diff mbox

ata: fix Marvell SATA driver dependencies

Message ID 5138028.BSQbEP7RRB@amdc1032 (mailing list archive)
State New, archived
Headers show

Commit Message

Bartlomiej Zolnierkiewicz March 26, 2014, 4:19 p.m. UTC
Make sata_mv host driver depend on PCI || ARCH_KIRKWOOD || ARCH_MVEBU ||
ARCH_ORION5X config options as Marvell SATA support covers both Marvell
PCI devices and Marvell Kirkwood, Armada 370/XP and Orion5x SoCs (for
non-PCI devices the driver to work requires suitable device tree node to
be defined).  Additionally allow the driver build if COMPILE_TEST config
option is set.

Cc: Jason Cooper <jason@lakedaemon.net>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Gregory Clement <gregory.clement@free-electrons.com>
Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
---
 drivers/ata/Kconfig |    2 ++
 1 file changed, 2 insertions(+)

Comments

Andrew Lunn March 26, 2014, 4:23 p.m. UTC | #1
On Wed, Mar 26, 2014 at 05:19:02PM +0100, Bartlomiej Zolnierkiewicz wrote:
> Make sata_mv host driver depend on PCI || ARCH_KIRKWOOD || ARCH_MVEBU ||
> ARCH_ORION5X config options as Marvell SATA support covers both Marvell
> PCI devices and Marvell Kirkwood, Armada 370/XP and Orion5x SoCs (for
> non-PCI devices the driver to work requires suitable device tree node to
> be defined).  Additionally allow the driver build if COMPILE_TEST config
> option is set.

Hi Bartlomiej

I _think_ there are a few powerpc platforms which can also make use of
this driver.

     Andrew
Tejun Heo March 26, 2014, 4:29 p.m. UTC | #2
On Wed, Mar 26, 2014 at 05:19:02PM +0100, Bartlomiej Zolnierkiewicz wrote:
> Make sata_mv host driver depend on PCI || ARCH_KIRKWOOD || ARCH_MVEBU ||
> ARCH_ORION5X config options as Marvell SATA support covers both Marvell
> PCI devices and Marvell Kirkwood, Armada 370/XP and Orion5x SoCs (for
> non-PCI devices the driver to work requires suitable device tree node to
> be defined).  Additionally allow the driver build if COMPILE_TEST config
> option is set.
> 
> Cc: Jason Cooper <jason@lakedaemon.net>
> Cc: Andrew Lunn <andrew@lunn.ch>
> Cc: Gregory Clement <gregory.clement@free-electrons.com>
> Cc: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
> Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>

Patch doesn't apply to libata/for-3.15.

Thanks.
Bartlomiej Zolnierkiewicz March 26, 2014, 4:57 p.m. UTC | #3
Hi,

On Wednesday, March 26, 2014 05:23:02 PM Andrew Lunn wrote:
> On Wed, Mar 26, 2014 at 05:19:02PM +0100, Bartlomiej Zolnierkiewicz wrote:
> > Make sata_mv host driver depend on PCI || ARCH_KIRKWOOD || ARCH_MVEBU ||
> > ARCH_ORION5X config options as Marvell SATA support covers both Marvell
> > PCI devices and Marvell Kirkwood, Armada 370/XP and Orion5x SoCs (for
> > non-PCI devices the driver to work requires suitable device tree node to
> > be defined).  Additionally allow the driver build if COMPILE_TEST config
> > option is set.
> 
> Hi Bartlomiej
> 
> I _think_ there are a few powerpc platforms which can also make use of
> this driver.

Currently no powerpc platform uses this driver for non-PCI devices and
looking at commits history for powerpc dts files it seems that it also
hasn't happened in the past.

I think that the current patch is sufficient as we can always relax
dependencies later if the need for it shows up + there is a possibility
to (ab)use COMPILE_TEST for older kernels.

Best regards,
--
Bartlomiej Zolnierkiewicz
Samsung R&D Institute Poland
Samsung Electronics
Alan Cox March 26, 2014, 4:59 p.m. UTC | #4
On Wed, 26 Mar 2014 17:23:02 +0100
Andrew Lunn <andrew@lunn.ch> wrote:

> On Wed, Mar 26, 2014 at 05:19:02PM +0100, Bartlomiej Zolnierkiewicz wrote:
> > Make sata_mv host driver depend on PCI || ARCH_KIRKWOOD || ARCH_MVEBU ||
> > ARCH_ORION5X config options as Marvell SATA support covers both Marvell
> > PCI devices and Marvell Kirkwood, Armada 370/XP and Orion5x SoCs (for
> > non-PCI devices the driver to work requires suitable device tree node to
> > be defined).  Additionally allow the driver build if COMPILE_TEST config
> > option is set.
> 
> Hi Bartlomiej
> 
> I _think_ there are a few powerpc platforms which can also make use of
> this driver

The hardware can show up in just about anything so I think this patch
should be dropped
Bartlomiej Zolnierkiewicz March 26, 2014, 5:12 p.m. UTC | #5
Hi,

On Wednesday, March 26, 2014 04:59:04 PM One Thousand Gnomes wrote:
> On Wed, 26 Mar 2014 17:23:02 +0100
> Andrew Lunn <andrew@lunn.ch> wrote:
> 
> > On Wed, Mar 26, 2014 at 05:19:02PM +0100, Bartlomiej Zolnierkiewicz wrote:
> > > Make sata_mv host driver depend on PCI || ARCH_KIRKWOOD || ARCH_MVEBU ||
> > > ARCH_ORION5X config options as Marvell SATA support covers both Marvell
> > > PCI devices and Marvell Kirkwood, Armada 370/XP and Orion5x SoCs (for
> > > non-PCI devices the driver to work requires suitable device tree node to
> > > be defined).  Additionally allow the driver build if COMPILE_TEST config
> > > option is set.
> > 
> > Hi Bartlomiej
> > 
> > I _think_ there are a few powerpc platforms which can also make use of
> > this driver
> 
> The hardware can show up in just about anything so I think this patch
> should be dropped

In theory maybe but in practice it for sure won't show up if the vendor
has its own SATA solution and at least in the ARM world it seems frequent
to have its own AHCI variant nowadays.  Moreover even if the new usage
shows up it is not enough to enable driver for it to make it work as
you also need to add a node describing your SATA MV variant to your SoC
device tree file (which requires updating kernel source anyway).

Best regards,
--
Bartlomiej Zolnierkiewicz
Samsung R&D Institute Poland
Samsung Electronics
Alan Cox March 26, 2014, 5:36 p.m. UTC | #6
> shows up it is not enough to enable driver for it to make it work as
> you also need to add a node describing your SATA MV variant to your SoC
> device tree file (which requires updating kernel source anyway).

You still have to recompile your kernel to update your device tree..
how... quaint ;-). It's not a generic limitation but I'm not that fussed.
It seems a fair argument that there would be other changes made when this
was being done for a system, and probably new board stuff.

Alan
diff mbox

Patch

Index: b/drivers/ata/Kconfig
===================================================================
--- a/drivers/ata/Kconfig	2014-03-26 15:49:59.919101924 +0100
+++ b/drivers/ata/Kconfig	2014-03-26 17:02:08.187028122 +0100
@@ -273,6 +273,8 @@  config SATA_HIGHBANK
 
 config SATA_MV
 	tristate "Marvell SATA support"
+	depends on PCI || ARCH_KIRKWOOD || ARCH_MVEBU || ARCH_ORION5X || \
+		   COMPILE_TEST
 	select GENERIC_PHY
 	help
 	  This option enables support for the Marvell Serial ATA family.