On Wed, 4 Nov 2026 at 08:51, Francesco Dolcini <[email protected]> wrote:
> > The documented and expected behaviour is that a separate host computer
> > then uploads a U-Boot binary which then boots into fastboot mode,
> > allowing for storage devices to be flashed, usually by the 'uuu' tool.
>
> Why is this i.MX specific? We (Toradex) have exactly the same on other
> boards that does not use any NXP SoC.
[...]
> How can I opt out from it?
>
> Did you consider having a new boot command variable.
>
> if (bootcmd_usb_recovery && booting_from_usb_recover())
> env_set(bootcmd, bootcmd_usb_recovery)
>
>
> The issue I see with this simple proposal is the USB instance number,
> but maybe this can be solved?
Good points. And thinking more, I'm wondering if it doesn't make sense
for it to be done in a generic way.
Having this enabled across all i.MX platforms is probably not correct,
some would want to opt out (or simply never have been opted in).
And while there is some similarity amongst i.MX platforms, there are
other non-i.MX boards that want this behaviour.
And as we look at making it more generic, then we quickly end up
considering: what if it's not the first USB instance.. what if it's
not usb at all.. what if it's not fastboot but some other recovery
mechanism.. what if the condition is not boot media but a GPIO or
something..
And all that leads back towards: these are board-level / product-level
details. And ultimately it's just a one line boot command we want to
drop in place.
So I'm wondering if we should just leave this as a board-level thing
as it is now, and have each board continue to do:
if (device was booted in recovery mode)
adjust the bootcmd appropriately;
and I'll just fix up the obvious inconsistency with the way bootcmd
gets set for the FRDM-IMX93 in front of me.
Thoughts?
Thanks
Daniel
