On 3/20/2024 12:04 AM, Stefano Garzarella wrote:
On Mon, Feb 19, 2024 at 02:55:58AM +0800, Zhu Lingshan wrote:
This commit allows reporting the max size of any
single segment of virtio-block devices to user space.
Signed-off-by: Zhu Lingshan <lingshan.zhu@xxxxxxxxx>
---
drivers/vdpa/vdpa.c | 17 +++++++++++++++++
include/uapi/linux/vdpa.h | 1 +
2 files changed, 18 insertions(+)
diff --git a/drivers/vdpa/vdpa.c b/drivers/vdpa/vdpa.c
index d5ccb618de2b..ee1472d32b0d 100644
--- a/drivers/vdpa/vdpa.c
+++ b/drivers/vdpa/vdpa.c
@@ -956,6 +956,20 @@ vdpa_dev_blk_capacity_config_fill(struct sk_buff
*msg,
val_u64, VDPA_ATTR_PAD);
}
+static int
+vdpa_dev_blk_seg_size_config_fill(struct sk_buff *msg, u64 features,
+ const struct virtio_blk_config *config)
+{
+ u32 val_u32;
+
+ if ((features & BIT_ULL(VIRTIO_BLK_F_SIZE_MAX)) == 0)
+ return 0;
+
+ val_u32 = __virtio32_to_cpu(true, config->size_max);
+
+ return nla_put_u32(msg, VDPA_ATTR_DEV_BLK_CFG_SEG_SIZE, val_u32);
Should we call this VDPA_ATTR_DEV_BLK_CFG_SIZE_MAX ?
I have tried this before, but this macro is too long causing a odd
format in the header file.
I can change this since you point this out. This is uapi means we should
merge the
fix during the RC release before 6.9 official.
@MST, does the process work for you?
+}
+
static int vdpa_dev_blk_config_fill(struct vdpa_device *vdev,
struct sk_buff *msg)
{
@@ -973,6 +987,9 @@ static int vdpa_dev_blk_config_fill(struct
vdpa_device *vdev,
if (vdpa_dev_blk_capacity_config_fill(msg, &config))
return -EMSGSIZE;
+ if (vdpa_dev_blk_seg_size_config_fill(msg, features_device,
&config))
+ return -EMSGSIZE;
+
return 0;
}
diff --git a/include/uapi/linux/vdpa.h b/include/uapi/linux/vdpa.h
index 1bf69226cb96..586bce3c906a 100644
--- a/include/uapi/linux/vdpa.h
+++ b/include/uapi/linux/vdpa.h
@@ -57,6 +57,7 @@ enum vdpa_attr {
VDPA_ATTR_DEV_FEATURES, /* u64 */
VDPA_ATTR_DEV_BLK_CFG_CAPACITY, /* u64 */
+ VDPA_ATTR_DEV_BLK_CFG_SEG_SIZE, /* u32 */
/* new attributes must be added above here */
VDPA_ATTR_MAX,
--
2.39.3
