[LowerToHW] Add support for lower-to-core in firrtl.fprintf

[nanjo712]

Add support for lower-to-core in firrtl.fprintf

Follow up #10450 .

Part of #10131 .

[nanjo712]

Hi @fzi-hielscher. I tried implementing the handling of firrtl.fprintf based on your description in #10422, and now sim.get_file is in the same sim.triggered block as sim.proc.print, which uses it. I think this might be the form you want? Please correct me if I've misunderstood.

[fzi-hielscher]

Yep, that's what I meant. 👍
We'll also have to update the GetFileOp itself: Make it procedural and note in the description that it has an immediate side-effect (create the file). Maybe it would be easier to do the changes in the core dialects and SV lowering before adding it to the FIRRTL lowering?

[nanjo712]

That sounds reasonable. I'll first move on to a new PR to handle the changes to GetFileOp, and once that PR is complete, we'll return to this PR to continue.

[circt-bot]

Results of circt-tests run for 63fe07a compared to results for 7ffa508: no change to test results.

[circt-bot]

Results of circt-tests run for 444f21b compared to results for 63fe07a: no change to test results.

[nanjo712]

Gentle ping on this PR. The change on GetFileOp has been completed, so I think we can continue with this PR now. Please let me know if there’s anything else I should address.

[nanjo712]

Hi @uenoku, since you have reviewed related LowerToHW changes before, would you mind taking a look when you have time? I’d mainly like to confirm whether this direction for lowering firrtl.fprintf into the core/sim path looks reasonable.

[circt-bot]

Results of circt-tests run for 2a25f87 compared to results for 63880ae: no change to test results.