Will Dietz

FIRRTL Rationale

2

We should add a place to document the "Rationale" for FIRRTL's design, perhaps in the spirit of MLIR's rationales: https://mlir.llvm.org/docs/Rationale/ . This has been proposed and requested by a number...

Better to point to where the intrinsic was used, not where the wrapper is defined. ### Contributor Checklist - [x] Did you add Scaladoc to every public function/method? - [x]...

[Calyx] Verified-valid IR crashes in canonicalizer

1

Consider this snippet, created while trying to reduce example in #7050: ```mlir module attributes {calyx.entrypoint = "main"} { calyx.component @main(%go: i1 {go}, %clk: i1 {clk}, %reset: i1 {reset}) -> (%done:...

[FIRRTL] Remove support for printf-encoded verification, tests.

8

Add small test for load-bearing "UNR" behavior especially interaction with `--add-companion-assume`.

[Handshake] Fix canonicalizer not going through rewriter for RAUW.

4

Detected by -DMLIR_ENABLE_EXPENSIVE_PATTERN_API_CHECKS=ON. cc #7047.

Script to help us/developers check this more easily themselves (some amount of `-P0 -n20` or something to xargs helps make that fast, or thereabouts). Add CI step that runs this.

[Caylx] canonicalization results in invalid IR: ifOp w/empty else

1

Verifier checks that else region is non-empty if present. Current patterns can fail this verification, for example when the else region contains some control op that removes itself (`emptyControl`) the...

With input probes no longer being valid IR, this pass can be simplified. Paths are now append-only and can be resolved in a single post-order (considering all entry points /...

# Summary Using printf strings to encode verification operations is deprecated. This is a transparent change to Chisel users, as Chisel no longer uses this for assert/assume/cover. For the deprecation...