Re: [PATCH 4/7] call/invoke SATA ACPI suspend/resume functions

[Jeff Garzik <jgarzik@pobox.com>]

Randy Dunlap wrote:
> From: Randy Dunlap <randy_d_dunlap@linux.intel.com>
> 
> Add calls to ACPI methods for SATA drives.
> 
> Signed-off-by: Randy Dunlap <randy_d_dunlap@linux.intel.com>
> ---
>  drivers/scsi/libata-core.c |    4 ++++
>  drivers/scsi/libata.h      |   26 ++++++++++++++++++++++++++
>  2 files changed, 30 insertions(+)
> 
> --- linux-2615-rc5g3.orig/drivers/scsi/libata-core.c
> +++ linux-2615-rc5g3/drivers/scsi/libata-core.c
> @@ -1345,6 +1345,8 @@ void ata_dev_config(struct ata_port *ap,
>  
>  	if (ap->ops->dev_config)
>  		ap->ops->dev_config(ap, &ap->device[i]);
> +
> +	do_drive_SDD(ap, i);
>  }
>  
>  /**
> @@ -1784,6 +1786,8 @@ static void ata_set_mode(struct ata_port
>  	if (ap->flags & ATA_FLAG_PORT_DISABLED)
>  		return;
>  
> +	do_drive_update_taskfiles(ap);
> +
>  	if (ap->ops->post_set_mode)
>  		ap->ops->post_set_mode(ap);
>  

Mostly OK.

1) The function names are awful.  I would suggest ata_acpi_push_id() and 
ata_acpi_exec_tfs() instead.

2) The call to do_drive_update_taskfiles() should be done in 
ata_bus_probe(), not inside ata_set_mode() where it is essentially 
hidden amongst code that is dedicated to a single purpose, getting the 
data xfer timings.

	Jeff