[compiler-rt] Restore unsigned value in struct, use enum only in function #165048
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
compnerd
approved these changes
Oct 24, 2025
| if (Type != CPU_TYPE_MAX) | ||
| __cpu_model.__cpu_type = Type; | ||
| if (Subtype != CPU_SUBTYPE_MAX) | ||
| __cpu_model.__cpu_subtype = Subtype; |
There was a problem hiding this comment.
I don't follow, I think one of your comments wasn't posted
There was a problem hiding this comment.
Ugh, silly GitHub. I was thinking early return might be useful here.
There was a problem hiding this comment.
I don't see it, why have early return? It's possible for a case to set both type and subtype (e.g. DMR)
e-kud
reviewed
Oct 25, 2025
Comment on lines
680
to
684
| if (Type != CPU_TYPE_MAX) | ||
| __cpu_model.__cpu_type = Type; | ||
| if (Subtype != CPU_SUBTYPE_MAX) | ||
| __cpu_model.__cpu_subtype = Subtype; | ||
|
|
There was a problem hiding this comment.
I won't block the PR but I don't like this kind of refactor. Previously it was a pure static function that touched no context. Now it implicitly modifies the global variable. References to __cpu_type and __cpu_subtype were additional return values.
There was a problem hiding this comment.
I understand the downsides, but I believe in this file it's not gonna be a problem, especially how the function is local.
I can refactor later to not use the global like that, but the previous solution of passing a pointer to some fields of the global isn't much different than what the code is doing now
There was a problem hiding this comment.
I commented on my original PR that passing pointers to field while assuming a different signedness is not good. What about passing pointers to entire struct? That would required the struct to be named but I think we should be fine by that?
There was a problem hiding this comment.
Yeah, passing the pointer to the struct would be my preference as well.
There was a problem hiding this comment.
I've replaced the global with a pointer to struct
e-kud
approved these changes
Oct 26, 2025
| #ifndef _WIN32 | ||
| __attribute__((visibility("hidden"))) | ||
| #endif | ||
| struct __processor_model { |
There was a problem hiding this comment.
I'm not sure, probably this naming was chosen when it wasn't supposed to pass a struct as an argument (even for static functions). Maybe we need to rename it? But I'm ok with the current version.
compnerd
approved these changes
Oct 26, 2025
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/186/builds/13472 Here is the relevant piece of the build log for the reference |
|
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/95/builds/18958 Here is the relevant piece of the build log for the reference |
varun-r-mallya
pushed a commit
to varun-r-mallya/llvm-project
that referenced
this pull request
Oct 27, 2025
…tion (llvm#165048) Typed enums are c23 features and are too new to be used. This PR restores the types in the `__processor_model` struct back to `unsigned int`, removes typed enums, and uses the enum in the function as a variable that's later assigned to a struct in order to prevent errors fixed initially here: llvm#164713 See llvm#165034 for more background
dvbuka
pushed a commit
to dvbuka/llvm-project
that referenced
this pull request
Oct 27, 2025
…tion (llvm#165048) Typed enums are c23 features and are too new to be used. This PR restores the types in the `__processor_model` struct back to `unsigned int`, removes typed enums, and uses the enum in the function as a variable that's later assigned to a struct in order to prevent errors fixed initially here: llvm#164713 See llvm#165034 for more background
Lukacma
pushed a commit
to Lukacma/llvm-project
that referenced
this pull request
Oct 29, 2025
…tion (llvm#165048) Typed enums are c23 features and are too new to be used. This PR restores the types in the `__processor_model` struct back to `unsigned int`, removes typed enums, and uses the enum in the function as a variable that's later assigned to a struct in order to prevent errors fixed initially here: llvm#164713 See llvm#165034 for more background
aokblast
pushed a commit
to aokblast/llvm-project
that referenced
this pull request
Oct 30, 2025
…tion (llvm#165048) Typed enums are c23 features and are too new to be used. This PR restores the types in the `__processor_model` struct back to `unsigned int`, removes typed enums, and uses the enum in the function as a variable that's later assigned to a struct in order to prevent errors fixed initially here: llvm#164713 See llvm#165034 for more background
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
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.
Typed enums are c23 features and are too new to be used. This PR restores the types in the
__processor_modelstruct back tounsigned int, removes typed enums, and uses the enum in the function as a variable that's later assigned to a global in order to prevent errors fixed initially here: #164713See #165034 for more background