Re[6]: RSBAC v1.1.1 problem


From: Keith Matthews <keith_m@sweeney.demon.co.uk>
Subject: Re[6]: RSBAC v1.1.1 problem
Date: Wed, 4 Apr 2001 13:13:29 +0100 (BST)

Next Article (by Date): WWW.RSBAC.DE Christine Hall
Previous Article (by Date): Re: Rule Set Based Access Control (RSBAC) Stephen Smalley
Top of Thread: Re[6]: RSBAC v1.1.1 problem Keith Matthews
Next in Thread: Re: Re[6]: RSBAC v1.1.1 problem Amon Ott
Articles sorted by: [Date] [Author] [Subject]


On Tue, 3 Apr 2001 14:09:29 +0200 Amon Ott <Amon Ott <ao@rsbac.org>> wrote:

> On Die, 03 Apr 2001 Keith Matthews wrote:
>   =20
> > Having got that bit to run I still cannot deal with things. Strange
> > message are in the logs, for example it is claiming to initialise itsel=
f
> > as v 1.1.0, (I have rechecked that it was the 1.1.1 tarball) and most o=
f
> > the modules are reporting that they cannot find files (e.e. Dev ACI cou=
ld
> > not be read).
>=20
> This is really weird. I never had that. Did you possibly extract the 1.1.=
1
> tarball over an existing RSBAC 1.1.0 kernel without a make dep afterwards=
?
> Certainly not.
>=20

OK, I've tracked this one down. When I first had the short patch problem I
'reverted' back to 1.1.0 and tried that, booting once or twice. Having
realised where the problem with the patch was I rebuilt the whole kernel
from scratch using 1.1.1. However, through something I do not yet
understand, I came up with a slightly different name for the new kernel,
hence the old one was not deleted.

As I am using GRUB on the box (LILO causes VFS panics if allowed to boot
after a timeout) and had set things to point at the actual kernel images
not simple 'vmlinuz' GRUB was still booting the 1.1.0 kernel (Duh !!).
Would not have happened anyway without the name change of course.

It was only when I ran ls -l on /boot and checked times that I spotted it.

I'm still having trouble getting pcmcia to work, and it is not
permissions. Pcmcia is started correctly as part of init, but it fails as
I have version details set on modules and it claims the versions are
wrong. A bit of checking has confirmed that they are wrong (or different
at least) as the pcmcia build seems to be taking the version information
from the VERSION,PATCHLEVEL,SUBLEVEL,EXTRAVERSION variables in the
makefile, but the kernel build is appending '-rsbac' to that when
generating the kernel version for inclusion.=20

A short term fix is obviously to remove the version checking, but clearly
long-term some measure of agreement between rsbac and pcmcia is required.
Which one is actually 'wrong' I have no idea as I have no knowledge of the
standards (conventions ?) for this area.


--
Keith Matthews

Frequentous Consultants  - Linux Services,=20
=09=09Oracle development & database administration


-
To unsubscribe from the rsbac list, send a mail to
majordomo@rsbac.org with
unsubscribe rsbac
as single line in the body.

Next Article (by Date): WWW.RSBAC.DE Christine Hall
Previous Article (by Date): Re: Rule Set Based Access Control (RSBAC) Stephen Smalley
Top of Thread: Re[6]: RSBAC v1.1.1 problem Keith Matthews
Next in Thread: Re: Re[6]: RSBAC v1.1.1 problem Amon Ott
Articles sorted by: [Date] [Author] [Subject]


Go to Compuniverse LWGate Home Page.