RPi4b 8GB 13.0-Current, XHCI broken, wrong U-Boot ?

[Klaus Küchemann]

> Am 05.01.2021 um 23:42 schrieb Emmanuel Vadot <manu at bidouilliste.com>:
> 
> I know (or at least hope) that you're trying to help, so let's start
> 2021 on a good base.
Happy new year, manu ! :-)

> Here I'm talking about u-boot itself, not FreeBSD, not booting from
> USB on the RPI4, U-Boot, only U-Boot.
> Right now, using the latest taggued release of rpi-firmware
> (20201201), there is no raspberrypi,firmware-reset compatible node in
> the dtb so U-Boot cannot init the usb controller. That doesn't mean
> that the RPI cannot boot from usb, it can and will load u-boot, but
> when u-boot will look for it's "root device" (the usb drive) this will
> fail as the usb controller couldn't be init.
> This also means that when we boot from SD card, we cannot access usb
> drives from U-Boot, as, again, the usb controller wasn't init.

And all that was the reason for https://reviews.freebsd.org/D26853 
because u-boot CAN and there IS a firmware-reset-compatible node in D26853

At least we have talked about a simple patch in D26853 in 5 million lines of text, while the reset-patch 
itself is 10 lines or so :-)
I found it funny that D26853 is even used in articles found by google to get the RPI4 up to date,
And it is validated by users.

We now also know that current rpi-org-upstream has problems in bcm2711-dtb AND in start4.elf,
while FreeBSD wouldn’t have any problem with that crappy chaos there if we simply patch the reset-node(all is then compatible 
with your current embedded-firmware-port).
While true: keeping an eye on what rpi.org is doing and will do is always good,
But we would have finished the issue with a simple patch of  bcm2711-dtb .
and we wouldn't have to wait for them upstream to repair something in order to break something else again. 
therefore bcm2711 as source code instead of binaries from elsewhere where we don’t know what happens.
O.K., this is state of today-keeping an eye on it is always good(e.g. yesterday a problem report for netboot- problems
belonging to firmware, while unclear for now if it’s a genet-driver-problem).
The genet-driver has also a little problem to `netif restart` -initialization & dhclient, so possible that this happens in netboot  …
we will see..,

Regards

K.