Hi Massimo,
On 10/5/2026 1:44 PM, Massimo Pegorer wrote:
> Hi Jonas,
>
> Il giorno dom 4 ott 2026 alle ore 20:54 Jonas Karlman
> <[email protected]> ha scritto:
>>
>> Hi Massimo,
>>
>> On 10/2/2026 6:36 PM, Massimo Pegorer wrote:
>>> The configuration options SPL_BSS_START_ADDR and SPL_BSS_MAX_SIZE are
>>> guarded by SPL_HAS_BSS_LINKER_SECTION and SPL_BSS_LIMIT options in the
>>> common/spl/Kconfig file where they are originally defined. Addition of
>>> a second config statement without 'depends on' guard causes these two
>>> options to be put in the configuration even when the guard is not
>>> selected, which is unneeded and sounds wrong.
>>>
>>> Signed-off-by: Massimo Pegorer <[email protected]>
>>> ---
>>> arch/arm/mach-rockchip/Kconfig | 2 ++
>>> 1 file changed, 2 insertions(+)
>>>
>>> diff --git a/arch/arm/mach-rockchip/Kconfig b/arch/arm/mach-rockchip/Kconfig
>>> index 1a2e7847c9e..5ed774159be 100644
>>> --- a/arch/arm/mach-rockchip/Kconfig
>>> +++ b/arch/arm/mach-rockchip/Kconfig
>>> @@ -836,11 +836,13 @@ config SPL_HAS_BSS_LINKER_SECTION
>>> default y if ARM64
>>>
>>> config SPL_BSS_START_ADDR
>>> + depends on SPL_HAS_BSS_LINKER_SECTION
>>
>> This change is causing make to prompt for SPL_BSS_MAX_SIZE when building
>> e.g. tinker-s-rk3288_defconfig.
>>
>> Maximum size in memory allocated to the SPL BSS (SPL_BSS_MAX_SIZE) [] (NEW)
>>
>>> default 0x63f80000 if SPL_TEXT_BASE = 0x60000000
>>> default 0x43f80000 if SPL_TEXT_BASE = 0x40000000
>>> default 0x03f80000 if SPL_TEXT_BASE = 0x00000000
>>>
>>> config SPL_BSS_MAX_SIZE
>>> + depends on SPL_BSS_LIMIT
>>> default 0x8000 if SPL_BSS_START_ADDR = 0x63f80000
>>> default 0x8000 if SPL_BSS_START_ADDR = 0x43f80000
>>> default 0x8000 if SPL_BSS_START_ADDR = 0x03f80000
>>
>> Possible adding something like following?
>>
>> default 0x8000 if !SPL_HAS_BSS_LINKER_SECTION
>
> What about selecting SPL_BSS_MAX_SIZE default based on SPL_TEXT_BASE instead
> of
> SPL_BSS_START_ADDR?
Unsure, the original intent of the BSS_START_ADDR dependency was to only
affect boards that had transitioned to use the common bss and stack
addresses introduced in the commit 008ba0d56d00 ("rockchip: Add common
default bss and stack addresses").
Not fully sure about current migration state for AArch64 targets, I do
not have any rk3368 boards so never tried to migrate targets for that
SoC to use these common addresses. Most/all other AArch64 targets should
now use these defaults so it is likely these checks can be simplified.
>
> This seems to me more straightforward. As this configuration states, there are
> cases where SPL_HAS_BSS_LINKER_SECTION (and thus SPL_BSS_START_ADDR) is not
> needed, but SPL_BSS_MAX_SIZE is. Therefore, I would avoid a default definition
> based on SPL_BSS_START_ADDR.
It is likely these complex conditions have overplayed its need now that
most AArch64 targets have been migrated to use these defaults.
>
> On the other hand, I've just noticed that the value is the same in the three
> SPL_BSS_MAX_SIZE default statements, so it would probably be even clearer to
> have something like (not tested yet):
>
> config SPL_BSS_MAX_SIZE
> default 0x8000 if SPL_BSS_LIMIT
>
> What do you think? Thanks for your revision.
Please check current values used for ARCH_ROCKCHIP boards using the
qconfig.py tool. If all boards use this value, we can likely just move
this default to normal "config SPL_BSS_MAX_SIZE" in common/spl/Kconfig.
default 0x8000 if ARCH_ROCKCHIP
One goal should be to continue migrate as many boards/SoCs to use the
common bss and stack addresses. It should mostly be older ARMv7 SoCs
that I do not have access to that never have been migrated. I do have
access to rk322x boxes, and have some early work for that SoC, but too
much other work to try and upstream before circling back to rk322x :-)
Regards,
Jonas
>
> Regards,
> Massimo
>
>>
>> Regards,
>> Jonas
