Basically it makes semodule -u file.pp, install file.pp if it does not exist. This matches the rpm syntax, and allows us too update/install many packages with a transaction without know whether the package is updated or installed.
Currently we can only do a -i which could hammer a newwer version.
commit 3a5ed0fdf42200d0efd6cb1064eab91d2eb5ca52
Author: Dan Walsh <dwalsh@redhat.com>
Date: Mon Aug 24 11:36:41 2009 -0400
i Upgrade patch
Add code to semanage_direct_commit() to notice that the disable_dontaudit
flag has been changed and rebuild the policy if so.
Currently, libsemanage doesn't notice that the disable_dontaudit flag is
set so it does not rebuild the policy. semodule got around this by calling
semanage_set_rebuild() explicitly, but libsemanage should really notice
that this has changed and rebuild appropriately.
On Mon, 2009-08-24 at 10:57 -0400, Chris PeBenito wrote:
> On Mon, 2009-08-24 at 10:04 -0400, Stephen Smalley wrote:
> > On Mon, 2009-08-24 at 09:54 -0400, Chris PeBenito wrote:
> > > I took the current release of libsemanage and added the patch to add a
> > > bzip blocksize option[1]. The modules in my store were already
> > > compressed with the stock release. I put bzip-blocksize=0 in my
> > > semanage.conf and I do semodule -B and get:
> > >
> > > libsepol.module_package_read_offsets: wrong magic number for module
> > > package: expected 0xf97cff8f, got 0x39685a42 (No such file or
> > > directory).
> > > libsemanage.semanage_load_module: Error while reading from module
> > > file /etc/selinux/strict/modules/tmp/modules/apm.pp. (No such file or
> > > directory).
> > > semodule: Failed!
> > >
> > > If I do semodule -l, it will also get the magic number error. If I
> > > remove the blocksize option, it works again. I was able to reinsert all
> > > of the modules to get it working again with the blocksize 0 option.
> > >
> > > [1] http://userspace.selinuxproject.org/trac/changeset/ee9827000137fed2d3300124115fc1572acafe2f
> >
> > Yes, that's what I would expect. The expectation is that either one
> > would set that option before installing the policy for the first time,
> > or that one completely re-installs the policy after setting that option.
>
> Can we have a little better handling of this case? I don't mind
> reinstalling the policy, but the error messages aren't helpful. In
> addition, with semodule -l being broken, I have to look into the module
> store to see what modules are installed or guess.
Seems like it is just as easy to just support pre-existing compressed
modules, see below.
Explicitly probe for the bzip2 magic string prefix and fall through to
BZ2_bzReadOpen() if the module is bzipped even if bzip-blocksize=0.
Thus bzip-blocksize=0 will prevent any further compression of
subsequently installed/updated modules, but will continue to function
with existing compressed modules.
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Remove the support for hard linking files in semanage_copy_file, as it
is unsafe and can leave the active store corrupted if something goes
wrong during the transaction. It also can leave the installed policy
files with incorrect file modes or security contexts.
To do this safely, we would need to change all functions that write to
the sandbox files to first unlink the destination file. This was done
in the original patch for the write_file helper but not for other cases.
It would need to be done for all functions that open.*O_CREAT or
fopen.*w on a file in the sandbox.
We also don't want this applied to the installed policy files, as they
need to be created with appropriate file modes and security contexts
that may differ from the sandbox files. At present, the hard link
support will only affect the installed policy files when they are first
created; afterward the link() call will always fail with EEXIST since
they are not unlinked prior to installation (nor would that be safe as
it could leave the system without a policy - rename would make more
sense in that situation). If we were to re-introduce hard link support,
we ought to use different helpers or flags for installing the policy
files than for copying the active store to the temporary sandbox to
avoid affecting both.
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Allow the administrator to customize the bzip block size and "small"
flag via semanage.conf. After applying you can add entries like these
to your /etc/selinux/semanage.conf to trade off memory vs disk space
(block size) and to trade off memory vs runtime (small):
bzip-blocksize=4
bzip-small=true
You can also disable bzip compression altogether for your module store
via:
bzip-blocksize=0
The semanage.conf entries are now validated against legal value ranges
at handle creation time.
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Currently any changes made to the policy which require committing a handle cause dontaudit rules to be re-enabled. This is confusing, and frustrating for users who want to edit policy with dontaudit rules turned off. This patch allows semanage to remember the last state of the dontaudit rules and apply them as default whenever a handle is connected. Additionally other functions may check for the file semanage creates to determine if dontaudit rules are turned on. This knowledge can be useful for tools like SETroubleshoot which may want to change their behavior depending on the state of the dontaudit rules. In the event that a the file cannot be created a call to commit will fail.
Signed-off-by: Christopher Pardy <cpardy@redhat.com>
[sds: Removed duplicate from other patch and cleaned up style.]
[sds: Changed uses of semanage_fname to semanage_path.]
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
This patch adds a SWIG specification file for ruby bindings for libsemanage.
The spec file is almost identical to the python SWIG file with the exception
that all list generating typemaps have been removed and the python related
functions have been replaced with the corresponding ruby ones. Finally the
Makefile is modified to be able to build the new bindings. Something to note is
that on 64-bit systems ruby.h might be found somewhere under /usr/lib64 instead
of /usr/lib so LIBDIR=/usr/lib64 will be needed to build the ruby bindings from
source.
Below is an example using the ruby bindings and produces the similar output
to semodule -l
#!/usr/bin/ruby
require "semanage"
handle = Semanage.semanage_handle_create
Semanage.semanage_select_store(handle, "targeted", Semanage::SEMANAGE_CON_DIRECT)
Semanage.semanage_connect(handle)
module_info = Semanage.semanage_module_list(handle)
modules = Array.new()
module_info[2].times do |n|
temp_module = Semanage.semanage_module_list_nth(module_info[1], n)
mod_string = Semanage.semanage_module_get_name(temp_module).to_s + " " \
+ Semanage.semanage_module_get_version(temp_module).to_s
modules.push(mod_string)
end
puts "List of Installed Modules"
modules.each do |str|
puts str
end
Signed-off-by: David P. Quigley <dpquigl@tycho.nsa.gov>
Add a semanage_mls_enabled() interface to libsemanage so that
semanage/seobject can be rewritten to use it to test whether MLS is
enabled for a given policy store rather than checking the runtime MLS
enabled status, which can be misleading when using semanage on a
SELinux-disabled host or when using semanage on a store other than the
active one. Sample usage:
from semanage import *
handle = semanage_handle_create()
rc = semanage_connect(handle)
rc = semanage_mls_enabled(handle)
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
The current libsemanage code does not correctly add lines which include
USER in them into the homedir_template, for example:
/tmp/gconfd-USER -d system_u:object_r:ROLE_tmp_t
This line was included in the past since it has ROLE. However, with the
switch to UBAC separations, the line has changed to:
/tmp/gconfd-USER -d system_u:object_r:user_tmp_t
and is no longer included. The follwing patch fixes.
Signed-off-by: Joshua Brindle <method@manicmethod.com>
Email: dwalsh@redhat.com
Subject: libsemage patch to not compile modules for seusers and fcontext
Date: Wed, 10 Sep 2008 10:30:08 -0400
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Ivan Gyurdiev wrote:
>
>>> I'm a little unclear on what this is doing - can you clarify?
>>>
>> This is clearing the existing seusers.final file, otherwise delete was
>> not working.
>>
> I think the previous code was doing more - it was merging the local file
> with the shipped base package file, like this:
>
> data = extract_file_from_policy_package( )
> write_file ( "seusers.final", data )
> if ( data != null ) {
> seusers.clear_cache() // thereby forcing reload from
> seusers.final when cache() is called again (in merge_components)
> } else {
> seusers.clear()
> }
>
> It's also doing this three times (once for fcontexts, once for seusers,
> once for seusers_extra).
> The problem is that you're skipping the link_sandbox call, which builds
> the base package, containing this information.
>
> Ivan
>
>
Ok I found some problems with the previous patch and did some code
reuse. I added a function that only read base.pp in order to handle the
base user_extra and seusers problem.
Signed-off-by: Joshua Brindle <method@manicmethod.com>
