Bump LLVM to 560b72c0408a8f7e4340a1d4197b164a14cd30b0.

[mikeurbach]

No description provided.

[mikeurbach]

We're overdue on an LLVM bump coming off the holidays, so wanted to get the ball rolling.

I'm seeing a few failures, potentially related to the end of the DialectConversion changes (mentioned here):

********************
Failed Tests (3):
  CIRCT :: Conversion/VerifToSMT/verif-to-smt.mlir
  CIRCT :: Dialect/ESI/lowering.mlir
  CIRCT :: Dialect/HW/flatten-io.mlir

cc @mortbopet and @maerhart who helped the last time these tests had issues. I haven't started triaging yet, but will look more closely.

[mortbopet]

@mikeurbach ESI and flatten-io tests should now be fixed. It was (again, as in #7976) a case of an existing issue that wasn't actually being checked for. Namely, that exploded hw.structs in a parent hw.module of a hw.instance weren't being re-materialized through hw.struct_create to be fed to an unmodified hw.extern_module.

[fabianschuiki]

LGTM!

[mikeurbach]

Thanks @mortbopet and @maerhart! I will plan to merge at the end of the work day US time.