scripts: add missing -exit flag to openroad -python call

[proppy]

Fixes #1387

[proppy]

Looking at the CI results, there doesn't seems to be a good way to pass the -exit flag for python script that consume arguments.

If placed after a -python SCRIPTNAME that takes arguments (https://github.com/The-OpenROAD-Project/OpenLane/actions/runs/3099512086/jobs/5018790378) it is processed by the python script itself:

Usage: lefutil.py get_metal_layers [OPTIONS] [LEFS]...
Try 'lefutil.py get_metal_layers --help' for help.

Error: No such option: -e

if placed before -python (https://github.com/The-OpenROAD-Project/OpenLane/actions/runs/3099596986/jobs/5018960704) is it not recognized as an valid option:

Unknown option: -e
usage: openroad [option] ... [-c cmd | -m mod | file | -] [arg] ...

Marking this PR as draft until we resolve all discussions on https://github.com/The-OpenROAD-Project/OpenROAD/pull/2288.

[donn]

Pardon, but would you please take a look at regression test tests/1007?

[proppy]

@donn is there a way to pull the logs for the failing test from the CI? or does one need to run the same test locally?

[proppy]

Here is the failure I get when running the regression suite locally:

OpenLane 💀 cat ./regression_results/issue_regression_1007.log
OpenLane bf0e83e509d0810d50d5d3579af61ce8916a8504
All rights reserved. (c) 2020-2022 Efabless Corporation and contributors.
Available under the Apache License, version 2.0. See the LICENSE file for more details.

[INFO]: Using configuration in 'tests/1007/config.tcl'...
[INFO]: PDK Root: /usr/local/google/home/proppy/src/github.com/The-OpenROAD-Project/OpenLane/pdks
[INFO]: Process Design Kit: sky130A
[INFO]: Standard Cell Library: sky130_fd_sc_hd
[INFO]: Optimization Standard Cell Library: sky130_fd_sc_hd
[INFO]: Run Directory: /openlane/tests/1007/runs/RUN_2022.09.28_02.17.08
[INFO]: Preparing LEF files for the nom corner...
[INFO]: Preparing LEF files for the min corner...
[INFO]: Preparing LEF files for the max corner...
[ERROR]: during executing openroad script /openlane/scripts/openroad/insert_buffer.tcl
[ERROR]: Log: tests/1007/runs/RUN_2022.09.28_02.17.08/logs/routing/0-insert_buffer.log
[ERROR]: Last 10 lines:
OpenROAD 5e557eeb13dcbc19a167fb7c00d06006db9c9433 
This program is licensed under the BSD-3 license. See the LICENSE file for details.
Components of this program may be licensed under more restrictive licenses which must be honored.
incompatible database schema revision
child process exited abnormally

[proppy]

incompatible database schema revision

seems to come from https://github.com/The-OpenROAD-Project/OpenROAD/blob/816cfe95a0c90dc7fe0395c000327b3aa6babf96/src/odb/src/db/dbDatabase.cpp#L308-L314

so I'd suspect that https://github.com/The-OpenROAD-Project/OpenLane/blob/master/tests/1007/in.odb has to be regenerated?

[proppy]

Looks that the failure is due to a recent schema version bump: https://github.com/The-OpenROAD-Project/OpenROAD/commit/6bea861cdb640856a643d8e7c42cb64feed3ab0a

That happened since the last time openroad_app got bumped https://github.com/The-OpenROAD-Project/OpenLane/commit/ae089d8da840644f77d60eeeae7695c3f1948d7f

[proppy]

I was able to migrate the .def file from version 56 to version 57 using the following methodology:

• install the most recent build of openroad preceding https://github.com/The-OpenROAD-Project/OpenROAD/pull/2267 (schema version 56) from https://anaconda.org/LiteX-Hub/openroad/files
• read in.odb using the openroad -python; write a def
• install the first build of openroad following https://github.com/The-OpenROAD-Project/OpenROAD/pull/2288 (schema version 57) from https://anaconda.org/LiteX-Hub/openroad/files
• read the def; write the updated in.odb file

see this notebook: https://colab.research.google.com/gist/proppy/a60669f58245da3dd55646c525850f52/migrate-opendb-versions.ipynb

After that the failing test now passes:

Running test case: 1007 (log: ./regression_results/issue_regression_1007.log)
1007 completed successfully.

@donn @maliberty @QuantamHD let me know if there is a better way to handle ODB schema conversion.

[proppy]

@maliberty @donn @QuantamHD relaunched the CI the arm64v8 which timeout'ed after 6h.

[proppy]

CI failure

Done.

[proppy]

@donn did you manage to work around the arm64 build timeout issue?

[donn]

@proppy I just manually built and pushed the arm64 binaries for OpenROAD