Which is the minimum set of files to make a pynq SD card for a ciustom board?
Do you mean the minimum files you need to provide or what can be left out of the PYNQ image?
You can find info on the image build here:
I’ve been reading that a number of times but I find it incomplete. Other tutorials don’t seem to help.
I have a linux VM with all the Xilinx tools, cloned the Pynq repo. I can build for the supported boards.
My objective is to have pynq on a different board (with ultrascale+, although as a test I also have a zynq-7000).
I understand I need a board folder in a board repo (done), where I have a spec file (done) and some folders (base, notebook, petalinux_bsp). I think the way to go is to copy files from an example board and modify them, but there are quite a few and some of them (bitstreams, for example) aren’t modifiable, that’s why I’m asking if it’s just a subset of files what I have to re-use and others are generated.
I tried sourcing some tcl’s in Vivado. They create a project and a block diagram for the example board. I think I can retarget the project and BD for my board, save the tcl’s and use them for my board. Is that the way/ a way to do it? But still confused about which and how many files do I need.
so, in short, what does the command make BOARD = my_board need?
You will need the BSP, but notebooks should be optional.
Base contains the “default” bitstream. This is what the Vivado project is that is being built. You can modify this to your own project for your board, or provide your own bitstream.
Have you read this?:
I might, being in the pynq repo. Probably I stopped when seeing it’s similar to the page in readthedocs. now I just noticed there are a bit more details… I’ll give it a go later on.
It looks like, all I need, is an hdf and bitstream. I can just start with vivado and export these. Then it’s pretty easy, isn’t it?
wow, I managed ‘make’ to run on just an hdf and bitstream. It takes many hours, though… probably will never do that again…