Re: [PATCH] common: spl: spl_nand: make CONFIG_SYS_NAND_U_BOOT_OFFS optional in spl_nand_get_uboot_raw_page()

[Weijie Gao <weijie.gao@mediatek.com>]

Hi Tom,

On Fri, 2026-04-10 at 08:44 -0600, Tom Rini wrote:
> On Fri, Apr 10, 2026 at 04:49:33PM +0800, Weijie Gao wrote:
> 
> > As spl_nand_get_uboot_raw_page() is a weak function, when a board
> > wants to
> > override it, the CONFIG_SYS_NAND_U_BOOT_OFFS becomes meaningless
> > within
> > this function.
> > 
> > This patch allows spl_nand_get_uboot_raw_page() being compiled
> > without
> > CONFIG_SYS_NAND_U_BOOT_OFFS being defined.
> > 
> > Note: this is available only when both CONFIG_SPL_NAND_RAW_ONLY and
> > CONFIG_SYS_NAND_U_BOOT_OFFS_REDUND are not defined.
> > 
> > Signed-off-by: Weijie Gao <weijie.gao@mediatek.com>
> > ---
> >  common/spl/spl_nand.c | 4 ++++
> >  1 file changed, 4 insertions(+)
> 
> I think the bug here is that when SYS_NAND_U_BOOT_LOCATIONS isn't
> defined, we should still be able to override
> spl_nand_get_uboot_raw_page
> (with an implementation that does not use
> CONFIG_SYS_NAND_U_BOOT_OFFS).
> Can you please re-examine the code with that in mind? Thanks!
> 

So I come up with the following idea:
1. With my current path, replace all reference to
   CONFIG_SYS_NAND_U_BOOT_OFFS with spl_nand_get_uboot_raw_page()
2. Add a new weak function, spl_nand_get_uboot_raw_page_redund(), to
   wrap the CONFIG_SYS_NAND_U_BOOT_OFFS_REDUND macro, and do the same
   modification as spl_nand_get_uboot_raw_page() do.

I'll send new patch later

Weijie