<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body text="#000000" bgcolor="#FFFFFF">
Le 13/05/16 à 17:16, Yuri D'Elia a écrit :<br>
<blockquote cite="mid:87bn4aouo5.fsf@wavexx.thregr.org" type="cite">
<pre wrap="">On Fri, May 13 2016, Laurent Bigonville <a class="moz-txt-link-rfc2396E" href="mailto:bigon@debian.org"><bigon@debian.org></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Again this is supposed to happen at early boot, and at this stage, only
PID1 exists. So I doubt there is a lot of concurrent processes at that time.
</pre>
</blockquote>
<pre wrap="">But this is not checked in the source.
In fact, this behavior will happen irregardless of the boot stage.</pre>
</blockquote>
<br>
It doesn't do anything if /proc is already mounted, which is IMHO
expected on a fully booted machine and even container.<br>
<br>
<blockquote cite="mid:87bn4aouo5.fsf@wavexx.thregr.org" type="cite"><br>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap="">Even if the fix is simply the removal of the mountpoint, I consider the
solution broken by design.
</pre>
</blockquote>
<pre wrap="">What about mounting /proc really early?
</pre>
</blockquote>
<pre wrap="">I can say the same about initramfs. Can't initramfs just mount /proc
sooner and fix the problem correctly?
</pre>
</blockquote>
<br>
initramfs in debian is, like I said, already doing so.<br>
<br>
SELinux upstream is probably supporting more usecases, apparently
not yours.<br>
<br>
[...]<br>
<br>
<blockquote cite="mid:87bn4aouo5.fsf@wavexx.thregr.org" type="cite">
<pre wrap="">I, honestly, expected someone that understand the issue to help and
chime to report it upstream.
</pre>
</blockquote>
<br>
But you didn't really explained me what the usecase was and why you
were not able to mount /proc early...<br>
<br>
Anyway, I already reported it upstream, let's see if the come with a
patch.<br>
</body>
</html>