[ot][spam][crazy] The Trials of Controlling a Programmer

[Undiscussed Horrific Abuse, One Victim of Many]

this morning i got back on porting serialice, which lets you try
firmware without reflashing the chip (and step through it). serialice
is part of coreboot but has a .com main page:
https://www.serialice.com/Main_Page . there's work in their issue repo
to merge it into coreboot that needs a volunteer to clean it up and
finish.

so, i transferred my serialice work to the main system so i could test
it without a working programmer. i laboriously added debugging and
test statements to bisect it with known-working code. in the end, it
turns it out _already works fine_. it just plain works, and i can
debug my cpu with the cpu using what i already wrote.

it is _incredibly easy_ to port a new board to serialice. all you need
to do is initialise the cpu and the serial port, nothing else, not
even the ram, and there is already code in serialice for common serial
port cpus and cpu families. if you're a little familiar with
assembler, you might be able to port a new board in a single day by
disassembling the factory rom. the boot vector is 16 bytes prior to
the end of the flash, and sometimes there is a southbridge vector 16
bytes prior to that. (at least that was how my old amd64 board is, and
the code it uses is shared across many other boards).

for me now, if i want to look at new bioses, the biggest impediment is
gaining access to the flash chip. i dunno how to do that yet.

anyway, with the logic analyser projects on the side to make the
actual programmer work, next step for me with serialice is to see if
my old build was actually working and my eyes were crossed, or if
there is a reason it wasn't, to discern what the difference is, since
the source code appears the same.