[U-Boot] [PATCH 0/2] arm: socfpga: Enable EFI support for DE0-nano-SoC disto boot

[U-Boot] [PATCH 0/2] arm: socfpga: Enable EFI support for DE0-nano-SoC disto boot

[Frank Kunz]

This enables EFI support for DE0-nano-SoC distro boot.

The patches will apply on top of Dalon Westergreens "[PATCH v3 0/8] arm: socfpga: Move to using distro boot" series.

Tested on openSUSE openSUSE Tumbleweed ARM JeOS image.

Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>

Frank Kunz (2):
  arm: socfpga: Enable abort for DE-nano-SoC SPL uboot load from MMC
  arm: socfpga: Enable part command and EFI partition support for
    DE0-nano-SoC

 configs/socfpga_de0_nano_soc_defconfig | 7 ++++---
 include/configs/socfpga_de0_nano_soc.h | 2 ++
 2 files changed, 6 insertions(+), 3 deletions(-)

--
2.11.0

[U-Boot] [PATCH 1/2] arm: socfpga: Enable abort for DE-nano-SoC SPL uboot load from MMC

[Frank Kunz]

This allows the SPL to scan the MMC for a valid uboot image on a second
sector location defined by CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR when
the default location "CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR + first
partition offset" does not contain a valid uboot image.

Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
---
:100644 100644 dd5933d43c... bd3e83ddea... M	include/configs/socfpga_de0_nano_soc.h
 include/configs/socfpga_de0_nano_soc.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/configs/socfpga_de0_nano_soc.h b/include/configs/socfpga_de0_nano_soc.h
index dd5933d43c..bd3e83ddea 100644
--- a/include/configs/socfpga_de0_nano_soc.h
+++ b/include/configs/socfpga_de0_nano_soc.h
@@ -12,6 +12,8 @@
 #define CONFIG_FAT_WRITE
 #define CONFIG_HW_WATCHDOG
 
+#define CONFIG_SPL_ABORT_ON_RAW_IMAGE
+
 /* Memory configurations */
 #define PHYS_SDRAM_1_SIZE		0x40000000	/* 1GiB */
 
-- 
2.11.0

[U-Boot] [PATCH 2/2] arm: socfpga: Enable part command and EFI partition support for DE0-nano-SoC

[Frank Kunz]

Those features are used by distro boot with efi boot.

Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
---
:100644 100644 b122135690... c22b54b48d... M	configs/socfpga_de0_nano_soc_defconfig
 configs/socfpga_de0_nano_soc_defconfig | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/configs/socfpga_de0_nano_soc_defconfig b/configs/socfpga_de0_nano_soc_defconfig
index b122135690..c22b54b48d 100644
--- a/configs/socfpga_de0_nano_soc_defconfig
+++ b/configs/socfpga_de0_nano_soc_defconfig
@@ -4,11 +4,11 @@ CONFIG_SYS_MALLOC_F_LEN=0x2000
 CONFIG_TARGET_SOCFPGA_TERASIC_DE0_NANO=y
 CONFIG_SPL_STACK_R_ADDR=0x00800000
 CONFIG_DEFAULT_DEVICE_TREE="socfpga_cyclone5_de0_nano_soc"
-CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"
 CONFIG_FIT=y
 CONFIG_SYS_CONSOLE_IS_IN_ENV=y
 CONFIG_SYS_CONSOLE_OVERWRITE_ROUTINE=y
 CONFIG_SYS_CONSOLE_ENV_OVERWRITE=y
+CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"
 CONFIG_VERSION_VARIABLE=y
 # CONFIG_DISPLAY_BOARDINFO is not set
 CONFIG_SPL=y
@@ -21,6 +21,7 @@ CONFIG_CMD_ASKENV=y
 CONFIG_CMD_GREPENV=y
 # CONFIG_CMD_FLASH is not set
 CONFIG_CMD_MMC=y
+CONFIG_CMD_PART=y
 CONFIG_CMD_SF=y
 CONFIG_CMD_SPI=y
 CONFIG_CMD_I2C=y
@@ -37,6 +38,8 @@ CONFIG_CMD_EXT4_WRITE=y
 CONFIG_CMD_FAT=y
 CONFIG_CMD_FS_GENERIC=y
 CONFIG_CMD_UBI=y
+CONFIG_EFI_PARTITION=y
+# CONFIG_SPL_EFI_PARTITION is not set
 CONFIG_SPL_DM=y
 CONFIG_DFU_MMC=y
 CONFIG_DM_GPIO=y
@@ -59,5 +62,3 @@ CONFIG_G_DNL_MANUFACTURER="terasic"
 CONFIG_G_DNL_VENDOR_NUM=0x0525
 CONFIG_G_DNL_PRODUCT_NUM=0xa4a5
 CONFIG_USE_TINY_PRINTF=y
-CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_USE_PARTITION_TYPE=y
-CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_PARTITION_TYPE=0xa2
\ No newline at end of file
-- 
2.11.0

[U-Boot] [PATCH 1/2] arm: socfpga: Enable abort for DE-nano-SoC SPL uboot load from MMC

[Marek Vasut]

On 03/05/2017 01:54 PM, Frank Kunz wrote:
> This allows the SPL to scan the MMC for a valid uboot image on a second
> sector location defined by CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR when
> the default location "CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR + first
> partition offset" does not contain a valid uboot image.
>
> Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
> ---
> :100644 100644 dd5933d43c... bd3e83ddea... M	include/configs/socfpga_de0_nano_soc.h
>  include/configs/socfpga_de0_nano_soc.h | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/include/configs/socfpga_de0_nano_soc.h b/include/configs/socfpga_de0_nano_soc.h
> index dd5933d43c..bd3e83ddea 100644
> --- a/include/configs/socfpga_de0_nano_soc.h
> +++ b/include/configs/socfpga_de0_nano_soc.h
> @@ -12,6 +12,8 @@
>  #define CONFIG_FAT_WRITE
>  #define CONFIG_HW_WATCHDOG
>
> +#define CONFIG_SPL_ABORT_ON_RAW_IMAGE
> +
>  /* Memory configurations */
>  #define PHYS_SDRAM_1_SIZE		0x40000000	/* 1GiB */
>
>
This looks like a horrid hack , didn't Dalon add support into U-Boot SPL 
to scan all partitions for the final u-boot image ?

[U-Boot] [PATCH 2/2] arm: socfpga: Enable part command and EFI partition support for DE0-nano-SoC

[Marek Vasut]

On 03/05/2017 01:54 PM, Frank Kunz wrote:
> Those features are used by distro boot with efi boot.
>
> Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
> ---
> :100644 100644 b122135690... c22b54b48d... M	configs/socfpga_de0_nano_soc_defconfig
>  configs/socfpga_de0_nano_soc_defconfig | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
>
> diff --git a/configs/socfpga_de0_nano_soc_defconfig b/configs/socfpga_de0_nano_soc_defconfig
> index b122135690..c22b54b48d 100644
> --- a/configs/socfpga_de0_nano_soc_defconfig
> +++ b/configs/socfpga_de0_nano_soc_defconfig
> @@ -4,11 +4,11 @@ CONFIG_SYS_MALLOC_F_LEN=0x2000
>  CONFIG_TARGET_SOCFPGA_TERASIC_DE0_NANO=y
>  CONFIG_SPL_STACK_R_ADDR=0x00800000
>  CONFIG_DEFAULT_DEVICE_TREE="socfpga_cyclone5_de0_nano_soc"
> -CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"
>  CONFIG_FIT=y
>  CONFIG_SYS_CONSOLE_IS_IN_ENV=y
>  CONFIG_SYS_CONSOLE_OVERWRITE_ROUTINE=y
>  CONFIG_SYS_CONSOLE_ENV_OVERWRITE=y
> +CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"

Is this move needed ? (no, it's not)

>  CONFIG_VERSION_VARIABLE=y
>  # CONFIG_DISPLAY_BOARDINFO is not set
>  CONFIG_SPL=y
> @@ -21,6 +21,7 @@ CONFIG_CMD_ASKENV=y
>  CONFIG_CMD_GREPENV=y
>  # CONFIG_CMD_FLASH is not set
>  CONFIG_CMD_MMC=y
> +CONFIG_CMD_PART=y
>  CONFIG_CMD_SF=y
>  CONFIG_CMD_SPI=y
>  CONFIG_CMD_I2C=y
> @@ -37,6 +38,8 @@ CONFIG_CMD_EXT4_WRITE=y
>  CONFIG_CMD_FAT=y
>  CONFIG_CMD_FS_GENERIC=y
>  CONFIG_CMD_UBI=y
> +CONFIG_EFI_PARTITION=y
> +# CONFIG_SPL_EFI_PARTITION is not set
>  CONFIG_SPL_DM=y
>  CONFIG_DFU_MMC=y
>  CONFIG_DM_GPIO=y
> @@ -59,5 +62,3 @@ CONFIG_G_DNL_MANUFACTURER="terasic"
>  CONFIG_G_DNL_VENDOR_NUM=0x0525
>  CONFIG_G_DNL_PRODUCT_NUM=0xa4a5
>  CONFIG_USE_TINY_PRINTF=y
> -CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_USE_PARTITION_TYPE=y
> -CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_PARTITION_TYPE=0xa2

+CC Dalon ...

> \ No newline at end of file
>

[U-Boot] [PATCH 1/2] arm: socfpga: Enable abort for DE-nano-SoC SPL uboot load from MMC

[Dalon Westergreen]

On Sun, 2017-03-05 at 18:16 +0100, Marek Vasut wrote:
> On 03/05/2017 01:54 PM, Frank Kunz wrote:
> > 
> > This allows the SPL to scan the MMC for a valid uboot image on a second
> > sector location defined by CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR when
> > the default location "CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR + first
> > partition offset" does not contain a valid uboot image.
> > 
> > Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
> > ---
> > :100644 100644 dd5933d43c... bd3e83ddea... M	include/configs/socfpga_
> > de0_nano_soc.h
> >  include/configs/socfpga_de0_nano_soc.h | 2 ++
> >  1 file changed, 2 insertions(+)
> > 
> > diff --git a/include/configs/socfpga_de0_nano_soc.h
> > b/include/configs/socfpga_de0_nano_soc.h
> > index dd5933d43c..bd3e83ddea 100644
> > --- a/include/configs/socfpga_de0_nano_soc.h
> > +++ b/include/configs/socfpga_de0_nano_soc.h
> > @@ -12,6 +12,8 @@
> >  #define CONFIG_FAT_WRITE
> >  #define CONFIG_HW_WATCHDOG
> > 
> > +#define CONFIG_SPL_ABORT_ON_RAW_IMAGE
> > +
> >  /* Memory configurations */
> >  #define PHYS_SDRAM_1_SIZE		0x40000000	/* 1GiB */
> > 
> > 
> This looks like a horrid hack , didn't Dalon add support into U-Boot SPL 
> to scan all partitions for the final u-boot image ?

I did.  The 0xA2 partition can be located on any partition number now.  I
believe Frank places the SPL image at a 64KB offset from the start of the
sdcard to skip the MBR/GPT.  But per our previous debate, it is not my
favored implementation as it relies on the bootrom to load the SPL once.

Also, i mentioned before that i would like to hold off on this patch set and
let Ley Foon et al finish the A10 stuff.

--dalon

> _______________________________________________
> U-Boot mailing list
> U-Boot at lists.denx.de
> http://lists.denx.de/listinfo/u-boot

[U-Boot] [PATCH 1/2] arm: socfpga: Enable abort for DE-nano-SoC SPL uboot load from MMC

[Marek Vasut]

On 03/05/2017 06:36 PM, Dalon Westergreen wrote:
> On Sun, 2017-03-05 at 18:16 +0100, Marek Vasut wrote:
>> On 03/05/2017 01:54 PM, Frank Kunz wrote:
>>>
>>> This allows the SPL to scan the MMC for a valid uboot image on a second
>>> sector location defined by CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR when
>>> the default location "CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR + first
>>> partition offset" does not contain a valid uboot image.
>>>
>>> Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
>>> ---
>>> :100644 100644 dd5933d43c... bd3e83ddea... M	include/configs/socfpga_
>>> de0_nano_soc.h
>>>  include/configs/socfpga_de0_nano_soc.h | 2 ++
>>>  1 file changed, 2 insertions(+)
>>>
>>> diff --git a/include/configs/socfpga_de0_nano_soc.h
>>> b/include/configs/socfpga_de0_nano_soc.h
>>> index dd5933d43c..bd3e83ddea 100644
>>> --- a/include/configs/socfpga_de0_nano_soc.h
>>> +++ b/include/configs/socfpga_de0_nano_soc.h
>>> @@ -12,6 +12,8 @@
>>>  #define CONFIG_FAT_WRITE
>>>  #define CONFIG_HW_WATCHDOG
>>>
>>> +#define CONFIG_SPL_ABORT_ON_RAW_IMAGE
>>> +
>>>  /* Memory configurations */
>>>  #define PHYS_SDRAM_1_SIZE		0x40000000	/* 1GiB */
>>>
>>>
>> This looks like a horrid hack , didn't Dalon add support into U-Boot SPL
>> to scan all partitions for the final u-boot image ?
>
> I did.  The 0xA2 partition can be located on any partition number now.  I
> believe Frank places the SPL image at a 64KB offset from the start of the
> sdcard to skip the MBR/GPT.  But per our previous debate, it is not my
> favored implementation as it relies on the bootrom to load the SPL once.
>
> Also, i mentioned before that i would like to hold off on this patch set and
> let Ley Foon et al finish the A10 stuff.

OK, although I think the A10 will take a bit longer, so there's no point 
in blocking this simple(r) patchset ... just don't wind up in deadlock ...

[U-Boot] [PATCH 2/2] arm: socfpga: Enable part command and EFI partition support for DE0-nano-SoC

[Frank Kunz]

Am 05.03.2017 um 18:17 schrieb Marek Vasut:
> On 03/05/2017 01:54 PM, Frank Kunz wrote:
>> Those features are used by distro boot with efi boot.
>>
>> Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
>> ---
>> :100644 100644 b122135690... c22b54b48d... M   
>> configs/socfpga_de0_nano_soc_defconfig
>>  configs/socfpga_de0_nano_soc_defconfig | 7 ++++---
>>  1 file changed, 4 insertions(+), 3 deletions(-)
>>
>> diff --git a/configs/socfpga_de0_nano_soc_defconfig
>> b/configs/socfpga_de0_nano_soc_defconfig
>> index b122135690..c22b54b48d 100644
>> --- a/configs/socfpga_de0_nano_soc_defconfig
>> +++ b/configs/socfpga_de0_nano_soc_defconfig
>> @@ -4,11 +4,11 @@ CONFIG_SYS_MALLOC_F_LEN=0x2000
>>  CONFIG_TARGET_SOCFPGA_TERASIC_DE0_NANO=y
>>  CONFIG_SPL_STACK_R_ADDR=0x00800000
>>  CONFIG_DEFAULT_DEVICE_TREE="socfpga_cyclone5_de0_nano_soc"
>> -CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"
>>  CONFIG_FIT=y
>>  CONFIG_SYS_CONSOLE_IS_IN_ENV=y
>>  CONFIG_SYS_CONSOLE_OVERWRITE_ROUTINE=y
>>  CONFIG_SYS_CONSOLE_ENV_OVERWRITE=y
>> +CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"
> 
> Is this move needed ? (no, it's not)

It is not needed, but "make savedefconfig" decided to do that. I do not
know how to overcome it.

> 
>>  CONFIG_VERSION_VARIABLE=y
>>  # CONFIG_DISPLAY_BOARDINFO is not set
>>  CONFIG_SPL=y
>> @@ -21,6 +21,7 @@ CONFIG_CMD_ASKENV=y
>>  CONFIG_CMD_GREPENV=y
>>  # CONFIG_CMD_FLASH is not set
>>  CONFIG_CMD_MMC=y
>> +CONFIG_CMD_PART=y
>>  CONFIG_CMD_SF=y
>>  CONFIG_CMD_SPI=y
>>  CONFIG_CMD_I2C=y
>> @@ -37,6 +38,8 @@ CONFIG_CMD_EXT4_WRITE=y
>>  CONFIG_CMD_FAT=y
>>  CONFIG_CMD_FS_GENERIC=y
>>  CONFIG_CMD_UBI=y
>> +CONFIG_EFI_PARTITION=y
>> +# CONFIG_SPL_EFI_PARTITION is not set
>>  CONFIG_SPL_DM=y
>>  CONFIG_DFU_MMC=y
>>  CONFIG_DM_GPIO=y
>> @@ -59,5 +62,3 @@ CONFIG_G_DNL_MANUFACTURER="terasic"
>>  CONFIG_G_DNL_VENDOR_NUM=0x0525
>>  CONFIG_G_DNL_PRODUCT_NUM=0xa4a5
>>  CONFIG_USE_TINY_PRINTF=y
>> -CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_USE_PARTITION_TYPE=y
>> -CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_PARTITION_TYPE=0xa2
> 
> +CC Dalon ...

Similar case here. "make savedefconfig" found that those two parameters
are already set by arch/arm/mach-socfpga/Kconfig and decided to remove
them here, I guess.

> 
>> \ No newline at end of file
>>
> 
> 

[U-Boot] [PATCH 1/2] arm: socfpga: Enable abort for DE-nano-SoC SPL uboot load from MMC

[Frank Kunz]

Am 05.03.2017 um 18:36 schrieb Dalon Westergreen:
> On Sun, 2017-03-05 at 18:16 +0100, Marek Vasut wrote:
>> On 03/05/2017 01:54 PM, Frank Kunz wrote:
>>>
>>> This allows the SPL to scan the MMC for a valid uboot image on a second
>>> sector location defined by CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR when
>>> the default location "CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR + first
>>> partition offset" does not contain a valid uboot image.
>>>
>>> Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
>>> ---
>>> :100644 100644 dd5933d43c... bd3e83ddea... M	include/configs/socfpga_
>>> de0_nano_soc.h
>>>  include/configs/socfpga_de0_nano_soc.h | 2 ++
>>>  1 file changed, 2 insertions(+)
>>>
>>> diff --git a/include/configs/socfpga_de0_nano_soc.h
>>> b/include/configs/socfpga_de0_nano_soc.h
>>> index dd5933d43c..bd3e83ddea 100644
>>> --- a/include/configs/socfpga_de0_nano_soc.h
>>> +++ b/include/configs/socfpga_de0_nano_soc.h
>>> @@ -12,6 +12,8 @@
>>>  #define CONFIG_FAT_WRITE
>>>  #define CONFIG_HW_WATCHDOG
>>>
>>> +#define CONFIG_SPL_ABORT_ON_RAW_IMAGE
>>> +
>>>  /* Memory configurations */
>>>  #define PHYS_SDRAM_1_SIZE		0x40000000	/* 1GiB */
>>>
>>>
>> This looks like a horrid hack , didn't Dalon add support into U-Boot SPL 
>> to scan all partitions for the final u-boot image ?
> 
> I did.  The 0xA2 partition can be located on any partition number now.  I
> believe Frank places the SPL image at a 64KB offset from the start of the
> sdcard to skip the MBR/GPT.  But per our previous debate, it is not my
> favored implementation as it relies on the bootrom to load the SPL once.

I put the u-boot at 4x64kiB (sector 0x200) offset on the sdcard. With
CONFIG_SPL_ABORT_ON_RAW_IMAGE config enabled the spl scans first for the
configured partition in the partition table as implemented by Dalon, if
that fails it tries to find the u-boot at sector 0x200 (as configured by
CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_SECTOR). Without that config the spl
endless retries to find u-boot in the configured partition, which is not
there when using GPT.

> 
> Also, i mentioned before that i would like to hold off on this patch set and
> let Ley Foon et al finish the A10 stuff.
> 
> --dalon
> 
>> _______________________________________________
>> U-Boot mailing list
>> U-Boot at lists.denx.de
>> http://lists.denx.de/listinfo/u-boot
> 

[U-Boot] [PATCH 2/2] arm: socfpga: Enable part command and EFI partition support for DE0-nano-SoC

[Marek Vasut]

On 03/05/2017 07:37 PM, Frank Kunz wrote:
> Am 05.03.2017 um 18:17 schrieb Marek Vasut:
>> On 03/05/2017 01:54 PM, Frank Kunz wrote:
>>> Those features are used by distro boot with efi boot.
>>>
>>> Signed-off-by: Frank Kunz <mailinglists@kunz-im-inter.net>
>>> ---
>>> :100644 100644 b122135690... c22b54b48d... M
>>> configs/socfpga_de0_nano_soc_defconfig
>>>  configs/socfpga_de0_nano_soc_defconfig | 7 ++++---
>>>  1 file changed, 4 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/configs/socfpga_de0_nano_soc_defconfig
>>> b/configs/socfpga_de0_nano_soc_defconfig
>>> index b122135690..c22b54b48d 100644
>>> --- a/configs/socfpga_de0_nano_soc_defconfig
>>> +++ b/configs/socfpga_de0_nano_soc_defconfig
>>> @@ -4,11 +4,11 @@ CONFIG_SYS_MALLOC_F_LEN=0x2000
>>>  CONFIG_TARGET_SOCFPGA_TERASIC_DE0_NANO=y
>>>  CONFIG_SPL_STACK_R_ADDR=0x00800000
>>>  CONFIG_DEFAULT_DEVICE_TREE="socfpga_cyclone5_de0_nano_soc"
>>> -CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"
>>>  CONFIG_FIT=y
>>>  CONFIG_SYS_CONSOLE_IS_IN_ENV=y
>>>  CONFIG_SYS_CONSOLE_OVERWRITE_ROUTINE=y
>>>  CONFIG_SYS_CONSOLE_ENV_OVERWRITE=y
>>> +CONFIG_DEFAULT_FDT_FILE="socfpga_cyclone5_de0_nano_soc.dtb"
>>
>> Is this move needed ? (no, it's not)
>
> It is not needed, but "make savedefconfig" decided to do that. I do not
> know how to overcome it.

Try git checkout -p to selectively clear bits from the changes.
Or git add -p ...

>>
>>>  CONFIG_VERSION_VARIABLE=y
>>>  # CONFIG_DISPLAY_BOARDINFO is not set
>>>  CONFIG_SPL=y
>>> @@ -21,6 +21,7 @@ CONFIG_CMD_ASKENV=y
>>>  CONFIG_CMD_GREPENV=y
>>>  # CONFIG_CMD_FLASH is not set
>>>  CONFIG_CMD_MMC=y
>>> +CONFIG_CMD_PART=y
>>>  CONFIG_CMD_SF=y
>>>  CONFIG_CMD_SPI=y
>>>  CONFIG_CMD_I2C=y
>>> @@ -37,6 +38,8 @@ CONFIG_CMD_EXT4_WRITE=y
>>>  CONFIG_CMD_FAT=y
>>>  CONFIG_CMD_FS_GENERIC=y
>>>  CONFIG_CMD_UBI=y
>>> +CONFIG_EFI_PARTITION=y
>>> +# CONFIG_SPL_EFI_PARTITION is not set
>>>  CONFIG_SPL_DM=y
>>>  CONFIG_DFU_MMC=y
>>>  CONFIG_DM_GPIO=y
>>> @@ -59,5 +62,3 @@ CONFIG_G_DNL_MANUFACTURER="terasic"
>>>  CONFIG_G_DNL_VENDOR_NUM=0x0525
>>>  CONFIG_G_DNL_PRODUCT_NUM=0xa4a5
>>>  CONFIG_USE_TINY_PRINTF=y
>>> -CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_USE_PARTITION_TYPE=y
>>> -CONFIG_SYS_MMCSD_RAW_MODE_U_BOOT_PARTITION_TYPE=0xa2
>>
>> +CC Dalon ...
>
> Similar case here. "make savedefconfig" found that those two parameters
> are already set by arch/arm/mach-socfpga/Kconfig and decided to remove
> them here, I guess.
>
>>
>>> \ No newline at end of file
>>>
>>
>>
>