[rsz] Change RSZ-0075 makeBufferedNet warning to debug message by openroad-ci · Pull Request #9503 · The-OpenROAD-Project/OpenROAD

openroad-ci · 2026-02-20T07:44:12Z

Summary

Changed makeBufferedNet failed for driver from WARNING (RSZ-0075) to debug message
This message is an internal state that users cannot act upon, causing log spam during repair_timing operations
Updated 3 test golden files to remove the now-suppressed warning output

Resolves

[WARNING RSZ-0075] makeBufferedNet failed for driver core/rob/core/rob/_99958_/CON #5988
make [WARNING RSZ-0075] makeBufferedNet failed for driver debug message #8988

Test plan

Incremental build passes
All 189 rsz regression tests pass (clone_flat, clone_hier, repair_setup_vt_swap2 verified)

🤖 Generated with Claude Code

Occurrences of `sta_->cmdMode()` should be read as TODOs for proper multi-mode handling later. Signed-off-by: Martin Povišer <povik@cutebit.org>

Now that the command has different behavior (drops constraints) it was breaking the tests. Signed-off-by: Martin Povišer <povik@cutebit.org>

Signed-off-by: Martin Povišer <povik@cutebit.org>

The makeBufferedNet failure is an internal state that users cannot act upon. Change it from a warning (RSZ-0075) to a debug message to eliminate log spam during repair_timing operations. Resolves The-OpenROAD-Project#5988, The-OpenROAD-Project#8988 Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Jaehyun Kim <jhkim@precisioninno.com>

gemini-code-assist

Code Review

This pull request appropriately changes the makeBufferedNet failed for driver warning (RSZ-0075) to a debug message. As noted in the description, this warning is unactionable for users and contributes to log spam, so demoting it to a debug message is a sensible improvement. The implementation correctly replaces logger_->warn with the debugPrint macro, and the test golden files have been updated accordingly to reflect this change. The pull request is well-executed and ready to merge.

github-actions · 2026-02-20T07:49:40Z

clang-tidy review says "All clean, LGTM! 👍"

oharboe · 2026-02-20T09:17:49Z

😌

jhkim-pii · 2026-02-20T13:01:57Z

@precisionmoon @povik

Related ORFS PR: The-OpenROAD-Project/OpenROAD-flow-scripts#3900

Remove pre-placement buffer attempt.
Remove estimate_parasitics -placement call in the floorplan stage.

@maliberty This OR PR does not affect QoR.

Related net lookup for modnets existed in two variants: slow which performed consistency checking and fast without checking. Transition isPower/isGround to using the fast variant, and rename the slow variant to emphasize its checking purpose. Signed-off-by: Martin Povišer <povik@cutebit.org>

maliberty · 2026-02-20T17:04:27Z

src/rsz/test/repair_setup_vt_swap2.ok

-[WARNING RSZ-0075] makeBufferedNet failed for driver _570_/SN
-[WARNING RSZ-0075] makeBufferedNet failed for driver _570_/CON
-[WARNING RSZ-0075] makeBufferedNet failed for driver _570_/SN
-[WARNING RSZ-0075] makeBufferedNet failed for driver _315_/Y


Do you understand the nature of the failure? I am wondering if we are hiding a bug.

It's expected when the net has unplaced pins

I'm wondering why we have a test with unplaced pins then.

The problematic test cases did not use read_def or placement.
I updated the test cases to use read_def or do placement.

Provide an alias for a method renamed in OpenSTA 3.0 to fix some bazel-orfs tests. Signed-off-by: Martin Povišer <povik@cutebit.org>

Signed-off-by: Martin Povišer <povik@cutebit.org>

The RSZ-0075 "makeBufferedNet failed" message is restored from debug back to warning. The root cause of excessive warnings was that test scripts created instances via read_verilog/link_design without placement, causing makeSteinerTree to return nullptr for all unplaced pins. Fix three test scripts to ensure proper placement: - clone_flat.tcl: use read_def instead of read_verilog/link_design (DEF already contains both netlist and placement) - clone_hier.tcl: add initialize_floorplan, global_placement, and detailed_placement to place verilog-created instances - repair_setup_vt_swap2.tcl: add placement flow before repair_timing Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Jaehyun Kim <jhkim@precisioninno.com>

github-actions · 2026-02-22T10:56:39Z

clang-tidy review says "All clean, LGTM! 👍"

povik · 2026-02-22T11:05:11Z

src/rsz/test/clone_hier.tcl

+source $tracks_file
+place_pins -hor_layers $io_placer_hor_layer \
+  -ver_layers $io_placer_ver_layer
+global_placement -skip_nesterov_place


Please load an already placed design. Doing placement as part of the test makes it sensitive to gpl changes and can burden gpl development. We have existing rsz tests doing this which we need to fix.

Addresses review feedback to avoid coupling RSZ tests to GPL changes. Tests now load pre-placed DEF files instead of running global_placement. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Jaehyun Kim <jhkim@precisioninno.com>

github-actions · 2026-02-22T15:26:23Z

clang-tidy review says "All clean, LGTM! 👍"

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Jaehyun Kim <jhkim@precisioninno.com>

github-actions · 2026-02-23T05:56:33Z

clang-tidy review says "All clean, LGTM! 👍"

precisionmoon · 2026-02-23T06:17:57Z

Besides regressions, do we have any RSZ-0075 errors in ORFS CI tests? I remember seeing these errors in many asap7 tests like ethmac, mock-cpu, jpeg-lvt, etc. Are these OK to suppress?

jhkim-pii · 2026-02-23T07:35:00Z

Besides regressions, do we have any RSZ-0075 errors in ORFS CI tests? I remember seeing these errors in many asap7 tests like ethmac, mock-cpu, jpeg-lvt, etc. Are these OK to suppress?

@precisionmoon
The-OpenROAD-Project/OpenROAD-flow-scripts#3900 addresses the warnings in ORFS designs.

RSZ-0075 warnings occurred at floorplan stage due to the unplaced pins. Buffering at floorplan stage seems not that useful, so I removed buffer strategy. I will study about applying gain-based buffering (which does not require placed pins) in floorplan stage to see if it can give any meaningful benefit (As @povik said, the buffers inserted at floorplan stage will be removed in placement stage by remove_buffer. So, not sure if it can give a meaningful gain).

povik and others added 6 commits February 19, 2026 16:51

Update to build with multi-mode STA

1a601fa

Occurrences of `sta_->cmdMode()` should be read as TODOs for proper multi-mode handling later. Signed-off-by: Martin Povišer <povik@cutebit.org>

Drop sta::network_changed from tests

0b2e66e

Now that the command has different behavior (drops constraints) it was breaking the tests. Signed-off-by: Martin Povišer <povik@cutebit.org>

Regolden tests after STA update

c332706

Signed-off-by: Martin Povišer <povik@cutebit.org>

Update flow metrics after STA update

962c5e0

Signed-off-by: Martin Povišer <povik@cutebit.org>

rsz: Resolve compile warnings

a998009

Signed-off-by: Martin Povišer <povik@cutebit.org>

openroad-ci assigned jhkim-pii Feb 20, 2026

gemini-code-assist bot reviewed Feb 20, 2026

View reviewed changes

openroad-ci mentioned this pull request Feb 20, 2026

[rsz] Change RSZ-0075 warning to debug + remove pre-placement buffer attempt The-OpenROAD-Project/OpenROAD-flow-scripts#3900

Open

3 tasks

jhkim-pii requested review from maliberty, povik and precisionmoon February 20, 2026 12:58

maliberty reviewed Feb 20, 2026

View reviewed changes

povik and others added 4 commits February 20, 2026 18:21

dbSta: Provide sta::endpoint_path_count alias

21d54bd

Provide an alias for a method renamed in OpenSTA 3.0 to fix some bazel-orfs tests. Signed-off-by: Martin Povišer <povik@cutebit.org>

Pull fix for write_timing_model crash

805a1d9

Signed-off-by: Martin Povišer <povik@cutebit.org>

Update mock-array script for new STA

1b3a3b6

Signed-off-by: Martin Povišer <povik@cutebit.org>

povik reviewed Feb 22, 2026

View reviewed changes

jhkim-pii and others added 2 commits February 23, 2026 12:20

Merge origin/master (STA MCMM update) and regolden RSZ tests

84cb249

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Jaehyun Kim <jhkim@precisioninno.com>

Merge staging STA MCMM submodule and regolden RSZ tests

168f0de

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com> Signed-off-by: Jaehyun Kim <jhkim@precisioninno.com>

Comments

Conversation

openroad-ci commented Feb 20, 2026

Summary

Resolves

Test plan

Uh oh!

gemini-code-assist bot left a comment

Choose a reason for hiding this comment

Code Review

Uh oh!

github-actions bot commented Feb 20, 2026

Uh oh!

oharboe commented Feb 20, 2026

Uh oh!

jhkim-pii commented Feb 20, 2026

Uh oh!

maliberty Feb 20, 2026

Choose a reason for hiding this comment

Uh oh!

povik Feb 20, 2026

Choose a reason for hiding this comment

Uh oh!

maliberty Feb 22, 2026

Choose a reason for hiding this comment

Uh oh!

jhkim-pii Feb 22, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Feb 22, 2026

Uh oh!

povik Feb 22, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jhkim-pii Feb 23, 2026

Choose a reason for hiding this comment

Uh oh!

povik Feb 23, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Feb 22, 2026

Uh oh!

github-actions bot commented Feb 23, 2026

Uh oh!

precisionmoon commented Feb 23, 2026

Uh oh!

jhkim-pii commented Feb 23, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

povik Feb 22, 2026 •

edited

Loading