[PATCH v4 05/23] j721e: dts: binman: Package tiboot3.bin, sysfw.itb, tispl.bin, u-boot.img

Neha Malcom Francis n-francis at ti.com
Fri May 19 08:39:29 CEST 2023

Hi Andrew

On 18/05/23 22:05, Andrew Davis wrote:
> On 5/18/23 9:26 AM, Neha Malcom Francis wrote:
>> By providing entries in the binman node of the device tree, binman will
>> be able to find and package board config artifacts generated by
>> TIBoardConfig with sysfw.bin and generate the final image sysfw.itb.
>> It will also pick out the R5 SPL and sign it with the help of TI signing
>> entry and generate the final tiboot3.bin.
>> Entries for A72 build have been added to k3-j721e-binman.dtsi to
>> generate tispl.bin and u-boot.img.
>> Support has been added for both HS-SE(SR 1.1), HS-FS(SR 2.0) and GP 
>> images
>> In HS-SE, the encrypted system firmware binary must be signed along with
>> the signed certificate binary.
>> tiboot3.bin and sysfw-j721e_sr1_1-hs.itb: For HS-SE devices
>> tiboot3.bin_fs and sysfw-j721e_sr2-hs-fs.itb: For HS-FS devices
> Do we actually need "tiboot3.bin_fs"? It should be identical to 
> "tiboot3.bin".
>> tiboot3.bin_unsigned and sysfw-j721e-gp-evm.itb: For GP devices
> I've been thinking about these filenames and what we use as the defaults.
> We have 3 revisions {sr1, sr1_1, sr2) to choose from and each has 3
> types {"gp", "hs", "hs-fs"}. Which gives 9 combos. For which the type
> GP the does not care about revision (the revisions only effected HS parts,
> so all revisions look the same for software on GP parts). The revision
> SR1.0 was never widely released on HS-FS nor HS-SE, so those two combos
> can also be ignored. This gives:
> sysfw-j721e-gp.itb
> sysfw-j721e_sr1_1-hs.itb
> sysfw-j721e_sr1_1-hs-fs.itb
> sysfw-j721e_sr2-hs.itb
> sysfw-j721e_sr2-hs-fs.itb
> So two are missing.

Agreed, will need thosYes we can target e, thanks!

>> <filename>.bin/img: For HS devices
>> <filename>.bin_unsigned/img_unsigned: For GP devices
> I'm not sure we need these two "_unsigned" images, the signature is
> trival to strip out if needed and the signed version boots just
> fine on GP devices as SPL knows how to strip out the signature
> during boot when it detects it is running on a GP device.
> None of the above is a big deal and nothing worth holding up this
> series over, I'll send patches later after this is taken into -next.

Yes, based on further comments we can target adding this in v5 or 
another patch later on.

> Andrew

Thanking You
Neha Malcom Francis

More information about the U-Boot mailing list