On Tue, May 20, 2014 at 11:18:11AM +0200, Sebastian Hesselbarth wrote:

On 05/20/2014 11:04 AM, Antoine T?nart wrote:

quoted

The Marvell Berlin AHCI has all his specific in the PHY driver. It then
only need to use the libahci functions to work properly.

If it is that generic, ..

quoted

Add its compatible into the libahci_platform driver.

Signed-off-by: Antoine T?nart <redacted>
---
 drivers/ata/ahci_platform.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/ata/ahci_platform.c b/drivers/ata/ahci_platform.c
index ef67e79944f9..bc050aabf206 100644
--- a/drivers/ata/ahci_platform.c
+++ b/drivers/ata/ahci_platform.c

@@ -76,6 +76,7 @@ static const struct of_device_id ahci_of_match[] = {
 	{ .compatible = "snps,exynos5440-ahci", },
 	{ .compatible = "ibm,476gtr-ahci", },
 	{ .compatible = "snps,dwc-ahci", },
+	{ .compatible = "marvell,berlin-ahci", },

.. why have a Marvell-specific compatible?

Well, the 3 other compatibles seemed as generic as the marvell one. Just
following what was done before here.

How about "generic-ahci" instead, like we have for other fooHCIs
already?

That would avoid an endless list of compatibles :)

Antoine

-- 
Antoine T?nart, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com