[rsbac] fatal problems with rsbac 1.2.0 and samba

[Michal Medvecky]

Hello,

since installed rsbac 1.2.0 - 2.4.18 kernel, I'm experiencing weird problem.

I use to mount my roommate's computer drive via samba. When using kernel with
rsbac, when I try to access the smb volume, process dies, and stays in "D"
state. It happens with any process (xmms, ls, find, ..., thus it's not process-dependent)

The only log that happens is (dmesg):

Unable to handle kernel paging request at virtual address e0000000
 printing eip:
c018bbb3
*pde = 00000000
Oops: 0000
CPU:    0
EIP:    0010:[<c018bbb3>]    Not tainted
EFLAGS: 00210282
eax: 6cb80e02   ebx: 5594eb9e   ecx: fba123fd   edx: 7c24cf9a
esi: e0000000   edi: d4c1de00   ebp: d4c1de98   esp: d4c1ddb0
ds: 0018   es: 0018   ss: 0018
Process find (pid: 520, stackpage=d4c1d000)
Stack: c0140910 d4c1de68 e18a6274 00000000 00000000 00000000 00000000 d6490440 
       d9e487c0 00dce2ea 0001017a 00dce2ea 00000000 00000000 d61ec000 00000004 
       00000000 00000000 00000001 00000006 c018a445 d92fbec0 d4c1dfb0 c0140910 
Call Trace: [<c0140910>] [<c018a445>] [<c0140910>] [<c0140910>] [<c018a4dc>] 
   [<c0140910>] [<c018b3bb>] [<c0140910>] [<c0140563>] [<c0140910>] [<c0140acf>]
 
   [<c0140910>] [<c013feb0>] [<c0106c5f>] 

Code: 0f b6 06 46 89 c2 c1 e2 04 01 da c1 e8 04 01 c2 8d 04 92 8d 

My machine remains running, but with deadlocked process(es) and I can't use any other fs mounted via network.

This does not happen (= never happens) on 2.4.18 without rsbac. 

I'm using debian woody with latest updates and pre-patched kernel from amon ott's www.rsbac.org/kernels/.

Any idea what's wrong?

Thanks,

Michal Medvecky

-- 
#define GNU_MESSAGE 1
#define GNUPG_KEY_URL "http://slovakia.sh.cvut.cz/~misko/gpg.key"