<html>
<head>
<base href="https://llvm.org/bugs/" />
</head>
<body><table border="1" cellspacing="0" cellpadding="8">
<tr>
<th>Bug ID</th>
<td><a class="bz_bug_link
bz_status_NEW "
title="NEW --- - SLP causes a miscompile"
href="https://llvm.org/bugs/show_bug.cgi?id=31243">31243</a>
</td>
</tr>
<tr>
<th>Summary</th>
<td>SLP causes a miscompile
</td>
</tr>
<tr>
<th>Product</th>
<td>libraries
</td>
</tr>
<tr>
<th>Version</th>
<td>trunk
</td>
</tr>
<tr>
<th>Hardware</th>
<td>PC
</td>
</tr>
<tr>
<th>OS</th>
<td>All
</td>
</tr>
<tr>
<th>Status</th>
<td>NEW
</td>
</tr>
<tr>
<th>Severity</th>
<td>normal
</td>
</tr>
<tr>
<th>Priority</th>
<td>P
</td>
</tr>
<tr>
<th>Component</th>
<td>Scalar Optimizations
</td>
</tr>
<tr>
<th>Assignee</th>
<td>unassignedbugs@nondot.org
</td>
</tr>
<tr>
<th>Reporter</th>
<td>michael.v.zolotukhin@gmail.com
</td>
</tr>
<tr>
<th>CC</th>
<td>llvm-bugs@lists.llvm.org
</td>
</tr>
<tr>
<th>Classification</th>
<td>Unclassified
</td>
</tr></table>
<p>
<div>
<pre>Created <span class=""><a href="attachment.cgi?id=17705" name="attach_17705" title="Reproducer">attachment 17705</a> <a href="attachment.cgi?id=17705&action=edit" title="Reproducer">[details]</a></span>
Reproducer
Reproducer (slp_bug.ll is attached):
<span class="quote">> llc -O0 slp_bug.ll -o good.s
> clang -m32 -o good good.s</span >
<span class="quote">> opt -slp-vectorizer slp_bug.ll -o after_slp.ll -S
> llc -O0 after_slp.ll -o bad.s
> clang -m32 -o bad bad.s</span >
<span class="quote">> ./good
> echo $?</span >
106
<span class="quote">> ./bad
> echo $?</span >
0
The bug is that we sign-extend values, which should be zero-extended instead.
For instance, here:
%10 = zext <4 x i8> %9 to <4 x i32> ; Original value is zero-extended
%11 = trunc <4 x i32> %10 to <4 x i8> ; Now we truncate it...
%12 = extractelement <4 x i8> %11, i32 0 ; extract first element...
%13 = sext i8 %12 to i32 ; and sign-extend it.
...
The issue seems to be introduced in r258929.</pre>
</div>
</p>
<hr>
<span>You are receiving this mail because:</span>
<ul>
<li>You are on the CC list for the bug.</li>
</ul>
</body>
</html>