<html>
  <head>
    <meta content="text/html; charset=windows-1252"
      http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    <div class="moz-cite-prefix">On 02/26/2015 03:09 AM, Anastasia
      Stulova wrote:<br>
    </div>
    <blockquote cite="mid:000001d051b4$b806c940$28145bc0$@arm.com"
      type="cite">
      <p class="MsoNormal"><span style="color:#1F497D">Hi Matt,<o:p></o:p></span></p>
      <p class="MsoNormal"><span style="color:#1F497D"><o:p> </o:p></span></p>
      <p class="MsoNormal"><span style="color:#1F497D">Thanks for your
          comments. <o:p></o:p></span></p>
      <p class="MsoNormal"><span style="color:#1F497D"><o:p> </o:p></span></p>
      <p class="MsoNormal"><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif";color:#1F497D">> </span><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif"">I think this needs a test that
          includes checking the emitted IR.<br>
          <br>
        </span><span style="font-size:12.0pt;font-family:"Times New
          Roman","serif";color:#1F497D"><o:p></o:p></span></p>
      <p class="MsoNormal"><span style="color:#1F497D">No problem. I
          will add that!<o:p></o:p></span></p>
      <p class="MsoNormal"><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif""><br>
        </span><span style="font-size:12.0pt;font-family:"Times New
          Roman","serif";color:#1F497D">> </span><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif"">Would it be reasonable to fix
          this more generally for all builtins? I know this is a problem
          for many others that accept pointers</span><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif";color:#1F497D"><o:p></o:p></span></p>
      <p class="MsoNormal"><span style="color:#1F497D"><o:p> </o:p></span></p>
      <p class="MsoNormal"><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif""><br>
        </span><span style="font-size:12.0pt;font-family:"Times New
          Roman","serif";color:#1F497D">This fix is not
          specific to c11 compare exchange atomic, but to atomics with a
          pointer as second parameter. Do you mean generalization for
          other parameters here? Could you tell me what atomic functions
          should I be looking at?<o:p></o:p></span></p>
      <p class="MsoNormal"><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif";color:#1F497D"> <o:p></o:p></span></p>
      <p class="MsoNormal"><span
          style="font-size:12.0pt;font-family:"Times New
          Roman","serif";color:#1F497D">Anastasia</span></p>
    </blockquote>
    I mean generalization for other builtins with pointer operands. For
    example, Tom is currently trying to fix __builtin_nanf for OpenCL.
    Strings in OpenCL are in the constant address space, but the type
    checking of the builtin assumes implicitly assumes the pointer
    should be address space 0, so the builtin is incorrectly rejected<br>
    <br>
    -Matt<br>
  </body>
</html>