On Thursday, March 09, 2017 03:08:29 PM Russell King - ARM Linux wrote:

On Thu, Mar 09, 2017 at 03:57:37PM +0100, Bartlomiej Zolnierkiewicz wrote:

quoted

On Thursday, March 09, 2017 02:51:47 PM Russell King - ARM Linux wrote:

quoted

On Thu, Mar 09, 2017 at 03:42:49PM +0100, Bartlomiej Zolnierkiewicz wrote:

quoted

Hi,

On Thursday, March 09, 2017 01:57:15 PM Russell King - ARM Linux wrote:

quoted

On Thu, Mar 09, 2017 at 03:24:56PM +0300, Sergei Shtylyov wrote:

quoted

On 03/09/2017 03:20 PM, Sekhar Nori wrote:

quoted

[...]

quoted

quoted

quoted

diff --git a/arch/arm/mach-davinci/board-dm644x-evm.c

b/arch/arm/mach-davinci/board-dm644x-evm.c
index 023480b75244..60a1f23890cd 100644

--- a/arch/arm/mach-davinci/board-dm644x-evm.c
+++ b/arch/arm/mach-davinci/board-dm644x-evm.c

@@ -744,7 +744,7 @@ static int davinci_phy_fixup(struct phy_device

*phydev)
    return 0;
}

-#define HAS_ATA        IS_ENABLED(CONFIG_BLK_DEV_PALMCHIP_BK3710)
+#define HAS_ATA        IS_ENABLED(CONFIG_PATA_BK3710)

  I think it would be more correct to check for both libata and IDE
drivers here...

As I understand, the plan is to remove the IDE driver soon.

   I'm not sure DaveM would support any removals in drivers/ide/. He has
explicitly expressed his will to maintain the IDE driuver forever.

As I keep saying, there are ARM machines where the IDE driver works but
the PATA driver doesn't.  EBSA110's PCMCIA slots for instance.

Please be more specific, I'll be happy to help with migrating this
machines to libata PATA.

BTW the example you're giving has never worked with the old IDE
driver in the upstream kernel.

That's where you are wrong.  It worked for 10+ years with no
modifications in the upstream IDE driver or ide-cs driver - only the
patch I quoted was required.

..and the patch you quoted is in which upstream kernel version exactly?

I don't claim that it works with the upstream kernel as a whole.  I claim
that it works with the ide-cs driver _unmodified_ in the upstream kernel,
but with a simple two line patch which is private to the ARM architecture.

Please pay attention to what I write, thanks.

Out-of-tree patches are on their own, as kernel evolves they need
to adapt, nothing new here.  For 10+ years you were just extremely
lucky with your out-of-tree patch. 8)

Anyway, if you ever get to running current upstream on your EBSA110
please ping me and I will provide you with the draft pata_pcmcia
patch to make EBSA110 work with libata.

Best regards,
--
Bartlomiej Zolnierkiewicz
Samsung R&D Institute Poland
Samsung Electronics