Re: [PATCH 2/2] block: add BLKSETDESCZONE ioctl for Zoned Block Devices

From: Javier GonzÃlez
Date: Tue Jul 07 2020 - 04:43:25 EST


On 03.07.2020 09:44, Matias Bjorling wrote:
-----Original Message-----
From: Javier GonzÃlez <javier@xxxxxxxxxxx>
Sent: Monday, 29 June 2020 21.39
To: Damien Le Moal <Damien.LeMoal@xxxxxxx>
Cc: Matias Bjorling <Matias.Bjorling@xxxxxxx>; axboe@xxxxxxxxx;
kbusch@xxxxxxxxxx; hch@xxxxxx; sagi@xxxxxxxxxxx;
martin.petersen@xxxxxxxxxx; Niklas Cassel <Niklas.Cassel@xxxxxxx>; Hans
Holmberg <Hans.Holmberg@xxxxxxx>; linux-scsi@xxxxxxxxxxxxxxx; linux-
kernel@xxxxxxxxxxxxxxx; linux-block@xxxxxxxxxxxxxxx; linux-
nvme@xxxxxxxxxxxxxxxxxxx
Subject: Re: [PATCH 2/2] block: add BLKSETDESCZONE ioctl for Zoned Block
Devices

On 29.06.2020 01:00, Damien Le Moal wrote:
>On 2020/06/29 8:01, Matias Bjorling wrote:
>> The NVMe Zoned Namespace Command Set adds support for associating
>> data to a zone through the Zone Descriptor Extension feature.
>>
>> To allow user-space to associate data to a zone, add support through
>> the BLKSETDESCZONE ioctl. The ioctl requires that it is issued to a
>> zoned block device, and that it supports the Zone Descriptor
>> Extension feature. Support is detected through the the
>> zone_desc_ext_bytes sysfs queue entry for the specific block device.
>> A value larger than zero communicates that the device supports the
>> feature.

Have you considered the possibility of adding this an action to a IOCTL that
looks like the zone management one we discussed last week? We would start
saving IOCTLs already if we count the offline transition and this one.

Yes, I considered it. Damien and Christoph have asked for it to separate ioctls.

Ok.



>>
>> The ioctl associates data to a zone by issuing a Zone Management Send
>> command with the Zone Send Action set as the Set Zone Descriptor
>> Extension.
>>
>> For the command to complete successfully, the specified zone must be
>> in the Empty state, and active resources must be available. On
>> success, the specified zone is transioned to Closed state by the
>> device. If less data is supplied by user-space then reported by the
>> the Zone Descriptor Extension size, the rest is zero-filled. If more
>> data or no data is supplied by user-space, the ioctl fails.
>>
>> To issue the ioctl, a new blk_zone_set_desc data structure is defined.
>> It has following parameters:
>>
>> * the sector of the specific zone.
>> * the length of the data to be associated to the zone.
>> * any flags be used by the ioctl. None is defined.
>> * data associated to the zone.
>>
>> The data is laid out after the flags parameter, and it is the
>> caller's responsibility to allocate memory for the data that is
>> specified in the length parameter.
>>
>> Signed-off-by: Matias BjÃrling <matias.bjorling@xxxxxxx>
>> ---
>> block/blk-zoned.c | 108 ++++++++++++++++++++++++++++++++++
>> block/ioctl.c | 2 +
>> drivers/nvme/host/core.c | 3 +
>> drivers/nvme/host/nvme.h | 9 +++
>> drivers/nvme/host/zns.c | 11 ++++
>> include/linux/blk_types.h | 2 +
>> include/linux/blkdev.h | 9 ++-
>> include/uapi/linux/blkzoned.h | 20 ++++++-
>> 8 files changed, 162 insertions(+), 2 deletions(-)
>>
>> diff --git a/block/blk-zoned.c b/block/blk-zoned.c index
>> 81152a260354..4dc40ec006a2 100644
>> --- a/block/blk-zoned.c
>> +++ b/block/blk-zoned.c
>> @@ -259,6 +259,50 @@ int blkdev_zone_mgmt(struct block_device *bdev,
>> enum req_opf op, } EXPORT_SYMBOL_GPL(blkdev_zone_mgmt);
>>
>> +/**
>> + * blkdev_zone_set_desc - Execute a zone management set zone
descriptor
>> + * extension operation on a zone
>> + * @bdev: Target block device
>> + * @sector: Start sector of the zone to operate on
>> + * @data: Pointer to the data that is to be associated to the zone
>> + * @gfp_mask: Memory allocation flags (for bio_alloc)
>> + *
>> + * Description:
>> + * Associate zone descriptor extension data to a specified zone.
>> + * The block device must support zone descriptor extensions.
>> + * i.e., by exposing a positive zone descriptor extension size.
>> + */
>> +int blkdev_zone_set_desc(struct block_device *bdev, sector_t sector,
>> + struct page *data, gfp_t gfp_mask)
>
>struct page for the data ? Why not just a "void *" to allow for
>kmalloc/vmalloc data ? And no length for the data ? This is a bit odd.
>
>> +{
>> + struct request_queue *q = bdev_get_queue(bdev);
>> + sector_t zone_sectors = blk_queue_zone_sectors(q);
>> + struct bio_vec bio_vec;
>> + struct bio bio;
>> +
>> + if (!blk_queue_is_zoned(q))
>> + return -EOPNOTSUPP;
>> +
>> + if (bdev_read_only(bdev))
>> + return -EPERM;
>
>You are not checking the zone_desc_ext_bytes limit here. You should.
>> +
>> + /* Check alignment (handle eventual smaller last zone) */
>> + if (sector & (zone_sectors - 1))
>> + return -EINVAL;
>
>The comment is incorrect. There is nothing special for handling the
>last zone in this test.
>
>> +
>> + bio_init(&bio, &bio_vec, 1);
>> + bio.bi_opf = REQ_OP_ZONE_SET_DESC | REQ_SYNC;
>> + bio.bi_iter.bi_sector = sector;
>> + bio_set_dev(&bio, bdev);
>> + bio_add_page(&bio, data, queue_zone_desc_ext_bytes(q), 0);
>> +
>> + /* This may take a while, so be nice to others */
>> + cond_resched();
>
>This is not a loop, so you do not need this.
>
>> +
>> + return submit_bio_wait(&bio);
>> +}
>> +EXPORT_SYMBOL_GPL(blkdev_zone_set_desc);
>> +
>> struct zone_report_args {
>> struct blk_zone __user *zones;
>> };
>> @@ -370,6 +414,70 @@ int blkdev_zone_mgmt_ioctl(struct block_device
*bdev, fmode_t mode,
>> GFP_KERNEL);
>> }
>>
>> +/*
>> + * BLKSETDESCZONE ioctl processing.
>> + * Called from blkdev_ioctl.
>> + */
>> +int blkdev_zone_set_desc_ioctl(struct block_device *bdev, fmode_t
mode,
>> + unsigned int cmd, unsigned long arg) {
>> + void __user *argp = (void __user *)arg;
>> + struct request_queue *q;
>> + struct blk_zone_set_desc zsd;
>> + void *zsd_data;
>> + int ret;
>> +
>> + if (!argp)
>> + return -EINVAL;
>> +
>> + q = bdev_get_queue(bdev);
>> + if (!q)
>> + return -ENXIO;
>> +
>> + if (!blk_queue_is_zoned(q))
>> + return -ENOTTY;
>> +
>> + if (!capable(CAP_SYS_ADMIN))
>> + return -EACCES;
>> +
>> + if (!(mode & FMODE_WRITE))
>> + return -EBADF;
>> +
>> + if (!queue_zone_desc_ext_bytes(q))
>> + return -EOPNOTSUPP;
>> +
>> + if (copy_from_user(&zsd, argp, sizeof(struct blk_zone_set_desc)))
>> + return -EFAULT;
>> +
>> + /* no flags is currently supported */
>> + if (zsd.flags)
>> + return -ENOTTY;
>> +
>> + if (!zsd.len || zsd.len > queue_zone_desc_ext_bytes(q))
>> + return -ENOTTY;
>
>This should go into blkdev_zone_set_desc() as well so that in-kernel
>users are checked. So there may be no need to check this here.
>
>> +
>> + /* allocate the size of the zone descriptor extension and fill
>> + * with the data in the user data buffer. If the data size is less
>> + * than the zone descriptor extension size, then the rest of the
>> + * zone description extension data buffer is zero-filled.
>> + */
>> + zsd_data = (void *) get_zeroed_page(GFP_KERNEL);
>> + if (!zsd_data)
>> + return -ENOMEM;
>> +
>> + if (copy_from_user(zsd_data, argp + sizeof(struct blk_zone_set_desc),
>> + zsd.len)) {
>> + ret = -EFAULT;
>> + goto free;
>> + }
>> +
>> + ret = blkdev_zone_set_desc(bdev, zsd.sector, virt_to_page(zsd_data),
>> + GFP_KERNEL);
>> +free:
>> + free_page((unsigned long) zsd_data);
>> + return ret;
>> +}
>> +
>> static inline unsigned long *blk_alloc_zone_bitmap(int node,
>> unsigned int nr_zones)
>> {
>> diff --git a/block/ioctl.c b/block/ioctl.c index
>> bdb3bbb253d9..b9744705835b 100644
>> --- a/block/ioctl.c
>> +++ b/block/ioctl.c
>> @@ -515,6 +515,8 @@ static int blkdev_common_ioctl(struct block_device
*bdev, fmode_t mode,
>> case BLKCLOSEZONE:
>> case BLKFINISHZONE:
>> return blkdev_zone_mgmt_ioctl(bdev, mode, cmd, arg);
>> + case BLKSETDESCZONE:
>> + return blkdev_zone_set_desc_ioctl(bdev, mode, cmd, arg);
>> case BLKGETZONESZ:
>> return put_uint(argp, bdev_zone_sectors(bdev));
>> case BLKGETNRZONES:
>> diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
>> index e961910da4ac..b8f25b0d00ad 100644
>> --- a/drivers/nvme/host/core.c
>> +++ b/drivers/nvme/host/core.c
>> @@ -776,6 +776,9 @@ blk_status_t nvme_setup_cmd(struct nvme_ns *ns,
struct request *req,
>> case REQ_OP_ZONE_FINISH:
>> ret = nvme_setup_zone_mgmt_send(ns, req, cmd,
NVME_ZONE_FINISH);
>> break;
>> + case REQ_OP_ZONE_SET_DESC:
>> + ret = nvme_setup_zone_set_desc(ns, req, cmd);
>> + break;
>> case REQ_OP_WRITE_ZEROES:
>> ret = nvme_setup_write_zeroes(ns, req, cmd);
>> break;
>> diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
>> index 662f95fbd909..5bd5a437b038 100644
>> --- a/drivers/nvme/host/nvme.h
>> +++ b/drivers/nvme/host/nvme.h
>> @@ -708,6 +708,9 @@ int nvme_report_zones(struct gendisk *disk,
>> sector_t sector, blk_status_t nvme_setup_zone_mgmt_send(struct
nvme_ns *ns, struct request *req,
>> struct nvme_command *cmnd,
>> enum nvme_zone_mgmt_action action);
>> +
>> +blk_status_t nvme_setup_zone_set_desc(struct nvme_ns *ns, struct
request *req,
>> + struct nvme_command *cmnd);
>> #else
>> #define nvme_report_zones NULL
>>
>> @@ -718,6 +721,12 @@ static inline blk_status_t
nvme_setup_zone_mgmt_send(struct nvme_ns *ns,
>> return BLK_STS_NOTSUPP;
>> }
>>
>> +static inline blk_status_t nvme_setup_zone_set_desc(struct nvme_ns *ns,
>> + struct request *req, struct nvme_command *cmnd) {
>> + return BLK_STS_NOTSUPP;
>> +}
>> +
>> static inline int nvme_update_zone_info(struct gendisk *disk,
>> struct nvme_ns *ns,
>> unsigned lbaf)
>> diff --git a/drivers/nvme/host/zns.c b/drivers/nvme/host/zns.c index
>> 5792d953a8f3..bfa64cc685d3 100644
>> --- a/drivers/nvme/host/zns.c
>> +++ b/drivers/nvme/host/zns.c
>> @@ -239,3 +239,14 @@ blk_status_t
nvme_setup_zone_mgmt_send(struct
>> nvme_ns *ns, struct request *req,
>>
>> return BLK_STS_OK;
>> }
>> +
>> +blk_status_t nvme_setup_zone_set_desc(struct nvme_ns *ns, struct
request *req,
>> + struct nvme_command *c)
>> +{
>> + c->zms.opcode = nvme_cmd_zone_mgmt_send;
>> + c->zms.nsid = cpu_to_le32(ns->head->ns_id);
>> + c->zms.slba = cpu_to_le64(nvme_sect_to_lba(ns, blk_rq_pos(req)));
>> + c->zms.action = NVME_ZONE_SET_DESC_EXT;
>> +
>> + return BLK_STS_OK;
>> +}
>> diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
>> index ccb895f911b1..53b7b05b0004 100644
>> --- a/include/linux/blk_types.h
>> +++ b/include/linux/blk_types.h
>> @@ -316,6 +316,8 @@ enum req_opf {
>> REQ_OP_ZONE_FINISH = 12,
>> /* write data at the current zone write pointer */
>> REQ_OP_ZONE_APPEND = 13,
>> + /* associate zone desc extension data to a zone */
>> + REQ_OP_ZONE_SET_DESC = 14,
>>
>> /* SCSI passthrough using struct scsi_request */
>> REQ_OP_SCSI_IN = 32,
>> diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index
>> 2ed55055f68d..c5f092dd5aa3 100644
>> --- a/include/linux/blkdev.h
>> +++ b/include/linux/blkdev.h
>> @@ -370,7 +370,8 @@ extern int blkdev_report_zones_ioctl(struct
block_device *bdev, fmode_t mode,
>> unsigned int cmd, unsigned long arg);
extern int
>> blkdev_zone_mgmt_ioctl(struct block_device *bdev, fmode_t mode,
>> unsigned int cmd, unsigned long arg);
>> -
>> +extern int blkdev_zone_set_desc_ioctl(struct block_device *bdev,
fmode_t mode,
>> + unsigned int cmd, unsigned long arg);
>> #else /* CONFIG_BLK_DEV_ZONED */
>>
>> static inline unsigned int blkdev_nr_zones(struct gendisk *disk) @@
>> -392,6 +393,12 @@ static inline int blkdev_zone_mgmt_ioctl(struct
block_device *bdev,
>> return -ENOTTY;
>> }
>>
>> +static inline int blkdev_zone_set_desc_ioctl(struct block_device *bdev,
>> + fmode_t mode, unsigned int cmd,
>> + unsigned long arg)
>> +{
>> + return -ENOTTY;
>> +}
>> #endif /* CONFIG_BLK_DEV_ZONED */
>>
>> struct request_queue {
>> diff --git a/include/uapi/linux/blkzoned.h
>> b/include/uapi/linux/blkzoned.h index 42c3366cc25f..68abda9abf33
>> 100644
>> --- a/include/uapi/linux/blkzoned.h
>> +++ b/include/uapi/linux/blkzoned.h
>> @@ -142,6 +142,20 @@ struct blk_zone_range {
>> __u64 nr_sectors;
>> };
>>
>> +/**
>> + * struct blk_zone_set_desc - BLKSETDESCZONE ioctl requests
>> + * @sector: Starting sector of the zone to operate on.
>> + * @flags: Feature flags.
>> + * @len: size, in bytes, of the data to be associated to the zone.
>> + * @data: data to be associated.
>> + */
>> +struct blk_zone_set_desc {
>> + __u64 sector;
>> + __u32 flags;
>> + __u32 len;
>> + __u8 data[0];
>> +};

Would it make sense to add nr_sectors if the host wants to associate the same
metadata to several zones. The use case would be the grouping of larger zones
in software.

I believe we get into atomicity concerns if we do ranges, and action
only supports a single zone. I like to align with the ZNS API as much
as possible, and let the user-space application track errors per set
desc ext action.

The atomicity concerns should be the same as current zone management
using nr_sectors, and these are already being supported for open, close,
etc.

TBH, the comment is most about making sure that the IOCTL is extendable
from conception.



>> +
>> /**
>> * Zoned block device ioctl's:
>> *
>> @@ -158,6 +172,10 @@ struct blk_zone_range {
>> * The 512 B sector range must be zone aligned.
>> * @BLKFINISHZONE: Mark the zones as full in the specified sector range.
>> * The 512 B sector range must be zone aligned.
>> + * @BLKSETDESCZONE: Set zone description extension data for the zone
>> + * in the specified sector. On success, the zone
>> + * will transition to the closed zone state.
>> + * The 512B sector must be zone aligned.
>> */
>> #define BLKREPORTZONE _IOWR(0x12, 130, struct blk_zone_report)
>> #define BLKRESETZONE _IOW(0x12, 131, struct blk_zone_range)
>> @@ -166,5 +184,5 @@ struct blk_zone_range {
>> #define BLKOPENZONE _IOW(0x12, 134, struct blk_zone_range)
>> #define BLKCLOSEZONE _IOW(0x12, 135, struct blk_zone_range)
>> #define BLKFINISHZONE _IOW(0x12, 136, struct blk_zone_range)
>> -
>> +#define BLKSETDESCZONE _IOW(0x12, 137, struct blk_zone_set_desc)
>> #endif /* _UAPI_BLKZONED_H */
>>
>
>How to you retreive an extended descriptor that was set ? I do not see
>any code doing that. Report zones is not changed, but I would leave
>that one as is and implement a BLKGETDESCZONE ioctl & in-kernel API.
>

In any case, this is needed. I also could not find a way to read the extended
descriptor back.

Thanks for the feedback.

Besides the users I have in mind, do you have users for which you need
the ability for user-space to access zone descriptor extensions data?
Is this a need to have or nice to have feature from your point of view?

At this moment most of the users we have in mind are user-space
applications that are zone-aware, so I would way this is necessary.

This said, I understand if you prioritize pushing the code that enables
in-kernel users and then re-iterate on this.

Thanks,
Javier