https://redmine.graphics-muse.org/https://redmine.graphics-muse.org/favicon.ico?16278048512014-12-16T21:46:11ZGraphics Muse Issue TrackerPiBox - Bug #417: videofe must be run as root or it doesn't workhttps://redmine.graphics-muse.org/issues/417?journal_id=15142014-12-16T21:46:11ZHammelmjhammel@graphics-muse.org
<ul><li><strong>Status</strong> changed from <i>New</i> to <i>In Progress</i></li><li><strong>% Done</strong> changed from <i>0</i> to <i>50</i></li></ul><p>I ran various tests and finally, after using strace to run omxplayer, I found that the problem was that omxplayer couldn't open /dev/vchiq, which was root.root and 660. If I changed this device to group nobody then omxplayer worked when videofe was run as nobody.nobody.</p>
<p>So now the question is: should I change the group to nobody for vchiq, change the perms to 664 (if that works) or create a completely different user for that device? If changing it to 664 is enough, that would be the best solution. I don't know if I need write access to the device for omxplayer.</p>
<p>If that doesn't work I think just switching to group nobody would be sufficient. That can be done as part of the postinstall for the omxplayer package.</p>
<p>Note: just tried 644 @ root.root and it failed. So the group has to be nobody.</p> PiBox - Bug #417: videofe must be run as root or it doesn't workhttps://redmine.graphics-muse.org/issues/417?journal_id=15152014-12-16T22:05:05ZHammelmjhammel@graphics-muse.org
<ul><li><strong>% Done</strong> changed from <i>50</i> to <i>60</i></li></ul><p>I added the device file group change to the postinst script but that isn't sufficient. The /dev/vchiq device is remade on reboots by the kernel. So the init processing has to handle the device change.</p> PiBox - Bug #417: videofe must be run as root or it doesn't workhttps://redmine.graphics-muse.org/issues/417?journal_id=15162014-12-16T22:26:36ZHammelmjhammel@graphics-muse.org
<ul><li><strong>Status</strong> changed from <i>In Progress</i> to <i>Closed</i></li><li><strong>% Done</strong> changed from <i>60</i> to <i>100</i></li></ul><p>The fix is to change /etc/mdev.conf in the core to include the following line:</p>
<pre>
vchiq root:nobody 660
</pre>
<p>I've tested this on the target and it works. Change committed in pibox core and pushed upstream.</p>
<p>Closing issue.</p>