[patch 1/1] kirkwood: Add iconnect support

[patch 1/1] kirkwood: Add iconnect support

[Arnaud Patard (Rtp)]

Andrew Lunn <andrew@lunn.ch> writes:

Hi,

>> > + 
>> > +#define ICONNECT_BLINK_HALF_PERIOD   100
>> > +
>> > +static int iconnect_blink_set(unsigned gpio, int state,
>>
>>__init ?
>
> Hi Jason
>
> It is correct without __init. This function will get called when you
> play with /sys/class/leds/*/trigger.
>
> However, this is the second public instance of this function, it is
> also in mach-orion5x/dns323-setup.c.

yeah, I know. I initially planned to send a patchset adding iconnect
support and putting this function into something like gpio.c but I
forgot about it when converting my first iconnect patch into DT. Any
objection about sending a patch doing that once I manage to merge the
iconnect support (and of course, in this case, feel free to remind me
of it if I forget to do it ) ?


Arnaud

[patch 1/1] kirkwood: Add iconnect support

[Jason Cooper]

On Wed, Apr 18, 2012 at 02:53:40PM +0200, Arnaud Patard wrote:
> Andrew Lunn <andrew@lunn.ch> writes:
> 
> Hi,
> 
> >> > + 
> >> > +#define ICONNECT_BLINK_HALF_PERIOD   100
> >> > +
> >> > +static int iconnect_blink_set(unsigned gpio, int state,
> >>
> >>__init ?
> >
> > Hi Jason
> >
> > It is correct without __init. This function will get called when you
> > play with /sys/class/leds/*/trigger.
> >
> > However, this is the second public instance of this function, it is
> > also in mach-orion5x/dns323-setup.c.
> 
> yeah, I know. I initially planned to send a patchset adding iconnect
> support and putting this function into something like gpio.c but I
> forgot about it when converting my first iconnect patch into DT. Any
> objection about sending a patch doing that once I manage to merge the
> iconnect support (and of course, in this case, feel free to remind me
> of it if I forget to do it ) ?

I'd prefer to have the fixes first, then base new stuff off of the fix.

thx,

Jason.

[patch 0/1] Kirwood: Add Iomega Iconnect support

[Arnaud Patard (Rtp)]

The following patch is adding support for Iomaga Iconnect kirkwood box. It has
tested against arm-soc#for-next of some days ago and with ARM_APPENDED_DTB as
the uboot I have on this device doesn't support device tree.
The DTS file only handles the serial port as it seems it's the only device
with OF support [ rtc-mv is converted too but it's already handled by
kirkwood.dtsi ]


Arnaud Patard

[patch 1/1] kirkwood: Add iconnect support

[Arnaud Patard (Rtp)]

An embedded and charset-unspecified text was scrubbed...
Name: iconnect-dt.patch
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20120417/5acbebf8/attachment.ksh>

[patch 1/1] kirkwood: Add iconnect support

[Jason Cooper]

On Tue, Apr 17, 2012 at 12:05:06AM +0200, Arnaud Patard wrote:
> Add support for Iomega Iconnect system.
> 
> Signed-off-by: Arnaud Patard <arnaud.patard@rtp-net.org>

Thanks for the patch, comments below.

> Index: arm-soc/arch/arm/mach-kirkwood/Kconfig
> ===================================================================
> --- arm-soc.orig/arch/arm/mach-kirkwood/Kconfig	2012-04-16 08:50:58.321398110 +0200
> +++ arm-soc/arch/arm/mach-kirkwood/Kconfig	2012-04-16 14:07:10.440563405 +0200
> @@ -58,6 +58,12 @@ config MACH_DREAMPLUG_DT
>  	  Say 'Y' here if you want your kernel to support the
>  	  Marvell DreamPlug (Flattened Device Tree).
>  
> +config MACH_ICONNECT_DT
> +	bool "Iomega Iconnect (Flattened Device Tree)"
> +	select ARCH_KIRKWOOD_DT
> +	help
> +	  Say 'Y' here to enable Iomega Iconnect support.
> +
>  config MACH_TS219
>  	bool "QNAP TS-110, TS-119, TS-119P+, TS-210, TS-219, TS-219P and TS-219P+ Turbo NAS"
>  	help
> Index: arm-soc/arch/arm/mach-kirkwood/Makefile
> ===================================================================
> --- arm-soc.orig/arch/arm/mach-kirkwood/Makefile	2012-04-16 08:50:58.349398109 +0200
> +++ arm-soc/arch/arm/mach-kirkwood/Makefile	2012-04-16 14:04:50.660569553 +0200
> @@ -22,3 +22,4 @@ obj-$(CONFIG_MACH_T5325)		+= t5325-setup
>  obj-$(CONFIG_CPU_IDLE)			+= cpuidle.o
>  obj-$(CONFIG_ARCH_KIRKWOOD_DT)		+= board-dt.o
>  obj-$(CONFIG_MACH_DREAMPLUG_DT)		+= board-dreamplug.o
> +obj-$(CONFIG_MACH_ICONNECT_DT)		+= board-iconnect.o
> Index: arm-soc/arch/arm/mach-kirkwood/board-iconnect.c
> ===================================================================
> --- /dev/null	1970-01-01 00:00:00.000000000 +0000
> +++ arm-soc/arch/arm/mach-kirkwood/board-iconnect.c	2012-04-16 14:05:51.304566885 +0200
> @@ -0,0 +1,187 @@
> +/*
> + * arch/arm/mach-kirkwood/board-iconnect.c
> + *
> + * Iomega i-connect Board Setup
> + *
> + * This file is licensed under the terms of the GNU General Public
> + * License version 2.  This program is licensed "as is" without any
> + * warranty of any kind, whether express or implied.
> + */
> +
> +#include <linux/kernel.h>
> +#include <linux/init.h>
> +#include <linux/platform_device.h>
> +#include <linux/of.h>
> +#include <linux/of_address.h>
> +#include <linux/of_fdt.h>
> +#include <linux/of_irq.h>
> +#include <linux/of_platform.h>
> +#include <linux/mtd/partitions.h>
> +#include <linux/mv643xx_eth.h>
> +#include <linux/gpio.h>
> +#include <linux/leds.h>
> +#include <linux/spi/flash.h>
> +#include <linux/spi/spi.h>
> +#include <linux/spi/orion_spi.h>
> +#include <linux/i2c.h>
> +#include <linux/input.h>
> +#include <linux/gpio_keys.h>
> +#include <asm/mach/arch.h>
> +#include <mach/kirkwood.h>
> +#include "common.h"
> +#include "mpp.h"
> +
> +static struct mv643xx_eth_platform_data iconnect_ge00_data = {
> +	.phy_addr	= MV643XX_ETH_PHY_ADDR(11),
> +};
> +
> +static struct gpio_led iconnect_led_pins[] = {
> +	{
> +		.name		= "led_level",

This name sounds like a led attribute, not it's name.  Is it correct?

> +		.gpio		= 41,
> +		.default_trigger = "default-on",
> +	}, {
> +		.name		= "power:blue",
> +		.gpio		= 42,
> +		.default_trigger = "timer",
> +	}, {
> +		.name		= "power:red",
> +		.gpio		= 43,
> +	}, {
> +		.name		= "usb1",
> +		.gpio		= 44,
> +	}, {
> +		.name		= "usb2",
> +		.gpio		= 45,
> +	}, {
> +		.name		= "usb3",
> +		.gpio		= 46,
> +	}, {
> +		.name		= "usb4",
> +		.gpio		= 47,
> +	}, {
> +		.name		= "otb",
> +		.gpio		= 48,
> +	},
> +};

Are usb[1-4] different hcis? are they all ehci?  what color are they?

> +
> +#define ICONNECT_BLINK_HALF_PERIOD	100
> +
> +static int iconnect_blink_set(unsigned gpio, int state,

__init ?

> +	unsigned long *delay_on, unsigned long *delay_off)
> +{
> +	if (delay_on && delay_off && !*delay_on && !*delay_off)
> +		*delay_on = *delay_off = ICONNECT_BLINK_HALF_PERIOD;
> +
> +	switch (state) {
> +	case GPIO_LED_NO_BLINK_LOW:
> +	case GPIO_LED_NO_BLINK_HIGH:
> +		orion_gpio_set_blink(gpio, 0);
> +		gpio_set_value(gpio, state);
> +		break;
> +	case GPIO_LED_BLINK:
> +		orion_gpio_set_blink(gpio, 1);
> +	}
> +
> +	return 0;
> +}
> +
> +static struct gpio_led_platform_data iconnect_led_data = {
> +	.leds		= iconnect_led_pins,
> +	.num_leds	= ARRAY_SIZE(iconnect_led_pins),
> +	.gpio_blink_set	= iconnect_blink_set,
> +};
> +
> +static struct platform_device iconnect_leds = {
> +	.name	= "leds-gpio",
> +	.id	= -1,
> +	.dev	= {
> +		.platform_data	= &iconnect_led_data,
> +	}
> +};
> +
> +static unsigned int iconnect_mpp_config[] __initdata = {
> +	MPP12_GPIO,
> +	MPP35_GPIO,
> +	MPP41_GPIO,
> +	MPP42_GPIO,
> +	MPP43_GPIO,
> +	MPP44_GPIO,
> +	MPP45_GPIO,
> +	MPP46_GPIO,
> +	MPP47_GPIO,
> +	MPP48_GPIO,
> +	0
> +};

Some comments here would be helpful, 41-48 are gpio_led, what are 12 and
35?

> +
> +static struct i2c_board_info __initdata iconnect_board_info[] = {
> +	{
> +		I2C_BOARD_INFO("lm63", 0x4c),
> +	},
> +};
> +
> +static struct mtd_partition iconnect_nand_parts[] = {
> +	{
> +		.name = "flash",
> +		.offset = 0,
> +		.size = MTDPART_SIZ_FULL,
> +	},
> +};

I was going to suggest moving the partition definitions to here, but
once orion_nand is fdt aware, we'll specify the partitions in the dts
anyway.

Unless someone has an objection, I say we leave the hackish commandline
part table in the dtb and convert it in the dtb once orion_nand had
devicetree bindings.

> +
> +/* yikes... theses are the original input buttons */
> +/* but I'm not convinced by the sw event choices  */
> +static struct gpio_keys_button iconnect_buttons[] = {
> +	{
> +		.type		= EV_SW,
> +		.code		= SW_LID,
> +		.gpio		= 12,
> +		.desc		= "Reset Button",
> +		.active_low	= 1,
> +		.debounce_interval = 100,
> +	}, {
> +		.type		= EV_SW,
> +		.code		= SW_TABLET_MODE,
> +		.gpio		= 35,
> +		.desc		= "OTB Button",
> +		.active_low	= 1,
> +		.debounce_interval = 100,
> +	},
> +};
> +
> +static struct gpio_keys_platform_data iconnect_button_data = {
> +	.buttons	= iconnect_buttons,
> +	.nbuttons	= ARRAY_SIZE(iconnect_buttons),
> +};
> +
> +static struct platform_device iconnect_button_device = {
> +	.name		= "gpio-keys",
> +	.id		= -1,
> +	.num_resources	= 0,
> +	.dev        = {
> +		.platform_data  = &iconnect_button_data,
> +	},
> +};
> +
> +void __init iconnect_init(void)
> +{
> +	kirkwood_mpp_conf(iconnect_mpp_config);
> +	kirkwood_nand_init(ARRAY_AND_SIZE(iconnect_nand_parts), 25);
> +	kirkwood_i2c_init();
> +	i2c_register_board_info(0, iconnect_board_info,
> +		ARRAY_SIZE(iconnect_board_info));
> +
> +	kirkwood_ehci_init();
> +	kirkwood_ge00_init(&iconnect_ge00_data);
> +
> +	platform_device_register(&iconnect_button_device);
> +	platform_device_register(&iconnect_leds);
> +}
> +
> +static int __init iconnect_pci_init(void)
> +{
> +	if (of_machine_is_compatible("iomega,iconnect"))
> +		kirkwood_pcie_init(KW_PCIE0);
> +	return 0;
> +}
> +subsys_initcall(iconnect_pci_init);
> +
> Index: arm-soc/arch/arm/boot/dts/kirkwood-iconnect.dts
> ===================================================================
> --- /dev/null	1970-01-01 00:00:00.000000000 +0000
> +++ arm-soc/arch/arm/boot/dts/kirkwood-iconnect.dts	2012-04-16 08:59:12.033376388 +0200
> @@ -0,0 +1,26 @@
> +/dts-v1/;
> +
> +/include/ "kirkwood.dtsi"
> +
> +/ {
> +	model = "Iomega Iconnect";
> +	compatible = "iomega,iconnect", "mrvl,kirkwood-88f6281", "mrvl,kirkwood";

This should be "iom,iconnect-MODELNUM", "iom,iconnect", "mrvl,...

And, correct elsewhere as needed.

> +
> +	memory {
> +		device_type = "memory";
> +		reg = <0x00000000 0x10000000>;
> +	};
> +
> +	chosen {
> +		bootargs = "console=ttyS0,115200n8 earlyprintk mtdparts=orion_nand:0xc0000 at 0x0(uboot),0x20000 at 0xa0000(env),0x300000 at 0x100000(zImage),0x300000 at 0x540000(initrd),0x1f400000 at 0x980000(boot)";
> +		linux,initrd-start = <0x4500040>;
> +		linux,initrd-end   = <0x4800000>;

does this need to be specified?  why doesn't the bootloader pass the
initrd properly?

thx,

Jason.

> +	};
> +
> +	ocp at f1000000 {
> +		serial at 12000 {
> +			clock-frequency = <200000000>;
> +			status = "ok";
> +		};
> +	};
> +};
> Index: arm-soc/arch/arm/mach-kirkwood/Makefile.boot
> ===================================================================
> --- arm-soc.orig/arch/arm/mach-kirkwood/Makefile.boot	2012-04-16 08:50:58.329398110 +0200
> +++ arm-soc/arch/arm/mach-kirkwood/Makefile.boot	2012-04-16 08:51:03.873397866 +0200
> @@ -3,3 +3,5 @@ params_phys-y	:= 0x00000100
>  initrd_phys-y	:= 0x00800000
>  
>  dtb-$(CONFIG_MACH_DREAMPLUG_DT) += kirkwood-dreamplug.dtb
> +dtb-$(CONFIG_MACH_ICONNECT_DT) += kirkwood-iconnect.dtb
> +
> Index: arm-soc/arch/arm/mach-kirkwood/board-dt.c
> ===================================================================
> --- arm-soc.orig/arch/arm/mach-kirkwood/board-dt.c	2012-04-16 08:50:58.337398109 +0200
> +++ arm-soc/arch/arm/mach-kirkwood/board-dt.c	2012-04-16 08:51:03.881397865 +0200
> @@ -55,11 +55,15 @@ static void __init kirkwood_dt_init(void
>  	if (of_machine_is_compatible("globalscale,dreamplug"))
>  		dreamplug_init();
>  
> +	if (of_machine_is_compatible("iomega,iconnect"))
> +		iconnect_init();
> +
>  	of_platform_populate(NULL, kirkwood_dt_match_table, NULL, NULL);
>  }
>  
>  static const char *kirkwood_dt_board_compat[] = {
>  	"globalscale,dreamplug",
> +	"iomega,iconnect",
>  	NULL
>  };
>  
> Index: arm-soc/arch/arm/mach-kirkwood/common.h
> ===================================================================
> --- arm-soc.orig/arch/arm/mach-kirkwood/common.h	2012-04-16 08:50:58.309398111 +0200
> +++ arm-soc/arch/arm/mach-kirkwood/common.h	2012-04-16 08:51:03.905397864 +0200
> @@ -57,6 +57,12 @@ void dreamplug_init(void);
>  #else
>  static inline void dreamplug_init(void) {};
>  #endif
> +#ifdef CONFIG_MACH_ICONNECT_DT
> +void iconnect_init(void);
> +#else
> +static inline void iconnect_init(void) {};
> +#endif
> +
>  
>  /* early init functions not converted to fdt yet */
>  char *kirkwood_id(void);
> 
> 
> 
> _______________________________________________
> linux-arm-kernel mailing list
> linux-arm-kernel at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

[patch 1/1] kirkwood: Add iconnect support

[Arnaud Patard (Rtp)]

Jason Cooper <jason@lakedaemon.net> writes:

> On Tue, Apr 17, 2012 at 12:05:06AM +0200, Arnaud Patard wrote:
>> Add support for Iomega Iconnect system.
>> 
>> Signed-off-by: Arnaud Patard <arnaud.patard@rtp-net.org>
>
> Thanks for the patch, comments below.
>
>> Index: arm-soc/arch/arm/mach-kirkwood/Kconfig
>> ===================================================================
>> --- arm-soc.orig/arch/arm/mach-kirkwood/Kconfig	2012-04-16 08:50:58.321398110 +0200
>> +++ arm-soc/arch/arm/mach-kirkwood/Kconfig	2012-04-16 14:07:10.440563405 +0200
>> @@ -58,6 +58,12 @@ config MACH_DREAMPLUG_DT
>>  	  Say 'Y' here if you want your kernel to support the
>>  	  Marvell DreamPlug (Flattened Device Tree).
>>  
>> +config MACH_ICONNECT_DT
>> +	bool "Iomega Iconnect (Flattened Device Tree)"
>> +	select ARCH_KIRKWOOD_DT
>> +	help
>> +	  Say 'Y' here to enable Iomega Iconnect support.
>> +
>>  config MACH_TS219
>>  	bool "QNAP TS-110, TS-119, TS-119P+, TS-210, TS-219, TS-219P and TS-219P+ Turbo NAS"
>>  	help
>> Index: arm-soc/arch/arm/mach-kirkwood/Makefile
>> ===================================================================
>> --- arm-soc.orig/arch/arm/mach-kirkwood/Makefile	2012-04-16 08:50:58.349398109 +0200
>> +++ arm-soc/arch/arm/mach-kirkwood/Makefile	2012-04-16 14:04:50.660569553 +0200
>> @@ -22,3 +22,4 @@ obj-$(CONFIG_MACH_T5325)		+= t5325-setup
>>  obj-$(CONFIG_CPU_IDLE)			+= cpuidle.o
>>  obj-$(CONFIG_ARCH_KIRKWOOD_DT)		+= board-dt.o
>>  obj-$(CONFIG_MACH_DREAMPLUG_DT)		+= board-dreamplug.o
>> +obj-$(CONFIG_MACH_ICONNECT_DT)		+= board-iconnect.o
>> Index: arm-soc/arch/arm/mach-kirkwood/board-iconnect.c
>> ===================================================================
>> --- /dev/null	1970-01-01 00:00:00.000000000 +0000
>> +++ arm-soc/arch/arm/mach-kirkwood/board-iconnect.c	2012-04-16 14:05:51.304566885 +0200
>> @@ -0,0 +1,187 @@
>> +/*
>> + * arch/arm/mach-kirkwood/board-iconnect.c
>> + *
>> + * Iomega i-connect Board Setup
>> + *
>> + * This file is licensed under the terms of the GNU General Public
>> + * License version 2.  This program is licensed "as is" without any
>> + * warranty of any kind, whether express or implied.
>> + */
>> +
>> +#include <linux/kernel.h>
>> +#include <linux/init.h>
>> +#include <linux/platform_device.h>
>> +#include <linux/of.h>
>> +#include <linux/of_address.h>
>> +#include <linux/of_fdt.h>
>> +#include <linux/of_irq.h>
>> +#include <linux/of_platform.h>
>> +#include <linux/mtd/partitions.h>
>> +#include <linux/mv643xx_eth.h>
>> +#include <linux/gpio.h>
>> +#include <linux/leds.h>
>> +#include <linux/spi/flash.h>
>> +#include <linux/spi/spi.h>
>> +#include <linux/spi/orion_spi.h>
>> +#include <linux/i2c.h>
>> +#include <linux/input.h>
>> +#include <linux/gpio_keys.h>
>> +#include <asm/mach/arch.h>
>> +#include <mach/kirkwood.h>
>> +#include "common.h"
>> +#include "mpp.h"
>> +
>> +static struct mv643xx_eth_platform_data iconnect_ge00_data = {
>> +	.phy_addr	= MV643XX_ETH_PHY_ADDR(11),
>> +};
>> +
>> +static struct gpio_led iconnect_led_pins[] = {
>> +	{
>> +		.name		= "led_level",
>
> This name sounds like a led attribute, not it's name.  Is it correct?
>

I know. In fact, this gpio allows to change the brightness of all leds
(either low luminosity or high luminosity). I'm not sure how to handle
it in a better way so I've reproduced what the vendor did. Better ideas
are welcomed :)

>> +		.gpio		= 41,
>> +		.default_trigger = "default-on",
>> +	}, {
>> +		.name		= "power:blue",
>> +		.gpio		= 42,
>> +		.default_trigger = "timer",
>> +	}, {
>> +		.name		= "power:red",
>> +		.gpio		= 43,
>> +	}, {
>> +		.name		= "usb1",
>> +		.gpio		= 44,
>> +	}, {
>> +		.name		= "usb2",
>> +		.gpio		= 45,
>> +	}, {
>> +		.name		= "usb3",
>> +		.gpio		= 46,
>> +	}, {
>> +		.name		= "usb4",
>> +		.gpio		= 47,
>> +	}, {
>> +		.name		= "otb",
>> +		.gpio		= 48,
>> +	},
>> +};
>
> Are usb[1-4] different hcis? are they all ehci?  what color are they?

There are 4 usb ports and each port has a (blue) led near it, that's why
they're called usb{1,2,3,4}. Will add the color for next patch version.

>
>> +
>> +#define ICONNECT_BLINK_HALF_PERIOD	100
>> +
>> +static int iconnect_blink_set(unsigned gpio, int state,
>
> __init ?
>

I don't think so. The leds-gpio driver can be module so putting it as
__init will make it unavalaible for the module.

>> +	unsigned long *delay_on, unsigned long *delay_off)
>> +{
>> +	if (delay_on && delay_off && !*delay_on && !*delay_off)
>> +		*delay_on = *delay_off = ICONNECT_BLINK_HALF_PERIOD;
>> +
>> +	switch (state) {
>> +	case GPIO_LED_NO_BLINK_LOW:
>> +	case GPIO_LED_NO_BLINK_HIGH:
>> +		orion_gpio_set_blink(gpio, 0);
>> +		gpio_set_value(gpio, state);
>> +		break;
>> +	case GPIO_LED_BLINK:
>> +		orion_gpio_set_blink(gpio, 1);
>> +	}
>> +
>> +	return 0;
>> +}
>> +
>> +static struct gpio_led_platform_data iconnect_led_data = {
>> +	.leds		= iconnect_led_pins,
>> +	.num_leds	= ARRAY_SIZE(iconnect_led_pins),
>> +	.gpio_blink_set	= iconnect_blink_set,
>> +};
>> +
>> +static struct platform_device iconnect_leds = {
>> +	.name	= "leds-gpio",
>> +	.id	= -1,
>> +	.dev	= {
>> +		.platform_data	= &iconnect_led_data,
>> +	}
>> +};
>> +
>> +static unsigned int iconnect_mpp_config[] __initdata = {
>> +	MPP12_GPIO,
>> +	MPP35_GPIO,
>> +	MPP41_GPIO,
>> +	MPP42_GPIO,
>> +	MPP43_GPIO,
>> +	MPP44_GPIO,
>> +	MPP45_GPIO,
>> +	MPP46_GPIO,
>> +	MPP47_GPIO,
>> +	MPP48_GPIO,
>> +	0
>> +};
>
> Some comments here would be helpful, 41-48 are gpio_led, what are 12 and
> 35?

They're used as buttons. See gpio-keys stuff declared later.
 
>
>> +
>> +static struct i2c_board_info __initdata iconnect_board_info[] = {
>> +	{
>> +		I2C_BOARD_INFO("lm63", 0x4c),
>> +	},
>> +};
>> +
>> +static struct mtd_partition iconnect_nand_parts[] = {
>> +	{
>> +		.name = "flash",
>> +		.offset = 0,
>> +		.size = MTDPART_SIZ_FULL,
>> +	},
>> +};
>
> I was going to suggest moving the partition definitions to here, but
> once orion_nand is fdt aware, we'll specify the partitions in the dts
> anyway.

I put the partitions in the dts as it's where it'll be defined in the end.

>
> Unless someone has an objection, I say we leave the hackish commandline
> part table in the dtb and convert it in the dtb once orion_nand had
> devicetree bindings.
>
>> +
>> +/* yikes... theses are the original input buttons */
>> +/* but I'm not convinced by the sw event choices  */
>> +static struct gpio_keys_button iconnect_buttons[] = {
>> +	{
>> +		.type		= EV_SW,
>> +		.code		= SW_LID,
>> +		.gpio		= 12,
>> +		.desc		= "Reset Button",
>> +		.active_low	= 1,
>> +		.debounce_interval = 100,
>> +	}, {
>> +		.type		= EV_SW,
>> +		.code		= SW_TABLET_MODE,
>> +		.gpio		= 35,
>> +		.desc		= "OTB Button",
>> +		.active_low	= 1,
>> +		.debounce_interval = 100,
>> +	},
>> +};
>> +
>> +static struct gpio_keys_platform_data iconnect_button_data = {
>> +	.buttons	= iconnect_buttons,
>> +	.nbuttons	= ARRAY_SIZE(iconnect_buttons),
>> +};
>> +
>> +static struct platform_device iconnect_button_device = {
>> +	.name		= "gpio-keys",
>> +	.id		= -1,
>> +	.num_resources	= 0,
>> +	.dev        = {
>> +		.platform_data  = &iconnect_button_data,
>> +	},
>> +};
>> +
>> +void __init iconnect_init(void)
>> +{
>> +	kirkwood_mpp_conf(iconnect_mpp_config);
>> +	kirkwood_nand_init(ARRAY_AND_SIZE(iconnect_nand_parts), 25);
>> +	kirkwood_i2c_init();
>> +	i2c_register_board_info(0, iconnect_board_info,
>> +		ARRAY_SIZE(iconnect_board_info));
>> +
>> +	kirkwood_ehci_init();
>> +	kirkwood_ge00_init(&iconnect_ge00_data);
>> +
>> +	platform_device_register(&iconnect_button_device);
>> +	platform_device_register(&iconnect_leds);
>> +}
>> +
>> +static int __init iconnect_pci_init(void)
>> +{
>> +	if (of_machine_is_compatible("iomega,iconnect"))
>> +		kirkwood_pcie_init(KW_PCIE0);
>> +	return 0;
>> +}
>> +subsys_initcall(iconnect_pci_init);
>> +
>> Index: arm-soc/arch/arm/boot/dts/kirkwood-iconnect.dts
>> ===================================================================
>> --- /dev/null	1970-01-01 00:00:00.000000000 +0000
>> +++ arm-soc/arch/arm/boot/dts/kirkwood-iconnect.dts	2012-04-16 08:59:12.033376388 +0200
>> @@ -0,0 +1,26 @@
>> +/dts-v1/;
>> +
>> +/include/ "kirkwood.dtsi"
>> +
>> +/ {
>> +	model = "Iomega Iconnect";
>> +	compatible = "iomega,iconnect", "mrvl,kirkwood-88f6281", "mrvl,kirkwood";
>
> This should be "iom,iconnect-MODELNUM", "iom,iconnect", "mrvl,...
>
> And, correct elsewhere as needed.

ok. fixed.

>
>> +
>> +	memory {
>> +		device_type = "memory";
>> +		reg = <0x00000000 0x10000000>;
>> +	};
>> +
>> +	chosen {
>> +		bootargs = "console=ttyS0,115200n8 earlyprintk mtdparts=orion_nand:0xc0000 at 0x0(uboot),0x20000 at 0xa0000(env),0x300000 at 0x100000(zImage),0x300000 at 0x540000(initrd),0x1f400000 at 0x980000(boot)";
>> +		linux,initrd-start = <0x4500040>;
>> +		linux,initrd-end   = <0x4800000>;
>
> does this need to be specified?  why doesn't the bootloader pass the
> initrd properly?

oh, I does give the right values. I was wondering if the system was
still booting without CONFIG_ARM_ATAG_DTB_COMPAT set and with theses
values set here.

Thanks,
Arnaud