Re: [PATCH] block: ublk: enable zoned storage support

From: Niklas Cassel
Date: Fri Feb 24 2023 - 10:54:05 EST

Next message: Björn Töpel: "Re: [PATCH v8 1/3] riscv: Introduce CONFIG_RELOCATABLE"
Previous message: Peter Rosin: "Re: [PATCH 8/9] i2c: mux: Convert all drivers to new .probe() callback"
In reply to: Andreas Hindborg: "[PATCH] block: ublk: enable zoned storage support"
Next in thread: Niklas Cassel: "Re: [PATCH] block: ublk: enable zoned storage support"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

Hello Andreas,

On Fri, Feb 24, 2023 at 01:59:50PM +0100, Andreas Hindborg wrote:
> Add zoned storage support to ublk: report_zones and operations:
> - REQ_OP_ZONE_OPEN
> - REQ_OP_ZONE_CLOSE
> - REQ_OP_ZONE_FINISH
> - REQ_OP_ZONE_RESET
>
> This allows implementation of zoned storage devices in user space. An
> example user space implementation based on ubdsrv is available [1].
>
> [1] https://github.com/metaspace/ubdsrv/commit/14a2b708f74f70cfecb076d92e680dc718cc1f6d
>
> Signed-off-by: Andreas Hindborg <a.hindborg@xxxxxxxxxxx>
> ---

Did you try to build this patch with CONFIG_BLK_DEV_ZONED disabled?

I got the following build errors when building it on top of vanilla v6.2
when CONFIG_BLK_DEV_ZONED is disabled:

drivers/block/ublk_drv.c: In function ‘ublk_dev_param_basic_apply’:
drivers/block/ublk_drv.c:221:28: error: ‘struct gendisk’ has no member named ‘nr_zones’
221 | ub->ub_disk->nr_zones = p->dev_sectors / p->chunk_sectors;
| ^~
drivers/block/ublk_drv.c: In function ‘ublk_dev_param_zoned_apply’:
drivers/block/ublk_drv.c:244:17: error: implicit declaration of function ‘disk_set_max_active_zones’; did you mean ‘bdev_max_active_zones’? [-Werror=implicit-function-declaration]
244 | disk_set_max_active_zones(ub->ub_disk, p->max_active_zones);
| ^~~~~~~~~~~~~~~~~~~~~~~~~
| bdev_max_active_zones
drivers/block/ublk_drv.c:245:17: error: implicit declaration of function ‘disk_set_max_open_zones’; did you mean ‘bdev_max_open_zones’? [-Werror=implicit-function-declaration]
245 | disk_set_max_open_zones(ub->ub_disk, p->max_open_zones);

> drivers/block/ublk_drv.c | 145 ++++++++++++++++++++++++++++++++--
> include/uapi/linux/ublk_cmd.h | 18 +++++
> 2 files changed, 157 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/block/ublk_drv.c b/drivers/block/ublk_drv.c
> index 6368b56eacf1..9741b4b1647b 100644
> --- a/drivers/block/ublk_drv.c
> +++ b/drivers/block/ublk_drv.c
> @@ -20,6 +20,7 @@
> #include <linux/major.h>
> #include <linux/wait.h>
> #include <linux/blkdev.h>
> +#include <linux/blkzoned.h>
> #include <linux/init.h>
> #include <linux/swap.h>
> #include <linux/slab.h>
> @@ -51,10 +52,12 @@
> | UBLK_F_URING_CMD_COMP_IN_TASK \
> | UBLK_F_NEED_GET_DATA \
> | UBLK_F_USER_RECOVERY \
> - | UBLK_F_USER_RECOVERY_REISSUE)
> + | UBLK_F_USER_RECOVERY_REISSUE \
> + | UBLK_F_ZONED)
>
> /* All UBLK_PARAM_TYPE_* should be included here */
> -#define UBLK_PARAM_TYPE_ALL (UBLK_PARAM_TYPE_BASIC | UBLK_PARAM_TYPE_DISCARD)
> +#define UBLK_PARAM_TYPE_ALL (UBLK_PARAM_TYPE_BASIC | UBLK_PARAM_TYPE_DISCARD \
> + | UBLK_PARAM_TYPE_ZONED)
>
> struct ublk_rq_data {
> struct llist_node node;
> @@ -212,6 +215,11 @@ static void ublk_dev_param_basic_apply(struct ublk_device *ub)
> set_disk_ro(ub->ub_disk, true);
>
> set_capacity(ub->ub_disk, p->dev_sectors);
> +
> + if (IS_ENABLED(CONFIG_BLK_DEV_ZONED) &&
> + ub->dev_info.flags & UBLK_F_ZONED && p->chunk_sectors) {
> + ub->ub_disk->nr_zones = p->dev_sectors / p->chunk_sectors;
> + }

Kernel coding style says to not unnecessarily use braces where a single
statement will do:
https://www.kernel.org/doc/html/latest/process/coding-style.html#placing-braces-and-spaces

> }
>
> static void ublk_dev_param_discard_apply(struct ublk_device *ub)
> @@ -227,6 +235,19 @@ static void ublk_dev_param_discard_apply(struct ublk_device *ub)
> blk_queue_max_discard_segments(q, p->max_discard_segments);
> }
>
> +static void ublk_dev_param_zoned_apply(struct ublk_device *ub)
> +{
> + const struct ublk_param_zoned *p = &ub->params.zoned;
> +
> + if (IS_ENABLED(CONFIG_BLK_DEV_ZONED) &&
> + ub->dev_info.flags & UBLK_F_ZONED) {
> + disk_set_max_active_zones(ub->ub_disk, p->max_active_zones);
> + disk_set_max_open_zones(ub->ub_disk, p->max_open_zones);
> + /* We do not support zone append yet */
> + //blk_queue_max_zone_append_sectors(q, zone_size);

Remove commented out code.
It can be added once that feature is added.

> + }
> +}
> +
> static int ublk_validate_params(const struct ublk_device *ub)
> {
> /* basic param is the only one which must be set */
> @@ -268,6 +289,9 @@ static int ublk_apply_params(struct ublk_device *ub)
> if (ub->params.types & UBLK_PARAM_TYPE_DISCARD)
> ublk_dev_param_discard_apply(ub);
>
> + if (ub->params.types & UBLK_PARAM_TYPE_ZONED)
> + ublk_dev_param_zoned_apply(ub);
> +
> return 0;
> }
>
> @@ -361,9 +385,18 @@ static void ublk_free_disk(struct gendisk *disk)
> put_device(&ub->cdev_dev);
> }
>
> +#if IS_ENABLED(CONFIG_BLK_DEV_ZONED)

Here you use "#if IS_ENABLED(CONFIG_BLK_DEV_ZONED)"
yet further down you use
"#ifdef CONFIG_BLK_DEV_ZONED".

IS_ENABLED() is usually used when testing for a Kconfig within
a normal C conditional, e.g. if (IS_ENABLED()), where possible,
in order to avoid ugly C preprocessor conditionals completely.

However, if you have to use C preprocessor conditionals, then
#ifdef CONFIG_SOMETHING
is the most common way to conditionally include something.

> +static int ublk_report_zones(struct gendisk *disk, sector_t sector,
> + unsigned int nr_zones, report_zones_cb cb,
> + void *data);
> +#endif
> +
> static const struct block_device_operations ub_fops = {
> - .owner = THIS_MODULE,
> - .free_disk = ublk_free_disk,
> + .owner = THIS_MODULE,
> + .free_disk = ublk_free_disk,
> +#if IS_ENABLED(CONFIG_BLK_DEV_ZONED)
> + .report_zones = ublk_report_zones,
> +#endif
> };
>
> #define UBLK_MAX_PIN_PAGES 32
> @@ -499,7 +532,7 @@ static int ublk_unmap_io(const struct ublk_queue *ubq,
> {
> const unsigned int rq_bytes = blk_rq_bytes(req);
>
> - if (req_op(req) == REQ_OP_READ && ublk_rq_has_data(req)) {
> + if ((req_op(req) == REQ_OP_READ || req_op(req) == REQ_OP_DRV_IN) && ublk_rq_has_data(req)) {
> struct ublk_map_data data = {
> .ubq = ubq,
> .rq = req,
> @@ -566,6 +599,26 @@ static blk_status_t ublk_setup_iod(struct ublk_queue *ubq, struct request *req)
> case REQ_OP_WRITE_ZEROES:
> ublk_op = UBLK_IO_OP_WRITE_ZEROES;
> break;
> +#ifdef CONFIG_BLK_DEV_ZONED
> + case REQ_OP_ZONE_OPEN:
> + ublk_op = UBLK_IO_OP_ZONE_OPEN;
> + break;
> + case REQ_OP_ZONE_CLOSE:
> + ublk_op = UBLK_IO_OP_ZONE_CLOSE;
> + break;
> + case REQ_OP_ZONE_FINISH:
> + ublk_op = UBLK_IO_OP_ZONE_FINISH;
> + break;
> + case REQ_OP_ZONE_RESET:
> + ublk_op = UBLK_IO_OP_ZONE_RESET;
> + break;
> + case REQ_OP_DRV_IN:
> + ublk_op = UBLK_IO_OP_DRV_IN;
> + break;
> + case REQ_OP_ZONE_APPEND:
> + /* We do not support zone append yet */
> + fallthrough;
> +#endif
> default:
> return BLK_STS_IOERR;
> }
> @@ -612,7 +665,8 @@ static void ublk_complete_rq(struct request *req)
> *
> * Both the two needn't unmap.
> */
> - if (req_op(req) != REQ_OP_READ && req_op(req) != REQ_OP_WRITE) {
> + if (req_op(req) != REQ_OP_READ && req_op(req) != REQ_OP_WRITE &&
> + req_op(req) != REQ_OP_DRV_IN) {
> blk_mq_end_request(req, BLK_STS_OK);
> return;
> }
> @@ -1493,6 +1547,73 @@ static struct ublk_device *ublk_get_device_from_id(int idx)
> return ub;
> }
>
> +#ifdef CONFIG_BLK_DEV_ZONED
> +static int ublk_report_zones(struct gendisk *disk, sector_t sector,
> + unsigned int nr_zones, report_zones_cb cb,
> + void *data)
> +{
> + struct ublk_device *ub;
> + unsigned int zone_size;
> + unsigned int first_zone;
> + int ret = 0;
> +
> + ub = disk->private_data;
> +
> + if (!(ub->dev_info.flags & UBLK_F_ZONED))
> + return -EINVAL;
> +
> + zone_size = disk->queue->limits.chunk_sectors;
> + first_zone = sector >> ilog2(zone_size);
> +
> + nr_zones = min(ub->ub_disk->nr_zones - first_zone, nr_zones);
> +
> + for (unsigned int i = 0; i < nr_zones; i++) {
> +

No need for empty line here.

> + struct request *req;
> + blk_status_t status;
> + struct blk_zone info;

Since this is allocated on the stack, does it need memset()?

null_blk seems to memset blk_zone before the loop, but not inside the loop:
https://github.com/torvalds/linux/blob/v6.2/drivers/block/null_blk/zoned.c#L206

So I'm a bit confused...

If e.g. cb() is blkdev_copy_zone_to_user()
https://github.com/torvalds/linux/blob/v6.2/block/blk-zoned.c#L318

It will copy whatever the kernel allocated.
So I guess that since you can't guarantee that the ublk user space driver
will fill in all fields in the zone when receiving a REQ_OP_DRV_IN,
I assume that you do need to do a memset() in order to not potentially
leak kernel space data.

> +
> + req = blk_mq_alloc_request(disk->queue, REQ_OP_DRV_IN, 0);
> +
> + if (IS_ERR(req)) {
> + ret = PTR_ERR(req);
> + goto out;
> + }
> +
> + req->__sector = sector;
> +
> + ret = blk_rq_map_kern(disk->queue, req, &info, sizeof(info),
> + GFP_KERNEL);
> +
> + if (ret)
> + goto out;
> +
> + status = blk_execute_rq(req, 0);
> + ret = blk_status_to_errno(status);
> + if (ret)
> + goto out;
> +
> + blk_mq_free_request(req);
> +
> + ret = cb(&info, i, data);
> + if (ret)
> + goto out;
> +
> + /* A zero length zone means don't ask for more zones */
> + if (!info.len) {
> + nr_zones = i;
> + break;
> + }
> +
> + sector += zone_size;
> + }
> + ret = nr_zones;
> +
> + out:
> + return ret;
> +}
> +#endif
> +
> static int ublk_ctrl_start_dev(struct io_uring_cmd *cmd)
> {
> struct ublksrv_ctrl_cmd *header = (struct ublksrv_ctrl_cmd *)cmd->cmd;
> @@ -1535,6 +1656,15 @@ static int ublk_ctrl_start_dev(struct io_uring_cmd *cmd)
> if (ret)
> goto out_put_disk;
>
> + if (IS_ENABLED(CONFIG_BLK_DEV_ZONED) &&
> + ub->dev_info.flags & UBLK_F_ZONED) {
> + disk_set_zoned(disk, BLK_ZONED_HM);
> + blk_queue_required_elevator_features(disk->queue, ELEVATOR_F_ZBD_SEQ_WRITE);
> + ret = blk_revalidate_disk_zones(disk, NULL);
> + if (ret)
> + goto out_put_disk;
> + }
> +
> get_device(&ub->cdev_dev);
> ret = add_disk(disk);
> if (ret) {
> @@ -1673,6 +1803,9 @@ static int ublk_ctrl_add_dev(struct io_uring_cmd *cmd)
> if (!IS_BUILTIN(CONFIG_BLK_DEV_UBLK))
> ub->dev_info.flags |= UBLK_F_URING_CMD_COMP_IN_TASK;
>
> + if (!IS_ENABLED(CONFIG_BLK_DEV_ZONED))
> + ub->dev_info.flags &= ~UBLK_F_ZONED;
> +
> /* We are not ready to support zero copy */
> ub->dev_info.flags &= ~UBLK_F_SUPPORT_ZERO_COPY;
>
> diff --git a/include/uapi/linux/ublk_cmd.h b/include/uapi/linux/ublk_cmd.h
> index 8f88e3a29998..074b97821575 100644
> --- a/include/uapi/linux/ublk_cmd.h
> +++ b/include/uapi/linux/ublk_cmd.h
> @@ -78,6 +78,10 @@
> #define UBLK_F_USER_RECOVERY (1UL << 3)
>
> #define UBLK_F_USER_RECOVERY_REISSUE (1UL << 4)
> +/*
> + * Enable zoned device support
> + */
> +#define UBLK_F_ZONED (1ULL << 5)
>
> /* device state */
> #define UBLK_S_DEV_DEAD 0
> @@ -129,6 +133,12 @@ struct ublksrv_ctrl_dev_info {
> #define UBLK_IO_OP_DISCARD 3
> #define UBLK_IO_OP_WRITE_SAME 4
> #define UBLK_IO_OP_WRITE_ZEROES 5
> +#define UBLK_IO_OP_ZONE_OPEN 10
> +#define UBLK_IO_OP_ZONE_CLOSE 11
> +#define UBLK_IO_OP_ZONE_FINISH 12
> +#define UBLK_IO_OP_ZONE_APPEND 13
> +#define UBLK_IO_OP_ZONE_RESET 15
> +#define UBLK_IO_OP_DRV_IN 34
>
> #define UBLK_IO_F_FAILFAST_DEV (1U << 8)
> #define UBLK_IO_F_FAILFAST_TRANSPORT (1U << 9)
> @@ -214,6 +224,12 @@ struct ublk_param_discard {
> __u16 reserved0;
> };
>
> +struct ublk_param_zoned {
> + __u64 max_open_zones;
> + __u64 max_active_zones;
> + __u64 max_append_size;
> +};
> +
> struct ublk_params {
> /*
> * Total length of parameters, userspace has to set 'len' for both
> @@ -224,10 +240,12 @@ struct ublk_params {
> __u32 len;
> #define UBLK_PARAM_TYPE_BASIC (1 << 0)
> #define UBLK_PARAM_TYPE_DISCARD (1 << 1)
> +#define UBLK_PARAM_TYPE_ZONED (1 << 2)
> __u32 types; /* types of parameter included */
>
> struct ublk_param_basic basic;
> struct ublk_param_discard discard;
> + struct ublk_param_zoned zoned;
> };
>
> #endif
> --
> 2.39.2
>

Next message: Björn Töpel: "Re: [PATCH v8 1/3] riscv: Introduce CONFIG_RELOCATABLE"
Previous message: Peter Rosin: "Re: [PATCH 8/9] i2c: mux: Convert all drivers to new .probe() callback"
In reply to: Andreas Hindborg: "[PATCH] block: ublk: enable zoned storage support"
Next in thread: Niklas Cassel: "Re: [PATCH] block: ublk: enable zoned storage support"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]