Lukas Sommer

IIRC, the (my) reasoning was a different one: We already have a `import`-command that supports all these flags and implements the corresponding behavior. In addition to that, we have the...

The distinction between global and contextual flags is a design decision that is surely debatable, but I do see why the distinction was made, basically there are two options: *...

> So this message leaves a lot of questions open.. I agree, this message has to be improved. But this has nothing to do with the fact, that there are...

Yes, you are wright. Even if we decided to remove the distinction between global and contextual flags, this could only happen as part of a major overhaul of the CLI....

I would like to keep the ˋimportˋ-command, because it allows users to import cores into the ˋcoreˋ-directory without having to know the directory structure. Besides that, ˋimportˋ with evaluation allows...

Hi @forflo, so far we have mainly observed this behavior with two HDL-implemented PEs. Usually, the problem is that one PE has an ongoing write request (e.g., AXI4 burst write)...

Hi @forflo, I think for the kind of deadlock we mainly had in mind for this issue, one would need two write ports. From the code snippet above there only...

Hi @forflo, thanks for investigating this, very interesting insights! Just out of curiosity and maybe also as a future reference for other users: How did you enforce the write-request to...

Comment by @jahofmann: My preferred solution would always use the last interrupt as a fall back interrupt. For example if 20 interrupts are available, interrupts 0 through 18 are dedicated...

What do you think about adding a possibility to program bitstreams programmatically? That might be useful if you use the TaPaSCo API as a base for some other runtime, e.g....