[LowerIntmodules] Fix EICG_wrapper intrinsic lowering, swap en/test_en.

[dtzSiFive]

The extmodule and intrinsic disagree on port ordering, swap in LowerIntmodules.

Thanks to @prithayan for reporting this!

[fabianschuiki]

🥳 LGTM! I'm wondering whether the port order of EICG_wrapper is fixed, or if there may be different definitions that have totally different port order. Great to have the lowering complain about those cases 👍

[prithayan]

Thanks for the quick fix.

[dtzSiFive]

🥳 LGTM! I'm wondering whether the port order of EICG_wrapper is fixed, or if there may be different definitions that have totally different port order. Great to have the lowering complain about those cases 👍

I'm leaning towards keeping it this way for now since grabbing and checking each by name appropriately is a pain and this is already something we're close to phasing out.

But you're right that's how it should work.

FWIW:
We never supported that and checked names in a particular order in the past, and blindly wired to it assuming the order 😓 .
Not that that's a great reason to keep that behavior...

We DO emit them in the other order in SeqToSV, so that's fun.