<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <p>In the scenario you gave, mul is a binary operator, 0 is not. 
      Calling tryReassociateBinary on the operand would be suspicious. 
      <br>
    </p>
    %1 = mul nsw i8 42, %0<br>
    <br>
    (I'd also not expect '42' to be a binary operator in the above. 
    Having i8 42 be an argument to tryReassociateBinaryOp would also
    seem wrong.)<br>
    <br>
    <div class="moz-cite-prefix">On 03/08/2018 02:43 PM, Evgeny
      Stupachenko wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CAOvf_xwP0gNyimhy4kCJfFeiejtxk54QzwJWrc24Tat4tWZcKQ@mail.gmail.com">
      <div dir="ltr">Well, is
        <div><br>
          <div>
            <span
style="color:rgb(34,34,34);font-family:arial,sans-serif;font-size:12.8px;font-style:normal;font-variant-ligatures:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;background-color:rgb(255,255,255);text-decoration-style:initial;text-decoration-color:initial;float:none;display:inline"><span> </span>%1
              = mul nsw i8 0, %0</span><br>
          </div>
        </div>
        <div><span style="font-size:12.8px"><br>
          </span></div>
        <div><span style="font-size:12.8px">binary operator?</span><br>
        </div>
        <div>It has Zero SCEV.</div>
        <div><br>
        </div>
      </div>
      <div class="gmail_extra"><br>
        <div class="gmail_quote">On Thu, Mar 8, 2018 at 1:57 PM, Philip
          Reames <span dir="ltr"><<a
              href="mailto:listmail@philipreames.com" target="_blank"
              moz-do-not-send="true">listmail@philipreames.com</a>></span>
          wrote:<br>
          <blockquote class="gmail_quote" style="margin:0 0 0
            .8ex;border-left:1px #ccc solid;padding-left:1ex">Huh?  Zero
            is not a binary operator.  So how'd it get here?  This patch
            looks wrong or at least incomplete.<span class="HOEnZb"><font
                color="#888888"><br>
                <br>
                Philip</font></span>
            <div class="HOEnZb">
              <div class="h5"><br>
                <br>
                <br>
                On 03/06/2018 06:17 PM, Evgeny Stupachenko via
                llvm-commits wrote:<br>
                <blockquote class="gmail_quote" style="margin:0 0 0
                  .8ex;border-left:1px #ccc solid;padding-left:1ex">
                  Author: evstupac<br>
                  Date: Tue Mar  6 18:17:08 2018<br>
                  New Revision: 326861<br>
                  <br>
                  URL: <a
                    href="http://llvm.org/viewvc/llvm-project?rev=326861&view=rev"
                    rel="noreferrer" target="_blank"
                    moz-do-not-send="true">http://llvm.org/viewvc/llvm-pr<wbr>oject?rev=326861&view=rev</a><br>
                  Log:<br>
                  Add early exit on reassociation of 0 expression.<br>
                  <br>
                  Summary:<br>
                  <br>
                  Before the patch a try to reassociate ((v * 16) * 0) *
                  1 fall into infinite loop<br>
                  <br>
                  Reviewers: pankajchawla<br>
                  <br>
                  Differential Revision: <a
                    href="http://reviews.llvm.org/D41467"
                    rel="noreferrer" target="_blank"
                    moz-do-not-send="true">http://reviews.llvm.org/D41467</a><br>
                  <br>
                  From: Evgeny Stupachenko <<a
                    href="mailto:evstupac@gmail.com" target="_blank"
                    moz-do-not-send="true">evstupac@gmail.com</a>><br>
                                            <<a
                    href="mailto:evgeny.v.stupachenko@intel.com"
                    target="_blank" moz-do-not-send="true">evgeny.v.stupachenko@intel.co<wbr>m</a>><br>
                  <br>
                  Added:<br>
                       llvm/trunk/test/Transforms/Na<wbr>ryReassociate/pr35710.ll<br>
                  Modified:<br>
                       llvm/trunk/lib/Transforms/Sca<wbr>lar/NaryReassociate.cpp<br>
                  <br>
                  Modified: llvm/trunk/lib/Transforms/Scal<wbr>ar/NaryReassociate.cpp<br>
                  URL: <a
href="http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Transforms/Scalar/NaryReassociate.cpp?rev=326861&r1=326860&r2=326861&view=diff"
                    rel="noreferrer" target="_blank"
                    moz-do-not-send="true">http://llvm.org/viewvc/llvm-pr<wbr>oject/llvm/trunk/lib/Transform<wbr>s/Scalar/NaryReassociate.cpp?<wbr>rev=326861&r1=326860&r2=<wbr>326861&view=diff</a><br>
                  ==============================<wbr>==============================<wbr>==================<br>
                  --- llvm/trunk/lib/Transforms/Scal<wbr>ar/NaryReassociate.cpp
                  (original)<br>
                  +++ llvm/trunk/lib/Transforms/Scal<wbr>ar/NaryReassociate.cpp
                  Tue Mar  6 18:17:08 2018<br>
                  @@ -429,6 +429,9 @@ NaryReassociatePass::tryReasso<wbr>ciateGEPAt<br>
                      Instruction *NaryReassociatePass::tryReass<wbr>ociateBinaryOp(BinaryOperator
                  *I) {<br>
                      Value *LHS = I->getOperand(0), *RHS =
                  I->getOperand(1);<br>
                  +  // There is no need to reassociate 0.<br>
                  +  if (SE->getSCEV(I)->isZero())<br>
                  +    return nullptr;<br>
                      if (auto *NewI = tryReassociateBinaryOp(LHS, RHS,
                  I))<br>
                        return NewI;<br>
                      if (auto *NewI = tryReassociateBinaryOp(RHS, LHS,
                  I))<br>
                  <br>
                  Added: llvm/trunk/test/Transforms/Nar<wbr>yReassociate/pr35710.ll<br>
                  URL: <a
href="http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Transforms/NaryReassociate/pr35710.ll?rev=326861&view=auto"
                    rel="noreferrer" target="_blank"
                    moz-do-not-send="true">http://llvm.org/viewvc/llvm-pr<wbr>oject/llvm/trunk/test/Transfor<wbr>ms/NaryReassociate/pr35710.ll?<wbr>rev=326861&view=auto</a><br>
                  ==============================<wbr>==============================<wbr>==================<br>
                  --- llvm/trunk/test/Transforms/Nar<wbr>yReassociate/pr35710.ll
                  (added)<br>
                  +++ llvm/trunk/test/Transforms/Nar<wbr>yReassociate/pr35710.ll
                  Tue Mar  6 18:17:08 2018<br>
                  @@ -0,0 +1,19 @@<br>
                  +; NOTE: Assertions have been autogenerated by
                  utils/update_test_checks.py<br>
                  +; RUN: opt < %s -nary-reassociate -S | FileCheck
                  %s<br>
                  +<br>
                  +; The test check that compilation does not fall into
                  infinite loop.<br>
                  +<br>
                  +define i8 @foo(i8 %v) local_unnamed_addr #0 {<br>
                  +; CHECK-LABEL: @foo(<br>
                  +; CHECK-NEXT:  region.0:<br>
                  +; CHECK-NEXT:    [[TMP0:%.*]] = mul nsw i8 16,
                  [[V:%.*]]<br>
                  +; CHECK-NEXT:    [[TMP1:%.*]] = mul nsw i8 0,
                  [[TMP0]]<br>
                  +; CHECK-NEXT:    [[TMP2:%.*]] = mul nsw i8 1,
                  [[TMP1]]<br>
                  +; CHECK-NEXT:    ret i8 [[TMP2]]<br>
                  +;<br>
                  +region.0:<br>
                  +  %0 = mul nsw i8 16, %v<br>
                  +  %1 = mul nsw i8 0, %0<br>
                  +  %2 = mul nsw i8 1, %1<br>
                  +  ret i8 %2<br>
                  +}<br>
                  <br>
                  <br>
                  ______________________________<wbr>_________________<br>
                  llvm-commits mailing list<br>
                  <a href="mailto:llvm-commits@lists.llvm.org"
                    target="_blank" moz-do-not-send="true">llvm-commits@lists.llvm.org</a><br>
                  <a
                    href="http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-commits"
                    rel="noreferrer" target="_blank"
                    moz-do-not-send="true">http://lists.llvm.org/cgi-bin/<wbr>mailman/listinfo/llvm-commits</a><br>
                </blockquote>
                <br>
              </div>
            </div>
          </blockquote>
        </div>
        <br>
      </div>
    </blockquote>
    <br>
  </body>
</html>