[all-commits] [llvm/llvm-project] a323df: Don't sink ptrtoint/inttoptr sequences into non-no...

[Tim Besard via All-commits]

  Branch: refs/heads/main
  Home:   https://github.com/llvm/llvm-project
  Commit: a323dfc0152a11c6934695b1ae67920dae9fad9b
      https://github.com/llvm/llvm-project/commit/a323dfc0152a11c6934695b1ae67920dae9fad9b
  Author: Tim Besard <tim at juliacomputing.com>
  Date:   2022-07-16 (Sat, 16 Jul 2022)

  Changed paths:
    M llvm/lib/CodeGen/CodeGenPrepare.cpp
    A llvm/test/Transforms/CodeGenPrepare/NVPTX/dont-introduce-addrspacecast.ll

  Log Message:
  -----------
  Don't sink ptrtoint/inttoptr sequences into non-noop addrspacecasts.

In https://reviews.llvm.org/D30114, support for mismatching address
spaces was introduced to CodeGenPrepare's optimizeMemoryInst, using
addrspacecast as it was argued that only no-op addrspacecasts would be
considered when constructing the address mode. However, by doing
inttoptr/ptrtoint, it's possible to get CGP to emit an addrspace
that's not actually no-op, introducing a miscompilation:

define void @kernel(i8* %julia_ptr) {
  %intptr = ptrtoint i8* %julia_ptr to i64
  %ptr = inttoptr i64 %intptr to i32 addrspace(3)*

  br label %end
end:

  store atomic i32 1, i32 addrspace(3)* %ptr unordered, align 4
  ret void
}

Gets compiled to:

define void @kernel(i8* %julia_ptr) {
end:
  %0 = addrspacecast i8* %julia_ptr to i32 addrspace(3)*
  store atomic i32 1, i32 addrspace(3)* %0 unordered, align 4
  ret void
}

In the case of NVPTX, this introduces a cvta.to.shared, whereas
leaving out the %end block and branch doesn't trigger this
optimization. This results in illegal memory accesses as seen in
https://github.com/JuliaGPU/CUDA.jl/issues/558

In this change, I introduced a check before doing the pointer cast
that verifies address spaces are the same. If not, it emits a
ptrtoint/inttoptr combination to get a no-op cast between address
spaces. I decided against disallowing ptrtoint/inttoptr with
non-default AS in matchOperationAddr, because now its still possible
to look through multiple sequences of them that ultimately do not
result in a address space mismatch (i.e. the second lit test).