<html>
  <head>
    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
  </head>
  <body text="#000000" bgcolor="#FFFFFF">
    <div class="moz-cite-prefix">On 4/24/2017 11:55 AM, Andrew Kelley
      wrote:<br>
    </div>
    <blockquote type="cite"
cite="mid:CACXN+=SiF-j2a5eQ4TceCJv7Px8_QRfT9MvOLRzwpJRXz4C8Fg@mail.gmail.com">
      <div dir="ltr"><br>
        <div class="gmail_extra"><br>
          <div class="gmail_quote">On Mon, Apr 24, 2017 at 2:53 PM,
            Friedman, Eli <span dir="ltr"><<a
                href="mailto:efriedma@codeaurora.org" target="_blank"
                moz-do-not-send="true">efriedma@codeaurora.org</a>></span>
            wrote:<br>
            <blockquote class="gmail_quote" style="margin:0 0 0
              .8ex;border-left:1px #ccc solid;padding-left:1ex"><span
                class="">On 4/23/2017 10:10 AM, Andrew Kelley via
                llvm-dev wrote:<br>
              </span>
              <blockquote class="gmail_quote" style="margin:0 0 0
                .8ex;border-left:1px #ccc solid;padding-left:1ex"><span
                  class="">
                  Here is a module:<br>
                  <br>
                  @vals = internal unnamed_addr global [20000000 x i32]
                  undef, align 4<br>
                  <br>
                </span><span class="">
                  LLVM puts the global in the .data section, and results
                  in a 77MB .o file of mostly zeroes. Why does this
                  variable not go in the .bss section?<br>
                  <br>
                </span></blockquote>
              <br>
              I think it's just an oversight; it doesn't matter for
              clang because it never emits globals like that.  See
              TargetLoweringObjectFile::getK<wbr>indForGlobal for the
              relevant logic.</blockquote>
            <div><br>
            </div>
            <div>I see, thanks. Would a patch be welcome which changed
              this behavior? Or I suppose I could emit zeroinitializer
              instead of undef. But it seems like undef is the more
              correct value.</div>
          </div>
          <br>
        </div>
      </div>
    </blockquote>
    <p>Sure, patch welcome.<br>
    </p>
    <p>-Eli<br>
    </p>
    <pre class="moz-signature" cols="72">-- 
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, a Linux Foundation Collaborative Project</pre>
  </body>
</html>