<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 2/6/2015 6:52 AM, Rafael Espíndola
wrote:<br>
</div>
<blockquote
cite="mid:CAG3jReLc3=88_kNdGu=XBok9XJMWnd8zQwr+=ei48988M6gzNg@mail.gmail.com"
type="cite">
<pre wrap="">On 5 February 2015 at 19:12, Shankar Easwaran <a class="moz-txt-link-rfc2396E" href="mailto:shankare@codeaurora.org"><shankare@codeaurora.org></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">You need to keep the group if you are referring to one function in the
group. Its a bug with gold/gnu if its throwing away the whole group.
</pre>
</blockquote>
<pre wrap="">
No, it is not.
Please point to the part of the ELF spec that supports your statement.
</pre>
</blockquote>
The resolver regards the group as one unit and the garbage
collection should also treat them as one unit.<br>
<br>
The message from Cary here has the snip.<br>
<br>
<a class="moz-txt-link-freetext" href="https://sourceware.org/ml/binutils/2012-01/msg00119.html">https://sourceware.org/ml/binutils/2012-01/msg00119.html</a><br>
<br>
<snip><br>
<br>
Even better, to my mind, we could use (both COMDAT and non-COMDAT)<br>
group sections to tie the code and debug info (and static data, and<br>
eh_frame info, etc.) together into a single group. <b>Garbage
collection</b><b><br>
</b><b>could then treat each such group as an atomic unit: any
reference to a</b><b><br>
</b><b>group member section makes the whole group reachable</b>.
Back when we<br>
added COMDAT group sections to the ELF format, we had in mind<br>
something like this, which is why each group section begins with a<br>
flag word with GRP_COMDAT set to identify a COMDAT group.<br>
<br>
</snip><br>
<br>
Shankar Easwaran<br>
<pre class="moz-signature" cols="72">--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by the Linux Foundation</pre>
</body>
</html>