All of lore.kernel.org
 help / color / mirror / Atom feed
From: pcclark@nps.edu (Paul Clark)
To: refpolicy@oss.tresys.com
Subject: [refpolicy] MLS file upgrade
Date: Fri, 3 Feb 2012 14:12:17 -0800	[thread overview]
Message-ID: <4F2C5BC1.8080602@nps.edu> (raw)
In-Reply-To: <4F2BE031.30502@tresys.com>



On 2/3/12 5:25 AM, Christopher J. PeBenito wrote:
> On 02/02/12 13:26, Paul Clark wrote:
>>
>> It now is reporting a 'relabelto' error.  I changed the 2nd MLS 'mlsconstrain' rule in 'policy/mls' that applies to 'relabelto' from
>>      ( h1 dom h2 )
>> to
>>     ( l1 domby l2)
> Thats not going to do anything in this case, as in your test, l1 == h1 and l2 == h2, since both your subject and object are single level.
>
It is my understanding that
     h1 = clearance of the process
     h2 = high level of the file/dir
     l1 = current level of the process
     l2 = low level of the file/dir
The original (h1 dom h2) test made sense in the original policy because 
it ensured that a process could not create a multilevel directory that 
was outside the clearance of the process.  For files, it was another way 
of ensuring that a process could not upgrade a file.  I changed that 
rule to (l1 domby l2) so that a process could upgrade a file, but it 
made sure that the new level of the file was >= the old level (i.e., it 
made sure it wasn't a downgrade).
> I did a quick look through the MLS constraints, and I /think/ this will achieve what you want.  It should remove all of the upgrading restrictions, except the relabelfrom one.  If you wanted to upgrade a file but your process doesn't dominate the object level (eg subject is s0 and object is s1) you'd have to remove the relabelfrom permission from the MLS constraint for all the file classes.
>
> diff --git a/policy/mls b/policy/mls
> index 0e8474b..9a8011b 100644
> --- a/policy/mls
> +++ b/policy/mls
> @@ -68,10 +68,6 @@ gen_levels(mls_num_sens,mls_num_cats)
>   mlsconstrain { file lnk_file fifo_file } { create relabelto }
>   	( l2 eq h2 );
>
> -# new file labels must be dominated by the relabeling subjects clearance
> -mlsconstrain { dir file lnk_file chr_file blk_file sock_file fifo_file } relabelto
> -	( h1 dom h2 );
> -
>   # the file "read" ops (note the check is dominance of the low level)
>   mlsconstrain { dir file lnk_file chr_file blk_file sock_file fifo_file } { read getattr execute }
>   	(( l1 dom l2 ) or
> @@ -108,12 +104,10 @@ mlsconstrain dir { add_name remove_name reparent rmdir }
>
>   # the file upgrade/downgrade rule
>   mlsvalidatetrans { dir file lnk_file chr_file blk_file sock_file fifo_file }
> -	((( l1 eq l2 ) or
> -	  (( t3 == mlsfileupgrade ) and ( l1 domby l2 )) or
> +	((( l1 domby l2 ) or
>   	  (( t3 == mlsfiledowngrade ) and ( l1 dom l2 )) or
>   	  (( t3 == mlsfiledowngrade ) and ( l1 incomp l2 ))) and
> -	 (( h1 eq h2 ) or
> -	  (( t3 == mlsfileupgrade ) and ( h1 domby h2 )) or
> +	 (( h1 domby h2 ) or
>   	  (( t3 == mlsfiledowngrade ) and ( h1 dom h2 )) or
>   	  (( t3 == mlsfiledowngrade ) and ( h1 incomp h2 ))));
>
> @@ -121,12 +115,10 @@ mlsvalidatetrans { dir file lnk_file chr_file blk_file sock_file fifo_file }
>   # has used setfscreate (note that both the high and low level of the object
>   # default to the process sensitivity level)
>   mlsconstrain { dir file lnk_file chr_file blk_file sock_file fifo_file } create
> -	((( l1 eq l2 ) or
> -	  (( t1 == mlsfileupgrade ) and ( l1 domby l2 )) or
> +	((( l1 domby l2 ) or
>   	  (( t1 == mlsfiledowngrade ) and ( l1 dom l2 )) or
>   	  (( t1 == mlsfiledowngrade ) and ( l1 incomp l2 ))) and
> -	 (( l1 eq h2 ) or
> -	  (( t1 == mlsfileupgrade ) and ( l1 domby h2 )) or
> +	 (( l1 domby h2 ) or
>   	  (( t1 == mlsfiledowngrade ) and ( l1 dom h2 )) or
>   	  (( t1 == mlsfiledowngrade ) and ( l1 incomp h2 ))));
>
Thanks for taking the time to do this.  I installed these changes, but I 
still get a 'relabelto' error.  I don't understand it.  Could it be that 
the relabelto error is coming form another module?  How can I tell?  Is 
it a problem with the type difference between the process and the file, 
and not an MLS problem?

Here is a condensed output from sealert:

node=sel13.ern.nps.edu type=AVC msg=audit(1328305466.993:55): avc:  
denied  { relabelto } for  pid=2088 comm="chcon" name="testfile" 
dev=sda1 ino=526347 scontext=user_u:user_r:user_t:s0 
tcontext=user_u:object_r:user_home_t:s1 tclass=file

node=sel13.ern.nps.edu type=SYSCALL msg=audit(1328305466.993:55): 
arch=40000003 syscall=226 success=no exit=-13 a0=86118d0 a1=71a185 
a2=8612c70 a3=1f items=0 ppid=2069 pid=2088 auid=501 uid=501 gid=501 
euid=501 suid=501 fsuid=501 egid=501 sgid=501 fsgid=501 tty=pts1 ses=1 
comm="chcon" exe="/usr/bin/chcon" subj=user_u:user_r:user_t:s0 key=(null)

      reply	other threads:[~2012-02-03 22:12 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-02-02  1:09 [refpolicy] MLS file upgrade Paul Clark
2012-02-02 15:23 ` Christopher J. PeBenito
2012-02-02 18:26   ` Paul Clark
2012-02-03 13:25     ` Christopher J. PeBenito
2012-02-03 22:12       ` Paul Clark [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4F2C5BC1.8080602@nps.edu \
    --to=pcclark@nps.edu \
    --cc=refpolicy@oss.tresys.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.