[PATCH 00/12] rockchip: rk3399: Migrate to use TPL on bob and kevin

[Jonas Karlman]

Hi Simon,

On 2024-12-13 04:09, Simon Glass wrote:
> Hi Jonas,
> 
> On Thu, 12 Dec 2024 at 16:57, Jonas Karlman <jonas at kwiboo.se> wrote:
>>
>> This is a continuation of an RFC [1] sent back in August.
>>
>> All Rockchip aarch64 boards, beside bob and kevin, use TPL to initialize
>> DRAM and SPL to load FIT with U-Boot proper and TF-A.
>>
>> Bob and kevin currently does not use TPL and instead initialize DRAM
>> in SPL and directly after tries to load a FIT with only U-Boot proper.
>>
>> This series migrate bob and kevin to also use TPL to uniform TPL/SPL use
>> for all Rockchip aarch64 boards.
>>
>> Features and changes in this series:
>> - Use of Rockchip aarch64 common bss and stack addresses
>> - SHA256 checksum/signature check of FIT images in SPL
>> - Use of dcache to speed up checksum check in SPL
>> - SDMA use for eMMC access to speed up read/write in U-Boot proper
>> - FIT location moved from 0x40000 to 0xE0000 in SPI flash
>>
>> Changes sinc RFC:
>> - Include pinctrl props and enable regulators based on DT in SPL
>> - Drop bob and kevin board specific code from common rk3399.c
>>
>> Please note that I do not have any bob or kevin so this has only been
>> compile tested and the original RFC [1] was runtime tested by Simon.
>>
>> [1] https://patchwork.ozlabs.org/patch/1967517/
>>
>> Jonas Karlman (12):
>>   rockchip: rk3399: Fix TPL build of bob and kevin
>>   rockchip: rk3399-gru: Use SYS_SPI_U_BOOT_OFFS value in offset prop
>>   rockchip: rk3399-gru: Move SPI flash payload offset for bob and kevin
>>   rockchip: rk3399-gru: Fix include of TPL in u-boot.rom image
>>   rockchip: rk3399-gru: Use TPL with common bss and stack addresses
>>   rockchip: rk3399-gru: Include binman generated FIT in u-boot.rom image
>>   rockchip: rk3399: Drop common Kconfig options already implied
>>   rockchip: rk3399-gru: Remove unused nodes from xPL control FDT
>>   rockchip: rk3399-gru: Enable dcache and signature validation in SPL
>>   rockchip: rk3399-gru: Use SDMA mode to read from eMMC
>>   rockchip: rk3399-gru: Include pinctrl and regulators in SPL
>>   rockchip: rk3399: Drop unneeded bob and kevin board specific code
>>
>>  arch/arm/dts/rk3399-gru-u-boot.dtsi    | 29 ++++++++++++++++++++++++--
>>  arch/arm/dts/rk3399-u-boot.dtsi        | 14 +++++++++++--
>>  arch/arm/mach-rockchip/rk3399/Kconfig  | 12 -----------
>>  arch/arm/mach-rockchip/rk3399/rk3399.c | 23 +-------------------
>>  configs/chromebook_bob_defconfig       | 23 ++++++--------------
>>  configs/chromebook_kevin_defconfig     | 23 ++++++--------------
>>  6 files changed, 52 insertions(+), 72 deletions(-)
> 
> Thank you for doing this!
> 
> I'm not sure if you have access to your own denx tree. If so, kevin
> and bob are in the sjg lab. If not, you can sign up at sjg.u-boot.org
> to get access to my tree.

I do not have access to my own tree.

> 
> This series eems to work for me, except for the final commit, which
> hangs with no output.

Strange, guess we can just skip the last patch, should I send a v2
without the last patch?

I wonder if it is not setting io_vsel or the regulators that is causing
an issue. Setting io_vsel is probably delayed until U-Boot proper
instead of being set very early in SPL (due to no IO-domain driver in
SPL), and enable of regulators may happen later in SPL with the last
patch applied.

There is also some code in board/google/gru/gru.c that mention a delay
related to the regulators may be needed, however board_early_init_f() is
never called as following can be found in .config:

  # CONFIG_BOARD_EARLY_INIT_F is not set

Will you be able to test with just the rk_setreg(&grf->io_vsel, 1 << 0);
and remove the spl_gpio_output/set_pull parts of the last patch before I
send a v2?

Regards,
Jonas

> 
> Regards,
> Simon