pdn: add check for instance halo to ensure it does not overlap rows

[gadfort]

Summary

Users keep running into issues where the PDN halo differs from the cut rows halo, causing issues/errors when PDN is inserting vias. This adds a check to ensure the powergrid instance halos do not overlap the follow pins on the standard cells.

Type of Change

• Bug fix
• Breaking change

Impact

If the powergrid setup is correct, no impact is expected, but if the grid has an overlap issue, this will break.

Verification

• I have verified that the local build succeeds (./etc/Build.sh).
• I have run the relevant tests and they pass.
• My code follows the repository's formatting guidelines.
• I have included tests to prevent regressions.
• I have signed my commits (DCO).

[gadfort]

@mole99 FYI

[gemini-code-assist]

Code Review

This pull request introduces a halo overlap check for macro instances in the PDN generator. It adds hasHalo(), suggestHalo(), and checkHalo() methods to InstanceGrid to detect when a macro's halo overlaps with standard cell rows and suggests a reduced halo size to resolve the conflict. Feedback on the changes highlights a potential underestimation in row bloat calculation due to integer division truncation when a followpin strap has an odd width, suggesting rounding up instead.

[maliberty]

Fix the tcl lint. I think the bazel failure is unrelated

[gadfort]

@maliberty it looks like the tinyrocket needs to have its powergrid grid fixed, but when I do I get metrics failures. How do I update those?

[gadfort]

@maliberty also I should mention most of the runtime of this test is actually global routing attempting to disable NDRs.

[maliberty]

@maliberty it looks like the tinyrocket needs to have its powergrid grid fixed, but when I do I get metrics failures. How do I update those?

I assume you mean the one in OR and not the one in ORFS. save_flow_metrics and save_flow_metrics_limits