[apparmor] [patch] fixes for LP: #775785, full /tmp reload issue

[Marc Deslauriers]

On Wed, 2011-08-24 at 12:01 -0700, Steve Beattie wrote:
> On Thu, Aug 18, 2011 at 06:24:41PM -0700, Kees Cook wrote:
> > On Mon, Aug 15, 2011 at 04:03:42PM -0700, Steve Beattie wrote:
> > > Simplification, I believe. I keep meaning to remove a lot of the
> > > historical cruft in the upstream initscripts, but it hasn't percolated
> > > to the top of the work queue.
> > > 
> > > I would like it to happen so that Ubuntu can use the upstream scripts,
> > > because I'd like the functions file to be shared across distros so
> > > that we only need to fix the majority of initscript bugs in one place.
> > 
> > The initscript fork is entirely my fault. :( In an earlier Ubuntu release
> > (Karmic?) we had some very specific design goals about boot speed and
> > script locations (it was split between initramfs and early boot at one
> > point and had to have no deps on /usr, etc, etc), and it seemed easier
> > to do a "minimal" fork of the init scripts to handle all of Ubuntu's
> > requirements at the time instead of breaking all the other distros and/or
> > suddenly ditching all the historical checks. Since then, things have
> > split out a bit more sanely, and I have hope that we're getting close
> > to being able to have a common initscript again.
> 
> Hrm, no deps on /usr may be a bit of an issue with my patch; awk (on
> Ubuntu and OpenSUSE) is installed in /usr/bin/.
> 
> Sigh.
> 
> Granted, it's only for reloading profiles, which shouldn't typically
> occur in an early boot situation where /usr wouldn't be mounted. And
> 'sort' is also a /usr/bin/ utility, which already exists in the script.

Having a separate /usr isn't supported on Linux anymore anyway...or at
least it breaks everything...

http://thread.gmane.org/gmane.comp.sysutils.systemd.devel/1337

Marc.