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

[Emmanuel Vadot]

On Sun, 3 Jan 2021 13:36:39 +0100
Klaus Küchemann <maciphone2 at googlemail.com> wrote:

> 
> > Am 03.01.2021 um 11:57 schrieb Emmanuel Vadot <manu at bidouilliste.com>:
> > 
> > 
> > Check the RPI image
> > (https://download.freebsd.org/ftp/snapshots/ISO-IMAGES/13.0/FreeBSD-13.0-CURRENT-arm64-aarch64-RPI-20201224-3cc0c0d66a0-255241.img.xz)
> 
> I did and now I?m sure that you don?t own an RPI4-8GB(because the image doesn?t boot from USB/SSD)..
> But the good news is that you have everything done well in ?organizing? the firmware-port.

 Ok, one thing at a time please. I was just answering your mail about
the release image not using the port and using some hardlink whatever
files.
 I do own a RPI4-8GB but indeed I've never tested usb boot on it.

> With this last image the start4.elf & fixup4.dat is fine ! So:good work.
> So it?s all about : https://reviews.freebsd.org/D26853 
> which fixes pcie/usb-bootup-firmware-issues.( tested by me & Andy for now).

 And again please provide an overlay for this with an explanation of
what and why things are needed.
 We will not replace a dts in sys/gnu/ with some other one coming from
a different upstream.

> > Am 03.01.2021 um 11:57 schrieb Emmanuel Vadot <manu at bidouilliste.com>:
> >> So the question is WHICH files should differ from sysutils/rpi-firmware.
> > 
> > None.
> 
>  it?s  bcm2711-rpi-4-b.dtb which(currently) has to differ if we want to boot straight up from USB/SSD .
> I say `currently` because that could change upstream. But why struggle with upstream if we can fix it ourselves?

 We have to do both, fix it ourselves and working with upstream.
 Now that being said please remember that RPI doesn't care about u-boot
so they might don't care about this problem.

> 
> > Am 03.01.2021 um 12:43 schrieb Andy McClements <ajm at ip-ether.net>:
> > 
> > ?..
> > I am happy to say that, WITH PATCH D26853, all of these will successfully allow FreeBSD 20201224 to boot from USB SSD. Also that the vanilla 20201224 img (which I have checked includes the 20201201 firmware) also works fine, WITH PATCH D26853.
> > ?.
> 
> 
> That?s awesome  news, Andy 
> and many many thanks for all your extensive tests !
> So we can probably say that currently we don?t have to struggle with different silicon-version, very good news.
> 
> And while #0XJTAG-Boss Rob privately is booting from a poor uSD-fingernail himself,
> the nicest present he gave us is the USB/SSD-bootup ;-) , superf**king fast, 10 times faster(measured) than uSD, great!
> 
> 
> Regards
> 
> K.
> 
>  
>  
> 
> 
> 
> 
> 


-- 
Emmanuel Vadot <manu at bidouilliste.com>