<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/68338>68338</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            Conversion of _Float16 for variadic functions doesn't seem to work
        </td>
    </tr>

    <tr>
      <th>Labels</th>
      <td>
            new issue
      </td>
    </tr>

    <tr>
      <th>Assignees</th>
      <td>
      </td>
    </tr>

    <tr>
      <th>Reporter</th>
      <td>
          sh1boot
      </td>
    </tr>
</table>

<pre>
    Converting `_Float16` to `double` to pass to a regular function looks fine, but doing the conversion for a variadic function (like `printf()` does not work.

`__fp16` gets converted to `double` just fine in both cases.

https://godbolt.org/z/r5o7djxbd

Observed on RISC-V and aarch64.
</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJxkkstu3DoMhp9G3ggZyPR94UVOAgNnVaAFug0ki7aV0YgDiZ60ffpC7hRpm40IXcj_Fz_qlNwaEEfR_Cea50LvvFEc01YaIi4M2e_jE4UbRnZhlaJVL5MnzWUrWiWZ8oml3Xi87686pRy1jLjuXke57GFmR0F6onOSiwso4EmanaWlXJQ3lPOhkfKzhaLU8qaj09bN7-kCeu_OmBWv0QVeBPQChqxrCZMMxPKN4vkk1LNQj_e1VS8vy_WX3RU53ZUY7Qf3r3viw550QRriTc46Yfqr3sZ8TaJ6FDAJmFayhjyfKK4Cph8CpthQZ1-_Gftn0ieTMN7QSgry8_9fnh6-Sh2s1DrOW1ufCjtWdqgGXeBYtkPbQK9UWWwjdBYQlm429bLoBkpbNks3L7rq0dZNV7gRFFSlUk3Z1W1dnobOYNUMtlNaQd1VolZ40c6fvL9dss3CpbTj2PZV1RdeG_TpQA8Q8E0elwIgT0Icc86D2dckauVd4vRehR17vA_GAY0W-XswDoAf8KUDUhDQsUyIl9z9TKvYox__6arjbTenmS4Cpix5Dw_XSK84s4DpMJoETMdHfgYAAP__Om7kAA">