64-bit RPi4B u-boot hangup with modern rpi firmware: some information (but investigative-toolbox limited)

Mark Millard marklmi at yahoo.com
Wed Oct 14 04:43:27 UTC 2020


On 2020-Oct-13, at 20:43, Klaus Cucinauomo <maciphone2 at googlemail.com> wrote:
> 
> Am 14.10.2020 um 03:13 schrieb Mark Millard via freebsd-arm <freebsd-arm at freebsd.org>:
>> 
>> ...64-bit RPi4B u-boot hangup with modern rpi firmware..
> 
> ????
> there is no u-boot-hangup with modern firmware(at least not on my board),
> the machine is hung up by using armstubs….
> first figured out by Mark Millard … :-)

The full set of observations that have been reported so far
still have no root cause found and do not well justify
assigning blaim, as far as I can tell.

Below are notes indicating that attributing everything to
armstub8-gic.bin may be a not-well-justified attribution
on the current evidence . . .

A) Robert C. provided a backtrace of a hangup in his context.
The call chain was of u-boot routines, not armstub8-gic.bin
code, not rpi firmware code. (By contrast, what I've done can
not even show that u-boot ever was started after it was loaded.
The same goes for armstub8-gic.bin : I've no direct evidence
of my own that it was ever started, just loaded. I depend on
Robert's report for the evidence that both were started.)

Note: I was able to validate the part of the backtrace that
Robert published was coherent with the u-boot 2020.10 source
code.

B) If I substitute some specific older rpi firmware, leaving
armstub8-gic.bin and u-boot.bin alone, the 4 GiBYte RPi4B
suddenly boots just fine. (But this may be tied to the
relationships among the 3, not specifically to an error in
the more modern rpi firmware.)

C) Any combination of the 3 parts (rpi firmware, armstub8-gic.bin,
u-boot) could be mishandling memory ownership in one or more ways.
But until the actual mishandling(s) involved were found,
attributing cause from available evidence is problematical. (B)
illustrates that.

I'll remind that I get the hangup with modern rpi firmware and
u-boot 2020.10 when all USB parts are empty. Also with a USB3
SSD only boot on a 4 GiByte RPi4B, no microsd card.

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



More information about the freebsd-arm mailing list