<html>
  <head>
    <meta content="text/html; charset=utf-8" http-equiv="Content-Type">
  </head>
  <body bgcolor="#FFFFFF" text="#000000">
    That part has been up for separate review for a while.  There was
    never really a consensus reached.  Looking back over the thread, it
    looks like we'd all ended up at the same place (Andy's last
    comment), but hadn't drawn a conclusion as to the patch itself.  <br>
    <pre wrap=""><a class="moz-txt-link-freetext" href="http://reviews.llvm.org/D4608">http://reviews.llvm.org/D4608</a></pre>
    <br>
    <div class="moz-cite-prefix">On 10/09/2014 01:39 PM, Kevin
      Modzelewski wrote:<br>
    </div>
    <blockquote
cite="mid:CAO=oM6sSEiJAguNJUb-uW4mB=c_=MeOqYS6HBeUYHap-NmVrgg@mail.gmail.com"
      type="cite">
      <div dir="ltr">Oh interesting, that seems to work -- assuming that
        other people like the approach, would you mind factoring out
        that part of your statepoint patch?</div>
      <div class="gmail_extra"><br>
        <div class="gmail_quote">On Thu, Oct 9, 2014 at 9:19 AM, Philip
          Reames <span dir="ltr"><<a moz-do-not-send="true"
              href="mailto:listmail@philipreames.com" target="_blank">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">
            <div bgcolor="#FFFFFF" text="#000000">
              <div>On 10/08/2014 09:40 PM, Kevin Modzelewski wrote:<br>
              </div>
              <blockquote type="cite">
                <pre>Hi atrick,

It didn't seem like there was anything special that had to be done besides
just defining the new intrinsic with the different return type.

It seems inelegant to create a new intrinsic for a new return type, but
we're running into a case where we'd like to return a double through a patchpoint
and I'm not sure of a better way to do that.</pre>
              </blockquote>
              You might want to take a look at how we defined the
              gc_result intrinsics in the statepoint patch which is
              currently up for review.  With some changes to the
              intrinsic name mangling, you can define an intrinsic per
              "family" of types.  The four we had to add were: void,
              int, pointer, float.  (float here means any floating point
              type, int means any integer type, etc..)
              <blockquote type="cite">
                <pre><a moz-do-not-send="true" href="http://reviews.llvm.org/D5696" target="_blank">http://reviews.llvm.org/D5696</a>

Files:
  docs/StackMaps.rst
  include/llvm/IR/Intrinsics.td
  lib/CodeGen/SelectionDAG/FastISel.cpp
  lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
  lib/CodeGen/SelectionDAG/SelectionDAGISel.cpp
  lib/IR/Verifier.cpp
  lib/Target/AArch64/AArch64TargetTransformInfo.cpp
  lib/Target/X86/X86TargetTransformInfo.cpp
  test/CodeGen/X86/patchpoint.ll
</pre>
                <br>
                <fieldset></fieldset>
                <br>
                <pre>_______________________________________________
llvm-commits mailing list
<a moz-do-not-send="true" href="mailto:llvm-commits@cs.uiuc.edu" target="_blank">llvm-commits@cs.uiuc.edu</a>
<a moz-do-not-send="true" href="http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits" target="_blank">http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits</a>
</pre>
              </blockquote>
              <br>
            </div>
          </blockquote>
        </div>
        <br>
      </div>
    </blockquote>
    <br>
  </body>
</html>