Skip to content
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

default on LLVM_IAS=1 #1434

Closed
nickdesaulniers opened this issue Aug 2, 2021 · 9 comments
Closed

default on LLVM_IAS=1 #1434

nickdesaulniers opened this issue Aug 2, 2021 · 9 comments
Assignees
Labels
feature-request Not a bug per-se [FIXED][LINUX] 5.15 This bug was fixed in Linux 5.15

Comments

@nickdesaulniers
Copy link
Member

now that we have LLVM_IAS=1 enabled for MIPS in CI, that leaves just s390 and powerpc without LLVM_IAS=1 support. Let's turn it on by default and have uses opt out explicitly via LLVM_IAS=0.

@nickdesaulniers nickdesaulniers added the feature-request Not a bug per-se label Aug 2, 2021
@nickdesaulniers nickdesaulniers self-assigned this Aug 2, 2021
@nathanchance
Copy link
Member

Something to keep in mind is that ARCH=arm and ARCH=riscv did not become buildable with LLVM_IAS=1 until LLVM 13.0.0. ARCH=arm64 and ARCH=x86_64 are buildable with LLVM_IAS=1 since LLVM 11.0.0 so a rhetorical question: should this be version dependent?

@nickdesaulniers
Copy link
Member Author

It's a good point and question.

should this be version dependent?

I think we can simply require folks set LLVM_IAS=0 CROSS_COMPILE=arm-linux-gnueabihf- etc for those older toolchains. This encourages folks to upgrade their tools to have nice things; in this case, a shorter command line.

@nickdesaulniers
Copy link
Member Author

@nickdesaulniers nickdesaulniers added the [PATCH] Submitted A patch has been submitted for review label Aug 2, 2021
@nathanchance
Copy link
Member

We should also check the various CI systems and see what versions of LLVM they are using along with how they are building the kernel, otherwise that patch could end up breaking them.

@nickdesaulniers
Copy link
Member Author

I mentioned this to @gctucker on Tuesday's KernelCI meeting. If they're not using LLVM=1 when building s390 or ppc, they should be fine.

Perhaps worthwhile to reach out to 0day folks as well.

@nathanchance
Copy link
Member

If they're not using LLVM=1 when building s390 or ppc, they should be fine.

Unless I am reading that patch wrong, that impacts CC=clang too.

@nickdesaulniers
Copy link
Member Author

v2: https://lore.kernel.org/lkml/20210806172701.3993843-1-ndesaulniers@google.com/

(probably should have added a Link tag to here)

@nickdesaulniers nickdesaulniers added [PATCH] Accepted A submitted patch has been accepted upstream and removed [PATCH] Submitted A patch has been submitted for review labels Aug 10, 2021
@nickdesaulniers
Copy link
Member Author

  1. 52cc02b
  2. f12b034

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
feature-request Not a bug per-se [FIXED][LINUX] 5.15 This bug was fixed in Linux 5.15
Projects
None yet
Development

No branches or pull requests

2 participants