Re: [PATCH 3/5] libata-dev: Let ata_hsm_move() work with both irq-pio and polling pio

[Jeff Garzik <jgarzik@pobox.com>]

Albert Lee wrote:
> +static int ata_hsm_move(struct ata_port *ap, struct ata_queued_cmd *qc,
> +			 u8 status, int in_wq)
>  {
> +	unsigned long flags = 0;
> +	int poll_next;
> +
>  	WARN_ON((qc->flags & ATA_QCFLAG_ACTIVE) == 0);
>  
> +	/* Make sure ata_qc_issue_prot() does not throw things
> +	 * like DMA polling into the workqueue. Notice that
> +	 * in_wq is not equivalent to (qc->tf.flags & ATA_TFLAG_POLLING).
> +	 */
> +	WARN_ON(in_wq != ((qc->tf.flags & ATA_TFLAG_POLLING) ||
> +			  (ap->hsm_task_state == HSM_ST_FIRST &&
> +			   ((qc->tf.protocol == ATA_PROT_PIO &&
> +			     (qc->tf.flags & ATA_TFLAG_WRITE)) ||
> +			    (is_atapi_taskfile(&qc->tf) &&
> +			     !(qc->dev->flags & ATA_DFLAG_CDB_INTR))))));


applied patches 1-5 to the irq-pio branch.

Please provide a follow-up patch which makes the above test more 
readable.  For example:

static inline int things_ok_in_wq(input...)
{
	if (test 1)
		return 0;
	if (test 2)
		return 0;
	return 1;	
}