Closed TaoBi22 closed 5 days ago
Sorry for the back-and-forth!
Looks like we do link the Verif dialect into SVTransforms, so maybe this is good to go. Going to circle back (and maybe others take a look/have thoughts).
Thanks for pushing on this! :+1:.
Sorry for the back-and-forth!
Looks like we do link the Verif dialect into SVTransforms, so maybe this is good to go. Going to circle back (and maybe others take a look/have thoughts).
Thanks for pushing on this! 👍.
No worries, thanks for taking a look! Strictly only one of this and #7247 is needed to break the cycle I need to break (so I can have Verif link in Seq), just figured it was worth tidying them both up while I was at it (although #7247 has the same case with a LowerSeqHLMem dep on SV that's linked for in CIRCTSeqTransforms).
Looks like we do link the Verif dialect into SVTransforms, so maybe this is good to go.
I agree with this analysis. Linking the transform/conversion libraries against Verif should be enough since the dialect library does not seem to use anything from Verif. Thus, the dependency of CIRCTSV
on CIRCTVerif
can go.
Thanks again for looking into this @dtzSiFive!
Same situation as #7247, SV depends on Verif but that seems to be unnecessary as long as om-linker and HWToBTOR2 CMakeLists are updated to link in Verif