Re: [PATCH v2 2/2] scsi: donot increase scsi_device's iorequest_cnt if dispatch failed

[Mike Christie <michael.christie@oracle.com>]

On 10/21/22 6:56 PM, Wenchao Hao wrote:
> If scsi_dispatch_cmd() failed, the scsi command did not send to disks,
> so it would never done from LLDs.
> 
> scsi scsi_queue_rq() would return BLK_STS_RESOURCE if
> scsi_dispatch_cmd() failed, the related request would be requeued, and
> the timeout of this request would not fired any more, so no one
> would increase iodone_cnt which matches with this increase of
> iorequest_cnt.
> 
> Signed-off-by: Wenchao Hao <haowenchao@huawei.com>
> ---
>  drivers/scsi/scsi_lib.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c
> index 8b89fab7c420..71edb9ffbe16 100644
> --- a/drivers/scsi/scsi_lib.c
> +++ b/drivers/scsi/scsi_lib.c
> @@ -1469,8 +1469,6 @@ static int scsi_dispatch_cmd(struct scsi_cmnd *cmd)
>  	struct Scsi_Host *host = cmd->device->host;
>  	int rtn = 0;
>  
> -	atomic_inc(&cmd->device->iorequest_cnt);
> ->  	/* check if the device is still usable */
>  	if (unlikely(cmd->device->sdev_state == SDEV_DEL)) {
>  		/* in SDEV_DEL we error all commands. DID_NO_CONNECT
> @@ -1764,6 +1762,7 @@ static blk_status_t scsi_queue_rq(struct blk_mq_hw_ctx *hctx,
>  		goto out_dec_host_busy;
>  	}
>  
> +	atomic_inc(&cmd->device->iorequest_cnt);
>  	return BLK_STS_OK;
>  
>  out_dec_host_busy:

Reviewed-by: Mike Christie <michael.christie@oracle.com>