Hi Jon,
Thank you for helping me out with this. I tried out possibly all the solutions that others have used to work around the electrical issue of the VCU118 board. I tried switching the resistors of the SD card adaptor, reducing the voltage of VADJ, and reducing the frequency of SD clock using the SD clock divider. Unfortunately, none of them were able to reliably solve the problem. The boot sequence stops at random places, which I assume is because of image corruption. I tried out all these solutions on several boards as well, but still no luck. So I'm currently working on loading the image through PCIe.
The PCIe integrated design I'm working on, I think is pretty similar to the F1 design. I'm currently using all the parameters that are used to build OpenPiton for VCU118, and replacing the memory controller with the F1 memory controller that gives a master AXI interface that can access DRAM. I exported the system as an IP and connected the IP to a DRAM memory controller through an AXI interconnect block. I connected an XDMA AXI bridge to the interconnect so that I have an interface to DRAM. I then copy the image by memory-mapping the DRAM address space through PCIe and reset OpenPiton. This approach also did not work out for some reason. I have several questions about this approach.
- Do you see any fundamental issues with this setup?
- Could the memory-zeroing logic be overwriting the loaded image? From reviewing the RTL, I believe it should not.
- What address range should the DRAM AXI interface be mapped to for the VCU118 system? I mapped it to 80000000 because the address translation from the piton interconnect to AXI did not seem to change
- Any additional feedback or debugging suggestions would be greatly appreciated.
Thank you as always,
Yasas