Hi,
I try to run “make boot_files BOARDS=ZCU102” to get pynq image file for zcu102. But when I run this command, it show some error message like the figure.
petalinux version : 2020.1
vitis and vivado version :2020.1
pynq version 2.6
Did you provide the BSP for the ZCU102?
We don’t generate a PYNQ image for this board, so you need to treat this like a custom board if you want to build the image.
Cathal
@cathalmccabe Thanks for your reply.
I clone the github code from GitHub - Xilinx/PYNQ: Python Productivity for ZYNQ and I download zcu102.bsp file from Xilinx. Then I copy zcu104.spec and modify them to sutie the zcu102.
There are what I modify:
ARCH_ZCU102 := aarch64
BSP_ZCU102 := xilinx-zcu102-v2020.1-final.bsp
STAGE4_PACKAGES_ZCU102 := ethernet
Do you have solution for above problem?
Follow this link to see what need to pay attention during Ultra-Scale series ZYNQ build:
Remember read the post completely and understand before you execute each part.
Enjoy ~
Hi @CharlesChen1024,
Is there a particular reason you need such an old pynq version? We’ve not built a 2.6 image in a while.
The top of your screenshot states that the build tried to access a repository and failed. Was there any more info on what it was trying to access? It’s possible that some of the urls that old builds relied on are dead by now.
Thanks
Shawn
Shawn if this is the case, are PYNQ going to discontinue some of the Old revision some how?
This could make user know from the first place and reduce the time on setting up environment and have a better experience.
Thank you
@briansune Thanks for your reply.
@skalade Thanks for your reply.
No particular reason. I just follow the tutorial to build the PYNQ image for zcu102. This tutorial used Xilinx version and pynq version are 2020.1 and 2.6 respectively. I will try the new version.
@CharlesChen1024 The steps should be very similar, the only difference with the latest sdbuild flow is we mandate the usage of prebuilt rootfs and sdist, which should significantly speed up your build time!
Hi @briansune, there’s very few reasons for a new user to use an old versions of PYNQ unless, as Charles stated, they were following a tutorial of some sort (which itself might be outdated). We can’t really delete github history.
Thanks
Shawn
Shawn, I think the major reason is not bounded to tutorial but the Vivado and PetaLinux version of the company or cooperation. As telling IT or head to update software usually require a huge procedure or approval.
Above that there are strong and straight reason why old PYNQ good or dead in some how.
I am not suggesting to remove the old trunk revision.
Maybe a weekly build automation to see it is still alive should do the job (Suggestion).
However it will be reason enough to state 3 years software support ground.
So 2.6, 2.7, 3.0 → 2019, 2020, 2022 is target of support.
Bests
Hi @CharlesChen1024 ,
I am also trying to build an image for the zcu102 v2.6. Did you manage to do that?
Thanks,
Gio
Hi @gioregg ,
I did not manage the issue, but I successfully built an PYNQ image for the zcu102 by fllowing version:
Pynq version: 2.7
Petalinux version : 2020.2
Xilinx tool version: 2020.2
Maybe you can try it.
Sincerely,
Charles
