Patchwork ahci: Turn off DMA engines when there's no device attached

login
register
mail settings
Submitter Matthew Garrett
Date March 10, 2010, 8:32 p.m.
Message ID <1268253150-24050-1-git-send-email-mjg@redhat.com>
Download mbox | patch
Permalink /patch/47283/
State Not Applicable
Delegated to: David Miller
Headers show

Comments

Matthew Garrett - March 10, 2010, 8:32 p.m.
According to section 10.3.1 of the AHCI spec, PxCMD.ST must not be set
unless there's a device attached. Following this saves us a measurable
quantity of power and does not impair hotplug support. Based on a patch
by Kristen Carlson Accardi.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Cc: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
---
 drivers/ata/ahci.c |   19 +++++++++++++++++++
 1 files changed, 19 insertions(+), 0 deletions(-)
Matthew Garrett - March 24, 2010, 2:19 p.m.
Any feedback on this?

On Wed, Mar 10, 2010 at 03:32:30PM -0500, Matthew Garrett wrote:
> According to section 10.3.1 of the AHCI spec, PxCMD.ST must not be set
> unless there's a device attached. Following this saves us a measurable
> quantity of power and does not impair hotplug support. Based on a patch
> by Kristen Carlson Accardi.
> 
> Signed-off-by: Matthew Garrett <mjg@redhat.com>
> Cc: Kristen Carlson Accardi <kristen.c.accardi@intel.com>
> ---
>  drivers/ata/ahci.c |   19 +++++++++++++++++++
>  1 files changed, 19 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c
> index 6bd930b..ff08317 100644
> --- a/drivers/ata/ahci.c
> +++ b/drivers/ata/ahci.c
> @@ -1016,11 +1016,30 @@ static int ahci_scr_write(struct ata_link *link, unsigned int sc_reg, u32 val)
>  	return -EINVAL;
>  }
>  
> +static int ahci_is_device_present(struct ata_port *ap)
> +{
> +	void __iomem *port_mmio = ahci_port_base(ap);
> +	u8 status = readl(port_mmio + PORT_TFDATA) & 0xff;
> +
> +	/* Make sure PxTFD.STS.BSY and PxTFD.STS.DRQ are 0 */
> +	if (status & (ATA_BUSY | ATA_DRQ))
> +		return 0;
> +
> +	/* Make sure PxSSTS.DET is 3h */
> +	status = readl(port_mmio + PORT_SCR_STAT) & 0xf;
> +	if (status != 3)
> +		return 0;
> +	return 1;
> +}
> +
>  static void ahci_start_engine(struct ata_port *ap)
>  {
>  	void __iomem *port_mmio = ahci_port_base(ap);
>  	u32 tmp;
>  
> +	if (!ahci_is_device_present(ap))
> +		return;
> +
>  	/* start DMA */
>  	tmp = readl(port_mmio + PORT_CMD);
>  	tmp |= PORT_CMD_START;
> -- 
> 1.6.6.1
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ide" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
>
Jeff Garzik - March 24, 2010, 2:43 p.m.
On 03/24/2010 10:19 AM, Matthew Garrett wrote:
> On Wed, Mar 10, 2010 at 03:32:30PM -0500, Matthew Garrett wrote:
>> According to section 10.3.1 of the AHCI spec, PxCMD.ST must not be set
>> unless there's a device attached. Following this saves us a measurable
>> quantity of power and does not impair hotplug support. Based on a patch
>> by Kristen Carlson Accardi.
>>
>> Signed-off-by: Matthew Garrett<mjg@redhat.com>
>> Cc: Kristen Carlson Accardi<kristen.c.accardi@intel.com>
>> ---
>>   drivers/ata/ahci.c |   19 +++++++++++++++++++
>>   1 files changed, 19 insertions(+), 0 deletions(-)


(please don't top-post, it makes replying more difficult)

> Any feedback on this?

It should be fine, but I have to rebase it onto the latest upstream ahci 
(libata-dev.git#upstream), so it's been waiting on that...

	Jeff


--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jeff Garzik - March 28, 2010, 4:39 a.m.
On 03/10/2010 03:32 PM, Matthew Garrett wrote:
> According to section 10.3.1 of the AHCI spec, PxCMD.ST must not be set
> unless there's a device attached. Following this saves us a measurable
> quantity of power and does not impair hotplug support. Based on a patch
> by Kristen Carlson Accardi.
>
> Signed-off-by: Matthew Garrett<mjg@redhat.com>
> Cc: Kristen Carlson Accardi<kristen.c.accardi@intel.com>
> ---
>   drivers/ata/ahci.c |   19 +++++++++++++++++++
>   1 files changed, 19 insertions(+), 0 deletions(-)

applied manually to libata-dev.git#upstream, due to large code movements 
in creation of new libahci kernel module


--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jeff Garzik - March 29, 2010, 9:30 p.m.
On 03/10/2010 03:32 PM, Matthew Garrett wrote:
> +static int ahci_is_device_present(struct ata_port *ap)
> +{
> +	void __iomem *port_mmio = ahci_port_base(ap);
> +	u8 status = readl(port_mmio + PORT_TFDATA)&  0xff;
> +
> +	/* Make sure PxTFD.STS.BSY and PxTFD.STS.DRQ are 0 */
> +	if (status&  (ATA_BUSY | ATA_DRQ))
> +		return 0;
> +
> +	/* Make sure PxSSTS.DET is 3h */
> +	status = readl(port_mmio + PORT_SCR_STAT)&  0xf;
> +	if (status != 3)
> +		return 0;
> +	return 1;


Note that this last test, reading SATA Control Register (SCR) Status, 
does not account for in-transition status, ie. devices that were just 
plugged in.

	Jeff


--
To unsubscribe from this list: send the line "unsubscribe linux-ide" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/drivers/ata/ahci.c b/drivers/ata/ahci.c
index 6bd930b..ff08317 100644
--- a/drivers/ata/ahci.c
+++ b/drivers/ata/ahci.c
@@ -1016,11 +1016,30 @@  static int ahci_scr_write(struct ata_link *link, unsigned int sc_reg, u32 val)
 	return -EINVAL;
 }
 
+static int ahci_is_device_present(struct ata_port *ap)
+{
+	void __iomem *port_mmio = ahci_port_base(ap);
+	u8 status = readl(port_mmio + PORT_TFDATA) & 0xff;
+
+	/* Make sure PxTFD.STS.BSY and PxTFD.STS.DRQ are 0 */
+	if (status & (ATA_BUSY | ATA_DRQ))
+		return 0;
+
+	/* Make sure PxSSTS.DET is 3h */
+	status = readl(port_mmio + PORT_SCR_STAT) & 0xf;
+	if (status != 3)
+		return 0;
+	return 1;
+}
+
 static void ahci_start_engine(struct ata_port *ap)
 {
 	void __iomem *port_mmio = ahci_port_base(ap);
 	u32 tmp;
 
+	if (!ahci_is_device_present(ap))
+		return;
+
 	/* start DMA */
 	tmp = readl(port_mmio + PORT_CMD);
 	tmp |= PORT_CMD_START;