Hi Eric and Tim, On 16:15 Sun 06 Sep , Eric Chung wrote: > On Wed, Sep 2, 2026 at 4:52 PM Tim Ouyang <[email protected]> wrote: > > > > On Mon, Aug 24, 2026 at 04:50:36PM +0800, Eric Chung wrote: > > > Add logic to determine the current boot device by reading the > > > SoC's configuration registers, rather than using a hardcoded > > > default. > > > > > > Signed-off-by: Eric Chung <[email protected]> > > > Reviewed-by: Yao Zi <[email protected]>
.. > > > > Hi Eric, > > > > I previously said that I found no obvious issues, but after taking > > another look, I noticed that read_boot_mode() is called twice: first > > from board_init_f() and later from spl_boot_device(). > > > > The first call returns a BOOT_DEVICE_* value, which board_init_f() > > writes back to BOOT_DEV_FLAG_REG. The second call then incorrectly > > interprets that value as a BOOT_MODE_* value. > > > > For example, BOOT_DEVICE_SPI is written as 8, which does not match > > BOOT_MODE_SPI (0x55d), so the second call falls back to MMC1. > > > > Please keep the BOOT_MODE_* and BOOT_DEVICE_* encodings separate. > > Sorry for missing this in my initial review. > > -- > > Best regards, > > Tim Ouyang > > Hi Tim, > > Thanks for catching this issue. I've fixed it and submitted a new > version. Please review and merge the new version. > with spacemit custodian tree created [1], I think it's more proper to route patches via there, I will review the patch in next few days and setup CI to test.. [1] https://git.u-boot-project.org/u-boot/custodians/u-boot-spacemit -- Yixun Lan (dlan)
