Re: [PATCH] binman: Fill in most of the imx8mimage and imx8mcst tests

[Marek Vasut <marex@denx.de>]

On 2/15/25 2:38 PM, Simon Glass wrote:
> Hi Marek,
> 
> On Sat, 8 Feb 2025 at 14:27, Marek Vasut <marex@denx.de> wrote:
>>
>> On 2/7/25 1:49 AM, Simon Glass wrote:
>>> Hi Marek,
>>>
>>> On Thu, 6 Feb 2025 at 13:52, Marek Vasut <marex@denx.de> wrote:
>>>>
>>>> On 2/6/25 1:42 PM, Simon Glass wrote:
>>>>
>>>> Hi,
>>>>
>>>>>>> FAILED (errors=1)
>>>>>>>
>>>>>>> So where is the SRK_1_2_3_4_table.bin file?
>>>>>> See the tool documentation:
>>>>>>
>>>>>> https://gitlab.apertis.org/pkg/imx-code-signing-tool/-/blob/debian/3.4.0+dfsg-2/docs/CST_UG.pdf?ref_type=tags
>>>>>>
>>>>>> 3.1.2 Running the hab4_pki_tree script Example
>>>>>> 3.1.2.1 Running the hab4_pki_tree script in interactive mode
>>>>>>
>>>>>> The hab4_pki_tree.sh script should generate all those files.
>>>>>
>>>>> But with binman we want to avoid vendor scripts, etc.
>>>>
>>>> I believe the script internally runs openssl to generate that file.
>>>>
>>>>> Could someone take a look at updating tools/binman/btool/cst.py or
>>>>> similar to build the tool from source? There are examples of others
>>>>> that do this, e.g. bootgen.py
>>>> Is this relevant to finalizing this patch ?
>>>
>>> Well, I don't see how the tests can pass if we can't build/run the tools needed.
>> Maybe you can try to use the shell scripts to generate the missing bin
>> file ? I suspect the script does some openssl invocation, which can be
>> reproduced in binman ?
> 
> But 'binman tool -f <toolname>' is supposed to build/fetch the tool.
> So any such building should happen in that bintool.
> 
> One of the main goals of binman is to reduce the work needed to create
> a working image. If people have to go spelunking around the Internet,
> build it themselves, etc. that is not good.
What you are asking me for is not the tool itself, but key material, 
which is generated by these scripts.