Uninitialized ports on IsuseUnitBanked

[seldridge]

(I'm mostly just opening issues to keep track of this stuff. I have a lead on what's going on here.)

Compiling the Large BOOM variants (from the tarball) is tripping uninitialized detection:

[firtool]   Running "firrtl.module(firrtl-expand-whens,firrtl-sfc-compat)"
chipyard.harness.TestHarness.LargeBoomV4Config.fir:473939:5: error: sink "col_0.io_rob_head" not fully initialized in "IssueUnitBanked"
    inst col_0 of IssueUnitAgeMatrix_3 @[generators/boom/src/main/scala/v4/exu/issue-units/issue-unit-banked.scala 33:13]
    ^
chipyard.harness.TestHarness.LargeBoomV4Config.fir:473939:5: error: sink "col_0.io_rob_pnr_idx" not fully initialized in "IssueUnitBanked"
    inst col_0 of IssueUnitAgeMatrix_3 @[generators/boom/src/main/scala/v4/exu/issue-units/issue-unit-banked.scala 33:13]
    ^
chipyard.harness.TestHarness.LargeBoomV4Config.fir:474652:5: error: sink "col_1.io_rob_head" not fully initialized in "IssueUnitBanked"
    inst col_1 of IssueUnitAgeMatrix_4 @[generators/boom/src/main/scala/v4/exu/issue-units/issue-unit-banked.scala 33:13]
    ^
chipyard.harness.TestHarness.LargeBoomV4Config.fir:474652:5: error: sink "col_1.io_rob_pnr_idx" not fully initialized in "IssueUnitBanked"
    inst col_1 of IssueUnitAgeMatrix_4 @[generators/boom/src/main/scala/v4/exu/issue-units/issue-unit-banked.scala 33:13]
    ^
chipyard.harness.TestHarness.LargeBoomV4Config.fir:475365:5: error: sink "col_2.io_rob_head" not fully initialized in "IssueUnitBanked"
    inst col_2 of IssueUnitAgeMatrix_5 @[generators/boom/src/main/scala/v4/exu/issue-units/issue-unit-banked.scala 33:13]
    ^
chipyard.harness.TestHarness.LargeBoomV4Config.fir:475365:5: error: sink "col_2.io_rob_pnr_idx" not fully initialized in "IssueUnitBanked"
    inst col_2 of IssueUnitAgeMatrix_5 @[generators/boom/src/main/scala/v4/exu/issue-units/issue-unit-banked.scala 33:13]
    ^

I think the issue is that this was an upstream BOOM bug that was fixed in: riscv-boom/riscv-boom@383f241#diff-15fca8e71cf557d910646ef8eab4a097bd686fd5c5b26166406171e153600766R43-R49

@TaoBi22: How did you generate these tarballs and can they be generated again with that patch? It looks like Chipyard has not pulled in that patch via a submodule update to BOOM.

[TaoBi22]

Thanks for looking into this! Should be regenerable with the patch, they were produced from the upstream BOOM configs with some tweaks to make them synthesizable (just realised I need to push these to a fork at some point), so I'll bump BOOM and regenerate.