[all-commits] [llvm/llvm-project] 4f131b: [IR] Require index width to be ule pointer width (...

[Nikita Popov via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: 4f131b0d2214b5fbaef43d130e576b40a20bda32
      https://github.com/llvm/llvm-project/commit/4f131b0d2214b5fbaef43d130e576b40a20bda32
  Author: Nikita Popov <npopov at redhat.com>
  Date:   2023-10-26 (Thu, 26 Oct 2023)

  Changed paths:
    M llvm/docs/LangRef.rst
    M llvm/lib/IR/DataLayout.cpp
    M llvm/test/Analysis/ScalarEvolution/ptrtoint-constantexpr-loop.ll
    A llvm/test/Assembler/invalid-datalayout-index-size.ll
    M llvm/test/Transforms/InferAlignment/ptrmask.ll
    M mlir/test/Target/LLVMIR/Import/data-layout.ll
    M mlir/test/Target/LLVMIR/data-layout.mlir

  Log Message:
  -----------
  [IR] Require index width to be ule pointer width (#70015)

I don't think there is a use case for having an index type that is wider
than the pointer type, and I'm not entirely clear what semantics this
would even have.

Also clarify the GEP semantics to explicitly say how they interact with
the index type width.