feat(Zvdot4a8i): add extension and instruction data

[Maanvi212006]

Summary

Adds YAML data for the Zvdot4a extension (vector 4-element, 8-bit integer dot-product accumulate), previously known as Zvqdotq. The extension and its instructions were renamed following ARC feedback; see the discussion on #505 and the renaming PR riscv/riscv-opcodes#402.

This adds:

• spec/std/isa/ext/Zvdot4a.yaml — extension definition
• spec/std/isa/inst/Zvdot4a/ — all seven instructions:
  • vdot4a.vv / vdot4a.vx (signed × signed)
  • vdot4au.vv / vdot4au.vx (unsigned × unsigned)
  • vdot4asu.vv / vdot4asu.vx (signed vs2 × unsigned vs1/rs1)
  • vdot4aus.vx (unsigned vs2 × signed rs1 — vector-scalar only, since
    the vector-vector combination isn't separately encoded and can be
    obtained by swapping operand order into vdot4asu.vv)

Encodings were taken directly from the merged
riscv/riscv-opcodes#402 (extensions/unratified/rv_zvdot4a).

Notes for reviewers

• operation() is left empty for all seven instructions, per the guidance in Add Zvdot4a8i extension #505 — happy to fill in IDL semantics in this PR or a
  follow-up if preferred.
• requirements.extension is set to Zve32x (the minimal embedded vector base), by analogy with how Zve32f requires Zve32x. Please confirm this is the correct/intended prerequisite for this extension (as opposed to V).
• state: development was used since the extension isn't ratified yet (target ratification 26Q4 per the RVI Jira ticket). Let me know if a different value is preferred.
• Ratification/spec reference: [Zvdot4a8i] Extension for vector dot product (4-element, 8-bit integer) Version 0.1 riscv-isa-manual#2576.

Testing

• rake test:inst_encodings — passed, no encoding conflicts.
• rake gen:resolved_arch — resolves cleanly.
• rake test:sorbet — no type errors.
• rake test:udb:unit — full suite passes (3583 runs, 0 failures).

Closes #505

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 68.51%. Comparing base (40ac592) to head (2d745b3).
⚠️ Report is 25 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2015   +/-   ##
=======================================
  Coverage   68.51%   68.51%           
=======================================
  Files          53       53           
  Lines       27916    27918    +2     
  Branches     3620     3620           
=======================================
+ Hits        19126    19129    +3     
+ Misses       7269     7265    -4     
- Partials     1521     1524    +3     

Flag	Coverage Δ
idlc	69.62% <ø> (+0.01%)	⬆️
udb	66.84% <ø> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[Paul Clarke (ThinkOpenly)]

I rebased this. It's very close to ready. CI is complaining about the "golden" instruction appendix not matching what would be generated with this PR merged. Please update:

./do chore:update_golden_appendix

Then commit, push here.

[Jordan Carlin (jordancarlin)]

I rebased this. It's very close to ready. CI is complaining about the "golden" instruction appendix not matching what would be generated with this PR merged. Please update:

./do chore:update_golden_appendix

Then commit, push here.

We should update the autofix job so that this is automatically updated in PRs. Many of the other golden files are already updated this way.

[Maanvi212006]

Hi Jordan Carlin (@jordancarlin)
On your note about the autofix job — would it be alright if I took a stab at fixing that separately (bringing the golden instruction appendix in line with however the other golden files are auto-updated in PRs)? I don't want to step on it if it's already earmarked, but happy to file an issue and then open a PR for it if useful

[Jordan Carlin (jordancarlin)]

Maanvi212006 feel free to open a PR updating the autofix.ci workflow.

[Paul Clarke (ThinkOpenly)]

CI failing again due to the instruction appendix. Could you update that again?

[Paul Clarke (ThinkOpenly)]

need another rebase, or update the instruction appendix once more.