All of lore.kernel.org
 help / color / mirror / Atom feed
* Silly audit2allows
@ 2008-02-25 18:11 Bill Chimiak
  2008-02-25 19:06 ` Joe Nall
  2008-02-25 20:11 ` Daniel J Walsh
  0 siblings, 2 replies; 4+ messages in thread
From: Bill Chimiak @ 2008-02-25 18:11 UTC (permalink / raw)
  To: selinux

Three things:
1. If one does the audit2allow ...
	checkmodule -M -m -o mynewmodule.mod mynewmodule.te
	semodule_package -o mynewmodule.pp -m mynewmodule.mod
	semodule -i mynewmodule.pp

How does one undo that if mynewmodule.te is a stupid policy?
Doesn't the semodule make that part of the policy on every boot?

2. As a selinux wannabee and an selinux enthusiast, I want more of
my coworkers to use selinux.  They are highly resistant and usually
have selinux=0 or enforce=0 on their boot commands.
Having a list of dumb audit2allow rules would be most helpful so 
I could explain to them how to use selinux without it being too cumbersome.
I know, a lot depends on the situation, but some should make one nervous,

For example, if one saw the following:
	allow unconfined_t root_t:file { read write append create};
one should be very nervous (I would think).  

There are other suggestions that I think you all see that might 
make you all chuckle.  I would like a list of chucklers so I do not
accidentally become a comedian.

3. Are any of these potentially dangerous (my apologies if this is a stupid 
request)?
allow automount_t unlabeled_t:dir search;
allow fsdaemon_t urandom_device_t:chr_file read;
allow groupadd_t devpts_t:chr_file { read write };
allow httpd_t default_t:dir search;
allow insmod_t src_t:dir search;
allow irqbalance_t user_home_t:dir search;
allow ldconfig_t var_t:dir write;
allow pam_console_t file_t:dir read;
allow semanage_t devpts_t:chr_file { read write };
allow setfiles_t devpts_t:chr_file { read write };
allow useradd_t devpts_t:chr_file { read write };


Thank you for your time and effort.
-- 
William Chimiak
Laboratory for Telecommunications Sciences
8080 Greenmead Road
College Park, MD
240-949-2778

--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Silly audit2allows
  2008-02-25 18:11 Silly audit2allows Bill Chimiak
@ 2008-02-25 19:06 ` Joe Nall
  2008-02-25 20:11 ` Daniel J Walsh
  1 sibling, 0 replies; 4+ messages in thread
From: Joe Nall @ 2008-02-25 19:06 UTC (permalink / raw)
  To: w.chimiak; +Cc: selinux


On Feb 25, 2008, at 12:11 PM, Bill Chimiak wrote:

> Three things:
> 1. If one does the audit2allow ...
> 	checkmodule -M -m -o mynewmodule.mod mynewmodule.te
> 	semodule_package -o mynewmodule.pp -m mynewmodule.mod
> 	semodule -i mynewmodule.pp
>
> How does one undo that if mynewmodule.te is a stupid policy?

semodule -r mynewmodule

(no .pp)

joe

--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Silly audit2allows
  2008-02-25 18:11 Silly audit2allows Bill Chimiak
  2008-02-25 19:06 ` Joe Nall
@ 2008-02-25 20:11 ` Daniel J Walsh
  2008-02-25 20:50   ` Christopher J. PeBenito
  1 sibling, 1 reply; 4+ messages in thread
From: Daniel J Walsh @ 2008-02-25 20:11 UTC (permalink / raw)
  To: w.chimiak; +Cc: selinux

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Bill Chimiak wrote:
> Three things:
> 1. If one does the audit2allow ...
> 	checkmodule -M -m -o mynewmodule.mod mynewmodule.te
> 	semodule_package -o mynewmodule.pp -m mynewmodule.mod
> 	semodule -i mynewmodule.pp
> 
> How does one undo that if mynewmodule.te is a stupid policy?
> Doesn't the semodule make that part of the policy on every boot?
> 
> 2. As a selinux wannabee and an selinux enthusiast, I want more of
> my coworkers to use selinux.  They are highly resistant and usually
> have selinux=0 or enforce=0 on their boot commands.
> Having a list of dumb audit2allow rules would be most helpful so 
> I could explain to them how to use selinux without it being too cumbersome.
> I know, a lot depends on the situation, but some should make one nervous,
> 
> For example, if one saw the following:
> 	allow unconfined_t root_t:file { read write append create};
> one should be very nervous (I would think).  
> 
> There are other suggestions that I think you all see that might 
> make you all chuckle.  I would like a list of chucklers so I do not
> accidentally become a comedian.
> 
> 3. Are any of these potentially dangerous (my apologies if this is a stupid 
> request)?
> allow automount_t unlabeled_t:dir search;
automount trying to mount a file system that SELinux Policy/Kernel does
not understand.  Potentially dangerous
> allow fsdaemon_t urandom_device_t:chr_file read;
fsdaemon reading /dev/urandom - Not dangerous
> allow groupadd_t devpts_t:chr_file { read write };
groupadd read/write of a generice pty. Not dangerous, since what
groupadd can do is far more dangerous.
> allow httpd_t default_t:dir search;
http is trying to read a directory that has the default label.  Probably
need to label it httpd_sys_content_t
> allow insmod_t src_t:dir search;
modutils command searching /src.  Highly unusual.  Might be executing
the command while sitting in /src directory.  Seems apps like to getcwd
when starting up.
> allow irqbalance_t user_home_t:dir search;
irqbalance is trying to search home directories.  Not something I like
to allow, since my homedirectories contain important information like
passwords and credit card data.  Might be the same reason as insmod_t
wants to search /src

> allow ldconfig_t var_t:dir write;
Probably a labeling problem.
> allow pam_console_t file_t:dir read;
Unlabeled file, potentially major labeling problem on your system.  This
means you have a file that was created on a machine that was running
without SELinux.  Should relabel.
> allow semanage_t devpts_t:chr_file { read write };
> allow setfiles_t devpts_t:chr_file { read write };
> allow useradd_t devpts_t:chr_file { read write };
All three of these are trying to read/write pty that has generic label.
 Nothing to worry about since these domains can do much more interesting
damage.
> 
> 
> Thank you for your time and effort.

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.8 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iEYEARECAAYFAkfDINgACgkQrlYvE4MpobMCvwCeIBCVp3h6UtvLj0xiXKgZrLFj
DfQAoJmTJRZvDGmZBpGHmEMbRaJ5tjep
=Rzwe
-----END PGP SIGNATURE-----

--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.

^ permalink raw reply	[flat|nested] 4+ messages in thread

* Re: Silly audit2allows
  2008-02-25 20:11 ` Daniel J Walsh
@ 2008-02-25 20:50   ` Christopher J. PeBenito
  0 siblings, 0 replies; 4+ messages in thread
From: Christopher J. PeBenito @ 2008-02-25 20:50 UTC (permalink / raw)
  To: Daniel J Walsh; +Cc: w.chimiak, selinux

On Mon, 2008-02-25 at 15:11 -0500, Daniel J Walsh wrote:
> Bill Chimiak wrote:
> > 3. Are any of these potentially dangerous (my apologies if this is a stupid 
> > request)?

> > allow fsdaemon_t urandom_device_t:chr_file read;
> fsdaemon reading /dev/urandom - Not dangerous

Right, not dangerous.  Generally it could only be a problem if you were
concerned about the domain draining all of the entropy in order to DoS
apps that use /dev/random.

> > allow groupadd_t devpts_t:chr_file { read write };
> groupadd read/write of a generice pty. Not dangerous, since what
> groupadd can do is far more dangerous.
[...]
> > allow semanage_t devpts_t:chr_file { read write };
> > allow setfiles_t devpts_t:chr_file { read write };
> > allow useradd_t devpts_t:chr_file { read write };
> All three of these are trying to read/write pty that has generic label.
>  Nothing to worry about since these domains can do much more interesting
> damage.

That's true if you look at as the domain doing malicious things to the
terminal.  Another way to look at it would be that these privileged
domains could be influenced by malicious data they read from a user's
terminal.  So if you don't care about the latter than its not a problem.
If you're running all unconfined users, then you trust the users and it
definitely isn't a problem.

-- 
Chris PeBenito
Tresys Technology, LLC
(410) 290-1411 x150


--
This message was distributed to subscribers of the selinux mailing list.
If you no longer wish to subscribe, send mail to majordomo@tycho.nsa.gov with
the words "unsubscribe selinux" without quotes as the message.

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2008-02-25 20:52 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-02-25 18:11 Silly audit2allows Bill Chimiak
2008-02-25 19:06 ` Joe Nall
2008-02-25 20:11 ` Daniel J Walsh
2008-02-25 20:50   ` Christopher J. PeBenito

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.