[PATCH 05/14] nvmet: get rid of the extra variable

[Logan Gunthorpe]

On 2020-08-10 12:54 p.m., Chaitanya Kulkarni wrote:
> In function nvmet_passthru_execute_cmd() the return value of the
> function nvmet_passthru_map_sg() is not used to determine stauts i.e.
> we can safely ignore the actual cause of error that is returned and just
> check for binary status pass or fail.
> 
> The prior check to this function call can be combined safely with the &&
> so that we can get rid of the extra variable and second if condition.
> 
> Remove the extra variable and adjust the code accordingly to have only
> one if condition.
> 
> Signed-off-by: Chaitanya Kulkarni <chaitanya.kulkarni at wdc.com>
> ---
>  drivers/nvme/target/passthru.c | 10 +++-------
>  1 file changed, 3 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/nvme/target/passthru.c b/drivers/nvme/target/passthru.c
> index 58377420874f..9a175b8b9d57 100644
> --- a/drivers/nvme/target/passthru.c
> +++ b/drivers/nvme/target/passthru.c
> @@ -221,7 +221,6 @@ static void nvmet_passthru_execute_cmd(struct nvmet_req *req)
>  	struct request *rq = NULL;
>  	u32 effects;
>  	u16 status;
> -	int ret;
>  
>  	if (likely(req->sq->qid != 0)) {
>  		u32 nsid = le32_to_cpu(req->cmd->common.nsid);
> @@ -242,12 +241,9 @@ static void nvmet_passthru_execute_cmd(struct nvmet_req *req)
>  		goto out_put_ns;
>  	}
>  
> -	if (likely(req->sg_cnt)) {
> -		ret = nvmet_passthru_map_sg(req, rq);
> -		if (unlikely(ret)) {
> -			status = NVME_SC_INTERNAL;
> -			goto out_put_req;
> -		}
> +	if (likely(req->sg_cnt) && unlikely(nvmet_passthru_map_sg(req, rq))) {
> +		status = NVME_SC_INTERNAL;
> +		goto out_put_req;
>  	}

IMO, that's really ugly and not at all an improvement.

Logan