[PATCH] D93586: [InstCombine] use poison as placeholder for undemanded elems

[Juneyoung Lee via Phabricator via cfe-commits]

aqjune created this revision.
aqjune added reviewers: spatel, lebedev.ri, efriedma.
Herald added subscribers: kerbowa, hiraditya, nhaehnle, jvesely.
aqjune requested review of this revision.
Herald added projects: clang, LLVM.
Herald added subscribers: llvm-commits, cfe-commits.

Currently undef is used as a don’t-care vector when constructing a vector using a series of insertelement.
However, this is problematic because undef isn’t undefined enough.
Especially, a sequence of insertelement can be optimized to shufflevector, but using undef as its placeholder makes shufflevector a poison-blocking instruction because undef cannot be optimized to poison.
This makes a few straightforward optimizations incorrect, such as:

  ;  https://bugs.llvm.org/show_bug.cgi?id=44185 
  
  define <4 x float> @insert_not_undef_shuffle_translate_commute(float %x, <4 x float> %y, <4 x float> %q) {
    %xv = insertelement <4 x float> %q, float %x, i32 2
    %r = shufflevector <4 x float> %y, <4 x float> %xv, <4 x i32> { 0, 6, 2, undef }
    ret <4 x float> %r ; %r[3] is undef
  }
  =>
  define <4 x float> @insert_not_undef_shuffle_translate_commute(float %x, <4 x float> %y, <4 x float> %q) {
    %r = insertelement <4 x float> %y, float %x, i32 1
    ret <4 x float> %r ; %r[3] = %y[3], incorrect if %y[3] = poison
  }
  
  Transformation doesn't verify!
  ERROR: Target is more poisonous than source

I’d like to suggest

1. Using poison as insertelement’s placeholder value (IRBuilder::CreateVectorSplat should be patched too)
2. Updating shufflevector’s semantics to return poison element if mask is undef

Note that poison is currently lowered into UNDEF in SelDag, so codegen part is okay.
m_Undef() matches PoisonValue as well, so existing optimizations will still fire.

The only concern is hidden miscompilations that will go incorrect when poison constant is given.
A conservative way is copying all tests having `insertelement undef` & replacing it with `insertelement poison` & run Alive2 on it, but it will create many tests and people won’t like it. :(

Instead, I’ll simply locally maintain the tests and run Alive2.
If there is any bug found, I’ll report it.

Relevant links: https://bugs.llvm.org/show_bug.cgi?id=43958 , http://lists.llvm.org/pipermail/llvm-dev/2019-November/137242.html


Repository:
  rG LLVM Github Monorepo

https://reviews.llvm.org/D93586

Files:
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector-constrained.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector2-constrained.c
  clang/test/CodeGen/SystemZ/builtins-systemz-zvector2.c
  llvm/lib/Transforms/InstCombine/InstCombineSimplifyDemanded.cpp
  llvm/test/CodeGen/AMDGPU/vector-alloca-bitcast.ll
  llvm/test/Transforms/InstCombine/AArch64/tbl1.ll
  llvm/test/Transforms/InstCombine/ARM/tbl1.ll
  llvm/test/Transforms/InstCombine/X86/clmulqdq.ll
  llvm/test/Transforms/InstCombine/X86/x86-addsub.ll
  llvm/test/Transforms/InstCombine/X86/x86-avx512.ll
  llvm/test/Transforms/InstCombine/X86/x86-insertps.ll
  llvm/test/Transforms/InstCombine/X86/x86-masked-memops.ll
  llvm/test/Transforms/InstCombine/X86/x86-muldq.ll
  llvm/test/Transforms/InstCombine/X86/x86-pack.ll
  llvm/test/Transforms/InstCombine/X86/x86-pshufb.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse2.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse41.ll
  llvm/test/Transforms/InstCombine/X86/x86-sse4a.ll
  llvm/test/Transforms/InstCombine/X86/x86-vec_demanded_elts.ll
  llvm/test/Transforms/InstCombine/X86/x86-vector-shifts.ll
  llvm/test/Transforms/InstCombine/X86/x86-vpermil.ll
  llvm/test/Transforms/InstCombine/X86/x86-xop.ll
  llvm/test/Transforms/InstCombine/bitcast-bigendian.ll
  llvm/test/Transforms/InstCombine/bitcast-vec-canon.ll
  llvm/test/Transforms/InstCombine/bitcast.ll
  llvm/test/Transforms/InstCombine/broadcast.ll
  llvm/test/Transforms/InstCombine/cast.ll
  llvm/test/Transforms/InstCombine/getelementptr.ll
  llvm/test/Transforms/InstCombine/insert-const-shuf.ll
  llvm/test/Transforms/InstCombine/insert-extract-shuffle.ll
  llvm/test/Transforms/InstCombine/masked_intrinsics.ll
  llvm/test/Transforms/InstCombine/minmax-fold.ll
  llvm/test/Transforms/InstCombine/pr2645-0.ll
  llvm/test/Transforms/InstCombine/shuffle_select.ll
  llvm/test/Transforms/InstCombine/shufflevector-div-rem.ll
  llvm/test/Transforms/InstCombine/sub-of-negatible.ll
  llvm/test/Transforms/InstCombine/trunc.ll
  llvm/test/Transforms/InstCombine/vec_demanded_elts.ll
  llvm/test/Transforms/InstCombine/vec_shuffle.ll
  llvm/test/Transforms/InstCombine/vector-casts.ll
  llvm/test/Transforms/InstCombine/vector_insertelt_shuffle.ll
  llvm/test/Transforms/LoopVectorize/X86/invariant-load-gather.ll
  llvm/test/Transforms/LoopVectorize/induction.ll
  llvm/test/Transforms/LoopVectorize/reduction-inloop-pred.ll
  llvm/test/Transforms/LoopVectorize/reduction-inloop.ll
  llvm/test/Transforms/PhaseOrdering/X86/horiz-math.ll
  llvm/test/Transforms/SLPVectorizer/AArch64/getelementptr.ll
  llvm/test/Transforms/SLPVectorizer/X86/alternate-cast.ll

-------------- next part --------------
A non-text attachment was scrubbed...
Name: D93586.312947.patch
Type: text/x-patch
Size: 217383 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20201219/516e55e9/attachment-0001.bin>