[PATCH v0 1/2] libahci: Implement the function restart_engine to restart the port dma engine.

[PATCH v0 1/2] libahci: Implement the function restart_engine to restart the port dma engine.

[Tejun Heo]

On Sat, Jun 07, 2014 at 02:58:53AM +0530, Suman Tripathi wrote:
> This patch implements the function restart_engine function to add the flexibility to  restart the port dma engine from the libata framework.
> 
> Signed-off-by: Loc Ho <lho@apm.com>
> Signed-off-by: Suman Tripathi <stripathi@apm.com>
> ---
>  drivers/ata/libahci.c  | 11 +++++++++++
>  include/linux/libata.h |  1 +
>  2 files changed, 12 insertions(+)
> 
> diff --git a/drivers/ata/libahci.c b/drivers/ata/libahci.c
> index b986145..5f9a13e 100644
> --- a/drivers/ata/libahci.c
> +++ b/drivers/ata/libahci.c
> @@ -70,6 +70,7 @@ static int ahci_scr_read(struct ata_link *link, unsigned int sc_reg, u32 *val);
>  static int ahci_scr_write(struct ata_link *link, unsigned int sc_reg, u32 val);
>  static unsigned int ahci_qc_issue(struct ata_queued_cmd *qc);
>  static bool ahci_qc_fill_rtf(struct ata_queued_cmd *qc);
> +static int ahci_restart_engine(struct ata_port *ap);
>  static int ahci_port_start(struct ata_port *ap);
>  static void ahci_port_stop(struct ata_port *ap);
>  static void ahci_qc_prep(struct ata_queued_cmd *qc);
> @@ -178,6 +179,7 @@ struct ata_port_operations ahci_ops = {
>  #endif
>  	.port_start		= ahci_port_start,
>  	.port_stop		= ahci_port_stop,
> +	.restart_engine		= ahci_restart_engine,

Umm... you're adding something very ahci-specific as general libata
callback.  I haven't looked at the second patch yet but don't think
this is gonna work out.

Thanks.

-- 
tejun

[PATCH v0 2/2] ata: Fix the dma state machine lockup for the PIO mode commands.

[Tejun Heo]

On Sat, Jun 07, 2014 at 02:58:54AM +0530, Suman Tripathi wrote:
> @@ -5072,6 +5072,16 @@ int ata_qc_complete_multiple(struct ata_port *ap, u32 qc_active)
>  		if (qc) {
>  			ata_qc_complete(qc);
>  			nr_done++;
> +			/*
> +			 * Some controller unable to clear the BSY bit after
> +			 * receiving the PIO Setup FIS from device resulting
> +			 * the DMA state to go into CMFatalErrorUpdate state.
> +			 * So need to restart the dma engine to get the
> +			 * controller out of this state.
> +			 */
> +			if ((ap->flags & ATA_HORKAGE_BROKEN_PIO_CMD) &&
> +			    (qc->tf.protocol == ATA_PROT_PIO))
> +				ap->ops->restart_engine(ap);

Why aren't you doing this by wrapping irq handler?

Thanks.

-- 
tejun