Re: [PATCH] scsi: sd: add runtime pm to open / release

[Bart Van Assche <bvanassche@acm.org>]

On 2020-06-23 04:10, Martin Kepplinger wrote:
> This add a very conservative but simple implementation for runtime PM
> to the sd scsi driver:
> Resume when opened (mounted) and suspend when released (unmounted).
> 
> Improvements that allow suspending while a device is "open" can
> be added later, but now we save power when no filesystem is mounted
> and runtime PM is enabled.
> 
> Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
> ---
>  drivers/scsi/sd.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
> index d90fefffe31b..fe4cb7c50ec1 100644
> --- a/drivers/scsi/sd.c
> +++ b/drivers/scsi/sd.c
> @@ -1372,6 +1372,7 @@ static int sd_open(struct block_device *bdev, fmode_t mode)
>  	SCSI_LOG_HLQUEUE(3, sd_printk(KERN_INFO, sdkp, "sd_open\n"));
>  
>  	sdev = sdkp->device;
> +	scsi_autopm_get_device(sdev);
>  
>  	/*
>  	 * If the device is in error recovery, wait until it is done.
> @@ -1418,6 +1419,9 @@ static int sd_open(struct block_device *bdev, fmode_t mode)
>  
>  error_out:
>  	scsi_disk_put(sdkp);
> +
> +	scsi_autopm_put_device(sdev);
> +
>  	return retval;	
>  }
>  
> @@ -1441,6 +1445,8 @@ static void sd_release(struct gendisk *disk, fmode_t mode)
>  
>  	SCSI_LOG_HLQUEUE(3, sd_printk(KERN_INFO, sdkp, "sd_release\n"));
>  
> +	scsi_autopm_put_device(sdev);
> +
>  	if (atomic_dec_return(&sdkp->openers) == 0 && sdev->removable) {
>  		if (scsi_block_when_processing_errors(sdev))
>  			scsi_set_medium_removal(sdev, SCSI_REMOVAL_ALLOW);

My understanding of the above patch is that it introduces a regression,
namely by disabling runtime suspend as long as an sd device is held open.

Bart.