[FIRRTL] Ensure all types and attributes are walkable #9007
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
seldridge
approved these changes
Sep 24, 2025
Member
seldridge
left a comment
seldridge
left a comment
There was a problem hiding this comment.
Only lightly reviewed. Looks good.
Member
There was a problem hiding this comment.
Lightly reviewed. This looks good.
| Builder builder{&context}; | ||
| }; | ||
|
|
||
| TEST_F(TypeWalkTest, BundleTypeWalk) { |
Member
There was a problem hiding this comment.
This kind of test could be golfed down to a single field. Not critical.
Member
|
Also, thank you for figuring out that this was the issue and fixing this! |
uenoku
reviewed
Sep 24, 2025
uenoku
approved these changes
Sep 24, 2025
Upstream MLIR requires type and attribute storages to define a `getAsKey` function, and for the return type to be something that implements the corresponding attribute/type walk. Most of our FIRRTL types and attributes predate these requirements. This PR is mostly busy work, adding `getAsKey` to the types and attributes that don't yet have it, and changing `std::pair` to `std::tuple`. (For some reason upstream type/attribute walking works on tuples, but not pairs.) Also add unit tests for attributes and types that ensure the walks actually work. These tweaks allow circt-reduce to be a lot more clever about which ops it touches, because it can see a lot more symbol uses now. I also expect the `InnerSymbolDCE` pass to potentially benefit from this, at least if inner symbols are weirdly nested within FIRRTL attributes and types. Most of this was done by Claude Sonnet 4, with a lot of handholding.
fabianschuiki
force-pushed
the
fschuiki/fix-type-walk
branch
from
9d894d8 to
4496357
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Upstream MLIR requires type and attribute storages to define a
getAsKeyfunction, and for the return type to be something that implements the corresponding attribute/type walk. Most of our FIRRTL types and attributes predate these requirements. This PR is mostly busy work, addinggetAsKeyto the types and attributes that don't yet have it, and changingstd::pairtostd::tuple. (For some reason upstream type/attribute walking works on tuples, but not pairs.) Also add unit tests for attributes and types that ensure the walks actually work.These tweaks allow circt-reduce to be a lot more clever about which ops it touches, because it can see a lot more symbol uses now. I also expect the
InnerSymbolDCEpass to potentially benefit from this, at least if inner symbols are weirdly nested within FIRRTL attributes and types.Most of this was done by Claude Sonnet 4, with a lot of handholding.