| 1 |
Dominik J. Fischer wrote: |
| 2 |
|
| 3 |
> Obviously, the SELinux base policy even does not provide rules for those |
| 4 |
> devices initialized at bootup. |
| 5 |
|
| 6 |
Assuming these are *only* happening when you first boot up, |
| 7 |
and don't continue to happen, these errors are mostly harmless. |
| 8 |
|
| 9 |
They are happening during the boot process, prior to having |
| 10 |
udev up and the /dev partition populated correctly. Before |
| 11 |
that, your /dev nodes are mislabeled, so the boot scripts |
| 12 |
attempts to write to /dev/console, /dev/null, etc. are |
| 13 |
generating errors. The only device *needed* for Gentoo to |
| 14 |
boot before udev is running is /dev/null, which you can fix |
| 15 |
by doing this: |
| 16 |
|
| 17 |
# mkdir /mnt/fakeroot |
| 18 |
# mount -o bind / /fakeroot |
| 19 |
# cd /fakeroot/dev |
| 20 |
# setfiles -r /fakeroot \ |
| 21 |
/etc/selinux/strict/contexts/files/file_contexts \ |
| 22 |
. |
| 23 |
# cd / |
| 24 |
# umount /fakeroot |
| 25 |
|
| 26 |
You will probably still get a few audit messages (about |
| 27 |
/dev/console) but its nothing that is required for Gentoo to |
| 28 |
boot. |
| 29 |
|
| 30 |
--Mike |
Archives