[PATCH v8 4/4] efi_vars: Implement SPI Flash store

[Michal Simek]

On 2/15/26 09:48, Heinrich Schuchardt wrote:
> On 2/11/26 16:56, Michal Simek wrote:
>> From: Shantur Rathore <i at shantur.com>
>>
>> Currently U-Boot uses ESP as storage for EFI variables.
>> Devices with SPI Flash are used for storing environment with this
>> commit we allow EFI variables to be stored on SPI Flash.
>>
>> https://github.com/rhboot/efivar
>> is not updated to support this location that's why you can mimic it by
>> running:
>> dd if=/sys/firmware/efi/efivars/VarToFile-b2ac5fc9-92b7-4acd-aeac-11e818c3130c \
>>     of=/tmp/vars skip=4 bs=1
>> flashcp /tmp/vars /dev/mtdX
>>
>> where mtdX should match location defined by
>> CONFIG_EFI_VARIABLE_SF_OFFSET/CONFIG_EFI_VAR_BUF_SIZE.
>>
>> Signed-off-by: Shantur Rathore <i at shantur.com>
>> Signed-off-by: Michal Simek <michal.simek at amd.com>
>> Tested-by: Neil Armstrong <neil.armstrong at linaro.org> # on AML-S905D3-CC
>> Acked-by: Ilias Apalodimas <ilias.apalodimas at linaro.org>
>>
>> ---
>>
>> Changes in v8:
>> - Add missing EFI_VARIABLE_SF_STORE dependency to EFI_VARIABLE_SF_DEVICE_INDEX
>>
>> Changes in v7:
>> - sed -i 's/efi_var_from/efi_var_from_storage/g'
>>
>> Changes in v6:
>> - sed -i 's/efi_var_read/efi_var_from/g'
>> - sed -i 's/efi_var_write/efi_var_to_storage/g'
>>
>> Changes in v4:
>> - Extend Kconfig description
>> - Extend commit message and describe efivar missing part
>> - use unify methods for reading/writing variable
>>
>> Changes in v3:
>> - Fixed compiler warnings.
>>
>> Changes in v2:
>> - Refactored efi_var_file to move common parts out as requested
>> - Changed ifdefs to use CONFIG_IS_DEFINED
>> - Fixed typos
>>
>>   lib/efi_loader/Kconfig      | 35 +++++++++++++-
>>   lib/efi_loader/Makefile     |  1 +
>>   lib/efi_loader/efi_var_sf.c | 92 +++++++++++++++++++++++++++++++++++++
>>   3 files changed, 127 insertions(+), 1 deletion(-)
>>   create mode 100644 lib/efi_loader/efi_var_sf.c
>>
>> diff --git a/lib/efi_loader/Kconfig b/lib/efi_loader/Kconfig
>> index 579eed658801..0f6f927ddaf1 100644
>> --- a/lib/efi_loader/Kconfig
>> +++ b/lib/efi_loader/Kconfig
>> @@ -124,6 +124,24 @@ config EFI_VARIABLE_FILE_STORE
>>         Select this option if you want non-volatile UEFI variables to be
>>         stored as file /ubootefi.var on the EFI system partition.
>> +config EFI_VARIABLE_SF_STORE
>> +    bool "Store non-volatile UEFI variables in SPI Flash"
>> +    depends on SPI_FLASH
>> +    help
>> +      Select this option if you want non-volatile UEFI variables to be
>> +      stored in SPI Flash.
>> +
>> +      Define CONFIG_EFI_VARIABLE_SF_OFFSET as offset in SPI Flash to use as
>> +      the storage for variables. CONFIG_EFI_VAR_BUF_SIZE defines the space
>> +      needed.
>> +
>> +      Note that SPI Flash devices have a limited number of program/erase
>> +      cycles. Frequent updates to UEFI variables may cause excessive wear
>> +      and can permanently damage the flash device, particularly on SPI NAND
>> +      or low-end SPI NOR parts without wear leveling. This option should be
>> +      used with care on such systems, and is not recommended for platforms
>> +      where UEFI variables are updated frequently.
>> +
>>   config EFI_MM_COMM_TEE
>>       bool "UEFI variables storage service via the trusted world"
>>       depends on OPTEE
>> @@ -153,7 +171,7 @@ endchoice
>>   config EFI_RT_VOLATILE_STORE
>>       bool "Allow variable runtime services in volatile storage (e.g RAM)"
>> -    depends on EFI_VARIABLE_FILE_STORE
>> +    depends on EFI_VARIABLE_FILE_STORE || EFI_VARIABLE_SF_STORE
> 
> Hello Michal,
> 
> If the backend store is SPI flash, we should not publish the variable 
> "RTStorageVolatile" at runtime as currently defined.
> 
> For the background see this commit for the efivar library:
> 
> https://github.com/rhboot/efivar/commit/68daa04654acbe1bbaa17ebfc23c371b39e69c6b
> 
> The first three patches look correct to me and I will add them to a merge 
> request for efi-next.

Do I read it that when I don't allow to select EFI_RT_VOLATILE_STORE you will be 
fine with the patch itself?


> The usage of efi_var_mem_ins() and efi_var_mem_del() when SetVariable() does not 
> change anything is not flash-friendly. Michal, do you want to have a look at 
> this area?

That's known limitation describe in EFI_VARIABLE_SF_STORE Kconfig Note.

Thanks,
Michal