I procrastinated for way too long and apparently now I'm running low on time to do some frivolous stunt hacking
step 0: my microsd card died??
status: I downloaded the wrong image, because apparently an x86 raspbian exists for some reason?
I am good engineer I swear
a bunch of waiting for slow SD cards and connecting wires later and I'm back to where I left off *checks calendar* over a year ago
RPi3B+ acting as JTAG adapter talking to the PXA255 on the target
status: I copied the SDRAM init code from the existing bootloader. it seems to work? I should in theory be able to load a custom binary into RAM using GDB at this point. (debugging this chip is extremely unreliable and I'm not sure why yet)
I can run my own code! I can't output anything yet, but I can use a classic embedded debugging technique of "write debug strings into a memory buffer."
one objective is to access the internal sd card slot so I can load more code faster and with fewer bit errors than via jtag
• • •
Missing some Tweet in this thread? You can try to
force a refresh
Fun hardware discovery of the day: so Atmel has their ATF15xx CPLD family, and this chip is compatible enough with Altera's MAX 7000 series to the point where Atmel has a tool, POF2JED, that can convert Altera bitstreams without recompiling.
I always assumed that this came about via some kind of licensing or second-sourcing agreement, but then I decided to actually take a look at the strings in the POF2JED binary.
It turns out that the POF2JED binary has some secret debugging capability (that cannot be turned on by default) that dumps Altera POF files into a human-readable form.