Skip to contentSkip to conversation

virtio_blk: add VIRTIO_BLK_F_LIFETIME feature support

PATCH Superseded Alvaro Karsz 8 participants 26 messages Dec 5, 2022 – Dec 19, 2022

Subsystems (6)

v3 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH v3 1/1 virtio_blk: add VIRTIO_BLK_F_LIFETIME feature support

Implement the VIRTIO_BLK_F_LIFETIME feature for VirtIO block devices.

This commit introduces a new ioctl command, VBLK_LIFETIME.

VBLK_LIFETIME ioctl asks for the block device to provide lifetime information by sending a VIRTIO_BLK_T_GET_LIFETIME command to the device.

lifetime information fields:

- pre_eol_info: specifies the percentage of reserved blocks that are consumed.
optional values following virtio spec:
*) 0 - undefined.
*) 1 - normal, < 80% of reserved blocks are consumed. *) 2 - warning, 80% of reserved blocks are consumed. *) 3 - urgent, 90% of reserved blocks are consumed.

- device_lifetime_est_typ_a: this field refers to wear of SLC cells and is provided in increments of 10used, and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11 are reserved.

- device_lifetime_est_typ_b: this field refers to wear of MLC cells and is provided with the same semantics as device_lifetime_est_typ_a.

The data received from the device will be sent as is to the user. No data check/decode is done by virtblk.

--
v2:
- Remove (void *) casting.
- Fix comments format in virtio_blk.h.
- Set ioprio value for legacy devices for REQ_OP_DRV_IN operations.

v3:
- Initialize struct virtio_blk_lifetime to 0 instead of memset.
- Rename ioctl from VBLK_LIFETIME to VBLK_GET_LIFETIME.
- Return EOPNOTSUPP insted of ENOTTY if ioctl is not supported.
- Check if process is CAP_SYS_ADMIN capable in lifetime ioctl.
- Check if vdev is not NULL before accessing it in lifetime ioctl.
--

---
 drivers/block/virtio_blk.c      | 106 ++++++++++++++++++++++++++++++--
 include/uapi/linux/virtio_blk.h |  32 ++++++++++
 2 files changed, 133 insertions(+), 5 deletions(-)
drivers/block/virtio_blk.c +101 -5
index 19da5defd73..392d57fd6b7 100644
@@ -101,6 +101,18 @@ static inline blk_status_t virtblk_result(struct virtblk_req *vbr)
 	}
 }
 
+static inline int virtblk_ioctl_result(struct virtblk_req *vbr)
+{
+	switch (vbr->status) {
+	case VIRTIO_BLK_S_OK:
+		return 0;
+	case VIRTIO_BLK_S_UNSUPP:
+		return -EOPNOTSUPP;
+	default:
+		return -EIO;
+	}
+}
+
 static inline struct virtio_blk_vq *get_virtio_blk_vq(struct blk_mq_hw_ctx *hctx)
 {
 	struct virtio_blk *vblk = hctx->queue->queuedata;
@@ -218,6 +230,7 @@ static blk_status_t virtblk_setup_cmd(struct virtio_device *vdev,
 	u32 type;
 
 	vbr->out_hdr.sector = 0;
+	vbr->out_hdr.ioprio = cpu_to_virtio32(vdev, req_get_ioprio(req));
 
 	switch (req_op(req)) {
 	case REQ_OP_READ:
@@ -244,15 +257,14 @@ static blk_status_t virtblk_setup_cmd(struct virtio_device *vdev,
 		type = VIRTIO_BLK_T_SECURE_ERASE;
 		break;
 	case REQ_OP_DRV_IN:
-		type = VIRTIO_BLK_T_GET_ID;
-		break;
+		/* type is set in virtblk_get_id/virtblk_ioctl_lifetime */
+		return 0;
 	default:
 		WARN_ON_ONCE(1);
 		return BLK_STS_IOERR;
 	}
 
 	vbr->out_hdr.type = cpu_to_virtio32(vdev, type);
-	vbr->out_hdr.ioprio = cpu_to_virtio32(vdev, req_get_ioprio(req));
 
 	if (type == VIRTIO_BLK_T_DISCARD || type == VIRTIO_BLK_T_WRITE_ZEROES ||
 	    type == VIRTIO_BLK_T_SECURE_ERASE) {
@@ -459,12 +471,16 @@ static int virtblk_get_id(struct gendisk *disk, char *id_str)
 	struct virtio_blk *vblk = disk->private_data;
 	struct request_queue *q = vblk->disk->queue;
 	struct request *req;
+	struct virtblk_req *vbr;
 	int err;
 
 	req = blk_mq_alloc_request(q, REQ_OP_DRV_IN, 0);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
+	vbr = blk_mq_rq_to_pdu(req);
+	vbr->out_hdr.type = cpu_to_virtio32(vblk->vdev, VIRTIO_BLK_T_GET_ID);
+
 	err = blk_rq_map_kern(q, req, id_str, VIRTIO_BLK_ID_BYTES, GFP_KERNEL);
 	if (err)
 		goto out;
@@ -508,6 +524,85 @@ static int virtblk_getgeo(struct block_device *bd, struct hd_geometry *geo)
 	return ret;
 }
 
+/* Get lifetime information from device */
+static int virtblk_ioctl_lifetime(struct virtio_blk *vblk, unsigned long arg)
+{
+	struct request_queue *q = vblk->disk->queue;
+	struct request *req = NULL;
+	struct virtblk_req *vbr;
+	struct virtio_blk_lifetime lifetime = {};
+	int ret;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	/* The virtio_blk_lifetime struct fields follow virtio spec.
+	 * There is no check/decode on values received from the device.
+	 * The data is sent as is to the user.
+	 */
+
+	/* This ioctl is allowed only if VIRTIO_BLK_F_LIFETIME
+	 * feature is negotiated.
+	 */
+	if (!virtio_has_feature(vblk->vdev, VIRTIO_BLK_F_LIFETIME))
+		return -EOPNOTSUPP;
+
+	req = blk_mq_alloc_request(q, REQ_OP_DRV_IN, 0);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	/* Write the correct type */
+	vbr = blk_mq_rq_to_pdu(req);
+	vbr->out_hdr.type = cpu_to_virtio32(vblk->vdev, VIRTIO_BLK_T_GET_LIFETIME);
+
+	ret = blk_rq_map_kern(q, req, &lifetime, sizeof(lifetime), GFP_KERNEL);
+	if (ret)
+		goto out;
+
+	blk_execute_rq(req, false);
+
+	ret = virtblk_ioctl_result(blk_mq_rq_to_pdu(req));
+	if (ret)
+		goto out;
+
+	/* Pass the data to the user */
+	if (copy_to_user((void __user *)arg, &lifetime, sizeof(lifetime))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+out:
+	blk_mq_free_request(req);
+	return ret;
+}
+
+static int virtblk_ioctl(struct block_device *bd, fmode_t mode,
+			 unsigned int cmd, unsigned long arg)
+{
+	struct virtio_blk *vblk = bd->bd_disk->private_data;
+	int ret;
+
+	mutex_lock(&vblk->vdev_mutex);
+
+	if (!vblk->vdev) {
+		ret = -ENXIO;
+		goto exit;
+	}
+
+	switch (cmd) {
+	case VBLK_GET_LIFETIME:
+		ret = virtblk_ioctl_lifetime(vblk, arg);
+		break;
+	default:
+		ret = -EOPNOTSUPP;
+		break;
+	}
+
+exit:
+	mutex_unlock(&vblk->vdev_mutex);
+	return ret;
+}
+
 static void virtblk_free_disk(struct gendisk *disk)
 {
 	struct virtio_blk *vblk = disk->private_data;
@@ -520,6 +615,7 @@ static void virtblk_free_disk(struct gendisk *disk)
 static const struct block_device_operations virtblk_fops = {
 	.owner  	= THIS_MODULE,
 	.getgeo		= virtblk_getgeo,
+	.ioctl		= virtblk_ioctl,
 	.free_disk	= virtblk_free_disk,
 };
 
@@ -1239,7 +1335,7 @@ static unsigned int features_legacy[] = {
 	VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
 	VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY, VIRTIO_BLK_F_CONFIG_WCE,
 	VIRTIO_BLK_F_MQ, VIRTIO_BLK_F_DISCARD, VIRTIO_BLK_F_WRITE_ZEROES,
-	VIRTIO_BLK_F_SECURE_ERASE,
+	VIRTIO_BLK_F_SECURE_ERASE, VIRTIO_BLK_F_LIFETIME,
 }
 ;
 static unsigned int features[] = {
@@ -1247,7 +1343,7 @@ static unsigned int features[] = {
 	VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
 	VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY, VIRTIO_BLK_F_CONFIG_WCE,
 	VIRTIO_BLK_F_MQ, VIRTIO_BLK_F_DISCARD, VIRTIO_BLK_F_WRITE_ZEROES,
-	VIRTIO_BLK_F_SECURE_ERASE,
+	VIRTIO_BLK_F_SECURE_ERASE, VIRTIO_BLK_F_LIFETIME,
 };
 
 static struct virtio_driver virtio_blk = {
include/uapi/linux/virtio_blk.h +32 -0
index 58e70b24b50..e755d62d2ea 100644
@@ -40,6 +40,7 @@
 #define VIRTIO_BLK_F_MQ		12	/* support more than one vq */
 #define VIRTIO_BLK_F_DISCARD	13	/* DISCARD is supported */
 #define VIRTIO_BLK_F_WRITE_ZEROES	14	/* WRITE ZEROES is supported */
+#define VIRTIO_BLK_F_LIFETIME	15 /* Storage lifetime information is supported */
 #define VIRTIO_BLK_F_SECURE_ERASE	16 /* Secure Erase is supported */
 
 /* Legacy feature bits */
@@ -165,6 +166,9 @@ struct virtio_blk_config {
 /* Get device ID command */
 #define VIRTIO_BLK_T_GET_ID    8
 
+/* Get lifetime information command */
+#define VIRTIO_BLK_T_GET_LIFETIME 10
+
 /* Discard command */
 #define VIRTIO_BLK_T_DISCARD	11
 
@@ -206,6 +210,30 @@ struct virtio_blk_discard_write_zeroes {
 	__le32 flags;
 };
 
+/* Get lifetime information struct for each request */
+struct virtio_blk_lifetime {
+	/*
+	 * specifies the percentage of reserved blocks that are consumed.
+	 * optional values following virtio spec:
+	 * 0 - undefined
+	 * 1 - normal, < 80% of reserved blocks are consumed
+	 * 2 - warning, 80% of reserved blocks are consumed
+	 * 3 - urgent, 90% of reserved blocks are consumed
+	 */
+	__le16 pre_eol_info;
+	/*
+	 * this field refers to wear of SLC cells and is provided in increments of 10used,
+	 * and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11
+	 * are reserved
+	 */
+	__le16 device_lifetime_est_typ_a;
+	/*
+	 * this field refers to wear of MLC cells and is provided with the same semantics as
+	 * device_lifetime_est_typ_a
+	 */
+	__le16 device_lifetime_est_typ_b;
+};
+
 #ifndef VIRTIO_BLK_NO_LEGACY
 struct virtio_scsi_inhdr {
 	__virtio32 errors;
@@ -219,4 +247,8 @@ struct virtio_scsi_inhdr {
 #define VIRTIO_BLK_S_OK		0
 #define VIRTIO_BLK_S_IOERR	1
 #define VIRTIO_BLK_S_UNSUPP	2
+
+/* Virtblk ioctl commands */
+#define VBLK_GET_LIFETIME	_IOR('r', 0, struct virtio_blk_lifetime)
+
 #endif /* _LINUX_VIRTIO_BLK_H */
Signature · 1 line
2.32.0
maintainer
#1 lore raw
Quoted 22 lines · Alvaro Karsz

Implement the VIRTIO_BLK_F_LIFETIME feature for VirtIO block devices.

This commit introduces a new ioctl command, VBLK_LIFETIME.

VBLK_LIFETIME ioctl asks for the block device to provide lifetime information by sending a VIRTIO_BLK_T_GET_LIFETIME command to the device.

lifetime information fields:

- pre_eol_info: specifies the percentage of reserved blocks that are consumed.
optional values following virtio spec:
*) 0 - undefined.
*) 1 - normal, < 80% of reserved blocks are consumed. *) 2 - warning, 80% of reserved blocks are consumed. *) 3 - urgent, 90% of reserved blocks are consumed.

- device_lifetime_est_typ_a: this field refers to wear of SLC cells and is provided in increments of 10used, and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11 are reserved.

Quoted 6 lines

- device_lifetime_est_typ_b: this field refers to wear of MLC cells and is provided with the same semantics as device_lifetime_est_typ_a.

The data received from the device will be sent as is to the user. No data check/decode is done by virtblk.

Is this based on some spec? Because it looks pretty odd to me. There can be a pretty wide range of two/three/etc level cells with wildly different ranges of durability. And there's really not a lot of slc for generic devices these days, if any.

Signature · 1 line
Jens Axboe
#2 lore raw

Hi,

Quoted 4 lines

Is this based on some spec? Because it looks pretty odd to me. There can be a pretty wide range of two/three/etc level cells with wildly different ranges of durability. And there's really not a lot of slc for generic devices these days, if any.

Yes, this is based on the virtio spec
https://docs.oasis-open.org/virtio/virtio/v1.2/csd01/virtio-v1.2-csd01.html
section 5.2.6
_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

maintainer
#3 lore raw
Quoted 10 lines · Alvaro Karsz

Hi,

Quoted 4 lines

Is this based on some spec? Because it looks pretty odd to me. There can be a pretty wide range of two/three/etc level cells with wildly different ranges of durability. And there's really not a lot of slc for generic devices these days, if any.

Yes, this is based on the virtio spec
https://docs.oasis-open.org/virtio/virtio/v1.2/csd01/virtio-v1.2-csd01.html
section 5.2.6

And where did this come from?

Signature · 7 lines
Jens Axboe


_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
#4 lore raw
Quoted 8 lines · Jens Axboe
Quoted 7 lines · Alvaro Karsz

Hi,

Quoted 4 lines

Is this based on some spec? Because it looks pretty odd to me. There can be a pretty wide range of two/three/etc level cells with wildly different ranges of durability. And there's really not a lot of slc for generic devices these days, if any.

Quoted 5 lines
Quoted 3 lines

Yes, this is based on the virtio spec
https://docs.oasis-open.org/virtio/virtio/v1.2/csd01/virtio-v1.2-csd01.html
section 5.2.6

And where did this come from?

Here's the commit log from the spec:
In many embedded systems, virtio-blk implementations are backed by eMMC or UFS storage devices, which are subject to predictable and measurable wear over time due to repeated write cycles.

For such systems, it can be important to be able to track accurately the amount of wear imposed on the storage over time and surface it to applications. In a native deployments this is generally handled by the physical block device driver but no such provision is made in virtio-blk to expose these metrics for devices where it makes sense to do so.

This patch adds support to virtio-blk for lifetime and wear metrics to be exposed to the guest when a deployment of virtio-blk is done over compatible eMMC or UFS storage.

Cc Enrico Granata as well.

Quoted 3 lines
Signature · 1 line
Jens Axboe

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#5 lore raw

The original definitions for these fields come from JESD84-B50, which is what eMMC storage uses. It has been a while, but I recall UFS doing something pretty similar.
Systems that don't have a well defined notion of durability would just not expose the flag (e.g. a spinning disk), and going for what eMMC/UFS expose already would make implementations fairly seamless for a lot of common embedded scenarios.

Of course, if you see room for improvement to the spec, I'd be very interested in hearing your thoughts

Thanks,
- Enrico

Thanks,
- Enrico

Quoted 43 lines · Michael S. Tsirkin
Quoted 8 lines · Jens Axboe
Quoted 7 lines · Alvaro Karsz

Hi,

Quoted 4 lines

Is this based on some spec? Because it looks pretty odd to me. There can be a pretty wide range of two/three/etc level cells with wildly different ranges of durability. And there's really not a lot of slc for generic devices these days, if any.

Quoted 5 lines
Quoted 3 lines

Yes, this is based on the virtio spec
https://docs.oasis-open.org/virtio/virtio/v1.2/csd01/virtio-v1.2-csd01.html
section 5.2.6

And where did this come from?

Here's the commit log from the spec:
In many embedded systems, virtio-blk implementations are backed by eMMC or UFS storage devices, which are subject to predictable and measurable wear over time due to repeated write cycles.

For such systems, it can be important to be able to track accurately the amount of wear imposed on the storage over time and surface it to applications. In a native deployments this is generally handled by the physical block device driver but no such provision is made in virtio-blk to expose these metrics for devices where it makes sense to do so.

This patch adds support to virtio-blk for lifetime and wear metrics to be exposed to the guest when a deployment of virtio-blk is done over compatible eMMC or UFS storage.

Cc Enrico Granata as well.

Quoted 3 lines

--
Jens Axboe

#6 lore raw
Quoted 8 lines · Enrico Granata

The original definitions for these fields come from JESD84-B50, which is what eMMC storage uses. It has been a while, but I recall UFS doing something pretty similar.
Systems that don't have a well defined notion of durability would just not expose the flag (e.g. a spinning disk), and going for what eMMC/UFS expose already would make implementations fairly seamless for a lot of common embedded scenarios.

Has it been considered there might be non-embeded deployments which virtio-blk frontend can also benefit from this feature and they can have more flash cell types than present in the approved spec ?

Quoted 9 lines

Of course, if you see room for improvement to the spec, I'd be very interested in hearing your thoughts

Thanks,
- Enrico

Thanks,
- Enrico

From a quick look :-

The struct virtio_blk_lifetime is too narrow for sure only supporting two types, it at least needs to support available flash cell types to start with and the members needs to be renamed to reflect the actual type for more clarity.

hope this helps ...

-ck

maintainer
#7 lore raw
In reply to Michael S. Tsirkin
Quoted 29 lines · Michael S. Tsirkin
Quoted 8 lines · Jens Axboe
Quoted 7 lines · Alvaro Karsz

Hi,

Quoted 4 lines

Is this based on some spec? Because it looks pretty odd to me. There can be a pretty wide range of two/three/etc level cells with wildly different ranges of durability. And there's really not a lot of slc for generic devices these days, if any.

Quoted 5 lines
Quoted 3 lines

Yes, this is based on the virtio spec
https://docs.oasis-open.org/virtio/virtio/v1.2/csd01/virtio-v1.2-csd01.html
section 5.2.6

And where did this come from?

Here's the commit log from the spec:
In many embedded systems, virtio-blk implementations are backed by eMMC or UFS storage devices, which are subject to predictable and measurable wear over time due to repeated write cycles.

For such systems, it can be important to be able to track accurately the amount of wear imposed on the storage over time and surface it to applications. In a native deployments this is generally handled by the physical block device driver but no such provision is made in virtio-blk to expose these metrics for devices where it makes sense to do so.

Quoted 7 lines

This patch adds support to virtio-blk for lifetime and wear metrics to be exposed to the guest when a deployment of virtio-blk is done over compatible eMMC or UFS storage.

Cc Enrico Granata as well.

Alvaro sent me this one privately too. To be honest, I don't like this patch very much, but that's mostly because the spec for this caters to a narrow use case of embedded flash. It's not a generic storage thing, it's just for mmc/ufs and the embedded space. That's a missed opportunity. And either it'll become mostly unused, or it'll actually be used and then extended at some point.

While I'm not a fan at all, if you guys are willing to take it in virtio-blk, then I won't stand in your way. I would rename it though to more specifically detail what it deals with.

Signature · 1 line
Jens Axboe
#8 lore raw
In reply to Jens Axboe
Quoted 27 lines · Jens Axboe
Quoted 26 lines · Alvaro Karsz

Implement the VIRTIO_BLK_F_LIFETIME feature for VirtIO block devices.

This commit introduces a new ioctl command, VBLK_LIFETIME.

VBLK_LIFETIME ioctl asks for the block device to provide lifetime information by sending a VIRTIO_BLK_T_GET_LIFETIME command to the device.

lifetime information fields:

- pre_eol_info: specifies the percentage of reserved blocks that are consumed.
optional values following virtio spec:
*) 0 - undefined.
*) 1 - normal, < 80% of reserved blocks are consumed. *) 2 - warning, 80% of reserved blocks are consumed. *) 3 - urgent, 90% of reserved blocks are consumed.

- device_lifetime_est_typ_a: this field refers to wear of SLC cells and is provided in increments of 10used, and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11 are reserved.

- device_lifetime_est_typ_b: this field refers to wear of MLC cells and is provided with the same semantics as device_lifetime_est_typ_a.

Quoted 8 lines
Quoted 2 lines

The data received from the device will be sent as is to the user. No data check/decode is done by virtblk.

Is this based on some spec? Because it looks pretty odd to me. There can be a pretty wide range of two/three/etc level cells with wildly different ranges of durability. And there's really not a lot of slc for generic devices these days, if any.

This is exactly what I said on initial version about new types ...

-ck

maintainer
#9 lore raw
In reply to Alvaro Karsz
Quoted 6 lines · Alvaro Karsz

Implement the VIRTIO_BLK_F_LIFETIME feature for VirtIO block devices.

This commit introduces a new ioctl command, VBLK_LIFETIME.

VBLK_LIFETIME ioctl asks for the block device to provide lifetime information by sending a VIRTIO_BLK_T_GET_LIFETIME command to the device.

s/VBLK_LIFETIME/VBLK_GET_LIFETIME

for the above.

Signature · 7 lines
Jens Axboe


_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
reviewer Reviewed-by on v3 1/1
#10 lore raw
In reply to Alvaro Karsz

On Mon, Dec 05, 2022 at 06:20:34PM +0200, Alvaro Karsz wrote:

I don't like that the ioctl lifetime struct is passed through little-endian from the device to userspace. The point of this new ioctl is not to be a passthrough interface. The kernel should define a proper UABI struct for the ioctl and handle endianness conversion. But I think Michael is happy with this approach, so nevermind.

include/uapi/linux/virtio_blk.h struct virtio_scsi_inhdr { · line 252

Quoted 7 lines
include/uapi/linux/virtio_blk.h +3 -0
@@ -219,4 +247,8 @@ struct virtio_scsi_inhdr {
 #define VIRTIO_BLK_S_OK		0
 #define VIRTIO_BLK_S_IOERR	1
 #define VIRTIO_BLK_S_UNSUPP	2
+
+/* Virtblk ioctl commands */
+#define VBLK_GET_LIFETIME	_IOR('r', 0, struct virtio_blk_lifetime)

Does something include <linux/ioctl.h> for _IOR()? Failure to include the necessary header file could break userspace applications that include <linux/virtio_blk.h>.

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#11 lore raw
Quoted 20 lines · Stefan Hajnoczi

On Mon, Dec 05, 2022 at 06:20:34PM +0200, Alvaro Karsz wrote:

I don't like that the ioctl lifetime struct is passed through little-endian from the device to userspace. The point of this new ioctl is not to be a passthrough interface. The kernel should define a proper UABI struct for the ioctl and handle endianness conversion. But I think Michael is happy with this approach, so nevermind.

include/uapi/linux/virtio_blk.h struct virtio_scsi_inhdr { · line 252

Quoted 7 lines
include/uapi/linux/virtio_blk.h +3 -0
@@ -219,4 +247,8 @@ struct virtio_scsi_inhdr {
 #define VIRTIO_BLK_S_OK		0
 #define VIRTIO_BLK_S_IOERR	1
 #define VIRTIO_BLK_S_UNSUPP	2
+
+/* Virtblk ioctl commands */
+#define VBLK_GET_LIFETIME	_IOR('r', 0, struct virtio_blk_lifetime)

Does something include <linux/ioctl.h> for _IOR()? Failure to include the necessary header file could break userspace applications that include <linux/virtio_blk.h>.

Quoted 1 line

Good point. I think it's preferable to add a new header for this stuff. virtio_blk_ioctl.h ? Have that pull in linux/ioctl.h Also VIRTIO_BLK_IOCTL_GET_LIFETIME
might be a better name to avoid confusion and collisions. And s/Virtblk/virtio-blk/

Signature · 1 line
MST
#13 lore raw
In reply to Alvaro Karsz
Quoted 13 lines · Alvaro Karsz
include/uapi/linux/virtio_blk.h +13 -0
+/* Get lifetime information struct for each request */
+struct virtio_blk_lifetime {
+	/*
+	 * specifies the percentage of reserved blocks that are consumed.
+	 * optional values following virtio spec:
+	 * 0 - undefined
+	 * 1 - normal, < 80% of reserved blocks are consumed
+	 * 2 - warning, 80% of reserved blocks are consumed
+	 * 3 - urgent, 90% of reserved blocks are consumed
+	 */
+	__le16 pre_eol_info;
+	/*
+	 * this field refers to wear of SLC cells and is provided in increments of 10used,

include/uapi/linux/virtio_blk.h struct virtio_blk_discard_write_zeroes { · line 235

Quoted 10 lines
include/uapi/linux/virtio_blk.h +10 -0
+	 * and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11
+	 * are reserved
+	 */
+	__le16 device_lifetime_est_typ_a;
+	/*
+	 * this field refers to wear of MLC cells and is provided with the same semantics as
+	 * device_lifetime_est_typ_a
+	 */
+	__le16 device_lifetime_est_typ_b;
+};

Why does the above data structure only refer to SLC and MLC but not to TLC or QLC?

How will this data structure be extended without having to introduce a new ioctl?

Thanks,

Bart.

#14 lore raw

Hi Bart,

Quoted 2 lines

Why does the above data structure only refer to SLC and MLC but not to TLC or QLC?

This has been discussed before.
The data structure follows the virtio spec
(https://docs.oasis-open.org/virtio/virtio/v1.2/csd01/virtio-v1.2-csd01.html)

Quoted 2 lines

How will this data structure be extended without having to introduce a new ioctl?

There are no more lifetime parameters defined in the virtio spec. Please note that this is a virtio block specific ioctl, not a block generic one.

#15 lore raw
In reply to Bart Van Assche

Bart,

Quoted 31 lines · Bart Van Assche
Quoted 18 lines · Alvaro Karsz
include/uapi/linux/virtio_blk.h +13 -0
+/* Get lifetime information struct for each request */
+struct virtio_blk_lifetime {
+    /*
+     * specifies the percentage of reserved blocks that are consumed.
+     * optional values following virtio spec:
+     * 0 - undefined
+     * 1 - normal, < 80% of reserved blocks are consumed
+     * 2 - warning, 80% of reserved blocks are consumed
+     * 3 - urgent, 90% of reserved blocks are consumed
+     */
+    __le16 pre_eol_info;
+    /*
+     * this field refers to wear of SLC cells and is provided in 

increments of 10used,
+     * and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11
+     * are reserved
+     */

Quoted 8 lines

+    __le16 device_lifetime_est_typ_a;
+    /*
+     * this field refers to wear of MLC cells and is provided with the same semantics as
+     * device_lifetime_est_typ_a
+     */
+    __le16 device_lifetime_est_typ_b;
+};

Why does the above data structure only refer to SLC and MLC but not to TLC or QLC?

Quoted 7 lines

How will this data structure be extended without having to introduce a new ioctl?

Thanks,

Bart.

We already discussed that see :-

https://lists.linuxfoundation.org/pipermail/virtualization/2022-November/063742.html

-ck

#16 lore raw
In reply to Alvaro Karsz

This just seems like a horrible interface. And virtio-blk should be a simple passthrough and not grow tons of side-band crap like this.

If you want to pass through random misc information use virtio-scsi or nvme with shadow doorbell buffers.

maintainer Acked-by Christoph Hellwig
#17 lore raw
Quoted 5 lines · Christoph Hellwig

This just seems like a horrible interface. And virtio-blk should be a simple passthrough and not grow tons of side-band crap like this.

If you want to pass through random misc information use virtio-scsi or nvme with shadow doorbell buffers.

Christoph you acked the spec patch adding this to virtio blk:

Still not a fan of the encoding, but at least it is properly documented now:

Did you change your mind then? Or do you prefer a different encoding for the ioctl then? could you help sugesting what kind?

Thanks,

Signature · 6 lines
MST

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
#18 lore raw
Quoted 9 lines · Michael S. Tsirkin

Christoph you acked the spec patch adding this to virtio blk:

Still not a fan of the encoding, but at least it is properly documented now:

Did you change your mind then? Or do you prefer a different encoding for the ioctl then? could you help sugesting what kind?

Well, it is good enough documented for a spec. I don't think it is a useful feature for Linux where virtio-blk is our minimum viable paravirtualized block driver.

#19 lore raw
Quoted 6 lines · Christoph Hellwig
Quoted 5 lines · Michael S. Tsirkin

Christoph you acked the spec patch adding this to virtio blk:

Still not a fan of the encoding, but at least it is properly documented now:

Quoted 8 lines
Quoted 4 lines

Did you change your mind then? Or do you prefer a different encoding for the ioctl then? could you help sugesting what kind?

Well, it is good enough documented for a spec. I don't think it is a useful feature for Linux where virtio-blk is our minimum viable paravirtualized block driver.

No idea what this means, sorry. Now that's in the spec I expect (some) devices to implement it and if they do I see no reason not to expose the data to userspace.

Alvaro could you pls explain the use-case? Christoph has doubts that it's useful. Do you have a device implementing this?

Signature · 1 line
MST
#20 lore raw
Quoted 2 lines

Alvaro could you pls explain the use-case? Christoph has doubts that it's useful. Do you have a device implementing this?

Our HW exposes virtio devices, virtio block included. The block device backend can be a eMMC/uSD card.

The HW can report health values (power, temp, current, voltage) and I thought that it would be nice to be able to report lifetime values as well.

#21 lore raw
Quoted 9 lines · Alvaro Karsz
Quoted 2 lines

Alvaro could you pls explain the use-case? Christoph has doubts that it's useful. Do you have a device implementing this?

Our HW exposes virtio devices, virtio block included. The block device backend can be a eMMC/uSD card.

The HW can report health values (power, temp, current, voltage) and I thought that it would be nice to be able to report lifetime values as well.

Why not use block layer passthru interface to fetch this instead of adding non-generic IOCTL ?

-ck

#22 lore raw
In reply to Michael S. Tsirkin

Michael,

Quoted 12 lines · Michael S. Tsirkin
Quoted 11 lines · Christoph Hellwig
Quoted 9 lines · Michael S. Tsirkin

Christoph you acked the spec patch adding this to virtio blk:

Still not a fan of the encoding, but at least it is properly documented now:

Did you change your mind then? Or do you prefer a different encoding for the ioctl then? could you help sugesting what kind?

Quoted 8 lines
Quoted 3 lines

Well, it is good enough documented for a spec. I don't think it is a useful feature for Linux where virtio-blk is our minimum viable paravirtualized block driver.

No idea what this means, sorry. Now that's in the spec I expect (some) devices to implement it and if they do I see no reason not to expose the data to userspace.

Even if any device implements is it can always use passthru commands. See below for more info...

Quoted 3 lines

Alvaro could you pls explain the use-case? Christoph has doubts that it's useful. Do you have a device implementing this?

From what I know, virtio-blk should be kept minimal and should not
add any storage specific IOCTLs or features that will end up loosing its generic nature.

The IOCTL we are trying to add is Flash storage specific which goes against the nature of generic storage and makes it non-generic. In case we approve this it will open the door for non-generic code/IOCTL in the virtio-blk and that needs to be avoided.

For any storage specific features or IOCTL (flash/HDD) it should be using it's own frontend such as virtio-scsi or e.g. nvme and not virtio-blk.

Hope this helps.

-ck

#23 lore raw
Quoted 31 lines · Chaitanya Kulkarni

Michael,

Quoted 12 lines · Michael S. Tsirkin
Quoted 11 lines · Christoph Hellwig
Quoted 9 lines · Michael S. Tsirkin

Christoph you acked the spec patch adding this to virtio blk:

Still not a fan of the encoding, but at least it is properly documented now:

Did you change your mind then? Or do you prefer a different encoding for the ioctl then? could you help sugesting what kind?

Quoted 8 lines
Quoted 3 lines

Well, it is good enough documented for a spec. I don't think it is a useful feature for Linux where virtio-blk is our minimum viable paravirtualized block driver.

No idea what this means, sorry. Now that's in the spec I expect (some) devices to implement it and if they do I see no reason not to expose the data to userspace.

Even if any device implements is it can always use passthru commands. See below for more info...

Quoted 3 lines

Alvaro could you pls explain the use-case? Christoph has doubts that it's useful. Do you have a device implementing this?

Quoted 8 lines

From what I know, virtio-blk should be kept minimal and should not
add any storage specific IOCTLs or features that will end up loosing its generic nature.

The IOCTL we are trying to add is Flash storage specific which goes against the nature of generic storage and makes it non-generic. In case we approve this it will open the door for non-generic code/IOCTL in the virtio-blk and that needs to be avoided.

Wrt these fields that horse has bolted, it's in the spec.

Quoted 7 lines

For any storage specific features or IOCTL (flash/HDD) it should be using it's own frontend such as virtio-scsi or e.g. nvme and not virtio-blk.

Hope this helps.

-ck

I don't understand what you are suggesting, sorry. It's a hardware device. It can't just "switch to a different frontend".

Signature · 6 lines
MST

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization
#24 lore raw
In reply to Alvaro Karsz
Quoted 281 lines · Alvaro Karsz

Implement the VIRTIO_BLK_F_LIFETIME feature for VirtIO block devices.

This commit introduces a new ioctl command, VBLK_LIFETIME.

VBLK_LIFETIME ioctl asks for the block device to provide lifetime information by sending a VIRTIO_BLK_T_GET_LIFETIME command to the device.

lifetime information fields:

- pre_eol_info: specifies the percentage of reserved blocks that are consumed.
optional values following virtio spec:
*) 0 - undefined.
*) 1 - normal, < 80% of reserved blocks are consumed. *) 2 - warning, 80% of reserved blocks are consumed. *) 3 - urgent, 90% of reserved blocks are consumed.

- device_lifetime_est_typ_a: this field refers to wear of SLC cells and is provided in increments of 10used, and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11 are reserved.

- device_lifetime_est_typ_b: this field refers to wear of MLC cells and is provided with the same semantics as device_lifetime_est_typ_a.

The data received from the device will be sent as is to the user. No data check/decode is done by virtblk.

--
v2:
- Remove (void *) casting.
- Fix comments format in virtio_blk.h.
- Set ioprio value for legacy devices for REQ_OP_DRV_IN operations.

v3:
- Initialize struct virtio_blk_lifetime to 0 instead of memset.
- Rename ioctl from VBLK_LIFETIME to VBLK_GET_LIFETIME.
- Return EOPNOTSUPP insted of ENOTTY if ioctl is not supported.
- Check if process is CAP_SYS_ADMIN capable in lifetime ioctl.
- Check if vdev is not NULL before accessing it in lifetime ioctl.
--

---
 drivers/block/virtio_blk.c      | 106 ++++++++++++++++++++++++++++++--
 include/uapi/linux/virtio_blk.h |  32 ++++++++++
 2 files changed, 133 insertions(+), 5 deletions(-)
drivers/block/virtio_blk.c +101 -5
index 19da5defd73..392d57fd6b7 100644
@@ -101,6 +101,18 @@ static inline blk_status_t virtblk_result(struct virtblk_req *vbr)
 	}
 }
 
+static inline int virtblk_ioctl_result(struct virtblk_req *vbr)
+{
+	switch (vbr->status) {
+	case VIRTIO_BLK_S_OK:
+		return 0;
+	case VIRTIO_BLK_S_UNSUPP:
+		return -EOPNOTSUPP;
+	default:
+		return -EIO;
+	}
+}
+
 static inline struct virtio_blk_vq *get_virtio_blk_vq(struct blk_mq_hw_ctx *hctx)
 {
 	struct virtio_blk *vblk = hctx->queue->queuedata;
@@ -218,6 +230,7 @@ static blk_status_t virtblk_setup_cmd(struct virtio_device *vdev,
 	u32 type;
 
 	vbr->out_hdr.sector = 0;
+	vbr->out_hdr.ioprio = cpu_to_virtio32(vdev, req_get_ioprio(req));
 
 	switch (req_op(req)) {
 	case REQ_OP_READ:
@@ -244,15 +257,14 @@ static blk_status_t virtblk_setup_cmd(struct virtio_device *vdev,
 		type = VIRTIO_BLK_T_SECURE_ERASE;
 		break;
 	case REQ_OP_DRV_IN:
-		type = VIRTIO_BLK_T_GET_ID;
-		break;
+		/* type is set in virtblk_get_id/virtblk_ioctl_lifetime */
+		return 0;
 	default:
 		WARN_ON_ONCE(1);
 		return BLK_STS_IOERR;
 	}
 
 	vbr->out_hdr.type = cpu_to_virtio32(vdev, type);
-	vbr->out_hdr.ioprio = cpu_to_virtio32(vdev, req_get_ioprio(req));
 
 	if (type == VIRTIO_BLK_T_DISCARD || type == VIRTIO_BLK_T_WRITE_ZEROES ||
 	    type == VIRTIO_BLK_T_SECURE_ERASE) {
@@ -459,12 +471,16 @@ static int virtblk_get_id(struct gendisk *disk, char *id_str)
 	struct virtio_blk *vblk = disk->private_data;
 	struct request_queue *q = vblk->disk->queue;
 	struct request *req;
+	struct virtblk_req *vbr;
 	int err;
 
 	req = blk_mq_alloc_request(q, REQ_OP_DRV_IN, 0);
 	if (IS_ERR(req))
 		return PTR_ERR(req);
 
+	vbr = blk_mq_rq_to_pdu(req);
+	vbr->out_hdr.type = cpu_to_virtio32(vblk->vdev, VIRTIO_BLK_T_GET_ID);
+
 	err = blk_rq_map_kern(q, req, id_str, VIRTIO_BLK_ID_BYTES, GFP_KERNEL);
 	if (err)
 		goto out;
@@ -508,6 +524,85 @@ static int virtblk_getgeo(struct block_device *bd, struct hd_geometry *geo)
 	return ret;
 }
 
+/* Get lifetime information from device */
+static int virtblk_ioctl_lifetime(struct virtio_blk *vblk, unsigned long arg)
+{
+	struct request_queue *q = vblk->disk->queue;
+	struct request *req = NULL;
+	struct virtblk_req *vbr;
+	struct virtio_blk_lifetime lifetime = {};
+	int ret;
+
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+
+	/* The virtio_blk_lifetime struct fields follow virtio spec.
+	 * There is no check/decode on values received from the device.
+	 * The data is sent as is to the user.
+	 */
+
+	/* This ioctl is allowed only if VIRTIO_BLK_F_LIFETIME
+	 * feature is negotiated.
+	 */
+	if (!virtio_has_feature(vblk->vdev, VIRTIO_BLK_F_LIFETIME))
+		return -EOPNOTSUPP;
+
+	req = blk_mq_alloc_request(q, REQ_OP_DRV_IN, 0);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
+	/* Write the correct type */
+	vbr = blk_mq_rq_to_pdu(req);
+	vbr->out_hdr.type = cpu_to_virtio32(vblk->vdev, VIRTIO_BLK_T_GET_LIFETIME);
+
+	ret = blk_rq_map_kern(q, req, &lifetime, sizeof(lifetime), GFP_KERNEL);
+	if (ret)
+		goto out;
+
+	blk_execute_rq(req, false);
+
+	ret = virtblk_ioctl_result(blk_mq_rq_to_pdu(req));
+	if (ret)
+		goto out;
+
+	/* Pass the data to the user */
+	if (copy_to_user((void __user *)arg, &lifetime, sizeof(lifetime))) {
+		ret = -EFAULT;
+		goto out;
+	}
+
+out:
+	blk_mq_free_request(req);
+	return ret;
+}
+
+static int virtblk_ioctl(struct block_device *bd, fmode_t mode,
+			 unsigned int cmd, unsigned long arg)
+{
+	struct virtio_blk *vblk = bd->bd_disk->private_data;
+	int ret;
+
+	mutex_lock(&vblk->vdev_mutex);
+
+	if (!vblk->vdev) {
+		ret = -ENXIO;
+		goto exit;
+	}
+
+	switch (cmd) {
+	case VBLK_GET_LIFETIME:
+		ret = virtblk_ioctl_lifetime(vblk, arg);
+		break;
+	default:
+		ret = -EOPNOTSUPP;
+		break;
+	}
+
+exit:
+	mutex_unlock(&vblk->vdev_mutex);
+	return ret;
+}
+
 static void virtblk_free_disk(struct gendisk *disk)
 {
 	struct virtio_blk *vblk = disk->private_data;
@@ -520,6 +615,7 @@ static void virtblk_free_disk(struct gendisk *disk)
 static const struct block_device_operations virtblk_fops = {
 	.owner  	= THIS_MODULE,
 	.getgeo		= virtblk_getgeo,
+	.ioctl		= virtblk_ioctl,
 	.free_disk	= virtblk_free_disk,
 };
 
@@ -1239,7 +1335,7 @@ static unsigned int features_legacy[] = {
 	VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
 	VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY, VIRTIO_BLK_F_CONFIG_WCE,
 	VIRTIO_BLK_F_MQ, VIRTIO_BLK_F_DISCARD, VIRTIO_BLK_F_WRITE_ZEROES,
-	VIRTIO_BLK_F_SECURE_ERASE,
+	VIRTIO_BLK_F_SECURE_ERASE, VIRTIO_BLK_F_LIFETIME,
 }
 ;
 static unsigned int features[] = {
@@ -1247,7 +1343,7 @@ static unsigned int features[] = {
 	VIRTIO_BLK_F_RO, VIRTIO_BLK_F_BLK_SIZE,
 	VIRTIO_BLK_F_FLUSH, VIRTIO_BLK_F_TOPOLOGY, VIRTIO_BLK_F_CONFIG_WCE,
 	VIRTIO_BLK_F_MQ, VIRTIO_BLK_F_DISCARD, VIRTIO_BLK_F_WRITE_ZEROES,
-	VIRTIO_BLK_F_SECURE_ERASE,
+	VIRTIO_BLK_F_SECURE_ERASE, VIRTIO_BLK_F_LIFETIME,
 };
 
 static struct virtio_driver virtio_blk = {
include/uapi/linux/virtio_blk.h +28 -0
index 58e70b24b50..e755d62d2ea 100644
@@ -40,6 +40,7 @@
 #define VIRTIO_BLK_F_MQ		12	/* support more than one vq */
 #define VIRTIO_BLK_F_DISCARD	13	/* DISCARD is supported */
 #define VIRTIO_BLK_F_WRITE_ZEROES	14	/* WRITE ZEROES is supported */
+#define VIRTIO_BLK_F_LIFETIME	15 /* Storage lifetime information is supported */
 #define VIRTIO_BLK_F_SECURE_ERASE	16 /* Secure Erase is supported */
 
 /* Legacy feature bits */
@@ -165,6 +166,9 @@ struct virtio_blk_config {
 /* Get device ID command */
 #define VIRTIO_BLK_T_GET_ID    8
 
+/* Get lifetime information command */
+#define VIRTIO_BLK_T_GET_LIFETIME 10
+
 /* Discard command */
 #define VIRTIO_BLK_T_DISCARD	11
 
@@ -206,6 +210,30 @@ struct virtio_blk_discard_write_zeroes {
 	__le32 flags;
 };
 
+/* Get lifetime information struct for each request */
+struct virtio_blk_lifetime {
+	/*
+	 * specifies the percentage of reserved blocks that are consumed.
+	 * optional values following virtio spec:
+	 * 0 - undefined
+	 * 1 - normal, < 80% of reserved blocks are consumed
+	 * 2 - warning, 80% of reserved blocks are consumed
+	 * 3 - urgent, 90% of reserved blocks are consumed
+	 */
+	__le16 pre_eol_info;
+	/*
+	 * this field refers to wear of SLC cells and is provided in increments of 10used,
+	 * and so on, thru to 11 meaning estimated lifetime exceeded. All values above 11
+	 * are reserved
+	 */
+	__le16 device_lifetime_est_typ_a;
+	/*
+	 * this field refers to wear of MLC cells and is provided with the same semantics as
+	 * device_lifetime_est_typ_a
+	 */
+	__le16 device_lifetime_est_typ_b;
+};
+
 #ifndef VIRTIO_BLK_NO_LEGACY
 struct virtio_scsi_inhdr {
 	__virtio32 errors;

include/uapi/linux/virtio_blk.h struct virtio_scsi_inhdr { · line 254

Quoted 8 lines
include/uapi/linux/virtio_blk.h +4 -0
@@ -219,4 +247,8 @@ struct virtio_scsi_inhdr {
 #define VIRTIO_BLK_S_OK		0
 #define VIRTIO_BLK_S_IOERR	1
 #define VIRTIO_BLK_S_UNSUPP	2
+
+/* Virtblk ioctl commands */
+#define VBLK_GET_LIFETIME	_IOR('r', 0, struct virtio_blk_lifetime)
+
 #endif /* _LINUX_VIRTIO_BLK_H */

So this makes the header not self-contained: you need to pull in ioctl.h. And that's a bit annoying to people who are used to just have spec defines in this header. Maybe add a new one virtio_blk_ioctl.h ?

And I'd still like to change VBLK_ prefix here to something like VIRTIO_BLK_IOCTL_

And maybe document that this is just for specific type of backend device, and mention the types which benefit, in code comment if not in the ioctl name.

Thanks!

Quoted 2 lines
Signature · 1 line
2.32.0

_______________________________________________
Virtualization mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/virtualization

#25 lore raw

Hi Michael,
Sorry, I had no time to create a new version.
I'll do it today.

Alvaro