to do it
>>>>>> properly would require also making sure commits that remove
>>>>>> "DM_TYPE_NVME_BIO_BASED", like 8d47e65948dd ("dm mpath: remove
>>>>>> unnecessary NVMe branching in favor of scsi_dh checks") are applied --
>>>>>> basically any lingering references to DM_TYPE_NVME_BIO_BASED need to
>>>>>> be removed.
>>>>>>
>>>>>> The commit header for 8d47e65948dd documents what
>>>>>> DM_TYPE_NVME_BIO_BASED was used for.. it was dm-mpath specific and
>>>>>> "nvme" mode really never got used by any userspace that I'm aware of.
>>>>>>
>>>>>> Sadly I currently don't have the time to do this backport for all N
>>>>>> stable kernels... :(
>>>>>>
>>>>>> But if that backport gets out of control: A simpler, albeit stable@
>>>>>> unicorn, way to resolve this is to simply revert 9c37de297f65 and make
>>>>
>>>> 9c37de297f65 can not be reverted in 5.4 and older because it isn't there,
>>>> and trying to apply it results in conflicts which at least I can not
>>>> resolve.
>>>>
>>>>>> it so that DM-mpath and DM core just used bio-based if "nvme" is
>>>>>> requested by dm-mpath, so also in drivers/md/dm-mpath.c e.g.:
>>>>>>
>>>>>> @@ -1091,8 +1088,6 @@ static int parse_features(struct dm_arg_set *as, struct multipath *m)
>>>>>>
>>>>>>                           if (!strcasecmp(queue_mode_name, "bio"))
>>>>>>                                   m->queue_mode = DM_TYPE_BIO_BASED;
>>>>>> 			else if (!strcasecmp(queue_mode_name, "nvme"))
>>>>>> -                               m->queue_mode = DM_TYPE_NVME_BIO_BASED;
>>>>>> +                               m->queue_mode = DM_TYPE_BIO_BASED;
>>>>>>                           else if (!strcasecmp(queue_mode_name, "rq"))
>>>>>>                                   m->queue_mode = DM_TYPE_REQUEST_BASED;
>>>>>>                           else if (!strcasecmp(queue_mode_name, "mq"))
>>>>>>
>>>>>> Mike
>>>>>>
>>>>>
>>>>> Ok, please submit a working patch for the kernels that need it so that
>>>>> we can review and apply it to solve this regression.
>>>>>
>>>>
>>>> So, effectively, v5.4.y and older are broken right now for use cases
>>>> with dm on NVME drives.
>>>>
>>>> Given that the regression does affect older branches, and given that we
>>>> have to revert this patch to avoid regressions in ChromeOS, would it be
>>>> possible to revert it from v5.4.y and older until a fix is found ?
>>>
>>> I obviously would prefer to not have this false-start.
>>>
>> The false start has already happened since we had to revert the patch
>> from chromeos-5.4 and older branches.
> 
> OK, well this is pretty easy to fix in general.  If there are slight
> differences across older trees they are easily resolved.  Fact that
> stable@ couldn't cope with backporting 9c37de297f65 is.. what it is.
> 
> But this will fix the issue on 5.4.y:
> 
> From: Mike Snitzer <snitzer@kernel.org>
> Date: Wed, 15 Jun 2022 14:07:09 -0400
> Subject: [5.4.y PATCH] dm: remove special-casing of bio-based immutable singleton target on NVMe
> 
> Commit 9c37de297f6590937f95a28bec1b7ac68a38618f upstream.
> 
> There is no benefit to DM special-casing NVMe. Remove all code used to
> establish DM_TYPE_NVME_BIO_BASED.
> 
> Signed-off-by: Mike Snitzer <snitzer@kernel.org>

This patch passes our tests after I removed the unused variable.

Tested-by: Guenter Roeck <linux@roeck-us.net>

Thanks a lot for the backport!

Guenter

> ---
>   drivers/md/dm-table.c         | 32 ++----------------
>   drivers/md/dm.c               | 64 +++--------------------------------
>   include/linux/device-mapper.h |  1 -
>   3 files changed, 7 insertions(+), 90 deletions(-)
> 
> diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
> index 06b382304d92..81bc36a43b32 100644
> --- a/drivers/md/dm-table.c
> +++ b/drivers/md/dm-table.c
> @@ -872,8 +872,7 @@ EXPORT_SYMBOL(dm_consume_args);
>   static bool __table_type_bio_based(enum dm_queue_mode table_type)
>   {
>   	return (table_type == DM_TYPE_BIO_BASED ||
> -		table_type == DM_TYPE_DAX_BIO_BASED ||
> -		table_type == DM_TYPE_NVME_BIO_BASED);
> +		table_type == DM_TYPE_DAX_BIO_BASED);
>   }
>   
>   static bool __table_type_request_based(enum dm_queue_mode table_type)
> @@ -929,8 +928,6 @@ bool dm_table_supports_dax(struct dm_table *t,
>   	return true;
>   }
>   
> -static bool dm_table_does_not_support_partial_completion(struct dm_table *t);
> -
>   static int device_is_rq_stackable(struct dm_target *ti, struct dm_dev *dev,
>   				  sector_t start, sector_t len, void *data)
>   {
> @@ -960,7 +957,6 @@ static int dm_table_determine_type(struct dm_table *t)
>   			goto verify_bio_based;
>   		}
>   		BUG_ON(t->type == DM_TYPE_DAX_BIO_BASED);
> -		BUG_ON(t->type == DM_TYPE_NVME_BIO_BASED);
>   		goto verify_rq_based;
>   	}
>   
> @@ -999,15 +995,6 @@ static int dm_table_determine_type(struct dm_table *t)
>   		if (dm_table_supports_dax(t, device_not_dax_capable, &page_size) ||
>   		    (list_empty(devices) && live_md_type == DM_TYPE_DAX_BIO_BASED)) {
>   			t->type = DM_TYPE_DAX_BIO_BASED;
> -		} else {
> -			/* Check if upgrading to NVMe bio-based is valid or required */
> -			tgt = dm_table_get_immutable_target(t);
> -			if (tgt && !tgt->max_io_len && dm_table_does_not_support_partial_completion(t)) {
> -				t->type = DM_TYPE_NVME_BIO_BASED;
> -				goto verify_rq_based; /* must be stacked directly on NVMe (blk-mq) */
> -			} else if (list_empty(devices) && live_md_type == DM_TYPE_NVME_BIO_BASED) {
> -				t->type = DM_TYPE_NVME_BIO_BASED;
> -			}
>   		}
>   		return 0;
>   	}
> @@ -1024,8 +1011,7 @@ static int dm_table_determine_type(struct dm_table *t)
>   	 * (e.g. request completion process for partial completion.)
>   	 */
>   	if (t->num_targets > 1) {
> -		DMERR("%s DM doesn't support multiple targets",
> -		      t->type == DM_TYPE_NVME_BIO_BASED ? "nvme bio-based" : "request-based");
> +		DMERR("request-based DM doesn't support multiple targets");
>   		return -EINVAL;
>   	}
>   
> @@ -1714,20 +1700,6 @@ static int device_is_not_random(struct dm_target *ti, struct dm_dev *dev,
>   	return q && !blk_queue_add_random(q);
>   }
>   
> -static int device_is_partial_completion(struct dm_target *ti, struct dm_dev *dev,
> -					sector_t start, sector_t len, void *data)
> -{
> -	char b[BDEVNAME_SIZE];
> -
> -	/* For now, NVMe devices are the only devices of this class */
> -	return (strncmp(bdevname(dev->bdev, b), "nvme", 4) != 0);
> -}
> -
> -static bool dm_table_does_not_support_partial_completion(struct dm_table *t)
> -{
> -	return !dm_table_any_dev_attr(t, device_is_partial_completion, NULL);
> -}
> -
>   static int device_not_write_same_capable(struct dm_target *ti, struct dm_dev *dev,
>   					 sector_t start, sector_t len, void *data)
>   {
> diff --git a/drivers/md/dm.c b/drivers/md/dm.c
> index 37b8bb4d80f0..3c45c389ded9 100644
> --- a/drivers/md/dm.c
> +++ b/drivers/md/dm.c
> @@ -1000,7 +1000,7 @@ static void clone_endio(struct bio *bio)
>   	struct mapped_device *md = tio->io->md;
>   	dm_endio_fn endio = tio->ti->type->end_io;
>   
> -	if (unlikely(error == BLK_STS_TARGET) && md->type != DM_TYPE_NVME_BIO_BASED) {
> +	if (unlikely(error == BLK_STS_TARGET)) {
>   		if (bio_op(bio) == REQ_OP_DISCARD &&
>   		    !bio->bi_disk->queue->limits.max_discard_sectors)
>   			disable_discard(md);
> @@ -1340,10 +1340,7 @@ static blk_qc_t __map_bio(struct dm_target_io *tio)
>   		/* the bio has been remapped so dispatch it */
>   		trace_block_bio_remap(clone->bi_disk->queue, clone,
>   				      bio_dev(io->orig_bio), sector);
> -		if (md->type == DM_TYPE_NVME_BIO_BASED)
> -			ret = direct_make_request(clone);
> -		else
> -			ret = generic_make_request(clone);
> +		ret = generic_make_request(clone);
>   		break;
>   	case DM_MAPIO_KILL:
>   		if (unlikely(swap_bios_limit(ti, clone))) {
> @@ -1732,51 +1729,6 @@ static blk_qc_t __split_and_process_bio(struct mapped_device *md,
>   	return ret;
>   }
>   
> -/*
> - * Optimized variant of __split_and_process_bio that leverages the
> - * fact that targets that use it do _not_ have a need to split bios.
> - */
> -static blk_qc_t __process_bio(struct mapped_device *md, struct dm_table *map,
> -			      struct bio *bio, struct dm_target *ti)
> -{
> -	struct clone_info ci;
> -	blk_qc_t ret = BLK_QC_T_NONE;
> -	int error = 0;
> -
> -	init_clone_info(&ci, md, map, bio);
> -
> -	if (bio->bi_opf & REQ_PREFLUSH) {
> -		struct bio flush_bio;
> -
> -		/*
> -		 * Use an on-stack bio for this, it's safe since we don't
> -		 * need to reference it after submit. It's just used as
> -		 * the basis for the clone(s).
> -		 */
> -		bio_init(&flush_bio, NULL, 0);
> -		flush_bio.bi_opf = REQ_OP_WRITE | REQ_PREFLUSH | REQ_SYNC;
> -		ci.bio = &flush_bio;
> -		ci.sector_count = 0;
> -		error = __send_empty_flush(&ci);
> -		bio_uninit(ci.bio);
> -		/* dec_pending submits any data associated with flush */
> -	} else {
> -		struct dm_target_io *tio;
> -
> -		ci.bio = bio;
> -		ci.sector_count = bio_sectors(bio);
> -		if (__process_abnormal_io(&ci, ti, &error))
> -			goto out;
> -
> -		tio = alloc_tio(&ci, ti, 0, GFP_NOIO);
> -		ret = __clone_and_map_simple_bio(&ci, tio, NULL);
> -	}
> -out:
> -	/* drop the extra reference count */
> -	dec_pending(ci.io, errno_to_blk_status(error));
> -	return ret;
> -}
> -
>   static blk_qc_t dm_process_bio(struct mapped_device *md,
>   			       struct dm_table *map, struct bio *bio)
>   {
> @@ -1807,8 +1759,6 @@ static blk_qc_t dm_process_bio(struct mapped_device *md,
>   		/* regular IO is split by __split_and_process_bio */
>   	}
>   
> -	if (dm_get_md_type(md) == DM_TYPE_NVME_BIO_BASED)
> -		return __process_bio(md, map, bio, ti);
>   	return __split_and_process_bio(md, map, bio);
>   }
>   
> @@ -2200,12 +2150,10 @@ static struct dm_table *__bind(struct mapped_device *md, struct dm_table *t,
>   	if (request_based)
>   		dm_stop_queue(q);
>   
> -	if (request_based || md->type == DM_TYPE_NVME_BIO_BASED) {
> +	if (request_based) {
>   		/*
> -		 * Leverage the fact that request-based DM targets and
> -		 * NVMe bio based targets are immutable singletons
> -		 * - used to optimize both dm_request_fn and dm_mq_queue_rq;
> -		 *   and __process_bio.
> +		 * Leverage the fact that request-based DM targets are
> +		 * immutable singletons - used to optimize dm_mq_queue_rq.
>   		 */
>   		md->immutable_target = dm_table_get_immutable_target(t);
>   	}
> @@ -2334,7 +2282,6 @@ int dm_setup_md_queue(struct mapped_device *md, struct dm_table *t)
>   		break;
>   	case DM_TYPE_BIO_BASED:
>   	case DM_TYPE_DAX_BIO_BASED:
> -	case DM_TYPE_NVME_BIO_BASED:
>   		dm_init_congested_fn(md);
>   		break;
>   	case DM_TYPE_NONE:
> @@ -3070,7 +3017,6 @@ struct dm_md_mempools *dm_alloc_md_mempools(struct mapped_device *md, enum dm_qu
>   	switch (type) {
>   	case DM_TYPE_BIO_BASED:
>   	case DM_TYPE_DAX_BIO_BASED:
> -	case DM_TYPE_NVME_BIO_BASED:
>   		pool_size = max(dm_get_reserved_bio_based_ios(), min_pool_size);
>   		front_pad = roundup(per_io_data_size, __alignof__(struct dm_target_io)) + offsetof(struct dm_target_io, clone);
>   		io_front_pad = roundup(front_pad,  __alignof__(struct dm_io)) + offsetof(struct dm_io, tio);
> diff --git a/include/linux/device-mapper.h b/include/linux/device-mapper.h
> index a53d7d2c2d95..60631f3abddb 100644
> --- a/include/linux/device-mapper.h
> +++ b/include/linux/device-mapper.h
> @@ -28,7 +28,6 @@ enum dm_queue_mode {
>   	DM_TYPE_BIO_BASED	 = 1,
>   	DM_TYPE_REQUEST_BASED	 = 2,
>   	DM_TYPE_DAX_BIO_BASED	 = 3,
> -	DM_TYPE_NVME_BIO_BASED	 = 4,
>   };
>   
>   typedef enum { STATUSTYPE_INFO, STATUSTYPE_TABLE } status_type_t;

--
dm-devel mailing list
dm-devel@redhat.com
https://listman.redhat.com/mailman/listinfo/dm-devel


  parent reply	other threads:[~2022-06-16 23:22 UTC|newest]

Thread overview: 68+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-03 17:42 [PATCH 5.4 00/34] 5.4.197-rc1 review Greg Kroah-Hartman
2022-06-03 17:42 ` [PATCH 5.4 01/34] lockdown: also lock down previous kgdb use Greg Kroah-Hartman
2022-06-03 17:42 ` [PATCH 5.4 02/34] x86/pci/xen: Disable PCI/MSI[-X] masking for XEN_HVM guests Greg Kroah-Hartman
2022-06-03 17:42 ` [PATCH 5.4 03/34] staging: rtl8723bs: prevent ->Ssid overflow in rtw_wx_set_scan() Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 04/34] Input: goodix - fix spurious key release events Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 05/34] tcp: change source port randomizarion at connect() time Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 06/34] secure_seq: use the 64 bits of the siphash for port offset calculation Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 07/34] media: vim2m: Register video device after setting up internals Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 08/34] media: vim2m: initialize the media device earlier Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 09/34] ACPI: sysfs: Make sparse happy about address space in use Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 10/34] ACPI: sysfs: Fix BERT error region memory mapping Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 11/34] pinctrl: sunxi: fix f1c100s uart2 function Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 12/34] net: af_key: check encryption module availability consistency Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 13/34] net: ftgmac100: Disable hardware checksum on AST2600 Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 14/34] i2c: ismt: Provide a DMA buffer for Interrupt Cause Logging Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 15/34] drivers: i2c: thunderx: Allow driver to work with ACPI defined TWSI controllers Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 16/34] assoc_array: Fix BUG_ON during garbage collect Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 17/34] cfg80211: set custom regdomain after wiphy registration Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 18/34] drm/i915: Fix -Wstringop-overflow warning in call to intel_read_wm_latency() Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 19/34] exec: Force single empty string when argv is empty Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 20/34] netfilter: conntrack: re-fetch conntrack after insertion Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 21/34] crypto: ecrdsa - Fix incorrect use of vli_cmp Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 22/34] zsmalloc: fix races between asynchronous zspage free and page migration Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 23/34] dm integrity: fix error code in dm_integrity_ctr() Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 24/34] dm crypt: make printing of the key constant-time Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 25/34] dm stats: add cond_resched when looping over entries Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 26/34] dm verity: set DM_TARGET_IMMUTABLE feature flag Greg Kroah-Hartman
2022-06-10  4:22   ` Oleksandr Tymoshenko
2022-06-10  5:15     ` Greg KH
2022-06-10  8:10       ` Oleksandr Tymoshenko
2022-06-10 15:11       ` Mike Snitzer
2022-06-10 15:11         ` [dm-devel] " Mike Snitzer
2022-06-13  9:13         ` Greg KH
2022-06-13  9:13           ` [dm-devel] " Greg KH
2022-06-15 14:36           ` Guenter Roeck
2022-06-15 14:36             ` [dm-devel] " Guenter Roeck
2022-06-15 15:29             ` Mike Snitzer
2022-06-15 15:29               ` [dm-devel] " Mike Snitzer
2022-06-15 17:50               ` Guenter Roeck
2022-06-15 17:50                 ` [dm-devel] " Guenter Roeck
2022-06-15 20:02                 ` Mike Snitzer
2022-06-15 20:02                   ` [dm-devel] " Mike Snitzer
2022-06-15 20:40                   ` Guenter Roeck
2022-06-15 20:40                     ` [dm-devel] " Guenter Roeck
2022-06-15 23:59                   ` Guenter Roeck
2022-06-15 23:59                     ` [dm-devel] " Guenter Roeck
2022-06-16 23:22                   ` Guenter Roeck [this message]
2022-06-16 23:22                     ` Guenter Roeck
2022-06-20 11:44                   ` Greg KH
2022-06-20 11:44                     ` [dm-devel] " Greg KH
2022-06-21 16:35                     ` [5.4.y PATCH v2] dm: remove special-casing of bio-based immutable singleton target on NVMe Mike Snitzer
2022-06-21 16:35                       ` [dm-devel] " Mike Snitzer
2022-06-23 15:48                       ` Greg KH
2022-06-23 15:48                         ` [dm-devel] " Greg KH
2022-06-23 16:00                       ` Patch "dm: remove special-casing of bio-based immutable singleton target on NVMe" has been added to the 5.4-stable tree gregkh
2022-06-23 16:00                         ` [dm-devel] " gregkh
2022-06-03 17:43 ` [PATCH 5.4 27/34] raid5: introduce MD_BROKEN Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 28/34] HID: multitouch: Add support for Google Whiskers Touchpad Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 29/34] tpm: Fix buffer access in tpm2_get_tpm_pt() Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 30/34] tpm: ibmvtpm: Correct the return value in tpm_ibmvtpm_probe() Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 31/34] docs: submitting-patches: Fix crossref to The canonical patch format Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 32/34] NFS: Memory allocation failures are not server fatal errors Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 33/34] NFSD: Fix possible sleep during nfsd4_release_lockowner() Greg Kroah-Hartman
2022-06-03 17:43 ` [PATCH 5.4 34/34] bpf: Enlarge offset check value to INT_MAX in bpf_skb_{load,store}_bytes Greg Kroah-Hartman
2022-06-04 12:21 ` [PATCH 5.4 00/34] 5.4.197-rc1 review Sudip Mukherjee
2022-06-04 17:31 ` Naresh Kamboju
2022-06-04 18:54 ` Guenter Roeck
2022-06-06  1:08 ` Samuel Zou

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4dea87ca-b478-829e-2337-3d82725b6aec@roeck-us.net \
    --to=linux@roeck-us.net \
    --cc=dm-devel@redhat.com \
    --cc=gregkh@linuxfoundation.org \
    --cc=keescook@chromium.org \
    --cc=ovt@google.com \
    --cc=regressions@lists.linux.dev \
    --cc=sarthakkukreti@google.com \
    --cc=snitzer@kernel.org \
    --cc=snitzer@redhat.com \
    --cc=stable@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.