dbSta: SpefReader name-escape regression for DEF+SPEF flow + OpenSTA submodule bump#10298
Conversation
Signed-off-by: dsengupta0628 <dsengupta@precisioninno.com>
Signed-off-by: dsengupta0628 <dsengupta@precisioninno.com>
|
clang-tidy review says "All clean, LGTM! 👍" |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request adds a regression test, spef_def_names, to verify SpefReader name-escape lookups for netlists defined in DEF files. The update includes new test assets (DEF, SPEF, TCL), expected output, and build system integration in BUILD and CMakeLists.txt, alongside a subproject commit update for sta. The reviewer recommended sorting the test lists alphabetically in the build files to enhance maintainability.
| "sta3", | ||
| "sta4", | ||
| "sta5", | ||
| "spef_def_names", |
There was a problem hiding this comment.
The test list in ALL_TESTS is not sorted alphabetically. While not strictly required, maintaining an alphabetical order improves maintainability and makes it easier to locate tests.
| sta3 | ||
| sta4 | ||
| sta5 | ||
| spef_def_names |
There was a problem hiding this comment.
The test list in or_integration_tests is not sorted alphabetically. Maintaining alphabetical order improves maintainability.
|
clang-tidy review says "All clean, LGTM! 👍" |
Signed-off-by: dsengupta0628 <dsengupta@precisioninno.com>
|
clang-tidy review says "All clean, LGTM! 👍" |
Signed-off-by: dsengupta0628 <dsengupta@precisioninno.com>
|
clang-tidy review says "All clean, LGTM! 👍" |
Signed-off-by: dsengupta0628 <dsengupta@precisioninno.com>
|
clang-tidy review says "All clean, LGTM! 👍" |
3 participants
Summary
Bumps the OpenSTA submodule to pick up
The-OpenROAD-Project/OpenSTA#356 ,
which adds flat-name fallbacks in
SpefReader::findInstanceRelative/findNetRelative.Adds a regression in
src/dbSta/test/that exercises the fix end-to-end.Motivation
read_deffollowed byread_spefproduces hundreds to thousands ofspurious
STA-1648 instance ... not foundandSTA-1650 net ... not foundwarnings on real designs whose flat netlist contains:/in the leaf(e.g.
u_sx1_iq/g1496206), where OpenDB stores the verbatim DEFstring but
Network::findInstanceRelativeinterprets the/as ahierarchy boundary;
bus[0]) where OpenDB stores[0]literally but theSPEF writer follows the SPEF grammar and emits
bus\[0\].On a customer testcase (~772k components), this accounted for >1k
spurious STA-1648 warnings; the upstream fix drops that to zero.
Type of Change
Impact
[How does this change the tool's behavior?]
Verification
./etc/Build.sh).Related Issues
[Link issues here]