Re: [RFC PATCH v3 3/3] nvme: wire up support for async passthrough

From: Keith Busch <kbusch@kernel.org>
To: Kanchan Joshi <joshi.k@samsung.com>
Cc: axboe@kernel.dk, hch@lst.de, chaitanya.kulkarni@wdc.com,
	io-uring@vger.kernel.org, linux-nvme@lists.infradead.org,
	anuj20.g@samsung.com, javier.gonz@samsung.com,
	nj.shetty@samsung.com, selvakuma.s1@samsung.com
Subject: Re: [RFC PATCH v3 3/3] nvme: wire up support for async passthrough
Date: Wed, 17 Mar 2021 09:45:50 -0700	[thread overview]
Message-ID: <20210317164550.GA4162742@dhcp-10-100-145-180.wdc.com> (raw)
In-Reply-To: <20210316140126.24900-4-joshi.k@samsung.com>

On Tue, Mar 16, 2021 at 07:31:26PM +0530, Kanchan Joshi wrote:
> @@ -1179,6 +1278,20 @@ static int nvme_submit_user_cmd(struct request_queue *q,
>  			req->cmd_flags |= REQ_INTEGRITY;
>  		}
>  	}
> +	if (ioucmd) { /* async handling */
> +		u32 effects;
> +
> +		effects = nvme_command_effects(ns->ctrl, ns, cmd->common.opcode);
> +		/* filter commands with non-zero effects, keep it simple for now*/

You shouldn't need to be concerned with this. You've wired up the ioucmd
only to the NVME_IOCTL_IO_CMD, and nvme_command_effects() can only
return 0 for that.

It would be worth adding support for NVME_IOCTL_IO_CMD64 too, though,
and that doesn't change the effects handling either.

> +		if (effects) {
> +			ret = -EOPNOTSUPP;
> +			goto out_unmap;
> +		}
> +		nvme_setup_uring_cmd_data(req, ioucmd, meta, write);
> +		blk_execute_rq_nowait(ns ? ns->disk : NULL, req, 0,
> +					nvme_end_async_pt);
> +		return 0;
> +	}

_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme