Linux kernel 3.0 + SELinux problem
Jerry James
loganjerry at gmail.com
Wed Jun 8 22:52:16 UTC 2011
I'm having some kind of problem with SELinux on the Rawhide 3.0
kernels. The boot process gets stuck loading the SELinux policy over
and over again. I get a long series of messages like this for a few
minutes:
[ timestamp] type=1403 audit(various numbers): policy loaded
auid=4294967295 ses=4294967295
Then something times out, I think. It always scrolls by too quickly
for me to read it, but it looks like a typical stuck process kernel
backtrace. Then I get some variety, and start seeing an endless
parade of these:
[ timestamp] type=1403 audit(various numbers): policy loaded
auid=4294967295 ses=4294967295
[ timestamp] SELinux: 2048 avtab hash slots, 223865 rules.
[ timestamp] SELinux: 2048 avtab hash slots, 223865 rules.
[ timestamp] SELinux: 9 users, 13 roles, 3663 types, 193 bools, 1
sens, 1024 cats
[ timestamp] SELinux: 81 classes, 223865 rules
Well, at least I guess it's endless. I've let it go for as long as 10
minutes in the hope that something else would happen. I've tried both
kernel 3.0-0.rc1.git0.2 and kernel 3.0-0.rc2.git0.1 and have the same
problem with both. I touched /.autorelabel and rebooted with the last
2.6.39 kernel, but that didn't help. The only way I can boot these
kernels is to use "selinux=0" on the boot line.
I'm seeing this on 2 virtual machines, one x86_64 and one i686, both
with fully updated Rawhide as of today and (almost) the same set of
packages installed. Both "yum upgrade" and "package-cleanup
--orphans" show nothing to do. On both, after booting with selinux=0,
"systemctl --failed" lists 0 units. Both started life as F-14
machines, became F-15 Alpha and then F-15 Beta boxes, and were
upgraded to Rawhide after the release of F-15. It's possible some
configuration got screwed up along the way.
If anyone has a theory about what's going on, I'm all ears. Thanks,
--
Jerry James
http://www.jamezone.org/
More information about the devel
mailing list