Closed hansfbaier closed 1 year ago
Now all tests pass. Interestingly enough, the downconverter expects different timing of signals than the normal and upconverting ones.
The CI still fails here because it depends on https://github.com/enjoy-digital/litex/pull/1674
Oh, I still forgot to account for the fact that avalon is byte addressed. What do you think would be the best way of going about this?
This PR now depends on https://github.com/enjoy-digital/litex/pull/1697 and the CI should pass once that has been merged. Also the burst tests pass now. This is ready for review.
Verified on the FPGA
Litescope trace
Waitstates are quite high. 15 clock cycles :/ But using zero L2 cache at this point. Will try with 8k.
Oh the glitches were probably from something else writing in scaler address space. Moved the scaler up a couple of megabytes, now the scaler video is without glitches!
@enjoy-digital The CI failures do not seem to be related to the changes made here. Is the CI broken?
@hansfbaier: Good work! Will review it tomorrow. I'll also look at the CI.
Thanks @hansfbaier, this is merged. I'll fix CI if failing.
The current implementation passes all tests except the downconverting ones. I compared with the wishbone tests and I could not find what I am missing. The wdata port just does nothing, and I can't understand why. See the skipped tests, which are the failing ones (simulator hangs).