[PATCH] Reject bitcasts between address spaces with different sizes

Eli Friedman eli.friedman at gmail.com
Mon Jul 22 18:50:27 PDT 2013


Is the change to test/Transforms/InstCombine/memset2.ll still necessary?

-Eli


On Mon, Jul 22, 2013 at 3:49 PM, Matt Arsenault
<Matthew.Arsenault at amd.com> wrote:
>   Remove leftover LLParser change.
>
>   Remove redundant global use checking.
>
>   Add note to LangRef that non-zero address spaces by default are assumed to be the same size/alignment as address space 0.
>
> Hi eli.friedman,
>
> http://llvm-reviews.chandlerc.com/D1083
>
> CHANGE SINCE LAST DIFF
>   http://llvm-reviews.chandlerc.com/D1083?vs=2952&id=2953#toc
>
> Files:
>   docs/LangRef.rst
>   lib/IR/Verifier.cpp
>   test/Transforms/InstCombine/memset2.ll
>   test/Verifier/bitcast-address-space-nested-global-cycle.ll
>   test/Verifier/bitcast-address-space-nested-global.ll
>   test/Verifier/bitcast-address-space-through-constant-inttoptr-inside-gep-instruction.ll
>   test/Verifier/bitcast-address-space-through-constant-inttoptr.ll
>   test/Verifier/bitcast-address-space-through-gep-2.ll
>   test/Verifier/bitcast-address-space-through-gep.ll
>   test/Verifier/bitcast-address-space-through-inttoptr.ll
>   test/Verifier/bitcast-address-spaces.ll
>   test/Verifier/bitcast-vector-pointer-as.ll



More information about the llvm-commits mailing list