On Mon, Nov 15, 2021 at 12:52:00PM +0800, yadong.qi@intel.com wrote:

The Linux block layer shares the DISCARD queue limits with SECDISCARD.
That's different from BLKZEROOUT (QEMU's WRITE_ZEROES). This is a Linux
implementation detail but I guess virtio-blk can share the DISCARD
limits with SECDISCARD too...

> @@ -622,6 +628,7 @@ static int virtio_blk_handle_request(VirtIOBlockReq *req, MultiReqBuffer *mrb)
>      unsigned out_num = req->elem.out_num;
>      VirtIOBlock *s = req->dev;
>      VirtIODevice *vdev = VIRTIO_DEVICE(s);
> +    bool is_secdiscard = false;
>  
>      if (req->elem.out_num < 1 || req->elem.in_num < 1) {
>          virtio_error(vdev, "virtio-blk missing headers");
> @@ -722,6 +729,9 @@ static int virtio_blk_handle_request(VirtIOBlockReq *req, MultiReqBuffer *mrb)
>       * VIRTIO_BLK_T_OUT flag set. We masked this flag in the switch statement,
>       * so we must mask it for these requests, then we will check if it is set.
>       */
> +    case VIRTIO_BLK_T_SECDISCARD & ~VIRTIO_BLK_T_OUT:
> +        is_secdiscard = true;
> +        __attribute__((fallthrough));

The DISCARD case doesn't use __attribute__((fallthrough)) so this is
inconsistent. QEMU doesn't use __attribute__((fallthrough)) so I suggest
dropping this.

> diff --git a/include/standard-headers/linux/virtio_blk.h b/include/standard-headers/linux/virtio_blk.h
> index 2dcc90826a..c55a07840c 100644
> --- a/include/standard-headers/linux/virtio_blk.h
> +++ b/include/standard-headers/linux/virtio_blk.h
> @@ -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_SECDISCARD	15	/* WRITE ZEROES is supported */

The comment is copy-pasted from WRITE_ZEROES.