-
Notifications
You must be signed in to change notification settings - Fork 14k
[TargetParser] Increase MAX_SUBTARGET_FEATURES to 384 #144326
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
There are 314 features in RISC-V backend, which is about to exceed the maxinum 320 as there are some ongoing new extensions. We increase the `MAX_SUBTARGET_FEATURES` to 384 so that we won't surprise anyone.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM.
tangent question: now reading how nvm, we want to value-initialize it from TableGen-erated code.MAX_SUBTARGET_FEATURES
/ MAX_SUBTARGET_WORDS
are used, anyone know why we didn't use BitVector
to implement FeatureBitset
?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I really hate this affects every target, but LGTM
LLVM Buildbot has detected a new failure on builder Full details are available at: https://lab.llvm.org/buildbot/#/builders/55/builds/12881 Here is the relevant piece of the build log for the reference
|
There are 314 features in RISC-V backend, which is about to exceed the maxinum 320 as there are some ongoing new extensions. We increase the `MAX_SUBTARGET_FEATURES` to 384 so that we won't surprise anyone.
There are 314 features in RISC-V backend, which is about to exceed
the maxinum 320 as there are some ongoing new extensions.
We increase the
MAX_SUBTARGET_FEATURES
to 384 so that we won'tsurprise anyone.