On Tue Jul 28, 2026 at 4:46 PM IST, Simon Glass wrote:
> Hi Anshul,
>
> On 2026-07-17T10:42:51, Anshul Dalal <[email protected]> wrote:
>> arm: dts: k3-am625: add template for tiboot3.bin
[snip]
>> diff --git a/arch/arm/dts/k3-am625-binman.dtsi
>> b/arch/arm/dts/k3-am625-binman.dtsi
>> @@ -0,0 +1,123 @@
>> + tiboot3_am625_hs_fs_template: template-tiboot3-hs-fs {
>> + symlink = "tiboot3.bin";
>> + ti-secure-rom {
>
> This bakes a policy into the template - that hs-fs is the default
> tiboot3 - but the hs and gp templates deliberately don't. All three
> callers (sk, verdin, phycore) already set symlink = "tiboot3.bin" on
> the hs-fs node themselves, and on the r5 beagleplay the symlink lives
> on the gp node. Would it be OK to drop this line so all three
> templates are symmetric?
Yeah, putting symlink in the template is not the right approach. I'll
fix it in the v2.
>
>> diff --git a/arch/arm/dts/k3-am625-r5-beagleplay.dts
>> b/arch/arm/dts/k3-am625-r5-beagleplay.dts
>> @@ -33,38 +33,7 @@
>> tiboot3-am62x-gp-evm.bin {
>> filename = "tiboot3-am62x-gp-evm.bin";
>> symlink = "tiboot3.bin";
>> - ti-secure-rom {
>> - content = <&u_boot_spl_unsigned>, <&ti_fs_gp>,
>> - <&combined_tifs_cfg_gp>, <&combined_dm_cfg_gp>;
>
> ...
>> - load = <0x43c00000>;
>
> The original hardcodes load = <0x43c00000> whereas the template uses
> load = <CONFIG_SPL_TEXT_BASE>. These resolve to the same value today
> (the ARCH_K3 && !ARM64 default is 0x43c00000, and no R5 defconfig here
> overrides it), so it is a no-op - but please mention it in the commit
> message under 'No functional change is intended' so the next reader
> doesn't have to chase Kconfig defaults.
>
This value is specific to the SoC and will not change for AM62x but I
will add a note in the commit message for the change anyways.
>> diff --git a/arch/arm/dts/k3-am625-binman.dtsi
>> b/arch/arm/dts/k3-am625-binman.dtsi
>> @@ -0,0 +1,123 @@
>> + tiboot3_am625_gp_template: template-tiboot3-gp {
>> + ti-secure-rom {
>> + content = <&u_boot_spl_unsigned>, <&ti_fs_gp>,
>> + <&combined_tifs_cfg_gp>, <&combined_dm_cfg_gp>;
>
> Just to check - the property ordering inside ti-secure-rom here
> (content-sbl / load / content-sysfw / load-sysfw / … then sw-rev /
> keyfile) differs from the hs and hs-fs templates, which group all
> content-* first and load-* last. Both work, but since you are
> consolidating anyway it would read better if the three followed the
> same layout. What do you think?
>
Yeah, that's fair. I'll re-order the gp template to match the other two.
Thanks for the review,
Anshul
