[PATCH 0/2] bootstd: rauc: fix doc and info msg to reflect real flow - test cases

[Martin Schwan]

Hi Simon,

sorry, that I didn't reply to your personal message.

I had the tests for bootmeth_rauc on my to-do list and I am working on
it right now. It took me some time to get the existing tests in
test/boot/ to work (test instructions and environment are not easy to
understand from the documentation). I hope to have some patches soon.

Regards,
Martin Schwan

On Thu, 2025-10-02 at 19:25 +0200, Andreas Pretzsch wrote:
> On Thu, 2025-10-02 at 10:30 -0600, Simon Glass wrote:
> > Hi,
> > 
> > On Wed, 1 Oct 2025 at 17:22, Andreas Pretzsch <apr at cn-eng.de>
> > wrote:
> > > For the RAUC bootmethod, newly introduced in U-Boot 2025.10
> > > [...]
> > > 
> > While we are talking about this bootmeth, does anyone know what
> > happened re a test for it?
> 
> Guess you refer to test code for bootmeth-rauc. Honestly, no idea.
> 
> I just came to notice that there now is an official way, so bumped
> the
> U-Boot version to 2025.10 in a customer project. And stumbled upon
> some
> details, hence the fixes. I've got some more in the workqueue (mostly
> improve printouts what went wrong and extending documentation a bit),
> but none of it will be ready before 2025.10 release. Needs polish and
> testing. I'll send them in hopefully in time for 2025.11 merge
> window.
> 
> A quick look in test/boot/bootmeth.c shows some tests for extlinux,
> efi, vbe simple. I guess the list check would be expandable for rauc,
> too. The fixed numbering might need some massage, but surely doable.
> 
> But the real thing is to test the boot flow itself, and especially
> the
> switchover and fallback handling. Which in principle is comparable
> for
> all methods, but differ in detail. I can see you already wrote quite
> extensive test cases here, respect for that! However I haven't looked
> into them yet. But pretty sure this will be a task of days, not
> hours.
> 
> While I'm happy to assist here, I realistically won't find the time
> to
> act as the main implementor.
> But maybe Martin Schwan already has some snippets prepared here ?
> Or we even might pull in the RAUC developers, to some extent at least
> ?
> I know their codebase includes an outstanding amount of test
> coverage,
> so there is appreciation for testing at least.
> 
> Best regards,
>   Andreas