u-boot-rpi* has the existing "make first page reserved" code (shown); this appears to be where any fix would go (it works without FreeBSD changes)

[Mark Millard]

On 2020-Feb-14, at 07:32, Klaus Küchemann <maciphone2 at googlemail.com> wrote:

> 
>> Am 14.02.2020 um 12:20 schrieb Mark Millard <marklmi at yahoo.com>:
> . . .
> 
>> Mark Millard  :… I also do not see a communication path for the size to be reported
>> to u-boot so that it could automatically adjust.
> 
> The communication-path is to track the u-boot upstream first.

https://github.com/gonzoua/rpi3-psci-monitor ( which is used to
build FreeBSD's armstub8*.bin files for sysutils/rpi-firmware )
and u-boot are not tied together. Updating to a more modern
u-boot does not change that of itself.

U-boot is not necessarily the most appropriate path for
armstub8*.bin to provide something that eventually indicate to
the FreeBSD kernel to avoid the pages that hold the content.

> . . .

> Nevertheless, Mark, thanks again investigating to retrack the root cause of RPI4 breakage.
> And all that u-boot stuff is absolutely no reason to stop support the RPI4(if possible),
> Just hook up your RPi4-version which boots and continue patching that messy gadget.
> But please in the (fbsd-) upstream ...

The root-cause investigation was for the folks
reporting FreeBSD as broken on their RPi3's/4's
for head as of -r356776 . I've not been trying
to make new things work on the RPi4B. The RPi4B
is just what I happened to have access to that
could be used for the investigation. Workarounds
were a natural part of the investigation.

As things are, if I'm to try to get a patch into
FreeBSD, or something used by FreeBSD, it would
more likely be for issues on old PowerMacs
(powerpc64 and 32-bit powerpc). (But it is
questionable that I'd have enough sustained-focus
time to have an acceptable code update anytime
soon.)

===
Mark Millard
marklmi at yahoo.com
( dsl-only.net went
away in early 2018-Mar)