<table border="1" cellspacing="0" cellpadding="8">
    <tr>
        <th>Issue</th>
        <td>
            <a href=https://github.com/llvm/llvm-project/issues/82903>82903</a>
        </td>
    </tr>

    <tr>
        <th>Summary</th>
        <td>
            Merge 99c457dc2ef395872d7448c85609f6cb73a7f89b into 18.x
        </td>
    </tr>

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

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

    <tr>
      <th>Reporter</th>
      <td>
          brad0
      </td>
    </tr>
</table>

<pre>
    Unbreak *tf builtins for hexfloat (#82208)

This re-lands cc0065a7d082f0bd322a538cf62cfaef1c8f89f8 in a way that
keeps existing targets working.

---------

Original commit message:
#68132 ended up removing
__multc3 & __divtc3 from compiler-rt library builds that have
QUAD_PRECISION but not TF_MODE due to missing int128 support.
I added support for QUAD_PRECISION to use the native hex float long double representation.

---------
</pre>
<img width="1px" height="1px" alt="" src="http://email.email.llvm.org/o/eJxsUk2P5CYQ_TX4UuoWLmw3HHyYZLalOWwmH7vnFoayTQaDBbhn5t9H7u4oq2h9MKJKvHr13tM5uykQ9az9hbXPld7KHFM_JG15NUT72X8PQyL9BgyfygjD5nxxIcMYE8z0MfqoCzCUDIVE5JKhYvyZ8af7_9vsMiQ6eB1sBmM471p9slziyAcrEHUrpBk7NKOmsTZylGqU4AJoeNefUGZd7khvRGsG-nC5uDBB0WmikuE9pjcXpuOPQw__fj8WX5ObXNAeTFwWV2ChnPVETDz6DEUna4FAwZKFbYVES7y6MN37l8uy-WIEMOzgcrHuul_GFJcdcXWe0iEV8G5IOn3ehLL5xh9mfaU7yB_fn54vv__55deXv15ef4NhKxBigW_ny9fX5y9gN4ISYXE57zu6UGqUkLd1jak8VnwBbXeCj-rNiP_BlghbJigzQdDFXWl3Cu5W-RgmsHEbPEGiNVGmUHRxMfxcwsr2wiqhdEV9feJSigYVr-Z-UEJRK4WuOTUdIre2FlwNHUkuVGMq1yPHhiO2vKlrwY-G11LJrh2MQj6eWtZwWrTzR--vyzGmqXI5b9RLVFxUXg_k8y2XiIHe4dZkiHtMU7-_OQzblFnDvcsl_4dSXPHUf6U0EShlmvZkDdIoVCtPaE9NI41sO67GzgwnoU-jVMOudYRaHj-qLfl-LmXNezTwzPA8uTJvw9HEheF5H_M4DmuKf5MpDM83cpnh-Ub-nwAAAP__eo0JPw">