[OE-core] Regression bug: dbus messagebus user generation is wrong

[Richard Purdie]

On Thu, 2011-12-22 at 11:26 +0100, Koen Kooi wrote:
> Op 22 dec. 2011, om 10:10 heeft Martin Jansa het volgende geschreven:
> 
> > On Thu, Dec 22, 2011 at 09:49:01AM +0100, Bernhard Guillon wrote:
> >> Hi all,
> >> we switched the way we chown /usr/libexec/dbus-daemon-launch-helper
> >> with this commit:
> >> 
> >> commit 46e6c3fa8034b12d178d605f3f5d7efe69671a13
> >> Author: Otavio Salvador <otavio at ossystems.com.br>
> >> Date:   Fri Oct 21 02:49:51 2011 +0000
> >> 
> >>     dbus: use useradd class to allow use in read-only filesystems
> >> 
> >>     Move creation of required user/groups to useradd class thus allowing
> >>     use with read-only filesystems and booting the initial boot.
> >> 
> >>     Signed-off-by: Otavio Salvador <otavio at ossystems.com.br>
> >> 
> >> 
> >> We changed the owner ship of the user on the device before this commit. 
> >> Now we change the owner ship on build time which is wrong. If we change 
> >> the owner ship on buildtime we cannot know the gid of the messagebus 
> >> group. On my system it is e.g 117 but on my targtet it is 997. There is no 
> >> standard mapping for gid to names so we need to switch back to change the 
> >> group on the device.
> > 
> > See http://bugzilla.pokylinux.org/show_bug.cgi?id=1711
> > but it's still broken in some cases (on my buildhost 3 from 5 machines
> > are broken, 2 are working after last opkg changes).
> > 
> > Good test is to enable testlab and run this in your deploy/images directory
> > find . ! -type l -name \*tar.gz`; do 
> >  echo $i; 
> >  tar -tvf $i | grep dbus-daemon-launch; 
> >  tar --numeric-owner -tvf $i | grep dbus-daemon-launch; 
> >  tar xzvpf $i ./etc/group; 
> >  grep messagebus ./etc/group; 
> > done | tee -a image.test
> > 
> > And working image needs to have not only messagebus as owner but also
> > matching GID in 2nd line and last line from packaged /etc/group. With
> > latest opkg fixes it seems to work, but only if there is messagebus in
> > packaged /etc/group at all, reported here:
> > http://lists.linuxtogo.org/pipermail/openembedded-core/2011-December/014619.html
> > but probably got lost in other issues with new opkg, like this one
> > http://lists.linuxtogo.org/pipermail/openembedded-core/2011-December/014654.html
> > (if D is not needed to be set from opkg then why was offlineroot_varname.patch
> > added in first place? and I have updated this after
> > offlineroot_varname.patch removal and new version is in shr branch)
> > 
> >> I can send a patch which reenables the postinstall owner ship change if 
> >> you like.
> > 
> > Such patch is already on ML from me, now I'm using .bbappend from my
> > layer 
> > http://git.shr-project.org/git/?p=meta-smartphone.git;a=commit;h=c08dfd5f5559902f6966648f75d6c6bfd2597a5b
> 
> I'm willing to put that into meta-oe till the problem is 100% fixed.

What I really need is help in understanding what is still breaking. If
someone can show me where things are going wrong I can help (and have
done so) but "there is a problem somewhere" isn't going to help get this
fixed :(

Cheers,

Richard