<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<p>Sorry for the delay on this... Please find my initial patch
proposal at <a class="moz-txt-link-freetext" href="https://reviews.llvm.org/D112942">https://reviews.llvm.org/D112942</a>.</p>
<p>/Jonas</p>
<p><br>
</p>
<div class="moz-cite-prefix">On 2021-10-07 19:39, Ulrich Weigand
wrote:<br>
</div>
<blockquote type="cite"
cite="mid:OF4C67E272.90CF8F33-ONC1258767.0060E216-C1258767.006109A0@ibm.com">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<p><tt><font size="2">"Reid Kleckner" <a class="moz-txt-link-rfc2396E" href="mailto:rnk@google.com"><rnk@google.com></a> wrote
on 06.10.2021 19:00:44:<br>
> On Wed, Oct 6, 2021 at 9:15 AM Ulrich Weigand
<a class="moz-txt-link-rfc2396E" href="mailto:Ulrich.Weigand@de.ibm.com"><Ulrich.Weigand@de.ibm.com></a> wrote:</font></tt><br>
<tt><font size="2">>> As I commented in another mail,
unfortunately this is not true.<br>
>> To ensure correct codegen in all cases, we need to
know *which*<br>
>> extension the callee expects: zero- or
sign-extension. The<br>
>> backend cannot know this without being told by the
frontend.</font></tt><br>
<br>
<tt><font size="2">> Maybe this is just a wording thing, but
I do believe my statement is<br>
> true: the backend is absolutely free to guess if it
wants to, and <br>
> requiring attributes doesn't guarantee correctness. It
may just move<br>
> the bug to the frontend. I acknowledge that there are
risks to <br>
> guessing. I just think it is important to consider
alternative <br>
> designs and weigh the pros and cons.</font></tt><br>
<br>
<tt><font size="2">I guess it's indeed just wording, and we
agree on the substance.</font></tt><br>
<br>
<tt><font size="2">Thanks,</font></tt><br>
<tt><font size="2">Ulrich</font></tt><br>
<br>
<br>
</p>
</blockquote>
</body>
</html>