<html>
<head>
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<p><tt>Hi all,</tt></p>
<p><tt>I'm interested in participating this GSoC with llvm. I would
like to contribute on the project idea "DebugInfo should not
effect codegen". Over the past few days, I have gone through
some bugs reported at bugzilla that cause different codegen
behavior of same program. <br>
</tt></p>
<p><tt>Bugs I have been looking at are: <br>
</tt></p>
<p><tt> 1. <a moz-do-not-send="true"
href="https://bugs.llvm.org/show_bug.cgi?id=37306">[fuzzDI]
-O1 + `-g` cause the generated code to change.</a><br>
</tt></p>
<p><tt> 2. <a moz-do-not-send="true"
href="https://bugs.llvm.org/show_bug.cgi?id=37728">Make llvm
passes debug info invariant</a><br>
</tt></p>
<p><tt> 3. <a moz-do-not-send="true"
href="https://bugs.llvm.org/show_bug.cgi?id=37727">LoopVectorize:
different IR generation when debug info is present</a><br>
</tt></p>
<p><tt> 4. <a moz-do-not-send="true"
href="https://bugs.llvm.org/show_bug.cgi?id=37714">InstCombine:
test/Transforms/InstCombine/call-guard.ll looks different with
debug info present</a><br>
</tt></p>
<p><tt>I'm currently working on finding the root cause of these
issues. I have some questions regarding the scope of the
project.</tt> </p>
<pre> 1) Whether the scope is finding as many issue as possible and fix it or whether I can design some protocol like stuff which impose constrains on optimization passes to interact with debug
info metadata, by this way we can far more generalize this idea and make a clear distinction between opt passes which operate on llvm ir & debug metadata.
I highly appreciate your thoughts on it.
Thanks
</pre>
<pre class="moz-signature" cols="72">--
Have a great day!
PreeJackie
</pre>
</body>
</html>