Hi,

I'm building a Gentoo system (x86) using the rsbac-sources kernel. I've never used RSBAC, so I figured I'd try it out. The install well exceptionally well, and I configured the kernel and compiled it according to this:

http://www.gentoo.org/proj/en/harden...quickstart.xml

I read the Introduction as per the guide's suggestion, and I enabled PAX as per the suggestion as well.

Now when I boot up (since the very first reboot...this is not something that has developed into a problem recently), I get this:

I'm posting here because the machine I'm installing on has had a few distros (including Gentoo) run on it quite successfully. Gentoo + RSBAC (or more likely, my lack of understanding and experience with RSBAC) is the combination that breaks the machine.

I've spent the past few hours trying to dig up information on Google, but the closest thread I could come up with was from kernel.org...a discussion of how this error message has been fixed on x86_64 as of 2.6.19 (again, I'm running x86).

I'm not asking anyone to step me through this, I just want to know if anyone has seen this error message before and what could cause it (RSBAC initial restrictions on filesystems? compilation issues?) (the kernel.org thread stated that it was a gcc bug that caused it on x86_64) so I know where to start...I did notice an inline function error during the kernel compile (function was inlined after declaration, I believe is what it said).

Thanks in advance...

edit - I compiled a 2.6 profile (2.6.19) (the link above covers 2.6 and 2.4)

Never seen that one before myself but I remember reading posts about it. One of the replies in that thread mentions adding boot arg "call_trace=old" could "fix" things (see http://lkml.org/lkml/2006/8/22/225). BTW this is not a Linux Security forum issue (not a security vulnerability or risk), more like something for the Linux General forum. After all more "general" than the kernel you can't get with GNU/Linux :-]. Lets move the thread after your reply.

Ahh, sorry. My bad, and thanks for the info!

I just tried adding "call_trace=old", and it gave me quite a bit more information...seems that rsbacd is the one causing the problems (as I suspected), and it appears to cr4p out immediately after initialization...I'm going to try toying with make.conf then rebuild and see if perhaps I can get this machine working...

Again, thank you for the tip!

To update this issue, I read every post in that link you sent, unSpawn, and it appears the functionality I was having difficulty with was something new (introduced around the 2.6.18 kernel?) and that thread had quite a bit of discussion about how to go about fixing it. If I wasn't on a time pinch (the server is a VoIP server for a gaming crew...haha), I'd dig in and try to figure out what was going on...but alas, as it stands, using a 2.6.14 kernel (most recent pre-2.6.18 kernel that's listed as stable in the Gentoo repository) "fixed" the problem. Thanks for pointing me in the right direction, man!

Good to see you found the fix all by yourself and thanks for updating the thread, it's definately appreciated.