Re: [PATCH v3 10/19] LoongArch: Boot Image bits

On Fri, Aug 21, 2026 at 10:20:48AM -0600, Tom Rini wrote:
> On Fri, Aug 21, 2026 at 08:42:10AM +0000, Yao Zi wrote:
> 
> > From: Jiaxun Yang <[email protected]>
> > 
> > Implement loading and booting functions for LoongArch
> > standard kernel image as per spec.
> > 
> > LoongArch kernel do expect us to fake a efi systemtable

I occasionally noticed this typo ('a' -> 'an'), will fix it.

> > for passing fdt to kernel, we don't need to implement any
> > EFI functions for kernel because it won't look into anything
> > beside devicetree from that table if we tell kernel we are
> > not efi compatible by setting a0 boot argument to zero.
> > 
> > Link: https://docs.kernel.org/arch/loongarch/booting.html
> > Signed-off-by: Jiaxun Yang <[email protected]>
> > Signed-off-by: Yao Zi <[email protected]>
> [snip]
> > diff --git a/arch/loongarch/Kconfig b/arch/loongarch/Kconfig
> > index ac4be2c837c0..5654de5a58c7 100644
> > --- a/arch/loongarch/Kconfig
> > +++ b/arch/loongarch/Kconfig
> > @@ -56,4 +56,7 @@ config STACK_SIZE_SHIFT
> >  config OF_BOARD_FIXUP
> >  	default y if OF_SEPARATE
> >  
> > +config LIB_BOOTM
> > +	select CRC32
> 
> If this is needed, it's a generic problem (and to be fixed outside this
> series) where LIB_BOOTM (and possibly LIB_BOOTI/Z since you know what to
> look for) need to select CRC32 and SPL_LIB_BOOTM SPL_CRC32. Thanks.

I think the dependency is specific to LoongArch, since CRC32 is
necessary only because we need to fake an EFI system table, which is
CRC32-checksum'ed, to pass FDT to the kernel on LoongArch. AFAIK other
architectures don't require such treatment, so I keep this
LoongArch-only.

Should I include a comment here to explain the dependency?

> -- 
> Tom

Regards,
Yao Zi